1: \documentclass[12pt]{article}
2: \usepackage{epsf}
3: \usepackage{amsmath}
4: \usepackage{amssymb}
5: \usepackage{amsfonts}
6:
7: % titlepage defs
8: %
9: \newcommand{\uiaddress}
10: {{\small\it Department of Physics, University of Illinois, Urbana, IL 61801}}
11: \newcommand{\ucaddress}
12: {{\small\it Enrico Fermi Institute, University of Chicago, Chicago, IL 60637}}
13: \newcommand{\email}[1]{\thanks{e-mail: \tt#1}}
14: \newcommand{\preprint}
15: {\begin{flushright}\begin{small}
16: ILL-(TH)-00-07\\ EFI-2000-21\\ hep-th/0006168\\
17: \end{small}\end{flushright}
18: }
19:
20: % shortcuts
21: \newcommand{\eqr}[1]{(\ref{eq:#1})}
22: \newcommand{\tr}{{\rm tr}}
23: \newcommand{\qcomm}[2]{\left[#1,#2\right]_q}
24: \newcommand{\vev}[1]{\langle#1\rangle}
25:
26: % new letters
27: \newcommand{\field}[1]{\mathbb{#1}}
28: \newcommand{\ZZ}{{\field Z}}
29: \newcommand{\CC}{{\field C}}
30: \newcommand{\RR}{{\field R}}
31:
32: % figures without captions (give full figfile name,
33: % followed by xsize in cm
34: \newcommand{\littlefig}[2]{
35: \epsfxsize=#2cm
36: \epsfbox{#1}
37: }
38: % figures (give full figfile name
39: % followed by xsize in cm, followed by caption.
40: \newcommand{\myfig}[3]{\begin{figure}[ht]
41: \begin{center}
42: \leavevmode
43: \epsfxsize=#2cm
44: \epsfbox{#1}
45: \end{center}
46: \caption{#3}
47: \label{fig:#1}
48: \end{figure}}
49:
50: \newcommand{\ack}[1]{Pfft!: {\bf #1}}
51:
52: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
53:
54: \begin{document}
55: %\setlength{\textheight}{8.8in}
56:
57: \begin{titlepage}
58: \title{
59: \preprint\vspace{1.5cm}
60: Non-Commutative Moduli Spaces,\\ Dielectric Tori and T-duality
61: }
62: \author{
63: David Berenstein,\email{berenste@pobox.hep.uiuc.edu}
64: Vishnu Jejjala,\email{vishnu@pobox.hep.uiuc.edu}\\
65: \uiaddress\\
66: and\\
67: Robert G. Leigh\email{rgleigh@uiuc.edu}\\
68: \ucaddress\\ {\small \it and} \uiaddress
69: \\
70: }
71: \maketitle
72:
73: \begin{abstract}
74: We review and extend recent work on the application of the non-commutative
75: geometric framework to an interpretation of the moduli space of vacua of
76: certain deformations of $N=4$ super Yang-Mills theories. We present a
77: simple worldsheet calculation that reproduces the field theory results
78: and sheds some light on the dynamics of the D-brane bubbles. Different
79: regions of moduli space are associated with $D5$-branes of various topologies;
80: singularities in the moduli space are associated with topology change.
81: $T$-duality on toroidal topologies maps between mirror string realizations of
82: the field theory.
83: \end{abstract}
84: \end{titlepage}
85:
86: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
87:
88: \section{Introduction}\label{sec:intro}
89:
90: The $N=4$ super Yang-Mills theories in four dimensions possess\cite{LS}
91: exactly marginal deformations
92: \begin{equation}\label{eq:margsup}
93: W=\tr \left( \phi_1\phi_2\phi_3-q\phi_2\phi_1\phi_3\right)+
94: \lambda\tr \left(\phi_1^3+\phi_2^3+\phi_3^3\right)
95: \end{equation}
96: The $q$-deformation is of particular interest. In its presence, we find
97: $F$-term constraints
98: \begin{equation}\label{eq:constr}
99: \qcomm{\phi_1}{\phi_2}=0,\ \ \ \
100: \qcomm{\phi_2}{\phi_3}=0,\ \ \ \
101: \qcomm{\phi_3}{\phi_1}=0
102: \end{equation}
103: The solutions of these equations determine the moduli space of
104: supersymmetric vacua of the theory. Eqs. \eqr{constr}
105: give relations in the algebra of $N\times N$ matrices. This algebra
106: is non-commutative, and thus the moduli space has a non-commutative
107: geometric interpretation.
108: Points in this non-commutative geometry are defined to be irreducible
109: representations of the algebra \eqr{constr}, up to equivalences (amounting
110: to gauge transformations on the brane worldvolume). This definition
111: is in accord\footnote{In particular, each irreducible representation
112: ${\cal A}\stackrel{\mu}{\to} M_N(\CC)\to0$ provides a maximal ideal, $ker\ \mu$.
113: Maximal ideals, as in algebraic geometry, correspond to points.}
114: with the usual definitions in non-commutative geometry.
115: This structure persists in general, for the superpotential
116: \eqr{margsup}, but also for arbitrary relevant single-trace perturbations as well.
117:
118: In Ref. \cite{BJL1}, we presented a discussion of non-commutative
119: moduli space for a variety of deformations.
120: Moduli space is built
121: of direct sums of irreducible representations of the algebra, and
122: given a complete classification of the irreducible representations,
123: it is possible to describe the moduli space as a non-commutative
124: version of the symmetric product.
125:
126: As well, we noted that
127: these theories may be obtained from two distinct classes of string
128: geometries. First, one may obtain these models via $\ZZ_m\times\ZZ_m$
129: orbifolds with discrete torsion\cite{D,DF,BL}, which is related to
130: $q$, for $q^m = 1$. Second, the models are obtained
131: through deformations of the
132: near-horizon geometry, $AdS_5\times S^5$, dual to the $N=4$ field
133: theory. These two models are related by mirror symmetry, which may
134: be traced to a T-duality on torus fibrations of the 5-sphere of the
135: near-horizon geometry. The significance of this torus manifests
136: itself when one considers $D3$-branes in background $NS$ and $RR$
137: fields. We will describe these phenomena in detail below.
138:
139: The principal effect here is that $D3$-branes in these backgrounds
140: become $D5$-branes of non-trivial topology. This phenomenon has
141: appeared in several guises over the past few years. First, it
142: appeared in Ref. \cite{BFSS} wherein membranes of Matrix theory were
143: understood to be built from $D0$-branes. Spherical membranes in Matrix
144: theory were considered in Ref. \cite{KTay,BC}.
145: More
146: recently, Myers\cite{RM} has shown that $Dp$-branes in $RR$
147: backgrounds in string theory carry a $D(p+2)$-brane dipole
148: moment, and thus may be thought of as $D(p+2)$-branes with
149: topology $\RR^p\times S^2$. This effect comes into play, for example,
150: for relevant
151: deformations\cite{PS} within the $AdS/CFT$ correspondence.
152: One of the difficult issues is to
153: understand this phenomenon for a single $Dp$-brane, as then the
154: non-commutative geometry that one obtains is trivial, and one can
155: just as well regard the brane as pointlike, as it seems to have no structure.
156: Moreover, in this regime the $\alpha'$
157: corrections are not suppressed by $1/N$, and the DBI action is also
158: unreliable.
159:
160: One purpose of this letter is to try to understand this phenomenon of
161: brane bubbling into spheres for a
162: single $Dp$-brane by using worldsheet methods.
163: We begin with a detailed account of the non-commutative
164: moduli space of vacua. We then consider the AdS/CFT realization of
165: these field theories. Here we are interested in a probe calculation,
166: and so small deformations in the geometry can be taken into account
167: systematically using worldsheet methods.
168: This fills gaps in the discussion of Ref. \cite{BJL1}:
169: the existence of certain massless modes can be proven and is important
170: for the realization of the non-commutative moduli space.
171: $Dp$-branes may be thought
172: of as $D(p+2)$-branes of various topologies. There is a T-dual orbifold
173: description; singularities in the latter description, such as the fractionation
174: of branes, corresponds directly to the degenerations of these
175: topologies.
176:
177: \section{General Features of the Moduli Space}
178:
179: Let us begin with a description of the moduli space for the
180: $q$-deformed theory, where $q^n=1$ for some integer $n$. When
181: $q$ is such a root of unity, there are branches of moduli space
182: which do not otherwise exist.
183: First, we note that the center of the algebra
184: is generated by the elements $x=\phi_1^n$, $y=\phi_2^n$, $z=\phi_3^n$
185: and $w=\phi_1\phi_2\phi_3$, and that these satisfy the matrix
186: relation
187: \begin{equation}
188: (-w)^n+xyz=0
189: \end{equation}
190: Thus the center of the algebra reproduces the commutative moduli
191: space, the orbifold upon which we expect point-like
192: $D$-branes to propagate. We will use orbifold language to describe
193: the moduli space here, but there is also a mirror description
194: which we will give details of in the next section.
195: The orbifold space has singularities along
196: complex lines where two of the $x,y,z$ vanish.
197: \myfig{modspq.eps}{4}{Moduli space of $q$-deformed theory.}
198:
199: Next, we look for representations of the algebra
200: \eqr{constr}. In the bulk of the orbifold, where at least
201: two of the $x,y,z$ do not vanish, we find an $n$-dimensional\footnote{We
202: assume that $n\le N$.} representation\footnote{This representation has
203: been chosen so that $\phi_3$ is diagonal at the singularity $x=y=0$. At
204: other singularities, one may parameterize in a different way. This may
205: be thought of in terms of patches.} $R(a,b,c)$
206: \begin{eqnarray}
207: \phi_1&=&aQ\\
208: \phi_2&=&bQ^{-1}P^{-1}\\
209: \phi_3&=&cP
210: \end{eqnarray}
211: where $a,b,c$ are arbitrary complex numbers and $P,Q$
212: satisfy $\qcomm{P}{Q}=0$. An explicit representation is
213: \begin{equation}\label{eq:PQ}
214: P = \begin{pmatrix}1&0&0 &\dots& 0 \\ 0& q & 0
215: & \dots&0 \\ 0& 0& q^2&\dots&0 \\
216: \vdots&\vdots&\vdots& \ddots&\vdots \\
217: 0&0&0&\dots&
218: q^{n-1}\end{pmatrix},\ \ \ \
219: Q = \begin{pmatrix}0&0& \dots&0&1\\
220: 1&0&\dots&0&0 \\
221: 0&1&\dots&0&0 \\
222: \vdots&\vdots&\vdots&\ddots&\vdots\\
223: 0&0&\dots&1&0
224: \end{pmatrix}
225: \end{equation}
226:
227: The individual (diagonal) elements of $P$ represent fractional branes, as
228: can be seen by going to the singularities. Indeed, at
229: for example $x=y=0$, we find $n$ one dimensional representations
230: $R(0,0,c)$, $R(0,0,qc)$ $\ldots,$ $R(0,0,q^{n-1}c)$ which are all
231: distinct. As one approaches the singularity from the bulk,
232: the bulk representation
233: $R(a,b,c)$ becomes reducible, and decomposes into the direct sum
234: \begin{equation}\label{eq:decompose}
235: \lim_{a, b\to 0} R(a,b,c) =
236: R(0,0,c)\oplus R(0,0,qc)\oplus\dots R(0,0,q^{n-1}c)
237: \end{equation}
238: For $n<N$, we can build representations by composing these $n$-dimensional
239: representations, together perhaps with one-dimensional
240: representations associated to the fractional branes.
241: When $q$ is not a root of unity, the representations
242: that we have given do not exist, and the moduli space is reduced
243: considerably, so only the fractional branes survive.
244:
245: The decomposition (\ref{eq:decompose}) may be represented graphically as
246: in Fig. \ref{fig:quiver}. Note that this diagram is precisely the quiver
247: diagram for the local singularity. Each node in the quiver represents one
248: of the one-dimensional representations, and the lines connecting them represent
249: the vevs of fields which vanish at the singularity.
250: \myfig{quiver.epsf}{4}{Quiver diagram for reducible representations.}\label{fig:quiver}
251:
252:
253: \subsection{Relevant Deformations}
254:
255: It is interesting to see what happens to this moduli space when
256: relevant deformations are added. We consider first the case of
257: a rank one mass term $W=\frac{1}{2}m\phi_3^2$. When $q=1$, it is
258: clear what happens: at scales below $m$, we integrate out the
259: field $\phi_3$, and the dimension of the moduli space is reduced
260: from $3N$ to $2N$. This is also true for $q=-1$.
261: For $q^n=1, n>2$, the mass term does not lift
262: the moduli space in this way. Instead, it deforms it to\footnote{The
263: Casimir $w$ is shifted to
264: $w=\phi_1\phi_2\phi_3+t\phi_3^2$.}
265: \begin{equation}\label{eq:onemassv}
266: (-w)^n+xyz=-t^n z^2
267: \end{equation}
268: where $t=m/(1-q^2)$.
269: The representation $R(a,b,c)$ is also suitably deformed
270: to
271: \begin{eqnarray}
272: \phi_1&=&aQ\\
273: \phi_2&=&bQ^{-1}P^{-1}+dQ^{-1}P\\
274: \phi_3&=&cP
275: \end{eqnarray}
276: where $ad(q^2-1)=mc$. Thus, although $\phi_3$ is determined,
277: there is a third independent complex parameter -- the moduli
278: space has not decreased in dimension by the addition of the
279: mass term.
280: Note that one can solve for $\phi_3$ by the $F$-term
281: equations and thus ``integrate'' it out of the superpotential.
282: However, this procedure is misleading in this case, because
283: $\phi_1,\phi_2$ have more degrees of freedom than one na\"\i vely
284: expects.
285:
286: The mass term does resolve one of the singular complex lines;
287: indeed, the variety \eqr{onemassv} is singular at $x=w=z=0$ and
288: $y=w=z=0$. Along the line $x=y=0$, the representation is no
289: longer reducible, as it was for $m=0$.
290: \myfig{modspqm.eps}{4}{Moduli space of $q$-deformed theory, with rank one
291: mass.}
292:
293:
294: One may also consider higher rank mass terms. One such rank three
295: mass term may be written (when $q\neq1$), through a
296: holomorphic field
297: redefinition, as
298: \begin{equation}
299: W=\tr\left( \phi_1\phi_2\phi_3 - q\phi_2\phi_1\phi_3 +
300: \frac m2\phi_3^2+\zeta_3\phi_3\right)
301: \end{equation}
302: This further deforms the Casimir $w$, resulting in a deformed
303: commutative geometry
304: \begin{equation}
305: xyz +(-w)^n = - t^n z^2+ t_\zeta^n z
306: \end{equation}
307: where $t_\zeta=\frac{\zeta_3}{q-1}$. Again, the
308: irreducible representation
309: has three independent complex parameters
310: \begin{eqnarray}
311: \phi_1&=&aQ\\
312: \phi_2&=&bQ^{-1}P^{-1}+dQ^{-1}P+eQ^{-1}\\
313: \phi_3&=&cP
314: \end{eqnarray}
315: where $ad(q^2-1)=mc$ and $ae(q-1)=\zeta_3$.
316: The two complex lines of singularities that were present for $\zeta_3=0$
317: are smoothed out to the hyperboloid
318: \begin{equation}\label{eq:sing3m}
319: xy=t_\zeta^n,\ \ \ z=w=0
320: \end{equation}
321: Along this singularity, one can take the $n$th root of eq.\eqr{sing3m},
322: and obtain one-dimensional representations (corresponding to fractional
323: branes).
324: \myfig{modspqmz.eps}{5}{Moduli space of $q$-deformed theory, with rank three
325: mass.}
326:
327: In addition, there are isolated representations that are not
328: covered by the above solution. For example, there are representations
329: for which $\phi_1^n=0$ non-trivially. This can happen for nilpotent $\phi_1$
330: where $\phi_1^k=0$ for any $1\le k\le n$. For $k<n$, these representations may be
331: thought of as $q$-deformed $SL(2)$ representations, as we may bring
332: the algebra to the form (for $q\neq 0,1$).
333: \begin{eqnarray}
334: \qcomm{A_+}{A_-}&=&2A_0, \label{eq:algo}\\
335: \qcomm{A_0}{A_+}&=&A_+,\\
336: \qcomm{A_-}{A_0}&=&A_-\label{eq:algt}
337: \end{eqnarray}
338: Each of these $n-1$ special branches occur at a fixed
339: value of $z$ along $x=y=0$.
340:
341: The rank three mass deformation that we have considered here is special,
342: in that it is equivalent to mass eigenvalues $(m,-m,m')$. Other rank
343: three mass terms lead to algebras which are different from
344: \eqr{algo}--\eqr{algt} as one cannot diagonalize the mass terms without
345: altering the form of the $q$-commutator. Nevertheless, the physics is
346: very similar\cite{BJL1} to the case discussed here.
347:
348:
349: \section{Spheres and Tori}
350:
351: Next, we consider the moduli space of these field theories by
352: looking at the dual near-horizon string theory. The orbifold
353: field theory has a description in terms of the near-horizon
354: geometry $AdS_5\times S_5/\Gamma$ with discrete torsion encoded
355: in the boundary conditions of massless twisted states. Here we
356: are interested instead with the deformation of the $N=4$ theory.
357: In the AdS/CFT correspondence\cite{M,W,GKP} we are instructed to
358: interpret deformations of a superpotential in terms of
359: non-normalizable modes of supergravity states. The $q$-deformation
360: is induced\cite{KRN} by elements of the ${\bf 45}$ of $SU(4)_R$ which appears
361: as the second harmonic of the anti-symmetric tensor modes,
362: $G_{(3)}=F_{(3)}-\tau H_{(3)}$.
363: As the deformation
364: preserves conformal invariance (indeed, $G_{(3)}$ is independent of
365: the $AdS_5$ radial coordinate), the geometry is of the form
366: $AdS_5\times {\cal N}$ for some ${\cal N}$.
367:
368: As we have reviewed earlier, $D$-branes in background $RR$ fields
369: attain multipole moments of branes of higher dimension.
370: Fractional $D3$-branes should be thought of as
371: $D5$-branes of topology $\RR^4\times S^2$, oriented in a certain
372: way.
373: There are actually two effects in the present case.
374: First, the 5-brane is electrically charged
375: with respect to $F_{(7)}\sim *_4 \tilde F_{(3)}$,
376: which has support on $\RR^4\times D^3$,
377: where $D^3$ is bounded by the $S^2$. This field is responsible
378: for the di-electric effect. Secondly, there is a
379: background $H_{(3)}^{NS}$. This is an $AdS_5$ scalar, and so has all
380: three indices along ${\cal N}$. Thus, $H_{(3)}$ contributes to the energy
381: of a 5-brane if the $D^3$ has components along ${\cal N}$. Taking both
382: fields into account, we then expect that the radius of the disc $D^3$
383: is oriented partially in the $AdS_5$ direction, and partially along
384: ${\cal N}$. As shown in Ref. \cite{BJL1}, energy considerations
385: from the DBI action suggest, when the deformation is small,
386: that the 2-sphere is of size
387: \begin{equation}
388: \vev{r}\sim (c_{NS}+g_{str}c_R)(4\pi^2\alpha')^2
389: \end{equation}
390: where $\int_{D^3}H_{(3)}=c_{NS}r^3$, $\int_{D^3}\tilde F_{(3)}=c_{R}r^3$.
391: Thus the 2-spheres are expected to be of string size, and one should be hesitant
392: in accepting the geometric picture described here. Nevertheless, we will continue
393: with this way of speaking, as it provides useful intuition for the applications
394: studied here.
395:
396: We are interested in the new branches of moduli space which appear
397: near the singularities. Near such a singularity, when $q^n=1$, the
398: background which corresponds to the
399: $q$-deformation consists of $H_{(3)}$ only.
400: As a result, the di-electric
401: effect vanishes due to the orientation of the background field.
402:
403: Small $D5$-branes with spherical topology correspond to fractional branes in
404: the orbifold dual. If the spheres intersect each other, then
405: semi-classically we expect that there are massless string modes stretching
406: between
407: them. However, the radii of the 2-spheres are small in string units, and
408: thus this intuition may not necessarily be reliable. Moreover, two spheres
409: coming
410: together have locally opposite orientations, and thus correspond to a
411: brane-antibrane system where one expects complicated dynamics.
412: In the following subsection,
413: we will show using a worldsheet computation that massless modes are
414: present if we
415: tune the separation of $D3$-branes appropriately. This occurs at finite
416: separation in the presence of a non-zero $G_{(3)}$ field.
417: Finally, in Section 4, we will argue that the
418: existence of these massless modes gives
419: full consistency with field theory expectations.
420:
421: \subsection{Massless States}
422:
423: We will demonstrate the existence of massless states between 5-brane spheres by
424: returning to the description in terms of $D3$-branes in weak backgrounds.
425: In particular, we will show that massless states arise when the $D3$-branes are
426: held at fixed non-zero separation. This will be done at lowest order in the
427: background expansion, but it is clear that the massless states persist in
428: stronger backgrounds.
429:
430: Note that, from the field theory perspective, we expect to see
431: new massless modes when the branes are at special
432: locations. Indeed, from the $q$-deformed superpotential,
433: there are
434: off-diagonal masses (for $\phi_3={\rm diag}(a_1,a_2)$)
435: proportional to $a_1-qa_2$ and $a_2-qa_1$. Thus, massless states are
436: present for $|a_1|=|a_2|$ if we have
437: $|q|=1$. It is only when $q$ is a root of unity that the new branch in moduli
438: space can open up, but the massless states between two $D$-branes persist.
439:
440: In Ref. \cite{BJL1}, we
441: simply assumed that these massless states are present because of field theory
442: considerations.
443: There are several ways in which this effect may
444: be seen from worldsheet computations. Here, we will
445: consider the equation of motion
446: of an open-string fermion. As is well-known, the Dirac equation is obtained
447: by integrating the BRST current around the vertex operator. In this case,
448: the background makes its presence felt through contact terms with the
449: BRST operator.\myfig{brstloop.epsf}{3}{Disc diagram for fermion EOM.}
450: At lowest order in the $RR$ background,
451: the contact terms have been worked out in Ref. \cite{LBQ,LB2};
452: the $NSNS$ contact
453: terms are standard. We will denote the separation of the branes by a
454: vector $a^\mu$, of length $a$ and direction orthogonal to the branes. There is
455: of course a linear contribution in $a^\mu$ to the mass of the stretched string
456: state. The Dirac equation then takes the form
457: \begin{equation}\label{eq:Dirac}
458: \left(i\nabla\cdot\Gamma+\frac{1}{2\pi\alpha'}a\cdot\Gamma+
459: H_{\mu\nu\rho}\Gamma^{\mu\nu\rho}
460: +gF_{\mu\nu\rho}\Gamma_{||}\Gamma^{\mu\nu\rho}\right)\psi=0
461: \end{equation}
462: The fields $H_{(3)}$ and $F_{(3)}$ have identical orientations in the
463: background corresponding to the $q$-deformation. In an appropriate
464: coordinate system near the singularity
465: $x=y=0$, $G_{(3)}$ is oriented along $d\theta\wedge
466: (d\bar x \wedge dx-d\bar y\wedge dy)$ and $H_{(3)}/g$ ($F_{(3)}$)
467: is the imaginary (real) part. It is then easy to see that
468: the $\Gamma$-matrix structure in eq. \eqr{Dirac} has half of its
469: eigenvalues equal to zero when $a=0$. Also, by aligning $a^\mu$
470: appropriately, here along $z$, there are zero eigenvalues for a
471: particular value of $|a|$ proportional to the field strength. This
472: corresponds to a chiral fermion in four dimensions. The value of
473: $|a|$ at which this occurs agrees with the Born-Infeld calculation
474: presented above (up to vertex normalizations that we have not considered
475: carefully).
476:
477: \subsection{Bulk branes and Tori}
478:
479: Given the presence of massless modes at the intersections of 2-spheres,
480: it is natural to ask under what conditions one can turn on vevs for these
481: modes. These vevs would have an interpretation in terms of smoothing the
482: singularity of joined spheres. However, by consideration of appropriate
483: string diagrams (or the equivalent field theory superpotential), we see
484: that there is a potential preventing this condensation in general, consistent
485: with supersymmetry. There is however a special configuration, corresponding
486: to a linear combination of blowup modes, which evades this potential. This
487: occurs when we have $n$ 2-spheres joined and wrapping around the 5-sphere,
488: as is suggested by Fig. \ref{fig:spheres.epsf}.
489: \myfig{spheres.epsf}{5}{The $n$-pinched torus as a string of fractional
490: 5-brane spheres.} This should be compared to the quiver diagram of
491: Fig. \ref{fig:quiver}. Turning on the vev referred to above corresponds to
492: smoothing the structure in Fig. \ref{fig:spheres.epsf} out to a 2-torus, and is equivalent, in the
493: orbifold language, to the process
494: of moving a brane off of the singularity into the bulk. Thus a bulk
495: 3-brane should be thought of as a $D5$-brane with topology $\RR^4\times T^2$.
496: The space ${\cal N}$ (the deformed $S^5$) should then be thought of
497: topologically in terms of a torus fibration over $S^3$.
498: Large tori are governed
499: by the Born-Infeld action and are extrema of this action.
500: Indeed, one can show that near the degeneration, the Born-Infeld action is
501: independent of the size of the torus; to see this, one needs only the first-order
502: analysis.
503:
504: This description is related to the orbifold theory
505: by T-duality.
506: The T-duality extends in a natural way to non-commutative geometry.
507: On one hand,
508: we are instructed to think of an irreducible representation of the
509: non-commutative
510: algebra as a point in a non-commutative (moduli) space\cite{BJL1}. On the other
511: hand, we often think of this very same algebra as that of the non-commutative
512: torus, as in \cite{BFSS}.
513: There is no tension between these two concepts;
514: they are just mirror to each other.
515:
516: \subsection{Degenerations}
517:
518: To further explore the duality between the two descriptions,
519: let us consider from a geometric standpoint
520: the various degenerations of the toroidal 5-branes, and
521: how these processes are related to singularities in the orbifold
522: moduli space. In the massless case, as we have stated, the degenerations
523: in the moduli space correspond directly to brane fractionation.
524:
525: With a rank one mass term, one of the singular lines is removed; algebraically,
526: the representation of the algebra is no longer reducible there. Topologically,
527: the torus pinches at those points in the moduli space, but only once, so
528: no moduli are available
529: to separate the surface.
530: On a torus fibration, this fiber is singular and becomes
531: a sphere with two points identified.
532:
533: With a rank three mass term, the singularity is a hyperboloid,
534: and again, along this locus, the torus degenerates to a collection of spheres.
535: When a brane approaches the locus of the special isolated
536: branches (at fixed $z$, $x=y=0$), it may be able to split
537: into different localized branes, although we have not
538: been able to check this assertion precisely. In this case, we would get a
539: degeneration into a collection of spheres, but no new moduli.
540:
541: One can in fact interpret the matrix representations topologically. Diagonal
542: entries, such as in the matrix $P$, correspond to spheres, while non-zero
543: off-diagonal elements can be thought of in terms of a tube joining different
544: spheres. The genus of the resulting Riemann surface is encoded in the
545: number of off-diagonal elements of the generators. Degenerations may be
546: understood in this language in terms of the vanishing of matrix elements.
547:
548: \section{Final remarks and outlook}
549:
550: In this letter we have seen various features of the interactions between
551: $D$-branes and weakly deformed backgrounds. The semi-classical
552: configuration of a $D(p+2)$-brane wrapping an $S^2$ provides us with
553: useful intuition and is consistent with worldsheet results. In a
554: sense however, the branes are truly pointlike, in that the spheres may
555: pass through each other without deformation.
556:
557: The effect of the background on the world-sheet is quantum-mechanical, and
558: presumably may be thought of in terms of a shift in
559: the zero point energy and modings of stretched strings.
560:
561: From a dual perspective,
562: the masslessness of these states arises because the fractional branes
563: are coincident.
564: The relation between these two different points of view, when the geometry
565: can be well understood, corresponds to $T$-duality.
566: Such dualities are expected to persist to more general situations.
567:
568:
569:
570:
571: \bigskip
572: \noindent {\bf Acknowledgments:} We wish to thank M. Strassler and
573: A. Hashimoto for discussions. RGL thanks the Enrico Fermi Institute at the
574: University of Chicago for hospitality while much of this work was carried out.
575: Work supported in part by U.S.
576: Department of Energy, grant DE-FG02-91ER40677 and
577: an Outstanding Junior Investigator Award.
578:
579: %\bibliography{deformation} \bibliographystyle{uiuchept}
580: \providecommand{\href}[2]{#2}\begingroup\raggedright\begin{thebibliography}{10}
581:
582: \bibitem{LS}
583: R.~G. Leigh and M.~J. Strassler, ``{{\it Exactly marginal operators and duality
584: in four-dimensional N=1 supersymmetric gauge theory,}}'' {\em Nucl. Phys.}
585: {\bf B447} (1995) 95--136,
586: \href{http://xxx.lanl.gov/abs/hep-th/9503121}{{hep-th/9503121}}.
587:
588: \bibitem{BJL1}
589: D.~Berenstein, V.~Jejjala, and R.~G. Leigh, ``{{\it Marginal and Relevant
590: Deformations of N=4 Field Theories and Non-Commutative Moduli Spaces of
591: Vacua,}}'' \href{http://xxx.lanl.gov/abs/hep-th/0005087}{{hep-th/0005087}}.
592:
593: \bibitem{D}
594: M.~R. Douglas, ``{{\it D-branes and discrete torsion,}}''
595: \href{http://xxx.lanl.gov/abs/hep-th/9807235}{{hep-th/9807235}}.
596:
597: \bibitem{DF}
598: M.~R. Douglas and B.~Fiol, ``{{\it D-branes and discrete torsion. II,}}''
599: \href{http://xxx.lanl.gov/abs/hep-th/9903031}{{hep-th/9903031}}.
600:
601: \bibitem{BL}
602: D.~Berenstein and R.~G. Leigh, ``{{\it Discrete torsion, AdS/CFT and
603: duality,}}'' {\em JHEP} {\bf 01} (2000) 038,
604: \href{http://xxx.lanl.gov/abs/hep-th/0001055}{{hep-th/0001055}}.
605:
606: \bibitem{BFSS}
607: T.~Banks, W.~Fischler, S.~H. Shenker, and L.~Susskind, ``{{\it M-theory as a
608: matrix model: A conjecture,}}'' {\em Phys. Rev.} {\bf D55} (1997) 5112--5128,
609: \href{http://xxx.lanl.gov/abs/hep-th/9610043}{{hep-th/9610043}}.
610:
611: \bibitem{KTay}
612: D.~Kabat and W.~Taylor, ``{{\it Spherical membranes in matrix theory,}}'' {\em
613: Adv. Theor. Math. Phys.} {\bf 2} (1998) 181,
614: \href{http://xxx.lanl.gov/abs/hep-th/9711078}{{hep-th/9711078}}.
615:
616: \bibitem{BC}
617: D.~Berenstein and R.~Corrado, ``{{\it Matrix Theory on ALE Spaces and Wrapped
618: Membranes,}}''
619: \href{http://xxx.lanl.gov/abs/hep-th/9803048}{{hep-th/9803048}}.
620:
621: \bibitem{RM}
622: R.~C. Myers, ``{{\it Dielectric-branes,}}'' {\em JHEP} {\bf 12} (1999) 022,
623: \href{http://xxx.lanl.gov/abs/hep-th/9910053}{{hep-th/9910053}}.
624:
625: \bibitem{PS}
626: J.~Polchinski and M.~J. Strassler, ``{{\it The string dual of a confining
627: four-dimensional gauge theory,}}''
628: \href{http://xxx.lanl.gov/abs/hep-th/0003136}{{hep-th/0003136}}.
629:
630: \bibitem{M}
631: J.~Maldacena, ``{{\it The large-N limit of superconformal field theories and
632: supergravity,}}'' {\em Adv. Theor. Math. Phys.} {\bf 2} (1998) 231--252,
633: \href{http://xxx.lanl.gov/abs/hep-th/9711200}{{hep-th/9711200}}.
634:
635: \bibitem{W}
636: E.~Witten, ``{{\it Anti-de Sitter space and holography,}}'' {\em Adv. Theor.
637: Math. Phys.} {\bf 2} (1998) 253--291,
638: \href{http://xxx.lanl.gov/abs/hep-th/9802150}{{hep-th/9802150}}.
639:
640: \bibitem{GKP}
641: S.~S. Gubser, I.~R. Klebanov, and A.~M. Polyakov, ``{{\it Gauge theory
642: correlators from non-critical string theory,}}'' {\em Phys. Lett.} {\bf B428}
643: (1998) 105, \href{http://xxx.lanl.gov/abs/hep-th/9802109}{{hep-th/9802109}}.
644:
645: \bibitem{KRN}
646: H.~J. Kim, L.~J. Romans, and P.~van Nieuwenhuizen, ``{{\it The Mass Spectrum Of
647: Chiral N=2 d = 10 Supergravity On $S^5$,}}'' {\em Phys. Rev.} {\bf D32}
648: (1985) 389.
649:
650: \bibitem{LBQ}
651: D.~Berenstein and R.~G. Leigh, ``{{\it Quantization of Superstrings in
652: Ramond-Ramond Backgrounds,}}''
653: \href{http://xxx.lanl.gov/abs/hep-th/9910145}{{hep-th/9910145}}.
654:
655: \bibitem{LB2}
656: D.~Berenstein and R.~G. Leigh, ``{{\it Superstring perturbation theory and
657: Ramond-Ramond backgrounds,}}'' {\em Phys. Rev.} {\bf D60} (1999) 106002,
658: \href{http://xxx.lanl.gov/abs/hep-th/9904104}{{hep-th/9904104}}.
659:
660: \end{thebibliography}\endgroup
661:
662:
663: \end{document}
664:
665: