1: 

2: \documentclass[12pt]{article}

3: \usepackage[dvips]{graphicx}

4: \usepackage{latexsym}

5: \topmargin -20mm

6: \textwidth 160mm

7: \textheight 220mm

8: \evensidemargin 0mm

9: \oddsidemargin 0mm

10: \parskip=\medskipamount

11: \def\baselinestretch{1.2}

12: \arraycolsep 2pt

13: 

14: 

15: \def \const {{\rm const}}

16: \def \un{\underline}

17: %%

18: %

19: %Calligraphic Lettersq

20: \newcommand {\cA}{{\cal A}}

21: \newcommand {\cB}{{\cal B}}

22: \newcommand {\cC}{{\cal C}}

23: \newcommand {\cD}{{\cal D}}

24: \newcommand {\cE}{{\cal E}}

25: \newcommand {\cF}{{\cal F}}

26: \newcommand {\cG}{{\cal G}}

27: \newcommand {\cH}{{\cal H}}

28: \newcommand {\cI}{{\cal I}}

29: \newcommand {\cJ}{{\cal J}}

30: \newcommand {\cK}{{\cal K}}

31: \newcommand {\cL}{{\cal L}}

32: \newcommand {\cM}{{\cal M}}

33: \newcommand {\cN}{{\cal N}}

34: \newcommand {\cO}{{\cal O}}

35: \newcommand {\cP}{{\cal P}}

36: \newcommand {\cQ}{{\cal Q}}

37: \newcommand {\cR}{{\cal R}}

38: \newcommand {\cS}{{\cal S}}

39: \newcommand {\cT}{{\cal T}}

40: \newcommand {\cU}{{\cal U}}

41: \newcommand {\cV}{{\cal V}}

42: \newcommand {\cW}{{\cal W}}

43: \newcommand {\cX}{{\cal X}}

44: \newcommand {\cY}{{\cal Y}}

45: \newcommand {\cZ}{{\cal Z}}

46: %

47: %Boldface Letters

48: \newcommand{\bA}{{\bf A}}

49: \newcommand{\bB}{{\bf B}}

50: \newcommand{\bC}{{\bf C}}

51: \newcommand{\bD}{{\bf D}}

52: \newcommand{\bE}{{\bf E}}

53: \newcommand{\bF}{{\bf F}}

54: \newcommand{\bG}{{\bf G}}

55: \newcommand{\bH}{{\bf H}}

56: \newcommand{\bI}{{\bf I}}

57: \newcommand{\bJ}{{\bf J}}

58: \newcommand{\bK}{{\bf K}}

59: \newcommand{\bL}{{\bf L}}

60: \newcommand{\bM}{{\bf M}}

61: \newcommand{\bN}{{\bf N}}

62: \newcommand{\bO}{{\bf O}}

63: \newcommand{\bP}{{\bf P}}

64: \newcommand{\bQ}{{\bf Q}}

65: \newcommand{\bR}{{\bf R}}

66: \newcommand{\bS}{{\bf S}}

67: \newcommand{\bT}{{\bf T}}

68: \newcommand{\bU}{{\bf U}}

69: \newcommand{\bV}{{\bf V}}

70: \newcommand{\bW}{{\bf W}}

71: \newcommand{\bX}{{\bf X}}

72: \newcommand{\bY}{{\bf Y}}

73: \newcommand{\bZ}{{\bf Z}}

74: %

75: \def\a{\alpha}

76: \def \bi{\bibitem}

77: \def \ci{\cite}

78: \def\b{\beta}

79: \def\c{\chi}

80: \def\d{\delta}

81: \def\e{\epsilon}

82: \def\f{\phi}

83: \def\g{\gamma}

84: \def\G{\Gamma}

85: \def\i{\iota}

86: \def\j{\psi}

87: \def\k{\kappa}

88: \def\l{\lambda}

89: \def\m{\mu}

90: \def\n{\nu}

91: \def\o{\omega}

92: \def\p{\pi}

93: \def\q{\theta}

94: \def\r{\rho}

95: \def\s{\sigma}

96: \def\t{\tau}

97: \def\u{\upsilon}

98: \def\x{\xi}

99: \def\z{\zeta}

100: \def\D{\Delta}

101: \def\F{\Phi}

102: \def\J{\Psi}

103: \def\L{\Lambda}

104: \def\O{\Omega}

105: \def\P{\Pi}

106: \def\Q{\Theta}

107: \def\S{\Sigma}

108: \def\U{\Upsilon}

109: \def\X{\Xi}

110: \def\tr{{\rm tr}}

111: %%%%%%%%%%%%%%%%%%%%%%%%%%%

112: \newcommand{\ad}{{\dot{\alpha}}}                           %new

113: \newcommand{\bd}{{\dot{\beta}}}                            %new

114: \newcommand{\ve}{\varepsilon}                            %new

115: \newcommand{\cDB}{{\bar\cD}}                            %new

116: 

117: \newcommand{\ab}{{\a\b}}

118: \newcommand{\abd}{{\ad\bd}}

119: \renewcommand{\aa}{{\a\ad}}

120: \newcommand{\bb}{{\b\bd}}

121: \newcommand{\pa}{\partial}                           %new

122: \newcommand{\hf}{\frac12}

123: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

124: \newcommand{\abar}{\bar{a}}

125: \newcommand{\bbar}{\bar{b}}

126: \newcommand{\cbar}{\bar{c}}

127: %

128: \newcommand{\vf}{\varphi}

129: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

130: \newcommand{\sect}[1]{\setcounter{equation}{0}\section{#1}}

131: \renewcommand{\theequation}{\thesection.\arabic{equation}}

132: %

133: \newcommand{\be}{\begin{equation}}

134: \newcommand{\ee}{\end{equation}}

135: \newcommand{\bea}{\begin{eqnarray}}

136: \newcommand{\eea}{\end{eqnarray}}

137: \newcommand{\non}{\nonumber}

138: %

139: \newcommand{\1}{\underline{1}}

140: \newcommand{\2}{\underline{2}}

141: \newcommand{\au}{\underline{a}}

142: \newcommand{\bu}{\underline{b}}

143: \newcommand{\cu}{\underline{c}}

144: \newcommand{\du}{\underline{d}}

145: %

146: 

147: %%%%%%%%%%%%%%%%

148: %%%%%%%%%%%%%%%%

149: \begin{document}

150: %%%%%%%%%%%%%%%%

151: %%%%%%%%%%%%%%%%

152: 

153: \begin{titlepage}

154: \thispagestyle{empty}

155: 

156: \begin{flushright}

157: hep-th/0310025 \\

158: October, 2003 \\

159: \end{flushright}

160: \vspace{5mm}

161: 

162: \begin{center}

163: {\Large \bf

164: On the two-loop four-derivative quantum corrections\\

165: in 4D N = 2 superconformal field theories}

166: \end{center}

167: %\vspace{3mm}

168: 

169: \begin{center}

170: {\large S. M. Kuzenko and   I. N. McArthur}\\

171: \vspace{2mm}

172: 

173: \footnotesize{

174: {\it School of Physics, The University of Western Australia\\

175: Crawley, W.A. 6009, Australia}

176: } \\

177: {\tt  kuzenko@cyllene.uwa.edu.au},~

178: {\tt mcarthur@physics.uwa.edu.au} \\

179: \end{center}

180: \vspace{5mm}

181: 

182: \begin{abstract}

183: \baselineskip=14pt

184: In  $\cN=2, 4$ superconformal field theories in 

185: four space-time dimensions, 

186: the quantum corrections with four derivatives are believed 

187: to be severely constrained by non-renormalization theorems.  

188: The strongest of these is the conjecture formulated 

189: by Dine and Seiberg  in hep-th/9705057 that such terms 

190: are generated only at one loop. In this note, using 

191: the background field formulation 

192: in $\cN=1$ superspace,  we test the Dine-Seiberg proposal 

193: by comparing  the two-loop $F^4$ quantum corrections 

194: in two different superconformal 

195: theories with the same gauge group $SU(N)$: 

196: (i) $\cN=4$ SYM (i.e. $\cN=2$ SYM with a single adjoint 

197: hypermultiplet); (ii) $\cN=2$ SYM with $2N$ hypermultiplets

198: in the fundamental. According to the Dine-Seiberg conjecture, 

199: these theories should yield identical two-loop $F^4$ contributions

200: from all the supergraphs  involving quantum hypermultiplets, 

201: since the pure $\cN=2$ SYM and ghost sectors are identical

202: provided the same gauge conditions are chosen.

203: We explicitly evaluate the relevant two-loop  supergraphs and 

204: observe that the $F^4$ corrections generated  

205: have different large $N$ behaviour in the two theories under 

206: consideration. Our results are in  conflict 

207: with the Dine-Seiberg conjecture. 

208: \end{abstract}

209: 

210: 

211: 

212: \vfill

213: \end{titlepage}

214: 

215: \newpage

216: \setcounter{page}{1}

217: 

218: \renewcommand{\thefootnote}{\arabic{footnote}}

219: \setcounter{footnote}{0}

220: %%%%%%%%%%%%%%%%%%%%%%%%%%%

221: \sect{Introduction}

222: Some time ago, we developed 

223: a manifestly covariant approach 

224: for evaluating multi-loop quantum corrections

225: to low-energy effective actions within the 

226: background field formulation \cite{KM}. 

227: This approach is  applicable to ordinary gauge theories

228: and  to supersymmetric Yang-Mills theories

229: formulated in superspace. Its power is not restricted

230: to computing just the counterterms --  it is well suited

231: for deriving finite quantum corrections in 

232: the framework of the derivative 

233: expansion. 

234: As a simple application of the techniques 

235: developed in \cite{KM}, 

236: we have recently derived  \cite{KM2}

237: the two-loop (Euler-Heisenberg-type) 

238: effective action for $\cN=2$ supersymmetric QED 

239: formulated in  $\cN=1$ superspace. 

240: 

241: The work of \cite{KM2} has brought a surprising outcome

242: regarding one particular conclusion drawn 

243: in \cite{BKO} on the basis of the background field 

244: formulation in $\cN=2$ harmonic superspace \cite{BBKO}.

245: According to \cite{BKO},  no super $F^4$ (four-derivative) 

246: quantum corrections occur at two loops in generic $\cN=2$ 

247: super Yang-Mills theories on the Coulomb branch, 

248: in particular in $\cN=2$ SQED.

249: However, by explicit calculation carried out in \cite{KM2}, 

250: it was shown that  a non-vanishing 

251: two-loop $F^4$ correction does occur in $\cN=2$ SQED. 

252: It was also shown in \cite{KM2}

253: that  the analysis in  \cite{BKO} 

254: contained a subtle loophole related

255: to the intricate structure of harmonic supergraphs. 

256: A more careful treatment of two-loop harmonic supergraphs

257: given in \cite{KM2} leads to the same non-zero $F^4$  term 

258:  in $\cN=2$ SQED at two loops as that derived 

259: using the $\cN=1$ superfield formalism.

260: 

261: The work of \cite{BKO} provided perturbative two-loop 

262: support  to the famous Dine-Seiberg 

263: non-renormalization conjecture

264: \cite{DS} that the 

265: $\cN=2$ supersymmetric 

266: four-derivative term\footnote{The 

267: functional (\ref{N=2F4}) was originally introduced 

268: in \cite{dWGR}. It is a unique $\cN=2$ superconformal 

269: invariant  in the family of non-holomorphic actions 

270: of the form $\int {\rm d}^4 x {\rm d}^8 \q \,H(W, {\bar W})$

271: introduced for the first time in \cite{Hen}.

272: More general (higher-derivative) superconformal 

273: invariants of the  $\cN=2$ Abelian vector multiplet   

274:  were given in \cite{BKT}.} 

275: \bea

276:  \int {\rm d}^4 x {\rm d}^8 \q \,

277: \ln W \,\ln {\bar W} 

278: \label{N=2F4}

279: \eea

280: is one-loop exact on the Coulomb branch of  $\cN=2,4$ 

281: superconformal theories.\footnote{The one-loop 

282: $F^4$ quantum corrections in $\cN=2,4$ 

283: SQFTs were computed in 

284: \cite{GR,BK2,LvU}.} 

285: It is known that the Dine-Seiberg  conjecture is  

286: well supported by non-perturbative considerations

287: \cite{DKMSW,BFMT}.

288: But since the two-loop $F^4$ conclusion of \cite{BKO}

289: is no longer valid, it seems important 

290: to carry out  an  independent calculation of the two-loop $F^4$ 

291: quantum corrections in $\cN=2$ superconformal theories.  

292: It is the aim of the present note to provide 

293: such a calculation.  As will be demonstrated below, 

294: the Dine-Seiberg  conjecture

295: is not fully supported  at the  perturbative level.

296: 

297: To test the Dine-Seiberg conjecture,

298: we consider 

299: two different $\cN=2$ superconformal 

300: theories with the same gauge group $SU(N)$: 

301: (i) $\cN=4$ SYM or, equivalently,  

302: $\cN=2$ SYM with a single adjoint hypermultiplet; 

303: (ii) $\cN=2$ SYM with $2N$ hypermultiplets

304: in the fundamental. At the quantum level, 

305: with the same gauge conditions chosen, 

306: these theories are identical in the pure $\cN=2$ SYM 

307: and ghost sectors. The difference between them 

308: occurs only in the sector involving quantum 

309: hypermultiplets. If the Dine-Seiberg conjecture holds, 

310: then since the pure $\cN=2$ SYM 

311: and ghost sectors are identical, 

312: these theories should yield identical two-loop $F^4$ contributions

313: from all the supergraphs  with quantum hypermultiplets.

314: However, as will be shown below by  direct calculations,

315: the relevant two-loop $F^4$ contributions have 

316: different large $N$ behaviour in the theories 

317: under  consideration.\footnote{To test the Dine-Seiberg 

318: conjecture, we do not need the two-loop $F^4$ contribution

319: from the pure $\cN=2$ SYM 

320: and ghost sectors. It will be discussed in a separate paper.}  

321: 

322:  ${}$From the point of view of $\cN=1$ supersymmetry,

323: the chiral superfield strength $W$ of the 

324: $\cN=2$ vector multiplet  is known 

325: to consist of a chiral  scalar $\f$ and a constrained chiral 

326: spinor $W_\a$, the latter being 

327: the $\cN=1$ vector multiplet field strength. 

328: When reduced to $\cN=1$ superspace, 

329: the functional (\ref{N=2F4}) is given by a sum of several terms, 

330: of which the leading (in a derivative expansion) term is 

331: \be

332: \U = \int {\rm d}^8 z \, \frac{ W^\a W_\a

333: {\bar W}_\ad {\bar W}^\ad  }{\f^2{\bar \f}^2 }~,

334: \label{F4}

335: \ee 

336: while the other terms involve derivatives of $\f$ and $\bar \f$.

337: If one uses the $\cN=1$ superspace formulation 

338: for $\cN=2$ superconformal field theories, it is 

339: typically sufficient to compute quantum corrections 

340: of the form (\ref{F4})  in order to restore their

341: $\cN=2$ completion (\ref{N=2F4}).

342: 

343: This note is organized as follows. 

344: Section 2 contains the necessary setup

345: regarding $\cN=2$ superconformal 

346: field theories and their background field

347: quantization (for supersymmetric 't Hooft gauge)

348:  in $\cN=1$ superspace. In section 3 we work out a useful 

349: functional representation for two-loop supergraphs

350: with quantum hypermultiplets.

351: In section 4 we describe,  following \cite{KM,KM2}, 

352: the exact superpropagators in a special $\cN=2$ 

353: vector multiplet background which is 

354: extremely simple but perfectly suitable for

355: computing quantum corrections of the form (\ref{F4}).

356: Sections 5 and 6 form the (somewhat technical) core of this paper. 

357: In section 5 we evaluate the two-loop $F^4$ corrections

358: in $\cN=2$ SYM with $2N$ hypermultiplets

359: in the fundamental. This consideration is extended in section 

360: 6 to the case of $\cN=4$ SYM. 

361: Finally, in section 7 we compare the two-loop corrections 

362: in the large $N$ limit for the two theories being 

363: studied.  Some aspects of the cancellation of divergences 

364: are  discussed  in the appendix. 

365: 

366: 

367: \sect{\mbox{$\cN = 2$} SYM setup}

368: 

369: The classical action of an $\cN=2$ superconformal 

370: field theory, 

371: $S_{\rm SCFT} = S_{\rm vector} + S_{\rm hyper}$,

372: consists of two parts: (i) the pure $\cN=2$ SYM action 

373: \bea

374: S_{\rm vector} &=& \frac{1}{g^2}\,{\rm tr}_{\rm F}

375: \Big( \int {\rm d}^8 z \, \F^\dagger \F

376: +  \int {\rm d}^6 z \, \cW^\a \cW_\a 

377: \Big)~;

378: \label{n=2pure-sym}

379: \eea

380:  (ii) the hypermultiplet action

381: \bea 

382: S_{\rm hyper} = 

383:  \int {\rm d}^8 z \, \Big( \cQ^\dagger  \,\cQ 

384: +  \tilde{\cQ}^\dagger   \tilde{\cQ} \Big)

385:  - {\rm i}  \int {\rm d}^6 z \, \tilde{\cQ}^{\rm T} \F \cQ 

386: + {\rm i}  \int {\rm d}^6 {\bar z}\,  \cQ^\dagger 

387: \F^\dagger \overline{\tilde{\cQ}} ~.

388: \label{hyper}

389: \eea

390: Here $\F$, $Q$ and $\tilde{Q}$ are {\it covariantly chiral}

391: superfields which transform, respectively, in the following 

392: representations of the gauge group:

393: (1) the adjoint; (2) a  representation $\rm R$; and

394: (3) its conjugate  ${\rm R}_{\rm c}$. 

395: The covariantly chiral superfield strength $\cW_\a$ 

396: is associated with  the gauge  covariant derivatives

397: \be

398: \cD_A = (\cD_a, \cD_\a , {\bar \cD}^\ad ) 

399: = D_A +{\rm i}\, \G_A~, \qquad 

400: \G_A = \G^\m_A (z) T_\m~, \qquad 

401: (T_\m)^\dagger = T_\m~, 

402: \ee

403: where $D_A$ are the flat covariant 

404: derivatives\footnote{Our $\cN=1$ notation 

405: and conventions correspond to \cite{BK}.}, 

406: and $\G_A$ the superfield connection taking its values 

407: in the Lie algebra of the gauge group.  

408: The   gauge covariant derivatives satisfy the following algebra:

409: \bea

410: & \{ \cD_\a , \cD_\b \} 

411: = \{ {\bar \cD}_\ad , {\bar \cD}_\bd \} =0~, \qquad 

412: \{ \cD_\a , {\bar \cD}_\bd \} = - 2{\rm i} \, \cD_{\a \bd}~, \non \\

413: & [ \cD_\a , \cD_{\b \bd}] = 2 {\rm i} \ve_{\a \b}\,{\bar \cW}_\bd ~, 

414: \qquad 

415: [{\bar \cD}_\ad , \cD_{\b \bd}] = 2{\rm i} \ve_{\ad \bd}\,\cW_\b ~ , 

416: \non \\

417: & [ \cD_{\a \ad}, \cD_{\b \bd} ] 

418: = - \ve_{\a \b}\, {\bar \cD}_\ad {\bar \cW}_\bd 

419: -\ve_{\ad \bd} \,\cD_\a \cW_\b~. 

420: \label{N=1cov-der-al}

421: \eea

422: The spinor field strengths $\cW_\a$ and ${\bar \cW}_\ad$ 

423: obey the Bianchi  identities

424: \be

425: {\bar \cD}_\ad \cW_\a =0~, \qquad 

426: \cD^\a \cW_\a = {\bar \cD}_\ad {\bar \cW}^\ad~.

427: \ee

428: The condition under which the $\cN=2$ theory 

429: is finite is  

430: \be

431: {\rm tr}_{\rm Ad} \,\F^2 ~=~ {\rm tr}_{\rm R} \, \F^2~.

432: \label{finita}

433: \ee

434: It is assumed that in the action (\ref{n=2pure-sym}) 

435:  the superfields $\F$ and $\cW_\a$ 

436: are given in the fundamental (or defining) 

437: representation of the gauge group, 

438: with  the corresponding generators

439: normalized such that  

440: ${\rm tr}_{\rm F} \,(T_\m\, T_\n) = \d_{\m \n}$.

441: 

442: To quantize the theory, we will use the $\cN=1$ 

443: background field formulation \cite{GGRS} and 

444: split the dynamical variables into background 

445: and quantum ones, 

446: \bea

447:  \F ~ \to ~ \F +\vf ~, \qquad \cQ ~ &\to & ~ \cQ +q ~, 

448: \qquad 

449: \tilde{\cQ} ~ \to ~ \tilde{\cQ}+ \tilde{q} ~, \non \\

450:   \cD_\a ~ \to ~ {\rm e}^{-v} \, \cD_\a \, {\rm e}^v~, 

451: \quad && \quad 

452: {\bar \cD}_\ad ~ \to ~ {\bar \cD}_\ad~,

453: \eea

454: with lower-case letters used for 

455: the quantum superfields. 

456: In this paper, we are not interested in the 

457: dependence of the effective action on 

458: the hypermultiplet superfields, 

459: and therefore we set $\cQ = \tilde{\cQ} =0$

460: in what follows. 

461: After the background-quantum splitting, 

462: the action (\ref{n=2pure-sym}) turns into

463: \bea

464: S_{\rm vector} &=& \frac{1}{g^2}\,{\rm tr}_{\rm F}

465: \Big( \int {\rm d}^8 z \, (\F +\vf)^\dagger \,

466: {\rm e}^v \, (\F +\vf) \, {\rm e}^{-v}

467: +  \int {\rm d}^6 z \, \bW^\a \bW_\a 

468: \Big)~,

469: \label{bqs-vm}

470: \eea

471: where 

472: \bea

473: \bW_\a &=& - {1\over 8} {\bar \cD}^2 \Big( 

474: {\rm e}^{-v}\, \cD_\a \,{\rm e}^{v} \cdot 1 \Big)

475: = \cW_\a - {1\over 8} {\bar \cD}^2 \Big( 

476: \cD_\a v- \hf [v, \cD_\a v] \Big) ~+~ O(v^3)~. 

477: \eea 

478: The hypermultiplet  action (\ref{hyper}) takes the form 

479: \bea 

480: S_{\rm hyper} = 

481:  \int {\rm d}^8 z \, \Big( q^\dagger \, {\rm e}^v \,q 

482: +  \tilde{q}^\dagger \, {\rm e}^{-v^{\rm T}} \, \tilde{q} \Big)

483:  - {\rm i}  \int {\rm d}^6 z \, \tilde{q}^{\rm T} 

484: (\F + \vf)  q 

485: + {\rm i}  \int {\rm d}^6 {\bar z}\,  q^\dagger 

486: (\F +\vf)^\dagger \overline{\tilde{q}} ~.

487: \label{bqs-hyper}

488: \eea

489: 

490: It is advantageous to use 

491: $\cN=1$ supersymmetric 't Hooft gauge 

492: (a special case of the supersymmetric $R_\x$-gauge

493: introduced in \cite{OW} and further developed in \cite{BBP})

494: which is specified by the nonlocal gauge conditions 

495: \bea

496: -4 \chi \,&= &{\bar \cD}^2 v +

497: [\F, ({\Box_+})^{-1} {\bar \cD}^2 \vf^\dagger ] 

498: = {\bar \cD}^2 v +

499: [\F, {\bar \cD}^2  ({\Box_-})^{-1}  \vf^\dagger ] 

500: ~, \non \\

501: -4 \chi^\dagger  &= &\cD^2 v -

502: [\F^\dagger , ({\Box_-})^{-1} \cD^2 \vf ] 

503: = \cD^2 v -

504: [\F^\dagger , \cD^2  ({\Box_+})^{-1}  \vf ] ~.

505: \eea

506: Here the covariantly chiral d'Alembertian, $\Box_+$,

507: is defined by

508: \bea 

509: \Box_+ &=& \cD^a \cD_a - \cW^\a \cD_\a -\hf \, (\cD^\a \cW_\a)~, 

510: \quad

511: \Box_+ \J = {1 \over 16} \, {\bar \cD}^2 \cD^2 \J ~, \quad 

512: {\bar \cD}_\ad \J =0~,

513: \eea

514: for a covariantly chiral superfield $\J$.

515: Similarly, the  covariantly antichiral 

516: d'Alembertian, $\Box_-$, is defined by

517: \bea 

518: \Box_- &=& \cD^a \cD_a + {\bar \cW}_\ad {\bar \cD}^\ad 

519: +\hf \, ({\bar \cD}_\ad  {\bar \cW}^\ad)~, 

520: \quad

521: \Box_- {\bar \J} = {1 \over 16} \, \cD^2 {\bar \cD}^2  {\bar \J} ~, 

522: \quad  \cD_\a {\bar \J} =0~,

523: \eea

524: for a covariantly antichiral superfield $\bar \J$.

525: The gauge-fixing functional\footnote{In this paper, 

526: the explicit structure of the ghost sector is not 

527: required.}  is

528: \be

529: S_{\rm GF} = -\frac{1}{g^2}\,{\rm tr}_{\rm F}

530:  \int {\rm d}^8 z \, \chi^\dagger \,\chi~.

531: \ee

532: The quantum quadratic part of 

533: $S_{\rm vector} + S_{\rm GF}$ is 

534: \bea 

535: S^{(2)}_{\rm vector} + S_{\rm GF} 

536:  &=&\, ~ \frac{1}{g^2}\,{\rm tr}_{\rm F}

537:  \int {\rm d}^8 z \,

538: \Big( \vf^\dagger \,\vf 

539: - [\F^\dagger , [\F, \vf^\dagger ]] \, 

540: \frac{1}{\Box_+} \vf \Big) \non \\

541: &-&\frac{1}{2g^2}\,{\rm tr}_{\rm F} \int {\rm d}^8 z \,

542: v \, \Big(  \Box_{\rm v}  v -[\F^\dagger, [\F, v]] \Big) 

543: ~+~ \dots 

544: \label{quad-prel}

545: \eea 

546: where the dots stand for the terms with derivatives

547: of the background (anti)chiral superfields 

548:  $\F^\dagger$ and $\F$.

549: The vector d'Alembertian,

550: $\Box_{\rm v}$,  is defined by

551: \bea 

552: {\Box}_{\rm v} 

553: &=& \cD^a \cD_a - \cW^\a \cD_\a +{\bar \cW}_\ad {\bar \cD}^\ad \\

554: &=& -\frac{1}{8} \cD^\a {\bar \cD}^2 \cD_\a 

555: +{1 \over 16} \{ \cD^2 , {\bar \cD}^2 \} 

556: -\cW^\a \cD_\a -\hf  (\cD^\a \cW_\a) \non \\

557: &=& 

558:  -\frac{1}{8} {\bar \cD}_\ad \cD^2 {\bar \cD}^\ad 

559: +{1 \over 16} \{ \cD^2 , {\bar \cD}^2 \} 

560: +{\bar \cW}_\ad {\bar \cD}^\ad +\hf({\bar \cD}_\ad {\bar \cW}^\ad ) ~.

561: \non 

562: \eea

563: The quantum quadratic part of $S_{\rm hyper} $ is

564: \bea 

565: S^{(2)}_{\rm hyper} = 

566:  \int {\rm d}^8 z \, \Big( q^\dagger  \,q 

567: +  \tilde{q}^\dagger  \, \tilde{q} \Big)

568:  +  \int {\rm d}^6 z \, \tilde{q}^{\rm T}\, \cM_{\rm R}  \,q 

569: +   \int {\rm d}^6 {\bar z}\,  q^\dagger \,

570: \cM_{\rm R}^\dagger \, \overline{\tilde{q}} ~.

571: \label{hyper-quad}

572: \eea

573: Here the operator $\cM$ is defined by 

574: \bea

575: \cM_{\rm D} \,\S &=& -{\rm i}\, \F\, \S~,

576: \eea

577: for a superfield $\S$ transforming 

578: in some representation D of the gauge group.

579: 

580: The background superfields will be  chosen to form 

581: a special on-shell $\cN=2$ vector multiplet in the Cartan 

582: subalgebra of the gauge group:

583: \be

584: [\F , {\bar \F} ] ~= ~ \cD^\a \cW_\a = 0~, 

585: \qquad \cD_\a \F ~=~0~.

586: \ee

587: Such a background configuration is convenient 

588: for computing those corrections to the effective action 

589: which do not contain derivatives of $\F$ and $\F^\dagger$.

590: Now, the action (\ref{quad-prel}) becomes

591: \bea 

592: S^{(2)}_{\rm vector} + S_{\rm GF} 

593:  &=& \frac{1}{g^2}\,{\rm tr}_{\rm F}

594:  \int {\rm d}^8 z \,

595: \Big( 

596: \vf^\dagger 

597: \frac{1}{\Box_+} 

598: (\Box_+ - |\cM_{\rm Ad}|^2 ) \vf 

599: - \hf 

600:  v( \Box_{\rm v} - |\cM_{\rm Ad}|^2) v \Big)~.

601: \label{vector-quad}

602: \eea

603: 

604: The Feynman propagators 

605: associated with the actions

606: (\ref{vector-quad}) and (\ref{hyper-quad}) 

607: can be expressed via a single Green's function

608: in  different representations of the gauge group.

609: Such a Green's function,

610: $G^{({\rm D})} (z,z')$,  originates 

611: in the following auxiliary  model

612: \be

613: S^{({\rm D})} =

614:  \int {\rm d}^8 z \,

615: \S^\dagger

616: ( \Box_{\rm v} - |\cM_{\rm D}|^2) \S~,

617: \label{S-action}

618: \ee

619: which describes the dynamics of an unconstrained complex 

620: superfield $\S$

621: transforming in some representation D

622: of the gauge group. 

623: The relevant Feynman propagator reads

624: \be

625: G^{({\rm D})}(z,z') = 

626: {\rm i}\,\langle 0| T \, \Big(\S(z)\,  \S^\dagger (z') \Big) |0\rangle

627: \equiv  {\rm i}\,\langle \S(z) \, \S^\dagger (z')\rangle

628: \ee

629: and satisfies the equation

630: \be

631: \Big(\Box_{\rm v} - |\cM_{\rm D}|^2  \Big) 

632: \, G^{({\rm D})}(z,z') 

633: = - {\bf 1}\,\d^8 (z-z')~.

634: \label{green}

635: \ee

636: The Feynman propagators in the model (\ref{vector-quad})

637: are

638: \bea

639: { {\rm i} \over g^2} \, \langle  v (z)\, v^{\rm T} (z') \rangle &=& 

640: -   G^{({\rm Ad})}(z,z') ~, \non \\

641: { {\rm i} \over g^2} \, \langle  \vf (z)\,  \vf^\dagger (z') \rangle &=& 

642: {1 \over 16} {\bar \cD}^2 \cD'^2\, 

643: G^{({\rm Ad})}(z,z') ~, \\

644:  \langle  \vf (z)\, \vf^{\rm T} (z') \rangle &=& 

645: \langle  {\bar \vf} (z)\, \vf^\dagger (z') \rangle

646: =0~. \non 

647: \eea

648: It is understood here that $v$ and $\vf$ 

649: are column-vectors, and not matrices as in the

650: preceding consideration.

651: To formulate the Feynman propagators 

652: in the model (\ref{hyper-quad}), it is useful 

653: to introduce the notation 

654: \bea

655: {\bf q} = 

656: \left(

657: \begin{array}{c}

658: q\\

659: \tilde{q}  

660: \end{array}

661: \right)~, \qquad 

662: {\bf q}^\dagger = \Big( q^\dagger, \,\tilde{q}^\dagger \Big)~.  

663: \eea

664: Then, the Feynman propagators read

665: \bea

666:  {\rm i}  \, \langle  {\bf q}  (z)\,  {\bf q} ^\dagger (z') \rangle &=& 

667: {1 \over 16} {\bar \cD}^2 \cD'^2\, 

668: G^{({\rm R} \oplus {\rm R}_{\rm c})}(z,z') ~, \non \\

669: {\rm i}  \, \langle  q  (z)\,  \tilde{q} ^{\rm T} (z') \rangle &=& 

670: \cM_{\rm R}^\dagger \,  

671: G^{({\rm R})}_+(z,z') 

672: ~, \\

673: {\rm i}  \, \langle  

674: \overline{\tilde{q}}  (z) \, q ^\dagger (z') \rangle &=&  

675: \cM_{\rm R} \, 

676: G^{({\rm R})}_-(z,z') ~,

677: \non  

678: \eea

679: where the covariantly chiral ($G_+$) and antichiral

680: ($G_-$) Green's functions are  related to $G$ as 

681: follows:

682: \bea

683: G_+(z,z') &=& 

684: -{1 \over 4} {\bar \cD}^2 G(z,z') 

685: = -{1 \over 4} {\bar \cD}'^2 G(z,z') ~,\non \\

686: G_-(z,z') &=& 

687: -{1 \over 4} \cD^2 G(z,z') 

688: = -{1 \over 4}\cD'^2 G(z,z') ~.

689: \label{chiral}

690: \eea

691: 

692: \sect{Functional representation for two-loop supergraphs

693: with quantum hypermultiplets}

694: 

695: The interactions for the quantum hypermuliplets are: 

696: \bea

697: && S_{\rm int} = \int {\rm d}^8 z \,  v^\m

698: {\bf q}^\dagger \,\cT_\m \,

699: {\bf q}

700: + \hf  \int {\rm d}^8 z \,  v^\m v^\n

701: {\bf q}^\dagger \,\cT_\m \cT_\n \,

702: {\bf q} \non \\

703: && \quad - { {\rm i} \over 2} \int {\rm d}^6 z \, \vf^\m \,

704: {\bf q}^{\rm T} 

705: \left(

706: \begin{array}{c c}

707: 0 &  T_\m{}^{\rm T} \\

708: T_\m& 0  

709: \end{array}

710: \right)  

711: {\bf q} 

712: + { {\rm i} \over 2} \int {\rm d}^6 {\bar z} \, 

713: {\bar \vf}^\m \,

714: {\bf q}^\dagger

715: \left(

716: \begin{array}{c c}

717: 0 &  T_\m{}^{\rm T} \\

718: T_\m& 0  

719: \end{array}

720: \right)  

721: {\bar {\bf q} }

722: ~,

723: \eea

724: where 

725: \bea 

726: \cT_\m = \left(

727: \begin{array}{c c}

728: T_\m & 0 \\

729: 0& -T_\m{}^{\rm T}  

730: \end{array}

731: \right)

732: \eea

733: are the generators of the representation 

734: ${\rm R}\oplus {\rm R}_{\rm c}$.

735: 

736: There are four two-loop supergraphs  

737: with quantum hypermultiplets, 

738: and they are depicted in Figures 1--4.

739: \begin{figure}[!htb]

740: \begin{center}

741: \includegraphics{DSfig1}

742: \caption{Two-loop supergraph I}

743: \end{center}

744: \end{figure}

745: \begin{figure}[!htb]

746: \begin{center}

747: \includegraphics{DSfig2}

748: \caption{Two-loop supergraph II}

749: \end{center}

750: \end{figure}

751: 

752: The  contributions from the first two supergraphs 

753: can be combined in the form 

754: \bea

755: \G_{\rm I +II} &=& -{{\rm i} \over 2^9 } \int {\rm d}^8 z 

756: \int {\rm d}^8 z' \,  \langle  v^\m (z) \,v^\n (z') \rangle  \non \\

757: &\times &

758: \, {\rm  tr} \left\{ \cT_\m

759: \Big( {\bar \cD}^2 \cD^2 \, 

760: G^{({\rm R} \oplus {\rm R}_{\rm c})}(z,z') \Big) \, \cT_\n

761:  [{\bar \cD}'^2, \cD'^2 ] \,

762: G^{({\rm R} \oplus {\rm R}_{\rm c})}(z',z) \right\}~,

763: \label{I+II}

764: \eea

765: where we have used the identities \cite{KM2}

766: \be

767: {\bar \cD}^2 \, G(z,z') = {\bar \cD}'^2 \, G(z,z') ~, \qquad 

768: \cD^2 \, G(z,z') = \cD'^2 \, G(z,z')~.

769: \label{transpar}

770: \ee

771: As in  \cite{KM2}, the above expression can be

772: considerably simplified.

773: The representation ${\rm R} \oplus {\rm R}_{\rm c}$ is real, 

774: \bea

775: -\cT_\m{}^{\rm T} = \s_1 \, \cT_\m \, \s_1~, 

776: \qquad 

777: \s_1= \left(

778: \begin{array}{c c}

779: 0 \, &\, {\bf 1} \\

780: {\bf 1} &0  

781: \end{array}

782: \right)~.

783: \label{real-1}

784: \eea

785: On the same grounds, the relevant Green's function 

786: obeys the following reality property:

787: \be

788: \Big( G^{({\rm R} \oplus {\rm R}_{\rm c})}(z',z)\Big)^{\rm T}

789: = \s_1 \, G^{({\rm R} \oplus {\rm R}_{\rm c})}(z,z')\, \s_1~. 

790: \label{real-2}

791: \ee

792: These relations, the symmetry property 

793: \be

794:  \langle  v^\m (z)\, v^\n (z') \rangle 

795: =  \langle  v^\n (z') v^\m (z) \rangle ~,

796: \ee

797: and a simple consequence  of (\ref{transpar}), 

798: \be 

799: [{\bar \cD}^2, \cD^2 ] \,G(z,z') 

800: = - [{\bar \cD}'^2, \cD'^2 ] \, G(z,z') ~, 

801: \ee

802: allow one to turn (\ref{I+II}) into 

803: \bea

804: \G_{\rm I +II} &=& -{{\rm i} \over 2^{10} } \int {\rm d}^8 z 

805: \int {\rm d}^8 z' \,  \langle  v^\m (z) \,v^\n (z') \rangle  \non \\

806: &\times &

807: \, {\rm  tr} \left\{ \cT_\m

808: \Big([ {\bar \cD}^2, \cD^2] \, 

809: G^{({\rm R} \oplus {\rm R}_{\rm c})}(z,z') \Big) \, \cT_\n

810:  [{\bar \cD}'^2, \cD'^2 ] \,

811: G^{({\rm R} \oplus {\rm R}_{\rm c})}(z',z) \right\}~,

812: \label{I+II-new}

813: \eea

814: Taking into account eqs.  

815: (\ref{real-1}) and (\ref{real-2}) once again, 

816: one ends up with 

817: \bea

818: \G_{\rm I +II} &=& -{{\rm i} \over 2^{9} } \int {\rm d}^8 z 

819: \int {\rm d}^8 z' \,  \langle  v^\m (z) \,v^\n (z') \rangle  \non \\

820: &\times &

821: \, {\rm  tr} \left\{ T_\m

822: \Big([ {\bar \cD}^2, \cD^2] \, 

823: G^{({\rm R})}(z,z') \Big) \, T_\n

824:  [{\bar \cD}'^2, \cD'^2 ] \,

825: G^{({\rm R} )}(z',z) \right\}~.

826: \label{I+II-final}

827: \eea

828: The following identity

829: \be

830: \frac{1}{16} [\cD^2 , {\bar \cD}^2 ] =   

831: \frac{\rm i}{4} {\bar \cD}_\ad \cD^{\a \ad } \cD_\a 

832: -\frac{\rm i}{4} \cD_\a \cD^{\a \ad } {\bar \cD}_\ad~,

833: \ee

834:  turns out  to be very useful  when computing the action 

835: of the commutators of covariant derivatives in 

836: (\ref{I+II-final}) on  the Green's functions. 

837: 

838: \begin{figure}[!htb]

839: \begin{center}

840: \includegraphics{DSfig3}

841: \caption{Two-loop supergraph III}

842: \end{center}

843: \end{figure}

844: 

845: The supergraph in Fig. 3

846: leads to the following contribution

847: \bea

848: \G_{\rm III} &=& {{\rm i} \over 16} \int {\rm d}^8 z 

849: \int {\rm d}^8 z' \,  \langle  v^\m (z) \, v^\n (z') \rangle  \non \\

850: &\times &  {\rm  tr} \left\{ T_\m \,\F^\dagger \,

851: \Big( {\bar \cD}^2  \, G^{({\rm R} )}(z,z') \Big) \,

852: T_\n \, \F \,  \cD'^2  \,G^{({\rm R} )}(z',z) \right\}~.

853: \label{III}

854: \eea

855: 

856: \begin{figure}[!htb]

857: \begin{center}

858: \includegraphics{DSfig4}

859: \caption{Two-loop supergraph IV}

860: \end{center}

861: \end{figure}

862: 

863: Finally, the supergraph in Fig. 4

864: leads to the following contribution

865: \bea

866: \G_{\rm IV} &=& -{{\rm i} \over 16} \int {\rm d}^8 z 

867: \int {\rm d}^8 z' \,  \d^8(z-z')\,

868: \langle  v^\m (z) \, v^\n (z') \rangle  

869: \non \\

870: &\times &  

871: {\rm  tr} \left\{ T_\m \,T_\n \,

872: {\bar \cD}^2  \cD'^2 G^{({\rm R} )}(z,z') 

873: \right\}~.

874: \label{IV}

875: \eea

876: 

877: \sect{Exact superpropagators} 

878: ${}$For computing quantum corrections of the form

879: (\ref{F4}), it is sufficient to consider 

880: a very special  type of background field configuration

881: specified by the constraint

882: \be

883: \cD_\a \cW_\b = 0~.

884: \label{glueball}

885: \ee

886: This is the simplest representative of 

887: background vector multiplets for which 

888: all Feynman superpropagators are known exactly 

889:  \cite{KM,KM2}. 

890: 

891: ${}$For the Green's function 

892: $G \equiv G^{({\rm R} )}$, 

893: we introduce the Fock-Schwinger 

894: proper-time representation 

895: \be

896: G(z,z') = 

897: {\rm i} \int\limits_0^\infty {\rm d}s \, K(z,z'|s) \, 

898: {\rm e}^{ -{\rm i} (|\cM|^2 -{\rm i}\ve ) s }~, 

899: \qquad   \ve \to +0~.

900: \label{proper-time-repr}

901: \ee

902: The corresponding heat kernel reads

903: \be

904: K(z,z'|s) = -\frac{\rm i}{(4 \pi s)^2} \, 

905: {\rm e}^{ {\rm i}\r^2 /4s } \, 

906: \d^2(\z  -{\rm i} s \,\cW) \,

907: \d^2({\bar \z}  + {\rm  i}s \,{\bar \cW} ) 

908: \, I(z,z')~,  

909: \label{real-kernel}

910: \ee

911: where the supersymmetric two-point function 

912: $\z^A(z,z') =- \z^A(z',z)=(\r^a , \z^\a, {\bar \z}_\ad)$

913: is defined as follows: 

914: \be

915: \r^a = (x-x')^a - {\rm i} (\q-\q') \s^a {\bar \q}' 

916: + {\rm i} \q' \s^a ( {\bar \q} - {\bar \q}') ~, \quad

917: \z^\a = (\q - \q')^\a ~, \quad

918: {\bar \z}_\ad =({\bar \q} -{\bar \q}' )_\ad ~. 

919: \label{two-point}

920: \ee

921: 

922: The parallel displacement propagator,

923: $I(z,z')$, is uniquely specified by 

924: the following requirements:\\

925: (i) the gauge transformation law

926: \be

927:  I (z, z') ~\to ~

928: {\rm e}^{{\rm i} \t(z)} \,  I (z, z') \,

929: {\rm e}^{-{\rm i} \t(z')} ~

930: \label{super-PDO1}

931: \ee

932: with respect to  a  gauge ($\t$-frame) transformation 

933: of  the covariant derivatives

934: \be

935: \cD_A ~\to ~ {\rm e}^{{\rm i} \t(z)} \, \cD_A\,

936: {\rm e}^{-{\rm i} \t(z)}~, \qquad 

937: \t^\dagger = \t ~, 

938: \label{tau}

939: \ee

940: with the gauge parameter $\t(z)$ being arbitrary modulo 

941: the reality condition imposed;\\

942: (ii) the equation  

943: \be

944: \z^A \cD_A \, I(z,z') 

945: = \z^A \Big( D_A +{\rm i} \, \G_A(z) \Big) I(z,z') =0~;

946: \label{super-PDO2}

947: \ee

948: (iii) the boundary condition 

949: \be 

950: I(z,z) ={\bf 1}~.

951: \label{super-PDO3}

952: \ee

953: These imply the important relation

954: \be

955: I(z,z') \, I(z', z) = {\bf 1}~,

956: \label{collapse}

957: \ee

958: as well as 

959: \be

960: \z^A \cD'_A \, I(z,z') 

961: = \z^A  \Big( D'_A \,I(z,z') 

962:  - {\rm i} \, I(z,z') \, \G_A(z') \Big) =0~.

963: \ee

964: 

965: ${}$For the background (\ref{glueball}), 

966: the parallel displacement propagator is completely 

967: specified by the properties:

968: \bea

969: I(z',z) \, \cD_{\a \ad} I(z,z') &=&

970: -{\rm i} ( \z_{\a} \bar{\cW}_{\ad} 

971: + \cW_\a \, \bar{\z}_{\ad}   ) ~, \non \\

972: I(z',z) \,  \cD_{\a} I(z,z') &=&

973: - \frac{{\rm i}}{2} \, \r_{\a \ad} 

974: \bar{\cW}^{\ad} 

975: + \frac13 ( \z_{\a} \bar{\z} \bar{\cW}

976: + \bar{\z}^2  \cW_{\a})~, \\

977: I(z',z) \,  {\bar \cD}_{\ad} I(z,z') &=&

978: - \frac{{\rm i}}{2} \, \r_{\a \ad} \cW^{\a} 

979: - \frac13 ( {\bar \z}_\ad \z \cW  + \z^2 {\bar \cW}_\ad)~. \non

980: \eea

981: 

982: The heat kernel corresponding 

983: to the chiral Green's function 

984: $G_+$ (\ref{chiral}) is

985: \bea

986: K_+(z,z'|s) &=&-{1 \over 4} {\bar \cD}^2 K(z,z'|s) \non \\

987:  &=& -\frac{\rm i}{(4 \pi s)^2} \, 

988: {\rm e}^{ {\rm i}\r^2 /4s } \, 

989: \d^2(\z  -{\rm i} s \,\cW) \,

990: {\rm e}^{ \frac{{\rm i}}{6} s\,\cW^2 \,({\bar \z} +{\rm i}s\, {\bar \cW})^2}

991: %{\rm e}^{ -\frac{{\rm i}}{2} \r^a 

992: %\cW   \s_a ({\bar \z} +{\rm i}s\, {\bar \cW})}

993: \, I(z,z')~. 

994: \label{chiral-glueball}

995: \eea

996: It is an instructive exercise to check, using the properties

997: of the parallel displacement propagator, 

998: that $K_+(z,z'|s) $ is covariantly chiral in both arguments. 

999: 

1000: The supersymmetric theories that we are going 

1001: to study below  are free of ultraviolet divergences. 

1002: This does not mean that individual 

1003: (say, two-loop) supergraphs are all finite; only their sum,  

1004: at any loop order, has 

1005: to be finite. To deal with UV divergent supergraphs, 

1006: we will adopt  supersymmetric dimensional regularization 

1007: via dimensional reduction \cite{GGRS}. 

1008: All manipulations with the gauge covariant derivatives 

1009: (D-algebra) have to be completed in four dimensions. 

1010: At a final stage, the bosonic part of the heat kernel 

1011: (\ref{real-kernel}) is to be continued to $d$ dimensions

1012: using the prescription

1013: \bea

1014: \frac{\rm i}{(4 \pi {\rm i} s)^2} \, 

1015: {\rm e}^{ {\rm i}\r^2 /4s }  \quad \longrightarrow \quad 

1016: \frac{\rm i}{(4 \pi {\rm i} s)^{d/2}} \, 

1017: {\rm e}^{ {\rm i}\r^2 /4s } ~,  \qquad \quad 

1018: d= 4-\ve~.

1019: \eea

1020: It is assumed that loop space-time integrals

1021: are done in $d$ dimensions, 

1022: using the following integration rules:

1023: \bea

1024: \frac{\rm i}{(4 \pi {\rm i} )^{d/2}} 

1025: \int {\rm d}^d \r \, {\rm e}^{ {\rm i} C \r^2 /4 } 

1026: &=&  C^{-d/2}~, 

1027: \non  \\

1028:  \frac{\rm i}{(4 \pi {\rm i} )^{d/2}} 

1029: \int {\rm d}^d \r \, \r_a \r_b\, 

1030: {\rm e}^{ {\rm i} C \r^2 /4 } 

1031: &=& 2\,{\rm i} \, \eta_{ab}\,

1032:  C^{-(d/2+1)}~,

1033: \label {int-rules}

1034: \eea

1035: with $C$ a positive parameter. 

1036: 

1037: \sect{$SU(N)$ SYM with $2N$ hypermultiplets

1038: in the fundamental}

1039: ${}$From now on, we choose the gauge group to be $SU(N)$.

1040: Lower-case Latin letters from the middle of the alphabet, 

1041: $i,j,\dots$, 

1042: will be used to denote matrix elements in the fundamental, 

1043: with the convention $i =0,1,\dots, N-1 \equiv 0, \un{ i}$.

1044: We choose a Cartan-Weyl basis 

1045: to consist of the elements:

1046: \be 

1047: H_I = \{ H_0, H_{\un{I}}\}~, \quad  \un{I} = 1,\dots, N-2~, 

1048: \qquad \quad E_{ij}~, \quad i\neq j~. 

1049: \label{C-W}

1050: \ee 

1051: The basis elements in the fundamental representation 

1052: are defined similarly to \cite{Georgi}, 

1053: \bea

1054: (E_{ij})_{kl} &=& \d_{ik}\, \d_{jl}~, \non \\

1055: (H_I)_{kl} &=& \frac{1}{\sqrt{(N-I)(N-I-1)} }

1056: \Big\{ (N-I)\, \d_{kI} \, \d_{lI} - 

1057: \sum\limits_{i=I}^{N-1} \d_{ki} \, \d_{li} \Big\} ~,

1058: \eea

1059: and are characterized by the properties

1060: \be

1061: {\rm tr}_{\rm F} (H_I\,H_J) = \d_{IJ}~, 

1062: \qquad 

1063: {\rm tr}_{\rm F} (E_{ij}\,E_{kl}) = \d_{il}\,\d_{jk}~, 

1064: \qquad {\rm tr}_{\rm F} (H_I \,E_{kl}) =0~.

1065: \ee

1066: 

1067: The $\cN=2$ background vector multiplet is chosen to be

1068: \be

1069: \F = \f \, H_0~, \qquad \cW_\a = W_\a \, H_0~,

1070: \ee

1071: Its characteristic feature is that it leaves

1072: the subgroup $U(1) \times SU(N-1) \subset SU(N)$ 

1073: unbroken, where $U(1)$ is associated with $H_0$

1074: and  $SU(N-1)$ is  generated by 

1075: $\{ H_{\un{I}}, \, E_{ \un{i} \un{j}} \}$.

1076: In evaluating the supergraphs, we 

1077: consider $\f$ and $W_\a$ to be constant.

1078: This suffices for our purposes.

1079: 

1080: The mass matrix is 

1081: \be

1082: |\cM|^2 = {\bar \f}\f \, (H_0)^2~, 

1083: \ee

1084: and therefore a superfield's mass is determined 

1085: by  its $U(1)$ charge with respect to $H_0$. 

1086: With the notation 

1087: \be

1088: e_{\rm f} = \sqrt{ { N-1 \over N}}  ={1 \over e_{\rm a} }~,

1089: \ee

1090: the $U(1)$ charges of all quantum superfields 

1091: are given in the table.

1092: \begin{center}

1093: \begin{tabular}{ | c || c | c || c | c| |c| c| c|}  \hline

1094: superfield & $q_0$ & $q_{\un{i}}$ & $\tilde{q}_0 $ &

1095: $\tilde{q}_{\un{i}} $ & $v^{0\, \un{i} }$ & $v^I$ &

1096: $v^{ \un{i} \,\un{j} }$\\ \hline 

1097: $U(1) $ charge & $e_{\rm f} $ & $e_{\rm f} - e_{\rm a} $ &

1098: $-e_{\rm f} $ & $e_{\rm a} - e_{\rm f} $ & $e_{\rm a} $ 

1099: &0&0\\  \hline

1100: \end{tabular} \\

1101: ${}$\\

1102: Table  1: $U(1)$ charges of superfields

1103: \end{center}

1104: As can be  seen, all fundamental hypermultiplet superfields are 

1105: massive. For the adjoint superfields\footnote{Since the basis 

1106: (\ref{C-W}) is not orthonormal, 

1107: ${\rm tr}_{\rm F} (T_\m \, T_\n)  =g_{\m \n} \neq \d_{\m \n}$, 

1108: it is necessary to keep track of the Cartan-Killing metric when 

1109: working with adjoint vectors. For any elements

1110: $u=u^\m T_\m $ and $v=v^\m T_\m $ of the Lie algebra, 

1111: we have $u\cdot v =  {\rm tr}_{\rm F} (u\,v) =u^\m \,v_\m$, 

1112: where $v_\m = g_{\m\n} v^\n$

1113: ($v_I =v^I$, $v_{ij}= v^{ji}$).}

1114: \be

1115: v = v^I \, H_I + v^{ij} \, E_{ij} \equiv v^\m \,T_\m~,  

1116: \qquad i \neq j ~, 

1117: \ee

1118: there are $2(N-1)$ massive superfields 

1119: ($v^{0 \un{i}}$ and their conjugates $v^{ \un{i}0}$), 

1120: while the remaining  $(N-1)^2$ superfields, $v^I$ and 

1121: $v^{\un{i} \,\un{j}} $,  are free massless.

1122: This follows from the identity 

1123: \be

1124: [H_0 , E_{ij}] ~=~ \sqrt{N \over N-1}\,  \Big( 

1125: \d_{0i}\, E_{0j} - \d_{0j}\, E_{i0} \Big)~. 

1126: \ee

1127: 

1128: Let us denote by $\bG^{(e)}(z,z')$ the Green's function 

1129: (\ref{green}) in the special  case when 

1130: the gauge group is $U(1)$ generated by $H_0$, 

1131: and the quantum superfield $\S$ in (\ref{S-action})

1132: carries $U(1)$ charge $e$, 

1133: $H_0 \,\S = e\, \S$ (in particular, the mass matrix is 

1134: $|\cM|^2 = e^2 \, {\bar \f} \f$). 

1135: The Green's function has the proper-time representation

1136: \be

1137: \bG^{(e)}(z,z') = 

1138: {\rm i} \int\limits_0^\infty {\rm d}s \, 

1139: \bK^{(e)}(z,z'|s) \, 

1140: {\rm e}^{ -{\rm i} ( e^2  {\bar \f} \f  -{\rm i}\ve ) s }~, 

1141: \qquad   \ve \to +0~,

1142: \label{U(1)GF}

1143: \ee

1144: where the  heat kernel is 

1145: \be

1146: \bK^{(e)}(z,z'|s) = -\frac{\rm i}{(4 \pi s)^2} \, 

1147: {\rm e}^{ {\rm i}\r^2 /4s } \, 

1148: \d^2(\z  -{\rm i} es \,W) \,

1149: \d^2({\bar \z}  + {\rm  i}es \,{\bar W} ) 

1150: \, I^{(e)}(z,z')~.

1151: \label{U(1)HK}

1152: \ee

1153: ${}$For the $\cN=2$ background vector multiplet chosen, 

1154: all the Feynman propagators are expressed via such 

1155: $U(1)$ Green's functions. 

1156: 

1157: In the remainder of this section, we 

1158: specialize to  the case of $\cN=2$ SYM

1159: with $2N$ hypermultiplets in the fundamental 

1160: representation of $SU(N)$. This theory is finite

1161: since the finiteness condition 

1162: (\ref{finita}) is satisfied 

1163: due to the well-known $SU(N)$ identity

1164: \be

1165: {\rm tr}_{\rm Ad} \,\J^2 ~=~ 2N\, {\rm tr}_{\rm F} \, \J^2~, 

1166: \qquad  \quad \J \in sl(N)~. 

1167: \ee

1168: 

1169: \subsection{Evaluation of $\G_{\rm I+II}$ }

1170: We now turn to evaluating $\G_{\rm I+II}$. 

1171: In accordance with (\ref{I+II-final}),

1172: it is necessary to analyze the expression 

1173: \bea

1174: 2N\,  \langle  v^\m (z) \,v^\n (z') \rangle  

1175:  &{\rm  tr}_{\rm F} & \left\{ T_\m

1176: \Big([ {\bar \cD}^2, \cD^2] \, 

1177: G^{({\rm F})}(z,z') \Big) \, T_\n

1178:  [{\bar \cD}'^2, \cD'^2 ] \,

1179: G^{({\rm F} )}(z',z) \right\} \non \\

1180: = 2N \sum\limits_{I} 

1181: \langle  v^I (z) \,v^I (z') \rangle  

1182: & {\rm  tr}_{\rm F} & \left\{ H_I

1183: \Big([ {\bar \cD}^2, \cD^2] \, 

1184: G^{({\rm F})}(z,z') \Big) \, H_I 

1185:  [{\bar \cD}'^2, \cD'^2 ] \,

1186: G^{({\rm F} )}(z',z) \right\} 

1187: \label{I+II-fund1}\\

1188: + 2N\sum\limits_{i\neq j} 

1189: \langle  v^{ij}  (z) \,v^{ji} (z') \rangle  

1190: &{\rm  tr}_{\rm F} &\left\{ E_{ij}

1191: \Big([ {\bar \cD}^2, \cD^2] \, 

1192: G^{({\rm F})}(z,z') \Big) \, E_{ji} 

1193:  [{\bar \cD}'^2, \cD'^2 ] \,

1194: G^{({\rm F} )}(z',z) \right\} ~, \non 

1195: \eea

1196: where the factor $2N$ relates to the presence 

1197: of $2N$ hypermultiplets.

1198: The expression in the second line can be 

1199: simplified on the basis of 

1200: the following observations: (i) the propagator 

1201: $G^{({\rm F} )}$ is diagonal; (ii)

1202: the massless adjoint  propagators are identical, 

1203: \be 

1204: \langle  v^I (z) \,v^I (z') \rangle  = 

1205: \langle  v^{\un{i}\, \un{j}}  (z) \,v^{\un{j}\,\un{i}} (z') \rangle  

1206: = {\rm i}\,g^2 \,{\bf G}^{(0)} (z,z')~.

1207: \ee

1208: with ${\bf G}^{(0)} (z,z')$ the free massless Green's function.

1209: Then, the second line of (\ref{I+II-fund1}) becomes

1210: \be 

1211: 2{\rm i}N \,g^2\,\bG^{(0)} (z, z')  \, 

1212: \sum\limits_{I}  

1213:  {\rm  tr}_{\rm F}  \left\{ (H_I)^2

1214: \Big([ {\bar \cD}^2, \cD^2] \, 

1215: G^{({\rm F})}(z,z') \Big) \,

1216:  [{\bar \cD}'^2, \cD'^2 ] \,

1217: G^{({\rm F} )}(z',z) \right\}~. 

1218: \ee

1219: In the fundamental representation of $SU(N)$,  

1220: \be 

1221: \sum\limits_{I} (H^{({\rm F})}_I)^2 ~=~ { {N-1} \over N} \, {\bf 1}~.

1222: \ee 

1223: This gives

1224: \bea

1225: && 2N \sum\limits_{I} 

1226: \langle  v^I (z) \,v^I (z') \rangle  

1227:  {\rm  tr}_{\rm F}  \left\{ H_I

1228: \Big([ {\bar \cD}^2, \cD^2] \, 

1229: G^{({\rm F})}(z,z') \Big) \, H_I 

1230:  [{\bar \cD}'^2, \cD'^2 ] \,

1231: G^{({\rm F} )}(z',z) \right\} \non \\

1232: &=& 2 {\rm i}\,( N-1) \, g^2\,

1233: \bG^{(0)} (z,z')  \, 

1234: \Big\{ \Big([ {\bar \cD}^2, \cD^2] \, 

1235: \bG^{(e_{\rm f})}(z,z') \Big) \,

1236:  [{\bar \cD}'^2, \cD'^2 ] \,

1237: \bG^{(e_{\rm f} )}(z',z) \\

1238: &&\qquad \qquad +~ (N-1) 

1239: \Big([ {\bar \cD}^2, \cD^2] \, 

1240: \bG^{(e_{\rm f}-e_{\rm a} )}(z,z') \Big) \,

1241:  [{\bar \cD}'^2, \cD'^2 ] \,

1242: \bG^{(e_{\rm f} -e_{\rm a})}(z',z) \Big\} ~. \non 

1243: \eea

1244: 

1245: 

1246: To transform the expression in the 

1247: third  line of  (\ref{I+II-fund1}),  

1248: we notice 

1249: \bea

1250: && {\rm  tr}_{\rm F} \left\{ E_{ij} \Big([ {\bar \cD}^2, \cD^2] \, 

1251: G^{({\rm F})}(z,z') \Big) \, E_{ji}  [{\bar \cD}'^2, \cD'^2 ] \,

1252: G^{({\rm F} )}(z',z) \right\} \non \\

1253: &&\quad = \Big([ {\bar \cD}^2, \cD^2] \, 

1254: G^{({\rm F})}(z,z') \Big)_{jj} \, 

1255: \Big(  [{\bar \cD}'^2, \cD'^2 ] \,

1256: G^{({\rm F} )}(z',z) \Big)_{ii}~, 

1257: \eea

1258: as immediately follows from the definition of $E_{ij}$.

1259: This leads to 

1260: \bea

1261: && \sum\limits_{i\neq j} 

1262: \langle  v^{ij}  (z) \,v^{ji} (z') \rangle  

1263: {\rm  tr}_{\rm F} \left\{ E_{ij}

1264: \Big([ {\bar \cD}^2, \cD^2] \, 

1265: G^{({\rm F})}(z,z') \Big) \, E_{ji} 

1266:  [{\bar \cD}'^2, \cD'^2 ] \,

1267: G^{({\rm F} )}(z',z) \right\}   \\

1268: &=& {\rm i} \, (N-1)g^2\,  \bG^{(e_{\rm a} )}(z,z') \,

1269: \Big([ {\bar \cD}^2, \cD^2] \, 

1270: \bG^{(e_{\rm f} -e_{\rm a})}(z,z') \Big) \,

1271: [{\bar \cD}'^2, \cD'^2 ] \,

1272: \bG^{(e_{\rm f} )}(z',z)  ~+~

1273:  (z \leftrightarrow z') \non \\

1274: &&+ {\rm i}\, (N-1)(N-2) g^2\,

1275: \bG^{(0)} (z, z') \, 

1276: \Big([ {\bar \cD}^2, \cD^2] \, 

1277: \bG^{(e_{\rm f} -e_{\rm a})}(z,z') \Big) \,

1278: [{\bar \cD}'^2, \cD'^2 ] \,

1279: \bG^{(e_{\rm f} - e_{\rm a} )}(z',z) ~. \non 

1280: \eea

1281: 

1282: As should be clear from the above consideration,

1283: the evaluation of $\G_{\rm I+II} $ amounts to computing 

1284: a functional of the form 

1285: \be

1286: \int {\rm d}^8 z  \int {\rm d}^8 z' \,  

1287: \bG^{(e_1)} (z, z') \, 

1288: \Big([ {\bar \cD}^2, \cD^2] \, 

1289: \bG^{(e_2)}(z,z') \Big) \,

1290: [{\bar \cD}'^2, \cD'^2 ] \,

1291: \bG^{(e_1 + e_2 )}(z',z) ~, 

1292: \label{block}

1293: \ee 

1294: for some charges $e_1$ and $e_2$. 

1295: ${}$For all the Green's functions, we introduce

1296: the proper-time representation (\ref{U(1)GF}). 

1297: Due to the explicit structure of the heat 

1298: kernel, eq. (\ref{U(1)HK}), the first 

1299: multiplier in (\ref{block}) contains a 

1300: Grassmann delta-function,  

1301: $$

1302: \d^2(\z  -{\rm i} e_1s \,W) \,

1303: \d^2({\bar \z}  + {\rm  i}e_1s \,{\bar W} ) ~, 

1304: $$

1305: which allows us to do the integral over

1306: $\q'$.  Next, the second and third multipliers in 

1307: (\ref{block}) can be evaluated 

1308: (in $d$ dimensions) as follows:

1309: \bea

1310: \frac{1}{16} [ {\bar \cD}^2, \cD^2 ] \, 

1311: \bK^{(e)}(z,z' | s) 

1312: & \approx  & \phantom{-}

1313: \frac{\rm i}{(4 \pi s)^2} \, 

1314: {\r  ^{\a \ad} \over s}

1315: (\z -{\rm i} es W)_\a   

1316:  (\bar{\z} +{\rm i} es {\bar W})_\ad  \,

1317: {\rm e}^{ {\rm i} \r^2/4s} \, I^{(e)}(z,z')  \non \\

1318: \longrightarrow

1319: & & - \frac{\rm i}{(4 \pi {\rm i}s)^{d/2}} \, 

1320: {\r  ^{\a \ad} \over s}

1321: (\z -{\rm i} es W)_\a   

1322:  (\bar{\z} +{\rm i} es {\bar W})_\ad  \,

1323: {\rm e}^{ {\rm i} \r^2/4s} \, I^{(e)}(z,z') ~,\non 

1324: \eea

1325: where we have omitted all terms of at least third order 

1326: in the Grassmann variables  $\z_\a, \,{\bar \z}_\ad$ and 

1327: $W_\a, \, {\bar W}_\ad$ 

1328: as they  do not contribute to (\ref{block}). 

1329: Now, the parallel displacement propagators 

1330: associated with the three Green's functions

1331: in (\ref{block}) simply annihilate each other. 

1332: Finally , the integral over $x'$ in (\ref{block}) 

1333: can easily be done if one  first replaces the bosonic variables

1334: $\{x,x'\} \to \{x,\r\}$ and then applies eq. 

1335: (\ref{int-rules}). 

1336: Of special importance is the fact that 

1337: the functional 

1338: \be

1339: \J = {1\over 2^8}

1340: \int {\rm d}^8 z  \int {\rm d}^8 z' \,  

1341: \bG^{(0)} (z, z') \, 

1342: \Big([ {\bar \cD}^2, \cD^2] \, 

1343: \bG^{(e)}(z,z') \Big) \,

1344: [{\bar \cD}'^2, \cD'^2 ] \,

1345: \bG^{(e )}(z',z) 

1346: \label{FUN-1}

1347: \ee 

1348: is finite (so we set $d = 4$) and does not

1349: depend on the charge $e$,

1350: for the background field configuration chosen,

1351: \bea

1352: \J &= & {4e^4 \over (4\p)^4 }

1353: \int {\rm d}^8 z  \, W^2 {\bar W}^2 \, 

1354: \int\limits_{0}^{\infty} 

1355: {\rm d}s_1  

1356: \int\limits_{0}^{\infty}

1357:  {\rm d}s_2 

1358: \int\limits_{0}^{\infty} 

1359: {\rm d}s_3  \,

1360: {\rm e}^{-e^2 \f {\bar \f} (s_2+s_3)} \,

1361: \frac{s_1\,s_2{}^2 \,s_3{}^2} 

1362: {(s_2s_3 + s_1s_2 +s_1s_3)^3} \non \\

1363: &=& {1 \over 3}\, {1 \over (4\p)^4 }

1364: \int {\rm d}^8 z  \, 

1365: { W^2 {\bar W}^2  \over (\f {\bar \f})^2}

1366: =   {1 \over 3}\, {1 \over (4\p)^4 } \, \U~,

1367: \label{FUN-2}

1368: \eea

1369: see \cite{KM2} for more details.

1370: 

1371: The computational scheme outlined 

1372: leads to the final result\footnote{In this paper, 

1373: we do not evaluate all of the proper-time 

1374: integrals,  such as $I_{\rm I+II} $. 

1375: We are only interested in their large $N$ behaviour 

1376: and in their singular parts. This is why we freely 

1377: set $d=4$ in  finite multiplicative factors, such as

1378: $(\f {\bar \f})^{-d/2}$. No mass scale 

1379: is required because the total contribution is finite. }: 

1380: \be

1381: \G_{\rm I+II} = \frac{g^2}{(4\p)^4} \, 

1382: \U \Big\{ {1\over 3} \,N(N-1)^2  

1383: +  8\,N(N-1)\, I_{\rm I+II} \Big\} ~,

1384: \ee

1385: where 

1386: \bea

1387:  I_{\rm I+II}  = \int\limits_{0}^{\infty} {\rm d}s_1 

1388: {\rm d}s_2

1389: {\rm d}s_3 \, 

1390: {\rm e}^{-[{\rm e}_{\rm a}^2 s_1 

1391: + ({\rm e}_{\rm a}-{\rm e}_{\rm f})^2 s_3

1392: + {\rm e}_{\rm f}^2 s_2]} \, 

1393: \frac{ s_1 

1394: [{\rm e}_{\rm a} s_1 + ({\rm e}_{\rm a}-{\rm e}_{\rm f})s_3]^2

1395: [{\rm e}_{\rm a} s_1 + {\rm e}_{\rm f}s_2 ]^2}

1396: {[s_2s_3 +s_1s_2 +s_1 s_3]^{d/2 + 1} }

1397: \label{I-I+II}

1398: \eea

1399: is a divergent integral

1400: in the limit $\ve = 4-d \to 0$. 

1401: 

1402: To isolate the divergence in (\ref{I-I+II}),

1403: we first rescale the  integral 

1404: \bea

1405:  I_{\rm I+II}  &=& {(N-1)^4 \over N^2} 

1406: \int\limits_{0}^{\infty} {\rm d}t_1  

1407: {\rm d}t_2

1408: {\rm d}t_3 \, 

1409: {\rm e}^{-[ t_1 + t_2 + t_3]} \,

1410: \frac{ t_1 

1411: [t_1 + Nt_3]^2

1412: [t_1 + {N \over N-1} t_2 ]^2}

1413: {[t_1 t_2 + (N-1)^2t_1t_3 + N^2 t_2 t_3]^{d/2 + 1} }~.

1414: \eea

1415: Now,  it is advantageous to  

1416: introduce new variables \cite{Zinn}:

1417: \be

1418: t_1 = s\,t\,u~, \qquad 

1419: t_2 =s\,t\,(1-u)~, \qquad 

1420: t_3 =s\,(1-t)~, 

1421: \ee

1422: with the important properties $s = t_1 +t_2+t_3$, 

1423: $s\,t = t_1 +t_2$ and 

1424: \be 

1425: \int\limits_{0}^{\infty} {\rm d}t_1 

1426: \int\limits_{0}^{\infty} {\rm d}t_2 

1427: \int\limits_{0}^{\infty} {\rm d}t_3 \cdots 

1428: ~=~\int\limits_{0}^{\infty} {\rm d}s

1429: \int\limits_{0}^{1} {\rm d} t 

1430: \int\limits_{0}^{1} {\rm d} u \, s^2 \,t  \cdots

1431: \ee

1432: The integral over $s$ factorizes and it is convergent, 

1433: \be

1434: \int\limits_{0}^{\infty} {\rm d}s \,

1435: s^{5-d}\, {\rm e}^{-s} = 1 +O(\ve)~.

1436: \ee

1437: As a result , we obtain

1438: \bea

1439: I_{\rm I+II}  &=& \left({N-1 \over N} \right)^2

1440: \int\limits_{0}^{1} {\rm d} t 

1441: \int\limits_{0}^{1} {\rm d} u

1442: \frac{ t^{3-d/2} u [N-u]^2 [N-t(N-u)]^2}

1443: {[(N-u)^2(1-t) +u(1-u) ]^{d/2 +1} } ~+~{\rm finite}~.

1444: \eea

1445: The divergent part of $I_{\rm I+II}$

1446: turns out to be 

1447: \be

1448: (I_{\rm I+II})_{\rm div} = 

1449: \left(-1 +4 \, {N-1 \over N^2} \right)\,

1450:  {1 \over \ve} ~.

1451: \label{I-I+II-div}

1452: \ee

1453: 

1454:  

1455: \subsection{Evaluation of $\G_{\rm III}$ }

1456: 

1457: The evaluation of $\G_{\rm III}$ is very similar 

1458: to that of $\G_{\rm I+II}$ just described. 

1459: Therefore, we simply give the final result: 

1460: \be

1461: \G_{\rm III} = \frac{g^2}{(4\p)^4} \, 

1462: \U \Big\{ {2\over 3} \,N(N-1)^2  

1463: -  4\,(N-1)\, I_{\rm III} \Big\} ~,

1464: \ee

1465: where 

1466: \bea

1467:  I_{\rm III}  = \int\limits_{0}^{\infty} {\rm d}s_1 

1468: {\rm d}s_2

1469: {\rm d}s_3 \, 

1470: {\rm e}^{-[{\rm e}_{\rm a}^2 s_1 

1471: + ({\rm e}_{\rm a}-{\rm e}_{\rm f})^2 s_3

1472: + {\rm e}_{\rm f}^2 s_2]} \, 

1473: \frac{  

1474: [{\rm e}_{\rm a} s_1 + ({\rm e}_{\rm a}-{\rm e}_{\rm f})s_3]^2

1475: [{\rm e}_{\rm a} s_1 + {\rm e}_{\rm f}s_2 ]^2}

1476: {[s_2s_3 +s_1s_2 +s_1 s_3]^{d/2 } }

1477: \label{I-III}

1478: \eea

1479: is a divergent integral

1480: in the limit $\ve = 4-d \to 0$. 

1481: Its divergent part proves to be

1482: \be

1483: (I_{\rm III})_{\rm div} =  

1484: 4 \, {N-1 \over N} \,

1485:  {1 \over \ve} ~.

1486: \ee

1487: 

1488: \subsection{Evaluation of $\G_{\rm IV}$ }

1489: It remains to evaluate  $\G_{\rm IV}$. 

1490: As is seen from  its defining  expression (\ref{IV}), 

1491: $ \G_{\rm IV}$ involves a vector propagators 

1492: at coincident points,  

1493: $\langle  v^\m (z) \, v^\n (z) \rangle $. 

1494: The latter is non-trivial only for the 

1495: massive  superfields, 

1496: \bea 

1497: \langle  v^I (z) \, v^I (z) \rangle &=&

1498: \langle  v^{\un{i} \un{j} }  (z) \, 

1499: v^{\un{j} \un{i} } (z) \rangle =0~, \non \\

1500: \langle  v^{0 \un{i}  }  (z) \, 

1501: v^{\un{i}0 } (z) \rangle &=& 

1502: \langle  v^{\un{i} 0 }  (z) \, 

1503: v^{0  \un{i} } (z) \rangle

1504: =-{ g^2 \over 8 \p^2  } 

1505: \left( {N-1 \over N} \right) \,

1506: \frac{ W^2 {\bar W}^2 }{ (\f {\bar \f})^3 } ~. 

1507: \eea

1508: Thus, we can rewrite $ \G_{\rm IV}$ in the form 

1509: \bea

1510: \G_{\rm IV} &=& {{\rm i}g^2  \over 128 \p^2 }

1511: \left({N-1 \over N} \right) 

1512:  \int {\rm d}^8 z 

1513: \frac{ W^2 {\bar W}^2 }{ (\f {\bar \f})^3 } \,

1514: {\rm  tr} \left\{ \{ E_{0 \un{i} }, E_{\un{i} 0} \} \,

1515: {\bar \cD}^2  \cD'^2 G^{({\rm R} )}(z,z') \Big|_{z'=z}

1516: \right\}~.

1517: \label{IV-mod}

1518: \eea

1519: 

1520: In the superconformal theory under consideration, 

1521: the quantum correction  (\ref{IV-mod}) is

1522: \bea

1523: \G_{\rm IV} &=& {{\rm i}g^2  \over 64 \p^2 }

1524: (N-1 ) 

1525:  \int {\rm d}^8 z 

1526: \frac{ W^2 {\bar W}^2 }{ (\f {\bar \f})^3 } \,

1527: {\rm  tr}_{\rm F} 

1528: \left\{ \{ E_{0 \un{i} }, E_{\un{i} 0} \} \,

1529: {\bar \cD}^2  \cD'^2 G^{({\rm F} )}(z,z') \Big|_{z'=z}

1530: \right\}~.

1531: \eea

1532: Its direct evaluation gives

1533: \be

1534: \G_{\rm IV} = -4

1535: \frac{g^2}{(4\p)^4} \, 

1536: \U \,(N-1) \Big\{ N -  2\,{N-1 \over N} \Big\}

1537: \, I_{\rm IV}  ~,

1538: \ee

1539: where 

1540: \be

1541:  I_{\rm IV}  = \int\limits_{0}^{\infty} {\rm d}s

1542:  \,s^{-d/2} \,  {\rm e}^{-s} 

1543: \label{IV-int}

1544: \ee

1545: is a a divergent integral

1546: in the limit $\ve = 4-d \to 0$, 

1547: \be

1548: (I_{\rm IV})_{\rm div} =  

1549: - {2 \over \ve} ~. 

1550: \ee

1551: It is easy to check that 

1552: \be

1553: \Big( \G_{\rm I+II} +

1554: \G_{\rm III} +

1555: \G_{\rm IV}  \Big)_{\rm div} = 0 ~,

1556: \ee

1557: consistent with the finiteness of the theory.

1558: 

1559: 

1560: \sect{$\cN=4$ SYM}

1561: 

1562: We now turn to evaluating to the two-loop supergraphs

1563: with quantum hypermultiplets in the $\cN=4$ super Yang-Mills

1564: theory which is simply $\cN=2$ SYM with a single hypermultiplet 

1565: in the adjoint.

1566: 

1567: \subsection{Evaluation of $\G_{\rm I+II}$ }

1568: We start by analyzing 

1569: $\G_{\rm I+II}$ 

1570: in the case of the adjoint representation. 

1571: According to (\ref{I+II-final}), we have to compute  

1572: \bea

1573: &&  \langle  v^\m \,v'^\n \rangle  

1574:  {\rm  tr}_{\rm Ad}  \left\{ T_\m

1575: \hat{G} \, T_\n \, \hat{G}'  \right\} 

1576: \label{I+II-ad1}

1577: \eea

1578: where we have introduced the following 

1579: condensed  notation:

1580: $$

1581:  \langle  v^\m \,v'^\n \rangle  

1582: = \langle  v^\m (z) \,v^\n (z') \rangle ~, \quad 

1583: \hat{G} =  [ {\bar \cD}^2, \cD^2] \,  G^{({\rm Ad})}(z,z')~,

1584: \quad 

1585: \hat{G}' = [ {\bar \cD}'^2, \cD'^2] \,  G^{({\rm Ad})}(z',z)~.

1586: $$

1587: Relative to the basis 

1588: $T_\m = (H_I, E_{0 \un{i} }, E_{\un{i} 0} , E_{ \un{i} \un{j} })$, 

1589: the hypermultiplet operator 

1590: $\hat{G} = (\hat{G}^\l{}_\r) $ in (\ref{I+II-ad1})

1591: has a diagonal structure, 

1592: \be

1593: \hat{G} = {\rm diag} \Big(\hat{\bG}^{(0)}\, {\bf 1}_{N-1} , ~

1594: \hat{\bG}^{(e_{\rm a} )}\, {\bf 1}_{N-1}, ~

1595: \hat{\bG}^{(-e_{\rm a} )}\, {\bf 1}_{N-1}, ~

1596: \hat{\bG}^{(0)}\, {\bf 1}_{(N-1)(N-2)} \Big)~,

1597: \label{I+II-ad12}

1598: \ee

1599: with the $U(1)$ charges given explicitly. 

1600: The evaluation of (\ref{I+II-ad1}) will be based on 

1601: considerations of charge conservation. 

1602: At each vertex ($z$ or $z'$), the total charge must be zero. 

1603: The possible charges in the adjoint representation are:

1604: $0, \pm e_{\rm a}$. Therefore, there are 

1605:  contributions to (\ref{I+II-ad1})

1606: of the two different types: (i) one line is neutral,

1607: and hence free massless, while the other two lines carry charges 

1608: $\pm   e_{\rm a}$; (ii) all three lines are neutral, and hence 

1609: free massless. The case (ii) can safely be ignored since 

1610: no dependence on the background fields is present.

1611: With such considerations in mind, we first separate

1612: the contributions to (\ref{I+II-ad1}) with neutral

1613: and charged gauge field lines:

1614: \bea

1615: &&  \sum\limits_{I} \langle  v^I \,v'^I \rangle  

1616:  {\rm  tr}_{\rm Ad}  \left\{ H_I

1617: \hat{G} \,   H_I \,\hat{G}'  \right\} 

1618: + \sum\limits_{\un{i} \neq \un{j}} 

1619: \langle  v^{\un{i}\un{j} }  \,v'^{\un{j} \un{i} }  \rangle  

1620:  {\rm  tr}_{\rm Ad}  \left\{ E_{\un{i} \un{j} }

1621: \hat{G} \,   E_{ \un{j} \un{i} }  \,\hat{G}'  \right\} \non \\

1622: &+&   

1623: \sum\limits_{\un{i} } 

1624: \langle  v^{0 \un{i} }  \,v'^{ \un{i} 0  }  \rangle  

1625:  {\rm  tr}_{\rm Ad}  \left\{ E_{0 \un{i}  }

1626: \hat{G} \,  E_{ \un{i} 0 }  \,\hat{G}'  \right\}  

1627: + \sum\limits_{\un{i} } 

1628: \langle  v^{ \un{i}0 }  \,v'^{ 0\un{i}  }  \rangle  

1629:  {\rm  tr}_{\rm Ad}  \left\{ E_{ \un{i} 0 }

1630: \hat{G} \,  E_{ 0\un{i}  }  \,\hat{G}'  \right\}  ~.

1631: \label{I+II-ad3}

1632: \eea

1633: Since the propagators 

1634: $\langle  v^I \,v'^I \rangle  =

1635: \langle  v^{\un{i}\un{j} }  \,v'^{\un{j} \un{i} }  \rangle 

1636: ={\rm i}\,g^2 {\bf G}^{(0)} (z,z')$

1637: are free massless, both $\hat{G}$ and $\hat{G}'$ 

1638: in the first line of (\ref{I+II-ad3}) should be charged. 

1639: In the second line of (\ref{I+II-ad3}), one of the

1640: $\hat{G}$ and $\hat{G}'$ should be  neutral,

1641: while  the other is charged. 

1642: We will analyze separately 

1643: the contributions  appearing in (\ref{I+II-ad3}).

1644: 

1645: Let $T^{({\rm Ad})}_\m$ be the matrix generators 

1646: in the adjoint representation, 

1647: \be

1648: [T_\m ,  T_\n ] = T_\l \, 

1649: (T^{({\rm Ad})}_\m )^\l{}_\n ~.

1650: \ee

1651: Since $H_I$, $\hat{G}$ and $\hat{G}'$ are diagonal, 

1652: the first term in (\ref{I+II-ad3}) becomes 

1653: \bea

1654: && \sum\limits_{I} \langle  v^I \,v'^I \rangle  

1655:  {\rm  tr}_{\rm Ad}  \left\{ H_I

1656: \hat{G} \,   H_I \,\hat{G}'  \right\} 

1657: = \sum\limits_{I} \langle  v^I \,v'^I \rangle  

1658:  {\rm  tr}_{\rm Ad}  \left\{ (H_I)^2

1659: \hat{G} \,  \hat{G}'  \right\} \non \\

1660: &=& {\rm i} \,g^2\,{\bf G}^{(0)} (z,z') \Big\{ 

1661: \hat{\bG}^{(e_{\rm a} )}\,\hat{\bG}'^{(e_{\rm a} )}

1662: + \hat{\bG}^{(-e_{\rm a} )}\,\hat{\bG}'^{(-e_{\rm a} )}\Big\}

1663: \sum\limits_{I} \sum\limits_{\un{i} }  

1664: (H^{({\rm Ad})}_I)^{0\un{i}} {}_ {0\un{i}} \,

1665: (H^{({\rm Ad})}_I)^{0\un{i}} {}_ {0\un{i}} ~, 

1666: \eea

1667: where we have used the identity

1668: \be

1669: (H^{({\rm Ad})}_I)^{0\un{i}} {}_ {0\un{i}} 

1670: =-(H^{({\rm Ad})}_I)^{\un{i}0} {}_ {\un{i}0} ~.

1671: \ee

1672: The group-theoretic factor in the last expression 

1673: is easy to evaluate:

1674: \be

1675: \sum\limits_{I} \sum\limits_{\un{i} } 

1676: (H^{({\rm Ad})}_I)^{0\un{i}} {}_ {0\un{i}} \,

1677: (H^{({\rm Ad})}_I)^{0\un{i}} {}_ {0\un{i}}  = 2(N-1) ~.

1678: \label{GT-1}

1679: \ee

1680: 

1681: Let us turn to the second term in (\ref{I+II-ad3}), 

1682: \be

1683: \sum\limits_{\un{i} \neq \un{j}} 

1684: \langle  v^{\un{i}\un{j} }  \,v'^{\un{j} \un{i} }  \rangle  

1685:  {\rm  tr}_{\rm Ad}  \left\{ E_{\un{i} \un{j} }

1686: \hat{G} \,   E_{ \un{j} \un{i} }  \,\hat{G}'  \right\} 

1687: ={\rm i} \, g^2\, {\bf G}^{(0)} (z,z') \,

1688: \sum\limits_{\un{i} \neq \un{j}} 

1689:  {\rm  tr}_{\rm Ad}  \left\{ E_{\un{i} \un{j} }

1690: \hat{G} \,   E_{ \un{j} \un{i} }  \,\hat{G}'  \right\} ~.

1691: \ee

1692: Since both the hypermultiplets must be 

1693: massive and of opposite charge, 

1694: for this expression we get 

1695: \bea

1696: {\rm i} \,g^2\, {\bf G}^{(0)} (z,z') \Big\{ 

1697: \hat{\bG}^{(e_{\rm a} )}\,\hat{\bG}'^{(e_{\rm a} )}

1698: + \hat{\bG}^{(-e_{\rm a} )}\,\hat{\bG}'^{(-e_{\rm a} )}\Big\}

1699:  \sum\limits_{\un{i} \neq \un{j}  }

1700:  \sum\limits_{\un{k} ,\, \un{l}  } 

1701: (E^{({\rm Ad})}_{\un{i} \un{j} })^{0\un{k}} {}_ {0\un{l}} \,

1702: (E^{({\rm Ad})}_{\un{i} \un{j} })^{ 0 \un{l} } {}_ {0 \un{k} } ~,

1703: \eea

1704: where the following identity 

1705: \be

1706: (E^{({\rm Ad})}_{\un{i} \un{j} })^{\un{k}0} {}_ {\un{l}0} =

1707: -(E^{({\rm Ad})}_{\un{i} \un{j} })^{0\un{l}} {}_ {0\un{k}} 

1708: \ee

1709: has been used. The group-theoretic factor in the last expression 

1710: is also easy to evaluate:

1711: \be 

1712: \sum\limits_{\un{i} \neq \un{j}  }

1713:  \sum\limits_{\un{k} ,\, \un{l}  } 

1714: (E^{({\rm Ad})}_{\un{i} \un{j} })^{0\un{k}} {}_ {0\un{l}} \,

1715: (E^{({\rm Ad})}_{\un{i} \un{j} })^{ 0 \un{l} } {}_ {0 \un{k} }

1716: =(N-1)(N-2)~.

1717: \label{GT-2}

1718: \ee

1719: 

1720: As a result, the first and second  terms in  

1721: (\ref{I+II-ad3}) lead  to the following contribution 

1722: \be 

1723: {\rm i} \,N(N-1) \,g^2\, {\bf G}^{(0)} (z,z') \Big\{ 

1724: \hat{\bG}^{(e_{\rm a} )}\,\hat{\bG}'^{(e_{\rm a} )}

1725: + \hat{\bG}^{(-e_{\rm a} )}\,\hat{\bG}'^{(-e_{\rm a} )}\Big\}~.

1726: \label{I+II-ad4}

1727: \ee

1728: 

1729: We now turn to the third term in (\ref{I+II-ad3}). 

1730: Since $\langle  v^{0 \un{i} }  \,v'^{ \un{i} 0  } \rangle

1731: = {\rm i} \,g^2{\bf G}^{(e_{\rm a})} (z,z') $ 

1732: is a massive propagator of charge $+ e_{\rm a}$, 

1733: one of the hypermultiplet propagators

1734: must be massive of charge $- e_{\rm a}$,

1735: with the other must be free neutral. 

1736: \bea

1737: && \qquad \qquad 

1738: \sum\limits_{\un{i} } 

1739: \langle  v^{0 \un{i} }  \,v'^{ \un{i} 0  }  \rangle  

1740:  {\rm  tr}_{\rm Ad}  \left\{ E_{0 \un{i}  }

1741: \hat{G} \,  E_{ \un{i} 0 }  \,\hat{G}'  \right\}   \\

1742: &=& {\rm i} \,g^2\, {\bf G}^{(e_{\rm a})} (z,z') \,

1743: \hat{\bG}^{(-e_{\rm a} )}\,\hat{\bG}'^{(0 )}

1744: \Bigg\{ 

1745: \sum\limits_{\un{i}, \, \un{j} }

1746: \sum\limits_{I }

1747: (E^{({\rm Ad})}_{0\un{i} })^I {}_ {0\un{j}} \,

1748: (E^{({\rm Ad})}_{\un{i} 0 })^{ 0 \un{j} } {}_ {I } 

1749: + \sum\limits_{\un{i} ,\,\un{j} }

1750: \sum\limits_{\un{k}\neq \un{l} }

1751: (E^{({\rm Ad})}_{0\un{i} })^{\un{k}\un{l} } {}_ {0\un{j}} \,

1752: (E^{({\rm Ad})}_{\un{i} 0 })^{ 0 \un{j} } {}_ {\un{k} \un{l} }

1753: \Bigg\}    \non \\

1754: &+& {\rm i} \,g^2\, {\bf G}^{(e_{\rm a})} (z,z') 

1755: \hat{\bG}^{(0 )}\,

1756: \hat{\bG}'^{(-e_{\rm a} )}

1757: \Bigg\{ 

1758: \sum\limits_{\un{i} ,\, \un{j}}

1759: \sum\limits_{I }

1760: (E^{({\rm Ad})}_{0\un{i} })^{0\un{j}}{}_I \,

1761: (E^{({\rm Ad})}_{\un{i} 0 })^I{}_{ 0 \un{j} } 

1762: + \sum\limits_{\un{i},\, \un{j} }

1763: \sum\limits_{\un{k}\neq \un{l} }

1764: (E^{({\rm Ad})}_{0\un{i} })^{0\un{j}}{}_{\un{k}\un{l} } \,

1765: (E^{({\rm Ad})}_{\un{i} 0 })^{\un{k}\un{l} }{}_{ 0 \un{j} } 

1766: \Bigg\}~. \non 

1767: \eea

1768: Using the symmetry properties of the structure constants, 

1769: the group-theoretical factors here can be related to those 

1770: which occur in eqs. (\ref{GT-1}) and (\ref{GT-2}):

1771: \bea

1772: \sum\limits_{\un{i}, \, \un{j} }

1773: \sum\limits_{I }

1774: (E^{({\rm Ad})}_{0\un{i} })^I {}_ {0\un{j}} \,

1775: (E^{({\rm Ad})}_{\un{i} 0 })^{ 0 \un{j} } {}_ {I } 

1776: &=& 

1777: \sum\limits_{\un{i} , \,\un{j}}

1778: \sum\limits_{I }

1779: (E^{({\rm Ad})}_{0\un{i} })^{0\un{j}}{}_I \,

1780: (E^{({\rm Ad})}_{\un{i} 0 })^I{}_{ 0 \un{j} } 

1781: =2(N-1)~,  \\

1782: \sum\limits_{\un{i} ,\, \un{j} }

1783: \sum\limits_{\un{k}\neq \un{l} }

1784: (E^{({\rm Ad})}_{0\un{i} })^{\un{k}\un{l} } {}_ {0\un{j}} \,

1785: (E^{({\rm Ad})}_{\un{i} 0 })^{ 0 \un{j} } {}_ {\un{k} \un{l} }

1786: &=& 

1787: \sum\limits_{\un{i} ,\,\un{j}}

1788: \sum\limits_{\un{k}\neq \un{l} }

1789: (E^{({\rm Ad})}_{0\un{i} })^{0\un{j}}{}_{\un{k}\un{l} } \,

1790: (E^{({\rm Ad})}_{\un{i} 0 })^{\un{k}\un{l} }{}_{ 0 \un{j} } 

1791: =(N-1)(N-2)~. \non 

1792: \eea

1793: 

1794: As a result, the third  term in  

1795: (\ref{I+II-ad3}) leads  to the following contribution 

1796: \be 

1797: {\rm i} \,N(N-1) \,g^2\, 

1798: {\bf G}^{(e_{\rm a})} (z,z') \,

1799: \Big\{ \hat{\bG}^{(-e_{\rm a} )}\,\hat{\bG}'^{(0 )}

1800: + \hat{\bG}^{(0 )}\,

1801: \hat{\bG}'^{(e_{\rm a} )} \Big\} ~.

1802: \label{I+II-ad5}

1803: \ee

1804: 

1805: On the base of the above considerations, 

1806: one can readily arrive at the final expression for

1807: $\G_{\rm I+II}$:

1808: \be

1809: \G_{\rm I+II} = N(N-1)\,

1810: \frac{g^2}{(4\p)^4} \, 

1811: \U  \Big\{ {1\over 3} 

1812: +  8 \, \hat{I}_{\rm I+II} \Big\} ~,

1813: \ee

1814: where 

1815: \bea

1816:  \hat{I}_{\rm I+II}  = \int\limits_{0}^{\infty} {\rm d}s_1 

1817: {\rm d}s_2

1818: {\rm d}s_3 \, 

1819: {\rm e}^{-[ s_1 +  s_2]} \, 

1820: \frac{ s_1^3 

1821: [ s_1 + s_2 ]^2}

1822: {[s_2s_3 +s_1s_2 +s_1 s_3]^{d/2 + 1} }

1823: \label{I-I+II-ad}

1824: \eea

1825: is a divergent integral

1826: in the limit $\ve = 4-d \to 0$. 

1827: This integral follows from 

1828: (\ref{I-I+II})

1829: in the limit  $e_{\rm a} =e_{\rm f}=1$

1830: or, equivalently, $N\to \infty$. 

1831: Therefore, the divergent part of $\hat{I}_{\rm I+II}$  

1832: can be read off from (\ref{I-I+II-div}),

1833: \be

1834: (\hat{I}_{\rm I+II})_{\rm div} = 

1835: - {1 \over \ve} ~.

1836: \ee

1837: 

1838: \subsection{Evaluation of $\G_{\rm III}$ }

1839: 

1840: The evaluation of $\G_{\rm III}$ is very similar 

1841: to that of $\G_{\rm I+II}$ just described. 

1842: Therefore, we simply give the final result: 

1843: \be

1844: \G_{\rm III} = {2 \over 3} \,N(N-1)\,

1845: \frac{g^2}{(4\p)^4} \,\U ~.

1846: \ee

1847: No divergences are present. 

1848: 

1849: \subsection{Evaluation of $\G_{\rm IV}$ }

1850: 

1851: It remains to evaluate $\G_{\rm IV}$ which 

1852: is determined by eq. (\ref{IV-mod}).

1853: In  supersymmetric dimensional regularization, 

1854: we have 

1855: \bea

1856: {1 \over 16} \, 

1857: {\bar \cD}^2  \cD'^2 

1858: \bG^{(e )}(z,z') \Big|_{z'=z} &=& 

1859: \frac{(e^2 \, \f {\bar \f})^{d/2 -1} }{(4\p)^{d/2} }\,

1860: \int\limits_{0}^{\infty}

1861: \frac{ {\rm d} s }{s^{d/2} } \, {\rm e}^{-s}~, \qquad 

1862: e \neq 0~, \non \\

1863: {1 \over 16} \, 

1864: {\bar \cD}^2  \cD'^2 

1865: \bG^{(0 )}(z,z') \Big|_{z'=z} &=& 

1866: 0~.

1867: \label{6-20}

1868: \eea

1869: The second relation here is actually a consequence 

1870: of one of the fancy properties of 

1871: dimensional regularization (see, e.g. \cite{Zinn})

1872: \be

1873: \int  \frac{ {\rm d}^d p }{p^2} = 0 \quad 

1874: \Longleftrightarrow \quad 

1875: \int\limits_{0}^{\infty}

1876: \frac{ {\rm d} s }{s^{d/2}} =0~.

1877: \label{SUSY-dim-reg}

1878: \ee

1879: Therefore, in the expression

1880: \be

1881: {\rm  tr}_{\rm Ad} \left\{ \{E_{0 \un{i} }, E_{\un{i} 0} \} \,

1882: {\bar \cD}^2  \cD'^2 G^{({\rm Ad} )}(z,z') \Big|_{z'=z}

1883: \right\}~,

1884: \ee

1885: which occurs in (\ref{IV-mod}), 

1886: we should take into account the massive modes only. 

1887: This amounts to computing the 

1888: following group-theoretic factor

1889: \be

1890: \sum\limits_{\un{i} , \, \un{j} } \left(

1891: \Big( \{E^{({\rm Ad})}_{0 \un{i} }, E^{({\rm Ad})}_{\un{i} 0} \} 

1892: \Big)^{0 \un{j} }{}_{0 \un{j} }

1893: + \Big( \{E^{({\rm Ad})}_{0 \un{i} }, E^{({\rm Ad})}_{\un{i} 0} \} 

1894: \Big)^{\un{j} 0}{}_{\un{j} 0} \right)= 2N(N-1)~.

1895: \ee

1896: As a result, we obtain 

1897: \be

1898: \G_{\rm IV} = -4 N(N-1)\,

1899: \frac{g^2}{(4\p)^4} \, 

1900: \U  \, I_{\rm IV}  ~,

1901: \ee

1902: with 

1903: $ I_{\rm IV}  $ given in (\ref{IV-int}).

1904: It is seen that the divergent parts of

1905: of $  \G_{\rm I+II} $ and $ \G_{\rm IV} $ cancel each other, 

1906: \be

1907: \Big( \G_{\rm I+II}  +

1908: \G_{\rm IV}  \Big)_{\rm div} = 0 ~,

1909: \ee

1910: consistent with the finiteness of the theory.

1911: An alternative treatment of the cancellation of divergences

1912: is given in the Appendix.

1913: 

1914: \sect{Discussion}

1915: As pointed out in the Introduction, the two $\cN =2$ superconformal

1916: field theories with gauge group $SU(N)$ considered 

1917: in this paper differ only in the hypermultiplet sector 

1918: --- one contains a single hypermultiplet in the

1919: adjoint representation, the other contains $2N$ hypermultiplets 

1920: in the fundamental representation. 

1921: If the Dine-Seiberg conjecture holds,

1922: then the two-loop $F^4$ contributions 

1923: to the effective action must vanish in both theories. 

1924: This would necessitate a cancellation 

1925: of the $F^4$ corrections between the pure $\cN=2$ SYM, ghost 

1926: and hypermultiplet sectors in both theories, implying that both 

1927: theories should yield identical two-loop $F^4$ contributions 

1928: in the hypermultiplet sector.

1929: 

1930: By explicit calculation, we have found  the following two-loop

1931: $F^4$ contributions, $ \G_{\rm I+II} +  \G_{\rm III} +\G_{\rm IV} $,

1932: from the hypermultiplet sector. For the case of

1933: $\cN=2$ SYM with 

1934: $2N$ hypermultiplets in the fundamental: 

1935: \bea

1936: \frac{g^2\, \U}{(4\p)^4} \,

1937: \Big\{ N(N-1)(N-2) &+&N(N-1) \non \\

1938: +  8N(N-1)  I_{\rm I+II} 

1939: &-& 4(N-1)  I_{\rm III} 

1940: -4(N-1)  (N -  2\,{N-1 \over N} )I_{\rm IV} \Big\} ~,

1941: \label{EA1}

1942: \eea

1943: where the integrals $I_{\rm I+II},$ $I_{\rm III}$ and $I_{\rm IV}$

1944: are given in equations (\ref{I-I+II}), (\ref{I-III}) and (\ref{IV-int})

1945: respectively. For the case of $\cN=4$ SYM:

1946: \be

1947: \frac{g^2\,\U}{(4\p)^4} \,

1948: \Big\{  N (N-1)

1949: +  8 N (N-1) \, \hat{I}_{\rm I+II}  

1950: -4  N(N-1) \,

1951:   I_{\rm IV}\Big\} ~,

1952: \label{EA2}

1953: \ee

1954: where the integrals $\hat{I}_{\rm I+II}$  and $I_{\rm IV}$

1955: are given in equations (\ref{I-I+II-ad}) and (\ref{IV-int})

1956: respectively.

1957: 

1958: In the large $N$ limit, all of the

1959: integrals contained in the expressions

1960: (\ref{EA1}) and (\ref{EA2})

1961:  become independent 

1962: of $N,$ as the charges $e_{\rm a}$ and $e_{\rm f}$

1963: approach 1. 

1964: With this observation, it is clear that these  

1965: $F^4$ contributions have different large $N$ behaviour. 

1966: The leading term in  (\ref{EA1}) 

1967: is of order $N^3,$ while the leading term

1968: in (\ref{EA2})  is of order $N^2.$

1969: This is inconsistent with the Dine-Seiberg conjecture, which would

1970: require identical leading large $N$ behaviour for the two theories.

1971: 

1972: 

1973: It is instructive to examine the source of the difference

1974: in the large $N$

1975: behaviour of the two theories, 

1976: which is due to the presence of a

1977:  leading $N^3$ contribution in the case 

1978: of $\cN=2$ SYM with $2N$

1979: hypermultiplets in the fundamental. This contribution

1980:  comes from  the diagrams of type I, II and III

1981: in which the $\cN=2$ vector multiplet 

1982: propagator (that is, $\langle v\, v \rangle$ 

1983: or $\langle \vf \, \vf^\dagger \rangle$)

1984: is massless and corresponds to one of the 

1985: unbroken $SU(N-1)$ generators $E_{\un{i} \un{j} }$,

1986: and the two hypermultiplet propagators are massive

1987: with the same mass $\bar{\phi} \phi/N(N-1)$.

1988: In the large $N$ limit, these

1989: hypermultiplets become massless

1990: and decouple from the background (as

1991: their $U(1)$ charges, $\pm 1/ \sqrt{ N(N-1))}$,  

1992: vanish), and so

1993: one might at first sight expect these diagrams not to

1994: contribute terms proportional to $\Upsilon$ in the

1995: large $N$ limit. However, the situation is more subtle, 

1996: because they really decouple only for $N= \infty$.

1997: The point is that the magnitude of the

1998: $U(1)$ charge on each of the hypermultiplet lines is the same. 

1999: Since all charge dependence occurs in the form $e W$ or $e \phi,$ 

2000: it cancels out of  the terms proportional to

2001: $W^2 \bar{W}^2/ (\f \bar{\phi} )^2$, see eqs. 

2002: (\ref{FUN-1}) and (\ref{FUN-2}).

2003: As a result, the contribution from these diagrams

2004:  survives in the large $N$ limit. 

2005: There is a combinatoric factor of $2N (N-1) (N-2),$ as

2006: there are $2N$ hypermultiplets and 

2007: $(N-1)(N-2)$ massless vectors

2008:  $v^{\un{i} \un{j}}.$

2009: 

2010: 

2011: There remains a (pretty solid)

2012: hope that the Dine-Seiberg conjecture

2013: holds, at least in the large $N$ limit, for those $\cN=2$ 

2014: superconformal theories which possess supergravity duals, 

2015: in particular: (i) $\cN=4$ SYM; 

2016: (ii) $USp(2N)$ gauge theory  with a traceless antisymmetric

2017: hypermultiplet and four fundamental hypermultiplets

2018: \cite{ASST}; (iii) quiver gauge theories \cite{quiver}. 

2019: This is based upon the AdS/CFT correspondence. 

2020: Maximal supersymmetry should also play a crucial role 

2021: in the case of $\cN=4$ SYM.

2022: Otherwise one would be forced to re-consider 

2023: numerous conclusions drawn on the basis of 

2024: this conjecture, for instance, in \cite{BPT,KMT}. 

2025: Explicit two-loop calculations of $F^4$ corrections

2026: in such theories are therefore 

2027: extremely desirable and can be 

2028: carried out using the techniques developed in the present 

2029: paper in conjunction with some ideas given in \cite{KMT}. 

2030: 

2031: \vskip.5cm

2032: 

2033: \noindent

2034: {\bf Acknowledgements:}\\

2035: We are grateful to Joseph Buchbinder, Jim Gates

2036: and Arkady Tseytlin for  comments.

2037: This work is supported in part by the Australian Research

2038: Council and  UWA research grants.

2039: 

2040: \begin{appendix}

2041: 

2042: \sect{Cancellation of divergences}

2043: To handle ill-defined two-loop  integrals, 

2044: we employed supersymmetric regularization 

2045: via dimensional reduction. Its use allowed us, 

2046: in a safe yet simple way, to make sure that 

2047: no divergences are present. On the other hand, 

2048: the absence of divergences indicates that there

2049: should exist a manifestly finite form for the effective 

2050: action  directly in four space-time dimensions.  

2051: Here we elaborate on such a form 

2052: in the case of $\cN=4$ SYM. 

2053: 

2054: The second and third terms in the two-loop 

2055: contribution (\ref{EA2}) contain the  

2056: proper-time integrals $\hat{I}_{\rm I+II} $

2057: (\ref{I-I+II-ad}) and $ I_{\rm IV} $  (\ref{IV-int}), 

2058: each of which diverges in $d=4$. 

2059: Nevertheless, let us try to evaluate 

2060: the joint contribution coming  from 

2061: the second and third terms in  (\ref{EA2})

2062: in $d=4$.  Since we no longer use 

2063: supersymmetric dimensional regularization, 

2064: the integral  $ I_{\rm IV} $ has to be

2065: modified as follows

2066: \be

2067: I_{\rm IV} \quad \longrightarrow \quad

2068: \hat{I}_{\rm IV}  = \int\limits_{0}^{\infty} 

2069: { {\rm d}s \over s^{d/2} } \,  {\rm e}^{-s} 

2070: +\int\limits_{0}^{\infty} 

2071: { {\rm d}s \over s^{d/2} }~. 

2072: \label{IV-int-til}

2073: \ee

2074: The second term here 

2075: is generated by those supergraphs

2076: of type IV

2077: which involve the structure in the second line 

2078: of (\ref{6-20}). This term 

2079: cannot be ignored anymore, 

2080: since the identity (\ref{SUSY-dim-reg})

2081: holds only in the framework of 

2082: supersymmetric dimensional regularization.

2083: The sum of divergent integrals is 

2084: \bea 

2085: \Big( 2\hat{I}_{\rm I+II} - 

2086: \hat{I}_{\rm IV} \Big)\Big|_{d=4} 

2087: = 2\int\limits_{0}^{\infty} {\rm d}s \,

2088: {\rm d}t \,

2089: {\rm d}u \, 

2090: \frac{ t^3 \,

2091: ( s + t )^2}

2092: {(st  +su +tu)^{3 }}\,

2093: {\rm e}^{-( s +  t)} 

2094: -  \int\limits_{0}^{\infty} 

2095: { {\rm d}s \over s^2 }

2096: \,  {\rm e}^{-s} 

2097: -  \int\limits_{0}^{\infty} 

2098: { {\rm d}s \over s^2 }

2099: ~.

2100: \eea

2101: In the first term on the right, 

2102: one can easily do the $u$-integral:

2103: \bea

2104: \int\limits_{0}^{\infty} {\rm d}s \,

2105: {\rm d}t \,

2106: {\rm d}u \, 

2107: \frac{ t^3 \,

2108: ( s + t )^2}

2109: {(st  +su +tu)^{3 }}\,

2110: {\rm e}^{-( s +  t)} 

2111: &=& \hf   \int\limits_{0}^{\infty} {\rm d}s \,

2112: {\rm d}t \,

2113: \frac{ t \,

2114: ( s + t )}

2115: {s ^{2 }}\,

2116: {\rm e}^{-( s +  t)} \non \\

2117: = \hf   \int\limits_{0}^{\infty} {\rm d}t \,

2118: t\,{\rm e}^{-t} 

2119: \int\limits_{0}^{\infty} { {\rm d}s \over s} \,  

2120: {\rm e}^{-s} 

2121: &+& 

2122: \hf   \int\limits_{0}^{\infty} {\rm d}t \,

2123: t^2 \,{\rm e}^{-t} 

2124: \int\limits_{0}^{\infty} { {\rm d}s \over s^2} \,  

2125: {\rm e}^{-s} ~.

2126: \eea

2127: As a result, one gets 

2128: \bea 

2129: \Big( 2\hat{I}_{\rm I+II} - \hat{I}_{\rm IV} \Big)\Big|_{d=4} 

2130: = \int\limits_{0}^{\infty}  {\rm d}s \,

2131: { {\rm d} \over {\rm d}  s} \,

2132: \left\{

2133: \frac{1}{s} \,  ( 1 - {\rm e}^{-s} ) \right\}

2134: = -1~.

2135: \eea

2136: %\be 

2137: %\Big( 2\hat{I}_{\rm I+II} - I_{\rm IV} \Big)\Big|_{d=4} 

2138: %= \G(0) + 2\,\G(-1)  -\G(-1) = 0~, 

2139: %\ee

2140: %if  the well-known 

2141: %property $\G(x+1) = x\,  \G(x)$ of the Euler 

2142: %gamma-function is applied 

2143: %to formally represent $\G(0) = -\G(-1)$.

2144: This shows that the second and third terms 

2145: in  (\ref{EA2}) provide a finite contribution

2146: to the effective action.

2147: 

2148: \end{appendix}

2149: 

2150: \begin{thebibliography}{99}

2151: 

2152: \bibitem{KM}

2153: S.~M.~Kuzenko and I.~N.~McArthur,

2154: ``On the background field method beyond one loop: 

2155: A manifestly covariant  derivative expansion 

2156: in super Yang-Mills theories,''

2157: JHEP {\bf 0305} (2003) 015 [arXiv:hep-th/0302205].

2158: 

2159: \bibitem{KM2}

2160: S.~M.~Kuzenko and I.~N.~McArthur, 

2161: ``Low-energy dynamics in N = 2 super QED: Two-loop approximation,''

2162: JHEP {\bf 0310} (2003) 029

2163: [arXiv:hep-th/0308136].

2164: 

2165: \bibitem{BKO}

2166: I.~L.~Buchbinder, S.~M.~Kuzenko and B.~A.~Ovrut,

2167: ``On the D = 4, N = 2 non-renormalization theorem,''

2168: Phys.\ Lett.\ B {\bf 433} (1998) 335

2169: [arXiv:hep-th/9710142].

2170: 

2171: \bibitem{BBKO}

2172: I.~L.~Buchbinder, E.~I.~Buchbinder, S.~M.~Kuzenko and B.~A.~Ovrut,

2173: ``The background field method for N = 2 super Yang-Mills theories 

2174: in  harmonic superspace,''

2175: Phys.\ Lett.\ B {\bf 417} (1998) 61

2176: [arXiv:hep-th/9704214].

2177: 

2178: \bibitem{DS}

2179: M.~Dine and N.~Seiberg,

2180: ``Comments on higher derivative operators in some 

2181: SUSY field theories,''

2182: Phys.\ Lett.\ B {\bf 409} (1997) 239

2183: [arXiv:hep-th/9705057].

2184: 

2185: \bibitem{dWGR}

2186: B.~de Wit, M.~T.~Grisaru and M.~Ro\v{c}ek,

2187: ``Nonholomorphic corrections to the one-loop 

2188: N=2 super Yang-Mills action,''

2189: Phys.\ Lett.\ B {\bf 374} (1996) 297

2190: [arXiv:hep-th/9601115].

2191: 

2192: \bibitem{Hen}

2193: M.~Henningson,

2194: ``Extended superspace, higher derivatives 

2195: and SL(2,Z) duality,''

2196: Nucl.\ Phys.\ B {\bf 458} (1996) 445

2197: [arXiv:hep-th/9507135].

2198: 

2199: \bibitem{BKT}

2200: I.~L.~Buchbinder, S.~M.~Kuzenko and A.~A.~Tseytlin,

2201: ``On low-energy effective actions in N = 2,4 superconformal theories 

2202: in  four dimensions,''

2203: Phys.\ Rev.\ D {\bf 62} (2000) 045001

2204: [arXiv:hep-th/9911221].

2205: 

2206: \bibitem{GR}

2207: F.~Gonzalez-Rey and M.~Ro\v{c}ek,

2208: ``Nonholomorphic N = 2 terms in N = 4 SYM: 

2209: 1-loop calculation in N = 2  superspace,''

2210: Phys.\ Lett.\ B {\bf 434} (1998) 303

2211: [arXiv:hep-th/9804010];

2212: F.~Gonzalez-Rey, B.~Kulik, I.~Y.~Park and M.~Ro\v{c}ek,

2213: ``Self-dual effective action of N = 4 super-Yang-Mills,''

2214: Nucl.\ Phys.\ B {\bf 544} (1999) 218

2215: [arXiv:hep-th/9810152].

2216: 

2217: \bibitem{BK2}

2218: I.~L.~Buchbinder and S.~M.~Kuzenko,

2219: ``Comments on the background field method 

2220: in harmonic superspace:  

2221: Non-holomorphic corrections in N = 4 SYM,''

2222: Mod.\ Phys.\ Lett.\ A {\bf 13} (1998) 1623

2223: [arXiv:hep-th/9804168];

2224: E.~I.~Buchbinder, I.~L.~Buchbinder and S.~M.~Kuzenko,

2225: ``Non-holomorphic effective potential in N = 4 SU(n) 

2226: SYM,'' Phys.\ Lett.\ B {\bf 446} (1999) 216

2227: [arXiv:hep-th/9810239].

2228: 

2229: \bibitem{LvU}

2230: D.~A.~Lowe and R.~von Unge,

2231: ``Constraints on higher derivative operators 

2232: in maximally supersymmetric  gauge theory,''

2233: JHEP {\bf 9811} (1998) 014

2234: [arXiv:hep-th/9811017].

2235: 

2236: \bibitem{DKMSW}

2237: N.~Dorey, V.~V.~Khoze, M.~P.~Mattis, M.~J.~Slater and W.~A.~Weir,

2238: ``Instantons, higher-derivative terms, 

2239: and nonrenormalization theorems 

2240: in  supersymmetric gauge theories,''

2241: Phys.\ Lett.\ B {\bf 408} (1997) 213

2242: [arXiv:hep-th/9706007].

2243: 

2244: \bibitem{BFMT}

2245: D.~Bellisai, F.~Fucito, M.~Matone and G.~Travaglini,

2246: ``Non-holomorphic terms in N = 2 SUSY 

2247: Wilsonian actions and RG equation,''

2248: Phys.\ Rev.\ D {\bf 56} (1997) 5218

2249: [arXiv:hep-th/9706099].

2250: 

2251: \bibitem{BK} I.~L.~Buchbinder and S.~M.~Kuzenko,

2252: {\it Ideas and Methods of Supersymmetry and

2253: Supergravity or a Walk Through Superspace},

2254: IOP, Bristol, 1998.

2255: 

2256: \bibitem{GGRS}

2257: S.~J.~Gates, M.~T.~Grisaru, M.~Ro\v{c}ek and W.~Siegel,

2258: {\it Superspace, or One Thousand 

2259: and One Lessons in Supersymmetry},

2260: Benjamin/Cummings, 1983 [arXiv:hep-th/0108200].

2261: 

2262: \bibitem{OW}

2263: B.~A.~Ovrut and J.~Wess,

2264: ``Supersymmetric $R_\x$ gauge and radiative symmetry breaking,''

2265: Phys.\ Rev.\ D {\bf 25} (1982) 409;

2266: P.~Binetruy, P.~Sorba and R.~Stora,

2267: ``Supersymmetric S covariant $R_\x$ gauge,''

2268: Phys.\ Lett.\ B {\bf 129} (1983) 85.

2269: 

2270: \bibitem{BBP}

2271: A.~T.~Banin, I.~L.~Buchbinder and N.~G.~Pletnev,

2272: ``On low-energy effective action in N = 2 super 

2273: Yang-Mills theories on  non-abelian background,''

2274: Phys.\ Rev.\ D {\bf 66} (2002) 045021

2275: [arXiv:hep-th/0205034];

2276: ``One-loop effective action for N = 4 SYM theory 

2277: in the hypermultiplet  sector: Leading low-energy 

2278: approximation and beyond,''

2279: Phys.\ Rev.\ D {\bf 68} (2003) 065024

2280: [arXiv:hep-th/0304046].

2281: 

2282: \bi{Georgi} H. Georgi, {\it Lie Algebras in Particle Physics: From 

2283: Isospin to Unified Theories}, Benjamin/Cummings, 1982.

2284: 

2285: \bi{Zinn} J. Zinn-Justin, {\it Quantum Field Theory and 

2286: Critical Phenomena}, Oxford University Press, 1989.

2287: 

2288: \bibitem{ASST}

2289: O.~Aharony, J.~Sonnenschein, S.~Yankielowicz and S.~Theisen,

2290: ``Field theory questions for string theory answers,''

2291: Nucl.\ Phys.\ B {\bf 493} (1997) 177

2292: [arXiv:hep-th/9611222]; 

2293: M.~R.~Douglas, D.~A.~Lowe and J.~H.~Schwarz,

2294: ``Probing F-theory with multiple branes,''

2295: Phys.\ Lett.\ B {\bf 394} (1997) 297

2296: [arXiv:hep-th/9612062];

2297: O.~Aharony, J.~Pawelczyk, S.~Theisen and S.~Yankielowicz,

2298: ``A note on anomalies in the AdS/CFT correspondence,''

2299: Phys.\ Rev.\ D {\bf 60} (1999) 066001

2300: [arXiv:hep-th/9901134].

2301: 

2302: \bibitem{quiver}

2303: M.~R.~Douglas and G.~W.~Moore,

2304: ``D-branes, quivers, and ALE instantons,''

2305: arXiv:hep-th/9603167; 

2306: C.~V.~Johnson and R.~C.~Myers,

2307: ``Aspects of type IIB theory on ALE spaces,''

2308: Phys.\ Rev.\ D {\bf 55} (1997) 6382

2309: [arXiv:hep-th/9610140];

2310: S.~Kachru and E.~Silverstein,

2311: ``4d conformal theories and strings on orbifolds,''

2312: Phys.\ Rev.\ Lett.\  {\bf 80} (1998) 4855

2313: [arXiv:hep-th/9802183].

2314: 

2315: \bibitem{BPT}

2316: I.~L.~Buchbinder, A.~Y.~Petrov and A.~A.~Tseytlin,

2317: ``Two-loop N = 4 super Yang Mills effective action 

2318: and interaction  between D3-branes,''

2319: Nucl.\ Phys.\ B {\bf 621} (2002) 179

2320: [arXiv:hep-th/0110173].

2321: 

2322: \bibitem{KMT}

2323: S.~M.~Kuzenko, I.~N.~McArthur and S.~Theisen,

2324: ``Low energy dynamics from deformed conformal symmetry 

2325: in quantum 4D N = 2 SCFTs,''

2326: Nucl.\ Phys.\ B {\bf 660} (2003) 131

2327: [arXiv:hep-th/0210007].

2328: 

2329: \end{thebibliography}

2330: 

2331: 

2332: \end{document}

2333: 

2334: %21.10.03