1: %\documentclass[12pt]{article}
2: \documentclass[twocolumn,prl,aps,floatfix]{revtex4}
3: %\usepackage{showkeys}
4: %\usepackage{floatfix}
5: \usepackage{psfig}
6: \topmargin -1cm
7: \newcommand{\beq}{\begin{equation}}
8: \newcommand{\eeq}{\end{equation}}
9: \newcommand{\beqn}{\begin{eqnarray}}
10: \newcommand{\eeqn}{\end{eqnarray}}
11: \newcommand{\bs}{\mbox{\boldmath $s$}}
12: \newcommand{\bb}{\mbox{\boldmath $b$}}
13: \def\la{\langle}
14: \def\ra{\rangle}
15: \def\mb{\,\mbox{mb}}
16: \def\pb{\,\mbox{pb}}
17: \def\fm{\,\mbox{fm}}
18: \def\GeV{\,\mbox{GeV}}
19: \def\MeV{\,\mbox{MeV}}
20: \def\pt{$ p_{t}$ }
21: \def\xf{$ x_F$ }
22: \def\rt{$
23: \sqrt\tau$ }
24: \def\u{$\Upsilon$ }
25: \def\JP{$\psi'$ }
26: \def\J{$J/\psi$ }
27: \def\up{$\Upsilon^{\prime}$ }
28: \def\upp{$\Upsilon^{\prime\prime}$ }
29: \def\D{$^2H$}
30:
31:
32: \begin{document}
33: \title{Nuclear Broadening of Transverse Momentum in Drell-Yan Reactions}
34: \author{
35: M.B.~Johnson$^a$,
36: B.Z.~Kopeliovich$^{b,c,d}$,
37: M.J.~Leitch$^{a}$,
38: P.L.~McGaughey$^{a}$,
39: J.M.~ Moss$^{a}$\\
40: I.K.~Potashnikova$^{b,c}$,
41: Ivan~Schmidt$^{b}$}
42: \affiliation{$^a$ Los Alamos National Laboratory, MS H846, Los Alamos,
43: NM 87545, USA\\
44: $^b$ Departamento de F\'isica, Universidad Tecnica Federico Santa
45: Mar\'ia, Valpara\'iso, Chile\\
46: $^c$ Joint Institute for Nuclear Research, Dubna, Russia\\
47: $^d$ Institute of Theoretical Physics, Heidelberg University,
48: Heidelberg, Germany}
49:
50: %\date{\today}
51:
52: \begin{abstract}
53: Data for Drell-Yan (DY) processes on nuclei are currently available from
54: fixed target experiments up to the highest energy of $\sqrt{s}=40\GeV$.
55: The bulk of the data cover the range of short coherence length, where the
56: amplitudes of the DY reaction on different nucleons do not interfere. In
57: this regime, DY processes provide direct information about broadening of
58: the transverse momentum of the projectile parton experiencing
59: initial-state multiple interactions. We revise a previous analysis of data
60: from the E772 experiment and perform a new analysis of broadening
61: including data from the E866 experiment at Fermilab. We conclude that the
62: observed broadening is about twice as large as the one found previously.
63: This helps to settle controversies that arose from a comparison of the
64: original determination of broadening with data from other experiments and
65: reactions.
66:
67: \smallskip
68:
69: \noindent
70: PACS: 13.85.Qk; 24.85.+p
71: \end{abstract}
72: \maketitle
73:
74: \vspace*{0.5cm}
75:
76: \section{Introduction}
77:
78: Understanding how partons propagate in a nucleus reveals precious
79: information about the early stages of hadronization that is difficult to
80: obtain by other means. This understanding also plays a role in using
81: partons as a sensitive probe of matter produced in relativistic heavy ion
82: collisions. One of the important observables is broadening of the
83: transverse momentum of produced particles due to multiple interactions of
84: a parton in the medium. This effect can be studied for the nuclear medium
85: in Drell-Yan reactions \cite{na10,e772,mmp} and deep-inelastic scattering
86: \cite{hermes,jlab} on nuclear targets. The former is an especially clean
87: tool, since the produced dilepton, which has no final state interactions,
88: carries undisturbed information on the transverse momentum of the parton
89: initiating the reaction.
90:
91: A parton propagating through a nuclear medium experiences Brownian motion
92: in the transverse momentum plane \cite{pirner}, with the increase of the
93: mean transverse momentum squared being proportional to the length of the
94: path times the medium density $\rho(z)$,
95: \beq
96: \Delta\la p_T^2(L)\ra = 2C(s)\int\limits_0^L dz\,\rho(z),
97: \label{10}
98: \eeq
99: where $z$ is the longitudinal coordinate. The energy dependent
100: dimensionless factor $C$ characterizes multiple interactions of the parton
101: with bound nucleons. In the light-cone dipole description of the
102: broadening process \cite{dhk,jkt,at}) $C$ is related to the small
103: separation behavior of the dipole-nucleon cross section,
104: \beq
105: C(s)=\left.\frac{d\sigma_{\bar qq}(r_T,s)}
106: {dr_T^2}\right|_{r_T=0}\ .
107: \label{20}
108: \eeq
109: The cross section $\sigma_{\bar qq}(r_T,s)$ introduced in \cite{zkl}
110: varies with transverse $\bar qq$ separation $\vec r_T$ and energy. This
111: quantity is difficult to calculate, since it involves nonperturbative
112: effects, but it can be fitted to data for the photoabsorption cross
113: section and the proton structure function $F_2(Q^2,x)$, which probe a wide
114: variety of separations and energies \cite{gbw,kst2}. Since each of the
115: scatterings is accompanied by a loss of energy, there is a close
116: connection between the multiple scattering that leads to momentum
117: broadening and initial state gluonic energy loss \cite{bdmps,vitev}.
118: It was found, however, that this induced energy loss effects the DY
119: process much less than the vacuum gluon radiation initiated by the
120: interaction on the nuclear surface \cite{e-loss}.
121:
122: In the parton model, a single interaction of a quark propagating through
123: the medium can be viewed as a quark-gluon correlation in a nucleus in its
124: infinite momentum frame. The corresponding twist-4 term \cite{qs} cannot
125: be calculated, but is rather fitted to data. This phenomenology has not
126: been successful so far, with the parameter extracted from data varying by
127: a factor of 20, depending on assumptions \cite{guo} and also demonstrating
128: a strong process dependence \cite{ww}.
129:
130:
131: The first observation of nuclear enhancement of large transverse momenta
132: was made in hadron-nucleus collisions \cite{cronin}. A sensitive tool to
133: measure the transverse momentum broadening of quarks propagating through
134: nuclei is heavy dilepton production in the Drell-Yan reaction. The leptons
135: have no interaction in the nuclear medium and thus carry undisturbed
136: information about the transverse momentum of the quark. Nuclear broadening
137: of the mean transverse momentum squared of the dilepton is defined as
138: \beq
139: \Delta\la p_T^2\ra = \la p_T^2\ra_A -
140: \la p_T^2\ra_N\ .
141: \label{22}
142: \eeq
143: Perturbative QCD predicts $d\sigma_{DY}/dp_T^2\propto 1/p_T^4$ (for
144: transversely polarized virtual photons) at large $p_T$. Therefore, the
145: mean value $\la p_T^2\ra$ is divergent, and the broadening of
146: Eq.~(\ref{22}) might be ill defined (see discussion in \cite{krtj}). Of
147: course, the experiment imposes a cut off. Thus, the experimentally
148: observed broadening is finite, but it may depend on the experimental
149: acceptance. This divergence, however, sometimes cancels in the broadening
150: for several reasons. First of all, perturbative QCD predicts no nuclear
151: effects at large $p_T$, i.e., $d\sigma_{DY}(pA)/dp_T^2 \to
152: A\,d\sigma_{DY}(pp)/dp_T^2$ (see, e.g., \cite{knst}). Secondly, nuclear
153: shadowing is expected to vanish at large $p_T$, especially at an energy as
154: low as $800\GeV$, the energy of our current analysis. In this case, the
155: divergent high-$p_T$ part of the integrations cancel in (\ref{22}), and
156: one can measure the nuclear broadening effect within a restricted interval
157: of $p_T$ where the data exhibit nuclear effects.
158:
159:
160: An analysis of broadening in the Drell-Yan reaction was presented in
161: \cite{mmp} using data from the E772 experiment at Fermilab. The results of
162: the analysis for $\Delta\la p_T^2\ra_{DY}$ are depicted in
163: Fig.~\ref{e772-old} as function of $A$. \\[0.2cm]
164: \begin{figure}[tbh]
165: \centerline{\psfig{figure=a-dep.eps,width=7cm}}
166: \protect\caption{Broadening of transverse momentum in production of
167: heavy dileptons (squares), $J/\Psi$ (circles) and $\Upsilon$ (triangles)
168: in $pA$ collisions at $800\GeV$. The data are from the E772, E789, and
169: E866 experiments at Fermilab \cite{e772,e772a,e789,e866-pt}. The curves
170: are results of a fit with the parametrization $\Delta\la p_T^2\ra
171: =D\left[(A/2)^{1/3}-1\right]$.}
172: \label{e772-old}
173: \end{figure}
174:
175: Following Eq.~(\ref{10}), one can parametrize the $A$-dependence as
176: \beq
177: \Delta\la p_T^2\ra =D\left[(A/2)^{1/3}-1\right]
178: \label{25}
179: \eeq
180: A fit to the data points for the Drell-Yan reaction in Fig.~\ref{e772-old}
181: results in $D=0.027\GeV^2$, which corresponds to $C=2$ in (\ref{10}).
182:
183: The same figure presents nuclear broadening in $J/\Psi$ and $\Upsilon$
184: production at the same energy. Comparing with this data and results
185: available at lower energies, one can make the following observations: (i)
186: the magnitude of broadening in reactions of heavy dilepton and quarkonium
187: production differ by a factor of 5, while one would expect the Casimir
188: factor $9/4$. Indeed, heavy quarkonia should be produced via gluons, which
189: interact in the medium more strongly than quarks. (ii) The errors for
190: Drell-Yan data are small compared to $J/\Psi$ production, whereas the
191: latter is measured with much higher statistics. (iii) Comparison with
192: broadening measured at lower energies presented in Fig.~\ref{e-dep}
193: suggests that broadening in heavy quarkonium production should rise with
194: energy, whereas it is seen to be rather flat, or even perhaps drop, at
195: high energies in the Drell-Yan reaction. \\[0.2cm]
196:
197: \begin{figure}[tbh]
198: \centerline{\psfig{figure=s-dep.eps,width=7cm}}
199: \protect\caption{Energy dependence of the parameter $D$ in (\ref{25})
200: fitted to data for dilepton and heavy quarkonium production in $pA$ and
201: $\pi A$ collisions.}
202: \label{e-dep}
203: \end{figure}
204: The latter observation is difficult to explain, since the cross section
205: controlling the multiple interactions in the medium rises with energy.
206: (iv) The dipole approach, Eq.~(\ref{20}), based on a phenomenological color dipole
207: cross section fitted to data, predicts the factor $C$ \cite{jkt} to be
208: about twice as large as one suggested by the data presented in
209: Fig.~\ref{e772-old}. \\
210:
211: In this paper we revise the result of the previous analysis \cite{mmp} of
212: the E772 data for Drell-Yan nuclear ratios, combining them with the
213: brand-new data \cite{webb} from the E866 experiment for Drell-Yan cross
214: section in $pD$ collisions. The improvement in the analysis arises from
215: the fact that the E866 experiment provides more accurate $pD$ cross
216: section data at small $p_T$ than that available from E772. Additionally,
217: it was assumed with no good justification that the same simple
218: parametrization of the $p_T$-dependence is valid for both $pD$ and $pA$
219: collisions, which dictated a specific shape for the ratios and led to a
220: substantial underestimate of the errors for $p_T$-broadening. Our new
221: analysis is less model-biased, but it results in somewhat larger errors.
222: We also perform here an analysis of new data from the E866 experiment,
223: which leads to a broadening about twice as large as in \cite{mmp},
224: although with larger experimental errors. This result resolves the
225: problems stated earlier.
226:
227:
228: \section{Light-cone dipole description of nuclear broadening}
229:
230: When a projectile scatters from a nucleus, one can treat the target nucleons as independent with some degree of accuracy if viewed in the rest frame of the nucleus, since the bound nucleons are sufficiently well separated in that frame.
231: However, when boosted to the infinite momentum frame the same reaction looks different since the nucleons may start
232: communicating via their parton clouds at small Bjorken $x$. Indeed the
233: Lorentz contraction is proportional to $1/x$, therefore it is much
234: stronger at large, than at small $x$. As a result, the parton clouds are
235: still separated for $x>x_c$, but overlap and interact at smaller $x$,
236: $x<x_c$. The border line $x_c$ can be estimated as,
237: \beq
238: x_c \sim \frac{1}{m_N\,R_A}\ ,
239: \label{26}
240: \eeq
241: where $R_A$ is the nuclear radius. Note that this is only an order of
242: magnitude estimate. There is a numerical factor in (\ref{26}) that varies
243: considerably for different species of partons and the way the
244: parton distribution is probed \cite{krt2}.
245:
246: The interaction between partons at $x<x_c$ leads to a modification of
247: their transverse momentum distribution: suppression of small momenta,
248: enhancement at intermediate, and no effect at large transverse momenta.
249: This phenomenon is known as the saturation effect \cite{glr}, or color
250: glass condensate \cite{mv}. One should keep in mind, however, that onset
251: of this effect happens only at $x<x_c$.
252:
253: The partonic interpretation of hard reactions at small Bjorken $x$ is
254: known to be Lorentz non-invariant. Deep-inelastic scattering (DIS) looks
255: like absorption of the virtual photon by a quark (antiquark) that belongs
256: to the target in the infinite momentum frame of the latter. The reaction looks,
257: however, quite different in the target rest frame, as the interaction with a
258: $\bar qq$ fluctuation of the high-energy virtual photon. Of course, all
259: observables, including the structure function $F_2(x,Q^2)$, which is
260: proportional to the total photon-target cross section, are Lorentz
261: invariant.
262:
263: The Drell-Yan reaction is usually interpreted as annihilation of a quark
264: with an antiquark belonging to the beam and target respectively (or vice
265: versa), $\bar qq \to \bar ll$ \cite{dy}. While this is correct in the rest
266: frame of the dilepton, this interpretation of a heavy dilepton production
267: is not applicable in target rest frame. Indeed, in order to satisfy the
268: kinematics annihilating with the projectile, the target parton must move
269: with momentum $p_L\sim p_T/x_2$, where $x_2$ in standard notation is
270: the Bjorken $x$ of the target parton. At small $x_2$, $x_2\ll1$, this would mean
271: presence of highly energetic partons in the stationary target. The proper
272: space-time interpretation in this case is analogous to the
273: Weitz\"acker-Williams mechanism of electromagnetic bremsstrahlung. Namely,
274: a beam quark (antiquark) develops a fluctuation, $q\to q\gamma^*$ (with
275: $\gamma^*\to\bar ll$), which interacts with the target freeing the
276: dilepton\cite{k,kst1}.
277:
278: What looked like overlap of parton clouds at small $x_2$ in the infinite
279: momentum frame, now takes the form of interference between the amplitudes
280: of a parton interacting with different nucleons. On the same footing it can
281: also be interpreted as the lifetime of the $|q\gamma^*\ra$ fluctuation,
282: given by the uncertainty principle
283: \beq
284: l_c=\frac{1}{q_L} =
285: \frac{2E_q}{M^2_{q\gamma^*}-m_q^2}\ , \label{28}
286: \eeq
287: where $q_L$ is the longitudinal momentum transfer in the $q\to q\gamma^*$
288: fluctuation. The effective mass squared of the fluctuation is
289: $M_{q\gamma^*}^2=M_{\gamma^*}^2/\alpha+m_q^2/(1-\alpha)+k_T^2/\alpha(1-\alpha)$,
290: where $\alpha$ is the fraction of the light-cone momentum of the parent
291: quark taken by the dilepton ($\gamma^*$), and $k_T$ is the $\gamma^*-q$
292: relative transverse momentum. Hereafter we do not discriminate between
293: coherence time and length, assuming that the quark has the speed of light.
294: Evidently, the same length $\sim1/q_L$ defines the maximal longitudinal
295: distance between two scattering centers that can interfere. This is why
296: the distance, Eq.~(\ref{28}), is also called coherence length.
297:
298:
299: In the case of a nuclear target, the coherence length is of great importance.
300: If this time interval is small compared to the mean spacing of nucleons in
301: the nucleus, dileptons are radiated in the Bethe-Heitler regime, i.e., with
302: no interference between the amplitudes of radiation from different
303: nucleons. At the another extreme, $l_c\gg R_A$, the Landau-Pomeranchuk
304: phenomenon governs the radiation. Namely, interferences suppress radiation
305: with small transverse momenta, enhance it at intermediate values of few
306: GeV, and make no changes at higher momentum transfer. In both cases, the
307: linear dependence on the path length, or nuclear thickness,
308: Eq.~(\ref{10}), holds, although with different factors $C$ \cite{krtj}.
309: The borderline between the two regimes, $l_c=R_A$, corresponds to the
310: relation for $x_c$, Eq.~(\ref{26}).
311:
312:
313: \subsection{Short coherence length}
314:
315: The mechanism of nuclear broadening depends on how long the coherence
316: length is compared to the nuclear size. In the regime of short coherence
317: length, $l_c\ll R_A$, the dilepton fluctuation appears for a very short
318: time deep inside the nucleus and is immediately released on mass shell by
319: the interaction, which is assumed to be the same as on a free nucleon.
320: Therefore, this interaction does not lead to any change in the transverse
321: momentum dependence, i.e., it provides no broadening. However, the initial
322: state interactions of the projectile quark affect the transverse momentum
323: distribution of the quark, which acquires the shape \cite{jkt}
324: \beqn
325: \frac{dN_q}{d^2p_T} &=&
326: \int d^2r_1\,d^2 r_2\,e^{
327: i\,\vec p_T\,(\vec r_1 - \vec r_2)}\,
328: \Omega^q_{in}(\vec r_1,\vec r_2)\nonumber\\
329: &\times& e^{-{1\over2}\,
330: \sigma^N_{\bar qq}(\vec r_1-\vec r_2,x)\,T_A(b)}\ .
331: \label{30}
332: \eeqn
333: Here $\Omega^q_{in}(\vec r_1,\vec r_2)$ is the density matrix describing
334: the impact parameter distribution of the quark in the incident hadron.
335: We assume a Gaussian shape,
336: \beq
337: \Omega^q_{in}(\vec r_1,\vec r_2) =
338: \frac{\la p_0^2\ra}{\pi}\,
339: e^{-{1\over2}(r_1^2+r_2^2)
340: \la p_0^2\ra}\ ,
341: \label{33}
342: \eeq
343: where $\la p_0^2\ra$ is the mean value of the primordial
344: transverse momentum squared of the quark.
345:
346: The transverse momentum of the radiated dilepton reflects the broadening acquired by the quark in the nuclear medium. However, the effect is weaker than for the quark, since the
347: transverse momenta of the dilepton and quark are connected by the
348: relation,
349: \beq
350: q_T^{\bar ll} = \alpha\,p_T^q\ ,
351: \label{35}
352: \eeq
353: where $\alpha < 1$ is the fractional momentum of the $\bar ll$. This
354: relation clearly demonstrates that the effect of broadening cannot be
355: translated into a modification of the quark distribution function (as is
356: frequently done), in the regime of short coherence length. Thus, it
357: confirms the breakdown of $k_T$-factorization imposed by the initial state
358: interaction, as suggested in \cite{bbl}.
359:
360:
361:
362: \subsection{\bf No-shadowing sum rule}
363:
364: To the extent that the coherence length for the Drell-Yan reaction is
365: short compared to the internucleon spacing in the nucleus, no shadowing is
366: possible. This means that in spite of the nuclear modification of the
367: transverse momentum dependence of Drell-Yan cross section, the number of
368: quarks is conserved. This fact can be represented in the form of a sum
369: rule,
370: \beq
371: \int d^2p_T\,\frac{d\sigma_{DY}(pA)}{d^2p_T} =
372: A\,\int d^2p_T\,\frac{d\sigma_{DY}(pN)}{d^2p_T}\ .
373: \label{40}
374: \eeq
375: The nuclear ratios measured in the E772 and E866 experiments at $800GeV$
376: depicted in Fig.~\ref{pt-e772} have a tendency to satisfy this constraint.
377: Indeed, the ratios are below one at small $p_T$, but rise above one at
378: larger momentum transfers.
379: \begin{figure}[tbh]
380: \centerline{\psfig{figure=ratios.ps,width=8cm}}
381: \protect\caption{The nucleus-to-deuterium and -beryllium ratios of
382: dilepton production cross sections as function of transverse momentum. The
383: data are from the E772 \cite{e772} and E866 \cite{e866-pt} experiments at
384: $800\GeV$.}
385: \label{pt-e772}
386: \end{figure}
387: The weak nuclear suppression observed at small $p_T$, where the absolute
388: cross section is maximal, seems to be compensated by a stronger
389: enhancement at larger $p_T$ in accordance with the sum rule
390: Eq.~(\ref{40}). Data in the small $p_T < 1-2\GeV$ domain, where the main
391: part of the cross section comes from, are most important for broadening.
392: In this region the E866 data in Fig. 3 demonstrate stronger nuclear
393: suppression than the E772 data. The observed deviation from unity at small
394: $p_T$, which is responsible for broadening, is about twice as large in
395: E866 as in E772 data.
396:
397: Notice that the $p_T$-dependence of the ratios in Fig.~\ref{pt-e772} has
398: the typical shape predicted for the regime of saturation (of quarks), or
399: color glass condensate \cite{mv}. Therefore, it might be tempting to
400: interpret such data as an observation of this phenomenon. We warn again
401: that this would be a wrong conclusion, since coherence and saturation are
402: impossible in the short coherence length regime; initial state interaction
403: effects imitate saturation.
404:
405:
406: \subsection{Long coherence length}
407:
408: In this regime, $l_c\gg R_A$, the amplitudes $\gamma^*\to\bar ll$ for
409: radiating a heavy photon from different nucleons interfere with each
410: other. This interference, which modifies the $p_T$ distribution of
411: radiation, is known as the Landau-Pomeranchuk effect and is a genuine
412: effect of saturation.
413:
414: Alternatively, one may say that a long-lived fluctuation $q\to\gamma^*q$
415: that propagates through the entire nucleus may be freed as a result of the
416: multiple interactions, in contrast to what happens in the regime of short
417: $l_c$, where the fluctuation is freed after an interaction with a single
418: nucleon. These multiple interactions also lead to a different type of
419: broadening mechanism. Its origin can be understood from the fact that most
420: of the $q\to\gamma^*q$ fluctuations in the quark are invisible in the
421: sense that their interactions with the target are so weak that the target
422: has restricted resolving power or ability to discriminate a hard
423: $q\gamma^*$ fluctuation, with large intrinsic transverse momentum $k_T$,
424: from a bare quark. However, because of the larger momentum transfer
425: allowed by the multiple interactions in the nuclear medium, the resolution
426: provided by the nucleus is greater than that of a free nucleon. The
427: nucleus can therefore free harder fluctuations than a free nucleon. This
428: is the source of nuclear broadening in the large coherence length regime.
429:
430: The same effect can be interpreted within the color dipole approach in
431: terms of color filtering. The nucleus filters out projectile dipoles of
432: smaller size than is possible on a free nucleon, leading to larger
433: transverse momenta. In this case, instead of Eq.~(\ref{30}), the
434: transverse momentum distribution of dileptons radiated by a quark
435: interacting with a nucleus is given by the factorized dipole formula
436: \cite{kst1},
437: \beqn
438: && \frac{d\sigma(qA\to \bar ll X)}
439: {d^2q_T} =
440: \int d^2b\int d^2r_1d^2r_2\,
441: e^{i\vec q_T(\vec r_1-\vec r_2)}\,
442: \nonumber\\ &\times&
443: \Psi_{q\gamma^*}^\dagger(\vec r_1,\alpha)\,
444: \Psi_{q\gamma^*}(\vec r_2,\alpha)
445: \left[1 - e^{-{1\over2}\sigma^N_{\bar qq}(r_1,x_2)T_A(b)}
446: \right.\nonumber\\ &-& \left.
447: e^{-{1\over2}\sigma^N_{\bar qq}(r_2,x_2)T_A(b)} +
448: e^{-{1\over2}\sigma^N_{\bar qq}(\vec r_1-\vec r_2,x_2)T_A(b)}
449: \right]\ .
450: \label{80}
451: \eeqn
452: Here $\Psi_{q\gamma^*}(\vec r,\alpha)$ is the pQCD calculated light-cone
453: distribution amplitude, which describes the dependence of the
454: $|q\gamma^*\ra$ Fock component on transverse separation and fractional
455: momentum. The nuclear thickness function at impact parameter $b$ is given
456: by the integral of the nuclear density along the quark trajectory,
457: $T_A(b)=\int_{-\infty}^\infty dz\,\rho_A(b,z)$.
458:
459: The ratio of the nuclear to nucleon cross sections of the Drell-Yan
460: reaction calculated in \cite{kst1} has a shape typical for the Cronin
461: effect. Such a modification of the transverse momentum dependence is a
462: genuine effect of saturation (or its onset). In spite of the similarity
463: with the analogous results in the short $l_c$ regime, the mechanisms and
464: their formulations in Eqs.~(\ref{30})-(\ref{80}) are different.
465: Additionally, the magnitude of the Cronin effect in the long $l_c$ regime
466: is smaller \cite{krtj}.
467:
468:
469: \section{Analysis of E772/866 data for Drell-Yan reaction}
470:
471: Here we determine the nuclear broadening $\Delta\la p_T^2\ra$ of dileptons
472: produced on different nuclear targets relying on data for $p_T$-dependent
473: ratios depicted in Fig.~\ref{pt-e772}. These include the
474: nucleus-to-deuterium ratios $R^{A/D}$ from the E772 \cite{e772}
475: experiment, and nucleus-to-beryllium ratios $R^{A/Be}$ from the E866
476: experiment \cite{e866-pt}, both at $800\GeV$. These results are quite
477: reliable since they are least affected by systematic uncertainties. The
478: ratios $R_i$ are available in 7 bins in $p_T$, $i=1-7$. Correspondingly,
479: one can calculate the broadening, Eq.~(\ref{22}), as
480: \beq
481: \Delta\la p_T^2\ra =
482: \frac{\sum_i \sigma_i\,R_i\,p_i^2}{\sum_i \sigma_i\,R_i}
483: - \frac{\sum_i \sigma_i\,p_i^2}{\sum_i \sigma_i}\ ,
484: \label{60}
485: \eeq
486: where for each $i$-th bin
487: \beq
488: \sigma_i=\int\limits_{(p_T)_i^{min}}^{(p_T)_i^{max}}
489: d^2p_T\,\frac{d\sigma_{DY}^N}{d^2p_T}\ ,
490: \label{70}
491: \eeq
492: with $p_i^2$ the mean value of $p_T^2$ within the $i$-th bin. In addition
493: to the seven $p_T$-intervals in data depicted in Fig.\ref{pt-e772}, we
494: have included an eighth interval in Eq.~(\ref{60}) covering
495: $(p_T)_7^{max}<p_T<\infty$ and assumed that $R_8=1$ in accordance with
496: perturbative QCD expectations \cite{knst} (see, however, next section).
497: Since the cross section falls steeply at large $p_T$, this point should
498: not substantially affect the results.
499:
500: The differential cross section for the Drell-Yan reaction on a free
501: nucleon, $d\sigma_{DY}^N/d^2p_T$, is more affected by systematic
502: uncertainties than the ratio is. We rely on the recently published
503: results of the E866 experiment for the differential cross section in $pp$
504: and $pD$ collisions \cite{webb}. We use this data in (\ref{70}) for the
505: analyses of all data presented in Fig.~\ref{pt-e772}.
506:
507: We parametrize the differential cross section of the Drell-Yan reaction as
508: \beq
509: \frac{d\sigma}{dp_T^2} = N\,
510: \frac{\left(1+\frac{p_T^2}{\lambda_1^2}\,
511: e^{-p_T^2/\lambda_2^2}\right)}
512: {(1+p_T^2/\lambda_3^2)^n}\ .
513: \label{90}
514: \eeq
515: The factor in the numerator is introduced to describe a possible forward
516: minimum in the cross section indicated by data \cite{webb}.
517:
518: We found that the shape of the $p_T$ dependence does not vary with
519: dilepton mass and $x_2$, as one can see in Fig.~\ref{fit}.
520: \begin{figure}[tbh]
521: \centerline{\psfig{figure=pd.eps,width=8.3cm}}
522: \protect\caption{Differential cross section of Drell-Yan reaction in
523: proton-deuteron collisions. Data are from the E866 experiment
524: \cite{e866-pt} at $800\GeV$ for different intervals of dilepton effective
525: mass. Curves show the result of the global fit with common parameters,
526: Eq.~(\ref{90}), controlling the shape of the cross section.}
527: \label{fit}
528: \end{figure}
529: A global fit to Drell-Yan data from $pD$ collisions measured by the E866
530: experiment for different $M$ and $x_2$ bins \cite{webb} with common
531: parameters $n$, $\lambda_i$ and different normalization factors $N$, led
532: to quite a good description, $\chi^2 = 83.3$ with $N_{d.o.f.}=52$. The
533: values of the parameters are $\lambda_1=0.78 \pm 0.08$; $\lambda_2=0.65
534: \pm 0.04$; $\lambda_3=3.31 \pm 0.21$; $n=7.00\pm0.66$. The results of the
535: fit and the data are depicted in Fig.~\ref{fit}. The error bars include
536: both statistical and systematic errors added in quadrature.
537:
538: Note that at large $p_T$ the $pD$ cross section in Eq.~(\ref{90}) is
539: falling as $p_T^{-14}$, much steeper than one may expect from perturbative
540: QCD. This happens due to restrictions imposed by finiteness of energy,
541: namely, both $x_1$and $x_2$ rise with $p_T$ towards the kinematic limit
542: causing an additional suppression of the cross section \cite{krtj}.
543: Additionally, available data cover a rather restricted interval of $p_T <
544: 4\GeV$.
545:
546: With the cross section in Eq.~(\ref{90}) we performed an analysis of data
547: from both experiments, E772 and E866. While the ratios measured in E772
548: are free from major systematic uncertainties, the absolute cross sections
549: are substantially less reliable. Indeed, comparison of the differential
550: cross sections of the Drell-Yan process measured for $p-D$ collisions in
551: the two experiments reveals substantial differences. Therefore, we rely on
552: the $p-D$ cross section from the E866 experiment. The results of such a
553: combined analysis of the E772 ratios are depicted in Fig.~\ref{results} by
554: closed squares.
555: \begin{figure}[tbh]
556: \centerline{\psfig{figure=results.ps,width=8cm}}
557: \protect\caption{Results of the present analysis for broadening in the
558: Drell-Yan reaction on different nuclei. Closed squares correspond to E772
559: data for $C,\ Ca,\ Fe$ and $W$. Open squares correspond to E866 data for
560: $Fe$ and $W$. For comparison, broadening for $J/\Psi$ and $\Upsilon$, the
561: same as in Fig.~\ref{e772-old}, are also shown.}
562: \label{results}
563: \end{figure}
564:
565: The ratios measured in the E866 experiment are for a nucleus relative to
566: beryllium. We use the chain relation between the two ratios,
567: \beq
568: R_{A/D}=R_{A/Be}\,R_{Be/D}\ ;
569: \label{99}
570: \eeq
571: however, the beryllium-to-deuterium ratio is unfortunately lacking.
572: Nevertheless, this cannot be a significant correction, since beryllium is
573: a light nucleus. Indeed, the carbon to deuterium ratio depicted in
574: Fig.~\ref{pt-e772} is compatible with no nuclear effects at all. If we
575: simply fix $R_{Be/D}=1$, the broadening given in Eq.~(\ref{60}) will be
576: underestimated. However, one can do better by relying on a theoretically
577: predicted value for $R_{Be/D}$ \cite{mikk} with Eq.~(\ref{30}). Of course,
578: this introduces an uncertainty related to model dependence, but this
579: uncertainty should not exceed few percent of the observed broadening. The
580: results of this analysis of the E866 data are depicted in
581: Fig.~\ref{results} by empty squares.
582:
583: Notice that the cross section ratios are affected by systematic
584: uncertaities much less than the absolute cross sections. The ratios
585: measured in both experiments have mainly an overall systematic uncertainty
586: $1\%$ in E866 \cite{e866-pt}, and $2\%$ in E772 \cite{e772} measurements.
587: One can see from Eq.~(\ref{60}) that an overall variation of all $R_i$
588: leaves the broadening unchanged. Therefore we could disregard the
589: systematic errors of the ratios.
590:
591: If we fit the data in Fig.~\ref{results} assuming a linear dependence on
592: $A^{1/3}$, in accordance with (\ref{25}), and do this separately for the
593: E772 and E866 data, we obtain,
594: \beqn
595: D(E772)=(0.029 \pm 0.008)\,(\GeV/c)^2\ ;
596: \nonumber\\
597: D(E866)=(0.059 \pm 0.009)\,(\GeV/c)^2\ .
598: \label{100}
599: \eeqn
600:
601: The value of $D(E772)$ is close to the result of previous analysis
602: \cite{mmp} depicted in Fig.~\ref{e772-old}, although with a considerably
603: larger error. However, the E866 data suggest a much stronger broadening.
604: Nevertheless, one may consider the results Eq.~(\ref{100}) as nearly
605: consistent since they lie within two standard deviations. If the
606: broadening according to the E866 data were actually about twice as large
607: as that given by the previous analysis \cite{mmp} of the E772 data, this
608: would settle the contradiction with theoretical expectations and
609: conclusions drawn from the data for the broadening of $J/\Psi$ and
610: $\Upsilon$ mentioned above.
611:
612:
613: \section{Beyond the data}
614:
615: Although QCD predicts no nuclear effects at high $p_T$, the data in
616: Fig.~\ref{pt-e772} show considerable nuclear enhancement at intermediate
617: values of $p_T$. This is the well-known Cronin effect first observed for
618: hadronic beams \cite{cronin}. It is not clear from the data how far in
619: $p_T$ this effect can propagate. The assumption made above for the last
620: unobserved $p_T$ interval, $R_8=1$, is probably incorrect, therefore the
621: broadening effect shown in Fig.~\ref{results} could be be underestimated.
622: The ad hoc shape of $R(p_T)$ used in \cite{mmp} is difficult to justify
623: and does not suggest a solution to the problem.
624:
625:
626: In order to get a hint about how much of the effect might have been
627: missed, one may make use of a theoretically calculated value for $R_8$. We
628: rely on the dipole formalism, Eqs.~(\ref{30})-(\ref{35}), from which one
629: is able to calculate the $p_T$ dependence of the Drell-Yan reaction on
630: proton and nuclear targets in a parameter-free way and explain the data in
631: Fig.~\ref{pt-e772} well with no adjustment of parameters. The dipole
632: approach also provided the first quantitative explanation of the Cronin
633: effect in available data for hadron-nucleus collisions and correctly
634: predicted the effect for RHIC \cite{knst}. Therefore we believe that a
635: calculated value of $R_8$ should be rather reliable. Doing this,
636: Eq.~(\ref{60}) leads to somewhat larger values for the broadening.
637:
638: The fit of Eq.~(\ref{25}) to this data leads to new values of parameter $D$,
639: \beqn
640: \tilde D(E772)=(0.038 \pm 0.008)\,(\GeV/c)^2\ ;
641: \nonumber\\
642: \tilde D(E866)=(0.069 \pm 0.009)\,(\GeV/c)^2\ .
643: \label{120}
644: \eeqn
645:
646: This procedure involving theoretical calculations may be considered as a
647: means to estimate the systematic uncertainty. The sign (positive) we
648: obtain for the uncertainty is certainly correct, and its magnitude is not
649: large. Note also that these corrections are correlated for the two sets of
650: data, and therefore one must apply them consistently to both the the
651: results of the E772 and E866 experiments.
652:
653: \section{Broadening versus coherence length}
654:
655:
656: The mechanism of broadening is expected to correlate with the coherence
657: length \cite{jkt,knst,krtj} as described above. The coherence length of
658: the Drell-Yan process is controlled by the value of $x_2$, $l_c\sim 1/m_N
659: x_2$ \cite{e-loss}. Thus, at large $x_2$ the dilepton is produced inside
660: the nucleus and only half of the nuclear thickness contributes to
661: broadening. However, at small $x_2$ the Drell-Yan pair is radiated outside
662: the nucleus and the entire nuclear thickness is effective. It would be a
663: spectacular observation to see a manifestation of this effect in data.
664:
665: Using recent data from the E866 experiment for the $p_T$-dependent ratios
666: binned in $x_2$ \cite{e866-pt,e866-x2} we can also test the predicted
667: $x_2$-independence of broadening. We have performed a similar analysis of
668: the $Fe/Be$ and $W/Be$ ratios in different $x_2$ bins and the results are
669: depicted in Fig.~\ref{x2-dep}
670: \begin{figure}[tbh]
671: \centerline{\psfig{figure=x2-dep-1.ps,width=8cm}}
672: \protect\caption{Broadening of the mean transverse momentum squared in
673: the Drell-Yan reaction measured \cite{e866-pt,e866-x2} on tungsten and
674: iron targets relative to beryllium, $\Delta\la p_T^2\ra = \la
675: p_T^2\ra_A-\la p_T^2\ra_{Be}$.}
676: \label{x2-dep}
677: \end{figure}
678:
679: Unfortunately, within the rather large statistical errors one cannot draw
680: a definite conclusion regarding a possible variation of broadening with
681: $x_2$.
682:
683: \section{Conclusions}
684:
685: Analysis of data from the E866 experiment for the DY reaction on nuclear
686: targets has resulted in a determination of the magnitude of the nuclear
687: broadening parameter $D(E866)$ given in (\ref{100}) and (\ref{120}), which
688: is about twice as large as the one suggested by the E772 data. This result
689: helps to resolve the controversies that were a partial motivation for this
690: analysis. One of these was the expectation that the observed broadening in
691: DY would be a factor of $4/3$ less than that observed in $J/\Psi$ and
692: $\Upsilon$, as dictated by the Casimir factor. The updated result is
693: consistent with this expectation and also with the expectation of a rising
694: energy dependence of broadening. And, last but not least, the value of
695: $D(E866)$ agrees with the parameter-free theoretical prediction of
696: \cite{jkt}.
697:
698: This new result also agrees with broadening observed by the HERMES
699: experiment at HERA in hadron production in deep-inelastic scattering on
700: nuclei \cite{hermes}. Indeed, both measurements are well described
701: within the same color-dipole approach \cite{with-with}.
702:
703: Since most of the data from both experiments E772 and E866 correspond to
704: the regime of short coherence length, the observed broadening is directly
705: connected via the relation in Eq.~(\ref{35}) to the broadening of a quark
706: propagating through the nuclear medium. In the regime of long coherence
707: length the mechanism of broadening changes and takes the form of color
708: filtering in hard coherent scatterings. To see a possible variation of
709: broadening as function of coherence length, we performed an analysis of
710: data additionally binned in $x_2$. Unfortunately, the restricted
711: statistics did not allow us to draw a definite conclusion from this
712: analysis.
713:
714: Broadening of the transverse momentum of a parton in a medium leads to
715: induced gluon radiation, i.e., to induced energy loss. This is the basis
716: of the effect called jet quenching which is considered to be a sensitive
717: probe for properties of matter created in heavy ion collisions.
718:
719: Note that broadening itself is much less dependent on models for
720: hadronization, than jet quenching. Therefore it is a better and more
721: certain probe.
722:
723:
724: \vspace{2mm}
725:
726: \noindent {\bf Acknowledgments}: This work was supported in part by U.S.
727: Department of Energy, the Research Ring "Center of Subatomic Studies"
728: (Chile), Fondecyt (Chile) grants 1030355, 1050519 and 1050589, and by DFG
729: (Germany) grant PI182/3-1.
730:
731: \begin{thebibliography}{99}
732:
733: \bibitem{na10} NA10 Collaboration, P.~Bordalo et al., Phys. Lett. B{\bf
734: 193}, 373 (1987).
735:
736: \bibitem{e772} E772 Collaboration, E772 Collaboration, D.M.~Alde et al.,
737: Phys. Rev. Lett. {\bf 66}, 2285 (1991).
738:
739: \bibitem{mmp} P.L.~McGaughey, J.M.~Moss, and J.-Ch.~Peng, NUCOLEX 99,
740: Wako, Japan, 1999 (hep-ph/9905447); Ann. Rev. Nucl. Part. Sci. {\bf 49},
741: 217 (1999).
742:
743: \bibitem{hermes} HERMES Collaboration, A. Airapetian et al., Eur. Phys. J.
744: C{\bf 20}, 479 (2001).
745:
746: \bibitem{jlab} W.~Brooks, talk presented at the Workshop "Parton
747: Propagation through Strongly Interacting Systems", ECT$^*$, Trento,
748: 26 September - 7 October, 2005.
749:
750: \bibitem{pirner} J. Hufner, Y. Kurihara, H.J. Pirner, Phys. Lett.
751: {\bf B215}, 218 (1988).
752:
753: \bibitem{dhk} J.~Dolejsi, J.~H\"ufner and B.Z.~Kopeliovich,
754: Phys. Lett. {\bf B312}, 235 (1993).
755:
756: \bibitem{jkt} M.B.~Johnson, B.Z.~Kopeliovich and A.V.~Tarasov, Phys.\
757: Rev.\ C {\bf 63}, 035203 (2001).
758:
759: \bibitem{at} A.~Accardi and D.~Treleani, Phys. Rev. {\bf D64}, 116004 (2001).
760:
761: \bibitem{zkl} B.Z.~Kopeliovich, L.I.~Lapidus, and A.B.~Zamolodchikov, Sov.
762: Phys. JETP Lett. {\bf 33}, 595 (1981); Pisma v Zh. Exper. Teor. Fiz. {\bf
763: 33}, 612 (1981).
764:
765: \bibitem{gbw} K.~Golec-Biernat and M.~W\"usthoff, Phys. Rev. D{\bf 59}
766: 014017(1999).
767:
768: \bibitem{kst2} B.Z.~Kopeliovich, A.~Sch\"afer and A.V.~Tarasov, Phys. Rev.
769: {\bf D62} 054022 (2000).
770:
771: \bibitem{bdmps} R.~Baier, Yu.L.~Dokshitzer, A.H.~Mueller, S. Peigne and
772: D.~Schiff, Nucl. Phys. B {\bf 484}, 265 (1997).
773:
774: \bibitem{vitev} I.~Vitev, Phys. Lett. {B562}, (2003) 36.
775:
776: \bibitem{e-loss} M.B.~Johnson {\it et al.} [FNAL E772 Collaboration],
777: Phys.\ Rev.\ Lett.\ {\bf 86}, 4483 (2001); Phys.Rev. C{\bf 65}, 025203
778: (2002).
779:
780: \bibitem{qs} M.~Luo, J.W.~Qiu, and G.~Sterman, Phys. Rev. {\bf D50}, 1951
781: (1994).
782:
783: \bibitem{guo} X.~Guo, J.~Qiu, X.~Zhang, Phys. Rev. Lett. {\bf 84} 5049
784: (2000).
785:
786: \bibitem{ww} E.~Wang and X.-N.~Wang, Phys. Rev. Lett. {\bf 89} 162301
787: (2002).
788:
789: \bibitem{cronin} D.~Antreasyan et al., Phys. Rev. D {\bf 19}, 764 (1979).
790:
791: \bibitem{krtj} B.~Z.~Kopeliovich, J.~Raufeisen, A.~V.~Tarasov and
792: M.~B.~Johnson, Phys.Rev. C{\bf 67}, 014903 (2003).
793:
794: \bibitem{knst} B.Z.~Kopeliovich, J.~Nemchik, A.~Sch\"afer and
795: A.V.~Tarasov, Phys. Rev. Lett. {\bf 88}, 232303 (2002).
796:
797: \bibitem{e772a} E772 Collaboration, D.M.~Alde et al., Phys. Rev. Lett.
798: {\bf 64}, 2479 (1990); M.~Leitch et all., Nucl. Phys. A{\bf 522}, 351c
799: (1991).
800:
801: \bibitem{e789} E789 Collaboration, M.J. Leitch et al., Phys. Rev. D{\bf
802: 52}, 4251 (1995).
803:
804: \bibitem{E866-psi} E866 Collaboration, M.J. Leitch et al., Phys. Rev.
805: Lett. {\bf 84}, 3256 (2000).
806:
807: \bibitem{webb} E866 Collaboration, J.C. Webb et al., hep-ex/0302019, to
808: appear in Phys. Rev. Lett.
809:
810: \bibitem{mikk} M.B.~Johnson, B.Z.~Kopeliovich, and I.~Schmidt,
811: hep-ph/0701015.
812:
813: \bibitem{krt2} B.Z.~Kopeliovich, J.~Raufeisen, and A.V.~Tarasov, Phys.
814: Rev. C{\bf 62}, 035204 (2000).
815:
816: \bibitem{glr} L.V.~Gribov, E.M.~Levin and M.G.~Ryskin, Nucl.
817: Phys. B {\bf 188}, 555 (1981); Phys. Rep. {\bf 100},1 (1983).
818:
819: \bibitem{mv} L.~McLerran and R.~Venugopalan, Phys. Rev. D {\bf 49}, 2233
820: (1994); D {\bf 49}, 3352 (1994).
821:
822: \bibitem{dy} S.~D.~Drell and T.~Yan, Phys.\ Rev.\ Lett.\ {\bf 25}, 316
823: (1970) [Erratum-ibid.\ {\bf 25}, 902 (1970)].
824:
825: \bibitem{k} B.\ Z.\ Kopeliovich, proc.\ of the workshop Hirschegg '95:
826: Dynamical Properties of Hadrons in Nuclear Matter, Hirschegg January
827: 16-21, 1995, ed.\ by H.\ Feldmeyer and W.\ N\"orenberg, Darmstadt, 1995,
828: p.\ 102 (hep-ph/9609385).
829:
830: \bibitem{kst1} B.Z.~Kopeliovich, A.~Sch\"afer and A.V.~Tarasov, Phys.\
831: Rev.\ C {\bf 59}, 1609 (1999) (see extended version in hep-ph/9808378).
832:
833: \bibitem{bbl} G.T.~Bodwin, S.J.~Brodsky, and G.P.~Lepage, Phys. Rev. D
834: {\bf 39}, 3287 (1989).
835:
836: \bibitem{e866-pt} M.A. Vasiliev et al., Phys. Rev. Lett. {\bf 83},
837: 2304 (1999).
838:
839: \bibitem{e866-x2} See plots and tables in the E866 web page,
840: http://p25ext.lanl.gov/e866/papers/papers.html
841:
842: \bibitem{with-with} B.Z.~Kopeliovich, J.~Nemchik, E.~Predazzi,
843: A.~Hayashigaki, Nucl. Phys. A{\bf 740}, 211 (2004).
844:
845:
846: \end{thebibliography}
847:
848: \end{document}
849: