1: \input harvmac
2: \input epsf
3: %\input srctex.sty
4:
5: \def\figin{\epsfcheck\figin}\def\figins{\epsfcheck\figins}
6: \def\epsfcheck{\ifx\epsfbox\UnDeFiNeD
7: \message{(NO epsf.tex, FIGURES WILL BE IGNORED)}
8: \gdef\figin##1{\vskip2in}\gdef\figins##1{\hskip.5in}% blank space instead
9: \else\message{(FIGURES WILL BE INCLUDED)}%
10: \gdef\figin##1{##1}\gdef\figins##1{##1}\fi}
11: \def\DefWarn#1{}
12: \def\figinsert{\goodbreak\topinsert}
13: \def\ifig#1#2#3#4{\DefWarn#1\xdef#1{fig.~\the\figno}
14: \writedef{#1\leftbracket fig.\noexpand~\the\figno}%
15: \figinsert\figin{\centerline{\epsfxsize=#3mm \epsfbox{#2}}}
16: \bigskip\medskip\centerline{\vbox{\baselineskip12pt
17: \advance\hsize by -1truein\noindent\footnotefont{\sl Fig.~\the\figno:}\sl\ #4}}
18: \bigskip\endinsert\noindent\global\advance\figno by1}
19:
20: \def\C{{\bf C}}
21: \def\R{{\bf R}}
22: \def\Z{{\bf Z}}
23:
24: \def\a{\alpha}
25: \def\b{\beta}
26: \def\g{\gamma}
27: \def\G{\Gamma}
28: \def\e{\epsilon}
29: \def\h{\widehat}
30: \def\th{\theta}
31: \def\Th{\Theta}
32: \def\k{\kappa}
33: \def\l{\lambda}
34: \def\L{\Lambda}
35: \def\m{\mu}
36: \def\n{\nu}
37: \def\r{\rho}
38: \def\s{\sigma}
39: \def\t{\tau}
40: \def\f{\phi}
41: \def\F{\Phi}
42: \def\w{\omega}
43: \def\W{\Omega}
44: \def\v{\varphi}
45: \def\z{\zeta}
46:
47: \def\P{{\bf P}}
48:
49: \def\d{\partial}
50: \def\dbar{{\overline\partial}}
51: \def\inv{^{-1}}
52: \def\Tr{{\rm Tr}}
53: \def\hf{{1\over 2}}
54: \def\cO{{\cal O}}
55: \def\cC{{\cal C}}
56: \def\cF{{\cal F}}
57: \def\cN{{\cal N}}
58: \def\cW{{\cal W}}
59: \def\cG{{\cal G}}
60: \def\cS{{\cal S}}
61: \def\cE{{\cal E}}
62:
63: \def\({\left(}
64: \def\){\right)}
65: \def\<{\left\langle}
66: \def\>{\right\rangle}
67:
68: \def\Im{{\rm Im}\;}
69: \def\mbar{\overline{m}}
70: \def\psibar{\overline{\psi}}
71: \def\Fbar{\overline{\F}}
72: \def\Fb{\overline{\F}}
73: \def\Ft{\widetilde{\F}}
74:
75: \def\Phib{\Fbar}
76: \def\Db{{\overline{D}}}
77: \def\lb{{\overline{\l}}}
78: \def\Qb{{\overline{Q}}}
79: \def\thb{{\overline{\th}}}
80: \def\mod{\ {\rm mod}\ }
81:
82: \def\gives{{$\Longrightarrow$}}
83: \def\ad{{\dot{\alpha}}}
84: \def\bd{{\dot{\beta}}}
85: \def\Dslash{D}
86: \def\Nt{{\widetilde{N}}}
87: \def\vol{{\rm vol}}
88:
89: \def\Del{\nabla}
90: \def\Delb{\overline{\nabla}}
91: \def\Delp{\nabla_{+}}
92: \def\Delm{\nabla_{-}}
93: \def\Delpd{\nabla_{\pd}}
94: \def\Delmd{\nabla_{\md}}
95: \def\Delpp{\nabla_{\pp}}
96: \def\Delmm{\nabla_{\mm}}
97: \def\Delsq{\nabla^2}
98: \def\Delbsq{{\bar{\nabla}}^2}
99: \def\Dela{\nabla_{\alpha}}
100: \def\Delad{\nabla_{\dot{\alpha}}}
101: \def\DelA{\nabla_A}
102: \def\Deltsq{{\tilde{\Del}}^2}
103: \def\frac#1#2{{#1 \over #2}}
104: \def\em{\it}
105:
106: %\def\Box{\Delta}
107:
108: \def\sqr#1#2{{%\vcenter
109: {\vbox{\hrule height.#2pt
110: \hbox{\vrule width.#2pt height#1pt \kern#1pt
111: \vrule width.#2pt} \hrule height.#2pt}}}}
112:
113: \def\Box{{\,\sqr66\,}}
114:
115: \lref\wit{
116: E.~Witten, ``Topological sigma models,'' Commun.\ Math.\ Phys.\ {\bf
117: 118}, 411 (1988); ``Mirror manifolds and topological field theory,''
118: arXiv:hep-th/9112056. }
119:
120: \lref\katz{
121: D.A.~Cox and S.~Katz, {\it Mirror Symmetry and Algebraic Geometry},
122: Mathematical Surveys and Monographs, AMS, Providence, 1999.
123: }
124:
125: \lref\mirb{
126: {\it Mirror Symmetry}, Clay Mathematics Monographs, Vol. 1, C.~Vafa
127: and E.~Zaslow, eds., AMS-CMI, Providence, 2003.
128: }
129:
130: \lref\kkv{
131: S.~Katz, A.~Klemm and C.~Vafa, ``M-theory, topological strings and
132: spinning black holes,'' Adv.\ Theor.\ Math.\ Phys.\ {\bf 3}, 1445
133: (1999) [arXiv:hep-th/9910181]. }
134:
135: \lref\gova{
136: R.~Gopakumar and C.~Vafa, ``M-theory and topological strings. I, II''
137: arXiv:hep-th/9809187, arXiv:hep-th/9812127.}
138:
139: \lref\osv{
140: H.~Ooguri, A.~Strominger and C.~Vafa, ``Black hole attractors and the
141: topological string,'' Phys.\ Rev.\ D {\bf 70}, 106007 (2004)
142: [arXiv:hep-th/0405146].
143: }
144:
145: \lref\andy{
146: D.~Gaiotto, A.~Strominger and X.~Yin, ``New connections between 4D and
147: 5D black holes,'' arXiv:hep-th/0503217.}
148:
149: \lref\ionv{
150: A.~Iqbal, N.~Nekrasov, A.~Okounkov and C.~Vafa, ``Quantum foam and
151: topological strings,'' hep-th/0312022.}
152:
153: \lref\nv{
154: A.~Neitzke and C.~Vafa, ``Topological strings and their physical
155: applications,'' arXiv:hep-th/0410178. }
156:
157: \lref\bcov{
158: M.~Bershadsky, S.~Cecotti, H.~Ooguri and C.~Vafa, ``Kodaira-Spencer
159: theory of gravity and exact results for quantum string amplitudes,''
160: Commun.\ Math.\ Phys.\ {\bf 165}, 311 (1994),
161: [arXiv:hep-th/9309140].}
162:
163: \lref\narain{
164: I.~Antoniadis, E.~Gava, K.S.~Narain and T.R.~Taylor, ``Topological
165: amplitudes in string theory,'' Nucl.\ Phys.\ B\ {\bf 413}, 162 (1994)
166: [arXiv:hep-th/9307158]. }
167:
168: \lref\mnop{
169: D. Maulik, N. Nekrasov, A. Okounkov and R. Pandharipande, "Gromov-Witten
170: theory and Donaldson-Thomas theory, I-II", math.AG/0312059;
171: math.AG/0406092. }
172:
173: \lref\orv{
174: A.~Okounkov, N.~Reshetikhin and C.~Vafa,
175: ``Quantum Calabi-Yau and classical crystals,'' hep-th/0309208.}
176:
177: \lref\DT{
178: S.K.~Donaldson and R.P.~Thomas, ``Gauge theory in higher dimensions,'' in
179: {\it The Geometric Universe; Science, Geometry, And The Work Of Roger
180: Penrose}, Oxford University Press, 1998.}
181:
182: \lref\dvv{
183: R.~Dijkgraaf, E.~Verlinde and M.~Vonk,
184: {``On the partition sum of the NS five-brane,''} hep-th/0205281.}
185:
186: \lref\vw{
187: C.~Vafa and E.~Witten,
188: ``A Strong coupling test of S duality,''
189: Nucl.\ Phys.\ B {\bf 431}, 3 (1994)
190: [arXiv:hep-th/9408074].
191: }
192:
193: \lref\tobe{
194: M.~Cheng, R.~Dijkgraaf, J.~Manschot and E.~Verlinde, work in progress, to
195: appear.}
196:
197: \lref\nov{
198: N.~Nekrasov, H.~Ooguri and C.~Vafa,
199: ``S-duality and topological strings,''
200: JHEP {\bf 0410}, 009 (2004)
201: [arXiv:hep-th/0403167].
202: }
203:
204: \lref\dmmv{
205: R.~Dijkgraaf, J.~M.~Maldacena, G.~W.~Moore and E.~P.~Verlinde,
206: ``A black hole Farey tail,''
207: arXiv:hep-th/0005003.
208: }
209:
210:
211: \lref\gsy{
212: D.~Gaiotto, A.~Strominger and X.~Yin,
213: `` From AdS3/CFT2 to Black Holes/Topological Strings,''
214: hep-th/0602046.
215: }
216:
217: %%%%%%%%%%%%%%%%%%% TITLE PAGE %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
218:
219: \Title{\vbox{\baselineskip11pt
220: \hbox{HUTP-06/A001}
221: \hbox{ITFA-2006-05} }}
222: {\vbox{
223: \centerline{M-theory and a Topological String Duality
224: } }}
225: \centerline{
226: Robbert Dijkgraaf,$\!{}^{1,2}$ Cumrun Vafa$^3$ and Erik Verlinde$^1$}
227: \medskip
228: \medskip
229: \medskip
230: %\medskip
231: \vskip 8pt
232: \centerline{\it $^1$Institute for Theoretical Physics, and $\,\,{}^2$KdV
233: Institute for Mathematics,}
234: \centerline{\it University of Amsterdam, Valckenierstraat 65,
235: 1018 XE Amsterdam, The Netherlands.}
236: \medskip
237: \centerline{\it
238: $^3$Jefferson Physical Laboratory, Harvard University, Cambridge, MA
239: 02138, USA.}
240: \medskip
241: \medskip
242: %\medskip
243: \medskip
244: %\bigskip
245: \noindent
246: We show how the topological string partition function, which is known
247: to capture the degeneracies of a gas of BPS spinning M2-branes in
248: M-theory compactified to 5 dimensions, is related to a 4-dimensional
249: D-brane system that consists of single D6-brane bound to
250: lower-dimensional branes. This system is described by a topologically
251: twisted $U(1)$ gauge theory, that has been conjecturally identified
252: with quantum foam models and topological strings. This also explains,
253: assuming the identification of Donaldson-Thomas invariants with this
254: $U(1)$ gauge theory, the conjectural relation between DT invariants
255: and topological strings. Our results provide further mathematical
256: evidence for the recently found connection between 4d and 5d black
257: holes.
258:
259: \medskip
260:
261: \Date{February 2006}
262:
263: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
264: \newsec{Introduction}
265:
266: A-model topological string \wit\ has been well studied from physical
267: and mathematical perspectives (see \refs{\katz,\mirb} for a detailed
268: review of the subject as well as \nv\ for a more recent update). It
269: is a source of inspiration for quite a number of diverse areas on the
270: physical and mathematical fronts. On the physics side it captures,
271: worldsheet instantons for sigma models, relevant configurations for
272: superpotential corrections for superstrings as well as counting the
273: BPS black hole degeneracies for 5d black holes \gova\ and more
274: recently connected to melting of crystals as a model for quantum
275: gravitational foam and a $U(1)$ topologically twisted Yang-Mills
276: theory on the Calabi-Yau \ionv . On the mathematical front it is
277: known as the study of Gromov-Witten invariant, which has been defined
278: rigorously and studied extensively. It has also recently been
279: conjecturally connected to Donaldson-Thomas (DT) invariants \mnop.
280:
281:
282: The relation between the physics and mathematical developments has
283: been basically clear: The worldsheet instanton corrections captured by
284: A-model topological strings is simply what one means by Gromov-Witten
285: invariant computations. The quantum gravitational foam captured by
286: the $U(1)$ topologically twisted Yang-Mills theory, is believed
287: (though not proven) to be the physical formulation of Donaldson-Thomas
288: invariants. So to complete the missing link we have to connect the 5d
289: BPS spinning black hole degeneracies with the $U(1)$ topologically
290: twisted Yang-Mills theory on the Calabi-Yau. This would be relating
291: what is called the Gopakumar-Vafa (GV) invariants, with the DT invariants.
292: It is the aim of this brief note to explain this link from an M-theory
293: perspective.
294:
295: The idea is rather simple: the topological string partition function
296: has an interpretation in M-theory, where it can be reformulated in
297: terms of the integer GV invariants \gova, that calculate the
298: BPS-degeneracies of M2-branes with angular momentum wrapped on
299: 2-cycles in the Calabi-Yau.
300:
301: On the other hand, the DT invariants, or the partition function of the
302: (presumably equivalent) $U(1)$ topologically twisted super-Yang-Mills
303: theory, compute the bound states of wrapped D2 and D0-branes to a
304: single D6-brane, that completely covers the Calabi-Yau. To relate the
305: two set of invariants we make use of the recent observation of \andy\
306: that connects the physics of 4d and 5d black holes.
307:
308: The configuration of D-branes relevant for the DT calculation can be
309: lifted to M-theory as follows. In M-theory the D2-branes simply become
310: M2-branes. The D6-brane located at a fixed point in the spatial $\R^3$
311: is represented by a Taub-NUT geometry. Asymptotically this looks like
312: $\R^3 \times S^1$, where the circle is the extra 11${}^{\rm th}$
313: dimension. At the center the geometry is smooth and given by
314: $\R^4$. The D0-brange charge is associated with a $U(1)$ isometry of
315: the Taub-Nut. At infinity this is the momentum along the circle, but
316: at the center it becomes a rotation. Therefore, the D-brane system can
317: be identified with a collection of spinning M2-branes in the TN
318: geometry.
319:
320: In this paper we will argue that the degeneracy of the D-brane bound
321: state, which by a most reasonable conjecture is computed by the DT
322: invariant, is equal to the degeneracy of a free gas of M2-branes. So,
323: in this way are able to relate the second quantized Hilbert space of
324: 5d BPS states with the DT invariants and thereby complete the circle
325: of ideas.
326:
327: The plan for this paper is as follows: In section 2 we review the
328: relation between A-model topological strings and BPS degeneracies for
329: 5d black holes. In section 3 we review the conjectures relating
330: topological strings with the $U(1)$ topologically twisted gauge
331: theory, which is believed to be the physical formulation of DT
332: invariants. In section 4 we connect these two viewpoints by
333: considering an extra D6-brane wrapping the Calabi-Yau and thus derive
334: the conjectural relation between $U(1)$ topologically twisted gauge
335: theory ({\it i.e.} conjecturally DT invariants) and the topological
336: A-model.
337:
338: \newsec{A-model topological string and 5d spinning black holes}
339:
340: It has been known that topological strings compute certain amplitudes
341: in compactifications of superstrings on Calabi-Yau manifolds down to
342: four dimensions \refs{\bcov,\narain}. In particular
343: the genus $g$ amplitude $F_g$ compute corrections to the ${\cal N}=2$
344: effective action of the form
345: %
346: \eqn\tops{\int d^4x \, d^4\theta \ F_g(t) {\cal W}^{2g-2}, }
347: %
348: where $t$ are the K\"ahler moduli and $\cal W$ is the graviphoton
349: superfield. By viewing these corrections as the contribution of
350: D2 and D0-branes, and interpreting this in the context of
351: embedding IIA theory in M-theory, it was argued in \gova\ that
352: $F_g$ encodes the degeneracy of BPS spinning 5d black holes
353: corresponding to M2-branes wrapped over 2-cycles of the
354: Calabi-Yau. In particular, it was argued that the Kahler moduli
355: dependence of topological string captures the charge dependence of
356: these BPS states and the string coupling constant dependence of
357: the amplitudes captures the spin content of the BPS states.
358:
359: The starting point of the GV calculation \gova\ is a compactification
360: of M-theory on a Calabi-Yau 3-fold $X$ down to 5 dimensions. One is
361: interested in computing a thermal partition function of BPS-states and
362: hence the time is Euclidean and compactified on a $S^1$. The idea
363: behind the GV calculation is to identify this thermal circle with the
364: 11${}^{th}$ M-theory direction. The 5D angular momentum couples to the
365: field strength $F_{gph}$ of the graviphoton field. Since the
366: counting of BPS-states corresponds to an index, the result is
367: independent of the period $\beta$ of the Euclidean time-circle, that
368: plays the role of the IIA string coupling constant. In this way we can
369: relate the perturbative picture at $\b =0$ to the strong coupling
370: limit $\b = \infty$.
371:
372: In the limit $\beta\to 0$ the M2-branes are mapped on to IIA
373: fundamental strings. Due to the presence of the graviphoton field
374: strength the string world sheet calculation reduces precisely to
375: topological string theory, and counts the contribution of world sheet
376: instantons. The topological string coupling constant is equal to the
377: combination
378: $$
379: \lambda=\beta F_{gph},
380: $$
381: which is kept fixed in the $\beta \to 0$ limit. Hence the
382: topological string partition function that computes the BPS quantities
383: receives contributions of all genera and even non-perturbative
384: effects.
385:
386: Alternatively, in the $\beta\to\infty$ limit, the light objects are
387: the D2 and D0-branes, and their bound states. The free energy
388: $F(t,\lambda)$ of the type IIA string can thus be rewritten as a sum
389: of contributions of a Schwinger-type calculation for each of these
390: D2-D0 bound states. Here the D0-brane charge corresponds to the
391: momentum along the Euclidean time-circle, and hence has no direct
392: physical interpretation from the four-dimensional space-time
393: perspective. The computation of \gova\ is based on integrating out
394: these charged particles that become light in the strong coupling
395: limit, where the type IIA theory should be reformulated in
396: M-theory. Therefore this Schwinger calculation essentially takes place
397: in 5 dimensions. A note of caution that we will return to: the
398: D0-brane charge that is being summed over in these virtual loops has
399: an interpretation as the momentum along the thermal circle. The
400: degeneracies are independent of this D0-brane charge, which should
401: {\it not} be confused with the spin of the M2-branes.
402:
403:
404: \subsec{Five-dimensional BPS degeneracies}
405:
406: Consider a compactification of M-theory on a Calabi-Yau threefold $X$
407: down to 5 dimensions. We are interested in the BPS states in this
408: theory that are charged under the $U(1)$'s and correspond to M2-branes
409: wrapped over two-cycles of the Calabi-Yau. Let
410: %
411: $$Q\in H_2(X,{\bf Z})$$
412: %
413: denote a generic charge of such an M2-brane. Such states will also
414: carry spin. The spatial rotation group in 5 dimensions is
415: $SO(4)=SU(2)_L\times SU(2)_R$. We can label a BPS state, in addition
416: to its charge $Q$, by its spin. Let $(m_L,m_R)$ denote the spin
417: content $(2j^3_L,2 j^3_R)$ of the highest spin state in a given BPS
418: multiplet. Let $N_Q^{(m_L,m_R)}$ denote the number of such BPS
419: states. Define
420: %
421: \eqn\NQm{
422: N_Q^m=\sum_{m_R}(-1)^{m_R} N_Q^{(m_L,m_R)}.
423: }
424: %
425: This quantity is an invariant in the sense that, if one changes the
426: hypermultiplet moduli, this index does not change. The numbers
427: $N_Q^{(m_L,m_R)}$ can only change if the left and right spin contents
428: pair up, which leaves this index invariant since it only counts the
429: net number. Note that $N_Q^m$ may be a positive or a negative
430: integer. As discussed in \kkv\ the spin content of an M2-bound state is
431: related to the topology of the holomorphic curve in the CY. A single
432: M2-brane wrapping a curve of genus $r$ gives a particular (reducible)
433: representation with maximal spin $r$. For our purpose it is useful to
434: concentrate only on the $j_L^3$ quantum number. So, in principle the
435: numbers $N_Q^m$ can receive contributions of curves of arbitrary high
436: genus. Nevertheless, we will assume that $N_Q^m$ are finite integers.
437:
438: Now we are ready to rewrite the topological string amplitudes in terms
439: of $N_Q^m$ following \gova, see also \nv. The topological string
440: partition function, that has a perturbative expansion of the form
441: %%
442: \eqn\topstring{
443: Z(\lambda,t) = \exp \sum_{g\geq 0} \lambda^{2g-2} F_g(t),
444: }
445: %%
446: depends on the Kahler moduli $t \in H^2(X,{\bf R})$ as well as on the
447: topological string coupling constant $\lambda$. Then the result of
448: \gova\ is that this partition function has an alternative form
449: %
450: \eqn\mafo{
451: Z(\lambda, t)=\prod_{Q,m} \Bigl[\prod_{n\geq 0} (1-e^{\l(n+m)+ t \cdot Q})^n
452: \Bigr]^{N_Q^m}
453: }
454: %
455: It has the following interpretation in terms of a gas of M2-branes in
456: five dimensions: First, one notices that, after substituting
457: relation \NQm\ in the above formula, each spin content
458: $m_L,m_R$ contributes a factor
459: %
460: $$
461: \prod_{n_1,n_2\geq 0} \left(
462: 1-e^{\l_1(m_1+n_1+\half) } e^{\l_2(m_2+n_2+\half)+
463: t \cdot Q} \right)^{\pm 1}.
464: $$
465: %
466: Here the quantum numbers $m_1,m_2$ are given by
467: $$
468: m_1 = {1\over 2}(m_L + m_R),\qquad
469: m_2 = {1\over 2}(m_L-m_R),
470: $$
471: and we introduced additional parameters $\l_1,\l_2$ that are eventually set to
472: $\l_1=\l_2=\l$. The parameter $t$ acts as a potential that couples to the
473: charge $Q$ of the M2-brane. The $\pm 1$ in the exponent depends on
474: whether the (top element) of the BPS state is a boson or
475: fermion. Tracing back the derivation, one finds that the quantum
476: numbers $m_1,m_2$ refer to the intrinsic spins of the BPS-state.
477: The numbers $n_1,n_2$ indicate the orbital helicities in each of the
478: two planes of ${\bf R}^4$.
479:
480: More precisely, for each BPS state we get a field $\Phi(z_1,z_2)$
481: where $z_1,z_2$ denote the coordinates of ${\bf R}^4={\bf C}^2$.
482: Because of the BPS condition this field has a holomorphic expansion
483: %
484: $$
485: \Phi(z_1,z_2)=\sum_{n_1,n_2} \alpha_{n_1,n_2} z_1^{n_1}z_2^{n_2}.
486: $$
487: %
488: These holomorphic wave functions are the familiar lowest Landau levels
489: that describe a charged particle in a constant self-dual graviphoton
490: field. These wave functions are normalizable, since the proper
491: normalization includes a Gaussian factor. Therefore, these wave
492: functions are localized in the $\R^4$ plane.
493:
494: Each factor in the above partition function can now be viewed as
495: counting the second quantized states of the corresponding quantum
496: field $\Phi$, where the oscillators $\alpha_{n_1,n_2}$ could be
497: bosonic or fermionic, depending on the internal spin of the field
498: $\Phi$.
499:
500:
501: \newsec{D-brane bound states and DT invariants}
502:
503: Let us now turn to the counting of D-brane states in IIA string
504: theory. Consider a single D6-brane wrapped on the Calabi-Yau manifold
505: $X$. On the brane there is a maximally supersymmetric $6+1$
506: dimensional $U(1)$ gauge theory, that is naturally topologically
507: twisted. This gauge theory has non-trivial topological sectors that
508: correspond to bound states with lower-dimensional branes.
509:
510: The Witten index of this field theory,
511: $$
512: {\rm Tr}\left[(-1)^F e^{-\beta H} \right],
513: $$
514: is the partition function on $S^1 \times X$ and, within a given
515: topological sector, this should computes the number of D-brane bound
516: states. Mathematicians have considered the corresponding moduli spaces
517: and the associated invariants (roughly the Euler numbers of these
518: moduli spaces) are known as Donaldson-Thomas invariants \DT. It is
519: conjectured in \ionv\ that the quantum gauge theory exactly computes these
520: mathematical invariants.
521:
522: By taking the $\beta \to 0$ limit of the index we obtain a
523: six-dimensional $U(1)$ topological field theory, that was conjectured
524: to be the same as the topological string partition function
525: \refs{\ionv,\mnop}. This second conjecture was based on the discovery
526: of a deep relationship between the statistical mechanical model of
527: crystal melting and topological string amplitudes \orv . Notice that,
528: since we shrink the thermal circle, it is most natural to perform a
529: $T$-duality along that circle. In this way we obtain immediately a
530: Euclidean D5-brane wrapping the CY, whose world-volume theory gives
531: the six-dimensional abelian gauge theory that we consider here. (For
532: the case of the B-model, the relation with the world-volume theory of
533: the 5-brane was also made in \dvv.)
534:
535: By considering a non-trivial gauge bundle, the topological field
536: theory on the D6-brane naturally includes bound states with lower
537: dimensional branes. We will
538: label the charges of the D6-D4-D2-D0 system as
539: $$
540: (p_0,p,q,q_0) \in H^0(X) \oplus H^2(X) \oplus H^4(X) \oplus H^6(X).
541: $$
542: Mathematically, the possible D-brane charges are
543: given by the Chern classes of the corresponding gauge bundle $\cE$. In
544: the case of a $U(1)$ theory the bundle $\cE$ is of rank one and
545: necessarily has to be an ideal sheaf. The physical charges are given
546: by the Chern characters
547: %%
548: $$
549: ch_k(\cE) = {1\over k! (2\pi)^k} {\rm Tr} \,F^k
550: \in H^{ev}(X,{\bf Q}).
551: $$
552: %%
553: These Chern characters can be expressed in terms of the integer Chern
554: classes $c_i(\cE) \in H^{ev}(X,\Z)$. For the case of a single D6-brane
555: with $ch_0(\cE)=rk(\cE)=1$ the charges are given by
556: %%%
557: \eqn\charges{
558: \eqalign{
559: D6:\ \ & p_0 = rk =1, \cr
560: D4:\ \ & p = c_1, \cr
561: D2:\ \ & q = ch_2 = -c_2 + {1\over 2} c_1^2 - {1\over 48} p_1(X), \cr
562: D0:\ \ & q_0 = 2 ch_3 = c_3 - c_1 c_2 + {1\over 3} c_1^3
563: - {1\over 24} p_1(X) c_1. \cr
564: }}
565: %%%%%%%%
566: Here we have also included the gravitational correction coming from
567: the first Pontryagin class $p_1(X) \sim {\rm Tr}\, R^2$ of the CY
568: space.
569:
570: In the absense of D4-branes, the coupling to the D2 and D0-branes is
571: given by the action
572: %
573: $$
574: S =\lambda \int_X c_3(F)+\int k\wedge ch_2(F)
575: $$
576: %
577: where $[k]= t$ gives the Kahler class of the CY and $\l$ will be
578: identified with the topological string coupling.
579: The resulting partition function has an expansion in terms of the integer
580: Donaldson-Thomas invariants $DT_{q,q_0}$
581: $$
582: Z(\lambda,t) = \sum_{q,q_0} DT_{q,q_0} \ e^{\l q_0 +t \cdot q}
583: $$
584: The conjecture of \refs{\ionv,\mnop} is that this equals the topological
585: string partition \topstring. The aim of this paper is to explain this
586: relation by connecting directly the D-brane state counting to the
587: 5-dimensional BPS counting of \gova. An alternative physical
588: explanation for this conjecture, based on quantum gravitational foam
589: for the A-model topological string, was given in \ionv. This independent
590: explanation provides additional support for the arguments presented in the
591: following section.
592:
593: Although the DT invariants are usually defined in the case $c_1=0$,
594: {\it i.e.} without D4-branes, they can in principle also be determined
595: for the case $c_1 \not = 0$. This is a rather trivial generalization,
596: because a non-trivial $c_1$ does not change the nature of the
597: singularities of the bundle. This flux is carried by smooth line
598: bundles with non-trivial topologies, so intuitively it can be thought
599: of as being smeared out over the CY space. The invariants only depends
600: on the moduli space of the singularities, which are captured by the
601: Chern classes $c_2$ and $c_3$. So, turning on a D4-brane magnetic
602: charge
603: $$
604: p=c_1(\cE) \in H^4(X,\Z).
605: $$
606: does not change the value of the invariant, but it merely changes the
607: definition of what we call D2 and D0-brane charges $(q,q_0)$, as we
608: can read off from formula \charges,
609: %%%%
610: \eqn\chargeshift{
611: \eqalign{
612: q & \to q + {1\over 2} p^2,\cr
613: q_0 & \to q_0 - q\cdot p - {1\over 6} p^3 -{1\over 48} p_1(X) p. \cr
614: }}
615: %%%%
616: The fact that the D-brane degeneracy is independent of this shift,
617: can also be obtained by duality arguments \andy.
618:
619:
620: \newsec{D-brane degeneracies from a gas of M2-branes}
621:
622:
623: As we explained in section 2.1, the topological string partition
624: function has an interpretation in terms of a gas of five-dimensional
625: black holes. We now like to argue that this same partition function
626: also gives the degeneracies of the four-dimensional D-brane system of
627: section 3, and hence is related to Donaldson-Thomas
628: theory. Furthermore, instead of counting a gas of 5d BPS states, it
629: enumerates the states of a single BPS object in four dimensions.
630:
631: \subsec{Lift to M-theory}
632:
633: Let us start with the IIA D-brane system of the previous section, without
634: any D4-branes for the moment. The thermal partition function of this
635: system is computed in the ten-dimensional background
636: %%
637: \eqn\bg{
638: X \times \R^3 \times S^1,
639: }
640: %%%
641: with $X$ the Calabi-Yau three-fold and $S^1$ the thermal circle of
642: radius $\b$. In
643: the lift to M-theory the single D6-brane gets represented by a
644: $k=1$ Taub-NUT geometry. So, the eleven-dimensional background is
645: $$
646: X \times TN \times S^1,
647: $$
648: The D2 and D0-branes lift to M2-branes and KK-momenta.
649:
650: This is precisely the configuration used in \andy\ to connect
651: 4d and 5d BPS computations.
652: The Taub-NUT space can be considered as an $S^1$ fibration over $\R^3$,
653: where the circle shrinks to zero at the origin (the location of the
654: D6-brane) and attains a finite
655: radius $R$ at infinity. The metric can be written with $\chi \in S^1$,
656: and ${\vec x}\in \R^3$, as
657: %%
658: \eqn\tnmetric{
659: ds^2_{TN} = R^2 \left[ {1\over V}(d\chi
660: +
661: \vec{A} \cdot d{\vec x})^2 + V d{\vec x}^2 \right],
662: }
663: %%
664: with
665: $$
666: V = {1 + {1\over |{\vec x}|}}, \qquad \vec{\nabla}\times
667: \vec{A} = \vec{\nabla} V.
668: $$
669:
670: Since we now have an M-theory compactification with two (asymptotic)
671: circles, we can take either one of these as the eleventh
672: direction. The two inequivalent reductions to IIA string theory give
673: rise to the two interpretations of the partition function that we want
674: to relate in this paper. The first reduction along the TN circle gives
675: back the starting configuration of D-branes in the background \bg. The
676: second reduction uses the thermal circle and gives a $X \times TN$
677: geometry. We will argue that this leads to the GV interpretation of
678: the topological string partition function.
679:
680: Note that the exchange of the two circles can be done entirely within
681: the IIA theory. It is the well-known ``9-11 flip'' that is obtained as
682: the duality transformation $T\cdot S \cdot T$, with $T$ being the
683: $T$-duality on the remaining $S^1$ and $S$ the strong-weak coupling
684: duality of the IIB string. Under this transformation the D6-brane in
685: our starting configuration gets mapped to the TN geometry as
686: follows. The first $T$-duality turns it into a Euclidean D5 wrapping
687: the CY. The $S$-duality makes this into a NS5-brane, which then by the
688: second (transversal) $T$-duality becomes the Taub-NUT. Similarly, the
689: D2-branes become Euclidean D1-branes and subsequently fundamental
690: strings. Finally, the D0-branes become momentum modes along the $S^1$
691: of the TN fibration.
692:
693: The TN metric is a hyper-K\"ahler geometry with a $U(1) \times
694: SO(3)$ isometry group, that acts in the obvious way on the variables
695: $\chi$ and ${\vec x}$. Near the origin the geometry is smooth, and
696: locally given by $\R^4$. This is most easily seen by writing the
697: metric on $\R^3$ in spherical coordinates $(r,\theta,\phi)$. After
698: making the substitution $r=\rho^2$ and using that near the origin $V
699: \sim \rho^{-2}$ one recognizes the flat $\R^4$-metric in spherical
700: coordinates $(\rho,\chi,\theta,\phi)$, where the angular part is
701: represented by the Hopf fibration of $S^3$ over $S^2$: here $\chi$ is
702: the coordinate along the fiber, while $\theta$ and $\phi$ parametrize
703: the base. The isometries of the Taub-NUT geometry leave the origin
704: invariant and act as $SO(4)=SU(2)_L\times SU(2)_R$ rotations in
705: $\R^4$. The $SO(3)$ rotations of the $S^2$-base correspond to
706: $SU(2)_R$, while the $U(1)$ that shifts $\chi$ is contained in
707: $SU(2)_L$. This in particular implies that the $U(1)$ quantum number
708: is given by an angular momentum quantum number at the origin. This
709: suggest the following re-interpretation of the D0-brane quantum
710: number. It equals the KK momentum at infinity, but in the interior it
711: is identified with the {\it total} angular momentum. This receives
712: contributions from both the intrinsic spin $m$ of the M2-branes and
713: from their orbital angular momentum $n$. This identification of the
714: 4d D0-brane charge with the 5d angular momentum was one of the
715: essential observations of \andy.
716:
717: \subsec{Physical derivation}
718:
719:
720: \ifig\tn{tn.eps}{100}{
721: The small and large radius limits of the Taub-NUT geometry interpolate
722: between $(a)$ a bound state of 4d D-branes and $(b)$ a free gas of spinning
723: 5d M2-branes.}
724:
725: From an M-theory perspective the perturbative gauge theory description
726: of the D-brane system naturally arises in the limit of small TN radius
727: $R$. In the $R\to 0$ limit the M2-branes are clustered together in
728: the origin and form a single bound state from the IIA point of view as
729: illustrated in \tn$(a)$. This is in accordance with the well-known fact
730: that codimension-four branes (in this case D6 and D2) can form a bound
731: state.
732:
733:
734: The idea is now to consider the opposite limit of large radius $R$,
735: where we make use of the fact that the M-theory partition function
736: with supersymmetric boundary conditions is independent of $R$. As we
737: can see from the metric \tnmetric, changing the radius is equivalent
738: to rescaling the entire geometry. So, we are effectively zooming in to
739: the center of the Taub-NUT, see \tn$(b)$. In the $R\to\infty$ limit
740: the M2-branes disperse and form a free gas of charged, spinning
741: BPS-particles in $\R^4$. This is where we make contact with the GV
742: computation.
743:
744: We are now in a position to give a four-dimensional interpretation of
745: the GV partition function. First of all, the D2-charge $q$ of the 4d
746: bound state is given by the sum of the charges $Q_i$ of the individual
747: M2-branes in the 5d gas
748: $$
749: q = \sum_i Q_i.
750: $$
751: Similarly, the D0-brane charge $q_0$ is equal to sum of the individual
752: angular momenta $\ell_i$, which for each M2-brane is the sum of the
753: internal and orbital spin quantum numbers $m_i$ and $n_i$
754: $$
755: q_0 = \sum_i \ell_i,\qquad \ell_i = m_i + n_i.
756: $$
757: Therefore, the GV partition function \mafo, when interpreted as
758: counting the D-brane degeneracies, should be written as
759: %
760: \eqn\dbrane{
761: Z(\lambda, t)=\prod_{Q,\ell} (1-e^{\l \ell + t \cdot
762: Q})^{c_Q^\ell} }
763: %
764: where the multiplicities
765: $$
766: c_Q^\ell=\sum_{n=0}^\ell n N_Q^{\ell-n}
767: $$
768: can in some way be regarded as counting the number of ``irreducible''
769: D2-D0 states inside the D6. In fact, this formula is a generalization
770: of a similar result obtained for the counting of D4-D0 bound states on
771: a 4-manifold $M$. As has been shown in \vw\ for a single
772: D4-brane with $N$ D0-branes the number of bound states is given by the
773: Euler number of the instanton moduli space (or more precisely in this
774: case the Hilbert scheme of points on $M$). These
775: degeneracies $d_N$ are encoded in the generating function
776: $$
777: \sum_N d_N e^{tN} = \prod_{k>0}\left(1 - e^{tk}\right)^{-\chi(M)}
778: $$
779: which can be obtained as a special case of \dbrane\ for the case
780: of $X = M \times T^2$.
781:
782: As explained, the D-brane partition function should also be captured
783: by the Donaldson-Thomas theory. Combining the above ingredients, we
784: see that the conjectured relation between the GV-invariants and the
785: DT-invariants is a direct consequence of the connection between 4d and
786: 5d black holes of \andy. Conversely, the independent mathematical
787: evidence for the relation between these two mathematical invariants
788: should be regarded as supporting the 4d-5d connection.
789:
790:
791: \subsec{Inclusion of D4-branes}
792:
793: We have seen in section 3 it is very natural to include also D4-branes
794: in the gauge theory. When we appropriately shift the charges $q$ and
795: $q_0$ this will not change the degeneracies. This is also clear from
796: the M-theory perspective. Here the D4-branes will be represented as
797: M5-branes that wrap a 4-cycle in the CY space. They are therefore
798: string-like objects in the Taub-NUT geometry that wind around the
799: $S^1$ fibers. Since that circle is contractible in the core of the TN,
800: these M5-branes are not supported by the topology and can be
801: contracted to zero size leaving behind a flux for the 4-form field
802: strength $G$. In the scaling limit where we zoom in to the interior of
803: the Taub-NUT geometry, this flux becomes diluted and has no effect on
804: the microscopic counting of the black hole states. However, the
805: definition of the asymptotic charges $q$ and $q_0$ gets modified, due
806: to the presence of the Chern-Simons term in the M-theory effective
807: action. This analysis has been performed in \andy, where exactly the
808: relations \chargeshift\ were found. It is gratifying that there is
809: such a direct relation between the quantities in the gauge theory and
810: the effective supergravity. Perhaps this observation can eventually
811: be extended to a full gauge theory/gravity correspondence.
812:
813:
814: \newsec{Additional Remarks}
815:
816: The connection between the topological string and the counting of
817: D-brane bound states that we have discussed here is different from
818: that of the OSV conjecture \osv\ . First of all, the present relation
819: is exact, but holds only for a single D6-brane. On the other hand, in
820: the OSV conjecture the relation with the topological string only
821: arises asymptotically in the limit for large magnetic charges, {\it
822: i.e.} D4 and D6. This regime is outside the scope of this present
823: paper. Furthermore, in the OSV conjecture the square of the
824: topological string partition function appears. Finally, the role of
825: the topological string coupling constant $\l$ and moduli $t$ is
826: different.
827:
828: Our M-theory derivation of the topological string duality makes use of
829: two different reductions to the IIA string: one along the circle in
830: Taub-NUT, one along the thermal circle. The GV relation with the
831: topological string \gova\ arises from the latter reduction. In this
832: paper the first was used to obtain the D-brane system. But also in
833: this case a role is played by the topological string, namely along the
834: lines of the OSV conjecture. Hence, these two occurrences of the
835: topological string are related by the $TST$ duality chain that
836: interchanges the two circles. This role of S-duality has been
837: anticipated in \refs{\nov,\osv} and was a crucial ingredient in the
838: recent paper \gsy.
839:
840: To use our exact counting formula for the computation of the entropy
841: of 4d black holes, we have to extend the present analysis from the
842: case of a single D6-brane to a large number of D6-branes. One expects
843: that the full answer is invariant under the $S$-duality and that
844: leading growth of the number of states is captured by the OSV formula,
845: that appears after the duality map, very much like the Cardy formula
846: in conformal field theory. This picture is very analogous to the
847: ``black hole Farey tail'' of \dmmv, and is also suggested by the
848: analysis of the $(0,4)$ CFT in \gsy. Semi-classically, the two
849: appearances of the topological string are separated by a
850: Hawking-Page-type phase transition in the dual gravitational system
851: from a thermal gas to a black hole \tobe.
852:
853: \medskip
854: \centerline{\bf Acknowledgments}
855:
856: We would like to thank M. Cheng, J. Manschot, A. Strominger for
857: fruitful discussions. R.D. and E.V. wish to thank the Harvard Physics
858: Department for kind hospitality. The research of R.D. and E.V. was
859: supported by a NWO Spinoza grant and the FOM program {\it String
860: Theory and Quantum Gravity}. The research of C.V. was supported in
861: part by NSF grants PHY-0244821 and DMS-0244464.
862:
863: \listrefs
864:
865: \end
866:
867: