1: \documentclass{elsart}
2: %\documentclass[fleqn,12pt]{article}
3: \usepackage{latexsym,amsmath,tabularx,amssymb,bm}
4: %\usepackage{amssymb}
5: \usepackage{epsfig}
6: \usepackage{citesort}
7: %\usepackage[notref,notcite]{showkeys}
8:
9: \long\def\comment#1{ }
10:
11: \setlength{\textwidth}{16cm}
12: \setlength{\textheight}{22.9cm}
13: \setlength{\oddsidemargin}{0.26cm}
14: \setlength{\evensidemargin}{0.26cm}
15: \setlength{\topmargin}{-0.04cm}
16:
17:
18: %\setlength{\evensidemargin}{0cm} \setlength{\oddsidemargin}{0cm}
19: %\setlength{\textwidth}{15.5cm} \setlength{\textheight}{21.4cm}
20: \usepackage{graphicx}
21: \usepackage{amssymb}
22: \vbadness=5000 \hbadness=5000 \hfuzz=30pt
23: \parindent 20pt
24: \def\Bs{{\textbackslash}}
25: \usepackage{makeidx}
26: \usepackage{graphicx}
27: \usepackage{multicol}
28:
29:
30:
31: \def\simge{\mathrel{%
32: \rlap{\raise 0.511ex \hbox{$>$}}{\lower 0.511ex \hbox{$\sim$}}}}
33: \def\simle{\mathrel{
34: \rlap{\raise 0.511ex \hbox{$<$}}{\lower 0.511ex \hbox{$\sim$}}}}
35:
36: %\input{macro1}
37: %\usepackage[french]{babel}
38: %\usepackage[T1]{fontenc}
39: %\usepackage[latin1]{inputenc}
40: %\nofiles
41:
42: \newcommand \beq{\begin{eqnarray}}
43: \newcommand \eeq{\end{eqnarray}}
44: \newcommand{\del}{\partial}
45:
46: \begin{document}
47: \begin{flushright}
48: ~\vspace{-1.25cm}\\
49: {\small\sf ECT*-- 05-02}
50: \end{flushright}
51: \vspace{0.8cm}
52: \begin{frontmatter}
53:
54: \parbox[]{16.6cm}{ \begin{center}
55:
56: \title{A new method to solve the Non Perturbative Renormalization Group equations}
57:
58: \author[ect]{J.-P. Blaizot \thanksref{cnrs}},
59: \author[montevideo]{Ram\'on M\'endez Galain\thanksref{email2}},
60: \author[montevideo]{Nicol\'as Wschebor\thanksref{email}}
61:
62: \address[ect]{$ECT*$, Villa Tambosi, Strada delle Tabarelle 286,
63: I-38050 Villazzano(TN), Italy}
64:
65: \address[montevideo]{ Instituto de F\'{\i}sica, Facultad de Ingenier\'{\i}a,
66: J.H.y Reissig 565, 11000
67: Montevideo, Uruguay}
68:
69: \thanks[cnrs]{Membre du Centre National de la Recherche Scientifique
70: (CNRS), France.}
71:
72: \thanks[email2]{email: mendezg@fing.edu.uy}
73:
74: \thanks[email]{email: nicws@fing.edu.uy}
75:
76:
77: \date{\today}
78: \vspace{0.8cm}
79: \begin{abstract}
80: We propose a method to solve the Non Perturbative Renormalization Group equations
81: for the $n$-point functions. In leading order, it
82: consists in solving the equations obtained by closing the infinite
83: hierarchy of equations for the $n$-point functions. This is
84: achieved: i) by exploiting the decoupling of modes and the analyticity of the $n$-point functions at small momenta: this allows us
85: to neglect some momentum dependence of the vertices entering the
86: flow equations; ii) by relating vertices at zero momenta to
87: derivatives of lower order vertices with respect to a constant
88: background field. Although the approximation is not controlled by
89: a small parameter, its accuracy can be systematically improved.
90: When it is applied to the $O(N)$ model, its leading order is exact in
91: the large $N$ limit; in this case, one
92: recovers known results in a simple and direct way, i.e., without introducing
93: an auxiliary field.
94:
95: \end{abstract}
96: \end{center}}
97:
98: \end{frontmatter}
99: \newpage
100:
101:
102: \section{Introduction}
103:
104:
105: In the last ten years a considerable amount of effort has been
106: devoted to the study of the Non Perturbative Renormalization Group
107: (NPRG), and to the development of new approximation schemes to solve
108: the corresponding infinite hierarchy of equations for the $n$-point
109: functions \cite{Tetradis94,Ellwanger94a,Morris94b,Morris94c}.
110: In this context, the so-called ``derivative expansion" has been
111: widely used. It defines a systematic approach that has led to
112: many successful applications in a variety of domains
113: \cite{Berges00,Bagnuls:2000ae,Delamotte:2004zg}. However, the
114: derivative expansion is a good approximation to $n$-point
115: functions only when the external momenta are smaller than the
116: lowest mass in the problem. In particular, for massless theories,
117: the derivative expansion provides information only on the
118: $n$-point functions (and their derivatives) at zero momenta.
119: This is not enough for applications which require the knowledge
120: of the full momentum dependence of the $n$-point functions. In these cases,
121: a new approximation scheme is necessary.
122:
123: To our knowledge, all efforts in this direction \cite{truncation}
124: have been based on various forms
125: of the early proposal by Weinberg \cite{weinberg73}:
126: one {\it truncates} the infinite tower of flow equations for the $n$-point
127: functions considering only vertices up to a given number of legs,
128: possibly using various ansatzs for some of them.
129: This leads to approximations similar to those
130: used when solving Schwinger-Dyson equations \cite{alkofer}.
131: However, despite the fact that very encouraging results have been obtained
132: problems remain with truncations around zero
133: external fields \cite{convergence}.
134:
135:
136: It is then interesting to look for alternative schemes to solve NPRG at finite
137: external momenta and it is
138: the purpose of this Letter to present a method to calculate correlation functions at arbitrary
139: momenta within the NPRG. As it is the case for the derivative expansion,
140: the proposed strategy takes into account,
141: at each order, an infinite number of vertices.
142: The accuracy is expected to be comparable to that achieved with the
143: derivative expansion when it is used to calculate the effective potential.
144: The method exploits two properties of the NPRG: the decoupling of
145: short wave-length modes, and the
146: analyticity of the $n$-point functions at small momentum
147: (guaranteed by the infrared regulator), in order to neglect some
148: of the momentum dependence in the vertices which enter the flow
149: equations. Then, it becomes possible to close the infinite
150: hierarchy of equations by calculating vertices at zero momenta as
151: derivatives of lower order vertices with respect to a constant
152: background field. Although the approximation is not controlled by
153: any small parameter, its accuracy can be systematically improved, in
154: a way similar to what can be done in the derivative expansion.
