Quantum Field Theory (2019-20) Toby Wiseman

Problem sheet 1 Solutions

Qu. 1 Solution
Using kg one can solve c = 3.00 ⇥ 108 m s 1
and ~ = 6.63 ⇥ 10 34
kg m2 s 1

for meters and seconds finding;

m = 4.52 ⇥ 1041 kg 1

s = 1.36 ⇥ 1050 kg 1

and then noting that the volume of 1kg of water is 1m3 , and assume 5 minutes
' 300s to drink a cup of tea, one finds;

1 kg water mass 1 kg = 1 kg (!)

1 kg water volume 1 m3 = 9 ⇥ 10124 kg 3
cup of tea 300 s = 4.1 ⇥ 1052 kg 1

Now solving c = 1 and ~ = 1 together with eV = 1.60 ⇥ 10 19

kg m2 s 2 , one
can eliminate all units in terms of eV to find;

kg = 5.63 ⇥ 1035 eV
m = 8.04 ⇥ 105 eV 1

s = 2.41 ⇥ 1014 eV 1

and then one finds,

1 kg water mass 1 kg = 5.6 ⇥ 1035 eV

1 kg water volume 1 m3 = 5.2 ⇥ 1017 eV 3

cup of tea 300 s = 7.2 ⇥ 1016 eV 1

1
Qu. 2 Solution
The ⇤µ ⌫ preserve ⌘µ⌫ so that,
µ µ ⌫
⌘↵ = ( ↵ + ✏v ↵ ) + ✏v ⌫ ⌘µ⌫ + O(✏2 )
µ ⌫ µ ⌫ ⌫ µ
= ↵ ⌘µ⌫ + ✏ ↵v + v ↵ ⌘µ⌫ + O(✏2 )
= ⌘↵ + ✏ v ⌫ ⌘↵⌫ + v µ ↵ ⌘µ + O(✏2 )
= ⌘↵ + ✏ (v↵ + v ↵ ) + O(✏2 )

and thus vµ⌫ = v[µ⌫] ie. it is antisymmetric.

Then we have

⇤µ ⌫ = µ
⌫ + ✏v µ ⌫
µ
= ⌫ + ✏⌘ µ↵ v↵⌫
µ
= ⌫ + ✏⌘ µ↵ v[↵⌫]

where v[↵⌫] is a general antisymmetric matrix. Now taking,

i
⇤µ ⌫ = µ
⌫ !↵ (M ↵ )µ ⌫ + . . . , (M ↵ )µ⌫ = i ↵
µ ⌫
↵
⌫ µ
2
we have,
i
⇤µ ⌫ = µ
⌫ + !↵ ⌘ µ⇢ (M ↵ )⇢⌫
2
µ 1 µ⇢ ↵ ↵
= ⌫ + !↵ ⌘ ⇢ ⌫ ⌫ ⇢
2
µ 1 µ⇢
= ⌫ + ⌘ (!⇢⌫ !⌫⇢ )
2
µ µ⇢
= ⌫ + ⌘ ![⇢⌫]

so we see that indeed (M ↵ )µ ⌫ is the correct generator, and the parameter

!µ⌫ is simply v µ ⌫ = ⌘ µ↵ !↵⌫ .
We may confirm that ⇤boost and ⇤rot are Lorentz matrices by explicit
matrix multiplication to show they preserve the Minkowski metric; for ex-
ample,
0 1 0 1 0 1
chr shr 0 0 1 0 0 0 chr shr 0 0
B C
shr chr 0 0 C B 0 B 1 0 C B
0 C B shr chr 0 0 C
⇤Tboost · ⌘ · ⇤boost = B
@ 0 · · C
0 1 0 A @ 0 0 1 0 A @ 0 0 1 0 A
0 0 0 1 0 0 0 1 0 0 0 1

2
 0 1 0 1
chr shr 0 0 chr shr 0 0
B shr chr 0 0 C B +shr chr 0 0 C
= B
@ 0
C·B C
0 1 0 A @ 0 0 1 0 A
0 0 0 1 0 0 0 1
0 1
cosh2 r sinh2 r 0 0 0
B 0 cosh r + sinh2 r
2
0 0 C
= B
@
C
0 0 1 0 A
0 0 0 1
0 1
1 0 0 0
B 0 1 0 0 C
= B
@ 0
C=⌘
0 1 0 A
0 0 0 1

Consider the rotation for small ✓ – we have,

0 1 0 1
1 0 0 0 0 0 0 0
B 0 1 ✓ 0 C B 0 0 1 0 C
⇤rot ' B
@ 0
C + O(✓2 ) = 1 + ✓ B C + O(✓2 )
✓ 1 0 A @ 0 1 0 0 A
0 0 0 1 0 0 0 0

and comparing with

⇤µ ⌫ = µ
⌫ + ✏v µ ⌫

we see,
0 1
0 0 0 0
B 0 0 1 0 C
vµ ⌫ = ✓B
@ 0
C
1 0 0 A
0 0 0 0

Now !µ⌫ = ⌘µ↵ v ↵ ⌫ , and since ⌘ij = ij , then,

0 1
0 0 0 0
B 0 0 1 0 C
!µ⌫ = ✓ B@ 0
C
1 0 0 A
0 0 0 0

which we note can be written !ij = ✓✏ij3 , with !0i = 0.

3
 Now consider linearising the boost for small rapidity r, and noting that
cosh r ⇠ 1 + O(r2 ), sinh r ⇠ r + O(r3 ) then,
0 1 0 1
1 r 0 0 0 1 0 0
B r 1 0 0 C B C
⇤boost = B C + O(r2 ) = 1 + r B 1 0 0 0 C + O(r2 )
@ 0 0 1 0 A @ 0 0 0 0 A
0 0 0 1 0 0 0 0

so that,
0 1
0 1 0 0
B 1 0 0 0 C
vµ ⌫ = rB
@ 0
C
0 0 0 A
0 0 0 0

Then !µ⌫ = ⌘µ↵ v ↵ ⌫ so using matrix notation;

0 1 0 1 0 1
1 0 0 0 0 1 0 0 0 1 0 0
B 0 1 0 C B
0 C B 1 0 0 0 C B 1 0 0 0 C
!µ⌫ = r B
@ 0 0 · C=B C
1 0 A @ 0 0 0 0 A @ 0 0 0 0 A
0 0 0 1 0 0 0 0 0 0 0 0

and so !01 = r = !10 with other components zero.

Let us compute the commutator;

([M↵ , M⇢ ])µ ⌫ = (M ↵ )µ (M ⇢ ) ⌫ (M ⇢ )µ (M ↵ ) ⌫
↵ µ ⇢ ⇢ µ ↵
= (M ) (M ) ⌫ (M ) (M ) ⌫

Now we note that;

(M↵ )µ⌫ = ⌘ µ⌧ ⌘ ⌫ (M↵ )⌧

= ⌘ µ⌧ ⌘ ⌫ (i) ⌧↵ ↵
⌧
↵µ ⌫ ↵⌫ µ
= i ⌘ ⌘ ⌘ ⌘

so,

([M↵ , M⇢ ])µ ⌫ = i ⌘ ↵µ ⌘ ⌘↵ ⌘ µ
i ⇢
⌫
⇢
⌫

i (⌘ ⇢µ ⌘ ⌘⇢ ⌘ µ) i ↵
⌫
↵
⌫

= ⌘ ↵µ ⌘ ⇢
⌘ ↵⇢ ⌘ µ
⌫ + ⌘ ↵µ ⌘ ⌘↵ ⌘ µ ⇢
⌫

4
 + (⌘ ⇢µ ⌘ ↵
⌘ ⇢↵ ⌘ µ ) ⌫ ⌘ ⇢µ ⌘ ⌘⇢ ⌘ µ ↵
⌫
= ⌘ ↵µ ⌘ ⇢ ⌫ + ⌘ ↵⇢ ⌘ µ ⌫ + ⌘ ⌘ ⌫⇢ ⌘ ↵ ⌘ µ ⌫⇢ ↵µ

+⌘ ⇢µ ⌘ ↵ ⌫ ⌘ ⇢↵ ⌘ µ ⌫ ⌘ ⇢µ ⌘ ⌫↵ + ⌘ ⇢ ⌘ µ ↵
⌫
= ⌘ ↵⇢ ⌘ µ ⌫ ⌘ µ ⌫ + ⌘ ↵ ⌘ ⇢µ ⌫ ⌘ µ ⌫⇢
⌘ (⌘ ⇢µ ↵
⌫ ⌘ ↵µ ⌫⇢ ) ⌘ ⇢
(⌘ ↵µ ⌫ ⌘ µ ↵
⌫)
⇣
= i ⌘ ↵⇢ i⌘ µ ⌫ ⌫ ⌘ ↵ i⌘ µ ⇢
⌫
⇢
⌫
⌘
µ ↵ ⇢ ⇢ ↵ ⇢ µ ↵ ↵
+⌘ i⌘ ⌫ ⌫ ⌘ i⌘ ⌫ ⌫

= i ⌘ ↵⇢ (M )µ ⌫ ⌘ ↵ (M ⇢ )µ ⌫ ⌘ ⇢ (M ↵ )µ ⌫ + ⌘ (M ↵⇢ )µ ⌫

Finally we try (Ji )µ ⌫ = + 12 ✏ijk (M jk )µ ⌫ ; then,

1
([Ji , Jj ])µ ⌫ = ✏iab ✏jmn ([Mab , Mmn ])µ ⌫
4
1
= i✏iab ✏jmn ⌘ am (M bn )µ ⌫ ⌘ an (M bm )µ ⌫ ⌘ bm (M an )µ ⌫ + ⌘ bn (M am )µ ⌫
4
1
= i✏iab ✏jmn am (M bn )µ ⌫ an
(M bm )µ ⌫ bm
(M an )µ ⌫ + bn (M am )µ ⌫
4
1
= i ✏iab ✏jan (M bn )µ ⌫ ✏iab ✏jma (M bm )µ ⌫ ✏iab ✏jbn (M an )µ ⌫ + ✏iab ✏jmb (M am )µ ⌫
4
1
= i ✏aib ✏ajn (M bn )µ ⌫ + ✏aib ✏ajm (M bm )µ ⌫ + ✏bia ✏bjn (M an )µ ⌫ + ✏bia ✏bjm (M am )µ ⌫
4
= i✏aib ✏ajn (M bn )µ ⌫
bn µ
= i ( ij bn in bj ) (M ) ⌫
= i(M ji )µ ⌫
= i(M ij )µ ⌫
ij µ
noting that ij (M ) ⌫ = 0 due to antisymmetry of M . But,

✏ijk ✏kab (M ab )µ ⌫ = ✏kij ✏kab (M ab )µ ⌫

ab µ
= ( ia jb ib ja ) (M ) ⌫
= (M ij )µ ⌫ (M ji )µ ⌫
= 2(M ij )µ ⌫

5
so we see,
1
([Ji , Jj ])µ ⌫ = i(M ij )µ ⌫ = i✏ijk ✏kab (M ab )µ ⌫
2
= i✏ijk (Jk )µ ⌫

One might think it confusing these have spacetime indices, µ, ⌫, not just
space, but if µ = 0 or ⌫ = 0 these generators are trivial so there is no con-
tradiction.

Now consider just these spatial components;

1
(Ji )mn = ✏ijk (Mjk )mn
2
1
= ✏ijk ⌘ mµ i µj nk j k
n µ
2
1
= ✏ijk ( mp )i pj nk j
n
k
p
2
i mp
= ✏ijk pj nk j k
n p
2
i mp
= (✏ipn ✏inp )
2
= i mp ✏ipn

6
Qu. 3 Solution

(i) Consider the Taylor expansion;

1
f (x a) = f (x) aµ @µ f (x) + ( aµ )( a⌫ )@µ @⌫ f (x) + . . .
2
and comparing to,

f (x a) = f (x) iaµ Pµ f (x) + . . .

and since this must hold for any function we see that,

aµ @µ f (x) = iaµ Pµ f (x) , Pµ = i@µ

Now as partial derivatives commute, we immediately derive,

[Pµ , P⌫ ] = [@µ , @⌫ ] = 0

(ii) Let us do the same for a Lorentz transformation. Then from the last
question,
✓✓ ◆ ◆
1 µ i ↵ µ
f (⇤ x) = f ⌫ + !↵ (M ) ⌫ + . . . x⌫
2
✓ ◆
µ i ↵ µ ⌫
= f x + !↵ (M ) ⌫ x + . . .
2
i
= f (xµ ) + !↵ (M ↵ )µ ⌫ x⌫ @µ f (x) + . . .
2
µ 1
= f (x ) !↵ ⌘ µ⇢ ( ⇢↵ ⌫ ↵ ⌫
⌫ ⇢ )x @µ f (x) + . . .
2
µ i
= f (x ) !↵ ⌘ µ⇢ ( ⇢↵ ⌫ ↵ ⌫
⌫ ⇢ )x ( i)@µ f (x) + . . .
2
µ i
= f (x ) !↵ ( ⇢↵ ⌫ ↵ ⌫ µ⇢
⌫ ⇢ )x ⌘ Pµ f (x) + . . .
2
i
= f (xµ ) !↵ ( ⇢↵ ⌫ ↵ ⌫ ⇢
⌫ ⇢ )x P f (x) + . . .
2
i
= f (xµ ) + !↵ (x↵ P x P ↵ )f (x) + . . .
2

7
by again Taylor expanding, and using (M ↵ )µ ⌫ = i⌘ µ⇢ ( ↵
⇢ ⌫
↵
⌫ ⇢)
from the question above. This should equal
i
f (⇤ 1 x) = f (xµ ) !↵ J ↵ f (x) + . . .
2
so we see,

J↵ = x↵ P x P↵

Let us check the algebra;

[J ↵ , J ⇢ ] = [x↵ P x P ↵ , x⇢ P x P ⇢]
= [x↵ P , x⇢ P ] [x↵ P , x P ⇢ ] [x P ↵ , x⇢ P ] + [x P ↵ , x P ⇢ ]

Now (where f is a function) we have,

[x↵ P , x⇢ P ]f = x↵ P (x⇢ P f ) x⇢ P (x↵ P f )

= x↵ (@ x⇢ @ f ) + x⇢ (@ x↵ @ f )
= x↵ (@ x⇢ )@ f + x⇢ (@ x↵ )@ f x↵ x⇢ @ @ f + x⇢ x↵ @ @ f
= x↵ ( ⇢ )@ f + x⇢ ( ↵ )@ f
⇢ ↵ ↵ ⇢
= i x P x P f

so that,

[x↵ P , x⇢ P ] = i ⌘ ⇢ x↵ P ⌘ ↵ x⇢ P

Then,

[J ↵ , J ⇢ ] = i ⌘ ⇢ x↵ P ⌘ ↵ x⇢ P + i ⌘ x↵ P ⇢ ⌘ ↵⇢ x P
+i ⌘ ⇢↵ x P ⌘ x⇢ P ↵ i ⌘ ↵x P ⇢ ⌘ ⇢x P ↵
= i⌘ ↵⇢ x P x P + i⌘ ⇢
(x P ↵ x↵ P )
+i⌘ ↵ x⇢ P x P⇢ i⌘ (x⇢ P ↵ x↵ P ⇢ )
= i⌘ ↵⇢ J + i⌘ ⇢
(J ↵ ) + i⌘ ↵ J ⇢
i⌘ (J ↵⇢ )
= i ⌘ ↵⇢ J ⌘ ⇢J ↵ ⌘↵ J ⇢
+ ⌘ J ↵⇢

which is the previous algebra, replacing M ↵ ! J ↵ .

8
 Now lets consider;

[P µ , J ↵ ] = P µ J ↵ J↵ P µ
= P µ x↵ P x P ↵ + x↵ P x P↵ Pµ
= P µ x ↵ P + x ↵ P Pµ + Pµ x P ↵ x P ↵ P µ
= (P µ x↵ x↵ P µ ) P + P µ x x Pµ P↵
⇥ ⇤
= [x↵ , P µ ] P x ,Pµ P↵

Now noting that (for a function f ),

[xµ , P⌫ ] f = ixµ @⌫ f + i@⌫ (xµ f ) = ixµ @⌫ f + i(@⌫ xµ )f + ixµ @⌫ f

= i(@⌫ xµ )f = i ⌫µ f

so (as usual) we have [xµ , P⌫ ] = µ

⌫, and raising the index,

[xµ , P ⌫ ] = i⌘ µ⌫

and so then,
⇥ ⇤
[P µ , J ↵ ] = [x↵ , P µ ] P x ,Pµ P↵
= i ⌘ µ↵ P ⌘µ P ↵

(iii) Now consider the quantum operator P̂ µ . Under a Lorentz transforma-

tion we have,
1 i ˆ↵ i ˆ↵
Û (⇤)P̂ µ Û (⇤) = e+ 2 !↵ J P̂ µ e 2 !↵ J
✓ ◆ ✓ ◆
i ˆ i
= ↵
1 + !↵ J + . . . P̂ µ
1 !↵ Jˆ↵ + . . .
2 2
i
= P̂ µ !↵ [P̂ µ , Jˆ↵ ] + . . .
2
But using the (operator version of the) Poincare algebra we have,
⇣ ⌘
µ ˆ↵ µ↵ µ ↵
[P̂ , J ] = i ⌘ P̂ ⌘ P̂

so that,
1 ⇣ ⌘
1
Û (⇤)P̂ µ Û (⇤) = P̂ µ + !↵ ⌘ µ↵ P̂ ⌘ µ P̂ ↵ + . . .
2
9
 1
= P̂ µ + !↵ ⌘ ↵µ ⌫ ⌘ µ ⌫↵ P̂ ⌫ + . . .
2
i
= P̂ µ !↵ (M ↵ )µ ⌫ P̂ ⌫ + . . .
✓ 2 ◆
µ i
= ⌫ !↵ (M ) ⌫ + . . . P̂ ⌫
↵ µ
2

but from Qu 2. we have that,

i
⇤µ ⌫ = µ
⌫ !↵ (M ↵ )µ ⌫ + . . .
2

ie. P̂ µ is indeed transforming as a vector should.

10
Qu. 4 Solution

1
(i) Firstly from Û (a)x̂i Û (a) = x̂i + ai we have,

x̂i Û (a) = Û (a)(x̂i + ai )

and so on any position eigenstate |y j i,

x̂i Û (a)|y j i = Û (a)(x̂i + ai )|y j i = Û (a)(y i + ai )|y j i = (y i + ai )Û (a)|y j i

and so x̂i (Û (a)|y j i) = (y i + ai )(Û (a)|y j i) and thus we see,

Û (a)|xj i = |xj + aj i

(ii) Now consider hxi |Û (a)|pj i;

⇣ ⌘†
i 1 i † a)k pk i
hx |Û (a)|pj i = Û (x)|x i |pj i = |xi ai i |pj i = hxi ai ||pj i = ei(x j

But this means,

⇣ ⌘
iak pk kp iak pk
hxi |Û (a)|pj i = e eix k i
j =e hxi ||pj i

Now since this holds for any hxi | this implies,

iak pk
Û (a)|pj i = e |pj i

which we can write using the operator p̂ as,

iak p̂k
Û (a)|pj i = e |pj i

but as this holds for any |pj i this implies,

iak p̂k
Û (a) = e

(iii) Let us use induction; for n = 1 the LHS is,

⇥ i j ⇤ ⇥ ⇤
x̂ , a p̂j = aj x̂i , p̂j = iaj i
j = iai

11
which indeed gives the RHS for n = 1. Assume for induction the
statement holds for n. Then consider;
⇥ i j n+1 ⇤ ⇥ ⇤
x̂ , (a p̂j ) = x̂i , (aj p̂j )n (aj p̂j )
⇥ ⇤ ⇥ ⇤
= x̂i , (aj p̂j )n (aj p̂j ) + (aj p̂j )n x̂i , (aj p̂j )
= inai (aj p̂j )n 1 (aj p̂j ) + iai (aj p̂j )n
= i(n + 1)ai (aj p̂j )n

using [A, BC] = [A, B]C + B[A, C], and hence the statement holds also
for (n + 1). Thus by induction it is true for all n 1.
So then,
h i X ( i)n
i ip̂j aj i
x̂ , e = x̂ , (p̂j aj )n
n=0
n!
X ( i)n ⇥ ⇤
= x̂i , (p̂j aj )n
n=0
n!
X ( i)n X ( 1)n 1
ip̂j aj
= inai (aj p̂j )n 1
= ai (aj p̂j )n 1
= ai e
n=0
n! n=0
(n 1)!

so that,
1 j ip̂j aj
Û (a)x̂i Û (a) = e+ip̂j a x̂i e
j ip̂j aj j ip̂j aj
= e+ip̂j a [x̂i , e ] + e+ip̂j a e x̂i
j ip̂j aj
= e+ip̂j a (ai e ) + x̂i
= x̂i + ai

12
Qu. 5 Solution

(i) Since we have, U (R) 1 x̂i U (R) = Ri j x̂j then,

x̂i U (R) = U (R) Ri j x̂j

so then,

(x̂i U (R))|xi = U (R)Ri j x̂j |xi = U (R)Ri j xj |xi = Ri j xj U (R)|xi

and so we see, x̂i (U (R)|xi) = Ri j xj (U (R)|xi), so,

U (R)|xi = |Ri j xj i = |R · ~xi

(ii) Now for an infinitessimal rotation;

Ri j x̂j = x̂i ink (Jk )i j x̂j + . . .

= x̂i ink ( i im ✏kmj )x̂j + . . .
= x̂i + im ✏mkj nk x̂j + . . .

Now consider,
k k
U (R) 1 x̂i U (R) = ein L̂k x̂i e in L̂k
= x̂i ink [x̂i , L̂k ] + . . .

so using L̂i = ✏ijk x̂j p̂m mk

then,

U (R) 1 x̂i U (R) = x̂i i[x̂i , nk L̂k ] + . . .

= x̂i ink [x̂i , ✏kmn x̂m p̂j jn ] + . . .
= x̂i ink ✏kmn jn x̂i x̂m p̂j x̂m p̂j x̂i + . . .
= x̂i ink ✏kmn jn m
x̂ [x̂i , p̂j ] + . . .
= x̂i ink ✏kmn jn m i
x̂ i j + . . .
= x̂i + in
✏nkm nk x̂m + . . .

which indeed agrees.

13
(iii) Let us check the commutator;
h i
L̂i , L̂j = ✏iab bm ✏jcd dn [x̂a p̂m , x̂c p̂n ]
= ✏iab bm ✏jcd dn
(x̂a p̂m x̂c p̂n x̂c p̂n x̂a p̂m )
= ✏iab bm ✏jcd dn
(x̂a [p̂m , x̂c ]p̂n x̂a x̂c p̂m p̂n x̂c [p̂n , x̂a ]p̂m + x̂c x̂a p̂n p̂m )
= ✏iab bm ✏jcd dn
( x̂a [x̂c , p̂m ]p̂n + x̂c [x̂a , p̂n ]p̂m x̂a x̂c p̂m p̂n + x̂a x̂c p̂m p̂n )
= ✏iab bm ✏jcd dn
( ix̂a m c
p̂n + ix̂c na p̂m )
bc dn a
= i✏iab ✏jcd x̂ p̂n + bm ad x̂c p̂m
bc dn a ad bm c
= i + ✏bia ✏cjd x̂ p̂n ✏aib ✏djc x̂ p̂m
dn a bm c
= i ( ij ad id aj ) x̂ p̂n ( ij bc ic bj ) x̂ p̂m
bn a
= i ib aj x̂ p̂n + ia bj bn x̂a p̂n
bn a
= i ( ia bj ib aj ) x̂ p̂n
km bn a
= i ✏kij ✏mab x̂ p̂n
= i km ✏kij L̂m = i✏ijk L̂k

which indeed is the correct algebra.

14
Qu. 6 Solution

(i) If it were Hermitian so â = ↠then,

[â, ↠] = [â, â] = ââ ââ = 0

but we know that [â, ↠] 6= 0 so it cannot be Hermitian.

(ii) Consider,
†
N̂ † = ↠â = (â)† (↠)† = ↠â = N̂

Then,

[N̂ , â] = [↠â, â] = ↠ââ â↠â

= [↠, â]â + â↠â â↠â
= [â, ↠]â = â

and similarly,

[N̂ , ↠] = [↠â, ↠] = ↠â↠↠↠â

= ↠[â, ↠] + ↠↠â ↠↠â
= ↠[â, ↠] = â†

(iii) So,

N̂ ↠|ni = [N̂ , ↠]|ni + ↠N̂ |ni

= ↠|ni + ↠n|ni
= (n + 1)↠|ni

Hence we must have ↠|ni = k|n + 1i for some (complex) constant k,
since it is an eigenstate of N̂ with eigenvalue (n + 1). To determine k
we just consider the normalization,
✓ ◆† ✓ ◆
1 † 1 †
1 = hn + 1|n + 1i = â |ni â |ni
k k
1
= hn|â↠|ni
|k|2

15
 1 1
= 2
hn|[â, ↠]|ni + 2 hn|↠â|ni
|k| |k|
1 1
= 2
hn|1̂|ni + 2 hn|N̂ |ni
|k| |k|
1 1
= hn|ni + 2 nhn|ni
|k|2 |k|
1
= (1 + n)
|k|2
so that,
p
k =↵ n+1

where |↵| = 1 is a complex phase. So indeed we have,

p
↠|ni = ↵ n + 1|n + 1i

Likewise,

N̂ â|ni = [N̂ , â]|ni + âN̂ |ni

= â|ni + ân|ni
= (n 1)â|ni

so that â|ni = k 0 |n 1i, and again we consider the normalization to

obtain k 0 ;
✓ ◆† ✓ ◆
1 1
1 = hn 1|n 1i = â|ni â|ni
k0 k0
1
= hn|↠â|ni
|k 0 |2
1
= hn|N̂ |ni
|k 0 |2
1
= nhn|ni
|k 0 |2
1
= n
|k 0 |2
so now,
p
k 0 = ↵0 n

16
 where |↵0 | = 1 is again a complex phase, giving,
p
â|ni = ↵0 n|n 1i
p
(iv) Taking |0i then using the previous answer, â|ni = ↵0 n|n 1i,
p
â|0i = ↵0 0| 1i = 0 ⇥ | 1i = 0
p
Now using ↠|ni = ↵ n + 1|n + 1i and choosing ↵ = 1 and relabeling
n ! n 1,
1
|ni = p ↠|n 1i
n
and so,

↠↠↠â†

|ni = p ⇥ p ⇥ . . . ⇥ p ⇥ p |0i
n n 1 2 1
† n
(â )
= p |0i
n!

(v) Consider an eigenstate |ni so that N̂ |ni = n|ni, but n is not necessarily
integer so we can’t just use the form above. Then,

n = nhn|ni = hn|N̂ |ni = hn|↠â|ni

Now writing | i = â|ni, then h | = | i† = (â|ni)† = hn|↠so,

n = (hn|↠)(â|ni) = h | i 0

using the fact that the Hilbert space norm is positive definite.
Let us show n must be an integer. Suppose for contradiction there
p exists
0
a state with positive but non-integer n. Then using â|ni = ↵ n|n 1i,
we could lower the excitation number as,
p p p
âp |ni = (↵0 )p n ⇥ n 1 ⇥ . . . ⇥ n p + 1|n pi
p p
wherepwhat ever we choose for p we note the factors n ⇥ n 1 ⇥
. . . ⇥ n p + 1 are all non-zero (since n was not integer). Then tak-
ing p > n we conclude that âp |ni is proportional to a state |n pi of

17
negative excitations numbers (n p) < 0. But these cannot exist as
we have just shown and hence this is a contradiction.

[ Note that if n is a positive integer, then this arguement no longer goes

through
p p since for p pn there will be a factor of zero in the product
n ⇥ n 1 ⇥ . . . ⇥ n p + 1. ]

18
Qu. 7 Solution

(i) If it were Hermitian so b̂ = b̂† then,

{b̂, b̂† } = {b̂, b̂} = 0
but we know that {b̂, b̂† } =
6 0 so it cannot be Hermitian.
(ii) We have that,
b̂|0i = 0
so we can’t get new states by acting on the vacuum with b̂. By acting
with b̂† we get a new state, |1i ⌘ b̂† |0i. However suppose we try to get
more?
Consider trying,
1⇣ † † ⌘
b̂† |1i = b̂† b̂† |0i = b̂ b̂ + b̂† b̂† |0i
2
1 n † †o
= b̂ , b̂ |0i = 0
2
so that doesn’t work. How about,
n o
b̂|1i = b̂b̂† |0i = b̂, b̂† |0i b̂† b̂|0i
n o
†
= b̂, b̂ |0i
= |0i
but (as in the bosonic case) we see this just ‘lowers’ us back to the
vacuum.
So the only states we can generate from the vacuum by acting with b̂
and b̂† are the vacuum itself and |1i.
(iii) Again defining N̂ = b̂† b̂ then using the above,
⇣ ⌘
†
N̂ |0i = b̂ b̂|0i = 0 = 0|0i
⇣ ⌘
N̂ |1i = b̂† b̂|1i = b̂† |0i = |1i = 1|1i

so indeed N̂ |ni = n|ni for n = 0, 1.

19