µ

Exercise 4.10 Show that under a Lorentz transformation of the coordinates : xµ ! x0 = ⇤µ⌫ x⌫ we have:
0 0
(i) = ,
0 5 0
(ii) = det(⇤) 5 ,
0 µ 0
(iii) = (⇤)µ⌫ ⌫ ,
0 µ 5 0
(iv) = det(⇤)(⇤)µ⌫ ⌫ 5 , where
5
<latexit sha1_base64="EtJ9BlK1L87bKPpUs10TvSL9jhE=">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</latexit>

and the
= i Det(⇤)
, relationship
⌘ . 5 0 1 2 3 † 0

Solution: We denote a spinor transformation matrix by S, i.e. under x ! ⇤x a spinor transforms as !S .

(i) Using the fact that 0 S † 0 = S 1 we get
¯0 0
= †
S† 0
S = ¯ 0
S† 0
S = ¯S 1
S = ¯ .
5 0 1 2 3
(ii) Note that =i can also be written as
i 5
"µ⌫⇢ µ ⌫ ⇢ , (4.1) =
n!
where in four space-time dimensions n = 4, since it is anti-symmetric under any odd permutation of the -matrices.
Here, the symbol "µ⌫⇢ denotes simply the standard antisymmetric non-covariant Levi-Civita symbol, "0123 = "1023 =
· · · = +1, which vanishes if any two indices are the same. The symbol with upper indices is the same symbol. Hence
the product
µ̃ ⌫
"µ⌫⇢ "µ̃˜⌫ ⇢˜
˜
= +4! ˜ ⇢˜ ˜
[µ ⌫ ⇢ ] , (4.2)
where the notation [. . . ] denotes complete antisymmetrization of the respective indices, that is
j1 j1
i1 ... ip
1 j1 ...jp j1 ...jp
S[i1 ...ip ] ⌘ i1 ...ip Sj1 ...jp , i1 ...ip ⌘ ... ,
p! jp jp
i1 . . . ip

with . . . denoting the appropriate p ⇥ p determinant (p = 4 in our case).

Multiplying both sides of (4.1) by "↵ and using (4.2), we immediately obtain:
[↵ ]
= +"↵ 0 1 2 3
(4.3)
1 1 µ
In the following lines we insert factors of S S = 1 to make use of the identity S S = ⇤µ ⌫ ⌫
:
¯0 5 0
= †
S S† 0 5

= ¯ 0 † 0 5
S S
¯
= S 1 5
S
i
= "µ⌫⇢ ¯(S 1 µ S)(S 1 ⌫ S)(S 1 ⇢ S)(S 1
S)
n!
i
= "µ⌫⇢ ⇤µ µ̃ ⇤⌫ ⌫˜ ⇤⇢ ⇢˜⇤ ˜ ¯ µ̃ ⌫˜ ⇢˜ ˜ (4.4)
n!
i
= "µ⌫⇢ "µ̃˜⌫ ⇢˜
˜
⇤µ µ̃ ⇤⌫ ⌫˜ ⇤⇢ ⇢˜⇤ ˜ ¯ 0 1 2 3 (4.5)
n!
1
= "µ⌫⇢ "µ̃˜⌫ ⇢˜
˜
⇤µ µ̃ ⇤⌫ ⌫˜ ⇤⇢ ⇢˜⇤ ˜ ¯ 5
n!
= det(⇤) ¯ 5 .

In going from (4.4) to (4.5) we used the fact that the expression is anti-symmetric under any exchange of indices
1
µ $ ⌫ etc. and therefore also under any exchange of the indices with tilde µ̃ $ ⌫˜, and thus we used (4.3). The last
line follows from the definition of the determinant of an N -dimensional matrix,
1 a1 ...aN
det(A) = " "b1 ...bN Aa1 b1 . . . AaN bN , (4.6)
n!
where the Latin indices here are viewed in Euclidean space 1 .

1 µ
(iii) Using the identity S S = ⇤µ ⌫ ⌫
we have
¯0 µ 0
= †
S† 0 µ
S = ¯ 0
S† 0 µ
S = ¯S 1 µ
S = ⇤µ ⌫ ¯ ⌫
.
1
(iv) Using the previous results and repeated insertions of SS = 1 we get
¯0 µ 5 0
= ¯S 1 µ 5
S = ¯S 1 µ
SS 1 5
S = det(⇤)⇤µ ⌫ ¯ µ 5
.

Exercise 4.11 Show that the Lagrangian (4.53) is invariant under Lorentz transformations.

Solution: Since the mass is a Lorentz invariant. the second term in the Dirac Lagrangian L = ¯(i · @ m)
is invariant by the transformation laws found in 4.10. Remembering that under an active Lorentz transformation
x ! x0 = ⇤x a spinor field transforms as (x) ! 0 (x) = S (⇤ 1 ) we have

@µ ( (x)) ! @µ ( 0 (x)) = @µ (S (⇤ 1
)) = S(⇤ 1
)µ (@ )(⇤ 1
x).