STONY BROOK UNIVERSITY

DEPARTMENT OF PHYSICS AND ASTRONOMY

Graduate Placement Exam Part 2, Aug. 25, 2010
General Instructions: This exam is for incoming graduate students who wish to
demonstrate mastery in one or more areas of the graduate core curriculum, in order
to skip one or more of the rst-year courses. Do two of the three problems in either
or both areas.
Each solution should typically take less than 45 minutes.
Use one exam book for each problem, and label it carefully with the problem topic
and number and your name. Make sure to do every part of the problems you choose.
You may use a one page help sheet, a calculator, and with the proctors approval, a
foreign language dictionary. No other materials may be used.
Quantum Mechanics 1
Calculation the transmission coecient of the following 1D scatterer:
() = 1 () + 2 ( )
as a function of the particles energy and nd its maximum value.
Solution:
Writing the scattering wavefunction with the appropriate scattering conditions at
, and matching the derivatives at = 0, we get
=

1
1,2
,
where

=
1,2
1 2 exp +(1 1 )(1 2 ) exp 2
2

To nd the maximum transmission, one can view the denominator of this fraction as
the squared magnitude of the dierence of two vectors with lengths

1 2 and(1 1 )(1 2 ) = (1 1 2 )2 + (1 + 2 )2 1 2 .
By the proper choice of the particles energy, i.e. parameter , the angle = 2+
const between these two vectors may be made zero, minimizing the denominator and
maximizing the transparency:
1
=
2
( (1 1 2 ) + (1 + 2 )2 1 2 )2
1

Quantum Mechanics 2
Calculate the dierential and full cross sections of a spherical scatterer:
{
0 , for <
() =
0, otherwise
in the Born approximation.
Solution:
For the spherically symmetric scatterer, one can use

20
20

(, ) = 2
sin = 2 3 (sin cos ).
0

Thus,

( )2

0
= 4 2 3 (sin cos )2 .

In contrast to the previous problem in this case the scattering intensity has the axial
symmetry about the axis of the incident wave propagation, so that the total cross
section may be calculated as

= 2
sin , = 2 sin .
2
0
As recommended in Landau and Lifshitz, the easiest way to calculate this integral is
to notice that since

= cos , and = 2 2 sin cos = 2 sin ,

2
2
2
we can replace sin with / 2 , and write

8 ( 0 )2 2 (sin cos )2
8 ( 0 )2 2 (sin cos )2
= 2
.
= 2

2
6

2
5
0
0
Working out the integral, one has nally
2 ( 0 2 )2 [
8()2 + 1 cos 4 4 sin 4 ]
= 2
1

2
32()4
Quantum Mechanics 3
A 1D quantum harmonic oscillator of frequency is initially in its ground state. At
time = 0, a constant force is abruptly applied to the oscillator. Express the forcerelated term in the Hamiltonian through the creation/annihilation operators , ,
and using these operators nd
2

a. the state () of the oscillator at > 0 in the basis {} of the new energy
eigenstates, and
b. the time dependence of average coordinate () and momentum () of the
oscillator.
Solution:
(a) Using the standard expression of the oscillator coordinate in terms of the creation/annihilation operators , , one can write the -related term in the Hamiltonian as:

)
(
+ .
(1)
= =
2
Then the total Hamiltonian of the oscillator at > 0 is:

)
(
+ = + const,
(2)
=
2
where
=,

.
(2 3 )1/2

Since the force was switched on abruptly, the state (0) of the oscillator right after
switching on of the force coincides with the initial ground state: (0) = 0, and
satises the following condition in terms of the new creation/annihilation operators
, :
(0) = (0) ,
i.e. it represents a coherent state. It can be expanded

(0) =

in the basis {} of the new energy eigenstates of the oscillator:

1
= ( ) 0 ,
!
where the new ground state is dened as usual by
0 = 0 .
The expansion coecients are found directly:
1
( )
= (0) = 0 (0) = 0(0) ,
!
!
3

and after proper normalization

( )
2
= /2 .
!
Taking into account the standard time evolution of the energy eigenstates, we obtain
the wavefunction of the oscillator in the basis of the new energy eigenstates:
() =

[ ]
2

/2 .
!

(3)

(b) Using the wavefunction () and the standard properties of the creation/annihilation
operators one nds:
()() = , () () = ,
and from this, the average coordinate () and momentum ():

(1 cos ) ,
(4)
(2 2 cos ) =
() =
2
2

() =
( ) = sin .
(5)
2

Note that these are exactly the classical equations, in agreement with Ehrenfests
theorem.
Statistical Mechanics 1
Derive the prediction from the Landau-Ginzburg mean eld formalism for the critical exponents as 0+ and 0 for the (isothermal, zero-eld)
susceptibility (e.g. of a ferromagnet) in a second-order phase transition.
Solution:
The second-order phase transition occurs for = , = 0. Here the LandauGinzberg expansion for the Helmholtz free energy is ( =temp, =magnetism)
(, ) = 0 + 2 2 + 4 4
where 4 > 0. The coecient 2 changes sign at the critical temperature, , for the
second-order phase transition. So we write it as 2 = 21 ( ) with 21 > 0 and
higher order terms in ( ). Now the external eld satises
=

= 22 + 44 3 = 2(2 + 24 2 )

and the isothermal compressibility is given by =

1
=

2
=
= 22 + 124 2

2
4

so

The zero-eld susceptibility is given by ( = 0). If > , then setting = 0

= 0,
i.e. there is no magnetism. If < , then for near , 2 = 2 /(24 ), i.e.
= 21 ( )/(24 ). This implies that the critical exponent for magnetization
is = 1/2. For , consider 0+ . Then 1
,=0 = 221 ( ) so
(, = 0) = 1/(221 ( )). The critical exponent for the susceptibiblity is
dened as
{
( ) , ( ) 0+
(, = 0) =

( ) , ( ) 0
So from the above calculation, we read o that = 1 is the prediction of the LandauGinzberg formalism for > . For < , we substitute 2 = 2 /(24 ) in the

expression for 1
and thus
, so (, = 0) 1/(421 ( )) as

= 1.
Statistical Mechanics 2

The Hamiltonian for the 1D Ising model may be written as = , , where

the classical Ising spin variable = 1 on each site , and denotes nearestneighbor pairs of sites. Derive the internal energy per site and the entropy per site
for this model on a one-dimensional lattice in the thermodynamic limit.
Solution:
For the zero-eld Ising model, the partition function is

=
exp

with = and = and this sum over nearest neighbors. The energy
can be calculated exactly by a transfer matrix method or by summing the small
series. The resulting Gibbs free energy (, ) = (, 0) (for the zero eld case) is
= with the value per site = (2 cosh ) where = .
The internal energy = / with the value per site
= / = tanh
for , so

= tanh + (2 cosh )

This can also be written as

=

= [ tanh + (2 cosh )].

N.B.: The transfer matrix is < exp >
Statistical Mechanics 3
The Debye temperature of silicon is 645 K and its melting temperature is
= 1.69 103 K. Derive the functional dependence on temperature of the molar
5

specic heat of silicon in the Debye model, for 10 K (for this part of the problem,
all that is asked for is the dependence, not the prefactor). Be sure to explain the
various assumptions of the model and the steps in the calculation. Give a rough
sketch of the molar specic heat as a function of for 0 800 K, labelling the
horizontal and vertical axes carefully. What is the approximate value of this specic
heat at 800 K and why?
Solution:
This is described in detail in e.g. Reif, Fundamentals of Thermal and Statistical
Physics. The model approximates lattice vibrations in the harmonic oscillator approximation. The internal energy

3 ()/(exp 1)
0

where = 1/ . Using =

2 ()/(exp 1)

in which is dened by = . The specic heat is given by

2
2
2 ()2()/(exp 1)2
=
=
3

Converting to dimensionless form with the integral becomes

1
3
4 exp /( 1)2
0
The dening = = / allows us to see that for << the upper limit
of the integral can be replaced with . Then for silicon with = 645 for silicon
and at = 10, so << and the result becomes
3 3 .
For >> quantum eects are small, so we use classical equipartition theorem for
the oscillator
1
1
= 2 + 2
2
2
(with 3 translation degrees of freedom and 3 potential energy degrees of freedom).
Each contribution gets 12 so = 3 and = 3 . The molar specic
heat is then = 3 with the ideal gas constant.