Mathematical Techniques III (PHY 317) Answers For Exercise Class 8

Mathematical Techniques III
(PHY 317)
Answers for Exercise Class 8
Answer to Exercise 1.
Let us consider the function
f(t) = e
t
2
/2
.
By denition, its Fourier transform is given by
f() =
1
2
_

e
t
2
/2
e
it
dt .
In order to perform this integral we complete the square in the expo-
nential:
2
t
2
it =
2
_
t
2
+
2i
t
_
=
2
_
t +
i
_
2

2
2
.
Therefore,
f() =
1
2
e
2
/2
_

e
(t+i/)
2
/2
dt .
To perform the integral we interpret it as a limit
lim
e
(t+i/)
2
/2
dt
and think of the integral for nite as the parametrised contour integral
_
e
z
2
/2
dz ,
where
is the contour parametrised by z(t) = t +i/ for t [, ].

In order to compute this integral, let us consider the closed contour
,
obtained by adding three line segments to
, as shown below:
+ i/
+ i/
6
-
?
The integrand exp(z
2
/2) is entire, so that by the integral theorem,
_
e
z
2
/2
dz = 0 .
1
This implies that
_
e
z
2
/2
dz =
_
_
e
z
2
/2
dz .
We now claim that the integrals along
vanish in the limit .

Indeed, notice that
is parametrised by z(t) = it/ for t

[0, 1]. Therefore,
_
e
z
2
/2
dz =
_
1
0
i
e
(it/)
2
/2
dt
=
i
2
/2
_
1
0
e
it+
2
t
2
/2
dt .
But now notice that
_
1
0
e
it+
2
t
2
/2
dt
_
1
0
e
it+
2
t
2
/2
dt =
_
1
0
e
2
t
2
/2
dt e
2
/2
,
whence

e
z
2
/2
dz
||
2
/2
e
2
/2
,
which goes to zero as . In summary,
lim
e
z
2
/2
dz = lim
e
z
2
/2
dz .
Along
, z(t) = t, so that
lim
e
z
2
/2
dz =
_

e
t
2
/2
dt .
This integral is elementary to perform: we let u =
_
2
t, so that
_

e
t
2
/2
dt =
_
2
e
u
2
du =
_
2
.
Therefore putting it all together,
f() =
1
2
_
2
2
/2
=
1
2
e
2
/2
.
2
By denition, the Dirac delta function is the Fourier transform of 1.
Therefore since lim
0
f(t) = 1 for all t, the limit 0 of its Fourier
transform

f is the delta function. This gives a way to construct func-
tions which approximate the delta function: simply by considering the
Fourier transforms of functions which approach the constant function
1.
This exercise contains the mathematical underpinning of the Heisen-
berg uncertainty principle. The Fourier transform relates the position
x and momentum p representations of the quantum mechanical Hilbert
space. If we prepare a gaussian state in the x representation of the
form
(x) e
x
2
/2
,
in the momentum representation, the state has the form
(p) e
p
2
/2
.
In other words, the widths of distribution are inversely proportional,
so that roughly speaking x p is a constant. Therefore the more
peaked our initial state around a given x, the wider the momentum
distribution, and viceversa. In the limit that the state is concentrated
at a single point (the delta function), all momenta are equally likely:
the constant function 1.
Parsevals identity tells us that the Fourier transform preserves (up to
a constant factor) the square norm of the function:
1
2
_

|f(t)|
2
dt =
_

f()|
2
d .
The proof is simple:
_

f()|
2
d =
_

1
2
_

f(t) e
it
dt
2
d
=
1
4
2
_

__

f(t) e
it
dt
_ __

f(s)
e
is
ds
_
d .
Now we interchange the order of integrations and do the integral rst
to obtain
_

f()|
2
d =
1
4
2
_

e
i(ts)
d f(t)f(s)
dt ds .
3
The integral gives 2(ts), which allows us to perform the s integral
to obtain the answer:
_

f()|
2
d =
1
2
_

(t s)f(t)f(s)
dt ds
=
1
2
_

|f(t)|
2
dt .
This exercise illustrates the power of the Fourier transform in solving
dierential equations. The idea behind any sort of useful transform is to
allow us to solve a complicated problem by mapping it to an equivalent
problem whose solution is simpler, solving the simpler problem, and
then mapping back. In particular the Fourier transform will allow us
to turn dierential equations into algebraic equations which are readily
solved.
Consider then the diusion equation
d
2
(t)
dt
2

2
(t) = q (t) .
We will rst take the Fourier transform of this equation and obtain an
algebraic equation for the Fourier transform

() of (t). The Fourier
transform of the rst term in the equation is obtained integrating by
parts:
1
2
_

d
2
(t)
dt
2
e
it
dt =
2
() ,
where we have dropped the boundary terms since we assume that
and its derivatives go to zero as |t| . The Fourier transform of
the second term is simply
2
(). Finally the transform of the right-

hand side of the equation is simply q/2. As a result the transformed
equation becomes
(
2
+
2
)
() =
q
2
,
which can be readily solved for

():
() =
q
2 (
2
+
2
)
.
In order to solve for (t) we simply invert the Fourier transform:
(t) =
q
2
_

2
+
2
e
it
d ,
4
which can be evaluated using the residue theorem as was done in the
lecture. If t > 0 we must close the contour in the upper half-plane and
we pick up the residue for the simple pole = +i, whereas if t < 0
we close the contour in the lower half-plane and we pick up the residue
for the pole at = i. If t = 0 we can close the contour either way.
Therefore,
_

2
+
2
e
it
d =
_
2 i Res(+i) for t 0; and
2 i Res(i) for t 0.
Computing the residues, we see that
Res(+i) =
e
t
2i
and Res(i) =
e
t
2i
.
Putting it all together, we have
(t) =
q
2

_
2 i
1
2i
e
t
for t 0, and
2 i
1
2i
e
t
for t 0;
=
q
2
e
|t|
.
5

Mathematical Techniques III (PHY 317) Answers For Exercise Class 8

Uploaded by

Copyright:

Available Formats

Mathematical Techniques III (PHY 317) Answers For Exercise Class 8

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Mathematical Techniques III (PHY 317) Answers For Exercise Class 8

Uploaded by

Copyright:

Available Formats

Mathematical Techniques III

is the contour parametrised by z(t) = t +i/ for t [, ].

vanish in the limit .

is parametrised by z(t) = it/ for t

(). Finally the transform of the right-

You might also like