1: \input harvmac

2: \input epsf

3: %\draftmode

4: 

5: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

6: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

7: %	Emil Martinec's macros

8: %

9: %

10: \def\nextline{\hfil\break}

11: 

12: 

13: \noblackbox

14: %

15: 

16: % Something to deal with sub-sub-sections

17: 

18: \def\unlockat{\catcode`\@=11}

19: \def\lockat{\catcode`\@=12}

20: 

21: \unlockat

22: % Something to deal with sub-sub-sections

23: 

24: \def\newsec#1{\global\advance\secno by1\message{(\the\secno. #1)}

25: \global\subsecno=0\global\subsubsecno=0\eqnres@t\noindent

26: {\bf\the\secno. #1}

27: \writetoca{{\secsym} {#1}}\par\nobreak\medskip\nobreak}

28: %

29: \global\newcount\subsecno \global\subsecno=0

30: \def\subsec#1{\global\advance\subsecno

31: by1\message{(\secsym\the\subsecno. #1)}

32: \ifnum\lastpenalty>9000\else\bigbreak\fi\global\subsubsecno=0

33: \noindent{\it\secsym\the\subsecno. #1}

34: \writetoca{\string\quad {\secsym\the\subsecno.} {#1}}

35: \par\nobreak\medskip\nobreak}

36: %

37: \global\newcount\subsubsecno \global\subsubsecno=0

38: \def\subsubsec#1{\global\advance\subsubsecno by1

39: \message{(\secsym\the\subsecno.\the\subsubsecno. #1)}

40: \ifnum\lastpenalty>9000\else\bigbreak\fi

41: \noindent\quad{\secsym\the\subsecno.\the\subsubsecno.}{#1}

42: \writetoca{\string\qquad{\secsym\the\subsecno.\the\subsubsecno.}{#1}}

43: \par\nobreak\medskip\nobreak}

44: 

45: \def\subsubseclab#1{\DefWarn#1\xdef

46: #1{\noexpand\hyperref{}{subsubsection}%

47: {\secsym\the\subsecno.\the\subsubsecno}%

48: {\secsym\the\subsecno.\the\subsubsecno}}%

49: \writedef{#1\leftbracket#1}\wrlabeL{#1=#1}}% Macros for boxes

50: \lockat

51: 

52: %

53: 

54: \def\ie{{\it i.e.}}

55: \def\eg{{\it e.g.}}

56: \def\cf{{\it c.f.}}

57: \def\etal{{\it et.al.}}

58: \def\etc{{\it etc.}}

59: 

60: \def\gpg{g^{-1} \p g}

61: \def\pgp{\pb g g^{-1}}

62: \def\hone{h^{(1)} }

63: 

64: %% MORE MACROS

65: \def\CM {{\cal M}}

66: \def\CN {{\cal N}}

67: \def\cO{{\cal O}}

68: \def\CO {{\cal O}}

69: \def\codim{{\mathop{\rm codim}}}

70: \def\cok{{\rm cok}}

71: \def\coker{{\mathop {\rm coker}}}

72: \def\cP{{\cal P}}

73: \def\CP {{\cal P }}

74: \def\CQ {{\cal Q }}

75: \def\CE{{\cal E }}

76: \def\CV{{\cal V }}

77: \def\CZ {{\cal Z }}

78: \def\CS {{\cal S }}

79: \def\ch{{\rm ch}}

80: \def\CY{{\cal Y }}

81: \def\diff{{\rm diff}}

82: \def\Diff{{\rm Diff}}

83: \def\Det{{\rm Det}}

84: \def\gpg{g^{-1} \p g}

85: \def\ib{{\bar i}}

86: \def\zb {\bar{z}}

87: \def\wb {\bar{w}}

88: \font\manual=manfnt \def\dbend{\lower3.5pt\hbox{\manual\char127}}

89: \def\danger#1{\smallskip\noindent\rlap\dbend%

90: \indent{\bf #1}\par\vskip-1.5pt\nobreak}

91: \def\c{\cdot}

92: \def\IZ{\relax\ifmmode\mathchoice

93: {\hbox{\cmss Z\kern-.4em Z}}{\hbox{\cmss Z\kern-.4em Z}}

94: {\lower.9pt\hbox{\cmsss Z\kern-.4em Z}}

95: {\lower1.2pt\hbox{\cmsss Z\kern-.4em Z}}\else{\cmss Z\kern-.4em

96: Z}\fi}

97: \def\half{{1\over 2}}

98: \def\sdtimes{\mathbin{\hbox{\hskip2pt\vrule height 4.1pt depth -.3pt

99: width

100: .25pt

101: \hskip-2pt$\times$}}}

102: \def\p{\partial}

103: \def\pb{\bar{\partial}}

104: 

105: \def\clb#1#2#3#4#5#6{\pmatrix{#1 & #2 & #3\cr

106: #4 & #5 & #6\cr} }

107: 

108: \def\CJ {{\cal J}}

109: \def\CM {{\cal M}}

110: \def\CN {{\cal N}}

111: \def\cO{{\cal O}}

112: \def\CO {{\cal O}}

113: \def\CX{{\cal X}}

114: \def\codim{{\mathop{\rm codim}}}

115: \def\cok{{\rm cok}}

116: \def\coker{{\mathop {\rm coker}}}

117: \def\cP{{\cal P}}

118: \def\CP {{\cal P }}

119: \def\CQ {{\cal Q }}

120: \def\CE{{\cal E }}

121: \def\CV{{\cal V }}

122: \def\CZ {{\cal Z }}

123: \def\CS {{\cal S }}

124: \def\ch{{\rm ch}}

125: \def\Det{{\rm Det}}

126: \def\DET{{\rm DET}}

127: 

128: \def\gpg{g^{-1} \p g}

129: \def\ib{{\bar i}}

130: 

131: % more macros, alphabetically

132: 

133: \def\Aut{{\rm Aut}}

134: 

135: \def\Hom{{\rm Hom}}

136: \def\imp{$\Rightarrow$}

137: \def\IZ{\relax\ifmmode\mathchoice

138: {\hbox{\cmss Z\kern-.4em Z}}{\hbox{\cmss Z\kern-.4em Z}}

139: {\lower.9pt\hbox{\cmsss Z\kern-.4em Z}}

140: {\lower1.2pt\hbox{\cmsss Z\kern-.4em Z}}\else{\cmss Z\kern-.4em

141: Z}\fi}

142: \def\IB{\relax{\rm I\kern-.18em B}}

143: \def\IC{{\relax\hbox{$\inbar\kern-.3em{\rm C}$}}}

144: \def\ID{\relax{\rm I\kern-.18em D}}

145: \def\IE{\relax{\rm I\kern-.18em E}}

146: \def\IF{\relax{\rm I\kern-.18em F}}

147: \def\IG{\relax\hbox{$\inbar\kern-.3em{\rm G}$}}

148: \def\IGa{\relax\hbox{${\rm I}\kern-.18em\Gamma$}}

149: \def\IH{\relax{\rm I\kern-.18em H}}

150: \def\II{\relax{\rm I\kern-.18em I}}

151: \def\IK{\relax{\rm I\kern-.18em K}}

152: \def\IP{\relax{\rm I\kern-.18em P}}

153: \def\IQ{\relax\hbox{$\inbar\kern-.3em{\rm Q}$}}

154: \def\IX{{\bf X}}

155: 

156: \def\jb{{\bar j}}

157: \def\lieg{{\underline{\bf g}}}

158: \def\liet{{\underline{\bf t}}}

159: \def\liek{{\underline{\bf k}}}

160: \def\inbar{\,\vrule height1.5ex width.4pt depth0pt}

161: \def\Map{{\rm Map}}

162: 

163: \def\mod{{\rm mod}}

164: \def\p{\partial}

165: \def\pb{{\bar \p}}

166: \def\pgp{\pb g g^{-1}}

167: \def\Pic{{\rm Pic}}

168: 

169: \font\cmss=cmss10 \font\cmsss=cmss10 at 7pt

170: \def\IR{\relax{\rm I\kern-.18em R}}

171: \def\pbar{\bar{\p}}

172: \def\sdtimes{\mathbin{\hbox{\hskip2pt\vrule

173: height 4.1pt depth -.3pt width .25pt\hskip-2pt$\times$}}}

174: \def\Tr{{\rm Tr}}

175: \def\vol{{\rm vol}}

176: \def\Vol{{\rm Vol}}

177: \def\wb{{\bar{w}}}

178: \def\wzwt{$WZW_2$}

179: \def\wzwf{$WZW_4$}

180: \def\ymt{$YM_2$}

181: \def\zb {{\bar{z}}}

182: % Macros for boxes

183: %

184: %\def\boxit#1{\vbox{\hrule\hbox{\vrule\kern8pt

185: %\vbox{\hbox{\kern8pt}\hbox{\vbox{#1}}\hbox{\k

186: %\hbox{$\displaystyle #1$}\kern8pt}\kern8pt\vrule}\hrule}}}

187: %

188: %%% MACROS FOR BOX BOUNDARY CONDS

189: %%% FROM KAWAI ET AL

190: 

191: %\def\makeblankbox#1#2{\hbox{\lower\dp0\vbox{\hidehrule{#1}{#2}%

192: %   \kern -#1% overlap rules

193: %   \hbox to \wd0{\hidevrule{#1}{#2}%

194: %      \raise\ht0\vbox to #1{}% vrule height

195: %      \lower\dp0\vtop to #1{}% vrule depth

196: %      \hfil\hidevrule{#2}{#1}}%

197: %   \kern-#1\hidehrule{#2}{#1}}}%

198: %}%

199: %\def\hidehrule#1#2{\kern-#1\hrule height#1 depth#2 \kern-#2}%

200: %\def\hidevrule#1#2{\kern-#1{\dimen0=#1\advance\dimen0 by #2\vrule

201: %    width\dimen0}\kern-#2}%

202: %\def\openbox{\ht0=1.2mm \dp0=1.2mm \wd0=2.4mm  \raise 2.75pt

203: %\makeblankbox {.25pt} {.25pt}  }

204: %\def\qed{\hskip 8mm \openbox}

205: %\def\abs#1{\left\vert #1 \right\vert}

206: %\def\bun#1/#2{\leavevmode

207: %   \kern.1em \raise .5ex \hbox{\the\scriptfont0 #1}%

208: %   \kern-.1em $/$%

209: %   \kern-.15em \lower .25ex \hbox{\the\scriptfont0 #2}%

210: %}

211: %\def\row#1#2{#1_1,\ldots,#1_#2}

212: %\def\apar{\noalign{\vskip 2mm}}

213: %\def\blackbox{\hbox{\vrule height .5ex width .3ex depth -.3ex}}

214: %\def\nord{{\textstyle {\blackbox\atop\blackbox}}}

215: %\def\ts{\,}

216: %\def\opensquare{\ht0=3.4mm \dp0=3.4mm \wd0=6.8mm  \raise 2.7pt

217: %\makeblankbox {.25pt} {.25pt}  }

218: 

219: %%%%%%%%%%%%%%%%%%%%%%%

220: 

221: \def\sector#1#2{\ {\scriptstyle #1}\hskip 1mm

222: \mathop{\opensquare}\limits_{\lower 1mm\hbox{$\scriptstyle#2$}}\hskip 1mm}

223: 

224: \def\tsector#1#2{\ {\scriptstyle #1}\hskip 1mm

225: \mathop{\opensquare}\limits_{\lower 1mm\hbox{$\scriptstyle#2$}}^\sim\hskip 1mm}

226: %%%

227: %%%

228: 

229: %% ANOTHER SET OF MACROS

230: 

231: \def\jb{{\bar j}}

232: \def\Lie{{\rm Lie}}

233: \def\lieg{{\underline{\bf g}}}

234: \def\lieh{{\underline{\bf h}}}

235: \def\liet{{\underline{\bf t}}}

236: \def\liek{{\underline{\bf k}}}

237: \def\inbar{\,\vrule height1.5ex width.4pt depth0pt}

238: \def\ndt{\noindent}

239: \def\p{\partial}

240: \def\pab{\pb_{\bar A} }

241: \def\pb{{\bar \p}}

242: \def\pgp{\pb g g^{-1}}

243: \font\cmss=cmss10 \font\cmsss=cmss10 at 7pt

244: \def\IR{\relax{\rm I\kern-.18em R}}

245: \def\pbar{\bar{\p}}

246: \def\rank{{\rm rank}}

247: \def\sdtimes{\mathbin{\hbox{\hskip2pt\vrule

248: height 4.1pt depth -.3pt width .25pt\hskip-2pt$\times$}}}

249: \def\ub{{\bar{u}}}

250: \def\vol{{\rm vol}}

251: \def\Vol{{\rm Vol}}

252: \def\wb{{\bar{w}}}

253: \def\wzwt{$WZW_2$}

254: \def\wzwf{$WZW_4$}

255: \def\ymt{$YM_2$}

256: \def\zb {{\bar{z}}}

257: 

258: 

259: %% new macros

260: 

261: \def\sst{\scriptscriptstyle}

262: \def\tst#1{{\textstyle #1}}

263: \def\frac#1#2{{#1\over#2}}

264: \def\coeff#1#2{{\textstyle{#1\over #2}}}

265: \def\half{\frac12}

266: \def\hf{{\textstyle\half}}

267: \def\ket#1{|#1\rangle}

268: \def\bra#1{\langle#1|}

269: \def\vev#1{\langle#1\rangle}

270: \def\d{\partial}

271: 

272: \def\inbar{\,\vrule height1.5ex width.4pt depth0pt}

273: \def\IC{\relax\hbox{$\inbar\kern-.3em{\rm C}$}}

274: \def\IR{\relax{\rm I\kern-.18em R}}

275: \def\IP{\relax{\rm I\kern-.18em P}}

276: \def\Z{{\bf Z}}

277: \def\One{{1\hskip -3pt {\rm l}}}

278: \def\nth{$n^{\rm th}$}

279: %

280: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

281: %

282: \catcode`\@=11

283: \def\slash#1{\mathord{\mathpalette\c@ncel{#1}}}

284: \overfullrule=0pt

285: \def\AA{{\cal A}}

286: \def\BB{{\cal B}}

287: \def\CC{{\cal C}}

288: \def\DD{{\cal D}}

289: \def\EE{{\cal E}}

290: \def\FF{{\cal F}}

291: \def\GG{{\cal G}}

292: \def\HH{{\cal H}}

293: \def\II{{\cal I}}

294: \def\JJ{{\cal J}}

295: \def\KK{{\cal K}}

296: \def\LL{{\cal L}}

297: \def\MM{{\cal M}}

298: \def\NN{{\cal N}}

299: \def\OO{{\cal O}}

300: \def\PP{{\cal P}}

301: \def\QQ{{\cal Q}}

302: \def\RR{{\cal R}}

303: \def\SS{{\cal S}}

304: \def\TT{{\cal T}}

305: \def\UU{{\cal U}}

306: \def\VV{{\cal V}}

307: \def\WW{{\cal W}}

308: \def\XX{{\cal X}}

309: \def\YY{{\cal Y}}

310: \def\ZZ{{\cal Z}}

311: \def\lam{\lambda}

312: \def\eps{\epsilon}

313: \def\vareps{\varepsilon}

314: \def\underrel#1\over#2{\mathrel{\mathop{\kern\z@#1}\limits_{#2}}}

315: \def\lapprox{{\underrel{\scriptstyle<}\over\sim}}

316: \def\lessapprox{{\buildrel{<}\over{\scriptstyle\sim}}}

317: \catcode`\@=12

318: 

319: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

320: \def\sdtimes{\mathbin{\hbox{\hskip2pt\vrule height 4.1pt depth -.3pt width

321: .25pt \hskip-2pt$\times$}}}

322: %

323: \def\bra#1{\left\langle #1\right|}

324: \def\ket#1{\left| #1\right\rangle}

325: \def\vev#1{\left\langle #1 \right\rangle}

326: \def\det{{\rm det}}

327: \def\tr{{\rm tr}}

328: \def\mod{{\rm mod}}

329: \def\sinh{{\rm sinh}}

330: \def\cosh{{\rm cosh}}

331: \def\sgn{{\rm sgn}}

332: \def\det{{\rm det}}

333: \def\exp{{\rm exp}}

334: \def\sh{{\rm sh}}

335: \def\ch{{\rm ch}}

336: 

337: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

338: % new defs:

339: \def\st{\scriptstyle}

340: \def\gop{g_{\rm op}}

341: \def\gcl{g_{\rm cl}}

342: \def\vt#1#2#3{ {\vartheta[{#1 \atop  #2}](#3\vert \tau)} }

343: \def\ceff{c_{\rm eff}}

344: \def\l{\ell}

345: \def\vv{{\bf v}}

346: \def\tc{{\tilde c}}

347: \def\tx{{\tilde x}}

348: \def\ra{{\rightarrow}}

349: \def\HJ{{\rm Hirzebruch-Jung}}

350: \def\mn{{\rm mod}\; n}

351: \def\Xbar{{\bar X}}

352: \def\eff{{\rm eff}}

353: 

354: 

355: 

356: 

357: %%%%%%%%%%%%%%%%%%%%

358: %%%%%%%%%%%%%%%%%%%% SOME OF GREG'S MACROS

359: %%%%%%%%%%%%%%%%%%%%

360: %%%%%%%%%%%%%%%%%%%%

361: 

362: 

363:  \def\p{\partial}

364: %\def\IC{{\bf C}}

365: %\def\IR{{\bf R}}

366: 

367: %\def\IC{{\bf C}}

368: \def\a{\alpha}

369: \def\b{\beta}

370: \def\e{\epsilon}

371: \def\CS{{\cal S}}

372: \def\CV{{\cal V}}

373: \def\mod{{\rm mod}}

374: 

375: \def\re{{\rm Re}}

376: 

377: 

378:  \def\p{\partial}

379: %\def\IC{{\bf C}}

380: %\def\IR{{\bf R}}

381: 

382: %\def\IC{{\bf C}}

383: %\def\IZ{{\bf Z}}

384: \def\a{\alpha}

385: \def\b{\beta}

386: \def\e{\epsilon}

387: \def\CE{{\cal E}}

388: \def\CS{{\cal S}}

389: \def\CU{{\cal U}}

390: \def\CV{{\cal V}}

391: \def\mod{{\rm mod}}

392: 

393: \def\re{{\rm Re}}

394: \def\I{{\cal I}}

395: \def\It{{\tilde {\cal I}}}

396: 

397: 

398: 

399: %

400: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

401: 

402: %% END MACROS

403: %%

404: 

405: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

406: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

407: %%%

408: %%% References

409: %%%

410: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

411: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

412: 

413: 

414: %\AdamsSV

415: \lref\aps{

416: A.~Adams, J.~Polchinski and E.~Silverstein,

417: ``Don't panic! Closed string tachyons in ALE space-times,''

418: JHEP {\bf 0110}, 029 (2001)

419: [arXiv:hep-th/0108075].

420: %%CITATION = HEP-TH 0108075;%%

421: }

422: 

423: 

424: %\HarveyWM

425: \lref\hkmm{

426: J.~A.~Harvey, D.~Kutasov, E.~J.~Martinec and G.~Moore,

427: ``Localized tachyons and RG flows,''

428: arXiv:hep-th/0111154.

429: %%CITATION = HEP-TH 0111154;%%

430: }

431: 

432: %\DixonQV

433: \lref\DixonQV{

434: L.~J.~Dixon, D.~Friedan, E.~J.~Martinec and S.~H.~Shenker,

435: ``The Conformal Field Theory Of Orbifolds,''

436: Nucl.\ Phys.\ B {\bf 282}, 13 (1987).

437: %%CITATION = NUPHA,B282,13;%%

438: }

439: 

440: %\MartinecWG

441: \lref\mm{

442: E.~J.~Martinec and G.~Moore,

443: ``On decay of K-theory,''

444: arXiv:hep-th/0212059.

445: %%CITATION = HEP-TH 0212059;%%

446: }

447: 

448: %\VafaRA

449: \lref\VafaRA{

450: C.~Vafa,

451: ``Mirror symmetry and closed string tachyon condensation,''

452: arXiv:hep-th/0111051.

453: %%CITATION = HEP-TH 0111051;%%

454: }

455: 

456: \lref\mt{

457: %\MinwallaHJ

458: %\lref\MinwallaHJ{

459: S.~Minwalla and T.~Takayanagi,

460: ``Evolution of D-branes under closed string tachyon condensation,''

461: JHEP {\bf 0309}, 011 (2003)

462: [arXiv:hep-th/0307248].

463: %%CITATION = HEP-TH 0307248;%%

464: }

465: 

466: 

467: 

468: %\DavidVM

469: \lref\DavidVM{

470: J.~R.~David, M.~Gutperle, M.~Headrick and S.~Minwalla,

471: ``Closed string tachyon condensation on twisted circles,''

472: JHEP {\bf 0202}, 041 (2002)

473: [arXiv:hep-th/0111212].

474: %%CITATION = HEP-TH 0111212;%%

475: }

476: 

477: %\DouglasHQ

478: \lref\DouglasHQ{

479: M.~R.~Douglas and B.~Fiol,

480: ``D-branes and discrete torsion. II,''

481: arXiv:hep-th/9903031.

482: %%CITATION = HEP-TH 9903031;%%

483: }

484: %\BerkoozIS

485: \lref\BerkoozIS{

486: M.~Berkooz and M.~R.~Douglas,

487: ``Five-branes in M(atrix) theory,''

488: Phys.\ Lett.\ B {\bf 395}, 196 (1997)

489: [arXiv:hep-th/9610236].

490: %%CITATION = HEP-TH 9610236;%%

491: }

492: %\BerkoozKM

493: \lref\BerkoozKM{

494: M.~Berkooz, M.~R.~Douglas and R.~G.~Leigh,

495: ``Branes intersecting at angles,''

496: Nucl.\ Phys.\ B {\bf 480}, 265 (1996)

497: [arXiv:hep-th/9606139].

498: %%CITATION = HEP-TH 9606139;%%

499: }

500: %\BilloYB

501: \lref\bcr{

502: M.~Billo, B.~Craps and F.~Roose,

503: ``Orbifold boundary states from Cardy's condition,''

504: JHEP {\bf 0101}, 038 (2001)

505: [arXiv:hep-th/0011060].

506: %%CITATION = HEP-TH 0011060;%%

507: }

508: 

509: \lref\dm{

510: %\lref\DouglasSW{

511: M.~R.~Douglas and G.~W.~Moore,

512: ``D-branes, Quivers, and ALE Instantons,''

513: arXiv:hep-th/9603167.

514: %%CITATION = HEP-TH 9603167;%%

515: }

516: 

517: %\HoriCK

518: \lref\HoriCK{

519: K.~Hori, A.~Iqbal and C.~Vafa,

520: ``D-branes and mirror symmetry,''

521: arXiv:hep-th/0005247.

522: %%CITATION = HEP-TH 0005247;%%

523: }

524: 

525: %\HoriIC

526: \lref\HoriIC{

527: K.~Hori,

528: ``Linear models of supersymmetric D-branes,''

529: arXiv:hep-th/0012179.

530: %%CITATION = HEP-TH 0012179;%%

531: }

532: 

533: 

534: 

535: %\MaldacenaSN

536: \lref\MaldacenaSN{

537:   J.~Maldacena, G.~W.~Moore, N.~Seiberg and D.~Shih,

538:   ``Exact vs. semiclassical target space of the minimal string,''

539:   JHEP {\bf 0410}, 020 (2004)

540:   [arXiv:hep-th/0408039].

541:   %%CITATION = HEP-TH 0408039;%%

542: }

543: 

544: 

545: %\MartinecWG

546: \lref\MartinecWG{

547: E.~J.~Martinec and G.~Moore,

548: ``On decay of K-theory,''

549: arXiv:hep-th/0212059.

550: %%CITATION = HEP-TH 0212059;%%

551: }

552: 

553: %\MartinecTZ

554: \lref\MartinecTZ{

555: E.~J.~Martinec,

556: ``Defects, decay, and dissipated states,''

557: arXiv:hep-th/0210231.

558: %%CITATION = HEP-TH 0210231;%%

559: }

560: 

561: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

562: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

563: %%% 

564: %%% References 

565: %%%

566: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

567: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

568: 

569: \lref\fulton{W. Fulton, {\it Introduction to Toric Varieties},

570: Annals of Mathematics Studies, vol. 131; Princeton Univ. Press (1993).} 

571: %

572: \lref\bpv{W.~Barth, C.~Peters, A.~Van de Ven, {\it Compact Complex Surfaces};

573: Springer-Verlag (1984).} 

574: %

575: \lref\stevens{J. Stevens, ``On the versal deformation

576: of cyclic quotient singularities'',

577: in {\it Singularity theory and its applications, part I},

578: LNM 1462 pp.302-319.}

579: %

580: \lref\ishii{A. Ishii, ``On McKay correspondence

581: for a finite small subgroup of GL(2,C)'',

582: to appear in J. Reine Ang. Math.

583: (available at \nextline

584: http://www.kusm.kyoto-u.ac.jp/preprint/preprint2000.html).}

585: %

586: \lref\wunram{J. Wunram, ``Reflexive modules on quotient surface

587: singularities'', Math. Ann. {\bf 279}, 583 (1988).}

588: %

589: \lref\riemenschneider{O. Riemenschneider,

590: ``Special representations and the two-dimensional McKay correspondence''

591: (available at \nextline

592: http://www.math.uni-hamburg.de/home/riemenschneider/hokmckay.ps).}

593: %

594: %\AdamsSV

595: \lref\AdamsSV{

596: A.~Adams, J.~Polchinski and E.~Silverstein,

597: ``Don't panic! Closed string tachyons in ALE space-times,''

598: JHEP {\bf 0110}, 029 (2001)

599: arXiv:hep-th/0108075.

600: %%CITATION = HEP-TH 0108075;%%

601: }

602: %

603: %\HarveyWM

604: \lref\HarveyWM{

605: J.~A.~Harvey, D.~Kutasov, E.~J.~Martinec and G.~Moore,

606: ``Localized tachyons and RG flows,''

607: arXiv:hep-th/0111154.

608: %%CITATION = HEP-TH 0111154;%%

609: }

610: %

611: %\VafaRA

612: \lref\VafaRA{

613: C.~Vafa,

614: ``Mirror symmetry and closed string tachyon condensation,''

615: arXiv:hep-th/0111051.

616: %%CITATION = HEP-TH 0111051;%%

617: }

618: %

619: %\MartinecTZ

620: \lref\MartinecTZ{

621: E.~J.~Martinec,

622: ``Defects, decay, and dissipated states,''

623: arXiv:hep-th/0210231.

624: %%CITATION = HEP-TH 0210231;%%

625: }

626: %

627: %\MorrisonFR

628: \lref\MorrisonFR{

629: D.~R.~Morrison and M.~Ronen Plesser,

630: ``Summing the instantons: Quantum cohomology 

631: and mirror symmetry in toric varieties,''

632: Nucl.\ Phys.\ B {\bf 440}, 279 (1995)

633: hep-th/9412236.

634: %%CITATION = HEP-TH 9412236;%%

635: }

636: %

637: 

638: 

639: 

640: 

641: 

642: %\CecottiVB

643: \lref\CecottiVB{

644: S.~Cecotti and C.~Vafa,

645: ``Exact results for supersymmetric sigma models,''

646: Phys.\ Rev.\ Lett.\  {\bf 68}, 903 (1992)

647: hep-th/9111016.

648: %%CITATION = HEP-TH 9111016;%%

649: }

650: %

651: %\WittenYC

652: \lref\WittenYC{

653: E.~Witten,

654: ``Phases of N = 2 theories in two dimensions,''

655: Nucl.\ Phys.\ B {\bf 403}, 159 (1993)

656: hep-th/9301042.

657: %%CITATION = HEP-TH 9301042;%%

658: }

659: %

660: %\HoriCK

661: \lref\HoriCK{

662: K.~Hori, A.~Iqbal and C.~Vafa,

663: ``D-branes and mirror symmetry,''

664: arXiv:hep-th/0005247.

665: %%CITATION = HEP-TH 0005247;%%

666: }

667: %

668: %\HoriKT

669: \lref\HoriKT{

670: K.~Hori and C.~Vafa,

671: ``Mirror symmetry,''

672: arXiv:hep-th/0002222.

673: %%CITATION = HEP-TH 0002222;%%

674: }

675: %

676: %\HoriFJ

677: \lref\HoriFJ{

678: K.~Hori,

679: ``Mirror symmetry and some applications,''

680: arXiv:hep-th/0106043.

681: %%CITATION = HEP-TH 0106043;%%

682: }

683: 

684: 

685: %\HoriIC

686: \lref\HoriIC{

687: K.~Hori,

688: ``Linear models of supersymmetric D-branes,''

689: arXiv:hep-th/0012179.

690: %%CITATION = HEP-TH 0012179;%%

691: }

692: %

693: %\HellermanBU

694: \lref\HellermanBU{

695: S.~Hellerman, S.~Kachru, A.~E.~Lawrence and J.~McGreevy,

696: ``Linear sigma models for open strings,''

697: JHEP {\bf 0207}, 002 (2002)

698: arXiv:hep-th/0109069.

699: %%CITATION = HEP-TH 0109069;%%

700: }

701: %

702: %\LercheUY

703: \lref\LercheUY{

704: W.~Lerche, C.~Vafa and N.~P.~Warner,

705: ``Chiral Rings In N=2 Superconformal Theories,''

706: Nucl.\ Phys.\ B {\bf 324}, 427 (1989).

