资源描述:
《Solution of the Anomaly Puzzle in SUSY Gauge Theories and the Wilson Operator Expansion》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库。
NuclearPhysicsB277(1986)456-486North-Holland,AmsterdamSOLUTIONOFTHEANOMALYPUZZLEINSUSYGAUGETHEORIESANDTHEWILSONOPERATOREXPANSIONM.A.SHIFMAN*andA.I.VAINSHTEINInstituteofNuclearPhysics,Novosibirsk90,USSRReceived23April1986Thepresentpapercompletesaseriesofworksonflfunctionsandtheanomalyprobleminsupersymmetrictheories.Exactexpressionsfortheflfunctionsareobtainedwithintheframeworkofstandardperturbationtheory.ThekeyobservationisthattheWilsoneffectiveactionSw(~)doesnotcoincidewiththesumofvacuumloopsintheexternalfieldF(/O.Thedifferenceisduetoinfraredeffects.Thecoefficient1/g2infrontoftheoperatorW2inSwisrenormalizedonlyatone-looplevel(extensionofthenon-renormalizationtheoremforF-terms).Thisfactresultsintheone-loopformoftheanomalousoperatorequationforthesupercurrent(generalizationoftheAdler-Bardeentheorem).ThefullGell-Mann-Lowfunctionemergesafterpassingtomatrixelementsoftheoperators.Thequantityenteringobservableamplitudesdiffersfrom1/g2by5"ilnZiwherethefactorsZidescriberenormalizationofthefields.(InthissensetheZfactorsofthematterfieldsbecomeobservable.)Wediscusstherelationwithcalculationsoftheinstantontype.1.IntroductionItiswell-knownthattheperturbativeseriesinsupersymmetric(SUSY)models[1]possessmiraculousproperties.Thus,forFtermstheloopcorrectionsareabsentatall(theso-callednon-renormalizationtheorems[2]),whiletheGell-Mann-LowfunctionsinN=2gaugetheoriesareexhaustedbythefirstloop[3].InthepresentworkwewilldiscusscalculationoftheeffectiveactioninN=1SUSYgaugetheories,i.e.inparticular,thegaugecouplingconstantrenormalization.Intheliteraturetheterm"effectiveaction"isused,actually,intwodistinctsenses.Accordingtooneprocedurewecalculatevacuumloopsinexternal(back-ground)fields.ThefunctionaloftheexternalfieldsobtainedinthiswayF(/~)isoftencalled"effectiveaction",althoughmorepreciseisanothername-thegenera-torofone-particle-irreduciblevertices.Wewillsticktothelatterterminology.ThesecondconstructionisacalculationoftheeffectiveactionSw(~t)hlaWilson[4].ThedifferencebetweenSw(/~)andY(#)isduetothefactthatinthevacuumloopsforSwwekeeponlythecontributionofvirtualmomentap>/~.TheactionSw(~t)*ITEP,Moscow117259.0550-3213/86/$03.50@ElsevierSciencePublishersB.V.(North-HollandPhysicsPublishingDivision) M.A.Shifman,A.LVainshtein/Anomalypuzzle457isthenormalactionwithrespecttothelow-frequencyfields.ThesubscriptWintroducedaboveemphasizesthedistinctionbetweenthetwonotations.Thus,withintheframeworkoftheWilsonprocedurewedealwiththenormaloperatorproductexpansion.InordertopassfromSwtoFonemusttakematrixelementsofexp(iSw(/~))eir(~)=(eiSw(~)).Letusemphasizethatthedifferencebetweenthetwodefinitionsisduetothecontributionoftheinfrareddomainp~(pisthemomentumflowinginloops).Inparticular,inSUSYgaugetheoriesthecoefficientsinfrontofW~W~(1/g2wheregisthegaugeconstant)aredifferentinSwandF(startingfromthesecondloop).WewillshowthatthecoefficientofW2inSwisrenormalizedonlyatoneloop.TheconventionalgaugecouplingisdefinedfromF.Two-loopandhigherordercontributionstothelattercouplingcorrespondintheWilsonlanguagetocalculationofsomematrixelements.Quiteclearly,itisconvenienttostudytheultravioletbehaviourofthetheoryintermsofSw.Moreover,insolvingtheanomalyproblemtheuseofSwbecomesanecessity,notjustaquestionofconvenience,ifwearegoingtodealwiththeanomalyequationsintheoperatorform.AsareflectionofthefactthatthecoefficientofW2inSwisrenormalizedonlyatone-looptheanomalyinthesupertraceD~J~containstheoperatorW2withtheone-loopcoefficient.Justthisobservationsolvestheproblemofhigherorders.Theissuestoucheduponabovehavearatherlonghistory.Letussketchthesituationwhichwillallowustoformulatetheresultsinamoreconcreteform.Wehaveknown,already,forafewyearstheexactexpressionfortheGell-Mann-Lowfunctionfl(a)inSUSYnon-abelianmodelswithmatter,Ot23T(G)-EiT(R,)(1-fl(a)=-1-(T(G)a/21r)'(1)wherethesumonther.h.s,runsoverallmattermultiplets,7iistheanomalousdimensionoftheithmattersuperfield,dinZ(/t)a7i(a)-C2(Ri)--+...,din~~rC2(R~)isthequadraticCasimiroperator,TaTa=C2(Ri)I,whilethecoefficientsT(R,)determinethenormalizationofthegenerators:Tr(T~Tb)=T(R)8~b" 458M.A.Shifman,A.LVainshtein/AnomalypuzzleWehaveintroducedT(G)=T(Adjoint).Recallthatdim(Ri)T(Ri)=C2(Ri)dim(Adjoint)"Insupersymmetricgluodynamics(hereafterreferredtoasSSYM),inparticular,theexact/3functionisfixedunambiguously,a23T(G)/3(a)2~r1-(T(G)a/2~r)"(2)Eqs.(1),(2)havebeenobtainedinref.[5]withintheframeworkofinstantoncalculus.Fromtheverybeginning,however,itwasclearthatthereshouldexistadirectderivationinthestandardperturbationtheory.Thefactthatthe/3functioncanbewritteninasimpleform-forinstance,itreducestoageometricalprogres-sion(2)inSSYM-couldnotbeaccidental,andhenceonehastogiveanswerstotheobviousquestions:Howdorelationslike(1)and(2)emergeinperturbationtheory?Whatpropertiesofthetheoryareresponsibleforthespecificstructureoftheaseriesforfl(a)?Dothesepropertiesshowupinotherquantities?Apartialanswer,primarilytothefirstquestion,hasbeengiveninrefs.[6,7].Theseworksreproduceeqs.(1),(2)withnoreferencetoinstantons.Hereweshifttheemphasisfromthecomputationalaspecttotheconceptualone.AnotherlineofresearchwhichalsomadeitnecessaryforustocarryoutthisworkisthenotoriousanomalyprobleminSUSYtheories.About10yearsagoFerraraandZuminonoticedthattheclassicalsupercurrentS~andtheenergy-momentumtensorO~wereconnectedwitheachotherbyasupersymmetrytransformation[8].Inref.[8]aprescriptionhasbeenformulatedaccordingtowhichonecouldconstructthesupermultipletJ~,includingS~,O~andtheaxialcurrentG"Allthreeobjects,a~,S~and0~,areclassicallyconservedandhaveanomaliesatthequantumlevel.Grisaruhasindicated[9]thatifa~,S~and0,,enteroneandthesamesupermultiplet,justthesamepropertyshouldbeinherentinthecorrespondinganomalies.Inref.[9]ithasbeendemonstratedthatthisisindeedthecaseatone-looplevel.Theproblemarisesattwoloops.TheAdler-Bardeentheorem[10],establishingtheone-loopnatureofO~a~,seemstocontradictthemultiloopexpressionforthetraceoftheenergy-momentumtensorG-4aG~G~,(3)whichisusuallyquotedintheliterature.ManypapersaredevotedtoattemptsofsolvingtheanomalypuzzleinSUSYtheories.Alistofreferences,whichisfarfrom M.A.Shifman,A.LVainshtein/Anomalypuzzle459beingcomplete,includesmorethan10papers[11-22].OneofthefirstdetailedinvestigationswasundertakenbyPiguetandSibold[11].Unfortunately,inspiteofdefiniteprogressnorealbreakthroughhasbeenachieved.Letuselucidateourbasicassertionsconcerning/3functionsandanomaliesbyasimpleexample-supersymmetricelectrodynamics(SQED).Theactionofthemodelcanbewrittenas:Sw4e2(#)1fd4xdOWZ(,)fd4xd40(T.e.r+~e_.V),(41whereWisthesupergeneralizationofthestrengthtensorWo=-~D2&V=iX,(XL)--0,D(XL)--iO~F,~(XL)+020~X~(x~),(XL)~a=X~a--2iO~ga"(5)T(xL,O)andU(xL,O)arechiralmattersuperfieldswithcharges+1and-1,respectively.Theaction(4)istobeunderstoodinthesenseofWilson,i.e.alloperatorsonther.h.s,arenormalizedat#and1/eZ(/.t)andZ(/t)arethecorrespondingcoefficientfunctions.ThemasstermmTUIFisomittedineq.(4),since#isassumedtobemuchlargerthanm,/t>>m.Themaximalvalueof/~isequaltoMo,theultravioletcutoffparameter.Atthispointtheaction(4)isjusttheoriginalSQEDactionandthecoefficients1/eZ(Mo),Z(Mo)arebareparameters.Forarbitrary/~thecoefficientfunctionsaredeterminedbynormalgraphsofperturbationtheoryconstructedstartingfromtheoriginalactionwiththefollowingconstraint-andthisisthemostcrucialpointforus-thattheintegrationdomainovermomentakinallloopsislimitedbythecondition/~"-7]'>+soz7-,(22)wherefortheZfactorwehaveZ(-~)=1-(1+½f)a°lnM°rr/~(23)NoticethatZdependsonthegaugeofthephotonfieldwhosepropagatorischosenintheform~=e2(_g~.+~--~-)--~'k~k~~1InthisapproximationF(/t)superficiallycoincideswithSw(/~)sincethephotonicmatrixelementof(Z-1)(~rp)*~,cpistobetakenintoaccountonlyinthetwo-looporder.Letusproceednowtoatwo-loopanalysis.Inthetwo-loopapproximationthecoefficientinfrontofFainSwisdeterminedbythediagramoffig.1.Letussingle 464M.A.Shifman,A.I.Vainshtein/AnomalypuzzleFig.1.Thetwo-loopcontributiontoSw(~)inscalarelectrodynamics.Thesolidline-thescalarparticlepropagatorintheexternalfield;wavyline-thephotonpropagator.Fig.2.Bycuttingoffthephotonlineinfig.1wearriveatthephotonpolarizationoperator/-/~,,.WeareinterestedinthecoefficientinfrontofF~aF~intheoperatorexpansionforH,,.outintegrationoverthevirtualphotonandperformitattheveryend.Then,beforethislastintegration,calculationofSw(/~)isequivalenttoacalculationofthephotonpolarizationoperatorinoneloop(fig.2).Morestrictly,oneneedstofindonlyonetermintheoperatorexpansionforH,.,namelyCF.BF.~.Thecoefficientofthistermisfiniteandwell-defined.Thenthelastintegrationoverthephotonmomentumkwillyieldalogarithmicintegralofthetypefd4k/k4whichcanbesimplycutofffromaboveatM0andfrombelowat/~.(Forfurtherdetailsseeref.[27].)Specifically,inthexrepresentation~(=)eff=-2IfMId4xi~.~,(x)(F2)()X(24);-~.(53)Hereallquantitiesarematricesinthecolourspace,(p~,).b=i(8.bov+sf:~bAc"""y]~=_GL-ab__facbGcTosimplifytheexpressionswehavescrappedthesubscript"ext"fortheexternalfields.Theformalismweexploitisexplicitlygaugeinvariantwithrespecttotheexternalfield.Asfarasthequantumfieldgaugeisconcerned,eq.(53)impliestheFeynmangauge,.~ggf..=-~(~,a,)1a2.Needlesstosay,thefinalanswershouldbeindependentofthequantumfieldgauge.Wewillreturntothisissuelater.Now,letusexpandthepropagatorineq.(53)inpowersofG/P2.ThezerothordertermsinGdropsoutbecauseofcontractionwith%~r8.Thesecond-orderterm111containstwoG'scontractedoveroneindex;hencethereisnowayofgettingtheonlystructure,q~,G,q3G,#~,determiningthelongitudinalpartofKz.Thus,weareleftwiththelinearinGterm:11=-8g2Trco,our(xlPr~Gz¢-~lx).(54) M.A.Shifman,A.I.Vainshtein/Anomalypuzzle475Insteadofdirectcomputationof(54)(whichpresentsnodifficulties,though)onecancomparethisexpressionwiththematrixelementofthespinoraxialcurrenta,,whoseanomalyiswell-known.Forthespinorcurrent1(at`)=-½ig2Trco,o~Trspin(x17t`Y5--Ix)1=-½ig2Trco,ourTrspin(XlYt`Y5~p2_~G,~#o,~Ix>"Asinthepreviouscaseonecanconvinceoneselfthatinthelongitudinalpartof(at`)onlythelinearinoGtermsurvives,11(55)Asimpleinspectionofeqs.(54)and(55)showsthepresenceinjustthesamepoleasin(at,>,(qJq2)GG,butwithanadditionalfactor4.Sincea'a<°3~at'>-4~r(G;~G~)ext'(561wegetfor~-<0t`K/.t>=~,-t`vt`t,lex,~"{'-wherewehaveaddedtheunittermfromtheclassicalpart.Eqs.(56)and(57)implyfortheGcomponentofW2(seeeq.(49)):T(G)a~T(G)a(TrW21G)=W2aextf~1+--(58)=Ic+2~-)"Now,keepinginmindsupersymmetry,weseethattheone-looppieceofrelation(46)isreproduced.Whatremainstobedoneistodemonstrateindependenceofthequantumfieldgauge.Inarbitrarygaugethepropagatorofthefielda~hastheform-@t,~(x,y)=(x[(P2g~,_2G~,~-~Pt`P~)-l[x>.(59) 476M.A.Shifman,A.LVainshtein/AnomalypuzzleOnecanreadilycheckthefollowingoperatorequalityPt,(p2g~.v-2G~++)=P2Pv-1-i~vayv.(60)Iftheexternalfieldisassumedtosatisfytheequationsofmotion,~vGy~=O,asrequiredintheexternalfieldmethod,thesecondtermonther.h.s,dropsout,andtheGreenfunctionisrepresentableintheclosedform:1]~1N.+(x,y)=(xlez-2G.~+1-,~P,,-~-iP,,Iy).ReturningnowtothematrixelementofK.(seeeq.(53))wewritethe~dependentpartasfollows2~~1(61)Formally,duetogaugeinvariancewithrespecttotheexternalfield,(x]p-4e~lx)mustbeproportionalto-@,Gv,,whichiszero.However,inthekinematicsconsidered(k12=k22=0)theexpression(xlP4pvlx)isill-definedintheinfrared(itcontainsl/k2).Forregularizationonecanintroduceaninfraredmassmtothequantumfielda~.Inotherwords,inallpropagatorsP2~(p2_m2)1.Itisassumedthatm2<>Mv).Withinthistechniquetherearenocorrectionsofsecondandhigherordersin0,a~.TheassertionasitstandsreferstotheamplitudeswithexternalmomentapintheintervalMF>>p>>Mv,i.e.itbearstheoperatorcharacterinourlanguage*.Inref.[20]wehaveextendedthetwo-limittechniquetosupersymmetryandfoundthatthesituationwith0~wasjustthesameaswitha~a~.Inotherwords,0~isexhaustedbyoneloopinthetwo-limitsense.Theanswer,however,didnotsatisfyussinceofmostinterestistheone-limitregularization(evolutiontop<0).Ind=4-e,apartfromW2^,thereexitsanotheroperatorgaugeinvariantwithrespecttotheexternalfield,I'I',whereI"isaconnection,andthedoublecaretaccordingto[21,22]denotesprojectiononthe"additional"edimensions.Theanswerforthetwo-loopdiagramobtainedin[21,22]reducestotheoperator(C/e)fdaxd40~2,andnottotheoperator(C/e)fd4xd20W2appearingintheone-loopgraph.ThentheauthorshaveusedthefactthatinSRDR~,2~,2=_eWZ.Accordingtothepicturedevelopedherethesolutionoftheanomalyproblemdoesnotrequireintroductionoftwoaxialcurrents,twooperatorsGG,etc.More-over,thetwocurrentsconsideredinref.[22]actuallydiffernotbyanultravioletconstant,butbyaninfraredsingularnon-localexpression.Asamanifestation,thedifferenceatAB-a~SSfromref.[22]couldnotbewritteninthelimite~0.Inourlanguagethesituationiseasilyexplainable:inessence,thetwo-loopcomputationofref.[22]isacomputationofthematrixelementoffd20W2withintheSRDRprocedure.Thematrixelementiscompletelysaturatedintheinfrareddomain.AsregardstotheissueofdifferentschemesfortheoperatorGG,themainpointisnotthedistinctionbetweenGGinthedifferentschemesbutthedistinctionbetweentheoperatoranditsmatrixelement.Thelatterisfixedunambiguously.Hereitwillbeinordertoexplaintowhichrenormalizationschemetheflfunctionsquotedineqs.(1),(2),(9)refer.OurdefinitionisclosetotheMOMscheme.Specifically,wefixthegaugecoupling[g2(/Q]forsomeexternalfieldmomentump-/zandexpressg2(~)intermsofthebarechargeandtheultravioletcut-off.Thus,wegetarelationbetweengoandM0.Nosubtractionsaremadeatintermediatestages.Thelatterpointseemstoexplainthedisagreementbetweenourthree-loopcoefficientineq.(2)andthatfoundin[41].Inconclusion,letusmentionthepaper[7]whichpresentstwoperturbativederivationsofeq.(1).Onederivationwasbasedonaninfraredregularizationina M.A.Shifman,A.LVainshtein/Anomalypuzzle485boxofafinitevolume.Althougheq.(7)fromref.[7]fortherelationbetween1/[a]and1/a0iscorrect,themotivationusedinitsderivation(seeeq.(6)in[7])is,literallyspeaking,unjustified.TheauthorsaregratefultoV.Chernyak,D.Kazakov,Y.Kogan,V.Novikov,A.Smilga,V.SokolovandV.Zakharovforusefuldiscussions.References[1]A.GolfandandE.Likhtman,ZhETFPis'ma13(1971)452;D.V.VolkovandV.P.Akulov,Phys.Lett.B46(1973)109;J.WessandB.Zumino,Nucl.Phys.B70(1974)39[2]B.Zumino,Nucl.Phys.B89(1975)535;P.West,Nucl.Phys.B106(1976)219;M.Grisaru,W.SiegelandM.Ro~ek,Nucl.Phys.B159(1979)429[3]P.S.Howe,K.S.StelleandP.S.West,Phys.Lett.124B(1983)55[4]K.WilsonandJ.Kogut,Phys.Reports12(1974)75[5]V.Novikov,M.Shifman,A.VainshteinandV.Zakharov,Nucl.Phys.B229(1983)381;preprintITEP-75(1985);Phys.Lett.B166(1986)329[6]M.Shifman,A.VainshteinandV.Zakharov,preprintITEP-89(1985);Phys.Lett.B166(1986)334[7]M.Shifman,A.VainshteinandV.Zakharov,preprintITEP-137(1985);Yad.Fiz.43(1986)1602[8]S.FerraraandB.Zumino,Nucl.Phys.B87(1975)207[9]M.Grisaru,inRecentdevelopmentsingravitation(CargeseLectures,1979)eds.M.Levy,S.Deser(Plenum,1979)p.130[10]S.L.AdlerandW.A.Bardeen,Phys.Rev.182(1969)1517[11]T.E.Clark,O.PiguetandK.Sibold,Nucl.Phys.B143(1978)445B172(1980),201;O.PiguetandK.Sibold,Nucl.Phys.B196(1982)428,447[12]D.R.T.Jones,Phys.Lett.123(1983)45[13]M.T.GrisaruandP.West,Nucl.Phys.B254(1985)249[14]A.Vainshtein,V.Zakharov,V.NovikovandM.Shifman,ZhETFPis'ma40(1984)161[15]P.Breitenlohner,D.MaisonandK.Stelle,Phys.Lett.134B(1984)63[16]D.R.T.JonesandL.Mezincescu,Phys.Lett.136B(1984)283,138B(1984)242[17]D.Espriu,OxfordUniv.preprint57/84(1984)[18]D.Kazakov,ZhETFPis'ma41(1985)272[19]D.R.T.Jones,L.MezincescuandP.West,Phys.Lett.151B(1985)219[20]V.Novikov,M.Shifman,A.VainshteinandV.Zakharov,Phys.Lett.157B(1985)169[21]M.T.Grisaru,B.MilewskiandD.Zanon,Phys.Lett.157B(1985)174[22]M.T.Grisaru,B.MilewskiandD.Zanon,Nucl.Phys.B266(1986)589[23]K.Konishi,Phys.Lett.135B(1984)439[24]T.E.Clark,O.PiguetandK.Sibold,Nucl.Phys.159B(1979)1[25]K.KonishiandK.Shizuya,PisaUniversitypreprintIFUP-TH-5/85(1985)[26]M.T.GrisaruandD.Zanon,Nucl.Phys.B252(1985)578;M.T.GrisaruandB.Milewski,Phys.Lett.155B(1985)357[27]M.ShifmanandA.Vainshtein,preprintITEP-146(1985);Yad.Fiz.44(1986)498[28]M.T.Grisaru,W.SiegelandH.Ro~ek,Nucl.Phys.B159(1979)429;S.J.Gates,M.T.Grisaru,M.Ro~ekandW.Siegel,Superspace(Benjamin,Reading,Mass.,1983)[29]A.D.DolgovandV.I.ZakharovB12(1971)68[30]G.'tHooft,inRecentdevelopmentsingaugetheories,eds.G.'tHooftetal.(Plenum,NY,1980)p.241[31]A.Belavin,A.Polyakov,A.ShwartzandY.Tyupkin,Phys.Lett.59B(1975)85[32]G.'tHooft,Phys.Rev.D14(1976)3432;L.S.Brown,R.D.Carlitz,D.B.CreamerandC.Lee,Phys.Rev.D17(1978)1583 486M.A.Shifman,A.LVainshtein/Anomalypuzzle[33]A.D'AddaandP.DiVecchia,Phys.Lett.73B(1978)162[34]G.'tHooft,Comm.Math.Phys.81(1981)267[35]E.Witten,Nucl.Phys.B202(1982)253[36]E.CohenandC.Gomez,Phys.Rev.Lett.52(1984)237[37]W.Fischler,H.P.Nilles,J.Polchinskietal.,Phys.Rev.Lett.47(1981)757[38]M.J.DuffandP.Nieuwenhuizen,Phys.Lett.94B(1980)179;W.Siegel,Phys.Lett.103B(1981)107;M.J.Duff,inSupergravity'81,eds.S.Ferrara,J.C.Taylor(CambridgeUniversityPress,1982)p.112[39]R.E.Kallosh,ZhETFPis'ma37(1983)503[40]M.T.Grisaru,N.K.Nielsen,W.SiegelandD.Zanon,Nucl.Phys.B247(1984)157[41]L.V.AvdeevandO.V.Tarasov,Phys.Lett.112B(1982)356