Professional Documents
Culture Documents
Chapter 12
12.1 The elements of the set (which we want to show form a group) are
all 2 2 matrices V which are unitary (i.e. V V = VV = I2) and
y y
= V1I2(V1) = I2
y
using rst the unitarity of V2 and then that of V1. So the product of
any two such matrices is a unitary 2 2 matrix. Similarly, det(V1V2) =
(detV1)(detV2) = 1 using the `det V=1' condition for each matrix. So
the product of any two such matrices is a unitary 2 2 matrix with
determinant +1, and therefore the product also belongs to the same
set, verifying property (i).
(ii) Matrix multiplication is associative.
(iii) The identity element is I2.
(iv) The inverse of any element V is V , by the unitarity property.
y
Hence
detVin = (1 + i11)(1 + i22) + 1221
1 + i(11 + 22)
to rst order in , so that the condition detVin = 1 reduces to 11 +
22 = 0, which is just Tr = 0.
12.3 The commutation relations can be checked by straightforward matrix
multiplication with the 2 2 matrices.
12.4 We have
1 1 1 1 1 1
Ti Tj ] = 2 (1)i + 2 (2)i + : : : 2 (A)i 2 (1)j + 2 (2)j + : : : 2 (A)j
1 1 1 1 1 1
= 2 (1)i 2 (1)j + 2 (2)i 2 (2)j + : : : 2 (A)i 2 (A)j
1 1 1
= iijk 2 (1)k + 2 (2)k + : : : 2 (A)k
= iijk Tk (1)
where in the second line we used the fact that the 's associated with
dierent nucleons commute, and in the third line we used equation
(12.28) for each of the independent 's.
12.5 The (j k) element (j =row label, k=column label) of T1(1) is ;i1jk. The
only non-zero entries are for j = 2 k = 3, giving (T1(1))23 = ;i, and for
j = 3 k = 2, giving (T1(1))32 = i. Hence
0 1
0 0 0
T1(1) = B
@ 0 0 ;i CA :
0 i 0
Similarly, 0 1
0 0 i
T2(1) = B
@ 0 0 0 CA
;i 0 0
and 0 1
0 ;i 0
T3(1) = B
@ i 0 0 CA :
0 0 0
It is then a matter of straightforward matrix multiplication to check
(12.47) for the case i j k = 1 2 3.
12.6 We expand equation (12.61):
0
vi = q (12 ; i =2)i (12 + i =2)q
y
= vi ; i 12 q j j iq + i 12 q ij j q
y y
= vi + i 21 j q (ij ; j i)q
y
(2)
as required, where in the second line we have dropped the term quadratic
in the 's, and used 12i = i12 = 12i12 = i, and a summation on
j = 1 2 3 is understood.
12.7 A matrix H is Hermitian if the condition
Hji = Hij
e f c
This becomes
i(;i=2)^q = iT^i( 2 ) q^]
1
2 d (x )
Z
= 21 (^u(x)d^(x) + d^(x)^u(x)) d3x
y y
u^ elds anticommute with other u^ elds, and with d^ elds the only
non-zero anti-commutator is (c.f. (7.44))
y
fu^(x t) u^(y t)g = (x ; y) :
Notice that the elds in this anti-commutator are at equal times t.
On the other hand, the elds in the commutator (A) are apparently
at dierent times x0 1and y0. However, equation (12.102) assures us
that the operators T^ ( 2 ) are in fact independent of time, and so we are
free to choose x0 in (A), and thus in (B) also, equal to y0. With this
understanding, in the second term in (B) the factor u^ (y) can be anti-
commuted rst past u^(x) and then past d^(x), giving +^u (y)d^(x)^u(x),
y y
which cancels with the fourth term of (B). The rst term in (B) can be
written as
u^(x)d^(x)^u (y) = ;u^(x y0)^u (y y0)d^(x y0)
y y
which is therefore the value of the commutator (B). Inserting this into
the left hand side of (A) gives
Z
; 21 3(x ; y) d^(x y0) d3x = ; 12 d^ (y y0)
which is the same as the right hand side, as was to be shown.
12.11 We introduce the notation q^a for the 8-component eld q^ of (12.86): `a'
runs over the two (isospin) values, a = 1 being the `up' component, and
a = 2 being the `down' component, while ` ' runs over the four Dirac
components in u^ and d^. Thus q^11 means the rst spinor component of
u^, and q^32 means the third spinor component of d^, for example. With
this notation, the only non-vanishing anti-commutator is
fq^a(x t) q^b(y t)g = ab3(x ; y):
y
We have, then,
Z
T^i( 2 ) = q^a(x)(i=2)abq^b(x) d3x
1
y
and Z
^Tj( 21 ) = q^c(y)(j =2)cd q^d(y) d3y:
y
The commutator on the left hand side of (12.103) is
1 Z d3x Z d3y ( ) ( ) (^q (x)^q (x)^q (y)^q (y);q^ (y)^q (y)^q (x)^q (x)): (C)
y y
i ab j cd a b d d b
y y
4 c c a
Note that quantities such as (i )ab are single numbers (here, the (a b)
element of the matrix i), so they can be freely moved to any convenient
position in the expression. Also, since the T^ ( 21 )'s are time-independent,
we are free to choose x0 = y0, and we shall assume this has been done,
without indicating it explicitly. Consider the two products of four eld
operators in (C). The rst one may be written as
q^a(x)^qb(x)^qc(y)^qd(y) = q^a(x);q^c(y)^qb(x) + bc3(x ; y)]^qd(y)
y y y y
The rst term here cancels with the rst term of the preceding expres-
sion, so that (C) becomes
1 Z Z d3x d3y ( ) ( ) ^q (x)^q (y) ; q^ (y)^q (x) ] 3(x ; y)
i ab j cd a d bc b ad
y y
4 c
where x0 = y0 is understood. Now note that
(i)ab(j )cdbc = (i)ab(j )bd = (ij )ad
while
(i)ab(j )cdad = (j i)cb:
Thus the expression (C) is
1 Z Z d3x d3y ^q (x)( ) q^ (y) ; q^ (y)( ) q^ (x)] 3(x ; y):
i j ad d j i cb b
y y
4 a c
In the second term, the indices c and b are both summed over, and
can be replaced by a and d, respectively, as in the rst term. The
expression
y
q^a(x)(ij )adq^d(x)
is just
q^ (x)(ij )^q(x)
y
with the explicit isospin and (Dirac) spinor indices inserted, and simi-
larly for the j i term, so nally (C) is
ZZ Z
d x d y q^ (x)i=2 j =2]^q(y) (x;y) = iijk d3x q^ (x)(k =2)^q(x)
3 3 y 3 y
Chapter 13
13.1 For the in nitesimal case, (13.34) is
1 W = i (@ (1 + i g (x)))(1 ; i g (x))
0
2 g 2 2
+ (1 + i g (x)) 1 W (1 ; i g (x))
2 2 2
and expanding this to rst order in gives
1 W i ig @ (x) + 1 W
0
2 g2 2
+ ( 2 ii(x))( 2 j Wj ) + ( 21 j Wj)( ;2ig ii(x))
i g 1
= ; 21 @ (x) + 12 W + ig 2i 2j ; 2j 2i i(x)Wj
= ; 12 @ (x) + 12 W + ig iijk 2k i(x)Wj:
The last term may be re-written as ;gkij 2k i(x)Wj, which is the triple
scalar product ;g 12 (x) W . Equating the components of 21 on
both sides then leads to (13.26).
13.2 Multiplying out the left hand side of (13.79) we obtain
(@ + ieA)(@
+ ieA
) ; (@ + ieA )(@
+ ieA
)
= @ @
+ ie(@ A )
+ ieA @
+ ieA@
;e2AA
; @ @
; ie(@ A)
; ieA@
; ieA @
+ e2A A
= ie(@ A ; @ A)
= ieF
as required.
13.3 The left hand side of (13.82) is
(@ + ig W )(@
( 21 )+ ig W
( 21 ));(@ + ig W )(@
( 21 )+ ig W
( 12 )):
2 2 2 2
After expanding this out some terms cancel and we are left with
ig (@ W ;@ W )
( 21 );g2 1 W 1 W ; 1 W 1 W
( 21 ):
2 2 2 2 2
The second term is
i j j i
;g 2 2 ; 2 2 Wi Wj
( 21 ) = ;g2iijk 2k WiWj
( 21 ) = i2g (;gW W )
( 12 )
2
0
A = A + A:
Then expanding the previous equation, and neglecting terms of second
order (remembering that A is itself rst order), we have
@
+ ig2s ( @ (x))
+ ig2s ( (x))@
+ ig2s ( A )
+ ( i2gs )2( A )( (x))
+ ig2s ( A)
= @
+ ig2s ( A)
+ ig2s ( (x))@
+ ( ig2s )2( (x))( A )
After some cancellations we are left with
( 2a Aa)
= ;( 2a @ a(x))
+igs( 2b b(x))( 2c Ac);( 2c Ac)( 2b b (x)]
:
The last term here is
igs 2b 2c ]b(x)Ac
= igsifbca 2a b(x)Ac
= ;gs 2a fabcb (x)Ac
and so we have
( 2a Aa)
= ;( 2a @ a(x))
; gs ( 2a fabcb (x)Ac)
which is satis ed if (13.90) holds. (We have checked (13.90) for the spe-
cial case of the covariant derivative (13.87) acting on an SU(3) triplet,
but (13.90) is in fact the general in nitesimal gauge transformation for
the octet of gauge elds Aa.)
13.7 We need to evaluate
(@ + ig2s A )(@
+ ig2s A
) ; (@ + i2gs A )(@
+ i2gs A
):
After cancellations we are left with
D D ]
= igs=2(@ A ;@ A )
+(igs)2( 2 A)( 2 A );( 2 A )( 2 A)]
:
The second term is
(igs)2 2b 2c ]AbAc
= (igs)2ifabc 2a AbAc
:
Hence
D D ]
= igs 2a Fa
(c.f. (13.82)) where Fa is as given in (13.95).
13.8 The ingoing X with four-momentum k2 in gure 13.9 is equivalent to an
outgoing X with four-momentum ;k2 we take the polarization vectors
to be real. Then (13.118) is (note that X should be replaced by X )
Z
;i(;ie) 2(e+ik2 x) f@ (3 e ik3 x1 e ik1 x) + 3 e ik3 x@ 1 e ik1 xg d4x
; ; ; ;
The rst term becomes, after a partial integration,
Z
;e 2(;@ e ik2 x)f3 1 e i(k1+k3 ) xg d4 x
; ;
Z
;e (21 )(ik2 3 )e i(k1 +k2 +k3 ) x d4x
;
Chapter 14
14.1 (a) Let's write the transformation (14.14) as
= W
0
where and run over the three colour indices (which we will call just
1, 2 and 3) and where as usual the repeated index is summed over.
We have a similar equation for each of
and
in (14.2), and so
the result of applying an SU(3) transformation to (14.2) is
=
W W W
:
0 0 0
3
and
;2 p
p3 = p3 + 2i p
3 83
0
and similarly for q2 and q3. Inserting these expressions into p2q3 ; p3 q2
0 0 0 0 0 0
we nd
p2q3;p3q2 = p2 q3;p3q2; 2i (1+i2)(p3q1;p1q3); 2i (3+ p1 8)(p2 q3;p3q2):
0 0 0 0
3
Setting Q1 = (p2 q3 ; p3q2) and Q2 = (p3q1 ; p1 q3) we recognize this as
the rst component of the transformation law
Q = (1 ; i =2)Q
0
as required.
14.2 As usual, the repeated indices and are all summed over, each
taking the values 0, 1, 2, 3. The symbol is only non-zero if all its
indices are dierent, when (see below) it has the value +1 or -1. The
rst index () may be chosen arbitrarily i.e. in 4 ways, then the next in
3 ways, the next in 2, leaving the fourth xed, which gives 24 dierent
non-zero symbols. Consider the 6 in which `0' is the rst index these
contribute
0123F^ 01F^ 23 + 0132F^ 01F^ 32 + 0213F^ 02F^ 13
0231F^ 02F^ 31 + 0312F^ 03F^ 12 + 0321F^ 03F^ 21:
0123 is de ned to be +1, and in general is equal to +1 for all
choices of indices which can be reached from `0123' by an even number
of pairwise interchanges it has the value -1 for those choices which are
reached from `0123' by an odd number of interchanges. Thus 0231 =
+1 0321 = ;1, etc. We also have, for example,
1 1
0
The unpolarized cross section will involve the usual sums and averages
over the fermion spins, which will lead to the familiar trace expression
Tru(p1)u(p1)u(p1)u(p1)] = Tr6 p16 p1] = 4p1 p1
0 0 0 0
in the massless limit. Similarly the spin sums for the other fermion yield
a factor 4p2 p2. On the other hand, t^ = (p1 ;p1)2 = ;2p1 p1 = ;2p2 p2,
0 0 0 0
also in the massless limit. Hence the numerator factors from the spin
sums cancel the 1=t^2 factor from the exchange propagator.
14.4 (i) The equality we have to show is in fact the (s l) matrix element of
the matrix product
(r) (r) + (b) (b) + (g) (g) ()
y y y
on the left hand side, and of the unit 3 3 matrix on the right hand
side. Evaluating () using the given colour wavefunctions we nd
0 1 0 1 0 1
X 1 0 0
y B C B
(c2) (c2) = @ 0 A (100) + @ 1 A (010) + @ 0 C
C B A (001)
c2 0 0 1
0 1 0 1 0 1
1 0 0 0 0 0 0 0 0
= B @ 0 0 0 CA + B@ 0 1 0 CA + B@ 0 0 0 CA
0 0 0 0 0 0 0 0 1
= I
as required.
(ii) In (14.72), since the matrix indices have all been indicated explicitly,
quantities such as ad(c3) and m(c1) for example are just numbers and
4 c3 d 3 c 3 c2 s 2 l 2 c1 m 1 r 1 c rs d lm
y
= 14 cdslrm(c )rs(d)lm
= 14 (c )rl(c)lr
= 14 Tr(c c )
(summed over c), as required.
(iii) It is straightforward to check that all the matrices c (c =
1 2 : : : 8) satisfy
Tr(c )2 = 2 no sum on c:
Summing the eight (equal) contributions then gives 16 as required.
14.5 First note that the eect of the discussion after (13.127) is to assure us
that for gluons, just as for photons (see the discussion in section 8.6.3
following equation (8.166)), polarization sums can be performed using
X
(p ) (p ) = ;g : ()
The quantity we have to evaluate is then
1 X jMj2 = e2ae4gs2 1 X v(k ) u(k )u(p ) (6 p1 + 6 p3) v(p )
1
4 spins Q4 4 spins 2 1
2p1 p3 2
;u(p1) (6 2pp2 + p6 p3) v(p2)] () ()
2 3
;v(p2) (6 2pp2 + 6pp3) u(p1)
2 3
+v(p2 ) 2p1 p6 p3) u(p1)]u(k1) v(k2)
( 6 p +
1 3
after doing the usual tricks with the dagger operation and 0 as in sec-
tion 8.2.3. The quantity u(k1) v(k2) is a number (not a matrix) and
can be moved through the other factors and placed next to v(k2) u(k1).
The expression
X 1 v(k ) u(k )u(k ) v(k )
2 1 1 2
spins of e and e
+ ; 2
then yields, in the massless limit,
1 Tr6 k 6 k ]
2 2 1
= 2k1 k2 + k1 k2 ; k1 k2 g ]
which we call L (k1 k2) in (14.56). Apart from the overall factor
e2ae4gs2=Q4, the remaining part of the expression comes from multiplying
out the two .....] brackets, which involve the quark wavefunctions and
will produce the hadron tensor H , and making use also of (). There
are four terms arising from the product of the square brackets. If, for
the moment, we write the square brackets in shorthand notation as
(13)-(23)]-(23)+(13)], we see that the four terms split into two of the
form (13)(13) and (23)(23), and two of the form -(13)(23) and -(23)(13).
Consider then the `(13)(13)' term, which is
X 1 (6 p1 + 6 p3) v(p )v(p ) (6 p1 + 6 p3) u(p )](;g )
u
( p 1 )
2p1 p3 2 2 2p1 p3 1
quark spins 2
= ; 21 Tr 6 p1 (6 2pp1 + p6 p3) 6 p2 (6 2pp1 + p6 p3) ]:
1 3 1 3
We use (J.3) from Appendix J of volume 1, and the cyclic property of
the Trace, to reduce this to
1 Tr(6 p + 6 p )6 p (6 p + 6 p ) 6 p ]:
4(p p )2 1 3 1 1 3 2
1 3
Since (6 p1 = (6 p3
)2 )2 = 0 in the massless limit, this becomes
1 Tr6 p 6 p 6 p 6 p ]:
4(p1 p3)2 3 1 3 2
Using J.6 and (6 p3)2 = 0 this reduces to
1 1 Tr6 p 6 p ] = 1 L (p p ):
(p1 p3 ) 2 3 2 (p1 p3) 2 3
The `(23)(23)' term is similar, and is
X 1 (6 p2 + 6 p3) v(p )v(p ) (6 p2 + 6 p3) u(p )](;g )
u
( p 1 )
2p2 p3 2 2 2p2 p3 1
quark spins 2
= ; 21 Tr6 p1 (6 2pp2 + 6pp3) 6 p2 (6 2pp2 + p6 p3) ]
2 3 2 3
which reduces, following steps similar to those carried out for the rst
term, to
1 L (p p ):
(p p ) 1 3
2 3
We turn now to the `-(13)(23)' term which is
X
; 21 u(p1) (6 2pp1 + 6pp3) v(p2)v(p2) (6 p22p + 6pp3) u(p1)](;g )
quark spins 1 3 2 3
2] 2 2
(q ) = 8e i 0
0
( s) A( s(jq2j) ; s )
where A > 0. Thus near s equation (15.42) becomes
hand, if we start from a value of q2 such that s(jq2j) < s , the right
hand side of (1) will be negative, and so s(jq2j) will increase as q2 goes
to zero, and will get pulled towards s . In either case, s(jq2j) ends up
at s as q2 ! 0.
!
s (jq 2j) ; s
B s ln(jq j= ) = ln (jq2j) ; :
2 2
s s
0 0
;
(4)2
= 1 2 (1 + =2 ln(4) + : : :) 2 ; + : : : (1 ; =2 ln(
) + : : :)
(4)
= (41 )2 2 ; + : : : (1 + =2 ln(4) ; =2 ln(
) + : : :)
1 2
= (4)2 + ln(4) ; ln(
) ; + : : :
as required.
15.7 The dimensions of the elds can be determined from the usual require-
ment that the Action (in d dimensions) is dimensionless. This is easily
done from the kinetic terms. Consider a bosonic eld , for example,
with a kinetic term Z
ddx @@ :
Let d be the mass dimension of then that of @@ is 2d + 2,
while that of the integration measure is ;d. It follows that we must
have
d = (d ;2 2) :
Similarly, for a fermionic eld
with kinetic term
i6 @
, we nd
d = (d ;2 1) :
Now consider the QED interaction
Z
e
A6
ddx:
If the mass dimension of e is de , for this to be dimensionless requires
de + 2(d ; 1)=2 + d=2 ; 1 ; d = 0
and hence
de = (4 ; d)=2 = =2:
15.8 We can illustrate the eect of changing the renormalization scheme
from MS to MS by considering the QED case. We rst evaluate 2]
(q )
2
using dimensional regularization. From (15.74), (15.75), and (15.79) we
obtain 2
(q ) = ;bQED ; + ln 4 ; ln
2] 2
16 Z1
= ; 3 dz (zn 2 + zn 3 + : : : 1) + (zn + zn 1 + : : :z2]
; ; ;
"0 n+1 n 3 ! z2 #1
16 z z z n 1 z n 2
= ; 3 n + 1 + n + 2 n ; 1 + n ; 2 + ::: 3 + 2 + z
;
z ;
1 1 1 0
16 1 1
= ; 3 n + 1 ; n + 2 n + n ; 1 + ::: 3 + 2 + 2 1 1
Chapter 16
16.1 Let's write
1 NX1 1;
2in1 (1 2 )=N1 = N1
NX
1 1;
ein1
N1 n1 =0 e
;
1 n1 =0
where
= 2(1 ; 2)=N1 :
The summation is a simple GP, whose sum is
1 ; eiN1 : (S)
1 ; ei
Now
N1 = 2(1 ; 2):
Consider the case 1 6= 2: then 1 ; 2 is an integer, so that the
numerator of (S) vanishes, and moreover the maximum value of 1 ; 2
is N1 ; 1, so that the denominator does not vanish. Hence (S) is zero
for 1 6= 2. On the other hand, when 1 = 2, we have = 0, and all
the N1 terms in the summation are equal to 1, and (bearing in mind
the normalizing factor 1=N1 ) the result is 1, thus verifying (16.9).
16.2 We have
1 Z dx m22(x)
2
X X 1 2i1n1=N1 ~ 2i2n1 =N1 ~
= 12 m2a e (1)e (2)
n1 1 2 N1 a
X " X #
1
= 2 m2 1 e2n1(1 +2 )=N1 ~(1)~(2)
1 2 N 1 n1
1 X
= 2 m2 1 2 ~(1)~(2) using (16.9)
;
1 2
1 X
= 2 m2 ~(1)~(;1)
1
as in (16.10). In (16.13) the argument `1' has been replaced by the
equivalent variable k1 = 21=L.
16.3 The mass term in (16.17) is essentially the same as in the bosonic case,
equation (16.10), except that the complex conjugation in
will mean
that the factor `exp2i1n1=N1]' will be replaced by `exp;2i1n1=N1 ]',
and hence the function is 1 2 rather than 1 2 , so that
~ and
~
are both evaluated at k1 (note that
~(;k1 ) in (16.19) should be re-
;
2 n1 a 1 2 N1a
; N1 a e 2i1 (n1+1)=N1
~(1)1Ee2i2n2 =N1
~(2)
;
1
i
= a sin(k1 a)
~(k1 )1E
~(k1 )
as in (16.19) (with the same correction to `
~(k1 )'), noting that k1 a =
21a=(N1a) = 21=N1 .
16.4 (16.25) gives
Zx
O(x y) = (x)exp ie y A(x ) dx (y):
y 0 0
= dlim (x )exp
y
f; i eA ( x )d xg ( ( x ) + d dx) ; ]=dx
y
x 0
! dx
= dlim y
d
(x)(1 ; ieA(x)dx)( + dx dx) ; ]=dx y
x 0
!
as in (16.28).
16.5 We evaluate (16.59) as follows. We have
Z
I 1 = exp ; 2m (q2 ; q1)2 + (q1 ; qi)2] dq1:
The exponent is
; 2m (q2)2 ; 2q2q1 + (q1)2 + (q1)2 ; 2q1qi + (qi)2]
= ; m (q1)2 ; q1(q2 + qi) + (q2)2=2 + (qi)2=2]
= ; m fq1 ; (q2 + qi)=2g2 + (q2)2=2 + (qi)2=2 ; (q2 + qi)2=4]
= ; m (q1 )2 + (q2 ; qi)2=4]
0
Z
I 1 = exp ;4m ( q 2 ; q i )2 exp;m(q1 )2=] dq1
0 0
12 ;m
= m exp 4 ( q 2 ; q i)2
The exponent is
; 4m (q2)2 ; 2q2qi + (qi)2] ; m2 (q ) ; 2q q + (q ) ]
32 3 2 2 2
But the rst of these is also the value of the exponent in the con g-
uration f; ; ; ; ;::::;g in which all the spins have the value -1,
since all the products sj sj+1 will then be equal to +1. Similarly, the
con guration f+ ; ; ; ; ; ; ; : : : ;g will lead to the same value of
the exponent in the second case, because here also s1s2 = ;1, and all
other products =+1. In short, each product of spin variables sj sj+1
takes the value +1 twice, once when both sj and sj+1 equal +1, and
once when they both equal -1 similarly, each product takes the value
-1 twice, when one variable is +1 and the other is -1, and vice versa.
Thus the possible values of the exponent are K1 + K2 + K3 + : : :KN 1,
;K1 + K2 + K3 + : : : KN 1, K1 ; K2 + K3 + : : : KN 1, ;K1 ; K2 +
;
+e K1 K2+K3 +:::+KN ;1 KN
; ;
i
+ : : : + e K1 K2 K3 ::: KN ;1KN
; ; ; ; ;
N sn f g n=1
On the other hand, we have
@ZN @ X "NX1 # ;
X ;
= spsp+1exp Kn sn sn+1
f sn g n=1
from which the stated result for hsp sp+1i follows.
In the same way, when we dierentiate the exponent in ZN with re-
spect to Kp Kp+1 : : : Kp+j 1 (note that `Kp+j ' should be replaced by
;
In this product the repeated pairs s2p+1 s2p+2 : : :s2p+j 1 may all be re-
placed by +1, leaving spsp+j as required.
;
ZN = 2 (cosh Kn )
n=1
and so
@ZN = Z tanh K
N p
@Kp
(since the dierentiation replaces cosh Kp by sinh Kp ), and generally
the j derivatives will give
ZN tanh Kp tanh Kp+1 tanh Kp+2 : : : tanh Kp+j :
Hence when all the J 's are equal, we have
hspsp+j i = (tanh K )j
which can be written as
e ja=
;
where a
= ; lntanh K )] :
When T ! 1, we have K ! 0 and tanh K ! K , so that ! lnaK
which tends to zero logarithmically. When T ! 0, we have K ! 1
j j
Chapter 17
17.1 We take the cube of volume # to have sides of length L, so that
# = L3. Periodic boundary conditions in each of the three spa-
tial directions imply that the components of k take discrete values
kx = 2nx=L ky = 2ny =L kz = 2nz =L (cf equation (E.11) in volume
1, and the paragraph following equation (16.6)), so that we may write
k = 2n=L. The three-dimensional generalization of equation (E.12)
is then
1 Z e ik xeik0 xd3x = 1 Z e i2n=L xei2n0=L xd3x = 0
#
;
#
;
nn
where k = 2n =L we could, however, just as well (and somewhat
0 0
2m Z
X X
= 21m d3x 11 ^ak e ik x(;ik) 11 a^k0 (ik ) eik x:
y ; 0
0
#2 k # 2 k0
The integral over x of the product of the two plane waves gives #kk0 ,
by the result of the preceding paragraph the delta function then allows
the sum over k to be performed, leading to the result
0
1 X k2a^ ^a
2m k k k
y
2#2
1 2 1 2
e; ik xe ik y k y k x:
1 ; 2
0
ei 2
0
ei 1
We introduce relative and centre of mass coordinates via
r = x ; y and R = (x + y)=2
which imples
x = R + r=2 and y = R ; r=2
and ZZ Z Z
d3x d3y ! d3R d3r:
Substituting for x and y in terms of R and r, the exponentials become
e ik1 (R+r=2)e 0 ik20 (R r=2)eik2 (0R r0=2)eik1 (R+r=2)
0 0
; ; ; ;
and the integral over r yields v(jk1 ; k1j), verifying the second term0
k1 + k2. The zero-momentum pieces are as follows. (a) We can have all2
0 0
2#
= 2# 1 v(0)fN + X ^a a^ ]2 ; X a^ ^a ]2g
k k
k k k k
y y
0
0 0
2# 1 v(0)N 2
k k
and have neglected the term with no N0 factor. We have therefore
derived the second term on the right hand side of (17.35).
A third zero-momentum contribution in the interaction term in (17.29)
is (c), in which k2 = k1 = 0 k1 = k2 6= 0. This contributes
0 0
1 X v(jk j)N a^ a^ :
1 0 k1 k1
0 y
2# k1
There is also a similar term in which k1 = k2 = 0 k2 = k1 6= 0, which 0 0
contributes the same. Thus these two terms together give (changing
the summation variables to k)
N0 X v(k)^a a^
kk
0 y
# k
which accounts for the second term on the right hand side of (17.36),
after replacing N0 by N .
Finally, there is the term (d) in which k1 = k2 = 0 k1 = ;k2 6= 0, 0 0
which contributes
1 X v(jk j)N a^ 0 a^ 0
2 0 k2 k2
0 0
2# k0 ;
which is the fourth term on the right hand side of (17.35), after replac-
ing N0 by N (and changing the dummy summation variable) and there
is the term (e) in which k1 = k2 = 0 k1 = ;k2 6= 0 which contributes
0 0
1 X v(jk j)N a^ a^
1 0 k1 k1
0 y y
2# k1 ;
which is the third term on the right hand side of (17.35), after replacing
N0 by N .
17.3 Please note that on page 206 `v(l)' should be replaced throughout by
`v(l)', where l is the modulus of l.
For the LHS of (17.42), we need to evaluate the commutator
X X
Ek a^k a^k + 12 N# v(k)(^ak ^a k + ^aka^ k )
0 y 0 y y
k k ; ;
where we have omitted the constant term in (17.35) since this obviously
commutes with all operators. Consider rst the Ek terms. We have
^ak ^ak a^l ] = a^k kl
y y y
and
^ak^ak ^a l] = ;a^k k l
y
; ;
;
For the terms multiplying N=# we use
^aka^ k a^l ] = 0
y y
;
y
and
^ak ^a k ^a l] = ;a^kkl ; a^ k k l
y y
; ;
y y
; ;
together with
^aka^ k a^ l ] = 0
; ;
and
^aka^ k a^l] = a^kk l + a^ k kl
;
y
; ;
17.4 (a) Let A^ = 12 (^a2 ; a^ 2) then U^ = expA^]. Now note that A^ = ;A^
y y
(b) We have
dI^ = dU^ a^U^ 1 + U^ ^a dU^ 1 : ;
;
d d d
For dU^=d we note that
U^ = expA^] = 1 + A^ + 12 2A^2 + : : :
and so
dU^ = A^ + A^2 + 2 A^3 + : : :
d 2
= A^ expA^] = expA^] A:^
Similarly,
dU^ 1 = ;A^ + A^2 ; 2 A^3 + : : :
;
d 2
= ;A exp;A] = ;exp;A^] A:
^ ^ ^
Hence
dI^ = expA^]f 1 (^a2 ; ^a 2) ^a]gexp;A^]:
y
d 2
The ^a ^a] part of the commutator is zero, and the other part is
2
y
^a a^ 2] = 2^a :
y
Hence
dI^ = expA^] 1 2^a exp;A^] = J^ :
y
d 2
Similarly,
dJ^ = expA^]f 1 (^a2 ; ^a 2) ^a ]gexp;A^]:
y y
d 2
This time the commutator gives 2^a, so that
dJ^ = I^
d
and the second result now follows.
(c) The general solution of the second order dierential equation for I^
is
I^ = cosh X^ + sinh Y^
where X^ and Y^ are operators, to be determined by the initial conditions
for I^ (i.e. the conditions at = 0). We have
I^=0 = U^=0 a^ U^=01
;
and
U^=0 = 1 = U^=01 :
;
I^.
17.5 We have, for the kinetic terms,
^
@^ = p1 (@h^ )exp;i^=v] + p1 (v + ^h) ;iv@ exp;i^=v]
2 2
and
^
@^ = p1 (@^h)expi^=v] + p1 (v + h^ ) i@v expi^=v]
y
2 2
and so the terms which are quadratic in the elds are
@^ @ ^ = 21 @^h@ h^ + 12 @^@ ^:
y
But from (17.80) we have v = 2= 12 . Substituting this into the pre-
ceding expression, we see that the term linear in ^h cancels, and the
constant and quadratic terms are as given in (17.85).
17.6 Please note that the exponent in (17.107) should be
exp;i =(2v)]:
The kinetic terms are derived as in the preceding problem, but because
the exponent contains ^ =2 rather than ^ the ^ term has a factor 1/8
rather than 1/2 the H^ term is the same as the ^h term in (17.85). The
potential terms are the same as in (17.85), with h^ ! H^ , leading to
(17.108).
17.7 We begin by checking that, for instance,
f
^k
^k0 g = 0:
Substituting from the rst equality in (17.111) (replacing ck by c^k ),
the LHS is
fuk c^k ; vk c^ k uk0 c^k0 ; vk0 c^ k0 g:
y
;
y
The anticommutator of two c^'s or two c^ 's vanishes, and we are left
y
with
;uk vk0 fc^k c^ k0 g ; vk uk0 fc^ k c^k0 g:
y
;
y
dealing with f
^k
^ kg, which can be evaluated as
;
; ;
y
; ;
Similarly, f
^k
^k0 g = 0. On the other hand,
y y
f
^k
^k0 g = fuk c^k ; vk c^k uk0 c^k0 ; vk0 c^k0 g
y y y
=
^kfkl ;
^l
^kg ;
^l
^k
^k
y y y y
=
^kkl ; (
^k
^l +
^l
^l )
^k
y y y y y
=
^kkl y
kX
= k (^ck c^k c^l ; c^l c^k c^k )
y y y y
k
X
= k (^ck kl ; fc^k c^l + c^l c^k gc^k )
y y y y y
k
= lc^l : y
Similarly, we nd X
k c^k c^k ^c l ] = ;lc^ l :
y
k
; ;
kkX
;
;
k k 0
; ;
; ;
= c^ i : ;
One might worry about another term, in which k = ;i, but of course
if both of these are included then the total is just c^ ic^i + c^ic^ i which
vanishes so only the one contribution is required. In general, then, we
; ;
nd
X X X
^ck0 c^ k0 c^ k c^k c^l ] = c^k0 c^ k0 c^ k kl = ( c^k0 c^ k0 )^c l
y y y y y y y
kk0 k0k k0
; ; ;
; ; ;
of cos l c^l and the second term in (17.109) is thus (relabelling the
y
summation) X
;V cos l c^k c^ k c^ l : y y
k ; ;
X
+V sin l c^ k c^k c^l : y
k
;
k k
; ; ;
Chapter 18
18.1 (18.24) is the same as (12.103) (compare (12.109) and (18.23)), the
derivation of which was set as problem 12.11. We shall verify (18.26),
using the same method as given in the solution to problem 12.11.
We have Z
Qi5 = q^a(x)(5) (i=2)abq^b(x) d3x
^ y
and Z
Qj5 = q^
c(y)(5)
(j =2)cd q^d(y) d3y:
^ y
= ;q^
c(y)^qa(x)^qd(y)^qb(x) + q^a(x)^qd(y)
bc3(x ; y):
y y y
The rst term here cancels with the rst term of the preceding expres-
sion, so that our commutator becomes
1 ( ) ( ) ( ) ( ) Z Z d3x d3y ^q (x)^q (y) ;q^ (y)^q (x) ] 3(x;y)
4 i ab j cd 5 5
d
bc
c b ad
y y
a
4 a c
In the second term, the indices c and b are both summed over, and can
be replaced by a and d, respectively, as in the rst term, and similarly
can be replaced by . The expression
q^a(x)(ij )adq^d(x)
y
is just
q^ (x)(ij )^q(x)
y
with the explicit isospin and (Dirac) spinor indices inserted, and simi-
larly for the j i term, so the result is
Z Z
d3x q^ (x)i=2 j =2]^q(x) = iijk d3x q^ (x)(k =2)^q(x)
y y
using (12.28), which is precisely the right hand side of (18.26), as re-
quired.
18.2 For (18.28) we have
Q^ iR Q^ jR] = 14 Q^ i + Q^ 15 Q^ j + Q^ j5]
= 14 fQ^ i Q^ j ] + Q^ i Q^ j5] + Q^ i5 Q^ j ] + Q^ i5 Q^ j5]g
= 41 fiijk Q^ k + iijk Q^ k5 ; ijik Q^ k5 + iijk Q^ k g
= iijk 21 (Q^ k + Q^ k5) = iijk Q^ kR
using (18.24)-(18.26) and the antisymmetry of the symbol. Similar
steps lead to (18.29) and (18.30).
18.3 The Lagrangian of (18.12) is
q^iD
6 ^ q^
in the doublet notation of (18.18). We write q^ in terms of q^R and q^L
via 1 + 5 1 ; 5
q^ = 2 q^ + 2 q^ q^R + q^L:
On the other hand,
" ! # 1 ; 5 1 + 5
q^L = (1 ; 5 ) y
q q q
^ 0 = ^ 0 = ^ 0
2
y y
2 2
and similarly for q^R. Now consider
q^L q^R = q^ 0 1 + 5 1 + 5 q^ = q^ 0 1 ; 5 1 + 5 q^ = 0
y y
2 2 2 2
and similarly for q^R q^L. So the only surviving terms in q^iD 6 ^ q^ are as in
(18.39).
18.4 The invariant amplitude is
M = ;GF cos Cu (p2) (1 ; 5)v(p1)p f
= ;GF cos Cf u (p2)(6 p1 + 6 p2)(1 ; 5)v(p1)
= ;GF cos Cf u (p2)(1 + 5)6 p1v(p1)
= GF cos Cf mu (p2)(1 + 5)v(p1)
neglecting the neutrino mass. Using the formula for the 1 ! 2 decay
rate given in appendix L, and expression (K.3) for the two-body phase
space given in appendix K, we obtain for the total rate
XZ 1 2 d p1
3 d3p2 (2)44(p ; p ; p ):
;= jMj 1 2
spins 2m (2)32E1 (2)32E2
The sum over spins gives
Tr(1+5 )v(p1)v(p1 )(1;5)u(p2)u(p2)] = Tr(1+5)(6 p1;m)(1;5 )6 p2]:
The term in m vanishes since (1 + 5)(1 ; 5) = 0, while the term in
6 p1 gives
Tr(1 + 5)6 p1(1 ; 5)6 p2] = Tr6 p1(1 ; 5)26 p2]
= 2Tr6 p16 p2] ; 2Tr6 p156 p2]
= 8p1 p2
using equations (J.30) and (J.36). The spatial part of the four-dimensional
delta function allows the integral over p2 (say) to be done. Since the
matrix element and the phase space factors are Lorentz invariant, we
can evaluate the integral in any frame, and it is obviously convenient
to choose the rest frame of the pion, in which the three-momenta of the
neutrino and the muon are equal and opposite. The remaining delta
function is then
(m ; jpj ; (p2 + m2)1=2)
where p is the three-momentum of the muon, while d3p1 reduces to
4jpj2djpj also, we have E1 = (m2 + p2)1=2 and E2 = jpj. To perform
the integral over jpj, we write
W = jpj + (m2 + p2)1=2:
Then
dW = 1 + jpj = W:
djpj (m + p )
2 2 1 = 2 E1
Hence
djpj = EW1 dW
and the delta function is (m ; W ). Putting all these results together
gives
; = m 1 G2 f 2 m2 cos2 p p jpj:
2 F C 1 2
For p1 p2 we have
p2 = m2 = (p1 + p2)2 = p21 + p22 + 2p1 p2 = m2 + 2p1 p2
whence
p1 p2 = (m2 ; m2)=2:
To determine jpj we use the energy-conservation relation
m = jpj + (m2 + p2)1=2:
Taking jpj to the LHS and squaring, we nd
2 ; m2
jpj = m2m :
The result (18.53) then follows.
18.5 We give two ways of approaching this problem.
First, we can use the equations of motion for q^ and q^ to calculate
@^ji5(x) directly, as follows:
@^ji5(x) = @(q^(x) 5 2i q^(x)) = (@q^(x)) 5 2i q^(x);q^(x)5 2i (@q^(x)):
The equation of motion for q^ is iD 6 ^ q^ = mq^, which can be written as
@q^(x) = ;igs =2 A q^ ; imq^:
For @q^(x), we take the dagger of the last equation, multiply from the
right by 0, and use 0 0 = (see (8.59)) to obtain
y
=
^i6 @
^
showing that this ( rst) term in (18.72) is invariant as usual, only
terms of rst order in the in nitesimal parameter have been retained,
and we have used f5 0g = f5 g = 0. It is also clear - since the
gradients don't act on the 's - that if we prove that ^ 2 +^ 2 is invariant,
it will follow that 12 @^ @ ^ + 12 @^ @ ^ is also invariant.
(ii) ^ 2 + ^ 2 transforms to
(^ ; ^ )2 + (^ + ^ )2
= ^ 2 ; 2 ^ ^ + ^ 2 + 2 ^ ^ = ^ 2 + ^ 2
showing that it is invariant we have used the fact that the ^ and ^
elds commute. ^ 5i
^^i,
(iii) We are left with the terms multiplying gNN . The rst, i
transforms to
i
^ (1 + i =2) 0 5i(1 ; i =25 )
^(^i + i^ )
y
= i
^ 5
^ ^ + i
^ 5
^^
;
^ ( =2)5 05i
^^i +
^ 05i( =2)5
^^i: (A)
y y
Using 52 = 1 and 5 05 = ; 0, the last two terms here combine to
give
^ i ( =2)
^^i =
^ 1 (j i + ij )
^( =2)i
^^i +
^ j ^i
2
=
^
^ ^ (B)
where we have used i j + j i = 2ij (see problem 4.4(a) for the same
relation written for the matrices). The second term multiplying gNN
is
^
^^ which transforms to
^ (1 + i =25 ) 0(1 ; i =25 )
^(^ ; ^ )
y
=
^
^^ ;
^
^ ^ ; i
^ 5
^^ :
The rst term here combines with i
^ 5
^ ^ of (A) to give the original
interaction term involving gNN , while the other two terms cancel with
the second term in (A) and with that in (B). Hence this interaction term
is also invariant. The other interactions are included in the invariant
(^2 + ^ 2)2.
18.7 Note that in this problem the parameter 2 in (18.72) is taken to be
negative, so that the potential has the form of (18.83) with the `new'
2 positive. Also note that in the second line of equation (18.88), 2
should be replaced by ;2, with 2 > 0.
The terms involving gNN , and the gradients, are straightforward. The
remaining terms (from ;V ( )) are (we work with the classical po-
tential for convenience - hats can be added afterwards)
1 2(v2 + 2v + 2 + 2) ; 1 (v2 + 2v + 2 + 2)2
0 0 0 0
2 4
= 21 2v2 + 2v + 12 2 ( 2 + 2)
0 0
;v2 2 = ;2 2
0 0
where k1 k2 are the 4-momenta of the 's, 1 2 are their polarizations,
and
are Lorentz indices. The rate is
1 XZ 1 d 3 k1 d 3 k2
; = 2m jMj 2 2E (2)3 2E (2)3 (2)44(p ; k1 ; k2)
2
1 2 1 2
where E1 E2 are the energies of the 's, p is the 4-momentum of the
pion, and the extra factor of 1/2 arises from the identity of the 's (see
page 389 of volume 1, just after the formula for the 1 ! 2 decay rate).
When we form jMj2, we get the rather complicated-looking expression
1 (k1 1)2(k2 2)1 (k1 1)2 (k2 2)k1k2 k1
k2 :
The sums over the polarization labels are done using (see section 8.6.3)
X
1 (k1 1)1 (k1 1) = ;g
1
and similarly for the sum on 2, so that the previous expression reduces
to
g g k1k2 k1
k2 : (C)
We now need a formula for the product of two four-dimensional sym-
bols, when both their rst two indices are `contracted' via the g sym-
bols. The required formula is
g g = ;2(g
g ; gg
)
which may be checked by examining special cases (note that 0123 =
+1) see also problem 20.7. Inserting this result in (C) gives
;2((k1)2(k2)2 ; (k1 k2)2) = 2(k1 k2)2
where we have used the masslessness conditions k12 = k22 = 0. Squaring
the 4-momentum conservation condition gives
p2 = m2 = (k1 + k2)2 = 2k1 k2
and hence k1 k2 = m2 =2. The phase space integrals are most con-
veniently done in the rest frame of the pion (as in problem 18.4), the
energy part of the delta function then becoming just (m ;2jkj), where
k is the 3-momentum of one of the photons in this frame. Putting the
details together leads to (18.101).
Taking = 1=137, f = 93 MeV, and m = 135 MeV, we nd
; = 7:64 10 6 MeV. The lifetime is given by h =;, and using h =
;
2
Since the 4-dimensional symbol is antisymmetric under the inter-
change of any pair of indices, we can replace
(@A^ (x) ; @ A^(x))(@ A^ (x) ; @ A^ (x))
by
4 (@A^ (x))(@ A^ (x)):
To evaluate the matrix element, we need
^ XZ d3k
A (x) = (2 ) (k) (k )e ik x + ^(k) (k )eik x]
3(2Ek )1=2 ^
; y
where Ek = jkj (see (7.101)) A^ (x) has a similar expansion but with
an integration variable k and a polarization label . Also,
0 0
2
XX Z Z dk
3 d3k i(k+k0 ) x
0
0
f1 02 (2)63(k ; k1)3(k ; k2) + 2 01 (2)63(k ; k2)3(k ; k1g:
0 0
Consider the rst term in the f::::g bracket. The k and k integrals 0
Z e 2
4
d xe i(k 1 +k2 q ) x
4 (k ) (k )k k
1 1 1 2 2 2 1 2
;
32 2
This is clearly the same as the previous term but with the labels `1'
and `2' exchanged, as would be expected. The two terms are in fact
identical, as must follow from Bose symmetry. We can check this by
re-labelling, in the second term, the index as , and v.v., and the
index as
, and v.v. The second term is then the same as the rst,
apart from the symbol which has now become but this equals
, since they are related by an even number of index interchanges.
Hence nally the factor e2=82 becomes e2=42, and equating the result
to the RHS of the rst equation of this solution leads to (18.103).
Chapter 19
19.1 We have
!
(;k2 + M 2)g + k k ] ;g + k k =M 2
k2 ; M 2
= g g ; g k k =M 2 ; kk2 k; M
g k k2k =M 2
2 + k2 ; M 2
= g ; kMk2 ; k2k;kM k k k =M
2 2
2 + k2 ; M 2 :
The last two terms here combine to give k k =M 2 , and the result fol-
lows.
19.2 The required Lorentz transformation has the form (for the 4-vector
(x0 x1 x2 x3))
x0 = (x0 +
x3)
0
x3 = (x3 +
x0)
0
where
= jkj=E and = E=M . Under this transformation, the
components of the 4-vector krest = (M 0) transform to
k0 = M = E
and
k3 =
M = jkj
as required. The components of (krest = 0) = (0 0 0 1) transform
to
0(k = 0) =
= jkj=M
and
3(k = 0) = = E=M
while the transverse components in the x1 and x2 directions are unaf-
fected by this transformation along the x3-axis.
19.3 We proceed by checking dierent values for the indices and explic-
itly. For example, consider = = 0. In this case,
X 0
(k )0 (k ) = 0(k = 0)0 (k = 0)
=0 1
=0 1
2 2
and so
@ ^ = p1 (v + h^ ) (;i@ ^=v) e i^=v + p1 @ h^ e i^=v
; ;
2 2
1 1
@ ^ = p (v + ^h) (i@ ^=v) ei=v + p @ h^ ei^=v :
y ^
2 2
Hence
= iq (^ @ ^ ; (@ ^ )^) ; 2q 2A^ ^ ^
^jem y y y
= i2q (v + h^ )f;i(v + ^h)@ ^=v + @ ^hg ; fi(v + ^h)@ ^=v + @ ^hg(v + ^h)]
;q2A^ (v + ^h)2:
The terms in @ ^h cancel, and the terms in A^ and @ ^ (omitting prod-
ucts of elds) combine to give (19.46).
19.5 With the ansatz (19.55) the Lagrangian of (19.40) is
L^H = 12 @ h^ +iqA^(v+^h)] @h^ +iqA^(v+^h)]; 41 F^ F^ ; 161 (v+^h)4+ 21 2(v+^h)2:
y
8 1 2 8 3
as required.
Chapter 20
20.1 We have
up un = up(1 )un
y
E +m p
for jpj
m, and similarly for un. Then the `1' contribution in up un
becomes just
y
pn
which vanishes if the spin states of p and n are dierent, while it is
easy to check that the `' part is zero, in this approximation.
20.2 Please note that there are two misprints in this question: a `5' has been
omitted from both bilinears. We consider the behaviour of
^1 (x t) 5
^2(x t)
under P^ . The = 0 component is
^15
^2, which is a pseudoscalar as
y
showing that they form the components of an axial vector (see (20.13)).
20.3 We require N such that
! ! !
( E ; j p j) 0 1
(E p )
2m = N m2 y y
1 0 ;j j
! (E p ) ! m
= N 2 (E ;mjpj) m
;j j
y y
= 2N 2 (E ;mjpj)
taking to be normalized to = 1 as usual. Hence
y
N2 = E m
2
; jpj = E + jpj
as required.
20.4 The spinor v satis es (6 p + m)v = 0, which is (in the representation
used in section 20.4)
! !
m E+p = 0
E;p m
which gives
= ; (E +m p) :
Hence (noting what is said above (20.54)) we have
(E p ) !
v(p = +1) = N ; ;j
m
j
;
;
which leads to
N2 = E m
2
; jpj = E + jpj:
v(p = +1) is then
q (E p ) !
E + jpj ; m
;j j
v(p = +1) = ;
0 q 1 ;
= @ q
; E ; jpj A;
E + jpj ;
i2
so that under C
! ;i2
! i2
and under CP
! CP = i2
! CP = ;i2
Now note that 1 = 1 2 = ;2 3 = 3, while 21 = ;12, and
23 = ;32. These relations imply that the previous equation can
be written as
E (i20) = ; p(i20)
which is just
E (0)CP = ; p(0)CP:
as required. The CP-invariance of the second equation in (20.56) can
be checked similarly.
20.6 (a) We have
52 = (i)2 0 1 2 3 0 1 2 3
= +( 0)2 1 2 3 1 2 3
= ( 1)2 2 3 2 3
= ; 2 3 2 3 = ( 2)2( 3)2 = 1:
Then
(1 + 5)(1 ; 5) = 1 ; (5)2 = 0:
(b) Consider f5 0g:
f5 0g = i 0 1 2 3 0 + ( 0)2 1 2 3]
= i;( 0)2 1 2 3 + ( 0)2 1 2 3]
= 0 (3)
and similarly for the other components of . The other results follow
easily.
20.7 (a) There are only four non-zero possibilities:
1212 1221 2112 and 2121
of which the rst and fourth have the value +1, and the other two have
the value -1. Thus when i = k and j = l we must get +1, while when
i = l and j = k we must get -1, and for all other possibilities (eg i = j )
we must get zero. It can be easily checked that these requirements are
met by the expression ik jl ; iljk .
Setting k = i and summing over i (using the summation convention)
gives
(iijl ; ilji) = 2jl ; jl = jl:
Setting j = l and summing over j gives jj = 2.
(b) Multiplying out the determinant gives
ijk lmn = il(jmkn ; jnkm) ; im (jlkn ; jn kl)+ in (jlkm ; jm kl):
Setting i = l and summing over i gives
ijk imn = 3(jm kn ; jnkm ) ; (jmkn ; jnkm) + (jnkm ; jmkn )
= (jm kn ; jnkm)
= jm jn
km kn
as required. Setting now j = m and summing over j gives
ijk ijn = jj kn ; jnkj = 2kn
and nally setting k = n and summing over k gives
ijk ijk = 2kk = 6:
(c) By analogy with the 2- and 3-dimensional cases, we expect that the
product of two 4-dimensional 's will have the form
g g g g
= S gg gg gg gg
g g g g
where S is +1 or -1, to be determined by looking at a special case.
From covariance, we know that the only tensors available to construct
the RHS are the g's, and we also know that the expression must be
antisymmetric under interchange of any pair of the indices
and
of the rst symbol, and also under interchange of any pair of the
indices and of the second symbol: the determinant automat-
ically builds in these antisymmetry requirements (as in the previous
cases in (a) and (b)).] The special case = = 0
= = 1 = =
2 = = 3 yields
1 0 0 0
0 ; 1
01230123 = 1 = S 0 0 ;1 0 = ;S
0 0
0 0 0 ;1
and hence S = ;1:
We are now interested in successive `contractions' of the indices between
the two symbols. The rst is
=
g:
Using the formula for the product of two 's (with S = ;1) we obtain
g g g g
= ; gg gg gg gg :
g g g g
Note, by the way, that we could just as well write for g, etc, as we
have in fact done in (20.107). Consider the expansion of this determi-
nant about the rst row. The rst term is
g g g
;g g
g
g
:
g g g
The value of g is 4 (remember, it is summed over = 0 1 2 3, and
g00 = g00 = 1 g11 = ;1 = ;g11 etc.) The second term in the expansion
of the 4 4 determinant is
g b g g g g
+g g
b
g
= g
g
g
:
g b g g g g
The third and fourth terms can be similarly worked out, and are found
to be equal to the second term, so that the result for this `once-
contracted' product is
g g g
= ; g
g
g
:
g g g
Proceeding in the same way, we nd the twice-contracted result (com-
pare (20.107))
g
= ;2 g g
g
and from this the thrice-contracted result
= ;6g
and nally
= ;24:
20.8 The tensor N is de ned in (20.97). Using (20.100), (20.101) and
(20.102) we obtain
N = 8k k + k k + (q2=2)g ] ; 8i
kg k
g :
0 0 0
= 64kk p
q (g
g ; g g
) = 64(k p)(k q) ; (k q)(k p)]
0 0 0
spins
+ 4mem (me + m )2gA2 ; (me ; m )2gV2 ]:
It is interesting to note that the signs of gV and gA are irrelevant here.
For the phase-space integration we use (m ; W ) where
W = (p2 + m2e)1=2 + (p2 + m2 )1=2 E1 + E2
so that
dW = jpjW :
djpj E E 1 2
The phase-space integral can then be done via
jpjdjpj(m ; W ) = EW 1 E2
dW(m ; W ):
Finally, we have
m ; (p2 + m2 )1=2 = (p2 + m2e)1=2
and by squaring this relation one nds
jpj = ((m + m )2 ; m2e)((m ; m )2 ; m2e)=(4m2 )]1=2:
The ratio of the rates can then be evaluated. For m
me, the sum
over spins of the squared amplitude is approximately proportional to
m2e, giving a strong suppression of the positron mode relative to the +
mode, as before, independently of the numerical values of gV and gA .
20.10 (a) From (20.158) we have
d() = G2F 4k1 p 4M N W() 2kd(2k)3
2 3 0
4kk cos2 , where we are (rather loosely) using k and k here for the
0 0
term is therefore
32kk sin2(=2) W1() :
0
M2 2
The ` term is
; 2M8 2
kk p
q W3(): 0
M2 3 M 3
2k 0
To convert from dk d cos to dQ2d , we use
0
which implies
dQ2d = 2kk dk d cos :
0 0
dxdy 2 F 2 1
+ (k M
+ k ) yk sin2(=2)W ()g:
0
0
3
k sin2(=2) = Q2 = Q2 = xy
0
M 4Mk 2M 2k 2
(as stated in the problem). In addition, y(k + k ) = (k + k ; ky)=k =
0
2 (1 ; y=2): Also,
k y cos2(=2) = k k cos2(=2) = k k (1 ; sin2(=2)) = kk ; k Mxy
0
0 0 0
2
Mxy
= (1 ; y) ; k 2 :
Neglecting the term in M here (which amounts to taking 0 in
the high-energy limit, in the laboratory system), and using s = 2Mk,
together with the Bjorken scaling behaviour, leads to the stated ex-
pression for d2()=dxdy.
(c) Using the C-G relation we obtain
d2() = G2F sF ()(1 ; y) + 1 F ()y2 + F ()(1 ; y=2)yx]
dxdy 2 2 2 2 3
( )
= G2F sF2()(1 ; y) + y2=2 + xF(3) y(2 2; y) ]
2
0 F2 1
G 2
= 2F sF2 @
( ) 1 + (1 ; y ) 2 xF ( ) 1 ; (1 ; y )2
+ (3) A:
2 F2 2
20.11 Consider the process
(k) + u(p) ! + d(p )
; 0
for which the matrix element is (by analogy with (20.93), but using a
slightly dierent labelling of the spinors)
;iGFu(k ) (1 ; 5)u (k)vu(p) (1 ; 5)vd(p ):
0 0
present case is that, in the second Trace in equation (20.95), the order
of p and p is reversed (see (R)) note that since we are neglecting
0
masses, uu = vv: This means that, while the ` ; ' tensor N stays
;
the same, the antiquark tensor E is the same as (20.109) but with p
and p interchanged. Since their product multiplies the symbol, which
0
and hence
u2 = (1 ; y)2
s2
and
d ( q) = G2F (1 ; y)2
dt
as required. Note that in the CM frame, u = ;2p2(1+cos ), where p is
the CM 3-momentum and is the CM scattering angle it follows that
the amplitude vanishes in the backward direction = , as discussed
below gure 20.5.
(b) We have
d ( q) = G2F :
dQ2
Following steps analogous to those in section 9.2, we can write this as
d2 ( q) = G2F ( ; Q2=2Mx)
dQ2d
G 2x
= F (x ; Q2=2M )
using property (E.29) of the function. Converting to the variables
x and y via the Jacobian known from problem 20.10(b), and using
s = 2Mk, leads to
d2 ( q) = G2F sx(x ; Q2=2M )
dQ2d
as required. The q cross section follows similarly.
(c) From problem 20.10(c) we have
d2() = G2F sF ()(1 ; y + y2=2) + xF ()(y ; y2=2)]
dxdy 2 2 3
Chapter 21
21.1 (a) For p = jpj(sin 0 cos ), we have
!
cos sin
p^ = sin ; cos :
Setting !
+ = b a
we have ! ! !
cos sin a = a
sin ; cos b b
which leads to
a(1 ; cos ) = b sin and a sin = b(1 + cos ):
Changing to half-angles, we see that both these relations become simply
a sin =2 = b cos =2
so that a suitable (normalized) solution is
!
cos = 2
+ = sin =2 :
Similarly (or simply using orthogonality) is;
!
; sin
= cos =2 :
;
= 2
011
B 0 CC
u( ) = E 1=2 B
;
B@ 0 CA
1
and so 001
B 0 CC
(1 ; 5)u( ) = B
;
B@ 0 CA :
2
For v() we take = and obtain
0 0 1
B 1 CC
v() = E 1=2 B
B@ ;1 CA :
0
It is then simple to check that the = 0 and = z or 3 components
of v( )(1 ; 5)u( ) both vanish. On the other hand, noting that
;
= (
; ) and hence
! !
0
x = ; 0 ; and
y = ; 0 ;
x y 0
x y
we nd that the = x component is ;2E and the = y component is
2iE .
We now consider the other half of the matrix element. For u(e ) we;
take (E), and for v(e) we take (E) with replaced by + , since this
is the opposite direction to that of the electron in the CM frame. This
gives 0 1
0
B 0 CC
(1 ; 5)v(e) = BB@ ;2 cos =2 CA
;2 sin =2
and
0 0 1
!B CC
x(1 ; 5)v(e) = 0x ;0 B 0
B@ ;2 cos =2 CA
x
;2 sin =2
0 0 1
B 0 CC
= BB@ 2 sin =2 CA :
2 cos =2
It follows that
u (e )
x(1 ; 5)v(e) = 2E (cos2 =2 ; sin2 =2) = 2E cos :
y ;
Similarly, we nd
u (e )
y (1 ; 5)v(e) = 2iE (cos2 =2 + sin2 =2) = 2iE:
y ;
g 1 1 ; 5 1 + 5
^ ^
= ;Z cos flf 2 (; cos W + sin W) 2 lf + sin Wlf
2 2 ^ 2 ^ ^lf g
W 1 ; 5 1 + 5 2
^ g
^ 1
= ;Z cos flf (; 2 + sin W) 2 + sin Wlf ^ ^
2 lf g:
2 2
W
The terms in sin2 W combine to give just sin2 W^lf ^lf , which can
alternatively be written as ; sin2 WQ^lf ^lf for this case, while the
`;1=2' can be written as t3. This contribution therefore reduces to
g 1 ; 5
;Z cos flf t3 2 ^lf ; sin2 W Q^lf ^lf g: (M)
^
^
W
For the neutral eld with 3 = +1, the part involving Z^ is
g g 1 ; 5
;Zlf 2 cos W + 2 sin W ] 2 ^lf
^ ^
0
g 1 1 ; 5
^ ^
= ;Z cos 2 lf ^lf : (N)
W 2
In this case we can replace the `1/2' by t3 (M) and (N) together then
yield (22.36).
22.2 Note that `(22.56)' should read `(22.58)']. The rate is
X g2 Z
;W = 2M1 31 j ( p u
) (p ) 1 ; 5 v (p )j2
1 2
W spins 2 2
d3p1 d3p2 (2)4(4)(p ; p ; p )
1 2
(2)32E1 (2)32E2
where is the polarization label of the W, which has 4-momentum p,
and p1 and p2 are the 4-momenta of the e and e. We neglect lepton
;
masses. The `1/3' arises from averaging over the W polarizations. The
total energy W is
W = 2jpj
where p is the CM momentum of the e , so that (in the CM) E1 =
;
E2 = jpj, and
dW = 2djpj:
The phase space term then reduces to
jpj2(dW=2)4 (W ; 2jpj) = (1=8):
424jpj2
The spin-summed squared matrix element is
1 X 1 Tr 6 p (1 + ) 6 p ]:
3 2 2 5 1
The average over polarizations gives (from (21.29) - note that `MW'
should read `MW2 ')
!
1 ;g + p p : (P)
3 M2 W
In the Trace, the 5 part will give (see (J.37)) something antisymmetric
in and , which will be contracted with the symmetric quantity (P),
so it won't contribute. The non-5 part of the Trace is (see (J.31))
Tr 6 p2 6 p1] = 4p2 p1 + p1 p2 ; g p1 p2]:
Contracting this with (P) gives
4 ;2p p + 4p p + 2 (p1 p)(p2 p) ; p2 p p ]:
1 2 1 2
3 Mw2 Mw2 1 2
From p = p1 +p2 we obtain ( neglecting lepton masses) p1 p2 = (p2)=2 =
(MW2 )=2, p p1 = p p2 = (m2W)=2. Putting the pieces together yields
(22.58).
22.3 We have
! ! ! ^ !
^ = 1 0 ^ ^C = i2^ = 0 1 0 1 (v + H
^) = 2 :
p
;1 0
2 (v + H )
1 (v + H
p p
2 0
So (22.125) becomes
1 (v + H ^ !
L^ = aij (u^Lid^Li ) 2 0 p ) u^Rj
!
0
+ bij (u^Lid^Li) 1 (v + H^ ) d^Rj + h:c:
2
p
1
= aij p u^Li(v + H^ )^uRj + bij p1 d^Li(v + H^ )d^Rj + h:c
2 2
which is easily seen to be the same as (22.127), with the identi cations
(22.128).
22.4 First, note that from (22.130) and (22.131) we have (repeated indices
are summed over)
u^Li = UL(u)i u^L
y
and so
y
u^Lj = u^L UL(u)j
y
and similarly for the other elds. Consider now the rst term on the
RHS of (22.133), involving the `L' elds. The neutral part of this is
(u) !
^ 3 ^
;(u^L ULj dL ULj ) (g 2 W + g 2 yB) U (d) d^
(u) (d) 0
1 ^
y
U L j u
^ L
y
Lj L
g (u)y
= ; 2 u^L 6 W^ (ULj ULj )^uL ; d^L 6 W^ (ULj ULj )d^L]
(u) (d)
y (d)
where a is the SU(2) index, and (0,i) are the 4-vector indices. Also, the
RHS of (22.150) and (22.151) should be be pre xed by a minus sign.
Making use of (22.148), we have
C^ P^ (^uL 0Vudd^L)(C^ P^ ) 1 = u^T 0 2 (;i) 0Vud (i) 2 0d^LT
y ; y y y
Now
2T = 2 0T = 0 = 0 and 2 = ; 2
When we form (D^ ^) (D^ ^) it's clear that the result has to be real, so
y
that we need only write down the real terms in the product. These are
1 (0 1)f@ ^ @ ^ + ( @ ^ )( @ ^ ) + (i)
2
; 2g @^ ^ W^ ; g2 ^ W^ @^ (ii)
+ 2g @^ ^ W
^ + g ^ W^ @ ^ (iii)
2
; 2 @ ^ (W ^ ) ; 2g (W^ ^ ) @ ^ (iv)
g ^
!
g 2
^ ^
+ 4 ( W ^ )( W ^ )g 1 : (vii) 0
g 2
= 8 W^ (W^ (^ ^ ) ; ^ (W^ ^ ))]
= g8 (W
2
^ W^ )(^ ^ ) ; (W^ ^ )(W^ ^ )]: (f)
The second term of (f) cancels with the second term in line (v), and
the rst term in (f) combines with (e) to give
g2 W^ W^ (^2 + ^ ^ ): (g)
8
The contributions (a), (b), (c), (d) and (g) then give the (corrected)
result for (22.183). Note that since all these terms do not involve the
matrices - that is, they multiply the unit matrix in internal space -
they can be taken outside the matrix product, which simply reduces to
!
0
(0 1) 1 = 1:
22.8 Taking the vertex from (22.136), and approximating Vtb by unity, the
rate is
1 1 X g2 1 ; 5 u(p )j2
;t = 2m 2 2 j (k )u ( p 2 )
2 1
t spins
p
d3 2 k
d3 (2)4(4)(p ; p ; k)
1 2
(2)32E2 (2)32Ek
where p1 is the 4-momentum of the top, p2 that of the bottom, and k
are the 4-momentum and polarization label of the W. The `1/2' before
the spin sum arises from the average over the top spins. The sum over
the fermion spins of the squared matrix element is
Tr 1 ;2 5 (6 p1 + mt) 1 +2 5 (6 p2 + mb)]
= Tr 1 ;2 5 f 1 ;2 5 6 p1 + 1 +2 5 mtg (6 p2 + mb)]
= Tr 1 ;2 5 6 p1 (6 p2 + mb)]
= 12 Tr 6 p1 6 p2] = 2p1 p2 + p2 p1 ; g p1 p2] (S)
where we have used (J.25), (J.31) and (J.36) the contribution from the
term with four -matrices and a 5 will (from (J.37)) be antisymmet-
ric in and , and will vanish when contracted with the symmetric
quantity
X
(k ) (k ) = (;g + k k =MW2 ): (W)
Contracting (S) with (W) gives
2(;g + k k =MW2 )(p1 p2 + p2 p1 ; g p1 p2)
= 2;2p1 p2 + 4p1 p2 + 2(p1 k)(p2 k)=MW2 ; k2p1 p2 =MW2 ]
= 2p1 p2 + 2(p1 k)(p2 k)=MW2 ]
We evaluate the dot products by using the 4-momentum conservation
equation,
p1 = p2 + k
thus
p1 k = (p21 + k2 ; p22)=2 = (m2t + MW2 ; m2b)=2
p1 p2 = (p21 + p22 ; k2)=2 = (m2t + m2b ; MW2 )=2
and
p2 k = (p21 ; p22 ; k2)=2 = (m2t ; m2b ; MW2 )=2:
q 2 2
In
q 2the CM frame, the total energy W is W = E 2 + E k = mb + p +
MW + p2. So
dW = jpj(E2 + Ek ) djpj:
EE 2 k
Also,
jpj = f(mt + MW )2 ; m2bgf(mt ; MW )2 ; m2bg]1=2=2mt :
Performing the phase space integral in the now familiar way (using
(mt ; W )), we nd
;t = MW2GpF (m2t+m2b;MW2 )+(m2t+MW2 ;m2b)(m2t;MW2 ;m2b)=MW2 ]jpj
2
4mt 2
where we have used (22.29). From this we obtain the numerical values
;W = 1:65 GeV, and W 4 10 25 s. ;
2MH 12
d k1
3 d3k2 (2)4(4)(p ; k ; k )
1 2
(2)32E1 (2)32E2
where p is the 4-momentum of the H. The polarization sum is
(;g + k1k1 =MW2 )(;g + k2k2 =MW2 )
= (4 ; k22=MW2 ; k12=MW2 + (k1 k2)2=MW4 )
= (4 ; 2 + (m2H ; 2MW2 )2=(4MW4 ))
= 4m
4
M 4 (1 ; 4MW =mH + 12MW =mH):
H 2 2 4 4
W
In the CM frame E1 = E2 = (MW2 + p2)1=2, and the total energy is
W = 2(MW2 + p2)1=2, so that
djpj = E1dW=(2jpj)
and
jpj = (m2H ; 4MW2 )1=2=2:
Performing the phase space integral in the usual way, these results lead
to (22.203). Note that there is a misprint in (22.203): the rst factor
in round brackets should read
4M 2 !1=2
1 ; m2 W :
H