1: \documentclass[nohyper,12pt,letterpaper]{JHEP3}
2: \usepackage{epsfig}
3:
4: \author{Robert de Mello Koch$^{1,3}$, Jeff Murugan$^{2,4}$, Jelena Smolic$^{1}$
5: and Milena Smolic$^{1}$\\
6: ${}^{1}$Department of Physics and Centre for Theoretical Physics,\\
7: University of the Witwatersrand, Wits, 2050, South Africa\\
8: \qquad\\
9: ${}^{2}$Department of Physics, Brown University\\
10: Providence, RI02912 , USA \\
11: \qquad\\
12: ${}^{3}$Stellenbosch Institute for Advanced Studies,\\
13: Stellenbosch, South Africa\\
14: \qquad\\
15: ${}^4$Department of Mathematics and Applied Mathematics,\\
16: University of Cape Town, Private Bag, Rondebosch 7700, South Africa\\
17: \qquad\\
18: E-mail: \email{robert@neo.phys.wits.ac.za, jeff@now.het.brown.edu, smolicj@science.pg.wits.ac.za, msmolic@webmail.co.za}}
19:
20:
21: \abstract{
22: In this article we study a pp-wave limit of the Lunin-Maldacena background. We show that
23: the relevant string theory background is a homogeneous pp-wave. We obtain the string
24: spectrum. The dual field theory is a deformation of
25: ${\cal N}=4$ super Yang-Mills theory. We have shown that, for a class of operators, at
26: $O(g_{YM}^2)$ and at leading order in $N$, all contributions to the anomalous dimension
27: come from $F$-terms. We are able to identify the operator in the
28: deformed super Yang-Mills which is dual to the lowest string mode. By studying the
29: undeformed theory we are able to provide some evidence, directly in the field theory,
30: that a small set of nearly protected operators decouple.
31: We make some comments on operators in the Yang-Mills theory that are dual to excited
32: string modes.}
33:
34: \preprint{Brown-Het-1446\\ Wits-CTP-022}
35:
36: \title{Deformed PP-waves from the Lunin-Maldacena Background}
37:
38: \keywords{AdS/CFT correspondence, BMN Limit, PP Waves}
39:
40: \def \Tr{\mbox{Tr\,}}
41:
42: \begin{document}
43:
44: \section{Introduction}
45: The AdS/CFT correspondence\cite{AdSCFT} relates string theory on
46: negatively curved spacetime and large $N$ quantum field theories.
47: The correspondence is a strong/weak coupling duality in the 't Hooft
48: coupling of the field theory. At large $N$ and large 't Hooft coupling,
49: both quantum gravity and curvature corrections in the string theory
50: are supressed. The dual gauge theory however is strongly coupled.
51: For small 't Hooft coupling and large $N$, the gauge theory coupling
52: is small, but curvature corrections in the string
53: theory are not. Computations that can be carried out on
54: both sides of the correspondence necessarily involve quantities that
55: are not corrected or receive small corrections, allowing weak coupling
56: results to be extrapolated to strong coupling. A very interesting class
57: of observables of this type are the near BPS operators discovered by
58: Berenstein, Maldacena and Nastase\cite{BMN}, which are dual to excited string modes.
59: Indeed, since these operators are not BPS, the BMN limit of the gauge
60: theory reproduces genuinely stringy physics, via the AdS/CFT correspondence.
61:
62: In a recent article\cite{LM}, Lunin and Maldacena (LM) have studied $\beta$-deformations
63: of the ${\cal N}=4$ super Yang-Mills theory, and have identified the
64: corresponding gravitational deformation of the AdS$_5\times$S$^5$ background.
65: The field theory deformation is obtained by
66: making the following replacement in the superpotential
67:
68: \begin{equation}
69: \Tr \left(\hat{\Phi}^1\hat{\Phi}^2\hat{\Phi}^3 -
70: \hat{\Phi}^1\hat{\Phi}^3\hat{\Phi}^2\right)\to
71: \Tr \left(e^{i\pi\gamma}\hat{\Phi}^1\hat{\Phi}^2\hat{\Phi}^3 -
72: e^{-i\pi\gamma}\hat{\Phi}^1\hat{\Phi}^3\hat{\Phi}^2\right).
73: \label{fdef}
74: \end{equation}
75:
76: \noindent
77: The deformed field theory has ${\cal N}=1$ supersymmetry and is invariant
78: under a $U(1)\times U(1)$ non-${\cal R}$ symmetry. The ${\cal N}=4$ theory is dual
79: to string theory in the AdS$_5\times$S$^5$ geometry, which contains a two torus.
80: The isometries of the two torus match with the $U(1)\times U(1)$ field theory symmetry.
81: Denote the metric of this two torus by $g$ and the NS-NS two form (which is of
82: course zero in the undeformed theory) by $B$. The deformation of the dual gravitational
83: theory is obtained by replacing
84:
85: $$\tau =B+i\sqrt{g}\to \tau_\gamma ={\tau\over 1+\gamma\tau}.$$
86:
87: \noindent
88: The AdS$_5$ factor is unchanged which is expected because (\ref{fdef}) is a marginal
89: deformation. Studies of the AdS/CFT correspondence for this deformation are likely
90: to produce interesting results for at least two reasons. Firstly, it is important to
91: generalize the AdS/CFT correspondence to less supersymmetric examples. Secondly, since
92: this background has a continuous adjustable deformation parameter, it may be possible to
93: define new scaling limits.
94:
95: A study of semiclassical string states provided important insights into the BMN
96: limit\cite{GKP}. Motivated by this, semiclassical string states in the LM background
97: were recently compared to a class of gauge theory scalar operators\cite{FRT}.
98: The 1-loop anomalous dimensions of these operators are described by an integrable
99: spin chain and match beautifully with the energies of the semiclassical string states.
100: Further, by employing the Lax pair for strings
101: in the LM background\cite{F}, the Landau-Lifschitz action associated to the one-loop spin
102: chain was recovered. This indicates that the integrable structures in the gauge
103: theory and the string theory match. Further analysis of the relevant spin chain is
104: given in \cite{Beisert}. For further recent insights into the gauge/string
105: correspondence for these (and other) new examples see\cite{BK}.
106:
107: The logic employed by Lunin and Maldacena to obtain the gravitational theory dual to the
108: deformed field theory can be extended in a number of ways. Recently, instead of deforming
109: the ${\cal N}=4$ super Yang-Mills theory, deformations of ${\cal N}=1$ and ${\cal N}=2$
110: theories have been considered\cite{GN}. Further, deformations of eleven
111: dimensional geometries of the form AdS$_4\times$Y$_7$ with $Y_7$ a seven dimensional
112: Sasaki-Einstein\cite{CA},\cite{Lee} or weak $G_2$ or tri-Sasakian\cite{Lee}
113: space have been considered.
114:
115: In this article we are interested in studying a pp-wave limit of the LM background.
116: There are a number of interesting pp-wave limits that can be taken. Each of these limits
117: allows us to probe different {\it stringy} aspects of the correspondence, and are thus
118: worthy of study. One such limit was in fact already considered in \cite{NP}.
119: We will be considering a different pp-wave
120: limit, to provide further independent support for and insight into the correspondence
121: of \cite{LM}.
122:
123: The paper is organized as follows: in section 2 we describe the pp-wave limit that we
124: consider here. We obtain the metric and B field by taking an appropriate limit of the
125: results in \cite{LM}. The resulting background is a homogeneous
126: pp-wave\cite{HPP}. It is known that the
127: string sigma model in this background can be solved exactly\cite{SHPP}. We
128: provide this analysis in
129: section 3. In section 4 we study the dual gauge theory and consider the question of
130: how to define the
131: near-BPS operators with anomalous dimensions which reproduce the spectrum of the string
132: sigma model. We are able to argue that, in the large $N$ limit and at one loop, one can
133: ignore gluon exchange, self energy insertions and $D$-term contributions. This allows
134: a significant simplification of the analysis. We are able to identify the operator dual
135: to the lowest string mode. For $\gamma=0$ we are able to find near-BPS operators which
136: reproduce the spectrum of the string sigma model. Our results are consistent with the
137: expected decoupling of a small set of nearly protected operators.
138: In the $\gamma\ne 0$ case, we are able argue that there is a set of nearly protected
139: operators whose spectrum of anomalous dimensions is independent of $\gamma$ in agreement
140: with the string theory result. We also find near BPS operators for small values of the
141: ${\cal R}$ charge $J$ and find that their anomalous dimension does depend on $\gamma$.
142: Section 5 is reserved for a discussion of our results.
143:
144: \section{PP-wave Limit of the Lunin-Maldacena Geometry}
145:
146: In this section we will take the pp-wave limit of the LM background. Our goal is to
147: obtain the spectrum of free strings in this background. To write down the relevant
148: string sigma model, we need only the metric and the B field. Thus, we do not consider
149: the RR-fluxes $C_2$ and $C_4$ which are also non-zero in the LM background.
150:
151: The metric is\cite{LM}
152:
153: \begin{eqnarray}
154: \nonumber
155: ds^2 &=&R^2\left(
156: -dt^2\cosh^2\rho +d\rho^2 +\sinh^2\rho d\Omega_3^2+
157: \sum_i d\mu_i^2\right.\\
158: &+& \left. G\sum_{i=1}^3\mu_i^2 d\phi_i^2
159: +\gamma^2 \mu_1^2\mu_2^2\mu_3^2 G\left(\sum_i d\phi_i\right)^2
160: \right).
161: \end{eqnarray}
162:
163: \noindent
164: where
165:
166: $$\mu_1=\cos\alpha,\qquad\mu_2 =\sin\alpha\cos\theta,\qquad
167: \mu_3 =\sin\alpha\sin\theta,$$
168: $$G={1\over 1+\gamma^2
169: (\mu_1^2\mu_2^2 +\mu_1^2\mu_3^2 +\mu_3^2\mu_2^2)}.$$
170:
171: \noindent
172: It is useful to use the angles $\psi$, $\varphi_1$ and $\varphi_2$, defined by
173:
174: $$\phi_1 =\psi-\varphi_2,\qquad\phi_2=\psi+\varphi_1+\varphi_2,\qquad
175: \phi_3=\psi-\varphi_1 .$$
176:
177: \noindent
178: The parameter $\gamma$ is the deformation parameter. We will perform the
179: Penrose limit using the null geodesic $\tau =\psi$, with
180: $\alpha_0=\cos^{-1}{1\over\sqrt{3}}$ and $\theta_0={\pi\over 4}.$ We set
181:
182: $$\theta ={\pi\over 4}+\sqrt{2\over 3}{x^1\over R} ,\qquad
183: \alpha =\alpha_0 -{x^2\over R},\qquad
184: \rho={r\over R}$$
185: $$\varphi_1={x^{3\prime}\over R},\qquad
186: \varphi_2={x^{4\prime}\over R},\qquad
187: t=x^+ +{x^- \over R^2},\qquad \psi ={x^- \over R^2}-x^+ , $$
188:
189: $$ x^3=\sqrt{2\over 3+\gamma^2}\left(x^{3\prime}+{1\over 2}x^{4\prime}\right),\qquad
190: x^4=\sqrt{3\over 2(3+\gamma^2)}x^{4\prime},$$
191:
192: \noindent
193: and take the limit $R\to\infty$. The pp-wave metric we obtain is
194:
195: \begin{eqnarray}
196: \nonumber
197: ds^2 &=& -4dx^+ dx^- -\left[r^2+{4\gamma^2\over 3+\gamma^2}
198: \left((x^1)^2 + (x^2)^2\right)\right](dx^+)^2 +dr^2 +r^2 d\Omega_3^2\\
199: &+&(dx^1)^2+(dx^2)^2+ (dx^3)^2 +(dx^4)^2
200: +{4\sqrt{3}\over\sqrt{3+\gamma^2}}\left(x^1dx^3+x^2 dx^4
201: \right) dx^+.
202: \end{eqnarray}
203:
204: To obtain the string sigma model, we will also need the $B$ field in the pp-wave limit.
205: We find
206:
207: $$B_{\varphi_1\varphi_2}{\cal D}\varphi_1\wedge{\cal D}\varphi_2=
208: G\gamma gR^2{\cal D}\varphi_1\wedge{\cal D}\varphi_2,$$
209:
210: \noindent
211: where
212:
213: $${\cal D}\varphi_1=d\varphi_1-d\psi +{3\mu_1^2\mu_2^2\over g}d\psi, \qquad
214: {\cal D}\varphi_2=d\varphi_2-d\psi +{3\mu_3^2\mu_2^2\over g}d\psi.$$
215:
216: \noindent
217: Taking the pp-wave limit as above, we find the following $B$ field
218:
219: $$ B={\gamma\over\sqrt{3}}d x^3\wedge d x^4
220: +{2\gamma\over\sqrt{3+\gamma^2}}
221: (x^2 d x^3\wedge dx^+ + x^1 dx^+\wedge dx^4 ),$$
222:
223: \noindent
224: and the following field strengths
225:
226: $$ H_{23+}={2\gamma\over\sqrt{3+\gamma^2}},
227: \qquad H_{14+}=-{2\gamma\over\sqrt{3+\gamma^2}}.$$
228:
229: \noindent
230: Thus, the field strength is {\it null} as it should be in the pp-wave limit.
231:
232: \section{Strings in the PP-wave Limit of the Lunin-Maldacena Geometry}
233:
234: Given the metric and $B$ fields written down in the previous section, in this
235: section we consider the resulting string sigma model. We show that this background
236: corresponds to a homogeneous pp-wave\cite{HPP} and are thus able to use existing
237: results\cite{SHPP} to obtain the spectrum.
238:
239: We will be working in lightcone gauge. The string worldsheet action is
240: (we are dropping the fermions from our analysis)
241: $$S=-{1\over 4\pi\alpha'}\int d^2\sigma \left[\sqrt{\eta}\eta^{ab}g_{\mu\nu}
242: \partial_a X^\mu\partial_b X^\nu +\epsilon^{ab}B^{NS}_{\mu\nu}\partial_a X^\mu
243: \partial_b X^\nu +\alpha'\sqrt{\eta}\phi(x)R\right],$$
244:
245: \noindent
246: with $R$ the scalar curvature on the worldsheet, $\eta^{ab}$ is the worldsheet
247: metric and $\eta=|\det {\eta_{ab}}|$.
248: We will choose $\sqrt{\eta}\eta^{ab}$ diagonal with $\sqrt{\eta}\eta^{00}=-1$ and
249: $\sqrt{\eta}\eta^{11}=1$. After shifting
250:
251: $$ x^- \to x^-+{\sqrt{3}\over 2\sqrt{3+\gamma^2}}
252: (x^1x^3+x^2x^4),$$
253:
254: \noindent
255: the metric becomes
256:
257: \begin{eqnarray}
258: ds^2 &=&
259: -4dx^+ dx^- -\left[\sum_{i=5}^8(x^i)^2+{4\gamma^2\over 3+\gamma^2}
260: \left((x^1)^2 + (x^2)^2\right)
261: \right](dx^+)^2 \nonumber\\
262: &+&\sum_{i=1}^8 (dx^i)^2
263: +{2\sqrt{3}\over\sqrt{3+\gamma^2}}\left(x^1dx^3 -x^3dx^1 +x^2 dx^4-x^4dx^2
264: \right) dx^+ .\nonumber
265: \end{eqnarray}
266:
267: \noindent
268: This metric corresponds to a homogeneous pp-wave\cite{HPP}. The sigma model for this
269: background has been considered in \cite{SHPP}; we will review the relevant results here.
270: In the gauge $x^+ =\tau$, we obtain the following Lagrangian density (we take $\sigma$ to
271: run from $0$ to $\pi$ and set $\alpha'={1\over 2\pi}$)\footnote{We thank
272: T. Mateos for pointing out an error in the next formula, which appeared in an
273: earlier version of this draft.}
274:
275: \begin{eqnarray}
276: {\cal L}
277: &=&-2{\partial x^-\over\partial \tau}
278: -{1\over 2}\left[\sum_{i=5}^8(x^i)^2+{4\gamma^2\over 3+\gamma^2}
279: ((x^1)^2+(x^2)^2)\right]+{2\gamma\over\sqrt{3+\gamma^2}}\left(
280: x^2{\partial x^3\over\partial\sigma}-x^1{\partial x^4\over\partial\sigma}\right)
281: \nonumber \\
282: \nonumber
283: &+&{\sqrt{3}\over \sqrt{3+\gamma^2}}\left(
284: x^1{\partial x^3\over\partial\tau}-x^3{\partial x^1\over\partial\tau}+
285: x^2{\partial x^4\over\partial\tau}-x^4{\partial x^2\over\partial\tau}
286: \right)-{1\over 2}\sum_{i =1}^8\partial_a x^i\partial^a x^i .
287: \end{eqnarray}
288:
289: \noindent
290: To quantize the theory, compute the canonical momenta
291:
292: $$p^1(\tau,\sigma)=\dot{x}^1 (\tau,\sigma)
293: -\sqrt{3 \over 3+\gamma^2}x^3,\qquad
294: p^2(\tau,\sigma)=\dot{x}^2 (\tau,\sigma)
295: -\sqrt{3 \over 3+\gamma^2}x^4,$$
296: $$p^3(\tau,\sigma)=\dot{x}^3 (\tau,\sigma)
297: +\sqrt{3 \over 3+\gamma^2}x^1,\qquad
298: p^4(\tau,\sigma)=\dot{x}^4 (\tau,\sigma)
299: +\sqrt{3 \over 3+\gamma^2}x^2,$$
300: $$p^k(\tau,\sigma)={\partial {\cal L}\over\partial \dot{x}^k}=\dot{x}^k (\tau,\sigma)
301: \qquad k=5,6,7,8,$$
302:
303: \noindent
304: and impose the equal time commutation relations
305:
306: $$\big[p^k(\tau,\sigma),x^j(\tau,\sigma')\big]=-i\delta^{jk}\delta (\sigma-\sigma').$$
307:
308: \noindent
309: The Hamiltonian is
310:
311: \begin{eqnarray}
312: H={1\over 2}\int_0^\pi d\sigma \left[\sum_{k=1}^8
313: \left(p^k p^k+{\partial x^k\over\partial\sigma}{\partial x^k\over\partial\sigma}\right)
314: +{3\over 3+\gamma^2}\sum_{k=1}^4 (x^k)^2+\sum_{k=5}^8(x^k)^2 +
315: {4\gamma^2\over 3+\gamma^2}\sum_{k=1}^2 (x^k)^2\right.\nonumber\\
316: \left. -{4\gamma\over\sqrt{3+\gamma^2}}\left(x^2{\partial x^3\over\partial\sigma}
317: -x^1{\partial x^4\over\partial\sigma}\right)+{2\sqrt{3}\over\sqrt{3+\gamma^2}}
318: \left(p^1x^3+p^2x^4-p^3x^1-p^4x^2\right)\right] .\nonumber
319: \end{eqnarray}
320:
321: \noindent
322: Notice that the modes corresponding to $x^5,x^6,x^7,x^8$ have masses that do not depend
323: on $\gamma$, i.e. they are unaffected by the deformation. This is not unexpected, since
324: these coordinates come from the AdS$_5$ part of the space which does not participate in
325: the deformation. We will, from this point on, consider only $x^1,x^2,x^3,x^4$.
326:
327: The Heisenberg equations of motion are
328:
329: $${\partial^2 x^i\over\partial t^2}-{\partial^2 x^i\over\partial \sigma^2}
330: +f^{ij}{\partial x^j\over\partial t}+h^{ij}{\partial x^j\over\partial \sigma}
331: +k_i x^i=0,$$
332:
333: \noindent
334: where
335:
336: $$f^{ij}=\left[\matrix{
337: 0 &0 &-2\sqrt{3\over 3+\gamma^2} &0\cr
338: 0 &0 &0 &-2\sqrt{3\over 3+\gamma^2}\cr
339: 2\sqrt{3\over 3+\gamma^2} &0 &0 &0\cr
340: 0 &2\sqrt{3\over 3+\gamma^2} &0 &0}\right],$$
341:
342: \noindent
343: and
344:
345: $$h^{ij}=2\left[\matrix{
346: 0 &0 &0 &{\gamma\over \sqrt{3+\gamma^2}}\cr
347: 0 &0 &-{\gamma\over \sqrt{3+\gamma^2}} &0\cr
348: 0 &{\gamma\over \sqrt{3+\gamma^2}} &0 &0\cr
349: -{\gamma\over \sqrt{3+\gamma^2}} &0 &0 &0}\right],$$
350: $$k_1=k_2={4\gamma^2\over 3+\gamma^2},\qquad k_3=k_4=0.$$
351:
352: \noindent
353: To solve these equations introduce the mode expansions
354:
355: $$ x^i(t,\sigma)=\sum_{n=-\infty}^\infty x_n^i (t)e^{2in\sigma}.$$
356:
357: \noindent
358: Reality of $x^i(t,\sigma)$ is encoded (as usual) in
359:
360: $$ x_n^i=(x_{-n}^i)^*.$$
361:
362: \noindent
363: The equations of motion now become the following equation for the modes
364:
365: $${\partial^2 x^i_n\over\partial t^2}+4n^2 x^i_n
366: +f^{ij}{\partial x^j_n\over\partial t}+2in h^{ij} x^j_n
367: +k_i x^i_n=0.$$
368:
369: \noindent
370: Following \cite{SHPP} we now make the following ansatz
371:
372: $$x_n^i(t)=a^{(n)}_j A^{(n)}_{ij}e^{i\omega^{(n)}_j t}.$$
373:
374: \noindent
375: $a^{(n)}_j$ will be a destruction/annihilation operator; $A^{(n)}_{ij}$ is a unitary transformation
376: diagonalizing the equation of motion; $\omega^{(n)}_j$ is our spectrum. Plugging this into
377: the equations of motion we find
378:
379: $$\left(-(\omega_k^{(n)})^2\delta^{ij}+4n^2\delta^{ij}+if^{ij}\omega_k^{(n)}+2inh^{ij}+k_i\delta^{ij}
380: \right)a_k^{(n)}A_{jk}^{(n)}e^{i\omega_k^{(n)}t}=0.$$
381:
382: \noindent
383: The condition for a nontrivial solution is
384:
385: $$\det \left(-(\omega_k^{(n)})^2\delta^{ij}+4n^2\delta^{ij}+if^{ij}\omega_k^{(n)}+2inh^{ij}+k_i\delta^{ij}
386: \right)=0,$$
387:
388: \noindent
389: which leads to the following quartic equation
390:
391: $$\omega^4 -(4+8n^2)\omega^2 +16n^4 =0.$$
392:
393: \noindent
394: It is solved by
395:
396: $$\omega
397: =1\pm\sqrt{1+4n^2}.$$
398:
399: \noindent
400: This is in perfect agreement with \cite{Toni}. Notice that the spectrum is independent
401: of the deformation parameter $\gamma$. The fact that the spectrum is independent
402: of $\gamma$ is unexpected. Evidently, the $\gamma$ dependence in the $B$ field exactly
403: compensates for the $\gamma$ dependence of the geometry.
404:
405: \section{Dual Field Theory Analysis}
406:
407: In this section, we will study the field theory obtained after deforming
408: the superpotential (fields with a hat $\hat\Phi$ denote superfields; fields without a
409: hat $\Phi$ denote the Higgs fields - the bosonic bottom component of $\hat\Phi$)
410:
411: $$ \Tr \left(\hat{\Phi}^1\hat{\Phi}^2\hat{\Phi}^3 -
412: \hat{\Phi}^1\hat{\Phi}^3\hat{\Phi}^2\right)\to
413: \Tr \left(e^{i\pi\gamma}\hat{\Phi}^1\hat{\Phi}^2\hat{\Phi}^3 -
414: e^{-i\pi\gamma}\hat{\Phi}^1\hat{\Phi}^3\hat{\Phi}^2\right).$$
415:
416: \noindent
417: We consider only the Higgs fields. Our goal is to construct
418: operators dual to the string modes
419: discussed in section 3.1; these will be built from the Higgs fields. The kinetic
420: terms and $D$ terms for the Higgs fields are invariant under the deformation.
421: The usual $F$ terms are however now replaced by
422:
423: $$V=\Tr( \Big|\big[\Phi^2 ,\Phi^3\big]_\gamma\Big|^2
424: + \Big|\big[ \Phi^3,\Phi^1\big]_\gamma\Big|^2
425: + \Big|\big[ \Phi^1,\Phi^2\big]_\gamma\Big|^2 ),$$
426:
427: \noindent
428: where
429:
430: $$\big[A ,B\big]_\gamma\equiv e^{i\pi\gamma}AB -e^{-i\pi\gamma}BA .$$
431:
432: \noindent
433: In the undeformed theory\cite{Motl}, when computing correlators of traces
434: in the case that each trace
435: involves only $\Phi^1$, $\Phi^2$ and $\Phi^3$ or
436: $\bar\Phi^1$, $\bar\Phi^2$ and $\bar\Phi^3$,
437: one does not need to consider $D$-term contributions, self energy corrections
438: or gluon exchange at order $g_{YM}^2$ in Yang-Mills perturbation theory. (See
439: \cite{Penati} for useful superspace techniques.) We
440: will argue that this is also true in the deformed theory, at leading order in $N$.
441: Using this insight,
442: we construct operators in the Yang-Mills theory that are dual to the vacuum of
443: the sigma model. Next we study operators dual to excited string
444: states in the undeformed ($\gamma=0$) theory. Finally, we reconsider
445: this question in the deformed theory.
446:
447: \subsection{Only $F$-terms contribute}
448:
449: In this section we will consider correlators of the form
450:
451: $$\langle {\cal O}\bar{\cal O}\rangle,$$
452:
453: \noindent
454: where ${\cal O}$ is a trace of $k$ Higgs fields
455:
456: $${\cal O}=f_{i_1 i_2\cdots i_k}\Tr \left( \Phi^{i_1}\Phi^{i_2}\cdots
457: \Phi^{i_k}\right),$$
458:
459: \noindent
460: and the indices $i_1,...,i_k\in \{1,2,3\}$. We do not assume anything about
461: the coefficient $f_{i_1 i_2\cdots i_k}$.
462: In the undeformed theory\cite{Motl}, one argues that the $D$-terms, gluon exchange
463: and self energy corrections are all flavor blind at one-loop. Consequently, if we
464: are working to one loop order, we could replace
465:
466: $${\cal O}\to\Tr \left((\Phi^1)^k\right).$$
467:
468: \noindent
469: The result of \cite{HFS} tells us that the correlator
470: $\langle\Tr \left((\Phi^1)^k\right)\Tr \left((\bar\Phi^1)^k\right)\rangle$ receives no
471: radiative corrections at $O(g_{YM}^2)$, so the result follows.
472:
473: When we deform the theory, the $D$-term contributions and gluon exchange contributions
474: are unchanged. $F$-term contributions to the self energy need to be considered
475: carefully because the $F$-terms are affected by the deformation. The $F$-terms can be
476: split into two pieces
477:
478: $$ V_F = V_{inv}+V_{def},$$
479:
480: \noindent
481: where
482:
483: $$V_{inv}= 2\Tr (\Phi^1\Phi^2\bar\Phi^2\bar\Phi^1 + \Phi^2\Phi^1\bar\Phi^1\bar\Phi^2),$$
484: $$V_{def}=-2\Tr ( e^{-2\pi i\gamma}\Phi^2\Phi^1\bar\Phi^2\bar\Phi^1 +
485: e^{2\pi i\gamma}\Phi^1\Phi^2\bar\Phi^1\bar\Phi^2).$$
486:
487: \noindent
488: The self energy contribution coming from $V_{inv}$ will be the same as in the
489: undeformed theory; self energy contribution coming from $V_{def}$ will not. The
490: Feynman diagrams corresponding to self energy contributions, coming from these
491: two vertices, are shown below. (A) shows the contribution from $V_{inv}$ and
492: (B) the contribution from $V_{def}$. Since (B) is a non-planar diagram, it can
493: be dropped at large $N$ and consequently, the only contribution to the self energy
494: coming from the $F$-terms is invariant under the deformation to leading order in $N$.
495: Thus, for the correlators that we are considering, one does not need to consider $D$-term
496: contributions,
497: self energy corrections or gluon exchange, at order $g_{YM}^2$ in Yang-Mills perturbation
498: theory and at leading order in $N$.
499:
500: \begin{center}
501: \begin{figure}[h]{\psfig{file=ppfig.ps,width=7cm,height=6cm}
502: \caption{This plot shows the Feynman diagrams corresponding to self energy
503: contributions coming from the $F$-terms. (A) shows the contribution from $V_{inv}$ and
504: (B) the contribution from $V_{def}$. (A) is $O(g_{YM}^2 N^3)$; (B) is $O(g_{YM}^2 N)$.}}
505: \end{figure}
506: \end{center}
507:
508: \subsection{Operators dual to the Vacuum}
509:
510: The operator dual to the vacuum of the string sigma model is a BPS operator.
511: Thus, we expect that the $U(1)_{\cal R}$ charge of this operator is equal to
512: its conformal dimension, and further that it is not charged under the $U(1)\times U(1)$
513: symmetry of the field theory. This follows because our pp-wave limit is taken by boosting
514: along $\psi$; there is no momentum in the $\varphi_1,\varphi_2$ directions.
515: The charges and dimension of the three Higgs fields are
516:
517: $$\matrix{ &U(1) &U(1) &U(1)_{\cal R}\equiv J &\Delta\cr
518: \Phi^1 &0 &-1 &1 &1\cr
519: \Phi^2 &1 &1 &1 &1\cr
520: \Phi^3 &-1 &0 &1 &1}$$
521:
522: \noindent
523: We will explicitly construct the operator dual to the vacuum for small values of
524: $J$. This will allow us to extract a rule that gives the correct operator for all $J$.
525:
526: For $J=3$, there are two independent loops out of which the operator dual to the
527: vacuum could be constructed
528:
529: $${\cal O}_1=\Tr (\Phi^1\Phi^2\Phi^3),\qquad {\cal O}_2 =\Tr (\Phi^1\Phi^3\Phi^2).$$
530:
531: \noindent
532: Using the two point function of the Higgs fields
533: (indices $a,b,c,d=1,...,N$ are color labels;
534: indices $j,k=1,2,3$)
535:
536: $$\langle \bar\Phi^j_{ab}(x)\Phi^k_{cd}(0)\rangle =\delta^{jk}\delta_{ad}\delta_{bc}
537: {1\over 4\pi^2 |x|^2}\equiv \delta^{jk}\delta_{ad}\delta_{bc}{a\over |x|^2},$$
538:
539: \noindent
540: we compute the planar contribution to
541:
542: \begin{equation}
543: \langle\bar{\cal O}_i(x_1) V_F(y){\cal O}_j(x_2)\rangle =
544: {\cal M}_{ij} {a^5\over |x_1-y|^4 |x_2-y|^4 |x_1-x_2|^2}N^4.
545: \label{keycor}
546: \end{equation}
547:
548: \noindent
549: We use the above correlator to define the matrix ${\cal M}$.
550: The result is
551:
552: $${\cal M}^T =
553: \left[\matrix{6 &-6e^{-2\pi i\gamma}\cr -6e^{2\pi i\gamma} &6}\right].$$
554:
555: \noindent
556: The matrix ${\cal M}$ has a single zero eigenvalue. The operator dual to
557: the vacuum is given by that linear combination which corresponds to the
558: zero eigenvalue - it is the two point function of this linear combination
559: that is not corrected, as expected for a BPS operator. There is a single zero
560: eigenvalue which implies that this state is unique. Our operator is
561:
562: $$ {\cal O}_\gamma =\Tr\left( \Phi^1\Phi^2\Phi^3\right)+e^{-2\pi i\gamma}\Tr \left(
563: \Phi^1\Phi^3\Phi^2\right).$$
564:
565: \noindent
566: It has dimension $\Delta =3$, $U(1)_{\cal R}$ charge $J=3$ and is neutral under
567: $U(1)\times U(1)$. Notice that when $\gamma=0$ our deformed operator recovers the
568: expected BPS operator of the undeformed case and further that ${\cal O}_\gamma $ respects
569: the $Z_3$ symmetry (which acts as a cyclic permutation of the three Higgs fields) of the
570: deformed theory.
571:
572: Next, consider $J=6$. In this case, our operator is a linear combination of 16 loops
573:
574: $${\cal O}_1=\Tr\left(\Phi^1 \Phi^1 \Phi^2 \Phi^2 \Phi^3 \Phi^3\right)\,\,
575: {\cal O}_2=\Tr\left(\Phi^1 \Phi^1 \Phi^2 \Phi^3 \Phi^2 \Phi^3\right)\,\,
576: {\cal O}_3=\Tr\left(\Phi^1 \Phi^1 \Phi^2 \Phi^3 \Phi^3 \Phi^2\right)$$
577: $${\cal O}_4=\Tr\left(\Phi^1 \Phi^1 \Phi^3 \Phi^2 \Phi^2 \Phi^3\right)\,\,
578: {\cal O}_5=\Tr\left(\Phi^1 \Phi^1 \Phi^3 \Phi^2 \Phi^3 \Phi^2\right)\,\,
579: {\cal O}_6=\Tr\left(\Phi^1 \Phi^1 \Phi^3 \Phi^3 \Phi^2 \Phi^2\right)$$
580: $${\cal O}_7=\Tr\left(\Phi^1 \Phi^2 \Phi^1 \Phi^2 \Phi^3 \Phi^3\right)\,\,
581: {\cal O}_8=\Tr\left(\Phi^1 \Phi^2 \Phi^1 \Phi^3 \Phi^2 \Phi^3\right)\,\,
582: {\cal O}_9=\Tr\left(\Phi^1 \Phi^2 \Phi^1 \Phi^3 \Phi^3 \Phi^2\right)$$
583: $${\cal O}_{10}=\Tr\left(\Phi^1 \Phi^2 \Phi^2 \Phi^1 \Phi^3 \Phi^3\right)\,\,
584: {\cal O}_{11}=\Tr\left(\Phi^1 \Phi^2 \Phi^2 \Phi^3 \Phi^1 \Phi^3\right)\,\,
585: {\cal O}_{12}=\Tr\left(\Phi^1 \Phi^2 \Phi^3 \Phi^1 \Phi^2 \Phi^3\right)$$
586: $${\cal O}_{13}=\Tr\left(\Phi^1 \Phi^2 \Phi^3 \Phi^1 \Phi^3 \Phi^2\right)\,\,
587: {\cal O}_{14}=\Tr\left(\Phi^1 \Phi^2 \Phi^3 \Phi^2 \Phi^1 \Phi^3\right)\,\,
588: {\cal O}_{15}=\Tr\left(\Phi^1 \Phi^3 \Phi^1 \Phi^3 \Phi^2 \Phi^2\right)$$
589: $${\cal O}_{16}=\Tr\left(\Phi^1 \Phi^3 \Phi^2 \Phi^1 \Phi^3 \Phi^2\right).$$
590:
591: \noindent
592: These operators were selected by requiring that they have $\Delta =J=6$,
593: and zero $U(1)\times U(1)$ charge. We again want to identify the linear combination
594: of these operators that is BPS. As for the case $J=3$, we do this by looking for the
595: linear combination whose two point function does not receive corrections
596: at $O(g_{YM}^2)$ and at leading order in $N$. By studying the correlator (\ref{keycor})
597: we can read off ${\cal M}$; null vectors of ${\cal M}$ are then natural candidate
598: BPS operators. In this case, again at leading order in $N$ we find
599:
600: $${\cal M}^T=2\left[
601: \matrix{
602: 3 &b &0 &0 &0 &0 &b &0 &0 &0 &b &0 &0 &0 &0 &0\cr
603: b^* &5 &b &b &0 &0 &0 &b &0 &0 &0 &0 &0 &b &0 &0\cr
604: 0 &b^* &4 &0 &b &0 &b^* &0 &b &0 &0 &0 &0 &0 &0 &0\cr
605: 0 &b^* &0 &4 &b &0 &0 &0 &0 &0 &b^* &0 &0 &0 &b &0\cr
606: 0 &0 &b^* &b^* &5 &b &0 &b^* &0 &0 &0 &0 &0 &b^* &0 &0\cr
607: 0 &0 &0 &0 &b^* &3 &0 &0 &b^* &0 &0 &0 &0 &0 &b^* &0\cr
608: b^* &0 &b &0 &0 &0 &5 &b &0 &b &0 &0 &b &0 &0 &0\cr
609: 0 &b^* &0 &0 &b &0 &b^* &6 &b &0 &0 &b^* &0 &0 &0 &b\cr
610: 0 &0 &b^* &0 &0 &b &0 &b^* &5 &b^* &0 &0 &b^* &0 &0 &0\cr
611: 0 &0 &0 &0 &0 &0 &b^* &0 &b &4 &b^* &0 &0 &0 &b &0\cr
612: b^* &0 &0 &b &0 &0 &0 &0 &0 &b &5 &0 &b &b &0 &0\cr
613: 0 &0 &0 &0 &0 &0 &0 &2b &0 &0 &0 &6 &2b &2b &0 &0\cr
614: 0 &0 &0 &0 &0 &0 &b^* &0 &b &0 &b^* &b^* &6 &0 &b &b\cr
615: 0 &b^* &0 &0 &b &0 &0 &0 &0 &0 &b^* &b^* &0 &6 &b &b\cr
616: 0 &0 &0 &b^* &0 &b &0 &0 &0 &b^* &0 &0 &b^* &b^* &5 &0\cr
617: 0 &0 &0 &0 &0 &0 &0 &2b^* &0 &0 &0 &0 &2b^* &2b^* &0 &6}
618: \right]$$
619:
620: \noindent
621: where $b=-e^{-2\pi i\gamma}$ and $b^*=-e^{2\pi i\gamma}$.
622: Again, ${\cal M}$ has a single zero eigenvalue, so that there is an
623: operator whose two point function does not get corrected at $O(g_{YM}^2)$
624: and it is again unique. For $J=9$ there are 188 basis loops. In this case
625: ${\cal M}$ again has a single null vector, so that we again have a unique candidate
626: BPS operator.
627:
628: By studying the candidate BPS operators for $J=3,6,9$ we have been able to
629: identify a rule which allows us to write down a candidate BPS operator for
630: {\it any} $J$. To write down our rule, we call the following exchanges
631:
632: $$\Phi^1\Phi^2\to\Phi^2\Phi^1,\quad {\rm or}\quad
633: \Phi^2\Phi^3\to\Phi^3\Phi^2,\quad {\rm or}\quad
634: \Phi^3\Phi^1\to\Phi^1\Phi^3 ,$$
635:
636: \noindent
637: {\it even exchanges} and the exchanges
638:
639: $$\Phi^2\Phi^1\to\Phi^1\Phi^2,\quad {\rm or}\quad
640: \Phi^3\Phi^2\to\Phi^2\Phi^3,\quad {\rm or}\quad
641: \Phi^1\Phi^3\to\Phi^3\Phi^1 ,$$
642:
643: \noindent
644: {\it odd exchanges}. Consider for illustration the case with $J=6$.
645: To construct the vacuum state, we start from the
646: loop $\Tr\left(\Phi^1 \Phi^1 \Phi^2 \Phi^2 \Phi^3 \Phi^3\right)$ and perform a sequence
647: of even and odd exchanges until we generate the full 16 operators generated above.
648: For each odd exchange we append the factor $\alpha^*=e^{2\pi\gamma i}$, and for each even
649: exchange we append the factor $\alpha =e^{-2\pi\gamma i}$. Thus, for example, since
650:
651: $$\Tr\left(\Phi^1 \Phi^1 \Phi^2 \Phi^3 \Phi^3\Phi^2\right)$$
652:
653: \noindent
654: is obtained from
655:
656: $$\Tr\left(\Phi^1 \Phi^1 \Phi^2 \Phi^2 \Phi^3 \Phi^3\right)$$
657:
658: \noindent
659: by performing two even exchanges
660:
661: $$\Tr\left(\Phi^1 \Phi^1 \Phi^2 \Phi^2 \Phi^3 \Phi^3\right)\to
662: \Tr\left(\Phi^1 \Phi^1 \Phi^2 \Phi^3 \Phi^2\Phi^3\right)\to
663: \Tr\left(\Phi^1 \Phi^1 \Phi^2\Phi^3 \Phi^3\Phi^2\right),$$
664:
665: \noindent
666: we know that it will have a phase of $\alpha^2 =(e^{-2\pi i\gamma})^2$.
667: Following this rule, we find the following operator
668:
669: \begin{eqnarray}
670: {\cal O}_{def}&=&
671: \Tr\left(\Phi^1 \Phi^1 \Phi^2 \Phi^2 \Phi^3 \Phi^3\right)+
672: \alpha\Tr\left(\Phi^1 \Phi^1 \Phi^2 \Phi^3 \Phi^2 \Phi^3\right)
673: \nonumber\\
674: &+&\alpha^2 \Tr\left(\Phi^1 \Phi^1 \Phi^2 \Phi^3 \Phi^3 \Phi^2\right)+
675: \alpha^2 \Tr\left(\Phi^1 \Phi^1 \Phi^3 \Phi^2 \Phi^2 \Phi^3\right)
676: \nonumber\\
677: &+&\alpha^3 \Tr\left(\Phi^1 \Phi^1 \Phi^3 \Phi^2 \Phi^3 \Phi^2\right)+
678: \alpha^4 \Tr\left(\Phi^1 \Phi^1 \Phi^3 \Phi^3 \Phi^2 \Phi^2\right)
679: \nonumber\\
680: &+&\alpha\Tr\left(\Phi^1 \Phi^2 \Phi^1 \Phi^2 \Phi^3 \Phi^3\right)+
681: \alpha^2 \Tr\left(\Phi^1 \Phi^2 \Phi^1 \Phi^3 \Phi^2 \Phi^3\right)
682: \nonumber\\
683: &+&\alpha^3\Tr\left(\Phi^1 \Phi^2 \Phi^1 \Phi^3 \Phi^3 \Phi^2\right)+
684: \alpha^2 \Tr\left(\Phi^1 \Phi^2 \Phi^2 \Phi^1 \Phi^3 \Phi^3\right)
685: \nonumber\\
686: &+&\alpha\Tr\left(\Phi^1 \Phi^2 \Phi^2 \Phi^3 \Phi^1 \Phi^3\right)+
687: \alpha\Tr\left(\Phi^1 \Phi^2 \Phi^3 \Phi^1 \Phi^2 \Phi^3\right)
688: \nonumber\\
689: &+&\alpha^2 \Tr\left(\Phi^1 \Phi^2 \Phi^3 \Phi^1 \Phi^3 \Phi^2\right)+
690: \alpha^2 \Tr\left(\Phi^1 \Phi^2 \Phi^3 \Phi^2 \Phi^1 \Phi^3\right)
691: \nonumber\\
692: &+&\alpha^3 \Tr\left(\Phi^1 \Phi^3 \Phi^1 \Phi^3 \Phi^2 \Phi^2\right)+
693: \alpha^3 \Tr\left(\Phi^1 \Phi^3 \Phi^2 \Phi^1 \Phi^3 \Phi^2\right).
694: \nonumber
695: \end{eqnarray}
696:
697: \noindent
698: for $J=6$. Notice that when $\gamma=0$, this again reduces to a BPS operator
699: of the undeformed theory.
700: As a second example, here are the first few terms for the $J=9$
701: operator dual to the vacuum
702:
703: \begin{eqnarray}
704: {\cal O}&=& \Tr \left( (\Phi^1)^3(\Phi^2)^3(\Phi^3)^3\right)+
705: \alpha\Tr \left((\Phi^1)^3(\Phi^2)^2\Phi^3\Phi^2(\Phi^3)^2\right)+
706: \alpha^2\Tr \left((\Phi^1)^3(\Phi^2)^2(\Phi^3)^2\Phi^2\Phi^3\right)
707: \nonumber\\
708: &+&\alpha^3\Tr \left((\Phi^1)^3(\Phi^2)^2(\Phi^3)^3\Phi^2\right)+...
709: \nonumber
710: \end{eqnarray}
711:
712: \noindent
713: There are a total of 188 terms in the above sum.
714:
715: One may worry that our prescription to obtain the operator dual to the vacuum is not
716: well defined. What is at stake here, is the fact that this
717: prescription might be ambiguous. If there is more than one sequence of even and odd
718: exchanges that will produce a particular word, we must check that each distinct
719: sequence of exchanges
720: assigns the same phase. In specific examples, we have checked that this is indeed
721: the case.
722:
723: \subsection{Operators Dual to Excited String Modes in the Undeformed Theory}
724:
725: In this subsection, we set $\gamma=0$. Lets us consider the original pp-wave
726: limit of \cite{BMN}. Towards this end, imagine taking the pp-wave limit by boosting
727: along the $\Phi^2$ direction (instead of along $\psi$). Define
728:
729: \begin{equation}
730: \tilde {\cal O}_{(n)}=\Tr (\Phi^1(\Phi^2)^n\Phi^3 (\Phi^2)^{J-n}).
731: \label{loops}
732: \end{equation}
733:
734: \noindent
735: which have two point functions
736:
737: $$\langle \tilde{\cal O}_{(n)}(x_1)\bar{\tilde{\cal O}}_{(m)}(x_2)\rangle
738: =\delta_{mn}{N^{J+2} a^{J+2}\over |x_1-x_2|^{2J+4}}.$$
739:
740: \noindent
741: To obtain the anomalous dimensions of the
742: $\tilde {\cal O}_{(n)}$ we need to diagonalize $Q$ where
743:
744: $$\langle\tilde {\cal O}_{(i)}(x_1)V_F(y)\bar{\tilde {\cal O}}_{(j)}(x_2)\rangle
745: =Q_{ij}{N^{J+3}a^{J+4}\over |x_1-x_2|^{2J}|x_1-y|^4 |x_2-y|^4}.$$
746:
747: \noindent
748: At leading order in $N$, we find
749:
750: $$ Q=2\left[
751: \matrix{
752: 3 &-2 &0 &0 &0 &\cdots &0 &0 &-1 \cr
753: -2 &4 &-2 &0 &0 &\cdots &0 &0 &0 \cr
754: 0 &-2 &4 &-2 &0 &\cdots &0 &0 &0 \cr
755: 0 &0 &-2 &4 &-2 &\cdots &0 &0 &0 \cr
756: : &: &: &: &: &\cdots &: &: &: \cr
757: 0 &0 &0 &0 &0 &\cdots &-2 &4 &-2 \cr
758: -1 &0 &0 &0 &0 &\cdots &0 &-2 &3}\right].$$
759:
760: \noindent
761: The eigenvalues of $Q$ determine the anomalous dimensions of
762: $\tilde {\cal O}_{(n)}$. The eigenvectors of $Q$ determine the operators that
763: are dual to excited string modes.
764:
765: Since the undeformed theory has an $SO(6)$ rotational invariance, the anomalous dimensions
766: of the loops discussed above should agree with the anomalous dimensions of the loops obtained
767: in the pp-wave limit we are interested in. We will show that this is indeed the case.
768:
769: If we look at the equation (\ref{loops}) one can think that the fields $\Phi^1$ define
770: a lattice, and that Yang-Mills interaction can be described in terms of the fields
771: $\Phi^2$ and $\Phi^3$ ``hopping on this lattice". In our pp-wave limit, there isn't a
772: field which is singled out to play the r\^ole of a lattice. However, in analogy to
773: (\ref{loops}), define
774:
775: $${\cal O}_{(n)}\equiv C_{i_1 i_2 i_3 \cdots i_J}\Tr (
776: \Phi^1\Phi^{i_1}\cdots \Phi^{i_n}\Phi^3 \Phi^{i_{n+1}}\cdots \Phi^{i_J}).$$
777:
778: The first thing we need to compute is the overlap
779: $\langle{\cal O}_{(l)}\bar{\cal O}_{(k)}\rangle .$
780: Introduce the notation (repeated indices summed as usual)
781:
782: $$ C^2\equiv C_{i_1 i_2 \cdots i_J}C_{i_1 i_2 \cdots i_J}.$$
783:
784: \noindent
785: The operator dual to the sigma model vacuum relevant for our pp-wave limit has an equal number
786: of $\Phi^1$s, $\Phi^2$s and $\Phi^3$s. This implies that (the exact location of the
787: indices that are 1s or 2s or 3s is unimportant because $C$ is a symmetric tensor)
788:
789: $$C_{1 i_2 \cdots i_J}C_{1 i_2 \cdots i_J}=
790: C_{2 i_2 \cdots i_J}C_{2 i_2 \cdots i_J}=
791: C_{3 i_2 \cdots i_J}C_{3 i_2 \cdots i_J}={1\over 3}C^2,$$
792: $$C_{1 2 \cdots i_J}C_{1 2 \cdots i_J}=
793: C_{2 3 \cdots i_J}C_{2 3 \cdots i_J}=
794: C_{3 1 \cdots i_J}C_{3 1 \cdots i_J}=AC^2,$$
795: $$C_{1 1 i_3 \cdots i_J}C_{1 1 i_3 \cdots i_J}=
796: C_{2 2 i_3 \cdots i_J}C_{2 2 i_3 \cdots i_J}=
797: C_{3 3 i_3 \cdots i_J}C_{3 3 i_3 \cdots i_J}=BC^2,$$
798:
799: \noindent
800: where
801:
802: $$ 6A+3B=1.$$
803:
804: \noindent
805: It is now simple to argue that ($k\ne l$; in the second equation below,
806: $l$ is {\it not} summed)
807:
808: $$\langle{\cal O}_{(l)}(x_1)\bar{\cal O}_{(k)}(x_2)\rangle =
809: \left[(J-2)AC^2 + {2\over 3}C^2+\delta_{l,J-k}AC^2\right] N^{J+2}
810: {a^{J+2}\over |x_1-x_2|^{2J+4}} ,$$
811: $$\langle{\cal O}_{(l)}(x_1)\bar{\cal O}_{(l)}(x_2)\rangle =
812: \left[ (J-1)AC^2 + C^2\right]N^{J+2}
813: {a^{J+2}\over |x_1-x_2|^{2J+4}}.$$
814:
815: \noindent
816: At large $J$ we can write
817:
818: $$\langle{\cal O}_{(l)}(x_1)\bar{\cal O}_{(k)}(x_2)\rangle =
819: \left[(J-2)AC^2 + {2\over 3}C^2\right] N^{J+2}
820: {a^{J+2}\over |x_1-x_2|^{2J+4}} .$$
821:
822: \noindent
823: Thus, ($k$ and $l$ unrestricted)
824:
825: \begin{eqnarray}
826: \langle{\cal O}_{(l)}(x_1)\bar{\cal O}_{(k)}(x_2)\rangle
827: &\equiv& M_{lk}{N^{J+2} a^{J+2}\over |x_1-x_2|^{2J+4}}
828: \nonumber\\
829: &=&\left[
830: \left((J-2)AC^2 +{2C^2\over 3}\right)L +(A+{1\over 3})C^2 I\right]_{lk}
831: {N^{J+2} a^{J+2}\over |x_1-x_2|^{2J+4}},
832: \nonumber
833: \end{eqnarray}
834:
835: \noindent
836: where $L$ is a matrix with a 1 in every single entry and $I$ is the identity matrix.
837: In what follows we will need the eigenvectors and eigenvalues of $M_{kl}$. There are $J$
838: eigenvectors that have the form (the first $n$ entries are 1s; $n=1,2,...,J$)
839:
840: $$|n\rangle ={1\over \sqrt{n^2+n}}
841: \left[\matrix{1\cr 1\cr :\cr 1\cr -n\cr 0\cr :\cr 0}\right].$$
842:
843: \noindent
844: These have eigenvalue $(A+{1\over 3})C^2$. There is a single eigenvector of the form
845:
846: $$ |J+1\rangle ={1\over\sqrt{J+1}}\left[\matrix{1\cr 1\cr :\cr :\cr 1}\right].$$
847:
848: \noindent
849: This eigenvector has eigenvalue
850: $(J+1)\left((J-2)AC^2 +{2C^2\over 3}\right) +(A+{1\over 3})C^2.$
851: These eigenvalues and eigenvectors can be used to define the new operators
852: ${\cal K}_{(n)}$ that have a diagonal two point function. Explicitly we have
853:
854: $${\cal K}_{(n)}={\langle n|_l {\cal O}_l\over\sqrt{\lambda_n}},$$
855:
856: \noindent
857: and
858:
859: $$\bar {\cal K}_{(n)}={\bar{\cal O}_l|n\rangle_l\over\sqrt{\lambda_n}}.$$
860:
861: \noindent
862: These operators have two point function
863:
864: $$\langle\bar {\cal K}_{(n)}(x_1){\cal K}_{(m)} (x_2)\rangle =
865: \delta_{mn}{N^{J+2} a^{J+2}\over |x_1-x_2|^{2J+4}}.$$
866:
867: \noindent
868: To determine the anomalous dimensions for this set of operators at $O(g_{YM}^2)$ we compute
869:
870: $$\langle{\cal O}_{(i)}(x_1)V_F(y)\bar{\cal O}_{(j)}(x_2)\rangle
871: =H_{ik}M_{kj}{N^{J+3}a^{J+4}\over |x_1-x_2|^{2J}|x_1-y|^4 |x_2-y|^4}.$$
872:
873: \noindent
874: At large $N$ we obtain
875:
876: $$ H=2\left[\matrix{
877: 3 &-2 &0 &0 &0 &\cdots &0 &0 &-1 \cr
878: -2 &4 &-2 &0 &0 &\cdots &0 &0 &0 \cr
879: 0 &-2 &4 &-2 &0 &\cdots &0 &0 &0 \cr
880: 0 &0 &-2 &4 &-2 &\cdots &0 &0 &0 \cr
881: : &: &: &: &: &\cdots &: &: &: \cr
882: 0 &0 &0 &0 &0 &\cdots &-2 &4 &-2 \cr
883: -1 &0 &0 &0 &0 &\cdots &0 &-2 &3}\right].$$
884:
885: \noindent
886: Using this result, it is a simple matter to demonstrate
887:
888: $$\langle{\cal K}_{(i)}(x_1)V_F(y)\bar{\cal K}_{(j)}(x_2)\rangle
889: =K_{ij}{N^{J+3}a^{J+4}\over |x_1-x_2|^{2J}|x_1-y|^4 |x_2-y|^4},$$
890:
891: \noindent
892: where
893:
894: $$ K_{nm}=\sqrt{\lambda_m\over\lambda_n}\langle n|H|m\rangle .$$
895:
896: \noindent
897: It is the eigenvalues of $K$ that determine the anomalous dimensions. The eigenvectors
898: of $K$ give the corresponding operators dual to excited string modes.
899: The prefactor $\sqrt{\lambda_m\over\lambda_n}$ differs from 1 only if $m=J+1$
900: or if $n=J+1$. As a consequence, noting that
901:
902: $$\langle J+1|H=0=H|J+1\rangle,$$
903:
904: \noindent
905: we see that we can write
906:
907: $$ K_{nm}=\langle n|H|m\rangle .$$
908:
909: \noindent
910: This implies that $K$ and $H$ are related by a unitary
911: transformation and hence we may as well solve the
912: eigenvalue problem for $H$. Since $Q$ and $H$ are identical matrices, this
913: demonstrates that the spectrum of our pp-wave limit agrees with the spectrum of the pp-wave
914: limit taken in \cite{BMN}, as expected from the rotational invariance of the background.
915: This agreement between the two computations gives us confidence
916: that we have indeed identified the operators dual to excited string states.
917:
918: A few comments are in order. In our analysis, we have focused on $J+1$ operators. If
919: we write down the full set of operators with specific $U(1)_{\cal R}$
920: charge $J+2$ and $U(1)\times U(1)$ charge
921: equal to $(1,1)$ we find many more than just $J+1$ operators. Indeed, for
922: $J=6$ ($J=9$) we have kept only 7 (10) operators out of a possible 70 (1050) operators
923: with the correct quantum numbers. For $J=3$ and $J=6$ we have checked explicitly,
924: using the full set of loops,
925: that the $J+1$ BMN operators we have obtained by keeping only this subset of $J+1$
926: operators do indeed provide operators with a definite
927: anomalous dimension at $O(g_{YM}^2)$. Further, we checked
928: that the anomalous dimension we obtained
929: agrees with the anomalous dimension obtained when the full class of operators is
930: considered.
931:
932: This decoupling of a small set of nearly protected states has been used in
933: both \cite{BMN} and \cite{Berenstein}. Understanding this decoupling directly
934: in the relevant quantum field theory is an important problem. The analysis of this
935: section provides some insight into this decoupling in the field theory. The usual
936: argument\cite{BMN},\cite{Berenstein} involves taking a limit in which all states
937: that are not nearly protected have a very large energy and hence decouple. In this
938: subsection we have seen that, at this order in perturbation theory, the potential
939: coupling between the nearly protected states and other states vanishes.
940:
941: \subsection{Operators Dual to Excited String Modes in the Deformed Theory}
942:
943: In this section we will study operators dual to excited string modes for both
944: large $J$ and small $J$. This allows us to verify the $\gamma$ independence of
945: the large $J$ spectrum and further, that this is no longer the case at finite $J$.
946:
947: Consider the large $J$ limit. First, we build the ``background" on which the
948: impurities move. The background is built from an even number of $\Phi^1$, $\Phi^2$
949: and $\Phi^3$ fields. Start by selecting one of the Higgs fields from which the background
950: is to be composed. Place a second background Higgs field to the left of this first one
951: and let it hop over the first, assigning phases for even and odd exchanges as in section
952: 4.2. Place a third background Higgs field to the left of the two terms generated,
953: and let it hop all the way to the right, generating a total of 6 terms. Continue until
954: all background Higgs fields have been selected. As an example, if we wanted to build
955: the background out of one $\Phi^1$, one $\Phi^2$ and one $\Phi^3$, we would find go
956: through the following steps
957:
958: \begin{eqnarray}
959: \nonumber
960: \Phi^1\to \Phi^2 \Phi^1 +e^{2\pi i\gamma}\Phi^1\Phi^2\to
961: \Phi^3\Phi^2 \Phi^1 &+&e^{2\pi i\gamma}\Phi^2\Phi^3 \Phi^1 +
962: \Phi^2\Phi^1 \Phi^3
963: +e^{2\pi i\gamma}\Phi^3\Phi^1\Phi^2\\
964: \nonumber
965: &+&\Phi^1\Phi^3\Phi^2
966: +e^{2\pi i\gamma}\Phi^1\Phi^2\Phi^3 .
967: \end{eqnarray}
968:
969: \noindent
970: Selecting the background fields in a different order may change the
971: overall (and hence arbitrary) phase of the above operator. By building the operator
972: in this way, each exchange term we add by hand will be matched by an exchange
973: performed by the potential, with an {\it opposite sign} so that this indeed
974: builds a BPS state. This is not quite exact, because we did not consider the exchange
975: that will swap the last and first Higgs field. However, we expect that neglecting this
976: exchange is justified in the leading order of a large $J$ expansion. Notice that if
977: the above operator is now traced, it will not in general reduce to the BPS
978: state we identified in section 4.2. This can be traced back to our neglect of the
979: exchange of the first and last Higgs fields.
980:
981: We will now describe how to build excited string states with two impurities. For the
982: impurities take $\Phi^1$ and $\Phi^3$. Let $\Phi^3$ hop
983: into the $n$th position using the same rules for hopping as above. The operator
984: obtained in this way is ${\cal O}^\gamma_{n}$. For each ${\cal O}_n^\gamma$
985: let $\Phi^1$ hop into the $m$th position. Call the resulting operator
986: ${\cal O}_{n,m}^\gamma$. Now define ($p=0,1,...,J$)
987:
988: $${\cal O}^\gamma_{(p)}=\sum_{n,m}\Tr\Big({\cal O}_{n,m}^\gamma\Big)\delta_{m-n,p},$$
989:
990: \noindent
991: where the delta function sets $m-n=p\, {\rm mod}\, J$.
992: It is now a simple task to show that
993:
994: $$ \langle {\cal O}^\gamma_{(i)}(x_1)V_F(y)\bar{\cal O}^\gamma_{(j)}(x_2)\rangle
995: =H^\gamma_{ik}M^\gamma_{kj}{N^{J+3}a^{J+4}\over |x_1-x_2|^{2J}|x_1-y|^4|x_2-y|^4},$$
996:
997: \noindent
998: where
999:
1000: $$ \langle {\cal O}^\gamma_{(i)}(x_1)\bar{\cal O}^\gamma_{(j)}(x_2)\rangle
1001: =M^\gamma_{ij}{N^{J+2}a^{J+2}\over |x_1-x_2|^{2J+4}}.$$
1002:
1003: \noindent
1004: When computing these correlators, we sum over all contractions except the contractions
1005: involving the fields that were at the endpoints of ${\cal O}_{n,m}^\gamma$; this
1006: should give the correct answer in the large $J$ limit.
1007: In the above, we have
1008:
1009: $$ H^\gamma_{ik}=8\delta_{ik}-4\delta_{i+1 k}-4\delta_{i k+1},$$
1010:
1011: \noindent
1012: This looks the same as $H$ of section 4.2 except that we don't have the
1013: -1 elements in $H_{0,J}$ and $H_{J,0}$. In the large $J$ limit, we expect (and have verified
1014: numerically) that the precise
1015: details of these terms are unimportant, so that $H^\gamma$
1016: has the same spectrum as $H$ in section 4.3. Our proposal for the BMN operators is then
1017: to build them using the eigenvectors of $H^\gamma$. Thus, we see that the spectrum
1018: of anomalous dimensions coincides with the spectrum of anomalous dimensions of
1019: the undeformed ($\gamma=0$) theory, in perfect agreement with the string theory
1020: prediction.
1021:
1022: This conclusion assumes that the eigenvalues of $H^\gamma$ determine the anomalous
1023: dimensions of the operators we consider. In the undeformed case we were able to argue
1024: that this is indeed the case by studying the eigenvalues and eigenvectors of $M_{ij}$.
1025: To prove this assumption in the deformed case, we would need to provide the
1026: corresponding study for $M_{ij}^\gamma$. Although our assumption seems reasonable, we
1027: have not proved that it is indeed correct.
1028:
1029: We now consider the small $J$ limit. For small values of $J$,
1030: we can work with the full set of loops
1031: ${\cal O}_i$ that have $U(1)\times U(1)$ charge (1,1) and $U(1)_{\cal R}$ charge
1032: $J+2$. The ${\cal O}_i$ are chosen to have two point function
1033:
1034: $$ \langle {\cal O}_i\bar{\cal O}_j\rangle\propto\delta_{ij}$$
1035:
1036: \noindent
1037: at large $N$ and $O(g_{YM}^0)$. Computing correlators of the form
1038:
1039: $$\langle{\cal O}_{i}(x_1)V_F(y)\bar{\cal O}_{j}(x_2)\rangle
1040: =T_{ij}{N^{J+3}a^{J+4}\over |x_1-x_2|^{2J}|x_1-y|^4 |x_2-y|^4},$$
1041:
1042: \noindent
1043: the matrix $T$ determines the operators with a definite anomalous dimension and the
1044: anomalous dimension itself, to $O(g_{YM}^2)$. We find that for $\gamma=0.1$ and $J=3$
1045: the smallest eigenvalue of $T$ is 0.07843... and for $J=6$, the smallest eigenvalue is
1046: 0.04124... The string theory prediction of section 3, which corresponds to infinite $J$,
1047: is that this smallest eigenvalue should be zero. The fact that the smallest eigenvalue
1048: is non-zero is a clear indication that we can't compare our finite $J$ field theory
1049: results with the string theory results. We have also developed an expansion for $T$
1050: in terms of $\gamma$. It is then possible (using the results of the appendix) to
1051: develop a perturbative expansion (treating $\gamma$ as a small number) for the anomalous
1052: dimension. We find that the $O(\gamma )$ term vanishes.
1053: One could in principle develop this perturbation to even
1054: higher orders. Reproducing this perturbation series directly in the string theory would
1055: be an interesting exercise.
1056:
1057: \section{Summary}
1058: We have taken a pp-wave limit of the Lunin-Maldacena background. The resulting
1059: geometry is that of a homogeneous plane wave. The spectrum of the string
1060: is independent of the deformation parameter $\gamma$. In the
1061: dual gauge theory, we have argued that for the class of operators we consider, at
1062: $O(g_{YM}^2)$ and at leading order in $N$, all contributions to the anomalous dimension
1063: come from $F$-terms. We have identified the operator in the deformed super Yang-Mills
1064: which is dual to the sigma model vacuum state. For the undeformed theory, we have
1065: been able to identify a set of operators dual to excited string modes. Further, these
1066: operators are a small fraction of the total number of operators with the correct
1067: quantum numbers to participate. This sheds some light on the important issue of
1068: decoupling a small set of nearly protected states\cite{BMN},\cite{Berenstein}.
1069: For the deformed theory, we proposed a set of operators dual to excited string modes,
1070: for large $J$. The anomalous dimensions of these operators are independent of $\gamma$
1071: in perfect agreement with the string theory spectrum.
1072: For finite $J$, at order $g_{YM}^2$, the anomalous dimensions we have computed do
1073: depend on $\gamma$. It would be interesting to reproduce this dependence in the
1074: string theory, presumably by adding ${1\over J}$ corrections.
1075:
1076: $$ $$
1077:
1078: \noindent
1079: {\it Acknowledgements:}
1080: We would like to thank T. Mateos for pointing out an error in our string
1081: spectrum which appeared in a previous draft.
1082: The work of RdMK, JS and MS is supported by NRF grant number Gun 2047219.
1083: JM is supported by an overseas postdoctoral fellowship of the NRF (South Africa).
1084:
1085: $$ $$
1086:
1087: \appendix{\bf Appendix: Eigenvalue Problem}
1088:
1089: In this appendix we solve the eigenvalue problem of the operator $H$ introduced in
1090: section 4.3. Denoting the components of the eigenvectors
1091:
1092: $$ H|i\rangle=\lambda_i|i\rangle ,$$
1093:
1094: \noindent
1095: by
1096:
1097: $$ |i\rangle =\left[\matrix{v_0\cr v_1\cr :\cr v_{J-1}\cr v_J}\right],$$
1098:
1099: \noindent
1100: we have
1101:
1102: \begin{equation}
1103: -4v_{n-1}+8v_n -4v_{n+1}=\lambda v_n
1104: \label{evone}
1105: \end{equation}
1106:
1107: \noindent
1108: for $1\le n\le J-1$ and
1109:
1110: \begin{equation}
1111: 3v_0-2v_1 -v_J=\lambda v_0\qquad 3v_J-2v_{J-1}-v_0=\lambda v_J.
1112: \label{evtwo}
1113: \end{equation}
1114:
1115: \noindent
1116: Make the ansatz
1117:
1118: $$ v_n=Ae^{ikn}+Be^{-ikn}.$$
1119:
1120: \noindent
1121: Then, (\ref{evone}) implies $\lambda =8-8\cos (k)$, and (\ref{evtwo}) implies that the
1122: allowed values of $k$ solve
1123:
1124: $$ {\rm Imag}\left[(\lambda-3+2e^{-ik}+e^{-iJk})
1125: (3e^{ikJ}-2e^{i(J-1)k}-1-\lambda e^{iJk})\right]=0,$$
1126:
1127: \noindent
1128: where ${\rm Imag}$ stands for the imaginary part. It is now a simple exercise to determine
1129: $A$ in terms of $B$ using (\ref{evtwo}). $B$ is determined by
1130: the normalization of the eigenvector.
1131:
1132: \begin{thebibliography}{30}
1133: \parskip-2pt
1134:
1135: \bibitem{AdSCFT}
1136: J. Maldacena, ``The large N limit of superconformal field theories and supergravity,"
1137: Adv. Theor. Math. Phys. {\bf 2} 231 (1998), {\tt hep-th/9711200};\\
1138: %%CITATION = HEP-TH 9711200;%%%
1139: S. Gubser, I.R. Klebanov and A.M. Polyakov, ``Gauge Theory Correlators from Non-critical
1140: String Theory," Phys. Lett.{\bf B428} (1998) 105, {\tt hep-th/9802109};\\
1141: %%CITATION = HEP-TH 9802109;%%%
1142: E. Witten, ``Anti-de Sitter Space and Holography,"
1143: Adv. Theor. Math. Phys. {\bf 2} (1998) 253, {\tt hep-th/9802150};\\
1144: %%CITATION = HEP-TH 9802150;%%%
1145: Ofer Aharony, Steven S. Gubser, Juan M. Maldacena, Hirosi Ooguri and Yaron Oz,
1146: ``Large N Field Theories, String Theory and Gravity,"
1147: Phys. Rept. {\bf 323} (2000) 183, {\tt hep-th/9905111}.
1148: %%CITATION = HEP-TH 9905111;%%%
1149:
1150: \bibitem{BMN}
1151: D. Berenstein, J.M. Maldacena and H. Nastase,
1152: ``Strings in Flat Space and pp-waves from N=4 super Yang-Mills,"
1153: JHEP {\bf 0204}, 013 (2002), {\tt hep-th/0202021}.
1154: %%CITATION = HEP-TH 0202021;%%%
1155:
1156: \bibitem{LM}
1157: O. Lunin and J.M. Maldacena, ``Deforming Field Theories with $U(1)\times U(1)$ global
1158: symmetry and their gravity duals," {\tt hep-th/0502086}.
1159: %%CITATION = HEP-TH 0502086;%%%
1160:
1161: \bibitem{GKP}
1162: S.S. Gubser, I.R. Klebanov and A.M. Polyakov, ``A Semi-classical Limit of the Gauge/String
1163: Correspondence," Nucl. Phys. {\bf B636}, 99 (2002), {\tt hep-th/0204051};\\
1164: %%CITATION = HEP-TH 0204051;%%%
1165: S. Frolov and A.A. Tseytlin, ``Multi-spin string solutions in AdS$_5\times$S$^5$ and beyond,"
1166: Nucl. Phys. {\bf B668}, 77 (2003), {\tt hep-th/0304255};\\
1167: %%CITATION = HEP-TH 0304255;%%%
1168: A.A. Tseytlin, ``Semiclassical Quantization of Superstrings: AdS$_5\times$S$^5$ and beyond,"
1169: Int. J. Mod. Phys. {\bf A18}, 981 (2003), {\tt hep-th/0209116};\\
1170: %%CITATION = HEP-TH 0209116;%%%
1171: A.A. Tseytlin, ``Spinning Strings and AdS/CFT Duality," {\tt hep-th/0311139}.
1172: %%CITATION = HEP-TH 0311139;%%%
1173:
1174: \bibitem{FRT}
1175: S.A. Frolov, R. Roiban and A.A. Tseytlin, ``Gauge-string duality for superconformal
1176: deformations of $N=4$ Super Yang-Mills Theory," {\tt hep-th/0503192}.
1177: %%CITATION = HEP-TH 0503192;%%%
1178:
1179: \bibitem{F}
1180: S.A. Frolov, ``Lax Pair for Strings in Lunin-Maldacena Background," {\tt hep-th/0503201}.
1181: %%CITATION = HEP-TH 0503201;%%%
1182:
1183: \bibitem{Beisert}
1184: N. Beisert and R. Roiban, ``Beauty and the Twist: The Bethe Ansatz for Twisted
1185: ${\cal N}=4$ SYM," {\tt hep-th/0505187}.
1186: %%CITATION = HEP-TH 0505187;%%%
1187:
1188: \bibitem{BK}
1189: J. Gomis and H. Ooguri, ``Penrose Limits of N=1 Gauge Theories," Nucl. Phys.
1190: {\bf B635} (2002) 106, {\tt hep-th/0202157};\\
1191: %%CITATION = HEP-TH 0202157;%%%
1192: Dominic Brecher, Clifford V. Johnson, Kenneth J. Lovis and Robert C. Myers,
1193: ``Penrose Limits, Deformed PP-waves and the String Duals of N=1 Large N Gauge Theory,"
1194: {\bf JHEP} 0210:008, 2002 {\tt hep-th/0206045};\\
1195: %%CITATION = HEP-TH 0206045;%%%
1196: H. Dimov, V. Filev, R.C. Rashkov and K.S. Viswanathan, ``Semiclassical Quantization of
1197: Rotating Strings in Pilch-Warner Geometry," Phys. Rev. {\bf D68} 066010, 2003,
1198: {\tt hep-th/0304035};
1199: %%CITATION = HEP-TH 0304035;%%%
1200: S. Benvenuti and M. Kruczenski, ``Semiclassical Strings in Sasaki-Einstein Manifolds
1201: and Long Operators in N=1 Gauge Theories," {\tt hep-th/0505046};\\
1202: %%CITATION = HEP-TH 0505046;%%%
1203: S. Benvenuti and M. Kruczenski, ``From Sasaki-Einstein Spaces to quives via BPS
1204: geodesics: Lpqr," {\tt hep-th/0505206}.
1205: %%CITATION = HEP-TH 0505206;%%%
1206:
1207: \bibitem{GN}
1208: U. Gursoy and C. Nunez, ``Dipole Deformations of N=1 SYM and Supergravity Backgrounds
1209: with $U(1)\times U(1)$ Global Symmetry," {\tt hep-th/0505100}.
1210: %%CITATION = HEP-TH 0505100;%%%
1211:
1212: \bibitem{CA}
1213: C. Ahn and J.F. Vazquez-Portiz, ``Marginal Deformations with $U(1)^3$ Global Symmetry,"
1214: {\tt hep-th/0505168}.
1215: %%CITATION = HEP-TH 0505168;%%%
1216:
1217: \bibitem{Lee}
1218: J.P. Gauntlett, S. Lee, T Mateos and D. Waldram, ``Marginal Deformations of Field Theories
1219: with AdS$_4$ Duals," {\tt hep-th/0505207}.
1220: %%CITATION = HEP-TH 0505207;%%%
1221:
1222: \bibitem{NP}
1223: V. Niarchos and N. Prezas, ``BMN Operators for ${\cal N}=1$ superconformal Yang-Mills
1224: theories and associated string backgrounds," JHEP {\bf 0306}, 015 (2003),
1225: {\tt hep-th 0212111}.
1226: %%CITATION = HEP-TH 0212111;%%%
1227:
1228: \bibitem{HPP}
1229: G. Papadopoulos, J.G. Russo and A.A. Tseytlin, ``Solvable Model of Strings in a time
1230: dependent plane-wave background," Class. Quant. Grav. {\bf 20}, 969 (2003),
1231: {\tt hep-th/0211289};\\
1232: %%CITATION = HEP-TH 0211289;%%%
1233: M. Blau and M. O'Loughlin, ``Homogeneous Plane Waves," Nucl. Phys. {\bf B654}, 135 (2003),
1234: {\tt hep-th/0212135}.
1235: %%CITATION = HEP-TH 0212135;%%%
1236:
1237: \bibitem{SHPP}
1238: R.R. Metsaev, ``Type IIB Green-Schwarz Superstring in Plane Wave Ramond-Ramond Background,"
1239: Nucl. Phys. {\bf B625} 70 (2002), {\tt hep-th/0112044};\\
1240: %%CITATION = HEP-TH 0112044;%%%
1241: R.R. Metsaev and A.A. Tseytlin, ``Exactly Solvable Model of Superstring in Plane Wave
1242: Ramond-Ramond Background," Phys. Rev. {\bf D65} 126004 (2002), {\tt hep-th/0202109};\\
1243: %%CITATION = HEP-TH 0202109;%%%
1244: J.G. Russo and A.A. Tseytlin, ``On Solvable Models of Type IIB Superstring in
1245: NS-NS and R-R Plane Wave Backgrounds," JHEP {\bf 0204} 021 (2002), {\tt hep-th/0202179};\\
1246: %%CITATION = HEP-TH 0202179;%%%
1247: M. Blau, M. O'Loughlin, G. Papadopoulos and A.A. Tseytlin, ``Solvable Models of Strings in
1248: Homogeneous Plane Wave Backgrounds," Nucl. Phys. {\bf B673}, 57 (2003), {\tt hep-th/0304198.}
1249: %%CITATION = HEP-TH 0304198;%%%
1250:
1251: \bibitem{Toni}
1252: T. Mateos, ``Marginal deformations of N=4 SYM and Penrose limits with continuum spectrum,"
1253: {\tt hep-th/0505243}.
1254: %%CITATION = HEP-TH 0505243;%%%
1255:
1256: \bibitem{Motl}
1257: N.R. Constable, D.Z. Freedman, M. Headrick, S. Minwalla, L. Motl, A. Postnikov and W. Skiba,
1258: ``PP-wave String Interactions from perturbative Yang-Mills Theory,"
1259: {\tt hep-th/0205089}.
1260: %%CITATION = HEP-TH 0205089;%%%
1261:
1262: \bibitem{Penati}
1263: S. Penati, A. Santambrogio and Daniela Zanon, ``Two point functions
1264: of Chiral Operators in N=4 SYM at order $g^4$," JHEP {\bf 9912}
1265: 006,(1999), {\tt hep-th/9910197};\\
1266: %%CITATION = HEP-TH 9910197;%%%
1267: Silvia Penati, Alberto Santambrogio and Daniela Zanon, ``More on
1268: Correlators and Contact Terms in N=4 SYM at order $g^4$,"
1269: Nucl. Phys. {\bf B593} 651-670 (2001), {\tt hep-th/0005223};\\
1270: %%CITATION = HEP-TH 0005223;%%%
1271: Silvia Penati and Alberto Santambrogio, ``Superspace Approach to
1272: Anomalous Dimensions in N=4 SYM," Nucl. Phys. {\bf B614} 367-387
1273: (2001) {\tt hep-th/0107071};\\
1274: %%CITATION = HEP-TH 0107071;%%%
1275: Alberto Santambrogio and Daniela Zanon, ``Exact Anomalous Dimensions
1276: of N=4 Yang-Mills Operators with Large R Charge," Phys. Lett. {\bf B545} 425-429 (2002) {\tt hep-th/0206079}.
1277: %%CITATION = HEP-TH 0206079;%%%
1278:
1279: \bibitem{HFS}
1280: E. D'Hoker, D.Z. Freedman and W.Skiba, ``Field Theory tests for correlators in the
1281: AdS/CFT Correspondence," Phys. Rev. {\bf D59} 045008 (1999), {\tt hep-th/9807098}.
1282: %%CITATION = HEP-TH 9807098;%%%
1283:
1284: \bibitem{Berenstein}
1285: D. Berenstein, {\it A Toy Model for the AdS/CFT Correspondence,} {\tt hep-th/0403110}
1286: %%CITATION = HEP-TH 0403110;%%%
1287:
1288: \bibitem{DHK}
1289: N. Dorey, T.J. Hollowood and S.P. Kumar, ``$S$ Duality of the Leigh-Strassler Deformation
1290: via Matrix Models,"
1291: JHEP {\bf 0212} 003 (2002), {\tt hep-th/0210239}.
1292: %%CITATION = HEP-TH 0210239;%%%
1293:
1294:
1295: \end{thebibliography}
1296: \end{document}
1297:
1298:
1299: