Escolar Documentos
Profissional Documentos
Cultura Documentos
H(2,C), cH» pla") = 2,0" _ [a =03 atic? (1.111) © \at ia? 2% +23 I1.3 SL(2,C), Dotted and Undotted Indices 57 where the four-vector is (2") 2°,x), (zz) = (2°,-x). Proposition 1,22: The map inverse to Eq. (1.111) pot: H(2,€) —> My is given by the following trace relation pt sHQ,€) — Ma, X esp \X)act= atk [xo]. (1.112) Proof: Using Eqs. (1.111) and (1.109a) we obtain: har [x04] = Lay [e,0"@) 7 it [o’a] 2, = 5 + Inlay = oh, Proposition 1.23: The determinant of p(x) € H(2,C) is given by: det X = det p(y) = a,c" = 2. (1.113) Proof: Using the definition of the map p, i.e. Eq. (1.111), we have: wag gly ‘S) al ig? 2423 det X = det p(z,) = det ( = (2°)? = (2)? = nwt” = 2°, The Hermiticity of the matrix X = p(z“) is seen immediately from the Hermiticity of the Pauli matrices X=xt=x*T, (1.114) We now consider the action of SL(2,C) on H(2, C), which is called the adjoint representation of the group SL(2,C). This representation is defined in the following way: ad: SL(2,C) —+ Aut(H(2,C)), M a: (1.115)58 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors with M' = adM(X) := MXMt, M,Mte SL(2,C), (1.116) where Aut(H(2,C)) is the automorphism group of H(2,C), which is isomor- phic to GL(H(2,€),C) [19]. We have to verify therefore that adM(X) € H(2,C), ie. we have to show that adM(X) is Hermitian. Using Eq. (1.116) we have [adM(X)]' = (MxMt)' = MxtMt = MXMt =adM(X), since X € H(2,C). From the adjoint representation (1.116) we can derive the index structure of the Pauli matrices, as indicated in Eq. (1.105). We had p(z") = X = 2,04, which transforms under the adjoint representation of SL(2,C) as: X'=MXMt, or ao" = Mo"a,M}. Using the index notation of the SL(2,C)-matrices M and Mt we conclude ao") 44 = Ma2(0") pav(M")? ;, (1.117) and we see that the Pauli matrices must carry a dotted and an undotted index. The matrices o# therefore map F into F, and similarly the matrices co” map F into F Next we calculate the determinant of a transformed Hermitian matrix. Using Eqs. (1.111) and (1.116) we have wia! = det(X') = det(MXM") = det(M) det (X) det(M') = det(X) = a,c", (1.118) since M € SL(2,C) the determinant of M equals 1. Thus the quadratic form 2," = det(X) is left invariant under transformations of the adjoint representation of SL(2,€). We now have the following construction. Starting from a Minkowski four- vector « € M4 we construct an Hermitian 2 x 2-matrix X € H(2,C). The determinant of this matrix is simply the Minkowski product of z, with itself.1.3 SL(2,C), Dotted and Undotted Indices 59 We then go to the adjoint representation of SL(2,C) on H(2,C€) and derive the result that the determinant is left invariant under SZ(2,C). If we apply the map p~! to the transformed Hermitian matrix, we obtain a four-vector z,, such that a,c!" = xy". We can demonstrate these steps schematically in the following way my 2 H(2,€) 24 H(2,€) 25 my, (1.119) that is: t4——> p(a,) = 2,0!» ad M——+ p(X") = «t, =X =MXmMt =X. This transformation from M4 to Mg is, of course, simply a Lorentz transforma- tion which transforms four-vectors into four-vectors leaving the Minkowski square, i.e. the line-element, invariant. Now from Eq. (1.1) we have al = Myx. On the other hand we can also write, using Eqs. (1.112), (1-116) and (1.111): att = Lay[x'o"] = br[Mxmtat] = p0y[Ma,o" Mio" = y[Mo' Mita", = 51 [o"Mo M2". (1.120) Comparing this expression with Eq. (1.1) above, we obtain: (1.121) Equation (1.121) is the explicit form of the group homomorphism SL(2,C) —+ Li. Proposition 1.24: The following properties hold: (i) Equation (1.121) is a group homomorphism, i.e. V Mi, M2 € SL(2,C) we have: AM, (M1)AY (M2) = AY (Mi M2), (1.122) and (Av*)M (4) = AFM"). (1.128)60 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors (ii) The matrix A, defined by Eq. (1.121), is an element of the restricted Lorentz group L., i.e. we have to show: det [A“,(M)] = +1 (1.124) and A°(M) >1. (1.125) Proof: We first prove Eq. (1.122). Using Eq. (1.121) we have a AY (M))AY (M2) = 7D [e"Mio,M{) Tr(a" MoopM}) = ED [Ma" mo.) 19" MzopM]] = 7 (Mtoe) 59" MopMiy; 1 F i o qleaton)”? (ov) ga )A4 [Moo pM3] 4 4 1 B ; 5 lion}? 56,4554 [Moo pM] 44 1 ar AA + a 5lMion Mi] [MropM3] 44 Tea ; = 5 [Miz*M.Meo,Ms)*; 1 - pi [Mia"M, Macy Mj) 1a eo = 5 B[a"( Me)o,(MiMe)*] = A¥(Mi M2), again using Eq. (1.121). A direct consequence of Eq. (1.122) is the following result: (A71)*)(M) = AH,(M~), (1.126) which we can prove as follows. Using Eqs. (1.122) and (1.109a): AM (MAY (Mo!) 2?) A (MM) = MAA sz0) fee (1.109a) 1 hayfateg] U2 3 ont, = 64, In order to prove Eq. (1.124) it suffices to recall that A*(Lsz2,c)) =1.3 SL(2,C), Dotted and Undotted Indices 61 and therefore det (A",(2s202,0))) = det (5",) = 1. Furthermore, det is a continuous function of A", so that det(A",(M)) must be +1 as M runs through SL(2,C). Finally, Eq. (1.125) may be shown as follows.?° Using Eq. (1.121) we have 1 1 A%(M) = 5 Tr[a°MooM!] = 5 Tr[MM"], since F = 0° = 09 = Nexo. Now let M € SL(2,€) and take U € U(2,€) such that UMUt =D, where D is diagonal. Then (Umut)! = Dt, or uMtut = Dt. Now 1 = det M = det(UMU) = det D = ayaa, where a1,a2 € C are the eigenvalues of M. Similarly 1 = det Mt = det(UM'U") = det Dt = aja}. Using the fact that a trace is invariant under similarity transformations we can write Tr[MMt) = Tr[UMM'Ut} = Tr[UMUtUMtU'] = Tr[DD'] = |a,|? + fag)’. ‘Taking into account that a) - az = af - a} = 1, we can write 4 1 T[MM1] = ar? + = ee la| las)? We now set = [af eR. 26We introduce a unitary matrix U which diagonalizes M.62 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors Then: +1 j= St With this relation we can show that Tr MM? > 2. Suppose we have the opposite, i.e. Tr[Mut Tr[MM!] <2. Then : e+1 <2 c which implies (e-1)? <0. But this is impossible for any c € R,. Hence we conclude Tr[MM‘*] > 2, and hence AY)(M) = 51e(MMMt] > 1. Thus, A“,(M) as defined by Eq. (1.121) is, together with Eq. (1.124), a restricted Lorentz transformation. Our next step is to derive the map inverse to that of Eq. (1.121), i.e. given any Lorentz matrix A“, the question is: How can we construct the corresponding SLZ(2,C)-matrix M = M(A). We start by considering the action of the restricted Lorentz group on Minkowski space M4: cH= At’, chem, AEL. (1.127) However, we know that the action of SL(2,€) on Hermitian 2 x 2-matrices is given by the adjoint representation: X'=MXMt, MEéSL(2,C), X €H(2,C), or using relation (1.111): aio" =Mayo"Mt, ao, = Mx’o,Mt. Using Eq. (1.127), MY a’o, = Mx’o,Mt, so that Mo, = MoyM'. (1.128)1.3 SL(2,C), Dotted and Undotted Indices 63 Proposition 1.25: For every M € SL(2,C) o.Mo" = 2Tr[M] Lox2 (1.129) Proof: This can be shown by explicit verification. Muitiplying Eq. (1.128) by &” and summing over v, thereby using Eq. (1.129), we obtain for M, Mt € SL(2,C): Mo,0" = Mo, Mio" = M(2Tr[M]) 19,0, (1.130) and so M(A) = Moyo". 1 IMT Taking the determinant of Eq. (1.130) we obtain: det (A",040") = det{M(2TrM")] = det M - (2TrMt)? - det 12.2. Using det M = 1, since M € SL(2,C), we can express TrMt as a function of A, ive. f2 otret = +{det{A",0,0"] } ‘ The final expression for M as a function of A is therefore: (1.131) In components, Eq. (1.131) reads 1 : M4? (A) = +-————p M (oy) an(0")”?. {det [A",0,0°]} Thus the Lorentz indices y,v are effectively exchanged via 0,7 for spinor indices A, B. Equations (1.121) and (1.131) show the connection between the two groups SL(2,C) and Li. An important point is that both SL(2,C)-matrices M and —M lead to the same Lorentz transformation as can be seen from formula (1.121). In Eq. (1.131) this fact corresponds to the + signs. Thus the correspondence A <> +M defines a two-valued representation of the64 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors restriced Lorentz group, which is called the spinor representation. This leads us to the identification Li, & SL(2,€)/Z2, (1.132) and so SL(2,C) is the universal covering group of L', which means the SL(2,€) is simply connected [104]. 1.3.4 The Fierz-Reordering Formula The Fierz reordering formula exchanges the order of anticommuting spinors. Proposition 1.26: The following relation holds for spinors 7%, ¢, x: i (GP)X 4 = (GO? R) (Hou) 4» (1.133) Proof: Using Eqs. (1.92a), (1.84), (1.86) we find: (OW)xX4 =o? bBX, = Peace jpx? = -ecng eye gpx” = cay Pei px° = 6405 MP ecneg py e?X”. Using Eq. (1.110) and finally Eq. (1.106a) we arrive at: 1 : (GW)X4 = 300) Aly Cecre nny oP? il * BH i = yl) ie" Op) py ecnen nV OX” i : = 50") salu) eH” Bd" pwroPX? 1 i = 59 (on aoX Yo") a4 1 ~ = 4 (404%) (0%) For further Fierz reordering formulas see Eqs. (1-101), (6.90), (6.91), and (6.92).1.3 SL(2,C), Dotted and Undotted Indices 65 1.3.5 Further Calculations with Spinors We begin by deriving some more formulas which are very useful in calcula- tions with spinors. Proposition 1.27: Let ¢ and x be spinors and o,a” given by Eqs. (1.105) and (1.106a), then the following identiy holds: (¢0%X) = —(xo"#) (1.134) Proof: We start with the expression on the left: ) (1928) 4 89) (AB bey M 2 Me Me : (¢0"X) ia (0%) ax? ta on(o") spe Xe C3 1,106: | fa = —bpeP (alt). pe®oXe | ad ~op(-a") PX 2) x(a") P by = -(Za"6). Proposition 1.28: Under Hermitian conjugation the binary spinor expressions ¢o"X and (@¢) transform according to: (¢o"%)! = (xo"), (1.135a) (69)' =6¢. (1.135b) Proof: The first expression, Eq. (1.135a), is $L(2,C) covariant. There- fore, with the definition of the {-operation (Got)! = (xtoMt gt) = xIA(0Ht gt? = x4(oM) 4a = (x08). Here we use: VAr= yt and gi8 = 8". It may be observed that the Grassmann property, Eqs. (1.94) and (1.95), has not been used because (¢x)! is defined as Xt ¢t. Similarly, Eq. (1.135b) can be demonstrated as follows: (66) °2 (gat =o1,'? = 35. From Eq. (1.135a) we see that yo“ is Hermitian.66 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors Proposition 1.29: As can be expected, the expression (¢o/X) transforms as a four- vector under the Lorentz group, i.e. (¢'o"X’) = A",(M) ($o"X)- (1.136) Proof: We have: (Worx) C2) — xara") = -x4(o")4? oy ~(M*) Px p(o")4? Mp Coc = -X_(Mtt)? (4? My de = ~Xa" (MN? 5(0")4P My Ec bv —F¥HP Corea MI)? ("PM yb E males i(o")" Mp“(ov)ea] (Ra(@")?? 40) a -har[Mi4 Mor] (xa") = A¥,(M)($o"X), using Eqs. (1.121) and (1.134). Proposition 1.30: The following relations can be shown to hold: ge = (oP aa? (608) = ba (08), (1.137) (80¥6)(00%8) = jn” (80) (00). Proof: Using Eqs. (1.92a) and (1.92b) we have (808) (608) O4(o!) 498 (oop = -04(04) 30°" (0 o30” a = -(0") 45(0"\op0400?0” a Flo aalo"ope*eP? (00) (80) Co i ) gn(@")°4 (06) (88) oo ite 2).1.3 SL(2,C), Dotted and Undotted Indices 67 Proposition 1.31: The generators of the group SL(2,C) in the spinor representa- tions (1/2,0) and (0,1/2) are given by (see Eq. (1.131)) ol = Floto - ova"), (1.138a) at = =(GHa" — aa"), (1.138b) (om) t = oH, (1.138c) We will need these two-dimensional spinor representations of the Lorentz generators later for the construction of the complex two-dimensional grading of the Poincaré algebra. In components, Eqs. (1.138a) and (1.138b) read: ft tie A a (ot) 4? = Z[(o") 4a(0")*? = (0) 440")? ], (ay, = L@AA(0") 4p — YAO!) a5) Proof of Eq. (1.138a): For an infinitesimal Lorentz transformation (w;; infinitesimal) we have: Moya” = Kyoto” 1 wor Woz wos _—[-eo -1 wi wis bn 7 ote wp. —wi2 —1 wg —w3 wig ang —1/ 4 — 2(wo3 + iwy2) —2(wo1 + twas) +2i(wo2 + iws1) —2(wo1 + twas) —2i(wo2 + iwsi) 4 + 2(wo3 + twa) Hence: det (A“,0,0") = 16 + Ow’), and, omitting terms of order w?: He N= [det(a",0,0")| = The relationship between an element A’, of the Lie group SO(1,3;R) and an element M“” of the Lie algebra so(1,3;R) is AP, (w) = [exr(-juum”)]"68 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors For infinitesimal parameters w,,, the element M(A) € SL(2,C) is M(A) = FM lwo" = 55 (tor ~ fee M! oP = ven = Seat, where 1 aM t= py (MM) 0°97. Now 2 = Mgr” = (pr + Wpr)2", and since every homogeneous linear transformation group possesses the triv- ial self-representation in the four-space spanned by the z,, we have: ww (MM) = Wor. This equation implies for the M’s: (ery? = i (nt a”, — nPPdH,). Of course, from Eq. (1.11) we know that the generators of the Lorentz group have this form. Hence a se a a ey at = 4. (tes, — #8, opa? = ¥ (ator — oot). In a similar way, a4” is obtained by considering M* as a function of the Lorentz transformation A. The matrices 0” and @” obey the commutation relations of the Lorentz algebra so(1,3;R), i.e. Eq. (1.14). Proposition 1.32: The generators of rotations Ji = 1,2,3, and Lorentz boosts K;,j =1,2,3, defined in Eqs. (1.19) and (1.20), i.e. - i= yee, Ki = Mo can be written: = 7 j a F = SA yo = Sota, (1.139a) (1.139)1.3 SL(2,C), Dotted and Undotted Indices 69 Proof: pee ‘i oF eel & Fe = Seigno® = Sein (o93* — oFat) WY) Hein. (—oFoh + ota!) i i 7 1 1 iE gcse ila, = zenilon = 3h = 9% = — ein [07,04] where we used 3 DY egee™ = 26. jk=1 Relation (1.139b) is shown as follows: K;=0%= 5 (oa -o'@°) where we used again Eq. (1.108). Proposition 1.33: The o” are selfdual, i.e. ol = Jet? ope, (1.140a) whereas the 4” are anti-selfdual, i.e. ay = ~ peo pn: (1.140b) Proof: In order to prove the first of these relations consider (i) p=0,v = i;4 = 1,2,3: t 1.108) @ oP! = £ (oa! — ata?) OY) 4 4 On the other hand with i, j,k = 1,2,3: 1 1 lt ipa: a ijk Ee Oijk pee Hae Opa = FE Ojk = we Tos, OKO. 2i po = gg Moje = 30g (Ose — OH i) = go (290 +0403) = — per [o5, 04) 04; t 1 vise, vl = — = tsk; = — Xk Ce te jKIe 0, we obtain: hy OuUw = 0, (1.167a) which can be written, using Eqs. (1.159) and (1.153), (ae ("2 a0) ($3) ae and therefore the 4 x 4 matrix equation (1.166b) decouples into two 2 x 2 matrix equations: i(o"d,) > — mo = 0, (1.167b) i(G*0,) 4 — mp = 0, (1.167c) i.e. for m — 0 we have: a ARG, oy 6A, al g)Be =(8)48 (4 2) pe + 458, ~ (073) 46 (0) 4g - (0) ge “tape, (@) ac -(@0?)4, +64, Using a” as matrices and (0?)? = 1x2, we obtain: Seco 0 vat} 0 ah 4 gd] = Bae Cc ¢, We now investigate the individual y-matrices. p=0: (Re) aa = Yab(Y) se Yea" 1 64 (07)an\ (58° 0 2 (tei i. ( 9 »,) ( BP + (0%) sal)? (0) an + (*)av -3( —(g2)4P AD ~(8) 46 (92), | 5A, ) “(aie M4) -Co 2) again using Eq. (1.108), i.e. ? if 8% © )as\( 9 (o')ee\( 5c? we) “3 \_(g2)4B -A, (eye 0 ~(#)¢ 7 = -o? and (07)? = Lox2. (0?) 49(#1)8? ee (ae 7 (we? —(a')AD +. (6)4B(a1) peg ag (0?) 49 )°°(0 op ~ ("an ) Oop + PMO) ps98 CHAPTER 1 Lorentz and Poincaré Group, SL(2,€) and Spinors 1 ( (07a) 4? — (070?) 4? (0?a10") 4p - ae) <2 =(@)4? + ole’ g)Ad -(@0)4 + (#0), mle (o!0? — o?a!) 4? tet ate} (-# a (0!0? — 070) 2i(o5) 4 -(ealo? +0) ap (o 8 oe 2i(o)A 5 vin D 1 a) oe (on a) a 0 ) oe ean) Ale 2 re aA 2i(0°)4, 0 io), 1( &? Jas 0 (*)be fe? (ew 9 —(#2)48 64, (eae 0 — (REP 5 io (0 2) AB : (wybD — (e8 Vago? . (© (#2)42 + (22/48 (0?) palo?) (07) 43)? ep ae | =)? op + (49 (07) pp, 1 0 -(0? = 0°30?) 45 a9! 2 — 5292G2)AD 0 ) Again using Eq. (1.108), we obtain 2x2 -0? - 0°a?o? = 0°(0?)? = 0°, since (02)? = 1x2. Therefore: 1 0 —2(07) 0-0? (dai (aero 0") = (et) pak (Re) aa = Yao 7b) y-Yea™ oe bas (0) a8 0 (0%)pe a7) —(62)48 -A, (wy Be 0 ee1.4 Dirac and Majorana Spinors 99 u i (0764) 4? ~ (0332). (00) 49 - (0) ay ) By (BAP 4 (G%o3q2)4D ~(380?)4, + (#o8)4, aa (mene oat ) 2 0 -((07,04)*5, _ 1 f-2i(o') 4? 0 2 0 -2i(0')4 _ Ce 0 ) = va 0 alder Now the connection between the Weyl representation and the Dirac rep- resentation is given by Eq. (1.173a), Tp =X7'TyX. The Dirac representation and the Majorana representation are related via Eq. (1.180), Ty =YPpY"l. Thus Tu =YX'TwXxY", or, using Y = Y~!, we obtain the similarity transformation that connects the Wey] representation and the Majorana representation: Ty = (XY)7'Pw(XY). (1.182) With L (-42x2 0° ) —1\T -1yt oo =(x-')7 =(x7y, 3 (7? 2) ey = 0) de ioyo ae + -1 Y=— =Ytey ee i) a we obtain: tae Gap 53° (one XarYoe=3\ _coyas _ 5A gyic _gb 2 \-(a) -54, } \-@) -58, 1 (548559 ~ 0) 4n@)P° -447(0") pe ~ (0°) 458% 2 \ -(99)4B 5,6 + 64, (5?)P° -(59)4B(G?) gg + 04,55", ey sitar +(07)4° (0) 46 - (ae ~ 2\ _(@yde 4 eae ~(0?)A, +54, :100 CHAPTER1 Lorentz and Poincaré Group, SL(2,C) and Spinors which is usually written in matrix form as: 1 (-1+0? -o?-1 xy=3 (47% 751): (1.183) The inverse, (XY)~!, is a eT lft o?\ /-1 -o® loy-ty-be (XY) YUCX z ( 2 ) ( 0 4) +077 —o° a Fo + :) > (1.184) which is usually written -1_1/-1+4+067? -1-¢? t => = ) (xY) z (he 1 0? 41 (xy), (1.185) or (xy) =y-'x-t = ytxt = (xy), However, for a clear index structure it seems to be advantageous to use the explicit form Eq. (1.184). Of course, Eq. (1.185) is the same matrix as Eq. (1.184) since as matrices: F = loxe, Ce Oa and so on. Next we check that he = (XY) My (XY) 0 0 o Ww=\x o}° We perform this calculation in the compact matrix form of Eq. (1.185): 7 1 (-1+0? -1-07\ (0 1 2p? = 1 (XY WY) = 3 (Ch -1 4) ( 6) (2 -o? 1-0? ) _1((-1 = 07)(o? - 1) + (-14 9?)(-1 - 07) 74 (1 - 0?)(o? - 1) + (0? +1)? —(1 + 0?)(-0? — 1) + (-1+0°)(1 - 0?) (1 - 0?)(-0? - 1) + (-0? - 1)(1 ~ 0”) _1(0 40%) _ (0 o\_ 19 fade 0 eo 0) In the last step we made use of Eq. (1.178). The other 7-matrices can be checked in a similar fashion. with1.4 Dirac and Majorana Spinors 101 1.4.4 Charge Conjugation, Dirac and Weyl Representations The charge conjugation matrix appears in Dirac theory*? in the following way. The Dirac theory implies the existence of electrons and positrons, par- ticles with the same mass but opposite charges, which obey the same equa- tion. The Dirac equation must therefore admit a symmetry corresponding to the interchange of particles and antiparticles. We thus seek a transformation vo YP, which reverses the sign of the charge, so that the Dirac spinor Y obeys the Dirac equation (:@-eA-m)¥=0, A= Ayr", (1.186) in the presence of the electromagnetic vector potential A,, whereas the charge conjugated spinor Y° obeys*> (i P+eA—m)¥°=0. (1.187) ‘The Dirac equation coupled minimally to the electromagnetic field is [7 (id, — eA,) — m]¥ = 0. Taking the complex conjugate we obtain: [y"* (id, — €A,) — m]¥* = 0. ‘Transposition yields WL (i By ~eAn) ~ mi] = 0, Wty? Pitty? (i By -eAy) Wy" (-i dn ~eAy) — m] = 0, where is the Dirac adjoint, and Potty? = yf. ‘Taking again the transpose we obtain =T [-7"T (i8, + eA,) — mW = 0. See e.g. the discussions in A. Pais [86], p. 381, and J.D. Bjorken and S.D. Drell [17]. *8For further discussions of these aspects see M.D. Scadron [103], Chap. 3.B.102. CHAPTER 1 Lorentz and Poincaré Group, SL({2,C) and Spinors Multiplying this equation by a 4 x 4-matrix C from the left and inserting C-'C in front of U", we get: C[-1"" (id) + e4,) — mlO“'0" =0, [-Cy""O"! (18, + eA,) — m|CW" =0. This equation can be identified with Eq. (1.187) provided we set: w= C0'," (1.188) except for a phase factor. In addition we have to demand that in any repre- sentation of the y-matrices CyTOnh = 7". (1.189) The matrix C is called the charge conjugation matriz. It suffices to construct the charge conjugation matrix in some particular representation; the unitary transformation which transforms to another representation then gives the matrix C’ in this new representation. We consider several representations. The Charge Conjugation Matrix in the Dirac Representation In the Dirac representation (1.56) the charge conjugation matrix C may be taken as j 0 oe Cp = inn =i (@ a ) . (1.190) Proposition 1.46: The charge conjugation matrix Cp defined in Eq. (1.190) pos- sesses the following properties: Cp = -C5! = -ch, = -CB. (1.191) Proof: We first verify Cp = —C5. Consider ara eRe —(Co)ab(C) be e a (caryee “r) tl I eee: aL i o1.4 Dirac and Majorana Spinors 103 (1.108) (a 0 ) OG. This shows that Cp = —C>}. Next we evaluate the Hermitian conjugate as follows: aay Nt gt (Chu =|3( rae ()aa\| =i . Mare aa (=) 9 (07) 0 0 -(0%) a5 “ (cosie 7) se using the Hermiticity of the Pauli matrices and ? = —o? from Eq, (1.108). Finally, the antisymmetry of the charge conjugation matrix Cp is shown as follows: 0 -(%)4,\'_.f 0 27) 45 (Co) ay =# (corpse oe) rt (cerryae : ee) 0 seat a 0 2 = (geryis ")*) = (Lierae ° 0) 0 fie at’ at (cosy | = ~(Co) ay Here we made use of Eq. (1.108) and the antisymmetry of the 0? matrix. Proposition 1.47: The charge conjugation matrix Cp, defined in Eq. (1.190), sat- isfies relation (1.189), i.e. Cop! Cp! = Yh. Proof: First we attach suffixes D to Eq. (1.189) into the following form. Corb! Op! =~}. Using Eq. (1.191) this can be written as . -CpYp' Cb = ~Y-104. CHAPTER 1 Lorentz and Poincaré Group, SL({2,C) and Spinors Transposition gives . ChyYBCD = Yb» and using the properties (1.191) again yields CpypCo = yee Consider the case ps = 0, i.e. (Cp)a(Yb)be(Co)ea = Next we work out the case y= 1: 1 : Cle1.4 Dirac and Majorana Spinors 105 Using Eq. (1.190) we have: 0 -(0%),; 0 ives (Co)an(rb)te(Cr)et = — (capa “ae (apse 7 je) 0 (92). : (een 7) : -( oe aslo" (2)49(0!)go(o")?? 0 a -( 0 : ne) ae (@ala?)AP 0 = 0 —(0?010?) 45 (o?at92)AP 0 a 7 =- (Gave (o .*”) (cate? = -o') ease _tityia & 40) (0 = ot) = (Dea since 0 qe 0 ity. Oa (piso ao = ((mao i 0”) 0 ("ap ee (#T)Ad 0 : Hence we have demonstrated that CorbCr = (1b) Next we evaluate the case ps = 2: 0 2: i (bas= (gayi 9M) Using Eq. (1.190): 0 ~(0?) ag 0 (0)b6 (Cp)av(Yb)te(Cd)ea = - (geo 0 (ease 0 ) tee“)106 CHAPTER1 Lorentz and Poincaré Group, $L(2,C) and Spinors {0 (2) 48° cp (3?)49 (0?) ge(@V°P 0 . 0 (078? o) 45 (F02a2)AD 0 = (gran a : Now the transpose of 77, is: 0 2) .\T 0 elie GoD aa we = ((aro a 0”) Then: oe 0 ee (Co)aalrb)te(Co)ea =~ oC. ae -e ae so (gs40 ] Now 0 ayn 0 (2) aa OB Yaa = (asy4o a 0) = (ary : ) (aie 4) (atin1.4 Dirac and Majorana Spinors 107 using Eq. (1.108) and 0% = 037. Hence . CoybCp = (yb) | as had to be shown. The Charge Conjugation Matrix in the Weyl Representation We transform the charge conjugation matrix Eq. (1.190) from the Dirac rep- resentation to the Weyl representation by using Eqs. (1.173a) and (1.173b); thus Cw = XCpX7! =iXyh7hXx7! ary = a i) CC) -io’ = iXyp X71 X49) x7! We prove Eq. (1.192) in the submatrix formulation, avoiding cumbersome indices. Thus Cw=i(-t °°) (9 -2\ (-1 -0° WE 2\-09 -1)\o? 0 J\O -1 i (Goze - 0°? -o? — aie) © 2100? +e? ao? + Fo" et (2on 08) (doe 0 ~2\0 -207) ~\ 0 -io?}* Remark: The correct form of the charge conjugation matrix in the Weyl representation, possessing the correct index structure, is 250 to*e 0 Cw = ( ; ia’) (1.193) with: 7 (Cw) i(0?a®) , 0 W) ab = fe . 0 i(a*o°)*, Of course, the matrix (1.193) is the same as Eq. (1.192), since 3° and o° are unit matrices, which are usually ignored in the literature. Proposition 1.48: The charge conjugation matrix Cy in the Weyl representation also satisfies Cw = -Cy} = -Ci) = -ch,, (1.194)108 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors and Cwrty Cy = —y. (1.195) Proof: Equations (1.194) and (1.195) can be checked directly, using Eq. (1.192) and Eq. (1.159), or alternatively with the help of Eqs. (1.173a) and (1.173b). Thus using first Eq. (1.192) and then Eq. (1.191) we have: Cw = XCpX~! = ~XC05!X7! = —(XOpX"') | = CF}. Also Cw OS? xcpx7) 2? xh xe! = (x7 ep x") * (1.182) ~(X@pX7!)* (2.192) -cy. The last property in Eq. (1.194) is shown as follows: Cw = XCpX = -Xchx" = -(x~Hopxtyt = —(XCpx7')' = -chy. Finally we verify Bq. (1.195). We know that in the Dirac representation we have: Coy,C5! = -78"- Hence: CorbCp' = —Ibs = X(CovgCp)X-1 = -X (76) X73, = XCpX7XyhX-'XCGIX7) = -X (7p) XT, 3 XCpX1 XY XA(XCpX' | = — (XoXo) and so ee T Cw Cay = -Hy - Further properties of the C-matrix Proposition 1.49: Independently of any representation, the charge conjugation ma- trix satisfies the following relations:14 Dirac and Majorana Spinors 109 CYC! = (°)", (1.196) O(ryH}O7} = (Py#)" (1.197) Proof: We first show Eq. (1.196). Thus, since Eq. (1.162a) is independent of the particular representation we have: oC =ic Py pot =i0yC"Cy'o" cro“ oxo" = if yt pst ( with (1.184)) = i(Py2y'y)™ = i(y yy)" 7 =(%)', using the basic Clifford algebra relation (1.160), which is valid in any repre- sentation. Equation (1.197) can be shown in a similar way: CoP yH#O7! = CCE HOM = = (9) (4) = -(9"')" = (oT with Eq. (1.164). 1.4.5 Majorana Spinors We consider a Dirac four-spinor in the Wey] representation. According to Eq. (1-156) we have: _(¢ ba . W= (§).- a ) eFoF"*. (1.198) For the index calculus to make sense, rows and columns must have the index structure of (i), 3), amt w),= (52) respectively. Thus (es us) és) _ My?d0+ Mygb” MAB MA, v M4865 4 MAD”110 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors and M, Map ~ ; eae ace (#4, wie ms) = (94mg? +B? 94M, 8 + Bam). B. We now observe that the relationship between left-handed and right-handed Weyl spinors may be written = PHO )PA, 84 = 80). (1.199) ‘The consistency of these relations with Eq. (1.63), i.e. wad, (1.200) follows from the fact that the 0° and @° matrices are 2 x 2 unit matrices. It may be noted that 44 is a two-component column vector, whereas WA isa two-component row; similarly P' is a column and a is a row. The consistency of Eqs. (1.199) and (1.200) can also be seen by calculat- ing the Dirac conjugate of W, i.e. Using Eq. (1.200) we have: Tw == (4s BD (so, B°) = Gi =, 34): Note that since Eq. (1.200) does not preserve a consistent index structure the matrix representation of 7° must be used. On the other hand, using Eq. (1.199) we have: a 0 0) Ww = Vip = (64%, Ba) ae te 38) 1.199) - = (Gi@) AB, g4*(0°) 4 A) OL (YF, by (1.201a) in agreement with the previous result. We note that Ty = (5) : (1.201b) The charge conjugate of a Dirac spinor in the Weyl representation is defined as in Eq. (1.188) with the charge conjugation matrix Eq. (1. 192), te. Sy = CwTw.1.4 Dirac and Majorana Spinors M1 Thus using Eq. (1.193) and Eqs. (1.201a), (1.201b), we have: (Yi), = (Cw) aT), ' (eye ay (3) (7 mt 3) Now (irony? = ( and (i@o°) ig = ( Hence (io?a°) 4? = erc(io?a®)? = eace®? = 54°, and (ia?o°)* pa 4° (igo) =e ea = Hence: (Viv), = (5) (1.202) Thus, charge conjugation flips ¢ and y. A Majorana spinor is a four-component Dirac spinor which satisfies (here in the Weyl representation) Uy = Vy. (1.203) Using Eq. (1.202) and the explicit form of a Dirac spinor in the Weyl repre- sentation, i.e. Eq. (1.156), we obtain eee) ee o e ‘Thus for a Majorana spinor in the Weyl representation we can write: wi = (5) . (1.204) Thus a Majorana spinor has only two independent complex components and is therefore equivalent to a two-component Weyl spinor or a real Dirac spinor.112. CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors 1.4.6 Calculations with Dirac Spinors It is useful to know the connection between the four-component Dirac for- malism and the two-component Weyl formalism. The use of the Weyl repre- sentation of the four-component Dirac formalism has certain advantages. We use the following notation for Dirac spinors: ba w=("), 1.205; (%) 2 where the subscripts + distinguish between Weyl spinors which are elements of representation spaces F (+) and F* (—) respectively. Then according to Eq. (1.201a), the Dirac conjugate of Eq. (1.205) is: =(¥4, D4). (1.206) Proposition 1.50: The following relations hold for Dirac spinors Y and y: (Ux), = (W-x+) 2 + 4X-)os (1.207a) (Bx), = —(H-x+)2 + PX Jo» (1.207b) (Y'x), = ©. met (v-0"X_) 5, (1.207e) (Fy? x) = (d-0" XJ — G45" x4) 95 (1.2074) (Woy), = (v-o8' “a+ (B74 R Jos (1.207e) where : of =o [yt]. (1.207f) Proof: (a) On using Eqs. (1.92a), (1.92b), and (1.93), we obtain for Eq. (1.207a): (Fa), = (4, Baa) (54) = vera + Bat = (b-x+)o + (OX) o- (b) With Eqs. (1.92a), (1.92b), and (1.93), we can prove Eq. (1.207b): Tye T.45 A moa 0 (xe (GPx) 4 = Vortoxe = (44, Ya) je 4 a B e = (04, Ba) (2G) —Whxea + Bad = —(Y-x4)2 + (BX )o»14 Dirac and Majorana Spinors 113 (c) Proof of Eq, (1.207¢): a 0 "ap (Gtx), = Vorkxe =” (WA B a) Ce 7 7) (S ) (nyA 0 = (47a) (Cag ) PX B = V0!) sax! + D4 Pxee = (bo F_)y + G.T"x4)o» again using Eqs. (1.92a), (1.92b), and (1.93). (d) Proof of Eq. (1.207d): (Go' x) 4 = Va Vea teXe : 0 (o#)4an\ (~o2° 9) (xse = (v4, ¥,4) (ee u) ( 0 #,) (<2) “a 0 a a = (#2, 44) (cone 7 aK 3) . *) axe = (4, B,) (“ oes ) = V8(0") 5X2 - Bia) xee = (pot X_)y = (GF x4) a» again using Eqs. (1.92a), (1.92b), and (1.93). (e) Proof of Eq. (1.207e): Before we demonstrate Eq. (1.207e), we discuss Eq, (1.2078), ie. ("y= G(71),,= *(otent, ~ hers) (gic OMe (amyae “) (ere ")114. CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors i oe y's) using Eqs. (1.138a) and (1.138b). With this expression one can directly verify Eq. (1.207e): aa o! 0 Won), (8, (MH cana,) (A) = JaPX+B = (v4, Bea) (oe wg) = VA (oh) aP xia + B44 gee = (H-08"x4)9 + (G472"X-)o» using Eqs. (1.92a), (1.92b), and (1.93). 1.4.7 Calculations with Majorana Spinors As explained in Sec. 1.4.5, a Majorana spinor has the characteristic property that we can replace the dotted Weyl spinor by the complex conjugate of the undotted one. In the terminology of the previous section this means that we can replace y4 by v4 and @° by Bs. Thus we can drop the suffixes + in Eq. (1.205) and write Un= (3!) i (1.208) In this case, Eqs. (1.207a) to (1.207e) read: (Gurxar) 4 = (bx)2 + (PX)o> (1.209a) (GarPx), = —(4x)2 + (PRos (1.209b) (Gur'xur), = GF"x+) 2 + (YOR) »» (1.209c) (Emr'Pxar), = (¥o"®X)_ — (Fe"x)y, (1.2094) (Guo’xm) 4 = (Yos"x)y + (072). (1.209e) Proposition 1.51: The following relation holds for Majorana spinors: (Gaerty ar) (Em PWa) = 0! (Cua) (1.210)1.4 Dirac and Majorana Spinors 115 Proof: Using Eq. (1.2094) we have: (Eu) (Eure va) (1.2094) [- Ge"), + (vor) | [- Go"), 7 (¥o"),| 29 [(otD), + (o").] [(Uo"D), + (Wo"),| = 4(Yo"B),(¥o"'B), OE” 20h” (W¥) 9B D>. For the present case Eq. (1.209a) is (Waa), = (¥%)o+ (b4),. Taking the square of this expression we obtain: (Gar ¥an),)? = (WY). + GV)” = (WH) 2+ (WO) 2 + %).(0¥) 2 + (FV) .(0%)y = 204) (FF), since as will be shown below: (Ve) (0%). = (UD) o(FH)y = 9, and (W¥)20¥)2 = PVo(WY)o- Hence it follows that (Gu Var) Gary’ Var) = 0” (Ea Yan) §- We now show explicitly that (y)3 = 0: (¥)2 = (H¥), (2), O29 (yAga) (vn) 039) Ao cpac doen = ACP bobadodn = Ppp bat + 72 yadda te dodo + edi dedi de = vovrdodr — dod dive — dridedati + didedide-116 CHAPTER 1 Lorentz and Poincaré Group, SL(2,C) and Spinors Taking into account the fact that #4 are Grassmann numbers, we obtain (cf. discussion following Eq. (1.96)) (wo); = 41d vod = 0, since for Grassmann numbers 4) {vv} =0 implies biti + bith = 0, ie th = A similar calculation can be carried out for (0%);. We now verify that (B¥)_(¥d), = (YH) o(¥9).- To this end consider PV a(00), = Fil Ws = Ta Tn = VDT ba = Vdd O" = (¥0), (OV We prove one more formula. Proposition 1.52: The following relation holds for Majorana spinors: ¥ 2g 2 (Gu Uo), = —- Gua) (1.211) Proof: Using Eq. (1.209b) we have: (Futa)3 = {-)a# +@B} = (WW)} - 204) (B9)_ + FD = -2(U0), (09), = -{(6v)3 + 2(0v) (8). + G3} -(Gu¥m)y, where in the last step we used Eq. (1.209a).Chapter 2 No-Go Theorems and Graded Lie Algebras 2.1 The Theorems of Coleman-Mandula and Haag, Lopuszanski, Sohnius We now discuss the two theorems already referred to in the Introduction.! 2.1.1 The Theorem of Coleman—Mandula ‘The following theorem? was established by S. Coleman and J. Mandula [23]: Let G be a connected symmetry group of the S-matrix, ie. a group whose generators commute with the S-matrix, and make the following five assumptions: (i) Lorentz invariance: G contains a subgroup which is locally isomorphic to the Poincaré group. (ii) Particle finiteness: All particle types correspond to positive- energy representations of the Poincaré group. For any finite mass M, there is only a finite number of particles with mass less than M. 14.0. Barut and R. Raczka [8], p. 43, also give considerations concerning the unifica- tion of the Poincaré algebra with internal symmetry algebras. Stronger group theoretical lemmas are given on p. 629, and it is pointed out (p. 630) that the infinite-parameter Lie algebra associated with noncompact dynamical groups (which lead to infinite particle multiplets (cf. pp. 411, 609)) does not contradict these theorems. See also the discussion in S. Weinberg [118] : 7118 CHAPTER 2 No-Go Theorems and Graded Lie Algebras (iii) Weak elastic analyticity: Elastic scattering amplitudes are analytic functions of centre-of-mass energy squared s and invariant momentum transfer squared ¢ in some neighbour- hood of the physical region, except at normal thresholds. (iv) Occurrence of scattering: Let |p) and |p’) be any two one- particle momentum eigenstates, and let |p,p') be the two- particle state constructed from these. Then T\p,p') # 0, where 7 is the T-matrix defined by $= 1-6(2n)*6'(py - pT except, perhaps, for certain isolated values of $. In simpler terms this assumption means: Two plane waves scatter at almost any energy. (v, Technical assumption: The generators of G, considered as integral operators in momentum space, have distributions for their kernels. Then the group G is locally isomorphic to the direct product of a compact symmetry group and the Poincaré group. We recall briefly some basic results of scattering theory. The Hilbert space H is the direct sum of an infinite number of subspaces, #.e. H=MM OHM. Here H denotes the n-particle subspace. It is a subspace of the direct product (symmetric or antisymmetric in accordance with the generalized exclusion principle) of n Hilbert spaces, each being isomorphic to HW, The S-matrix is a unitary operator on H. A unitary operator U on # is said to be a symmetry transformation of the S-matrix if (i) U transforms one-particle states into one-particle states, (ii) U acts on many-particle states as if they were tensor products of one- particle states, (iii) U commutes with S.