hep-th0405102/b.tex
1:    \documentclass[12pt]{article}
2:         %
3:    \usepackage{amssymb}
4:    \usepackage{amsmath}
5:    \usepackage{subfigure}
6:          \usepackage{epsfig}
7:           %
8:          \textwidth 15.5cm \hoffset  -1.2cm%
9:          \textheight 20.8cm %
10:          \topmargin -0.5cm %
11:          \baselineskip=17pt
12:          \renewcommand{\theequation}{{\protect \thesection.\arabic{equation}}}
13: 
14:          \def\href#1#2{#2}
15:          %
16:          \def\IP{\relax{\rm I\kern-.18em P}}
17:          \newcommand{\beq}{\begin{equation}} 
18:          \newcommand{\eeq}{\end{equation}}
19:          \newcommand{\beqa}{\begin{eqnarray}}
20:          \newcommand{\eeqa}{\end{eqnarray}}
21:          \def\tQ{{\hat Q}} 
22:          \def\bj{{\bar \jmath}} 
23:          \def\cA{{\cal A}}
24:          \def\cV{{\cal V}}
25:          \def\cH{{\cal H}}
26:          \def\cJ{{\cal J}} 
27:          \def\vn{{\vec{n}}}\def\mod{{\rm mod}}
28:          \def\CN{{\cal N}}
29:          \def\QN{\mathbb{N}}
30:          \def\QC{\mathbb{C}}
31:          \def\QH{\mathbb{H}}
32:          \def\QM{\mathbb{M}}
33:          \def\QR{\mathbb{R}}
34:          \def\QX{\mathbb{X}}
35:          \def\QZ{\mathbb{Z}}
36:          \def\QP{\mathbb{P}}
37:          \def\S{{\cal S}} 
38:          \def\orb{{\mbox{\rm \scriptsize orb}}}
39:          \def\even{{\mbox{\rm \scriptsize even}}}
40:          \def\dt{\delta t}
41:          \def\dr{\delta \rho}
42:          \def\df{\delta \phi}
43:          \def\snoll{sinh(\rho_0)}
44:          \def\cnoll{cosh(\rho_0)}
45:          %%%%%%%%%%%%%%%%%%%%%%  COMMANDS %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
46:          %%% Equations
47:          \def\be{ \begin{equation}}\def\ee{ \end{equation}}
48:          \def\ba{ \begin{eqnarray}}\def\ea{ \end{eqnarray}}
49:          \renewcommand{\theequation}{{\thesection.\arabic{equation}}}
50:          \def\nn{\nonumber}   \def\ssstyle{\scriptscriptstyle}
51:          %%% Sections
52:          %\renewcommand{\section}[1]{\section #1 \setcounter{equation}{0}}
53:          %%% Special Symbols
54:          \def\End{{\mbox{\it End\/}}}
55:          \def\C{\mathbb{C}} \def\Z{\mathbb{Z}}
56:    \def\R{\mathbb{R}}\def\T{\mathbb{T}}
57:          \def\o{\otimes} \def\bz{\mbox{\bf z}}
58:          \def\bN{{\sf \bf N}}
59:          \def\cedille#1{\setbox0=\hbox{#1}\ifdim\ht0=1ex \accent'30 #1%
60:           \else{\ooalign{\hidewidth\char'30\hidewidth\crcr\unbox0}}\fi}
61:          \def\gaw{Gaw\cedille edzki}
62:          %%% Calligraphic Letters 
63:          \def\p{\partial}
64:          \def\Tr{\mbox{\rm tr}}
65:          \def\Ad{\mbox{\rm Ad}} \def\ad{\mbox{\rm ad}} 
66:          \def\mathC{\C}
67:          \def\mathR{\R}
68:          \def\V{{\cal V}}
69:          \def\W{{\cal W}}   \def\M{{\cal M}}   \def\G{{\cal G}} 
70:          %%% Greek Letters
71:          \def\a{\alpha }\def\b{\beta }  \def\c{\gamma }   
72:          \def\d{\delta }\def\e{\epsilon}\def\ve{\varepsilon}
73:          \def\k{\kappa} \def\s{\sigma}
74:          \def\ew{\hspace*{-1mm}}   \def\ppe{\hspace*{-2.5mm}}
75:          \def\bb{\hspace*{-1mm}}
76:          \newcommand{\note}[1]{\raisebox{1ex}{{\footnotesize \sf #1}}}
77:          \newcommand{\rnote}[1]{\raisebox{1ex}{{\hspace*{-3mm} \scriptsize\sf#1}}
78:             \hspace*{-4mm}}
79:          \newcommand{\Fus}[6]{F_{{\scriptstyle #1},{\scriptstyle #2}}
80:            \hspace*{.3mm}\displaystyle{[} \ew \begin{array}{ll} {\scriptstyle #3 }
81:            \ppe & {\scriptstyle #4} \ppe \\[-2mm] {\scriptstyle #5}\ppe &
82:            {\scriptstyle #6}\ew \end{array}\displaystyle{]}}
83:          \newcommand{\CG}[6]{\displaystyle{[} \,\ew \begin{array}{lll} 
84:            {\scriptstyle #1} \ppe
85:            & {\scriptstyle #2} \ppe & {\scriptstyle #3} \ew \\[-2mm]
86:          {\scriptstyle
87:            #4} \ppe & {\scriptstyle #5}\ppe & {\scriptstyle #6}
88:    \ew\end{array}
89:            \displaystyle{]}}
90:          \newcommand{\Br}[6]{B_{{\scriptstyle #1}{\scriptstyle #2}}
91:            \hspace*{.3mm}\displaystyle{[} \ew \begin{array}{ll} {\scriptstyle #3 }
92:            \ppe & {\scriptstyle #4} \ppe \\[-2mm] {\scriptstyle #5}\ppe &
93:            {\scriptstyle #6}\ew \end{array}\displaystyle{]}}
94:          \newcommand{\SJS}[6]{ \displaystyle{\{ } \ew \begin{array}{lll} 
95:            {\scriptstyle #1} \ppe  &
96:            {\scriptstyle #2} \ppe & {\scriptstyle #3}
97:            \ppe \\[-2mm]{\scriptstyle #4}  \ppe & {\scriptstyle #5} \ppe &
98:           {\scriptstyle #6} \ew \end{array} \displaystyle{\} } }
99:          \renewcommand{\vert}[3]{\displaystyle{(} \begin{array}{rcl}
100:           \ppe & {\scriptstyle #1} \ppe & \ppe \\[-2mm] \ew {\scriptstyle
101:    #2} 
102:            \ppe &  \ppe & {\scriptstyle
103:           #3}\ew \end{array} \displaystyle{)}} 
104:          \def\Mat{{\mbox{\rm Mat}}}
105:          \def\sq{{\sf q}}
106:          \def\ik{{\sf k}}%\def\ik{{\rm k}}
107:          \def\min{{\mbox{\rm min\/}}}
108:          \def\cH{{\cal H}}
109:          \def\Im{{\mbox{\rm Im}}}
110:          \def\nno{\newline\noindent}
111:          \def\id{{\rm id}}
112:          \def\tr{{\rm tr}}
113:          \def\cS{{\cal S}}
114:          \def\tF{{\rm F}}
115:          \def\tCS{{\rm CS}}
116:          \def\tf{{f}}
117:          \def\tL{{\rm L}}
118:          \def\cC{{\cal C}}
119: 
120:          \def\cN{{N}}
121:          \def\itY{{\rm Y}} 
122:          \def\tA{{\rm A}} 
123:          \def\astk{\, , \, } 
124:          %%%%%%%%%%%%%%%%%%%%%%  PARAMETERS %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
125:          \newtheorem{theo}{Theorem} 
126:          \newtheorem{defn}{Definition}
127:          \newtheorem{ex}{Example}   
128:          \newtheorem{rem}{Definition and Lemma}
129:          \newtheorem{prop}[theo]{Proposition}
130:          \newtheorem{coro}[theo]{Corollary}
131:          %%%%%%%%%%%%%%%%%%%%%%  TITLEPAGE %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
132:          %%%%%%%%%%%%%%%%%%%%%%  DOCUMENT %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
133:          \begin{document}
134: 
135: 
136: {}~ \hfill\vbox{\hbox{hep-th/0405102}
137: \hbox{UUITP-14/04}
138: }\break
139: 
140: \vskip 1.cm
141: 
142: \centerline{\large \bf Pulsating Strings on $AdS_5 \times S^5$}
143: \vspace*{1.5ex}
144: 
145: 
146: \vspace*{4.0ex}
147: 
148: \centerline{\large \rm M. Smedb\"ack\footnote{
149: mikael.smedback@teorfys.uu.se}}
150: \vspace*{2.5ex}
151: \centerline{\large \it Department of Theoretical Physics}
152: \centerline{\large \it Box 803, SE-751 08 Uppsala, Sweden}
153: \vspace*{3.0ex}
154: 
155: \centerline{\large July 12, 2004}
156: \vspace*{3.0ex}
157: 
158: \vspace*{4.5ex}
159: \medskip
160: \bigskip\bigskip
161: \centerline {\bf Abstract}
162: 
163: We find the anomalous dimension and the conserved charges of an R-charged 
164: string 
165: pulsating on $AdS_5$.
166: The analysis is performed both on the gauge and string side,
167: where we find agreement at the one-loop level.
168: Furthermore, the solution is shown to be related by analytic continuation 
169: to a string which is pulsating on $S^5$, thus providing an example of
170: the close relationship between the respective isometry groups.
171:          
172: \bigskip
173: 
174: \vfill \eject
175: \baselineskip=17pt
176: 
177: 
178: 
179:          %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
180: 
181: %         \tableofcontents
182: %         \newpage
183: 
184:          \section{Introduction}\label{intro}
185: The AdS/CFT conjecture \cite{AdS1,AdS2,AdS3}
186: has lead to a better understanding of both conformal gauge theories 
187: as well as string theory in curved spaces.
188: Within this framework, the seminal work of \cite{BMN} included a discussion of
189: operators of the form Tr$Z^{J_1}W^{J_2}+\cdots$ (built up from the scalars
190: $Z$ and $W$ of the $\mathcal{N}=4$ SYM supermultiplet) 
191: where $J_1 \ll J_2$. 
192: The dots indicate other permutations of the fields $Z$ and $W$ inside
193: the trace, and in general these states mix among themselves under scaling; 
194: only certain linear combinations are eigenstates to the scaling
195: operator.
196: Semiclassical string configurations which usually go beyond
197: the BMN limit (e.g. by taking both $J_1$ and $J_2$ to be large)
198: have since been studied extensively
199: \cite{semiclass1,semiclass2,semiclass3,semiclass4,semiclass5,
200: semiclass6,semiclass7,semiclass8,semiclass9,semiclass10,
201: semiclass11,semiclass12,semiclass14,semiclass15}
202: (see also \cite{scalso1,scalso2,scalso3,scalso4}),
203: and are 
204: reviewed in
205: \cite{semiclass13}.
206: 
207: The observation of \cite{YM1} that the matrix of anomalous 
208: dimensions
209: could be mapped to an integrable Bethe spin chain \cite{bethe}
210: simplified and extended the studies of the corresponding gauge theory  
211: \cite{YM2,YM3,YM4,YM7,YM8,YM9,YM10,YM12,YM13,YM15,YM17,YM18}
212: (see also \cite{YMalso1,YMalso2}),
213: reviewed in \cite{faddeev}. The original results of \cite{YM1} were
214: restricted to the group $SO(6)$ at 1-loop level, but were later
215: extended to the full 1-loop $SU(2,2|4)$ chain \cite{YM5,YM6},
216: taking advantage of previous results on integrability in QCD
217: amplitudes \cite{QCD1,QCD2} and the QCD dilatation operator
218: \cite{QCDdilop1,QCDdilop2,QCDdilop3}
219: (see also \cite{QCDmixed1,QCDmixed2,QCDmixed3,
220: QCDintegrab1,QCDintegrab2}).
221: Progress on higher orders in closed subsectors has also been made 
222: \cite{YM2extra,YM11,YM14,YM16,priv}.
223: 
224: The integrable spin chain formulation exposes the conserved charges.
225: Conserved charges in the sigma model were first discussed in
226: \cite{HCC1,HCC2} (see also \cite{HCC3,HCC4}).
227: Progress on relating the conserved charges on either side
228: to each other by viewing (subsectors of) both sides of the duality as 
229: an integrable system was made in \cite{HCC5,HCC6,HCC7,HCC8} 
230: (see also \cite{Roiban}). 
231: The work on finding descriptions of the AdS/CFT duality
232: in terms of integrable systems
233: are reviewed in \cite{review}.
234: 
235: In this paper, we will analyse a string pulsating on $AdS_5$ and 
236: whose centre of mass is revolving on $S^5$, 
237: both from the gauge and string side (assuming
238: large quantum numbers).
239: From the AdS/CFT conjecture, we expect that the anomalous dimension
240: of the corresponding operator will coincide with the first order energy
241: correction on the string side.
242: Another motivation for studying this configuration is that 
243: the conserved charges on either side
244: of the duality can be matched explicitly using integrability.
245: 
246: A third motivation
247: is that our solution will be shown to
248: be related by an analytic continuation to the solution of \cite{YM9}
249: for a string pulsating and revolving on $S^5$. This
250: provides
251: an example of the close mathematical relationship between
252: the isometry groups of $AdS_5$ and $S^5$; $SO(4,2)$ and $SO(6)$,
253: respectively.
254: Such relations were discussed in \cite{YM7}, where a first example
255: was given; a string rotating in two planes on $S^5$ was shown to
256: be related by analytic continuation to a string whose centre of mass
257: is revolving in one plane
258: on $S^5$ and with one spin in $AdS_5$.
259: One may speculate that tying together seemingly different solutions
260: in this way may help in providing a bridge between
261: duality checks at the level of individual solutions and higher-level
262: checks. An example of the latter is the recent analysis of the
263: duality at the level of actions \cite{al1,al2,al3,al4,al5,al6}.
264: 
265: We will analyse the case at hand from the gauge side and string side
266: in sections \ref{gauge} and \ref{string}, respectively.
267: In section \ref{conserved} we exhibit the conserved charges on the
268: string side. Our conclusions are presented in section \ref{conclusions}.
269: 
270: %%%%%%%%%%%%%%%%%%%%%%%%%%%
271: 
272: \section{Gauge Side}\label{gauge}
273: \setcounter{equation}{0}
274: 
275: In this section, 
276: we will consider operators of the form
277: Tr$(D\bar{D})^B Z^J$, which are charged under $SO(2,2)$.
278: Here, $D \equiv D_1 +iD_2$,
279: (where $D_i$ are covariant derivatives) and $Z$ is one of the three
280: complex scalars of the $\mathcal{N}=4$ supermultiplet.
281: Individual operators are formed by 
282: linear combinations of different distributions
283: of the $D$'s and $\bar{D}$'s over the $Z$'s.
284: In general 
285: mixing occurs under scaling within the full $SO(4,2)$.
286: However, in the semiclassical limit it turns out that it
287: {\it will} be sufficient
288: to consider the bosonic subgroup $SO(2,2)$ \cite{priv},
289: cf. what happens in the $SO(6)$ case \cite{YM9,HCC8}.
290: The mapping of the matrix of anomalous
291: dimensions to a Hamiltonian of a spin chain will then allow us to
292: find the the eigenvalues of the diagonalized system.
293: 
294: The simple roots of $SO(2,2)$ are
295: $\overrightarrow{\alpha}_1 = (1,1)$ and
296: $\overrightarrow{\alpha}_2 = (1,-1)$.
297: In the infinite-dimensional representation of highest weight
298: $\overrightarrow{w} = (-1,0)$,
299: the Bethe equations are
300: \begin{equation}
301:   \left( \frac{u_{q,i}+i\overrightarrow{\alpha}_q \cdot \overrightarrow{w}/2}
302:               {u_{q,i}-i\overrightarrow{\alpha}_q \cdot \overrightarrow{w}/2}
303:   \right)^L =
304:   \prod_{j \neq i}^{n_q} 
305:     \frac{u_{q,i}-u_{q,j}+i\overrightarrow{\alpha}_q \cdot \overrightarrow{\alpha}_q/2}
306:          {u_{q,i}-u_{q,j}-i\overrightarrow{\alpha}_q \cdot \overrightarrow{\alpha}_q/2}
307:   \prod_{q' \neq q} \prod_j^{n_{q'}}
308:     \frac{u_{q,i}-u_{q',j}+i\overrightarrow{\alpha}_q \cdot \overrightarrow{\alpha}_q'/2}
309:          {u_{q,i}-u_{q',j}-i\overrightarrow{\alpha}_q \cdot \overrightarrow{\alpha}_q'/2},
310: \end{equation}
311: as written in \cite{YM1} for an arbitrary Lie group, and
312: \begin{equation}\label{bethe}
313:   \begin{split}
314:     \left( \frac{u_{1,i}-i/2}{u_{1,i}+i/2} \right)^L
315:     & = \prod_{j \neq i}^{n_1} \frac{u_{1,i}-u_{1,j}+i}{u_{1,i}-u_{1,j}-i} \\
316:     \left( \frac{u_{2,i}-i/2}{u_{2,i}+i/2} \right)^L
317:     & = \prod_{j \neq i}^{n_2} \frac{u_{2,i}-u_{2,j}+i}{u_{2,i}-u_{2,j}-i}
318:   \end{split}
319: \end{equation}
320: for $SO(2,2)$.
321: The anomalous dimension is
322: \begin{equation}\label{andim}
323:   \gamma = \frac{\lambda}{8\pi^2} \left( 
324:             \sum_{i=1}^{n_1} \frac{1}{u_{1,i}^2 + 1/4}
325:            +\sum_{i=1}^{n_2} \frac{1}{u_{2,i}^2 + 1/4} \right) .
326: \end{equation}
327: As indicated, there are $n_q$ roots of the type $q$.
328: The form of the operator we are looking for is 
329: Tr$(D\bar{D})^B Z^J$,
330: so the number of sites is $L=J$. 
331: The two root types essentially correspond to creation of 
332: $D$'s and $\bar{D}$'s,
333: respectively, so we set $n \equiv n_1 = n_2 = B$.
334: 
335: Assuming that the number of roots is large (so that they can
336: be approximated by a continuous distribution) 
337: in the thermodynamic 
338: limit (i.e. a large number of sites $L$)
339: the log of the Bethe equation for the first type of
340: root (after a rescaling $u \rightarrow uL$) is
341: \begin{equation}\label{thermo}
342:   \frac{2}{\alpha} \left( -\frac{1}{u} + 2\pi m \right) =
343:   2 -\hspace{-0.48cm}\int_C \frac{\sigma(u')du'}{u-u'},
344: \end{equation}
345: where the line through the integral sign 
346: indicates that the singularity at $u'=u$ is resolved by taking
347: the principal value of the integral.
348: The contour $C$ is defined by the support of the root density $\sigma(u')$
349: and its endpoints are $a$ and $b$.
350: We have defined $\alpha \equiv n/L$. The integer $m$
351: corresponds to different branches of the $\log$. 
352: The root density
353: is normalized as
354: \begin{equation}\label{norm}
355:   \int_C \sigma(u') du' = 2.
356: \end{equation}
357: Reading (\ref{thermo}) as a force balancing equation, we conclude
358: that the roots are repelled from each other but attracted to the point
359: $u=1/2 \pi m$. We therefore expect that the roots will spread out along
360: the contour $C$ passing through this point.
361: 
362: Performing an inverse Hilbert transform on (\ref{thermo}), the
363: root density is
364: \begin{equation}
365:   \sigma(u) = - \frac{1}{\pi^2 \alpha} \left[ (u-a)(u-b) \right]^{1/2}
366:     -\hspace{-0.48cm}\int_C du' \left( \frac{1}{u'} - 2\pi m \right) \frac{1}{u'-u}
367:     \frac{1}{ \left[(u'-a)(u'-b) \right]^{1/2} }.
368: \end{equation}
369: The multivalued function $[\cdots]^{1/2}$ has a cut along the segment
370: of the real axis coinciding with the contour $C$.
371: Calculating the integral by deforming the contour and picking
372: up the residue at $u'=0$, we get
373: \begin{equation}\label{sigma}
374:   \sigma(u) = - \frac{i}{\pi \alpha u \sqrt{ab}} 
375:                 \left[ (u-a)(u-b) \right]^{1/2}.
376: \end{equation}
377: 
378: The endpoints $a$ and $b$ of the contour $C$ are determined by inserting
379: (\ref{sigma}) into equations (\ref{thermo}) and (\ref{norm}).
380: This results in the two equations
381: \begin{equation}
382:   \begin{split}
383:     \sqrt{ab} & = \frac{1}{2\pi m} \\
384:         a + b & = \frac{1+2 \alpha}{\pi m}
385:   \end{split}.
386: \end{equation}
387: In particular, this means that for non-negative (i.e. physical) values
388: of $\alpha$, the endpoints of the contour will lie on the positive
389: real axis (for positive $m$) and the contour will pass
390: through the point $u=1/2 \pi m$, as expected.
391: 
392: Now define the resolvent
393: \begin{equation}
394:   W(u') \equiv \int_C   du \frac{\sigma(u)}{u'-u}.
395: \end{equation}
396: By deforming the contour and picking up residues at $u=0$, $u=u'$
397: and $u=\infty$, the resolvent becomes \cite{YM9,HCC6}
398: \begin{equation}\label{resolvent}
399:   -\alpha W(u') = \frac{1}{u'} \left[ 1 -   
400:     \sqrt{(1-2\pi mu')^2-2\alpha(4\pi mu')} \right] + \pi m.
401: \end{equation}
402: The square root denotes the branch which coincides with the principal
403: branch for small $u'$.
404: One of the virtues of the resolvent is that it determines the
405: first part of the
406: anomalous dimension (\ref{andim}) in the thermodynamic limit:
407: \begin{equation}\label{anom}
408:   \gamma_1 = -\frac{\lambda \alpha}{16 \pi^2 L} W'(0)
409: \end{equation}
410: Inserting (\ref{resolvent}) into (\ref{anom}), we get
411: \begin{equation}
412:   \gamma_1 = + \frac{\lambda m^2}{2L} \alpha (1 + \alpha).
413: \end{equation}
414: According to (\ref{bethe}), the two root types behave symmetrically
415: and do not interact. Due to the trace condition, which in this case
416: takes the form
417: \begin{equation}\label{trace}
418:   \prod_{j=1}^{j=L} 
419:     \frac{(u_{1,j}+i/2)(u_{2,j}+i/2)}{(u_{1,j}-i/2)(u_{2,j}-i/2)} = 1,
420: \end{equation}
421: the second type of roots spread out along along
422: the segment $[-b,-a]$ of the negative real axis, 
423: so (\ref{andim}) becomes
424: \begin{equation}
425:   \gamma = \gamma_1 + \gamma_2 = 2\gamma_1 
426:          = + \frac{\lambda m^2}{L} \alpha (1 + \alpha)
427:          = + \frac{\lambda m^2 B}{J^2} \left( 1 + \frac{B}{J} \right).
428: \end{equation}
429: The conformal dimension is $\Delta = 2B+J$, in terms of which
430: \begin{equation}\label{anomdim}
431:   \gamma = m^2 \lambda \frac{\Delta^2-J^2}{4J^3}.
432: \end{equation}
433: 
434: 
435: %%%%%%%%%%%%%%%%%%%%%%%%%%%%
436: 
437: 
438:          \section{String Side}\label{string}
439:          \setcounter{equation}{0}
440: In \cite{semiclass4}, a string pulsating on $AdS_5$ or $S^5$ was considered.
441: In \cite{YM9}, the latter configuration 
442: was generalized to include a rotation
443: in one plane on $S^5$.
444: In this section, we will consider the closely related 
445: sigma model description of
446: a string pulsating on $AdS_5$ and whose centre of mass is
447: revolving on $S^5$.
448: Restricting the motion to the subspace $AdS_3 \times S^1$
449: means that the isometry group contains a factor isomorphic to
450: $SO(2,2)$, hence matching the set of operators considered in
451: section \ref{gauge}.
452: 
453: The metric on $AdS_5 \times S^5$ will be written
454: \begin{equation}
455:   \begin{split}
456:   ds^2_{AdS_5} & = d\rho^2 -\cosh^2 \rho dt^2 + \sinh^2 \rho
457:          ( d \theta^2 + \cos^2 \theta d\Phi_1^2 + \sin^2 \theta d\Phi_2^2 ) \\
458:   ds^2_{S^5}   & = d \gamma^2 + \cos^2 \gamma d\phi_3^2 + \sin^2 \gamma
459:          ( d\Psi^2 + \cos^2 \Psi d\phi_1^2 + \sin^2 \Psi d\phi_2^2)
460:   \end{split}.
461: \end{equation}
462: We will use the notation $\phi \equiv \phi_3$.
463: The relevant metric for the $AdS_3 \times S^1$ subspace of the full space is
464: \begin{equation}
465:   ds^2 = d\phi^2 + d\rho^2 - \cosh^2 \rho dt^2 + \sinh^2 \rho d\theta^2.
466: \end{equation}
467: We assume that the string is wrapped around the azimuthal angle
468: $\theta$ on $AdS_3$. We then use the ansatz
469: \begin{equation}
470:   \phi=\phi(\tau), \rho=\rho(\tau), t=\tau, \theta=m \sigma.
471: \end{equation}
472: The integer $m$ allows for multi-wrapping.
473: We will consider $t$ and $\theta$ to be gauge fixed.
474: The Nambu-Goto action is
475: \begin{equation}
476:     S = -m\sqrt{\lambda} \int dt \sinh \rho 
477:         \sqrt{\cosh^2 \rho -\dot{\phi}^2 - \dot{\rho}^2}.
478: \end{equation}
479: The energy $\pi_t \propto H$ and the spin $\pi_\phi$ are conserved.
480: The dynamical momenta are
481: \begin{equation}
482:     \pi_\rho = \frac{m \sqrt{\lambda}\dot{\rho}\sinh \rho}
483:                      {\sqrt{\cosh^2 \rho -\dot{\phi}^2 - \dot{\rho}^2}}
484: \end{equation}
485: and
486: \begin{equation}
487:     \pi_\phi = \frac{m \sqrt{\lambda}\dot{\phi}\sinh \rho}
488:                      {\sqrt{\cosh^2 \rho -\dot{\phi}^2 - \dot{\rho}^2}}.
489: \end{equation}
490: The Hamiltonian is then given by
491: \begin{equation}
492:       H^2 = (\pi_\rho \dot{\rho} + \pi_{\phi} \dot{\phi} - L)^2 
493:           = \cosh^2 \rho (\pi_\rho^2 + \pi_\phi^2 + m^2 \lambda \sinh^2 \rho).
494: \end{equation}
495: Following \cite{YM9}, we now consider the term $V(\rho) = m^2 \lambda \cosh^2 \rho \sinh^2 \rho$ 
496: to be a perturbation.
497: A Hermitian form of the unperturbed Hamiltonian operator acting on a 
498: wave function is then
499: $\hat{H}_0^2 \Psi(\rho) = 
500:   \cosh \rho (\hat{\pi}_\rho^2 + \hat{\pi}_\phi^2) \cosh \rho \Psi(\rho)$, i.e.
501: \begin{equation}\label{startekv}
502:   \Delta^2 \Psi(\rho) = -(\cosh \rho) \nabla^2 (\cosh \rho) \Psi(\rho)
503:                       + J(J+4) \cosh^2 \rho \Psi(\rho),
504: \end{equation}
505: where
506: \begin{equation}
507:   \nabla^2 = \frac{1}{\sinh^3 \rho \cosh \rho} 
508:              \frac{\partial}{\partial \rho}
509:              \sinh^3 \rho \cosh \rho \frac{\partial}{\partial \rho}.
510: \end{equation}
511: $J$ and $\Delta$ are non-negative integers.
512: Introducing $x=\frac{1}{\cosh^2 \rho}$ and restricting to even
513: integers $J=2j$, $\Delta=2a$ transforms (\ref{startekv}) into
514: \begin{equation}
515:   -\frac{x^{7/2}}{1-x} \frac{d}{dx} \frac{(1-x)^2}{x} \frac{d}{dx}
516:     \frac{1}{x^{1/2}} \Psi(x) + j(j+2) \Psi(x) - a^2 \Psi(x) = 0.
517: \end{equation}
518: The power series ansatz
519: \begin{equation}\label{ansatz}
520:   \Psi(x) = \sum_{\lambda=0}^{\lambda=\infty} a_\lambda x^{k+\lambda}
521: \end{equation}
522: results in an indicial equation with two solutions. The solution
523: which keeps all terms finite on the interval $0 \le x \le 1$ is
524: $k = j+5/2 $. Hence, the recursion relation becomes
525: \begin{equation}
526:   a_\lambda = - a_{\lambda-1} \frac{(a+1+\lambda+j) (a-1-\lambda-j)}
527:                                    {\lambda (\lambda+2+2j)},
528: \end{equation}
529: whose solution is
530: \begin{equation}
531:   a_p = (-1)^p
532:   \left(\hspace{-0.20cm}
533:   \begin{array}{cc}
534:     a-j-2 \\
535:     p
536:   \end{array}
537:   \hspace{-0.20cm} \right)
538:   \frac{(a+j+1+p)!}{(p+2j+2)!}.
539: \end{equation}
540: Inserting this into (\ref{ansatz}) gives the wave functions,
541: whose normalized forms are
542: \begin{equation}
543:   \Psi(x) = \frac{2\sqrt{a(a-j-1)}}{(a-j-1)!\sqrt{a+j+1}}
544:     \frac{1}{x^{j-1/2}} \left( \frac{d}{dx} \right)^{a-j-1}x^{a+j+1}
545:     (1-x)^{a-j-2}.
546: \end{equation}
547: The first order correction to the energy is
548: \begin{equation}
549:   \Delta E^2 = \int d\tau \Psi(x) V(x) \Psi(x)
550:              = \frac{2a(a+j+1)(a-j-1)}{(j+1)(2j+1)(2j+3)},
551: \end{equation}
552: where $d\tau$ is the volume element.
553: The energy to first order and for large quantum numbers is then
554: $E=\Delta+\gamma$, where
555: \begin{equation}\label{energycorr}
556:   \gamma = m^2 \lambda \frac{\Delta^2-J^2}{4J^3}.
557:  \end{equation}
558: This agrees with the expected anomalous dimension (\ref{anomdim}).
559: 
560: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
561: 
562: 
563: \section{Conserved Charges}\label{conserved}
564: 
565: On the gauge side, the mapping of the matrix of anomalous
566: dimensions to a Hamiltonian of an integrable spin chain
567: immediately provides all conserved charges
568: in terms of
569: the resolvent. Following \cite{HCC6}, 
570: the recent paper \cite{HCC7} demonstrates
571: that for the R-charge assignment $(J_1,J_2,J_2)$ on $S^5$,
572: the charges on the string side precisely match those on
573: the gauge side (at the 1-loop level).
574: In our case, the 1-loop resolvent 
575: is (\ref{resolvent}). On the string side, the
576: corresponding generator is essentially given by the
577: quasi-momentum, as discussed in \cite{HCC8}. In this section,
578: we will follow the procedure outlined in \cite{HCC8} to
579: exhibit the quasi-momentum for the case at hand.
580: 
581: We are assuming that the string is moving on an $AdS_3 \times S^1$
582: subspace. The $AdS_3$ space can be described as the hypersurface
583: $-X_1^2 -X_2^2 + X_3^2 + X_4^2 = 1$ in $\mathbb{R}^4$. Defining $W \equiv X_1 + iX_2$
584: and $Z \equiv X_3 +iX_4$, this space can be equivalently described
585: as an $SU(1,1)$ group manifold using the map
586: \begin{equation}
587:   \left(
588:   \begin{array}{cc}
589:     Z       & W        \\
590:     \bar{W} & \bar{Z}  \\
591:   \end{array}
592:   \right)
593:   = g \in SU(1,1).
594: \end{equation}
595: As an ansatz for the string pulsating on $AdS_3$ and revolving on $S^1$
596: we use
597: \begin{equation}
598:   \begin{array}{cc}
599:     W & = \sinh \rho e^{i \theta} \\
600:     Z & = \cosh \rho e^{i t}
601:   \end{array}.
602: \end{equation}
603: In this section we will use the Polyakov action in unit gauge,
604: \begin{equation}\label{action}
605:   \begin{array}{cc}
606:     S & = \frac{\sqrt{\lambda}}{4\pi} \int d\sigma d\tau 
607:         \left[
608:           \partial Z \partial \bar{Z} - \partial W \partial \bar{W} 
609:           -(\partial X_5)^2
610:         \right] \\
611:       & = - \frac{\sqrt{\lambda}}{4\pi} \int d\sigma d\tau
612:           \left[
613:             \frac{1}{2} \mbox{Tr} (g^{-1}\partial_\alpha g)^2
614:             +(\partial X_5)^2
615:           \right]
616:   \end{array}
617: \end{equation}
618: (hence choosing $X_3$ and $X_4$ to be time-like).
619: In this description, we will no longer consider $t(\tau)$ to be gauge
620: fixed.
621: By the equations of motion, $\phi \equiv X_5 = Q \tau$.
622: The action is invariant under constant shifts along the
623: circle, so $Q\sqrt{\lambda} \equiv J$ is the conserved charge corresponding to the spin.
624: 
625: 
626: The action (\ref{action}) is also invariant under constant left and right
627: shifts of the group elements, $g \rightarrow hg$ and $g \rightarrow gh$.
628: The corresponding charge is the energy 
629: \begin{equation}
630:   E = +\sqrt{\lambda} Q_l = -\sqrt{\lambda}Q_r = -\sqrt{\lambda} \dot{t} \cosh^2 \rho.
631: \end{equation}
632: In the following, we will restrict our considerations to the
633: time $\tau$ when $\rho(\tau)=t(\tau)=0$. Then
634: \begin{equation}
635:     \frac{E^2}{\lambda} = \dot{t}^2 = \dot{\rho}^2 + Q^2,
636: \end{equation}
637: where the last equality follows from the constraint corresponding
638: to fixing the gauge in the Polyakov action.
639: 
640: Defining $\partial_{\pm} \equiv \partial_\tau \pm \partial_\sigma$
641: and currents $j_{\pm} \equiv g^{-1} \partial_{\pm} g$, it follows from the
642: constraint $Z\bar{Z}-W \bar{W}=\det(g)=1$ that
643: \begin{equation}
644:   0 = \partial_+ j_- - \partial_- j_+ + [j_+,j_-].
645: \end{equation}
646: This coincides with the consistency condition $[L,M]=0$ for the linear
647: problem $L\Psi = M\Psi = 0$, where
648: \begin{equation}
649:   \begin{array}{cc}
650:     L & = \partial_\sigma +\frac{1}{2} 
651:         \left( \frac{j_+}{1-x} - \frac{j_-}{1+x} \right) \\
652:     M & = \partial_\tau +\frac{1}{2} 
653:         \left( \frac{j_+}{1-x} + \frac{j_-}{1+x} \right)
654:   \end{array}.
655: \end{equation}
656: Explicitly, the first equation is
657: \begin{equation}\label{sigmaprob}
658:   \partial_\sigma \Psi = \frac{x}{x^2-1}
659:   \left(
660:   \begin{array}{cc}
661:     iQ_l                   & \dot{\rho}e^{im\sigma} \\
662:     \dot{\rho}e^{-im\sigma} & -iQ_l
663:   \end{array}
664:   \right)
665:   \Psi
666: \end{equation}
667: Considering $\Psi$ to be a vector of the type
668: \begin{equation}
669:   \Psi = \left(
670:   \begin{array}{cc}
671:     A e^{ip_{+} \sigma/2\pi} \\
672:     B e^{ip_{-} \sigma/2\pi}
673:   \end{array}
674:   \right)
675: \end{equation}
676: provides a family of solutions\footnote{
677: The ansatz $p_\pm(x)=a(x) \pm \pi m$ is helpful.} 
678: to (\ref{sigmaprob}), provided that
679: the condition
680: \begin{equation}
681:   \left[ \frac{x^2-1}{2\pi} (p_{\pm} \mp \pi m) \right]^2
682:   + (x\dot{\rho})^2
683:   = \left[ \frac{m(x^2-1)}{2} - xQ_l \right]^2
684: \end{equation}
685: is satisfied.
686: Consequently, each root $p_\pm(x)$ will be double-valued.
687: Subtracting the poles from one of the sheets of $p_-$
688: (with a branch cut along the positive real axis),
689: the resolvent is
690: \begin{equation}
691:   G(x) = \frac{2\pi}{x^2-1} \left[
692:          -xQ + \left\{ \left[ \frac{m}{2}(x^2-1)-xQ_l \right]^2 - (x\dot{\rho})^2 \right\}^{1/2}
693:          \right]
694:          -\pi m.
695: \end{equation}
696: Rescaling $x \rightarrow 4\pi Qx$, the leading contribution for large
697: quantum numbers ($\lambda \rightarrow 0$ in $Q=J/\sqrt{\lambda}$ and
698: $Q_l = E/\sqrt{\lambda}$) is
699: \begin{equation}\label{resolvent_string}
700:   -G_0(x) = \frac{1}{2x} \left[
701:            1-\sqrt{(1-2\pi mx)^2 - 2\alpha (4\pi mx)}
702:            \right] + \pi m.
703: \end{equation}
704: The square root denotes the branch which coincides with the principal
705: branch for small $x$.
706: It is proportional to the resolvent (\ref{resolvent}) on the gauge theory
707: side.
708: Since the charges are generated by the odd part of the resolvent,
709: this shows that the charges on the gauge and string side match.
710: 
711: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
712: 
713: 
714: \section{Conclusions}\label{conclusions}
715: 
716: We considered a string pulsating on $AdS_5$ and revolving on $S^5$.
717: The anomalous dimension (\ref{anomdim}) agrees with the first
718: order energy correction (\ref{energycorr}), as expected from
719: the AdS/CFT conjecture. In terms of 
720: $\alpha \equiv \frac{n}{J} = \frac{\Delta-J}{2J}$, these
721: results become
722: \begin{equation}\label{energycorr}
723:   \gamma = m^2 \lambda \frac{1}{J} \alpha (1+\alpha).
724: \end{equation}
725: 
726: Consider analytically continuing this result to the unphysical region
727: $\alpha < 0$ by $\Delta \rightarrow - J_1$ and
728: $J \rightarrow -L$. This takes 
729: %$-2\alpha \rightarrow \alpha_{EMZ} \equiv 1-\frac{J_1}{L}$, i.e.
730: $\alpha \rightarrow -\alpha_{EMZ} \equiv \frac{J_1-L}{2L}$, i.e.
731: \begin{equation}\label{energycorr}
732:   \gamma \rightarrow m^2 \lambda \frac{1}{L} \alpha_{EMZ} (1-\alpha_{EMZ}).
733: \end{equation}
734: This is the result\footnote{
735: Our definition of $\alpha_{EMZ}$ differs by a factor of 2 from that of
736: \cite{YM9}.}
737: of \cite{YM9} for a string pulsating and revolving
738: on $S^5$, which together with our result provides a complete
739: description for all real values of $\alpha$; for $\alpha < 0$,
740: the string pulsates and revolves on $S^5$. As $\alpha$ is turned
741: to positive values\footnote{
742: Note that we assumed large quantum numbers.
743: As $J \rightarrow 0$, 
744: the thermodynamic limit is no longer valid.
745: The behaviour of the anomalous dimension 
746: in the strong coupling region is discussed in \cite{semiclass4}.
747: A similar phenomenon occurs in \cite{YM7}.
748: }, 
749: the string starts pulsating on $AdS_5$ instead
750: (while still revolving on $S^5$).
751: On the gauge side, the corresponding operator forms are
752: Tr$(Z\bar{Z})^{(L-J_1)/2} X^{J_1}$ and
753: Tr$(D\bar{D})^{(\Delta-J)/2} Z^J$, respectively.
754: 
755: This description can be compared to the extension of \cite{YM7}
756: to the results of \cite{YM4}. 
757: In \cite{YM4}, operators of the form Tr$Z^{J_1}W^{J_2}$ were considered, 
758: corresponding to strings rotating in two planes on $S^5$.
759: It was shown in \cite{YM7} that the replacements
760: $J_1 + J_2 \rightarrow - J$, $J_2 \rightarrow S$ and 
761: $\gamma \rightarrow - \gamma$ turned the system 
762: of Bethe equations\footnote{
763: In the present case and 
764: on the level of Bethe equations in the thermodynamic limit, 
765: taking $\alpha \rightarrow -\alpha_{EMZ}$ turns (\ref{thermo})
766: into the corresponding equation in \cite{YM9}.
767: }
768: and anomalous dimensions
769: into a description
770: of operators of the form Tr$D^S Z^J$, corresponding to strings
771: rotating both on $AdS_5$ and $S^5$.
772: 
773: Let us also mention that the presence of an integrable structure on both 
774: sides of the duality
775: is manifested in our case by the agreement of the corresponding generators of 
776: conserved charges, (\ref{resolvent}) and (\ref{resolvent_string}).
777: For the string pulsating on $S^5$, considered in \cite{YM9}, 
778: the corresponding check was carried out in \cite{HCC8}.
779: 
780: %-----------------------------
781: 
782: 
783:          \bigskip
784:          \noindent
785:         {\bf Acknowledgments:}
786:          I am very grateful to J.\ Minahan for helpful discussions
787:          and support during the course of this work, and for
788:          useful comments
789:          on the manuscript. I would also like to thank  
790:          J.\ Engquist, L.\ Freyhult and K.\ Zarembo for conversations.
791: 
792: 
793: %%%%%%%%%%%%%%%%%%%%
794: 
795: %        I am very grateful to J.\ Minahan
796: %        for several enlightening discussions and many useful comments on the
797: %        manuscript. I would also like to thank
798: %        L.\ Freyhult, J.\ Gregory, F.\ Kristiansson, U.\ Lindstr\"om
799: %        and V.\ Schomerus for conversations. 
800: 
801: 
802: 
803: %         \begin{thebibliography}{99}
804:            \footnotesize
805: 
806:            \bibliographystyle{unsrt}
807:            \bibliography{references_rev2}
808:            
809: %          \bibitem{superspinchain}
810: %            N.\ Beisert, M.\ Staudacher,
811: %            {\sl The N=4 SYM Integrable Super Spin Chain}, 
812: %            hep-th/0307042%
813: 
814: %         \bibitem{spectroscopy}
815: %            N.\ Beisert, S.\ Frolov, M.\ Staudacher, A.\ Tseytlin,
816: %            {\sl Precision Spectroscopy of AdS/CFT}, 
817: %            hep-th/0308117%
818: 
819: %         \end{thebibliography}
820: 
821: 
822:          \end{document}
823: