Scribd
Upload
73 views

Eigenfunctions of Laplacian

The document discusses eigenfunctions of the Laplacian operator. It provides examples of eigenfunctions on the 2-torus and 2-sphere. Any square-integrable function on these spaces can be expanded using eigenfunction series. Applications include solving PDEs and stationary quantum states. Open problems relate to characterizing eigenvalues/eigenfunctions and determining properties of domains based on spectral data.

Uploaded by

artherisk
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
73 views

Eigenfunctions of Laplacian

The document discusses eigenfunctions of the Laplacian operator. It provides examples of eigenfunctions on the 2-torus and 2-sphere. Any square-integrable function on these spaces can be expanded using eigenfunction series. Applications include solving PDEs and stationary quantum states. Open problems relate to characterizing eigenvalues/eigenfunctions and determining properties of domains based on spectral data.

Uploaded by

artherisk
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

Eigenfunctions of laplacian

D. Jakobson (McGill)
email: [email protected]
September 13, 2007
The Laplacian of a function f is given by
f = div(gradf).
An eigenfunction with eigenvalue 0 sat-
ises
f +f = 0.
Example 1: R
2
.
f =

2
f
x
2
+

2
f
y
2
.
Periodic eigenfunctions on the 2-torus T
2
:
f(x 2, y 2) = f(x, y). They are
sin(m x+n y), cos(m x+n y), = m
2
+n
2
.
Fact: any square-integrable function F(x, y)
on T
2
(s.t.

T
2
|F(x, y)|
2
dxdy < ), can be
expanded into Fourier series,
F =
+

m,n=
a
m,n
sin(mx+ny)+b
m,n
cos(mx+ny).
Example 2: sphere S
2
= {(x, y, z) : x
2
+ y
2
+
z
2
= 1}. Spherical coordinates: (, ) [0, ]
[0, 2], where x = sincos , y = sinsin, z =
cos .
f =
1
sin
2

2
f

2
+
cos
sin

+

2
f

2
Eigenfunctions are called spherical harmonics:
Y
m
l
(, ) = P
m
l
(cos )(a cos(m) +b sin(m)).
Here = l(l + 1); P
m
l
, |m| l is associated
Legendre function,
P
m
l
(x) =
(1)
m
2
l
l!
(1 x
2
)
m
2
d
l+m
dx
l+m

(x
2
1)
l

.
Any square-integrable function F on S
2
can be
expanded in a series of spherical harmonics.
The same is true on any compact (e.g. closed
and bounded) curved surface S: any square-
integrable function F on S can be expanded in
a series of eigenfunctions of .
Similar results hold in higher dimensions, and
for domains with boundary.
Applications: Solving partial dierential equa-
tions like heat equation u(x, t)/t = c
x
u(x, t)
and wave equation
2
u(x, t)/t
2
= c
x
u(x, t).
Stationary solutions of Schrodinger equation
or pure quantum states.
Inverse problems: suppose you know some eigen-
values and eigenfunctions; describe the domain
S (related problems appear in radar/remote
sensing, x-ray/MRI, oil/gas/metal exploration
etc).
Mathematical Problems:
Determine the smallest > 0 for a given sur-
face S (its bass note), and the corresponding
eigenfunction.
Can you hear the shape of a drum: can two
dierent domains S have the same spectrum,
e.g. the collection of all {0
1

2
. . .}?
Count the eigenvalues: N(T) = #{
j
< T}.
How fast does N(T) grow as T ?
Example: 2-torus T
2
:
m,n
= m
2
+ n
2
. Let
T = R
2
. Then
N(R
2
) = #{(m, n) : m
2
+n
2
< R
2
} =
#{(m, n) :

m
2
+n
2
< R}.
How many lattice points are inside the circle of
radius R? Leading term is given by the area:
N(R
2
) = R
2
+E(R), (1)
where E(R) is the remainder.
Question: How big is E(R)? Conjecture (Hardy):
for any > 0,
E(R) < C() R
1/2+
, as R .
Best known estimate (Huxley, 2003):
E(R) < C R
131/208
(logR)
2.26
.
Note: 131/208 = 0.629807....
An analogue of (1) holds for very general do-
mains; it is called Weyls law (Weyl, 1911).
Much less is known about E(R).
Questions about eigenfunctions: Let
f +f = 0, -large (high energy).
Where is f concentrated, i.e. describe {(x, y) :
|f(x, y)| is large}?
Ex: on some domains with boundary, whis-
pering gallery eigenfunctions concentrate near
the boundary.
Nodal sets: study {(x, y) : f(x, y) = 0}. This
will generally be a curve, or a union of curves.
First pictures: Chladni plates (E. Chladni, 18th
century; see google video links on my home
www-page).
Ex: On T
2
, function f(x, y) = sin(mx) sin(ny)
vanishes on a rectangular grid:
{(x, y) : x = j/m, or y = k/n}.
In general, much less is known about nodal
sets.

You might also like