155:
156: The leading order of the proposed approximation scheme shares some of
157: the nice properties of the leading order of the derivative
158: expansion, the so-called local potential approximation (LPA): it
159: is exact at one loop order, and it also reproduces the exact
160: large $N$ limit of the $O(N)$ model. Of course, the LPA provides
161: only information on momentum independent quantities, such as the
162: effective potential, whereas the present method does not have this
163: limitation. In fact, as an illustration, we use it to solve the NPRG equations for
164: the full momentum dependent $n$-point functions of the $O(N)$
165: model in the large $N$ limit. This can be done directly, i.e.,
166: without having to introduce an auxiliary
167: field. To our knowledge, this is the first time that, within the NPRG,
168: such a calculation is done for this model.
169:
170: The method presented here is an improvement of the one that we have applied in a
171: previous paper to the calculation of the transition temperature of a weakly repulsive Bose gas
172: \cite{Blaizot:2004qa}. Compared to the method proposed in
173: \cite{Blaizot:2004qa}, the present one is conceptually simpler as it
174: involves a single approximation. Its numerical implementation will be discussed in a
175: separate publication \cite{BMW05_3}.
176:
177:
178: \section{The NPRG and the derivative expansion}
179:
180:
181: Let us start by recalling some basic features of the NPRG. Although most of the arguments in
182: this paper have a wider range of applicability, we shall consider a scalar field
183: theory defined by an Euclidean action $S$ in $d$ dimensions, of
184: the generic form:
185: \begin{equation}\label{eactON}
186: S = \int {\rm d}^{d}x\,\left\lbrace{ 1 \over 2} \left(\del_\mu
187: \varphi(x)\right)^2 +V(\varphi)\right\rbrace \,,
188: \end{equation} where $V(\varphi)$ is a polynomial in $\varphi$. It
189: is understood that the parameters of the action (hidden in
190: $V(\varphi)$) and the field normalization are fixed at a
191: microscopic momentum scale $\Lambda$ (which may be infinite). The
192: NPRG equations relate the classical action to the full effective
193: action. This relation is obtained by controlling the magnitude of
194: long wavelength field fluctuations with the help of an infrared
195: cut-off, which is implemented
196: \cite{Tetradis94,Ellwanger94a,Morris94b,Morris94c} by adding to
197: the classical action (\ref{eactON}) a regulator of the form
198: \begin{equation}
199: \Delta S_\kappa[\varphi] =\frac{1}{2} \int \frac{{\rm d}^dp}{(2\pi)^d}
200: R_\kappa(p)
201: \varphi(p)\varphi(-p),
202: \end{equation}\normalsize
203: where $R_\kappa$ denotes a family of ``cut-off functions''
204: depending on a parameter $\kappa$. The role of $\Delta S_\kappa$
205: is to suppress the fluctuations with momenta $p\simle \kappa$,
206: while leaving unaffected the modes with $p\simge \kappa$. Thus,
207: typically $ R_{\kappa}(p)\to\kappa^2$ when $ p \ll \kappa$, and
208: $R_{\kappa}(p)\to 0$
209: when $ p\simge \kappa$.
210: There is a large
211: freedom in the choice of $R_\kappa(p)$, abundantly discussed in the literature
212: \cite{Ball95,Comellas98,Litim,Canet02}.
213:
214:
215: We denote the effective action in the presence of the regulator by
216: $\Gamma_\kappa[\phi]$, where $\phi$ is the average field, $\phi(x)=\left\langle \varphi(x)
217: \right\rangle$. When $\kappa \to \Lambda$ quantum fluctuations are suppressed and $\Gamma_\Lambda[\phi]$ coincides with the classical action. As $\kappa$ decreases,
218: more and more quantum fluctuations are taken into account and, as $\kappa\to 0$, $\Gamma_{\kappa=0}[\phi]$ becomes the usual effective action $\Gamma[\phi]$. In other words, as
219: $\kappa$ decreases from $\Lambda$ to $0$, $\Gamma_\kappa$
220: interpolates between the classical action and the full
221: effective action (see e.g. \cite{Berges00}). The variation
222: with $\kappa$ of $\Gamma_\kappa[\phi]$ is governed by the
223: following flow
224: equation \cite{Tetradis94,Ellwanger94a,Morris94b,Morris94c}:
225: \begin{equation}
226: \label{NPRGeq}
227: \partial_\kappa \Gamma_\kappa[\phi]=\frac{1}{2}\int \frac{d^dq}{(2\pi)^d} \partial_\kappa R_\kappa(q^2)
228: \left[\Gamma_\kappa^{(2)}+R_\kappa\right]^{-1}_{q,-q},
229: \end{equation}
230: where $\Gamma_\kappa^{(2)}$ is the second derivative of $\Gamma_\kappa$ w.r.t. $\phi$ (see Eq.~(\ref{gamma2}) below).
231: Eq.~\ref{NPRGeq} is the master equation of the NPRG. Its right
232: hand side has the structure of a one loop integral, with one
233: insertion of $\partial_\kappa R_\kappa(q^2)$.
234:
235:
236:
237: As well known \cite{Zinn-Justin:2002ru}, the effective action $\Gamma[\phi]$ is the generating functional
238: of the one-particle irreducible $n$-point functions. Similarly, for a given value of $\kappa$, we define the $n$-point functions $\Gamma_\kappa^{(n)}$:
239: \begin{equation}
240: \Gamma_\kappa^{(n)}(p_1,\dots,p_n)=\int d^dx_1\dots\int d^dx_{n-1}
241: e^{i\sum_{j=1}^n p_jx_j}\frac{\delta^n\Gamma_\kappa}{\delta\phi(x_1)
242: \dots \delta\phi(x_n)}.
243: \end{equation}
244: By deriving eq.~(\ref{NPRGeq}) with respect to $\phi$, and then
245: letting the field be constant, one gets the flow equations for all $n$-point
246: functions in a constant background field $\phi$. These equations can
247: be represented diagramaticaly by one loop diagrams
248: with dressed vertices and propagators (see e.g. \cite{Berges96}). For instance, the
249: flow of the 2-point
250: function in a constant external field reads:
251: \beq
252: \label{gamma2}
253: \partial_\kappa\Gamma_\kappa^{(2)}(p,-p;\phi)&=&\int
254: \frac{d^dq}{(2\pi)^d}\partial_\kappa R_k(q)\left\{G_\kappa(q^2;\phi)\Gamma_\kappa^{(3)}(p,q,-p-q;\phi)\right. \nonumber \\
255: &&\times G_\kappa((q+p)^2;\phi)\Gamma_\kappa^{(3)}(-p,p+q,-q;\phi)G_\kappa(q^2;\phi) \nonumber \\
256: &&\left.-\frac{1}{2}G_\kappa(q^2;\phi)\Gamma_\kappa^{(4)}(p,-p,q,-q;\phi)G_\kappa(q^2;\phi)\right\},
257: \eeq where
258: \begin{equation}\label{G-gamma2}
259: G^{-1}_\kappa (q^2;\phi) = \Gamma^{(2)}_\kappa (q,-q;\phi) +
260: R_\kappa(q^2).
261: \end{equation}The corresponding diagrams contributing to the flow are shown
262: in Fig.~\ref{2-point-diagrams}.
263:
264: \begin{figure}
265: \begin{center}
266: \includegraphics[scale=.5] {Pescado2.eps} \hspace{20mm}
267: \includegraphics[scale=.5] {Pulpo2.eps}
268: \end{center}
269: \caption{The two diagrams contributing to the flow of the 2-point vertex. The lines represent dressed propagators, $G_\kappa$. The
270: cross represents an insertion of $\partial_\kappa R_k$. The vertices denoted by black dots are $\Gamma^{(3)}_\kappa$ and $\Gamma^{(4)}_\kappa$. \label{2-point-diagrams}}
271: \end{figure}
272:
273:
274: Flow equations for the $n$-point functions do not close: for example, in order to solve eq.~(\ref{gamma2}) one needs
275: the 3- and the 4-point functions, $\Gamma_\kappa^{(3)}$ and $\Gamma_\kappa^{(4)}$ respectively. At this point we observe that because of the shape of the regulator, only internal momenta $q$ smaller than $\kappa$ contribute to the flow, i.e., to the integral in the r.h.s. of eq.~(\ref{gamma2}). One refers to this property as to the decoupling of high momentum modes. Besides, the regulator insures also that all vertices are smooth functions of momenta.
276: Suppose then that one wants to calculate $n$-point
277: functions at small external momenta, say, $p_i^2 \simle
278: \kappa^2$. Then all momenta are small.
279: This, together with the fact that the $n$-point functions are smooth functions of the momenta, make it
280: possible to expand the $n$-point functions $\Gamma_\kappa^{(n)}$ in the r.h.s. of the flow equations in terms of $q^2/\kappa^2$ and $p^2/\kappa^2$, or equivalently in terms of the derivatives of the field.
281:
282:
283: Such considerations are at the basis of the derivative expansion, which is usually formulated in terms of an ansatz for the effective action, rather than in terms of an approximation for the $n$-point functions.
284: Its zero order, the LPA, assumes that the effective action has the form:
285: \begin{equation}\label{gammaLPA} \Gamma_k^{LPA}[\phi]=\int d^dx \left\{
286: \frac{1}{2}\partial_{\mu}\phi\partial_{\mu}\phi+V_k(\phi)\right\}.
287: \end{equation}
288: The derivative term here is simply the one appearing in the
289: classical action, and $V_k(\phi)$ is the effective potential. The
290: flow equation for $V_\kappa$ is a closed equation that is easily
291: obtained by assuming that the field $\phi$ is constant in
292: eq.~(\ref{NPRGeq}):
293: \begin{equation}
294: \label{pot}
295: \partial_\kappa V_\kappa (\phi)=\frac{1}{2}\int \frac{d^dq}{(2\pi)^d} \partial_\kappa R_\kappa(q^2) \; G_\kappa^{LPA}(q^2;\phi),
296: \end{equation}
297: with $G_\kappa^{LPA}(q^2;\phi)$ given by Eq.~(\ref{G-gamma2}) in which
298: \begin{equation}
299: \label{prop}
300: \Gamma_\kappa^{(2)}(q,-q;\phi)=q^2+\del ^2V_\kappa/\del\phi^2,
301: \end{equation}
302: as obtained from Eq.~(\ref{gammaLPA})
303:
304: Higher order corrections to the LPA include terms in the effective action with an
305: increasing number of derivatives. Although there is no formal
306: proof of convergence,
307: the expansion exhibits quick apparent
308: convergence if the
309: regulator $R_\kappa(q^2)$ is appropriately chosen \cite{Litim01,Canet02,Canet03}. In practice, the LPA reproduces well the physical quantities dominated by small momenta (such as the effective potential or
310: critical exponents) in all theories where it has been tested (see,
311: for example, \cite{Berges00,Delamotte:2004zg}). Higher order corrections
312: improve the results. The expansion has been pushed
313: up to third order \cite{Canet03}, yielding critical exponents in
314: the Ising universality class as good as those obtained with the
315: best accepted methods.
316:
317: \section{A new approximation scheme }
318:
319: The derivative expansion, as described above, strictly makes sense only for momenta
320: not much larger than $\sqrt{\kappa^2+m_\kappa^2}$, where $m_\kappa$ is the running mass. Thus, at criticality, and in the physical limit $\kappa \to 0 $, it provides information only on $n$-point functions and their derivatives at zero momenta. However, by focusing on the $n$-point functions rather than the effective
321: action, we can generalize slightly the arguments on which the derivative expansion is based in order to set up a much more powerful approximation scheme. We observe that:
322: i) the momentum $q$ circulating in the loop integral of a flow equation
323: is limited by $\kappa$; ii) the
324: smoothness of the $n$-point functions allows us to make an expansion in powers of $q^2/\kappa^2$, independently of the value of the external momenta $p$. Now, a typical $n$-point function entering a flow equation is of the from $\Gamma^{(n)}_\kappa(p_1,p_2,...,p_{n-1}+q,p_n-q;\phi)$, where $q$ is the loop momentum. The proposed approximation scheme, \emph{in its leading order}, will then consist in neglecting the $q$-dependence of such vertex functions:
325: \begin{equation}\label{approx}
326: \Gamma^{(n)}_\kappa(p_1,p_2,...,p_{n-1}+q,p_n-q;\phi)\sim
327: \Gamma^{(n)}_\kappa(p_1,p_2,...,p_{n-1},p_n;\phi).
328: \end{equation}
329: Note that this approximation is a priori well justified.
330: Indeed, when all the external momenta $p_i=0$, eq.~(\ref{approx})
331: is the basis of the LPA which, as stated above, is a good
332: approximation. When the external momenta $p_i$ begin to grow, the
333: approximation in eq.~(\ref{approx}) becomes better and better, and it is
334: trivial when all momenta are much larger than $\kappa$. With this
335: approximation, eq.~(\ref{gamma2}) for instance becomes:
336: \begin{eqnarray}
337: \label{gamma2app}
338: \partial_\kappa\Gamma_\kappa^{(2)}(p,-p;\phi)&=&\int
339: \frac{d^dq}{(2\pi)^d}\partial_\kappa R_k(q^2)\left\{G_\kappa(q^2;\phi)\Gamma_\kappa^{(3)}(p,0,-p;\phi)\right. \nonumber \\
340: &&\times G_\kappa((q+p)^2;\phi)\Gamma_\kappa^{(3)}(-p,p,0;\phi)G_\kappa(q^2;\phi) \nonumber \\
341: &&\left.-\frac{1}{2}G_\kappa(q^2;\phi)\Gamma_\kappa^{(4)}(p,-p,0,0;\phi)G_\kappa(q^2;\phi)\right\}
342: \end{eqnarray}
343: Note that we do not also assume $q=0$ in the propagators. The reason for this will become clear shortly.
344:
345:
346: Now comes the second ingredient of the approximation scheme, which
347: exploits the advantage of working with a non vanishing background
348: field: vertices evaluated at zero external momenta can be
349: written as derivatives of vertex functions with a smaller number
350: of legs (this observation can be found, in a different
351: context, in \cite{Tetradis94} and in \cite{Golner:1998sr}):
352: \begin{equation}
353: \label{faireq=0}
354: \Gamma_\kappa^{(n+1)}(p_1,p_2,...,p_n,0;\phi)=\frac{\partial
355: \Gamma_\kappa^{(n)}(p_1,p_2,...p_n;\phi)} {\partial \phi}.
356: \end{equation}
357: To prove this relation, expand $\Gamma_k[\phi]$ around an arbitrary constant field
358: $\phi^0$:
359: \begin{eqnarray}
360: \label{seriedelafonctionnelle} \Gamma_\kappa[\phi]&=&\sum_n
361: \frac{1}{n!}\int d^dx_1 ... d^dx_n\, [\phi(x_1)-\phi^0]...
362: [\phi(x_n)-\phi^0]\,\Gamma^{(n)}_\kappa(x_1,...,x_n;\phi_0) .
363: \end{eqnarray}
364: Since $\Gamma_\kappa[\phi]$ does not depend on $\phi^0$, $\del\Gamma_\kappa[\phi]/\del \phi^0=0$.
365: Taking the derivative of eq.~(\ref{seriedelafonctionnelle}) with respect to $\phi_0$ one then gets:
366: \begin{equation} \int
367: d^dy\,
368: \Gamma_\kappa^{(n+1)}(x_1,x_2,...,x_n,y,\phi^0)=\frac{\partial
369: \Gamma_\kappa^{(n)}(x_1,x_2,...x_n,\phi^0)} {\partial \phi^0}
370: ,\end{equation}from which eq.~(\ref{faireq=0}) follows after a
371: Fourier transform. By exploiting eq.~(\ref{faireq=0}), one easily
372: transforms eq.~(\ref{gamma2app}) into a {\it closed equation} (recall
373: that $G_\kappa$ and $\Gamma_\kappa^{(2)}$ are related by
374: eq.~(\ref{G-gamma2})):
375: \begin{eqnarray}\label{2pointclosed}
376: &&\partial_\kappa\Gamma_\kappa^{(2)}(p^2;\phi)=\int
377: \frac{d^dq}{(2\pi)^d}\,\partial_\kappa R_\kappa(q^2) \; G^2_\kappa(q^2;\phi) \nonumber \\
378: &&
379: \times \left\{ \left( \frac{\partial \Gamma_\kappa^{(2)}(p,-p;\phi)}
380: {\partial \phi} \right)^2 G_\kappa ((p+q)^2;\phi)
381: \; - \; \frac{1}{2}\frac{\partial^2 \Gamma_\kappa^{(2)}(p,-p;\phi)}
382: {\partial \phi^2}\right\}.
383: \end{eqnarray}
384:
385: It is interesting to emphasize the similarity of this equation with
386: eq.~(\ref{pot}): both are closed equations because
387: the vertices appearing in the r.h.s. have been expressed
388: as derivatives of the function in the l.h.s..
389: This is the main result of this paper.
390:
391: The approximation scheme presented here is similar to that used in \cite{Blaizot:2004qa}. There also the momentum dependence of the vertices was neglected in the leading order. However further approximations were needed in order to close the hierarchy. The progress realized here is to bypass these extra approximations by working in a constant background field.
392:
393: The construction of closed equations for the $n$-point
394: functions with arbitrary $n$ $ (n \geq 3$) follows the same lines as that of the equation for the 2-ppoint function.
395: The flow equation for $\Gamma_\kappa^{(n)}$ involves all $m$-point
396: functions with $m\le n+2$. Since the r.h.s. of the flow equations
397: have the structure of one loop integrals, the contributions
398: involving $\Gamma_\kappa^{(n+1)}$ and $\Gamma_\kappa^{(n+2)}$
399: are of the types shown in Fig.~\ref{n-point-diagrams}. When the loop momentum
400: entering these vertices is taken to be zero, in line with eq.~(\ref{approx}), the vertex of order $n+1$ has one leg at zero
401: momentum and the vertex of order $n+2$ has two; thus, according to eq.~(\ref{faireq=0}), they can be written as derivatives of the $n$-point function $\Gamma_\kappa^{(n)}$. It follows that
402: the equation for $\Gamma_\kappa^{(n)}$ is a closed equation,
403: assuming of course that all other functions $\Gamma_\kappa^{(m)}$
404: with $m<n$ are determined similarly.
405:
406:
407: \begin{figure}
408: \begin{center}
409: \includegraphics[scale=.5] {Pescado.eps}\hspace{25mm}
410: \includegraphics[scale=.5] {Pulpo.eps}
411: \end{center}
412: \caption{The two diagrams contributing to the flow of the
413: $n$-point vertex and containing vertices with more than $n$ legs.
414: \label{n-point-diagrams}}
415: \end{figure}
416:
417:
418: One can go beyond the leading order approximation, based on
419: eq.~(\ref{approx}), in the following way. Focusing on
420: eq.~(\ref{gamma2}) for $\Gamma_\kappa^{(2)}(p,-p;\phi)$ one can solve simultaneously the flow equations of
421: $\Gamma_\kappa^{(2)}$, $\Gamma_\kappa^{(3)}$ and $\Gamma_\kappa^{(4)}$, with no
422: approximation in the flow equation for $\Gamma_\kappa^{(2)}$, but using
423: (\ref{approx}) in the right-hand-side of the flow equations for
424: $\Gamma_\kappa^{(3)}$ and $\Gamma_\kappa^{(4)}$. In this way one would determine
425: $\Gamma_\kappa^{(3)}$ and $\Gamma_\kappa^{(4)}$ with a ``leading order"
426: precision and $\Gamma_\kappa^{(2)}$ with a ``next-to-leading order" one.
427: By iterating the procedure, which amounts to including more equations, one would get better approximations
428: for a larger number of $n$-point functions. It is easily verified that such an iterative approximation scheme
429: reproduces
430: perturbation theory at high momenta: for instance, if the bare
431: vertices of the theory are momentum independent, the leading order
432: approximation contains the exact one loop (this is only true if we keep
433: $q$ in the propagators, as we did in (\ref{gamma2app})). More generally,
434: a simple analysis shows that to get the expression of a given
435: $n$-point functions at order $m$ loop, one has to consider all
436: equations up to that for the $n+2m-2$-point function. (In a theory with
437: momentum dependent bare vertices, perturbation theory is also
438: reproduced, but more flow equations are required to match
439: perturbation theory at a given number of loops.) Of course, such a scheme is not only accurate at high momenta, where it reproduces perturbation theory, but also in the small momentum, possibly critical, regime, where its accuracy is at least comparable
440: to that of the derivative expansion. In practice, the scheme that we have described may become rapidly cumbersome. However, it can be simplified by using a further approximation consisting in a systematic expansion in $q^2/\kappa^2 $
441: of the $n$-point functions in the right-hand-side of flow equations. This further approximation preserves the essential property of leading to a closed set of equations. We have solved the flow equation for the 2-point function in leading order and obtained results which are comparable to those obtained in \cite{Blaizot:2004qa}. More complete numerical studies will be presented in a forthcoming publication.
442:
443:
444: \section{The $O(N)$ model in the large $N$ limit}
445:
446: We turn now to the $O(N)$ model in the large $N$ limit, where analytical results can be obtained. Our goal here is twofold. First,
447: to prove that, in its leading order, our approximation is in fact
448: exact in the large $N$ limit. Second, to solve analytically the
449: NPRG equations and find the $n$-point functions in a direct way,
450: i.e., without introducing an auxiliary field, as commonly done
451: \cite{Moshe:2003xn}.
452:
453: The large $N$ limit of the $O(N)$ model within the NPRG has been
454: thoroughly studied in \cite{D'Attanasio97}, where it is shown
455: that the effective action takes the form:
456: \begin{equation}\label{largeNlimit} \Gamma_\kappa[\phi]=\int d^dx
457: \left\{\frac{1}{2}\partial_{\mu}\phi^a\partial_{\mu}\phi^a\right\}+\hat\Gamma_\kappa[\rho],
458: \end{equation} where $\rho(x)=\phi^a(x)\phi^a(x)/2$, $a=1,\dots , N$. Using
459: this expression, the authors of Ref.~\cite{D'Attanasio97} show
460: that the LPA flow equation for the effective potential is exact.
461: Moreover, they find an interesting analytical solution of the
462: equation. However, the use of (\ref{largeNlimit}) in the flow
463: equation of a $n$-point function at zero field does not give a
464: closed equation \cite{Blaizot05}. In this section we obtain a
465: system of closed equations that can be solved analytically.
466:
467:
468: We shall again focus, for the sake of illustration, on the 2-point
469: function. Its exact flow equation in an external field is the
470: generalization of eq.~(\ref{gamma2})) to the $O(N)$ model:
471: \begin{eqnarray}
472: \label{gamma2champnonnul}
473: \partial_\kappa\Gamma_{ab}^{(2)}(p,-p;\kappa;\phi)&=&\int
474: \frac{d^dq}{(2\pi)^d}\partial_\kappa R_\kappa(q^2)\left\{G_{ij}(q^2;\kappa;\phi)\Gamma_{ajk}^{(3)}(p,q,-p-q;\kappa;\phi)\right. \nonumber \\
475: &&\times G_{kl}((q+p)^2;\kappa;\phi)\Gamma_{blm}^{(3)}(-p,p+q,-q;\kappa;\phi)G_{mi}(q^2;\kappa;\phi) \nonumber \\
476: &&\left.-\frac{1}{2}G_{ij}(q^2;\kappa;\phi)\Gamma_{abjk}^{(4)}(p,-p,q,-q;\kappa;\phi)G_{ki}(q^2;\kappa;\phi)\right\}.
477: \end{eqnarray}
478:
479:
480: At this point, it is useful to recall the structure of the $n$-point functions
481: of the $O(N)$ model in the large N limit. These are obtained by taking the functional
482: derivatives of (\ref{largeNlimit}), letting the background field
483: to be constant, and then taking the Fourier transform. We obtain in this way
484: the following expressions for the 2-, 3- and 4-point vertices:
485: \beq \label{gamma2Ngrand}
486: \Gamma_{ab}^{(2)}(p,-p;\kappa;\phi)=(p^2+V'_\kappa
487: (\rho))\delta_{ab}+\phi_a\phi_b\hat{\Gamma}^{(2)}_\kappa(p,-p;\rho), \eeq
488: \beq
489: \Gamma_{abc}^{(3)}(p_1,p_2,p_3;\kappa;\phi)&=&\phi_a\delta_{bc}\hat{\Gamma}_\kappa^{(2)}(p_1,-p_1;\rho)+\phi_b\delta_{ac}\hat{\Gamma}_\kappa^{(2)}(p_2,-p_2;\rho)\nonumber \\
490: &+&\phi_c\delta_{ab}\hat{\Gamma}_\kappa^{(2)}(p_3,-p_3;\rho)
491: +\phi_a\phi_b\phi_c\hat{\Gamma}_\kappa^{(3)}(p_1,p_2,p_3;\rho),
492: \eeq
493: \beq\label{Gamma4}
494: &&\Gamma_{abcd}^{(4)}(p_1,p_2,p_3,p_4;\kappa;\phi)=\delta_{ab}\delta_{cd}\hat{\Gamma}_\kappa^{(2)}(p_1+p_2,-p_1-p_2;\rho)\nonumber \\
495: &&+\delta_{ac}\delta_{bd}\hat{\Gamma}_\kappa^{(2)}(p_1+p_3,-p_1-p_3;\rho)
496: +\delta_{ad}\delta_{bc}\hat{\Gamma}_\kappa^{(2)}(p_1+p_4,-p_1-p_4;\rho)\nonumber \\
497: &&+\delta_{ab}\phi_c\phi_d\hat{\Gamma}_\kappa^{(3)}(p_1+p_2,p_3,p_4;\rho)
498: +\delta_{ac}\phi_b\phi_d\hat{\Gamma}_\kappa^{(3)}(p_1+p_3,p_2,p_4;\rho)\nonumber \\
499: &&+\delta_{ad}\phi_b\phi_c\hat{\Gamma}_\kappa^{(3)}(p_1+p_4,p_2,p_3;\rho)
500: +\delta_{bc}\phi_a\phi_d\hat{\Gamma}_\kappa^{(3)}(p_2+p_3,p_1,p_4;\rho) \nonumber \\
501: &&+\delta_{bd}\phi_a\phi_c\hat{\Gamma}_\kappa^{(3)}(p_2+p_4,p_1,p_3;\rho)
502: +\delta_{cd}\phi_a\phi_b\hat{\Gamma}_\kappa^{(3)}(p_3+p_4,p_1,p_2;\rho)\nonumber \\
503: &&+\phi_a\phi_b\phi_c\phi_d\hat{\Gamma}_\kappa^{(4)}(p_1,p_2,p_3,p_4;\rho) \; .
504: \eeq
505: As for the propagator, it can be written in terms of its
506: longitudinal and transverse components:
507: \begin{equation}
508: G_{ab}(p^2;\phi)=G_T(p^2;\rho)\left(\delta_{ab}-\frac{\phi_a\phi_b}{2\rho}\right)
509: +G_L(p^2;\rho)\frac{\phi_a\phi_b}{2\rho},
510: \end{equation}
511: with
512: \begin{eqnarray}
513: \label{propagatorsTL}
514: G_T^{-1}(p^2;\kappa;\rho)&=&p^2+V_\kappa'(\rho)+R_\kappa(p^2)\;,\nonumber \\
515: G_L^{-1}(p^2;\kappa;\rho)&=&p^2+V_\kappa'(\rho)+2\rho \hat{\Gamma}_\kappa^{(2)}(p,-p;\rho) +R_\kappa(p^2).
516: \end{eqnarray}
517:
518: We can now replace these expressions in
519: eq.~(\ref{gamma2champnonnul}) and keep only the leading terms in
520: $1/N$. To do so, notice that the
521: non trivial large $N$ limit of the $O(N)$ model is obtained when
522: the effective action $\Gamma_\kappa$ is considered to be of order
523: $N$ and the $\phi$ field of order $\sqrt{N}$ (see for example
524: \cite{D'Attanasio97}). Then, the 2-point function is of order 1,
525: the 3-point function of order $1/\sqrt{N}$ and the 4-point
526: function of order $1/N$. Simple counting of $N$ factors in the
527: flow equation (\ref{gamma2champnonnul}) gives a right hand side of order $1/N$. Thus, the
528: only way of having both sides of order 1 is to collect an explicit
529: $N$ factor from the trace of an identity matrix. The
530: surviving terms yield:
531: \begin{eqnarray}
532: \label{gamma2Ngrand+}
533: &&\partial_\kappa\Gamma_{ab}^{(2)}(p,-p;\kappa;\phi)=N\int
534: \frac{d^dq}{(2\pi)^d}\partial_\kappa R_\kappa(q^2)G_T^2(q^2;\kappa;\rho)
535: \left\{-\frac{1}{2}\delta_{ab}\hat{\Gamma}_\kappa^{(2)}(0,0;\rho)\right.\nonumber \\
536: &\qquad&\qquad\left.+\phi_a\phi_b\left(\hat{\Gamma}_\kappa^{(2)}(p,-p;\rho)\right)^2 G_T((p+q)^2;\kappa;\rho)
537: -\frac{1}{2}\phi_a\phi_b
538: \hat{\Gamma}_\kappa^{(3)}(p,-p,0;\rho)\right\}.
539: \end{eqnarray}
540:
541: Now comes the important observation: the vertices contributing to
542: the flow in the large-$N$ limit are $q$-independent. That means
543: that our leading order approximation becomes exact in the
544: large-$N$ limit of the $O(N)$ model, as anticipated. Notice that
545: for this to be true, we need to keep the $q$ dependence of the
546: propagators in the flow equations. However, in the large-N limit, the momentum
547: dependence of the transverse propagator (the only one appearing in
548: (\ref{gamma2Ngrand+})) is simply the bare
549: one, as shown in eq.~(\ref{propagatorsTL}).
550:
551: Substituting in eq.~(\ref{gamma2Ngrand+}) the expression (\ref{gamma2Ngrand}) of the 2-point function
552: and performing the tensor decomposition,
553: one obtains the system of
554: equations:
555: \begin{eqnarray}\label{closedeqnsON}
556: \partial_\kappa V'_\kappa(\rho)&=&-\frac{N}{2}V''_\kappa(\rho)\int \frac{d^dq}{(2\pi)^d}\partial_\kappa R_\kappa(q^2)G_T^2(q^2;\kappa;\rho) \nonumber \\
557: \partial_\kappa \hat{\Gamma}_\kappa^{(2)}(p,-p;\rho)&=&N\int \frac{d^dq}{(2\pi)^d}\partial_\kappa
558: R_\kappa(q^2)G_T^2(q^2;\kappa;\rho) \nonumber \\
559: &&\times\left\{\left(\hat{\Gamma}_\kappa^{(2)}(p,-p;\rho)\right)^2 G_T((p+q)^2;\kappa;\rho)
560: -\frac{1}{2} \frac{\partial \hat{\Gamma}_\kappa^{(2)}(p,-p;\rho)}{\partial \rho}\right\}
561: \end{eqnarray}
562: where we have used, as in eq.~(\ref{faireq=0}),
563: \begin{equation}
564: \hat{\Gamma}_\kappa^{(2)}(0,0;\rho)=V''_\kappa (\rho) ,\qquad
565: \hat{\Gamma}_\kappa^{(3)}(p,-p,0;\rho)=\frac{\partial \hat{\Gamma}_\kappa^{(2)}(p,-p;\rho)}{\partial \rho},
566: \end{equation}
567: which allowed us to close the equations. In the equations above, $V'_\kappa(\rho)$ and $V''_\kappa(\rho)$ denote respectively the first and second derivative of the effective potential with respect to $\rho$.
568:
569:
570: The first of eqs.~(\ref{closedeqnsON}) is simply the derivative of the equation for
571: the potential in the large-N limit which was solved analytically
572: using two different methods \cite{Tetradis95,D'Attanasio97}. The
573: second equation, which, to our knowledge, has not been presented
574: before in the literature, can be solved also using any of these
575: two methods. Here we present the solution using the method of
576: \cite{Tetradis95}.
577:
578: To this aim, we set $W=V'_\kappa (\rho)$, and use $(\kappa,W)$ as
579: independent variables instead of $(\kappa,\rho)$. In order to do
580: so, we introduce the inverse function relating $\rho$ to $W$:
581: \begin{equation}
582: \rho=f_\kappa(W) \end{equation} so that:
583: \begin{equation}\label{relations}
584: \partial_\kappa f_\kappa (W)=-f'_\kappa (W)\partial_\kappa W(\rho) ,\qquad
585: f'_\kappa (W)=1/W'(\rho),
586: \end{equation}
587: where $f'$ and $W'$ are the derivatives of $f$ and $W$ with
588: respect to their respective arguments. Making the change of
589: variables in eq.~(\ref{closedeqnsON}), and using the relations
590: (\ref{relations}), one gets:
591: \begin{eqnarray}
592: \partial_\kappa f_\kappa(W)&=&\frac{N}{2}\int \frac{d^dq}{(2\pi)^d}\frac{\partial_\kappa R_\kappa (q^2)}{(q^2+W+R_\kappa (q^2))^2} \nonumber \\
593: \partial_\kappa \left(\hat{\Gamma}_\kappa^{(2)}(p,-p;\rho)\right)^{-1}
594: &=&-N\int \frac{d^dq}{(2\pi)^d}\frac{\partial_\kappa R_\kappa (q^2)}{(q^2+W+R_\kappa (q^2))^2}
595: \frac{1}{(q+p)^2+W+R_\kappa ((q+p)^2)} \nonumber \\
596: \end{eqnarray}
597: At this point we observe that the right hand sides of both the
598: expressions above are total derivatives:
599: \begin{eqnarray}
600: &&\int \frac{d^dq}{(2\pi)^d}\frac{\partial_\kappa R_\kappa (q^2)}{(q^2+W+R_\kappa (q^2))^2}
601: =- \partial_\kappa \int \frac{d^dq}{(2\pi)^d}\frac{1}{q^2+W+R_\kappa (q^2)}\nonumber \\
602: &&\int \frac{d^dq}{(2\pi)^d}\frac{\partial_\kappa R_\kappa (q^2)}{(q^2+W+R_\kappa (q^2))^2}\,\frac{1}{(q+p)^2+W+R_\kappa ((q+p)^2)} \nonumber \\
603: &&=-\frac{1}{2}\partial_\kappa \int \frac{d^dq}{(2\pi)^d}\frac{1}{q^2+W+R_\kappa (q^2)}\,\frac{1}{(q+p)^2+W+R_\kappa ((q+p)^2)}.\nonumber \\
604: \end{eqnarray}
605: This allows us to integrate analytically the equations. For the potential, one obtains the well known result \cite{Tetradis95,Berges00}.
606: By imposing the initial condition that for $\kappa =\Lambda$ the 4-point function is the bare vertex, we get from eq.~(\ref{Gamma4})
607: \begin{equation}
608: \hat{\Gamma}_{\kappa=\Lambda}^{(2)}(p,-p;\rho)=\frac{u}{3},
609: \end{equation}
610: corresponding to a $\phi^4$ interaction in the classical action, of the form $(u/4!)( \phi_a\phi_a)^2$. We then obtain the
611: longitudinal part of the 2-point function (in the limit $\Lambda\to\infty$):
612: \begin{equation}
613: \hat{\Gamma}_\kappa^{(2)}(p,-p;\rho)=\frac{u}{3}
614: \frac{1}{1+\frac{Nu}{6}\int \frac{d^dq}{(2\pi)^d}\frac{1}{q^2+W+R_\kappa (q^2)}\frac{1}{(q+p)^2+W+R_\kappa ((q+p)^2)}}.
615: \end{equation}
616: This is the usual sum of bubbles in large $N$, which can be found using other methods \cite{Moshe:2003xn}. Notice however that our solution follows directly from the NPRG without introducing any auxiliary field. The dependence on $W$ reflects the presence of the background field: $W$ plays here the role of a (field dependent) ``mass term''.
617:
618: The same method can be applied to the other $n$-point functions.
619: The calculations are straightforward, but the explicit results
620: involve lengthy expressions. Let us just give here the result for
621: the 3-point function: \beq
622: \hat\Gamma^{(3)}_\kappa(p_1,p_2,p_3;W) &=&N\hat\Gamma^{(2)}_\kappa(p_1,-p_1;W) \hat\Gamma^{(2)}_\kappa(p_2,-p_2;W) \hat\Gamma^{(2)}_\kappa(p_3,-p_3;W)
623: \nonumber \\
624: &&\times\int \frac{d^dq}{(2\pi)^d} G_T(q^2;\kappa;\rho)
625: G_T((q+p_1)^2;\kappa;\rho)G_T((q-p_3)^2;\kappa;\rho) .\nonumber \\
626: \eeq
627:
628: \section{Conclusions and perspectives}
629:
630: In this paper we have presented an approximation scheme to solve
631: the NPRG equations and obtain the
632: $n$-point functions for any external momenta.
633: Our proposal is as general and systematic as the
634: derivative expansion: results can be systematically improved as
635: one can write, at any order, a closed set of equations.
636: We have
637: provided here, as an illustration, an application to the
638: $O(N)$ model in the
639: large $N$ limit: there, the method turns out to be exact at leading order,
640: and it provides an economical procedure to obtain the analytical expressions
641: of the $n$-point functions at arbitrary momenta. Clearly, however, the method is more general. It is based on
642: an approximation that can be exported to other field theories.
643: Work on such extensions, in particular to gauge theories, is
644: underway. This, as well as numerical studies for physically
645: interesting problems, will be reported in forthcoming
646: publications.
647:
648:
649: \bibliographystyle{unsrt}
650: \begin{thebibliography}{10}
651:
652: %\cite{Tetradis:1993ts}
653: \bibitem{Tetradis94}
654: N.~Tetradis and C.~Wetterich,
655: %``Critical exponents from effective average action,''
656: Nucl.\ Phys.\ B {\bf 422} (1994) 541.
657: %[arXiv:hep-ph/9308214].
658: %%CITATION = HEP-PH 9308214;%%
659:
660:
661: \bibitem{Ellwanger94a}
662: Ulrich Ellwanger,
663: %\newblock Flow equations for {N} point functions and bound states.
664: Z. Phys. C62 (1994) 503--510.
665:
666: \bibitem{Morris94b}
667: T.~R. Morris,
668: %\newblock The {E}xact renormalization group and approximate solutions.
669: Int. J. Mod. Phys. A9 (1994) 2411--2450.
670:
671: \bibitem{Morris94c}
672: T.~R. Morris,
673: %\newblock Derivative expansion of the exact renormalization group.
674: Phys. Lett. B329 (1994) 241--248.
675:
676:
677: \bibitem{Berges00}
678: J. Berges, N. Tetradis and C. Wetterich,
679: Phys. Rept. 363 (2002) 223--386.
680:
681: \bibitem{Bagnuls:2000ae}
682: C~Bagnuls and C~Bervillier,
683: Phys. Rept. 348(2001) 91.
684:
685: %\cite{Delamotte:2004zg}
686: \bibitem{Delamotte:2004zg}
687: B.~Delamotte and L.~Canet,
688: \emph{What can be learnt from the nonperturbative renormalization group?},
689: arXiv:cond-mat/0412205.
690: %%CITATION = COND-MAT 0412205;%%
691:
692: \bibitem{weinberg73}
693: S.~Weinberg, Phys. Rev. {\bf D8} (1973) 3497.
694:
695: \bibitem{truncation} U. Ellwanger, Z. Phys., {\bf C62} (1994), 503;
696: U. Ellwanger and C. Wetterich, Nucl. Phys. {\bf B423} (1994), 137;
697: U.~Ellwanger, M.~Hirsch and A.~Weber,
698: %``The heavy quark potential from Wilson's exact renormalization group,''
699: Eur.\ Phys.\ J.\ C {\bf 1} (1998) 563;
700: J.~M.~Pawlowski, D.~F.~Litim, S.~Nedelko and L.~von Smekal,
701: %``Infrared behaviour and fixed points in Landau gauge QCD,''
702: Phys.\ Rev.\ Lett.\ {\bf 93} (2004) 152002;
703: J.~Kato,
704: %``Infrared non-perturbative propagators of gluon and ghost via exact
705: %renormalization group,''
706: arXiv:hep-th/0401068; C.~S.~Fischer and H.~Gies,
707: %``Renormalization flow of Yang-Mills propagators,''
708: JHEP {\bf 0410} (2004) 048.
709:
710: \bibitem{alkofer} For a review, see
711: R. Alkofer and L. von Smekal, Phys. Rep. {\bf 353} (2001), 281.
712:
713: \bibitem{convergence} T.~R.~Morris,
714: %``On truncations of the exact renormalization group,''
715: Phys.\ Lett.\ B {\bf 334} (1994) 355.
716: %[arXiv:hep-th/9405190].
717: %%CITATION = HEP-TH 9405190;%%
718:
719:
720: \bibitem{Blaizot:2004qa}
721: J.~P.~Blaizot, R.~M.~Galain and N.~Wschebor,
722: \emph{Non Perturbative Renormalization Group, momentum dependence of $n$-point
723: functions and the transition temperature of the weakly interacting Bose
724: gas}, arXiv:cond-mat/0412481.
725: %%CITATION = COND-MAT 0412481;%%
726:
727: \bibitem{BMW05_3}
728: J.~P.~Blaizot, R.~M.~Galain and N.~Wschebor,
729: \emph{A study of the self-energy of the scalar field within the non perturbative renormalization group}, in preparation .
730:
731:
732: \bibitem{Ball95}
733: R.D. Ball, P.E. Haagensen, J.I. Latorre, and E.~Moreno, Phys.
734: Lett. B347 (1995) 80.
735:
736: \bibitem{Comellas98}
737: J.~Comellas,
738: Nucl. Phys. B509 (1998) 662.
739:
740: \bibitem{Litim}
741: D.Litim, Phys. Lett. {\bf B486}, 92 (2000); Phys. Rev. {\bf D64},
742: 105007 (2001); Nucl. Phys. {\bf B631}, 128 (2002);
743: Int.J.Mod.Phys. {\bf A16}, 2081 (2001).
744:
745: \bibitem{Canet02}
746: L. Canet, B. Delamotte, D. Mouhanna and J. Vidal.
747: %\newblock Optimization of the derivative expansion in the nonperturbative
748: % renormalization group.
749: Phys. Rev. D67 (2003) 065004.
750:
751:
752:
753:
754: %\cite{Zinn-Justin:2002ru}
755: \bibitem{Zinn-Justin:2002ru}
756: J.~Zinn-Justin,
757: \emph{Quantum field theory and critical phenomena},
758: Int.\ Ser.\ Monogr.\ Phys.\ {\bf 113} (2002) 1.
759: %%CITATION = IMPHA,113,1;%%
760:
761:
762: \bibitem{Berges96}
763: J.~Berges, N.~Tetradis, and C.~Wetterich,
764: %\newblock Critical equation of state from the average action.
765: Phys. Rev. Lett. 77 (1996) 873--876.
766:
767: \bibitem{Litim01}
768: Daniel~F. Litim,
769: %\newblock Optimised renormalization group flows.
770: Phys. Rev. D64 (2001)105007.
771:
772: \bibitem{Canet03}
773: L. Canet, B. Delamotte, D. Mouhanna and J. Vidal,
774: %\newblock Nonperturbative renormalization group approach to the ising model: a
775: % derivative expansion at order $\partial^4$.
776: Phys. Rev. B68 (2003) 064421.
777:
778: %\cite{Golner:1998sr}
779: \bibitem{Golner:1998sr}
780: G.~R.~Golner,
781: {\it Exact renormalization group flow equations for free energies and N-point
782: functions in uniform external fields},
783: arXiv:hep-th/9801124.
784: %%CITATION = HEP-TH 9801124;%%
785:
786: %\cite{Moshe:2003xn}
787: \bibitem{Moshe:2003xn}
788: M.~Moshe and J.~Zinn-Justin,
789: %``Quantum field theory in the large N limit: A review,''
790: Phys.\ Rept.\ {\bf 385}, 69 (2003)
791: % [arXiv:hep-th/0306133].
792: %%CITATION = HEP-TH 0306133;%%
793:
794:
795: \bibitem{D'Attanasio97}
796: M. D'Attanasio and T.~R. Morris,
797: %\newblock Large {N} and the renormalization group.
798: Phys. Lett. B409 (1997) 363--370.
799:
800: \bibitem{Blaizot05}
801: J.-P. Blaizot, R.~Mendez Galain, and N. Wschebor, {\it Non
802: perturbative renormalization group and momentum dependence of
803: $n$-point functions}, in preparation.
804:
805:
806: \bibitem{Tetradis95}
807: N.~Tetradis and D.~F. Litim,
808: %\newblock Analytical solutions of exact renormalization group equations.
809: Nucl. Phys. B464 (1996) 492--511.
810:
811: \end{thebibliography}
812:
813: \end{document}
814:
815:
816:
817: %%%%%%%%%%%%%%%%%%%%%%
818: