1: % V1 19 NOV 05
2: % V2 16 MAR 06
3: % V3 24 APR 06
4: % V4 11 AUG 06
5: % V5 22 AUG 06
6: %
7: %
8: % REVTEX 4
9: %
10: \documentclass[prd,superscriptaddress,twocolumn,showpacs,amsmath,%
11: preprintnumbers]{revtex4}
12: %
13: \usepackage{bm}
14: \usepackage{graphicx}
15: %
16: % GENERAL DEFINITIONS
17: %
18: \newcommand{\beq}{\begin{equation}}
19: \newcommand{\eeq}{\end{equation}}
20: \newcommand{\be}{\begin{eqnarray}}
21: \newcommand{\ee}{\end{eqnarray}}
22: %
23: \begin{document}
24: %
25: \author{L.~Frankfurt}
26: \affiliation{School of Physics and Astronomy, Tel Aviv University,
27: Tel Aviv, Israel}
28: %
29: \author{C.E.~Hyde--Wright}
30: \affiliation{Old Dominion University, Norfolk, VA 23529, USA}
31: %
32: \author{M.~Strikman}
33: \affiliation{Department of Physics, Pennsylvania State University,
34: University Park, PA 16802, USA}
35: %
36: \author{C.~Weiss}
37: \affiliation{Theory Center, Jefferson Lab, Newport News, VA 23606, USA}
38: %
39: \title{Generalized parton distributions and rapidity gap survival \\
40: in exclusive diffractive $pp$ scattering}
41: %
42: \begin{abstract}
43: We propose a new approach to the problem of rapidity gap survival (RGS)
44: in the production of high--mass systems ($H$ = dijet, heavy quarkonium,
45: Higgs boson) in double--gap exclusive diffractive $pp$ scattering,
46: $pp \rightarrow p + \text{(gap)} + H + \text{(gap)} + p$.
47: It is based on the idea that hard and soft interactions proceed over
48: widely different time-- and distance scales and are thus approximately
49: independent. The high--mass system is produced in a hard scattering
50: process with exchange of two gluons between the protons. Its amplitude
51: is calculable in terms of the gluon generalized parton distributions (GPDs)
52: in the protons, which can be measured in $J/\psi$ production in exclusive
53: $ep$ scattering. The hard scattering process is modified by soft spectator
54: interactions, which we calculate in a model--independent way in terms of
55: the $pp$ elastic scattering amplitude. Contributions from inelastic
56: intermediate states are suppressed. A simple geometric picture of the
57: interplay of hard and soft interactions in diffraction is obtained.
58: The onset of the black--disk
59: limit in $pp$ scattering at TeV energies strongly suppresses diffraction
60: at small impact parameters and is the main factor in determining the
61: RGS probability. Correlations between hard and soft interactions
62: (\textit{e.g.}\ due to scattering from the long--range pion field of
63: the proton, or due to possible short--range transverse correlations
64: between partons) further decrease the RGS probability.
65: We also investigate the dependence of the diffractive cross section on the
66: transverse momenta of the final--state protons (``diffraction pattern'').
67: By measuring this dependence one can perform detailed tests of the
68: interplay of hard and soft interactions, and even extract information
69: about the gluon GPD in the proton. Such studies appear to be feasible
70: with the planned forward detectors at the LHC.
71: \end{abstract}
72: %
73: \keywords{Quantum chromodynamics, diffraction, generalized
74: parton distributions, Higgs boson search}
75: %
76: \pacs{12.38.-t, 13.85.-t, 13.85.Dz, 14.80.Bn}
77: %
78: \preprint{JLAB-THY-06-533}
79: %
80: \maketitle
81: %
82: \section{Introduction}
83: Hard processes in high--energy $pp$ scattering are important both
84: as a laboratory for studying strong interaction dynamics and the
85: parton structure of the proton, and as one of the main tools
86: in the search for new heavy particles. Of particular interest are
87: so--called diffractive processes, in which the produced high--mass
88: system (dijet, heavy particle) is separated from the projectile
89: fragments by large rapidity gaps. Double--gap exclusive processes
90: (\textit{i.e.}, without breakup of the protons)
91: \beq
92: pp \;\; \rightarrow \;\; p + \text{(gap)} + H + \text{(gap)} + p,
93: \label{exclusive_diffraction}
94: \eeq
95: are considered as an option for the Higgs boson search at the LHC
96: \cite{Schafer:1990fz,Bialas:1991wj,Dokshitzer:1991he,%
97: Khoze:2000wk,Kaidalov:2003fw}.
98: Such processes have lower cross section than inclusive
99: double--gap processes (with breakup of one or both protons),
100: but offer better chances for detection, and for
101: determining the mass of the produced particle and possibly even
102: its quantum numbers; see Ref.~\cite{Kaidalov:2003fw} and
103: references therein. Double--gap exclusive processes
104: (\ref{exclusive_diffraction}) also appear to be an effective method
105: for producing heavy quarkonia and investigating their properties.
106:
107: From the point of view of strong interactions, double--gap exclusive
108: events (\ref{exclusive_diffraction}) arise as the result of
109: an interesting interplay of ``hard'' (involving momentum transfers
110: much larger than the typical hadronic mass scale)
111: and ``soft'' (momentum transfers of the order of the typical hadronic
112: mass scale) interactions. The high--mass system is produced in a
113: hard scattering process, involving the exchange of two gluons
114: between the protons. The requirement of the absence of QCD radiation
115: ensures the localization of this process in space and time.
116: This alone, however, is not sufficient to guarantee a diffractive event.
117: One must also require that the soft interactions between
118: the spectator systems do not lead to particle production.
119: This results in a suppression of the cross section as compared to
120: the hard scattering process alone, the so--called
121: rapidity gap survival (RGS) probability. While not directly
122: observable, this quantity plays a central role in the discussion
123: of hard diffractive processes and their use in new particle searches.
124:
125: Diffractive final states are most favorably produced in scattering
126: at large impact parameters (peripheral scattering), where the chances
127: for the spectator systems not to interact inelastically are large.
128: Conversely, this means that the selection of diffractive events
129: changes the effective impact parameters as compared to inclusive
130: events with the same hard scattering process. This effect is
131: essential for understanding the physical mechanism of RGS. In this sense,
132: RGS is a manifestation of a general quantum--mechanical
133: phenomenon --- the postselection of certain initial--state configurations
134: by conditions imposed on the final state \cite{Nussinov}.
135:
136: The concept of RGS should also be viewed in the context of QCD
137: factorization for the production of heavy particles in
138: $pp$ scattering \cite{Bjorken:1992er}. QCD factorization
139: was formally proved for inclusive scattering, $pp \rightarrow H + X$.
140: A crucial element in this is the cancellation of initial--state
141: and final--state QCD radiation. This cancellation becomes incomplete
142: if additional conditions, such as rapidity gaps, are imposed
143: on the hadronic final state. The introduction of the RGS probability
144: can be seen as an attempt to ``restore'' QCD factorization for
145: diffractive processes at the phenomenological level,
146: in a form analogous to the inclusive case.
147:
148: In this paper, we propose a new approach to RGS in double--gap exclusive
149: hard diffractive processes. It is based on the idea that hard and soft
150: interactions are approximately independent because they happen over widely
151: different time-- and distance scales. We implement this idea in the
152: framework of a partonic description of the proton, along the lines
153: of Gribov's parton picture of high--energy hadron--hadron
154: scattering \cite{Gribov:1973jg}. In the approximation where hard
155: and soft interactions are considered to be completely independent,
156: one can regard the hard scattering process as a ``local operator''
157: in partonic states, and RGS appears as the ``renormalization''
158: of this operator due to soft interactions. At the amplitude level,
159: this leads to an absorption correction to the
160: hard production process due to elastic rescattering;
161: contributions from inelastic intermediate states are suppressed
162: because of the different character of the states accessible in hard and
163: soft interactions. At the cross section level, we recover
164: a simple ``geometric'' expression for the RGS probability, which was
165: suggested on the basis of heuristic arguments in
166: Refs.~\cite{Frankfurt:2004kn,Frankfurt:2004ti,Frankfurt:2005mc}.
167: In addition to providing a transparent physical picture of RGS,
168: this expression can readily be evaluated in terms of two phenomenological
169: ingredients, both of which can be probed in independent measurements:
170: %
171: \begin{itemize}
172: %
173: \item
174: The gluon generalized parton distribution (GPD) in the proton;
175: more precisely its $t$--dependence (``two--gluon formfactor''),
176: whose Fourier transform describes the transverse spatial
177: distribution of gluons. Information about it
178: comes from measurements of hard exclusive processes in $ep$ scattering,
179: in particular $J/\psi$ photoproduction (HERA, FNAL).
180: %
181: \item
182: The $pp$ elastic scattering amplitude at high energies;
183: in particular its profile function in the impact parameter
184: representation. It is known from fits to $pp/\bar pp$ total and
185: elastic cross section data up to the Tevatron energy, and constrained
186: by general arguments (black--disk limit, see below)
187: \end{itemize}
188: %
189:
190: The framework provided by our partonic approach to RGS allows us
191: to take into account two basic facts about the dynamics of hard
192: and soft interactions at high energies, which turn out to have a
193: decisive influence on the numerical value of the RGS probability.
194: These are:
195: %
196: \begin{itemize}
197: %
198: \item \textit{Small transverse radius of hard interactions.}
199: The radius of the transverse distribution of hard gluons in the proton
200: is significantly smaller than the transverse radius of soft interactions
201: in high--energy $pp$ collisions (``two--scale picture''). This
202: basic fact explains many qualitative features of hard exclusive
203: diffractive processes, such as the effective impact parameters
204: in diffractive events, the order--of--magnitude of the RGS probability,
205: and the pattern of the transverse momentum dependence of the cross section.
206: %
207: \item \textit{Black--disk limit in high--energy $pp$ scattering.}
208: Parametrizations of the data as well as general theoretical arguments
209: indicate that the profile function of $pp$ elastic scattering becomes
210: ``black'' at small impact parameters at energies above the
211: Tevatron energy, $\sqrt{s} > 2\, \text{TeV}$. This circumstance
212: makes the description of $pp$ scattering at small impact parameters at
213: LHC energies practically model--independent, and is essential for the
214: stability of numerical predictions for the RGS.
215: %
216: \end{itemize}
217: The evidence supporting these statements is described in detail in
218: Ref.~\cite{Frankfurt:2005mc}, and summarized in
219: Sections~\ref{sec:transverse_soft} and \ref{sec:transverse_hard} below.
220:
221: Our partonic approach also allows us to describe the effects of correlations
222: between hard and soft interactions on the RGS probability. Such
223: correlations can arise from various dynamical mechanisms,
224: \textit{e.g.}\ from the long--range pion field of the proton,
225: or from possible short--range transverse correlations between
226: hard partons, as suggested by the Tevatron CDF data on inclusive $pp$
227: scattering with multiple hard processes \cite{Abe:1997bp}.
228: We find that the inclusion of such correlations decreases the RGS
229: probability compared to the independent interaction approximation.
230: While these effects cannot be calculated in a completely
231: model--independent way, they are important both for our general
232: understanding of the mechanism of RGS, and for obtaining reliable
233: numerical estimates of the RGS probability.
234: These effects clearly merit further study.
235:
236: A unique feature of exclusive diffractive processes is that the
237: interplay of hard and soft interactions can be studied experimentally,
238: by measuring the dependence of the cross section on the transverse
239: momenta of the final--state protons. The modification of the hard
240: scattering amplitude by soft elastic rescattering can be viewed
241: as an interference phenomenon, which gives rise to a distinctive
242: ``diffraction pattern'' in the final--state transverse momenta.
243: By measuring this dependence in exclusive diffractive processes
244: with relatively large cross section, such as dijet production,
245: one can perform a variety of tests of the diffractive reaction mechanism,
246: and extract information about the transverse radii of hard and soft
247: interactions and their energy dependence. In Higgs production, measurements
248: of the transverse momentum dependence would allow one to obtain additional
249: information about the parity of the produced particle \cite{Kaidalov:2003fw}.
250: Experimentally, such studies appear to be feasible with the planned forward
251: detectors at the LHC \cite{Albrow:2005ig} and the
252: Tevatron \cite{Albrow:2005fw}.
253:
254: Detailed studies of RGS in diffractive $pp$ scattering were done
255: in a model of soft interactions based on eikonalized Pomeron
256: exchange \cite{Khoze:2000wk,Kaidalov:2003fw}. We show here
257: that, in the approximation where hard and soft interactions are
258: considered to be independent, the RGS probability unambiguously
259: follows from QCD and can be calculated in a model--independent way.
260: Nevertheless, our numerical results for the RGS probability in this
261: approximation turn out to be of the same order of magnitude as those
262: reported in Refs.~\cite{Khoze:2000wk,Kaidalov:2003fw}, which can be
263: attributed to the fact that the Pomeron parametrization of the $pp$
264: elastic amplitude reproduces the approach to the BDL at small impact
265: parameters. The agreement between our numerical results and
266: theirs at the quantitative level is somewhat accidental, being due to the
267: fact that in the calculation of Refs.~\cite{Khoze:2000wk,Kaidalov:2003fw}
268: the effect from inelastic intermediate states (which in our approach
269: are seen to be strongly suppressed because of the small overlap of states
270: accessible via hard and soft interactions) is partly compensated by
271: the choice of a larger value of the $t$--slope of the gluon GPD; see
272: Section~\ref{sec:cross_section} for details. Finally, our partonic
273: approach can naturally be extended to include correlations between
274: hard and soft interactions, which have a potentially large numerical
275: effect on the RGS probability.
276:
277: This paper is organized as follows. In Section~\ref{sec:transverse_soft}
278: we review the information about the transverse structure of soft interactions
279: from $pp$ elastic scattering, and the approach to the black--disk
280: limit at central impact parameters.
281: In Section~\ref{sec:transverse_hard} we discuss the properties
282: of the proton's gluon GPD at small $x$ and summarize our knowledge
283: of the transverse spatial distribution of hard partons in the proton.
284: Section~\ref{sec:gap_survival} describes the basic framework of our
285: approach to RGS. We outline the properties of the hard scattering
286: amplitude, describe the theoretical formulation of the independence
287: of hard and soft interactions, and obtain a master expression for
288: the diffractive amplitude combining hard and soft interactions.
289: We then explain the suppression of inelastic intermediate states,
290: and evaluate the diffractive amplitude in terms of the gluon GPD
291: and the $pp$ elastic amplitude.
292: In Section~\ref{sec:cross_section} we use our result for the
293: amplitude in the independent interaction approximation
294: to calculate the RGS probability. We recover a simple geometric
295: expression for the RGS probability and discuss the effective impact
296: parameters in exclusive diffraction. We then evaluate the RGS probability
297: numerically, estimate the uncertainty of the numerical predictions
298: due to the phenomenological input, and emphasize the crucial role of the
299: black--disk limit in stabilizing the numerical predictions.
300: We also comment on the results for the RGS obtained within
301: the eikonalized Pomeron model for soft interactions \cite{Khoze:2000wk}
302: from the perspective of our approach. In Section~\ref{sec:beyond}
303: we discuss various effects beyond the approximation of independence
304: hard and soft interactions in exclusive diffraction.
305: We first point out that hard screening corrections may reduce
306: the diffractive cross section beyond the RGS probability due to soft
307: interactions. We then discuss the effect of correlations between
308: hard and soft interactions on the RGS probability, considering
309: two specific mechanisms --- diffractive scattering from the long--range
310: pion field, and short--range transverse correlations between partons.
311: In Section~\ref{sec:differential} we work out the dependence of the
312: exclusive diffractive cross section on the final proton transverse momenta.
313: We discuss which experimentally observable features of this dependence
314: furnish useful tests of the diffractive reaction mechanism,
315: and how one can extract information about the gluon GPD.
316: In Section~\ref{sec:discussion} we summarize our results.
317: We comment on the implications for the Higgs boson search, and
318: on the experimental feasibility of measuring the transverse momentum
319: dependence of exclusive diffraction with the planned forward
320: detectors at the LHC.
321: %
322: \section{Black--disk limit in $pp$ elastic scattering}
323: \label{sec:transverse_soft}
324: Information on the transverse radius of strong interactions at
325: high energies comes mostly from measurements of the $t$--dependence
326: of the differential cross section for $pp$ and $\bar p p$ elastic scattering.
327: Combining these data with those on the $pp/\bar pp$ total cross section,
328: and implementing theoretical constraints following from the unitarity
329: of the $S$--matrix, one can reconstruct the complex $pp$ elastic scattering
330: amplitude, $T_{\text{el}}(s, t)$; see \textit{e.g.}\
331: Refs.~\cite{Block:1998hu,Bourrely:2002wr,Islam:2002au}.
332: At high energies, $s \gg |t|$, angular momentum conservation in the CM
333: frame implies that the scattering
334: amplitude is effectively diagonal in the impact parameter of the
335: colliding $pp$ system. It is convenient to represent it as a Fourier
336: integral over a transverse coordinate variable, $\bm{b}$,
337: \be
338: T_{\text{el}} (s, t = -\bm{\Delta}_\perp^2)
339: &=& \frac{i s}{4\pi} \int d^2 b \;
340: e^{-i (\bm{\Delta}_\perp \bm{b})}
341: \; \Gamma (s, \bm{b}) ,
342: \label{Gamma_def}
343: \ee
344: where $\Gamma$ is the (dimensionless) profile function.
345: One can then express the elastic, total, and inelastic
346: (total minus elastic) $pp$ cross sections in terms of the
347: profile function as
348: \begin{equation}
349: \left. \begin{array}{l}
350: \sigma_{\text{el}}(s) \\[1ex]
351: \sigma_{\text{tot}}(s) \\[1ex]
352: \sigma_{\text{inel}}(s)
353: \end{array}
354: \right\}
355: \;\; = \;\; \int d^2 b \; \times
356: \left\{ \begin{array}{l}
357: |\Gamma (s,\bm{b})|^2 , \\[1ex]
358: 2 \, \text{Re} \, \Gamma (s,\bm{b}) , \\[1ex]
359: \left[ 1 - |1-\Gamma (s,\bm{b})|^2 \right] .
360: \end{array}
361: \right.
362: \label{unitarity}
363: \end{equation}
364: The functions on the R.H.S.\ describe the distribution of the respective
365: cross sections over $pp$ impact parameters, $b \equiv |\bm{b}|$
366: \footnote{For $pp$ scattering without transverse polarization effects,
367: which we consider here,
368: the profile function depends only on the modulus of the transverse
369: coordinate, $b = |\bm{b}|$. We nevertheless regard it as a function
370: of the vector variable, $\bm{b}$, to facilitate later usage when we
371: consider convolutions in the transverse coordinate variable.}.
372: In particular, we note that the combination
373: \begin{equation}
374: |1 - \Gamma (s, \bm{b})|^2
375: \label{P_noin}
376: \end{equation}
377: can be interpreted as the probability for ``no inelastic interaction''
378: in a $pp$ collision at impact parameter $b$; this combination plays
379: an important role in our calculation of the RGS probability (see
380: Section~\ref{sec:cross_section} below) \footnote{Notice
381: that $1 - \Gamma(s, \bm{b})$ can be interpreted as the impact
382: parameter representation of the $S$--matrix element, which is related
383: to the scattering amplitude by Eqs.~(\ref{S_minus_1_me}) and
384: (\ref{S_minus_1_t}).}.
385: A measure of the transverse size of the proton is the logarithmic
386: $t$--slope of the elastic $pp$ cross section at $t = 0$,
387: \beq
388: B \;\; \equiv \;\; \frac{d}{dt} \left[
389: \frac{d\sigma_{\text{el}} /dt \, (t)}{d\sigma_{\text{el}} /dt \, (0)}
390: \right]_{t = 0}.
391: \label{B_zero}
392: \eeq
393: At high energies, where the elastic amplitude is predominantly
394: imaginary, and $\Gamma$ is real, $B$ is equal to
395: half the average squared impact parameter in the total
396: $pp$ cross section,
397: \beq
398: B \;\; \approx \;\;
399: \frac{\langle b^2 \rangle_{\text{tot}}}{2} \;
400: \;\; \equiv \;\; \frac{1}{2} \frac{\int d^2 b \, b^2 \,
401: 2 \, \text{Re} \, \Gamma (s,\bm{b})}
402: {\int d^2 b \; 2 \, \text{Re} \, \Gamma (s,\bm{b})} ,
403: \eeq
404: which may be associated with the transverse area of the
405: individual protons. The data show that the slope increases
406: with the CM energy as
407: \be
408: B(s) \;\; =\;\; B(s_0) \; + \; 2 \alpha^{\prime}\; \ln (s/s_0),
409: \label{alpha_prime}
410: \ee
411: where $\alpha^{\prime} \approx 0.25 \, \text{GeV}^{-2}$.
412: In the Pomeron exchange parametrization of the $pp$ elastic amplitude
413: this constant is identified with the slope of the Pomeron trajectory.
414:
415: In Gribov's parton picture of high--energy hadron--hadron
416: interactions \cite{Gribov:1973jg}, the transverse size of the proton
417: in $pp$ elastic scattering can be directly associated with the average
418: transverse radius squared of the distribution of soft partons
419: mediating the soft interactions,
420: \beq
421: B \;\; = \;\; \langle \rho^2 \rangle_{\text{soft}}.
422: \eeq
423: Here and in the following, we use $\rho \equiv |\bm{\rho}|$ to denote
424: the transverse distance of partons from the center of the proton,
425: and $b = |\bm{b}|$ for the impact parameter of the $pp$ collision.
426: The growth of the proton's transverse size with energy is explained
427: as the result of random transverse displacements in the successive
428: decays generating the distribution of soft partons (Gribov diffusion).
429: Below we shall compare this distribution of soft partons to the
430: distribution of hard partons probed in hard exclusive processes
431: (see Section~\ref{sec:transverse_hard}).
432:
433: Parametrizations of the available data indicate that at energies above
434: the Tevatron energy,
435: $\sqrt{s} \gtrsim \sqrt{s_{\text{Tevatron}}} = 2\, \text{TeV}$,
436: the profile function at small impact parameters approaches
437: \beq
438: \Gamma (s, b) \;\; \rightarrow \;\; 1
439: \hspace{3em} \text{for} \;\; b < b_0 (s) .
440: \eeq
441: This corresponds to unit probability for inelastic scattering
442: for impact parameters $b < b_0 (s)$, \textit{cf.}\ Eqs.~(\ref{unitarity})
443: and (\ref{P_noin}), similar to the scattering of a pointlike object from a
444: black disk of radius $b_0$, and is referred to as the black--disk limit (BDL)
445: \cite{Frankfurt:2003td,Frankfurt:2004ti,Frankfurt:2005mc}.
446:
447: The approach to the BDL in central $pp$ scattering at high energies
448: is a general prediction of QCD, independent of detailed assumptions
449: about the dynamics. Studies of the interaction of small--size color
450: dipoles with hadrons, based on QCD factorization in the leading
451: $\log Q^2$ approximation, show that the BDL is attained at high energies
452: as a result of the growth of the gluon
453: density at small $x$ due to DGLAP evolution \cite{Frankfurt:2003td}.
454: This result can be used to estimate the interaction of
455: leading projectile partons with the small--$x$ gluons in the target
456: in $pp$ scattering; one finds that there is no chance for the
457: projectile wavefunction to remain coherent in small impact parameter
458: scattering at TeV energies \cite{Frankfurt:2004ti,Frankfurt:2005mc}.
459: Similar reasoning allows one to predict the growth of the size of the
460: black region, $b_0$, with $s$ \cite{Frankfurt:2004ti,Frankfurt:2005mc}.
461: As a by--product, these arguments explain why the observed
462: coefficient in the Froissart formula for the total cross sections is
463: significantly smaller than that derived from the general principles
464: of analyticity of the amplitude in momentum transfer and unitarity of the
465: $S$--matrix \cite{Martin:1962rt}. We note that the need for the approach
466: to the BDL in high--energy scattering at central impact parameters was
467: understood already in the pre--QCD period within the Pomeron calculus,
468: where it was noted that this phenomenon resolves the apparent contradiction
469: between the formulae of the triple--Pomeron limit and the unitarity of the
470: $S$--matrix, especially in models where the Pomeron intercept,
471: $\alpha_P(0)$, exceeds unity \cite{Marchesini:1976hw}.
472:
473: %
474: % FIGURE
475: %
476: \begin{figure}
477: %
478: \includegraphics[width=8cm]{comp_wnoin.eps}
479: %
480: \caption[]{The probability distribution for no inelastic
481: interaction, $|1 - \Gamma (s, \bm{b})|^2$, Eq.~(\ref{P_noin}),
482: as a function of $b \equiv |\bm{b}|$,
483: at the LHC energy ($\sqrt{s} = 14 \, \text{TeV}$),
484: as computed with different parametrizations of the
485: $pp$ elastic scattering amplitude. Solid line: Parametrization of
486: Islam \textit{et al.}\ \cite{Islam:2002au} (``diffractive part'' only).
487: Dashed line: Exponential parametrization, Eq.~(\ref{Gamma_gaussian}),
488: with $\Gamma_0 = 1$ (BDL), \textit{cf.}\ Eq.~(\ref{Gamma_0_BDL}), and
489: $B = 21.8 \, \text{GeV}^{-2}$.}
490: \label{fig:gamma}
491: \end{figure}
492: %
493: For our studies of diffractive $pp$ scattering it will be useful
494: to have a simple analytic parametrization of the $pp$ elastic amplitude
495: at the LHC energy, which incorporates the approach to the BDL at small
496: impact parameters. The $t$--dependence of the $pp$ elastic scattering
497: cross section for $|t| \lesssim 1 \, \text{GeV}^2$ over the measured
498: energy range is reasonably described by an exponential shape,
499: \beq
500: \frac{d\sigma_{\text{el}}}{dt} \;\; \propto \;\; \exp \left[ B(s) \, t
501: \right] ,
502: \eeq
503: where $B(s)$ represents an effective slope, to be distinguished from the
504: ``exact'' slope at $t = 0$, Eq.~(\ref{B_zero}). A parametrization of the
505: $pp$ elastic amplitude which reproduces this dependence is
506: \beq
507: T_{\text{el}} (s, t) \;\; = \;\; \frac{is}{8\pi} \; \sigma_{\text{tot}}(s)
508: \; \exp \left[ \frac{B(s) \, t}{2} \right] ,
509: \label{A_exponential}
510: \eeq
511: corresponding to
512: \be
513: \Gamma (s, \bm{b}) \;\; = \;\;
514: \Gamma_0 (s) \; \exp \left[ -\frac{\bm{b}^2}{2 B(s)} \right]
515: \label{Gamma_gaussian}
516: \ee
517: with
518: \beq
519: \Gamma_0 (s) \;\; \equiv \;\; \Gamma (s, \bm{b} = 0)
520: \;\; = \;\; \frac{\sigma_{\text{tot}}(s)}{4\pi B(s)} .
521: \eeq
522: Equation~(\ref{A_exponential}) takes into account that the
523: amplitude at high energies is predominantly imaginary, and satisfies
524: the optical theorem for the total cross section,
525: $\sigma_{\text{tot}}(s) = (8\pi /s) \, \text{Im}\, T_{\text{el}} (s, t = 0)$.
526: We may now incorporate the constraint of the BDL at small impact parameters
527: by replacing
528: \beq
529: \Gamma_0 \;\; \rightarrow 1.
530: \label{Gamma_0_BDL}
531: \eeq
532: The value of $B$ we determine by comparing the profile function
533: (\ref{Gamma_gaussian}) with phenomenological parametrizations
534: of the data, extrapolated to the LHC energy, which gives
535: \beq
536: B \;\; \approx \;\;
537: 20 \, \text{GeV}^{-2} \hspace{3em} (\sqrt{s} = 14 \, \text{TeV}) .
538: \label{B_LHC}
539: \eeq
540: In particular, with $B = 21.8 \, \text{GeV}^{-2}$ we obtain excellent
541: agreement with the Regge parametrization of Ref.~\cite{Khoze:2000wk}.
542: Figure~\ref{fig:gamma} shows the probability for no inelastic interaction,
543: $|1 - \Gamma (s, \bm{b})|^2$, Eq.~(\ref{P_noin}), computed with the
544: phenomenological parametrization of Ref.~\cite{Islam:2002au} and
545: our exponential parametrization incorporating the BDL,
546: Eqs.~(\ref{Gamma_gaussian}) and (\ref{Gamma_0_BDL}).
547: One sees that the simple exponential parametrization is a
548: reasonable overall approximation to the phenomenological parametrization
549: over the $b$--range shown in Fig.~\ref{fig:gamma}.
550: %
551: \section{Transverse spatial distribution of gluons}
552: \label{sec:transverse_hard}
553: Information about the transverse structure of hard interactions
554: comes from studies of hard exclusive processes in $ep$ scattering,
555: such as meson electroproduction or virtual Compton scattering.
556: Such processes probe the GPDs in the proton, whose Fourier transform
557: with respect to the transverse momentum transfer to the proton
558: describes the spatial distribution of quarks and gluons
559: in the transverse plane; see Refs.~\cite{Diehl:2003ny,Belitsky:2005qn}
560: for a review. In this section we summarize what is
561: known about the gluon GPD at small $x$ from theoretical
562: considerations, and from measurements of $J/\psi$ photoproduction
563: and other processes at HERA and in fixed--target experiments.
564:
565: The gluon GPD can be formally defined as the transition matrix
566: element of the twist--2 QCD gluon operator between proton states of
567: different momenta, $p$ and $p'$. Physically, it describes the amplitude
568: for a fast--moving proton to ``emit'' and ``absorb'' a gluon with given
569: longitudinal momenta, with transverse momenta (virtualities)
570: integrated over up to some hard scale, $Q^2$, and a certain invariant
571: momentum transfer to the proton, $t \equiv (p' - p)^2$.
572: The choice of longitudinal momentum
573: variables is a matter of convention. Instead of the initial and final
574: gluon momentum fractions (with respect to the initial proton momentum),
575: $x$ and $x'$, we use as independent variables the initial gluon
576: momentum fraction, $x$, and the fractional longitudinal momentum transfer
577: to the proton (``skewness'') \footnote{In most of the recent literature the
578: skewness is defined such that $2\xi = x' - x$. In the present context
579: it is convenient to omit the factor 2 and define the skewness
580: as in Eq.~(\ref{xi_def}).}
581: \beq
582: \xi \;\; \equiv \;\; x - x',
583: \label{xi_def}
584: \eeq
585: and denote the gluon GPD by
586: \beq
587: H_g (x, \xi, t; Q^2) .
588: \eeq
589: In the limit of zero momentum transfer, the gluon GPD reduces to the
590: usual gluon momentum density in the proton \footnote{In the present context
591: it is convenient to define the gluon GPD as corresponding to the gluon
592: momentum density, $x G(x)$, not the number density, $G(x)$, in the limit
593: of zero momentum transfer.}
594: \beq
595: H_g (x, \xi = 0, t = 0; Q^2) \;\; = \;\; x G(x, Q^2 ) .
596: \label{H_g_forward}
597: \eeq
598: For discussing the $t$--dependence of the gluon GPD
599: it is convenient to write it in the form
600: \beq
601: H_g (x, \xi, t; Q^2) \;\; = \;\; H_g (x, \xi, t = 0; Q^2) \;
602: F_g (x, \xi, t; Q^2),
603: \label{twogl_def}
604: \eeq
605: where the function $F_g$ is known as the ``two--gluon formfactor'' of
606: the proton and satisfies $F_g (t = 0) = 1$. Note that the two--gluon
607: formfactor still depends on $x$ and $\xi$, \textit{i.e.},
608: Eq.~(\ref{twogl_def}) does not imply naive factorization of the
609: $t$--dependence from that on the partonic variables.
610:
611: The dependence of the gluon GPD on the QCD scale, $Q^2$, is governed
612: by the QCD evolution equations. In applications to production of
613: fixed--mass systems in high--energy $ep$ or $pp$ collisions with
614: $M^2 \ll s$ (such as Higgs boson production at the LHC) we shall be
615: interested in the gluon GPD in the region where
616: \beq
617: x, x' \;\; \ll \;\; 1 ,
618: \eeq
619: while at the same time $Q^2$ is much larger than the typical hadronic
620: mass scale, $Q^2 \gg 1\, \text{GeV}^2$. In this region the gluon GPD
621: can be calculated by applying QCD evolution to a ``primordial''
622: distribution at a low scale, $Q_0^2$, in which one neglects the
623: skewness, $\xi = 0$, or $x = x'$ (diagonal approximation)
624: \cite{Frankfurt:1997ha,Shuvaev:1999ce}. QCD evolution degrades
625: the individual gluon momentum fractions with increasing $Q^2$,
626: while their difference, $\xi = x - x'$, remains
627: fixed by kinematics, being equal to the longitudinal momentum
628: transfer to the proton; as a result, the primordial GPD at the
629: low scale is sampled in the region $|x - x'| \ll x, x'$,
630: where the diagonal approximation is justified. In the diagonal
631: approximation, the GPD at $t = 0$ is completely determined
632: by the usual gluon density, \textit{cf.}\ Eq.~(\ref{H_g_forward}),
633: leaving only the $t$--dependence (two--gluon formfactor) and its
634: correlation with $x$ up to modeling. This approximation
635: makes for a great simplification in applying GPDs at small $x$,
636: and has been used extensively in the analysis of exclusive
637: electroproduction processes; see Ref.~\cite{Frankfurt:2005mc}
638: for a review.
639:
640: For $x \ll 1$, the two--gluon formfactor permits a
641: simple interpretation in terms of a spatial distribution of gluons
642: in the proton. For $\xi \ll 1$, the invariant momentum transfer
643: is dominated by the transverse momentum transfer between the
644: proton states,
645: \beq
646: t \;\; \approx \;\; -\bm{\Delta}_\perp^2,
647: \hspace{3em} \bm{\Delta}_\perp \;\; \equiv \;\;
648: \bm{p}'_\perp - \bm{p}_\perp .
649: \eeq
650: The two--gluon formfactor can be represented as a
651: Fourier integral over a transverse coordinate variable, $\bm{\rho}$,
652: \be
653: F_g (x, \xi, t = -\bm{\Delta}_\perp^2; Q^2) &=&
654: \int d^2 \rho \; e^{-i (\bm{\Delta}_\perp \bm{\rho})}
655: \nonumber \\
656: &\times& F_g (x, \xi, \bm{\rho}; Q^2) .
657: \label{rhoprof_def}
658: \ee
659: For economy of notation we use the same symbol for the
660: two--gluon formfactor and its Fourier transform, distinguishing
661: the two functions by their arguments. The $\bm{\rho}$--dependent
662: function describes the spatial distribution of gluons
663: in the proton in the transverse plane; see Ref.~\cite{Burkardt:2002hr}
664: for a review. For $\xi = 0$ (\textit{i.e.}, $x' = x$) it is positive
665: definite and can be interpreted probabilistically as the gluon density
666: at transverse position $\bm{\rho}$ \cite{Pobylitsa:2002iu};
667: for $\xi \neq 0$ it describes the non-diagonal transition matrix element
668: of the gluon density \cite{Diehl:2002he}. A measure of the gluonic transverse
669: size of the nucleon for given $x$ and $Q^2$ is the average of $\rho^2$,
670: which is proportional to the $t$--slope of the two--gluon formfactor
671: at $t = 0$,
672: \be
673: \langle \rho^2 \rangle_g \, (x, Q^2) &\equiv&
674: \int d^2 \rho \; \rho^2 \; F_g (x, \, \xi = 0, \, \bm{\rho}; \, Q^2 )
675: \\
676: &=&
677: 4 \, \frac{\partial F_g}{\partial t} \, (x, \, \xi = 0, \, t = 0; \, Q^2).
678: \ee
679:
680: The two--gluon formfactor of the nucleon, and hence the transverse
681: spatial distribution of gluons, can directly be extracted from the
682: $t$--dependence of the differential cross section for hard exclusive
683: vector meson production processes probing the gluon GPD.
684: QCD factorization implies that the $t$--dependence of the cross
685: section resides solely in the gluon GPD,
686: \beq
687: \left( \frac{d\sigma}{dt} \right)^{\gamma^* p \rightarrow V p}
688: \;\; \propto \;\; F_g^2 (x, \xi, t; Q^2) ,
689: \eeq
690: up to small higher--twist corrections due to the finite size of the
691: produced vector meson \cite{Frankfurt:2005mc}. In particular, the
692: $t$--slope at $t = 0$ is proportional to the proton's average
693: gluonic transverse size,
694: \beq
695: B_g \;\; \equiv \;\; \frac{d}{dt} \left[
696: \frac{d\sigma/dt \, (t)}{d\sigma/dt \, (0)}
697: \right]^{\gamma^* p \rightarrow V p}_{t = 0}
698: \;\; = \;\; \frac{\langle \rho^2 \rangle_g}{2} .
699: \eeq
700: A crucial test of the
701: applicability of QCD factorization is provided by the observed
702: convergence of the $t$--slopes of various gluon--dominated
703: vector meson production processes ($J/\psi, \rho, \phi$) at large
704: $Q^2$; see Ref.~\cite{Frankfurt:2005mc} for a detailed discussion.
705:
706: A particularly clean probe of the two--gluon formfactor is
707: $J/\psi$ photoproduction, the $t$--dependence of which has been
708: measured in experiments at the HERA collider
709: ($x \sim 10^{-2} - 10^{-4}$) \cite{Aktas:2005xu,Chekanov:2004mw},
710: the FNAL fixed--target experiment
711: ($\langle x \rangle \sim 5 \times 10^{-2}$) \cite{Binkley:1981kv},
712: and a number of fixed--target experiments at lower energies
713: ($x \sim 10^{-1}$); see Refs.~\cite{Frankfurt:2002ka,Strikman:2004km}
714: for a review of the data. This process probes the two--gluon formfactor
715: at an effective scale $Q^2 \approx 3 \, \text{GeV}^2$. Analysis of the data,
716: combined with theoretical investigations, has produced a rather detailed
717: picture of the gluonic transverse size of the nucleon and its
718: $x$--dependence \cite{Frankfurt:2005mc}. For $x \sim 0.1 - 0.3$,
719: the gluonic transverse size suggested by the fixed--target data is
720: $\langle \rho^2 \rangle_g \approx 0.25 \, \text{fm}^2$, close
721: to $2/3$ times the proton's axial charge radius, $\langle r^2 \rangle_A$.
722: Between $x \sim 10^{-1}$ and $x \sim 10^{-2}$, $\langle \rho^2 \rangle_g$
723: increases by $\sim 30\%$. This can be explained by the contribution of the
724: nucleon's pion cloud to the gluon density at large transverse distances,
725: $\rho \sim 1/(2 M_\pi )$, which is dynamically suppressed for
726: $x > M_\pi / M_N$ and reaches its full strength for
727: $x \ll M_\pi / M_N$ \cite{Strikman:2003gz}. Finally, over
728: the HERA range, $x \sim 10^{-2} - 10^{-4}$, the gluonic
729: transverse size exhibits a logarithmic growth with $1/x$,
730: \beq
731: \langle \rho^2 \rangle_g \;\; = \;\;
732: \langle \rho^2 \rangle_g (x_0) \; + \; 4 \, \alpha'_g \, \ln\frac{x_0}{x}
733: \hspace{2em} (x < x_0 \approx 10^{-2}),
734: \label{alpha_g}
735: \eeq
736: with a rate, $\alpha'_g$, considerably smaller than that governing the
737: growth of the proton's transverse size in $pp$ elastic scattering,
738: which is dominated by soft interactions,
739: \beq
740: \alpha'_g \;\; \ll \;\; \alpha' .
741: \label{alpha_prime_comparison}
742: \eeq
743: A recent analysis of the H1 data finds $\alpha'_g = 0.164 \pm 0.028
744: (\mbox{stat}) \pm 0.030 (\mbox{syst}) \, \text{GeV}^{-2}$ for
745: $J/\psi$ photoproduction
746: and $0.019 \pm 0.139 (\mbox{stat}) \pm 0.076 (\mbox{syst}) \,
747: \text{GeV}^{-2}$
748: for electroproduction \cite{Aktas:2005xu};
749: an analysis of ZEUS electroproduction data
750: quotes $\alpha'_g = 0.07 \pm 0.05 (\mbox{stat}) { }^{+0.03}_{-0.04}
751: (\mbox{syst}) \, \text{GeV}^{-2}$ \cite{Chekanov:2004mw}, significantly
752: smaller than the soft value $\alpha' = 0.25 \, \text{GeV}^{-2}$.
753: The smaller rate of growth of the nucleon's size in hard interactions
754: can qualitatively be explained by the suppression of Gribov diffusion
755: in the decay of hard (highly virtual) partons as compared to soft partons.
756:
757: %
758: % FIGURE
759: %
760: \begin{figure}
761: %
762: \includegraphics[width=7cm]{twoscale.eps}
763: %
764: \caption[]{The ``two--scale picture'' of the transverse
765: structure of the proton in high--energy collisions.}
766: \label{fig:twoscale}
767: %
768: \end{figure}
769: %
770: A crucial observation is that the transverse area occupied by
771: partons with $x \gtrsim 10^{-1}$ is much smaller than the
772: transverse area associated with the proton in soft interactions
773: (see Fig.~\ref{fig:twoscale}),
774: \beq
775: \langle \rho^2 \rangle_g \, (x \gtrsim 10^{-1})
776: \;\; \ll \;\; \langle \rho^2 \rangle_{\text{soft}},
777: \eeq
778: or
779: \beq
780: 2 B_g \;\; \ll \;\; B .
781: \eeq
782: In high--energy $pp$ collisions with hard partonic processes one
783: is thus dealing with a two--scale picture of the transverse structure
784: of the proton. Moreover, when considering the production of a heavy
785: particle with fixed mass, $m_H$, in a partonic process with
786: $x_{1, 2} \sim m_H / \sqrt{s}$, the ``soft'' area
787: of the proton increases with $s$ faster than the ``hard'' area
788: (which changes as a result of the decrease of $x$),
789: because $\alpha' > \alpha'_g$, \textit{cf.}\
790: Eq.~(\ref{alpha_prime_comparison}). Thus, the difference of the
791: two areas becomes even more pronounced with increasing energy.
792:
793: %
794: % FIGURE
795: %
796: \begin{figure}
797: %
798: \includegraphics[width=8cm]{fg_dip_exp.eps}
799: %
800: \caption[]{Comparison of the dipole (solid line) and exponential
801: (dashed line) parametrizations of the two--gluon formfactor, with
802: the parameters related by Eq.~(\ref{dip_exp}). Shown is the
803: squared two--gluon formfactor, $F_g^2 (t)$, for both parametrizations,
804: as corresponds to the $t$--dependence of the cross section for
805: $J/\psi$ photoproduction (details see text).}
806: \label{fig_fg_dip_exp}
807: \end{figure}
808: %
809: For our studies of hard processes in diffractive $pp$ scattering
810: we require a parametrization of the $t$--dependence of the
811: two--gluon formfactor, \textit{viz.}\ the shape of the transverse
812: spatial distribution of gluons. The $x$--values probed in
813: Higgs production at central rapidities are $x \sim 10^{-2}$
814: at the LHC energy. Taking into account the effect of DGLAP evolution,
815: even larger values of $x$ are probed when parametrizing the two--gluon
816: formfactor at the $J/\psi$ production scale, $Q^2 \sim 3 \, \text{GeV}^2$
817: (for a general discussion of the effect of DGLAP evolution on the transverse
818: spatial distribution of gluons, see Ref.~\cite{Frankfurt:2003td}).
819: We thus need to look at the $J/\psi$ photoproduction data at
820: $x \gtrsim 10^{-2}$, which are probed in fixed--target experiments.
821:
822: Theoretical arguments suggest that the two--gluon formfactor at
823: $x \gtrsim 10^{-1}$ should be close to the axial formfactor,
824: which is well described by a dipole form (we omit all arguments
825: except $t$),
826: \beq
827: F_g (t) \;\; = \;\; \frac{1}{(1 - t/m_g^2)^2} ,
828: \label{twogl_dipole}
829: \eeq
830: with $m_g^2 \approx 1 \, \text{GeV}^2$ \cite{Frankfurt:2002ka}.
831: The corresponding transverse spatial distribution of gluons is given by
832: \beq
833: F_g (\bm{\rho}) \;\; = \;\; \frac{m_{g}^2}{2\pi}
834: \; \left(\frac{m_{g} \rho}{2}\right) \; K_1 (m_{g} \rho ) .
835: \label{spatial_dipole}
836: \eeq
837: We also consider an exponential parametrization
838: of the two--gluon formfactor,
839: \beq
840: F_g (t) \;\; = \;\; \exp (B_g t/2) ,
841: \label{twogl_exponential}
842: \eeq
843: corresponding to
844: \beq
845: F_g (\bm{\rho} ) \;\; = \;\;
846: \frac{\exp \left[ -\bm{\rho}^2 / (2 B_g)\right]}
847: {2 \pi B_g} .
848: \label{spatial_gaussian}
849: \eeq
850: The relation between the parameters of the dipole and exponential
851: parametrization which would follow from identifying
852: $\langle \rho^2 \rangle = 4 \, dF_g / d t (t = 0)$
853: is $B_g = 4/m_g^2$. Better overall agreement between the squared
854: formfactors for $|t| < 1\, \text{GeV}^2$ is obtained for somewhat
855: smaller values of $B_g$. Matching the squared formfactors at
856: $|t| = 0.5\, \text{GeV}^2$ we obtain
857: \beq
858: B_g \;\; = \;\; \frac{3.24}{m_g^2} ,
859: \label{dip_exp}
860: \eeq
861: see Figure~\ref{fig_fg_dip_exp}. It was shown in Ref.~\cite{Strikman:2004km}
862: that both the dipole with $m_g^2 = 1.1\,\text{GeV}^2$ and the
863: exponential with $B_g = 3.0 \, \text{GeV}^2$ given by Eq.~(\ref{dip_exp})
864: describe well the $t$--dependence of the data from the FNAL E401/E458
865: experiment at $\langle E_\gamma \rangle = 100\; {\rm GeV}$ in which the
866: recoiling proton was detected \cite{Binkley:1981kv}. We also note
867: that this value of $B_g$ is consistent with what
868: one obtains from the extrapolation of the HERA data towards larger
869: $x$, using Eq.~(\ref{alpha_g}) with the measured $\alpha'_g$.
870: We shall use the dipole, Eqs.~(\ref{twogl_dipole}), with
871: $m_g^2 = 1 \, \text{GeV}^2$ and the exponential,
872: Eq.~(\ref{twogl_exponential}), with $B_g = 3.24 \, \text{GeV}^{-2}$,
873: as our standard parametrizations for calculations in the kinematics of
874: Higgs production at the LHC below; comparison between the two will
875: allow us to estimate the uncertainty of our numerical predictions with
876: respect to the shape of the two--gluon formfactor.
877: %
878: \section{Theory of rapidity gap survival}
879: \label{sec:gap_survival}
880: %
881: We now outline the basic steps in the calculation of the
882: amplitude of double--gap exclusive diffractive processes
883: (\ref{exclusive_diffraction}),
884: and develop the physical picture of RGS.
885: The underlying idea of our approach is that hard and soft interactions
886: are approximately independent, because they happen over widely different
887: distance-- and time scales.
888: %
889: \subsection{Hard scattering process}
890: In the first step, one calculates the amplitude for double--gap
891: diffractive production of the high--mass system due to hard interactions.
892: For definiteness, we shall refer in the following to Higgs boson production,
893: keeping in mind that the discussed mechanism applies to production
894: of other high--mass states as well (dijets, heavy quarkonia, \textit{etc.}).
895: According to electroweak theory, the Higgs boson is produced
896: predominantly through its coupling to gluons via a quark loop;
897: for a review and references see Ref.~\cite{Gunion:1992hs}.
898: In contrast to inclusive production,
899: the amplitude for double--gap diffractive production is
900: in the lowest order in the QCD running coupling constant, $\alpha_s$,
901: given by the exchange of two gluons with vacuum quantum numbers
902: in the $t$--channel (see Fig.~\ref{fig:hard}). The Higgs boson
903: is radiated from one of the gluon lines. The role of the second
904: exchanged gluon is to neutralize the color charge in order to avoid gluon
905: bremsstrahlung. However, global color neutrality alone is not sufficient.
906: To suppress radiation, one must require that color be screened
907: locally in space--time. Conversely, this means that the selection
908: of a diffractive process, without accompanying radiation, guarantees
909: some degree of localization of the exchanged system.
910:
911: Operationally, the localization of the exchanged two--gluon system is
912: ensured by Sudakov formfactors, which suppress configurations with
913: low virtualities prone to emit gluon bremsstrahlung. The actual
914: calculation of the hard scattering amplitude including Sudakov
915: suppression is a challenging problem, which was addressed in
916: various approximations in Refs.~\cite{Khoze:1997dr,Khoze:2000cy}.
917: Fortunately, for our purposes we do not need to solve this problem
918: at a fully quantitative level, as only a few qualitative aspects of
919: the hard scattering process turn out to be essential for the
920: physics of RGS.
921:
922: To discuss the hard scattering process, it is natural to perform
923: a Sudakov decomposition of the four--momenta, using the
924: initial proton momenta, $p_1$ and $p_2$, as basis vectors, with
925: $2 (p_1 p_2) = s$ (we neglect the proton mass). As the transverse
926: momenta of the final--state protons are small compared to the
927: Higgs mass, we can expand the final proton four--momenta as
928: \be
929: p_1' &=& (1 - \xi_1) p_1 \; + \; p_{1\perp}' ,
930: \nonumber \\
931: p_2' &=& (1 - \xi_2) p_2 \; + \; p_{2\perp}' ,
932: \ee
933: where $(p_{1\perp}', p_1) = (p_{1\perp}', p_2) = 0$ \textit{etc.},
934: and $\xi_{1, 2}$ parametrize the longitudinal momentum loss
935: [\textit{cf.}\ Eq.~(\ref{xi_def}) and the footnote before it],
936: \beq
937: \xi_{1, 2} \;\; = \;\; \frac{m_H}{\sqrt{s}} \, e^{\pm y},
938: \label{xi_kinematics}
939: \eeq
940: where $y$ is the rapidity of the produced Higgs boson.
941: Assuming a Higgs mass of the order of $100 - 200 \, \text{GeV}$,
942: the typical values of $\xi_{1, 2}$ are of the order of $10^{-2}$
943: for production at central rapidities at the LHC
944: ($\sqrt{s} = 14 \, \text{TeV}$).
945:
946: %
947: % FIGURE
948: %
949: \begin{figure}
950: %
951: \includegraphics[width=5cm]{hard.eps}
952: %
953: \caption[]{The hard scattering process in double--gap exclusive diffractive
954: Higgs boson production (\ref{exclusive_diffraction}). Two gluons
955: are exchanged between the protons. The gluon--Higgs coupling is
956: indicated as a local vertex. The upper and lower blobs
957: denote the gluon--proton scattering amplitude, which can be
958: calculated in terms of the gluon GPD in the proton.}
959: \label{fig:hard}
960: %
961: \end{figure}
962: %
963: Consider now the two--gluon exchange process of Fig.~\ref{fig:hard}
964: as a Feynman diagram, in which the upper and lower blobs denote the
965: gluon--proton scattering amplitudes, to be specified in more detail
966: below. The four--momenta of the gluons
967: coupling to the Higgs we parametrize as
968: \be
969: k_1 &=& x_1 p_1 \; + \; x_2' p_2 \; + \; k_\perp - p_{1\perp}',
970: \label{k_1_Sudakov}
971: \\
972: k_2 &=& x_1' p_1 \; + \; x_2 p_2
973: \; - \; k_\perp - p_{2\perp}' .
974: \label{k_2_Sudakov}
975: \ee
976: The four--momentum of the screening gluon then follows from
977: four--momentum conservation,
978: \beq
979: k \;\; = \;\; x_1' p_1 \; + \; x_2' p_2 \; + \; k_\perp ,
980: \eeq
981: where
982: \be
983: x_1' &\equiv& \xi_1 - x_1 ,
984: \\
985: x_2' &\equiv& \xi_2 + x_2 .
986: \ee
987: We want to identify the dominant region of integration in the
988: loop integral. First, analogy with inclusive production of heavy particles
989: at central rapidities suggests that the momentum fractions of the
990: gluons producing the Higgs boson (with respect to their parent protons)
991: are practically the same as given by the naive parton model estimate,
992: in which one neglects the transverse momenta and virtualities
993: of the annihilating gluons,
994: \beq
995: x_{1,2} \;\; \sim \;\; \frac{m_H}{\sqrt{s}} \, e^{\pm y} .
996: \label{x_12}
997: \eeq
998: That is, the momentum fractions of the annihilating gluons are equal
999: to the protons' fractional longitudinal momentum loss,
1000: $x_{1, 2} \approx \xi_{1, 2}$. Second, in the case of double--gap
1001: diffractive production the Sudakov formfactor associated with the
1002: Higgs boson vertex, which accounts for the absence of gluon
1003: bremsstrahlung, restricts the (spacelike) virtualities of the
1004: annihilating gluons and their transverse momenta to values of the
1005: order of some ``intermediate'' hard scale,
1006: \beq
1007: -k_{1, 2}^2, \; -k_\perp^2 \;\; \sim \;\; Q_{\text{int}}^2,
1008: \label{k2_Q2}
1009: \eeq
1010: with
1011: \beq
1012: m_H^2/4 \;\; \gg \;\; Q_{\text{int}}^2 \;\; \gg \;\; \Lambda_{\text{QCD}}^2 .
1013: \label{Q2_def}
1014: \eeq
1015: This was demonstrated explicitly in Ref.~\cite{Khoze:1997dr}, where
1016: the distribution of transverse momenta in the loop integral was studied
1017: in a model which included the LO Sudakov formfactor associated
1018: with the $ggH$ vertex; see also Ref.~\cite{Khoze:2000cy}
1019: \footnote{The Sudakov formfactor associated
1020: with the $ggH$ vertex accounts for the absence of gluon bremsstrahlung
1021: with transverse momenta $|k_\perp| < |k_{\perp, \text{rad}}| < m_H/2$,
1022: radiated from the annihilating gluons.
1023: The requirement of absence of radiation from the screening gluon
1024: results in an additional shift of the $k_\perp$ distribution towards
1025: larger values. However the amplitude for radiation from this line
1026: does not contain factors $\alpha_s \ln (k_\perp^2 /m_H^2)$, and is
1027: beyond the accuracy of the calculation of Ref.~\cite{Khoze:1997dr}.}.
1028: Expressing now the virtualities of the annihilating gluons, $k_{1, 2}^2$,
1029: in terms of the decompositions (\ref{k_1_Sudakov}) and
1030: (\ref{k_2_Sudakov}), neglecting the proton transverse momenta
1031: relative to $k_\perp$, we find that Eq.~(\ref{k2_Q2}) implies
1032: \beq
1033: x_{1, 2}' \;\; = \;\; \frac{k_{2, 1}^2 - k_\perp^2}{x_{2, 1} s}
1034: \;\; \sim \;\; \frac{Q_{\text{int}}^2}{x_{2, 1} s}
1035: \;\; \sim \;\; \frac{Q_{\text{int}}^2}{m_H \sqrt{s}}
1036: \;\; \ll \;\; x_{1, 2} ,
1037: \label{x_12_prime}
1038: \eeq
1039: \textit{i.e.}, the energy and longitudinal momentum fraction
1040: of the screening gluon are substantially smaller than those of
1041: the annihilating gluons. The screening gluon does not ``belong to''
1042: any of the two protons; its momentum is predominantly transverse,
1043: and it has spacelike virtuality,
1044: \beq
1045: -k^2 \;\; \approx \;\; -k_\perp^2 \;\; \sim \;\; Q_{\text{int}}^2 .
1046: \eeq
1047: In the annihilating gluons, on the other hand, longitudinal and
1048: transverse momenta contribute in equal amounts to the virtuality,
1049: \beq
1050: -k_{1,2}^2 \;\; \approx \;\; x_{1, 2} x_{2, 1}' s - k_\perp^2
1051: \;\; \sim \;\; Q_{\text{int}}^2 .
1052: \eeq
1053: To summarize, the hard scattering process takes the form of the
1054: exchange of two gluons with comparable virtualities $\sim Q_{\text{int}}^2$,
1055: and transverse momenta $\sim Q_{\text{int}}$, between the protons.
1056: Of the two gluons, one carries substantial longitudinal momentum fraction
1057: of the proton, $\sim m_H/\sqrt{s}$, and annihilates with the
1058: corresponding other to make the Higgs, the other gluon represents
1059: a ``Coulomb--like'' exchange with small momentum fraction
1060: $\sim Q_{\text{int}}^2/(m_H \sqrt{s})$.
1061:
1062: The important point about the two--gluon exchange process
1063: is the appearance of the intermediate hard scale, $Q_{\text{int}}^2$,
1064: governing the virtualities and transverse momenta of the exchanged gluons.
1065: This allows us to make a crucial simplification in the description
1066: of the gluon--proton scattering amplitudes, which we have not yet
1067: specified so far. Namely, we argue that, in a partonic description
1068: of the proton, the gluon--proton scattering amplitude is dominated
1069: by the two gluons coupling to the same parton \cite{Brodsky:1994kf}.
1070: This approximation is analogous to the assumption of dominance of the
1071: ``handbag graph'' in virtual Compton scattering at large photon virtuality,
1072: $Q^2$, which is well established and forms the basis of QCD factorization
1073: for this process \cite{Mueller:1998fv,Abramowicz:1995hb,%
1074: Ji:1996nm,Radyushkin:1997ki,Collins:1998be}.
1075: In this approximation, the gluon--proton scattering amplitude,
1076: which is predominantly imaginary at high energies, can be calculated
1077: in terms of the generalized parton distributions (GPDs) --- here,
1078: predominantly, gluon distributions --- in the protons.
1079:
1080: We do not attempt to calculate the absolute normalization of the
1081: amplitude for double--gap hard diffractive
1082: production through two--gluon exchange in terms of the GPDs; doing so
1083: would require a substantially more accurate evaluation of the
1084: two--gluon exchange graph than the qualitative estimates presented
1085: above. Fortunately, for the theory of RGS, the only
1086: information we require (in addition to the qualitative properties of
1087: the hard process derived above) is the dependence of the double--gap
1088: hard diffractive amplitude on the transverse momentum transfers to the
1089: protons, $p_{1\perp}$ and $p_{2\perp}$. For sufficiently large scales
1090: $Q_{\text{int}}^2$, this dependence is described by the GPDs,
1091: even if the gluon momentum fractions in the hard
1092: amplitude and the virtuality of the exchanges are subject to
1093: integration over a certain range, and determined only in
1094: order--of--magnitude. Thus, we can state that the
1095: $p_{1\perp}$-- and $p_{2\perp}$ dependence of the double--gap
1096: hard diffractive amplitude is proportional to
1097: \beq
1098: T_{\text{hard}} \;\; \propto \;\;
1099: H_g (x_1, \xi_1, t_1; Q^2) \; H_g (x_2, \xi_2, t_2; Q^2) ,
1100: \label{A_hard_GPD}
1101: \eeq
1102: where
1103: \beq
1104: t_1 \; = \; p_{1\perp}^{\prime 2} \; < 0,
1105: \hspace{2em}
1106: t_{2} \; = \; p_{2\perp}^{\prime 2} \; < 0
1107: \eeq
1108: are the invariant momentum transfers to the proton. Here the longitudinal
1109: momentum transfers, $\xi_{1, 2}$, are kinematically fixed by
1110: Eq.~(\ref{xi_kinematics}), while parton momentum fractions,
1111: $x_{1,2}$, are determined by Eq.~(\ref{x_12}) with accuracy
1112: given by Eq.~(\ref{x_12_prime}). The resolution scale, $Q^2$,
1113: at which the GPD needs to be taken here, is parametrically of
1114: the order $Q_{\text{int}}^2$ [\textit{cf.}~Eq.~(\ref{Q2_def})],
1115: but numerically substantially larger,
1116: \beq
1117: Q^2 \;\; \sim \;\; \text{several times} \;\;\; Q_{\text{int}}^2 .
1118: \eeq
1119: This follows from the fact that, by convention, $Q^2$ determines
1120: the upper limit of the transverse momenta in the parton
1121: distribution, while $Q_{\text{int}}^2$ is a measure of the
1122: dominant (average) values in the distribution, which for a
1123: $1/k_\perp^2$ distribution is significantly lower than the upper limit.
1124:
1125: %
1126: % FIGURE
1127: %
1128: \begin{figure*}
1129: %
1130: \includegraphics[width=13cm]{parton.eps}
1131: %
1132: \caption[]{Schematic illustration of hard and soft interactions
1133: in the parton picture of double--gap exclusive diffractive $pp$ scattering.
1134: (a) The hard scattering process producing the large--mass system
1135: (Higgs, dijet) is represented by a local operator in parton degrees
1136: of freedom. (b) Hard and soft interactions are approximately
1137: independent since they happen over widely different time-- and
1138: distance scales.}
1139: \label{fig:hardsoft}
1140: \end{figure*}
1141: %
1142: To conclude this discussion, some comments concerning QCD evolution
1143: and the modeling of the GPDs are in order. First, in Higgs boson production
1144: at the LHC we are dealing with the gluon GDP at $x, \xi \sim 10^{-2}$
1145: and $x' \ll x$, where it can legitimately be calculated by applying QCD
1146: evolution to a diagonal GPD at a lower scale (\textit{cf.}\ the discussion
1147: in Section~\ref{sec:transverse_hard}).
1148: Second, at LHC energies the typical momentum fraction of the
1149: screening gluon reaches values $x_{1,2}' \sim 10^{-6}$
1150: for $Q^2 \sim \text{few GeV}^2$, \textit{cf.}\ Eq.~(\ref{x_12_prime}).
1151: For such low values of $x$ the use of DGLAP evolution in principle
1152: requires justification. However, as explained in detail in
1153: Ref.~\cite{Frankfurt:2005mc}, for such values of $x$ the kinematic
1154: conditions still restrict the actual number of radiated gluons to a few,
1155: so that NLO DGLAP and resummed BFKL evolution give similar results,
1156: see Ref.~\cite{Salam:2005yp} for a review. In this sense, the use of
1157: GPDs generated by DGLAP evolution seems to be appropriate.
1158: Third, one may wonder about the breakdown of the pQCD calculation
1159: of the hard reaction amplitude due to the growth of the gluon density
1160: at small $x$ within the DGLAP approximation --- the BDL,
1161: see Ref.~\cite{Frankfurt:2005mc} and references therein.
1162: This should not be a problem in the present context,
1163: since the BDL affects only collisions at central impact
1164: parameters, where the diffractive amplitude is anyway
1165: suppressed due to the vanishing rapidity gap survival
1166: probability (see Sec.~\ref{sec:cross_section} below).
1167: %
1168: \subsection{Combining hard and soft interactions}
1169: \label{subsec:combining}
1170: In the second step, we formalize the interplay of hard and soft
1171: interactions in the amplitude of the hadronic diffractive process,
1172: Eq.~(\ref{exclusive_diffraction}). To this end, we invoke
1173: the parton picture of the proton wavefunction, as developed
1174: by Gribov \cite{Gribov:1973jg}. We consider the process
1175: (\ref{exclusive_diffraction}) in the CM frame, in which the
1176: two protons in the initial state have longitudinal momenta
1177: $\pm \sqrt{s}/2$, and zero transverse momentum,
1178: $\bm{p}_{1\perp}, \bm{p}_{2\perp} = 0$. Since $m_H \ll \sqrt{s}/2$,
1179: angular momentum conservation implies that the reaction amplitude
1180: is approximately diagonal in the
1181: transverse coordinates of the colliding protons (\textit{i.e.}, in impact
1182: parameter) as in two--body elastic scattering at high energies. We thus
1183: consider partonic configurations centered around the transverse centers of
1184: the two protons, in which the partons carry fractions of the
1185: longitudinal momentum of the respective proton.
1186: We may regard the hard scattering process as
1187: an operator in the basis of these partonic states, denoted by
1188: $\hat V_{\text{hard}}$. Soft interactions, which build up the partonic
1189: wavefunctions, are governed by a soft Hamiltonian, $\hat H_{\text{soft}}$.
1190: While we do not know their explicit form, we can nevertheless state some
1191: important properties of these operators:
1192: %
1193: \begin{enumerate}
1194: %
1195: \item $\hat V_{\text{hard}}$ is local in time (instantaneous)
1196: on the typical timescale of soft interactions,
1197: %
1198: \item $\hat V_{\text{hard}}$ is local in transverse position
1199: on the distance scale over which the transverse spatial distribution
1200: of partons in the protons changes due to soft interactions,
1201: %
1202: \item $\hat V_{\text{hard}}$ preserves the number of partons,
1203: since scattering of both gluons from the same parton dominates in the
1204: hard regime.
1205: %
1206: \item $\hat V_{\text{hard}}$ preserves the helicity of the colliding partons,
1207: because in a perturbative gauge theory the dominant contribution to the
1208: interaction of partons over large rapidity intervals comes from
1209: the parton helicity--conserving component of the propagator of
1210: the exchanged gluon \cite{Gribov:2003nw}.
1211: %
1212: \end{enumerate}
1213: %
1214: As a consequence of these properties, we conclude that the operator of
1215: the hard reaction commutes with the Hamiltonian of soft interactions,
1216: \beq
1217: [ \hat V_{\text{hard}}, \, \hat H_{\text{soft}} ] \;\; = \;\; 0.
1218: \label{hard_soft_commute}
1219: \eeq
1220: This is the mathematical statement of the independence of hard and soft
1221: interactions in diffraction. A schematic illustration of this picture is
1222: given in Fig.~\ref{fig:hardsoft}. We shall refer to
1223: Eq.~(\ref{hard_soft_commute}) and the formulas derived from it
1224: as the independent interaction approximation.
1225:
1226: The amplitude for the double--gap exclusive diffractive process
1227: (\ref{exclusive_diffraction}) is then determined by the
1228: matrix element
1229: \beq
1230: T_{\text{diff}} \;\; = \;\;
1231: \langle p_1' p_2'|
1232: \; \hat S_{\text{soft}} (\infty, 0) \; \hat V_{\text{hard}} \;
1233: \hat S_{\text{soft}} (0, -\infty) \;
1234: | p_1 p_2 \rangle ,
1235: \eeq
1236: where
1237: \beq
1238: \hat S_{\text{soft}} (t_2, t_1) \;\; \equiv \;\;
1239: \text{T} \; \int_{t_1}^{t_2} dt \, \exp (i t \hat H_{\text{soft}} )
1240: \eeq
1241: is the time evolution operator due to soft interactions (we have put the
1242: time of the hard interaction at $t = 0$). Because of
1243: Eq.~(\ref{hard_soft_commute}), the operator $\hat V_{\text{hard}}$ commutes
1244: with the soft time evolution operator. Using the property
1245: \beq
1246: \hat S_{\text{soft}} (\infty, 0) \, \hat S_{\text{soft}} (0, -\infty)
1247: \;\; = \;\;
1248: \hat S_{\text{soft}} (\infty, -\infty)
1249: \;\; \equiv \;\; \hat S_{\text{soft}} ,
1250: \eeq
1251: where $\hat S_{\text{soft}}$ is the $S$--matrix due to soft
1252: interactions, we obtain
1253: \beq
1254: T_{\text{diff}}
1255: \;\; = \;\;
1256: \langle p_1' p_2' | \; \hat V_{\text{hard}} \; \hat S_{\text{soft}} \;
1257: | p_1 p_2 \rangle .
1258: \label{RGS_master}
1259: \eeq
1260: Thus, the amplitude is expressed in terms of the observable matrix
1261: elements of the soft $S$--matrix, and the operator $\hat V_{\text{hard}}$,
1262: calculable in QCD at the partonic level. Eq.~(\ref{RGS_master})
1263: is the fundamental expression for discussing the physics
1264: of RGS within our approach.
1265:
1266: \subsection{Suppression of inelastic diffraction}
1267: \label{subsec:inelastic}
1268: In the next step, we evaluate the amplitude for the
1269: double--gap exclusive diffractive process based on Eq.~(\ref{RGS_master}),
1270: by inserting ``intermediate'' states (actually, states at $t = \infty$)
1271: between the operators. In principle, one needs to sum over all
1272: diffractive states (elastic and inelastic) produced by the
1273: operator $\hat V_{\text{hard}}$. An important question is which states can
1274: give large contributions to the matrix element. In fact, it turns out
1275: that the different preferences of hard and soft interactions severely
1276: restrict the range of states which can effectively contribute.
1277:
1278: Simple arguments show that large--mass diffractive states should
1279: make a negligible contribution to Eq.~(\ref{RGS_master}). If the
1280: two hard gluons in the hard interaction are attached to two different
1281: partons in the proton, the inelastic states predominantly produced
1282: are two jets and gluon bremsstrahlung. It is virtually impossible
1283: to produce such states in soft interactions, hence they cannot
1284: contribute to Eq.~(\ref{RGS_master}). If the two hard gluons are attached
1285: to the same parton, the cross section of inelastic diffraction is small
1286: for small $t$ because of the small overlap integral with the inelastic
1287: state (most of the overlap is with the elastic state), while for large $t$
1288: one produces a single parton with transverse momentum
1289: $p_\perp \sim \sqrt{-t}$, which again
1290: is a state difficult to reach through soft interactions. In addition,
1291: for $t \neq 0$ soft diffraction at LHC energies is known to be dominated
1292: by the spin--flip amplitude, which further suppresses the overlap integral.
1293: Together, this restricts the possible mass range of diffractively produced
1294: states to $M^2_{\text{diff}} \sim \text{few GeV}^2$.
1295:
1296: For a more quantitative estimate, we suppose that the state produced
1297: through inelastic diffraction has the form
1298: $|pp\rangle + \epsilon | pX \rangle$, where the state $X$
1299: is different from the proton, and $\epsilon$ is a small correction.
1300: We can then estimate $\epsilon$ from the Schwarz inequality:
1301: \beq
1302: \frac{\epsilon}{2} \;\; = \;\;
1303: \sqrt \frac{ \sigma_{\text{soft}}(pp\to Xp) \;
1304: \sigma_{\text{hard}}(pp\to Xp)}
1305: {\sigma_{\text{soft}}(pp\to pp) \;
1306: \sigma_{\text{hard}}(pp\to pp)} ,
1307: \eeq
1308: where $\sigma_{\text{hard}}(pp\to pX)$ is the cross section for the
1309: production of the state $| pX \rangle$ by the operator
1310: $\hat V_{\text{hard}}$.
1311: Analysis of the Tevatron data (for a review, see
1312: Ref.~\cite{Goulianos:2005ac})
1313: shows that the fraction of diffractive events in soft collisions decreases
1314: with increasing energy, and that the distribution over the excitation mass
1315: is $\propto 1/M_X^2$. As a result, we expect that at the LHC energy
1316: $\epsilon \le 2 \cdot 10^{-2}$. Thus, the diffractively produced state
1317: is actually the $|pp\rangle$ state, and the contributions from
1318: inelastic diffraction are small.
1319:
1320: The small overlap between hard and soft diffraction can also be understood
1321: as the result of the different impact parameter dependence of both types
1322: of processes. Hard diffraction occurs mostly at small
1323: impact parameters, $b^2 \sim B_g$. Soft diffraction, because of the
1324: approach to the BDL, occurs mostly at large impact parameters,
1325: $b^2 \sim B$, which, moreover, rapidly grow with the collision energy.
1326: We note again that the peripheral nature of soft diffraction
1327: was established already within Reggeon field theory, where it was
1328: found that the BDL solves the consistency problem associated with the
1329: triple Reggeon formula \cite{Marchesini:1976hw}.
1330:
1331: The restriction to the $pp$ intermediate state turns Eq.~(\ref{RGS_master})
1332: into a tool for quantitative evaluation of the diffractive amplitude
1333: and the RGS. In particular, with the $pp$ intermediate state we
1334: can approximate the matrix element of the soft--interaction
1335: time evolution operator by that of the full $S$--matrix,
1336: \textit{i.e.}, the $pp$ elastic scattering amplitude, which
1337: is known experimentally; see Section~\ref{sec:transverse_soft}.
1338: For this approximation to be legitimate it is crucial that
1339: scattering at small impact parameters turns out to be strongly suppressed
1340: due to the approach to the BDL in $pp$ elastic scattering,
1341: as will be seen from the results of Section~\ref{sec:cross_section}.
1342: The diffractive process is thus dominated by large impact parameters,
1343: where $pp$ elastic scattering is dominated by soft interactions.
1344:
1345: In the studies of double--gap exclusive diffraction based on the
1346: pomeron model of Ref.~\cite{Khoze:2000wk}, inelastic intermediate
1347: states were effectively included by way of a two--component formalism
1348: (however, no explicit non-diagonal ``transition'' GPDs were introduced).
1349: We have argued here that these contributions are strongly suppressed,
1350: because of the small overlap of states accessible in hard and soft
1351: interactions. We shall comment on the implications of this for
1352: the numerical values of the RGS probability
1353: in Section~\ref{subsec:numerical}
1354:
1355: %
1356: % FIGURE
1357: %
1358: \begin{figure*}
1359: %
1360: \includegraphics[width=14cm]{mom.eps}
1361: %
1362: \caption{The amplitude for double--gap exclusive hard diffraction in momentum
1363: representation, Eqs.~(\ref{amp_momentum})--(\ref{amp_momentum_t}).
1364: The first term is the amplitude of the hard reaction alone,
1365: the second term the correction due to soft elastic rescattering.
1366: Only the transverse momenta of the protons are indicated;
1367: the momentum transfer due to soft elastic scattering is
1368: $\bm{\Delta}_\perp$.}
1369: \label{fig:mom}
1370: \end{figure*}
1371: %
1372: \subsection{Evaluation of the diffractive amplitude}
1373: \label{subsec:evaluation}
1374: It remains to actually evaluate the matrix element (\ref{RGS_master})
1375: with $|pp\rangle$ intermediate states, using the specific form
1376: of the hard scattering amplitude and the $pp$ elastic
1377: scattering amplitude. We insert a set of $pp$ intermediate
1378: states in the form
1379: \beq
1380: \int\! \frac{d^3 p_1^{''}}{(2\pi )^3 \sqrt{s}} \;
1381: \int\! \frac{d^3 p_2^{''}}{(2\pi )^3 \sqrt{s}} \;
1382: |p_1^{''} p_2^{''} \rangle \langle p_1^{''} p_2^{''} | ,
1383: \eeq
1384: where we have approximated the energy of the individual protons
1385: by $\sqrt{s}/2$. The matrix element of the operator
1386: $\hat V_{\text{hard}}$ between the two--proton states is,
1387: by definition, given by [\textit{cf.}\ Eqs.~(\ref{A_hard_GPD})]
1388: \be
1389: \langle p_1^{'} p_2^{'} | \; \hat{V}_{\text{hard}} \; | p_1^{''} p_2^{''}
1390: \rangle
1391: &=& \kappa (s, \xi_1, \xi_2 ) \nonumber \\
1392: &\times& F_g (x_1, \xi_1, \tilde t_1; Q^2) \nonumber \\
1393: &\times& F_g (x_2, \xi_2, \tilde t_2; Q^2) ,
1394: \ee
1395: where
1396: \be
1397: \tilde t_1 &\equiv&
1398: -(\bm{p}_{1\perp}^{'} - \bm{p}_{1\perp}^{''})^2 ,
1399: \\
1400: \tilde t_2 &\equiv&
1401: -(\bm{p}_{2 \perp}^{'} - \bm{p}_{2 \perp}^{''})^2 .
1402: \ee
1403: The factor
1404: \be
1405: \kappa (s, \xi_1, \xi_2 ) &\equiv& C_{\text{hard}} \nonumber \\
1406: &\times& H_g (x_1, \xi_1, t_1 = 0) \nonumber \\
1407: &\times& H_g (x_2, \xi_2, t_2 = 0)
1408: \label{kappa}
1409: \ee
1410: represents the symbolic result for the absolute normalization of amplitude
1411: of the hard scattering process; it contains the amplitude of the two--gluon
1412: exchange process, $C_{\text{hard}}$, including the information about the
1413: $ggH$ coupling given by the electroweak theory, as well as the information
1414: about the gluon GPD in the colliding protons at $t = 0$.
1415: The information about the transverse momentum dependence of the amplitude
1416: is contained in the two--gluon formfactors, $F_g$,
1417: \textit{cf.}\ Eq.(\ref{twogl_def}). Furthermore, we replace
1418: in Eq.~(\ref{RGS_master})
1419: \beq
1420: \hat S_{\text{soft}}
1421: \;\; \rightarrow \;\; \hat S \;\; = \;\; 1 \; + \; (\hat S - 1),
1422: \label{S_decomposition}
1423: \eeq
1424: and use the fact that the matrix element of the operator
1425: $\hat S - 1$ is given by
1426: \be
1427: \langle p_1^{''} p_2^{''} | \; \hat{S} - 1 \; | p_1 p_2 \rangle
1428: &=& i (2\pi)^4 \; \delta^{(4)} (p_1^{''} + p_2^{''} - p_1 - p_2)
1429: \nonumber \\
1430: &\times& (4\pi) \, T_{\text{el}} (s, t) ,
1431: \label{S_minus_1_me}
1432: \ee
1433: with
1434: \beq
1435: t \;\; = \;\; - (\bm{p}_{1\perp}^{''} - \bm{p}_{1\perp})^2
1436: \;\; = \;\; - (\bm{p}_{2\perp}^{''} - \bm{p}_{2\perp})^2 .
1437: \label{S_minus_1_t}
1438: \eeq
1439: Finally, taking into account that at high energies the
1440: energy--conserving delta function in Eq.~(\ref{S_minus_1_me})
1441: effectively conserves longitudinal momentum, and combining the
1442: contributions from the two terms in Eq.~(\ref{S_decomposition}),
1443: we obtain
1444: \be
1445: \lefteqn{T_{\text{diff}} (\bm{p}_{1\perp}', \bm{p}_{2\perp}')
1446: \;\; = \;\; \int\frac{d^2 \Delta_\perp}{(2\pi)^2}} &&
1447: \nonumber \\
1448: &\times& \kappa \; F_g \left( x_1, \xi_1, \tilde t_1; \, Q^2 \right) \;
1449: F_g \left( x_2, \xi_2, \tilde t_2; \, Q^2 \right)
1450: \nonumber \\
1451: &\times & \left[ (2\pi)^2 \delta^{(2)}(\bm{\Delta}_\perp)
1452: \; + \; \frac{4\pi i}{s} T_{\text{el}} (s, t)
1453: \right] ,
1454: \label{amp_momentum}
1455: \ee
1456: where now
1457: \be
1458: \tilde t_1 &\equiv& -(\bm{p}_{1\perp}' - \bm{\Delta}_\perp)^2 , \,
1459: \\
1460: \tilde t_2 &\equiv& -(\bm{p}_{2\perp}' + \bm{\Delta}_\perp)^2 ,
1461: \\
1462: t &\equiv& -\bm{\Delta}_\perp^2 .
1463: \label{amp_momentum_t}
1464: \ee
1465: This result has a simple interpretation (see Fig.~\ref{fig:mom}).
1466: The first term in the
1467: brackets represents the amplitude of the hard reaction alone.
1468: The second term represents the contribution in which the
1469: hard reaction is accompanied by soft elastic rescattering
1470: with transverse momentum transfer $\bm{\Delta}_\perp$.
1471: The total amplitude is the coherent superposition of
1472: the two contributions. We note that the form of this result
1473: is analogous to the well--known absorption corrections
1474: in Regge phenomenology, in which an elementary Regge pole amplitude
1475: is modified by elastic rescattering.
1476:
1477: It is instructive to convert the result (\ref{amp_momentum}) to
1478: the transverse coordinate representation.
1479: Substituting the Fourier representation
1480: of the gluon GPDs, Eq.~(\ref{rhoprof_def}), and the representation
1481: of the $pp$ elastic amplitude in terms of the profile function,
1482: Eq.~(\ref{Gamma_def}), and using standard Fourier transform
1483: manipulations, we obtain
1484: \be
1485: \lefteqn{T_{\text{diff}} (\bm{p}_{1\perp}', \bm{p}_{2\perp}')
1486: \;\; = \;\; \int d^2 b \int d\rho_1 \int d\rho_2}
1487: && \nonumber \\
1488: &\times& \delta^{(2)} (\bm{b} - \bm{\rho}_1 + \bm{\rho}_2)
1489: \; e^{-i (\bm{p}_{1\perp}' \bm{\rho}_1)
1490: -i (\bm{p}_{2\perp}' \bm{\rho}_2)}
1491: \nonumber \\
1492: &\times& \kappa \;
1493: F_g \left( x_1, \xi_1, \bm{\rho}_1; \, Q^2 \right)
1494: \; F_g \left( x_2, \xi_2, \bm{\rho}_2; \, Q^2 \right)
1495: \nonumber \\
1496: &\times&
1497: \left[ 1 - \Gamma(s, \bm{b}) \right] .
1498: \label{T_rho}
1499: \ee
1500: Here the scattering amplitude is represented as the coherent superposition
1501: of amplitudes corresponding to $pp$ scattering at given transverse
1502: displacement (impact parameter), $\bm{b}$.
1503: The amplitude for the hard process is proportional
1504: to the product of the transverse spatial gluon transition densities
1505: at positions $\bm{\rho}_{1, 2}$ relative to the centers of the
1506: respective protons, with the three transverse vectors satisfying
1507: the triangular condition $\bm{\rho}_1 - \bm{\rho}_2 = \bm{b}$
1508: (see Fig.~\ref{fig:coord}). The modifications due to elastic
1509: rescattering now take the
1510: form of a multiplication of the hard scattering amplitude with
1511: the ``absorption factor,'' $1 - \Gamma(s, b)$. Note that
1512: the modulus squared of this factor can be interpreted as the
1513: probability for ``no inelastic interaction'' in $pp$ scattering
1514: at a given impact parameter, \textit{cf.}\ Eq.~(\ref{P_noin}).
1515: This interpretation will be explored further in Sec.~\ref{sec:cross_section}.
1516: %
1517: % FIGURE
1518: %
1519: \begin{figure}
1520: \includegraphics[width=5cm]{coord.eps}
1521: \caption{Illustration of the transverse coordinate representation
1522: of the diffractive amplitude,
1523: Eq.~(\ref{T_rho}). The hard scattering process is local in transverse space.
1524: The centers of the colliding protons are displaced by the distance
1525: $b = |\bm{b}|$, and $\bm{\rho}_{1, 2} = |\bm{\rho}_{1, 2}|$ are the
1526: distances from the centers to the point of the hard process.}
1527: \label{fig:coord}
1528: \end{figure}
1529: %
1530:
1531: Our partonic approach allows us to calculate the amplitude for double--gap
1532: exclusive diffraction in a model--independent way in terms of the gluon
1533: GPD and the phenomenological $pp$ elastic scattering amplitude; see
1534: Eqs.~(\ref{amp_momentum}--\ref{T_rho}). In Ref.~\cite{Kaidalov:2003fw}
1535: such processes were studied using a model of elastic $pp$ scattering
1536: which included the enhanced eikonal series of single Pomeron exchanges
1537: and the triple--Pomeron vertex to describe the soft spectator interactions.
1538: The expression for the amplitude in the case of a single Pomeron exchange
1539: in that model (and without inelastic intermediate states) would formally
1540: coincide with our expressions (\ref{amp_momentum}--\ref{T_rho}).
1541: Whether the same is true for the full amplitude in that model is
1542: less clear; \textit{cf.}\ the discussion of the numerical results
1543: in Section~\ref{subsec:numerical} below.
1544: %
1545: \section{The rapidity gap survival probability}
1546: \label{sec:cross_section}
1547: We now use our general result for the amplitude of double--gap exclusive
1548: diffractive processes in the independent interaction approximation,
1549: Eq.~(\ref{T_rho}), to calculate the RGS probability for such processes.
1550: At this level of approximation, we shall recover a simple ``geometric''
1551: expression for the RGS probability, which permits a probabilistic
1552: interpretation and was heuristically derived in
1553: Refs.~\cite{Frankfurt:2004kn,Frankfurt:2004ti,Frankfurt:2005mc}.
1554: We discuss the impact parameter dependence of RGS and stress the
1555: the crucial role of the BDL in stabilizing the numerical estimates
1556: ensuring a model--independent result.
1557: %
1558: \subsection{Impact parameter representation}
1559: \label{subsec:geometry}
1560: In order to compute the cross section for double--gap exclusive diffractive
1561: production of a given state at fixed rapidity, we integrate the
1562: modulus squared of the amplitude (\ref{T_rho}) over the
1563: final proton transverse momenta. By standard Fourier transform
1564: manipulations we obtain
1565: \be
1566: \sigma_{\text{diff}} &=& \text{(kinematic factors)} \; \times
1567: \int \frac{d^2 p_{1\perp}'}{(2\pi )^2} \;
1568: \int \frac{d^2 p_{2\perp}'}{(2\pi )^2} \;
1569: \nonumber \\
1570: &\times &
1571: \left| T_{\text{diff}} (\bm{p}_{1\perp}', \bm{p}_{2\perp}') \right|^2
1572: \\
1573: &=& \text{(const)} \; \times \; \int d^2 \rho_1 \; \int d^2 \rho_2 \;
1574: F^2_g (\bm{\rho}_1) \; F^2_g (\bm{\rho}_2) \;
1575: \nonumber \\
1576: &\times &
1577: \left| 1 - \Gamma (\bm{\rho}_2
1578: - \bm{\rho}_1) \right|^2
1579: \label{sigma_full}
1580: \ee
1581: (for brevity we suppress all arguments except the transverse
1582: coordinates in $F_g$ and $\Gamma$).
1583: The RGS probability due to soft interactions \cite{Bjorken:1992er},
1584: by definition, is given by the ratio of the cross section of the
1585: physical double--gap
1586: diffractive process (\ref{sigma_full}) to the cross section of the
1587: hypothetical process with the same hard scattering subprocess but
1588: with no soft spectator interactions, corresponding to
1589: expression (\ref{sigma_full}) with $\Gamma \equiv 0$,
1590: \beq
1591: S^2 \;\; \equiv \;\;
1592: \frac{\sigma_{\text{diff}} \, (\text{physical}) \phantom{xxi}}
1593: {\sigma_{\text{diff}} \, (\text{no soft interactions})} .
1594: \label{survb_def}
1595: \eeq
1596: We can cast this ratio in a simple form. We rewrite
1597: the convolution integral in Eq.~(\ref{sigma_full}) by inserting
1598: unity in the form (\textit{cf.}\ Fig.~\ref{fig:coord})
1599: \beq
1600: \int d^2 b \; \delta^{(2)} (\bm{b} + \bm{\rho}_1 - \bm{\rho}_2 ) ,
1601: \eeq
1602: and introduce a normalized impact parameter distribution,
1603: \be
1604: \lefteqn{
1605: P_{\text{hard}} (\bm{b}) \;\; \equiv \;\;
1606: \int d^2 \rho_1 \; \int d^2 \rho_2 \;
1607: \delta^{(2)} (\bm{b} + \bm{\rho}_1 - \bm{\rho}_2 ) }
1608: && \nonumber \\
1609: &\times&
1610: \frac{F^2_g (\bm{\rho}_1)}
1611: {\displaystyle \left[ \int d^2 \rho_1' \; F^2_g (\bm{\rho}_1') \right]}
1612: \; \frac{F^2_g (\bm{\rho}_2)}
1613: {\displaystyle \left[ \int d^2 \rho_2' \; F^2_g (\bm{\rho}_2') \right]} ,
1614: \label{P_hard}
1615: \ee
1616: which satisfies
1617: \beq
1618: \int d^2 b \; P_{\text{hard}} (\bm{b}) \;\; = \;\; 1 .
1619: \label{P_hard_normalization}
1620: \eeq
1621: In terms of this distribution the RGS probability (\ref{survb_def})
1622: is expressed as
1623: \beq
1624: S^2 \;\; = \;\;
1625: \int d^2 b \; P_{\text{hard}} (\bm{b}) \; |1 - \Gamma (\bm{b})|^2 .
1626: \label{survb}
1627: \eeq
1628: This result agrees with the expression for the RGS probability derived
1629: heuristically in
1630: Refs.~\cite{Frankfurt:2004kn,Frankfurt:2004ti,Frankfurt:2005mc}
1631: \footnote{Note that the distribution $P_{\text{hard}} (\bm{b})$ of
1632: Eq.~(\ref{P_hard}) (overlap integral of squared spatial distribution
1633: of gluons) is different from the distribution $P_4(\bm{b})$
1634: (square of overlap integral of spatial distribution of gluons),
1635: which was introduced in Ref.~\cite{Frankfurt:2003td} to describe
1636: double dijet production
1637: in inclusive high--energy $pp / \bar p p$ scattering. The two distributions
1638: coincide only in the case of an exponential parametrization of the two--gluon
1639: formfactor, which corresponds to a Gaussian dependence of the
1640: formfactors on transverse momenta and coordinates; see
1641: Eqs.~(\ref{spatial_gaussian}) and (\ref{P_hard_gaussian}).
1642: In general, it is not correct to replace $P_{\text{hard}} (\bm{b})$
1643: by $P_4(\bm{b})$, as was done in
1644: Refs.~\cite{Frankfurt:2004kn,Frankfurt:2004ti}.}.
1645: For the comparison of our result for the RGS probability
1646: with that obtained with the pomeron model of Ref.~\cite{Kaidalov:2003fw}
1647: we refer to Section~\ref{subsec:numerical} below;
1648: see also the comments at the end of Sec.~\ref{subsec:evaluation}.
1649:
1650: Expression~(\ref{survb}) for the RGS probability
1651: allows for a simple probabilistic interpretation. Consider a $pp$ collision
1652: at given impact parameter, $b = |\bm{b}|$. Since the hard two--gluon
1653: exchange process is effectively local in transverse space, the probability
1654: for it to happen is proportional to the product of the
1655: squared transverse spatial distributions of gluons in the two colliding
1656: protons, integrated over the transverse plane,
1657: as given by the numerator of Eq.~(\ref{P_hard}).
1658: Consider now a hypothetical sample of $pp$ events with the
1659: two--gluon induced hard scattering process,
1660: but an otherwise arbitrary (non-diffractive) final state. By the laws of
1661: probability, the distribution of impact parameters in this sample is given
1662: by the normalized distribution $P_{\text{hard}}(\bm{b})$, Eq.~(\ref{P_hard}).
1663: A diffractive event results if the spectator systems of the two protons
1664: do not interact inelastically. The probability for this to happen
1665: in a $pp$ collision at fixed $b$ is given by $|1 - \Gamma (\bm{b})|^2$,
1666: \textit{cf.}\ Eq.~(\ref{P_noin}),
1667: in analogy to the well--known formula for inelastic scattering in
1668: non-relativistic theory \cite{LLIII}. The RGS probability,
1669: which is defined as the fraction of diffractive events in the sample of
1670: all events containing the same hard scattering process, is then
1671: given by the average of this function with the normalized
1672: $b$--distribution in the sample, Eq.~(\ref{survb}).
1673:
1674: It needs to be stressed that the impact parameter of a single
1675: $pp$ event is not observable, being a microscopic quantity beyond the
1676: reach of any experimental apparatus. In the above arguments, the
1677: impact parameter plays the role of a randomly chosen external parameter.
1678: However, using information about the transverse spatial distribution
1679: of gluons in the proton from independent measurements, we can calculate
1680: the probability for certain hard processes in a $pp$ collision as a
1681: function of the impact parameter, and thus infer the distribution
1682: of impact parameters in a sample of events with the
1683: same hard process. This logic was used in Ref.~\cite{Frankfurt:2003td}
1684: to devise a trigger on central collisions in inclusive $pp$ scattering
1685: by requiring hard dijet production at small rapidities. Here we
1686: use the same strategy to model soft spectator interactions
1687: in double--gap exclusive diffractive $pp$ scattering.
1688:
1689: %
1690: % FIGURE
1691: %
1692: \begin{figure}
1693: %
1694: \includegraphics[width=8cm]{survb_gauss.eps}
1695: %
1696: \caption[]{The integrand (impact parameter distribution)
1697: in the RGS probability, Eq.~(\ref{survb}),
1698: for Higgs boson production at the LHC energy. Dashed line:
1699: $b$--distribution of the hard two--gluon exchange, $P_{\text{hard}} (b)$,
1700: Eq.~(\ref{P_hard}), evaluated with the
1701: exponential parametrization of the two--gluon formfactor,
1702: Eq.~(\ref{twogl_exponential}) with $B_g = 3.24\, \text{GeV}^{-2}$.
1703: Solid line: The product $P_{\text{hard}} (b) |1 - \Gamma (\bm{b})|^2$,
1704: evaluated with the exponential parametrization,
1705: Eq.~(\ref{Gamma_gaussian}), with $B = 21.8 \, \text{GeV}^{-2}$.
1706: The vanishing of $|1 - \Gamma (\bm{b})|^2$,
1707: at small $b$, \textit{cf.}\ Fig.~\ref{fig:gamma}, strongly suppresses
1708: contributions from small impact parameters. Note that the plot
1709: shows $2\pi b$ times the functions of impact parameter; the
1710: small--$b$ part of the dashed curve [distribution $P_{\text{hard}} (b)$]
1711: would be close to the left boundary of the plot and was omitted for
1712: better legibility. The RGS probability, $S^2$,
1713: is given by the area under the solid curve.}
1714: \label{fig:survb}
1715: \end{figure}
1716: %
1717: The integrand in Eq.~(\ref{survb}) describes the effective distribution
1718: of impact parameters in a sample of double--gap diffractive events,
1719: and reflects the interplay of hard and soft interactions at the
1720: cross section level. The probability for the hard process,
1721: $P_{\text{hard}}(\bm{b})$, favors small impact parameters, which maximize
1722: the overlap of the large--$x$ gluon distributions in the protons,
1723: and vanishes for $b^2 \gg 1/B_g$. The probability for no inelastic soft
1724: interactions, $|1 - \Gamma (\bm{b})|^2$, favors large impact
1725: parameters, which increase the chances for the protons to stay intact,
1726: and vanishes for $b^2 \ll 1/B$ where $pp$ scattering approaches the
1727: BDL. The product of the two probabilities
1728: is suppressed both at small and at large $b$,
1729: and thus concentrated at intermediate values of $b$.
1730:
1731: This point can nicely be illustrated with the Gaussian
1732: parametrizations of the transverse spatial distribution of gluons,
1733: Eq.~(\ref{spatial_gaussian}), and the $pp$ elastic profile function,
1734: Eq.~(\ref{Gamma_gaussian}). With the Gaussian form (\ref{spatial_gaussian}),
1735: the convolution integral in Eq.~(\ref{P_hard}) can be computed
1736: analytically,
1737: \beq
1738: P_{\text{hard}} (\bm{b}) \;\; = \;\;
1739: \frac{\exp\left[ -\bm{b}^2 / (2B_g) \right]}{2 \pi B_g} .
1740: \label{P_hard_gaussian}
1741: \eeq
1742: This function is shown by the dashed line in Fig.~\ref{fig:survb}.
1743: The integrand of Eq.~(\ref{survb}) is given by
1744: \be
1745: \lefteqn{P_{\text{hard}} (\bm{b}) \; |1 - \Gamma (\bm{b})|^2} &&
1746: \nonumber \\
1747: &=& \frac{1}{2 \pi B_g} \exp \left( -\frac{\bm{b}^2}{2B_g} \right)
1748: \; \left[ 1 - \exp \left( -\frac{\bm{b}^2}{2B} \right) \right]^2 ,
1749: \label{combined_gaussian}
1750: \ee
1751: and is shown by the solid line in Fig.~\ref{fig:survb}.
1752: It is suppressed both for $b^2 \ll 1/B$ (because of the ``blackness'' of
1753: the $pp$ amplitude) and for $b^2 \gg 1/B_g$ (because of the vanishing of
1754: the overlap of the two gluon distributions), and thus concentrated at
1755: intermediate values of $b$. The maximum of $2\pi b$ times
1756: the combined distribution is at
1757: \beq
1758: b^2 \;\; \approx \;\; 5 B_g \hspace{3em} (B_g \ll B) .
1759: \label{b2_max}
1760: \eeq
1761: We see that within our two--scale picture of the transverse structure
1762: of hard and soft interactions, \textit{cf.}\ Fig.~\ref{fig:twoscale}, the
1763: dominant impact parameters in double--gap exclusive diffractive processes
1764: are determined by $B_g$ --- the smaller of the
1765: two areas ---, but may be numerically large because of a
1766: large numerical factor. The RGS probability,
1767: Eq.~(\ref{survb}), is given by the integral of $2\pi b$ times
1768: Eq.~(\ref{combined_gaussian}) (\textit{i.e.}, the area under the
1769: solid curve in Fig.~\ref{fig:survb}), and can be computed
1770: analytically,
1771: \beq
1772: S^2 \;\; = \;\; \frac{2 B_g^2}{(B + B_g) (B + 2 B_g)}
1773: \;\; \approx \;\; \frac{2 B_g^2}{B^2} \hspace{3em} (B_g \ll B) .
1774: \label{S2_exponential_gen}
1775: \eeq
1776: The gap survival probability is of the order $(B_g / B)^2$,
1777: \textit{i.e.}, it is proportional to the square of the ratio
1778: of the transverse area occupied by hard gluons to the area
1779: corresponding to soft interactions. Thus, our two--scale picture
1780: offers a parametric argument for the smallness of the rapidity gap
1781: survival probability.
1782:
1783: The approach to the black--disk limit in $pp$ scattering at
1784: high energies, \textit{i.e.}, the fact that
1785: $\Gamma (\bm{b}) \rightarrow 1$ at small $b$, plays a crucial
1786: role in determining the numerical value of the RGS probability
1787: and ensuring stability of our calculation with respect to
1788: variation of the parameters.
1789: A small deviation of the profile function from unity at $b = 0$,
1790: of the form $\Gamma (\bm{b} = 0) = 1 - \epsilon$ with $\epsilon \ll 1$,
1791: would change the result for the gap survival probability to
1792: \beq
1793: S^2 \;\; \rightarrow \;\; S^2|_{\text{BDL}} \; + \; \epsilon^2
1794: \eeq
1795: [here we have used that $B_g \ll B$, and that the integral of
1796: $P_{\text{hard}}$ is unity, \textit{cf.}\ Eq.~(\ref{P_hard_normalization})].
1797: The approach to the BDL effectively eliminates $\Gamma (\bm{b} = 0)$ as a
1798: parameter in our calculation. We stress again that the experimental
1799: data as well as theoretical arguments indicate that the BDL is indeed
1800: reached in $pp$ scattering at small impact parameters at the LHC energy.
1801: %
1802: \subsection{Numerical estimates}
1803: \label{subsec:numerical}
1804: %
1805: % FIGURE
1806: %
1807: \begin{figure}
1808: %
1809: \includegraphics[width=8cm]{s2_sdep.eps}
1810: %
1811: \caption[]{The RGS probability for double--gap exclusive diffractive processes,
1812: Eq.~(\ref{survb}), as a function of the squared CM energy, $s$.
1813: The Tevatron and LHC energies are marked by arrows.
1814: Shown are the results obtained with the dipole parametrization
1815: of the two--gluon formfactor (\ref{twogl_dipole}), for different
1816: values of the mass parameter, $m_g^2$. The value $m_g^2 = 1 \, \text{GeV}^2$
1817: is appropriate for Higgs boson production at the LHC at central rapidities.
1818: The profile function of the $pp$ elastic amplitude was taken from
1819: Ref.~\cite{Islam:2002au} (\textit{cf.}\ Fig.~\ref{fig:gamma}).}
1820: \label{fig:s2_sdep}
1821: %
1822: \end{figure}
1823: %
1824: For a numerical estimate of the gap survival probability we
1825: evaluate Eq.~(\ref{survb}) with the
1826: dipole parametrization of the two--gluon formfactor,
1827: Eq.~(\ref{twogl_dipole}), and the parametrization of the $pp$
1828: elastic amplitude of Ref.~\cite{Islam:2002au}. For Higgs production
1829: at the LHC ($\sqrt{s} = 14 \, \text{TeV}$)
1830: at central rapidities the momentum fractions of the annihilating
1831: gluons are $x_{1, 2} \sim 10^{-2}$ (at a scale $Q^2 \ll m_H^2$).
1832: For such values of $x$ in principle the contributions of the nucleon's
1833: pion cloud to the gluon density at transverse distances
1834: $\rho \sim 1/(2 M_\pi)$ need to be taken into account; see
1835: Section~\ref{sec:transverse_hard}. As we shall explain
1836: below, these contributions to the
1837: gluon density involve correlations in the nucleon wavefunction,
1838: which effectively reduce their contribution to RGS, and should not
1839: be included in the estimate based on Eq.~(\ref{survb}).
1840: We therefore use in our estimate at the LHC energy the simple dipole
1841: formfactor with $m_g^2 \approx 1\, \text{GeV}^2$, which does not
1842: include the pion cloud contribution. With this choice of parameters
1843: Eq.~(\ref{survb}) gives for the RGS probability for Higgs production
1844: at the LHC
1845: \beq
1846: S^2 \;\; = \;\; 0.027 .
1847: \label{S2_LHC_num}
1848: \eeq
1849: The energy dependence of the RGS probability is shown in
1850: Fig.~\ref{fig:s2_sdep}, for various values of the mass parameter
1851: of the two--gluon formfactor, $m_g^2$. At the Tevatron energy
1852: ($\sqrt{s} = 1.9 \, \text{TeV}$), the gluon momentum fractions
1853: $x_{1, 2}$ are of the order $\sim 10^{-1}$, for which the pion cloud
1854: contributions to the gluon density are naturally absent. While a mass
1855: parameter $m_g^2 = 1 \, \text{GeV}^2$ is still reasonable in this situation,
1856: even higher values $m_g^2$ might be relevant in this case. Taking into
1857: account this effective change of $m_g^2$ with $s$ via the
1858: $x_1, x_2$--dependence, the actual variation of the RGS probability
1859: between the LHC and the Tevatron energies is less pronounced than
1860: it appears from the fixed--$m_g^2$ curves in Fig.~\ref{fig:s2_sdep}.
1861:
1862: %
1863: % FIGURE
1864: %
1865: \begin{figure}
1866: %
1867: \includegraphics[width=8cm]{s2_mdep.eps}
1868: %
1869: \caption[]{Dependence of the RGS probability for double--gap exclusive
1870: diffractive Higgs production at the LHC, Eq.~(\ref{survb}), on the mass
1871: parameter in the dipole parametrization of the two--gluon formfactor,
1872: $m_g^2$, Eq.~(\ref{twogl_dipole}). Also shown are the results obtained
1873: with the exponential parameterization, Eq.~(\ref{twogl_exponential}),
1874: with $B_g = 3.24 / m_g^2$, \textit{cf.}\ Eq.~(\ref{dip_exp}) and
1875: Fig.~\ref{fig_fg_dip_exp}. The profile function of the $pp$ elastic
1876: amplitude in both cases is the one of Ref.~\cite{Islam:2002au}.}
1877: \label{fig:s2_mdep}
1878: %
1879: \end{figure}
1880: %
1881: Figure~\ref{fig:s2_mdep} (solid line) shows the dependence of the
1882: RGS probability at the LHC energy
1883: on the mass parameter in the dipole parametrization of the
1884: two--gluon formfactor, $m_g^2$, Eq.~(\ref{twogl_dipole}).
1885: The value of $S^2$ strongly increases with decreasing $m_g^2$,
1886: \textit{i.e.}, with increasing radius of the
1887: the transverse spatial distribution of gluons,
1888: similar to the behavior found with the simple exponential
1889: parametrizations, Eq.~(\ref{S2_exponential_gen}).
1890: To illustrate the sensitivity of our numerical
1891: predictions to the shape of the two--gluon formfactor,
1892: we also show the results for $S^2$ obtained when replacing the
1893: dipole formfactor by the exponential, Eq.~(\ref{twogl_exponential}),
1894: with $B_g = 3.24 / m_g^2$, \textit{cf.}\ Eq.~(\ref{dip_exp}).
1895: (Figure~\ref{fig:s2_mdep}, dashed line).
1896: One sees that the numerical values are rather different,
1897: in spite of the apparent similarity of the two formfactors over a
1898: wide range, $|t| \lesssim 1 \, \text{GeV}^2$ (\textit{cf.}\
1899: Fig.~\ref{fig_fg_dip_exp}). Comparing the two curves of
1900: Figure~\ref{fig:s2_mdep}, we estimate the uncertainty of our
1901: numerical prediction for the RGS probability due to the uncertainty
1902: of the shape of the two--gluon formfactor to be at least
1903: $\sim 30\%$.
1904:
1905: In a similar way, we can estimate the uncertainty of the RGS
1906: probability due to the uncertainty of the profile of the $pp$ elastic
1907: amplitude. Comparing the numerical values obtained from
1908: Eq.~(\ref{survb}) with the parametrization of Ref.~\cite{Islam:2002au}
1909: and with the simple exponential parametrization, Eqs.~(\ref{A_exponential})
1910: and (\ref{Gamma_gaussian}), for the same two--gluon formfactor,
1911: we again find variations of the order of $\sim 30\%$.
1912:
1913: The relatively high sensitivity of the numerical estimates
1914: to the shape of the two--gluon formfactor and the profile function
1915: of the $pp$ elastic amplitude can be understood as a result of the
1916: peculiar interplay of hard and soft interactions in Eq.~(\ref{survb}).
1917: The different impact parameter dependence of hard and soft interactions
1918: (\textit{cf.}\ Fig.~\ref{fig:survb} and the discussion above)
1919: essentially eliminates contributions from the
1920: regions corresponding to the ``bulk'' of the individual distributions,
1921: $P_{\text{hard}} (\bm{b})$ and $|1 - \Gamma (\bm{b})|$, allowing
1922: for significant strength only in an intermediate region of
1923: impact parameters, where there is considerable sensitivity to
1924: the shape of the distributions. This seems to be a principal
1925: feature of estimates of the RGS probability based on Eq.~(\ref{survb}).
1926:
1927: Detailed numerical studies of the RGS probability were
1928: made within the eikonalized Pomeron model for soft interactions,
1929: see Ref.~\cite{Khoze:2000wk} for a summary. We would like to comment
1930: on these estimates from the perspective of our approach. As we
1931: already noted, the approach of Ref.~\cite{Khoze:2000wk} includes
1932: contributions from inelastic intermediate states (albeit without
1933: introducing non-diagonal ``transition'' GPDs). We have argued that
1934: within the independent interaction approximation these contributions
1935: are strongly suppressed and should not be included,
1936: see Section~\ref{subsec:inelastic}. Nevertheless, the numerical
1937: result for the RGS probability in Higgs boson production in
1938: double--gap diffraction at the LHC quoted in Ref.~\cite{Kaidalov:2003fw},
1939: $S^2 = 0.023$ \footnote{The value $S^2 = 0.026$
1940: quoted in Ref.~\cite{Kaidalov:2003fw} is obtained when taking into
1941: account corrections to the hard scattering amplitude resulting from
1942: the proton transverse momenta, which are not included in our approach.
1943: The difference between the two values is immaterial for the present
1944: discussion.}, is rather similar to our estimate (\ref{S2_LHC_num}).
1945: Note that Ref.~\cite{Kaidalov:2003fw} ascribes an uncertainty of
1946: $\sim 50\%$ to this value; we have estimated a similar uncertainty for
1947: our result due to the combined uncertainties in the profile function
1948: and the two--gluon formfactor (see above). It is interesting to ask
1949: why the results are so similar when the two approaches differ in
1950: their treatment of inelastic diffraction. In order to clarify this
1951: question, we have evaluated our expression for the RGS
1952: probability, Eq.~(\ref{survb}), with the profile function of the $pp$
1953: elastic amplitude obtained within the model of Ref.~\cite{Khoze:2000wk},
1954: which was kindly provided to us by M.~Ryskin; we emphasize that this is
1955: not the same as evaluating the expression for the RGS probability given in
1956: Ref.~\cite{Khoze:2000wk}. Using the exponential form of the two--gluon
1957: formfactor (\ref{twogl_exponential}) with parameters
1958: $B_g = (4, \, 5.5, \, 10.1 ) \, \text{GeV}^{-2}$
1959: we obtain in this way $S^2 = (0.042, \, 0.069, \, 0.157)$, which should be
1960: compared to the results quoted in Ref.~\cite{Khoze:2000wk},
1961: $S^2 = (0.02, \, 0.04, \, 0.11)$. One sees that our Eq.~(\ref{survb})
1962: gives systematically larger values than the approach of
1963: Ref.~\cite{Khoze:2000wk} for the same profile function and the same
1964: two--gluon formfactor. This difference should be attributed to the
1965: effect of inelastic diffraction. What is then quoted as the final estimate
1966: of $S^2$ depends on the preferred value of $B_g$. The RGS probability strongly
1967: decreases with $B_g$, \textit{cf.}\ Eq.~(\ref{S2_exponential_gen})
1968: and Fig.~\ref{fig:s2_mdep}. For Higgs production at the LHC we use a value
1969: of $B_g = 3.24 \, \text{GeV}^{-2}$ (corresponding to a mass
1970: parameter in the dipole parametrization of $m_g^2 = 1\,\text{GeV}^2$),
1971: which is based on analysis of the $J/\psi$ photoproduction data over a
1972: wide range of energies and takes into account the effects of QCD evolution
1973: (\textit{cf}.\ Section~\ref{sec:transverse_hard}).
1974: With this value we obtain $S^2 = 0.030$ for the exponential two-gluon
1975: formfactor and the profile function of Ref.~\cite{Khoze:2000wk}.
1976: This value of $B_g$ is lower than the range of values considered in
1977: Ref.~\cite{Khoze:2000wk} (our $B_g = 2 b$ in the notation of that paper).
1978: The value of $B_g$ for Higgs production at the LHC taken in
1979: Ref.~\cite{Kaidalov:2003fw} is $B_g = 4 \, \text{GeV}^{-2}$,
1980: which results in $S^2 = 0.02$ in their model. One sees that the different
1981: values of $B_g$ partly compensate the differences due to the
1982: treatment of inelastic diffraction in the two approaches.
1983: We thus conclude that the similarity of the final numerical
1984: estimate of Ref.~\cite{Kaidalov:2003fw} with our results
1985: is somewhat accidental. In any case, the differences between
1986: the numerical results of both approaches are within the
1987: estimated uncertainties.
1988:
1989: Potentially more important than the uncertainties of our calculation
1990: of the RGS probability within the independent interaction approximation
1991: are effects of possible correlations between hard and soft interactions.
1992: These effects can naturally be incorporated into our partonic picture,
1993: and further decrease the RGS probability compared to the
1994: independent interaction approximation
1995: (see Section~\ref{subsec:correlations}).
1996: %
1997: \section{Beyond the independent interaction approximation}
1998: \label{sec:beyond}
1999: Our treatment of RGS so far was based on the idea that
2000: hard and soft interactions are approximately independent because
2001: they proceed over widely different time-- and distant scales.
2002: It is clear that this approximation has certain limitations,
2003: concerning both the range of its applicability and its accuracy.
2004: In this Section we discuss various physical effects which
2005: violate the assumption of independence of hard and soft interactions
2006: and give rise to corrections to the estimates of the RGS probability
2007: of Section~\ref{sec:cross_section}. These are (a) the increase of
2008: hard screening corrections with energy, (b) correlations
2009: between hard and soft interactions. These corrections have not been
2010: considered in previous treatments of RGS in
2011: Refs.~\cite{Kaidalov:2003fw,Khoze:1997dr,Khoze:2000cy}.
2012: %
2013: \subsection{Hard screening corrections}
2014: The independent interaction approximation relies on the assumption
2015: of widely different characteristic scales of hard and soft processes.
2016: However, the difference between the two scales tends to decrease,
2017: and may even disappear, at high energies, because of the fast increase
2018: of the amplitudes of hard processes with energy.
2019: One example of this effect is the BDL in $pp$ elastic scattering
2020: at central impact parameters, which can be explained both on the basis
2021: of hard and soft interactions (see Section~\ref{sec:transverse_soft}).
2022: In diffractive scattering, the increase of hard amplitudes at
2023: high energies leads to a reduction of the RGS probability relative
2024: to the estimates presented in Section~\ref{sec:cross_section}.
2025:
2026: One specific mechanism which can lead to qualitative changes of the
2027: hard scattering process at high energies is ``local'' absorption of
2028: the hard scattering amplitude for diffractive production.
2029: This corresponds to the familiar attenuation of the hard gluon exchange
2030: diagram for Higgs production, but by a two--gluon ladder (see
2031: Fig.~\ref{fig:hardscreen}). Here the two--gluon ladder is attached to the
2032: partons which emitted the gluons involved in the Higgs production.
2033: Actually, in order to regularize the infrared divergences present in the
2034: two--gluon ladder we may consider instead the amplitude for the scattering
2035: of two colorless dipoles, in which the typical virtuality of the
2036: ``constituents'' is $k_\perp^2$; a second parton with such virtuality
2037: is anyway present as a result of QCD evolution.
2038: The amplitude of the additional interaction is then suppressed
2039: relative to the original hard amplitude by a factor
2040: $\alpha_s^2 (k_\perp^2 /\tilde k_\perp^2) (x_0/x)^{\lambda}$,
2041: where $\tilde k_\perp$ is the characteristic gluon transverse
2042: momentum in the absorptive ladder. The characteristic
2043: $\tilde k_\perp^2$ increases with energy, and $\lambda \approx 0.2 - 0.25$ at
2044: $Q^2 = 4\, \text{GeV}^2$ \cite{Ciafaloni:2003rd,Altarelli:2003hk}.
2045: Ultimately, the amplitude of the additional interaction would thus
2046: reach a strength comparable to the maximal one (BDL), resulting in
2047: complete suppression of RGS.
2048:
2049: %
2050: % FIGURE
2051: %
2052: \begin{figure}
2053: %
2054: \includegraphics[width=8cm]{hardscreen.eps}
2055: %
2056: \caption[]{A typical diagram describing hard screening corrections
2057: to the hard amplitude for double--gap diffractive production.
2058: Shown is only the generic structure of the partonic ladders;
2059: the dominant contribution comes from gluons.}
2060: \label{fig:hardscreen}
2061: %
2062: \end{figure}
2063: %
2064: A semi-realistic estimate, using the eikonal
2065: approximation for the scattering of the colorless dipoles and
2066: assuming $\lambda \approx 0.2-0.25$, suggests that at the LHC energy
2067: the additional suppression is significant (a factor 2--3)
2068: for Higgs boson production by valence gluons, but much less so
2069: for processes initiated by valence quarks; at the Tevatron energy
2070: the effect should be much less pronounced for both gluons and quarks.
2071: In order to evaluate this additional suppression at LHC energies
2072: quantitatively, one needs to follow the space--time evolution of the
2073: production of small--$x$, large--virtuality partons, and must not restrict
2074: the discussion to the gluon GPDs in the individual protons.
2075:
2076: We emphasize that the screening effect described here is not included
2077: in the definition of the RGS probability due to soft interactions,
2078: as it corresponds to a modification of the probability of
2079: finding two gluons at small transverse distances without
2080: reference to the spectator interactions. Detailed calculations
2081: of the RGS probability including this effect require modeling
2082: of the color connections in the parton wavefunctions of the protons.
2083: Another perturbative screening correction to the hard process,
2084: due to diffraction into high--mass states (with the mass increasing
2085: with the collision energy) was considered in Ref.~\cite{Bartels:2006ea},
2086: which leads to a complementary suppression of the amplitude for exclusive
2087: double--gap diffraction.
2088: %
2089: \subsection{Correlations between hard and soft interactions}
2090: \label{subsec:correlations}
2091: In our treatment of RGS in Section~\ref{subsec:combining} we have not
2092: taken into account effects due to correlations between partons in the
2093: wavefunctions of the colliding protons. Generally speaking, in the
2094: presence of such correlations the selection of particular configurations
2095: by the hard scattering process changes soft interactions as compared to
2096: ``average'' configurations. The neglect of this change is basic
2097: to the approximation (\ref{hard_soft_commute}), in which hard and
2098: soft interactions are assumed to commute also with respect to transverse
2099: degrees of freedom. A consistent treatment of correlation effects requires
2100: a much more detailed description of hard and soft interactions and the
2101: parton--hadron interface than the one given in Section~\ref{sec:gap_survival}.
2102: Here we would like to discuss this problem at the qualitative level,
2103: preparing the ground for a future in-depth investigation.
2104:
2105: As a pedagogical example illustrating the effect of correlations
2106: on the RGS probability, let us consider double--gap diffractive
2107: proton--deuteron ($pd$) scattering --- possibly including the quasi--elastic
2108: channel --- within the framework of Eq.~(\ref{survb}).
2109: The profile function for $pd$ elastic
2110: scattering is significantly smaller than that for $pp$ scattering
2111: even for energies at which $pp$ scattering is close to the black--disk
2112: limit,
2113: \beq
2114: \Gamma^{dp} \;\; \ll \;\; \Gamma^{pp} \;\; \lesssim \;\; 1.
2115: \label{Gamma_deuteron}
2116: \eeq
2117: This follows from Eq.~(\ref{unitarity}) when noting that the $dp$
2118: total cross section is approximately equal to twice the $pp$ cross section,
2119: while the spatial size of the deuteron is much larger than the proton
2120: radius, and can be understood simply
2121: as the result of a ``dilution'' of the blackness of the individual
2122: protons due to their transverse motion in the deuteron bound state.
2123: At the same time, the transverse spatial distribution of gluons is
2124: now characterized by the transverse size of the deuteron.
2125: Because of Eq.~(\ref{Gamma_deuteron}), the factor $|1 - \Gamma^{dp}|^2$
2126: in Eq.~(\ref{survb}) is always of order unity, and no significant
2127: suppression takes place as it does in $pp$ scattering. Thus, one would
2128: conclude that the RGS probability is much larger in $dp$ than in $pp$
2129: scattering, which is clearly nonsensical. The paradox is resolved
2130: by noting that hard and soft interactions in diffractive $pd$ scattering
2131: are highly correlated. By considering events with a hard process
2132: one is effectively selecting configurations
2133: in which the projectile proton scatters from one of the nucleons.
2134: Soft interactions in these configurations are significantly
2135: larger than in $pd$ scattering in average configurations, in which
2136: there is a substantial chance of the projectile ``missing'' the
2137: nucleons in the deuteron \footnote{It is worth noting here that diffraction
2138: in this case does not result from the presence of the fluctuations of the
2139: cross section strength. As a result, the cross section of diffraction is zero
2140: at $t = 0$. At the same time, the diffractive cross section
2141: ($dp\rightarrow pnp$) is not suppressed for $|t| \ge 1/R_d^2$,
2142: illustrating that diffraction cannot be described within the logic
2143: of the eigenstate scattering formalism \cite{Good:1960ba,Miettinen:1978jb}
2144: for finite $t$. Also, there is no simple relation in this case between
2145: the Fourier transform of the diffractive amplitude (impact parameter
2146: representation) and the actual values of $b$ in the process;
2147: the Fourier transform would suggest
2148: $\langle b^2 \rangle \propto 1/B$ [$B$ is the slope of the $pp$
2149: elastic cross section, Eq.~(\ref{B_zero})], while in reality
2150: $\langle b^2 \rangle \propto R_d^2$.}.
2151:
2152: The deuteron example shows that transverse correlations
2153: in the wavefunctions can qualitatively change the picture of RGS.
2154: In particular, positive spatial correlations between hard partons
2155: and the opacity for soft interactions decrease the RGS probability
2156: compared to the uncorrelated estimate based on Eq.~(\ref{survb}).
2157: In connection with $pp$ scattering, Eq.~(\ref{exclusive_diffraction}),
2158: we now discuss the effects of two types of transverse correlations:
2159: (a) long--distance correlations due to scattering from the
2160: proton's long--range pion field (``pion cloud''), (b) short--distance
2161: correlations related to parton clustering in ``constituent quarks''.
2162:
2163: A distinctive contribution to diffractive $pp$ scattering at small
2164: $\bm{p}_{1\perp}'$ results from the process in which a soft pion,
2165: emitted and absorbed by proton 1, scatters diffractively from proton 2.
2166: This contribution could properly be calculated using the known pion--nucleon
2167: coupling, and applying Eqs.~(\ref{amp_momentum})--(\ref{amp_momentum_t})
2168: to the $\pi p$ diffractive amplitude, with the two--gluon
2169: formfactor in the pion and the profile function of $\pi p$ elastic scattering.
2170: It is generally small, for two reasons. First, the coupling
2171: of the soft pion to the nucleon is small because it is the Goldstone boson
2172: of spontaneously broken chiral symmetry. Second, softness of the pion implies
2173: that its longitudinal momentum
2174: fraction in the proton be small, $y \lesssim m_\pi / m_p$. In Higgs
2175: production at the LHC, where $x_{1, 2} \sim 10^{-2}$, this puts the
2176: momentum fraction of the gluons in the pion at relatively large values,
2177: $z \sim x_{1, 2}/y \sim 10^{-1}$, where the gluon distribution
2178: is not enhanced by DGLAP evolution.
2179:
2180: In the partonic picture, the pion cloud contribution represents
2181: the result of specific correlations between hard and
2182: soft partons in the proton wavefunction. This observation has an
2183: interesting consequence for the estimate of the RGS probability
2184: for $pp$ diffractive scattering based on Eq.~(\ref{survb}).
2185: Namely, the gluon distribution in the proton receives a
2186: contribution from the pion cloud at transverse distances
2187: $\rho \sim 1/(2 m_\pi)$ and momentum fractions
2188: $x \lesssim m_\pi / m_N$ \cite{Strikman:2003gz}. Including this
2189: contribution in Eq.~(\ref{survb}) would be inconsistent,
2190: since a hard process involving these gluons in the pion cloud
2191: should be accompanied by a very specific modification of the
2192: soft interactions, which is not accounted for in Eq.~(\ref{survb}).
2193: For this reason, we have not included an explicit pion cloud
2194: contribution in the two--gluon formfactor parametrization used
2195: in our estimate leading to the value (\ref{S2_LHC_num}).
2196: More precisely, if we knew the ``physical'' gluon GPD,
2197: which by definition includes the pion cloud contribution,
2198: we would need to remove this contribution before using
2199: the GPD in Eq.~(\ref{survb}), and thus obtain a lower value
2200: for the RGS probability. This is just another example of the
2201: general rule that correlations lower the
2202: RGS probability compared to the independent interaction
2203: approximation, Eq.~(\ref{survb}).
2204:
2205: %
2206: % FIGURE
2207: %
2208: \begin{figure}
2209: %
2210: \includegraphics[width=8cm]{twocomp_fg_rho2.eps}
2211: %
2212: \caption[]{The two--component parametrization of the
2213: transverse spatial distribution of gluons including the pion cloud,
2214: Eqs.~(\ref{fg_twocomp}) and (\ref{fg_cloud}). Shown are the radial
2215: distributions $2\pi \rho^3 F_g (\rho )$,
2216: the integral of which determines the average gluonic transverse
2217: size, $\langle \rho^2 \rangle$. For the sake of comparison,
2218: the figure shows the combined (bulk + cloud) distribution
2219: without adjustment of the normalization factor ($N = 1$),
2220: so that the bulk contribution is the same as in the
2221: case of no pion cloud.}
2222: \label{fig:twocomp_fg_rho2}
2223: %
2224: \end{figure}
2225: %
2226: To estimate the correction resulting from the removal of the
2227: pion cloud contribution from the gluon GPD, we evaluate
2228: Eq.~(\ref{survb}) with a ``two--component'' parametrization
2229: of the two--gluon formfactor (\textit{viz}.\ the transverse spatial
2230: distribution of gluons) including the pion cloud.
2231: We write
2232: \beq
2233: F_g (\rho ) \;\; = \;\; N \left[ F_{\text{bulk}} (\rho ) \; + \;
2234: F_{\text{cloud}} (\rho ) \right] ,
2235: \label{fg_twocomp}
2236: \eeq
2237: where $N$ is a factor which ensures overall normalization
2238: to $\int d^2\rho F_g(\rho) = 1$. The ``bulk'' contribution to the
2239: two--gluon formfactor we parametrize by the dipole formfactor with
2240: $m_g^2 = 1 \, \text{GeV}^2$, Eq.~(\ref{twogl_dipole}),
2241: the ``cloud'' contribution as
2242: \beq
2243: F_{\text{cloud}} (\rho ) \;\; = \;\;
2244: C_{\text{cloud}} \; \frac{\rho^2}{\rho^2 + \rho_0^2} \;
2245: \frac{e^{-2 m_\pi \rho}}{2 m_\pi \rho} .
2246: \label{fg_cloud}
2247: \eeq
2248: This form is essentially the asymptotic form of the gluon
2249: density at $\rho \gtrsim 1/(2 m_\pi)$ for $x \ll m_\pi / m_N$
2250: (``Yukawa tail'') \cite{Strikman:2003gz}, regularized at
2251: small $\rho$ such as to avoid a large contribution in the
2252: bulk region; the parameter $\rho_0^2$ is chosen of the
2253: order $\langle \rho^2 \rangle_{g, \text{bulk}}
2254: = 8/m_g^2 = 0.3\, \text{fm}^2$. The coefficient $C_{\text{cloud}}$
2255: we determine such that the inclusion of the cloud contribution
2256: increases the overall gluonic transverse size of the proton by 30\% ,
2257: which is the value found in the calculation of
2258: Ref.~\cite{Strikman:2003gz} (based on the phenomenological
2259: gluon distribution in the pion) and supported by the
2260: $J/\psi$ photoproduction data; see Section~\ref{sec:transverse_hard}.
2261: Figure~\ref{fig:twocomp_fg_rho2} shows the transverse spatial distributions
2262: for ``bulk only'' and ``bulk + cloud'', multiplied by
2263: $2\pi \rho^3$, the integral of which determines
2264: $\langle \rho^2 \rangle_g$. Figure~\ref{fig:twocomp_survb} shows
2265: the impact parameter distributions in the RGS probability
2266: (\textit{cf.}\ Fig.~\ref{fig:survb}) obtained in the two cases.
2267: One sees that removal of the pion cloud contribution indeed reduces
2268: the RGS probability. The numerical effect turns out to be rather
2269: small,
2270: \beq
2271: \frac{S^2 \text{(bulk only)}}{S^2 \text{(bulk + cloud)}}
2272: \;\; = \;\; 0.94 .
2273: \eeq
2274: This can be explained by the fact that the pion cloud contribution
2275: to the gluon density is noticeable compared to the bulk only at
2276: transverse distances $\rho \gg 1/(2 m_\pi) = 0.7 \, \text{fm}$,
2277: while the RGS integral is dominated by rather short distances,
2278: $\rho_{\text{eff}} \sim b_{\text{eff}}/2 \sim 0.4 \, \text{fm}$,
2279: \textit{cf}\ Figs.~\ref{fig:coord} and \ref{fig:survb}.
2280: %
2281: % FIGURE
2282: %
2283: \begin{figure}
2284: %
2285: \includegraphics[width=8cm]{twocomp_survb.eps}
2286: %
2287: \caption[]{The integrand (impact parameter distribution)
2288: in the rapidity gap survival probability, Eq.~(\ref{survb}),
2289: for the two--component form of the transverse spatial distribution
2290: of gluons. The parameters are the same as in Fig.~\ref{fig:twocomp_fg_rho2}.}
2291: \label{fig:twocomp_survb}
2292: %
2293: \end{figure}
2294: %
2295:
2296: Corrections to the rapidity gap survival probability as given by
2297: Eq.~(\ref{survb}) can also come from transverse short--distance
2298: correlations in the proton wavefunction (correlation length $\ll$
2299: proton size). The Tevatron CDF data on $\bar pp$ collisions with
2300: multiple hard processes (three jet plus photon production) \cite{Abe:1997bp}
2301: indicate the presence of significant correlations between the
2302: transverse positions of hard partons over distances
2303: $r \sim 0.3\, \text{fm}$ \cite{Frankfurt:2004kn}.
2304: Given one hard parton with $x \gtrsim 0.05$ at a certain transverse position,
2305: it is much more likely to find in the proton wavefunction
2306: a second hard parton $x \gtrsim 0.05$ within a distance $\sim r$
2307: than elsewhere in the transverse plane. As a result, in $pp$ events with
2308: (at least) one hard process the probability for a second hard interaction is
2309: substantially larger (by a factor of $\sim 2$) than it would be without
2310: correlations. One may also suppose that the local cross section density
2311: (opacity) for soft inelastic interactions is higher near
2312: the position of a hard parton than elsewhere.
2313: Such correlations would result in a higher probability for
2314: inelastic interactions in $pp$ events with a hard process
2315: (such as the hard two--gluon exchange in double--gap diffractive
2316: production) as compared to generic $pp$ collisions, and would
2317: thus decrease the RGS probability compared to Eq.~(\ref{survb}).
2318:
2319: The corrections to Eq.~(\ref{survb}) due to short--distance correlations
2320: can be viewed as an effective reduction of the size of the diffractively
2321: scattering system, from the proton radius to the size of the transverse
2322: correlation, $r$. The corrections could be particularly large in
2323: the situation when the correlated areas are ``black spots,'' while the
2324: proton overall is still ``gray'' because of the dilution by
2325: the transverse motion. This situation would in a sense correspond to
2326: the above example of the deuteron,
2327: with the proton now playing the role of the deuteron.
2328: A quantitative description of these effects would require
2329: detailed modeling of the correlation between hard partons
2330: and the opacity for soft inelastic interactions, including an
2331: analysis of $pp$ elastic scattering allowing local fluctuations
2332: in opacity, which are outside of the scope of the present paper.
2333:
2334: A model of the proton accounting for short--distance correlations
2335: between partons is the so--called chiral quark--soliton model
2336: \cite{Diakonov:1987ty}, which describes the proton as a system of
2337: constituent quarks bound by a classical pion field,
2338: see Ref.~\cite{Diakonov:2002fq} for a review. This model implements
2339: the short--distance scale related to the spontaneous breaking of
2340: chiral symmetry, which appears here as the ``size'' of the constituent
2341: quark, $r \sim 0.3\, \text{fm}$ (in the Euclidean formulation of QCD
2342: this scale can be associated with the average instanton size in the vacuum).
2343: This model provides a consistent description of the
2344: twist--2 quark-- and antiquark distributions in the nucleon
2345: at the scale $\mu \sim r^{-1} \approx 600\, \text{MeV}$ \cite{Diakonov:1996sr}.
2346: It also suggests that the gluons at this scale are ``packaged'' inside the
2347: constituent quarks and antiquarks \cite{Diakonov:2002fq}. Assuming
2348: perturbative QCD evolution to be applicable starting from this scale
2349: one would thus expect significant correlations between the positions
2350: of hard quarks and gluons and the opacity for soft interactions
2351: over a transverse size $r \sim 0.3 \, \text{fm}$.
2352: Incorporating these correlations
2353: into the description of the BDL in high--energy $pp$ scattering
2354: and the theory of RGS in diffractive processes is an important
2355: problem for future studies.
2356:
2357: To summarize, the dynamical mechanisms which we have discussed here
2358: (and indeed all mechanism which we are aware of) give rise to
2359: positive correlations between the transverse position of hard partons
2360: and the opacity for inelastic interactions. We can thus say with some
2361: confidence that estimates based on the independent interaction
2362: approximation, Eq.~(\ref{survb}), represent an upper bound
2363: on the RGS probability.
2364: %
2365: \section{Transverse momentum dependence}
2366: \label{sec:differential}
2367: The interplay of hard and soft interactions in exclusive double--gap
2368: diffraction not only causes the suppression of the integrated
2369: cross section summarized in the RGS probability, but also
2370: gives rise to a distinctive dependence of the cross section
2371: on the final proton transverse momenta. By observing this
2372: ``diffraction pattern'' one can perform detailed tests of the
2373: diffractive reaction mechanism, and even extract information
2374: about the two--gluon formfactors of the colliding protons.
2375: The transverse momentum dependence also contains information
2376: about the quantum numbers (parity) of the produced system
2377: \cite{Kaidalov:2003fw}. We consider here production of a $0^+$ system,
2378: for which the hard scattering amplitude depends on the proton transverse
2379: momenta only through $|\bm{p}_{1\perp}'|$ and $|\bm{p}_{2\perp}'|$,
2380: and the diffractive amplitude is given by Eq.~(\ref{amp_momentum});
2381: we comment on production of $0^-$ systems at the end of this section.
2382:
2383: %
2384: % FIGURE
2385: %
2386: \begin{figure*}
2387: %
2388: \includegraphics[width=16.5cm]{diffp.eps}
2389: %
2390: \caption[]{Transverse momentum dependence of the cross section
2391: for double--gap exclusive diffraction (\ref{exclusive_diffraction}).
2392: The plots show the modulus squared of the amplitude, $|T_\text{diff}|^2$,
2393: with $\kappa = 1$, as a function of $\bm{p}_{2\perp}'$,
2394: for three fixed values of $\bm{p}_{1\perp}'$ (chosen to point in
2395: $x$--direction, $p_{1y}' = 0$). The upper row of plots
2396: show the contours of constant values of $|T_\text{diff}|^2$
2397: on a logarithmic scale as functions of $p_{2x}'$ and $p_{2y}'$
2398: (in units of GeV). The lower row of plots shows the profile
2399: along the $p_{2y}' = 0$ axis. Shown are the results calculated
2400: with the exponential parametrizations of the two--gluon formfactor
2401: and the $pp$ elastic amplitude, Eq.~(\ref{T_diff_gauss_p1_p2}).
2402: The diffraction pattern in $\bm{p}_{2\perp}'$ evolves from a
2403: rotationally symmetric one for $\bm{p}_{1\perp}' = 0$ to a
2404: two--centered one for large $|\bm{p}_{1\perp}'|$.}
2405: \label{fig:diffp}
2406: \end{figure*}
2407: %
2408: For a quick orientation over the transverse momentum dependence,
2409: we evaluate the amplitude
2410: with the exponential parametrizations of the two--gluon formfactor,
2411: Eqs.~(\ref{twogl_exponential}), and the $pp$ elastic
2412: scattering amplitude, Eq.~(\ref{Gamma_gaussian}).
2413: In this case the convolution integral
2414: in Eq.~(\ref{amp_momentum}) reduces to a Gaussian integral,
2415: and we obtain a closed expression for the amplitude,
2416: \be
2417: \lefteqn{T_{\text{diff}} (\bm{p}_{1\perp}', \bm{p}_{2\perp}') } &&
2418: \nonumber \\[1ex]
2419: &=&
2420: \kappa \; \exp \left( - \frac{B_{g1} \bm{p}_{1\perp}^{'2}}{2}
2421: - \frac{B_{g2} \bm{p}_{2\perp}^{'2}}{2} \right)
2422: \nonumber \\
2423: &\times &
2424: \left\{ 1 - \frac{B}{B_{\text{tot}}}
2425: \exp \left[ \frac{(B_{g1} \, \bm{p}_{1\perp}'
2426: - B_{g2} \, \bm{p}_{2\perp}')^2}
2427: {2 B_{\text{tot}}} \right]
2428: \right\} ,
2429: \label{T_diff_gauss_p1_p2}
2430: \ee
2431: where $B_{g1} \equiv B_g (x_1)$ and $B_{g2} \equiv B_g (x_2)$
2432: are the slopes corresponding to the momentum fractions
2433: $x_{1, 2} \sim \xi_{1, 2}$, and
2434: \beq
2435: B_{\text{tot}} \;\; \equiv \;\; B_{g1} + B_{g2} + B .
2436: \eeq
2437: [As a check, we note that the integral of the square of
2438: this expression, divided by the corresponding expression for
2439: $B = 0$ (no soft interactions), reproduces the result for the RGS
2440: probability obtained in the coordinate space calculation,
2441: Eq.~(\ref{S2_exponential_gen}).] The amplitude (\ref{T_diff_gauss_p1_p2})
2442: vanishes trivially for large $|\bm{p}_{1\perp}|$ or $|\bm{p}_{2\perp}|$,
2443: independently of the directions of the momentum vectors.
2444: In addition, it has a zero at finite values of the transverse
2445: momenta, namely at
2446: \be
2447: \lefteqn{(B_{g1} \, \bm{p}_{1\perp}' - B_{g2} \, \bm{p}_{2\perp}')^2 }
2448: &&
2449: \nonumber \\
2450: &=& 2 B_{\text{tot}}
2451: \; \ln \frac{B_{\text{tot}}}{B}
2452: \nonumber \\
2453: &\approx & 2 (B_{g1} + B_{g2})
2454: \hspace{3em} (B \gg B_{g1}, B_{g2}) .
2455: \label{zero}
2456: \ee
2457: This zero arises because of the destructive interference of
2458: the amplitude of the hard scattering process alone and the
2459: amplitude including soft elastic rescattering,
2460: \textit{cf.}\ Fig.~\ref{fig:mom}, and directly reflects
2461: the interplay of hard and soft interactions. It leads to a dip
2462: in the diffractive cross section, and thus to a typical
2463: ``diffraction pattern'' in the dependence on the transverse momentum
2464: of the first proton, $\bm{p}_{1\perp}'$, at fixed transverse momentum
2465: of the second proton, $\bm{p}_{2\perp}'$. Figure~\ref{fig:diffp}
2466: shows this diffraction pattern in the kinematics of Higgs production
2467: at the LHC at zero rapidity, for which $B_{g1} = B_{g2}$.
2468: One sees that the pattern
2469: in $\bm{p}_{2\perp}'$ evolves from a rotationally symmetric one for
2470: $\bm{p}_{1\perp}' = 0$ to a two--centered one for large $|\bm{p}_{1\perp}'|$.
2471: This basic feature of the diffractive cross section does not depend on the
2472: details of the parametrization of the two--gluon formfactor; similar
2473: results are obtained with the dipole parametrization,
2474: \textit{cf.}\ the detailed comparison below.
2475:
2476: %
2477: % FIGURE
2478: %
2479: \begin{figure}
2480: %
2481: \includegraphics[width=8cm]{phidepn.eps}
2482: %
2483: \caption[]{Angular dependence of the cross section
2484: for double--gap diffractive production of a $0^+$ system.
2485: The plot shows the modulus squared of the amplitude,
2486: $|T_{\text{diff}}|^2$, Eq.~(\ref{amp_momentum}), with $\kappa = 1$.
2487: The numbers above the curves indicate the values of
2488: $|\bm{p}_{1\perp}'| = |\bm{p}_{1\perp}'|$. The kinematics
2489: corresponds to Higgs boson production at the LHC.}
2490: \label{fig:phidep}
2491: %
2492: \end{figure}
2493: %
2494: The diffraction pattern of Fig.~\ref{fig:diffp} implies a
2495: strong angular dependence of the cross section, which, moreover,
2496: changes with the magnitude of the transverse momenta,
2497: $|\bm{p}_{1\perp}'|$ and $|\bm{p}_{2\perp}'|$. This is illustrated
2498: in Figure~\ref{fig:phidep}, which shows the dependence
2499: of $|T_{\text{diff}}|^2$ on the
2500: angle between the transverse momenta, $\phi$, for various
2501: values of $|\bm{p}_{1\perp}'| = |\bm{p}_{2\perp}'|$.
2502: For small values of the transverse momenta the cross section
2503: is maximal at zero angle; for large values (where the cross sections
2504: as a function of angle runs through the diffractive dip) it is maximal
2505: at $\phi = \pi$. This dependence needs to be taken into account when
2506: attempting to maximize the diffractive cross section in the search
2507: for new particles.
2508:
2509: %
2510: % FIGURE
2511: %
2512: \begin{figure*}
2513: %
2514: \begin{tabular}{cc}
2515: \includegraphics[width=7cm]{rdep.eps}
2516: &
2517: \includegraphics[width=7cm]{pdep.eps}
2518: \end{tabular}
2519: %
2520: \caption[]{\textit{Left:}
2521: Dependence of the cross section for double--gap diffractive
2522: $0^+$ production on the square of the difference of the proton transverse
2523: momenta, $\bm{r}_\perp^2$, for $\bm{P}_\perp = 0$ (\textit{i.e.},
2524: $\bm{p}_{1\perp}' = -\bm{p}_{2\perp}' = 2\bm{r}_\perp$).
2525: The kinematics corresponds to production of a system with
2526: $M = 140\, \text{GeV}$ at the LHC energy at zero rapidity
2527: ($x_1 = x_2 \sim 10^{-2}$).
2528: The plot shows the modulus squared of the amplitude,
2529: $|T_{\text{diff}}|^2$, Eq.~(\ref{amp_momentum}), with $\kappa = 1$,
2530: obtained with the exponential and dipole parametrizations of
2531: the two--gluon formfactor. For the exponential parametrization,
2532: the position of the
2533: diffractive zero is given by Eq.~(\ref{zero}).
2534: \textit{Right:} Dependence on the square of the sum of the
2535: transverse momenta, $\bm{P}_\perp^2$, for $\bm{r}_\perp = 0$ (\textit{i.e.},
2536: $\bm{p}_{1\perp}' = \bm{p}_{2\perp}' = \bm{P}_\perp$).
2537: Also shown is $F_g^4 (t)$ for the
2538: two parametrizations, which would be the $t$--dependence of the
2539: amplitude without soft spectator interactions.}
2540: \label{fig:rpdep}
2541: %
2542: \end{figure*}
2543: %
2544: An interesting questions is which specific features of the transverse
2545: momentum dependence of the diffractive cross section could be used to
2546: test the diffractive reaction mechanism, and possibly extract information
2547: about the two--gluon formfactors of the colliding protons. Such studies
2548: would be feasible in diffractive dijet production, which has a relatively
2549: large cross section and also allows one to vary the invariant mass of the
2550: diffractively produced system, and thus the effective values of the gluon
2551: momentum fraction, $x_1$ and $x_2$. To address this question,
2552: we again start from the explicit expression for the amplitude obtained
2553: with the exponential parametrization of the two--gluon formfactor,
2554: Eq.~(\ref{T_diff_gauss_p1_p2}). We rewrite it in terms of
2555: the center--of--mass and relative transverse momenta of the
2556: final--state protons,
2557: \be
2558: \bm{P}_\perp &\equiv& (\bm{p}_{1\perp}' + \bm{p}_{2\perp}')/2,
2559: \nonumber \\
2560: \bm{r}_\perp &\equiv& \bm{p}_{1\perp}' - \bm{p}_{2\perp}' ,
2561: \ee
2562: and obtain
2563: \be
2564: \lefteqn{T_{\text{diff}} (\bm{p}_{1\perp}', \bm{p}_{2\perp}') } &&
2565: \nonumber \\[1ex]
2566: &=&
2567: \kappa \; \exp \left[ - \frac{B_{g1} + B_{g2}}{2} \left(
2568: \bm{P}_\perp^2 + \frac{\bm{r}_\perp^2}{4} \right)
2569: \right. \nonumber \\
2570: && \left.
2571: \phantom{\exp} \;
2572: - \frac{B_{g1} - B_{g2}}{2} (\bm{P}_\perp \bm{r}_\perp) \right]
2573: \nonumber \\
2574: &\times &
2575: \left\{ 1 - \frac{B}{B_{\text{tot}}}
2576: \exp \left[
2577: \frac{(B_{g1} - B_{g2})^2}{2 B_{\text{tot}}} \bm{P}_\perp^2
2578: \right.\right.
2579: \nonumber \\
2580: && + \frac{(B_{g1} + B_{g2})^2}{2 B_{\text{tot}}} \frac{\bm{r}_\perp^2}{4}
2581: \nonumber \\
2582: && \left. \left.
2583: + \frac{(B_{g1} - B_{g2}) (B_{g1} + B_{g2})}{2 B_{\text{tot}}}
2584: (\bm{P}_\perp \bm{r}_\perp) \right]
2585: \right\} .
2586: \label{T_diff_P_r}
2587: \ee
2588: For production at zero rapidity, for which $x_1 = x_2 \equiv x$,
2589: and $B_{g1} = B_{g2} \equiv B_g$, this simplifies to
2590: \be
2591: \lefteqn{T_{\text{diff}} (\bm{p}_{1\perp}', \bm{p}_{2\perp}') } &&
2592: \nonumber \\[1ex]
2593: &=&
2594: \exp \left( -B_g \bm{P}_\perp^2 - \frac{B_g \bm{r}_\perp^2}{4} \right)
2595: \nonumber \\
2596: &\times &
2597: \left[ 1 - \frac{B}{B_{\text{tot}}}
2598: \exp \left(
2599: \frac{B_g^2}{B_{\text{tot}}} \bm{r}_\perp^2 \right) \right] ,
2600: \label{T_diff_P_r_sym}
2601: \ee
2602: where now $B_{\text{tot}} = B + 2 B_g$. In this case the amplitude
2603: does not depend on the variable $(\bm{P}_\perp \bm{r}_\perp)
2604: = (\bm{p}_{1\perp}^{'2} - \bm{p}_{2\perp}^{'2})/2$, which is
2605: an obvious consequence of its symmetry with respect to the
2606: interchange of $\bm{p}_{1\perp}^{'}$ and $\bm{p}_{2\perp}^{'}$.
2607: Furthermore, one sees that the dependences of Eq.~(\ref{T_diff_P_r_sym})
2608: on $\bm{P}_\perp^2$ and $\bm{r}_\perp^2$ are very different.
2609: The dependence on $\bm{P}_\perp^2$ is monotonic and governed by
2610: the parameter $B_g$ alone; it essentially
2611: probes the square of the two--gluon formfactors of the colliding protons.
2612: The dependence on $\bm{r}_\perp^2$, however, is governed by both
2613: $B_g$ and $B$, and exhibits the diffractive zero (\ref{zero});
2614: it reflects the interplay of hard and soft interactions.
2615:
2616: The qualitative differences between the $\bm{P}_\perp$ and $\bm{r}_\perp$
2617: dependence of the amplitude are not specific to the exponential
2618: parametrization of the two--gluon formfactor, and can be used
2619: to test the diffractive reaction mechanism and extract information
2620: about the two--gluon formfactor. Figure~\ref{fig:rpdep} (left plot)
2621: shows the squared modulus of the amplitude (for $\kappa = 1$)
2622: as a function of $\bm{r}_\perp^2$,
2623: for $\bm{P}_\perp = 0$, in the kinematics corresponding to
2624: dijet production with $x = 10^{-2}$ at the LHC. One sees that the
2625: dependence in the forward peak (near $\bm{r}_\perp^2 = 0$), and even
2626: the position of the diffractive dip, are not very different for the
2627: two parametrizations. Experimental observation of this structure would
2628: thus constitute a stringent test of the diffractive reaction mechanism.
2629:
2630: Figure~\ref{fig:rpdep} (right plot)
2631: shows the dependence of the cross section for
2632: double--gap diffractive $0^+$ production on $\bm{P}_\perp^2$, for
2633: $\bm{r}_\perp = 0$. Also shown is the fourth power of the
2634: two--gluon formfactor, $F_g^4 (t)$ for the two parametrizations,
2635: which would be the $t$--dependence of the amplitude without soft
2636: spectator interactions. For the exponential parametrization the
2637: $t$--dependence of the full diffractive amplitude is identical
2638: to that of $F_g^4 (t)$, \textit{cf.}\ Eqs.~(\ref{T_diff_P_r_sym})
2639: and (\ref{twogl_exponential}); one sees that the two dependences
2640: are similar also for the dipole parametrization. The different
2641: normalization of the two sets of curves reflects the RGS probability,
2642: \textit{cf.}\ Sec.~\ref{subsec:geometry}.
2643:
2644: The position of the diffractive zero in the $\bm{r}_\perp^2$--dependence
2645: (see Fig.~\ref{fig:rpdep}, left plot) is correlated with the slope of the
2646: monotonic $\bm{P}_\perp^2$--dependence of the diffractive cross section;
2647: both essentially reflect the two--gluon formfactor of the colliding
2648: protons. A sensible strategy for the analysis of double--gap
2649: diffractive dijet
2650: production would be to first establish the existence of the
2651: diffractive zero in $\bm{r}_\perp^2$ at $\bm{P}_\perp = 0$,
2652: and then extract $B_g$ from the $\bm{P}_\perp^2$--dependence of the
2653: diffractive cross section at $\bm{r}_\perp = 0$, where the
2654: cross section is maximal. In the next step, one could change the
2655: dijet mass (\textit{i.e.}, the momentum fractions $x_1 = x_2 = x$
2656: in the two--gluon formfactor) and verify whether both the position
2657: of the zero and the $\bm{P}_\perp^2$--slope change proportionately,
2658: and whether the rate of change with $\ln x$ is consistent with
2659: the value of $\alpha'_g$ at the relevant scale, \textit{cf.}\
2660: Eq.~(\ref{alpha_g}).
2661:
2662: The $x_{1, 2}$--dependence of the two--gluon formfactor in the protons
2663: can be probed more directly by extending the measurements of diffractive
2664: production to non-zero rapidity, $y \neq 0$, corresponding to different
2665: momentum fractions of the annihilating gluons in the two protons,
2666: \beq
2667: x_{1, 2} \;\; = \;\; x \, e^{\pm y} ,
2668: \hspace{3em} x \;\; \equiv \;\; \frac{m_H}{\sqrt{s}}.
2669: \label{x_1_2_from_y}
2670: \eeq
2671: In this case the two--gluon formfactors of the two protons are
2672: different, and the diffractive cross section is no longer
2673: invariant under exchange of the final proton transverse momenta,
2674: $\bm{p}_{1\perp}' \leftrightarrow \bm{p}_{2\perp}'$. One sees that
2675: the expression for the amplitude obtained with the exponential
2676: parametrization, Eq.~(\ref{T_diff_P_r_sym}), acquires a
2677: dependence on $(\bm{P}_\perp \bm{r}_\perp) =
2678: (\bm{p}_{1\perp}^{'2} - \bm{p}_{2\perp}^{\prime 2})/2$, which is controlled
2679: by the difference of the slopes, $B_{g1} - B_{g2}$. With the
2680: $x$--dependence of the individual slopes given by
2681: \be
2682: B_{g1} &\equiv& B_g (x_1) \;\; = \;\;
2683: B_g (x) + 2\alpha'_g \ln \frac{x_1}{x} , \\
2684: B_{g2} &\equiv& B_g (x_2) \;\; = \;\;
2685: B_g (x) + 2\alpha'_g \ln \frac{x_2}{x} ,
2686: \ee
2687: their difference is directly proportional to the rapidity
2688: [\textit{cf.}\ Eq.~(\ref{x_1_2_from_y})],
2689: \beq
2690: B_{g1} - B_{g2} \;\; = \;\; 2 \alpha'_g y .
2691: \eeq
2692: For small rapidities, $|y| \lesssim 1$, and because of the relatively
2693: small value of $\alpha'_g$ this difference is substantially smaller
2694: than the central value, $B_g (x)$, and the dependence of the
2695: amplitude on $(\bm{P}_\perp \bm{r}_\perp)$
2696: can be treated in first--order approximation.
2697: This implies that the cross section depends practically
2698: linearly on $(\bm{P}_\perp \bm{r}_\perp)$ for $|y| \lesssim 1$.
2699:
2700: %
2701: % FIGURE
2702: %
2703: \begin{figure}
2704: %
2705: \begin{tabular}{c}
2706: \includegraphics[width=8cm]{asym_025.eps} \\
2707: \includegraphics[width=8cm]{asym_05.eps}
2708: \end{tabular}
2709: %
2710: \caption[]{The cross section asymmetry (\ref{asym}) as a function of
2711: rapidity, $y$. Shown are the results obtained with the exponential and the
2712: dipole parametrizations of the two--gluon formfactor,
2713: for two representative values of $\alpha'_g$. In this calculation only
2714: the $y$--dependence of the cross section arising from the convolution
2715: integral is taken into account, not the $y$--dependence arising from the
2716: overall normalization, $\kappa$.}
2717: \label{fig:asym}
2718: %
2719: \end{figure}
2720: %
2721: A convenient observable to measure is the asymmetry of the
2722: cross section with respect to $\bm{r}_\perp \rightarrow -\bm{r}_\perp$
2723: and $\bm{P}_\perp \rightarrow \bm{P}_\perp$ (\textit{i.e.},
2724: $\bm{p}_{1\perp}' \leftrightarrow \bm{p}_{2\perp}'$) at fixed
2725: $x$ and $y \neq 0$ (\textit{i.e.}, fixed $x_1 \neq x_2$),
2726: \be
2727: A &\equiv&
2728: \frac{
2729: \sigma_{\text{diff}}(\bm{p}_{1\perp}', \bm{p}_{2\perp}') -
2730: \sigma_{\text{diff}}(\bm{p}_{2\perp}', \bm{p}_{1\perp}')}
2731: {\sigma_{\text{diff}}(\bm{p}_{1\perp}', \bm{p}_{2\perp}') +
2732: \sigma_{\text{diff}}(\bm{p}_{2\perp}', \bm{p}_{1\perp}')} ;
2733: \label{asym}
2734: \ee
2735: alternatively, one could exchange the rapidities and leave the
2736: transverse momenta the same. This asymmetry is odd in $y$ and vanishes
2737: linearly for $y \rightarrow 0$. For $|y| \lesssim 1$, it is practically
2738: is proportional to $(\bm{P}_{\perp}\bm{r}_{\perp}) =
2739: (\bm{p}_{1\perp}^{\prime 2} - \bm{p}_{2\perp}^{\prime 2})/2$.
2740: When calculating the asymmetry at finite $y$, we have to take into
2741: account that, in general, also the overall normalization of the cross
2742: section, $\kappa (s, \xi_1, \xi_2)$, changes with $y$, because of the
2743: change of arguments in the $t = 0$ gluon GPDs of the protons,
2744: see Eq.~(\ref{kappa}). However, this change relative to the
2745: value at $y = 0$ is of second order in $y$ (the changes in the
2746: arguments of the gluon densities cancel each other to first order)
2747: and can be neglected for small $y$. This implies that for $y \ll 1$.
2748: the asymmetry is of the form
2749: \be
2750: A &\sim & C y \alpha'_g (\bm{P}_{\perp}\bm{r}_{\perp}) ,
2751: \ee
2752: where the constant, $C$, is calculable solely from the convolution
2753: integral of the two--gluon formfactor and the $pp$ elastic amplitude,
2754: and does not contain information on the gluon densities. For finite $y$,
2755: the asymmetry is still proportional to
2756: $\alpha'_g (\bm{P}_{\perp}\bm{r}_{\perp})$, but the coefficient
2757: is a more complicated function of $y$, which depends also on the
2758: gluon densities in the colliding protons.
2759: Figure~\ref{fig:asym} shows the theoretical
2760: estimate of the asymmetry as obtained with the exponential and the dipole
2761: parametrizations of the two--gluon formfactor, as a function of the rapidity,
2762: $y$, for two representative values of $\alpha'_g$. In this calculation,
2763: for simplicity, we have taken into account only the $y$--dependence
2764: of the cross section arising from the RGS integral, not the
2765: $y$--dependence arising from the overall normalization, $\kappa$.
2766: The latter is $\propto y^2$ at small $y$ but may be numerically
2767: important at $y \sim 1$; the curves in this region are shown
2768: for illustrative purposes only.
2769:
2770: In the above discussions we have considered production of a $0^+$ system.
2771: The cross section for production of a $0^{-}$ state is significantly
2772: suppressed compared to $0^+$. This is because in the hard scattering process
2773: only one gluon polarization state gives a large contribution in the
2774: LO approximation \cite{Gribov:2003nw}, and from one gluon polarization
2775: it is impossible to build a parity--conserving amplitude for the production
2776: of a $0^{-}$ state.
2777:
2778: Our discussion of the transverse momentum dependence in this Section
2779: is based on the approximation of independent hard and soft interactions,
2780: Eqs.~(\ref{amp_momentum}--\ref{T_rho}). One should expect that the
2781: inclusion of correlations between hard and soft interactions,
2782: as described in Section~\ref{sec:beyond}, would modify not
2783: only the RGS probability but also the transverse momentum
2784: dependence of double--gap exclusive diffraction. This interesting
2785: question will be addressed elsewhere.
2786: %
2787: \section{Discussion and outlook}
2788: \label{sec:discussion}
2789: In this paper we have outlined an approach to RGS in double--gap
2790: exclusive diffractive processes in $pp$ scattering based on the
2791: idea that hard and soft interactions are approximately independent
2792: because they proceed over widely different time--and distance scales.
2793: We have shown that this idea can be practically implemented in
2794: the framework of Gribov's parton picture of high--energy scattering,
2795: and gives rise to a conceptually clear and quantitative description
2796: of RGS.
2797:
2798: In the independent interaction approximation, where correlations
2799: between hard and soft interactions are neglected,
2800: the RGS probability can be expressed in a model--independent fashion
2801: in terms in two phenomenological ingredients --- the gluon GPD in the
2802: proton, and the $pp$ elastic scattering amplitude. At this level we
2803: recover a simple geometric picture of the interplay of hard and soft
2804: interactions in the impact parameter representation. The fact that the
2805: $pp$ elastic amplitude at high energies approaches the BDL at TeV
2806: energies suppresses small impact parameters and ensures
2807: dominance of peripheral scattering in double--gap diffraction.
2808: This is crucial both for justifying the approximations made
2809: in our derivation, and for determining the numerical value
2810: of the RGS probability. Our numerical results for the RGS probability
2811: are somewhat lower than those obtained previously within the eikonalized
2812: Pomeron model of soft interactions; the agreement of Eq.~(\ref{S2_LHC_num})
2813: with the best estimate quoted in Ref.~\cite{Kaidalov:2003fw} is accidental
2814: and due to the fact that these authors assume a larger radius of the
2815: transverse spatial distribution of gluons with $x \sim 10^{-2}$.
2816:
2817: Our numerical prediction for the RGS probability in double--gap exclusive
2818: diffractive Higgs boson production at the LHC ($m_H = 100 - 200 \, \text{GeV},
2819: \sqrt{s} = 14\, \text{TeV}$) in the independent interaction approximation
2820: is $S^2 \approx 0.03$. According to the detailed calculations of
2821: Ref.~\cite{Kaidalov:2003fw},
2822: this would put the estimated cross section for Higgs production
2823: in the minimal supersymmetric model (MSSM) with detection through
2824: the $b\bar b$ mode at $\text{Br}(H \rightarrow b\bar b) \,
2825: \sigma_{\text{diff}}(0^+) \sim \text{few} \times 10 \, \text{fb}$,
2826: which should be accessible at the LHC; see Ref.~\cite{Kaidalov:2003fw}
2827: for details.
2828:
2829: Our partonic approach to RGS also allows us to discuss the effect
2830: of correlations between hard and soft interactions on the RGS
2831: probability. Such correlations lower the RGS probability compared
2832: to the independent interaction approximation. In the case of
2833: long--distance correlations due to the proton's pion cloud,
2834: we estimated this effect to be of the
2835: order of $\sim 10\%$. Potentially more important are short--distance
2836: correlations, \textit{e.g.}\ those suggested by the phenomenological
2837: concept of a ``constituent quark'' structure of the proton).
2838: Such correlations can be regarded as a
2839: change of the effective size of the diffractively scattering systems,
2840: and could reduce the predictions for the RGS probability by a
2841: substantial factor. In view of its importance for the Higgs boson
2842: search at the LHC this problem clearly requires further theoretical
2843: study. It can in principle also be addressed experimentally,
2844: by ``measuring'' the RGS probability in double--gap processes for which
2845: the hard scattering amplitude is known, such as dijet production.
2846:
2847: An important ingredient in our description of RGS is the
2848: profile function of the complex $pp$ elastic scattering amplitude.
2849: This underscores the importance of the planned measurements of $pp$
2850: elastic scattering and total cross sections in the TOTEM experiment
2851: at the LHC \cite{unknown:1997xu},
2852: as well as at RHIC \cite{Bultmann:2005na}. In addition to providing input
2853: for more accurate estimates of the RGS probability in diffraction,
2854: such measurements would allow us to further explore the fascinating
2855: new regime of the BDL in high--energy hadron--hadron scattering.
2856:
2857: Measurements of the transverse momentum dependence of double--gap
2858: exclusive diffractive processes with large cross section
2859: (dijet production) would allow one to perform detailed studies of the
2860: diffractive reaction mechanism. Following the strategy outlined in
2861: Section~\ref{sec:differential}, once the reaction mechanism has
2862: been established, one could even use such processes to extract information
2863: about the transverse spatial distribution of gluons in the colliding
2864: protons, including its change with $x$. Such studies would complement
2865: the information on the two--gluon formfactor obtained from vector meson
2866: production at HERA or a future electron--ion collider (EIC). Eventually,
2867: using QCD evolution as well as models of nucleon structure,
2868: these data on the transverse spatial distribution of gluons could
2869: also be correlated with the planned measurements of quark GPDs
2870: in hard exclusive processes in $ep$ scattering in fixed--target
2871: experiments (HERMES, JLab 12 GeV, COMPASS). One of the
2872: advantages of Gribov's parton picture of hard and soft interactions
2873: is precisely that it unifies the description of hadron--hadron
2874: and electron/photon--hadron scattering at high energies.
2875: Other ways to probe GPDs in $pp$ scattering with hard processes
2876: (non-diffractive) have been described in Ref.~\cite{Frankfurt:2004kn}.
2877:
2878: We would like to comment on some of the experimental
2879: aspects of measurements of the transverse momentum
2880: dependence of double--gap exclusive diffraction with the
2881: proposed forward detectors at the LHC. Such measurements require
2882: good energy resolution in order to guarantee exclusivity
2883: and determine the mass of the diffractively produced system,
2884: as well as sufficient transverse momentum resolution to
2885: map the $\bm{p}_{1\perp}, \bm{p}_{2\perp}$ distributions.
2886: An important experimental problem is that the intrinsic transverse
2887: momentum distribution in the beams at the interaction point (IP)
2888: puts a lower bound on the transverse momentum transfers that can
2889: be resolved. This distribution is determined by the beam optics,
2890: and thus closely correlated with the luminosity. The proposed 420 m
2891: forward detectors for the CMS and ATLAS experiments
2892: \cite{Albrow:2005ig,Cox:DIS2006}
2893: can tag protons in the range $0.002 \le \xi \le 0.015$; in the TOTEM
2894: experiment at CMS with detectors at 200 m the range will be extended to
2895: $\xi<0.1$ \cite{Whitmore:DIS2006}. Both detectors can obtain a longitudinal
2896: and transverse momentum resolution comparable to the beam distributions.
2897: At a luminosity of $10^{33} \, \text{cm}^{-2} s^{-1}$ with $\beta^* = 0.5 \,
2898: \text{m}$, and a one-sigma normalized emittance $\epsilon = 3.75\times 10^{-6} \, \pi \,
2899: m$, the one-sigma angular spread of the beams at the IP is $8 \, \mu r$,
2900: corresponding to a transverse momentum spread of $56 \, \text{MeV}/c$.
2901: This sets the scale for experimental smearing of the transverse momentum
2902: distributions. The TOTEM
2903: experiment also envisages running at substantially larger $\beta^*$ values
2904: (18, 90 and 1540 m). These values will reduce the transverse angular spread
2905: of the beams at the IP by $\sqrt{\epsilon/(\pi\beta^*)}$, but with a
2906: concomitant reduction in luminosity. Given that the typical scale of
2907: the transverse momentum distributions is $m_g \approx 1\, \text{GeV}$,
2908: it seems feasible to make detailed and precise measurements of the
2909: transverse momentum distributions with the LHC420 and TOTEM detectors
2910: even when running in high--luminosity mode.
2911: %
2912: \begin{acknowledgments}
2913: We thank V.~Khoze, R.~Orava, and M.~Ryskin for useful discussions.
2914: M.~Ryskin kindly made available to us a numerical parametrization
2915: of the $pp$ elastic amplitude of Ref.~\cite{Khoze:2000wk}.
2916:
2917: Notice: Authored by Jefferson Science Associates, LLC under U.S.\ DOE
2918: Contract No.~DE-AC05-06OR23177. The U.S.\ Government retains a
2919: non-exclusive, paid-up, irrevocable, world-wide license to publish or
2920: reproduce this manuscript for U.S.\ Government purposes.
2921: Supported by other DOE contracts and the Binational Science
2922: Foundation (BSF).
2923: \end{acknowledgments}
2924: %
2925: %
2926: \begin{thebibliography}{99}
2927: %
2928: %
2929: \bibitem{Schafer:1990fz}
2930: A.~Schafer, O.~Nachtmann and R.~Schopf,
2931: %``Production Of Higgs Particles In Diffractive Hadron Hadron Collisions,''
2932: Phys.\ Lett.\ B {\bf 249}, 331 (1990).
2933: %%CITATION = PHLTA,B249,331;%%
2934: %
2935: %
2936: \bibitem{Bialas:1991wj}
2937: A.~Bialas and P.~V.~Landshoff,
2938: %``Higgs production in p p collisions by double pomeron exchange,''
2939: Phys.\ Lett.\ B {\bf 256}, 540 (1991).
2940: %%CITATION = PHLTA,B256,5
2941: %
2942: %
2943: \bibitem{Dokshitzer:1991he}
2944: Y.~L.~Dokshitzer, V.~A.~Khoze and T.~Sjostrand,
2945: %``Rapidity gaps in Higgs production,''
2946: Phys.\ Lett.\ B {\bf 274}, 116 (1992).
2947: %%CITATION = PHLTA,B274,116;%%
2948: %
2949: %
2950: \bibitem{Khoze:2000wk}
2951: V.~A.~Khoze, A.~D.~Martin and M.~G.~Ryskin,
2952: %``Soft diffraction and the elastic slope at Tevatron and LHC energies: A
2953: %multi-pomeron approach,''
2954: Eur.\ Phys.\ J.\ C {\bf 18}, 167 (2000)
2955: [arXiv:hep-ph/0007359].
2956: %%CITATION = HEP-PH 0007359;%%
2957: %
2958: %
2959: \bibitem{Kaidalov:2003fw}
2960: A.~B.~Kaidalov, V.~A.~Khoze, A.~D.~Martin and M.~G.~Ryskin,
2961: %``Central exclusive diffractive production as a spin parity analyser: From
2962: %hadrons to Higgs,''
2963: Eur.\ Phys.\ J.\ C {\bf 31}, 387 (2003)
2964: [arXiv:hep-ph/0307064].
2965: %%CITATION = HEP-PH 0307064;%%
2966: %
2967: %
2968: \bibitem{Nussinov} S.~Nussinov, private communication to L.~F.
2969: %
2970: %
2971: \bibitem{Bjorken:1992er}
2972: J.~D.~Bjorken,
2973: %``Rapidity gaps and jets as a new physics signature in very high-energy
2974: %hadron hadron collisions,''
2975: Phys.\ Rev.\ D {\bf 47}, 101 (1993).
2976: %%CITATION = PHRVA,D47,101;%%
2977: %
2978: %
2979: \bibitem{Gribov:1973jg}
2980: V.~N.~Gribov,
2981: %``Space-Time Description Of Hadron Interactions At High Energies,''
2982: arXiv:hep-ph/0006158.
2983: %
2984: %
2985: \bibitem{Frankfurt:2005mc}
2986: L.~Frankfurt, M.~Strikman and C.~Weiss,
2987: %``Small-x physics: From HERA to LHC and beyond,''
2988: Ann.\ Rev.\ Nucl.\ Part.\ Sci.\ {\bf 55}, 403 (2005)
2989: [arXiv:hep-ph/0507286].
2990: %%CITATION = HEP-PH 0507286;%%
2991: %
2992: %
2993: \bibitem{Frankfurt:2004kn}
2994: L.~Frankfurt, M.~Strikman and C.~Weiss,
2995: %``3D parton imaging of the nucleon in high-energy p p and p A collisions,''
2996: Annalen Phys.\ {\bf 13}, 665 (2004)
2997: [arXiv:hep-ph/0410307].
2998: %%CITATION = HEP-PH 0410307;%%
2999: %
3000: %
3001: \bibitem{Frankfurt:2004ti}
3002: L.~Frankfurt, M.~Strikman, C.~Weiss and M.~Zhalov,
3003: %``Transverse structure of strong interactions at LHC: From diffraction to new
3004: %particle production,''
3005: arXiv:hep-ph/0412260.
3006: %%CITATION = HEP-PH 0412260;%%
3007: %
3008: %
3009: \bibitem{Abe:1997bp}
3010: F.~Abe {\it et al.} [CDF Collaboration],
3011: % ``Measurement of double parton scattering in $\bar{p}p$ collisions at
3012: %$\sqrt{s} = 1.8$ TeV,''
3013: Phys.\ Rev.\ Lett.\ {\bf 79}, 584 (1997).
3014: %%CITATION = PRLTA,79,584;%%
3015: %
3016: %\bibitem{Abe:1997xk}
3017: F.~Abe {\it et al.} [CDF Collaboration],
3018: % ``Double parton scattering in $\bar{p}p$ collisions at $\sqrt{s} = 1.8
3019: %$TeV,''
3020: Phys.\ Rev.\ D {\bf 56}, 3811 (1997).
3021: %%CITATION = PHRVA,D56,3811;%%
3022: %
3023: %
3024: \bibitem{Albrow:2005ig}
3025: M.~G.~Albrow {\it et al.},
3026: %``FP420: An R&D proposal to investigate the feasibility of installing
3027: %proton tagging detectors in the 420-m region at LHC,''
3028: CERN-LHCC-2005-025
3029: %\href{http://www.slac.stanford.edu/spires/find/hep/www?r=cern-lhcc-2005-025}
3030: %{SPIRES entry}
3031: %
3032: %
3033: \bibitem{Albrow:2005fw}
3034: M.~G.~Albrow {\it et al.},
3035: %``A search for the Higgs boson using very forward tracking detectors with
3036: %CDF,''
3037: arXiv:hep-ex/0511057.
3038: %%CITATION = HEP-EX 0511057;%%
3039: %
3040: %
3041: \bibitem{Block:1998hu}
3042: M.~M.~Block, E.~M.~Gregores, F.~Halzen and G.~Pancheri,
3043: % ``Photon proton and photon photon scattering from nucleon nucleon forward
3044: %amplitudes,''
3045: Phys.\ Rev.\ D {\bf 60}, 054024 (1999)
3046: [arXiv:hep-ph/9809403].
3047: %%CITATION = HEP-PH 9809403;%%
3048: %
3049: %
3050: \bibitem{Bourrely:2002wr}
3051: C.~Bourrely, J.~Soffer and T.~T.~Wu,
3052: % ``Impact-picture phenomenology for pi+- p, K+- p and p p, anti-p p elastic
3053: %scattering at high energies,''
3054: Eur.\ Phys.\ J.\ C {\bf 28}, 97 (2003)
3055: [arXiv:hep-ph/0210264].
3056: %%CITATION = HEP-PH 0210264;%%
3057: %
3058: %
3059: \bibitem{Islam:2002au}
3060: M.~M.~Islam, R.~J.~Luddy and A.~V.~Prokudin,
3061: Mod.\ Phys.\ Lett.\ A {\bf 18}, 743 (2003).
3062: %%CITATION = HEP-PH 0210437;%%
3063: %
3064: %
3065: \bibitem{Frankfurt:2003td}
3066: L.~Frankfurt, M.~Strikman and C.~Weiss,
3067: %``Dijet production as a centrality trigger for p p collisions at LHC,''
3068: Phys.\ Rev.\ D {\bf 69}, 114010 (2004)
3069: [arXiv:hep-ph/0311231].
3070: %%CITATION = HEP-PH 0311231;%%
3071: %
3072: %
3073: \bibitem{Martin:1962rt}
3074: A.~Martin,
3075: %``Unitarity And High-Energy Behavior Of Scattering Amplitudes,''
3076: Phys.\ Rev.\ {\bf 129}, 1432 (1963).
3077: %%CITATION = PHRVA,129,1432;%%
3078: %
3079: %
3080: \bibitem{Marchesini:1976hw}
3081: G.~Marchesini and E.~Rabinovici,
3082: %``Diffractive Production Amplitudes For Alpha (P) > 1,''
3083: Nucl.\ Phys.\ B {\bf 120}, 253 (1977);
3084: %%CITATION = NUPHA,B120,253;%%
3085: %\bibitem{Marchesini:1978kp}
3086: G.~Marchesini, E.~Rabinovici and L.~Trentadue,
3087: %``High Mass Diffraction In The Supercritical Region,''
3088: Nucl.\ Phys.\ B {\bf 147}, 41 (1979).
3089: %%CITATION = NUPHA,B147,41;%%
3090: %
3091: %
3092: \bibitem{Diehl:2003ny}
3093: M.~Diehl,
3094: %``Generalized parton distributions,''
3095: Phys.\ Rept.\ {\bf 388}, 41 (2003)
3096: [arXiv:hep-ph/0307382].
3097: %%CITATION = HEP-PH 0307382;%%
3098: %
3099: %
3100: \bibitem{Belitsky:2005qn}
3101: A.~V.~Belitsky and A.~V.~Radyushkin,
3102: %``Unraveling hadron structure with generalized parton distributions,''
3103: Phys.\ Rept.\ {\bf 418}, 1 (2005)
3104: [arXiv:hep-ph/0504030].
3105: %%CITATION = HEP-PH 0504030;%%
3106: %
3107: %
3108: \bibitem{Frankfurt:1997ha}
3109: L.~Frankfurt, A.~Freund, V.~Guzey and M.~Strikman,
3110: %``Nondiagonal parton distributions in the leading logarithmic
3111: %approximation,''
3112: Phys.\ Lett.\ B {\bf 418}, 345 (1998)
3113: [Erratum-ibid.\ B {\bf 429}, 414 (1998)]
3114: [arXiv:hep-ph/9703449].
3115: %%CITATION = HEP-PH 9703449;%%
3116: %
3117: %
3118: \bibitem{Shuvaev:1999ce}
3119: A.~G.~Shuvaev, K.~J.~Golec-Biernat, A.~D.~Martin and M.~G.~Ryskin,
3120: %``Off-diagonal distributions fixed by diagonal partons at small x and xi,''
3121: Phys.\ Rev.\ D {\bf 60}, 014015 (1999)
3122: [arXiv:hep-ph/9902410].
3123: %%CITATION = HEP-PH 9902410;%%
3124: %
3125: %
3126: \bibitem{Burkardt:2002hr}
3127: M.~Burkardt,
3128: Int.\ J.\ Mod.\ Phys.\ A {\bf 18}, 173 (2003).
3129: %
3130: %
3131: \bibitem{Pobylitsa:2002iu}
3132: P.~V.~Pobylitsa,
3133: Phys.\ Rev.\ D {\bf 66}, 094002 (2002).
3134: %
3135: %
3136: \bibitem{Diehl:2002he}
3137: M.~Diehl,
3138: Eur.\ Phys.\ J.\ C {\bf 25}, 223 (2002).
3139: %
3140: %
3141: \bibitem{Binkley:1981kv}
3142: M.~Binkley {\it et al.},
3143: Phys.\ Rev.\ Lett.\ {\bf 48}, 73 (1982).
3144: %%CITATION = PRLTA,48,73;%%
3145: %
3146: %
3147: \bibitem{Aktas:2005xu}
3148: A.~Aktas {\it et al.} [H1 Collaboration],
3149: %``Elastic J/psi production at HERA,''
3150: arXiv:hep-ex/0510016.
3151: %%CITATION = HEP-EX 0510016;%%
3152: %
3153: %
3154: \bibitem{Chekanov:2004mw}
3155: S.~Chekanov {\it et al.} [ZEUS Collaboration],
3156: %``Exclusive electroproduction of J/psi mesons at HERA,''
3157: Nucl.\ Phys.\ B {\bf 695}, 3 (2004)
3158: [arXiv:hep-ex/0404008].
3159: %%CITATION = HEP-EX 0404008;%%
3160: %
3161: %
3162: \bibitem{Frankfurt:2002ka}
3163: L.~Frankfurt and M.~Strikman,
3164: Phys.\ Rev.\ D {\bf 66} (2002) 031502.
3165: %
3166: %
3167: \bibitem{Strikman:2004km}
3168: M.~Strikman and C.~Weiss, in Proceeedings of the XII International
3169: Workshop on Deep Inelastic Scattering (DIS 2004), Strbske Pleso,
3170: Slovakia, Apr.~14-18, 2004 [arXiv:hep-ph/0408345].
3171: %
3172: %
3173: \bibitem{Strikman:2003gz}
3174: M.~Strikman and C.~Weiss,
3175: %``Chiral dynamics and the growth of the nucleon's gluonic
3176: %transverse size at small x,''
3177: Phys.\ Rev.\ D {\bf 69}, 054012 (2004)
3178: [arXiv:hep-ph/0308191].
3179: %%CITATION = HEP-PH 0308191;%%
3180: %
3181: %
3182: \bibitem{Gunion:1992hs}
3183: J.~F.~Gunion, H.~E.~Haber, G.~L.~Kane and S.~Dawson,
3184: ``The Higgs hunter's guide,''
3185: Frontiers in Physics, 80, Addison-Wesley (1990);
3186: Errata
3187: %``Errata for the Higgs hunter's guide,''
3188: arXiv:hep-ph/9302272.
3189: %%CITATION = HEP-PH 9302272;%%
3190: %
3191: %
3192: \bibitem{Khoze:1997dr}
3193: V.~A.~Khoze, A.~D.~Martin and M.~G.~Ryskin,
3194: %``The rapidity gap Higgs signal at LHC,''
3195: Phys.\ Lett.\ B {\bf 401}, 330 (1997)
3196: [arXiv:hep-ph/9701419].
3197: %
3198: %
3199: \bibitem{Khoze:2000cy}
3200: V.~A.~Khoze, A.~D.~Martin and M.~G.~Ryskin,
3201: %``Can the Higgs be seen in rapidity gap events at the Tevatron or the
3202: %LHC?,''
3203: Eur.\ Phys.\ J.\ C {\bf 14}, 525 (2000)
3204: [arXiv:hep-ph/0002072].
3205: %%CITATION = HEP-PH 0002072;%%
3206: %
3207: %
3208: \bibitem{Brodsky:1994kf}
3209: S.~J.~Brodsky, L.~Frankfurt, J.~F.~Gunion, A.~H.~Mueller and M.~Strikman,
3210: %``Diffractive leptoproduction of vector mesons in QCD,''
3211: Phys.\ Rev.\ D {\bf 50}, 3134 (1994)
3212: [arXiv:hep-ph/9402283].
3213: %%CITATION = HEP-PH 9402283;%%
3214: %
3215: %
3216: \bibitem{Mueller:1998fv}
3217: D.~Mueller, D.~Robaschik, B.~Geyer, F.~M.~Dittes and J.~Horejsi,
3218: %``Wave functions, evolution equations and evolution kernels from light-ray
3219: %operators of {QCD},''
3220: Fortsch.\ Phys.\ {\bf 42}, 101 (1994)
3221: [arXiv:hep-ph/9812448].
3222: %%CITATION = HEP-PH 9812448;%%
3223: %
3224: %
3225: \bibitem{Abramowicz:1995hb}
3226: H.~Abramowicz, L.~Frankfurt and M.~Strikman,
3227: %``Interplay of hard and soft physics in small x deep inelastic processes,''
3228: Surveys High Energ.\ Phys.\ {\bf 11}, 51 (1997)
3229: [arXiv:hep-ph/9503437].
3230: %%CITATION = HEP-PH 9503437;%%
3231: %
3232: %
3233: \bibitem{Ji:1996nm}
3234: X.~D.~Ji,
3235: %``Deeply-virtual Compton scattering,''
3236: Phys.\ Rev.\ D {\bf 55}, 7114 (1997)
3237: [arXiv:hep-ph/9609381].
3238: %%CITATION = HEP-PH 9609381;%%
3239: %
3240: %
3241: \bibitem{Radyushkin:1997ki}
3242: A.~V.~Radyushkin,
3243: %``Nonforward parton distributions,''
3244: Phys.\ Rev.\ D {\bf 56}, 5524 (1997)
3245: [arXiv:hep-ph/9704207].
3246: %%CITATION = HEP-PH 9704207;%%
3247: %
3248: %
3249: \bibitem{Collins:1998be}
3250: J.~C.~Collins and A.~Freund,
3251: %``Proof of factorization for deeply virtual Compton scattering in {QCD},''
3252: Phys.\ Rev.\ D {\bf 59}, 074009 (1999)
3253: [arXiv:hep-ph/9801262].
3254: %%CITATION = HEP-PH 9801262;%%
3255: %
3256: %
3257: \bibitem{Salam:2005yp}
3258: G.~P.~Salam,
3259: %``Asymptotics and preasymptotics at small x,''
3260: arXiv:hep-ph/0501097 .
3261: %%CITATION = HEP-PH 0501097;%%
3262: %%Cited 6 times in SPIRES-HEP
3263: %
3264: %
3265: \bibitem{Gribov:2003nw}
3266: V.~N.~Gribov,
3267: ``The theory of complex angular momenta: Gribov lectures on theoretical
3268: physics,''
3269: Cambridge University Press, Cambridge, UK (2003).
3270: %
3271: %
3272: \bibitem{Goulianos:2005ac}
3273: K.~Goulianos,
3274: in: Proceedings of the 11th International Conference on Elastic
3275: and Diffractive Scattering: Towards High Energy Frontiers: The 20th
3276: Anniversary of the Blois Workshops, Chateau de Blois, Blois, France,
3277: 15--20 May 2005,
3278: %``Twenty years of diffraction at the Tevatron,''
3279: arXiv:hep-ph/0510035.
3280: %%CITATION = HEP-PH 0510035;%%
3281: %
3282: %
3283: \bibitem{LLIII} L.~D.~Landau and E.~M.~Lifshits, in:
3284: Course of Theoretical Physics, Vol.\ III: Quantum Mechanics,
3285: Pergamon Press, Oxford, 1973.
3286: %
3287: %
3288: \bibitem{Ciafaloni:2003rd}
3289: M.~Ciafaloni, D.~Colferai, G.~P.~Salam and A.~M.~Stasto,
3290: %``Renormalisation group improved small-x Green's function,''
3291: Phys.\ Rev.\ D {\bf 68}, 114003 (2003)
3292: [arXiv:hep-ph/0307188].
3293: %%CITATION = HEP-PH 0307188;%%
3294: %
3295: %
3296: \bibitem{Altarelli:2003hk}
3297: G.~Altarelli, R.~D.~Ball and S.~Forte,
3298: %``An anomalous dimension for small x evolution,''
3299: Nucl.\ Phys.\ B {\bf 674}, 459 (2003)
3300: [arXiv:hep-ph/0306156].
3301: %%CITATION = HEP-PH 0306156;%%
3302: %
3303: %
3304: \bibitem{Bartels:2006ea}
3305: J.~Bartels, S.~Bondarenko, K.~Kutak and L.~Motyka,
3306: % ``Exclusive Higgs boson production at the LHC: Hard rescattering
3307: %corrections,''
3308: Phys.\ Rev.\ D {\bf 73}, 093004 (2006)
3309: [arXiv:hep-ph/0601128].
3310: %%CITATION = HEP-PH 0601128;%%
3311: %
3312: %
3313: \bibitem{Good:1960ba}
3314: M.~L.~Good and W.~D.~Walker,
3315: %``Diffraction dissociation of beam particles,''
3316: Phys.\ Rev.\ {\bf 120}, 1857 (1960).
3317: %%CITATION = PHRVA,120,1857;%%
3318: %
3319: %
3320: \bibitem{Miettinen:1978jb}
3321: H.~I.~Miettinen and J.~Pumplin,
3322: %``Diffraction Scattering And The Parton Structure Of Hadrons,''
3323: Phys.\ Rev.\ D {\bf 18}, 1696 (1978).
3324: %%CITATION = PHRVA,D18,1696;%%
3325: %
3326: %
3327: \bibitem{Diakonov:1987ty}
3328: D.~Diakonov, V.~Y.~Petrov and P.~V.~Pobylitsa,
3329: %``A Chiral Theory Of Nucleons,''
3330: Nucl.\ Phys.\ B {\bf 306}, 809 (1988).
3331: %%CITATION = NUPHA,B306,809;%%
3332: %%Cited 249 times in SPIRES-HEP
3333: %
3334: %
3335: \bibitem{Diakonov:2002fq}
3336: D.~Diakonov,
3337: %``Instantons at work,''
3338: Prog.\ Part.\ Nucl.\ Phys.\ {\bf 51}, 173 (2003)
3339: [arXiv:hep-ph/0212026].
3340: %%CITATION = HEP-PH 0212026;%%
3341: %%Cited 71 time in SPIRES-HEP
3342: %
3343: %
3344: \bibitem{Schafer:1996wv}
3345: T.~Schafer and E.~V.~Shuryak,
3346: %``Instantons in QCD,''
3347: Rev.\ Mod.\ Phys.\ {\bf 70}, 323 (1998)
3348: [arXiv:hep-ph/9610451].
3349: %%CITATION = HEP-PH 9610451;%%
3350: %%Cited 370 times in SPIRES-HEP
3351: %
3352: %
3353: \bibitem{Diakonov:1996sr}
3354: D.~Diakonov, V.~Petrov, P.~Pobylitsa, M.~V.~Polyakov and C.~Weiss,
3355: %``Nucleon parton distributions at low normalization point in the
3356: %large N(c) limit,''
3357: Nucl.\ Phys.\ B {\bf 480}, 341 (1996)
3358: [arXiv:hep-ph/9606314];
3359: %%CITATION = HEP-PH 9606314;%%
3360: %%Cited 62 times in SPIRES-HEP
3361: %
3362: %
3363: %\bibitem{Diakonov:1997vc}
3364: %D.~Diakonov, V.~Y.~Petrov, P.~V.~Pobylitsa, M.~V.~Polyakov and C.~Weiss,
3365: %``Unpolarized and polarized quark distributions in the
3366: %large-N(c) limit,''
3367: Phys.\ Rev.\ D {\bf 56}, 4069 (1997)
3368: [arXiv:hep-ph/9703420].
3369: %%CITATION = HEP-PH 9703420;%%
3370: %%Cited 55 times in SPIRES-HEP
3371: %
3372: %
3373: \bibitem{unknown:1997xu}
3374: TOTEM Collaboration,
3375: %``Totem Cross Section: Elastic Scattering And Diffraction Dissociation
3376: %At The Lhc,''
3377: CERN-LHCC-1997-049
3378: %
3379: %
3380: \bibitem{Bultmann:2005na}
3381: S.~Bultmann {\it et al.},
3382: %``First measurement of A(N) at s**(1/2) = 200-GeV in polarized proton
3383: %proton elastic scattering at RHIC,''
3384: Phys.\ Lett.\ B {\bf 632}, 167 (2006)
3385: [arXiv:nucl-ex/0507030];
3386: %%CITATION = NUCL-EX 0507030;%%
3387: %\cite{Bueltmann:2003gq}
3388: %\bibitem{Bueltmann:2003gq}
3389: S.~L.~Bueltmann {\it et al.},
3390: %``First measurement of proton proton elastic scattering at RHIC,''
3391: Phys.\ Lett.\ B {\bf 579}, 245 (2004)
3392: [arXiv:nucl-ex/0305012];
3393: %%CITATION = NUCL-EX 0305012;%%
3394: %
3395: H.~Spinka {\it et al.}, ``Physics with Tagged Forward Protons with
3396: the STAR Detector at RHIC,'' proposal to STAR, unpublished, June 8, 2005.
3397: %
3398: %
3399: \bibitem{Cox:DIS2006} B.~Cox, Talk presented at
3400: XIV International Workshop on Deep Inelastic Scattering (DIS2006),
3401: Tsukuba City, Japan, 20 -- 24 Apr.~2006.
3402: %
3403: %
3404: \bibitem{Whitmore:DIS2006} J.~Whitmore, Talk presented at
3405: XIV International Workshop on Deep Inelastic Scattering (DIS2006),
3406: Tsukuba City, Japan, 20 -- 24 Apr.~2006.
3407: %
3408: %
3409: \end{thebibliography}
3410: %
3411: %
3412: %\newpage
3413: %
3414: %
3415: %\widetext
3416: %\tableofcontents
3417: %
3418: \end{document}
3419: %
3420: