Schwartz - QFT and the Standard Model To Chin Yu

Problem 2.1 We can write the transformation as perturbation series

x0 = x + vt + δx(2)
t0 = t + δt(1) + δt(2)

The perturbation terms can be obtained via order-by-order comparison

t02 − x02 = t2 − x2
(2tδt(1) − 2tvx) + (2tδt(2) + [δt(1) ]2 − 2xδx(2) − v 2 t2 ) = 0

which gives
δt(1) = vx
v2 t v2 x
2t(δt(2) − ) + 2x( − δx(2) ) = 0
2 ( 2
2
δt(2) = v2 t
2 (x, t are arbitrary)
δx(2) = v 2x

If we expand the full transformations, we have

v2
x0 = (x + vt)(1 + + O(v 4 ))
2
v2 x
= x + vt + + O(v 3 )
2
v2 t
t0 = t + vx + + O(v 3 )
2
which agree with the results obtained from perturbation.

Problem 2.2

E = 7 TeV
E
γ= ≈ 7000
m
β ≈ 0.9999999898
v − c = (1 − β)c ≈ 3 ms−1

For the relative velocity we have

2β
vrel = c=c
1 + β2

Send comments to [email protected] 1

Schwartz - QFT and the Standard Model To Chin Yu

Problem 2.3 The energy of the CMB photon is given by

Eγ ≈ kTCM B = 0.00023 eV
At threshold, the nal products should be at rest in the center of momentum frame. By mo-
mentum conservation we have
Pp + Pγ = Pp0 + Pπ0
m2p + 2Eγ (Ep + pp ) = (mp + mπ )2
q (mp + mπ )2 − m2p
Ep + Ep2 − m2p = =A
2Eγ
Ep2 − m2p = A2 − 2AEp + Ep2
A2 + m2p
Ep =
2A
≈ 3.1 × 1020 eV

To nd the energy of the outgoing proton, we rst nd its speed via
pp + Eγ = γβ(mp + mπ )
q
Eγ + Ep2 − m2p
γβ = =B
mp + mπ
p
γ = B2 + 1
≈ 2.6 × 1011

Thus
Ep0 = γmp ≈ 2.6 × 1020 eV

Problem 2.4 Yes, rotate about y -axis by π and then apply P .

Problem 2.5 Typical X-ray energy is on the order of 1 keV which is much larger than the
ionization energy of the electron on the order of 10 eV. Therefore for most cases we can neglect the
binding energy of the electron.
We can nd the frequency of the reected X-ray via momentum conservation
Pγ + Pe = Pγ0 + Pe0
(Pe0 )2 = (Pγ + Pe − Pγ0 )2
m2e = m2e + 2Pe · Pγ − 2Pe · Pγ0 − 2Pγ · Pγ0
0 = me Eγ − me Eγ0 − Eγ Eγ0 (1 − cos θ)
Eγ
Eγ0 = Eγ
1 + me (1 − cos θ)

which looks like the following when plotted

Send comments to [email protected] 2

Schwartz - QFT and the Standard Model To Chin Yu

If the electron mass goes to zero, the only possible solution becomes Eγ0 = Eγ and θ = 0 because
the photon cannot have zero energy. In other words the interaction between the photon and the
electron will be turned o in this limit.
In classical EM the frequency of the outgoing radiation produced by the electron is the same
as the driving frequency which is the frequency of the incoming photon. Therefore the frequency
distribution is a constant function.

Problem 2.6
Z ∞ Z ∞  δ(k 0 − ω ) δ(k 0 + ω ) 
k k
dk 0 δ((k 0 )2 − (k2 + m2 ))θ(k 0 ) = dk 0 + θ(k 0 )
−∞ −∞ 2ωk 2ωk
∞
δ(k 0 − ωk )
Z
= dk 0
0 2ω
1
=
2ωk

where ωk = k2 + m2 .
p

The Jacobian for Lorentz tranformation is J = | det Λ| = 1. Therefore the measure d4 k is

Lorentz invariant.
We can consider the following integral
Z ∞
I= d4 k δ(k 2 − m2 ) θ(k 0 )
−∞

Both the measure and the integrand are manifestly Lorentz invariant, thus I itself must be
Lorentz invariant. We can write I in another form
Z Z
I= d3 k dk 0 δ(k 2 − m2 ) θ(k 0 )

d3 k
Z
=
2ωk
which gives us the desired result.

Send comments to [email protected] 3

Schwartz - QFT and the Standard Model To Chin Yu

Problem 2.7

† † † † † †
∂z (e−za aeza ) = −a† (e−za aeza ) + (e−za aeza )a†
† †
= e−za (aa† − a† a)eza
=1
−za† za†
e ae =z+a
† † †
aeza = zeza + eza a

Thus
†
a|zi = aeza |0i
†
= z|zi + eza a|0i
= z|zi

hn|n̂|zi = hn|a† a|zi

√
nhn|zi = z nhn − 1|zi
z
cn = √ cn−1
n
zn
cn = √ c0
n!
zn
=√
n!

where cn = hn|zi, and c0 = h0|eza |0i = h0|(1 + za† + ...)|0i = h0|0i = 1.

†

To calculate the uncertainties, we rst evaluate the following quantities

X X |z|2n 2
hz|zi = |cn |2 = = e|z|
n!
X √ X |z|2(n−1) 2
hz|a|zi = cn c∗n−1 n=z = ze|z|
(n − 1)!
X 2
hz|n̂|zi = cn c∗n n = |z|2 e|z|
X 2
hz|aa† |zi = cn c∗n (n + 1) = (|z|2 + 1)e|z|
X p 2
hz|a2 |zi = cn c∗n−2 n(n − 1) = z 2 e|z|

Thus
z 2 + 2|z|2 + 1 + (z ∗ )2 − (z + z ∗ )2 1
∆q 2 = hq 2 i − hqi2 = =
2mω 2mω
2 2 2 −mω 2 2 ∗ 2 ∗ 2 mω
∆p = hp i − hpi = [z − 2|z| − 1 + (z ) − (z − z ) ] =
2 2

Send comments to [email protected] 4

Schwartz - QFT and the Standard Model To Chin Yu

Thus
1
∆p∆q =
2
Assume |wi = bn |ni is an eigenstate of a† , then we have
P

X √
a† |wi = bn n + 1|n + 1i
X X √
wbn |ni = bn−1 n|ni
√
n
bn = bn−1
√ w
n!
= b0 n
w
=0

In the last line we used b0 = 0. This is because 0 = h0|a† |wi = wb0 . Since all the coecients
vanish, a† does not have any eigenstate.

Send comments to [email protected] 5

Schwartz - QFT and the Standard Model To Chin Yu

Problem 3.1 Each additional derivative simply produces a minus sign after integration by parts,
therefore it is easy to see that the equation of motion is
∂L ∂L ∂L
− ∂µ + ∂µ ∂ν + etc. = 0
∂φ ∂(∂µ φ) ∂(∂µ ∂ν φ)

Problem 3.2 We can write the innitesimal Lorentz transform as

1 (αβ)
Λµν = δµν + Jµν ωαβ
2
where ωαβ is the rotation or boost parameter in the α − β plane. The factor of 1/2 is to avoid
double-counting (e.g. 1 − 2 and 2 − 1 are the same rotation).
For example, for an innitesimal rotation in the 1 − 2 plane we have
   
1 0 0 0 0 0 0 0
0 1 −ω12 0  = δνµ + ω12 0 0 −1 0
Λµν = 
 

0 ω12 1 0 0 1 0 0
0 0 0 1 0 0 0 0
 
0 0 0 0
(12)
0 0 1 0
Jµν =   = δµ1 δν2 − δν1 δµ2
0 −1 0 0
0 0 0 0

There are C24 = 6 such generators (3 rotations and 3 boosts), and in general we can write
(αβ)
Jµν = δµα δνβ − δνα δµβ
Equipped with this formula, we can now calculate the Noether current corresponding to Lorentz
symmetry.

 ∂L 
∂µ Lδxµ = ∂µ ∂λ φn δxλ
∂(∂µ φn )
 ∂L 
(αβ) ν
∂ µ LJ (αβ)
µν x ν
= ∂ µ
∂ λ
φ n Jλν x
∂(∂ µ φn )
∂L (αβ)
∂µ( ∂ λ φn Jλν xν − LJµν
(αβ) ν (αβ) µν
x ) = −LJµν δ
∂(∂ µ φn )
(αβ)
∂ µ (Tµλ Jλν xν ) = 0
∂ µ (Tµα xβ − Tµβ xα ) = 0
Kαβµ = Tµα xβ − Tµβ xα

The third line we used the fact that an antisymmetric tensor contracted with a symmetry tensor
gives zero.
Now we want to evaluate this current for a free massive scalar theory. We rst calculate the
stress-momentum tensor.
1 1
Tµν = −∂µ φ∂ν φ + gµν (∂λ φ)2 − gµν m2 φ2 = Tνµ
2 2

Send comments to [email protected] 1

Schwartz - QFT and the Standard Model To Chin Yu

and also
1 1
∂ µ Tµν = −∂ 2 φ∂ν φ − ∂ µ φ∂µ ∂ν φ + ∂ν (∂λ φ)2 − ∂ν (m2 φ2 )
2 2
= −∂ 2 φ∂ν φ − m2 φ∂ν φ − ∂ µ φ∂µ ∂ν φ + ∂ λ φ∂ν ∂λ φ
=0
and
Kαβµ = xβ (∂µ φ∂α φ + gµα (∂λ φ)2 − gµα m2 φ2 ) − xα (∂µ φ∂β φ + gµβ (∂λ φ)2 − gµβ m2 φ2 )
To check that the current satises the continuity equation, we take the derivative.
∂µ Kαβµ = Tµα δβµ − Tµβ δαµ
= Tβα − Tαβ
=0
Next we want to calculate the charge corresponding to the boost.
K0i0 = T00 xi − T0i x0
= Exi − P i t
Z
Qi = Exi d3 x − Pi t

= E x̄i − Pi t
Qi Pi
x̄i = + t
E E
Thus physically the conservation law means that the center of energy of the system x̄i moves in
a straight line.
The nal part of the problem follows trivially from the Heisenberg equation.
dQi ∂Qi
= i[H, Qi ] +
dt ∂t
∂Qi
i = [H, Qi ]
∂t
In general, since Qi is a linear combination of position and momentum, [Qi , H] 6= 0. Therefore
Qi is not an invariant of the equation of motion.

Problem 3.3 By adding a total derivative the energy-momentum tensor changes by

∂∂λ X λ
δTµν = ∂ν φ − gµν ∂λ X λ
∂(∂ µ φ)
Z
∂
δQ = (∂λ X λ )φ̇ − ∂λ X λ d3 x
∂ φ̇
Z
= −2 ∂λ X λ d3 x
Z
= −2∂0 X 0 d3 x

=0

Send comments to [email protected] 2

Schwartz - QFT and the Standard Model To Chin Yu

where the last equality is due to the fact that X µ must vanish at the boundary.
For electromagnetism, the energy-momentum tensor is
∂L 1
Tµν = µ
∂ν Aλ + gµν F 2
∂(∂ Aλ ) 4
1
= −Fµλ ∂ν Aλ + gµν F 2
4
The rst term is generally not symmetry.
Assume T̃µν is our new symmetrized tensor, then we must have
∂∂σ X σ ∂∂σ X σ
T̃{µν} = −Fµλ ∂ν Aλ + Fνλ ∂µ Aλ + ∂ν Aλ − ∂µ Aλ
∂(∂µ Aλ ) ∂(∂ν Aλ )
Thus we must have
∂∂σ X σ
= Fµλ
∂(∂µ Aλ )
For which we can choose
1 στ
Xσ = F Aτ
2
which gives
∂∂σ X σ ∂ 1
= ( F στ ∂σ Aτ )
∂(∂ µ Aλ ) ∂(∂µ Aλ ) 2
= Fµλ

and thus leads to a symmetric energy-momentum tensor.

Problem 3.4 The Feynman diagram is given by

where we have 5 points x, u, v, y, z, t. The point v can be on either the upper or lower branch,
so we have 2 diagrams in total.
Using the Feynman rules given in the chapter, we have
Z
2
h2 (x) = − λ d4 u d4 v d4 y d4 z d4 t Π(x, u)Π(u, t)Π(u, v)Π(v, y)Π(v, z)J(y)J(z)J(t)
Z
− λ2 d4 u d4 v d4 y d4 z d4 t Π(x, u)Π(u, y)Π(u, v)Π(v, t)Π(v, z)J(y)J(z)J(t)
Z
2
= − 2λ d4 u d4 v d4 y d4 z d4 t Π(x, u)Π(u, t)Π(u, v)Π(v, y)Π(v, z)J(y)J(z)J(t)

Send comments to [email protected] 3

Schwartz - QFT and the Standard Model To Chin Yu

From perturbation theory we have

u h2 = 2λh1 (u)h0 (u)
Z Z
= 2λ2 d4 v d4 y d4 z Π(u, v)Π(v, y)Π(v, z)J(y)J(z) d4 t Π(u, t)J(t)
Z
h2 = −2λ2 d4 u d4 v d4 y d4 z d4 t Π(x, u)Π(u, t)Π(u, v)Π(v, y)Π(v, z)J(y)J(z)J(t)

So the two results match.

Problem 3.5 The equation of motion is

λ 3
−φ + m2 φ − φ =0
3!
λ
m2 c = c3
3! r
6
c = 0, ± m
λ
r
6 3 m4
V (0) = 0, V (± m) = −
λ 2 λ
The non-zero solutions are the ones corresponding to the ground state.
The only state that respect the Z2 symmetry is the state φ = −φ = 0. The ground state picks
out either the positive or the negative direction so that symmetry is broken.
If we expand around the ground state, we get
1 1 1 λ
L = − (c + π)π + m2 c2 + m2 cπ + m2 π 2 − (c4 + 4c3 π + 6c2 π 2 + 4cπ 3 + π 4 )
2 2 2 4!
We see that now there are some odd powers of π in the Lagrangian that breaks the Z2 symmetry.
The equation of motion is
λ 2 λ λ
−π + m2 π − c π − cπ 2 − π 3 = 0
2! 2! 3!
The Z2 transformation of φ corresponds to π → −π − 2c. The Lagrangian is obviously invariant
in this transformation (because this is just φ → −φ!).

Problem 3.6 The equation of motion is given by the Proca equation.

∂µ Fµν + m2 Aν = Jν

Taking divergence we have

∂µ ∂ν Fµν + m2 ∂ν Aν = 0
∂ ν Aν = 0

The last line we use the fact the Fµν is anti-symmetric so the double derivative vanishes.

Send comments to [email protected] 4

Schwartz - QFT and the Standard Model To Chin Yu

The equation for A0 for a point charge is

∂µ (∂µ A0 − ∂0 Aµ ) + m2 A0 = eδ (3) (x)
( + m2 )A0 = eδ (3) (x)
e
A0 = δ (3) (x)
−∆ + m2
d3 k
Z
e
A0 = eik·x
(2π)3 k 2 + m2
Z ∞
e k2 eikr − e−ikr
= dk
(2π)2 0 k 2 + m2 ikr
Z ∞ Z ∞
e h k ikr k −ikr
i
= dk e − dk e
4π 2 ir 0 k 2 + m2 0 k 2 + m2
Z ∞
e k
= dk eikr
4π 2 ir −∞ k 2 + m2
where in the second line we used the fact that the potential is static so its time derivative
vanished.
Evaluating this with contour integration gives
e im −mr
A0 = e
2πr 2im
e −mr
= e
4πr
In the limit m → 0 (massless photon), we recover the Coulomb potential A0 = e/4πr, as
expected.
The main dierence between the Yukawa potential and the Coulomb potential is that the Yukawa
potential has a characteristic range given by R ∼ 1/m.
For this to be a candidate for force between proton, the characteristic range should be on the
order of the proton radius, which corresponds to a mass of m ∼ 1 fm−1 ≈ 200 MeV.
Let us try to substitute the gauge constraint into the Lagrangian.
1 1
L = − (∂µ Aν ∂µ Aν − ∂µ Aν ∂ν Aµ ) + m2 A2µ − Aµ Jµ
2 2
1 1 2 2
= (Aν Aν − Aν ∂ν ∂µ Aµ ) + m Aµ − Aµ Jµ
2 2
1
= Aµ ( + m2 )Aµ − Aµ Jµ
2
where in the second line we have integrated by parts.
The equation of motion becomes
Aµ + m2 Aµ = Jµ
The constraint now becomes
( + m2 )(∂µ Aµ ) = 0
0=0
In the original Lagrangian the mass acts as a Lagrange multiplier. If we turn o the mass the
constraint will vanish.

Send comments to [email protected] 5

Schwartz - QFT and the Standard Model To Chin Yu

Problem 3.7 The Lagrangian has the unit of energy per volume, i.e. mass power is 4. Thus,
looking at the rst term, we know that h2 has unit of energy per length so h has mass power 1.
Therefore a = b = −1.
The rst order solution is just
MP−1
l m
h(1) = −
4πr
The appearance of the additional 4π is just because some factor of 4π has been dropped in the
Lagrangian. We can drop the factor of 4π and instead write
MP−1
l m
h(1) = −
r
The second-order correction is given by
h(2) = MP−1
l (h
(1) 2
)
h(2) ∼ MP−1
l (h
(1) 2
)
MP−3
l m
2
= 2
r
The orbital frequency and the Newtonian potential u = MP−1l h are related by
∂u
ω2 R = | |r=R
∂r
GN mSun
ω2 =
R3
ω ∼ 10 s
−7 −1

The correction to ω is given by

1
δω = δ(ω 2 )
2ω
1 G2N m2Sun
=
ω c2 R4
∼ 10−14 s−1
∼ 10 arcsec/century

Note that factors of c has to be restored as needed to produce the correct dimension.
The eect from other planets can be estimated to be
1 X Gmi
δω ∼
ω ri3
∼ 10−12 s
∼ 103 arcsec/century

If we derive (3.91) from (3.90), we get an additional term ∼ MP−1l hh. This term is the same
order as MP−1l h2 . For order-of-magnitude estimates dropping a few terms with the same order
doesn't really matter.

Send comments to [email protected] 6

Schwartz - QFT and the Standard Model To Chin Yu

Problem 3.9 Let us substitute the gauge condition ∂µ Aµ into the Lagrangian a la problem 3.6
to get
1
L= Aµ Aµ − Jµ Aµ
2
The equation of motion is given by
Aµ = Jµ

Substituting this back to the Lagrangian gives

1
L = − Aµ Aµ
2
d4 k d4 k 0
Z
1 1 0
=− 4 4
Jµ (k) 2 Jµ (k 0 )ei(k+k )x
2 (2π) (2π) k
d4 k
Z
1 1
∼− Jµ (k) 2 Jµ (−k)
2 (2π)4 k

Choosing kµ = (ω, k, 0, 0), we have, in momentum space,

kµ Jµ (k) = 0
ωJ0 (k) = kJ1 (k)

Therefore we can rewrite the Lagrangian as

1
L ∼ Jµ (k) Jµ (−k)
k2
1 h ω2 i
= 2 2 J0 (k)J0 (−k) − 2 J0 (k)J0 (−k) − J2 (k)J2 (−k) − J3 (k)J3 (−k)
ω −k k
1 1 h i
= − 2 J0 (k)J0 (−k) − 2 J 2 (k)J 2 (−k) + J 3 (k)J3 (−k)
k ω − k2
Thus the time derivative of J0 disappeared.

d4 k d4 k 0 1
Z
1 0
L∼ 4 4 2 J0 (k)J0 (k 0 )ei(k+k )x + · · ·
2 (2π) (2π) k
d3 k d3 k 0
Z
1 1 0
= 3 3
J0 (t, k) 2 J0 (t, k0 )e−i(k+k )·x + · · ·
2 (2π) (2π) k
If we transform back to position space, we note that the rst term couples J0 at dierent points
in space but at the same moment in time, thus being non-local. However, this non-local degree of
freedom is not physical as it can be removed by further gauge xing. For example choosing the
gauge A0 = 0 will remove all appearances of J0 in the Lagrangian.

Send comments to [email protected] 7

Schwartz - QFT and the Standard Model To Chin Yu

Problem 4.1 Both electrons and muons couple to the photon eld with the same strength. Thus
the interaction can be written as
Z
1
V = e d3 x ψe φψe + ψµ φψµ
2
Note that here we have neglected the dierence between particle and anti-particle but for a
rough argument like this it doesn't really matter.
Again the rst order term is zero because both initial and nal states contain no photon.
The retarded and advanced intermediate states are |n(R) i = |φγ i and |n(A) i = |ψe1 ψe2 φγ ψµ3 ψµ4 i.
The relevant matrix elements are
Z
(R) 1
Vni = ehφγ | d3 x ψe φψe |ψ1e ψ2e i
2
= e(2π)3 δ (3) (p1 + p2 − pγ )
Z
(R) 1
Vf n = ehψ3µ ψ4µ | d3 x ψµ φψµ |φγ i
2
= e(2π)3 δ (3) (p3 + p4 − pγ )
(A)
Vni = hψe1 ψe2 φγ ψµ3 ψµ†4 |V |ψe1 ψe2 i
= hφγ ψµ3 ψµ†4 |V |0ihψe1 ψe†2 |ψe1 ψe†2 i
= e(2π)3 δ (3) (p3 + p4 + pγ )
(A)
Vf n = e(2π)3 δ (3) (p1 + p2 + pγ )

Therefore the transfer matrix element is

Z h δ (3) (p + p − p )δ (3) (p + p − p ) i
(R) 3 4 γ 1 2 γ
Tf i = e2 dpγ (2π)6
Ei − En
1
= e2
E1 + E2 − Eγ
2 1
=e 0
E − Eγ

Z h δ (3) (p + p + p )δ (3) (p + p + p ) i
(A) 3 4 γ 1 2 γ
Tf i = e2 dpγ (2π)6
Ei − En
1
= e2
E1 + E2 − (E1 + E2 + Eγ + E3 + E4 )
1
= e2
−(E3 + E4 ) − Eγ
1
= e2 0
−E − Eγ

where E 0 = E1 + E2 = E3 + E4 . If we let k = (E 0 , pγ ) be the virtual o-shell momentum of the

Send comments to [email protected] 1

Schwartz - QFT and the Standard Model To Chin Yu

photon, then the transfer matrix is simply

(R) (A)
Tf i = Tf i + Tf i
2Eγ
= e2
E 02− Eγ2
 e2 
= 2Eγ
k2

Send comments to [email protected] 2

Schwartz - QFT and the Standard Model To Chin Yu

Problem 5.1

X d3 pf d3 pB 1 1
dΠLIP S = (2π)4 δ 4 ( p)
(2π)3 (2π)3 2Ef 2EB
Z
dΩ 1 1
= 2
δ(Ef + EB − Ei − mA )p2f dpf
16π Ef EB

To proceed we change variable from pf to x(pf ) = Ef + EB − Ei − mA using

dx dEf dEB pf pf + pi cos θ

= + = +
dpf dpf dpf Ef EB
dΩ h 2 1 i
dΠLIP S = p
16π 2 f EB pf + Ef pf + Ef pi cos θ
dΩ h 1 i
= p pf
16π 2 EB + Ef (1 + pfi cos θ)
dσ 1 h 1 ip
f
= 2 pi |M|2
dΩ 64π mA EB + Ef (1 + pf cos θ) pi

where we have used vi = pi /Ei .

Problem 5.2 See problem 2.6

Problem 5.3 Let us work in the rest frame of the decaying muon and take the direction of the
outgoing electron neutrino to be the z -axis.

X d3 pe d3 pνe d3 pνµ 1 1 1
dΠLIP S = (2π)4 δ 4 ( p)
(2π)3 (2π)3 (2π)3 2Ee 2Eνe 2Eνµ
Z
1 4πpe dpe EdEdΩ
=
8(2π)5
p
p2e + E 2 + 2pe E cos θ
Z Z 1
1 d cos θ
= 3
δ(m − pe − E − p νµ )pe dp e EdE p
4(2π) −1 pe + E 2 + 2pe E cos θ
2
Z
1
= δ(m − pe − E − pνµ )dpe EdE
4(2π)3
1
= EdE
4(2π)3
Z m/2
32G2F
Γ= (m2 − 2mE)E 2 dE
8(2π)3 0
G2 m5
= F 3
192π
Note that the upper bound for E is m/2 because otherwise there is no way to balance the
momentum such that sum of momentum is zero.

Send comments to [email protected] 1

Schwartz - QFT and the Standard Model To Chin Yu

If we put the muon mass m = 106 MeV, we get

Γ = 3 × 10−16 MeV
τ = Γ−1 = 6.5 × 1017 fm
= 2.18 µs

The percentage discrepancy is around 1%, which might be due to non-zero electron mass (me ∼
0.5 MeV ∼ 1% mµ ).

Problem 5.5 The classical Rutherford cross-section is given by

dσ Z 2 α2
= 2 sin4 (θ/2)
dΩ 4EK

where α = 4π is the ne-structure constant. This assumes that the alpha particle is non-
2
e

relativistic and the nucleus is heavy enough that we can neglect the recoil.
We replace e2 by 2Ze2 and me by mα to get
dσ 16Z 2 α2 m2α
=
dΩ k4

The momentum of the virtual photon is just equal to the change of momentum in the alpha
particle which is given by
kµ = mα v(0, cos θ − 1, sin θ, 0)
k 2 = 2m2α v 2 (1 − cos θ)
= 8mα EK sin2 (θ/2)

Therefore
dσ Z 2 α2
= 2 sin4 (θ/2)
dΩ 4EK
which matches with the classical expression.
There is no real reason why the leading contribution from both classical and quantum mechanics
should match so there is no way to know ahead of time. Rutherford was very lucky in this regard.
For electron-electron (Møller) scattering this formula breaks down because we can no longer
neglect the recoil and treat the other electron as a static eld. Also there are important contributions
coming from the spin structure of the electrons.

Send comments to [email protected] 2

Schwartz - QFT and the Standard Model To Chin Yu

Problem 6.1 We rst calculate the case for ∆t = 0 and promote our answer using Lorentz
invariance.

d4 k
Z
i
DF (x1 − x2 ) = eik(x1 −x2 )
(2π)4 k 2 − m2 + i
d3 k dk 0
Z
i 0
= eik (t1 −t2 ) e−ik·(x1 −x2 )
(2π)3 2π (k 0 )2 − ω 2 + i
d3 k
Z
1
= e−iω|t1 −t2 | e−ik·(x1 −x2 )
(2π)3 2ω + i
Z
dk k
=i e−iω|∆t| (e−ikr − eikr )
(2π)2 2rω + i
Z
dk k
= √ sin kr
(2π)2 r k 2 + m2 + i
m
= K1 (mr)
4π 2 r + i
1
→
4π 2 r2 + i
Therefore by Lorentz symmetry the propagator must be
1 1
DF = − =− 2
4π 2 (∆t2 − r2 ) + i 4π (x1 − x2 )2 + i

Problem 6.2

d3 k 1
Z
0
DR = θ(x − y ) 0
(e−ik(x−y) − eik(x−y) )
(2π)3 2ωk
d3 k 1 ik·(x−y) −iωk (x0 −y0 )
Z
0 0
0
= θ(x − y )0
e (e − eiωk (x −y ) )
(2π)3 2ωk
d3 k ik·(x−y) dk 0 −1
Z Z
0 0 0
= e e−ik (x −y )
(2π)3 2πi (k 0 + i)2 − ωk2
d4 k e−ik(x−y)
Z
=i
(2π)4 (k 0 + i)2 − k 2 − m2

where we have use the mathematical identity

(
1 −iωk (x0 −y 0 ) 0
−y 0 )
dk 0 −1 − eiωk (x ) x0 − y 0 > 0
Z
2ωk (e
0 0 0
e−ik (x −y ) =
2πi (k 0 + i)2 − ωk2 0 x0 − y 0 ≤ 0
1 0 0 0 0
= θ(x0 − y 0 ) (e−iωk (x −y ) − eiωk (x −y ) )
2ωk
Similarly we have

Send comments to [email protected] 1

Schwartz - QFT and the Standard Model To Chin Yu

d3 k 1
Z
0
DA = θ(y − x ) 0
(e−ik(x−y) − eik(x−y) )
(2π)3 2ωk
d3 k 1 ik·(x−y) iωk (y0 −x0 )
Z
0 0
0
= θ(y − x ) 0
3
e (e − e−iωk (y −x ) )
(2π) 2ωk
d3 k ik·(x−y) dk 0
Z Z
1 0 0 0
= 3
e 0 2 2 eik (y −x )
(2π) 2πi (k − i) − ωk
Z 4 −ik(x−y)
d k e
= −i
(2π)4 (k 0 − i)2 − k 2 − m2

Problem 6.3 The operator O can be written in terms of its matrix elements:
XZ
O= dn qdm p |q1 q2 ...qn iOnm (q1 , ..., qn ; p1 , ..., pm )hp1 p2 ...pm |
nm
We claim this is equivalent to
XZ
O= dn qdm p a†q1 ...a†qn ap1 ...apm Cnm (q1 , ..., qn ; p1 , ..., pm )
nm

Note that the ordering of q 's and p's does not matter because all the creation operators commute
with each other and all the annihilation operators commute with each other. Thus both Onm and
Cnm have the same permutation symmetry in their arguments1 . This is important for the proof to
work.
We have to show that we can reproduce all the matrix elements correctly. First it is obvious that
we can reproduce the vacuum expectation by choosing C00 = O00 . Now assume we have chosen
Cnm such that all matrix elements of order up to n = N and m = M are correctly reproduced.
Let us consider the case when n = N + 1, m = M . For simplicity we have dropped normalization
factors which does not aect the result.
XZ
O(N +1)M (q, qN +1 ; p) = h0|aq1 ... aqN +1 dn q 0 dm p0 a†q0 ...a†qn0 ap01 ...ap0m Cnm (q10 , ..., qn0 ; p01 , ..., p0m ) |pi
1
nm

On the right hand side there will be two type of terms. The rst type involves the commutators
[aqN +1 , a†q0 ] which gives terms involving C(N +1)M . There are (N + 1)!M ! such terms due to the
i
dierent combinations and although the dierent combinations result in dierent ordering among
the q 's and the p's, the value of C(N +1)M for all these permutations of the momenta is the same.
The second group of terms involves commutators [aqN +1 , a†pi ] between operators in the left and right
states. The remaining operators will pair up resulting in terms including Cnm of order n ≤ N, m ≤
M which have been already xed due to our assumption. Therefore we have

O(N +1)M (q; p) = (N + 1)!M !C(N +1)M (q; p) + (xed terms)

which gives us the correct choice for C(N +1)M (q; p). The case where n = N and m = M + 1 is
almost identical. Therefore by induction the claim is true.
1 The proof also works for anti-commutating operators thanks to this because Onm and Cnm will pick up the
same minus sign which then will cancel out.

Send comments to [email protected] 2

Schwartz - QFT and the Standard Model To Chin Yu

Problem 7.1

M=g

b
i i i
iM ∼ (ig)3
k12 2 2
− m + iε k2 − m + iε k3 − m2 + iε
2 2

We need to conserve momentum. Take clockwise as the positive direction for the internal
momenta k1 , k2 , k3 , we have

p1 + k3 = k1
p2 + k2 = k1
p3 + k3 = k2

Thus, let k3 = k , we have

d4 k
Z
1 1 1
iM = −g 3
(2π)4 (p1 + k)2 − m2 + iε (p3 + k)2 − m2 + iε k 2 − m2 + iε

c In position space we have

Z
3
hφ(a)φ(b)φ(c)i = −ig d4 x d4 y d4 z DF (a, x)DF (x, y)DF (y, b)DF (y, z)DF (z, c)DF (z, x)

d By the LSZ formula (we suppress iε for clarity)

Send comments to [email protected] 1

Schwartz - QFT and the Standard Model To Chin Yu

h Z ih i h i  Z 
4 −ip1 a 2 3
iM = i d a e (a + m ) ... ... − ig d4 x d4 y d4 z DF (a, x)DF (x, y)DF (y, b)DF (y, z)DF (z, c)DF (z, x)
b c
Z
3 4 4 4 4 ip1 x
= −ig d xd xd yd ze DF (x, y)eip2 y DF (y, z)eip3 z DF (z, x)

d4 k1
Z
1 1 1
= −g 3 d4 x... ... eip1 x eik1 (x−y) eip2 y eik2 (y−z) eip3 z eik3 (z−x)
(2π)4 k12 − m2 k22 − m2 k32 − m2
Z
1 1 1
= −g 3 d4 k1 ... δ (4) (p1 + k1 − k3 )δ (4) (p2 − k1 + k2 }δ (4) (p3 + k3 − k2 ) 2 2 2
k1 − m k2 − m k3 − m2
2 2
Z
1 1 1
= −g 3 δ (4) (p1 − p2 − p3 ) d4 k
(p1 − k)2 − m2 (p3 − k) − m2 k 2 − m2
d4 k
Z
3 1 1 1
= −g
(2π) (p1 + k) − m (p3 + k) − m k − m2
4 2 2 2 2

which is exactly the same as the expression in (b). In the last line we dropped the delta function
of total momentum.

Problem 7.2 Using the LSZ formula gives us

M6pt = λ

X
δ (4) ( p)M3pt = g 2 δ (4) (p1 − p3 − p4 )δ (4) (p2 − p5 − p6 ) + (permutations)
The 3-point amplitude has an additional delta function, as pointed out in section 7.3.2.

Problem 7.3

a We have the usual s,t and u channel scatterings.

b The s channel is forbidden due to charge conservation.

c
im2e e2 im2e e2
iM = −
t u

d The connected pair of electrons must have the same spin. Therefore there are 2×2 = 4
combinations for each of t and u processes. Out of these processes only 2 are common to both
and hence can interfere. The remaining 8 processes with unequal number of up/down spins are
forbidden.

Send comments to [email protected] 2

Schwartz - QFT and the Standard Model To Chin Yu

e
h1 1 1  2 i
|M|2 = m4e e4 + −
u2 t2 2 ut
4 4h
me e 1 1 1 i
= 4 2
+ 2
− 2
4p (1 − cos θ) (1 + cos θ) 1 − cos θ
(cross-section?)

Problem 7.4

a Neglecting the disconnected graphs, we have

etc.
The black dot denotes mass coupling m2 .

G0 = DF (x, y)
Z
i
= d4 k eik(x−y) 2
k
Z
2 4
G1 = −im d u DF (x, u)DF (u, y)
Z  i 2
= −im2 d4 k eik(x−y) 2
k
Z
2 2 4 4
G2 = (−im ) d u d v DF (x, u)DF (u, v)DF (v, y)
Z  i 3
= (−im2 )2 d4 k eik(x−y) 2
k
etc.

c Summing over the series gives

Z
i X i n
G= d4 k 2
− im2 2
k n k
Z
i 1
= d4 k 2 2
k 1− mk2
Z
i
= d4 k
k2 − m2
which is just the massive propagator.

Send comments to [email protected] 3

Schwartz - QFT and the Standard Model To Chin Yu

d The equation for the propagator is

G + m2 G = −iδ (4) (x − y)

The zeroth order solution is Z

i
G(0) = d4 k eik(x−y)
k2
The rst order solution is

G(1) + m2 G(0) = 0
−m2 (0)
G(1) = G

i m2
Z
G(1) = d4 k eik(x−y) 2 2
k k
All higher orders are given by the recursion

−m2 (n)
G(n+1) = G

which leads to Z  m2 n i
G(n) = d4 k eik(x−y)
k2 k2
which again sums to the massive propagator.

Problem 7.5

Problem 7.8

1 1 1 1 1
L = − ee − µ( + m2µ )µ − νµ νµ − νe νe − W ( + m2W )W + g(µνµ W + eνe W )
2 2 2 2 2

b We work in the center of momentum frame and take the direction of the outgoing muon
neutrino as the z -axis.

i
iM = (ig)2
(pµ − pνµ )2 − m2W
g4  2(m2µ − mµ Eνµ ) 
|M|2 = 4 1 +
mW m2W

Send comments to [email protected] 4

Schwartz - QFT and the Standard Model To Chin Yu

c M is dimensionless so g appears to have dimension of mass. We can make it dimensionless

by inserting mµ .

g 4 m4µ  2(m2µ − mµ Eνµ ) 

|M|2 = 4 1+
mW m2W

The decay rate is (after integrating over Eνµ )

g 4 m5µ  1 3m2µ 
Γ= +
192π 3 m4W 2 4m2W

d To give an order of magnitude estimate of mW , let us take g ∼ 0.1. Ignoring the second
term in the bracket we can estimate mW to be

m
W
4 m5µ T
≈
g 192π 3 2
mW ≈ 250g GeV
mW ∼ O(10) GeV

e We take the ratio of lifetimes to have

 m 5 Tµ
τ
=
mµ Tτ
mτ ≈ 2500 MeV

Γ = 17.8%Γtot
mτ ≈ 1767 MeV

g We can measure the decay width of muon up to leading order to get the ratio g/mW and
then measure the decay width of tau up to NLO to t out mW and thus g. Of course we can also
measure both widths up to NLO, which would require much higher precision. Assuming we know
the mass of muon and tau to high precision, either of these will allow us to t both g and mW ,
rather than just the ratio. The minimum precision we need is on the order of

∆T  m 2
τ
∼ ∼ 0.05%
T mW

Problem 7.7

Send comments to [email protected] 5

Schwartz - QFT and the Standard Model To Chin Yu

a Due to the external legs the symmetry factor is 1. To write down the amplitude using Feyn-
man rules we rst need to fully implement all the momentum conservations at vertices. Denoting
the external momenta as pi 's (all outgoing except p1 ) and the internal momenta as ki 's, we have

p1 = k1 + k2 + k3
p2 = k1 − k4 − k5
p3 = k3 + k5 + k6
p4 = k2 + k4 − k6
This is 4 equations for 6 unknowns but only 3 equations are linearly independent due to total
momentum conservation. Writing everything in terms of k1 , k2 , k3 and pi 's, we have

1
k4 = (p1 − k1 − k2 − k3 )
2
1
k5 = (p3 + p4 − p2 + k1 − k2 − k3 )
2
1
k6 = (p3 − p4 + p2 − k1 + k2 − k3 )
2
Integrating over the undetermined momenta, we have

d4 k1 d4 k2 d4 k3 1 1 1 1 1 1
Z
iM = −λ4
(2π)12 k12 k22 k32 k42 k52 k62
with k4 , k5 , k6 given above.

b The symmetry factor simply corresponds to permuatation of the vertices, i.e. S = 4!. Thus,

λ4 d4 k1 d4 k2 d4 k3 1 1 1 1 1 1
Z
iM = −
4! (2π)12 k12 k22 k32 k42 k52 k62
with all pi 's set to zero.

Problem 7.9

a For s-channel the mediating particle has momentum s = (p1 + p2 )2 thus

1
|M|2 ∼ | |2
s − m2 + imΓ
1 1
σ∼
s (s − m )2 + m2 Γ2
2

b Let us take the constant of proportionally to be m6 by changing to suitable units, and let
2
x = s/m
1 1
σ(x) =  2
x Γ
(x − 1)2 + m

Γ/m large the cross-section looks 1/x for small x and decays as x−3 for large x (left below).
For
For Γ/m small the cross-section has a sharp peak around x = 1 (right below).

Send comments to [email protected] 6

Schwartz - QFT and the Standard Model To Chin Yu

1
M=
p2 − m2 + iε
ε
Im(M) = − lim
ε→0 (p2 − m2 )2 + ε2
= −πδ(p2 − m2 )

where we have use the limit representation of delta function given by limx→0 x/(x2 +t2 ) = πδ(t).
Therefore when the particle is o-shell the imaginary part is zero.

d Assume that the amplitude of the loop diagram is Mloop = A + iΣ. Then we have

(dressed)
DF = DF + DF (iMloop )DF + DF (iMloop )DF (iMloop )DF + ...
i 1
= 2
p − m20 1 + pA+iΣ
2 −m2

i
=
p2 − (m20 − A) + iΣ
i
=
p2 − m2 + iΣ
The real part of the loop diagram leads to some renormalization of the mass which we are not
interested in. The imaginary part leads to a decay width given by

Γ = Σ/m

e We can interpret the results of parts c and d using virtual particles. When a particle is
propagating it is constantly turning into some virtual particles and back due to interaction with the
non-trivial vacuum. For example in part d φ is constantly turning into ψ 's and back to φ. When
the kinematics allows, the virtual particles can go on-shell and materialize. Physically then one
observes a decay from the original particle to these newly materialized particles. Mathematically,
as we have seen in part c, the on-shelling leads to a non-zero imaginary part in the amplitude.
This in turn lead to a decay width in the propagator as shown in part d. We also see in part d
that we expect the decay width to be proportional to the strength of interaction. This makes sense
because the stronger the interaction with the virtual sea the higher the chance that the original
particle will turn into the decay product.

Send comments to [email protected] 7