1: \documentclass[12pt, letterpaper]{JHEP3}

2: 

3: \usepackage{amssymb, amsmath, amsopn, amsthm}

4: \usepackage{epsfig}

5: %\usepackage{latexsym}

6: %\usepackage{graphics}

7: %\usepackage{epsfig}

8: 

9: \title{Nongeometric Flux Compactifications}

10: 

11: \author{Jessie Shelton, Washington Taylor, and Brian Wecht \\

12: {Center for Theoretical Physics} \\ {Massachusetts Institute of Technology} \\

13: {Cambridge, MA 02139, U.S.A.} \\ {\tt jshelton\ {\rm at}\ mit.edu, wati\

14:   {\rm at}\ mit.edu, bwecht\ {\rm at}\ mit.edu}}

15: 

16: \abstract{We investigate a simple class of type II string

17: compactifications which incorporate nongeometric ``fluxes'' in

18: addition to ``geometric flux'' and the usual $H$-field and R-R fluxes.

19: These compactifications are nongeometric analogues of the twisted

20: torus.  We develop T-duality rules for NS-NS geometric and

21: nongeometric fluxes, which we use to construct a superpotential for

22: the dimensionally reduced four-dimensional theory.  The resulting

23: structure is invariant under T-duality, so that the distribution of

24: vacua in the IIA and IIB theories is identical when nongeometric

25: fluxes are included.  This gives a concrete framework in which to

26: investigate the possibility that generic string compactifications may

27: be nongeometric in any duality frame.  The framework developed in

28: this paper also provides some concrete hints for how mirror symmetry

29: can be generalized to compactifications with arbitrary $H$-flux, whose

30: mirrors are generically nongeometric.}

31: 

32: %\keywords{Flux compactification,  Moduli stabilization, String vacua}

33: \preprint{hep-th/0508133, MIT-CTP-3673}

34: 

35: \newcommand{\NP}{{ Nucl.\ Phys.\ }}

36: \newcommand{\AP}{{Ann.\ Phys.\ }}

37: \newcommand{\PL}{{Phys.\ Lett.\ }}

38: \newcommand{\PR}{{Phys.\ Rev.\ }}

39: \newcommand{\PRP}{{Phys.\ Rep.\ }}

40: \newcommand{\CMP}{{Comm.\ Math.\ Phys.\ }}

41: \newcommand{\MPL}{{Mod.\ Phys.\ Lett.\ }}

42: \newcommand{\PRL}{{Phys.\ Rev.\ Lett.\ }}

43: \newcommand{\IJMP}{{Int.\ J.\ Mod.\ Phys.\ }}

44: 

45: \newcommand{\ket}[1]{\mathchoice{

46:     {\left|{#1}\right\rangle}}{|{#1}\rangle}{|{#1}\rangle}{|{#1}\rangle}}

47: \newcommand{\bra}[1]{\left\langle{#1}\right|}

48: \newcommand{\ibra}[2]{\sideset{#2}{}\la{#1}\rvert}

49: \newcommand{\ld}{\lambda}

50: \newcommand{\g}{\gamma}

51: \newcommand{\G}{\Gamma}

52: \newcommand{\pd}{\partial}

53: 

54: \newcommand{\half}{\frac{1}{2}}

55: \newcommand{\quarter}{\frac{1}{4}}

56: 

57: \newcommand{\cA}{\mathcal{A}}

58: \newcommand{\cB}{\mathcal{B}}

59: \newcommand{\cC}{\mathcal{C}}

60: \newcommand{\cD}{\mathcal{D}}

61: \newcommand{\cE}{\mathcal{E}}

62: \newcommand{\cF}{\mathcal{F}}

63: \newcommand{\cG}{\mathcal{G}}

64: \newcommand{\cH}{\mathcal{H}}

65: \newcommand{\cI}{\mathcal{I}}

66: \newcommand{\cJ}{\mathcal{J}}

67: \newcommand{\cK}{\mathcal{K}}

68: \newcommand{\cL}{\mathcal{L}}

69: \newcommand{\cM}{\mathcal{M}}

70: \newcommand{\cN}{\mathcal{N}}

71: \newcommand{\cO}{\mathcal{O}}

72: \newcommand{\cP}{\mathcal{P}}

73: \newcommand{\cQ}{\mathcal{Q}}

74: \newcommand{\cR}{\mathcal{R}}

75: \newcommand{\cS}{\mathcal{S}}

76: \newcommand{\cT}{\mathcal{T}}

77: \newcommand{\cU}{\mathcal{U}}

78: \newcommand{\cV}{\mathcal{V}}

79: \newcommand{\cW}{\mathcal{W}}

80: \newcommand{\cX}{\mathcal{X}}

81: \newcommand{\cY}{\mathcal{Y}}

82: \newcommand{\cZ}{\mathcal{Z}}

83: 

84: \newcommand{\bN}{\mathbb{N}}

85: \newcommand{\bC}{\mathbb{C}}

86: \newcommand{\bP}{\mathbb{P}}

87: \newcommand{\bZ}{\mathbb{Z}}

88: \newcommand{\bR}{\mathbb{R}}

89: 

90: \newcommand{\td}{\tilde}

91: \renewcommand{\Re}{{\rm Re}\,}

92: \renewcommand{\Im}{{\rm Im}\,}

93: \newcommand{\vol}{{\rm vol}}

94: \newcommand{\sgn}{{\rm sgn}\,}

95: \newcommand{\beq}{\begin{equation}}

96: \newcommand{\eeq}{\end{equation}}

97: \newcommand{\tc}{\td{c}}

98: \newcommand{\er}[1]{(\ref{eq:#1})}

99: 

100: 

101: \DeclareMathOperator{\Det}{Det}

102: \DeclareMathOperator{\Log}{Log}

103: \DeclareMathOperator{\la}{\langle}

104: \DeclareMathOperator{\ra}{\rangle}

105: \DeclareMathOperator{\Tr}{Tr}

106: \DeclareMathOperator{\Str}{Str}

107: 

108: \begin{document}

109: \tableofcontents

110: \section{Introduction}

111: \label{sec:}

112: 

113: Since the early days of string theory, it has been clear that there

114: are many possible ways in which to compactify the various perturbative

115: superstring and supergravity theories from ten or eleven dimensions to

116: four space-time dimensions.  For example, compactifying any

117: ten-dimensional string theory on a Calabi-Yau complex three-fold leads

118: to a supersymmetric theory of gravity coupled to light fields in the

119: remaining four macroscopic space-time dimensions.  Moduli

120: parameterizing the size and shape of the Calabi-Yau appear as massless

121: scalar fields in the four-dimensional theory.  Understanding and

122: classifying the range of possible compactifications is an important

123: part of the program of relating superstring theory to observed

124: phenomenology and cosmology.  In recent years, compactifications with

125: topologically quantized fluxes wrapping compact cycles on the

126: compactification manifold have become a subject of much interest,

127: following the work of \cite{flux-c, gvw, drs}.  The topological fluxes

128: produce a 

129: potential for the scalar moduli, and can thus ``stabilize'' some or

130: all of the moduli to take specific values \cite{gkp, kst, fp}.  Once

131: fluxes are added to the system, however, the geometric structure of

132: the compactification manifold may also become more general.  Recent

133: work has addressed the generalization to superstring 

134: compactifications on non-Calabi-Yau geometries \cite{Hitchin,glmw,fmt,su3}.

135: 

136: The goal of the present work is to take the study of flux

137: compactifications one step further, by including ``compactifications''

138: which cannot be described by a geometric ten-dimensional space-time

139: manifold.  It was argued in \cite{kstt} that nongeometric flux

140: compactifications arise naturally as configurations which are T-dual

141: to known geometric supersymmetric flux compactifications.  To be

142: specific, consider for example a compactification on a six-torus $T^6$

143: with NS-NS 3-form flux $H_{abc}$ on some three-cycle, where indices

144: $a, b, \ldots\in\{1, 2, \ldots, 6\}$ take values in the compact

145: directions.  Under a single T-duality, say in direction $a$, this

146: $H$-flux is mapped to ``geometric flux'' associated with a twist in

147: the torus topology.  In the presence of this geometric flux, the

148: metric on the twisted torus acquires a contribution which can be

149: written as $(dx^a - f^a_{bc}x^c dx^b)^2$, where $f^a_{bc}$ is

150: integrally quantized and characterizes the ``geometric flux'' of the

151: compactification.  This kind of ``twisted torus'' has been studied as

152: a type of Scherk-Schwarz compactification for many years \cite{ss, early-ss-reduction, km},

153: and is considered in the context of flux compactification in

154: \cite{glmw,Schulz,dkpz, x1, hre, vz,cfi}, among others.   

155: Even in the presence of

156: geometric flux $f^a_{bc }$, however, as was pointed out in

157: \cite{kstt}, we can perform another T-duality on direction $b$, since

158: the metric can be chosen to be independent of the coordinate $x^b$.

159: Carrying out this T-duality explicitly leads to a dual ``torus'',

160: which is locally geometric, but which cannot be described globally in

161: terms of a fixed geometry, due to the appearance of a nongeometric

162: duality transformation in the boundary conditions which patch together

163: local descriptions of the compactification space.  Nongeometric spaces

164: of this type were considered in \cite{dh, Hellerman:2002ax, fww, x2, gh}.  In this paper we

165: label the nongeometric flux resulting from T-duality $T_b$ by

166: $Q^{ab}_c$, and we determine how fluxes $Q^{ab}_c$ can be incorporated in

167: the superpotential for a simple T-duality invariant class of IIA and

168: IIB compactifications.  Note that although we use T-duality to

169: determine the role of the fluxes in the superpotential, a generic

170: configuration with fluxes $H, f, Q$ turned on cannot be T-dualized to

171: a completely geometric compactification.

172: 

173: After performing the second T-duality to a

174: configuration with nongeometric flux $Q^{ab}_c$, there is no apparent

175: residual

176: isometry around direction $c$.  Na\"ively it does not seem that a

177: further T-duality can be performed.  Nonetheless, we find a structure

178: suggesting that there is some meaning which can be given to a T-dual

179: flux of this type, which we label $R^{abc}$.  While we do not have an

180: explicit presentation of a nongeometric compactification with such

181: nongeometric fluxes, it seems that this structure should have some

182: meaning in any background-independent formulation of the theory.

183: The situation here is analogous to that for R-R fluxes.  The Buscher

184: rules \cite{Buscher, bho, mo} for T-duality

185: act on the $p$-form R-R fields $A^{(p)}$, and therefore cannot be used

186: to explicitly construct configurations with R-R flux $F^{(0)}$, just

187: as they cannot be used to construct $R^{abc}$, which acts formally as an NS-NS

188: 0-form flux.   In the case of $F^{(0)}$, it is necessary to use

189: T-duality rules which act directly on the R-R fluxes

190: \cite{ght,Hassan,kst} to map $F^{(1)} \rightarrow F^{(0)}$. 

191: Acting on the integrated R-R fluxes, these T-duality rules

192: take

193: \begin{equation}

194: F_{x\alpha_1 \cdots \alpha_p}

195: \stackrel{T_x}{\longleftrightarrow}

196: F_{\alpha_1 \cdots \alpha_p} \,.

197: \label{eq:ft}

198: \end{equation}

199: The T-duality rules we construct in this paper for nongeometric

200: fluxes, which take

201: \begin{equation}

202: H_{abc} \stackrel{T_a}{\longleftrightarrow} f^a_{bc}

203: \stackrel{T_b}{\longleftrightarrow} Q^{ab}_c

204: \stackrel{T_c}{\longleftrightarrow} R^{abc},

205: \label{eq:t-chain}

206: \end{equation}

207: can be thought of as an extension of (\ref{eq:ft}) to a general class

208: of integral NS-NS fluxes.  The structure we find here for the simple

209: toroidal example suggests that in addition to $H$-form flux, both

210: geometric and nongeometric fluxes can be thought of as additional

211: algebraic structure added to a given string background.  In the

212: geometric context, this kind of structure for geometric fluxes as data

213: which decorate a Calabi-Yau seems to arise naturally when the mirror

214: of a Calabi-Yau with $H$-flux has a geometric description \cite{glmw,

215: fmt}.  Another perspective on the geometric and nongeometric structures

216: arising from the T-dualities in (\ref{eq:t-chain}) is given in

217: \cite{Mathai}.

218: 

219: The approach taken in this paper is as follows: In Section

220: \ref{sec:superpotential}, we incorporate nongeometric fluxes by

221: using

222: T-duality and coordinate transformations to construct the complete set

223: of terms which may appear in the superpotential for a class of

224: compactifications based on a symmetric $T^6 = (T^2)^3$ with all $T^2$

225: components identical.  The resulting simple polynomial superpotential

226: subsumes the previously known superpotentials for geometric

227: compactifications of the IIA and IIB theories, and extends them to a

228: T-duality invariant form which includes both nongeometric $Q^{ab}_c$

229: and $R^{abc}$ types of ``fluxes.''  In Section

230: \ref{sec:interpretation} we give a more detailed discussion of the

231: interpretation of $Q^{ab}_c$ and $R^{abc}$ fluxes, and provide an

232: explicit description of ``T-folds'' \cite{x2} with $Q$-structure.  In Section

233: \ref{sec:constraints} we discuss constraints on the fluxes, which

234: arise both from tadpole cancellation requirements in the presence of

235: orientifolds, and from Bianchi-type identities.

236: Finally, in Section \ref{sec:conclusions} we conclude and discuss some

237: directions for future research.

238: 

239: 

240: 

241: \section{Nongeometric fluxes and the superpotential}

242: \label{sec:superpotential}

243: 

244: In this section we use various dualities to explicitly construct a

245: simple low-energy effective theory governing a class of geometric and

246: nongeometric compactifications.  We begin by considering the

247: compactification of type IIA and type IIB string theory on a torus

248: $T^6= (T^2)^3$, where each $T^2$ factor represents an identical torus.

249: We can think of this as a compactification on a $T^6$ with additional

250: discrete symmetries imposed.  We impose a symmetry under $\bZ_2$ which

251: reflects the first two 2-tori under (-1, -1, -1, -1, 1, 1) and a

252: further symmetry under a $\bZ_3$ which rotates the tori $T^2_{(1)}

253: \rightarrow T^2_{(2)} \rightarrow T^2_{(3)} \rightarrow T^2_{(1)}$.

254: In the IIB theory we then have a single complex structure modulus

255: $\tau$ parameterizing the complex structure of the $T^2$, a single

256: K\"ahler modulus $U$ containing the $C_4 $ modulus and the scale of

257: the $T^2$ and an axiodilaton $S$.  In the IIA theory which arises

258: after 3 T-dualities (one on each $T^2$), $\tau$ becomes the K\"ahler

259: modulus and $U$ becomes the complex structure modulus \cite{vz,cfi}.

260: Flux compactifications of this type were considered in type IIB in

261: \cite{kst,fp,x3,dgkt1} and in IIA in \cite{vz,cfi}.  A slightly more

262: general model was also considered in these papers, where the three

263: complex structure and K\"ahler moduli are allowed to vary

264: independently by imposing a second $\bZ_2$ symmetry (1, 1, -1, -1, -1,

265: -1) instead of the $\bZ_3$ we use here.  This model can be seen as a

266: special case of that $T^6/\bZ_2^2$ model.  It is straightforward to

267: generalize the considerations here to that more general model, with

268: slightly more algebra.

269: 

270: We wish to include various kinds of fluxes on the $T^6$.  When an

271: orientifold is included to cancel tadpoles,

272: these

273: fluxes lead generally to a low-energy effective ${\cal N} = 1$ supergravity

274: theory in four dimensions which has a superpotential $W(\tau, U, S)$, a K\"ahler

275: potential $K (\tau, U, S)$, and a resulting potential for the moduli

276: given by

277: \begin{equation}

278: V = e^K \left(\sum_{i,j, = \{\tau, U, S \}} K^{ij} D_i W \overline{D_j

279:   W} - 3 |W|^2 \right)\,,

280: \end{equation}

281: where $K^{ij}$ is the inverse of $K_{ij} =  \partial_i

282: \bar{\partial_{j}} K$.  The construction of  flux compactifications

283: with

284: orientifolds was developed in \cite{gvw,drs}, and applied to the IIB

285: theory in \cite{gkp} and many subsequent papers, and to the IIA theory

286: in

287: \cite{dkpz, vz,cfi,x4, x5, dgkt2}.

288: 

289: An important caveat which must be taken into account when describing

290: flux compactifications through the dimensionally reduced

291: four-dimensional theory is that the low-energy four-dimensional

292: supergravity action is only valid when the moduli acquire masses which

293: are small compared to those of fields such as higher string modes,

294: winding modes, and

295: Kaluza-Klein modes which are neglected in the dimensional reduction.

296: This issue must be addressed in any study of flux compactifications.

297: In the class of vacua we consider here, which are not geometric, and

298: for which ten-dimensional supergravity is not a valid approximation,

299: this question becomes even more subtle.  For the present, we will

300: simply describe the superpotential for the four-dimensional

301: supergravity theory as a function of the degrees of freedom associated

302: with the original moduli on the torus.  This allows us at least to

303: characterize the topological features of the nongeometric fluxes in

304: which we are interested.  We leave a more detailed study of the regime

305: of validity of the low-energy theory in the presence of nongeometric

306: fluxes to further work.

307: 

308: The  particular symmetric torus $T^6 = (T^2)^3$ model

309: we are interested in here was studied in \cite{kst, x3}  and explicitly

310: solved

311: in \cite{dgkt1} for IIB

312: compactifications with R-R and NS-NS form field flux.  In this case the

313: superpotential is given by

314: \begin{equation}

315: W_{{\rm IIB}} = P_1 (\tau) + SP_2 (\tau)

316: \label{eq:w-IIB}

317: \end{equation}

318: where $P_{1, 2} (\tau)$ are cubic polynomials in $\tau$.  The K\"ahler

319: potential is

320: \begin{equation}

321: K = -3\ln (-i (\tau -\bar{\tau}))

322:  -\ln (-i (S -\bar{S}))

323:  -3\ln (-i (U -\bar{U})) \,.

324: \label{eq:Kaehler}

325: \end{equation}

326: 

327: In \cite{vz,cfi} this model was studied for the IIA theory, where in

328: addition to NS-NS and R-R form field fluxes, geometric flux was

329: also allowed.  In this case the K\"ahler potential is again

330: (\ref{eq:Kaehler}), while the superpotential is

331: \begin{equation}

332: W_{{\rm IIA}} = P_1 (\tau) + SP_2 (\tau) + UP_3 (\tau),

333: \label{eq:w-IIA}

334: \end{equation}

335: with $P_1$ again cubic, but with $P_{2, 3}$ now linear in $\tau$.

336: 

337: In order to consider a complete  T-duality invariant family of flux

338: compactifications we must extend somewhat the nature of allowed

339: fluxes.  As discussed in the Introduction, we must include not

340: only geometric fluxes but also some structures we interpret as

341: ``nongeometric fluxes''.

342: Simply using T-duality and coordinate symmetries, we can proceed to

343: construct the full duality-invariant superpotential $W$, identifying the

344: fluxes corresponding to each term in $W$.  We now proceed to directly

345: present this superpotential, which is one of the main results of this paper,

346: after which we give a more detailed discussion of how the various

347: terms are derived through dualities.  The later sections of the paper

348: discuss the interpretation of the fluxes which appear in this

349: potential, and constraints on these quantized fluxes.

350: 

351: We claim that the full potential for the symmetric torus in both the

352: IIA and IIB theories is given by

353: \begin{equation}

354: W_{{\rm complete}} = P_1 (\tau) + SP_2 (\tau) + UP_3 (\tau),

355: \label{eq:w-complete}

356: \end{equation}

357: where now all three of $P_{1, 2, 3} (\tau)$ are cubic polynomials.

358: The coefficients in these polynomials are given in the IIB theory by

359: (integrally quantized)

360: NS-NS and R-R fluxes $\bar{H}_{abc}, \bar{F}_{abc}$ 

361: (denoting the integral number of units of flux of, {\it e.g.},

362: $F_{abc}$ by $\bar{F}_{abc}$)

363: and 

364: also by ``nongeometric''

365: fluxes $Q^{ab}_c$ (which each can individually arise as the T-dual on

366: direction $b$ of the geometric

367: flux  $f^a_{bc}$).  In the IIA theory, the coefficients include 

368: (integrally quantized) R-R

369: $p$-form fluxes $F^{(0)}, \bar{F}_{ab}, \bar{F}_{abcd}, \bar{F}_{abcdef}$, as well as

370: NS-NS 3-form flux $\bar{H}_{abc}$, geometric fluxes $f^a_{bc}$,

371: nongeometric fluxes $Q^{ab}_c$, and further nongeometric fluxes

372: $R^{abc}$ (which can individually be seen formally as the T-dual on $c$ of

373: $Q^{ab}_c$).  In the next section we discuss the interpretation of

374: these nongeometric fluxes in more detail.  For now, however, we will

375: simply show how dualities determine which fluxes arise as which

376: coefficients in the superpotential (\ref{eq:w-complete}).

377: 

378: 

379: To make the discussion more explicit, we label coordinates $1, 3, 5$

380: on the $T^6$ with indices $\alpha, \beta, \gamma$ and coordinates $2,

381: 4, 6$ with indices $i, j, k$.  The IIB torus is taken to have an

382: O3-plane filling four-dimensional space-time, so that all internal coordinates are

383: odd under the orientifold reflection $\Omega$.  To get to the IIA

384: theory we T-dualize on the dimensions $1, 3, 5$, in that order,

385: so that the resulting

386: IIA O6-plane extends along  these dimensions with indices $\alpha,

387: \beta, \gamma$.  In the following table, % Table~\ref{t:coefficients}

388: we list the fluxes associated with each term in the superpotential

389: (\ref{eq:w-complete}) in both IIA and IIB.  

390: %\TABLE{

391: \begin{center}

392: \begin{tabular}{ || c  || c || c || c ||}

393: \hline

394: \hline

395: Term & IIA flux & IIB flux & integer flux\\

396: \hline

397: \hline

398: $1 $& $ \bar{F}_{\alpha i \beta j \gamma k}$& $ \bar{F}_{ ijk} $& $  a_0 $\\

399: \hline

400: $\tau $& $ \bar{F}_{\alpha i \beta j} $& $\bar{F}_{ ij \gamma} $& $   a_1 $\\

401: \hline

402: $\tau^2 $& $ \bar{F}_{\alpha i} $& $\bar{F}_{i \beta \gamma} $& $  a_2 $\\

403: \hline

404: $\tau^3 $& $ F^{(0)}  $& $\bar{F}_{\alpha \beta \gamma} $& $  a_3 $\\

405: \hline

406: $S $& $ \bar{H}_{ijk} $& $ \bar{H}_{ ijk} $& $  b_0$\\

407: \hline

408: $U $& $ \bar{H}_{\alpha \beta k} $& $  Q^{\alpha \beta}_k $& $  c_0 $\\

409: \hline

410: $S\tau $& $ f^\alpha_{j k} $& $ \bar{H}_{\alpha jk} $& $  b_1 $\\

411: \hline

412: $U\tau $& $ f ^ j_{k\alpha}, f^i_{\beta k}, f^\alpha_{\beta \gamma} $& $ 

413: Q ^{\alpha j}_k, Q^{i \beta}_k, Q^{\beta \gamma}_\alpha $&

414: $\check{c}_1,  \hat{c}_1,\tilde {c}_1 $\\

415: \hline

416: \hline

417: $S \tau^2 $& $ Q^{\alpha \beta}_k $& $ \bar{H}_{ i \beta \gamma} $& $  b_2 $\\

418: \hline

419: $U \tau^2 $& $  Q ^{\gamma i}_\beta, Q^{i \beta}_\gamma, Q^{ij}_k $& $ Q ^{i\beta}_\gamma, Q^{\gamma i}_\beta,

420: Q^{ij}_k $& $\check{c}_2,  \hat{c}_2,\tilde{c}_2 $\\

421: \hline

422: $S\tau^3 $& $  R^{\alpha \beta \gamma} $& $ \bar{H}_{\alpha \beta \gamma} $& $  b_3 $\\

423: \hline

424: $U \tau^3 $& $ R^{ij \gamma} $& $  Q^{ij}_{\gamma} $& $c_3 $\\

425: \hline

426: \hline

427: \end{tabular}

428: \vskip0.5cm

429: {\small {\bf Table 1}:\,\,Fluxes appearing as coefficients of terms in the superpotential.} 

430: \end{center}

431: 

432: To be explicit about the index structure in this table, notice

433: that the orbifold projection we have chosen implies that all objects with

434: three indices must have one index on each $T^2$; our convention

435: in the table

436: is that these indices are ordered cyclically by $T^2$ in the fashion

437: indicated by the Greek and Latin indices; thus for example $Q^{\alpha

438:   \beta}_k = Q^{13}_6 = Q^{35}_2 = Q^{51}_4$. The IIA R-R forms

439: that survive the orbifold projection must have pairs of indices

440: extending on both dimensions of each

441: $T^2$ on which there are any indices.  We denote here by $\alpha i,

442: \beta j, \gamma k$ pairs of indices on the same torus; thus, for

443: example $\bar{F}_{\alpha i} = \bar{F}_{12}=\bar{F}_{34}=\bar{F}_{56}$.

444: Additionally, all fluxes in the table are antisymmetric in their upper

445: indices as well as in their lower indices.  

446: Note that while $f$ and $R$ fluxes do not appear on the IIB side of

447: the table, this is a consequence of the fact that all dimensions on

448: the $T^6$ are odd under the orientifold reflection, and $f$ and $R$

449: require an even number of odd indices.  In more general orientifolds,

450: IIB compactifications would also admit $f$ and $R$ fluxes.

451:   

452: The resulting full superpotential in the symmetric torus model

453: is

454: \begin{eqnarray}

455: W & = & a_0 - 3a_1 \tau + 3a_2 \tau^2 - a_3 \tau^3\\ & &

456:  \hspace{0.2in} + S (-b_0 + 3b_1 \tau - 3b_2 \tau^2 + b_3 \tau^3)

457: \nonumber\\ & &

458:  \hspace{0.2in} + 3 U (c_0 + (\hat{c}_1 +\check{ c}_1 - \tilde{c}_1)

459:  \tau - (\hat{c}_2 +\check{ c}_2 - \tilde{c}_2) \tau^2 -c_3

460:  \tau^3). \nonumber

461: \end{eqnarray}

462: 

463: At this point let us comment briefly on the nature of the integrally

464: quantized fluxes appearing in the table.  On the standard geometric

465: torus, by $\bar{H}_{ijk}, \bar{F}_{\alpha i}, \ldots$ we simply mean

466: the number of units of $H$ or $F$ integrated over the appropriate

467: cycle on the torus.  In the presence of geometric flux such as

468: $f^a_{bc}$ this is slightly more subtle, but can still be made

469: explicit.  There is a natural basis of Einbeins $\eta^a$ 

470: satisfying $d\eta^a = f^a_{bc}\eta^b \wedge \eta^c$

471: \cite{kstt}, which we may use to obtain integrally quantized fluxes such as

472: $F^{(2)}= \bar{F}_{ab} \eta^a \wedge \eta^b$ for any pair $a, b$, even

473: when the corresponding R-R flux is not in the cohomology of the

474: manifold.  For example, turning on units of flux $ \bar{H}_{123}= 1,

475: \bar{F}_4 = M$ on a standard 4-torus gives a configuration which is

476: taken by T-duality to $f^1_{23} = 1, \bar{F}_{14} = M$ on the dual

477: torus with geometric flux.  While the resulting R-R flux $F^{(2)}$ is

478: not in the cohomology (since the 1-cycle is trivial in homotopy),

479: there is still a nontrivial integral quantization, as we see from this

480: explicit T-duality.  There are constraints due to tadpole cancellation

481: and integrated Bianchi identities which we discuss in Section 4; these

482: place linear constraints on the R-R fluxes in a fixed geometric

483: background with nonzero $f^a_{bc}$.  This is presumably how the

484: K-theory description of R-R charges \cite{K-theory} continues to be

485: valid in the case with geometrical fluxes (possibly related

486: considerations appeared in \cite{3}).  We do not have a complete

487: understanding of how this works in detail, however, particularly when

488: there is torsion in the cohomology.  In the cases with nongeometric

489: fluxes $Q^{ab}_c$ and $R^{abc}$, we do not have a specific and

490: concrete interpretation of the meaning of the quantized fluxes, but

491: from the approach we take here it seems natural to associate integral

492: fluxes $\bar{H}_{abc}, \bar{F}_{a_1 \cdots a_p}$ with every cycle on

493: the original torus; these fluxes appear in the superpotential and are

494: constrained by the identities we compute in Section 4.  These fluxes

495: are perhaps best interpreted as some ``dressing'' added to the basic

496: topological structure of the geometrical torus, in a way which might

497: naturally generalize to other Calabi-Yau manifolds.

498: 

499: Let us now discuss the detailed derivation of the arrangement of

500: fluxes in the table above.  As discussed above, the IIB NS-NS and R-R

501: fluxes appearing in $P_1$ and $P_2$ are already known

502: \cite{gvw,drs,kst}, as are the IIA NS-NS, R-R, and geometric fluxes

503: appearing in $P_1$ and the linear parts of $P_2$ and $P_3$

504: \cite{dkpz, vz, cfi}.  We need to complete the story using T-duality

505: and coordinate symmetry.  Our conventions for the action of T-duality

506: on topological R-R and NS-NS fluxes, including the nongeometric

507: fluxes, is that T-duality removes or adds an index to the first

508: position of an R-R flux, $T_x:\bar{F}_{i_1 \cdots i_n} \leftrightarrow

509: \bar{F}_{x i_1 \cdots i_n}$, and acts on NS-NS fluxes by either

510: raising the first lower index or lowering the last upper index, so for

511: example $T_b:f^{a}_{bc} \leftrightarrow Q^{ab}_c$.  Since we are only

512: interested in the action of T-duality on the topological part of the

513: fluxes, additional moduli-dependent terms which appear in the local

514: T-duality rules \cite{Buscher,bho,ght,Hassan,kst} are not relevant to this

515: discussion.

516: 

517: We begin by noting that the first eight lines in the table all contain

518: known fluxes in the IIA picture.  Each of these fluxes can be

519: T-dualized directly to IIB.  The R-R fluxes transform to the known IIB

520: R-R flux coefficients.  The $S$ and $S \tau$ terms are associated with

521: NS-NS $H$-fluxes and geometric fluxes which transform to the known

522: $H$-fluxes in the IIB theory.  The $U$ and $U \tau$ terms can

523: also be T-dualized, and lead to new coefficients in the IIB theory

524: associated with nongeometric fluxes $Q^{ab}_c$ for various values of

525: $a, b, c$.  Thus, we can use T-duality to complete the picture for the

526: first 8 lines in the table.

527: 

528: To proceed further we note that in the IIB model, there is no

529: geometric distinction between $\alpha$ and $i$ indices, as the

530: O3-plane does not extend in any of the directions on the $T^6$.  Thus,

531: by switching the roles of the $\alpha$ and $i$ indices in defining the

532: complex structure, we exchange $\alpha \leftrightarrow i$, etc..  This

533: exchange takes $1 \leftrightarrow \tau^3, \tau \leftrightarrow \tau^2$

534: in the superpotential, and allows us to identify the remaining

535: $Q^{ab}_c$ coefficients associated with $U \tau^2, U \tau^3$

536: in the IIB superpotential.

537: 

538: This derivation relies only on duality transformations which can be

539: performed explicitly (at least for each individual flux), and

540: therefore is a rigorous demonstration that the IIB theory has the full

541: superpotential (\ref{eq:w-complete}) with all coefficients in the

542: cubic polynomials $P_{1, 2, 3} (\tau)$ associated with well-defined

543: geometric and nongeometric fluxes.  A more detailed discussion of the

544: nongeometric fluxes  $Q^{ab}_c$ appears in the following section.

545: 

546: To complete the story on the IIA side, we would like to carry our

547: results from IIB back to IIA using the 3-fold T-duality on the

548: complete IIB superpotential.  The $S \tau^2$ and $U \tau^2$ terms are

549: associated in the IIB theory with fluxes which transform to

550: nongeometric fluxes $Q^{ab}_c$ on the IIA side.  This extends the

551: superpotential constructed in \cite{vz, cfi} to include nongeometric

552: fluxes of the $Q$-form.  Note, however, that the terms $S \tau^3$ and

553: $U \tau^3$ are associated with fluxes in the IIB theory which in the

554: IIA theory must take the form of a ``T-dual'' on direction $c$ of a

555: nongeometric flux $Q^{ab}_c$.  We do not know how to carry out such a

556: transformation explicitly.  It seems, however, that for duality

557: invariance to be complete, we must introduce a new type of

558: nongeometric flux in the IIA theory, labeled $R^{abc}$.  We discuss

559: the possible interpretation of these new topological nongeometric

560: fluxes in the next section.  Including these fluxes leads to a

561: superpotential (\ref{eq:w-complete}) which is manifestly T-duality

562: invariant.  Note that with this complete set of fluxes, not only can

563: we go from the IIB theory with an O3-plane to the IIA theory with an

564: O6-plane, but we can also perform a complete 6-fold duality on the IIA

565: theory.  This duality flips the complex structure on each $T^2$, again

566: taking $1 \leftrightarrow \tau^3, \tau \leftrightarrow \tau^2$.

567: Again, for this to be an invariance of $W$ we must include the fluxes

568: $R^{\alpha \beta \gamma}, R^{ij \gamma}$, which in this case are the

569: 6-fold T-duals of $\bar{H}_{ijk}, \bar{H}_{\alpha \beta k}$.

570: 

571: This completes our construction of the duality-invariant

572: superpotential for orientifold compactifications of

573: the generalization of the twisted torus in type IIA

574: and IIB string theory.  As we have seen, nongeometric fluxes appear as

575: coefficients of various terms in this superpotential.  In the

576: following sections we will discuss the interpretation of these

577: nongeometric fluxes and topological constraints on possible values of

578: these fluxes.

579: 

580: Given the superpotential we have computed here, it is straightforward

581: in principle

582: to choose integral fluxes $a_0, \ldots$ (subject to constraints which

583: we will discuss in Section 4) and to solve the 

584: equations of motion.

585: Given the superpotential (\ref{eq:w-complete}) and the K\"ahler

586: potential (\ref{eq:Kaehler}), the equations for a supersymmetric

587: vacuum in the four-dimensional theory are

588: \begin{equation}

589: D_\tau W = D_SW = D_UW = 0,

590: \end{equation}

591: where

592: \begin{equation}

593: D_AW = \partial_AW +  (\partial_AK)W \,.

594: \end{equation}

595: For generic flux coefficients in the superpotential (\ref{eq:w-complete}),

596: the equations for $S$ and $U$ are equivalent to

597: \begin{eqnarray}

598: P_1 (\tau) + \bar{S}P_2 (\tau) +UP_3 (\tau)& = &  0 \label{eq:es}\\

599: P_1 (\tau) + S P_2 (\tau) + \left({2\over 3} U +{1\over 3}\bar{U}\right) P_3 (\tau) & = &  0  \label{eq:eu}\,.

600: \end{eqnarray}

601: The remaining ($\tau$) equation is

602: \begin{equation}

603: (\tau -\bar{\tau})  \partial_\tau W -3 W = 0  \label{eq:et}\,.

604: \end{equation}

605: We defer a detailed analysis of solutions of these equations to a

606: forthcoming paper \cite{stw-2}, but we will make a few brief comments

607: here regarding the space of solutions.  Clearly, the space of SUSY

608: solutions to these equations will include all type IIB and IIA flux

609: vacua on the geometric symmetric torus, as well as possibly a large

610: number of vacua with geometric and nongeometric fluxes, which may

611: generically have no geometric duals.  In \cite{dgkt1}, a family of

612: supersymmetric IIB vacua on the symmetric torus with $W = 0$ was

613: identified, corresponding to flux compactifications with vanishing

614: cosmological constant.  These vacua all have nonvanishing $b_2$ or

615: $b_3$ and therefore are nongeometric in the IIA picture.  Indeed, it

616: is easy to see from (\ref{eq:es}-\ref{eq:et}) that there are no

617: geometric $W = 0$ solutions in the IIA theory, since this would

618: require either Im $S = 0$ (which is unphysical) or $P_2 =0$.  The

619: vanishing of $P_2$ in turn implies that either $b_0 = b_1 = 0$ (which

620: violates the tadpole condition (\ref{eq:tadpole1}), which we derive in

621: general in Section 4), or $\tau = b_0/3 b_1$ is real (which is again

622: unphysical).  Thus, admitting nongeometric fluxes expands the set of

623: Minkowski IIA flux vacua on the symmetric torus from the empty set to

624: a nonzero set of vacua.  In \cite{cfi} it was argued that when

625: geometrical fluxes are allowed, there are an infinite family of AdS

626: vacua in the IIA torus model.  Assuming this result is correct, this

627: shows that including nongeometric fluxes on the IIB side extends the

628: finite set of geometric SUSY Minkowski vacua by an infinite set of AdS

629: vacua.  Beyond these already known results, it seems that generic

630: nongeometric flux configurations may have no geometric duals, but may

631: nevertheless lead to acceptable SUSY flux compactifications.  A more

632: detailed analysis of this issue will appear in \cite{stw-2}.

633: 

634: 

635: \section{Interpretation of nongeometric fluxes}

636: \label{sec:interpretation}

637: 

638: In this section we describe in greater detail the structure of

639: ``T-folds'' with nongeometric $Q$-fluxes, and speculate about the

640: nature of $R$-fluxes.  So far, our treatment of these fluxes has been

641: fairly formal.  We have essentially defined a set of T-duality

642: transformation rules for generalized NS-NS fluxes on the torus through

643: (\ref{eq:t-chain}), analogous to the T-duality rules for R-R fluxes.

644: In this section we discuss the nongeometric fluxes $Q^{ab}_c$ and

645: $R^{abc}$ in greater detail, and discuss when compactifications with such

646: extra structure can be described at least locally geometrically.  We

647: also comment on the world-sheet description of compactifications with

648: nongeometric fluxes.

649: 

650: 

651: In order to develop some intuition for the nongeometric fluxes $Q

652: ^{ab}_c$, let us discuss a simple example where only the flux $Q

653: ^{ab}_c$ is present.  We will construct this configuration by step by

654: step using T-duality, as in \cite{kstt, Schulz}, beginning from a

655: square three-torus with metric \beq ds ^ 2 = dx ^ 2+dy ^ 2+dz ^ 2 \eeq

656: and $N$ units of $H$-flux, $\bar{H}_{xyz}=N$. We are free to choose a

657: gauge where \beq

658: \label{eq:b-example}

659: B_{xy} = N z.  \eeq We can think of this configuration as a $T ^ 2$

660: parameterized by $x$ and $y$, fibered over a circle with coordinate

661: $z$. The NS-NS degrees of freedom coming from reduction on $T ^ 2_{xy} $ are

662: the complex structure modulus of the torus, $\tau$, and the K\"ahler

663: modulus $\rho = B_{xy} + i\vol_{xy}$. The perturbative duality group

664: of the theory reduced on $T ^ 2_{xy} $ is, up to discrete factors, $SL

665: (2, \bZ)_\tau \times SL (2, \bZ)_\rho$.  The presence of $N$ units of

666: $H$-flux is described in this language as a nontrivial monodromy

667: $\rho\rightarrow\rho + N$ as $z\rightarrow z+1$.

668: 

669: Since it will be useful in our discussion of the nongeometric fluxes,

670: let us take a moment to develop this $SL(2,\bZ)$ description of

671: compactification with $H$-flux a little further. This will help in understanding a general 

672: set of quadratic constraints on the fluxes which we derive in the next section.  In an ordinary

673: dimensional reduction, we take the fields to be independent of the

674: coordinates of the compact directions.  In a reduction with $H$-flux,

675: there is nontrivial coordinate dependence in the gauge potentials.

676: Dimensional reduction in the presence of flux can thus be understood

677: as a class of Scherk-Schwarz generalized dimensional reduction

678: \cite{ss, km}.  From this point of view, the $z$-dependence

679: of the fields is understood by specifying an $z$-dependent element

680: $M(z) \in SL(2,\bR)_\rho$ which has the desired monodromy

681: $\left ( \begin{matrix} 1 & N \cr 0 & 1

682: \end{matrix}

683:  \right)$

684: when $z\rightarrow z + 1$ \cite{llp,hull-98}. 

685: The theory reduced on the three-torus must be independent of $z$, so in the present example $M(z)$ 

686: can be at most linear in $z$. Our choice of gauge \er{b-example} for $B$ gives us 

687: \beq

688: \label{eq:mz}

689: M(z)= \left(

690: \begin{matrix}

691: 1 & Nz \cr

692: 0 & 1

693: \end{matrix}

694: \right )_{\rho};

695: \eeq

696: other choices of $M(z)$ with the same monodromy are possible and lead to reduced theories

697: which are equivalent under field redefinitions \cite{hull-98}. The reduction ansatz for an arbitary field $\phi$ is

698: then $\phi(z)= \left[M(z)\right]_\phi \phi_0$, where

699: $\left[M(z)\right]_\phi $ is the appropriate representation of

700: $M(z)$, and $\phi_0$ is a vector in this representation containing the

701: degrees of freedom analogous to zero-modes in this background.

702: 

703: Let us consider the case where the degrees of freedom are R-R field strengths; this will be useful in understanding how turning on topological NS-NS fluxes affects the topological R-R fluxes.  For illustration, consider the topological fluxes $\bar{F}_{wxy}$ and $\bar{F}_w$ in type IIB.  Here $w$ denotes some compact direction transverse to the three-torus.  These degrees of freedom transform under $SL(2,\bZ)_\rho$ as

704: \beq

705: \left(\begin{array}{c} \bar{F}_{w xy}\\\bar{F}_w\end{array}\right)\rightarrow\left(\begin{array}{cc} a & b\\c & d\end{array}\right)\left(\begin{array}{c} \bar{F}_{w xy}\\\bar{F}_w\end{array}\right).

706: \eeq

707: Therefore, using the matrix \er{mz} to describe the presence of $H$-flux gives us

708: \beq

709: \left(\begin{array}{c} {F}_{w xy}(z)\\ {F}_w(z)\end{array}\right)=\left(\begin{array}{cc} 1 & Nz\\ 0 & 1\end{array}\right)\left(\begin{array}{c} \bar{F}_{w xy}\\\bar{F}_w\end{array}\right),

710: \eeq

711: so

712: \beq

713: F ^ {(3)} = (\bar{F}_{wxy} +Nz\, \bar{ F}_w) dw\, dx\, dy,

714: \eeq

715: from which we can recover the familiar Bianchi identity

716: \beq 

717: dF ^ {(3)} =-N\bar{ F}_w \, dw\, dx\, dy\, dz =-\bar{ F} ^{(1)}\wedge\bar{H}.

718: \eeq

719: As usual, integrating this equation on the $(w, x, y, z)$-cycle leads to the constraint

720: \beq

721: \bar{F}_{[w} \bar{H}_{xyz]} = 0.

722: \eeq

723: Note that when other fluxes are present, including geometric and nongeometric fluxes, 

724: this constraint will acquire other terms.

725: The point of view we used to obtain this constraint will prove very useful in obtaining the analogous constraint terms in the presence of geometric flux and nongeometric $Q$-flux, as we will demonstrate below.

726: 

727: 

728: Following \cite{kstt}, we take this square three-torus with $H$-flux and perform a T-duality on the $x$ direction. This yields a twisted $T^3$ with

729: $f^x_{yz} = N$, which in this gauge has the metric 

730: \beq

731: \label{eq:twisted-torus}

732: ds^2 =  (dx-Nz dy)^2 + dy^2 + dz^2

733: \eeq

734: and $B=0$. The nontrivial monodromy  is now $\tau \rightarrow \tau -N$, realized

735: by the action of the matrix

736:  \beq

737: M(z)= \left(

738: \begin{matrix}

739: 1 & -Nz \cr

740: 0 & 1

741: \end{matrix}

742: \right )_{\tau}

743: \eeq

744: on the fields in the theory. 

745: 

746: Consider now the behavior of Ramond-Ramond fluxes in this background. For concreteness,

747: consider the fluxes $\bar{F}_{wx}$ and $\bar{F}_{wy}$ in type IIA. These transform under $SL(2,\bZ)_{\tau}$ as

748: \beq

749: \left(\begin{array}{c}

750:   \bar{F}_{w y}\\\bar{F}_{wx}\end{array}\right)\rightarrow\left(\begin{array}{cc} a &

751:     b\\c & d\end{array}\right)\left(\begin{array}{c}

752:      \bar{ F}_{wy}\\\bar{F}_{wx}\end{array}\right). 

753: \eeq

754: Therefore, in this background, we find $F ^ {(2)}$ is appropriately expanded as 

755: \beq

756: F ^ {(2)} = (\bar{F}_{wy} - Nz \bar{F}_{wx}) dw\, dy+ \bar{F}_{wx} dw\, dx.

757: \eeq

758: This reproduces, from another point of view, the expansion of $F ^{(2)}$ on a twisted torus in 

759: a basis of globally defined 1-forms $\{\eta^x= d x-Nzdy,  \eta^y=dy\}$, where we define $\bar{F}_{ab} $ through $F ^ {(2)} = \bar{F}_{ab}\eta ^ a\wedge\eta ^ b$

760: \cite{kstt}. 

761: The Bianchi identity for $F ^ {(2)}$ is as usual

762: \beq

763: d F ^ {(2)} =-N dz\, \bar{F}_{wx} dw\, dy =-\bar{ F}_{wx} f ^ x_{yz} dw\, dy\, dz\,

764: \eeq

765: which we may freely integrate over the non-twisted $(w, y, z) $-cycle  to obtain the constraint term 

766: (in the absence of other fluxes)

767: \beq

768:  \bar{F}_{x[w}f^x_{ yz]} = 0.

769: \eeq

770: 

771: 

772: As the metric \er{twisted-torus} does not depend on $y$, we may perform another T-duality in the $y$ direction to  arrive at a $T^3$ with nongeometric flux $Q^{xy}_z = N$. The metric

773: on this background is 

774: \beq

775: ds ^ 2 = {1\over 1+N ^ 2 z ^ 2}\left( dx ^ 2+dy ^ 2\right) + dz^2

776: \eeq

777: and the $B$-field becomes

778: \beq

779: B_{xy} ={Nz\over 1+N ^ 2 z ^ 2}.

780: \eeq

781: In this background, the nontrivial monodromy is

782: ${1\over\rho}\rightarrow{1\over\rho} + N$ as $z\rightarrow z + 1$,

783: which mixes the metric and the $B$-field of the two-torus.  

784: Given our particular choice of gauge, the presence of the non-zero $Q$-flux is described by the $SL(2,\bR)$ matrix

785: \beq

786: \left(\begin{array}{cc} 1 & 0\\Nz & 1\end{array}\right)_\rho.

787: \eeq

788: 

789: Now consider the behavior of Ramond-Ramond fields in this background. Returning to type IIB with the field strengths $\bar{F}_{wxy}$ and $\bar{F}_w$ turned on, we can see that in the presence of the $Q $-flux, we must write $F ^ {(1)} $ as

790: \beq

791: F ^{(1)} = (\bar{ F}_w +Nz\, \bar{ F}_{wxy}) dw,

792: \eeq

793: and therefore we obtain the modified Bianchi identity

794: \beq 

795: dF ^ {(1)} = -N \bar{ F}_{w xy} dw\, dz = -Q ^{xy}_z \bar {F}_{xyw} dw\, dz .

796: \eeq

797: Further, as there are no nongeometric twists on the

798: directions $ z$ and $w$, we may regard $ dF ^ {(1)}$ as a

799: two-form without ambiguity.  Integrating this two-form over the $(z,

800: w)$ 2-cycle, we find the constraint term (again in the absence of other fluxes)

801: \beq

802: \bar{F}_{xy  [w}Q^{xy}_{z]} = 0.

803: \eeq

804: This is an example of a new kind of constraint, which will be important to take into account in constructing  nongeometric flux compactifications; we will find the general constraints incorporating this type of term in the next section via T-duality.

805: 

806: There is a natural question that arises at this point: 

807: Given a nonzero $Q^{xy}_z$,

808: can one

809: meaningfully perform another T-duality in the $z$-direction? By

810: analogy with the above, we would expect that such an

811: operation raises another

812: index, producing the object we have denoted $R^{xyz}$. In Section 2, we have

813: argued for the existence of this object in the IIA theory

814: by asserting that the

815: four dimensional superpotential in

816: IIA must agree through T-duality with

817: that of the IIB theory, so it would be nice to be

818: able to see this directly arise from T-dualizing $Q_z^{xy}$. 

819: 

820: The simplest situation in which one would expect the $R^{abc}$ terms

821: to appear is in a $T^3$ compactification with a single type of flux,

822: where we would hope to 

823: have a sequence of T-duality transformations on NS-NS fluxes given by

824: \beq

825: \bar{H}_{xyz} \stackrel{T_x}{\longleftrightarrow} f^x_{yz}

826: \stackrel{T_y}{\longleftrightarrow} Q^{xy}_z

827: \stackrel{T_z}{\longleftrightarrow} R^{xyz} 

828: \eeq

829: As we have discussed above,

830: the first two steps in this procedure can be implemented while

831: thinking of the $T^3$ as a $T^2$ bundle over an $S^1$, but this

832: interpretation breaks down in the final step because of the necessary

833: $z$-dependence of the metric and $B$-field describing $Q^{xy}_z$. 

834: We leave the ten-dimensional interpretation of this flux for future

835: work.  

836: For practical purposes such as computing flux vacua and the properties

837: of the low-energy effective action, we can certainly make progress

838: without an explicit description of these nongeometric $R$-fluxes; all

839: we need for such computations is the

840: four-dimensional superpotential which uses these objects. 

841: Eventually, however, for a full understanding of vacua with 

842: $R$-fluxes, we need some way of explicitly describing such

843: compactifications in string theory.

844: 

845: It is possible that the $R^{xyz}$ have no interpretation using

846: conventional notions of local spacetime, and in this sense are truly

847: nongeometric.  Studying these objects may help us to better

848: understand how both geometric and nongeometric structures may emerge

849: from a fundamental formulation of string theory.  Indeed, we can argue

850: that a truly background-independent formulation of string theory (such

851: as string field theory) must include backgrounds with nongeometric

852: $R$-fluxes as follows: Imagine we have a complete,

853: background-independent formulation of both type IIA and IIB string

854: theory.  The formulation of IIA and IIB in a standard toroidal

855: compactification background with no fluxes must be equivalent.  In the

856: IIB background there are SUSY flux vacua with the flux $H_{\alpha

857: \beta \gamma}$ turned on; many explicit examples of such vacua were

858: described in \cite{kst, dgkt1}.  Such a background should be connected

859: to the background without fluxes in a nonperturbative fashion in the

860: complete formulation of the IIB theory.  But the

861: background-independent descriptions of the IIA and IIB theories must

862: be equivalent.  Thus, there must be a nonperturbative procedure in the

863: IIA model for going from the vacuum without fluxes to the vacuum with

864: $R$-flux.  This indicates that $R$-fluxes must have meaning in a

865: background-independent formulation of the theory.  Of course, one

866: might prefer the IIB description without $R$-flux when it is

867: available, but generically there may be vacua where both the IIA and

868: IIB descriptions have $R$-flux, in which case a more explicit

869: description of the nature of these nongeometric fluxes would be highly

870: desirable.

871: 

872: It is also worth thinking about the worldsheet interpretation of these

873: spaces. Many nongeometric compactifications that have been studied in

874: the past \cite{dh, Hellerman:2002ax, fww} have been claimed to be

875: related to asymmetric orbifolds \cite{Narain:1986qm}. One particular

876: type of compactification features nontrivial twists by an element of

877: the U-duality group when going around a circular direction \cite{dh,

878: fww}. On general grounds, it was shown in \cite{dh} that in such

879: examples some moduli are fixed by requiring that the vacuum lie at the

880: minimum of a Scherk-Schwarz potential; this is essentially the same as

881: saying that the moduli will always lie at fixed points of the

882: U-duality twist.

883: 

884: 

885: It is not, however, necessarily true that the kinds of nongeometric

886: vacua we present in this paper have asymmetric orbifold

887: descriptions. We can see from the $T^3$ example above that the

888: monodromy $1/\rho \rightarrow 1/\rho + N$ has no fixed points, and as

889: such does not fall into the class of vacua considered in

890: \cite{dh,fww}. As such, we have no particular reason to believe that

891: the compactifications we consider here generically have asymmetric

892: orbifold descriptions.  It may be possible to construct some more

893: general asymmetric CFT in which U-duality twists are incorporated at

894: the level of the world-sheet action. Although the $f$-, $Q$-, and $R$-fluxes are all in the NS-NS sector, and thus may have descriptions with nontrivial boundary conditions implemented on a conventional string worldsheet, this would presumably require

895: developing some new technology to understand fully.

896: 

897: 

898: \section{Constraints}

899: \label{sec:constraints}

900: 

901: In this section we discuss constraints on the fluxes.  These

902: constraints arise in two closely related ways.  

903: First, the new fluxes contribute to the tadpole

904: constraints associated with R-R fields.  Second, 

905: the new fluxes contribute to the Bianchi identities for the

906: R-R and NS-NS fields.  In this section we first derive the general R-R

907: and NS-NS constraints on fluxes using T-duality, and then specialize

908: to the particular toroidal compactification of interest in this paper.

909: 

910: 

911: \subsection{R-R tadpole and Bianchi identity constraints}

912: 

913: In this subsection we discuss the constraints on the fluxes coming

914: from the Ramond-Ramond tadpole conditions and Bianchi identities.  

915: The simplest constraint arises  in the IIB theory, where

916: there is a tadpole for the R-R four-form field $A_4$ arising from the

917: Chern-Simons term $\int A_4 \wedge H_3 \wedge F_3$.  

918: $A_4$ is also sourced by local D3-brane and O3-plane contributions.

919: Integrating this tadpole constraint over a six-dimensional compactification

920: manifold gives the topological constraint

921: \beq

922:  \bar{F}_{[abc} \bar{H}_{def]} + {\rm local} = 0, \label{eq:fh}

923: \eeq

924: 

925: By repeated application of T-duality, including both geometric fluxes

926: $f^a_{bc}$ and nongeometric fluxes $Q^{ab}_c$ and $R^{abc}$, in the

927: absence of local sources we have the R-R constraints

928: 

929: \begin{eqnarray}

930: \bar{F}_{[abc} \bar{H}_{def]} & = &  0\label{eq:R-R1}\\

931: \bar{F}_{x [abc}f^x_{ de]}- \bar{F}_{[ab} \bar{H}_{cde]} & = &  0\label{eq:R-R2}\\

932: \bar{F}_{xy [a b c}Q^{xy}_{d]}

933:  -3 \bar{F}_{x[ab}f^x_{cd]}- 2\bar{F}_{[a} \bar{H}_{ bcd]} & = &  0\label{eq:R-R3}\\

934: \bar{F}_{xyz[a b c]} R^{xyz}  -9\bar{F}_{xy [a b}Q^{xy}_{c]}

935:  -18 \bar{F}_{x[a}f^x_{ bc]}+6F^{(0)}\bar{H}_{[abc]} & = &  0 \label{eq:R-R4}\\

936: \bar{F}_{xyz[a b]} R^{xyz}  + 6\bar{F}_{xy  [a}Q^{xy}_{b]}

937: -6\bar{F}_xf^x_{[ab]} & = &

938: 0

939: \label{eq:R-R5}\\

940: \bar{F}_{xyza} R^{xyz}  -3\bar{F}_{xy}Q^{xy}_a & = &  0\label{eq:R-R6}\\

941: \bar{F}_{xyz} R^{xyz} & = &  0 . \label{eq:R-R7}

942: \end{eqnarray}

943: 

944: If we restrict to the geometric context, these constraints are just

945: versions of the standard Bianchi identity $(d + H)F = 0$. The individual $FH$, $Ff$, and $FQ$ terms

946: appearing in these constraints were demonstrated explicitly in the simple example discussed in the previous section.

947: As mentioned previously, in the presence of fixed geometric fluxes, the constraints 

948: \er{R-R1}-\er{R-R5}

949: give linear conditions on integrally quantized fluxes $\bar{F}$ which

950: may not be in the cohomology.

951: The extra terms with $Q$'s and $R$'s are additional contributions from

952: nongeometric fluxes.  While we have derived these constraints from

953: T-duality on the torus, we expect that there may be a much more

954: general class of compactifications in which these constraints apply.

955: 

956: In our toroidal compactification, we have a set of O3-planes in the IIB model

957: which sets the RHS of (\ref{eq:R-R1}) to  16 $ \times\left(\begin{array}{c}

958: 6 \\3

959: \end{array} \right)^{-1}$, where the last factor comes from the

960: combinatorial factors associated with $F \wedge H$.  In the IIA model,

961: the corresponding O6-planes set the RHS of (\ref{eq:R-R4}) to 16

962: $\times$ 6

963: when the free indices $a, b, c$ run over the directions $i, j, k$.  In

964: terms of the integer coefficients $a_0, \ldots,$ the resulting tadpole

965: constraint is the same in both models and is \beq

966: \label{eq:tadpole1}

967: a_0 b_3-3 a_1 b_2+3 a_2b_1-a_3b_0 = 16.

968: \eeq

969: 

970: There is only one further R-R constraint relevant for our model, which

971: comes from (\ref{eq:R-R5}) for the IIB model and again from

972: (\ref{eq:R-R4}) in the IIA model.  This constraint becomes

973: \beq

974: \label{eq:tadpole2}

975: a_0 c_3+a_1 (\check{c}_2 + \hat{c}_2-\tilde{c}_2)

976: -a_2 (\check{c}_1 + \hat{c}_1 -\tilde{c}_1)-a_3c_0 = 0.

977: \eeq

978: All remaining tadpole constraints from (\ref{eq:R-R1}--\ref{eq:R-R7}) are satisfied automatically in the

979: particular background we are considering here.

980: 

981: 

982: \subsection{NS-NS Bianchi identity constraints}

983: 

984: We can carry out a similar analysis of the NS-NS Bianchi identities

985: from T-duality.

986: In a geometric compactification, the NS-NS fluxes must satisfy

987: \beq

988: %\label{eq:fh}

989: f ^ x_{[ab} \bar{H}_{cd]x} = 0,

990: \eeq

991: which comes from the Bianchi identity $dH = 0$.  

992: Using T-duality, we find the set of NS-NS constraints

993: \begin{eqnarray}

994: \label{eq:ns-1}

995: \bar{H}_{x[ab}f^x_{cd]} & = &  0\\

996: \label{eq:ns-2}

997: f^a_{x[b}f^x_{cd]} + \bar{H}_{x[bc}Q^{ax}_{d]} & = &  0\\

998: \label{eq:ns-3}

999: Q^{[ab]}_xf^x_{[cd]}-4f^{[a}_{x[c}Q^{b]x}_{d]} + \bar{H}_{x [cd]} R^{[ab]x} & = &  0\\

1000: \label{eq:ns-4}

1001: Q^{[ab}_xQ^{c]x}_d + f^{[a}_{xd} R^{bc]x} & = &  0\\

1002: \label{eq:ns-5}

1003: Q^{[ab}_x R^{cd]x} & = &  0.

1004: \end{eqnarray}

1005: Finally, in order for the $f$- and $Q$-fluxes to be individually T-dual to $H$-flux, they must satisfy 

1006: \beq

1007: f ^ x_{xa} = 0 = Q ^{ax}_x.

1008: \eeq

1009: 

1010: Equations (\ref{eq:ns-1}--\ref{eq:ns-5}) have a

1011: nice interpretation in the four-dimensional effective theory. Ignoring

1012: the R-R fields for the purposes of this discussion, in a reduction on

1013: $T ^ 6$ without flux, the four-dimensional supergravity theory

1014: contains a gauge sector with gauge group $U (1) ^ {12}$ coming from

1015: the 10-dimensional metric and $B$-field.  As noted in \cite{ss} and

1016: developed in \cite{km}, adding geometric NS-NS fluxes

1017: $\bar{H}_{abc}$ and $f ^ a_{bc}$ to the compactification makes the

1018: gauge algebra of the four-dimensional theory nonabelian; the fluxes

1019: appear as structure constants of the gauge algebra.  Denoting

1020: generators descending from 10-dimensional diffeomorphism invariance as

1021: $Z_m$ and generators descending from the 10-dimensional gauge symmetry

1022: of $B$ as $X ^ m$, the Lie algebra of the compactified theory is

1023: \cite{km}

1024: \begin{eqnarray}

1025: {[Z_a, Z_b]} & = & \bar{H}_{abc} X^ c+f ^ c_{ab} Z_c\\

1026: \label{eq:commute-1}

1027: {[Z_a, X ^ b]} & = &-f ^ b_{ac} X ^ c \\ 

1028: \label{eq:commute-2}

1029: {[X ^ a, X ^ b]} & = & 0.

1030: \end{eqnarray}

1031: The Jacobi identities of this algebra then yield the purely geometric portion of the NS-NS constraints (\ref{eq:ns-1}--\ref{eq:ns-2}).  This algebra may be written in a form which is manifestly covariant under the perturbative duality group $O(6,6,\bZ)$ \cite{km, hre}.

1032: 

1033: By acting on the four-dimensional theory with elements of $O(6, 6,\bZ) $  corresponding to T-duality, we may deduce how to modify this gauge algebra in the presence of nongeometric fluxes $Q, R$.  We find 

1034:  the commutators (\ref{eq:commute-1}--\ref{eq:commute-2}) are modified in the obvious way,

1035: \begin{eqnarray}

1036: {[Z_a, X ^ b] } & = &-f ^ b_{ac} X ^ c +Q ^{bc}_aZ_c\\

1037: {[X ^ a, X ^ b]} & = & Q ^{ab}_c X ^ c +R ^{abc} Z_c.

1038: \end{eqnarray}

1039: The Jacobi identities of this, fully general, algebra now reproduce the full set of NS-NS constraints (\ref{eq:ns-1}--\ref{eq:ns-5}).

1040: 

1041: When applied to our toroidal compactification, the constraints

1042: (\ref{eq:ns-1}--\ref{eq:ns-5}) lead to a number of conditions on the

1043: integer coefficients $b_0, \ldots$.  The first set of  conditions

1044: arises from \er{ns-2} in IIB and yields

1045: \begin{eqnarray}

1046: \label{eq:bc1}

1047: c_0 b_2-\td{c}_1 b_1+\hat{c}_1 b_1-\check{c}_2 b_0 & = & 0\\

1048: \check{c}_1 b_3-\hat{c}_2 b_2+\td{c}_2 b_2-c_3 b_1 & = & 0

1049: \label{eq:bc2}\\

1050: c_0 b_3-\td{c}_1 b_2+ \hat{c}_1 b_2-\check{c}_2 b_1 & = & 0

1051: \label{eq:bc3}\\

1052: \label{eq:bc4}

1053: \check{c}_1 b_2-\hat{c}_2 b_1+\td{c}_2 b_1-c_3 b_0 & = & 0;

1054: \end{eqnarray}

1055: as well as parallel constraints in which all hats and checks are

1056: switched through $\hat{c}_i \leftrightarrow\check{ c}_i$.  In IIA

1057: these constraints come from, in order, \er{ns-2}, \er{ns-4}, and (the

1058: last two) \er{ns-3}.  

1059: For instance, we obtain \er{bc1} in IIB from setting $a=\beta, b=\gamma, c=j$, and

1060: $d=k$ in \er{ns-2}; the others follow similarly. 

1061: The second set of conditions arises from

1062: \er{ns-4} in IIB and yields

1063: \begin{eqnarray}

1064: c_0\td{c}_2-\check{c}_1 ^ 2+\td{c}_1\hat{c}_1-\hat{c}_2 c_0 & = & 0

1065: \label{eq:cc1}\\

1066: c_3\td{c}_1-\check{c}_2 ^ 2 +\td{c}_2\hat{c}_2-\hat{c}_1 c_3 & = & 0\\

1067: c_3 c_0-\check{c}_2\hat{c}_1 +\td{c}_2\check{c}_1-\hat{c}_1\td{c}_2 & = & 0\\

1068: \hat{c}_2\td{c}_1-\td{c}_1\check{c}_2 +\check{c}_1\hat{c}_2-c_0c_3 & =

1069: & 0

1070: \label{eq:cc4},

1071: \end{eqnarray}

1072: as well as the parallel constraints with hats and checks switched.  In

1073: IIA, these constraints again come from \er{ns-2}, \er{ns-4}, and (the

1074: last two) \er{ns-3}.

1075: 

1076: We can simplify these conditions significantly by subtracting each

1077: equation from its parallel counterpart with hats and checks switched.

1078: {}From (\ref{eq:cc1}-\ref{eq:cc4}), we find the conditions

1079: \begin{eqnarray}

1080: c_1 \Delta_1 & = &  c_0 \Delta_2\\

1081: c_3 \Delta_1 & = &  c_2 \Delta_2\\

1082: \tilde{c}_2 \Delta_1 & = & - \tilde{c}_1 \Delta_2\\

1083: (2\tilde{c}_2 + \hat{c}_2 + \check{c}_2) \Delta_1 & = &  

1084: (2\tilde{c}_1 + \hat{c}_1 + \check{c}_1) \Delta_2

1085: \end{eqnarray}

1086: where

1087: \begin{eqnarray}

1088: c_1 & = & (\tilde{c}_1 + \hat{c}_1 + \check{c}_1)\\

1089: c_2 & = & (\tilde{c}_2 + \hat{c}_2 + \check{c}_2)\\

1090: \Delta_i & = & \hat{c}_i-\check{c}_i, \;\;\;\;\; i \in\{1, 2\} \,.

1091: \end{eqnarray}

1092: Assuming both $\Delta$'s are nonzero allows us to rewrite equations

1093: (\ref{eq:cc1}-\ref{eq:cc4}) in terms of the three components of

1094: $c_1$.  All 4 equations reduce to the same quadratic

1095: \begin{equation}

1096: 3 \tilde{c}_1^2 + 3 \tilde{c}_1 (\hat{c}_1 +\check{ c}_1) +

1097: \hat{c}_1^2 +\check{ c}_1^2 + \hat{c}_1\check{ c}_1 = 0.

1098: \end{equation}

1099: This equation has no real solution for $\tilde{c}_1$ unless $\hat{c}_1

1100: =\check{ c}_1$, so $ \hat{c}_1$ and $\check{c}_1$ can be identified.

1101: A similar argument demonstrates that, even after setting

1102: $\hat{c}_1 =\check{ c}_1$, we must have $\hat{c}_2 =\check{ c}_2$.

1103: Thus, the full set of constraints is just 

1104: (\ref{eq:bc1}-\ref{eq:bc4}) and (\ref{eq:cc1}-\ref{eq:cc4}), with

1105: $\hat{c}_i =\check{ c}_i$.

1106: 

1107: The equality $\hat{c}_i =\check{ c}_i$ implies a convenient anti-symmetry property of the

1108: $Q^{ab}_c$ and $f ^ a_{ bc}$ in our model.  Given the equality $Q^{i \beta}_k =

1109: Q^{\alpha j}_k$, we may through cyclic permutation of the tori obtain $Q^{\alpha j}_k =

1110: Q^{\beta k}_i = -Q^{k \beta}_i$.  One may show similarly that

1111: $Q^{ab}_c$ is antisymmetric under exchange of any upper and lower index,

1112: provided that both indices are of the same kind (Greek or Latin); the

1113: same is true for $f^a_{bc}$. Note, however, that neither $Q^{ab}_c$

1114: nor $f^a_{bc}$ is fully antisymmetric in all three indices, since we

1115: are not free to exchange Latin and Greek indices.

1116: 

1117: Given the simplification

1118: $\hat{c}_i =\check{ c}_i$, the constraints

1119: (\ref{eq:bc1}-\ref{eq:bc4}) and (\ref{eq:cc1}-\ref{eq:cc4}) can be

1120: simplified further.  In particular,

1121: (\ref{eq:bc1}) and (\ref{eq:bc4}) become equivalent, and

1122: (\ref{eq:bc2}) and (\ref{eq:bc3}) become equivalent when the

1123: constraints on the $c$'s are imposed.  Details of the

1124: parameterization of solutions to these constraints will be presented

1125: along with solutions of the SUSY vacuum equations in \cite{stw-2}.

1126: 

1127: We close this section with a brief discussion of S-duality.  The IIB

1128: theory is invariant under an S-duality symmetry which exchanges the

1129: fluxes $F_{abc}$ and $H_{abc}$ (with a change of sign in one

1130: direction), while taking $S \rightarrow -1/S$.  This has the effect in

1131: the superpotential of exchanging the integral flux parameters $a_i

1132: \leftrightarrow b_i$.  We expect that it should be possible to combine

1133: this S-duality transformation with T-duality to get a larger U-duality

1134: group under which our 4D theory is invariant.  The constraint

1135: (\ref{eq:tadpole1}) is indeed invariant under S-duality.  The

1136: remaining constraints, however, provide a puzzle.  The equation

1137: (\ref{eq:tadpole2}) is precisely the sum of the independent $bc$

1138: constraints (\ref{eq:bc3}) and (\ref{eq:bc4}) when $a$ and $b$ are

1139: exchanged.  Thus, the constraints are not obviously incompatible with

1140: S-duality, but also

1141: do not precisely match.  

1142: This apparent mismatch in constraints presumably arises from the fact

1143: that the $Q$'s actually transform nontrivially under S-duality, since

1144: they generically mix the  $B$-field and the metric.

1145: Indeed, the mismatch can be seen directly by noting that in

1146: the $FQ$ term in \er{R-R5}

1147: and the $HQ$ term in \er{ns-2} the free indices

1148: (and the number of constraints) do not match.  

1149: It is clearly crucial to better

1150: understand the

1151: effects of S-duality on nongeometric fluxes. We leave this as an open

1152: question for future work.

1153: 

1154: 

1155: \section{Conclusions}

1156: \label{sec:conclusions}

1157: 

1158: In this paper we have developed a framework for systematically

1159: describing nongeometric NS-NS fluxes in the context of a simple

1160: toroidal compactification of type II string theory.  Like R-R fluxes,

1161: the geometric and nongeometric NS-NS fluxes act in some sense as

1162: $p$-forms on a canonically chosen space-time, here $T^6$, and

1163: transform under T-duality by adding and removing lower indices through

1164: \begin{equation}

1165: H_{abc} \stackrel{T_a}{\longleftrightarrow} f^a_{bc}

1166: \stackrel{T_b}{\longleftrightarrow} Q^{ab}_c

1167: \stackrel{T_c}{\longleftrightarrow} R^{abc}  \,.

1168: \label{eq:t-chain-2}

1169: \end{equation}

1170: While we do not have a complete mathematical description of these

1171: objects, at least on the torus we can take (\ref{eq:t-chain-2}) as a

1172: definition of how these nongeometric fluxes transform.  The $f^a_{bc}$

1173: fluxes correspond to geometrical fluxes defining a ``twisted torus''

1174: \cite{ss,km, dh}.  The $Q^{ab}_c$ fluxes describe compactifications on

1175: locally geometric spaces with nongeometric global boundary conditions,

1176: such as previously discussed in \cite{dh, fww, x2, gh}.  We can

1177: explicitly carry out 

1178: T-duality from $H \rightarrow f \rightarrow Q$ using standard Buscher

1179: T-duality rules, so our discussion here is on well-established

1180: ground.  We cannot, however, use the Buscher rules to T-dualize

1181: $Q^{ab}_c \rightarrow R^{abc} $, just as the Buscher rules on R-R

1182: potentials $A^{( p)}$ cannot lead to a direct construction of the R-R

1183: 0-form flux $F^{(0)}$.  In this sense, the last T-duality in

1184: (\ref{eq:t-chain-2}) must at this point be taken as a formal definition.

1185: 

1186: The need to include nongeometric fluxes of the $R$-type becomes clear

1187: in our construction of the superpotential describing the moduli of the

1188: toroidal compactification to four dimensions.  We have used concrete T-duality

1189: constructions to understand how the $Q$'s extend the superpotential in

1190: the type IIB case, where there are no $R$'s allowed in our particular

1191: orientifold compactification due to parity constraints.  The

1192: consistency of the IIA and IIB pictures then forces us to the

1193: conclusion that the nongeometrical fluxes $R^{abc}$ must be included

1194: on the IIA side.  An important open question is whether these $R$-type

1195: fluxes admit a locally geometric description like the $Q$-type fluxes.

1196: 

1197: In this paper we have focused on nongeometric fluxes associated with a

1198: toroidal compactification.  It is natural to ask how this structure

1199: generalizes to other Calabi-Yau manifolds.  The structure we have

1200: found indicates that nongeometric fluxes may be thought of as some

1201: additional data which decorates the structure of some particular

1202: Calabi-Yau geometry.  This could naturally lead to a generalization of

1203: mirror symmetry, in which a Calabi-Yau in IIA or IIB decorated with

1204: one set of general NS-NS $H$, geometric, and nongeometric fluxes is

1205: mapped through mirror symmetry to the mirror Calabi-Yau in IIB or IIA

1206: decorated with the dual set of NS-NS fluxes.  In particular, in the

1207: picture of mirror symmetry as T-duality on a toroidal fibration

1208: \cite{syz}, $H$-flux with one leg on the $T^3$ fiber would map to

1209: geometric $f$-flux, $H$-flux with two legs on the $T^3$ fiber would

1210: map to nongeometric $Q$-flux, and $H$-flux wrapping the $T^3$ fiber

1211: would map to $R$-flux on the mirror Calabi-Yau.  The situation where

1212: the mirror of a Calabi-Yau with $H$-flux is geometrical (i.e., the

1213: $H$-flux has 0 or 1 legs on the $T^3$) has recently been described in

1214: detail in \cite{glmw, fmt}, following a suggestion in \cite{Vafa},

1215: using the generalized Calabi-Yau geometry developed by Hitchin; it

1216: would be very interesting to understand whether there is a precise way

1217: of extending that work to the nongeometric context considered here.

1218: The generalized tadpole and integrated Bianchi identities we derived

1219: in this paper should be valid in a more general context than just the

1220: toroidal model considered here, and may provide a good starting point

1221: for the concrete generalization of the picture presented in this

1222: paper.

1223: 

1224: It would also be nice to understand how S-duality fits into this

1225: framework.  As we have discussed at the end of Section 4, it is

1226: natural to expect that the framework we discuss here should be

1227: invariant under a full U-duality group generated by T-duality and

1228: S-duality transformations.  The constraints we have found on the

1229: geometric and nongeometric fluxes seem compatible with S-duality, but

1230: are not manifestly invariant, so some additional structure may be

1231: needed to form the full U-duality invariant picture.  We leave the

1232: resolution of this question as an outstanding problem for future work.

1233: 

1234: In this paper we have focused on a set of essentially topological

1235: features of string compactifications characterized by a general set of

1236: NS-NS fluxes.  We have described the interplay between these integral

1237: fluxes and a set of degrees of freedom (the torus moduli) chosen by

1238: considering the light degrees of freedom in the particular background

1239: without fluxes.  As we change fluxes, in different regions of flux

1240: space other stringy degrees of freedom will become light, as

1241: discussed for example in \cite{gp}.  Thus, in many cases the low-energy

1242: effective theory described by the superpotential we have computed

1243: here will not give a complete description of the physics.  This is an

1244: issue with any classification of flux vacua, but is more acute here

1245: where we do not necessarily have tools to assess the validity of the

1246: low-energy theory when all nongeometric fluxes are turned on.  Indeed,

1247: these nongeometric flux compactifications may generically appear at

1248: sub-string scales where the supergravity approximation is not valid;

1249: since these compactifications also have R-R fluxes, and, even in the

1250: locally geometric case, have complicated boundary conditions, we

1251: currently have no way of describing these backgrounds precisely 

1252: using perturbative string theory.  It is

1253: clearly important to understand better how the compactifications we

1254: describe here can be understood in terms of some fundamental

1255: formulation of string/M-theory.  In the

1256: full theory, the fluxes we have described here should be a

1257: useful tool for classifying and understanding string backgrounds.  In

1258: some cases, such as those dual to geometric compactifications in which

1259: the low-energy effective description is valid, we know by duality that the

1260: low-energy effective description given by the superpotential we have

1261: computed in terms of nongeometric fluxes will still be valid.  It is

1262: likely that there are other backgrounds which have no geometric dual,

1263: in which this low-energy description is still valid, though these

1264: backgrounds will be significantly harder to identify.

1265: 

1266: An obvious application of the formalism developed in this paper is to

1267: classify the full landscape of type II compactifications on tori

1268: with general NS-NS fluxes.  The first step in this program would be to

1269: determine the vacua arising from the superpotential we have computed

1270: here, after which it is necessary to determine corrections to the

1271: classical vacuum, including those from other fields which may become

1272: light as mentioned above.  We have explicitly computed the

1273: superpotential for the simplest model with 3 moduli, as well as all

1274: constraints on the fluxes.  Solutions to this superpotential will

1275: include not only all geometric IIA and IIB flux compactifications in

1276: this class, but also compactifications which involve nongeometric

1277: fluxes either in one or both pictures.  A more detailed analysis of

1278: the solution space for this model is currently underway and will be

1279: reported elsewhere \cite{stw-2}.  Unless there is some unexpected

1280: general obstruction to the solution of the SUSY equations, it may be

1281: possible to use methods developed here to demonstrate conclusively

1282: that generic string vacua are nongeometric, increasing yet further

1283: the size of the enormous haystack known as the ``string landscape'',

1284: in which we hope to find a compactification correctly describing our

1285: world's phenomenology and cosmology.

1286: 

1287: 

1288: \begin{center}

1289: \bf{Acknowledgements}

1290: \end{center}

1291: \medskip

1292: We would like to thank Katrin Becker, Oliver DeWolfe, Michael Douglas,

1293: Andrew Frey, Shamit Kachru, David Kutasov, Greg Moore, Li-Sheng Tseng,

1294: and Daniel Waldram for useful conversations, and other participants in

1295: the Summer 2005 Aspen Supercosmology workshop for helpful comments.

1296: Thanks to O.\ De Wolfe, S.\ Kachru, and G.\ Moore for comments on a

1297: preliminary version of this paper.  WT would like to thank National

1298: Taiwan University for hospitality during part of this work, and the

1299: Aspen Center for Physics for hospitality during the completion of this

1300: work, This work was supported by the DOE under contract

1301: \#DE-FC02-94ER40818. BW is additionally supported by NSF grant

1302: PHY-00-96515.

1303: 

1304: \vspace*{0.2in}

1305: 

1306: 

1307: 

1308: 

1309: \begin{thebibliography}{10}

1310: 

1311: % section 1 references

1312: 

1313: \bibitem{flux-c} % catchall for flux compactifications

1314: A.\ Strominger, ``Superstrings with torsion,''

1315:   Nucl.\ Phys.\ B {\bf 274}, 253 (1986);

1316: %x lookup 9510227

1317:   J.~Polchinski and A.~Strominger,

1318:   ``New Vacua for Type II String Theory,''

1319:   Phys.\ Lett.\ B {\bf 388}, 736 (1996)

1320:   {\tt hep-th/9510227};

1321: %x lookup 9605053

1322:   K.~Becker and M.~Becker,

1323:   ``M-Theory on Eight-Manifolds,''

1324:   Nucl.\ Phys.\ B {\bf 477}, 155 (1996)

1325:   {\tt hep-th/9605053};

1326: %x lookup 9610151

1327:   J.~Michelson,

1328:   ``Compactifications of type IIB strings to four dimensions with  non-trivial

1329:   classical potential,''

1330:   Nucl.\ Phys.\ B {\bf 495}, 127 (1997)

1331:   {\tt hep-th/9610151};

1332: %x lookup 0004103

1333:   B.~R.~Greene, K.~Schalm and G.~Shiu,

1334:   ``Warped compactifications in M and F theory,''

1335:   Nucl.\ Phys.\ B {\bf 584}, 480 (2000)

1336:   {\tt hep-th/0004103};

1337: %x lookup 0009211

1338:   M.~Grana and J.~Polchinski,

1339:   ``Supersymmetric three-form flux perturbations on AdS(5),''

1340:   Phys.\ Rev.\ D {\bf 63}, 026001 (2001)

1341:   {\tt hep-th/0009211};

1342: %x lookup 0012152

1343:   G.~Curio and A.~Krause,

1344:   ``Four-flux and warped heterotic M-theory compactifications,''

1345:   Nucl.\ Phys.\ B {\bf 602}, 172 (2001)

1346:   {\tt hep-th/0012152};

1347: %x lookup 0012213

1348:   G.~Curio, A.~Klemm, D.~Lust and S.~Theisen,

1349:   ``On the vacuum structure of type II string compactifications on  Calabi-Yau

1350:   spaces with H-fluxes,''

1351:   Nucl.\ Phys.\ B {\bf 609}, 3 (2001)

1352:   {\tt hep-th/0012213};

1353: %x lookup 0103068

1354:   M.~Haack and J.~Louis,

1355:   ``M-theory compactified on Calabi-Yau fourfolds with background flux,''

1356:   Phys.\ Lett.\ B {\bf 507}, 296 (2001)

1357:   {\tt hep-th/0103068};

1358: %x lookup 0107044

1359:   K.~Becker and M.~Becker,

1360:   ``Supersymmetry breaking, M-theory and fluxes,''

1361:   JHEP {\bf 0107}, 038 (2001)

1362:   {\tt hep-th/0107044};

1363: %x lookup 0202168

1364:   J.~Louis and A.~Micu,

1365:   ``Type II theories compactified on Calabi-Yau threefolds in the presence  of

1366:   background fluxes,''

1367:   Nucl.\ Phys.\ B {\bf 635}, 395 (2002)

1368:   {\tt hep-th/0202168}.

1369: 

1370: 

1371: 

1372: \bibitem{gvw}

1373:   S.~Gukov, C.~Vafa and E.~Witten,

1374:   ``CFT's from Calabi-Yau four-folds,''

1375:   Nucl.\ Phys.\ B {\bf 584}, 69 (2000)

1376:   [Erratum-ibid.\ B {\bf 608}, 477 (2001)]

1377:   {\tt hep-th/9906070}.

1378: 

1379: 

1380: \bibitem{drs}

1381: %x lookup 9908088

1382:   K.~Dasgupta, G.~Rajesh and S.~Sethi,

1383:   ``M theory, orientifolds and G-flux,''

1384:   JHEP {\bf 9908}, 023 (1999)

1385:   {\tt hep-th/9908088}.

1386: 

1387: 

1388: 

1389: \bibitem{gkp}

1390: %x lookup 0105097

1391:   S.~B.~Giddings, S.~Kachru and J.~Polchinski,

1392:   ``Hierarchies from fluxes in string compactifications,''

1393:   Phys.\ Rev.\ D {\bf 66}, 106006 (2002)

1394:   {\tt hep-th/0105097}.

1395: 

1396: 

1397: 

1398: \bibitem{kst}

1399:   S.~Kachru, M.~B.~Schulz and S.~Trivedi,

1400:   ``Moduli stabilization from fluxes in a simple IIB orientifold,''

1401:   JHEP {\bf 0310}, 007 (2003)

1402:   {\tt hep-th/0201028}.

1403: 

1404: \bibitem{fp}

1405: %x lookup 0201029

1406:   A.~R.~Frey and J.~Polchinski,

1407:   ``N = 3 warped compactifications,''

1408:   Phys.\ Rev.\ D {\bf 65}, 126009 (2002)

1409:   {\tt hep-th/0201029}.

1410: 

1411: 

1412:   

1413: \bibitem{Hitchin}

1414: N.\ Hitchin,

1415: ``Generalized Calabi-Yau manifolds,'' {\tt  math.dg/0209099}

1416: 

1417: 

1418: 

1419: \bibitem{glmw}

1420: %x lookup 0211102

1421:   S.~Gurrieri, J.~Louis, A.~Micu and D.~Waldram,

1422:   ``Mirror symmetry in generalized Calabi-Yau compactifications,''

1423:   Nucl.\ Phys.\ B {\bf 654}, 61 (2003)

1424:   {\tt hep-th/0211102};

1425: %x lookup 0212278

1426:   S.~Gurrieri and A.~Micu,

1427:   ``Type IIB theory on half-flat manifolds,''

1428:   Class.\ Quant.\ Grav.\  {\bf 20}, 2181 (2003)

1429:   {\tt hep-th/0212278}.

1430: 

1431: 

1432: \bibitem{fmt}

1433: %x lookup 0311122

1434:   S.~Fidanza, R.~Minasian and A.~Tomasiello,

1435:   ``Mirror symmetric SU(3)-structure manifolds with NS fluxes,''

1436:   Commun.\ Math.\ Phys.\  {\bf 254}, 401 (2005)

1437:   {\tt hep-th/0311122}.

1438: 

1439: 

1440: \bibitem{su3} 

1441: %x lookup 0211118

1442:   G.~L.~Cardoso, G.~Curio, G.~Dall'Agata, D.~Lust, P.~Manousselis and G.~Zoupanos,

1443:   ``Non-Kaehler string backgrounds and their five torsion classes,''

1444:   Nucl.\ Phys.\ B {\bf 652}, 5 (2003)

1445:   {\tt hep-th/0211118};

1446: %x lookup 0301161

1447:   K.~Becker, M.~Becker, K.~Dasgupta and P.~S.~Green,

1448:   ``Compactifications of heterotic theory on non-Kaehler complex manifolds.

1449:   I,''

1450:   JHEP {\bf 0304}, 007 (2003)

1451:   {\tt hep-th/0301161};

1452: %x lookup 0304001

1453:   K.~Becker, M.~Becker, K.~Dasgupta and S.~Prokushkin,

1454:   ``Properties of heterotic vacua from superpotentials,''

1455:   Nucl.\ Phys.\ B {\bf 666}, 144 (2003)

1456:   {\tt hep-th/0304001};

1457: %x lookup 0306088

1458:   G.~L.~Cardoso, G.~Curio, G.~Dall'Agata and D.~Lust,

1459:   ``BPS action and superpotential for heterotic string compactifications  with

1460:   fluxes,''

1461:   JHEP {\bf 0310}, 004 (2003)

1462:   {\tt hep-th/0306088};

1463: %x lookup 0310021

1464:   G.~L.~Cardoso, G.~Curio, G.~Dall'Agata and D.~Lust,

1465:   ``Heterotic string theory on non-Kaehler manifolds with H-flux and  gaugino

1466:   condensate,''

1467:   Fortsch.\ Phys.\  {\bf 52}, 483 (2004)

1468:   {\tt hep-th/0310021};

1469: %x lookup 0310058

1470:   K.~Becker, M.~Becker, P.~S.~Green, K.~Dasgupta and E.~Sharpe,

1471:   ``Compactifications of heterotic strings on non-Kaehler complex  manifolds.

1472:   II,''

1473:   Nucl.\ Phys.\ B {\bf 678}, 19 (2004)

1474:   {\tt hep-th/0310058};

1475: %x lookup 0310245

1476:   M.~Serone and M.~Trapletti,

1477: ``String vacua with flux from freely-acting orbifolds,''

1478:   JHEP {\bf 0401}, 012 (2004)

1479:   {\tt hep-th/0310245};

1480: %x lookup 0312221

1481:   M.~Becker and K.~Dasgupta,

1482:   ``Kaehler versus non-Kaehler compactifications,''

1483:   {\tt hep-th/0312221};

1484: %x lookup 0403288

1485:   M.~Becker, K.~Dasgupta, A.~Knauf and R.~Tatar,

1486:   ``Geometric transitions, flops and non-Kaehler manifolds. I,''

1487:   Nucl.\ Phys.\ B {\bf 702}, 207 (2004)

1488:   {\tt hep-th/0403288};

1489: %x lookup 0404107

1490:   A.~R.~Frey,

1491:   ``Notes on SU(3) structures in type IIB supergravity,''

1492:   JHEP {\bf 0406}, 027 (2004)

1493:   {\tt hep-th/0404107};

1494: %x lookup 0406137

1495:   M.~Grana, R.~Minasian, M.~Petrini and A.~Tomasiello,

1496:   ``Supersymmetric backgrounds from generalized Calabi-Yau manifolds,''

1497:   JHEP {\bf 0408}, 046 (2004)

1498:   {\tt hep-th/0406137};

1499: %x lookup 0408121

1500:   S.~Gurrieri, A.~Lukas and A.~Micu,

1501:   ``Heterotic on half-flat,''

1502:   Phys.\ Rev.\ D {\bf 70}, 126009 (2004)

1503:   {\tt hep-th/0408121};

1504: %x lookup 0408192

1505:   S.~Alexander, K.~Becker, M.~Becker, K.~Dasgupta, A.~Knauf and R.~Tatar,

1506:   ``In the realm of the geometric transitions,''

1507:   Nucl.\ Phys.\ B {\bf 704}, 231 (2005)

1508:   {\tt hep-th/0408192};

1509: %x lookup 0409008

1510:   A.~Micu,

1511:   ``Heterotic compactifications and nearly-Kaehler manifolds,''

1512:   Phys.\ Rev.\ D {\bf 70}, 126002 (2004)

1513:   {\tt hep-th/0409008};

1514: %x lookup 0505177

1515:   T.~House and E.~Palti,

1516:   ``Effective action of (massive) IIA on manifolds with SU(3) structure,''

1517:   Phys.\ Rev.\ D {\bf 72}, 026004 (2005)

1518:   {\tt hep-th/0505177};

1519: %x lookup 0505264

1520:   M.~Grana, J.~Louis and D.~Waldram,

1521:   ``Hitchin functionals in N = 2 supergravity,''

1522:   {\tt hep-th/0505264};

1523: %x lookup 0507173

1524:   B.~de Carlos, S.~Gurrieri, A.~Lukas and A.~Micu,

1525:   ``Moduli stabilisation in heterotic string compactifications,''

1526:   {\tt hep-th/0507173}.

1527: 

1528: 

1529: 

1530: 

1531: \bibitem{kstt}

1532:   S.~Kachru, M.~B.~Schulz, P.~K.~Tripathy and S.~P.~Trivedi,

1533:   ``New supersymmetric string compactifications,''

1534:   JHEP {\bf 0303}, 061 (2003)

1535:   {\tt hep-th/0211182}.

1536: 

1537: 

1538: \bibitem{ss}

1539:   J.~Scherk and J.~H.~Schwarz,

1540:   ``How To Get Masses From Extra Dimensions,''

1541:   Nucl.\ Phys.\ B {\bf 153}, 61 (1979).

1542: 

1543: 

1544: \bibitem{early-ss-reduction}

1545:   E.~Bergshoeff, M.~de Roo, M.~B.~Green, G.~Papadopoulos and P.~K.~Townsend,

1546:   ``Duality of Type II 7-branes and 8-branes,''

1547:   Nucl.\ Phys.\ B {\bf 470}, 113 (1996)

1548:   {\tt hep-th/9601150};

1549:   %%CITATION = HEP-TH 9601150;%%

1550:   P.~M.~Cowdall, H.~Lu, C.~N.~Pope, K.~S.~Stelle and P.~K.~Townsend,

1551:   ``Domain walls in massive supergravities,''

1552:   Nucl.\ Phys.\ B {\bf 486}, 49 (1997)

1553:   {\tt hep-th/9608173}.

1554:   %%CITATION = HEP-TH 9608173;%%

1555: 

1556: 

1557: 

1558: \bibitem{km}

1559:   N.~Kaloper and R.~C.~Myers,

1560:   ``The O(dd) story of massive supergravity,''

1561:   JHEP {\bf 9905}, 010 (1999)

1562:   {\tt hep-th/9901045}.

1563: 

1564: 

1565: 

1566: 

1567:   \bibitem{Schulz}

1568: %x lookup 0406001

1569:   M.~B.~Schulz,

1570:   ``Superstring orientifolds with torsion: O5 orientifolds of torus fibrations

1571:   and their massless spectra,''

1572:   Fortsch.\ Phys.\  {\bf 52}, 963 (2004)

1573:   {\tt hep-th/0406001};

1574: %x lookup 0412270

1575:   M.~B.~Schulz,

1576:   ``Calabi-Yau duals of torus orientifolds,''

1577:   {\tt hep-th/0412270}.

1578: 

1579: 

1580: \bibitem{dkpz}

1581:   J.~P.~Derendinger, C.~Kounnas, P.~M.~Petropoulos and F.~Zwirner,

1582:   ``Superpotentials in IIA compactifications with general fluxes,''

1583:   Nucl.\ Phys.\ B {\bf 715}, 211 (2005)

1584:   {\tt hep-th/0411276};

1585: %x lookup 0503229

1586:   J.~P.~Derendinger, C.~Kounnas, P.~M.~Petropoulos and F.~Zwirner,

1587:   ``Fluxes and gaugings: N = 1 effective superpotentials,''

1588:   Fortsch.\ Phys.\  {\bf 53}, 926 (2005)

1589:   {\tt hep-th/0503229}.

1590: 

1591: 

1592: 

1593: \bibitem{x1}

1594: %x lookup 0502066

1595:   G.~Dall'Agata and S.~Ferrara,

1596:   ``Gauged supergravity algebras from twisted tori compactifications with

1597:   fluxes,''

1598:   Nucl.\ Phys.\ B {\bf 717}, 223 (2005)

1599:   {\tt hep-th/0502066}.

1600: 

1601: 

1602: 

1603: 

1604: \bibitem{hre} 

1605:   C.~M.~Hull and R.~A.~Reid-Edwards,

1606:   ``Flux compactifications of string theory on twisted tori,''

1607:   {\tt hep-th/0503114}.

1608: 

1609: 

1610: \bibitem{vz}

1611:   G.~Villadoro and F.~Zwirner,

1612:   ``N = 1 effective potential from dual type-IIA D6/O6 orientifolds with

1613:   general fluxes,''

1614:   JHEP {\bf 0506}, 047 (2005)

1615:   {\tt hep-th/0503169}.

1616: 

1617: \bibitem{cfi}

1618:   P.~G.~Camara, A.~Font and L.~E.~Ibanez,

1619:   ``Fluxes, moduli fixing and MSSM-like vacua in a simple IIA orientifold,''

1620:   {\tt hep-th/0506066}.

1621:   

1622: 

1623: \bibitem{dh}

1624:   A.~Dabholkar and C.~Hull,

1625:   ``Duality twists, orbifolds, and fluxes,''

1626:   JHEP {\bf 0309}, 054 (2003)

1627:   {\tt hep-th/0210209}.

1628:   

1629:   

1630:   \bibitem{Hellerman:2002ax}

1631:   S.~Hellerman, J.~McGreevy and B.~Williams,

1632:   ``Geometric constructions of nongeometric string theories,''

1633:   JHEP {\bf 0401}, 024 (2004)

1634:   {\tt hep-th/0208174}.

1635:   %%CITATION = HEP-TH 0208174;%%

1636: 

1637: 

1638: 

1639: \bibitem{fww}

1640:   A.~Flournoy, B.~Wecht and B.~Williams,

1641:   ``Constructing nongeometric vacua in string theory,''

1642:   Nucl.\ Phys.\ B {\bf 706}, 127 (2005)

1643:   {\tt hep-th/0404217}.

1644: 

1645: \bibitem{x2}

1646: %x lookup 0406102

1647:   C.~M.~Hull,

1648:   ``A geometry for nongeometric string backgrounds,''

1649:   {\tt hep-th/0406102}.

1650: 

1651: \bibitem{gh}

1652: %x lookup 0506092

1653:   J.~Gray and E.~Hackett-Jones,

1654:   ``On T-folds, G-structures and supersymmetry,''

1655:   {\tt hep-th/0506092}.

1656: 

1657: 

1658: 

1659: \bibitem{Buscher}

1660:   T.~H.~Buscher,

1661:   ``A Symmetry Of The String Background Field Equations,''

1662:   Phys.\ Lett.\ B {\bf 194}, 59 (1987).

1663:   %%CITATION = PHLTA,B194,59;%%

1664: 

1665: \bibitem{bho}

1666: E.\ Bergshoeff, C.\ Hull and T.\ Ortin,

1667: ``Duality in the type-II superstring effective action''

1668:   Nucl.\ Phys.\ B {\bf 451}, 547 (1995)

1669: 

1670: \bibitem{mo}

1671:   P.~Meessen and T.~Ortin,

1672:   ``An Sl(2,Z) multiplet of nine-dimensional type II supergravity theories,''

1673:   Nucl.\ Phys.\ B {\bf 541}, 195 (1999)

1674:   {\tt hep-th/9806120}.

1675: 

1676: 

1677: \bibitem{ght}

1678:   M.~B.~Green, C.~M.~Hull and P.~K.~Townsend,

1679:   ``D-Brane Wess-Zumino Actions, T-Duality and the Cosmological Constant,''

1680:   Phys.\ Lett.\ B {\bf 382}, 65 (1996)

1681:   {\tt hep-th/9604119}.

1682: 

1683: 

1684: 

1685: \bibitem{Hassan}

1686: S.\ F.\ Hassan,

1687:   ``T-duality, space-time spinors and R-R fields in curved backgrounds,''

1688:   Nucl.\ Phys.\ B {\bf 568}, 145 (2000)

1689:   {\tt hep-th/9907152}.

1690: 

1691: 

1692: 

1693: 

1694: 

1695: 

1696: 

1697: \bibitem{Mathai}

1698: %x lookup 0306062

1699:   P.~Bouwknegt, J.~Evslin and V.~Mathai,

1700:   ``T-duality: Topology change from H-flux,''

1701:   Commun.\ Math.\ Phys.\  {\bf 249}, 383 (2004)

1702:   {\tt hep-th/0306062};

1703: %x lookup 0312052

1704:   P.~Bouwknegt, J.~Evslin and V.~Mathai,

1705:   ``On the topology and H-flux of T-dual manifolds,''

1706:   Phys.\ Rev.\ Lett.\  {\bf 92}, 181601 (2004)

1707:   {\tt hep-th/0312052};

1708: %x lookup 0401168

1709:   V.~Mathai and J.~M.~Rosenberg,

1710:   ``T-duality for torus bundles via noncommutative topology,''

1711:   Commun.\ Math.\ Phys.\  {\bf 253}, 705 (2004)

1712:   {\tt hep-th/0401168};

1713: %x lookup 0409073

1714:   V.~Mathai and J.~M.~Rosenberg,

1715:   ``On mysteriously missing T-duals, H-flux and the T-duality group,''

1716:   {\tt hep-th/0409073};

1717: %x lookup 0412092

1718:   P.~Bouwknegt, K.~Hannabuss and V.~Mathai,

1719:   ``Nonassociative tori and applications to T-duality,''

1720:   {\tt hep-th/0412092};

1721: %x lookup 0508084

1722:   V.~Mathai and J.~Rosenberg,

1723:   ``T-duality for torus bundles with H-fluxes via noncommutative topology. II:

1724:   The high-dimensional case and the T-duality group,''

1725:   {\tt hep-th/0508084}.

1726: 

1727: %-------;

1728: 

1729: % section 2 references

1730: 

1731: \bibitem{x3}

1732: %x lookup 0301139

1733:   P.~K.~Tripathy and S.~P.~Trivedi,

1734:   ``Compactification with flux on K3 and tori,''

1735:   JHEP {\bf 0303}, 028 (2003)

1736:   {\tt hep-th/0301139}.

1737: 

1738: \bibitem{dgkt1}

1739:   O.~DeWolfe, A.~Giryavets, S.~Kachru and W.~Taylor,

1740:   ``Enumerating flux vacua with enhanced symmetries,''

1741:   JHEP {\bf 0502}, 037 (2005)

1742:   {\tt hep-th/0411061}.

1743: 

1744: 

1745: 

1746: \bibitem{x4}

1747: %x lookup 0411279

1748:   S.~Kachru and A.~K.~Kashani-Poor,

1749:   ``Moduli potentials in type IIA compactifications with RR and NS flux,''

1750:   JHEP {\bf 0503}, 066 (2005)

1751:   {\tt hep-th/0411279}.

1752: 

1753: 

1754: 

1755: \bibitem{x5}

1756: %x lookup 0412277

1757:   T.~W.~Grimm and J.~Louis,

1758:   ``The effective action of type IIA Calabi-Yau orientifolds,''

1759:   Nucl.\ Phys.\ B {\bf 718}, 153 (2005)

1760:   {\tt hep-th/0412277}.

1761: 

1762: 

1763: 

1764: \bibitem{dgkt2}

1765:   O.~DeWolfe, A.~Giryavets, S.~Kachru and W.~Taylor,

1766:   ``Type IIA moduli stabilization,''

1767:   {\tt hep-th/0505160}.

1768: 

1769: 

1770: 

1771: \bibitem{K-theory}

1772: %x lookup 9710230

1773:   R.~Minasian and G.~W.~Moore,

1774:   ``K-theory and Ramond-Ramond charge,''

1775:   JHEP {\bf 9711}, 002 (1997)

1776:   {\tt hep-th/9710230};

1777: %x lookup 9810188

1778:   E.~Witten,

1779:   ``D-branes and K-theory,''

1780:   JHEP {\bf 9812}, 019 (1998)

1781:   {\tt hep-th/9810188};

1782: %x lookup 9812135

1783:   P.~Horava,

1784:   ``Type IIA D-branes, K-theory, and matrix theory,''

1785:   Adv.\ Theor.\ Math.\ Phys.\  {\bf 2}, 1373 (1999)

1786:   {\tt hep-th/9812135};

1787: %\bibitem{Moore:1999gb}

1788:   G.~W.~Moore and E.~Witten,

1789:   ``Self-duality, Ramond-Ramond fields, and K-theory,''

1790:   JHEP {\bf 0005}, 032 (2000)

1791:   {\tt hep-th/9912279};

1792:   %%CITATION = HEP-TH 9912279;%%

1793: %\cite{Diaconescu:2000wy}

1794: %\bibitem{Diaconescu:2000wy}

1795:   D.~E.~Diaconescu, G.~W.~Moore and E.~Witten,

1796:   ``E(8) gauge theory, and a derivation of K-theory from M-theory,''

1797:   Adv.\ Theor.\ Math.\ Phys.\  {\bf 6}, 1031 (2003)

1798:   {\tt hep-th/0005090};

1799:   %%CITATION = HEP-TH 0005090;%%

1800: %\cite{Diaconescu:2000wz}

1801: %\bibitem{Diaconescu:2000wz}

1802:   D.~E.~Diaconescu, G.~W.~Moore and E.~Witten,

1803:   ``A derivation of K-theory from M-theory,''

1804:   {\tt hep-th/0005091};

1805:   %%CITATION = HEP-TH 0005091;%%

1806: %\cite{Freed:2000ta}

1807: %\bibitem{Freed:2000ta}

1808:   D.~S.~Freed,

1809:   ``Dirac charge quantization and generalized differential cohomology,''

1810:   {\tt hep-th/0011220};

1811:   %%CITATION = HEP-TH 0011220;%%

1812: %

1813:   G.~W.~Moore and N.~Saulina,

1814:   ``T-duality, and the K-theoretic partition function of typeIIA  superstring

1815:   theory,''

1816:   Nucl.\ Phys.\ B {\bf 670}, 27 (2003)

1817:   {\tt hep-th/0206092};

1818:   %%CITATION = HEP-TH 0206092;%%

1819: %x lookup 0312033

1820:   V.~Mathai and H.~Sati,

1821:   ``Some relations between twisted K-theory and E(8) gauge theory,''

1822:   JHEP {\bf 0403}, 016 (2004)

1823:   {\tt hep-th/0312033}.

1824: 

1825: 

1826: 

1827: \bibitem{3}

1828: %x lookup 0103170

1829:   J.~de Boer, R.~Dijkgraaf, K.~Hori, A.~Keurentjes, J.~Morgan, D.~R.~Morrison and S.~Sethi,

1830:   ``Triples, fluxes, and strings,''

1831:   Adv.\ Theor.\ Math.\ Phys.\  {\bf 4}, 995 (2002)

1832:   {\tt hep-th/0103170}.

1833: 

1834: 

1835: 

1836: \bibitem{stw-2}

1837: J.\ Shelton, W.\ Taylor and B.\ Wecht, to appear.

1838: 

1839:   

1840: % section 3 references

1841: 

1842: \bibitem{llp}

1843:   I.~V.~Lavrinenko, H.~Lu and C.~N.~Pope,

1844:   ``Fibre bundles and generalised dimensional reductions,''

1845:   Class.\ Quant.\ Grav.\  {\bf 15}, 2239 (1998)

1846:   {\tt hep-th/9710243}.

1847:   %%CITATION = HEP-TH 9710243;%%

1848: 

1849: \bibitem{hull-98}

1850:   C.~M.~Hull,

1851:   ``Massive string theories from M-theory and F-theory,''

1852:   JHEP {\bf 9811}, 027 (1998)

1853:   {\tt hep-th/9811021}.

1854:   %%CITATION = HEP-TH 9811021;%%

1855: 

1856:   

1857: 

1858: 

1859:   \bibitem{Narain:1986qm}

1860:   K.~S.~Narain, M.~H.~Sarmadi and C.~Vafa,

1861:   ``Asymmetric Orbifolds,''

1862:   Nucl.\ Phys.\ B {\bf 288}, 551 (1987).

1863:   %%CITATION = NUPHA,B288,551;%%

1864:   

1865:   \bibitem{syz}

1866:   A.~Strominger, S.~T.~Yau and E.~Zaslow,

1867:   ``Mirror symmetry is T-duality,''

1868:   Nucl.\ Phys.\ B {\bf 479}, 243 (1996)

1869:   {\tt hep-th/9606040}.

1870:   %%CITATION = HEP-TH 9606040;%%

1871:   

1872:   \bibitem{Vafa}

1873: %x lookup 0008142

1874:   C.~Vafa,

1875:   ``Superstrings and topological strings at large N,''

1876:   J.\ Math.\ Phys.\  {\bf 42}, 2798 (2001)

1877:   {\tt hep-th/0008142}.

1878: 

1879: \bibitem{gp}

1880:   A.~Giveon, M.~Porrati and E.~Rabinovici,

1881:   ``Target space duality in string theory,''

1882:   Phys.\ Rept.\  {\bf 244}, 77 (1994)

1883:   {\tt hep-th/9401139}.

1884: 

1885: 

1886: 

1887: 

1888: \end{thebibliography}

1889: 

1890: \end{document}

1891: 

1892: 

1893: 

1894: 

1895: 

1896: 

1897: