Derivation of the BET and Langmuir Isotherms
October 5, 2011
1 Langmuir Isotherm
Assumptions used for the Langmuir isotherm
1. Gaseous molecules behave ideally
2. Only one monolayer forms
3. All sites on the surface are equivalent
4. No adsorbate - adsorbate interactions
5. An adsorbed molecule is immobile
Then, for the following reaction
1k
A(g) + S(surface)
AS(surface)
k1
we can say
rate of adsorption = k1 PN(1-)
rate of desorption = k1 N
where P is the pressure above the surface, N is the total number of sites,
and
number of sites occupied na
= =
number of sites available nm
na is the number of molecules adsorbed, and nm is the number of molecules
in one monolayer. N(1-) is the number of empty sites and N is the number
of occupied sites.
1
We can then say that the rate of change of is
d
= rate of adsorption rate of desorption
dt
= k1 PN(1 ) k1 N
d
At equilibrium dt =0 so
k1 PN
=
k1 PN + k1 N
KP
=
KP + 1
k1
where K = k1 is the binding equilibrium constant with units of P1 .
1.2 Langmuir Isotherm
1
K=10
0.8
K=1
na/nm
0.6
K=0.1
0.4
0.2
K=0.01
0
-1 4 9 14
P (atm)
Figure 1: Langmuir isotherm plot of na /nm vs P (atm) for several dierent
values of the equilibrium binding constant, K.
Figure 1 shows a graph of na /nm as a function of pressure in atmospheres
for several dierent values of K, the binding equilibrium constant. Figure 2
shows a cartoon representation of how the alumina surface is being filled. For
the Langmuir isotherm the alumina surface can be treated as being made of
2
a bunch of identical slots, each of which can hold one nitrogen molecule. At
low pressures the nitrogen molecules begin to fill the slots. As the pressure
increases, more nitrogen gets adsorbed as can be seen in the near linear
low pressure part of the na /nm vs P curve. As the pressure continues to
increase, the slots fill up and eventually the surface is covered. At this point
increasing the pressure does not increase the amount of nitrogen adsorbed.
This occurs graphically on the flat part of the na /nm vs P curve, where
increasing the pressure does not change the amount of nitrogen adsorbed.
From this figure we can also see that as the binding equilibrium constant
increases (as binding becomes more favorable), the pressure where saturation
occurs decreases.
Langmuir isotherm
Alumina Surface:
All sites are considered equivalent
Each site can hold one N2
molecule. As P increases the N2 N2 N2
number of molecules
adsorbed increases
Once the surface is full,
increasing P does not increase N2 N2 N2 N2 N2 N2
the number of molecules
adsorbed
Figure 2: Cartoon of nitrogen adsorbing onto an alumina surface assuming
a Langmuir isotherm.
3
2 BET Isotherm
2.1 Assumptions
While the Langmuir isotherm provides a simple picture for looking at sur-
face adsorption, it is not an accurate representation of how nitrogen actually
adsorbs to the surface. This is because more than one nitrogen molecule can
adsorb to each site on the alumina. A better approximation for this process
is to use the BET Isotherm (named for its inventors, Brunauer, Emmett,
and Teller) which allows for multiple nitrogen molecules to adsorb to each
alumina site.
The assumptions used to derive the BET isotherm are
1. Gaseous molecules behave ideally
2. Multiple nitrogen molecules can be adsorbed to each site
3. Each adsorbed molecule provides a site for the adsorption of the molecule
in the layer above it
4. All sites on the surface are equivalent
5. No adsorbate - adsorbate interactions
6. An adsorbed molecule is immobile
7. Nitrogen in the second and higher layers are assumed to be liquid like
2.2 Derivation
Figure 3 shows a cartoon of what an alumina surface would look like under
the BET isotherm assumptions. We define 0 , 1 , 2 , n ... where
# of sites with n adsorbed molecules
n = (1)
# of sites total
(Note the definition is the number of sites with n adsorbed molecules, not
with at least n adsorbed molecules, see Figure 3.) We can also define
kn - rate of condensation to form the nth layer
kn - rate of evaporation to form the nth layer
P - pressure above the surface
N - total number of sites
4
&
T0=3/14
T1=4/14
T2=4/14
T3=2/14 N2
T4=1/14
N2 N2 N2
N2 N2 N2 N2 N2 N2 N2
N2 N2 N2 N2 N2 N2 N2 N2 N2 N2 N2
Figure 3: Cartoon of nitrogen adsorbing onto an alumina surface assuming
a BET isotherm.
At equilibrium
dn
=0 (2)
dt
for all n. For the bare surface we can write
d0
= rate of evaporation f rom 1st layer rate of condensation onto bare surf ace
dt
0 = N k1 1 N k1 P 0
k1
1 = P 0 (3)
k1
where N1 is the number of sites with 1 layer of nitrogen on it and N0 is the
number of sites with no nitrogen adsorbed. (Note this is the same answer
as for the Langmuir isotherm where 1 =(1-0 ).)
For surface sites with only one molecule adsorbed we can write
5
d1
= rate of condensation onto bare surf ace + rate of evaporation f rom 2nd layer
dt
rate of condensation onto the 1st layer rate of evaporation f rom the 1st layer
0 = k1 P 0 N + k2 N 2 k2 P N 1 k1 N 1 (4)
combining equations 3 and 4 gives
k2 2 = k2 P 1
k2 k2 k1
2 = P 1 = P 2 0 (5)
k2 k2 k1
For the second layer
d2
= rate of condensation onto 1st layer + rate of evaporation f rom 3rd layer
dt
rate of condensation onto the 2nd layer rate of evaporation f rom the 2nd layer
0 = k2 P 1 N + k3 N 3 k3 P N 2 k2 N 2 (6)
Plugging equation 5 into 6 gives
k3 3 = k3 P 2
k3 k3 k2 k1
3 = P 2 = P 3 0 (7)
k3 k3 k2 k1
Generalizing to the ith layer gives
ki i = ki i1 P
ki
i = P i1
ki
ki ki1 ki2 k1 i
= ... P 0 (8)
ki k(i1) k(i2) k1
Since for i>1 the BET isotherm assumes the nitrogen is liquid like, we can
also say that
k2 k3 k4 ki = k1
k2 k3 k4 ki = k1
where ki is the rate of adsorption onto a liquid nitrogen surface and ki is
the rate of desorption o a liquid nitrogen [Link] if we define x=P kki
i
we can write
k1 1 k1
i = xi1 P 0 = xi P 0 ; i = 0, 1 (9)
k1 x k1
6
P k1
and if we define c= xk 1
, we can say
i = xi c0 ; i = 0, 1 (10)
If we then define the total surface coverage, , such that
na total # of molecules adsorbed
= =
n total # of sites
we can write
na = ii N
i=0
because each site has i molecules on it and
n= i N
i=0
so that
ii N ii c0 ixi
i=0 i=1 i=1
= = =
i N 0 + i 0 + c0 xi
i=0 i=1 i=1
If we use
x
xi =
1x
i=1
then
d i d i d x x
ixi = x ixi1 = x x =x x =x =
dx dx dx 1 x (1 x)2
i=1 i=1 i=1 i=1
plugging this in for we obtain
x cx
c0 (1x) 2 (1x)2 cx 1 cx
= x = 1x+cx = =
0 + c0 1x (1x)
1 x 1 x + cx (1 x)(1 + (c 1)x)
7
2.3 Making Physical Sense of the Constants
The constant x was defined as x=P(ki /ki ) where ki and ki are the ad-
sorption rate and desorption rate, respectively, o a liquid nitrogen surface.
If we consider a liquid nitrogen surface in equilibrium with the gas state the
condition for equilibrium is
ki N P = ki N
or ki P=ki . The equilibrium pressure for this system, P, is just the vapor
pressure for pure nitrogen, Po , at the temperature the experiment is being
done. Therefore x is just P/Po , the pressure over the equilibrium vapor
pressure of liquid nitrogen at the experimental temperature.
The constant c was defined as (Pk1 )/(xk1 ) where k1 is the rate of ad-
sorption onto the bare alumina surface and k1 is the rate of desorption o
the surface. Plugging in what we know for x, c becomes (Po k1 )/(k1 ), a unit
less number. Using this and Equation 3, c=1 /0 , the ratio of the number
of adsorbed molecules on the first layer to the number of empty sites. This
ratio should be
n1 des H o vap H o
c= =e RT
n0
where des H o is the enthalpy of desorption and vap H o is the enthalpy of
vaporization. This equation comes about because n0 , the number of bare
des H o
spots, is proportional to e RT (if desorption is energetically favorable
o
(des H <0) n0 is larger) and the number of molecules in the first mono-
vap H o
layer, is proportional to e RT (if vaporization is energetically favorable
o
(vap H <0) n1 is larger). Notice that c is large when the enthalpy of
desorption from the monolayer is large compared with the enthalpy of va-
porization.
2.4 Limitations of the BET Isotherm
The BET isotherm is accurate when P/Po is between 0.05 and 0.3, pres-
sures at which only a few complete layers have formed. At low pressures the
BET isotherm does not do a good job at describing adsorption because in
deriving the isotherm it was assumed that all the sites on the alumina surface
are equivalent. This is not really the case and when only a few molecules
have adsorbed, eects from inhomogeneities in the surface become more
8
important. As more molecules adsorb these inhomogeneities average out.
At high pressures, the cracks and roughness of the surface cause the BET
isotherm to fail. This is especially true for very porous materials where the
indents have a width of only a few monolayers. This is because the cracks
can only hold a few monolayers and as the cracks fill up, the eective surface
area of the solid where the nitrogen adsorbs can change.
3 References
Atkins, P.W. Physical Chemistry 2nd edition. San Francisco: W. H. Free-
man and Company, 1978.
Atkins, P.W. Physical Chemistry 6th edition. San Francisco: W. H. Free-
man and Company, 1998.
Derivation of the Langmuir and BET Isotherms
([Link]
Shoemaker, D. P., Garland, C. W., & Nibler, J. W. Experiments in Physical Chemistry 5th edition.
New York: McGraw-Hill, 1989.
Surface Adsorption
([Link] 91/images/stories/[Link])
