INDEX

S. No Topic Page No
Week 1
1 Partition, Riemann intergrability and One example 1
2 Partition, Riemann intergrability and One example (Contd.) 11
3 Condition of integrability 23
4 Theorems on Riemann integrations 32
5 Examples 40
Week 2
6 Examples (Contd.) 49
7 Reduction formula 61
8 Reduction formula (Contd.) 72
9 Improper Integral 86
10 Improper Integral (Contd.) 98
Week 3
11 Improper Integral (Contd.) 108
12 Improper Integral (Contd.) 118
13 Introduction to Beta and Gamma Function 129
14 Beta and Gamma Function 140
15 Differentiation under Integral Sign 152
Week 4
16 Differentiation under Integral Sign (Contd.) 163
17 Double Integral 174
18 Double Integral over a Region E 186
19 Examples of Integral over a Region E 196
20 Change of variables in a Double Integral 207
Week 5
21 Change of order of Integration 217
22 Triple Integral 227
23 Triple Integral (Contd.) 236
24 Area of Plane Region 248
25 Area of Plane Region (Contd.) 259
Week 6
26 Rectification 270
27 Rectification (Contd.) 280
28 Surface Integral 290
29 Surface Integral (Contd.) 299
30 Surface Integral (Contd.) 309
Week 7
31 Volume Integral, Gauss Divergence Theorem 320
32 Vector Calculus 332
33 Limit, Continuity, Differentiability 342
34 Successive Differentiation 353
35 Integration of Vector Function 364
Week 8
36 Gradient of a Function 376
37 Divergence & Curl 386
38 Divergence & Curl Examples 396
39 Divergence & Curl important Identities 406
40 Level Surface Relevant Theorems 416
Week 9
41 Directional Derivative (Concept & Few Results) 426
42 Directional Derivative (Concept & Few Results) (Contd.) 436
43 Directional Derivatives, Level Surfaces 446
44 Application to Mechanics 454
45 Equation of Tangent, Unit Tangent Vector 465
Week 10
46 Unit Normal, Unit binormal, Equation of Normal Plane 476
47 Introduction and Derivation of Serret-Frenet Formula, few results 486
48 Example on binormal, normal tangent, Serret-Frenet Formula 495
49 Osculating Plane, Rectifying plane, Normal plane 504
50 Application to Mechanics, Velocity, speed , acceleration 513
Week 11
51 Angular Momentum, Newton's Law 524
52 Example on derivation of equation of motion of particle 533
53 Line Integral 543
54 Surface integral 552
55 Surface integral (Contd.) 561
Week 12
56 Green's Theorem & Example 570
57 Volume integral, Gauss theorem 581
58 Gauss divergence theorem 591
59 Stoke's Theorem 602
60 Overview of Course 610
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 01
Partition, Riemann intergrability and One example

Hello students. Today, we are going to begin our very first lecture on this Integral and
Vector Calculus. And as I have told you that we will basically start with the integral
calculus section at first, and then we will move to the vector calculus part. So, to begin
with just to give you an overview of the course, I have already done that in the introductory
video.

(Refer Slide Time: 00:49)

So, we will first start with the partition, and concepts of Riemann integral, and properties
of Riemann integrable functions and so on. So, today we will start with the partition. And
we will try to explain the concepts of Riemann integrable functions.

1
(Refer Slide Time: 01:22)

So, let us so first of all an integral calculus up until your plus 2 level, you may be familiar
with the integral of the type ∫ 𝑓(𝑥)𝑑𝑥. So, this is basically an indefinite integral. And you
𝑏
also have done something like integral from ∫𝑎 𝑓(𝑥)𝑑𝑥.. So, this is basically definite
integral. And we know that if we integrate this function, so this function 𝑓(𝑥) they most
of the time they are really nicely behaved functions. So, you can do integration or you can
do differentiation of these functions, because they are very well behaved in a way.

And one such example could be that let us say example 1. So, you have integral of 𝑥 square
𝑥3
𝑑𝑥. And if you integrate, then it is basically + 𝑐, where 𝐶 is any arbitrary constant. So,
3

you can see that the answer of this integral on the left hand side is basically a very nicely
behaved function. So, 𝑥 3 is a continuous as well as differentiable function, and 𝐶 is
basically a constant function.

So, it is always something goes in the direction of finding the derivative of this right hand
𝑥3
side. So, if we differentiate the right hand side + 𝑐, then this will be basically 𝑥 2 ,
3

because the derivative of the constant is 0, and derivative of the first term is basically 𝑥 2 .

So, it is always goes in a way that how can we find the derivative of this function 𝑥 2 ,
here. So, whatever function you have in the integral. So, these functions these 𝑓(𝑥) here,
they are called as integrand. So, whatever integrand, you have in the integral they can be

2
a differentiation of some other function. And these functions 𝑓(𝑥), they are sometimes
called as anti-derivative. So, in a way you have to find out a function, which is a derivative
of this function 𝑓(𝑥), and that will be your answer to your integral. So, this is what we did
up until our plus 2 level.

And now, we will basically look into the Riemann integral. So, those integrals were quite
simple, where you had to find out the derivative of the function 𝑓(𝑥) in the integral. In
Riemann integral, we will look at integrals in a little bit different way. And we will start
first of all with partition.

(Refer Slide Time: 04:18)

So, just to start with partition of a set so, by partition of a set, what do we mean by that? I
said to start with, we will assume let [𝑎, 𝑏] be a closed and bounded interval. So, of course
it is a closed interval and 𝑎 and 𝑏 are finite.

Then a partition of 𝑎, 𝑏 is basically a finite ordered set 𝑃, which is defined as let us say
𝑥0 , 𝑥1 , 𝑥2 up to 𝑥𝑛 of points of 𝑎, 𝑏, which means that is a is equals to 𝑥0 , 𝑥0 less than 𝑥1 ,
𝑥1 less than 𝑥2 , and so on. So, it is basically finite, and it is also ordered. So, you have 𝑥0
less than 𝑥1 less than 𝑥2 and so on. So, such kind of finite ordered set 𝑃 is basically called
as the partition of 𝑎, 𝑏 of this closed interval 𝑎, 𝑏.

For example, we can consider for example 𝑃 = 0,1,2 up to 5 is a partition of 𝑎 to 𝑏 this

1 1
is 1𝑠𝑡 example. Second example could be 𝑃 = 0, 2 , 3 so here 𝑎 and 𝑏 so for sorry here 𝑎 is

3
 1 1 1 1
0, and 𝑏 is 5. Similarly, for example 2 if we consider 𝑃 = 0, 2 , 3 , 4 , 5 , … 1 , then this is a

partition of 𝑎 to 𝑏, where 𝑎 is 0, and 𝑏 is 1. So, you can consider any type of partition for
this interval 𝑎 to 𝑏.

And the family of all the partitions of 𝑎, 𝑏 is denoted by ҏ excuse me let me erase this.
So, it is denoted by ҏ of 𝑎, 𝑏. So, this is basically the family of all partitions of the close
interval 𝑎, 𝑏. Now, if we let me go to a new page.

(Refer Slide Time: 08:57)

Now, if we assume that 𝑃 is one such partition of the closed interval [𝑎, 𝑏, where our 𝑃 is
basically consists of 𝑥0 , 𝑥1 , 𝑥2 up to 𝑥𝑛 . And such that such that 𝑎 is 𝑥0 less than 𝑥1 less
than 𝑥2 up to 𝑥𝑛 . Then this 𝑃, divides the interval [𝑎, 𝑏] into non-overlapping intervals
non-over lapping intervals 𝑥0 to 𝑥1 , 𝑥1 to 𝑥2 ,… up to 𝑥𝑛−1 to 𝑥𝑛 , where our 𝑥0 is 𝑎 and
𝑥𝑛 is 𝑏.

So, if we consider a partition 𝑃 of this interval [𝑎, 𝑏], which is basically our given closed
interval. And if this partition 𝑃 is defined by this set 𝑥0 , 𝑥1 , 𝑥2 up to 𝑥𝑛 , then this 𝑃 will
divide this closed interval into non-overlapping sub intervals. So, if you take the union of
these of these sub intervals, then in that case you will basically get your closed interval
[𝑎 𝑏].

Now, this term here non-overlapping, it means that these a sub intervals 𝑥0 to 𝑥1 , 𝑥1 to
𝑥2 , I mean they do not intersect, I mean they do not have any points common except for

4
the endpoints, so that means 𝑥0 to 𝑥1 , so where this interval 𝑥0 to 𝑥1 , ends you have
interval 𝑥1 to 𝑥2 , beginning from there. But, they do not have any other common point, so
that means these intervals they are non-overlapping.

And in other words it means that if you have your interval 𝑎 to 𝑏 let us say, then in that
case your 𝑥0 to 𝑥1 , will be here, then 𝑥1 to 𝑥2 , will be beginning from here and so on.
Then you have 𝑥𝑛−1, to 𝑥𝑛 , which is basically your point 𝑏, so that is how these sub
intervals are defined.

We can look into one example. So, one such example could be the partition 𝑃 the
1 1 1 1
partition 𝑃 = 0, 2 , 3 , 4 , 5 , … 1 . Let us say divides 0 , 1 into non-overlapping subintervals,
1 1 1 1 1 1
sub intervals given by so those sub intervals are 0, 2 , sorry 4 , 3, 2 and so is 0, 4, 3 or we
1 2 3
can have this as 0 , 4 , 4, and then that this one will be much more interesting. ah
4

1 2 3
So, let us say you have 0 , 4 , 4, 4, and then I am taking the end point as 1; just to make the

concepts, so little bit clear, instead of complicating it so 1. Then in that case the
1 1 2 2 3 3
subintervals can be given by 0 ,4 , 4 to 4, then 4 to 4 and finally 4 to 1. So, this partition 𝑃
1 2 3
which has points as 0 , 4 , 4, 4 and 1, we will divide the interval [0,1] into non-overlapping

sub intervals given by these 4 subintervals. So, this is an example of a partition for the
interval [0,1] .

Now, that we have cleared out the concept of partition, we will basically look into the
Riemann integrability, because Riemann integrability involves these concepts of partition
and sub intervals. But, before we do the Riemann integrability, we will look into what is
the upper sum and the lower sum sometimes they are also called as well Riemann upper
sum or Riemann lower sum, so to do that we will start with a new page actually.

5
(Refer Slide Time: 15:59)

So, again so our point of discussion is upper sum and lower sum. So, upper sum and lower
sum again, we will assume that let [𝑎, 𝑏] be a closed interval. So, just to give you an idea
when we say a closed interval that means, both the endpoints of that interval are included.
And when we say an open interval, then the endpoints are not included, and then you write
it open interval something like this so this is the way to write an open interval all right.

Now, in this case we will always start, we will always take a closed interval unless we are
told otherwise. Now, here we take [𝑎, 𝑏] as a closed interval. And f mapping from [𝑎, 𝑏]
to R be a bounded function in R. The when we say bounded function, it means that it is
bounded from above, it is bounded from below, otherwise we will say that f is a bounded
function from above that means, it may not be bounded from below and vice versa
basically.

Now, for example when we say that a bounded function 𝑓(𝑥) equals to let us say 𝑥 2
between the interval [1,2] is a bounded function. So, in that case the upper bound would
be 4, and the lower bound would be 1, so that is what we mean by the bounded function.
And we say let 𝑃 be a partition of f of sorry not f, but it is the partition for the closed
interval [𝑎, 𝑏] I beg your pardon. So, it is the partition of the closed interval [a, b], where
our partition 𝑃 has the points 𝑥0 , 𝑥1 and so on up to 𝑥𝑛 . And again it follows that 𝑎 =
𝑥0 < 𝑥1 < 𝑥2 < ⋯ < 𝑥𝑛−1 < 𝑥𝑛 = 𝑏.

6
So, since now f is bounded on 𝑎, 𝑏, then it must be bounded on 𝑥𝑟−1 to 𝑥𝑟 , where 𝑟 can
be 1, 2, 3 up to 𝑛. Here the idea is I mean if you have a function 𝑓, which is bounded on
the whole interval 𝑎, 𝑏, then certainly it is bounded on every subintervals. And so here
𝑥𝑟−1 , 𝑥𝑟 can for every value of our 𝑟 can be of any one of those sub intervals. And if the
function 𝑓 is bounded on the entire interval, then obviously it will be bounded on that one
small sub interval.

For example, let us say we have 𝑓 (𝑥) equals to 𝑥 2 and our closed interval [𝑎, 𝑏] is equals
to [0 ,1], so where 𝑎 is 0, and 𝑏 is 1. Then the partition 𝑃, we can consider as 0, we divide
2 3 1 2 3
it in a four interval, so , then and sorry 0, , then , then 4, and then 1 then. This
4 4 4 4

partition 𝑃 divides this interval 0,1 into non-overlapping sub intervals.

And you can see that if we pick any one of these sub intervals, so here the possible sub
1 1 2 2 3 3
intervals are 0, 4 , then we have 4 to 4, then we have to 4, and we finally have 4 to 1.
4

So, these are the four sub intervals for this closed interval [𝑎, 𝑏].

And if you consider any one of these sub intervals, you can see right away that this function
𝑓 is bounded, it is mainly because of the fact that the function 𝑓 is bounded on the whole
interval [𝑎, 𝑏]. So, it is it will be always bounded on one of these sub intervals. So, similar
condition holds here as well. And from here we can also see that if the function 𝑓 is
bounded on the entire interval 𝑎, 𝑏, then in that case you have an upper bound and a lower
bound for this interval [𝑎, 𝑏].

However, if you consider these sub intervals, then on each of these sub intervals you have
an upper bound and then you have a lower bound. So, you can see that on this interval 0,4,
1 1 2
the lower bound is 0 and the upper bound is . However, in the interval 4 ,4 your lower
16
1 1
bound is , and upper bound is . So, similarly on every one of these sub intervals, you
16 4

will get an upper bound and then you will get a lower bound.

So, now we will define basically these upper bounds and lower bounds for the whole
interval 𝑎, 𝑏 and also for these small sub intervals. So, for this function 𝑓, now we are
back to this definition,. We write let 𝑀 equals to supremum of all the 𝑥 in 𝑎, 𝑏 such that
super over all the 𝑥 supreme of the function 𝑓(𝑥) for all 𝑥 in the closed interval [𝑎, 𝑏].
And small 𝑚 is equals to infimum of the function 𝑓(𝑥) for all 𝑥 in the closed interval [𝑎, 𝑏].

7
So, this supremum and infimum are so all the upper bound and the lower bound for this
function 𝑓(𝑥). on the interval 𝑎 , 𝑏.

Similarly, now for these subintervals 𝑥𝑟−1 to 𝑥𝑟 ; we define capital 𝑀𝑟 r as the supremum
of the function 𝑓(𝑥) on the interval 𝑥𝑟−1 to 𝑥𝑟 .And 𝑚𝑟 is equals to infimum of all the 𝑥
infimum of the function 𝑓(𝑥) for all the x in 𝑥𝑟−1 to 𝑥𝑟 , where r is running from 1, 2, 3 up
to 𝑛 all right ok. This page is full, so we go to the next page.

So, from so in the previous page so here basically what we see is as I was talking in this
example 𝑓(𝑥) = 𝑥 2 . So, here in this case, we see that on every subinterval we have a lower
bound and we have an upper bound, but at the same time for this interval [0,1] the whole
interval, we have a lower bound as well as we have an upper bound.

So, now there is a relation between the lower bound and upper bound on the whole interval.
And the lower bound and upper bound on each one of these sub intervals, which means
1
the lower bound on this interval on this small sub interval 0 , 4 , the lower bound is 0.
1 2 1
However, the lower bound on these intervals to 4 , it is 16. Similarly, lower bound on
4
1
this interval is 1 by so 4 and so on, so that means, the lower bound on this whole interval

[0,1] will be either less than or equal to the lower interval on or at the lower bound on all
these sub intervals does that make sense.

So, the lower bound for this whole interval [0,1] is either equal to the lower bound on
these sub intervals in this case for the first interval, because the lower bound for the whole
1
interval is 0 and the lower bound for this is smaller sub interval 0 , is also 0. So, it is
4

equal to actually, but as we proceed along these sub intervals. We can see that the lower
bound for this whole interval 0 comma 1 is either equal or it is less than, because on all of
1 1
these sub intervals the lower bound is not 0, it is either 16 or or something like that, so
4

that means, there is a less or equal to relation between the lower bounds. And that relation
can be written as in this fashion.

8
(Refer Slide Time: 27:25)

So, we can write that relation as a small 𝑚 is lesser equal to small 𝑚𝑟 . So, the lower bound
of the function 𝑓 on the whole interval will be lesser equal to the lower bound of the
function f on all of these subintervals. And of course, the lower bound on all of these
subintervals will be lesser equal to the upper bound on all of these subintervals. So, this
from here to here it is obvious. Now, upper bound on all of these subintervals will certainly
be less than upper bound on the whole interval because, when we talk about the upper
bound, we are taking the maximum value possible for the function 𝑓(𝑥) on that interval
[0,1].

And since the interval 0,1 was divided into 4 sub intervals for that particular example. We
are taking the maximum value of the function, and it can be attained on any one of those
sub intervals. And basically, we are taking that particular value as the supremum or as the
upper bound for that function 𝑓. However, in rest of the sub intervals of course, we have
an upper bound of course, the function 𝑓 has a maximum value at some point, but that
upper bound is always lesser equal to the upper bound of the function 𝑓 on the whole
interval [0,1].

So, that is why the upper bounds on all of these sub intervals will be lesser equal to the
upper bound of the function 𝑓 on the whole interval, so that upper bound is denoted by 𝑀.
And here we have r running from 1, 2, 3 up to 𝑛. So, this is a very vital relation.

9
And this relation says that the lower bound of the function 𝑓 on the closed interval [𝑎, 𝑏]
will be lesser equal to the lower bound of the function 𝑓 on all the sub intervals, which is
lesser equal to the upper bound of the function 𝑓 on all the sub intervals, which is less than
or equal to the upper bound of the function f on the whole interval [𝑎, 𝑏] for 𝑟 running
from 1, 2, 3 up to 𝑛. And we will conclude this lecture until here. And in the next lecture,
we will begin with upper sum and lower sum, and then we will introduce the concepts of
Riemann integrablility.

10
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 02
Partition, Riemann intergrability and One example (Contd.)

(Refer Slide Time: 00:25)

Hello students. So, up until last lecture we lived into an inequality of this type, where we
had the lower bound of a function f defined on [a,b] is less or equal to the lower bound on
all the sub intervals, which is less than or equal to of course, the upper bound of the
function f on all the sub intervals which is less than or equal to the upper bound of the
function f on [a,b].

Let us name this relation. So, let us call this relation as equation 1. Now, we define the
upper sum of the function f corresponding to the partition P. As we write

𝑈(𝑃, 𝑓) = 𝑀1 (𝑥1 − 𝑥0 ) + 𝑀2 (𝑥2 − 𝑥1 )+. . . +𝑀𝑛 (𝑥𝑛 − 𝑥𝑛−1 )

So, you can write it as in the form

∑ 𝑀𝑟 (𝑥𝑟 − 𝑥𝑟−1 )
𝑟=1

11
So, this is our upper sum, similarly we can define our lower sum. The lower sum is given
by L(P,f) and you may have guessed it will be given by

𝑚1 (𝑥1 − 𝑥0 ) + 𝑚2 (𝑥2 − 𝑥1 )+. . . +𝑚𝑛 (𝑥𝑛 − 𝑥𝑛−1 )

we can use the summation concept and then it can be written as

∑ 𝑚𝑟 (𝑥𝑟 − 𝑥𝑟−1 )
𝑟=1

So, these two are basically called as the upper sum and the lower sum. Now, you can see
that each partition

𝑃 ∈ ℘[𝑎, 𝑏]

determines these two numbers U (P, f) and L (P, f) isn’t it. So, for every partition P we
will be able to get these points 𝑥0 𝑥1 𝑥2 𝑥3 up to 𝑥𝑛 and based on those points we can be
able to obtain our sub intervals 𝑥0 to 𝑥1 , 𝑥1 to 𝑥2 and so on.

And on each one of these sub intervals, we can be able to find our lower bound for the
function f and the upper bound for the function f, mainly because the function f is
bounded on [a,b]. And based on that upper bound and lower bound on all of those sub
intervals we can be able to calculate the upper sum, which is given by
𝑀1 (𝑥1 − 𝑥0 ) + 𝑀2 (𝑥2 − 𝑥1 )+. . . +𝑀𝑛 (𝑥𝑛 − 𝑥𝑛−1 )

and we can be able to calculate our lower sum, which is given by 𝑚1 (𝑥1 − 𝑥0 ) +
𝑚2 (𝑥2 − 𝑥1 )+. . . +𝑚𝑛 (𝑥𝑛 − 𝑥𝑛−1 )

So, each partition and this in this family of partitions will determine these two numbers
U (P, f) and L (P, f), you can see that they are basically numbers because upper bound
and lower bound are numbers and 𝑥1 , 𝑥2 , 𝑥3 are points on real line. So, basically the
difference is also a number. So, ultimately we are obtaining two numbers with the help
of this partition P alright.

(Refer Slide Time: 05:29)

12
So, now what we have is so from inequality 1 where is that let us go back to the previous
slides ok. So, in this slide from inequality one we have

𝑚 ≤ 𝑚𝑟 ≤ 𝑀𝑟 ≤ 𝑀

So, we will write that inequalities so,

𝑚 ≤ 𝑚𝑟 ≤ 𝑀𝑟 ≤ 𝑀

and we will multiply this whole inequality by

(𝑥𝑟 − 𝑥𝑟−1 ).

Let us multiply so, as I was saying that (𝑥𝑟 − 𝑥𝑟−1 ), they are both points on real line. So,
their difference is also a real number basically. Now, even if you have a negative interval;
that means [-1,-2] and if you divided into equal subintervals you will still get this
difference as a positive. So, (𝑥𝑟 − 𝑥𝑟−1 ) is always a positive number and that is why when
you multiplied the inequality here, this inequality sign did not change alright.

So, next we will take the summation on both sides. So, this is true for every r running from
1 to n. So, if I take summation on both sides so, this will be

𝑛 𝑛 𝑛 𝑛

∑ 𝑚(𝑥𝑟 − 𝑥𝑟−1 ) ≤ ∑ 𝑚𝑟 (𝑥𝑟 − 𝑥𝑟−1 ) ≤ ∑ 𝑀𝑟 (𝑥𝑟 − 𝑥𝑟−1 ) ≤ ∑ 𝑀(𝑥𝑟 − 𝑥𝑟−1 )

𝑟=1 𝑟=1 𝑟=1 𝑟=1

13
So, here in this sum on the left hand side here, in this sum small m is independent of r,
basically because small m is the lower bound on the whole interval [a,b]. So, it is not
relevant to each one of those sub intervals. So, you can take that m out of the sub interval
and then it will be basically

𝑛 𝑛 𝑛 𝑛

𝑚 ∑(𝑥𝑟 − 𝑥𝑟−1 ) ≤ ∑ 𝑚𝑟 (𝑥𝑟 − 𝑥𝑟−1 ) ≤ ∑ 𝑀𝑟 (𝑥𝑟 − 𝑥𝑟−1 ) ≤ 𝑀 ∑(𝑥𝑟 − 𝑥𝑟−1 )

𝑟=1 𝑟=1 𝑟=1 𝑟=1

also we take capital M out of the interval, because capital M is also independent of r.

Now, here what we are basically doing we are basically summing the difference of all
the sub intervals so; that means

(𝑥1 − 𝑥0 ) + (𝑥2 − 𝑥1 ) + (𝑥3 − 𝑥2 ) +. . .

.So, if we if we expand this summation then you will be able to notice that except
𝑥𝑛 − 𝑥0 all other points cancels out. So, if you expand the summation, we will be able to
see that and ultimately we will be left with
𝑥𝑛 − 𝑥0 . Similarly, here this term is defined as the lower sum L(P, f) if we look into the
previous slide. So, here so L(P, f) is defined by this sum here.

So, we can write it as L(P,f) and this sum is defined as U(P,f) so, U(P,f). And here it is
basically 𝑀(𝑥𝑛 − 𝑥0 )

(Refer Slide Time: 10:04)

14
And this (𝑥𝑛 − 𝑥0 ) we know that 𝑥𝑛 is our point b and 𝑥0 is our point a is less or equal to
L(P,f) is less or equal to U(P,f), which is less or equal to M(b-a).

So, here what we have is basically an inequality which is also important. So, this inequality
says that your lower sum and your upper sum, will be bounded by m(b-a). So, basically
the lower bound of the function f on [a,b] times the difference of the endpoints basically
and the upper bound is M(b-a) which means that upper bound of the function f times, the
difference of the endpoints will be an upper bound for your lower sum as well as the upper
sum.

So, this is the second inequality which we needed to establish, before we go to the Riemann
integrable functions. Now, based on this we can define two numbers so, the first number
is so, here we have two sets of real numbers. So, one is as we said every partition P
determines U(P,f) and we have L(P,f) for every partition P in that family of partition [a,b].

So, basically we have two sets of real numbers for every partition P in that family of
partitions and the supremum of this set L(P,f) such that P is in ℘[𝑎, 𝑏] is called as the
lower integral of f on [a,b]. And it is denoted by
𝑏
∫ 𝑓(𝑥)𝑑𝑥
−𝑎

we put a small minus sign here just to signify that it is a lower integral f(x)dx, or you
can shorten it by
𝑏
∫ 𝑓𝑑𝑥
−𝑎

or we can even shorten it by

𝑏
∫ 𝑓
−𝑎

So, when I write

𝑏
∫ 𝑓
−𝑎

it basically means that integral from a to b lower integral of course, f(x)dx. This is just
one of the notations to save time in a way now that we have the lower integral we can also
define the upper integral, for the upper integral lets go to the next page ok.

15
(Refer Slide Time: 14:19)

The infimum of the set of all those U(P,f) such that P is in ℘[𝑎, 𝑏] is called as the upper
integral of f on [a,b]. And we can write this upper integral and it is denoted by
−𝑏
∫ 𝑓(𝑥)𝑑𝑥
𝑎

we put a dash sign or minus sign in the upper limit f (x)dx, or we can write it as
−𝑏
∫ 𝑓𝑑𝑥
𝑎

, or to save time we can even write

−𝑏
∫ 𝑓
𝑎

So, based on the upper sum and lower sum, we can be able to define the upper integral
and the lower integral, they are also sometimes called as upper Riemann integral and
upper lower Riemann integral.

Now, that the definition of the upper integral and lower integral are given and we have
also given the notations, we can define the Riemann integrable function. So, definition a
𝑏
function f, or a bounded function f on [a,b] is said to be Riemann integrable if both ∫−𝑎 𝑓𝑑𝑥

and,

−𝑏
∫ 𝑓𝑑𝑥
𝑎

16
lets follow one notation. So, I am removing this dx here so, lower integral and upper
integral exist and they satisfy this relation
𝑏 −𝑏
∫ 𝑓𝑑𝑥 = ∫ 𝑓𝑑𝑥
−𝑎 𝑎

So, the lower integral and the upper integral has to be equal, then such functions are
called as Riemann integrable function and the common value of this is called the
Riemann integral of f on [a,b] and it is denoted by
𝑏
∫ 𝑓(𝑥)𝑑𝑥
𝑎

we can write it f (x)dx or we can shorten it like before simply by

𝑏
∫ 𝑓𝑑𝑥
𝑎

or
𝑏
∫ 𝑓
𝑎

So, in order to talk about the Riemann integrability, you see that we had to first look into
the concepts of partition that, how you get the partition of a closed interval, then from that
partition we formed non overlapping sub intervals based on those sub intervals. We could
be able to define those lower bounds and upper bounds not only for the function on the
whole interval, but also on there was sub intervals. And for every such partition you will
get a different type of non overlapping sub intervals and then you get different types of
lower bounds and upper bounds on those sub intervals.

Now, once we have those lower bounds and upper bounds, we were able to obtain the
lower sum and the upper sum. And based on those lower sum and upper sum we can be
able to obtain this lower integral and the upper integral. And if those lower integral and
upper integral are same then in that case we say that the function is Riemann integrable,
here the upper sum and the lower sum they depend heavily on what kind of partition you
are choosing.

So, for every partition P of this family of partitions, we can be able to obtain this upper
sum and lower sum. So, these are basically two sets of real numbers depending on the
partition P and based on those two real numbers, which is lower sum and upper sum we
can be able to define the lower integral and the upper integral for the function f.

17
And if those lower integrals and upper integrals are same, then in that case the function is
said to be Riemann integrable and their common value; that means, when they are equal
then that particular value is called as the Riemann integral of the function f on [a,b]. And
we use a how to say a simplified notation in a way simply

𝑏
∫ 𝑓𝑑𝑥
𝑎

So, this is just how to say a preparation to define the Riemann integral or Riemann
integrable functions on [a,b]. We will look into now an example, where we will show that
a function is Riemann integrable. Before I proceed any further I have provided you a list
of references, we can where you can look into for the details the things which I am
teaching.

Most of a time I will try to follow my own lecture notes and sort of prepared my own
lecture notes, which I will follow and I will also try to give you some examples based on
the concepts which I am teaching and hopefully that will help. So, now we will look into
one example, where we will show that a given function is Riemann integral or Riemann
integrable or not.

(Refer Slide Time: 21:36)

So, let us start with example 1 so, statement let [a,b] be a closed and bounded interval in
set of all real numbers ℝ and c is any arbitrary point or number in ℝ. And let f mapping

18
from [a,b] to R be defined as f(x)=c ∀ x in [a,b], then prove that f is Riemann integrable
alright.

So, we will look into the solution, before we start solving this problem. We first have to
just gather the ingredient what do we need to show that Riemann integrability. So, we have
to show that the lower integral is equal to the upper integral. Now, in order to obtain the
lower integral or upper integral, we first have to obtain the lower sum and the upper sum.

And before we can obtain the lower sum and upper sum, we first have to find out a partition
for [a,b] based on that we have to find out the lower bound and the upper bound for this
function f, on those sub intervals as well as on [a,b]. Here we are a little bit in luck because
the given function is constant. So, for the constant function it does not matter what kind of
partition, or what kind of points you were choosing the value would remain always the
same so; that means, the lower bound and the upper bound would always be same not only
on that closed interval, but also on those sub intervals as well. However, just to make things
clear or the concept clear, we are going to solve this example.

So, let us see first of all f is bounded on [a,b] and let us take a partition P of [a,b] as
[𝑥0 , 𝑥1 ]. . 𝑥

sorry so, {𝑥0 , 𝑥1 , . . . 𝑥𝑛 }

where
𝑎 = 𝑥0 < 𝑥1 <. . . < 𝑥𝑛 = 𝑏

and up; obviously, the sub intervals would be

[𝑥0 , 𝑥1 ], [𝑥1 , 𝑥2 ],. . . , [𝑥𝑛−1 , 𝑥𝑛 ]

And let capital M be the supremum of the function f(x) on [a,b] and small m is the
infimum of the function f(x) on [a,b]. Similarly we define capital
𝑀𝑟 as the supremum of f(x) on [𝑥𝑟−1 , 𝑥𝑟 ] . And a small
𝑚𝑟 as the infimum of the function f(x) on the interval [𝑥𝑟−1 , 𝑥𝑟 ] , by the way if I do not
write
𝑥 ∈ [𝑥𝑟−1 , 𝑥𝑟 ]

like the way, I have written here it always means that x is in [𝑥𝑟−1 , 𝑥𝑟 ] where r is
running from 1 2 3. . . n.

So, since f is a constant function its upper bound and the lower bound would always be
same, not only that it will be same on that I mean on those sub intervals.

19
(Refer Slide Time: 26:55)

So, since f is a constant function our capital M, the upper bound is basically the constant
c our lower bound on the whole interval [a,b] is again c our upper bound on each one of
those sub intervals is c and our lower bound on each one of those sub intervals is again c,
where r is running from 1 2 3 up to n.

So, based on that I can calculate my upper sum U(P,f), which is basically

∑ 𝑀𝑟 (𝑥𝑟 − 𝑥𝑟−1 )
𝑟=1

So,
𝑀𝑟 is basically c so, c will come outside of the summation because it is independent of r.
And here we will have (𝑥𝑟−1 , 𝑥𝑟 ) . Now, again this is similar to what we did earlier. So, if
you expand the summation, then you will basically have (𝑥𝑛 − 𝑥0 ) left. So, we will have
(𝑥𝑛 − 𝑥0 )

left 𝑥𝑛 is basically our b and 𝑥0 is basically our a, similarly we can calculate our lower
sum which is
𝑛

∑ 𝑚𝑟 (𝑥𝑟 − 𝑥𝑟−1 )
𝑟=1

this is
𝑛

𝑐 ∑(𝑥𝑟 − 𝑥𝑟−1 )
𝑟=1

20
Similarly this will also give
𝑐(𝑥𝑛 − 𝑥0 )

And therefore, we will obtain

𝑐(𝑏 − 𝑎) alright.

So, now let P be let us so,let us consider the set of all partition the set
℘[𝑎, 𝑏] of all partitions of [a,b], then basically it follows that, the set L(P,f), where all such
P is in ℘[𝑎, 𝑏]. And the set U(P,f) where P is in
℘[𝑎, 𝑏] , I mean they are both basically how to say the singleton set 𝑐(𝑏 − 𝑎).

Because it does not matter what kind of partition, you will choose since it is a constant
function, we will always obtain 𝑐(𝑏 − 𝑎) . And therefore, the least the how to say the
supremum or the infimum that is the least upper bound or the least or the how to say the
greatest lower bound, will be always 𝑐(𝑏 − 𝑎).

So, like we are talking about the supremum of L(P,f) and infimum of U(P,f)

(Refer Slide Time: 30:42)

If we take for this these two sets actually, it will always be so, what I am trying to say
supremum of all such L(P,f) such that P is in
℘[𝑎, 𝑏] ,will be 𝑐(𝑏 − 𝑎) and infimum of all such U(P,f) such that P is in
℘[𝑎, 𝑏] is again 𝑐(𝑏 − 𝑎) so; that means the lower integral.

So, that is the lower integral and the upper integral are same which means

21
 𝑏 −𝑏
∫ 𝑓 = ∫ 𝑓 = 𝑐(𝑏 − 𝑎)
−𝑎 𝑎

And since they have a common value which is basically 𝑐(𝑏 − 𝑎) this is the Riemann
integral of the function f on [a,b].

So, this means that 𝑐(𝑏 − 𝑎) is the Riemann integral of f on [a,b] where f is the constant
function and it is denoted or we can just leave it like that. So, for this constant function f
is equals to c we were able to show that this constant function is Riemann integrable on
[a,b] for that of course, we needed to calculate the upper sum and lower sum which was
basically 𝑐(𝑏 − 𝑎). And if we take the supremum of all such upper sum as sorry of all such
lower sum and the infimum of all such upper sums.

For all I mean how to say set of all these partitions, then you always get it has a 𝑐(𝑏 − 𝑎)
We will always get it as c times b minus a. And since the lower integral and the upper
integral is same, this common value is basically the Riemann integral of this function f on
[a,b], we will look into one more example in the next lecture. And we will conclude today’s
lecture on this example.

So, thank you for your attention.

22
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 03
Condition of integrability

Hello students. So up until last class we saw the definitions of partition of a set and we
also looked into the in how to say, Riemann integration and Riemann integrable functions,
upper limit, lower limit sum, upper limit sum, and things like that.

So, today basically we will extend these concepts of Riemann integrable functions; there
are some how to say integrability conditions for a function to be Riemann integrable and
we will also try to look into a fundamental theorem of integral calculus if time permits in
today’s lecture.

(Refer Slide Time: 01:12)

So, we in the last class we saw that a partition P which is basically given by 𝑥0 𝑥1 up to
𝑥𝑛 of a closed interval [a,b] such that 𝑎 = 𝑥0 < 𝑥1 < 𝑥2 . . . < 𝑥𝑛 = 𝑏. So, this is one such
partition of the set of the of the closed interval [a,b]

Now, for this partition P we can define the norm of this partition. So, the norm of this
partition is defined as; so the norm of P is defined by the length of the greatest interval
greatest of all the subintervals of all the sub intervals [𝑥𝑖−1 , 𝑥𝑖 ] for i running from 1 2 3 up

23
to n so; that means, norm of P, so norm is usually denoted by ||𝑃|| how to say bar notation,
so norm of P is basically the greatest. So, we can write as

. ||𝑃|| = max(𝑥1 − 𝑥0, 𝑥2 − 𝑥1 , . . . , 𝑥𝑖 − 𝑥𝑖−1 , . . . , 𝑥𝑛 − 𝑥𝑛−1 )

= max (𝑥𝑖 − 𝑥𝑖−1)

1≤𝑖≤𝑛

So, this is the way we define the norm of this partition P, you can also use a notation
something like 𝜇(𝑃). So, we can write it as 𝜇(𝑃) and the norm of the partition P you can
be denoted by this symbol 𝜇(𝑃) as well

So, now that the definition of norm of this partition is cleared, we will look into 1 or 2
integrability condition or to be very precise Riemann integrability condition in this case.
So, let me start the new page.

(Refer Slide Time: 04:05)

So, the first one is: so Riemann integral as the limit of a sum. So, this particular definition
or the representation of a Riemann integral will give motivation to the definite integral.
So, we will see that what do we mean by it. So, the statement goes like this, so we can
write a small theorem.

So, the theorem says that if f is a bounded and integrable function over [a,b] then for every
𝜖 positive there corresponds or there exists a 𝛿 positive such that for every partition 𝑃 =
{𝑎 = 𝑥0 , 𝑥1 , . . . 𝑥𝑛 = 𝑏} of norm ||𝑃|| ≤ 𝛿; and for every choice of 𝜉𝑟 ∈ [𝑥𝑟−1, 𝑥𝑟 ]. So, we

24
choose this point 𝜉𝑟 in any r-th interval [𝑥𝑟−1, 𝑥𝑟 ] . We have, what do we have? So, we
have

𝑛 𝑏
│ ∑ 𝑓(𝜉𝑟 )(𝑥𝑟 − 𝑥𝑟−1 ) − ∫ 𝑓(𝑥)𝑑𝑥│ < 𝜖
𝑟=1 𝑎

So, what does this theorem actually saying? So, this theorem says that, if we have a
bounded function f which is also integrable, Riemann integrable over the interval [a,b]
then for every 𝜖 positive. So, this 𝜖 is an arbitrary chosen small positive number there
corresponds a 𝛿 > 0 such that for every partition P whose norm ||𝑃|| < 𝛿 we can find a
point 𝜉𝑟 in the r-th interval such that this value is less than 𝜖.

So, you see 𝜖 is an arbitrary chosen positive number. So, this integral is nothing but it can
be expressed as ∑𝑛𝑟=1 𝑓(𝜉𝑟 )(𝑥𝑟 − 𝑥𝑟−1 ) . So, this integral here it can be represented
as ∑𝑛𝑟=1 𝑓(𝜉𝑟 )(𝑥𝑟 − 𝑥𝑟−1 ) . The proof of this theorem is a little bit long. So, we will not
get into the proof of this of such theorems since, it is a course meant for BSc and BTech
students. So, we will not get into the proof, but the flavor of this theorem, so what does it
mean I will try to explain that.

(Refer Slide Time: 08:53)

So, let us look into this some. So, let us write; so let us look into this some. So, let us write,
let us say S(P,f) So, sum over the partition P for the function f we write it as
∑𝑛𝑟=1 𝑓(𝜉𝑟 )(𝑥𝑟 − 𝑥𝑟−1 ) So, this sum where this 𝜉𝑟 is any arbitrary point between [𝑥𝑟−1, 𝑥𝑟 ]

25
. So, the sum basically this sum S(P,f) is called the Riemann sum I beg your pardon. So,
this sum is called the Riemann sum of f over [a,b] with respect to the partition P of [a,b].

Now, the above theorem can be seen in this following way that if a function f, or in other
words it can be seen as if a function f is bounded and integrable over the interval [a,b],
𝑏
then what we have is we have integral sorry; what we have is we have ∫𝑎 . beg your pardon,
f (x)dx is basically lim .or the norm of P going to 0 such that this limit S P f is actually the
𝜇(𝑃)

value of that definite integral.

So, if you make the how to say ||𝑃|| → 0; that means, ||𝑃|| or the how to say, the length
of those intervals when we say that it is going to 0; that means, you are actually taking the
number of subintervals as very small. And then in that case if you look at this figure, let
us say this is my f (x) and that is x=a and that is x equal x=b. So, doing this definite integral
is basically we are calculating this area. if we are doing integral from a to b.

Now, if I take ||𝑃|| norm of partition of P going to 0; that means, I am basically making
the number of subintervals really small and small. That means, this n here these number
of subintervals are basically I mean getting larger and larger. So that means, I am taking
how to say many large,a really large number of subintervals and if you take really how to
say if you take your 𝑥1 here then, in that case what I am doing is the this area here., This
area here it is basically if you take 𝑥1 , if you take a really large number of subintervals
then in that case these small areas are actually converting into a rectangle. And here in this
formula here 𝑓(𝜉𝑟 )(𝑥𝑟 − 𝑥𝑟−1 ) this is basically if you forget about the summation then
this is basically the area of 1 such rectangle.

𝑏
So, you are actually dividing integral ∫𝑎 𝑓(𝑥)𝑑𝑥 into some of the areas of really how to
say large number of rectangles. And if you sum all those areas of rectangles then that will
𝑏
actually give you the ∫𝑎 .. So, that when you make 𝜇(𝑃) goes to 0 you are actually making
n tends to infinity. We are actually making here n tends to infinity and if you make n tends
to infinity then basically you are increasing the number of subintervals. And then in that
case this in on every sub interval you are basically calculating the area of a rectangle. And
if you sum all those areas and that will give you the area of the curve of the function f,
between the points x=a to x=b. And that is what we mean by this integral on the right hand
side.

26
So, if you make 𝑛 → ∞ of this Riemannian sum then that will actually give you the definite
𝑏
integral or the integral of the function ∫𝑎 𝑓(𝑥)𝑑𝑥 and this is the how to say the vital point
which we wanted to make that how Riemann sum how this sum here is related to the
definite integral. So, the proof of the theorem is a little bit complicated, but you can be
able to find the proof in any standard integral calculus book where they have shown the
proof with the help of upper sum and lower sum it is not that difficult and you can be able
to go through that proof. However, for our lecture we would try not to get into those details,
but just to give you an idea how Riemann sum is in related to definite integral that is what
I was trying to do.

Now, that is we know how Riemann sum is related to definite integral we will jump to a
new or how to say a new result where we look into the condition of integrability.

(Refer Slide Time: 16:01)

So, condition of Riemann integrability, so we can formulate this in terms of another small
result. So, theorem 2 and it says that; a necessary and sufficient condition for the Riemann
integrability of a bounded function f is that for every 𝜖 > 0 there exists 𝛿 > 0 such that,
for every partition P of [a,b] with ||𝑃|| ≤ 𝛿 we have U(P,f) which is our upper sum minus
L(P,f) which is our lower sum is less than 𝜖.

So, that means if you are given a bounded function f and if you would like to see whether
that function is Riemann integrable or not then we have to calculate the upper sum and the
lower sum. In the last class we have seen that how we calculate the upper sum and lower

27
sum and if the difference of the upper sum and lower sum is less than this chosen 𝜖, then
in that case that particular function is said to be Riemann integrable. This makes sense
because the upper sum U(P,f) if you take how to say the infimum of all the U (P,f) over
the partition P, then in that case that will give you the upper integral sum. And if you take
the supremum of all such L(P,f) such that now for this partition P then that will give you
the lower integral sum and the difference basically we say that the function is Riemann
integrable when the lower integral sum and upper lower integral sum and upper integral
sum are same.

So, basically if you say that the difference is less than 𝜖. That means, the function is
Riemann integrable and your lower integral and upper integral are same. So, in a way this
condition actually makes sense and based on which we can say that the given function is
Riemann integrable or not. Next we have is several properties of Riemann integrable
function so for example we know that if a function is bounded. And then if it satisfies this
condition 𝑈(𝑃, 𝑓) − 𝐿(𝑃, 𝑓) < 𝜖 then it is Riemann integrable, but other properties the
function needs to have if you want to say that whether the function is Riemann integrable
or not.

So, we will look into a few properties and how to say some results related to Riemann
integrable function.

(Refer Slide Time: 20:31)

28
So, let us go to the first property. So, I can write it as properties of Riemann integrable
functions. So, the first property is or you can write it in terms of as theorem. So, the first
property or the first theorem in this regard is every continuous function continue sorry
every continuous function is Riemann integrable.

So, this makes sense because every continuous function is always bounded and if the
function is bounded then all we have to show that whether 𝑈(𝑃, 𝑓) − 𝐿(𝑃, 𝑓) < 𝜖 or not
and if it does then in that case the function is Riemann integrable. The proof is a little bit
how to say well it is not long it is just that for the proof you can look into any book here I
am just how to say listing some of the important properties of Riemann integrable
functions. For the proofs I would recommend to look into the references which I have
shown you in the first class and there you can be able to find the proofs of these theorems.

The second property is if a function f is monotonic on [a,b] then it is Riemann integrable.

So, I am hoping that all of you know what does a monotonic function mean. So, a
monotonic function f a function f is said to be said to be monotonic. So, a function can be
monotonically increasing or monotonically decreasing or it can be strictly increasing or
strictly decreasing. So, here monotonically increasing on [a,b] if for if 𝑓(𝑥1 ) ≥ 𝑓(𝑥2 ) for
𝑥1 ≥ 𝑥2 where 𝑥1 and 𝑥2 are both 2 points in [a,b] monotonically increasing and similarly
we can define the monotonically decreasing part.

So, a function is said to be monotonically decreasing mono sorry so it is said to be

monotonically decreasing if 𝑓(𝑥1 ) ≤ 𝑓(𝑥2 ) for 𝑥1 ≥ 𝑥2 where 𝑥1 and 𝑥2 are 2 points in
this interval [a,b]. So, here we can see that if a function is monotonic. So, let us assume
that it is monotonically increasing on [a,b], then in that case of course, it is bounded. And
if it is bounded then all we have to show is that whether 𝑈(𝑃, 𝑓) − 𝐿(𝑃, 𝑓) < 𝜖 or not
and then that will ensure the Riemann integrability.

Similarly we can assume that the function is monotonically decreasing and if it is

monotonically decreasing, then of course in that case also it will be bounded. And we again
try to show that 𝑈(𝑃, 𝑓) − 𝐿(𝑃, 𝑓) < 𝜖 and then the function will become Riemann
integrable. So, the proof of this theorem can also be found in those in those books which I
have listed as reference next property is I can list it as let us say theorem 3.

29
(Refer Slide Time: 25:26)

So, if f is a Riemann integrable function on [a,b] and c is any arbitrary point between the
interval [a,b] then f is Riemann integrable on [a,c] and [c,b] and we can write the integral
𝑏
∫𝑎 𝑓(𝑥)𝑑𝑥 as the sum of these 2 integrals. So that means, you can take as many points as
you please in your closed interval [a,b] and the sum of those integrals on each one of those
sub intervals will be the integral of the how to say whole of the function f on the whole
interval [a,b]. So, here not only c you can take as many points as possible in between that
interval [a,b] and the result would still be true. Next theorem we have or the property we
have is if a function f is Riemann integrable on the closed interval [a,b] then 𝑓 2 is also
Riemann integrable.

So, if you have and similarly you can continue, so if 𝑓 2 is Riemann integrable and if then
in that case you can be able to show that 𝑓 3 or 𝑓 4 they are also Riemann integrable so that
means, this property carries forward. And again for the proof of this theorem I would
recommend you to look into the references which I have mentioned.

Next, as we have seen that for the Riemann integrability, how we can connect the Riemann
integrability with Riemannian sum and how it is connected with the definite integral. We
also saw that the necessary and sufficient condition for a function to be Riemann integrable
and some properties associated with Riemann integrable function.

So, we will stop this lecture at here. And in the next class we will again continue with the
properties of Riemann integrable function.

30
Thank you.

31
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 04
Theorems on Riemann integration

Hello students, so up until last class we saw the Riemann integrable functions, and how
they are connected with the Riemannian sum, and how Riemannian sum is connected with
definite integral. And we also looked into some properties of Riemann integrable function.
Today, we will list few more properties of Riemann integrable functions, and then we go
to fundamental theorem of integral calculus, and what do we mean by anti derivatives and
primitive.

(Refer Slide Time: 00:52)

So, as I was mentioning about the properties, so another property, property 5 of Riemann
integrability is if f is a Riemann integrable function, then |𝑓| is also a Riemann integrable
𝑏 𝑏
function. And we have |∫𝑎 𝑓(𝑥)𝑑𝑥| ≤ ∫𝑎 |𝑓(𝑥)|𝑑𝑥 . So, and so that means the mod of
function f is also Riemann integrable, and it satisfies this inequality.

Third sorry the 6th property in this regard is its little bit how to say towards the direction
of what we have already studied in limit and limit continuity and differentiability of two
functions. So, we know that if f and g are both how to say continuous, then sum of the

32
continuous functions would also be continuous or how to say if you are calculating the
limit then in that case if lim 𝑓 and lim 𝑔 exist than the sum of their limits or difference of
. .

their limits or product of their limits also exist.

So, here in this case also we have the similar theorem. So, if f and g are two Riemann
integrable functions, then sum of these two functions is also Riemann integrable. The
difference of these two functions would also be Riemann integrable. And the product of
these two functions would also be Riemann integrable. For the quotient you need to have
how to say the nonzero condition on the function g that g has to be nonzero throughout
that interval [a,b], and then we can talk about this Riemann integrability.

So, it is pretty much along the same line of limit and continuity and how to say properties
of sum of two functions or difference of two functions and things like that. And again for
the proof of this theorem I would recommend you to look into the books which I
recommended. Next in Riemann integrable functions, we have something called let me go
to a new page.

(Refer Slide Time: 04:12)

The primitive, we sometimes also call it as anti-derivative. So, the definition is if f is a

differentiable function, such that 𝐹 ′ = 𝑓 on the closed interval [a,b], then the capital
function F is called the anti derivative or the primitive whichever you prefer primitive of
the function f. And sometimes we write it, and this function capital F is also called the
integral function of f.

33
So, what do we mean by this? So, it means that we can write the function f, this small f,
let me use a different notation. So, we can write this function small f as let us say
𝑥
∫𝑎 𝑓(𝑡)𝑑𝑡. So, the capital F(x) can be expressed as the integral of the function small f. And
if you take the derivative on both sides, then in that case the right hands that are the left
hand side will be 𝐹 ′ (x) which is equals to the function f(x); and that capital F based on this
definition is called as the anti-derivative or the integral function of this small function f.

And if you have, but you see that in order to write this expression, we need to have some
special conditions on this function small f and capital F. We cannot just simply write every
function capital F as this integral form. So, the very first criteria in this regard is every
continuous function. So, every continuous function f possesses a primitive that means, if f
is a continuous function, then it can be written as this integral form. So, the very first
criteria we need to have for the function small f is that it needs to be continuous. And then
you can be able to write the capital F as the integral form of this small f.

(Refer Slide Time: 08:45)

Now, we can put this in a small theorem. So, the theorem goes like this. So, the first
theorem is, if a function f is bounded and integrable on [a,b], then the function F defined
𝑥
as capital 𝐹(𝑥) = ∫𝑎 𝑓(𝑡)𝑑𝑡 is continuous on [a,b]. And if further if f is continuous on the
closed interval [a,b], then capital F is the primitive or you can write 𝐹 ′ (𝑥) = 𝑓(𝑥).So, if
you have the bounded function f small f, then in that case this continuity criteria is pretty
much clear because in order to show the continuous function you take F(x) at a certain

34
 𝑥 𝑐
point you take F(x)-F(c)=∫𝑎 𝑓(𝑡)𝑑𝑡 − ∫𝑎 𝑓(𝑡)𝑑𝑡 and from there f(t) is bounded. So, it will
be pretty much obvious to so the continuity of the function F, a capital F on this closed
interval [a,b]. However, if your small function f is continuous then in that case your capital
F is actually the anti-derivative or the primitive of this small f(x). And from this theorem
in a way or from this definition in a way, the first the fundamental theorem of integral
calculus is motivated.

So, from here we can state our very important theorem of Riemann integration which is
basically fundamental theorem of integral calculus.

(Refer Slide Time: 12:36)

So, fundamental theorem of integral calculus says that let us say I write it as theorem. So,
I can write it in terms of theorem. The theorem says that a function f is bounded and
integrable on the closed interval [a,b], and there exists a function F such that 𝐹 ′ (𝑥) = 𝑓(𝑥)
𝑏
on the closed interval [a,b], then we can be able to write ∫𝑎 𝑓(𝑥)𝑑𝑥 = 𝐹(𝑏) − 𝐹(𝑎). So,
that means, if we have a function small f(x) which is basically bounded and Riemann
integrable on the closed interval [a,b], and if we have 𝐹 ′ (𝑥) = 𝑓(𝑥) then in that case
𝑏
∫𝑎 𝑓(𝑥)𝑑𝑥 = 𝐹(𝑏) − 𝐹(𝑎)

So, this is also in a way our Newtonian integral. So, Newtonian integral is also about
finding an anti-derivative or finding how does a primitive of the function, so that you can
be able to write it as a how to say integral of that function. So, what I mean is, so if we

35
 2
have something like 𝐼 = ∫1 𝑥 5 𝑑𝑥, then this is basically my small f(x). And I need to find
a function capital F such that 𝐹 ′ (𝑥) = 𝑓(𝑥) . So, this is all about finding an anti-derivative.
And if we can be able to find a 𝐹 ′ (𝑥) of this type such that 𝐹 ′ (𝑥) = 𝑓(𝑥), then in that
case the integral would be nothing but the difference of that capital F at the point 2 minus
the value of that capital F at the point 1.

𝑥6
So, in this case, it is very easy to find out this 𝐹 ′ (𝑥) . So, here our 𝐹 ′ (𝑥) would be .
6

So, this is our 𝐹 ′ (𝑥) . And if I take the derivative of that 𝐹 ′ (𝑥), then we will obtain 𝑥 5
and this 𝑥 5 is our small f (x) here and you see this is our capital F(x). And if I this is our
26 16
capital F(x), then it is basically − and that will be the answer.
6 6

So, based on this how to say fundamental theorem of integral calculus, we can see that
how a function and its anti-derivative are connected. And it is how to say one of the
important theorems in integral calculus. And it is also very interesting, and I hope you
were able to understand this. I would also try to give you some examples. So, first let me
state that two last theorems of this Riemann integrable functions before we jump into any
example.

(Refer Slide Time: 16:08)

So, the first theorem is; the first theorem which I am talking about is first mean value
theorem. So, the first mean value theorem says that if f and g are so it is more or less like
the mean value theorem which we studied in differential calculus.

36
So, in the differential calculus also you have a how to say a function differentiable function
on an open interval (a,b) which is continuous on the closed interval [a,b], then in that case
𝑓(𝑏)−𝑓(𝑎)
you can be able to find a point c in between the interval (a,b) such that 𝑓 ′ (𝑐) = .
𝑏−𝑎

So, here it is pretty much the same. So, we say that f and g are two Riemann integrable
functions on the closed interval [a,b], and g keeps the same sign over [a,b], then there
exists 𝜇 lying between the bounds of f such that the product of the integration
𝑏 𝑏
∫𝑎 𝑓(𝑥)𝑔(𝑥)𝑑𝑥 = 𝜇 ∫𝑎 𝑔(𝑥)𝑑𝑥.

So, this is the first mean value theorem we says that if you have two Riemann integrable
functions f and g, and the function g keeps the same sign over the interval [a,b], then you
have a value basically 𝜇 which is lying in between the upper bound and the lower bound
of the function f. So, the function f is assumed to be bounded because it is Riemann
integrable and this point 𝜇 can be in between anywhere in of the lower bound and upper
𝑏 𝑏
bound, then the sum of then the product of the integral ∫𝑎 𝑓(𝑥)𝑔(𝑥)𝑑𝑥 = 𝜇 ∫𝑎 𝑔(𝑥)𝑑𝑥.
And this is basically our first mean value theorem.

So, here you can see that instead of calculating the product of the integral, we have
replaced the function value of the function f with its value 𝜇. So, this 𝜇 is attained
somewhere. So, 𝜇 is attained at any point on this interval [a,b]. So, it can be at I do not
know 𝑥 = 𝑥2 or 𝑥3 any such point in the closed interval [a,b] and that value 𝜇 is basically
𝑏
multiplied with the integral ∫𝑎 𝑔(𝑥)𝑑𝑥 and that will actually give us the value of the
product of the integral on the left hand side.

So, this is in a way a mean value theorem. So, we are taking any value of the function f
between the upper bound and lower bound. The next theorem in this regard is the second
mean value theorem, which states that,So, the second mean value theorem says that if
𝑏 𝑏
∫𝑎 𝑓(𝑥)𝑑𝑥 and ∫𝑎 𝑔(𝑥)𝑑𝑥 both exists, and f is monotonic on the interval [a,b], then there
𝑏 𝑐
exists a c in the closed interval [a,b] such that ∫𝑎 𝑓(𝑥)𝑔(𝑥)𝑑𝑥 = 𝑓(𝑎) ∫𝑎 𝑔(𝑥)𝑑𝑥 +
𝑏
𝑓(𝑏) ∫𝑐 𝑔(𝑥)𝑑𝑥.

37
(Refer Slide Time: 21:44)

So, this is an another interesting theorem. So, we have both Riemann integrable function
f and g so that these two definite integral exist. And the function f is actually monotonic
on this closed interval [a,b], then in that case we have a point c lying between [a,b]. So, all
we need is the monotonicity of the function f, then we can be able to find a function c in
𝑏
this closed interval [a,b] such that the product of the integral here ∫𝑎 𝑓(𝑥)𝑔(𝑥)𝑑𝑥 =
𝑐 𝑏
𝑓(𝑎) ∫𝑎 𝑔(𝑥)𝑑𝑥 + 𝑓(𝑏) ∫𝑐 𝑔(𝑥)𝑑𝑥 .

So; that means, in order to evaluate the product of the integral, we do not need to evaluate
the product all we have to find the value of the function f at the endpoints. And then just
𝑐 𝑏
evaluating the integral ∫𝑎 𝑔(𝑥)𝑑𝑥 and ∫𝑐 𝑔(𝑥)𝑑𝑥 , we can be able to find the value of this
definite integral on the left hand side. So, this is an another how to say important theorem
in this regard. So, to the syllabus which I gave you earlier, and there we in the first lecture
we were sort of planning to cover the partition, definition of Riemann integrable functions
and some properties and then mean value theorems.

And so far we have covered the theoretical part. And since it is a very extensive topic I
would recommend you to look into the books of (Refer Time: 23:48), S. K Mapa, also
Santhinarayan for to know these things in detail. I will also try to work out two or three
examples probably just to give you an idea and that what do we, how does it mean by these
integrability conditions, and the app how to say one or two examples on fundamental
theorem of integral calculus.

38
Also most of my lectures I will since I prefer to write it here and work out the examples or
theorems here, but some of my lectures would also include a PowerPoint presentation, and
I will try to make things clear in those a Power Point presentations as well. So, it will be a
mixture of both this writing on the board as well as a PPT. And in the next lecture we will
look into a few examples related to Riemann integrability and fundamental theorem of
integral calculus. And today, we will stop our lecture here.

Thank you.

39
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 05
Examples

Hello students, so welcome to this class. So, up until last lecture, we looked into Riemann
Integrable functions and their definition and several properties associated with a Riemann
integrable function. For example, if your function is continuous, then it is Riemann
integrable; if it is monotonic, whether it is monotonic increasing or decreasing, then in that
case also it was Riemann integrable on the given interval.

We also looked into the definition of first of the fundamental theorem of integral calculus,
we also looked into the definition or the statement of first mean value theorem and the
second mean value theorem. The theory of Riemann integrable function is anyways too
extensive, but I try to compressed it in a you know how to say shorter format.

Now, based on the theories which are the theorems which we learned in previous classes,
today, we are going to work out few examples. And we will see that how we can be able
to show that whether a given function is Riemann integrable or not, how we can calculate
upper integral sum and lower integral sum of a given function.

(Refer Slide Time: 01:59)

40
So, today’s lecture will be all about the examples on Riemann integration. So, let me start
1
all right. So, our 1st example is; for the integral ∫0 𝑥𝑑𝑥 , we will calculate so for the
integral this, find the upper and lower integral sum, lower integral sum on [0,1] by dividing
it into 3 sub intervals. So, that means, our partition will have four points and those four
points, so we will constitute our how to say three subintervals.

Now, it is obvious that if you divide the interval [0 ,1] into 3 sub intervals, then your
1 2 3
partition points will be 0, 3, 3 and , which is 1. So, let P be the partition of [0,1], such
3
1 2 3
that our partition P will be 0, 3 , 3 and3 , which is 1 such that and so this is basically our
1 1 1 2
partition P. And the sub intervals will be [0, 3]. So, the sub intervals are; [0, 3],[3 , 3] and
2
[3 , 1].

Now, if we look into the definition of our upper integral sum and lower integral sum, which
are basically U(P,f), and L(P,f). So, upper integral sum U(P,f) is given by a formula of this
type. Since in our case n is the number of subintervals and in our case n is basically 3,
because we have 3 sub intervals. So, ∑3𝑟=1 𝑀𝑟 (𝑥𝑟 − 𝑥𝑟−1 ) This is our upper integral sum.

And our lower integral sum L(P,f) will be ∑𝑛𝑟=1 𝑚𝑟 (𝑥𝑟 − 𝑥𝑟−1 ) , so in this case n is 3. So,
we will substitute n=3 𝑚𝑟 (𝑥𝑟 − 𝑥𝑟−1 ). Note that 𝑀𝑟 and 𝑚𝑟 are the supremum and
infimum or upper bound and lower bound of the function f, which is x here, in these
individual sub intervals. So, let us name this equation as equation 1 and let us call this
equation as equation 2.

41
(Refer Slide Time: 06:21)

So, next we have to calculate the upper bound in each one of these sub intervals. So, we
1
have 𝑓(𝑥) = 𝑥 and from here; our capital 𝑀1 is basically 𝑓(3), because 𝑓(𝑥) = 𝑥 is a

monotonically increasing function in the interval [0,1].

1
And therefore, the maximum value of this function 𝑓(𝑥) would be attend at the point ,
3

because at all the other points at x=0, it the function will attain it is minimum value and so
on. So, the maximum value of this function 𝑓(𝑥) = 𝑥 will be attend at the end point 𝑥 =
1 1 1
. So, what will be the maximum value or the upper bound of this function at 3 , it is 3 .
3

2
Next, we can calculate 𝑀2 . 𝑀2 will be again attained at this end point 3. So, it will be
2
attained at 3 , because we have the function is a monotonically increasing and the points

are increasing again, so in this case the maximum value or the upper bound will be attained
2 2 2
at this end point. So, we can calculate the maximum value as 𝑓(3) and 𝑓(3) is again 3.

And 𝑀3 is the maximum value of the function f, which is again attained that x=1. So, at
𝑓(1), so this will be 1.

Similarly, we can calculate 𝑚1 . So, we can write similarly 𝑚1 equals to the infimum of
the value at x=0, because in this interval the infimum or the lower bound is attained at x=0.
So, we can write f (0) ;f (0) is 0.

42
Next, we can calculate 𝑚2 , which is the infimum or the lower bound. So, this is again
attained at let us go back to this slide. So, the lower bound of this function f will be attained
1 1
in this interval at the point 𝑥 = 3. So, we can calculate the lower bound at 𝑥 = 3, which is
1 2
. And that and 𝑚3 , so 𝑚3 is the lower bound of this function f in the interval [3 , 1]. So,
3
2 2
this will be attained at 𝑓 = 3, which is basically 3.

So, now that we have all these ingredients, which we have to calculate 𝑥1 − 𝑥0 . So, 𝑥1 is
1 1 2 1 1
basically 3 − 0, which is3 . Similarly, 𝑥2 − 𝑥1 will be 3 − 3, so this is again 3. And 𝑥3 − 𝑥2
2 1
will be 1 − 3, so this is basically 3.

So, since we have divided the interval into three equal parts. These are basically the length
of every sub interval and it is pretty much clear that if you have divided and interval into
three equal parts, then in that case the length of each one of these sub intervals would be
1 1 1
. Because, your interval is [0,1], so that is what we are getting, 𝑥1 − 𝑥0 = 3 , 𝑥2 − 𝑥1 = 3
3
1
and 𝑥3 − 𝑥2 = 3.

So, now we are ready to calculate our upper integral sum. So, upper integral sum
is ∑3𝑟=1 𝑀𝑟 (𝑥𝑟 − 𝑥𝑟−1 ) . So, if we substitute the above values, then it will be 𝑀1 . So, first
of all we can expand this summation 𝑀1 (𝑥1 − 𝑥0 ) + 𝑀2 (𝑥2 − 𝑥1 ) + 𝑀3 (𝑥3 − 𝑥2 ) is it
clear.

1 1 2 1
So, now 𝑀1 is 3 times (𝑥1 − 𝑥0 ) is also 3, 𝑀2 is 3 , (𝑥2 − 𝑥1 )is again 3, 𝑀3 is 1 and this 1
1 1 1 2 2
is again 3. So, we take 3 common, this one is 3 + 3 + 1, and this will be ultimately 3.

And similarly, we can calculate L(P, f) which is our lower integral sum. So, this can be
given as ∑𝑛𝑟=1 𝑚𝑟 (𝑥𝑟 − 𝑥𝑟−1 ) , we expand in the similar fashion 𝑚1 (𝑥1 − 𝑥0 ) +
1 1
𝑚2 (𝑥2 − 𝑥1 ) + 𝑚3 (𝑥3 − 𝑥2 ) . So, our 𝑚1 is 0 and (𝑥1 − 𝑥0 ) is 3+ 𝑚2 is 3 and (𝑥2 − 𝑥1 )is
1 2 1 1 1 2
again 3 , then 𝑚3 is 3 times 3. So, we can take 3 common, this is 3 + 3 , so ultimately we
1 1 2
will get3 . So, this is our upper integral sum 𝐿(𝑃, 𝑓) = 3, and 𝑈(𝑃, 𝑓) = 3 are our

respective lower integral sum and upper integral sum, which we wanted to calculate.

43
So, this is how we can calculate the upper integral sum and lower integral sum, it is we
have seen several times that all we have to do is to find a partition of the given closed
interval and you need to calculate the upper bound and lower bound of the function in
those sub intervals and then put it in this formula. So, this is basically the formula. And
that will give you the U (P, f) and L (P, f) of a given function. So, I hope you were able to
understand this example and we will probably have such kind of examples in your exercise
or in your assignments and hopefully that will make the concepts even more clear all right.

(Refer Slide Time: 13:51)

So, now that we have looked into L(P, f) and U(P, f) example. let us consider an example,
where we actually say whether our given function is Riemann integrable or not. So,
1 𝑓𝑜𝑟 1 ≤ 𝑥 ≤ 2 3
example 2; example 2 is if 𝑓(𝑥) = { , then evaluate ∫1 𝑓(𝑥)𝑑𝑥.
2 𝑓𝑜𝑟 2 ≤ 𝑥 ≤ 3

So, first of all the given function is f (x)=1 between [1,2] and f(x)=2 between [2,3]. So,
obviously it is a discontinuous function at x =2. Now, having just one point of discontinuity
that would not create any kind of problem, I mean you can still talk about the Riemann
integrability. Remember all we have to have is whether the function is first of all the
function is bounded or not. And second of all I mean if it is monotonic, then in that case
also we can say that the function is Riemann integrable.

So, first of all here so, here our f (x) is bounded on [1,3] and it is monotone; and it is
monotone there right, which is obvious. So, between [1,2] it is 1 and then between [2,3] it

44
is 2. So, obviously, it is a monotone function with one point of discontinuity at the point
x=2.

And therefore, the theorem or the properties which we studied earlier that if the function
is monotonic, then in that case it is Riemann integrable. So, by that theorem you may have
to look into the number. So, then by previous theorem, I would expect you to write that
theorem number here, f(x) is Riemann integrable right on [1,3] just looked into that
theorem, which say that a monotonic function is Riemann integrable.

So, with the help of that theorem, we can say that this function is at least Riemann
integrable that means, this integral here it makes sense that if someone asks us to evaluate
such kind of integral, first of all we have to make sure that whether the function is integral
or not, whether the integral exists at all exists or not.

So, since by this theorem the integral exists. We can now calculate this integral. So, how
3
do we calculate? So, to calculate we can write the integral ∫1 𝑓(𝑥)𝑑𝑥, now remember 2 is
the point of discontinuity. So, we divide the integral into 2 sub integrals at the point x=2.
2 3
So, ∫1 𝑓(𝑥)𝑑𝑥 + ∫2 𝑓(𝑥)𝑑𝑥.

Now, between [1,2] f(x) is 1 and between [2,3] f(x) is 2. So, if we calculate, then this is
2 3
basically ∫1 𝑑𝑥 + 2 ∫2 𝑑𝑥.. So, this will be (2-1) + (6-2), so ultimately it is 5 sorry this one
is 4 so, it will be 4 , so 6-4 and therefore, this will be 3, so the answer is 3 yes.

So, you see first of all we have to check whether the given function is bounded and
monotonic or not or whether it is continuous or not, so that we can at least talk about that
the function is Riemann integrable or the integral at all exists. And once we have that then
we can calculate the integral the way we please. So, this is how to say of one such example,
where we use those theorems. Now, next we will move to let us go to new page.

45
(Refer Slide Time: 19:13)

Now, here an example; so, another example is given a function f(x) defined by

𝑥 2 𝑓𝑜𝑟 0 ≤ 𝑥 ≤ 1 2
𝑓(𝑥) = { . So, then we have to find out what is so evaluate ∫0 𝑓(𝑥)𝑑𝑥?
√𝑥 𝑓𝑜𝑟 1 ≤ 𝑥 ≤ 2

So, here our given function f is 𝑥 2 between [0,1] and √𝑥 between [1,2]. So, obviously this
function is continuous from [0,2]. And therefore, we can use that continuity theorem that
every continuous function is Riemann integrable. So, we can write 𝑥 2 and √𝑥 are
respectively integral or Riemann integrable in their respective range, since they are both
continuous right, it is very obvious to see that so that means, the function f is continuous.

And if the function f is continuous, then this function is Riemann integrable. So, also f is
continuous on [0,2]. So, at least the continuity part is settled. And if the function is
continuous on that interval [0,2], then obviously it is integrable or Riemann integrable.

So, let us calculate the Riemann integral. So, since we have two different functions from
[0,2], we can divide the range in 2 parts f(x)dx. So, we have divided the range or the limit
how to say the interval [0,2] into two parts. So, after division, it looks like this. So, one
whole integral is divided into two sub integrals.

1 2 𝑥3
Now, we can write it as ∫0 𝑥 2 𝑑𝑥 + ∫1 √𝑥𝑑𝑥. So, this will be from 0 to 1 + this will be
3
3
2
so √𝑥. So, it will be 3 𝑥 2 yes from 1 to 2. So, then we can calculate these and after

46
 3
1 2 2 4√2 1
calculating, we will be able to obtain this as 3 + 3 22 − 3, so this will be I believe − 3.
3

So, we can calculate whatever this is so I leave this up to you.

So, you can see here the given function f was actually given by how to say two different
functions. So, f is of course it is continuous, but it is given with the help of two different
functions. And all we have to check, first of all whether they are continuous or not? If they
are not continuous or if there is even one point of discontinuity, then we check I mean
whether they are monotonic or not. And if they are monotonic and bounded functions, then
they are Riemann integrable.

However, in this case we are very lucky, because both the functions are continuous in their
respective range and therefore the function f is continuous. And if the function f is
continuous, then it is Riemann integrable. So, here I can write a small point f is Riemann
integrable so f is Riemann integrable and now we can calculate the integral from [0,2]. So,
and this will be whatever you get after the calculation here that will be the answer. So,
4√2 1
− 3 would be your answer. So, this is one such example.
3

Now, what I am talking about is with this discontinuity part that let us say, you have one
point of discontinuity or if you have two points of discontinuity, then what will happen to
the remain integrability of the function. And in order to do that we have a small theorem,
which talks about these finite points of discontinuity. And even though if you have finite
points of discontinuity, you can still talk about the Riemann integrability.

47
(Refer Slide Time: 25:49)

So, let us state a very small result all our properties. So, the result goes like this, I would
list it as property all right so, property so it says that any bounded function, which is
continuous except for finite number of points, so that means, it is not continuous at these
points is Riemann integrable.

So, if you have a bounded function, which is considered continuous except for finite
number of points that means, there are finite number of points, where the function is not
continuous even in that case your given function would be Riemann integrable. There is a
little bit generalized form of this theorem. So, the generalized form goes like this.

If a function f(x) is bounded on the closed interval [a,b] and the set of it is points of
discontinuity; has only a finite number of limit points, then also f is Riemann integrable
on [a,b]. So, that means, if you have a given bounded function on a closed interval [a,b]
there and the set of points, where the function is discontinuous has only a finite number of
limit points and then in that case the function is also Riemann integrable. So, it is in a way
a modified version of the previous theorem.

And we will see via some example in our next lecture, what do we mean by these two
theorems and where how can we apply them. So, today we will stop here and in the next
lecture, we will continue with the examples on Riemann integration.

48
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 06
Examples (Contd.)

Hello students. We will continue our lecture with the examples on Riemann integration.
So, in the last class we looked into the theorem on if a function has finite number of
discontinuity or if the set of points of discontinuity of the function has only a finite number
of limit points then the function f is also Riemann integrable.

(Refer Slide Time: 00:40)

So, what do we mean by these 2 theorems we will look into by an example.

49
(Refer Slide Time: 00:54)

So, let us consider this example here.

(Refer Slide Time: 01:00)

So this will be our example 4. Show that the function

1 1 1
𝑓𝑜𝑟 < 𝑥 ≤ 2𝑛 , 𝑛 = 0,1,2 … 𝑠𝑜 𝑜𝑛
𝑓(𝑥) = { 2𝑛 2𝑛−1
0 𝑓𝑜𝑟 𝑥 = 0

Now, is integrable on [0,1] although it has infinite number of points of discontinuity. So,
here the problem states that if you have a function f(x) given in this fashion then we have

50
to show that it is Riemann integrable although it has an infinite number of points of
discontinuity. So, let us see whether it has infinite number of discontinuity or not. So, here
1 1
let us say when n=1 so first of all f(0)=0 when n=0 so sorry here it is 2𝑛+1to 2𝑛 yes, so there
1
was a small correction. So, when n= 0 then it is 𝑓(𝑥) = 1 𝑓𝑜𝑟 < 𝑥 ≤ 1 then 𝑓(𝑥) =
2
1 1 1 1 1 1
𝑓𝑜𝑟 < 𝑥 ≤ 2 and dot dot and so on it will be 𝑓(𝑥) = 2𝑛−1 𝑓𝑜𝑟 < 𝑥 ≤ 2𝑛−1 and so
2 4 2𝑛

on. So, it will continue.

So that means if we see 𝑓 when 𝑥 = 0 the function 𝑓(𝑥) = 0 actually. So, here instead of
having 0 you can write it as 𝑓(𝑥) just to keep the same notation. So, when 𝑥 = 0 , 𝑓 = 0
1 1 1 1
when 𝑥 is between 2 and 1 then 𝑓(𝑥) = 1 when 𝑥 is between 4 to 2 then 𝑓(𝑥) = 2 and so

on. So that means, the function 𝑓 it is how to say monotonic. So, we are getting the value
1 half and so on, so that means, the value is continuously decreasing. So, the function 𝑓 is
basically monotonic. And secondly, we can see that at 𝑥 = 0 the function is discontinuous
1 1
at 𝑥 = 2 it is again discontinuous 𝑥 = 4 it is again discontinuous and so on. So that means

the function has infinite number of points of discontinuity because n is running up to

1
infinity. So, the function is always discontinuous at the point this 2𝑛 type.

Now, we know that the function 𝑓 is monotonic, we can see that from here and also the
function is bounded, because the maximum value it can attain is 1 and the lowest value it
1
can attain is 0. So, when n goes to infinity then this will become and so sorry when n
2𝑛
1
goes to infinity then 2𝑛 will go to 0, and consequently the maximum value of this function

f can attain is 1 and the minimum value it can attend is 0. So, it is also bounded.

51
(Refer Slide Time: 06:13)

So, the function 𝑓 is bounded and monotonic on [0,1] the given interval in the problem
and it has infinite points of discontinuity alright which we just saw.

So, it has infinite point of discontinuity that means, we can write it in other words f is
continuous on [0,1] except at the set of points what are those set of points? 𝑥 = 0, 𝑥 =
1 1 1 1
, 𝑥 = 22 , 𝑥 = 23 up to 𝑥 = 2𝑛 and so on, which has only one limiting point so this our
2
1 1 1 1
limit point. So, this sequence 0, 2 , 22 , 23 dot dot and so on up to and so on it has one
2𝑛

limiting point which is 0. And since in the previous theorem it says that if for the function
𝑓 the set of points of its discontinuity has only a finite number of limit points then in that
case the function is Riemann integrable

So, here in our case the function 𝑓 has only 1 limit point which is 0 and this is obviously,
a finite number of points. So, then in that case we can use that property or we by the help
of that property whatever number we can give this function f is Riemann integrable on the
interval [0,1]. So, just using this property 2 in this case we can be able to say that the given
function f here is Riemann integrable. And now we can calculate the value of this Riemann
1
1 1
integral. So, the value is integral ∫0 ., we have ∫02 𝑓(𝑥)𝑑𝑥 + ∫1 𝑓(𝑥)𝑑𝑥 , next we will divide
2

this we will continue doing this. So, what we will do here is basically; now, we write this

52
 1
as half to sorry integral from. So, here 𝑓(𝑥) is missing. So, ∫1 𝑓(𝑥)𝑑𝑥 then we can write
2
1
1
this one as 2 and here I can write ∫21 𝑓(𝑥)𝑑𝑥 and I can put a small limit.
2𝑛

1
So, lim ∫21 𝑓(𝑥)𝑑𝑥 So, when 𝑛 → ∞ this whole thing will go to 0 and therefore, it will
𝑛→∞ 2𝑛
1
be ∫02 𝑓(𝑥)𝑑𝑥.

(Refer Slide Time: 11:16)

1
Now we can write this as ∫1 𝑓(𝑥)𝑑𝑥 . So, we can write
2

1 1 1
1
2𝑛−1 2𝑛−2 2
∫ 𝑓(𝑥)𝑑𝑥 + lim [∫ 𝑓(𝑥)𝑑𝑥 + ∫ 𝑓(𝑥)𝑑𝑥 +. . . + ∫ 𝑓(𝑥)𝑑𝑥
1 𝑛→∞ 1 1 1
2 2𝑛 2𝑛−1 22

1 1
Now from to 1 what is the value of the function from to 1 the function is 1 and here it
2 2

will be integral from.

Now, we can write this whole thing in terms of other way around. So,

1 1 1
1
21 24 1 2𝑛−1 1
∫ 1𝑑𝑥 + lim [∫ 𝑑𝑥 + ∫ 𝑑𝑥 +. . . + ∫ 𝑑𝑥
1 𝑛→∞ 1 2 1 4 1 2𝑛−1
2 22 23 2𝑛

53
 So, this can be written as

1 1 1 1 1 1 1
+ lim [ ( − 2 ) + ( 2 − 3 ) +. . .]
2 𝑛→∞ 2 2 2 4 2 2

it will continue. So, if we calculate this whole thing then after simplification we will end
up with

1 1 1 1 1 𝑛−1 1
= 2 + lim [23 + 2 . 24 + . . . + (22 ) ]
𝑛→∞ 2

1 1 1 1 𝑛−1
= lim 2 [1 + 22 + 24 +. . . + (22 ) ]
𝑛→∞

1
1 1−( )𝑛 1 4 2
4
= 2 lim 1 = 2.3 = 3
𝑛→∞ 1−
4

(Refer Slide Time: 14:58)

54
(Refer Slide Time: 15:40)

So, this will be the answer of this problem So, here the given function has infinite point of
discontinuity and even though it has so many points of discontinuity all those points of
discontinuity has only 1 limiting point. Therefore, the function is Riemann integrable with
that property. And now we can calculate this integral the definite integral by using a simple
geometric sum formula. So, if we use that formula and when we do n go to infinity then in
2
that case the value of the integral is 3. So, this is one such result where we use that previous

theorem.

(Refer Slide Time: 18:00)

55
Next we can look into another example. So, an another example would be show that
3
∫0 [𝑥]𝑑𝑥 exists. So, first of all I hope most of you know what is a box function is. So let us
write the box function. So, here f(x) is [x]. So, what would be a [x] if x is between 0 and
1 So, box function is basically the integral part of the function; that means, it is sort of like
a floor function.

So, if the value of x is for example, if the value of x is I do not know 1.5 then in that case
box of x will be 1. So, you always take the integral part of an out of the function. So, here

𝑓(𝑥) = [𝑥] = 0 𝑖𝑓 0 ≤ 𝑥 < 1

= 1 𝑖𝑓 1 ≤ 𝑥 < 2

= 2 𝑖𝑓 2 ≤ 𝑥 < 3

= 3 𝑖𝑓 𝑥 = 3

So, now, here in this case what we have is a box function whose Riemann integrability
we have to confirm. So, obviously, the function is Riemann integrable because it is
bounded between 0 to 3 not only that it is also monotonic 0 to 3.

So, it is always a Riemann integrable function. So, f is Riemann integrable on [0,3] right.
3
Next we will calculate ∫0 𝑓(𝑥)𝑑𝑥 . So, we divide the interval into 3 sub intervals

1 2 3
∫ 𝑓(𝑥)𝑑𝑥 + ∫ 𝑓(𝑥)𝑑𝑥 + ∫ 𝑓(𝑥)𝑑𝑥
0 1 2

1
let us go to new page ∫0 .

56
(Refer Slide Time: 20:37)

So, between 0 to 1 f(x) is 0 between 1 to 2 f(x) is 1 and between 2 to 3 between 2 to 3 f(x)

is 2. So, if we calculate then this will be 0+1+6-4, so, ultimately 3. So, that is the value of
the integral. So, it is Riemann integrable and the value is 3 alright. Next, we can look into
2
an example of type this here. So, let us consider an another example evaluate ∫0 |1 − 𝑥|𝑑𝑥.
So, here our function f(x) is |1 − 𝑥| for x between [0,2].

1 − 𝑥 𝑓𝑜𝑟 0 ≤ 𝑥 ≤ 1
So, that is in other words 𝑓(𝑥) = {
𝑥 − 1 𝑓𝑜𝑟 1 ≤ 𝑥 ≤ 2

because your mod function is sort of like a positive function and if our x is between 0 and
1 then in that case this mod function will remain as 1-x and if our x is between 1 and 2
then in that case |1-x| will be written as x-1 So, just to keep the value of positive or the
value absolute in a way. So, let us see if the function is continuous or not. So, here the
function f is definitely continuous between 0 to 1 the first part this part here and the second
part x-1 this 1 is also continuous between 1 to 2

Now, the only thing which we have to make sure here in this case is that whether a function
is continuous or monotone or something and then we can talk about certain integrate built
in. So, since the function is continuous between 0 to 2 it is obviously, Riemann integrable
on [0,2]. So, we can say that since it is continuous I am not writing that part it is it is
obvious I guess that the function is continuous and from here we can say that the function
is Riemann Integrable on [0,2]. And in order to evaluate the Riemann integrable of f(x) we

57
just write the function into 2 sub interval a sub integrals and the integral into 2 sub integrals
2
So, ∫1 𝑓(𝑥)𝑑𝑥 and what is the value of the function from 0 to 1? The value of the function
is 1-x and the value of the function between 1 to 2 is x-1 and I believe you can be able to
evaluate this integral

So, that will be the value of the Riemann integral for the function f. So, this was another
example where you have to how to say break the mod and get the individual functions and
check whether that whether those individual functions are continuous or bounded or
monotonic and then you can talk about the Riemann integrability. Alright, the next
example which we will look into as is to apply the mean value theorem.

So, in order to apply the mean value theorem so, I am sorry.

(Refer Slide Time: 25:32)

So, another example show that for 𝐾 2 < 1,

1
𝜋 2 𝑑𝑥 𝜋 1
≤∫ ≤
6 2 2 2
0 √(1 − 𝑥 )(1 − 𝑘 𝑥 ) 6 2
√1 − 𝐾
4

So, here we have to show that this integral is bounded between these 2 points.

So, in order to apply the mean value theorem we will start with the first mean value
theorem. So, let us see whether we can apply the first mean value theorem or not. So, the

58
first mean value theorem says that if we have how to say, if we have 2 functions f and g.
And if the function g maintains the same sign on the interval a to b then in that case you
can write the integral of 2 functions as the value of the one function multiplied by the
integral of the second function. So, here our 2 functions, so 1 function we can choose as,
1 1
so let us choose 𝑓(𝑥) = √1−𝐾2 and 𝑔(𝑥) = √1−𝑥 2.
𝑥2

So, as for the statement of first fundamental theorem of integral invalid sorry as first mean
value sorry as per the statement of the first mean value theorem f and g both needs to be
1
bounded. So obviously, they are bounded in the interval [0,2] because 𝐾 2 < 1, So, this

quantity is bounded and of course, this function keeps the same sign. So, whatever the
1
value is between [0,2] , this function g here keeps the same sign. And therefore, it satisfies

that criteria of first mean value theorem. And by first mean value theorem I will write it as
MVT first mean value theorem we have

So, it satisfies all the criteria of the first mean value theorem and therefore, by mean value
1
theorem we have 0 ≤ 𝜉 ≤ 2 such that

1 1
1
2 𝑑𝑥 1 2 𝑑𝑥 1
∫ = ∫ = [sin−1 𝑥]20
0 √(1 − 𝑥 2 )(1 − 𝑘 2 𝑥 2 ) √1 − 𝑘 2 𝜉 2 0 √1 − 𝑥 2 √1 − 𝑘 2 𝜉 2

(Refer Slide Time: 29:52)

59
 1 𝜋
Now, this is equals to . Next, so hence
√1−𝑘 2 𝜉 2 6

1
𝑑𝑥 𝜋 1
∫0 √(1−𝑥 2 )(1−𝑘 2 𝑥 2 ) =
2
6 √1−𝑘 2 𝜉 2

And if I take 𝜉 = 1, and 𝜉 equals to a sorry if I take 𝜉 = 0 , so for 𝜉 = 0 this integral, so

let us call it as I.

𝜋 𝜋 1
So,𝐼 = 6 . 1. So, this is basically 6 and if I take 𝜉 = 2, so 𝜉 can attend the balance between
1 𝜋 1
only 2 points 0 and half. So, if I take 𝜉 = 2 then this will be 1 sorry times . So,
6 𝐾
√1−( )2
2

from let us say this one is equation 1 and this one is 2 and this one is 3. So, from 1, 2 and
3 we will have from 1, 2 and 3 we will have

1
𝜋 2 𝑑𝑥 𝜋 1
≤∫ ≤
6 2 2 2
0 √(1 − 𝑥 )(1 − 𝑘 𝑥 ) 6 2
√1 − 𝐾
4

, which is what we needed to prove.

So, with the help of first mean values we can be able to show that this integral is bounded
between these 2 values. And we will look into in the next class probably one application
of the second mean value theorem and then we will move to the reduction formula. So, for
today we will stop here and in the next lecture we will start with some other examples.

Thank you.

60
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 07
Reduction formula

Hello students. So today we are going to start a new topic in our integral calculus section.
So, up until last class we looked into Riemann integration, Riemann integrable functions
as a fundamental theorem of integral calculus, mean value theorems, and we also worked
out few examples. I will also try to include some problems in your assignment sheet, and
I will also provide the solution, so that you will get to practice some problems from
Riemann integration.

Now, today we will start with reduction formula, and we will derive some formula based
on different types of integrals, so where we use the reduction formula. So, what do we
mean by reduction formula actually?

(Refer Slide Time: 01:21)

So, first of all in a integral calculus if I ask you to integrate a function of a type let us say
a function of type let us say if it is ∫ 𝑠𝑖𝑛𝑥𝑑𝑥 then its relatively simple to integrate. Even if
it is∫ 𝑠𝑖𝑛2 𝑥𝑑𝑥 , then it would also be quite simple to integrate you have to multiply by 2,
and then you adjust the factor two and then you do the integration. But suppose if you have

61
an integral of type ∫ 𝑠𝑖𝑛10 𝑥𝑑𝑥, or even worse if you have ∫ 𝑠𝑖𝑛100 𝑥𝑑𝑥 , then in that case
that it would not be that much straightforward to apply our traditional methods of
integration.

So, in such cases you need some kind of how to say iterative formula or some kind of an
induction formula that will how to say give you an easy way to calculate these type of
integrals not only with one trigonometrical function, you could also have ∫ 𝑠𝑖𝑛4 𝑥𝑐𝑜𝑠 8 𝑥𝑑𝑥.
And if you are asked to evaluate an integral of this type then even in this case it will be a
little bit tedious to evaluate this integral using our traditional method. So, then in that case
you need induction formula to evaluate these integrals.

𝑑𝑥 𝑑𝑥
You can also have integrals of type ∫ (𝑥 2 +𝑎2 )𝑛 or ∫ (𝑎+𝑏𝑐𝑜𝑠𝑥)𝑛 So, all these type of integrals

can be put into some kind of reduction formula. And we will see how with the help of
reduction formula, we can actually calculate these integrals very easily. So, let us start with
our first reduction formula. So, let us go into a new page.

(Refer Slide Time: 03:39)

So, suppose we want to derive a reduction formula for ∫ 𝑥 𝑛 𝑒 𝑎𝑥 𝑑𝑥, where n is a positive
integer. So, when I say positive integer, it usually means n is, so that n is 1, 2, 3 dot dot
and so on. And in short I will try to write n belongs to z positive. So, if I write z positive
that basically mean that we are talking about set of all positive integers not 0, which does
not include 0.

62
Now, you may have seen an integral of this type. So, instead of n it could be 1 or 2 or any
positive integer. So, if you have any positive integer sitting at x to the power sitting at the
place of n, then how do you evaluate actually. So, instead of doing the traditional
integration by parts, we will derive a reduction formula that will help us evaluate for any
n. So, let us start.

So, our given integral 𝐼 = ∫ 𝑥 𝑛 𝑒 𝑎𝑥 𝑑𝑥. Now, since we have n here I can put a small n here
i.e. 𝐼𝑛 = ∫ 𝑥 𝑛 𝑒 𝑎𝑥 𝑑𝑥, and so where n is a positive integer. Let us write in short ok. So now,
I will do in the integration by parts. So, I will do basically integration by parts. So, if I do
𝑥 𝑛 𝑒 𝑎𝑥 𝑒 𝑎𝑥
integration by parts, then this will become − ∫ 𝑛𝑥 𝑛−1 𝑑𝑥 and then this will
𝑎 𝑎
𝑥 𝑛 𝑒 𝑎𝑥 𝑛
become − 𝑎 ∫ 𝑥 𝑛−1 𝑒 𝑎𝑥 𝑑𝑥.
𝑎

So, now if you see this term here; this is nothing but 𝐼𝑛−1 . So, if I take here, if I take in
place of n, if I take (n-1) then this formula is basically ∫ 𝑥 𝑛−1 𝑒 𝑎𝑥 𝑑𝑥. So that means, this
𝑥 𝑛 𝑒 𝑎𝑥 𝑛
will be equal to − 𝑎 𝐼𝑛−1 . So, this is basically our reduction formula. Now, this
𝑎
𝑥 𝑛 𝑒 𝑎𝑥 𝑛
implies 𝐼𝑛 = − 𝑎 𝐼𝑛−1 . So, this is our required reduction formula all right.
𝑎

(Refer Slide Time: 07:31)

Now, let us say we want to apply this reduction formula to find, so example 1. So, let us
say we want to use that reduction formula to find ∫ 𝑥 4 𝑒 𝑎𝑥 𝑑𝑥. So, as you can see if I go
with the traditional integration by parts formula, then this will be a little bit how to say

63
extensive in a way I mean it will require some time to calculate this integral, instead we
will use the reduction formula. So, how we can write solution?

𝑥 4 𝑒 𝑎𝑥 4
So, first of all n=4, so we write 𝐼4 = ∫ 𝑥 4 𝑒 𝑎𝑥 𝑑𝑥. And this can be written as − 𝑎 𝐼3 .
𝑎

Next, we have to calculate 𝐼3 . So, let us calculate 𝐼3 . I can name this equation as equation
𝑥 3 𝑒 𝑎𝑥 3
1. Now, 𝐼3 will be ∫ 𝑥 3 𝑒 𝑎𝑥 𝑑𝑥, and then this can be written as − 𝑎 𝐼2 . Then again 𝐼2
𝑎
𝑥 2 𝑒 𝑎𝑥 2 𝑥𝑒 𝑎𝑥 1
can be written as , − 𝑎 𝐼1 then 𝐼1 . And again 𝐼1 equals to − 𝑎 𝐼0 .
𝑎 𝑎

(Refer Slide Time: 09:57)

Now, I can calculate 𝐼0 . What is 𝐼0 ? 𝐼0 is ∫ 𝑥 0 𝑒 𝑎𝑥 𝑑𝑥. So, 𝑥 0 is 1. Then we are left with
𝑒 𝑎𝑥
∫ 𝑒 𝑎𝑥 𝑑𝑥 and ∫ 𝑒 𝑎𝑥 𝑑𝑥 = 𝑎
. So now, we will go backwards. And we can write we can
𝑥 4 𝑒 𝑎𝑥 4 𝑥 3 𝑒 𝑎𝑥 3
write from 1, we have 𝐼4 = − 𝑎 𝐼3 . What is our 𝐼3 ? 𝐼3 is − 𝑎 𝐼2 . What is our
𝑎 𝑎
𝑥 2 𝑒 𝑎𝑥 2
𝐼2 ? 𝐼2 is − 𝑎 𝐼1 .
𝑎

𝑥𝑒 𝑎𝑥 1 𝑒 𝑎𝑥
And what is our 𝐼1 ? 𝐼1 is − 𝑎 𝐼0 ; and that 𝐼0 = ∫ 𝑒 𝑎𝑥 𝑑𝑥 = , I will close all these
𝑎 𝑎

brackets. So, basically if I multiply + a constant c of course, so at every step we will get a
constant. So, since it is a definite integral at every step, we will get a constant. And we can
finally, write a big constant at the end. And then we can write this as

𝑥 4 𝑒 𝑎𝑥 4𝑥 3 𝑒 𝑎𝑥 12𝑥 2 𝑒 𝑎𝑥 24𝑥𝑒 𝑎𝑥 24𝑒 𝑎𝑥

− + − + +𝑐
4 𝑎2 𝑎3 𝑎4 𝑎5

64
.

So, this is our how to say this answer or the solution of this problem. So, as you can see
that instead of doing integration by parts again and again. We basically use the reduction
formula to calculate the value of this integral. So, this is our required induction formula
for any form of ∫ 𝑥 𝑛 𝑒 𝑎𝑥 𝑑𝑥. So, you can play with any power here. So, instead of 𝑥 4 , we
can even take 𝑥 20 . And then you just calculate 𝐼20 , 𝐼19 , 𝐼18 , 𝐼17 dot dot up to 𝐼0 , and then
you just put them together and that will give you the value of 𝐼20 actually. So, this one was
how to say pretty straightforward.

(Refer Slide Time: 13:49)

Next let us say you have you have an integral of type; so reduction formula for ∫ 𝑠𝑖𝑛𝑛 𝑥𝑑𝑥.
So, in order to find the reduction formula for this, we can write again 𝐼𝑛 = ∫ 𝑠𝑖𝑛𝑛 𝑥𝑑𝑥 .
Now, we can do the integration by parts. So, if I do the integration by parts, I can first
write this as ∫ 𝑠𝑖𝑛𝑛−1 𝑥𝑠𝑖𝑛𝑥𝑑𝑥 . And then I will use integration by parts.

So, if I use integration by parts here, this will reduce to 𝑠𝑖𝑛𝑛−1 𝑥(−𝑐𝑜𝑠𝑥) −
∫(𝑛 − 1)𝑠𝑖𝑛𝑛−2 𝑥𝑐𝑜𝑠𝑥(−𝑐𝑜𝑠𝑥)𝑑𝑥 . So, it is like differentiation of the first function which
was 𝑠𝑖𝑛𝑛−1 𝑥 times integration of the second function which was sine x. So, hence we are
getting minus of cos x. So, this can be written as −𝑠𝑖𝑛𝑛−1 𝑥𝑐𝑜𝑠𝑥 + (𝑛 −
1) ∫ 𝑠𝑖𝑛𝑛−2 𝑥𝑐𝑜𝑠 2 𝑥𝑑𝑥 And we can write

65
 𝐼𝑛 = −𝑠𝑖𝑛𝑛−1 𝑥𝑐𝑜𝑠𝑥 + (𝑛 − 1) ∫ 𝑠𝑖𝑛𝑛−2 𝑥(1 − 𝑠𝑖𝑛2 𝑥)𝑑𝑥

= −𝑠𝑖𝑛𝑛−1 𝑥𝑐𝑜𝑠𝑥 + (𝑛 − 1) ∫ 𝑠𝑖𝑛𝑛−2 𝑥𝑑𝑥 − (𝑛 − 1) ∫ 𝑠𝑖𝑛𝑛 𝑥𝑑𝑥

So, now I can write this as let us go to a new page. So, here this is ∫ 𝑠𝑖𝑛𝑛−2 𝑥𝑑𝑥. So, this
is basically my 𝐼𝑛−2 . So, if I take instead of I, if I take 𝐼𝑛−2 , then this will be basically
∫ 𝑠𝑖𝑛𝑛−2 𝑥𝑑𝑥. So, this is our 𝐼𝑛−2 , and this is again our 𝐼𝑛 . So, I will bring 𝐼𝑛 on the left
hand side.

(Refer Slide Time: 17:07)

So, this will become

𝑛𝐼𝑛 = 𝑠𝑖𝑛𝑛−1 𝑥𝑐𝑜𝑠𝑥 + (𝑛 − 1)𝐼𝑛−2

So, if I divide by n on both sides, since n is a positive integer, we can do that

𝑠𝑖𝑛𝑛−1 𝑥𝑐𝑜𝑠𝑥 (𝑛 − 1)
𝐼𝑛 = + 𝐼𝑛−2
𝑛 𝑛

So, this is our required reduction formula for ∫ 𝑠𝑖𝑛𝑛 𝑥𝑑𝑥 . So, if you have instead of n, if
you have let us say 20 or even 10. Then in that case we just have to put n=10 or 20 whatever
it is and then we keep on calculating 𝐼𝑛 then 𝐼𝑛−1 and 𝐼𝑛−2 dot dot so until we get here 𝐼0 .
And then we can easily calculate 𝐼0 and that will basically give us the answer of the
required reduction formula.

66
 𝜋
Now, suppose we need to calculate ∫02 𝑠𝑖𝑛𝑛 𝑥𝑑𝑥 . So, this is our required integral formula,
𝜋
we can write 𝑛 ∈ ℤ+ . Suppose, we are asked to calculate ∫02 𝑠𝑖𝑛𝑛 𝑥𝑑𝑥 . So, what, what this
𝜋
𝑠𝑖𝑛𝑛−1 𝑥𝑐𝑜𝑠𝑥 2
would be this would be basically [ ]0 . And this will be a definite integral of
𝑛
𝜋 𝜋
course, I am writing 𝐼𝑛−2, but this is basically my definite integral, and at 𝑥 = 𝑐𝑜𝑠 2 is
2

0 and at x=0 sine 0.

So, basically that one is also 0. So, the first term will actually vanish and then we are left
with
(𝑛 − 1)
𝐼𝑛 = 𝐼𝑛−2
𝑛

𝜋
where n is a positive integer. So, if you are given this range from 0 to 2 , then the required

reduction formula would be

(𝑛 − 1)
𝐼𝑛 = 𝐼𝑛−2
𝑛

some people also denote it by j. So, instead of using I not to confuse, with this I they
introduce j, but that does not matter here. So, for example, let us say if we are asked to
𝜋
calculate ∫02 𝑠𝑖𝑛5 𝑥𝑑𝑥 , then this will be basically our, so suppose if we are asked to
calculate this. So, let us call it as 𝐼5 .

(Refer Slide Time: 20:31)

67
 5−1 4
And 𝐼5 would be basically. So, 𝐼5 would be basically 𝐼5−2 . So, this is basically 5 𝐼3 and
5
𝜋
2
𝐼3 is 3-1. So, 3 𝐼1 . So, what is our 𝐼1 . So, our 𝐼1 is ∫0 𝑠𝑖𝑛𝑥𝑑𝑥. So, integral of sinx is -cos
2

𝜋 𝜋
x 0 to 2 . So, cos 2 is 0 and cos0 is 1. So, this is basically plus 1, because we have a minus
4 42 8
here. So, thus our 𝐼5 = 5 𝐼3 which is 5 3 𝐼1, and this will be 15, because 𝐼1 is 1.

So, this is our required answer to these reduction formulas instead of n=5, you can have
n=15. And you just proceed in the similar manner. You may have noticed a trick here that
when n is odd that means if you have n to the power 5, 7, 13 or 15, then you will basically
end up with calculating 𝐼1 . And if n is even, that means, if you have sine to the power 6,
10, 14, then in that case you will end up with 𝐼0 . And sine to the power 0 is basically 1.
𝜋
And then you just calculate ∫02 𝑑𝑥, and that will give you the answer. So, if n is odd, then
you end up with calculating 𝐼1 ; if n is even, then you are end up calculating 𝐼0 at the end,
so that is the trick here.

Now, that we have the reduction formula for sinx, the reduction formula for cos x; so
reduction formula for ∫ 𝑐𝑜𝑠 𝑛 𝑥𝑑𝑥 . So, this one is ∫ 𝑐𝑜𝑠 𝑛 𝑥𝑑𝑥 . So, this follows on the
similar footsteps basically. So, the way we have calculated the induction formula for
∫ 𝑠𝑖𝑛𝑛 𝑥𝑑𝑥 , we can also calculate the induction formula for∫ 𝑐𝑜𝑠 𝑛 𝑥𝑑𝑥 . You just have to
write it as 𝑐𝑜𝑠 𝑛−1 𝑥𝑐𝑜𝑠𝑥 like the way we did for the sine. So, I leave this task to the
students. And I am pretty sure you can be able to do it.

So, I am just writing the formula. So, the formula would be from here it would imply our
reduction formula would be

𝑐𝑜𝑠 𝑛−1 𝑥𝑠𝑖𝑛𝑥 (𝑛 − 1)

𝐼𝑛 = + 𝐼𝑛−2
𝑛 𝑛

𝜋 𝜋
And, if we have let us say ∫02 𝑐𝑜𝑠 𝑛 𝑥𝑑𝑥 , like we saw in the ∫02 𝑠𝑖𝑛𝑛 𝑥𝑑𝑥 , then from here
(𝑛−1)
our 𝐼𝑛 would be just 𝐼𝑛−2 . So, it is pretty much the same actually like the formula we
𝑛
𝜋
looked for in ∫02 𝑠𝑖𝑛𝑛 𝑥𝑑𝑥 .

68
So, we can also solve the similar examples like we did for the sinx part. So, I am going to
leave those examples for∫ 𝑐𝑜𝑠 𝑛 𝑥𝑑𝑥 . You can look into any book which I put into the
references and there you will be able to find these kinds of examples. Next is how to say
a little bit interesting. So, instead of having just one trigonometrical function, you can have
the product of trigonometrical functions, and then how you deal with calculating their
integral.

(Refer Slide Time: 25:01)

So, let us say you have a reduction, you have an integral of the following type. So, we are
calculating the reduction formula for ∫ 𝑠𝑖𝑛𝑚 𝑥𝑐𝑜𝑠 𝑛 𝑥𝑑𝑥 , where m and n are both positive
integers. So, the reason we are taking the positive integers is that I mean this is a very vital
point. So, the reason we are taking a positive integer is that here we are always ending up
with either 𝐼0 or 𝐼1 . But if we take fractions, then we might end up with some kind of
fraction here or the how to say this whole reduction formula will turn into a totally
different; it may not follow how to say this type of nice integral representation.

So, in order to put it in a nice representation form like we have here. So, these nice
representations would not be how to say straightforward and that is one of the reason we
are picking we are choosing basically we are choosing m & n as the positive integer. So,
as I was saying a reduction formula, for ∫ 𝑠𝑖𝑛𝑚 𝑥𝑐𝑜𝑠 𝑛 𝑥𝑑𝑥 , where m and n are both
positive integers.

69
So, to derive the formula let us write since now we have two integers. So, I can write
𝐼𝑚,𝑛 = ∫ 𝑠𝑖𝑛𝑚 𝑥𝑐𝑜𝑠 𝑛 𝑥𝑑𝑥 all right. And then we take this, this whole thing as the product
of 𝑐𝑜𝑠 𝑛−1 𝑥(𝑠𝑖𝑛𝑚 𝑥𝑐𝑜𝑠𝑥). And then I choose this function as u and this entire function as
v. And based on this, now I can apply the induction of the how to say it is the integration
by parts. So, now, I can apply integration by parts.

So, to do that if you calculate the integration by parts part, then this is basically 𝑐𝑜𝑠 𝑛−1 𝑥
times integration of the first function unchanged integration of the second function. So, if
I integrate the second function, then you can see that cosxdx is basically how to say it
derivative of sinx, so the cosx is derivative of sinx. So, if I substitute sinx=z, then this will
𝑧 𝑚+1
turn into cosxdx= d z. And ultimately we will be able to obtain . And since our z is
𝑚+1
𝑠𝑖𝑛𝑚+1 𝑥
sinx it will reduce to . So, basically to calculate the integral of this thing we use the
𝑚+1

method of substitution.

Now, minus the derivative of the first function, so (𝑛 − 1)𝑐𝑜𝑠 𝑛−2 𝑥(−𝑠𝑖𝑛𝑥); and then
𝑠𝑖𝑛𝑚+1 𝑥
again if I integrate, then this will reduce to 𝑑𝑥. So, if I play with all these how to
𝑚+1

say functions, and if I use those how to say formulas, then in that case this will basically
𝑐𝑜𝑠𝑛−1 𝑥𝑠𝑖𝑛𝑚+1 𝑥 𝑠𝑖𝑛𝑚+2 𝑥
reduce to + (𝑛 − 1) ∫ 𝑐𝑜𝑠 𝑛−2 𝑥 . So, this will be 𝑠𝑖𝑛𝑚+2 𝑥 , and
𝑚+1 𝑚+1

𝑠𝑖𝑛𝑚+2 𝑥 can be written as𝑐𝑜𝑠 𝑛−2 𝑥𝑠𝑖𝑛𝑚 𝑥𝑠𝑖𝑛2 𝑥 . And now I can write 𝑠𝑖𝑛2 𝑥 as 1 − 𝑐𝑜𝑠 2 𝑥
and then this will become

(Refer Slide Time: 29:59)

70
𝑠𝑖𝑛𝑚+1 𝑥𝑐𝑜𝑠𝑛−1 𝑥 𝑛−1
. And this will become + 𝑚+1 ∫ 𝑐𝑜𝑠 𝑛−2 𝑥𝑠𝑖𝑛𝑚 𝑥 (1 − 𝑐𝑜𝑠 2 𝑥)𝑑𝑥. Now, if I
𝑚+1

see if here, so this is basically if I take the first term, then this is basically 𝐼𝑚,𝑛−2 and this
will be again −𝐼𝑚,𝑛 . So, I can write this as

𝑠𝑖𝑛𝑚+1 𝑥𝑐𝑜𝑠 𝑛−1 𝑥 𝑛 − 1 𝑛−1

+ 𝐼𝑚,𝑛−2 − 𝐼
𝑚+1 𝑚+1 𝑚 + 1 𝑚,𝑛

And if I take everything on the, if I take everything on the on the right hand left hand side,
𝑛−1
then this will become 𝐼𝑚,𝑛 [1 + 𝑚+1] . And then on the right hand side, we will have
𝑠𝑖𝑛𝑚+1 𝑥𝑐𝑜𝑠𝑛−1 𝑥 𝑛−1
+ 𝑚+1 𝐼𝑚,𝑛−2 . So, next we will divide by (m+1) on both sides. And we will
𝑚+1

take (m+n) on the left hand side on the denominator on the right hand side. So, this will
1
become 𝑚+𝑛 [𝑠𝑖𝑛𝑚+1 𝑥𝑐𝑜𝑠 𝑛−1 𝑥 + (𝑛 − 1)𝐼𝑚,𝑛−2 ].

So, this is basically our required reduction formula for product of two trigonometrical
functions of this type. And we will use this formula to calculate how to say the product of
two trigonometrical functions integral of the product of two trigonometrical functions. And
we will see that in our next class.

Thank you.

71
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 08
Reduction formula (Contd.)

Hello students, we will continue our this lecture with the Reduction formula and we will
start from where we left off. So, we were trying to derive the reduction formula for the
product of 2 trigonometrical functions and we basically started with integral of type 𝐼𝑚,𝑛 =
∫ 𝑠𝑖𝑛𝑚 𝑥𝑐𝑜𝑠 𝑛 𝑥𝑑𝑥

(Refer Slide Time: 00:38)

I believe here there was a small error, so it should be plus and then we ultimately obtained
this reduction formula.

72
(Refer Slide Time: 00:48)

So now, we will see an example where we can apply this reduction formula, so an another
example.

(Refer Slide Time: 00:54)

So, find ∫ 𝑠𝑖𝑛4 𝑥𝑐𝑜𝑠 2 𝑥𝑑𝑥. So, the solution would go something like this, so we can write
𝐼4,2 = ∫ 𝑠𝑖𝑛4 𝑥𝑐𝑜𝑠 2 𝑥𝑑𝑥.

Now, 𝐼4,2 can be written as sin to the power, it can be written as 𝑠𝑖𝑛𝑚+1 𝑥, so, m+1would
be 5. So, where is that formula? So, here it will be m+1. So,

73
 𝑠𝑖𝑛5 𝑥𝑐𝑜𝑠 2 𝑥 2 − 1
𝐼4,2 = + 𝐼
6 6 4,0

𝑠𝑖𝑛5 𝑥𝑐𝑜𝑠2 𝑥 1
So, here we will have + 6 𝐼4,0 . So, 𝐼4,0 , next 𝐼4,0 is basically ∫ 𝑠𝑖𝑛4 𝑥𝑑𝑥.
6

Now, 𝐼4,0 and here we have ∫ 𝑠𝑖𝑛4 𝑥𝑑𝑥 will again fall into that ∫ 𝑠𝑖𝑛𝑛 𝑥𝑑𝑥 formula and
∫ 𝑠𝑖𝑛𝑛 𝑥𝑑𝑥 formula was ∫ 𝑠𝑖𝑛𝑛−1 𝑥𝑠𝑖𝑛𝑥𝑑𝑥 . So, this is basically

𝑠𝑖𝑛3 𝑥𝑐𝑜𝑠𝑥 3
𝐼4,0 = ∫ 𝑠𝑖𝑛4 𝑥𝑑𝑥 = + 𝐼2,0
4 4

(Refer Slide Time: 03:44)

So, 𝐼2,0 can be written as ∫ 𝑠𝑖𝑛2 𝑥𝑑𝑥 and ∫ 𝑠𝑖𝑛2 𝑥𝑑𝑥 can again follow the same what to say
𝑠𝑖𝑛𝑥𝑐𝑜𝑠𝑥
integral reduction formula. So, this can be written as sin 2 minus 1, so basically +
2
1
𝐼 and 𝐼0,0 is nothing, but ∫ 𝑠𝑖𝑛0 𝑥𝑑𝑥 and ∫ 𝑠𝑖𝑛0 𝑥𝑑𝑥 is basically simply x.
2 0,0

So, I can write it this whole thing as

𝑠𝑖𝑛5 𝑥𝑐𝑜𝑠𝑥 1 3 𝑠𝑖𝑛𝑥𝑐𝑜𝑠𝑥 𝑥

𝐼4,2 = + [ { + }] + 𝑐
6 6 4 2 2

So, you will obtain whatever after the multiplication you will get from here.

74
So, this is one another way to calculate the product of 2 trigonometrical functions and you
can play with any power here. So, you can play with any power here you please sorry, here
you can play with any power of power of sin x and cos x here. So, and then you just apply
this formula, this one here and that will give you the required answer for that particular
integral and this is one such iterative formula.

(Refer Slide Time: 05:55)

Next we will derive the reduction formula of type, let us say ∫ 𝑡𝑎𝑛𝑛 𝑥𝑑𝑥 where n is again
a positive integer. So, to do that I can again write 𝐼𝑛 because of that n here and then I will
write ∫ 𝑡𝑎𝑛𝑛 𝑥𝑑𝑥 and I can split this∫ 𝑡𝑎𝑛𝑛 𝑥𝑑𝑥 as ∫ 𝑡𝑎𝑛𝑛−2 𝑥𝑡𝑎𝑛2 𝑥𝑑𝑥 . And, then you
just 𝑡𝑎𝑛2 𝑥 write as a (𝑠𝑒𝑐 2 𝑥 − 1) like we know from the trigonometrical formula and
then we just proceed as before.

So, we write this one as (𝑠𝑒𝑐 2 𝑥 − 1) here and then you will basically end up with an
iterative formula of type

𝐼𝑛 = ∫ 𝑡𝑎𝑛𝑛 𝑥𝑑𝑥 = ∫ 𝑡𝑎𝑛𝑛−2 𝑥(𝑠𝑒𝑐 2 𝑥 − 1)𝑑𝑥 = ∫ 𝑡𝑎𝑛𝑛−2 𝑥𝑠𝑒𝑐 2 𝑥𝑑𝑥 − ∫ 𝑡𝑎𝑛𝑛−2 𝑥𝑑𝑥

and here if I substitute 𝑡𝑎𝑛𝑥 = 𝑧 then in that case that 𝑑𝑧 = 𝑠𝑒𝑐 2 𝑥𝑑𝑥 . And, we integrate
𝑡𝑎𝑛𝑛−1 𝑥
and then this will become − 𝐼𝑛−2 . So, this is basically the reduction formula for
𝑛−1

∫ 𝑡𝑎𝑛𝑛 𝑥𝑑𝑥.I mean all you have to do is play with some trigonometrical formula it is not
that tricky

75
And similarly if you have some range for the definite integral that is if it is
𝜋
∫04 𝑡𝑎𝑛𝑛 𝑥𝑑𝑥 then you just have to integrate this part and keep on calculating depending
on the power of n here. So, if you have n =4 then this will become (4-2) and then again in
the next step it will become (2-0) and if it is odd then you will end up with 𝐼1 at the end.
So, it is pretty much on the similar footsteps of ∫ 𝑠𝑖𝑛𝑛 𝑥𝑑𝑥 and ∫ 𝑐𝑜𝑠 𝑛 𝑥𝑑𝑥 formula. So, I
leave the examples for this direction formula up to the students you can look into any book.

(Refer Slide Time: 09:00)

Similarly, if you have let us say if you have other trigonometrical functions for example if
you have 𝐼𝑛 = ∫ 𝑐𝑜𝑡 𝑛 𝑥𝑑𝑥 then you can again write 𝑐𝑜𝑡 𝑛 𝑥𝑑𝑥 as 𝑐𝑜𝑡 𝑛−2 𝑥𝑐𝑜𝑡 2 𝑥𝑑𝑥 𝑑𝑥 and
𝑐𝑜𝑡 2 𝑥 can be written as 𝑐𝑜𝑠𝑒𝑐 2 𝑥 − 1 and then you follow the similar steps like before. So,
ultimately here the reduction formula would be

𝑐𝑜𝑡 𝑛−1 𝑥
𝐼𝑛 = ∫ 𝑐𝑜𝑡 𝑛 𝑥𝑑𝑥 = − 𝐼𝑛−2
𝑛−1

So, this will be the reduction formula for ∫ 𝑐𝑜𝑡 𝑛 𝑥𝑑𝑥 .

Similarly, one can have 𝐼𝑛 = ∫ 𝑠𝑒𝑐 𝑛 𝑥𝑑𝑥 . So, here also you write 𝑠𝑒𝑐 𝑛 𝑥 as 𝑠𝑒𝑐 𝑛−2 𝑥𝑠𝑒𝑐 2 𝑥
and we substitute how to say 𝑠𝑒𝑐 2 𝑥 as how to say 1 + 𝑡𝑎𝑛2 𝑥 and then we use some how
𝑠𝑒𝑐 𝑛−2 𝑥𝑡𝑎𝑛𝑥
to say basic method of substitution. So, this will again give you 𝐼𝑛 = +
𝑛−1
𝑛−2
𝐼 . I am just writing this because these are very straightforward all you have to do is
𝑛−1 𝑛−2

76
 𝑛−2
do some basic integration and this will be 𝐼 and use some basic algebraic
𝑛−1 𝑛−2

operations.

So, similarly we have the reduction formula for ∫ 𝑠𝑒𝑐 𝑛 𝑥𝑑𝑥 you can also have in the
similar fashion you can also have ∫ 𝑐𝑜𝑠𝑒𝑐 𝑛 𝑥𝑑𝑥 and you can write 𝑐𝑜𝑠𝑒𝑐 𝑛 𝑥
as 𝑐𝑜𝑠𝑒𝑐 𝑛−2 𝑥𝑐𝑜𝑠𝑒𝑐 2 𝑥 . And, then you use again some trigonometrical formula and you
𝑐𝑜𝑠𝑒𝑐 𝑛−2 𝑥𝑐𝑜𝑡𝑥 𝑛−2
will basically end up with 𝐼𝑛 = + 𝑛−1 𝐼𝑛−2 . Here you can see a pattern that
𝑛−1

say ∫ 𝑠𝑖𝑛𝑛 𝑥𝑑𝑥 and ∫ 𝑐𝑜𝑠 𝑛 𝑥𝑑𝑥 follows the same formula reduction formula, similarly
∫ 𝑡𝑎𝑛𝑛 𝑥𝑑𝑥 and ∫ 𝑐𝑜𝑡 𝑛 𝑥𝑑𝑥 follow the same trigonometrical formula and also ∫ 𝑠𝑒𝑐 𝑛 𝑥𝑑𝑥
and∫ 𝑐𝑜𝑠𝑒𝑐 𝑛 𝑥𝑑𝑥 also follow pretty much the same trigonometrical formula.

So, instead, so when you have when instead of 𝑠𝑒𝑐 𝑛 𝑥 if you have 𝑐𝑜𝑠𝑒𝑐 𝑛 then we just
replace this secx by cosecx and we replace tanx by cotx and then you will pretty much end
up with a similar type of formula. And as I was talking the derivation of these formulas
are not difficult, so I will leave those task to the students for practice. Next we have
reduction formula of type let us go to a new page.

(Refer Slide Time: 12:29)

So, next we have a reduction formula of type let us say, reduction formula. So, as you are
seeing that by the help of these reduction formulas we are deriving some kind of iterative
schemes that will help us calculate very complicated powers of functions involved in your
integral and you just have to put in those reduction formula you just I mean and you just

77
calculate if it is either known I to the 𝐼10 and then you may have to calculate 𝐼8 𝐼6 𝐼4 𝐼2 . And,
then 𝐼0 and ultimately you just put those in that formula and that will give you the value
of that 𝐼10 for example, or if it is an odd one then you can just calculate how to say let us
say if it is 𝐼11 , then you calculate 𝐼9 𝐼7 𝐼5 𝐼3 and then ultimately 𝐼1 and that will give you the
answer.

So, even to you initially have a very complicated or a difficult integral or a product of 2
functions to you just have to use these formula and that will give you the answer. So, in a
way a reduction formula is a nice shortcut or a nice tool that will help you calculate those
difficult integral. Next we have a function of type this. So, let us write 𝐼𝑚,𝑛 =
∫ 𝑐𝑜𝑠 𝑚 𝑥𝑠𝑖𝑛𝑛𝑥𝑑𝑥 . So, basically I will integrate by parts. So, I will integrate by parts and
if I integrate by parts then this will be

𝑐𝑜𝑠 𝑚 𝑥𝑐𝑜𝑠𝑛𝑥 𝑚
− − ∫ 𝑐𝑜𝑠 𝑚−1 𝑥(−𝑠𝑖𝑛𝑥)(−𝑐𝑜𝑠𝑛𝑥)𝑑𝑥
𝑛 𝑛

Now, if you remember the formula for sin(𝑛 − 1) 𝑥, so the formula sin(𝑛 − 1) 𝑥 for is
equals to 𝑠𝑖𝑛𝑛𝑥𝑐𝑜𝑠𝑥 − 𝑐𝑜𝑠𝑛𝑥𝑠𝑖𝑛𝑥. So, I will use this formula here and then this whole
thing will ultimately reduce to

𝑐𝑜𝑠 𝑚 𝑥𝑐𝑜𝑠𝑛𝑥 𝑚
− − [∫ 𝑐𝑜𝑠 𝑚 𝑥𝑠𝑖𝑛𝑛𝑥𝑑𝑥 − ∫ 𝑐𝑜𝑠 𝑚−1 𝑥𝑠𝑖𝑛(𝑛 − 1)𝑥𝑑𝑥]
𝑛 𝑛

(Refer Slide Time: 16:44)

78
So this will reduce to

𝑐𝑜𝑠 𝑚 𝑥𝑐𝑜𝑠𝑛𝑥 𝑚 𝑚
− − ∫ 𝑐𝑜𝑠 𝑚 𝑥𝑠𝑖𝑛𝑛𝑥𝑑𝑥 + ∫ 𝑐𝑜𝑠 𝑚−1 𝑥𝑠𝑖𝑛(𝑛 − 1)𝑥𝑑𝑥
𝑛 𝑛 𝑛

So, this is basically our 𝐼𝑚.𝑛 and this 1 is basically 𝐼𝑚−1,𝑛−1.

𝑐𝑜𝑠𝑚 𝑥𝑐𝑜𝑠𝑛𝑥 𝑚 𝑚
So, I can write it as− − 𝐼 + 𝐼 . So, if I take 𝐼𝑚.𝑛 on the left hand
𝑛 𝑛 𝑚.𝑛 𝑛 𝑚−1,𝑛−1

side and if I do multiply both sides by n then ultimately we will end up with

[−𝑐𝑜𝑠 𝑚 𝑥𝑐𝑜𝑠𝑛𝑥 + 𝑚𝐼𝑚−1,𝑛−1 ]

𝐼𝑚.𝑛 =
𝑚+𝑛

and this is our required reduction formula for 𝐼𝑚.𝑛 .

So, if we have ∫ 𝑐𝑜𝑠 𝑚 𝑥𝑠𝑖𝑛𝑛𝑥𝑑𝑥 our required reduction formula would be of this type. So,
again instead of m if you have 10 and instead of n if you have 20, then you just have to put
here and then do the calculation and ultimately you will be able to end up with either 𝐼0.0
or 𝐼1.1 or 𝐼1.0, I mean a relatively simple integral to evaluate basically and just how to say
reverse engineers, so that engineer all those things; that means, you just have to put
everything backwards and then that will give you the answer.

So, this is one such reduction formula. Similarly, instead of 𝑐𝑜𝑠 𝑚 𝑥 you can have
𝑠𝑖𝑛𝑚 𝑥𝑐𝑜𝑠𝑛𝑥 and then you proceed in the similar fashion and then you will probably end
up with a similar formula of this type. Next we will look into formula of type let us say
formula of type.

79
(Refer Slide Time: 19:57)

𝑑𝑥
So, reduction formula ∫ (𝑥 2 +𝑎2)𝑛 where n is a positive integer so. Next what we will do
𝑑𝑥
we will write 𝐼𝑛 = ∫ (𝑥 2 +𝑎2 )𝑛 and what we will do is basically we will write this function.

So, we will write this function as integration by parts we will do integration by parts. So,
1 1
our first function is (𝑥 2 +𝑎2 )𝑛 and the second function is 1 So, this will be (𝑥 2 +𝑎2 )𝑛 and the

second function is 1, so integration of the second function would be simply x minus

differentiation of the first function.

𝑛(2𝑥)
So, this will be ∫ (𝑥 2 +𝑎2 )𝑛+1 𝑥𝑑𝑥 all right. So, this can be written as

𝑥 𝑥 2 𝑑𝑥
+ 2𝑛 ∫ 2
(𝑥 2 + 𝑎2 )𝑛 (𝑥 + 𝑎2 )𝑛+1

right.

80
(Refer Slide Time: 21:48)

Now, I can write

𝑥 𝑥 2 + 𝑎2 − 𝑎2
+ 2𝑛 ∫ 2 𝑑𝑥
(𝑥 2 + 𝑎2 )𝑛 (𝑥 + 𝑎2 )𝑛+1

So, here we have minus and then this integral, so this will basically turn this whole thing
into a plus, so this is plus and if I go back then this will be plus and then this will be a
minus. So, this is

𝑥 𝑥 2 + 𝑎2 𝑑𝑥
2 2 𝑛
+ 2𝑛 ∫ 2 2 𝑛+1
𝑑𝑥 − 2𝑛𝑎2 ∫ 2
(𝑥 + 𝑎 ) (𝑥 + 𝑎 ) (𝑥 + 𝑎2 )𝑛+1

𝑥
So, this is basically (𝑥 2 +𝑎2 )𝑛 + 2𝑛𝐼𝑛 − 2𝑛𝑎2 𝐼𝑛+1 . So, I can write

𝑥
2𝑛𝑎2 𝐼𝑛+1 = + (2𝑛 − 1)𝐼𝑛
(𝑥 2 + 𝑎2 )𝑛

because this is this whole expression is equals to 𝐼𝑛 . So, I am bringing this on the left hand
side and I am bringing 𝐼𝑛 on the right hand side, so we will end up with this.

81
Next we will divide the both sides by 2𝑛𝑎2 and since we need the reduction formula for
𝐼𝑛 , so basically I will substitute n equals to (n-1) on both sides.

(Refer Slide Time: 24:06)

So, this will yield

𝑥 2(𝑛 − 1) − 1
𝐼𝑛 = + 𝐼
2(𝑛 − 1)𝑎2 (𝑥 2 + 𝑎2 )𝑛−1 2(𝑛 − 1)𝑎2 𝑛−1

So, this will be our required reduction formula, let me verify. So, n-1 yes, so this will be
𝑑𝑥
our required reduction formula for ∫ (𝑥 2 +𝑎2 )𝑛 , so a function of this type all right. So, if

you are given n equals to let us say n=10 then you just have to substitute n equal to 10 here
and then you keep on calculating 𝐼9 , 𝐼8 , 𝐼7 , up to 𝐼0 and you just substitute the value of
𝐼9 , 𝐼8 , 𝐼7 up to 𝐼0 and that will give you the value of 𝐼10 basically So, this is one such
reduction formula.

Some people right prefer to write it in a rather how to say simpler form. So, you just keep
this thing on the left hand side and you just write

𝑥
2(𝑛 − 1)𝐼𝑛 = + (2𝑛 − 3)𝐼𝑛−1
(𝑥 2 + 𝑎2 )𝑛−1

So, this will be your required reduction formula, yes, from 𝐼𝑛 . Next you can have functions
of type algebraic multiplied by trigonometrical functions.

82
So, our next reduction formula would be of type let us say reduction formula for
∫ 𝑥 𝑛 𝑠𝑖𝑛𝑥𝑑𝑥. So, to find the reduction formula for this function I can write 𝐼𝑛 =
∫ 𝑥 𝑛 𝑠𝑖𝑛𝑥𝑑𝑥 and here basically we have to do an integration by parts. So, we do integration
by parts for, so the first function will remain unchanged integration of second function
would be

−𝑥 𝑛 𝑐𝑜𝑠𝑥 + ∫ 𝑛𝑥 𝑛−1 𝑐𝑜𝑠𝑥𝑑𝑥

and then we do the integration by parts again for this part.

(Refer Slide Time: 27:34)

And if we do the integration by parts for this one, then you will basically obtain

= −𝑥 𝑛 𝑐𝑜𝑠𝑥 + 𝑛 [𝑥 𝑛−1 𝑠𝑖𝑛𝑥 − ∫(𝑛 − 1)𝑥 𝑛−2 𝑠𝑖𝑛𝑥𝑑𝑥]

= −𝑥 𝑛 𝑐𝑜𝑠𝑥 + 𝑛𝑥 𝑛−1 𝑠𝑖𝑛𝑥 − 𝑛(𝑛 − 1) ∫ 𝑥 𝑛−2 𝑠𝑖𝑛𝑥𝑑𝑥

So, this one is our 𝐼𝑛−2. So, this is our 𝐼𝑛−2 So, I can write −𝑥 𝑛 𝑐𝑜𝑠𝑥 + 𝑛𝑥 𝑛−1 𝑠𝑖𝑛𝑥 −
𝑛(𝑛 − 1)𝐼𝑛−2 .

So, this is our so this is our required reduction formula for 𝐼𝑛 of this type. You can have a
𝜋 𝜋
integral ∫02 𝑥 𝑛 𝑠𝑖𝑛𝑥𝑑𝑥 let us say an example So, you can have integral ∫02 𝑥 𝑛 𝑠𝑖𝑛𝑥𝑑𝑥 . So,
you can see that will basically be left with ∫ 𝑥 𝑛−1 𝑠𝑖𝑛𝑥𝑑𝑥 . So, we will be left with n times.

83
So, if I substitute in there. So, we will have 𝐼𝑛 minus sorry, so there should not be any d x
here.

Now, if I substitute so there should not be any dx here. Now if I substitute x =0 then the
𝜋 𝜋
cos0 is 1 and if I substitute 𝑥 = 2 , so 𝑐𝑜𝑠 2 is equals to 0, so the cos0 is 1, but x=0. So, the
𝜋 𝜋
first so at 0 this will vanish and at 2 , 𝑐𝑜𝑠 2 is 0; so this one will again vanish and therefore,
𝜋 𝜋
and as far as this one is concerned, so at 2 this will remain. So, this will be ( 2 )𝑛−1 and at

x=0 sin vanishes. So, this at x=0 term will not exist and then this one will be 𝑛(𝑛 − 1)𝐼𝑛−2 .

So, I can bring and this one is plus and so, I can bring this whole thing on the left hand
𝜋
side. So, 𝐼𝑛 + 𝑛(𝑛 − 1)𝐼𝑛−2 = 𝑛( 2 ) 𝑛−1 So, this will be basically the answer to this
𝜋
reduction formula for the integral ∫02 𝑥 𝑛 𝑠𝑖𝑛𝑥𝑑𝑥 . So, this is how we apply this induction
formula instead of n you can have n=50 and then you just have to calculate 𝐼48 , 𝐼46 and so
on and then just put it back into this formula and that will give you the answer.

So, instead of sinx you can have any trigonometrical function and you just have to find in
some way to up to write that trigonometrical part as 𝐼𝑛−2 to get the formula of type 𝐼𝑛−2
or 𝐼𝑛−1 and then we will basically obtain a nice representation of this type. So, it is all
about playing with the trigonometrical functions and some formulas and you sort of obtain
the required reduction formula. So, it is not that difficult.

And as far as the examples are concerned just substitute the how to say range of integration
range of this definite integral and then you just calculate or play with these algebraic
formulas. And as far as these induction formulas are concerned at the end of it you can
refer to if you prefer to write a constant just write the constant otherwise it is understood
that there will a constant be here when you calculate let us say for n =10 when you calculate
𝐼0 or 𝐼2 ultimately would obtain a constant.

So, it is up to you whether you write this constant here or not because even if you write
the constant here and at the end of it when you are calculating 𝐼2 or 𝐼0 then you will obtain
another constants, a constant plus constant is a new constant. So, we prefer to write it as c.
So, it is totally up to you whether you want to write the c or not it is under understood that
it will have a c here.

84
So, we will stop today’s lecture here and in the next lecture I will try to derive 1 or 2 more
reduction formulas and afterwards we will move to the improper integral. And in your
assignment I will include some problems from reduction formula just to make the concept
clear. So, I hope this lecture was fruitful to you and I look forward to your next lecture.

Thank you.

85
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 09
Improper Integral

Hello students. So, up until last class, we looked into reduction formulas. We also derived
several types of reduction formulas involving trigonometrical functions, involving
algebraic functions. And we also tried to work out few examples. And I hope I was I gave
you enough how to say introduction, and also the flavour of the topic related to reduction
formula. You might also be able to get some problems in your assignment sheet, and I will
also provide the solution. So, hopefully the concept would be much clearer, when you
work out few examples by your own.

(Refer Slide Time: 01:34)

There might be one another type of reduction formula, which is of type ∫ 𝑥 𝑚 (𝑎 +

𝑏𝑥 𝑛 )𝑝 𝑑𝑥. So, you may come across algebraic expressions or integrand of this type. So,
let me change the colour first. So, in such cases what you can do, like we did earlier you
can substitute 𝑢 = (𝑎 + 𝑏𝑥 𝑛 )𝑝 . And then we can choose our function 𝑣 = 𝑥 𝑚 , and then
we just use the integration by parts formula.

86
And after doing some calculation at the end, we will be able to obtain our required
reduction formula of this type. So, our required reduction formula can be obtained by doing
the integration by parts on u and v. And basically, you need to find some kind of iterative
reduction formula to calculate any such problem of this type. So, for example one possible
example can be let us say, 𝐼 = ∫ 𝑥10 (𝑎 + 𝑏𝑥)6 𝑑𝑥. So, this is one such example, which
falls into this category and we basically how to say find the reduction formula of this type
of integral here to calculate an integral of this type.

So, I mean this is one such example, we can have lot of other types of integrals and based
on which you can derive some kind of reduction formula. So, the bottom line is that in
case of reduction formulas all you have to do is to find out some kind of iterative scheme
or some kind of how to say an n type formula for every n, you can be able to calculate a
certain integral. So, I leave this reduction formula topic here. And since we have very
compressed time to cover a lot of things so, I will move on to our next topic, which is
Improper Integral. And we will work out few examples on reduction formula in our
assignment sheet.

(Refer Slide Time: 04:15)

So, let us go to our next topic, which is improper integral, so that is our next topic in
agenda.

87
(Refer Slide Time: 04:28)

So, as you can see we were supposed to reduction formula and the derivation. And the next
topic, we are supposed to cover improper integrals, their convergence and some tests that
will confirm the convergence of an improper integral.

So, let us go back to our topic here. So, when we say improper integral, what do we mean
by it? The term improper itself somehow signifies that something that is not good in a way
or that is not proper in a way. And, something which is not good in case of a function or
in case of an integral is that a function may be having a discontinuity or a function is not
differentiable. So, those kind of properties are somehow related to something is not good
with a function.

So, here in this case it means exactly the same. So, when we say an improper integral, so
𝑏
it could mean that, we know that a definite integral can be written in this form ∫𝑎 𝑓(𝑥)𝑑𝑥.
And when we say that this integral is improper that means, something is not good with this
integral, it means that either the lower limit is ∞ or the upper limit is ∞ or both the lower
limit and upper limits are ∞ or the lower limit and upper limits are fine, but the function
is discontinuous or unbounded at a point in between these two points.

So, in this case f(x) is unbounded at x equals to c in the interval [a,b]. So, do you see that
what do I mean, I mean it means that there is some kind of issue either with the limit points
like here, like in the first three sets of definite integral or there is some kind of issue with
the function itself in the given interval. And if such kind of things happen with an integral,

88
then such type of integral are called as an improper integral or integral which are not
behaving nicely.

So, we can put this in a small statement. So, the small statement is types of improper
integral. So, when we say improper integral, what are their types, so or what do we actually
mean by them. So, the types of improper integrals are the 1st type is the interval increases,
the interval increases without any limit. So, this could be one problem. Like here in this
case or in this case the interval is increasing up to infinity. So, these type of integrals can
be of type improper integral or the 2nd type is the integrand.

So, the integrand is basically the function f(x). The integrand is or the integrand has a finite
number of infinite discontinuities. So, any one of these two things can happen with our
function or with our integral and then such type of integrals are called as the improper
integral.

(Refer Slide Time: 08:55)

So, let us look into the type-1 so type-1 type of improper integral. So, type-1 type of
integral is so in this case we have three types of unboundedness on the range of integral.

So, the first one is limit let us say our lower integral is infinity. So, we have first type of
unboundedness of this type. So, here if a function f(x) is bounded and integrable in 𝐴 ≤
𝑏
𝑥 ≤ 𝑏, where for every A capital 𝐴 < 𝑏. And lim ∫𝐴 𝑓(𝑥)𝑑𝑥 exists, then we say that
𝐴→−∞

89
 𝑏 𝑏
∫−∞ 𝑓(𝑥)𝑑𝑥 exists or it is convergent. And we write ∫−∞ 𝑓(𝑥)𝑑𝑥 is basically
𝑏
lim ∫𝐴 𝑓(𝑥)𝑑𝑥 .
𝐴→−∞

So, the thing is if the interval is unbounded means if we have the range of integral (-∞,b],
then in that case what we do is basically, we get rid of that -∞ part by putting it in terms
𝑏
of limit. And then we basically evaluate this integral here ∫𝐴 𝑓(𝑥)𝑑𝑥 , and afterwards we
pass on the limit. And that will be exactly the same value of this integral. So, for this type
of integral, where the lower limit is minus infinity, we basically calculate this limit and
this limit is actually however the value of the integral.

(Refer Slide Time: 12:15)

In type-1 category another type of integral could be of this type, let us say we have
∞
∫𝑎 𝑓(𝑥)𝑑𝑥. Then the definition pretty much goes the same that if f(x) is a bounded
function, let us go to the previous slide. So, if f(x) is a bounded function and integrable
from a to capital B for every B greater than a.

So, here if f is bounded and integrable on 𝑎 ≤ 𝑥 ≤ 𝐵 for every 𝐵 > 𝑎 and this
𝐵
lim ∫𝑎 𝑓(𝑥)𝑑𝑥 exists, then we say that this integral is convergent. And we can
𝐵→∞
𝐵
write lim ∫𝑎 𝑓(𝑥)𝑑𝑥 is equal to the value of that integral right. So, it follows pretty much
𝐵→∞

the same definition like the definition 1.

90
 +∞
And another type of improper integral could be ∫−∞ 𝑓(𝑥)𝑑𝑥 . So, here both the upper
limits and lower limits are infinity. So, then in that case, what we will do? We will choose
point, so we will write if f is bounded and integrable on 𝐴 ≤ 𝑥 ≤ 𝑎for 𝑎 > 𝐴and on capital
𝑎 𝐵
B 𝐵 ≥ 𝑥 ≥ 𝑎 for every 𝐵 > 𝑎. And lim ∫𝐴 𝑓(𝑥)𝑑𝑥 And lim ∫𝑎 𝑓(𝑥)𝑑𝑥 .
𝐴→−∞ 𝐵→∞

(Refer Slide Time: 15:11)

+∞
They both exist for 𝐴 < 𝑎 < 𝐵. Then we say that this integral ∫−∞ 𝑓(𝑥)𝑑𝑥 convergent
or exist, so this into this integral will exist or convergent. And the value of this integral
+∞
∫−∞ 𝑓(𝑥)𝑑𝑥 , can be split into two sub integrals, so addition of two sub integrals.

So, you basically see what we are doing? So, we had−∞ and +∞ at both the end points,
so what we have done? We have chosen a point in between −∞ to +∞, so that now we
have infinity at only one of the endpoints. And then it falls into the definition of one or the
definition of two, so that is what we are basically doing?

So, we are dividing the whole interval into two parts. And then we are using the definition
one and two to show that these two limits these two individual limits. So, here I can write
𝑎 𝐵
lim , so these two individual limits lim ∫𝐴 𝑓(𝑥)𝑑𝑥 + lim ∫𝑎 𝑓(𝑥)𝑑𝑥 and if this limit
𝐴→−∞ 𝐴→−∞ 𝐵→∞

exists, then the value of the original integral would exist.

So, with the limit points these are the three types of how to say improper integral, we can
come across. Now, let us evaluate an example. So, we have given the how to say a few

91
definitions, and now we are going to as work out few examples. So, let us go to a new
page.

(Refer Slide Time: 17:30)

Example, sometimes I am not good at keeping track of the examples, so I will just write E
∞ 𝑑𝑥
X ok. So, the very first question is does the improper integral ∫0 exist. So, of course
1+𝑥 2

it is an improper integral, because you have the upper limit as infinity.

Now, if the upper limit is infinity, then it falls into the definition 2 criteria. And therefore,
we know that the value of this integral let us write it as I the value of this integral would
𝐵 𝑑𝑥
actually be equal to this lim ∫0 .
𝐵→∞ 1+𝑥 2

𝑑𝑥
Now, integral of 1+𝑥 2 is tan−1 𝑥, so we will write tan−1 𝑥here, then integral from 0 to B.

This will turn into lim [tan−1 𝐵 − tan−1 0] now tan−1 0 is 0. And tan−1 𝐵 , when B goes
𝐵→∞
𝜋 𝜋
to infinity will go to 2 , so the value of the integral is 2 . And therefore, this improper

integral, it converges or it exists. So, from here we can say that I exist or it I is convergent
or it is I is convergent all right. So, this was a fairly simple example to verify.

92
(Refer Slide Time: 19:58)

Next let us consider an example of type integral let us say verify the convergence or
0
existence, we can use either one of these terms, existence of integral ∫−∞ 𝑒 𝑥 𝑑𝑥.

So, let us start. Now, this one falls into the category of type-1 one. So, we can write 𝐼 =
0
∫−∞ 𝑒 𝑥 𝑑𝑥. Now, we know that from definition 1, where is that so here. So, we know that
if we have an integral of this type, then the value of that integral will be the value of this
0
limit here. So, we will do exactly the same thing. So, I can write lim ∫𝐴 𝑒 𝑥 𝑑𝑥.
𝐴→−∞

Now, if you integrate 𝑒 𝑥 , then will obtain basically 𝑒 𝑥 integral from A to 0. Now, this can
be written as lim [1 − 𝑒 𝐴 ] all right.
𝐴→−∞

So, now when 𝐴 → −∞ , the 1st term is constant, so it will remain unaffected. And the
2nd term, when A goes to minus infinity basically this whole thing will go to 0, because e
to the power minus infinity will be 1 by e to the power infinity, and that will be basically
0. So, this will be (1-0), so 1. And from here, we can say that I is convergent or I exist
convergent all right. So, this is one such example, where you can use the definition of type-
1. Now, let us play with this example this integral here a little bit. And then we can say
that the integral is actually divergent.

93
(Refer Slide Time: 22:33)

So, let us consider an another example, verify the convergence the convergence of integral
∞
∫0 𝑒 𝑥 𝑑𝑥. So, we have to verify the convergence of this integral here. So, this one is again
of type-2 type-1 2, so that means, where the upper limit is infinity.

∞
So, I can write our integral as I, which is equals to ∫0 𝑒 𝑥 𝑑𝑥. And then I can use the limit
𝐵
definition, so I can write it ∫0 𝑒 𝑥 𝑑𝑥 . And then this can be written as lim , the integral of
𝐵→∞
𝑥 𝑥
𝑒 would remain 𝑒 and this one will be from 0 to B.

Now, we have lim [𝑒 𝐵 − 𝑒 0 ], 𝑒 0 is again 1. And here when B goes to infinity 𝑒 𝐵 will
𝐵→∞

also diverge. So, this whole thing will diverge. So, it diverges that means, this increases,
how to say beyond all bounds actually as B goes to infinity that means, this integral I here,
this integral I here, it does not converge to a finite limit. So, it does not converge into any
finite number or the limit does not exist. And therefore, from here our integral I is divergent
or it does not exist. So, you see just by playing with the limits same integrand can be shown
to be divergent all right.

94
(Refer Slide Time: 24:49)

∞ 𝑑𝑥
Now, let us consider an another example. Let us say evaluate ∫1 , if it converges. So,
𝑥2
∞ 𝑑𝑥 𝐵 𝑑𝑥
here I can write I as ∫1 . And now I can write it as lim ∫1 And if I integrate the
𝑥2 𝐵→∞ 𝑥2
1
integrand, then I will end up with [− 𝑥]1𝐵 .

1 1
And this one can ultimately be written as lim [− 𝐵 + 1]. So, when B goes to infinity 𝐵 will
𝐵→∞

go to 0, and therefore will end up with 1. And this implies that our integral I is again
convergent. So, any one of these examples if we consider, then we can see that if either
one of the limits is infinity.

Then all we have to do is to consider the limit form of that of that integral, and then evaluate
the integral like the way we do, and then finally pass on the limit, and the value of that
limit will be the value of the integral. So, if it is finite, then in that case the integral is
convergent; and if it is not, then it is divergent. And an integral is can never oscillate, so a
integral is not oscillatory, so either it is divergent or it is convergent. And in order to do
that we can verify in this fashion or we will also look into some tests that will assure the
convergence.

95
(Refer Slide Time: 27:08)

Now, and another example could be of type, where both of our limits. So, let us write
verify the convergence. So, verify the convergence of this integral here. So, I hope you got
the statement, what I am trying to say. So, verify the convergence of this integral.

∞ 2
So, let us write this integral as 𝐼 = ∫−∞ 𝑥𝑒 −𝑥 𝑑𝑥 ; so, again this integrand here. We
basically choose any point between −∞ to∞ , so that we can split this integral into 2
0 2 ∞ 2
halves. So, I choose ∫−∞ 𝑥𝑒 −𝑥 𝑑𝑥 + ∫0 𝑥𝑒 −𝑥 𝑑𝑥 . We have to make sure that in this
interval the function is not unbounded.

So, whatever the point, I choose now in between those two points the function should not
be unbounded here also. So, if you see −∞ between to 0, it does not matter whatever we
substitute, it will always remain finite and similarly from 0 to ∞ . So, we have to choose
this point in such a way that the function or the integrand remains bounded.

Now, I can call this one as 𝐼1 , and I can call this one as 𝐼2 . And then I evaluate 𝐼1 , so next
0 2
we can evaluate 𝐼1 , which is ∫−∞ 𝑥𝑒 −𝑥 𝑑𝑥 . And if I follow the previous definition, then I
0 2
can write it as lim ∫𝐴 𝑥𝑒 −𝑥 𝑑𝑥 . This is a traditional method of substitution example.
𝐴→−∞

So, here if I substitute 𝑥 2 = 𝑧, then basically I will obtain 𝑥𝑑𝑥, and then you do basically
1 0
the integral integration here. And this will reduce to lim ∫ 𝑒 −𝑧 𝑑𝑧 yeah.
2 𝐴2
𝐴→−∞

96
 1 2
And if I substitute x=0, so this will be limit A goes to −∞, this will be 2 (−1 + 𝑒 −𝐴 ).

Now, when A goes to minus infinity, then this will be actually e to the power −∞.

(Refer Slide Time: 30:40)

1 1
So, ultimately this will be our the value will be 2so this will be our − 2. And similarly, we
∞ 2
can evaluate similarly we can evaluate 𝐼2 = ∫0 𝑥𝑒 −𝑥 𝑑𝑥 , we can write it as
𝐵 2
lim ∫0 𝑥𝑒 −𝑥 𝑑𝑥. And if we follow the similar steps like 𝐼1 , then this will give us the value
𝐵→∞
1
as 2.

1 1
So, therefore our capital I is basically 𝐼1 + 𝐼2 . Now, 𝐼1 is − 2, and 𝐼1 is 2 , so ultimately the

value is 0, so that means, the value of this integral is nothing but 0. And since 0 is finite,
we can say that I is convergent or the integral I exist.

So, we have seen different types of examples in this case, and where the both lower limit
upper limit or both the limits can be 0, and we also worked out for example. So, in the next
lecture, we will look into some integrals of type-2, I will give you an explanation what do
we mean by type-2. So, we will stop here for today, and see you in next lecture.

Thank you.

97
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture - 10
Improper Integral (Contd.)

Hello students. So, we will start our next section of Improper Integral. And in this section
we will basically look into the improper integral of type II. And the type II type of improper
integral are nothing but where you have the discontinuity in the function f.

(Refer Slide Time: 00:45)

So, to define that let us write type II type of improper integral, so if the function f has
𝑏
infinite discontinuity only at left hand end point a. Then by ∫𝑎 𝑓(𝑥)𝑑𝑥; we shall mean that
𝑏
lim+ ∫𝑎−𝜀 𝑓(𝑥)𝑑𝑥, where 0 < 𝜀 < 𝑏 − 𝑎. So that means, if we have an infinite
𝜀→0

discontinuity let us say at the left endpoint. At the left endpoint then in that case we add a
little bit epsilon where epsilon is a arbitrary small positive real number to the left endpoint
and then this will give us the value of the integral.

So, we just get rid of that point where the discontinuity happens and then we try to evaluate
the integral and afterwards we make epsilon go to 0, and that limit will actually be the limit
of our improper integral.

98
Similarly let us say if we have a discontinuity. So, I am avoiding the whole statement. So,
what I am trying to say is that if we have infinite discontinuity, so infinite discontinuity at
the point x=b or the right hand endpoint then in that case the value of the integral
𝑏 𝑏−𝜀
∫𝑎 𝑓(𝑥)𝑑𝑥 = lim+ ∫𝑎 𝑓(𝑥)𝑑𝑥. So that means, we again we cannot go beyond the
𝜀→0

interval b. So, we have to stay inside the interval otherwise the range of integral would be
beyond the interval [a,b].

So, what we are doing is we just get how to say get rid of that point in a way that we
substract a small epsilon from b, and that will give us the value of the integral b minus
epsilon doing this b minus epsilon here. And that will, evaluating this limit will give us
𝑏
the value of the integral ∫𝑎 𝑓(𝑥)𝑑𝑥.

𝑏
And another type would be let us say we have an interval ∫𝑎 𝑓(𝑥)𝑑𝑥 . So, this interval
will have a point. So, f(x) has a discontinuity, so within this interval f(x) has discontinuity
at x=c. So that means, in between that interval a to b there exist a point c where the function
is not continuous.

(Refer Slide Time: 04:35)

So, there exist a point c where the function is not continuous. And then what we will do?
𝑏 𝑐 𝑏
We will write the integral ∫𝑎 𝑓(𝑥)𝑑𝑥 = ∫𝑎 𝑓(𝑥)𝑑𝑥 + ∫𝑐 𝑓(𝑥)𝑑𝑥. Like we did for the
limits where both the limits were infinite. So, here also the point of discontinuity we can
break the integral into the points of discontinuity.

99
 𝑐−𝜀 𝑏
And now here I can write lim+ ∫𝑎 𝑓(𝑥)𝑑𝑥, and here I can write lim+ ∫𝑐+𝜀 𝑓(𝑥)𝑑𝑥 or
𝜀→0 𝜀→0
𝑏
lim ∫ 𝑓(𝑥)𝑑𝑥. So, and if both of these 2 limits exist then in that case that will give us
𝛿→0+ 𝑐+𝛿

the value of the integral.

And suppose if you have more than one point of discontinuity let us say if there is a point
𝑐 𝑑 𝑏
called d then we basically write ∫𝑎 . + ∫𝑐 . + ∫𝑑 .so that means, if we have finite number of
infinite discontinuity we can break the integral into finite number of sub integrals. And at
each one of these sub integrals we put the limit of this type lim+ .and then we evaluate those
𝜀→0

finite sub integrals and that will give us the value of the integral.

So, now let us work out few examples based on the above definitions. So, our first example
1 𝑑𝑥
would be evaluate 𝐼 = ∫0 if it converges. So, here we can see that 0 is the point of
𝑥

infinite discontinuity. So, note that 0 is the point of infinite discontinuity or only point of
infinite discontinuity we can use the term only is the only point of infinite discontinuity.
1 𝑑𝑥
And so I can write my I as lim+ ∫0+𝜀 .
𝜀→0 𝑥

(Refer Slide Time: 07:24)

So, then this will be lim+[log 𝑒 𝑥]1𝜀 and then this can be written as lim+ [log 𝑒 1 − log 𝑒 𝜀]
𝜀→0 𝜀→0

Now, log 𝑒 1 is 0 and log 𝑒 𝜀; now log 𝑒 1 is 0, but this here; so − log 𝑒 𝜀. So, when epsilon
goes to 0 this log 𝑒 𝜀 will go to minus infinity. So that means, the whole term will go to

100
+∞. So, this log 𝑒 𝜀 will go to −∞ as 𝜀 → 0+ . So, this whole thing will go to a +∞ that
means, this whole thing is divergent. So, it diverges and from here we can say that our
integral I is divergent or does not exists, all right.

Next, let us consider an example where we can at least show the convergence. So, let us
1 𝑑𝑥
consider 𝐼 = ∫0 √1−𝑥 2
. So, we have to always identify the type of improper integral. So,

whether you have the discontinuity in the function or whether you have improperness in
the limit.

So, once you identify that, so here in this case the limits are all right because the limits are
finite 0 to 1 however, we can see that the function is discontinuous at x=1. As I was saying
there is a strong possibility you may have a function which is discontinuous at 2 points or
3 points and so on and then in that case we have to break this integral here into that many
number of sub integrals.

1−𝜀 𝑑𝑥
So, here what would happen is, we will first get rid of this discontinuous ∫0 √1−𝑥 2
.

(Refer Slide Time: 10:03)

1−𝜀 𝑑𝑥
Now, we can write this as lim+ ∫0 √1−𝑥 2
. Here we can put a small comment that, note
𝜀→0

that 1 is the only point of discontinuity, all right.

101
So, next we can write it as the value of the integral. So, here we have a √1 − 𝑥 2 . So, this
1−𝜀
will be basically sin−1 𝑥. So, [sin−1 𝑥]1−𝜀
0 . So, ∫0 .and then we substitute the limits. So,
this will be [sin−1 (1 − 𝜀) − sin−1 0]. So, sin−1 (1 − 𝜀) when 𝜀 goes to 0 this will be
𝜋
basically sin−1 1 minus sin−1 0 and sin−1 1 is 2 . So, therefore, the value of the integral I
𝜋
is actually 2 and from here we can say that our integral I is convergent, all right.

So, this is a one such example where we have subtracted the epsilon from the upper limit
and we just have to calculate the integral and then pass on that epsilon goes to 0 and that
will give us the value of the integral I.

(Refer Slide Time: 11:56)

So, we have covered the examples of both of the types. Next, another example could be
1 𝑑𝑥
let us say of this type. So, evaluate integral ∫−1 𝑥 3 if it exists. So, first of all let us see the

type of this integral.

So, here the range of integration is [-1,1]. So, the range is pretty much fine because they
are finite. Now, at 𝑥 = −1 and 𝑥 = 1 this function is also not unbounded. So, we can say
that the function is not unbounded between [-1,1], however that is not true because at x=0
which is a point between [-1,1] this function is unbounded. So, now, it falls into the
category of this type. So, it falls into the category of this type, this here. And then we have
to split the whole interval integral into 2 sub integrals. So, let us go back to that slide.

102
So, here we can write note that x =0 is the point of infinite discontinuity, all right. And
0 𝑑𝑥 1 𝑑𝑥
then I can write my integral I as ∫−1 𝑥 3 + ∫0 and now I can write these 2 integrals as
𝑥3

𝐼1 + 𝐼2 , where is 𝐼1 this one and 𝐼2 is this one.

0 𝑑𝑥
So, next we will look into our integral 𝐼1 . So, 𝐼1 is ∫−1 𝑥 3 . So, this can be written as

lim .and since the discontinuity has the upper limit I will subtract a little bit epsilon from
𝜀→0+
0−𝜀 𝑑𝑥
there lim+ ∫−1
𝜀→0 𝑥3

(Refer Slide Time: 14:37)

1
So, then the value of the integral, the value of the integral would be [− 2𝑥 2 ]−𝜀
−1 and if I

1 1
substitute the limit then this will be lim+ [− 2𝜀2 + 1]. Now, when 𝜀 → 0+ this 2𝜀2 this will
𝜀→0

not exist. So, this will diverge to infinity. So, this whole thing, so this will diverge to
infinity, so this whole thing diverges.

And if since we are writing the 2 sub integrals, if any one of them if both of them are
convergent then the whole thing is convergent, but if any one of them is divergent and the
other one is convergent then the whole thing will again be divergent and if both of them
are divergent then the whole thing will again be divergent.

So, here our 𝐼1 is divergent similarly we can see that the integral 𝐼2 , 𝐼2 which is basically
1 𝑑𝑥 1 1 𝑑𝑥
∫0 which is equals to lim+ ∫0+𝜀.so that means, +𝜀 so that means, + ∫𝜀 And if we
𝑥3 𝜀→0 𝑥3

103
 1
evaluate this integral then here we will obtain basically lim+ .. This one will be − 2 +
𝜀→0
1
and then this one is also divergent. So, this is also diverges.
2𝜀 2

So that means, therefore, our 𝐼 = 𝐼1 + 𝐼2 they are both divergent and this implies that our
integral I is altogether divergent or it does not exist, all right. So, just splitting the interval
into 2 sub intervals we evaluate each one of the sub intervals and if any one of them is
divergent or if both of them are divergent then in that case that the whole interval I would
be divergent and if both of them are convergent then the whole into integral I would be
convergent. So, this is one way to check whether the given integral is convergent or not.

(Refer Slide Time: 17:40)

2 𝑑𝑥
Next, similarly another example could be evaluate ∫−1 if it exists. So, I will just give
𝑥

some hints for this example. So, I can write I as limit epsilon goes to or first let us write
the integral into 2 sub integrals. So, here again 0 is the point of discontinuity we can see
0 𝑑𝑥 2 𝑑𝑥
that clearly. So, I can write it as ∫−1 + ∫0 , and then we can write it as
𝑥 𝑥
−𝜀 𝑑𝑥 2 𝑑𝑥
lim+ ∫−1 + lim+ ∫𝜀 .
𝜀→0 𝑥 𝜀→0 𝑥

And then we just evaluate each one of these individual integrals and then we can be able
to see whether one of them or both of them are divergent or convergent, based on that we
can draw our conclusion. So, this is pretty much straight forward example.

104
(Refer Slide Time: 19:39)

So, next we will look into integral of type test of convergence. So, next we will look into
test of convergence. So, suppose you are given an integral and if it is not straightforward
to evaluate via doing those limit procedures or via how to say the traditional method to
calculate the improper integral, then we look for some tests that will help us reduce
calculating those limits and integrals and just using these tests we will assure whether our
given improper integral is convergent or not. So, one such test is called comparison test.

𝐵
And it says that if f(x) be a non-negative integrable function when 𝑥 ≥ 𝑎 and ∫𝑎 𝑓(𝑥)𝑑𝑥 is
∞
bounded above for every B>a then integral ∫𝑎 𝑓(𝑥)𝑑𝑥 will converge otherwise it will
diverge to +∞. So, what it says is that if we have a non-negative function f(x) which is
integrable for 𝑥 ≥ 𝑎, and integral from a to B where capital B can be treated as the upper
limit or upper limit of this range of integration is bounded then for every B>a then this
∞
integral ∫𝑎 𝑓(𝑥)𝑑𝑥 will also converge, otherwise it will diverge. So, this is the
comparison test of let us say, comparison test of type 1 or simply just comparison test 1,
we can write it as comparison test 1. So, this is first comparison test.

And second comparison test would be, the second type would be if f(x) and g(x) be
integrable functions when 𝑥 ≥ 𝑎, such that 0 ≤ 𝑓(𝑥) ≤ 𝑔(𝑥) then, integral
∞ ∞ ∞
∫𝑎 𝑓(𝑥)𝑑𝑥 converges if integral∫𝑎 𝑔(𝑥)𝑑𝑥 converges and integral ∫𝑎 𝑔(𝑥)𝑑𝑥 diverges
∞
if integral ∫𝑎 𝑓(𝑥)𝑑𝑥 diverges.

105
So, what does it mean is if we have 2 functions f(x) and g(x) and this follow some kind of
comparison. So, this here this inequality is nothing, but a comparison. So, if 𝑓(𝑥) ≤
𝑔(𝑥) then in that case if the function g(x) is convergent, then in that case we can say that
the if the integral of the function g(x) is convergent, then we can say that the integral of
the function f(x) is convergent and if the integral of the function f(x) is divergent then the
integral of the function g(x) is also divergent. So, these are the 2 comparison test in a way
where you compare the function f(x) or the given integral of f(x) with integral of g(x) by
this fashion.

So, you have to make sure that whether 𝑓(𝑥) ≤ 𝑔(𝑥) or not. We need to find out a function
basically g(x), that will sort of how to say satisfy this inequality. And based on that we can
check whether the function g(x) is convergent or not and with the help of which we can be
able to say that whether the function f(x) is convergent or not. So, this is one such type of
comparison test.

(Refer Slide Time: 25:56)

There is a third type, third type is limit test, it is called limit test. So, it says that let f(x)
and g(x) be integrable functions, for 𝑥 ≥ 𝑎 and g(x) be positive, g(x) be positive then if
𝑓(𝑥) ∞ ∞
lim = 𝜆 ≠ 0 then integral ∫𝑎 𝑓(𝑥)𝑑𝑥 and integral ∫𝑎 𝑔(𝑥)𝑑𝑥 both converge
𝑥→∞ 𝑔(𝑥)

absolutely or diverge. So, they are both either convergent or they are both divergent.

106
So, this is a third type of comparison test where you basically evaluate this limit and if that
limit is non-zero then in that case we can talk about the convergence or divergence of these
2 integrals. So, either they will converse together or they will diverse together. So, all we
have to do is to find a function g(x) such that we can have a non-zero limit and then we
can talk about the convergence.

So, based on these 3 types of comparison test we will work out few examples to make the
concept clear and will do that in our next lecture. So, I will look forward to it and see you
in next class.

Thank you.

107
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture - 11
Improper Integral (Contd.)

Hello students. So, upon till last class we looked into different types of Improper Integrals,
where you have the either the upper limit or the lower limit as infinity or you have some
kind of infinity discontinuity in the function itself. This all those types of integrals are
categorized as improper integral. We also worked out few examples and we also looked
into some test or we defined some test which will assure the convergence of an improper
integral.

Before working out examples on those test I will give you one or two more examples about
improper integral where, the convergence depends on certain type of factor involving in
the integral itself. So, let us see what do I mean by those examples.

(Refer Slide Time: 01:19)

∞ 𝑑𝑥
So, to start with example; test the convergence of integral of type ∫0 , where n is any let
𝑥𝑛
∞ 𝑑𝑥
us say here we have ∫𝑎 where n is any positive number. So now, in this case what we
𝑥𝑛
∞ 𝑑𝑥
have is one of the limits as infinity. So, I will write this integral as 𝐼 = ∫𝑎 where n is a
𝑥𝑛

positive number. So, for we do not know for what value of n this integral would converge

108
or diverge. So, at the moment we have to just test the convergence and see for what values
we can get the convergence.

𝐵 𝑑𝑥
So now, if I write it in our traditional improper integral forms. So, I will write as lim ∫𝑎 ,
𝐵→∞ 𝑥𝑛

now evaluating this integral is fairly easy. So, we know the value of the integral would be
1
𝑥1−𝑛 and then I will close this times. So, this the integral would run from so, arrange
1−𝑛
1 1 1
would run from a to B. And, now if I substitutes I will get 1−𝑛 lim [𝐵𝑛−1 − 𝑎𝑛−1 ], where n
𝐵→∞

is any positive number.

Now, we can see here that this integral so, first of all it does not matter whether B goes to
infinity or not, if n is 1 then in that case this here would be undefined. So, if n is 1 then
this here would be undefined. So, our very first criteria is that n cannot be 1; now let us
check whether so, the first observation; first observation is n cannot be 1.

(Refer Slide Time: 04:12)

So, this is our first observation let us say, I can write it here as observation. So, we have
got at least one value of n where this integral would not be defined or that limit would not
exists. Now, 2nd observation is, let us go back to the previous slide. Now, 2nd observation
is if B goes to infinity and if n <1 then in that case this will become B to the power
something negative and that will go in the numerator.

109
And, then it will become B to the power something positive and since it B goes to infinity
this whole thing will go to infinity if n <1. So, in that case this n cannot be less than 1
because, if n is less than 1 then this whole thing will go to infinity. So, the 2nd observation
is if n is less than 1, then the limit diverges or limit diverges to plus infinity. So; that means,
if n ≤1 then in that case; this limit here would not exists.

And therefore, if we want to test the convergence therefore, the limit exists only when or
only for n>1; let us check. So, if n > 1 then this will be 1 by B to the power some positive
and if B goes to infinity then this term will go to 0, a is any for the moment let us assume
that a is also any positive number. So, a is any positive number and if n is positive then
this whole thing would also be finite and ultimately the limit would exists. So; that means,
here this integral I this implies that so, or this means that; the integral I; exists only when
n >1. So; that means, it is convergent exists or we can also write it is convergent for n >1.

You see in this particular example, not only we had to be careful about the improperness
of the integral, we also have to be careful about the ranges for this n for which we can see
whether it is convergent or divergent . So, this was a very nice example where, we could
see that for n >1 this integral I, this improper integral was convergent and for n ≤1 this
integral I is also divergent. So, that is we have verified here.

And another example in this context could be you may come across these type of examples
𝑏 𝑑𝑥
in your study. So, test the convergence of integral ∫𝑎 . So, again here so, here none
(𝑥−𝑎)𝑛

of the limits are infinity. So, the limits are finite, but the function is creating a problem.
So, at the left end point this function has infinite discontinuity. So, we can write that here
a is the only point of infinite discontinuity and another thing is we have to be careful about
this n. So, for what values of n this integral would be convergent or would be divergent.

So, a is the only point of discontinuity. So, we know that what we do in this situations, we
𝑏 𝑑𝑥
first of all write I and then the value this integral ∫𝑎 would be equal to value of these
(𝑥−𝑎)𝑛
𝑏 𝑑𝑥
of the limit lim+ ∫𝑎+𝜀 (𝑥−𝑎)𝑛. So, we add a small epsilon just to get away from the
𝜀→0

discontinuity. So, we add a little bit epsilon in that left end point and then we write this as
𝑏 𝑑𝑥
∫𝑎+𝜀 (𝑥−𝑎)𝑛 and now we can integrate.

110
(Refer Slide Time: 09:41)

So, if we integrate we can write this as lim+ .; once we integrate this will reduce to
𝜀→0
1 1
lim [ (𝑥 − 𝑎)1−𝑛 ]𝑏𝑎+𝜀 And, then lim [(𝑏 − 𝑎)1−𝑛 − 𝜀 1−𝑛 ] alright.
𝜀→0+ 1−𝑛 1−𝑛 𝜀→0+

So, now what we can see again like the previous example, if n is 1 so, we can write as
small observation. So, 1st observation is if n is 1 then this whole thing will be undefined
or this whole thing would not exist. So, n cannot be 1 that is the first observation. Now,
what is the 2nd observation here? if n>1 then this whole thing will become negative. So,
epsilon to the power negative and if it is epsilon to the power negative then we can write
it as 1 by epsilon to the power that number that real number. Now, if epsilon goes to 0 then
the whole 1 by epsilon will go to infinity and then this whole limit will become undefined.
So that means, n cannot be greater than 1 because, then in that case this limit will diverge
to infinity.

So, the 2nd observation is n cannot be greater than 1 like we wrote down here in the
previous example. So, n cannot be greater than 1 because, since for n greater than 1 the
limit diverges to plus infinity; now what happens for n less than 1. So, if n is less than 1
then in that case of course, this is a positive quantity if n is less than 1 then this one will
also be defined. And, if n is less than 1 then this will be epsilon to the power some positive
real number and when epsilon goes to 0 epsilon to the power that positive real number will
go to 0. So, then in that case this limit will be defined.

111
So, we can write therefore, for n less than 1 the limit let us call this as equation 1, the limit
in equation 1 exists. This means that; the integral I, the integral or the improper integral I
is convergent for n less than 1. And of course, it will be divergent for all n ≥1. So, this one
was also very nice and interesting example, where we had a discontinuity in the function
at one of the end points. And, that integral I mean the convergence of that integral involved
this n as a factor.

So, this n plays a very important role when we talked about the convergence of this
improper integral and for n less than 1, this improper integral is convergent where as if n
is greater or equal to 1 we just saw that it will be divergent. So, this is an another type of
examples where, how to say the convergence of the integral depended on some other
factors involved in the integrand.

(Refer Slide Time: 14:10)

So, this is something I wanted to show you all and now we are ready to look into the
examples of based on comparison test. So, we defined the comparison test as of this type.
So, the first comparison test says that if you have two functions f(x) and g(x) and if they
satisfy this type of inequality. And, integral of this g(x) is convergent, then in that case we
can talk about the convergence of the integral involving f(x) and then there was a limit test
where we evaluate the limit.

So, first of all we will look into we will look into the example on comparison test. So, test
∞ 𝑐𝑜𝑠𝑥𝑑𝑥
the convergence of integral ∫0 . So, this one was our comparison test of inequality
1+𝑥 2

112
type and the second one is comparison test of limit type. So, let us see what kind of
comparison test we can use here. So, looking at cosx we know that; cosx≤1 for all x in R.

So, since we know this it is pretty much intuitive to use comparison test of inequality type
because, sometimes looking at the integral you can able to make out that what kind of
comparison test I need to use. So, just look at this integral and looking at the integrand you
may get an idea that since, cosx≤1 I may have to use comparison test of inequality type.

𝑐𝑜𝑠𝑥 1 𝑐𝑜𝑠𝑥
And in order to do so, let us define f(x) is 1+𝑥 2 and g(x) as 1+𝑥 2. So, then our f(x) is 1+𝑥 2.
1
And, from here f(x)≤g(x)= right because, cosx is less or equal to 1 So, we have this
1+𝑥 2

inequality. So, we found a function g(x); which is what to say dominating this function
f(x). So, which is acting like an upper bound for this function f(x); so, the first criteria
satisfied. And, the second criteria is to check whether this integral g(x) is convergent or
∞ 𝑑𝑥 𝐵 𝑑𝑥
not. So, ∫0 right. And, this can be written as lim ∫0 .
1+𝑥 2 𝐵→∞ 1+𝑥 2

We have just worked out an example like this in our previous class on improper integral.
𝜋
So, this is basically tan−1 𝑥 and when B tends to infinity this whole thing will go to 2 . So
∞ 𝜋
that means, this integral ∫0 𝑔(𝑥)𝑑𝑥 is convergent and it is converging to and it is also
2

dominating the function f(x). So, from here we can say that so, we can put all this into any
∞
statement. So, the statement is integral ∫0 𝑔(𝑥)𝑑𝑥 is convergent or exists and also 𝑓(𝑥) ≤
∞
𝑔(𝑥). So, from here we can say that integral ∫0 𝑓(𝑥)𝑑𝑥 is also convergent.

So that means, just looking at the function g(x) and its properties we can talk about the
convergence of this function f(x). So, all we have do is to find an inequality like this. And,
if we can able to find or locate an inequality like this we can talk about its convergence
and that will assure the convergence of the given improper integral.

113
(Refer Slide Time: 19:57)

Now, let us look into an another example which is basically based on the limit type. So,
the limit type example is of this type. So, example 2 test the convergence, I am avoiding
this statement and I am just writing the integral. So, we have to test the convergence of the
1 𝑒 −𝑥 𝑑𝑥
integral ∫0 . So, we have to test the convergence of this integral. Now, this integral is
𝑥 1−𝑎

only undefined or has infinite discontinuity for a<1 at the point x=0.

Because, if a >1 then this whole thing will go to numerator and then this integrand is
defined, I mean if does not have any kind of infinite discontinuity. So, it is only creating
problem when a < 1 so, let us write it. So, the solution here 0 is the only point of infinite
discontinuity for a <1. Because, for a ≥1 this integrand is always defined.

𝑒 −𝑥
So, now we will only check the convergence for a <1. So, let 𝑓(𝑥) = 𝑥 1−𝑎. And, we take
1
g(x) as then we can evaluate the limit the point, where we have the discontinuity
𝑥 1−𝑎
𝑓(𝑥) 𝑒 −𝑥
which is 0. So, let us write lim+ 𝑔(𝑥), then this is basically lim+ 𝑥 1−𝑎 . 𝑥1−𝑎 . So, this whole
𝑥→0 𝑥→0
−𝑥
thing will reduce to lim+ 𝑒 which is 1≠ 0.
𝑥→0

So, in the comparison test of limit type the limit first of all has to be non-zero. So, this is
a non-zero limit and this is basically 𝜆 from the definition of the comparison test of limit
type. Now, first thing is assured. Now, second thing is next we have to check whether
1
integral ∫0 𝑔(𝑥)𝑑𝑥 is convergent or not, because if they do then in that case both f(x) and

114
 1 1 𝑑𝑥
g(x) would converge together or diverse together. So, this one will be ∫0 𝑔(𝑥)𝑑𝑥 is ∫0 .
𝑥 1−𝑎
1 𝑑𝑥
So, this integral it can be seen since a <1, then in that case we have lim+ ∫𝜀 .
𝜀→0 𝑥 1−𝑎

Then we do the integration like we did before and then we substitute epsilon in the value
of the integral and then we make epsilon goes to 0. So, this is very straight forward to do
here and we can be able to see that this integral here converges, it is very straight forward
to see that this integral converges if (1 − 𝑎) < 1. So that means, 𝑎 > 0 am I right. So, a
1
goes on that side so, a >1. So, which means that; integral ∫0 𝑔(𝑥)𝑑𝑥 is convergent for 𝑎 >
0 so, for all 𝑎 > 0.

(Refer Slide Time: 25:08)

And hence by comparison test by comparison; so, I am just writing the short form of it.
1 𝑒 −𝑥
So, by comparison test of limit type; integral ∫0 𝑓(𝑥)𝑑𝑥. So, f(x) was 𝑥 1−𝑎 is convergent;

for 𝑎 > 0. So, although the function has 0 has the point of infinite discontinuity for a <1,
we just showed that the integral is always convergent why I am comparison test of limit
type for all 𝑎 > 0. So, this is one such case where we use comparison test of limit type,
we also saw a comparison test of inequality type, we can work out one or two more
examples.

So, let me just write the examples and I am pretty sure you can be able to do it by yourself,
I will just give some hint. So, we have to check the convergence of integral of this type.

115
 1
So, from here we can easily establish these inequalities. So, 0 ≤ 𝑒 𝑥+1 because, it is always
1
positive for all x. And for all x between 0 to infinity and then this one ≤ 𝑒 𝑥 This is also

very straight forward and then this one is ≤ 𝑒 −𝑥 .

Now, if I assume this as our function g(x) and this is our function f(x) then in that case all
∞
we have to check is whether this integral ∫0 𝑔(𝑥)𝑑𝑥 is convergent or not from the
comparison test of inequality type. Because, we have got the inequality, all we have to
show is that this integral is convergent or not and it is very straight forward. So, from here
it is very straight forward and easy to show that this integral here is convergent
∞
∫0 𝑒 −𝑥 𝑑𝑥 is convergent. And, if this integral is convergent; then from the comparison test
∞ 𝑑𝑥
of inequality type we can write that ∫0 is also convergent.
𝑒 𝑥+1

So, handling examples of this type basically, I can also give you examples where we have
limit to evaluate the limit very quickly. So, let us say we have something like integral form.
𝜋 √𝑥
example of this type integral ∫0 𝑑𝑥 . So, here we will basically divide this integral. let
𝑠𝑖𝑛𝑥
1 √𝑥 𝜋 √𝑥
us say from between two sub integral. So, ∫0 𝑑𝑥 +∫1 𝑑𝑥. And, then what we do is
𝑠𝑖𝑛𝑥 𝑠𝑖𝑛𝑥

in order to check the convergence at 0 because, 0 is the only point where this function is
having infinite discontinuity. And, in the second sub integral 𝜋 is the only point where this
function have infinite discontinuity.

(Refer Slide Time: 29:01)

116
So,we can check the convergence first of all at 0. So, convergence at 0 and in order to do
√𝑥 1 𝑓(𝑥)
that we consider our function 𝑓(𝑥) = 𝑠𝑖𝑛𝑥 and 𝑔(𝑥) = . And, then we do lim+ 𝑔(𝑥). And
√𝑥 𝑥→0

we can be able to see that this limit is basically 1 and we can also be show that integral
1
∫0 𝑔(𝑥)𝑑𝑥 is convergent and therefore, the first sub integral would be convergent.

Similarly, we can check the convergence at 𝜋 and in order to do that I will take my function
√𝑥 1
𝑓(𝑥) = 𝑠𝑖𝑛𝑥. And I will take my function 𝑔(𝑥) = 𝑥−𝜋, I can again evaluate the limit and it
𝑓(𝑥)
can be shown. I can be able to show that this lim− 𝑔(𝑥) = −√𝜋. And I can also be able to
𝑥→𝜋
𝜋
show that integral ∫1 𝑔(𝑥)𝑑𝑥 is actually divergent. So, this I can be able to show that.

So, since this one is divergent our original sub integral this one will be divergent; however,
our sub integral 𝐼1 is convergence. So, convergence plus divergence means that the whole
integral is divergent. So, from here I can write 𝐼 = 𝐼1 + 𝐼2 . And, since 𝐼2 is divergent and
𝐼1 is convergent the whole thing will be divergent. Therefore, by this comparison test of
limit type I was being able to show that the given integral I is divergent. And, we will stop
our lecture for today here. And, in the next class we will start with new test and we will
work out few examples as well.

Thank you.

117
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture - 12
Improper Integral (Contd.)

Hello students. So, up until last class we looked into comparison test and examples
involving comparison tests. There is or there are a couple of more tests in the context of
Improper Integral where we use them to talk about the convergence of those improper
integral. So, I believe we have covered sufficient examples on comparison tests and you
will be able to work out few more examples in your [shine/assignment] assignment. And
today we will start with 𝜇-test.

(Refer Slide Time: 00:59)

So, test of convergence using 𝜇- test. So, the 𝜇-test for convergence. So, the statement goes
∞
like this let f(x) be an integrable function for 𝑥 > 𝑎. Then integral ∫𝑎 𝑓(𝑥)𝑑𝑥 converges
∞
if lim 𝑥 𝜇 𝑓(𝑥) = 𝜆 for 𝜇 > 1 and ∫𝑎 𝑓(𝑥)𝑑𝑥 diverges. If lim 𝑥 𝜇 𝑓(𝑥) = 𝜆 which is non
𝑥→∞ 𝑥→∞

zero or ±∞ for 𝜇 ≤ 1.

So, if 𝜇 ≥ 1 then we can say that the given integral improper integral is convergent if it is
not if 𝜇 ≤ 1, but we are still getting the finite value then in that case the given integral is

118
divergent. So, let us see a few examples where we can talk about the convergence of the
given improper integral.

So, example 1 test so, let us consider the similar example to test the convergence of
∞ 𝑐𝑜𝑠𝑥𝑑𝑥 𝑐𝑜𝑠𝑥
∫0 . So, then in that case we can write it so, the solution. So, let 𝑓(𝑥) = 1+𝑥 2 and if
1+𝑥 2
𝑐𝑜𝑠𝑥
I take lim 𝑥 𝜇 𝑓(𝑥) then this will be lim 𝑥 𝜇 1+𝑥 2 . And if I take 𝜇 > 2 or how to say at least
𝑥→∞ 𝑥→∞

𝜇 = 2, then in that case this whole thing will be equal to 1.

(Refer Slide Time: 05:13)

𝑐𝑜𝑠𝑥
So, if I take 𝜇 = 2 . So, let us say if I take lim 𝑥 2 1+𝑥 2 for 𝜇 = 2. Then in that case I can
𝑥→∞
1
be able to write this as 1 by. So, this is basically lim 1 𝑐𝑜𝑠𝑥 and this whole thing
𝑥→∞ 1+ 2
𝑥

𝑥 2 𝑓(𝑥) and this whole thing would converse to or for the time being.

If I assume that instead of cosx if I let us say just start with just to make the example a
little bit simpler because I do not want to complicate it. So, let us start with just 1 instead
of co x if I take just 1 then what would happen. Then if I take 1 here let us say instead of
cosx I have 1 then it will be a little bit straightforward. So, instead of cosx lets say I have
1
1. So, then in that case this will be lim 1 and then the whole limit will be 1 for 𝜇 = 2.
𝑥→∞ 1+ 2
𝑥

119
So, the message which I want to convey is lim 𝑥 𝜇 𝑓(𝑥) = 𝜆 . A lambda which is basically
𝑥→∞

our 1 for 𝜇 = 2 which is greater than 1. So, having this limit here which is of course, finite
for 𝜇 = 2 which is of course, greater than 1, we can be able to apply this 𝜇-test.

So, I just wanted to show you with a simple example. So, having cosx would complicate
1
a little bit. So, let us not go there first of all. So, if I have 1+𝑥 2 we have already worked out

this example several times, but I would like to show the application of 𝜇- test here. So, let
us say you have the improper integral of this time. So, we multiply by 𝑥 𝜇 to the function
f(x) and then I am evaluating this limit.

So, if I choose a 𝜇 = 2 which is greater than 1 then in that case the integral which is in this
∞ ∞ 𝑑𝑥
case. So, then in that case the integral ∫0 𝑓(𝑥)𝑑𝑥 which was integral ∫0 is convergent
1+𝑥 2

by 𝜇- test. So, this is one such example of 𝜇- test for this simple problem. Next we can see
an another example, let us see that.

(Refer Slide Time: 08:33)

So, now we can go to a bit complicated examples.

120
(Refer Slide Time: 08:51)

So, let us consider example and also I would like to point out that in previous class, we
𝑐𝑜𝑠𝑥 𝑐𝑜𝑠2 𝑥
considered an example of this type . It is better to put here. So, if I put 𝑐𝑜𝑠 2 𝑥
1+𝑥 2 1+𝑥 2

here. So, we can test the convergence of this problem so, not the cosx, but 𝑐𝑜𝑠 2 𝑥 so, that
it is rather more how to say then in contrast with the comparison test of inequality type.
∞ 𝑐𝑜𝑠2 𝑥𝑑𝑥
So, the original problem should be stated as ∫0 . Then in that case it will be in
1+𝑥 2

contrast with the with the comparison test of inequality type.

Now, that was just one remark or comment on my previous lecture. Next example is we
test the convergence so, I am not writing the whole statement. So, we test the convergence
∞ 2
of ∫0 𝑒 −𝑥 𝑑𝑥. So, we want to test the convergence of this integral here. So, let us define
2 2
our function f(x) as 𝑒 −𝑥 and next we want to evaluate let us say the limit 𝐿 = lim 𝑥 𝜇 𝑒 −𝑥 .
𝑥→∞

So, here in this case we have a problem at 𝑥 = ∞ because, the upper limit is ∞ and I
2
write 𝐿 = lim 𝑥 𝜇 𝑒 −𝑥 . Now, for the time being if I take 𝑥 2 instead of 𝜇 if I take 2 then
𝑥→∞
2 𝑥2
this is 𝐿 = lim 𝑥 2 𝑒 −𝑥 for 𝜇 = 2. Then this can be written as 𝐿 = lim 2 and next this
𝑥→∞ 𝑥→∞ 𝑒 𝑥
∞ ∞
one is basically ∞ form if you remember from L’Hospital rule. So, this is basically ∞ form.

121
So, I can differentiate both numerator and denominator and this will be written as 𝐿 =
2𝑥
lim 2 , then differentiation of 𝑥 2 would be 2x. And then I can differentiate the whole
𝑥→∞ 2𝑥𝑒 𝑥
1
thing again and this will be lim 2 2 .
𝑥→∞ 𝑒 𝑥 +𝑥2𝑥𝑒 𝑥

1
So, when x goes to ∞ this whole thing will be ∞. So, this whole thing will go to 0 so; that

means, the limit or the 𝜆 of this limit is it is a finite number and it is going to 0 for 𝜇 =
2 which is; obviously, greater than 1. And if it is greater than 1 then from the convergence
of 𝜇-test we can say that the integral is convergent. So, from here we can say that integral
∞ 2
∫0 𝑒 −𝑥 𝑑𝑥 is convergent by 𝜇-test. So, you see it is always about guessing the correct
value for 𝜇.

So, for what value of 𝜇 you can be able to show that the limit is finite and if it is finite then
we have to see for what values of 𝜇 is that happening. So, if it is happening for 𝜇 > 1,
then only the integral is convergent like in this case and if it is happening for 𝜇 ≤ 1 then
in that case the integral is divergent. So, choosing the correct value of 𝜇 plays an important
role here.

(Refer Slide Time: 13:02)

We can so, there is another example where we have the divergence part. So, let us consider
3 3
∞ 𝑥2 𝑥2
the integral ∫0 3𝑥 2 +5 𝑑𝑥 . So, obviously, here we choose f(x) as our 3𝑥 2 +5 and since one of

122
the limit is ∞ that is why it is an improper integral. So, let us find the limit or 𝜆 from our
𝜇 convergence test. So, 𝜆 is basically lim 𝑥 𝜇 𝑓(𝑥).
𝑥→∞

3
1
𝑥2 1
So, then this is basically 𝜆 = lim 𝑥 2 for 𝜇 = 2which is; obviously, less than 1. Then
𝑥→∞ 3𝑥 2 +5

𝑥2 1
this will reduce to lim and this can be written as lim 5 .
𝑥→∞ 3𝑥 2 +5 𝑥→∞ 3+ 2
𝑥

1 1
So, ultimately this value will be 3 for 𝜇 = 2 which is less than 1. Now, this falls into this

category that if 𝜇 ≤ 1 and if the value is non 0 then in that case the integral is divergent.
∞
So, from here we can write that integral from ∫0 𝑓(𝑥)𝑑𝑥 is divergent for 𝜇 < 1 or by 𝜇-
test for 𝜇 < 1 or we can write by 𝜇-test.

So, that is the way we show the divergence as I was saying having this value of 𝜇 is a very
vital in order to ensure the convergence or divergence of the given improper integral. We
will work out one last example of an integral which is very vital in integral calculus.

(Refer Slide Time: 15:40)

So, the example goes like this we will of course, look into a gamma function and beta
∞
function in the next class. But we see that the integral ∫0 𝑥 𝑛−1 𝑒 −𝑥 𝑑𝑥 which is defined as
gamma function is also an improper integral. And here I am just writing n is positive I am
not writing the range for which n is this improper integral is convergent or not we will test
its convergence now.

123
So, for the time being let us take 𝑛 > 0 and then we will say for what values of n this
improper integral is convergent. So, let us test its convergence. Now, let me write 𝐼 =
∞
∫0 𝑥 𝑛−1 𝑒 −𝑥 𝑑𝑥 . So, first of all if 𝑛 > 1, then this integral this integrand is defined.

However, this is still an improper integral because you have one of the limits we have one
of the limits or the upper limit as infinity. Now, when 𝑛 < 1 then in that case this whole
thing will start creating problem because then in that case 0 will be the point of infinite
discontinuity. Since it will become x to the power minus something and then that will
come at the denominator and then when x goes to or at 𝑥 = 0 that denominator that
particular part will be undefined.

So, at 𝑥 = 0 this function or this integrand has an infinite discontinuity. So, we can write
that 0 is the only point of discontinuity for 𝑛 < 1 right next let us take 𝑓(𝑥) = 𝑥 𝑛−1 𝑒 −𝑥 .
1
So, then I can write my integral I as the sum of two integrals ∫0 𝑥 𝑛−1 𝑒 −𝑥 𝑑𝑥 which is the
∞
given function f(x) then ∫1 𝑓(𝑥)𝑑𝑥.

So, this one is our integral 𝐼2 this one is our integral 𝐼1 . So, for integral 𝐼1 we have infinite
discontinuity at 𝑥 = 0 for 𝑛 < 1. So, let us proceed so, convergence for 𝐼1 at 𝑥 = 0. So, 𝐼1
1
is integral ∫0 𝑥 𝑛−1 𝑒 −𝑥 𝑑𝑥 .

(Refer Slide Time: 19:11)

124
 1
Now, in this case I choose let 𝑔(𝑥) = 𝑥 1−𝑛. Then what will be my limit of the quotient of

these two functions. So, it is 0 where we have then infinite discontinuity I can write
𝑓(𝑥) 𝑥 𝑛−1 𝑒 −𝑥
and this will be lim+ 1 . So, this will be basically 1 so, 𝑒 −0 . So, this will this will
𝑔(𝑥) 𝑥→0
𝑥1−𝑛

lead to 1 which is; obviously, non zero and what is this limit here g(x) I can write next.

1 1 𝑑𝑥
So, next this limit here integral ∫0 𝑔(𝑥)𝑑𝑥 = ∫0 . So, this integral has an infinite point
𝑥 1−𝑛

of discontinuity or has a point of infinite discontinuity at 𝑥 = 0. So, we can write it

1 𝑑𝑥
lim+ ∫𝜖 the things which we are doing and in the previous classes.
𝜖→0 𝑥 1−𝑛

So, it will be fairly easy to check its convergence and I am skipping that step and I am
writing it for you that this integral converges if and only if. So, if and only if 1 − 𝑛 < 1
so; that means, 𝑛 > 0 we just worked out an example like that. So, this integral converges
only when n is positive so; that means, from here we can say that. So, we have a limit test
which says that the limit is non 0 and we were able to show that this integral converges
1
only when n is positive; that means, the integral ∫0 𝑓(𝑥)𝑑𝑥 is convergent for 𝑛 > 0.

So, even though at 𝑥 = 0 we had a point of infinite discontinuity for 𝑛 < 1 we can be able
to show that the integral is convergent the first sub integral this sub integral is convergent
for all 𝑛 > 0. So, the first part is settled now let us go to the convergence of 𝐼2 at 𝑥 = ∞.
Because in case of 𝐼2 only at 𝑥 = ∞ is the improperness in the limit the function is always
defined because even though if 𝑛 < 1 this function will never have a point of infinite
∞
discontinuity in that interval [1,∞]. So, our 𝐼2 is ∫1 𝑥 𝑛−1 𝑒 −𝑥 𝑑𝑥.

125
(Refer Slide Time: 22:42)

And from here we know that let us define let 𝑓(𝑥) = 𝑥 𝑛−1 𝑒 −𝑥 and g(x)=. So, we do not
have to define f(x) is actually this one. So, I can modify this a little bit since f(x) equals to
1
this and let 𝑔(𝑥) = 𝑥 2 .

Now, from inequality or from our basic knowledge of inequalities I can be able to write
𝑒 𝑥 > 𝑥 𝑛+1 for given n and for large value of x. Now this can be written as 𝑒 −𝑥 < 𝑥 −𝑛−1.

Now, if I multiply both sides by 𝑥 𝑛−1 then in that case this will be since x is running from
0 to ∞ this inequality will remain unchanged. And therefore, this will be 𝑥 𝑛−1 𝑒 −𝑥 <
1
𝑥 −2 which is . So, this is my 𝑔(𝑥) this is my 𝑓(𝑥) and; obviously, this is a positive
𝑥2

quantity so, we have the first part of our comparison test of inequality type.

Now, all we have to do is to check the value of this integral. So, next we have to check the
values of all or how to say convergence of this integral. So, what is the value of this integral
∞ 𝑑𝑥 𝐵 𝑑𝑥 1
∫1 . So, I can be able to write lim ∫1 and then this can be written as lim [− 𝑥]1𝐵 .
𝑥2 𝐵→∞ 𝑥2 𝐵→∞

And then we substitute the limit and when 𝐵 → ∞ the first one will go to 0 therefore, the
value of the integral would be 1 which is non zero and this is true for all n so, true for all
∞
n positive. So; that means, 𝐼2 = ∫1 𝑥 𝑛−1 𝑒 −𝑥 𝑑𝑥 is convergent at 𝑥 = ∞ for all 𝑛 > 0.

126
(Refer Slide Time: 25:48)

And hence from here, we can say that 𝐼 = 𝐼1 + 𝐼2 is convergent for all 𝑛 > 0. Because 𝐼1 is
convergent for all 𝑛 > 0 and 𝐼2 is convergent for all 𝑛 > 0 therefore, their sum is
convergent for all 𝑛 > 0. That means, the given gamma function which was
∞
∫0 𝑥 𝑛−1 𝑒 −𝑥 𝑑𝑥 is convergent for all 𝑛 > 0. And this is a very vital result in integral
calculus and of course, it is also our next topic in agenda.

1
Similarly we have an integral of type this let us say 𝐼 = ∫0 𝑥 𝑚−1 (1 − 𝑥)𝑛−1 𝑑𝑥 and here
we can see that 0 and 1 can be the points of discontinuities for certain values of m and n.
And for this integral we can also talk about such convergence and the range for m and n
for which this integral is convergent.

And we will see this integral in our next chapter because, this is also in our syllabus and
this integral is called as beta function. So, this integral converges for 𝑚 > 0 and 𝑛 > 0.
And this integral is also a type of improper integral and because it creates some kind of
infinite discontinuity at the, for this integral at the point 0 and 1.

And we can put it in the similar fashion by dividing the integral into two sub integrals and
then we look into the values of m and n for which we can talk about the convergence of
this improper integral. So, will probably do that in our next class and I hope I was able to
give you sufficient examples and theories on improper integral will work out for examples

127
in your assignment sheet as well. So, thank you for today and I look forward to your next
class.

128
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 13
Introduction to Beta and Gamma Function

Hello, students. So, in the last class we looked into the concepts of Improper Integrals and
how do you show that those improper integrals are convergent or not. So, we know that
we call an integral as an improper integral when you have some kind of how to say problem
with the lower limit or upper limit; for example, if any one of the limits are −∞ or +∞ or
if you have some kind of infinite discontinuity in your integrand then those integrals are
categorized as improper integral and.

Then, there are different-different methods with the help of which you can be able to show
their convergence, then there are several tests and we also worked out few tests with the
help of which we can be able to show the convergence of an improper integral.

(Refer Slide Time: 01:19)

So, that was up until the last lecture. Today, we will jump into a new topic in our integral
calculus section which is basically beta and gamma function.

So, will basically look into this topic beta and gamma function here.

129
(Refer Slide Time: 01:35)

So, we will basically look into beta and gamma function today and we will look into their
properties, the convergence and some other results related to beta and gamma function.

So, let us start with the definition of gamma function.

(Refer Slide Time: 02:08)

So, to start with this is our sorry chapter 3. Chapter 3 is Beta and Gamma function; beta
and gamma function and we will basically start with gamma functions.

130
 ∞
So, the integral, it is defined as ∫0 𝑥 𝑛−1 𝑒 −𝑥 𝑑𝑥 , where 𝑛 > 0 is called a gamma function
and it is denoted by Γ. And, so, this is a Greek letter capital Γ and that is we will write
∞
Γ(n) = ∫0 𝑥 𝑛−1 𝑒 −𝑥 𝑑𝑥 , where n is a positive number real number actually.

And, in the previous class we saw that this integral the gamma integral or gamma function
is convergent for all 𝑛 > 0. So, we can make a remark 1: Γ(n) is convergent for 𝑛 > 0,
we just saw in the previous class. So, this is first remark and there are several properties
that we can associate with this gamma integral and some of which are.

So, let me write it as properties. So, the first property is Γ(1) = 1. This one is pretty
straightforward to see. So, to prove this, so, let us say this is our property – 1. So, to prove
∞ ∞
this I can write Γ(1) = ∫0 𝑥1−1 𝑒 −𝑥 𝑑𝑥 . Now, this can be written as ∫0 𝑥 0 𝑒 −𝑥 𝑑𝑥 and
∞
therefore, we will end up with ∫0 𝑒 −𝑥 𝑑𝑥 .

∞
Now, this is a classic example of improper integral. So, we can write it as lim ∫𝐵 𝑒 −𝑥 𝑑𝑥 .
𝐵→∞

(Refer Slide Time: 05:37)

Now, after integration we will obtain lim [−𝑒 −𝑥 ]𝐵0 then this will be lim [−𝑒 −𝐵 + 𝑒 0 ]. So,
𝐵→∞ 𝐵→∞
−𝐵
when B goes to ∞, 𝑒 will basically equal to 0. So, this term will go to 0 and since this
term is a constant, it will remain unaffected by the limit and therefore, we will have 𝑒 0 ,
which is basically one and this is the first property that the value of Γ(1) will be 1. So, this
property is verified.

131
Next, we will prove that second property is property 2 is Γ(𝑛) = (𝑛 − 1)Γ(𝑛 − 1). So, let
us see how we can prove this. So, first of all by definition so, by definition we have Γ(n) =
∞
∫0 𝑥 𝑛−1 𝑒 −𝑥 𝑑𝑥 .

𝐵
So, this can be written as lim ∫0 𝑥 𝑛−1 𝑒 −𝑥 𝑑𝑥 and now, we will integrate this by parts. So,
𝐵→∞

if I integrate this by parts let me write in a small bracket integration by parts. So, if I
integrate this by parts then what will happen we will have

𝐵
lim [{−𝑥 𝑛−1 𝑒 −𝑥 }𝐵0 − (𝑛 − 1) ∫0 𝑥 𝑛−2 (−𝑒 −𝑥 )𝑑𝑥].
𝐵→∞

(Refer Slide Time: 08:19)

So, here we will have

𝐵
lim [{−𝐵 𝑛−1 𝑒 −𝐵 + 0𝑛−1 𝑒 −0 } + (𝑛 − 1) ∫ 𝑥 𝑛−2 𝑒 −𝑥 𝑑𝑥]
𝐵→∞ 0

So, now, here so, we have bracket here and this one can be put in a curly bracket. So, now,
when 𝐵 → ∞ this whole term will go to 0, because this will go to infinity and this one will
go to 0. So, ultimately the product will go to 0 and this is anyway 0. So, we are left
𝐵
(𝑛 − 1) lim ∫0 𝑥 (𝑛−1)−1 𝑒 −𝑥 𝑑𝑥. So, in short we can be able to write (𝑛 −
𝐵→∞
∞
1) ∫0 𝑥 (𝑛−1)−1 𝑒 −𝑥 𝑑𝑥.

132
Now, this is nothing, but this definition here. So, if I replace n by (𝑛 − 1), then this will
become the definition for Γ(𝑛 − 1), right. So, let us replace Γ(𝑛) by Γ(𝑛 − 1), then in that
case this one will be the definition of Γ(𝑛 − 1) which is basically our Γ(𝑛) =
(𝑛 − 1)Γ(𝑛 − 1) and therefore, this is what we needed to prove, alright. So, the second
property is established.

(Refer Slide Time: 10:34)

Now, we have the third property. So, property 3 which is basically Γ(𝑛) = (𝑛 − 1)! ,
where n is positive integer. So, you know what a factorial means for example, 5!
=5.4.3.2.1. So, that is what we mean by factorial. So, you multiply the number and then it
is previous number one by one and that will give you the factorial of that particular number.

So, here in this case we say that Γ(𝑛) = (𝑛 − 1)! . So, this is very straightforward to see.
So, by property 2 we have Γ(𝑛) = (𝑛 − 1)Γ(𝑛 − 1). Now, I use this how to say iterative
formula to write Γ(𝑛 − 1) = (𝑛 − 2)Γ(𝑛 − 2), right.

So, this formula for Γ(𝑛 − 1) one can be replaced with a formula of this type. So, that is
what I am doing I can write Γ(𝑛) = (𝑛 − 1)(𝑛 − 2)Γ(𝑛 − 2) . Similarly, I will proceed
and I can be able to write Γ(𝑛) = (𝑛 − 1)(𝑛 − 2). . . Γ(𝑛 − (𝑛 − 1)) and this one will
become Γ(1) and the value of Γ(1) will be finally,. So, the value of Γ(1) would be 1 times
Γ(1) and value of Γ(1). So, here is Γ(𝑛) = (𝑛 − 1)(𝑛 − 2). . .1 .

133
So, you see we have basically(𝑛 − 1)(𝑛 − 2)(𝑛 − 3). . .1 and therefore, this can be
written as (𝑛 − 1)! and this is what we needed to prove that Γ(𝑛) = (𝑛 − 1)! , right. So,
we can see that this Γ(𝑛) follows this kind of formula (𝑛 − 1)!.

So, after gamma function we will look into the definition of beta function and we will also
try to analyze its properties that how do we define also whether we can prove it is
convergence or not things like that.

(Refer Slide Time: 14:20)

So, let us start our next topic beta function.

1
So, an integral of type ∫0 𝑥 𝑚−1 (1 − 𝑥)𝑛−1 for 𝑚 > 0 and 𝑛 > 0 is called as the beta
1
function and it is denoted by Β(𝑚, 𝑛) = ∫0 𝑥 𝑚−1 (1 − 𝑥)𝑛−1. Some people also write it as
𝛽 instead of writing B, they use the symbol 𝛽 because it is called as the beta function, but
we will write B for this integral.

And, here we can see that this is an improper integral because when 𝑛 < 1, then in that
case this one will become (1 − 𝑥) to the power something negative and if one if we have
(1 − 𝑥) to the power something negative then that will come in the denominator so, that
it will become (1 − 𝑥) to the power something positive and at 𝑥 = 1, we will have a point
of discontinuity.

134
So, at 𝑥 = 1 when 𝑛 < 1we will have a point of infinite discontinuity similarly at 𝑥 =
0 when 𝑚 < 1 then we have a point of infinite discontinuity. So, for both m and n less
than 1 at the point 𝑥 = 0 and 𝑥 = 1 we have the point of infinite discontinuity. So,
Β(𝑚, 𝑛) is an improper integral.

Now, in order to show it is convergence we will follow the steps what we did in case of
gamma function and we can be able to show that this beta function is convergent. So, we
1
1 1
split the integral ∫0 𝑓(𝑥)𝑑𝑥 = ∫02 𝑓(𝑥)𝑑𝑥 + ∫1 𝑓(𝑥)𝑑𝑥. So, that we have only one point
2

of discontinuity in each one of these sub-integrals and then we show the convergence at 0
and convergence at 1. And, we can be able to see that both of these two sub-integrals are
convergent and therefore, the whole beta function will be convergent. So, I leave that task
up to the students and you can look into any book for the proof itself it is follows the
similar same steps like we did for the gamma function.

Today, we are going to analyze some properties of beta function. So, let us do that. So,
first property. So, property – 1 which is Β(𝑚, 𝑛) = Β(𝑛, 𝑚) . So, that means, you can
switch the indices and it will still remain the same, let us see. So, the proof or the solution
1
would go like this. So, Β(𝑚, 𝑛) = ∫0 𝑥 𝑚−1 (1 − 𝑥)𝑛−1 where our m and n are both positive
numbers.

Now, if I substitute 𝑥 = (1 − 𝑧) then in that case we will have 𝑑𝑥 = −𝑑𝑧 and when 𝑥 =
0, then 𝑧 = 1 and when 𝑥 = 1 then 𝑧 = 0. Therefore, our Β(𝑚, 𝑛) would reduce to integral
0
− ∫1 (1 − 𝑧)𝑚−1 𝑧 𝑛−1 𝑑𝑧

Now, since we know a small property of integral calculus that if you have
𝑏 𝑎
− ∫𝑎 𝑓(𝑥)𝑑𝑥 then this will be equals to integral ∫𝑏 𝑓(𝑥)𝑑𝑥 so, that means, this minus will
get reabsorbed basically when you are changing the range of integration in a way. So, here
we had a to b if we switch the limit points from a to b to b to a and in that case this minus
will get reabsorbed.

1
So, will do the same thing here and then this will turn into ∫0 𝑧 𝑛−1 (1 − 𝑧)𝑚−1 and if you
look at the definition of beta function then instead of x we have z and instead of m we have
n, instead of n we have m. So, z is basically a variable. So, it does not matter whether we
are using x or z it does not matter. However, you can see that instead of m we have n, and

135
instead of n we have m. So, that means, this is nothing, but the definition of
Β(𝑛, 𝑚) because the indices are switched.

So, here what we have is Β(𝑛, 𝑚) actually and this shows that your Β(𝑚, 𝑛) = Β(𝑛, 𝑚) for
all m and n positive. So, the first property is taken care of.

(Refer Slide Time: 20:44)

Now, we will look into the property – 2. So, the second property says Β(𝑚, 𝑛) =
𝜋
2 ∫02 𝑐𝑜𝑠 2𝑚−1 𝜃𝑠𝑖𝑛2𝑛−1 𝜃𝑑𝜃 . So, that means, instead of having an algebraic expression we
now, have a trigonometrical expression in the integrand and from looking at this
trigonometrical expression it is very straightforward that we have to do some kind of
substitution for x in our formula.

1
So, let us try to prove that. So, we know that Β(𝑚, 𝑛) = ∫0 𝑥 𝑚−1 (1 − 𝑥)𝑛−1, right. where
m and n are both positive. So, now substitute or we put 𝑥 = 𝑐𝑜𝑠 2 𝜃 then from here 𝑑𝑥 =
−2𝑐𝑜𝑠𝜃 sin 𝜃𝑑𝜃. And, therefore, our

0
Β(𝑚, 𝑛) = −2 ∫𝜋 𝑐𝑜𝑠 2𝑚−2 𝜃(1 − 𝑐𝑜𝑠 2 𝜃)𝑛−1 cos 𝜃 sin 𝜃𝑑𝜃 .
2

136
 𝜋
.So, let us see what would be the limit. So, here, when 𝑥 = 0 then 𝜃 = 2 and when 𝑥 =
0
1 then 𝜃 = 0. So, this can be written as integral −2 ∫𝜋 𝑐𝑜𝑠 2𝑚−2 𝜃𝑠𝑖𝑛2𝑛−2 𝜃 cos 𝜃 sin 𝜃𝑑𝜃
2
0
So, this will term into turn into −2 ∫𝜋 𝑐𝑜𝑠 2𝑚−2 𝜃𝑠𝑖𝑛2𝑛−2 𝜃 cos 𝜃 sin 𝜃𝑑𝜃.
2

Now, we can bring one cos 𝜃 here and then (2𝑚 − 2 + 1) will become (2𝑚 − 1). So, this
0
one will become −2 ∫𝜋 𝑐𝑜𝑠 2𝑚−1 𝜃𝑠𝑖𝑛2𝑛−1 𝜃and now, we will use the formula or the
2

𝑏
property which we just mentioned here that if you have − ∫𝑎 𝑓(𝑥)𝑑𝑥, then it can be written
𝑎
as ∫𝑏 𝑓(𝑥)𝑑𝑥.

(Refer Slide Time: 24:16)

So, I can use that property and write the integral Β(𝑚, 𝑛) as
𝜋
2 ∫02 𝑐𝑜𝑠 2𝑚−1 𝜃𝑠𝑖𝑛2𝑛−1 𝜃𝑑𝜃 and this is what we needed to prove, right. So, these formula
say our algebraic expression can be reduced into a trigonometrical expression for this beta
function.

𝑝+1 𝑞+1
Now, we can also be able to write next property. So, property – 3:Β ( , )=
2 2
𝜋
2 ∫02 𝑠𝑖𝑛𝑝 𝑥𝑐𝑜𝑠 𝑞 𝑥𝑑𝑥. We will need these properties to solve some examples. So, let us see
1 𝑝+1 𝑞+1
𝑝+1 𝑞+1 −1 −1
how we can prove it. So, we know our Β ( , ) = ∫0 𝑥 2 (1 − 𝑥) 2 𝑑𝑥 .
2 2

137
And, now, we will substitute 𝑥 = 𝑐𝑜𝑠 2 or x equals to. So, we have 𝑠𝑖𝑛𝑝 𝑥. So, we
substitute 𝑥 = 𝑠𝑖𝑛2 𝜃 . So, we substitute 𝑥 = 𝑠𝑖𝑛2 𝜃, then this will reduce to 𝑑𝑥 =
𝜋
2 sin 𝜃 cos 𝜃 𝑑𝜃. And, when x is 0, we will have then 𝜃 = 0 and when x is 1 then 𝜃 = 2 .

𝜋 𝑝+1−2
𝑝+1 𝑞+1 2( )
So, our Β( , ) will reduce to integral ∫02 (𝑠𝑖𝑛𝜃) 2 (1 −
2 2
𝑞+1−2
𝑠𝑖𝑛2 𝜃) 2 2 sin 𝜃 cos 𝜃 𝑑𝜃.

So, like previous example from here we will obtain basically 𝑐𝑜𝑠 2 𝜃. 𝑐𝑜𝑠 2 𝜃 so, square and
2 this will get cancelled and like the 2 and 2 here gets cancelled.

(Refer Slide Time: 27:55)

𝜋
So, we will basically end up with 2 ∫02 𝑠𝑖𝑛𝑝−1 𝜃𝑐𝑜𝑠 𝑞−1 𝜃 sin 𝜃 cos 𝜃 𝑑𝜃. So, one sin will
come here. So, it will turn into 𝑠𝑖𝑛𝑝 𝜃 1 cos will come here and it will turn into 𝑐𝑜𝑠 𝑞 𝜃 .
𝜋
So, let us write them. So, we will have 2 ∫02 𝑠𝑖𝑛𝑝 𝜃𝑐𝑜𝑠 𝑞 𝜃𝑑𝜃 .

Since 𝜃 is just a variable if I substitute 𝜃 = 𝑧 or 𝜃 = 𝑥 it does not matter what I substitute.

So, if I substitute 𝜃 = 𝑧 or x whatever I please the value of x will be the value of 𝜃 . So,
𝜋 𝜋
that means, when 𝜃 is 0, x is 0 when 𝜃 is 2 , x is 2 and from here 𝑑𝜃 = 𝑑𝑥. So, everything

will remain same and we can do that mainly because 𝜃 is just a variable here. So, let us

138
 𝜋
change the variables and then this will reduce to 2 ∫02 𝑠𝑖𝑛𝑝 𝑥𝑐𝑜𝑠 𝑞 𝑥𝑑𝑥 and this is what we
𝜋
𝑝+1 𝑞+1
needed to prove. So, Β ( , ) = 2 ∫02 𝑠𝑖𝑛𝑝 𝑥𝑐𝑜𝑠 𝑞 𝑥𝑑𝑥.
2 2

So, this was one another property of beta function and in the next class we will actually
work out the examples related to beta and gamma function where we will use these
1 3 3 5
properties to evaluate the value of Γ (2) or Γ (4) or the product of Γ (4) times Γ (8)so,

examples like that. So, we can be able to evaluate the value of how do say gamma of some
fraction or beta of some fraction. So, we will do such things in our next class and.

Thank you for your time.

139
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 14
Beta and Gamma Function

Hello, students. So, in the last class we looked into the definition of Beta and Gamma
Functions and we also tried to derive some properties related to these two special type of
integrals. And, in this lecture we will actually work out few examples and we will also
derive some properties of these two integrals and we will see how we can apply those in
those properties to solve some problems.

(Refer Slide Time: 00:53)

So, in order to start with let us. So, this is the topic which we are going to cover today;
beta and gamma function, their properties and if time permits then we will start with
differentiation under the integral sign. So, let us begin.

140
(Refer Slide Time: 01:17)

So, in the last class we saw that the gamma function is defined as Γ(𝑛) =
∞
∫0 𝑥 𝑛−1 𝑒 −𝑥 𝑑𝑥 where, n is a positive number. And, beta function is defined as
Β(𝑚, 𝑛) equals to or you can also write 𝛽(𝑚, 𝑛); it is up to you whatever notation you
1
prefer integral ∫0 𝑥 𝑚−1 (1 − 𝑥)𝑛−1 𝑑𝑥, where m and n are both positive numbers. So, both
of these two integrals are of course, improper integral for certain values of m and n and
we have shown that these two improper integrals are also convergent for this m, n and
considered to be positive.

Now, there is a very nice relation between beta and gamma function although it does not
seem like just looking at these two integrals because, one of them involves 𝑒 −𝑥 . Whereas,
the other one does not involve any kind of exponential function although we will see that
these two integrals are actually related. However, the proof is not that simple and if we get
into the proof then it will be slightly complicated. So, I will skip the proof, but I am going
to write the relation because it will help us solve several problems for those of the students
who are interested in the proof you can look into the books which I have recommended. It
is not that difficult is that it is to lengthy and keeping the time constraint in mind it is really
not worth doing in the class. However, the proof is interesting and I would suggest to look
into them look into those books.

141
 Γ(𝑚)Γ(𝑛)
So, the relation between beta and gamma function is given by Β(𝑚, 𝑛) = , where
Γ(𝑚+𝑛)

m and n are both positive numbers. So, this is the first relation between beta and gamma
1 1
function and just if let us say if we are asked to calculate the value of let us say Β (2 , 2) .
1 1
Γ( )Γ( ) 1
2 2
Then in that case we can write it as and if we know the value of Γ(2), then we can
Γ(1)
1 1
be able to calculate the value of Β (2 , 2).

1
So, let us see how we can calculate the value of Γ(2). So, this is our first relation. Now,

the second relation is between the trigonometric function and gamma function. So, this
𝜋 𝑝+1 𝑞+1
Γ( )Γ( )
says that ∫02 𝑠𝑖𝑛𝑝 𝑥𝑐𝑜𝑠 𝑞 𝑥𝑑𝑥 = 2 2
𝑝+1 𝑞+1 . So, this is an another relation between beta
2Γ( + )
2 2

between gamma function and the trigonometrical function. So, how do we prove this
relation? So, the proof of this relation is fairly simple.

(Refer Slide Time: 05:16)

So, the solution or the proof of that relation goes like this. If you remember we have shown
𝜋
that Β(𝑝, 𝑞) = 2 ∫02 𝑐𝑜𝑠 2𝑝−1 𝑥𝑠𝑖𝑛2𝑞−1 𝑥𝑑𝑥. So, we had proved this relation.

Now, if I substitute in place of p, (𝑝 + 1)and in place of q, (𝑞 + 1) then this will reduce

𝜋
to so, this will reduce to 2 ∫02 𝑐𝑜𝑠 2𝑝+1 𝑥𝑠𝑖𝑛2𝑞+1 𝑥𝑑𝑥. And, or we can take the other relation,
we do not need to take this relation. we can take the other relation between beta and gamma

142
function. The other relation is we have derived this one. So, the other relation is beta p
𝑝+1 𝑞+1
plus 1. We have this relation which we have derived in the previous class Β ( , )=
2 2
𝜋
2 ∫02 𝑠𝑖𝑛𝑝 𝑥𝑐𝑜𝑠 𝑞 𝑥𝑑𝑥.

So, this relation we have already derived in the previous class and this can be written
𝜋
𝑝+1 𝑞+1
as2 ∫02 𝑠𝑖𝑛𝑝 𝑥𝑐𝑜𝑠 𝑞 𝑥𝑑𝑥 = Β ( , ) and now, here I will use the relation – 1 and based
2 2
𝑝+1 𝑞+1
Γ( )Γ( )
on relation – 1, I can write 2
𝑝+1 𝑞+1
2
. So, here we have used via relation – 1. Let us see
Γ( + )
2 2

what is relation – 1.

𝑝+1 𝑞+1
So, in relation – 1, Β(𝑚, 𝑛) can be written as if I replace m by and n by then I can
2 2
𝑝+1 𝑞+1
Γ( )Γ( )
2 2
write 𝑝+1 𝑞+1 . So, this is the relation which we have used here and this can be written
Γ( + )
2 2

𝜋 𝑝+1 𝑞+1
1 Γ( 2 )Γ( 2 )
as integral ∫02 𝑠𝑖𝑛𝑝 𝑥𝑐𝑜𝑠 𝑞 𝑥𝑑𝑥 = 2 . 𝑝+1 𝑞+1 .
Γ( + )
2 2

So, this is the required relation and here we know that gamma is defined for all the positive
numbers. So, in order to have this one as positive, so, Γ(𝑛) is defined for all 𝑛 > 0. So, in
𝑝+1 𝑝+1
order to have this one positive I can write > 0 and > 0 that is our 𝑝 > −1 and 𝑞 >
2 2

−1. So, this relation is true. So, here I can write the condition that 𝑝 > −1 and 𝑞 > −1.
So, for p and q both greater than minus 1 this relation holds. Now, let us see how we can
1
obtain the value of Γ(2).

143
(Refer Slide Time: 09:40)

1 1
So, to obtain the value of Γ (2) that will be our relation – 3. So, Γ (2) =√𝜋 , alright. So,
𝑝+1 𝑞+1
from previous relation we have Β ( , ) or we can take Β(𝑚, 𝑛) relation. So, from the
2 2
Γ(𝑚)Γ(𝑛) 1
previous relation we know that is Β(𝑚, 𝑛) basically . So, here if I take 𝑚 = 2 =
Γ(𝑚+𝑛)
1 1
1 1 Γ( )Γ( )
2 2
𝑛 then this will become Β (2 , 2) = 1 1 .
Γ( + )
2 2

So, the denominator will become Γ(1) and we know that the value of Γ(1) is 1. So, I can
1
(Γ( ))2 1 1
2
write this one as . So, I am not writing that 1 and Β (2 , 2) can be written as
1
1 1 1 1
−1 𝑑𝑥
integral∫0 𝑥 2 (1 − 𝑥)2−1. So, this will be basically integral ∫0 .
√𝑥√1−𝑥

Now, this is a classical example of a method of substitution. So, I substitute 𝑥 = 𝑠𝑖𝑛2 𝜃 and
then you do some trigonometrical simplification which I am pretty sure you can be able to
do that.

144
(Refer Slide Time: 11:51)

𝜋 𝜋
2 sin 𝜃 cos 𝜃𝑑𝜃
So, here ultimately you will end up with integral ∫02 . So, this will be 2 ∫02 𝑑𝜃.
sin 𝜃 cos 𝜃
𝜋 1
So, ultimately we will obtain 2 × 2 . So, this is 𝜋; that means, we have (Γ (2))2 = 𝜋 and
1
from here it implies that Γ (2) = √𝜋 and this is what we needed to proof. So, the value of
1
Γ (2) 𝑖𝑠 √𝜋.

So, this is also a very important relation in gamma function and we often how to say require
1
the value of Γ (2) while calculating any type of in how to say integral or a formula. Next,

our next formula is we call it as Legendre’s duplication formula. Of course, it is proof is a

little bit extensive or a little bit lengthy. So, I will avoid the proof. However, I am just
going to write the formula. So, the formula says that show that √𝜋Γ(2𝑚) =
1
22𝑚−1 Γ(𝑚)Γ (𝑚 + 2)and this is our required duplication from where m is any positive

number.

Now, this formula here I mean in order to prove this we start with the definition of gamma
function and then we do some trigonometrical substitution; it is not very complicated, it is
just a little bit lengthy. So, I leave this proof up to the students and also you can look into
the book of Santi Narayan and other authors where you can find this trivial proof there.
What I am interested is to show the application of these formulas that how you can
1
calculate different values of gamma; for example, we know Γ(1) is 1, but what is Γ (2)?

145
And, you see with the application of this result here we can be able to calculate the value
1
of Γ (2).

So, these formulas although they are proof is a little bit lengthy, they are very handy in
1 1 1 1 1
order to obtain different values of Γ (2) or Β (2 , 2) and 4 or 2, things like that. So, these

formulas are quite handy and we are basically interested in their application. So, let us see
with the help of this formula what else we can do. So, our next result is or example is show
1 3
that let us say show that Γ (4) Γ (4) = √2𝜋 . So, let us see how we can prove it. So, the
1 1
solution here we take 𝑚 = 4. So, take 𝑚 = 4 in the duplication formula.

1
1 1 1 3
So, if I substitute 𝑚 = 4 what would happen. So, we will obtain √𝜋Γ (2) = 2−2 Γ(4)Γ(4).

(Refer Slide Time: 16:37)

1 3 1 1
So, this will result into Γ ( ) Γ ( ) = √2√𝜋 Γ( ) . So, Γ( ) is again √𝜋 and this one is √2.
4 4 2 2

So, this is √2 × √𝜋 × √𝜋 is equal to √2𝜋. So, you see although we needed to prove
1 3
Γ (4) Γ (4) we did not use at all the integral definition of gamma we just used this formula.

So, in order to getting into the integral and then trying to prove something which could
have been a little bit how to say complicated or extensive in a way we just use this formula
and with the help of which the whole proof was like three lines long. So, as you can see

146
these formulas are proving to be very handy in order to solve any type of gamma formula
here.

So, next example could be our next example is the example so, show that integral
∞ 2 √𝜋
∫0 𝑒 −𝑥 = 2
a solution. So, here when we see on the right hand side if it is √𝜋; that
1
means, it must result in to Γ (2) in some way. So, the basic instinct would say that it has
1 1
to be Γ (2) and; that means, here it would somehow lead to Γ (2) times some factor; that
1 1 1
means 2. So, it has to be in some way Γ (2) and let us see whether we can obtain Γ (2) or

not.

∞ 2
So, let us say we have 𝐼 = ∫0 𝑒 −𝑥 to our in given integral or the left hand side. So, this
is our left hand side I can also write left hand side or I that that is not a problem here. Now,
let us substitute 𝑥 2 = 𝑧. So, then in that case this will be 2𝑥𝑑𝑥 = 𝑑𝑧 and when x is 0, z is
1 ∞ 1
0, when x is ∞ z is ∞. So, from here we will have 𝐼 = 2 ∫0 𝑒 −𝑧 𝑧 𝑑𝑧.
√

1 ∞ 1
−1 1 1
So, I can write this as ∫ 𝑧 2 𝑒 −𝑧 𝑑𝑧, is it not? Because (2 − 1)is − 2 and then that
2 0
1
𝑧 −2 will come into denominator and then it will turn into a positive power. So, I can be
1 1
able to write this thing in this way now this is the definition of Γ (2) if I take 𝑚 = 2 then
1
this is nothing, but Γ (2) actually from the definition of gamma function.

1 1 √𝜋
So, this is our Γ (2) is it not? And we know that Γ (2) is √𝜋 𝑖. 𝑒. and this is what we
2

needed to prove here. So, you see using just some simple formulas, we are not getting into
√𝜋
any complicated calculation, we can be able to prove that this integral is actually . So,
2

this is one another application of gamma integral there is. There is one more example that
we can that we can prove.

147
(Refer Slide Time: 21:02)

1
So, it says express integral ∫0 𝑥 𝑚 (1 − 𝑥 𝑛 )𝑝 𝑑𝑥 in terms of beta function and hence
1
evaluate integral ∫0 𝑥 5 (1 − 𝑥 3 )3 𝑑𝑥.

1
So, solution so, first of all I will give an integral let me write it as 𝐼 = ∫0 𝑥 𝑚 (1 − 𝑥 𝑛 )𝑝 𝑑𝑥 .
So, if I substitute. So, first of all in order to have this in terms of beta function, this has to
be x to the power m minus something and this has to be (1 − 𝑥) to the power p minus
something or n minus something; so, we have to get rid of this x to the power and if we
want to express it as a beta function.

So, I substitute 𝑥 𝑛 = 𝑧, then this will be 𝑛𝑥 𝑛−1 𝑑𝑥 = 𝑑𝑧 and when x is 0, z is 0 when x is

1, z is 1. So, this will remain as it is. So, the limits will remain unchanged; this will be
1 1 𝑚 1−𝑛
∫ 𝑧 𝑛 (1 − 𝑧)𝑝 𝑧
𝑛 0
𝑛 𝑑𝑧 because, we have to find out the value of x to the power. So, this
1
𝑥 𝑛−1 will come in the, and denominator and then in that case x would be 𝑧1−𝑛 . So, it will
𝑛−1
be 𝑧 𝑛 and then the whole thing will go in the numerator and then it will turn into
something like this.

1 1 𝑚+1−1
Now, this can be written as 𝑛 ∫0 𝑧 𝑛 (1 − 𝑧)𝑝 𝑑𝑧, is it not? Now, this is our 𝑥 𝑚−1 in the

beta function formula and this is (1 − 𝑥)𝑛 .

148
(Refer Slide Time: 24:01)

1 𝑚+1
So, I can write this whole thing as Β( , 𝑝 + 1); because in beta function this was
𝑛 𝑛

supposed to be (𝑝 + 1), but since it is p; that means, I can be able to write it as

(𝑝 + 1 − 1)and hence you have here (𝑝 + 1), alright. So, that means, I was able to or we
were able to explain express this integral here in terms of a beta function, alright. So, let
1 1 𝑚+1
me write it as integral ∫0 𝑥 𝑚 (1 − 𝑥 𝑛 )𝑝 𝑑𝑥 = 𝑛 Β( , 𝑝 + 1).
𝑛

Now, we are supposed to calculate the value of this. So, in order to calculate the value of
1
the given integral take 𝑚 = 5 and 𝑛 = 3and 𝑝 = 3. So, then we have ∫0 𝑥 5 (1 −
1 5+1
𝑥 3 )3 𝑑𝑥 equals to. So, our 𝑛 = 3 and Β( , 3 + 1). So, this is this. So, I can be able to
3 3
1
write 3 Β(2,4).

1 Γ(2)Γ(4)
Now, I will use the beta and gamma relation here. So, then this will reduce to 3 .
Γ(2+4)

1 Γ(2)Γ(4)
So, this will reduce to3 .
Γ(6)

149
(Refer Slide Time: 25:54)

And, now I can be able to write this whole thing using that gamma factorial relation; so,
1 1Γ(1)3!Γ(1)
that means, 3 .
5!Γ(1)

1 1.3.2.1
So, all the Γ(1)’s will get how to say will be 1. So, this will be 3 . 5.4.3.2.1. So, this will get
1
cancelled; 2, 2 cancels. So, the answer will be and then we also have here I believe 3, sorry.
1
So, we also have here. So, we will have 1! Γ(1) and this one will be 1 again so, 3 and
3
1 1
then this, ok; so, 60. So, this is the value of that integral. So, integral ∫0 𝑥 5 (1 − 𝑥 3 )3 𝑑𝑥 =
1
and this is what we needed to show.
60

So, here we saw that if we are asked to evaluate a certain integral of this type we just have
to use this formula here the 1 which we have derived, put the values of m and n and p to
obtain the given problem and then we just have to calculate these gamma functions which
does not even involve calculating an integral, we just have to remember some formulas.
So, like Γ(n) = (n − 1)! . So, Γ(1) = 1!, Γ(4) = 3! and Γ(6) = 5! and with the help of
which we can be able to calculate these gamma function this integral here.

So, similarly there can be some other problems as well and I believe you can be able to
solve them I will also include those problems in our assignment sheet. So, that you can be
able to practice more and probably I will solve one more example from gamma function

150
in our next lecture and then we will close this topic and start with differentiation under the
integral sign.

So, thank you for your time and I will see you in the next class.

151
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 15
Differentiation under Integral Sign

Hello students. So, in the last class, we worked out few examples on beta and gamma
function. Today, I am going to give one more example where you can express an integral
in terms of gamma function and then we will move to the next topic which is a
Differentiation under the Integral Sign.

(Refer Slide Time: 00:37)

So, to start with, let me write the example. So, the example reads as show that
1 1 2 1
integral ∫0 √1 − 𝑥 4 𝑑𝑥 = 12 √𝜋 (Γ (4))2. So, here you can see that the left hand side does

not at all resemble with gamma integral; however, the answer can be expressed as gamma
function. So, how can we do that? Let us see. So, the solution we will assume that our
1
integral 𝐼 = ∫0 √1 − 𝑥 4 𝑑𝑥.

So, here I will take 𝑥 2 = 𝑠𝑖𝑛𝜃 . So, then in that case this will be 2𝑥𝑑𝑥 = cos 𝜃𝑑𝜃 and
𝜋
when x is 0, then 𝜃 is also 0 and when x is 1, then 𝜃 is 2 . So, based on this our integral I

152
 𝜋 𝜋 1
cos 𝜃𝑑𝜃 1
would reduce to ∫02 cos 𝜃 2√sin 𝜃 . So, this is basically 2 ∫02 𝑠𝑖𝑛−2 𝜃𝑐𝑜𝑠 2 𝜃𝑑𝜃. Now, we will
𝜋 𝑝+1 𝑞+1
𝑝+1 𝑞+1 1
) = 2 ∫02 𝑐𝑜𝑠 2( )−1
𝑥𝑠𝑖𝑛2( )−1
use that Β ( , 2 2 𝑥𝑑𝑥.
2 2

(Refer Slide Time: 02:56)

So, basically we will use this formula here. So, the formula which I am talking about is
1
this one. So, here.So, instead of p, I will choose − 2 and instead of q, I will choose 2.

(Refer Slide Time: 03:26)

153
 1
11 2+1 −2+1
So, then this whole thing will reduce to Β( , 2 ). So, I can write since
22 2
𝜋 𝑝+1 𝑞+1
𝑝+1 𝑞+1 1
) = 2 ∫02 𝑐𝑜𝑠 2( )−1
𝑥𝑠𝑖𝑛2( )−1
Β( , 2 2 𝑥𝑑𝑥.
2 2

𝑝+1 𝑞+1
So, this is the formula which we are using basically, so 2 ( ) − 1 and sin of 2 ( )−
2 2

1.

1 1
So, ultimately instead of q, I am taking − 2. So, this will reduce to − 2 ok. And instead of
3
p, I am taking 2. So, this will be 2 and 2, 2 will get cancelled so, ultimately 2, yes. So, I am
1 3 1
using this formula here. Now, I can be able to write it as 4 Β(2 , 4). So, this will be our beta

1 by 4 and now based on which, I can use that first relation which says that this will be
3 1
1 Γ(2)Γ(4)
.
4 Γ(3+1)
2 4

3 1
1 Γ(2)Γ(4) 3 3
So, this can be written as 4 7 . So, this Γ (2), so we note that Γ (2) can be written as
Γ( )
4

1 1 1
Γ (1 + 2) and then this can be written as, 2 Γ (2) and, yes. Now, so this can be written as
1 1
Γ (2). So, this whole thing can be now converted into, so this whole thing can be now,
2
1 1
1 Γ( )Γ( )
2 4
so this is our 1 by 4 as always; so, this is our 4 and then, we have 3 1 .
Γ( + )
2 4

(Refer Slide Time: 07:49)

154
 1
√𝜋Γ(4) 3
So, this can be written as 7 3 . So, here what we are doing is (𝑛 − 1). So, (2 − 1) =
8( −1)Γ( )
4 4

1 1
and then, we have Γ (2). So, that is the formula which I was using. So, we can either
2

write it or we do not write it. So, the formula is Γ(𝑛) = (𝑛 − 1)Γ(𝑛 − 1). So, this is what
we are using here and then, we will use this formula again here; so, this one.

1
√𝜋 Γ(4) √𝜋 1 3
Now, this is . . Now, we will have and we will have Γ (4) Γ (4) = √2𝜋. So, will
8 3Γ(3) 6
4 4

3 √2𝜋
Γ (4) be 1 . So, that will go at the numerator and then this whole thing will turn into a
Γ( )
4

square. So, we will basically have √2. So, we will basically have this here and this will
1 2 1
turn into √ (Γ( )2 . So, this is what we needed to prove. It is just some algebraic
12 𝜋 4

calculation which you can be able to do by yourself.

So, here we see that although this integral here, had no connection with gamma function,
whatsoever we can be able to reduce into a gamma function first of all into a beta function
formula. And, then we can be able to reduce this whole thing into a gamma function
formula and with the help of which we can be able to use some gamma function properties
and then we can be able to able to derive the required right hand side. So, this is what we
needed to prove.

(Refer Slide Time: 10:44)

155
So, similarly you may come across with a different types of integral in gamma function
example and based on which you can be able to calculate how to say those integrals using
the function of gamma, properties of gamma function. So, I will include some more
examples in your assignment sheet just for you to have a look at them and I am pretty sure
you will be able to solve such problems involving integrals and expressing them in terms
of gamma function.

So, the next topic in our third chapter is the differentiation under the integral sign. So, this
is also very interesting topic and not only that you will need it here in the integral calculus,
but you will be able to see this particular formula has a lot of applications in pde’s or also
when you are solving how to say, ordinary differential equations where you when. So, for
example, you might come across with certain results where you need to differentiate an
integral where which is not usual what we do in our differential calculus.

So, in differential calculus, we are usually given a function of this type, 𝑦 = 𝑓(𝑥) and
𝑑𝑦
when we say differentiate, then we basically differentiate like = 𝑓 ′ (𝑥)and if when
𝑑𝑥

𝑓(𝑥) is a function of one variable. Let us say, 𝑓(𝑥) can be our 𝑥 5 + sin 𝑥 + 𝑒 𝑥 some other
function dot dot and so on. So, we can be able to differentiate this function because it is a
function of one variable.

Now, now if we have a function of two variables, let us says 𝑧 = 𝑓(𝑥, 𝑦), then we usually
𝜕𝑓 𝜕𝑓 𝜕2 𝑓 𝜕2 𝑓
have partial derivatives. So, we can be able to calculate 𝜕𝑥 and 𝜕𝑦 or 𝜕𝑥 2 , 𝜕𝑥 2 dot dot and

so on. So that means, if you have a function of one variable, function of two variables,
then doing the differentiation is quite easy and we have. So, the function needs to be how
to say smooth in a way, but differentiation of an integral is not that straightforward.

So, it is not only the function of one variable or two variables, the smooth function you
can differentiate. You can also differentiate an integral, but in order to do that, it involves
some kind of special properties of that function and we will look into those properties. So,
𝑏
first of all, if I write 𝐼 = ∫𝑎 𝑓(𝑥)𝑑𝑥, so here x is a variable and we do not have any
parameter. So, 𝑓(𝑥) is a variable when I integrate then it will always result in to constant
or we can say that it has a constant function. So, a constant is also a function which assumes
the same value. So, if we are on the x and y plane and it will assume always the same

156
value. And so, 𝑦 = 𝑓(𝑥) = to a constant. So, a constant function is also function, but it is
constant at every point.

So, when you integrate a function of one variable, we obtain a constant function and that
constant function is always differentiable. So, there is no how to say a special thing to look
𝑏
at. However, if you have an integral of this type, let us say 𝐼 = ∫𝑎 𝑓(𝑥, 𝑦)𝑑𝑦, then after
you integrate, what do we obtain? We do not obtain a constant, we obtain a function of x
mainly because here variable x is involved. So, after we are done with the integration, we
will be able to obtain a function of x 𝜑(𝑥).

Now, if that 𝜑(𝑥) is smooth; that means, if the derivative of the function 𝜑(𝑥) with respect
to x. So, if the differentiation of 𝜑(𝑥) with respect to x exists, then in that case, we can be
able to differentiate this integral I here with respect to x. And this function is differentiable
with respect to x depends highly on the differentiability of this function with respect to x
because, if this function is not differentiable with respect to x, then this one will also be
not differentiable with respect to x.

So, in order to have the derivative, in order to differentiate this function, we need to have
the different stability of this function with respect to x. And in other words, the partial
derivative of this function x of this function f with respect to x must exist. Not only that,
in order to have it to be existing on a given domain of integration, it has to be continuous
throughout that interval so that it will exist at the point. So, that means, this function here
𝑓𝑥 must be, so must exist and continuous on some rectangular domain, on some domain R.
So, I can put this whole thing in a statement.

157
(Refer Slide Time: 16:44)

So, let us put this whole thing in a statement or a theorem. So, the theorem reads as, the
𝑏
theorem says let 𝜑(𝑦) = ∫𝑎 𝑓(𝑥, 𝑦)𝑑𝑥. So, now, we are in the function of two variable
case, 𝑓(𝑥, 𝑦)𝑑𝑥 where our 𝑓(𝑥, 𝑦) is a continuous function of x y in the rectangle 𝑅 =
{(𝑥, 𝑦): 𝑎 ≤ 𝑥 ≤ 𝑏, 𝑐 ≤ 𝑦 ≤ 𝑑}. So that means, we have a rectangular domain of this type
where the function f is defined. So, this is let us say. So, this is our a, b, c and d and this is
our rectangle R where the function is defined or where we are performing the
differentiation or integration.

Now, and 𝑓𝑦 (𝑥, 𝑦), so like I was saying in the in the previous slide the partial derivative
must exist and it is continuous in R, then the derivative of 𝜑 will exist and is equal to
𝑏 𝜕𝑓(𝑥,𝑦) 𝑑𝜑(𝑦)
integral ∫𝑎 . That is 𝜑 ′ (𝑦) which is basically is equals to integral
𝜕𝑦 𝑑𝑦
𝑑 𝑏 𝑏 𝜕
∫ 𝑓(𝑥, 𝑦)𝑑𝑥
𝑑𝑦 𝑎
equals to integral ∫𝑎 𝜕𝑦
f(x,y)dx ; that means, if the function f is continuous

in the rectangle R which is given in this fashion and if the partial derivative with respect
to y for this function f exists and if it is continuous in R, then the derivative of the function
𝜑 can be given in this fashion and you can bring the derivative inside the integral. And
when we bring the derivative inside the integral, then it will turn into a partial derivative
𝑑
which makes sense because in case of function of two variable we do not have , we
𝑑𝑦
𝜕 𝜕
always have 𝜕𝑦 or 𝜕𝑥.

158
So, once we bring the differentiations inside the integral, it will turn into a partial derivative
and this is our first formula for the differentiation under the integral sign. So, we are doing
the differentiation under the integral signs. And this is the integral sign which we were
talking about and this is our differentiation. So, you see if our integrand has some special
properties, we can even differentiate the resulting integral with respect to the parameter.
So, here y is our parameter and x is the variable of this integral and at the end, we can
differentiate this integral with respect to the parameter. So, here y is basically our
parameter, all right.

So, next proof of this theorem is, we are also skipping because it is not in the scope of this
lecture because it is a little bit how to say extensive in a way and we will mostly focused
on working or some examples. If you are interested and you can definitely look into those
books where they have used the epsilon delta definition which I am sure you already know
about. And by using some inequalities and epsilon delta definition you can be able to prove
this theorem. So, we leave the proof up to reader or up to the students. And at first, we will
see the application of this theorem which we just stated. Let us name it as theorem 1.

(Refer Slide Time: 22:30)

So, first example is, using differentiation under the integral sign, prove that integral
1 𝑥 𝑦 −1
∫0 𝑑𝑥 = log(1 + 𝑦). So, here we have to use the differentiation under the integral
𝑙𝑜𝑔𝑥

sign because it is specifically said in the statement. So, now, let us consider, 𝐼 =
1 𝑥 𝑦 −1
∫0 𝑑𝑥. So, first of all, after integrating this integral, we will obtain a function which
𝑙𝑜𝑔𝑥

159
is a function of y only because x is our variable. So, we are integrating with respect to x
and once the integration is done, we will end up with a function of y only, all right. So, I
1 𝑥 𝑦 −1
can write it as 𝐼(𝑦) = ∫0 𝑑𝑥
𝑙𝑜𝑔𝑥

Now, what we will do? We will differentiate both sides with respect to y because 𝑥 𝑦 , is
also a differentiable function. So, we can be able to differentiate this 𝑓𝑥 with respect to y.
So, if we can differentiate with respect to y and also it does not involve any kind of how
to say bad behaving functions in a way that partial derivative with respect to y would also
be continuous. So, we can differentiate this function now. So, let us differentiate.
𝑑𝐼(𝑦)
Differentiate with respect to y. So, what would happen we would obtain, =
𝑑𝑦
1 𝜕 𝑥 𝑦 −1
∫0 [ ] 𝑑𝑥.
𝜕𝑦 log 𝑥

1 𝜕 𝑥𝑦 𝜕 1
So, when we are differentiating, it will be∫0 [𝜕𝑦 (log 𝑥) − 𝜕𝑦 (log 𝑥)]𝑑𝑥 So,
𝜕 1
now, 𝜕𝑦 (log 𝑥) = 0 because we are differentiating with respect to y and this function does

not involve any variable with of y. So that means, this will be treated as a constant and
differentiation of constant is always 0.

(Refer Slide Time: 25:22)

1 𝑥 𝑦 log 𝑥
So, we will obtain ∫0 𝑑𝑥 after differentiation. So, this is what we will obtain. Now,
log 𝑥

we can integrate both sides. So, if I, so we can integrate this one not both sides, we will

160
 1 𝑥 𝑦+1 1
integrate this one and this will reduce to ∫0 𝑥 𝑦 𝑑𝑥 = [ 𝑦+1 ]10So, this will be . So; that
1+𝑦
𝑑𝐼 1 𝑑𝑦
means,𝑑𝑦 = 1+𝑦. So, we will have 𝑑𝐼 = 1+𝑦. This is basically our ordinary differential

equation, alright. So, at the end, we are obtaining an ordinary differential equation.

So, if I integrate this will turn into 𝐼(𝑦) = log 𝑒 (𝑦 + 1) + log 𝑒 𝑐. So, this can be written as
𝐼(𝑦) = log 𝑒 𝑐(𝑦 + 1). Now, I need to obtain the value of c. So, what is 𝐼(0)? So, let us
call it as equation 2 and I will call the original problem by equation 1. So, I will call this
relation, not the original problem but I will call this relation as 1. So, what is 𝐼(0)? ? So,
1 𝑥 0 −1
𝐼(0) is nothing but integral ∫0 𝑑𝑥 ; now, 𝑥 0 is 1 and 1 minus 1 is 0.
log𝑒 𝑥

(Refer Slide Time: 27:22)

So, the value of 𝐼(0) is 0 and substituting 𝐼(0) = 0 in 2, putting 𝐼(𝑦) = 0 in 2, we will
obtain we will obtain 0 = log 𝑒 𝑐. So, therefore, 𝑐 = 𝑒 0 which is basically one and we will
use this value of c in 2. So, from 2, we will have 𝐼(𝑦) = log 𝑒 (𝑦 + 1). 1. So, this is basically
log 𝑒 (𝑦 + 1)and 𝐼(𝑦) is nothing but our given integral, right. So, this is what we started
1 𝑥 𝑦 log 𝑥
with 𝐼(𝑦) = ∫0 𝑑𝑥 and this is what we needed to prove.
log 𝑥

So, you see just using the formula of differentiation under the integral sign, we can be able
to calculate the value of this integral without doing any complicated method of substitution
or anything. So, this is a very nice tool which helps us calculate the value of an integral by
solving some kind of ordinary differential equation like here. And, in the next class we

161
will look into a very important theorem of differential under integral sign which is Leibniz
rule of differentiation under the integral sign. So, I will stop here for today and I will look
forward to your next class.

162
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 16
Differentiation under Integral Sign (Contd.)

Hello students. So, up until last class we looked into Differentiation under the Integral
Sign, where we learnt about if you have an integral of this type let us say, 𝐼=
𝑏
∫𝑎 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦 then in that case how we can differentiate this integral.

(Refer Slide Time: 00:32)

𝑏
So, here so we do not need to have dy. So, first let us 𝐼 = ∫𝑎 𝑓(𝑥, 𝑦)𝑑𝑥 . So, if we integrate
𝑏
this integral then in that case we will obtain a function of y 𝐼(𝑦) = ∫𝑎 𝑓(𝑥, 𝑦)𝑑𝑥 . So,
basically this integral is nothing, but a function of y and after the integration, we can talk
about the differentiability of this function I(y). And if we want to differentiate this integral
𝑑𝐼 𝑏
let us say if we are doing 𝑑𝑦 then in that case, we can write it as integral ∫𝑎 𝑓(𝑥, 𝑦)𝑑𝑥 and

after the after we perform the differentiation, we will basically bring the differentiation
inside the integral and it can be written in this fashion.

𝑑𝐼 𝑑 𝑏 𝑏
𝜕
= ∫ 𝑓(𝑥, 𝑦)𝑑𝑥 = ∫ 𝑓(𝑥, 𝑦)𝑑𝑥
𝑑𝑦 𝑑𝑦 𝑎 𝑎 𝜕𝑦

163
Now, bringing this differential inside the integral is not that straightforward, I mean in
order to do that we need to have some special properties for this function f. And we saw
in the previous class that that this function f(x,y) needs to be continuous on a rectangular
domain R, and also the partial derivative of this function f needs to be continuous with
respect to y. So, based on that we also solved an example we will continue with the similar
topic which is differentiation under the integral sign, and we will now state an important
theorem in that respect. So, the theorem is called as Leibnitz’s.

(Refer Slide Time: 02:28)

Leibniz rule of differentiation under the integral sign so, Leibniz was a German
mathematician.

Now, the statement goes like this. Let so, this can be basically view as a theorem; we can
consider it as theorem 2. So, let 𝑓(𝑥, 𝑦) and 𝑓𝑥 (𝑥, 𝑦) be continuous on the rectangle R,
which is defined as {(𝑥, 𝑦): 𝑎 ≤ 𝑥 ≤ 𝑏, 𝑐 ≤ 𝑦 ≤ 𝑑}.

And let 𝜑, 𝜓: [𝑎, 𝑏] → [𝑐, 𝑑] be both differentiable in [𝑎, 𝑏], then the integral let us 𝑔(𝑥) =
𝜓(𝑥)
∫𝜑(𝑥) 𝑓(𝑥, 𝑦)𝑑𝑦 is derivable or differentiable or we can also write a differentiable on the
𝑑 𝜓(𝑥)
interval [𝑎, 𝑏], and our derivative can be given as 𝑔′ (𝑥) = 𝑑𝑥 ∫𝜑(𝑥) 𝑓(𝑥, 𝑦)𝑑𝑦 =
𝜓(𝑥)
𝑓(𝑥, 𝜓(𝑥))𝜓′ (𝑥) − 𝑓(𝑥, 𝜑(𝑥))𝜑 ′ (𝑥) + ∫𝜑(𝑥) 𝑓𝑥 (𝑥, 𝑦)𝑑𝑦

164
So, first of all if we look at this integral; obviously, if we integrate with respect to y and
if the limits are both functions of x, then in that case after substitution will basically obtain
a function of x, and then we can talk about it its differentiability. So, to talk about its
differentiability makes sense, because we ultimately obtain a function of x only.

Now, in order to do the differentiation on this interval [𝑎, 𝑏], this function f needs to have
these 2 important properties that it needs to be continuous and also its a partial derivative
with respect to x also needs to be continuous. And then in that case we can talk about such
difference ability and rule of differentiation is we first substitute 𝑓(𝑥, 𝜓(𝑥)) in the place
of y we substitute the upper limit which is 𝑓(𝑥, 𝜓(𝑥))𝜓′ (𝑥) − 𝑓(𝑥, 𝜑(𝑥))𝜑′ (𝑥) +
𝜓(𝑥)
∫𝜑(𝑥) 𝑓𝑥 (𝑥, 𝑦)𝑑𝑦 .

Now here, whether we would like to see, whether this Leibniz rule of differentiation under
the integral sign is actually valid when the limits are constant or not. So, if I substitute 𝜑 =
𝑎 & 𝜓 = 𝑏 then b is constant. So, 𝜑 ′ (𝑥) will be 0 and 𝜑 is constant and in this place of 𝜑
we have a. So, 𝜑 is constant and therefore, 𝜑 ′ (𝑥) = 0 as well.

So, the first 2 term will vanish and then in that case we are left with integral
𝑏
∫𝑎 𝑓𝑥 (𝑥, 𝑦)𝑑𝑦 so, which is basically our theorem 1 for differentiation under the integral
sign. Remember theorem 1 was something which deal with constant limits so; that means
both the limits lower and upper limits where constant in case of theorem 1 and in case of
theorem 2, you have both the limits as a function of x. So, this is a general rule of
differentiation under the integral sign or sometimes they it is also called as Leibniz rule of
differentiation.

So, for the proof we are not concerned with that, because the proof is relatively long and
its out of the scope of this syllabus, it is just that knowing the statement is very important
because we require this statement quite often in our work. And when we also solve some
exercises and some problems, we definitely require this formula to solve those problems.

So, now we are ready with solving the examples. So, in order to do so, let us start with our
first problem. So, to start with our first problem let us consider example 1.

165
(Refer Slide Time: 09:03)

So, find the differentiation of the following integral with respect to x on [0,1] × [0,3]. So,
that is our rectangle 𝑅 = {(𝑥, 𝑦): 0 ≤ 𝑥 ≤ 1,0 ≤ 𝑦 ≤ 3}.

𝑥4
So, we need to find that differentiation of the integral given as 𝐼(𝑥) = ∫𝑥 2 (𝑥 2 + 𝑦 2 )𝑑𝑥all
right. So, here our function f is so, let us go to the solution. So, of course, here we will
apply the Leibniz rule of differentiation. So, first of all we see that both the functions let
us start this interval by 0 as well. So, that we have a bigger interval. So, [0,1] and [0,3].
So, of course, these 2 functions are continuous both the upper limits and lower limits in
this interval, they also map from [0,1] to [0, 3]. And this function this (𝑥 2 + 𝑦 2 ) since it
is an algebraic function in a way, it will be continuous and not only that partial derivative
with respect to x is also continuous in this domain. So, all those conditions for the Leibniz
rule of differentiation is satisfied. So, we do not have to worry.

Now, or we can write at least one line that, the conditions for Leibniz rule of differentiation
𝑑𝐼 𝑑 𝑥4
is satisfied. So, we have 𝑑𝑥 = 𝑑𝑥 ∫𝑥 2 (𝑥 2 + 𝑦 2 )𝑑𝑥and then this will be

𝑥 4
2
𝑑 4 𝑑 2 𝜕 2
(𝑥 + (𝑥 4 )2 2 2
(𝑥 ) − (𝑥 + (𝑥 ) 2 (𝑥 ) + ∫ (𝑥 + 𝑦 2 )𝑑𝑥
𝑑𝑥 𝑑𝑥 𝑥 2 𝜕𝑥

166
(Refer Slide Time: 13:27)

𝑥4
and then this can be written as (𝑥 2 + 𝑥 8 )4𝑥 3 − (𝑥 2 + 𝑥 4 )2𝑥 + ∫𝑥 2 2𝑥𝑑𝑥 .

4
So, this is basically 4𝑥 3 (𝑥 2 + 𝑥 8 ) − 2𝑥(𝑥 2 + 𝑥 4 ) + [𝑥 2 ]𝑥𝑥2 . So, we substitute 𝑥 2 to
𝑥 4 and then whatever is the limit we just write the limit here. So, we substitute 𝑥 2 . So, this
will be (𝑥 8 − 𝑥 4 ) and we just calculate and that will be the differentiation of the integral.

So, this is how we calculate in this case the differentiation of an integral. So, what we do,
we just apply the Leibniz rule of differentiation. So, 𝜑(𝑥) to 𝜓(𝑥) and then we integrate
with respect to y I am sorry. So, here we are integrating with respect to y. So, this will be
4
2xdy and then we can write this as 4𝑥 3 (𝑥 2 + 𝑥 8 ) − 2𝑥(𝑥 2 + 𝑥 4 ) + 2𝑥[𝑦]𝑥𝑥2 sorry. The
integration here the integral has to be with respect to y then in that case we get the integral
with respect to x yes.

So, if we integrate with respect to y, then we will basically obtain y here and this is of
course, 2x and then in that case this will be 2x times y and the limit we can substitute 𝑥 2
to 𝑥 4 and then we calculate this is basically a calculation of a simple algebraic expression.
So, this is how we calculate the differentiation of an integral and next we work out few
more examples just to get the concept a little bit more clear.

167
So, example 2 or I would say 3. So, assuming the validity of differentiation under the
∞ 𝛼2
2 1
integral sign, show that integral ∫0 𝑒 −𝑥 cos 𝛼𝑥 𝑑𝑥 = 2 √𝜋𝑒 − 4 , where 𝛼 is any positive

number.

𝛼2
∞ 2 1 −
So, we have to show at this integral ∫0 𝑒 −𝑥 cos 𝛼𝑥 𝑑𝑥 = 2 √𝜋𝑒 4 . So, here it is very

clear that 𝛼 is the parameter since we are integrating with respect to x. So, we can write
∞ 2
here 𝐼(𝛼) = ∫0 𝑒 −𝑥 cos 𝛼𝑥 𝑑𝑥 So, this is given. Now, let us differentiate both sides with
respect to 𝛼.

(Refer Slide Time: 17:59)

𝑑𝐼
So, this will be 𝑑𝛼 and our partial derivative will go inside of the integral. So, this will be
𝑑𝐼 𝑑𝐼
all right. So, this is our 𝑑𝛼, and now we integrate by parts both sides, then we obtain 𝑑𝛼 =
∞ 2
− ∫0 𝑥𝑒 −𝑥 sin 𝛼𝑥 𝑑𝑥. So, we integrate by parts and then this will reduce to
1 2 1 ∞ 2
[2 𝑒 −𝑥 sin 𝛼𝑥]∞
0 − 2 𝛼 ∫0 𝑒
−𝑥
cos 𝛼𝑥 𝑑𝑥.

So, here we are integrating this whole thing by parts and considering sin 𝛼𝑥 as the first
2
function, and 𝑥𝑒 −𝑥 as the second function, we will obtain this integral here, this is
basically a part of method of substitution and on. So, it is pretty straightforward and from
here when x is ∞ then this whole thing will go to 0 and when x is 0, then again this whole
thing will go to 0. So, the first term will not play any role and the second term can be

168
 𝑑𝐼 𝛼
written as 𝐼(𝛼), because this is our 𝐼(𝛼) we started with. Therefore, our 𝑑𝛼 = − 2 𝐼(𝛼) is

basically this ordinary differential equation.

𝑑𝐼 𝛼
So, if we integrate both sides then in that case this will be = − 2 𝑑𝛼 . So, if I integrate
𝐼
𝛼2
then this will be log 𝑒 𝐼 = − + 𝑐.
4

(Refer Slide Time: 20:40)

𝛼2
And this can be written as 𝐼(𝛼) = 𝑒 𝑐− 4 alright.

So, now we find out the value of I(0). So, when 𝛼 is 0 then this will be 𝑒 𝑐−0 so, basically
𝑒 𝑐 . So, let us call this as equation number 1 this as equation number 2. So, we have to
evaluate the value of C and I(0) on the left hand side basically I(0) is when I substitute𝛼 =
∞ 2
0, then this is basically cos0 and cos0 is 1. So, we have 𝐼(0) = ∫0 𝑒 −𝑥 cos 0 𝑑𝑥 so, cos0
∞ 2
is 1. So, this will be integral ∫0 𝑒 −𝑥 𝑑𝑥, and in our previous lecture we just proved that
1
Γ( )
2
this is nothing, but our .
2

√𝜋
So; that means, . So, this relation we already proved in previous topic. So, combining
2

these 2 we can see that I can see that C will be. So, if I choose C as let us say log 𝑒 𝐶. then

169
 𝛼2
𝛼2
in that case this will be log 𝑒 𝐼 = − + log 𝑒 𝐶 . So, this will be 𝐼(𝛼) = 𝐶𝑒 − 4 . So, this
4
𝛼2
will be it the C will come here and this will be 𝐼(𝛼) = 𝐶𝑒 − 4 . .

So, we can it is all about playing with the arbitrary constant. So, we can play with this
arbitrary constant as we please. So, we can write it as 𝐼(0) = 𝐶𝑒 −0 . So, this is basically
our C. So, our C is basically I so; that means, from 2 and 3. So, from 2 and 3 we have from
√𝜋
2 and 3 𝐼(0) = 𝐶 = . So, from here we can write from 1 we can write 𝐼(𝛼) =
2
𝛼2
√𝜋 −
𝑒 4 and this is what we needed to prove.
2

So, you see by considering this integral. So, this integral here it is also very important to
notice whether this falls under the category of differentiation under the integral sign or not.
So, if it does falls into that category, then in that case we have to identify the parameter it
is usually the variable with respect to which we are not integrating. So, here we are
integrating with respect to x. So, 𝛼 is the parameter and then after the integration we will
obtain a function which is a function of 𝛼 only, and therefore, we can write it as 𝐼(𝛼).

Now, we differentiate this 𝐼(𝛼) with respect to 𝛼, and we obtain this integral here then we
integrate this integral by parts and then we obtain this here, and we can evaluate the limit
and we realize that the limit in the first term vanishes therefore, the second term will reduce
𝑑𝐼 𝛼
to = − 2 𝑑𝛼, and next we integrate and after integrating choosing a constant we know
𝐼

that from our plus 2 level that we choose this constant in such a way that you can use those
logarithmic formulas and then you can write it here as e to the power that whatever you
have on the right hand side.

So, instead of choosing C, I chose a logarithmic to the base e and then I can bring this log
𝐼
here and then this will be log 𝑒 𝐼 − log 𝑒 𝐶. So, it will be log 𝑒 𝐶 and then on the right hand
𝛼2
side we will obtain 𝐶𝑒 − 4 . Now we need to evaluate the value of C. So, what we do? We
√𝜋
substitute 𝛼 = 0 and therefore, this will reduce to and by substituting the value of C
2

here we will obtain require 𝐼(𝛼), and this is what we needed to prove in our example. So,
in this problem this is what we needed to prove. So, it is not that how to say complicated
it might get sometimes lengthy, but solving this type of problem would not be that much
complicated.

170
Next so, let me give you 1 or 2 more examples, but I would assume that solving those
problems can also be fairly easy because they do not involve such a complicated
calculation or something.

(Refer Slide Time: 26:57)

𝑎 log(1+𝑎𝑥) 1
So, another example could be show that integral ∫0 𝑑𝑥 = 2 log(1 + 𝑎2 ) tan−1 𝑎.
1+𝑥 2

So, again if you look at this integral it might look a little bit complicated, and on the right
hand side we have a value of the integral, which sort of involves log and tan−1 .

𝑎 log(1+𝑎𝑥)
So, I can write 𝐼 = ∫0 𝑑𝑥. Now again here we are integrating dx. So, here we are
1+𝑥 2

integrating with respect to x so; that means, a here will be considered as the parameter and
we have an a in the integrand and as well as in the upper limit as well. So, this 𝐼(𝑎) =
𝑎 log(1+𝑎𝑥)
∫0 𝑑𝑥 is not that simple to differentiate here we have to use the Leibniz rule of
1+𝑥 2

differentiation, before we use theorem 1 of integral under the differentiation under the
integral sign, but in this case we have to use the Leibniz rule.

𝑑𝐼
So, the Leibniz rule can be applicable here. So, which is which can be given as
𝑑𝑎
𝑑 𝑎 log(1+𝑎𝑥)
∫ 1+𝑥 2 𝑑𝑥
𝑑𝑎 0
and this can be written as log of x y. So, instead of y I substitute a. So,
log𝑒(1+𝑎.𝑎) 𝑑 log𝑒 (1+𝑎.0) 𝑑 𝑎 1 𝑥
this will be (𝑎) − (0) + ∫0 . 𝑑𝑥
1+𝑎2 𝑑𝑎 1+02 𝑑𝑎 1+𝑎𝑥 1+𝑥 2

right.

171
So, this will be the value of the differentiation of this integral with respect to x here and
now we have to again calculate like we did before. So, this will reduce to this will reduce
to a function with respect to a. So, here this will reduce to a yeah and then we basically
formulate the ode. So, let me go to the next page. So, we will basically obtain an ode of
following type.

(Refer Slide Time: 32:02)

𝑑𝐼 1 𝑎 tan−1 𝑎
So, the ode will look like = 2(1+𝑎2 ) log𝑒 (1 + 𝑎2 ) + . So, this is the required
𝑑𝑎 1+𝑎2

ordinary differential equation that we will obtain.

It involves a slight simplification or calculation here, but since it is relatively big I leave
this to the students and ultimately we will obtain an ordinary differential equation of this
type, and from here we will basically obtain I(0) like we did before, because we will obtain
a constant and then we substitute the value of the constant to solve this ordinary differential
equation, and that will give us the value of this integral I which we started with.

So, based on these calculation, we see that we can be able to evaluate this type of integral,
which is basically involves another parameter and this type of integral do come up quite
often in engineering sciences also, and there we might need to do the differentiation under
the integral sign and this theorem 1 and theorem 2 which is basically Leibniz rule of
differentiation will prove to be very handy, and I will include some more examples in your
assignment sheet and I look forward to your next class.

172
Thank you.

173
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 17
Double Integral

Hello, students. So, up until previous class we looked into the first three topics of our
integral calculus section.

(Refer Slide Time 00:31)

So, they were basically – a partition and concept of Riemann integral, as a fundamental
theorem of integral calculus mean value theorems, then we looked into a reduction
formula, derivation of different types of reduction formula, then we also looked into
improper integral and their convergence and finally, we looked into beta and gamma
functions and differentiation under the integral sign.

174
(Refer Slide Time 01:03)

So, the next topic in our agenda is about double integrals change of order of integration
and things like that. So, to start with so, today, we will start with double integrals. So, to
start with, what do we mean by double integral and how do we define a double integral
and things like that.

(Refer Slide Time 01:41)

So, we know that a function for the function of one variable; let us say a function f which
is defined and bounded on a closed interval [𝑎, 𝑏] then our definite integral was given by

175
 𝑏
integral ∫𝑎 𝑓(𝑥)𝑑𝑥. So, this is how we define the definite integral and this actually gives
us the area between the point a to b bounded by this curve 𝑓(𝑥).

Now, an interesting point one could ask in this context is that what happens for the function
of two variables. So, instead of having a function let us say for one variable if you have a
function of two variable say 𝑓(𝑥, 𝑦) then what would happen with the integration of such
functions because we know that for the function of one variable we can be able to check
actually its limit, its continuity, differentiability and at the same time we can also perform
the integration.

So, we know from differential calculus that for the function of two variables as well we
can be able to check its limit continuity, partial derivative things like that. So, what about
the integrability of a function of two variables and how do we actually define that?

So, since we have two variables here, x and y we need to find the intervals for these two
variables; that means, if you have two variables then you must be having two intervals
where these two variables actually belong to. So, let us say x and y. So, I am going to give
how to say a formal definition in a way that let 𝑎 ≤ 𝑥 ≤ 𝑏 and 𝑐 ≤ 𝑦 ≤ 𝑑 such that the
rectangle R is given by [𝑎, 𝑏] × [𝑐, 𝑑].

So, what does this mean is if you draw this on x and y axis so, here we have the point a,
here we have the point b, we have the point let us say c and here we have the point d. So,
this is basically our rectangle R and we say that if f is a bounded function on R then we
𝑏 𝑑
denote the double integral as ∫𝑎 ∫𝑐 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦.

Now, what do we mean by this integral I am going to write it down in probably next couple
of minutes, but here we have a range of integration a to b which is for the variable x and
range of integration from c to d which is for the variable y. So, do not mix it up. So, when
we write a to b that is for x and when we write c to d then that is for the variable y.

So, let us give a formal definition for this double integral and, for the time being we will
perform the double integral on the rectangle only.

176
(Refer Slide Time 06:01)

So, double integral on the rectangle R and R is our [𝑎, 𝑏] × [𝑐, 𝑑]. So, that is pretty much
the same.

So, the definition goes like this; let f be a bounded function of two variables of course, of
two variables x and y over a rectangle R which is given by [𝑎, 𝑏] × [𝑐, 𝑑] and also let 𝑃1 as
the partition of the interval for the variable x. So, 𝑃1 = {𝑎 = 𝑥0 < 𝑥1 < 𝑥2 <. . . < 𝑥𝑛 =
𝑏} and 𝑃2 = {𝑐 = 𝑦0 < 𝑦1 < 𝑦2 <. . . < 𝑦𝑛 = 𝑑} .

So, let this and this 𝑃1 and 𝑃2 be the partitions of [𝑎, 𝑏] and [𝑐, 𝑑]respectively. So, what we
are doing is basically if we have this interval let us say this is x, this is y, so, this is my a,
b, c, d. So, what we are doing is we are taking the partition of sorry c and d; d is here. So,
a, b, c, d. Now, if I follow that the same terminology. So, a b and c d, yes.

177
(Refer Slide Time 08:45)

So, a, b and this one is c, d, yes. So, this one is c and this one is d alright.

So, now what we have is here is ; that means, we are taking the partition of these interval
for the variable x. So, like 𝑥0 , 𝑥1 , 𝑥2 , 𝑥3 dot dot up to 𝑥𝑛 and then we are taking the
partition for the variable y for the interval [𝑐, 𝑑] where we have the variable y. So, we take
𝑦0 , 𝑦1 , 𝑦2 up to 𝑦𝑛 = 𝑑. Now, this 𝑥0 to 𝑥1 this is a very tiny rectangle, but here 𝑥0 to 𝑥1
and 𝑦0 to 𝑦1 is one such tiny rectangle, then we have 𝑦1 to 𝑥1 to 𝑥2 and 𝑦1 to 𝑦2 will be
another rectangle. So, in a way here we will have if we have. So, here we will keep a
rectangle otherwise it will be a square. So, we have 𝑦1 , 𝑦2 , 𝑦3 up to 𝑦𝑚 .

So, that means, then here we have n subintervals for the variable x and then we have m
sub intervals for the variable y. So, in a way we have m times n. So, mn number of
rectangles here so, here we will have these two partitions. These two partitions these two
partitions divide the rectangle R into m times n sub rectangles, alright because we have n
number of intervals in this direction m number of intervals in direction. So, the number of
rectangles will be m times n sub rectangles.

And, the area of this sub rectangle that is let us say our sub rectangle let us write it as
∆𝑅𝑖𝑗 : [𝑥𝑖−1 , 𝑥𝑖 ] × [𝑦𝑗−1 , 𝑦𝑗 ] where 𝑖 = 1,2,. . . , 𝑛 and 𝑗 = 1,2,3,. . . , 𝑚 alright. So, this
is basically our R-th how to say rectangle in a way and it is given by this fashion.

178
(Refer Slide Time 12:35)

Now, since we have defined the how to say the partition of that rectangle we denote the
subdivision of R into mn sub rectangles, by P. So, that means, this one this here we denoted
by P. So, P is sort of like how to say a partition of rectangles basically.

So, in case of function of one variable we had P as a partition where we had

(𝑛 + 1) number of points and that partition P divided the interval into n subintervals. Here
in this case the partition P for the function of two variable divides the whole rectangle R
into mn number of sub rectangles that is what we have written here. And, let 𝑀𝑖𝑗 and 𝑚𝑖𝑗 be
the upper and lower bounds of f in ∆𝑅𝑖𝑗 .

Then we write upper integral some or 𝑈(𝑃, 𝑓) = ∑𝑛𝑖=1 ∑𝑚

𝑗=1 𝑀𝑖𝑗 (𝑥𝑖 − 𝑥𝑖−1 )(𝑦𝑗 − 𝑦𝑗−1 ).

And, we defined 𝐿(𝑃, 𝑓) = ∑𝑛𝑖=1 ∑𝑚

𝑗=1 𝑚𝑖𝑗 (𝑥𝑖 − 𝑥𝑖−1 )(𝑦𝑗 − 𝑦𝑗−1 ) .

So, these two are the upper integral sum and lower integral sum. Now, if you remember
the function of one variable where we defined the integral we actually took the infimum
of all these upper integral sum. So, that gave us the upper integral and then we took the
supremum of for all 𝐿(𝑃, 𝑓) then that gave us actually the lower integral.

So, here we will have the infimum or greatest lower bound of the set of upper integral
−
sums is called upper integral and it is denoted by integral ∬𝑅 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦 =
𝑏− 𝑑−
∫𝑎 ∫𝑐 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦.

179
So, that is basically our upper integral. Similarly, you can define the lower integral as well
by taking the supremum over all 𝐿(𝑃, 𝑓).

(Refer Slide Time 16:39)

.
And, it will be denoted by rectangle R ∬𝑅 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦 or one can write integral
−

𝑏 𝑑
∫𝑎 ∫𝑐 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦 and this is our basically lower integral.
− −

And, like function of one variable when we have so, if we have

𝑏 𝑑 𝑏− 𝑑−
∫ ∫ 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦 = ∫ ∫ 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦
𝑎− 𝑐− 𝑎 𝑐

if they are equal then we say the double integral of f exists on the rectangle R and it is
. 𝑏 𝑑
denoted by integral ∬𝑅 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦 or you can write integral ∫𝑎 ∫𝑐 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦.

So, that is how we define formally the double integral of a function f which is bounded on
the rectangle [𝑎, 𝑏] × [𝑐, 𝑑]. So, it is how to say a nice way to define, but there is another
question here that how do we evaluate the double integral. So, it is not only just the
definition, but at the same time we have to address that how do we calculate the double
integral.

180
(Refer Slide Time 18:56)

So, in order to calculate the double integral the calculation of double integral basically can
be done. The calculation of double integral over R; so, if the double integral how do we
𝑏 𝑑
calculate. So, if the double integral ∫𝑎 ∫𝑐 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦 exists and integral from
𝑏 𝑏 𝑑
∫𝑎 𝑓(𝑥, 𝑦)𝑑𝑥 also exists for each y in [𝑐, 𝑑] then integral ∫𝑎 ∫𝑐 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦 =
𝑑 𝑏
∫𝑐 [∫𝑎 𝑓(𝑥, 𝑦)𝑑𝑥]𝑑𝑦 which means that if this integral exists; that means, if for every y if
the integration with respect to x exists then in that case which we can evaluate this double
integral.

So, first we evaluate this integral with respect to x only where y will be treated as constant
and after the integration then we substitute the values of x and then we will basically obtain
a function of y only and afterwards we integrate with respect to the function y only because
other than that you have rest of the things as constant and then that will be actually the
value of that double integral. So, here we first integrated with respect to x and then we
integrated with respect to y.

Similarly, so, let us call it as a point – 1. Similarly, if the double integral so, all these
𝑏 𝑑
language would be same. So, if integral ∫𝑎 ∫𝑐 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦 exist so, if this double integral
𝑑
exists and integral ∫𝑐 𝑓(𝑥, 𝑦)𝑑𝑦 also exists for each x in [𝑎, 𝑏], then we can write integral

𝑏 𝑑 𝑏 𝑑
∫ ∫ 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦 = ∫ [∫ 𝑓(𝑥, 𝑦)𝑑𝑦]𝑑𝑥
𝑎 𝑐 𝑎 𝑐

181
So, that means, if we have this double integral exist here and if the integration with respect
to y also exists, where x is treated as a constant then we can be able to evaluate this double
integral in this fashion. So, first we integrate with respect to y treating x as constant and
then we integrate the rest of the result with respect to x because then we will have only a
function of x. So, either we integrate with respect to y or we integrate with respect to x
first and then we integrate whatever the variable is left afterwards and if the double integral
exists then these two repeated integral would also be same.

So, if because you see if the double integral exists; that means, this side has a value. So,
then in that case you will have the value for this side as well and the same value should be
𝑑 𝑏
equal to this side as well so; that means, integral ∫𝑐 [∫𝑎 𝑓(𝑥, 𝑦)𝑑𝑥]𝑑𝑦 =
𝑏 𝑑
∫𝑎 [∫𝑐 𝑓(𝑥, 𝑦)𝑑𝑦]𝑑𝑥 . So, these two repeated integral will be equal if this double integral
exist. The proof is a little bit lengthy and it is actually out of the scope of this lecture.

So, instead of going into the theoretical proof of this result we will omit that and will now
look into few examples where we can actually apply this result where we can calculate the
double integral of a given function defined on a rectangle.

So, let us start with our first example.

(Refer Slide Time 24:16)

Example 1. So, evaluate integral over the left angle R; you can also use only single integral
symbol and if you write R then that is also how to say a way to say that it is basically a

182
double integral defined on the rectangle R. So, we do not necessarily have to write a double
integral or double integral R all the time, alright.

.
So, integral ∫𝑅(𝑥 2 + 2𝑥)𝑑𝑥𝑑𝑦. So, that means, our rectangle looks like so, 𝑅 =
[0,1] × [0,2]. So, this is our how to say range of integration. So, that is our rectangle R
and this is where we have to perform the integration, alright.

. 1 2
So, solution: so, let us write ∫𝑅(𝑥 2 + 2𝑥)𝑑𝑥𝑑𝑦. Now, I can write R as ∫𝑥=0 ∫𝑦=0(𝑥 2 +

2𝑥)𝑑𝑥𝑑𝑦. So, now, what we would do? We would integrate with respect to y first, because
here this is obviously, a very nice behaving function. There is a polynomial function in a
way and it is definitely defined and bounded on this rectangle R. So, this double integral
will exist. So, we can actually do that repeated integral property here.

So, let us first treat x as constant. So, I am going to integrate with respect to y. So, if we
1
integrate with respect to y then this will be ∫𝑥=0[𝑥 2 𝑦 + 𝑦 2 ]20 𝑑𝑥 , right because it will be
𝑦2
× 2 will cancel out integral value from 0 to 2.
2

(Refer Slide Time 27:02)

1
And, if I substitute the value then this will be ∫0 [2𝑥 2 + 4]𝑑𝑥 and after if we integrate this
2 2 14
here then it will be [3 𝑥 3 + 4𝑥]10 and this will be (3 + 4), right. So, ultimately . So, if we
3

183
integrate with respect to y first; so, here we see that we integrated with respect to y first
14
and then we integrated with respect to x and our answer is ..
3

So, now let us verify what happens if we integrate with respect to x first and then we
integrate with respect to y. Because, our conclusion is if the double integral exists then
both the repetitive integral must be equal. So, let us see what happens? Alternatively, so,
our result is done here. So, basically we have found the answer it is just that we want to
verify whether we will obtain the same result or not. So, alternatively we can do
1 2
∫𝑥=0 ∫𝑦=0(𝑥 2 + 2𝑥)𝑑𝑥𝑑𝑦.

2 1
So, here if I integrate with respect to x first then I can write it ∫0 [∫0 (𝑥 2 + 2𝑦)𝑑𝑥]𝑑𝑦 and
2 𝑥3
if I integrate with respect to x then this will be ∫0 [ 3 + 2𝑥𝑦]10 𝑑𝑦 and this will be integral
2 1 𝑦 2
∫0 (3 + 2𝑦) 𝑑𝑦 and if I integrate this then this will be [ 3 + 𝑦 2 ]20 . So, this will be 3 + 4;
14
that means, .
3

So, even if we integrate with respect to x first by treating y as constant we would obtain
the similar result and this is basically happening because this double integral exists on this
rectangle R. Sometimes it is also intuitive just looking at the integral you can be able to
make out whether that integral would exist on that interval or not. So, the first criteria is it
has to be bounded on that interval and then we can at least talk about that the double
integral exists.

And, then afterwards it does not matter whether you integrate with respect to x first treating
y as constant and then integrating with respect to y or integrating with respect to y first
treating x as constant and then you integrate with respect to x at the end. Both the repeated
integral would give you the same result like here and you do not have to do both of them
you just have to calculate at least just one of the repetitive integral and that will be a
required answer.

So, we saw an introduction to the double integral in this lecture and we also worked out
an example, where we showed that about the repeated integral will be equal if the integral
double integral exists. We will stop today lecture here and then in the next class we will
continue with some further examples.

184
Thank you.

185
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 18
Double Integral over a Region E

Hello students. So, in the last class, we started with the introduction of double integral, we
also worked out one example. In today’s lecture, we will continue with further examples
actually just to make the concept a little bit more clear in double integral.

(Refer Slide Time 00:40)

So, let us work out an another example. So, example 2, here it says evaluate,
. 𝑑𝑥𝑑𝑦
∬𝑅 (𝑥+𝑦+1)2 where the rectangle R is defined as [0,1] × [0,1]. So, here basically our given

domain or how to say sorry range of integration is basically 0 to 1, 0 to 1. So, this is our
. 𝑑𝑥𝑑𝑦
rectangle. Now, so let us write 𝐼 = ∬𝑅 (𝑥+𝑦+1)2 . So, I can write it as integral
1 1 𝑑𝑥𝑑𝑦
∫𝑥=0 ∫𝑦=0 (𝑥+𝑦+1)2.

So, now I can do that repeated integral calculation here. So, first I will treat y as constant
and we will integrate with respect to x first. So, if I integrate with respect to x first, so let
1 1 𝑑𝑥
us treat y as constant and then I can use that big bracket ∫𝑦=0[∫𝑥=0 (𝑥+𝑦+1)2 ]𝑑𝑦.

186
Now, as we have how to say told several times that when we integrate with respect to x,
then we treat the variable y as constant. So, here this (𝑦 + 1)is basically a constant and
1 𝑑𝑥 𝑑𝑥
then this integral here ∫0 is nothing but that our traditional ∫ 𝑥 2 integral.
(𝑥+𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡)2

(Refer Slide Time 03:11)

1 1
And in order to evaluate that, we simply write integral ∫0 [− ]1 𝑑𝑦, which we already
𝑥+𝑦+1 0

know from our traditional integral calculus results,

1 1 1
And now, we can substitute the value of x here. So, this will be − ∫0 [𝑦+2 − 𝑦+1] 𝑑𝑦. So,
1 1 1
we can rearrange it a little bit, so this will be ∫0 [𝑦+1 − 𝑦+2] 𝑑𝑦 , is not it. I just put the

minus sign inside the bracket and if we integrate now, then this will be
[log 𝑒 (𝑦 + 1) − log 𝑒 (𝑦 + 2)] 10 So, this will be log 𝑒 2 − log 𝑒 1 − log 𝑒 3 + log 𝑒 2.

So, we will basically be having 2 log 𝑒 2 − log 𝑒 3. You can further simplify, but I am just
leaving the result up to here. This value is 1 which we know from the logarithmic results.
So, this is our required result. So, see here also we perform the same repeated integral
property that we integrate with respect to x first and then we integrate with respect to y
like here, all right?

187
(Refer Slide Time 05:28)

So, now, let us consider another example which is a little bit difficult but not too difficult
.
actually. So, evaluate integral over the rectangle R ∬𝑅 𝑦𝑒 𝑥𝑦 𝑑𝑥𝑑𝑦 ; 𝑅 = [0, 𝑎] × [0, 𝑏].
So, this must be example 3. It is pretty much the similar interval like the previous example.
So, I am not drawing that, that area of integration. It is pretty much straight forward here.

So, now I am going to write I as integral over the rectangle R. So, 𝐼 =

. 𝑎 𝑏
∬𝑅 𝑦𝑒 𝑥𝑦 𝑑𝑥𝑑𝑦 = ∫𝑥=0 ∫𝑦=0 𝑦𝑒 𝑥𝑦 𝑑𝑥𝑑𝑦. This double integral also exist on this rectangle R.

Just looking at this integrand, you can be able to tell. And next, we see that we have
𝑎 𝑏
∫𝑥=0 ∫𝑦=0 𝑦𝑒 𝑥𝑦 𝑑𝑥𝑑𝑦. So, if I substitute xy as z let us say by treating y as constant, then this

will be basically our usual method of substitution formula.

So, let us see how we can how we can do that or we can separate this. So, there are several
ways to do this. So, let us separate this y and then I will integrate with respect to x first
𝑏 𝑎
and if we integrate with respect to x first, then this will be ∫𝑦=0 𝑦[∫𝑥=0 𝑒 𝑥𝑦 𝑑𝑥]𝑑𝑦. So, this
𝑏 𝑒 𝑥𝑦 𝑎
will be ∫𝑦=0 𝑦[ ] 𝑑𝑦, then this y, this y will get cancelled and then we will have
𝑦 𝑥=0
𝑏
∫𝑦=0(𝑒 𝑎𝑦 − 1)𝑑𝑦.

𝑒 𝑎𝑦 𝑒 𝑎𝑏 1
So, when we integrate this one this will be [ − 𝑦]𝑏𝑦=0. We will have ( − 𝑏) − 𝑎 =
𝑎 𝑎
1
(𝑒 𝑎𝑏 − 𝑎𝑏 − 1).
𝑎

188
 1
So, basically our result is 𝑎 (𝑒 𝑎𝑏 − 𝑎𝑏 − 1) , all right. So, yeah now, it is correct.

So, you see that you are just looking at the integrand, we see that this integrand is definitely
bounded on this interval [𝑎, 𝑏] and based on that, I can now treat this in double integral by
the method of this repeated integral. So, I started with treating this integrand as first of all
integrating with respect to x keeping y as constant and after I got this function with respect
to y, I integrated it and then this is our final result.

So, this is another example where we did repeated integration for a double integral.

(Refer Slide Time 10:38)

We can consider one more example before we look into the change of order of integration.
.
So, it is evaluate, integral over the rectangle R ∬𝑅 𝑥 sin(𝑥 + 𝑦)𝑑𝑥𝑑𝑦 where our rectangle
𝜋
𝑅 = [0, 𝜋] × [0, 2 ]. So, again just look at this integrand, it is basically x times R

trigonometrical function. So, of course, sin actually. So, the integrand is bounded on this
on this rectangle R and as a result of which, this double integral will exist and now we will
treat this double integral by the by that repeated integral method.

𝜋
. 𝜋
So, let us write it as 𝐼 = ∬𝑅 𝑥 sin(𝑥 + 𝑦)𝑑𝑥𝑑𝑦 = ∫𝑥=0 ∫𝑦=0
2 𝑥𝑠𝑖𝑛(𝑥 + 𝑦)𝑑𝑥𝑑𝑦 . So, first of

all, I will integrate with respect to y. So, let us see what happens. It is totally depend on
you whether you would like to integrate with respect to x first or with respect to y. So, here
in this case, we choose to integrate with respect to y first.

189
 𝜋
𝜋
So, here I can write ∫𝑥=0 𝑥[∫𝑦=0
2 𝑠𝑖𝑛(𝑥 + 𝑦)𝑑𝑦]𝑑𝑥. So, here y is basically a constant at the

moment. So, it does not play any role here and we will just integrate with sorry here x is a
constant. So, x does not play any role here. We will basically integrate with respect to y.
𝜋
𝜋
So, ∫𝑥=0 𝑥[− cos(𝑥 + 𝑦)]02 𝑑𝑥 .

(Refer Slide Time 13:14)

𝜋 𝜋
So, now this will be integral ∫𝑥=0 𝑥[− cos (2 + 𝑥) + cos 𝑥]𝑑𝑥 . So, this will be
𝜋
∫𝑥=0(𝑥 sin 𝑥 + cos 𝑥)𝑑𝑥 = 𝜋 − 2. So, cos will change to sin and since sin is positive sorry,
cos is negative here, so negative and negative will turn into a positive. So, here we will
have (𝑥 sin 𝑥 + cos 𝑥).

And now, we will basically integrate with respect to x here and we will do that by doing
that integration by parts for the first term and we will basically integrate the second term.
So, the second term will result into integral of cos x and integral of cosx is basically sinx
and then from 0 to 𝜋, both the values will be 0 and this one can be treated by integration
by parts. So, if you do all those things, so I leave these in this part up to the students
because it is a basic trigonometric integration which you have done in your plus 2 level.

So, if you do the integration by parts, then ultimately you would obtain as (𝜋 − 2). So,
that will be your final result. So, here also we started with a given double integral, here we
started with treating x as constant and we integrated with respect to y first and afterwards,

190
we integrated with respect to x. And after if you do this integration by parts, then ultimately
you will obtain (𝜋 − 2) as the answer.

So, this was one such example where we use this repeated integral method.

(Refer Slide Time 15:53)

1
0≤𝑥<𝑦≤1𝑦2
Our next example is for the function f defined by 𝑓(𝑥, 𝑦) = { 1
. So,
− 𝑥2 0 ≤ 𝑦 < 𝑥 ≤ 1
1 1 1 1
then show that integral ∫0 𝑑𝑥 ∫0 𝑓(𝑥, 𝑦)𝑑𝑦 ≠ ∫0 𝑑𝑦 ∫0 𝑓(𝑥, 𝑦)𝑑𝑥.

So that means, the double integral or to be very precise the repeated integral are not equal.
And if the repeated integral are not equal, then that case, the double integral does not exist
between the interval between the interval [0,1] × [0,1]. So, let us see how we can evaluate
this. So, solution here, it is better to write as 0 and then, let us include the point 0; basically
the endpoints. So, here (0,0) is the only point of infinite discontinuity because the function
is unbounded at the point (0,0).

1
So, the function is unbounded at the point (0,0). So, then integral ∫0 𝑓(𝑥, 𝑦)𝑑𝑦 =
𝑥 1 1 1 𝑦 1 1 1 1
− ∫0 𝑑𝑦 + ∫𝑥 𝑑𝑦 = −[𝑥 2 ]0𝑥 − [𝑦] = − 𝑥 + 𝑥 − 1 = −1.
𝑥2 𝑦2 𝑥

So, we basically divided this interval [0,1] by [0, 𝑥]. So, we have used these two values.

191
(Refer Slide Time 19:36)

𝑦
Now, when we integrate, then the first term will be − 𝑥 2 integral value will be from x

running from 1 from 0 sorry, y running from 0 to x plus we have 1 by 𝑦 2 .

1
So, this will be − 𝑦 and then y will be running from x to 1. So, if I substitute the value,
1 1
then this will − 𝑥 + 𝑥 − 1 = −1 .

So, ultimately the value is −1 and terefore, from here if I want to evaluate, the left hand
1 1 1
side. So,∫0 𝑑𝑥 ∫0 𝑓(𝑥, 𝑦)𝑑𝑦 = ∫0 (−1)𝑑𝑥 = −[𝑥]10 = −1. So, this is basically −1.

1
Now, so we have found the value as −1. Now, integral ∫0 𝑓(𝑥, 𝑦)𝑑𝑥, so what will be the
𝑦 1
value of this one? So, I can again write integral ∫0 𝑓(𝑥, 𝑦)𝑑𝑥 + ∫𝑥=𝑦 𝑓(𝑥, 𝑦)𝑑𝑥. So,
𝑦 1 1 1
integral ∫𝑥=0 𝑦 2 𝑑𝑥 + ∫𝑥=𝑦 (− 𝑥 2 ) 𝑑𝑥.

all right.

192
(Refer Slide Time 22:13)

1 1
And now, if we integrate like we did before, then we will basically obtain +1−𝑦 =
𝑦

1 and then this answer is 1. So, therefore, from here if we have integral
1 1 1
∫0 𝑑𝑦 ∫0 𝑓(𝑥, 𝑦)𝑑𝑥 = ∫0 1𝑑𝑦 = [𝑦]10 = 1. So, this will be 1.

1 1 1 1
So, that means, ∫0 𝑑𝑦 ∫0 𝑓(𝑥, 𝑦)𝑑𝑥 ≠ ∫0 𝑑𝑥 ∫0 𝑓(𝑥, 𝑦)𝑑𝑦.

So that means, these two repeated integrals are not equal which we wanted to prove and
from here we can also say that, so remark a small remark; the double integral
1 1
∫0 ∫0 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦 does not exist. And the reason for it not existing is that this function is
unbounded or it has an infinite point of discontinuity at the point (0,0) and that is why
double integral does not exist or they these two repeated integrals they are not same. So,
this is a very nice example where we can see that if you have even a little bit problem in
your integrand your function f, then your double integral would not exist.

Although your repeated integral may or may not exist, but they will not be same and at the
same time, your double integral how to say would not exist because if it exists, then both
these repeated integrals where is that, they also should be same, but it is not happening in
this case. So, this is one example which shows that and it is very interesting actually.

So, next, we will look into calculation of double integral. So, not only on the rectangle R,
but on a certain region, so up until now, we had a very nice rectangle where we could write

193
 𝑏 𝑑
integral ∫𝑎 ∫𝑐 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦 where the integral was evaluated on a rectangle R. But, what
would happen if you have a region or a certain type of domain?

(Refer Slide Time 25:14)

So, calculation of double integral over a region let us say E over a region E, if I write R,
then we might confuse with rectangle.

So, let us write the region E. So, the statement goes like this. If f is a continuous function,
f is of course a function of two variable on a domain E which is bounded by. So, the domain
is now given by which is bounded by the curves 𝑦 = 𝜑(𝑥) and 𝑦 = 𝜓(𝑥).

So, these two are the curves that are defining the region or the domain E where 𝑎 ≤ 𝑥 ≤
𝑏. where 𝜑 and 𝜓 are continuous and 𝜓(𝑥) ≤ 𝜑(𝑥) ∀𝑥 ∈ [𝑎, 𝑏], then the double integral
.
over the region E or over the domain E can be given as, integral ∬𝐸 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦 =
𝑏 𝜑(𝑥)
∫𝑥=𝑎[∫𝜓(𝑥) 𝑓(𝑥, 𝑦)𝑑𝑦]𝑑𝑥. So, this is the way we define the double integral on a region E.

Similarly, if our function f is continuous on a domain E which is bounded by the curve

𝑥 = 𝜑(𝑦) and 𝑥 = 𝜓(𝑦). So, if we have the curves 𝑥 = 𝜑(𝑦) and 𝑥 = 𝜓(𝑦) where 𝑐 ≤
𝑦 ≤ 𝑑, and both 𝜑 and 𝜓 are continuous ∀𝑦 ∈ [𝑐, 𝑑], then our double integral can be
. 𝑑 𝜓(𝑦)
written as integral ∬𝐸 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦 = ∫𝑦=𝑐[∫𝑥=𝜑(𝑦) 𝑓(𝑥, 𝑦)𝑑𝑥]𝑑𝑦 we integrate with

respect to x, substitute the value and then integrate with respect to y.

194
So, either you can integrate with respect to x first by putting the values of these two curves,
but for the variable x and then, we integrate with respect to y at then because we will have
only functions of y or we can first integrate with respect to y by putting, the values for y
as 𝜓(𝑥) and 𝜑(𝑥) and then we integrate with respect to x. So, either way would do. We
will stop here for today. And in the next lecture, we will basically start with the examples
where instead of rectangle; we will have regions to perform the double integral.

Thank you.

195
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 19
Example of Integral over a Region E

Hello, students. So, up until last class we looked into double integrals, how do we calculate
the double integral on a rectangular domain and we worked out several examples. Now,
we also introduced about the concepts of a double integral on a bounded region. So, instead
of having a rectangle if we have a bounded region, then how can we calculate the double
integral in such cases. So, I have also showed you the idea how do we do the calculation.

Today, we will work out a few examples just to make those ideas are slightly more clearer.

(Refer Slide Time: 00:57)

So, today basically we are still continuing with the double integral part. So, we will start
with this double integral and if time permits, then we shift to change the order of
integration, alright. So, let us worked out a few examples. Let us try to work out few
examples, alright.

196
(Refer Slide Time: 01:27)

So, the first example is example one. So, the first example is: evaluate the integral on a
bounded region; let us say E we are avoiding R for the region because by R be also meant
.
rectangle. So, we will write ∬𝐸 𝑥 3 𝑦 2 𝑑𝑥𝑑𝑦, where E is the region given by 𝑥 2 + 𝑦 2 ≤ 𝑎2 ,
where a is any positive number.

So, here we have to evaluate or we have to calculate this double integral over the
circle 𝑥 2 + 𝑦 2 = 𝑎2 , where a is basically the radius of the circle. So, that means, our
region of integration so, if we have x and y, so, our region of integration would be the
circle, where a is the radius. So, we have to perform the area in this region alright.

Now, if we have to perform the integration in this region. So, in order to do that, we have
to first of all find out the range for x and the range for y. So, we will write here, the given
region is the circle 𝑥 2 + 𝑦 2 ≤ 𝑎2 . So, obviously, x is varying from −𝑎 to +𝑎; similarly, y
is varying from −𝑎 to +𝑎.

So, we can write the range for y as −√𝑎2 − 𝑥 2 ≤ 𝑦 ≤ √𝑎2 − 𝑥 2 . So, this will be the range

for y and then we can guess the range for a. So, the range for a will be the range or the
variable x, the range for x is so; this is the range for our variable x.

197
So, now we have the range for the variable x which is −𝑎 to a and range for the variable y
𝑎 √𝑎2 −𝑥 2
𝐼 = ∫−𝑎 ∫−√𝑎2 −𝑥 2 𝑥 3 𝑦 2 𝑑𝑥𝑑𝑦. So, it is very easy to see I mean how we are deriving the
range for x and y. So, obviously, x is varying from −𝑎 to 𝑎.

So, the range for x is already known to us. Now, y will vary if x is known then y will vary
from here to here. So, this is the how to say the domain for the variable y. So, y will vary
from this point and then it will cover all these points and then it will go up to here. So, that
is how we are drawing we are how to say calculating the range for y.

So, now we substitute the range for x and range for y and that is our area or the region of
integration. So, that is our region for into a region of integration. So, we substituted or we
replaced the region E by the limits of x and y. And, now from here like we saw in the
previous lecture that we separate the variable now we separate the how to say here the
integral with respect to x and with respect to y.

𝑎 √𝑎2 −𝑥 2
So, if we separate, then it will be ∫−𝑎 𝑥 3 𝑑𝑥 ∫−√𝑎2 −𝑥 2 𝑦 2 𝑑𝑦.

(Refer Slide Time: 06:15)

𝑎 1 2 2
And, this can be evaluated as integral ∫−𝑎 𝑥 3 𝑑𝑥[3 𝑦 3 ]√𝑎 −𝑥
−√𝑎2 −𝑥 2
. So, if we substitute the value
3
2 𝑎
then this will be 3 ∫−𝑎 𝑥 3 (𝑎2 − 𝑥 2 )2 𝑑𝑥, alright. So, this is the function of x and now we

have to evaluate this integral.

198
An another interesting point in this integrand is that if I substitute for 𝑥 = −𝑥 then that
because this will behave as an odd function. So, this whole thing is actually an odd
function, and we know a small result from integral calculus which says that if f is an odd
𝑎
function then integral ∫−𝑎 𝑓𝑑𝑥 = 0.

So, this results so, this is a very well known result from the integral calculus and since we
have an odd function and since the range of integration is from −𝑎 to a the value of this
2
integral will be 0. So, I can directly write × 0.so, ultimately 0. So, that means, integral
3
.
∬𝐸 𝑥 3 𝑦 2 𝑑𝑥𝑑𝑦 = 0, where E was the given circle basically. So, this is how we evaluate on
the double integral over a given region and in this case of course, the answer is 0.

(Refer Slide Time: 08:43)

So, let us see an another example where we have to work out a little bit the area of the
region of integration in a way. So, let us consider another example. So, evaluate integral
∬(𝑥 2 + 𝑦 2 )𝑑𝑥𝑑𝑦 over the domain or over the region bounded by 𝑥 2 = 𝑦 and 𝑦 2 = 𝑥. So,
this is our how to say area the domain of integration.

So, let us first draw the region. So, this is my x, this is my y that is 0. So, we have a first
parabola which is 𝑦 2 = 𝑥; so, obviously, it will pass through the origin and then we have
second parabola 𝑥 2 = 𝑦 which will also pass through the origin. So, this is the parabola
𝑦 2 = 𝑥 and then we have another parabola which is again passing through the origin. So,

199
something like 𝑥 2 = 𝑦 . So, this is. So, I am not very good at drawing. So, roughly this is
what our second parabola will look like.

Now, the domain of integration is bounded by these two parabola. So, it is bounded by
these two parabola means the region which is common between these two parabolas. So,
if you can see that this parabola and this parabola enclosed this region here. So, this is our
region E or the domain of integration and they intersect these two parabola they intersect
at certain point.

So; that means, our domain of integration or the domain or the region E here let us say;
region E is basically starting from here to here and of course, it is bounded by these two
parabola. So, how do we calculate this area or this point of intersection?

So, here the two parabolas 𝑥 2 = 𝑦 and 𝑦 2 = 𝑥 intersect at the following points. So, if we
want to find the point of intersection we basically substitute. So, we have 𝑦 2 = 𝑥 and we
substitute for y we substitute 𝑥 2 . So, this will turn out to be 𝑥 4 = 𝑥.

So, from here we can do some factorization. So, this will be 𝑥(𝑥 − 1)(𝑥 2 + 𝑥 + 1) = 0,
right, So, from here these two will give us the real values of x. So, from here we will
get 𝑥 = 0, 𝑥 = 1 and this will give us some I would say an imaginary value. So, if we
solve this equation if we make this one equals to 0, then in that case we will get imaginary
values for x. So, we are not interested in that. So, we will take only 𝑥 = 0 and 𝑥 = 1 as
the as the two solutions. So, this equation is a fourth order equation. So, it will definitely
have four solutions, but the other two are imaginary. So, we are only interested in real
solutions.

So, the real solutions are 0 and 1. And, when 𝑥 = 0 so, from here so, when 𝑥 = 0 we have
𝑦 = 0 and when 𝑥 = 1, then we have 𝑦 = 1. So, that means, the two point of intersections
are (0,0) and (1,1) . So, at these two points these two parabola intersect. So, now we have
the point of intersection and we can now calculate the limits. Let us call these two points
as O and P. So, they intersect at the point O which is (0,0) and the point P which is (1,1).

So, now, here we have and let us say this is our point Q. anyway. So, this is my point Q
which is how to say any point on this lower curve. So, of course, at the point so, we have
𝑦 = 𝑥 2 at the point Q and on the upper point let us say P, Q R. So, and 𝑦 = √𝑥 at the point
R, right. So, on this how to say lower lying parabola we have 𝑦 = 𝑥 2 . So, that is what I

200
have written that it is 𝑦 = 𝑥 2 at the point Q. So, Q will be any arbitrary point on this
parabola and it will always be 𝑦 = 𝑥 2 .

However, any point on this upper parabola which we have 𝑦 = √𝑥. So, that is where the y
varies. So, y is varying from 𝑥 2 to √𝑥 and if I put and x varies from 0 to 1. So, that is x is
varying from 0 to 1, alright.

(Refer Slide Time: 15:25)

So, let us see how we calculate the double integral. So, instead of writing the region, now
1 √𝑥
I am writing the range for x. 𝐼 = ∫𝑥=0 ∫𝑥 2 (𝑥 2 + 𝑦 2 )𝑑𝑥𝑑𝑦 . So, now, we can again do that
separation thing. So, we will first integrate with respect to y.

1 √𝑥
So, let us take out ∫𝑥=0 𝑑𝑥 ∫𝑥 2 (𝑥 2 + 𝑦 2 )𝑑𝑦 and if we integrate so, when we integrate this
1 𝑦 3 √𝑥
will yield ∫𝑥=0[𝑥 2 𝑦 + ] 2 𝑑𝑥 and after integrating and after substituting we will obtain
3 𝑥
3
1 1
∫0 [𝑥 2 (√𝑥 − 𝑥 2 ) + 3 (𝑥 2 − 𝑥 4 )] 𝑑𝑥 and this is just how to say an algebraic expression to

integrate. I am pretty sure you have done such kind of integration in your plus 2 level. So,
I will leave this part up to the students and once you integrate you will ultimately obtain
6
. So, that is the required answer.
35

So, here let us go to the previous slide. So, here the how to say region or the domain was
bounded by these two parabola. So, drawing this figure always helps them and it is a

201
suggestion that whenever you come across a problem like that it is always advisable to
first draw the domain of integration.

Because from there you can be able to draw the conclusion that what will be my range of
x and what will be my range of y, like we did here. And, then just substitute the values of
replace the region E by the values of x and y like we did here and then we just do the
normal integration and that will give you the answer.

So, this was another example where we calculated the double integral bounded by a region.
We will see one or two more examples like that before we jump to our new topic which is
change of order of integration.

(Refer Slide Time: 18:15)

.
So, let us consider another example. Evaluate integral ∬𝐸 𝑦𝑑𝑥𝑑𝑦 over the part of the plane
bounded by the line 𝑦 = 𝑥 and the parabola 𝑦 = 4𝑥 − 𝑥 2 .

So, here the given curves are slightly tricky. So, this is not our traditional parabola, this is
a slightly different form of parabola, but it is still a parabola and of course, the region is
bounded between or the domain of integration is bounded between this parabola and the
straight line. So, as I was saying that it is always helpful to draw the domain of integration
first.

202
So, let us draw. So, that is our x-axis and this is our y-axis that is the origin. Here we can
write this sentence here 𝑦 = 4𝑥 − 𝑥 2 . So, we can be able to write it as (𝑥 − 2)2 = −(𝑦 −
4). So, the vertex of this parabola is at the point (2,4).

So, let us draw the parabola first. So, let us say this is my vertex and since we have 𝑥 2 =
−4𝑎𝑦 type. So, that means, it will be as minus. So, it will be an inverted parabola. So, it
will go something like this x and it the region is bounded between the points this as straight
line and the parabola.

So, let us draw a straight line passing through the origin. So, this is the region of integration
and of course, they intersect at two points the first one is here; I can call it as let us say P
and this region is Q, any point on this straight line is R and any point on this parabola is S.
So, this is our 𝑦 = 𝑥 and this is our (𝑥 − 2)2 = 4 − 𝑦 , alright.

So, now we have to guess the values for these we have to calculate not guess, but we have
to calculate these two point of intersection. So, here this is our parabola and it intersects
with the line 𝑦 = 𝑥 at the following points. So, we have 4𝑥 − 𝑥 2 = 𝑦 and we substitute
for 𝑦 = 𝑥 here. So, 4𝑥 − 𝑥 2 = 𝑦 = 𝑥 now if I bring this x here. So, this is 3𝑥 − 𝑥 2 = 0and
this can be written as 𝑥(3 − 𝑥) = 0.

So, from here we will basically obtain 𝑥 = 0 and 𝑥 = 3 and when x is 0, when x is 0, y is
0 and when x is 3, then y is 3.

(Refer Slide Time: 22:27)

203
So, that means, this point of intersection was actually origin. So, they are intersecting at
the origin. So, this origin O we can rename it as P, so, we no need to rename twice. So,
they are basically intersecting at the point (0,0); we just saw here and then we they are
intersecting at the point (3,3) which we also saw.

So, they are intersecting at the point (0,0) and they are intersecting at the point (3,3) and
this is (0,0) point is named as P the point this (3,3) point is named as Q now we have the
area or sorry we have the point of intersection, we need to find out the limits for the
variable x and y so, that we can replace this region E with those limits, alright.

So, here any point on this straight line is given by 𝑦 = 𝑥. So, we can write the point like
we did in the previous example, so, we can write the point we have what we have named
it we have named it R. So, the point R is R at the point R we can write at the point R we
have 𝑦 = 𝑥 and at the point S, right? So, at the point S we have 𝑦 = 4𝑥 − 𝑥 2 . So, we have
𝑦 = 4𝑥−𝑥 2 .

So, then our double integral will be first we replace the region E by the limits of x and y,
3 4𝑥−𝑥 2
so, 𝐼 = ∫𝑥=0 ∫𝑥 𝑦𝑑𝑥𝑑𝑦alright.

3 4𝑥−𝑥 2
So, again we do the separation. So, first we write this as integral ∫𝑥=0 𝑑𝑥 ∫𝑥 𝑦𝑑𝑦. Now,
3 𝑦2 2
if I integrate then this will be ∫𝑥=0[ 2 ]4𝑥−𝑥
𝑥 .

(Refer Slide Time: 26:01)

204
 1 3
And, this can be written as 2 ∫0 (𝑥 4 − 8𝑥 3 + 15𝑥 2 )𝑑𝑥 right. And, then we just use the

integration. So, it is the basic how to say integration of an algebraic function. So, I leave
to evaluate this integral up to the students and if you evaluate then at the end you will
54
ultimately get . So, this is how we calculate the area bounded by this parabola and this
5

is straight line.

So, remember, the rule of thumb is that you always need to draw the figure first. So, once
you draw the figure first it will become very clear that what would be our domain of
integration because you do need that domain of integration in order to replace this E with
those limits for x and y. So, here in this case we draw the domain first and we can see that
for this inverted parabola and this is straight line they intersect. So, what are the two points
there are intersecting we can solve actually and after solving we obtained that they intersect
at the point (0,0) and the point (3,3).

So, now that we have the point of intersection we know the range for the variable x and y
and then we basically look at the variable y it is always suggestible to integrate with respect
to y first and then do the integration with respect to x and in order to do the integration
with respect to y first you need to find out the range for y here the range for y is basically
y is varying in this direction, right.

So, y is varying in this direction and the domain of integration is bounded between these
two curves. So, first if y is running between these two how to say these two curves in a
way, so, we can take any arbitrary point on this straight line and at those points we will
have 𝑦 = 𝑥 and if we take any arbitrary point on that parabola then at those points we will
have 𝑦 = 4𝑥 − 𝑥 2 . So, that is what we have done, ok.

So, here there is a small error, we can correct it. So, 4𝑥 − 𝑥 2 , sorry. So, this is basically
4𝑥 − 𝑥 2 this is also 4𝑥 − 𝑥 2 . And then just calculating this algebraic expression will give
you the required answer. So, this is the way we evaluate the integral over a certain region
or over a bounded region in a way. So, it is not only that you always have a rectangles for
the double integral, you may also have regions bounded regions basically and you that is
how you basically evaluate those integrals you just have to draw the domain and then guess
the range for the variable x and y and there are several examples which we can work out,
but there are more or less of the similar type.

205
So, it is the nothing that I can teach is just that you have to practice a lot of examples like
that, but the basic idea is same. I will include some problems like that in your assignments
for you to practice and for you to clear your doubts and we will also look how do you say
provide the solutions of them as well and hopefully it will help you clear out the doubts if
you have at all, in this double integral over a region or over a rectangular domain.

So, we will stop today’s lecture here and in the next class we will begin with change of
order of integration.

Thank you.

206
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 20
Change of variables in a Double Integral

Hello students. So, in the last class we looked into the examples of double integrals on
rectangular as well as unbounded regions. Today we will start with a change of order of
integration and Jacobean transformation. So, how you can how to say transform a double
integral into a double integral using Jacobean transformations and all.

So, we will start with a basic theory, the proof of which we will avoid, but I will just give
you an idea that how you do the change of variables in a double integral or how you change
the order of integration. So, let us start change of variables in a double integral.

(Refer Slide Time: 01:04)

So, the reason why we do the change of variable is that, many double integrals can be
evaluated easily if we do the change of variables.

So, sometimes you may have a very complicated double integral given to you and it might
seem that it is very difficult to do that, but if you do some certain change of variables, then
the overall integral might reduce to a very simple expression and that will be easy to
evaluate and of course, after doing the change of variables, your answer will still remain

207
the same. So, it would not happen that the value of that double integral would change. So,
it would still remain the same; however, doing that change of variable actually saved us
from lot of trouble actually.

So, let us see. Suppose we have a bounded function. So, let 𝑓(𝑥, 𝑦) be a continuous
function on a domain or region on a domain E, and let the variables x and y in the double
.
integral ∬𝐸 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦 be changed to u and v by following relation. So, what is that
relation? So, we are using 𝑥 = 𝜑(𝑢, 𝑣) and 𝑦 = 𝜓(𝑢, 𝑣). So, these are some kind of
transformation, where 𝜑 and 𝜓 are continuous functions on E, with continuous first order
partial derivative all right; partial derivatives in a certain region 𝐸1 of uv plane.

So, since we are transforming the variable x and y to the variable u and v so; obviously,
from the region E we will get transferred to a region 𝐸1 because for the variable u and v
after doing the transformation they might belong to a different region. So, they might
bound a different region; however, the dimension will not change. So, x and y they signify
the two dimensional geometry. So, then in that case once you are using the transformation,
we will get away the transformation function as 𝜑(𝑢, 𝑣), and they will get again transferred
to a two dimensional domain. Of course, the region might be two dimensional geometry,
but of course, the region might be different and that is why we are saying that these you
have 𝜑 and 𝜓, they are how to say they are continuous and they have continuous partial
derivative in a certain region 𝐸1 so, because they are now in uv plane.

(Refer Slide Time: 06:04)

208
 .
So, then the integral ∬𝐸 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦 can be changed let us write in a new page integral
.
∬𝐸 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦 . So, we are changing 𝑥 = 𝜑(𝑢, 𝑣). So, let us write
.
∬𝐸 𝑓(𝜑(𝑢, 𝑣), 𝜓(𝑢, 𝑣))|𝐽|𝑑𝑢𝑑𝑣 however, we will have an external factor here which is
1

this |𝐽| and this is not actually |𝐽| this has a meaning. So, this has a meaning which is so, J
is called the Jacobian of the transformation and when I say transformation, I basically
mean this here.

So, this is my transformation and J is the Jacobian of that transformation which is given
𝜕𝑥 𝜕𝑥 𝜕𝑥 𝜕𝑥
𝜕(𝑥,𝑦)
by given by 𝐽 = 𝜕(𝑢,𝑣)and its basically (𝜕𝑢
𝜕𝑦
𝜕𝑣
𝜕𝑦
) and |𝐽| = |𝜕𝑢
𝜕𝑦
𝜕𝑣
𝜕𝑦
|, is basically
𝜕𝑢 𝜕𝑣 𝜕𝑢 𝜕𝑣
𝜕𝑥 𝜕𝑥

determinant of this matrix here. So, that is |𝜕𝑢

𝜕𝑦
𝜕𝑣
𝜕𝑦
| .
𝜕𝑢 𝜕𝑣

So, this is basically our Jacobian matrix. So, first of all when we are doing how to say
change of variable, we have to find out these transformation. So, we are the one who has
to calculate these two transformation and once we get these two transformation, then we
have to calculate the Jacobian determinant so, that we can do the change of variable. But
before that we cannot simply put any transformation here. So, that transformation should
also have these two properties. So, it should be a continuous function on 𝐸1 and it should
have continuous first order partial derivatives as well.

So, once we have all these ingredients and we can do the change of variable in our integral.
Now we might or one might ask that change of how do we change.

209
(Refer Slide Time: 09:05)

So, an example interesting example could be; change the variable let us say any we have
.
an integral ∬𝐸 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦 to the polar coordinate from Cartesian coordinate; where E is
given by 𝑥 2 + 𝑦 2 ≤ 𝑟 2; r can be any a b c whatever radius we prefer.

So; obviously, for a circle we substitute 𝑥 = 𝑟 cos 𝜃 and we substitute 𝑦 = 𝑟 sin 𝜃. So,
from here we will have. So, we can calculate the Jacobean. So, let us calculate the Jacobean
determinant. The Jacobean determinant is determinant of del x. So, our new variables are
r and 𝜃. So, I can write this as a function of r and 𝜃. So, 𝜑(𝑟, 𝜃) and this one is basically
𝜓(𝑟, 𝜃), and what is our 𝜑(𝑟, 𝜃)? Its 𝑟 cos 𝜃 and what is our 𝜓(𝑟, 𝜃), ? It is 𝑟 sin 𝜃 .

𝜕𝑥 𝜕𝑥
𝜕𝑟 𝜕𝜃 𝜕𝑥 𝜕𝑥 𝜕𝑥
So, we have |𝐽| = |𝜕𝑦 𝜕𝑦
|. So, what is my ? would be cos 𝜃 simply will be
𝜕𝑟 𝜕𝑟 𝜕𝜃
𝜕𝑟 𝜕𝜃
𝜕𝑦 𝜕𝑦
−𝑟 sin 𝜃 and then 𝜕𝑟 will be sin 𝜃 only and 𝜕𝜃 will be 𝑟 cos 𝜃. So, if we evaluate this
cos 𝜃 −𝑟 sin 𝜃
determinant | | , then this will be 𝑟(𝑐𝑜𝑠 2 𝜃 + 𝑠𝑖𝑛2 𝜃).
sin 𝜃 𝑟 cos 𝜃

So, ultimately we will obtain r. Therefore, our integral I which is now transferred into a
.
polar coordinate, our new region 𝐸1 then 𝐼 = ∬𝐸 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦 =
.
∬𝐸 𝑓(𝑟 cos 𝜃, 𝑟 sin 𝜃)𝑟𝑑𝑟𝑑𝜃. This 𝐸1 can be replaced by 0 to r and from 𝜃 can be replaced
1

by 0 to 2𝜋.

210
So, we can replace the c 1 by the limits of r and 𝜃 and this is the required transformed or
changed integral from Cartesian coordinate system to polar coordinate system. And using
this kind of formula. So, here one can be able to transform or change the variable of a
given double integral basically. So, let us work out and another example. So, this one is
actually based on polar coordinate system.

(Refer Slide Time: 12:47)

.
So, its evaluate integral ∬𝐸 sin 𝜋(𝑥 2 + 𝑦 2 )𝑑𝑥𝑑𝑦 where E is given by 𝑥 2 + 𝑦 2 ≤ 1, yes.
So, here our domain of integration is basically this circle with radius one right. So, this is
our region E sorry this is x this is y. So, that is origin.

Now, we can see that here we have sin 𝜋(𝑥 2 + 𝑦 2 )𝑑𝑥𝑑𝑦 . If we want to evaluate using our
traditional method like finding the limit for x which is 0 to 1 and the limit for y is
−√1 − 𝑥 2 to √1 − 𝑥 2 , then it might become a little bit difficult. However, if we do the
change of variable from Cartesian to the polar coordinate system it might be beneficial.

So, let us see whether doing change of variable helps or not, that traditional method will
always work, but its just that it might get tedious or it might get difficult at some point or
the calculation basically. However, doing that the change of variable might make our life
a little bit easier. So, let us see. So, substitute 𝑥 = 𝑟 cos 𝜃 and 𝑦 = 𝑟 sin 𝜃; where our r is
running from 0 to 1 because the radius is 1 and 𝜃 will run from 0 to 2𝜋 right.

211
Now, we know from the previous example, that |𝐽| = 𝑟 So, here we can do the similar
calculation which I am leaving up to the students and I am just going to use that answer,
which is |𝐽| = 𝑟 = 1. So, let us write just 1 and now we have sin 𝜋 yes.

.
So, now we have integral 𝐼 = ∬𝐸 sin 𝜋(𝑥 2 + 𝑦 2 )𝑑𝑥𝑑𝑦 . So, now, since I am transforming
this whole integral into polar coordinate system, I can write the range for r which is 0 to 1,
I can write range for 𝜋 which is a 𝜃 which is 0 to 2𝜋 and then I have
1 2𝜋
∫𝑟=0 ∫𝜃=0 sin 𝜋(𝑟 2 𝑐𝑜𝑠 2 𝜃 + 𝑟 2 𝑠𝑖𝑛2 𝜃)𝑟𝑑𝑟𝑑𝜃 and. So, this value will remain r as it is,
because we are taking r as a variable. So, this will become 𝑟𝑑𝑟𝑑𝜃 . So, we take 𝑟 2 common.
So, it will be 𝑐𝑜𝑠 2 𝜃 + 𝑠𝑖𝑛2 𝜃 then that value will be 1 and then we are left with
1 2𝜋
∫𝑟=0 ∫𝜃=0 sin 𝜋𝑟 2 𝑟𝑑𝑟𝑑𝜃. So, here it is a very simple integral to evaluate we can substitute
𝑡 = 𝑟 2.

(Refer Slide Time: 17:02)

2𝜋 1
So, putting 𝑡 = 𝑟 2 and we separate these two integrals. So, ∫0 𝑑𝜃 ∫𝑟=0 sin 𝜋𝑟 2 𝑟𝑑𝑟𝑑𝜃. So,
1
let us put 𝑡 = 𝑟 2 , then 𝑑𝑡 = 2𝑟𝑑𝑟 and then this whole thing will reduce − 2𝜋 or I can write

one more step.

1 2𝜋 𝜋
So, this can be written 2 ∫0 𝑑𝜃 ∫0 sin 𝜋𝑡𝑑𝑡 and if I integrate then this will reduce to
1 2𝜋
− 2𝜋 ∫0 [cos 𝜋𝑡]01 𝑑𝜃.

212
So, we will have cos 𝜋𝑡 and this will be from 0 to 1 d t, now this can be written as
1 2𝜋 1 2𝜋 1
− 2𝜋 ∫0 [cos 𝜋 − cos 0]𝑑𝜃. So, basically− 2𝜋 [−2] ∫0 𝑑𝜃 So, this will be 𝜋 × 2𝜋. So, the

answer is ultimately 2. So, here as you can see that initially we were given a very
complicated integral to evaluate, but just by changing the variables; that means, changing
the variables from Cartesian coordinate system to polar coordinate system, we can how to
say calculate this whole integral in a lot more simpler manner.

So, this just in reduced to sin 𝜋𝑟 2 𝑟𝑑𝑟𝑑𝜃 and calculating that one, it was very easy because
we substituted 𝑡 = 𝑟 2 which will give you 𝑑𝑡 = 2𝑟𝑑𝑟, then you can adjust this factor here
and the rest of the calculation will become fairly easy. Now not only always polar
coordinates, but one can also try to how to say transform them into a spherical coordinate
system. If we have a triple integral, then you can also transform them to or change the
variables to spherical polar coordinate system, which we will learn later on.

So, this is a very handy tool actually to a calculate difficult integrals easily. Next we can
work out and another example. So, an another example could be let us say prove that.

(Refer Slide Time: 20:46)

. 2 −𝑦 2 𝜋 2
So, we will solve this example ∬𝐸 𝑒 −𝑥 𝑑𝑥𝑑𝑦 = 4 (1 − 𝑒 −𝑅 ) also in the similar fashion

the way we did before edges that we are practicing one more problem. So, where E is the
region defined by 𝑥 ≥ 0, 𝑦 ≥ 0 and 𝑥 2 + 𝑦 2 ≤ 𝑅 2. So, here there is something new
actually in this problem.

213
So, first of all let us draw our region. So, if we draw our region, then we have a circle. So,
let us draw this circle and that is our radius R; however, we are also given two separate
conditions which is x is positive and y is positive; that means x is positive. So, we are not
interested in this half and then y is positive. So, we are also not interested.So, then our
region is bounded by this circle and then these two lines x equals to 0 and y equals to 0.
So, this is our region E right.

So, in this region of course, we have r is running from this small r is running from 0 to
𝜋
capital R up to the radius, but then 𝜃 is running from 0 to only because we are not
2

completing the whole circle we are just sticking to this half only. So, in that case we have
𝜋
𝜃 running between 0 to 2 . So, let us provide this transformation. So, substitute or put 𝑥 =

𝑟 cos 𝜃 and 𝑦 = 𝑟 sin 𝜃, where r is running between 0 to capital R and 𝜃 is running between
𝜋
0 to 2 right.

. 2 −𝑦 2
So, now we write this integral here, 𝐼 = ∬𝐸 𝑒 −𝑥 𝑑𝑥𝑑𝑦. So, when we transform it to the
. 2 −𝑦 2
region 𝐸1 , we will write this region in 𝐸1 in a minute we can write∬𝐸 𝑒 −𝑥 𝑑𝑥𝑑𝑦 =
. 2 𝑐𝑜𝑠2 𝜃+𝑟 2 𝑠𝑖𝑛2 𝜃)
∬𝐸 𝑒 −(𝑟 |𝐽|𝑑𝑟𝑑𝜃 . So, we can write instead of this Jacobian we can simply
1
. 2 𝑐𝑜𝑠2 𝜃+𝑟 2 𝑠𝑖𝑛2 𝜃)
write r i.e. ∬𝐸 𝑒 −(𝑟 𝑟𝑑𝑟𝑑𝜃 .
1

So, now here we will replace this region 𝐸1 with the limits for r and 𝜃. So,
𝜋
𝑅 2
∫𝑟=0 ∫𝜃=0
2 𝑒 −𝑟 𝑟𝑑𝑟𝑑𝜃 right.

So, now I can substitute instead of 𝑟 2 , I can substitute t again and like what we did in the
in the previous example we can also do the same thing here and ultimately we will end up
with.

214
(Refer Slide Time: 24:55)

𝜋
𝑅 2
So, we will first write ∫𝜃=0
2 𝑑𝜃 ∫0 𝑟𝑒 −𝑟 𝑑𝑟 and then we substitute for 𝑟 2 = 𝑡 and then we
transform this whole thing into function of 𝑒 −𝑡 and then this will ultimately give us
𝜋
1 2
− 2 ∫02 [𝑒 −𝑟 ]𝑅0 𝑑𝜃 yeah.

𝜋
1 2 𝜋1 2
So, this will be − 2 ∫02 (𝑒 −𝑅 − 1)𝑑𝜃 . So, this will be (𝑒 −𝑅 − 1). So, ultimately we
22
𝜋 2
will have 4 (𝑒 −𝑅 − 1)and I believe this is what we needed to prove yes.

So, see initially it was a little bit how to say complicated in a way to evaluate this integral,
but just using some transformation which was how to say polar coordinate system which
transferred this Cartesian coordinate system into a polar coordinate system we were able
2 −𝑦 2
to reduce this whole 𝑒 −𝑥 into a fairly simple method of substitution integral and by
doing that metal substitution our required result will be this.

And of course, using these transformation of course, makes life quite easy and there will
be a lot of situations where we come across a very complicated integral double integral,
but just using some simpler transformation we can be able to reduce that complicated
integral into a fairly simple one and this is one such example. We will restrict ourselves to
practicing examples on this topic up to here, in the next class we will start with change of
order of integration. So, today we saw a change of variables using Jacobian transformation,

215
in the next class we will see change of order of integration and triple integral and that we
conclude this section as well. So, I look forward to your next class.

Thank you.

216
 Integral and Vector Calculus
Dr. Hari Shankar Mahato
Department of Mathematics
Indian Institutes of Technology, Kharagpur

Lecture – 21
Change of order of Integration

Hello students. So, up until last class we looked into the concepts of changing the variables
in a double integral; we also learnt about Jacobian transformation.

(Refer Slide Time: 00:33)

So, we actually looked into the fact that if we have an integral of let us say integral
.
∬𝐸 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦, where the integration is uncertain region E. Then by using some sort of
transformation, we can be able to transform the surface integral to a region E 1 and here
.
we will have ∬𝐸 𝑓(𝑢, 𝑣)𝑑𝑢𝑑𝑣 . So, we can change the variables. So the, of the integrand
1

and by doing this change of variables, it will actually make our life a little bit easier
because the transformed integral. There is a strong possibility can be a simpler integral and
then, it will be easy to compute this integral than computing this difficult one and of course,
we have to find out the range for this integration.

So, it will change accordingly based on the transformation and then we just perform, we
just calculate the Jacobian and then we calculate this integral. And in most of the cases,
this actually comes out to be very simpler integral to compute and the answer would

217
exactly be same. Now, besides change of variable of integrals sometimes it is also wise to
change the order of integration. So, suppose if we have range of let us say
𝑏 𝜓(𝑥)
∫𝑥=𝑎 ∫𝑦=𝜑(𝑥) 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦.

So, it can be possible that this integrand is so complicated or how to say not so common
that we cannot integrate it properly. So, it is not falling into one of the traditional how to
say integral calculation. So, we cannot simply do the integration here. One such example
2 2 2
could be let us say we have something like this; ∫𝑦=0 ∫𝑥=𝑦 𝑒 𝑥 𝑑𝑥 𝑑𝑦.

So, now this integral let us say for example, here the integrand is not so common because
in order to integrate this even with respect to x or with respect to with respect to y, I mean
with respect to y it will be easy; but with respect to x we do not have any anti derivatives
here or we cannot express e to the power x square in some method of substitution way. So,
calculating this integral will not be straightforward and then, the only option we will have
is to transform this integral into a simpler integral by doing the change of order.

So, if by doing the change of order, if we can be able to obtain a rather simpler integral;
then, that can be evaluated and of course, the value of those two integral would be same.
So, this is one such scenario. Now, let us work out an example where we can see that by
changing the order of integration, we will still obtain the same answer.

(Refer Slide Time: 04:22)

218
So, let us start with the following example. It says find the area between the region 𝑥 2 =
𝑦 and 𝑦 = 2 − 𝑥. So, we have to calculate the area between these two curves. So, as I
always say it is always suggestible to draw the area first. So, let us draw it. So, this is our
x; this is our y. So, we have 𝑥 2 = 𝑦 . So, it should be something like this a parabola passing
through origin. So, it must touch the origin as well alright and then, we have 𝑥 +𝑦 = 2 So,
let us draw this line as well. So, it must look something like this.

So, this is my line 𝑥 +𝑦 = 2 and this is my parabola 𝑥 2 = 𝑦 . So, this is the area bounded
between them. Let us call this as point O. Then, this is A; this is my B; this is C; this is D.
Now, since the parabola and the straight line, they are intersecting or the area is bounded
between them. So, they must meet at some point.

So, in order to calculate that we have here the parabola 𝑥 2 = 𝑦 and 𝑦 = 2 − 𝑥 will meet
at the following points. So, how do we calculate? We just substitute 𝑦 = 𝑥 2 in this
equation. So, it will be 2 − 𝑥 = 𝑦 = 𝑥 2 and therefore, this will become 𝑥 2 + 𝑥 − 2. So,
the roots of this equation will be x equals to we can factor it and then, it will be 𝑥 = 1 and
another root will be, I can substitute x equals to 1.

So, one root is 1 and another root will be; if I substitute 2, then this will be 4. So, another
root will be -2 alright. So, I have -2 here and I have 1 let us say 1 here. So, they are meeting
at the point x coordinates are -2 and 1 and when x is -2. Then, y is 4 and when x is 1; then
y is 1. So that means, they are intersecting at the point (1, 1).

So, they meet at (1,1) alright and (-2,4). So, these are the two points they are meeting.
Now, let us calculate the area in this way. So, we first integrate with respect to y, and then,
with respect to x. So, that is how we are maintaining the order. So, if we integrate with
respect to x later. So, then in that case x will vary from -2. So, we can write simply -2 to 1
and y we can write so, y the area bounded between these two curves; so for these two
curves x is varying from -2 to 1 alright.

So, that is the range for x and for y, y is varying between these two curves; so that means,
it is above this parabola but below the straight line. So, it is above the parabola, below the
straight line. So, above the parabola; that means, we have 𝑦 = 𝑥 2 and below that straight
1 2−𝑥
line is 2-x ∫𝑥=−2 ∫𝑥 2 𝑑𝐴, A denotes the area. So, now, we have x running from -2 to 1

219
and 𝑥 2 running from 2 − 𝑥 and we are integrating with respect to y first and then, we are
integrating with respect to x. So, that is what we are doing.

(Refer Slide Time: 09:48)

1 2−𝑥
So, if we integrate with respect to y first, then we will end up with integral ∫−2 ∫𝑥 2 𝑑𝑦𝑑𝑥.
So, integral with respect to dy. So, y and then ranges would be 𝑥 2 to 2 − 𝑥 dy and then we
1
substitute the value; so, it will be ∫−2(2 − 𝑥 − 𝑥 2 )𝑑𝑥. and after integrating and substituting
9
the value, we will be able to obtain 2; so, ultimately 4.5. So, that is the area bounded

between these two curves and not only that we integrate it with respect to y first and then,
we integrate it with respect to x. So, that is the order we followed actually.

So, alternatively now let us change the order. Alternatively, let us change the order of
integration; that means, that is we integrate with respect to x first and then with respect to
y. So, what will happen? So, what will happen is the following. So, we still have let us say
integral over the region E; I am calling this region. So, I am calling this region as E. Now,
if we want to integrate with respect to x first and then with respect to y, we have to see
how x is varying. So, first of all they intersect at the point x equals to 1, y equals to 1.

So, let us draw a line parallel to x axis. So, this is my straight line y running from y =1.
So, here we can see that y is varying from 0 to 1 right. So, y is varying from 0 to 1 and x
will vary from this corner to this corner. So that means, −√𝑦 to √𝑦. So, y is from 0 to 1;

220
x is from −√𝑦 to −√𝑦. So, let us write it. So, we separate the integral into these two sub
regions.

So, this is my first sub region; let us mark this way and this is my second sub region, this
is the second sub region and this is the first sub region alright. So, we will write here d A.
1 √ 𝑦
So, we will write it as integral ∫𝑦=0 ∫𝑥=− 𝑑𝑥𝑑𝑦. Now, we are in this region. So, in this
√𝑦

region y is varying from 1 to 4 right.

So, y is varying from 1 to 4 and x will vary from y equals to x will vary from y is equals
to−√𝑦 to this line right because that is where the x is bounded. So, we will now write, y

is varying from 1 to 4 and x is varying from −√𝑦 to 2 − 𝑦 dx dy and then, we calculate.

(Refer Slide Time: 13:53)

1 √ 𝑦 4 2−𝑦
So, if we calculate, then we have basically integral ∫𝑦=0 ∫𝑥=− 𝑑𝑥𝑑𝑦 + ∫𝑦=1 ∫𝑥=− 𝑑𝑥𝑑𝑦
√𝑦 √𝑦

So, if we substitute, then this will be y running from 0 to 1 this will be √𝑦 .

1 4
So, basically 2 ∫𝑦=0 √𝑦𝑑𝑦 + ∫𝑦=1(2 − 𝑦 + √𝑦)𝑑𝑦. So, if we integrate; then, ultimately
9
here also we will obtain 2; so, 4.5. So, we saw that just by changing the order of integration,

we do not get different answer because the area where we are actually doing the integration
is fixed. It is just that we are changing the order of integration; that means, instead of

221
integrating with respect to x first, we are integrating with respect to y. But that will not
alter the value of the overall integral.

Because then in that case we are getting a different area, which will not be the the actual
area bounded by the curve that was given to us originally. So, even if you do the change
of variables or change of order of integration, the area of the region bounded by these two
curves or the integral which you are evaluating has to be same and it is same all the time.
Now we see that sometimes we come across with integrals which are really difficult to
evaluate and then, just using this tool that is changing the order of integration we might
end up getting a little bit in easier integral to evaluate. For example, in this case we had to
evaluate the area bounded between these two curves.

So, you see if we are integrating with respect to y first and then, integrating with respect
to x then we are calculating a very simple integral and that is the answer. But if we change
the order of integration, then in that case we are basically calculating two sub integrals.
So, this one and this one and we also have to do some how to say some effort to basically
calculate these two domains. So, sometimes it is useful sometimes it is not, it is just that it
is from the intuition you have to understand whether the integral can be integrable or not
and then, we use the tool. We will work out to one or two more examples just to make an
idea clear. So, let us look into an another example.

(Refer Slide Time: 17:05)

222
 1 1−𝑥
So, example 2, calculate or evaluate; ∫𝑥=0 ∫𝑦=0 𝑐𝑜𝑠(1 − 𝑦)2 𝑑𝑦𝑑𝑥. So, now, solution; first

of all, let us draw this region; the y=0 x=0. So, here our x is running from 0 to 1 and y is
running from 0 to 1-x. So, it is basically a straight line 𝑥 + 𝑦 = 1. So, let us draw that line.
So, this is that point of intersection (1,1) sorry as (1,0) and this is (0,1) all right.

So, that is basically our given region of integration. So, I am calling it as E. This is my
point A and this is my point B. So, here just looking at this integral, let us write it as I. It
is very straight forward to see that it is not possible to integrate this because even if you
want to substitute 1 − 𝑦 = 𝑧 or something, then in that case this 𝑐𝑜𝑠(1 − 𝑦)2 𝑑𝑦, it cannot
be reduced to a very simpler integral. So, this is quite I have to say a complicated form.
Although, it looks simple; but it is not very easy to integrate this. So, then in that case the
only way out we have is to change the order of this integration and see whether we can do
something about this integral here.

So, first of all if we change the order of integration; so what will happen? So, changing the
order of integration; so first of all, we write the area the region of integration; order of
integration y, the region of integration lies in the strip between x=0, x =1 and is determined
by y values starting at 0 and increasing to the line 𝑦 = 1 − 𝑥. So, so that is the area of
integration or region of integration. Now, we see that if we instead of changing instead of
integrating with respect to y, if we integrate with respect to x first; then, in that case the
same region, our x will vary from then 0. So, x will start varying from 0 to 1 − 𝑦 and y
will vary from again 0 to 1.

So, it is not really changing the limits here. So, if we change the order. So, changing the
1 1−𝑦
order of integration our I will look ∫𝑦=0 ∫𝑥=0 𝑐𝑜𝑠(1 − 𝑦)2 𝑑𝑥𝑑𝑦 because the integrand

would remain same. It is just that we have changed the order. So, if we change the order,
then in that case since here we do not have two sub regions like parabola and straight line
and all that. Here, we just have this straight line and the usual coordinate x axis.

So, it is very easy to calculate the range for x and range for y. So, now, we have
1 1−𝑦
∫𝑦=0 ∫𝑥=0 𝑐𝑜𝑠(1 − 𝑦)2 𝑑𝑥𝑑𝑦 . So, when we integrate; now when we integrate, then it will

be integral 0 to 1.

223
(Refer Slide Time: 22:26)

So, y running from 0 to 1 𝑐𝑜𝑠(1 − 𝑦)2 and then here we will have integral x running from
1
0 to 1-y dx and then, dy. So, this will be ∫𝑦=0(1 − 𝑦)𝑐𝑜𝑠(1 − 𝑦)2 𝑑𝑦𝑑𝑥 . Now, we can

substitute this (1 − 𝑦)2 = 𝑧. So, now, we can substitute (1 − 𝑦)2 = 𝑧. So, then this will
be 2(1 − 𝑦)𝑑𝑦 = −𝑑𝑧 and if I substitute the value.

1 0
So, when y is 0. z is 1 and when y is 1, z is 0 and here I have − 2 ∫1 𝑐𝑜𝑠𝑧𝑑𝑧 =
1 1 1 1 𝑠𝑖𝑛1
∫ 𝑐𝑜𝑠𝑧𝑑𝑧 = 2 [𝑠𝑖𝑛𝑧]10 = 2 [𝑠𝑖𝑛1 − 𝑠𝑖𝑛0] =
2 0 2

𝑠𝑖𝑛1
So, is the required answer.
2

So, just in this example also we saw that. So, this is basically how to say calculation error,
I know I am pretty sure you can be able to fix that. So, here in this example also initially
if we look at the problem, it is very clear that we cannot be able to integrate this function
by with our traditional method.

Because there is no dz; there is no how to say that anti derivative form here. So, we cannot
substitute (1 − 𝑦)2 and get something and get something compensated from here or
something. So, then in that case, we have to the only option have is to change the order of
the integration. So, if we are changing the order of integration then, we have to find the
range for the variable x and y. So, that is what we did and another good thing about this
example was x was running from 0 to 1 minus y as y was running from 0 to 1 − 𝑥. So, it

224
was very fairly easy to calculate the range for the variable x and for the variable y it is just
interchanging.

So, now y is varying from 0 to 1. So, that is what we have written here alright and if we
have that; so then, in that case we integrate with respect to x first and then we have that z
how to say method of substitution formula working. So, we substitute (1 − 𝑦)2 = 𝑧 and
then we just do some simple calculation. So, that is what your answer is. So, this is one
such example where how to say changing the order of integration made our life a lot easier
actually. Now, it may be possible that sometimes changing the order of integration may
not be that useful.

(Refer Slide Time: 26:52)

2 2 2
So, what I mean by that is let us consider an another example 𝐼 = ∫0 ∫𝑥=𝑦 𝑒 𝑥 𝑑𝑥𝑑𝑦. So,

here we have our area of integration as 0 to so x is 2 and 𝑦 = 𝑥. So, this is this is now 𝑦 =
𝑥 and this is the area of integration right so E. So, this is my area of integration and. So,
2 𝑥 2
after changing the area of integration, we can write 𝐼 = ∫𝑥=0 ∫𝑦=0 𝑒 𝑥 𝑑𝑦 𝑑𝑥

So, now, we can be able to integrate this integral with respect to y first. So, if we want to
2 2
integrate then this will be integral ∫𝑥=0 𝑥𝑒 𝑥 𝑑𝑥. Now, we can substitute this and then we
get the value out of it.

225
However, if we had started with let us say; instead of this, we can have a problem where
2 𝑦 2
we can start with let us say this example if we had integral 𝐼 = ∫𝑦=0 ∫𝑥=0 𝑒 𝑥 𝑑𝑦 𝑑𝑥, where

is that?

So, if we have started with y running from 0 to 2 and x running from let us change this x
running from 0 to y. Then, in that case if we change the order of this integral, it would
2 2 2
converge to integral ∫𝑥=0 ∫𝑦=𝑥 𝑒 𝑥 𝑑𝑥 𝑑𝑦 and then, this will reduce to integral
2 2 2
∫0 (2𝑒 𝑥 − 𝑥𝑒 𝑥 )𝑑𝑥 which is a rather complicated integral to evaluate than evaluating
this integral.

So, just starting with a different form of integral, we can be able to see that always
changing the order of integration may not help. So, sometimes it is useful sometimes it is
not, but most of the time it is useful. So, as in this example, it becomes a rather complicated
to integrate especially the first term. So, sometimes we do not have to change the integral,
but most of the time we do.

So, we will conclude this section here today and in the next class we will start with area
and how do we calculate the area under a curve and several other things related to area
volume and surface integral and we will work out more examples based on this topic in
our assignment section, assignment sheets.

So, thank you for attention today and I will see you in the next class.

226
 Integral and Vector Calculus
Dr. Hari Shankar Mahato
Department of Mathematics
Indian Institutes of Technology, Kharagpur

Lecture – 22
Triple Integral

Hello students. So, up until last class we looked into change of variables in an integral and
also change of order of integration, where we can reduce a very complicated integral to a
rather simpler form by using either change of variables or change of order of integration.
Sometimes it is useful, sometimes it is not. So, most of the time we have to judge from our
intuition, that whether changing the order of integration or changing the order of variable
would be helpful or not.

Sometimes it also looks like the integral is quite complicated and there is no way we can
be able to integrate using the traditional method of integration. Then, in that case changing
the order of integration or changing the variable of the integration would be the only way
out there.

So, we also worked out few examples in the previous class and today, we will start with
something called Triple Integral. So, we know that in a; so, this is according to your
syllabus. So, if we look into your syllabus then today we are going to cover the triple
integral.

(Refer Slide Time: 01:37)

227
So, today we will be covering basically the triple integral and so, let us go back to the
topic. So, we will start with Triple integral alright. So, now, like in case of function of one
variable let us say f(x) which is defined on an interval [a, b]; then, in that case we can be
𝑏
able to talk about integral of this type 𝐼 = ∫𝑎 𝑓(𝑥)𝑑𝑥 or if we want to integrate along a
curve, then we can also do that here. Now, if we have a function let us say of two variables.

So, this was the first type of integral we saw. Now we have the second type of integral
which is basically if we have a function of two variables say x and y on a rectangle let us
say [𝑎, 𝑏] × [𝑐, 𝑑]; this can also be a region. So, it can happen that instead of rectangle we
have a region. However, the concept would work similarly. So, the integral can be defined
𝑏 𝑑
as ∫𝑎 ∫𝑐 𝑓(𝑥, 𝑦)𝑑𝑥 𝑑𝑦. Now, how do we do the integration that is something we have
already seen.

Now, suppose we know that we also have three-dimensional geometries and three-
dimensional figures and in such cases we can also talk about triple integrals. So, triple
integral deals with the surfaces which are in 3 D. So, here in this case we have the range
or the domain of integration are basically two-dimensional geometry. So, instead of this
rectangle we can also have a region which is in 2 D setting, but instead of 2 D setting, we
can also have a 3 D setting.

So, suppose if we have a function f x y z defined on a rectangular parallelepiped

[𝑎, 𝑏] × [𝑐, 𝑑] × [𝑒, 𝑓]. So that means, now we have a rectangular parallelepiped. Here, in
this case also we can talk about a triple integral. So, we write it in this fashion;
𝑏 𝑑 𝑓
∫𝑎 ∫𝑐 ∫𝑒 𝑓(𝑥, 𝑦, 𝑧)𝑑𝑥𝑑𝑦𝑑𝑧. So, triple integral comes from the fact that we have used 3
integral symbol. So, hence the triple integral; sometimes people also prefer to call it as
volume integral, this is also fine and this dx dy dz is basically the volume element.

So, this dx dy and dz are the volume element and here of course, how we evaluate the
triple integral we will see in a minute. Now, we know that in case of function of one
variable, we could be able to define that upper integration sum, lower integration sum;
from there we defined the upper integral and a lower integral accordingly and if the upper
integral and the lower integral are same. Then, we can say that the function is integrable
or Riemann integrable. In case of function of two variables, we saw that similar concept
works. So, in case of function of two variable we can also be able to calculate the upper

228
integral sum and lower integral sum and from there we can calculate upper integral and
lower integral and if they are same; then the function is said to be integrable. On that
rectangular region R are now in case of function of 3 variables, the concept still holds and
it goes in the following way.

(Refer Slide Time: 05:40)

So, the definition we will start basically with the definition. So, integration of a bounded
function on a rectangular, let us say parallelepiped. So, it is given by 𝑅 =
[𝑎, 𝑏] × [𝑐, 𝑑] × [𝑒, 𝑓].. By the way this f here and the function f, they are two different
things. So, this f here is basically a real number and that f denotes the function of x y z
alright and suppose we consider.

So, let us consider the 3 partition let us consider the 3 partitions; say 𝑃0 , 𝑃1 and 𝑃2 such
that we have 𝑃0 = {𝑥0 = 𝑎 < 𝑥1 < 𝑥2 . . . < 𝑥𝑛−1 < 𝑥𝑛 = 𝑏} . Similarly, we can write
𝑃1 = {𝑦0 = 𝑐 < 𝑦1 < 𝑦2 . . . < 𝑦𝑛−1 < 𝑦𝑛 = 𝑑}. And similarly, 𝑃2 = {𝑧0 = 𝑒 < 𝑧1 <
𝑧2 . . . < 𝑧𝑛−1 < 𝑧𝑛 = 𝑓}.

Now, these 3 partition; give rise to a new partition say P of the rectangular parallelepiped
which divides rectangular parallelepiped.

which divides R into. So, by the way we can have this one as 𝑥𝑛 ; otherwise it will be a
square. So, we take it as a rectangular parallelepiped. So, 𝑥𝑛 , 𝑦𝑚 & 𝑧𝑝 into m n p sub

229
rectangular parallelepiped. So, it will divide the whole rectangular parallelepiped into m n
times p sub rectangle parallelepiped.

(Refer Slide Time: 09:30)

And where sub rectangular parallelepiped; [𝑥𝑟−1 , 𝑥𝑟 ] × [𝑦𝑠−1 , 𝑦𝑠 ] × [𝑧𝑡−1 , 𝑧𝑡 ] where 𝑟 =

1,2,3,. . . , 𝑛; 𝑠 = 1,2,3,. . . , 𝑚; 𝑡 = 1,2,3, . . . , 𝑝.

So, this is the number of how to say sub rectangular parallelepipeds by which for this
interval this partition P divides actually. Now we write let 𝑤𝑟𝑠𝑡 = (𝑥𝑟 − 𝑥𝑟−1 )(𝑦𝑠 −
𝑦𝑠−1 )(𝑧𝑡 − 𝑧𝑡−1 ) alright which is the volume of this rectangle parallelepiped this one. So,
that is the volume of this rectangular parallelepiped this here.

So, it is basically length into breadth into height. and now we write let 𝑀𝑟𝑠𝑡 and 𝑚𝑟𝑠𝑡 be
the upper and lower bound of the function f on this sub rectangle let us say on the sub
rectangular parallelepiped; sometimes this is spelling might be wrong.

So, do not worry about it you can parallelepiped. So, you can correct that; so, this spelling
is not an issue alright. So, now, that we have the upper bound we have defined, we have
denoted the upper bound and lower bound, I can be able to write my upper integral sum
which is 𝑈(𝑃, 𝑓),which is basically sum of all the upper bounds or I can write instead of
writing this notation. So, I was going to write the whole thing together, but let us write it
𝑝
separately; ∑𝑛𝑟=1 ∑𝑚
𝑠=1 ∑𝑡=1 𝑀𝑟𝑠𝑡 𝑤𝑟𝑠𝑡 . So, this is my upper integral sum.

230
 𝑝
I can write 𝐿(𝑃, 𝑓) = ∑𝑛𝑟=1 ∑𝑚
𝑠=1 ∑𝑡=1 𝑚𝑟𝑠𝑡 𝑤𝑟𝑠𝑡 is basically lower integral sum alright. So,

this is how we define the upper integral sum and lower integral sum. Now, as for the
function of 1 variable or function of 2 variables, the infimum of this set and so, infimum
of all such upper integral sum for all the partitions for I mean that is basically called as the
upper integral sum and lower integral sum.

(Refer Slide Time: 13:30)

So, basically the upper integral sum on this rectangle let us say R rectangular
−
parallelepiped; let us say ∭𝑅 𝑓(𝑥, 𝑦, 𝑧)𝑑𝑥𝑑𝑦𝑑𝑧 is basically the infimum of set of all sums.

So, 𝑈(𝑃, 𝑓) infimum all such 𝑈(𝑃, 𝑓) where for all partitions P is called as the upper
integral sum and similarly the lower integral sum let us write
.
∭𝑅 𝑓(𝑥, 𝑦, 𝑧)𝑑𝑥𝑑𝑦𝑑𝑧 equals to 𝑠𝑢𝑝𝐿(𝑃, 𝑓). So, this is basically 𝐿(𝑃, 𝑓) and this is called
−

as lower integral. Now from our classical theory, we know that the upper integral if the
upper integral and the lower integral let us write here.

−
So, this R means rectangular parallelepiped alright. So,∭𝑅 𝑓(𝑥, 𝑦, 𝑧)𝑑𝑥𝑑𝑦𝑑𝑧 =
.
∭𝑅 𝑓(𝑥, 𝑦, 𝑧)𝑑𝑥𝑑𝑦𝑑𝑧 , if they are equal then this actually implies that the function f is
−

triple integrable over the rectangular parallelepiped and the common value is basically
denoted by rectangular integral over the rectangular parallelepiped
.
∭𝑅 𝑓(𝑥, 𝑦, 𝑧)𝑑𝑥𝑑𝑦𝑑𝑧 is called the triple integral. It is called the triple integral of f on the
rectangular parallelepiped R.

231
So, this is how we basically define in general the triple integral of a function f on a
rectangular parallelepiped R like we did for the function on a rectangular domain in
𝑅 2 actually. Similarly, instead of having a rectangular parallelepiped, we can have a
region. So, we can have double triple integral.

(Refer Slide Time: 16:09)

So, calculation of triple integral on a rectangular on a region actually. So, we can have
something like calculation of triple integral on a bounded domain. So, like bounded region,
here we have bounded domain E. So, the definition goes like this. If f is a continuous
function; on a domain E bounded by the surfaces 𝑧 = 𝜑(𝑥, 𝑦) and 𝑧 = 𝜓(𝑥, 𝑦), then 𝑦 =
𝑔(𝑥), 𝑦 = ℎ(𝑥)& 𝑥 = 𝑎 & 𝑥 = 𝑏. Then,

. 𝑏 ℎ(𝑥) 𝜑(𝑥,𝑦)
∬𝐸 𝑓(𝑥, 𝑦, 𝑧)𝑑𝑥𝑑𝑦𝑑𝑧 = ∫𝑥=𝑎 ∫𝑦=𝑔(𝑥) ∫𝑧=𝜓(𝑥,𝑦) 𝑓(𝑥, 𝑦, 𝑧)𝑑𝑥𝑑𝑦𝑑𝑧.

So, that is how basically we will calculate the triple integral over a bounded region.

So, first of all we need to identify by what curve the region E is formed actually. So, that
can be calculated or sometimes it is also given. So, we can be able to find out that and the
bounded region and then, the triple integral over that bounded region can be given in this
fashion. So, we can see that from this form which is that first we are integrating with
respect to z. So, we are substituting for z and then, we will get a function of x and y only.
So, then we will integrate with respect to y to get the function of x only and then, we
integrate with respect to x. So, that is how that is the kind of flow chart we are following

232
here, alright. So, these are the two basic definitions in a way for us to calculate the triple
integral.

So, either the given domain can be a rectangular domain or it can be a region and if it is a
region, then we have to find out these curves actually. So, these curves here and then, we
this is how we calculate the double triple integral. So, now, that we have the definition; we
will try to work out few examples.

(Refer Slide Time: 19:35)

So, let us start with our first example. So, the first example is. So, evaluate or express let
us say; evaluate or express, better we use the word express. 𝐼=
.
∭𝑅 𝑓(𝑥, 𝑦, 𝑧)𝑑𝑥𝑑𝑦𝑑𝑧 over the sphere 𝑥 2 + 𝑦 2 + 𝑧 2 = 𝑎2 as a triple integral, or repeated
integral. So, how can we express this integral as a triple integral? So, first of all the given
region is a sphere.

So obviously, & a is a finite positive number; so obviously, it is a bounded domain. So, I

do not know if we can be able to draw the 3 D figures correctly, but let us try to do that
anyways. So, this is basically our sphere X Z and Y. So, this is the radius a and of course,
it is a bounded region. So, we have to first find out the limit for x. So, here x is varying
from -a to +a.

Now what about z? So, z is basically so, if we see the sphere and the z axis going through
the center of it, then in that case z is basically varying from −√𝑎2 − 𝑥 2 − 𝑦 2 to

233
+√𝑎2 − 𝑥 2 − 𝑦 2 . So, z is varying from this point; so, this point to this point. So that
means, this point to this point; that means, its slightly difficult to figure out the 3 D
geometry, but its varying from −√𝑎2 − 𝑥 2 − 𝑦 2 .

So, −√𝑎2 − 𝑥 2 − 𝑦 2 𝑡𝑜 + √𝑎2 − 𝑥 2 − 𝑦 2 and y is varying from square root of if we

take 𝑧 = 0, then into 2 dimensional geometry is it is basically a circle 𝑥 2 + 𝑦 2 = 𝑎2 ;
where, y is varying from −√𝑎2 − 𝑥 2 , to +√𝑎2 − 𝑥 2 basically. So, this is our required how
to say range for the variable x y and z and now, we can write that integral 𝐼 =
.
∭𝑅 𝑓(𝑥, 𝑦, 𝑧)𝑑𝑥𝑑𝑦𝑑𝑧 = integral over the region R which is basically so here we will have
dx dy dz.

𝑎 √𝑎2 −𝑥 2 √𝑎2 −𝑥 2 −𝑦2

And this one will be ∫𝑥=−𝑎 ∫𝑦=−√𝑎2 −𝑥2 ∫𝑧=−√𝑎2 −𝑥 2 −𝑦2 𝑓(𝑥, 𝑦, 𝑧)𝑑𝑥𝑑𝑦𝑑𝑧. So, this is the

required form for this integral in terms of how to say triple integral.

Well, now we have the limit of the integration. Of course, if we are given a function, then
this will give us the volume bounded by this region E for that function f. So, basically if
we have the function f, then we can be able to calculate this volume integral quite easily.
So, now, that we have seen the first example.

(Refer Slide Time: 24:48)

234
Let us go to our second example. So, here it says evaluate ∭(𝑥 + 𝑦 + 𝑧)2 𝑑𝑥𝑑𝑦𝑑𝑧 ;
where, the region is 𝑥 ≥ 0;𝑦 ≥ 0;𝑧 ≥ 0;𝑥 + 𝑦 + 𝑧 ≤ 1. So, this one; now here we have
basically a region, which is 𝑥 + 𝑦 + 𝑧 ≤ 1. So, we can draw it something like this.

Now, we have to consider the fact that, x is positive. So that means, all the points of the
region which are lying on the positive side of the x axis. Similarly, y is positive. So, we
have to consider all the points which is lying on the positive side of the y axis and then we
have to consider the points 𝑧 ≥ 0; that means, all the points which are lying on the positive
side of z axis. And then, 𝑥 + 𝑦 + 𝑧 ≤ 1 is the upper part of the plane that is how to say
forming a bounded region basically.

So, here z varies from 0 to 1 − 𝑥 − 𝑦, y varies from 0 to 1 − 𝑥 and x varies from 0 to 1.

So, we have got the limit of the integration and now, we can calculate the integral. So,
basically triple integral x running from 0 to 1, y running from let us write these integrals a
bit apart so that we can write these things neatly; y running from 0 to 1 − 𝑥 and z running
1 1−𝑥 1−𝑥−𝑦
from 0 to 1 − 𝑥 − 𝑦 ; 𝐼 = ∫𝑥=0 ∫𝑦=0 ∫𝑧=0 (𝑥 + 𝑦 + 𝑧 + 1)2 𝑑𝑥 𝑑𝑦 𝑑𝑧 and now, what we

will do is, now we will integrate this integrand here.

So, this will require some time to and perform this integral. So, we are running out of time
in this lecture. So, we will continue with a similar example in our next lecture, where we
will evaluate this integral and we can calculate this triple integral. So, I thank you for your
time today and I will see you in the next class.

Thank you.

235
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 23
Triple Integral (Contd.)

Hello students. So, in the last class, we started with Triple Integral, and we also looked
into the basic definition, and the concept of triple integral. We also started with one
example which turned out to be a little bit long. So, we how to say stopped it in the
midway. So, we will continue with the similar example, and we will try to finish it in this
lecture. And then we will also look into some more examples motivated from triple
integral.

(Refer Slide Time: 00:51)

.
So, we started with basically this example here. Evaluate 𝐼 = ∭𝐸(𝑥 + 𝑦 + 𝑧 +
1)2 𝑑𝑥𝑑𝑦𝑑𝑧 which is basically the volume element where the region is given by this here.
So, this one was basically our region of integration. So, from here we calculated the
range for z. So, obviously the z is varying from 0 to 1 − 𝑥 − 𝑦, because we start with all
the points which are lying on the positive side of the z-axis.

Similarly, y will vary from 0 to 1 − 𝑥, because we again start with the points which are
lying on the positive side of y-axis. And similarly,our x will vary from 0 to 1, because

236
we are starting for all the points which are lying on the positive side of x-axis. So, now
that we have the limit, we can substitute for these values on this region of integration it is
better to write E here, just to signify that we are talking about a region.

(Refer Slide Time: 02:13)

Now, we have substituted the limits as you can see here, and then we will integrate one
by one. So, first we will start with integrating with respect to variable z. So, let us start
with integrating with respect to variable z. So, this is basically
1 1−𝑥 1−𝑥−𝑦
∫0 𝑑𝑥 ∫0 𝑑𝑦 ∫0 (𝑥 + 𝑦 + 𝑧 + 1)2 𝑑𝑧 all right.

So, when we are integrating this integrand with respect to z that means, x and y are both
treated as constant here, so because there is no there is no x and y variable involving in
this element dz, so that means we treat x y, and of course 1 as a constant here, and then
this is basically (𝑎 + 𝑥)2 definitely integral. So, for such integrals we know how to do
1 1−𝑥 1
that how to do the integration, so this will reduce to ∫0 𝑑𝑥 ∫0 [ (𝑥 + +𝑧 +
3
1−𝑥−𝑦
1)3 ]0 𝑑𝑦 all right. So, this will be the integral of this integration of this integration
here.

1 1 1−𝑥
Now, 3 ∫0 𝑑𝑥 ∫0 [8 − (1 + 𝑥 + 𝑦)3 ]𝑑𝑦, and now we substitute the value. So, when z

is1 − 𝑥 − 𝑦, so then this will be basically two right, so 8 minus when z is 0, then this is
basically (1 + 𝑥 + 𝑦)3 dy. Now, we integrate with respect to y only. So, again when we

237
integrate with respect to y, then x will be treated as constant, and 1 and 8 are of course
treated as constant.

1 1 (1+𝑥+𝑦)4 1−𝑥
So, we integrate so let us integrate, then this will be 3 ∫0 [8𝑦 − ]0 𝑑𝑥 all right,
4
1 1 24
and then we substitute the value. So, this will be basically ∫ [8(1 − 𝑥) −
3 0 4
+
(1+𝑥)4
] 𝑑𝑥 all right.
4

(Refer Slide Time: 04:56)

1 1
So, we will basically obtain after simplification, 3 [4𝑥 − 4𝑥 2 + 20 (𝑥 + 1)5 ]1𝑥=0 integral

sorry the value of the function of the integral evaluated at 0 to 1. So, here basically we
calculate (1 + 𝑥)4 using the binomial theorem, so it will be 𝑥 4 + 4𝑥 3 +. and so on.

And then we integrate with respect to x, so this is a fairly easy how to say integral to
evaluate. So, I am pretty sure you can be able to do this integral, because it is a basically
an algebraic expression. And after you do the integration, you will obtain a result like
1 31
this. And then we substitute the value, and it will result in60 (32 − 1). So, basically60, so

that is the value of this integral evaluated between this region actually.

So, here first we had to find out the range for the variable x, y, and z. And afterwards we
just integrated with respect to z first, with respect to y second, and then with respect to x
third. So, this is the order which is preferred by a most of the people, so they initially try

238
with integrating with respect to z, and then y, and then x. So, we will work out few more
examples for you to get a custom with triple integral.

So, let us go to our second example all right. So, the second example is evaluate 𝐼 =
.
∬𝐸(𝑥 2 + 𝑦 2 + 𝑧 2 )𝑑𝑥𝑑𝑦𝑑𝑧 over the sphere 𝑥 2 + 𝑦 2 + 𝑧 2 ≤ 𝑎2 . So, we have to evaluate
this integral here, so that is the integrand basically, so this integral here over the region
E. And this region E is basically that sphere which is of course a bounded domain, when
a is finite.

And if we want to draw this sphere, so I am not a really good how to say really good at
drawing, so this is basically 0, y, z, and x, so that is my radius a, and this rectangular I
sorry this is this region E, where we have to perform the integral. So, now in this region
if we substitute, so let us substitute 𝑥 = 𝑟𝑠𝑖𝑛𝜃𝑐𝑜𝑠𝜑, 𝑦 = 𝑟𝑠𝑖𝑛𝜃𝑠𝑖𝑛𝜑 & 𝑧 = 𝑟𝑐𝑜𝑠𝜃. So, if
we take 𝑥 2 + 𝑦 2 + 𝑧 2 , then this will actually reduce to this . 𝑥 2 + 𝑦 2 + 𝑧 2 = 𝑎2

(Refer Slide Time: 09:05)

Now, to do the change of variables actually, we know that we need to calculate the
𝜕(𝑥,𝑦)
Jacobian. So, for the function of two variable, the Jacobian was given as |𝐽| = 𝜕(𝑢,𝑣). So,

further that was for the function of two variables.

Now, if you want to have the Jacobian for the function of three variable, then in that case
𝜕(𝑥,𝑦,𝑧)
we basically write|𝐽| = 𝜕(𝑢,𝑣,𝑤). So, this is how to say a standard way of converting a one

239
function of one variable to the function of second variable in the integral, and for in order
to do that we need to calculate the Jacobian which is given in this fashion. So, this is
𝜕𝑥 𝜕𝑥 𝜕𝑥
𝜕𝑢 𝜕𝑣 𝜕𝑤
basically|| 𝜕𝑦 𝜕𝑦 𝜕𝑦 |.
𝜕𝑢 𝜕𝑣 𝜕𝑤 |
𝜕𝑧 𝜕𝑧 𝜕𝑧
𝜕𝑢 𝜕𝑣 𝜕𝑤

Now, here u is basically r, v is basically 𝜃, and z is basically 𝜑. So, if we differentiate x,

y and z with respect to r, 𝜃 and 𝜑 , then basically we can calculate this whole
determinant, and it will result in 𝑟 2 𝑠𝑖𝑛𝜃. So, this task is very simple. And therefore, I
leave this one to the students to calculate. So, we will basically obtain 𝑟 2 𝑠𝑖𝑛𝜃.

.
And therefore, I can write my integral 𝐼 = ∬𝐸(𝑥 2 + 𝑦 2 + 𝑧 2 )𝑑𝑥𝑑𝑦𝑑𝑧 . So, if I transform
this integral into a polar one, then in that case this will be r running from 0 to a, because
that is the range for the variable r or the radius. And theta will run from 0 to 𝜋. And phi
will run from 0 to 2𝜋 all right.

And this will be 𝑥 2 + 𝑦 2 + 𝑧 2 . So, if I substitute for 𝑥 = 𝑟𝑠𝑖𝑛𝜃𝑐𝑜𝑠𝜑 , then it will be

𝑟 2 𝑠𝑖𝑛2 𝜃𝑐𝑜𝑠 2 𝜑 . And similarly, it for y it will be 𝑟 2 𝑠𝑖𝑛2 𝜃𝑠𝑖𝑛2 𝜑, and then z will be z
will be 𝑟 2 𝑐𝑜𝑠 2 𝜃. So, we substitute all those things, and then this will be reduced to
𝑎 𝜋 2𝜋 𝑎 𝜋 2𝜋
∫𝑟=0 ∫𝜃=0 ∫𝜑=0 𝑟 2 |𝐽|𝑑𝑟𝑑𝜃𝑑𝜑. And then this will be ∫𝑟=0 ∫𝜃=0 ∫𝜑=0 𝑟 4 𝑠𝑖𝑛𝜃𝑑𝑟𝑑𝜃𝑑𝜑.

(Refer Slide Time: 12:37)

240
 𝑎 𝜋 2𝜋
And then this whole thing will reduce to integral ∫𝑟=0 ∫𝜃=0 ∫𝜑=0 𝑟 4 𝑠𝑖𝑛𝜃𝑑𝑟𝑑𝜃𝑑𝜑 , so the

same thing here. And now we integrate with respect to 𝜑 first. So, we integrate with
𝑎 𝜋 2𝜋
respect to 𝜑, and then this will be ∫𝑟=0 ∫𝜃=0 𝑟 4 𝑠𝑖𝑛𝜃(∫𝜑 𝑑𝜑) 𝑑𝑟𝑑𝜃, so the value of 𝜑 will

be 2𝜋 basically. So, I am taking that 2𝜋 outside, and then here we will have
𝑎 𝜋
2𝜋 ∫𝑟=0 ∫𝜃=0 𝑟 2 𝑠𝑖𝑛𝜃𝑑𝑟𝑑𝜃 .

Next we will integrate with respect to 𝜃 first, and then in that case this will reduce to
𝑎
2𝜋 ∫𝑟=0 𝑟 4 [−𝑐𝑜𝑠𝜃]𝜋0 𝑑𝑟. So, here let us write r running from 0 to a, when we integrate
with respect to 𝜃 first, it will be so r square. So, sorry we will have 𝑟 4 right yes so 𝑟 4 .

And when we integrate with respect to 𝜃, and this will be a – 𝑐𝑜𝑠𝜃 will vary from 0 to 𝜋
dr and when 𝑐𝑜𝑠𝜃 when 𝑐𝑜𝑠𝜃 when 𝜃 is 𝜋, then 𝑐𝑜𝑠𝜋 is −1, then this one will be plus
1. And this one will be cos0. So, cos0 is 1, and then minus minus plus. So, this will again
𝑎 𝑟5
be 4𝜋 ∫0 𝑟 4 𝑑𝑟. And when we integrate this one, then this will be 4𝜋[ 5 ]𝑎𝑟=0 . So, this will
4
be 5 𝜋𝑎5 . So, this is the required answer for the integral for the triple integral which we

started with.

And we can see that so here we can see that we were given a bounded region first of all.
So, from the bounded region, it is also very convenient to use a pull of spherical polar
coordinate system. So, here we have a sphere basically. And for this is sphere, we can
use this spherical polar coordinate substitution, so that is the 𝑟𝑠𝑖𝑛𝜃𝑐𝑜𝑠𝜑, 𝑟𝑠𝑖𝑛𝜃𝑠𝑖𝑛𝜑and
𝑟𝑐𝑜𝑠𝜃, of course this point lies on the sphere.

Now, we substitute so next we calculate the Jacobian, and while we have the Jacobian
value. We can do the how to say change of variable or transformation, and by changing
the variable we have next new variable as dr d𝜃 d𝜑, and then here you will have a
Jacobian J, and just substitute the value of the Jacobian. And then integrate with respect
to 𝜑 first, then with respect to 𝜃, and then with respect to r, and that will be our required
answer.

And this is the area which is basically bounded by sorry this is the how to set the value of
the integrand given bounded by that region or that sphere 𝑥 2 + 𝑦 2 + 𝑧 2 ≤ 𝑎2 . So, this
was one such example of triple integral.

241
(Refer Slide Time: 16:27)

Similarly, we can also have example of type let us say example-3,evaluate

.
∭𝑉 𝑥 2 𝑑𝑥𝑑𝑦𝑑𝑧where V is the sphere𝑥 2 + 𝑦 2 + 𝑧 2 ≤ 1 is it 1 yes. So, now again here
we are given a sphere, which is basically again a bounded sphere. So, a bounded region
we have in our hand, and we need to evaluate this integral here.

.
So, the integral I is given by ∭𝑉 𝑥 2 𝑑𝑥𝑑𝑦𝑑𝑧. And obviously, V is a bounded region
where x, y, and z can be transformed into a spherical polar coordinate system. So, we can
have 𝑥 = 𝑟𝑠𝑖𝑛𝜃𝑐𝑜𝑠𝜑, 𝑦 = 𝑟𝑠𝑖𝑛𝜃𝑠𝑖𝑛𝜑 & 𝑧 = 𝑟𝑐𝑜𝑠𝜃 .

So, now we can have dx dy dz calculated, then on the right hand side we will have dr d𝜃
and d𝜑 calculated. And we substitute the value of x here, which is 𝑟 2 𝑠𝑖𝑛2 𝜃𝑐𝑜𝑠 2 𝜑. So,
ultimately this whole thing will reduce to r varying from 0 to 1, and 𝜃 is varying from 0
to 𝜋, and 𝜑 is varying from 0 to 2𝜋, then we have 𝑥 2 which is basically 𝑟 2 𝑠𝑖𝑛2 𝜃𝑐𝑜𝑠 2 𝜑.
And then dx dy dz, which can be transformed into a |J|dr d𝜃 and d𝜑 all right.

So, now we can calculate this Jacobian here. So, in order to calculate this Jacobian, we
will do the similar trick, what we did in the previous example, and this will be basically
𝑟 2 𝑠𝑖𝑛𝜃 was it the same thing yes. So, we will obtain 𝑟 2 𝑠𝑖𝑛𝜃 . So, substitute the value
of𝑟 2 𝑠𝑖𝑛𝜃 , and then this whole thing will reduce to a basically a trigonometrical
expression. and then we just integrate with respect to 𝜑 first, and then with respect to 𝜃,
and then with respect to r.

242
(Refer Slide Time: 19:52)

So, let us try to do that, so what we do? We first separate the function 𝜑. So, 𝐼 =
1 𝜋 2𝜋
∫𝑟=0 ∫𝜃=0 ∫𝜑=0 𝑟 2 𝑠𝑖𝑛2 𝜃𝑐𝑜𝑠 2 𝜑|𝐽|𝑑𝑟 𝑑𝜃 𝑑𝜑. And if we integrate with respect to r first,
1 2𝜋 𝜋
then this will be ∫ 𝑐𝑜𝑠 2 𝜑𝑑𝜑 ∫𝜃=0 𝑠𝑖𝑛3 𝜃[𝑟 5 ]10 𝑑𝜃 and this is basically dr so this is
5 𝜑=0

basically d𝜃d𝜑.

So, now we do substitution for r equals to 0; and r equals to 1, so this term will remain as
1 2𝜋 𝜋
it is and then here we will have 5 ∫0 𝑐𝑜𝑠 2 𝜑𝑑𝜑 ∫0 𝑠𝑖𝑛3 𝜃𝑑𝜃 . Here we can either use the

induction formula or we can use the trignometrical formula, where this will reduce to
𝑠𝑖𝑛3𝜃 − 3𝑠𝑖𝑛𝜃 and from there we can do some calculation, and we can be able to find
out the value of this integral with respect to 𝜃.

And then we integrate then we are just a factor two here, so that we will get 2𝑐𝑜𝑠 2 𝜑, and
then we can write 2𝑐𝑜𝑠 2 𝜑 is 𝑐𝑜𝑠2𝜑 + 1, and from there we can be able to calculate this
integral as well or we can use the induction formula. So, on both of them we can either
use the induction formula or use our traditional 𝑐𝑜𝑠2𝜑, 𝑠𝑖𝑛3𝜑 results.

81𝜋2
And then after doing the simplification, we can be able to obtain , so this is
5223
4𝜋
basically 15
. And this is the required answer of our problem where is that this here. So,

here the given integrand was 𝑥 2 𝑑𝑥𝑑𝑦𝑑𝑧, so basically doing this the simple method of
substitution.

243
We can be able to reduce the whole integral into a rather simpler one, and then we just
use some trigonometrical formula like 𝑐𝑜𝑠2𝜃 or 𝑠𝑖𝑛2𝜃or 𝑐𝑜𝑠2𝜑 𝑐𝑜𝑠3𝜃& all, and that
will give us the required value of this integral, because now we are integrating basically
two different integrals, because of their integrand.

So, this one we can do that at home, and this one also we can do that basically by
ourselves, and just what to say writing 2𝑐𝑜𝑠 2 𝜑 as that formula expression of 𝑐𝑜𝑠2𝜑 + 1,
and we can write 𝑠𝑖𝑛3 𝜃 as 𝑠𝑖𝑛3𝜃 formula we can be able to calculate these two integrals
separately. And then substitute them, and that will give you the answer So, you will
reduce them to a trigonometrical form, it is not that difficult all right. And this was the
second example or third example basically which we covered.

(Refer Slide Time: 23:46)

We can work out one more example if we have some time, so let us consider an another
example. So, an another example could be without having to reduce it into a
trigonometrical form could be so we can start with a let us consider this problem here.
.
So, example-3 evaluate 𝐼 = ∭𝑅 √1 − 𝑥 2 − 𝑦 2 − 𝑧 2 , where R is the region interior to the
sphere 𝑥 2 + 𝑦 2 + 𝑧 2 = 1.

So, solution here also what we will do basically is R is the region, which is the interior of
the sphere. So, again the region R is bounded by the sphere S here let us say. So, we can
draw the so this is our z-axis, this is our y-axis, this is our x-axis, and that is my sphere,
this is the radius 1, and this whole is the region R so this is the region R.

244
And if we substitute again 𝑥 = 𝑟𝑠𝑖𝑛𝜃𝑐𝑜𝑠𝜑, 𝑦 = 𝑟𝑠𝑖𝑛𝜃𝑠𝑖𝑛𝜑 & 𝑧 = 𝑟𝑐𝑜𝑠𝜃 ., then this will
.
basically this integral 𝐼 = ∭𝑅 √1 − 𝑥 2 − 𝑦 2 − 𝑧 2 .

So, when we are transforming it into how to say a spherical polar coordinate system or r
will run from 0 to 1, 𝜃 will run from 0 to 𝜋, and 𝜑 will run from 0 to 2𝜋, then we
substitute for x y axis x y and z and here basically the value of the integral. So, when we
substitute 𝑥 2 = 𝑟 2 𝑠𝑖𝑛2 𝜃𝑐𝑜𝑠 2 𝜑, and y equals to so the value of this integrand basically
will be 0.

So, this here, so this problem will not be that much how to say interesting for us. So,
what we can do is we can impose conditions like 𝑥 ≥ 0, 𝑦 ≥ 0, 𝑧 ≥ 0 , then in that case
we are actually in the first quadrant. And then for the first quadrant, we can have
𝜋 𝜋
1
∫𝑟=0 ∫𝜃=0
2
∫𝜑=0
2 √1 − 𝑟 2 |𝐽|𝑑𝑟 𝑑𝜃 𝑑𝜑.

So, we just made the problem a little bit more interesting by imposing these three
conditions. So, if we impose these three conditions that means, x will lie in the first
quadrant, y will lie again in the second quadrant, and z will also lie in the so y will also
lie in the first quadrant, and z will also lie in this first quadrant, so that means, they are
both lying in the positive quadrant of x, y, and z-axis.

And then in that case 𝜃 will go from 0 to 2𝜋, and 𝜑 will also go from 0 to 2𝜋, so that is
what we have written here. And now this Jacobian we know that is basically 𝑟 2 𝑠𝑖𝑛𝜃. So,
𝜋 𝜋
1
let us write that and this will be ∫𝑟=0 ∫𝜃=0
2
∫𝜑=0
2 𝑟 2 𝑠𝑖𝑛𝜃√1 − 𝑟 2 𝑑𝑟 𝑑𝜃 𝑑𝜑. And now we

can evaluate these three integrals, so basically the triple integral one by one.

245
(Refer Slide Time: 29:15)

𝜋 𝜋
1
So, what we can do, we can write ∫02 𝑑𝜑 ∫02 𝑠𝑖𝑛𝜃𝑑𝜃 ∫0 𝑟 2 √1 − 𝑟 2 𝑑𝑟. So, this is what we
will get, and we can calculate this one from our usual trigonometrical results. So, we
again substitute 𝑟 = 𝑠𝑖𝑛𝑡, and then from there we can try to calculate this for this
function this integral here, calculating these two would not be difficult.

𝜋 𝜋
So, this one is basically 2 , and this one is −𝑐𝑜𝑠𝜃, so 𝑐𝑜𝑠 2 is 0, and this so this basically
𝜋
will give us 1. So, the value of this integral is 1, this one is 2 . And the calculation of this

one should be fairly easy using some trigonometrical results.

𝜋 1
And ultimately, we will obtain the value as . 1. 4 , this one is 1, and this one is just you
2

just calculate the value and substitute here that will that will give us the required value.
𝜋
So, this is 2 , this is 1. And you have to calculate the value of this part to substitute here,

and that will be the answer of this integral. So, I am leaving the calculation up to the
students, because it is a very simple example.

So, here in this case also our given region of integration was this sphere bounded by
three other conditions that the all the points should lie on the positive side of x-axis, y-
axis, and z-axis that means, we are in the first quadrant. And therefore, your r 𝜃and
𝜋 𝜋
𝜑would vary from 0 to 1, , 2 and respectively. And then we substituted the values for
2

246
x, y, z using spherical polar coordinate system. And then we just have to do this simple
calculation, so and that will give you the required answer.

So, here again in this example, we saw that how we calculate the triple integral on a
bounded region. And we can also work out several examples of this type. All you have to
do is to find out the region or the domain for the variables x, y, and z for which we can
do the v integral actually. And once you have the region, then doing the integral or
performing the triple integral would be fairly easy as we saw in these examples.

So, we will try to include some more examples in your assignment sheet, so that you can
be able to practice them, and be perfect at them. And we will stop our triple integral part
here. And in the next class, we will start with our area plane regions. So, thank you for
your attention today, and I will look forward to you in the next class.

247
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 24
Area of Plane Region

Hello students. So, up till last class we looked into the triple integral, we also tried to
worked out few examples related to triple integral and how one can basically how to say
evaluate these triple integral on a bounded region in 3 dimensional geometry. So, now,
that we have covered our double and triple integral, we will go to the application part of
integral calculus, where we calculate the area of plane regions. So, we do some
rectification, we calculate surface integral, volume integral and things like that.

So, today we will start with area of plane region. So, this is basically our new topic
today, we will start with our new topic and excuse me.

(Refer Slide Time: 01:23)

So, this is basically today’s topic which we start with. So, let us go to the introduction
part. So, what do we mean by area bounded between two points and a curve? So, we
know that in our usual integral calculations when we say that the area bounded between
two points.

248
(Refer Slide Time: 01:51)

Let us say x equals to. So, this is our y axis this is our x.

So, let us say x equals to a and x equals to b and if this is a curve then let us say f(x) then
the area bounded between these two points and the curve is basically this area here and it
𝑏
can be given by this integral ∫𝑎 𝑓(𝑥)𝑑𝑥. So, that is our usual definite integral formula for
the area bounded between two points and the curve. And similarly if we have a curve of
let us say this type. So, this one was the curve of 𝑦 = 𝑓(𝑥), now suppose if you have a
curve of this type 𝑥 = 𝑓(𝑦)and if we if you want to calculate the area bounded between
the points let us say y=c to y=d, then in that case your area bounded between these two
𝑑
points and the curve can be given by integral ∫𝑐 𝑓(𝑦)𝑑𝑦.

So, regardless whether you are integrating with respect to x or integrating with respect to
y, it just depends on the context what kind of curve you are given and what type of points
you are given all right. So, this is just to how to say a definition, where we calculate the
area between two points and a curve. Now suppose you are given curves like. So, we
𝑥2 𝑦2
might be given curves like ellipse, let us say 𝑎2 + 𝑏2 = 1 or we can be given a hyperbola.
𝑥2 𝑦2
So, this is basically a hyperbola 𝑎2 − 𝑏2 = 1 and similarly several other curve.

So, what kind of area is bounded by this curve here and between two points? Let us say x
equals to given two points actually for this hyperbola. So, we can be given any curve and

249
some points where we from using that we can be able to calculate the area bounded
between the points and the given curve. So, that is what we basically mean by areas of
plane regions actually. So, we will perform several examples, where we calculate the
how to say area bounded between these curves and the points and we will see how much
is that area basically.

So, for example, area bounded between an ellipse. So, inside the ellipse, the area of
bounded by this ellipse is actually 𝜋𝑎𝑏. So, we will see that how we can calculate that.

(Refer Slide Time: 04:47)

Now we can also calculate area between two closed curves, or simply curves they do not
need to be closed. So, let us say two curves. So, between two curves; so, let the two
curves, I have to write the definition. So, let two curves 𝑦 = 𝜑1 (𝑥) and 𝑦 = 𝜑2 (𝑥)
intersect at two points say (a,c) and (b,d).

So, they are intersecting at two points and let between these points, the first curve lies
above the second curve. So; that means, we have x axis we have y axis. So, that is my
𝑥 = 𝑎 this is my 𝑥 = 𝑏 and suppose we have a curve like this and then we have a curve
like this. So, this point here is basically (b,d) and this point here is basically (a,c) and this
is the. So, this is basically 𝑦 = 𝜑1 (𝑥) and this 𝑦 = 𝜑2 (𝑥) and that is the area bounded
between these two curves all right.

250
So, if this is the area bounded between these two curves, then it can be given by. So,
such as this, such as given in the figure then the area between the curves is equals to area
under the first curve. So, we first calculate the area under the first curve and then we
subtract the area under the second curve and that will give us the area between them.

So, this is area under the second curve right. So, how do we calculate the area under the
first curve? So, it is between these two points and then 𝜑1 (𝑥) and the area under the
𝑏
second curve is again between these two points and then 𝜑2 (𝑥)𝑑𝑥 𝑖. 𝑒. ∫𝑎 𝜑1 (𝑥)𝑑𝑥 −
𝑏
∫𝑎 𝜑2 (𝑥)𝑑𝑥 all right and then we just calculate the value of the individual sub integrals
and subtract them and that will give us the area bounded between these two curves.

So, that is another way to calculate the areas between these two curves. So, now, let us
consider an example and see how we actually calculated.

(Refer Slide Time: 08:41)

𝑥2 𝑦2
So, example 1, let us say find the area bounded by the ellipse + 𝑏2 = 1. So, we know
𝑎2

that if we substitute 𝑎 = 𝑏, then an ellipse reduces to a circle and for a circle the area is
𝜋𝑎2 ; however, what can be the area for an ellipse? As I told you that the answer is 𝜋𝑎𝑏,
but how do we actually calculate it we will see just now. So, first of all let us draw an
ellipse.

251
So, this is our x axis, this is our y axis and if we draw the ellipse, as I told you I am really
not good at drawings. So, this is our ellipse, that is basically the semi major axis this is
semi minor axis or this length is a this length is b. Now we know that an ellipse is a
symmetric along both the axis’s.

So, if we can be able to calculate the area in any one of these quadrants, then we just
multiplied by 4. So, whatever the area is here it will be the similar area here, it will be
similar area here and then it will be similar area here. So, we do not have to calculate
over the entire ellipse we just calculate for one of these portions of the ellipse and then
we just multiplied by 4 all right. So, let us see how we can do that. We can write it that
the ellipse is symmetrical about both the axes or coordinate axes so, that.

So, that the two axes divide it into 4 portions, whose areas are equal. So, now, we will
basically integrate this region. So, we will basically integrate the area bounded between
this curve and these two points. So, x=0 to x=1. So, the area let us say this is O A and B;
𝑎
so, the area of the region. So, area of the region OABO is equals to ∫0 𝑦𝑑𝑥and our y is

𝑥2 𝑎 𝑥2 𝑏 𝑎
basically 𝑏 √1 − 𝑎2dx right. So, this will be ∫0 𝑏√1 − 𝑎2 𝑑𝑥 = 𝑎 ∫0 √𝑎2 − 𝑥 2 . Now this

we know that from our usual integral calculus formula that is this will somehow be I
mean the answer will involve sin−1 𝑥.

(Refer Slide Time: 12:43)

252
 𝑏 𝑥√𝑎2 −𝑥 2
So, we can write now the value of that integrals. So, it will be [ +
𝑎 2
𝑎2 𝑥
sin−1 𝑎] 𝑎0 So, now, if I substitute the values then this will reduce to. So, when x is a
2
𝑏 𝑎2 𝑎
this is 0 when x is 0 then this is again 0 now this one is 𝑎 [ 2 sin−1 𝑎 − 0] . So, ultimately
𝜋 𝑏 𝑎2 𝜋
this is 0 this is sin−1 1; sin−1 1 is 2 . So, ultimately we will have .. So, this will be
𝑎 2 2
𝜋𝑎𝑏 𝜋𝑎𝑏
. So, from here the area of OABO is basically .
4 4

Now since we just talked about that ellipses symmetric along both the axes. So, it is
symmetric along the both the axes so; that means, if we calculate the area of any one of
these portions, then we just multiply it by 4 and then it will be the entire area of the
ellipse. So, to calculate the area of the ellipse, so, this implies the area of the ellipse is
𝜋𝑎𝑏
equals to 4 times area of the portion OABO so; that means, 4 times so; that
4

means, 𝜋𝑎𝑏 all right. So, that is basically the required area of an ellipse our area bounded
by the ellipse.

So, this is one such example where we can calculate the area of a function or of a given
curve bounded between two points and then we multiplied by 4 and that that will give us
the area of the entire portion or bounded by that curve. So, that is how we calculate the
area of this plane region, next we consider an another example.

(Refer Slide Time: 15:11)

253
So, prove that or we can basically say evaluate. So, we will just evaluate that will make it
more interesting. So, evaluate the area of the region bounded by the curve 𝑎4 𝑦 2 =
𝑥 5 (2𝑎 − 𝑥) right.

So, let us write the solution. So, that is my x axis, this is origin and that is my y axis.
Now we have 𝑥 5 (2𝑎 − 𝑥) so; that means, and here we have 𝑦 2 so; that means,;
obviously, the point (0,0) lies on this curve because if we substitute x=0 and y=0. So, the
point (0,0) lies on this curve. So, that is pretty much straight forward. So, one of the
points on this curve is (0,0) now we have x equals we have 2𝑎 − 𝑥. So, when y is 0 then
basically x=2a. So, the curve, so, this is basically x=2a. So, of course, it passes through
the point (0,0) and then it also pass through the point (2a,0) because when y is 0 x is 2a.

So, it passes through the point x=2a and y=0 and now we have 𝑦 2 . So, for y negative
and y positive, it will have the same value so; that means, it also how to say it also lies I
mean in a way both sides of x axis in a way so; that means, it forms basically a loop. So,
what I am trying to say is that it forms some kind of looping. So, it will look something
like this and not exactly of this figure, but it will looks something. So, a little bit down
here and then up here and then a little bit down here and then up here.

So, that is how it will look like and this is the region bounded by this, let us call this as E.
So, this is the region that will be bounded by this by this curve. So, the curve we can
write now the curve consists of loop lying between the line x=0 and x=2a and it is
√𝑥 5 (2𝑎−𝑥)
symmetrical along the x axis because the values of y are ∓ so; that means, y is
𝑎2

lying both above the x axis and below the x axis.

So, that is how basically it will form how to say a loop. So, it it has some part lying
above the x axis and line below the x axis. So, this is this is what we mean here. So, the
required area is equals to 2 times area of the region let us give it a name OA say some
point that here is B. So, of the region OABO in a way and the area of the region this
much only. So, area of this. So, we just calculate the area of this region and then we just
2𝑎 1 5
multiply by 2. So, the area of this region only will be 2 ∫0 𝑥 2 √2𝑎 − 𝑥𝑑𝑥 all right.
𝑎2

254
(Refer Slide Time: 20:25)

5
2 2𝑎
So; that means, we have 𝑎2
∫0 𝑥 2 √2𝑎 − 𝑥𝑑𝑥all right. And to evaluate the integral now

we have a how to say square root form. So, we substite 𝑥 = 2𝑎𝑠𝑖𝑛2 𝜃. So, we just
substitute 𝑥 = 2𝑎𝑠𝑖𝑛2 𝜃. So, that will end up getting some kind of 𝑐𝑜𝑠 2 𝜃 that will how
to say neutralize the square root. So, that is what we will do here and ultimately if we do
that then the whole thing will reduce to. So, the whole thing will reduce to there is no
such symbol. So, the whole thing will reduce to
𝜋 5
2
𝑎2
∫0 (2𝑎)2 𝑠𝑖𝑛5 𝜃√2𝑎𝑐𝑜𝑠𝜃4𝑎𝑠𝑖𝑛𝜃𝑐𝑜𝑠𝜃𝑑𝜃 .
2

𝜋
So, ultimately this whole thing will reduce to 64𝑎2 ∫02 𝑠𝑖𝑛6 𝜃𝑐𝑜𝑠 2 𝜃𝑑𝜃 all right. So,
𝑠𝑖𝑛6 𝜃𝑐𝑜𝑠 2 𝜃 now this falls into our reduction formula category. So, we have already
evaluated integrals of this type in our reduction formula type situation. So, this will be
5311𝜋
64𝑎2 8 6 4 2 2 .

5
So, if we calculate the whole thing then it will be 4 𝜋𝑎2 and that is basically our required

area of this where is that of this region. So, of this entire region. So, we did not have to
calculate the area of the whole region since it is symmetric about the x axis, we just
calculated the area of the of the above region and then multiplying it by 2 it will give us
the area of the whole region, and that is what we have done and the required answer is

255
5
𝜋𝑎2 . So, this is how we calculate the area of this plane region. Next we have a curve
4

that is given something like this.

(Refer Slide Time: 23:21)

Let us consider an another example. So, sometimes instead of having a Cartesian

coordinate expression you may have a polar coordinate expression and the curves like a
cycloid or lemniscate of Bernoulli things like that, they are expressed in a they can be
expressed in terms of a polar coordinate system as well.

So, let us look at one such curve. So, find the area enclosed between one arc of the
cycloid 𝑥 = 𝑎(𝜃 − 𝑠𝑖𝑛𝜃), 𝑦 = 𝑎(1 − 𝑐𝑜𝑠𝜃) and its base. So, basically if you look into
any two 2D coordinate geometry book then in that case there you can be able to see all
these nice figures and the figure for cycloid can be given by something like this here let
us say A and P. So, just one arc and its base. So, this is the base and that is the one arc all
right.

Next, so, this is Y 0 A P all right.So, to describe the first arc of the cycloid theta varies
from 0 to 2𝜋 and the coordinates of O and A are (0,0) and (2a𝜋,0). So, here this point is
(2a𝜋,0). and this one is (0,0). So, this is basically the two points where we how to say
evaluate the area of this cycloid. So, let us do that. So, the required area is basically
2𝑎𝜋
integral ∫0 𝑦𝑑𝑥.

256
(Refer Slide Time: 26:53)

So, we can write it as 2a𝜋. So, we have basically y equals integrals. So, this is basically
2a𝜋. So, we have 𝑦 = 𝑎(1 − 𝑐𝑜𝑠𝜃)right and then we have dx, but since this is in 𝜃 and
that is in dx. So, we have to somehow transform this x into a 𝜃 form. So, we know that x
is basically 𝑎(𝜃 − 𝑠𝑖𝑛𝜃) all right. So, 𝜃 − 𝑠𝑖𝑛𝜃. So, from here we have 𝑑𝑥 =
𝑎(1 − 𝑐𝑜𝑠𝜃) is that correct and when x is 0 then in that case 𝜃 is 0 and when x is a then
𝜃 is 2𝜋. So, we can write this here like which is basically the required area.

2𝑎𝜋
So, that required area is basically ∫0 𝑎(1 − 𝑐𝑜𝑠𝜃)𝑎(1 − 𝑐𝑜𝑠𝜃)𝑑𝜃 all right. So, now,
2𝑎𝜋
we can break this whole thing into 𝑎2 ∫0 (1 − 2𝑐𝑜𝑠𝜃 + 𝑐𝑜𝑠 2 𝜃)𝑑𝜃, then we are just a
plus𝑐𝑜𝑠 2 𝜃 then we are just a 2 here by adjusting a half and this will reduce to
2𝑎𝜋 1+𝑐𝑜𝑠2𝜃
𝑎2 ∫0 (1 − 2𝑐𝑜𝑠𝜃 + ) 𝑑𝜃 and now we can integrate this individual term one by
2

one. So, this is fairly simple and I am pretty sure you can be able to do that.

So, this will whole thing will reduce to 3𝜋𝑎2 is the required answer. So, in this case we
were given a cycloid. So, from here first of all we need to find out how does that curve
look like. So, sometimes you also need to practice with curves. So, that just looking at
the curve we can make out this is the way this curve should look like. And to describe
the first arc of the cycloid we can see that theta will vary from 0 to 2𝜋, and the
coordinates for O and the point A are basically at (0,0) and (2𝑎𝜋, 0) and the required
area would be then this one here.

257
So, this is our required region or the area which we want to calculate. So, of course, x
will vary from 0 to 2𝑎𝜋 and y dx. So, this is the y part. So, y dx and we substitute the
value of y here and dx can be calculated from here which we have done in the second
page. So, here I have forgotten a 𝑑𝜃. So, 𝑑𝜃 and then we just substitute the value of dx
here. So, this a square will come here and we will have a 1 − 𝑐𝑜𝑠 2 𝜃. So, it is just a
simple trigonometrical calculation and I am pretty sure you can be able to do that and
ultimately we will obtain 3𝜋𝑎2 which is the required area of the cycloid and its base
between the cycloid and its base.

So, will stop here for today, but in our next class will start with a new example related to
area of plane regions and will how to say practice may be at least one or two more
examples to clarify that concept and then will move to a rectification if time permits. So,
thank you for your attention and look forward to your next class.

258
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 25
Area of Plane Region (Contd.)

Hello students. So, in the last class we looked into the concepts of how calculating Area
of Plane Regions where we have also looked into the formula how do we calculate the area
which is bounded between 2 different curves, we also worked out few examples. Now in
area of plane regions there are several other problems which we can actually work out with
and today we will look into some more examples where to just to make the concepts of
this area of plane regions a bit clear. So, just as a recapitulation we will start with an another
example.

(Refer Slide Time: 01:03)

So, let us consider an another example. So, let us say example 1 for today and this example
says that find the whole area of the curve 𝑥𝑦 2 = 𝑎2 (𝑎 − 𝑥) and the y axis. So, here in this
case the given curve here is basically is this one and we have to calculate the area between
this curve and the whole of y axis. So, this curve would more or less will look like
something of this type.

So, it will look like this is x axis and this is y axis and then it would look somewhat of this
type. So, this is our y axis and we have to calculate the area between this curve and the

259
entire y axis. So, of course, this curve is symmetric at the point along the x axis. So, it is a
so, what we have to do? We have to calculate the area of this region up to y axis and then
we just multiply it by 2. So, that will give us the area between this curve and the y axis.

So, there are certain properties we can write. Say the curve the equation of the curve is
𝑥𝑦 2 = 𝑎2 (𝑎 − 𝑥) and first property is the curve is and the first property is it is the curve
is symmetric along or about x axis and it contains even power of y. The curve passes
through what is the point that it is passing through (a,0) and we so, these are the 2 properties
of this curve. So, then to calculate the area, so, the required area is equal to 2 times area
bounded between the curve and the upper half of x axis. So, this is this area basically.

So, that is the area bounded between this curve and the upper half of x axis and we can
calculate this one in the following way. So, in this case since we are calculating from how
to say from 0 to ∞ in a way because y is varying from as x is varying from actually 0 to
∞ because we are going along the y axis. So, the equation of y axis is x and y axis is 𝑥 =
∞
0 and here we have x will go until infinity. So, we can write 2 ∫0 𝑥𝑑𝑦

(Refer Slide Time: 05:44)

∞ 𝑎3
And this can be written 2 ∫0 𝑑𝑦 and from here we can write dy. Now from here this
𝑎2 +𝑦 2
1 𝑦 𝜋
will reduce to 2𝑎3 [𝑎 tan−1 𝑎]∞ −1
0 now when tan ∞ then this value will be and
2
1
when tan−1 0 then in that case the value will be 0. So, we will have 2𝑎3 𝑎 tan−1 ∞ .

260
 𝜋
So, that is 2𝑎2 2 . So, the value will be 𝜋𝑎2 . So, this is the required area that is being

bounded between this curve. So, we can write this point as (a,0). So, between this curve
and the origin and to calculate that we just have to calculate the upper half and multiply it
by 2 because this curve is symmetric along the x axis. So, this is another example where
we calculated the area of a plane region bounded between 2 curves. We can consider an
another example.

So, to talk about an another example we start with let us say a curve like this; so, example
2.

(Refer Slide Time: 07:15)

So, show that the area between 𝑎2 𝑥 2 + 𝑏 2 𝑦 2 = 1 and 𝑏 2 𝑥 2 + 𝑎2 𝑦 2 = 1 where 0 < 𝑎 <
4 𝑎
𝑏 is 𝑎𝑏 tan−1 𝑏. So, here we have to be a little bit careful. So, it is not that; so, first of all

we have to see what kind of curve they are.

𝑥2 𝑦2
So, we can see clearly that if we divide both sides by 𝑎2 𝑏 2 then this will be + 𝑎2 =
𝑏2
1 𝑥2 𝑦2 1
and similarly this will be + 𝑏2 = 𝑎2 𝑏2 . So, then in that case it is basically an
𝑎2 𝑏 2 𝑎2

equation of an ellipse actually. So, this is our equation of an ellipse and we can actually
𝑥2 𝑦2
write instead, in fact, we don’t have to multiply anything with we can write 1 + 1 =
𝑎2 𝑏2

1 and that will be a sufficient.

261
So, we don’t have to divide anything. So, it is pretty much obvious that if we can write it
𝑥2 𝑦2
as 𝑎2 𝑥 2 + 𝑏 2 𝑦 2 = 1 then from here it will be 1 + 1 = 1. So, this will basically be our
𝑎2 𝑏2

how to say square of the length of the semi-major axis and square of the length of the semi
𝑥2 𝑦2
minor axis. Similarly, we can write this one as 1 + 1 = 1so, this is our another equation
𝑏2 𝑎2

of an another ellipse. Now we note that 𝑎 < 𝑏.

1 1
So, since 𝑎 < 𝑏 from here it implies that 𝑎 > 𝑏. So, then in that case in the first case now

we are ready to draw our area bounded between these 2 curves. So, this is our x axis this
1 1
is our y axis that is the origin. So, if 𝑎 > 𝑏 then in that case this will act as the length of the

square of semi-major axis and this will act as the semi minor axis. So, our first ellipse will
1 1
look like this and since > , so, this ellipse will be inverted. So, something like this.
𝑎2 𝑏2

So, the area bounded between them is this much and since they contain x square and y
square; that means, they are symmetrical in a way. So, it basically means that we have to
calculate the area of any one of these parts and then multiply it by 4. So, we don’t have to
calculate the area of all these sections altogether, we just have to calculate the area of any
one of these small how to say sections and then that will give a multiplied by 4 and then it
will give the total area bounded by these 2 ellipse. So, we can call it as A B C. So, OABC
area of OABC will be the multiplied by 4 will be the area of the entire region.

So to calculate that we can calculate this point of intersection and if we calculate that point
of intersection, so, what we basically do? We substitute the value of 𝑥 2 from here to here
and we calculate the value of y and similarly we substitute the value of y then we get the
value of x.

262
(Refer Slide Time: 11:38)

So that means, the point of intersection, we get first point of intersection. So, we get
1 1
solving the 2 ellipse equation we get one point of intersection as (√𝑎2 , √𝑎2 )
+𝑏2 +𝑏 2

So, these are the one point of intersection up there of these 2 ellipse. Similarly, we will get
the point of intersection of this point and it will surely give you the 4 different point of
intersection because once you are substituting the value of x here than in that case you are
getting the 2 roots of y; that means, you are getting 2 values of y positive and negative.

Similarly, for each value of y then we will get 2 values of x and then we have basically 4
combinations.

So, x and y both positive, x is negative y is positive then x and y both negative and then x
is negative y is positive. So, we will basically get 4 different types of combination here;
that means, 4 different types of point of intersection all right. So, we are just focused in
this first quadrant because we will multiply the area of the first quadrant by 4 ok.

So, now, that we have the point of intersection we will write now the required area now
the required area is equals to 4 times area of I have named it OABC or so, let me say it
OABCO. So, that is the area and area of OABCO is basically 4 times area of if we call this
point of intersection let us say B then in that case it will be so, then in that case it will be
area let us say of this curve and then this curve.

263
So, suppose this point is P, I am calling this point as P so; that means, we are basically
taking the area of the entire region as the area of 2 different area bounded by these 2 curves
because this is the first curve and then this is the second curve. So that means, we will
integrate first from here to here and then here to here. So, 2 x points; so that means, we
will get basically area of what is that? Area of OPBC; so OPBC plus area of PAB OABC
and then area of PAB right. And for area of PAB we have to now we will calculate the
area individually. So, for area of PAB here this is the part of so, PAB is basically this 1
here and this is the part of our inverted ellipse and from here we have to calculate the value
of y actually.

So, if we calculate the value of y then that will be so, for the area of PAB the limits are
given in the equation from; so, what are the limits? So, the limits are from this point to this
1
point; that means, this point is basically A and this point is basically OP √𝑎2 +𝑏2
So, from

O P to PA right; so, from here to here and then here to here. So, the limit will be from OM
to OC and so, this one is basically O O P to O A. So, the limits are basically OP to OA and
1
OP is basically √𝑎2 +𝑏2
and OA is basically our a and from here we can solve y. So, y is
1
basically 𝑏 √1 − 𝑎2 𝑥 2 . So, this is our y actually. So, now, that we have y and we have our

limits, so, we can calculate the area of the region PAB.

(Refer Slide Time: 17:58)

264
 𝑎
So, the area of the region PAB is basically ∫ 1 𝑦𝑑𝑥. So, here we will have
√𝑎2 +𝑏2

𝑎 1
∫ 1 √1 − 𝑎2 𝑥 2 𝑑𝑥 and similarly the area of this region, similarly we will calculate the
𝑏
√𝑎2 +𝑏2

1
area of this region by integrating from 0 to √𝑎2 and then we will take the curve y from
+𝑏2

here and now we have to find the integration of this thing here. So, we have to calculate
the integral. So, this is pretty straightforward. So, we will take out a square from here.

𝑎 𝑎 1
So, then it will be if I take 𝑎2 from there then it will be 𝑏 ∫ 1 √ − 𝑥 2 and then we will
𝑎2
√𝑎2 +𝑏2

use the known formula of integral calculus and if we integrate this then this will actually
reduce to something called then this will actually reduce to and the first part will reduce to
1 𝑥
sinx times 𝑎2 (𝑎2 − 𝑥 2 ) by a plus sin−1 1 . So, this can be evaluated very easily and
𝑎

similarly we can evaluate the area of this part as I showed you that we have to now integrate
from the origin to this point and this point is 1 by a square plus b square and then we will
use this curve.

So, the value will come from here. And now in this case it is 𝑏 2 𝑥 2 and then 𝑎2 𝑦 2 . So, this
is√1 − 𝑏 2 𝑥 2 and we will divide it by a and sorry so, this one is 𝑏 2 𝑥 2 and we have divided
𝑏
by a. So, this is 𝑎 and this here is 𝑏 2 . So, ultimately we will integrate with respect to x here

and then that will give us the required value of the integral and similarly we will integrate
this part and that will give us the area of this part and then we will sum them. So, then it
will give the area of the whole ellipse of this section and then we will multiply by 4.

So, when we multiply by 4 that will be the required area of the whole region bounded by
the ellipse.

265
(Refer Slide Time: 21:07)

So, the required area would be 4 times area of PAB plus what was that another one was
1 𝑎
area of OPBC and that will give us actually 4 times. So, this will come out to be 𝑎𝑏 tan−1 𝑏.

So, it is a bit lengthy example, but I am pretty sure you got the idea what I was trying to
say. So, this is an another way to calculate the area of 2 of the region bounded by 2 plane
curves.

So, here in this case the 2 curves were ellipse and you just have to in always remember
that in this case in these type of examples you always have to draw the figure first. So, if
you can be able to draw the figure correctly then first you have to notice where are they
intersecting. If they are intersecting, then calculate the point of intersection or at first you
can solve them and then you will get the point of intersection.

And once you get the point of intersection you sort of can draw the figure and if you draw
the figure then you will be able to see that whether you need to calculate the area only
above the one curve or you have to involve both the curves and if you are involving the
both the curves and you have to guess the limits correctly like we did in this case and once
you have the limits correctly then calculating the am area would be pretty much straight
forward all right.

So, next point we have in agenda is sectorial area; so, sectorial area. This is motivated from
a polar coordinate system. So, sectorial area means if 𝑟 = 𝑓(𝜃) be the equation of a curve
in a polar coordinate system, then the area of the sector enclosed by the curve and the two

266
radial vectors or two radii sometimes they are also called as radial vectors 𝜃 = 𝛼 and 𝜃 =
1 𝛽
𝛽 is 2 ∫𝛼 𝑟 2 𝑑𝜃.

We will not go into the proof, but what we mean is. So, suppose this is our x axis in a way
so, and suppose this is a given curve, now this point here is the point P and this is our point
A and that is our B.

So, this angle XOA is basically angle 𝛼 and XOB is basically our angle 𝛽. So, that sectorial
area means the area bounded between OA and OB and this curve. So, this area and this
radii vector and the curve so, that particular area so, this area is basically called as the
1 𝛽
sectorial area and it is given by 2 ∫𝛼 𝑟 2 𝑑𝜃. This is very useful when a curve is given in the

polar coordinate system.

We will see that what do we mean by it. So, if a curve is given in a polar coordinate system,
we can calculate the sectorial area using this formula.

(Refer Slide Time: 25:59)

So, let us see an example. So, let us say we have find the area of the cardioide 𝑟 = 𝑎(1 −
𝑐𝑜𝑠𝜃). So, the curve first of all if we draw the how to say a cardioide. So, I am pretty sure
you may have seen the figure for the cardioide before. So, it looks something like this all
right. So, that is the origin. Let us say this is the point A and that is how it looks like and
it is symmetrical the curve is symmetrical about the initial line. So, the initial line is

267
basically 𝜃 = 0. So, if 𝜃 is 0 then in that case we have cos 0 as 1 and this 1 will be (1-1) 0
so, r is also 0.

So, that means, that this is our initial line where 𝜃 is 0 and the curve is actually
symmetrical. So, the required area will then be just 𝜃 from 0 to 𝜋 because it is symmetrical
along this initial line. So, we just multiply it by 2 and that will give us the area of this entire
𝜋 1
cardioide. So, the required area is equal ∫0 𝑟 2 𝑑𝜃 and since the formula is we write
2
1 𝜋
∫ 𝑟 2 𝑑𝜃 and since the required area is 2 times of the area of any one of these sections;
2 0
1 𝜋
so, we have to also write multiplied by 2 here 2 × ∫0 𝑟 2 𝑑𝜃 .
2

𝜋
So, then in that case this 2 and this 2 will get cancelled and we will have∫0 𝑟 2 𝑑𝜃 . Now r
𝜋
is 𝑎(1 − cos 𝜃). So, here we will have∫0 𝑎2 (1 − cos 𝜃)2 𝑑𝜃 . So, this will turn out to be
𝜋
𝑎2 ∫0 (1 − 2𝑐𝑜𝑠𝜃 + 𝑐𝑜𝑠 2 𝜃 )𝑑𝜃.

1 𝜋 1+cos 2𝜃
Now I can write 2 here and then this will reduce to 𝑎2 ∫0 (1 − 2𝑐𝑜𝑠𝜃 + )𝑑𝜃. Now
2

evaluating this integral is fairly easy, I am pretty sure you can be able to evaluate. So, this
1 𝑠𝑖𝑛2𝜃
will become 𝑎2 [𝜃 − 2𝑠𝑖𝑛𝜃 + (𝜃 + )]𝜋0 and then you substitute the value of 𝜃.
2 2

3𝜋 3𝑎2 𝜋
So, ultimately you will be able to obtain 𝑎2 × . So, this is basically . So, that is the
2 2

area bounded by this cardioide between I mean in this region actually. So, you see instead
of converting this into a Cartesian coordinate system and then trying to draw this curve
and also finding out the limits and all we just had to remember the shape of this curve and
from the shape we can see that the curve is symmetrical about the initial line and then we
can calculate the area or the sectorial area in this sense by just writing the limit 0 to 𝜋 and
then multiplying it by 2 so, that will have the entire area and then you just do some simple
calculation and that will give you the required answer. So, this is a very interesting example
where we actually did not have to take help of the Cartesian coordinate system. We can
consider an another example.

268
(Refer Slide Time: 30:40)

𝜋
So, let us see. So, find the area of the curve 𝑟 2 = 𝑎2 cos 2𝜃. So, the value of 𝜃 = ± 4 will
𝜋 𝜋
make 𝑟 = 0. So, for 𝜃 = − 4 it will be 0 r will be 0 and 𝜃 = + 4 r will be 0.

So, the curve is symmetrical about the initial line and then the required area of the curve
is equals to 4 times because this curve is will look like something of this type. So, this is
our initial line. So, that is the point A let us say this is point B. So, we are basically
𝜋
1
calculating this area and then we are multiplying it by 4. So, 4 × 2 ∫04 𝑟 2 𝑑𝜃.

𝜋 𝜋 𝜋
sin 2𝜃 4
So, this is basically 2 × ∫04 𝑟 2 𝑑𝜃 = 2 ∫04 𝑎2 cos 2𝜃 𝑑𝜃and then this 2𝑎2 [ ]0 =
2
𝜋
𝑠𝑖𝑛
2𝑎2 . So, that means, this will be ultimately 𝑎2 yes. So, the required area of this, this is
2
2

also called as a Lemniscate of Bernoulli and the curve actually looks like this.

So, in order to calculate the area we have to see for what value of theta this is becoming
how to say r becoming 0 and that will actually give us the limit from where to where we
have to integrate and the area of the entire curve will be given by 4 times the area of these
small sections. So, 4 times the area of this region is 4 times this region and then the area
of the entire Lemniscate of Bernoulli will be 𝑎2 . So, this is another way where you can
calculate the area of a curve which is given in terms of polar coordinates. So, we will stop
here for today and we will continue with our further examples in our next lecture.

269
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 26
Rectification

Hello students. So, in the last class, we looked into the concepts of area of plane regions;
where we calculated the area of curve given by the equation either 𝑦 = 𝑓(𝑥) or 𝑟 =
𝑓(𝜃) which basically mean that, we either calculated the area bounded between a curve
and two points or the sectorial area where, we also worked out several examples. And I
will also try to include some examples in your assignment sheet so, that they can practice
them.

Today our topic in agenda is Rectification. So, what do we mean by rectification of a curve
and how do we calculate it. So, we will look into the definition of a rectification of a curve
and then we will work out few examples just to make the concepts clear related to this
topic.

(Refer Slide Time: 01:22).

So, let us start. So, today we will start with rectification all right.

270
(Refer Slide Time: 01:30)

So, let us start; so, rectification alright. So, what do we mean by rectification? Rectification
basically mean that, we will determine basically so, we determine the length of the arcs of
plane curves whose equations are given in Cartesian, parametric Cartesian which means
𝑥 = 𝜑(𝑡), 𝑦 = 𝜓(𝑡); so, parametric Cartesian or polar co-ordinate system alright.

So, this process is called the rectification. So, the process of determining the length of the
arcs of plane curve whose the equation is given, either it can be a Cartesian equation. So,
something like 𝑥 2 + 𝑦 2 = 1 which is basically the equation of a circle with unit radius or
parametric Cartesian.

So, we can write, 𝑥 = 𝑐𝑜𝑠𝑡 and 𝑦 = 𝑠𝑖𝑛𝑡 or polar co-ordinate system. So, for polar
coordinate system, we can write 𝑥 = 𝑐𝑜𝑠𝜃, 𝑦 = 𝑠𝑖𝑛𝜃, where 0 ≤ 𝜃 ≤ 2𝜋. So, any one of
them can be considered and this process is called the rectification of a curve.

So, like the way we calculated the area bounded between a curve and two points. So, the
area bounded between a curve so, the area bounded between a curve let us say, 𝑦 =
𝑏
𝑓(𝑥) and two points 𝑥 = 𝑎, 𝑥 = 𝑏 is given by we know that, ∫𝑎 𝑦𝑑𝑥 and y equals to
basically 𝑓(𝑥)𝑑𝑥. So, that is how we calculate the area.

Now; similarly, the rectifier how to say the rectification or the length of the arc can be
calculated in the following way. So, the length of the arc of the curve, 𝑦 = 𝑓(𝑥) included
between two points say, 𝑥 = 𝑎 and 𝑥 = 𝑏 is given by so, this is our x axis, this is our y

271
axis, that is the point let us say, x equals to a that is the point let us say, x equals to b and
this is our curve.

(Refer Slide Time: 05:46)

So, we are basically calculating the length of the arc. So, just this length and this length of
𝑏 𝑑𝑦 𝑏
this arc can be given by integral ∫𝑎 √1 + ( )2 𝑑𝑥 or we can also write ∫𝑎 √1 + {𝑓 ′ (𝑥)}2.
𝑑𝑥

So, that is the formula for giving the arc length of this curve y equals to f x.

So, you remember for so, that is 𝑥 = 𝑎 and for this point, this is 𝑥 = 𝑏; so, the area is
𝑏 𝑏
basically given by ∫𝑎 𝑓(𝑥)𝑑𝑥 and the length of the arc will be given by ∫𝑎 √1 + {𝑓 ′ (𝑥)}2 .
So, that is the formula for calculating the arc length. The proof is slightly big and we will
ignore the proof here, we are basically interested in the formula and then working out some
examples.

So, if we have an equation of a curve given in the Cartesian co-ordinate system;𝑦 = 𝑓(𝑥),
then the formula to calculate the length of the arc will be given in this fashion. Now, we
just studied that the equation cannot be given always in the Cartesian co-ordinate system;
they can also be given in the parametric Cartesian or polar coordinate system so, how we
can calculate the arc length.

So, there are formulas for such co-ordinate systems as well. So, we write other expression
for calculating arc lengths. So, what are those other formulas? So, the first one is let s be

272
the length of the arc; so, this is basically s length of the arc of the curve included between
the points, 𝑥 = 𝑎, 𝑥 = 𝑏, then what we will have? We will have we know that from co-
ordinate geometry, that or and also from the differential calculus, that 𝑑𝑠 =
√(𝑑𝑥)2 + (𝑑𝑦)2 ; so, this is something we already know.

𝑑𝑠 𝑑𝑦
So, we can write 𝑑𝑥 = √1 + (𝑑𝑥 )2and this can actually be written as so, from here we can

𝑑𝑦
also write let us say,𝑑𝑠 = √1 + (𝑑𝑥 )2 𝑑𝑥 . And now, if we integrate from point a to

𝑏 𝑏 𝑑𝑦
b,∫𝑎 𝑑𝑠 = ∫𝑎 √1 + (𝑑𝑥 )2 𝑑𝑥 .

𝑏 𝑑𝑦
So, this is actually (the value of s at b)-(value of s at a)= ∫𝑎 √1 + (𝑑𝑥 )2 𝑑𝑥 . So, this is

basically our arc of how to say so, if that is the length if that is the length so, here basically
the value of s at b will be the whole length and the value of s will be actually at 𝑥 = 𝑎 a is
0.

So, this is basically the arc of let us say, we can call it as ab; so, this is basically our length
of the arc ab equals to this formula. So, that is basically this formula here. So, this is equals
to basically this formula and we can see that, the calculation for the arc length which is
basically s can be given by this formula.

(Refer Slide Time: 11:41)

273
Similarly, if we express here instead of dividing it by dx, we can divide it by dy and then
we will still obtain the similar formula. So, the second form will be 𝑑𝑠 =
𝑑 𝑑𝑥
∫𝑐 √1 + (𝑑𝑦)2 𝑑𝑦 and that will give us the arc length for the curve. We can also have in

terms of d s dt so, this will be integral from let us say, t is equals to a to t, is equals to b.
This we have to determine what are the limits and in case if the given curve is in parametric
form.

𝑡=𝑏 𝑡=𝑏 𝑑𝑥 𝑑𝑦
So, then in that case, the parametric form will be ∫𝑡=𝑎 𝑑𝑠 = ∫𝑡=𝑎 √( 𝑑𝑡 )2 + ( 𝑑𝑡 )2. So, this

is this formula is for the so, this is when the given curve is of 𝑥 = 𝑓(𝑦) form, this is for
the parametric, this is for the parametric Cartesian and the last form; the fourth form
basically will be for the polar one. So, if we have ds, then we know that 𝑑𝑠 =
𝑑𝑟
√𝑟 2 + ( )2 𝑑𝜃.
𝑑𝜃

And then here we will have a 𝑑𝜃 and we can integrate for theta running from I do not
𝜃=𝛽 𝜃=𝛽 𝑑𝑟
know, certain point 𝛼 to 𝜃 = 𝛽 i.e. ∫𝜃=𝛼 𝑑𝑠 = ∫𝜃=𝛼 √𝑟 2 + (𝑑𝜃)2 𝑑𝜃 let us say, and this

will give us the arc length of the curve which is given by the equation 𝑟 = 𝑓(𝜃) which is
actually the polar setting.

So, depending on the type of the equation that is given to us, we can use either one of the
formulas. Say, if it is 𝑦 = 𝑓(𝑥) , then we use the formula 1, if it is 𝑥 = 𝑓(𝑦) we use the
formula 2, and if it is 𝑥 = 𝜑(𝑡) and 𝑦 = 𝜓(𝑡), then we will use this formula and if we have
𝑟 = 𝑓(𝜃); that means, if we have some polar co-ordinate expression, then in that case we
will use the last formula. And based on that, we can be able to calculate the length of the
arc between two points.

So, earlier we calculated the area between the two points, now we are calculating the length
of the arc of which is between two points actually. So, let us work out a few examples so
to do some example, let us go to a new page.

274
(Refer Slide Time: 14:50)

So, suppose our given equation is example 1. So, the problem is find the arc length or
length of the arc, it is the same so, arc length of the parabola our 𝑥 2 = 4𝑎𝑦. So, let us say
we have 𝑥 2 = 4𝑎𝑦 measured from the vertex of one extremity of the latus rectum, all
right. So, in this case, the abscissa or the x point basically, the abscissa of the extremity L,
of the latus rectum is 2a.

So, the latus rectum is basically given by this way. So, this is our x axis, this is our y axis,
that is the origin and this is our parabola passing through the origin all right and this is our
lattice rectum. So, L and that is 2a so, this point is a and this is our s. And let us say, we
have this point as A and this is our point B all right. So, this is so, we have to calculate the
arc length from here to here. So, and this is basically our latus rectum extremity L of the
latus rectum.

So, now, we have to calculate the length of the arc from this point to this point. So, we can
see that clearly, that here the required length will be given by so, the required length will
be given by because the equation is of type 𝑥 = 𝑓(𝑦). So, the given equation is 𝑥 2 = 4𝑎𝑦.
𝑥2 𝑑𝑦 2𝑥 𝑥
So, we can write it as 𝑦 = 4𝑎; so, from here, our = 4𝑎. So, this is basically 2𝑎.
𝑑𝑥

So, the required length so, the required arc length will be from 0 to 2a; so, integral
2𝑎 𝑑𝑦
∫0 √1 + (𝑑𝑥 )2 𝑑𝑥 so, this is basically 2a that is a. So, it is 0 to 2a and that is 1 extremity

and this is our arc length. So, 0 to 2a dy dx measured from the vertex, from the vertex of

275
1 extremity. So, the vertex from so, this is the whole extremity L of the latus rectum. And
here we are measuring from the vertex to 1 extremity so; that means, we will calculate
basically this arc length all right. So, that is why we are integrating from 0 to 2a ok.

So, this is a hidden part in the question. So, vertex of 1 extremity actually; so, we have to
start from the vertex here and then that is the length we need to calculate so from 0 to 2a.
𝑑𝑦 𝑥
And now, we have as , we substitute there. So, this will be integral
𝑑𝑥 2𝑎
1 2𝑎
∫ √4𝑎2
2𝑎 0
+ 𝑥 2 𝑑𝑥 .

(Refer Slide Time: 19:20)

2𝑎
1 𝑥√𝑥 2 +4𝑎2 𝑥
And we can calculate this one; so, this will be 2𝑎 [ + 2𝑎2 sinh−1 2𝑎] . So, this is
2 0

some kind of integral calculus formula.

So, I am pretty sure you have done this before so, you just write the whole formula here.
And when we substitute, the value as 𝑥 = 2𝑎 and 𝑥 = 𝑎; so, then this will reduce to
1
[2√2𝑎2 + 2𝑎2 sinh−1 1]. So, sinh−1 1, sinh−1 0 will be 0; and therefore, we will be left
2𝑎

with 4𝑎2 to cancel yes.

So, therefore, the that will basically give us the required how to say the length of the arc
or arc length. You can further simplify if you want here, but we will just how to say leave
the answer up to here. So, one can also write the value of sinh−1 1, but this is this will be
the answer.

276
So, you see the given parabola here, was of inverted shape and we were asked to measure
the length from the vertex to one of the extremities of the length of the latus rectum. So,
this is the latus rectum and this point the extremity point is 2a, a and from there we can
calculate the length by integrating from 0 to 2a, by integrating from 0 to 2a and we use
this formula. So, that will give us the required arc length of the given parabola.

(Refer Slide Time: 21:39)

Similarly, we can calculate; we can consider an another example. Example 2 let us say;
𝑒 𝑥 −1
so, an another example it says, find the length of the curve 𝑦 = log 𝑒 𝑒 𝑥 +1 from 𝑥 = 1 to

𝑥 = 2.

So, we are here not concerned about the figure because it will take some time to draw this
figure; it is just that we are given the curve and we are also given the points where we have
to do the, where we have to calculate the length of this curve. So, this can be a curve of I
am not sure, but let us say hypothetically, this can be a curve of something like this and
then this is our point 𝑥 = 𝑎 and this is our point 𝑥 = 𝑎 and this is our point 𝑥 = 2 and we
are basically interested in calculating this arc length, all right.

𝑒 𝑥 −1 𝑑𝑦
So, to do that, the given curve is 𝑦 = log 𝑒 𝑒 𝑥 +1. So, from here we will calculate 𝑑𝑥 equals
𝑑𝑦
to so, 𝑑𝑥 will be we can also write this one as 𝑦 = log 𝑒 (𝑒 𝑥 − 1) + log 𝑒 (𝑒 𝑥 + 1).

277
 𝑑𝑦 𝑒𝑥 𝑒𝑥
So, now, doing the derivative would be fairly easy. So, it will be = 𝑒 𝑥 −1 − 𝑒 𝑥 +1. So,
𝑑𝑥
2𝑒 𝑥
this whole thing will comprise to 𝑒 2𝑥 −1 all right; so, that is the derivative.

Now, the required arc length;. So, this will be from our known formula, we have 𝑦 =
𝑓(𝑥) type situation and if we have 𝑦 = 𝑓(𝑥) type situation, then the arc length is x running
𝑥=2 𝑑𝑦 𝑑𝑦
from a to b. So, our a is 1, b is 2 ∫𝑥=1 √1 + (𝑑𝑥 )2 𝑑𝑥, we have already calculated in
𝑑𝑥

2 2𝑒 𝑥
equation 1 let us say this is our equation 1. So, from 1, we will have ∫1 √1 + (𝑒 2𝑥 −1)2 𝑑𝑥.

We can write it as from 1 all right.

(Refer Slide Time: 25:10)

2 4𝑒 𝑥
So, if we do the square here and then this will reduce to integral ∫1 √1 + (𝑒 2𝑥 −1)2 𝑑𝑥 and

2 𝑒 2𝑥 +1 2 𝑒 𝑥 +𝑒 −𝑥
this can be simplified to integral ∫1 𝑑𝑥 this can be further simplified to ∫1 𝑑𝑥.
𝑒 2𝑥 −1 𝑒 𝑥 −𝑒 −𝑥

We are just dividing by 𝑒 𝑥 or we are multiplying both numerator and denominator by 𝑒 −𝑥 .

Now, we see that this term here is basically the derivative of this term. So, if we substitute
this one equals to z, sorry this is 1, if we substitute this one equals to z, then this will be
the derivative; this term will be the derivative of this one.

278
 𝑑𝑧
So, we can simply write, we know that if we have something like ; so, we write logz
𝑧

right. So, that is the formula we are using here. So,[log 𝑒 |𝑒 𝑥 − 𝑒 −𝑥 |]12. So, this will reduce
to log 𝑒 |𝑒 2 − 𝑒 −2 | − log 𝑒 |𝑒 − 𝑒 −1 |.

𝑒 2 −𝑒 −2
So, this is nothing but, log 𝑒 and if we multiply both numerator and denominator
𝑒−𝑒 −1

by 𝑒 2 , then in that case this will reduce to log of, if we multiply both numerator and
𝑒(𝑒 4 −1) 1
denominator by 2, then this will reduce to log 𝑒 𝑒 2 (𝑒 2 −1). So, this will reduce to log 𝑒 (𝑒 + 𝑒).

So, this is the required answer; I mean we can leave up to here, it is not a problem and that
is how we calculate basically the arc length for this given curve. So, you just have to guess
the form first of all. So, whether it is given as 𝑦 = 𝑓(𝑥), 𝑥 = 𝑓(𝑦), 𝑥 = 𝜑(𝑡), 𝑦 = 𝜓(𝑡)or
a polar system. And from there you have to find out the points where you need to calculate
the length.

So, once you have these ingredients you pretty much know what kind of formula we have
𝑑𝑦
to use so, the in this case I calculated , because the curve is already given in terms of
𝑑𝑥

𝑦 = 𝑓(𝑥) form and from there, this is our required formula to calculate the arc length, we
substituted every value and then we just did some integration.

Now, this is just a calculation part, you can leave your answer up to here and that is still
correct. Simplifying it to a certain form that is not our concern here, how we calculate the
arc length that is our concern. So, this is how you calculate the arc length.

So, we will stop today’s class here and in the next class we will start working on some few
examples which might also include parametric Cartesian or polar coordinate system and
we calculate the length of those arcs. So, thank you for attention and I look forward to you
in the next class.

279
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 27
Rectification (Contd.)

Hello students. So, in the last class we started with Rectification of a curve which basically
means that calculating the arc length. And we worked out few examples we also give
looked into several formulas, some of which we are going to implement today and we
worked out 1 or 2 examples where we saw if we are given a curve let us say 𝑦 = 𝑓(𝑥),
then in that case how we can calculate the arc length between 2 points 𝑥 = 𝑎 and 𝑥 = 𝑏.
So, today we are going to do few more examples and today we will start with a parametric
Cartesian a form where we calculate the arc length.

(Refer Slide Time: 01:05)

So, let us start with our problem. So, example 1: so, here our problem is find the arc length
or the length of the arc its the same thing the arc length of the curve 𝑥 = 𝑡, 𝑦 = 𝑡 2 from
the point 𝑡 = 0 to 𝑡 = 1 all right. So, if we are already given the curve and we are also
given the point where we need to do the integration, then we need only to draw the curve
because we are already provided enough information to perform that calculation of the arc
length.

280
So, here we are given the curve as the given curve is 𝑥 = 𝑡 and 𝑦 = 𝑡 2 . So, from here we
𝑑𝑥 𝑑𝑦
can write as 1 and as 2t. So, therefore, the required arc length is integral
𝑑𝑡 𝑑𝑡

𝑡=1 𝑑𝑥 𝑑𝑦 1
∫𝑡=0 √( 𝑑𝑡 )2 + ( 𝑑𝑡 )2 𝑑𝑡. So, here we have ∫0 √1 + 4𝑡 2 𝑑𝑡 and now this has become our

simple and integral calculus problem. So, we know that from one of those formulas we
can be able to write this integral into a nice formula here, and then we just substitute 𝑡 =
0 and 𝑡 = 1 and that will give us the required answer.

So, I will give you that task up to the students, because this is from here it is pretty much
straight forward. So, this is how you calculate the arc length if we are given the Cartesian
parametric Cartesian system. We can consider an another example of similar source.

(Refer Slide Time: 03:51)

So, example let us say 2. So, find the arc length of the curve let us say 𝑥 = 𝑐𝑜𝑠𝑡. So, and
𝑦 = 𝑠𝑖𝑛𝑡 for 𝑡 = 0 to 𝑡 = 1 . So, here also not only that we know this curve, we also have
all the vital information to calculate the arc length. So, here we have 𝑥 = 𝑐𝑜𝑠𝑡 and 𝑦 =
𝑑𝑥
𝑠𝑖𝑛𝑡 all right. So, from here we can very easily calculate our which is basically
𝑑𝑡
𝑑𝑦
−𝑠𝑖𝑛𝑡 and our would be simply 𝑐𝑜𝑠𝑡 and now we are supposed to calculate the arc
𝑑𝑡

length.

281
 𝑡=1 𝑑𝑥 𝑑𝑦
So, the required arc length is basically integral ∫𝑡=0 √( 𝑑𝑡 )2 + ( 𝑑𝑡 )2 𝑑𝑡. So, this is integral
1
∫0 √𝑠𝑖𝑛2 𝑡 + 𝑐𝑜𝑠 2 𝑡𝑑𝑡. So, this one we can integrate we dont have to put it off. So, 𝑠𝑖𝑛2 𝑡 +
1
𝑐𝑜𝑠 2 𝑡 = 1and there is ∫0 √1𝑑𝑡 and . So, basically we will have t at 1.

So, 1 √1 can also be taken as −1 because we have ±1 as √1, but we cannot take minus
here because length is a positive quantity. So, we cannot have −1. So, −1 does not make
any sense. So, that is why we have taken only 1 as √1 all right and therefore, we got the
length as a positive number. So, this is a how to say a small essence of mathematics I
would say, where you have to be a little bit cautious what you choose as a value and here
is the one such example that we cannot take the negative value we always have to take the
positive value. So, this is how we calculate the arc length of a curve which is given in
parametric Cartesian system and of course, the points are given.

So, the calculation with is fairly simple. Next type of formula is basically of polar
coordinate system. So, let us work out few examples based on the polar coordinate system.
So, will do we proceed in the following way.

(Refer Slide Time: 07:17)

So, example let us say 3. So, rectify or calculate the arc length whatever. So, rectify the
curve 𝑥 = 𝑎(𝜃 + 𝑠𝑖𝑛𝜃) and 𝑦 = 𝑎(1 − 𝑐𝑜𝑠𝜃) . So, how do we do the rectification?

282
So, this is our given curve right my figures are little bit how to say upside down mainly
because I am not a good painter in a way or good at drawing. So, let us draw this curve.
So, it looks something like this. Now about the drawing of the polar coordinate curves or
any Cartesian curves that is a different chapter. So, you may have to look into a chapter on
how to draw the curves. So, there are formulas or methods where you learn about how you
can draw a curve, but that is not basically our concern here. So, here we will we are
assuming that the reader is already familiar with the concepts of how to draw a curve. So,
this is how our curve will look like.

So, this is A and this point is 𝐴′ all right. So, now, we first have to guess the limit because
if we have equation of a curve given in terms of polar coordinate system. So, the
rectification formula involves the integration for 𝜃 = 𝛼 to 𝜃 = 𝛽; that means, we need to
identify these 2 values of 𝜃 these 2 angles we need to identify so, that we can do the
integration. So, here we will try to do the same thing we will first try to identify the range
for 𝜃 that is from where to where it is varying alright. So, as a point let us say any point
on this arc, as a point moves from one end 𝐴′ to other end A of the one arc, the parameter
𝜃 increases basically; increases from −𝜋 to +𝜋 of course, and the parameter 𝜃 is 0 for the
vertex O all right.

So, this is vertex O that is here and 0 is like 0 real number all right. So, as the arc is
symmetrical about OY. So, the arc is I did not draw the symmetric figure, but it is
symmetrical. So, the arc is symmetrical along OY or about OY whichever you prefer to
use about or along OY. The required arc length is basically our arc AO𝐴′ which is 2 times
arc, O𝐴′ right or 2 times arc OA.

So, any one of the arc and then you multiply by 2 and that will give you the length of the
whole arc. So, we can do the calculation for any one of the let us say sub arcs all right. So,
now, we have 𝑥 = 𝑎(𝜃 + 𝑠𝑖𝑛𝜃) .So, then in that case if we are calculating the arc length
for any one of these sub arcs, we can basically have theta running from 0 to 𝜋 all right. So,
𝑑𝑥
𝑎(𝜃 + 𝑠𝑖𝑛𝜃) and we have 𝑦 = 𝑎(1 − 𝑐𝑜𝑠𝜃). So, from here we will have =
𝑑𝜃
𝑑𝑦
𝑎(1 + 𝑐𝑜𝑠𝜃) and = 𝑎𝑠𝑖𝑛𝜃 . So, basically sine theta, is not it? So, that is what we get
𝑑𝜃

doing the differentiation and now we know our formula where we have to substitute all of
these.

283
(Refer Slide Time: 13:01)

So, the formula says that the required arc length AO𝐴′ equals to 2 times the arc length let
𝜋 𝑑𝑥 𝑑𝑦
us say OA. So, then in that case it is 2 ∫𝜃=0 √(𝑑𝜃)2 + (𝑑𝜃)2 𝑑𝜃. Now this is
𝜋
2 ∫𝜃=0 √𝑎2 (1 + 𝑐𝑜𝑠𝜃)2 + 𝑎2 𝑠𝑖𝑛2 𝜃𝑑𝜃. So, if we how to say if we calculate this whole
𝜋 𝜃
thing, then it will turn out to be 2𝑎 × 2 ∫𝜃=0 cos 𝑑𝜃. This involves some very simple
2

trigonometrical calculation and I am pretty sure you can be able to do that.

𝜋 𝜃
So, ultimately we will be able to obtain 4𝑎 ∫𝜃=0 cos 2 𝑑𝜃 and once we integrate this thing.
𝜃 𝜋
So, it will reduce to 4𝑎 × 2[sin 2]𝜋0 . So, when 𝜃 = 𝜋 , then in that case this sin 2 and when
𝜋 𝜋
𝜃 is 0 then sin0 is 0. So, 4𝑎 × 2 × sin 2 and then sin 2 is 1. So, 1; that means, 8a. So, that

is the required arc length from here to here although we calculated just from here to here
by varying the 𝜃 from 0 to 𝜋.

So, this is one way where we can calculate the arc length who is how to say polar Cartesian
a polar representation is given. Although it is not entirely polar it is actually parametric
Cartesian representation because the equations are given in 𝑥 = 𝜑(𝜃) and 𝑦 = 𝜓(𝜃). So,
instead of t they are using 𝜃 and the limits for 𝜃, we are calculating basically just looking
at the curve. So, it is not entirely a polar coordinate representation its actually a parametric
Cartesian representation. So, this is one such example where we can use that formula. Now

284
let us actually work out an example where we will use polar coordinate system. So, let us
go to an another example.

(Refer Slide Time: 16:03)

Example I don’t know 4. So, I lost the track. So, it could be example 4. So, find the
perimeter or arc length or rectify they are all the same thing

So, find the arc length or let us say perimeter. So, perimeter means actually the whole arc
length in a way. So, the entire arc length so, perimeter of the cardioid 𝑟 = 𝑎(1 − 𝑐𝑜𝑠𝜃).
So, the solution; so, how to draw the cardioid that is a different issue and that is a different
topic to be studied. So, I recommend to read some books where they have addressed this
issue that how you can draw a curve. Now the cardioid will look something like this, this
is our A that is O all right now this curve. So, this curve is symmetrical around the initial
line OX and therefore, the perimeter is double the length of therefore, its perimeter
therefore, its perimeter is double the length of the arc lying above the initial line.

So, we just calculate the arc length above the initial line and we multiply multiplied by 2
and that will give us the perimeter or the entire arc length of the cardioid of the initial line.
𝑑𝑟 𝑑𝑟
So, the given equation is 𝑟 = 𝑎(1 − 𝑐𝑜𝑠𝜃). So, from here we can calculate our 𝑑𝜃 and 𝑑𝜃 is

𝑎 sin 𝜃 its better to write some sentences. So, the given equation of the curve is this one
and then we do the differentiation alright. And 𝜃 is basically varying from 0 to 𝜋. So, 0 to

285
𝜋 to cover the entire arc length. Now the required arc length is perimeter. So, let us use the
term perimeter.

So, the required perimeter, equals to 2 times length of the upper arc. Now the length of the
𝜋 𝑑𝑟
upper arc will be calculated by 2 ∫𝜃=0 √𝑟 2 + (𝑑𝜃)2 𝑑𝜃. This is where we are using the polar

coordinate form for calculating the arc length. So, we have used that formula here all right.
Now we substitute these 2 values. So, let us substitute the values this will be 𝑟 2 .

(Refer Slide Time: 20:33)

𝜋
So, 2 ∫𝜃=0 √𝑎2 (1 − cos 𝜃)2 + 𝑎2 𝑠𝑖𝑛2 𝜃𝑑𝜃 .

𝜋 𝜃
So, from here we will basically obtain 2 ∫𝜃=0 2𝑎 sin 2 𝑑𝜃 . This is a very simple

trigonometric calculation which I am leaving up to the students 𝑑𝜃and then this 1 will be
𝜃 𝜋 𝜋
4𝑎[−2 cos 2] . So, this will be actually −8𝑎[cos 2 − cos 0].
0

𝜋
So, cos 2 is 0 and cos0 is 1. So, this is basically 8a. So, that is the required arc length or

the perimeter of this cardioid this one here. So, here the given equation was actually in
polar coordinate system and using this equation, we can be able to. So, using this equation
we were able to calculate this arc length here and multiply it by 2, that is here and that will
give you the whole parameter. So, that is basically the required perimeter of the cardioid.

286
So, we can work out an another problem, just to get more familiar with this topic actually.
So, let me consider an another example. so, find the arc length of the curve given by
cardioid 𝑟 = 𝑎(1 − cos 𝜃) and it is find the arc length of the curve given by cardioid and
it is divided by the line 4𝑟 cos 𝜃 = 3𝑎. So, this is the required. So, the cardioid basically
is divided by the line this one here and then the arc length we have to calculate the arc
length of a I mean when this line is dividing the cardioid.

(Refer Slide Time: 24:11)

So, if we draw this curve. So, unless we draw this curve we will not be able to understand
the problem. So, let us draw this curve. So, this is the curve. So, that is our origin and this
𝜋
is let us say CB and that is basically 𝜃 = 3 . So, the given curve and the straight line are

𝑟 = 𝑎(1 + cos 𝜃) and 4𝑟 cos 𝜃 = 3𝑎 respectively. So, basically this is the line and this is
the curve and they intersect one another at B and C.

So, from here we can therefore, therefore, the point of intersection can be calculated by
substituting r equals to. So, from here we can substitute the value of r here . So, that is
3𝑎
= 𝑎(1 + cos 𝜃) and from here we will basically obtain a quadratic equation
4 cos 𝜃

4𝑐𝑜𝑠 2 𝜃 + 4 cos 𝜃 − 3 = 0 and if we solve this equation, then we will basically obtain
1 1 𝜋
cos 𝜃 = 2 and if cos 𝜃 = 2 then 𝜃 is 3 .

287
 𝜋
So, they are intersecting at the point 𝜃 = and we have to calculate the value of r So, we
3
𝜋
can substitute for 𝜃 = here that will give us the value of r and another value of cos 𝜃 is
3
3
ignored because the other value is cos 𝜃 = − 2 we can write the other value.

(Refer Slide Time: 26:53)

3
The other valuecos 𝜃 = − 2 is inadmissible. So, we have ignored that value and for 𝜃 =
𝜋
we can also calculate the value of r. So, that is the point where they are intersecting one
3

another and the required arc length of the arc let us say. So, this is the arc for which we
have to calculate the arc length.

So, the required arc length ABC we have to write the correct order. So, CAB. So, CAB
equals to 2 times arc length what is the arc length then upper half because it is symmetrical.
So, the arc length AB. So, since it is symmetrical we just write 2 times arc length AB. So,
𝜋
𝑑𝑟 𝑑𝑟
this is basically 2 ∫𝜃=0
3 √𝑟 2 + ( )2 . So, we know that our 𝑟 2 = 𝑎2 (1 − cos 𝜃)2 , and is
𝑑𝜃 𝑑𝜃

𝑑𝑟
very simple to calculate from here. So, 𝑑𝜃 is sorry.

𝑑𝑟
So, here it is minus. So, is basically 𝑎 sin 𝜃. So, this is
𝑑𝜃
𝜋
2 ∫𝜃=0
3 √𝑎2 (1 − cos 𝜃)2 + 𝑎2 𝑠𝑖𝑛2 𝜃𝑑𝜃 and if you do the simplification then we will
𝜋
𝜃
basically obtain here 2 ∫𝜃=0
3 2𝑎 sin 2 𝑑𝜃. And we basically do the integration like we did

288
before and then we will get our required answer. So, here we saw that we had to calculate
the arc length of the curve given by the cardioid and it is divided by the line.

So; that means, this part of the cardioid this part of the cardioid or this part of the arc we
needed to calculate since it is symmetrical. So, we just calculate the one side and multiply
it by 2. So, that is what we are doing here and here its fairly just its pretty much just a
simple calculation like we did before and that will give us the required arc length between
these 2 curves.

So, I tried to cover enough examples from all of these type of coordinate system Cartesian,
parametric Cartesian and polar coordinate system and we will continue with our next topic,
which is basically surface integral in our next class and I will try to include some examples
on rectification in our assignment sheet and I look forward to you in your next class.

Thank you.

289
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 28
Surface Integral

Hello students. So, in the last class we practiced rectification of curves which basically
means that, we calculated the length of the arc and we also saw that we can calculate the
length of the arc for different types of course; in given in different forms actually. So,
either you will be given 𝑦 = 𝑓(𝑥) form or 𝑥 = 𝑓(𝑦) form or 𝑟 = 𝑓(𝜃). So, whatever the
form maybe we can implement that particular type of formula, to calculate the length of
the arc between certain point to a certain point.

So, those kind of examples we practiced and we also saw that in what situation what kind
of formula we can use. So, today we will start with basically an application side of integral
calculus. So, when I say application, it basically means that we will be doing some line
integral surface integral and the volume integral. So, that is the final or maybe 1 chapter
before the final chapter. So, the final chapter is basically calculating the moment of inertia
and things like that, but before that we will practice this line integral, volume integral and
surface integral problems.

So, today we will start with the line integral. This is also motivated from the vector calculus
as well. So, when we go through the topics on vector calculus, we will come across the
similar topics where there also we will practice line integral, surface integral and volume
integral, but that will be more towards the vector calculus part. And today and probably in
the next lecture we will look into more from the integral calculus perspective.

290
(Refer Slide Time: 02:27)

So, let us start with line integral. So, line integrals basically mean that a curve. So, we give
a small definition. So, a curve in 𝑅 2 is the set of points 𝑐 = {(𝑥, 𝑦): 𝑥 = 𝜑(𝑡) 𝑎𝑛𝑑 𝑦 =
𝜓(𝑡)𝑤ℎ𝑒𝑟𝑒 𝑎 ≤ 𝑡 ≤ 𝑏}. So, for example, circle is a curve in 𝑅 2 and its equation can be
given by 𝑥 = cos 𝑡 and 𝑦 = sin 𝑡 for t running between 0 to 2𝜋.

So, that circle is a curve which can be expressed in this form and we say that if we assume
that let us say and if we assume that, the function f,𝜑 and 𝜓 are continuous and 𝜑 possesses
. 𝑏
a continuous derivative 𝜑 ′ (𝑡), then integral ∫𝑐 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦 = ∫𝑎 𝑓(𝜑(𝑡), 𝜓(𝑡))𝜑 ′ (𝑡)𝑑𝑡

So; that means, basically what we are doing is, we have a curve let us say something like
this and we are basically calculating this integral, integral of this function along this curve
c. So, the name this curve is basically our curve c. So, here in this definition of the curve,
we can write 𝑥 = 𝜑(𝑡) and 𝑦 = 𝜓(𝑡) for a certain values of t and we replace the x and y
in this integral on the left hand side with 𝜑(𝑡) and 𝜓(𝑡) and this here is basically the range
for t, then this integral is equal to the integral of this function here in of course, parameter
t and t is varying from a to b.

So, that is what we basically mean by the calculation of the integral of a function f along
this curve c from let us say point t equals to a to, from a point a to point b. So, we have a
point certain point here let us say here x equals to some point, and then this will be given
by that [𝑎, 𝑏] to [𝑐, 𝑑] then in that case, the integral of the curve, integral of the function f

291
along this curve from [𝑎, 𝑏] to [𝑐, 𝑑] will be given by the formula on the right hand side,
where we have to express now, the equation of the curve in terms of a parametric equation.

So, that is the main task I would say here is to find the parametric equation of the curve
and based on that we can calculate the limit, from where to where the t is varying and then
just substitute on the function and then we can calculate the right hand side, and that will
give us the required how to say line integral of the function f. And next this is somehow
also how to say motivated from the vector calculus sense, but we will work on that part
when we actually look into the vector calculus part. So, how do we calculate these type of
integrals? So, the integral calculation can be done in the following way.

(Refer Slide Time: 07:21)

. 𝑑𝑥
So, let us start with our very first example. So, this example says evaluate integral ∫𝑐 𝑥+𝑦,

where c is the curve and here we have only dx actually. So, here we will have only d x and
.
yes. So, we do not need y actually. So, integral ∫𝑐 𝑓(𝑥, 𝑦)𝑑𝑥𝑑𝑦 =
𝑏
∫𝑎 𝑓(𝜑(𝑡), 𝜓(𝑡))𝜑 ′ (𝑡)𝑑𝑡 here.

. 𝑑𝑥
And now ∫𝑐 𝑥+𝑦 where c is the curve 𝑥 = 𝑎𝑡 2 and 𝑦 = 2𝑎𝑡 where 0 ≤ 𝑡 ≤ 2. So, we

already have the parametric representation of the curve and for that we do not need actually
the calculation to calculate the parametric representation of a certain curve. This curve is
basically as we can see its a parabola because if we do 𝑦 2 . So, since 𝑦 2 = 4𝑎2 𝑡 2 .

292
So, this can be written as 4𝑎𝑥. So, the given curve is a parabola and since it is of standard
shape. So, we do not need to draw this parabola as well. Now we have our given function
1 1
f is 𝑓(𝑥, 𝑦) = 𝑥+𝑦. So, this will be actually and if I follow that formula. So, we
𝑎𝑡 2 +2𝑎𝑡

need to calculate also the 𝜑 ′ (𝑡), in this sense our 𝜑(𝑡) = 𝑎𝑡 2 and 𝜓(𝑡) = 2𝑎𝑡. So, that is
𝜑(𝑡) and 𝜓(𝑡) the 2 functions.

So, here 𝜑(𝑡) is basically 𝑎𝑡 2 . So, from here we will have 𝜑 ′ (𝑡) is basically 2𝑎𝑡and
. 𝑑𝑥 2 2𝑎𝑡𝑑𝑡
therefore, our required integral ∫𝑐 𝑥+𝑦 = ∫𝑡=0 𝑎𝑡 2 +2𝑎𝑡. Now, this will be basically integral
2 𝑑𝑡
2 ∫𝑡=0 𝑡+2.

So, once we integrate it will be 2[log 𝑒 (𝑡 + 2)]2𝑡=0 . So, when we substitute t equals to 2
then it will be log4, and when we substitute t is equals to 0 then it will be log2 and if we
do some simple calculation, then it will lead to 2 log 𝑒 2. So, this is the required answer or
the evaluation of this integral along this curve c. So, that is what I mean by line integrals.

So, we are actually integrating along a curve. So, sort of like following a kind of like a
line, it is not entirely a line, but the name is somehow motivated about following a certain
curve, along a certain curve if I am walking along a certain curve, then what will be the
integral of a function along that curve from a certain point to a certain point. So, that is
what we mean by line integral and this is what we are doing here. So, here in this case we
have a function and we need to calculate its integral along this parabola from the point 𝑡 =
0 to 𝑡 = 2.

So, if I substitute t equals to 0; that means, (0,0) so; that means, we are calculating the
integral from the origin and when I substitute t equals to 2, then its 4a so; that means, we
need to calculate the integral from (0,0) to (4a,4a), I hope it is 4a. So, yeah 4a to 4a along
that curve. So, that is what we mean by line integral here. We will work out few more
examples just to make the concept a little bit clear. So, let us proceed.

293
(Refer Slide Time: 12:49)

.
So, here we have evaluate integral ∫𝑐(𝑥 2 𝑑𝑥 + 𝑥𝑦𝑑𝑦) right taken along the line segment
line segment from (1,0) to (0,1). So, we have to calculate the line integral of this curve
taken along the line segment (1,0) and (0,1). So, first of all let us draw this curve. So, the
equation; so, the line segment is (1,0) and another to the point (0,1) so; that means, if I join
these two points let me call it as A and this 1 is B. So, we are basically calculating the line
integral along this line segment.

So, the line segment or the straight line; joining (1,0) and (0,1) is given by, what will be
the equation? Is given by 𝑥 + 𝑦 = 1 so; that means, this is the equation of this straight line.
So, if I put 𝑦 = 𝑡 then in that case x will be (1 − 𝑡). So, the parametric equation or we do
not have to do the parametric equation here, when we are walking along this line or let us
do it anyways the parametric equation is let us say when x is t. So, when 𝑥 = 𝑡 then y will
be (1 − 𝑡)alright where 0 less or equal to.

So, where t is; so, it would be better if we take this one the calculation point of view, you
would understand what I am trying to say. So, if I take 𝑦 = 𝑡 then 𝑥 = (1 − 𝑡)where 0 ≤
𝑡 ≤ 1 so; that means, when t is 0 the parametric the point is (1,0). So, this point and when
t is 1 then its (0,1). So, this point and all the other points for every value of t will lie on
this a straight line. So, that is the parametric representation.

294
Now, let us go to this curve. So, here we have 𝑥 = (1 − 𝑡) and 𝑦 = 𝑡So, that is the
parametric equation. So, to be very precise we can write it, this is not the equation this is
just a coordinate. So, this is our parametric equation if we write it correctly. So, that is
basically the parametric equation, the earlier what we have written was just a point. So,
.
this is the parametric equation. Now we will integrate ∫𝑐(𝑥 2 𝑑𝑥 + 𝑥𝑦𝑑𝑦).

. .
So, first of all we will integrate ∫𝑐 𝑥 2 𝑑𝑥 + ∫𝑐 𝑥𝑦𝑑𝑦 alright. So, we can write it as integral
. 𝑑𝑥 . 𝑑𝑦 𝑑𝑥 𝑑𝑥 𝑑𝑦
∫𝑐 𝑥 2 𝑑𝑡 𝑑𝑡 + ∫𝑐 𝑥𝑦 𝑑𝑡 𝑑𝑡. Now what is 𝑑𝑡
? 𝑑𝑡
would be −1 and 𝑑𝑡
will be 1. So, let us
𝑑𝑦 𝑑𝑥
substitute the value of 𝑥 2 value of xy and and .
𝑑𝑡 𝑑𝑡

(Refer Slide Time: 18:03)

1 𝑑𝑥
So, this will reduce to ∫𝑡=0(−(1 − 𝑡)2 𝑑𝑡 + (1 − 𝑡)𝑡. 1𝑑𝑡) right. So, 𝑥 2 and then sorry.
𝑑𝑡
𝑑𝑥
So, 𝑑𝑡 is −1. So, let us write that −1. So, there will be a minus here and then dt then plus
𝑑𝑦
xy. So, (1 − 𝑡)𝑡. So, let us write (1 − 𝑡)𝑡 and then is 1 and then again dt. So, this will
𝑑𝑡
1 1
be integral ∫𝑡=0(−1 + 2𝑡 − 𝑡 2 + 𝑡 − 𝑡 2 )𝑑𝑡 . So, this will be ∫𝑡=0(−2𝑡 2 + 3𝑡 − 1)𝑑𝑡.

So, from here we can do the rest of the calculation. So, you see we substituted the value of
x which is (1 − 𝑡). So, that is what we did and x times y so, y is t. So, (1 − 𝑡)𝑡 and then
𝑑𝑦 𝑑𝑥
𝑑𝑡
is again 1 alright. So, we substituted the value. So, this will be 𝑥 2 is d x. So, 𝑑𝑡
is −1.

295
So,−𝑥 2 ; so, this is −(1 − 𝑡)2 𝑑𝑡 + (1 − 𝑡)𝑡. 1𝑑𝑡 and then we just break this formula. So,
its −1 + 2𝑡 − 𝑡 2 and then this one is +𝑡 − 𝑡 2

So, if we integrate this whole thing then we will obtain basically whatever. So, that will
be the required answer. So, this is how we do the line integral for this type of problem,
similarly if we want we can also have this curve as. So, if instead of this line segment we
can also be given let us say some circle. So, taken along a circle 𝑥 = cos 𝑡 instead of a line
segment, let me say it is a circle 𝑥 = cos 𝑡 and 𝑦 = sin 𝑡 where t is between 0 to 𝜋 let say.
So, then in that case you have to substitute 𝑥 = cos 𝑡 and 𝑦 = sin 𝑡here and then do the
similar thing.

So, it just playing with what kind of curve that is given to us. So, its not that complicated
alright we will work out maybe 1 or 2 more examples. So, let us do that.

(Refer Slide Time: 21:27)

.
So, here we have example 3rd, show that integral 𝐼 = ∫𝑐[(𝑥 − 𝑦)2 𝑑𝑥 + (𝑥 −, and let us
take a simpler examples instead of going straight to the complicated one, we can start with
another simple example.

.
So, this one integral ∫𝑐(𝑦𝑑𝑥 − 𝑥𝑑𝑦); the reason why I am considering this example is to
show you a slight trick here and also the statement of the problem. So, that is why I am
choosing this example because its a little bit interesting compared to the traditional

296
examples. So, where c is the ellipse 𝑥 = 𝑎 cos 𝑡 and 𝑦 = 𝑏 sin 𝑡 taken in the clockwise
direction. So, when we say clockwise direction; that means, the seconds key in the clock
it moves or the minute key they move like; they move in this fashion.

So, that is our clockwise direction and if it rotates in the in the opposite direction then its
called the anti clockwise direction. So, that means, t is actually instead of moving from 0
to 2𝜋 it is moving from 2𝜋 to 0 and having this particular thing written here anti clockwise
or clockwise and its plays a very important role. So, you have to pay very close attention
what is the range I mean in which range t is varying, whether it is 0 to 1 or 1 to 0? So, I
mean of course, when we say that 1 to 0 theoretically or conceptually it does not make
sense because it will always go from 0 to 1 not from 1 to 0, but it does happen and from
and when we actually use those limits in your integral you may have to play with the
integral to correct that limit or; that means, you can do 0 to 1 at some point.

Because you might get a minus of integral and then you can use the formula which is
𝑏 𝑎
− ∫𝑎 𝑓(𝑥)𝑑𝑥 = ∫𝑏 𝑓(𝑥)𝑑𝑥. So, some kind of integral calculus formula implementation
might be there, but always you have to pay very close attention to this statement, whether
it says clockwise or whether it says anti clockwise. So, here in this case it says clockwise
and that was the intention of taking this example.

Now, we will see how we can solve this line integral. So, to solve this we just write 𝐼 =
. 𝑑𝑥
∫𝑐(𝑦𝑑𝑥 − 𝑥𝑑𝑦) and x is given. So, we can easily calculate 𝑑𝑡
, we just write
. 𝑑𝑥 𝑑𝑦 0
∫𝑐 (𝑦 𝑑𝑡 − 𝑥 𝑑𝑡 ) 𝑑𝑡 the limit first ∫𝑡=2𝜋(𝑏 sin 𝑡 × −𝑎𝑠𝑖𝑛𝑡 − 𝑎 cos 𝑡 𝑏 sin 𝑡)𝑑𝑡.

So, what we have done is basically we have done something like this alright. So, that is
0
what we are doing at the moment and then it is integral −𝑎𝑏 ∫𝑡=2𝜋(𝑠𝑖𝑛2 𝑡 + 𝑐𝑜𝑠 2 𝑡)𝑑𝑡. Now
as I was saying we can use some integral calculus formula to revert the limit. So, its
2𝜋
𝑎𝑏 ∫0 𝑑𝑡 and this is the kind of correction I was talking about. So, although you had
higher value to lower value this in the limits at some point you might be able to correct
that and this is what we will get. So, the answer is 2𝑎𝑏𝜋.

So, that is actually the line integral of this function or of this integrand along this ellipse
alright. So, this is what we mean here now we can consider a complicated example,
afterwards we will close this line integral chapter.

297
(Refer Slide Time: 27:04)

.
So, the fourth example is, evaluate integral ∫𝑐[(𝑥 + 𝑦 2 )𝑑𝑥 + (𝑥 2 − 𝑦)𝑑𝑦] taken in the
clockwise direction, along a closed curve 𝑦 3 = 𝑥 2 and the line from (0,0) to (1,1).

So, first of all we have a closed curve between this and this closed curve and the line this.
So, let us first draw the curve. So, first of all we have a closed curve between this curve
and the line. So, let us draw the line. So, here somewhere we can have the point (1,1) and
𝑦 3 = 𝑥 2 this curve will be given by something like this. So, you can look into any book
how to draw this curve. So, this is our 𝑦 3 = 𝑥 2 and this is our 𝑦 = 𝑥 and since we are in
the clockwise direction. So, if we are in the clockwise direction; that means, we are moving
from this direction to this direction and this is our closed curve.

So, the curve c consists of the arc OA. So, let us call this point OA and the line AO. So,
now, that we have our curve this is our curve c, we will stop here for this problem, we will
continue with a similar problem in our next class, where we see how we can evaluate this
integral and we will. So, this will involve a slight how to say kind of a trick or conceptual
knowledge in a way. So, we will see that in our next class and will stop here today.

Thank you.

298
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 29
Surface Integral (Contd.)

Hello students. So, in the last class we started with the line integral and today we will
continue with probably with our last example which we started and then, we switch to a
Surface Integral.

(Refer Slide Time: 00:33)

So, in the last class we started with this example where we needed to evaluate this integral
along the closed curve given by the line from (0,0) to (1,1). So, we can actually draw the
figure and the line from (0,0) to (1,1) is a straight line passing through the origin. So, this
is our line and the curve is given in this fashion and they intersect of course at the point (1,
1).

So, if we substitute 𝑦 = 𝑥 here, then it will be basically 𝑥 3 − 𝑥 2 = 0. So, from there we

can be able to calculate the point of intersection as 0 and 1 and when x is 0, then y is 0 and
when x is 1, then y is 1. So, these are the two points where they intersect. Now, that we
have the curve and the we have the line, so we basically have our closed curve and now
we do that calculation. So, how can we do that? So, of course here the line integral can be

299
done in two steps. So, first we do the line integral along this curve and then, we do the line
.
integral along this curve. So, let us see. So, we have ∫𝑐[(𝑥 + 𝑦 2 )𝑑𝑥 + (𝑥 2 − 𝑦)𝑑𝑦]. So,
.
we can write this one as the line integral ∫𝑎𝑙𝑜𝑛𝑔 𝑡ℎ𝑒 𝑙𝑖𝑛𝑒 𝐴𝑂[(𝑥 + 𝑦 2 )𝑑𝑥 + (𝑥 2 − 𝑦)𝑑𝑦] +
.
∫𝑎𝑙𝑜𝑛𝑔 𝑡ℎ𝑒 𝑎𝑟𝑐 𝑂𝐴[(𝑥 + 𝑦 2 )𝑑𝑥 + (𝑥 2 − 𝑦)𝑑𝑦]right because we have to divide it into two

segments. Now, along the line 𝑦 = 𝑥 I can substitute 𝑦 = 𝑥 here and x will vary from
basically 0 to 1. So, let us do that.

So, I can do that along the line AO I can do 1, so, x is varying from 1 to 0.

(Refer Slide Time: 03:11)

4 1
0 1 2
So, we can write it as ∫1 [(𝑥 + 𝑥 2 )𝑑𝑥 + (𝑥 2 − 𝑥)𝑑𝑥] + ∫0 [(𝑥 + 𝑥 3 ) (3 𝑥 3 ) 𝑑𝑥 +
2
(𝑥 2 − 𝑥 3 ) 𝑑𝑥

and if we integrate the whole thing, so of course integrating would not be that much
difficult because we basically have an algebraic expression and I am pretty sure you can
be you all can be able to do that. So, we will avoid that doing that here. So, if you do the
1
whole calculation, then you will ultimately obtain 84, so, this is the required answer. Now,

this example is in particular interesting because here we have to divide the line integrals
into two segments basically.

300
So, first we integrated with respect to this line segment and then, we integrated with respect
to this arc here and the sum will actually give you the integral along that whole closed
curve because they are closed and that is basically our answer is. So, this is how we
calculate the line integral. So, in case of line integral. You always have to identify the
curve and its parametric representation or the range for the variable x and y and based on
that which either in this case in the previous example we did not have to put the parametric
representation.

We managed with the x and y representation itself, but if you need to put the parametric
representation of the curve, we can also do that. We saw those one or two examples like
that and then, you integrate with respect to the parameter t and use the limits of the
parameter t, otherwise you can also integrate with respect to x and y and divide the given
a closed curve. If it is composed of two different segments, then you divide the closed
curve into two segments and then, you do the integration like the previous example.

So, the line integrals are not that difficult and they are also very interesting. So, I am hoping
you would how to say find it interesting when you practice it by your own and we will also
try to include some formulas or assignments, basically not formulas assignments in your
assignment sheet for you to practice, all right. So, more or less most of the examples are
of similar types.

So, we will skip these examples and now I am in my lecture note. Now, I am going to start
with our next topic which is basically surface integral and in surface integral here will
evaluate some surface integral obviously and also we will evaluate some volume integrals.
In our vector calculus part we will do the same thing actually. So, in vector calculus part
also we will evaluate surface and volume integral and of course, those will be motivated
from the vector calculus point of view.

So, there we will say Gauss Divergence theorem, Stokes theorem, Greens theorem and
here we will stick to our traditional surface and volume integrals. If time permits we might
look into the Gauss Divergence theorem or Stokes theorem in the integral calculus itself.
If not then we will cover those topics in our vector calculus part which will I am hoping to
start it next one or two lectures after actually.

301
So, let us start with the surface integral. So, surface in 𝑅 3 , so a curve or a surface in 𝑅 3 is
a vector valued function whose domain is a subset of 𝑅 2 and the range is a subset of 𝑅 3 ,
all right.

So, if 𝑓, 𝑔, ℎ are three real valued functions defined on a domain D subset of 𝑅 2 such that
𝑥 = 𝑓(𝑢, 𝑣), 𝑦 = 𝑔(𝑢, 𝑣) and ℎ = 𝑧(𝑢, 𝑣).

(Refer Slide Time: 10:15)

So, then the set then the set (𝑓(𝑢, 𝑣), 𝑔(𝑢, 𝑣), ℎ(𝑢, 𝑣)) is a surface in 𝑅 3 . So, all of this, so
this will give the point x, this will give the point y and this will give the point cell. So, the
collection of all these points will actually represent a surface in 𝑅 3 and we will actually do
the integration on these surfaces. So, let us see how we can do that and in our case whatever
surface we work with, they are always smooth.

So, that means all of them will have a continuous partial derivative so with respect to these
variables actually. So, now how can we calculate the surface area of a surface.So, the
formula is the area of a smooth surface, so, that means they have continuous partial
derivatives 𝑥 = 𝑓(𝑢, 𝑣), 𝑦 = 𝑔(𝑢, 𝑣) and set 𝑧 = ℎ(𝑢, 𝑣) is defined as
. 𝜕(𝑦,𝑧) 𝜕(𝑧,𝑥) 𝜕(𝑥,𝑦)
∬𝐷 √[𝜕(𝑢,𝑣)]2 + [𝜕(𝑢,𝑣)]2 + [𝜕(𝑢,𝑣)]2 if we simplify this formula. So, if we consider the

surface. So, if we let us say so these are basically the Jacobians, but we have to simplify
this formula.

302
Now, if we write the surface as 𝑧 = 𝑓(𝑥, 𝑦). So, if this is the given surface or we can write
𝑥 = 𝑢, 𝑦 = 𝑣 and then 𝑧 = 𝑓(𝑢, 𝑣) so again 𝑥 = ℎ(𝑢, 𝑣), 𝑦 = 𝑔(𝑢, 𝑣) and 𝑧 = 𝑓(𝑢, 𝑣),
then in that case we can calculate these individual Jacobians. So, calculating these
individual Jacobians will not be difficult, we have already done the calculation. So, this
𝜕(𝑦,𝑧)
for example, is 𝜕(𝑢,𝑣), then, we do the calculation of that determinant that will give us the

first Jacobian, then, similarly the 2nd Jacobian, similarly the 3rd Jacobian.

So, if we do all those calculation, then we will basically obtain let us say if we will basically
obtain the surface integral which I am going to write as.

(Refer Slide Time: 14:41)

. 𝜕𝑧 𝜕𝑧
𝐼𝑠 = ∬𝐷 √1 + (𝜕𝑥)2 + (𝜕𝑦)2, it is a formula. where D is the projection of the surface on

xy plane, right and if I do the projection on let us say zx plane, so then in that case this will
𝜕𝑧
be .
𝜕𝑥

So, this is the projection on del z plane. Now, if the given surface is 𝑥 = 𝑔(𝑦, 𝑧), then
again I can substitute 𝑦 = 𝑢 and 𝑧 = 𝑣, then it will be 𝑥 = 𝑔(𝑢, 𝑣)and therefore, the
. 𝜕𝑥 𝜕𝑥
required form will be ∬𝐷 √1 + (𝜕𝑧 )2 + (𝜕𝑦)2 𝑑𝑥𝑑𝑦where D is the projection on yz plane

. 𝜕𝑦 𝜕𝑦
and similarly if the surface is let us say 𝑦 = ℎ(𝑧, 𝑥), then 𝐼𝑠 = ∬𝐷 √1 + (𝜕𝑥 )2 + ( 𝜕𝑧 )2,

where D is the projection is the projection on zx plane.

303
So, here it also depends on in what form you are given the equation. So, if we have the
given surface as 𝑥 = 𝑓(𝑦, 𝑧), then we basically. So, if the given formula is 𝑧 = 𝑓(𝑥, 𝑦),
then we basically obtain the first formula and then, we use the first formula. Actually if
we are given the equation of the curve as 𝑥 = 𝑔(𝑦, 𝑧), then in that case we will use the 2nd
formula and if it is given 𝑦 = ℎ(𝑧, 𝑥), then we use the 3rd formula. So, it also depends on
the fact that in which form you are given the surface and based on that we can use either
one of these formulas. So, I think it was a little bit abstract. So, let us start with the first
example.

(Refer Slide Time: 18:27)

So, example 1; Let us find the surface area of a curve which is or a surface which is already
known to us. So, find the surface area of the surface 𝑥 2 + 𝑦 2 + 𝑧 2 = 𝑎2 . Now, just looking
at this curve or at the surface we can easily say what kind of surface it is. It is our sphere,
all right. So, we know that. So, instead of surface I can write now sphere now that we all
know. Now, we know that surface area of a sphere is 4𝜋𝑎2 . So, let us see whether we can
be able to obtain using these integral calculus formula or not.

So, let us start. So, we have 𝑥 2 + 𝑦 2 + 𝑧 2 = 𝑎2 . So, from here I can be able to write 𝑧 =
√𝑎2 − 𝑥 2 − 𝑦 2 , right. I am not taking the negative value, so, let us take only the positive
value of z, right. So, this is basically my 𝑓(𝑥, 𝑦) isn't it and now if I have the formula in
terms of 𝑧 = 𝑓(𝑥, 𝑦), then we can be able to obtain the 1st formula for the surface integral.

304
 𝜕𝑧 𝜕𝑦
So, for the 1st formula I need to calculate where is that and there is . So, let us
𝜕𝑥 𝜕𝑥
𝜕𝑧 𝜕𝑧 −2𝑥
calculate 𝜕𝑥 . So, this will be 𝜕𝑥 = .
2√𝑎2 −𝑥 2 −𝑦 2

So, this is basically 2x and there is a 2 here, so, 2 2 will get cancelled. So, this is
𝜕𝑧 −𝑥 𝜕𝑧 −2𝑦
basically = . Similarly I can calculate 𝜕𝑦 = . So, this is
𝜕𝑥 √𝑎2 −𝑥 2 −𝑦 2 2√𝑎2 −𝑥2 −𝑦 2
−𝑦
, right. So, now based on this formula; based on this formula I can be able to
√𝑎2 −𝑥2 −𝑦 2
𝜕𝑧 𝜕𝑧
calculate 1 + (𝜕𝑥)2 + (𝜕𝑦)2.

𝑥2 𝑦2
So, this is basically 1 + 𝑎2 −𝑥 2 −𝑦 2 + 𝑎2 −𝑥2 −𝑦 2. So, if I take this if I calculate this, then this
𝑎2
will be 𝑎2 −𝑥 2 −𝑦 2 .

(Refer Slide Time: 21:35)

So, now the surface area of the sphere is twice the surface area of the upper half, right and
further the projection of this sphere on xy plane is, right. So, we have taken the upper half
because the surface area of the whole sphere would be just the double of the surface area
of the upper half and if we do the projection on xy plane, then it will basically be a circle.

So, initially it is an upper half of this sphere, but once we do the projection then in that
case it is basically a circle which is given by this equation. So, the radius will remain same,
it is not like if we do the projection, then the radius would change. So, the radius would

305
still remain same and this is the equation of our circle. Now, our required surface area will
be two times R is the projection 𝑥 2 + 𝑦 2 = 𝑎2 . So, let us write 𝐼𝑠 =
. 𝜕𝑧 𝜕𝑧 . 𝑎
2 ∬𝑅 √1 + (𝜕𝑥)2 + (𝜕𝑦)2 𝑑𝑥𝑑𝑦. So, this is basically 2 ∬𝑅 𝑑𝑥𝑑𝑦.
√𝑎2 −𝑥2 −𝑦 2

Now, this is just our plane integral calculus for the function of two variables. This is we
have already done. So, this type of integral calculation for the function of two variable we
have already done. So, here we basically take help of the how to say polar coordinates
actually.

(Refer Slide Time: 24:15)

So, let us substitute putting 𝑥 = 𝑟 cos 𝜃 and 𝑦 = 𝑟 sin 𝜃, where 0 ≤ 𝑟 ≤ 𝑎 and 0 ≤ 𝜃 ≤

𝑎 2𝜋 𝑎
2𝜋, then the surface integral 𝐼𝑠 = 2 ∫𝑟=0 ∫𝜃=0 √𝑎2 |𝐽|𝑑𝑟𝑑𝜃.
−𝑟 2 𝑐𝑜𝑠2 𝜃−𝑟 2 𝑠𝑖𝑛2 𝜃

𝑎 2𝜋 𝑟𝑑𝑟𝑑𝜃
So, this will be integral 2𝑎 ∫𝑟=0 ∫𝜃=0 √𝑎2 . We have already seen it many times how we
−𝑟 2

can calculate the Jacobian for this transformation and therefore, we will end up with
2𝜋 𝑎 𝑟𝑑𝑟
2𝑎 ∫𝜃=0 𝑑𝜃 ∫𝑟=0 √𝑎2 . So, now we separate 𝜃 and r because they are not mixed up
−𝑟 2

together.

306
(Refer Slide Time: 26:07)

2𝜋 𝑎 𝑟𝑑𝑟
So, 2𝑎 ∫𝜃=0 𝑑𝜃 ∫𝑟=0 √𝑎2 . Now, I substitute 𝑎2 − 𝑟 2 = 𝑧 and then, in that case we will
−𝑟 2

have −2𝑟𝑑𝑟 = 𝑑𝑧 and that can be handled very easily. So, calculating this integral would
not be difficult. So, this will be 2𝑎 × 2𝜋[−√𝑎2 − 𝑟 2 ]𝑎𝑟=0. So, this is basically 4𝑎𝜋[−0 +
𝑎]. So, that means 4𝜋𝑎2 . So, this is the required surface area of this sphere. So, where is
that problem?

So, calculating the surface area of this sphere we tried to express the variable z as a function
𝜕𝑧 𝜕𝑧
of x and y. So, now we are in that 𝑧 = 𝑓(𝑥, 𝑦) form. So, for that we need and 𝜕𝑦. So,
𝜕𝑥

we substituted the value here and then, that is our this square and then, in the formula also
we need to take the projection. So, before doing the projection we first need to find out the
surface area. How I mean what kind of concept do we need to use? So, since it is a sphere
the surface area of the whole sphere be would be twice the surface area of the upper half.

So, that is what we have done and R is the projection. So, when we take the projection of
the upper half on xy plane, it will be a circle with similar radius. So, that is what we have
written and then, we are just using this how does a conversion into a polar coordinate
system. So, this is 𝑥 = 𝑟 cos 𝜃 and 𝑦 = 𝑟 sin 𝜃, where 0 ≤ 𝑟 ≤ 𝑎 and 0 ≤ 𝜃 ≤ 2𝜋 and
then, we just substitute the value, do some simple calculation that will give us the required
integral as 4𝜋𝑎2 .

307
So, here in this example we saw that how we can calculate the surface area and that
calculation match with our traditional how to say a concept of surface area of a sphere. So,
we will definitely work out one or two more examples on calculating the surface area of a
surface. We will probably see some examples where we have two intersecting surfaces
and how we can calculate the surface area. We will do all those things in our next class
and then, we will switch to volume integral. So, I thank you for your attention and I will
see you in the next class.

308
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 30
Surface Integral (Contd.)

Hello students. So, in the last class we were looking into the Concepts of Surface Integral
and we also started with one two examples. Since, it is one of the important topics in
integral calculus, we will practice maybe a couple of more examples on surface integral
and then we move to the volume integral.

Of course, surface integral has two aspects basically. So, one can look to them from the
integral calculus point of view and also from the vector calculus point of view. So, we will
also work out a few surface integrals when we look into vector calculus part, but right now
we are more analyzing it from the integral calculus point of view.

(Refer Slide Time: 01:00)

So, let us start with our first example for today. So, our first example states that. So,
example 1,find the surface area of that part of the surface of the cylinder 𝑥 2 + 𝑦 2 =
𝑎2 which is cut out by the cylinder 𝑥 2 + 𝑧 2 = 𝑎2 .

So, basically we have to find out the surface area of the first cylinder cut off by the second
cylinder. So, the first cylinder has the base in xy plane and the second cylinder has the

309
base in xz plane. So, of course, one cylinder is like this, another one is like this, so they
will intersect somewhere and we have to find out the area which is basically how to say
intersected or cut off by these two cylinders.

So, this is our given cylinder; so, the given surface is basically this one here. So, let us
write the equation of the given surface is 𝑦 = √𝑎2 − 𝑥 2 . So, I take 𝑥 2 on the other side
and we can calculate. So, the surface is basically this base line in xy plane. So, basically
𝜕𝑦 𝜕𝑦 𝜕𝑦 −2𝑥
we will do 1 + (𝜕𝑥 )2 + ( 𝜕𝑧 )2. So, will be , this one is 02 . So, this will be 1 +
𝜕𝑥 2√𝑎2 −𝑥 2
−2𝑥 𝑥2 𝑎2
[ ]2 + 02 . So, this will ultimately end 1 + 𝑎2 −𝑥 2 = 𝑎2 −𝑥 2 . So, we will get this value
2√𝑎2 −𝑥2

equals to this thing.

(Refer Slide Time: 04:01)

Now, we see that our given surface is symmetrical about the plane 𝑦 = 0. So, of course, it
is symmetrical because we have 𝑥 2 + 𝑦 2 = 𝑎2 . So, it does not matter whatever value we
put for x, we will get always how to say the same value for y and that means, that the
cylinder is lying on the both half of the plane 𝑦 = 0. So, it is symmetrical on the both sides
of the plane 𝑦 = 0 that means, along x axis, along x z actually. So, along the plane basically
𝑦 = 0.

310
And now, that we have the symmetrical part, hence the required surface area, so we will
. 𝜕𝑦 𝜕𝑦
take the projection on xz plane, all right surface area is 𝑆 = 2 ∬𝑅 √1 + (𝜕𝑥 )2 + ( 𝜕𝑧 )2. So,

when we take the projection on xz plane.

Now we will have here a 2, because this the plane is symmetrical along the y. So, we have
to just calculate the surface area above the xz plane and just multiplied by 2 and that will
give us the whole surface area because of this symmetric condition. So, we have 2 here
and if I take the square root so over the region R I will have a and then in the denominator,
. 𝑎
I will have 2 ∬𝑅 √𝑎2 𝑑𝑥𝑑𝑧.
−𝑥 2

Now, the projection on the xz plane will be projection on the xz plane for this cylinder
. 𝑑𝑥𝑑𝑧
here is this one actually and therefore, we can write 2𝑎 ∬𝑅 √𝑎2 and our region R is we
−𝑥 2

will just write the equation. Region or surface whatever you would like to call. So, our
surface R or region R is this one.

(Refer Slide Time: 07:05)

𝑎 𝑑𝑥 √𝑎2 −𝑥 2
And this can be written as integral 2𝑎 × 4 ∫0 √𝑎2 −𝑥 2
∫𝑧=0 𝑑𝑧, positive quadrant will be

used and then we put a 4 here basically. So, we put a 4 here. Because now the projection
is taken in the xz plane, and xz plane what will happen is it is basically a circle and since
circle is in a way symmetrical along x and z axis then in that case we do not have to

311
calculate the how to say the surface area along the whole circle we just take one of the
quadrants and then multiplied by 4. So, that is what we are doing.

We, instead of calculating over the whole 𝑥 2 + 𝑧 2 = 𝑎2 we just calculate on the first
quadrant and then we multiply the whole thing by 4 and that will give you the whole
surface area. So, this is also convenient because now we are only focused in the first
quadrant and in the first quadrant if we draw the curve like this, so this is our first quadrant
and that is the equation of this circle. Let us say, so x plane and z plane.

So, in the first quadrant our x is varying from 0 to a. So, this is our x limit and then z will
vary from 0 to √𝑎2 − 𝑥 2 . So, z will vary from 0 to √𝑎2 − 𝑥 2 . So, that is what we have
𝑎 𝑑𝑥 2 −𝑥 2
written as the limit of z. So, now, this will result in 8𝑎 ∫𝑥=0 √𝑎2 [𝑧]√𝑎
0 .
−𝑥 2

𝑎 √𝑎2 −𝑥 2
So, this will become 8𝑎 ∫𝑥=0 √𝑎2 𝑑𝑥 . So, both will cancel out and then it will become
−𝑥 2
𝑎
8𝑎 ∫0 𝑑𝑥 and then this will be 8𝑎[𝑥]𝑎0 . So, this will be basically 8𝑎2 .

So, here we had to pay a very close attention that the two cylinders, I mean first of all the
we have to look at the equation of the cylinder along which plane it is symmetric, and then
we just have to calculate the surface area on the upper half and multiplied by 2 that will
give you the area surface area of the whole cylinder.

Now, it is being intersected or it is being cut off by some other cylinder. So, depending on
what type of equation that cylinder has we can use the projection in that plane and even in
that plane here we have use the property that its basically a circle and then we do not have
to calculate along the whole circle in a way or disk in a way. We just have to look into our
first quadrant and then multiplied by 4 that is what we are doing here. And this is your
required answer. So, this is what we are doing here and that is why required answer.

So, this is how we calculate the surface area in this problem. We will work out a similar
problem like that, before we move on to our next topic.

312
(Refer Slide Time: 11:01)

So, let us start with our next example. So, our next example is, show that the area of the
3
𝑥2 𝑦2 𝑥2 𝑦2 2
surface of the paraboloid + = 2𝑧 inside the cylinder, 𝑎2 + 𝑏2 = 𝑘 is 3 𝜋𝑎𝑏[(1 + 𝑘)2 -
𝑎 𝑏

1].

So, instead of having two regular cylinders we have one of them as paraboloid and it is
being how to say in a way it is cut off or it is inside that cylinder given by this equation.
So, here the equation of the cylinder the base basically lies in xy plane and based on that
𝑥2 𝑦2
first of all the given surface, we can write the given surface is + = 2𝑧.
𝑎 𝑏

𝜕𝑧 𝜕𝑧
So, from here we obtain the given surface is this, full stop. We obtain 1 + (𝜕𝑥)2 + (𝜕𝑦)2.
𝜕𝑧 𝜕𝑧 𝑥2 𝑦2
Now, from here if I calculate𝜕𝑥 and 𝜕𝑦 then this will be 1 + 𝑎2 + 𝑏2 , right, yes. So, this is

what we obtain here.

Now, the required surface area, since our ellipse lies in xy plane we take the projection on
xy plane. So, the required surface area S can be written as, so
. 𝜕𝑧 𝜕𝑧
∬𝑅 √1 + (𝜕𝑥)2 + (𝜕𝑦)2 𝑑𝑥𝑑𝑦, all right.

313
(Refer Slide Time: 14:21)

So, now let us substitute the values here. So, this will be over the region R,
. 𝑥2 𝑦2
∬𝑅 √1 + 𝑎2 + 𝑏2 , all right.

So, now, what we can do in first of all instead of how to say calculating this integral in x
and y variable, what we will do, we will do some change of variables here. And, if we do
the change of variables basically doing some method of substitution we can be able to
reduce this whole thing into a rather simpler integral. So, how we can do that? We can
substitute, so we put 𝑥 = 𝑎 tan 𝜃 cos 𝜑 & 𝑦 = 𝑏 tan 𝜃 sin 𝜑.

Since, we have integral of two variable we have to assume the substitution in two variables
as well and from here I can calculate our dx dy. So, those things can be calculated easily.
𝑥2 𝑦2
Now, what we can do? We will basically see that √1 + 𝑎2 + 𝑏2 is equals to

√1 + 𝑡𝑎𝑛2 𝜃(𝑠𝑖𝑛2 𝜑 + 𝑐𝑜𝑠 2 𝜑)and this will reduce to the √1 + 𝑡𝑎𝑛2 𝜃. So, this is basically
𝑠𝑒𝑐 2 𝜃 so that means, sec 𝜃, all right. So, this is our sec 𝜃.

𝑥2 𝑦2
And x square by a square and our surface 𝑎2 + 𝑏2 = 𝑘 will transform to 𝑡𝑎𝑛2 𝜃(𝑠𝑖𝑛2 𝜑 +

𝑐𝑜𝑠 2 𝜑) = 𝑘. So, this is our given surface, due to this change of variable the surface will
get also changed in that variable and therefore, this will be 𝑡𝑎𝑛2 𝜃 = 𝑘 And from here we
will obtain 𝜃 = tan−1 √𝑘, all right. So, we have got the value of 𝜃.

314
(Refer Slide Time: 16:52)

And now, if we do the change of variable then we also have to calculate the Jacobian you
remember. From the change of variable in our how to say change of variables part
basically, where if we wanted to change the variable we can change it, but we also have to
introduce a Jacobian.

𝜕(𝑥,𝑦)
So, we can calculate the Jacobian and Jacobian is basically 𝐽 = 𝜕(𝜃,𝜑). So, this is basically
𝜕𝑥 𝜕𝑥
𝜕𝜃 𝜕𝜑
|𝜕𝑦 𝜕𝑦
|. So, if we calculate this whole thing then it will be 𝑎𝑏𝑠𝑒𝑐 2 𝜃𝑡𝑎𝑛𝜃, all right. So,
𝜕𝜃 𝜕𝜑

we have got our Jacobian how to say this integrand as sec 𝜃 tan as sec 𝜃 and we have also
got our 𝜃. So, now, let us go back to our required surface integral.

So, this is this was our required surface integral and we have basically 𝑆 =
. 𝑥2 𝑦2
∬𝑅 √1 + 𝑎2 + 𝑏2 . So, now we are doing the change of variable in xy plane. So, in xy plane

our 𝜃 would vary from 0 to 2𝜋 and the 𝜑 will vary from 0 to 2𝜋. So, this is the limit for
2𝜋 𝜃
the 𝜑 and our 𝜃 would vary from 0 to 𝜃, ∫𝜑=0 𝑑𝜑 ∫0 sec 𝜃 𝑎𝑏𝑠𝑒𝑐 2 𝜃 tan 𝜃 𝑑𝜃, all right.

So, now, here what we can do when we integrate this one. So, this will become
tan−1 𝑘
2𝜋𝑎𝑏 ∫0 𝑠𝑒𝑐 3 𝜃 tan 𝜃 𝑑𝜃 And here we will do some method of substitution, so if I
substitute 𝑧 = tan 𝜃 then it will 𝑠𝑒𝑐 2 𝜃 or if we substitute say 𝑧 = sec 𝜃 then it basically

315
the tan 𝜃. So, by doing some method of substitution this integral would reduce to basically
integral 𝑠𝑒𝑐 2 𝜃 evaluated at 0 to tan−1 𝑘.

(Refer Slide Time: 19:39)

So, here is basically method of substitution. So, we will keep sec 𝜃 tan 𝜃 together and then
we substitute sec 𝜃 = 𝑧 . So, that will give us sec 𝜃 tan 𝜃 𝑑𝜃 = 𝑑𝑧 and then this whole
thing will reduce to just 𝑑𝑠𝑒𝑐 2 𝜃 and then that will give us the required integral.

2𝜋𝑎𝑏 −1 𝑘
Now, here I can substitute [𝑠𝑒𝑐 3 𝜃]tan
0 can be written as 1 + 𝑡𝑎𝑛2 𝜃 and then whole
3
2𝜋𝑎𝑏
to the power 3. And then if I write 𝑠𝑒𝑐 2 𝜃 as 1 + 𝑡𝑎𝑛2 𝜃 then I have [(1 +
3
3 −1 𝑘
𝑡𝑎𝑛2 𝜃)2 ]tan
0 here, yes.

3
2
So, this is tan−1 𝑘 and now I am substituting the value. So, 3 𝜋𝑎𝑏[(1 + 𝑘)2 − 1]. So, this

will be our required answer. So, this is what we needed to prove.

The calculation is little bit tricky and but what I am doing here in the calculation is not
relevant to the complexity of the surface integral. So, this is basically for you to practice
how you can do such kind of calculation. So, when you see a complicated integral like this
all you have to do is the think of some alternative way. Here in this case we have used
method of change of variables.

316
So, when we change the variables we basically get our variables change and then the
Jacobian. So, this is our Jacobian which I have substituted here that is the integrand and
then we have 𝑑𝜃𝑑𝜑 where 𝜑 limit is 0 to 2𝜋 and 𝜃 limit is 0 to tan−1 𝑘. And then we have
done some simple calculation which you can do on your own time.

So, in this surface integral part we have seen certain type of how to say integrals where
your one surface is been cut out by another surface. And then surface area which is
common between these two surfaces needs to be calculated. So, it is like in our plane curve
region where we had one curve is being cut off by another curve, and then we have to
calculate the common area bounded by these two curves or enclosed by these two curves.
So, its smaller less the same thing but in a higher dimension. So, kind of like a
generalization of those problems.

Now, we can work out one last example because this is indeed an interesting topic.

(Refer Slide Time: 23:16)

So, let us work out our very last example on the surface integral section. So, find the
surface area of the paraboloid 𝑥 2 + 𝑦 2 = 𝑎𝑧 which lies between the planes between the
planes 𝑧 = 0 and 𝑧 = 𝑎. So, of course, the plane is lying in z equals, so the plane is given
by 𝑧 = 0 and 𝑧 = 𝑎 .

So, those are the two points where those are the two planes basically in between we have
our paraboloid then we have to calculate that surface area lying between these two planes.

317
And to do that basically we will have the given surface, yes. So, 𝑧 = 0 and 𝑧 = 𝑎 is
basically a plane in xy plane and we will basically take the projection in xy plane actually.
So, let us do that.

𝜕𝑧 𝜕𝑧 𝜕𝑧
So, the given surface is 𝑥 2 + 𝑦 2 = 𝑎𝑧, then our 1 + (𝜕𝑥)2 + (𝜕𝑦)2. So, if I do 𝜕𝑥 then this
4𝑥 2 4𝑥 2 4𝑦 2 1
will be and then this one will be 1 + + . So, if I take 𝑎2 square common and this
𝑎2 𝑎2 𝑎2
1
will be 𝑎2 (𝑎2 + 4(𝑥 2 + 𝑦 2 ) , all right. So, this here.

Now projecting on the plane z equals to a gives. So, if I project this here on the plane 𝑧 =
𝑎, then in that case then our region R will be 𝑥 2 + 𝑦 2 = 𝑎2 because if you project the
parabola like this then on xy plane it will be basically a circle, all right.

So, that is the circle we will obtain and thus our surface area will be, integral over the
region R we are taking the projection on xy plane. So, always look for the equation. What
type of equation is given? Whether it is given in xz variable or yz variable or xy variable
or even if the equation is given in terms of planes then in what sense they are meaning and
based on that we have to decide or projection. So, here 𝑧 = 0, 𝑧 = 𝑎 are basically equation
of xy plane and therefore, we have to take the projection in xy plane, all right.

. 𝜕𝑧 𝜕𝑧 1 .
So, 𝑆 = ∬𝑅 √1 + (𝜕𝑥)2 + (𝜕𝑦)2 𝑑𝑥𝑑𝑦 = 𝑎 ∬𝑅 √𝑎2 + 4(𝑥 2 + 𝑦 2 𝑑𝑥𝑑𝑦, all right.

(Refer Slide Time: 27:05)

318
 1 2𝜋 𝑎
Now, this can be written 𝑎 ∫0 ∫0 √𝑎2 + 4𝑟 2 𝑟𝑑𝑟𝑑𝜃, where 𝑥 = 𝑟 cos 𝜃 and 𝑦 = 𝑟 sin 𝜃,

where r is between 0 to a and 𝜃 is between 0 to 2𝜋.

Because in this case instead of focusing on one quadrant we can how to say rotate our
𝜃 from 0 to 2𝜋 and that will give us the entire surface area over the entire circle and this
will be basically 𝑑𝑟𝑑𝜃 here and since we also have to incorporate a Jacobian then it will
be r here.

So, now this is our required change of variable its how to say changed integral. So, we
have basically changed the variables using 𝑥 = 𝑟 cos 𝜃 and 𝑦 = 𝑟 sin 𝜃 . So, this is the
integrand, this is the integrand, that is your Jacobian, that is you how to say 𝑑𝑟𝑑𝜃.

2𝜋 𝑎
Now, if we perform then basically we take 𝑑𝜃 here, so this will be 𝑎
∫0 𝑟√𝑎2 + 4𝑟 2 𝑑𝑟.

And to calculate this integral we substitute 𝑟 2 = 𝑧, then this will be 2𝑟𝑑𝑟 = 𝑑𝑧 and then
we will use some integral calculus formula.

So, calculating this integral is not difficult. You can do that by your own. So, ultimately
3
2𝜋 2
you will obtain [(𝑎2 + 4𝑟 2 )2 ]𝑎0 . So, this will be 4 and then to this. So, this will be
𝑎 3.8
3 3
𝜋
[(5𝑎2 )2 − (𝑎2 )2 ].
6𝑎

3
𝜋
So, ultimately, we will have 6𝑎 𝑎[52 𝑎2 − 𝑎2 ] and one a is getting cancelled. So, we will
3
𝜋
obtain 𝑎2 [52 − 1]. So, the simplification is up to you. So, how you do the simplification?
6

Now, this is one another interesting example where instead of having it cut off by some
other surface we just had two planes in between we have our paraboloid and then we
needed to calculate the surface area. So, these are some problems which you can encounter
in surface integral part. And I hope I try to cover as many examples as I could. I will also
include some examples in your assignment sheet and hopefully you will be able to work
on them.

So, I thank you for your attention for today. And, I will see you in the next class.

319
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institutes of Technology, Kharagpur

Lecture – 31
Volume Integral, Gauss Divergence Theorem

Hello students. So, in the class before the last class as well we introduced the concepts of
surface integral and we also tried to cover as many good examples as possible; where we
saw that how you calculate a surface area which is common between two surfaces or where
two surfaces are cutting each other at some place and also, the one surface is lying in
between two planes. So, we also worked out examples like that.

Now in today's class we will cover our next topic which is basically Volume Integral. So,
we started with line integral, then surface integral, area between two curves and today we
will start with a volume integral.

(Refer Slide Time: 01:13)

So, what do we mean by volume Integral? So, let us start with. So, volume integrals are
basically integral of these types. So, triple integral let us say and I write the volume or the
.
region where we are doing the integration ∭𝑉 𝑓𝑑𝑉 . So, this is how we denote our volume
integral volume integral. Now, in order to calculate these type of integral we I will show
that how you can do it. So, first of all we can write this one as triple integral or one can

320
also prefer to write just one integral and then, V and it depending upon how you are writing
.
this dv. So, some people prefer to write it as ∭𝑉 𝑓𝑑𝑥𝑑𝑦𝑑𝑧, some people write simply dV.
So, if you write 3 variables and if you just write one integral, that means that it is a volume
integral.

So, basically any triple integral is like a volume integral where you are calculating actually
the volume enclosed by that particular surface or even if nothing is given, then basically
you are calculating the volume within that range let us say. So, for example if we if our
1 1 1
𝑓 = 1 and if our volume has limits ∫0 ∫0 ∫0 𝑑𝑥𝑑𝑦𝑑𝑧, then we are basically calculating the
volume of this surface here. So, where we can draw it as something like this 𝑥 = 0 to 1
and then, y is 0 to 1 and then, 𝑧 = 0 to 1, so, basically this cuboid in a way. So, that is one
way we mean by what we mean by volume integral.

So, let us consider an actual how to say example where we calculate the volume of a
surface. So, first example find or calculate find the integral. Let us say I am writing 𝐼 =
∭(𝑥 2 + 𝑦 2 + 𝑧 2 )𝑑𝑥𝑑𝑦𝑑𝑧 over the sphere 𝑥 2 + 𝑦 2 + 𝑧 2 ≤ 𝑎2 , all right. So, basically we
have to calculate this integral here over this surface. So, basically a volume integral of this
integrand over this surface. So, of course everything is then how to say nice figure form,
we have 𝑥 2 + 𝑦 2 + 𝑧 2 ≤ 𝑎2 .

So, we can think of it as let us say our circle situation where we had 𝑥 2 + 𝑦 2 = 𝑎2 and
then, in that case we used to substitute 𝑥 = 𝑟 cos 𝜃 , 𝑦 = 𝑟 sin 𝜃 and then, we transform the
whole integral into an integral with r and 𝜃 using Jacobian and other basically change of
variables. So, here in this case also we will do some substitution. So, we substitute 𝑥 =
𝑟 cos 𝜃 cos 𝜑 because we are in function of 3 variable. We have a spherical polar
coordinate system and the spherical polar coordinate system involves the r,𝜃 and 𝜑, all
right.

So, we substitute 𝑥 = 𝑟 cos 𝜃 cos 𝜑 , 𝑦 = 𝑟 sin 𝜃 sin 𝜑 . Let us take it as sin 𝜃. 𝑦 =

𝑟 sin 𝜃 sin 𝜑 and 𝑧 = 𝑟 cos 𝜃. So, where 𝜃 would vary from 0 to 𝜋, 𝜑 would vary from 0
to 2𝜋 and the r would vary from 0 to a. So, we can see all those things from our spherical
polar coordinate system.

321
(Refer Slide Time: 06:13)

Now, we will start with our integral I. So, our integral I is basically integral over the sphere
.
let us say which is given. So, 𝐼 = ∭𝑉(𝑥 2 + 𝑦 2 + 𝑧 2 )𝑑𝑥𝑑𝑦𝑑𝑧 and if I change the variables,
then in that case it will be a volume integral again, but in terms of r ,𝜃 and 𝜑 and this one
will be 𝑟 2 . So, first term will give a, so, if we do the square, then we will take
𝑠𝑖𝑛2 𝜃 common and then, it will be 𝑐𝑜𝑠 2 𝜑 + 𝑠𝑖𝑛2 𝜑 that will be 1 and then, again
𝑟 2 common.

So, it will be 𝑠𝑖𝑛2 𝜃 + 𝑐𝑜𝑠 2 𝜃 and then, that will again be 1. So, it will be basically 𝑟 2 and
if I substitute if I transform dx dy dz into another variable involving r,𝜃and 𝜑, then this
will be 𝑟 2 sin 𝜃 𝑑𝑟𝑑𝜃𝑑𝜑. So, our Jacobian determinant is basically 𝑟 2 sin 𝜃 all right. So,
. 𝑎 𝜋 2𝜋
this will become integral ∭𝑉(𝑟 2 )𝑟 2 sin 𝜃 𝑑𝑟𝑑𝜃𝑑𝜑 = ∫𝑟=0 𝑟 4 𝑑𝑟 ∫𝜃=0 sin 𝜃 𝑑𝜃 ∫𝜑=0 𝑑𝜑.

So, 𝜑 is totally independent of r and 𝜃.

𝑟5
So, we can integrate and put the value 2𝜋[ 5 ]𝑎𝑟=0 [− cos 𝜃]𝜋0 . So, cos 𝜋 is −1 and minus
𝑎5
minus plus and minus minus plus cos0, so, again plus; so, basically 2. So, this is 4𝜋 =
5
4
𝜋𝑎5 . So, this is our required volume integral, our triple integral. So, here in this case
5

basically the given equation was a sphere, the given equation of the wall of the surface
was just sphere.

322
So, it is fairly easy to come up with this substitution and then, just substitute for the variable
x y and z and use some Jacobian that will give us the required answer as this one, all right.
We can consider another example. So, another example could be let me find an interesting
example which is yes.

(Refer Slide Time: 09:23)

. 𝑥2 𝑦2 𝑧2
So, now another example is compute 𝐼 = ∭𝑉 √1 − 𝑎2 − 𝑏2 − 𝑐 2 𝑑𝑥𝑑𝑦𝑑𝑧 taken over the

𝑥2 𝑦2 𝑧2
region + 𝑏2 + 𝑐 2 ≤ 1 all right. So, in this case we have the ellipsoid basically. So,
𝑎2

previously we had a sphere and in this case we have that ellipsoid. So, what we will have
is the given region is an ellipsoid.

𝑥 𝑦 𝑧
So, substituting = 𝑟 sin 𝜃 cos 𝜑 , 𝑏 = 𝑟 sin 𝜃 sin 𝜑 and = 𝑟 cos 𝜃. So, basically if I
𝑎 𝑐
𝑥2
substitute we can see it as a circle as a sphere as well. So, instead of considering if I
𝑎2
𝑥 𝑥 𝑦 𝑧
consider 𝑎 as a variable let us say, then in that case 𝑎 = 𝑢 as the 𝑏 = 𝑣 and 𝑐 = 𝑤 , then it

is basically 𝑢2 + 𝑣 2 + 𝑤 2 ≤ 1. So, then again it says sphere.

𝑥
So, that is what we are doing. We are taking the substitution for 𝑎 not x and similarly for
𝑦 𝑧
and 𝑐 . So, now, I can write it as 𝑥 = 𝑎𝑟 sin 𝜃 cos 𝜑, 𝑦 = 𝑏𝑟 sin 𝜃 sin 𝜑and 𝑧 = 𝑐𝑟 cos 𝜃.
𝑏

, right. So, now I will do the change a variable. So, here we will have the volume integral
V and if I do the change a variable, then I mean this whole thing will reduce to basically

323
 .
𝐼 = ∭𝑉 √1 − 𝑟 2 |𝐽|𝑑𝑟𝑑𝜃𝑑𝜑 because we will be having this a 𝑐𝑜𝑠 2 + 𝑠𝑖𝑛2 formula and
everything will keep getting reduced to 1 and dxdydz.

So, here we will have our Jacobian determinant and then dr d theta d phi. So, the Jacobian
determinant can be given by |𝐽| = 𝑎𝑏𝑐𝑟 2 sin 𝜃; this we can calculate easily.

(Refer Slide Time: 12:37)

So, Jacobian determinant can be calculated easily and this is 𝑎𝑏𝑐𝑟 2 sin 𝜃. So, I will
substitute all these values in our volume integral, right. So, let us substitute all these values
1 𝜋 2𝜋
in our volume integral I am calling it as 𝐼 = 𝑎𝑏𝑐 ∫𝑟=0 ∫𝜃=0 ∫𝜑=0 √1 − 𝑟 2 𝑟 2 sin 𝜃 𝑑𝑟𝑑𝜃𝑑𝜑.

Now, we can have this a separate.

So, we keep r integrals at one place, 𝜃 integrals in one place and 𝜑 integrals are one place.
So, since there are no variables involving 𝜑 in the integrands, it will be simply
1 𝜋
2𝜋𝑎𝑏𝑐 ∫𝑟=0 𝑟 2 √1 − 𝑟 2 ∫𝜃=0 sin 𝜃 𝑑𝜃 here and then, this will be − cos 𝜃 and just like
previous example my cos 𝜋 and minus cos ultimately this will reduce to 2 ×
1
2𝜋𝑎𝑏𝑐 ∫𝑟=0 𝑟 2 √1 − 𝑟 2 𝑑𝑟.

So, now calculating this integral is fairly simple. So, we just have to use some kind of
substitution. So, we substitute 𝑟 = sin 𝑡 and then, just transform this limits and ultimately
11𝜋
we will be able to obtain, so, its not very complicated. So,4𝜋𝑎𝑏𝑐 4 2 2 . So,

324
 𝜋2
ultimately 𝑎𝑏𝑐. So, this is the required volume integral of this integrand here taken over
4

the region given by this ellipsoid.

So, this is how we basically calculate the volume integral. So, volume integral is not that
complicated in a way because if someone has worked out a line integral and area between
the curve or surface integral, then volume integral is basically very how to say small
generalization. It is not too complicated that is maybe small is not the right word, but its
not very complicated once you are perfect with the area between the curves or area between
different surfaces.

Now in volume integral, there is a very important theorem. Of course one can view it more
from the vector calculus point of view, but we can also see this theorem from integral
calculus point of view and that is called a Gauss Divergence theorem.

So, Gauss Divergence theorem gives us a tool to transform any surface integral into a
volume integral. So, if you are given a surface integral which may appear to you a little bit
complicated, then instead of solving the surface integral if it is in certain particular form,
then you can transform the surface integral into a volume integral. So, how we can do that
let us see in our statement.

(Refer Slide Time: 16:15)

So, Gauss divergence theorem; so it says that if a 3 dimensional regular domain let us call
it as V is bounded by a smooth orientated surface lets us say S and f, g, h are 3 functions

325
with continuous partial derivatives. That means, that is 𝑓𝑥 , 𝑔𝑦 𝑎𝑛𝑑 ℎ𝑧 are continuous,
alright.

So, the first order with continuous maybe first order it should be written first order partial
.
derivatives at each point of V and S, then our volume integral ∭𝑉(𝑓𝑥 + 𝑔𝑦 +
.
ℎ𝑧 )𝑑𝑥𝑑𝑦𝑑𝑧 = ∬𝑆(𝑓𝑑𝑦𝑑𝑧 + 𝑔𝑑𝑧𝑑𝑥 + ℎ𝑑𝑥𝑑𝑦) ; where the surface integral is taken over
the exterior of the surface S. That means, the normal or basically on the outer side of this
surface S and if that.

So, this integral basically this Gauss Divergence theorem basically gives you a tool that
can help you transform a surface integral into a volume integral. So, we have to make sure
that these functions f, g and z f, g and h not z f, g and h they are they have a continuous
partial derivative and if they do then in that case you see you have a function f, but if they
have continuous partial derivative just looking at them, you can be able to understand
whether they have continuous partial derivatives or not its very simple. So, even if they
involve lot of complicated algebraic expressions when you take 𝑓𝑥 , then it is actually
always one derivative one powerless.

So, it could be a rather simpler expression to deal with then dealing with this surface
integral, right. So, with the help of Gauss Divergence theorem we can always be able to
transform the surface integral into a volume integral and most probably I mean in most of
the cases it happens that this volume integral is very simple to evaluate compared to this
surface integral. So, this is the only theorem from vector calculus in a way we will learn
in our integral calculus part and we will work out few examples to see its application that
whether actually a given surface integral can be handled using a Gauss Divergence
theorem or not.

326
(Refer Slide Time: 21:03)

.
So, let us start with our first example. So, evaluate let us say 𝐼𝑠 = ∬𝑆(𝑥𝑑𝑦𝑑𝑧 + 𝑦𝑑𝑧𝑑𝑥 +
𝑧𝑑𝑥𝑑𝑦) S being the outer side of the cube [0, 𝑎] × [0, 𝑎] × [0, 𝑎]. So, that is where the
surface integral is taking place on the outer side. Now, of course it may not be a very
complicated expression. So, here we have x, y and z and we can just integrate how to say
need surfaces and then, that will give us the surface integral, but at the end of the day we
have to integrate on each surfaces. Ultimately it will become a very big example or big
calculation to do.

So, what we can do here, we will apply Gauss Divergence theorem because obviously x,
y and z are continuous functions and they also have continuous first order partial
derivatives. So, these f g and h have continuous partial derivatives. So, I can write
here 𝑓(𝑥, 𝑦, 𝑧) = 𝑥, 𝑔(𝑥, 𝑦, 𝑧) = 𝑦, ℎ(𝑥, 𝑦, 𝑧) = 𝑧 . So, they can be function of x y z as well
here in this case we are lucky that they are only functions of x y and z respectively here.

That is what we have. Then, our partial derivative 𝑓𝑥 = 1, 𝑔𝑦 = 1 & ℎ𝑧 = 1, then by Gauss
.
Divergence theorem we will have surface integral ∬𝑆(𝑓𝑑𝑦𝑑𝑧 + 𝑔𝑑𝑧𝑑𝑥 + ℎ𝑑𝑥𝑑𝑦) =
.
∬𝑉(𝑓𝑥 + 𝑔𝑦 + ℎ𝑧 )𝑑𝑥𝑑𝑦𝑑𝑧.

𝑎 𝑎 𝑎
So, now volume integral ∫𝑥=0 ∫𝑦=0 ∫𝑧=0(1 + 1 + 1)𝑑𝑥𝑑𝑦𝑑𝑧. So, now we will integrate

how to say first with respect to. So, we will basically write here dx. So, that will go into
each basically individual integral and this will then reduce to sorry so, this is not dx dy.

327
So, 𝑓𝑥 dx and then, dx dy dz is here yes. So, that is I was wondering, so, why it is looking
that way?

So, we have dx dy dz and we have 1 here, 1 here, 1 here.

(Refer Slide Time: 24:39)

𝑎 𝑎 𝑎
So, this will ultimately be 3 ∫0 𝑑𝑥 ∫0 𝑑𝑦 ∫0 𝑑𝑧. So, ultimately we will obtain a a a. So, 3
times a, that means 𝑎3 so, 3𝑎3 . So, this is the required volume integral. So, here yes the
formula is correct. So, here you can see we could have calculated the surface integral is
just that S being the outer surface of the cube. So, we have to go on every surface one by
one, one by one, so, there could be a possibility that we may have to evaluate 8 types of
surface integral.

So, instead of doing that we just use Gauss Divergence theorem because our functions are
nicely behaving and we can calculate the volume integral and the answer of which is as
same as the answer of the surface integral. So, as you can see it is proven to be a very
efficient tool in order to handle these kind of integrals. We can consider one more example.

328
(Refer Slide Time: 25:45)

.
And let us consider another example. So, evaluate 𝐼 = ∬𝑆(𝑥 3 𝑑𝑦𝑑𝑧 + 𝑦 3 𝑑𝑧𝑑𝑥 +
𝑧 3 𝑑𝑥𝑑𝑦), where S is the sphere. Our surface is basically a sphere 𝑥 2 + 𝑦 2 + 𝑧 2 = 𝑎2 .

So, obviously here we can do some substitution, but before that if we want to evaluate the
surface integral over the surface, it will certainly be little quite complicated. So, instead of
doing that what we can do is we can use Gauss Divergence theorem. So, by Gauss
Divergence theorem obviously we can see that this is our f, this is our g and this is our h.
So, since they are all algebraic or quadratic function, they will have continuous first order
partial derivatives and if we use the Gauss Divergence theorem, then this 𝐼𝑆 = 𝐼𝑉 =
.
∭𝑉(3𝑥 2 + 3𝑦 2 + 3𝑧 2 )𝑑𝑥𝑑𝑦𝑑𝑧.

So, that 𝑔𝑦 that means 3𝑦 2 and then ℎ𝑧 = 3𝑧 2 and then, here we will have dx dy dz. So,
.
basically 3 ∭𝑉(𝑥 2 + 𝑦 2 + 𝑧 2 )𝑑𝑥𝑑𝑦𝑑𝑧, all right and now we basically use that
substitution 𝑥 = 𝑟 sin 𝜃 cos 𝜑 , 𝑦 = 𝑟 sin 𝜃 sin 𝜑 , 𝑧 = 𝑟 cos 𝜃. So, then this integral will
reduce to two volume integrals.

329
(Refer Slide Time: 28:03)

𝑎 𝜋 2𝜋
So, this will reduce to 3 ∫𝑟=0 ∫𝜃=0 ∫𝜑=0 𝑟 2 𝑟 2 sin 𝜃 𝑑𝑟𝑑𝜃𝑑𝜑or we can modify this example

a little bit, so that just to make things a little bit easier. So, this will be 3 3 3 𝑥 2 + 𝑦 2 + 𝑧 2 ;.
So, then in that case we will have 𝑟 2 𝑠𝑖𝑛2 𝜃(𝑐𝑜𝑠 2 𝜑 + 𝑠𝑖𝑛2 𝜑) + 𝑟 2 𝑐𝑜𝑠 2 𝜃. So, ultimately it
will reduce to a very nice simple formula.

So, we can do that and then, this will be 0 to 2𝜋, then this is 𝑟 2 times the value of the
Jacobian will be 𝑟 2 sin 𝜃 and we will have 𝑑𝑟𝑑𝜃𝑑𝜑. So, now we can put 𝑑𝜑here, so, this
𝜋 𝑎
will be 3 × 2𝜋 ∫𝜃=0 sin 𝜃 𝑑𝜃 ∫𝑟=0 𝑟 4 𝑑𝑟.

12
So, ultimately we can evaluate this whole thing and it will reduce to 𝜋𝑎5 . So, here you
5

can see just looking at the integral. So, 𝑥 3 , 𝑦 2 and 𝑧 3 just for the simplicity we have
considered 𝑥 3 , 𝑦 2 and 𝑧 3 So, just looking at the integral we can realize that it could be a
little bit complicated to do the surface integral.

However, if we use the Gauss Divergence theorem, then this whole thing will reduce to a
very simple integral where we just use some method of substitution; or change of variables
to obtain a rather simpler integral and this is very easy to calculate.

So, there could be a possibility that this surface integral may have lead to a complicated
calculation, but just by the help of Gauss Divergence theorem we finish the problem in
like 5 or 6 steps. So, the Gauss Divergence theorem is proven to be a very important tool

330
in integral calculus and we will also see in vector calculus that how with the help of Gauss
Divergence theorem we can reduce a surface integral to a volume integral more from the
vector calculus point of view.

So, I will stop here for today and in the next class, we will continue with our next topic
and I will try to include some examples based on this Gauss Divergence and you know in
your assignments sheet and I look forward to your next class.

Thank you.

331
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institutes of Technology, Kharagpur

Lecture – 32
Vector Calculus

Hello students. So, in the previous class we looked into the concepts of volume integral
and before that class we worked out the examples on surface integral. For the time being
we go to the Vector Calculus part because that is also very important and at the end I will
try to cover the center of gravity chapter where we have some formulas, from integral
calculus that will help us calculate center of gravity and moment of inertia.

But for the moment, I will start with the vector calculus section which is also very
interesting and you can be able to relate the things which you have learnt in your scalar
functions like limits continuity and differentiability, integration of scalar functions which
we were doing up until last class.

So, in the vector calculus as well these things are there and we will see how we actually
define the vector function and its limits continuity differentiability things like that. So,
today we are going to start with this chapter here; so, this here.

(Refer Slide Time: 00:26)

332
So, scalar sorry so, limit continuity differentiability of vector functions. So, this is what
we are going to start with all right.

(Refer Slide Time: 01:42)

So, let me start a new page and I am going to increase the thickness a little bit yes. So,
today’s topic is vector functions. So, what do we mean by vector functions?

So, usually the function of scalar variable or scalar functions are written as 𝑦 =
𝑓(𝑥) where x is the independent variable and y is the dependent variable. Possible
examples could be 𝑦 = 𝑓(𝑥) = 𝑥 2 + 2𝑥 + 3 or 𝑓(𝑥) = sin 𝑥 or 𝑒 𝑥 any such functions
which we have learnt basically. So, those are the scalar function of one variable.

Now how do we define the function of; how to say it is a vector functions actually? So, I
mean we know that vector quantity they have directions as well as the magnitude. So, how
do you define functions related to that?

So, we can give a formal definition for a vector function. So, the formal definition goes
like this. So, let D be a subset of set of all real numbers and if for every t in D, we can
associate or we associate by some rule. So, what is this rule? I will define, but for the time
being let us write it as by some rule a unique vector say 𝑓⃗(𝑡), then this rule defines a
vector function all right; of the scalar variable t. And here f is basically a vector quantity
or a vector function for the variable t.

333
So, f can be called as basically a vector function. So, one possible; so, we represent this.
So, since a vector quantity has three directions. So, i, j and k we can be able to with the
help of i j and k we can be able to write a vector quantity or a vector basically. So, here
this vector function 𝑓⃗(𝑡) can also be represented with respect to those unit vectors i , j and
k. So, we represent or we write 𝑓⃗(𝑡) = 𝑓1 (𝑡)𝑖̂ + 𝑓2 (𝑡)𝑗̂ + 𝑓3 (𝑡)𝑘̂ where 𝑖̂, 𝑗̂, 𝑘̂ are unit
vectors. So, basically along x axis y axis and z axis so, they are perpendicular to each other
and they are the unit vectors.

So, here 𝑓1 (𝑡) is a function of t, 𝑓2 (𝑡) is a scalar function of t, 𝑓3 (𝑡) is a scalar function of
t. So, all of them are basically scalar function of t and while writing with the help of i , j
and k they constitute a vector function. So, one possible example could be let us say 𝑟⃗ =
𝑓⃗(𝑡). So, we write basically the vector function by 𝑟⃗ and 𝑟⃗ is a function of t and we write
say 𝑟⃗ = 𝑓⃗(𝑡) = 2𝑡 2 𝑖̂ + sin 𝑡 𝑗̂ + cos 𝑡 𝑘̂ ok. So, this is one such example for let us say 𝑡 ∈
[1,2].

So, this could be our a vector function. So, like this we can define several types of vector
functions and they are how to say how we represent them using i j and k their domain well;
for example, where the variable t belongs to things like that all right.

(Refer Slide Time: 07:10)

So, now what do we mean by let us say scalar field and vector field. So, scalar field and
vector field what do we mean by them? So, scalar field is basically; so, if to each point

334
𝑃(𝑥, 𝑦, 𝑧) of a region R let us say of a region R in space, there corresponds a unique scalar
𝑓(𝑃), then f is called a scalar point function and we say that a scalar field f is defined on
R is fine yes.

So, if for every point P in that region R; if we can correspond a unique scalar 𝑓(𝑃). So,
then in that case that is unique scalar is called as the scalar point function and we say that
a scalar field f is defined on R. So, similarly and so for example, we have 𝑓(𝑥, 𝑦, 𝑧). So,
for example, if we have 𝑓(𝑥, 𝑦, 𝑧) = 𝑥 2 𝑦 3 − 3𝑧 2 . So, this defines a scalar field. all right.

So, next we define the vector field. So, if to each point 𝑃(𝑥, 𝑦, 𝑧) of a region D let us say,
in space there corresponds a unique vector 𝑓(𝑃), then f is called a vector point function.
And we say that a vector field f is defined on D. So, I defined on D. So, that is how we
define a vector field.

(Refer Slide Time: 11:03)

And a possible example could be let us say 𝑓(𝑡) = 3𝑡 2 𝑖̂ − 2𝑡𝑗̂ + 𝑒 𝑡 𝑘̂ . So, this defines a
vector field on the given domain actually. So, we can have any domain and then this will
actually defining a vector field all right. So, that is how we define the vector field.

Next we have is so, we know in the similar fashion we can define all sorts of vector
functions. So, defining the vector function and having its domain of definition is really not
a very complicated thing. So, those things are can be done very easily.

335
Now the next topic we know that from a function of scalar variable actually after the
function we have we when we study usually the limit our first scalar function. So, here
also we will basically look into the limit of a vector function. So, how do we define it what
do we mean by it and we will work out some examples. So, let us start limit of a vector
function not functions. So, usually we know that the the function 𝑓⃗(𝑡) = 𝑓1 (𝑡)𝑖̂ + 𝑓2 (𝑡)𝑗̂ +
𝑓3 (𝑡)𝑘̂ right.

So, when we pass the limit on the left hand side Do, let us we have lim 𝑓⃗(𝑡) and I am
𝑡→𝑡0

passing the limit on the left hand side. So, what would happen to the right hand side? How
do we define the limits on this 𝑓1 (𝑡)𝑖̂, 𝑓2 (𝑡)𝑗̂, 𝑓3 (𝑡)𝑘̂? So, the idea to define the limit is
basically to define the limit in every component. So, we define the limit in every
component all right. So, here we are defining the limits. Now if I define the limits then this
should be looking like lim [𝑓1 (𝑡)𝑖̂ + 𝑓2 (𝑡)𝑗̂ + 𝑓3 (𝑡)𝑘̂]; so, this is the required way to define
𝑡→𝑡0

the limit.

So, we passed the limit on individual components lim 𝑓1 (𝑡) 𝑖̂ + lim 𝑓2 (𝑡) 𝑗̂ + lim 𝑓3 (𝑡) 𝑘̂
𝑡→𝑡0 𝑡→𝑡0 𝑡→𝑡0

So, suppose after making t goes to 𝑡0 our 𝑓1 (𝑡0 )𝑖̂ + 𝑓2 (𝑡0 )𝑗̂ + 𝑓3 (𝑡0 )𝑘̂. So, this is how we
define the limits on individual how to say components.

(Refer Slide Time: 14:47)

336
So, we can look into an example. So, find the limit of let us say 𝑓⃗(𝑡) = 𝑒 𝑡 𝑖̂ − 𝑒 −2𝑡 𝑗̂ +
𝑒 3𝑡 𝑘̂ as 𝑡 → 0 all right. So, solution would be we will start like lim ..
𝑡→𝑡0

So, that 𝑡0 is actually 0 lim 𝑓⃗(𝑡) = lim(𝑒 𝑡 𝑖̂ − 𝑒 −2𝑡 𝑗̂ + 𝑒 3𝑡 𝑘̂). So, now, we will pass the
𝑡→𝑡0 𝑡→0

limits in individual functions. So, lim 𝑒 𝑡 𝑖̂ − lim 𝑒 −2𝑡 𝑗̂ + lim 𝑒 3𝑡 𝑘̂ right. And now we will
𝑡→0 𝑡→0 𝑡→0

pass the limit on individual functions. So, let us pass the limiting the first function. So,
when t goes to 0 𝑒 0 will go to 1.

So, i when again t goes to 0 𝑒 −2𝑡 will go to 1. So, this is 𝑒 −2𝑡 as t goes to 0 is 1 and then
here also it will be 1. So, ultimately this thing 𝑖̂ − 𝑗̂ + 𝑘̂ so, that will be the limit of this
vector valued function or vector function not vector valued, but a vector function. So,
similarly you can work out some more examples. So, I am writing them, I think it is very
straightforward. So, you may try to do them by your own. So, for example, find the limit
𝜋
of let us say 𝑓(𝑡) = sin 𝑡 𝑖̂ + cos 𝑡 𝑗̂ + (𝑡 + 1)𝑘̂ as 𝑡 → 2 .

𝜋 𝜋
So, when 𝑡 → 2 we will pass the limit. Then this will be 1, this will be 0, this will be 2 + 1.
𝜋
So, ultimately the answer will be 𝑖̂ + ( 2 + 1)𝑘̂. So, basically the component of j so, our

component of any unit vectors if it is 0, then we do not write that vector. So, we simply
avoid that vector that unit vector basically that or that particular component and we write
the remaining 2 components as a plus symbol and it is understood that the component of
the missing unit vector is basically 0 and that is why you are not writing it.

𝜋 𝜋 𝜋
So, here in this case at 𝑡 = 2 this is becoming cos 2 . So, cos 2 is 0. So, basically we do not

write the jth component. So, we just write ith and kth component all right and.

337
(Refer Slide Time: 18:14)

There are some other ways to denote a vector function or to denote a vector quantity. So,
far we have been writing 𝑓⃗(𝑡) = 𝑓1 (𝑡)𝑖̂ + 𝑓2 (𝑡)𝑗̂ + 𝑓3 (𝑡)𝑘̂ , but we can also write. So, there
are several other ways to so, we can also write 𝑓⃗(𝑡) = (𝑓1 (𝑡), 𝑓2 (𝑡), 𝑓3 (𝑡)). So, like a point
in 3 dimension or we can simply write𝑓⃗(𝑡) = (𝑓1 , 𝑓2 , 𝑓3 ) .

So, all these are just equivalent ways to write this vector function in commas basically. So,
it is up to you and in several books you may find the first notation, if you may find the
second notation or you may find the third notation.

There can be some other notations as well they basically mean this first line here that you
have component wise addition of these scalar functions multiplied by this unit vectors i j
and k respectively. And also you should not change the order. So, if you
write (𝑓1 (𝑡), 𝑓2 (𝑡), 𝑓3 (𝑡)) ; that means, the component of i or the coefficient of i is 𝑓1
coefficient of j is 𝑓2 and coefficient of k is 𝑓3 .

So, do not how to say play with the order of this notation all right. So, in a function of a
scalar variable or in scalar functions, we learnt about limits, but we also learnt about their
epsilon delta definition. So, epsilon delta definition is something like for scalar function is
if we have 𝑓(𝑥) → 𝑙 as 𝑥 → 𝑎, then this implies that, we usually used to write that for
every 𝜀 > 0 there exist a 𝛿 > 0 such that we had |𝑓(𝑥) − 𝑙| < 𝜀 whenever 0 < |𝑥 − 𝑎| <
𝛿.

338
So, this was our epsilon delta definition for the function of scalar variable when 𝑓(𝑥) →
𝑙 as 𝑥 → 𝑎. Now in case of a vector functions basically, we can have the similar epsilon
delta definition as well.

(Refer Slide Time: 20:51)

So, I can write 𝜀 − 𝛿 limit definition of a vector function all right. So, the definition goes
like this a vector function 𝑓⃗(𝑡) → 𝐼 when 𝑡 → 𝑎 if for any 𝜀 > 0. However, small or
arbitrarily small; however, small there exists 𝛿 > 0 or a positive number 𝛿 such that our

|𝑓⃗(𝑡) − 𝐼| < 𝜀 whenever 0 < |𝑡 − 𝑎| < 𝛿

So, |𝑡 − 𝑎| < 𝛿. So, and if this if this is true, then in that case we write simply lim 𝑓⃗(𝑡) = 𝐼.
𝑡→𝑎

So, this is basically our epsilon delta definition for the limit of a function. So, if we say
that the function f has a limit let us say I, then in that case for every positive ok

So, for any positive 𝜀; however, small there exist a positive number 𝛿 such that
|𝑓⃗(𝑡) − 𝐼| < 𝜀 whenever 0 < |𝑡 − 𝑎| < 𝛿

and we basically write lim 𝑓⃗(𝑡) = 𝐼 . So, this is this is how we define the limit of a function
𝑡→𝑎

using the 𝜀 − 𝛿 definition.

339
(Refer Slide Time: 24:05)

Similarly so, like limit of scalar functions we had several properties. In case of limit of
vector functions, we will also have several properties. So, I can just summarize them in a
small theorem. So, if 𝑓(𝑡) and 𝑔(𝑡) are two vector functions of scalar variable say t then
the following holds.

So, what are the following? First one is lim (𝑓⃗(𝑡) ± 𝑔⃗(𝑡)) = lim 𝑓⃗(𝑡) ± lim 𝑔⃗(𝑡). So,
𝑡→𝑡0 𝑡→𝑡0 𝑡→𝑡0

the limit of the sum is equal to sum of the limit all right. Now similarly if you have a
subtraction so, the limit of the difference is equals to difference of the limit. So, this is also
true. Now if we have let us say product. So, the product of two vector function which could
be scalar product or a dot product, then it will be so, if you have scalar product we simply
do not have product in vector functions. We either have scalar product which is also called
as dot product or the cross product.

So, if you have let us say a dot product of two vectors functions and their limit will be

lim (𝑓⃗(𝑡). 𝑔⃗(𝑡)) = lim 𝑓⃗(𝑡) . lim 𝑔⃗(𝑡). So that means, limit of the dot product is equal
𝑡→𝑡0 𝑡→𝑡0 𝑡→𝑡0

to dot product of the limits similarly limit of the cross product is lim (𝑓⃗(𝑡) × 𝑔⃗(𝑡)) =
𝑡→𝑡0

lim 𝑓⃗(𝑡) × lim 𝑔⃗(𝑡) and the third and fourth property is basically lim (𝜑(𝑡)𝑓⃗(𝑡)) =
𝑡→𝑡0 𝑡→𝑡0 𝑡→𝑡0

lim 𝜑(𝑡) lim 𝑓⃗(𝑡) where 𝜑 is a scalar function.

𝑡→𝑡0 𝑡→𝑡0

340
So; that means, if you have a vector function multiplied by scalar function and if you take
the limit, then basically limit of this scalar function times limit of the vector function. So,
that is what we get after taking the limit.

(Refer Slide Time: 27:25)

And fourth result is lim |𝑓⃗(𝑡)| = | lim 𝑓⃗(𝑡)|. So, limit of the absolute value is equals to
𝑡→𝑡0 𝑡→𝑡0

absolute value of the limit. So, whatever we will get after calculation. So, these are the
results which also holds for the vector functions and the proof is we will basically prove it
using the epsilon delta definition and it follows on the similar arguments based on the
results on a scalar function.

So, I am not proving these results and also they are not important from the context of the
syllabus. So, these are the results which we will be using as quite often and I mean for the
interested readers you can look into any vector calculus book suggested by me and there
you might be able to find the proof as well, but these are the results which are valid in case
of vector functions as well.

Now in the next class we will start with the continuity of a vector function and probably
will also introduce the concept of differentiability of a vector function. So, I will stop here
for today and I will see you in the next class.

Thank you.

341
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 33
Limit, Continuity, Differentiability

Hello students. So, in the previous class, we started with the vector calculus part of our
syllabus. And we introduced the concepts of vector function, and limit of a vector function,
how do we define the limit using epsilon delta definition, and some results which are
basically motivated from the function of scalar variables.

So, like I was saying in the previous class that like the function of scalar like the scalar
functions, sometimes I am confusing these two terms basically during the pronunciation,
so do not get confused. So, during the scalar functions, we introduce the concepts of a
continuity, differentiability, so in case of vector functions as well, we have those concepts
valid. So, let me introduce the concept of continuity, and then we will go to the definition
of differentiability.

(Refer Slide Time: 01:22)

So, what do we mean by continuous function, so continuous vector function all right. So,
a vector function 𝑓⃗ is said to be continuous for a value or at a point for a value 𝑡0 or at a
point 𝑡0 of t if 𝑓⃗(𝑡0 ) is defined.

342
And for any given arbitrary small positive number 𝜀 there corresponds or there exists a
positive number 𝛿 such that our |𝑓⃗(𝑡) − 𝑓⃗(𝑡0 )| < 𝜀, whenever |𝑡 − 𝑡0 | < 𝛿.

And if this is true, then the function 𝑓⃗ is said to be continuous at the point 𝑡 = 𝑡0 . Here in
case of limit we were only interested what is the value, when t is going to 𝑡0 . So, we were
not interested the value of the vector function at the particular point 𝑡 = 𝑡0 .

But, in case of continuity like scalar functions, we are interested in the value of the function
at the point 𝑡 = 𝑡0 as well. Because, if the function |𝑓(𝑡) − 𝑓(𝑡0 ) < 𝜀 , if the absolute
value is less than 𝜀, only then we can say that the function 𝑓(𝑡) is continuous at the point
𝑡 = 𝑡0 . So, in this case we actually want to know, what is the value of that vector function
at the point 𝑡 = 𝑡0 . So, it is pretty much motivated from the continuity definition of scalar
functions, and it is not that different actually.

However, here when we evaluate this mod, we have to consider the individual component,
then take the absolute value, and try to obtain this how to say criteria alright.

(Refer Slide Time: 04:42)

And for example, we can have let us say a vector function of type, I am writing 𝑟⃗(𝑡) =
𝜋
𝑓⃗(𝑡) = 𝑡 3 𝑖̂ + 4𝑡 2 𝑗̂ + sin 𝑡 𝑘̂ then verify the continuity of 𝑓⃗ at 𝑡 = 2 .

343
 𝜋 𝜋 𝜋
So, of course the function is continuous at 𝑡 = 2 , because at 𝑓⃗ ( 2 ) at 𝑡 = 2 , this basically
𝜋
𝑓⃗ ( 2 ) will get a value, then we will do that epsilon delta definition, and we can be able to

show that the value can be made arbitrary small. And here of course you can have an
interval instead of having this point, because this is an algebraic function, this is an
algebraic function, this is a trigonometrical function. So, you can even choose the interval
𝜋
as [0, 2 ],and this vector value this vector function is continuous throughout this interval as

well. So, it is a really nice behaving vector function. And therefore, it is continuous even
in that interval also.

So, now what we will do is we will define the differentiability of a vector function. So, let
us define differentiability or derivative let us write derivative of a vector function alright.
So, like function of a scalar functions, we can also differentiate the vector functions as
well, but in which context do we mean.

So, for example a function like 𝑦 = 𝑓(𝑥) = 𝑥10 is 10 times differentiable, so it is a very
nicely behaving function, basically a polynomial function. And you can be able to
differentiate at least 10 times or if because the exponent is 10. However, if we had a
1
function like, let us say 𝑦 = 𝑓(𝑥) = 𝑥 . And suppose you want to check its differentiability

at 𝑥 = 0, then of course it is not a differentiable function at 𝑥 = 0. And mainly because it

is also not continuous at 𝑥 = 0.

So, similarly we can have similar kind of situations, in case of vector functions as well.
So, what way we can define the differentiability of a vector function is the question here.
So, in case of scalar functions, you remember we used to take 𝑥 + 𝛿𝑥 , and the value of
𝑓(𝑥+𝛿𝑥)−𝑓(𝑥)
is defined as the derivative, so here also we will do the similar thing.
𝛿𝑥

So, the definition would be so suppose if we have a curve like this, so this is basically our
𝑓⃗(𝑡). So, this is the vector function, and that is the origin, and this is my vector 𝑟⃗, and this
is the point P. Now, this curve can go like this, and then this will be point Q all right the
point P and Q. And at the point P we have the point, we have the vector 𝑟⃗. And at the point
Q, we have the vector 𝑟⃗ + 𝛿𝑟⃗ all right. So, this is how we define this.

344
And let us now put everything in words. So, let 𝑟⃗ = 𝑓⃗(𝑡) be a vector function of the scalar
variable t. And we define 𝑟⃗ + 𝛿𝑟⃗ = 𝑓⃗(𝑡 + 𝛿𝑡), so that means, from here if I write 𝑟⃗ +
𝛿𝑟⃗ − 𝑟⃗ = 𝛿𝑟⃗ = 𝑓⃗(𝑡 + 𝛿𝑡) − 𝑓⃗(𝑡), so that is this small how to say movement from P to Q.
So, this small arc is our 𝛿𝑟⃗, and the 𝛿𝑟⃗ is defined in this way.

𝛿𝑟⃗ 𝑓⃗(𝑡+𝛿𝑡)−𝑓⃗(𝑡)
Now, if we consider this vector so consider the vector = .
𝛿𝑡 𝛿𝑡

(Refer Slide Time: 10:26)

𝛿𝑟⃗ 𝑓⃗(𝑡+𝛿𝑡)−𝑓⃗(𝑡)
And if I make 𝛿𝑡 → 0 so if 𝛿𝑡 → 0, then lim = lim exist. So, or if 𝛿𝑡 → 0,
𝛿𝑡→0 𝛿𝑡 𝛿𝑡→0 𝛿𝑡

so instead of then we can write such that so such that this exist, then the value of the limit
𝑑𝑟⃗
is called the derivative of the vector function 𝑓⃗, and it is denoted by 𝑑𝑡 .

So, when 𝛿𝑡 → 0, then this right hand side basically, if the limit exist is called as the
𝑑𝑟⃗
derivative of the vector function 𝑓⃗, and it is denoted by all right. So, for example, if we
𝑑𝑡
𝑑𝑟⃗
have a function let us say 𝑟⃗(𝑡) = 𝑓⃗(𝑡) = 𝑡 2 𝑖̂ + 𝑡 4 𝑗̂ − sin 𝑡 𝑘̂ , then it is derivative =
𝑑𝑡

2𝑡𝑖̂ + 4𝑡 3 𝑗̂ − cos 𝑡 𝑘̂ ok. So, we basically differentiate every component.

So, when it comes to differentiating a vector function, which is given by 𝑟⃗ = 𝑓⃗(𝑡) =

𝑑𝑟⃗ 𝑑𝑓⃗(𝑡) 𝑑𝑓 𝑑𝑓
𝑓1 (𝑡)𝑖̂ + 𝑓2 (𝑡)𝑗̂ + 𝑓3 (𝑡)𝑘̂. Then in that case the derivative is 𝑑𝑡 = 𝑑𝑡 = 𝑑𝑡1 𝑖̂ + 𝑑𝑡2 𝑗̂ +
𝑑𝑓3
𝑘̂. So, these are all functions of t.
𝑑𝑡

345
So, this is how we differentiate a vector values function. So, we differentiate every
component, and if any one of the component does not exist at a given point, then in that
case the function the vector function is not differentiable. So, suppose we are evaluating
1
the differentiability at 𝑡 = 0, and we have instead of 𝑡 4 , we have 𝑡 . Then in that case of

course, the second component is not differentiable, and therefore the overall function
vector function would not be differentiable.

So, if we want to talk about the differentiability of the vector function, we have to talk
about the differentiability of each individual component or each individual scalar function
involved in that vector function, so that is how we define the differentiability of a vector
function all right.

(Refer Slide Time: 14:24)

Now, like function of one variable or scalar functions 𝑦 = 𝑓(𝑥), we had 𝑦 ′ = 𝑓 ′ (𝑥), 𝑦 2 =
𝑓 ′′ (𝑥), 𝑦 3 = 𝑓 ′′′ (𝑥),. . . , 𝑦 𝑛 = 𝑓 𝑛 (𝑥) all right.

So, similarly we have such rule for the vector function as well. So, if we have 𝑟⃗ =
𝑑𝑟 2
𝑑 𝑟 3
𝑑 𝑟
𝑓⃗(𝑡), 𝑑𝑡 = 𝑓 ′ (𝑡), 𝑑𝑡 2 = 𝑓 ′′ (𝑡), 𝑑𝑡 3 = 𝑓 ′′′ (𝑡) and dot dot so on.

So, you can do successive derivative for the vector functions as well. So, you can
differentiate a vector functions as many times as you want, if it is differentiable that many
times all right. So, this is how we define the derivative of a vector function or successive
derivative. So, this is basically called successive derivative.

346
And like continuity or limit rule, we also have rules for some difference product of vector
functions, and they are derivative. So, theorem-2 you may call. So, let 𝑎⃗(𝑡) and 𝑏⃗⃗(𝑡) be
two differentiable vector functions.

𝑑 𝑑𝑎⃗⃗ ⃗⃗
𝑑𝑏
Then so the first rule is (𝑎⃗ ± 𝑏⃗⃗) = 𝑑𝑡 ± 𝑑𝑡 , so that means differentiation of the sum is
𝑑𝑡

equal to sum of the differentiation all right so. Secondly, if you have subtraction, so
differentiation of the difference is equal to the difference of the derivative. So, derivative
of the difference is equal to difference of the derivative.

Third is the dot product. So,

𝑑
(𝑎
⃗. ⃗⃗) = 𝑎⃗. 𝑑𝑏⃗⃗ + 𝑑𝑎⃗⃗ . 𝑏⃗⃗, it is like differentiation of the
𝑏
𝑑𝑡 𝑑𝑡 𝑑𝑡

product of two functions. So, in case of differentiation of product of two function, which
𝑑 𝑑𝑓 𝑑𝑔
was something like (𝑓(𝑥)𝑔(𝑥)) = 𝑑𝑥 𝑔 + 𝑓 𝑑𝑥 . So, we used to differentiate one
𝑑𝑥

function leaving another. And then differentiation of second function leaving the first
function. So, it is the same rule here.

And since we have a dot product dot product is commutative. So, we can play with this a
little bit, and we can arrange this dot product, so because of the commutativity of the dot
𝑑 ⃗⃗
𝑑𝑏 𝑑𝑎⃗⃗
product. We can write 𝑑𝑡 (𝑎⃗. 𝑏⃗⃗) = 𝑎⃗. 𝑑𝑡 + 𝑏⃗⃗. 𝑑𝑡 , so that is the dot product.

(Refer Slide Time: 18:02)

347
 𝑑 𝑑𝑎⃗⃗ 𝑑𝑏 ⃗⃗
And if you have let us say the cross product 𝑑𝑡 (𝑎⃗ × 𝑏⃗⃗) = 𝑑𝑡 × 𝑏⃗⃗ + 𝑎⃗ × 𝑑𝑡 . So, this is the

cross product rule. And we have to remember that here we cannot change the order,
𝑑𝑎⃗⃗
because if we write 𝑏⃗⃗ × 𝑑𝑡 , then we have to put a minus sign. So, 𝑎⃗ × 𝑏⃗⃗ = −𝑏⃗⃗ × 𝑎⃗, so that

is that is that is what we have to take care of here.

Now, similarly if we have let us say a scalar function which is multiplied by vector
function, so I have 𝜑(𝑡) which is also a differentiable scalar function. And then I have a
𝑑
vector function say 𝑎⃗(𝑡). So, then in that case it can be written as (𝜑(𝑡)𝑎⃗(𝑡)) =
𝑑𝑡
𝑑𝜑 𝑑𝑎⃗⃗
𝑎⃗(𝑡) + 𝜑(𝑡) 𝑑𝑡 all right. So, it is just that there would not be any kind of dot or cross
𝑑𝑡

here, but it is just that the function 𝜑 is multiplied in usual multiplication sense.

So, when I say usual multiplication sense, it means that if we have a 𝜑𝑎⃗(𝑡), so let us say
our function 𝑎⃗(𝑡) = 𝑡 2 𝑖̂ + 𝑡 4 𝑗̂. And my 𝜑 function is something say a constant basically
say 2, then in that case it will be a usual multiplication of these two scalar quantities. So,
it is 2𝑡 2 𝑖̂ + 𝑡 4 𝑗̂ , so it will not be like a dot product. So, dot product will usually give you
a real number actually, but in this case you will still get a vector function all right.

Now, if you have let us say instead of a scalar function and a vector function, we can also
𝑑
have (𝑎⃗ × (𝑏⃗⃗ × 𝑐⃗)let me write this formula properly. So, we can also have something
𝑑𝑡

like (𝑎⃗ × (𝑏⃗⃗ × 𝑐⃗), I cannot have (𝑎⃗. (𝑏⃗⃗. 𝑐⃗), because then 𝑏⃗⃗. 𝑐⃗ is a scalar function, and then
it is the same definition here all right.

𝑑
So, we are not writing that formula. So, here in this case, it will be (𝑎⃗ × (𝑏⃗⃗ × 𝑐⃗)) =
𝑑𝑡

𝑑𝑎⃗⃗ ⃗⃗
𝑑𝑏
× (𝑏⃗⃗ × 𝑐⃗) + 𝑎⃗ × ( 𝑑𝑡 × 𝑐⃗) all right. So, again we have to take care of the order. So, here
𝑑𝑡

𝑏⃗⃗ × 𝑐⃗, and then I have to make sure that this the cross product is here, and then we have a
cross product here. So, this is how we will do the cross product for derivative of the cross
product of three vector functions.

So, obviously we treat these two as one vector, and this one as another vector, and based
𝑑𝑎⃗⃗
on that we can do this derivative. So, basically we differentiated times 𝑏⃗⃗ × 𝑐⃗, and then
𝑑𝑡
⃗⃗
𝑑𝑏
I used instead of 𝑏⃗⃗ in this formula, I am using 𝑏⃗⃗ × 𝑐⃗. So, then our 𝑎⃗ × ( 𝑑𝑡 × 𝑐⃗) we will get

348
common, and then we do the derivative one 𝑏⃗⃗ × 𝑐⃗. And that will give us the whole cross
derivative of the cross product of 𝑎⃗ × (𝑏⃗⃗ × 𝑐⃗).

So, there are several other formulas of derivative of vector functions, we can look into any
text book. And there you can find all sorts of formulas, and for the interested readers, you
can consult any one of those vector calculus book for more formulas. I have just listed few,
because we might be using them at some point in the future, and therefore it is a good
collection to have.

Now, that we defined these formulas, we can also try to prove them, but it is just that the
proof is very straightforward, and it is also present in those textbooks. So, we will not do
that for the time being.

(Refer Slide Time: 23:09)

Next we will try to prove some basic result, so let us say remark-1. So, we know that
derivative of a constant function is 0, so that is a very well-known result in our function of
one variable. Now, what is the derivative of a constant vector so like constant function, we
also have a scalar functions, we also have constant vector functions.

So, suppose we have 𝑟⃗ = 𝑓⃗(𝑡) = 𝑐⃗ a constant function, so whose component are 𝑐1 𝑖̂ +

𝑑𝑟⃗
𝑐2 𝑗̂ + 𝑐3 𝑘̂, so what is ? So, 𝑐1 , 𝑐2 , 𝑐3 are all real numbers. So, they are all real numbers,
𝑑𝑡

and then we have to calculate the derivative of this vector function.

349
So, solution we have 𝑟⃗ = 𝑓⃗(𝑡) = 𝑐⃗. So, from here 𝑟⃗ + 𝛿𝑟⃗ = 𝑓⃗(𝑡 + 𝛿𝑡) = 𝑐⃗ . And
therefore, from here we will have 𝛿𝑟⃗ = 𝑓⃗(𝑡 + 𝛿𝑡) − 𝑓⃗(𝑡) = ⃗⃗
0.

𝛿𝑟⃗ ⃗0⃗
So, from here we will get lim = lim = lim ⃗0⃗ = ⃗0⃗. So, limit of a constant function
𝛿𝑡→0 𝛿𝑡 𝛿𝑡→0 𝛿𝑡 𝛿𝑡→0

is basically constant that particular constant, so here it is 0.

𝑑𝑟⃗
And therefore, = ⃗0⃗ which is a zero vector, I can put a vector sign here, so which is
𝑑𝑡

basically a zero vector. So, like derivative of a scalar function else derivative of a a
constant in function of in scalar functions gives us constant. In here in case of is a vector
functions, the derivative of a constant vector will give us a zero vector. So, this result a
also holds true from, what we what we have learned in our scalar functions.

(Refer Slide Time: 26:02)

Now, the next thing is the next property is we can have it again as a remark. The necessary
we can say that the necessary another remark say two the necessary and sufficient
𝑑𝑎⃗⃗
condition of a function to be constant is that ⃗⃗.
=0
𝑑𝑡

So, the necessary condition we have already proved that if we have 𝑎⃗(𝑡) = 𝑐⃗ is equals to
𝑑𝑎⃗⃗
a constant function, then from here 𝑑𝑡 = ⃗0⃗ , this is we have already seen, just like couple

of minutes ago, so this is true.

350
Now, the condition is sufficient, so that means the sufficient condition, to prove the
𝑑𝑎⃗⃗
sufficient condition. We will have 𝑑𝑡 = ⃗0⃗, now the vector function a has three components.
𝑑𝑎1 𝑑𝑎2 𝑑𝑎3
So, let us write 𝑖̂ + 𝑗̂ + 𝑘̂ = ⃗0⃗, now this is zero vector actually. So, this is
𝑑𝑡 𝑑𝑡 𝑑𝑡

basically our zero vector. So, this zero vector can be written as 0𝑖̂ + 0𝑗̂ + 0𝑘̂, where each
component is 0. So, we can write each component and 0.

(Refer Slide Time: 28:04)

And now we equate the coefficients of 𝑖̂, 𝑗̂and 𝑘̂. So, if we equate the coefficients of 𝑖̂,
𝑑𝑎1 𝑑𝑎2 𝑑𝑎3
𝑗̂ and 𝑘̂, then this will give = 0, = 0, = 0 So, from here if we integrate both
𝑑𝑡 𝑑𝑡 𝑑𝑡

sides, then we will get 𝑎1 (𝑡) = 𝑐1 , 𝑎2 (𝑡) = 𝑐2 , 𝑎3 (𝑡) = 𝑐3 constant say, so that means,
𝑎1 (𝑡), 𝑎2 (𝑡) & 𝑎3 (𝑡) are basically constant.

So, from here our vector 𝑎⃗(𝑡) = 𝑎1 (𝑡)𝑖̂ + 𝑎2 (𝑡)𝑗̂ + 𝑎3 (𝑡)𝑘̂ = 𝑐1 𝑖̂ + 𝑐2 𝑗̂ + 𝑐3 𝑘̂. Since,
𝑐1 , 𝑐2 , 𝑐3 are all constant, this is basically a constant vector. So, the condition is sufficient
as well.

So, you see like a scalar functions, if the function is given to be constant, then it is
derivative is 0. In case of vector function as well, if you have a constant vector, then it is
derivative would also be 0. And it is necessary and sufficient condition, so that means if
the derivative is 0, then the function vector function has to be a constant function. And if
the vector function is a constant function, then its derivative has to be 0.

351
So, we learned this basic result in vector calculus. And in our next class, we will try to
work out some examples on derivative of a vector function, and then we move to our next
topic. So, I will stop here for today.

And I thank you for your attention, and I look forward to you next class.

352
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 34
Successive Differentiation

Hello students. So, in the previous class, we started with the concepts of vector calculus,
where we introduced how to say idea of limit continuity and differentiability of a vector
function, like we already know those properties for a scalar function. We also prove very
basic result which is sort of like a generalization from a result on scalar function, which is
basically that if you have a constant vector then its differentiation would be 0 and if you
have a differentiation of a vector as 0 then the vector must be constant.

So, this is also true for the scalar functions. So, we tried to generalize that result for the
vector functions just to give you an idea that a lot of properties from the differentiation of
scalar functions, would be true for the differentiation of vector functions as well and today,
I am going to list some of these properties and I will try to prove at least one or two results,
and I will leave the rest of the results up to you, because they basically involve some
properties from vector algebra and differentiation of a scalar function. So, I am pretty sure
you can be able to solve the rest of the properties.

(Refer Slide Time: 01:47).

353
So, let us start with our first result, I am going to write as our next topic as properties of
differentiability, properties on differentiation. Let us call it as properties on differentiation
of vector functions. So, the first property is vector functions. So, the first property is for
today actually.

So, if 𝑎⃗ is a differentiable vector, actually differentiable vector function of the scalar

𝑑 𝑑|𝑎⃗⃗| 𝑑𝑎⃗⃗ 𝑑|𝑎⃗⃗|
variable t, then the first result is 𝑑𝑡 (𝑎⃗2 ) = 2|𝑎⃗| and the second result is 𝑎⃗. 𝑑𝑡 = |𝑎⃗| .
𝑑𝑡 𝑑𝑡

So, here mod is basically the absolute value so; that means, |𝑎⃗| = √𝑎1 2 + 𝑎2 2 + 𝑎3 2 .

So, that is what we actually mean by |𝑎⃗|. So, let us try to prove the first result. So, the first
result. So, we have here 𝑎⃗2 = 𝑎⃗. 𝑎⃗. So, that is what we mean by the square of the vector.

So, 𝑎⃗2 = 𝑎⃗. 𝑎⃗ = |𝑎⃗||𝑎⃗| cos 0 So, cos 0 is basically 1 so; that means, this is |𝑎⃗|2 . So, we
can call it as a scalar quantity. So, this is basically a scalar quantity and if we do the
𝑑
differentiation then this will be basically 2 𝑑𝑡 (|𝑎⃗|2 ) will actually convert this whole thing
𝑑|𝑎⃗⃗|
into a scalar quantity and this will be 2|𝑎⃗| .
𝑑𝑡

So, that is the first result and similarly, we can prove the second result as well. So, for the
𝑑 𝑑 𝑑
second result, we will have, 𝑑𝑡 of so, for the second result we have 𝑑𝑡 (|𝑎⃗|2 ) = 𝑑𝑡 (𝑎⃗. 𝑎⃗) =
𝑑𝑎⃗⃗ 𝑑𝑎⃗⃗
𝑎⃗. 𝑑𝑡 + 𝑑𝑡 . 𝑎⃗.Now, we know that differentiation of vector|𝑎⃗|2 is equals to this here. So,

we can use the result from section, from say from part a here. So, the left hand side of this
one will be equal to this thing.

354
(Refer Slide Time: 06:03)

𝑑|𝑎⃗⃗| 𝑑𝑎⃗⃗ 𝑑𝑎⃗⃗

So, we can write it as 2|𝑎⃗| = 𝑎⃗. 𝑑𝑡 + 𝑑𝑡 . 𝑎⃗. Now, dot product is commutative. So, this
𝑑𝑡
𝑑|𝑎⃗⃗| 𝑑𝑎⃗⃗ 𝑑𝑎⃗⃗ 𝑑𝑎⃗⃗
can be written as 2|𝑎⃗| = 𝑎⃗. 𝑑𝑡 + 𝑑𝑡 . 𝑎⃗ = 2𝑎⃗. 𝑑𝑡 . So, 2 will get cancelled and therefore,
𝑑𝑡
𝑑|𝑎⃗⃗| 𝑑𝑎⃗⃗
we will have |𝑎⃗| = 𝑎⃗. 𝑑𝑡 .
𝑑𝑡

So, this is what we needed to prove in the second result. You can also call it, let us say
|𝑎⃗| = 𝑎, without any vector notation. So, this a is basically scalar. So, instead of writing
𝑑 𝑑𝑎
|𝑎⃗|, we can replace |𝑎⃗| by a. So, then in that case this one will be (|𝑎⃗|2 ) = 2𝑎 . So,
𝑑𝑡 𝑑𝑡
𝑑𝑎
this dot is usually the multiplication dot. So, this is not a dot product. So, 𝑎 𝑑𝑡 . So, like
𝑑𝑎⃗⃗ 𝑑𝑎
scalar dot and a dot vector 𝑎⃗. 𝑑𝑡 = 𝑎 𝑑𝑡 .

So, it is better to denote the modulus or the magnitude of this vector by a scalar a. So, that
there would not be any confusion; however, these properties would still remain true for
this vector function 𝑎⃗, all right.

355
(Refer Slide Time: 07:59)

Now, the next result is or property 2 let us say is, if 𝑎⃗ has constant length, let us say constant
length or fixed magnitude, constant length means; it has fixed magnitude. Then vector 𝑎⃗
𝑑𝑎⃗⃗
and vector are perpendicular.
𝑑𝑡

So, this is the perpendicular notation are, perpendicular to each other provided, the
𝑑𝑎⃗⃗
absolute or the magnitude value | 𝑑𝑡 | ≠ 0. So, of course, the condition is sufficient as

necessary as well as sufficient. So, let us start with the fact that the 𝑎⃗ is constant. So, let 𝑎⃗
has constant length. So, if 𝑎⃗ has constant length then what does that mean? |𝑎⃗| = 𝑎 where
a is a positive real number including 0.

So, positive set of, where a is a non negative real number all right. So, and that is basically
a constant. Now, from here, we will have |𝑎⃗|2 = 𝑎2 all right and if I differentiate both
𝑑𝑎⃗⃗
sides with respect to t. So, this will be 2 𝑎⃗. 𝑑𝑡 = 0, we just saw that right. In the previous

result, we just saw that, if we differentiate, this is basically, this |𝑎⃗|2 and then from there
𝑑𝑎⃗⃗
we will basically obtain 𝑎⃗. 𝑑𝑡 = 0 right.

So, if I do that, then in that case the right hand side since, it is constant it will be 0 and
𝑑𝑎⃗⃗
therefore, from here we will have 𝑎⃗. 𝑑𝑡 = 0 and this means that these two vectors are
𝑑𝑎⃗⃗
perpendicular to one another, because from here we will get |𝑎⃗| | 𝑑𝑡 | and from here what

356
 𝑑𝑎⃗⃗
we will get? We will get |𝑎⃗| | 𝑑𝑡 | cos 𝜃 = 0 . Now, 𝑎⃗ has constant length. So, |𝑎⃗| cannot

be 0, because this is positive basically. So, |𝑎⃗| cannot be 0 and we have assumed that
𝑑𝑎⃗⃗
| 𝑑𝑡 | ≠ 0.

So, this is not 0. So, the only possibility is cos 𝜃 = 0. So, from here we will get cos 𝜃 =
𝑑𝑎⃗⃗ 𝜋
0.Since |𝑎⃗| ≠ 0 and | 𝑑𝑡 | ≠ 0 . So, from here cos 𝜃 = 0 , we basically get that 𝜃 = 2 ; that

means, these two vectors, are perpendicular to one another. So, therefore, the condition is
necessary and we can also prove that the condition is sufficient.

(Refer Slide Time: 11:29).

𝜋
So, let us go to the next slide. So, from here what we basically get is, 𝜃 = 2 . So; that means,
𝑑𝑎⃗⃗
the vectors. So, this implies that the vectors the vector 𝑎⃗ and the vector are
𝑑𝑡

perpendicular to one another. Next, we have is other way around. So, let us assume that
𝑑𝑎⃗⃗ 𝑑𝑎⃗⃗ 𝑑𝑎⃗⃗
𝑎⃗ & they are perpendicular. So, 𝑎⃗. 𝑑𝑡 = 0 and from here we will get equals to, we
𝑑𝑡 𝑑𝑡

will get from here.

𝑑|𝑎⃗⃗|
This will be equal to if I go to the previous slide. So, this will be equal to |𝑎⃗| = 0.
𝑑𝑡

357
So, from the previous result I can use this here and sorry, this here all right and this means
1 𝑑
that I can be able to write this as (|𝑎⃗|)2 = 0 and from here we will basically obtain
2 𝑑𝑡

|𝑎⃗|2 = 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 and if |𝑎⃗|2 = 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 are basically constant.

So, from here we can write |𝑎⃗| = √𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡. So, this will be again a constant, so that
means, if they are perpendicular to one another then in that case |𝑎⃗| would also be constant.
So, therefore, the condition is also sufficient so; that means, if the dot product is 0 then in
that case we can also say that the vector is of constant length.

So, of course, this condition is necessary as well as sufficient, provided this is true all right.
So, this were the two basic results that we proved here we can also try to prove some more
results, but I think they are all analogous and motivated from the vector algebra or some
differentiation of the scalar functions.

(Refer Slide Time: 14:59)

So, we will skip those, but I am going to write those properties. So, if 𝑎⃗ is a differentiable
𝑑 𝑑𝑎⃗⃗ 𝑑2 𝑎⃗⃗
function with respect to t and then we will have (𝑎⃗ × 𝑑𝑡 ) = 𝑎⃗ × 𝑑𝑡 2 mainly, because
𝑑𝑡
𝑑𝑎⃗⃗
when you take here then the 2 vectors will become parallel.
𝑑𝑡

So, this is obviously, true thus a fourth property is the necessary and sufficient condition.
It says that the necessary, the fourth property is the necessary and sufficient condition for

358
 𝑑𝑎⃗⃗
a vector 𝑎⃗(𝑡) to have constant direction is 𝑎⃗ × 𝑑𝑡 = ⃗0⃗ . So, this is also nice property for a

vector to have a constant direction.

So, that was for the constant length, this one is for the constant direction and this is very
𝑑𝑎⃗⃗
how to say obvious to prove. So, you just have to use some derivative result for 𝑎⃗ × 𝑑𝑡 and

from there you can be able to obtain this. So, it is really not complicated I can also list
some more properties for example, if you, so, if a curve is given by let us say 𝑟⃗(𝑡) = 𝑎⃗𝑡 +
2
𝑑 𝑟⃗
𝑏⃗⃗ constant then it satisfies 𝑑𝑡 2 = ⃗0⃗.

So, this is also not complicated to show. So, we just differentiate with respect to t. So,
derivative of a constant vector would be 0 and then you differentiate again with respect to
t, then you will have constant vector 𝑎⃗ that derivative of that constant vector would again
𝑑2 𝑟⃗
be 0 and then you will be basically obtain this = ⃗⃗
0.
𝑑𝑡 2

So, like that if you go through some vector calculus books which I have listed in the
references, you come across a lot of results which have basically motivated from vector
algebra and differentiation of scalar functions. So, I leave those results up to you for
practice and hopefully I will try to include some examples in your assignment sheets as
well all right.

So, next we will practice some problems from differentiation of vectors. So, these were
some results. Now, we will practice some problems, before we move on to the next topic.

359
(Refer Slide Time: 18:38)

So, let us practice a few examples. So, example 1; if our 𝑟⃗(𝑡) = (𝑡 + 1)𝑖̂ +
𝑑𝑟⃗ 𝑑2 𝑟⃗
(𝑡 2 + 𝑡 + 1)𝑗̂ + (𝑡 3 + 𝑡 2 + 𝑡 + 1)𝑘̂ then find and . So, now, we are practicing few
𝑑𝑡 𝑑𝑡 2

examples all right. So, we have a given vector function 𝑟⃗(𝑡).

So, let us write the vector function 𝑟⃗(𝑡) again, which is 𝑟⃗(𝑡) = (𝑡 + 1)𝑖̂ + (𝑡 2 + 𝑡 + 1)𝑗̂ +
(𝑡 3 + 𝑡 2 + 𝑡 + 1)𝑘̂ . So, these are all the components along the x y and z directions. Now,
𝑑𝑟⃗
we have to find out when we study about application of vector calculus.
𝑑𝑡

𝑑𝑟⃗ 𝑑2 𝑟⃗
We will get to know what does this 𝑑𝑡 mean and how similarly? What does 𝑑𝑡 2 mean, but
𝑑𝑟⃗ 𝑑2 𝑟⃗
so for at the moment, we will just take to the basic calculation of and 𝑑𝑡 2 . So, this one
𝑑𝑡

will be just a differentiation component wise with respect to t. So, if I differentiate this
one, then differentiation of 1 is 0, differentiation of t would be 1.

𝑑𝑟⃗
So, this will be = 𝑖̂ + (2𝑡 + 1)𝑗̂ + (3𝑡 2 + 2𝑡 + 1)𝑘̂ and when I differentiate again with
𝑑𝑡
𝑑 𝑟⃗ 2 𝑑 𝑟⃗ 2
respect to t. So, this will be 𝑑𝑡 2 = 0𝑖̂ + 2𝑗̂ + (6𝑡 + 2)𝑘̂ . So, 𝑑𝑡 2 = 0𝑖̂ + 2𝑗̂ + (6𝑡 + 2)𝑘̂ =

2[𝑗̂ + (3𝑡 + 1)𝑘̂] .

𝑑2 𝑟⃗
So, this is the required answer for 𝑑𝑡 2 . So, that is how you basically calculate the derivative

of the given vector function.

360
(Refer Slide Time: 21:28)

Let us practice an another example. So, suppose if 𝑟⃗(𝑡) = sin 𝑡 𝑖̂ + cos 𝑡 𝑗̂ + 𝑡𝑘̂. Then
𝑑𝑟⃗ 𝑑2 𝑟⃗ 𝑑𝑟⃗ 𝑑2 𝑟⃗
find(𝑖) 𝑑𝑡 (𝑖𝑖) 𝑑𝑡 2 (𝑖𝑖𝑖) | 𝑑𝑡 | (𝑖𝑣)| 𝑑𝑡 2 |. So, the solution would look like this, our given

function 𝑟⃗(𝑡) = sin 𝑡 𝑖̂ + cos 𝑡 𝑗̂ + 𝑡𝑘̂ .

𝑑𝑟⃗ 𝑑𝑟⃗ 𝑑𝑟⃗ 𝑑𝑟⃗

Now, , what is our ? So, would be = cos 𝑡 𝑖̂ − sin 𝑡 𝑗̂ + 𝑘̂ . So, from here
𝑑𝑡 𝑑𝑡 𝑑𝑡 𝑑𝑡
𝑑𝑟⃗
our | 𝑑𝑡 | = √𝑐𝑜𝑠 2 𝑡 + 𝑠𝑖𝑛2 𝑡 + 12 . So, 𝑐𝑜𝑠 2 𝑡 + 𝑠𝑖𝑛2 𝑡 = 1 and then 12 is 1. So, this is
𝑑2 𝑟⃗
basically, √2 and again we will calculate from here.
𝑑𝑡 2

𝑑2 𝑟⃗ 𝑑2 𝑟⃗
So, this will be = − sin 𝑡 𝑖̂ − cos 𝑡 𝑗̂ + 0𝑘̂. So, basically = − sin 𝑡 𝑖̂ − cos 𝑡 𝑗̂ and
𝑑𝑡 2 𝑑𝑡 2
𝑑2 𝑟⃗
from here we can calculate our |𝑑𝑡 2 | = √(− sin 𝑡)2 + (− cos 𝑡)2 = √1 = 1.

So, ultimately 1. So, that is the required answer and it cannot be ±1, because we are
talking about the magnitude and the magnitude cannot be negative. So, that is why we are
always taking the positive value. So, this is the required solution to this problem all right.

361
(Refer Slide Time: 24:10)

An another example of this type could be that you might be asked to evaluate something
𝑑𝑟⃗
like this. So, suppose if 𝑟⃗(𝑡) = cos 𝑛𝑡 𝑖̂ + sin 𝑛𝑡 𝑗̂, then show that 𝑟⃗ × 𝑑𝑡 = 𝑛𝑘̂ all right or
𝑑𝑟⃗
you might be asked to, so that 𝑟⃗. 𝑑𝑡 is equals to something else.

𝑑𝑟⃗
So, here we have 𝑟⃗(𝑡) = cos 𝑛𝑡 𝑖̂ + sin 𝑛𝑡 𝑗̂. So, from here we can calculate our 𝑑𝑡 , because
𝑑𝑟⃗ 𝑑𝑟⃗
we are needed to take the cross product of 𝑟⃗ and 𝑑𝑡 . So, here = −𝑛 sin 𝑛𝑡 𝑖̂ + 𝑛 cos 𝑛𝑡 𝑗̂.
𝑑𝑡
𝑑𝑟⃗
Now, we will calculate the cross product 𝑟⃗ × 𝑑𝑡 . So, this will be

𝑖̂ 𝑗̂ 𝑘̂
| cos 𝑛𝑡 sin 𝑛𝑡 0|.
−𝑛 sin 𝑛𝑡 𝑛 cos 𝑛𝑡 0

We know that how we calculate the cross product. So, we take the determinant and then
𝑑𝑟⃗
here we will have component of 𝑟⃗, the component of or the kth component is again 0.
𝑑𝑡

So, if you break this determinant then you will find that both the components 𝑖̂ and 𝑗̂ are
0, because their coefficients are 0.

So, 𝑖̂ and 𝑗̂ would not occur and simply we will have 𝑘̂(𝑛𝑐𝑜𝑠 2 𝑛𝑡 + 𝑛𝑠𝑖𝑛2 𝑛𝑡). So, if I take
n common then this will be 𝑛𝑘̂(𝑐𝑜𝑠 2 𝑛𝑡 + 𝑠𝑖𝑛2 𝑛𝑡) So, that is one and therefore, this will
be 𝑛𝑘̂ and this is what we needed to prove.

362
So, that you might come across examples, which are of this type that you are given a vector
𝑑2 𝑟⃗ 𝑑2 𝑟⃗
function and then you will be asked to evaluate 𝑟⃗ × 𝑑𝑡 2 or 𝑟⃗. 𝑑𝑡 2 or you will be given an
𝑑2 𝑟⃗ 𝑑2 𝑟⃗ 𝑑2 𝑟⃗
expression 𝑑𝑡 2 + 𝑟⃗ × 𝑑𝑡 2 + 𝑟⃗. 𝑑𝑡 2 and then you will be given an expression or you will be

asked to find a certain expression.

So, just remember that all you have to do some simple differentiation with respect to t of
the vector function 𝑟⃗ and then just calculate those expressions. So, substitute the values,
simplify them and that will give you the required answer. So, this type of examples would
not be that difficult and I am pretty sure, you can be able to solve all of these by doing
some simple differentiation.

So, we may try to solve maybe, one or two examples more little bit different than these in
our next class and I am hoping that you would find them interesting. So, thank you for
your attention for today and I look forward to your next class.

363
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture - 35
Integration of Vector Function

Hello students. So, in the previous class we started doing some examples on differentiation
of vector functions with respect to a scalar variable. In our case it is t and we tried to say
solve some problems based on simple differentiation rule and using some simple vector
algebra results. Today I will try to solve one more example which is somehow a little bit
more interesting in a way that it will give you more idea that how to differentiate a vector
function and then, we will move to our next topic.

(Refer Slide Time: 01:07)

So, let us state our problem. Example, I believe it is 4 in this section, but I am not sure
about it. So, you may have to follow the whole I am how to say tutorial. So, if 𝑟⃗ is a vector
function, the vector function of a scalar variable t and 𝑎⃗ is a constant vector m is a constant,
then differentiate the following

𝑑𝑟⃗ 𝑑𝑟⃗
(𝑎)𝑟⃗. 𝑎⃗ (𝑏)𝑟⃗ × 𝑎⃗ (𝑐)𝑟⃗ × (𝑑)𝑟⃗.
𝑑𝑡 𝑑𝑡
1 𝑑𝑟⃗ 𝑟⃗+𝑎⃗⃗ 𝑟⃗×𝑎⃗⃗
(𝑒)𝑟⃗ 2 + 2 (𝑓)𝑚( )2 (𝑔) 2 2 (ℎ)
𝑟⃗ 𝑑𝑡 𝑟⃗ +𝑎⃗⃗ 𝑟⃗.𝑎⃗⃗

364
. So, obviously we will differentiate with respect to t. So, what are we differentiating
basically?

So, we are differentiating 𝑟⃗. 𝑎⃗ . So, of course I would not solve all of these 8 sub problems,
but I will consider at least solving 3 of them. So, let us start with a solution. So, let us start
with a problem (a).

So, the problem (a) is I am going to call it as 𝑓⃗(𝑡). So, let us call this function as 𝑓⃗(𝑡)
where 𝑓⃗(𝑡) = 𝑟⃗(𝑡). 𝑎⃗, where 𝑎⃗ is a constant vector. So, if I am differentiating this function,
this is like our whole vector function. See if I am differentiating this vector function with
respect to t.

𝑑𝑓⃗ 𝑑 𝑑𝑟⃗ 𝑑𝑎⃗⃗

So, this will be = 𝑑𝑡 [𝑟⃗(𝑡). 𝑎⃗] = . 𝑎⃗ + 𝑟⃗. 𝑑𝑡 . Now, it is given that 𝑎⃗ is a constant
𝑑𝑡 𝑑𝑡

vector. So, obviously it’s derivative with respect to t would be 0. So, this will be 0 alright
and dot product of 𝑟⃗ with zero vector would again be a zero vector and if we add zero
vector here, then it would remain unchanged.

𝑑𝑓⃗ 𝑑 𝑑𝑟⃗
So, we can write it as = 𝑑𝑡 [𝑟⃗(𝑡). 𝑎⃗] = 𝑎⃗. 𝑑𝑡 and since dot product is commutative I just
𝑑𝑡
𝑑𝑟⃗ 𝑑𝑟⃗
exchange 𝑎⃗ where . 𝑎⃗ = 𝑎⃗. 𝑑𝑡 . So, this is the required dot product, that is the required
𝑑𝑡

derivative also of this function 𝑟⃗. 𝑎⃗, alright. Now, next let me consider let me consider
problem (c).

(Refer Slide Time: 05:11)

365
 𝑑𝑟⃗
So, problem (c) is let me call it again as a function 𝑓⃗(𝑡) = 𝑟⃗ × 𝑑𝑡 . So, obviously this is
𝑑𝑓⃗ 𝑑 𝑑𝑟⃗
also function of t. So, if I differentiate then this will be = 𝑑𝑡 (𝑟⃗ × 𝑑𝑡 ) and now, if I
𝑑𝑡

𝑑𝑓⃗ 𝑑 𝑑𝑟⃗ 𝑑𝑟⃗ 𝑑𝑟⃗ 𝑑2 𝑟⃗

differentiate then this will be = 𝑑𝑡 (𝑟⃗ × 𝑑𝑡 ) = × 𝑑𝑡 + 𝑟⃗ × 𝑑𝑡 2 . Now, these two are
𝑑𝑡 𝑑𝑡

same vectors. So, we know that 𝑎⃗ × 𝑎⃗, we know that 𝑎⃗ × 𝑎⃗ = ⃗0⃗ that means cross product
of a vector with itself is equal to ⃗0⃗ because they are considered as to be collinear or a
parallel. So, in a way a if two vectors are parallel, then the cross product is ⃗0⃗.

So, 𝑎⃗ × 𝑎⃗ it is considered as to be parallel and therefore, the cross product is ⃗0⃗. So, the
𝑑𝑓 ⃗
𝑑 𝑟⃗ 2
cross product of first term is ⃗0⃗ and the second term will remain as it is, 𝑑𝑡 = 𝑟⃗ × 𝑑𝑡 2 alright.

So, this is how we differentiate the term (c) or the problem (c). Now, let us consider
problem (e).

(Refer Slide Time: 06:47)

So, I am just skipping one problem, mainly because then the next problem is simpler in a
way. Now, 𝑟⃗ here 𝑓(𝑡) before I call this function as vector or scalar let me just write the
1
right-hand side. So, the right-hand side is 𝑓(𝑡) = 𝑟⃗ 2 + 𝑟⃗2 , right.

So, from vector calculus we know that if 𝑟⃗ is equals if we have 𝑟⃗ as a vector, then 𝑟⃗ 2 =
𝑟⃗. 𝑟⃗ and that is basically 𝑟⃗ 2 = 𝑟⃗. 𝑟⃗ = |𝑟⃗|2. So, instead of writing 𝑟⃗ 2 I can let |𝑟⃗|2. Now,
|𝑟⃗| is a scalar function. So, I can simply write as a 𝑟 2 . So, this |𝑟⃗| = 𝑟. So, remember this

366
scalar r and this vector 𝑟⃗, there are two different things because this is a scalar function of
t whereas, this is a vector function of t. It is just that |𝑟⃗|. That means, the magnitude or
absolute value in a way or magnitude let us call it magnitude of 𝑟⃗ is equals to a scalar r
where scalar r is also is also a function of t, but it is no longer a vector. So, there is a big
difference between a vector 𝑟⃗ and scalar r although I am using the same notation. So, that
means instead of writing |𝑟⃗|2 I can replace it with scalar 𝑟 2 .

1
So, let me write it as scalar 𝑓(𝑡) = 𝑟 2 + 𝑟 2. Now this is a scalar function of t. So, instead

of writing vector it is basically a scalar function. So, writing 𝑓(𝑡) makes sense. Now, I am
𝑑𝑓 𝑑𝑟 2 𝑑𝑟
differentiating with respect to t. So, this will be 𝑑𝑡 = 2𝑟 𝑑𝑡 − 𝑟 3 𝑑𝑡 . So, if I take 2 common
𝑑𝑓 𝑑𝑟 2 𝑑𝑟 1 𝑑𝑟
this will be = 2𝑟 𝑑𝑡 − 𝑟 3 𝑑𝑡 = 2(𝑟 − 𝑟 3 ) 𝑑𝑡 . Now, this r is basically |𝑟⃗|, I can write it
𝑑𝑡
𝑑𝑓 𝑑𝑟 2 𝑑𝑟 1 𝑑𝑟 1 𝑑|𝑟⃗|
as 𝑑𝑡 = 2𝑟 𝑑𝑡 − 𝑟 3 𝑑𝑡 = 2 (𝑟 − 𝑟 3 ) 𝑑𝑡 = 2(|𝑟⃗| − |𝑟⃗|3 ) . So, this is the required result,
𝑑𝑡

alright.

Now, which problem should be solved? Let me choose; so, this is simple; this is not
complicated. So, m is constant. So, m will come out of the differentiation and then, this
𝑑𝑟 𝑑2 𝑟
will be 2 . So, this is simple. I can consider probably (g). So, let us solve the problem
𝑑𝑡 𝑑𝑡 2

(g).

(Refer Slide Time: 09:57)

367
So, the problem (g) is let me call it as 𝑓(𝑡) first and then, we will see whether it is the
scalar function or a vector function. So,𝑟⃗ 2 + 𝑎⃗2 . So, again we have 𝑟⃗ vector r square. So,
from the previous problem we can write it as just scalar 𝑟 2 and this one is vector 𝑎⃗2 which
we can write it as simply 𝑎2 , but 𝑎⃗ is a constant vector, alright.

𝑟⃗+𝑎⃗⃗ 𝑟⃗+𝑎⃗⃗
So, this can be written as 𝑓⃗(𝑡) = 𝑟⃗2 +𝑎⃗⃗2 = 𝑟 2 +𝑎2 𝑤ℎ𝑒𝑟𝑒|𝑟⃗| = 𝑟, |𝑎⃗| = 𝑎. Both are scalar,

alright and since 𝑟⃗ is a vector,𝑎⃗ is a vector I can now write 𝑓⃗(𝑡) as a vector, alright. Now,
we can differentiate with respect to t, alright. So, the differentiation with respect to t would
go something like this. So, it would not be simple. It will be slightly lengthy to do the
𝑑𝑓⃗ 𝑑 𝑟⃗+𝑎⃗⃗
calculation. So, let us write = 𝑑𝑡 (𝑟 2 +𝑎2) and now here we will apply the product rule.
𝑑𝑡
𝑑 1 1 𝑑
So, the product rule would state that we will have ( ) (𝑟⃗ + 𝑎⃗) + 𝑟 2 +𝑎2 𝑑𝑡 (𝑟⃗ + 𝑎⃗).
𝑑𝑡 𝑟 2 +𝑎2

So, now when we differentiate this one, it will be.

(Refer Slide Time: 11:43)

2𝑟 𝑑𝑟 1 𝑑𝑟⃗ 𝑟(𝑟⃗+𝑎⃗⃗) 𝑑𝑟
So, we will have −(𝑟⃗ + 𝑎⃗) 2(𝑟 2 +𝑎2 ) 𝑑𝑡 + 𝑟 2 +𝑎2 ( 𝑑𝑡 + ⃗⃗
0). So, this will be − 𝑟 2 +𝑎2 𝑑𝑡 +
1 𝑑𝑟⃗ 1 1
. So, that means I can take common and this one will be [−𝑟(𝑟⃗ +
𝑟 2 +𝑎2 𝑑𝑡 𝑟 2 +𝑎2 𝑟 2 +𝑎2
𝑑𝑟 𝑑𝑟⃗
𝑎⃗) 𝑑𝑡 + 𝑑𝑡 ].

𝑑𝑟 𝑑𝑟⃗
So, there is a big difference between 𝑑𝑡
and the vector 𝑑𝑡
, alright. So, we should not be

confused with them. So, this is the required, notice the result in a way. So, we can leave it

368
up to there. So, it is not a matter problem, all right. So, now we have this result here. Let
me go back to the problems which we had earlier. This one can also be treated in the in the
similar fashion like we treated the problem (g).

So, you see I am just to have a look at these results and you will be able to see that in case
of vector functions you first have to make sure that the function is differentiable and if it
is a differentiable function, then all the rules for the differentiation of scalar functions hold
true in case of vector functions as well and all you have to do is to use some basic results
from vector algebra and differentiation of scalar functions to prove these results. So, it
really did not involve any kind of complicated calculations. I basically just do some simple
how to say results from vector algebra and I am hoping that you also be able to solve these
problems.

So, we will move to our next topic. So, the next topic in agenda is to introduce the concept
of vector integration. So, first of all vector function, limit, continuity, differentiability,
successive differentiation, dot product rule, cross product rule, addition, subtraction,
multiplication by scalar function and all that. So, all these properties what we have learned
basically for scalar function are holding true for vector functions as well.

So, we have seen limit continuity differentiability for vector functions. Now, from scalar
function we also have a concept motivated which is called integration. So, integration of
scalar function similarly is integration of vector function.

(Refer Slide Time: 15:29)

369
So, let us define the integration of vector function. So, let small 𝑓⃗(𝑡) is a vector and of
course, if we adding the definite vector integration in a way.

So, as sorry indefinite vector integration, then the integration of a vector would always
result in how to say a vector function and if we are doing the definite vector integration,
then the integration of a vector function would result into a vector. So, it is not like
integrating a vector function definite or indefinite integral would result in a constant scalar
or something. So, it will always end up being a vector, whether it is a vector function or
vector constant vector. That would depend what type of integration are we doing, alright.

𝑑
So, let 𝑓⃗(𝑡) and 𝐹⃗ (𝑡) be two vector functions, such that 𝑑𝑡 𝐹⃗ (𝑡) = 𝑓⃗(𝑡), then 𝐹⃗ (𝑡) is called

the indefinite integral of 𝑓⃗(𝑡) with respect to t and we write integral ∫ 𝑓⃗(𝑡)𝑑𝑡 = 𝐹⃗ (𝑡) +
𝑐⃗, where 𝑐⃗ is a constant vector, alright and yes. So, this is how we write the vector
integration for two vector functions and you can see that it is pretty much motivated or
sort of like a generalization of integration of scalar function.

So, there we also had integration ∫ 𝑓(𝑥)𝑑𝑥 = 𝐹(𝑥) + 𝑐. So, here 𝑐⃗ is a constant vector.
So, let us consider a few examples and of course, if we have a definite integral let us say
𝑏 𝑏
∫𝑎 𝑓⃗(𝑡)𝑑𝑡, then obviously we write integral ∫𝑎 𝑓⃗(𝑡)𝑑𝑡 = [𝐹⃗ (𝑡)]𝑏𝑎 . So, this will be 𝐹⃗ (𝑏) −

𝐹⃗ (𝑎) alright. So, where 𝑎 ≤ 𝑡 ≤ 𝑏. So, if we are integrating with respect to t between the
range a to b, then in that case this will be our required definite integral. So, let us solve
few examples. So, our first example if 𝑓⃗(𝑡) = (𝑡 − 𝑡 2 )𝑖̂ + 2𝑡 3 𝑗̂ − 3𝑘̂. So, then find first
2
∫ 𝑓⃗ (𝑡)𝑑𝑡 and second ∫1 𝑓⃗(𝑡)𝑑𝑡, alright.

370
(Refer Slide Time: 19:39)

So, let us solve the definite integral first. So, 𝐼 = ∫ 𝑓⃗(𝑡)𝑑𝑡 = ∫[(𝑡 − 𝑡 2 )𝑖̂ + 2𝑡 3 𝑗̂ −
3𝑘̂]𝑑𝑡 and when we are doing the integration, so the integral like the differentiation where
the differentiation went to the each component in case of integral as well the integral will
go to each component.

So, the integral will be will be done on the first component. So, ∫(𝑡 − 𝑡 2 )𝑑𝑡 𝑖̂ +
𝑡2 𝑡3 2
∫ 2𝑡 3 𝑑𝑡 𝑗̂ − ∫ 3𝑑𝑡 𝑘̂ and then, when we integrate this will result in as ( 2 − 3 ) 𝑖̂ + 4 𝑡 4 𝑗̂ −
1 𝑡 1
3𝑡𝑘̂ + 𝑐⃗. So, this is or I can simplify it a little bit. So, 𝐼 = 𝑡 2 (2 − 3) 𝑖̂ + 2 𝑡 4 𝑗̂ − 3𝑡𝑘̂ + 𝑐⃗.

So, this is the required integral for the given vector function 𝑓⃗(𝑡) and as you can see it is
basically doing the integration of three scalar functions. So, you just send your integral to
three different components the component of 𝑖̂, 𝑗̂ and 𝑘̂ and do the scale and integration,
alright.

So, this is our required integral and now we can calculate our definite integral. So,
2 1 𝑡 1
∫1 𝑓⃗(𝑡)𝑑𝑡 = [𝑡 2 (2 − 3)12 ]𝑖̂ + [2 𝑡 4 ]12 𝑗̂ − [3𝑡]12 𝑘̂.

So, ultimately value of constant at 2 and value of constant and 1 would again remain
constant and then, constant minus constant would be 0. So, 𝑐⃗ would no longer appear and
5 15
if you solve this whole thing, then you will end up getting − 6 𝑖̂ + 2
𝑗̂ − 3𝑘̂. So, this is the

required answer of this integral of 𝑓⃗(𝑡)𝑑𝑡 between the range 1 to 2. So, this is how we

371
solve the problems from my integral in the concept of integral calculus and let us solve
one more example.

(Refer Slide Time: 23:11)

So, this one is see you might be getting the idea that as we did the differentiation of vector
function, the differentiation of the integral of vector function is pretty much following the
similar rules. It is just that here we are doing the integration that is it is nothing special.
So, let us consider our next problem.

2 𝑑2 𝑟⃗
So, if 𝑟⃗(𝑡) = 5𝑡 2 𝑖̂ + 𝑡𝑗̂ − 𝑡 3 𝑘̂, then prove that integral ∫1 (𝑟⃗ × 𝑑𝑡 2 ) 𝑑𝑡 = −14𝑖̂ + 75𝑗̂ −

15𝑘̂. So, the solution would go like this. Here 𝑟⃗(𝑡) = 5𝑡 2 𝑖̂ + 𝑡𝑗̂ − 𝑡 3 𝑘̂ . So, we have to
𝑑2 𝑟⃗ 𝑑𝑟⃗
calculate 𝑑𝑡 2 . So, first we will calculate 𝑑𝑡 . So, this will be 10𝑡𝑖̂ + 𝑗̂ − 3𝑡 2 𝑘̂ , then we have
𝑑2 𝑟⃗
to calculate . So, this will be 10𝑖̂ − 6𝑡𝑘̂, right and then, we will calculate the cross
𝑑𝑡 2
𝑑2 𝑟⃗ 𝑑2 𝑟⃗
product of 𝑟⃗ and . So, cross product of 𝑟⃗ and . So, how do we calculate? So, in order
𝑑𝑡 2 𝑑𝑡 2
𝑑2 𝑟⃗
to calculate that we will simply write 𝑟⃗ × 𝑑𝑡 2 .

372
(Refer Slide Time: 25:13)

𝑑2 𝑟⃗
So, our given vector 𝑟⃗ × 𝑑𝑡 2 = (5𝑡 2 𝑖̂ + 𝑡𝑗̂ − 𝑡 3 𝑘̂) × 10𝑖̂ − 6𝑡𝑘̂ . So, calculating this one is

𝑖̂ 𝑗̂ 𝑘̂
𝑑2 𝑟⃗ 3̂
not difficult. So. 𝑟⃗ × 𝑑𝑡 2 2 ̂
= (5𝑡 𝑖̂ + 𝑡𝑗̂ − 𝑡 𝑘) × (10𝑖̂ − 6𝑡𝑘) = |5𝑡 2 𝑡 −𝑡 3 | So,
10 0 −6𝑡
ultimately this if you simplified this whole thing, then we probably end up getting an
expression.

So, we will probably end up getting an expression something like 𝑖̂(−6𝑡 2 ) +

𝑗̂[−10𝑡 3 + 30𝑡 3 ] + 𝑘̂[−10𝑡]. So, ultimately we will obtain −6𝑡 2 𝑖̂ + 20𝑡 3 𝑗̂ − 10𝑡𝑘̂ and
now we will do the integration.

So, now we will do the integration from 1 to 2, right. We were asked to evaluate the integral
2 𝑑2 𝑟⃗ 2
from 1 to 2. So,∫1 (𝑟⃗ × 𝑑𝑡 2 )𝑑𝑡. So, this will be integral ∫1 [ − 6𝑡 2 𝑖̂ + 20𝑡 3 𝑗̂ −

10𝑡𝑘̂]𝑑𝑡 and if we do the integration, then we will ultimately obtain.

373
(Refer Slide Time: 27:41)

This as [−2𝑡 3 ]12 𝑖̂ + [5𝑡 4 ]12 𝑗̂ − [5𝑡 2 ]12 𝑘̂, alright. So, integrating individual components and
now we will substitute the value evaluate and then, you will probably get −14𝑖̂ + 75𝑗̂ −
15𝑘̂.

So, this is what we needed to prove. So, you remember we had to differentiate an
expression like that in the differentiation part. So, similarly we had in this problem we
𝑑2 𝑟⃗
have to integrate this 𝑟⃗ × 𝑑𝑡 2 and this one is a definite integral problem. So, we integrated

between 1 to 2 and once we got our required integral, we just substituted the value of t at
t equals to 2 and t equals to 1 and this is the required answer which we needed to needed
to find.

So, integration of vector function is more or less similar to integration of scalar functions
and like differentiation of vector functions, we follow the similar results of integration of
scalar functions for the integration of vector functions. So, it is all of like generalizing the
concepts of scalar functions to the vector functions and we see that all of them holds true
in this case. It is just that in case of vector calculus, you are just doing the differentiation
or integration component wise.

So, we are not doing just for one function, we are doing it for the every functions involved
in each component basically. I will try to solve maybe 1 or 2 more examples based on

374
integration of vector functions and afterwards we will move to our next topic which is
basically Divergence Gradient and Car.

So, we will stop here for today and I thank you for attention. I will see you in the next
class.

375
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology Kharagpur

Lecture – 36
Gradient of a Function

Hello students, so up until last class we looked into differentiation of a vector and also
integration of a vector. So, it is basically some are motivated from the integration of scalar
𝑡=𝑏
function. So, we saw that if we have an integral of type let us say ∫𝑡=𝑎 𝑓⃗(𝑡)𝑑𝑡if we have a
vector.

(Refer Slide Time: 00:37)

𝑡=𝑏
Let us say ∫𝑡=𝑎 𝑓⃗(𝑡)𝑑𝑡 where the function 𝑓⃗(𝑡) = 𝑥(𝑡)𝑖̂ + 𝑦(𝑡)𝑗̂ + 𝑧(𝑡)𝑘̂, then that case

we basically substitute this in this vector 𝑓⃗(𝑡) as this one here in this integral and we
integrate component wise. So for example, if we have an integral of type let us say
2
∫1 𝑓⃗(𝑡)𝑑𝑡 where our 𝑓⃗(𝑡) = 𝑡 2 𝑖̂ + 2𝑡𝑗̂ + sin 𝑡 𝑘̂ . So, then in that case we just integrate we

substitute this 𝑓⃗(𝑡) here in this integral and then we integrate component wise with respect
to t and that would give us the integral of this vector function.

Now this is called as integration of a vector, but in couple of chapters later we will come
to the topic of vector integration which is a slightly different. So, we will also learn about

376
those things. Now today I am going to introduce another important topic which is quite
relevant from vector calculus point of view.

(Refer Slide Time: 02:09)

So, today’s topic is Gradient, Divergence and Curl of a vector function. So, by the way
gradient is taken on a scalar function and the divergence and curl will be taken or is taken
on a vector function. So, it is not like gradient of a vector function it is always like gradient
of a scalar function and the divergence and curl of a vector function. Now to give you
formal definition it goes like this. First of all what do we mean by how to say partial
derivative of a vector function.

So, in case of scalar functions we saw that if we have a function of 2 variables then in that
case we can talk about. So, let us say if we have a function of 2 variables 𝑧 = 𝑓(𝑥, 𝑦) then
𝜕𝑧 𝜕𝑧
in this case in scalar case we can be able to talk about and 𝜕𝑦. So, similarly in case of
𝜕𝑥

vector functions suppose our vector function 𝑟⃗ = 𝑓⃗(𝑥, 𝑦, 𝑧), then in that case here I can be
able to talk about the partial derivative of this vector function. So, how do we basically
define the partial derivative of a vector function.

So, let me give you a formal definition, so definition is suppose 𝑟⃗ is a vector function
depending on more than one variable and scalar variable of course and let 𝑟⃗ = 𝑓⃗(𝑥, 𝑦, 𝑧),
that is 𝑟⃗ be function of x y and z. So, basically 𝑟⃗ is a vector function of 3 scalar variables

377
x y and z. So, you can imagine a vector function of type 3𝑥 2 𝑧𝑖̂ + 𝑦 2 𝑧𝑗̂ + 𝑧 2 𝑥𝑘̂ So, that
can be one possible vector function.

So, any vector function of the scalar variable 𝑟⃗ can be of that particular function ok. Now
then the partial derivative of 𝑟⃗ with respect to x is given by how do we define, we define
𝛿𝑟⃗ 𝑓⃗(𝑥+𝛿𝑥,𝑦,𝑧)−𝑓⃗(𝑥,𝑦,𝑧)
it as 𝛿𝑥 = lim
𝛿𝑥→0 𝛿𝑥

. So, this is the this is how we define the partial derivative of 𝑟⃗ which is basically a vector
function of 3 scalar variables x y z and similarly we can define the partial derivative of 𝑟⃗
with respect to y, then we add a small element 𝛿𝑦 like we usually define for the scalar
function.

So, for the scalar functions also we define the derivative in the similar fashion and when
𝛿𝑟⃗
we define the derivative with respect to z we write 𝛿𝑧 and then limit 𝛿𝑧 → 0 and then we

small increment in the variable z. So, this is how we define the partial derivative of 𝑟⃗ with
respect to x, y and z. If we want to define the successive derivative.

(Refer Slide Time: 06:57)

𝜕2 𝑟⃗ 𝜕 𝜕𝑟⃗ 𝜕2 𝑟⃗ 𝜕 𝜕𝑟⃗ 𝜕2 𝑟⃗
So, I mean successive partial derivative; that means, 𝜕𝑥 2 = 𝜕𝑥 (𝜕𝑥) , 𝜕𝑦 2 = 𝜕𝑦 (𝜕𝑦) , 𝜕𝑥𝜕𝑦 =
𝜕 𝜕𝑟⃗
( ).
𝜕𝑥 𝜕𝑦

378
so this is how we define the mixed derivative all right. Now that we have introduced the
concepts of partial derivative of a vector function we can actually introduce an operator
and that operator is called as Gradient of a vector function.

So, how do we define it so let 𝑓(𝑥, 𝑦, 𝑧) be a scalar function and it is differentiable at each
point (x, y, z) in a certain region of space. Then the gradient of f written as ∇𝑓 or simply
𝜕𝑓 𝜕𝑓 𝜕𝑓 𝜕𝑓 𝜕𝑓 𝜕𝑓
grad f is defined as ∇𝑓 = 𝜕𝑥 𝑖̂ + 𝜕𝑦 𝑗̂ + 𝜕𝑧 𝑘̂, some people also write it as (𝜕𝑥 , 𝜕𝑦 , 𝜕𝑧 ).

So, instead of writing 𝑖̂, 𝑗̂ and 𝑘̂ we can write it as a triplet and this symbol here is called
𝜕 𝜕 𝜕
as nabla. So, usually in vector calculus ∇≡ 𝜕𝑥 𝑖̂ + 𝜕𝑦 𝑗̂ + 𝜕𝑧 𝑘̂. So, this is our nebula and
𝜕 𝜕 𝜕 𝜕 𝜕 𝜕
some people also use the notation of triplet, so ∇≡ 𝜕𝑥 𝑖̂ + 𝜕𝑦 𝑗̂ + 𝜕𝑧 𝑘̂ = (𝜕𝑥 , 𝜕𝑦 , 𝜕𝑧) . So,

this is the notation for nabla. So, usually when we say that gradient of 𝑓 then we are
actually of charging this operator.

So, this operator is called as nabla and when you are charging on a scalar function then it
will be called as a gradient of a scalar function and obviously it is a vector quantity.

So, gradient of a scalar function is a vector quantity and when of course, in this case you
must have all the partial derivatives to be existing for that scalar function and we can also
note that gradient of a scalar field defines a vector field, that means gradient of a scalar
point function defines a vector field actually.

So, if 𝑓 is a scalar point function which we defined earlier, then ∇𝑓 is basically a vector
point function or if 𝑓 defines a scalar field then ∇𝑓 defines a vector field alright.

379
(Refer Slide Time: 11:33)

Now, let us calculate example. So, find ⃗∇⃗𝑓 where 𝑓 is obviously a scalar function. So,
where 𝑓(𝑥, 𝑦, 𝑧) = 3𝑥 2 𝑦𝑧, so that is our given function and we need to find ⃗∇⃗𝑓 here. So,
𝜕𝑓 𝜕𝑓 𝜕𝑓 𝜕𝑓
let us do that ⃗∇⃗𝑓 = 𝜕𝑥 𝑖̂ + 𝜕𝑦 𝑗̂ + 𝜕𝑧 𝑘̂ . So now, if I do for this function then it will be
𝜕𝑥

6𝑥𝑦𝑧 because we are differentiating partially with respect to x only.

So, y and z will be treated as constant when we are differentiating with respect to x only
⃗⃗𝑓 = 6𝑥𝑦𝑧𝑖̂ + 3𝑥 2 𝑧𝑗̂ + 3𝑥 2 𝑦𝑘̂.
times ∇

So, this is the required ⃗∇⃗𝑓 and that is how we calculate. So, based on this you can calculate
the gradient of any scalar function provided it has a partial derivatives that exist ah. So,
we can consider an another example; let us say to find ⃗∇⃗𝑓 where 𝑓(𝑥, 𝑦, 𝑧) =
𝑥 3 sin 𝑦 cos 𝑧.

So here obviously, this is a differentiable function at least with respect to x, y and z

𝜕𝑓 𝜕𝑓 𝜕𝑓
individually. So, then we can write ⃗∇⃗𝑓 = 𝜕𝑥 𝑖̂ + 𝜕𝑦 𝑗̂ + 𝜕𝑧 𝑘̂.

So now, when we differentiate 𝑓 with respect to x only then it will ⃗∇⃗𝑓 =

3𝑥 2 sin 𝑦 cos 𝑧 𝑖̂ + 𝑥 3 cos 𝑦 cos 𝑧 𝑗̂ − 𝑥 3 sin 𝑦 sin 𝑧 𝑘̂ , so that is how we basically
calculate gradient of a scalar function alright.

380
Now there are certain results that are I mean how to say valid for 2 functions basically. So,
suppose if you have 2 addition of 2 functions, then if you charge the gradient on that sum
of 2 functions what would happen to the sum. So, like we did for the differentiation or a
limit and continuity for the sum of 2 functions or difference of 2 functions and things like
that. So, I would call them as properties.

(Refer Slide Time: 15:11)

So, properties of the gradient of a vector function, so what are the properties. So, statement
is let 𝑓 and 𝑔 be 2 functions of scalar variable x, y and z and ⃗∇⃗𝑓 and ⃗∇⃗𝑔 exists. So, they
are defined in a way then exists actually.

Then the first property is ⃗∇⃗(𝑓 ± 𝑔) = ⃗∇⃗𝑓 ± ⃗∇⃗𝑔. So that means, a gradient of the sum is
equal to sum of the gradient, similarly if we have a difference here, then gradient of the
difference is equal to difference of the gradient all right. So, let us write a minus sign here
I can try to prove addition and then the subtraction follows in the similar way, so let us try
to show at least one result.

𝜕 𝜕 𝜕 𝜕𝑓 𝜕𝑔 𝜕𝑓 𝜕𝑔
So, ⃗∇⃗(𝑓 + 𝑔) = (𝜕𝑥 𝑖̂ + 𝜕𝑦 𝑗̂ + 𝜕𝑧 𝑘̂)(𝑓 + 𝑔) . So, this would be 𝜕𝑥 𝑖̂ + 𝜕𝑥 𝑖̂ + 𝜕𝑦 𝑗̂ + 𝜕𝑦 𝑗̂ +
𝜕𝑓
̂ + 𝜕𝑔 𝑘̂ .
𝑘
𝜕𝑧 𝜕𝑧

𝜕𝑓
Now we gather f terms at one place and g terms at one place, so this will be (𝜕𝑥 𝑖̂ +
𝜕𝑓 𝜕𝑓
𝑗̂ ̂ ) + (𝜕𝑔 𝑖̂ + 𝜕𝑔 𝑗̂ + 𝜕𝑔 𝑘̂ ).
𝑘
𝜕𝑦 𝜕𝑧 𝜕𝑥 𝜕𝑦 𝜕𝑧

381
So, now this is basically our ⃗∇⃗𝑓 that is how we defined the ⃗∇⃗𝑓. So, I can write this as ⃗∇⃗𝑓 +
⃗∇⃗𝑔. So, I can write ⃗∇⃗𝑔 and this is the required result. So, gradient of the sum is equal to
sum of the gradient similarly we can prove the gradient of the difference is equals to the
difference of the gradient. So, the proof is not that complicated.

(Refer Slide Time: 18:35)

Now, we can also have a result or property that gradient or gradient of the product or we
also write as by nabla symbol. So, gradient of the product 𝑔𝑟𝑎𝑑(𝑓𝑔) = ⃗∇⃗𝑓𝑔 = 𝑓∇
⃗⃗𝑔 +
⃗⃗𝑓. So, it is the similarly like a rule like differentiation of 2 functions.
𝑔∇

So, the first function would remain unchanged the differentiation of the second function
and then second function would remain unchanged and the differentiation of the first
function. So, it follows very much the similar exactly the similar rule like the
differentiation of 2 functions.

So, this proof is also very easy and I am leaving this up to the students and I am pretty sure
you can be able to solve it. The third property is let 𝑓 and 𝑔 are two vector functions such
that their ⃗∇⃗𝑓 and ⃗∇⃗𝑔 exist, then we can write gradient of the quotient.

𝑓 ⃗⃗𝑓−𝑓∇
𝑔∇ ⃗⃗𝑔
⃗⃗ ( ) =
So, the gradient of the quotient would be ∇ . So, this is how we define the
𝑔 𝑔2

gradient of the quotient and of course the proof is done and that complicated to say because
here we have a quotient.

382
 1
So, if you differentiate with respect to x, y or z, then it will always be 𝑔2 and then you are

doing the differentiation for the rest of the function. So, it is a fairly simple to prove it
might be a little bit lengthy but of course, you will get to practice a little bit more in this
sense, but the proof is very straightforward so I am leaving this task up to the students.

So, they can be able to solve it I am pretty sure about it and then we have an another small
result. or the fourth property that the necessary and the sufficient condition for a scalar
point function to be constant, is that ⃗∇⃗𝑓 = ⃗0⃗. You always have to pay very close attention
that gradient is a vector function, but divergence is a scalar function which we will define
in probably in our next lecture. However, again the curl of a vector function is again a
vector.

So, divergence is a scalar function divergence of a vector function is a scalar, however the
gradient of a scalar function and curl of a vector function they are all vector quantities. So,
you have to pay very close attention that whether it is a ⃗⃗
0 or just 0 all right. So, the
necessary condition is that if we have a constant function then the gradient would be ⃗⃗
0.
So, it is very easy so necessary condition this implication. So, it is very easy to see if 𝑓 is
constant. So, let 𝑓(𝑥, 𝑦, 𝑧) = 𝑎 𝑐𝑜𝑠𝑠𝑡𝑎𝑛𝑡 = 𝐶.

𝜕𝑓 𝜕𝑓 𝜕𝑓
So, I call this constant as C, so obviously from here we will have = 𝜕𝑦 = =
𝜕𝑥 𝜕𝑧

0 because the function is constant. So, partial derivative with respect to x, y and z would
always be 0 and therefore from here if I write ⃗∇⃗𝑓 = ⃗0⃗. So, ⃗∇⃗𝑓 = ⃗0⃗ and since all these
components there is a coefficients are 0. So, 0𝑖̂ + 0𝑗̂ + 0𝑘̂ this will be actually a zero
vector. So, that condition is necessary and that has been established alright. Now let us go
the reverse way.

383
(Refer Slide Time: 22:55)

So, the reverse way suppose ⃗∇⃗𝑓 = ⃗⃗

0. So, then what would happen, what would happen is
𝜕𝑓 𝜕𝑓 𝜕𝑓
we can write it as 𝜕𝑥 𝑖̂ + 𝜕𝑦 𝑗̂ + 𝜕𝑧 𝑘̂ = ⃗0⃗ = 0𝑖̂ + 0𝑗̂ + 0𝑘̂ .

So now, this is sort of like an identity so the coefficients must be equal so; that means, the
coefficient of 𝑖̂ must be equal to coefficient of 𝑖̂ on the right hand side coefficient of 𝑗̂ on
the left hand side must be equal to coefficient of 𝑗̂ on the right hand side and similarly for
𝜕𝑓 𝜕𝑓 𝜕𝑓
the coefficient of 𝑘̂ . So, from here I can write 𝜕𝑥 = 0, 𝜕𝑦 = 0, =0 .
𝜕𝑧

So, if I integrate on both sides of these equations, so if I integrate then 𝑓 will be a function
of y z only. So, I can write 𝑓(𝑥, 𝑦, 𝑧) = 𝑔(𝑦, 𝑧), if I integrate the second equation then
𝑓(𝑥, 𝑦, 𝑧) = ℎ(𝑧, 𝑥) unless the partial derivative cannot be 0. So, unless the function 𝑓 is
a function of h and z x only the partial derivative cannot be 0 and we cannot write constant
because we do not know whether the whether the function 𝑓 is a constant or not. So, even
if the function 𝑓 is a function of z and x only it is partial derivative can still be 0 without
the function of being a constant.

So, once you integrate you should not write constant you should always write the function
of z and x all right and when we integrate 𝑓 then g h let us say I have k, then integrating
the third equation will give me 𝑓(𝑥, 𝑦, 𝑧) = 𝑘(𝑥, 𝑦). Now here we have a problem one
function cannot be equals to 3 different functions unless the function it is elf is constant
because the function cannot be just a function of y and z, it cannot be function of z and x

384
and it cannot be function of x and y. So, the only possibility for this function 𝑓 to be
existing is that 𝑓(𝑥, 𝑦, 𝑧) = 𝑎 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡 = 𝐶.

So that means, a constant function so that means, it must be it equal to to C. So, unless this
function 𝑓 is constant these 3 possibilities cannot happen, so if 𝑓 is equals how to say
𝑔(𝑦, 𝑧) if 𝑓 is a function of 𝑔(𝑦) as a function of y z, z x and x y. So, from the first solution
we can see that 𝑓 is independent of x from second equation we see that 𝑓 is independent
of y from third equation we see that 𝑓 is independent of z.

So, this one and that means that 𝑓 is independent of x, y and z and since 𝑓 is a function of
a 3 variable so the only possibility 𝑓 has is to be a constant. So, it cannot be any other
⃗⃗𝑓 = ⃗⃗
function so that means if ∇ 0 then the function has to be constant and if the function
⃗⃗𝑓 = ⃗0⃗.
is constant and obviously ∇

So, this is what we needed to prove and these are some simple results which we can
definitely practice and we will practice some more examples on gradient of a function and
we will also introduce the concept of Divergence of a vector function in our next class.

So, I thank you for your attention and I will see you in the next class.

385
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 37
Divergence & Curl

Hello students. So, in the previous class, we introduced the concepts of gradient of a scalar
function, we also talked about its properties and we also introduced a worked out one or
two examples. And in today’s lecture, we will actually continue working on the examples,
and then I will try to introduce the concept of divergence of a vector function. So, let us
work out one or two more examples before we can actually go to the divergence.

(Refer Slide Time: 00:49)

So, example let us say one for today if 𝑓(𝑥, 𝑦, 𝑧) = 3𝑥 2 𝑦 − 𝑦 3 𝑧 2 , then find gradient of f
at the point (1, −2,1). So, here our given a scalar function is this one, and we have to
calculate gradient at a certain point. So, let us calculate or let us see how we can do that.

⃗ 𝑓 = 𝜕𝑓 𝑖̂ + 𝜕𝑓 𝑗̂ + 𝜕𝑓 𝑘̂ = 6𝑥𝑦𝑖̂ +
So, gradient of f is denoted by this symbol, ∇ 𝜕𝑥 𝜕𝑦 𝜕𝑧

(3𝑥 2 − 3𝑦 2 𝑧 2 )𝑗̂ + (−2𝑦 3 𝑧)𝑘̂. So, this is the required how to say gradient of f.

And now, if you want to calculate the gradient of f at the point (1, −2,1). So, we just
⃗ 𝑓 𝑎𝑡 (1, −2,1) = (6.1. −2)𝑖̂ + 3(12 −
substitute these values here, so it will be ∇

386
(−2)2 . (−1)2 )𝑗̂ − 2(−2)3 . (−1)𝑘̂. So, we just substituted the value of x, y and z in the
gradient of f.

So, gradient of f is basically how to say is directed outward to a given surface. So, if you
have a surface f, 𝑓(𝑥, 𝑦, 𝑧) will actually signify a surface; and the gradient of f is actually
to direct it towards the normal. So, it is actually perpendicular to the surface. And at every
point you can be able to calculate this gradient of f. So, on this surface 𝑓(𝑥, 𝑦, 𝑧) gradient
of f is directed towards the normal, it is perpendicular to the surface. So, you can be able
to calculate the points where where the gradient of f or what gradient of f is perpendicular
to that surface. At a certain point this whatever we get has a calculation, so that will be
actually gradient of f that is perpendicular at the point (1, −2,1) to the surface 𝑓(𝑥, 𝑦, 𝑧).

So, if I do the calculation then we will ultimately obtain −12𝑖̂ − 9𝑗̂ − 16𝑘̂. So, there is the
required gradient of f at a certain point (1, −2,1).

(Refer Slide Time: 04:33)

Similarly we can calculate let us say if we have suppose 𝑟 = 𝑥𝑖̂ + 𝑦𝑗̂ + 𝑧𝑘̂ such that |𝑟| =
√𝑥 2 + 𝑦 2 + 𝑧 2 = 𝑟, then the first thing that we have to show ⃗∇𝑓(𝑟) = 𝑓 ′ (𝑟)∇
⃗ 𝑟.

So, let us see how we can prove that all right. So, in order to show this result, we can write
⃗∇𝑓(𝑟), now r is this how to say this here. So, 𝑟 is √𝑥 2 + 𝑦 2 + 𝑧 2 . So, I can write ⃗∇𝑓(𝑟) =

⃗ 𝑓(√𝑥 2 + 𝑦 2 + 𝑧 2 ) = 𝜕𝑓 𝑖̂ + 𝜕𝑓 𝑗̂ + 𝜕𝑓 𝑘̂ .
∇ 𝜕𝑥 𝜕𝑦 𝜕𝑧

387
 2𝑥𝑓 ′ (𝑟) 2𝑦𝑓 ′ (𝑟)
So, now, when I am doing partial derivative, it will be 𝑖̂ + 𝑗̂ +
2√𝑥 2 +𝑦 2 +𝑧 2 2√𝑥 2 +𝑦 2 +𝑧 2

2𝑧𝑓 ′ (𝑟)
𝑘̂.
2√𝑥 2 +𝑦2 +𝑧 2

1
So, here I can be able to write (𝑥𝑖̂ + 𝑦𝑗̂ + 𝑧𝑘̂)𝑓 ′ (𝑟) and then this.
√𝑥 2 +𝑦2 +𝑧 2

And similarly, 𝑓 ′ (𝑟), then this and 𝑓 ′ (𝑟), then this because this one would remain 𝑓 ′ (𝑟)
of that and then we differentiate r..

(Refer Slide Time: 08:35)

Or instead of going into this complicated calculation what we can do is we can write simply
⃗ 𝑓(𝑟) = 𝜕𝑓(𝑟) 𝑖̂ + 𝜕𝑓(𝑟) 𝑗̂ + 𝜕𝑓(𝑟) 𝑘̂. So, this will be rather must simpler. So, when we
∇ 𝜕𝑥 𝜕𝑦 𝜕𝑧
𝜕𝑟 𝜕𝑟
differentiate, we will differentiate with respect to r. So, 𝑓 ′ (𝑟) 𝜕𝑥 𝑖̂ + 𝑓 ′ (𝑟) 𝜕𝑦 𝑗̂ +
𝜕𝑟
𝑓 ′ (𝑟) 𝜕𝑧 𝑘̂ =. So, it is like product also.

𝑑𝑧 𝑑𝑢
If we have 𝑧 = 𝑓(𝑢). So, when we differentiate. So, 𝑑𝑥 = 𝑓 ′ (𝑢) 𝑑𝑥 . So, we can write it as
𝑑𝑢 𝑑𝑧 𝑑𝑢
times. So, here we can write 𝑑𝑥 = 𝑓 ′ (𝑢) 𝑑𝑥 . So, similar thing we are doing here. So, this
𝑑𝑥

one is rather much simpler. So, I can take 𝑓 ′ (𝑟) common and then this one will be
𝜕𝑟 𝜕𝑟 𝜕𝑟
𝑓 ′ (𝑟)[𝜕𝑥 𝑖̂ + 𝜕𝑦 𝑗̂ + 𝜕𝑧 𝑘̂].

388
 𝜕𝑟 𝜕𝑟 𝜕𝑟
So, this will be basically , , . So, this is our vector gradient of r all right. So,
𝜕𝑥 𝜕𝑦 𝜕𝑧

⃗ 𝑓(𝑟) = 𝑓 ′ (𝑟)∇
∇ ⃗ 𝑟 where r is basically a scalar function. So, this is this is how we prove
this identity.

(Refer Slide Time: 10:49)

And the second result will be let us go to the second result. So, let us consider we have
1 𝑟
something like ⃗∇(𝑟 ). So, what will be the result? So, it will be − 𝑟 2 . So, this is scalar
1
function r all right. Now, so if I write ⃗∇(𝑟 ) where r is a scalar function.

𝜕 1 𝜕 1 𝜕 1 1 𝜕𝑟 1 𝜕𝑟
So, this can be written as 𝑖̂ 𝜕𝑥 (𝑟 ) + 𝑗̂ 𝜕𝑦 (𝑟 ) + 𝑘̂ 𝜕𝑧 (𝑟 ). So, this will be − 𝑟 2 𝜕𝑥 𝑖̂ − 𝑟 2 𝜕𝑦 𝑗̂ −
1 𝜕𝑟
̂ . So, if I take − 1 common, it will be − 1 [𝜕𝑟 𝑖̂ + 𝜕𝑟 𝑗̂ + 𝜕𝑟 𝑘̂ ]. Now, r is basically
𝑘
𝑟2 𝜕𝑧 𝑟2 𝑟 2 𝜕𝑥 𝜕𝑦 𝜕𝑧

this r is basically our √𝑥 2 + 𝑦 2 + 𝑧 2 .

1 𝜕 𝜕
So, I can be able to write it as − 𝑟 2 [𝜕𝑥 (√𝑥 2 + 𝑦 2 + 𝑧 2 )𝑖̂ + 𝜕𝑦 (√𝑥 2 + 𝑦 2 + 𝑧 2 )𝑗̂ +
𝜕
(√𝑥 2 + 𝑦 2 + 𝑧 2 )𝑘̂] all right.
𝜕𝑧

389
(Refer Slide Time: 13:09)

So, now, when we differentiate this whole thing it will be converted into
1 𝑥 𝑦 𝑧
− 𝑟2 [ 𝑖̂ + 𝑗̂ + 𝑘̂]. Now, this is our |𝑟|. So, I can be able to
√𝑥 2 +𝑦 2 +𝑧 2 √𝑥 2 +𝑦 2 +𝑧 2 √𝑥 2 +𝑦 2 +𝑧 2
1 𝑥 𝑦 𝑧
write − 𝑟 2 [𝑟 𝑖̂ + 𝑟 𝑗̂ + 𝑟 𝑘̂ ] .

1 1
So, I can take 𝑟 common, and then this will be − 𝑟 3 (𝑥𝑖̂ + 𝑦𝑗̂ + 𝑧𝑘̂ ). And this is nothing but
𝑟
− 𝑟 3 . So, I think here it was 𝑟 3 all right. So, this one has to be 𝑟 3 . So, here we can see that

by doing just some simple partial derivative, just calculating some partial derivative, we
can be able to obtain the results.

And gradient divergence curl is nothing but just doing some partial derivatives it might get
a slightly complicated at some places, but it is just that at the end of the day we are doing
𝜕𝑓 𝜕𝑓 𝜕𝑓
some partial derivatives to calculate these , 𝑎𝑛𝑑 the three components of the
𝜕𝑥 𝜕𝑦 𝜕𝑧

given scalar function or a vector function. And with the help of which we can be able to
calculate the gradient divergence or curl of a vector function.

So, sometimes the calculation might get a little bit tedious, but you know what I mean
actually. So, here for example, in this case it is not that we cannot be able to obtain, this is
just that it was leading to a complicated situation. So, I decided that we can be able to do
it in a much simpler way than doing it in a complicated way. So, then we considered this.

390
So, this is an alternative method. And with this method, we can be able to solve the
problem one in a rather simpler way than going into a complicated way all right. And this
is our example 2. So, there can be some other forms to show that instead of this a gradient
1
of 𝑟 , we may be having something like I do not know let me give you an another example.

(Refer Slide Time: 15:57)

⃗ 𝑟 𝑛 = 𝑛𝑟 𝑛−2 𝑟. So, here you just have to do some

So, third example could be calculate ∇
partial derivative play with the exponent of n, there is nothing much complicated here. It
will just have to be really good at calculating these partial derivatives and that will be
pretty much it. So, I am leaving this problem up to the students; I am pretty sure you can
be able to solve it.

And we will move on to our next problem or yes. So, we will move on to our next how to
say topic, and this is basically how to say divergence sorry. So, the next topic is divergence
of a vector function.. Now, how do we define the divergence of a vector function, let me
get the how to say formal definition.

⃗ be a differentiable vector function, then the

So, the divergence of a vector function is let 𝑉
⃗ is given by we write it with the same nabla symbol, ∇
divergence of 𝑉 ⃗ .𝑉
⃗ . So, here since
⃗ as a vector function when we are charging this operator nabla which is itself a
we have 𝑉
vector quantity, we have to have some kind of how to say operation here. So, either it is a
dot product or either it is a cross product because that is how we how to say multiply or

391
come or charge one vector on another. So, either it has to be a dot product or it has to be a
cross product all right.

So, divergence deals with the dot product. And when we learned about curl then in that
⃗ = ( 𝜕 𝑖̂ + 𝜕 𝑗̂ + 𝜕 𝑘̂) . 𝑉
case it will be a cross product all right. So, 𝑑𝑖𝑣𝑉 ⃗.
𝜕𝑥 𝜕𝑦 𝜕𝑧

⃗ since it is a vector quantity or vector function, it also must have three components
So, 𝑉
𝜕 𝜕 𝜕 𝜕𝑉1 𝜕𝑉2 𝜕𝑉3
let me call them as (𝜕𝑥 𝑖̂ + 𝜕𝑦 𝑗̂ + 𝜕𝑧 𝑘̂) . (𝑉1 𝑖̂ + 𝑉2 𝑗̂ + 𝑉3 𝑘̂) = + + . So,
𝜕𝑥 𝜕𝑦 𝜕𝑧

ultimately we are getting a scalar function. So, this is a scalar function so that means,
gradient of a scalar function is a vector quantity, whereas the divergence of a vector
⃗ is actually a scalar
function is a scalar quantity. So, here we see that the divergence of 𝑉
function all right. And whatever we get after calculating that is a different thing.

(Refer Slide Time: 20:09)

Now we can also define, the curl of a vector. So, as I was saying that when we are charging
⃗ be a vector function, then the curl of 𝑓
nabla on a vector quantity, so let us call it as let 𝑉
is defined as. So, here in this case also the new operator nabla will be charged on the vector
⃗.
𝑉

But since nabla is a vector quantity charging on a vector function, there has to be some
kind of operator or operation acting in between. Now, the first operation which is basically
dot product is resolved for the divergence. So, the second operation that is left is basically

392
 ⃗ is called as the curl. And this is nothing, but (𝑖̂ 𝜕 +
⃗ ×𝑉
this cross product. And this ∇ 𝜕𝑥
𝜕 𝜕
𝑗̂ 𝜕𝑦 + 𝑘̂ 𝜕𝑧) × (𝑉1 𝑖̂ + 𝑉2 𝑗̂ + 𝑉3 𝑘̂).

𝑖̂ 𝑗̂ 𝑘̂
𝜕 𝜕 𝜕
So, from here we can calculate the cross product as | |. So, if we break this
𝜕𝑥 𝜕𝑦 𝜕𝑧
𝑉1 𝑉2 𝑉3
𝜕𝑉 𝜕𝑉2 𝜕𝑉 𝜕𝑉3 𝜕𝑉 𝜕𝑉
course or this determinant, then this will be 𝑖̂ ( 𝜕𝑦3 − ) + 𝑗̂ ( 𝜕𝑧1 − ) + 𝑘̂( 𝜕𝑥2 − 𝜕𝑦1 ).
𝜕𝑧 𝜕𝑥

⃗ whose components are 𝑉1 . 𝑉2 , 𝑉3.

So, this is the required curl of a vector function 𝑉

So, in case of scalar function when we are dividing the gradient, we are obtaining a scalar
function a vector function. But in case of divergence, it is a scalar function. And again in
case of curl, it is a vector function. So, this is how we define the curl of a vector.

(Refer Slide Time: 22:59)

And we have two quick definitions. So, the first definition is solenoidal vector. So, a vector
⃗ is said to be solenoidal if 𝑑𝑖𝑣𝑉
𝑉 ⃗ = 0 . And conservative vector and then we have another
⃗ is said to be irrotational if
definition which is basically irrotational vector. So, a vector 𝑉
⃗ = ⃗0 all right. So, we have a solenoidal vector field and irrotational vector fields.
⃗∇ × 𝑉

⃗ 𝑑𝑖𝑣𝑉
So, if 𝑉 ⃗ = ⃗0 , then it will
⃗ = 0 , then we is said to be solenoidal vector. And if ⃗∇ × 𝑉
be called as irrotational and from divergence and gradient, we can also define a very

393
important operator in vector calculus or in partial differential equations which is something
of very important.

So, what happens if you charge let us say divergence on the gradient of a scalar function.
So, what would this be f is a scalar function. all right. So, what would happen let us see.
⃗ 𝑓) = ⃗∇. (𝜕𝑓 𝑖̂ + 𝜕𝑓 𝑗̂ + 𝜕𝑓 𝑘̂)the first gradient would give us this that is pretty
So, ⃗∇. (∇ 𝜕𝑥 𝜕𝑦 𝜕𝑧

much straightforward.

𝜕𝑓 𝜕𝑓 𝜕𝑓
Now, let us break this gradient here. So, 𝜕𝑥 𝑖̂ + 𝜕𝑦 𝑗̂ + 𝜕𝑧 𝑘̂ . So, now, we have a dot product.
𝜕 𝜕
And if we have a dot product, then the dot product would look like (𝜕𝑥 𝑖̂ + 𝜕𝑦 𝑗̂ +
𝜕 𝜕𝑓 𝜕𝑓 𝜕𝑓
𝑘̂) . (𝜕𝑥 𝑖̂ + 𝜕𝑦 𝑗̂ + 𝜕𝑧 𝑘̂), because this partial derivative will be charged here. So, this will
𝜕𝑧

𝜕2 𝑓 𝜕2 𝑓 𝜕2 𝑓
be 𝜕𝑥 2 + 𝜕𝑦 2 + 𝜕𝑧 2 .

And this second order partial derivative of f with respect to x, y and z, can simply be
written as ∇2 𝑓or some people write it as ∆𝑓. So, sometimes it is called as delta. This and
this is nothing but a Laplace operator. So, this ∆, or this ∇2 is called as the Laplace
operator. And the Laplace equation, so when you have ∇2 𝑓 = 0 or ∇2 𝑓 = 𝑔 or whatever
function you have then such equations are called as Laplace equation which is very
important in diffusion reaction equations in and when they often comes as a Poisson’s
equation.

So, basically it has a very how to say deep applied mathematical connection. So, this
Laplace or ∇2 or ∆ is an important operator and not only in the vector calculus, but in other
⃗ . as a divergence on
fields as well. And how do we obtain it we basically charged this ∇
⃗ 𝑓 a vector function, so this one here. And this is our required
the which is actually ∇
Laplace operator.

So, now, that we have introduced the concepts of divergence and curl. In our next class,
we will practice some more examples motivated from divergence and curl of a vector and
we will also try to show when a vector field is conservative, when a vector field is
solenoidal or irrotational, and then we move to our next topic which is level surfaces and
directional derivative.

394
So, I thank you for your attention. And I will see you in the next class.

395
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 38
Divergence & Curl Examples

Hello students. So, in the last class we introduced the concepts of gradient of a scalar
function and Divergence and Curl of a vector function. We also worked out few examples
on a gradient of a scalar function; how do you calculate it and then we gave definitions of
divergence and curl of a vector function.

Today we will practice few more examples, but basically, we will start with the examples
on divergence and the curl of a vector function and we will see how we calculate them and
there are some results which are sort of important from these two concepts. So, let us start
with our a how to say examples on divergence and curl.

(Refer Slide Time: 01:09)

⃗⃗. 𝑓⃗ = 𝜕𝑓1 + 𝜕𝑓2 +

So, yesterday we introduced that if 𝑓⃗ is a vector function, then 𝑑𝑖𝑣𝑓⃗ = ∇ 𝜕𝑥 𝜕𝑦
𝜕𝑓3
where 𝑓⃗ is a vector function and it has three components 𝑓⃗=(𝑓1 , 𝑓2 , 𝑓3 ) and each
𝜕𝑧

𝑓1 , 𝑓2 , 𝑓3 are functions of x, y and z all right.

396
 𝜕𝑓 𝜕𝑓 𝜕𝑓 𝜕𝑓 𝜕𝑓 𝜕𝑓
Similarly the 𝑐𝑢𝑟𝑙𝑓⃗ = 𝑖̂ ( 𝜕𝑦3 − 𝜕𝑧2 ) + 𝑗̂ ( 𝜕𝑧1 − 𝜕𝑥3 ) + 𝑘̂( 𝜕𝑥2 − 𝜕𝑦1 ). You can be able to

calculate with the help of that determinant. So, this is basically our how to say curl of a
vector function 𝑓⃗ .

Now motivated from these 2, this how to say formulas or definitions, we will first start
with our example one. So, let 𝑟⃗ = 𝑥𝑖̂ + 𝑦𝑗̂ + 𝑧𝑘̂. So, when I write 𝑖̂ so; that means, it is a
⃗⃗. 𝑟⃗ and ∇
unit vector all right and then find ∇ ⃗⃗ × 𝑟⃗ right. So, let us try to find. So, solution let

us try to find ⃗∇⃗. 𝑟⃗. So, ⃗∇⃗. 𝑟⃗ would go like this we know that from the formula 𝑟⃗ has 3
𝜕𝑟1 𝜕𝑟2 𝜕𝑟3
components; so, 𝑟1 , 𝑟2 , 𝑟3 . So, I can be able to write this formula + + now 𝑟1 is
𝜕𝑥 𝜕𝑦 𝜕𝑧
𝜕𝑥 𝜕𝑦 𝜕𝑧
⃗⃗. 𝑟⃗ = 3 if 𝑟⃗ is given
basically our x. So (𝜕𝑥 + 𝜕𝑦 + 𝜕𝑧) = 1 + 1 + 1 = 3,. So, ultimately ∇

as this vector here.

⃗⃗ × 𝑟⃗ = 𝑖̂ (𝜕𝑟3 − 𝜕𝑟2 ) +
⃗⃗ × 𝑟⃗. So, if I calculate the ∇
Now, we will calculate the ∇ 𝜕𝑦 𝜕𝑧
𝜕𝑟 𝜕𝑟3 𝜕𝑟 𝜕𝑟
𝑗̂ ( 𝜕𝑧1 − ) + 𝑘̂( 𝜕𝑥2 − 𝜕𝑦1 ) . If you really want to, then you can say let 𝑟⃗ = 𝑥𝑖̂ + 𝑦𝑗̂ +
𝜕𝑥

𝑧𝑘̂ = 𝑟1 𝑖̂ + 𝑟2 𝑗̂ + 𝑟3 𝑘̂ result that you can do.

𝜕𝑧 𝜕𝑦 𝜕𝑥 𝜕𝑧 𝜕𝑦 𝜕𝑥
So, now, ⃗∇⃗ × 𝑟⃗ = 𝑖̂ (𝜕𝑦 − 𝜕𝑧 ) + 𝑗̂ (𝜕𝑧 − 𝜕𝑥) + 𝑘̂ (𝜕𝑥 − 𝜕𝑦) = 𝑖̂(0 − 0) + 𝑗̂(0 − 0) +

𝑘̂(0 − 0) = ⃗0⃗ .

So, ultimately we will get a ⃗⃗

0. So that means, if 𝑟⃗ is equals to this vector here than in that
⃗⃗ × 𝑟⃗ = ⃗0⃗. So,
⃗⃗. 𝑟⃗ = 3 ∇
case our divergence and curl would be given in this fashion. So, ∇
that is how we basically calculate this a divergence and curl let us consider a few more
examples.

397
(Refer Slide Time: 05:27)

So, just so we can be thorough in this thing. So, example let us say 2 and the example
states that if our vector function 𝑓⃗(𝑥, 𝑦, 𝑧) = 𝑥 2 𝑦𝑖̂ − 2𝑥𝑧𝑗̂ + 2𝑦𝑧𝑘̂ then find first
(𝑖)𝑑𝑖𝑣𝑓⃗ (𝑖𝑖)𝑐𝑢𝑟𝑙𝑓⃗ (𝑖𝑖𝑖)𝑐𝑢𝑟𝑙(𝑐𝑢𝑟𝑙𝑓⃗) . So, the thing is here curl is basically curl of a
vector function is again a vector quantity.

So, we can talk about its curl again. So, unless the quantity is not a vector function, we
cannot talk about its curl. But as soon as soon as the quantity becomes or a function
becomes a vector function will not becomes if it is a vector function, then in that case we
can always talk about its curl and that is what is that is what is happening in this case.

𝑓⃗ is a vector function, but once we charge the curl it again the else a vector function and
therefore, we can again charge curl on it which will again give us a vector function and we
can again charge curl on it. So, as long as the function or the I mean we have a vector
function, we can always talk about its curl. However, in case of divergence, we cannot
charge divergence again because the divergence of f would yield of scalar function

So, once we get a scalar function we can talk about its gradient, but we cannot talk about
its divergence. So, what could happen is, when we charge divergence on a vector function,
it would give us a scalar function and then we charge a gradient. And once we charge the
gradient, then it would become a vector function and then we can again charge gradient as
a divergence and then it will again yield a scalar function.

398
So, divergence gradient divergence gradient and so, on; however, in case of curl every
time we can charge a curl on it. So, these are some small aspects of vector calculus and I
am pretty sure you may have got an idea that, how to deal with these operators all right.
So, let us start calculating the first how to say 𝑑𝑖𝑣𝑓⃗.

So, solution divergence of f or let me start by writing so just start with writing here. So, I
am just writing here 𝑓1 (𝑥, 𝑦, 𝑧) = 𝑥 2 𝑦, 𝑓2 (𝑥, 𝑦, 𝑧) = −2𝑥𝑧, 𝑓3 (𝑥, 𝑦, 𝑧) = 2𝑦𝑧 all right.
So, I just wrote the 3 components just for our how to say seek. Now we have 𝑑𝑖𝑣𝑓⃗ =
⃗∇⃗. 𝑓⃗ = 𝜕 (𝑥 2 𝑦) + 𝜕 (−2𝑥𝑧) + 𝜕 (2𝑦𝑧) = 2𝑥𝑦 + 0 + 2𝑦 = 2𝑦(𝑥 + 1).
𝜕𝑥 𝜕𝑦 𝜕𝑥

and that is the required 𝑑𝑖𝑣𝑓⃗ all right. So, this is what we obtained as a 𝑑𝑖𝑣𝑓⃗.

Now, we can calculate the 𝑐𝑢𝑟𝑙𝑓⃗. So, 𝑐𝑢𝑟𝑙𝑓⃗ I can write it as ∇

⃗⃗ × 𝑓⃗ . I prefer to write in

𝜕
terms of there is operators instead of writing in words. Now we have 𝑖̂ (𝜕𝑦 (2𝑦𝑧) −

𝜕 𝜕 𝜕 𝜕 𝜕
(−2𝑥𝑧)) + 𝑗̂ ( (𝑥 2 𝑦) − (2𝑦𝑧)) + 𝑘̂ ( (−2𝑥𝑧) − (𝑥 2 𝑦)).
𝜕𝑧 𝜕𝑧 𝜕𝑥 𝜕𝑥 𝜕𝑧𝑦

all right. Now doing this partial derivatives is very simple and I leave that task up to the
students.

(Refer Slide Time: 11:21)

399
So, here at the end you will basically obtain your answer as (2𝑧 + 2𝑥)𝑖̂ − 0𝑗̂ + (−2𝑧 −
𝑥 2 )𝑘̂. So, ultimately we will obtain 2(𝑧 + 𝑥)𝑖̂ − (𝑥 2 + 2𝑧)𝑘̂. So, this is our required
⃗∇⃗ × 𝑓⃗. So, I can write it as ⇒ ⃗∇⃗ × 𝑓⃗ = (2𝑧 + 2𝑥)𝑖̂ − 0𝑗̂ + (−2𝑧 − 𝑥 2 )𝑘̂.

So, ⃗∇⃗ × 𝑓⃗ and our problem 3 is 𝑐𝑢𝑟𝑙(𝑐𝑢𝑟𝑙𝑓⃗) right so; that means, we basically have
⃗⃗ × 𝑓⃗) = ⃗∇⃗ × [2(𝑥 + 𝑧)𝑖̂ − (𝑥 2 + 2𝑧)𝑘̂]. So, here we have to identify our 𝑓1 , 𝑓2 and
⃗∇⃗ × (∇

𝑓3 and then put everything back in this formula, and we would obtain our required result
all right. So, I am leaving that task up to the students.

So, doing this is not complicated so, you can we can identify the 3 components and just
follow the previous formula and that will give you the required result all right.

(Refer Slide Time: 13:02)

Now we will consider a slightly different types of examples. So, let us start with them. So,
here this says that determine the constant a so that the following vector field is solenoidal.
⃗⃗ (𝑥, 𝑦, 𝑧) = (𝑥 + 3𝑦)𝑖̂ + (𝑦 − 2𝑧)𝑗̂ + (𝑥 + 𝑎𝑧)𝑘̂.
So, what is that? So, 𝑉

So, we have to determine this a here so, that this vector field will be a solenoidal all right.
So, this is our given vector field. Now we know that from the definition in the previous
⃗⃗ = 0 so; that means, we have to make.
class that a vector field is solenoidal when 𝑑𝑖𝑣𝑉
⃗⃗ = 0 and from there
So, we assume that the vector field is solenoidal and we make 𝑑𝑖𝑣𝑉
⃗⃗ = 0 actually occurs all right.
we have to find out that for what value of a that 𝑑𝑖𝑣𝑉

400
So, we basically solve for a equals to some constant and that will be the required how to
say the answer or the required value of a for which this vector field is solenoidal. So, let
⃗⃗ . So, here we can again identify our 3
us start with calculating the divergence of 𝑉
components 𝑉1 = 𝑥 + 3𝑦, 𝑉2 = 𝑦 − 2𝑧, 𝑉3 = 𝑥 + 𝑎𝑧 . So, each component is a function of
x, y, z. The correct way to write is actually 𝑉1 (𝑥, 𝑦, 𝑧) all right; however, you are allowed
to write 𝑉1 only as well, but the correct way to write is 𝑉1 (𝑥, 𝑦, 𝑧) because it is actually a
function of x, y, z each component basically.

⃗⃗ . So, 𝑑𝑖𝑣𝑉
Now we calculate 𝑑𝑖𝑣𝑉 ⃗⃗ = ∇ ⃗⃗ = 𝜕𝑉1 + 𝜕𝑉2 + 𝜕𝑉3 = 1 + 1 + 𝑎 = 𝑎 + 2. Now
⃗⃗. 𝑉
𝜕𝑥 𝜕𝑦 𝜕𝑧

⃗⃗ were a solenoidal, then 𝑑𝑖𝑣𝑉

if 𝑉 ⃗⃗ = 0 because that is a definition and now if 𝑑𝑖𝑣𝑉
⃗⃗ = 0
⃗⃗ is a
; that means, 𝑎 + 2 = 0. So, from here we will obtain 𝑎 = −2. So, for 𝑎 = −2 𝑉
solenoidal field all right.

So, after calculating the versions of we will start with the assumption that we have a
⃗⃗ = 0 and from there we calculate 𝑎 = −2 all
solenoidal field and then in that case 𝑑𝑖𝑣𝑉
right. So, this is our required value for a all right. Now let us move on to our next example.

(Refer Slide Time: 17:03)

So, our next example says that show that the following vector is irrotational. So, our given
⃗⃗ (𝑥, 𝑦, 𝑧) = (sin 𝑦 + 𝑧)𝑖̂ + (𝑥 cos 𝑦 − 𝑧)𝑗̂ + (𝑥 − 𝑦)𝑘̂. So, we have a given
vector is 𝑉
vector function or a vector field and we have to show that this vector function is actually

401
irrotational. So, from the definition of irrotational which we learnt in the previous lecture,
our vector function is irrotational when that curl of that vector function is ⃗0⃗.

⃗⃗ = ⃗0⃗ and then it will be a irrotational field. So, let us

So, here we have to show that 𝑐𝑢𝑟𝑙𝑉
first calculate or identify the 3 components and then we calculate curl of f all right. So,
here our 𝑉1 (𝑥, 𝑦, 𝑧) = sin 𝑦 + 𝑧, 𝑉2 (𝑥, 𝑦, 𝑧) = 𝑥 cos 𝑦 − 𝑧, 𝑉3 (𝑥, 𝑦, 𝑧) = 𝑥 − 𝑦 all right.
⃗⃗. So, 𝑐𝑢𝑟𝑙𝑉
Now we calculate the 𝑐𝑢𝑟𝑙𝑉 ⃗⃗ = 𝑖̂ [ 𝜕 (𝑥 − 𝑦) − 𝜕 (𝑥 cos 𝑦 − 𝑧)] +
⃗⃗ = ⃗∇⃗ × 𝑉
𝜕𝑥 𝜕𝑧
𝜕 𝜕 𝜕 𝜕
𝑗̂ [𝜕𝑧 (sin 𝑦 + 𝑧) − 𝜕𝑥 (𝑥 − 𝑦)] + 𝑘̂[𝜕𝑥 (𝑥 cos 𝑦 − 𝑧) − 𝜕𝑦 (sin 𝑦 + 𝑧)]

⃗⃗ = ∇
is not it all right. So, now, when we calculate so, 𝑐𝑢𝑟𝑙𝑉 ⃗⃗ = 𝑖̂ [ 𝜕 (𝑥 − 𝑦) −
⃗⃗ × 𝑉
𝜕𝑥
𝜕 𝜕 𝜕 𝜕 𝜕
(𝑥 cos 𝑦 − 𝑧)] + 𝑗̂ [ (sin 𝑦 + 𝑧) − (𝑥 − 𝑦)] + 𝑘̂ [ (𝑥 cos 𝑦 − 𝑧) − (sin 𝑦 +
𝜕𝑧 𝜕𝑧 𝜕𝑥 𝜕𝑥 𝜕𝑦

𝑧)] = 𝑖̂[−1 + 1] + 𝑗̂[1 − 1] + 𝑘̂[cos 𝑦 − cos 𝑦]

So, basically we have 0𝑖̂ + 0𝑗̂ + 0𝑘̂. So, ultimately we obtain a ⃗0⃗. So, from here, we have
⃗⃗ = ⃗0⃗ and this implies that 𝑉
𝑐𝑢𝑟𝑙𝑉 ⃗⃗ = ⃗0⃗ , then
⃗⃗ is an irrotational vector field. So, if the 𝑐𝑢𝑟𝑙𝑉
⃗⃗ is
the vector field is said to be irrotational and we proved that the given vector function 𝑉
actually irrotational.

⃗⃗ = 0, then it is solenoidal
So, all you have to do is remember these formulas that if the 𝑑𝑖𝑣𝑉
⃗⃗ = ⃗⃗
and if 𝑐𝑢𝑟𝑙𝑉 0, then it is a irrotational and you just have to verify or if you were asked
to calculate the value of a certain parameter for which it is either divergence or a either
solenoidal or irrotational, then you basically do what we did in the previous example all
right.

Now, let us just consider a few more examples; some remarks are also to be addressed.

402
(Refer Slide Time: 22:26)

So, let us consider our first remark suppose 𝑓⃗(𝑥, 𝑦, 𝑧) equals to a constant vector all right;
then find ⃗∇⃗. 𝑓⃗ and ⃗∇⃗ × 𝑓⃗. So, the thing is solution or the thing is if 𝑓⃗ is a constant, then in
that case when you charge this in nabla operator nabla operator has partial derivatives
𝜕 𝜕 𝜕
,. , Now how it has been charged depending on whether you are charging the
𝜕𝑥 𝜕𝑦 𝜕𝑧

divergence or whether you are charging the curl, but whatever that vector function is it
will always be differentiated with a with respect to x ,y and z.

So, if your 𝑓⃗ is a constant function I mean, it will always be differentiated with respect to
x ,y and z and the minute you differentiate every component of that constant vector, it will
lead to a ⃗0⃗ because a constant vector differentiation of it will always be 0. So, regardless
whether you are calculating divergence or curl, the answer would always be 0.

In the first case it will be scalar 0 and in the second case, it would be a ⃗0⃗ because in case
of curl we get a vector whose every component is actually 0. So, it is very simple to see
and I am leaving this result up to the students. So, divergence of 𝑓⃗ is actually 0 and curl
of 𝑓⃗ is actually a ⃗0⃗. So, these 2 things are quite straightforward to see and I am leaving
these up to the students all right.

403
Next we consider an example. Evaluate ⃗∇⃗ × 𝑓⃗ where our 𝑓⃗ (𝑥, 𝑦, 𝑧), it is an little bit
interesting function; it is a slightly interesting function. So, 𝑓⃗ (𝑥, 𝑦, 𝑧) = 𝑒 𝑥𝑦𝑧 (𝑖̂ + 𝑗̂ +

𝑘̂) so, here we have to calculate the 𝑐𝑢𝑟𝑙𝑓⃗ .

So, let us go to the solution. So, 𝑐𝑢𝑟𝑙𝑓⃗ we know that the formula. So, here our 3
components are 𝑒 𝑥𝑦𝑧 , 𝑒 𝑥𝑦𝑧 , 𝑒 𝑥𝑦𝑧 because they are all same. So, I can write here
𝑓1 (𝑥, 𝑦, 𝑧) = 𝑒 𝑥𝑦𝑧 , 𝑓2 (𝑥, 𝑦, 𝑧) = 𝑒 𝑥𝑦𝑧 , 𝑓3 (𝑥, 𝑦, 𝑧) = 𝑒 𝑥𝑦𝑧 . So, now, we calculate 𝑐𝑢𝑟𝑙𝑓⃗.
𝜕 𝜕 𝜕 𝜕
So, 𝑐𝑢𝑟𝑙𝑓⃗ = ⃗∇⃗ × 𝑓⃗ = 𝑖̂ [𝜕𝑦 (𝑒 𝑥𝑦𝑧 ) − 𝜕𝑧 (𝑒 𝑥𝑦𝑧 )] + 𝑗̂ [𝜕𝑧 (𝑒 𝑥𝑦𝑧 ) − 𝜕𝑥 (𝑒 𝑥𝑦𝑧 )] +
𝜕 𝜕
𝑘̂ [𝜕𝑥 (𝑒 𝑥𝑦𝑧 ) − 𝜕𝑦 (𝑒 𝑥𝑦𝑧 )].

So, now, when we differentiate 𝑒 𝑥𝑦𝑧 with respect to y so, this will yield 𝑖̂(𝑒 𝑥𝑦𝑧 𝑥𝑧 −
𝑒 𝑥𝑦𝑧 𝑥𝑦) + 𝑗̂(𝑒 𝑥𝑦𝑧 𝑥𝑦 − 𝑒 𝑥𝑦𝑧 𝑦𝑧) + 𝑘̂(𝑒 𝑥𝑦𝑧 𝑦𝑧 − 𝑒 𝑥𝑦𝑧 𝑥𝑧). So, now, we will take it to the
power x, why is that common? So, here it will be x times e to the power.

(Refer Slide Time: 27:35)

So, this will be basically 𝑒 𝑥𝑦𝑧 and then we have 𝑒 𝑥𝑦𝑧 [𝑥(𝑦 − 𝑧)𝑖̂ + 𝑦(𝑧 − 𝑥)𝑗̂ + 𝑧(𝑥 − 𝑦)𝑘̂.
Did I do it correctly? all right.

So, this will be (𝑦 − 𝑥) this will be (𝑦 − 𝑥) and this will be (𝑥 − 𝑧).So, let us correct it.
So, 𝑒 𝑥𝑦𝑧 [𝑥(𝑦 − 𝑧)𝑖̂ + 𝑦(𝑥 − 𝑧)𝑗̂ + 𝑧(𝑦 − 𝑥)𝑘̂ all right; this is the required 𝑐𝑢𝑟𝑙𝑓⃗ for this

404
given vector function 𝑓⃗. Now today we try to cover as many examples as possible. We
will probably continue with our examples in our next class and I look forward to it.

Thank you for attention.

405
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 39
Divergence & Curl important Identities

Hello students. So, in the previous class we were solving some examples on Divergence
and a Curl of a vector function. Today we will continue with those examples and we will
also address some important in some way they are important because you have say let say
dot product of two vectors functions and when you are charging gradient onto them, how
the formula would look like or if you have one scalar function one vector function and you
are charging divergence on them, then how the formula would look like things like that.
Because the not only in vector calculus, but these formulas will also be used or would be
of some use in mechanics and in other branches of applied mathematics.

So, it is always nice to know these formulas and today we would try to derive at least one
or two of them. So, today we will start with our first example and then we will move on to
those formulas so, just to recapitulation.

(Refer Slide Time: 01:17)

So, example one for today show that or find the value. So, show that Laplacian of
𝑥
𝑖. 𝑒. ∇2 (𝑟 3 ) = 0. So, remember our 𝑟⃗ = 𝑥𝑖̂ + 𝑦𝑗̂ + 𝑧𝑘̂. So, this is our r and when we

406
say 𝑟 = |𝑟⃗| = √𝑥 2 + 𝑦 2 + 𝑧 2 . So, in vector calculus or in vector algebra whatever you
are studying its always how to say common to know that r is basically this (x,y,z) vector
and |𝑟⃗| = √𝑥 2 + 𝑦 2 + 𝑧 2 that is the length of this vector 𝑟⃗. So, just keep in mind all right.

Now this ∇2 ; it means that double square not double square nabla square is basically your
Laplacian or Laplace operator. So, this is called as Laplace operator and we know that
𝜕2 𝜕2 𝜕2
Laplace operator we know that this Laplace operator is nothing, but ∇2 = 𝜕𝑥 2 + 𝜕𝑦 2 + 𝜕𝑧 2;

that means, partial derivative of with respect to x twice partial derivative with respect to y
twice and partial derivative with respect to z twice. And on this Laplace operator, we
always charge a scalar function because Laplace operator itself is a scalar operator in a
way.

𝑥
So, that is why here you have a scalar function ; 𝑟 3 is a scalar function, x is a scalar
𝑟3
𝜕2
function all right. So, let us start with calculating our . So, basically what we have is
𝜕𝑥 2
𝑥 𝜕2 𝜕2 𝜕2 𝑥
the Laplacian of. So,what we have is here . ∇2 (𝑟 3 ) = (𝜕𝑥 2 + 𝜕𝑦 2 + 𝜕𝑧 2 ) (𝑟 3 ) all right.

So that means, we will charge this on every how to say partial derivative. So, we have
𝜕2 𝑥 𝜕2 𝑥 𝜕2 𝑥
(𝑟 3 ) + 𝜕𝑦 2 (𝑟 3 ) + 𝜕𝑧 2 (𝑟 3 ) all right. So, let us first calculate this one. So, we can call it
𝜕𝑥 2

𝜕2 𝑥 𝜕 𝜕 𝑥 𝜕 1
as equation 1. So, ( ) = 𝜕𝑥 (𝜕𝑥 (𝑟 3 )) all right. So, when we calculate [ .1−
𝜕𝑥 2 𝑟 3 𝜕𝑥 𝑟 3

3 𝜕𝑟
𝑥 𝑟 4 𝜕𝑥] . So, we treat it as a product of 2 functions.

𝜕𝑟
right yes because 𝑟 = √𝑥 2 + 𝑦 2 + 𝑧 2 . So that means, first of all it will become =
𝜕𝑥
2𝑥 𝑥
= 𝑟 . So, what we are doing is here let me expand this thing a little bit so just
2√𝑥 2 +𝑦2 +𝑧 2

𝜕𝑟 𝑥
to make things clear to the you. So, then it will become = all right.
𝜕𝑥 𝑟

Because first we differentiate it with respect to r and then it became then sorry then it
𝜕𝑟
So, first we differentiate it with respect to r and now we are differentiating r with respect
𝜕𝑥

to x in a way and our r is if you look our r is √𝑥 2 + 𝑦 2 + 𝑧 2 . So, when we differentiate

2𝑥
with respect to x, it will be because then you differentiate with 𝑥 2 with respect
2√𝑥 2 +𝑦 2 +𝑧 2

to x.

407
(Refer Slide Time: 06:38)

So, this whole thing will turn into very nicely. This whole thing will turn into the
𝜕2 𝑥 𝜕 1 3 𝜕𝑟 𝜕 1 3 𝑥 𝜕 1 3𝑥 2 3 𝑥 6𝑥
(𝑟 3 ) = 𝜕𝑥 [𝑟 3 . 1 − 𝑥 𝑟 4 𝜕𝑥] = 𝜕𝑥 [𝑟 3 . 1 − 𝑥 𝑟 4 . 𝑟 ] = 𝜕𝑥 [𝑟 3 − ] = − 𝑟4 𝑟 − 𝑟5 +
𝜕𝑥 2 𝑟5
15𝑥 2 𝜕𝑟 3𝑥 6𝑥 15𝑥 2 𝑥 9𝑥 15𝑥 3
= − 𝑟5 − 𝑟5 + . 𝑟 = − 𝑟5 + .
𝑟 6 𝜕𝑥 𝑟6 𝑟7

So, I wrote r it is very simple to see and it is also very simple to calculate. I am sure you
can be able to do that.

𝜕2
So, that is just the first component. Now similarly we can calculate 𝜕𝑦 2

(Refer Slide Time: 08:51)

408
 𝜕2 𝑥 3𝑥 15𝑥𝑦 2
So, similarly we can calculate similarly 𝜕𝑦 2 (𝑟 3 ) = − 𝑟 5 + and sorry here it was that
𝑟7

𝜕2 𝑥 3𝑥 15𝑥𝑧 2
given vector function ( ) = − 𝑟5 + all right.
𝜕𝑧 2 𝑟 3 𝑟7

𝜕2 𝜕2 𝜕2
So, now, let us add all these how to say all these 𝜕𝑥 2 , 𝜕𝑦 2 , 𝜕𝑧 2 term. So, putting above values
𝑥 9𝑥 15𝑥 3
in 1 so, then our Laplacian of x by r cube would result into ∇2 (𝑟 3 ) = − 𝑟 5 + +=
𝑟7
3𝑥 15𝑥𝑦 2 3𝑥 15𝑥𝑧 2 15𝑥 15𝑥 15𝑥 15𝑥 15𝑥
− 𝑟5 + − 𝑟5 + =− + (𝑥 2 + 𝑦 2 + 𝑧 2 ) = − + . 𝑟2 = − +
𝑟7 𝑟7 𝑟5 𝑟7 𝑟5 𝑟7 𝑟5
15𝑥
= 0.
𝑟5

So, we will have this thing this is equal to 0. So, they and this is what we needed to prove
right. So, this is what we needed to prove and therefore, we have the answer. So, it is all
about doing partial derivative carefully and I am pretty sure you can be able to solve these
types of problems all right. Now I think we have covered enough examples on calculation
of gradient divergence and curl of several types of vector and scalar functions and now we
move on to some vector identities all right.

409
(Refer Slide Time: 11:57)

So, let us move on to some vector identities or let us say important vector identities ok.
So, the first identity is and we assume that all these vector functions have continuous
partial derivatives.

So, these are all vector functions of variables x, y and z and they all have continuous partial
derivatives so, that we can at least talk about their gradient, divergence and curl. So, I am
not going to write all those things, I will just write the vector identities. So, suppose we
have 𝐴⃗ and 𝐵
⃗⃗ as the 2 vector functions, then ⃗∇⃗. (𝐴⃗ + 𝐵
⃗⃗ ) = ⃗∇⃗. 𝐴⃗ + ⃗∇⃗. 𝐵
⃗⃗, I wrote already this

result.

I do not remember, but I think I may have written this and of course, doing this is not
complicated. So, we basically do so divergence of the sum is equal to sum of the
divergence. So, from here we charge the divergence on both sides and then we separate
divergence for 𝐴⃗ divergence of for 𝐵
⃗⃗ and then that will give you the required answer all
right.

So, here this one proving this one is not complicated; let me how to say show you at least
one of them. So, 𝐴⃗ is a vector function. So, 𝐴⃗ has 3 components 𝐴1 𝑖̂ + 𝐴2 𝑗̂ + 𝐴3 𝑘̂. And 𝐵
⃗⃗
⃗⃗ = 𝐵1 𝑖̂ + 𝐵2 𝑗̂ + 𝐵3 𝑘̂ . So, when we talk about 𝐴⃗ + 𝐵
has 3 components so, 𝐵 ⃗⃗ so; that

means, we will sum that component. So,(𝐴1 + 𝐵1 )𝑖̂ + (𝐴2 + 𝐵2 )𝑗̂ + (𝐴3 + 𝐵3 )𝑘̂. And

410
now when we charge the divergence on both sides, then this will become ⃗∇⃗. (𝐴⃗ + 𝐵
⃗⃗ ) =
𝜕 𝜕 𝜕
(𝐴1 + 𝐵1 ) + (𝐴2 + 𝐵2 ) + (𝐴3 + 𝐵3 ).
𝜕𝑥 𝜕𝑦 𝜕𝑧

𝜕𝐴 𝜕𝐴2 𝜕𝐴3 𝜕𝐵 𝜕𝐵2 𝜕𝐵3

So, we can separate the term ( 𝜕𝑥1 + + ) + ( 𝜕𝑥1 + + ) = ⃗∇⃗. 𝐴⃗ + ⃗∇⃗. 𝐵
⃗⃗. So
𝜕𝑦 𝜕𝑧 𝜕𝑦 𝜕𝑧

now, this is nothing, but our ⃗∇⃗. 𝐴⃗ right and this is nothing, but our ⃗∇⃗. 𝐵
⃗⃗ and this is what we
needed to prove. So, divergence of the sum is equals to sum of the divergence this result
follows.

(Refer Slide Time: 15:23)

Now second result is let us say we have ⃗∇⃗. (𝜑𝐴⃗) = 𝜑∇

⃗⃗. 𝐴⃗ + ⃗∇⃗𝜑. 𝐴⃗ . Now since we had a

dot product here, we can write 𝐴⃗. ⃗∇⃗𝜑 or ⃗∇⃗𝜑. 𝐴⃗ that is all fine, but if it was if it were a curl
here, then this formula cannot be how to say played with.

So, this is basically divergence of 𝜑 times 𝐴⃗ and proving this result would also be not
difficult. So, we multiply every component of 𝐴⃗ with 𝜑 and then we just follow the product
rule of 2 scalar functions and you equate the terms. So, you keep the terms for 𝜑 at one
place and term for 𝐴⃗ at one place, you take something common and that will give you the
required result. So, proving this one would also be not difficult and it is very
straightforward. Third result is curl of now this is interesting; now ⃗∇⃗ × (𝜑𝐴⃗) =
⃗⃗ × 𝐴⃗) + ⃗∇⃗𝜑 × 𝐴⃗ all right.
𝜑(∇

411
 ⃗⃗ × 𝐴⃗) + ⃗∇⃗𝜑 × 𝐴⃗. So, this is our required result and here I cannot write 𝐴⃗ × ⃗∇⃗𝜑 , 𝐴⃗
So, 𝜑(∇

cross product with ⃗∇⃗𝜑 . So, that I cannot write because then in that case, I have to put
minus sign here. So, here I can write 𝐴⃗. ⃗∇⃗𝜑, but here I cannot write 𝐴⃗ × ⃗∇⃗𝜑 because as I
said this formula cannot be played with all right.

Now next we have suppose instead of one scalar function if we have 2 vector functions.
So, I have 2 vector functions, I do not have any scalar function. So, this can be proved
⃗⃗ dot dot
easily. Now when we have 2 vector functions and then in that case this will be 𝐵
product with sorry.

So, when we have so let us start with the dot product. So, when we have dot here , I will
⃗⃗. (𝐴⃗ × 𝐵
come to the curl formula. So, when we have dot here, then ∇ ⃗⃗ ) = 𝐵 ⃗⃗ × 𝐴⃗) −
⃗⃗ . (∇

𝐴⃗. (∇
⃗⃗ × 𝐵
⃗⃗ ) so, the minute I take dot product it will become a scalar function here also it

will become a scalar function. So, ultimately we are getting a scalar function. Proving this
one is also not difficult here we basically calculate the cross product of 𝐴⃗ and 𝐵
⃗⃗.

So, that will be like 𝑖̂, 𝑗̂, 𝑘̂ j k and then component of 𝐴⃗, component of 𝐵
⃗⃗ you break that
determinant and charge the gradient and then separate the terms basically and that will
give you the required answer. So, since they are a little bit tedious, but obvious to do I am
avoiding them; however, they are very important I mean it is very nice how to say
important to know these formulas, because they might be useful at some point all right.
Next we have is ⃗∇⃗ × (𝐴⃗ × 𝐵 ⃗⃗ . ⃗∇⃗)𝐴⃗ − 𝐵
⃗⃗ ) = (𝐵 ⃗⃗ . 𝑑𝑖𝑣𝐴⃗ − (𝐴⃗. ⃗∇⃗)𝐵
⃗⃗ + 𝐴⃗. 𝑑𝑖𝑣𝐵
⃗⃗.

so; that means, here we are obtaining a vector quantity. So, 𝐴⃗ × 𝐵

⃗⃗ is a vector and its curl
is also vector.

So, it is very obvious that the right hand side should also be a vector all right. So, right
⃗⃗ dot nabla.
hand side should also be a vector so; that means, the formula would read as 𝐵
⃗⃗ dot nabla would become scalar however, when you charge a vector then it would
So, 𝐵
⃗⃗ times divergence of 𝐴⃗. So, divergence of 𝐴⃗ would be scalar
become a vector. Similarly 𝐵
and then you multiply by a vector. So, ultimately this is also vector. 𝐴⃗ dot nabla is a scalar,
⃗⃗ and then in that case it will become a vector and divergence of
but in when you charge 𝐵
⃗⃗ is a scalar, but multiplying with a vector 𝐴⃗ is again vector.
𝐵

412
So; that means, from the scalar and vector point of view these 2 inqualities are settled.
Now we have to prove this. So, what you can do? You can extract the components I mean
you so, first calculate the curl of 𝐴⃗ and 𝐵
⃗⃗ and then whatever curl you get you charge nabla
and then you get a certain expression, then similarly you calculate every term on the right
hand side and show that they are same. So, of course, this will be a slightly complicated
or may be time taking in a way, but it is not difficult. So, it is not difficult or it is not tough
it is just tedious or too long to do, but if you remember the formula that is fine. So, this is
another formula from the divergence and curl point of view.

Our next formula is 𝑔𝑟𝑎𝑑(𝐴⃗. 𝐵 ⃗⃗ . ⃗∇⃗)𝐴⃗ + (𝐴⃗. ⃗∇⃗)𝐵

⃗⃗ ) = (𝐵 ⃗⃗ × 𝑐𝑢𝑟𝑙𝐴⃗ + 𝐴⃗ × 𝑐𝑢𝑟𝑙𝐵
⃗⃗ + 𝐵 ⃗⃗ . Again

this the result is not complicated; it is just that you have to calculate the left hand side and
right hand side and show that they are same. So, of course, calculating the left hand side
is simpler compared to the right hand side. So, you basically calculate 𝐴⃗. 𝐵
⃗⃗.

So, it is 𝐴1 𝐵1 + 𝐴2 𝐵2 + 𝐴3 𝐵3 and then you charge the gradient and then you obtain an
expression and you do the similar thing on the left hand and the right hand side and show
that they are same all right. This will also be a little bit tedious or long to do, but it is
simple.

(Refer Slide Time: 23:01)

Our next formula is 𝑑𝑖𝑣(𝑔𝑟𝑎𝑑𝜑) so; that means, we have 𝑑𝑖𝑣(𝑔𝑟𝑎𝑑𝜑). And this is
⃗⃗𝜑) = ⃗∇⃗. (𝜕𝜑 𝑖̂ + 𝜕𝜑 𝑗̂ + 𝜕𝜑 𝑘̂). And when you charge this
nothing, but 𝑑𝑖𝑣(𝑔𝑟𝑎𝑑𝜑) = ⃗∇⃗. (∇ 𝜕𝑥 𝜕𝑦 𝜕𝑧

413
 𝜕𝜑 𝜕𝜑 𝜕𝜑
⃗⃗. (∇
divergence, then it will become 𝑑𝑖𝑣(𝑔𝑟𝑎𝑑𝜑) = ∇ ⃗⃗𝜑) = ∇
⃗⃗. ( 𝑖̂ + 𝑗̂ + 𝑘̂) =
𝜕𝑥 𝜕𝑦 𝜕𝑧

𝜕2 𝜑 𝜕2 𝜑 𝜕2 𝜑
+ 𝜕𝑦 2 + 𝜕𝑧 2 = ∇2 𝜑 . So, ultimately this is Laplacian of 𝜑.
𝜕𝑥 2

So, basically 𝑑𝑖𝑣(𝑔𝑟𝑎𝑑𝜑) is actually Laplacian of 𝜑 right; very simple to solve and now
this is interesting find 𝑐𝑢𝑟𝑙(𝑔𝑟𝑎𝑑𝜑). So, the solution 𝑐𝑢𝑟𝑙(𝑔𝑟𝑎𝑑𝜑); that is we have to
𝑖̂ 𝑗̂ 𝑘̂
𝜕 𝜕 𝜕
find 𝑐𝑢𝑟𝑙(𝑔𝑟𝑎𝑑𝜑) = ∇ ⃗⃗𝜑) = || 𝜕𝑥
⃗⃗ × (∇ 𝜕𝑦
|
𝜕𝑧 |..
𝜕𝜑 𝜕𝜑 𝜕𝜑
𝜕𝑥 𝜕𝑦 𝜕𝑧

𝜕2 𝜑 𝜕2 𝜑 𝜕2 𝜑 𝜕2 𝜑
So, if you break this determinant then it will be 𝑖̂ (𝜕𝑦𝜕𝑧 − 𝜕𝑧𝜕𝑦) + 𝑗̂ (𝜕𝑧𝜕𝑥 − 𝜕𝑥𝜕𝑧) +
𝜕 𝜑2 𝜕 𝜑 2
𝑘̂ (𝜕𝑥𝜕𝑦 − 𝜕𝑦𝜕𝑥) = 𝑖̂[0 − 0] + 𝑗̂[0 − 0] + 𝑘̂[0 − 0] = ⃗0⃗. So, ultimately a ⃗0⃗ so; that means,

curl of divergence of 𝜑 regardless of what type of function 𝜑 is this will always be 0 all
right. So, it is always a ⃗⃗
0 Similarly we can show that or we can prove that 𝑑𝑖𝑣(𝑐𝑢𝑟𝑙𝐴⃗); so,
⃗⃗ × 𝐴⃗).
similarly we can prove the ⃗∇⃗. (∇

Let us write in words 𝑑𝑖𝑣(𝑐𝑢𝑟𝑙𝐴⃗) = 0 . So, first we calculate the curl and then when you
take the divergence you will always get del del x of del A to del y minus del A 3 del z. So,
basically you will differentiate with respect to a variable for which you will end up getting
a constant which you end up getting a 0 partial derivative in a way.

So, this can also be proved very easily and here in this case, you will actually obtain the
similar type of partial derivatives. So, here in this case you do not obtain how to say 0
partial derivative, you basically obtain partial derivatives which are same and therefore,
they will cancel out. So, it will be done like that and the final result is 𝑐𝑢𝑟𝑙(𝑐𝑢𝑟𝑙𝐴⃗).

414
(Refer Slide Time: 26:42)

So, you have 𝑐𝑢𝑟𝑙(𝑐𝑢𝑟𝑙𝐴⃗) = ⃗∇⃗ × (∇

⃗⃗ × 𝐴⃗) = ⃗∇⃗(∇
⃗⃗. 𝐴⃗) − ⃗∇⃗(∇
⃗⃗. 𝐴⃗) = ⃗∇⃗(∇
⃗⃗. 𝐴⃗) − ∇2 𝐴⃗.all

right. So, this is also quite easy to prove. So, first we calculate the left hand side and then
we calculate the right hand side and we show that they are same. So, let us see if it makes
sense or not. So, we have 𝑐𝑢𝑟𝑙(𝑐𝑢𝑟𝑙𝐴⃗) so; that means, this will result into a vector and
here, it will also give us a vector and this one would also give us a vector because here we
have a vector and ultimately we will obtain this result all right.

Yes. So, this is basically it is better to write this as this way. So, it is not exactly a
Laplacian. So, we write it as this way. So, then you have a scalar, then this will become a
vector and then you have a scalar then this again becomes a vector. So, this and now I get
it. So, we can write it as Laplacian yes. So, we can write it as Laplacian of a vector 𝐴⃗.

So, this is the required how to result and you just calculate the left hand side and right hand
side and show them. They are same and that this would complete our require our how to
say collection of ten identities there are some more results we will not get into them and
in the next class we will probably practice few more examples maybe a couple of more on
say gradient divergence and curl of a vector function and then we move on to our next
topic which is directional derivative of a vector function.

So, I will stop here for today and I look forward to you in our next class.

Thank you.

415
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 40
Level Surface Relevant Theorems

Hello students. So, in the last class, we started with the examples on divergence, gradient
and curl. So, up until last class we looked into three important operators in vector calculus.
And they are divergence of a vector function gradient of a scalar function, and a curl of a
vector function. And we also worked out few examples, where we saw how we can
calculate these three operators.

(Refer Slide Time: 01:02)

Like if you were given a scalar function let us say f, so then in that case how we can
calculate. So, if we are given a scalar function f, which is a function of let us say three
scalar variables 𝑓(𝑥, 𝑦, 𝑧), then basically the divergence of this scalar function f is denoted
by sorry gradient of the scalar function will be denoted by this ⃗∇𝑓. And this nabla is a
vector quantity. So, we put a vector sign here.

𝜕𝑓 𝜕𝑓 𝜕𝑓
⃗ 𝑓 = 𝑖̂ + 𝑗̂ + 𝑘̂ . So, this is a vector quantity this gradient.
And this is basically ∇ 𝜕𝑥 𝜕𝑦 𝜕𝑧

And if we are given a vector function 𝑓 (𝑥, 𝑦, 𝑧), then we know that since we are charging
this nabla, which is a vector quantity on to a vector function. So, then in that case there

416
can be two possible operations. So, either it can be a dot product or it can be a cross
product.

𝜕𝑓1 𝜕𝑓2 𝜕𝑓3

So, if it is a dot product, then this is called divergence, and it is nothing but + + .
𝜕𝑥 𝜕𝑦 𝑑𝑧
𝜕𝑓1 𝜕𝑓 𝜕𝑓
Some people also prefer to write it as + 𝜕𝑥2 + 𝑑𝑥3 , it is depending on the fact that
𝜕𝑥1 2 3

whether you are choosing your coordinate axes as x, y, z or 𝑥1 , 𝑥2 , 𝑥3 , it really does not
make any difference. And if we charge let us say cross product, then this can be calculated
𝑖̂ 𝑗̂ 𝑘̂
𝜕 𝜕 𝜕
as | |.
𝜕𝑥 𝜕𝑦 𝜕𝑧
𝑓1 𝑓2 𝑓3

So, of course a vector function has three components, so we write it in this way. And this
is basically divergence this is our divergence, and this is our gradient, and this one is
actually our curl of a vector function 𝑓. So, here I was supposed to write three components
of the vector function 𝑓. So, 𝑓1 , 𝑓2 , and 𝑓3 . And here we also use the same components 𝑓1 ,
𝑓2 , and 𝑓3 .

So, we worked out few examples based on these three operators, and we also calculated
somehow to say a vector function whether it is solenoidal or irrotational. So, we did some
examples like that as well. Now, using this gradient of a scalar function. There are two
new topics that we are going to introduce at the moment.

So, the first one is actually it is called directional derivative. So, we will start with today
directional derivative. So, in case of scalar function, we calculated the derivative of a given
scalar function. So, for a scalar function let us say 𝑦 = 𝑓(𝑥) if it is differentiable, then we
𝑑𝑦
just differentiate 𝑑𝑥 , and we write the derivative of this scalar function as 𝑓 ′ (𝑥).

Now, in case of vector functions, let us say we have 𝑟 = 𝑓(𝑡), when we differentiate this
one we can talk about a direction in which this vector function can be differentiable. So,
whether it is differentiable along how to say x, y, x-axis, y-axis or z-axis or whether it is it
is differentiable along any vector, which is not x, y or z-axis.

So, we can talk about its derivative along a certain vector actually, and that is derivative is
called as the directional derivative. So, we will give the formal definition of directional

417
derivative, but before that we will start with the definition of level surfaces. So, what do
we mean by level surfaces, and we will try to see one or two results based on the level
surfaces, before we move to the to the directional derivative all right.

(Refer Slide Time: 05:18)

So, let us start with level surfaces. So, this is also in our syllabus, and I think we can learn
a little bit about it. So, the formal definition goes like this. Let 𝑓(𝑥, 𝑦, 𝑧)be a scalar field.
So, we have defined the scalar field in our previous lecture, so or maybe not in previous,
but one or two lectures before, so you can look into the definition of scalar field.

So, let 𝑓(𝑥, 𝑦, 𝑧) be a scalar field over a region R, then the points satisfying an equation of
the type 𝑓(𝑥, 𝑦, 𝑧) = 𝑐, where c is an arbitrary constant constitutes a family of surfaces in
three- dimensional space. And the family of these surfaces are called as level surfaces.

So, f is a scalar field in a region R, and the points which are actually satisfying this equation
𝑓(𝑥, 𝑦, 𝑧) = 𝑐, where c is can be any arbitrary constant. So, for every constant c will get a
how to say will get a surface basically, so 𝑓(𝑥, 𝑦, 𝑧) = 1, then it will give one surface, then
𝑓(𝑥, 𝑦, 𝑧) = 2, then it will give another surface.

So, the family of all these surfaces, they are called as the level surfaces. So, this and of
course this family of surfaces, they are in three- dimensional space. So, because c is any
arbitrary constant, we will get different types of surfaces actually and that constitutes a

418
family, so that family is called as level surface. And that is our definition for the level
surface.

Now, there is a very small result in this regard. So, the result goes like this. So, this is a
small theorem. So, let 𝑓(𝑥, 𝑦, 𝑧) be a scalar field over a region R, then through any point
of R there passes one and only one level surface. So, it says that suppose you have a scalar
field 𝑓(𝑥, 𝑦, 𝑧) over a region R, then every point on this region R, there can be only one
level surface that will pass through that point. So, there cannot be two level surfaces that
will pass through the same point. So, we will prove that at any particular point on that
surface R one and only one level surface will pass through that point.

(Refer Slide Time: 10:54)

So, to begin with let (𝑥1 , 𝑦1 , 𝑧1 ) be any point of the region R, then the level surface
𝑓(𝑥, 𝑦, 𝑧) passes through that point because 𝑓(𝑥, 𝑦, 𝑧) is a level surface in R. So, if we have
a point (𝑥1 , 𝑦1 , 𝑧1 ) in R, then 𝑓(𝑥, 𝑦, 𝑧) = 𝑐 will pass through that point, so that is this
𝑓(𝑥, 𝑦, 𝑧) = 𝑐 will satisfy the point (𝑥1 , 𝑦1 , 𝑧1 ) will satisfy that equation, and this will be
𝑓(𝑥1 , 𝑦1 , 𝑧1 ) = 𝑐1 right.

Now, suppose the level surfaces 𝑓(𝑥, 𝑦, 𝑧) = 𝑐 and 𝑓(𝑥1 , 𝑦1 , 𝑧1 ) = 𝑐1 passes through the
same point (𝑥1 , 𝑦1 , 𝑧1 ). So, 𝑓(𝑥, 𝑦, 𝑧) = 𝑐 is the equation of one level surface. And
𝑓(𝑥1 , 𝑦1 , 𝑧1 ) = 𝑐1 is the equation of this second level surface all right or second surface.

419
Now, if we assume that both of these two level surfaces passes through the same point
(𝑥1 , 𝑦1 , 𝑧1 ), then they both satisfy then the point (𝑥1 , 𝑦1 , 𝑧1 ) must satisfy both the level
surfaces, so that means when I substitute (𝑥1 , 𝑦1 , 𝑧1 ) in 𝑓(𝑥, 𝑦, 𝑧) = 𝑐, then that equation
will also hold true. And if I substitute, and 𝑓(𝑥1 , 𝑦1 , 𝑧1 ) = 𝑐1, then that equation will also
hold true. So, if both of these two-level surfaces passes through this point (𝑥1 , 𝑦1 , 𝑧1 ), then
the point must satisfy these two equations.

So, what will happen is we will have 𝑓(𝑥1 , 𝑦1 , 𝑧1 ) = 𝑐1 from here. And if it satisfy this
equation, then it will be 𝑓(𝑥1 , 𝑦1 , 𝑧1 ) = 𝑐2 . We will get a different constant, because the
level surfaces different. And so ideally I mean we should get a different constant. So, j and
now we see that the equation since 𝑓(𝑥, 𝑦, 𝑧) has a unique value at (𝑥1 , 𝑦1 , 𝑧1 ), because the
point is same.

So, if the point is same, and the equation of the level surface which was 𝑓(𝑥, 𝑦, 𝑧) = 𝑐 is
same in a way, so that means the value at the point (𝑥1 , 𝑦1 , 𝑧1 )must also be same. So, we
can write since 𝑓(𝑥, 𝑦, 𝑧) has a unique value at (𝑥1 , 𝑦1 , 𝑧1 ), therefore the constants must be
same. So, if the equation 𝑓(𝑥, 𝑦, 𝑧) = 𝑐 has a unique value at any point (𝑥1 , 𝑦1 , 𝑧1 ), because
it cannot have two different values, it cannot have 𝑐1, and then it cannot have 𝑐2 , so because
the equation of the level surface is 𝑓(𝑥, 𝑦, 𝑧) = 𝑐 is same.

So, if the point is also same, then it cannot yield two different values then in that case it is
a contradiction. So, therefore, our constants must be same, so that means,𝑐1 = 𝑐2 . And this
implies that hence only one level surface passes through the point (𝑥1 , 𝑦1 , 𝑧1 ). So, what I
am trying to say is that you choose two different points say (𝑥1 , 𝑦1 , 𝑧1 ) and (𝑥2 , 𝑦2 , 𝑧2 ),
and since those two points are in the region R we substitute those 2 points in the equation
𝑓(𝑥, 𝑦, 𝑧) = 𝑐, so that means, we have according to this here we will have 𝑓(𝑥1 , 𝑦1 , 𝑧1 ) =
𝑐1 , and 𝑓(𝑥2 , 𝑦2 , 𝑧2 ) = 𝑐2 that is what we are trying to do here.

But if it passes through the same point, then we will have 𝑓(𝑥1 , 𝑦1 , 𝑧1 ) = 𝑐1 , and
𝑓(𝑥1 , 𝑦1 , 𝑧1 ) = 𝑐2 . But the points the level surface 𝑓(𝑥, 𝑦, 𝑧) must have a unique value
because the point is same. So, if you assume the same point, so the same level surface that
is passing through the point cannot yield to different values. So, 𝑐1 cannot be different
from 𝑐2 . And therefore, the only possibility we have is that 𝑐1 = 𝑐2.

420
And if you have 𝑐1 = 𝑐2, that means, that every level surface can pass through point in R
only once I mean they cannot how to say only one level surface can pass through the point
(𝑥1 , 𝑦1 , 𝑧1 ), so that means, when you have a certain point in the region R one and only
level surface the will pass through that that particular point and this is this is what we
wanted to prove all right.

So, there cannot be two level surfaces that will pass through this point
(𝑥1 , 𝑦1 , 𝑧1 ).Similarly, you can choose any arbitrary point, so since (𝑥1 , 𝑦1 , 𝑧1 ) was chosen
arbitrarily this result will true for all the points in R all right. So, these type of results are
also quite common in level surfaces.

(Refer Slide Time: 17:35)

Now, another theorem which is a little bit more interesting; so, you remember we
calculated the gradient of a scalar function. So, gradient of a scalar function we usually
write in this way, but what does it mean actually what it does to a surface let us say f(x, y,
z)=c. So, when we calculate the gradient of this of this surface, that means we can simply
write it as 𝜑(𝑥, 𝑦, 𝑧) = 𝑓(𝑥, 𝑦, 𝑧) − 𝑐 , and then calculating gradient of 𝜑 is like
calculating gradient of f because gradient of c will be 0.

So, what it this gradient of f mean here I mean to this surface 𝑓(𝑥, 𝑦, 𝑧) = 𝑐 and that
actually is its quite how to say interesting or important to know that this gradient of f is
actually normal to the surface 𝑓(𝑥, 𝑦, 𝑧) = 𝑐. And there this is also one of the important
results in vector calculus that what do we actually mean by a gradient of a scalar function.

421
And it actually means that a gradient of the scalar function is normal to the surface
𝑓(𝑥, 𝑦, 𝑧) or that is scalar function equals to the equals to constant. So, we will try to prove
⃗ 𝑓 is a vector normal to the surface
that result. So, let me state that theorem. So, ∇
⃗ 𝑓 is a vector normal to the surface
𝑓(𝑥, 𝑦, 𝑧) = 𝑐 or proof, we can write also as proof. So, ∇
𝑓(𝑥, 𝑦, 𝑧) = 𝑐, where c is a constant all right.

So, let us see. So, let 𝑟 = 𝑥𝑖̂ + 𝑦𝑗̂ + 𝑧𝑘̂. So, always as I have told you whenever you write
𝑟 in vector calculus, it is always the point (x, y, z). So, 𝑥𝑖̂ + 𝑦𝑗̂ + 𝑧𝑘̂ be the position vector,
of any point P(x, y, z) on the level surface of x sorry level surface it is not a vector function.
So, level surface 𝑓(𝑥, 𝑦, 𝑧) = 𝑐. So, that means, if we have let us say x-axis, y-axis and z-
axis and let us say if this is our level surface. So, and this is our point P. Of course, the
level surface it is not looking like a level surface it has to be a surface, I drew it like a
curve, but it is not a curve. We have to draw it in a proper manner that, so that it looks like
a surface.

Since, I am not good at drawing, I just drew like that. But remember this has to be a surface
all right. And now this is so this P, so this is basically our vector. And this is our vector 𝑟
𝑂𝑃 = 𝑥𝑖̂ + 𝑦𝑗̂ + 𝑧𝑘̂, and it can be
and this vector 𝑟 is equals to if you write vector 𝑟 = ⃗⃗⃗⃗⃗
written as (𝑥𝑖̂ + 𝑦𝑗̂ + 𝑧𝑘̂ ) all right ok. So, P(x, y, z) be any point on the level surface. So,
this is my equation of the level surface 𝑓(𝑥, 𝑦, 𝑧) = 𝑐 all right.

And now I assume another point let us say Q here, and let Q be any point. So, Q be any
point on the level surface as I can write 𝑥 + 𝛿𝑥. So, it is a small increment. So, I can
assume at no somewhere either here or here it is up to you. So, I assume let us say here as
a point 𝑄(𝑥 + 𝛿𝑥, 𝑦 + 𝛿𝑦, 𝑧 + 𝛿𝑧). So, it says, it is a neighboring point on the level surface
𝑓(𝑥, 𝑦, 𝑧) = 𝑐. And it is a neighboring point to the point P. Then the position vector then
⃗⃗⃗⃗⃗⃗ or the position vector of Q can be given by (𝑥 + 𝛿𝑥)𝑖̂ + (𝑦 + 𝛿𝑦)𝑗̂ + (𝑧 + 𝛿𝑧)𝑘̂.
𝑂𝑄

422
(Refer Slide Time: 23:31)

So, I can write it as (𝑥𝑖̂ + 𝑦𝑗̂ + 𝑧𝑘̂) + (𝛿𝑥𝑖̂ + 𝛿𝑦𝑗̂ + 𝛿𝑧𝑘̂) = 𝑟 + 𝛿𝑟

⃗⃗⃗⃗ . So, I write

(𝑥𝑖̂ + 𝑦𝑗̂ + 𝑧𝑘̂) as 𝑟 and I write this (𝛿𝑥𝑖̂ + 𝛿𝑦𝑗̂ + 𝛿𝑧𝑘̂) as 𝛿𝑟

⃗⃗⃗⃗ . So, both of them are vector

⃗⃗⃗⃗ all right. So, now P to Q this vector,

quantity. And this small increment is denoted by 𝛿𝑟
⃗⃗⃗⃗⃗ = 𝑂𝑄
is 𝑃𝑄 ⃗⃗⃗⃗⃗⃗ − 𝑂𝑃 ⃗⃗⃗⃗ ) − 𝑟 = 𝛿𝑟
⃗⃗⃗⃗⃗ = (𝑟 + 𝛿𝑟 ⃗⃗⃗⃗ .

⃗⃗⃗⃗⃗ tends to the tangent at P right to the level surface at P

Now, as Q tends to P, the line 𝑃𝑄
⃗⃗⃗⃗ = 𝑑𝑥𝑖̂ + 𝑑𝑦𝑗̂ + 𝑑𝑧𝑘̂ lies on the tangent plane to the surface at P. So, if we
.Therefore, 𝑑𝑟
make Q goes to P, then in that case the line ⃗⃗⃗⃗⃗
𝑃𝑄 will actually be a tangent to this level
surface at the point P all right, so that is what we have written here.

So, as Q goes to P, the line ⃗⃗⃗⃗⃗

𝑃𝑄 tends to tangent at the point P to the level surface. And
therefore, this ⃗⃗⃗⃗
𝑑𝑟 = 𝑑𝑥𝑖̂ + 𝑑𝑦𝑗̂ + 𝑑𝑧𝑘̂ lies on the tangent plane. So, when Q goes to P, this
(𝑟 + ⃗⃗⃗⃗
𝑑𝑟) will tends to ⃗⃗⃗⃗
𝑑𝑟 and this 𝑥 + 𝑑𝑥 will tend to 𝑑𝑥 say and so on, and that will
actually lie on the tangent plane to the surface at the point P. And therefore, from the
differential calculus we know that. So, what do we have from differential calculus, we
𝜕𝑓 𝜕𝑓 𝜕𝑓
have 𝑑𝑓 = 𝜕𝑥 𝑑𝑥 + 𝜕𝑦 𝑑𝑦 + 𝜕𝑧 𝑑𝑧.

𝜕𝑓 𝜕𝑓
So, this is basically if I want to write it, then this can be written as (𝜕𝑥 𝑖̂ + 𝜕𝑦 𝑗̂ +
𝜕𝑓
𝑘̂) . (𝑑𝑥𝑖̂ + 𝑑𝑦𝑗̂ + 𝑑𝑧𝑘̂) .
𝜕𝑧

423
(Refer Slide Time: 27:27)

So, I can write this as, so I can write it here. So, this can be written as ⃗∇𝑓. ⃗⃗⃗⃗
𝑑𝑟 right. And
since the equation of the level surface or of the surface simply we do not have to write
levels. So, of the surface is 𝑓(𝑥, 𝑦, 𝑧) = 𝑐. So, from here 𝑑𝑓 = 0 right, it is a differential
of this 𝑓(𝑥, 𝑦, 𝑧) = 𝑐 because right hand side is a constant. And from here this df can be
written as ⃗∇𝑓. ⃗⃗⃗⃗
𝑑𝑟 = 0, and this one is ⃗⃗⃗⃗
𝑑𝑟. So, if the dot product of 2 vectors as 0; then, they
must be perpendicular to one another.

So, from here I can write this ⃗∇𝑓 is a vector perpendicular to ⃗⃗⃗⃗
𝑑𝑟. And if it is perpendicular
⃗⃗⃗⃗ , then this 𝑑𝑟
to 𝑑𝑟 ⃗⃗⃗⃗ if we look here, so this 𝑑𝑟
⃗⃗⃗⃗ we draw the conclusion that 𝑑𝑟
⃗⃗⃗⃗ is actually
⃗⃗⃗⃗ , but
⃗ 𝑓 that is perpendicular to 𝑑𝑟
lying on the tangent plane, so that means, you have ∇
⃗⃗⃗⃗ lies on the tangent plane and then this ∇
since 𝑑𝑟 ⃗ 𝑓 must be perpendicular to the tangent
plane or to the to the to the surface 𝑓(𝑥, 𝑦, 𝑧) = 𝑐.

⃗⃗⃗⃗ , and hence and therefore, perpendicular

⃗ 𝑓 will actually be perpendicular to 𝑑𝑟
So, this ∇
⃗ 𝑓 is
to the tangent plane at the point P to the surface 𝑓(𝑥, 𝑦, 𝑧) = 𝑐. And this implies that ∇
normal to the surface 𝑓(𝑥, 𝑦, 𝑧) = 𝑐.

⃗ 𝑓 or 𝛿𝑓 is actually perpendicular or normal

So, here we prove that gradient of a function ∇
to the surface 𝑓(𝑥, 𝑦, 𝑧) = 𝑐. So, you remember in the beginning when I introduced the
concepts of gradient of a function I mentioned that the gradient of a function has a physical
meaning. And this is what the physical meaning is that if you have a scalar function or a

424
surface given as 𝑓(𝑥, 𝑦, 𝑧) = 𝑐, then when you calculate the gradient as grad f then in that
⃗ 𝑓 is a vector perpendicular to the surface 𝑓(𝑥, 𝑦, 𝑧) = 𝑐 . And this result proves
case that ∇
that how to say theorem.

So, we will stop here for in this class. And I will try to introduce the concepts of directional
derivative in our next class, and then we work out few examples. So, today’s class was a
bit more theoretical as I told you right in the beginning that I will introduce the theory
wherever it is necessary and wherever it is interesting and so that is what I try to do in
today’s class. And in the next class, we will start with some examples on directional
derivative; we will introduce the concepts of directional derivative. And hopefully you
will find it interesting.

So, thank you for attention and I look forward to you in your next class.

425
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 41
Directional Derivatives (Concept & Few Results)

Hello, students. So, in the previous class, we introduced the concept of level surfaces and
we also saw two results that when you have a level surface on a region R, then at every
point of R only one level surface and that can pass through that point in R and we also saw
another result which says that gradient of a function is perpendicular to the surface
𝑓(𝑥, 𝑦, 𝑧) = 𝑐.

So, now that we have those two basic results today we will start with the concepts of
directional derivative and we will try to work out few examples based on directional
derivative. So, we will start with a formal definition what do they actually mean.

(Refer Slide Time: 01:06)

So, directional derivative of a scalar function, alright. So, the formal definition goes like
this let 𝑓(𝑥, 𝑦, 𝑧) defines a scalar field in a region R and let P be any point in this region
and suppose Q is a point in this region in the neighbourhood of the point P in the direction
of a given unit vector 𝑎̂.

426
So, this is very important we always have to have a unit vector when we are calculating
𝑓(𝑄)−𝑓(𝑃)
the directional derivative of a given unit vector 𝑎̂ . Then, lim if it exists is called
𝑄→𝑃 𝑃𝑄

the directional derivative of f at the point P in the direction of 𝑎̂.

So, basically what it means is you have a you have a point P on a surface 𝑓(𝑥, 𝑦, 𝑧) it goes
to C in a region R and then you have another point let us say Q in the neighbourhood of
the point P. So, when Q approaches to P then if this limit exists then this limit is actually
called as the directional derivative in the direction of this unit vector 𝑎̂. So, this point Q
which we have chosen so, this point Q is of course, is a point in the neighbourhood is of
course, a point in the neighbourhood of the point P, but it is also in the direction of a unit
vector 𝑎̂ .

So, it is a point in the direction of in the direction of a unit vector 𝑎̂ and then we calculate
this limit and if this limit exists, then we say that the then we say that this limit is actually
the directional derivative is the directional derivative of the function f at the point P in the
direction of 𝑎̂ , alright.

(Refer Slide Time: 05:32)

So, basically what it means is so, physical meaning or what it means is interpretation we
can write interpretation of directional derivative. So, I write it as DD. So, interpretation is
let P be any point. So, what we are saying is let P be any point (𝑥, 𝑦, 𝑧)and Q be the point

427
(𝑥 + 𝛿𝑥, 𝑦 + 𝛿𝑦, 𝑧 + 𝛿𝑧) and suppose, PQ which is a very small arc length is 𝛿𝑠. So, it is
a very small arc length and we assume that the length is 𝛿𝑠 alright.

So, suppose 𝑃𝑄 = 𝛿𝑠 then 𝛿𝑠 is a small element at P in the direction of 𝑎̂ and if 𝛿𝑓 =

𝑓(𝑥 + 𝛿𝑥, 𝑦 + 𝛿𝑦, 𝑧 + 𝛿𝑧) − 𝑓(𝑥, 𝑦, 𝑧). So, this is our point Q this is our point P. So, 𝛿𝑓 =
𝜕𝑓
𝑓(𝑥 + 𝛿𝑥, 𝑦 + 𝛿𝑦, 𝑧 + 𝛿𝑧) − 𝑓(𝑥, 𝑦, 𝑧) = 𝑓(𝑄) − 𝑓(𝑃) then represents the average
𝜕𝑠

rate of change of f per unit distance in the direction of 𝑎̂.

And, now, the directional derivative of f at the point P in the direction of 𝑎̂ is basically
𝑓(𝑄)−𝑓(𝑃) 𝜕𝑓 𝑑𝑓 𝑑𝑓
lim = lim = . So, this is nothing, but our because when 𝛿𝑠 → 0 , this
𝑄→𝑃 𝑃𝑄 𝑄→𝑃 𝜕𝑠 𝑑𝑠 𝑑𝑠
𝑜𝑟
𝛿𝑠→0
𝑑𝑓
whole thing we will converge to 𝑑𝑠 .

(Refer Slide Time: 09:04)

And, so, basically it represents the rate of change of f with respect to distance at P in the
direction of 𝑎̂ which is a unit vector. So, what it means physically is that we have two
neighbouring points on the surface 𝑓(𝑥𝑦, 𝑧) = 𝑐 and then that is small increment in the
function f or from the point P to Q is denoted by 𝛿𝑓and 𝛿𝑠. So, that a small increment in
the function f we write it as 𝛿𝑓 which is basically (𝑓(𝑄) − 𝑓(𝑃)) and the small element
which is this PQ is basically 𝛿𝑠.

428
 𝜕𝑓
So, when 𝛿𝑠 → 0 is actually the average rate of change of f per unit distance. So, how
𝜕𝑠
𝜕𝑓
much the function is changing how it is changing per unit distance is given by in the
𝜕𝑠

direction of the vector 𝑎̂.

So, we always have a unit vector along which we are calculating the rate of change of the
function f and the directional derivative of the function f at this point P is basically when
𝑓(𝑄)−𝑓(𝑃) 𝑑𝑓
lim = 𝑑𝑠 and at the point P we have this 𝛿𝑠 this surface element is that a small
𝑄→𝑃 𝑃𝑄

element at the point P in the direction of 𝑎̂. So, that is what we are calculating here. So,
this is basically our rate of change of the function f per unit distance in the direction of 𝑎̂ at
the point P.

So, this is represent the rate of change of the function f with respect to distance or per unit
distance at the point P in the direction of 𝑎̂. So, 𝑎̂ is a unit vector at the point P and this
𝑑𝑓
is actually denoting our rate of change of the function f. So, this is what we mean
𝑑𝑠

physically by the directional derivative. So, at a certain point you have a unit vector and
you need to calculate the rate of change of that function along that unit vector. So, that is
what simple and an in simple words it mean.

So, you have a unit vector at a point P on a surface 𝑓(𝑥, 𝑦, 𝑧) = 𝑐. So, your directional
derivative is actually the rate of change of that function 𝑓(𝑥, 𝑦, 𝑧) with respect to distance
of course, in the direction of that unit vector at the point P. So, or whichever point it is
where you are calculating the directional derivative and this is what we mean physically
or the physical interpretation of the direction and derivative.

𝑑𝑓
Now, that we have stated is basically that rate of change of f with respect to distance
𝑑𝑠

how do we calculate the directional derivative do we really have to differentiate

𝑑𝑓 𝑑𝑓
𝑓(𝑥, 𝑦, 𝑧) = 𝑐 as 𝑑𝑠 or do we have to calculate to calculate the directional derivative or
𝑑𝑠

is there any other formula to calculate the directional derivative so, that we will now prove,
in terms of a small theorem. So, whether we calculate this thing or whether there is some
other tool that will help us calculate the directional derivative. So, that is something we
are going to see now.

429
So, let me put a small theorem here. So, today it is theorem 1, and it says that the directional
derivative of a scalar field f at a point 𝑃(𝑥, 𝑦, 𝑧) in the direction of a unit vector 𝑎̂ is given
𝑑𝑓
by 𝑑𝑠 = ⃗∇𝑓𝑃 . 𝑎̂ . So, we have to calculate the gradient of the function at the point face. So,

I can put a P here and that basically says that the gradient of the function has to be
calculated at the point P times the unit vector 𝑎̂.

𝑑𝑓
So, that means, instead of calculating we can just using this theorem the directional
𝑑𝑠

derivative of the function f at the point P in the direction of a is given by ⃗∇𝑓 at the point P
𝑑𝑓
dot product with 𝑎̂. So, this = ⃗∇𝑓𝑃 . 𝑎̂.
𝑑𝑠

𝑑𝑓
So, this is a very small result and we will try to prove that and see how this thing this 𝑑𝑠 =

⃗ 𝑓𝑃 . 𝑎̂ .
∇

(Refer Slide Time: 14:45)

So, the proof or the solution. So, let 𝑓(𝑥, 𝑦, 𝑧) be a scalar field in the region R and let
𝑃(𝑥, 𝑦, 𝑧) be any arbitrary point in the region R. So, 𝑃(𝑥, 𝑦, 𝑧) belongs to R. So, then I can
𝑂𝑃 or the position vector of the point P is equals to 𝑥𝑖̂ + 𝑦𝑗̂ + 𝑧𝑘̂, alright.
write ⃗⃗⃗⃗⃗

Now, if S denotes the distance of P from some fixed point A in the direction of 𝑎̂ then
𝑑𝑟
𝛿𝑥 denotes a small element at P in the direction of 𝑎̂ and therefore, is a unit vector at
𝑑𝑠
𝑑𝑟
the point P in this direction that is 𝑑𝑠 = 𝑎̂.

430
So, what we are doing is we are assuming the distance of P from any fixed point A in the
direction of 𝑎̂. So, in the direction of 𝑎̂ we assume any arbitrary point let us say A, and S
denotes the distance and then in that case a small element or a small increment in the point
P in the direction of 𝑎 then that small increment is denoted by 𝛿𝑥. So, if we are moving
along the surface and from P to let us say Q or P to a we are going then that small increment
is basically our 𝛿𝑥 and then in the direction of 𝑎 of course,.

𝑑𝑟
So, therefore, basically the rate of change of 𝑟 with respect to the distance is a unit
𝑑𝑠
𝑑𝑟
vector in the direction of this unit vector 𝑎̂. So, this basically 𝑑𝑠 is a unit vector at the point
𝑑𝑟 𝑑𝑥
P in the direction of 𝑎̂, alright. So, but our 𝑟 = 𝑥𝑖̂ + 𝑦𝑗̂ + 𝑧𝑘̂. So, from here = 𝑖̂ +
𝑑𝑠 𝑑𝑠
𝑑𝑦 𝑑𝑟
𝑗̂ + 𝑑𝑠 𝑘̂ our alright.
𝑑𝑠

𝜕𝑓 𝜕𝑓 𝜕𝑓
⃗ 𝑓. 𝑎̂, what is this? So, this one will be 𝑎1 + 𝑎2 + 𝑎3 right. So, this 𝑎
So, now our ∇ 𝜕𝑥 𝜕𝑦 𝜕𝑧

𝑑𝑟 𝑑𝑥 𝑑𝑦
actually. So, this 𝑎1 , 𝑎2 , 𝑎3 will be actually we can write it now. So, 𝑑𝑠 = 𝑖̂ + 𝑑𝑠 𝑗̂ +
𝑑𝑠
𝑑𝑟
𝑘̂ = 𝑎̂ . So, that means, if I write this one as 𝑎̂, so, the components of 𝑎̂ will be
𝑑𝑠

𝑎1 , 𝑎2 , 𝑎3 .

𝑑𝑥 𝑑𝑦 𝑑𝑧 𝜕𝑓 𝑑𝑥 𝜕𝑓 𝑑𝑦
So, 𝑎1 = , 𝑎2 = and 𝑎3 = 𝑑𝑠 . So, I am substituting them here. So, 𝜕𝑥 . 𝑑𝑠 + 𝜕𝑦 . 𝑑𝑠 +
𝑑𝑠 𝑑𝑠
𝜕𝑓 𝑑𝑧 1 𝜕𝑓 𝜕𝑓 𝜕𝑓
. . So, since dz and 𝜕𝑧 they are a small element. So, ( 𝑑𝑥 + 𝜕𝑦 𝑑𝑦 + 𝜕𝑧 𝑑𝑧) =
𝜕𝑧 𝑑𝑠 𝑑𝑠 𝜕𝑥
1 𝑑𝑓
𝑑𝑓 = .
𝑑𝑠 𝑑𝑠

431
(Refer Slide Time: 20:38)

𝑑𝑓
Now, we know that 𝑑𝑠 is the directional derivative at the point P in the direction of 𝑎. So,

that is what physically it means from the definition of directional derivative.

So, this is our directional derivative of the function f at the point P in the direction of 𝑎.
𝑑𝑓
So, what we have is we have ⃗ 𝑓𝑃 . 𝑎̂ and if we want to calculate a directional
=∇
𝑑𝑠

derivative, then we can calculate the gradient of the function f at the point P, this is very
important times the unit vector 𝑎̂. So, this is the required formula to calculate directional
derivative, alright. So, and this is what we wanted to prove.

So, the direction derivative of a scalar field at the point P in the direction of a unit vector
𝑎̂ is given in this fashion and just assuming some basic results from the differential calculus
and not assuming, but using those results from the differential calculus we can be able to
𝑑𝑓
show that the direction derivative = ⃗∇𝑓𝑃 . 𝑎̂ and this is what is equal to our directional
𝑑𝑠

derivative, alright.

So, now we will try to work out few examples before we go to our next topic. So, let us
calculate the directional derivative we will work out few examples on level surfaces,
alright.

432
(Refer Slide Time: 22:47)

So, in the previous class we started with level surfaces and in today’s class we started with
directional derivative. So, we will try to solve the examples on both of these two topics.
So, the first one is first example find a unit normal vector to the level surface 𝑥 2 𝑦 + 2𝑥𝑧 =
4 at the point 𝑃(2, −2,3). So, our given level surfaces 𝑥 2 𝑦 + 2𝑥𝑧 = 4 and we have to
calculate the unit normal vector at the point. So, let us put a P here at the point 𝑃(2, −2,3).

So, what we are going to do? First of all from the results on a level surfaces we know that
gradient of f is a normal to this level surface. So, here it says that find a unit normal. So,
we have to find a normal to this surface given here and then dividing it with it is magnitude
will give us the unit normal, alright.

So, those two steps are clear. So, first of all we have to calculate its gradient. So, here the
equation of the level surface is 𝑓(𝑥, 𝑦, 𝑧) = 𝑥 2 𝑦 + 2𝑥𝑧 − 4 , alright and we know that
⃗ 𝑓 is perpendicular or normal to the surface f. Then, first of all we calculate the ∇
∇ ⃗ 𝑓 would
𝜕
⃗ . (𝑥 2 𝑦 + 2𝑥𝑧 − 4). So, calculating this one is not difficult. So, (𝑥 2 𝑦 + 2𝑥𝑧 − 4)𝑖̂ +
be ∇ 𝜕𝑥
𝜕 𝜕
(𝑥 2 𝑦 + 2𝑥𝑧 − 4)𝑗̂ + (𝑥 2 𝑦 + 2𝑥𝑧 − 4)𝑘̂ = (2𝑥𝑦 + 2𝑧)𝑖̂ + 𝑥 2 𝑗̂ + 2𝑥𝑘̂. So, this is the
𝜕𝑦 𝜕𝑧

required gradient of f.

And, now, the ⃗∇𝑓 at the point P and P is (2, −2,3). So, if I substitute then it will be
[2.2. (−2) + 2.3]𝑖̂ + [22 ]𝑗̂ + 2.2𝑘̂. So, this will be ultimately an on −2𝑖̂ + 4𝑗̂ + 4𝑘̂ and
therefore, this ⃗∇𝑓.

433
(Refer Slide Time: 26:41)

So, the ⃗∇𝑓 or −2𝑖̂ + 4𝑗̂ + 4𝑘̂ is normal, or perpendicular to the surface is normal or
perpendicular to the surface 𝑓(𝑥, 𝑦, 𝑧)alright, but we have to find out a unit normal or unit
̂
−2𝑖̂+4𝑗̂ +4𝑘
vector perpendicular to this. So, to give the unit normal we write it as =
√4+16+16
̂
−2𝑖̂+4𝑗̂ +4𝑘
.
2√10

So, ultimately this will be −2𝑖̂ + 4𝑗̂ + 4𝑘̂ and then this will be 32 + 40. So, we can write
2√10 is a unit normal or unit vector perpendicular to the surface 𝑓(𝑥, 𝑦, 𝑧) = 𝑐 and
𝑓(𝑥, 𝑦, 𝑧) is basically our 𝑥 2 𝑦 + 2𝑥𝑧 = 4. So, that c is 4.

So, this is the required this is the weight this one is a 16, 16, 32 + 4, 36 ok. So, this has to
̂
−2𝑖̂+4𝑗̂ +4𝑘 ̂
−2𝑖̂+4𝑗̂ +4𝑘
be = . So, this is the required unit normal perpendicular to the surface
√4+16+16 6

𝑓(𝑥, 𝑦, 𝑧) = 𝑐 and you see we had a equation of a level surface here in this example we
had a question of a level surface.

So, we wrote this level surface as 𝑓(𝑥, 𝑦, 𝑧) = 𝑐 and then we had to calculate the gradient
because gradient is a vector normal to the level normal to the surface 𝑓(𝑥, 𝑦, 𝑧) = 𝑐 and
then we calculated the gradient at the point P and then we divided it with its magnitude
and that gave was the unit normal or unit vector perpendicular to the surface 𝑥 2 𝑦 + 2𝑥𝑧 =
4 at the point (2, −2,3) you also have to make sure whether this point lies on that level

434
surface on that level surface or not because we can do all this calculation but, always make
sure that this point lies there.

So, if I substitute there then this will be 4. So, 22 . (−2) + 2.2.3 = −8 + 12 = 4 to also of
course, this point lies on the level surface. So, before you calculate anything you also have
to make sure the given point P lies on the level surface or not. So, here in this case it does
lying on that level surface and therefore, the gradient of f is actually normal to that level
surface at the point P and if you divide it with its magnitude then that is the unit normal;
⃗ 𝑓̂, alright.
here I can write it as ∇

So, today we saw one example motivated from the level surface and unit normal. In our
next class we will continue with the examples on level surfaces and directional derivative
just to make the concepts clear and I look forward to you in your next class.

Thank you.

435
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 42
Directional Derivative (Concept & Few Results) (Contd.)

Hello students. So, in the previous class we started with examples actually based on level
surfaces and how do we calculate the normal on a given surface and also we introduced
the concepts of directional derivative.

So, today we will practice few examples based on a Directional Derivative and afterwards
if time permits then we move to our next topic. So, to start with let me start with one
example on directional derivative.

(Refer Slide Time: 00:55)

So, today our example one; so, the example goes like this. So, find the directional
derivative of the scalar function 𝑓(𝑥, 𝑦, 𝑧). So, we have a function in 3 dimension geometry
this function is 𝑓(𝑥, 𝑦, 𝑧) = 𝑥 2 𝑦𝑧 + 4𝑥𝑧 2 at the point. So, as we know we always have to
calculate the directional derivative at a certain point and at the same time we also need to
have a unit vector towards in the direction of which we will calculate the directional
derivative.

436
So, at the point what is the point here? The point is (1, −2,1) in the direction. So, this is
our direction of the vector 2𝑖̂ − 𝑗̂ − 2𝑘̂ alright; 𝑖̂, 𝑗̂, 𝑘̂ are the unit vectors. So, let me adjust
my chair first alright; so, here; so now we are going to solve this.

So, here the given scalar function the given function is 𝑓(𝑥, 𝑦, 𝑧) = 𝑥 2 𝑦𝑧 + 4𝑥𝑧 2 . And in
the previous class we saw the direction derivative of a given scalar function at a point P is
given as gradient of that function at the point P, times the unit vector in the direction of
which we are calculating the directional derivative. So, here first of all we need to calculate
the gradient of this function at the point P; so, this is our point P.

⃗ 𝑓 which is a vector quantity and ∇

So, we will calculate ∇ ⃗ 𝑓 = 𝜕𝑓 𝑖̂ + 𝜕𝑓 𝑗̂ +
⃗ 𝑓 would be ∇
𝜕𝑥 𝜕𝑦
𝜕𝑓
𝑘̂. So, if I do partial derivative with respect to x then this will be ⃗∇𝑓 = (2𝑥𝑦𝑧 + 4𝑧 2 )𝑖̂ +
𝜕𝑧

[𝑥 2 𝑧]𝑗̂ + (𝑥 2 𝑦 + 8𝑥𝑧)𝑘̂

alright; so that is our ⃗∇𝑓. But we need to calculate the gradient of the function at the point
P because that is where we are calculating the directional derivative. So, what is the value
of ⃗∇𝑓 at the point P? So, this is one of the notations to write ⃗∇𝑓 𝑎𝑡 𝑡ℎ𝑒 𝑝𝑜𝑖𝑛𝑡 (1, −2, −1).
So, the value is we substitute all these things [2.1. (−2). (−1) + 4(−1)2 ]𝑖̂+. . ... So, this is
very simple to do and I am pretty sure you can be able to do this simple calculation ; 𝑖̂ plus
you substitute the value for the rest of the variables here in 𝑗̂ and 𝑘̂.

So, ultimately if you calculate then at the end you will be able to obtain 8𝑖̂ − 𝑗̂ − 10𝑘̂ . So,
that is the gradient of the function f at the point P and it is being calculated for substituting
𝑥 = 1, 𝑦 = −2, 𝑧 = −1 in this expression here alright; so that is our gradient of f.

Now, in order to calculate the directional derivative we need the unit vector. So, here in
this case we have the given vector, but is it a unit vector? It is; obviously, not because if it
is a unit vector then the modulus of this vector should be 1. So, that is what we mean by
unit vector that the mod of that vector is equals to 1, but in this case it is certainly not. So,
for this vector we will obtain a unit vector and we will see how we do that.

437
(Refer Slide Time: 05:57)

So, the vector the vector in the direction of which the directional derivative D D is
supposed to be calculated; is given by. So, what is the vector? We have 𝑎 = 2𝑖̂ − 𝑗̂ − 2𝑘̂.
So, this is my normal this, this how to say a normal vectors not that normal. So, this is
basically the given vector do not focus on normal or anything else. So, this is the given
vector from here I have to obtain a unit vector.

So, the unit vector will be obtained by dividing the vector 𝑎; by dividing the vector 𝑎 with
̂
2𝑖̂−𝑗̂ −2𝑘 ̂
2𝑖̂−𝑗̂ −2𝑘 1
its magnitude right. So, the given vector is 𝑎̂ = √22 = = 3 (2𝑖̂ − 𝑗̂ − 2𝑘̂ ).
+12 +22 √9

now it is a unit vectors. If we take the how to say magnitude of this 𝑎̂ then this will yield
1; so; that means, this is a unit vector. Therefore, the required directional derivative of f at
P; what is our P? P is (1, −2, −1) in the direction of 𝑎̂; is given by directional derivative
𝐷𝑓𝑎̂ = ⃗∇𝑓𝑃 . 𝑎̂.

Here instead of 𝑎̂; I will write a because we are calculating basically in the direction of 𝑎.
So, it does not matter whether we write 𝑎̂; it is just that if we write 𝑎 here the in the formula
we should be careful that we always write the unit vector because here we have a unit
vector always in the formula. So, the directional derivative I have chosen a notation
𝐷𝑓 although the notation does not play a big role here, you can ignore the notation and use
your own notation for the directional derivative; this is not universal in a way this is just

438
my notation to signify that the directional derivative of the function f in the direction of
𝑎̂ alright.

⃗ 𝑓𝑃 . 𝑎̂ = (8𝑖̂ − 𝑗̂ − 10𝑘̂). 1 (2𝑖̂ − 𝑗̂ − 2𝑘̂). Now, we calculate this

So, now here we have ∇ 3
1
dot product and this will be 3 [16 + 120] isn’t it? Yes.

37
So, we will ultimately obtain ; so directional derivative of the function f at the point P in
3
37
the direction of this vector 𝑎 is given as . So, of course, it is a scalar quantity and its
3

calculated in this fashion alright. So, this was our first example we will practice few I mean
like 2 or 3 more examples just to make the concept clear.

(Refer Slide Time: 10:16)

Next let us consider an example find the directional derivative; of a scalar point function;
f in the direction of coordinate axes; so, this is our example 2. So, it is also a very
interesting example we have to calculate the directional derivative of a function f in the
direction of coordinate axes. So; that means, if I draw my coordinate axes; so, this is X
axis this is my Y axis and this is Z axis right.

So, we have to calculate directional derivative in the direction of coordinate axess; that
means, in the direction of X axis in the direction of Y axis and in the direction of Z axis.
So, when we are calculating in the direction of X axis the directional derivative. So, then
in that case what will be our unit vector.

439
So, if we are calculating the directional derivative in the direction of X axis; that means,
the unit vector is basically the vector along this direction is actually 𝑖̂ because X axis is
perpendicular to YZ plane and the unit vector for this X axis in a way is actually 𝑖̂. So,𝑖̂ is
the unit vector in the direction of X axis and our 𝑎̂ previously in the formula is actually 𝑖̂.
So, 𝑖̂ is the unit vector which is perpendicular to YZ plane and it is along the direction of
X axis and therefore, we use it as a normal in this case alright.

So, let us see; what is our direction derivative in the direction of X axis first. So, we have
to calculate along the coordinate axes; that means, we have to calculate the direction
derivative along the X axis, along the Y axis and along the Z axis. So, we have to calculate
the directional derivative along all these 3 axes alright; so, let us see.

𝜕𝑓 𝜕𝑓 𝜕𝑓
⃗ 𝑓𝑃 = 𝑖̂ + 𝑗̂ + 𝑘̂ alright.
let me write for the given scalar point function 𝑓(𝑥, 𝑦, 𝑧), ∇ 𝜕𝑥 𝜕𝑦 𝜕𝑧

So, that is our gradient of f and at any point P; so that point P is (𝑥, 𝑦, 𝑧). So, I can write it
as P and then I can write P out here.

So, this is a gradient of the function f at any point P(𝑥, 𝑦, 𝑧). and therefore, the directional
⃗ 𝑓𝑃 . 𝑎̂ = ∇
derivative along X axis is; so, I will write 𝐷𝑓𝑃 = ∇ ⃗ 𝑓𝑃 . 𝑖̂. So, along X axis the unit

vector is given as 𝑖̂ alright. So, we will write ⃗∇𝑓𝑃 . 𝑖̂. .

𝜕𝑓 𝜕𝑓 𝜕𝑓
Now, if we take the dot product of (𝜕𝑥 𝑖̂ + 𝜕𝑦 𝑗̂ + 𝜕𝑧 𝑘̂) . 𝑖̂; then in that case only first term

will survive the rest of the term will be 0 because the component of 𝑗̂ and 𝑘̂ is 0 in case of
this vector. So, here in case of this vector; 𝑗̂ and 𝑘̂ components are 0. So, when we take the
𝜕𝑓
dot product only 𝜕𝑥 we will survive alright.

𝜕𝑓 𝜕𝑓 𝜕𝑓 𝜕𝑓
So, (𝜕𝑥 𝑖̂ + 𝜕𝑦 𝑗̂ + 𝜕𝑧 𝑘̂) . 𝑖̂ . So, if I take the dot product then this will yield . So, that is
𝜕𝑥

the direction derivative of the function f at any point P in the direction of X axis alright.
Similarly the directional derivative of the function f at the point P along Y axis is we will
𝜕𝑓
proceed in the similar fashion and this will be basically 𝐷𝑓𝑃 = 𝜕𝑦. And similarly the
𝜕𝑓 𝜕𝑓
directional derivative along Z axis for the function f at the point P is 𝜕𝑧 ; so, 𝐷𝑓𝑃 = .
𝜕𝑧

So, basically in this case we got 3 directional derivative and each one of them are directed
along along three different directions. So, the first one is directed directional derivative

440
along X axis the second one is directional derivative along Y axis and third one is
directional derivative along Z axis. And you just have to know that 𝑖̂, 𝑗̂ and 𝑘̂ are the 3 unit
vectors along X, Y and Z axes respectively and rest of the things are very easy.

So, this is another interesting example. So, you might be given a function f has 𝑥 2 𝑦 +
3𝑦 2 𝑧 and then you will be asked to calculate the directional derivative along X axis. So,
you just have to calculate these gradients and based on that you just take the dot product.
𝜕𝑓
And if you are asked to calculate just along the X axis then you just calculate 𝜕𝑥 and that

will be a required answer at a point P. So, this is an interesting example which I wanted to
show; we will move on to our next example; example 3.

(Refer Slide Time: 17:22)

So, find the directional derivative of the function 𝑓(𝑥, 𝑦, 𝑧) = 𝑥 2 − 𝑦 2 + 2𝑧 2 at the point;
at the point 𝑃(1,2,3) in the direction of the line PQ; where Q is the point (5,0,4) alright.

So, here we are asked to calculate the directional derivative of the function f which is given
at the point P. But instead of giving a vector in the direction of which we are supposed to
calculate the directional derivative we are given an another point. And the vector which
you obtain by joining P to Q along which we have to calculate the directional derivative.

Now if you are familiar with vector algebra if you have two points; let us say P and Q then
⃗⃗⃗⃗⃗ = 𝑂𝑄
the vector 𝑃𝑄 ⃗⃗⃗⃗⃗⃗ − 𝑂𝑃
⃗⃗⃗⃗⃗ . So, you take the distance from the origin and then you take
the; for Q you take the distance for P from the origin and then you subtract them and that

441
will give you the required vector⃗⃗⃗⃗⃗⃗
𝑃𝑄 or P to Q. So, that is some that is exactly what we
are going to do here.

So, first of all in order to calculate the directional derivative we need to calculate the
gradient of the given function; so, that is where we start from. So, here the given function
𝜕𝑓 𝜕𝑓 𝜕𝑓
is 𝑓(𝑥, 𝑦, 𝑧) = 𝑥 2 − 𝑦 2 + 2𝑧 2 . Then ⃗∇𝑓 = 𝜕𝑥 𝑖̂ + 𝜕𝑦 𝑗̂ + 𝜕𝑧 𝑘̂ = 2𝑥𝑖̂ − 2𝑦𝑗̂ + 4𝑧𝑘̂. So, that

is the required gradient of the function f at the point P and in general the gradient of the
function is given by this one.

Now, we have to calculate the gradient of the function at the point P. So, the point P is
(1,2,3). So; that means, we have ⃗ 𝑓𝑃(1,2,3) = 2.1𝑖̂ − 2.2𝑗̂ + 4.3𝑘̂ = 2𝑖̂ − 4𝑗̂ +
∇
12𝑘̂ alright. So, this is the required gradient of the function f at the point P. Now, we have
to calculate the direction derivative along the direction of P to Q.

(Refer Slide Time: 21:30)

So, the vector or the direction vector is P to Q alright. So, P to Q can be given that is let
⃗⃗⃗⃗⃗ = 𝑂𝑄
me write this in a clear manner. So, that is 𝑃𝑄 ⃗⃗⃗⃗⃗ = (5𝑖̂ + 0𝑗̂ + 4𝑘̂) − 𝑖̂ − 2𝑗̂ −
⃗⃗⃗⃗⃗⃗ − 𝑂𝑃

3𝑘̂ = 4𝑖̂ − 2𝑗̂ + 𝑘̂ = 𝑎 alright. So, along direction along this vector P to Q we have to
calculate the directional derivative, I will call it as let us say a vector. So, say ⃗⃗⃗⃗⃗
𝑃𝑄 is our
vector 𝑎 alright

442
 𝑎⃗ ̂
4𝑖̂−2𝑗̂ +𝑘 ̂
4𝑖̂−2𝑗̂ +𝑘
So, from here I can write say. So, from here 𝑎̂ = |𝑎⃗| = = alright. And now
√16+4+1 √21

the required directional derivative of f at the point P is given ⃗∇𝑓𝑃 . 𝑎̂.

So, gradient of f at the point P is 2𝑖̂ − 4𝑗̂ + 12𝑘̂. So, this is ∇

⃗ 𝑓𝑃 . 𝑎̂ = (2𝑖̂ − 4𝑗̂ +
̂
4𝑖̂−2𝑗̂ +𝑘
12𝑘̂). . So, whatever you get up to the simplification. So, here you just have to take
√21
28 4
the dot product and ultimately we will be able to obtain 21
√21 = 3 √21. So, that is the
required directional derivative of constant f in the direction of P to Q

So, here the direction vector let us say is not given and we had to calculate the direction
⃗⃗⃗⃗⃗ = 𝑂𝑄
vector by just; I mean how to say writing 𝑃𝑄 ⃗⃗⃗⃗⃗⃗ − 𝑂𝑃
⃗⃗⃗⃗⃗ you just substitute the value,
take the difference and that is a required direction vector. And from there we have to
calculate the unit vector should be divided it by its magnitude and that gave us the give us
⃗ 𝑓𝑃 . 𝑎̂ and
𝑎̂. And therefore, the required directional derivative of f at the point P is the ∇
that is your required answer alright.

(Refer Slide Time: 24:58)

Now, we will calculate another example so, what is; so we know that directional
derivatives is the rate of change of function f with respect to the distance in the direction
of a unit vector.

443
So, D f D s is the directional derivative or the rate of change of the function f with respect
to the distance or length in the direction of a vector 𝑎̂ or a whatever we would like to say
⃗ 𝑓𝑃 . 𝑎̂ alright.
basically 𝑎̂ So, from that formula the directional derivative is nothing, but ∇

So, this directional derivative it will be maximum in what direction I mean; what is the
maximum value of this directional derivative and in which direction would it be
maximum? So, the thing is the maximum value would be attained if I mean if we consider
this 𝑎̂ as a unit vector. So, it is actually a unit vector, but if let us say if that is 𝐷𝑓 = ⃗∇𝑓. 𝑎̂ =
⃗
⃗ 𝑓. ∇𝑓 = |∇
∇ ⃗ 𝑓|; that means, its maximum along the direction of gradient basically.
⃗ 𝑓|
|∇

So, instead of considering 𝑎 is equals to any arbitrary vector a vector if we consider a⃗ =

⃗∇𝑓 . So, that is what I am trying to say that if we consider a⃗ = ⃗∇𝑓 then in that case 𝑎̂ =
⃗∇𝑓
⃗ 𝑓|.
and when we take the dot product then it is will be basically |∇
|∇⃗ 𝑓|

So; that means, the maximum value is attained when the directional derivative is equals to
⃗∇𝑓
the |∇⃗𝑓| . So, that is the maximum value. So, what I am trying to say is that if we have

𝐷𝑓 = ⃗∇𝑓𝑃 . 𝑎̂ if I write 𝑎̂. So, 𝑎̂ is basically a unit vector along which we are calculating the
⃗ 𝑓; then in that case this will be |∇
direction derivatives of 𝑎̂ is equals to ∇ ⃗ 𝑓| and there is the

this is the maximum value for the directional derivative.

So, in the direction of gradient of f the maximum value is attained. So, basically the
maximum value of the directional derivative is attained in the direction of gradient of f.
So, instead of taking any vector 𝑎 if you take gradient of f itself alright. So, you have a
⃗
⃗ 𝑓 and here instead of any 𝑎 if as 𝑎̂; if we take ∇𝑓 . ∇
given surface if you take 𝑎 as ∇ ⃗𝑓,
⃗ 𝑓|
|∇

⃗ 𝑓| as the directional derivative.

then we will basically left with |∇

And that is the maximum value of the directional derivative or maximum rate of change
of the function 𝑢 = 𝑥𝑦𝑧 2 at the point (1,0,3) given as a scalar function. So, this is a very
vital result that in the direction of maximum rate of change. So, that direction is basically
the direction of the gradient of f and this is the message I am trying to convey.

⃗ 𝑓; so, along which the function

So, the direction of maximum rate of change is basically ∇
is how to say has the greatest or attains the maximum rate of change. And that can be

444
basically the directional derivative along the direction of maximum rate of change; it is
⃗ 𝑓| alright.
given by |∇

So, we will do the same thing here; so basically, we will try to solve this example in our
next class. And we will stop here for today and we will continue with the same example
in our next class. We will try to solve this and then we will continue with some more
examples on directional derivative.

So, I thank you for your attention for today and I will see you in the next class.

445
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture - 43
Directional Derivatives, Level Surfaces

Hello students. So, today we will again continue with our examples on Directional
Derivative. It is a very interesting topic. So, that is why, I am trying to cover as many
examples as possible motivated from Directional Derivative and Level Surfaces, tangent
planes and things like that. So today, we will continue with the example which we left off
last time. So, in the last class, we had a problem on directional derivative which was this
one.

(Refer Slide Time: 00:48)

So, what is the greatest rate of change for the function 𝑢 = 𝑥𝑦𝑧 2 at the point (1,0,3), all
right. So, let us start with this problem here. So, the given scalar function or the givens
⃗ 𝑢 = 𝜕𝑢 𝑖̂ + 𝜕𝑢 𝑗̂ + 𝜕𝑢 𝑘̂. So,
function, simply is 𝑢(𝑥, 𝑦, 𝑧) = 𝑥𝑦𝑧 2 . So, then here, then ∇ 𝜕𝑥 𝜕𝑦 𝜕𝑧
𝜕𝑢
what is 𝜕𝑥 ? It is 𝑦𝑧 2 𝑖̂ + 𝑥𝑧 2 𝑗̂ + 2𝑥𝑦𝑧𝑘̂, all right. So, that is our ⃗∇𝑢. And now, at the point

⃗ 𝑢𝑃 = 0𝑖̂ + 32 𝑗̂ + 0𝑘̂ = 9𝑗̂. So, basically 9𝑗̂, all right

⃗𝑢=∇
therefore, at the point (1,0,3), ∇
⃗ 𝑢 at the point P. Now, we have to calculate
so that is our so basically 9𝑗̂, so that is our ∇

446
the greatest rate of change, our rate of increase of u; that means, we have to calculate the
maximum value.

So, we have to calculate the maximum value of f at the point, (1,0,3). So, greatest rate of
change or greatest rate of increase is nothing but the maximum value of 𝐷𝑓𝐷 s because
DfD is the rate of change of f all right. So, this greatest rate of increase is nothing but the
maximum value of f at the point (1,0,3).

Now, as I was speaking in the previous class that the maximum value of 𝐷𝑓𝐷 s will be
attained if we consider the vector a as gradient of f. So, if we are calculating the directional
derivative along the direction of gradient, then the I mean that is the basically the
maximum rate of change or maximum value of 𝐷𝑓𝐷 s will be attained.

(Refer Slide Time: 04:11)

∇𝑢 ⃗
So; that means, here, e what we will have is we will have ⃗∇𝑢𝑃 . |∇⃗𝑢𝑃 | . So, if we consider
𝑃

⃗𝑢
∇
the vector 𝑎̂ = |∇⃗𝑢𝑃 |, so that is where a maximum value would be attained. And here, we
𝑃

⃗ 𝑢𝑃 | because this will be

can take the dot product. So, ultimately we will be left with |∇
⃗ 𝑢𝑃 |2
|∇
and ⃗ 𝑢𝑃 | = |∇
one of them we will get cancel out. So, we will have basically |∇ ⃗ 𝑢|, I
⃗ 𝑢𝑃 |
|∇

am, so I am just avoiding P now.

447
 ⃗ 𝑢|at the point P and this value is nothing but |9𝑗̂| right. We
So, we will basically obtain |∇
have already calculated this value here. So, I substitute gradient of u at the point P as let
me write it has |9𝑗̂| = 9 because it will be √02 + 92 + 02 = 9. So, the second component
is92 . So, square root out of square root, it will become again line and the rest of the term
will be one. So, ultimately that is our required maximum or greatest rate of increase for
the function f at the point P equals to (1,0,3).

So, this is the required answer. So, this is a very vital thing to know that in which direction
the maximum rate of change is occurring and that is the direction of actually gradient of f.
So, if instead of any other vector, if you consider the vector a as gradient of f, then along
the direction of gradient of f, the maximum rate of change or the maximum value of 𝐷𝑓𝐷
s would be at the end.

So, you do not have to calculate any other thing if it is says that calculate the directional
derivative along the direction of maximum rate of change. So, all you have to do is
calculate the gradient of the function at the point P and then take it is mod and that will be
the required answer, like in this case what we have obtained all right.

(Refer Slide Time: 06:34)

Now, let us consider an another example. I have lost the track. So, I am just writing
example 4, all right. Now, in this case, show that the directional derivative of a scalar point
function at any point along any tangent line to the level surface at that point is 0.

448
So, what here says is that we have a scalar point function, let us say f and we want to
calculate the directional derivative along any tangent line to the level surface. So, if we
have any tangent line to the level surface and if we are calculating the directional derivative
along that tangent line, then in that case, that directional derivative would be 0. So, this
would be true for any kind of level surface 𝑓(𝑥, 𝑦, 𝑧) = 𝑐. If we calculate the direction
derivative along the tangent line, then in that case it will be always 0.

So, let us see how we can prove that. So, we will start with, let 𝑓(𝑥, 𝑦, 𝑧) be a scalar point
function and let 𝑎̂ be a unit vector along a tangent line to the level surface and the equation
of the level surfaces 𝑓(𝑥, 𝑦, 𝑧) = 𝑐. So, we assume that 𝑓(𝑥, 𝑦, 𝑧) be there scalar point
function and 𝑎̂ with a unit vector along the direction of tangent line to the surface
𝑓(𝑥, 𝑦, 𝑧) = 𝑐. So, this is our required level surface. So, this is up to this we assumed.

So, now we know that gradient of f is a normal vector at any point of the surface
𝑓(𝑥, 𝑦, 𝑧) = 𝑐 all right. Therefore, the vectors ⃗∇𝑓 and 𝑎̂ must be perpendicular to one
another. So, this is very simple. If 𝑎̂ is the unit vector and if 𝑎̂ is along the direction of
tangent line, so we have a vector along the tangent line and ⃗∇𝑓 .

⃗ 𝑓. 𝑎̂ where 𝑎̂
Now, remember to calculate the directional derivative; we always need that ∇
is the unit vector in the direction of which we are calculating the directional derivative.
⃗ 𝑓 is normal to this level surface
Now, if 𝑎̂ is along the direction of tangent line and since ∇
⃗ 𝑓 must be normal or must be perpendicular to 𝑎̂ because
𝑓(𝑥, 𝑦, 𝑧) = 𝑐 , then in that case ∇
𝑎̂ is in is along the direction of tangent line and we have a perpendicular vector on the
surface.

So, it is a perpendicular to the tangent line as well and therefore, it is perpendicular to that
𝑎̂ as well. So, that is what we are saying. Therefore, the vectors ⃗∇𝑓 and 𝑎̂ must be
⃗ 𝑓. 𝑎̂ =
perpendicular to one another, but from the condition of perpendicularity, we have ∇
0 because they are perpendicular to one another, their dot product must be 0.

449
(Refer Slide Time: 11:58)

And if their dot product is 0, then the directional derivative of f is given by D f at the point
P is equals to ⃗∇𝑓. 𝑎̂ = 0. And therefore, the directional derivative of the function f along
any tangent line in the direction of any tangent line is always it goes to 0 because gradient
of f and that vector along the tangent line are mainly perpendicular to each other and
therefore, this is our required result. Next problem is motivated from the level surface and
the tangent plane normal to the surface, all right.

So, let me write the problem. So, find the equation of the tangent plane and normal to the
surface2𝑥𝑧 2 − 3𝑥𝑦 − 4𝑥 = 7 at the point (1, −1,2) . So, the given equation of the
surfaces, so we have to calculate the tangent plane and the normal for this surface. So,
these now we are doing both types of examples there.

So, the examples motivated from directional derivative and examples motivated from level
surfaces. So, I am trying to solve examples on both of these two topics one by one. So, the
given surface is 𝑓(𝑥, 𝑦, 𝑧) = 2𝑥𝑧 2 − 3𝑥𝑦 − 4𝑥 = 7, so we can calculate first of all
⃗ 𝑓 = 𝜕𝑓 𝑖̂ + 𝜕𝑓 𝑗̂ + 𝜕𝑓 𝑘̂ = (2𝑧 2 − 3𝑦)𝑖̂ + (−3𝑥)𝑗̂ + (4𝑥𝑧)𝑘̂. So, this is
gradient of f. So, ∇ 𝜕𝑥 𝜕𝑦 𝜕𝑧

our gradient of f.

And now, we have to calculate the tangent plane and the normal at the point (1, −1,2). So,
calculating normal is not difficult because we know that gradient of f is normal to the
surface gradient is normal to the surface 𝑓(𝑥, 𝑦, 𝑧) = 𝑐. So, basically this gradient will act

450
as a normal for the given function 𝑓(𝑥, 𝑦, 𝑧). And ⃗∇𝑓𝑃(1,−1,2) = 7𝑖̂ − 3𝑗̂ + 8𝑘̂ . So, this is
the required normal for the function f at the point (1, −1,2) all right, but calculating
tangent plane is little bit how to say time taking.

(Refer Slide Time: 16:35)

So, we now, calculate the tangent plane. So, if 𝑅⃗ = 𝑋𝑖̂ + 𝑌𝑗̂ + 𝑍𝑘̂ be the position vector
of any point, let us say 𝑃(𝑋, 𝑌, 𝑍) on the tangent plane right.

So, since I am using 𝑅⃗ , is better to use capital X, capital Y, capital Z.. So, then the vector
(𝑅⃗ − 𝑟) is perpendicular. So, this is from coordinate geometry perpendicular to the vector
⃗ 𝑓 all right. So, we have basically a tangent. So, we have basically this vector 𝑅⃗ = 𝑋𝑖̂ +
∇
𝑌𝑗̂ + 𝑍𝑘̂ be the position vector of any point (𝑋, 𝑌, 𝑍) on the tangent plane and then the
vector (𝑅⃗ − 𝑟) ⊥ ∇
⃗ 𝑓 because anything you have on the tangent plane will be by default

perpendicular to gradient of f because gradient of f is normal to the function 𝑓(𝑥, 𝑦, 𝑧) =

𝑐 . So, if you have a tangent plane on 𝑓(𝑥, 𝑦, 𝑧) = 𝑐, then that tangent plane will always
be perpendicular to the gradient of f. So, even if you take this difference, it will always be
perpendicular all right.

So, now 𝑅⃗ = 𝑋𝑖̂ + 𝑌𝑗̂ + 𝑍𝑘̂ all right and minus what is this 𝑟? The 𝑟 is given as this point.
So, we are calculating the tangent plane at a certain point P and that point is this
one,(1, −1,2). So, I can write 𝑟 = 𝑖̂ − 𝑗̂ + 2𝑘̂ right and ok. Here, then the equation of the
tangent plane. So, this I, I forgot to write this.

451
So, the equation of the tangent plane is 𝑔𝑟𝑎𝑑𝑓. (𝑅⃗ − 𝑟) = 0. So, this is the required

equation of the tangent plane. So, now, I am writing (𝑅⃗ − 𝑟). 𝑔𝑟𝑎𝑑𝑓 = 0 ⇒
[(𝑋𝑖̂ + 𝑌𝑗̂ + 𝑍𝑘̂ ) − (𝑖̂ − 𝑗̂ + 2𝑘̂)]. (7𝑖̂ − 3𝑗̂ + 8𝑘̂) = 0 is this one because dot product is
commutative.

So, we can actually exchange them. And now, if I calculate this, so this will be 7(𝑋 − 1) −
3(𝑌 + 1) + 8(𝑍 − 2) = 0. So, the required equation therefore, so this is the equation of
the tangent plane and we gave the normal vector. Now, we want to write the equation of
that normal vector.

𝑋−1
So, the required equation of the normal to the surface at the point (1, −1,2) is 𝜕𝑓 =
𝜕𝑥

𝑌−1 𝑍−1 𝜕𝑓 𝜕𝑓 𝜕𝑓 𝑋−1 𝑌−1 𝑍−1

𝜕𝑓 = 𝜕𝑓 and now = 7, 𝜕𝑦 = −3, 𝜕𝑧 = 8 And this is so = = , basically
𝜕𝑥 7 −3 8
𝜕𝑦 𝜕𝑧

these two so; this is equation of the tangent plane, this is the equation of the normal and
these two what we had to calculate.

So, calculating this tangent plane our normal vector and those are not complicated, it is
just that we have to use some tricks from vector calculation and also from coordinate
geometry too, so that we can be able to write the equation of the tangent plane and the
question of the normal. So, this was an another example where we were supposed to
calculate the tangent plane and the normal.

(Refer Slide Time: 23:05)

452
You can be asked I can write an another example, but I will of course, leave it as a task for
the student. So, example 5, so find the equations of the normal and the tangent plane for
the surface 𝑓(𝑥, 𝑦, 𝑧) = 𝑥𝑦𝑧 = 4 at the point (1,2,2).

So, here in this example also, we have to calculate the equation of the normal and the
tangent plane for the surface 𝑓(𝑥, 𝑦, 𝑧) = 𝑥𝑦𝑧 = 4 at the point (1,2,2). So, here in this
case also, we will calculate the gradient and then we calculate the equation of the tangent
plane and from gradient, we have to calculate and in such a way, so in the gradient itself
we have to calculate the equation of the tangent plane in such a way that, we can be able
to write this thing here (𝑅⃗ − 𝑟). 𝑔𝑟𝑎𝑑𝑓 so that we can be able to obtain the equation of the
tangent plane in this fashion.

And then, the equation of the normal to the surface at this point (1, −1,2) can be given by
𝑋−1 𝑌−1 𝑍−1 𝜕𝑓 𝜕𝑓 𝜕𝑓
this 𝜕𝑓 = 𝜕𝑓 = 𝜕𝑓 and just equate the coefficients for 𝜕𝑥 , 𝜕𝑦 , 𝜕𝑧 and then divide them
𝜕𝑥 𝜕𝑦 𝜕𝑧

by 7, −3,8 and that will give you the required equation of the normal and the tangent plane
is already given.

So, this is how we calculate the tangent plane and normal of a given scalar function f at a
certain point P. And in the next class, I will try to solve one more example on directional
derivative before we move on next topic. So, today, I will stop here and I thank you for
your attention and I look forward to your next class.

Thank you.

453
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 44
Application to Mechanics

Hello students, so, in the last class we were solving some examples based on level surfaces
and directional derivative. So, in the level surfaces we sort of tried to cover a few examples
where we can calculate the normal tangent plane and things like that. So, I did try to show
you 1 or 2 examples and the rest of the problems can be solved in the similar fashion.

So, you may try to look into some books which I have listed in the references and I try to
solve some examples. We will also try to include some problems in your assignment sheet.
So, that you can be able to practice them and we also solved few examples on directional
derivatives.

So, today I will solve one more example on directional derivative just to make the concept
clear and then we will move on to our next topic which is basically the Application of
Vector Calculus in applied mathematics or in Mechanics. So, that would also be an
interesting thing to study in this context all right.

(Refer Slide Time: 01:28)

454
So, let us start with our very last example on the directional derivative chapter. So, find
the directional derivative for the function 𝜑(𝑥, 𝑦, 𝑧) = 𝑥𝑦𝑧 = 4 at the point (2,2,2) in the
direction of (𝑖̂ + 𝑗̂ + 𝑘̂)all right. So, the solution. So, first of all we know that when we are
given a scalar function 𝜑(𝑥, 𝑦, 𝑧)𝑜𝑟 𝑓(𝑥, 𝑦, 𝑧). So, the very first thing that we do is to
⃗ 𝜑 or ∇
calculate the ∇ ⃗ 𝑓.

𝜕𝜑 𝜕𝜑 𝜕𝜑
So, the given function is 𝜑(𝑥, 𝑦, 𝑧), then ⃗∇𝜑 = 𝑖̂ + 𝜕𝑦 𝑗̂ + 𝑘̂ = 𝑦𝑧𝑖̂ + 𝑥𝑧𝑗̂ + 𝑥𝑦𝑘̂ all
𝜕𝑥 𝜕𝑧

right.

(Refer Slide Time: 03:29)

So, then the ⃗∇𝜑𝑃(2,2,2) = 4𝑖̂ + 4𝑗̂ + 4𝑘̂ = 4(𝑖̂ + 𝑗̂ + 𝑘̂ ) all right. Now, the given direction
vector or the given vector or I also prefer to call it as direction vector, because we have to
calculate the directional derivative along the direction of this vector ok. So, the given
vector 𝑎 is equals to what do we have? We have 𝑖̂ + 𝑗̂ + 𝑘̂ .

𝑎⃗ ̂
𝑖̂+𝑗̂ +𝑘 ̂
𝑖̂+𝑗̂ +𝑘
So, then our 𝑎̂ = |𝑎⃗| = √12 = all right.
+12 +12 √3

So, the required direction derivative; or we can write it as 𝐷𝐷 = 𝐷𝜑𝑃 = ⃗∇𝜑𝑃 . 𝑎̂ =

1 4 4
4(𝑖̂ + 𝑗̂ + 𝑘̂). (𝑖̂ + 𝑗̂ + 𝑘̂) = (12 + 12 + 12 ) = . 3 = 4√3
√3 √3 √3

. So, that is the required directional derivative of the given scalar function 𝜑 in this case in
the direction of the vector (𝑖̂ + 𝑗̂ + 𝑘̂). And we just followed the traditional how to say

455
method to calculate the directional derivative and this is how we obtain the directional
derivative of the function 𝜑 all right.

So, I will stop with the examples on directional derivative, because I have tried to cover
as many examples as possible. And now we move on to our next topic, which are basically
an application or some applications not an application, but some applications of vector
calculus in mechanics in a way and in elementary differential geometry and in mechanics.

So, we will now learn the concepts of a tangent normal binormal, there is a very nice
formula called as Serret Frenet formula that shows that how you connect in vector calculus
the normal, the tangent and the binormal of a curve at a certain point P. So, we will now
move on to those topics and today we will start with elementary how to say difference as
geometric concepts and then we move to those tangent normal and binormal concepts all
right.

(Refer Slide Time: 07:13)

So, basically so, we will start with today with this Serret; tangent, normal, binormal, Serret-
Frenet formula all right. So, this is what we start with today. So, before we start with
tangent, normal and binormal we give some basic definitions and how to say idea of what
do we mean by a parametric representation of a curve, how do we define a curve in a space
and then we slowly move on to these topics all right. So, let me go back to my notepad ok.

456
(Refer Slide Time: 07:50)

So, now, what do we mean by a curve in space? So, a curve in space; a formal definition
goes like this; basically, a curve is an aggregate; of points whose co-ordinates are functions
of a single variable.

So, thus the equation; 𝑥 = 𝑥(𝑡), 𝑦 = 𝑦(𝑡) & 𝑧 = 𝑧(𝑡); represents a curve in space and the
variable t is called a parameter and for each value of t with a certain range. So, t has a
certain range let us say, 𝑎 ≤ 𝑡 ≤ 𝑏, where 𝑎, 𝑏 ∈ ℝ corresponds for each value of t there
corresponds to a definite point 𝑃(𝑥, 𝑦, 𝑧) of the curve right.

So, for example, if I want to write how to say, let us say this is our curve in 2 dimensional
space; here they are talking about 3 dimensional space I am just for the sake of explanation
I am just taking this example. So, let us assume that our curve C is basically this circle.
So, how do we write this curve in space using this formula? So, I can write 𝑥 = 𝑥(𝑡). So,
𝑥(𝑡) would be let us say cos 𝑡 and 𝑦 = 𝑦(𝑡) = sin 𝑡 and I can choose 0 ≤ 𝑡 ≤ 2𝜋 all right.

So, if I choose 𝑡 = 0, then we have 𝑥 = cos 0 & 𝑦 = sin 0. So, basically (1,0); obviously,
𝜋
this (1,0) point lies on this circle, I can choose 𝑡 = 2 and then in that case 𝑥 =
𝜋 𝜋
cos 2 = 0 & 𝑦 = sin 2 = 1, then the point (0,1) lies on the circle and so on. So that means,

for every value of t we get a unique point P. So, well not unique, but there exists for every
value of t there corresponds a point P on the curve because for 𝑡 = 2𝜋 we get the same
point actually. So, for 𝑡 = 2𝜋 and 𝑡 = 0 we are getting the same point.

457
So, for every value of t, there corresponds a definite point. So, we must have a point on
that curve and that point on the curve is obtained for a certain value of the parameter t. So,
for every t; so, t has a range and for every t will obtain a point on that given curve. And
that curve is basically called as a curve in space. We can also have a sphere 𝑥 2 + 𝑦 2 +
𝑧 2 = 1 and then this equation would change. It would be cost, sint and then cost.

So, cost cost; cost sint and then again, so, basically we have to use 2 different variables.
So, then in that case it will be cost and then some other variable and. So, for a sphere it
will be costsin t and then we will have costcost and then.

So, you basically you got the idea how; how you formulate. So, 𝑥(𝑡) = cos 𝑡 sin 𝑡 and then
𝑦(𝑡) = cos 𝑡 cos 𝑡 and then 𝑧(𝑡) = sin 𝑡. So, then in that case 𝑐𝑜𝑠 2 𝑡 is so, that will be one.
So, again for a sphere you can give the formula in this fashion. So, that is one way to define
a curve in space, if the given curve is actually a sphere and the parametric. This is also
called as the parametric representation is given by if it is a sphere, then it will be given in
this fashion, if it is a circle then is given in this fashion all right.

Now, and if we want to have the circle as a curve in space, let us say in 3 d then we write
it as 𝑥 = cos 𝑡, 𝑦 = sin 𝑡 & 𝑧 = 0. Now we have a 3 dimensional representation. So, this
is again a curve in space where 𝑧 = 0. So, these are all tricks basically. So, we have a 2
dimensional circle and if you want to represent it in a 3 dimensional sense then basically
the z component is 0.

So, you just write 𝑧 = 0 and 𝑥 = cos 𝑡, 𝑦 = sin 𝑡 and that is your curve in space whose z
component is 0. So, that is how we write this curve in space all right. Now, that we have a
curve in space, we can actually be able to write, So, let me put it in a nice sentence.

458
(Refer Slide Time: 14:58)

So, in the language of vectors a curve can be represented; by an equation of type 𝑟 = 𝑓 (𝑡).
So, basically the equation of the curve in space can be written as 𝑟 = 𝑓(𝑡) . So, let us take
our circle case. So, for example, for that circle case I can write 𝑟 = 𝑓 (𝑡) and f has 3
components. So, I can write 𝑓 (𝑡) = 𝑓1 (𝑡)𝑖̂ + 𝑓2 (𝑡)𝑗̂ + 𝑓3 (𝑡)𝑘̂ = cos 𝑡 𝑖̂ + sin 𝑡 𝑗̂ + 0𝑘̂ =
cos 𝑡 𝑖̂ + sin 𝑡 𝑗̂. This is actually a function of t and that is why we are writing 𝑓(𝑡). So, 𝑓
is a vector function of a scalar variability.

So, you see our circle can also be written as 𝑟 = 𝑓(𝑡). So, in the language of vectors
⃗⃗⃗⃗⃗ .
actually, a curve can be represented by an equation 𝑟 = 𝑓 (𝑡) all right. Where 𝑟 = 𝑂𝑃
So, let me draw a figure. So, if I draw a figure I can be able to write it as. So, this is my x
axis, this is my y axis, this is origin this is z and suppose this is our curve f all right.

So, this is my curve f, and suppose this is the point 𝑃(𝑥, 𝑦, 𝑧) this is my point 𝑄(𝑥 +
𝛿𝑥. 𝑦 + 𝛿𝑦, 𝑧 + 𝛿𝑧). So, I am going to make them worse and this one is another vector.
⃗⃗⃗⃗⃗⃗ , 𝑂𝑃
So, there is our vector 𝑂𝑄 ⃗⃗⃗⃗⃗ and then this is our vector P to Q all right. Now this is 𝑟
and this is 𝑟 + 𝛿𝑟 all right.

Now, choosing 3 fixed directions now, I am just trying to put it in a nice word. So, choosing
3 fixed directions 𝑖̂, 𝑗̂ and 𝑘̂ mutually perpendicular to one another that means, 𝑖̂ is
perpendicular to 𝑗̂, 𝑗̂ is perpendicular to 𝑘̂ and then 𝑘̂ is perpendicular to 𝑖̂. So, they are

459
mutually perpendicular to one another, we may be able to write, we may express the
equation. Let us say, this is our equation star analytically as.

⃗⃗⃗⃗⃗ = 𝑥𝑖̂ + 𝑦𝑗̂ + 𝑧𝑘̂ = 𝑥(𝑡)𝑖̂ + 𝑦(𝑡)𝑗̂ + 𝑧(𝑡)𝑘̂ right. And this is which is
So, 𝑟 = 𝑂𝑃
equivalent which is equivalent to the 3 scalar equations given by 𝑥 = 𝑥(𝑡), 𝑦 =
𝑦(𝑡) 𝑎𝑛𝑑 𝑧 = 𝑧(𝑡) the equation from which we started originally. So, this is the scalar
equation for a curve in space.

Now, for every point on that curve in space we are associating a vector, let us say P is any
arbitrary point on that curve then from origin we are associating a vector OP and that is
given as 𝑟; 𝑟 is the position vector of a point P on that curve. And basically, 𝑟 can be with
the help of 𝑖̂, 𝑗̂ & 𝑘̂, r can be written as 𝑟 = 𝑥𝑖̂ + 𝑦𝑗̂ + 𝑧𝑘̂ , but x, y and z are all functions
of t. So, we write 𝑟 = 𝑥(𝑡)𝑖̂ + 𝑦(𝑡)𝑗̂ + 𝑧(𝑡)𝑘̂ and this is in a way how to say a way to
write a curve in terms of vector and that curve is actually a curve in space.

So, just using this 𝑟 = 𝑓 (𝑡), where 𝑓(𝑡) has 3 components 𝑥(𝑡), 𝑦(𝑡) & 𝑧(𝑡) , we are how
to say transforming that scalar representation we are sort of finding an alternative way to
write that scalar representation in a vector form. So, basically instead of writing 𝑥 =
𝑥(𝑡), 𝑦 = 𝑦(𝑡) 𝑎𝑛𝑑 𝑧 = 𝑧(𝑡), we are just writing 𝑟(𝑡) = 𝑥(𝑡)𝑖̂ + 𝑦(𝑡)𝑗̂ + 𝑧(𝑡)𝑘̂. And that
vector representation 𝑟(𝑡) equals to that expression is same as the scalar equation.

So, that scalar equation and that vector equation both are the same thing it is just that its
an alternative way to write the same curve in space. And based on that now we introduce
several concepts like if we had a scalar equation 𝑥 = 𝑥(𝑡), 𝑦 = 𝑦(𝑡) 𝑎𝑛𝑑 𝑧 = 𝑧(𝑡) since it
is a curve in space we can about tangent, we can talk about normal.

Similarly, for this vector representation the way we have did the vector representation that
is 𝑥(𝑡)𝑖̂ + 𝑦(𝑡)𝑗̂ + 𝑧(𝑡)𝑘̂ . We can still be able to talk about tangent normal and some other
things as well; we will see what are those things. So, its not completely different its the
same thing, is just that the way we are expressing it using the vectors is slightly different,
but it is convenient its very convenient to write the equation of a curve in space in a vector
form than writing it in a scalar form. And we will see what are its benefits over the time.
And now, as I was saying that we can actually be able to define the tangent and normal.
So, we will start with the definition of this tangent how do we define the tangent for the
curve in space given by this vector form all right.

460
(Refer Slide Time: 22:45)

So, let us start, tangent to a curve at a point P. So, we refer to the same figure and when Q
tends to P, then this will actually be how to say along the direction of a tangent to this
curve all right. So, the formal definition would be the tangent line PT at a point P; of a
curve is the limiting position; of the secant PQ joining P to a neighbouring point Q, when
Q approaches P along the curve.

So, this is we know already from our previous topics. So, when it is basically the tangent
is nothing but a limiting position. So; that means, when Q tends to P this will actually be
along the direction of tangent or this is actually a tangent in a way not along the, but it is
exactly the tangent. So, when Q approaches P, then it will not how to say go through this
curve, it will actually be a tangent. So, it will actually be touching the curve and that is that
will happen when we are making Q going to P. So, it is basically a limiting approach all
right or a limiting case.

So; that means, how do we define the tangent? Basically, in terms of vector. So, we first
saw that; so, how do we define the tangent? So, to define the tangent in terms of vector we
take two points P and Q which we have already done. So, the points P and Q are basically
(𝑥, 𝑦, 𝑧) and (𝑥 + 𝛿𝑥, 𝑦 + 𝛿𝑦, 𝑧 + 𝛿𝑧). So, this one is for the vector 𝑟 and this one is 𝑟 +
𝛿𝑟 all right.

461
So, now then our vector 𝑟 is ⃗⃗⃗⃗⃗
𝑂𝑃 all right and 𝑟 = ⃗⃗⃗⃗⃗
𝑂𝑃 = 𝑓(𝑡). So, that is how we are
giving the equation and for the point Q, 𝑟 + 𝛿𝑟 is the position of the point P at a time 𝑡 +
𝛿𝑡.

So, at time t it is simply (𝑥, 𝑦, 𝑧) and at time 𝑡 + 𝛿𝑡 it is now (𝑥 + 𝛿𝑥, 𝑦 + 𝛿𝑦, 𝑧 + 𝛿𝑧). so,
that is the point Q. And in terms of the function 𝑓 , I can be able to write 𝑟 + 𝛿𝑟 = 𝑓(𝑡 +
𝛿𝑡) . So, at the time 𝑡 + 𝛿𝑡 the point has moved along the curve to the point Q and it is
given as this way all right.

So, our line ⃗⃗⃗⃗⃗ ⃗⃗⃗⃗⃗⃗ − ⃗⃗⃗⃗⃗

𝑃𝑄 = 𝑂𝑄 𝑂𝑃 = 𝑓 (𝑡 + 𝛿𝑡)−𝑓 (𝑡) all right and we divide by 𝛿𝑡. So, then this
⃗⃗⃗⃗⃗
𝑃𝑄 𝑓 (𝑡+𝛿𝑡)−𝑓(𝑡) ⃗⃗⃗⃗⃗
𝑃𝑄 𝑓 (𝑡+𝛿𝑡)−𝑓 (𝑡)
will imply = right. And now we take lim = lim all right.
𝛿𝑡 𝛿𝑡 𝛿𝑡→0 𝛿𝑡 𝛿𝑡→0 𝛿𝑡

So, we will have this thing here. Now, this quantity P to Q is nothing but that 𝛿𝑟 right, so,
P to Q is nothing but 𝛿𝑟.

(Refer Slide Time: 28:15)

𝛿𝑟 𝑓 (𝑡+𝛿𝑡)−𝑓 (𝑡) 𝑑𝑟
So, I can write lim = lim . So, this is nothing but our = 𝑓 ′ (𝑡). So, this
𝛿𝑡→0 𝛿𝑡 𝛿𝑡→0 𝛿𝑡 𝑑𝑡

we already know from the differentiation. So, this is the derivative of the vector function
f.

462
 𝑑𝑟
So that means or 𝑓 ′ (𝑡) is parallel to the tangent PT right of the curve 𝑟 = 𝑓(𝑡) at the
𝑑𝑡
𝑑𝑟
point P, where t is the parameter. And therefore, 𝑅⃗ = 𝑟 + 𝜆 . So, 𝑅⃗ can be written as,
𝑑𝑡
𝑑𝑟
𝑅⃗ − 𝑟 = 𝜆 𝑑𝑡

(Refer Slide Time: 30:26)

So, this equation represents the equation of the tangent line PT, where 𝜆 is any arbitrary
constant; 𝑟 is the position vector of P and 𝑅⃗ is the position vector of any point on the
tangent line.

𝑑𝑟
So; that means, in this equation we have shown that is basically parallel to the tangent
𝑑𝑡

PT of the curve 𝑟 = 𝑓 (𝑡). So that means, the position vector of any point on the tangent
𝑑𝑟
plane or on the tangent line is denoted by 𝑅⃗ and this is basically 𝑟 + 𝜆 , So; that means,
𝑑𝑡
𝑑𝑟
we can be able to write the equation of the tangent line as 𝑅⃗ − 𝑟 = 𝜆 .
𝑑𝑡

𝑑𝑟
So, this is basically saying that the tangent is parallel to and this 𝑅⃗ and 𝑟 are basically
𝑑𝑡

nothing but this is the position vector of the point P. And that is the position vector of any
𝑑𝑟
point on the tangent plane and is basically the derivative of the vector function 𝑟 with
𝑑𝑡

respect to t and 𝜆 is a constant.

463
So, this is how we give the equation of a tangent line at a point P for the curve 𝑟 = 𝑓 (𝑡).
So, today we will stop at here, in the next class we will work out few examples that how
we calculate the tangent line although we saw in the previous chapter on level surfaces
how we calculate the tangent line. So, perhaps we will do one more example. And then we
will introduce the concepts of normal and binormal of a vector function and then we will
try to derive this Serret Frenet formula.

So thank you for attention and I look forward to your next class.

464
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 45
Equation of Tangent, Unit Tangent Vector

Hello, students. So, up until last class we looked into the topics of tangents and we also
worked out examples based on a gradient divergence curve, we calculated directional
derivative and things like that. And in the last class, we started looking into more like
differential geometry aspect of vector calculus. So, like tangent, normal, by normal and
then we will look into the Serret-Frenet formula.

(Refer Slide Time: 01:03)

𝑑𝑟
So, in the last class we derived a formula for the tangent as (𝑅⃗ − 𝑟) × 𝑑𝑡 = ⃗0 where our 𝑟

is the position vector; of point P.

So, if you have a curve like let us say in three-dimensional geometry this is x-axis, this is
our y-axis, this is z-axis, that is origin. So, if you have a curve like this. So, let us say this
is my point P and this is my point Q and the position vector of the point P is 𝑟. So, 𝑟 is the
position vector of a point P and the 𝑅⃗ is the position vector of any point on the tangent line
and the equation of the curve is of course, given by 𝑟 = 𝑓 (𝑡).

465
So, if you want to calculate the equation of the tangent line then in that case we calculate
𝑑𝑟 𝑑𝑟
for the given curve we calculate and then taking (𝑅⃗ − 𝑟) × 𝑑𝑡 = ⃗0 will give us the
𝑑𝑡

required equation of the tangent to that curve 𝑟 = 𝑓(𝑡).

Now, we can also express this equation in terms of a Cartesian coordinate system. So, we
are familiar with the equation of the tangent in three dimensional geometry to a surface
and so, we can actually express this equation in terms of Cartesian coordinate system as
well. So, how do we do that?. So, we can write 𝑟 = 𝑥𝑖̂ + 𝑦𝑗̂ + 𝑧𝑘̂ that came.

𝑑𝑟 𝑑𝑥 𝑑𝑦 𝑑𝑧
So, let us write that and then from here I can calculate 𝑑𝑡 = 𝑖̂ + 𝑑𝑡 𝑗̂ + 𝑑𝑡 𝑘̂ or I can also
𝑑𝑡
𝑑𝑟 𝑑𝑥 𝑑𝑦 𝑑𝑧
write 𝑑𝑡 = 𝑖̂ + 𝑗̂ + ̂ = (𝑑𝑥 , 𝑑𝑦 , 𝑑𝑧). So, of course, if you can either it is up to you
𝑘
𝑑𝑡 𝑑𝑡 𝑑𝑡 𝑑𝑡 𝑑𝑡 𝑑𝑡

whether you write in terms of 𝑖̂, 𝑗̂, 𝑘̂ or you write in terms of as a triplet, alright.

So, the equation of the tangent line; at the point P is given by. So, if you remember in the
𝑑𝑟 𝑑𝑥 𝑑𝑦 𝑑𝑧
last class we derived something like 𝑅⃗ = 𝑟 + 𝜆 𝑑𝑡 = (𝑥, 𝑦, 𝑧) + 𝜆 ( 𝑑𝑡 , 𝑑𝑡 , 𝑑𝑡 ). So, from
𝑑𝑥 𝑑𝑦 𝑑𝑧
here I can write (𝑋, 𝑌, 𝑍) = (𝑥, 𝑦, 𝑧) + 𝜆 ( 𝑑𝑡 , 𝑑𝑡 , 𝑑𝑡 ).

Now, from here I will take this triplet on the left hand side and then we will equate the
coefficients of 𝑖̂, 𝑗̂ and 𝑘̂ from both sides.

(Refer Slide Time: 05:45)

466
 𝑋−𝑥 𝑌−𝑦 𝑍−𝑧
So, ultimately we will obtain 𝑑𝑥 = 𝑑𝑦 = 𝑑𝑧 = 𝜆 and here (𝑋, 𝑌, 𝑍) is any point; on
𝑑𝑡 𝑑𝑡 𝑑𝑡

the tangent line; at 𝑝(𝑥, 𝑦, 𝑧).

So, this is the required equation of the tangent line in terms of Cartesian coordinate system
and you may have seen this type of equation in three-dimensional geometry for the
equation of a line in 3D. So, this is just an another form in terms of a Cartesian coordinate
𝑑𝑟
system and for the vector form we know its (𝑅⃗ − 𝑟) × 𝑑𝑡 = ⃗0.

So, now let us work out an example where we see how we can calculate the tangent line
at a given point. So, example find the equation of the tangent line to the space curve 𝑥 =
2
𝑡, 𝑦 = 𝑡 2 , 𝑧 = 3 𝑡 3 or we can write 𝑟 is equal to or do not write or we will do something

about it at the point 𝑡 = 1, right. So, we have to find the equation of the tangent line to the
space curve given by these equations at the point 𝑡 = 1.

So, first of all we can write, 𝑟 = 𝑓 (𝑡) = 𝑥𝑖̂ + 𝑦𝑗̂ + 𝑧𝑘̂ = (𝑥, 𝑦, 𝑧). 𝑟 = 𝑓 (𝑡) = 𝑥𝑖̂ + 𝑦𝑗̂ +
2
𝑧𝑘̂ = 𝑡𝑖̂ + 𝑡 2 𝑗̂ + 3 𝑡 3 𝑘̂ right. And now that we have this equation of the curve in terms of

vector as a vector form I can calculate the equation of the tangent at the point P or at a
point when where 𝑡 = 1, first in terms of the vector and the second we will calculate as a
Cartesian form.

𝑑𝑟
So, for the vector (𝑅⃗ − 𝑟) × 𝑑𝑡 = ⃗0. So, that is the vector form equation for the tangent
𝑑𝑟 𝑑𝑟
line. So, first of all we need to calculate . So, what is our 𝑑𝑡 ?
𝑑𝑡

467
(Refer Slide Time: 09:13)

𝑑𝑟 𝑑𝑟 𝑑𝑟
So, if we like to calculate 𝑑𝑡 then it will be 𝑑𝑡 = 𝑖̂ + 2𝑡𝑗̂ + 3𝑡 2 𝑘̂ . So, from here [ 𝑑𝑡 ]𝑡=1 =
2
𝑖̂ + 2𝑗̂ + 3𝑘̂, right. So, that is the tangent and at the point 𝑡 = 1, so, 𝑟(𝑡)𝑡=1 = 𝑖̂ + 𝑗̂ + 3 𝑘̂ .
𝑑𝑟 2
Therefore, the required vector equation will be (𝑅⃗ − 𝑟) × 𝑑𝑡 = ⃗0 ⇒ [𝑅⃗ − (1,1 3)] ×

(1,2,3) = ⃗0.

So, if you want we can write this 𝑅⃗ = 𝑋𝑖̂ + 𝑌𝑗̂ + 𝑍𝑘̂. So, this is basically
2
[(𝑋 − 1, 𝑌 − 1, 𝑍 − 3] × (1,2,3) = ⃗0. So, this is the required equation of the tangent line

at the point P to the space curve given by that equation 𝑟(𝑡) = to this equation and that is
𝑑𝑟
basically cross product with the tangent equals to ⃗0. So, this is how we calculate the
𝑑𝑡

equation of the tangent line in terms of the in the vector form.

468
(Refer Slide Time: 11:28)

Now, we can also calculate it in Cartesian form. So, in the Cartesian form we have to
𝑑𝑟 𝑑𝑥 𝑑𝑦 𝑑𝑧
substitute 𝑡 = 1 in , then from there [ 𝑑𝑡 ]𝑡=1 = 1, [ 𝑑𝑡 ]𝑡=1 = 2, [𝑑𝑡 ]𝑡=1 = 3 right and
𝑑𝑡
𝑋−𝑥 𝑌−𝑦 𝑍−𝑧
therefore, the required equation would be = =
1 2 3

and where (x, y, z) is any point is any point on the tangent line on the tangent line at
2
𝑃(1,1, 3).

2
So, at 𝑡 = 1 the point P is given by (1,1, 3).and that (𝑋, 𝑌, 𝑍) is any point on the tangent
2
line at the point this. So, basically which was(𝑋, 𝑌, 𝑍) as (1,1, 3) and that is the required

and then we substitute for (𝑋, 𝑌, 𝑍) here. So, this is the required equation of the tangent
plane at the point P and of course, this is in the Cartesian form. So, either you can express
it in terms of the vector form or in terms of the Cartesian form that is up to you and yeah
similarly we can find equation of the tangent line for any arbitrary curve. We can have a
curve something like x equals to. So, this example ends here , but we can have something
like 𝑥 = cos 𝑡 , 𝑦 = sin 𝑡 , 𝑧 = 𝑡 and let us say we want to find the equation of the tangent
𝜋
line at the point 𝑡 = 2 .

So, we will follow the similar steps and then use the same formula for the tangent line and
you can be able to calculate the tangent line in the vector form or in the Cartesian form.
You can also have something like 𝑥 = 𝑡 2 , 𝑦 = 𝑡 4 , 𝑧 = 1 and you are asked to calculate the

469
equation of the tangent line at 𝑡 = 1. So, just follow the similar formula what we did before
and that will give you the required equation of the tangent line either in vector or in the
Cartesian form.

So, the examples would not be that much complicated when you are asked to calculate the
tangent line for a given curve it might involve doing some derivative which may have
some higher powers or something like that, but pretty sure you can be able to do it. So,
now we move to our move to our how to say a little bit more into differential geometry
aspect, but it is not too much differential geometry . It is somehow involving our tangents
and normal and the Serret-Frenet formula. So, keeping those in mind we have to study
these concepts. So, let me introduce that via a small statement.

(Refer Slide Time: 15:26)

So, basically now we are going to study about tangent line at a point P of the curve 𝑟 =
𝑓(𝑠) or basically we are interested in unit tangent vector. So, s is the arc length basically.
So, s that t is the parameter, but when we are using s its pretty much understood that s
denotes the arc length. So, suppose you have a given equation of a curve in terms of arc
length and from there how we define the unit tangent vector or first of all the tangent
afterwards the unit tangent vector we will see that here alright.

So, I am going to state a very important theorem, but we will avoid the proof because the
proof is slightly lengthy and it is out of the scope of this course, but interested readers they
can look into the books that have suggested in the references for the proof, alright. So, the

470
statement goes like this. So, if vector 𝑟 = 𝑓(𝑠) be the vector equation of a smooth curve
in terms of s which is basically the arc length.

𝑑𝑟 𝑑𝑟
So, s as a parameter then 𝑑𝑠 = 𝑡̂,the unit tangent vector. So, remember 𝑑𝑡 is I mean in our
𝑑𝑟
previous class we learnt the 𝑑𝑡 is actually the tangent vector in a way. So, it actually gives

us how to say that I have the tangent vector on that curve at a certain point.

Now, if we want to calculate the unit tangent vector technically if you want to calculate
⃗
𝑑𝑟
𝑑𝑟
the unit tangent vector then we have to divide with its magnitude right i.e. 𝑡̂ = 𝑑𝑡
⃗
𝑑𝑟
. So,
𝑑𝑡 | |
𝑑𝑡

𝑑𝑟
denotes the tangent to the curve 𝑟 = 𝑓(𝑡). So, if we want to calculate the unit tangent
𝑑𝑡
𝑑𝑟
vector we must divide by its magnitude, but this theorem says that if we have an
𝑑𝑡
𝑑𝑟
equation 𝑟 = 𝑓 (𝑠) of a smooth curve then the unit tangent vector will be given by 𝑑𝑠 . So,

if we differentiate 𝑟 with respect to the arc length as a parameter then in that case the
differentiation the first order differentiation will give us the unit tangent vector.

𝑑𝑟
And therefore, from here what we can write is so we can write that . So, from here this
𝑑𝑠

is basically the unit tangent vector and from here I can write unit tangent vector 𝑡̂ =
𝑑𝑟 𝑑𝑡 𝑑𝑟 𝑑𝑠 𝑑𝑟
× 𝑑𝑠. So, basically 𝑑𝑡 = 𝑡̂ 𝑑𝑡 , alright. So, this is another result. So, that means, our 𝑑𝑡 the
𝑑𝑡

original how to say a tangent vector which was not a unit vector can be given by
𝑑𝑠
differentiation of this arc length s with respect to t times the unit tangent vector 𝑡̂ . So,
𝑑𝑡

this is a very nice result to remember that the differentiation of the curve with respect to
the arc length is the unit tangent vector.

Now, from here this formula actually gives us the equation of the tangent vector where the
equation of the curve is given in terms of s. So, here we see that the equation of the curve
is given in terms of s and in this case and basically the tangent vector can be calculated at
𝑑𝑠
times the unit tangent vector 𝑡̂, alright. So, let us see how we can calculate this unit
𝑑𝑡

tangent vector 𝑡̂. So, I can start with an example.

471
(Refer Slide Time: 20:50)

So, first of all suppose the given equation of the curve is 𝑥 = 𝑡, 𝑦 = 𝑡 2 , 𝑧 = 𝑡 3 . So, that is
the given equation of the curve. We have to find the equation of a unit tangent vector. So,
𝑑𝑟
either we can calculate 𝑑𝑡 and then divide it with its magnitude or we can use that previous

formula. So, here I can write 𝑟 = 𝑡𝑖̂ + 𝑡 2 𝑗̂ + 𝑡 3 𝑘̂ . So, this I can always write because I can
write 𝑥𝑖̂ + 𝑦𝑗̂ + 𝑧𝑘̂.

𝑑𝑟 𝑑𝑟
Now, from here my = 𝑖̂ + 2𝑡𝑗̂ + 3𝑡 2 𝑘̂ . So, that is my 𝑑𝑡 , alright and we know that from
𝑑𝑡
𝑑𝑟 𝑑𝑠 𝑑𝑠
previous formula = 𝑡̂ 𝑑𝑡 , alright. So, 𝑡̂ 𝑑𝑡 , so, that I have written here. Therefore, from
𝑑𝑡
𝑑𝑟 𝑑𝑠 𝑑𝑠
here | 𝑑𝑡 | = |𝑡̂ 𝑑𝑡 | = |𝑡̂| |𝑑𝑡 | is it not? Now, |𝑡̂| = 1 because 𝑡̂ is a unit tangent vector. So,
𝑑𝑟 𝑑𝑠
it is magnitude is always 1, so, we have | 𝑑𝑡 | = |𝑑𝑡 |.

𝑑𝑠 𝑑𝑟
And therefore, |𝑑𝑡 | = | 𝑑𝑡 | = √1 + 4𝑡 2 + 9𝑡 4 , is it not? And now, I can write this here as

𝑑𝑟 𝑑𝑟⁄ 𝑖̂+2𝑡𝑗̂ +3𝑡 2 𝑘

̂ 𝑑𝑡
𝑑𝑡
𝑡̂. So, the tangent vector is 𝑡̂ = 𝑑𝑠 = 𝑑𝑠⁄ = √1+4𝑡 2 because this can be written as 𝑑𝑠.
𝑑𝑡 +9𝑡 4

I will bring it in the denominator and then it will become 𝑑𝑠⁄𝑑𝑡 and if it is in denominator

then 𝑑𝑠⁄𝑑𝑡 is this here.

472
 𝑑𝑟 𝑑𝑠
And therefore, | 𝑑𝑡 | = |𝑑𝑡 | is basically the mod or the positive value. So, it will always
𝑑𝑟 𝑑𝑠
remain positive and = 𝑖̂ + 2𝑡𝑗̂ + 3𝑡 2 𝑘̂ and |𝑑𝑡 |; that means, taking only the positive
𝑑𝑡

value √1 + 4𝑡 2 + 9𝑡 4 . So, this is our required how to say answer or the value of the unit
tangent vector to the given curve this here. Of course, as I said we could have divided this
weight it is not 6 its 9 actually yeah 9.

𝑑𝑟
So, we could have divided this 𝑑𝑡 with its magnitude and that would have given us the unit

tangent vector, but in case if you did not have t and if you had arc length as 𝑥 = 𝑠, 𝑦 = 𝑠 3 ,
or something; so, that means, the given equation is in terms of the arc length and then we
had to use a different formula to calculate the unit tangent vector. And if you have
something in terms of s then basically the derivative of that 𝑟 with respect to s would give
you the unit tangent vector.

So, the unit tangent vector the curve and the arc length is related with that relation that 𝑡̂ =
𝑑𝑟
and its also a very important relation to remember alright.
𝑑𝑠

(Refer Slide Time: 25:10)

And now, that we have introduced the concept of how to say in a unit tangent vector we
can move to our next topic which is normal at a point. So, normal plane at a point so,
normal or normal plane at a point; so, what is the definition? The definition is the plane

473
through a point P perpendicular to the tangent at P is called the normal plane at that point,
alright. So, usually that is what we mean by normal.

So, you have a curve and then you have a tangent at a certain point P. So, a line which is
perpendicular to the tangent at that particular point is called as the normal to the curve at
that point. So, usually a normal plane is like a generalization of the normal to a curve. So,
if you have a plane that is passing through a point P and if it is perpendicular to the tangent
plane at the point P, then that plane is called as a normal plane to the to the given surface
or in this case to the given curve at that particular point P alright.

So, let us see how we can express its equation so, we know that; so, we know that the
𝑑𝑟
equation of the tangent plane is given by (𝑅⃗ − 𝑟) × 𝑑𝑡 = ⃗0. So, this means that these two
𝑑𝑟
vectors are parallel, right. So, 𝑑𝑡 and (𝑅⃗ − 𝑟) are parallel to one another and therefore,
𝑑𝑟
their cross product is 0. So, if I write (𝑅⃗ − 𝑟). 𝑑𝑡 = 0, then (𝑅⃗ − 𝑟) is perpendicular to
𝑑𝑟
and that will actually help us to obtain the equation of the normal plane.
𝑑𝑡

So, if 𝑅⃗ denotes the position vector of any point on the normal at the point P whose
𝑑𝑟
position vector is 𝑟 then the vector (𝑅⃗ − 𝑟) and 𝑑𝑡 will be perpendicular to one another;
𝑑𝑟
that is (𝑅⃗ − 𝑟). 𝑑𝑡 = 0.

𝑑𝑟
So, 𝑑𝑡 is basically the tangent vector and if we choose 𝑅⃗ as any point on the normal which
𝑑𝑟
is normal at a point P then in that case (𝑅⃗ − 𝑟) will be perpendicular to 𝑑𝑡 . Obviously and
𝑑𝑟
therefore, that equation (𝑅⃗ − 𝑟). 𝑑𝑡 = 0 is basically the equation of the normal plane
𝑑𝑟
because and (𝑅⃗ − 𝑟) are perpendicular to one another and that is actually what normal
𝑑𝑡

plane mean that a plane which is perpendicular to tangent vector is actually a normal plane
at a point P to that curve. So, this equation is actually the equation of that normal plane at
a point P to the curve 𝑟 = 𝑓 (𝑡), alright. So, this is the equation of our normal plane.

So, in today’s lecture we saw how we derive the equation of a tangent plane at a point P
we worked out few examples we also gave the concept of a unit tangent vector and we
derived the normal at a point P to a given curve. In the next class we will start working

474
with some examples and I will also try to introduce the concept of unit normal and if time
permits the concept of binomial. So, I will stop here for today and I will continue with our
next topic in our next class.

Thank you.

475
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 46
Unit normal, Unit binormal, Equation of Normal Plane

Hello students. So, in the last class we learned about the concepts of a unit tangent vector
and the equation of a normal plane at a point P to a given curve 𝑟⃗ = 𝑓⃗(𝑡).

(Refer Slide Time: 00:33)

𝑑𝑟⃗
And our equation of the normal plane was something like (𝑅⃗⃗ − 𝑟⃗). 𝑑𝑡 = 0 . This is not, a

⃗⃗
0 because the dot product is always yields a scalar quantity.

Now, I can also write this equation in terms of the in terms of the Cartesian coordinate
system. So, how we are going to do that? We can write [(𝑋, 𝑌, 𝑍) −
𝑑𝑥 𝑑𝑦 𝑑𝑧 𝑑𝑥 𝑑𝑦 𝑑𝑧
(𝑥, 𝑦, 𝑧)]. ( , , ) = 0. So, I can take it as (𝑋 − 𝑥) 𝑑𝑡 + (𝑌 − 𝑦) 𝑑𝑡 + (𝑍 − 𝑧) 𝑑𝑡 = 0 .
𝑑𝑡 𝑑𝑡 𝑑𝑡

So, this is just some simple dot product formula. So, this is the required equation of a
normal plane at a given point P to the curve 𝑟⃗ = 𝑓⃗(𝑡), alright.

So, let us work out an example just to make things clear. So, given a space curve we take
2
the same example 𝑥 = 𝑡, 𝑦 = 𝑡 2 , 𝑧 = 3 𝑡 3 . So, here we have to find out the equation of the

476
normal plane. So, of course, we can write 𝑟⃗ = 𝑥𝑖̂ + 𝑦𝑗̂ + 𝑧𝑘̂ where all these variables are
2
function of t. So, I can write 𝑟⃗ = 𝑥𝑖̂ + 𝑦𝑗̂ + 𝑧𝑘̂ = 𝑡𝑖̂ + 𝑡 2 𝑗̂ + 3 𝑡 3 𝑘̂ .So, so far we are not

differentiating we are just writing the equation of the curve. So, let us write the equation
of the curve first.

2 𝑑𝑟⃗
So, this one is basically 𝑡 2 and 3 𝑡 3 . So, from here we can write = 𝑖̂ + 2𝑡𝑗̂ + 2𝑡 2 𝑘̂,
𝑑𝑡
𝑑𝑟⃗
alright. So, this is the required tangent vector .
𝑑𝑡

Now, from here we can calculate, how to say the equation of the normal plane in the vector
form. So, 𝑅⃗⃗ = 𝑋𝑖̂ + 𝑌𝑗̂ + 𝑍𝑘̂.

(Refer Slide Time: 03:47)

𝑑𝑟⃗
So, first of all we will write this thing here. So, dot product with , so dot product with
𝑑𝑡
𝑑𝑟⃗
, is (𝑅⃗⃗ − 𝑟⃗). (1,2𝑡, 2𝑡 2 ) = 0, alright. And from here we can choose let us say any point
𝑑𝑡

where we want to calculate the normal vector.

2
So, suppose for 𝑡 = 1, the point will be (1,1, 3), ok. So, let us calculate that normal plane
2
at the point 𝑡 = 1 . So, at the point 𝑡 = 1 the point P will be (1,1, 3), so I can write this as
2
[(𝑋, 𝑌, 𝑍) − (1,1, 3)] . (1,2,2) = 0. So, from here I can write (𝑋 − 1) + 2(𝑌 − 1) +
2
2 (𝑍 − 3) = 0. So, these are all capitals, alright.

477
So, this is basically the equation of the normal plane at a point 𝑡 = 1. So, that at 𝑡 = 1 the
2 𝑑𝑟⃗
point is (1,1, 3) and the derivative this 𝑑𝑡 is basically (1,2,2). So, just take the dot product

and that will give you the required equation of the normal plane. So, it is really not
complicated to calculate, we can leave it here or we can also express it in terms of the
Cartesian coordinate system, so that is up to us, alright. So, this is how we calculate the
normal plane at a certain point P to a curve.

Next like a unit tangent vector we will move to unit normal and binormal. So, what do we
mean by these things? So, the thing is we are slowly moving towards Serret Frenet formula,
alright.

(Refer Slide Time: 06:23)

So, first of all we know unit tangent vector, right. So, we know unit tangent vector which
𝑑𝑟⃗
is given by 𝑡̂. And we have shown that or we know that if we have a curve given in
𝑑𝑠
𝑑𝑟⃗
terms of the arc length then 𝑑𝑠 = 𝑡̂, and 𝑡̂ is a unit vector or unit tangent vector, alright.

𝑑𝑟⃗ 𝑑𝑟⃗
And we also know the equation of the tangent plane as (𝑅⃗⃗ − 𝑟⃗) × 𝑑𝑡 . So, 𝑑𝑡 can be replaced
𝑑𝑟⃗
with 𝑡̂ because we divide by its magnitude and then we multiply it on the right-hand
𝑑𝑡

side, so that will be 0, so ultimately we will have a 𝑡̂ here as well. So, this is not very
complicated to see, alright.

478
And next we define the unit normal. So, unit principal normal; so, we have normal plane
or normal now we are talking about unit normal 𝑛̂, alright. So, there is a small statement,
so the statement is suppose, the length of t is constant, so since how to say the length of
𝑑𝑡
the tangent in this unit tangent vector is constant then in that case 𝑑𝑠 the its derivative will
𝑑𝑡
be 0, and if it is not 0 then in that case 𝑑𝑠 will be perpendicular to t, alright.

𝑑𝑡
So, if this that the derivative of t with respect to s if it is not 0 then in that case it is
𝑑𝑠

perpendicular to t and therefore, it will be a normal at the point P. So, you have a tangent
𝑑𝑡 𝑑𝑡
vector and if we differentiate the tangent vector and if ≠ 0, then this 𝑑𝑠 will be actually
𝑑𝑠

perpendicular to the tangent vector t at the point P. So, what do we mean by this? Let me
write this in terms of statement.

𝑑𝑡
So, since the length of t is constant, its derivative 𝑑𝑠 if it is not 0 must be perpendicular to

t, alright. And therefore, it will be normal to the curve at the point P. So, if it is not 0 then
𝑑𝑡
in that case this 𝑑𝑠 will be perpendicular to the tangent this 𝑡̂ or unit tangent vector at the
𝑑𝑡
point P. And a directed line through P in the direction of 𝑑𝑠, is called the principal normal

to the curve at the point P.

𝑑𝑡
So; that means, if 𝑑𝑠 ≠ 0 then it is perpendicular to the curve at the point P and a directed
𝑑𝑡
line through P. So, if you have a directed line through P in the direction of , then that
𝑑𝑠

directed line will be called as a principal normal. So, and if 𝑛̂ is the unit vector so, we have
a very we have a small formula. So, if 𝑛̂ is the unit vector in the direction in the direction
𝑑𝑡
of the principal normal let us say 𝑛⃗⃗. So, then we may write = 𝜅𝑛̂. So, this is basically
𝑑𝑠

kappa.

479
(Refer Slide Time: 12:23)

𝑑𝑡 𝑑𝑡
So, you see 𝑛̂ is basically a unit vector in the direction of and this = 𝜅𝑛̂ and this
𝑑𝑠 𝑑𝑠
𝑑𝑡
means that they are parallel to one another; that means, 𝑛̂ is in the direction of 𝑑𝑠 and this

𝑛̂ is called as the unit principal normal. So, here where 𝜅 is a non-negative scalar and 𝑛̂ is
called a unit principle normal and kappa is called the curvature of the curve at the point P.
So, this is not actually K, this is kappa but it is sort of like a curlier notation. So, I leave it
up to you how do you want to denote this notation, but it is basically kappa. So, kappa is
𝑑𝑡
actually the curvature of the curve at the point P and 𝑑𝑠 = 𝜅𝑛̂ is the required formula for
𝑑𝑡 𝑑𝑡
and that is how it connects the unit principal normal with 𝑑𝑠 .
𝑑𝑠

𝑑𝑡
So, here we can see that the direction of the unit principal normal and 𝑑𝑠 are the same and
𝑑𝑡
if is perpendicular to the curve at the point P then in that case 𝑛̂ is actually also
𝑑𝑠

perpendicular to the curve at the point P, and it is a unit vector, so it is basically unit
principal normal alright. And this kappa is the curvature of the curve at the point P.

So, we have basically learnt about tangent vector, unit tangent vector and unit principal
normal, alright. Now, what we can do, we can write the unit binormal. So, we know tangent
vector, we know normal and now we learn about unit binormal. So, we introduce another
unit vector 𝑏̂ defined by 𝑏̂ = 𝑡̂ × 𝑛̂ where 𝑏̂, 𝑛̂ and 𝑡̂ form a right-handed system of

480
orthogonal unit vectors. So, that means, if you have this is x, this is y and this is z, and let
us say this is my curve then in that case and if this is the point P let us say.

So, I can write as this as I am not very good at drawing. So, this could be our 𝑡̂, this could
be our 𝑛̂, and let us say 𝑏̂ is in this direction so, 𝑏̂, 𝑡̂ and 𝑛̂ . So, that is actually our how to
say, our required this right handed orthogonal unit vector. So, this is how 𝑏̂, 𝑡̂ and 𝑛̂ are
connected, now to define it formally I write a directed line a directed line through P in the
direction of 𝑏̂ is called a binormal to the curve at the point P;

(Refer Slide Time: 17:19)

So, we may call 𝑏̂ as a unit binormal. So, 𝑏̂ is a unit vector and it is in the direction of. So,
any any directed any directed line through the point P in the direction of 𝑏̂ is called as the
binormal and then 𝑏̂ is called as a unit binormal or we call 𝑏̂ as a unit binormal.

So that means, when we have this coordinate system x, y and z, and if we have 𝑖̂ in this
direction, 𝑗̂ in this direction, and 𝑘̂ in this direction then we know that 𝑖̂ × 𝑗̂ = 𝑘̂, 𝑗̂ × 𝑘̂ =
𝑖̂ &𝑘̂ × 𝑖̂ = 𝑗̂. So, this we know for 𝑖̂, 𝑗̂ and 𝑘̂. Similarly, since 𝑏̂, 𝑡̂, 𝑛̂ they form how to
say a right handed system for unit vectors we can be able to write 𝑏̂ = 𝑡̂ × 𝑛̂, 𝑛̂ =
𝑡̂ × 𝑏̂, 𝑡̂ = 𝑛̂ × 𝑏̂ , alright. So, this is also true for these 3 vectors I can take 𝑏̂, 𝑛̂, 𝑡̂ or I can
take 𝑡̂, 𝑏̂ and 𝑛̂ then in that case it will yield a similar formula.

481
So, this is how normal unit normal and binormal normal or unit binormal, unit principle
normal and unit tangent vector are connected with one another. And this is a very important
formula to remember because I mean sometimes you might be given tangent and normal
and then you are asked to calculate binormal. So, we use the first formula. Similarly, you
could be given I do not know binormal and tangent and you will be asked to calculate the
normal, so we use this formula, alright.

So, now they are also called as moving tetrahedral. So, trihedral sorry not tetrahedral. So,
they are called as moving trihedral. So, they are 3 in number, so they are called as trihedral,
alright. And now there is something called torsion we will introduce the concept of for
torsion now. So, what is torsion? So, we have tangent normal and binormal, now we will
introduce the concept of torsion. So, an another topic.

So, why I am introducing all these things is because now I am slowly moving towards
Serret Frenet formula. So, we have these 3, now we will introduce the concept of torsion.
So, in the above relation we have defined that 𝑏̂ = 𝑡̂ × 𝑛̂. So, we have defined that, 𝑏̂ =
𝑑𝑏
𝑡̂ × 𝑛̂ , right and since 𝑏̂ is a unit vector, 𝑑𝑠 if not 0 must be perpendicular, to 𝑏̂. So, it is a
𝑑𝑏
unit vector; that means, its magnitude is 1, but if it is not a 0 vector then in that case 𝑑𝑠 the

derivative of 𝑏̂ with respect to arc length must be perpendicular to 𝑏̂, alright.

𝑑𝑏̂
And here if I write 𝑏̂ = 𝑡̂ × 𝑛̂ and if I differentiate both sides with respect to s, then 𝑑𝑠 =
𝑑𝑡̂ 𝑑𝑛̂ 𝑑𝑡̂
× 𝑛̂ − 𝑡̂ × 𝑑𝑠 , right. Now, = 𝜅𝑛̂, alright. So, I can write here 𝜅𝑛̂ , sorry kappa is a
𝑑𝑠 𝑑𝑠
𝑑𝑏̂ 𝑑𝑛̂
scalar quantity = 𝜅𝑛̂ × 𝑛̂ − 𝑡̂ × 𝑑𝑠 . Now, 𝑛̂ × 𝑛̂ = ⃗0⃗ because they are the I mean
𝑑𝑠

𝑎⃗ × 𝑎⃗ = ⃗0⃗. So, this is basically a ⃗0⃗. So, k times ⃗0⃗ is again a ⃗0⃗, now we subtract it, you add
𝑑𝑛̂
it in −𝑡̂ × 𝑑𝑠 , then it will remain the same.

482
(Refer Slide Time: 22:54)

𝑑𝑏̂ 𝑑𝑛̂ 𝑑𝑡̂

So, here we will obtain = −𝑡̂ × 𝑑𝑠 , alright. So, from here we can say that; so and
𝑑𝑠 𝑑𝑠

why there is a minus sign. So, there should not be any minus sign. So, there is a plus yes.
𝑑𝑛̂
So, this is +𝑡̂ × 𝑑𝑠 and then we have this thing so, of course its plus.

𝑑𝑏̂ 𝑑𝑏̂ 𝑑𝑛̂

Now, that means, from here we can write, this 𝑑𝑠 is perpendicular to 𝑡̂ because 𝑑𝑠 = 𝑡̂ × 𝑑𝑠 .
𝑑𝑏̂ 𝑑𝑛̂ 𝑑𝑏̂ 𝑑𝑛̂ 𝑑𝑏̂
Now, 𝑑𝑠 and they are along the same. So, 𝑑𝑠 and from here we can say that this 𝑑𝑠 is
𝑑𝑠 𝑑𝑠

actually perpendicular to this 𝑡̂ from this formula. So, from this formula we can be able to
𝑑𝑏̂
say that is perpendicular to 𝑡̂, but it is also perpendicular to 𝑏̂, we have stated here.
𝑑𝑠

So, it is perpendicular to 𝑏̂ and it is perpendicular to 𝑡̂; that means, and must therefore, be
parallel to 𝑛̂. So, you have a vector which is perpendicular to both 𝑏̂ and 𝑡̂. So, if it is
perpendicular to both 𝑏̂ and 𝑡̂; that means, it must be along the direction of 𝑛̂ because 𝑛̂ is
𝑑𝑏̂
the vector which is perpendicular to 𝑡̂ and 𝑏̂. So, the only possibility this has is to
𝑑𝑠
𝑑𝑏̂
become parallel with 𝑛̂ . So, if it is parallel to 𝑛̂ then I can write 𝑑𝑠 = −𝜁𝑛̂, and where 𝜁 is

a scalar called as the torsion of the curve at the point P. And the minus sign in this case is
taken, I mean this the minus sign has this purpose that I mean basically this minus sign is
taken because, how do I write it in a nice word.

483
So, the minus sign is taken because when 𝜁 is positive, so this is not t this is tau when 𝜁 is
𝑑𝑏̂
positive has the direction of −𝑛̂. Then, as P moves along the curve in the positive
𝑑𝑠

direction, 𝑏̂ revolves about 𝑡̂ in the same sense as a right handed screw system, advancing
in the direction of 𝑡̂.

(Refer Slide Time: 27:31)

So, it is taken purposely because in the direction of 𝑡̂, so that it maintains that
𝑑𝑏̂
perpendicularity condition that if you have 𝑑𝑠 = −𝜁𝑛̂. So, then in that case it will be along

the curve in the positive, so it will then as the point P moves along the curve then 𝑏̂ will
revolves about 𝑡̂ in the same sense as a right handed screw system. So, it will then form a
right handed screw system like we stated earlier. So, that is why we have taken a minus
𝑑𝑏̂
sign, alright. So, we have = −𝜁𝑛̂ and therefore, to summarize we have derived several
𝑑𝑠

types of formula.

484
(Refer Slide Time: 28:25)

𝑑𝑟⃗ 𝑑𝑡̂
So, first of all we have derived = 𝑡̂ alright, then we derived = 𝜅𝑛̂, and then we
𝑑𝑠 𝑑𝑠
𝑑𝑏̂
derived = −𝜁𝑛̂, alright. So, these are the 3 important formulas that we derived today.
𝑑𝑠

In the next class we will finally, be able to derive this Serret Frenet formula. We will also
work out few examples just to see why we need these formulas. So, today we will stop
here, and I thank you for your attention and then we will continue with our same topic in
the next class.

Thank you.

485
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 47
Introduction and Derivation of Serret-Frenet Formula, Few Results

Hello students. So, in the last class we started with the concepts of a little bit towords
differential geometry aspect of Vector Calculus. So, we introduced the concepts of
curvature and torsion and we also introduced how the unit tangent vector unit principle
normal and the unit binormal are connected with one another and we sort of derived some
relations.

(Refer Slide Time: 00:49)

𝑑𝑡̂
So, for example, we obtained relations like 𝑑𝑠 = 𝜅𝑛̂ where 𝜅 is the curvature and then we
𝑑𝑏̂
obtained relations like =, where 𝑡̂ is the unit tangent vector 𝑏̂ is the unit binormal. So,
𝑑𝑠
𝑑𝑏̂
= −ζ𝑛̂ where 𝑛̂ is the unit principal normal and 𝜁 is the torsion, we also obtain 𝑏̂ =
𝑑𝑠

𝑡̂ × 𝑛̂ I prefer to write a cap.

Because there are unit vectors, but you can also write a vector notation and make sure you
specify that there are unit vectors; so, for example, 𝑏̂ is a unit binormal. And similarly you
can write 𝑡̂ = 𝑏̂ × 𝑛̂ or 𝑛̂ × 𝑏̂ so, so whenever you are changing the order so, then in that

486
case you have to put a minus sign. And of course, these 3 vectors they form a right handed
screw system. So, they are mutually perpendicular to one another like the unit vectors
𝑖̂, 𝑗̂ and 𝑘̂ along the coordinate axes.

𝑑𝑏̂
So, we derived these 2 relations. So, we derived something like = −𝜁𝑛̂. So, this one
𝑑𝑠
𝑑𝑛̂ 𝑑𝑏̂
was also obtained. And now how do we obtain ? So, we have obtained excuse me
𝑑𝑠 𝑑𝑠
𝑑𝑛̂
and how do we obtain now .
𝑑𝑠

(Refer Slide Time: 02:38)

𝑑𝑛̂
So, in order to obtain the first of all we will write that right handed screw system 𝑛̂ =
𝑑𝑠

𝑏̂ × 𝑡̂. So, this is also part of that right handed screw system relation. And now we
differentiate both sides with respect to the arc length all right. So, let us see what happens
𝑑𝑛̂ 𝑑𝑏̂
differentiating with respect to s all right. So, when we differentiate this will be = ,I
𝑑𝑠 𝑑𝑠

am not writing cap for the moment, but it is understood all right just to save some time
𝑑𝑛 𝑑𝑏 𝑑𝑡 𝑑𝑏̂
now 𝑑𝑠 = × 𝑡 + 𝑏 × 𝑑𝑠. So, now 𝑑𝑠 = −𝜁𝑛̂ all right cross product with 𝑡̂ of course and
𝑑𝑠
𝑑𝑡
b times 𝑑𝑠 = 𝜅𝑛.

𝑑𝑛
So, this is basically 𝑑𝑠 = (−𝜁𝑛) × 𝑡 + 𝑏 × 𝜅𝑛 = −𝜁(𝑛 × 𝑡) + 𝜅(𝑏 × 𝑛), then in that case

if we do (𝑛 × 𝑡), then in that case it will be −(𝑛 × 𝑡) right. So, if you change the order of
that cross product then in that case you have to put a minus sign. So, here this is nothing

487
but minus so this is nothing, but 𝑏̂ = 𝑡̂ × 𝑛̂ = −𝑛̂ × 𝑡̂ ⇒ 𝑛̂ × 𝑡̂ = −𝑏̂. So, I can substitute
𝑛̂ × 𝑡̂ as −𝑏̂ so minus minus this will turn into a plus and Kappa times. So, this is not K
this is basically Kappa make sure you see that Greek letter in some literature.

So, then so basically any mathematical books on notations, there you can be able to find
this Kappa its a Greek letter and yes so this is Kappa and then we have b× 𝑛. So, we know
that 𝑏 = 𝑡 × 𝑛; 𝑛 = 𝑏 × 𝑡 and 𝑡 = 𝑛 × 𝑏 all right. So, let me just confirm if this is true so,
𝑡 = 𝑛 × 𝑏 that is correct. So, now that we have this here I can substitute 𝑏 × 𝑛 = −𝑡.

𝑑𝑛 𝑑𝑛
So, this will be = 𝜁𝑏 + 𝜅(𝑏 × 𝑛). So, this is basically we can write as = 𝜁𝑏 − 𝜅𝑡.
𝑑𝑠 𝑑𝑠
𝑑𝑛̂
So, therefore, = −𝜅𝑡̂ + 𝜁𝑏̂.
𝑑𝑠

(Refer Slide Time: 05:38)

𝑑𝑡̂ 𝑑𝑏̂ 𝑑𝑛̂

And if I write these 3 relations all together that is 𝑑𝑠 = 𝜅𝑛̂, 𝑑𝑠 = −𝜁𝑛̂ & − 𝜅𝑡̂ + 𝜁𝑏̂. So,
𝑑𝑠

these are the 3 relations that connects the derivative of 𝑡̂, 𝑏̂ and 𝑛̂. So, basically derivative
of unit tangent vector derivative of binormal and derivative of unit principal normal with
𝑛̂, 𝑡̂ and 𝑏̂.

𝑑𝑡̂ 𝑑𝑏̂ 𝑑𝑛̂

So, = 𝜅𝑛̂, 𝑑𝑠 = −𝜁𝑛̂ & − 𝜅𝑡̂ + 𝜁𝑏̂ and these 3 relations all together are called as
𝑑𝑠 𝑑𝑠

Serret Frenet’s formula. So, these are the required Serret Frenet formula. So, if you have
a let us say a right handed screw system of a unit tangent vector, unit principle normal and

488
unit binormal, then they are connected with that 𝑏̂ = 𝑡̂ × 𝑛̂ and then 𝑡̂ = 𝑛̂ × 𝑏̂ or
𝑏̂ × 𝑛̂ something like that.

So, that is the relation amongst the unit tangent vector unit principle normal and unit
binormal and the relation between their derivatives will be given in this fashion and this
relation is basically called as Serret Frenet formula. It is also very useful because if you
know the unit tangent vector, then from there you can be able to calculate your curvature.
And if you have a unit binormal given then from there you can be able to calculate, your
torsion and if you have how to say your torsion and your binormal. And all the other things
are given on the right hand side of this equation, then you can be able to calculate the
derivative of a unit principal normal.

So, this relation is very useful and its widely used in vector calculus and also in some parts
of mechanics as well. So, this is what we wanted to derive and its also part of our syllabus.
So, we are glad that we were able to derive this relation. So, this is one of the important
relations of vector calculus keep in mind so, we will put a star here all right.

Now that we have this relation of course, we would like to solve few examples. So, let us
start with our examples. So, how do we calculate these things 𝜅 and 𝜁?

So, there must be some kind of formula that would relate these derivatives with the curve
itself. So, the equation of the curve if we remember, its 𝑟⃗ = 𝑓⃗(𝑡) and from there we were
𝑑𝑟⃗ 𝑑𝑟⃗
calculating which is actually the tangent vector. So, we were calculating then from
𝑑𝑡 𝑑𝑡
𝑑𝑟⃗
there we can calculate 𝑑𝑠 and all that.

𝑑𝑡̂ 𝑑𝑡̂ 𝑑𝑏̂ 𝑑𝑛̂

So, this 𝑑𝑠 the obvious question is this 𝑑𝑠 , 𝑑𝑠 and must connect with some derivatives
𝑑𝑠
𝑑𝑟⃗ 𝑑𝑡̂
of 𝑟⃗ and nothing else. So, that we do not have to calculate those and and things like
𝑑𝑠 𝑑𝑠

that. We do not want to calculate those things we want to calculate the derivative of 𝑟⃗ with
respect to t only. And it could be first order derivative, it could be second order derivative,
it could be third order derivative whatever it is, we just want to express these parameters
like torsion and curvature in terms of the derivative of 𝑟⃗ we do not want to go to these
𝑑𝑡̂ 𝑑𝑏̂
, . So, how we can do that? Alright.
𝑑𝑠 𝑑𝑠

489
So, we will of course, use these formulas to derive that relation where only 𝑟⃗ and its
derivatives are connected with Kappa and tau all right.

(Refer Slide Time: 10:11)

So, let us go. So, the vector form, of a curve is 𝑟⃗ = 𝑓⃗(𝑡) right. And then we can be able to
write here 𝑥 = 𝑥(𝑡), 𝑦 = 𝑦(𝑡) & 𝑧 = 𝑧(𝑡). So, this is the Cartesian form, this is the vector
𝑑𝑟⃗
form all right. Now, if I differentiate this thing here, so then in that case this is 𝑑𝑡 .

𝑑𝑟⃗
So, we know that derivative can also be written as 𝑟̇ and this one is basically . So, we
𝑑𝑠
𝑑𝑟 𝑑𝑟 𝑑𝑠 𝑑𝑟
can be able to write = 𝑟̇ = 𝑑𝑠 × 𝑑𝑡 and = 𝑡̂. So, let us write this as unit tangent
𝑑𝑡 𝑑𝑠
𝑑𝑟 𝑑𝑠
vector. So, 𝑑𝑡 = 𝑡̂ 𝑑𝑡 = 𝑡̂𝑠̇ .

So, this is what we know. So, from here it follows that |𝑟̇ | = |𝑡̂𝑠̇| = |𝑡̂||𝑠̇ | all right. So,
derivative of s with respect to t and |𝑡̂| = 1 . So, since 𝑡̂ is a unit tangent vector we can
𝑑2 𝑟 𝑑𝑡̂
write |𝑡̂| = 1. So, this we can always write. Now, again we have 𝑟̈ = 𝑑𝑡 2 = 𝑑𝑠 × |𝑠̇ |2 + 𝑡̂𝑠̈.

𝑑𝑡̂ 𝑑𝑡̂
Now, 𝑑𝑠 if we go to the Serret Frenet formula 𝑑𝑠 = 𝜅𝑛̂. So, let us substitute that here. So,

this is 𝜅𝑛̂ and this is not cross product actually. So, do not confuse it with this cross
product, its just a usual multiplication. So, I do not know; let us leave it like that. So, 𝑟̈ =
𝑑2 𝑟 𝑑𝑡̂
= 𝑑𝑠 × |𝑠̇ |2 + 𝑡̂𝑠̈ = 𝜅𝑛̂|𝑠̇ |2 + 𝑡̂𝑠̈ all right and lastly we will calculate 𝑟⃛.
𝑑𝑡 2

490
(Refer Slide Time: 12:59)

𝑑3 𝑟⃗ 𝑑3 𝑟⃗ 𝑑𝑛̂ 3
So, 𝑟⃛ = and this will be basically if I go here then, this will be 𝑟⃛ = = 𝑠̇ +
𝑑𝑡 3 𝑑𝑡 3 𝑑𝑠
𝑑𝑡̂ 𝑑𝑠 𝑑2 𝑠 𝑑 𝑠 3 𝑑𝑠 𝑑 𝑠 2
+ 𝑡̂ 𝑑𝑡 3 + 2𝜅𝑛̂ 𝑑𝑡 𝑑𝑡 2 .
𝑑𝑠 𝑑𝑡 𝑑𝑡 2

𝑑𝑠 𝑑 𝑠 2
𝑑 𝑠 3 𝑑𝑠 𝑑 𝑠 2
So, let us put here; its (𝜁𝑏̂ − 𝜅𝑡̂)𝑠̇ 3 + 𝜅𝑛̂ 𝑑𝑡 𝑑𝑡 2 + 𝑡̂ 𝑑𝑡 3 + 2𝜅𝑛̂ 𝑑𝑡 𝑑𝑡 2 + 𝜅𝑛̂𝑠̇ 2 . So, let me

just match; if I have done it correctly, so, yes. So, now we rearrange the terms.

So, if we rearrange the terms; then we will obtain (𝑠⃛ − 𝑠̇ 𝜅 2 )𝑡̂ + (3𝑠̇ 𝑠̈ + 𝑠̇ 2 𝜅)𝑛̂ + 𝑠̇ 3 𝜅𝜁𝑏̂.

So, 2 will come at the front and then s dot and then, so, here we had; so, 2 will come at, at
the front. So, I am using basically the product rule. So, the other things are fine. So, there
was one more term missing.

It is a little bit complicated formula, but not really that much complicated. So, ultimately
𝑑𝑡̂
basically we will obtain . So, let me just match again. So, I am matching with my lecture
𝑑𝑠

notes.

So, so, here I can differentiate 𝜅 as well. So, I will obtain. So, there will be another
term, 𝜅𝑛̂𝑠̇ 2 So, it’s just how to say Kappa is also a function of 𝑛̂. So, we can differentiate
Kappa as well. So, ultimately we will obtain this expression all right.

491
So, 𝑟⃛ = (𝑠⃛ − 𝑠̇ 𝜅 2 )𝑡̂ + (3𝑠̇ 𝑠̈ + 𝑠̇ 2 𝜅)𝑛̂ + 𝑠̇ 3 𝜅𝜁𝑏̂ now its fine. So, basically at the end you
will obtain this relation it is a little bit complicated. And now, that we have this relation
here, we can write hence. So, if we take 𝑟̇ × 𝑟̈ , then we just know what is our 𝑟̇ ; 𝑟̇ = 𝑡̂𝑠̇ and
I am taking cross product with 𝑟̈ , so, basically 𝑟̇ × 𝑟̈ = 𝑡̂𝑠̇ × (𝜅𝑛̂𝑠̇ 2 + 𝑡̂𝑠̈ )

So, if we take the cross product, then this will be basically 𝑟̇ × 𝑟̈ = 𝑡̂𝑠̇ × (𝜅𝑛̂𝑠̇ 2 + 𝑡̂𝑠̈) =
𝜅𝑏̂𝑠̇ 3 + 0 = 𝜅𝑏̂𝑠̇ 3 . So, ultimately I can be able to write this as 𝜅𝑏̂𝑠̇ 3 + 𝑠̇ 𝑠̈(𝑡̂ × 𝑡̂) = 𝜅𝑏̂𝑠̇ 3 +
0 . So, ultimately 𝑟̇ × 𝑟̈ is 𝜅𝑏̂𝑠̇ 3 .

(Refer Slide Time: 20:30)

And if I take (𝑟̇ × 𝑟̈ ). 𝑟⃛ = 𝑠̇ 6 𝜅 2 𝜁 simply. And therefore, so, this is let us say my relation
number 2 and this is my relation number 1.

|𝑟̇ ×𝑟̈ |
So, from both of these 2 relations, what we have is? 𝜅 = right. And 𝑟̇ = 𝑡̂𝑠̇ ⇒ |𝑟̇ | =
𝑠̇ 3

|𝑠̇ | = 𝑠̇ right. So, 𝑠̇ is basically the arc length whose derivative is with respect to t.

alright, because it does not have many components. So, it just have one component, so,
|𝑟̇ ×𝑟̈ | |𝑟̇ ×𝑟̈ | 𝑑𝑠
this will remain just like s dot. So, here I can write 𝜅 = = . So, 𝑑𝑡 or 𝑠̇ simply is
𝑠̇ 3 |𝑟̇ |3

nothing but |𝑟̇ | all right. So, we have this thing here and now 𝜁 can be calculated as; 𝜁 =
(𝑟̇ ×𝑟̈ ).𝑟⃛
so, from relation number 2.
𝑠̇ 6 𝜅 2

492
So, this is from relation number 1. Now, this is from relation number 2 I have 𝜁 =
(𝑟̇ ×𝑟̈ ).𝑟⃛ |𝑟̇ |6
× |𝑟̇ ×𝑟̈ |2 here.
𝑠̇ 6

So, 𝑠̇ = |𝑟̇ |that is what we have here. So, basically these two will get cancelled and then
(𝑟̇ ×𝑟̈ ).𝑟⃛ |𝑟̇ |6 (𝑟̇ ×𝑟̈ ).𝑟⃛
we will obtain × |𝑟̇ ×𝑟̈ |2 = .
|𝑟̇ |6 |𝑟̇ ×𝑟̈ |2

|𝑟̇ ×𝑟̈ | (𝑟̇ ×𝑟̈ ).𝑟⃛

So, basically 𝜅 = and 𝜁 = . So, this is how we obtain the curvature and the
|𝑟̇ |3 |𝑟̇ ×𝑟̈ |2

torsion of the given curve and you can see here, that I mean now we do not have to
𝑑𝑏̂ 𝑑𝑛̂ 𝑑𝑡̂
calculate any , or or anything like that. We just have to differentiate the given
𝑑𝑠 𝑑𝑠 𝑑𝑠

curve 𝑟⃗ = 𝑓⃗(𝑡) with respect to t at least thrice, because for torsion you need to calculate
𝑟⃛; that means, derivative of r up to third order with respect to t.

So, we just have to calculate the time derivative or the derivative with respect to the
parameter t for the curve 𝑟⃗. And we just calculate these dot products cross product and
divided by this mod of cross product whole square and that will give us the torsion. And
|𝑟̇ ×𝑟̈ |
similarly, for the curvature we can calculate and that will give us the curvature.
|𝑟̇ |3

So, of course, it was a slightly how to say tricky here, its not tricky actually, you just have
to spend some time calculating this derivative. So, just remember Kappa is also a function
of s. So, when we were differentiating the first term, we should have for example, here we
𝑑𝑛̂
had a 𝜅𝑛̂𝑠̇ . So, 𝜅𝑛̂. So, 𝑑𝑠 and then there should have been a Kappa here; there should have

been a Kappa here and then when we are differentiating Kappa then 𝑛̂. 𝑠̂ will be unchanged.

𝑑𝑛̂
And now we will use the formula for from Serret Frenet formula. We substituted the
𝑑𝑠
𝑑𝑡̂ 𝑑𝑛̂
value for 𝑑𝑠 , 𝑑𝑠 and then we are just calculating the terms and we just take the cross product

with 𝑟̇ and 𝑟̈ that will give us this relation and we take the cross product of 𝑟̇ with 𝑟̈ , 𝑟⃛ that
will give us this relation and this comes from I and this comes from II.

So, from I and from II, these are the required relations. So, we see that when we want to
calculate the curvature and torsion of a given curve we really do not have to go through all
𝑑𝑛̂ 𝑑𝑏̂
those and . We just differentiate the curve with respect to the parameter t up to 3
𝑑𝑠 𝑑𝑠

times and then we calculate this dot product and cross product and that will give us the

493
required curvature and torsion of a given curve. So, we will practice a few examples based
on these formulas in our next class.

And I thank you for your attention and I look forward to you in your next class.

494
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 48
Example on binormal, normal tangent, Serret-Frenet Formula

Hello students. So, in the last class we started with derivation of Serret-Frenet formula
and then we derived a result which actually connects the time derivative or the derivative
with respect to the parameter of a given curve 𝑟⃗ = 𝑓⃗(𝑡) its with curvature and torsion.

𝑑𝑛̂ 𝑑𝑏̂
So, we saw that we really do not have to calculate , and all those things in order to
𝑑𝑠 𝑑𝑠

calculate the curvature and torsion we just have to differentiate the given curve with
respect to the parameter and we require derivative at least up to the third order to calculate
these things and it just makes our life a little bit easier instead of going through all those
𝑑𝑛̂ 𝑑𝑏̂
the and calculation.
𝑑𝑠 𝑑𝑠

So, today we will work out few examples. And we will see if we have some time then we
can move on to our next aspect of vector calculus which is application to mechanics, but
for the time being let us start with some examples alright. So, now, you might be asked in
your exam to prove few relations.

(Refer Slide Time: 01:27)

495
So, the first one is example. So, prove the relation 𝜅 = |𝑟̈ |, 𝑟̇ × 𝑟̈ = 𝜅𝑏̂, 𝑟⃛ =
𝜅 ′
𝜅(𝜁𝑏̂ − 𝜅𝑡̂) + 𝑛̂ , 𝑟̇ . (𝑟̈ × 𝑟⃛) = 𝜅 2 𝜁; so we will prove these relation.

So, for example, if you have a given curve 𝑟⃗ then we know that this formula is true. So,
we saw that in the previous class that this formula is true; however, you can actually
𝑑2 𝑟⃗ 𝑑2 𝑟⃗
calculate and |𝑑𝑡 2 | will also give you the curvature alright. So, let us see how we can
𝑑𝑡 2

do that; so the first relation is to prove this one. So, from Serret-Frenet formula we have,
𝑑𝑟⃗
first of all we have 𝑡̂ = .
𝑑𝑡

𝑑𝑡 𝑑𝑡
So, if we differentiate both sides then it will be 𝑑𝑠. So, it will be 𝑑𝑠 if I differentiate both
𝑑𝑡 𝑑2 𝑟 𝑑2 𝑟
sides. So, 𝑑𝑠 = now dt ds is from Serret-Frenet formula is 𝜅𝑛̂. So, it is 𝜅𝑛̂ = 𝑑𝑡 2 . So,
𝑑𝑡 2

now, if I take mod on both sides then in that case this will be 𝑟̈ which is basically |𝑟̈ | =
|𝜅𝑛̂| = 𝜅|𝑛̂| = 𝜅. And therefore, this is our 𝜅.

So, the first relation is 𝜅 = |𝑟̈ |, is proved all right our next relation is 𝑟̇ × 𝑟̈ . So, let us see
what is our 𝑟̈ .

(Refer Slide Time: 04:35)

𝑑2 𝑟
So, in the previous relation; so in the relation one we have shown that equals to what
𝑑𝑡 2
𝑑𝑟⃗
is that? 𝜅𝑛̂ and we know that 𝑑𝑡 is basically our t and wait wait wait this one we can take.

496
So, here when we have dot I just forgot to mention, that here dots denote differentiation
with respect to s right; so this one is 𝑑 2 ok.

So, here differentiation denotes the there is a dot denotes the differentiation with respect
to s, not with respect to t. So, earlier the differentiation where with respect to the parameter
t, but in this case the differentiation is with respect to the arc length s because we can also
write 𝑟 = 𝑓(𝑠) as a curve, where s is the arc length and then in that case these derivatives
𝑑2 𝑟
are basically derivatives with respect to s not with respect to t. So, of course, 𝜅 = 𝑑𝑠2 .

(Refer Slide Time: 05:55)

|𝑟̇ ×𝑟̈ |
However, in the previous class we saw that 𝜅 = . So, here the dots are with is to t. So,
|𝑟̇ |3

when the equation of the curve is 𝑟⃗ = 𝑓⃗(𝑡) alright, but in this case these dots are with
respect to s. So, that was something I sort of skipped to tell you it is just that these
𝑑2 𝑟⃗ 𝑑𝑟⃗
derivatives are with respect to s only alright. Now, we have = 𝜅𝑛̂, = 𝑡̂ alright.
𝑑𝑠2 𝑑𝑠

𝑑𝑟⃗ 𝑑 𝑟⃗ 2
So, when I take the cross product which is 𝑑𝑠 × 𝑑𝑠2 = 𝑡̂ × 𝜅𝑛̂ = 𝜅(𝑡̂ × 𝑛̂) ⇒ 𝑟̇ × 𝑟̈ = 𝜅𝑏̂;

so I can write 𝑏̂ here.

So, you see 𝑟̇ × 𝑟̈ = 𝜅𝑏̂ and these dots are taken with respect to s only. Now third relation
𝑑𝑟⃗
is 𝑟⃛. So, 𝑟⃛ is basically we know that 𝑑𝑠 = 𝜅𝑛̂ ; so I can differentiate twice.

497
 𝑑𝑘 𝑑𝑟⃗ 𝑑2 𝑟 𝑑3 𝑟 𝑑𝑘 𝑑𝑛̂
So, this will be 𝑑𝑠 , 𝑑𝑠 is basically t and 𝑑𝑠2 = 𝜅𝑛̂. So, 𝑑𝑠3 = 𝑛̂ + 𝜅 𝑑𝑠 .
𝑑𝑠

(Refer Slide Time: 08:37)

𝑑3 𝑟 𝑑𝑘 𝑑𝑛̂ 𝑑𝑘
So, now this is basically = 𝑛̂ + 𝜅 𝑑𝑠 = 𝑛̂ + 𝜅(−𝜅𝑡̂ + 𝜁𝑏̂) is from Serret-Frenet
𝑑𝑠3 𝑑𝑠 𝑑𝑠
𝑑𝑛̂ 𝑑 𝑟 3
formula we have = −𝜅𝑡̂ + 𝜁𝑏̂. So, this is ultimately 𝑑𝑠3 = 𝜅(𝜁𝑏̂ − 𝜅𝑡̂) + 𝜅 ′ 𝑛̂ and if I
𝑑𝑠

multiply both numerator and denominator with 𝑛̂. So, or we can leave it like that.

So, in the formula we do not have to take a division. So, in the formula we can leave it like
that. So, let us not divide it or anything; so this is just 𝜅 ′ 𝑛̂. So, 𝜅 ′ multiplied with 𝑛̂. So,
this is what we are required to prove in the third relation, and the fourth relation is basically
𝑟̇ . (𝑟̈ × 𝑟⃛) = 𝜅 2 𝜁. So, here all of them are taken with respect to s. So, 𝑟̇ = 𝑡̂, 𝑟̈ = 𝜅𝑛̂ and
𝑟⃛ = 𝜅(𝜁𝑏̂ − 𝜅𝑡̂) + 𝜅 ′ 𝑛̂.

So, all you have to do is calculate this cross product. So, we do the cross product of this
and this first and then cross product of this and this and all that. So, keep in mind that 𝑏̂ =
𝑡̂ × 𝑛̂ 𝑜𝑟 𝑛̂ × 𝑡̂ whatever it is. So, just use those formulas and whenever you are using them
use their proper sign whether you have to use 𝑏̂ × 𝑡̂ or whether you have to use
𝑡̂ × 𝑏̂ depending on that you will use the minus sign, and just substitute those signs here.
It is just some vector algebra calculation nothing more and when you substitute all those
things then it will simplify nicely into something like this.

498
So, here we have excuse me, So, we are calculating 𝑟̇ . (𝑟̈ × 𝑟⃛) = 𝑡̂. [𝜅𝑛̂ × [𝜅(𝜁𝑏̂ − 𝜅𝑡̂) +
𝜅 ′ 𝑛̂] = 𝑡̂. (𝜅 2 𝜁𝑡̂ + 𝜅 3 𝑏̂) = 𝜅 2 𝜁 ⇒ 𝑟̇ . (𝑟̈ × 𝑟⃛) = 𝜅 2 𝜁.

So, this is ultimately 𝜅 2 𝜁 and 𝑡̂. 𝑏̂ = 0 because they are mutually perpendicular to one
another. So, the other term is 0 and therefore, this is what we needed to prove. So, we had
to prove 𝑟̇ . (𝑟̈ × 𝑟⃛) = 𝜅 2 𝜁 and this is what we needed to prove. So, remember in mind that
here the dots are taken with respect to the arc length. So, instead of writing the dots we
could have taken, here a different notation let us say dash. So, dash dash dash and dash
here dash here dash. So, the dash s basically denotes the differentiation with respect to s.

So, instead of taking the dot we can take dash just to distinguish it that we are not taking
the derivative with respect to t instead we are taking derivative with respect to the arc
length although its not something to confuse about since I told you several times and, but
just for the clarity that in this example in order to calculate the curvature 𝜅 and the torsion
𝜁 we have used derivative with respect to s all right.

So, this is one another example where you are asked to prove some relations we might
may try to include examples motivated from this and now let us work out some examples;
so some more examples actually.

(Refer Slide Time: 12:59)

1
So, we will start with for the twisted cube or cubic 𝑥 = 2𝑡, 𝑦 = 𝑡 2 , 𝑧 = 3 𝑡 3 , if you are

given these then find 𝑡̂, 𝑛̂, 𝑏̂. So, they are all unit tangent vector(𝑡̂), unit principle

499
normal(𝑛̂), unit binormal(𝑏̂), kappa(𝜅), and tau(𝜁) at 𝑡 = 1. So, at 𝑡 = 1 our given point
1
can be basically (2,1, 3). So, the given equation of the curve is, the given curve is 𝑟⃗ =
1
𝑥(𝑡)𝑖̂ + 𝑦(𝑡)𝑗̂ + 𝑧(𝑡)𝑘̂ = 2𝑡𝑖̂ + 𝑡 2 𝑗̂ + 3 𝑡 3 𝑘̂.

So, first of all we have to calculate 𝑡̂. So, in order to calculate 𝑡̂ what we will do? We will
𝑑𝑟⃗
first do the differentiation with respect to t. So, = 2𝑖̂ + 2𝑡𝑗̂ + 𝑡 2 𝑘̂and we will
𝑑𝑡
2
𝑑 𝑟⃗
calculate 𝑑𝑡 2 = 𝑟̈ = 2𝑗̂ + 2𝑡𝑘̂.

(Refer Slide Time: 15:21)

𝑑3 𝑟⃗
And then we will calculate = 0𝑖̂ + 0𝑗̂ + 2𝑘̂ right. So, this is ultimately (0,0,2),
𝑑𝑡 3

previously it was (0,2,2) and this is (2,2𝑡, 2𝑡 2 ) alright. Now at the point 𝑡 = 1 r is [𝑟⃗]𝑡=1 is
1 𝑑𝑟⃗ 𝑑2 𝑟⃗
the point P is basically (2,1, 3) then the derivative is [ 𝑑𝑡 ]𝑡=1 = (2,2,1) then [𝑑𝑡 2 ]𝑡=1 =
𝑑3 𝑟⃗
(0,2,2) and [ ] = (0,0,2) alright.
𝑑𝑡 3 𝑡=1

So, from here the unit tangent vector; so first of all and we can also calculate these factors.
So, 𝑟̇ × 𝑟̈ = we will take the cross product of this and this which is fairly easy to calculate
I leave this task up to the students. So, 𝑟̇ × 𝑟̈ = (2, −4,4) I have the results calculated in
my lecture notes. So, I am just writing it from there and 𝑟̇ × 𝑟̈ . 𝑟⃛ = 8 all right.

500
So, now we have all these three ingredients. So, what are we going to do? So, this is the
dot product and then this is the cross product.

So, dot product and this is the cross product. So, because we cannot (𝑟̇ . 𝑟̈ ) × 𝑟⃛ yes. So,
because we cannot we cannot simply(𝑟̇ × 𝑟̈ ) . So, 𝑟̇ × 𝑟̈ . 𝑟⃛. So, (𝑟̇ × 𝑟̈ ). 𝑟⃛.

So, ultimately this formula is 𝑟̇ × 𝑟̈ . 𝑟⃛ alright. So, we have 𝑟̇ × 𝑟̈ . 𝑟⃛ yeah. So, there was a
slight modification in the calculation of 𝜁. So, this we have to keep in mind; so there is a
slight vector notation issue here.

(Refer Slide Time: 19:19)

So, I copied it wrong; so, it is basically 𝑟̇ × 𝑟̈ . So, I had to copy it 𝑟̇ × 𝑟̈ . 𝑟⃛ so the rest of the
things are same.

So, here I made a small error while copying the results from the previous slide. So, that
can happen to alright. So, 𝑟̇ × 𝑟̈ . 𝑟⃛ = 8. So, now, our unit tangent vector; so the first thing
⃗⃗
𝑑𝑟
(2,2,1) 1
that we will calculate is unit tangent vector which is 𝑡̂ and 𝑡̂ = 𝑑𝑡
⃗⃗
𝑑𝑟
= = 3 (2,2,1).
| | √4+4+1
𝑑𝑡

So, that is the unit tangent vector now that we have the unit tangent vector for unit binormal
for the binormal we have for the binormal we use the formula kappa. So, we use the
𝑑𝑏
formula for the binormal we use 𝑠̇ . So, we for the binormal we use . So, you know that
𝑑𝑠

binormal is given by this relation here 𝑡̂ × 𝑛̂.

501
(Refer Slide Time: 21:01)

So, 𝑡̂ is basically the unit tangent vector and 𝑛̂ is the unit unit normal. So, unit normal can
be given by or simply we just have to find a unit normal which is perpendicular or a unit
vector that is perpendicular to 𝑡̂.

So, the unit vector that is perpendicular to 𝑡̂ is (1, −2,2). So, I can write simply 𝑏̂ =
1
(1, −2,2) because then 𝑛̂ and 𝑏̂ at 𝑡̂ and 𝑏̂ will be perpendicular they are unit vectors and
3

therefore, it is a unit binormal and from here I can calculate my 𝑛̂.

1
So,𝑛̂ = 𝑏̂ × 𝑡̂ = 3 (−2,1,2). So, they are all perpendicular to one another. So, we first

calculated 𝑡̂, from there we can write our 𝑏̂ and based on the 𝑏̂ we can calculate our 𝑛̂ by
just simply calculating 𝑏̂ × 𝑡̂. So, we know our 𝑡̂ is this one, and we know our 𝑏̂ is this one.

So, just calculate their cross product one by three will come up front and then we will have
this as the as the required unit principle normal. So, we have these 2, now we can calculate
our kappa. So, kappa is basically from this relation let me go back to my lecture note and
|𝑟̇ ×𝑟̈ |
there I have 𝜅 = . So, what is my 𝑟̇ × 𝑟̈ ? So, (2, −4,4) so they are mod.
|𝑟̇ |3

|𝑟̇ ×𝑟̈ | |(2,−4,4)| √4+16+16 6 2

So, 𝜅 = = = 3 = 27 = 9
|𝑟̇ |3 |(2,2,1)|3 (√4+4+1)

and 𝜁 can be calculated as using this formula from our previous class from our previous
class.

502
 (𝑟̇ ×𝑟̈ ).𝑟⃛
So, this formula 𝜁 = alright. So, if I take the dot product with 𝑟⃛ what is our 𝑟⃛? So,
|𝑟̇ ×𝑟̈ |2

[𝑟⃛]𝑡=1 = (0,0,2).

(𝑟̇ ×𝑟̈ ).𝑟⃛ 8 8

So, if I take the dot product with this here I will only obtain 𝜁 = = |(2,−4,4)|2 = 62 =
|𝑟̇ ×𝑟̈ |2
8 2
=9.
36

So, therefore, we have calculated the kappa and tau. So, all you have to do is just use some
results what we have studied in Serret-Frenet formula use the appropriate relations to
𝑑𝑟⃗
calculate 𝑑𝑡 . And do not forget to use the perpendicularity condition or the relation between

𝑏̂, 𝜁and 𝑛̂ in half an hour I am sometimes how to say making a small errors in the
calculation, but I am pretty sure you are getting what I am trying to say.

𝑑𝑟⃗
So, you just have to calculate those use the relation between 𝑏̂, 𝑡̂,and 𝑛̂ and from there
𝑑𝑡

if you know one of the vectors you can calculate the other one by simply using the
perpendicularity condition. And from there you can calculate the third one using that
relation that 𝑏̂ = 𝑡̂ × 𝑛̂. And similarly 𝑛̂ = 𝑏̂ × 𝑡̂ and 𝑡̂ = 𝑏̂ × 𝑛̂.

So, if you know any one of them the calculating other 2 would not be difficult and to
calculate curvature and torsion you just have to differentiate the curve with respect to t if
the given equation is in terms of the parameter t. However, if it is given in terms of arc
length then we have also worked out an example where curvature and the derivative of r
with respect to arc length is connected. So, it depends what kind of curve equation is given
to you based on that you use that if you use the respective formula and then you can
calculate the curvature and torsion.

So, that is what we did in today’s class and in the next class we will probably do one or 2
more examples on Serret-Frenet formula and then we will move on to our next topic. So,
I will stop here for today and I look forward to you in your next class.

503
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 49
Osculating Plane, Rectifying plane, Normal plane

Hello students. So, in the last class we were looking into the concepts of a very important
formula in Vector Calculus which is Serret Frenets formula which actually connects the
derivative of tangent normal and binormal with tangent normal binormal curvature and
torsion. So, we derived that formula and we were also working our few examples today
we will continue with that formula and we will look into three different planes where those
three vectors lie actually.

Because we know that all we learnt in the previous class that the tangent vector, the normal
and the binormal they form a right-handed sort of like screw system. So, each one of them
are perpendicular to one another so; that means, the each one of them will lie in one plane
in their respective planes perpendicular to the other two planes in which the other two
vectors lie and we will write down the equations for those planes and then we will try to
work our few examples.

(Refer Slide Time: 01:29)

504
 𝑑𝑡
So, let me start we had the Serret Frenet formula as. So, we had the formula as, =
𝑑𝑠
𝑑𝑏 𝑑𝑛
𝜅𝑛. 𝑑𝑠 = −𝜁𝑛, 𝑑𝑠 = −𝜅𝑡 + 𝜁𝑏, where t b and n are the unit tangent vector unit binormal

and unit principal normal alright.

So, basically what we have is here is first of all the tangent line right. So, we have (i) unit
tangent vector or the tangent line, it is the next we have is the (ii) principal normal or unit
principal normal and the third thing that we have is the (iii) binormal right. So, these are
the three vectors that are connected with this equation.

Now, they form are sort of like a trihedral system or right handed screw system
with; so, these three vectors must be lying in three different planes mutually
perpendicular to one another. So, those three planes have also their names and the
names are like first one is, so the three mutually perpendicular planes in which the
above vectors lie or we can write determine the (i)osculating planes, second one is
(ii) normal plane and third one is (iii) rectifying plane alright.

So, basically the normal plane will contain the normal vector perpendicular to tangent line
and binormal, the oscillating plane we will contain your binormal basically and it will be
perpendicular to 𝑡̂ and 𝑛̂ and the rectifying plane we will contain the tangent sorry here
there is the c missing. So, rectifying plane we will contain the tangent vector and it will be
perpendicular to both b and n alright.

So, what we have is. So, if we want to draw a figure. So, the figure would be slightly. So,
what let me write in words just to make the things clear? So, what do we mean by
oscillating plane? So, the first definition is the oscillating plane to a curve to a curve at a
point P is the plane containing the tangent and the principle normal at P.

So, basically it will contain the principle normal and the tangent at the point P. So, the
equation and the equation of the osculating plane. So, basically it will contain both tangent
and normal and so osculating plane is actually perpendicular to the binormal yes.

So, the equation of the osculating plane is; so, it is not actually containing, but it is
containing two vectors 𝑡̂ and 𝑛̂ and its actually perpendicular to the binormal. So,
oscillating plane is perpendicular to binormal; however, it will contain the tangent and the
principle normal. So, now the equation of the oscillating plane is (𝑅⃗ − 𝑟). 𝑏̂ = 0.

505
(Refer Slide Time: 07:12)

Where our 𝑅⃗ is the position vector of any point on the plane and 𝑟 is the position vector of
the specified point P of the curve. So, 𝑟 is the position vector of any point on that oscillating
plane and 𝑟 is basically the position vector of a point P where we are calculating that
oscillating plane alright.

So, this oscillating plane will be perpendicular to the binormal and it will contain 𝑡̂ and ̂𝑛
. So, if it is containing 𝑡̂ and 𝑛̂ ; that means, it is obviously perpendicular to ̂𝑏 because 𝑡̂
and 𝑛̂ be are perpendicular to one another. So, if the vector is containing let us say if you
have something like a three-dimensional geometry and if we have let us say this, this and
this if this is my 𝑡̂ this is my 𝑛̂ and this is my 𝑏̂. So, if a vector is containing both 𝑡̂ and 𝑛̂
then in that case it is; obviously, perpendicular to 𝑏̂ and if a vector is containing 𝑛̂ and 𝑏̂
then it will be perpendicular to 𝑡̂ and if a vector is containing 𝑏̂ and 𝑡̂ then it will be
perpendicular to 𝑛̂. So, now, we will define the normal plane.

So, the second definition is the normal plane. So, the normal plane is the plane through P
perpendicular to the tangent line so; that means, it will contain 𝑏̂ and 𝑛̂ alright. So,
perpendicular to the tangent line hence containing normal and binormal at the point P. So,
and the equation of the normal plane is 𝑅⃗ . So, 𝑅⃗ is any arbitrary point on that normal
plane (𝑅⃗ − 𝑟). 𝑡̂ = 0 . So, 𝑟 is the position vector of the point P where we are calculating
the normal plane dot product with, since it is perpendicular to the tangent we will write
𝑡̂ equals to 0. So, this is the required equation of the normal plane.

506
And third one is rectifying plane. So, the rectifying plane is the plane through P
perpendicular to the normal and it contains. What are the lines it contains or what are the
vectors it contains? So, it will contain 𝑡̂ and 𝑏̂.

So, it will contain 𝑡̂ and 𝑏̂ and the equation is (𝑅⃗ − 𝑟). 𝑛̂ = 0. So, where 𝑅⃗ is any arbitrary
point on the rectifying plane and 𝑟 is the position vector of the point P and since it is
perpendicular to the normal we have the dot product equals to 0

So, going backwards this one is the equation of the osculating plane this one is the equation
of the normal plane and this one is the equation of the rectifying plane. So, if you are given
a equation of a curve in terms of 𝑟 = 𝑓 (𝑡) and if you are supposed to calculate any one of
these planes, let us say at a point 𝑡 = 𝑎. Then in that case we have to calculate tangent
binormal and normal and based on that of course, we can be able to give equations of any
one of these planes.

So, we will work out one or two examples where we calculate these planes and of course,
you can be able to express these equations in terms of the Cartesian coordinate. So, 𝑡̂ can
be written as so, this can be written as [(𝑋, 𝑌, 𝑍) − (𝑥, 𝑦, 𝑧)]. And if you have 𝑡̂ equals to
a given then in that case you can be able to calculate this (𝑥, 𝑦, 𝑧) and then you take dot
product with 𝑡̂ in terms of the Cartesian coordinate system and that will give you required
equation in the Cartesian form. So, we will see how we can do that alright. So, let me look
into my lecture note to get some examples alright.

507
(Refer Slide Time: 13:26)

So, we have an example here. So, I will start with the same example that we considered
yesterday. So, find the equation of oscillating plane, normal plane and the third one was
1
rectifying plane at the point 𝑡 = 1 for the curve 𝑥 = 2𝑡, 𝑦 = 𝑡 2 , 𝑧 = 3 𝑡 3 .

So, here we are given the equation in as x equals to y equals to and z equals to some
function of t and we have to calculate oscillating plane, normal plane and the rectifying
plane. And of course, for 𝑡 = 1 the point P is given. So, at the point P which can be
evaluated at 𝑡 = 1 we have to calculate the equations of all of these planes.

So, let us see how we can do that. So, first of all the equation of the curve can be written
as 𝑟(𝑡) = 𝑓(𝑡) = 𝑥(𝑡)𝑖̂ + 𝑦(𝑡)𝑗̂ + 𝑧(𝑡)𝑘̂ or I can write it as a triplet alright. So, for the
𝑑𝑟 𝑑𝑟
tangent we need 𝑑𝑡 in a way because if we have 𝑑𝑡 then that is basically the tangent vector

and then you divide it with its magnitude and then we will get the unit tangent vector.

𝑑𝑟
So, we first calculate . So, how do we calculate we differentiate both sides, I can also
𝑑𝑡
𝑑𝑟
write 𝑑𝑡 = 𝑟̇ = (2,2𝑡, 𝑡 2 ) and from here I can evaluate [𝑟̇ ]𝑡=1 = (2,2,1) and of course, the

point P would be at 𝑡 = 1 r at 𝑡 = 1 and that will be the point P next we can calculate 𝑟̈ .

So, now that we are calculating everything let us calculate 𝑟̈ = (0,2,2𝑡). So, [𝑟̈ ]𝑡=1 =
(0,2,2) and I can also calculate triple dot. So, I can calculate let me write it here in small.
So,𝑟⃛ = (0,0,2). So,[𝑟⃛]𝑡=1 = (0,0,2). So, these are the three values for 𝑟̇ , 𝑟̈ and 𝑟⃛ at 𝑡 = 1.

508
(Refer Slide Time: 17:34)

1
And the point P at 𝑡 = 1, 𝑃(𝑥(𝑡 = 1), 𝑦(𝑡 = 1), 𝑧(𝑡 = 1)) = 𝑃(2,1, 3). So, this is my

point P for 𝑡 = 1.

So, now I will calculate. So, first I am being asked to calculate the we are being asked to
calculate the equation of the osculating plane. So, we have all the ingredients, now we start
calculating the equation of the oscillating plane. So, the equation of the osculating plane
is if we go back, is (𝑅⃗ − 𝑟). 𝑏̂ = 0 . So, b is actually our binormal unit binormal and

rectifying plain is (𝑅⃗ − 𝑟). 𝑛̂ = 0 and normal plane is (𝑅⃗ − 𝑟). 𝑡̂ = 0 .

So, first of all we can be able to calculate our tangent. So, the required unit tangent vector
⃗
𝑑𝑟
𝑑𝑟
is 𝑡̂ = 𝑑𝑡
⃗
𝑑𝑟
. So, 𝑑𝑡 or 𝑟̇ is basically at 𝑡 = 1 is (2,2,1). So, at 𝑡 = 1 it is at t they required in
| |
𝑑𝑡

(2,2,1) 1
a tangent vector at 𝑡 = 1 or at P we can write at P. So, this is 𝑡̂ = = 3 (2,2,1).
√4+4+1

1
So, this is ultimately 3 (2,2,1). So, that is our required unit tangent vector. So, from here

the required binormal. vector at P is. So, do not write it is just because two is equals to
vector perpendicular to 𝑡̂.

So, I can write 1 by 2 and of course, I have to consider a unit vector. So, this will be 𝑏̂ =
1
(1, −2,2) and the required unit principal normal at P is 𝑛̂ = 𝑏̂ × 𝑡̂. So, we can calculate
3

509
 1
this cross product and this will ultimately yield 3 (−2,1,2). So, we have been able to

calculate the unit tangent vector unit binormal and unit principle normal. So, now, we have
all the ingredients to write the equation of the oscillating plane. So, let us do that.

So, first the equation of osculating plane is. So, we have (𝑅⃗ − 𝑟). 𝑛̂ = 0 oscillating plane.
For the osculating plane we need to have here our where is that equation. So, we need to
have here our 𝑏̂ or binormal actually. So, the binormal what is that. So, the binormal is
1
[(𝑋, 𝑌, 𝑍) − (2,1, 3)] . (1, −2,2) = 0.

So, I choose any point on the osculating plane as (𝑋, 𝑌, 𝑍), 𝑟 is basically the position
1
vector. So, I can write (𝑋 − 2, 𝑌 − 1, 𝑍 − 3) . (1, −2,2) = 0.

(Refer Slide Time: 22:27)

1 2 2
So, this will be (𝑋 − 2) − 2(𝑌 − 1) + 2 (𝑍 − 3) = 0 ⇒ 𝑋 − 2𝑌 + 2𝑍 = 2 − 2 + 3 = 3.

So, this is the required equation of the of the osculating plane. If you want then you can
keep all these 𝑋, 𝑌, 𝑍 on one side and constant on the other side.

2
So, we can simplify this equation a little bit it will be 𝑋 − 2𝑌 + 2𝑍 = 3, I do not know

from here to here if it is correct or not, but that will be the you can leave it here or you can
leave it here. So, this is the required equation of your oscillating plane in terms of X Y Z

510
you can also use a x,y, z and. So, this is capital Z capital Z capital Z. So, if you want you
can use this small x y z now and its totally safe to do that.

So, that is how we give the equation of the osculating plane, now let us go move on to our
next task which is calculating the equation of; what was that? equation of the normal plane.
So, the equation of the normal plane (𝑅⃗ − 𝑟). 𝑡̂ = 0.

1 1 1
⇒ [(𝑋, 𝑌, 𝑍) − (2,1, 3)] . 3 (2,2,1) = 0 ⇒ 2(𝑋 − 2) + 2(𝑌 − 1) + (𝑍 − 3) = 0 ⇒ 2𝑋 +
1 1 19
2𝑌 + 𝑍 = 4 + 2 + 3 = 6 + 3 = . So, you can check from here to here that is not an is it
3

is a difficult task to do. So, this is the required equation of our normal plane alright.

And now you can use the small x y z as well and the third task that is given to us is the
rectifying plane, the plane that is perpendicular to 𝑡̂ i guess.

(Refer Slide Time: 26:34)

So, the rectifying plane now it is perpendicular to 𝑛̂, rectifying plane (𝑅⃗ − 𝑟).𝑛̂ = 0 ⇒
1 1 1
[(𝑋, 𝑌, 𝑍) − (2,1, 3)] . 3 (2,2,1) = 0 ⇒ −2(𝑋 − 2) + (𝑌 − 2) + 2 (𝑍 − 3) = 0.

So, if we multiply both sides by minus and then we do some simplification you can be
able to obtain the equation of the rectifying plane. So, rectifying plane if we I just have to
check once, rectifying plane is dot 𝑛̂ yes.

511
So, you see for this given curve which is also called as to state to a state cube we were able
to obtain the equation of the osculating plane which is basically osculating plane is
basically a plane containing tangent and principal normal and it is perpendicular to the
binormal. And then we calculated the normal plane, for that given curve which is basically
a plane perpendicular to the tangent and then we calculated the rectifying plane which is a
plane containing 𝑡̂ and 𝑏̂ and perpendicular to 𝑛̂.

So, we just have to calculate these three vectors. So, if you know the tangent vector we
can easily calculate binormal or normal because they are mutually perpendicular to one
another and then the equation of the oscillating plane will be (𝑅⃗ − 𝑟). 𝑡̂ = 0. And then just
substitute the value of t and that will give us the required osculating plane; similarly the
normal plane can be given by this equation. So, we just substituted the value of P and the
tangent and then that is the equation and similarly the equation of the rectifying plane can
be given by this way.

You can also calculate, the length of the curve measured from t up to any point. So, the
length of the curve. So, this example was up to here we can also measure the length of the
curve from 𝑡 = 0 to any point. So, the length of the curve for the same problem from 𝑡 =
𝑡
0 to certain point let us say t to a point to any certain point t is given by 𝐿 = ∫0 |𝑟̇ |𝑑𝑡.

So, we will try to work out an example based on this in our next class. So, in today’s class
we were able to determine the oscillating plane normal plane and the rectifying plane of a
given curve. And we will stop here for today and then in the next class we will continue
with calculating the length of a given curve from a certain point to a certain point and I
look forward to you in your next class.

Thank you.

512
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 50
Application to Mechanics, Velocity Speed, Acceleration

Hello students. So, in the last class we were deriving the equations of oscillating planes,
rectifying planes and normal plane and then I at the end I started to show you a small
example where how you can calculate the length of a curve, but due to time constraint we
could not continue. So, today I will continue with the same example and then we move on
to our next example or the next topic we will see.

(Refer Slide Time: 00:48)

So, in the last class after deriving the equation of the rectifying plane I moved on to show
you how you can calculate the length of a curve from a certain point 𝑡 = 0 to a certain
𝑡
point let us say 𝑡 = 𝑡1 . So, this is basically given by integral ∫0 |𝑟̇ |𝑑𝑡. So, 𝑟̇ is basically
your, So, if you remember from integral calculus the length of a curve from a point 𝑡 =
𝑎 to 𝑡 = 𝑏 is basically L equals to sorry is basically L equals to let me go to a new page.

513
(Refer Slide Time: 01:29)

𝑡=𝑏 𝑑𝑥 𝑑𝑦
So, 𝐿 = ∫𝑡=𝑎 √( 𝑑𝑡 )2 + ( 𝑑𝑡 )2 𝑑𝑡. And how this is coming here is mainly because we know

that 𝑑𝑠 2 = 𝑑𝑥 2 + 𝑑𝑦 2 .

So, if you integrate both sides with respect to s actually. So, this will actually give you say
let us say 𝑡 = 𝑎 to 𝑡 = 𝑏 you are integrating. So, here we substitute 𝑑𝑠 =
𝑡=𝑏 𝑡=𝑏 𝑡=𝑏 𝑑𝑥 𝑑𝑦
√(𝑑𝑥)2 + (𝑑𝑦)2 ⇒ ∫𝑡=𝑎 𝑑𝑠 = ∫𝑡=𝑎 √(𝑑𝑥)2 + (𝑑𝑦)2 = ∫𝑡=𝑎 √( 𝑑𝑡 )2 + ( 𝑑𝑡 )2 𝑑𝑡. So, that

is all.

𝑡=𝑡 𝑑𝑟⃗ 𝑡=𝑡

So, now in case of vector calculus our 𝐿 = ∫𝑡=0 1 | 𝑑𝑡 | 𝑑𝑡 = ∫𝑡=0 1|𝑟̇ |𝑑𝑡 , So, this thing can

be written as; so, this thing can be written as instead of writing it again this thing can be
𝑑𝑟⃗
written as | 𝑑𝑡 | because 𝑟⃗ is a vector times dt or you can write it as a small notation simply

|𝑟̇ |𝑑𝑡. So, these two formulas are same because your 𝑟̇ is nothing, but the |𝑟̇ | is nothing,
but this thing.

1
So, in the example from the previous class where we had 𝑥 = 2𝑡, 𝑦 = 𝑡 2 , 𝑧 = 3 𝑡 3 if we

want to calculate let us say length of the curve from a point 𝑡 = 0 to a point t then that case
𝑡 𝑡
our 𝐿 = ∫𝑡=0|𝑟̇ |𝑑𝑡 = ∫𝑡=0 √4 + 4𝑡 2 + 𝑡 4 𝑑𝑡

514
(Refer Slide Time: 03:58)

𝑡
So, this can be written as ∫𝑡=0 √(𝑡 2 + 2)2 𝑑𝑡 yes and if we integrate then this is basically.

𝑡
So, first of all we have to take it out of the square root. So, this is ∫𝑡=0(𝑡 2 + 2)𝑑𝑡 = 2𝑡 +
1 3
𝑡 right. So, this is the required answer. So, just substitute instead of t if we substitute let
3
1 7 7
us say 𝑡 = 1 then this will be 2 + 3 = 3. So, 3 would be the required length and of course,

you can substitute the unit over there. So, to calculate the length of a curve all you have to
do is calculate 𝑟̇ and that 𝑟̇ will give you the required length of the curve if we use this
formula alright;

So, next I will give you one example to practice because it follows the same logic and
same method. So, its kind of tedious to do the same example again and again, but you can
practice at your own these examples. So, I am going to write one example for you to
practice.

515
(Refer Slide Time: 05:41)

So, example, find the unit tangent vector, unit principal normal or unit normal and unit
binormal unit will not add let us say I am including everything. So, unit binormal,
𝜋
osculating plane, normal plain and rectifying plane at 𝑡 = 4 for the curve 𝑥=

𝑎 cos 2𝑡, 𝑦 = 𝑎 sin 2𝑡, 𝑧 = 2𝑎 sin 𝑡 alright.

And so and you can also try to calculate find the length of the curve measured from 𝑡 =
𝜋
0 to 𝑡 = 4 . So, this is a very nice example to practice all the things that we have studied

before. So, first of all you can be able to write this x y z as a equation of a curve as 𝑟⃗ =
𝑑𝑟⃗ 𝑑2 𝑟⃗ 𝜋
𝑓⃗(𝑡), a vector equation and then from there you can calculate , and 𝑟⃛ after 𝑡 = 4 .
𝑑𝑡 𝑑𝑡 2

We probably do not need 𝑟⃛, but it is a good to calculate up to 𝑟⃛ for any given equation
because you might need at some point or may not need it and it is good to have it on your
paper.

So, now once we have those equations we can be able to. So, I can put a necessity to
calculate those things. So, I can also write curvature and torsion. So, now, you need those
derivatives. So, calculate unit tangent vector unit normal unit by normal curvature torsion
𝜋
osculating plane normal plane rectifying plane at 𝑡 = 4 . So, now, you need all those

derivatives and once you have those derivatives you can be able to calculate curvature and
torsion by using that 𝑟̇ × 𝑟⃛ and all that.

516
 𝑑𝑟⃗
So, use that formula to calculate the curvature and the torsion and then from we can be
𝑑𝑡

able to calculate our unit tangent vector by dividing it with its magnitude 𝑟̇ from there we
can be able to calculate binormal, normal using that 𝑛̂ = 𝑏̂ × 𝑡̂ or whatever it is. And from
there we will be able to calculate our osculating plane, rectifying plane and normal plane
the way I just showed you and in the previous class and finally, we can be able to calculate
𝜋
𝜋
the length of the curve from 𝑡 = 0 to 𝑡 = 4 by integrating ∫04 |𝑟̇ |𝑑𝑡. So, whatever 𝑟̇ we will

get we will take its magnitude and just integrated with respect to t and then substitute the
𝜋
value of 𝑡 = 0 and 𝑡 = 4 .

So, that is how you will be able to calculate all these terms here and it is a nice example to
practice for you so keep doing that. And now, we will move on to our next aspect of vector
calculus which is basically application of vector calculus in mechanics. So, that was the
one of the topics that I suggested in the syllabus. So, now, we will start with that.

(Refer Slide Time: 10:19)

So, our next topic is application. So, why do we have to study the application of vector
calculus in mechanics? The thing is whenever a particle moves it has a speed, but more
precisely it has a velocity because whenever it is moving and not only that it has a speed
it also has a direction that in which direction it is moving.

So, when once we talk about the direction that is when the vector quantity comes in. So, a
vector quantity is something that has magnitude as well as the direction, but a scalar

517
quantity has only the magnitude. So, when we talk about a motion of a particular or motion
of a body we always talk about its direction and when we talk about its direction that is
when the concept of vectors comes in. Now, that direction, so, a particle moving has a
speed and once we associate that speed with a direction then that is when we get a velocity.
So, velocity is something which has a direction as well as a magnitude. So, you see how
motion or a basically mechanics is connected to vectors.

So, we will start with those topics now. So, first of all what do we mean by velocity of a
particle? So, definition and how do we express it in terms of vectors alright. So, the
velocity of a particle relative to a suitable frame of reference is the time rate of change of
the position vector 𝑟⃗ of the particle relative to the given frame of reference.

So, that means, the velocity of a particle when a particle is moving is basically the time
rate of change of the position vector so; that means, if we talk about position vector we are
talking about actually the direction in a way and how the direction of the particle is
changing over time. So, that time rate of change of the position vector is actually called as
the velocity alright.

So, if we consider let us say if we have a curve this 𝑟⃗ = 𝑓⃗(𝑡) and suppose this is my
equation and that is origin that is the point P. So, this is our position vector 𝑟⃗ and if it is
moving along this curve with respect to t or with respect to time then that is actually the
velocity of that particle. So, we have the vector equation and that can be expressed as. So,
with the difference with the difference to a frame if a particle P has at any instance the
position vector as OP equals to 𝑟⃗ and if during the time interval ∆𝑡 the increment is ∆𝑟.

So, let us say the increment here is 𝑟⃗ + ∆𝑟 for any time 𝑡 + ∆𝑡. So, this position vector is
∆𝑟
let us say Q is 𝑟⃗ + ∆𝑟 . So, at any time interval ∆𝑡 the increment in ∆𝑟 then is the
∆𝑡
∆𝑟 𝑑𝑟⃗
average velocity of P relative to O during the interval ∆𝑡. And hence lim = = 𝑟̇ is
∆𝑡→0 ∆𝑡 𝑑𝑡

called as the velocity of the particle P or of the point P at that instant it is also called as
instantaneous velocity of P.

So, basically if we have a small increment for any time interval ∆𝑡 we have a small
∆𝑟
increment ∆𝑟 then basically is the average velocity. So, how much it has increased is
∆𝑡
∆𝑟
basically it is increased by ∆𝑟 and is actually the average velocity of P relative to this
∆𝑡

518
 𝑑𝑟⃗
origin O during the interval ∆𝑡. And therefore, we can write in terms of limit and this 𝑑𝑡 or

𝑟̇ is actually our velocity, at that point P at an instant t. So, basically if we have a given, if
we have a curve let us say 𝑟⃗ = 𝑓⃗(𝑡).

(Refer Slide Time: 17:28)

So, if we have a curve let us say 𝑟⃗ = 𝑓⃗(𝑡) then of course, P is called as the trajectory. So,
as the direction or the curve along which P is moving is called as the trajectory the locus
of P is basically called as the trajectory. So, the curve or the path it follows so that is
basically the trajectory of t and if we have a given curve 𝑟⃗ = 𝑓⃗(𝑡) as the equation of the

= 𝑓⃗̇ = 𝑉
𝑑𝑟⃗
curve. Then basically what we are doing is we are writing velocity ⃗⃗ and that
𝑑𝑡

is how we calculate the velocity of the given point P on a curve 𝑟⃗ = 𝑓⃗(𝑡) alright.

𝑑𝑟⃗ 𝑑𝑟⃗ 𝑑𝑠 𝑑𝑠
⃗⃗ = 𝑟̇⃗ =
Now, since we can write; 𝑉 = . = 𝑡̂ 𝑑𝑡 . So, s is the arc length and rate of
𝑑𝑡 𝑑𝑠 𝑑𝑡

change of arc length is not relevant to the direction. So, in which direction the arc length
is changing is not relevant. So, basically the rate of change of arc length is given by this
𝑑𝑠 𝑑𝑟⃗
and ⃗⃗ | then it is basically
is basically our unit tangent vector. So, you see if I take |𝑉
𝑑𝑡 𝑑𝑠
𝑑𝑠 𝑑𝑠
|𝑑𝑡 |. So, 𝑑𝑡 is rate of change of arc length.

𝑑𝑠
So, it will always be positive. So, we have 𝑑𝑡 and this positive thing can be denoted by V

⃗⃗ |
and this V is actually the speed. So, as we were saying speed is a scalar quantity. So, |𝑉

519
or the magnitude of V will actually give us the speed of the particle P along that curve or
𝑑𝑠
the speed of the point P along that curve and that can also be given by 𝑑𝑡 .

So, differentiation of the arc length with respect to t will give you the speed that how fast
that point P is moving along the arc length of the curve. So, speed of the particle or speed
of the point particle or point both are same in this context at the point P alright. So, this is
basically our velocity. So, this is our speed and now we can define our acceleration. So,
after speed we have acceleration.

So, acceleration, it is defined as it is time rate of change of velocity. So, it is basically time
rate of change of velocity and this acceleration vector this. So, first of all rate of change of
𝑟⃗ is called as velocity and now rate of change of velocity is called as acceleration. So, if it
the rate of change of velocity is increasing then in that case it will be acceleration, if it is
decreasing then in that case it will be deceleration.

So, although we use acceleration here in this context, but those are the two how to say
adjectives you use that is acceleration and deceleration. So, this acceleration vector let us
say of the particle P at some instant t is given by. So, basically we have as limit ∆𝑡. So,
∆𝑉 ⃗⃗
𝑑𝑉 𝑑 𝑑𝑟⃗ 𝑑2 𝑟⃗
time rate of change of velocity. So, 𝑎⃗ = lim = = 𝑑𝑡 ( 𝑑𝑡 ) = 𝑑𝑡 2 = 𝑟̈ .
∆𝑡→0 ∆t 𝑑𝑡

or simply 𝑟̈ . So, 𝑟̈ is actually our acceleration and it is also a vector quantity because 𝑟⃗ is
a vector quantity alright.

So, we have 𝑎⃗ = 𝑟̈ and this actually defines our acceleration. So, if we are asked to
calculate the acceleration of a moving particle along a curve 𝑟⃗ = 𝑓⃗(𝑡) all you have to do
just differentiate it twice with respect to the parameter whether it is t or 𝜃 whatever it is
and that will actually give you the acceleration. And if you are given a point let us say 𝑡 =
⃗⃗ =
1 then it will give you the acceleration at that point 𝑡 = 1. Now, in this equation 𝑉
𝑑𝑠
𝑡̂ 𝑑𝑡 = 𝑡̂𝑣.

520
(Refer Slide Time: 23:24)

⃗⃗ = 𝑡̂ 𝑑𝑠 . So, I can be able to write it as 𝑉

So, we have 𝑉 ⃗⃗ = 𝑡̂𝑣 . So, in this case this v is
𝑑𝑡
⃗⃗
⃗⃗ is vector. So, if I differentiate both sides then this will be 𝑑𝑉 = 𝑑𝑡̂ 𝑣 + 𝑡̂ 𝑑𝑣, so,
scalar this 𝑉 𝑑𝑡 𝑑𝑡 𝑑𝑡

this v is scalar.

So, now the unit tangent vector. So, the unit tangent vector 𝑡̂ may be regarded as a function
𝑑𝑡̂ 𝑑𝑡̂ 𝑑𝑠 𝑑𝑡̂
of the arc length s; then our = 𝑑𝑠 𝑑𝑡 right. So, and = 𝜅𝑛̂ from that Serret Frenets
𝑑𝑡 𝑑𝑠
𝑑𝑠 𝑑𝑡̂
⃗⃗ . So, this is = 𝜅𝑛̂𝑉
formula and 𝑑𝑡 = 𝑉 ⃗⃗ and if I substitute all these values here.
𝑑𝑡

So, this will reduce to equation let us say 1 from 1 and 2, we will have 1 and 2 we will
⃗⃗
𝑑𝑉 𝑑𝑡̂ 𝑑𝑡̂
have ⃗⃗ . 𝑎⃗ = 𝜅𝑛̂𝑉 2 +
which is basically our acceleration is equals to 𝑑𝑡 ; 𝑑𝑡 is actually 𝜅𝑛̂𝑉
𝑑𝑡
𝑑𝑣 𝑛̂𝑉 2𝑑𝑣
𝑡̂ 𝑑𝑡 = 𝜌 + 𝑡̂ 𝑑𝑡 .

1
where 𝜌 is the radius of curvature and kappa is just a curvature. So, I have used 𝜅 = 𝜌. So,
1 𝑉2
this is what I have used. So, 𝜅 = 𝜌 is taken here. So, this is basically so; V is our speed
𝜌
𝑑𝑣
by radius of curvature plus 𝑡̂ 𝑑𝑡 and this is our unit normal and.

521
So, what does it say that acceleration vector so what does it say is that acceleration vector.
So, you can see acceleration vector can be written as the combination of 𝑛̂ and 𝑡̂. So, 𝑛̂ and
𝑡̂ means, if we go to our previous definition.

(Refer Slide Time: 27:07)

So, basically the acceleration vector is given as the linear combination of the normal and
the tangent vector; that means, the acceleration vector is perpendicular to 𝑏̂, because if we
take here dot product with 𝑏̂ then in that case 𝑛̂. 𝑏̂ = 0 and 𝑡̂. 𝑏̂ = 0 and on the left hand
side we will have 𝑎⃗. 𝑏̂ = 0.

So, what does that mean? That mean that this vector 𝑎⃗ or the acceleration vector lies in the
osculating plane and that is by taking the dot product with be equals to 0. So, this relation
I can name it as let us say 3 equation a relation number 3. So, sorry relation number 3.

522
(Refer Slide Time: 27:58)

So, the relation 3 shows that vector 𝑎⃗ is in osculating plane because it is given as the
combination of 𝑛̂ and 𝑡̂. So, if we take the dot product then 𝑎⃗. 𝑏̂ = 0 lies in the or is in the
osculating plane.

⃗⃗ tangential component is
And the tangential and normal components of acceleration are 𝑉
𝑑𝑣 𝑉2
where v is a scalar and normal component is where V is the speed and 𝜌 is the radius
𝑑𝑡 𝜌

of curvature.

So, from this relation we can see clearly that the acceleration vector actually lies in the
plane of 𝑡̂ and 𝑛̂ and it is perpendicular to the vector by normal 𝑏̂ and so; that means, the
vector a lies in the osculating plane. And the component for the tangential and the normal
𝑑𝑣 𝑉2
components for 𝑎⃗ are given as 𝑑𝑡 where v is the speed and 𝜌
is the radius of curvature.

So, this is an another a nice formula to remember that the acceleration vector lies in the
osculating plane. So, we will stop here today and in the next class we will derive several
other how to say relations keeping vector calculus in mind and mechanics in mind and we
will see how vector calculus is widely used in mechanics and we will also try to work out
few examples if time permits. So, we will stop here for today and I will look forward to
you in your next class.

Thank you.

523
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture - 51
Angular Momentum, Newton's Law

Hello students. So, upon the last class, we looked into the application of vector calculus in
mechanics basically, so using the vector calculus, you can be able to express things like
velocity acceleration in some vector notation. So, today, we will continue with those things
and we will see there are several other concepts like momentum, angular momentum, areal
velocity or some equations of motion of some particle under some kind of law basically
can also be expressed using the concepts of Vector Calculus.

So, today, we will start with areal velocity and then we will move to momentum and then,
we will try to write the equation of motion of some particles. So, will see how we can do
that. So, to begin with, let me give you a formal definition of the Areal velocity.

(Refer Slide Time: 01:18)

So, when I say areal velocity, it basically mean, so areal velocity. So, the areal velocity
basically the areal, sorry I need to correct the definitions spelling basically, areal yes, a r e
a l velocity.

524
So, the areal velocity of a point P about the point O, the origin of the reference system is
the time rate of change, is the time rate of description or change basically, description of
⃗⃗⃗⃗⃗ . So, what do we mean by it? So, we basically mean
the vector area swept by the line 𝑂𝑃
that suppose we have a line, this is our reference line OX and this is our given curve. And
I said this basically will be our point P. So, this is our vector 𝑟 and this is our point let us
say Q and this is 𝑟 + 𝛿𝑟 and this is our angle 𝜃 and that is basically 𝜃 + 𝛿𝜃. And suppose
here, we have a tangent. So, this is actually N and V.

So, what it means is that the areal velocity of a point P, so this is our point P about the
point O, which is basically the origin for the reference system is the time rate of description
of the vector area. So, this is basically the vector area in this direction. So, this is our 𝑟 this
is 𝑟 + 𝛿𝑟. So, this is the vector area that is being swept out by the vector ⃗⃗⃗⃗⃗
𝑂𝑃 and that is
actually the description, the time rate description of the vector area is actually called as the
areal velocity.

So, how do we write in terms of the vector equation? So, we start with, let us say proof
basically, so a small proof. So, we consider a general case of motion. So, basically this
particle P is moving along this curve; however, we assume that consider a general case of
motion of the point P, when its path is not necessarily a plane curve, alright.

And now, let P and Q be the two positions of the particle, at the time t and 𝑡 + 𝛿𝑡, and let
𝑟 and 𝑟 + 𝛿𝑟 be their position vectors, alright. So, then if this area is ∆𝐴, so if this whole
area is ∆𝐴. So, if I call this area as ∆𝐴, so let me write if ∆𝐴 denotes the area of the triangle.
So, since 𝛿𝑟, is a very small arbitrary increment. So, then in that case this OPQ is actually
forming a triangle. So, it is not actually a curve this PQ. It is basically considered as a line
because this increment is considered to be very small increment of time. So, 𝛿𝑡, 𝛿𝑟 they
are considered in very small quantity. So, basically ∆𝐴 is the area of the triangle OPQ,
right.

1 1
So, we know that the area of a triangle is given by ∆𝐴 = 2 𝑟 × (𝑟 + 𝛿𝑟) = 2 𝑟 × 𝑟 +
1 1
𝑟 × 𝛿𝑟. So, since 𝑟 × 𝑟 = ⃗0, so this is ⃗0 + 2 𝑟 × 𝛿𝑟 alright. And now if I define this, so
2

the rate of change of, so the definition says the time rate of description of. So, the time rate
of this change of this area of this description of the vector area will be basically we have
∆𝐴
to take .
∆𝑡

525
(Refer Slide Time: 08:21)

Because that is what we are denoting the rate of change of the areas by the vector ⃗⃗⃗⃗⃗
𝑂𝑃. And
just for the notation point of view, I think it is a good idea to take it as small delta, just for
the sake of notation.

So, 𝛿𝑠, small delta, so we are basically in delta how to say notation, so delta, delta, this
delta, so I will also take it as small delta, alright. So, it is this basically notation but it
𝛿𝐴 1 𝛿𝑟
makes the whole thing look how to say nicer in a way. So, 𝛿𝑡 = 2 𝑟 × 𝛿𝑡 . Now, if I take

𝛿𝑡 goes to 0. So, proceeding with 𝛿𝑡 going to 0, we will have limit this thing will actually
𝑑𝐴 1 𝑑𝑟
reduce to, we do not have to write the limit, we can write simply as = 2 𝑟 × 𝑑𝑡 . So, that
𝑑𝑡

means, the required therefore, the required areal velocity; therefore, the required areal
velocity which is basically the time rate this 𝑑𝑡 is signifying the time rate of description of
the area swept by the vector ⃗⃗⃗⃗⃗
𝑂𝑃 right.

So, this is basically the time rate of description. So, the areal velocity is basically
1 𝑑𝑟
𝑟 × 𝑑𝑡 and this is the required areal velocity for this point P with whose position vector
2

is given by vector 𝑟 alright. And similarly, we can calculate, so here in my lecture note I
have also calculated the moment basically. So, how do we calculate the moment? So, this
proof is complete up to here. Let me also give you the moment. So, the moment you may
have studied these things in your physics course. So, the moment about O of the velocity
vector say v.

526
So, this is not the areal velocity, this is the usual velocity vector v considered localized
along the tangent line, at P is by definition. So, the moment of the particle about O of the
velocity vector basically. So, the moment of the velocity vector considered localized along
⃗ right. And its
the tangent vector and it is given by. So, by definition it is given by 𝑟 × 𝑉
direction, being at right angles or perpendicular basically right angles to the plane of 𝑟
⃗ , alright. So, it is actually at the right angle to the plane of 𝑟 and 𝑉
and 𝑉 ⃗ . And if we consider
this as the, so this is our N. So, if we consider this, so this as the perpendicular whose
length is small p.

(Refer Slide Time: 13:03)

So, if p is the length of the perpendicular ON, so I can write if small p is the length of the
⃗ is perpendicular
perpendicular, ON to the tangent at P. Then, in that case, since our 𝑟 × 𝑉
⃗ , I can be able to write 𝑟 × 𝑉
to the plane of 𝑟 and 𝑉 ⃗ is parallel to that normal. And that is
⃗ = 𝑝𝑣𝑛̂ because if 𝑟 × 𝑉
basically 𝑟 × 𝑉 ⃗ is perpendicular to the plane of 𝑟 and 𝑉
⃗ , then it
must be parallel to the normal you know way and these are basically the scalars that
signifies that it is basically a normal, alright.

So, here basically small v is the speed of the particle or of the point P, the point or particle
whatever you prefer from the particle P and 𝑛̂ is the unit vector in the direction of normal
⃗ . So that means, basically so; we can write thus, I can
to the plane of 𝑟 and vector 𝑉
⃗ = 𝑝𝑣𝑛̂.
write 𝑟 × 𝑉

527
 1
⃗ ) = 𝑝𝑣𝑛̂. So, this v is speed
So, that means, from here I can write two times, 2 × 2 × (𝑟 × 𝑉

actually. And this shows that, this is nothing but moment of v about O. So, this is actually
our moment of v about O given by this formula and that is actually twice the areal velocity
⃗ 𝑎𝑏𝑜𝑢𝑡 𝑜.
𝑖. 𝑒. 2(𝑎𝑟𝑒𝑎𝑙 𝑣𝑒𝑙𝑜𝑐𝑖𝑡𝑦) = 𝑝𝑣𝑛̂ = 𝑚𝑜𝑚𝑒𝑛𝑡 𝑜𝑓 𝑉

So, this is one of the important formulas of in this motion of a particle on a curve. And
here we can see that with the help of vector calculus we can be able to show that the areal
velocity is half the moment of the velocity of the particle at the point velocity of the particle
P about the point O, about the origin O. So, this is another important result that we can
express using the vector calculus, it would also help if you have some knowledge of
physics from your previous plus two level just to make these things a little bit more clear
alright.

Next is momentum and linear momentum and then we will also write a Newton’s second
law of motion using vector calculus so or vector notation basically.

(Refer Slide Time: 16:50)

So, what do we mean by momentum? So, the definition you must remember by momentum
⃗⃗ = 𝑚𝑉
M; which is a vector of a moving particle at any time, we mean the vector 𝑀 ⃗ where
⃗ is the velocity of the particle right. And in notation or we write
m is the mass and 𝑉
⃗⃗ = 𝑚𝑉
simply 𝑀 ⃗ . So, 𝑚 is the mass of the particle and we mean the vector 𝑉
⃗ where m is
⃗ is the velocity of the particle.
the mass of the particle and 𝑉

528
So, for example, if you have let us say an iron ball and if you just touch it on your hand
and it will not hurt you that much. So, suppose the weight of the iron ball is like 500 grams.
So, if you just touch it with your hand, then it will not hurt your hand, but if you throw it
that ball on your hand and there is a strong possibility that it will hurt your hand. The thing
is that since it got some velocity, the momentum is higher. So, when it is hitting, that is
when you are feeling the glow actually.

So, even though the mass is very huge if the velocity is not higher than in that case, the
momentum will always be small. So, that is why when something hit us or when someone
gets hit by something then in that case that at that point momentum is high basically
because the velocity was high. So, momentum is directly proportional to the velocity or it
depends on the velocity basically. So, higher the velocity is, higher the momentum would
be and m is the mass of course. So, of course, it also depends on mass. So, if your mass is
as big and if the velocity is higher, then the blow will be higher. And if the mass is small
and velocity is higher, then at that time also the low will be higher, but it will not be as
high as when you had a bigger mass. So, mass also plays an important role, but in case of
momentum, I would say that the velocity plays an important roles because velocity can
increase and decrease and based on that for a given mass, the momentum can also increase
or decrease alright.

Now, next we define the moment of a momentum. You probably know all these things,it
might be a recapitulation from your physics class, but I am just writing the advantages of
⃗⃗ be the momentum, of a particle P at any instant and O be
vector calculus here. So, if 𝑀
⃗⃗⃗⃗⃗ , so then the vector 𝑂𝑃
any point, then 𝑂𝑃 ⃗⃗⃗⃗⃗ × 𝑀
⃗⃗ . So, this is not necessarily origin.

So, that this can be any arbitrary point, so I could take Q which or let us let us keep it as
⃗⃗⃗⃗⃗ × 𝑀
O. So, 𝑂𝑃 ⃗⃗ = 𝐻
⃗ because this is a vector this is the vector. So, the cross product will
definitely be a vector is called the moment of momentum or sometimes they are also called
as angular momentum of the particle about, alright. So, this is called the angular
momentum of the particle about O.

Now, and so here basically so here ⃗⃗⃗⃗⃗

𝑂𝑃 is the position vector of the point P about O. So,
⃗⃗⃗⃗⃗ is the position vector and in case and in case O is the origin in case,
with respect to O, 𝑂𝑃

O is the origin, then this can be written as vector 𝑟 × 𝑀 ⃗ = 𝑚𝑟 × 𝑑𝑟. So, 𝑉

⃗⃗ = 𝑟 × 𝑚𝑉 ⃗ can
𝑑𝑡

529
 𝑑𝑟
be written as ⃗⃗ the momentum time
and this is basically our angular momentum. So, 𝑀
𝑑𝑡
𝑑𝑟 𝑑𝑟
cross product with or 𝑚𝑟 × 𝑑𝑡 is basically your angular momentum. So, if I go back to
𝑑𝑡

this formula where I have the areal velocity, so basically areal velocity and angular
velocity. So, I can take two times areal velocity. So, this is basically two times m times
areal velocity. So, that is basically our angular momentum alright.

And, so, the thing and another thing is angular momentum will be different for every point
that we choose this O. If we choose O as a different vector point, then in that case this
⃗⃗⃗⃗⃗ vector will change and then of course, our angular momentum will change. So, these
𝑂𝑃
are the small here and there details which you have to be careful about. So, here if we
choose instead of one a different point, then in that case this angular momentum will
change. So, yeah this is our moment of momentum. The next topic is Newton’s second
law of motion.

(Refer Slide Time: 24:12)

So, you probably heard about there are three laws of motion and when you say three laws
of motion it obviously, means that we are talking about Newton’s three laws of motion
because everything around us they obey Newton’s laws of motion. So, the first one we all
know that it is called law of inertia. So, if a body is in motion, then it will remain in the
motion. If it is in rest, then it will remain in the rest unless it is being how to say compelled
by some external force to change it is state. And the second law is that force is

530
perpendicular to, proportional to acceleration basically and and the third law is that to
every action there is an equal and opposite reaction.

So, these are the three basic laws and which has a huge, I mean which has huge, huge,
huge application everywhere around you. So, anything that is moving it has in one way or
another some application of our direct application of Newton’s law or laws of motion. So,
basically all three of them are important, but today we will just derive or learn the second
law of motion. So, the second law of motion says that the time rate. So, the statement is
the time rate of change of momentum of a particle is proportional to the impressed force
and takes place in the direction in which the force acts, alright. So, if 𝑃⃗ is the impressed
force or if 𝑃⃗ is the applied force, then in that case let us call 𝑃⃗ because if you are applying
a force, then it has a direction. Because if you apply a force and expects the chains to
happen in this direction, that will not happen. So, that, when you apply the force in a certain
direction, the change would also happen in that direction. So, the force is a vector quantity.
So, it is always having a direction.

𝑑𝑀 ⃗⃗
So, that is what 𝑃⃗ is the force which is chosen as a vector and 𝑃⃗ ∝ 𝑑𝑡 . Momentum is a

⃗ . So, I can I can change this 𝑃⃗ =

vector quantity because it involves velocity and 𝑀 = 𝑚𝑉
⃗
𝑑𝑉
𝐾𝑚 𝑑𝑡 = 𝐾𝑚𝑎. So, rate of change of velocity and I can write this as 𝐾𝑚𝑎 where K is the

constant is a scalar constant given constant of proportionality m is the mass and 𝑎 is the
acceleration.

So, next if we choose the unit of force, unit of force has one which produces in a unit mass,
unit acceleration, then our K will be constant. Then, K will be 1. So, of course, it is constant
then K will be 1 and hence we will have 𝑃⃗ = 𝑚𝑎. So, this is basically our Newton’s second
law of motion. So, that means, force is equals to mass into acceleration and this means that
acceleration produced in a motion of a particle of a constant mass has the same direction
as the force producing it.

So, it will happen, the change will happen in the direction of the force and this equation is
called as the equation of motion of a particle. So, this today, so this is an another
application of vector calculus or we saw that how with the help of vector calculus we can
be able to write Newton’s second law of motion. So, these are some of the applications of

531
vector calculus in the applied mathematics or in mechanics. In the next class, we will
continue with that and if time permits, then will move on to our next topic.

So, I thank you for attention today and I will see you in the next class.

532
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture - 52
Example on derivation of equation of motion of particle

Hello students. in the previous class, we were sort of looking at the application of vector
calculus in the mechanics and we try to derive equation of or formula for momentum,
angular momentum, Newton’s second law of motion. And we saw that the force 𝑃⃗ can be
written as m times acceleration where m is the mass of a particle and a P is the applied
force or impressed force and 𝑎 is the acceleration of the particle. So, today, we will try to
derive some of the equations of motion where we see how vector calculus is useful to write
the equation of motion of such particles are moving under some physical law and they will
try to derive their path and things like that.

So, let us start with.

(Refer Slide Time: 01:13)

So, in the previous class, we had this formula 𝑃⃗ = 𝑚𝑎 and this formula was obtained
which is basically called as Newton’s second law of motion. And sometimes, people also
call it as equation of motion of a particle.

533
(Refer Slide Time: 01:28)

Now, let us start with our very first example, example 1. So, motion under gravity, so,
motion under gravity means suppose you drop the particle or you threw a particle, then
after a while, when you throw a particle, then of course, it starts more or less in a straight
line manner, but over the time it follows a parabolic profile. So, be it mainly because the
particle because of the gravity falls on the ground. So, basically the gravity at that point is
acting as a force and it is pulling the particle down towards it.

So, it does not matter what kind of force you have applied it may go far or it may fall right
next to you, it will always follow a parabolic profile. And we will try to derive that
parabolic profile and will first see how we can do that. Suppose, we have a moving particle.
So, suppose if we have, if a moving particle of mass m be subjected to the action of gravity
alone, then the equation of motion, of the particle.

So, equation of motion as you remember 𝑃⃗ = 𝑚𝑎 . So, 𝑎 is the acceleration and 𝑃⃗ is the
𝑑2 𝑟
force and the force is given by 𝑚 𝑑𝑡 2 = −𝑚𝑔𝑘̂ where 𝑘̂ is the unit vector drawn vertically

upwards, alright.

𝑑2 𝑟
So, now, if we cancel m from both sides, then this is basically our = −𝑔𝑘̂. And if I
𝑑𝑡 2
𝑑𝑟
integrate both sides, then in that case this will be ⃗ = −𝑔𝑡𝑘̂ + 𝑏⃗. And if I integrate
=𝑉
𝑑𝑡

again then in that case this will be

534
 𝑡2
So,𝑟 = −𝑔 + 𝑏⃗𝑡 + 𝑐 . So, let me call it as equation number. So, this is our equation
2

number 1, this is our equation number 2, this is our equation number 3. So, I integrated
both sides of these equations and when you integrate, I am assuming you are familiar with
that ordinary differential equations. So, I integrate it with respect to t. So, will get a
constant vector 𝑏⃗ and then I again integrate it and then I got a constant vector 𝑐 . Now, we
have to find the values for 𝑏⃗ and 𝑐.

(Refer Slide Time: 05:42)

⃗ and the initial velocity

So, if, so when 𝑡 = 0, at time 𝑡 = 0 at that time the velocity 𝑉
⃗ . So, I can write
would be same. So, the initial velocity at 𝑡 = 0 is same as the velocity𝑉
⃗ =𝑢
𝑉 ⃗ . So, 𝑢
⃗ is my initial velocity. So, then from 2 we will have, what do we have?
⃗ , this term will be 0 and 𝑏⃗. So, this will be 𝑢
⃗ =𝑢
So, 𝑉 ⃗ = 𝑏⃗. So, if I use this in this equation,
so I can substitute 𝑢
⃗ here. Now, when 𝑡 = 0, now again when 𝑡 = 0, then the vector 𝑟 will
be 0 because when t equals to 0, then the particle is not moving. So, at that time basically
⃗⃗⃗ and then from 3, we will have
the position vector 𝑟 would also be 0. So, when 𝑡 = 0, 𝑟 = 0
0 and here we will have 0, 0. So, 𝑐 is also a ⃗0.

So, 𝑐 = ⃗0, not implication, but we can write that is 𝑐 = ⃗0 . Therefore, from 2, therefore,
1
sorry from 3 actually therefore, from 3 we will have 𝑟 = − 2 𝑔𝑡 2 𝑘̂ + 𝑢
⃗ 𝑡. So, this is the

required equation of the motion of the file. So, this is the required how to say locus of the

535
point P. So, I can write the locus of 𝑟 is a plane curve on the plane determined by the
⃗ and 𝑘̂ and the origin.
vectors 𝑢

⃗ and 𝑘̂ will determine the locus of 𝑟. And if we consider

So, the origin and the vector 𝑢
and so, this is the required equation of the particle P actually moving along the curve and
basically this is the equation of the curve along which the particle is moving. And if we
⃗ and 𝑘̂ as the x axis and y axis
consider, so one particular case, so if we consider 𝑢
respectively, then from this one is equation number 4, then from 4 we find that. So, if we
equate, so here I can substitute basically. So, if it is x axis, then this one will be (𝑢
⃗ , 0)and
if it is y axis, then this will be just 0 and 0 comma 𝑘̂ and then I can equate the coefficients.

(Refer Slide Time: 10:07)

1
⃗ |𝑡 and 𝑦 = − 2 𝑔𝑡 2 . So, eliminating t, so if I substitute for
So, basically we find that 𝑥 = |𝑢

t here, so this will be x by use of eliminating t will be. So, eliminating t would give 𝑦 =
1 𝑔
− 2 |𝑢⃗|2 𝑥 2 . So, this shows that the path is a parabola. So, that means, when you throw this

the stone or a particle like this, then so, if you throw like this, I do not know if you can be
able to see in that video not. So, if you throw like this, then basically your 𝑢
⃗ is x axis and
your 𝑘̂ is let us say y axis. So, then in that case, the particle is actually following a parabolic
path and that is what we were trying to establish.

So, you see using just some simple concepts of vector calculus. We did not even
complicate things, we were able to derive the equation although, I sorry the curve along

536
with the particle P or in that case and a stone P let us say move along that curve, alright.
So, this is one of the interesting applications of vector calculus basically in mechanics.
And let me see in my notes, if I have anything other, anything else interesting to show.

So, next we can derive the harmonic motion. So, let me give you a small example, what
do you mean by harmonic motion.

(Refer Slide Time: 12:09)

So, equation of motion for the harmonic motion, so basically the equation of motion of a
particle subjected to a force towards a fixed point and with magnitude proportional to its
distance, from the fixed point is given by.

So, basically what we have is, we have a let us say a closed curve. In this case, we are
choosing an ellipse alright. This is our origin O and let us say this is the point P and this is
x axis and this is our y axis and the magnitude is |𝑏⃗| and here in this case the magnitude
|𝑎| is |𝑎| and this is basically our mu times our the position vector, alright. So, what it
says is the equation of motion of a particle subjected to a force towards the fixed point.

So, this is our fixed point O and with magnitude towards a fixed point and with magnitude
𝑑2 𝑟
for point proportional to it is distance from that fixed point is given by 𝑑𝑡 2 = −𝜇𝑟 (𝜇 >

0). So, basically since it is, so this vector is the 𝑟. So, since it is subjected towards that
fixed point O; that means, it is subjected towards that fixed point; that means, that the

537
 𝑑2 𝑟
vector is in the opposite direction. So, basically is equals to it should have been ⃗⃗⃗⃗⃗
𝑂𝑃.
𝑑𝑡 2

⃗⃗⃗⃗⃗ and
But since it is subjected in the fixed direction O, then it should be ma basically 𝑃𝑂
⃗⃗⃗⃗⃗ .
that is basically −𝑂𝑃

So, when you change the direction of a vector, then in that case you have to put a minus
sign. So, that is why you have taken −𝜇𝑟 or we have taken −𝑚 𝜇𝑟where 𝑚𝜇 is the constant
basically. So, the general solution of this equation can be written as, so if you solve this
equation, then basically the general solution can be given by. So, if you solve this equation,
then in that case this is basically you can assume 𝑟 = 𝑎 cos(√𝜇𝑡) + 𝑏⃗ sin(√𝜇𝑡). So, if you
substitute this solutions here, then it will satisfy this equation. So, this is basically the
general solution where 𝑎 and 𝑏⃗ are the arbitrary vectors where 𝑎 and 𝑏⃗ are the arbitrary
vectors.

So, this equation is similar to what we remember from our ordinary differential equation.
𝑑2 𝑦
So, + 𝜇𝑦 = 0 . So, if you solve this equation, then we usually obtain a 𝑦 =
𝑑𝑥 2

𝑎 cos(√𝜇𝑥) + 𝑏 sin(√𝜇𝑥). So, it is the same thing. You can write 𝑟 = 𝑥𝑖̂ + 𝑦𝑗̂ and then
you equate the coefficients of 𝑖̂ and 𝑗̂ and then you solve the individual equations and then
you sum them to get the vector 𝑟. And then, you take (𝑎1 + 𝑎2 ) as a new vector and
(𝑏1 + 𝑏2 ) as the second vector and that is basically how you obtain this general solution.
It is not complicated but it is a little bit lengthy. So, I am pretty sure you can be able to do
that.

Now, this is our arbitrary vectors. And from this equation, basically if we differentiate then
𝑑𝑟
we will obtain ⃗ = −√𝜇𝑎 sin(√𝜇𝑡) + 𝑏⃗√𝜇 cos(√𝜇𝑡), alright. So, these are the
=𝑉
𝑑𝑡

formulas.

538
(Refer Slide Time: 18:08)

And if I choose at 𝑡 = 0, so when 𝑡 = 0 , then in that case when sorry. So, when 𝑡 = 0 then
⃗ is if I substitute 𝑡 = 0, then sin 0 = 0 and
basically our 𝑟 = 𝑎. So, in that case 𝑟 = 𝑎 and 𝑉
⃗ = √𝜇𝑏⃗. so, if the initial position and the velocity is
then this one will be cos 0 = 1. So,𝑉
given. we can fix up 𝑎 and 𝑏⃗.

So, if we have the velocity of the particle and the initial position, let us say that vector 𝑟,
then in that case we can fix up these constant 𝑎 and 𝑏⃗. And if the directions of of 𝑎 and 𝑏⃗
are taken as x axis and y axis then, so if they are taken as x axis and y axis, then we can
equate the first of all the coefficients from here and then it is 𝑥 = |𝑎| cos(√𝜇𝑡) & 𝑦 =
2
𝑥
|𝑏⃗| sin(√𝜇𝑡), alright. And this will be basically if I divide, then this will be basically 𝑎2 +
𝑦2
= 1 where 𝑎 = |𝑎| & 𝑏 = |𝑏⃗|. So, if you choose vector 𝑎 and vector 𝑏⃗, those are
𝑏2

basically our reference vector initially.

So, if you choose those reference vector as x axis and y axis, then basically the curve along
which the particle P is moving for which it is for which the it is subjected to a force directed
toward the origin, the path which it follows is basically and ellipse. So, this is the path or
this is the curve along which the particle is moving. This is somehow also related to our
planetary motion that it follows an elliptic path and the force which is basically in this case
the gravity or acting between sun and the planets. It is actually directed towards that fixed

539
point and therefore, the path along which this planet is moving is basically an ellipse or
the curve is basically an ellipse.

So, a and b are the length of the semi major axis and semi minor axis. So, this is an
important example which I chose to show you all. And do we have some other interesting
examples? Of course, I mean, so here I have a lot of examples actually, but if we try to
cover all of them, then they probably run out of time to cover the other parts of this labours.

So, let me give you an another example and I will leave the derivation of the equation for
optima of the curve up to the students. So, it is you have to derive the formula for that,
alright.

(Refer Slide Time: 22:06)

So, the third problem is motion under the gravity subjected to the resistance proportional
𝑑2 𝑟
to velocity. So, first of all motion under gravity we know how to write 𝑚 𝑑𝑡 2 = −𝑚𝑔𝑘̂ .

Now, the thing is that we have a resistance. So, that means when resistance against
velocity. So that means, the particle cannot move freely. It got some kind of resistance.
So, that a distance is actually a force which we have to subtract. So, it is a quantity or it is
a term that we have to subtract from the right hand side because now, you have got
𝑑2 𝑟
resistance. And since it is proportional to velocity, we can write it as 𝑚 𝑑𝑡 2 = −𝑚𝑔𝑘̂ −
𝑑𝑟
𝑚𝜇 𝑑𝑡 .

540
So, that is the velocity m is the mass and 𝑚𝜇 is some constant of proportionality. So, 𝑚𝜇
is the constant of proportionality alright. So, if you cancel m from both sides, then this is
2
𝑑 𝑟 𝑑𝑟
basically 𝑑𝑡 2 + 𝜇 𝑑𝑡 + 𝑔𝑘̂ = 0. So, this is your required vector equation or ordinary vector

differential equations I would say. Although there is no such thing, you can call it simply
a differential equation or let us say ordinary differential equation in terms of vectors and
from here you have to solve this vector 𝑟.

So, if we saw how to say write these equations as x, y, z and then if we try to solve it, then
we will obtain 𝑟 is equals to ultimately. So, here I have small answers. So, here we will
⃗
𝑢 𝑡𝑒 𝜇𝑡 𝑒 𝜇𝑡
obtain 𝑟 = 𝑒 −𝜇𝑡 [𝜇 𝑒 𝜇𝑡 − 𝑔𝑘̂ ( − )+𝑐
𝜇 𝜇2

So, 𝑐 is a vector which we have to determine. So, this is the required equation of the part
where we have to determine this constant 𝑐 . And this is another example where we have
used the vector calculus to derive the equation of the curve along which the particle is
moving.

So, bit like this, you can have several examples from vector calculus where from
mechanics where you will have the application of vector calculus. So, for example, here I
have the inverse law of attraction, then you can also write equations such as a planetary
motions, speed of a particle on any orbit, things like that. So, it is not just limited to these
two or three examples. There are like I mean thousands of examples where you can apply
the concepts of vector calculus all with the help of vector calculus, you can be able to write
their equation of motion and just try to solve to know the curve along which the particle is
moving, use some initial conditions things like that.

So, will probably stop this chapter here because I think I have covered enough examples
and gave you how to say enough idea where we use the vector calculus in applied
mechanics. So, we will move on to our next topic which is a vector integration like line
integral, surface integral and volume integral because we also have only 6 or 8 lectures
left. So, I will try to include some examples in your assignment sheet. And you can solve
them and you are also, I advise you to look into some vector calculus books where they
are addressing some problems from mechanics and practice them and I am pretty sure you
will be able to have a same learn some new things from there and it would not be any
problem for you.

541
So, thank you for attention for today and I look forward to you in your next class.

542
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 53
Line Integral

Hello students. So, in the previous class, we concluded the part where we saw how vector
calculus can be applicable to mechanics or applied mathematics and there we derived
several equations of motion for particles under certain laws of physics and we also derived
formulas for momentum, angular momentum and things like that.

So, now we are at a stage that we can move to the Integral part of the Vector Calculus
which is basically the line integral surface integral and volume integral and they basically
include Greens theorem, Gauss theorem and Stokes theorem. So, we will also learn about
these three important theorems of vector calculus.

So, today we are going to start with line integral. So, basically in vector calculus, any
integral along a curve is called as a line integral in a nutshell actually.

(Refer Slide Time: 01:20)

So, to start with, we will start with line integrals. So, any integral which is to be evaluated
along a curve is called a line integral.

543
So, basically we know that equation of a curve. So, suppose the equation of the curve is
given by 𝑟⃗(𝑡) = 𝑓⃗(𝑡) = 𝑥(𝑡)𝑖̂ + 𝑦(𝑡)𝑗̂ + 𝑧(𝑡)𝑘̂ be the given equation of the curve. And
suppose we are integrating, be the given equation of the curve joining two points. So, two
points which are let us say A at 𝑡 = 𝑡1 and the second point is B at 𝑡 = 𝑡2 . So, since these
are obvious, I am not drawing any figure alright. And suppose, we have a 𝐹⃗ so we have a
𝐹⃗ which is a function of x, y and z and it is given as 𝐹⃗ (𝑥, 𝑦, 𝑧) = 𝐹1 (𝑥, 𝑦, 𝑧)𝑖̂ +
𝐹2 (𝑥, 𝑦, 𝑧)𝑗̂ + 𝐹3 (𝑥, 𝑦, 𝑧)𝑘̂. So, these are all functions of x, y and z alright.

Now, we are evaluating this function. So, suppose this is a function is a vector function,
defined and continuous along c. And if small s denotes the arc length, of the curve C, then
𝑑𝑟⃗
= 𝑡̂ as we know is the unit tangent vector to the curve C at the point P whose position
𝑑𝑠
𝑑𝑟⃗
vector is 𝑟⃗, at a point P. And, then the component of F along the tangent is 𝐹⃗ . . So, this
𝑑𝑠

is the component and the integral basically and the integral of this component of 𝑑𝐹⃗ along
𝐵 𝑑𝑟⃗
C from A to B is given by, so basically we integrate ∫𝐴 [𝐹⃗ . 𝑑𝑠 ]𝑑𝑠.

(Refer Slide Time: 05:27)

So, basically this here times ds, so we are integrating along the arc length and this can be
𝐵
written as integral ∫𝐴 𝐹⃗ . 𝑑𝑟⃗. And if we want to write in terms of t, then we can be able to
𝑡=𝑡 𝐵 𝑑𝑟⃗
write this as integral ∫𝑡=𝑡 2 ∫𝐴 [𝐹⃗ . 𝑑𝑡 ]𝑑𝑡. So, basically the point A is attained at the point
1

𝑡 = 𝑡1 and the point B is attained at 𝑡 = 𝑡2 .

544
So, we divide both sides by dt. And so, we divide the sorry, not both sides we divide this
𝑑𝑟⃗ 𝑑𝑟⃗
and we multiply it by dt and then that is basically your integral this 𝐹⃗ . 𝑑𝑡 . We can also
𝑑𝑡
𝑡=𝑡 𝑑𝑥 𝑑𝑦 𝑑𝑧
write it in the Cartesian form so, this one will be ∫𝑡=𝑡 2 [𝐹1 𝑑𝑡 + 𝐹2 𝑑𝑡 + 𝐹3 𝑑𝑡 ]𝑑𝑡 . So, this
1

𝐵 𝑑𝑟⃗
is the required integral of this function ∫𝐴 𝐹⃗ . 𝑑𝑠 along from the point A to B.

So, in this case, we are actually integrating along a curve. So, C is the given curve and
which can be smoother, which can be piecewise smooth. So, when it is piecewise smooth,
then we integrate along the one piece, then we integrate along the second piece. So, like
in part we can do the integration. So, this is what we mean by a line integral along a curve
and when we talk about, so there is a small definition here.

So, in the line integral we can also talk about circulation. So, what do we mean by
circulation? So, if C is a simple closed curve that is a curve which does not intersect, itself
anywhere. So, there is no looping there. So, it is not intersecting itself anywhere, then the
.
tangent line; ∮𝐶 𝐹⃗ . 𝑑𝑟⃗ is called the circulation of 𝐹⃗ about C and it is often denoted by this
. .
symbol and we write it as ∮𝐶 𝐹⃗ . 𝑑𝑟⃗ = ∮𝐶(𝐹1 𝑑𝑥 + 𝐹2 𝑑𝑦 + 𝐹3 𝑑𝑧) . So, this is the required
definition of this circulation of F around the curve C, alright.

So, now we will try to solve some examples on the line integral because line integral makes
more sense when you actually try solving some examples than learning about the theory.
So, let us start with our first example.

545
(Refer Slide Time: 09:59)

So, when we solve the examples, the concept will become even more clear. So, evaluate,
.
the first example is evaluate integral over the curve ∫𝐶 𝐹⃗ . 𝑑𝑟⃗ where our 𝐹⃗ (𝑥, 𝑦) is basically.

So, we are in 2 D. So, 𝐹⃗ (𝑥, 𝑦) = 𝑥 2 𝑖̂ + 𝑦 3 𝑗̂ where and sorry not where and C is the arc of
the parabola, 𝑦 = 𝑥 2 , y plane from (0,0) to (1,1).

So, basically here we have a vector function 𝐹⃗ and we need to calculate the line integral
of this function 𝐹⃗ along the curve C, where C is the arc of the parabola from (0,0) to (1,1).
Now, here in this case and so, basically if we want to draw the figure, then this is our x,
this is our y and that is our parabola touching the origin and we have to calculate the line
integral along (0,0) and then (1,1) so along this arc. So, let us say O, A, So, along this arc
we have to evaluate the line integral for this function F.

So, now what we do? We form a parametric equation. So, that means, we involve a
parameter t and we find the range for that parameter t which will satisfy this parabolic
equation. So, let 𝑥 = 𝑡, 𝑦 = 𝑡 2 and if 𝑟⃗ is the position vector, of any point on C, then
our 𝑟⃗(𝑡) = 𝑥(𝑡)𝑖̂ + 𝑦(𝑡)𝑗̂ = 𝑡𝑖̂ + 𝑡 2 𝑗̂. And t will be, so when x is 0, t is 0. When x is 1, t is
1. Similarly, when y is 0, t is 0. When y is 1, t is 1. So, basically 0 ≤ 𝑡 ≤ 1, alright.

𝑑𝑟⃗ 𝑑𝑟⃗
Now, we will calculate from here, so our = 𝑖̂ + 2𝑡𝑗̂, alright. So, we have got our
𝑑𝑡 𝑑𝑡
𝑑𝑟⃗ . 1 𝑑𝑟⃗
. Thus integral over the curve C ∫𝑐 𝐹⃗ . 𝑑𝑟⃗ = ∫𝑡=0(𝑡 2 𝑖̂ + 𝑡 6 𝑗̂). 𝑑𝑡 𝑑𝑡 all right.
𝑑𝑡

546
 𝑑𝑟⃗ 1 𝑑𝑟⃗ 1
So, I will make it 𝑑𝑡 . So, then in that case, this will be ∫𝑡=0(𝑡 2 𝑖̂ + 𝑡 6 𝑗̂). 𝑑𝑡 𝑑𝑡 = ∫𝑡=0( 𝑡 2 𝑖̂ +
1 𝑡3 𝑡8 1 1 7
𝑡 6 𝑗̂). (𝑖̂ + 2𝑡𝑗̂)𝑑𝑡 = ∫𝑡=0(𝑡 2 + 2𝑡 7 )𝑑𝑡 = [ 3 + 2 8 ]1𝑡=0 = 3 + 4 = 12

(Refer Slide Time: 14:35)

so, this is the required answer.

So, when we integrate the given vector function 𝐹⃗ along this parabola from the point
(0,0) to (1,1), that will be our required answer. So, you see when, we are actually working
out the example we have you get to see what I actually mean by this line integral. So, it is
not always a line, I mean it is not something like a straight line. So, when we say a line
integral; that means, we are integrating along some curve which can be smooth, which can
be piecewise smooth, which can a simple closed curve things like that. So, these are the
certain properties our curve can have and based on that we can calculate the line integral.
We will work out few more examples just to make the concept a bit more clear.

So, now let us consider an another example. So, if 𝐹⃗ (𝑥, 𝑦) = 3𝑥𝑦𝑖̂ − 𝑦 2 𝑗̂, then evaluate
.
line integral ∫𝑐 𝐹⃗ . 𝑑𝑟⃗ where c is the curve in the looks y plane, 𝑦 = 2𝑥 2 from (0,0) to (1,1).
So, this example is also pretty much similar to the previous example. So, here the given
equation is again a parabola. So, we have to assume 𝑥 = 𝑡, 𝑦 = 2𝑡 2 . And at both of these
two points, the equation will satisfy. Therefore, t will run from 0 to 1 I and from there just
integrating like before will give you the required answer. So, this example is exactly like
the previous example.

547
(Refer Slide Time: 17:39)

So, I leave this example up to the students for them to practice. We will now move on to
our next example which is this one, example 3.

.
So, the example 3 is evaluate ∫𝑐(𝑥𝑑𝑦 − 𝑦𝑑𝑥) around the circle 𝑥 2 + 𝑦 2 = 1. So, here the
given equation of the curve c is basically a circle, 𝑥 2 + 𝑦 2 = 1 . So, this circle is actually
a simple closed curve because it is not forming any kind of loop. So, it is not intersecting
itself. It is actually a simple closed curve. So, this is nothing but the circulation of the
function 𝐹⃗ . It is also very easy to get the function 𝐹⃗ from here, let us see. So, the given
equation of the circle is 𝑥 2 + 𝑦 2 = 1. So, from here, from the equation of the curve, we
have x equals to so, basically the radius of the circle.

So, radius is 1 obviously. So, we have 𝑥 = cos 𝑡 , 𝑦 = sin 𝑡 where 0 ≤ 𝑡 ≤ 2𝜋 because we

have a closed circle. So, our t will be running from 0 to 2𝜋 and this function here 𝑥𝑑𝑦 −
𝑑𝑟⃗ 𝑑𝑟⃗
𝑦𝑑𝑥 = (−𝑦𝑖̂ + 𝑥𝑗̂). (𝑑𝑥𝑖̂ + 𝑑𝑦𝑗̂) = 𝐹⃗ . 𝑑𝑟⃗ = 𝐹⃗ . . So, this can be calculated very
𝑑𝑡 𝑑𝑡
𝑑𝑟⃗ 𝑑𝑥 𝑑𝑦 𝑑𝑟⃗
easily. So, is nothing but ( 𝑑𝑡 𝑖̂ + 𝑑𝑡 𝑗̂) which is = − sin 𝑡 𝑖̂ + cos 𝑡 𝑗̂. So, from here
𝑑𝑡 𝑑𝑡

what we will obtain is, so the equation of the curve can be written as or can be given by 𝑟⃗ =
𝑑𝑟⃗
𝑥(𝑡)𝑖̂ + 𝑦(𝑡)𝑗̂ , right. And now, we will take 𝑑𝑡 = − sin 𝑡 𝑖̂ + cos 𝑡 𝑗̂, alright.

548
(Refer Slide Time: 21:05)

So, now I will do the integration. So, now, we have integration over the curve
. . 𝑑𝑟⃗
∫𝑐(𝑥𝑑𝑦 − 𝑦𝑑𝑥) which can also be written as integral over the curve ∫𝑐 𝐹⃗ . 𝑑𝑡 . So, what is
2𝜋 2𝜋
our 𝐹⃗ ? ∫𝑡=0(− sin 𝑡 𝑖̂ + cos 𝑡 𝑗̂). (− sin 𝑡 𝑖̂ + cos 𝑡 𝑗̂)𝑑𝑡 = ∫0 𝑑𝑡 = 2𝜋. So, we will take
the dot product and that will result into 𝑠𝑖𝑛2 𝑡 + 𝑐𝑜𝑠 2 𝑡 = 1. So, ultimately the whole thing
will be 1.

And therefore, we will have integration of dt which is something like this right. And when
we integrate, then we substitute the value and then this is basically 2𝜋. So, the circulation
of the vector function 𝐹⃗ which is given by this way is actually equals to 2𝜋 along that
circle C.

So, sometimes you are not given the function 𝐹⃗ and there you have to actually obtain the
vector function 𝐹⃗ . So, it is also a little bit small trick. It is not actually complicated it is
very easy to see what can be our function 𝐹⃗ . So, just have a look at the given integral and
from there, you can be able to evaluate this will be my vector function 𝐹⃗ which I need to
𝑑𝑟⃗
take dot product with and from the given equation of the curve where you have to
𝑑𝑡

calculate the integral, you can be able to find out the equation of the curve in vector form.
𝑑𝑟⃗
And then, you calculate and then the rest of the things are same as in the example one.
𝑑𝑡

549
So, like we did in this example. So, we calculated the circulation of the function 𝐹⃗ which
was not given. So, we calculated the vector function 𝐹⃗ along the given curve c alright.

So, we can have examples similar to this, many examples similar to this actually. So, I am
just looking into my lecture note which one to consider. So, now what we will do, let us
consider and this example. So, here in this example, example 4 I believe. So, evaluate
.
∫𝑐 𝐹⃗ . 𝑑𝑟⃗ where 𝐹⃗ (𝑥, 𝑦, 𝑧) = 𝑥𝑦𝑖̂ + 𝑦𝑧𝑗̂ + 𝑧𝑥𝑘̂ and the given curve is 𝑟⃗ = 𝑡𝑖̂ + 𝑡 2 𝑗̂ +

𝑡 3 𝑘̂ where −1 ≤ 𝑡 ≤ 1. Now, this is not a very complicated example. So, from the given
equation, of the curve which is this one, we can be able to obtain our 𝑥(𝑡) = 𝑡, 𝑦 = 𝑡 2 , 𝑧 =
𝑡3.

𝑑𝑟⃗
And then, we substitute 𝑥(𝑡), 𝑦(𝑡) & 𝑧(𝑡) there basically in this equation and we do 𝑑𝑡 and
𝑑𝑟⃗ 𝑑𝑟⃗
then we calculate 𝑑𝑡 from here and then we substitute the value of 𝑑𝑡 and we integrate from

−1 to +1. So, here in this example, it is just that the equation I mean how to say in the
Cartesian form 𝑥(𝑡), 𝑦(𝑡) & 𝑧(𝑡) the equation of the curve is not given, but the vector
equation of the curve is given. So, from there extracting 𝑥(𝑡), 𝑦(𝑡), 𝑧(𝑡) is relatively
simple.

So, here, we see that in this example, you can actually be able to write 𝑥(𝑡), 𝑦(𝑡) +
𝑧(𝑡) although they are not given. So, you see these examples, I mean one thing is not given,
then you can be able to obtain the other thing. Similarly, the other thing is given, then you
can be able to obtain the first thing. So, it is all about doing practice and getting used to
the examples of this type they are not complicated they are just playing with some
parameters. That is, pretty much it and from here on onwards, you can be able to solve this
𝑑𝑟⃗
example very easily; just substitute these values here, do the from here and then
𝑑𝑡

integrate from minus 1 to plus 1 and then everything is like piece of cake alright. So, this
is also very simple, yet how to say a nice example.

550
(Refer Slide Time: 26:20)

Next, I am going to consider an example which is I personally like. So, here we have to
.
evaluate like before ∫𝐶 𝐹⃗ . 𝑑𝑟⃗ where our given vector function 𝐹⃗ (𝑥, 𝑦) = (𝑥 2 + 𝑦 2 )𝑖̂ −
2𝑥𝑦𝑗̂ where C is the curve, where C is the rectangle in the x, y plane;𝑥 = 𝑎, y is equal to
0, 𝑥 = 𝑎, x, 𝑦 = 𝑏 and x equals to 0.

So, the solution, so here as you can see we have a rectangle so, obviously, it is not a smooth
curve because then you have four corners here. So, if I draw this curve. So, that is our x
axis, this is origin, this is y axis. So, this is my 𝑥 = 0 line, this is 𝑥 = 𝑎 and that is 𝑦 =
0 and that is 𝑦 = 𝑏. So, I can write it as a, b, c. So, this is our curve right not the middle
portion the curve this one all right. So, now, it depends on us which direction are we
following? Are we calculating in this direction or are we calculating in this direction? So,
that is one thing.

Now, the second thing is that whatever we get along this direction, we just have to take
the reverse direction actually when we are on the upper side, alright. So, this is a very
interesting sample and we will actually try to solve this in our next class because it will
require some time to see the interesting part of this example. So, I will stop here for today
and in the next class we will continue with this example. And I thank you for your attention
and I look forward to your next class.

551
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 54
Surface integral

Hello students. So, in the previous class, we started with the integral aspects of Vector
Calculus and we started with line integral actually. So, I left off the previous lecture at a
very interesting example in my opinion actually. So, today we will continue with the
similar example.

(Refer Slide Time: 00:37)

So, we were given a vector function and we had to find out the line integral
.
∫𝐶 𝐹⃗ . 𝑑𝑟⃗ 𝑤ℎ𝑒𝑟𝑒 𝐹⃗ (𝑥, 𝑦) = (𝑥 2 + 𝑦 2 )𝑖̂ − 2𝑥𝑦𝑗̂ along this rectangle 𝑦 = 0, 𝑥 = 𝑎, 𝑦 =
𝑏 & 𝑥 = 0. So, we have assumed that the orientation is in this way. So, OA, AB and then
CO or yes CO. So, now let me write these points (𝑎, 0) and then this is (𝑎, 𝑏) and then
this is (0, 𝑏) alright. So, now, on OA x is varying, but y is 0 because the equation of the
line of the x axis is y equals to 0. So, on OA our y is 0, and therefore, 𝑑𝑦 = 0 and x varies
from 0 to a right;

So, 0 to a that is how x is varying. Now on AB our x is a and therefore, dx is 0 because x

is constant. So, dx is 0 and y varies from 0 to b right ok. Now, on BC we have 𝑦 = 𝑏 and

552
therefore, dy will be 0 and x varies from a to 0. And on CO our CO x is 0 so, basically dx
is 0 and y varies from b to 0 right. So, basically we have four smooth curves so, and along
all these four curves we have these four criterias.

So, now instead of calculating the line integral of the vector function 𝐹⃗ along this curve C
. . .
so, along this curve C, the line integral along the curve ∫𝐶 𝐹⃗ . 𝑑𝑟⃗ = ∫𝐶 𝐹⃗ . 𝑑𝑟⃗ + ∫𝐶 𝐹⃗ . 𝑑𝑟⃗ +
1 2
. .
∫𝐶 𝐹⃗ . 𝑑𝑟⃗ + ∫𝐶 𝐹⃗ . 𝑑𝑟⃗will be same because our curve C is actually the union of 𝐶1 , 𝐶2 , 𝐶3 , 𝐶4 .
3 4

So, here we can write union of 𝐶1 , 𝐶2 , 𝐶3 , 𝐶4 and from some of the properties from integral
calculus or Riemann integral this union can be transferred into a summation. So, it can be
changed into a summation. So, basically integral over the union of the domains is equals
to the sum of the sub integrals over each of these domains alright. So, that is what we are
doing here.

So, integral over 𝐶1 , 𝐶2 , 𝐶3 , 𝐶4 if you sum them then that will be the integral over the curve
C. So, that is the formula which we have used here and we can call this line as our curve
𝐶1 , we can call this line as our curve 𝐶2 , we can call this line as our curve 𝐶3 and we can
call this line as our curve 𝐶4 and we have to now evaluate these integrals separately. So,
what is the value of the integral along 𝐶1 ?

(Refer Slide Time: 04:27)

553
 . 𝑎 𝑎 𝑥 3 𝑎 3
So,∫𝐶 𝐹⃗ . 𝑑𝑟⃗ = ∫𝑥=0(𝑥 2 𝑖̂ + 0𝑗̂). (𝑑𝑥𝑖̂ + 0𝑗̂) = ∫𝑥=0 𝑥 2 𝑑𝑥 = [ 3 ]𝑎0 = 3 ; I believe yes. And
1

. 𝑏
similarly we can calculate the integral ∫𝐶 𝐹⃗ . 𝑑𝑟⃗ = ∫𝑦=0[(𝑎2 + 𝑦 2 )𝑖̂ − 2𝑎𝑦𝑗̂]. (0𝑖̂ + 𝑑𝑦𝑗̂) =
2

𝑏
−2𝑎 ∫𝑦=0 𝑦𝑑𝑦 = −𝑎𝑏 2 .

So, this is ultimately −𝑎𝑏 2 . And then similarly if we integrates along 𝐶3

. 0 𝑎
∫ 𝐹⃗ . 𝑑𝑟⃗ = ∫ [(𝑥 + 𝑏 ) 𝑖̂ − 2𝑥𝑏𝑗̂]. [𝑑𝑥𝑖̂ + 0𝑗̂] = − ∫ ( 𝑥 2 + 𝑏 2 )𝑑𝑥
2 2
𝐶3 𝑥=𝑎 𝑥=0

So, just do those formula changes. So, what we will get is basically this thing. So, when
we integrate, then this will ultimately give us or what we can do? We first change the
formula. So, this will be x running from 0 to a and then we write the entire dot product.

and similarly we integrate. the function along the curve 𝐶4 and then you sum them and
ultimately we will obtain the answer 𝐼 = −2𝑎𝑏 2 and this is the required answer. So, here
in this case instead of having one smooth curve, we basically had how to say a union of
four smooth curves and those four smooth curves actually needed to be evaluated. So,
basically the line integral along those four smooth curves needs to be evaluated separately.
We cannot do C in whole we have to do like 𝐶1 then integral on 𝐶2 , then integral on 𝐶3 ,
then integral on 𝐶4 . So, sometimes you might come across examples where the curve is
actually composed of a parabola and a straight line.

So, you basically have to divide your integral into 2 sub integrals. So, you first have to
integrate along that parabola and then integrate along that straight line alright. So, that is
something I wanted to show you. So, of course, there are examples like that and due to
time constraints we have to finish this course on time. So, we will move on to our next
topic which is basically surface integral and greens theorem.

554
(Refer Slide Time: 10:21)

So, basically what do we mean by surface integral? So, let us assume that this is our surface
S in three dimensional space. So, I call it as S and any integral so, it is like line integral
definition. So, any integral which is to be evaluated over a surface say S is called a surface
integral. So, basically what do we mean by is there is a very nice theory going on at the
back, but we generally write it as in this way.

So, we are doing the surface integral, we usually denoted as surface integral
. .
∫𝑆 𝐹⃗ . 𝑛̂𝑑𝑠 which can also be written as surface integral ∫𝑆 𝐹⃗ . 𝑛̂𝑑𝑠 . So, here what we actually

mean is that this 𝐹⃗ . 𝑛̂ is the component of 𝐹⃗ along the normal 𝑛̂ and this ds. So, on the
surface of this let us say this surface S, we basically have this is our vector and this is our
point P. This is our vector 𝐹⃗ and this is our vector or the normal 𝑛̂. So, this 𝐹⃗ . 𝑛̂ is actually
the normal component of 𝐹⃗ at the point P and if we and the integral of 𝐹⃗ . 𝑛̂, then over the
surface S. So, this is the integral of the function 𝐹⃗ . 𝑛̂ over the surface S is given in this
fashion. So, this is just like a notation. So, when you will be given as let us say a vector
function and then from there you have to calculate the normal 𝑛̂; so, that you can evaluate
.
this surface integral ∫𝑆 𝐹⃗ . 𝑛̂𝑑𝑠 alright.

Now, when we evaluate the surface integral, we have to remember certain ways that how
we can evaluate. So, this is of course, is given as an abstract definition that what do we
mean by this component of 𝐹⃗ integral of component of 𝐹⃗ along the normal 𝑛̂ over the

555
surface S. But when we need to evaluate this we need some kind of formula and that will
help us to evaluate. So, the formula is you have to take the projection of the surface either
on xy plane or yz plane or zx plane. So, depending on your projection basically your
evaluation will depend.

Of course, the answer would remain same its just that the complications are; I do not know
the way you solve the problem will depend on which plane you are taking the projection
.
alright. So, here we have ∫𝑆 𝐹⃗ . 𝑛̂𝑑𝑠 . So, the first projection let us say we are projecting on
. . 𝑑𝑥𝑑𝑦
XY plane. So, then this will be a surface integral ∫𝑆 𝐹⃗ . 𝑛̂𝑑𝑠 = ∫𝑆 𝐹⃗ . 𝑛̂ |𝑛̂.𝑘̂| right; So, this is

what we get when we take the projection on first projection or projection on XY plane.
Similarly you can take the projection on XZ plane and we can take the projection on YZ
.
plane F dot n sorry XZ. So, if we are taking the projections this has to be ∫𝑆 𝐹⃗ . 𝑛̂𝑑𝑠 =
. 𝑑𝑥𝑑𝑧 .
∫𝑆 𝐹⃗ . 𝑛̂ |𝑛̂.𝑗̂ | and when we are taking projection and YZ plane so, this has to be ∫𝑆 𝐹⃗ . 𝑛̂𝑑𝑠 =
. 𝑑𝑦𝑑𝑧
∫𝑆 𝐹⃗ . 𝑛̂ |𝑛̂.𝑖̂| . So, that trick too on YZ plane. So, the trick to remember these formulas is

that whenever you are taking projection on a certain plane so, that will be in the product
dxdz and then the remaining axis. So, if you are taking the projection on XZ plane then
basically y axis is perpendicular.

So, you take 𝑛̂. 𝑗̂. So, 𝑗̂ is the unit vector along the x is y. So, we take basically 𝑛̂. 𝑗̂ similarly
if you are taking the projection on XY plane, then your z axis is perpendicular to x XY
plane and therefore, you take 𝑛̂. 𝑘̂. And similarly if we are taking YZ plane then x axis is
perpendicular to yz plane and therefore, we take the projection we take the product as 𝑛̂. 𝑖̂.
So, it is a very nice and convenient way to remember these formulas alright. So, now we
will see how we can solve some examples. So, this is of course, they are given in a bit
abstract way we will see how we can work out an example. So, let us start alright.

556
(Refer Slide Time: 16:45)

.
So, to begin with so, evaluate ∫𝑆 𝐹⃗ . 𝑛⃗⃗𝑑𝑆 where our function 𝐹⃗ (𝑥, 𝑦, 𝑧) = 𝑦𝑧𝑖̂ + 𝑧𝑥𝑗̂ +

𝑥𝑦𝑘̂ where S is the part of the surface of this sphere 𝑥 2 + 𝑦 2 + 𝑧 2 = 1 which is in the first
octant. So, here in this example what we have is a given vector function 𝐹⃗ and here we
.
have to calculate the surface integral ∫𝑆 𝐹⃗ . 𝑛⃗⃗𝑑𝑆 and our surface along which we have to
integrate basically or on which we have to integrate is given by 𝑥 2 + 𝑦 2 + 𝑧 2 = 1.
However, there is a small cast that we have to just consider a surface that is in the first
octant.

So, if you consider let us say a circle in 2D, then it actually has 4 how to say quadrants.
So, and in case of a sphere since it is a three dimensional geometry, you will have actually
eight octants and we have to limit to the just the first octant alright. So, let us start. So, first
of all if we want to calculate 𝐹⃗ . 𝑛̂, we have to find out 𝑛̂. What is our 𝑛̂ and unless you
calculate this 𝑛̂ we cannot proceed any further. But we remember from the gradient
divergence and curl that in gradient of a function or the gradient of a given. Let us say you
have the equation of a surface then the gradient of that that F equals to that surface is
basically a normal to the surface. So, the gradient of F is actually a normal to the curve F
is equal to some constant or F equals to 0.

So, basically a vector normal to the surface S is given by S equals to ⃗∇⃗𝜑 where 𝜑(𝑥, 𝑦, 𝑧) =
𝑥 2 + 𝑦 2 + 𝑧 2 − 1 alright. So, this is the required equation of the basically 𝜑(𝑥, 𝑦, 𝑧) =

557
𝑥 2 + 𝑦 2 + 𝑧 2 − 1. So, now we do not have to write S here. So, a vector normal to the
⃗⃗𝜑 and this ∇
surface S is given by ∇ ⃗⃗𝜑 can be calculated now. So, ∇
⃗⃗(𝑥 2 + 𝑦 2 + 𝑧 2 − 1) .

So, this is basically 2(𝑥𝑖̂ + 𝑦𝑗̂ + 𝑧𝑘̂) alright. So, we take the gradient and that is what we
obtain. Now this is normal to the surface S, but 𝑛̂ is the unit outward drawn normal;
⃗⃗𝜑 is just the outward drawn normal.
however, ∇

So, we have to now calculate 𝑛̂. So, to calculate 𝑛̂ a unit normal to any point actually. So,
⃗⃗𝜑
∇
a unit normal to any point (𝑥, 𝑦, 𝑧) is basically 𝑛̂ = |∇⃗⃗𝜑|.

(Refer Slide Time: 21:43)

⃗⃗𝜑
∇ ̂)
2(𝑥𝑖̂+𝑦𝑗̂ +𝑧𝑘
So, 𝑛̂ = |∇⃗⃗𝜑| = = (𝑥𝑖̂ + 𝑦𝑗̂ + 𝑧𝑘̂). So, that is our normal 𝑛̂ and now we will
2√𝑥 2 +𝑦 2 +𝑧 2

take projection on XY plane. So, if we take the projection on XY plane. So, this is our
plane y and that is the first octant R because if you take the projection of first octant in the
sphere; if you take the projection on XY plane, then it will be a circle in the first octant.
So, taking projection and of course, a vector perpendicular to this XY plane is the k axis
right alright.

So, taking projection of S into the XY plane; so, take a projection of S into the XY plane
gives so, this is basically our unit normal and now taking projection of S into XY plane it
. . 𝑑𝑥𝑑𝑦
will be basically. So, our ∬𝑆 𝐹⃗ . 𝑛̂𝑑𝑆 = ∬𝑅(𝑦𝑧𝑖̂ + 𝑧𝑥𝑗̂ + 𝑥𝑦𝑘̂ ). (𝑥𝑖̂ + 𝑦𝑗̂ + 𝑧𝑘̂) |𝑛̂.𝑘̂| .

558
 𝑑𝑥𝑑𝑦
So, we have 𝑛̂ and we have 𝐹⃗ and now dx is basically ̂|
right. So, this is nothing, but
|𝑛̂.𝑘

so R is basically our after projection, the new the new idea basically or the new place
where we are doing the integration.

And here in this case if our 𝑛̂ is 𝑛̂ is 𝑘̂ basically so, 𝑛̂ is the unit normal perpendicular to
it. So, 𝑛̂ is the unit normal and the 𝑘̂ is the given vector perpendicular to XY plane. So, if
you take 𝑛̂. 𝑘̂ then this will reduce to just z. So, from here what we have is 𝐹⃗ . 𝑛̂ is if you
break the dot product, then this will be 3xyz and 𝑘̂ will be 𝑛̂. 𝑘̂ will be basically z right.
.
And therefore, if I substitute everything here then surface integral ∬𝑆 𝐹⃗ . 𝑛⃗⃗𝑑𝑆 =
. 𝑑𝑥𝑑𝑦 .
∬𝑅 3𝑥𝑦𝑧 𝑧
= ∬𝑅 3𝑥𝑦𝑑𝑥𝑑𝑦 right.

And now, what we will do is we substitute since we are in the 3 in 2 dimensional geometry
and it is basically sort of like a surface in this integral in the in the circle. This whatever
you want to call, we basically substitute 𝑥 = 𝑟 cos 𝜃 & 𝑦 = 𝑟 sin 𝜃.

(Refer Slide Time: 26:53)

So, basically we substitute 𝑥 = 𝑟 cos 𝜃 & 𝑦 = 𝑟 sin 𝜃 and therefore, the required line
integral sorry that requires surface integral. Let us say I am calling it as 𝐼𝑆 =
𝜋
1
3 ∫𝑟=0 ∫𝜃=0
2 (𝑟 cos 𝜃)(𝑟 sin 𝜃)𝑟𝑑𝑟𝑑𝜃 right.

559
So, if you multiply this whole thing and take the term for r on one side so, this
𝜋
1
3 ∫𝑟=0 𝑟 3 𝑑𝑟 ∫𝜃=0
2 𝑠𝑖𝑛𝜃 cos 𝜃𝑑𝜃 So, we integrate this thing here and then we integrate this
3 1 3
thing here and this will give us 4 . 2. So, this is ultimately 8. So, we are just a half with 2

here and then we do sin 2𝜃 and all that. So, this is the required answer.

So, in this case we basically have to find out the projection of the surface on a certain plane
and just having to know the projection you can be able to solve the rest of the example.
We will solve some more examples on surface integral in our next class just to make the
concept clear and then we will move on to volume integral. So, I will stop here for today
and we will continue with our examples on surface integral in our next class.

Thank you.

560
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 55
Surface integral (Contd.)

Hello students. So, yesterday we were looking into the concepts of or in our previous class,
we were looking into the concepts of surface integral and circulation of function around a
given curve and things like that. So, today we will continue practicing those examples and
then if time permits, then we will move on to our next topic which is basically a Green’s
theorem actually. So, the Green’s function is actually in some other topic.

So, we will look into Green’s theorem in the vector calculus context. So, first of all, let us
start with our example on circulation actually.

(Refer Slide Time: 01:01)

So, our first example is find the circulation of 𝐹⃗ around round the curve C where 𝐹⃗ =
𝑦𝑖̂ + 𝑧𝑗̂ + 𝑥𝑘̂ and C is the circle 𝑥 2 + 𝑦 2 = 1 and 𝑧 = 0. So, basically here we have to
find out the circulation of this function F around curve C where we have the given function
𝐹⃗ (𝑥, 𝑦, 𝑧) it goes to this.

So, of course, 𝐹⃗ is a function of x, y and z. So and C is the circle given by 𝑥 2 + 𝑦 2 =

1 and 𝑧 = 0. So, basically our curve is in the x, y plane and the given equation of the curve

561
is basically a circle. So, by definition of circulations, So, since it is a circle which is a
simple closed curve, so any line integral along that simple closed curve will be circulation.
.
So, by definition, the circulation of 𝐹⃗ around C is this notation ∮𝐶 𝐹⃗ . 𝑑𝑟⃗. So, this can be

written as where 𝑟⃗ = 𝑥𝑖̂ + 𝑦𝑗̂ + 𝑧𝑘̂. We know that 𝑟⃗ is always given by this way and
therefore, from here we will have 𝑑𝑟⃗ = 𝑑𝑥𝑖̂ + 𝑑𝑦𝑗̂ + 𝑑𝑧𝑘̂, right.

.
Now, here if I write integral ∮𝐶(𝑦𝑖̂ + 𝑧𝑗̂ + 𝑥𝑘̂). ( 𝑑𝑥𝑖̂ + 𝑑𝑦𝑗̂ + 𝑑𝑧𝑘̂). Now, since we are in
the xy plane, so 𝑧 = 0 and 𝑑𝑧 = 0. So, if I substitute z is 0 here and these are 0 here, then
.
this will lead to integral over C this is 0, So, ultimately we will have ∫𝐶 𝑦𝑑𝑥, since on C z
is 0 and this implies 𝑑𝑧 = 0. So, we will basically substitute 𝑧 = 0 here and 𝑑𝑧 = 0 here.
.
So, we will end up with integral ∫𝐶 𝑦𝑑𝑥.

(Refer Slide Time: 04:25)

Now, if I substitute 𝑥 = cos 𝜃 & 𝑦 = sin 𝜃, so this is basically integral over the curve C.
So, in this case the parameter theta would vary from 0 to 2𝜋. I am substituting 𝑥 =
cos 𝜃 & 𝑦 = sin 𝜃. So, this is sin 𝜃 and then if 𝑥 = cos 𝜃 ⇒ 𝑑𝑥 = − sin 𝜃𝑑𝜃. So, this will
2𝜋 1 2𝜋
be − ∫𝜃=0 𝑠𝑖𝑛2 𝜃𝑑𝜃. So, I can adjust − 2 ∫𝜃=0 2𝑠𝑖𝑛2 𝜃𝑑𝜃. And then, this can be written as
1 2𝜋 1 sin 2𝜃 2𝜋 2𝜋
− 2 ∫𝜃=0(1 − cos 2𝜃)𝑑𝜃 = − 2 [𝜃 − ] =− = −𝜋.
2 𝜃=0 2

2𝜋
So, sin 4𝜋 = 0,sin 0 = 0. So, this will be ultimately − 2
. So, that is −𝜋. So, this is the

required circulation of this function 𝐹⃗ around this curve C given by this circle. So, circle

562
is a simple closed curve. So, the line integral will be a circulation. So, this is how we
calculate the circulation of the function 𝐹⃗ , alright. Now, we will continue with our
examples on a surface integral.

(Refer Slide Time: 05:49)

So, let me write an example, let me write a problem here. So, evaluate integral
.
∬𝑆 𝐹⃗ . 𝑛̂𝑑𝑆 where 𝐹⃗ = 𝑧𝑖̂ + 𝑥𝑗̂ − 3𝑦 2 𝑧𝑘̂ where S is the surface of the cylinder 𝑥 2 + 𝑦 2 =
16 included in the first octant 𝑧 = 0 and 𝑧 = 5. So, basically 𝑥 2 + 𝑦 2 = 16. So, that is a
circle and then you have 𝑧 = 0 and 𝑧 = 5. So, it is a cylinder.

Now, if you consider the part which is included in the first octant; that means, the circle in
the first octant and the length 𝑧 = 0 to 𝑧 = 5 at the, I mean like the height of that cylinder.
So, whatever lies in the first octant of that cylinder; so, that is the surface S and our given
vector function 𝐹⃗ is this one. So, first of all if we have a given vector function 𝐹⃗ , and if
you have the surface S, then from there we have to find out the normal 𝑛⃗⃗. So, a vector
normal to the surface we, if you remember then yesterday we basically calculated
the ⃗∇⃗(𝑥 2 + 𝑦 2 − 16). That is our given surface.

So, a vector normal to the surface S is given by ⃗∇⃗(𝑥 2 + 𝑦 2 − 16) and this will be 2(𝑥𝑖̂ +
𝑦𝑗̂), yes. So, this is the required normal and from here the unit normal. Therefore, the unit
normal therefore, 𝑛̂ basically, unit normal at any point on S; will be given by a gradient of

563
 ⃗⃗𝜑
∇
let us say this one was our 𝜑 = 𝑥 2 + 𝑦 2 − 16. So, if this one was our 𝜑. So,𝑛̂ = |∇⃗⃗𝜑| =
2(𝑥𝑖̂+𝑦𝑗̂ ) 2(𝑥𝑖̂+𝑦𝑗̂ ) 𝑥𝑖̂+𝑦𝑗̂
= = .
√4(𝑥 2 +𝑦 2 ) 2×4 4

So, this is our unit normal 𝑛̂. Now, since we have a cylinder which is whose base is 𝑥 2 +
𝑦 2 = 16 and height is 5. So, we take a projection either on xz plane or yz plane.

(Refer Slide Time: 09:44)

.
So, if we take the projection on xz plane, so we have in surface integral ∬𝑆 𝐹⃗ . 𝑛̂𝑑𝑆 or let
. . 𝑑𝑥𝑑𝑧
us say 𝑛̂. So, this can be written as ∬𝑆 𝐹⃗ . 𝑛̂𝑑𝑆 = ∬𝑅 𝐹⃗ . 𝑛̂ |𝑛̂.𝑗̂| , right.

So, this is our required dot product where R is the and this one can be let us write as 𝑆1, so
where 𝑆1 is the projection or let us call it as R, so S for the surface and R for the region,
where R is the projection of S on xz plane. So, that is the projection on xz plane So, we
cannot take the projection of S on xy plane as the surface S is perpendicular to the xy
plane. So, we simply cannot take the projection because since the cylinder is perpendicular
to xy plane. So, if we take the projection, then it will be just a circle. So, I mean we exclude
that cylinder part here.

So, that is why we take the projection either on xz plane or yz plane. So, now our 𝑛̂ =
𝑥𝑖̂+𝑦𝑗̂ 𝑦 𝑦
. So, from here 𝑛̂. 𝑗̂ = 4 and |𝑛̂. 𝑗̂| = 4 right. And 𝐹⃗ . 𝑛̂ is basically 𝐹⃗ . 𝑛̂ is our
4

564
 𝑥𝑖̂+𝑦𝑗̂
(𝑧𝑖̂ + 𝑥𝑗̂ − 3𝑦 2 𝑧𝑘̂). ( ) this. So, when we take the dot product, so the k component is
4

0. So, that will be 0 and therefore, this will be 𝐹⃗ . 𝑛̂ = (𝑥𝑧 + 𝑥𝑦) right.

So, when we take the dot product, now we substitute all these values here. So, our required
. . (𝑥𝑧+𝑥𝑦) 𝑑𝑥𝑑𝑧
surface integral ∬𝑆 𝐹⃗ . 𝑛̂𝑑𝑆 = ∬𝑅 . 𝑦 .
4
4

(Refer Slide Time: 14:37)

Now, this can be written as if we simplify, then this will be written as

5 4 𝑥𝑧+𝑥√16−𝑥 2 5 4 𝑥𝑧
∫𝑧=0 ∫𝑥=0 √16−𝑥 2
𝑑𝑥𝑑𝑧 = ∫𝑧=0 ∫𝑥=0 (√16−𝑥 2 + 𝑥) 𝑑𝑥𝑑𝑧. So, we did the separation. So,

basically the partial fraction and now we can integrate this very easily. So, we basically do
𝑥2
the integration with respect to x first and so, this one will be and this one will be if we
2

substitute this equals to z. So, that will be just √𝑧 and all that.

So, doing a simplification from here is not complicated. So, we will basically at the end
5
obtain ∫0 (4𝑧 + 8)𝑑𝑧 which is equals to 90. So, this is the required surface integral. So,
here there is a small mix up. So, it is just that here it will be 𝑥 2 + 𝑦 2 , this 2 will come
outside. And since 𝑥 2 + 𝑦 2 = 16, so it will be 2 times 4. So, 2, 2 will get cancel. So, we
will obtain simply 4 yeah sorry about that mix up. And then when we take the dot product,
𝑦 𝑦
so then this will be 4 and this will be 𝑥𝑗̂ + 4. So, ultimately they get cancel. So, they really

do not affect anything. It is just that yeah, we have to be a little bit careful.

565
And now, substituting 𝑦 = √16 − 𝑥 2 and here also, then we do the partial fraction and
then this is ultimately 90. So, this is how we calculate the surface integral of a given
function 𝐹⃗ whose surface S is given by that circle and cylinder. So, that was another
interesting example. We can consider another example on surface integral before we move
to the Green’s theorem.

(Refer Slide Time: 16:37)

.
So, another example, evaluate ∬𝑆 𝐹⃗ . 𝑛̂𝑑𝑆 where 𝐹⃗ = 𝑦𝑖̂ + 2𝑥𝑗̂ − 𝑧𝑘̂ on S and S is the
surface of the plane 2𝑥 + 𝑦 = 6 in the first octant cut off by the plane 𝑧 = 4. So, here we
have a given vector function actually and S is the surface of the plane in the first octant cut
off by the plane 𝑧 = 4 right. So, that is the given equation of the surface.

So, now, first of all we will calculate the normal. So, a vector normal to the surface S is
⃗⃗(2𝑥 + 𝑦 − 6). So, this is ultimately (2𝑖̂ + 𝑗̂)and therefore, from
given by let us say,𝑛⃗⃗ = ∇
⃗⃗
𝑛 2𝑖̂+𝑗̂
here our unit normal is 𝑛̂ = |𝑛⃗⃗|. So, this will be , right. So, this is our required unit
√5

normal actually. And the given surface integral S can be written as so, since we have to
take a, projection. So, in this case also we take the projection either on xz or yz plane.

So, we cannot take the projection on xy plane because the surface S is, perpendicular to
the xy plane. So, here we take the surface integral and we take the projection. So, that is
. . 𝑑𝑥𝑑𝑧
the region R where the projection will be ∬𝑆 𝐹⃗ . 𝑛⃗⃗𝑑𝑆 = ∬𝑅 𝐹⃗ . 𝑛⃗⃗ |𝑛̂.𝑗̂ |

566
 1 2𝑦+2𝑥
So, now 𝑛̂ is given in this fashion. So, 𝑛̂. 𝑗̂ = right and 𝐹⃗ . 𝑛̂ = ; the 𝑘̂ component
√5 √5

is 0. So, dot product after taking the dot product, the 𝑘̂ component will not occur here. So,
that is what we will get 𝐹⃗ . 𝑛̂ and that is what we will get as 𝑛̂. 𝑗̂. So, we substitute everything
here.

(Refer Slide Time: 20:00)

. 2 𝑑𝑥𝑑𝑧 .
So, it will be let us say let us say 𝐼𝑆 = ∬𝑅 (𝑦 + 𝑥) 1 = 2 ∬𝑅(𝑦 + 𝑥)𝑑𝑥𝑑𝑧 =
√5
√5

. . 3 4
2 ∬𝑅(6 − 2𝑥 + 𝑥)𝑑𝑥𝑑𝑧 = 2 ∬𝑅(6 − 𝑥)𝑑𝑥𝑑𝑧 = 2 ∫𝑥=0 ∫𝑧=0(6 − 𝑥)𝑑𝑥𝑑𝑧.

(Refer Slide Time: 21:27)

567
 3 4
So, I can write the integral as 2 ∫𝑥=0 ∫𝑧=0(6 − 𝑥)𝑑𝑥𝑑𝑧 and we can integrate with respect
to z first. So, it will be 2 times 4 and then we integrate with respect to x. So, this will
3 9
2.4 ∫𝑥=0(6 − 𝑥)and. So, this will be 2.4 (6.3 − 2) = 108. So, ultimately, it will be 108.

So, that is the required surface integral.

So, like this here you can practice many more examples where you have a given vector
function. So, you take the projection on xy plane yz plane or zx plane. You have to find
out the normal 𝑛⃗⃗ which is perpendicular to the surface and then everything is just doing
some simple calculation. So, after doing the projection, you get to know what is the limit
for x or y or z depending on your projection and then just perform the integration it is from
then on onwards it will become very simple. So, I will move to our next topic because I
think we have a practiced enough examples on surface integral. And next, we will move
to our next topic which is basically Green’s theorem.

(Refer Slide Time: 22:49)

So, Green’s theorem states that let R be a closed bounded domain or closed bounded
region, in the xy plane whose boundary C consists of finitely many smooth curves, smooth
curves. And let M and N be the continuous functions of x and y having continuous partial
𝜕𝑀 𝜕𝑁 . 𝜕𝑁 𝜕𝑀
derivatives, and in R. Then so, basically then surface integral ∬𝑅( 𝜕𝑥 − )=
𝜕𝑦 𝜕𝑥 𝜕𝑦
.
∮𝐶(𝑀𝑑𝑥 + 𝑁𝑑𝑦).

568
Where the line integral, being taken along the entire boundary C of R such that R is on the
left side as one advances in the direction of integration. So, basically R is a closed bounded
region in the xy plane. And suppose we have two functions M and N which are continuous
𝜕𝑀 𝜕𝑁
functions of x and y which has continuous partial derivatives and in R.
𝜕𝑦 𝜕𝑥

. 𝜕𝑁 𝜕𝑀
Then, if it has continuous partial derivatives, then we can write ∬𝑅( 𝜕𝑥 − )=
𝜕𝑦
.
∮𝐶(𝑀𝑑𝑥 + 𝑁𝑑𝑦).

where the line integral is taken in such a way that when you walk along this curve C, then
your region are will always be on the left hand side. So, that means, it always have a anti
clockwise direction. So, you the way you are walking along the boundary C is such that
the region R which is being closed by the boundary by the curve C is always on the left
hand side. So that means, you always walk along how to say along the anti clockwise
direction and this is what this theorem says and all you need to have is that M and N should
have continuous partial derivatives of order 1 and then this identity would hold.

So, we can actually verify this identity, it is also one of the important theorems in vector
calculus. And we can be able to verify this by using some formulas and some calculus,
calculus results. So, I will stop here from for today and in the next class we will work out
one or two examples in Green’s theorem and we will see how it is, being solved. So, thank
you for your attention and I look forward to you in your next class.

569
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 56
Green’s Theorem & Example

Hello, students. So, in your previous class we started with Green’s theorem in vector
calculus.

(Refer Slide Time: 00:41)

So, we saw that if you have two functions say M and N which has continuous partial
derivatives with respect to x and y then in that case you can be able to write this identity
. 𝜕𝑁 𝜕𝑀 .
∬𝑅 ( 𝜕𝑥 − ) = ∮𝐶 (𝑀𝑑𝑥 + 𝑁𝑑𝑦) and you take the integration this line integral in such a
𝜕𝑦

way that the region R which is being enclosed by the curve C is always on the left hand
side; so, that means, in a way you are walking in an anticlockwise direction, alright.

So, that was the statement of for Green’s theorem today we will verify Green’s theorem
with some examples. The thing is why do we have to study these theorems like Green’s
theorem, Gauss theorem or Stokes’ theorem; so, you can see that on the left hand side you
have a surface integral, but on the right hand side you have a line integral. So, sometimes
it is not easy to evaluate this line integral and if your M and N they have continuous partial
derivatives then you can basically evaluate you just take their partial derivatives and then

570
you can evaluate their surface integral and most of the time it becomes a very simple
expression or you have a let us say surface integral and it has a complicated region and
other things. So, what you do you basically identify your M and N and then you evaluate
the line integral.

So, it is a very handy tool to convert a surface integral into a line integral or a line integral
into a surface integral and most of the time it actually make our life easier to evaluate either
one of them when their respective integrals are in surface or line and they are complicated
in a way. So, just to do this conversion it sometimes we may be able to evaluate the integral
easily, but right now we will just verify this theorem.

(Refer Slide Time: 02:31)

So, let us see how we can do that. So, in order to verify let me start with this example. So,
.
verify Green’s theorem in the plane for integral ∮𝐶(𝑥𝑦 + 𝑦 2 )𝑑𝑥 + 𝑥 2 𝑑𝑦 where C is the
closed curve of the region bounded by 𝑦 = 𝑥 and 𝑥 = 𝑦 2, alright.

So, first of all we in order to verify the Green’s theorem first of all we have to identify
what is our M and what is our N. So, let us look at the statement. So, in the statement it
.
says that you have a circulation or let us say line integral ∮𝐶(𝑀𝑑𝑥 + 𝑁𝑑𝑦) . So, here you
have dx and then this thing and then you have dy and then this thing, so; that means, this
must be our M and this must be our N. So, by Green’s theorem in the plane; so, what is
the plane? So, we have x y. So, this is our line 𝑦 = 𝑥 and then we have a parabola 𝑥 2 = 𝑦.
So, this must be our parabola alright.

571
So, we have to walk in such a way that the region R must be on the left hand side. So, that
means, we are walking in the anticlockwise direction and then this point of intersection let
us say P can be obtained. So, they basically intersect at two points; first at origin and then
the second one at P. So, by Green’s theorem in plane this we have to verify
. 𝜕𝑁 𝜕𝑀 .
∬𝑅 ( 𝜕𝑥 − ) = ∮𝐶 (𝑀𝑑𝑥 + 𝑁𝑑𝑦) , where 𝑀(𝑥, 𝑦) = 𝑥𝑦 + 𝑦 2 & 𝑁(𝑥, 𝑦) = 𝑥 2 , right. So,
𝜕𝑦

these are our M and N.

Now, the curves 𝑦 = 𝑥 and 𝑥 2 = 𝑦 intersect at; so, we can solve these two equations to
obtain the point of intersection. So, if you solve them then basically what we have to do is
substitute 𝑦 = 𝑥. So, this will be 𝑥 2 − 𝑥 = 0. So, the two possible points are 0 and 1 and
so, y is also 0 and 1 therefore, the intersect at (0,0) and (1,1). So, we just solve these two
equations and will be able to obtain these the point of insertion at (0,0) and (1,1)alright.
So, we have these two points of intersection.

(Refer Slide Time: 06:33)

𝜕𝑁 𝜕𝑀 𝜕𝑁
Now, in order to verify the Green’s theorem we first calculate 𝜕𝑥 and 𝜕𝑦 , alright. So, 𝜕𝑥 =
𝜕 𝜕𝑀 𝜕
(𝑥 2 ) = 2𝑥. So, this is basically 2x and = 𝜕𝑦 (𝑥𝑦 + 𝑦 2 ) = 𝑥 + 2𝑦, alright. Now, we
𝜕𝑥 𝜕𝑦
. 𝜕𝑁 𝜕𝑀 .
will calculate the surface integral. So,∬𝑅 ( 𝜕𝑥 − ) 𝑑𝑥𝑑𝑦 = ∬𝑅[2𝑥 − 𝑥 − 2𝑦]𝑑𝑥𝑑𝑦 =
𝜕𝑦
. 1 𝑥 1 𝑥
∬𝑅(𝑥 − 2𝑦)𝑑𝑥𝑑𝑦 = ∫𝑥=0 ∫𝑦=𝑥 2(𝑥 − 2𝑦)𝑑𝑥𝑑𝑦 = ∫𝑥=0[∫𝑦=𝑥 2(𝑥 − 2𝑦)𝑑𝑦]𝑑𝑥 .

572
 1 𝑥
So, we first integrate with respect to y. So, this will be ∫𝑥=0[∫𝑦=𝑥 2(𝑥 − 2𝑦)𝑑𝑦]𝑑𝑥. So, we

integrate with respect to y first, substitute the limit and then we integrate with respect to x.

(Refer Slide Time: 08:23)

So, ultimately we will obtain after integrating with respect to y and putting the values
1 𝑥5 𝑥4 1 1 1 1 1
∫0 (𝑥 4 − 𝑥 3 )𝑑𝑥 = [ 5 − ]
4 0
= − = − .. So, this will be basically − .
5 4 20 20

Now, that is this surface integral, now we will evaluate the line integral to match the values
whether they are same or not. So, for the circulation part we see that of course, it is a closed
curve, but it is a piecewise smooth curve actually. So, it is continuous from here to here
and then here to here. So, we actually evaluate the line integral along these two paths. So,
first we will evaluate along this path and then we will along this path. So, we did some an
example like this.

. .
So, if we go back. So, this is our ∮𝐶(𝑀𝑑𝑥 + 𝑁𝑑𝑦) = ∮𝐶(𝑥𝑦 + 𝑦 2 )𝑑𝑥 + 𝑥 2 𝑑𝑦 =
. .
∮𝐶 (𝑥𝑦 + 𝑦 2 )𝑑𝑥 + 𝑥 2 𝑑𝑦 + ∮𝐶 (𝑥𝑦 + 𝑦 2 )𝑑𝑥 + 𝑥 2 𝑑𝑦 . So, the part of the parabola is
1 2

considered as to be 𝐶1 and the part of the straight line is considered as to be 𝐶2 .

Now, along 𝐶1 basically we have 𝑦 = 𝑥 2 and therefore, 𝑑𝑦 = 2𝑥𝑑𝑥. So, I will substitute
𝑑𝑦 = 2𝑥𝑑𝑥 here and x will vary from 0 to 1 and similarly, along this line we have y is
equals to and along 𝐶2 we have 𝑦 = 𝑥. So, 𝑑𝑦 = 𝑑𝑥. So, we substitute 𝑦 = 𝑥 and 𝑑𝑦 =

573
 1 1
𝑑𝑥. So, let us call it as 𝐼𝐶 = ∫𝑥=0[(𝑥. 𝑥 2 + 𝑥 4 )𝑑𝑥 + 2𝑥 3 𝑑𝑥] + ∫𝑥=0[(𝑥 2 + 𝑥 2 )𝑑𝑥 +
1
𝑥 2 𝑑𝑥] = − 20.

Now, we have to evaluate this this line this integral where the limit is from 0 to 1. Let me
put everything in the bigger bracket. So, this is our given integral and if you evaluate this
1
whole thing, then it will be actually you will obtain as − 20. So, this is not very complicated

to obtain. So, of course, there will be a minus sign here because forgot to tell you because
we are going in the reverse direction. So, if we are going in the reverse direction; that
means, we are going from 1 to 0.

So, if we are going from 1 to 0, then I want to reverse the direction I want to go from 0 to
1, so, I have to put a minus sign here. So, this is an obvious thing that we have to remember
and keeping a minus sign here because initially it was supposed to be 1 to 0. So, now, we
1
are doing 0 to 1. So, a minus sign and then you evaluate and then basically obtain − 20.

So, you see this is how we verify the Green’s theorem.

So, it will be slightly lengthy all these theorem. So, Green’s theorem, Stokes’ theorem or
Gauss theorem if you want to verify then it will be lengthy because you have to check both
sides of the identity. So, here in this case we had to guess what is our M and N and from
𝜕𝑁 𝜕𝑀
there we had to calculate , 𝜕𝑦 and then substitute here and simplify the surface integral.
𝜕𝑥

Similarly, for the line integral I had to calculate the line integral along two different paths
and substitute here and just see whether the two sides are equal or not. So, since the both
sides are equal that means, that actually that Green’s theorem hold true in this case, alright.
So, this was an interesting example where you verify the Green’s theorem.

574
(Refer Slide Time: 13:15)

Let us consider an another example where we can use the Green’s theorem to simplify a
given integral which is slightly complicated.

.
So, example 2 evaluate by Green’s theorem ∫𝐶(𝑥 2 − cosh 𝑦)𝑑𝑥 + (𝑦 + sin 𝑥)𝑑𝑦, where
C is the rectangle with vertices (0,0), (𝜋, 0), (𝜋, 1) & (0,1) , right. So, we basically have a
region R. So, let me draw this region R xy. So, this is our rectangle. So, we have vertices
(0,0); (𝜋, 0) which is say A, then B is (𝜋, 1) and then this one is (0,1) alright. So, this is
our region R this is the curve C along which we have to calculate this line integral.

So, by Green’s theorem we know that now we will see first whether the Green’s theorem
is applicable or not. So, first of all we have (𝑥 2 − cosh 𝑦) which is obviously,
differentiable and I have continuous first order partial derivative and then we have
(𝑦 + sin 𝑥) which is again having continuous partial derivatives. So, that means, these are
our M and N and they both behave nicely. So, by Green’s theorem we have to write or
when we have been use.

. 𝜕𝑁 𝜕𝑀 .
So, we when we use Green’s theorem so, this will be ∬𝑅 ( 𝜕𝑥 − ) = ∮𝐶(𝑀𝑑𝑥 + 𝑁𝑑𝑦) .
𝜕𝑦

So, that means, since our M and N have a continuous first order partial derivatives I can
use Green’s theorem and therefore, this line integral can be converted into a surface
integral, alright. So, if I convert this line integral into a surface integral then our 𝑁(𝑥, 𝑦) =

575
 𝜕𝑁
𝑦 + sin 𝑥 𝑠𝑜 = cos 𝑥 and our 𝑀(𝑥, 𝑦) = 𝑥 2 − cosh 𝑦. So, if we differentiate then
𝜕𝑥
𝜕𝑀
= − sinh 𝑦, right, alright.
𝜕𝑦

(Refer Slide Time: 16:21)

.
So, then we substitute in this surface integral let us say 𝐼𝑅 = ∬𝑅(cos 𝑥 + sinh 𝑦)𝑑𝑥𝑑𝑦 =
𝜋 1
∫𝑥=0 ∫𝑦=0(cos 𝑥 + sinh 𝑦)𝑑𝑥𝑑𝑦. So, since we do not have any product of x and y or

something, so that means, here we do not have a function of y or here we do not have a
function of x.

So, it is fairly easy to integrate because then we separate the terms and once we separate
𝜋 𝜋
the terms then this will be basically integral ∫𝑥=0[𝑦 cos 𝑥 + cosh 𝑦]1𝑦=0 𝑑𝑥 = ∫𝑥=0[cos 𝑥 +
cosh 1 − 1]𝑑𝑥 = [sin 𝑥 + 𝑥 cosh 1 − 𝑥]𝜋0 = 𝜋(cosh 1 − 1).

So, you see initially we had a very complicated line integral to evaluate, but we just took
help of Green’s theorem which says that you can be able to use this identity only when M
and N have continuous first order partial derivatives with respect to y and x, I believe and
since our functions are behaving nicely I just calculated the partial derivatives and put it
on the left hand side of this formula and that gave us this nice result, alright. So, this is one
such situation, where we can use that theorem and obtain the required result.

576
(Refer Slide Time: 18:31)

Similarly, we can have an solve an another example. So, let us consider evaluate this
example evaluate by Green’s theorem. So, here it says specifically Green’s theorem
.
∮𝐶(cos 𝑥 sin 𝑦 − 𝑥𝑦) + sin 𝑥 cos 𝑦 𝑑𝑦 where C is the circle, 𝑥 2 + 𝑦 2 = 1, alright. So, here
again we are given a very complicated expression to evaluate and obviously, when you are
told that you have to use Green’s theorem then you really do not have any way out. So, we
see how we can use the Green’s theorem.

So, first of all we have (cos 𝑥 sin 𝑦 − 𝑥𝑦) & sin 𝑥 cos 𝑦 . So, they are all very nicely
behaving functions. So, obviously, they have continuous partial derivatives. So, we do not
have to worry about that. So, then we can write here 𝑀(𝑥, 𝑦) = (cos 𝑥 sin 𝑦 − 𝑥𝑦). So,
𝜕𝑀 𝜕𝑀
what will be our ? 𝜕𝑦 = cos 𝑥 cos 𝑦 − 𝑥, alright and 𝑁(𝑥, 𝑦) = sin 𝑥 cos 𝑦. So, this will
𝜕𝑦
𝜕𝑁 𝜕𝑁
be 𝜕𝑥 and = cos 𝑥 cos 𝑦.
𝜕𝑥

. 𝜕𝑁 𝜕𝑀 .
So, therefore, by Green’s theorem we have ∬𝑅 ( 𝜕𝑥 − ) = ∮𝐶(𝑀𝑑𝑥 + 𝑁𝑑𝑦) . So, you
𝜕𝑦

see we will basically obtain a very simple expression to solve. So, instead of working with
this complicated one we will obtain a very simple expression to sort or to work with.

577
(Refer Slide Time: 21:17)

. 𝜕𝑁 𝜕𝑀 .
So, let me write 𝐼𝑅 = ∬𝑅 ( 𝜕𝑥 − ) 𝑑𝑥𝑑𝑦 = ∬𝑅[cos 𝑥 cos 𝑦 + 𝑥 − cos 𝑥 cos 𝑦]𝑑𝑥𝑑𝑦 =
𝜕𝑦

. 1 √1−𝑥 2
∬𝑅 𝑥𝑑𝑥𝑑𝑦 = − ∫𝑥=0 ∫𝑦=−√1−𝑥 2 𝑥𝑑𝑥𝑑𝑦, right.

right because that is the range for the y and the range for the x is. So, the range for the x is
𝜕𝑁
0 to 1 or what we can do. So, here you can integrate and then we basically obtain so, 𝜕𝑥 .
𝜕𝑀
So, this is our M and this is our N. So, that M; so, this is my M. So, = cos 𝑥 cos 𝑦 − 𝑥.
𝜕𝑦

(Refer Slide Time: 23:03)

578
So, this is basically.

(Refer Slide Time: 23:45)

.
So, ultimately we will obtain here is ∬𝑅 𝑥𝑑𝑥𝑑𝑦, right. So, we will obtain
. 1 √1−𝑥2
basically ∬𝑅 𝑥𝑑𝑥𝑑𝑦 . So, either we can have − ∫𝑥=0 ∫𝑦=−√1−𝑥 2 𝑥𝑑𝑥𝑑𝑦 or we can substitute

𝑥 = 𝑟 cos 𝜃 and 𝑦 = sin 𝜃.

1 2𝜋
So, then in that case the surface integral 𝐼𝑅 = ∫𝑟=0 ∫𝜃=0(𝑟 cos 𝜃)(𝑟𝑑𝑟𝑑𝜃) =
1 2𝜋 1
∫𝑟=0 𝑟 2 𝑑𝑟 ∫𝜃=0 cos 𝜃 𝑑𝜃 = ∫𝑟=0 𝑟 2 𝑑𝑟 [sin 𝜃]2𝜋
0 = 0.

So, here you can see that for this given expression here where we had to calculate the line
integral for this curve C, I guess the function incorrectly. So, this is my M and this is our
𝜕𝑁
N. So, we have to calculate 𝜕𝑥
which is cos 𝑥 cos 𝑦 and then here we have to calculate
𝜕𝑀
which is cos 𝑥 cos 𝑦 − 𝑥.
𝜕𝑦

So, we substitute the whole thing in this expression, on the left hand side and then we do
some calculations and either we can proceed in this way or we can proceed in this way it
is up to us and we just substitute 𝑥 = 𝑟 cos 𝜃 & 𝑦 = 𝑟 sin 𝜃 and the ultimate answer is 0.
So, looking at this integral it would not have been easy to guess the limit that it will because
the value of the integral not the limit to the value of the integral that it will be 0. But, it is

579
just that taking help of Green’s theorem, we can be able to show that the value of the
integral is 0 by doing some simple partial derivatives.

So, you got the idea that what we have to do actually. There might be some errors here and
there while doing the calculation. I hope you would understand. But, yeah the message is
that whenever you have a complicated line integral given to you and if your M and N
seems to be having continuous partial derivatives then try to use Green’s theorem and there
is a strong possibility if you use the Green’s theorem then the whole integral will reduce
to a very simpler one and you just like in this case or in the previous case you just do some
simple known formula or known calculation to obtain the value.

So, today we tried to learn about Green’s theorem and we also saw its how to say efficiency
that how you can use it and how it makes our life easier while calculating the surface
integral or line integral. So, I will stop here for today now and in our next class we will
start with the Gauss divergence theorem. So, I thank you for your attention.

580
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture – 57
Volume integral, Gauss theorem

Hello, students. So, in the last class we practiced surface integral and we also learnt about
Green’s theorem where we could connect the line integral with the surface integral in a
way and if we have a certain form given let us say in the line integral then from there we
can guess the M and N part of this integral. And from there we can convert this line
integral, if the if M and N have continuous partial derivatives then we can convert this line
integral into a surface integral and more how to say in more situations where you can
actually convert a complicated line integral into a simple surface integral and vice versa.

So, you may have how to say complicated surface integral could be and if you see that you
can actually integrate it back to M and N then that line integral might become easier. So,
it works in both ways it is just that your M and N needs to have continuous partial
derivatives. So, we practiced a few examples motivated from this on Green’s theorem part
and now we will start with volume integral. So, we did surface integral Green’s function
and then now today we will practice some volume integral examples and then we will
move to Gauss divergence theorem, alright.

(Refer Slide Time: 01:43)

581
So, volume integral basically mean that suppose the statement goes like this. So, suppose
you have V is a volume bounded by a surface S suppose 𝑓(𝑥, 𝑦, 𝑧) is a single valued
function defined on V. So, that means, 𝑓(𝑥, 𝑦, 𝑧) is defined on capital V. So, it maps
actually from V to set of all real numbers in a way.

And when we write then the volume integral basically so, it follows the similar motivation
what we have learned for the function of one variable. So, in case of function of one
variable you have an interval then you divide this interval into n number of sub intervals
and then sum of the areas of each of these sub intervals when you sum them and when you
make n goes to infinity then that gives you actually the integral of that function defined in
that interval [𝑎, 𝑏].

So, its also the same here. So, you basically consider sub volumes on V and then you
consider the volume of those how to say sum of those sub volumes times f and then you
basically take n tends to infinity and that gives you actually the volume integral. So, the
motivation is pretty much same what we have learnt in case of function of one variable or
function of two variable. So, I am not writing all those things. I am just writing the notation
for the volume integral because our main target here to practice few examples because we
have very less number of lectures left.

So, then the volume integral of 𝑓(𝑥, 𝑦, 𝑧) is denoted by; or given by let us say I am
writing 𝐼𝑉 . So, I is for the integral, V is for the volume and then you write triple integrals.
.
So, triple integral is for the volume integral and we write 𝐼𝑉 = ∭𝑉 𝑓(𝑥, 𝑦, 𝑧)𝑑𝑉, alright.
so, if we divide this V into sub volumes then in that case on one of those sub volumes we
can take the length breadth and height as dx dy and dz. So, that is basically the volume
element in a way.

. .
So, this can be written as integral 𝐼𝑉 = ∭𝑉 𝑓(𝑥, 𝑦, 𝑧)𝑑𝑉 = ∭𝑉 𝑓(𝑥, 𝑦, 𝑧)𝑑𝑉. So, that is

basically our 𝑑𝑉 = 𝑑𝑥𝑑𝑦𝑑𝑧. So, this is the required volume element and if 𝐹⃗ is a vector
.
function say 𝐼𝑉 = ∭𝑉 𝐹⃗ . 𝑑𝑉
⃗⃗ . So, 𝐹⃗ is a vector function when we are doing the volume

integral, then we have to take this dot product is also an example of volume integral.

⃗⃗ we take this 𝑑𝑉
So, basically instead of 𝑑𝑉 ⃗⃗ vector, alright and this is also an example of
volume integral. So, that is how we write the volume integral. Now, let us see if we can
solve some examples. So, I am just looking for an example in my lecture note, yes.

582
(Refer Slide Time: 06:39)

.
So, to start with let me consider this example evaluate integral ∭𝑉 𝜑𝑑𝑉 where
𝜑(𝑥, 𝑦, 𝑧) = 45𝑥 2 𝑦 and V is the volume or a closed region bounded by the planes 4𝑥 +
2𝑦 + 𝑧 = 8 and 𝑥 = 0, 𝑦 = 0 and 𝑧 = 0.

So, obviously, drawing this plane is very easy because it is actually x equals to 0, y equals
to 0 and z equals to 0 and bounded by that plane 4𝑥 + 2𝑦 + 𝑧 = 8. So, you can actually
be able to draw this in 3D this plane and it is fairly a simple domain basically or a bounded
volume all right. So, here we if you want to evaluate the limit for let us say when y and z
are 0, then x is varying from 0 to 4. So, x intercept will be (0,0) and (4,0) sorry x intercept
would be (0,0) and (2,0) .

So, there is a 4 here excuse me, y intercept would be (0,0) and (0,4), 0 and z intercept
would be (0,0,8) in a way. So, you are getting the idea what I am trying to say. So, when
you draw this you can actually how does they get the x the point where it is intersecting
the x axis, the point where it is intersecting the y axis and the point where it is intersecting
the z axis based on that you can be able to draw this plane. So, it is a bounded domain and
now, if we want to evaluate the volume integral.

.
So, as I was saying we can write it as 𝐼𝑉 = ∭𝑉 45𝑥 2 𝑦𝑑𝑥𝑑𝑦𝑑𝑧. Now, to write the limits
2 8−4𝑥 8−4𝑥−2𝑦
for x, y and z how do we calculate the limits? So,∫𝑥=0 ∫𝑦=0 ∫𝑧=0 45𝑥 2 𝑦𝑑𝑥𝑑𝑦𝑑𝑧 =

583
 2 8−4𝑥 8−4𝑥−2𝑦 2 8−4𝑥
∫𝑥=0 ∫𝑦=0 45𝑥 2 𝑦𝑑𝑥𝑑𝑦(∫𝑧=0 𝑑𝑧) = ∫0 ∫0 45𝑥 2 𝑦(8 − 4𝑥 − 2𝑦)𝑑𝑥𝑑𝑦 =
2 8−4𝑥 2 𝑥2
∫𝑥=0(∫𝑦=0 [45𝑥 2 𝑦(8 − 4𝑥 − 2𝑦)] 𝑑𝑦)𝑑𝑥 = 45 ∫𝑥=0 3
(4 − 2𝑥)2 𝑑𝑥 = 128

(Refer Slide Time: 10:13)

2 8−4𝑥
So, ultimately we will obtain integral ∫0 ∫0 45𝑥 2 𝑦(8 − 4𝑥 − 2𝑦) 𝑑𝑥𝑑𝑦 and now, we
2 8−4𝑥
integrate with respect to y. So, this will be ∫𝑥=0(∫𝑦=0 [45𝑥 2 𝑦(8 − 4𝑥 − 2𝑦)] 𝑑𝑦)𝑑𝑥.

Sorry, about the brackets, but you can use the correct bracket. So, here it should be curly
bracket and then the big bracket, but using bracket is not the concern here, so, I am not
focusing on that thing.

And now we can integrate this and when we integrate with respect to y. So, this will
become. So, we multiply by 45 inside this bracket and then we do the integration. So,
2 𝑥2
everything is pretty straightforward and ultimately we will obtain 45 ∫𝑥=0 (4 −
3

2𝑥)2 𝑑𝑥 and when we integrate with respect to x then this whole thing will reduce to 128.
So, the required answer is 128.

So, other than doing some complicated arithmetic calculation there is really nothing much
in this example. So, what you really had to do just substitute the value of of this integrand
here and then guess the and then calculate the limit for z, y and x or z is running from 8 −

584
4𝑥 − 2𝑦 similarly for y similarly for x and then integrate individually and just do some
complicated calculations here. So, that is the only difficult part in here other than that this
example is pretty straightforward. So, this is an example or interesting example where we
calculated the volume integral next is the statement of Gauss divergence terms. So, let me
go to this statement, yes.

So, now if you have given as we saw in this example if you are given a function 𝜑 or
whatever it is and if you are given the volume V then you can be able to calculate the
volume integral. Similarly, if it is a vector function then you take the dot product
.
∭𝑉 𝐹⃗ . 𝑑𝑉
⃗⃗ and just calculate the limit substitute the limit and then that will be; pretty much

it. So, in either cases it is a very generic example.

Now, we go to a very important theorem in vector calculus and not only in vector calculus
it is used in other fields as well. For example, in partial differential equations in some
context you use these theorems in particular this divergence theorem.

(Refer Slide Time: 13:31)

So, let me give you the statement Gauss divergence theorem. So, the statement goes like
this. Suppose V is the volume bounded by a closed piecewise smooth surface S. So, it can
be a parallelepiped.

So, parallelepiped if you consider or a cube then in that case cube is like only piecewise
smooth in a way and you have the volume V enclosed inside the cube. So, examples like

585
that or you can have a sphere which does not have any discontinuity or something. So, you
can always consider a smooth surface as well or you can also consider a piecewise smooth.

Now, suppose 𝐹⃗ (𝑥, 𝑦, 𝑧) be a vector function of position which is continuous and has
.
⃗⃗ 𝑑𝑉 =
continuous first order partial derivatives; in V sorry on V then we can write ∭𝑉 𝑑𝑖𝑣𝑉
.
∬𝑆 𝐹⃗ . 𝑛̂𝑑𝑆, where 𝑛̂ is outward drawn unit normal vector to S.

So, that means, if you have a given volume V which is bounded by a piecewise smooth
surface actually say S and if you have a vector function 𝐹⃗ (𝑥, 𝑦, 𝑧) which is defined on V
then in that case of course, it also needs to have continuous first order partial derivatives
. .
so, that you can write this thing. Then in that case ∭𝑉 𝑑𝑖𝑣𝐹⃗ 𝑑𝑉 = ∬𝑆 𝐹⃗ . 𝑛̂𝑑𝑆 or in terms of
. .
⃗⃗. 𝐹⃗ 𝑑𝑉 = ∬ 𝐹⃗ . 𝑛̂𝑑𝑆.
notation you can write ∭𝑉 ∇ 𝑆

So, this is the required Gauss divergence formula and you need to have just a continuous
first order partial derivatives for this vector function 𝐹⃗ right, so, this is 𝐹⃗ . So, it is not V
this is𝐹⃗ .

. .
So, ∭𝑉 𝑑𝑖𝑣𝐹⃗ 𝑑𝑉 = ∬𝑆 𝐹⃗ . 𝑛̂𝑑𝑆 . So, similarly this is the divergence of F I thought it is the
. .
function V, but it is actually 𝐹⃗ . So, here you take a ∭𝑉 𝑑𝑖𝑣𝐹⃗ 𝑑𝑉 = ∬𝑆 𝐹⃗ . 𝑛̂𝑑𝑆 and similarly
. .
⃗⃗. 𝐹⃗ 𝑑𝑉 = ∬ 𝐹⃗ . 𝑛̂𝑑𝑆. So, this is a very
in terms of notations you can write this ∭𝑉 ∇ 𝑆

important theorem in vector calculus and it has also different how to say varieties in a way
I mean you can reformulate these the this equation in several other forms.

586
(Refer Slide Time: 18:29)

So, if we assume that F has 3 parts; so, alternatively alternative forms, so, if I assume that
our vector function F has 3 components 𝐹⃗ = 𝐹1 𝑖̂ + 𝐹2 𝑗̂ + 𝐹3 𝑘̂ let us say and if 𝛼, 𝛽 & 𝛾 are
the angles which outward drawn unit normal 𝑛̂ makes with positive direction with positive
directions of x, y and z axes.

Then obviously, we know that cos 𝛼 , cos 𝛽 and cos 𝛾 are the direction cosines. This is
from our 3D geometry you can have a look in any standard book on 3D geometry. So,
these are the direction cosines and of 𝑛̂ and we can be able to write and we have; and we
have; and we have and so, this is and we have our 𝑛̂ = cos 𝛼 𝑖̂ + cos 𝛽 𝑗̂ + cos 𝛾 𝑘̂.

So, we can be able to write our 𝑛̂ in this fashion. So, therefore, from here 𝐹⃗ . 𝑛̂ = 𝐹1 cos 𝛼 +
𝐹2 cos 𝛽 + 𝐹3 cos 𝛾 and our required Gauss divergence formula let us call it equation 1.
So, there from equation 1, we will have from 1, we will have volume integral divergence
of 𝐹⃗ . So, when I am charging the divergence operator on 𝐹⃗ then it will be
. 𝜕𝐹 𝜕𝐹2 𝜕𝐹 . .
∭𝑉( 𝜕𝑥1 + +. 𝜕𝑧3 )𝑑𝑥𝑑𝑦𝑑𝑧 = ∬𝑆 𝐹⃗ . 𝑛̂𝑑𝑆 = ∬𝑆(𝐹1 cos 𝛼 + 𝐹2 cos 𝛽 + 𝐹3 cos 𝛾)𝑑𝑆 =
𝜕𝑦
.
∬𝑆(𝐹1 𝑑𝑦𝑑𝑧 + 𝐹2 𝑑𝑧𝑑𝑥 + 𝐹3 𝑑𝑥𝑑𝑦)

So, this is the right hand side or the surface integral part. So, yes, so, this is how we express
this in terms of the Cartesian coordinate system.

587
So, now nothing is in the vector form everything is in the Cartesian coordinate system and
based on this given form we can express this volume integral this in this way. So, this
volume integral has a very wide application in vector calculus and in other fields as well.
So, will learn about those things later on, but today we will practice a few examples.

(Refer Slide Time: 22:51)

So, to start with let me consider our first example. So, for any closed surface let us consider
.
the first example for any closed surface closed surface S prove that ∬𝑆 𝑐𝑢𝑟𝑙𝐹⃗ . 𝑛̂ = 0. So,

this is what we have to prove. So, since we have 𝑐𝑢𝑟𝑙𝐹⃗ ; that means that the function 𝐹⃗ has
continuous partial derivatives and the function itself is continuous. So, we have to prove
this therefore, by divergence theorem or gauss divergence theorem what we will have
. . . .
∬𝑆 𝑐𝑢𝑟𝑙𝐹⃗ . 𝑛̂𝑑𝑆 = ∭𝑉 𝑑𝑖𝑣(𝑐𝑢𝑟𝑙𝐹⃗ )𝑑𝑉 = ∭𝑉 ∇
⃗⃗. (∇
⃗⃗ × 𝐹⃗ )𝑑𝑉 = ∭ 0𝑑𝑉 = 0
𝑉

right.

So, you see instead of doing this complicated calculation we use a Gauss divergence
theorem and we arrived from here to here and this is we know how to calculate. So, this
we calculate and by doing that our answer is 0. So, of course, using Gauss divergence
theorem is very how to say beneficial here.

588
(Refer Slide Time: 25:23)

Now, let me consider an another example. So, example let us say 2, if 𝐹⃗ (𝑥, 𝑦, 𝑧) = 𝑎𝑥𝑖̂ +
. 4
𝑏𝑦𝑗̂ + 𝑐𝑧𝑘̂ where a, b, c are constants show that ∬𝑆 𝐹⃗ . 𝑛̂𝑑𝑆 = 3 𝜋(𝑎 + 𝑏 + 𝑐), where S is

the unit sphere alright is the surface of course, S is the surface of unit sphere.

So, here we are given to evaluate the surface integral if we really want to do that we can
do that by taking 𝐹⃗ . 𝑛̂. So, we have to calculate 𝑛̂ from the surface of the sphere. So, the
equation of the sphere can be written 𝑥 2 + 𝑦 2 + 𝑧 2 = 1 from there we can calculate grad
of that expression and then we can calculate n we can calculate 𝑛̂ taking the projection and
all that. So, of course, we can do that.

However, if we want to use the divergence theorem then by divergence theorem we can
. . . .
write ∬𝑆 𝐹⃗ . 𝑛̂𝑑𝑆 = ∭𝑉 𝑑𝑖𝑣𝐹⃗ 𝑑𝑉 = ∭𝑉(𝑎 + 𝑏 + 𝑐)𝑑𝑉 = (𝑎 + 𝑏 + 𝑐) ∭𝑉 1𝑑𝑉 = (𝑎 +
4
𝑏 + 𝑐)𝑉 = 3 𝜋(𝑎 + 𝑏 + 𝑐)

Now, V is the volume of that sphere. So, when you are integrating the constant function;
that means, you are basically getting the volume V, right. So, this is what we will get
actually. So, well get basically the volume V, but the volume V is the volume of that sphere
4
and the volume of that unit sphere in this case would be 𝜋. 1. So, this is basically
3
4
𝜋(𝑎 + 𝑏 + 𝑐) you see just a simple application of divergence theorem has reduced our
3

effort in such an I mean nice manner.

589
So, we really do not have to calculate any n or do not have to take any projection or
anything like that we just had to use a Gauss divergence theorem and by that Gauss
divergence theorem everything reduced into a four line calculation. So, this divergence
theorem is proven to be a very handy tool and we will practice few more examples just to
show you how nice and convenient this and Gauss divergence theorem is.

So, today we will stop here and in our next class we will continue with the examples on
they were just theorem. So, I thank you for your attention and I will see you in the next
class.

590
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture - 58
Gauss divergence theorem

Hello students. So, in the previous class, we started with volume integral and then we gave
I gave you the different the statement of Gauss divergence theorem and then we practiced
one or two examples. But since it is a very important topic in vector calculus, we will
continue practicing few more examples on Gauss divergence theorem. So, let me start with
our first example.

(Refer Slide Time: 00:44)

So, example one for today.

.
So, our first example is evaluate ∬𝑆(𝑥 3 𝑑𝑦𝑑𝑧 + 𝑥 2 𝑦𝑑𝑧𝑑𝑥 + 𝑥 2 𝑧𝑑𝑥𝑑𝑦) where S is the
closed surface bounded by the planes 𝑧 = 0, 𝑧 = 0 and 𝑥 2 + 𝑦 2 = 𝑎2 . So, basically we
have a cylinder. So, here the given surface integral is this one. Of course, if you want to
evaluate the surface integral; it would be very complicated because we have to evaluate
these three terms actually and we have to calculate the limits for whenever we are
integrating.

591
So, here we have to calculate limits for x, y and z; for z it is given. So, its relatively
complicated, but if we look into the Cartesian form of Gauss divergence theorem there we
had ; so, we can go back.

(Refer Slide Time: 02:35)

And there we had 𝐹1 , 𝐹2 , 𝐹3 and we can write this thing as a.

So, now, if I look into the formula where is that if I look into this function here. So, this is
our 𝐹1 , this is our 𝐹2 this must be dz dx and this is our 𝐹3 .

So, from Cartesian form of divergence theorem, we have basically a surface integral
. . 𝜕𝐹 𝜕𝐹2 𝜕𝐹3
∬𝑆(𝑥 3 𝑑𝑦𝑑𝑧 + 𝑥 2 𝑦𝑑𝑧𝑑𝑥 + 𝑥 2 𝑧𝑑𝑥𝑑𝑦) = ∭𝑉 ( 𝜕𝑥1 + 𝜕𝑦
+ 𝜕𝑧
) 𝑑𝑥𝑑𝑦𝑑𝑧 where our 𝐹1 =

𝑥 3 , 𝐹2 = 𝑥 2 𝑦, 𝐹3 = 𝑥 2 𝑧.

.
So, if I substitute all these things here. So, this will be ∭𝑉(3𝑥 2 + 𝑥 2 + 𝑥 2 )𝑑𝑥𝑑𝑦𝑑𝑧 =
. 𝑎 √𝑎2 −𝑥 2 𝑏
5 ∭𝑉 𝑥 2 𝑑𝑥𝑑𝑦𝑑𝑧 = 5 ∫𝑥=−𝑎 ∫𝑦=−√𝑎2 −𝑥2 ∫𝑧=0 𝑥 2 𝑑𝑥𝑑𝑦𝑑𝑧 =
𝑎 √𝑎2 −𝑥2 𝑏
5.4 ∫𝑥=0 ∫𝑦=0 ∫𝑧=0 𝑥 2 𝑑𝑥𝑑𝑦𝑑𝑧.

592
(Refer Slide Time: 05:24)

So, this Cartesian form of Gauss divergence theorem actually made this term or this
integral this integral to reduce to a fairly simple integrand which is basically
.
5 ∭𝑉 𝑥 2 𝑑𝑥𝑑𝑦𝑑𝑧.

Now, we have to guess the limits. So, in order to guess the limits z is actually varying from
0 to b. So, guessing the limits for z is not complicated. Now y will vary from −√𝑎2 − 𝑥 2 to
√𝑎2 − 𝑥 2 and x will vary from −𝑎 to a.

So, basically we take both plus and minus value and x is varying from -a to plus a. So, first
of all we can integrate with respect to with respect to x, but here we can see that there is
symmetry in a way. So, instead of calculating the area in the whole in the whole circle, we
can actually calculate in one of the halves actually.

𝑎 √𝑎2 −𝑥 2 𝑏
So, this will become 5.4 ∫𝑥=0 ∫𝑦=0 ∫𝑧=0 𝑥 2 𝑑𝑥𝑑𝑦𝑑𝑧. Because whatever you get the area

in this one half in a way is actually the four time and multiplied by four times and that will
be there that will be the whole area in a way or volume integral in this case in the whole
cylinder. So, basically we took the four times of that volume in that one particular quadrant
in a way. So, therefore, we had to add a 4 here. Here now we can integrate.

𝑎 2 2
So, this is basically 20𝑏 ∫𝑥=0 𝑥 2 [𝑦]√𝑎
𝑦=0
−𝑥
𝑑𝑥 and or we can integrate with respect to x first;
sorry we are integrating with respect to y. and then this will be 20 b times.

593
 𝑎
So, this will be integral 20𝑏 ∫𝑥=0 𝑥 2 √𝑎2 − 𝑥 2 𝑑𝑥. Now we have to evaluate this integral.
So, evaluating this integral is not complicated, you just have to substitute 𝑥 = sin 𝜃 and
then convert the whole thing in some polar coordinates and then you will basically be able
to evaluate this integral; it just that it is slightly lengthy. So, I am leaving this task up to
the students. I am pretty sure you can be able to do that and finally, the answer would be
5
𝜋𝑎4 𝑏. So, this will be the answer.
4

So, other than guessing the limits the rest of the things are pretty much same what we have
studied before. So, obtaining this limit is not complicated. So, I am pretty sure you can be
able to do that. So, this is how we just using the Gauss divergence theorem, you see the
whole integral is simplified to a simple volume integral where of course, now here you
have to do some complicated calculation a slightly complicated. So, it is simple from here
to here and yeah this is one of the applications of Gauss divergence theorem we will
practice the next example now.

(Refer Slide Time: 10:06)

So, let me consider an another example. So, evaluate example 2, I guess. So, evaluate
.
surface integral ∬𝑆 𝑥 2 𝑑𝑦𝑑𝑧 + 𝑦 2 𝑑𝑧𝑑𝑥 + 2𝑧(𝑥𝑦 − 𝑥 − 𝑦)𝑑𝑥𝑑𝑦 where S is the surface of
the cube 0 ≤ 𝑥 ≤ 1,0 ≤ 𝑦 ≤ 1 & 0 ≤ 𝑧 ≤ 1. So, here of course, it again falls into that
Cartesian category of Gauss divergence theorem. And if we compare like previous
example, then this is our 𝐹1 = 𝑥 2 , 𝐹2 = 𝑦 2 , 𝐹3 = 2𝑧(𝑥𝑦 − 𝑥 − 𝑦). So, I am not writing all
those things; I am just writing by Gauss divergence theorem or simply by divergence

594
theorem sometimes I am using small d. So, do not get confused. So, you can write small d
or capital D; it is up to you.

So, our 𝐼𝑆 which is basically the surface integral will reduce to our volume integral 𝐼𝑉 =
. .
∭𝑉(2𝑥 + 2𝑦 + 2𝑥𝑦 − 2𝑥 − 2𝑦)𝑑𝑥𝑑𝑦𝑑𝑧 = 2 ∭𝑉 𝑥𝑦𝑑𝑥𝑑𝑦𝑑𝑧.

So, now, here we have to substitute the values for x y and z.

(Refer Slide Time: 12:18)

1 1 1
So, since it is a volume integral,2 ∫𝑥=0 ∫𝑦=0 ∫𝑧=0 𝑥𝑦𝑑𝑥𝑑𝑦𝑑𝑧. So, first we integrate with
1 1
respect to z and the value will be 2 ∫𝑥=0 ∫𝑦=0 𝑥𝑦𝑑𝑥𝑑𝑦. So, this is just 2 times surface
1 1 𝑥2
integral. Now we integrate with respect to y and then this will be 2. 2 ∫𝑥=0 𝑥𝑑𝑥 = [ 2 ]10 =
1
.
2

So, this is basically half. So, you see initially we had a very how to say complicated
expression and we had to evaluate the surface integral, but all of these things are just
algebraic expression and therefore, they must have continuous partial derivatives.

So, just applying the Gauss divergence theorem, we were able to obtain this form here and
from there just substitute the values of x y and z and then that will give us the required
answer which is half in this case. So, like this we can practice many examples. So, I have
some examples in my lecture note we can. So, let me consider an another example alright.

595
(Refer Slide Time: 14:00)

So, here we have an another interesting example; so, example 3 I believe. So, evaluate
.
surface integral of ∬𝑆 𝐹⃗ . 𝑛̂𝑑𝑆 over the entire surface of the region above the xy plane

bounded by the cone 𝑧 2 = 𝑥 2 + 𝑦 2 and the plane 𝑧 = 4. If this 𝐹⃗ = 4𝑥𝑧𝑖̂ + 𝑥𝑦𝑧 2 𝑗̂ + 3𝑧𝑘̂ .
So, here we have to evaluate the surface integral where we have the given vector function
and the given surface basically for this surface integral is actually the region above xy
plane bounded by this cone and the plane 𝑧 = 4 actually.

So, the terms in this vector function, they are all algebraic functions and basically in a way
they are product of xy and z square. So, they have continuous partial derivatives. So, we
.
can use Gauss divergence theorem. So, by divergence theorem; we have ∬𝑆 𝐹⃗ . 𝑛̂𝑑𝑆 =
.
∭𝑉 𝑑𝑖𝑣𝐹⃗ 𝑑𝑉.

So, 𝑑𝑖𝑣𝐹⃗ would be volume integral divergence. If we take the divergence here then this
.
will be ∭𝑉(4𝑧 + 𝑥𝑧 2 + 3)𝑑𝑥𝑑𝑦𝑑𝑧. You just have to take the divergence of this function
and then you obtain this limit here. Now in this case we have to get the limits for x y and
z and we have to remember that the volume enclosed by this cone and the plane has to be
4 𝑧 √𝑧 2 −𝑦 2
above xy plane in a way. So, the limit for z would be then; so,∫𝑧=0 ∫𝑦=−𝑧 ∫𝑥=−√𝑧 2 −𝑦2(4𝑧 +

𝑥𝑧 2 + 3)𝑑𝑥 𝑑𝑦 𝑑𝑧

596
. And now we can instead of calculating this whole integral, I can calculate in the upper
half.

(Refer Slide Time: 17:50)

4 𝑧 √𝑧 2 −𝑦 2
So, then in that case this will be 2 ∫𝑧=0 ∫𝑦=−𝑧 ∫𝑥=0 (4𝑧 + 3)𝑑𝑥 𝑑𝑦 𝑑𝑧.

Now if we assume that our if so, this function basically is an odd function in x. So,
therefore, that will be 0 and sum of these two function and so forth that plus 3 is an even
function. So, basically I have written a 2 here. So, that 𝑥𝑧 2 is vanished because of being
an odd function right.

So, that is taken care of here and now what we are going to do? We are going to integrate
4 𝑧
first with respect to x. So, this will reduce to 2 ∫𝑧=0 ∫𝑦=−𝑧(4𝑧 + 3)√𝑧 2 − 𝑦 2 𝑑𝑦 𝑑𝑧.

So, this will be or I can write this as 0 to z and then put a 2 here again. So, this is
4 𝑧 4 𝑦
2.2 ∫𝑧=0 ∫𝑦=0(4𝑧 + 3)√𝑧 2 − 𝑦 2 𝑑𝑦 𝑑𝑧 = 4 ∫𝑧=0(4𝑧 + 3)[ 2 √𝑧 2 − 𝑦 2 +
𝑧2 𝑦 𝑧 4 𝑧2
sin−1 𝑧 ]𝑦=0 𝑑𝑧 = 4 ∫𝑧=0(4𝑧 + 3) ( 2 sin−1 1) 𝑑𝑧 right dz.
2

𝜋 𝜋
So,sin−1 1 = 2 . So, I will take outside and then there is so, this 4 will be gone. And
2
4
therefore, we will have simply this integral 𝜋 ∫0 (4𝑧 + 3)𝑧 2 𝑑𝑧 = 320𝜋 here.

597
After integration so, ultimately if we do the integration, then this will be 320𝜋. So, you
see initially we had this surface integral to evaluate, we could have done that we could
have calculated the normal for this surface and then we could have taken the projection
and things like that, but it would have lead to a little bit complicated calculation. So, instead
of doing that, we took the help of Gauss divergence theorem and with the help of Gauss
divergence theorem, we can be able to see that we just have to calculate these limits and
then do this simple calculation use the odd and even property of this function here. So,
since it is an odd function that is why this term is vanished here and the rest of the functions
are even function.

So, we I put a 2 here, then I integrated with respect to x, then integrate it with respect to y
and then integrate it with respect to z. So, this is actually fairly simple to do instead of
doing that complicated surface integral. So, this is another example or application of Gauss
divergence theorem. You might also be asked to verify. So, when you are asked to verify
that is when the examples becomes very lengthy and hopefully let us assume that still you
will be not asked to verify. But sometimes you might and then in that case you have to
evaluate the both surface integral and the volume integral.

(Refer Slide Time: 22:36)

So, let me give you one example where you have to verify example 4; I believe. So, verify
the divergence theorem for 𝐹⃗ (𝑥, 𝑦, 𝑧) = (𝑥 2 − 𝑦𝑧)𝑖̂ + (𝑦 2 − 𝑧𝑥)𝑗̂ + (𝑧 2 − 𝑥𝑦)𝑘̂ taken

598
over the parallelepiped 0 ≤ 𝑥 ≤ 𝑎, 0 ≤ 𝑦 ≤ 𝑏, 0 ≤ 𝑧 ≤ 𝑐. So, the solution; so, first of all
let us draw our parallelepiped.

So, this is the origin that is my x axis that is y axis, this is z axis and if I draw then this is
all right and this is G, this is F A B. So, this is A B C D E F and G alright. So, we have 8
faces right yes all right. So, now, we have to verify.

. .
So, by divergence theorem, we verify that ∬𝑆 𝐹⃗ . 𝑛̂𝑑𝑆 = ∭𝑉 𝑑𝑖𝑣𝐹⃗ 𝑑𝑉 right. So, let me
.
calculate the right hand side. So, the right hand side Rhs equals ∭𝑉 𝑑𝑖𝑣𝐹⃗ 𝑑𝑉 .

. 𝜕 𝜕 𝜕
So, we basically have ∭𝑉[𝜕𝑥 (𝑥 2 − 𝑦𝑧) + 𝜕𝑦 (𝑦 2 − 𝑧𝑥) + 𝜕𝑧 (𝑧 2 − 𝑥𝑦)]𝑑𝑥𝑑𝑦𝑑𝑧 =
. 𝑎 𝑏 𝑐
2 ∭𝑉(𝑥 + 𝑦 + 𝑧)𝑑𝑥 𝑑𝑦𝑑𝑧 = 2 ∫𝑥=0 ∫𝑦=0 ∫𝑧=0(𝑥 + 𝑦 + 𝑧)𝑑𝑥 𝑑𝑦 𝑑𝑧.

(Refer Slide Time: 25:47)

So, I can write those limits. And then we integrate first with respect to x, then with respect
to y and then with respect to z.

So, it is a fairly easy thing to do and therefore, this will ultimately give you
𝑎2 𝑏𝑐 𝑎𝑏 2 𝑐 𝑎𝑏𝑐 2
2[ + + ] = 𝑎𝑏𝑐(𝑎 + 𝑏 + 𝑐). So, this is the required volume integral. Now
2 2 2

we have to verify whether the left hand side is also equal to that volume integral or not
now here is the interesting part. This is not a very simple surface integral to evaluate. Here
we have basically 8 surfaces.

599
So, on every surface of this parallelepiped, we have to evaluate the surface integral and on
every surface we have to calculate this unit normal and for every surface, we just substitute
that unit normal and then we do the calculation of this surface integral. So, let me give you
an example over the surface. So, basically our surface integral would be sum of 8 sub
surface integrals.

So, surface integral on this part surface integral on this part, this part that other side this
side and the downside of that parallelepiped so, basically there will be 8 surface integrals
sub surface integrals I would say and when you sum them then that is when you obtain
this surface integral here.

(Refer Slide Time: 27:59)

So, basically what would happen is let me give write this term is equals to there will be
. . . .
∬𝑆 𝐹⃗ . 𝑛̂𝑑𝑆 = ∬𝑆 𝐹⃗ . 𝑛̂𝑑𝑆 + ∬𝑆 𝐹⃗ . 𝑛̂𝑑𝑆 +. . . + ∬𝑆 𝐹⃗ . 𝑛̂𝑑𝑆 all right. So, let us call this
1 2 6

phase facing front is our 𝑆1 . So, over 𝑆1 or DEFG, our 𝑛̂ = 𝑖̂ and 𝑥 = 𝑎 right. So,
. 𝑐 𝑏
∬𝑆 𝐹⃗ . 𝑛̂𝑑𝑆 = ∫𝑧=0[∫𝑦=0[(𝑎2 − 𝑦𝑧)𝑖̂ + (𝑦 2 − 𝑧𝑎)𝑗̂ + (𝑧 2 − 𝑎𝑦)𝑘̂]. 𝑖̂𝑑𝑦𝑑𝑧 =
1

𝑐 𝑏 𝑐 2 𝑏2
∫𝑧=0 ∫𝑦=0(𝑎2 − 𝑦𝑧)𝑑𝑦𝑑𝑧 = 𝑎2 𝑏𝑐 − 4
.

Next we will integrate on the other side.

So, if DFG was this face, then we will be now integrate with this on the on the surface a
OBC and if we integrate on that other surface, then instead of taking i we will take minus

600
-i and proceed in the similar fashion. Of course, we have to take 𝑥 = 0 because that is the
plane 𝑥 = 0 substitute here. So, over let us say I am calling that one as 𝑆2 or AOCB our
. 𝑐 𝑏 𝑏 𝑐 2 2
∬𝑆 𝐹⃗ . 𝑛̂𝑑𝑆 = ∫𝑧=0 ∫𝑦=0 𝑦𝑧𝑑𝑦𝑑𝑧 = 4 .
2

So, you see when you sum these integrals, then there will be some cancellation.
. . . .
So, ∬𝑆 𝐹⃗ . 𝑛̂𝑑𝑆 = ∬𝑆 𝐹⃗ . 𝑛̂𝑑𝑆 + ∬𝑆 𝐹⃗ . 𝑛̂𝑑𝑆 +. . . + ∬𝑆 𝐹⃗ . 𝑛̂𝑑𝑆 = 𝑎𝑏𝑐(𝑎 + 𝑏 + 𝑐).
1 2 6

So, 12 then 3 4 and then 56 there will be 6 integrals. So, on these 6 surfaces we basically
obtain 6 integrals and then you sum them you will have some cancellation.

(Refer Slide Time: 31:30)

and when you write all these and then there will be some cancellation and when you sum
them, then its 𝑎𝑏𝑐(𝑎 + 𝑏 + 𝑐).

So, this is the required surface integral and therefore, this verification process shows that
the Gauss divergence theorem is verified. So, we will stop here for today and will continue
with our further integral so, in vector calculus terms. So, next we will look into stokes
theorem. So, I think we have practiced enough examples on Gauss divergence theorem.
So, next we will start looking into Gauss stokes theorem. So, I thank you for your attention
and I look forward to you in your next class.

601
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture - 59
Stoke's Theorem

Hello students. So, in the previous class, we started with Gauss divergence theorem and
we also looked into the statement of Gauss divergence theorem and how we can convert a
surface integral into a volume integral and then a volume integral back into the surface
integral. So, whichever is given, if in your question if you are asked to evaluate the other
one, you can use the Gauss divergence theorem. And since we have only two more lectures
left, today we will start with one example on Gauss divergence theorem and then we move
to our final theorem in vector calculus which is basically stokes theorem and unfortunately
we do not have any more lectures left.

So, we may not be able to practice examples from the integral calculus section or even
from the introduction of the vector calculus part, but we will try to solve at least two or
three examples on stokes theorem just to make the concepts clear. So, today, we will start
with one example on Gauss divergence theorem.

(Refer Slide Time: 01:31)

602
So, in the statement from the Gauss divergence theorem, you remember that you can write
if 𝐹⃗ is any vector function, then you can write the volume integral over the divergence of
. .
let us say our vector function is 𝐹⃗ , then in that case I can write ∭𝑉 𝑑𝑖𝑣𝐹⃗ 𝑑𝑉 = ∬𝑆 𝐹⃗ . 𝑛̂𝑑𝑆.
.
So, this is basically our, I am sorry. So, this one is just ∬𝑆 𝐹⃗ . 𝑛̂𝑑𝑆.

So, this is basically our Gauss divergence theorem. And with the help of Gauss divergence
term, you can be able to convert or if you are given to evaluate a surface integral, you can
just convert it into a volume integral and the things will become very straightforward. So,
we will solve one example based on this Gauss divergence theorem.

So, the first example is for today, so the first example is let us say by converting the surface
.
integral into a volume integral, evaluate surface integral ∬𝑆(𝑥 3 𝑑𝑦𝑑𝑧 + 𝑦 3 𝑑𝑧𝑑𝑥 +
𝑧 3 𝑑𝑥𝑑𝑦) where S is the surface of 𝑥 2 + 𝑦 2 + 𝑧 2 = 1. So, this is basically our sphere. And
we have to convert the surface integral which is to be evaluated on the surface of the sphere
into a volume integral and then we have to calculate the value of this surface integral. So,
we know that from Gauss divergence theorem, we know that, so we have learnt this
formula.

So, there are like two or three expressions for Gauss divergence theorem and one of them
is this one another one is with d x, d y, d y, d z and d y d x form. So, we are going to write
.
that form here. So, by divergence theorem we know that, ∬𝑆(𝐹1 𝑑𝑦𝑑𝑧 + 𝐹2 𝑑𝑧𝑑𝑥 +
. 𝜕𝐹 𝜕𝐹2 𝜕𝐹3
𝐹3 𝑑𝑥𝑑𝑦) = ∭𝑉( 𝜕𝑥1 + + )𝑑𝑥𝑑𝑦𝑑𝑧.
𝜕𝑦 𝜕𝑧

So, this is the required how to say form in terms of for Cartesian coordinates like x, y and
z. So, now, we can compare. So, you see this surface integral here and this surface integral
here, if we compare these two forms, then in that case our 𝐹1 (𝑥, 𝑦, 𝑧) = 𝑥 3 , 𝐹2 (𝑥, 𝑦, 𝑧) =
𝑦 3 , 𝐹3 (𝑥, 𝑦, 𝑧) = 𝑧 3 right. So, here V is the volume enclosed by the surface S.

Since 𝐹1 is a function of x only, y is 𝐹2 is a function of y only and 𝐹3 is a function of z

only, it makes our life a lot easier because it may happen that here 𝐹3 could be a product
of x y z and their power and things like that. So, then it would have become a little bit
complicated. But since here everything is just like a function of one variable only, then
this is a fairly simple example to do. So, now, by the divergence theorem let me name this
equation as 1. So, from equation 1 if we convert this surface integral, into a volume

603
integral, then it will take this form like an equation 1. And in equation 1, V is the volume
𝜕𝐹1 𝜕𝐹2
enclosed by this fair S which is given here and we have to evaluate = 3𝑥 2 , =
𝜕𝑥 𝜕𝑦
𝜕𝐹3
3𝑦 2 , = 3𝑧 2
𝜕𝑧

, alright. So, let us substitute all these values here.

(Refer Slide Time: 07:02)

So, from 1 and let me call this equation as 2. So, this is as 2. So, from 1 and 2, we have
.
volume integral over V 3 ∭𝑉(𝑥 2 + 𝑦 2 + 𝑧 2 )𝑑𝑥𝑑𝑦𝑑𝑧. Now, this is basically volume
integral and if we substitute this as 𝑥 2 + 𝑦 2 + 𝑧 2 = 1,

So, and then d x d y d z is basically dv, so that is the volume element. So, I can write it as
dv alright and this means that we are basically doing or calculating the volume of that
sphere right. So, if your vector function is 1 like in this case and I mean not vector function.
So, if your function overall is 1 in a way, then in that case here this means that you are
actually doing a volume or calculating the volume of the given surface. So, that is what
this mean and here in our case, the given surface is a sphere of unit radius.

So, in that case we can write 4 by 3 or what we can do is instead of substituting, there is
another method which we can follow.

604
(Refer Slide Time: 09:14)

So, what we can do? Instead of substituting this one, I can substitute just to make things a
little bit more clear, I can substitute 𝑥 = 𝑟 cos 𝜃 cos 𝜑 , 𝑦 = 𝑟 cos 𝜃 sin 𝜑 , 𝑧 = 𝑟 sin 𝜃.
.
Then in that case, this will be 3 ∭𝑉 𝑟 2 𝑟 2 sin 𝜃 𝑑𝑟𝑑𝜃𝑑𝜑.

1 𝜋 2𝜋
So, now for this V, what we can do is we can write 3 ∫𝑟=0 ∫𝜃=0 ∫𝜑=0 𝑟 4 sin 𝜃 𝑑𝑟𝑑𝜃𝑑𝜑. So,

we can integrate because they are not product they are not in a way involving 𝜃 and r
together. So, we can integrate with respect to r and we can integrate for 𝜃 for the integral
with respect to 𝜃.

1 𝜋 2𝜋 1
So that means, this will reduce to 3 ∫𝑟=0 ∫𝜃=0 𝑟 4 sin 𝜃 [∫0 𝑑𝜑] 𝑑𝑟 𝑑𝜃 = 3 × 2𝜋 × 2 5.

12𝜋
So, ultimately it will be . So, you see initially we had a very how to say complicated
5

expression where it is not that much complicated but still we had to calculate the surface
integral. So, in the question itself it says that convert the surface integral into a volume
integral. So, if we are asked to convert this surface integral to a volume integral, we just
have to use this divergence theorem. And in that divergence theorem if you compare, then
this is basically our 𝐹1 , this is basically 𝐹2 and this is basically 𝐹3 . And on the right hand
side of that divergence theorem, we had to calculate the partial derivatives. So, we
calculated the partial derivatives, we substituted these partial derivatives here and then we
are just substituting taking the help of a spherical polar coordinate system.

605
So, we substitute 𝑥 = 𝑟 cos 𝜃 cos 𝜑 , 𝑦 = 𝑟 cos 𝜃 sin 𝜑 , 𝑧 = 𝑟 sin 𝜃. So, the volume
element will be 𝑑𝑟𝑑𝜃𝑑𝜑(𝑟 sin 𝜃) and then we have here 𝑟 2 basically. So, we so this is our
basically volume element and that is the function with what we are getting from this 𝑥 2 +
𝑦 2 + 𝑧 2 . So, all together this is the 𝑟 4 , this is 𝑑𝑟𝑑𝜃𝑑𝜑(𝑟 sin 𝜃) and the limit for r is 0 to 1
for 𝜃 is 0 to 𝜋 and 𝜑 is 0 to 2𝜋. Then when you integrate, you obtain this integral the value
of this surface integral. So, that is how we use the Gauss divergence theorem.

So, whenever you are given a surface integral, you can use Gauss divergence theorem to
convert it into a volume integral and it will become very easy. Because when you are
evaluating surface integral, in some cases you may have a parallelepiped, cube or a cube
and then in that case, you have to do the surface integral on every surface. However, if you
convert that surface integral into a volume integral and the light will become much easier,
then in that case we just have to get the limits for x, y and z and do that calculate that
divergence and hopefully the whole thing will become very easy. So, this is one of the
examples motivated from Gauss divergence theorem. And now we move to our final
theorem which is basically stokes theorem. So, I am going to write the statement first.

(Refer Slide Time: 14:40)

So, Stoke’s theorem, so let me have a look at the statement in my lecture note. So, the
Stoke’s theorem says let S be a piecewise smooth open surface bounded by a piecewise
simple closed curve C and let 𝐹⃗ (𝑥, 𝑦, 𝑧) be a continuous vector function which has
continuous first order partial derivatives in a region of space which contains S in it is

606
 . . .
interial, then ∮𝐶 𝐹⃗ . 𝑑𝑟⃗ = ∬𝑆 𝑐𝑢𝑟𝑙𝐹⃗ . 𝑛̂𝑑𝑆 = ∬𝑆(∇
⃗⃗ × 𝐹⃗ ). 𝑛̂𝑑𝑆 or we can write it as where 𝑛̂

is a unit outward drawn normal alright. So, this is the required statement and of course S
so if you are walking along the curve C, then your surface S must lie on the left hand side.

So, that is how we choose the curve C that the orientation is always in the anticlockwise
direction. So, that is the property of this curve C and n is outward drawn normal. So, if
you are walking in the anticlockwise direction, then your surface will always lie on the
left. That is how we mean by closed, that is what we mean by this piecewise smooth
bounded by a simple closed curve C. And 𝑛̂ is actually the outward or normal on the
. .
surface S and this then in that case you can write the, so ∮𝐶 𝐹⃗ . 𝑑𝑟⃗ = ∬𝑆 𝑐𝑢𝑟𝑙𝐹⃗ . 𝑛̂𝑑𝑆 =
.
⃗⃗ × 𝐹⃗ ). 𝑛̂𝑑𝑆. So, this is also a very important theorem in vector calculus and also in
∬𝑆(∇
applied it several other branches of applied mathematics where we take help of this
theorem.

So, for example, in partial differential equations or even in fluid mechanics and these
theorems have a very wide application. So, this is required statement and we can also write
this theorem in a Cartesian form. So, I have it in my lecture notes here. So, if I write it as
a Cartesian form, then in that case we can write it as let. So, alternatively or a different
form basically, so alternative or a different form a different form. So, basically let 𝐹⃗ =
𝐹1 𝑖̂ + 𝐹2 𝑗̂ + 𝐹3 𝑘̂ and let 𝑛̂ the outward drawn normal, the outward drawn normal makes
angle 𝛼, 𝛽, 𝛾 with positive directions of x, y and z axes.

(Refer Slide Time: 20:15)

607
Then, we can write our normal 𝑛̂ = cos 𝛼 𝑖̂ + cos 𝛽 𝑗̂ + cos 𝛾 𝑘̂ and we know that 𝑐𝑢𝑟𝑙𝐹⃗ ,
𝑖̂ 𝑗̂ 𝑘̂
𝜕 𝜕 𝜕
if 𝐹⃗ has three components and 𝑐𝑢𝑟𝑙𝐹⃗ = | |.
𝜕𝑥 𝜕𝑦 𝜕𝑧
𝐹1 𝐹2 𝐹3

So, we basically take a dot product with this normal 𝑛̂ which is given here. So, on the left
.
hand side, basically we will have line integral. So, therefore, ∮𝐶(𝐹1 𝑖̂ + 𝐹2 𝑗̂ + 𝐹3 𝑘̂ ). (𝑑𝑥𝑖̂ +
.
𝑑𝑦𝑖̂ + 𝑑𝑧𝑘̂) = ∬𝑆(∇
⃗⃗ × 𝐹⃗ ). (cos 𝛼 𝑖̂ + cos 𝛽 𝑗̂ + cos 𝛾 𝑘̂)𝑑𝑆.

So, now, we can calculate the 𝑐𝑢𝑟𝑙𝐹⃗ easily and then we take the dot product and therefore,
.
the whole thing will reduce to and here it will be ∮𝐶(𝐹1 𝑑𝑥 + 𝐹2 𝑑𝑦 + 𝐹3 𝑑𝑧) =
. 𝜕𝐹 𝜕𝐹2 𝜕𝐹 𝜕𝐹3 𝜕𝐹 𝜕𝐹1
∬𝑆 [( 𝜕𝑦3 − 𝜕𝑧
) cos 𝛼 + ( 𝜕𝑧1 − 𝜕𝑥
) cos 𝛽 + ( 𝜕𝑥2 − 𝜕𝑦
) cos 𝛾 ]𝑑𝑆.

So, this is the required form of Stoke’s theorem in the Cartesian form. So, like these
theorem or Gauss divergence theorem, Stoke’s theorem also have also has very wide
application in vector calculus and we will practice a few examples motivated from stokes
theorem and then we will probably close this topic. So, let me just try to find out, ok. So,
here I have some examples.

(Refer Slide Time: 23:29)

.
So, example 1, I guess. So, first example is prove that ∮𝑐 𝑟⃗. 𝑑𝑟⃗ = 0. So, we have to show

that, so that line integral of this 𝑟⃗ is our given vector function in this case. So, 𝐹⃗ = 𝑟⃗ and

608
line integral over this curve C where C is again any piecewise smooth curve closed curve
. . .
of course, ∮𝐶 𝑟⃗. 𝑑𝑟⃗ = ∬𝑆(𝑐𝑢𝑟𝑙𝑟⃗). 𝑛̂𝑑𝑆 = ∬𝑆 0. 𝑛̂ 𝑑𝑆 = 0 right.

So, this is how we use the Gauss stokes theorem. Next, we can show that a similar type of
result. So, show that, so there are several show that type of problems which you can also
.
⃗⃗𝜑 . 𝑑𝑟⃗ = 0 . So, our given
solve with this with the help of stokes theorem. Show that ∫𝐶 𝜑∇

function 𝐹⃗ = 𝜑∇
⃗⃗𝜑.

. . .
⃗⃗𝜑 . 𝑑𝑟⃗ = ∬ 𝑐𝑢𝑟𝑙( 𝜑∇
So, here we can have ∫𝐶 𝜑∇ ⃗⃗𝜑). 𝑛̂𝑑𝑆 = ∬ [𝜑𝑐𝑢𝑟𝑙∇
⃗⃗𝜑 + ∇
⃗⃗𝜑 ×
𝑆 𝑆

⃗∇⃗𝜑]. 𝑛̂𝑑𝑆.

So, this is basically ⃗⃗

0 because 𝑎⃗ × 𝑎⃗ = ⃗⃗ ⃗⃗𝜑 = 0. So, this is also 0 and this is 0
0 and 𝜑𝑐𝑢𝑟𝑙∇
by definition. So, ultimately we have this as 0, this as 0. So, the integrand is 0 and
.
therefore,∬𝑆 0. 𝑛̂𝑑𝑆 = 0. and then we integrate over surface integral S let us say. So, this
.
⃗⃗𝜑 . 𝑑𝑟⃗ = 0.
is ultimately 0 so, ∫𝐶 𝜑∇

So, you see we did not have to go through any complicated calculation or something, we
just had to use Stoke’s theorem. That means, that you can convert your line integral into a
surface integral and then the rest of the simplification is pretty much straightforward. So,
these were the, to prove that examples that we solved today, but in our next class we will
consider at least one or two examples motivated from the stokes theorem where we might
need to verify the stokes theorem and we will start such examples in our next class.

So, I thank you for your attention today and I will see you in your next class.

609
 Integral and Vector Calculus
Prof. Hari Shankar Mahato
Department of Mathematics
Indian Institute of Technology, Kharagpur

Lecture - 60
Overview of Course

Hello students. So, in today’s lecture we will continue with some examples on Stokes
theorem and since today is our last lecture I apologize that I could not be able to cover
more examples on integral calculus section as I promised or showed you in our syllabus.
But I try to compute as many examples as I could also because the time is very limited and
because of that sometimes I also had to rush through and maybe one or two occasions we
could not be able to do more examples.

But I suggest you to follow any one of the textbooks or if possible, then follow all the
textbooks which I suggested in the reference what are the beginning of this course. And I
am pretty sure there you will be able to find a lot of examples just for you to practice and
get ready good at it and I am also providing a lot of assignments. So, try to solve those
assignments that will definitely help you clear out your doubts and would also help you to
become how to say very good at this particular subject.

So, today let us start with our last topic which is stokes theorem and I am going to solve
some examples and let us do that.

(Refer Slide Time: 01:40)

610
So, verify Stokes theorem for 𝐹⃗ (𝑥, 𝑦, 𝑧) = (2𝑥 − 𝑦)𝑖̂ − 𝑦𝑧 2 𝑗̂ − 𝑦 2 𝑧𝑘̂ where S is the upper
half surface of this sphere 𝑥 2 + 𝑦 2 + 𝑧 2 = 1 and C is its boundary. So, what we have
basically is a sphere and we have to verify the stokes theorem on the upper half of that
sphere so; that means, if you consider let us say x axis y axis and z axis, then it will it we
can consider the upper half as from 𝑧 = 0 till 𝑧 = 1 so; that means, that could be one
upper half or 𝑥 = 0 till 𝑥 = 1.

So, that can be another upper half. So, any one of the upper half basically and in this case
to do that first of all to verify actually the Stokes theorem; we write the statement. So, by
. .
Stokes theorem what do we have? We have line integral ∮𝐶 𝐹⃗ . 𝑑𝑟⃗ = ∬𝑆 𝑐𝑢𝑟𝑙𝐹⃗ . 𝑛̂𝑑𝑆 right.
So, we have to so that left hand side is equal to the right hand side.

.
So, let us start with the left hand side. So, its line integral ∮𝐶 𝐹⃗ . 𝑑𝑟⃗. So, here C is the
boundary in the xy plane and we are considering basically 𝑧 = 0 part. So, 𝑧 = 0 to 𝑧 =
1 and if C is the boundary then in xy plane then it is basically that circle right. So, that is
basically the circle and if you walk along that circle in the anti clockwise direction, then
your surface is falling on the left hand side, so, that makes sense.

So, let me invite one or two lines before I actually calculate this line integral. So, in order
to do that; so, let us write the boundary C of the surface S is the circle in xy plane of radius
unity and center origin and the parametric representation of that circle would be 𝑥 =
cos 𝑡 , 𝑦 = sin 𝑡 & 𝑧 = 0. So, now, the line integral and t is basically 𝑡 ∈ [0,2𝜋] alright.
. .
So, now, we have line integral ∮𝐶 𝐹⃗ . 𝑑𝑟⃗ = ∮𝐶 [ (2𝑥 − 𝑦)𝑖̂ − 𝑦𝑧 2 𝑗̂ − 𝑦 2 𝑧𝑘̂]. (𝑑𝑥𝑖̂ + 𝑑𝑦𝑗̂ +

𝑑𝑧𝑘̂]

611
(Refer Slide Time: 06:35)

.
Now if we take the dot product then this will be basically ∮𝐶(2𝑥 − 𝑦)𝑑𝑥. So, here 𝑧 =
0, 𝑑𝑧 = 0 because on the curve C, we have taken z as 0 right. So, we substitute 𝑧 = 0 here
.
and here and we substitute 𝑑𝑧 = 0. So, ultimately we will get ∮𝐶(2𝑥 − 𝑦)𝑑𝑥. So, that is
what we are obtaining here. And now we substitute 𝑥 = cos 𝑡 , 𝑦 = sin 𝑡. So, this will
2𝜋 𝑑𝑥 2𝜋 2𝜋 1
∫𝑡=0(2 cos 𝑡 − sin 𝑡) 𝑑𝑡 𝑑𝑡 = − ∫𝑡=0(2 cos 𝑡 − sin 𝑡)(sin 𝑡)𝑑𝑡 = − ∫𝑡=0 (sin 2𝑡 − 2 (1 −

cos 2𝑡)) 𝑑𝑡 = 𝜋.

So, we adjust 2 here dt and then we integrate term by term and ultimately we will obtain
𝜋 right. So, ultimately we will obtain 𝜋. Now we calculate; the right hand side.

612
(Refer Slide Time: 08:48)

𝑖̂ 𝑗̂ 𝑘̂
𝜕 𝜕 𝜕
⃗⃗ × 𝐹⃗ = |
So, first of all we have to calculate ∇ 𝜕𝑥 𝜕𝑦 𝜕𝑧
| = 𝑘̂. So, when we
2𝑥 − 𝑦 −𝑦𝑧 2 −𝑦 𝑧2

do the partial derivative, we will ultimately obtain a 𝑘̂ because rest of the term will be
either cancelled out or they will either be 0. So, they do not play any role and therefore,
the component of 𝑖̂, 𝑗̂ and 𝑘̂ will be 𝑖̂, 𝑗̂ will be 0 and 𝑘̂ will be one. So, that is why we are
only writing the unit vector 𝑘̂ and 𝑛̂ is the unit outward wrong normal on the surface S.

So, if our C is the curve basically that is enclosing that surface and if 𝑛̂ is the outward
drawn normal so, you have a sphere on the upper half of the z axis and then if you have a
unit on normal 𝑛̂, then you can choose 𝑘̂ axis itself as your unit normal right because 𝑘̂
axis is normal to that surface of the sphere and we can actually choose it instead of
calculating we can actually choose 𝑘̂ as the unit vector.

And therefore, we can write if 𝑆1 is the plane region bounded by the circle C, then,
.
⃗⃗ × 𝐹⃗ ). 𝑛̂𝑑𝑆 =
basically we have this as surface integral over 𝑆1. Let us say we have ∬𝑆(∇
.
⃗⃗ × 𝐹⃗ ). 𝑘̂𝑑𝑆 = ∬. 𝑘̂. 𝑘̂𝑑𝑆 = ∬. 𝑑𝑆 = 𝜋
∬𝑆 (∇
1 𝑆 1 𝑆 1

So, then we have basically 𝜋 . So, our left hand side was 𝜋 and our right hand side is also
𝜋. Therefore, if the left hand side and right hand side are both same then by Stokes
theorem, we can be able to say that the given integral or the given problem satisfies the

613
Stokes theorem which means that the surface integral and the line integral both are same
or both hold true all right. So, that is one way we can solve our problems motivated from
stokes theorem.

Similarly you can have several other problems from several other problems from Stokes
theorem where you either convert your surface integral into a line integral or you can
convert your line integral into a surface integral and just calculate either one of them. So,
if one expression is given to be complicated try to use the Stokes theorem to convert it into
a simpler one and see if you can calculate that one or not alright.

So, let me give you a quick recap of these syllabus. So, before I proceed any further.

(Refer Slide Time: 13:05)

So, we started so, we initially we started this integral and vector calculus course from
Riemann integral or from the concepts of Riemann integral. So, I know that in your plus 2
or even when you were preparing for other competitive exams, you must have learnt a
little bit about indefinite integrals and definite integral. Those are all very nice form of
integrals or sometimes they are also called as Newtonian integral. So, where you look for
the anti derivative and either by method of substitution or whatever methods are unknown,
you just use them to calculate the integral.

So, more or less since you were familiar with the integral of a function or integral of single
variable. We started with concepts of Riemann integral I gave you introduction of partition

614
concepts of Riemann integral; Riemann integrable function fundamental theorem of
integral calculus mean value theorems and things like that and based on which we will also
provide you some examples or assignments to practice.

Then we looked into reduction formula and derivation of different types of reduction
formula. So, sometimes you may be asked to calculate sin to the power 13 x dx. So, then
in that case you just have to use that reduction formula for sin to the power nx dx and from
that formula you can be able to calculate whatever power is given there.

So, the deduction formula is a very nice tool where you just see if there is some kind of
liquidity formula there, so, that you can come to a generalized formula. And then we learnt
about improper and proper and improper integral their convergence, we also learnt about
some tests to do the convergence all right. And afterwards we moved to beta and gamma
function, we looked into their properties we learnt about differentiation under the integral
sign; Leibnitz theorem and things like that.

(Refer Slide Time: 15:12)

Afterwards we looked into double integral, we also learnt about change of order of
integration, Jacobian transformation, then we moved to the applied path a part of integral
calculus where we learnt about rectification and the surface integral. We learned about
volume integrals and several other types of volume integrals the center of gravity and
moment of inertia. I did not cover that part because from integral calculus perspective I
think it, it might go a little bit in the different direction. So, if anyone of you are interested

615
you can look into those references; since I was focused to also cover the vector calculus
part I gave my attention to those two topics.

So, these are the two topics which I initially proposed, but then over the time I realized
that this might lead to problems from a different parts of mechanics or something. So, it is
better that we should not do it at the moment, but if any one of you are interested you
should just look into those books and you will be able to find these two topics there. And
then we moved to our vector calculus part where we looked into the limit continuity
differentiability of a vector function, arc length, partial derivative of vector function,
directional derivative, gradient divergence and curl operators. So, we learned about
different types of notions in vector calculus. We also practiced some examples.

(Refer Slide Time: 16:48)

Then we learnt about tangent normal and binormal and then we derived one of the
important formulas in vector calculus which is Serret Frenet formula we also looked into
the applied part of vector calculus where it is used and how to derive some equations of
motions and things like that. We then looked into 3 important theorems like Gauss
divergence theorem, Stokes theorem and the Green’s theorem.

So, up until previous class I did try to cover Green’s theorem, Gauss theorem and today
we are we learnt about Stokes theorem. We also tried to solve one example motivated from
the Stokes theorem. And in our 12th week or in our last week, I thought maybe I could
practice more examples from integral calculus or vector calculus, but due to the time

616
constraint unfortunately we cannot be able to practice more examples. But I would suggest
you to look into those books, there you will find plenty of examples for you to practice
and work out problems from our assignment sheet that will also help you get a sense of
this topic. And this is what we tried to cover in a 12 week session actually. Although I
personally feel that it might need a little bit more time to cover these two topics in detail;
however, we tried our best to do that.

So, just go through these lectures and try to get sense a feeling of the topic whatever we
learned and talked about and if you have any questions or doubts you, I believe we have a
weekly sessions where we can clear out the doubts. Also if you have any questions, you
can just go in to my home page and write me an email I will be more than happy to answer
your questions and try to help you with your queries. So, we do not have to just worry
about the fact that you can only contact me why this NPTEL format or something.

So, if you have any in general any questions or doubts on integral calculus or vector
calculus or which you could not be able to follow in this lecture series, then just write me
an email go to the IIT Kharagpur website, try to find out my profile there. You find my
email address and write me an email and I will be more than happy to answer your queries.

So, from my side I tried to cover as many topics as I could and I hope you will be able to
learn these things more clearly and I look forward to your questions your queries and yeah
good luck. Since we still have at least 5 minutes left so, let me just state one more example
from Stokes theorem and then we will move I will close this lecture. So, let me go back to
my alright.

617
(Refer Slide Time: 20:00)

So, since we have some time we can focus on one more example. So, example 2 from
Stokes theorem. So, verify stokes theorem 𝐹⃗ = (𝑥 2 + 𝑦 2 )𝑖̂ − 2𝑥𝑦𝑗̂ taken around the
rectangle bounded by 𝑥 = ±𝑎 and 𝑦 = 0 and 𝑦 = 𝑏 . So, this is just one more example for
us to practice. It is really not that much complicated it is a lengthy problem, but it is really
not complicated.

So, if we want to draw this rectangle, we can just do that 𝑥 = −𝑎, 𝑥 = 𝑎 and this is our x
axis, this is our y axis and then we have 𝑦 = 0 and we have 𝑦 = 𝑏 sorry. So, this is our
𝑥 = 𝑎 and our rectangle this is our rectangle. So, we our curve C will move like this. So,
this is our curve C alright and the equation of this line is 𝑥 = 𝑎, this line is 𝑦 = 𝑏, this line
is 𝑦 = 0 and this is 𝑥 = −𝑎. So, these are the 4 equations and let us call it as A B C and
D. So, first of all in order to verify we have to calculate the surface integral and the surface
integral can be calculated by so, we have.

𝑖̂ 𝑗̂ 𝑘̂
𝜕 𝜕 𝜕
So, we first need to calculate 𝑐𝑢𝑟𝑙𝐹⃗ . So,𝑐𝑢𝑟𝑙𝐹⃗ = | 𝜕𝑥 𝜕𝑦 𝜕𝑧
| = −4𝑦𝑘̂. And
𝑥2 + 𝑦2 −2𝑥𝑦 0
unit normal since this whole surface lies in the xy plane and out of drawn unit normal
would be any cost all right. Because z axis is obviously perpendicular to the xy plane and
you can choose an escape. So, we choose or we have or we get 𝑛̂ = 𝑘̂ since the given
surface is in xy plane.

618
(Refer Slide Time: 23:00)

.
⃗⃗ × 𝐹⃗ ). 𝑛̂𝑑𝑆 =
So, therefore, if we want to calculate the surface integral ∬𝑆(∇
. .
∬𝑆 −4𝑦𝑘̂. 𝑘̂𝑑𝑆 = −4 ∬𝑆 𝑦𝑑𝑆

𝑎 𝑏 𝑏2
So,−4 ∫𝑥=−𝑎 ∫𝑦=0 𝑦𝑑𝑥𝑑𝑦 = −4 2𝑎 = −4𝑎𝑏 2 .
2

So, that is the value of the surface integral. Now for the line integral; for the line integrals
C here, we have to be a little bit careful because our curve is composed of 4 curves; our
curve C is composed of 4 mini curves. So, this is our 𝐶1 , this is our 𝐶2 , this is our 𝐶3 and
this is our 𝐶4 so; that means, we can write this integral over C as the union of
. .
𝐶1 , 𝐶2 , 𝐶3 & 𝐶4 and union of 𝐶1 , 𝐶2 , 𝐶3 & 𝐶4 can be written as ∫𝐶 𝐹⃗ . 𝑑𝑟⃗ = ∫𝐶 𝐹⃗ . 𝑑𝑟⃗ +
1
. . .
∫𝐶 𝐹⃗ . 𝑑𝑟⃗ + ∫𝐶 𝐹⃗ . 𝑑𝑟⃗ + ∫𝐶 𝐹⃗ . 𝑑𝑟⃗. So, we have to evaluate all of these 4 integrals right.
2 3 4

So, when you first evaluate over 𝐶1 , you have to take care of the fact that the equation on
𝐶1 we have y goes to 0; then dy is 0 and you just; and you just integrate from −𝑎 to +𝑎.
So, that will be the value on 𝐶1 . Similarly when you are integrating on 𝐶2 , you have to take
𝑥 = 𝑎 and y is varying from 0 to b. So, and so, 𝑑𝑥 = 0,x is a.

So, we calculate the second integral. Similarly we calculate third and fourth, then you sum
all of them and you will be able to see that the required answer is −4𝑎𝑏 2 . And therefore,
your Stokes theorem will be verified. So, this was one more example where you could be
able to verify the Stokes theorem.

619
So, I hope these examples have cleared out your doubts based on Stokes theorem and I
believe we have exhausted the time. So, I will stop here for now and I hope you enjoyed
this course and I will look forward to your questions your suggestions, your remarks. If
you have anything and based on this course and I will try to answer them, I will try to reply
to you as soon as possible and of course, I am always reachable by email. So, do not
hesitate to contact me.

So, thank you for attention and have a good time.

620
 THIS BOOK IS
NOT FOR SALE
NOR COMMERCIAL USE

(044) 2257 5905/08 nptel.ac.in swayam.gov.in