707: %%CITATION = NUPHA,B324,427;%%

708: }

709: %

710: %\DouglasSW

711: \lref\DouglasSW{

712: M.~R.~Douglas and G.~W.~Moore,

713: ``D-branes, Quivers, and ALE Instantons,''

714: arXiv:hep-th/9603167.

715: %%CITATION = HEP-TH 9603167;%%

716: }

717: %

718: %\HarveyNA

719: \lref\HarveyNA{

720: J.~A.~Harvey, D.~Kutasov and E.~J.~Martinec,

721: ``On the relevance of tachyons,''

722: arXiv:hep-th/0003101.

723: %%CITATION = HEP-TH 0003101;%%

724: }

725: %

726: %\SenMD

727: \lref\SenMD{

728: A.~Sen,

729: ``Supersymmetric world-volume action for non-BPS D-branes,''

730: JHEP {\bf 9910}, 008 (1999)

731: arXiv:hep-th/9909062.

732: %%CITATION = HEP-TH 9909062;%%

733: }

734: %

735: %\SenXM

736: \lref\SenXM{

737: A.~Sen,

738: ``Universality of the tachyon potential,''

739: JHEP {\bf 9912}, 027 (1999)

740: arXiv:hep-th/9911116.

741: %%CITATION = HEP-TH 9911116;%%

742: }

743: %

744: %\KutasovQP

745: \lref\KutasovQP{

746: D.~Kutasov, M.~Marino and G.~W.~Moore,

747: ``Some exact results on tachyon condensation in string field theory,''

748: JHEP {\bf 0010}, 045 (2000)

749: arXiv:hep-th/0009148.

750: %%CITATION = HEP-TH 0009148;%%

751: }

752: %

753: %\GerasimovZP

754: \lref\GerasimovZP{

755: A.~A.~Gerasimov and S.~L.~Shatashvili,

756: ``On exact tachyon potential in open string field theory,''

757: JHEP {\bf 0010}, 034 (2000)

758: arXiv:hep-th/0009103.

759: %%CITATION = HEP-TH 0009103;%%

760: }

761: 

762: \lref\reid{

763: M.~Reid,

764: ``La correspondance de McKay,''

765: S\'eminaire Bourbaki, 52\`eme ann\'ee, novembre 1999, no. 867, 

766: to appear in Ast\'erisque 2000

767: arXiv:alg-geom/9911165. For further references see 

768: http://www.maths.warwick.ac.uk/$\scriptstyle\sim$miles/McKay/

769: }

770: 

771: %\ReidZY

772: \lref\ReidZY{

773: M.~Reid,

774: ``McKay correspondence,''

775: arXiv:alg-geom/9702016.

776: %%CITATION = ALG-GEOM 9702016;%%

777: }

778: 

779: %\MayrAS

780: \lref\MayrAS{

781: P.~Mayr,

782: ``Phases of supersymmetric D-branes on Kaehler 

783: manifolds and the McKay  correspondence,''

784: JHEP {\bf 0101}, 018 (2001)

785: arXiv:hep-th/0010223.

786: %%CITATION = HEP-TH 0010223;%%

787: }

788: 

789: %\AnselmiSM

790: \lref\AnselmiSM{

791: D.~Anselmi, M.~Billo, P.~Fre, L.~Girardello and A.~Zaffaroni,

792: ``Ale Manifolds And Conformal Field Theories,''

793: Int.\ J.\ Mod.\ Phys.\ A {\bf 9}, 3007 (1994)

794: arXiv:hep-th/9304135.

795: %%CITATION = HEP-TH 9304135;%%

796: }

797: 

798: %\BuscherQJ

799: \lref\BuscherQJ{

800: T.~H.~Buscher,

801: ``Path Integral Derivation Of Quantum Duality In Nonlinear Sigma Models,''

802: Phys.\ Lett.\ B {\bf 201}, 466 (1988).

803: %%CITATION = PHLTA,B201,466;%%

804: }

805: 

806: %\RocekPS

807: \lref\RocekPS{

808: M.~Rocek and E.~Verlinde,

809: ``Duality, quotients, and currents,''

810: Nucl.\ Phys.\ B {\bf 373}, 630 (1992)

811: arXiv:hep-th/9110053.

812: %%CITATION = HEP-TH 9110053;%%

813: }

814: 

815: %\DelaOssaXK

816: \lref\DelaOssaXK{

817: X.~De la Ossa, B.~Florea and H.~Skarke,

818: ``D-branes on noncompact Calabi-Yau manifolds: K-theory and monodromy,''

819: Nucl.\ Phys.\ B {\bf 644}, 170 (2002)

820: arXiv:hep-th/0104254.

821: %%CITATION = HEP-TH 0104254;%%

822: }

823: 

824: \lref\ito{

825: Y.~Ito,

826: ``Special McKay correspondence,''

827: arXiv:alg-geom/0111314.

828: }

829: 

830: \lref\morelli{

831: R. Morelli, ``K theory of a toric variety,'' Adv. in Math. {\bf 100}(1993)154

832: }

833: 

834: %\AspinwallXS

835: \lref\AspinwallXS{

836: P.~S.~Aspinwall and M.~R.~Plesser,

837: ``D-branes, discrete torsion and the McKay correspondence,''

838: JHEP {\bf 0102}, 009 (2001)

839: arXiv:hep-th/0009042.

840: %%CITATION = HEP-TH 0009042;%%

841: }

842: %\DiaconescuEC

843: \lref\DiaconescuEC{

844: D.~E.~Diaconescu and M.~R.~Douglas,

845: ``D-branes on stringy Calabi-Yau manifolds,''

846: arXiv:hep-th/0006224.

847: %%CITATION = HEP-TH 0006224;%%

848: }

849: %\DiaconescuBR

850: \lref\DiaconescuBR{

851: D.~E.~Diaconescu, M.~R.~Douglas and J.~Gomis,

852: ``Fractional branes and wrapped branes,''

853: JHEP {\bf 9802}, 013 (1998)

854: arXiv:hep-th/9712230.

855: %%CITATION = HEP-TH 9712230;%%

856: }

857: 

858: %\LercheVJ

859: \lref\LercheVJ{

860: W.~Lerche, P.~Mayr and J.~Walcher,

861: ``A new kind of McKay correspondence from non-Abelian gauge theories,''

862: arXiv:hep-th/0103114.

863: %%CITATION = HEP-TH 0103114;%%

864: }

865: 

866: %\WittenCD

867: \lref\WittenCD{

868: E.~Witten,

869: ``D-branes and K-theory,''

870: JHEP {\bf 9812}, 019 (1998)

871: arXiv:hep-th/9810188.

872: %%CITATION = HEP-TH 9810188;%%

873: }

874: %\GarciaCompeanRG

875: \lref\GarciaCompeanRG{

876: H.~Garcia-Compean,

877: ``D-branes in orbifold singularities and equivariant K-theory,''

878: Nucl.\ Phys.\ B {\bf 557}, 480 (1999)

879: arXiv:hep-th/9812226.

880: %%CITATION = HEP-TH 9812226;%%

881: }

882: %\DelaOssaXK

883: \lref\DelaOssaXK{

884: X.~De la Ossa, B.~Florea and H.~Skarke,

885: ``D-branes on noncompact Calabi-Yau manifolds: K-theory and monodromy,''

886: Nucl.\ Phys.\ B {\bf 644}, 170 (2002)

887: arXiv:hep-th/0104254.

888: %%CITATION = HEP-TH 0104254;%%

889: }

890: 

891: %\GovindarajanVI

892: \lref\GovindarajanVI{

893: S.~Govindarajan and T.~Jayaraman,

894: ``D-branes, exceptional sheaves and quivers on 

895: Calabi-Yau manifolds: From Mukai to McKay,''

896: Nucl.\ Phys.\ B {\bf 600}, 457 (2001)

897: arXiv:hep-th/0010196.

898: %%CITATION = HEP-TH 0010196;%%

899: }

900: %\GovindarajanEF

901: \lref\GovindarajanEF{

902: S.~Govindarajan, T.~Jayaraman and T.~Sarkar,

903: ``On D-branes from gauged linear sigma models,''

904: Nucl.\ Phys.\ B {\bf 593}, 155 (2001)

905: arXiv:hep-th/0007075.

906: %%CITATION = HEP-TH 0007075;%%

907: }

908: 

909: %\HeCR

910: \lref\HeCR{

911: Y.~H.~He,

912: ``On algebraic singularities, finite graphs and D-brane gauge theories:

913: A  string theoretic perspective,''

914: arXiv:hep-th/0209230.

915: %%CITATION = HEP-TH 0209230;%%

916: }

917: 

918: %\TakayanagiXT

919: \lref\TakayanagiXT{

920: T.~Takayanagi,

921: ``Tachyon condensation on orbifolds and McKay correspondence,''

922: Phys.\ Lett.\ B {\bf 519}, 137 (2001)

923: arXiv:hep-th/0106142.

924: %%CITATION = HEP-TH 0106142;%%

925: }

926: 

927: %\TomasielloYM

928: \lref\TomasielloYM{

929: A.~Tomasiello,

930: ``D-branes on Calabi-Yau manifolds and helices,''

931: JHEP {\bf 0102}, 008 (2001)

932: arXiv:hep-th/0010217.

933: %%CITATION = HEP-TH 0010217;%%

934: }

935: 

936: %\GovindarajanVI

937: \lref\GovindarajanVI{

938: S.~Govindarajan and T.~Jayaraman,

939: ``D-branes, exceptional sheaves and quivers on Calabi-Yau manifolds:

940: {}From  Mukai to McKay,''

941: Nucl.\ Phys.\ B {\bf 600}, 457 (2001)

942: arXiv:hep-th/0010196.

943: %%CITATION = HEP-TH 0010196;%%

944: }

945: 

946: %\BatyrevJU

947: \lref\BatyrevJU{

948: V.~V.~Batyrev and D.~I.~Dais,

949: ``Strong Mckay Correspondence, String Theoretic Hodge Numbers And

950: Mirror Symmetry,''

951: arXiv:alg-geom/9410001.

952: %%CITATION = ALG-GEOM 9410001;%%

953: }

954: 

955: %\ItoZX

956: \lref\ItoZX{

957: Y.~Ito and M.~Reid,

958: ``The McKay correspondence for finite subgroups of SL(3,C),''

959: arXiv:alg-geom/9411010.

960: %%CITATION = ALG-GEOM 9411010;%%

961: }

962: 

963: \lref\itonak{

964: Y.~Ito and H.~Nakajima,

965: ``McKay correspondence and Hilbert schemes in dimension three,''

966: arXiv:al-geom/9802120.

967: }

968: 

969: \lref\crawthesis{

970: A.~Craw,

971: ``The McKay correspondence and representations of the McKay quiver,''

972: Ph.D. thesis, University of Warwick; available at

973: http://www.math.utah.edu/~craw.

974: }

975: 

976: %\KachruAN

977: \lref\KachruAN{

978: S.~Kachru, S.~Katz, A.~E.~Lawrence and J.~McGreevy,

979: ``Mirror symmetry for open strings,''

980: Phys.\ Rev.\ D {\bf 62}, 126005 (2000)

981: arXiv:hep-th/0006047.

982: %%CITATION = HEP-TH 0006047;%%

983: }

984: 

985: 

986: %\GovindarajanEF

987: \lref\GovindarajanEF{

988: S.~Govindarajan, T.~Jayaraman and T.~Sarkar,

989: ``On D-branes from gauged linear sigma models,''

990: Nucl.\ Phys.\ B {\bf 593}, 155 (2001)

991: arXiv:hep-th/0007075.

992: %%CITATION = HEP-TH 0007075;%%

993: }

994: 

995: %\HellermanCT

996: \lref\HellermanCT{

997: S.~Hellerman and J.~McGreevy,

998: ``Linear sigma model toolshed for D-brane physics,''

999: JHEP {\bf 0110}, 002 (2001)

1000: arXiv:hep-th/0104100.

1001: %%CITATION = HEP-TH 0104100;%%

1002: }

1003: 

1004: %\GovindarajanKR

1005: \lref\GovindarajanKR{

1006: S.~Govindarajan and T.~Jayaraman,

1007: ``Boundary fermions, coherent sheaves and D-branes on Calabi-Yau

1008: manifolds,''

1009: Nucl.\ Phys.\ B {\bf 618}, 50 (2001)

1010: arXiv:hep-th/0104126.

1011: %%CITATION = HEP-TH 0104126;%%

1012: }

1013: 

1014: %\DistlerYM

1015: \lref\DistlerYM{

1016: J.~Distler, H.~Jockers and H.~J.~Park,

1017: ``D-brane monodromies, derived categories and boundary linear sigma

1018: models,''

1019: arXiv:hep-th/0206242.

1020: %%CITATION = HEP-TH 0206242;%%

1021: }

1022: 

1023: %\KatzGH

1024: \lref\KatzGH{

1025: S.~Katz and E.~Sharpe,

1026: ``D-branes, open string vertex operators, and Ext groups,''

1027: arXiv:hep-th/0208104.

1028: %%CITATION = HEP-TH 0208104;%%

1029: }

1030: 

1031: 

1032: %\MorrisonYH

1033: \lref\MorrisonYH{

1034: D.~R.~Morrison and M.~Ronen Plesser,

1035: ``Towards mirror symmetry as duality 

1036: for two dimensional abelian gauge  theories,''

1037: Nucl.\ Phys.\ Proc.\ Suppl.\  {\bf 46}, 177 (1996)

1038: arXiv:hep-th/9508107.

1039: %%CITATION = HEP-TH 9508107;%%

1040: }

1041: 

1042:  

1043: 

1044: \lref\kronheimer{PB. Kronheimer and H. Nakajima, 

1045: ``Yang-Mills instantons on ALE gravitational 

1046: instantons,'' Math. Ann. {\bf 288}(1990)263}

1047: 

1048: \lref\mooresegal{G. Moore and G. Segal, unpublished. The 

1049: material is available in lecture notes from the April 2002 Clay 

1050: School on Geometry and Physics, Newton Institute,  and 

1051: http://online.kitp.ucsb.edu/online/mp01/moore1.} 

1052: 

1053: \lref\iaslectures{E. Witten, in {\it Quantum Fields and Strings: A 

1054: Course for Mathematicians}, vol. 2, P. Deligne et. al. eds. 

1055: Amer. Math. Soc. 1999} 

1056: 

1057: 

1058: \lref\kapustinlectures{A. Kapustin, Lectures at the KITP workshop 

1059: on Mathematics and Physics, August, 2003.} 

1060: 

1061: %\SinYM

1062: \lref\SinYM{

1063: S.~J.~Sin,

1064: ``Comments on the fate of unstable orbifolds,''

1065: Phys.\ Lett.\ B {\bf 578}, 215 (2004)

1066: [arXiv:hep-th/0308028].

1067: %%CITATION = HEP-TH 0308028;%%

1068: }

1069: 

1070: %\LeeAR

1071: \lref\LeeAR{

1072: S.~g.~H.~Lee and S.~J.~H.~Sin,

1073: ``Chiral rings and GSO projection in orbifolds,''

1074: Phys.\ Rev.\ D {\bf 69}, 026003 (2004)

1075: [arXiv:hep-th/0308029].

1076: %%CITATION = HEP-TH 0308029;%%

1077: }

1078: %\LeeSS

1079: \lref\LeeSS{

1080: S.~Lee and S.~J.~Sin,

1081: ``Localized tachyon condensation and G-parity conservation,''

1082: arXiv:hep-th/0312175.

1083: %%CITATION = HEP-TH 0312175;%%

1084: }

1085: 

1086: %\GregoryYB

1087: \lref\GregoryYB{

1088: R.~Gregory and J.~A.~Harvey,

1089: ``Spacetime decay of cones at strong coupling,''

1090: Class.\ Quant.\ Grav.\  {\bf 20}, L231 (2003)

1091: [arXiv:hep-th/0306146].

1092: %%CITATION = HEP-TH 0306146;%%

1093: }

1094: 

1095: %\HeadrickYU

1096: \lref\HeadrickYU{

1097: M.~Headrick,

1098: ``Decay of C/Z(n): Exact supergravity solutions,''

1099: arXiv:hep-th/0312213.

1100: %%CITATION = HEP-TH 0312213;%%

1101: }

1102: 

1103: %\DixonIZ

1104: \lref\DixonIZ{

1105: L.~J.~Dixon and J.~A.~Harvey,

1106: ``String Theories In Ten-Dimensions Without Space-Time Supersymmetry,''

1107: Nucl.\ Phys.\ B {\bf 274}, 93 (1986).

1108: %%CITATION = NUPHA,B274,93;%%

1109: }

1110: 

1111: 

1112: %\MaldacenaKY

1113: \lref\MaldacenaKY{

1114: J.~M.~Maldacena, G.~W.~Moore and N.~Seiberg,

1115: ``Geometrical interpretation of D-branes in gauged WZW models,''

1116: JHEP {\bf 0107}, 046 (2001)

1117: [arXiv:hep-th/0105038].

1118: %%CITATION = HEP-TH 0105038;%%

1119: }

1120: 

1121: 

1122: %\MinwallaHJ

1123: \lref\MinwallaHJ{

1124: S.~Minwalla and T.~Takayanagi,

1125: ``Evolution of D-branes under closed string tachyon condensation,''

1126: JHEP {\bf 0309}, 011 (2003)

1127: [arXiv:hep-th/0307248].

1128: %%CITATION = HEP-TH 0307248;%%

1129: }

1130: 

1131: 

1132: %\KapustinRC

1133: \lref\KapustinRC{

1134: A.~Kapustin and Y.~Li,

1135: ``D-branes in topological minimal models: The Landau-Ginzburg approach,''

1136: arXiv:hep-th/0306001.

1137: %%CITATION = HEP-TH 0306001;%%

1138: }

1139: %\KapustinGA

1140: \lref\KapustinGA{

1141: A.~Kapustin and Y.~Li,

1142: ``Topological correlators in Landau-Ginzburg models with boundaries,''

1143: arXiv:hep-th/0305136.

1144: %%CITATION = HEP-TH 0305136;%%

1145: }

1146: %\KapustinKT

1147: \lref\KapustinKT{

1148: A.~Kapustin and D.~Orlov,

1149: ``Lectures on mirror symmetry, derived categories, and D-branes,''

1150: arXiv:math.ag/0308173.

1151: %%CITATION = MATH-AG 0308173;%%

1152: }

1153: 

1154: %\DouglasSW

1155: \lref\DouglasSW{

1156: M.~R.~Douglas and G.~W.~Moore,

1157: ``D-branes, Quivers, and ALE Instantons,''

1158: arXiv:hep-th/9603167.

1159: %%CITATION = HEP-TH 9603167;%%

1160: }

1161: 

1162: 

1163: 

1164: 

1165: \lref\segal{G.B.  Segal, ``Equivariant K-Theory,''  Publ. Math.

1166: IHES, {\bf 34}(1968)129}

1167: 

1168: %\MooreYT

1169: \lref\MP{

1170:   G.~W.~Moore and A.~Parnachev,

1171:   ``Localized tachyons and the quantum McKay correspondence,''

1172:   JHEP {\bf 0411}, 086 (2004)

1173:   [arXiv:hep-th/0403016].

1174:   %%CITATION = HEP-TH 0403016;%%

1175: }

1176: 

1177: %\HoriKT

1178: \lref\HV{

1179:   K.~Hori and C.~Vafa,

1180:   ``Mirror symmetry,''

1181:   arXiv:hep-th/0002222.

1182:   %%CITATION = HEP-TH 0002222;%%

1183: }

1184: 

1185: %\HoriCK

1186: \lref\HIV{

1187:  K.~Hori and C.~Vafa,

1188:   ``Mirror symmetry,''

1189:   arXiv:hep-th/0002222;

1190:   %%CITATION = HEP-TH 0002222;%% 

1191:  K.~Hori, A.~Iqbal and C.~Vafa,

1192:   ``D-branes and mirror symmetry,''

1193:   arXiv:hep-th/0005247;

1194:   %%CITATION = HEP-TH 0005247;%%

1195: }

1196: 

1197: \lref\boalch{Boalch, Oxford PhD thesis. Go to http://www.math.columbia.edu/~boalch/publications.html }

1198: 

1199: \lref\kaminski{D. Kaminski and R.B. Paris, ``Asymptotics of a class of multidimensional Laplace-type integrals. I. 

1200: Double integrals,'' Phil. Trans. R. Soc. London. A (1998){\bf 356}583} 

1201: 

1202: 

1203: 

1204: \lref\HV{Hori and Vafa, ``Proof'' } 

1205: %\lref\HIV{Hori, Iqbal, Vafa} 

1206: \lref\boalch{Boalch, Oxford PhD thesis. Go to http://www.math.columbia.edu/~boalch/publications.html } 

1207: %\lref\MP{Moore and Parnachev} 

1208: 

1209: \lref\kaminski{D. Kaminski and R.B. Paris, ``Asymptotics of a class of multidimensional Laplace-type integrals. I. 

1210: Double integrals,'' Phil. Trans. R. Soc. London. A (1998){\bf 356}583} 

1211:  

1212: %\HarveyNA

1213: \lref\HarveyNA{

1214:   J.~A.~Harvey, D.~Kutasov and E.~J.~Martinec,

1215:   ``On the relevance of tachyons,''

1216:   arXiv:hep-th/0003101.

1217:   %%CITATION = HEP-TH 0003101;%%

1218: }

1219: 

1220: %\MelnikovHQ

1221: \lref\MelnikovHQ{

1222:   I.~V.~Melnikov and M.~R.~Plesser,

1223:   ``The Coulomb branch in gauged linear sigma models,''

1224:   arXiv:hep-th/0501238; I.~V.~Melnikov and M.~R.~Plesser,

1225:   ``A-Model Correlators from the Coulomb Branch,''

1226:   arXiv:hep-th/0507187.

1227: }

1228: 

1229: %\MorrisonFR

1230: \lref\MNP{

1231:   D.~R.~Morrison, K.~Narayan and M.~R.~Plesser,

1232:   ``Localized tachyons in C(3)/Z(N),''

1233:   JHEP {\bf 0408}, 047 (2004)

1234:   [arXiv:hep-th/0406039].

1235:   %%CITATION = HEP-TH 0406039;%%

1236: }

1237: 

1238: %\MorrisonJA

1239: \lref\MorrisonJA{

1240:   D.~R.~Morrison and K.~Narayan,

1241:   ``On tachyons, gauged linear sigma models, and flip transitions,''

1242:   JHEP {\bf 0502}, 062 (2005)

1243:   [arXiv:hep-th/0412337].

1244:   %%CITATION = HEP-TH 0412337;%%

1245: }

1246: 

1247: %\SarkarRY

1248: \lref\SarkarRY{

1249:   T.~Sarkar,

1250:   ``On localized tachyon condensation in C**2/Z(n) and C**3/Z(n),''

1251:   Nucl.\ Phys.\ B {\bf 700}, 490 (2004)

1252:   [arXiv:hep-th/0407070].

1253:   %%CITATION = HEP-TH 0407070;%%

1254: }

1255: 

1256: %\OoguriVR

1257: \lref\OoguriVR{

1258:   H.~Ooguri, C.~Vafa and E.~P.~Verlinde,

1259:   ``Hartle-Hawking wave-function for flux compactifications,''

1260:   arXiv:hep-th/0502211.

1261:   %%CITATION = HEP-TH 0502211;%%

1262: }

1263: 

1264: %\McGreevyCI

1265: \lref\McGreevyCI{

1266:   J.~McGreevy and E.~Silverstein,

1267:   ``The tachyon at the end of the universe,''

1268:   arXiv:hep-th/0506130.

1269:   %%CITATION = HEP-TH 0506130;%%

1270: }

1271: 

1272: %\HorowitzVP

1273: \lref\HorowitzVP{

1274:   G.~T.~Horowitz,

1275:   ``Tachyon condensation and black strings,''

1276:   arXiv:hep-th/0506166.

1277:   %%CITATION = HEP-TH 0506166;%%

1278: }

1279: 

1280: %\HeadrickHZ

1281: \lref\HeadrickHZ{

1282:   M.~Headrick, S.~Minwalla and T.~Takayanagi,

1283:   ``Closed string tachyon condensation: An overview,''

1284:   Class.\ Quant.\ Grav.\  {\bf 21}, S1539 (2004)

1285:   [arXiv:hep-th/0405064].

1286:   %%CITATION = HEP-TH 0405064;%%

1287: }

1288: 

1289: \lref\dubrovinI{B. Dubrovin, ``Painleve' transcendents and 

1290: two-dimensional topological field theory,''

1291: arXiv:math.AG/9803107}

1292: \lref\dubrovinII{B. Dubrovin, ``Geometry and analytic theory of 

1293: Frobenius manifolds,'' arXiv:math\ .AG/9807034}

1294: \lref\guzzetti{D. Guzzetti, ``Stokes matrices and monodromy of the 

1295: quantum cohomology of

1296: projective spaces,'' arXiv:math.AG/9904099}

1297: \lref\ueda{K.Ueda, ``Stokes matrices for the quantum cohomologies of 

1298: Grassmannians,''

1299: arXiv:\ math.AG/0503355}

1300: \lref\arouxI{D. Auroux, L. Katzarkov, and D. Orlov, ``Mirror symmetry 

1301: for weighted projective

1302: planes and their noncommutative deformations,'' arXiv:math.AG/0404281}

1303: \lref\arouxII{D. Auroux, L. Katzarkov, and D. Orlov, ``Mirror symmetry 

1304: for Del Pezzo surfaces:

1305: vanishing cycles and coherent sheaves,'' arXiv:math.AG/0506166}

1306: 

1307: 

1308: %\DouglasHQ

1309: \lref\DouglasHQ{

1310: M.~R.~Douglas and B.~Fiol,

1311: ``D-branes and discrete torsion. II,''

1312: arXiv:hep-th/9903031.

1313: %%CITATION = HEP-TH 9903031;%%

1314: }

1315: %\BerkoozIS

1316: \lref\BerkoozIS{

1317: M.~Berkooz and M.~R.~Douglas,

1318: ``Five-branes in M(atrix) theory,''

1319: Phys.\ Lett.\ B {\bf 395}, 196 (1997)

1320: [arXiv:hep-th/9610236].

1321: %%CITATION = HEP-TH 9610236;%%

1322: }

1323: %\BerkoozKM

1324: \lref\BerkoozKM{

1325: M.~Berkooz, M.~R.~Douglas and R.~G.~Leigh,

1326: ``Branes intersecting at angles,''

1327: Nucl.\ Phys.\ B {\bf 480}, 265 (1996)

1328: [arXiv:hep-th/9606139].

1329: %%CITATION = HEP-TH 9606139;%%

1330: }

1331: 

1332: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1333: % paper starts here !!!

1334: 

1335: 

1336:  

1337: \Title{\vbox{\baselineskip12pt

1338: \hbox{hep-th/0507190}

1339: %\hbox{RUNHETC-2004-04}

1340: }}

1341: {\vbox{\centerline{Profiling the Brane Drain}

1342: \vskip.06in

1343: \centerline{in a Nonsupersymmetric Orbifold } 

1344: %\centerline{Quantum McKay Correspondence}

1345: }}

1346: \centerline{Gregory Moore and Andrei Parnachev}

1347: \bigskip

1348: \centerline{{\it Department of Physics, Rutgers University}}

1349: \centerline{\it Piscataway, NJ 08854-8019, USA}

1350:  \vskip.1in \vskip.1in \centerline{\bf Abstract}  

1351: \noindent

1352: %

1353: We study D-branes in a nonsupersymmetric orbifold of type $\IC^2/\Gamma$, perturbed 

1354: by a tachyon condensate, using a gauged linear sigma model. The RG flow has both 

1355: higgs and coulomb branches, and each branch supports different branes. 

1356: The coulomb branch branes account for the ``brane drain'' from the higgs branch, 

1357: but their precise relation to fractional branes has hitherto been unknown. 

1358: Building on the results of hep-th/0403016 we construct, in detail, the map between 

1359: fractional branes and the coulomb/higgs branch branes for two examples in the type 0 theory. 

1360:  This map depends on the phase of the tachyon condensate 

1361: in a surprising and intricate way. In the mirror 

1362: Landau-Ginzburg picture the dependence on the tachyon phase is manifested by 

1363: discontinuous changes in the shape of the D-brane. 

1364: 

1365: 

1366: \vfill

1367: 

1368: \Date{July 20, 2005}

1369:    

1370: %\draftmode

1371: 

1372: \newsec{Introduction and summary}

1373: 

1374: An important property of string theory is that it is 

1375: well-defined in the presence of certain spacetime 

1376: singularities which render general relativity and quantum 

1377: field theory ill-defined. Moreover, string theory 

1378: contains mechanisms for smoothing out spacetime singularities. 

1379: An interesting set of concrete examples  

1380: of this phenomenon   are spacetime nonsupersymmetric orbifolds of flat space 

1381: \AdamsSV. (See \MartinecTZ\HeadrickHZ\ for reviews).

1382: In such situations, the closed string spectrum contains tachyons whose

1383: wavefunctions are localized near the singular point; the resolution of singularities

1384: happens through the condensation of these tachyons.

1385: The presence of   $\NN=2$ worldsheet supersymmetry imposes

1386: constraints on the dynamics of the system, allowing one to  

1387:  understand the renormalization group (RG) flow \hkmm\ in a 

1388: way analogous to the understanding of open string tachyon 

1389: condensation  (see e.g. \HarveyNA). In this paper we assume that 

1390: the RG flow gives a good description of  condensation

1391: of localized closed string tachyons.

1392: The behavior of the system in the IR corresponds 

1393: to later times in the time evolution.

1394: 

1395: It is technically convenient to introduce the gauged linear 

1396: sigma model (GLSM), whose higgs branch in the ultraviolet is the

1397: nonsupersymmetric orbifold \VafaRA\mm.

1398: In the process of RG flow, the higgs branch resolves into

1399: a smooth \HJ\ space, which has a natural spacetime interpretation.

1400: The number of branes wrapping nontrivial two-cycles

1401: of the \HJ\ space is generally smaller than

1402: that of fractional branes, so naively a ``brane drain'' is taking place.

1403: However one must bear in mind that the infrared theory contains a coulomb branch 

1404: with isolated massive vacua.

1405: It has been suggested in \mm\ that D-branes wrapping  nontrivial

1406: cycles in the higgs branch, together with D-branes supported 

1407: at the vacua of the coulomb branch, are in one-to-one correspondence with 

1408: fractional branes in the orbifold theory.

1409: This picture has been sharpened in \MP\ where the open string

1410: Witten index was used to construct a map between the fractional

1411: branes and higgs and coulomb branch branes in the IR.\foot{

1412: The significance of the coulomb branch goes back to 

1413: \WittenYC\MorrisonFR\

1414: and was also recently emphasized 

1415: in \MelnikovHQ\ where the correlators in the topologically

1416: twisted A-model were found to have support precisely on the

1417: coulomb branch.}

1418: While \MP\ found the map in the case of type II string theory,

1419: a similar construction is possible in type 0 theory.

1420: We describe it in detail in Section 2.

1421: 

1422: One can study D-branes away from the conformal point 

1423: using techniques developed in \HIV.

1424: The mirror description of the GLSM is given by the Landau-Ginzburg (LG)

1425: theory with an inhomogeneous superpotential.

1426: Fractional branes localized at the orbifold fixed point

1427: preserve B-type supersymmetry, so they become A-type

1428: branes in the LG model.

1429: The latter are associated with critical points of the superpotential:

1430: the critical points away from the origin give rise to the coulomb

1431: branch branes, while those at the origin describe

1432: the higgs branch brane(s).

1433: In this paper we focus on the overlap of the boundary state

1434: with the identity operator (this is defined more

1435: precisely in Section 3).

1436: This quantity, which we call the generalized central charge, should

1437: in the first approximation be thought of as generalization of the D-brane mass to nonconformal

1438: theories.

1439: The generalized central charge 

1440: of a brane described by the boundary state $|B\rangle$ can be computed as an integral

1441: \eqn\gcci{  \langle B|1\rangle=\int\int {dx_1\over x_1} {dx_3\over x_3} \exp(-W)   }

1442: over the A-brane surface.

1443: In \gcci\ $x_1$ and $x_3$ are the LG fields, and $W$ is the superpotential. 

1444: To determine the shape of the A-brane, it is necessary to

1445: solve certain soliton equations.

1446: The set of all solutions is parametrized by a small circle (wavefront) around

1447: the critical point.

1448: The A-brane surface is traced by the wavefront evolving in 

1449: time \HIV.

1450: 

1451: We specialize to the case of $\IC^2/\IZ_{n(p)},\,p=1$ orbifolds, 

1452: whose minimal resolution contains a single non-trivial cycle.

1453: (The higgs branch in this case is $\OO(-n)\ra\IP^1$.)

1454: There is a single higgs branch brane wrapping the nontrivial cycle,

1455: and $n-2$ coulomb branch branes associated with the massive vacua of the

1456: superpotential.

1457: We find that the generic form of the A-brane surface resembles

1458: that of a propeller.

1459: Near the critical point, the wavefront is a small circle, whose segments

1460: at late times trace various quarter-planes (``wings'' of the ``propeller'').

1461: This property can be used to compute \gcci\ as a function of (complex) tachyon

1462: expectation value $w$; $w\ra 0$ corresponds to the orbifold (UV) limit,

1463: while $|w|\ra\infty$ describes the IR regime.

1464: As explained in Section 4, \gcci\ satisfies a GKZ equation, which 

1465: in the Calabi-Yau case is a Picard-Fuchs equation for the periods.

1466: A basis of nonconstant solutions of the GKZ equation

1467: is given by the integrals \gcci\ over the quarter planes.

1468: Linear combinations of these integrals, which we compute in Section 4, determine

1469: the generalized central charge of the A-brane whose wings asymptote to these quarter planes.

1470: 

1471: We analyze the behavior of \gcci\ in the simple cases of $n=3,4$ in Section 5.

1472: The intersection matrix and the quantum symmetry of the orbifold theory 

1473: %allows the 

1474: %identification of the coulomb branch branes with fractional branes.

1475: suffices to determine the map between the coulomb branch branes

1476: and the fractional branes.

1477: This map depends of the phase of the tachyon expectation value.

1478: %$w\ra 0$ corresponds to the orbifold (UV) limit,

1479: %while $|w|\ra\infty$ describes the IR regime. 

1480: Multiplication of $w$ by an $n$-th root of unity enforces the permutation

1481: symmetry of the fractional branes.

1482: In terms of the propeller surfaces, the asymptotics change

1483: discontinuously when the value of $w$ goes from one angular sector 

1484: of the complex plane to another.

1485: 

1486: As explained in Section 2, even after modding out by permutation symmetry

1487: of the fractional branes, in the type 0 theory there is more then one expression for the

1488: higgs branch brane which is consistent with the intersection matrix.

1489: In section 5 we find that each of the $n$ angular sectors which differ by a 

1490: permutation of fractional branes is further divided into smaller

1491: subsectors, where different expressions for the higgs branch brane are realized.

1492: The generalized central charge for all higgs branch branes has the same leading

1493: logarithmic behavior in the regime of large $|w|$.

1494: %The difference turns out to be non-perturbative (exponentially small) in $w$.

1495: %It is nevertheless very important near the orbifold point, where

1496: %some higgs branch branes can even become massless!

1497: %This does not happen in the type II case, where the expression 

1498: %for the higgs branch brane is unique (up to a permutation of fractional branes).

1499: 

1500: To summarize, this paper is organized as follows.

1501: In the next section we describe the orbifold, GLSM and

1502: its mirror LG model.

1503: We use the intersection matrix to relate

1504: fractional branes to the coulomb and higgs branch branes in the IR.

1505: In Section 3 we define the generalized central charge.

1506: In Section 4 we show that it solves the GKZ equation, and analyze

1507: the solutions.

1508: Section 5 is devoted to the detailed analysis of the $n=3$ and

1509: $n=4$ cases.

1510: We discuss our results and directions for future research in

1511: Section 6.

1512: Appendix A contains information on the construction of fractional

1513: brane boundary states.

1514: Appendix B is devoted to the numerical analysis of the shape

1515: of A-branes in the LG theory.

1516: 

1517: 

1518: 

1519: 

1520: 

1521: 

1522: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1523: \newsec{Fractional branes vs. higgs and coulomb branch branes } 

1524: In this section we start by reviewing the results of  \mm and \MP\

1525: where the fate of fractional branes was studied using the gauged

1526: linear sigma model.

1527: We recall the mirror description in terms of the Landau-Ginzburg theory,

1528: and give a first hint at the appearance of the fractional branes

1529: in this language.

1530: We study the intersection form for type II and type 0 orbifolds

1531: and use it to construct the linear map between the fractional

1532: branes and the LG branes.

1533: We discuss the $\IC^2/\IZ_{3(1)}$ and  $\IC^2/\IZ_{4(1)}$

1534: examples in detail.

1535: Part of this section is review material.

1536: A more detailed exposition can be found in \hkmm\mm\MP\MartinecTZ.

1537: The map between the fractional branes and the LG (or, equivalently, GLSM)

1538: branes is spelled out in detail, although such a map was

1539: constructed implicitly in \MP.

1540: The discussion of the intersection form in the type 0 theory is new.

1541: The intersection form in type 0 theory is important for the $\IC^2/\IZ_{3(1)}$

1542: example which is discussed at the end of this section and later in the paper. 

1543: %

1544: \subsec{Condensation of localized tachyons and the fate of the fractional branes.}

1545: 

1546: We consider type II or type 0 theory in 9+1 dimensions.

1547: The orbifolding by $\IZ_{n(p)}$ happens in the 67 and 89 planes,

1548: parametrized by complex coordinates $X^{(1)}$ and  $X^{(2)}$.

1549: The orbifold group is generated by

1550: \eqn\orbaction{ g=\exp\left({2\pi i\over n}(J_{67}+p J_{89})\right) ,}

1551: where $J_{67}$ and $J_{89}$ generate rotations in two complex planes.

1552: When there are fermions in the theory, $p$ is defined ${\rm mod} \, 2n $.

1553: We will take the fundamental domain to be $p\in (-n,n)$.

1554: The action of $g^n$ on the Ramond sector ground state is a multiplication by $(-1)^{p\pm 1}$,

1555: depending on chirality.

1556: When $p$ is even, this acts as $(-1)^F$ where $F$ is the spacetime fermion

1557: number.

1558: In type II, there is no bulk tachyon and there are closed string

1559: fermions in the bulk, hence $p$ must be odd \refs\AdamsSV.

1560: %

1561: In the NSR formalism, a useful ingredient in the theory is the operator \hkmm

1562: \eqn\defxj{

1563:   X^{(i)}_{s\over n}=\sigma_{s/n}^{(i)}\;\exp\left[i(s/n)(H^{(i)}-\tilde H^{(i)})\right];

1564: 	\qquad i=1,2; \qquad s=1,2,\cdots, n-1

1565: }

1566: where $H^{(i)},\tilde H^{(i)}$ are the bosonised left- and right-moving fermions and

1567: $\sigma_{s/n}$ is the bosonic twist $s$ operator \refs\DixonQV.

1568: In the following we will restrict our attention to the left movers.

1569: %and will drop the $+$ and $-$ subscripts.

1570: There are two possible inequivalent choices for

1571: the worldsheet $\NN=1$ supersymmetry generator in the theory.

1572: Correspondingly, there are two chiral rings

1573: which are BPS under these supersymmetries.

1574: The (c,c) ring vertex operators are

1575: \eqn\xj{X_{s\over n}^{(cc)}=X^{(1)}_{s\over n} X^{(2)}_{\left\{ {sp\over n} \right\}}       ,}

1576: where $\left\{ x \right\} \equiv x-\left[ x \right]$

1577: is the fractional part of $x$.

1578: The (c,a) ring vertex operators are

1579: \eqn\xjca{X_{s\over n}^{(ac)}=

1580:           X^{(1)}_{s\over n} \left( X^{(2)}_{1-\left\{ {sp\over n} \right\}}\right)^*       ,}

1581: The operation $p\ra -p$ corresponds to exchanging the $(c,c)$ and the $(c,a)$

1582: rings. 

1583: Therefore we can restrict ourselves to the theories with $p\in (0,n)$.

1584: 

1585: The generators of the (c,a) ring, denoted  $W_\alpha, \;\;\alpha=1\ldots r$ 

1586: form a collection of (in general) relevant operators. Turning these on 

1587: in the action induces RG flow to  

1588: the minimal resolution of the singularity \refs{\hkmm,\mm}.

1589: For the $\IC^2/\IZ_{n(p)}$ orbifold such a resolution is encoded

1590: in the continued fraction

1591: \eqn\cfdef{{n\over n-p}=a_1-{1\over a_2-{1\over a_3- \ldots}}\; := \; [a_1,a_2,\ldots a_r]      ,}

1592: Note the appearance of $n-p$ rather then $p$ in \cfdef, since we are

1593: talking about the (c,a) ring.

1594: The smooth space which appears after the minimal

1595: resolution of the singularity is called the \HJ\ manifold.

1596: The generators of the chiral ring are in one-to-one correspondence

1597: with the exceptional $\IP^1$'s of this space.

1598: Their intersection numbers are given by

1599: \eqn\exc{C_{\alpha\beta}=

1600:          -\delta_{\alpha,\beta-1}+a_\alpha \delta_{\alpha,\beta}-\delta_{\alpha,\beta+1}}

1601: %

1602: The (c,c) ring generators give rise to the resolution with 

1603: similar properties; one needs to substitute $n-p\ra p$ in \cfdef.

1604: In the type II theory one should bear in mind the existence of a chiral GSO projection.

1605: As explained in \MP, all generators $W_\alpha$ in the (c,a) ring

1606: survive the GSO projection if and only if all  $a_j$ are even integers.

1607: In the (c,c) ring at least one generator is projected out \MP.

1608: In the type 0 theory chiral operators in (c,a) and (c,c) 

1609: rings are not projected out by the diagonal GSO projection.

1610: Since $p\ra p+n$ does not affect the ring structure, but only

1611: affects the GSO projection, the closed string sector in type 0 theories  

1612: is unchanged under this operation.

1613: This means that type 0 theories with $p$ and $n-p$ are isomorphic: they are related

1614: by interchanging the (c,c) and the (c,a) rings.

1615: Put differently, in a type 0 theory with a given $p$, one should

1616: be able to resolve the singularity into two different \HJ\ spaces,

1617: whose intersection numbers correspond to continued

1618: fractions determined by both $p$ and $n-p$.

1619: In \MP\ it has been shown that branes in type II theory which wrap nontrivial two-cycles 

1620: in the \HJ\ space are given by linear combinations of the

1621: fractional branes.

1622: In this paper we will see that the situation in the type 0

1623: string theory is similar.

1624: Since there are two different resolutions, one can define two

1625: sets of branes wrapping the two-cycles in these spaces.

1626: Both sets are given by certain linear combinations of fractional branes.

1627: %We will see how different sets of D-branes see these resolutions. 

1628: 

1629: In \mm\ the gauged linear sigma model (GLSM) was used to shed light

1630: on the fate of fractional D-branes in the process of twisted

1631: tachyon condensation.

1632: The field content of the relevant GLSM involves

1633: $r$ abelian $\CN=2$ gauge fields

1634: $V_\alpha$, $\alpha=1,...,r$ coupled to $r+d$ 

1635: $\CN=2$ chiral matter fields

1636: $X_i$ with charges 

1637: \eqn\qcharges{ Q_{\alpha i} = -C_{\alpha i}    }

1638: with $C_{\alpha i}$ given by [compare with \exc]:

1639: \eqn\exci{C_{\alpha i}=

1640:          -\delta_{\alpha,i-1}+a_\alpha \delta_{\alpha,i}-\delta_{\alpha,i+1};\qquad i=0,\ldots,r+1}

1641: The field strengths of the gauge fields 

1642: are contained in twisted chiral superfields

1643: $\Sigma={1\over 2} \{\overline\CD,\CD^*\}$.

1644: The classical Lagrangian is

1645: \eqn\glsmact{

1646:   \CL = \int d^4\theta \;\left(\Xbar_i e^{2Q_{\alpha i} V_\alpha} X_i

1647: 	-\frac{1}{2e_{\alpha}^2}\bar\Sigma_\alpha \Sigma_\alpha\right)

1648: 	-\half\left(\int d^2\tilde\theta 

1649: 		\;t_{\alpha}\Sigma_\alpha+{\rm c.c.}\right)\ ,

1650: }

1651: %%%%

1652: %%%%

1653: where repeated indices are summed and 

1654: %

1655: \eqn\teebare{

1656: t_{\alpha } =\zeta_\alpha-i\theta_\alpha

1657: }

1658:  combines

1659: the Fayet-Iliopoulos (FI) parameter $\zeta$ and theta angle $\theta$

1660: for the $\alpha^{\rm th}$ gauge field; $d^2\tilde\theta$ is

1661: the twisted chiral superspace measure.  

1662: The renormalized FI parameters at the scale $\mu$ is

1663: %

1664: \eqn\FIren{

1665:   t_{\alpha,{\rm eff}}(\mu) = t_{\alpha,{\rm bare}} 

1666: 	+ \sum_{i=1}^{r+d} Q_{\alpha i} \log{\mu\over \Lambda} 

1667: }

1668: %

1669: where  $t_{\alpha,{\rm bare}} $ are bare parameters 

1670: defined at the momentum cutoff scale $\Lambda$. 

1671: Due to \qcharges, \exc,\  $t_{\alpha,{\rm bare}}\ra-\infty $

1672: in the UV.

1673: As explained in \mm,\ in this regime the higgs branch

1674: describes the $\IC^2/\IZ_{n(p)}$ orbifold.

1675: The theory also contains a coulomb branch, where the lowest

1676: components of $\Sigma_\alpha$ get expectation values.

1677: In the infrared, the higgs branch becomes the \HJ\ space \mm.

1678: We will call branes wrapping two-cycles in this resolved space ``higgs

1679: branch branes''.

1680: Another important property of the infrared physics is

1681: decoupling of the coulomb branch from the higgs branch.

1682: The former develops a set of massive isolated minima.

1683: Some fractional branes become B-branes ``supported'' at these coulomb 

1684: branch vacua, as twisted tachyons condense \mm.

1685: We call such branes ``coulomb branch branes''.

1686: 

1687: For our purposes it will be more convenient to look at the ``mirror''

1688: description of the coulomb branch that follows from the 

1689: approach to mirror symmetry using abelian duality of 2D gauge theory 

1690: of Morrison and Plesser

1691: \MorrisonFR\MorrisonYH. This can 

1692: be cast in terms of an effective Landau-Ginzburg theory \HIV. 

1693: (We follow the line of argument explained in \mm.)

1694: The chiral superfields $X^i$ are eliminated in favor of 

1695: the twisted scalar superfields $Y_i$.

1696: The twisted superpotential in the theory reads

1697: \eqn\horivafa{\eqalign{

1698:         \widetilde W&= \sum_{\alpha=1}^{r} \Sigma_\alpha

1699: \Biggl(\sum_{i=0}^{r+1} Q_{\alpha i} nY_i-t_\alpha(\mu)\Biggr)

1700:                         +\mu \sum_i \lambda_i e^{-nY_i}\cr}

1701: }

1702: %where

1703: %

1704: %\eqn\lamdef{

1705: %  \lambda_{\alpha}=\exp[C^{-1}_{\alpha\beta}t_\beta]

1706: %        \equiv exp[\zeta'_\alpha]

1707: %}

1708: %and $\lambda_{0},\lambda_{r+1}\equiv 1$.

1709: %

1710: %\eqn\teeflow{

1711: %t_{\alpha}(\mu):= t_{\alpha,{\rm bare}} 

1712: %+ \sum_{i=0}^{r+1} Q_{\alpha i} \log\bigl({\mu\over \Lambda}\bigr) \ .

1713: %}

1714: %

1715: Integrating out the $Y_i$ gives the effective superpotential

1716: of \WittenYC\MorrisonFR, while

1717: eliminating instead the $\Sigma_\alpha$ and $Y_\alpha$ 

1718: $\alpha=1,...,r$, gives

1719: (in terms of $u_{0}=(\mu\lambda_0)^{1/n}\exp[-Y_{0}]$ 

1720: and $u_{r+1}=(\mu\lambda_{r+1})^{1/n}\exp[-Y_{r+1}]$)

1721: \eqn\twspotl{

1722:   \widetilde W = u_0^n + u_{r+1}^n 

1723: 		+ \sum_{\alpha=1}^r

1724:                 \lambda_\alpha' u_0^{p_\alpha}u_{r+1}^{q_\alpha}\ ,

1725: }

1726: where 

1727: %

1728: \eqn\newlambds{

1729:   \lambda_{\alpha}' = \lambda_{\alpha} 

1730: 	\;\Lambda^{1-\Delta_{\alpha}}

1731: 	\;e^{t'_{\alpha,{\rm bare}}} 

1732: 	= \lambda_{\alpha}

1733: 		\;\mu^{1-\Delta_{\alpha}}

1734: 		\;e^{t'_{\alpha,\eff}(\mu)}

1735: \ .

1736: }

1737: %The massless noncompact region of the sigma model at

1738: %large $|X|$ is thus (by the second of equations \horivafa)

1739: %at large $Re(Y)$ and thus small $u$.

1740: %

1741: The scaling dimensions of the $\alpha$'s operator in the sum

1742: identifies it with the ring generator $W_\alpha$.

1743: Two important comments are in order

1744: 

1745: \item{(1)} The LG description involves dualizing

1746: the phases of $X_i$, and hence is not well defined 

1747: near the higgs branch, where $X_\alpha=0$.

1748: We do expect it to give a correct description of the

1749: coulomb branch though.

1750: As we will see later in the paper, we will be able to 

1751: recover the description of the higgs branch as well.

1752: The essential ingredient will be the identification of

1753: the higgs branch brane with a combination of the fractional

1754: branes with the help of intersection form.

1755: \item{(2)}

1756: The `mirror' $\IZ_n$ transformation

1757: \eqn\mirzn{

1758:   (u_0,u_{r+1})\sim(\omega u_0,\omega^{-p}u_{r+1})

1759: }

1760: leaves the effective superpotential \twspotl\ invariant --

1761: it fixes all the $\Sigma'_\alpha$.  Indeed it is a 

1762: {\it gauge symmetry} remnant of the duality transformation 

1763: and therefore we should quotient the LG model by its action.  

1764: 

1765: \noindent

1766: In this paper we will be concerned with the simplest 

1767: case of a continued fraction of length one.

1768: This corresponds to $p=1$, $r=1$, $a_1=n$.

1769: There are two fields, $x_1\equiv u_0$ and $x_3\equiv u_2$.

1770: The tachyon expectation value is determined by the

1771: parameter $w\equiv \lambda'_1$.

1772: The LG superpotential is

1773: \eqn\lgsp{ W=x_1^n+w x_1 x_3+x_3^n  }

1774: and the theory should be quotiented by 

1775: \eqn\lgq{ (x_1,x_3)\sim (e^{2\pi i\over n}x_1, e^{-{2\pi i\over n}}x_3) }

1776: 

1777: 

1778: 

1779: 

1780: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1781: %

1782: %

1783: %

1784: %

1785: \subsec{Intersection form and the map between LG and fractional branes}

1786: %

1787: We start by reviewing the results of \MP.

1788: There we considered the D-brane intersection form in type II string theory

1789: on the orbifold $\IC^2/\Gamma$ \refs{\DouglasHQ\BerkoozIS-\BerkoozKM}

1790: \eqn\defind{\I_{ab}=\tr_{R,ab} (-1)^F q^{L_0-{c\over 24}}   .}

1791: Here the trace is over the states of the open string suspended

1792: between D-branes which correspond to representations of $\Gamma$

1793: labeled by $a$ and $b$ and $F$ is the worldsheet fermion

1794: number.

1795: In the case of type II theory, this formula can be

1796: written as \MP

1797: \eqn\om{\I_{ab}={4 \over n}\sum_{s=0}^{n-1}\exp\left({2\pi i (b-a)s \over n}\right)

1798:                            \sin\left({\pi s\over n}\right)\sin\left({\pi s p\over n}\right) }

1799: %

1800: The first factor in the sum comes from the action of the group

1801: element on the Chan-Phaton factors, while the product of the $sin$'s 

1802: is due to the fermion zero modes in the R sector.

1803: It is not hard to evaluate \om: 

1804: \eqn\oma{\I_{ab}=\delta_{a-b-{1-p\over 2}}+\delta_{a-b+{1-p\over 2}}

1805:            -\delta_{a-b-{1+p\over 2}}-\delta_{a-b+{1+p\over 2}}         }

1806: where

1807: \eqn\defdm{\delta_a\equiv\delta_{a, 0\;\mn}.}

1808: Note that the arguments of delta functions in \oma\ are

1809: always integers, thanks to the requirement that $p$ is odd.

1810: The matrix $\I$ in \oma\ is written in the basis 

1811: \eqn\bbasic{ e_0,e_1,\ldots,e_{n-1}   }

1812: where $e_a$ is the $a$-th fractional brane\foot{The details of fractional

1813: brane construction in type 0 and type II string theories are summarized

1814: in Appendix A.}.

1815: The intersection matrix $\I$ is invariant under the cyclic

1816: permutation of fractional branes, $e_a\ra e_{a+1}$.

1817: In \MP\ we found a change of basis which block-diagonalizes

1818: $\I$; one of the two blocks is given precisely by the intersection

1819: matrix for the higgs branch branes $C_{\alpha\beta}$.

1820: However this block-diagonalization is clearly invariant under

1821: the cyclic permutation of the fractional branes; in other

1822: words, if certain expressions for the higgs branch branes

1823: \eqn\hone{ h_\alpha=\sum_a H_{\alpha a} e_a, \quad H_{\alpha a}\in \IZ  } 

1824: give rise to the intersection matrix $C_{\alpha\beta}$, then

1825: the expressions with the indices of fractional branes permuted,

1826: \eqn\htwo{ h_\alpha=\sum_a H_{\alpha a} e_{a+1}  } 

1827: are equally good.

1828: We will see later in the paper that, thanks to the

1829: dependence of the map between fractional and coulomb/higgs branch branes

1830: on the tachyon VEV, all possible cyclic permutation 

1831: of fractional branes are realized, depending on the phase of the

1832: tachyon expectation value.

1833: For now we assume the basis \bbasic\ for simplicity, keeping the permutation symmetry in 

1834: mind.

1835: In fact, it is convenient to switch to the basis 

1836: $\sum_a e_a,e_1,\ldots,e_{n-1}$ which effectively substitutes both the

1837: first row and the first column in $\I$ by a set of zeroes 

1838: (since the D0 brane, $D0=\sum_a e_a$, has zero intersection with any brane

1839: in the theory, including itself).

1840: Now we can omit the first row and column from $\I$-- it

1841: is this reduced matrix, denoted $\It$ which appears in the rest of this paper, unless

1842: stated otherwise.

1843: 

1844: We would like to generalize this discussion for type 0.

1845: The number of branes is now doubled, for there are

1846: branes which are labeled by the choice of sign 

1847: in the gluing conditions (see Appendix A for details).

1848: We will call these branes $\eta=+1$ or $\eta=-1$ branes\foot{

1849: These branes are often called electric and magnetic branes

1850: in the literature, but they are not electric/magnetic duals.

1851: In type 0 theories obtained as orbifolds by $(-1)^{F_{\rm spacetime}}$ 

1852: of the type II string, they are in fact fractional branes \MP.}.

1853: As explained in \MP, when $p$ is odd, the intersection form

1854: of the type 0 theory is simply obtained from the intersection

1855: form $\It$ of type II:

1856: \eqn\izero{ \It_0=\pmatrix{

1857: ~0& ~\It\cr

1858: ~\It^T& ~0 }

1859: }

1860: where $\It^T=\It$. 

1861: Note that fermionic degrees of freedom only exist on the intersections

1862: of branes of different types.

1863: It will be convenient to separate $\eta=+1$ and  $\eta=-1$

1864: fractional branes into

1865: two sets: the ones that are labeled by the ``special representation'' integer $e_{p_\alpha}$

1866: and the rest, $e_\nu$.

1867: The integers $p_\alpha$ are determined by the continued fraction $[a_1,a_2,\ldots a_r]$

1868: via the recursion relations \mm:

1869: \eqn\rrpq{

1870: p_{j-1}/p_j=\left[a_j,a_{j+1},\ldots,a_r\right],\qquad 1\le j\le r

1871: }

1872: with the initial conditions $p_{r+1}=0,\;p_r=1$.

1873: In the basis 

1874: \eqn\basise{  e^{(+)}_{p_1},\ldots,e^{(+)}_{p_r},\{e^{(+)}_{\nu}\},

1875:                      e^{(-)}_{p_1},\ldots,e^{(-)}_{p_r},\{e^{(-)}_{\nu}\}         }

1876: %

1877: we consider the following ansatz for the linear map between

1878: the LG branes (higgs and coulomb branch branes $h_\alpha$  and $c_\alpha$)

1879: and the fractional branes

1880: \eqn\az{ \pmatrix{~h^{(+)}_\alpha\cr~c^{(+)}_\alpha\cr~h^{(-)}_\alpha\cr~c^{(-)}_\alpha}=

1881: \pmatrix{

1882: ~A& ~0\cr

1883: ~0& ~B } \pmatrix{~e^{(+)}\cr e^{(-)}} = 

1884: \pmatrix{~1& ~a& ~0& ~0\cr ~0& ~1& ~0& ~0\cr ~0& ~0& ~1& ~b^T\cr ~0& ~0& ~0& ~1}

1885: \pmatrix{ ~e^{(+)}_{p_\alpha}\cr ~e^{(+)}_{\nu}\cr ~ ~e^{(-)}_{p_\alpha}\cr ~e^{(-)}_{\nu}  }

1886: } 

1887: where $a$ and $b$ are matrices whose entries are integers.

1888: We then require

1889: \eqn\cb{  \pmatrix{~A& ~0\cr ~0& ~B } \pmatrix{~0& ~\It\cr ~\It^T& ~0 } \pmatrix{~A^T& ~0\cr ~0& ~B^T }=

1890:    \pmatrix{~0& ~0& ~C& ~0\cr ~0& ~0& ~0& ~C'\cr ~C& ~0& ~0& ~0\cr ~0& ~C'& ~0& ~0}     }   

1891: One can solve for $a$, $b$ and $C$.

1892: The result is 

1893: \eqn\abc{   a=-x C'^{-1},\qquad b=-C'^{-1} y,\qquad C=C_1-x C'^{-1} y    }

1894: where $C_1$, $C'$, $x$ and $y$ are the components of $\It$:

1895: \eqn\icomps{   \It=\pmatrix{~C_1& ~x\cr ~y& ~C' }          }

1896: Note that the map determined by \abc\  

1897: is determined up to an addition/subtraction of any multiple of D0 branes,

1898: since the latter have zero intersection with any fractional brane. 

1899: 

1900: When $p$ is odd, the orbifold group is $\IZ_n$, one can define type II theory,

1901: and $\It=\It^T$.

1902: In this case $\It$ is given by the reduction of \oma, and $a=b^T$.

1903: In \MP\ it was shown that if in addition all $a_\alpha$ are even, then

1904: the entries of $a$ (and $b$) in \abc\ are integers 

1905: and $C$ computed in \abc\ coincides with \exc.

1906: The closed string sector of type 0 theory is invariant 

1907: under $p\ra p+n$.

1908: However eq. \om\ is not invariant under such a shift.

1909: This is a manifestation of the fact that one can define two sets

1910: of branes in the type 0 theory.

1911: These two types of branes will have intersection matrices

1912: corresponding to two different types of resolution, as discussed above.

1913: A simple set of examples considered in \MP\ is $p=1, n-1$.

1914: The intersection form for the $p=1$ case reproduces the Cartan matrix 

1915: which defines the supersymmetric ALE singularity.

1916: For $p=n-1$ there is a single higgs branch brane $h$ which wraps the base of $\OO(-n)\ra \IP^1$

1917: which is the \HJ\ manifold in this case.

1918: The change of basis in the $p=n-1$ case is nontrivial; here we quote 

1919: the result obtained in \MP:

1920: \eqn\cbpn{   h=e_1+2 e_2+\ldots+{n\over 2}e_{n\over 2}-

1921:     \left({n\over 2}-1\right)e_{{n\over 2}+1}-\ldots-e_{n-1}           }

1922: 

1923: Suppose now $p$ is even.

1924: (And, consequently, $n$ is odd as we restrict our 

1925: consideration to $n$, $p$ relatively prime.)

1926: In this case the orbifold group is $\IZ_{2n}$ and expression \om\

1927: for the unreduced matrix $\I_0$ should be modified to

1928: \eqn\omb{\I_{0,ab}={4 \over 2 n}\sum_{s=0}^{2n-1}\exp\left({2\pi i (a-b)s \over 2 n}\right)

1929:                            \sin\left({\pi s\over n}\right)\sin\left({\pi s p\over n}\right) }

1930: %

1931: The indices $a$, $b$ which label the branes now run from $0$ to $2n-1$.

1932: $n$ is substituted by $2n$ in the prefactor and in the first factor

1933: in the sum-- this is the result of the order of the orbifold

1934: group becoming $\IZ_{2n}$.

1935: The matrix element in \omb\ evaluates to zero whenever $a-b$ is even.

1936: This is an indication that $a$, $b$ now label not only different types

1937: of fractional branes, but also $\eta=+1$ and $\eta=-1$ types.

1938: We can define $a=2 a'$ and $b=2b'+1$ with $a'=0,\ldots,n-1$  and  $b'=0,\ldots,n-1$ 

1939: labeling $\eta=+1$ and $\eta=-1$ branes respectively.

1940: The intersection form is then of the form \izero, although 

1941: $\I$ is no longer a symmetric matrix:

1942: \eqn\inonsym{  \I_{a'b'}=\delta_{a'-b'+{p\over 2}}+\delta_{a'-b'+1-{p\over 2}}

1943:            -\delta_{a'-b'-{p\over 2}}-\delta_{a'-b'+1+{p\over 2}}         }

1944: To use \abc\ we still need to factor out the $D0$ brane by omitting

1945: the first row and the first column.

1946: At this point it is worth mentioning the following

1947: important issue that was not present in the type II theory.

1948: We can shift the $\eta=-1$ branes by $e^{(-)}_{a'}\ra e^{(-)}_{a'+1}$,

1949: without shifting the  $\eta=+1$ branes.

1950: This leads to permutation of columns, $\I_{a',b'}\ra \I_{a',b'+1}$.

1951: After the reduction to the subspace which does not contain a D0 brane,

1952: we obtain $n$ inequivalent matrices $\It$ this way.

1953: In this paper we mostly discuss the $p=1$ case for $n=3,4$.

1954: As explained in Section 5, all inequivalent matrices $\It$ obtained as

1955: described above, admit a block-diagonalization of the form \cb.

1956: The analysis for general $n$ and $p$ will be reported elsewhere.

1957: 

1958: 

1959: 

1960: %For any $n$ that we checked, $a$ and $b$ computed from \abc\ have only integer

1961: %entries and $C$ in \abc\ is given by \exc.

1962: %The expressions for electric and magnetic higgs branch branes appear to be

1963: %\eqn\hbbel{     h^{(el)}=e_1^{(el)}+2 e_2^{(el)}+\ldots+{n-1\over 2}e_{n-1\over 2}^{(el)}-

1964: %    {n-1\over 2}e_{{n-1\over 2}+1}^{(el)}-\ldots-e_{n-1}^{(el)}           }

1965: %and

1966: %\eqn\hbbm{     h^{(m)}=e_1^{(m)}+2 e_2^{(m)}+\ldots+{n+1\over 2}e_{n+1\over 2}^{(m)}-

1967: %    \left({n+1\over 2}-2\right)e_{{n+1\over 2}+1}^{(m)}-\ldots-e_{n-1}^{(m)}           }

1968: 

1969: 

1970: 

1971: 

1972: 

1973: 

1974: 

1975: 

1976: 

1977: 

1978: 

1979: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1980: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1981: \newsec{D-branes in the LG mirror and generalized central charges}

1982: 

1983: The superpotential of the LG model providing a mirror description of the GLSM \glsmact\

1984: has a set of critical points.

1985: According to \HIV\ each critical point gives rise to a

1986: D-brane with A-type boundary conditions.

1987: This D-brane, which we will sometimes refer to as an A-brane surface,

1988: is a Lagrangian submanifold of $\IC^2$ whose

1989: image in the $W$-plane is a half-line which starts at the critical value and extends

1990: in the positive real direction.

1991: The critical points away from $(x_1,x_3)=(0,0)$ correspond

1992: to the coulomb branch branes, while the single critical point

1993: at the origin gives rise to a higgs branch brane

1994: (recall that we specialize to the case of continued fraction of

1995: length one)

1996: %We should really consider the orbifold of the LG model, 

1997: %so taking quotient by \lgq\ is implied.

1998: 

1999: As explained in \HIV\ a practical way to determine 

2000: these surfaces is to consider the soliton equation

2001: \eqn\soleq{  {d x^i\over d\sigma}={1\over2} g^{i\bar j} \p_{\bar j} \bar W   }

2002: where $ g_{i\bar j}$ is the Kahler metric and $\sigma$ is the coordinate

2003: along the soliton trajectory.

2004: The soliton trajectory is supposed to originate from a critical point,

2005: where the Kahler metric is nonsingular.

2006: (This leads to problems when considering the higgs branch brane, whose

2007: critical point is at $(x_1,x_3)=(0,0)$ where the Kahler metric

2008: cannot even be reliably determined).

2009: The shape of the A-brane associated with a given critical point is

2010: the set of all trajectories satisfying \soleq.

2011: Near the critical point this shape can be easily determined.

2012: One needs to find the coordinates $u^i$ which diagonalize the system \soleq.

2013: In these coordinates, the solutions are $u^i=u^i_0 \exp(\lambda^i \sigma)$,

2014: so that $u^i\ra0$ as $\sigma\ra-\infty$.

2015: Consistency of the equations \soleq\ then forces $u^i_0 \in \IR$.

2016: A convenient way of parameterizing a set of solutions  \HIV\ is considering

2017: a sphere of small radius $\epsilon$

2018: \eqn\sphere{  \sum_i (u^i)^2=W(\phi^i_*+u^i)-W(\phi^i_*)=\epsilon^2,\qquad u^i\in \IR       }

2019: As $\sigma$ increases, this small sphere (``the wavefront'') will evolve; the surface traced by it

2020: in this process is the A-brane.

2021: In Appendix B we analyze these surfaces numerically.

2022: We find that in the models we consider they resemble a propeller,

2023: with various segments of the small circle developing into quarter-plane ``wings''.

2024: 

2025: An interesting object that one can consider in the LG models

2026: is an overlap of a RR ground state\foot{

2027: The correspondence is realized by performing the worldsheet path integral

2028: on the semiinfinitely long cigar with no insertions, but in the twisted theory,

2029: to produce a RR state.},

2030: corresponding to the identity operator,

2031: with the D-brane specified by the boundary state $| B\rangle$ \HIV.

2032: To make this object holomorphic, one needs to consider

2033: the formal limit $\bar W\ra 0$ \HIV.

2034: In this case, the overlap can be computed as an integral over the

2035: A-brane surface:

2036: \eqn\gcc{  \langle B|1\rangle=\int\int {dx_1\over x_1} {dx_3\over x_3} \exp(-W)   }

2037: As we will see in the next section, 

2038: this integral is convergent, since $W\ra +\infty$ in the asymptotic region.

2039: Moreover, in the next section we will also see that \gcc\ 

2040: satisfies the GKZ equation.

2041: In the supersymmetric case the solutions 

2042: would give rise, via mirror symmetry, to the integrals

2043: of the complexified Kahler form over the cycles in the

2044: higgs branch of the GLSM.

2045: This motivates us to call the quantity \gcc\ ``the generalized 

2046: central charge'' of the D-brane described by the boundary state $|B\rangle$.

2047: In the non-Calabi-Yau case of this paper,

2048: the higgs branch brane in the GLSM still has a geometric interpretation

2049: of a brane wrapping a two-cycle in the \HJ\ space. 

2050: 

2051: Note that the branes described above correspond to the B-branes in the original GLSM and orbifold

2052: theories with $\eta=1$.

2053: Before performing the GSO projection, the theory contains a second set of

2054: branes with $\eta=-1$, which preserve a different combination 

2055: of the worldsheet $\NN=2$ supersymmetry.

2056: Such branes are described by eq. \soleq\ with an additional

2057: minus sign in the right-hand side.

2058: The image of these branes in the $W$-plane would therefore be half-lines extending in the

2059: negative real direction.

2060: The definition \gcc\ would have to be modified accordingly.

2061: 

2062: 

2063: 

2064: 

2065: 

2066: 

2067: 

2068: 

2069: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2070: 

2071: \newsec{Generalized periods  for the $\IC^2/\IZ_{n(1)}$ orbifold } 

2072:   

2073: 

2074: In this section we consider the generalized periods for the  LG superpotential 

2075: \eqn\lgpot{

2076: W := a_1 x_1^n + a_2 x_1 x_3 + a_3 x_3^n 

2077: }

2078: %

2079: This section is rather technical. A summary is found in the final subsection.  

2080: 

2081: 

2082: \subsec{GKZ Equation} 

2083: 

2084: 

2085: Associated to a toric manifold is a canonically determined system of differential 

2086: equations, the GKZ system of differential equations. For toric hypersurfaces in 

2087: CY manifolds these equations are related to the Picard-Fuchs equations of the mirror 

2088: and therefore solutions are related to the periods of the mirror variety. 

2089: 

2090: In the present case, using the toric data of the $\IC^2/\IZ_{n(1)}$ manifold one

2091: finds the differential equation 

2092: %

2093: \eqn\gkzeqi{

2094: \biggl[ \Theta^2 -  z (n \Theta) (n \Theta +1) \cdots (n\Theta + n-1)\biggr] F(z)=0 

2095: }

2096: %

2097: with $\Theta = z{ d\over dz}$. 

2098: $z=\infty$ is a regular singular point, and corresponds to the orbifold point. 

2099: $z=0$ is an irregular singular point and corresponds to the ``IR limit.'' 

2100: 

2101: 

2102: To derive \gkzeqi\  one begins 

2103: with the fan  $v_1 = (0,1), v_2 = (1,0), v_3 = (n,-1)$ to produce the differential 

2104: operator 

2105: %

2106: \eqn\diffop{ 

2107: \CD:= (\p_{a_1} \p_{a_3} -  (-1)^n\p_{a_2}^n)    

2108: }

2109: %

2110: Now, again using the toric vectors one defines an invariant combination   $ z := a_1 a_3/a_2^n$.  

2111: When acting on a function depending on $a_i$ only through $z$ it is 

2112: straightforward to show that 

2113: %

2114: \eqn\simpdep{

2115:  \CD f(z) = {1\over a_1 a_3} \biggl[ \Theta^2 -  z (n \Theta) (n \Theta +1) \cdots (n\Theta + n-1)\biggr] F(z)

2116: }

2117: %

2118: This establishes \gkzeqi. 

2119: 

2120: A useful change of variable $w= z^{-1/n}=a_2/(a_1 a_3)^{1/n}$ brings the GKZ equation to the form: 

2121: %

2122: \eqn\gkzsimp{

2123: \biggl[ -\bigl(-{d\over dw} \bigr)^{n} + {1\over n^2} \bigl(w{d\over dw}\bigr)^2 \biggr]

2124:   f(w) =0 

2125: }

2126: %

2127: 

2128: 

2129: The $n$-dimensional space of solutions to this equation is the $n$-dimensional space of 

2130: generalized periods. The constant solution - which will be associated with the $D0$ brane - 

2131: is somewhat trivial and we define $\CV$ to be the space of nonconstant solutions which vanish at 

2132: $w=0$. Much of the work in   Section 5 of this paper will be 

2133: writing down different bases for $\CV$  and interpreting them physically.  

2134: One basis of solutions is obtained by straightforward application of the Frobenius technique. 

2135: This is: 

2136: %

2137: \eqn\fpp{

2138: \eqalign{

2139: \hat f_{-1}(w)&:=1 \cr

2140: \hat f_m(w) & := (-w)^{m+1} \sum_{k=0}^\infty (-w)^{nk} {\bigl(\Gamma(k + {m+1\over n})\bigr)^2\over \Gamma(kn+m+2) } 

2141: \qquad 0 \leq m \leq n-2\cr}

2142: }

2143: %

2144: Note that $\hat f_m(w)$ are entire functions in the $w$-plane for $n>2$ with an essential singularity at $w=\infty$. 

2145: It will be useful to extend the definition \fpp\ to include $\hat f_{n-1}(w)$, although this function does 

2146: {\it not} solve the GKZ equation. 

2147: 

2148: 

2149: \subsec{Integral representation} 

2150: 

2151: In the physical interpretation of the solutions it is very useful to have an integral 

2152: representation directly related to the path integral which computes the overlap between 

2153: RR groundstates and the branes. This solution has the schematic form 

2154: %

2155: \eqn\lgi{

2156: \int_{\gamma} e^{- W}  {dx_1\over x_1} {dx_3 \over x_3}

2157: }

2158: %

2159: where $\gamma$ is an appropriate contour, to be discussed in detail below. 

2160: 

2161: 

2162: Note that it is trivially true that for $W$ given by \lgpot\ and $\CD$ given by \diffop\

2163: %

2164: \eqn\lgiact{

2165: \CD \int_{\gamma} e^{- W}  {dx_1\over x_1} {dx_3 \over x_3} =0

2166: }

2167: %

2168: %where 

2169: %

2170: %\eqn\fixdee{

2171: %\CD = (\p_{a_1} \p_{a_3} -  (-1)^n\p_{a_2}^n)

2172: %}

2173: %

2174: %and    

2175: where $\gamma$ is any {\it fixed} chain of real dimension 2 in $\IC^* \times \IC^*$. 

2176: 

2177: 

2178: Now, the generalized periods should only be functions of the scaling variable $w$. 

2179: This can be arranged by exploiting the $\IC^*\times \IC^*$ action on $(x_1,x_3)$. 

2180: %By analytic continuation it suffices to establish solutions of the form \lgi\ for $a_i>0$. Then we need only 

2181: In fact, by analyticity, we need only consider $\IR^*_+ \times \IR^*_+$ invariant orbits. Any such orbit through $(\alpha,\beta) \in \IC^*\times \IC^*$ is of the 

2182: form 

2183: %

2184: \eqn\gami{

2185: \gamma_{\alpha,\beta} := \{ (t_1 \a, t_3 \b)\vert t_1,t_3>0 \}.

2186: }

2187: %

2188: One might be tempted to use the chains $\gamma_{\alpha,\beta}$ in \lgi\ to produce solutions of 

2189: \gkzsimp. There are two problems with this. 

2190:  First,  we must also regularize the logarithmic singularities 

2191: at the origin. Next,  we must 

2192: ensure convergence of the integral at $\infty$. The singularity at the origin is easily regularized 

2193: by taking   $t_i\geq \e$. Thus we consider  the contour integral:

2194:  %

2195: \eqn\regin{

2196: I_{\a,\b}(w;\e):= \int_\e^\infty {dt_1\over t_1} \int_\e^\infty {dt_3\over t_3} e^{-W} 

2197: }

2198: %

2199: where  

2200: %

2201: \eqn\neve{

2202: W = \a^n t_1^n + \a \b w t_1 t_3 + \b^n t_3^n

2203: }

2204: %

2205: with $t_1, t_3>0$ always. 

2206: 

2207: 

2208: We will use this basic integral to construct solutions. To  ensure convergence at infinity  we require

2209: $\re(\a^n)>0$ and $\re(\b^n)>0$. There are different sectors of the $(x_1,x_3)$ plane in which the 

2210: integrals converge. We will need a way of denoting these {\it convergent sectors}.  

2211: Define  $\a = \vert \a\vert e^{i \theta}$. Then if   $\a$ is in a convergent sector 

2212: there must exist an integer  $s_\a$ such  that 

2213: %

2214: \eqn\voncen{

2215: - {\pi \over 2n} <  \theta + {2\pi \over n} s_\a < {\pi \over 2n} 

2216: }

2217: %

2218: We will denote the convergent sectors \voncen\ in the $\alpha$ plane by $\CS_{s_\a}$. 

2219: Equation \voncen\ only defines $s_\a$ modulo $n$. If we choose the principal branch of the 

2220: logarithm then $s_\a$ is defined absolutely. We choose the fundamental domains:

2221: \foot{ This has one awkward feature for $n$ even: In this case 

2222: the angular sector containing the negative real axis is a convergent sector and $s$ is discontinuous 

2223: across the negative real axis, jumping from $s=-n/2$ just above the negative real axis to $s=n/2$

2224: just below the negative real axis. }

2225: %

2226: %

2227: \eqn\rangeess{

2228: \eqalign{ 

2229: & s= {n-1\over 2}, {n-3\over 2}, \dots, - {n-1\over 2} \qquad \qquad n \quad odd \cr

2230: & s= \half n ,   \dots, - \half n \qquad \qquad n \quad even \cr}

2231: }

2232: %

2233: See Figs. 1,2 for $n=3,4$. 

2234: \midinsert\bigskip{\vbox{{\epsfxsize=3in

2235:         \nobreak

2236:     \centerline{\epsfbox{xcn3.eps}}

2237:         \nobreak\bigskip

2238:     {\raggedright\it \vbox{

2239: {\bf Fig 1.}

2240: {\it The angular structure in the $\alpha$, $\beta$ 

2241: (equivalently, $x_1$, $x_3$) planes for $n=3$.  In the shaded sectors

2242: the integral \regin\ converges.}}}}}}

2243: \bigskip\endinsert

2244: %

2245: \midinsert\bigskip{\vbox{{\epsfxsize=3in

2246:         \nobreak

2247:     \centerline{\epsfbox{xcn4.eps}}

2248:         \nobreak\bigskip

2249:     {\raggedright\it \vbox{

2250: {\bf Fig 2.}

2251: {\it The angular structure in the $\alpha$, $\beta$ 

2252: (equivalently, $x_1$, $x_3$) planes for $n=4$.  In the shaded sectors

2253: the integral \regin\ converges.}}}}}}

2254: \bigskip\endinsert

2255: 

2256: 

2257: By using rescalings of $t_1,t_3$, the equation $\CD e^{-W}=0$, and  keeping track of boundary terms, and 

2258: \simpdep\  it is straightforward to show that 

2259: %

2260: \eqn\difsnac{

2261: a_1 a_3 \CD I_{\a,\b}(w;\e) =  {1\over n^2} e^{-W(t_1=\e, t_3 = \e)} 

2262: - {\a^n\over n} \int_\e^\infty dt_1 t_1^{n-1} e^{-W(t_1, t_3=\e) }

2263:  - {\b^n\over n} \int_\e^\infty dt_3 t_3^{n-1} e^{-W(t_1=\e, t_3) }

2264: }

2265: %

2266: Thus, the $I_{\a,\b}(w;\e)$ do {\it not} solve the differential equation. 

2267: Note, however, that the RHS of \difsnac\ has a smooth $\e\to 0$ limit, 

2268: given by $-1/n^2$

2269: %

2270: %\eqn\smthlm{

2271: %{1\over n^2}(1-1-1) = - {1\over n^2} 

2272: %}

2273: %

2274: 

2275: 

2276: We will form solutions from the $I_{\a,\b}$ by taking appropriate linear combinations

2277: of $I_{\a,\b}(w;\e)$ and taking the $\e\ra0$ limit. 

2278: In order to produce appropriate combinations let us investigate the $\epsilon \to 0$ 

2279: behavior.  Note that $I_{\a,\b}(w;\e)$ has a   convergent expansion in $w$.  

2280: If we expand the integrand

2281: in \regin\  in a power 

2282: series in $w$, only the first term has a divergence as $\e \to 0$. Let us define 

2283: $\hat I_{\a, \b}(w;\e)$ via:

2284: %

2285: \eqn\ddfihat{

2286: I_{\a, \b}(w;\e) =  \int_\e^\infty {dt_1\over t_1}e^{-\a^n t_1^n}  \int_\e^\infty {dt_3\over t_3} e^{-\b^n t_3^n}  + 

2287: \hat I_{\a, \b}(w;\e)

2288: }

2289: %

2290: The first term is divergent for $\epsilon \to 0$, and constant in $w$. 

2291: In fact, one has: 

2292: %

2293: \eqn\consttrm{

2294: \int_\e^\infty {dt_1\over t_1}e^{-\a^n t_1^n} = {1\over n} E_1(\epsilon^n \a^n) = - \log \e - {1\over n} \log \a^n - {\gamma\over n} + \CO(\e^n).

2295: }

2296: %

2297: Thus we conclude that  

2298: %

2299: \eqn\ilime{

2300: I_{\a,\b}(w;\e) \rightarrow (\log \e)^2  + K_{\a\b} \log \e  + {1\over n^2}(\log \a^n + \gamma)(\log \b^n + \gamma) + \hat I_{\a\b}(w;0) + \CO(\epsilon)

2301: }%

2302: %

2303: In particular 

2304: %

2305: \eqn\lieim{

2306: {d\over dw} I_{\a,\b}(w;\e)={d\over dw} \hat I_{\a,\b}(w;\e)

2307: }

2308: %

2309: has a smooth $\e\to 0$ limit. 

2310: 

2311: 

2312: It follows from \ilime\ that if we choose linear combinations 

2313: %

2314: \eqn\linec{

2315: \sum c_{\a\b} \hat I_{\a,\b}(w;0) 

2316: }

2317: %

2318: such that 

2319: %

2320: \eqn\linde{

2321: \sum c_{\a\b} =0  

2322: }

2323: %

2324: then we have a solution of the GKZ equation \gkzsimp. Indeed, we can take the limit $\e\to 0$ 

2325: directly from \ddfihat\   and 

2326: we  can compute the power series in 

2327: $w$ explicitly. Using

2328: %

2329: \eqn\simplin{

2330: \int_0^\infty {dt\over t} t^k e^{-\a^n t^n} = \a^{-k} e^{-2\pi i {k\over n} s_\a} {1\over n} \Gamma({k\over n}) \qquad k\geq 1

2331: }

2332: %

2333:   we arrive at 

2334: %

2335: \eqn\itoeff{

2336: \eqalign{

2337: \hat I_{\a,\b}(w;0) & = {1\over n^2} \sum_{k=1}^\infty (-w)^k {(\Gamma(k/n))^2\over k!} e^{-2\pi i {k\over n}(s_\a + s_\b)} \cr

2338:  & = {1\over n^2} \sum_{j=1}^n e^{-2\pi i {j\over n} (s_\a + s_\b)} \hat f_{j-1}(w) \cr}

2339: }

2340: %

2341: where $\hat f_m$ was defined in \fpp. 

2342: %The definition there can be extended to 

2343: %$\hat f_{n-1}$, although 

2344: Recall that $\hat f_{n-1}$ does {\it not} solve the differential equation \gkzsimp. This is in accord with \difsnac. 

2345: 

2346: 

2347: Let us note a few important properties of the functions $\hat I_{\a,\b}(w;0)$. First, it is clear from 

2348: \itoeff\ that  $\hat I_{\a,\b}(w;0) $ only depends on $\a,\b$ through the sector $\CS_{s_\a} \times \CS_{s_\b}$.  

2349: In fact $\hat I_{\a\b}$ only 

2350: depends on the combination $s_{\a\b} := (s_\a + s_\b) \mod~ n$. By definition, the function  $\hat I_s$ is $\hat I_{\a\b} $ 

2351: such that $s_{\a\b}=s \mod ~n$. Next, note that the functions

2352: %

2353: \eqn\newslkc{

2354: \sum c_{s} \hat I_{s}(w;0)

2355: }

2356: %

2357:  subject to $\sum c_{s}=0 $

2358: %is in fact an overcomplete {\it basis} of 

2359: span the space $\cal V$ of nonconstant solutions to the GKZ system.  

2360: Indeed, inverting the finite Fourier transform we see that the span of $\hat I_{\a,\b}(w;0)$ 

2361: is the span of $\hat f_0, \dots, \hat f_{n-1}$, and that $\hat f_m$ for $m<n-1$ are given by linear combinations 

2362: with $\sum c_{\a\b}=0 $.  Finally, note   that $\sum c_{\a\b} I_{\a\b}(w;\epsilon)$ only has a smooth limit for 

2363: $\sum c_{\a\b}=0$ and $\sum c_{\a\b} K_{\a\b}=0$. In this case 

2364: $\lim_{\e\to 0} \sum c_{\a\b} I_{\a\b} = \sum c_{\a\b} \hat I_{\a\b} $. 

2365: 

2366: \subsec{LG symmetry} 

2367: 

2368: 

2369: The ``LG symmetry'' or ``quantum symmetry'' plays a  fundamental role in what follows. 

2370: The point $w=0$ corresponds to the orbifold point, where there is a quantum symmetry that 

2371: permutes the fractional branes. From the action of this symmetry on the chiral ring 

2372: generator we see that the action of the  quantum symmetry is $w \to \omega w$, where $\omega  = e^{2\pi i /n}$, so a 

2373: generator of the quantum symmetry takes $\hat I_s \to \hat I_{s+1}$, that is: 

2374: %

2375: \eqn\lgsymm{

2376: \hat I_s(\omega^t w) = \hat I_{s-t}(  w)

2377: }

2378: %

2379: 

2380: 

2381:  

2382: 

2383: Thus the set of solutions  $\hat I_1-\hat I_0, \hat I_2-\hat I_1, \dots, \hat I_{n-1}-\hat I_{n-2}, \hat I_0 - \hat I_{n-1}$ 

2384: are cyclically permuted under the quantum symmetry. The sum of these solutions is $0$. Thus, this basis is 

2385: reminiscent of the   the space of fractional branes orthogonal to the $D0$ brane, and one might 

2386: be tempted to identify these with the periods of fractional branes.   Unfortunately, quantum monodromy is not 

2387: strong enough to guarantee this and we will see that in fact a more subtle basis corresponds to the 

2388: basis of fractional branes.  

2389: 

2390: 

2391: 

2392: \subsec{Critical points and propeller branes} 

2393: 

2394: 

2395: The integral representation \lgi\ is useful for investigating the asymptotics of the solutions

2396: via stationary phase.  

2397: %Let us consider the stationary phase approximation to \lgi. 

2398: In this section we set $a_1=a_3=1$ so 

2399: $$

2400: W = x_1^n + w x_1 x_3 + x_3^n.

2401: $$

2402: Of course, solving for $W'=0$ is the same as solving 

2403: for the LG vacua. We find there are $(n-1)^2$ solutions. $n(n-2)$ solutions are nonzero and come in 

2404: $(n-2)$ different ``Landau-Ginzburg 

2405: orbits'' of the quantum $\IZ_n$ symmetry $(x_1,x_3) \to (\omega x_1, \omega^{-1} x_3)$. 

2406: 

2407: We should simply write: 

2408: %

2409: \eqn\lgiiip{

2410: \eqalign{ 

2411: x_1 & = (-{w\over n})^{1\over n-2} e^{2\pi i  {\nu (n-1)\over n(n-2)} } \cr

2412: x_3 & = (-{w\over n})^{1\over n-2} e^{2\pi i   {\nu \over n(n-2)}  } \cr}

2413: }

2414: %

2415: with $\nu = 1, \dots, n(n-2)$.

2416: % Note this can be written 

2417: %

2418: %\eqn\ismplr{

2419: %(x_1,x_3)_\nu = (\chi_\nu(w) e^{2\pi i {\nu\over n}} , \chi_\nu(w) ) 

2420: %}

2421: %

2422: The LG $\IZ_n$ symmetry is $\nu \to \nu + (n-2)$. 

2423: The remaining critical point is at $(x_1,x_3)=(0,0)$. 

2424: 

2425: 

2426: 

2427: 

2428: The critical value of $W$ at \lgiiip\ is 

2429:  %

2430: \eqn\lgivp{

2431: W_\nu = (2-n) \bigl({-w\over n}\bigr)^{{n\over n-2}} e^{2\pi i {\nu  \over n-2}} 

2432: }

2433: %

2434:  Note that the different LG orbits are  separated by the value of $W$ on the orbit. 

2435: It is also useful to compute the Hessian: 

2436: 

2437: %

2438: \eqn\hess{

2439: W_\nu'' = - w\pmatrix{(n-1)e^{-2\pi i \nu/n} & - 1 \cr -1 & (n-1)e^{2\pi i \nu/n}\cr} 

2440: }

2441: %

2442:  Note that 

2443: $$

2444: \det W'' = w^2 n(n-2)

2445: $$

2446: does not depend on the sector. All critical points are Morse critical points. 

2447: 

2448: 

2449: Associated to each critical point is a vacuum state in the GLSM. Associated with 

2450: each vacuum is a (topological) D-brane.   To write the generalized period for the brane we 

2451: define $\Gamma_\nu$ to be the A-brane surface

2452: %soliton surface $\Gamma_\nu$ 

2453: defined in \HIV (see Section 3). In the math 

2454: literature these are known as  ``Lefshetz thimbles'' and in the context of this paper they 

2455: are the propeller branes. We then define

2456: %

2457: \eqn\coulbm{

2458: C_\nu(w):= \int_{\Gamma_\nu} e^{-W} {dx_1\over x_1}{dx_3\over x_3} 

2459: }

2460: %

2461: These are well-defined for the nonzero critical points. 

2462: The period $C_0$ for brane corresponding to $(x_1,x_3)=(0,0)$ cannot be obtained from 

2463: \coulbm\ because the latter has a logarithmic 

2464: singularity. 

2465: 

2466: 

2467: 

2468: Although the chains $\Gamma_\nu$ are not $\IR^* \times \IR^*$ invariant the functions 

2469: $C_\nu$ nevertheless satisfy the GKZ equation. The reason is that the chains $\Gamma_{\nu}$ 

2470: approach a linear combination of the chains $\gamma_{\a\b}$ at infinity, and, by Cauchy's theorem 

2471: we can deform contours in a compact region without changing the integral. In fact, by 

2472: examining the asymptotic behavior of the functions $C_{\nu}(w)$ at large $w$ 

2473: we find that $C_\nu$ with $\nu=0,1,\dots, n-2$  define a basis for the space $\CV$, and therefore there is a locally constant 

2474: matrix $M_{\nu s}$ such that 

2475: %

2476: \eqn\matrx{

2477: C_\nu = \sum_{s} M_{\nu s} \hat I_s 

2478: }

2479: %

2480: This matrix is only locally constant. It will be constant in angular sectors and will change 

2481: discontinuously across angular sectors because the solutions to an equation with an irregular singular 

2482: point exhibit Stokes' phenomenon. 

2483: 

2484: 

2485: 

2486: 

2487: \subsec{A  basis of asymptotic solutions} 

2488: 

2489: 

2490: Let us now consider the asymptotic behavior of the functions in $\CV$ at large $w$.

2491: From the saddle point formula the contribution of the critical point of type $\nu$ to $C_\nu$ is: 

2492:  %

2493: \eqn\spcontr{

2494: \epsilon_{\nu}  {1\over \sqrt{n^3(n-2)}} \bigl({-w\over n} \bigr)^{-n\over n-2} e^{-2\pi i {\nu\over n-2} } 

2495: \exp\biggl[ (n-2) (-{w\over n})^{n\over n-2} e^{2\pi i {\nu\over n-2}} 

2496: \biggr] 

2497: }

2498: %

2499: where $\epsilon_\nu = \pm 1$. This is the leading term in an asymptotic  

2500: expansion given by expanding the integrand of 

2501: %

2502: \eqn\asympsolii{

2503:  \eqalign{

2504: &  e^{-W_\nu} \int_{-\infty}^{+\infty} {dt_1 dt_3 \over (x_1  +t_1)

2505: (x_3  +t_3)} e^{-\half \pmatrix{t_1 & t_3\cr}W''\pmatrix{t_1\cr t_3\cr} } \cr

2506: & 

2507: e^{- \sum_{j=3}^n {n\choose j} x_1^{n-j} t_1^j + \sum_{j=3}^n {n\choose j} x_3^{n-j} t_3^j }\cr}

2508: }

2509: %

2510: in powers of $t_1, t_3$ and doing the gaussian integrals. 

2511: Here $(x_1,x_3)$ is the critical point we are expanding around. 

2512: Note that the ``interaction vertices'' in the Feynman diagram expansion depend on $\nu$. 

2513: The result is that \spcontr\ is multiplied by   a power series in $w^{-n/(n-2)}$.  

2514: 

2515: 

2516: 

2517: 

2518: These asymptotic expansions   give $n-2$ asymptotic solutions 

2519: to the GKZ equation, valid at large $w$. The last solution is the log solution from the 

2520: asymptotic expansion at $(0,0)$. 

2521: Thus, we have a basis of formal solutions: 

2522: %

2523: \eqn\asympbasis{

2524: \eqalign{

2525: b_0 & =  \log w - \sum_{k=1}^\infty {\Gamma(kn)\over (k!)^2} w^{-kn}  \cr 

2526: b_\nu & = {1\over \sqrt{n^3(n-2)}} \bigl({-w\over n} \bigr)^{-n\over n-2} e^{-2\pi i {\nu\over n-2} } 

2527: \exp\biggl[ (n-2) (-{w\over n})^{n\over n-2} e^{2\pi i {\nu\over n-2}} 

2528: \biggr] 

2529: p^{\nu}  \qquad \nu = 1, \dots, n-2 \cr}

2530: }

2531: %

2532: where $p^\nu$ is an asymptotic  series in $w^{-n/(n-2)}$. 

2533:  

2534: 

2535: 

2536: Note that, roughly speaking, half of the series $b_\nu$ are ``exponentially growing,'' and 

2537: half are ``exponentially decreasing.'' Thus, there is a filtration on $\CV$ given by the maximal 

2538: asymptotic growth as $w\to \infty$ along a fixed ray. We order the values of $\nu_i$ so that 

2539: %

2540: \eqn\filrta{

2541: \re (-{w\over n})^{n\over n-2} e^{2\pi i {\nu_1\over n-2}} < \re (-{w\over n})^{n\over n-2} e^{2\pi i {\nu_2\over n-2}}< \cdots 

2542: %< \re (-{w\over n})^{n\over n-2} e^{2\pi i {\nu_{n-2}\over n-2}}

2543: }

2544: %

2545: to produce the filtration 

2546: %

2547: \eqn\filtrate{

2548: \CF_1 \subset \CF_2 \subset \cdots \subset \CF_{n-1} = \CV 

2549: }

2550: %

2551: Note that there are $(n-1)$ steps in the filtration because of the logarithmic solution. 

2552: 

2553: 

2554: Our next task is to find out how the true solutions $\sum_s c_s \hat I_s(w)$ with $\sum_s c_s=0$ 

2555: fit into the filtration 

2556: \filtrate. 

2557: 

2558: 

2559: 

2560: 

2561: \subsec{Asymptotics of  $\hat I_s$: exponential growth } 

2562: 

2563:  

2564: The most natural way to investigate the asymptotics of $\hat I_s$ is to apply the 

2565: saddle-point technique to the integral representation

2566: \eqn\inthat{{d\over dw } \hat I_{\alpha\beta}=-\int_{\gamma_{\alpha\beta}}e^{-W}dx_1dx_3 .}

2567: Applying the saddle point technique is not straightforward. Care is needed in 

2568: establishing   which of the saddle points \lgiiip\ contribute to a 

2569: given integral $\hat I_s$. However, if a critical point lies in the convergent sector 

2570: containing the contour defining $\hat I_s$ then that critical point does contribute since 

2571: no large contour deformations are required. Using that rule alone we can learn some useful 

2572: facts about when $\hat I_s(w)$ has exponential growth. 

2573: 

2574: 

2575: Let us label the convergent sectors by $\CS_s$ where $s$ is defined, modulo $n$, by \voncen. 

2576: 

2577: 

2578: Depending on the phase of $w$ the critical points $(x_1,x_3)_\nu$ in 

2579: \lgiiip\ might or might not be in a convergent sector. Note that since $x_1^\nu= x_3^\nu e^{2\pi i {\nu\over n}}$, 

2580: it follows that 

2581: $x_1,x_3$ either both lie in a convergent or in a nonconvergent sector. 

2582: The condition to lie in a convergent sector is:  

2583:  %

2584: \eqn\secdep{

2585: \re\biggl(

2586: (-{w\over n})^{n\over n-2} e^{2\pi i \bigl( {\nu \over n-2} \bigr) } \biggr) > 0 

2587: }

2588: %

2589: This is precisely the criterion 

2590: that $W_\nu <0$, and hence the same as the criterion that the 

2591: contribution to an integral, if it exists, is always a {\it growing} 

2592: exponential. 

2593: 

2594: 

2595: Let us analyze more fully when the exponentially growing 

2596: critical points can contribute to an integral $I_{\a\b}$. 

2597: Let  

2598: %

2599: \eqn\ampphs{

2600: (-{w\over n})^{1\over n-2} e^{2\pi i \bigl( {\nu \over n(n-2)} \bigr) } = e^{i \psi }A 

2601: }

2602: %

2603: where $A>0$, and $-\pi < \psi < \pi$. Then from \secdep\ we know there is an $s_*$ with 

2604: %

2605: \eqn\vonceni{

2606: - {\pi \over 2n} <  \psi + {2\pi \over n} s_* < {\pi \over 2n} 

2607: }

2608: %

2609: Then the critical point \lgiiip\ is 

2610: in the sector: 

2611: %

2612: \eqn\insectr{

2613: (A e^{i \psi + 2\pi i\nu /n} , A e^{i \psi } ) \in \CS_{s_*+\nu} \times \CS_{ s_* }

2614: }

2615: %

2616: Thus, for $w$ such that \vonceni\ holds, $\hat I_{\a\b}$ has an exponential growth from a critical point iff 

2617: $s_{\a\b} = 2s_* +\nu  \mod ~n $. Moreover, the Landau-Ginzburg symmetry $\nu \to \nu+(n-2)$ relates the 

2618: growth of different functions $\hat I_s \to \hat I_{s+1}$. To exploit this consider  

2619: $\hat I_0(w)$. For $\vert \arg(-w)^{1/(n-2)} \vert < {\pi\over 2}$, that is, for 

2620: %

2621: \eqn\expgr{

2622: -{\pi \over 2} + {\pi\over n} < \arg(-w) < {\pi \over 2} - {\pi\over n}

2623: }

2624: %

2625: the critical point $\nu=n(n-2) \cong 0$ contributes to the integral. Therefore, $\hat I_0$ has exponential 

2626: growth in this sector. Now using $\hat I_s(w) = \hat I_0(\omega^{-s} w)$ we can make similar statements 

2627: about the other sectors. 

2628: 

2629: 

2630: As we shall see in the next section, in the sector complementary to \expgr\ $\hat I_0$ has in fact at most 

2631: logarithmic growth. In overlapping sectors of the type \expgr\ we can form linear combinations of the 

2632: $\hat I_s$ to produce functions with exponential growth slower than the leading one. 

2633: 

2634: 

2635: {\bf Examples} 

2636: 

2637: 

2638: \item{1.} $n=3$. The functions $\hat I_s$ have exponential growth 

2639: like $\CE=-\sqrt{27}w^{-3}\exp((-w/3)^3)$ for: 

2640: %

2641: \eqn\expnsg{

2642: \eqalign{

2643: \hat I_0: \quad & \quad   -{\pi\over 6} < \arg(-w) < {\pi \over 6} \cr

2644: \hat I_1: \quad & \quad   -{5\pi \over 6} < \arg(-w) < - {3\pi \over 6} \cr

2645: \hat I_2: \quad & \quad   {3\pi \over 6} < \arg(-w) < {5\pi \over 6}\cr}}

2646: %

2647: These are the shaded regions in Fig. 3 below.

2648: 

2649: 

2650: 

2651: \item{2.} $n=4$. We have exponential growth for 

2652: %

2653: \eqn\expnsg{

2654: \eqalign{

2655: \hat I_0: \quad & \quad   -{\pi\over 4} < \arg(-w) < {\pi \over 4} \cr

2656: \hat I_1: \quad & \quad   {\pi\over 4} < \arg(-w) < {3\pi \over 4} \cr

2657: \hat I_2: \quad & \quad   {3\pi\over 4} < \arg(-w) < {5\pi \over 4} \cr

2658: \hat I_3: \quad & \quad   {-3\pi\over 4} < \arg(-w) < {-\pi \over 4} \cr}

2659: }

2660: %

2661: See Fig. 4.

2662: 

2663: 

2664: 

2665: 

2666: \subsec{Asymptotics of $\hat I_s$: Coefficient of the logarithm} 

2667: 

2668: 

2669: In this section we introduce a different integral representation for the functions $\hat I_s$ 

2670: which, while only valid in part of the complex $w$ plane, is very useful for extracting 

2671: asymptotic behavior for $\vert w \vert \to \infty$. We apply a method described in  

2672:  \kaminski. \foot{

2673: This reference examines asymptotics of the integrals relevant to 

2674: the more general set of  $\IC^2/\IZ_{n(p)}$ models and might be useful in further 

2675: extensions of the present paper. } 

2676: 

2677: 

2678: 

2679: In what follows $\arg(z)$ always means the principal branch of the logarithm, so it is defined 

2680: for $z\in \IC - \IR^-$ and $\vert \arg(z)\vert < \pi$. 

2681: We begin with 

2682: %

2683: \eqn\ezee{

2684: e^{-z} = {1\over 2\pi i } \int_{\CC} \Gamma(u) z^{-u} du \qquad \vert \arg(z)\vert < \pi/2

2685: }

2686: %

2687: where $\CC$ is the contour $u= \epsilon + i y , y\in \IR, \epsilon>0$. Note that 

2688: %

2689: \eqn\gamms{

2690: \vert \Gamma(x+ iy ) \vert \sim \sqrt{2\pi} \vert y\vert^{-\half + x} e^{-{\pi \over 2} \vert y \vert} 

2691: }

2692: %

2693: for $y \to \pm \infty$ at any fixed $x$. Thus the integrand converges absolutely for 

2694: $\vert \arg(z)\vert < \pi/2$. 

2695: %We can close the contour in the LHP. 

2696: To prove \ezee\ note that we can close the contour in 

2697: the left half plane.

2698: 

2699: We apply this to $e^{-\a\b w t_1 t_3}$ in the integral representation 

2700: for $\hat I$ which follows from \regin. Thus we must require 

2701: %

2702: \eqn\newrpi{

2703: \vert \arg(\a\b w) \vert < {\pi \over 2} 

2704: }

2705: %

2706: and for such values of $\a,\b, w$ we define the integer $N(\a,\b, w)$ by 

2707: %

2708: \eqn\newrpii{

2709: -{\pi \over 2} <  \arg(\a) + \arg(\b) + \arg(w) + 2\pi N(\a,\b, w) < {\pi \over 2} 

2710: }

2711: %

2712: The integrals are absolutely convergent and we can exchange them and do the $t_1, t_3$ 

2713: integral. Using the generalization of \simplin\

2714: %

2715: \eqn\simplinb{

2716: \int_0^\infty {dt\over t} t^{-s} e^{-\a^n t^n} = \a^{s-1} e^{ {2\pi i s_\a\over n} (s-1)} {1\over n} \Gamma({1-s\over n}) \qquad \re(s)<1 

2717: }

2718: %

2719:  we find

2720: %

2721: \eqn\newrep{

2722: {d\over dw} \hat I_{\alpha\beta}(w;0) = - e^{-2\pi i s_{\a\b}/n} {1\over 2\pi i n^2} 

2723: \int_{\CC} \Gamma(u) \bigl(\Gamma({1-u\over n})\bigr)^2 w^{-u} 

2724: \exp\bigl[ 2\pi i u \bigl({s_{\a\b}\over n} - N(\a,\b, w) \bigr) \bigr] du

2725: }

2726: %

2727: This is valid in the region \newrpi. It is a good exercise to use \gamms\ to check that \newrep\ is 

2728: an absolutely convergent integral in this range. Indeed, this  

2729: condition guarantees absolute convergence of the integral in \newrep\ along 

2730: any contour of the form $x + i y$ for fixed $x$ with $y\in \IR$. 

2731: 

2732: 

2733: Now, using the property that $I_{\a\b}$ only depends on the wedge in which $\a,\b$ live 

2734: we can map out the range of validity for the integral representation \newrep. 

2735: It follows that, for $\hat I_0$, the domain of validity of \newrep\ is  the region 

2736: %

2737: \eqn\logreg{

2738: -{\pi \over 2} - {\pi \over n} < \arg(w) < {\pi \over 2} + {\pi \over n} 

2739: }

2740: %

2741: Note that this is perfectly complementary to the

2742: region \expgr.

2743: Using the LG symmetry we find that the formula \newrep\ for  $\hat I_s$ 

2744: holds in the     range

2745: %

2746: \eqn\rane{

2747: - {\pi \over 2 } + {\pi \over n}(2s-1) < \arg w <  {\pi \over 2 } + {\pi \over n}(2s+1) 

2748: }

2749: %

2750: %provided $n$ is sufficiently larger than $s$. 

2751: If $n$ is sufficiently larger then $s$ then $N=0$ throughout \rane.

2752: 

2753: 

2754: 

2755: The integral \newrep\ is useful because it allows us to obtain asymptotic expansions for $\hat I_s$ 

2756: in the region \rane.  While we   cannot close the $u$-contour integral in the  

2757: left half plane,

2758: we can displace the contour to the right, thanks to \gamms. In the process we pick up poles 

2759: from $u=1+kn$, $k=0,1,2,\dots$.  In this way we arrive at the   asymptotic expansion 

2760: %

2761: \eqn\asympttos{

2762: {d\over dw} \hat I_{\alpha\beta}(w;0)  \sim 2\pi i ({s_{\a\b}\over n} - N(\a,\b,w) ) \sum_{k=0}^\infty {\Gamma(1+kn)\over (k!)^2} 

2763: w^{-1-kn}  + \sum_{k=0}^\infty D_k(n) w^{-1-kn}

2764:   }

2765: %

2766: where the $D_k(n)$ are 

2767: functions of $k,n$ but are independent of $w$ and, crucially, are 

2768: independent  of $\a,\b$ and hence 

2769: cancel out when one forms combinations $\sum c_{\a\b} \hat I_{\a\b}$ 

2770: such that $\sum c_{\a\b}=0$. Integrating this formula we have the asymptotic expansion

2771: %

2772: \eqn\asymptosi{

2773: \eqalign{

2774: \hat I_{\a\b}(w;0) & \sim 2\pi i ({s_{\a\b}\over n} - N(\a,\b,w) )\biggl( \log w - \sum_{k\geq 1} {\Gamma(kn)\over (k!)^2} w^{-kn} \biggr)  + \CU \cr

2775: \CU & := - \half(\log w)^2 - {(n-2)\over n}\gamma \log w + c+ \sum_{k\geq 1} {\Gamma(kn)\over (k!)^2} (-kn w^{-kn }\log w + h_k w^{-kn}) \cr

2776: h_k & := -1 + {2\over n}(1+ {1\over 2} + \cdots + {1\over k} - \gamma) - \Psi(1+kn) \cr}

2777: }

2778: %

2779: where $c$ is a constant. 

2780: The important thing in this formula is that $\CU$ is 

2781: independent of $\alpha,\beta$ (and hence of $s$). 

2782: Note that the $s$-dependent term is nicely consistent with 

2783: the logarithmic dependence on $w$ in the expansion $\CU$,

2784: and the LG symmetry.

2785: The formula is valid in the LG images of \logreg. 

2786: The asymptotics  perfectly complement the region \expgr\ where the leading exponential dominates. 

2787: 

2788: 

2789: One important conclusion we can draw from \asymptosi\ is that  

2790: in those regions where \newrpi\ is simultaneously valid for all terms with 

2791: $c_{\a\b}\not=0$ and such that 

2792: $\sum  c_{\a\b} ({s_{\a\b}\over n} - N(\a,\b,w) ) =0 $

2793: then $\sum c_{\a\b} \hat I_{\a\b}$ will be an exponentially decaying solution. 

2794: 

2795: 

2796: \def\im{{\rm Im}}

2797: 

2798: The discussion in this section falls short of giving a complete description of the filtration \filtrate\ 

2799: in all angular sectors for general $n$ because we have not explained how to form linear combinations with 

2800: prescribed sub-exponential growth.  One can apply the saddle point technique to the integral \newrep\ 

2801: for linear combinations, such as $\hat I_s + \hat I_{-s} - 2\hat I_0$ for which the pole terms cancel. 

2802: One finds that $\hat I_s$ contributes an exponential behavior like 

2803: %

2804: \eqn\expgrwth{

2805:  \cases{ \exp\biggl[ - (n-2) \bigl({w\over n}\bigr)^{n\over n-2}   e^{-{2\pi i \over n-2}(s+1) }\biggr] & $\im w>0 $ \cr

2806: \exp\biggl[ - (n-2) \bigl({w\over n}\bigr)^{n\over n-2}   e^{-{2\pi i \over n-2}(s-1) }\biggr] & $\im w<0 $ \cr}

2807: }

2808: %

2809: in the intersection of the regions \rane. (We assume $n \gg \vert s\vert $ at this point.) 

2810: %Unfortunately, 

2811: %this is not sufficient to determine the filtration. 

2812: The term with the least rapid decay will then dominate.

2813: Careful application of this rule might  suffice to determine the

2814: full filtration \filtrate\ but we have not carried this out.

2815: There is also a   Mellin-Barnes representation of 

2816: the functions $\hat f_m$ but the saddle point technique applied to this representation  proves 

2817: inconclusive.  Nevertheless, the results we have presented here do suffice to give a rather complete picture of the 

2818: filtration for $n=3$ and $n=4$, as we describe in the next subsection. 

2819: 

2820: 

2821: \subsec{Examples: $n=3$ and $n=4$}

2822: 

2823: 

2824: Let us show how the above general results can give a   complete picture of the filtration 

2825: for the cases $n=3$ and $n=4$. Let 

2826: %

2827: \eqn\calee{

2828: \CE = -\sqrt{27} w^{-3} \exp\bigl(-{w\over 3}\bigr)^3

2829: }

2830: %

2831: Then we have

2832: %

2833: \eqn\cyou{

2834: \hat I_0 \sim \cases{ \CU & $-{5\pi \over 6} < \arg w < {5\pi \over 6} $\cr

2835:  \CE & $-{\pi \over 6} < \arg -w < {\pi \over 6} $\cr}

2836: }

2837: %

2838: %

2839: \eqn\cyou{

2840: \hat I_1 \sim \cases{ \CU + {2\pi i \over 3} \log w + \cdots & $-{\pi \over 6} < \arg w < \pi  $\cr

2841: \CU - {4\pi i \over 3} \log w + \cdots & $-\pi < \arg w < -{\pi \over 2}  $\cr

2842:  \CE & $-{\pi \over 2} < \arg -w < -{\pi \over 6} $\cr}

2843: }

2844: %

2845: %

2846: \eqn\cyou{

2847: \hat I_{-1} \sim \cases{ \CU + {4\pi i \over 3} \log w + \cdots & ${\pi \over 2} < \arg w < \pi  $\cr

2848: \CU - {2\pi i \over 3} \log w + \cdots & $-\pi < \arg w < {\pi \over 6}  $\cr

2849:  \CE & ${\pi \over 6} < \arg -w < {\pi \over 2} $\cr}

2850: }

2851: %

2852: 

2853: Now let us describe the corresponding filtrations. 

2854: There are only two steps in \filtrate. In the ``convergent sectors'' 

2855: $\re w^3 <0 $ the exponential solution, which is asymptotic to $\CE$, 

2856:  is growing and we have $\CF_0 \subset \CF_+$ where $\CF_0$ is 

2857: the 1-dimensional space spanned by the log solution. In the sectors with $\re w^3>0$, $\CE$ is decaying

2858: and we have the filtration $\CF_- \subset \CF_0$, where $\CF_-$ is the one-dimensional 

2859: space spanned by the exponentially decaying solution $\CE$. 

2860: 

2861: The first step in the filtration is given by 

2862: %

2863: \eqn\firsta{

2864: \CF_0 = {\rm Span}\cases{\hat I_0 - \hat I_{-1} \sim {2\pi i \over 3} \log w + \cdots & $-{3\pi \over 6} < \arg w < - {\pi \over 6} $\cr

2865: \hat I_1 - \hat I_{-1} \sim - {2\pi i \over 3} \log w + \cdots & $-{\pi \over 6} < \arg (-w) <  {\pi \over 6} $\cr

2866: \hat I_0 - \hat I_{1} \sim -{2\pi i \over 3} \log w + \cdots & ${\pi \over 6} < \arg w <  {3\pi \over 6} $\cr}

2867: }

2868: %

2869: for the sectors in which $\CE$ is exponentially growing. Note the three lines of \firsta\ are related by LG symmetry. 

2870: Similarly the first step in the filtration is given by  

2871: %

2872: \eqn\firstap{

2873: \CF_- = {\rm Span}\cases{2\hat I_0 - \hat I_{-1}- \hat I_{+1}  \sim cnst. \CE & $-{\pi \over 6} < \arg w < {\pi \over 6} $\cr

2874: 2\hat I_1 - \hat I_{-1}- \hat I_{0}  \sim cnst. \CE & ${3\pi \over 6} < \arg w < {5\pi \over 6} $\cr

2875: 2\hat I_{-1} - \hat I_{1}- \hat I_{0}  \sim cnst. \CE & $-{5\pi \over 6} < \arg w < -{3\pi \over 6} $\cr}

2876: }

2877: %

2878: for the sectors in which $\CE$ is exponentially decreasing.  

2879: 

2880: 

2881: Similarly, for $n=4$ we find: 

2882: %

2883: %

2884: \eqn\cyouf{

2885: \hat I_0 \sim \cases{ \CU & $-{3\pi \over 4} < \arg w < {3\pi \over 4} $\cr

2886:  \CE_0 & $-{\pi \over 4} < \arg(-w) < {\pi \over 4} $\cr}

2887: }

2888: %

2889: Here $\CE_0 \sim \sqrt{2} w^{-2}\exp[w^2/8] $ is the 

2890: growing exponential corresponding to the critical point with $\nu=0$ (or its LG images).

2891: %

2892: \eqn\cyouff{

2893: \hat I_1 \sim \cases{ \CU + {2\pi i \over 4} \log w + \cdots & $-{\pi \over 4} < \arg w < \pi  $\cr

2894: \CU - {3\pi i \over 4} \log w + \cdots & $-\pi < \arg w < -{3\pi \over 4}  $\cr

2895:  \CE_1 & $-{3\pi \over 4} < \arg w < -{2\pi \over 4} $\cr}

2896: }

2897: %

2898: Here $\CE_1 \sim \sqrt{2} w^{-2}\exp[-w^2/8] $ is the 

2899: growing exponential corresponding to the critical point with $\nu=1$ (or its LG images). 

2900: Similarly

2901: %

2902: \eqn\cyoufff{

2903: \hat I_{-1} \sim \cases{ \CU - {2\pi i \over 4} \log w + \cdots & $-\pi  < \arg w < {\pi \over 4} $\cr

2904: \CU +{3\pi i \over 4} \log w + \cdots & ${3\pi \over 4}  < \arg w < \pi   $\cr

2905:  \CE_1 & $ {\pi \over 4} < \arg  w < {3\pi \over 4} $\cr}

2906: }

2907: %

2908: %

2909: \eqn\cyoufv{

2910: \hat I_{2} \sim \cases{ \CU + i \pi  \log w + \cdots & ${\pi\over 4}   < \arg w <\pi  $\cr

2911: \CU -i \pi  \log w + \cdots & $-\pi   < \arg w < -{\pi \over 4}    $\cr

2912:  \CE_0 & $ -{\pi \over 4} < \arg  w < {\pi \over 4} $\cr}

2913: }

2914: %

2915: 

2916: Using these formulae we can specify bases for the 3-step filtration $\CF_- \subset \CF_0 \subset \CF_+$, where 

2917: $\CF_-$ has at most exponential decay, and $\CF_0$ has at most logarithmic growth. In the 

2918: sectors $\vert \arg w \vert < {\pi \over 4} $ we find that $\CF_- $ is generated by 

2919: $2 \hat I_0 - \hat I_1 - \hat I_{-1} \sim cnst. \CE_1$, while $\CF_0$ is spanned by 

2920: $\hat I_0 - \hat I_1$ and $\hat I_0 - \hat I_{-1}$. The filtrations in the other sectors are given by 

2921: applying the LG symmetry. 

2922: 

2923: \subsec{Stokes matrices}

2924: 

2925: 

2926: \def\CS{{\cal S}}

2927: \lref\balser{W. Balser, W.B. Jurkat, D.A. Lutz, {\it Birkhoff Invariants 

2928: and

2929: Stokes' Multipliers for Meromorphic Linear Differential Equations},

2930: Journal of Math. Analysis and Applications, {\bf 71}(1979)48-94}

2931: 

2932: \lref\dingle{R.B. Dingle, {\it Asymptotic Expansions: their derivation and

2933: interpretation}, Academic Press, 1973}

2934: 

2935: The differential equation \gkzsimp\ can be written as a first order 

2936: $n\times n$ matrix equation

2937: of the form:

2938: %

2939: \eqn\matrixfrm{

2940: \bigl({d\over dw} - A(w) \bigr) \Psi =0

2941: }

2942: %

2943: where

2944: %

2945: \eqn\apto{

2946: A(w) = \sum_{i=1}^{n-1} e_{i,i+1} + {(-1)^n\over n^2} w e_{n,2} + 

2947: {(-1)^n\over n^2} w^2 e_{n,3}

2948: }

2949: %

2950: and $e_{i,j}$ are matrix units. 

2951: We regard this as an equation for an $n\times n$ invertible matrix $\Psi$.

2952: The physical interpretation of $\Psi$ is that it is the matrix of 1-point 

2953: correlators of the elements of the chiral ring $\bigl(\IC[x_1,x_3]/(x_i \p_i W)\bigr)^{\IZ_n}$:

2954: %

2955: \eqn\chiralrng{

2956: \Psi_{ij} = \int_{\gamma_i} (x_1 x_3)^{j} e^{-W} {dx_1\over x_1} {dx_3 \over x_3} 

2957: }

2958: %

2959: At infinity there is a formal asymptotic solution. True $n \times n$ matrix

2960: solutions asymptotic to the fixed  formal solution can only be defined 

2961: in angular sectors.

2962: For sectors of sufficiently wide angle the true solution is unique. On 

2963: overlapping

2964: sectors two such solutions will be related by right multiplication by a 

2965: constant matrix

2966: known as a Stokes matrix.  For further details see, for examples, 

2967: \dingle\balser.

2968: 

2969: 

2970: The results of the previous sections allow one to determine the Stokes 

2971: matrices for

2972: $n=3,4$. 

2973: We work directly with a basis for $\cal V$, rather then $\Psi$.

2974: For $n=3$ we define a vector of formal solutions

2975: %

2976: \eqn\formalsl{

2977: \pmatrix{ \kappa_0 b_0 \cr \kappa_1 b_1 \cr}

2978: }

2979: %

2980: where $\kappa_0,\kappa_1$ are appropriate constants and $b_0, b_1$ are 

2981: defined in \asympbasis.

2982: We now introduce

2983: six sectors: $\CS_j:= \{ w\vert {\pi j\over 3} - {\pi \over 2}

2984: < \arg w < {\pi j\over 3} + {\pi \over 6} \},\, 1\leq j\leq6$. In each of these sectors 

2985: there is a {\it unique}

2986: basis of solutions asymptotic to \formalsl. These are:

2987: %

2988: \eqn\trusol{

2989: \eqalign{

2990: \CS_1:\qquad\qquad & \psi_1 = \pmatrix{ \hat I_0 - \hat I_1 \cr 2 \hat 

2991: I_0 - \hat I_1 - \hat I_{-1} \cr} \cr

2992: \CS_2: \qquad\qquad & \psi_2 = \pmatrix{ \hat I_0 - \hat I_1 \cr 2 \hat 

2993: I_1 - \hat I_0 - \hat I_{-1} \cr} \cr

2994: \CS_3: \qquad\qquad & \psi_3  = \pmatrix{ \hat I_1 - \hat I_{-1} \cr 2 

2995: \hat I_1 - \hat I_0 - \hat I_{-1} \cr} \cr

2996:  \cdots\qquad\qquad\,\,&\qquad\qquad\qquad\cdots \cr}

2997: }

2998: %

2999: The other sectors are obtained by LG symmetry. Then we have Stokes 

3000: matrices:

3001: %

3002: \eqn\sotks{

3003: \eqalign{

3004: \CS_1 \cap \CS_2:\qquad\qquad  & \psi_1 = \pmatrix{1&0 \cr 3 & 1\cr} 

3005: \psi_2 \cr

3006: \CS_2 \cap \CS_3: \qquad\qquad & \psi_2 = \pmatrix{1&-1\cr 0& 1\cr} 

3007: \psi_3 \cr

3008: \cdots\qquad\qquad\quad\,\,&\qquad\qquad\,\cdots \cr}

3009: }

3010: %

3011: and the remaining sectors are obtained by applying LG symmetry.

3012: Similarly, one can compute the Stokes matrices for $n=4$.

3013: 

3014: 

3015: 

3016: 

3017: 

3018: \subsec{Summary} 

3019: 

3020: 

3021: The generalized periods are solutions of the GKZ equation \gkzsimp. A basis of solutions can be written in 

3022: the integral representation \lgi\ \inthat\ for appropriate linear combinations of  regulated ``straightline contours'' \gami. 

3023:  These linear combinations should 

3024: be thought of as elements of the homology group $H_n(\IC^2, B)$ where $B$ is the region at 

3025: infinity where $ \re W \to +\infty$. By combining the ``Landau-Ginzburg symmetry'' 

3026: \lgsymm\ with saddle point techniques ( equations \lgiiip\lgivp\spcontr) and the integral 

3027: representation \newrep\ we are able to find the angular-sector-dependent asymptotic behavior of the 

3028: generalized periods, which exhibit Stokes' phenomenon.

3029: We gave a detailed analysis   for the cases of $n=3,4$ in \calee--\cyoufv\ and showed how 

3030: to compute the Stokes' matrices in \sotks.  

3031:  

3032: 

3033: 

3034: 

3035: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

3036: \newsec{Propeller branes vs. fractional branes: analysis for small $n$}

3037: In this section we analyze $\IC^2/\IZ_{n(p=1)}$ orbifolds for 

3038: $n=3,4$.

3039: Let us first summarize our findings.

3040: We look at the overlap of the Ramond ground state with the branes

3041: in the theory deformed by the addition of a tachyon vertex operator.

3042: This quantity is the direct analog of the central charge in

3043: the spacetime supersymmetric theories and we call it the ``generalized

3044: central charge'' or ``generalized period.''

3045: Earlier in the paper we explained that this quantity satisfies

3046: the GKZ equation.

3047: A convenient basis of non-constant solutions is provided by 

3048: combinations of $\hat I_s$ which satisfy  \newslkc.

3049: 

3050: 

3051: 

3052: As $x_1,x_3\ra \infty$, the A-brane surfaces defined in Section 3 

3053: asymptote to quarterplanes \gami\ used

3054: in the definition of $I_{\alpha,\beta}$ in \regin.

3055: %As we see below, the linear combination of $I_{\alpha,\beta}$ defined

3056: %by these asymptotics satisfies  \linec.

3057: More precisely, we find that 

3058: \eqn\cgccp{  \langle c_\nu|1\rangle = {\cal C}_1 +

3059:                   {\cal C}_2 (\hat I_{s+1}+\hat I_{s-1}-2\hat I_s)       }

3060: with the relation between $\nu$ and $s$ to be determined later.

3061: In eq. \cgccp\ ${\cal C}_1$ and ${\cal C}_2$ 

3062: are $w$-independent constants.

3063: We do not know the behavior of the contour $\Gamma_\nu$ in the 

3064: compact region. Because of the pole $1/x_1 x_3$ in the 

3065: measure we must allow for a   constant term  $C_1$ 

3066: which cannot be determined by our techniques\foot{

3067: Recall that the intersection matrix is also insensitive to

3068: the addition/subtraction of D0 branes, which corresponds to

3069: adding/subtracting an integer from \cgccp.}.

3070: In the following, we will adopt the notation 

3071: \eqn\cgcc{  \langle c_\nu|1\rangle \sim \hat I_{s+1}+\hat I_{s-1}-2\hat I_s     }

3072: which is meant to be equivalent to \cgccp.

3073: %This allows us to substitute $\hat I$'s instead of $I$'s.

3074: %Moreover, the LHS of \cgcc\ defined by \gcc\  can possibly differ from 

3075: %$ 2 I_s-I_{s+1}-I_{s-1} $ by a constant term, since the contours 

3076: %of integration differ in the compact region.

3077: The evidence for \cgcc\ comes from numerically solving the differential

3078: equations that define the profile of the A-brane.

3079: [We also observe that \cgcc\ is consistent with the saddle-point evaluation

3080: of the integral \gcc.]

3081: 

3082: The theory at the orbifold point contains a set of fractional branes 

3083: whose boundary states are described in appendix A.

3084: In the following we will provide evidence that the coulomb branch branes

3085: are equivalent to the fractional branes, up to a permutation.

3086: In fact, we propose

3087: \eqn\fbgcc{  \langle e_a|1\rangle ={\cal C}_2( \hat I_{a+1}+\hat I_{a-1}-2 \hat I_a)+{1\over n}  }

3088: (From the exact boundary state description we happen to know the

3089: value of the constant term for this overlap; it is equal to $1/n$.)

3090: The evidence for \fbgcc\ comes from the map between the higgs branch brane

3091: and the fractional branes, which is determined by the intersection matrix.

3092: In fact, for $n=3$ the map also implies that the coulomb branch brane is

3093: equal to the fractional brane up to permutations.

3094: Here is a consistency check for \fbgcc.

3095: Consider turning on the tachyon vertex operator $w$.

3096: The worldsheet action of the orbifold theory is modified by

3097: \eqn\modac{  \delta S\sim w\int dz d\bar z \int d^2\theta X_{1\over n}    .  } 

3098: The disk one-point function of the unit operator to   first nontrivial 

3099: order in $w$ is

3100: \eqn\onept{  \langle e_a|1\rangle\sim  w\, \exp\left(-{2\pi i a \over n}\right)+\OO(w^2)  }

3101: where we used the boundary state expression for the fractional brane

3102: to determine the one-point function, and we have dropped $n$-dependent

3103: normalization constants.

3104: Substituting the small $w$ expansion \itoeff\ into \fbgcc\ gives precisely

3105: \onept.

3106: 

3107: 

3108: \subsec{$n=3$}

3109: We first consider the case of $n=3$, $p=1$.

3110: In fact, the orbifold group is $Z_{N=2n}=Z_6$ due to the action

3111: on fermions; only the type 0 (not the type II) theory can be defined.

3112: There are three choices for the intersection matrix for fractional branes, given by \inonsym\

3113: and by matrices obtained from \inonsym\ by permutation of columns.

3114: %There are $n=3$ choices for the reduced (D0-brane omitted) intersection matrix $I$.

3115: \item{(1)} The complete, unreduced intersection matrix, obtained by a 

3116: permutation of columns from \inonsym\ is 

3117: \eqn\ithreeanr{ \I=  \pmatrix{~-2& ~1& ~1\cr ~1& ~-2& ~1\cr ~1& ~1& -2}   }

3118: The reduced matrix, with the D0 brane factored

3119: out, corresponds to the continued fraction

3120: determined by $n/(n-p)\ra[2,2]$.

3121: \eqn\ithreea{ \It=  \pmatrix{~-2& ~1\cr ~1& ~-2 }   }

3122: %

3123: The corresponding \HJ\ space is an ALE space. It is obtained from the orbifold

3124: by turning on the generators of the (c,c) ring.

3125: This is essentially equivalent to the spacetime supersymmetric case.

3126: There is no coulomb branch and the map between the fractional

3127: branes and the higgs branch branes is known via the McKay correspondence.

3128: \item{(2)} The unreduced matrix is obtained by a permutation of 

3129: columns in \ithreeanr.

3130: The reduced intersection form is determined by the continued fraction

3131: determined by $n/p\ra [3]$.

3132: \eqn\ithreeb{ \It=  \pmatrix{~1& ~-2\cr ~1& ~1 }   }

3133: According to \abc, the higgs and the coulomb branch branes are given by

3134: \eqn\hbbthree{  h=e_s-e_{s+1},\qquad c=e_{s+1}    }

3135: where the index $s$ is not determined at this stage.

3136: In fact, as we will see below, all permutations will be realized,

3137: depending on the argument of $w$.

3138: \item{(3)} Similar to the previous case, 

3139: the continued fraction is $n/p\ra [3]$ and

3140: \eqn\ithreec{ \It=  \pmatrix{~1& ~1\cr ~-2& ~1 }   }

3141: and therefore

3142: \eqn\hbbthreeb{  h=e_s+2e_{s+1},\qquad c=e_{s+1}    }

3143: 

3144: 

3145: 

3146: \noindent

3147: To proceed further, it will be convenient to have a map of the $w$-plane

3148: divided into angular sectors.

3149: This is shown in Fig. 3.

3150: %

3151: %

3152: \midinsert\bigskip{\vbox{{\epsfxsize=3in

3153:         \nobreak

3154:     \centerline{\epsfbox{n3p.eps}}

3155:         \nobreak\bigskip

3156:     {\raggedright\it \vbox{

3157: {\bf Fig 3.}

3158: {\it The angular structure in the $w$ plane. $w$ in the shaded sectors

3159: gives rise to a critical point with a negative value of $Re\, W$.

3160: Big blue arrows mark sectors which differ by permutations of 

3161: fractional branes. Dashed red and green arrows mark sectors which differ

3162: by the type of higgs branch brane}}}}}}

3163: \bigskip\endinsert

3164: %

3165: %

3166: According to \lgiiip, modulo the LG symmetry, there is a single critical point of the

3167: superpotential $W$.

3168: Its coordinates, and the value of $W$ are

3169: \eqn\cpthree{  x_1=x_3=-{w\over 3},\qquad W_*=\left({w\over 3}\right)^3  }

3170: Hence there are three angular sectors in the $w$-plane 

3171: where $W_*$ is negative and the integral \gcc\ can pick up 

3172: an exponentially growing contribution from the critical point.

3173: This happens when $(x_1,x_3)$ given by \cpthree\ fall into the

3174: regions (see \expnsg\ above)

3175: \eqn\dashthree{

3176: \eqalign{ {5\pi\over 6}\leq \arg(w)&\leq {7\pi\over 6}, \qquad  s=0   \cr

3177:           {3\pi\over 2}\leq \arg(w)&\leq {11\pi\over 6}, \qquad s=1   \cr

3178:           {\pi\over 6}\leq \arg(w)&\leq {\pi\over 2}, \qquad s=-1     \cr

3179:           }

3180: }

3181: These are the shaded sectors in Fig. 3.

3182: The value of $s$ in \dashthree\ specifies which $\hat I_s$ receives

3183: an exponentially growing contribution, in accord with the rule \insectr.

3184: In these sectors, the nonconstant solutions of the GKZ equation are spanned by the 

3185: two functions with the following leading asymptotics:

3186: \eqn\gkzthree{ 

3187: \eqalign{  {5\pi\over 6}\leq \arg(w)&\leq {7\pi\over 6}:\quad \hat I_{-1}-\hat I_1\sim {2\pi i\over3}\log w,\quad

3188:                                                         \hat I_0-I_{-1}\sim \exp(-W_*) \cr

3189:        {3\pi\over 2}\leq \arg(w)&\leq {11\pi\over 6}:\quad \hat I_0-\hat I_{-1}\sim {2\pi i\over3}\log w,\quad

3190:                                                          \hat I_1-I_{0}\sim \exp(-W_*) \cr

3191:        {\pi\over 6}\leq \arg(w)&\leq {\pi\over 2}:\quad \hat I_1-\hat I_{0}\sim {2\pi i\over3}\log w,\quad

3192:                                                          \hat I_{-1}-I_{1}\sim \exp(-W_*) \cr }

3193: }

3194: where $W_*$ is given by \cpthree\ and is a growing exponential.

3195: The precise form of the exponential growth is given in \calee, but

3196: we use $e^{-W_*}$ as a shorthand for $\CE$.

3197: Note that the three sectors are different by a simple permutation of

3198: indices of $\hat I$'s.

3199: We will see below that this is a general phenomenon which is rooted in

3200: the permutation symmetry of the fractional branes.

3201: 

3202: Consider now three sectors where $Re W_*$ is positive, and the contribution

3203: from the nonzero critical point is a decaying exponential.

3204: These are the unshaded sectors in Fig. 3.

3205: The leading contribution now comes from the critical point at $(x_1,x_3)=(0,0)$

3206: and is logarithmic.

3207: \eqn\gkzthreeb{ 

3208: \eqalign{  -{\pi\over 6}\leq \arg(w)&\leq {\pi\over 6}:\quad \hat I_{1}-\hat I_0\sim {2\pi i\over3}\log w,

3209:                    \quad  \langle c|1\rangle\sim \hat I_1+\hat I_{-1}-2\hat I_0\sim \exp(-W_*) \cr

3210:        {\pi\over 2}\leq \arg(w)&\leq {5\pi\over 6}:\quad \hat I_{-1}-\hat I_1\sim {2\pi i\over3}\log w,

3211:                   \quad  \langle c|1\rangle\sim \hat I_{-1}+\hat I_{0}-2\hat I_1\sim \exp(-W_*) \cr

3212:        {7\pi\over 6}\leq \arg(w)&\leq {3\pi\over 2}:\quad \hat I_{0}-\hat I_{-1}\sim {2\pi i\over3}\log w,

3213:                   \quad  \langle c|1\rangle\sim  \hat I_{0}+\hat I_{1}-2\hat I_{-1}\sim \exp(-W_*) \cr }

3214: } 

3215: In eq. \gkzthreeb\ we identified the exponentially decaying solution 

3216: with the generalized central charge of the coulomb branch brane.

3217: Indeed, we expect the latter to receive an exponentially decaying contribution

3218: from the critical point of $W$.

3219: Now both \hbbthree\ and \hbbthreeb\ state that the coulomb branch brane

3220: is the same as the fractional brane, up to a permutation.

3221: But we can fix the freedom with the help of \onept.

3222: This leads to \fbgcc.

3223: 

3224: The picture that we infer from \gkzthreeb\ is therefore the following.

3225: In the (undashed) sector  $-{\pi\over 6}\leq \arg(w)\leq {\pi\over 6}$, $c=e_0$.

3226: Rotating $\arg(w)$ by $2\pi/3$ enforces the permutation of fractional branes.

3227: E.g. in the sector ${\pi\over 2}\leq \arg(w)\leq {5\pi\over 6}$,  $c=e_1$ etc.

3228: What we do not yet know is how the transition between $c=e_0$ and $c=e_1$

3229: (and more generally, between  $c=e_s$ and $c=e_{s+1}$) happens.

3230: A natural scenario would be the following.

3231: The $w$ plane is divided into 3 angular regions, marked by the blue arrows in Fig. 3.

3232: In the region $-\pi/3\leq\arg(w)\leq\pi/3$, $c=e_0$, and $w\ra \exp(2\pi i /3)w$

3233: enforces the permutation of the fractional branes $e_s\ra e_{s+1}$.

3234: Unfortunately we cannot see this directly from the asymptotics of $\hat I$'s.

3235: The reason is that both $e_s$ and $e_{s+1}$ have the same exponentially

3236: growing asymptotics in the region where the transition happens.

3237: However there is a way to verify the picture proposed above.

3238: As explained before, the shape of the coulomb branch brane can be inferred from

3239: the solution of the soliton equations.

3240: The corresponding solutions emanate from the critical point of $W$ and

3241: run to infinity along the quarterplanes $\gamma_{\alpha\beta}$ used

3242: to define the $\hat I_{\alpha\beta}$'s

3243: (there are $n=3$ inequivalent choices of them).

3244: Solving the equations near the critical point and far away, in the asymptotic region,

3245: is easy.

3246: The more difficult question is matching the solutions in the two regions, i.e. 

3247: understanding which direction in the $w$ plane is chosen by a certain soliton trajectory,

3248: and how this choice depends on $\arg (w)$.

3249: In appendix B we analyze this question numerically.

3250: Our analysis confirms the picture described above.

3251: That is, there are three angular sectors in the $w$ plane, and

3252: the asymptotics of the A-brane surface emanating from the nonzero critical point

3253: (i.e. coulomb branch brane) jump as $w$ crosses the lines of $\arg(w)=\pi/3,\pi,-\pi/3$.

3254: Moreover, the asymptotics are consistent with \gkzthreeb.

3255: 

3256: So far we have determined the behavior of the generalized central charge for the coulomb

3257: branch brane as a function of $w$.

3258: The theory contains a single higgs branch brane, which wraps 

3259: the exceptional $\IP^1$  of $\OO(-3)\ra \IP^1$.

3260: Unfortunately we cannot analyze the soliton equations for this brane,

3261: since the LG model is unreliable near $(x_1,x_3)=(0,0)$ which would

3262: serve as a critical point.

3263: Nevertheless, we can use \hbbthree\ and \hbbthreeb\ together with \fbgcc\

3264: to compute the generalized central charge.

3265: Let us specialize the discussion to the angular sector $-\pi/6\leq\arg(w)\leq\pi/6$;

3266: the other two sectors differ by permutation.

3267: In this sector $c=e_0$, therefore the two possibilities for the higgs branch brane

3268: consistent with the intersection matrix  are

3269: $h^{(1)}=e_2-e_0$ and $h^{(2)}=e_2+2 e_0$.

3270: (Again, one must keep in mind that the expression for the the higgs and

3271: coulomb branch branes are obtained modulo the addition/subtraction

3272: of D0 branes)

3273: Using \fbgcc\ we obtain 

3274: \eqn\hbgccthree{  \langle h^{(1)}|1\rangle\sim3(\hat I_0-\hat I_{-1}),\quad                  

3275:                   \langle h^{(2)}|1\rangle\sim3(\hat I_1-\hat I_{0})       }

3276: We can identify the logarithmic solution with the higgs branch

3277: brane for the shaded sectors in the $w$ plane.

3278: This is where it is defined uniquely (in the unshaded sectors, 

3279: one can add an exponentially decaying solution without changing 

3280: the leading asymptotics).

3281: Comparing with \gkzthree\ we observe that for $-\pi/3\leq\arg(w)\leq-\pi/6$,

3282: $h^{(1)}=e_2-e_0$, while for $\pi/6\leq\arg(w)\leq\pi/3$, $h^{(2)}=e_2+2e_0$.

3283: The natural line of transition between the two happens at $\arg(w)=0$.

3284: (This would be a direct analog of the transition between $c=e_s$ and $c=e_{s+1}$).

3285: We cannot rigorously prove this point and leave

3286: further justification to the future\foot{ 

3287: We can solve the soliton equations {\it with the flat metric}

3288: similarly to what is described in appendix B.

3289: The solutions emanating from the $(x_1,x_3)=(0,0)$ critical point

3290: suggest that the corresponding A-brane surface indeed asymptotes to 

3291: the quarter planes defining $3(\hat I_0-\hat I_{-1})$ and $3(\hat I_1-\hat I_0)$

3292: and the transition between the two happens at $\arg(w)=0$.

3293: This is another strong piece of evidence in favor of the picture proposed.}.

3294: %It will hopefully receive more evidence in

3295: %the future.

3296: 

3297: Let us summarize.

3298: There are $n=3$ ``big'' angular sectors.

3299: In the sector $-\pi/3\leq\arg(w)\leq\pi/3$, the coulomb branch 

3300: brane $c=e_0$.

3301: Multiplication by $\exp(2\pi i/3)$ causes cyclic permutations

3302: of the fractional branes $e_0\ra e_1$ etc.

3303: In addition, each big sector is divided into two smaller ones.

3304: For example, in the angular sector $-\pi/3\leq\arg(w)\leq\pi/3$

3305: there are two possible higgs branch branes:

3306: \eqn\hbfbthreeb{

3307: \eqalign{    h^{(1)}&=e_2-e_0+\ell D0, \quad -{\pi\over3}\leq\arg(w)\leq0   \cr

3308:               h^{(2)}&=e_0-e_1+\ell' D0, \quad 0\leq\arg(w)\leq{\pi\over3}   \cr}

3309: }

3310: where $\ell$ and $\ell'$ are nonnegative integers.

3311: It is natural to assume that $\ell'=\ell$.

3312: The two branes in \hbfbthreeb\ then differ by a permutation.

3313: % and addition of

3314: %$D0$ branes.

3315: Note that both higgs branch branes in \hbfbthreeb\ have the same

3316: asymptotic behavior for the generalized central charge as a function

3317: of $w$:

3318: \eqn\gcchbbt{   \langle  h^{(1,2)}|1\rangle\sim 2\pi i\log w   }

3319: This is reminiscent of the mirror map for the spacetime non-supersymmetric case.

3320: %In fact, the difference between the two branes in \hbfbthree\ 

3321: %goes to a constant as $|w|\ra\infty$.

3322: %Near the orbifold point though, $ h^{(1)}$ and $ h^{(2)}$

3323: %at first sight look very different.

3324: When $\ell=0$ or $\ell'=0$, the branes in \hbfbthreeb\ become massless at the orbifold point, which

3325: generally signals the breakdown of the string perturbation theory.

3326: %Note however that one can write

3327: %\eqn\hbfbthreea{  h^{(2)}=e_2+2e_0=e_0-e_1+D0   }

3328: %where $D0=\sum_a e_a$ is a D0-brane.

3329: %However, as mentioned above, both the intersection matrix and our method

3330: %of computation of the generalized periods are insensitive to the appearance

3331: %of the D0 brane.

3332: It is possible however, that the states which appear in the Hilbert space

3333: of the orbifold theory have positive $\ell,\ell'$.

3334: % is not the first formula in \hbfbthree, but

3335: %rather 

3336: %\eqn\hbfbthree{

3337: %\eqalign{   , \quad -{\pi\over3}\leq\arg(w)\leq0   \cr

3338: %              h^{(2)}&=e_2+2e_0, \quad 0\leq\arg(w)\leq{\pi\over3}   \cr}

3339: %}

3340: %\eqn\hbfbthreeb{  h^{(1)}{}'=e_2-e_0+l D0=e_1+2 e_2+(l-1) D0,\qquad l\in \IZ_+   }

3341: %That is,  the higgs branch branes in the two subsectors differ by a permutation.

3342: 

3343: 

3344: 

3345: 

3346: 

3347: 

3348: 

3349: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

3350: \subsec{n=4}

3351: Our next example is $n=4$, $p=1$.

3352: In this case the orbifold group is  $Z_{N=n}=Z_4$ 

3353: and both type 0 and type II theory can be defined. 

3354: There are $n=4$ choices for the reduced intersection matrix $\It$.

3355: \item{(1)} The intersection form corresponds to the continued fraction

3356: determined by $n/(n-p)\ra[2,2,2]$.

3357: \eqn\ifoura{ \It=  \pmatrix{~-2& ~1& ~0\cr ~1& ~-2& ~1\cr ~0& ~1& ~-2 }   }

3358: %

3359: This describes spacetime-supersymmetric ALE space.

3360: In all other examples the intersection form is determined by the continued fraction

3361: determined by $n/p\ra [4]$.

3362: \item{(2)} 

3363: \eqn\ifourb{ \It=  \pmatrix{~1& ~-2& ~-1\cr ~0& ~1& ~-2\cr ~1& ~0& ~1 }   }

3364: The higgs branch brane is given by

3365: \eqn\hbbfour{  h^{(1)}=e_s+2 e_{s+1}+3 e_{s+2}    }

3366: %

3367: \item{(3)} 

3368: \eqn\ifourc{ \It=  \pmatrix{~0& ~1& ~-2\cr ~1& ~0& ~1\cr ~-2& ~1& ~0 }   }

3369: and therefore

3370: \eqn\hbbfourb{  h^{(2)}=e_s+2 e_{s+1}- e_{s+2}    }

3371: In fact, $\It$ in \ifourc\ is the intersection matrix for the 

3372: type II string, and \hbbfourb\ is the unique higgs branch brane

3373: in this case.

3374: %

3375: \item{(4)} 

3376: \eqn\ifourd{ \It=  \pmatrix{~1& ~0& ~1\cr ~-2& ~1& ~0\cr ~1& ~-2& ~1 }   }

3377: This is the transpose of \ifourb.

3378: The higgs branch brane is 

3379: \eqn\hbbfourd{  h^{(3)}=e_s-2 e_{s+1}- e_{s+2}    }

3380: 

3381: \noindent

3382: We also claim that in $(2)-(4)$

3383: \eqn\cbbfour{ c_1=e_{s+1},\, c_2=e_{s+2},\quad {\rm or}\quad 

3384:           c_1=e_{s+2},\, c_2=e_{s+1}   }

3385: The reasoning goes as follows.

3386: Below we will see that the generalized central charge for

3387: the coulomb branch branes is given by \cgcc.

3388: This formula satisfies (in a certain angular sector)

3389: \eqn\ccnu{  \langle c_k|1\rangle\sim w\, \exp\left({2\pi i (k-2)\over n}\right)   +\OO(w^2),

3390: \qquad k=1,2} 

3391: The intersection form implies that $c_k$ and $e_k$ are possibly

3392: related as

3393: \eqn\cerel{   \pmatrix{~e_s\cr ~e_{s+1}}=\pmatrix{~a& ~b\cr ~c& ~d} \pmatrix{~c_1\cr ~c_2},\qquad

3394:      ad-bc=1                      }

3395: The consistency with \onept\ implies that in this angular sector

3396: $e_0=c_2,e_1=c_1$ and, more generally, \cbbfour.

3397: (The alternative solution is $e_1=c_1,e_2=-c_2$ implies ``negative mass'' for the 

3398: $c_1$ brane and therefore should be discarded.

3399: It is also not consistent with the formula for the higgs branch brane.)

3400: %

3401: 

3402: What happens as the phase of $w$ is varied?

3403: The relevant angular sectors are depicted in Fig. 4.

3404: There are two critical points (modulo LG symmetry)

3405: \eqn\critfour{ \nu=0:\;\; (x_1,x_3)=( {\sqrt{-w}\over 2},{\sqrt{-w}\over 2});\qquad

3406:   \nu=1:\;\; (x_1,x_3)=( {\sqrt{-w}e^{\pi i \over4}\over 2},{\sqrt{-w}e^{3\pi i\over4}\over 2})} 

3407: The choice of sign in the square root does not matter, the difference

3408: amounts to the LG symmetry.

3409: We choose a fundamental domain to be $0\leq\arg(-w)<2\pi$ which maps into 

3410: $0\leq\arg(\sqrt{-w})<\pi$.

3411: For a given value of $w$, $Re W_*$ has different signs for the two critical

3412: points.

3413: One critical point contributes a growing exponential, while the other a decaying

3414: one.

3415: The $w$ plane is divided into four big sectors, which can labeled by

3416: $\nu=0,1$ and $s=0,1,2,3$, depending on which critical point contributes

3417: a growing exponential to which $\hat I_s$.

3418: %

3419: %

3420: \midinsert\bigskip{\vbox{{\epsfxsize=3in

3421:         \nobreak

3422:     \centerline{\epsfbox{n4p.eps}}

3423:         \nobreak\bigskip

3424:     {\raggedright\it \vbox{

3425: {\bf Fig 4.}

3426: {\it The angular structure in the $w$ plane for $n=4$. When $w$ is in the shaded sectors,

3427: the $\nu=0$ ($\nu=1$)critical point has negative (positive) value of $Re\, W$.

3428: This is reversed in the unshaded sectors.

3429: Big blue arrows mark sectors which differ by permutations of 

3430: fractional branes. Dashed red and green arrows mark sectors which differ

3431: by the type of higgs branch brane}}}}}}

3432: \bigskip\endinsert

3433: %

3434: %

3435: %

3436: \eqn\nus{  

3437: \eqalign{   -{\pi\over4}&< \arg(w)<{\pi\over4}:\quad \nu=0,\;s=2\cr

3438:             {\pi\over4}&< \arg(w)<{3\pi\over4}:\quad \nu=1,\;s=3\cr

3439:             {3\pi\over4}&< \arg(w)<{5\pi\over4}:\quad \nu=0,\;s=0\cr

3440:             {5\pi\over4}&< \arg(w)<{7\pi\over4}:\quad \nu=1,\;s=1\cr}

3441: }

3442: These big sectors differ from each other by a permutation 

3443: of fractional branes, in exact analogy to the $n=3$ case.

3444: The exponentially decaying solution can be identified unambiguously;

3445: the logarithmically growing solution is defined up to an addition of exponentially

3446: decaying piece:

3447: \eqn\logexpf{ 

3448: \eqalign{   -{\pi\over4}&< \arg(w)<{\pi\over4}:\quad \hat I_1-\hat I_3\sim{\pi i\over2}\log w,\quad

3449:                                             \hat I_1+\hat I_3-2\hat I_0\sim \exp(-W_1)  \cr

3450:             {\pi\over4}&< \arg(w)<{3\pi\over4}:\quad \hat I_2-\hat I_0\sim{\pi i\over2}\log w,\quad

3451:                                             \hat I_2+\hat I_0-2\hat I_1\sim \exp(-W_0)  \cr

3452:             {3\pi\over4}&< \arg(w)<{5\pi\over4}:\quad \hat I_3-\hat I_1\sim{\pi i\over2}\log w,\quad

3453:                                             \hat I_3+\hat I_1-2\hat I_2\sim \exp(-W_1)  \cr

3454:             {5\pi\over4}&< \arg(w)<{7\pi\over4}:\quad \hat I_0-\hat I_2\sim{\pi i\over2}\log w,\quad

3455:                                             \hat I_2+\hat I_0-2\hat I_1\sim \exp(-W_0)  \cr}

3456: }

3457: %

3458: Let us restrict to the big sector $ -{\pi\over4}< \arg(w)<{\pi\over4}$.

3459: The exponentially decaying solution is given by the first equation

3460: in \logexpf.

3461: According to \fbgcc, the respective coulomb branch brane $c_2$ must 

3462: be equal to $e_0$.

3463: Hence, there are two possibilities, consistent with \cbbfour.

3464: Either $c_1=e_1$, $c_2=e_0$ and $s=3$ in \cbbfour\ (remember that $s$ is

3465: defined mod $n$.)

3466: Then, $\langle c_2|1\rangle\sim \hat I_1+\hat I_3-2\hat I_0$,

3467: $\langle c_1|1\rangle\sim \hat I_2+\hat I_0-2\hat I_1$.

3468: In this case

3469: \eqn\casea{

3470: \eqalign{  \langle h^{(1)}|1\rangle&\sim 4(\hat I_2-\hat I_1)\sim \exp(-W_0)   \cr

3471:            \langle h^{(2)}|1\rangle&\sim 4(\hat I_1-\hat I_0)\sim 2\pi i \log(w)   \cr

3472:            \langle h^{(3)}|1\rangle&\sim 4(\hat I_0-\hat I_3)\sim 2\pi i \log(w)   \cr}

3473: }

3474: The generalized central charge for $h^{(1)}$ has a growing exponential

3475: behavior instead of the expected logarithmic behavior.

3476: Alternatively, it might happen that 

3477: $c_1=e_3$, $c_2=e_0$, $s=2$ in \cbbfour, and

3478: hence $\langle c_1|1\rangle\sim \hat I_0+\hat I_2-2\hat I_3$,

3479: $\langle c_2|1\rangle\sim \hat I_1+\hat I_3-2\hat I_0$ and

3480: \eqn\caseb{

3481: \eqalign{  \langle h^{(1)}|1\rangle&\sim 4(\hat I_1-\hat I_0)\sim 2\pi i \log(w)  \cr

3482:            \langle h^{(2)}|1\rangle&\sim 4(\hat I_0-\hat I_3)\sim 2\pi i \log(w)   \cr

3483:            \langle h^{(3)}|1\rangle&\sim 4(\hat I_3-\hat I_2)\sim \exp(-W_0)   \cr}

3484: }

3485: Now it is $h^{(3)}$ whose behavior is not consistent with the expectations.

3486: 

3487: In the Appendix B we analyze the behavior of the A-brane surfaces.

3488: The results suggest that the sector $-\pi/4\leq\arg(w)\leq\pi/4$ is

3489: subdivided into the two subsectors,

3490: similar to the $n=3$ case.

3491: Whenever $0\leq\arg(w)\leq\pi/4$, we have $c_2=e_0$, $c_1=e_1$,

3492: and $\langle h|1\rangle\sim 4(\hat I_1-\hat I_0)$.

3493: That is, in this subsector 

3494: \eqn\hbfbfoura{ h=h^{(2)}=e_3+2e_0-e_1, \qquad 0\leq\arg(w)\leq {\pi\over4}  }

3495: Whenever $-\pi/4\leq\arg(w)\leq 0$, we have $c_1=e_3$, $c_2=e_0$,

3496: and $\langle h|1\rangle\sim 4(\hat I_0-\hat I_3)$.

3497: In this subsector we again have

3498: \eqn\hbfbfourb{ h=h^{(2)}=e_2+2e_3-e_0, \qquad -{\pi\over4}\leq\arg(w)\leq 0 }

3499: %It is quite amazing to observe that $h^{(3)}$, whose mass is negative

3500: %at the orbifold point is never realized\foot{This seems

3501: %to vindicate our anzats \az. Note that the multiplication

3502: %of \hbbfourd\ by $-1$ makes it a positive mass brane and does not affect the 

3503: %intersection matrix.

3504: %The higgs branch brane then takes a form similar to \hbbfourb.}.

3505: Hence, among the three possible expressions for the

3506: higgs branch brane consistent with the intersection form, 

3507: only one is realized.

3508: It is the one which appears in the type II case.

3509: [As before, there is a freedom to add D0 branes to \hbfbfoura, \hbfbfourb]

3510: 

3511: In Appendix B we also analyze the shape of the brane emanating from

3512: the $\nu=1$ critical point.

3513: In the analysis above it has been assumed that in the

3514: angular sector $-\pi/4\leq\arg(w)\leq\pi/4$, the corresponding

3515: central charge is given by $\langle c|1\rangle\sim \hat I_1+\hat I_3-2 \hat I_0$,

3516: since the value of $W$ at the critical point is positive and

3517: we expect the generalized central charge to decay exponentially.

3518: This is confirmed by solving the soliton equations in Appendix B.

3519: 

3520: 

3521: 

3522: 

3523: 

3524: 

3525: 

3526: 

3527: 

3528: 

3529: 

3530: 

3531: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%5

3532: \newsec{Discussion}

3533: In this paper we studied A-branes in the LG model which

3534: describes the resolution of the spacetime non-supersymmetric $\IC^2/\IZ_{n(p=1)}$

3535: orbifold.

3536: The model has coulomb branch branes, supported at the minima

3537: of the superpotential $W$ away from the origin, and higgs branch 

3538: branes, associated with the critical point at the origin.

3539: The generalized central charge for these branes, defined in Section 3, 

3540: satisfies the GKZ equation.

3541: This fact, together with a knowledge of the critical points of $W$,

3542: and the numerical analysis of the A-brane shape, allows

3543: us to compute the central charges for both coulomb and higgs branch branes.

3544: The results have the expected asymptotic behavior (determined by the

3545: value of $W$ at the corresponding critical point).

3546: They are also consistent with the open string Witten index.

3547: That is, in the examples that we have studied, we can 

3548: identify coulomb branch branes and fractional branes using

3549: the intersection matrix and first order conformal

3550: perturbation theory at the orbifold point.

3551: The higgs branch brane is given by a linear combination

3552: of fractional branes, so its generalized central charge

3553: can be computed accordingly.

3554: This procedure gives the same result as integrating

3555: \gcc\ over the higgs branch brane propeller surface\foot{

3556: Numerical data suggests that the propeller surface

3557: depends only weakly  on the LG metric: its asymptotics,

3558: which are important for the value of the integral, seem

3559: to be metric-independent.}.

3560: This result has logarithmic asymptotics as $|w|\ra\infty$,

3561: as expected from the form of the integrand at the

3562: origin.

3563: Similar logarithmic behavior arises also in the

3564: spacetime supersymmetric case.

3565:   

3566: The complex $w$ plane, is divided into $n$ angular 

3567: sectors, related by the permutation of the fractional branes.

3568: Hence, it is the phase of the tachyon VEV that is responsible 

3569: for a particular ordering of fractional branes being realized.

3570: Moreover, each sector is further divided into subsectors, where

3571: different expressions for the higgs branch brane are valid.

3572: %These expressions differ by the terms which are non-perturbative in $1/|w|$,

3573: %but this difference can be sugnificant near the orbifold point.

3574: 

3575: 

3576: In writing formulae for the higgs branch brane one must bear in mind that

3577: our techniques are not powerful enough to distinguish between $h$ and

3578: $h  + \ell D0$ where $\ell \in \IZ$ and $D0 = e_0 + \cdots + e_{n-1}$.

3579: Recall that $D0$ is in the annihilator of the intersection form.

3580: In principle this ambiguity could be fixed by determining the

3581: constant $C_*$ in $\langle h \vert 1 \rangle = C_* + \sum c_s \hat I_s$

3582: using the propeller surface to determine \gcc. Unfortunately we have

3583: not been able to extract this constant. In the case $n=3$ we found

3584: the combination $h=e_2 - e_0$ in \hbfbthreeb\ when $\ell=0$. Taken at face value

3585: this would be a massless brane at the orbifold point, signaling a

3586: breakdown in string perturbation theory. For this reason we find

3587: positive values of $\ell$ in \hbfbthreeb\ more likely. Clearly, further work is needed here.

3588: 

3589: 

3590: 

3591: 

3592: 

3593: 

3594: 

3595: %In the $n=3$ case, the subtraction of the D0 brane 

3596: %from the higgs branch brane renders the latter massless

3597: %at the orbifold fixed point.

3598: %one of the higgs branes becomes massless

3599: %at the orbifold point. [The expression in terms

3600: %of the fractional branes is given by the first eq. in \hbfbthree]

3601: %The presence of a massless brane usually signifies the breakdown of the string

3602: %perturbation theory.

3603: %In the present case, however, it is not clear whether such a massless

3604: %state appears in the physical spectrum.

3605: % as it may form a bound 

3606: %state with a D0 brane.

3607: %Rather it might be the bound state which includes the D0 brane.

3608: 

3609: %However, we do not necessarily expect any divergence in the string amplitudes

3610: %for $\IC^2/\IZ_{n(n-1)}$ orbifold in type 0 theory.

3611: %The identification of the higgs branch brane with the linear combination

3612: %of the fractional branes is initially based on the self-intersection number for

3613: %the non-trivial cycle in the resolution.

3614: %This implies that the higgs branch brane is well defined in the

3615: %region where $w$ is large.

3616: %However, these expressions are supported by the formulae for

3617: %the generalized central charge of the higgs and coulomb branch branes.

3618: %The linear combinations in \hbfbthree, \hbfbfoura, \hbfbfourb, ensure

3619: %that the cenral charge of the higgs branch brane has a logarithmic 

3620: %asymptotics as $|w|\ra\infty$.

3621: %This seems to suggest that the brane whose mass vanishes at the

3622: %orbifold fixed point, is not subject to decay, and remains 

3623: %in the physical spectrum as $w$ is taken to zero.

3624: %It would be very interesting to identify other signatures

3625: %of this brane in the type 0 orbifold theory.

3626: 

3627: %There are other ``massless'' branes in the theory.

3628: %Approximately half of the coulomb branch branes are assosiated

3629: %with the critical points where the value of $Re\, W$ is positive.

3630: %Hence, their central charge is approaching zero exponentially fast.

3631: %The physical interpretation of this fact in the non-conformal

3632: %theory is not completely clear.

3633: %On the one hand, in the conformal (e.g. supersymmetric) case,

3634: %the generalized central charge as related to the mass of 

3635: %the respective brane, and its vanishing would signal

3636: %the appearance of a massless brane in the spectrum.

3637: Among the coulomb branch branes, approximately one half 

3638: have exponentially growing periods while the other half have exponentially

3639: decaying  periods.

3640: The block-diagonalization of the

3641: intersection matrix implies that the coulomb branch branes are

3642: decoupled from their higgs branch counterparts.

3643: Surprisingly, the two coulomb branch branes in $\IC^2/\IZ_{4(1)}$

3644: are not orthogonal to each other, even though the coulomb

3645: vacua are far separated in the IR!

3646: It would be interesting to analyze in detail the patterns

3647: at higher values of $n$ where many coulomb branch branes

3648: associated with critical points with both positive and 

3649: negative values of $Re\, W$ are present.

3650: This might be useful for the physical interpretation of \gcc\ in the non-conformal 

3651: case.

3652: 

3653: This work raises a number of technical issues which must be 

3654: solved in order to make further progress. 

3655: We used numerical analysis to determine the shape of

3656: the propeller branes.

3657: The wings depend on the angular sectors in the $w$ plane;

3658: the asymptotics change discontinuously as $w$ crosses the

3659: borders of these sectors.

3660: It would be nice to have some analytic technology to

3661: understand these phenomena better.

3662: In the $n=4$ case, some wings developed which did not seem

3663: to contribute to \gcc.

3664: This will probably be a persistent issue for

3665: higher $n$, and  understanding better the shape of the propeller

3666: brane and \gcc\ is important for making progress. Similarly, 

3667: we have made an important assumption that the asymptotics of the 

3668: propeller branes is independent of the choice of metric 

3669: $g_{i\bar j}$ used in \soleq. We have checked this numerically 

3670: for the metrics of interest here (see appendix B) but some 

3671: rigorous results concerning this would be most welcome. 

3672: Finally, in constructing a  map between the fractional

3673: branes and the coulomb branch branes, it is would be very useful

3674: to know the intersection form for the latter.

3675:  

3676: Understanding the structure of branes in more general

3677: spacetime non-supersymmetric orbifolds is another interesting

3678: direction.

3679: At present, the higgs branch brane can be expressed in

3680: terms of fractional branes for the $\IC^2/\IZ_{n(p)}$ orbifolds

3681: which have a resolution in type II theory \mm.

3682: Generalizing this result to type 0 theory for arbitrary

3683: $n$ and $p$ will probably involve understanding phase

3684: diagrams of multiple tachyons, and the space

3685: of possible higgs branch branes.

3686: It would also be worthwhile understanding spacetime nonsupersymmetric

3687: $\IC^3/\IZ_n$ orbifolds, where some new features appear already in

3688: the closed string sector \refs{\MNP\SarkarRY-\MorrisonJA}.

3689: 

3690: 

3691: The Stokes' phenomenon observed in this paper is very likely related to 

3692: that associated to general

3693: semisimple Frobenius manifolds in \dubrovinI\dubrovinII. It is possible 

3694: that some of the techniques

3695: used in \guzzetti\ueda\ can be applied to elucidate the behavior of 

3696: generalized periods and their

3697: Stokes matrices for general $\IC^2/\IZ_{n(1)}$ orbifolds, or even 

3698: $\IC^2/\IZ_{n(p)}$ orbifolds.

3699: We hope our considerations will be useful in understanding homological 

3700: mirror symmetry for non-Fano manifolds. In the Fano case Stokes' 

3701: matrices are related to the dimensions of Ext groups of exceptional collections 

3702: in the derived  category of the mirror \guzzetti\ueda. 

3703: (For recent progress in homological mirror symmetry in the Fano case see 

3704: \arouxI\arouxII.) In the examples of the  $\IC^2/\IZ_{n(p)}$ orbifolds 

3705: we are instead trying to formulate a quantum version of the McKay 

3706: correspondence, as explained in \mm\MP. One point which is currently 

3707: missing is an  analogous interpretation of the Stokes' matrices

3708: in the non-Fano case.

3709: 

3710: 

3711: 

3712: Stokes' phenomenon has recently played an important role in brane physics in the 

3713: context of minimal string theory \MaldacenaSN. It is interesting to contrast 

3714: that application with the present one. In both cases a ``brane partition function'' 

3715: satisfies a differential equation with an irregular singular point, and the 

3716: angular-sector-dependence of exponential growth and decay has important

3717: physical consequences. In both cases one can  use branes to probe the nature 

3718: of spacetime, and Stokes' phenomenon has important implications for  the resulting 

3719: spacetime picture. On the other hand, in \MaldacenaSN\ one works with the 

3720: {\it non-perturbative} brane amplitude (the Baker-Akhiezer function of the matrix 

3721: model, now also known as the ``FZZT partition function'' ) but the present paper 

3722: only makes use of the {\it perturbative} disk one-point function.  Nevertheless, 

3723: there is a common mathematical thread in both examples, since  both examples 

3724: are governed by a family of Landau-Ginzburg theories, and hence by a similar 

3725:  underlying structure of a Frobenius manifold. 

3726: 

3727: 

3728: In conclusion there is an amazingly rich structure in the D-branes of the

3729: $\IC^2/\IZ_{n(1)}$ orbifolds. 

3730: Surely it is even more intricate in the  $\IC^2/\IZ_{n(p)}$ and

3731: $\IC^n/\Gamma$ orbifolds.

3732: Elucidating this structure appears to be a challenging project.

3733: 

3734: 

3735: 

3736: 

3737: 

3738: 

3739: 

3740: 

3741: 

3742: 

3743: 

3744: 

3745: 

3746: 

3747: 

3748: 

3749: 

3750: 

3751: 

3752: 

3753: 

3754: 

3755: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

3756: \bigskip\bigskip\noindent{\bf Acknowledgements:}

3757: We would like to thank E. Diaconescu, R. Karp and R. Plesser 

3758: for discussions.

3759: We also thank B. Florea for initial participation in the project

3760: and for many useful discussions.

3761: This work was supported in part by  DOE grant DE-FG02-96ER40949.

3762: 

3763: 

3764: 

3765: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%5

3766: \appendix{A}{Boundary states for fractional branes}

3767: In this appendix we describe in more detail the

3768: construction of the boundary states which correspond to fractional 

3769: branes at the $\IC^2/\IZ_N$ orbifold.

3770: There are two cases: (1) $p$ odd, $N=n$; (2) $p$ even, $N=2n$.

3771: A useful reference is \bcr.

3772: The first step is constructing B-type Ishibashi boundary states

3773: \eqn\ishia{

3774: \eqalign{

3775:      (\alpha_{m-\nu}-{\tilde\alpha}_{-m+\nu})|s;\eta\rangle\rangle_{NSNS,RR}&=0,

3776:                 \qquad m=1,2,\ldots,\quad \nu={s\over n}   \cr

3777:      ({\bar\alpha}_{m+\nu}-{\tilde{\bar\alpha}}_{-m-\nu})|s;\eta\rangle\rangle_{NSNS,RR}&=0,

3778:                 \qquad m=0,1,\ldots,\quad    \cr}

3779: }

3780: and 

3781: \eqn\ishib{

3782: \eqalign{ (\psi_{r-\nu}+i\eta {\tilde\psi}_{-r+\nu})|s;\eta\rangle\rangle_{NSNS,RR}&=0,

3783:                 \qquad r={1\over 2}+\IZ(\IZ)\quad {\rm for\; NS(R) },\; r-\nu\geq 0  \cr

3784:           ({\bar\psi}_{r+\nu}+i\eta {\tilde{\bar\psi}}_{-r-\nu})|s;\eta\rangle\rangle_{NSNS,RR}&=0,

3785:                 \qquad  r+\nu\geq 0,\quad  \eta=\pm1 \cr}

3786: }

3787: One can verify that conditions \ishia\ and \ishib\ give rise to

3788: the B-boundary states of the $\NN=(2,2)$ superconformal theory:

3789: \eqn\abstates{

3790: \eqalign{ 

3791:                 (G_r^++i\eta {\tilde G}_{-r}^+)|s;\eta\rangle\rangle_{NSNS,RR}&=0,\quad

3792:                 (G_r^-+i\eta {\tilde G}_{-r}^-)|s;\eta\rangle\rangle_{NSNS,RR}=0    \cr

3793:                 (J_n+{\tilde J}_{-n})|s;\eta\rangle\rangle_{NSNS,RR}&=0              \cr}

3794: }

3795: %

3796: In the open string sector the basic ingredients are the characters

3797: with the insertion of the group element $g^s$ where $g$ is the generator of $\IZ_n$

3798: \eqn\chars{ \chi_{NS,R}^{s}(q_o)=\tr_{NS,R}\, g^s q_o^{L_0-{c\over 12}},\qquad

3799:           \chi_{NS,R}^{(-)^F;s}(q_o)=\tr_{NS,R}\, (-)^F g^s q_o^{L_0-{c\over 12}}   }

3800: Here $s$ runs from $1$ to $N$ and $q_o=\exp(2\pi i \tau_o)$.

3801: When $p$ is even and $N=2n$, the number of independent characters

3802: can be reduced to $n$ by the following identity

3803: \eqn\idchars{  \chi_{NS}^{s+n}(q_o)= \chi_{NS}^{s}(q_o),\qquad 

3804:              \chi_{NS}^{(-)^F;s+n}(q_o)= \chi_{NS}^{(-)^F;s}(q_o)    }

3805: and

3806: \eqn\idcharsb{  \chi_{R}^{s+n}(q_o)= -\chi_{R}^{s}(q_o),\qquad 

3807:              \chi_{R}^{(-)^F;s+n}(q_o)= -\chi_{R}^{(-)^F;s}(q_o)    }

3808: These identities follow from \orbaction.

3809: In the $NS$ sector $g^n=1$, while in the $R$ sector $g^n=(-)^{p+1}=-1$.

3810: The characters \chars\ have the following modular transformation properties 

3811: under $\tau_o\ra \tau_c=-1/\tau_o$:

3812: \eqn\mtchar{

3813: \eqalign{ \chi_{NS}^{s}(q_o)&=\sigma(s)\; {}_{NSNS}\langle\langle s;\pm|

3814:            q_c^{{1\over 2}(L_0+{\tilde L}_0-{c\over 12})} |s\pm\rangle\rangle_{NSNS} \cr

3815:         \chi_{NS}^{(-)^F;s}(q_o)&=-\sigma(s)\; {}_{RR}\langle\langle s;\pm|

3816:            q_c^{{1\over 2}(L_0+{\tilde L}_0-{c\over 12})} |s\pm\rangle\rangle_{RR} \cr

3817:         \chi_{R}^{s}(q_o)&=\sigma(s)\; {}_{NSNS}\langle\langle s;\pm|

3818:            q_c^{{1\over 2}(L_0+{\tilde L}_0-{c\over 12})} |s\mp\rangle\rangle_{NSNS} \cr

3819:         \chi_{R}^{(-)^F;s}(q_o)&=-\sigma(s)\; {}_{RR}\langle\langle s;\pm|

3820:            q_c^{{1\over 2}(L_0+{\tilde L}_0-{c\over 12})} |s\mp\rangle\rangle_{RR} \cr}

3821: }

3822: where 

3823: \eqn\sigmadef{ \sigma(s)=4\sin\left({\pi s\over n}\right)\sin\left({\pi s p\over n}\right)  }

3824: With the exception of the factor $\sigma(s)$, the modular transformation properties

3825: \mtchar\ are the same as those of the untwisted characters.

3826: The boundary states are constructed by requiring open-closed string duality.

3827: Consider $N=n$ case first.

3828: \eqn\ocdual{

3829: \eqalign{  \tr_{ab;NS}\; q_o^{L_0-{c\over 12}}&=

3830:             {1\over n}\sum_{s=0}^{n-1} \omega_n^{(a-b)s} \sigma(s) {}_{NSNS}\langle\langle s;\pm|

3831:            q_c^{{1\over 2}(L_0+{\tilde L}_0-{c\over 12})} |s\pm\rangle\rangle_{NSNS}         \cr

3832:   &={}_{NSNS}\langle a;\pm|q_c^{{1\over 2}(L_0+{\tilde L}_0-{c\over 12})} |b;\pm\rangle_{NSNS}  \cr}  

3833: }

3834: where we introduced the Cardy state $|a;\eta\rangle_{NSNS}$ and $\omega_n=\exp(2\pi i/n)$.

3835: Eq. \ocdual\ and its counterpart with $\tr_{ab;R}\; q_o^{L_0-{c\over 12}}$ in the RHS implies

3836: \eqn\cardya{|a;\eta\rangle_{NSNS}=\sum_{s=0}^{n-1}  \omega_n^{a s} \sqrt{\sigma(s)}\; 

3837:            |s;\eta\rangle\rangle_{NSNS}                     }

3838: Similarly,

3839: \eqn\cardyb{|a;\eta\rangle_{RR}=\sum_{s=0}^{n-1}  \omega_n^{a s} \sqrt{\sigma(s)}\; 

3840:            |s;\eta\rangle\rangle_{RR}                     }

3841: Type 0 theory admits two types of branes distinguished by the value of $\eta$:

3842: %which are sometimes called ``electric'' and ``magnetic'' branes:

3843: \eqn\tzerob{ |a;+\rangle={1\over\sqrt{2}}(|a;+\rangle_{NSNS}+|a;+\rangle_{RR}),\quad

3844:              |a;-\rangle={1\over\sqrt{2}}(-|a;-\rangle_{NSNS}+|a;-\rangle_{RR})      }

3845: In type II theory, only one combination is invariant under the GSO projection

3846: \eqn\typetwob{ |a;II\rangle={1\over 2}

3847:            (|a;+\rangle_{NSNS}-|a;-\rangle_{NSNS}+|a;+\rangle_{RR}+|a;-\rangle_{RR})       }

3848: The situation with $p$ even $n$ odd, where type II can not be defined and

3849: the orbifold group is $\IZ_N=\IZ_{2n}$ is a little bit more tricky.

3850: Now $a$ in the open string sector is forced to run from $0$ to $2n-1$ and the analog of 

3851: \ocdual\ is

3852: \eqn\ocduala{   \tr_{ab;NS}\; q_o^{L_0-{c\over 12}}=

3853:             {1\over 2n}\sum_{s=0}^{n-1} (1+(-)^{a-b})

3854:            \omega_{2n}^{(a-b)s} \sigma(s) {}_{NSNS}\langle\langle s;\pm|

3855:            q_c^{{1\over 2}(L_0+{\tilde L}_0-{c\over 12})} |s\pm\rangle\rangle_{NSNS}         }

3856: where we converted the sum over $s=0,\ldots,2n-1$ to the sum over   $s=0,\ldots,n-1$

3857: using \idchars\ and $\omega_{2n}=\exp(2\pi i/2 n)$.

3858: The closed string sector in type 0 is invariant under $s\ra s+n$, that

3859: is why there are $n$ Ishibashi states in \ocduala.

3860: The sum in \ocduala\ is zero unless $a-b$ is even.

3861: This prompts us to introduce two types of branes: the ones with $a=2a';\;\eta=+1$

3862: and the ones with $a=2a'+1;\;\eta=-1$.

3863: (Now, in addition to $\IZ_n$ quantum symmetry which permutes the branes, 

3864: there is a $\IZ_2$ symmetry which amounts to multiplying $\eta$ by $-1$.) 

3865: The boundary states are

3866: \eqn\cardyc{|a';+\rangle_{NSNS,RR}=\sum_{s=0}^{n-1}  \omega_n^{a' s} \sqrt{\sigma(s)}\; 

3867:            |s;+\rangle\rangle_{NSNS,RR}                     }

3868: and

3869: \eqn\cardyd{|a';-\rangle_{NSNS,RR}=\sum_{s=0}^{n-1}  \omega_n^{(a'+{1\over2}) s} \sqrt{\sigma(s)}\; 

3870:            |s;-\rangle\rangle_{NSNS,RR}     }

3871: These boundary states are consistent with \ocduala\ and its counterpart with 

3872: $(-)^F$ inserted in the LHS.

3873: They are also consistent with the corresponding open string expressions

3874: in the $R$ sector:

3875: \eqn\ocdualb{   \tr_{ab;R}\; q_o^{L_0-{c\over 12}}=

3876:             {1\over 2n}\sum_{s=0}^{n-1} (1+(-)^{a-b-1})

3877:            \omega_{2n}^{(a-b)s} \sigma(s) {}_{NSNS}\langle\langle s;+|

3878:            q_c^{{1\over 2}(L_0+{\tilde L}_0-{c\over 12})} |s-\rangle\rangle_{NSNS}         }

3879: and

3880: \eqn\ocdualc{   \tr_{ab;R}\; (-)^F q_o^{L_0-{c\over 12}}=

3881:             {1\over 2n}\sum_{s=0}^{n-1} (1+(-)^{a-b-1})

3882:            \omega_{2n}^{(a-b)s} \sigma(s) {}_{RR}\langle\langle s;+|

3883:            q_c^{{1\over 2}(L_0+{\tilde L}_0-{c\over 12})} |s-\rangle\rangle_{RR}         }

3884: 

3885: 

3886: 

3887: 

3888: 

3889: 

3890: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

3891: \appendix{B}{Shape of A-branes}

3892: In this appendix we analyze the shape of A-branes in the LG theory.

3893: As explained in \HIV\ these branes are Lagrangian surfaces

3894: whose image in the $W$ plane is a semi-infinite real line emanating

3895: from the critical point $\phi^i=\phi^i_*$ and going in the positive real direction.

3896: More practically, one is instructed to solve the soliton equation \soleq.

3897:  The shape of the A-brane associated with a given critical point is

3898: the set of all trajectories satisfying \soleq.

3899: Near the critical point, the set of solutions can be parametrized

3900: by a small sphere, as in \sphere.

3901: 

3902: Let us specialize to our LG model.

3903: The superpotential is

3904: \eqn\spotl{ W=x_1^n+w x_1 x_3+x_3^n  }

3905: To write the soliton equations we must make a choice of metric. The measure factor 

3906: in \gcc\ (which originates from a path integral) suggests that one should use the 

3907: metric $ds^2 = \vert{dx_1 \over x_1} \vert^2 + \vert{dx_3 \over x_3} \vert^2$ for 

3908: large $x_i$. This metric is inconvenient for displaying the results of the numerical 

3909: analysis because the soliton equations develop singularities at finite values of $\sigma$. 

3910: We have instead used the metric $ds^2 = \vert{x_1 dx_1  } \vert^2 + \vert{x_3 dx_3  } \vert^2$ 

3911: because this is the simplest metric for which there is no singularity at finite $\sigma$. 

3912: A tedious numerical check shows that the asymptotics of the propeller branes described below 

3913: is in fact independent of the choice of the metric. It would be very useful to establish 

3914: this rigorously. Some further comments on metric dependence can be found below. 

3915: 

3916: The soliton equations are

3917: \eqn\soleqlg{   {dx_1\over d\sigma}={1\over|x_1|^2}(n \bar x_1^{n-1}+\bar w \bar x_3),\qquad

3918:    {dx_3\over d\sigma}={1\over|x_3|^2}(n \bar x_3^{n-1}+\bar w \bar x_1)}

3919: The non-zero critical points are given by \lgiiip\ and the value of $W_*$

3920: by \lgivp.

3921: Substituting $x_1=x_1^{(\nu)}+\delta x_1, x_3=x_3^{(\nu)}+\delta x_3$ we have

3922: \eqn\spexp{  W=W_{\nu}+{w\over 2}\left( e^{-{2\pi i\nu\over n}}(1-n)\delta x_1^2

3923:                   +2\delta x_1 \delta x_3  + e^{2\pi i\nu\over n}(1-n)\delta x_3^2\right)   }

3924: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%5

3925: \subsec{n=3}

3926: For $n=3$ there is a single critical point with $\nu=0$, $x_1=x_3=-w/3$ and $W_0=w^3/27$.

3927: The change of coordinates which diagonalizes \spexp\ is

3928: \eqn\ccoord{ u={\delta x_1+\delta x_3\over\sqrt{2}},\qquad   v={\delta x_1-\delta x_3\over\sqrt{2}}  }

3929: which recasts \spexp\ as

3930: \eqn\spexpa{  W=W_{\nu}+|w| \exp(i\varphi_w+i\pi)\left({1\over 2}u^2+{3\over2}v^2\right)   }

3931: where we introduced $\varphi_w=\arg(w)$.

3932: The wavefront is a circle and can be parameterized by a single angle $\theta$

3933: \eqn\wf{  

3934: \eqalign{   x_1&=-{w\over3}+{\epsilon\over\sqrt{|w|}} e^{-i{\varphi_w+\pi\over 2}}\left(\cos\theta+

3935:                  {1\over\sqrt{3}}\sin\theta\right)   \cr

3936:              x_3&=-{w\over3}+{\epsilon\over\sqrt{|w|}} e^{-i{\varphi_w+\pi\over 2}}\left(\cos\theta-

3937:                  {1\over\sqrt{3}}\sin\theta\right)   \cr}

3938: }

3939: When $x_1,x_3$ are large, the $w x_1 x_3$ term in the superpotential \spotl\

3940: is negligible.

3941: Therefore in this regime the 

3942: equations for $x_1$ and $x_3$ decouple and the A-brane degenerates to the product of lines

3943: $x_1^n\in \IR_+$ and $x_3^n\in \IR_+$, which are used to define $\hat I$'s.

3944: Each of these surfaces is essentially a quarter plane.

3945: There are $n^2$ choices of defining lines, but

3946: the LG symmetry brings it down to $n$ inequivalent surfaces.

3947: Generally, a circle parametrized by $\theta$ splits into 

3948: several components; at large worldsheet time $\sigma$, each component 

3949: traces some quarter plane described above.

3950: Therefore the shape of the A-brane surfaces resembles that of a propeller,

3951: and we call these surfaces ``propeller branes''.

3952: The integral of \gcc\ over each wing of the propeller defines a function $\hat I_s$. 

3953: 

3954: By solving the soliton equations \soleqlg\ with the initial conditions

3955: \wf\ we can determine which combination of  $\hat I$'s corresponds to 

3956: a given coulomb branch brane.

3957: The relevant differential equation can be solved by Mathematica.

3958: In Fig. 5,6  we present a plot of $\arg(x_1), \arg(x_3)$ as functions 

3959: of $\sigma$ for several values of $\theta$.

3960: 

3961: 

3962: 

3963: %

3964: %

3965: \midinsert\bigskip{\vbox{{\epsfxsize=3in

3966:         \nobreak

3967:     \centerline{\epsfbox{n3hivx1.eps}}

3968:         \nobreak\bigskip

3969:     {\raggedright\it \vbox{

3970: {\bf Fig 5.}

3971: {\it Solution of soliton equation. Horizontal axis corresponds

3972: to $\sigma$. Vertical axis is $\arg(x_1(\sigma))/\pi$.

3973: The lines are computed for $w=\exp(0.05 \pi i)$, $\theta=0$ (orange),

3974: $\pi/2$ (pink), $3\pi/2$ (turquoise), $\pi$ (blue).

3975: They asymptote to $\pm2/3$ as $\sigma\ra\infty$.}}}}}}

3976: \bigskip\endinsert

3977: %

3978: %

3979: %

3980: %

3981: \midinsert\bigskip{\vbox{{\epsfxsize=3in

3982:         \nobreak

3983:     \centerline{\epsfbox{n3x3.eps}}

3984:         \nobreak\bigskip

3985:     {\raggedright\it \vbox{

3986: {\bf Fig 6.}

3987: {\it Solution of soliton equation. Horizontal axis corresponds

3988: to $\sigma$. Vertical axis is $\arg(x_3(\sigma))/\pi$.

3989: The lines are computed for $w=\exp(0.05 \pi i)$, $\theta=0$ (orange),

3990: $\pi/2$ (pink), $3\pi/2$ (turquoise), $\pi$ (blue).

3991: They asymptote to  $\pm2/3$ as $\sigma\ra\infty$.}}}}}}

3992: \bigskip\endinsert

3993: From Figs. 5,6 we infer that the pink [turquoise] line which corresponds to

3994: $\theta=\pi/2$ [$\theta=3\pi/2$] asymptotes to $(\arg(x_1),\arg(x_3))=(2\pi/3,-2\pi/3)$

3995: [ $(\arg(x_1),\arg(x_3))=(-2\pi/3,2\pi/3)$.]

3996: For $e^{i\theta}$ in most of the upper (lower) half plane, the asymptotics are similar to those

3997: of $\theta=\pi/2$ ($\theta=3\pi/2$).

3998: These two regions define two different representatives of $\hat I_0$

3999: related by the LG symmetry.

4000: For $\theta\approx 0$ the asymptotics corresponds to $\hat I_1$,

4001: while for $\theta\approx \pi$ we get $\hat I_{-1}$.

4002: Hence, we observe that the behavior of solutions is consistent

4003: with the identification 

4004: \eqn\pbnt{ \langle c|1\rangle\sim \hat I_1+\hat I_{-1}-2 \hat I_0, \qquad  -\pi/3\leq\arg(w)\leq\pi/3 }

4005: In fact, Figs. 5,6 do not qualitatively change in 

4006: the whole ``big'' angular sector $-\pi/3\leq\arg(w)\leq\pi/3$.

4007: The qualitative change happens when $w$ crosses the lines $\arg(w)=-\pi/3$

4008: and $\arg(w)=\pi/3$.

4009: The asymptotics of the coulomb branch brane changes in accord with 

4010: the rule proposed above: multiplying $w$ by $\exp(2\pi i/3)$ (rotating

4011: the ``big'' sector) corresponds to the permutation of the fractional

4012: branes.

4013: 

4014: In the discussion above we determined that the integral \gcc\ over

4015: the propeller surface originating at the coulomb branch critical point

4016: gives rise to \pbnt.

4017: This is because the propeller surface in question has four wings.

4018: Integrating \gcc\ over these wings gives rise to $I_1$, $I_{-1}$ 

4019: and $I_0$ (twice).

4020: Eq. \pbnt\ is then consistent with \linde, with 

4021: the one-point function for the fractional brane, and with

4022: the intersection matrix.

4023: Nevertheless it is desirable to have an independent way of

4024: determining the orientation of the wings, i.e. the signs in

4025: eq. \pbnt.

4026: Consider a wing defined by $(x_1,x_3)=(e^{2\pi i s_1\over n} t_1,e^{2\pi i s_3\over n} t_3)$

4027: for large positive $t_1$ and $t_3$.

4028: A soliton trajectory originating at a given value of $\theta$

4029: gives rise to a ray in $(t_1,t_3)$ plane.

4030: In the simplest scenario, the slope of this ray $\gamma=|x_1|/|x_3|$ is a monotonic 

4031: function of $\theta$.

4032: The orientation is then determined by the sign of $d\gamma/d\theta$.

4033: In Fig. 7 we plot $\gamma(\theta)$.

4034: There are four regions where this function is monotonic; it

4035: takes all values between zero and infinity.

4036: These four regions correspond to  $I_1$, $I_{-1}$ 

4037: and $I_0$ (twice).

4038: The signs of $d\gamma/d\theta$ are consistent with the signs that appear in

4039: \pbnt.

4040: %

4041: \midinsert\bigskip{\vbox{{\epsfxsize=3in

4042:         \nobreak

4043:     \centerline{\epsfbox{orientn3.eps}}

4044:         \nobreak\bigskip

4045:     {\raggedright\it \vbox{

4046: {\bf Fig 7.}

4047: {\it Orientation of the wings ($n=3$) is determined by $\gamma(\theta)$.

4048: Vertical axis is $\gamma\equiv|x_1|/|x_3|$.

4049: Horizontal axis is $\theta/\pi$.}}}}}}

4050: \bigskip\endinsert

4051: %

4052: It is also interesting to investigate the dependence of the solutions

4053: of soliton equations on the metric.

4054: %In the discussion above, we used the metric for LG variables which

4055: %is good as $|x_1|,|x_3|\ra\infty$.

4056: %Recall that the LG model was obtained from the GLSM by a T-duality

4057: %and subsequent change of variables $x_i=\exp(-Y_i)$.

4058: %The last step brings canonical kinetic terms to the form reflected in \soleqlg.

4059: One natural metric to try is the Euclidean metric, leading to 

4060: soliton equations of the form

4061: %

4062: \eqn\soleqtm{   {dx_1\over d\sigma}=(n \bar x_1^{n-1}+\bar w \bar x_3),\qquad

4063:    {dx_3\over d\sigma}=(n \bar x_3^{n-1}+\bar w \bar x_1)}

4064: We analyzed the solutions of these equations emanating from 

4065: a nonzero critical point.

4066: The technical difference with \soleqlg\ is that now $x_1$ or $x_3$

4067: run off to infinity at finite values of $\sigma$.

4068: Moreover, for generic values of $\theta$ only one of $x_1, x_3$

4069: runs off to infinity, and only for isolated values of $\theta$ 

4070: both of them do so.

4071: All of this makes eqs. \soleqtm\ more difficult to study numerically

4072: then eqs. \soleqlg.

4073: The numerical data suggests that solutions of \soleqtm\ asymptote

4074: to the surface that is homologous at the infinity to the one

4075: defined by  \soleqlg. We have also carried out similar checks of metric 

4076: independence for the metric 

4077: $ds^2 = \vert{dx_1 \over x_1} \vert^2 + \vert{dx_3 \over x_3} \vert^2$ 

4078: (this metric is the natural one to expect from the derivation of the 

4079: LG theory via $T$-duality). 

4080: All this suggests that the behavior of the A-brane surface depends only weakly

4081:  on the metric, and the asymptotics might stay the same 

4082: for a large class of metric deformations.

4083: It is now natural to try to  analyze the behavior of the higgs branch brane as well.

4084: The metric near the origin is not known, but if the asymptotics do not

4085: depend on it, we might as well use eqs. \soleqtm.

4086: The results confirm the picture described in the main text.

4087: Restricting to the big angular sector $-\pi/3<\arg(w)<\pi/3$,

4088: the solutions experience an abrupt change as $w$ crosses the line $\arg(w)=0$.

4089: Above this line, the solutions asymptote to the surface defining 

4090: $3(\hat I_1-\hat I_0)$, and below the line to $3(\hat I_0- \hat I_{-1})$. 

4091: 

4092: 

4093: 

4094: 

4095: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

4096: \subsec{n=4}

4097: Consider now $n=4$.

4098: We will restrict our analysis to the angular sector

4099: defined by $-\pi/4<\arg(w)<\pi/4$ (see Fig. 4).

4100: Other big sectors are related to this one by the LG symmetry.

4101: We studied the surface corresponding to the higgs branch brane (with the flat metric).

4102: The result is

4103: \eqn\hbf{ 

4104: \eqalign{   0<\arg(w)<{\pi\over4}:\quad &\langle h|1\rangle\sim 4(\hat I_1-\hat I_0) \cr

4105:            -{\pi\over4}<\arg(w)<0:\quad &\langle h|1\rangle\sim 4(\hat I_3-\hat I_0) \cr }

4106: }

4107: Consider now the contribution of the critical point with $\nu=0$,

4108: $x_1=x_3=\sqrt{-w}/2$, $W=-2 (w/2)^2$.

4109: (We denote the corresponding coulomb branch brane by $c_1$).

4110: As explained before, it contributes a growing exponential, hence

4111: it is necessary to solve the soliton equations to determine the

4112: asymptotics.

4113: Note that in the unshaded sectors in Fig. 4 this critical point 

4114: contributes a decaying exponential. 

4115: The integral \gcc\ over the coulomb branch branes associated with it,

4116: should give rise to the following overlaps:

4117: \eqn\oldec{

4118: \eqalign{  {\pi\over4}<\arg(w)<{3\pi\over4}:\quad \langle c_1|1\rangle&\sim

4119:                                    \hat I_2+\hat I_0-2 \hat I_1 \cr

4120: {5\pi\over4}<\arg(w)<{7\pi\over4}:\quad \langle c_1|1\rangle&\sim

4121:                                    \hat I_0+\hat I_2-2 \hat I_3 \cr}

4122: }

4123: Returning back to the soliton equations, the change of variable which 

4124: diagonalizes \spexp\ is

4125: \eqn\diagspf{ u=e^{i\varphi_w+i\pi\over2}(\delta x_1+\delta x_3),\qquad

4126: v=e^{i\varphi_w+i\pi\over2}(\delta x_1-\delta x_3)                   }

4127: which gives rise to

4128: \eqn\wfourcp{W=W_0+{|w|\over 2} (u^2+2 v^2)    }

4129: The wavefront near the critical point is 

4130: \eqn\wffoura{    

4131: \eqalign{   x_1&={\sqrt{-w}\over2}+{\epsilon\over\sqrt{|w|}} e^{-i{\varphi_w+\pi\over 2}}\left(\cos\theta+

4132:                  {1\over\sqrt{2}}\sin\theta\right)   \cr

4133:              x_3&={\sqrt{-w}\over2}+{\epsilon\over\sqrt{|w|}} e^{-i{\varphi_w+\pi\over 2}}\left(\cos\theta-

4134:                  {1\over\sqrt{2}}\sin\theta\right)   \cr}

4135: }

4136: The solutions exhibit more complicated behavior than what 

4137: we have seen so far.

4138: As mentioned above, whenever $w$ is in the unshaded sector,

4139: we expect the integral \gcc\ over the wings of the propeller

4140: surface to give \oldec.

4141: This is indeed what the numerical analysis tells us.

4142: Suppose $\pi/4\leq\arg(w)\leq3\pi/4$.

4143: Then integrating \gcc\ over the wings gives rise to

4144: the first equation of \oldec.

4145: Crossing the line $\arg(w)=\pi/4$ into the shaded sector

4146: does not change the solution.

4147: Figs. 8--10 contain the graphs of $\arg(x_1)$, $\arg(x_3)$

4148: as functions of $\theta$ for large worldsheet time $\sigma$.

4149: According to Figs. 8--10, the propeller surface emanating

4150: from the $\nu=0$ critical point has four wings.

4151: Integrating \gcc\ over these wings gives rise to the first

4152: equation in \oldec.

4153: (This is similar to the $n=3$ case)

4154: %

4155: %

4156: \midinsert\bigskip{\vbox{{\epsfxsize=3in

4157:         \nobreak

4158:     \centerline{\epsfbox{n4argsa.eps}}

4159:         \nobreak\bigskip

4160:     {\raggedright\it \vbox{

4161: {\bf Fig 8.}

4162: {\it Solution of soliton equation at $\epsilon=0.005$, $\sigma=8$, $w=w^{0.9\pi i/4}$.

4163: Horizontal axis is $\theta/\pi$. 

4164: Vertical axis is $\arg(x_i(\sigma))/\pi$. Red line is $i=1$. Green line $i=3$.}}}}}}

4165: %

4166: \bigskip\endinsert

4167: \midinsert\bigskip{\vbox{{\epsfxsize=3in

4168:         \nobreak

4169:     \centerline{\epsfbox{n4argsb.eps}}

4170:         \nobreak\bigskip

4171:     {\raggedright\it \vbox{

4172: {\bf Fig 9.}

4173: {\it Close-up view of Fig. 8}}}}}}

4174: %

4175: \bigskip\endinsert

4176: \midinsert\bigskip{\vbox{{\epsfxsize=3in

4177:         \nobreak

4178:     \centerline{\epsfbox{n4argsc.eps}}

4179:         \nobreak\bigskip

4180:     {\raggedright\it \vbox{

4181: {\bf Fig 10.}

4182: {\it Another close-up view of Fig. 8}}}}}}

4183: %

4184: \bigskip\endinsert

4185: %

4186: 

4187: To determine the orientation, we compute $\gamma(\theta)$

4188: for large worldsheet time $\sigma$, where $\gamma\equiv |x_1|/|x_3|$,

4189: as before.

4190: The result is presented in Fig. 11.

4191: It is essentially equivalent to the one in Fig. 7,

4192: as is consistent with the signs in \oldec.

4193: \midinsert\bigskip{\vbox{{\epsfxsize=3in

4194:         \nobreak

4195:     \centerline{\epsfbox{orientn4.eps}}

4196:         \nobreak\bigskip

4197:     {\raggedright\it \vbox{

4198: {\bf Fig 11.}

4199: {\it Orientation of the wings ($n=4$; $w=e^{0.9\pi i/4}$) is determined by $\gamma(\theta)$.

4200: Vertical axis is $\gamma\equiv|x_1|/|x_3|$.

4201: Horizontal axis is $\theta/\pi$.}}}}}}

4202: %

4203: \bigskip\endinsert

4204: %

4205: This   is not yet the end of the story, as we will see

4206: shortly.

4207: As $\arg(w)$ decreases past $\arg(w)\sim 0.6\pi/4$, the

4208: form of the graphs in Fig. 8 qualitatively changes.

4209: Fig. 12 provides an illustration for $\arg(w)=0.45\pi/4$, and

4210: Figs. 13, 14 give the close-up views.

4211: %

4212: %

4213: \midinsert\bigskip{\vbox{{\epsfxsize=3in

4214:         \nobreak

4215:     \centerline{\epsfbox{n4argsd.eps}}

4216:         \nobreak\bigskip

4217:     {\raggedright\it \vbox{

4218: {\bf Fig 12.}

4219: {\it Solution of soliton equation at $\epsilon=0.005$, $\sigma=8$, $w=e^{0.45\pi i/4}$.

4220: Horizontal axis is $\theta/\pi$. 

4221: Vertical axis is $\arg(x_i(\sigma))/\pi$. Red line is $i=1$. Green line $i=3$.}}}}}}

4222: % 

4223: %

4224: \bigskip\endinsert

4225: \midinsert\bigskip{\vbox{{\epsfxsize=3in

4226:         \nobreak

4227:     \centerline{\epsfbox{n4argse.eps}}

4228:         \nobreak\bigskip

4229:     {\raggedright\it \vbox{

4230: {\bf Fig 13.}

4231: {\it Close-up view of Fig. 12}}}}}}

4232: %

4233: \bigskip\endinsert

4234: \midinsert\bigskip{\vbox{{\epsfxsize=3in

4235:         \nobreak

4236:     \centerline{\epsfbox{n4argsf.eps}}

4237:         \nobreak\bigskip

4238:     {\raggedright\it \vbox{

4239: {\bf Fig 14.}

4240: {\it Another close-up view of Fig. 12}}}}}}

4241: %

4242: \bigskip\endinsert

4243: %

4244: %

4245: %

4246: Figs. 12--14 seem to suggests that the propeller surface develops two

4247: extra wings.

4248: The integral \gcc\ over these wings can cancel, 

4249: so Figs. 12--14 are still consistent with the expression for the central charge 

4250: in \oldec.

4251: Yet, they are also consistent with $\langle1|c\rangle\sim \hat I_2+2\hat I_1-3\hat I_0$\foot{

4252: Note that as $\arg(w)\ra 0$, the wings which give rise to $\hat I_1$ shrink

4253: but do not disappear until $w$ crosses the $\arg(w)=0$ line.

4254: On the contrary, the new wings, which may define extra $2 I_0$ disappear as $\arg(w)$ is

4255: increased over $\sim0.6\pi/4$.}

4256: To resolve this ambiguity, we look again at the $\gamma(\theta)$

4257: graph which determines the orientation.

4258: Corresponding graphs are presented in Figs. 15--17.

4259: \midinsert\bigskip{\vbox{{\epsfxsize=3in

4260:         \nobreak

4261:     \centerline{\epsfbox{orientn4a.eps}}

4262:         \nobreak\bigskip

4263:     {\raggedright\it \vbox{

4264: {\bf Fig 15.}

4265: {\it Orientation of the wings ($n=4$; $w=e^{0.45\pi i/4}$) is determined by $\gamma(\theta)$.

4266: Vertical axis is $\gamma\equiv|x_1|/|x_3|$.

4267: Horizontal axis is $\theta/\pi$.}}}}}}

4268: \bigskip\endinsert

4269: %

4270: \midinsert\bigskip{\vbox{{\epsfxsize=3in

4271:         \nobreak

4272:     \centerline{\epsfbox{orientn4c.eps}}

4273:         \nobreak\bigskip

4274:     {\raggedright\it \vbox{

4275: {\bf Fig 16.}

4276: {\it Close-up view of Fig. 15.}}}}}}

4277: \bigskip\endinsert

4278: %

4279: \midinsert\bigskip{\vbox{{\epsfxsize=3in

4280:         \nobreak

4281:     \centerline{\epsfbox{orientn4b.eps}}

4282:         \nobreak\bigskip

4283:     {\raggedright\it \vbox{

4284: {\bf Fig 17.}

4285: {\it Orientation of the wings ($n=4$; $w=e^{0.45\pi i/4}$) is determined by $\gamma(\theta)$.

4286: Vertical axis is $\gamma\equiv|x_1|/|x_3|$.

4287: Horizontal axis is $\theta/\pi$.}}}}}}

4288: \bigskip\endinsert

4289: %

4290: From Figs. 15--17 we immediately see that

4291: the orientation of the four wings which appeared in 

4292: Fig. 6 does not change as $\arg(w)$ is decreased.

4293: Therefore, we must associate 

4294: $\langle c_1|1\rangle \sim \hat I_2+\hat I_0-2 \hat I_1$, as before.

4295: The integrals over the new wings must cancel each other,

4296: instead of giving an extra $\hat I_0$.

4297: A similar situation happens for $\arg(w)<0$.

4298: 

4299: To summarize, in the sector $-\pi/4\leq\arg(w)\leq\pi/4$ we have

4300: [compare with \oldec]

4301: \eqn\cbsf{

4302: \eqalign{  0<\arg(w)<{\pi\over4}:\quad \langle c_1|1\rangle&\sim

4303:                                    \hat I_2+\hat I_0-2 \hat I_1 \cr

4304: -{\pi\over4}<\arg(w)<0:\quad \langle c_1|1\rangle&\sim

4305:                                    \hat I_0+\hat I_2-2 \hat I_3 \cr}

4306: }

4307: This implies

4308: \eqn\cbfbfs{

4309: \eqalign{  0<\arg(w)<{\pi\over4}:\quad c_2=e_0; c_1=e_1 \cr

4310: -{\pi\over4}<\arg(w)<0:\quad c_2=e_0; c_1=e_3           \cr}

4311: }

4312: 

4313: 

4314: Consider now the critical point labeled by $\nu=1$,

4315: giving rise to the coulomb branch brane

4316: denoted by $c_2$.

4317: The choice of coordinates which diagonalizes \spexp\ is

4318: now given by

4319: \eqn\diagno{  u=e^{{i\varphi_w\over2}+{i\pi\over4}}\left[

4320:    {1\over 2}\left( {3\over\sqrt{2}}-2\right)^{1/2}\delta x_3 

4321:    -{i\over2}\left({3\over\sqrt{2}}+2\right)^{1/2} \delta x_1 \right] }

4322: and

4323: \eqn\diagnov{  v=e^{{i\varphi_w\over2}+{i\pi\over4}}\left[

4324:    {1\over 2}\left( {3\over\sqrt{2}}+2\right)^{1/2}\delta x_3 

4325:    -{i\over2}\left({3\over\sqrt{2}}-2\right)^{1/2} \delta x_1 \right] }

4326: which corresponds to

4327: \eqn\wfourcpp{W=W_1+\sqrt{2} |w| (u^2+v^2)    }

4328: The wavefront near the critical point is

4329: \eqn\wffourb{    

4330: \eqalign{   x_1&={\sqrt{-w}\over2}e^{\pi i\over4}+

4331:   {\epsilon\over\sqrt{|w|}} e^{-{i \varphi_w\over2}-{i\pi\over 4}}

4332:     \left[  {i\over 2}\left( {3\over\sqrt{2}}+2\right)^{1/2}\cos\theta 

4333:    -{i\over2}\left({3\over\sqrt{2}}-2\right)^{1/2} \sin\theta \right]       \cr

4334:              x_3&={\sqrt{-w}\over2}e^{3\pi i\over4}+

4335:   {\epsilon\over\sqrt{|w|}} e^{-{i \varphi_w\over2}-{i\pi\over 4}}

4336:  \left[  -{1\over 2}\left( {3\over\sqrt{2}}-2\right)^{1/2}\cos\theta 

4337:    +{1\over2}\left({3\over\sqrt{2}}+2\right)^{1/2} \sin\theta \right]  \cr}

4338: }

4339: In the angular sector $-\pi/4\leq\arg(w)\leq\pi/4$, the propeller

4340: surface has four wings, and the integral \gcc\ reduces to

4341: $\langle c_2|1\rangle\sim \hat I_1+\hat I_3-2\hat I_0$.

4342: This is the unique exponentially decaying solution in this

4343: sector.

4344: This result is consistent with the fact that the critical

4345: value of $W$ at the $\nu=1$ critical point is positive,

4346: so we expect the corresponding central charge to decay

4347: exponentially.

4348: 

4349: 

4350: \listrefs

4351: 

4352: \end

4353: --------------040708060406070809070807--