1: \documentclass[12pt]{article}

2: \usepackage{epsfig}

3: %\input epsf

4: \usepackage{amsmath,amssymb,amsthm,amscd}

5: \usepackage{rotating}

6: 

7: % My default margin widths and so on unless overridden in the latex

8: %file

9: \setlength{\oddsidemargin}{0.25in}      % 1.25in left margin

10: \setlength{\evensidemargin}{0.25in}     % 1.25in left margin (even pages)

11: \setlength{\topmargin}{0.5in}           % 1in top margin

12: \setlength{\textwidth}{6.0in}           % 6.0in text - 1.25in rt  margin

13: \setlength{\textheight}{9in}            % Body ht for 1in margins

14: \addtolength{\topmargin}{-\headheight}  % No header, so compensate

15: \addtolength{\topmargin}{-\headsep}     % for header height and  separation

16: \setlength{\marginparwidth}{0.75in}

17: %\setlength{\marginparsep}(0.05 in}

18: % For my home printer

19: % \addtolength{\topmargin}{0.5 in}

20: 

21: 

22: 

23: \let\ex=\times          %\def\ShowPolyData{} %\def\ShowOtherConData{}

24: \def\IP{{\mathbb P}}

25: \def\re{{\rm e}}

26: \def\ZZ{{\mathbb Z}}

27: \def\IR{{\mathbb R}}

28: \def\IC{{\mathbb C}}

29: \def\bT{{\mathbb T}}

30: \def\IQ{{\mathbb Q}}

31: \def\bC{{\mathbb C}}

32: \def\IF{{\mathbb F}}

33: \def\WP{{\mathrm W\mathbb P}}

34: \let\D=\Delta \let\s=\sigma \let\Th=\Theta \let\S=\Sigma

35: \def\2{{1\over2}}

36: \let\<=\langle \let\>=\rangle

37: \def\new#1\endnew{{\bf #1}}

38: \def\ifundefined#1{\expandafter\ifx\csname#1\endcsname\relax}

39: \ifundefined{draftmode}\else    \input draftmode        \fi

40: \let\Msize=\footnotesize        \def\VL{\;\vrule\;}     \def\v{\nu^*}

41: \def\BM{\Msize\begin{matrix}}           \def\EM{\end{matrix}}

42: \def\MN M:#1 #2 N:#3 #4 {{(#1_{#2},#3_{#4})}}

43: \def\MNH M:#1 #2 N:#3 #4 H:#5,#6 [#7]{{(#1_{#2},#3_{#4})^{#5,#6}_{#7}}}

44: \newcommand{\um}{\phantom{-}}

45: \newcommand{\ds}{\displaystyle}

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

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

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

49: \def\dd{\mathrm{d}}

50: \def\cI{\mathcal{I}}

51: \def\sO{\mathscr{O}}

52: 

53: 

54: \DeclareMathOperator{\rank}{rk} \DeclareMathOperator{\ch}{c}

55: \def\CY{Calabi--Yau }

56: \def\Vol{\mathrm{Vol}}

57: \def\conv{\mathrm{conv}}

58: \def\r{\mathrm{r}}

59: \newcommand{\ri}{{\rm i}}

60: \newcommand{\tr}{{\rm Tr}}

61: \newcommand{\zb}{\bar z}

62: \newcommand{\refb}[1]{(\ref{#1})}

63: \newcommand{\p}{\partial}

64: \newcommand{\half}{{1\over 2}}

65: \newcommand{\sectiono}[1]{\section{#1}\setcounter{equation}{0}}

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

67: 

68: 

69: %%Useful symbols%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

70: \def\ket{\rangle}

71: \def\bra{\langle}

72: \def\CA{{\cal A}}

73: \def\CB{{\cal B}}

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

75: \def\CD{{\cal D}}

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

77: \def\CF{{\cal F}}

78: \def\CG{{\cal G}}

79: \def\CH{{\cal H}}

80: \def\CI{{\cal I}}

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

82: \def\CK{{\cal K}}

83: \def\CL{{\cal L}}

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

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

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

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

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

89: \def\CR{{\cal R}}

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

91: \def\CT{{\cal T}}

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

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

94: \def\CW{{\cal W}}

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

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

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

98: 

99: %macros

100: \newcommand{\todo}[1]{{\em \small {#1}}\marginpar{$\Longleftarrow$}}

101: \newcommand{\labell}[1]{\label{#1}}

102: \newcommand{\bbibitem}[1]{\bibitem{#1}\marginpar{#1}}

103: \newcommand{\llabel}[1]{\label{#1}\marginpar{#1}}

104: \newcommand{\dslash}[0]{\slash{\hspace{-0.23cm}}\partial}

105: 

106: % macros for the conical defect paper

107: \newcommand{\sphere}[0]{{\rm S}^3}

108: \newcommand{\su}[0]{{\rm SU(2)}}

109: \newcommand{\so}[0]{{\rm SO(4)}}

110: %\newcommand{\sl}[0]{{\rm SL(2,R)}}

111: \newcommand{\bK}[0]{{\bf K}}

112: \newcommand{\bL}[0]{{\bf L}}

113: \newcommand{\bR}[0]{{\bf R}}

114: \newcommand{\tK}[0]{\tilde{K}}

115: \newcommand{\tL}[0]{\bar{L}}

116: \newcommand{\tR}[0]{\tilde{R}}

117: 

118: 

119: \newcommand{\btzm}[0]{BTZ$_{\rm M}$}

120: \newcommand{\ads}[1]{{\rm AdS}_{#1}}

121: \newcommand{\eds}[1]{{\rm EdS}_{#1}}

122: \newcommand{\sph}[1]{{\rm S}^{#1}}

123: \newcommand{\cosm}[0]{R}

124: \newcommand{\hdim}[0]{\bar{h}}

125: \newcommand{\bw}[0]{\bar{w}}

126: \newcommand{\bz}[0]{\bar{z}}

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

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

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

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

131: \newcommand{\pat}{\partial}

132: \newcommand{\lp}{\lambda_+}

133: \newcommand{\bx}{ {\bf x}}

134: \newcommand{\bk}{{\bf k}}

135: \newcommand{\bb}{{\bf b}}

136: \newcommand{\BB}{{\bf B}}

137: \newcommand{\tp}{\tilde{\phi}}

138: \hyphenation{Min-kow-ski}

139: 

140: 

141: %%Commonly used constants and symbols%%%%%%%%%%%%%%%%%%%%%%%%%

142: \def\apr{\alpha'}

143: \def\a{\alpha}

144: \def\str{{str}}

145: \def\lstr{\ell_\str}

146: \def\gstr{g_\str}

147: \def\Mstr{M_\str}

148: \def\lpl{\ell_{pl}}

149: \def\Mpl{M_{pl}}

150: \def\varep{\varepsilon}

151: \def\del{\nabla}

152: \def\grad{\nabla}

153: \def\tr{\hbox{tr}}

154: \def\perp{\bot}

155: \def\half{\frac{1}{2}}

156: \def\p{\partial}

157: \def\perp{\bot}

158: \def\eps{\epsilon}

159: 

160: 

161: 

162: \begin{document}

163: \begin{titlepage}

164: {}~ \hfill\vbox{ \hbox{} }\break

165: 

166: 

167: \rightline{hep-th/0612125} \rightline{MAD-TH-06-12} \vskip 1cm

168: 

169: \centerline{\Large \bf Topological string theory on compact

170: Calabi-Yau:} \vskip 0.2 cm \centerline{\Large \bf modularity and

171: boundary conditions} \vskip 0.5 cm

172: 

173: %\centerline{\Large \bf Draft:} \vskip 0.2 cm \centerline{\Large \bf preliminary

174: %and unfinished} \vskip 0.5 cm

175: 

176: \long\def\symbolfootnote[#1]#2{\begingroup%

177: \def\thefootnote{\fnsymbol{footnote}}\footnote[#1]{#2}\endgroup}

178: 

179: %\renewcommand{\thefootnote}{\fnsymbol{footnote}}

180: \vskip 30pt \centerline{ {\large \rm Min-xin Huang

181: \symbolfootnote[1]{minxin@physics.wisc.edu},

182: Albrecht Klemm\symbolfootnote[7]{aklemm@physics.wisc.edu} and

183: Seth Quackenbush\symbolfootnote[2]{squackenbush@wisc.edu} }} \vskip .5cm \vskip

184: 30pt

185: \centerline{$^{\star}$ $^{\star\star}$ $^\dagger${\it Department of Physics, University of

186: Wisconsin}}

187: \centerline{$^{\star\star}${\it Department of Mathematics, University of

188: Wisconsin}}

189: \centerline{\it Madison, WI 53706, U.S.A.}

190: 

191: 

192: \setcounter{footnote}{0}

193: %\renewcommand{\thefootnote}{\arabic{footnote}}

194: \vskip 100pt

195: \begin{abstract}

196: The topological string partition function $Z(\lambda,t,\bar t)$

197: $=\exp(\lambda^{2g-2} F_g(t,\bar t))$ is calculated on a compact Calabi-Yau $M$. 

198: The $F_g(t,\bar t)$ fulfill the holomorphic anomaly

199: equations, which imply that $\Psi=Z$ transforms as a wave function

200: on the symplectic space $H^3(M,\mathbb{Z})$. This defines it

201: everywhere in the moduli space ${\cal M}(M)$ along with preferred

202: local coordinates. Modular properties of the sections $F_g$ as

203: well as local constraints from the 4d effective action allow us to

204: fix $Z$ to a large extent. Currently with a newly found 

205: gap condition at the conifold, regularity at the orbifold and the most naive bounds

206: from Castelnuovo's theory, we can provide the boundary data, which

207: specify $Z$, e.g. up to genus 51 for the quintic.

208: \end{abstract}

209: 

210: \end{titlepage}

211: \vfill \eject

212: 

213: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

214: 

215: \newpage

216: 

217: 

218: \baselineskip=16pt

219: 

220: \tableofcontents

221: 

222: 

223: \section{Outline}

224: 

225: Coupling topological matter to topological gravity is a key

226: problem in string theory. Conceptually most relevant is the

227: topological matter sector of the critical string as it arises e.g. in

228: Calabi-Yau compactifications. Topological string theory on

229: non-compact Calabi-Yau manifolds such as ${\cal O}(-3)\rightarrow

230: \mathbb{P}^2$ is essentially solved either by localization -

231: \cite{Klemm:1999gm} or large N-techniques~\cite{Aganagic:2003db}

232: and has intriguing connections to Chern-Simons

233: theory~\cite{Witten:1992fb}, open-closed string

234: duality~\cite{Gopakumar:1998ki}, matrix

235: models~\cite{Dijkgraaf:2002fc}, integrable hierarchies of

236: non-critical string theory~\cite{Aganagic:2003qj} and 2d

237: Yang-Mills theory~\cite{Vafa:2004qa}.

238: 

239: However, while local Calabi-Yau manifolds are suitable to study

240: gauge theories and more exotic field theories in 4d and specific

241: couplings to gravity, none of the techniques above extends to

242: compact Calabi-Yau spaces, which are relevant for important

243: questions in 4d quantum gravity concerning e.g. the properties of

244: 4d black holes~\cite{Ooguri:2004zv} and the wave function in mini

245: superspace~\cite{Ooguri:2005vr}.

246: 

247: 

248: Moreover, while the genus dependence is encoded in the

249: Chern-Simons and matrix model approaches in a superior fashion by

250: the ${1\over N^2}$-expansion, the moduli dependence on the

251: parameter $t$ is reconstructed locally and in a holomorphic limit,

252: typically by sums over partitions. This yields an algorithm, which

253: grows exponentially in the world-sheet degree or the

254: space-time instanton number.

255: 

256: As the total $F_g(t,\bar t)$ are modular invariant sections over

257: the moduli space ${\cal M}(M)$, they must be generated by a ring

258: of almost holomorphic modular forms.  This solves the dependence

259: on the moduli in the most effective way. In the following we will

260: show that space-time modularity, the holomorphic anomaly equations

261: of Bershadsky, Cecotti, Ooguri and Vafa, as well as boundary

262: conditions at various boundary components of the moduli space,

263: solve the theory very efficiently.

264: 

265: For compact (and non-compact) Calabi-Yau spaces mirror symmetry is

266: proven at genus zero. The modular properties that we need are also

267: established  at genus zero. Moreover it has been argued recently

268: that the  holomorphic anomaly recursions follow from categorical

269: mirror symmetry \cite{costello,kontsevich}. To establish  mirror

270: symmetry at higher genus, one needs merely to prove that the same

271: boundary data fix the $F_g(t,\bar t)$ in the $A$- and the

272: $B$-model.

273: 

274: 

275: \subsection{Extending the Seiberg-Witten approach to gravity}

276: Seiberg-Witten reconstructed the non-perturbative N=2 gauge gauge

277: coupling from meromorphic section over ${\cal M}(M)$ using their

278: modular properties and certain local  data from the effective

279: action at singular divisors of ${\cal M}(M)$. In~\cite{Huang}~we

280: reconsidered the problem of topological string on local Calabi-Yau

281: from the modular point of view and found that the singular

282: behaviour of the gravitational couplings is restrictive enough to

283: reconstruct them globally. This can be viewed as the most straightforward 

284: extension of the  Seiberg-Witten approach to gravitational

285: couplings.

286: 

287: Note that the problem of instanton counting in these cases is

288: solved either by geometric engineering, one of the techniques

289: mentioned above, or more directly by the localisation techniques

290: in the moduli space of gauge theory instantons by Nekrasov,

291: Nakajima et. al. It is nevertheless instructive to outline the

292: general idea in this simple setting\footnote{Maybe the simplest

293: example of the relation between modularity and the holomorphic

294: anomaly equations is provided by Hurwitz theory on elliptic curves

295: \cite{dijkgraafzagier}.}. We focused on the N=2 $SU(2)$

296: Seiberg-Witten case, but the features hold for any local

297: Calabi-Yau whose mirror is an elliptic curve with a meromorphic

298: differential~\cite{Aganagic:2006wq,Huang2}\footnote{With fairly

299: obvious generalizations for the cases where the mirror is a higher

300: genus curve. In this case the traditional modular forms of

301: subgroups $\Gamma$ of $\Gamma_0:=SL(2,\mathbb{Z})$ have to be

302: replaced by Siegel modular forms of subgroups of

303: $Sp(2g,\mathbb{Z})$~\cite{Aganagic:2006wq}.} and are as follows

304: \begin{itemize}

305: \item The genus $g$ topological string partition functions are given by

306: \begin{equation}

307: F^{(g)}(\tau, \bar \tau)=\xi^{2g-2} \sum_{k=0}^{3(g-1)}

308: \hat E_2^{k}(\tau,\bar \tau) c^{(g)}_k(\tau)\ .

309: \label{eq:generallocalform}

310: \end{equation}

311: Here $\hat E_2(\tau,\bar \tau):= E_2(\tau) + {6 i \over \pi(\bar \tau-\tau)}$ is the

312: modular invariant anholomorphic extension of the second Eisenstein series

313: $E_2(\tau)$ and the holomorphic `Yukawa coupling'

314: $\xi:=C^{(0)}_{ttt}={\partial \tau \over \partial t}$ is an

315: object of weight $-3$ under the modular $\Gamma\in \Gamma_0=SL(2,\mathbb{Z})$.

316: For example for pure $N=2$ $SU(2)$ gauge theory $\Gamma=\Gamma(2)$~\cite{Seiberg:1994rs}.

317: Modular invariance implies then that $c^{(g)}_k(\tau)$ are modular forms of

318: $\Gamma$ of weight $6(g-1)-2k$.

319: \item The simple anti-holomorphic dependence of (\ref{eq:generallocalform}) implies

320: that the only part in $F^{(g)}(\tau, \bar \tau)$ not fixed by the recursive anomaly

321: equations is the weight $6(g-1)$ holomorphic forms $c^{(k)}_0(\tau)$, which

322: are finitely generated as a weighted polynomial $c^{(g)}_0(\tau)=p_{6(g-2)}(k_1,\ldots,k_m)$

323: in the holomorphic generators $G_{k_1},\ldots,G_{k_m}$ of forms of $\Gamma$.

324: \item The finite data needed to fix the coefficients in $p_{6(g-2)}(k_1,\ldots,k_m)$

325: are provided in part by the specific leading behaviour of the $F^{(g)}$ at the conifold divisor

326: \begin{eqnarray} \label{thegap}

327: F^{(g)}_{\textrm{conifold}}=\frac{(-1)^{g-1}B_{2g}}{2g(2g-2)t_D^{2g-2}}+\mathcal{O}(t_D^0),

328: \end{eqnarray}

329: in special local coordinates $t_D$. The order of the leading term

330: was established in~\cite{BCOV}, the coefficient of the leading

331: term in \cite{Ghoshal:1995}, and the `gap condition,' i.e. the

332: vanishing of the following $2g-3$ negative powers in $t_D$

333: in~\cite{Huang}. This property in particular carries over to the

334: compact case and we can give indeed a string theoretic explanation

335: of the finding in~\cite{Huang}.

336: \item Further conditions are provided by the regularity of the

337: $F^{(g)}$ at orbifold points in ${\cal M}(M)$. These conditions

338: unfortunately turn out to be somewhat weaker in the global case

339: than in the local case.

340: \end{itemize}

341: 

342: Similar forms as (\ref{eq:generallocalform}) for the $F^{(g)}$ 

343: appear in the context of Hurwitz theory on elliptic 

344: curves~\cite{dijkgraafzagier}, of mirror symmetry in $K3$ 

345: fibre limits\cite{Marino:1998} and on rational 

346: complex surfaces~\cite{Hosono:1999qc,Hosono:2001gf}.

347: In the local cases, which have elliptic curves as mirror geometry,

348: we found~\cite{Huang,Huang2}, that the above conditions

349: (over)determine the unknowns in $p_{6(g-2)}(k_1,\ldots,k_m)$ and

350: solve the theory. This holds also for the gauge theories with

351: matter, which from geometric engineering point of view correspond

352: to local Calabi-Yau manifolds with several (K\"ahler)

353: moduli~\cite{Huang2}. Using the precise anholomorphic

354: dependence and restrictions from space-time modularity one can

355: iterate the holomorphic anomaly equation with an algorithm which

356: is exact in the moduli dependence and grows polynomially in

357: complexity with the genus.

358: 

359: Here we extend this approach further to compact Calabi-Yau spaces

360: and focus on the class of one K\"ahler moduli Calabi-Yau spaces

361: $M$ such as the quintic. More precisely we treat the class of one

362: modulus cases whose mirror $W$ has, parameterized by a suitable

363: single cover variable, a Picard-Fuchs system  with exactly three

364: regular singular points: The point of maximal monodromy, a

365: conifold point, and a point with rational branching. The latter can

366: be simply a $\mathbb{Z}_d$ orbifold point. This e.g. is the case

367: for the hypersurfaces where the string theory has an exact

368: conformal field theory description at this point in terms of an

369: orbifold of a tensor product of minimal $(2,2)$ SCFT field

370: theories, the so called the Gepner-model. For some complete

371: intersections there are massless BPS particles at the branch

372: locus, which lead in addition to logarithmic singularities.

373: 

374: We find a natural family of coordinates in which the conifold

375: expansion as well as the rational branched logarithmic

376: singularities exhibit the gap condition (\ref{thegap}). Despite

377: the fact that the modular group, in this case a subgroup of ${\rm

378: Sp}(h^3,\mathbb{Z})$, is poorly understood\footnote{Subgroups of

379: ${\rm Sp}(4,\mathbb{Z})$ in which the monodromy group of the

380: one-parameter models live have been recently specified

381: \cite{cyy}.}, we will see that the essential feature carry over to

382: the compact case. Modular properties, the ``gap condition'',

383: together with regularity at the orbifold, the leading behaviour

384: of the $F_g$ at large radius, and Castelenovo's Bound 

385: determine topological string on one modulus 

386: Calabi-Yau to a large extent.

387: 

388: 

389: 

390: \section{The topological B-model}

391: %---------------------------------------------------------

392: In this section we give a quick summary of the approach of

393: \cite{BCOVI,BCOV} to the topological B-model, focusing as fast as

394: possible on the key problems that need to be overcome: namely the

395: problem of integrating the anomaly equation efficiently and

396: the problem of fixing the boundary conditions.

397: 

398: 

399: \subsection{The holomorphic anomaly equations}

400: \label{holomorphicanomaly}

401: The definition of

402: $F^{(g)}$ is $F^{(g)}=\int_{ {\cal M}_g} \mu_g$ with measure on ${\cal M}_g$

403: \begin{equation}

404: \mu_g=\prod_{i=1}^{3g-3} {\rm d} m_i {\rm d} {\bar m}_{\bar \imath}

405: \left\langle \prod_{i,\bar \imath}

406: \int_\Sigma G_{zz} \mu^{(i)\, z}_{\bar z} {\rm d}^2 z

407: \int_\Sigma G_{\bar z \bar z} \mu^{(i)\, \bar z}_{z} {\rm d}^2 z \right\rangle \ .

408: \end{equation}

409: Here the Beltrami differentials $\mu^{(i)\, z}_{\bar z} {\rm d}

410: {\bar z}$ span $H^1(\Sigma,T\Sigma)$, the tangent space to ${\cal

411: M}_g$. The construction of the measure $\mu_g$ is strikingly

412: similar to the one for the bosonic string, once the BRST partner

413: of the energy-momentum tensor is identified with the

414: superconformal current $G_{zz}{\rm d } z$ and the ghost number

415: with the $U(1)$ charge \cite{MB}. $\left\langle \right\rangle$ is

416: to be evaluated in the internal $(2,2)$ SCFT, but it is easy to

417: see that it gets only contributions from the topological $(c,c)$

418: sector.

419: 

420: The holomorphic anomaly equation reads for $g=1$  \cite{BCOVI}

421: \begin{equation}

422: {\bar \partial}_{\bar k} \partial_m F^{(1)}={1\over 2}

423: {\bar C}_{\bar k}^{ij} C^{(0)}_{mij} + \left( {\chi\over 24} -1\right )

424: G_{\bar km}\ ,

425: \label{eq:anomalyg1}\

426: \end{equation}

427: where $\chi$ is the Euler number of the target space $M$, and for

428: $g>1$ \cite{BCOV}

429: \begin{equation}

430: {\bar \partial}_{\bar k} F^{(g)}={1\over 2}{\bar C}_{\bar k}^{ij}

431: \left(D_i D_j F^{(g-1)}+ \sum_{r=1}^{g-1}D_i F^{(r)}  D_j F^{(g-r)}\right)\ .

432: \label{eq:anomaly}

433: \end{equation}

434: The right hand side of the equations comes from the complex

435: co-dimension one boundary of the moduli space of the worldsheet

436: ${\cal M}_g$, which corresponds to pinching of handles. The key idea is  that $\bar \partial_{\bar

437: k} F^{(g)}=\int_{{\cal M}_g} \bar \partial \partial \lambda_g$,

438: where $\bar \partial \partial$ are derivatives on ${\cal M}_g$ so

439: that $\bar \partial_{\bar k} F^{(g)}=\int_{\partial {\cal M}_g}

440: \lambda_g$. The contribution to the latter integral is  from the

441: co-dimension one boundary $\partial {\cal M}_g$.

442: 

443: The first equation (\ref{eq:anomalyg1}) can be integrated using

444: special geometry up to a holomorphic function \cite{BCOVI}, which

445: is fixed by the consideration in Sect. \ref{boundaryconditions}.

446: 

447: The equations  (\ref{eq:anomaly}) are solved in BCOV using the

448: fact that due to

449: \begin{equation}

450: \bar D_{\bar i} \bar C_{\bar \jmath \bar k \bar l}=\bar D_{\bar \jmath} \bar C_{\bar

451: \imath \bar k \bar l}

452: \end{equation}

453: one can integrate

454: \begin{equation}

455: \bar C_{\bar \jmath \bar k \bar l}=e^{-2 K} \bar D_{\bar i}\bar D_{\bar \jmath}

456: \bar \partial_{\bar k} S

457: \end{equation}

458: as

459: \begin{equation}

460: S_{\bar \imath} =\bar \partial_{\bar \imath} S,\ \ \

461: S^j_{\bar \imath} =\bar \partial_{\bar \imath} S^j,\ \ \

462: \bar C_{\bar k}^{ij}= \partial_{\bar k} S^{ij}\ .

463: \label{eq:propagators}

464: \end{equation}

465: The idea is to write the right hand side of (\ref{eq:anomaly}) as

466: a derivative w.r.t. $\bar \partial_{\bar k}$. In the first step

467: one writes

468: \begin{equation}

469: \begin{array}{rl}

470: {\bar \partial}_{\bar k} F^{(g)}=&\ds{{\bar \partial}_{\bar k}

471: \left({1\over 2}S^{ij}

472: \left(D_i D_j F^{(g-1)}+ \sum_{r=1}^{g-1}D_i F^{(r)}  D_j F^{(g-r)}\right)\right)} \\[ 3 mm]

473: &\ds{-{1\over 2} S^{ij}{\bar \partial}_{\bar k}\left(D_i D_j F^{(g-1)}+

474: \sum_{r=1}^{g-1}D_i F^{(r)}  D_j F^{(g-r)}\right)\ .}

475: \end{array}

476: \label{int1:anomaly}

477: \end{equation}

478: With the commutator $R^l_{i\bar k j}=-\bar \partial_{\bar k}

479: \Gamma_{ij}^l=[D_i,\partial_{\bar k}]^l_j= G_{i\bar

480: k}\delta^l_j+G_{j\bar k}\delta^l_i-C^{(0)}_{ijm} {\bar C}_{\bar

481: k}^{ml}$ the  second term can be rewritten so that the $\bar

482: \partial_{\bar k}$ derivative acts in all terms directly on

483: $F^{(g)}$. Then using (\ref{eq:anomalyg1}, \ref{eq:anomaly}) with

484: $g'<g$ one can iterate the procedure, which produces an equation

485: of the form

486: \begin{equation}

487: {\bar \partial}_{\bar k} F^{(g)}=\bar \partial_{\bar k}

488: \Gamma^{(g)}(S^{ij},S^i,S, C^{(<g)}_{i_1,\ldots,i_n})\ ,

489: \end{equation}

490: where $\Gamma^{(g)}$ is a functional of  $S^{ij},S^i,S$ and $C^{(<g)}_{i_1,\ldots,i_n}$.

491: This implies that

492: \begin{equation}

493: F^{(g)}=\Gamma^{(g)}(S^{ij},S^i,S, C^{(<g)}_{i_1,\ldots,i_n})+c^{(g)}_0(t)\ ,

494: \label{fgformal}

495: \end{equation}

496: is a solution. Here $c^{(g)}_0(t)$ is the holomorphic ambiguity,

497: which is not fixed by the recursive procedure. It is holomorphic

498: in $t$ as well as modular invariant. The {\sl major conceptual

499: problem} of topological string theory on compact Calabi-Yau is to

500: find the {\sl boundary conditions} which fix $c^{(g)}_0(t)$. Note

501: that the problem is not well defined without the constraints from

502: modular invariance. Using the generalization of the gap condition

503: in Sect. \ref{boundaryconditions}, the behaviour of the orbifold

504: singularities in Sect. \ref{quinticorbifold} and Castelnuovo's

505: bound in Sect. \ref{quinticdbrane} we can achieve this goal to a

506: large extent.

507: 

508: 

509: Properties of the $\Gamma^{(g)}(S^{ij},S^i,S,

510: C^{(<g)}_{i_1,\ldots,i_n})$ are established using the auxiliary

511: action

512: \begin{equation}

513: Z=\int \dd x \dd \phi \exp(Y+\tilde W)

514: \label{auxilliaryaction}

515: \end{equation}

516: where

517: \begin{equation}

518: \begin{array}{rl}

519: \tilde W(\lambda,x,\phi,t,\bar t)=&\ds{\sum_{g=0}^\infty \sum_{m=0}^\infty

520: \sum_{n=0}^\infty {1\over m! n!} \tilde C^{(g)}_{i_1,\ldots,i_n,\phi^{m}}

521: x_{i_1}\ldots x_{i_n} \phi^m}\\

522: =&\ds{\sum_{g=0}^\infty  \sum_{n=0}^\infty

523: {\lambda^{2g -2}\over n!} C^{(g)}_{i_1,\ldots,i_n}x_{i_1}\ldots x_{i_n}(1-\phi)^{2-2g-n}}+\left({\chi\over 24}-1\right)

524: \log\left(1\over {1-\phi}\right) \ ,

525: \end{array}

526: \end{equation}

527: with $C^{(g)}_{i_1,\ldots,i_n}=D_{i_1}\ldots D_{i_n} F^{(g)}$  and

528: the ``kinetic term'' is given by

529: \begin{equation}

530: Y(\lambda , x, \phi;t,\bar t)=-{1\over 2 \lambda^2}

531: (\Delta_{ij}x^i x^j+ 2 \Delta_{i\phi}x^i \phi +\Delta_{\phi\phi}

532: \phi^2) + {1\over 2} \log\left(\det \Delta\over \lambda^2\right)\ .

533: \end{equation}

534: In \cite{BCOV} it was shown that $\exp(\tilde W)$ fulfills an

535: equation

536: \begin{equation}

537: {\partial \over \partial {\bar t_i}} \exp(\tilde W) =\left[ {\lambda ^2\over 2}

538: \bar C^{jk}_{\bar \imath} {\partial^2 \over \partial x^j \partial x^k} -

539: G_{\bar \imath j}  x^j {\partial \over \partial \phi} \right]\exp(\tilde W)\

540: \label{linearanomaly}

541: \end{equation}

542: that is equivalent to the holomorphic anomaly equations, by

543: checking the coefficients of the $\lambda$ powers, and $\exp(Y)$

544: fulfills

545: \begin{equation}

546: {\partial \over \partial {\bar t_{\bar \imath}}} \exp(Y) =\left[ -{\lambda ^2\over 2}

547: \bar C^{jk}_{\bar \imath} {\partial^2 \over \partial x^j \partial x^k} -

548: G_{\bar \imath j}  x^j {\partial \over \partial \phi} \right]\exp(Y) \

549:                 \end{equation}

550: implying that  $\Delta_{ij}$, $\Delta_{i\phi}$ and $\Delta_{\phi\phi}$ are the inverses

551: to the propagators $K^{ij}=-S^{ij}$, $K^{i\phi}:=-S^i$ and  $K^{\phi\phi}:=-2S$. A

552: saddle point expansion of $Z$ gives

553: \begin{equation}

554: \log(Z)=\sum_{g=2}^{\infty} \lambda^{2g-2}\left[F^{(g)}-

555: \Gamma^{(g)}(S^{ij},S^i,S, C^{(<g)}_{i_1,\ldots,i_n})\right]\ ,

556: \end{equation}

557: where  $\Gamma^{(g)}(S^{ij},S^i,S, C^{(<g)}_{i_1,\ldots,i_n})$ is

558: simply the Feynman graph expansion of the action

559: (\ref{auxilliaryaction}) with the vertices $\tilde

560: C^{(g)}_{i_1,\ldots,i_n,\phi^{m}}$ and the propagators above.

561: Moreover it can be easily shown that ${\partial \over

562: \partial_{\bar t_i}}Z=0$, which implies to all orders that

563: $F^{(g)}$ can be written as (\ref{fgformal}). This establishes the

564: reduction of the whole calculation to the determination of the

565: holomorphic modular invariant sections $c^{(g)}_0(t)\in {\cal

566: L}^{2g-2}$. However it also reflects the {\sl major technical

567: problem} in the approach of BCOV, namely that the procedure to

568: determine the recursive anholomorphic part {\sl grows

569: exponentially with the genus}. It has been observed in

570: \cite{Katz:1999} that in concrete cases the terms appearing in the

571: Feynman graph expansion are not functionally independent. This is a

572: hint for finitely generated  rings of anholomorphic modular

573: forms over ${\cal M}(M)$. Using the modular constraints

574: systematically in each integration step Yamaguchi and Yau

575: developed a recursive procedure for the quintic whose complexity grows

576: asymptotically only polynomially, see (\ref{growth}).

577: 

578: Since the B-model is 2d gravity coupled to 2d matter, let us

579: compare the situation with pure  2d gravity, where the objects of

580: interest are  correlation functions of $\tau_{d_i}=(2d_i+1)!!

581: c_1(L_i)^{d_i}$  which are forms on ${\overline {\cal M}_g}$

582: constructed from the descendent fields

583: \begin{equation}

584: F_g(t_0,t_1,\ldots )=\sum_{\{d_i\}} \langle \prod \tau_{d_i}\rangle_g

585: \prod_{r>0} {t_r^{n_r}\over n_r!}\ .

586: \end{equation}

587: Here $\{d_i\}$ are the set of all non-negative integers and  $n_r:={\rm Card}(i:d_i=r)$.

588: 

589: The linear second order differential equations (\ref{linearanomaly}) is the small

590: phase space analog of  the {\sl Virasoro constraints}

591: \begin{equation}

592: L_n Z=0, \qquad n\ge -1

593: \label{virasoroconstraints}

594: \end{equation}

595: on $Z=e^F$ with $F=\sum_{g=0}^\infty \lambda^{2 g-2} F_g$ the free

596: energy of 2d topological gravity~\cite{wittensurvey}. Indeed the

597: $L_n$ with $[L_n,L_m]=(n-m) L_{n+m}$ are second order linear

598: differential operators in the $t_i$. The {\sl non-linear KdV

599: Hierarchy}, which together with dilaton and string  equation are

600: equivalent to (\ref{virasoroconstraints})~\cite{wittensurvey}, and

601: correspond in the small phase space of the B-model to the

602: holomorphic anomaly equations

603: (\ref{eq:anomalyg1},\ref{eq:anomaly}). In the $A$-model approaches

604: to topological string on Calabi-Yau manifolds, such as relative

605: GW-theory, localisation or attempts to solve the theory via

606: massive $(2,2)$ models, the descendents are introduced according

607: to the details of the geometrical construction and then ``summed

608: away''.

609: 

610: The combinatorial cumbersome information in the descendent sums is

611: replaced in the B-model by the contraints from the modular group,

612: holomorphicity and boundary information from the effective 4d

613: action. As a consequence of this beautiful interplay between

614: space-time and world-sheet properties one needs only the small

615: phase space equations (\ref{eq:anomalyg1}, \ref{eq:anomaly},

616: \ref{linearanomaly}).

617: 

618: This approach requires the ability to relate various local

619: expansions of $F^{(g)}$ near the boundary of the moduli space.

620: Sensible local expansions (of terms in the effective action) are

621: in locally monodromy invariant coordinates. As explained

622: in~\cite{Aganagic:2006wq} these coordinates in various patches are

623: related by symplectic transformations on the phase space $H^3(M)$.

624: The latter extend as metaplectic transformations to the wave

625: function $\psi=Z$ of the topological string on the

626: Calabi-Yau~\cite{wittenwavefunction}, which defines the

627: transformation on the $F^{(g)}$. It will be important for us that

628: the real

629: polarisation~\cite{Verlinde:2004ck,Aganagic:2006wq,Gunaydin:2006bz}

630: defines an unique splitting (\ref{fgformal},\ref{fgdefinition}) of

631: local expansions of the $F^{(g)}$ in the anholomorphic modular

632: part determined by the anomaly equations and the holomorphic

633: modular part $c^{(g)}_0(t)$. Aspects of the wave functions

634: properties and the various polarizations have been further

635: discussed

636: in~\cite{Verlinde:2004ck,Aganagic:2006wq,Gunaydin:2006bz}.

637: 

638: 

639: 

640: 

641: 

642: %-------------------------------------------------------------

643: \subsection{Boundary conditions from light BPS states}

644: \label{boundaryconditions}

645: Boundaries in the moduli space ${\cal M}(M)$ correspond to degenerations

646: of the manifold $M$ and general properties of the effective action can be inferred

647: from the physics of the lightest states. More precisely the light states relevant

648: to the $F^{(g)}$ terms in the $N=2$ actions are the BPS  states.

649: Let us first discuss the boundary conditions for $F^{(1)}$ at the singular points

650: in  the moduli space.

651: \begin{itemize}

652: \item At the point of maximal unipotent monodromy in the mirror manifold

653: $W$, the K\"ahler areas, four, and  six volumes of the original

654: manifold $M$ are all large. Therefore the lightest string states

655: are the constant maps $\Sigma_g\rightarrow pt\in M$. For these

656: Kaluza-Klein reduction, i.e. a zero mode analysis of the A-twisted

657: non-linear $\sigma$-model is sufficient to calculate the leading

658: behaviour\footnote{The leading of $F^{(0)}$  at this point is

659: similarly calculated  and given in (\ref{prepotential}).} of

660: $F^{(1)}$ as~\cite{BCOVI}

661: \begin{equation}

662: F^{(1)}={t_i\over 24} \int c_2 \wedge J_i+ {\cal O}(e^{2 \pi i t})\ .

663: \label{F1}

664: \end{equation}

665: Here   $2 \pi i \,t_i={X^i\over X^0}$ are the canonical K\"ahler parameters,

666: $c_2$ is the second Chern class, and $J_i$ is the basis for the

667: K\"ahler cone dual to 2-cycles $C_i$ defining the $t_i:=\int_{C_i} \hat J= \int_{C_i} \sum_{i} t_i J_i$.

668: 

669: \item At the conifold divisor in the moduli space  ${\cal M}(W)$, $W$

670: develops a nodal singularity, i.e., a collapsing  cycle

671: with $S^3$ topology. As discussed in sect. \ref{quinticconifold} this

672: corresponds to the vanishing of the total volume of $M$. The leading

673: behaviour at this point is universally~\cite{Hosono:1994ax}

674: \begin{equation}

675: F^{(1)}= {1\over 12} \log(t_D)+O(t_D)\ .

676: \label{betaconifold}

677: \end{equation}

678: This leading behaviour has been physically explained as the effect of integrating out

679: a non-perturbative hypermultiplet, namely the  extremal black hole of~\cite{Strominger:1995cz}.

680: Its mass $\sim t_D$, see (\ref{massbh}), goes to zero at the conifold and it couples to

681: the $U(1)$ vector in the $N=2$ vectormultiplet, whose lowest component is the modulus $t_D$.

682: The factor $\frac{1}{12}$ comes from the gravitational one-loop $\beta$-function,  which describes the

683: running of the $U(1)$ coupling~\cite{Vafa:1995ta}.

684: A closely related  situation is the one of a shrinking lense space $S/G$.

685: As explained in \cite{GVbfg} one gets in this case several BPS hyper multiplets as the

686: bound states of wrapped D-branes, which modifies the  factor $\frac{1}{12}\rightarrow \frac{|G|}{12}$

687: in the one loop $\beta$-function (\ref{betaconifold}).

688: \item The gravitational $\beta$-function argument extends also to non-perturbative spectra arising at more

689: complicated singularities, e.g. with gauge symmetry enhancement and adjoint

690: matter~\cite{Klemm:1995kj}.

691: \end{itemize}

692: For the case of the one parameter families the above boundary information and the fact is

693: sufficient to fix the holopmorphic ambiguity in $F^{(1)}$.

694: 

695: To learn from  the effective action point of view about the higher

696: genus boundary behaviour, let us recall that the $F^{(g)}$ as in

697: $F(\lambda,t)=\sum_{g=1}^\infty \lambda^{2g-2} F^{(g)}(t)$ give

698: rise to the following term:

699: \begin{equation}

700: S^{N=2}_{1-loop}=\int {\rm d}^4 x R_+^2 F(\lambda ,t)\ ,

701: \label{N=2oneloop}

702: \end{equation}

703: where $R_+$ is the self-dual part of the curvature and we identify

704: $\lambda$ with $F_+$, the self-dual part of the graviphoton field

705: strength. As explained in \cite{GVI,GVII}, see \cite{MB} for a

706: review, the term is computed by a one-loop integral in a constant

707: graviphoton background, which depends only on the left

708: ($SO(4)=SU(2)_L\otimes SU(2)_R$) Lorentz quantum numbers of BPS

709: particles $P$ in the loop. The calculation is very similar to the

710: normal Schwinger-loop calculation. The latter computes the

711: one-loop effective action in an $U(1)$ gauge theory, which comes

712: from integrating out massive particles $P$ coupling to a constant

713: background $U(1)$ gauge field. For a self-dual background field

714: $F_{12}=F_{34}=F$ it leads to the following one-loop determinant

715: evaluation:

716: \begin{equation}

717: S^{S}_{1-loop}=\log\, \det \left(\nabla + m^2 + 2 e \, \sigma_L F\right)=

718: \int_{\epsilon}^\infty \frac{{\rm d} s}{s} \ds{{ {\rm Tr} (-1)^f \exp({- s m^2 }) \exp({-2 s e \sigma_L F})}\over

719: 4 \sin^2\ds{\left( s e F/2\right)}}\ .

720: \label{schwingerloop}

721: \end{equation}

722: Here the $(-1)^f$ takes care of the sign of the log of the

723: determinant depending on whether $P$ is a boson or a fermion, and

724: $\sigma_L$ is the Cartan element in the left Lorentz

725: representation of $P$. To apply this calculation to the $N=2$

726: supergravity case one notes, that the graviphoton field couples

727: to the mass, i.e., we have to identify $e=m$. The loop has two

728: $R_+$ insertions and an arbitrary, (for the closed string action

729: even) number, of graviphoton insertions. It turns out \cite{GVII}

730: that the only supersymmetric BPS states with the Lorentz quantum

731: numbers

732: \begin{equation}

733: \left[\left({\bf\frac{1}{2}},{\bf 0}\right)+2({\bf 0},{\bf 0})\right]\otimes {\cal R}

734: \label{lorentzrep}

735: \end{equation}

736: contribute to the loop. Here  ${\cal R}$ is an arbitrary Lorentz

737: representation of $SO(4)$. Moreover the two $R_+$ insertions are

738: absorbed by the first factor in the Lorentz representation

739: (\ref{lorentzrep}), and the coupling of the particles in the loop

740: to $F_+$ insertions in the $N=2$ evaluation works exactly  as in

741: the non-supersymmetric Schwinger-loop calculation above for $P$ in

742: the representation ${\cal R}$.

743: 

744: What are the microscopic BPS states that run in the loop? They are

745: related to non-perturbative RR states, which are the only charged

746: states in the Type II compactification. They come from branes

747: wrapping cycles in the Calabi-Yau, and as BPS states their masses

748: are proportional to their central charge (\ref{dbranecharge}). For

749: example, in the large radius in the type IIA string on $M$, the

750: mass is determined by integrals of complexified volume forms over

751: even cycles. E.g., the mass of a 2 brane wrapping a holomorphic

752: curve ${\cal C}_\beta\in  H^2(M,\mathbb{Z})$ is given by

753: \begin{equation}

754: m_\beta=\frac{1}{\lambda} \int_{ {\cal C}_\beta\in H^2(M,\mathbb{Z})} (iJ+B)=

755: \frac{1}{\lambda} 2 \pi i t\cdot \beta =:\frac{1}{\lambda}t_\beta\ .

756: \label{massd2}

757: \end{equation}

758: We note that $H^2(M,\mathbb{Z})$ plays here the role of the charge

759: lattice. In the type IIB picture the charge is given by integrals of the

760: normalized holomorphic $(3,0)$-form $\Omega$. In particular the mass of the extremal

761: black hole that vanishes at the conifold  is given by

762: \begin{equation}

763: m_{BH}=\frac{1}{\lambda \int_{A_D} \Omega} \int_{S^3} \Omega =:\frac{1}{\lambda}t_D\ ,

764: \label{massbh}

765: \end{equation}

766: where $A_D$ is a suitable non-vanishing cycle at the conifold. It

767: follows from the discussion in previous paragraph that with the

768: identification $e=m$ and  after a rescaling $s\rightarrow s

769: \lambda / e$ in (\ref{schwingerloop}), as well as absorbing $F$

770: into $\lambda$, one gets a result for (\ref{N=2oneloop})

771: \begin{equation}

772: F(\lambda, t)= \int_{\epsilon}^\infty \frac{{\rm d} s}{s} \ds{{

773: {\rm Tr} (-1)^f \exp({- s t }) \exp({-2 s  \sigma_L

774: \lambda})}\over 4 \sin^2\ds{\left( s \lambda /2\right)}}\ .

775: \label{n=2loop}

776: \end{equation}

777: Here $t$ are the regularized masses, c.f.

778: (\ref{massd2},\ref{massbh}) of the light particles $P$ that are

779: integrated out, $f$ is their spins in ${\cal R}$, and $\sigma_L$

780: is the Cartan element in the representation  ${\cal R}$.

781: 

782: At the large volume point one can related the relevant BPS states

783: actually to bound states of $D2$  with and infinite tower of $D0$

784: branes with quantized momenta along the $M$-theory circle.

785: Moreover the left spin content of the bound state can be related

786: uniquely to the genus of ${\cal C}_\beta$. This beautiful story

787: leads, as explained in \cite{GVI,GVII}, after summing over the to

788: momenta of the $D0$ states  to (\ref{schwingerloopd2d0}), which

789: together with Castelnuovo's bound for smooth curves leads to very

790: detailed and valuable boundary information as explained in Sec.

791: \ref{symplecticinvariants}. It is important to note that all

792: states are massive so that there are no poles in $F^{(g)}$ for

793: $g>1$. Hence the leading contribution is regular and can be

794: extracted from the constant $\beta=0$  contribution in

795: (\ref{schwingerloopd2d0}) as

796: \begin{eqnarray} \label{constantmap}

797: \lim_{t\rightarrow

798: \infty}F^{(g)}_{\textrm{A-model}}=\frac{(-1)^{g-1}B_{2g}B_{2g-2}}{2g(2g-2)(2g-2)!}\cdot \frac{\chi}{2}\ .

799: \label{constantmaps}

800: \end{eqnarray}

801: The moduli space of constant maps factors into tow components: the

802: moduli space of the world sheet curve $\Sigma_g$ and the location

803: of the image point, i.e. $M$. The measures on the components are

804: $\lambda^{3g-3}$ and $c_3(T_M)$ respectively, explaining the above

805: formula.

806: 

807: Let us turn to type IIB compactifications near the conifold. As it was checked

808: with the $\beta$-function  in~\cite{Vafa:1995ta} there is precisely one

809: BPS hypermultiplet with the Lorentz representation of the

810: first factor in (\ref{lorentzrep}) becoming massless at the conifold. In this case

811: the Schwinger-Loop calculation (\ref{n=2loop}) simply becomes

812: \begin{equation}

813: F(\lambda, t_D)=\int_{\epsilon}^\infty \frac{{\rm d} s}{s} \ds{{

814: \exp({- s t_D })} \over 4 \sin^2\ds{\left( s \lambda

815: /2\right)}}+{\cal O}(t_D^0) =\sum_{g=2}^\infty \left(\lambda\over

816: t_D\right)^{2g-2} \frac{(-1)^{g-1} B_{2g}}{2 g (2 g-2)}+{\cal

817: O}(t_D^0)\ . \label{gap}

818: \end{equation}

819: Since there are no other light particles, the above equation

820: (\ref{gap}) encodes all singular terms in the effective action.

821: There will be regular terms coming from other massive states. This

822: is precisely the gap condition.

823: 

824: 

825: 

826: %--------------------------------------------------------------------------------------

827: \section{Quintic}

828: \label{quintic}

829: 

830: We consider the familiar case of the quintic hypersurface in

831: $\mathbb{P}^4$. The topological string amplitudes $F^{(g)}$ were

832: computed up to genus 4 in \cite{BCOV, Katz:1999} using the

833: holomorphic anomaly equation, and fixing the holomorphic ambiguity

834: by various geometric data. It was also observed a long time ago

835: \cite{Ghoshal:1995} that the leading terms in $F^{(g)}$ around the

836: conifold point are the same as $c=1$ strings at the self-dual radius,

837: thus providing useful information for the holomorphic ambiguity. We

838: want to explore whether we can find coordinates in which 

839: the $F^{(g)}$ on the compact Calabi-Yau exhibit the gap structure 

840: around the conifold point that was recently found for local 

841: Calabi-Yau geometries~\cite{Huang}. In order to do this, it is useful to 

842: rewrite the topological string amplitudes

843: as polynomials \cite{Yamaguchi}. We briefly review the formalism

844: in \cite{Yamaguchi}.

845: 

846: The quintic manifold $M$ has one K\"ahler modulus $t$ and its mirror

847: $W$ has one complex modulus $\psi$ and is given by the 

848: equation~\footnote{Here for later convenience we use a slightly 

849: different notation from that of

850: \cite{Yamaguchi}. Their notation is simply related to ours by a

851: change of variable $\psi^5\rightarrow \psi$.}

852: \begin{equation}

853: W=x_1^5+x_2^5+x_3^5+x_4^5+ x_5^5-5\psi^\frac{1}{5}x_1x_2x_3x_4x_5=0\ .

854: \label{quintic} 

855: \end{equation} 

856: 

857: There is a relation between $t$ and $\psi$ known as the mirror 

858: map $t(\psi)$. The mirror map and the genus zero amplitude 

859: can be obtained using the Picard-Fuchs equation on $W$

860: \begin{eqnarray}

861: \{(\psi\partial_{\psi})^4-\psi^{-1}(\psi\partial_{\psi}-\frac{1}{5})

862: (\psi\partial_{\psi}-\frac{2}{5})(\psi\partial_{\psi}-\frac{3}{5})(\psi\partial_{\psi}-\frac{4}{5})\}\omega=0

863: \label{differentialoperator}

864: \end{eqnarray}

865: We can solve the equation as an asymptotic series around

866: $\psi\rightarrow \infty$.

867: %------------------------------------------------------------------------------------------------

868: \subsection{$\psi=\infty$ expansion and integer symplectic basis}

869: \label{integralsymplecticbasis}

870: Here we set the notation for the periods in the integer 

871: symplectic basis on $W$  and the relation of this basis to the 

872: D-brane charges on $M$. Eqs. (\ref{periods},\ref{prepotential}) and 

873: following, apply to all one parameter models.

874:  

875: The point $\psi=\infty$  has maximal unipotent monodromy and

876: corresponds to the large radius expansion of the mirror $M$

877: \cite{Candelas:1990rm}. In the variable $z={1\over 5^5\psi}$ and using the definitions

878: \begin{equation}

879: \omega(z,\rho):= \sum

880: _{n=0}^{\infty}\frac{\Gamma(5(n+\rho)+1)}{\Gamma^5(n+\rho+1)} z^{n

881: +\rho}\qquad D^k_\rho \omega:=\left.\frac{1}{(2\pi i)^k k!}\frac{\partial^k}{\partial^k \rho}\omega\right|_{\rho=0}

882: \end{equation}

883: one can write  the solutions~\cite{Hosono:1994ax}

884: \begin{equation}

885: \begin{array}{rl}

886: \omega_0&= \omega(z,0)=\sum _{n=0}^{\infty}

887: \frac{(5n)!}{(n!)^5(5^5\psi)^{n}}\\[ 2mm]

888: \omega_1&=D_\rho \omega(z,0)={1\over 2 \pi i} \left(\omega_0 \ln(z)+\sigma_1\right)\\[ 2mm]

889: \omega_2&= \kappa D^2_\rho \omega(z,\rho)-c \omega_0={\kappa\over 2\cdot  (2 \pi i)^2 }\left( \omega_0 \ln^2(z)+2 \sigma_1 \ln(z)+\sigma_2\right) \\[2 mm]

890: \omega_3&= \kappa D^3_\rho \omega(z,\rho)-c \omega_1+e \omega_0=\frac{\kappa}{6 \cdot  (2 \pi i)^3 }\left( \omega_0 \ln^3(z)+3 \sigma_1 \ln^2(z)+3 \sigma_2\ln(z)+\sigma_3\right)

891: \end{array}

892: \label{rawbasis}

893: \end{equation}

894: The constants $\kappa,c,e$ are topological intersection numbers,

895: see below. The $\sigma_k$ are also determined by

896: (\ref{differentialoperator}). To the first few orders,

897: $\omega_0=1+120 z+113400z^2+{\cal O}(z^3)$, $\sigma_1=770z+810225

898: z^2+{\cal O}(z^3)$, $\sigma_2=1150z+{4208175\over 2} z^2+{\cal

899: O}(z^3)$ and $\sigma_3=-6900z-{9895125\over 2}z^2+{\cal O}(z^3)$.

900: 

901: 

902: The solutions (\ref{rawbasis}) can be combined into the period vector

903: $\vec \Pi$ with respect to an integer symplectic basis\footnote{With $A^i\cap B_j=\delta^i_j$

904: and zero intersections otherwise.} $(A^i,B_j)$ of $H^3(W,\mathbb{Z})$

905: as follows \cite{Candelas:1990rm}:

906: \begin{equation}

907: \vec \Pi  =

908: \left(

909: \begin{array}{c}

910: \int_{B_1} \Omega\\

911: \int_{B_2} \Omega\\

912: \int_{A^1} \Omega\\

913: \int_{A^2} \Omega

914: \end{array}

915: \right)=

916: \left(

917: \begin{array}{c}

918: F_0\\

919: F_1\\

920: X_0\\

921: X_1\\

922: \end{array}

923: \right)=

924: \omega_0\left(

925: \begin{array}{c}

926: 2{\cal F}^{(0)} -t \partial_t {\cal F}^{(0)}\\

927: \partial_t {\cal F}^{(0)} \\

928: 1\\

929: t\\

930: \end{array}

931: \right)=\left(

932: \begin{array}{c}

933: \omega_3+c\, \omega_1+e\, \omega_0 \\

934: -\omega_2- a\, \omega_1+c \, \omega_0 \\

935: \omega_0\\

936: \omega_1 \\

937: \end{array}

938: \right)\ .

939: \label{periods}

940: \end{equation}

941: Physically, $t$ is the complexified area of a degree one curve and is related by the mirror map

942: \begin{eqnarray}

943: 2 \pi i t(\psi)&=&\int_{\cal C} (iJ+ B)=

944: {\omega_1\over \omega_0}=-\log(5^5\psi)+\frac{154}{625\psi}+\frac{28713}{390625\psi^{2}}+.. \label{mm}\\

945: {1\over z}=5^5 \psi&=&\frac{1}{q} + 770 + 421375\,q +

946: 274007500\,q^2 +\ldots \label{imm}

947: \end{eqnarray}

948: to $\psi$. In (\ref{imm}), we inverted (\ref{mm}) with  $q=e^{2 \pi i t}$.

949: The prepotential is given by

950: \begin{equation}

951: {\cal F}^{(0)}=-{\kappa\over 3!} t^3-{a\over 2}  t^2+ c t+\frac{e}{2}+ f_{inst}(q) \

952: \label{prepotential}

953: \end{equation}

954: where the instanton expansion $f_{inst}(q)$  vanishes in the large

955: radius  $q\rightarrow 0$ limit. The constants in

956: (\ref{periods},\ref{prepotential}) can be related to the classical

957: geometry of the mirror

958: manifold~\cite{Hosono:1994ax,Candelas:1990rm}. Denote by  $J\in

959: H_2(M,\mathbb{Z})$ the K\"ahler form which spans the one-dimensional

960: K\"ahler cone. Then $\kappa=\int_M J\wedge J\wedge J

961: $ is the classical triple intersection number, $c={1\over 24}

962: \int_M c_2(T_M) \wedge J$ is proportional to the  evaluation of

963: the second Chern class of the tangent bundle $T_M$ against $J$,

964: $e={\zeta(3)\over (2 \pi i)^3}\int_M c_3(T_M)$ is proportional to

965: the Euler number, and  $a$ is related to the topology of the divisor

966: $D$ dual to $J$. $A^1$ is topologically a $T_{\mathbb{R}}^3$,

967: while analysis at the conifold shows that the dual cycle $B_1$ has

968: the topology of an $S^3$.

969: 

970: The identification of the central charge formula for compactified

971: type II supergravity~\cite{Ceresole:1995jg} with the K-theory

972: charge of $D$-branes as objects $A\in

973: K(M)$~\cite{Minasian:1997mm,Cheung:1997az},

974: \begin{equation}

975: \vec Q\cdot \vec \Pi=-\int_M e^{-\hat J} {\rm ch}(A)\sqrt{{\rm td}( M )}=Z(A),

976: \label{dbranecharge}

977: \end{equation}

978: with $\hat J =t (i J+B)$ for the one-dimensional K\"ahler cone,

979: checks the $D$-brane interpretation of (\ref{periods})~\cite{Brunner:1999jq,Diaconescu:1999vp,Mayr:2000as}

980: in the $q\rightarrow 0$ limit on the classical terms $\kappa, a, c$, but misses the  ${\chi \zeta(3)\over (2 \pi i)^3}$

981: term. Based on their scaling with the area parameter $t$ the periods $(F_0,F_1,X_0,X_1)$

982: are identified with the masses of $(D_6,D_4,D_0,D_2)$ branes. For smooth intersections and

983: $D$ the restriction of the hyperplane class we can readily calculate $\kappa,a,c,\chi$

984: using the adjunction formula, see Appendix A.

985: 

986: %------------------------------------------------------------------------------------------------------

987: \subsection{Polynomial expansion of $F^{(g)}$}

988: \label{polynomialfg}

989: 

990: >From the periods,  or equivalently the prepotential ${\cal

991: F}^{(0)}$, we can compute the K\"ahler potential $K:=-\log

992: (i\left({\bar X}^{\bar a} F_a- X^{a} \bar F_{\bar a}\right))$ and

993: metric $G_{\psi\bar{\psi}}:=\partial_\psi {\bar\partial}_{\bar

994: \psi}\, K$ in the moduli space. The genus zero Gromov-Witten

995: invariants are obtained by expanding ${\cal F}^{(0)}$ in large

996: K\"ahler parameter in a power series in $q$, see Section

997: \ref{symplecticinvariants}.

998: 

999: 

1000: We use the notation of \cite{Yamaguchi} and introduce the following symbols:

1001: \begin{eqnarray}

1002: &&

1003: A_p:=\frac{(\psi\partial_{\psi})^pG_{\psi\bar{\psi}}}{G_{\psi\bar{\psi}}},

1004: ~~~B_p:=\frac{(\psi\partial_{\psi})^pe^{-K}}{e^{-K}}, ~~(p=1,2,3,\cdots) \nonumber \\

1005: && C:=C_{\psi\psi\psi}\psi^3, ~~~ X:=\frac{1}{1-\psi}=:-\frac{1}{\delta}

1006: \label{generators}

1007: \end{eqnarray}

1008: Here $C_{\psi\psi\psi}\sim \frac{\psi^{-2}}{1-\psi}$ is the three

1009: point Yukawa coupling, and $A:=A_1=-\psi\Gamma_{\psi\psi}^{\psi}$

1010: and $B:=B_1=-\psi\partial_\psi K$ are the Christoffel and K\"ahler

1011: connections. In the holomorphic limit $\bar{\psi}\rightarrow

1012: \infty$, the holomorphic part of the K\"ahler potential and the

1013: metric go like

1014: \begin{eqnarray}

1015: e^{-K}\sim \omega_0, ~~~~G_{\psi\bar{\psi}}\sim \partial_\psi t,

1016: \end{eqnarray}

1017: so in the holomorphic limit, the generators $A_p$ and $B_p$ are

1018: \begin{eqnarray}

1019: A_p=\frac{(\psi\partial_{\psi})^p(\partial_\psi t)}{\partial_\psi

1020: t}, ~~~B_p=\frac{(\psi\partial_{\psi})^p \omega_0}{ \omega_0}.

1021: \end{eqnarray}

1022: It is straightforward to compute them in an asymptotic series

1023: using the Picard-Fuchs equation. There are also some derivative

1024: relations among the generators,

1025: \begin{eqnarray}

1026: \psi\partial_\psi A_p=A_{p+1}-AA_p, ~~~ \psi\partial_\psi

1027: B_p=B_{p+1}-BB_p,~~~ \psi\partial_\psi X=X(X-1). \nonumber

1028: \end{eqnarray}

1029: The topological amplitudes in the ``Yukawa coupling frame'' are

1030: denoted as

1031: \begin{eqnarray}

1032: P_g:=C^{g-1}F^{(g)}, ~~~~

1033: P^{(n)}_g=C^{g-1}\psi^{n}C_{\psi^n}^{(g)}.

1034: \label{defPg}

1035: \end{eqnarray}

1036: The A-model higher genus Gromov-Witten invariants are obtained in

1037: the holomorphic limit $\bar{t}\rightarrow \infty$ with a familiar

1038: factor of $\omega_0$ as

1039: \begin{eqnarray} \label{10-14-2.8}

1040: F^{(g)}_{\textrm{A-model}} &=& \lim_{\bar{t}\rightarrow \infty}

1041: \omega_0^{2(g-1)}(\frac{1-\psi}{5\psi})^{g-1}P_g.

1042: \end{eqnarray}

1043: The $P^{(n)}_g$ are defined for $g=0, n\geq 3$, $g=1, n\geq 1$,

1044: and $g=2, n\geq 0$. We have the initial data

1045: \begin{eqnarray}

1046: P^{(3)}_{g=0}&=&1 \nonumber \\

1047: P^{(1)}_{g=1} &=&

1048: -\frac{31}{3}B+\frac{1}{12}(X-1)-\frac{1}{2}A+\frac{5}{3}

1049: \label{Pg01},

1050: \end{eqnarray}

1051: and using the Christoffel and K\"ahler connections we have the

1052: following recursion relation in $n$,

1053: \begin{eqnarray}

1054: P^{(n+1)}_g=\psi\partial_\psi

1055: P^{(n)}_g-[n(A+1)+(2-2g)(B-\frac{1}{2}X)]P^{(n)}_g.

1056: \end{eqnarray}

1057: 

1058: One can also use the Picard-Fuchs equation and the special

1059: geometry relation to derive the following relations among

1060: generators:

1061: \begin{eqnarray}

1062: B_4&=&2XB_3-\frac{7}{5}XB_2+\frac{2}{5}XB-\frac{24}{625}X \nonumber \\

1063: A_2&=& -4B_2-2AB-2B+2B^2-2A+2XB+XA+\frac{3}{5}X-1.

1064: \end{eqnarray}

1065: By taking derivatives w.r.t. $\psi$ one see that all higher $A_p$

1066: ($p\geq 2$) and $B_p$ ($p\geq 4$) can be written as polynomials of

1067: $A$, $B$, $B_2$, $B_3$, $X$. It is convenient to change variables

1068: from $(A,B,B_2,B_3,X)$ to $(u,v_1,v_2,v_3,X)$ as

1069: \begin{eqnarray}

1070: && B=u, ~~~ A=v_1-1-2u,~~~ B_2=v_2+uv_1, \nonumber \\

1071: && B_3=v_3-uv_2+uv_1X-\frac{2}{5}uX.

1072: \end{eqnarray}

1073: Then the  main result of \cite{Yamaguchi} is the following

1074: proposition.

1075: \\

1076: 

1077: \noindent{\bf Proposition}: Each $P_g$ ($g\geq 2$) is a degree

1078: $3g-3$ inhomogeneous polynomial of $v_1$, $v_2$, $v_3$, $X$, where

1079: one assigns the degree $1,2,3,1$ for $v_1, v_2, v_3, X$,

1080: respectively.

1081: \\

1082: 

1083: For example, at genus two the topological string amplitude is

1084: \begin{eqnarray}

1085: P_2 &=& \frac{25}{144} - \frac{625}{288} v_1 + \frac{25}{24} v_1^2

1086: - \frac{5}{24}v_1^3 - \frac{625}{36}v_2 + \frac{25}{6}v_1v_2 +

1087: \frac{350}{9}v_3 - \frac{5759}{3600}X  \nonumber \\ && -

1088: \frac{167}{720}v_1X + \frac{1}{6}v_1^2X - \frac{475}{12}v_2X +

1089: \frac{41}{3600}X^2- \frac{13}{288}v_1X^2 + \frac{X^3}{240}.

1090: \end{eqnarray}

1091: The expression for the $P_g$, $g=1,\ldots,12 $ for all models

1092: discussed in this paper can be obtained in a

1093: Mathematica readable form on \cite{webpage}, see ``Pgfile.txt.''

1094: 

1095: %--------------------------------------------------------------

1096: \subsection{Integration of the holomorphic anomaly equation}

1097: \label{integration}

1098: The anti-holomorphic derivative $\partial_{\bar{\psi}}B_p$ of

1099: $p\geq 2$ can be related to $\partial_{\bar{\psi}}B$

1100: \cite{Yamaguchi}. Assuming $\partial_{\bar{\psi}}A$ and

1101: $\partial_{\bar{\psi}}B$ are independent, one obtains two

1102: relations for $P^{(n)}_g$ from the BCOV holomorphic anomaly

1103: equation as the following:

1104: \begin{eqnarray}

1105: \frac{\partial P_g}{\partial u}&=&0 \label{BCOV1} \\

1106: (\frac{\partial}{\partial{v_1}}-u\frac{\partial }{\partial

1107: {v_2}}-u(u+X)\frac{\partial}{\partial{v_3}})P_g&=&-\frac{1}{2}(P^{(2)}_{g-1}+\sum_{r=1}^{g-1}P^{(1)}_rP^{(1)}_{g-r})

1108: \label{BCOV2}

1109: \end{eqnarray}

1110: The first equation (\ref{BCOV1}) implies there is no $u$

1111: dependence in $P_g$, as already taken into account in the main

1112: proposition of \cite{Yamaguchi}. The second equation (\ref{BCOV2})

1113: provides a very efficient way to solve $P_g$ recursively up to a

1114: holomorphic ambiguity. To do this, one writes down an ansatz for

1115: $P_g$ as a degree $3g-3$ inhomogeneous polynomial of $v_1$, $v_2$,

1116: $v_3$, $X$, plugs in equation (\ref{BCOV2}), and uses the lower

1117: genus results $P_r^{(1)}$, $P^{(2)}_r$ ($r\leq g-1$) to fix the

1118: coefficients of the polynomials. The number of terms in general

1119: inhomogeneous weighted polynomials in $(v_1$, $v_2$, $v_3,X)$ with

1120: weights $(1,2,3,1)$ is given by the generating function

1121: \begin{equation}

1122: {1\over {(1-x)^3 (1-x^2) (1-x^3)}}=\sum_{n=0}^\infty p(n) x^n\ .

1123: \end{equation}

1124: It is easy to see from the equation (\ref{BCOV2}) that the terms

1125: $v_1,\ldots v_1^{2g-4}$ vanish, which implies that the number of

1126: terms in  $P_g$ is

1127: \begin{equation}

1128: n_g= p(3 g-3)-(2 g-4),

1129: \end{equation}

1130: e.g. $n_g= 14, 62, 185, 435, 877, 1590, 2666, 4211, 6344$ for

1131: $g=1,\ldots,10$. Comparing with $\sum_{n=1}^\infty \tilde p(n)

1132: x^n={1\over (1-x)^5}$ it follows in particular that

1133: asymptotically

1134: \begin{equation}

1135: n_g \preceq (3g-3)^4 \ . \label{growth}

1136: \end{equation}

1137: 

1138: 

1139: Note that equation (\ref{BCOV2}) determines the term

1140: $\hat  P_g(v_1,v_2,v_3,X)$ in

1141: \begin{equation}

1142: P_g=:\hat P_g(v_1,v_2,v_3,X) + f^{(g)}(X)\

1143: \label{fgdefinition}

1144: \end{equation}

1145: completely. We can easily understand why the terms in the modular

1146: as well as holomorphic ambiguity $f^{(g)}(X)$ are not fixed, by

1147: noting that the equation (\ref{BCOV2}) does not change when we add

1148: a term proportional to $X^i$ to $P_g$. This ambiguity must have

1149: the form~\cite{Katz:1999,Yamaguchi}

1150: \begin{eqnarray} \label{10-14-2.16}

1151: f^{(g)}=\sum_{i=0}^{3g-3}a_iX^i\ .

1152: \end{eqnarray}

1153: The maximal power of $X$ is determined by (\ref{gap}). This

1154: follows from (\ref{defPg}) and the universal behaviour of $t_D\sim

1155: \delta+ {\cal O}(\delta^2)$ at the conifold (\ref{dualmirrormap}).

1156: Note in particular from (\ref{leadingconifold}) that $\hat  P_g$

1157: is less singular at that point. The minimal power in

1158: (\ref{10-14-2.16}) follows from (\ref{constantmaps}) for $\psi

1159: \sim \infty$ and the leading behaviour of the solutions in Sect.

1160: \ref{integralsymplecticbasis}.

1161: 

1162: We will try to fix these $3g-2$ unknown constants $a_i$

1163: ($i=0,1,2\cdots 3g-3$) by special structure of expansions of

1164: $F^{(g)}$ around the orbifold point $\psi\sim 0$ and the conifold

1165: point $\psi\sim 1$. Before proceeding to this, we note the

1166: constant term is fixed by the known leading coefficients in large

1167: complex structure modulus limit $\psi\sim \infty$ in

1168: \cite{Marino:1998, GVII, FP}.

1169: 

1170: The leading constant terms in the A-model expansion $\psi\sim

1171: \infty$ come from the constant map from the worldsheet to the

1172: Calabi-Yau (\ref{constantmap}). The large complex structure

1173: modulus behavior of $X$ is

1174: \begin{eqnarray}

1175: X\sim \frac{1}{\psi}\sim q=e^{2\pi it}.

1176: \end{eqnarray}

1177: So only the constant term $a_0$ in the holomorphic ambiguity

1178: contributes to leading term in A-model expansions

1179: (\ref{constantmap}), and is thus fixed. We still have $3g-3$

1180: coefficients $a_i$ ($i=1,2,\cdots, 3g-3$) to be fixed.

1181: 

1182: 

1183: 

1184: %------------------------------------------------------------------

1185: \subsection{Expansions around the orbifold point $\psi=0$}

1186: \label{quinticorbifold}

1187: To analyze the $F^{(g)}$ in a new region of the 

1188: moduli space we have to find the right choice of 

1189: polarization. To do this we analytically 

1190: continue the periods to determine the symplectic 

1191: pairing in the new region and pick the new 

1192: choice of conjugated varaibles.      

1193: 

1194: The solutions of the  Picard-Fuchs equation around the orbifold point$\psi\sim 0$

1195: are four power series solutions with the indices

1196: $\frac{1}{5},\frac{2}{5},\frac{3}{5},\frac{4}{5}$

1197: \begin{equation}

1198: \begin{array}{rl}

1199: \omega_{k}^{\rm orb}&=\ds{\psi^{\frac{k}{5}}\sum_{n=0}^\infty

1200: \frac{\left(\left[\frac{k}{5}\right]_n\right)^5}{\left[k\right]_{5 n}}\left(5^5 \psi\right)^n} \\[2 mm]

1201: &=\ds{-\frac{\Gamma(k)}{\Gamma^5\left(\frac{k}{5}\right)}\int_{{\cal

1202: C}_0} \frac{\dd s}{e^{2 \pi i s}-1}

1203: \frac{\Gamma^5\left(s+\frac{k}{5}\right)}{\Gamma(5 s+k)}

1204: \left(5^5 \psi\right)^{s+\frac{k}{5}} \ , \qquad k=1,\ldots, 4}\ .

1205: \label{orbifoldbasis}

1206: \end{array}

1207: \end{equation}

1208: \parbox{14cm} % Seitenbreite

1209: {

1210:    %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1211:    %% Abbildung contour  %%%%%%%%%%%%%%%%%%%%%%

1212:    %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1213: 

1214:    \begin{center}

1215:    \mbox{

1216:          % \begin{turn}{-90}  % Hier k"onnen Sie das Bild drehen

1217:                              % z.B. um -90 Grad

1218:              \epsfig{file=contour.eps,width=14cm}

1219:          % \end{turn}

1220: }

1221:    \end{center}

1222:    %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1223: }

1224: %

1225: \vskip -2mm

1226: \centerline{ $\qquad\qquad$ ${\cal C}_\infty$ for $|\psi|> 1$ $\qquad\qquad\qquad$ ${\cal C}_0$ for $|\psi|< 1$.}

1227: % nachfolgender Text

1228: \vskip 2mm

1229: 

1230: The Pochhammer symbol are defined as $[a]_n:=\frac{\Gamma(a+n)}{\Gamma(a)}$ and we

1231: normalized the first coefficient in  $\omega_{k}^{\rm orb}= \psi^\frac{k}{5}+{\cal O}(\psi^\frac{5+k}{5})$

1232: to one.  The expression in the first line is recovered from the integral representation

1233: by noting that the only poles inside ${\cal C}_0$ for which the integral

1234: converges for $|\psi|<0$ are from $g(s)=\frac{1}{\exp(2 \pi i s)-1}$, which behaves

1235: at $s^\eps_{-n}=n-\epsilon$, $n\in \mathbb{N}$ as $g(s^\eps_{-n})\sim-\frac{1}{2\pi i \epsilon}$.

1236: 

1237: Up to normalization this basis of solutions is canonically

1238: distinguished, as it diagonalizes the $\mathbb{Z}_5$ monodromy at

1239: $\psi=0$. Similar as for the $\mathbb{C}^3/\mathbb{Z}_3$

1240: orbifold~\cite{Aganagic:2006wq}, it can be viewed as a twist field

1241: basis. Here this basis is induced from the twist field basis of

1242: $\mathbb{C}^5/\mathbb{Z}_5$. As it was argued

1243: in~\cite{Aganagic:2006wq} for $\mathbb{C}^3/\mathbb{Z}_3$, this

1244: twist field basis provides the natural coordinates in which the

1245: $F^{(g)}$ near the orbifold point can be interpreted as generating

1246: functions for orbifold Gromov-Witten invariants. Following up on

1247: foundational work on orbifold Gromov-Witten theory \cite{CR} and

1248: examples in two complex dimensions~\cite{BGP} this prediction has

1249: been checked by direct computation of orbifold  Gromov-Witten

1250: invariants~\cite{CCIT} at genus zero. This provides a beautiful

1251: check on the global picture of mirror symmetry.

1252: 

1253: As explained in~\cite{Aganagic:2006wq} the relation between the

1254: large radius generating function of Gromov-Witten invariants and

1255: generating function of orbifold Gromov-Witten invariants is

1256: provided by the metaplectic transformation of the wave

1257: function~\cite{wittenwavefunction}. Since the change of the phase

1258: space variables from large radius to the orbifold the symplectic

1259: form is only invariant up to scaling one has to change the

1260: definition of the string coupling, which plays the role of

1261: $\hbar$ in the  metaplectic transformation.  These can be viewed

1262: as small phase space specialization of the  metaplectic

1263: transformation on the large phase space~\cite{GiventhalCoates}.

1264: 

1265: Using the modular invariance of the anholomorphic $\hat F^g(t,\bar

1266: t)$ it has been further shown in~\cite{Aganagic:2006wq} that this

1267: procedure of obtaining the transformed holomorphic wave function

1268: is simply equivalent to taking the holomorphic limit on the $\hat

1269: F^g(t,\bar t)$. This point was made~\cite{Aganagic:2006wq} for the

1270: local case. But the only point one  has to keep in mind for the

1271: global case is that $\hat F^g(t,\bar t)$ are globally well defined

1272: sections of the K\"ahler line bundle. i.e. one has to perform a

1273: K\"ahler transformation along with the holomorphic limit.

1274: 

1275: 

1276: We will now study  the transformation from the basis

1277: (\ref{periods}) to the basis (\ref{orbifoldbasis}) to make B-model

1278: prediction along the lines of~\cite{Aganagic:2006wq} with the

1279: additional K\"ahler transformation. Since the symplectic form

1280: $\omega$ on the moduli space is  invariant under monodromy and

1281: $\omega_k^{orb}$ diagonalizes the $\mathbb{Z}_5$ monodromy, we

1282: must have in accordance with the expectation from the orbifold

1283: cohomology $H^*(\mathbb{C}^5/\mathbb{Z}_5)=\mathbb{C}{\bf

1284: 1}_0\oplus\mathbb{C}{\bf 1}_1\oplus\mathbb{C}{\bf

1285: 1}_2\oplus\mathbb{C} {\bf 1}_3\oplus\mathbb{C}{\bf 1}_4$

1286: \begin{equation}

1287: \omega=\dd F_k\wedge \dd X_k= -\frac{5^4s_1}{6}\dd

1288: \omega^{orb}_4\wedge \dd \omega^{orb}_1+ \frac{5^4s_2}{2}\dd

1289: \omega^{orb}_3\wedge \dd \omega^{orb}_2\ .

1290: \end{equation}

1291: The rational factors above have been chosen to match constraints

1292: from special geometric discussed below.  Similarly the monodromy

1293: invariant K\"ahler potential must have the form

1294: \begin{equation}

1295: e^{-K}=\sum_{k=1}^4 r_k \omega_k^{\rm orb} \overline{\omega_k^{\rm orb}}\ .

1296: \label{eK}

1297: \end{equation}

1298: To obtain the $s_i,r_i$ by analytic continuation to the basis (\ref{periods}) we follow \cite{Candelas:1990rm} for the

1299: quintic and the generalisation in~\cite{Klemm:1992tx} for other cases and  note that the integral converges for

1300: $|\psi|>1$ due to the asymptotics of the $f(s)=\frac{\Gamma^5(s+k/5)}{\Gamma(5s+k)}$ term,

1301: when the integral is closed along ${\cal C}_\infty$~\cite{Candelas:1990rm}. At $s^\eps_{n}=-n-\epsilon$

1302: the  $g(s^\eps_{n})$ pole is compensated by the  $f(s^\eps_{n})$ zero and at

1303: $s^\eps_{n,k}= -n-k/5-\epsilon$ we note the expansions

1304: \begin{equation}

1305: \begin{array}{rl}

1306: g(s^\eps_{n,k})=&{\frac{\a^k}{1-\a^k}+

1307: \frac{2\pi i\a^k}{(1-\a^k)^2}\eps+

1308: \frac{(2\pi i)^2\a^k(1+\a^k)}{2(1-\a^k)^3}\eps^2+

1309: \frac{(2\pi i)^3\a^k(1+4\a^k+\a^{2k})}{6(1-\a^k)^4}\eps^3+{\cal O}(\eps^4)}\\[3 mm]

1310: f(s^\eps_{n,k})=&{\frac{\kappa \omega_0(n)}{\eps^4}+

1311: \frac{\kappa \sigma_1(n)}{\eps^3}+

1312: \frac{1}{\eps^2}\left(\frac{\kappa \sigma_2(n)}{2}+\frac{(2\pi i)^2c_{2J} \omega_0(n)}{24}\right)+}\\[3 mm]

1313: &{\frac{1}{\eps}\left(\frac{\kappa \sigma_3(n)}{6}+\frac{(2\pi i)^2 c_{2J}\sigma_1(n)}{24}+\chi \zeta(3)

1314: \omega_0(n)\right)+{\cal O}(\eps^0)}\\[3 mm]

1315: \left(5^5\psi\right)^{s^\eps_n}=&{z^n(1+\log(z)\eps+\frac{1}{2}\log(z)^2\eps^2+\frac{1}{6}\log(z)^3\eps^3+{\cal O}(\eps^4)})

1316: \end{array}.

1317: \end{equation}

1318: Here  $\alpha=\exp(2 \pi i/5)$. $\kappa=\int_MJ^3$,

1319: $c_{2J}=\int_M c_2 J$, and $\chi=\int_M c_3$ are calculated

1320: in App. \ref{intersection}~\footnote{In fact using the

1321: generalization of (\ref{orbifoldbasis}) in~\cite{Klemm:1992tx} it

1322: is easily shown that the combinatorics, which leads to

1323: (\ref{aconst}), are the same as the ones leading to the occurrence

1324: of the classical intersections here.}.  The $\omega_0(n)$,

1325: $\sigma_i(n)$ are coefficients of the series we encountered in

1326: sec. \ref{integralsymplecticbasis}. Performing the residue

1327: integration and comparing with (\ref{rawbasis},\ref{periods}) we

1328: get

1329: \begin{equation}

1330: \omega_k^{orb}=\frac{(2 \pi i)^4 \Gamma(k)}{\Gamma^5\left(\frac{k}{5}\right)}\left(

1331: \frac{\a^k F_0}{1-\a^k}-

1332: \frac{\a^k F_1}{(1-\a^k)^2}+

1333: \frac{5\a^k(\a^{2 k}-\a^k +1)X_0}{(1-\a^k)^4}+

1334: \frac{\a^k(8\a^{k}-3)X_1}{(1-\a^k)^3}\right)

1335: \label{matoi}

1336: \end{equation}

1337: It follows with $r_i=\frac{\Gamma^{10}\left(\frac{k}{5}\right)}{\Gamma^2(k)} c_i$

1338: that

1339: \begin{eqnarray}

1340: c_1&=&-c_4=\a^2 (1-\a)(2+\a^2+\a^3), \quad  c_3=-c_2=\a(2+\a-\a^2-2 \a^3)\nonumber \\

1341: s_1&=&s_2 =-\frac{1}{5^5(2\pi i)^3} \label{metaplectic}\\

1342: \left(\begin{array}{c}F_0\\ F_1\\X_0\\X_1 \end{array} \right)&=&

1343: \psi^{1/5}\frac{\a \Gamma^5\left(1\over 5\right)}{(2 \pi i)^4} \left(\begin{array}{c}

1344: (1-\a)(\a-1-\a^2)\\

1345: \frac{1}{5}(8-3 \a) (1-\a)^2\\

1346: (1-\a+\a^2)\\

1347: \frac{1}{5}(1-\a)^3 \end{array}\right)+{\cal O}(\psi^{2/5})\ . \label{leading}

1348: \end{eqnarray}

1349: Eq. (\ref{metaplectic}) implies that up to a rational rescaling of

1350: the orbifold periods the transformation of the wave functions from

1351: infinity to the orbifold is given by a metaplectic transformation

1352: with the same rescaling of the string coupling as for the

1353: $\mathbb{C}^3/\mathbb{Z}_3$ case in \cite{Aganagic:2003db}. Eq.

1354: (\ref{leading}) implies that there are no projective coordinates

1355: related to an  ${\rm Sp}(4,\mathbb{Z})$ basis, which would vanish

1356: at the orbifold. This means that there is no massless RR state in

1357: the K-theory charge lattice which  vanishes at the orbifold point.

1358: We note further that after rescaling of the orbifold periods the

1359: transformation (\ref{matoi}) can be chosen to lie in ${\rm

1360: Sp}(4,\mathbb{Z}[\a,\frac{1}{5}])$.

1361: 

1362: 

1363: We can define the analogue of mirror map at the orbifold point,

1364: \begin{eqnarray}

1365: s=\frac{\omega_2^{orb}}{\omega^{orb}_1}=\psi^{\frac{1}{5}}(1+\frac{13\psi}{360}+\frac{110069\psi^{2}}{9979200}

1366: +\mathcal{O}(\psi^{3}))

1367: \end{eqnarray}

1368: where  we use the notation $s$, as in~\cite{Aganagic:2006wq}, to

1369: avoid confusion with the mirror map in the large volume limit. We

1370: next calculate the genus zero prepotential at the orbifold point.

1371: For convenience let us rescale our periods $\hat \omega_{k-1}=

1372: 5^{3/2} \omega^{orb}_{k}$. The Yukawa-Coupling is transformed to

1373: the $s$ variables as

1374: \begin{equation}

1375: C_{sss}=\frac{1}{\hat \omega_0^2}

1376: \frac{5}{\psi^2(1-\psi)} \left(\frac{\partial \psi}{\partial s}\right)^3=

1377: 5 +\frac{5}{3} s^5 +\frac{5975}{6048} s^{10}+\frac{34521785}{54486432} s^{15}+{\cal O}(s^{20})  \ .

1378: \end{equation}

1379: A trivial consistency check of special geometry is that the genus

1380: zero prepotential $F^{(0)}=\int \dd s \int \dd s \int \dd s \

1381: C_{sss}$ appears in the periods $\hat \Pi^{orb}=( \hat

1382: \omega_0,\hat \omega_1,\frac{5}{2!} \hat \omega_2,-\frac{5}{3!}

1383: \hat \omega_3)^T$ as

1384: \begin{equation}

1385: \hat \Pi^{orb}=\hat \omega_0 \left(\begin{array}{c}

1386: 1\cr

1387: s\cr

1388: \partial_s F_{\textrm{A-orbf.}}^{(0)}\cr

1389: 2 F_{\textrm{A-orbf.}}^{(0)}- s\partial_s F_{\textrm{A-orbf.}}^{(0)}

1390: \end{array}

1391: \right)\ .

1392: \end{equation}

1393: This can be viewed also as a simple check on the lowest order 

1394: meta-plectic transformation of $\Psi$ which is just the Legendre 

1395: transformation. Note that the Yukawa coupling is 

1396: invariant under the $\mathbb{Z}_5$ which acts as

1397: $s\mapsto \alpha s$. $\mathbb{Z}_5$ implies further that there can be no

1398: integration constants, when passing from $C_{sss}$ to $F_0$ and the coupling

1399: $\lambda$ must transform with $\lambda \mapsto \alpha^\frac{3}{2} \lambda$ to

1400: render $F(\lambda,s,\bar s)$ invariant.

1401: 

1402: The holomorphic limit  $\bar{\psi}\rightarrow 0$  of K\"ahler

1403: potential and metric follows from (\ref{eK}) by extracting the

1404: leading anti-holomorphic behaviour. Denoting\footnote{This  is to

1405: make contact with the other one modulus cases. Of course if

1406: $a_1=a_2$ a log singularity appears and the formula does not

1407: apply.} by $a_k$ the leading powers of  $\omega_k^{orb}$  we find

1408: \begin{equation}

1409: \lim_{\bar \psi\rightarrow 0} e^{-K}=r_1 \bar \psi^{a_1}

1410: \omega^{orb}_1, \qquad \lim_{\bar \psi\rightarrow 0} G_{\psi\bar

1411: \psi}= \bar \psi^{a_2-a_1-1}

1412: \frac{r_2}{r_1}\left(\frac{a_2}{a_1}-1\right) \frac{\partial

1413: s}{\partial \psi} \ .

1414: \end{equation}

1415: Note that the constants and  $\bar \psi$ and its leading power

1416: are irrelvant for the holomorphic limit of the generators

1417: (\ref{generators})

1418: \begin{eqnarray}

1419: X&=& \frac{1}{1-\psi}=1+\psi+\psi^{2}+\mathcal{O}(\psi^{3})

1420: \nonumber \\ A&=& -\frac{4}{5}+\frac{13}{60}\psi+

1421: \frac{3551}{18144}\psi^2+\mathcal{O}(\psi^{3}) \nonumber \\

1422: B&=&

1423: \frac{1}{5}+\frac{1}{120}\psi+\frac{17}{4032}\psi^2+\mathcal{O}(\psi^{3})

1424: \nonumber \\

1425: B_2 &=& \frac{1}{25}+\frac{7}{600}\psi+\frac{1027}{100800}\psi^2

1426: +\mathcal{O}(\psi^{3}) \nonumber

1427: \\

1428: B_3 &=&

1429: \frac{1}{125}+\frac{43}{3000}\psi+\frac{1633}{72000}\psi^2+\mathcal{O}(\psi^{3}).

1430: \nonumber

1431: \end{eqnarray}

1432: 

1433: Using this information and integrating (\ref{Pg01}) we obtain the

1434: genus one free energy $F^{(1)}_{\textrm{A-orbf.}}=

1435: -\frac{s^5}{9}+\ldots$ The regularity of $F^{(1)}$, i.e. the

1436: absence of log terms, is expected as there are no massless BPS

1437: states at the Gepner-point. Because of this, the considerations in

1438: Sec.\ref{boundaryconditions} imply also that the higher genus

1439: amplitudes

1440: \begin{eqnarray} \label{fgorb}

1441: F^{(g)}_{\textrm{A-orbf.}} &=& \lim_{\bar{\tilde t}\rightarrow 0}

1442: (\hat \omega_0)^{2(g-1)}(\frac{1-\psi}{5\psi})^{g-1}P_g

1443: \end{eqnarray}

1444: have no singularity at the orbifold point $\psi\sim 0$. This is in

1445: accordance with the calculations in \cite{Katz:1999} and implies

1446: that $\frac{P_g}{\psi^{\frac{3}{5}(g-1)}}$ is regular at $\psi\sim

1447: 0$.

1448: 

1449: The situation for the higher genus amplitudes  for the compact

1450: ${\cal O}(5)$ constraint in $\mathbb{P}^4$ is considerably

1451: different from the one for the resolution ${\cal O}(-5)\rightarrow

1452: \mathbb{P}^4$ of the $\mathbb{C}^5/\mathbb{Z}_5$. In normal

1453: Gromov-Witten theory for genus $g>0$ on ${\cal O}(5)$ in

1454: $\mathbb{P}^4$ there is no bundle whose Euler class of its pullback

1455: from the ambient space $\mathbb{P}^4$ to the moduli space of

1456: maps gives rise to a suitable measure on ${\cal M}_{g,\beta}$ that

1457: counts the maps to the quintic. This is the same difficulties one

1458: has to face for higher genus calculation for the orbifold GW

1459: theory in ${\cal O}(5)$ in  $\mathbb{P}^4$, and it is notably

1460: different\footnote{Since the brane bound state cohomology at

1461: infinity can only be understood upon including higher genus

1462: information, see Sec. \ref{symplecticinvariants} and

1463: \ref{quinticdbrane}, the claims that one can learn essential

1464: properties about the D-branes of the quintic at small volume from

1465: the $\mathbb{C}^5/\mathbb{Z}_5$ orbifold might be overly

1466: optimistic.} from the equivariant GW theory on

1467: $\mathbb{C}^5/\mathbb{Z}_5$.

1468: 

1469: However we claim that our $F^{(g)}_{\textrm{A-orbf.}}$ predictions

1470: from the $B$-model computation contain the information about the

1471: light even RR states at the orbifold point in useful variables and

1472: could in principle be checked  in the A-model by some version of

1473: equivariant localisation. Below we give the first few order

1474: results. They are available to genus $20$ at \cite{webpage}.

1475: $$

1476: \begin{array}{rl}

1477: F^{(0)}_{\textrm{A-orbf.}}&=\frac{5\,s^3}{6} + \frac{5\,s^8}{1008} + \frac{5975\,s^{13}}{10378368} +

1478: \frac{34521785\,s^{18}}{266765571072}+\ldots\\ [2 mm]

1479: F^{(1)}_{\textrm{A-orbf.}}&= -\frac{s^5}{9} - \frac{163 \, s^{10}}{18144} - \frac{85031 \, s^{15}}{46702656} -

1480: \frac{6909032915\, s^{20}}{20274183401472} +\ldots \\[2 mm]

1481: F^{(2)}_{\textrm{A-orbf.}}&=\frac{155\,s^2}{18} - \frac{5\,s^7}{864} + \frac{585295\,s^{12}}{14370048} +

1482: \frac{1710167735\,s^{17}}{177843714048}+\ldots\\[2 mm]

1483: F^{(3)}_{\textrm{A-orbf.}}&=\frac{488305\,s^4}{9072} - \frac{3634345\,s^9}{979776} - \frac{1612981445\,s^{14}}{7846046208} -

1484:   \frac{2426211933305\,s^{19}}{116115777662976} +\ldots\\[2 mm]

1485: F^{(4)}_{\textrm{A-orbf.}}&=\frac{48550\,s}{567} + \frac{36705385\,s^6}{163296} + \frac{16986429665\,s^{11}}{603542016} +

1486:   \frac{341329887875\,s^{16}}{70614415872}+\ldots\\[2 mm]

1487: F^{(5)}_{\textrm{A-orbf.}}&=\frac{1237460905\,s^3}{224532} + \frac{108607458385\,s^8}{28740096} - \frac{2079654832074515\,s^{13}}{1553517149184} -

1488:   \frac{50102421421803185\,s^{18}}{438808843984896}+\ldots\\ [2 mm]

1489: \end{array}

1490: $$

1491: 

1492: The holomorphic ambiguity (\ref{10-14-2.16}) is a power series

1493: of $\psi$ starting from a constant term, so requiring

1494: $\frac{P_g}{\psi^{\frac{3}{5}(g-1)}}$ to be regular imposes

1495: \begin{equation}

1496: \lceil\frac{3}{5}(g-1)\rceil

1497: \end{equation}

1498: number of relations in $a_i$ in (\ref{10-14-2.16}), where

1499: $\lceil\frac{3}{5}(g-1)\rceil$ is the ceiling, i.e. the smallest

1500: integer greater or equal to $\frac{3}{5}(g-1)$. We note that the

1501: leading behaviour $\omega^{orb}_1\sim \psi^\frac{1}{d}$ with $d=5$

1502: for the quintic, which is typical for an orbifold point in compact

1503: Calabi-Yau, and which ``shields'' the singularity and diminishes

1504: the boundary conditions at the orbifold point from $g-1$ to

1505: $\lceil \frac{d-2}{d}(g-1)\rceil $. If this period is

1506: non-vanishing at $\psi=0$ and it is indeed simply a constant for

1507: all local cases~\cite{Huang2}, one gets $g-1$ conditions, which

1508: together with the gap condition at the conifold and the constant

1509: map information is already sufficient to completely solve the

1510: model. An example of this type is ${\cal O}(-3)\rightarrow

1511: \mathbb{P}^2$.

1512: 

1513: \subsection{Expansions around the conifold point $\psi=1$}

1514: \label{quinticconifold}

1515: An new feature of the conifold region is that there is an 

1516: choice in picking the polarization, but as we will show 

1517: the gap property is independent of this choice.       

1518: 

1519: A basis of solutions of the Picard-Fuchs equation around the

1520: conifold point $\psi-1=\delta\sim 0$ is the following

1521: \begin{equation}

1522: \vec \Pi_c=\left(

1523: \begin{array}{c}

1524: \omega_0^c\\

1525: \omega_1^c\\

1526: \omega_2^c\\

1527: \omega_3^c\\

1528: \end{array}

1529: \right)

1530: =\left(

1531: \begin{array}{l}

1532: 1+\frac{2\delta^3}{625}-\frac{83\delta^4}{18750}+\frac{757\delta^5}{156250}+\mathcal{O}(\delta^6)\\

1533: \delta-\frac{3\delta^2}{10}+\frac{11\delta^3}{25}-\frac{217\delta^4}{2500}+\frac{889\delta^5}{15625}+\mathcal{O}(\delta^6)\\

1534: \delta^2-\frac{23\delta^3}{30}+\frac{1049\delta^4}{1800}-\frac{34343\delta^5}{75000}+\mathcal{O}(\delta^6) \\

1535: \omega_1^c\log(\delta)-\frac{9\,d^2}{20} - \frac{169\,d^3}{450} + 

1536:   \frac{27007\,d^4}{90000} - 

1537:   \frac{152517\,d^5}{625000}+\mathcal{O}(\delta^6)

1538: \end{array}

1539: \right)

1540: \end{equation}

1541: Here we use the superscript ``c'' in the periods to denote them as

1542: solutions around the conifold point. We see that one of the

1543: solutions $\omega_1^c$ is singled out as it multiplies the log in

1544: the solution $\omega_3^c$. By a Lefshetz argument \cite{AGV} it

1545: corresponds to the integral over the vanishing $S^3$ cycle $B_1$

1546: and moreover a solution containing the log is the integral over

1547: dual cycle $A_1$. Comparing with

1548: (\ref{periods},\ref{dbranecharge}) shows in the Type IIA

1549: interpretation that the $D6$ brane becomes massless. To determine

1550: the symplectic basis we analytically continue the solutions

1551: (\ref{periods}) from $\psi=\infty$ and get

1552: \begin{equation}

1553: \left(

1554: \begin{array}{c}

1555: F_0\\[2 mm]

1556: F_1\\[2 mm]

1557: X_0\\[2 mm]

1558: X_1\\[2 mm]

1559: \end{array}

1560: \right)=

1561: \left(\begin{array}{rrrr}

1562: 0&{\sqrt{5}\over 2 \pi i}&0&0\\[1 mm]

1563: a-{11 i\over 2} g& b-{11i \over 2}h& c-{11i \over 2}r&0\\[1 mm]

1564: d&e&f&-{ \sqrt{5} \over (2 \pi i)^2}\\

1565: ig&ih& ir& 0\end{array}

1566: \right)\left(

1567: \begin{array}{c}

1568: \omega_0^c\\[2 mm]

1569: \omega_1^c\\[2 mm]

1570: \omega_2^c\\[2 mm]

1571: \omega_3^c\\[2 mm]

1572: \end{array}

1573: \right)

1574: \end{equation}

1575: Six of the real numbers

1576: $a,\ldots,r$ are only known numerically\footnote{ 

1577: $a= 6.19501627714957\ldots$,

1578: $b= 1.016604716702582\ldots$, 

1579: $c= -0.140889979448831\ldots$, 

1580: $d=1.07072586843016\ldots$, 

1581: $e= -0.0247076138044847\ldots$, 

1582: $g= 1.29357398450411\ldots$, 

1583: $h=\frac{ 2\,b\,g\,\pi -\left( {\sqrt{5}}\,d \right) }{2\,a\,\pi }$, 

1584: $r=\frac{5 + 16\,c\,g\,{\pi }^3}{16\,a\,{\pi }^3}$,

1585: $f=\frac{{\sqrt{5}}\,b + 8\,c\,d\,{\pi }^2}{8\,a\,{\pi }^2}$.}. 

1586: Nevertheless we can give the symplectic form exactly in the new basis 

1587: \begin{equation} 

1588: \dd F_k\wedge \dd X_k = -\frac{1}{(2 \pi i)^3} \left(\frac{5}{2} \dd \omega_2^c\wedge\dd \omega_0^c +  

1589: (-5)\dd \omega_3^c\wedge\dd \omega_1^c\right)\ .

1590: \label{symplecticconifold} 

1591: \end{equation} 

1592: 

1593: 

1594: The mirror map should be invariant under the conifold monodromy

1595: and vanishing at the conifold. The vanishing period has $D6$

1596: brane charge and is singled out to appear in the numerator of the

1597: mirror map. The numerator is not fixed up to the fact that 

1598: $\omega_3^c$ should not appear. The simplest mirror map compatible with 

1599: symplectic form (\ref{symplecticconifold}) is

1600: \begin{eqnarray} \label{dualmirrormap}

1601: t_D(\delta)&:=&\frac{\omega_1^c}{\omega_0^c}=\delta-\frac{3

1602: \delta^2}{10}+\frac{11\delta^3}{75}-

1603: \frac{9\delta^4}{100}+\frac{5839 t_D^5}{93750}

1604: +\mathcal{O}(t_D^6)\\

1605: \delta(t_D)&=&t_D+\frac{3 t_D^2}{10}+\frac{t_D^3}{30}+\frac{t_D^4}{200}+\frac{169 t_D^5}{375000}

1606: +\mathcal{O}(t_D^6)

1607: \end{eqnarray}

1608: We call this the dual mirror map and denote it $t_D$ to

1609: distinguish from the large complex structure modulus case.

1610: 

1611: 

1612: In the holomorphic limit $\bar{\delta}\rightarrow 0$, the Kahler

1613: potential and metric should behave as $e^{-K}\sim \omega_0^c$ and

1614: $G_{\delta\bar{\delta}}\sim \partial_\delta t_D$. We can find the

1615: asymptotic behavior of various generators,

1616: \begin{eqnarray}

1617: X&=& \frac{1}{1-\psi}=-\frac{1}{\delta}

1618: \nonumber \\

1619: A&=&

1620: -\frac{3}{5}-\frac{2}{25}\delta+\frac{2}{125}\delta^2-\frac{52}{9375}\delta^3+\mathcal{O}(\delta^4)

1621: \nonumber \\

1622: B&=&

1623: \frac{6}{625}\delta^2-\frac{76}{9375}\delta^3+\frac{611}{93750}\delta^4

1624: +\mathcal{O}(\delta^5)

1625: \nonumber \\

1626: B_2 &=&

1627: \frac{12}{625}\delta-\frac{16}{3125}\delta^2+\frac{82}{46875}\delta^3

1628: +\mathcal{O}(\delta^4)

1629: \nonumber \\

1630: B_3 &=&

1631: \frac{12}{625}+\frac{28}{3125}\delta-\frac{78}{15625}\delta^2

1632: +\mathcal{O}(\delta^3) \label{leadingconifold.}

1633: \end{eqnarray}

1634: Now we can expand

1635: \begin{eqnarray} \label{10-14-2.21}

1636: F^{(g)}_{\textrm{conifold}}=\lim_{\bar{\delta}\rightarrow

1637: 0}(\omega_0^c)^{2(g-1)}(\frac{1-\psi}{\psi})^{g-1}P_g

1638: \end{eqnarray}

1639: around the conifold point in terms of $t_D$ using the dual mirror

1640: map (\ref{dualmirrormap}). 

1641: 

1642: Remarkably it turns out that  

1643: shifts $\omega_0^c\rightarrow \omega_0^c+b_1\omega_1^c+b_2\omega_2^c$ 

1644: does not affect the structure we are interested in. The fact that the 

1645: $b_1$ shifts do not affect the amplitudes is reminiscent of the ${\rm

1646: SL}(2,\mathbb{C})$ orbit theorem~\cite{Schmidt,CKS} and is proven

1647: in App. \ref{sectioninvariance}. It is therefore reasonable to 

1648: state the results in the more general polarization and define $\hat \omega^c_0= \omega^c_0 

1649: + b_1 \omega^c_1 + b_2 \omega^c_2$. We first determine the genus

1650: $0$ prepotential checking consistency of the solutions with special 

1651: geometry. Defining $\hat t_D=\frac{\omega_1^c}{\hat \omega_0^c}$ and 

1652: $\Pi_{con}=(\hat \omega^c_0,\omega^c_1,\frac{5}{2}\omega^c_2,-5\omega^c_3)^T$ we get

1653: \begin{equation}

1654: \hat \Pi^{orb}=\hat \omega_0 \left(\begin{array}{c}

1655: 1\cr

1656: \hat t_D\cr

1657: 2 F_{\textrm{conif.}}^{(0)}- \hat t_D \partial_{\hat t_D} 

1658: F_{\textrm{conif.}}^{(0)}\cr

1659: \partial_{\hat t_D} F_{\textrm{conif.}}^{(0)}

1660: \end{array}

1661: \right)\ .

1662: \end{equation}

1663: We can now calculate the $F^{(g)}$ in the generalized polarization. Since the normalization of

1664: $\hat \omega_0^c$ is not fixed by the Picard-Fuchs equation, we 

1665: pick the  normalization $b_0=1$ that will be convenient later on\footnote{The $b_0$ dependence can be restored noting

1666: that each order in $\hat t_D$ is homogeneous in $b_i$.} Using the known expression for the ambiguity 

1667: at genus $2$, $3$ \cite{BCOV, Katz:1999,Yamaguchi} for $g=0-3$ and Castelnuovos bound for $g>3$ we 

1668: find the same interesting structure first observed in \cite{Huang}

1669: \begin{equation} 

1670: \begin{array}{rl} 

1671: F_{\textrm{conif.}}^{(0)}&=-\frac{5}{2} \log (\hat t_D) \hat t_D^2 +\frac{5}{12}\left( 1 - 6b_1 \right)\hat t^3_D + 

1672:   \left(\frac{5}{12} \left( b_1 - 3 b_2 \right)  -\frac{89}{1440}  - 

1673: \frac{5}{4}\,b_1^2\right) \hat t_D^4 + {\cal O}(\hat t_D^5)\\ [2 mm]

1674: F_{\textrm{conif.}}^{(1)}&=-\frac{\log (\hat t_D)}{12} + \left( \frac{233}{120} - \frac{113\,b_1}{12} \right) \,\hat t_D + 

1675:   \left(\frac{233\,b_1}{120} - \frac{113\,{b_1}^2}{24} - 

1676:      \frac{107 b_2}{12} -\frac{2681}{7200}\right) \,{\hat t_D}^2 + {{\cal O}(\hat t_D^3)}\\[2 mm]

1677: F_{\textrm{conif.}}^{(2)}&=\frac{1}{240 {\hat t_D}^2} - \left(\frac{120373}{72000}   + \frac{11413 b_2}{144}\right) + 

1678:   \left( \frac{107369}{150000} - \frac{120373\,b_1}{36000} + \frac{23533\,b_2}{720} - 

1679:      \frac{11413 b_1 b_2}{72} \right) \,\hat t_D + {{\cal O}(\hat t_D^2)}\\ [2 mm] 

1680: F_{\textrm{conif.}}^{(3)}&=\frac{1}{1008\,{\hat t_D}^4} - \left(

1681:    \frac{178778753}{324000000}   + \frac{2287087\,b_2}{43200} + 

1682:     \frac{1084235\,{b_2}^2}{864}\right) + {\cal O}(\hat t_D)\\[2 mm]

1683: F_{\textrm{conif.}}^{(4)}&=\frac{1}{1440\,{\hat t_D}^6} - \left( \frac{977520873701}{3402000000000}   

1684: + \frac{162178069379\,b_2}{3888000000} + 

1685:     \frac{5170381469\,{b_2}^2}{2592000} + \frac{490222589\,{b_2}^3}{15552}\right) + 

1686:   {\cal O}(\hat t_D) \cr.

1687: \end{array}

1688: \end{equation}

1689: As explained in sec. \ref{boundaryconditions} we expect this gap structure

1690: to be present at higher genus as in (\ref{thegap}), i.e.

1691: \begin{eqnarray}

1692: F^{(g)}_{\textrm{conifold}}&=&\frac{(-1)^{g-1}B_{2g}}{2g(2g-2)\hat t_D^{2g-2}}+\mathcal{O}(\hat t_D^0).

1693: \end{eqnarray}

1694: If this is true then it will impose $2g-2$ conditions on the

1695: holomorphic ambiguity (\ref{10-14-2.16}). We can further note that

1696: because the prefactor in (\ref{10-14-2.21}) goes like

1697: $\delta^{g-1}$, and the generator $X$ goes like $X\sim

1698: \frac{1}{\delta}$, the terms $a_iX^i$ in holomorphic ambiguity

1699: (\ref{10-14-2.16}) with $i\leq g-1$ do not affect the gap

1700: structure in (\ref{thegap}). Therefore the gap structure fixes the

1701: coefficients $a_i$ of $i=g, \cdots, 3g-3 $ in (\ref{10-14-2.16}),

1702: but not the coefficients $a_i$ with $i\leq g-1$. 

1703: Note that the choice of $b_i$, $i=0,1,2$ does not  affect the gap structure 

1704: at all.  In App. \ref{sectioninvariance} it is proven that the generators 

1705: of the modular forms in $P_g$ do not change under the shift $b_1$. 

1706: The effect of this shift is hence merely a K\"ahler gauge 

1707: transformation.   On the other hand we note that imposing the 

1708: invariance of the gap structure under the $b_2$ shift does not 

1709: give further conditions on the $a_i$, but it does affect 

1710: the subleading expansion.

1711: 

1712: 

1713: 

1714: 

1715: 

1716: \subsection{Fixing the holomorphic ambiguity: a summary of results}

1717: 

1718: Putting all the information together, let us do a counting of

1719: number of unknown coefficients. Originally we have $3g-2$

1720: coefficients in (\ref{10-14-2.16}). The constant map calculation

1721: \cite{Marino:1998, GVI, FP} in A-model large complex structure

1722: modulus limit $\psi\sim \infty$ fixes one constant $a_0$. The

1723: conifold expansion around $\psi\sim 1$ fixes $2g-2$ coefficients

1724: $a_i$ of $i=g, \cdots, 3g-3$, and the orbifold expansion around

1725: $\psi\sim 0$ further fixes $\lceil\frac{3(g-1)}{5}\rceil$

1726: coefficients. So the number of unknown coefficients at genus $g$

1727: is

1728: \begin{eqnarray}

1729: 3g-2-(1+2g-2+\lceil\frac{3(g-1)}{5}\rceil)=[\frac{2(g-1)}{5}].

1730: \label{boundarycountquintic}

1731: \end{eqnarray}

1732: This number is zero for genus $g=2,3$. So we could have computed

1733: the $g=2,3$ topological strings using this information, although

1734: the answers were already known. At $g\geq 4$ there are still

1735: some unknown constants. However, in the A-model expansion when we

1736: rewrite the Gromov-Witten invariants in terms of Gopakumar-Vafa

1737: invariants, one can in principle use the Castelnuovos bound to fix

1738: the $F^{(g)}$ up to genus 51. This will be shown in sect.

1739: \ref{quinticdbrane}.

1740: 

1741: 

1742: 

1743: \section{One-parameter Calabi-Yau spaces with three regular singular points}

1744: We  generalize the analysis for the quintic to other one K\"ahler

1745: parameter Calabi-Yau three-folds, whose mirror $W$ has a

1746: Picard-Fuchs-system with three regular singular points. Note that

1747: this type of CY has been completely classified~\cite{DoranMorgan}

1748: starting from the Riemann-Hilbert reconstruction of the

1749: Picard-Fuchs equations given the monodromies and imposing the

1750: special geometry property as well as integrality conditions on the

1751: solutions of the latter. There are 13 cases whose mirrors can be

1752: realized as hypersurfaces and complete intersections in weighted

1753: projective spaces with trivial fundamental group\footnote{We also 

1754: have some results on the cases with non-trivial fundamental group. 

1755: They are available on request.}, and are well-known

1756: in the mirror symmetry literature. A fourteenth case is

1757: related to a degeneration of a two parameter model as pointed out

1758: in~\cite{KKRS}. Five are obtained as a free discrete orbifold of

1759: the former CY.

1760: 

1761: 

1762: We focus on the $13$ former cases. In the notation of

1763: \cite{Katz:1999} these complete intersections of degree

1764: $(d_1,d_2,\cdots,d_k)$ in weighted projective spaces

1765: $\mathbb{P}^n(w_1,\dots,w_l)$, Calabi-Yau manifolds are

1766: abbreviated as $X_{d_1,d_2,\cdots,d_k}(w_1,\dots,w_l)$. For

1767: example, the familiar quintic manifold is denoted as $X_{5}(1^5)$.

1768: The list of 13 such examples is the following:

1769: \begin{eqnarray}

1770: &&X_{5}(1^5):

1771: \vec{a}=(\frac{1}{5},\frac{2}{5},\frac{3}{5},\frac{4}{5}),~~

1772: X_{6}(1^4,2):

1773: \vec{a}=(\frac{1}{6},\frac{2}{6},\frac{4}{6},\frac{5}{6}),~~

1774:  X_{8}(1^4,4):

1775: \vec{a}=(\frac{1}{8},\frac{3}{8},\frac{5}{8},\frac{7}{8}),\nonumber \\

1776: && X_{10}(1^3,2,5):

1777: \vec{a}=(\frac{1}{10},\frac{3}{10},\frac{7}{10},\frac{9}{10}),~~

1778: X_{3,3}(1^6):

1779: \vec{a}=(\frac{1}{3},\frac{1}{3},\frac{2}{3},\frac{2}{3}),\nonumber \\

1780: && X_{4,2}(1^6):

1781: \vec{a}=(\frac{1}{4},\frac{1}{2},\frac{1}{2},\frac{3}{4}),

1782: ~~X_{3,2,2}(1^7):

1783: \vec{a}=(\frac{1}{3},\frac{1}{2},\frac{1}{2},\frac{2}{3}),\nonumber \\

1784: && X_{2,2,2,2}(1^8):

1785: \vec{a}=(\frac{1}{2},\frac{1}{2},\frac{1}{2},\frac{1}{2}),~~

1786: X_{4,3}(1^5,2):

1787: \vec{a}=(\frac{1}{4},\frac{1}{3},\frac{2}{3},\frac{3}{4}),

1788: \nonumber \\

1789: &&X_{4,4}(1^4,2^2):

1790: \vec{a}=(\frac{1}{4},\frac{1}{4},\frac{3}{4},\frac{3}{4}),~~

1791: X_{6,2}(1^5,3):

1792: \vec{a}=(\frac{1}{6},\frac{1}{2},\frac{1}{2},\frac{5}{6}),

1793: \nonumber \\

1794: &&X_{6,4}(1^3,2^2,3):

1795: \vec{a}=(\frac{1}{6},\frac{1}{4},\frac{3}{4},\frac{5}{6}),~~

1796: X_{6,6}(1^2,2^2,3^2):

1797: \vec{a}=(\frac{1}{6},\frac{1}{6},\frac{5}{6},\frac{5}{6}).

1798: \nonumber

1799: \end{eqnarray}

1800: The examples satisfy the Calabi-Yau condition $\sum_i d_i=\sum_i

1801: w_i$ required by the vanishing of the first Chern class. The

1802: components of the vector $\vec{a}$ specify the Picard-Fuchs

1803: operators for the mirror manifolds with $h^{2,1}=1$,

1804: \begin{eqnarray} \label{PF-10-28-mh}

1805: \{(\psi\frac{\partial}{\partial

1806: \psi})^4-\psi^{-1}\prod_{i=1}^4(\psi\frac{\partial}{\partial

1807: \psi}-a_i)\}\Pi=0.

1808: \end{eqnarray}

1809: The indices of the Picard-Fuchs equation satisfy $\sum_{i=1}^4

1810: a_i=2$ and we have arranged $a_i$ in increasing order for later

1811: convenience.

1812: 

1813: There are three singular points in the moduli space. The maximally

1814: unipotent point is the large complex structure modulus limit

1815: $\psi\sim \infty$ that has three logarithmic solutions for the

1816: Picard-Fuchs equation. The conifold point is $\psi=1$ with three

1817: power series solutions and one logarithmic solution for the

1818: Picard-Fuchs equation. If the indices $a_i$ are not degenerate,

1819: the Picard-Fuchs equation around the orbifold point $\psi=0$ has

1820: four powers series solutions with the leading terms going like

1821: $\psi^{a_i}$. Each degeneration of the indices generates a

1822: logarithmic solution for the Picard-Fuchs equation around the

1823: orbifold point $\psi=0$.

1824: %-----------------------------------------------------------------------

1825: \subsection{The integration of the anomaly equation}

1826: We can straightforwardly generalize the formalism in

1827: \cite{Yamaguchi} to the above class of models.  The mirror map is

1828: normalized as $t= \log(\frac{\prod_i d_i^{d_i}}{\prod_i

1829: w_i^{w_i}}\psi)+\mathcal{O}(\frac{1}{\psi})$, so that the

1830: classical intersection number in the prepotential is $F^{(0)}=

1831: \frac{\kappa}{6}t^3+\cdots$ where $\kappa=\frac{\prod_i

1832: d_i}{\prod_i w_i}$. The generators of the topological string

1833: amplitudes are defined accordingly,

1834: \begin{eqnarray}

1835: &&A_p:=\frac{(\psi\partial\psi)^p

1836: G_{\psi\bar{\psi}}}{G_{\psi\bar{\psi}}},~~B_p:=\frac{(\psi\partial\psi)^p

1837: e^{-K}}{e^{-K}}, ~~(p=1,2,3,\cdots) \nonumber \\ &&

1838: C:=C_{\psi\psi\psi}\psi^3,~~X:=\frac{1}{1-\psi}.

1839: \end{eqnarray}

1840: Here the familiar three point Yukawa coupling is

1841: $C_{\psi\psi\psi}\sim \frac{\psi^{-2}}{1-\psi}$, and as in the

1842: case of the quintic we denote $A:=A_1$ and $B:=B_1$. The generators

1843: satisfy the derivative relations

1844: \begin{eqnarray}

1845: \psi\partial_\psi A_p=A_{p+1}-AA_p, ~~~ \psi\partial_\psi

1846: B_p=B_{p+1}-BB_p,~~~ \psi\partial_\psi X=X(X-1), \nonumber

1847: \end{eqnarray}

1848: and the recursion relations

1849: \begin{eqnarray}

1850: B_4 &=&

1851: (\sum_ia_i)XB_3-(\sum_{i<j}a_ia_j)XB_2+(\sum_{i<j<k}a_ia_ja_k)XB-(\prod_i

1852: a_i)X, \nonumber \\

1853: A_2 &=& -4B_2-2AB-2B+2B^2-2A+2XB+XA-r_0X-1,

1854: \end{eqnarray}

1855: where the first equation can be derived from the Picard-Fuchs

1856: equation (\ref{PF-10-28-mh}) and the second equation is derived

1857: from the special geometry relation up to a holomorphic ambiguity

1858: denoted by the constant $r_0$. One can fix the constant $r_0$ by

1859: expanding the generators around any of the singular points in the

1860: moduli space. For example, as we will explain, the asymptotic

1861: behaviors of various generators around the orbifold point are

1862: $A_p=(a_2-a_1-1)^p+\mathcal{O}(\psi)$,

1863: $B_p=a_1^p+\mathcal{O}(\psi)$, and $X=1+\mathcal{O}(\psi)$, so we

1864: find the constant is

1865: \begin{eqnarray} \label{constantr0-10-28}

1866: r_0=a_1(1-a_1)+a_2(1-a_2)-1.

1867: \end{eqnarray}

1868: The polynomial topological amplitudes $P_g$ are defined by

1869: $P^{(n)}_g:=C^{g-1}\psi^nC^{(g)}_{\psi^n}$, and satisfy the

1870: recursion relations with initial data

1871: \begin{eqnarray}

1872: P^{(3)}_{g=0}&=&1, \nonumber \\

1873: P^{(1)}_{g=1} &=&

1874: (\frac{\chi}{24}-2)B-\frac{A}{2}+\frac{1}{12}(X-1)+\frac{s_1}{2},

1875: \nonumber \\

1876: P^{(n+1)}_g &=& \psi\partial_\psi

1877: P^{(n)}_g-[n(A+1)+(2-2g)(B-\frac{X}{2})]P^{(n)}_g,

1878: \end{eqnarray}

1879: where $\chi$ is the Euler character of the Calabi-Yau space and

1880: $s_1=2 c-\frac{5}{6}$ is a constant that can be fixed by the

1881: second Chern class of the Calabi-Yau; see appendix A. We provide

1882: the list of constants for the 13 cases of one-parameter Calabi-Yau

1883: in the following table.

1884: 

1885: \begin{table}

1886: \begin{centering}

1887: \begin{tabular}{|r|r|r|r|r|r|r|r|}

1888: \hline CY &$X_5(1^5)$ & $X_6(1^4,2)$ & $X_8(1^4,4)$ &

1889: $X_{10}(1^3,2,5)$ & $X_{3,3}(1^6)$

1890: \\ \hline

1891: $\chi$ & $-200$ & $-204$ & $-296$ & $-288$ & $-144$ \\

1892: \hline $s_1$ & $\frac{10}{3}$ & $\frac{8}{3}$ & $\frac{17}{6}$ &

1893: $2$ & $\frac{11}{3}$ \\ \hline CY & $X_{4,2}(1^6)$ &

1894: $X_{3,2,2}(1^7)$ &$X_{2,2,2,2}(1^8)$ & $X_{4,3}(1^5,2)$ &

1895: $X_{4,4}(1^4,2^2)$  \\ \hline $\chi$ & $-176$ & $-144$  & $-128$ &

1896: $-156$ & $-144$  \\ \hline $s_1$  & $\frac{23}{6}$ &

1897: $\frac{25}{6}$ & $\frac{9}{2}$ & $\frac{19}{6}$ & $\frac{5}{2}$ \\

1898: \hline CY & $X_{6,2}(1^5,3)$ & $X_{6,4}(1^3,2^2,3)$ &

1899: $X_{6,6}(1^2,2^2,3^2)$ & &\\ \hline  $\chi$ & $-256$ & $-156$ &

1900: $-120$  &

1901: &\\ \hline  $s_1$ & $\frac{7}{2}$ & $\frac{11}{6}$ & $1$  & & \\

1902: \hline

1903: \end{tabular} \caption{Euler numbers and the constant $s_1$}

1904: \end{centering}

1905: \end{table}

1906: 

1907: After changing variables to a convenient basis from

1908: $(A,B,B_2,B_3,X)$ to $(u,v_1,v_2,v_3,X)$ as the following:

1909: \begin{eqnarray}

1910: &&B=u,~~A=v_1-1-2u,~~B_2=v_2+uv_1, \nonumber \\ &&

1911: B_3=v_3-uv_2+uv_1X-(r_0+1)u X. \nonumber

1912: \end{eqnarray}

1913: One can follow \cite{Yamaguchi} and use the BCOV formalism to show

1914: that $P_g$ is a degree $3g-3$ inhomogeneous polynomial of

1915: $v_1,v_2,v_3, X$ with the assigned degrees $1,2,3,1$ for $v_1,

1916: v_2, v_3, X$ respectively. The BCOV holomorphic anomaly equation

1917: becomes

1918: \begin{eqnarray}

1919: (\frac{\partial}{\partial{v_1}}-u\frac{\partial }{\partial

1920: {v_2}}-u(u+X)\frac{\partial}{\partial{v_3}})P_g

1921: =-\frac{1}{2}(P^{(2)}_{g-1}+\sum_{r=1}^{g-1}P^{(1)}_rP^{(1)}_{g-r}).

1922: \end{eqnarray}

1923: As in the case of the quintic, the holomorphic anomaly equation

1924: determines the polynomial $P_g$ up to a holomorphic ambiguity

1925: \begin{eqnarray}

1926: f=\sum_{i=0}^{3g-3} a_iX^i.

1927: \end{eqnarray}

1928: 

1929: \subsection{The boundary behaviour}

1930: The constant term $a_0$ is fixed by the known constant map

1931: contribution in the A-model expansion

1932: \begin{eqnarray}

1933: F^{g}_{\textrm{A-model}}=\lim _{t\rightarrow \infty}

1934: \omega_0^{2(g-1)}(\frac{1-\psi}{\kappa \psi})^{g-1}P_g,

1935: \end{eqnarray}

1936: where $\omega_0$ is the power series solution in the large complex

1937: structure modulus limit.

1938: 

1939: The main message here is that structure of the conifold expansion

1940: is universal. For all cases of Calabi-Yau spaces, the Picard-Fuchs

1941: equation around $z=\psi-1$ has four solutions that go like

1942: $\omega_0= 1+\mathcal{O}(z)$, $\omega_1= z+\mathcal{O}(z^2)$,

1943: $\omega_2=z^2+\mathcal{O}(z^3)$, and

1944: $\omega_4=\omega_1\log(z)+\mathcal{O}(z^4)$. We define a dual

1945: mirror map $t_D=\frac{\omega_1}{\omega_0}=z+\mathcal{O}(z^2)$,

1946: expand the topological string amplitudes in terms of $t_D$, and

1947: impose the following gap structure in the conifold expansion:

1948: \begin{eqnarray}

1949: F^{(g)}_{\textrm{conifold}} &=& \lim_{\bar{z}\rightarrow

1950: 0}\omega_0^{2(g-1)}(\frac{1-\psi}{\psi})^{g-1}P_g \nonumber \\

1951: &=&

1952: \frac{(-1)^{g-1}B_{2g}}{2g(2g-2)t_D^{2g-2}}+\mathcal{O}(t_D^0).

1953: \end{eqnarray}

1954: This fixes $2g-2$ coefficients in the holomorphic ambiguity.

1955: 

1956: On the other hand, we discover a rich variety of singularity

1957: structures around the orbifold point $\psi=0$. A natural

1958: symplectic basis of solutions of the Picard-Fuchs equation

1959: (\ref{PF-10-28-mh}) is picked out by the fractional powers of the

1960: leading terms. As in the case of quintic for our purpose we only

1961: need the first two solutions $\omega_0$, $\omega_1$. The leading

1962: behaviors are

1963: \begin{enumerate}

1964: \item If $a_2>a_1$, then we have

1965: $\omega_0=\psi^{a_1}(1+\mathcal{O}(\psi))$,

1966: $\omega_1=\psi^{a_2}(1+\mathcal{O}(\psi))$.

1967: \item If $a_2=a_1$, then we have

1968: $\omega_0=\psi^{a_1}(1+\mathcal{O}(\psi))$,

1969: $\omega_1=\omega_0(\log(\psi)+\mathcal{O}(\psi))$.

1970: \end{enumerate}

1971: In both cases we can define a mirror map around the orbifold point

1972: as $s=\frac{\omega_1}{\omega_0}$. Using the behaviors of

1973: K\"ahler potential $e^{-K}\sim \omega_0$ and metric

1974: $G_{\psi\bar{\psi}}\sim

1975: \partial_\psi s$ in the

1976: holomorphic limit, we can find the asymptotic expansion of various

1977: generators. In both cases $a_2>a_1$ and $a_2=a_1$, the leading

1978: behaviors are non-singular:

1979: \begin{eqnarray}

1980: A_p &=& (a_2-a_1-1)^p+\mathcal{O}(\psi) \nonumber \\

1981: B_p &=& a_1^p+\mathcal{O}(\psi).

1982: \end{eqnarray}

1983: This can be used to fix a holomorphic ambiguity

1984: (\ref{constantr0-10-28}) relating $A_2$ to other generators. On

1985: the other hand, since the constant (\ref{constantr0-10-28})  can

1986: be also derived at other singular points of the moduli space, this

1987: also serves as a consistency check that we have chosen the correct

1988: basis of solutions $\omega_0$, $\omega_1$ at the orbifold point.

1989: 

1990: The orbifold expansion of the topological string amplitudes are

1991: \begin{eqnarray}

1992: F^{(g)}_{\textrm{orbifold}}=\lim_{\bar{\psi}\rightarrow 0

1993: }\omega_0^{2(g-1)}(\frac{1-\psi}{\psi})^{g-1}P_g\sim

1994: \frac{P_g}{\psi^{(1-2a_1)(g-1)}}

1995: \end{eqnarray}

1996: We can expand the polynomial $P_g$ around the orbifold point

1997: $\psi=0$. Generically, $P_g$ is power series of $\psi$ starting

1998: from a constant term, so $F^{(g)}_{\textrm{orbifold}}\sim

1999: \frac{1}{\psi^{(1-2a_1)(g-1)}}$. Since $\sum_ia_i=2$ and $a_1$ is

2000: the smallest, we know $a_1\leq \frac{1}{2}$ and the topological

2001: string amplitude around orbifold point is generically singular.

2002: Interestingly, we find the singular behavior of the topological

2003: strings around the orbifold point is not universal and falls into

2004: four classes:

2005: 

2006: \begin{enumerate}

2007: \item This class includes all 4 cases of hypersurfaces and 4 other cases of complete intersections.

2008: They are the Calabi-Yau spaces $X_5(1^5)$, $X_6(1^4,2)$,

2009: $X_8(1^4,4)$, $X_{10}(1^3,2,5)$, $X_{3,3}(1^6)$,

2010: $X_{2,2,2,2}(1^8)$, $X_{4,4}(1^4,2^2)$, $X_{6,6}(1^2,2^2,3^2)$.

2011: For these cases the singularity at the orbifold point is cancelled

2012: by the series expansion of the polynomial $P_g$. The requirement

2013: of cancellation of singularity in turn imposes

2014: \begin{eqnarray}

2015: \lceil(1-2a_1)(g-1)\rceil

2016: \end{eqnarray}

2017: conditions on the holomorphic ambiguity in $P_g$. Taking into

2018: account the A-model constant map condition and the boundary

2019: condition from conifold expansion, we find the number of unknown

2020: coefficients at genus $g$ is

2021: \begin{eqnarray} \label{number1-10-28}

2022: [2a_1(g-1)].

2023: \end{eqnarray}

2024: Notice for the Calabi-Yau $X_{2,2,2,2}(1^8)$ the cancellation is

2025: trivial since in this case $a_1=\frac{1}{2}$ and topological

2026: strings around the orbifold point are generically non-singular, so

2027: this does not impose any boundary conditions.

2028: 

2029: 

2030: \item This class of Calabi-Yau spaces include $X_{4,2}(1^6)$ and

2031: $X_{6,2}(1^5,3)$. We find the singularity at the orbifold point is

2032: not cancelled by the polynomial $P_g$, but it has a gap structure

2033: that closely resembles the conifold expansion. Namely, when we

2034: expand the topological strings in terms of the orbifold mirror map

2035: $s=\frac{\omega_1}{\omega_0}$, we find

2036: \begin{eqnarray}

2037: F^{(g)}_{\textrm{orbifold}}=\frac{C_g}{s^{2(g-1)}}+\mathcal{O}(s^0)

2038: \end{eqnarray}

2039: where in our normalization convention, the constant is

2040: $C_g=\frac{B_{2g}}{2^{5g-4}(g-1)g}$ for the $X_{4,2}(1^6)$ model,

2041: and $C_g=\frac{B_{2g}}{3^{3(g-1)}4(g-1)g}$ for the

2042: $X_{6,2}(1^5,3)$ model. Since the mirror map goes like

2043: $s\sim \psi^{a_2-a_1}$, and the $P_g$ is a power series of

2044: $\psi$, this imposes

2045: \begin{eqnarray}

2046: \lceil 2(a_2-a_1)(g-1) \rceil

2047: \end{eqnarray}

2048: conditions on the holomorphic ambiguity in $P_g$. In both cases of

2049: $X_{4,2}(1^6)$ and $X_{6,2}(1^5,3)$, we have $a_2=\frac{1}{2}$, so

2050: the number of un-fixed coefficients is the same as the models in

2051: the first class, namely

2052: \begin{eqnarray}

2053: [2a_1(g-1)].

2054: \end{eqnarray}

2055: 

2056: \item This class of Calabi-Yau spaces include only the Calabi-Yau $X_{3,2,2}(1^7)$.

2057: For this model, the singularity around the orbifold point does not

2058: cancel, so there is no boundary condition imposed on the

2059: holomorphic ambiguity in $P_g$. At genus $g$ we are simply left

2060: with $g-1$ unknown coefficients.

2061: 

2062: \item This class of Calabi-Yau spaces include $X_{4,3}(1^5,2)$ and

2063: $X_{6,4}(1^3,2^2,3)$. For this class of models, the singularity at

2064: the orbifold point is partly cancelled. Specifically, we find

2065: $\frac{P_g}{\psi^{(1-2a_1)(g-1)}}$ is not generically regular at

2066: $\psi\sim 0$, but $\frac{P_g}{\psi^{(1-2a_2)(g-1)}}$ is always

2067: regular. This then imposes

2068: \begin{eqnarray}

2069: \lceil(1-2a_2)(g-1)\rceil

2070: \end{eqnarray}

2071: conditions on holomorphic ambiguity in $P_g$, and the number of

2072: unknown coefficients at genus $g$ is now

2073: \begin{eqnarray}

2074: [2a_2(g-1)].

2075: \end{eqnarray}

2076: 

2077: \end{enumerate}

2078: 

2079: It looks like we have the worst scenarios in the two models

2080: $X_{2,2,2,2}(1^8)$ and $X_{3,2,2}(1^7)$, where we essentially get

2081: no obvious boundary conditions at the orbifold point $\psi=0$.

2082: However after a closer examination, we find some patterns in the

2083: leading coefficients of the orbifold expansion as the followings.

2084: In our normalization convention, we find the leading constant

2085: coefficients of $X_{2,2,2,2}(1^8)$ model is

2086: \begin{eqnarray}

2087: F^{(g)}_{\textrm{orbifold}}=\frac{(21\cdot2^{2g-2}-5)(-1)^{g}B_{2g}B_{2g-2}}{2^{2g-3}g(2g-2)(2g-2)!}

2088: +\mathcal{O}(\psi)

2089: \end{eqnarray}

2090: whereas for $X_{3,2,2}(1^7)$ model the leading coefficients are

2091: \begin{eqnarray} \label{twoparticles-12-01}

2092: F^{(g)}_{\textrm{orbifold}}=

2093: \frac{(7\cdot2^{2g-2}-1)B_{2g}}{2^{4g-3}3^{2g-2}(g-1)g}

2094: \frac{1}{s^{2g-2}}+\mathcal{O}(\frac{1}{s^{2g-8}})

2095: \end{eqnarray}

2096: These leading coefficients provide one more boundary condition for

2097: the models $X_{2,2,2,2}(1^8)$ and $X_{3,2,2}(1^7)$, although this

2098: is not much significant at large genus. On the other hand, we

2099: observe the leading coefficients of $X_{2,2,2,2}(1^8)$ model are

2100: similar to the constant map contribution of Gromov-Witten

2101: invariants except the factor of $(21\cdot2^{2g-2}-5)$, whereas the

2102: leading coefficients of $X_{3,2,2}(1^7)$ model are similar to the

2103: conifold expansion except the factor of $(7\cdot2^{2g-2}-1)$.

2104: These non-trivial factors can not be simply removed by a different

2105: normalization of variables and therefore contain useful

2106: information. In fact in the latter case of $X_{3,2,2}(1^7)$ model,

2107: the factor of $(7\cdot2^{2g-2}-1)$ will motivate our physical

2108: explanations of the singularity in a moment.

2109: 

2110: 

2111: As shown in \cite{Katz:1999}, see also sec.

2112: \ref{symplecticinvariants}, for a fixed genus $g$, the

2113: Gopakumar-Vafa invariants $n^g_d$ are only non-vanishing when the

2114: degree $d$ is bigger than $a_g$, where $a_g$ is a model dependent

2115: number with weak genus dependence, in particular for large $g$ one

2116: has $d_{\textrm{min}}-1 = a_g \sim \sqrt{g}$. So long as the

2117: number of zeros in the low degree Gopakumar-Vafa invariants are

2118: bigger than the number of unknown coefficients that we determine

2119: above using all available boundary conditions, we have a

2120: redundancy of data to compute the topological strings recursively

2121: genus by genus, and are able to make non-trivial checks of our

2122: computations. For all of the 13 cases of one-parameter Calabi-Yau

2123: spaces, we are able to push the computation to very high genus. So

2124: far our calculations are limited only by the power of our

2125: computational facilities.

2126: 

2127: We now propose a ``phenomenological'' theory of the singularity

2128: structures at the point with rational branching. Our underlying philosophy is

2129: that a singularity of $F^{(g)}$ in the moduli space can only be

2130: generated if there are charged massless states near this point of

2131: moduli space. This is already familiar from the behaviors at

2132: infinity $\psi=\infty$ and conifold point $\psi=1$. At infinity

2133: $\psi=\infty$ the relevant charged states are massive $D2-D0$

2134: brane bound states and therefore $F^{(g)}$ are regular, whereas at

2135: the conifold point there is a massless charged state from a $D3$

2136: brane wrapping a vanishing 3-cycle, and this generates the gap

2137: like singularity at the conifold point as we have explained. We

2138: should now apply this philosophy to the much richer behavior at

2139: the orbifold point $\psi=0$. We discuss the four classes of models

2140: in the same order as mentioned above.

2141: \begin{enumerate}

2142: \item We argue for this class of models the $F^{(g)}$ are regular

2143: at the orbifold point because there is no massless charged state.

2144: A necessary condition would be that the mirror map parameter

2145: $s$ is non-zero at the orbifold point, since as we have

2146: learned there are D-branes wrapping cycles whose charge and mass

2147: are measured by $s$. This is clear for the complete

2148: intersection cases, namely $X_{3,3}(1^6)$, $X_{2,2,2,2}(1^8)$,

2149: $X_{4,4}(1^4,2^2)$, $X_{6,6}(1^2,2^2,3^2)$, because for these

2150: models the first two indices of the Picard-Fuchs equation is

2151: degenerate $a_2=a_1$, therefore the mirror map goes like

2152: \begin{eqnarray}

2153: s\sim \log{\psi}\rightarrow \infty

2154: \end{eqnarray}

2155: and we see that the D-branes are very massive and generate

2156: exponentially small corrections just like the situation at

2157: infinity $\psi=\infty$. As for the hypersurface cases $X_5(1^5)$,

2158: $X_6(1^4,2)$, $X_8(1^4,4)$, $X_{10}(1^3,2,5)$, we comment that

2159: although naively the mirror map goes like $s\sim

2160: \psi^{a_2-a_1}\rightarrow 0 $, there is a change of basis under

2161: which the generators are invariant as explained in Appendix

2162: \ref{sectioninvariance}, and which could make the periods goes

2163: like $\omega_0\sim\omega_1\sim \psi^{a_1}$, therefore the mirror

2164: map becomes finite at the orbifold point. This is also consistent

2165: with the basis at orbifold point we obtained by analytic

2166: continuation from infinity $\psi=\infty$. In fact we checked that the regularity 

2167: of $F^{(g)}$ is not affected when we take $s$ to be the ratio of generic 

2168: arbitrary linear combinations of the periods $\omega_i$, $i=0,1,2,3$.  

2169: 

2170: \item We argue this class of models have a conifold like structure

2171: because of the same mechanism we have seen for the conifold point

2172: $\psi=1$. This is consistent with the fact that the degenerate

2173: indices in these cases, e.g. models $X_{4,2}(1^6)$ and

2174: $X_{6,2}(1^5,3)$, are the middle indices, namely we have

2175: $a_2=a_3=\frac{1}{2}$. Therefore the Picard-Fuchs equation

2176: constrains one of periods to be a power series proportional to

2177: \begin{eqnarray}

2178: \omega_1\sim \psi^{\frac{1}{2}}

2179: \end{eqnarray}

2180: Since there is another period that goes like $\omega_0\sim

2181: \psi^{a_1}$, the mirror map goes like $s\sim

2182: \psi^{\frac{1}{2}-a_1}\rightarrow 0$ and it is not possible to

2183: change the basis in a way such that the mirror map is finite.

2184: Integrating out a charged nearly massless particle generates the

2185: gap like conifold singularity as we have explained.

2186: 

2187: \item For the model $X_{3,2,2}(1^7)$ we argue there are two

2188: massless states near the orbifold point. Since the middle indices

2189: also degenerate $a_2=a_3=\frac{1}{2}$, we can apply the same

2190: reasoning from the previous case and infer that the mirror map

2191: goes like $s\sim \psi^{\frac{1}{6}}$ and there must be at

2192: least one charged massless state from a D3 brane wrapping

2193: vanishing 3-cycle. However the situation is now more complicated.

2194: We do not find a gap structure in the expansion of

2195: $F^{(g)}_{\textrm{orbifold}}$, and the leading coefficients differ

2196: from the usual conifold expansion by a factor of $(7\cdot

2197: 2^{2g-2}-1)$ as observed in (\ref{twoparticles-12-01}). These can

2198: be explained by postulating that there are two massless particles

2199: in this case whose masses are $m$ and $2m$. This fits nicely with

2200: the $2^{2g-2}$ power in the intriguing factor of $(7\cdot

2201: 2^{2g-2}-1)$, and also explains the absence of gap structure by

2202: the possible interactions between the two massless particles.

2203: 

2204: \item Finally, we discuss the cases of models $X_{4,3}(1^5,2)$ and

2205: $X_{6,4}(1^3,2^2,3)$. The indices $a_i$ are not degenerate in

2206: these cases. What makes these models different from the

2207: hypersurface cases is the fact that the ratio $\frac{a_2}{a_1}$ is

2208: now not an integer in these cases. This makes it difficult to

2209: change the basis such that the mirror map is finite. We conjecture

2210: that the mirror map indeed goes like $s\sim \psi^{a_2-a_1}$ and

2211: therefore there exist massless particle(s) whose masses are

2212: proportional to $s$ and who are responsible for generating the

2213: singularity of $F^{(g)}$ at the orbifold point. This is consistent

2214: with our analytic continuation analysis in Appendix

2215: \ref{appendixD-12-06}. We note that a necessary non-trivial

2216: consequence of this scenario would be that the

2217: $F^{(g)}_{\textrm{orbifold}}$ is no more singular than

2218: $\frac{1}{s^{2g-2}}$, i.e. the product

2219: \begin{eqnarray}

2220: s^{2g-2}F^{(g)}_{\textrm{orbifold}}\sim

2221: \frac{P_g}{\psi^{(1-2a_2)(g-1)}}

2222: \end{eqnarray}

2223: should be regular. This is precisely what we observe

2224: experimentally.

2225: 

2226: 

2227: 

2228: 

2229: \end{enumerate}

2230: 

2231: 

2232: \section{Symplectic invariants at large radius}

2233: \label{symplecticinvariants} The coefficients of the large radius

2234: expansion of the $\cF^{(g)}=\lim_{{\bar t}\rightarrow \infty}

2235: F^{(g)}(t,\bar t)$  have an intriguing conjectural interpretation

2236: as symplectic invariants of $M$. First of all we have

2237: \begin{equation}

2238: \cF^{(g)}(q)=\sum_{\beta} r_\beta^{(g)} q^{\beta}\ ,

2239: \end{equation}

2240: where $r^{(g)}_\beta\in \mathbb{Q}$ are the {\sl Gromow-Witten

2241: invariants} of holomorphic maps. Secondly the {\sl Gopakumar-Vafa

2242: invariants}~\cite{GVII} count the cohomology of the $D_0-D_2$

2243: bound state moduli space, see also~\cite{Katz:1999},  and are

2244: related to the

2245: \begin{equation}

2246: \begin{array}{rcl}

2247: \cF(\lambda,t)&=&\ds{\sum_{g=0}^\infty \lambda^{2 g-2} \cF^{(g)}}(t)\\

2248: &=&\ds{{c(t)\over \lambda^2}+l(t)+\sum_{g=0}^\infty \sum_{\beta\in H_2(M,\ZZ)}\sum_{m=1}^\infty n_{\beta}^{(g)} {1\over m}

2249: \left(2 \sin {m \lambda \over 2}\right)^{2 g-2}  q^{\beta m}}\ .

2250: \end{array}

2251: \label{schwingerloopd2d0}

2252: \end{equation}

2253: Here $c(t)$ and $l(t)$ are some cubic and linear polynomials in $t$, which follow

2254: from the leading behaviour of $\cF^{(0)}$ and $\cF^{(1)}$  as explained in (\ref{prepotential})

2255: and (\ref{F1}).

2256: With $q_\lambda=e^{i\lambda}$ we can write a product form\footnote{Here we dropped the $\exp({c(t)\over \lambda^2}+l(t))$

2257: factor of the classical terms  at genus $0,1$.} for the partition function $Z^{\rm hol}=\exp(\cF^{\rm hol})$

2258: \begin{equation}

2259: Z^{\rm hol}_{\rm GV}(M,\lambda,q)=\prod_{\beta}\left[

2260: \left(\prod_{r=1}^\infty (1-q_\lambda^r q^\beta)^{r n_\beta^{(0)}}\right)

2261: \prod_{g=1}^\infty \prod_{l=0}^{2g-2}(1-q_\lambda^{ g-l-1} q^\beta)^{(-1)^{g+r}

2262: \left(2 g-2\atop l\right) n_\beta^{(g)}}\right]\

2263: \label{zhol}

2264: \end{equation}

2265: in terms of the {\sl Gopakumar-Vafa invariants} $n_\beta^{(g)}$. Based on the partition

2266: functions there is a conjectural relation of the latter to the {\sl Donaldson-Thomas invariants}

2267: $\tilde n_\beta^{(g)}$, which are invariants of the moduli space $I_k(M,\beta)$ of

2268: ideal sheaves $\cI$ on $M$.  Defining $Z^{\rm hol}_{\rm DT}(M,q_{\lambda},q)=

2269: \sum_{\beta,k\in {\ZZ}} {\tilde n}^{(k)}_{\beta} q_\lambda^k q^\beta$ one expects~\cite{DTGW}

2270: \begin{equation}

2271: Z^{\rm hol}_{\rm GV}(M,q_{\lambda},q)M(q_\lambda)^{\frac{\chi(M)}{2}}=Z^{\rm hol}_{\rm DT}(M,-q_{\lambda},q)\ ,

2272: \label{eq:GVDT}

2273: \end{equation}

2274: where the McMahon function  is defined as $M(q_\lambda):=\prod_{n\ge 0} {1\over (1-q_\lambda^n)^n}$.

2275: We will give below the information about the ${\cal F}^{(g)}$ in terms of the

2276: {\sl Gopakumar-Vafa invariants} and give a  more detailed account of the data

2277: of  the symplectic invariants on the webpage \cite{webpage}.

2278: 

2279: 

2280: \begin{table}

2281: \begin{centering}

2282: \begin{tabular}{|r|rrrrrr|}

2283: \hline

2284: g &d=1 &d=2 &d=3 &d=4 &d=5 &d=6  \\

2285: \hline

2286: \, 0& 2875& 609250& 317206375& 242467530000& \!229305888887625&\! 248249742118022000 \\

2287: 1& 0    &     0&    609250&   3721431625&  12129909700200&  31147299733286500 \\

2288: 2& 0    &     0&         0&       534750&     75478987900&    871708139638250 \\

2289: 3& 0    &     0&         0&         8625&       -15663750&     3156446162875  \\

2290: 4& 0    &     0&         0&            0&           49250&       -7529331750 \\

2291: 5& 0    &     0&         0&            0&            1100&         -3079125 \\

2292: 6& 0    &     0&         0&            0&              10&            -34500 \\

2293: 7& 0    &     0&         0&            0&               0&                0  \\

2294: \hline

2295: \end{tabular}

2296: \vskip 2 pt

2297: \begin{tabular}{|r|rrr|}

2298: \hline

2299: g &d=7 & d=8 &d=9 \\

2300: \hline

2301: 0& \!\!\! 295091050570845659250& \!\!\!\!\!\! 375632160937476603550000& \!\!\!\!\!\! 503840510416985243645106250\\

2302: 1& \!\!\! 71578406022880761750& \!\!\!\!\!\! 154990541752961568418125& \!\!\!\!\!\! 324064464310279585657008750 \\

2303: 2& 5185462556617269625& 22516841063105917766750& 81464921786839566502560125 \\

2304: 3& 111468926053022750& 1303464598408583455000& 9523213659169217568991500 \\

2305: 4& 245477430615250& 25517502254834226750& 507723496514433561498250 \\

2306: 5& -1917984531500& 46569889619570625& 10280743594493108319750 \\

2307: 6& 1300955250& -471852100909500& 30884164195870217250 \\

2308: 7& 4874000& 2876330661125&-135197508177440750 \\

2309: 8& 0& -1670397000& 1937652290971125 \\

2310: 9& 0& -6092500& -12735865055000\\

2311: 10& 0& 0& 18763368375\\

2312: 11& 0& 0& 5502750\\

2313: 12& 0& 0& 60375\\

2314: 13& 0& 0&  0\\

2315: \hline

2316: \end{tabular}

2317: \begin{tabular}{|r|rr|}

2318: \hline

2319: g &d=10 & d=11 \\

2320: \hline

2321: 0& \quad \,\,\,\,704288164978454686113488249750&  \quad \,\,\,\,\,   1017913203569692432490203659468875\\

2322: 1& 662863774391414096742406576300& 1336442091735463067608016312923750 \\

2323: 2& 261910639528673259095545137450& 775720627148503750199049691449750 \\

2324: 3& 52939966189791662442040406825 & 245749672908222069999611527634750 \\

2325: 4& 5646690223118638682929856600  &44847555720065830716840300475375\\

2326: 5& 302653046360802682731297875   & 4695086609484491386537177620000\\

2327: 6& 6948750094748611384962730     &267789764216841760168691381625\\

2328: 7& 40179519996158239076800       &7357099242952070238708870000\\

2329: 8& -25301032766083303150         &72742651599368002897701250\\

2330: 9& 1155593062739271425           &140965985795732693440000\\

2331: 10& -17976209529424700           &722850712031170092000 \\

2332: 11& 150444095741780              &-18998955257482171250\\

2333: 12& -454092663150                & 353650228902738500 \\

2334: 13&   50530375                   &-4041708780324500\\

2335: 14&  -286650                     &22562306494375\\

2336: 15&   -5700                      &-29938013250\\

2337: 16&    -50                       & -7357125\\

2338: 17&     0                        &  -86250\\

2339: 18&     0                        &      0\\

2340: \hline

2341: \end{tabular}

2342: \caption{BPS invariants $n^g_d$ on the Quintic hypersurface in $\mathbb{P}^4$. See also Table 3.}

2343: \end{centering}

2344: \end{table}

2345: 

2346: 

2347: 

2348: \subsection{Castelnuovo's theory and the cohomology of the BPS state moduli space}

2349: 

2350: Let us give checks of the numbers using techniques of algebraic

2351: geometry and the description of the BPS moduli space and its

2352: cohomology developed in \cite{GVII,Katz:1999}. The aim is to check

2353: the gap condition in various geometric settings, namely

2354: hypersurfaces and complete intersections in (weighted) projective

2355: spaces discussed before. According to \cite{GVII,Katz:1999} the

2356: BPS number of a given charge, i.e. degree $d$, can be calculated

2357: from cohomology of the moduli  space $\hat {\cal M}$ of a

2358: $D_2-D_0$ brane system. The latter is the fibration of the

2359: Jacobian $T^{2 \tilde g}$ of a genus ${\tilde g}$ curve over its

2360: moduli space of deformations ${\cal M}$ . Curves of arithmetic

2361: genus $g<{\tilde g}$ are degenerate curves, in the simplest case

2362: with $\delta={\tilde g}-g$ nodes. Their BPS numbers are calculated

2363: using the Euler numbers of relative Hilbert schemes ${\cal

2364: C}^{(i)}$ of the universal curve (${\cal C}^{(0)}={\cal M}$,

2365: ${\cal C}^{(1)}$ is the universal curve, etc) in simple situations

2366: as follows:

2367: \begin{equation}

2368: \begin{array}{rl}

2369: n_d^{g}&={\displaystyle n_d^{{\tilde g} -\delta}=(-1)^{{\rm

2370: dim}({\cal M})+\delta}\sum_{p=0}^\delta b( {\tilde g}-p,\delta-p)

2371: e({\cal C}^{(p)})},

2372: \\[ 3 mm]

2373: b(g,k)&={2\over k!} (g-1)\prod_{i=1}^{k-1} (2 g - (k+2) +i),\qquad   b(g,0)=0.

2374: \end{array}

2375: \label{bpsformula}

2376: \end{equation}

2377: 

2378: As explained in \cite{Katz:1999} curves in  projective spaces

2379: meeting the  quintic are either plane curves in $\mathbb{P}^2$,

2380: curves in $\mathbb{P}^3$, or $\mathbb{P}^4$. In all case one gets

2381: from Castelnuovo theory a bound on $g$, which grows for large $d$

2382: like  $g(d)\sim d^2$. For a detailed exposition of curves in

2383: projective space see~\cite{Harris}. Using this information one can

2384: determine which curves above is realized and contributes to the

2385: BPS numbers. These statements generalize to the  hypersurfaces and

2386: complete intersections with one K\"ahler modulus in weighted

2387: projective spaces. In particular the qualitative feature $g(d)\sim

2388: d^2$ of the bound for large $d$ carries over. We note for

2389: later convenience that to go from a smooth curve of genus $\tilde

2390: g$ to a curve with arithmetic genus $g=\tilde g-\delta$ by

2391: enforcing $\delta$ nodes we get from (\ref{bpsformula})

2392: \begin{equation}

2393: \begin{array}{rl}

2394: n^{\tilde g-1}_d&=(-1)^{{\rm dim}({\cal M})+1}\left( e({\cal C})+ (2\tilde g-2) e(\cal M)\right)\\

2395: n^{\tilde g-2}_d&=(-1)^{{\rm dim}({\cal M})+1}\left(e({\cal C}^{(2)})+(2 \tilde g-4) e({\cal C})+{1\over 2}(2 \tilde g-2) (2 \tilde g-5)e(\cal M)\right)\ . \\

2396: \end{array}

2397: \label{BPSspecial}

2398: \end{equation}

2399: 

2400: %----------------------------------------------------------------------

2401: \subsection{D-branes on the quintic}

2402: \label{quinticdbrane}

2403: One consequence of our global understanding of the $F^{(g)}$

2404: is that we can make detailed statements about the `number' of

2405: $D$-brane states for the quintic at large radius.

2406:  We focus on $d=5$, because

2407: there is a small numerical flaw in the analysis of

2408: \cite{Katz:1999}, while the right numerics confirms the gap

2409: structure quite significantly. In this case the complete

2410: intersection with multidegree $(1,1,5)$ is a plane curve with

2411: genus $\tilde g=(d-1)(d-2)/2=6$, while the other possibilities

2412: have at most genus $g=2$. Curves of $(g=\tilde g,d)=(6,5)$ are

2413: therefore smooth plane curves with $\delta=0$ and according to

2414: (\ref{bpsformula}) their BPS number is simply $n^g_d=(-1)^{{\rm

2415: dim} \cal M} e({\cal M})$. Since $d=5$ their moduli space is

2416: simply the moduli space of $\mathbb{P}^2$'s in $\mathbb{P}^4$,

2417: ${\cal M}$ is the Grassmannian\footnote{The space of

2418: $\mathbb{P}^k$'s in  $\mathbb{P}^n$ , which we call

2419: ${\mathbb{G}}(k,n)$, is also the space of $k+1$ complex

2420: dimensional subspaces in an $n+1$ dimensional complex vector

2421: space, which is often alternatively denote as $G(k+1,n+1)$.}

2422: ${\mathbb{G}}(2,4)$. Grassmannians ${\mathbb{G}}(k,n)$ have

2423: dimensions $(k+1)(n-k)$ and their Euler number can be calculated

2424: most easily by counting toric fixed points to be

2425: $\chi({\mathbb{G}}(k,n))=\left(n+1\atop k+1\right)$.  We get

2426: $n^6_5=(-1)^6 10=10$.

2427: 

2428: For the $(g,d)=(5,5)$ curves we have to determine the Euler number

2429: of the universal curve ${\cal C}$, which is a fibration $\pi:{\cal

2430: C} \rightarrow {\cal M}$ over ${\cal M}$. To get  an geometric

2431: model for ${\cal C}$ we consider the projection $\tilde \pi:{\cal

2432: C}\rightarrow X$. The fiber over a point $p\in X$ is the  set of

2433: $\IP^2$'s in $\IP^4$ which contain the point $p$. This is described

2434: as the space of $\mathbb{P}^1$'s in $\mathbb{P}^3$ i.e.

2435: ${\mathbb{G}}(1,3)$ with~\footnote{${\mathbb{G}}(1,3)$ is Plucker

2436: embedded in $\mathbb{P}^5$ as a quadric (degree 2). From the

2437: adjunction formula we also  get $\chi({\mathbb{G}}(1,3))=6$.}

2438: $\chi({\mathbb{G}}(1,3))=6$. As the fibration $\tilde \pi$ is

2439: smooth we obtain $e({\cal C})=\chi(X)

2440: \chi({\mathbb{G}}(1,3))=-200\cdot 6=-1200$. Applying  now

2441: (\ref{BPSspecial}) we get $n^{5}_5=(-1)^5(-1200+ (2 \cdot 6-2)

2442: 10)=1100$.

2443: 

2444: 

2445: The calculation of $n^4_4$ requires the  calculation  of  $e({\cal C}^{(2)})$. The model

2446: for ${\cal C}^{(2)}$ is constructed from the fibration  $\hat \pi:{\cal C}^{(2)}\rightarrow {\rm Hilb}^2(X)$

2447: as follows. A point in ${\rm Hilb}^2(X)$ are either two distinct points or one point of multiplicity $2$ with  distinct

2448: tangent direction.  In both cases the

2449: fiber over $P \in {\rm Hilb}^2(X)$ is an $\mathbb{P}^2$ passing though $2$ points in $\mathbb{P}^4$,

2450: which is a  $\mathbb{P}^2$. The fibration is smooth and it remains to calculate the Euler number

2451: of the basis.

2452: %Using the formula

2453: %\begin{equation}

2454: %\sum_{n=0}^\infty e({\rm Sym}^n X) q^n=\prod_{i=1}^\infty \left({1\over 1- q^n} \right)^{e(M)}

2455: %\end{equation}

2456: There are nice product formulas for the Euler number of symmetric

2457: products of surfaces modded out by $S_n$. For surfaces it is more

2458: cumbersome. We calculate the Euler number $e({\rm

2459: Sym}^2(X))=\left(-199\atop 2\right)$. ${\rm Hilb}^2(X)$ is the

2460: resolution of the  orbifold  ${\rm Sym}^2(X)$, which has the

2461: diagonal $X$ as fix point set. The resolution replaces each point

2462: in the fixed point set by $\mathbb{P}^2$. Simple surgery and  the

2463: smooth fibration structure of  ${\cal C}^{(2)}$ gives hence

2464: $e({\cal C}^{(2)})=3 (e({\rm Sym}^2(X)+(3-1)e(X))=58500,$ which by

2465: (\ref{BPSspecial}) yields  $n^{4}_5=58500+(2\cdot 6-4)(-1200)+35

2466: \cdot 10=49250$.

2467: 

2468: The approach becomes more difficult with the number of free points

2469: $\delta$ and at $\delta=4$ it is currently not know how to treat

2470: the singularities of the Hilbert scheme.

2471: 

2472: On the other hand smooth curves at the `edge' of the  Castelnuovo

2473: bound are of no principal problem. E.g., using adjunction for a

2474: smooth complete intersection of degree $(d_1,\ldots,d_r$) in a

2475: (weighted) projective space $\mathbb{WCP}^n(w_1,\ldots,w_{n+1})$

2476: in Appendix A, we calculate $\chi=(2-2g)$ and see that the degree

2477: 10 genus 16 curve is the complete intersection  $(1,2,5)$. The

2478: moduli space is calculated by counting the independent

2479: deformations of that complete intersection. The degree five

2480: constraint lies on the quintic, the linear constraint has five

2481: parameters. The identification by the $\mathbb{C}^*$ action of the

2482: ambient $\mathbb{P}^4$ shows that these parameters lie in a

2483: $\mathbb{P}^4$. This constraint allows to eliminate one variable

2484: from the generic quadratic constraint  which has hence $10$

2485: parameters and a $\mathbb{P}^9$ as moduli space. So we check in

2486: Tab. 2 the entry $n^{16}_{10}=(-1)^{13} 5\cdot 10=-50$.

2487: 

2488: Let us discuss the upper bound on the genus at which we can

2489: completely completely fix the $F_g$ given simply the bound

2490: (\ref{boundarycountquintic}). We claim that this bound is  $g\le

2491: 51$. At degree $20$ there is a smooth complete intersection curve

2492: $(1,4,5)$ of that genus. We first check that this is the curve of

2493: maximal genus in degree $20$. The Castelnuovos bound for curves in

2494: $\mathbb{P}^4$ shows that they have smaller genus~\cite{Harris}.

2495: We further see from the discussion in~\cite{Harris} that for

2496: curves in $\mathbb{P}^3$ not on quadric and a cubic, which would

2497: have the wrong degree, the Castelnuovos's bound is saturated for

2498: the complete intersection  $(1,4,5)$. For $g=51$

2499: (\ref{boundarycountquintic}) indicates that the gap, constant map

2500: contribution  and regularity at the orbifold fixes $131$ of the

2501: $151$ unknown coefficients in  (\ref{10-14-2.16}). The vanishing

2502: of $n^{51}_{d}=0$, $1\le d\le 19$ and the value of

2503: $n^{51}_{20}=(-1)^{ 4+ 34}

2504: \chi(\mathbb{P}^4)\chi(\mathbb{P}^{34}) =165$ for the Euler number

2505: of the moduli space of the smooth curve give us the rest of the

2506: data.

2507: 

2508: \parbox{14cm} % Seitenbreite

2509: {

2510:    %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2511:    %% Abbildung contour  %%%%%%%%%%%%%%%%%%%%%%

2512:    %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2513: 

2514:    \begin{center}

2515:    \mbox{

2516:          % \begin{turn}{-90}  % Hier k"onnen Sie das Bild drehen

2517:                              % z.B. um -90 Grad

2518:              \epsfig{file=castelnouvo.eps,width=14cm}

2519:          % \end{turn}

2520: }

2521:    \end{center}

2522:    %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

2523: }

2524: %

2525: \vskip -2mm\noindent {\em Castelnuovo's bound for higher genus

2526: curves on the quintic. The dashed line correspond roughly (up to taking the floor) 

2527: to the number of coefficients in $f_g$ (\ref{10-14-2.16}) which are not fixed by

2528: constant map contribution, conifold and orbifold boundary

2529: conditions.}

2530: % nachfolgender Text

2531: \vskip 2mm

2532: 

2533: 

2534: It is of course no problem to calculate  form the B-model the

2535: higher genus amplitudes to arbitrary degree. For completeness we

2536: report the first nontrivial numbers for $g=18-20$ in Tab. 3.

2537: 

2538: \begin{table}

2539: \begin{centering}

2540: \begin{tabular}{|r|rr|}

2541: \hline

2542: d &g=18 &g=19\\

2543: \hline

2544: \vdots& \vdots   &   \vdots \\

2545: 11& 0 &  0  \\

2546: 12&-3937166500  & -13403500  \\

2547: 13& 285683687197594125   &   -2578098061480250            \\

2548: 14&  -95076957496873268057250  &    2730012506820193210000            \\

2549: 15&  6438165666769014564325336250  &            -342304337102629200272769700   \\

2550: 16&  15209062594213864261318125134875  &    15209062594213864261318125134875           \\

2551: \hline

2552: \end{tabular}

2553: 

2554: \begin{tabular}{|r|r|}

2555: \hline

2556: d &g=20 \\

2557: \hline

2558: \vdots& \vdots \\

2559: 11& 0 \\

2560: 12& 0 \\

2561: 13& 10690009494250 \\

2562: 14& -59205862559233156250 \\

2563: 15& 15368486208424999875838025 \\

2564: 16& -1036824730393980503709247290500 \\

2565: \hline

2566: \end{tabular}

2567: \caption{Some higher degree genus 18-20 BPS numbers for the quintic. Note

2568: that we can calculate all Donaldson-Thomas invariants for $d=1,\ldots,12$

2569: exactly.}

2570: \end{centering}

2571: \end{table}

2572: 

2573: \subsection{D-brane states on hypersurfaces in weighted projective space}

2574: 

2575: Similarly, for the sextic in $\mathbb{P}^4(1^4,2)$, the degree

2576: $(1,2,6)$ complete intersection curve has genus $g=10$ and degree

2577: $d={\prod_i d_i\over \prod_k w_k}=6$ in the weighted projective

2578: space . Its moduli space is $\mathbb{P}^3$ for the degree one

2579: constraint, i.e., we can eliminate $x_4$ form the quadric and the 7

2580: coefficeints of the monomials $x_1^2, x_1x_2, x_1x_3, x_2^2, x_2

2581: x_3 ,x_3^2,x_5$ form a $\mathbb{P}^6$. This yields

2582: $n^{10}_6=(-1)^9 4 \cdot 7=-28$.

2583: 

2584: There are further checks for the octic in $\mathbb{P}^4(1^4,4)$ BPS

2585: invariants. The complete intersection $(1^2,8)$

2586: has total degree $2$ and genus $g=3$. The two linear constraints

2587: describe a $\mathbb{P}^1$ in $\mathbb{P}^3$ i.e. an

2588: $\mathbb{G}(1,3)$ with Euler number $6$ and dimension $4$, which

2589: yields $n^3_2=6$. Similarly we have a $g=7$ complete intersection

2590: $(1,2,8)$ of degree $4$, whose moduli space is  $\mathbb{P}^3$

2591: times $\mathbb{P}^5$ hence $n^7_4=24$.

2592: 

2593: For the degree 10 hypersurface in $\mathbb{P}^4(1^3,2,5)$ we check the BPS invariants:

2594: {}From the degree $(1,1,10)$ hypersuface of degree $1$ complete intersection with $g=2$. The moduli

2595: space of the linear constraints are just the one of  point in

2596: $\mathbb{P}^2$, i.e.$\mathbb{P}^2$, hence $n^2_1=3$. The degree

2597: $(1,2,10)$ complete intersection with total degree $2$ and genus

2598: $4$ has the moduli space of the linear constraint, which is

2599: $\mathbb{P}^2$ and of the quadratic constraint is $\mathbb{P}^3$

2600: (from the coefficients of the monomials $x_1^2,x_1 x_2,

2601: x^2_2,x_4$), yielding $n_2^4=- 12$. Finally  the $(1,3,10)$

2602: complete intersection with genus $7$ and degree $3$, has a moduli

2603: space $\mathbb{P}^2$ times $\mathbb{P}^5$ (from the coefficients

2604: of the monomials $x_1^3,x_1^2 x_2, x_1 x^2_2,x_1 x_4,x_2^3 ,x_2

2605: x_4$) and $n_3^7=- 18$.

2606: 

2607: These checks in different geometrical situations

2608: establish quite impressively the universality

2609: of the gap structure at the conifold expansion.

2610: 

2611: 

2612: 

2613: \subsection{D-branes on complete intersections}

2614: Here we summarize our results on one modulus complete

2615: intersections in (weighted) projective space. More complete

2616: results are available in \cite{webpage}. Again we can check many

2617: BPS invariants associated to the smooth curves.

2618: 

2619: Let us check e.g.in table 7 the $n_9^{10}=15$. According

2620: to (\ref{chi}) we see that at degree $9$ there is a smooth genus

2621: $10$ curve, given by a complete intersection of multi degree $(1^2,3^2)$ in

2622: $\mathbb{P}^5$. Their moduli space is the Grassmannian

2623: $\mathbb{G}(3,5)$, which has Euler number $15$ and dimension $8$,

2624: hence  $n_9^{10}=15$. In a very  similar way it can be seen that

2625: the $n_8^9$ comes form a complete intersection curve of degree

2626:  $(4,2,1^2)$ with the same moduli

2627: space, so $n_8^9=15$ in table 8. Grassmannians related to complete

2628: intersection are also identified with the moduli spaces of the

2629: following smooth curves: The total degree six curve of genus seven

2630: in table 11 is a CI of multi degree $(1^2,3,4)$. Its moduli space

2631: is a $\mathbb{G}(2,4)$ explaining $n_6^7=10$. The degree four

2632: curve of genus five in table 12 is of multi degree $(1^2,4^2)$ and

2633: has moduli space $\mathbb{G}(1,3)$ yielding $n_4^5=6$.  The degree

2634: four curve of genus five in table 13 is of multi degree

2635: $(1^2,2,6)$ and has moduli space $\mathbb{G}(2,4)$ yielding

2636: $n_4^5=10$.  The degree two curve of genus three in table 14 is of

2637: multi degree $(1^2,4,6)$ and has moduli space

2638: $\mathbb{G}(0,2)=\mathbb{P}^2$ yielding $n_2^3=3$. The moduli

2639: space of the degree $4$ genus seven curve $(1,2,4,6)$ is an

2640: $\mathbb{P}^2$ times the moduli space $\mathbb{P}^4$ of quadrics

2641: in $\mathbb{WCP}^3(1^2,2^2)$, so that $n_4^7=(-1)^6 3\cdot 5=15$.

2642: For the degree $(6,6)$ complete intersection in

2643: $\mathbb{WCP}^3(1^2,2^2,3^2)$, see table 15, we have a degree one

2644: genus two intersection $(1^2,6^2)$, whose moduli space is a point

2645: hence $n_1^2=1$, a degree two genus four intersection $(1,2,6^2)$,

2646: whose moduli space is $\mathbb{P}^1$ times the moduli space

2647: $\mathbb{P}^2$ of quadrics in $\mathbb{WCP}^2(1,2^2)$ hence

2648: $n_2^4=-6$, a degree three  genus seven intersection $(1,3,6^2)$,

2649: whose moduli space is $\mathbb{P}^1$ times the moduli space

2650: $\mathbb{P}^4$ of cubics in $\mathbb{WCP}^4(1,2^2,3^2)$ hence

2651: $n_3^7=-10$ and finally  a degree four genus eleven intersection

2652: $(1,4,6^2)$, whose moduli space is  $\mathbb{P}^1$ times the

2653: moduli space $\mathbb{P}^5$ of quadrics in

2654: $\mathbb{WCP}^4(1,2^2,3^2)$ hence $n_4^{11}=12$.

2655: 

2656: 

2657: There are many further checks that are somewhat harder to perform.

2658: E.g. we notice that there is a genus one degree three curve in the

2659: $(3,2,2)$ CI in $\mathbb{P}^6$, which comes from a complete

2660: intersection $(1^4,3)$. Now the moduli space of this complete

2661: intersection in  $\mathbb{P}^6$ is $\mathbb{G}(2,6)$. However not

2662: all $\mathbb{P}^2$ parametrized by $\mathbb{G}(2,6)$, which

2663: contain the cubic, are actually in the two quadrics of the

2664: $(3,2,2)$ CICY. We can restrict to those $\mathbb{P}^2$, which

2665: fulfill these constraints, by considering the simultaneous zeros

2666: of sections of two rank six bundles of quadratic forms on the

2667: moving $\mathbb{P}^2$. These are a number of points, which is

2668: calculated by the integral of the product of the Chern classes of

2669: these rank 6 bundles over  $\mathbb{G}(2,6)$. Indeed we obtain,

2670: for example with ``Schubert'' \cite{Schubert}

2671: \begin{equation}

2672: n_3^1=(-1)^0 \int_{\mathbb{G}(2,6)} c^2_6({\rm Sym}(2,Q))=64,

2673: \end{equation}

2674: which confirms the corresponding entry in table 9.

2675: 

2676: 

2677: 

2678: \section{Conclusions}

2679: 

2680: In this paper we solve the topological string B-model on compact

2681: Calabi-Yau $M$ using the modularity of the $F_g$,

2682: the wave function transformation property of $Z$ and the boundary

2683: information imposed by effective action considerations. The method

2684: pushes the calculation to unprecedented high genus amplitudes.

2685: E.g. for the quintic the boundary condition count

2686: (\ref{boundarycountquintic}) together with the simplest vanishing

2687: arguments at large volume fixes the amplitudes up to genus $g=51$.

2688: Beyond that the  prime mathematical problem to overcome in this

2689: region of the moduli space is understand the degeneration of more

2690: then four points in the relative Hilbert scheme of the universal

2691: curves in a threefold\footnote{As a motivation and check for the

2692: task to develop the  theory of Hilbert schemes for 3folds we

2693: calcultated the invariants explicitly to high genus. For the

2694: quintic to genus 20 and for all other the results up genus 12 are

2695: available at \cite{webpage}.}. Similar problems have been

2696: encountered in~\cite{Gaiotto:2006wm}, where it was suggested to

2697: fix a very similar ambiguity to a anholomorphic ${\rm

2698: SL}(2,\mathbb{Z})$-modular elliptic index of a $D4-D2-D0$ brane

2699: system~\cite{Gaiotto:2006wm,deBoer:2006vg,DennefMoore}. There one  uses

2700: ${\rm SL}(2,\mathbb{Z})$ invariance of the index and a dual dilute gas

2701: approximation in $AdS_3\times S^2\times M$ to fix the coeffcients 

2702: of the ring of modular forms. The construction of the moduli-space of the  

2703: $D4-D2-D0$ brane system uses rational GW invariants and implies non-trivial 

2704: relations among them~\cite{Gaiotto:2006wm}. Such considerations 

2705: could in principal provide further boundary conditions 

2706: at large radius.

2707: 

2708: 

2709: Our sharpest tool is the global control of $Z$ over ${\cal

2710: M}(M)$ and we  expect that by a closer analysis of the RR-spectrum

2711: at the orbifold of compact Calabi-Yau, we will be able to recover

2712: at least the $\lceil \frac{2}{d}(g-1)\rceil$ conditions that one

2713: loses relative to the local cases~\cite{Huang2} and solve the

2714: model completely. We obtained not only the Gromov-Witten, the

2715: Donaldson-Thomas and Gopakumar-Vafa invariants at infinity, but

2716: also the local expansion at the conifold, the Gepner point and

2717: other more exotic singularities with one or more massless states.

2718: The leading singular terms in the effective action reflect the

2719: massless states. The branch locus of the 13 parameter 

2720: models has an intriguing variety of such light spectra and 

2721: we can learn from the effective action about the singularity 

2722: and vice versa. Stability properties of theses states have 

2723: been analysed in App. \ref{appendixD-12-06}.

2724: 

2725: 

2726: Most importantly our exact expansions do contain further

2727: detailed information of the towers massive RR-states at these

2728: points. We described them in natural local variables. The 

2729: information from different genera should be of great value for the

2730: study of stable even D-brane bound states on compact Calabi-Yau as

2731: it is the content of the supersymmetric index of~\cite{GVII},

2732: which is protected under deformations of the complex structure.

2733: Non-compact Calabi-Yau such as the resolution of

2734: $\mathbb{C}^n/G$, with $G\in {\rm SL}(n,\mathbb{C})$ have

2735: no complex moduli. The issue does not arise and the situation is

2736: better understood, see e.g. \cite{Bridgeland,Aspinwall} for reviews.

2737: 

2738: One can also use the  explicit expansions  to study  the

2739: integrable theories that have been associated to the local

2740: expansion of the topological string on Calabi-Yau manifold, such

2741: as the $c=1$ string at the conifold or the quiver gauge theories

2742: at the orbifold, matrix models e.g. at the ADE singularities and new 

2743: ones for more exotic singularities such as the  

2744: branch points of the complete intersections Calabi-Yau 

2745: manifolds that we discussed here.

2746: 

2747: 

2748: The ability to obtain the imprint of the BPS spectrum on the

2749: effective action everywhere on the moduli space is of

2750: phenomenological interest as flux compactifications drive the

2751: theory to attractor points inside the moduli space.

2752: 

2753: Our expression are governed by the representation of the modular

2754: group of the Calabi-Yau on almost holomorphic forms, which we

2755: explicitly constructed from the periods, without having much of an

2756: independent theory about them. The simpler case of the torus

2757: suggest that such forms and their extensions should play a role

2758: in the study of of virtually any physical amplitude --- open or closed--- 

2759: in compactifications on the Calabi-Yau space, even as conjectured 

2760: in the hypermultiplet sector~\cite{Rocek:2005ij}.

2761: 

2762: One may finally wonder whether the topological string B-model is

2763: an integrable theory that is genuinely associated to this new and

2764: barely explored class of modular forms on Calabi-Yau spaces moduli

2765: spaces, whereas most known integrable models are associated to

2766: abelian varieties. As it was noted

2767: in~\cite{Verlinde:2004ck,Aganagic:2006wq, Gunaydin:2006bz}~in the

2768: complex moduli space extended by the dilaton, called extended

2769: phase space, one has one has rigid special K\"ahler geometry and

2770: many aspects of the sympletic transformations and its metaplectic

2771: realization are easier understood in the extended phase space.

2772: There are two maps $\Phi^{(i)}: {\cal M}\rightarrow T^{(k)}_{IJ}$,

2773: $I,J=1,\ldots,\frac{h^3}{2}$ from the complex moduli space to

2774: tensors in the extended phase space on which

2775: $\left(\begin{array}{cc} A&B\\ C&D \end{array}\right)\in {\rm Sp}

2776: (h^3,\mathbb{Z})$ acts projectively like $T^{(k)}\mapsto (A

2777: T^{(k)}+B)(C T^{k}+D)^{-1}$. For the holomorphic object

2778: $\tau_{IJ}=\partial_I \partial_J F^{(0)}=:T^{(1)}_{IJ}(t)$, which is 

2779: mostly discussed in this

2780: context of the metaplectic transfomations~\cite{Verlinde:2004ck,

2781: Aganagic:2006wq,Gunaydin:2006bz}, ${\rm Im}(\tau)$ is indefinite,

2782: while for the non-holomorphic object ${\cal N}_{IJ}={\bar

2783: \tau}_{IJ}+ 2 i \frac{{\rm Im}\tau_{IK} X^K {\rm Im}\tau_{IL}

2784: X^L}{X^L{\rm Im}\tau_{KL} X^L} =:T^{(2)}_{IJ}(t,\bar t)$ comes

2785: from the kinetic term in the 10d action whose reduction involves

2786: the Hodge-star on $M$. It's imaginary  part ${\rm Im}({\cal N})>0$

2787: is the kinetic term for the vector multiplets and is  hence

2788: positive definite. In other words ${\rm Im}\Phi^{(2)}$ defines a

2789: map to the Siegel upper space. $\Phi^{(2)}$ should relate Siegel

2790: modular forms for admittedly very exotic subgroups~\cite{cyy} of

2791: ${\rm Sp}(4,\mathbb{Z})$ to Calabi-Yau amplitudes. Such Siegel

2792: modular forms for abelian varieties  are also associated to $N=2$

2793: Seiberg-Witten (gauge) theories, while the modular forms on

2794: Calabi-Yau studied in this paper underline $N=2$ exact terms in

2795: $N=2$ supergravity. The map $\Phi^{(2)}$ could be a manifestation

2796: of a gravity-gauge theory correspondence for 4d theories with

2797: $N=2$ supersymmetry.

2798: 

2799: 

2800: It is no principal problem to generalize this to multi-moduli Calabi-Yau as

2801: long as the Picard-Fuchs equations are known. These have different, more

2802: general singularities with interesting local effective actions. In

2803: K3 fibrations which have at least two moduli, the modular properties are much

2804: better understood and in fact the ambiguity in the fiber is

2805: complete fixed heterotic string calculations. Moreover these cases

2806: have $N=2$ field theory limits, which contain further information,

2807: which might be sufficient to solve these models \cite{GKMW}.

2808: 

2809: 

2810: 

2811: 

2812: %----------------------------------------------------------------------------------------

2813: 

2814: 

2815: \vspace{0.2in} {\leftline {\bf Acknowledgments:}}

2816: 

2817: We thank M.~Aganagic, V.~Bouchard, T.~Grimm, S.~Katz, M.~Kontsevich, M.~Marino, C.~Vafa,

2818: S.~T.~Yau and D.~Zagier for discussions. Sheldon Katz helped us with

2819: the verifications of the BPS numbers and Cumrun Vafa with remarks

2820: on the draft. Don Zagier's comments on~\cite{Huang} triggered many ideas here.

2821: We thank the MSRI in Berkeley and AK thanks in particular the

2822: Simons Professorship Program. MH/AK thank also the Simons Workshops

2823: in Mathematics and Physics 05/06 for its hospitality.

2824: 

2825: \appendix

2826: 

2827: \section{Appendices}

2828: %----------------------------------------------------------------------------------

2829: \subsection{Classical intersection calculations using the adjunction

2830: formula}

2831: \label{intersection}

2832: 

2833: The adjunction formula\footnote{See \cite{fulton} for a

2834: pedagogical account of these matters.} for the total Chern class

2835: of a  for dimension $m=n-r$ smooth complete intersections  $M$ of

2836: multi degree  $d_1,\ldots,d_r$ in a weighted projective space

2837: $\mathbb{WCP}^n(w_1,\ldots,w_{n+1})$ is

2838: \begin{equation}

2839: c(T_M)=\sum_i c_i(T_M)=

2840: {c(T_{\mathbb{WCP}})\over c({\cal N})}= {\prod_{i=1}^{n+1} (1+w_i K)\over \prod_{k=1}^r( 1+ d_k K)}=\sum_i c_i K^i\ ,

2841: \label{adjunction}

2842: \end{equation}

2843: where   $c(T_{\mathbb{WCP}})=\sum_i c_i(T_{\mathbb{WCP}})=\prod_{i=1}^{n+1} (1+w_i K)$

2844: is the total Chern class  of the weighted projective space, $K$ is its K\"ahler class and

2845: $c({\cal N})=\prod_{k=1}^r( 1+ d_k K)$ is  the  total Chern class of the normal bundle.

2846: 

2847: Integration of a top form $\omega=x J^m$ with $J=K|_M$ over $M$ is obtained by integration

2848: along the normal direction as

2849: \begin{equation}

2850: \int_{M} \omega=\int_{\mathbb{WCP}} \omega \wedge c_r({\cal N}) = {x\over \prod_{k=1}^{n+1} w_k} \prod_{k=1}^r d_k \ .

2851: \label{integrationalg}

2852: \end{equation}

2853: Here we used the normalization  $\int_{\mathbb{WCP}} K^n= {1\over \prod_{k=1}^{n+1} w_k}$.

2854: This  yields the first line below:

2855: \begin{eqnarray}

2856: \kappa&=&\int_{M} J^m={\prod_{k=1}^r d_k\over \prod_{i=1}^{r+1} w_i} \label{kappa}\\

2857: \chi  &=& \int_M c_3(T_M)={c_3\over \prod_{k=1}^{n+1} w_k} \prod_{k=1}^r d_k \label{chi}\\

2858:   c  &=& {1\over 24} \int_M c_2\wedge J={1\over 24} {c_2\over \prod_{k=1}^{n+1} w_k} \prod_{k=1}^r d_k \label{cconst}\\

2859:   a  &=& {1\over 2}  \int_M i_* c_1(D)\wedge J= {1 \over 2}

2860: \int_{\mathbb{WCP}} {c(T_M)\over (1+J)}\wedge J^{r+1}={{\rm coeff} \left({c(T_M)\over (1+J)},J^{m-1}\right)

2861: \over 2\cdot \prod_{k=1}^{n+1} w_k }\label{aconst}

2862: \end{eqnarray}

2863: Combining (\ref{adjunction},\ref{integrationalg}) one gets the line 2

2864: and 3. The leading $t$ terms in $F_0$ can be obtained by

2865: calculating $Z(M)$  using (\ref{dbranecharge}), while the last

2866: line follows from the calculation of $Z(D)$ assuming that the

2867: $D_4$-brane is supported on $D$ the restriction of the hyperplane

2868: class\footnote{$a$ is physically less relevant, as it does not

2869: affect the effective action. Its value $a={11\over 2}$ obtained

2870: for the quintic from  (\ref{aconst}) checks

2871: with~\cite{Candelas:1990rm}} of ${\mathbb{WCP}}$ to $M$ and the Gysins

2872: formula for smooth embeddings \cite{fulton}.

2873: 

2874: \subsection{Tables of Gopakumar-Vafa invariants}

2875: 

2876: We list the tables of BPS invariants for all the Calabi-Yau models

2877: computed in this paper.

2878: 

2879: 

2880: %-----------------------------SEXTIC----------------------------------------------------------

2881: 

2882: \begin{table}

2883: \begin{centering}

2884: \begin{tabular}{|r|rrrrr|}

2885: \hline

2886: g &d=1 &d=2 &d=3 &d=4 &d=5  \\

2887: \hline

2888: \, 0& \, 7884&6028452& \, 11900417220& \, 34600752005688& \,  24595034333130080\\

2889: 1& 0   &   7884&   145114704&  1773044322885&  17144900584158168\\

2890: 2& 0   &     0&        17496&    10801446444&    571861298748384\\

2891: 3& 0   &     0&          576&      -14966100&      1985113680408\\

2892: 4& 0   &     0&            6&       -47304  &       -21559102992\\

2893: 5& 0   &     0&            0&              0&           22232340\\

2894: 6& 0   &     0&            0&              0&              63072\\

2895: 7& 0   &     0&            0&              0&                  0\\

2896: \hline

2897: \end{tabular}

2898: \vskip  5 pt

2899: \begin{tabular}{|r|rrr|}

2900: \hline

2901: g &d=6 & d=7 &d=8 \\

2902: \hline

2903: 0& \!\!\! 513797193321737210316&  \!\!\!\!\!\!\!\!\!\!\!2326721904320912944749252&

2904: \!\!\!\!\!\!\!\!\!\!\!11284058913384803271372834984\\

2905: 1& \!\!\! 147664736456952923604& \!\!\!\!\!\!\!\!\!\!\!1197243574587406496495592&

2906: \!\!\!\!\!\!\!\!\!\!\!9381487423491392389034886369\\

2907: 2&  \!\!\! 13753100019804005556&  \!\!\!\!\!\!\!\!\!\!\! 233127389355701229349884&

2908: \!\!\!\!\!\!\!\!\!\!3246006977306701566424657380\\

2909: 3&    411536108778637626&   19655355035815833642912& 561664896373523768591774196\\

2910: 4&      1094535956564124&     628760082868148062854& 48641994707889298118864544\\

2911: 5&       -18316495265688&       3229011644338336680& 1863506489926528403683191\\

2912: 6&          207237771936&        -18261998133124302& 20758968356323626025164\\

2913: 7&            -583398600&           513634614205788& 10040615124628834206\\

2914: 8&               -146988&            -8041642037676& 1129699628821681740\\

2915: 9&                 -3168&               54521267292& -38940584273866593\\

2916: 10&                  -28&                 -43329384&  904511824896888\\

2917: 11&                    0&                  -110376& -12434437188576\\

2918: 12&                    0&                     0 &     76595605884 \\

2919: \hline

2920: \end{tabular}

2921: \caption{BPS invariants $n^g_d$ on the Sextic hypersurface in $\mathbb{P}^4(1^4,2)$.}

2922: \end{centering}

2923: \end{table}

2924: 

2925: 

2926: 

2927: 

2928: %----------------------------------Octic-------------------------------------------------

2929: \begin{table}

2930: \begin{centering}

2931: \begin{tabular}{|r|rrrrr|}

2932: \hline

2933: g &d=1 &d=2 &d=3 &d=4 &d=5  \\

2934: \hline

2935: 0& 29504& 128834912&   1423720546880& 23193056024793312&467876474625249316800 \\

2936: 1& 0   &      41312&     21464350592&  1805292092705856&101424054914016355712 \\

2937: 2& 0   &        864&       -16551744&    12499667277744&  5401493537244872896 \\

2938: 3& 0   &          6&         -177024&     -174859503824&    20584473699930496 \\

2939: 4& 0   &          0&               0&         396215800&     -674562224718848 \\

2940: 5& 0   &          0&               0&            301450&       12063928269056 \\

2941: 6& 0   &          0&               0&              4152&         -86307810432 \\

2942: 7& 0   &          0&               0&                24&             37529088 \\

2943: 8& 0   &          0&               0&                 0&               354048 \\

2944: $\vdots$&$\vdots$&$\vdots$&$\vdots$& $\vdots$& \\

2945: \hline

2946: \end{tabular}

2947: \caption{BPS invariants $n^g_d$ on the Octic hypersurface in $\mathbb{P}^4(1^4,4)$.}

2948: \end{centering}

2949: \end{table}

2950: 

2951: 

2952: 

2953: %---------------------------------Degree Ten--------------------------------------------

2954: \begin{table}

2955: \begin{centering}

2956: \begin{tabular}{|r|rrrr|}

2957: \hline

2958: g &d=1 &d=2 &d=3 &d=4  \\

2959: \hline

2960: 0& 231200& 12215785600& 1700894366474400& 350154658851324656000

2961: %& 89338191421813572850115680

2962: \\

2963: 1&    280&   207680960&  161279120326840& 103038403740897786400

2964: %& 59221844124053623534387208

2965: \\

2966: 2&      3&    -537976 &    1264588024791&   8495973047204168640

2967: %& 14756017196840260164213441

2968: \\

2969: 3&      0&       -1656&     -46669244594&     61218893443516800

2970: %&  1453221695196526165828470

2971: \\

2972: 4& 0     &         -12&        630052679&     -2460869494476896

2973: %&    34858775918056957341231

2974: \\

2975: 5& 0     &           0&         -1057570&       145198012290472

2976: %&     -116969829757480707582

2977: \\

2978: 6& 0     &           0&            -2646&        -5611087226688

2979: %&       21889605001143026505

2980: \\

2981: 7& 0     &           0&              -18&         125509540304

2982: %&        -2168811495068153226

2983: \\

2984: 8& 0     &           0&                0&         -1268283512

2985: %&          168060629266907127

2986: \\

2987: $\vdots$&$\vdots$&$\vdots$&$\vdots$&$\vdots$ \\

2988: \hline

2989: \end{tabular}

2990: \caption{BPS invariants $n^g_d$ on the degree 10  hypersurface in $\mathbb{P}^4(1^3,2,5)$.}

2991: \end{centering}

2992: \end{table}

2993: 

2994: %----------------------------------bicubic -------------------------------------------------

2995: \begin{table}

2996: \begin{centering}

2997: 

2998: \begin{tabular}{|r|rrrrrr|}

2999: \hline

3000: g &d=1 &d=2 &d=3 &d=4 &d=5 &d=6 \\

3001: \hline

3002: \, 0& 1053 & 52812 & 6424326  & 1139448384 & 249787892583& 62660964509532\\

3003: 1& 0   &  0 & 3402 &5520393  & 4820744484&3163476682080\\

3004: 2& 0   &  0 & 0 & 0 &  5520393& 23395810338\\

3005: 3& 0   &  0 &  0&  0& 0& 6852978 \\

3006: 4& 0   &  0 &  0&  0& 0&  10206\\

3007: 5& 0   &  0 &  0&  0& 0&  0 \\

3008: \hline

3009: \end{tabular}

3010: \begin{tabular}{|r|rrr|}

3011: \hline

3012: g &d=7 &d=8 &d=9  \\

3013: \hline

3014: 0& 17256453900822009 & 5088842568426162960  & 1581250717976557887945 \\

3015: 1&  1798399482469092 &  944929890853230501  &  473725044069553679454 \\

3016: 2&   42200615912499  &   50349477671013600  &   47431893998882182563 \\

3017: 3&      174007524240 &     785786604262830  &    1789615720312984368 \\

3018: 4&          -484542  &       2028116431098  &      21692992151427138\\

3019: 5&           158436 &           -784819773  &         36760497856020\\

3020: 6&                0 &               372762  &           -61753761036\\

3021: 7&                0 &               6318    &             -5412348\\

3022: 8&                0 &                0      &                39033\\

3023: 9&                0 &                0      &                1170\\

3024: 10&               0 &                0      &                 15\\

3025: 11      &         0 &                0      &                  0\\

3026: \hline

3027: \end{tabular}

3028: \caption{$n^g_d$ for the degree  (3,3) complete intersection in $\mathbb{P}^5$.}

3029: \end{centering}

3030: \end{table}

3031: 

3032: 

3033: \begin{table}

3034: \begin{centering}

3035: \begin{tabular}{|r|rrrrrr|}

3036: \hline

3037: g &d=1 &d=2 &d=3 &d=4 &d=5 &d=6\\

3038: \hline

3039: \, 0& 1280& 92288& 15655168& 3883902528& 1190923282176& 417874605342336  \\

3040: 1& 0&0& 0& -672& 16069888& 174937485184\\

3041: 2& 0& 0& 0& -8& 7680& 12679552\\

3042: 3& 0& 0& 0& 0& 0& 276864\\

3043: 4& 0& 0& 0& 0& 0& 0\\

3044: \hline

3045: \end{tabular}

3046: \begin{tabular}{|r|rrr|}

3047: \hline

3048: g &d=7 &d=8 &d=9\\

3049: \hline

3050: 0& \! 160964588281789696&  \!\! 66392895625625639488&  \!\! 28855060316616488359936 \\

3051: 1&  19078577926517760& 14088192680381290336& 9895851364631438617600\\

3052: 2&  494602061689344& 853657285175383648& 1137794220513866498304\\

3053: 3&  2016330670592& 14859083841009280& 49286012311292922368\\

3054: 4&  -285585152& 37334304102560& 679351051885623552\\

3055: 5& 591360& -46434384200& 1103462757073920\\

3056: 6& 7680& -8285120& -4031209095680\\

3057: 7&0& 67208& 370290688 \\

3058: 8& 0& 1520& -2270720 \\

3059: 9& 0& 15& -25600 \\

3060: 10&0& 0& 0\\

3061: \hline

3062: \end{tabular}

3063: 

3064: \caption{$n^g_d$ for the degree (4,2) complete intersection in $\mathbb{P}^5$.}

3065: \end{centering}

3066: \end{table}

3067: 

3068: \begin{table}

3069: \begin{centering}

3070: \begin{tabular}{|r|rrrrrr|}

3071: \hline

3072: g &d=1 &d=2 &d=3 &d=4 &d=5 &d=6\\

3073: \hline

3074: 0& 720& 22428& 1611504& 168199200& 21676931712& 3195557904564  \\

3075: 1& 0& 0& 64& 265113& 198087264& 89191835056\\

3076: 2& 0& 0& 0& 0& 10080& 180870120\\

3077: 3& 0& 0& 0& 0& 0& -3696  \\

3078: 4& 0& 0& 0& 0& 0& -56 \\

3079: 5& 0& 0& 0& 0& 0& 0 \\

3080: \hline

3081: \end{tabular}

3082: \begin{tabular}{|r|rrr|}

3083: \hline

3084: g &d=7 &d=8 &d=9\\

3085: \hline

3086: 0& \,\, 517064870788848& \, 89580965599606752& \, 16352303769375910848  \\

3087: 1& 32343228035424&10503104916431241& 3201634967657293024 \\

3088: 2& 315217101456& 280315384261560& 178223080602086784 \\

3089: 3& 199357344& 1430336342574& 2915033921871456 \\

3090: 4& 30240& 194067288& 8888143990672 \\

3091: 5& 0& 795339& -233104896  \\

3092: 6& 0& 0& 4857552 \\

3093: 7& 0& 0& 384 \\

3094: 8& 0& 0&  0\\

3095: \hline

3096: \end{tabular}

3097: \caption{$n^g_d$ for the (3,2,2) complete intersection in $\mathbb{P}^6$.}

3098: \end{centering}

3099: \end{table}

3100: 

3101: \begin{table}

3102: \begin{centering}

3103: \begin{tabular}{|r|rrrrrrr|}

3104: \hline

3105: g &d=1 &d=2 &d=3 &d=4 &d=5 &d=6& d=7\\

3106: \hline

3107: 0& \,\,  512&\, 9728& \,416256& \,25703936& \,1957983744&  \, 170535923200&  \,16300354777600   \\

3108: 1& 0& 0& 0& 14752& 8782848& 2672004608& 615920502784 \\

3109: 2& 0& 0& 0& 0& 0& 1427968& 2440504320\\

3110: 3& 0&0& 0& 0& 0& 0& 86016 \\

3111: 4&0& 0& 0& 0& 0& 0& 0  \\

3112: \hline

3113: \end{tabular}

3114: \begin{tabular}{|r|rrrr|}

3115: \hline

3116: g &d=8 &d=9 &d=10 &d=11\\

3117: \hline

3118: 0& \!\!\!\!\!  1668063096387072& \!\!\!\!\!\!\!\!\! 179845756064329728& \!\!\!\!\!\!\!\!\!20206497983891554816& \!\!\!\!\!\!\!\!\!2347339059011866069504 \\

3119: 1& \!\!\!\!\!\!\!\!\!123699143093152& \!\!\!\!\!\!\!\!\!22984995833484288&  \!\!\!\!\!\!\!\!\! 4071465816864581632&\!\!\!\!\!\!\!\!\! 698986176207439627264\\

3120: 2&  1628589698304& 702453851520512& 236803123487243776&68301097513852719616\\

3121: 3& 2403984384& 4702943495168& 4206537025629952& 2482415474680798208\\

3122: 4& -37632& 2449622016& 16316531089408& 29624281509824512\\

3123: 5& -672& 258048& 2777384448& 73818807399424\\

3124: 6& 0& 0& 4283904& 1153891840\\

3125: 7& 0& 0& 0& 26348544\\

3126: 8& 0& 0& 0& 0\\

3127: \hline

3128: \end{tabular}

3129: \caption{ $n^g_d$ for the degree (2,2,2,2) complete intersection in $\mathbb{P}^7$.}

3130: \end{centering}

3131: \end{table}

3132: 

3133: 

3134: \begin{table}

3135: \begin{centering}

3136: \begin{tabular}{|r|rrrrrr|}

3137: \hline

3138: g &d=1 &d=2 &d=3 &d=4 &d=5 &d=6 \\

3139: \hline

3140: 0& 1944& 223560& 64754568& 27482893704& 14431471821504& 8675274727197720\\

3141: 1&  0& 27& 161248& 381704265& 638555324400& 891094220317561\\

3142: 2& 0& 0& 0& 227448& 3896917776& 20929151321496\\

3143: 3& 0& 0& 0& 81& 155520& 75047188236\\

3144: 4& 0& 0& 0& 0& 5832& -40006768\\

3145: 5& 0& 0& 0& 0& 0& 26757\\

3146: 6& 0& 0& 0& 0& 0& 816\\

3147: 7& 0& 0& 0& 0& 0& 10\\

3148: 8&  0& 0& 0& 0& 0& 0\\

3149: \hline

3150: \end{tabular}

3151: \caption{ $n^g_d$ for the degree (4,3) complete intersection in $\mathbb{WCP}^5(1^5,2)$.}

3152: \end{centering}

3153: \end{table}

3154: 

3155: 

3156: \begin{table}

3157: \begin{centering}

3158: \begin{tabular}{|r|rrrrrr|}

3159: \hline

3160: g &d=1 &d=2 &d=3 &d=4 &d=5 &d=6 \\

3161: \hline

3162: 0& 3712& 982464& 683478144& 699999511744& 887939257620352& 1289954523115535040\\

3163: 1& 0& 1408& 6953728& 26841854688& 88647278203648& 266969312909257728\\

3164: 2& 0& 0& 3712& 148208928& 2161190443904& 17551821510538560\\

3165: 3& 0& 0& 0& -12432& 7282971392& 362668189458048\\

3166: 4&0& 0&0& 384& -14802048& 773557598272\\

3167: 5& 0& 0& 0& 6& -22272& -7046285440\\

3168: 6& 0& 0& 0& 0& 0& 6367872\\

3169: 7& 0& 0& 0& 0& 0& 11264\\

3170: 8& 0& 0& 0& 0& 0& 0\\

3171: \hline

3172: \end{tabular}

3173: \caption{$n^g_d$ for the degree (4,4) complete intersection in $\mathbb{WCP}^5(1^4,2^2)$.}

3174: \end{centering}

3175: \end{table}

3176: 

3177: 

3178: \begin{table}

3179: \begin{centering}

3180: \begin{tabular}{|r|rrrrrr|}

3181: \hline

3182: g &d=1 &d=2 &d=3 &d=4 &d=5 &d=6 \\

3183: \hline

3184: 0&4992&\!\!\!\!\!\! 2388768& \!\!\!\!\!\! 2732060032&\!\!\!\!\!\! 4599616564224& \!\!\!\!\!\!9579713847066240& \!\!\!\!\!\! 22839268002374163616 \\

3185: 1&0& -504& 1228032& 79275664800& 633074010435840& 3666182351842338408 \\

3186: 2& 0& -4& 14976& -13098688& 3921835430016& 128614837503143532\\

3187: 3& 0& 0& 0& 87376& -5731751168& 482407033529880\\

3188: 4&0& 0& 0& 1456& -7098624& -3978452463012 \\

3189: 5&0& 0& 0& 10& -59904& 1776341072 \\

3190: 6& 0& 0& 0& 0& 0& 18680344\\

3191: 7&0& 0& 0& 0& 0& -7176 \\

3192: 8& 0& 0& 0&  0& 0& -36\\

3193: 9&0& 0& 0& 0& 0& 0\\

3194: \hline

3195: \end{tabular}

3196: \caption{ $n^g_d$ for the degree (6,2) complete intersection in $\mathbb{WCP}^5(1^5,3)$.}

3197: \end{centering}

3198: \end{table}

3199: 

3200: \begin{table}

3201: \begin{centering}

3202: \begin{tabular}{|r|rrrrr|}

3203: \hline

3204: g &d=1 &d=2 &d=3 &d=4 &d=5  \\

3205: \hline

3206: 0& \ \ 15552&\ \  27904176& \ \ 133884554688&\ \  950676829466832& \ 8369111295497240640\\

3207: 1& 8& 258344& 5966034472& 126729436388624& 2512147219945401752\\

3208: 2&0& 128& 36976576& 4502079839576& 264945385369932352\\

3209: 3&0& 3& -64432& 15929894952& 9786781718701824\\

3210: 4& 0& 0& -48& -272993052& 42148996229312 \\

3211: 5& 0& 0& 0& 800065& -592538522344\\

3212: 6& 0& 0& 0& 1036& 14847229472\\

3213: 7& 0& 0& 0& 15& -148759496\\

3214: 8& 0& 0& 0& 0& 160128\\

3215: 9& 0& 0& 0& 0& 96\\

3216: 10& 0& 0& 0& 0& 0\\

3217: \hline

3218: \end{tabular}

3219: \caption{$n^g_d$ for the  degree (6,4) complete intersection in

3220: $\mathbb{WCP}^5(1^3,2^2,3)$.}

3221: \end{centering}

3222: \end{table}

3223: 

3224: \begin{table}

3225: \begin{centering}

3226: \begin{tabular}{|r|rrrr|}

3227: \hline

3228: g &d=1 &d=2 &d=3 &d=4 \\

3229: \hline

3230: 0&67104& 847288224& 28583248229280& 1431885139218997920 \\

3231: 1& 360& 40692096& 4956204918600& 616199133098321280\\

3232: 2& 1& 291328& 254022248925& 102984983365762128 \\

3233: 3& 0& -928& 1253312442& 6925290146728800\\

3234: 4& 0& -6& -39992931& 104226246583368\\

3235: 5& 0& 0& 867414& -442845743788\\

3236: 6&0& 0& -1807& 53221926192\\

3237: 7& 0& 0& -10& -3192574724\\

3238: 8& 0& 0& 0& 111434794\\

3239: 9& 0& 0& 0& -1752454\\

3240: 10& 0& 0& 0& 3054\\

3241: 11&0& 0& 0& 12\\

3242: 12& 0& 0& 0& 0\\

3243: \hline

3244: \end{tabular}

3245: \caption{$n^g_d$ for the degree (6,6) complete intersection in $\mathbb{WCP}^5(1^2,2^2,3^2)$.}

3246: \end{centering}

3247: \end{table}

3248: 

3249: 

3250: \newpage

3251: %-----------------------------------------------------------------

3252: \subsection{Invariance of the generators under a change of the basis}

3253: \label{sectioninvariance}

3254: We have seen the topological strings can be written as polynomials

3255: of the generators $v_1$ , $v_2$, $v_3$, and $X$. In the

3256: holomorphic limit, these generators can be computed from the first

3257: two solutions $\omega_0$, $\omega_1$ of the Picard-Fuchs equation.

3258: In the followings we prove that under an arbitrary linear change

3259: of basis of $\omega_0$ and $\omega_1$, these generators and

3260: therefore the topological strings are actually invariant. This is

3261: true anywhere in the moduli space. In particular, this partly

3262: explains why the gap structure in the conifold expansion is not

3263: affected by a change of basis of $\omega_0$ as we observed in

3264: all cases.

3265: 

3266: Since $X=\frac{1}{1-\psi}$ is independent of the basis $\omega_0$

3267: and $\omega_1$ , it is trivially invariant. In the holomorphic

3268: limit, The Kahler potential and metric go like $e^{-K}\sim

3269: \omega_0$ and $G_{\psi\bar{\psi}}\sim

3270: \partial_{\psi}t$, where $t=\frac{\omega_1}{\omega_0}$ is the mirror

3271: map. The generators $u$ and $v_i$ are related to $A_i$ and $B_i$,

3272: which we recall were defined as

3273: \begin{eqnarray} &&A_p:=\frac{(\psi\partial\psi)^p

3274: G_{\psi\bar{\psi}}}{G_{\psi\bar{\psi}}},~~B_p:=\frac{(\psi\partial\psi)^p

3275: e^{-K}}{e^{-K}}, ~~(p=1,2,3,\cdots)

3276: \end{eqnarray}

3277: So a different normalization of basis $\omega_0$ of $\omega_1$, as

3278: well as a change of basis in $\omega_1\rightarrow \omega_1+b_1

3279: \omega_0$ obviously do not change the generators $A_i$ and $B_i$,

3280: and therefore the generators $u$ and $v_i$ are also invariant.

3281: 

3282: We now tackle the remaining less trivial situation, namely a

3283: change of basis in $\omega_0$ as the following

3284: \begin{eqnarray} \label{basis-11-04-001}

3285: \omega_0\rightarrow \tilde{\omega}_0=\omega_0+b_1\omega_1

3286: \end{eqnarray}

3287: where $b_1$ is an arbitrary constant. We denote the K\"ahler

3288: potential, metric, mirror map and various generators in the new

3289: basis by a tilde symbol. It is straightforward to relate them to

3290: variables in the original basis. We find the following relations

3291: for the mirror map

3292: \begin{eqnarray}

3293: s=\frac{\omega_1}{\tilde{\omega}_0}=\frac{t}{1+b_1t}

3294: \nonumber \\

3295: \partial_{\psi}s=\frac{\partial_\psi t}{(1+b_1t)^2}

3296: \end{eqnarray}

3297: and the generators $A$ and $B$

3298: \begin{eqnarray} \label{ABrelation-11-04-002}

3299: \tilde{A}&=&\frac{\psi\partial_\psi\tilde{G}_{\psi\bar{\psi}}}{\tilde{G}_{\psi\bar{\psi}}}=A-\frac{2b_1\psi\partial_\psi t}{1+b_1\psi} \nonumber \\

3300: \tilde{B}&=&\frac{\psi\partial_\psi\tilde{\omega}_0}{\tilde{\omega}_0}=B+\frac{b_1\psi\partial_\psi

3301: t}{1+b_1\psi}

3302: \end{eqnarray}

3303: So we find the generators $A$ and $B$, as well as the generator

3304: $u=B$ are \textit{not} invariant under the change of basis

3305: (\ref{basis-11-04-001}). However, we recall the generator $v_1$ is

3306: defined as

3307: \begin{eqnarray}

3308: v_1=1+A+2B

3309: \end{eqnarray}

3310: Using the equations in (\ref{ABrelation-11-04-002}) we find the

3311: generator $v_1$ is invariant, namely $\tilde{v}_1=v_1$. To see

3312: $v_2$ and $v_3$ are invariant, we use the derivative relations

3313: \begin{eqnarray}

3314: \psi\partial_\psi v_1 &=& -v_1^2-2v_2-(1+r_0)X+v_1X \label{dr-11-04-0031} \\

3315: \psi\partial_\psi v_2 &=& -v_1v_2+v_3 \label{dr-11-04-0032}

3316: \end{eqnarray}

3317: where $r_0$ is a constant that appears in the relation of

3318: generator $A_2$ to lower generators. These derivative relations

3319: are exact and independent of the choice of the basis in asymptotic

3320: expansion. We have show that $v_1$ and $X$ in the first equations

3321: (\ref{dr-11-04-0031}) are invariant under a change of the basis

3322: (\ref{basis-11-04-001}), therefore the generator $v_2$ appearing

3323: on the right hand side must be also invariant. Applying the same

3324: logic to the second equation (\ref{dr-11-04-0032}) we find the

3325: generator $v_3$ are also invariant.

3326: 

3327: Our proof explains why a change of basis like

3328: (\ref{basis-11-04-001}) does not change the gap structure around

3329: the conifold point and seems to be related to the $SL_2$ orbit

3330: theorem of \cite{Schmidt,CKS}. Under a change of basis, the mirror map at the

3331: conifold point is

3332: $\tilde{t}_D=\frac{\omega_0t_D}{\tilde{\omega}_0}$, and has the

3333: asymptotic leading behavior $\tilde{t}_D\sim t_D\sim

3334: \mathcal{O}(\psi)$. Recall in the holomorphic limit, the conifold

3335: expansion is

3336: \begin{eqnarray}

3337: F^{(g)}_{\textrm{conifold}} =

3338: \omega_0^{2(g-1)}(\frac{1-\psi}{\psi})^{g-1}P_g(v_1,v_2,v_3,X) ,

3339: \end{eqnarray}

3340: As we have shown the generators $v_i$ and therefore $P_g$ are

3341: invariant, so in the new basis

3342: \begin{eqnarray}

3343: \tilde{F}^{(g)}_{\textrm{conifold}}=

3344: (\frac{\tilde{\omega}_0}{\omega_0})^{2(g-1)}

3345: F^{(g)}_{\textrm{conifold}}=(\frac{t_D}{\tilde{t}_D})^{2(g-1)}

3346: F^{(g)}_{\textrm{conifold}}

3347: \end{eqnarray}

3348: It is clear if there is a gap structure in one basis

3349: $F^{(g)}_{\textrm{conifold}}=\frac{(-1)^{g-1}B_{2g}}{2g(2g-2)t_D^{2g-2}}+\mathcal{O}(t_D^0)$,

3350: the same gap structure will be also present in the other basis,

3351: \begin{eqnarray}

3352: \tilde{F}^{(g)}_{\textrm{conifold}}=\frac{(-1)^{g-1}B_{2g}}{2g(2g-2)\tilde{t}_D^{2g-2}}+\mathcal{O}(\tilde{t}_D^0)

3353: \end{eqnarray}

3354: The asymptotic expansion in sub-leading terms $\mathcal{O}(t_D^0)$

3355: and $\mathcal{O}(\tilde{t}_D^0)$ will be different and can be

3356: computed by the relation between $t_D$ and $\tilde{t}_D$.

3357: 

3358: Around the conifold point there is another power series solution

3359: to the Picard-Fuchs equation that goes like $\omega_2\sim

3360: \mathcal{O}(\psi^2)$. We also observe that the gap structure is

3361: not affected by a change of the basis

3362: \begin{eqnarray}

3363: \omega_0\rightarrow \omega_0+b_2\omega_2

3364: \end{eqnarray}

3365: It appears to be much more difficult to prove this observation,

3366: since now the generators $v_i$ are not invariant under this change

3367: of basis. A proof of our observation would

3368: depend on the specific details of the polynomial $P_g$, and

3369: probably requires a deeper conceptual understanding of the

3370: conifold expansion. We shall leave this for future investigation.

3371: %------------------------------------------------------------------

3372: 

3373: 

3374: \subsection{Symplectic basis, vanishing cycles and massless particles}

3375: \label{appendixD-12-06}

3376: 

3377: We can study in more details the analytic continuation of the

3378: symplectic basis of the periods to the orbifold point $\psi=0$.

3379: For the four hypersurface cases and two other complete

3380: intersection models $X_{4,3}(1^5,2)$ and $X_{6,2}(1^3,2^2,3)$, the

3381: indices $a_i$ ($i=1,2,3,4$) of the Picard-Fuchs equation are not

3382: degenerate at the orbifold point, so there are $4$ power series

3383: with the leading behavior of $\psi^{a_i}$, and the analytic

3384: continuation procedure is similar to the quintic case. In our

3385: physical explanation of the singularity structure of higher genus

3386: topological string amplitudes in these models, we claim that for

3387: the hypersurface cases there is no stable massless charged state around

3388: the orbifold point, whereas for models $X_{4,3}(1^5,2)$ and

3389: $X_{6,2}(1^3,2^2,3)$ there are nearly massless charged states of

3390: mass $m\sim \psi^{a_2-a_1}$. Since the charge and the mass of a

3391: D-brane wrapping cycle is determined by the mirror map parameter,

3392: which is the ratio of two symplectic periods, it is only possible

3393: to have massless particles if there is a rational linear

3394: combination of the periods with the leading behavior of

3395: $\psi^{a_k}$, $k>1$.

3396: 

3397: For a complete intersection of degree $(d_1,\cdots,d_r)$ in

3398: weighted projective space $\mathbb{WCP}^n(w_1,\cdots,w_{n+1})$

3399: with non-degenerate indices $a_k$, the natural basis of solutions

3400: at the orbifold point is

3401: \begin{eqnarray}

3402: \omega_k^{\textrm{orb}}=\psi^{a_k}\frac{\prod_{i=1}^r\Gamma(d_ia_k)}{\prod_{i=1}^{n+1}\Gamma(w_ia_k)}

3403: \sum_{n=0}^{\infty}(c_0\psi)^n\frac{\prod_{i=1}^{n+1}\Gamma(w_i(n+a_k))}{\prod_{i=1}^r\Gamma(d_i(n+a_k))},~~~

3404: k=1,2,3,4

3405: \end{eqnarray}

3406: where

3407: $c_0=\frac{\prod_{i=1}^rd_i^{d_i}}{\prod_{i=1}^{n+1}w_i^{w_i}}$.

3408: We can write the sum of the series as a contour integral\footnote{Further useful

3409: properties of the periods of the one parameter models have been established in~\cite{Lazaroiu:2000jx}. }

3410: enclosing the positive real axis and analytically continue to the negative

3411: real axis to relate the above basis of solutions to the known

3412: symplectic basis at infinity $\psi=\infty$. We find the following

3413: relation, generalizing the result for quintic case,

3414: \begin{eqnarray} \label{D.113-12-07}

3415: \omega_k^{\textrm{orb}}&=&(2\pi

3416: i)^4\frac{\prod_{i=1}^r\Gamma(d_ia_k)}{\prod_{i=1}^{n+1}\Gamma(w_ia_k)}\{\frac{\alpha_k

3417: F_0}{1-\alpha_k}-\frac{\alpha_k F_1}{(1-\alpha_k)^2}

3418: +\frac{\alpha_k[\kappa(1+\alpha_k)-2a(1-\alpha_k)

3419: ]}{2(1-\alpha_k)^3}X_1 \nonumber \\

3420: && +\frac{\alpha_k[12c(1-\alpha_k)^2

3421: +\kappa(1+4\alpha_k+\alpha_k^2)]}{6(1-\alpha_k)^4}X_0

3422:   \}

3423: \end{eqnarray}

3424: here $\alpha_k=\exp(2\pi ia_k)$ and $\kappa$, $c$, $a$ are from

3425: the classical intersection calculations in Appendix

3426: \ref{intersection}. We find for all cases the symplectic form and

3427: Kahler potential have the same diagonal behavior as the case of

3428: the quintic

3429: \begin{eqnarray}

3430: \omega=dF_k\wedge dX_k=s_1 d\omega_1^{\textrm{orb}}\wedge

3431: d\omega_4^{\textrm{orb}}+s_2 d\omega_2^{\textrm{orb}}\wedge

3432: d\omega_3^{\textrm{orb}}

3433: \end{eqnarray}

3434: and $e^{-K}=\sum_{k=1}^4 r_k \omega_k^{\textrm{orb}}

3435: \overline{\omega_k^{\textrm{orb}}}$, for some constants $s_1, s_2$

3436: and $r_k$.

3437: 

3438: It is straightforward to invert the transformation

3439: (\ref{D.113-12-07}) and study the asymptotic behavior of the

3440: geometric symplectic basis $(F_0,F_1,X_0,X_1)$ around the orbifold

3441: point. Generically a linear combination of the symplectic periods

3442: $(F_0,F_1,X_0,X_1)$ is proportional to the power of $\psi$ with

3443: the lowest index, namely, for generic coefficients $c_1, c_2, c_3,

3444: c_4$ we have

3445: \begin{eqnarray}

3446: c_1X_0+c_2X_1+c_3F_1+c_4F_0\sim \omega_1^{\textrm{orb}}\sim

3447: \psi^{a_1}

3448: \end{eqnarray}

3449: In order for massless particles to appear at the orbifold point,

3450: there must be a $Sp(4,\mathbb{Z})$ transformation of the

3451: symplectic basis such that one of periods goes to zero faster that

3452: the generic situation, namely we should have integer coefficients $n_k\in \mathbb{Z}$

3453: satisfying

3454: \begin{eqnarray} \label{D.115}

3455: n_1X_0+n_2X_1+n_3F_1+n_4F_0\sim \omega_2^{\textrm{orb}}\sim

3456: \psi^{a_k},~~~~k>1

3457: \end{eqnarray}

3458: 

3459: We find (\ref{D.115}) is impossible for three of the hypersurface

3460: cases $X_5(1^5)$, $X_8(1^4,4)$ and $X_{10}(1^3,2,5)$, but possible

3461: for the Sextic hypersurface $X_6(1^4,2)$ and the complete

3462: intersection cases $X_{4,3}(1^5,2)$ and $X_{6,2}(1^3,2^2,3)$.

3463: Specifically, for the Sextic hypersurface $X_6(1^4,2)$, the

3464: condition for (\ref{D.115}) with $k=2$

3465: \begin{eqnarray}

3466: n_1+3n_3=0, ~~~ 3n_1+4(n_2+n_4)=0

3467: \label{allowed1}

3468: \end{eqnarray}

3469: whereas for the complete intersections  $X_{4,3}(1^5,2)$ and

3470: $X_{6,2}(1^3,2^2,3)$, the conditions for (\ref{D.115}) are

3471: \begin{eqnarray}

3472: n_1+3n_3=0, ~~~ n_1=4n_2+8n_4,

3473: \label{allowed2}

3474: \end{eqnarray}

3475: and

3476: \begin{eqnarray}

3477: n_1+2n_3=0, ~~~ n_2+\frac{7}{24}n_3+n_4=0.

3478: \label{allowed3}

3479: \end{eqnarray}

3480: respectively.

3481: 

3482: This fact that there are massless integer charged states possible

3483: in the models $X_{4,3}(1^5,2)$ and  $X_{4,3}(1^5,2)$ is entirely

3484: consistent with our physical picture, which explained the singular

3485: behaviour of the $F^{(g)}$ form the effective action point of

3486: view. What is very interesting is the fact that the period

3487: degeneration at the branch point of the hypersurface models,

3488: $X_5(1^5)$,  $X_6(1^4,2)$, $X_8(1^4,4)$ and  $X_{10}(1^3,2,5)$  

3489: is at genus zero very similar to the cases $X_{4,3}(1^5,2)$ and  $X_{6,2}(1^3,2^2,3)$.

3490: In particular the periods of the models have no logarithmic singularities. 

3491: The leading behaviour of the higher genus expansion, which we obtain 

3492: from the global properties, indicates that the BPS states,  

3493: which are possibly massless by (\ref{allowed1},\ref{allowed2},\ref{allowed3}),   

3494: are stable in  $X_{4,3}(1^5,2)$  and  $X_{6,2}(1^3,2^2,3)$ models, 

3495: but  not stable in the  $X_6(1^4,2)$ model.  

3496: 

3497: 

3498: %although it seems possible, the actual geometric cycles at the

3499: %orbifold point do not shrink faster than they generically do.

3500: 

3501: 

3502: \begin{thebibliography}{10}

3503: 

3504: \bibitem{Klemm:1999gm}

3505:   A.~Klemm and E.~Zaslow,

3506:    ``Local mirror symmetry at higher genus,''

3507:   arXiv:hep-th/9906046.

3508:   %%CITATION = HEP-TH 9906046;%%

3509: 

3510: 

3511: \bibitem{Aganagic:2003db}

3512:   M.~Aganagic, A.~Klemm, M.~Marino and C.~Vafa,

3513:    ``The topological vertex,''

3514:   Commun.\ Math.\ Phys.\  {\bf 254}, 425 (2005)

3515:   [arXiv:hep-th/0305132].

3516:   %%CITATION = HEP-TH 0305132;%%

3517: 

3518: \bibitem{Witten:1992fb}

3519:   E.~Witten,

3520:    ``Chern-Simons Gauge Theory As A String Theory,''

3521:   Prog.\ Math.\  {\bf 133}, 637 (1995)

3522:   [arXiv:hep-th/9207094].

3523:   %%CITATION = HEP-TH 9207094;%%

3524: 

3525: \bibitem{Gopakumar:1998ki}

3526:   R.~Gopakumar and C.~Vafa,

3527:    ``On the gauge theory/geometry correspondence,''

3528:   Adv.\ Theor.\ Math.\ Phys.\  {\bf 3}, 1415 (1999)

3529:   [arXiv:hep-th/9811131].

3530:   %%CITATION = HEP-TH 9811131;%%

3531: 

3532: \bibitem{Dijkgraaf:2002fc}

3533:   R.~Dijkgraaf and C.~Vafa,

3534:    ``Matrix models, topological strings, and supersymmetric gauge theories,''

3535:   Nucl.\ Phys.\ B {\bf 644}, 3 (2002)

3536:   [arXiv:hep-th/0206255].

3537:   %%CITATION = HEP-TH 0206255;%%

3538: 

3539: 

3540: 

3541: 

3542: \bibitem{Aganagic:2003qj}

3543:   M.~Aganagic, R.~Dijkgraaf, A.~Klemm, M.~Marino and C.~Vafa,

3544:    ``Topological strings and integrable hierarchies,''

3545:   Commun.\ Math.\ Phys.\  {\bf 261}, 451 (2006)

3546:   [arXiv:hep-th/0312085].

3547:   %%CITATION = HEP-TH 0312085;%%

3548: 

3549: \bibitem{Vafa:2004qa}

3550:   C.~Vafa,

3551:    ``Two dimensional Yang-Mills, black holes and topological strings,''

3552:   arXiv:hep-th/0406058.

3553:   %%CITATION = HEP-TH 0406058;%%

3554: 

3555: \bibitem{costello}

3556: K.~Costello,

3557: ``Topological conformal field theories and Calabi-Yau categories,''

3558: [math.QA/0412149].

3559: 

3560: \bibitem{kontsevich}

3561: M.~Kontsevich, ``TQFTs and geometry of pre-Frobenius manifolds'' and  ``$A_\infty$  categories

3562: and  their (co)homology theories,'' Lecures at the ``Erwin Schroedinger Institute,'' Vienna 2006.

3563: 

3564: \bibitem{Ooguri:2004zv}

3565:   H.~Ooguri, A.~Strominger and C.~Vafa,

3566:    ``Black hole attractors and the topological string,''

3567:   Phys.\ Rev.\ D {\bf 70}, 106007 (2004)

3568:   [arXiv:hep-th/0405146].

3569:   %%CITATION = HEP-TH 0405146;%%

3570: 

3571: \bibitem{Ooguri:2005vr}

3572:   H.~Ooguri, C.~Vafa and E.~P.~Verlinde,

3573:    ``Hartle-Hawking wave-function for flux compactifications,''

3574:   Lett.\ Math.\ Phys.\  {\bf 74}, 311 (2005)

3575:   [arXiv:hep-th/0502211].

3576:   %%CITATION = HEP-TH 0502211;%%

3577: 

3578: 

3579: %\cite{Huang:2006si}

3580: \bibitem{Huang}

3581:   M.~x.~Huang and A.~Klemm,

3582:   ``Holomorphic anomaly in gauge theories and matrix models,''

3583:   arXiv:hep-th/0605195.

3584:   %%CITATION = HEP-TH 0605195;%%

3585: 

3586: \bibitem{Huang2}

3587:   M.~x.~Huang and A.~Klemm,

3588:   to appear.

3589:   %%CITATION = HEP-TH 0605195;%%

3590: 

3591: \bibitem{Seiberg:1994rs}

3592:   N.~Seiberg and E.~Witten,

3593:    ``Electric - magnetic duality, monopole condensation, and confinement in N=2

3594:     supersymmetric Yang-Mills theory,''

3595:   Nucl.\ Phys.\ B {\bf 426}, 19 (1994)

3596:   [Erratum-ibid.\ B {\bf 430}, 485 (1994)]

3597:   [arXiv:hep-th/9407087].

3598:   %%CITATION = HEP-TH 9407087;%%.

3599: 

3600: 

3601: %\cite{Aganagic:2006wq}

3602: \bibitem{Aganagic:2006wq}

3603:   M.~Aganagic, V.~Bouchard and A.~Klemm,

3604:   ``Topological strings and (almost) modular forms,''

3605:   arXiv:hep-th/0607100.

3606:   %%CITATION = HEP-TH 0607100;%%

3607: 

3608: 

3609: 

3610: \bibitem{Hosono:1999qc}

3611:   S.~Hosono, M.~H.~Saito and A.~Takahashi,

3612:   %``Holomorphic anomaly equation and BPS state counting of rational  elliptic

3613:   %surface,''

3614:   Adv.\ Theor.\ Math.\ Phys.\  {\bf 3} (1999) 177

3615:   [arXiv:hep-th/9901151].

3616:   %%CITATION = HEP-TH 9901151;%%

3617: 

3618: \bibitem{Hosono:2001gf}

3619:   S.~Hosono, M.~H.~Saito and A.~Takahashi,

3620:   %``Relative Lefschetz Action and BPS State Counting,''

3621:   arXiv:math.ag/0105148.

3622:   %%CITATION = MATH-AG 0105148;%%

3623: 

3624: \bibitem{dijkgraafzagier} R.~Dijkgraaf, ``Mirror Symmetry and elliptic curves,'' in {\sl

3625: The moduli Space of Curves}, Progr. Math. {\bf 129}, 149;

3626: K.~Kaneko and D.~B.~Zagier, ``A generalized Jacobi theta functions

3627: and quasimodular forms,'' ibidem 165.

3628: 

3629: 

3630: %\cite{Candelas:1990rm}

3631: \bibitem{Candelas:1990rm}

3632:   P.~Candelas, X.~C.~De La Ossa, P.~S.~Green and L.~Parkes,

3633:    ``A pair of Calabi-Yau manifolds as an exactly soluble superconformal

3634:   theory,''

3635:   Nucl.\ Phys.\ B {\bf 359}, 21 (1991).

3636:   %%CITATION = NUPHA,B359,21;%%

3637: 

3638: %\cite{Yamaguchi:2004bt}

3639: \bibitem{Yamaguchi}

3640:   S.~Yamaguchi and S.~T.~Yau,

3641:   ``Topological string partition functions as polynomials,''

3642:   JHEP {\bf 0407}, 047 (2004)

3643:   [arXiv:hep-th/0406078].

3644:   %%CITATION = HEP-TH 0406078;%%

3645: 

3646: 

3647: %\cite{Bershadsky:1993ta}

3648: \bibitem{BCOVI}

3649:   M.~Bershadsky, S.~Cecotti, H.~Ooguri and C.~Vafa,

3650:    ``Holomorphic anomalies in topological field theories,''

3651:   Nucl.\ Phys.\ B {\bf 405}, 279 (1993)

3652:   [arXiv:hep-th/9302103].

3653:   %%CITATION = HEP-TH 9302103;%%

3654: 

3655: %\cite{Bershadsky:1993cx}

3656: \bibitem{BCOV}

3657:   M.~Bershadsky, S.~Cecotti, H.~Ooguri and C.~Vafa,

3658:    ``Kodaira-Spencer theory of gravity and exact results for quantum string

3659:   amplitudes,''

3660:   Commun.\ Math.\ Phys.\  {\bf 165}, 311 (1994)

3661:   [arXiv:hep-th/9309140].

3662:   %%CITATION = HEP-TH 9309140;%%

3663: 

3664: \bibitem{Hosono:1994ax}

3665:   S.~Hosono, A.~Klemm, S.~Theisen and S.~T.~Yau,

3666:    ``Mirror symmetry, mirror map and applications to complete intersection

3667:   Calabi-Yau spaces,''

3668:   Nucl.\ Phys.\ B {\bf 433}, 501 (1995)

3669:   [arXiv:hep-th/9406055].

3670:   %%CITATION = HEP-TH 9406055;%%

3671: 

3672: \bibitem{MB} K.~Hori, S.~Katz, A.~Klemm, R.Pandharipande, R.~Thomas, C.~Vafa, R.~Vavil and E. Zaslow,

3673: {\sl Mirror Symmetry}, AMS (2003).

3674: 

3675: \bibitem{Ceresole:1995jg}

3676:   A.~Ceresole, R.~D'Auria, S.~Ferrara and A.~Van Proeyen,

3677:    ``Duality transformations in supersymmetric Yang-Mills theories coupled to

3678:    supergravity,''

3679:   Nucl.\ Phys.\ B {\bf 444}, 92 (1995)

3680:   [arXiv:hep-th/9502072].

3681:   %%CITATION = HEP-TH 9502072;%% and the $K$-theory

3682: 

3683: \bibitem{Minasian:1997mm}

3684:   R.~Minasian and G.~W.~Moore,

3685:    ``K-theory and Ramond-Ramond charge,''

3686:   JHEP {\bf 9711}, 002 (1997)

3687:   [arXiv:hep-th/9710230].

3688:   %%CITATION = HEP-TH 9710230;%%

3689: 

3690: \bibitem{Cheung:1997az}

3691:   Y.~K.~Cheung and Z.~Yin,

3692:    ``Anomalies, branes, and currents,''

3693:   Nucl.\ Phys.\ B {\bf 517}, 69 (1998)

3694:   [arXiv:hep-th/9710206].

3695:   %%CITATION = HEP-TH 9710206;%%

3696: 

3697: \bibitem{Brunner:1999jq}

3698:   I.~Brunner, M.~R.~Douglas, A.~E.~Lawrence and C.~Romelsberger,

3699:    ``D-branes on the quintic,''

3700:   JHEP {\bf 0008}, 015 (2000)

3701:   [arXiv:hep-th/9906200].

3702:   %%CITATION = HEP-TH 9906200;%%.

3703: 

3704: 

3705: \bibitem{fulton} W.~Fulton, ``Intersection Theory,'' Springer (1998).

3706: 

3707: \bibitem{AGV} V.~I.~ Arnold, S.~M.~ Gusein-Zade, A.~N.~ Varchencko, ``Singularities of diefferential maps I\& II,''

3708: Birkh\"auser, Basel (1985).

3709: 

3710: 

3711: \bibitem{Diaconescu:1999vp}

3712:   D.~E.~Diaconescu and C.~Romelsberger,

3713:   %``D-branes and bundles on elliptic fibrations,''

3714:   Nucl.\ Phys.\ B {\bf 574}, 245 (2000)

3715:   [arXiv:hep-th/9910172].

3716:   %%CITATION = HEP-TH 9910172;%%

3717: 

3718: \bibitem{Mayr:2000as}

3719:   P.~Mayr,

3720:    ``Phases of supersymmetric D-branes on Kaehler manifolds and the McKay

3721:    correspondence,''

3722:   JHEP {\bf 0101}, 018 (2001)

3723:   [arXiv:hep-th/0010223].

3724:   %%CITATION = HEP-TH 0010223;%%

3725: 

3726: %\cite{Katz:1999xq}

3727: \bibitem{Katz:1999}

3728:   S.~Katz, A.~Klemm and C.~Vafa,

3729:   ``M-theory, topological strings and spinning black holes,''

3730:   Adv.\ Theor.\ Math.\ Phys.\  {\bf 3}, 1445 (1999)

3731:   [arXiv:hep-th/9910181].

3732:   %%CITATION = HEP-TH 9910181;%%

3733: 

3734: %\cite{Ghoshal:1995wm}

3735: \bibitem{Ghoshal:1995}

3736:   D.~Ghoshal and C.~Vafa,

3737:   ``C = 1 String As The Topological Theory Of The Conifold,''

3738:   Nucl.\ Phys.\ B {\bf 453}, 121 (1995)

3739:   [arXiv:hep-th/9506122].

3740:   %%CITATION = HEP-TH 9506122;%%

3741: 

3742: %\cite{Marino:1998pg}

3743: \bibitem{Marino:1998}

3744:   M.~Marino and G.~W.~Moore,

3745:   ``Counting higher genus curves in a Calabi-Yau manifold,''

3746:   Nucl.\ Phys.\ B {\bf 543}, 592 (1999)

3747:   [arXiv:hep-th/9808131].

3748:   %%CITATION = HEP-TH 9808131;%%

3749: 

3750: %\cite{Klemm:2005pd}

3751: \bibitem{Klemm:2005pd}

3752:   A.~Klemm and M.~Marino,

3753:    ``Counting BPS states on the Enriques Calabi-Yau,''

3754:   arXiv:hep-th/0512227.

3755:   %%CITATION = HEP-TH 0512227;%%

3756: 

3757: 

3758: %\cite{Gopakumar:1998ii}

3759: \bibitem{GVI}

3760:   R.~Gopakumar and C.~Vafa,

3761:   ``M-theory and topological strings. I,''

3762:   arXiv:hep-th/9809187.

3763: 

3764: \bibitem{GVII}

3765:   R.~Gopakumar and C.~Vafa,

3766:  ``M-theory and topological strings. II,''

3767:    hep-th/9812127.

3768:   %%CITATION = HEP-TH 9809187;%%

3769: 

3770: %\cite{Faber:2000ma}

3771: \bibitem{FP}

3772:   C.~Faber and R.~Pandharipande,

3773:    ``Hodge Integrals and Gromov-Witten theory,''

3774:   arXiv:math.ag/9810173.

3775:   %%CITATION = MATH-AG 0002112;%%

3776: 

3777: %\cite{GKMW}

3778: \bibitem{GKMW}

3779: T.~Grimm, A.~Klemm, M.~Marino and M.~Weiss, work in progress.

3780: 

3781: \bibitem{Vafa:1995ta}

3782:   C.~Vafa,

3783:    ``A Stringy test of the fate of the conifold,''

3784:   Nucl.\ Phys.\ B {\bf 447}, 252 (1995)

3785:   [arXiv:hep-th/9505023].

3786:   %%CITATION = HEP-TH 9505023;%%

3787: 

3788: 

3789: \bibitem{wittensurvey} E.~Witten, ``Two-Dimensional Gravity and intersection theory on  moduli spaces,''

3790: Surveys in Diff. Geometry {\bf 1} (1991) 243.

3791: 

3792: 

3793: 

3794: \bibitem{Strominger:1995cz}

3795:   A.~Strominger,

3796:    ``Massless black holes and conifolds in string theory,''

3797:   Nucl.\ Phys.\ B {\bf 451}, 96 (1995)

3798:   [arXiv:hep-th/9504090].

3799:   %%CITATION = HEP-TH 9504090;%%

3800: 

3801: 

3802: \bibitem{GVbfg} R. Gopakumar and C. Vafa, ``Branes and fundamental Groups,''

3803:   [arXiv:hep-th/9712048].

3804: 

3805: 

3806: 

3807: \bibitem{Klemm:1995kj}

3808:   A.~Klemm and P.~Mayr,

3809:    ``Strong coupling singularities and non-Abelian gauge symmetries in  N = 2

3810:   string theory,'' Nucl.\ Phys.\ B {\bf 469}, 37 (1996)

3811:   [arXiv:hep-th/9601014].

3812: 

3813: 

3814: 

3815: \bibitem{Klemm:1992tx} A.~Klemm and S.~Theisen,

3816:    ``Considerations of one modulus Calabi-Yau compactifications: Picard-Fuchs

3817:    equations, Kahler potentials and mirror maps,''

3818:   Nucl.\ Phys.\ B {\bf 389}, 153 (1993)

3819:   [arXiv:hep-th/9205041].

3820:   %%CITATION = HEP-TH 9205041;%%

3821: 

3822: \bibitem{DTGW} D.~Maulik, N.~Nekrasov, A.~Okounkov and R.~Pandharipande,

3823:         {\sl Gromov-Witten theory and Donaldson-Thomas theory I,}

3824:         [arXiv:math.AG/0312059].

3825: 

3826: 

3827: 

3828: \bibitem{Verlinde:2004ck}

3829:   E.~P.~Verlinde,

3830:   ``Attractors and the holomorphic anomaly,''

3831:   arXiv:hep-th/0412139.

3832:   %%CITATION = HEP-TH 0412139;%%

3833: 

3834: 

3835: 

3836: \bibitem{Gunaydin:2006bz}

3837:   M.~Gunaydin, A.~Neitzke and B.~Pioline,

3838:   ``Topological wave functions and heat equations,''

3839:   arXiv:hep-th/0607200.

3840:   %%CITATION = HEP-TH 0607200;%%

3841: 

3842: \bibitem{cyy} Y.-H. Chen, Y. Yang, N. Yui, ``Monodromy of Picard-Fuchs

3843: differential equations for Calabi-Yau threefolds,'' [arXiv:math.AG/0605675].

3844: 

3845: 

3846: 

3847: \bibitem{CR} W.~Chen and Y.~Ruan, ``Orbifold Gromow-Witten theory'', Orbifolds in mathmatics and physics

3848: Madison WI 2001, Contemp. Math., vol 310, Amer. Math. Soc., Providence, RI, 2002 25-85.

3849: 

3850: 

3851: \bibitem{BGP} J.~Bryan, T.~Graber, R.~Pandharipande, ``The orbifold quantum

3852: cohomology of $C^2/Z_3 $ and Hurwitz-Hodge integrals,'' [arXiv:math.AG/0510335].

3853: 

3854: \bibitem{CCIT} T.~Coates, A.~Corti, H.~Iritani and H.-H. Tseng, ``Wall-Crossing in Toric Gromow-Witten Theory I:

3855: Crepant Examples'', [arXiv:math.AG/0611550].

3856: 

3857: 

3858: \bibitem{wittenwavefunction} E.~Witten, ``Quantum background independence

3859: in string theory,'' Nucl. Phys. B {\bf 372}, 187 (1992), [arXiv:hep-th/9306122].

3860: 

3861: \bibitem{GiventhalCoates} A.~B.~Giventhal, ``Symplectic geometry of Frobenius structures,''

3862: Frobenius manifolds, Aspects. Math., E36, Vieweg, Wiesbaden,

3863: (91-112) 2004, T.~Coates, ``Giventhals Lagrangian Cone and

3864: $S^1$-Equivariant Gromow-Witten Theory,'' [arXiv:math.AG.0607808].

3865: 

3866: \bibitem{DoranMorgan} C.~Doran and J~Morgan, ``Mirror Symmetry and Integral Variations of

3867:     Hodge Structure Underlying One Parameter Families of Calabi-Yau Threefolds,'' [arXiv:math.AG/0505272].

3868: 

3869: \bibitem{Lazaroiu:2000jx}

3870:   C.~I.~Lazaroiu,

3871:   ``Collapsing D-branes in one-parameter models and small/large radius

3872:    duality,''

3873:   Nucl.\ Phys.\ B {\bf 605}, 159 (2001)

3874:   [arXiv:hep-th/0002004].

3875:   %%CITATION = HEP-TH 0002004;%%

3876: 

3877: \bibitem{KKRS} A.~Klemm, M.~Kreuzer, E.~Riegler and E.~Scheidegger,

3878:   ``Topological string amplitudes, complete intersection Calabi-Yau spaces  and

3879:   threshold corrections,''

3880:   JHEP {\bf 0505}, 023 (2005)

3881:   [arXiv:hep-th/0410018].

3882:   %%CITATION = HEP-TH 0410018;%%

3883: 

3884: \bibitem{Schmidt} W.~Schmidt, ``Variaton of Hodge structure: The singularities of the period mapping,''

3885:         Inventiones Math.{\bf 22} (1973) 211-319.

3886: 

3887: \bibitem{CKS} E.~ Cattani, A.~Kaplan and W.~Schmidt, ``Degenerarions of Hodge Structures,''

3888: Annals of Math. {\bf 123} (1986)  457-535.

3889: 

3890: \bibitem{Harris} J.~Harris, ``Curves in Projective Space,'' Sem. de Math. Sup. Les Presses de l'University

3891: de Monteral (1982).

3892: 

3893: \bibitem{webpage} {\tt http://uw.physics.wisc.edu/$\sim$strings/aklemm/highergenusdata/}

3894: 

3895: \bibitem{Bridgeland}  T.~Bridgeland, ``Derived categories of coherent sheaves,''

3896: Survey article for ICM 2006,  [arXiv:math.AG/0602129].

3897: 

3898: \bibitem{Aspinwall}

3899:   P.~S.~Aspinwall,

3900:   ``D-branes on Calabi-Yau manifolds,''

3901:   arXiv:hep-th/0403166.

3902:   %%CITATION = HEP-TH 0403166;%%

3903: 

3904: 

3905: \bibitem{Schubert} S.~Katz, S. Str{\o}mme and J-M. {\O}kland: ``Schubert: A Maple package

3906: for intersection theory and enumerative geometry,'' {\tt http://www.uib.no/People/nmasr/schubert/}.

3907: 

3908: \bibitem{Gaiotto:2006wm}

3909:   D.~Gaiotto, A.~Strominger and X.~Yin,

3910:   ``The M5-brane elliptic genus: Modularity and BPS states,''

3911:   arXiv:hep-th/0607010.

3912:   %%CITATION = HEP-TH 0607010;%%

3913: 

3914: \bibitem{deBoer:2006vg}

3915:   J.~de Boer, M.~C.~N.~Cheng, R.~Dijkgraaf, J.~Manschot and E.~Verlinde,

3916:   ``A farey tail for attractor black holes,''

3917:   JHEP {\bf 0611}, 024 (2006)

3918:   [arXiv:hep-th/0608059].

3919:   %%CITATION = HEP-TH 0608059;%% using the dilute gas approximation in $AdS_3\times S^2\times M$.

3920: 

3921: \bibitem{DennefMoore} F.~Dennef and G.~Moore, ``From OSV to OSV,'' talk available at

3922: {\tt http://strings06.itp.ac.cn/} and work in progress.

3923: 

3924: 

3925: \bibitem{Rocek:2005ij}

3926:   M.~Rocek, C.~Vafa and S.~Vandoren,

3927:   ``Hypermultiplets and topological strings,''

3928:   JHEP {\bf 0602} (2006) 062

3929:   [arXiv:hep-th/0512206].

3930:   %%CITATION = HEP-TH 0512206;%%.

3931: 

3932: \end{thebibliography}

3933: 

3934: 

3935: 

3936: 

3937: 

3938: \end{document}

3939: