1.2.

SCATTERING

February 4, 2015
Lecture VII

1.2 Scattering
Scattering is perhaps the most important experimental technique for exploring the structure of
matter.
• From Rutherford’s measurement that informed the “planetary” in place of the plum pudding
model, of negative charges circulating about a compact positively charged nucleus. Of course
the energy of the alpha particle probe needed to be sufficient to penetrate the electron coud
so that it would see the full nuclear charge.
• Scattering of electrons from nuclei revealed the structure of first of the nuclei (protons and
neutrons) and then fractionally charged quarks, quark spins, quark momentum distriburtions
within the proton.
• Neutrino, nuclei scattering experiments demonstratd the weak coupling of quarks, via W and
Z0 bosons.
• Electron-positron scattering lead to a new field of heavyy quark spectroscopy, the study of
the bound states of charm and beauty quarks and anti-quarks.
• Total e+ e− cross section for hadron production was direct observation of color degree of
freedom, that quards come in 3 colors
• Forward backward asymmetry in e+ e− scattering revealed the weak neutral current coupling
of quarks as well as leptons.
• The ultimate scattering experiments at LHC have unearthed a Higgs, and perhaps ?

• xray scattering is used to determine the structure of crystals.

• And x-ray and low energy electron scattering and neutron scattering is the principle tool for
the study of condensed matter.
• Parity violation in weak interactions, CP violation in Kaon and B-meson decay, etc.

We have a nice neat picture today, the Standar Model. The experimantal evidence is the product
of mostly scattering experiments.

1.3 Lippman-Schwinger Equation

We start with the simplest description of scattering, that is, a free particle (plane wave) interacting
with a fixed and localized potential. Localized means that the potential falls off rapidly far from
the origin. And plane wave meaning a state with definite energy and momentum. While scattering
is inherently time dependent we begin by thinking about a time independent version, perhaps a
continuous stream of incident and then scattered particles. As this is quantum mechnaics, the
Schrodinger equation will play a big role. We will represent the scattering particle as a solution to

1
1.3. LIPPMAN-SCHWINGER EQUATION

SE. But what do we measure? Not a wave function. We measure a flux of scattered particles at
some angle, energy, spin? So we need a way to translate from wave function (solution to SE) to
flux.
Of course if we think about classical scattering, then we have an impact parameter. We solve
the equations of motion, and relate impact parameter to scattering angle. But that does not work
for QM.
The strategy is first to find a solution to the time independent Schrodinger equation for free
particle states, where that solution is a plane wave from the scattering center. Picture a plane wave
from −∞ at the left to +∞ to the right and a spherical wave centered around the origin of the
potential. We want the solution for a particular energy, namely the energy of the incident particle.
We start assuming elastic scattering. The energy of the scattered particle is fixed.
We start with the Schrodinger equation

(H0 + V )| ψi = E| ψi
→ (E − H0 )| ψi = V | ψi

where H0 is the kinetic energy of the free particle and V the scattering potential. Again, we imagine
an incoming plane wave, interaction with some potential, and then time independent solution. We
“solve” as follows
1
→ | ψi = V | ψi + | φi (1.1)
E − H0
where H0 | φi = E| φi. That way when V → 0, | ψi → | φi. Also, far from the scattering center,
| ψi → | φi. So far we have an operator equation. We can translate it to some concrete algebra by
inserting some basis states. But first we note that H0 has a continuous spectrum that will include
E, so we replace E with the complex variable E ± i. Then
1
→ | ψi = V | ψi + | φi (1.2)
E − H0 ± i
In the coordinate basis
1
hx | ψi = hx | V | ψi + hx | φi
E − H0 ± i
Z Z
1
= d3 x00 d3 x0 hx | | x0 ihx0 |V | x00 ihx00 | ψi + hx | φi
E − H0 ± i

For a local potential hx0 | V | x00 i = δ 3 (x0 − x00 )V (x0 )

1
hx | ψi = hx | V | ψi + hx | φi
E − H0 ± i
Z
1
= d3 x0 hx | | x0 ihx0 |V | x0 ihx0 | ψi + hx | φi
E − H0 ± i

Let’s try to evaluate

~2 1
G(x, x0 ) = hx | | x0 i
2m E − H0 ± i

2
1.3. LIPPMAN-SCHWINGER EQUATION

Z
~2 1
= d3 p00 d3 p0 hx | p0 ihp0 | | p00 ihp00 | x0 i
2m E − H0 ± i
Z
1
= ~2
d3 p0 hx | p0 ihp0 | p00 i 2 hp00 | x0 i
p − p0 2 ± i
Z i(x−x0 )·p0 /~ D E
1 3 0 e 00 3 0 00 e−ix·p/~
= d p where we use p | p = δ (p − p ), and hp | xi =
(2π)3 ~ p2 − p0 2 ± i (2π~)3/2
Z 0 0
1 ei|x−x |p cos θ/~
= d3 p 0 2
(2π) ~3
p − p0 2 ± i

Next integrate around θ

Z ∞ Z 0 0
0 1 02 0 ei|x−x |p cos θ/~
G(x, x ) = 2π p dp dφd(cos θ) 2
(2π)3 ~ 0 p − p0 2 ± i
Z ∞ 0 0 0 0
1 2π ~ 0 2 0 ei|x−x |p /~ − e−i|x−x |p /~
= − p dp
(2π)3 ~ i|x − x0 | 0 p0 p2 − p0 2 ± i
Z ∞ 0 0
i|x−x |k 0 0
1 1 0 0e − e−i|x−x |k
= − 2 k dk
8π i|x − x0 | −∞ k 2 − k 0 2 ± i
Z ∞ i|x−x0 |k0 0 0
1 1 0 0 e − e−i|x−x |k
= − 2 k dk
8π i|x − x0 | −∞ (k − k 0 ± i)(k + k 0 ± i)

Next do the contour integral. We first consider the contribution from the denominator E − H0 + i.
This can be written in terms of k where E = ~2 k 2 /2m and k 0 as (k + i0 − k 0 )(k + i0 + k 0 ) With
that in mind let’s rewrite our last expression just for the +i piece. Then
Z ∞ 0 0 0 0
0 1 1 ei|x−x |k − e−i|x−x |k
0 0
G(x, x ) = − 2 k dk
8π i|x − x0 | −∞ (k + i0 − k 0 )(k + i0 + k 0 )

There are poles at k 0 = k + i0 and k 0 = −k − i0 . (The poles for denominator E − H0 − i are
shown in Figure 1.2.) For the positive exponent we close in the upper half plane and include the
pole at k 0 = k + i. For the negative exponent close in the lower half plane and include the pole at
k 0 = −k − i. Each term contributes 0
ei|x−x |k
2πik
2k
Again for +ı we get
Z ∞ 0 0 0 0
0 1 1 0 0 ei|x−x |k − e−i|x−x |k
G(x, x ) = − 2 k dk
8π i|x − x0 | −∞ (k + i0 − k 0 )(k + i0 + k 0 )
1 2πi 0
= − 2 ei|x−x |k
8π i|x − x0 |
1 1 0
= − 0
ei|x−x |k
4π |x − x |

3
 1.3. LIPPMAN-SCHWINGER EQUATION
70 Version of October 25, 2006CHAPTER 7. THE CALCULUS OF RESIDUES

• − k + i! Γ
→
−R
•k − i! R

Figure 7.5: The closed

Figure contour
1.1: Poles C for theEintegral
for denominator in Eq. (7.41).
− H0 − i.

For E − H0 = i, we would get

Equivalently, instead of deforming the contour to avoid the singularity, one
can displace the singularity,G(x, x0 x→ 0 x0 ± 1 1
) = − i!. Then ei|x−x |k
0
(1.3)
! ∞ 4π!|x∞− x0 |
g(x) g(x)
We are interested in thedx
solution far from= thePscattering ± iπg(x
dx origin, since that is0 ),
where we will(7.39)
detect
x −the
−∞In fact,
the scattered particle. ∓ i! particle
x0scattered −∞will be x a−plane
x0 wave with definite momentum,
ifjust
g islikea the incident particle, but in a different direction. So let’s inspect G
regular function on the real axis. [Proof: Homework.]
|x − x0 | is large, and |x| is much greater than the range of the potential.
in the limit where

7.7 Example 4
G(x, x ) = 0
−
1
e i|x−x0 |p/~
∼ −
ei(x
2
+x0 2 −2x·x0 )1/2 p/~

4π|x − x0 | 4πr
Consider the integral ir(1−x·x0 /r 2 )p/~
e
∼ −
! ∞ 4πr iqx
e e
e 0 )p/~
irp/~ −i(x̂·x
I∼= − 2
dq, x > 0, (7.40)
−∞ q 4πr − k 2 + i!
which is0 important in quantum mechanics. We can replace this integral by the
Now if k = x̂|p|/~, that is k 0 is set to be in the outgoing x direction to the detector.
contour integral "
eiqx eikr −ix ·k 0 0
→2 xx 2 = − dq, e x > 0,
0 (7.41)
C q − k + i! 4πr
where
Finally the closed
we can writecontour C is shown in Fig. 7.5. The integral over the “infinite”
semicircle Γ is zero 2m Z
according to Jordan’s lemma.
hx | ψi =
√ By redefining !, but not
d3 x0 G(x, x0 )V (x0 )hx0 | ψi + hx | ki2
changing its sign, we ~write 2 the integral as (k = + k )
# Z $ %
" 2m eikr
hx | ψi = −1 2 0 −ix ·x̂p/~
d3 x1e
0
iqx
(x0 )hx0 | ψi + hx |eki
V iqx %
I= dq ~ 4πr e = 2πi %
q − (k − i!)
ikr Z + (k − i!)
q q − (k − i!) %
C 2m e 3 0 −ix ·k 0
0 0
0 q=−(k−i!)
hx | ψi = − d xe V (x )hx | ψi + hx | ki
~2 4πr πi −ikx
= − e , (7.42)
k
4
in the end taking ! → 0.

7.8 Example 5
We will consider two ways of evaluating
! ∞
dx
I= . (7.43)
1.3. LIPPMAN-SCHWINGER EQUATION

Figure 1.2: Red is the region of the scattering potential. The incident plane wave is headed in the
k direction. The scattered particle is observed at P. The direction of propagation of the outgoing
plane wave is x.

Z
2m eikr 0 0 eix·k
= − 2 d3 x0 e−ix ·k V (x0 )hx0 | ψi +
~ 4πr (2π)3/2
Z 0 0
!
1 eikr 2m(2π)3 e−ix ·k
= − d3 x0 V (x0 )hx0 | ψi + eix·k
(2π)3/2 r 4π~2 (2π)3/2

Define the scattering amplitude

Z
4π 2 m
f (k 0 , k) = − d3 x0 hk 0 | x0 iV (x0 ) x0 | ψ + (1.4)
~2
Then
 
1 eikr
hx | ψi ∼ f (k 0 , k) + eik·x
(2π)3/2 r

5
1.4. DIFFERENTIAL CROSS SECTION

1.4 Differential cross section

We found that the solution to the Schrodinger equation has the form
 ikr 
1 e 0 ik·x
hx | ψi ∼ f (k , k) + e
(2π)3/2 r
and that Z
0 4π 2 m
f (k , k) = d3 x0 hk 0 | x0 iV (x0 )hx0 | ψi (1.5)
~2
Not really much good since we need the solution to do the calculation, but we do learn something
1
about the form. We have been sloppy about normalization. Multiply by V − 2 where V is volume of
space. Then if the plane wave represents the incoming flux, we have incident flux v/V = ~k/mV .
The flux scattered radially outward is
v |f (k, k 0 )|2
.
V r2
Let dṄ be the number of particles scattered outward per unit time into the solid angle dΩ.
v |f (k, k 0 )|2 2
dṄ = r dΩ (1.6)
V r2
The differential cross section
dσ dṄ
= = |f (k, k 0 )|2 (1.7)
dΩ Inc Flux

1.4.1 Probability current

The scattered particle probability flux is
~
j = Im(ψ ∗ ∇ψ)
m
  ikr 
~ 1 f ∗ −ikr e eikr eikr
∼ Im e ik f− 2 f+ ∇f
m 8π 3 r r r r

At large r, all terms fall off faster than 1/r2 except the first. Note that ∇f will involve angular
derivatives that all have 1/r and then derivatives with respect to θ and φ. So very far away,
  ikr 
~ 1 f ∗ −ikr e
j ∼ Im e ik f
m 8π 3 r r
~ k|f 2 |
∼
m 8π 3 r2

The flux into the detector with area r2 dΩ will be Fdet = jr2 dΩ. The total incoming flux is
jinc = k (2π)~3 m . Then the rate of scattering into solid angle dΩ is
dσ j 2 r2
R = Finc dσ = jscat r2 dΩ → = = |f |2
dΩ jinc
That’s all well and good. But, probability is conserved. So j for the entire wave function ψ
integrated over the entire sphere must be zero.