Quantum Mechanics 2

Solution to problem set 10

Pay attention that in some of the questions, I use a different form of the Born approx-
imation. Carry out the k integration, and find out that this is exactly what we had in
class

1. We evaluate the scattering in the Born approximation, which is valid when the
kinetic energies are much larger than the binding energy. The Fourier transform of
the potential is
4π 2
V (~q) = 2
λ + q2
and the formula for the total cross section σ of electrons with initial wave vector ~k
is
2π Z d3 k 0 2 ~ 0 ~  h̄2 k 2 h̄2 k 02
!
V k −k δ −
h̄v (2π)3 2m 2m
This cross section is suitable for classical particles, without regard to spin. The
specification to the spin states S = 0, 1 is made below. Write d3 k 0 = k 02 dkdΩ The
differential cross section is found by taking the functional derivative of the cross
section with respect to the solid angle
dσ 1
= 2 2
dΩ a0 [λ + 2k (1 − cos θ)]2
2

where we have used the fact that v = 2h̄k/m and k is defined by E = h̄2 k 2 /2m The
 2
magnitudes of the vectors ~k and ~k 0 are the same, so ~k − ~k 0 = 2k 2 (1 − cos θ). All
of the dimensional factors are combined into the Bohr radius a0 . Now we consider
how this formula is altered by the spin of the electrons. Spin is conserved in the
scattering, so the pair of electrons has the same spin state before and after the
collision

(a) For S = 0 the two electrons are in a spin singlet which has odd parity. Hence,
the orbital state must have even parity. The initial and final orbital wave
functions are given below, along with the form of the matrix element. The
relative coordinate is ~r
1 h ~ ~
i
ψi = √ eik·~r + e−ik·~r
2
1 h i~k0 ·~r ~0
i
ψf = √ e + e−ik ·~r
2

1
    
hf | V |ii = V ~k 0 − ~k + V ~k 0 + ~k
" #2
dσ 1 1 1
= 2 2 2
+ 2 2
dΩ a0 λ + 2k (1 − cos θ) λ + 2k (1 + cos θ)
The matrix element has two factors
(b) For S = 1 the spins are in a triplet state which has even parity. The orbital
part of the wave function has odd parity. There is a minus sign between the
two terms in initial state instead of a plus sign, and ditto for the final wave
function. Now the differential cross section is
" #2
dσ 1 1 1
= 2 2 2
− 2 2
dΩ a0 λ + 2k (1 − cos θ) λ + 2k (1 + cos θ)
There is a relative minus sign between the two term in the matrix element
2. In the first Born approximation the scattering is proportional to the square of the
matrix element between initial and final states. If the initial wave vector is ~k and
the final one is ~k 0 set ~q = ~k 0 − ~k and evaluate (remember that the transfor of a
gaussian is a gaussian, and the second term can be written as ∇q~ of a transform of
a gaussian. Here these are implicit)
Z
V (~q) = d3 r ei~q·~r V (r)
!3/2 " #
π −q 2 /4µ i
= e A+ B~q · ~σ
µ 2µ
!3/2 " #
π −q 2 /4µ i
= e A+ B (qz σz + q+ σ− + q− σ+ )
µ 2µ
where

q± = qx ± iqy
1
σ± = (σx ± iσy ) (1)
2
and where we have written the transverse components of momentum in terms of
spin raising and lowering operators. The initial spin is pointing along the direction
of the initial wave vector ~k, which we define as the z−direction Let us quantize the
final spins along the same axis. Now consider how the three factors scatter the spins
(a) The term A is spin independent. It puts the final spin in the same state as the
initial spin.
(b) Bqz σz is a diagonal operator, so the final spin is also along the initial direction,
and this term has a value of Bqz .

2
 (c) Bq+ σ− flips the spin from +z to −z and contributes a matrix element of Bq+
to the final state with the spin reversed
(d) Bq− σ+ gives a matrix element of zero since the initial spin cannot be raised.
When we take the magnitude squared of each transition and sum over final
states, we get the factors for spins of
2 2
i i B 2q2
A+ Bqz + B (qx + iqy ) = A2 + .
2µ 2µ 4µ2
The differential cross section is written as
2π 4 h̄2 k 2 h̄2 k 02
!
dσ Z
02
2
0 −(~k−~k0 ) /2µ
= k dk e δ −
dΩ h̄vµ3 (2π)3 2m 2m
2 
" #
B  2
× A2 + 2 ~k − ~k 0
4µ
2
πm −k2 (1−cos θ)/µ 2 B 2 2
" #
= e A + 2 k (1 − cos θ) .
4h̄4 µ3 4µ
We have used the fact that energy is conserved, so k 0 = k.
3. (a) The particle scatters from the potential energy V (r) which is related to the
charge distribution
∇2 V (r) = −4πeρ(r)
4πe
V (q) = ρ(q)
q2
where V (q), ρ(q) are the Fourier transforms of V (r), ρ(r) respectively. The
differential cross section in the Born approximation is
dσ 1 Z
0 02

~k − ~k 0 2

= dk k δ (E k − E k 0) V
dΩ (2π)2 h̄vk
 
2 θ
4e2 m2 ρ 2k sin 2
=
h̄4
 4
2k sin θ 2

(b) In forward scattering we take θ → 0. we get

Z
ρ(q) = d3 rρ(r)ei~q·~r
(~q · ~r)2
Z " #
3
= d rρ(r) 1 + i~q · ~r − + ...
2
q2A
= −
6
3
 (the first term vanishes because the distribution is neutral, the second because
it is spherically symmetric, and the third termR gives the angular average since
it is ∝ dΩ (~q · ~r)2 = dΩq 2 r2 cos θ2 = q 2 r2 2π dθ cos2 θ = 2π q 2 r2 ) and so
R R
3

ρ2 (q) A
lim 2
=−
q→0 q 6
and the cross section for forward scattering is
dσ e2 m2 A2
lim =
θ→0 dΩ 9h̄4
(c) The charges in a hydrogen atom are the nucleus, which is taken as a delta
function at the origin, and the electron, which is given by the square of the
ground state wave function
h i
ρ (~r) = −e δ 3 (~r) − ψ100
2
(r)
" #
16
ρ(q) = −e 1 − 2
(4 + q 2 a20 )
A = 3ea20
dσ
lim = a20
θ→0 dΩ

4. To solve this problem properly, we have to be careful about the k’s and q’s. Remem-
ber, that k is the magnitude of the momentum ~k, of both incoming, and scattered
 
waves, and q is the “momentum transfer”. Its magnitude is q = 2k sin 2θ , and
is directly dependent on the angle of observation. That is why, even for the Born
approximation, we wrote the scattering amplitude as f (θ). Don’t be confused by
the fact that for the Born amplitude we integrated over all angles, so naively you
would think that f must be independent of θ. The spatial integration carried out
for the scattering amplitude was d3 r0 , and its θ0 6= θ.
So, the scattering amplitude for a central field in the Born approximation is given
by (we had this in class)
BORN
−2m Z ∞
f (θ) = 2 r sin(qr)V (r)dr
h̄ q 0
This approximation is valid for a sufficiently weak potential when all the phase shifts
δl << 1. The expression of f (θ) in terms of δl , is
1X
f BORN (θ) = (2l + 1) eiδl sinδl Pl (cos θ)
k l
1X
≈ (2l + 1) δl Pl (cos θ)
k l

4
 using the orthogonality relations of Legendre polynomials we find
Z 1
2δl
= f BORN (θ)Pl (cos θ) d (cos θ)
k −1
−2m Z ∞ Z 1
= r sin(qr)V (r)Pl (cos θ) d (cos θ) dr
h̄2 q 0 −1
−2m Z ∞ 2 Z 1
sin(qr)
= 2 r V (r) Pl (cos θ) d (cos θ) dr
h̄ 0 −1 qr
−2m Z ∞ 2 Z 1
sin(2k sin 2θ r)
= r V (r) Pl (cos θ) d (cos θ) dr
h̄2 0 −1 2k sin 2θ r
−2m Z ∞ 2
= 2jl (kr)r2 V (r) dr
h̄2 0
where the last equality is not trivial, and which gives us the final expression, if all
the phase shifts are small
−2mk Z ∞ 2
δl = jl (kr)r2 V (r) dr (2)
h̄2 0

5. (a) Let us use the known form for the Born scattering amplitude of a spherical
symmetric potential, and substitute in it the explicit form of the potential
−2m Z ∞
f (θ) = r sin(qr)V (r)dr
h̄2 q 0
−2mV0 R2
= sin(qr)
qh̄2
q
with q = 2k sin (θ/2), and k = 2mE/h̄2 . Therefore

dσ 4m2 V02 R4
= |f |2 = 2 sin2 (qr).
dΩ qh̄
The Bom approximation is applicable in cases where the scattering potential
can be considered as a perturbation, or, in our case

h̄2 k
V0 r sin(kr) << .
2m
This can be obtained in two limiting cases
2mV0 R
I h̄2 k
<< 1 high energies

2mV0 R
II h̄2
<< 1 low energies

5
 We note that the first condition (kR arbitrary) is less restrictive than the
second one (kr << 1 ). Equation (I) indicates that Bom approximation is
applicable for scattering at sufficiently high energies. Equation (II) shows on
the other hand that if kR << 1 , then the Bom approximation is valid for all
velocities v = h̄k/m. (in both cases one must, of course, consider scattering
from a relatively weak and short-ranged potential). We can also verify, that in
the low-energy limit (qR → 0) the Born scattering cross section is completely
isotropic.
q
(b) We use the result of the previous problem (2) and also jl (x) = π/2xJl+1/2 (x)
so
−pimR2 V0 2
δl = Jl+1/2 (kR).
h̄2
Using the asymptotic behavior
q
2
I Jl+1/2 (x)x→∞ → πx
sin(x − πl/2)
q
2 xl+1/2
II Jl+1/2 (x)x→∞ → π (2l+1)!!

in the expression for the phase shifts, we find for x → ∞

πmR2 V0 2
|δl | < << 1
h̄2 πkR
which coincides with the low energy validity of the Born approximation. Sim-
ilarly, for x → 0 we find that

−πmR2 V0 2 (kR)2l+1
δl ≈ (3)
h̄2 π [(2l + 1)!!]2

Hence, the condition δl << δ0 , coincides with the second condition of part (a).
(c) As we just found if kR << 1 then δl << δ0 . This result is in agreement with
the general analysis of partial wave expansion, which states that for finite-range
potential the main contribution to the scattering amplitude comes from values
of l < kR, where R is the range of the potential. Using (3) for s-wave scattering
(l = 0) we obtain
−2mR3 V0 k
δ0 ≈ sin δ0 ≈
h̄2
Thus, the leading term of the differential cross section is

dσ 4m2 R6 V02
=
dΩ h̄2

6
As expected from the previous discussion, the two cases coincide in the limit
qR → 0 . In this case, both the s-wave approximation and the Born approxi-
mation lead to a differential cross section that depends neither on the angle of
scattering nor on the energy of the incident particle.