1: \documentclass[10pt]{article}
2: \textwidth 450pt \textheight 600pt \oddsidemargin 14pt
3: \evensidemargin 14pt\topmargin -10pt \baselineskip 40pt
4: \parindent 20pt
5: \renewcommand{\thefootnote}{\fnsymbol{footnote}}
6: \usepackage{amssymb}
7: \usepackage{amsbsy}
8: \usepackage{amsmath}
9: \usepackage{amsfonts}
10: \usepackage{verbatim}
11: \usepackage[dvips]{graphicx}
12: \usepackage{mathenv}
13: \begin{document}
14: \begin{flushright}
15: BI-TP 2005/23\\
16: LPT-Orsay 05-53
17: \end{flushright}
18: \vspace{1.0cm} {\begin{center}
19:
20: \huge {\bf Has saturation physics been observed in deuteron-gold
21: collisions at RHIC ?}\end{center}}
22:
23: \vspace{1.0cm}
24:
25: \begin{center}
26: {Rudolf Baier $^{a}$, Yacine Mehtar-Tani $^{b,}$\footnote{E-mail
27: address: mehtar@th.u-psud.fr}, Dominique Schiff $^{b}$}
28:
29:
30: \vspace{0.5cm} $^a${\it Fakult\"at f\"ur Physik, Universit\"at
31: Bielefeld,\\D-33501,
32: Bielefeld, Germany } \\
33: $^b$ {\it Laboratoire de Physique Th\'eorique, Universit\'e de
34: Paris XI, \\B\^atiment 210,
35: 91405 Orsay Cedex, France} \\
36: \end{center}\vspace{0.5cm}
37:
38: \begin{abstract}
39: In the framework of the recently proposed saturation picture, we examine in a
40: systematic way whether the nuclear modification factor measured for d-Au
41: collisions at RHIC may be simply explained. The Cronin peak which is obtained
42: at mid-rapidity around $k_{\bot}\simeq 3$ GeV may be reproduced at the proper
43: height only by boosting the saturation momentum by an additional nuclear
44: component as already shown in the literature. In this respect, mid-rapidity
45: RHIC data cannot necessarily be seen as a probe of the saturation picture. The large rapidity ($\eta\simeq
46: 3$) region allows us to test the shape of the unintegrated gluon distribution
47: in the nucleus, investigating various parameterizations inspired by large
48: rapidity solutions (of the BFKL and) of the Balitsky-Kovchegov (BK) equation. A satisfactory
49: description of $R_{CP}$ at RHIC is
50: obtained in the BK picture.
51: \end{abstract}
52: \section{Introduction}
53:
54:
55:
56: Testing the saturation (Color Glass Condensate) picture
57: \cite{JKrev}, for the
58: initial state of deuteron-gold (dA) collisions against RHIC data has been a
59: subject of interest for some time. Two salient features have been observed
60: \cite{BRAHMS1,exp} concerning the behavior of the nuclear modification factor $R_{dA}$.
61: At mid-rapidity, the Cronin peak height depends on the centrality of the
62: collision. At large rapidity, the suppression predicted by quantum evolution is
63: observed and is bigger for smaller centralities. A number of papers have recently discussed the description of the Cronin enhancement \cite{BKW,AG,KKT1,KKT2,JMG,IIT} and the effect of small $x$ \cite{KKT1,KT,AAKSW,JJM,KLM,BGV1}. In this work, we examine in a
64: systematic way, how and if saturation and quantum evolution provide a
65: reasonable quantitative agreement with data \cite{BRAHMS1}.\\
66: The time is indeed appropriate to assess the predictability of
67: the saturation (CGC) picture. This endeavor has a mitigated
68: conclusion: as it turns, unavoidably, the saturation scale
69: introduced in the theory, does not have the proper size to
70: explain RHIC data at mid-rapidity. This conclusion is similar to the one stated in \cite{AG}. On the other hand, quantum
71: evolution as described by the theory gives the proper
72: suppression of the nuclear modification factor above the
73: saturation scale.
74:
75: In section 2.1, we calculate the hadron production cross section in dA at
76: mid-rapidity using the semi-classical approach and show the prediction
77: for the minimum-bias nuclear modification factor
78: $R_{dA}$ (and $R_{CP}$ for central versus peripheral collisions dA collisions)
79: in relation with other previous studies and comparing with data.\\
80: We then discuss, in section 2.2, quantum evolution. We first derive the expression for the
81: cross-section at leading log accuracy, including both gluon and quark
82: distributions within the deuteron. This expression is identical to eq. (22) in
83: \cite{dum}. We then present various parameterizations of the unintegrated gluon
84: distribution in the nucleus, inspired by large rapidity solutions of the BK
85: equation \cite{BK}
86: and show the comparison with representative data \cite{BRAHMS1}.
87: In section 3, the conclusion and outlook are given.
88:
89: \section{Hadron production in dA}\label{sec1}
90:
91: The nuclear modification factor $R_{dA}$ and the
92: $R_{CP}$(Central/Peripheral collisions)
93: ratio are defined as
94: \begin{equation}\label{RpA}
95: R_{dA}=\frac{1}{N_{coll}}\frac{\frac{dN^{dA\rightarrow hX}}{d\eta
96: d^{2}{\bf k} }}{\frac{dN^{pp\rightarrow hX}}{d\eta d^{2}{\bf k}
97: }},
98: \end{equation}
99: \begin{equation}
100: R_{CP}=\frac{N^{P}_{coll}\frac{dN^{dA\rightarrow hX}}{d\eta
101: d^{2}{\bf k} }\vert_{C}}{N^{C}_{coll}\frac{dN^{dA\rightarrow
102: hX}}{d\eta d^{2}{\bf k}}\vert_{P}}.
103: \end{equation}
104: ${\bf k}$ and $\eta$ are respectively the transverse momentum and
105: the pseudo-rapidity of the observed hadron. $N_{coll}$ is the
106: number of collisions in dA, it is roughly twice the number of
107: collisions in pA(proton-Gold). The centrality dependence of
108: $R_{dA}$ is related to the dependence of $N^{dA\rightarrow
109: hX}=d\sigma^{dA\rightarrow hX}/d^{2}{\bf b}$ and $N_{coll}({\bf
110: b})$ on the impact parameter of the collision. In this paper, we
111: address the predictions of the Color Glass Condensate for these
112: ratios. We always assume that cross-sections depend on the impact
113: parameter only through the number of participants which is
114: proportional to the saturation scale
115: \begin{equation}
116: Q_ {sA}^2({\bf b})\simeq Q_ {sA}^2(0)N _{part.Au}({\bf b})/N
117: _{part.Au}(0),
118: \end{equation}
119: where $N_{part.Au}$ is the number of participants in the gold
120: nucleus in d-Au collisions. This is coherent with the assumption
121: that $Q_ {sA}^2({\bf b}) \simeq (N _{part.Au}({\bf b})/2)Q_
122: {sp}^2$ such that $ Q_ {sA}^2({\bf b}=0)\simeq A^{1/3} Q_ {sp}^2$
123: \cite{KLN}.
124: We use Table 2 in \cite{BRAHMS1} which gives the number
125: of participants $N_{part}$ and the number of collisions $N_{coll}$
126: for several centralities.
127: \subsection{Semi-classical approach}
128: \label{sec11}
129:
130: We first deal with gluon production at
131: mid-rapidity for which different approaches have been proposed.
132: The inclusive cross section has been calculated in \cite{KovMul}
133: in a quasi classical approach of multiple rescattering inside the
134: nucleus, see also \cite{KKT1,KT}. Further confirmation has been
135: given in \cite{BGV1,DuMcl}. The
136: inclusive cross section, for a gluon with transverse momentum ${\bf k}$
137: and
138: rapidity $\eta=0$, is written as
139:
140: \begin{equation}\label{KKT}
141: \frac{d\sigma^{dA\rightarrow gX}}{d\eta d^{2}{\bf k} d^{2}{\bf
142: b}}=\frac{C_{F}}{\alpha_{s}\pi(2\pi)^{3}}\frac{1}{{\bf k}^{2}}\int
143: d^{2}{\bf B}\int d^{2}{\bf z} \nabla^{2}_{{\bf z}}n_{G}({\bf
144: z},{\bf b}-{\bf B})\nabla^{2}_{{\bf z}}N_{G}({\bf z},{\bf
145: b})e^{-i{\bf k}.{\bf z}},
146: \end{equation}
147: where $N_{G}({\bf z},{\bf b})$ is the forward scattering amplitude
148: of a gluon dipole off the nucleus, it contains all higher twists
149: in the semi classical approximation and $n_{G}({\bf z},{\bf
150: b}-{\bf B})$ is the forward scattering amplitude of a gluon dipole
151: off the deuteron at leading twist approximation. ${\bf B}$ and
152: ${\bf b}$ are the impact parameters of the deuteron and the gluon
153: with respect to the center of the nucleus. At mid-rapidity and
154: RHIC energies, it is legitimate to neglect quantum evolution.
155:
156: This approach provides a description of $N_{G}$ at all orders in
157: terms of the saturation scale $Q^{2}_{sA}=Q^{2}_{s}\varpropto
158: A^{1/3}\Lambda^{2}_{QCD}$. The CGC approach yields a
159: Glauber-Mueller form for the dipole forward scattering amplitude
160: \begin{equation}\label{NG}
161: N_{G}({\bf z},{\bf b})=1-{\bf exp}({-\frac{1}{8}{\bf
162: z}^2Q^{2}_{s}({\bf b})\ln\frac{1}{{\bf z}^2\Lambda^2}}),
163: \end{equation}
164: where the saturation scale is given by \cite{KovMul}
165: \begin{equation}\label{QS}
166: Q^{2}_{s}({\bf b})\ln\frac{1}{{\bf
167: z}^2\Lambda^2}=\frac{4\pi^{2}\alpha_{s}}{C_{F}}\rho T({\bf
168: b})xG(x,1/{\bf z}^{2}).
169: \end{equation}
170: $\rho$ is the nuclear density in the nucleus and $T({\bf
171: b})=2\sqrt{ R^2-{\bf b}^2}$ the nuclear profile function of a
172: spherical nucleus of radius $R$. $\Lambda$ is an infrared cut-off
173: of order $\Lambda_{QCD}$. The cross section can be rewritten in
174: the following simpler form
175: \begin{equation}\label{cross}
176: \frac{d\sigma^{dA\rightarrow gX}}{d\eta d^{2}{\bf k} d^{2}{\bf
177: b}}=\frac{C_{F}\alpha_{s}}{\pi^{2}}\frac{2}{{\bf
178: k}^{2}}\int_{0}^{~1/\Lambda}du\ln\frac{1}{u\Lambda}\partial_{u}[u\partial_{u}N_{G}(u,{\bf
179: b })]J_{0}(\vert {\bf k}\vert u),
180: \end{equation}
181: where $u=\vert {\bf z}\vert$.\\
182: \subsubsection{A model for minimum bias collisions}
183:
184: One defines the minimum bias cross-section as the average over the impact parameter of the collision, it may be written as
185: \begin{equation}
186: \frac{dN^{min.bias}}{d\eta d^{2}{\bf k}}=\langle \frac{d\sigma}{d\eta d^{2}{\bf
187: k}d^{2}{\bf b}}\rangle=\frac{1}{S_{A}} \int d^{2}{\bf b}\frac{d\sigma}{d\eta d^{2}{\bf
188: k}d^{2}{\bf b}}.
189: \end{equation}
190: $S_A=\pi R^2$ is the transverse area of the nucleus. All the centrality dependence in the cross-section is contained in $N_G({\bf z},{\bf b})$ as given by (\ref{NG}) and (\ref{QS}).
191: One can perform the ${\bf b}$ integral for $N_G$ yielding
192: \begin{equation}
193: \langle N_{G}({\bf z},Q_{s})\rangle=1+\frac{128}{{\bf z}^4 Q_{sC}^4 \ln^2
194: (1/{\bf z^2}\Lambda^2)}[(1+\frac{1}{8}{\bf z}^2Q_{sC}^2 \ln
195: (1/{\bf z}^2\Lambda^2))\exp\{-\frac{1}{8}{\bf z}^2 Q_{sC}^2 \ln
196: (1/{\bf z}^2\Lambda^2)\}-1],
197: \end{equation}
198: where $Q_{sC}^2=Q^2_s ({\bf b=0})$. The corresponding result for $R_{dA}$ is shown in Fig.
199: \ref{fig1} (a) taking $Q_{sC}^2=2$ GeV$^2$ and $\Lambda=0.2$ GeV. The proton-proton cross-section is calculated by using the Glauber-Mueller formula (\ref{NG}) with $Q_{sp}$. $R_{dA}$ shows a Cronin peak for $k_{\bot}$
200: in the range of $Q_{s}$. We have used a prescription such that
201: the region $z \sim 1/\Lambda$ does not affect the integral. For
202: that purpose, we make the replacement $\ln\frac{1}{{\bf
203: z}^2\Lambda^2} \rightarrow \ln(\frac{1}{{\bf
204: z}^2\Lambda^2}+a^{2})$ choosing $a=3$ \cite{BKW}.
205: A good approximation of the integral
206: over ${\bf b}$ is to choose for $Q_{s}$ the
207: average value $\langle Q^{2}_{s}({\bf
208: b})\rangle\equiv Q^{2}_{s.min-bias}=(2/3)Q_{sC}^2$, in which case
209: \begin{equation}
210: \label{Rb}
211: \langle \frac{d\sigma}{d\eta d^{2}{\bf
212: k}d^{2}{\bf b}}(Q^{2}_{s}({\bf b}))\rangle\simeq \frac{d\sigma}{d\eta d^{2}{\bf
213: k}d^{2}{\bf b}}(\langle Q^{2}_{s}({\bf b})\rangle).
214: \end{equation}
215: Actually, (\ref{Rb}) turns into an equality in the high $k_{\bot}$ region
216: and is quite good in the region of the Cronin peak. In Fig. \ref{fig1}(b) we see that the error is
217: maximum when ${\bf k}\lesssim Q_{s}$, reaching $10 \%
218: $. For a
219: cylindrical nucleus there is no ${\bf b}$ dependence in the cross
220: section and (\ref{Rb}) turns into an equality.
221: This tells us that the physics is the same whatever the
222: geometry of the nucleus \cite{KKT1}.
223: \subsubsection{Effects of fragmentation on the Cronin peak}
224:
225: To get the hadron cross-section we still have to convolute
226: (\ref{cross}) with the proper fragmentation functions
227: \begin{equation}\label{crossFF}
228: \frac{d\sigma^{dA\rightarrow hX}}{d\eta d^{2}{\bf k} d^{2}{\bf b}}
229: = \int \frac{dz}{z^{2}}D_{g}^{h}(z,Q_f
230: ^2)\frac{d\sigma^{dA\rightarrow gX}}{d\eta d^{2}{\bf q} d^{2}{\bf
231: b}}({\bf q}={\bf k}/z),
232: \end{equation}
233:
234: \begin{figure}[hbtp]
235: \begin{tabular}{c c }
236:
237: \qquad\includegraphics[width=6cm]{Fig1a.eps} &\qquad \includegraphics[width=6cm]{Fig1b.eps} \\
238: \qquad (a)& \qquad(b)
239: \end{tabular} \caption{(a) :
240: The nuclear modification factor $R_{dA}$ without fragmentation
241: functions (gluon production) for $Q^2_{sC}=2$ GeV$^2$. (b) :
242: Comparison between the average full min-bias calculation and the
243: calculation with average min-bias $Q_s$ for $Q^2_{sC}=2, 5$ and
244: $9$ GeV$^2$ (Full, dashed and dotted).} \label{fig1}
245: \end{figure}
246: where $Q_f$ is a large scale of order $k_{\bot}$ taken between 2
247: and 8 GeV. We use the fragmentation functions of \cite{kkp}. In
248: Fig.\ref{fig2}(a) we present the result for $R_{dA}$.
249:
250: \begin{figure}[hbtp]
251: \begin{tabular}{c c }
252:
253: \qquad\includegraphics[width=6cm]{Fig2a.eps} &\qquad \includegraphics[width=6cm]{Fig2b.eps} \\
254: \qquad (a)& \qquad(b)
255: \end{tabular}\caption{(a) :
256: Minimum bias $R_{dA}$ for charged hadron production for
257: $Q^2_{s.min-bias}=1.3, 3.33, 6$ and $9$ GeV$^2$ (lowest to highest
258: curve). (b) : Minimum bias $R_{dA}$ for neutral pions production
259: for $Q^2_{s.min-bias}=9$ GeV$^2$. The points are representative
260: data taken from \cite{BRAHMS1}.} \label{fig2}
261: \end{figure}
262:
263: \begin{figure}[h]
264: \centering
265: \includegraphics[width=8cm]{Fig3.eps}
266:
267: \caption{$R_{CP}$ for $Q^2_{sC}=9$ GeV$^2$ (thick lines) and
268: $Q^2_{sC}=2$ GeV$^2$ (thin lines). Full lines correspond to
269: central over peripheral collisions (full experimental dots).
270: Dashed lines correspond to semi-central over peripheral collisions
271: (empty experimental dots). Data from \cite{BRAHMS1}.}\label{fig3}
272: \end{figure}
273:
274: The fragmentation functions induce a flattening of the Cronin
275: peak toward 1, with a shift toward the small $k_{\bot}$'s. This
276: feature can be understood by comparing $R^g_{dA}$ without F.F's
277: (fragmentation functions) and $R^h_{dA}$ with F.F's, order by
278: order in powers of $Q_{s}^2/{\bf k}^{2}$ (or the number of
279: participants). Above the saturation scale, the following expansion
280: has a meaning
281: \begin{equation}
282: R^g_{dA}=1+R_{dA}^{(1)g}+...
283: \end{equation}
284: In the Leading Log approximation with respect to
285: $\ln(k_{\bot}/\Lambda)$, using (\ref{crossFF}), we approximate the
286: nuclear modification factor by
287: \begin{equation}
288: R_{dA}^h=1+\frac{\langle z^4 \rangle}{\langle z^2
289: \rangle}R_{dA}^{(1)g}+...
290: \end{equation}
291: where we have defined for $n\geqslant 2$
292: \begin{equation}
293: \langle z^n \rangle=\int_{z_0}^1 dz D(z,Q_f^2) z^n/\int_{z_0}^1 dz
294: D(z,Q_f^2).
295: \end{equation}
296: Since $z<1$ we always have $\langle z^n \rangle/\langle z^2
297: \rangle <1$. The rescattering terms are then less important and
298: $R_{dA}^h$ gets closer to 1. This is shown in Fig. \ref{fig2} (a) :
299: comparing with Fig. \ref{fig1} (a) we see the dramatic effect of
300: the fragmentation functions for $Q_{s.min-bias}^{2}=1.3$ GeV$^2$.
301: To get agreement with RHIC data we have to increase the value of
302: the saturation scale $Q_s$ and take $Q_{s.min-bias}^{2}\gtrsim 6$
303: GeV$^2$. This feature has already been mentioned in \cite{KKT2}.
304: In fact, $Q_s^2$ as defined in (\ref{QS}) is at most of the order
305: of $2$ GeV$^2$ for ${\bf b}=0$. The authors of \cite{KKT2} propose
306: to enhance it by adding an additional momentum due to
307: "non-perturbative" nuclear effects :
308: \begin{equation}\label{ansatz}
309: Q_s^2\rightarrow Q_s^2+\kappa^2 A^{1/3},
310: \end{equation}
311: with $\kappa^2=1$ GeV$^2$. This amounts to boosting $Q_s^2({\bf
312: b}=0)$ to $\sim9$ GeV$^2$. In fact, as shown in Fig. \ref{fig2} (a), we need even a larger saturation scale. On the other hand,
313: the way $R_{dA}$ is normalized (by estimating the proton-proton
314: cross-section using (\ref{NG})) is not fully convincing. The ratio
315: $R_{CP}$ does not suffer from the same uncertainty. It is shown in
316: Fig. \ref{fig3} for $Q_{sC}^2=9$ GeV$^2$ and $Q_{sC}^2=2$ GeV$^2$
317: for central and semi-central collisions. Taking into account the
318: experimental error-bars, the large value for $Q_{sC}$ is
319: definitely preferred and in agreement
320: with data.\\
321:
322: We should at this point remark that saturation physics in the
323: present stage is in the situation where the leading order
324: perturbative QCD description was, concerning large $k_{\bot}$
325: spectra in hadron-hadron collisions. The agreement with data could
326: only be obtained by implementing an intrinsic non-perturbative
327: transverse momentum for partons inside the hadron. The present
328: state seems to be that, at next-to-leading order, the perturbative
329: theory becomes predictive \cite{AF,Av} : a detailed NLO comparison with RHIC data for $pp\rightarrow\pi^0X$ is presented in \cite{Adler}. In this respect and in the
330: leading order perturbation QCD framework, the authors of ref.
331: \cite{AG} have used a traditional Glauber-Eikonal approach of
332: sequential multiple partonic scatterings with the implementation
333: of a large intrinsic $k_{\bot}$ in parton distribution functions
334: and they have obtained a good agreement with data for $\pi_0$
335: production in dA collisions.\\
336:
337: Staying in the saturation physics
338: framework, we may nonetheless try to go beyond the ad-hoc ansatz
339: (\ref{ansatz}) and modify the saturation picture, which is
340: exclusively based on hard multiple scatterings, by adding
341: non-perturbative scatterings, when a parton (gluon) is passing
342: through a nucleus. Following the Moli\`ere scattering theory
343: \cite{Bethe} extended to QCD \cite{BDMPS}, we define the
344: probability distribution for the scattered parton by
345: \begin{equation}\label{Nphi}
346: V({\bf k})=\frac{1}{\sigma}\frac{d\sigma}{d^2{\bf
347: k}}=\frac{1}{\pi}\{\frac{c}{\langle {\bf k}^2\rangle
348: }e^{-\frac{{\bf k}^2}{\langle {\bf k}^2\rangle
349: }}+\frac{(1-c)\Lambda^2}{({\bf k}^2+\Lambda^2)^2}\},
350: \end{equation}
351: with $\int d^2{\bf k} V({\bf k})=1$.\\
352:
353: A soft, non-perturbative, gaussian contribution is added to the
354: hard screened (by mass $\Lambda$) gluon exchange term. Solving the
355: kinetic master equation for the survival probability of the
356: propagating parton, an effective scale (up to logarithm) is
357: derived,
358: \begin{equation}
359: Q_s^2\rightarrow (c\frac{\langle {\bf
360: k}^2\rangle}{\Lambda^2}+(1-c))Q_s^2.
361: \end{equation}
362: Different from (\ref{ansatz}) the non-perturbative part is added
363: but
364: weighted by a factor $c<1$.\\
365:
366: In order to obtain an effective $Q_s^2\approx O(10\text{ GeV}^2)$
367: one has to add a significant non-perturbative part, e.g. for
368: $\langle {\bf k}^2\rangle\simeq 0.5$ GeV$^2$, $\Lambda=0.2$ GeV
369: and $c\simeq0.3$ indeed
370: \begin{equation}
371: Q_s^2\approx2\text{ GeV}^2\rightarrow Q_s^2\approx 9\text{ GeV}^2.
372: \end{equation}
373: The underlying picture of dominating soft parton interactions with
374: nucleons has strong implications, especially for the width of jet
375: broadening, i.e. the resulting transport coefficient for cold
376: matter $\hat{q}\simeq Q_s^2/(2R)$ becomes also rather large,
377: namely $\hat{q}\simeq 0.8$ GeV$^2$/fm (for gold)! But such a
378: strong jet broadening has not been observed so far \cite{arleo}.
379: \subsection{Forward rapidity and quantum evolution}
380: \label{sec12}
381: \subsubsection{Hadron production cross-section}
382:
383: As shown in \cite{KKT1,KT}, it is possible to rewrite eq. (\ref{KKT})
384: under a $k_{\bot}$-factorized form \cite{KovMul}, which is then generalized to
385: include the rapidity dependence. At leading twist, on defines the unintegrated gluon
386: distribution for the nucleus and the deuteron, respectively, as
387: \begin{equation}\label{Nphi}
388: \varphi_{A}(Y_{A},{\bf k},{\bf
389: b})=\frac{C_{F}}{\alpha_{s}(2\pi)^{3}}\int d^{2}{\bf
390: z}\nabla^{2}_{{\bf z}}N_{G}(Y_{A},{\bf z},{\bf b})e^{-i{\bf
391: k}.{\bf z}},
392: \end{equation}
393: and
394: \begin{equation}
395: \varphi_{p}(Y_{d},{\bf k},{\bf b}-{\bf
396: B})=\frac{C_{F}}{\alpha_{s}(2\pi)^{3}}\int d^{2}{\bf
397: z}\nabla^{2}_{{\bf z}}n_{G}(Y_{d},{\bf z},{\bf b}-{\bf
398: B})e^{-i{\bf k}.{\bf z}},
399: \end{equation}
400: with $Y_{A}=\ln(1/x_{A})=Y+\eta$ and $Y_{d}=\ln(1/x_{d})=Y-\eta$
401: the rapidities of the gluons merging respectively from the nucleus
402: and the deuteron and carrying the light cone momentum fractions
403: $x_{A}$ and $x_{d}$; $\eta$ is the rapidity of the produced gluon
404: measured in the forward deuteron direction and
405: $Y=\ln(\sqrt{s}/k_{\bot})$. One obtains the expression for the
406: cross-section as
407: \begin{equation}\label{kt}
408: \frac{d\sigma^{dA\rightarrow gX}}{d\eta d^{2}{\bf k} d^{2}{\bf
409: b}}=\frac{2\alpha_{s}}{ C_{F}}\frac{1}{{\bf k}^{2}}\int d^{2}{\bf
410: B}\int d^{2}{\bf q }\varphi_{d}({\bf q}-{\bf k},Y-\eta,{\bf
411: b-B})\varphi_{A}({\bf q},Y+\eta,{\bf b}).
412: \end{equation}
413: We will focus our analysis on the large forward rapidity region for the gluon where we can neglect the emission of
414: additional gluons in the wave function of the proton. Actually, in
415: this region the biggest value of $Y_d$ is reached at $\eta=0$
416: yielding $Y_{d}=Y$ which is not indeed large enough at the
417: energies of RHIC for the evolution to take place. However, the
418: opposite happens in the nucleus where $Y_{A}$ increases with
419: $\eta$ and we thus expect to probe the small $x$
420: regime of the nucleus wave function \footnote{A calculation of $R_{dA}$ has been recently performed \cite{Strik} in the framework of linearly factorized pQCD at NLO, taking into account nuclear leading twist shadowing for the partonic distributions in the nucleus. The observed suppression cannot be explained in this context.}. We have, taking $\varphi_{d}\simeq
421: 2\varphi_{p} $ at leading twist without quantum evolution
422: \cite{KovMul},
423: \begin{equation}
424: \int d^{2}{\bf b} \varphi_{d}({\bf k},Y-\eta\simeq 0,{\bf
425: b})=\frac{\alpha_{s}C_{F}}{\pi}\frac{2}{{\bf k}^{2}}.
426: \end{equation}
427: This allows us, assuming $\varphi_{A}$ smooth enough, to
428: approximate (\ref{kt}) at leading log accuracy, when $\vert {\bf
429: k}\vert\gg \Lambda$, by
430: \begin{equation}
431: \frac{d\sigma^{dA\rightarrow gX}}{d\eta d^{2}{\bf k} d^{2}{\bf
432: b}}=\frac{2\alpha_{s}}{ C_{F}}\frac{\varphi_{A}({\bf k
433: },Y+\eta,{\bf b })}{{\bf k }^{2}}\int d^{2}{\bf B }
434: \int_{\Lambda}^{\vert {\bf k}\vert} \varphi_{d}({\bf q
435: },Y-\eta\simeq0,{\bf B })d^{2}{\bf q }.
436: \end{equation}
437: Using the relation,
438:
439: \begin{equation}
440: xG(x_{d},{\bf k }^{2})=\int d^{2}{\bf B } \int_{\Lambda}^{\vert {\bf k}\vert }
441: \varphi_{d}({\bf q },Y-\eta \simeq 0,{\bf B })\frac{d^{2}{\bf q
442: }}{\pi},
443: \end{equation}
444: we end up with
445: \begin{equation}
446: \frac{d\sigma^{dA\rightarrow gX}}{d\eta d^{2}{\bf k} d^{2}{\bf
447: b}}=\frac{\alpha_{s}(2\pi)}{ C_{F}}\frac{\varphi_{A}({\bf
448: k},Y+\eta,{\bf b})}{{\bf k}^{2}}xG(x_{d},{\bf k}^{2}).
449: \end{equation}
450: Notice the collinear factorization of the gluon distribution in
451: the proton. For hadron production one has to add the valence quark
452: contribution which may be written \cite{JMG}
453: \begin{equation}
454: \frac{d\sigma^{dA\rightarrow qX}}{d\eta d^{2}{\bf k} d^{2}{\bf
455: b}}=\frac{\alpha_{s}(2\pi)}{ N_{c}}\frac{\varphi_{A}({\bf k}
456: ,Y+\eta,{\bf b})}{{\bf k}^{2}}xq_{V}(x_{d},{\bf k}^{2}).
457: \end{equation}
458: Convoluting with the fragmentation functions we get
459: \begin{equation}
460: \frac{d\sigma^{dA\rightarrow hX}}{d\eta d^{2}{\bf k} d^{2}{\bf
461: b}}=\frac{\alpha_{s}(2\pi)}{ C_{F}}\sum_{i=g,u,d}\int_{z_0}^1
462: dz\frac{\varphi_{A}({\bf k}/z,Y+\eta+\ln z,b)}{{\bf k}^{2}}[
463: f_{i}(x_{d}/z,{\bf k}^{2}/z^{2})D_{h/i}(z,{\bf k}^2)],
464: \end{equation}
465: where $ f_{u,d}(x,{\bf k}^{2})=(C_{F}/N_{c})xq_{u,d}(x,{\bf
466: k}^{2})$ and $f_{g}(x,{\bf k}^{2})=xG(x,{\bf k}^{2})$ are the
467: parton distributions inside the proton; $D_{h/i}(z,{\bf k})$ are
468: F.F's of the parton $i$ into hadron $h$, and
469: $z_0=(k_{\bot}/\sqrt{s})e^{\eta}$. Notice that this formula is
470: identical to eq. (22) of \cite{dum} in the leading twist
471: approximation. At forward rapidity, the longitudinal momentum
472: fraction $x_d$ carried by the parton inside the deuteron is of
473: order of one. This implies that the dynamics of the deuteron is
474: completely dominated by valence quarks. For numerical calculations
475: we use the GRV LO parton distribution functions inside the proton
476: from \cite{grv}, assuming that there is no significant difference
477: between the proton and the neutron for charged hadron production.
478: \subsubsection{The unintegrated gluon distribution}
479: \label{121}
480: In the physics of saturation the relevant observable is
481: the forward scattering amplitude of a quark-antiquark dipole off a
482: target (a nucleus in our case). It enters several processes at
483: high energy like DIS, photoproduction and hadron-hadron
484: scattering. Its quantum evolution has recently been the object of
485: many studies \cite{KKT1,KT,BK,munp,munp2,MuT,IIMc}. The Balitsky-Kovchegov (BK) equation \cite{BK} valid in the large
486: $N_{c}$ limit and in the mean field approximation provides a
487: tool to study the rapidity behavior of the gluon distribution in a large momentum range including the saturation region where the
488: effects of gluon recombination, taken into account in the
489: non-linear term, become important, instead of the BFKL equation
490: \cite{BFKL} which contains only gluon splitting (linear term)
491: yielding a power growth of the gluon distribution with respect to
492: the center of mass energy violating unitarity and the Froissart
493: bound. In this work, as a simple test of the theory, we focus on the region above the saturation scale. The BK equation reads in momentum space
494: \begin{equation}
495: \partial_{y}\tilde{N}({\bf k},y)=\frac{\alpha
496: N_{c}}{\pi}[\chi(-\frac{\partial}{\partial\ln{\bf
497: k}^2})\tilde{N}({\bf k},y)-\tilde{N}^{2}({\bf k},y)],
498: \end{equation}
499: where $\chi(\gamma)=2\psi(1)-\psi(\gamma)-\psi(1-\gamma)$ is the
500: BFKL kernel, and
501:
502: \begin{equation}\label{NN}
503: \tilde{N}({\bf k},{\bf b},y)=\int \frac{d^{2}{\bf z}}{{\bf
504: z}^{2}}N({\bf z},{\bf b},y)e^{i{\bf z}{\bf k}}.
505: \end{equation}
506: At large $y$, a solution has been found for the BK equation above and not too far from
507: the saturation scale in terms of travelling waves. It has a
508: geometric scaling behavior in the variable $L=\ln ({\bf
509: k}^{2}/Q_{s}^{2}({\bf b},y))$ when $y$ goes to infinity
510: \cite{munp,MuT}
511:
512: \begin{equation}\label{BKsol}
513: \tilde{N}(L,Y)=CL\exp[-\gamma_{s}L-\beta (y) L^{2}].
514: \end{equation}
515: C is an undetermined constant irrelevant in the present analysis
516: of $R_{dA}$ and $R_{CP}$, $\gamma_s\simeq 0.628$ is the anomalous
517: dimension of the BFKL dynamics in the geometric scaling region
518: \cite{MuT,IIMc}: $Q_s^2 \lesssim k_{\bot}^2 \ll
519: Q_s^2(Y)\exp{(1/\beta)} $ and
520: $\beta\equiv\beta(y)=(2\bar{\alpha}\chi''(\gamma_{s})y)^{-1}$. We
521: used the recent fit to the HERA data performed in \cite{IIM} where
522: $\beta=(2\lambda\kappa y)^{-1}$, $\lambda=0.25$, $\kappa=9.9$ and
523: $Q_{s}^2(y,0)=3/2Q^2_{s.min.bias}(y)=3/2A^{1/3}(x_0/x)^{\lambda}$
524: GeV$^2$ with $x_0=0,67.10^{-4}$ and $x=e^{-y}$ given here by
525: $x=k_{\bot}/\sqrt{s}e^{-\eta}$. There is no straightforward way to
526: link this asymptotic form to the one at mid-rapidity: this implies
527: that there is an overall constant coming from $Q_s$ which is
528: negligible at very large rapidity and far from the saturation
529: scale. We expect this constant to play a significant role at RHIC
530: energies. Then, one can fix $Q_{s}$ as shown above, and put all
531: the freedom into an additional constant ($L\rightarrow L+L_{0}$).
532: Thus,
533: \begin{equation}\label{Ntilde}
534: \tilde{N}(L,y)=C'(L+L_{0})\exp[-(\gamma_{s}+2\beta L_{0})L-\beta
535: L^{2}].
536: \end{equation}
537: Making the same approximations as in the BK equation, one can
538: write an expression for $N_{G}$ \cite{KKT1}
539: \begin{equation}
540: N_{G}({\bf z},y)=\frac{2C_{F}}{N_{c}}(2N({\bf z},y)-N^{2}({\bf
541: z},y)).
542: \end{equation}
543: In the leading twist approximation where one can neglect
544: non-linear terms, we find
545: \begin{equation}
546: N_{G}({\bf z},y)\simeq 2N({\bf z},y).
547: \end{equation}
548: Both distributions $\tilde{N}$ and $\varphi_{A}$ are linked to the
549: dipole scattering amplitude in (\ref{NN}) and (\ref{Nphi}). One
550: can thus eliminate the latter yielding
551: \begin{equation}
552: \varphi_{A}(L,y)=\frac{4N_{c}}{\alpha_{s}(2\pi)^{3}}\frac{d^2}{dL^2}\tilde{N}(L,y).
553: \end{equation}
554: At very large $y$ one can neglect the term which breaks scaling,
555: namely $\beta \ll 1$. So that the last expression reduces to the
556: simple form exhibiting an exact scaling behavior
557: \begin{equation}\label{ESphi}
558: \varphi_{A}(L,y)\varpropto(L+L_{0}-\frac{2}{\gamma_{s}})\exp[-\gamma_{s}L].
559: \end{equation}
560: We have fixed $L_{0}$ such that $\varphi_{A}$ has a maximum when
561: ${\bf k}^2=Q_{s}^{2}({\bf b},y)$ \cite{MuT,BMS} corresponding to
562: $L_{0}=3/\gamma_{s}$. This is the only free parameter of our
563: calculation, it exhibits the uncertainty in the value of $Q_s$. It
564: turns out that $R_{CP}$ is very sensitive to variations of $L_{0}$
565: at energies of RHIC. For the numerical study we choose three
566: different expressions for $\varphi_A$,
567: selecting various specific terms in (\ref{Ntilde}) :\\\\
568: i) The BFKL saturation-inspired form \cite{IIMc}, which violates
569: scaling, derived from
570:
571: \begin{equation}\label{BFKL}
572: \tilde{N}(L,y)_{BFKL}\varpropto\exp[-\gamma_{s}L-\beta L^{2}].
573: \end{equation}
574: ii) The BK exact-scaling form (\ref{ESphi}). \\
575: iii) Finally the full expression derived from (\ref{Ntilde}).\\
576: \\
577:
578: \begin{figure}[hbtp]
579: \centering
580: \begin{tabular}{c c c }
581:
582: \qquad\includegraphics[width=4cm]{Fig4a.eps} &\qquad \includegraphics[width=4cm]{Fig4b.eps} &\qquad \includegraphics[width=4cm]{Fig4c.eps}\\
583: \qquad (a)& \qquad(b)& \qquad(c)
584: \end{tabular} \caption{
585: $R_{CP}$ for the BK parametrization (thick lines) and the
586: BFKL + saturation form (thin lines) at different rapidities $\eta=1,
587: 2.2$ and $3.2$. Full lines correspond to central over peripheral
588: collisions (full experimental dots). Dashed lines correspond to
589: semi-central over peripheral collisions (empty experimental dots
590: ). Data from \cite{BRAHMS1}.} \label{fig4}
591: \end{figure}
592: \begin{figure}[hbtp]
593: \centering
594: {\includegraphics[width=4cm]{Fig5.eps}}
595:
596: \caption{$R_{CP}$ for the exact scaling BK form at $\eta=3.2$.
597: Full line : central/peripheral, dashed line:
598: semi-central/peripheral. } \label{fig5}
599: \end{figure}
600:
601: In Fig. \ref{fig4} we show the comparison between the BFKL +
602: saturation form (\ref{BFKL}) and the BK parametrization
603: (\ref{Ntilde}). We expect this comparison to be valid at large
604: enough $k_{\bot}$. The agreement with data is quite good for the
605: latter. With decreasing rapidity we would expect our formula to
606: break down, nevertheless the global features of the data are
607: reproduced even at $\eta=1$. The exact scaling form, shown in Fig.
608: \ref{fig5}, is a slowly varying function of $\eta$ and is too low
609: to describe the data. However, if the picture is right, for
610: increased rapidity ($\eta\simeq 5$ or $6$), data points should
611: match that shape. The fact that the saturation model has a
612: semi-quantitative agreement with data for $R_{CP}$ is essentially
613: due to the anomalous dimension since an approximate form is (when
614: $k_{\bot}\gtrsim Q_s$)
615: \begin{equation}
616: R_{CP}\simeq
617: \frac{N^{P}_{coll}}{N^{C}_{coll}}(\frac{N^C_{part}}{N^P_{part}})^{\gamma_{eff}}\simeq
618: (\frac{N^C_{part}}{N^P_{part}})^{\gamma_{eff}-1}.
619: \end{equation}
620: At forward rapidity
621: $\gamma_{eff}\simeq\gamma_s+\beta(\eta)\ln(k_{\bot}^2/Q_s^2)$ is a
622: decreasing function of $\eta$ and an increasing function of
623: $k_{\bot}$. This allows us to understand the qualitative behavior
624: shown by data and in particular the inversion of the centrality
625: dependence compared to mid-rapidity where $R_{CP}\gtrsim 1$
626: (Cronin enhancement) corresponding to $"\gamma_{eff}"\gtrsim 1$ .
627: At very large $\eta$ the anomalous dimension stabilizes at
628: $\gamma_{eff}=\gamma_s$, which could be tested at the LHC.
629:
630:
631:
632: \section{Conclusion and outlook}
633: To conclude we may reiterate that in our opinion, the saturation
634: picture as probed by mid-rapidity data for the nuclear
635: modification factor at RHIC energies cannot be seen as predictive.
636: Indeed, the CGC at mid-rapidity is based on leading order pQCD
637: calculations including multiple eikonal partonic rescatterings
638: inside the nucleus, giving rise to a saturation scale $Q_s^2$ at
639: most of order of 2 GeV$^2$ and it turns out that this is not
640: sufficient to describe RHIC data at mid-rapidity. It may be very
641: interesting to relate the increase of the saturation scale needed
642: to explain the dA data to the comparison of experimental
643: measurements with theory for observables such as jet broadening in
644: nucleus-nucleus (AA) collisions and also jet quenching, since the
645: saturation scale
646: determines their order of magnitude. \\
647:
648: At forward rapidity, a saturation inspired framework for quantum
649: evolution predicts the suppression as observed in data. This is a
650: leading twist effect driven by the anomalous dimension which is
651: the main source of the observed behavior. There is good evidence
652: that the answer to the question: "has saturation been observed at
653: RHIC ?" is positive and a definitive answer could be provided if
654: one measures, in a more precise way, the anomalous dimension as
655: predicted by the theory. To do so, one needs to go to larger
656: rapidities (energies) at LHC.
657: \\
658:
659: As one more step to improve further the status of the present
660: theoretical description let us remark that initial state
661: suppression effects may be also present in AA collisions at large
662: $\eta$ and large transverse momentum $k_{\bot}$. An interesting
663: quantity to be measured is the double ratio
664: $R(\eta_1,\eta_2,k_{\bot})$ with $\eta_1>\eta_2>0$, defined by
665: \begin{equation}
666: R(\eta_1,\eta_2,k_{\bot})=\frac{R_{CP}(\eta_1,k_{\bot})}{R_{CP}(\eta_2,k_{\bot})},
667: \end{equation}
668: as a function of $k_{\bot}$. For $\eta_1=2.2$ and $\eta_2=0$, this
669: ratio is measured for AA by the BRAHMS Collaboration \cite{Reta};
670: there is evidence of suppression becoming more important at
671: forward
672: rapidities, as observed in dA measurements.\\
673:
674: Because of the final state suppression in AA collisions one may
675: argue that this ratio is bounded by
676: \begin{equation}\label{eta12}
677: R(\eta_1,\eta_2,k_{\bot})<R(\eta_1,\eta_2,k_{\bot})_{initial}<1,
678: \end{equation}
679: where $R(\eta_1,\eta_2,k_{\bot})_{initial}$ for AA collisions may
680: be estimated for $R_{CP-initial}$ in a way done for dA collisions
681: (e.g. by assuming $k_{\bot}-$factorization). We expect it to be
682: quantitatively similar to the dA case (e.g. Figs. \ref{fig4}(b)
683: and (c)), which is well understood in the saturation picture for
684: $\eta_1>\eta_2>0$. The inequality (\ref{eta12}) is based on the
685: assumption that $R_{CP}$ in AA collisions may be factorized as
686: $R_{CP-initial}Q(k_{\bot},\eta)$: the quenching factor takes into
687: account the (radiative) final state interactions \cite{BaierQ}. It
688: depends, at fixed $k_{\bot}$ on the pathlength of the parton
689: propagating in a dense medium, such that $Q$ is decreasing with
690: increasing pathlength, i.e. it amounts to more quenching.
691: Assuming, on geometrical grounds, a longer path for $\eta_1$ than
692: for $\eta_2$, $\eta_1>\eta_2$, implies
693: $Q(k_{\bot},\eta_1)/Q(k_{\bot},\eta_2)<1$, and (\ref{eta12})
694: follows.\\\\\\
695: \begin{Large}
696: \textbf{Acknowledgements}
697: \end{Large}
698: \\\\
699: We acknowledge helpful comments by A. H. Mueller.
700:
701:
702:
703: \begin{thebibliography}{10}
704: \bibitem{JKrev} See the recent reviews "Saturation physics and Deuteron-Gold Collisions at RHIC", J.Jalilian-Marian and Y. V. Kovchegov, arXiv:hep-ph/0505052; J-P. Blaizot and F. Gelis, Nucl. Phys. A {\bf
705: 750}, 148 (2005); E. Iancu and R. Venugopalan, in "Quark Gluon
706: Plasma 3". (editors: R.C. Hwa and X. N. Wang. World Scientific.
707: Singapor), p.249-363, arXiv:hep-ph/0303204.
708:
709: \bibitem{BRAHMS1} I. Arsene {\it et al.} [BRAHMS Collaboration],
710: Phys. Rev. Lett. {\bf 93}, 242303 (2004); R. Debbe for the BRAHMS Collaboration,
711: arXiv:nucl-ex/0405018.
712: \bibitem{exp} J. Adams {\it et al.} [STAR Collaboration],
713: arXiv: nucl-ex/051009; B. B. Back {\it et al.} [PHOBOS
714: Collaboration], arXiv:nucl-ex/0410022; K. Adcox {\it et al.}
715: [PHENIX Collaboration], arXiv:nucl-ex/0410003; I. Arsene {\it et
716: al.} [BRAHMS Collaboration], arXiv:nucl-ex/0410020.
717:
718:
719: \bibitem{BKW} R. Baier, A. Kovner and U. A. Wiedemann, Phys. Rev. D {\bf 68}, 054009 (2003).
720: \bibitem{AG} A. Accardi and M. Gyulassy, Phys. Lett. B {\bf 586}, 244-253
721: (2004); A. Accardi, Acta. Phys. Hung. (to appear).
722: \bibitem{KKT1} D. Kharzeev, Y. V. Kovchegov and K. Tuchin,
723: Phys. Rev. D {\bf 68}, 094013 (2003).
724: \bibitem{KKT2} D. Kharzeev, Y. V. Kovchegov and K. Tuchin,
725: Phys. Lett. B {\bf 599}, 23-31 (2004).
726: \bibitem{JMG} F. Gelis and J. Jalilian-Marian, Phys. Rev. D {\bf 67}, 074019 (2003).
727: \bibitem{IIT} E. Iancu, K. Itakura and D. N. Triantafyllopoulos, Nucl. Phys. A {\bf 742}, 182-252,
728: (2004).
729: \bibitem{KT} Y. V. Kovchegov, K. Tuchin, Phys. Rev. D {\bf 65}, 074026 (2002).
730: \bibitem{AAKSW} J. L. Albacete, N. Armesto, A. Kovner, C. A. Salgado and U. A. Wiedemann, Phys. Rev. Lett. {\bf 92}, 082001 (2004).
731: \bibitem{JJM} J. Jalilian-Marian, Nucl. Phys. A {\bf 748}, 664-671 (2005).
732:
733: \bibitem{KLM} D. Kharzeev, E. Levin and L. D. McLerran, Phys.
734: Lett. B {\bf 561}, 93 (2003).
735: \bibitem{BGV1} J-P. Blaizot, F. Gelis and R. Venugopalan,
736: Nucl. Phys. A {\bf 743}, 13-56 (2004).
737:
738:
739: \bibitem{dum} A. Dumitru, A. Hayashigaki and J. Jalilian-Marian,
740: arXiv:hep-ph/0506308.
741: \bibitem{BK} I. Balitsky, Nucl. Phys. B {\bf 463}, 99 (1996); Y. V. Kovchegov, Phys. Rev. D {\bf 60}, 034008
742: (1999).
743: \bibitem{KLN} D. Kharzeev, E. Levin and M. Nardi,
744: Nucl. Phys. A {\bf 730}, 448-459 (2004), Erratum-ibid. A {\bf
745: 743}, 329-331 (2004).
746: \bibitem{KovMul} Y. V. Kovchegov and A. H. Mueller,
747: Nucl. Phys. B {\bf 529}, 451-479 (1998).
748: \bibitem{DuMcl} A. Dumitru and L. D. McLerran, Nucl. Phys. A {\bf 700}, 492-508 (2002).
749:
750:
751:
752: \bibitem{kkp} B. A. Kniehl, G. Kramer and B. P\"otter, Nucl. Phys. B {\bf 597}, 337-369 (2001).
753:
754: \bibitem{AF} P. Aurenche, M. Fontannaz, J.-Ph. Guillet, B. A. Kniehl and M. Werlen, Eur. Phys. J. C {\bf 13},
755: 347-355 (2000); M. Fontannaz, J.-Ph. Guillet and G. Heinrich, Eur.
756: Phys. J. C {\bf 22}, 303-315 (2001).
757: \bibitem{Av} F. Aversa, P. Chiappetta, M. Greco and J.-Ph. Guillet, Nucl. Phys. B {\bf 327}, 105 (1989); B. J\"ager, A. Sch\"afer, M. Stratmann and W. Vogelsang, Phys. Rev. D {\bf 67}, 054005 (2003); D. de Florian, Phys. Rev. D {\bf 67}, 054004 (2003).
758: \bibitem{Adler} S. S. Adler {\it et al.} [PHENIX Collaboration],
759: Phys. Rev. Lett. {\bf 91}, 241803 (2003).
760:
761: \bibitem{Bethe} H. A. Bethe, Phys. Rev. D {\bf 89}, 1256 (1953).
762: \bibitem{BDMPS} R. Baier, A. H. Mueller, Yu. L. Dokshitzer, S. Peign\'e and D. Schiff, Nucl. Phys. B {\bf 484}, 256 (1997).
763: \bibitem{Strik} V. Guzey, M. Strikman and W. Vogelsang, Phys. Lett. B {\bf 603}, 173-183 (2004).
764:
765: \bibitem{arleo} F. Arleo {\it et al.}, arXiv:hep-ph/0311131.
766:
767:
768: \bibitem{grv} M. Gluck, E. Reya and A. Vogt, Eur. Phys. J. C {\bf 5}, 461 (1998).
769:
770:
771:
772:
773: \bibitem{Jd} A. Dumitru and J. Jalilian-Marian, Phys. Lett. B {\bf
774: 547}, 15 (2002).
775:
776: \bibitem{munp} S. Munier and R. Peschanski, Phys. Rev. Lett. {\bf 91}, 232001 (2003).
777: \bibitem{munp2} S. Munier and R. Peschanski, Phys. Rev. D {\bf 69}, 034008 (2004).
778: \bibitem{MuT} A. H. Mueller and D. N. Triantafyllopoulos, Nucl. Phys. B {\bf 640}, 331-350 (2002).
779: \bibitem{IIMc} E. Iancu, K. Itakura and L. D. McLerran, Nucl. Phys. A {\bf 708}, 327 (2002).
780: \bibitem{BFKL} E. A. Kuraev, L. N. Lipatov and V. S. Fadin, Sov.
781: Phys. JETP {\bf 45}, 199 (1977) [Zh. Eksp. Teor. Fiz. {\bf 72},
782: 377 (1977)]; I. I. Balitsky and L. N. Lipatov, Sov. J. Nucl. Phys.
783: {\bf 28}, 822 (1978) [Yad. Fiz. {\bf 28}, 1597 (1978)].
784: \bibitem{IIM} E. Iancu, K. Itakura and S. Munier, Phys. Lett. B {\bf 590}, 199-208 (2004).
785: \bibitem{BMS} R. Baier, A. H. Mueller and D. Schiff, Nucl. Phys. A {\bf 741}, 358-380 (2004).
786: \bibitem{Reta} J. J. Gaardhoje [BRAHMS Collaboration], Nucl. Phys. A {\bf
787: 734}, 13 (2004).
788: \bibitem{BaierQ} R. Baier, A. H. Mueller, Yu. L. Dokshitzer and D. Schiff, JHEP 0109, 033 (2001).
789:
790:
791: \end{thebibliography}
792: \end{document}
793: