1

Greens Function for the Harmonic Oscillator

Consider the simple harmonic oscillator, whose motion is governed by the equation
d2 x
m 2 + kx = 0.
dx
We are familiar with its general solution
x = A sin !t + B cos !t,
q

k
where ! = m
. It would be nice if we could somehow use this result to find
the general solution for the forced harmonic oscillator

d2 x
+ kx = F (t).
dx2

The main task is to find a particular solution xp that satisfies this dierential
equation. Then the general solution is simply x = xp + A sin !t + B cos !t. But
finding a particular solution for an arbitrary function F (t) is not an easy task.
In order to make it less daunting, we define the function in a slightly unusual
way, using the picking out property of the Dirac-delta function.
Z 1
F (t) =
F (t0 ) (t t0 )dt0
1

Written in this form, the function F (t) can be thought of as a sum of delta
functions, each with magnitude F (t0 ). The advantage of writing it this way is
2
that we can exploit the fact that if x1 and x2 are solutions for m ddxx2 +kx = F1 (t)
2
2
and m ddxx2 + kx = F2 (t), then x1 + x2 is a solution for m ddxx2 + kx = F1 (t) + F2 (t).
Since F (t) can be expressed as a sum of delta functions, our task is reduced to
finding the particular solution for
m

d2 x
+ kx = (t
dx2

t0 ).

After finding a particular solution for this t0 element, we can just scale it by
F (t0 ) and add the solutions for every t0 to get the particular solution xp that
we want.
Lets plug in the general solution for the simple harmonic oscillator x =
A sin !t + B cos !t to the above dierential equation and determine A and B.
We may choose A = 1/m! and B = 01 and define
8
>
if t < t0
<0
x(t) = 1
if t = t0
>
: 1
t0 )) if t > t0
m! sin (!(t
1 Actually, B doesnt have to be 0. But we can still choose B to be zero because all we
need to do is to find just one particular solution that satisfies the equation

This satisfies the above equation. We have found the special solution(Greens
Function) G(t; t0 ) that we were looking for.
At this point, it is worth thinking about the physical meaning of Greens
Function. When there is no driving force, the RHS of the dierential equation
is always 0. Instead, imagine that at time t0 , somebody gives an infinitely
strong, but infinitesimally short push to our oscillator. The Greens function
describes the resulting motion. If there is another push at some dierent time t00 ,
then the resulting motion is the sum of the two Greens functions corresponding
to each push. So when we add up all the special solutions to get xp , we are
essentially adding up the eects of infinitely many infinitesimal push at each
time t0 .

Greens Function for Electrostatics

From Gausss law,

rE=

.
0

Electric field can be written as the negative gradient of electric potential.

rV

E=

Plugging this into the first equation gives

r2 V =

.
0

In two dimensional case, we can rewrite this as

@2V
@2V
+
=
2
@x
@y 2

(x, y)
.
0

In general, this dierential equation is not easy to solve because (x, y) can
be any arbitrary charge distribution. This is analogous to the forced harmonic
oscillator case, in which the F (t) term on the right hand side could be any
arbitrary function.
What plays an analogous role for the simple harmonic oscillator in this situation is the free space case(no electric charge), where (x, y) is 0 everywhere.
When there is no charge, the equation becomes
@2V
@2V
+
= 0.
@x2
@y 2
The general solution is
V (x, y) = A ln

x2 + y 2 + B ,

for (x, y) 6= (0, 0). For the same reason as before, we are interested in the case
where the RHS of our dierential equation is (r r0 ). If we could somehow
find a special solution that satisfies
@2V
@2V
+
= (r
@x2
@y 2

r0 ) ,

then we can just scale it by (r0 )/0 and add it up for each r0 to get the
particular solution for the original dierential equation.
In order to find the special solution(Greens function), I will move the origin
to r0 and define r r r0 . In polar coordinates, the dierential equation
becomes2

1 @
@V
r
= (r )
r @r
@r
and the general solution for free space becomes
V ( r ) = A ln

+B

r 6= 0.
If we choose A = 1/2 and B = 0, the above dierential equation is satisfied. We can check this by substituting V = (1/2) ln r into our dierential
equation. It is easy to see that the LHS becomes zero everywhere except at
the origin. At the origin, the LHS blows up, which is consistent with how delta
functions behave. Now we just have to check if the integral over the whole space
is 1.

Z Z 1
Z 1
1 d
dV
d
dV
r
r
r
d r d =
dr
dr
dr
0
1 r dr
1 dr

1
dV
= r
dr
1
for

For r > 0, ddV

r = A. V has to be an even function of r because negative
r just means opposite direction. The derivative of an even function is an odd
function. This means that for r < 0, ddV
r = A. Therefore, the integral above
evaluates to 2A. If A = 1/2, the integral is 1 as required. So the Greens
function for this dierential equation is
G( r ) =

1
ln
2

Greens Function for Schr

odinger Equation

We start with the two dimensional time-independent Schrodinger equation:

h2

@
@2
+ 2 + V (x, y) (x, y) = E (x, y)
2m @x2
@y
2 Im

assuming that there is no dependence, which is reasonable for electrostatics.

With the substitution k =

2mE/h, this can be written as

r2 + k 2

2m
V .
h2

If V = 0 everywhere(free particle), then the dierential equation becomes

r2 + k 2

= 0.

This is the homogeneous Helmholtz equation and its general solution is given
by
( r ) = AJ0 (k r ) + BY0 (k r )
where J0 and Y0 are the zeroth order Bessel functions of the first and second
kind. I have again shifted the origin to r0 in order to avoid writing r r0
repeatedly.
We are interested in the solution to the equation
2m
(r ).
h2
Before plugging in the general solution as we have done in earlier sections, let
us note that if k = 0, the equation becomes
r2 + k 2

@2
@2
2m
+
= 2 (r ).
2
2
@x
@y

h
It is exactly the equation that we encountered in section 2, only scaled on the
right hand side by 2m/
h2 . The Greens function for this equation would be
(r ) =

2m 1
ln
h2 2

Now, although k is not zero in our equation, the special solution(Greens

function) has to approach the above solution for very small k. If we choose
appropriate A and B, the solution will converge to a natural log for small k.
Bessel functions of the first and second kind can be written as
1
X
( 1)m (k r )2m
J0 (k r ) =
4m (m!)2
m=0
Y0 (k r ) =

2
kr
(ln (
) + )J0 (k r )

where

1
2 X ( 1)m (k r )2m
m=0
4m (m!)2

1 1
1
+ + + )
n!1
2 3
n
is the Euler constant. We want a linear combination of J0 (k r ) and Y0 (k r )
m
that goes to
ln r for small k. J0 goes to 0 for small k, so it is irrelevant.
h2
We may as well choose A = 0. We need B = m/2h2 to get the right limit.
Therefore,
m
G( r ) = 2 Y0 (k r ) .
2h
= lim ( ln(n) + 1 +

Things to add/fix
Greens function for electrostatics is 1/ r . Mine is ln

Physical meaning of Greens function(Sections 2 and 3)

Determining coefficients in section 1 needs more detail
It is not obvious that B = m/2h2 in section 3
Integral in section 2 is somewhat dubious