1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2: %
3: %
4: %
5: %
6: %
7: %
8: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
9:
10: \documentclass[12pt]{article}
11: \usepackage{amsmath}
12: \usepackage[psamsfonts]{amssymb}
13: \usepackage[dvips]{graphicx}
14: \usepackage[hypertex]{hyperref}
15: %\usepackage{showkeys}
16: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
17: % page size
18:
19: \setlength{\textwidth}{ 16truecm}
20: \setlength{\textheight}{24truecm}
21:
22:
23: \setlength{\oddsidemargin}{\paperwidth}
24: \addtolength{\oddsidemargin}{-\textwidth}
25: \setlength{\oddsidemargin}{0.5\oddsidemargin}
26: \addtolength{\oddsidemargin}{-1truein}
27: \setlength{\evensidemargin}{\oddsidemargin}
28: \setlength{\topmargin}{\paperheight}
29: \addtolength{\topmargin}{-\textheight}
30: \setlength{\topmargin}{0.5\topmargin}
31: \addtolength{\topmargin}{-\headheight}
32: \addtolength{\topmargin}{-\headsep}
33: \addtolength{\topmargin}{-1truein}
34:
35: \setlength{\jot}{3ex}
36: \renewcommand{\baselinestretch}{1.3}
37: \numberwithin{equation}{section}
38: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
39: % often used macro
40: \newcommand{\la}{\lambda}
41: \newcommand{\Zb}{\mathbb{Z}}
42: \newcommand{\Rb}{\mathbb{R}}
43: \newcommand{\Cb}{\mathbb{C}}
44: \newcommand{\Ncal}{\mathcal{N}}
45: \newcommand{\del}{\partial}
46: \DeclareMathOperator*{\Tr}{{\rm Tr}}
47: \DeclareMathOperator*{\tr}{{\rm tr}}
48: \DeclareMathOperator{\arccosh}{{\rm arccosh}}
49:
50: \newcommand{\vp}{\varphi}
51: \newcommand{\Fcal}{{\cal F}}
52: \newcommand{\Lcal}{{\cal L}}
53: \newcommand{\Hcal}{{\cal H}}
54: \newcommand{\thet}{\theta_k}
55: \renewcommand{\th}{\Theta}
56: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
57: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
58:
59: %\renewcommand{\baselinestretch}{1.3}
60: \renewcommand{\baselinestretch}{1.2}
61:
62: \begin{document}
63: %%% Title page %%%%%
64: \begin{titlepage}
65:
66: \renewcommand{\thefootnote}{\fnsymbol{footnote}}
67: \begin{flushright}
68: \begin{tabular}{l}
69: {\tt hep-th/0610275}\\
70: UT-06-22 \\
71: IHES/P/06/53 \\
72: \end{tabular}
73: \end{flushright}
74:
75: \vfill
76: \begin{center}
77: \font\titlerm=cmr10 scaled\magstep4
78: \font\titlei=cmmi10 scaled\magstep4
79: \font\titleis=cmmi7 scaled\magstep4
80: {\titlerm Correlator of Fundamental and Anti-symmetric \\
81: Wilson Loops in AdS/CFT Correspondence}
82: \vskip 1.9 truecm
83:
84: \noindent{{\large
85: Ta-Sheng Tai$^a$\footnote{E-mail:tasheng@hep-th.phys.s.u-tokyo.ac.jp}
86: and Satoshi Yamaguchi$^b$\footnote{E-mail:yamaguch@ihes.fr} }}
87: \bigskip
88: \vskip .9 truecm
89: \centerline{\it $^a$ Department of Physics, Faculty of Science,
90: University of Tokyo}
91: \centerline{\it Hongo, Bunkyo-ku, Tokyo 113-0033, JAPAN}
92: \bigskip
93: \centerline{\it $^b$ IHES, Le Bois-Marie, 35, Route de Chartres
94: F-91440 Bures-sur-Yvette, FRANCE}
95: \vskip .4 truecm
96: \end{center}
97: \vfill
98: \vskip 0.5 truecm
99:
100:
101: \begin{abstract}
102: We study the two circular Wilson loop correlator in which one is of
103: anti-symmetric representation, while the other is of fundamental
104: representation in 4-dimensional ${\cal N}=4$ super Yang-Mills theory.
105: This correlator has a good AdS dual, which is a system of a D5-brane and
106: a fundamental string. We calculated the on-shell action of the string,
107: and clarified the Gross-Ooguri transition in this correlator.
108: Some limiting cases are also examined.
109: \end{abstract}
110:
111:
112: \vfill
113: \vskip 0.5 truecm
114:
115: \setcounter{footnote}{0}
116: \renewcommand{\thefootnote}{\arabic{footnote}}
117: \end{titlepage}
118:
119: \newpage
120:
121: \tableofcontents
122: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
123: \section{Introduction and summary}
124:
125: The Wilson loop is one of the most interesting quantities which appeared in
126: the AdS/CFT correspondence
127: \cite{Maldacena:1997re}.
128: In
129: the computation of the expectation value of the Wilson loop, the stringy
130: effect (not just the supergravity) is essential. Actually, it is
131: calculated on the AdS side as the on-shell action of a classical
132: macroscopic string\cite{Rey:1998ik,Maldacena:1998im} in the
133: limit of large $N$ and large 't Hooft coupling $\lambda=4\pi g_s N=g^2_{YM} N$.
134:
135:
136:
137: Recently, certain D-brane pictures of Wilson loops
138: have drawn wide attention because
139: they are powerful in evaluating the Wilson loop of higher rank
140: representation. The D3-brane with electric flux, which has already been
141: investigated in \cite{Rey:1998ik}, is found to
142: describe Wilson loops of symmetric representations (or multiply wound
143: Wilson loops) by Drukker and Fiol \cite{Drukker:2005kx}. On the
144: other hand, the Wilson loop of anti-symmetric representation is
145: found to be described by a D5-brane with electric
146: flux\cite{Hartnoll:2006hr,Yamaguchi:2006tq,Gomis:2006sb}.
147: In both
148: cases, the results show successful agreement with the gauge theory side
149: calculations \cite{Erickson:2000af,Drukker:2000rr}.
150: Other further investigations along this line can be found in \cite{Rodriguez-Gomez:2006zz,Okuyama:2006jc,Hartnoll:2006is,
151: Hartnoll:2006ib,Chen:2006iu,Giombi:2006de}.
152:
153:
154: One of the interesting phenomena in the AdS picture of Wilson loops is the
155: phase transition, which occurs in the two Wilson loop correlator.
156: This is first anticipated by Gross and Ooguri \cite{Gross:1998gk}, and
157: further explored in
158: \cite{Zarembo:1999bu,Olesen:2000ji,Kim:2001td,Arutyunov:2001hs,Drukker:2005cu,Tsuji:2006zn,Ahn:2006px}.
159: If the two loops are close enough, the classical string of annulus
160: topology leads to the smallest action, which dominates the connected
161: correlator. By contrast, when the two loops get far enough,
162: two disk solution with massless propagating modes dominates the connected
163: correlator. Between them, there exists a critical point.
164:
165: In this paper, we study two concentric circular Wilson loop correlator
166: of different representations, i.e. one is fundamental and the other is
167: anti-symmetric%
168: \footnote{Anti-symmetric--anti-symmetric correlator
169: has been studied in \cite{Ahn:2006px}.}. Because the circular
170: anti-symmetric Wilson loop has a gravity dual as a D5-brane with
171: electric flux, this correlator can thus be recognized as a system of a
172: D5-brane and an F-string. More explicitly, it is realized as a
173: string which propagates between the AdS boundary and the
174: D5-brane.
175:
176: Here let us summarize the results of this paper. We calculated the correlator
177: by evaluating the on-shell action of the string worldsheet. This correlator also
178: exibits the phase transition. The phase structure is shown in figures~\ref{pi/2}, \ref{2pi/3} and \ref{pi/3}. The critical surface of the phase transition is represented by a solid line. There is also a critical surface represented by a dashed line where the annulus solution becomes unstable.
179:
180: The parameter $\th$ (direction in $S^5$) dependence is quite nontrivial and interesting. As easily seen from the figures \ref{pi/2}, \ref{2pi/3} and \ref{pi/3},
181: as well as from the expression in section 3, the correlator has
182: the symmetry $\th\to 2\theta_k-\th$, where $\theta_k$ is a function of $k$
183: as in eq.\eqref{thetak}. Especially, the point $\th=2\theta_k-\pi$
184: looks similar to the point $\th=\pi$. At $\th=\pi$, it is always in the disk
185: phase, since it is BPS even when two loops get on top of each other.
186: On the other hand, $\th=2\theta_k-\pi$ is not a priori related to the BPS configuration, but it is
187: always in the disk phase. This fact implies that overlapped loops at
188: $\th=2\theta_k-\pi$ bear a similar nature to the BPS loops.
189:
190: Another interesting feature related to $\th$ dependence is the maximum
191: of the critical distance. The critical distance is maximal at $\th=\theta_k$.
192: This is reasonable in the AdS picture
193: since the D5-brane sits at $\theta=\theta_k$. However,
194: it is rather misterious because
195: the origin of $\theta_k$ is quite different from $\th$ on the gauge theory side.
196: It should be interesting and useful to consider
197: the origin of $\theta_k$ more geometrically on the gause theory side.
198:
199: We also studied a couple of limiting cases. The solution
200: of two fundamental Wilson loop correlator first examined by
201: Zarembo\cite{Zarembo:1999bu} is recovered, meanwhile the Coulomb-like behavior
202: is found
203: when the separation is much smaller than the circle radii.
204:
205: As a future work, the Yang-Mills side analysis like
206: \cite{Zarembo:2001jp,Plefka:2001bu} seems to be interesting.
207: Correlators of Wilson loops of other representations by using the
208: D3-brane \cite{Drukker:2005kx} or supergravity solutions
209: \cite{Yamaguchi:2006te,Lunin:2006xr} are as well worthy of study.
210: Also, the extension to finite temperature cases is in progress.
211:
212: The rest of this paper is organized as follows. In sec.2, the detailed
213: setup is given. In sec.3, the AdS side calculation is performed
214: through a series of elliptic integrals. Finally, in sec.4,
215: we plotted the phase diagram, which visualizes clearly the Gross-Ooguri
216: transition.
217: Comments on these solutions are also provided, where known
218: results appear as limiting cases. Some definitions and algebra
219: involving elliptic integrals are added in the appendix \ref{app-a}.
220: The explicit forms of the Coulomb coefficients are shown in appendix
221: \ref{app-coulomb}.
222:
223:
224:
225:
226:
227:
228:
229:
230:
231:
232: \section{Setup}
233: In $\Ncal=4$ Super Yang-Mills theory, the Wilson loop is defined as
234: \begin{align}
235: W_{\mathsf{R}}(C):= \Tr_{\mathsf{R}}\big[
236: P \exp \oint_{C} d\tau (i A_{\mu}{\dot x}^{\mu}+\Phi_{I}n^{I}(\tau)|\dot x|)
237: \big],
238: \end{align}
239: where $A_{\mu}$ is the gauge field, $\Phi_I$'s are six scalar fields,
240: $C$ is a closed trajectory in 4-dimensional spacetime parameterized by
241: $\tau$, $n_I$ is a 6-dimensional unit vector, and $\mathsf{R}$ is a representation of SU($N$). In this paper, we
242: consider the following type of correlation function
243: \begin{align}
244: \left\langle W_{A_k}(C_1)W_{\Box}(C_2) \right\rangle. \label{cor}
245: \end{align}
246: We take the trajectories $C_1$ and $C_2$ as two circles. Both of them
247: are embedded in a 3-dimensional hyperplane of 4-dimensional space. They
248: are parallel and the centers of them are on the axis which is
249: orthogonal to the circles. If we introduce the Cartesian coordinate
250: $x_{\mu}$, $C_1$ is expressed via $\tau$ as
251: \begin{align}
252: x_{\mu}=(0,R_1 \cos \tau,R_1 \sin \tau,0),\qquad
253: n^{I}=(1,0,0,0,0,0),
254: \end{align}
255: while $C_2$ is
256: \begin{align}
257: x_{\mu}=(L ,R_2 \cos \tau,-R_2 \sin \tau,0),\qquad
258: n^{I}=(\cos \th,\sin \th,0,0,0,0),\qquad (0<L, ~ 0\le \th \le \pi).
259: \label{c1}
260: \end{align}
261: So parameters of this configuration are $R_1,R_2,L,\th$.
262:
263:
264:
265:
266:
267: The correlator \eqref{cor} factorizes in the leading term of
268: the $1/N$ expansion. But this term does not depend on parameters
269: $L,R_1,R_2,\th$. In order to see their dependence,
270: the subleading terms become relevant. In other words, what we have to consider
271: is the ``connected correlator''
272: \begin{align}
273: \left\langle W_{A_k}(C_1)W_{\Box}(C_2) \right\rangle_{conn}
274: =\left\langle W_{A_k}(C_1)W_{\Box}(C_2) \right\rangle-
275: \left\langle W_{A_k}(C_1) \right\rangle
276: \left\langle W_{\Box}(C_2) \right\rangle.
277: \end{align}
278: As we will see later, it is convenient to define the function
279: $V_{k}(R_1,R_2,L ,\th)$ as
280: \begin{align}
281: V_{k}(R_1,R_2,L ,\th):=
282: \frac{\left\langle W_{A_k}(C_1)W_{\Box}(C_2) \right\rangle-
283: \left\langle W_{A_k}(C_1) \right\rangle
284: \left\langle W_{\Box}(C_2) \right\rangle}{
285: \left\langle W_{A_k}(C_1) \right\rangle}.\label{Vk}
286: \end{align}
287: On the AdS side, this correlation function is described by a system of
288: a D5-brane and a fundamental string subject to certain boundary conditions.
289: The D5-brane is rigid because it is much heavier than the fundamental
290: string. By contrast, the string is flexible,
291: so various kinds of contributions should be considered.
292:
293:
294:
295: There are two possible leading contributions to $V_k$.
296: One is a disk bounded by $C_2$
297: with massless propagating modes connecting the F-string
298: and the D5-brane (see figure \ref{fig-disk}).
299: The other is
300: an annulus whose boundaries are
301: attached to the D5-brane and $C_2$ (see figure \ref{fig-annulus}).
302: These two contributions are of
303: the same order $\simeq{\cal O}(N^0)$ in $1/N$ expansion.
304: Their magnitudes are estimated by the on-shell action of
305: the annulus and the disk%
306: \footnote{The massless propagator contributes a factor of order ${\cal O}(\lambda^n)$
307: to the amplitude. While is compared to $\exp(-\sqrt{\lambda})$, it is thus
308: negligible in large $\lambda$ limit.}, respectively, in large $\lambda$ limit.
309: \begin{figure}
310: \begin{center}
311: \includegraphics[width=7truecm]{disk.eps}
312: \end{center}
313: \caption{The schematical picture of the disk solution with massless propagating
314: modes.
315: The disk solution is of order $1/g_s$, the massless propagator
316: is of order $g_s^2$ and the coupling to the D5-brane is of order
317: $1/g_s$. Totally, the contribution is of order $g_s^0$, which is the same
318: order as the annulus contribution.}
319: \label{fig-disk}
320: \end{figure}
321: \begin{figure}
322: \begin{center}
323: \includegraphics[width=7truecm]{annulus.eps}
324: \end{center}
325: \caption{The schematical picture of the annulus. Since the D5-brane is much
326: heavier than the fundamental string, the D5-brane is not deformed. The
327: contribution from this annulus solution is as well of order $g_s^0$.}
328: \label{fig-annulus}
329: \end{figure}
330: In summary, in large $\lambda$ limit, $V_k$ is approximated as
331: \begin{align}
332: V_{k}(R_1,R_2,L ,\th)=\exp(-S_{min}),
333: \label{logV}
334: \end{align}
335: where $S_{min}$ is the global minimum of the string action.
336: The classical on-shell action of the disk, which corresponds to a single
337: circular Wilson loop, has been calculated in \cite{Berenstein:1998ij}. The
338: result is
339: \begin{align}
340: S_{disk,min}=-\sqrt{\lambda}.
341: \label{sdisk}
342: \end{align}
343: The question now is the competition of
344: \eqref{sdisk} and that of the annulus. We will
345: investigate the classical annulus solution and
346: compare it with \eqref{sdisk} in this paper.
347:
348:
349:
350: Before doing this,
351: we can make some observation upon using the conformal symmetry.
352: There are certain constraints on the correlation function.
353: For example, we have a relation from the dilatation
354: \begin{align}
355: V_{k}(R_1,R_2,L ,\th)=V_{k}(a R_1,a R_2,a L ,\th),
356: \end{align}
357: where $a$ is a real positive parameter of the dilatation. We can also use
358: the special conformal transformation with a vector parameter $b^{\mu}$
359: \begin{align}
360: x'^{\mu}=\frac{x^{\mu}-b^{\mu} |x|^2}{1-2b^{\mu}x_{\mu}+|b|^2|x|^2}.
361: \label{sc}
362: \end{align}
363: This special conformal transformation \eqref{sc} along $x_1$, i.e.
364: $b^{\mu}=(h,0,0,0)$ makes
365: \begin{align}
366: \begin{aligned}
367: & V_{k}(R_1,R_2,L ,\th)=V_{k}(R_1',R_2',L ',\th),\\
368: & R_1'=\frac{R_1}{1+h^2 R_1^2},\qquad
369: R_2'=\frac{R_2}{1-2hL +h^2(L ^2+R_2^2)},\\
370: &L '=\frac{L -h(L ^2+R_2^2)}{1-2hL +h^2(L ^2+R_2^2)}
371: +\frac{hR_1^2}{1+h^2 R_1^2}.
372: \end{aligned}
373: \end{align}
374: Therefore, $V_{k}$ is a function of the invariant combination
375: $\eta$ as%
376: \footnote{This $\eta$ plays the same role as the ``cross ratio.''}
377: \begin{align}
378: V_{k}(R_1,R_2,L ,\th)=\widetilde V_{k}(\eta,\th),\qquad
379: \eta=\frac{R_1^2+R_2^2+L ^2}{2R_1R_2}.
380: \label{eta}
381: \end{align}
382: On the AdS side,
383: this simple dependence on $(\eta,\th)$
384: is not quite trivial. This is due to the introduction of the cutoff, which makes
385: the conformal symmetry not manifest. In order to check this conformal symmetry,
386: we keep the parameters $(R_1,R_2,L)$ generic in the calculation in the next section.
387: We will find that the constraint \eqref{eta} is actually satisfied.
388:
389:
390:
391:
392:
393:
394:
395:
396:
397:
398:
399: \section{Classical solution of the annulus}\label{sec-solve}
400: In this section, we will investigate the annulus solution.
401: We use similar techniques employed by \cite{Zarembo:1999bu,Olesen:2000ji,
402: Kim:2001td}.
403: We write down the solution and the corresponding on-shell action.
404: One can see the ``phase structure'' from the result.
405:
406: \subsection{Ansatz and equations of motion}
407: In this paper, we use the following notation for
408: $AdS_5\times S^5$ with radius $\sqrt{\alpha'}\lambda^{1/4}$
409: \begin{align}
410: ds^2=\frac{\alpha'\sqrt{\lambda}}{y^2}(dy^2+dr^2+r^2d\vp^2+dt^2+dx^2)
411: +\alpha'\sqrt{\lambda}(d\theta^2+\sin^2\theta d\Omega_4^2). \label{metric}
412: \end{align}
413: The boundary of $AdS_5$ is at $y=0$. The coordinate
414: $(r,\vp,t,x)$
415: at the boundary
416: is related to the Cartesian coordinate $x_{\mu}$
417: introduced in the previous section as
418: \begin{align}
419: x_1=x, \qquad x_2=r \cos \vp,\qquad x_3=r\sin \vp,\qquad x_4=t.
420: \end{align}
421: The string dual of $W_{A_k}(C_1)$ is a D5-brane
422: \cite{Yamaguchi:2006tq,Gomis:2006sb}.
423: Its explicit form\cite{Pawelczyk:2000hy,Camino:2001at} is expressed as
424: \begin{align}
425: r^2+y^2=R_1^2,\qquad \theta=\theta_k, \label{D5r}
426: \end{align}
427: where $R_1$ is the radius of $C_1$, and $\theta_k$ is related to $k$ as
428: \begin{align}
429: \frac{k}{N}=\frac{1}{\pi}\left(\theta_k-\frac12 \sin 2\theta_k\right).
430: \label{thetak}
431: \end{align}
432: The gauge field excitation on the D5-brane worldvolume has the
433: field strength
434: \begin{align}
435: \Fcal_{y\vp}=-i\cos\thet \frac{\alpha'\sqrt{\lambda} R_1}{y^2},
436: \qquad \Fcal:=2\pi \alpha' dA.
437: \label{D5A}
438: \end{align}
439: The worldsheet stretching between this D5-brane and $C_2$ is
440: symmetric under the rotation $\vp\to\vp+\alpha,\
441: \alpha$: real constant parameter. Hence,
442: we can assign this symmetry to our solution.
443: The reparameterization degrees of freedom enable one to set
444: the worldsheet coordinates as $(\sigma,\vp)$,
445: where $\vp$ is the same as the spacetime $\vp$.
446: The ansatz used is written as
447: \begin{align}
448: y&=y(\sigma),\qquad r=r(\sigma),\qquad x=x(\sigma),\qquad \theta=\theta(\sigma).\label{ansatz}
449: \end{align}
450: In this ansatz, we include $\sigma$ dependence of $\theta$
451: because the two boundaries are separated in $S^5$ in general, i.e.
452: the D5-brane is wrapped on an $S^4$ at $\theta=\theta_k$,
453: while $C_2$ sits at a point $\theta=\th$ in $S^5$. This kind of separation in $S^5$
454: is first considered extensively in \cite{Drukker:2005cu}.
455: Their method of solving the equation of motion is applicable to the problem here,
456: if one uses the conformal symmetry and simplifies the configuration. Here, however, we do not use the conformal symmetry to simplify the configuration. This is because one purpose of this calculation is to check the conformal symmetry.
457:
458: The bulk part comes from the Nambu-Goto action
459: \begin{align}
460: S_{bulk}=\frac{1}{2\pi \alpha'}\int d \sigma d\vp \sqrt{\det G},
461: \end{align}
462: where $G$ is the induced metric. Plugging \eqref{ansatz}
463: and integrating out $\vp$, we obtain
464: \begin{align}
465: &S_{bulk}=\sqrt{\la} \int d \sigma \frac{r}{y^2}
466: \sqrt{x'^2+r'^2+y'^2+y^2\theta'^2} ,\label{bulk-action}
467: \end{align}
468: where prime ${}'$ denotes the $\sigma$ derivative.
469:
470:
471:
472:
473: At both boundaries, we need to include
474: proper boundary terms. Let $\sigma=\sigma_1$ be the boundary on the D5-brane,
475: whereas $\sigma=\sigma_2$ be the boundary at $C_2$ with $\sigma_1 < \sigma_2$.
476: At $\sigma=\sigma_1$, the boundary is constrained on the rigid D5-brane in \eqref{D5r},
477: and should satisfy at least the following conditions
478: \begin{align}
479: r^2+y^2=R_1^2,\qquad x=0,\qquad \theta=\theta_k.
480: \label{bc1-1}
481: \end{align}
482: These are not all the conditions however.
483: Due to the gauge field excitation in \eqref{D5A},
484: we should add
485: \begin{align}
486: \begin{aligned}
487: S_{bdy,1}=i\oint_{\sigma=\sigma_1}
488: d\varphi ~A_{\varphi} ~ +(\text{constant})=\sqrt{\lambda}
489: \cos{\theta_k} \left(1-\frac{R_1}{y}\right),~
490: &&A_{\varphi}=\frac{1}{2\pi \alpha'}i \cos{\theta_k}
491: \frac{\alpha'\sqrt{\lambda} R_1}{y}.
492: \label{gaugevol}
493: \end{aligned}
494: \end{align}
495: The (constant) ensures that $S_{bdy,1}=0$ when $y=R_1$, where
496: the circle of $\vp$ shrinks to a point.
497: The boundary condition is derived from the variational principle.
498: The variation of the bulk action has boundary terms at $\sigma=\sigma_1$ as
499: \begin{align}
500: \delta S_{bulk}|_{bdy,1}
501: =&-p_r \delta r-p_y \delta y-p_x \delta x-p_{\theta}\delta \theta.
502: \label{bv}
503: \end{align}
504: The momenta $p$ are written as
505: \begin{align}
506: p_r=\sqrt{\lambda}\frac{r}{y^2}\frac{r'}{\sqrt{x'^2+r'^2+y'^2+y^2\theta'^2}},\\
507: p_y=\sqrt{\lambda}\frac{r}{y^2}\frac{y'}{\sqrt{x'^2+r'^2+y'^2+y^2\theta'^2}},\\
508: p_x=\sqrt{\lambda}\frac{r}{y^2}\frac{x'}{\sqrt{x'^2+r'^2+y'^2+y^2\theta'^2}},\\
509: p_{\theta}=\sqrt{\lambda}\frac{r\theta'}{\sqrt{x'^2+r'^2+y'^2+y^2\theta'^2}}.
510: \end{align}
511: Due to $r^2+y^2=R_1$,
512: it is convenient to rewrite \eqref{bv} as
513: \begin{align}
514: \delta S_{bulk}|_{bdy,1} &=-\frac12 \left(\frac{p_r}{r}+\frac{p_y}{y}\right)
515: \left(r\delta r+y\delta y\right)
516: -\frac12 \left(\frac{p_r}{r}-\frac{p_y}{y}\right)
517: \left(r\delta r-y\delta y\right)
518: -p_x \delta x-p_{\theta}\delta \theta.
519: \end{align}
520: On the other hand, the variation of $S_{bdy,1}$ is written as (using the
521: condition $r\delta r+y\delta y=0$)
522: \begin{align}
523: \delta S_{bdy,1}=-\frac12 \sqrt{\lambda}\cos \theta_k \frac{R_1}{y^3}
524: (r\delta r-y\delta y).
525: \end{align}
526: We obtain totally
527: \begin{align}
528: \delta S_{bulk}|_{bdy,1}+\delta S_{bdy,1}
529: &=-\frac12 \left(\frac{p_r}{r}+\frac{p_y}{y}\right)
530: \left(r\delta r+y\delta y\right)
531: -\frac12 \left(\frac{p_r}{r}-\frac{p_y}{y}
532: +\sqrt{\lambda} \cos \theta_k \frac{R_1}{y^3}\right)
533: \left(r\delta r-y\delta y\right)\nonumber\\
534: &\qquad-p_x \delta x-p_{\theta}\delta \theta.
535: \end{align}
536: The first, third and fourth term vanish because of
537: \eqref{bc1-1}. From the second term, it can be seen that
538: the additional boundary condition at $\sigma=\sigma_1$ is
539: \begin{align}
540: \frac{p_r}{r}-\frac{p_y}{y}+\sqrt{\lambda} \cos \theta_k \frac{R_1}{y^3}=0.
541: \label{tb}
542: \end{align}
543:
544:
545:
546: Next, let us turn to the other boundary at $\sigma=\sigma_2$. Naively, the area of
547: the minimal surface is infinite. It is necessary to
548: introduce a cutoff at
549: $y=\epsilon$ in order to regularize the action.
550: We follow \cite{Drukker:1999zq} and perform the
551: Legendre transformation. This gives a boundary term
552: \begin{align}
553: S_{bdy,2}=-p_y y-p_{\theta}(\theta-\th), ~~~~\text{at}~\sigma=\sigma_2.
554: \label{bt2}
555: \end{align}
556:
557: Now we solve the equations of motion subject to the above boundary
558: conditions. Each equation of motion of $r,y,x,\theta$, respectively,
559: derived from \eqref{bulk-action} can be listed as
560: \begin{align}
561: & \frac{1}{y^2}\sqrt{x'^2+r'^2+y'^2+y^2\theta'^2}-\del_{\sigma}
562: \left(\frac{r}{y^2}\frac{r'}{\sqrt{x'^2+r'^2+y'^2+y^2\theta'^2}}\right)=0,\label{eom-r}\\
563: &-2\frac{r}{y^3}\sqrt{x'^2+r'^2+y'^2+y^2\theta'^2}+\frac{r}{y}
564: \frac{\theta'^2}{\sqrt{x'^2+r'^2+y'^2+y^2\theta'^2}}
565: -\del_{\sigma}\left(\frac{r}{y^2}\frac{y'}{\sqrt{x'^2+r'^2+y'^2+y^2\theta'^2}}\right)=0,\label{eom-y}\\
566: & \del_{\sigma}\left(\frac{r}{y^2}\frac{x'}{\sqrt{x'^2+r'^2+y'^2+y^2\theta'^2}}\right)=0,\label{eom-x}\\
567: &\del_{\sigma}\left(\frac{r\theta'}{\sqrt{x'^2+r'^2+y'^2+y^2\theta'^2}}\right)
568: =0.\label{eom-theta}
569: \end{align}
570: Following \cite{Zarembo:1999bu,Olesen:2000ji},
571: we integrate \eqref{eom-x}, \eqref{eom-theta} and fix the reparameterization
572: of $\sigma$ as $x=\sigma$ to obtain
573: \begin{align}
574: & \frac{r}{y^2}\frac{1}{\sqrt{1+r'^2+y'^2+y^2\theta'^2}}=\ell,\label{ell}\\
575: & \frac{r\theta'}{\sqrt{1+r'^2+y'^2+y^2\theta'^2}}=m,\label{m}
576: \end{align}
577: where $\ell$ and $m$ are integration constants. We can assume $\ell >0$
578: without loss of generality. Plugging these into \eqref{eom-r} and
579: \eqref{eom-y}, we obtain
580: \begin{align}
581: \frac{r}{\ell^2y^4}-r''&=0,
582: \label{eom-r2}\\
583: -\frac{2r^2}{\ell^2 y^5}+\frac{m^2}{\ell^2 y^3}-y''&=0,
584: \label{eom-y2}
585: \end{align}
586: and they combine to be
587: \begin{align}
588: &0=1+r'^2+y'^2-\frac{r^2}{\ell^2y^4}+\frac{m^2}{\ell^2 y^2}.
589: \label{read}
590: \end{align}
591: Adding \eqref{eom-r2} multiplied by $r$ to
592: \eqref{eom-y2} multiplied by $y$ leads to
593: \begin{align}
594: & (r^2+y^2)''+2=0.
595: \end{align}
596: This equation is integrated to be
597: \begin{align}
598: &r^2+y^2+(x+c)^2=a^2,
599: \end{align}where
600: $a$ and $c$ are integration constants.
601: At $x=0$ (D5-brane side), $r^2+y^2=R_1^2$ leads to
602: \begin{align}
603: a^2-c^2=R_1^2,
604: \end{align}
605: while at $x=L $ (AdS boundary,) $r=R_2$ and $y=0$ lead to
606: \begin{align}
607: a^2-(L +c)^2=R_2^2.
608: \end{align}
609: Based on these, $a$ and $c$ can be expressed in terms of $R_1,R_2,L $ as
610: \begin{align}
611: c=\frac{R_1^2-R_2^2-L ^2}{2L },\qquad
612: a=\frac{\sqrt{(R_1^2+R_2^2+L ^2)^2-4R_1^2R_2^2}}{2L }.
613: \label{ac}
614: \end{align}
615: By using the trigonometric parameterization
616: \begin{align}
617: r=\sqrt{a^2-(x+c)^2}\cos\phi ,~~~~
618: y=\sqrt{a^2-(x+c)^2}\sin\phi , \qquad
619: 0\le \phi \le \frac{\pi}{2},
620: \label{tri}
621: \end{align}
622: \eqref{read} becomes
623: \begin{align}
624: \phi'=\pm \left(\frac{a}{a^2-(x+c)^2}\right)
625: \frac{\sqrt{\cos^2\phi-m^2\sin^2\phi-\ell^2a^2\sin^4\phi}}{\ell a \sin^2\phi}.
626: \label{prime}
627: \end{align}
628: Furthermore, the boundary condition \eqref{tb} can
629: be rewritten as a constraint on
630: $\phi_1:=\phi(0)$ and $\phi'(0)$ as
631: \begin{align}
632: \phi'(0)=\cos\theta_k \frac{\sin \phi_1}{\ell R_1^2\cos^2 \phi_1},
633: \label{phiprime0}
634: \end{align}
635: where we have used \eqref{tri}.
636: Also,
637: \eqref{phiprime0} and \eqref{prime} lead to
638: the expression of $\phi_1$ as
639: \begin{align}
640: \sin^2 \phi_1=\frac{-(m^2+\sin^2\theta_k)+\sqrt{(m^2+\sin^2\theta_k)^2
641: +4\ell^2a^2\sin^2\theta_k}}{2\ell^2 a^2}.\label{sin2phi1}
642: \end{align}
643: Finally, \eqref{m} can be rewritten as
644: \begin{align}
645: \theta'=\frac{m}{\ell(a^2-(x+c)^2)\sin^2\phi}
646: \label{thetap}
647: \end{align}
648: by using \eqref{ell} and \eqref{tri}.
649: This and \eqref{prime} together lead to
650: \begin{align}
651: \frac{d\phi}{d\theta}=\frac{\phi'}{\theta'}=\pm \frac{1}{m}
652: \sqrt{\cos^2\phi-m^2\sin^2\phi-\ell^2 a^2 \sin^4\phi}.
653: \label{dphidtheta}
654: \end{align}
655:
656:
657:
658:
659:
660:
661:
662:
663:
664:
665:
666:
667: Now, we have almost done. Note that \eqref{prime} and \eqref{dphidtheta} can
668: be easily integrated. The remaining subtlety is to choose
669: the right branch from them.
670: At $x=0$, we find that the sign of $\phi'(0)$ is determined
671: by the sign of $\cos \theta_k$ from \eqref{phiprime0}.
672: We will examine two cases: $\theta_k > \pi/2$ and $\theta_k < \pi/2$
673: in turn. The case of $\theta_k=\pi/2$ is realized as the limit of either case.
674:
675:
676:
677:
678:
679:
680:
681:
682:
683: \subsection{$\theta_k > \pi/2$: monotonically decreasing $\phi$}
684: When $\theta_k>\pi/2$, $\phi'(0)$ is
685: negative by virture of \eqref{phiprime0}, i.e. around $x=0$, $\phi$
686: decreases. Defining what inside the square root in \eqref{prime}
687: as a function of $\phi$~:
688: \begin{align}
689: \Hcal(\phi):=\cos^2\phi-m^2\sin^2\phi-\ell^2 a^2 \sin^4\phi,
690: \label{h}
691: \end{align}
692: we see that
693: $\Hcal(\phi)$ is a monotonically decreasing function on $0 \le \phi \le \pi/2$,
694: and $\Hcal(\phi_1)> 0$ according to \eqref{sin2phi1}.
695: Hence as $\phi$ decreases, $\Hcal$ becomes bigger and never reaches
696: the branching point $\Hcal=0$.
697: Therefore, $\phi(x)$ is a monotonically decreasing
698: function over $0\le x \le L $. We thus take the minus sign
699: in \eqref{prime}, \eqref{dphidtheta}, and then integrate
700: them to obtain
701: \begin{align}
702: & \int_{\phi_1}^{\phi}\frac{\ell a \sin^2\psi d\psi}{\sqrt{\cos^2 \psi-m^2\sin^2\psi-\ell^2 a^2\sin^4 \psi}}
703: =-\frac12 \log \frac{a+x+c}{a-x-c}+\frac12 \log\frac{a+c}{a-c}, \label{sol1p}\\
704: &\int_{\phi_1}^{\phi}\frac{m d\psi}{\sqrt{\cos^2 \psi-m^2\sin^2\psi-\ell^2 a^2\sin^4 \psi}}
705: =-\theta+\theta_k.\label{sol1t}
706: \end{align}
707: Note that $\phi(x)$
708: and $\theta(x)$
709: can be completely fixed by solving the above equations implemented
710: with the boundary conditions :
711: \begin{align}
712: \phi(L )=0,\qquad \theta(L )=\th,
713: \label{bcxi}
714: \end{align}
715: which in turn determine $m$ and $\ell$.
716: To see this more explicitly,
717: it is convenient to introduce
718: \begin{align}
719: f(m,\ell a)&=
720: \int_0^{\phi_1}\frac
721: {\ell a \sin^2\psi d\psi}{\sqrt{\cos^2 \psi-m^2\sin^2\psi -\ell^2 a^2\sin^4 \psi}} ,\label{f}\\
722: g(m,\ell a)&=\int_0^{\phi_1} \frac{m d\psi }{\sqrt{\cos^2 \psi-m^2\sin^2\psi
723: -\ell^2 a^2\sin^4 \psi}}.
724: \label{g}
725: \end{align}
726: Rewriting the L.H.S. of \eqref{sol1p} at $x=L $ via
727: \eqref{ac} into
728: \begin{align}
729: \begin{aligned}
730: -\frac{1}{2}\log\frac{a+L +c}{a-L -c}
731: +\frac{1}{2}\log\frac{a+c}{a-c}
732: &=-\log\frac{R_1^2+R_2^2+L ^2+\sqrt{(R_1^2 -R_2^2)^2+L ^4+2L ^2(R_1^2 +R_2^2)}}
733: {2R_1 R_2}\\
734: &=-\log\left[\eta+\sqrt{\eta^2-1}\right],
735: \end{aligned}
736: \end{align}
737: where $\eta$ is defined in \eqref{eta},
738: we can re-express \eqref{bcxi} as
739: \begin{align}
740: \cosh[f(m,\ell a)]=\eta,\qquad
741: g(m,\ell a)= \th-\theta_k.
742: \label{Theta}
743: \end{align}
744: Namely,
745: $m$ and $\ell a$ are determined in terms of $\eta,\th$ if exist.
746: Let us compute the on-shell action.
747: The bulk action \eqref{bulk-action} now takes the form
748: \begin{align}
749: &S_{bulk}=\sqrt{\la}\int_0^L dx \frac{r}{y^2}
750: \sqrt{1+r'^2+y'^2 +y^2\theta'^2}
751: =\sqrt{\la}\int^{\phi_1}_{\frac{\epsilon}{R_2}}
752: \frac{d\phi\cot^2\phi}{\sqrt{\cos^2\phi -m^2 \sin^2\phi -{\ell^2 a^2} \sin^4\phi}},
753: \end{align}
754: where we have put a cutoff at $y=\epsilon$. We ignore terms which vanish
755: when $\epsilon \to 0$. Also, there are two boundary terms
756: \begin{align}
757: &x=0, ~~~~S_{bdy,1}=\sqrt{\lambda }\cos\theta _k (1-\csc\phi_1),
758: \label{onbdy1}\\
759: &x=L, ~~~~S_{bdy,2}=-\sqrt{\lambda}\frac{R_2}{\epsilon},
760: \label{onbdy2}
761: \end{align}
762: from \eqref{gaugevol} and \eqref{bt2}, respectively.
763: The total on-shell action is $S_{tot}=S_{bulk}+S_{bdy,1}+S_{bdy,2}$.
764:
765:
766:
767:
768:
769:
770: Here we summarize the result of this subsection. When $\theta_k > \pi/2$,
771: the on-shell action can be written by using the formulas in appendix \ref{app-elliptic} as
772: \begin{align}
773: \begin{aligned}
774: \frac{S_{tot}}{\sqrt{\lambda}}=&
775: [(m^2+1)^2+4\ell^2 a^2]^{1/4}
776: \Bigg[-\cot\chi(\phi_1)\sqrt{1-\kappa^2\sin^2\chi(\phi_1)}-E(\chi(\phi_1),\kappa)\\
777: &\qquad \qquad\qquad \qquad
778: +(1-\kappa^2)F(\chi(\phi_1),\kappa)\Bigg]
779: +\cos\theta _k (1-\csc\phi_1).
780: \end{aligned}
781: \label{res1s}
782: \end{align}
783: Note that $\phi_1$ is defined in \eqref{sin2phi1}, whereas
784: $\chi(\phi), \kappa$ and $C$
785: are defined in \eqref{defi}. Here
786: $F$ and $E$ are the elliptic integrals of the first and second kind, respectively.
787: The integration constants $m, \ell a$ are determined by
788: \begin{align}
789: \arccosh \eta&=\frac{(m^2+1)+\sqrt{(m^2+1)^2+4\ell^2 a^2}}{
790: 2\ell a \left[(m^2+1)^2+4\ell^2 a^2\right]^{1/4}}
791: \left[
792: \Pi(\chi(\phi_1),C,\kappa)-F(\chi(\phi_1),\kappa)
793: \right],\label{res1f}\\
794: \th-\theta_k&=m[(m^2+1)^2+4\ell^2 a^2]^{-1/4}
795: F(\chi(\phi_1),\kappa),\label{res1g}
796: \end{align}
797: where $\Pi$ is the elliptic integral of the third kind.
798: We found that this result satisfies the conformal
799: symmetry condition indicated in \eqref{eta}.
800:
801:
802:
803:
804: \subsection{$\theta_k < \pi/2$: increasing and decreasing $\phi$}
805: When $\theta_k<\pi/2$, $\phi'(0)$ is positive by
806: virtue of \eqref{phiprime0}. Note that
807: $\phi$ keeps increasing from $x=0$
808: (during when $\Hcal$ in \eqref{h} decreases) untill
809: the branching point ($\Hcal=0$) is reached at $x=x_0$. This causes
810: a sign change in
811: \eqref{prime} and \eqref{dphidtheta} such that
812: $\phi$ starts to decrease all the way to zero at $x=L $.
813: Consequently, we take the plus branch for $x\le x_0$, and the minus branch
814: for $x \ge x_0$ in \eqref{prime} and \eqref{dphidtheta}.
815:
816:
817:
818:
819: It is convenient to define $\phi_0:=\phi(x_0)$,
820: which satisfies $\Hcal(\phi_0)=0$ and can be solved as
821: \begin{align}
822: \sin^2 \phi_0=\frac{-(m^2+1)+\sqrt{(m^2+1)^2+4\ell^2 a^2}}{2\ell^2 a^2}.
823: \label{phi0}
824: \end{align}
825: When $x\le x_0$, integrating \eqref{prime} and \eqref{dphidtheta} gives
826: \begin{align}
827: & \int_{\phi_1}^{\phi}\frac{\ell a \sin^2\psi d\psi}{\sqrt{\cos^2 \psi-m^2\sin^2\psi-\ell^2 a^2\sin^4 \psi}}
828: =\frac12 \log \frac{a+x+c}{a-x-c}-\frac12 \log\frac{a+c}{a-c}, \label{sol21p}\\
829: &\int_{\phi_1}^{\phi}\frac{m d\psi}{\sqrt{\cos^2 \psi-m^2\sin^2\psi-\ell^2 a^2\sin^4 \psi}}
830: =\theta-\theta_k.\label{sol21t}
831: \end{align}
832: On the other hand, when $x\ge x_0$,
833: \begin{align}
834: & \left(\int_{\phi_1}^{\phi_0}-\int_{\phi_0}^{\phi}\right)\frac{\ell a \sin^2\psi d\psi}{\sqrt{\cos^2 \psi-m^2\sin^2\psi-\ell^2 a^2\sin^4 \psi}}
835: =\frac12 \log \frac{a+x+c}{a-x-c}-\frac12 \log\frac{a+c}{a-c}, \label{sol22p}\\
836: &\left(\int_{\phi_1}^{\phi_0}-\int_{\phi_0}^{\phi}\right)\frac{m d\psi}{\sqrt{\cos^2 \psi-m^2\sin^2\psi-\ell^2 a^2\sin^4 \psi}}
837: =\theta-\theta_k.\label{sol22t}
838: \end{align}
839:
840: We again define
841: \begin{align}
842: \tilde f(m,\ell a)&=\left(\int_{\phi_1}^{\phi_0}+\int_{0}^{\phi_0}\right)
843: \frac{\ell a \sin^2\psi d\psi}{\sqrt{\cos^2 \psi-m^2\sin^2\psi -{\ell^2 a^2}\sin^4 \psi}} ,\\
844: \tilde g(m,\ell a)&=\left(\int_{\phi_1}^{\phi_0}+\int_{0}^{\phi_0}\right)
845: \frac{m d\psi }{\sqrt{\cos^2 \psi-m^2\sin^2\psi -{\ell^2 a^2}\sin^4 \psi}}
846: \end{align}
847: as before in order to incorporate the boundary condition \eqref{bcxi}.
848: \eqref{sol22p}, \eqref{sol22t} and \eqref{bcxi} lead to
849: \begin{align}
850: \eta=\cosh [\tilde f(m,\ell a)],\qquad
851: \tilde g(m,\ell a)= \th-\theta_k.
852: \label{Theta--}
853: \end{align}
854: Here, $m$ and $\ell a$ are determined in terms of $\eta,\th$ if exist.
855: The on-shell action $S_{tot}$ contains
856: the bulk part
857: \begin{align}
858: \begin{aligned}
859: S_{bulk}
860: &=\sqrt{\la}\left(\int_{\phi_1}^{\phi_0}
861: +\int_{\frac{\epsilon}{R_2}}^{\phi_0}\right)
862: \frac{d\phi\cot^2\phi}{\sqrt{\cos^2\phi -m^2 \sin^2\phi -{\ell^2 a^2} \sin^4\phi}}
863: \label{Sinde}
864: \end{aligned}
865: \end{align}
866: as well as boundary terms which are the same as \eqref{onbdy1} and \eqref{onbdy2}.
867:
868:
869:
870: We summarize the result of this subsection. When $\theta_k < \pi/2$,
871: the on-shell action can be written by using the formulas in appendix \ref{app-elliptic} as
872: \begin{align}
873: \begin{aligned}
874: \frac{S_{tot}}{\sqrt{\lambda}}=&
875: [(m^2+1)^2+4\ell^2 a^2]^{1/4}
876: \Bigg[\cot\chi(\phi_1)\sqrt{1-\kappa^2\sin^2\chi(\phi_1)}
877: \\&
878: +E(\chi(\phi_1),\kappa)
879: -(1-\kappa^2)F(\chi(\phi_1),\kappa)-2E(\kappa)
880: +2(1-\kappa^2)K(\kappa)\Bigg]
881: +\cos\theta _k (1-\csc\phi_1),
882: \end{aligned}
883: \label{res2s}
884: \end{align}
885: where $\phi_1$ is defined in \eqref{sin2phi1}, while $\chi(\phi), \kappa$, and
886: $C$ are defined in \eqref{defi}.
887: The integration constants $m, \ell a$ are determined by
888: \begin{align}
889: & \arccosh \eta=\frac{(m^2+1)+\sqrt{(m^2+1)^2+4\ell^2 a^2}}{
890: 2\ell a \left[(m^2+1)^2+4\ell^2 a^2\right]^{1/4}}
891: \left[
892: 2\Pi(C,\kappa)-2K(\kappa)
893: -\Pi(\chi(\phi_1),C,\kappa)+F(\chi(\phi_1),\kappa)
894: \right],\label{res2f}\\
895: & \th-\theta_k=m[(m^2+1)^2+4\ell^2 a^2]^{-1/4}[
896: 2K(\kappa)
897: -F(\chi(\phi_1),\kappa)].
898: \label{res2g}
899: \end{align}
900: Again, the conformal symmetry condition in
901: \eqref{eta} is satisfied.
902:
903:
904:
905:
906:
907:
908:
909:
910:
911:
912:
913: \section{Comments on the solutions}
914: \subsection{Gross-Ooguri phase transition}
915: Let us consider the Gross-Ooguri (GO) phase transition here. We compare
916: \eqref{res1s}-\eqref{res1g} (or \eqref{res2s}-\eqref{res2g}, when
917: $\theta_k < \pi/2$) and \eqref{sdisk}. We are also interested in
918: how far the annulus solution can stretch along $\eta$. For example, when $\theta_k > \pi/2$,
919: the annulus solution exists, if and only if
920: \eqref{res1f} and \eqref{res1g} can be solved by some $(m,\ell a)$
921: for a given $(\eta,\th)$ pair.
922:
923:
924:
925: Figures \ref{pi/2}, \ref{2pi/3} and \ref{pi/3} present the phase
926: diagrams plotted against the $\eta$-$\th$ plane at $\theta_k=\pi/2$, $\
927: 2\pi/3$ and $\ \pi/3$, respectively. The black solid line indicates the
928: critical line of the GO phase transition.
929: The red dashed line stands for the critical line where the annulus solution
930: becomes unstable.
931:
932: One finds that $\th=\pi$ is always in the disk phase. This
933: observation has a good explanation. When $\th=\pi$, two loops
934: get overlapped ($\eta=1$) so that this configuration becomes BPS. Therefore, the
935: annulus solution does not exist.
936:
937: The solution obtained in the previous section has a ``duality'' for a fixed
938: $k$, i.e. $(\th,\eta)\to(2\theta_k-\th,\eta)$.
939: Consequently, for $k\ge N/2$ ($\theta_k\ge \pi/2$),
940: $\th\le 2\theta_k-\pi$ is
941: always in the disk phase.
942: In paticular, the point $\th=2\theta_k-\pi,\ \eta=0$
943: bears a similar nature to the BPS point $\th=\pi,\ \eta=0$ in the approximation used here.
944:
945: In addition, $\th$ dependence of the phase diagram is interesting.
946: The critical distance marks a maximum at $\th=\theta_k$ when $\theta_k$
947: is fixed. This fact is quite reasonable from the AdS point of
948: view. However, it is rather mysterious from the gauge theory side, since
949: $\th$ and $\theta_k$ have rather different origins; $\th$ is
950: the direction of the scalar as in \eqref{c1}, while $\theta_k$ is
951: determined by the rank of the anti-symmetric representation as
952: in \eqref{thetak}. To consider this fact from the gauge theory is an
953: interesting future work.
954:
955:
956: \begin{figure}[htbp]
957: \begin{center}
958: \includegraphics{half-pi.eps}
959: \end{center}
960: \caption{Phase diagram on $\eta$-$\th$ plane with $\theta_k=\pi/2$. The
961: solid black line indicates the critical line of the GO transition.
962: Inside (outside) the solid line, the annulus (disk) solution dominates
963: the connected correlator. The red dashed line represents the critical
964: line where the annulus solution becomes unstable. The annulus solution
965: does not exist outside the red dashed line.} \label{pi/2}
966: \end{figure}
967:
968: \begin{figure}[htbp]
969: \begin{center}
970: \includegraphics{2-3rd.eps}
971: \end{center}
972: \caption{Phase diagram on $\eta$-$\th$ plane with $\theta_k=2\pi/3$.}
973: \label{2pi/3}
974: \end{figure}
975:
976:
977: \begin{figure}[htbp]
978: \begin{center}
979: \includegraphics{1-3rd.eps}
980: \end{center}
981: \caption{Phase diagram on $\eta$-$\th$ plane with $\theta_k=\pi/3$.}
982: \label{pi/3}
983: \end{figure}
984:
985:
986:
987: \subsection{$k=1$}
988: It is illuminating to check some limiting aspects of the above
989: solutions. When $k=1$, the $k$-th anti-symmetric representation reduces
990: to the fundamental one. Assuming $k=1$, one
991: may expect \eqref{res2s}-\eqref{res2g} reproduce the correlator of two
992: fundamental loops. We will see this is actually the case. It is
993: nontrivial in the sense that the D5-brane picture of the
994: anti-symmetric loop is valid only when $k$ is large and comparable to
995: $N$. From \eqref{thetak}, it is seen that
996: $\lim_{k\rightarrow1}\theta_{k}\cong \left(\frac{3\pi}{2N}\right)^{1/3}
997: \ll 1$. Due to $\theta_k\ll 1$, one finds $\phi_1 \ll 1$ by
998: using \eqref{sin2phi1}. This means that the endpoint on the D5-brane is
999: drawn closely to the AdS boundary.
1000:
1001:
1002:
1003:
1004: Taking these facts into account, we can express the results
1005: \eqref{res2s}-\eqref{res2g} for $k=1$ as
1006:
1007:
1008: \begin{align}
1009: \begin{aligned}
1010: S_{tot}=&
1011: \sqrt{\lambda}+2\sqrt{\lambda}[(m^2+1)^2+4\ell^2 a^2]^{1/4}
1012: \left[
1013: -E(\kappa)
1014: +(1-\kappa^2)K(\kappa)\right]\label{fk1}
1015: \end{aligned}
1016: \end{align}
1017: \begin{align}
1018: & \arccosh \eta=\frac{(m^2+1)+\sqrt{(m^2+1)^2+4\ell^2 a^2}}{
1019: \ell a \left[(m^2+1)^2+4\ell^2 a^2\right]^{1/4}}
1020: \left[
1021: \Pi(C,\kappa)-K(\kappa)
1022: \right],\label{res2f}\\
1023: & \th=2m[(m^2+1)^2+4\ell^2 a^2]^{-1/4}
1024: K(\kappa).
1025: \end{align}
1026:
1027:
1028:
1029:
1030:
1031: %\begin{align}
1032: % \arccosh \eta=
1033: % \frac{1+\sqrt{1+4\ell^2 a^2}}{\ell a(1+4\ell^2 a^2)^{1/4}}\left[
1034: %\Pi(C,\kappa)-K(\kappa)
1035: %\right], ~~~~
1036: %\kappa^2=\frac12+\frac{1}{2\sqrt{1+4\ell^2 a^2}},\label{fk1}
1037: %\end{align}
1038: %\begin{align}
1039: %S_{tot}=\sqrt{\lambda}+2\sqrt{\lambda} \left[1+4\ell^2 a^2\right]^{1/4}
1040: %\left[-E(\kappa)+(1-\kappa^2)K(\kappa)\right]\label{Sk1}.
1041: %\end{align}
1042:
1043: Recall that $S_{tot}$ is identified with $-\log V_k$ in \eqref{logV}.
1044: The first term in $S_{tot}$ is interpreted as the
1045: contribution from the denominator in \eqref{Vk}. The second term
1046: in \eqref{fk1} is
1047: completely the same as the one in the correlator of
1048: two fundamental Wilson loops,
1049: see \cite{Drukker:2005cu}.
1050: If one takes further the limit $\th\to 0$
1051: in the above expression, it reproduces the results of
1052: \cite{Zarembo:1999bu,Olesen:2000ji, Kim:2001td}.
1053:
1054:
1055:
1056:
1057:
1058:
1059:
1060:
1061:
1062:
1063:
1064:
1065:
1066:
1067: \subsection{Limit to anti-parallel lines}\label{sec-coulomb}
1068:
1069: When $R_1=R_2 \gg L $, the Wilson loop correlator
1070: reduces to the anti-parallel lines. In this case, one expects that the
1071: correlator exhibits the Coulomb's law as
1072: \begin{align}
1073: -\log V_{k}(R_1,R_2,L ,\th)\cong 2\pi R_2 \frac{M}{L },
1074: \end{align}
1075: where $M$ is a constant determined by $k$ and $\th$.
1076: In terms of $\ell$ and $m$, this limit is
1077: realized via $\ell a \to \infty$ with $\gamma=\frac{m^2}{2\ell a}$ fixed.
1078: The explicit form of the coefficient $M$ is written down in appendix
1079: \ref{app-coulomb}.
1080:
1081:
1082:
1083:
1084:
1085:
1086:
1087: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1088: \subsection*{Acknowledgements}
1089: We would like to thank Changhyun Ahn, Yosuke Imamura, Feng-Li Lin,
1090: Yutaka Matsuo, Nikita Nekrasov, Soo-Jong Rey and
1091: Jung-Tay Yee for useful discussions and comments. The work of S.Y. was
1092: supported in part by the European Research Training Network contract
1093: 005104 ``ForcesUniverse.''
1094: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1095:
1096:
1097:
1098:
1099: \appendix
1100: \section{Elliptic integrals}\label{app-a}
1101:
1102: \subsection{Definitions and some formulas of the standard elliptic integrals}
1103: \begin{align}
1104: &F(\vp,k):=\int_0^{\sin \vp} dz \frac{1}{\sqrt{(1-z^2)(1-k^2z^2)}}
1105: ,\\
1106: &E(\vp,k):=\int_0^{\sin \vp} dz \sqrt{\frac{1-k^2 z^2}{1-z^2}}
1107: ,\\
1108: &\Pi(\vp,C,k):=\int_0^{\sin\vp}dz
1109: \frac{1}{(1-C z^2)\sqrt{(1-z^2)(1-k^2 z^2)}}.
1110: \end{align}
1111: Complete elliptic integrals are
1112: \begin{align}
1113: K(k):=F(\pi/2,k),\qquad
1114: E(k):=E(\pi/2,k),\qquad
1115: \Pi(C,k):=\Pi(\pi/2,C,k).
1116: \end{align}
1117: When $C=k^2$, the third elliptic integral is related to the second one as
1118: \begin{align}
1119: \Pi(\vp,k^2,k)=\frac{1}{1-k^2}E(\vp,k)-\frac{k^2}{1-k^2}\frac{\sin\vp\cos\vp}{\sqrt{1-k^2\sin^2\vp}}.
1120: \end{align}
1121: Let $a,b,c,d$ be four real constants which satisfies $a<b<c<d$. We assume the real variable $u$ satisfies $b\le u \le c$.
1122: If two variable $u$ and $z$ are related as
1123: \begin{align}
1124: z^2:=\frac{(c-a)(u-b)}{(c-b)(u-a)},\quad
1125: \text{or} \quad
1126: u=\frac{(c-a)b-(c-b)a z^2}{(c-a)-(c-b)z^2},
1127: \end{align}
1128: there is a 1-form relation
1129: \begin{align}
1130: &\frac{du}{\sqrt{(u-a)(u-b)(u-c)(u-d)}}
1131: =\frac{2}{\sqrt{(c-a)(d-b)}}\frac{dz}{\sqrt{(1-z^2)(1-k^2z^2)}},\\
1132: &k^2:=\frac{(c-b)(d-a)}{(c-a)(d-b)}.
1133: \end{align}
1134: We obtain the following formula.
1135: \begin{align}
1136: & \int_{b}^{v}\frac{du}{\sqrt{(u-a)(u-b)(u-c)(u-d)}}
1137: =\frac{2}{\sqrt{(c-a)(d-b)}} F(\vp,k),\\
1138: & \sin\vp:=\frac{(c-a)(v-b)}{(c-b)(v-a)}.
1139: \end{align}
1140: \begin{align}
1141: &\int_{b}^{v}du\sqrt{\frac{(u-b)}{(u-a)(u-c)(u-d)}}
1142: =\frac{2(b-a)}{\sqrt{(c-a)(d-b)}} \left[\Pi(\vp,C,k)-F(\vp,k)\right],\\
1143: &C=\frac{c-b}{c-a}.
1144: \end{align}
1145: Let $\delta$ be a small number.
1146: \begin{align}
1147: \int_{b+\delta}^{v}
1148: &\frac{1}{(u-b)\sqrt{(u-a)(u-b)(u-c)(u-d)}}
1149: =\nonumber\\
1150: &\frac{2}{\sqrt{(c-a)(d-b)}}
1151: \Bigg\{
1152: \sqrt{\frac{c-a}{(c-b)(b-a)}}\frac{1}{\sqrt{\delta}}
1153: -\frac{(c-a)}{(c-b)(b-a)} \cot \vp\sqrt{1-k^2\sin^2 \vp}
1154: \nonumber\\&\qquad\qquad
1155: -\frac{(c-a)}{(c-b)(b-a)}E(\vp,k)
1156: +\frac{1}{c-b}F(\vp,k)
1157: \Bigg\}+(\text{terms vanishing when } \delta \to 0).
1158: \end{align}
1159: In order to derive this formula, the relation
1160: \begin{align}
1161: \frac{1}{z^2\sqrt{(1-z^2)(1-k^2z^2)}}
1162: =\frac{d}{dz}\left[\frac{-1}{z}\sqrt{(1-z^2)(1-k^2z^2)}\right]
1163: +\sqrt{\frac{1-k^2z^2}{1-z^2}}
1164: +\sqrt{(1-z^2)(1-k^2z^2)}
1165: \end{align}
1166: is useful.
1167:
1168:
1169:
1170:
1171:
1172:
1173:
1174: \subsection{Expressions of some integrals in terms of the standard elliptic integrals}
1175: \label{app-elliptic}
1176: \begin{align}
1177: &u=\sin^2 \psi,\qquad v=\sin^2 \phi,\qquad
1178: \beta_{\pm}=\frac{-(m^2+1)\pm \sqrt{(m^2+1)^2+4\ell^2 a^2}}{2\ell^2 a^2},
1179: \end{align}
1180: where $\beta_{-}<0\le v\le \beta_{+}<1$, and
1181: \begin{align}
1182: &\frac{d\psi}{\sqrt{\cos^2\psi-m^2\sin^2\psi-\ell^2 a^2\sin^4\psi}}
1183: =\frac{du}{2\ell a\sqrt{(u-\beta_{-})u(u-\beta_{+})(u-1)}}.
1184: \end{align}
1185: \begin{align}
1186: \kappa:=\sqrt{\frac{\beta_{+}(1-\beta_{-})}{\beta_{+}-\beta_{-}}},\qquad
1187: C:=\frac{\beta_{+}}{\beta_{+}-\beta_{-}},\qquad
1188: \chi(\phi):=\sin^{-1}\sqrt{\frac{(\beta_{+}-\beta_{-})\sin^2\phi}{\beta_{+}
1189: (\sin^2\phi-\beta_{-})}}.\label{defi}
1190: \end{align}
1191: \begin{align}
1192: \begin{aligned}
1193: &\int_{0}^{\phi}d\psi\frac{\ell a \sin^2\psi d\psi}
1194: {\sqrt{\cos^2\psi-m^2\sin^2\psi-\ell^2 a^2\sin^4\psi}}
1195: =\frac12 \int_{0}^{v}du\sqrt{\frac{u}{(u-\beta_{-})(u-\beta_{+})(u-1)}}\\
1196: &=\frac{(m^2+1)+\sqrt{(m^2+1)^2+4\ell^2 a^2}}{
1197: 2\ell a \left[(m^2+1)^2+4\ell^2 a^2\right]^{1/4}}
1198: \left[
1199: \Pi(\chi(\phi),C,\kappa)-F(\chi(\phi),\kappa)
1200: \right].
1201: \end{aligned}
1202: \end{align}
1203: \begin{align}
1204: \begin{aligned}
1205: &\int_{0}^{\phi}d\psi\frac{m d\psi}{\sqrt{\cos^2\psi-m^2\sin^2\psi-\ell^2 a^2\sin^4\psi}}
1206: =\frac1{2\ell a} \int_{0}^{v}du\frac{m}{\sqrt{(u-\beta_{-})u(u-\beta_{+})(u-1)}}\\
1207: &=m[(m^2+1)^2+4\ell^2 a^2]^{-1/4}
1208: F(\chi(\phi),\kappa).
1209: \end{aligned}
1210: \end{align}
1211: \begin{align}
1212: \begin{aligned}
1213: & \int_{\frac{\epsilon}{R_2}}^{\phi}d\psi\frac{\cot^2 \psi d\psi}
1214: {\sqrt{\cos^2\psi-m^2\sin^2\psi-\ell^2 a^2\sin^4\psi}}
1215: =\frac1{2\ell a} \int_{\frac{\epsilon^2}{R_2^2}}^{v}du\frac{1-u}{u\sqrt{(u-\beta_{-})u(u-\beta_{+})(u-1)}}\\
1216: &=\frac{R_2}{\epsilon}+[(m^2+1)^2+4\ell^2 a^2]^{1/4}
1217: \left[-\cot\chi(\phi)\sqrt{1-\kappa^2\sin^2\chi(\phi)}-E(\chi(\phi),\kappa)
1218: +(1-\kappa^2)F(\chi(\phi),\kappa)\right].
1219: \end{aligned}
1220: \end{align}
1221: When $\phi=\phi_0$, each integral becomes a complete elliptic integral, i.e.
1222: \begin{align}
1223: \int_{0}^{\phi_0}d\psi\frac{\ell a \sin^2\psi d\psi}{\sqrt{\cos^2\psi-m^2\sin^2\psi-\ell^2 a^2\sin^4\psi}}
1224: &=\frac{(m^2+1)+\sqrt{(m^2+1)^2+4\ell^2 a^2}}{
1225: 2\ell a \left[(m^2+1)^2+4\ell^2 a^2\right]^{1/4}}
1226: \left[
1227: \Pi(C,\kappa)-K(\kappa)
1228: \right],
1229: \label{Appi0}
1230: \end{align}
1231: \begin{align}
1232: \int_{0}^{\phi_0}d\psi\frac{m d\psi}{\sqrt{\cos^2\psi-m^2\sin^2\psi-\ell^2 a^2\sin^4\psi}}
1233: &=m[(m^2+1)^2+4\ell^2 a^2]^{-1/4}
1234: K(\kappa),
1235: \label{Appi1}
1236: \end{align}
1237: \begin{align}
1238: & \int_{\frac{\epsilon}{R_2}}^{\phi_0}d\psi\frac{\cot^2 \psi d\psi}
1239: {\sqrt{\cos^2\psi-m^2\sin^2\psi-\ell^2 a^2\sin^4\psi}}
1240: &=\frac{R_2}{\epsilon}+[(m^2+1)^2+4\ell^2 a^2]^{1/4}
1241: \left[-E(\kappa)
1242: +(1-\kappa^2)K(\kappa)\right].
1243: \label{Appi2}
1244: \end{align}
1245:
1246:
1247:
1248:
1249:
1250:
1251:
1252:
1253:
1254:
1255:
1256:
1257:
1258:
1259:
1260:
1261:
1262:
1263:
1264:
1265:
1266:
1267: \section{Explicit form of the coulomb coefficient}
1268: \label{app-coulomb}
1269: Here we write the explicit form mentioned in section \ref{sec-coulomb}.
1270:
1271: \begin{align}
1272: \begin{aligned}
1273: \beta_{\pm}&=\frac{-\gamma \pm \sqrt{\gamma^2+1}}{\ell a},\qquad
1274: \kappa^2=C=\frac{-\gamma+\sqrt{\gamma^2+1}}{2\sqrt{\gamma^2+1}},\qquad
1275: \sin^2\phi_1=\frac{-\gamma + \sqrt{\gamma^2+\sin^2\theta_k}}{\ell a},\\
1276: \chi(\phi_1)&=\sin^{-1}\sqrt{
1277: \frac{2(-\gamma+\sqrt{\gamma^2+\sin^2\theta_k})\sqrt{\gamma^2+1}}{
1278: (-\gamma+\sqrt{\gamma^2+1})(\sqrt{\gamma^2+\sin^2\theta_k}+\sqrt{\gamma^2+1})}}
1279: ,\qquad
1280: \eta=1+\frac12 \left(\frac{L }{R_2}\right)^2.
1281: \end{aligned}
1282: \label{anti}
1283: \end{align}
1284:
1285: \subsection{$\theta_k >\pi/2$}
1286: $\gamma$ is related to $\th$ by
1287: \begin{align}
1288: \th-\theta_k=\frac{\sqrt{\gamma}}{(\gamma^2+1)^{1/4}}
1289: F(\chi(\phi_1),\kappa).
1290: \end{align}
1291: The potential is
1292: \begin{align}
1293: S_{tot}=2\pi R_2 \frac{M}{L },
1294: \end{align}
1295: where the coefficient can be written as
1296: \begin{align}
1297: M=&\frac{1}{2\pi}\frac{\gamma+\sqrt{\gamma^2+1}}{\sqrt{2}(\gamma^2+1)^{1/4}}
1298: \left[\Pi(\chi(\phi_1),\kappa^2,\kappa)-F(\chi(\phi_1),\kappa)\right]\nonumber\\
1299: &\times\sqrt{\lambda}\Bigg[
1300: -\frac{\cos \theta_k}{\sqrt{-\gamma+\sqrt{\gamma^2+\sin^2\theta_k}}}
1301: \nonumber\\
1302: &
1303: +\sqrt{2}(\gamma^2+1)^{1/4}\left\{
1304: -\cot\chi(\phi_1)\sqrt{1-\kappa^2\sin^2\chi(\phi_1)}
1305: -E(\chi(\phi_1),\kappa)+(1-\kappa^2)F(\chi(\phi_1),\kappa)
1306: \right\}\Bigg].
1307: \end{align}
1308:
1309:
1310: \subsection{$\theta_k<\pi/2$}
1311:
1312: \begin{align}
1313: &\th-\theta_k=\frac{\sqrt{\gamma}}{(\gamma^2+1)^{1/4}}
1314: \left[2K(\kappa)-F(\chi(\phi_1),\kappa)\right] .
1315: \end{align}
1316:
1317:
1318: \begin{align}
1319: M=&\frac{1}{2\pi}\frac{\gamma+\sqrt{\gamma^2+1}}{\sqrt{2}(\gamma^2+1)^{1/4}}
1320: \left[
1321: 2\Pi(\kappa^2,\kappa)-\Pi(\chi(\phi_1),\kappa^2,\kappa)
1322: -2K(\kappa)+F(\chi(\phi_1),\kappa)
1323: \right]\nonumber\\
1324: &\times\sqrt{\lambda}\Bigg[
1325: -\frac{\cos \theta_k}{\sqrt{-\gamma+\sqrt{\gamma^2+\sin^2\theta_k}}}
1326: +\sqrt{2}(\gamma^2+1)^{1/4}\Big\{
1327: \cot\chi(\phi_1)\sqrt{1-\kappa^2\sin^2\chi(\phi_1)}
1328: \nonumber\\
1329: &
1330: -2E(\kappa)
1331: +E(\chi(\phi_1),\kappa)
1332: +2(1-\kappa^2)K(\kappa)
1333: -(1-\kappa^2)F(\chi(\phi_1),\kappa)
1334: \Big\}\Bigg].
1335: \end{align}
1336:
1337:
1338:
1339: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1340: \begin{thebibliography}{10}
1341:
1342: \bibitem{Maldacena:1997re}
1343: J.~M. Maldacena, {\it ``The large {N} limit of superconformal field theories
1344: and supergravity,''} Adv. Theor. Math. Phys. {\bf 2} (1998) 231--252,
1345: [\href{http://arXiv.org/abs/hep-th/9711200}{{\tt hep-th/9711200}}].
1346:
1347: \bibitem{Rey:1998ik}
1348: S.-J. Rey and J.-T. Yee, {\it ``Macroscopic strings as heavy quarks in large
1349: {N} gauge theory and anti-de {S}itter supergravity,''} Eur. Phys. J. {\bf
1350: C22} (2001) 379--394, [\href{http://arXiv.org/abs/hep-th/9803001}{{\tt
1351: hep-th/9803001}}].
1352:
1353: \bibitem{Maldacena:1998im}
1354: J.~M. Maldacena, {\it ``{W}ilson loops in large {N} field theories,''} Phys.
1355: Rev. Lett. {\bf 80} (1998) 4859--4862,
1356: [\href{http://arXiv.org/abs/hep-th/9803002}{{\tt hep-th/9803002}}].
1357:
1358: \bibitem{Drukker:2005kx}
1359: N.~Drukker and B.~Fiol, {\it ``All-genus calculation of {W}ilson loops using
1360: {D}-branes,''} JHEP {\bf 02} (2005) 010,
1361: [\href{http://arXiv.org/abs/hep-th/0501109}{{\tt hep-th/0501109}}].
1362:
1363: \bibitem{Hartnoll:2006hr}
1364: S.~A. Hartnoll and S.~Prem~Kumar, {\it ``Multiply wound {P}olyakov loops at
1365: strong coupling,''} Phys. Rev. {\bf D74} (2006) 026001,
1366: [\href{http://arXiv.org/abs/hep-th/0603190}{{\tt hep-th/0603190}}].
1367:
1368: \bibitem{Yamaguchi:2006tq}
1369: S.~Yamaguchi, {\it ``{W}ilson loops of anti-symmetric representation and
1370: {D5}-branes,''} JHEP {\bf 05} (2006) 037,
1371: [\href{http://arXiv.org/abs/hep-th/0603208}{{\tt hep-th/0603208}}].
1372:
1373: \bibitem{Gomis:2006sb}
1374: J.~Gomis and F.~Passerini, {\it ``Holographic {W}ilson loops,''} JHEP {\bf 08}
1375: (2006) 074, [\href{http://arXiv.org/abs/hep-th/0604007}{{\tt
1376: hep-th/0604007}}].
1377:
1378: \bibitem{Erickson:2000af}
1379: J.~K. Erickson, G.~W. Semenoff and K.~Zarembo, {\it ``Wilson loops in {N = 4}
1380: supersymmetric {Y}ang-{M}ills theory,''} Nucl. Phys. {\bf B582} (2000)
1381: 155--175, [\href{http://arXiv.org/abs/hep-th/0003055}{{\tt hep-th/0003055}}].
1382:
1383: \bibitem{Drukker:2000rr}
1384: N.~Drukker and D.~J. Gross, {\it ``An exact prediction of {N = 4} {SUSYM}
1385: theory for string theory,''} J. Math. Phys. {\bf 42} (2001) 2896--2914,
1386: [\href{http://arXiv.org/abs/hep-th/0010274}{{\tt hep-th/0010274}}].
1387:
1388: \bibitem{Rodriguez-Gomez:2006zz}
1389: D.~Rodriguez-Gomez, {\it ``Computing {W}ilson lines with dielectric branes,''}
1390: Nucl. Phys. {\bf B752} (2006) 316--326,
1391: [\href{http://arXiv.org/abs/hep-th/0604031}{{\tt hep-th/0604031}}].
1392:
1393: \bibitem{Okuyama:2006jc}
1394: K.~Okuyama and G.~W. Semenoff, {\it ``Wilson loops in {N} = 4 {SYM} and fermion
1395: droplets,''} JHEP {\bf 06} (2006) 057,
1396: [\href{http://arXiv.org/abs/hep-th/0604209}{{\tt hep-th/0604209}}].
1397:
1398: \bibitem{Hartnoll:2006is}
1399: S.~A. Hartnoll and S.~P. Kumar, {\it ``Higher rank {W}ilson loops from a matrix
1400: model,''} JHEP {\bf 08} (2006) 026,
1401: [\href{http://arXiv.org/abs/hep-th/0605027}{{\tt hep-th/0605027}}].
1402:
1403: \bibitem{Hartnoll:2006ib}
1404: S.~A. Hartnoll, {\it ``Two universal results for {W}ilson loops at strong
1405: coupling,''} \href{http://arXiv.org/abs/hep-th/0606178}{{\tt
1406: hep-th/0606178}}.
1407:
1408: \bibitem{Chen:2006iu}
1409: B.~Chen and W.~He, {\it ``On 1/2-{BPS} {W}ilson-'t {H}ooft loops,''}
1410: \href{http://arXiv.org/abs/hep-th/0607024}{{\tt hep-th/0607024}}.
1411:
1412: \bibitem{Giombi:2006de}
1413: S.~Giombi, R.~Ricci and D.~Trancanelli, {\it ``Operator product expansion of
1414: higher rank {W}ilson loops from {D}-branes and matrix models,''}
1415: \href{http://arXiv.org/abs/hep-th/0608077}{{\tt hep-th/0608077}}.
1416:
1417: \bibitem{Gross:1998gk}
1418: D.~J. Gross and H.~Ooguri, {\it ``Aspects of large {N} gauge theory dynamics as
1419: seen by string theory,''} Phys. Rev. {\bf D58} (1998) 106002,
1420: [\href{http://arXiv.org/abs/hep-th/9805129}{{\tt hep-th/9805129}}].
1421:
1422: \bibitem{Zarembo:1999bu}
1423: K.~Zarembo, {\it ``Wilson loop correlator in the {AdS/CFT} correspondence,''}
1424: Phys. Lett. {\bf B459} (1999) 527--534,
1425: [\href{http://arXiv.org/abs/hep-th/9904149}{{\tt hep-th/9904149}}].
1426:
1427: \bibitem{Olesen:2000ji}
1428: P.~Olesen and K.~Zarembo, {\it ``Phase transition in {W}ilson loop correlator
1429: from {AdS/CFT} correspondence,''}
1430: \href{http://arXiv.org/abs/hep-th/0009210}{{\tt hep-th/0009210}}.
1431:
1432: \bibitem{Kim:2001td}
1433: H.~Kim, D.~K. Park, S.~Tamarian and H.~J.~W. Muller-Kirsten, {\it
1434: ``{G}ross-{O}oguri phase transition at zero and finite temperature: Two
1435: circular {W}ilson loop case,''} JHEP {\bf 03} (2001) 003,
1436: [\href{http://arXiv.org/abs/hep-th/0101235}{{\tt hep-th/0101235}}].
1437:
1438: \bibitem{Arutyunov:2001hs}
1439: G.~Arutyunov, J.~Plefka and M.~Staudacher, {\it ``Limiting geometries of two
1440: circular {M}aldacena-{W}ilson loop operators,''} JHEP {\bf 12} (2001) 014,
1441: [\href{http://arXiv.org/abs/hep-th/0111290}{{\tt hep-th/0111290}}].
1442:
1443: \bibitem{Drukker:2005cu}
1444: N.~Drukker and B.~Fiol, {\it ``On the integrability of {W}ilson loops in
1445: {$AdS_5\times S^5$}: Some periodic ansatze,''} JHEP {\bf 01} (2006) 056,
1446: [\href{http://arXiv.org/abs/hep-th/0506058}{{\tt hep-th/0506058}}].
1447:
1448: \bibitem{Tsuji:2006zn}
1449: A.~Tsuji, {\it ``Holography of {W}ilson loop correlator and spinning
1450: strings,''} \href{http://arXiv.org/abs/hep-th/0606030}{{\tt
1451: hep-th/0606030}}.
1452:
1453: \bibitem{Ahn:2006px}
1454: C.~Ahn, {\it ``Two circular {W}ilson loops and marginal deformations,''}
1455: \href{http://arXiv.org/abs/hep-th/0606073}{{\tt hep-th/0606073}}.
1456:
1457: \bibitem{Zarembo:2001jp}
1458: K.~Zarembo, {\it ``String breaking from ladder diagrams in {SYM} theory,''}
1459: JHEP {\bf 03} (2001) 042, [\href{http://arXiv.org/abs/hep-th/0103058}{{\tt
1460: hep-th/0103058}}].
1461:
1462: \bibitem{Plefka:2001bu}
1463: J.~Plefka and M.~Staudacher, {\it ``Two loops to two loops in {N} = 4
1464: supersymmetric {Y}ang-{M}ills theory,''} JHEP {\bf 09} (2001) 031,
1465: [\href{http://arXiv.org/abs/hep-th/0108182}{{\tt hep-th/0108182}}].
1466:
1467: \bibitem{Yamaguchi:2006te}
1468: S.~Yamaguchi, {\it ``Bubbling geometries for half {BPS} {W}ilson lines,''}
1469: \href{http://arXiv.org/abs/hep-th/0601089}{{\tt hep-th/0601089}}.
1470:
1471: \bibitem{Lunin:2006xr}
1472: O.~Lunin, {\it ``On gravitational description of {W}ilson lines,''} JHEP {\bf
1473: 06} (2006) 026, [\href{http://arXiv.org/abs/hep-th/0604133}{{\tt
1474: hep-th/0604133}}].
1475:
1476: \bibitem{Berenstein:1998ij}
1477: D.~Berenstein, R.~Corrado, W.~Fischler and J.~M. Maldacena, {\it ``The operator
1478: product expansion for {W}ilson loops and surfaces in the large {N} limit,''}
1479: Phys. Rev. {\bf D59} (1999) 105023,
1480: [\href{http://arXiv.org/abs/hep-th/9809188}{{\tt hep-th/9809188}}].
1481:
1482: \bibitem{Pawelczyk:2000hy}
1483: J.~Pawelczyk and S.-J. Rey, {\it ``{R}amond-{R}amond flux stabilization of
1484: {D}-branes,''} Phys. Lett. {\bf B493} (2000) 395--401,
1485: [\href{http://arXiv.org/abs/hep-th/0007154}{{\tt hep-th/0007154}}].
1486:
1487: \bibitem{Camino:2001at}
1488: J.~M. Camino, A.~Paredes and A.~V. Ramallo, {\it ``Stable wrapped branes,''}
1489: JHEP {\bf 05} (2001) 011, [\href{http://arXiv.org/abs/hep-th/0104082}{{\tt
1490: hep-th/0104082}}].
1491:
1492: \bibitem{Drukker:1999zq}
1493: N.~Drukker, D.~J. Gross and H.~Ooguri, {\it ``{W}ilson loops and minimal
1494: surfaces,''} Phys. Rev. {\bf D60} (1999) 125006,
1495: [\href{http://arXiv.org/abs/hep-th/9904191}{{\tt hep-th/9904191}}].
1496:
1497: \end{thebibliography}
1498: \end{document}
1499: