1: \documentclass[11pt]{article}
2: \usepackage{epsfig}
3: \newcommand{\mbf}[1]{\mbox{\boldmath $#1$}}
4: \usepackage{graphicx}% Include figure files
5: \usepackage{amsmath}
6: \usepackage{dsfont}
7: \newcommand{\bk}{\mbf{k}}
8: \newcommand{\bq}{\mbf{q}}
9: \newcommand{\bp}{\mbf{p}}
10: \newcommand{\bb}{\mbf{b}}
11: \newcommand{\disc}{\textrm{disc}}
12: \newcommand{\cA}{{\cal A}}
13: \newcommand{\cT}{{\cal T}}
14: \newcommand{\cF}{{\cal F}}
15: \newcommand{\cN}{{\cal N}}
16: \newcommand{\cM}{{\cal M}}
17: \newcommand{\cD}{{\cal D}}
18: \newcommand{\cK}{{\cal K}}
19: \newcommand{\cG}{{\cal G}}
20: \newcommand{\tF}{\tilde{\cal F}}
21: \newcommand{\tA}{\tilde{\cal A}}
22: \newcommand{\tT}{\tilde{\cal T}}
23: %\input{xgs}
24: \renewcommand{\baselinestretch}{1.}
25: \textwidth 155mm
26: \textheight 240mm
27: \columnsep 38pt
28: \topmargin -30pt
29: \oddsidemargin 5pt
30: \pagestyle{empty}
31: \parsep 3pt plus 1pt minus 1pt
32: \title{\bf AGK Cutting Rules and Multiple Scattering\\in Hadronic Collisions}
33: \author{J.Bartels$^{a}$, M.Salvadore$^{a,b}$, and G.P.Vacca$^{b}$ \\
34: $^a$ II Inst. f. Theor. Physik, Univ. of Hamburg, Germany and\\
35: $^b$Dipartimento di Fisica and INFN, Univ. of Bologna, Italy }
36: \date{}
37: \pagestyle{myheadings}
38: \def\beq{\begin{equation}}
39: \def\eeq{\end{equation}}
40: \def\bea{\begin{eqnarray}}
41: \def\eea{\end{eqnarray}}
42: \def\noi{\noindent}
43: \def\o{\omega}
44: \begin{document}
45: \maketitle
46: \medskip
47: \noindent
48: {\bf Abstract:}\\
49: We discuss the AGK rules for the exchange of an arbitrary number
50: of reggeized gluons in perturbative QCD in the high energy limit.
51: Results include the cancellation of corrections to single jet and double
52: jet inclusive cross sections, both for hard and soft rescattering
53: contributions.
54:
55:
56: \section{Introduction}
57: More than 30 years ago, Abramovsky, Gribov, and Kancheli in their pioneering
58: paper ~\cite{AGK} have pointed out that, for high energy hadron hadron
59: scattering, the multiple exchange of Pomerons with intercept at or close to
60: unity leads to observable effects in multiparticle final states.
61: As an example, double Pomeron exchange predicts fluctuations of
62: the rapidity densities of produced particles; furthermore, in double
63: inclusive particle production multi-Pomeron exchange leads to long range
64: rapidity correlations. All these results suggests that, at very high energies,
65: hadron hadron scattering might exhibit some sort of critical behavior.
66: As to the theoretical basis of these predictions, only fairly general
67: properties of Regge theory have been used, especially features
68: of particle-reggeon vertices which are quite independent of any special
69: underlying quantum field theory. In particular, no reference has been made to
70: QCD.
71:
72: In recent years, the investigation of the small-$x$ region of hard
73: scattering processes has lead to interest in the BFKL
74: Pomeron~\cite{BFKL}
75: which, at short distances, describes the exchange of vacuum quantum
76: numbers at high energies. In connection with saturation and the color
77: glass condensate, both in DIS at HERA and in heavy ion collisions at
78: RHIC also the
79: multiple exchange of BFKL Pomerons is being addressed; a convenient framework
80: for this is the nonlinear Balitsky-Kovchegov (BK) equation.
81: At hadron colliders, there is more and more evidence
82: ~\cite{SS} that at high energies
83: multiple interactions of partons cannot be neglected.
84: All this naturally suggests to perform the AGK
85: analysis in the framework of perturbative QCD where the Pomeron
86: is described by the sum of BFKL ladder diagrams. In particular,
87: it will be of interest to understand how the presence of multiple
88: scattering affects, for example, the multiplicity of jets in the final state
89: or the rapidity correlation between two jets.
90: A special motivation for investigating multiple scattering comes from the
91: interest in saturation: most direct signals of the existence of saturation
92: are expected to be seen in specific features of final states.
93: Use of the AGK rules in pQCD has already been made in several papers
94: \cite{Braun,Kovchegov}.
95:
96: The theoretical basis for studying these rules in perturbative QCD
97: has been laid down in ~\cite{BR}. A clean theoretical environment for
98: studying high energy scattering processes in pQCD is provided by virtual
99: photons where the mass of the photon defines the momentum scale of the QCD
100: coupling.
101: The coupling of two BFKL Pomerons to a virtual photon has been analysed in
102: ~\cite{BW}; a generalization to three BFKL ladders has been investigated in
103: ~\cite{BE}. It is the structure of these couplings of BFKL Pomerons to
104: the external photon which provides the basis for performing the AGK analysis
105: in pQCD. The most interesting applications of the pQCD AGK-analysis
106: include deep inelastic electron proton scattering or multijet production in
107: $pp$ collisions: in both cases one needs multi-gluon couplings to the
108: proton. In the absence of any first-principle calculation, these couplings
109: have to be modelled, similarly to the initial conditions of the parton
110: densities. The perturbative analysis of the coupling of two (and more) BFKL
111: Pomerons to virtual photons suggests that these couplings are not
112: arbitrary: they obey certain symmetry requirements, which can
113: be expected to be valid also beyond perturbation theory. We therefore believe
114: that our understanding of the pQCD couplings will help to model, for example
115: in proton proton scattering, the coupling of $n$ gluons to the proton.
116:
117: In this paper we therefore begin with a brief review of the main results
118: of ~\cite{BR}, and we point out which features of
119: the couplings are crucial
120: for obtaining the AGK results\footnote{The issue of AGK rules
121: in connection with multiparticle interaction has already been addressed
122: in a different way in \cite{RT} without investigation of the
123: cancellations involved.}.
124: We then propose a generalization to
125: multiple BFKL exchange in hadron-hadron scattering.
126: Starting from these couplings, we then re-derive, for illustration,
127: the counting rules presented in the original AGK paper for an arbitrary
128: number of exchanged gluons,
129: and we apply these rules to single and double inclusive
130: jet production cross sections. The comparison with the original AGK paper
131: allows us to include into our analysis also soft rescattering corrections.
132: As a specific example, we consider the double
133: BFKL correction to the Mueller-Navelet jet cross section formula \cite{MN}.
134:
135: The paper is organized as follows. In section 2 we briefly review
136: the logic behind the AGK analysis, and we summarize the results
137: of ~\cite{BR}. This leads us to define a framework for
138: studying multiple BFKL
139: exchanges in hadron-hadron scattering. In section 3 we discuss the
140: counting rules for inclusive cross sections, and in sections 4 and 5
141: we turn to single jet and double jet inclusive cross sections. A few details
142: are put into an appendix.
143:
144: \section{Reggeon unitarity, energy cuts \`a la AGK,\\
145: and particle-Pomeron couplings in $\gamma^*\gamma^*$ scattering}
146: \subsection{The nonperturbative AGK rules}
147:
148: In this section we briefly review the AGK strategy and its application to
149: pQCD. We will conclude that the central task is the derivation and the
150: study of the coupling of four (or more) reggeized gluons to virtual photons.
151:
152: %figure--------- n-poles amplitude ----------------------
153: \begin{figure}[ht]
154: \begin{center}
155: \includegraphics[width=5cm]{fig/n-poles.eps}
156: \caption{Graphical representation of a multi-Regge poles contribution to
157: the elastic scattering amplitude. The zigzag lines represent pomerons.}
158: \label{n-poles}
159: \end{center}
160: \end{figure}
161: %figure--------------------------------------------------
162: The original AGK paper starts from a multi-Regge pole contribution
163: to the elastic scattering amplitude (Fig.\ref{n-poles}), written as a
164: Sommerfeld-Watson representation:
165: \beq
166: \cT_{AB}(s,t) =
167: \int \frac{d \o}{2 i} \xi(\o) s^{1+\o} \cF(\o,t).
168: \label{scat}
169: \eeq
170: with $\o=J-1$,
171: \beq
172: \label{signature}
173: \xi(\o)=\frac{\tau-e^{-i \pi \o}}{\sin \pi \o} =
174: i\frac{e^{-i \frac{\pi}{2} (\omega + \frac{1-\tau}{2})}}
175: {\cos [\frac{\pi}{2}(\omega + \frac{1-\tau}{2})]}
176: =i -[\tan{\frac{\pi}{2}\left( \omega+\frac{1-\tau}{2}\right)]},
177: \eeq
178: and $\tau = \pm 1$ being the signature.
179: The (real-valued) partial wave $\cF(\o,t)$ has singularities in the
180: complex $\o$-plane, and the multi-Regge exchange corresponds to a
181: particular branch cut. There is a general formula for the discontinuity
182: across this cut~\cite{ABSW} (Fig.\ref{t-disc}a):
183: \beq
184: \disc_{\o}^{(n)}[\cF(\o,t)] = 2 \pi i
185: \int \frac{d\Omega_n}{n!}~\gamma_{\{\beta_j\}}~
186: \cN_n^{A} (\{\bk_j\}; \o)~\cN_n^{B} (\{\bk_j\}; \o)
187: ~\delta(\o - {\Sigma}_j \beta_j)
188: \label{regunitarity}
189: \eeq
190: \beq
191: d\Omega_n = (2 \pi)^2\delta^2(\bq - {\Sigma}_j \bk_j)
192: \prod_{j=1}^n\frac{d^2\bk_j}{(2 \pi)^2}\nonumber
193: \eeq
194: where $\bk_j$ ($j=1,...,n$) denotes the transverse momentum of the $j$-th
195: Regge pole, $\bq$ is the sum over all transverse momenta
196: with $\bq^2 = -t$, and $\alpha(-\bk_j^2)= \alpha_j = 1 + \beta_j$
197: is the Regge pole trajectory function.
198: The factor which determines the overall sign has the form
199: \beq
200: \gamma_{\{\beta_j\}} = \Im \big[ -i \Pi_j (i\xi_j) \big] =
201: (-1)^{n-1}\frac{\cos \big[ \frac{\pi}{2} \textrm{$\sum_j$} \Big(\beta_j +
202: \frac{1-\tau_j}{2} \Big) \big] }{\textrm{$\prod_j$} \cos \big[\frac{\pi}{2}
203: \Big( \beta_j + \frac{1-\tau_j}{2} \Big)\big] }
204: \label{gammafactor}
205: \eeq
206: As an example, the contribution of two even-signature Regge poles (Pomerons)
207: with intercept close to unity is negative compared to the single pole
208: contribution.
209: Equation (\ref{regunitarity}) is a `reggeon unitarity equation':
210: it describes the contribution of the $n$-reggeon $t$-channel state to
211: the discontinuity in angular momentum of the partial wave $\cF$
212: (Fig.\ref{t-disc}).
213: In the same way as in a usual unitarity integral particles in the
214: intermediate state are to be taken on mass shell, in the reggeon unitarity
215: integral the reggeons of the intermediate state are on shell in reggeon
216: energy: as indicated in Fig.\ref{t-disc}c, the Regge pole is a bound state
217: of (at least) two particles, and the complex angular momentum of the two
218: particles is put equal to the trajectory function of the Regge pole.
219: %figure------ t-channel discontinuity -------------------
220: \begin{figure}[ht]
221: \begin{center}
222: \includegraphics[width=15cm]{fig/t-disc.eps}
223: \caption{Representation of the $t$-channel content of the elastic amplitude:
224: a) discontinuity across the $2$-poles cut in the $\o$-plane,
225: b) computation of the coupling $2\rightarrow$4 in the $t$-channel physical
226: region and
227: c) isolation of the Regge poles in the (12) and (34) channels.
228: The crosses in a), b) indicate that the lines are on shell in angular momentum
229: and $4$-momentum, resp.}
230: \label{t-disc}
231: \end{center}
232: \end{figure}
233: %figure--------------------------------------------------
234: The formula \eqref{regunitarity} contains the coupling of $n$ Regge poles to
235: the external particles, denoted by $\cN_n({\bk}_j,\omega)$.
236: In general, they are functions of $\omega$ and contain, for example poles
237: and cuts due to the exchange of Regge poles. This includes, in particular,
238: the possibility that the reggeons $i$ and $j$ form a composite state.
239: Depending on the structure of the $\cN_n$, it may be necessary to
240: replace, in (\ref{regunitarity}), the $\omega$ -dependent
241: $\delta$-function by $\prod_1^n \left(\int d \omega_j \delta(\omega_j -
242: \beta_j) \right) \delta (\omega - \sum_j \omega_j)$: in this case,
243: the vertex function $\cN_n$ may depend not only upon the total
244: angular momentum $\omega$, but also upon the $\omega_j$.
245: An example will be given futher below.
246: For these reasons, the vertex functions $\cN_n$ are more general than the
247: impact factors.
248:
249:
250: For later considerations it will be useful to say a few more
251: words about the origin of this formula.
252: Following the idea of Gribov, Pomeranchuk and Ter-Martirosian ~\cite{GPT}
253: one starts, in the simplest case, from the four particle intermediate state in the $t$-channel
254: unitarity equation in the physical region of the process $A+\bar{A}
255: \to B+\bar{B}$ (Fig.\ref{t-disc}b).
256: Above the 4-particle state, the $2 \to 4$ amplitude
257: for the process $A+\bar{A} \to 1+2+3+4$ appears.
258: This scattering amplitude (and the corresponding one below the
259: 4-particle state) is expanded in terms of
260: partial waves, inserted into the $t$-channel unitarity
261: equation. Defining analytic continuations in angular momentum and helicity
262: variables, using Sommerfeld-Watson transformations, and performing
263: parts of the $t$-channel unitarity phase space integrals, one moves from the
264: physical $t$-channel region to negative $t$-values. Retaining Regge poles
265: in the $(12)$ and $(34)$ channels (illustrated in Fig.\ref{t-disc}c),
266: one finally arrives at the discontinuity
267: of the $t$-channel partial wave across the two-reggeon cut, given by
268: \eqref{regunitarity}. This form of
269: $t$-channel unitarity is called 'reggeon unitarity':
270: in the Regge description of high energy scattering processes
271: the partial waves satisfy reggeon unitarity relations, and the single
272: pole, double, triple... exchanges can be isolated by computing
273: discontinuities across the corresponding cuts in the
274: angular momentum plane. The analytic form of the discontinuity equation
275: given in \eqref{regunitarity} is universal,
276: whereas the reggeon-particle couplings entering the equation, $\cN_n$, have
277: to be computed from the underlying theory.
278:
279: The central goal of the AGK analysis is the decomposition of the n-reggeon
280: exchange contribution in terms of $s$-channel intermediate states. To be
281: precise, one is interested in the total cross section, i.e. in
282: the absorptive part or, equivalently, in the discontinuity w.r.t. energy
283: of the scattering amplitude \eqref{scat}.
284: It is quite obvious that the
285: absorptive part of the amplitude (1) will consist of several different
286: contributions: each piece belongs to an energy cut line across
287: Fig.\ref{n-poles}, and there are several different ways of drawing
288: such energy-cutting lines. Each of them
289: belongs to a particular set of $s$-channel intermediate states.
290: For example, a cutting line between reggeons (Fig.\ref{2pcuts}a)
291: belongs to double diffractive production on both sides of the cut:
292: there is a rapidity gap between what is inside the upper blob and the
293: lower blob. When relating
294: this contribution with the full diagram in
295: Fig.\ref{n-poles} one requires a `cut version' of the reggeon
296: particle couplings
297: $\cN_n$. Similarly, the cut through a reggeon (Fig.\ref{2pcuts}b)
298: corresponds to a so-called
299: multiperipheral intermediate state, and another cut version of $\cN_n$
300: appears. The basis of the AGK analysis is the observation that, under very
301: general assumptions for the underlying dynamical theory,
302: the couplings $\cN_n$ are fully symmetric under the exchange
303: of reggeons, and all their cut versions are identical.
304: This property then allows to find simple relations between
305: the different cut contributions, and to derive a set of counting rules.
306: %figure---------- 2-pomeron cuts -----------------------
307: \begin{figure}[ht]
308: \begin{center}
309: \includegraphics[width=10cm]{fig/2pcuts.eps}
310: \caption{Some examples of the different $s$-channel cuts of the $2$-pomeron
311: contribution: a) diffractive and b) multiperipheral intermidiate states.}
312: \label{2pcuts}
313: \end{center}
314: \end{figure}
315: %figure--------------------------------------------------
316:
317: \subsection{The pQCD case}
318: From this brief review it follows that the central task of
319: performing the AGK analysis in pQCD requires the computation and study of
320: the coupling functions $\cN_n$. The simplest task is the study of the
321: two-Pomeron exchange. Since the BFKL Pomeron
322: is a composite state of two reggeized gluons, we have to start from the
323: exchange of four reggeized gluons (Fig.\ref{gamma-gamma}).
324: The blobs above and below denote the couplings $\cN$, and they
325: will be discussed below. They have a nontrivial content of reggeized gluons,
326: and, in particular, any pair of gluons $1,2,...,4$ can come from a composite
327: BFKL state contained in $\cN$. This diagram has to be compared
328: with Fig.\ref{n-poles}: the elementary Regge poles in pQCD are
329: the reggeized gluons, and the Pomeron appears as a composite state.
330: %figure----- gamma-gamma to 4 reggeized gluons ---------
331: \begin{figure}[ht]
332: \begin{center}
333: \includegraphics[width=5cm]{fig/gamma-gamma.eps}
334: \caption{Exchange of four reggeized gluons in pQCD. The wavy lines represent
335: reggeized gluons. The blobs above and below denote the functions
336: $\cN_4(\{\bk_j\},\omega)$, computed in pQCD.}
337: \label{gamma-gamma}
338: \end{center}
339: \end{figure}
340: %figure--------------------------------------------------
341: Rather than going in all detail through
342: the chain of arguments described before (which would lead us to the
343: study of a $t$-channel $2 \to 8$ process!), we mention a 'shortcut' path
344: which takes us, in an easier way, to the desired coupling of four reggeized
345: gluons to external particles. It is based upon the observation that the
346: same coupling which
347: appears in the reggeon unitarity equation (\ref{regunitarity}) for the
348: discontinuity across the four reggeon cut is also contained in the diffractive cross section formula
349: for low and high-mass diffraction (triple Regge limit): the process
350: (Fig.\ref{triplereg}) $\gamma^*+C \to X +C$ where $X$
351: sums over low and high mass diffractive states (with squared mass
352: $M_X^2$) of the incoming projectile $\gamma^*$.
353: %figure------------ triple Regge limit ------------------
354: \begin{figure}[ht]
355: \begin{center}
356: \includegraphics[width=15cm]{fig/triplereg.eps}
357: \caption{Six point amplitude for the computation of the high-mass diffraction
358: cross section in the triple Regge limit.}
359: \label{triplereg}
360: \end{center}
361: \end{figure}
362: %figure--------------------------------------------------
363: In the limit of large $M_X$ and large energy $s$ with the restriction
364: $M_X^2 \ll s$, this process is dominated by Pomeron exchanges in the
365: lower $t$ channels which in pQCD are BFKL ladders. By cutting
366: these $t$-channels (more precisely: by taking cuts across the angular
367: momentum variables in the lower $t$-channels), we arrive at the coupling
368: of four reggeized gluons to the external particle $\gamma^*$.
369: This is the method which was used in ~\cite{BW} for deriving the
370: coupling of four reggeized gluons to the virtual photon, named $\cD_4$.
371: It is defined to contain the energy denominator $1/(\omega - \sum_{j=1}^{4})$
372: for the four reggeized gluons; its amputated counterpart,
373: $C_4 = D_4 (\omega - \sum_{j=1}^{4})$ will lead to a pQCD model for the
374: coupling function $\cN_4$. The study of $D_4$, therefore,
375: provides the starting point for the AGK analysis in pQCD.
376: In the following we will discuss the function $D_4$, keeping in mind
377: that at the end we have to multiply by $(\omega - \sum_{j=1}^{4})$.
378: Before inserting this coupling into the discontinuity formula,
379: we have to address the complications arising from the color degree of freedom
380: and from the reggeization of the gluon.
381:
382: Let us recapitulate the main properties of $\cD_4$.
383: From the $t$-channel point of view it is natural to demand that the
384: reggeon-particle coupling
385: that enters the reggeon unitarity equation satisfies certain symmetry
386: requirements. In the simplest case, it should be symmetric under
387: the exchange of the reggeons, i.e. their momenta and color labels.
388: On the other hand, we know from the BFKL equation that there are two
389: different $t$-channel states. The Pomeron state which belongs to the color
390: singlet is completely symmetric; on the other hand, when the BFKL
391: amplitude is projected onto the antisymmetric color octet, it satifies the
392: bootstrap equation, and the two gluons
393: 'collapse' into one single gluon. It is therefore not unexpected that also
394: the four reggeon amplitude $\cD_4$ contains antisymmetric configurations which
395: satisfy bootstrap equations. One of the main results of the analysis
396: of ~\cite{BW} is the complete decomposition of $\cD_4$ into two pieces:
397: \beq
398: \cD_4=\cD_4^I + \cD_4^R
399: \label{D4decomp}
400: \eeq
401: Here the first term, $\cD_4^I$, is completely symmetric under the exchange of
402: any two gluons,
403: whereas the second one, $\cD_4^R$, is a sum of antisymmetric terms which,
404: as a result of bootstrap properties, can be expressed in terms of
405: two-gluon amplitudes, $\cD_2$. In a graphical way, $\cD_4^R$ is illustrated in
406: Fig.\ref{D4R}:
407: after making use of the bootstrap properties, $\cD_4^R$ is reduced to
408: a sum of $\cD_2$ functions. Under the exchange of the two reggeized gluons,
409: $\cD_2$ is symmetric. It is only after this decomposition has been
410: performed, and we have arrived at reggeon particle couplings with `good'
411: properties, that we can start with the AGK analysis.
412: %figure------------------- D4R --------------------------
413: \begin{figure}[ht]
414: \begin{center}
415: \includegraphics[width=15cm]{fig/D4R.eps}
416: \caption{Decomposition of the reggeizing part of the four reggeon amplitude
417: $\cD_4^R$ in term of reggeized gluons.}
418: \label{D4R}
419: \end{center}
420: \end{figure}
421: %figure--------------------------------------------------
422:
423: The construction of $\cD_4^I$ implies that the same function appears when,
424: in the AGK analysis, one computes different energy discontinuities of
425: $\gamma^*$-$\gamma^*$ scattering,
426: no matter where the cutting line runs, between gluon `1' and `2',
427: or between `2' and `3', or between `3' and `4'.
428: Since in pQCD the Pomeron is a bound state
429: of two gluons, the two-Pomeron state is formed, e.g., by the pairs $(12)$
430: and $(34)$. Thus cuts through the Pomeron or between
431: two Pomerons all lead to the same two-Pomeron particle vertex.
432: This means that $\cD_4^I$ satifies all the requirements listed in the
433: AGK paper.
434:
435: Starting from these functions $\cD_4^I$ and $\cD_4^R$, the investigation in
436: ~\cite{BR} has shown in some detail how the AGK counting rules work in
437: pQCD: the analysis has to be done
438: seperately for $\cD_4^I$ and $\cD_4^R$. For the first piece, we obtain
439: the counting arguments for the Pomerons (which is even-signatured)
440: given by AGK;
441: here the essential ingredient is the complete symmetry of $\cD_4^I$ under the
442: permutation of reggeized
443: gluons. In the latter piece, the odd-signature reggeizing gluons lead to
444: counting rules which are slightly different from those of the even signature
445: Pomeron: once the bootstrap properties have been invoked and $\cD_4^R$ is
446: expressed in terms of $\cD_2$ functions, cutting lines through the reggeized
447: gluon appear. Since it carries negative signature, the relative weight
448: between cut and uncut reggeon is different from the Pomeron. It should be
449: stressed, however, that the generalization of the AGK analysis
450: to odd signature reggeons is contained in the
451: AGK paper; in pre-QCD times, however, there was no
452: obvious reason for considering Regge poles other than the Pomeron.
453:
454: %figure--------------------------------------------------
455: \begin{figure}[ht]
456: \begin{center}
457: \includegraphics[height=3cm]{fig/Nn.eps}
458: \caption{Graphical illustration of the coupling
459: $\cN_n(\bk_1,a_1;\ldots;\bk_n,a_n;\o)$ of a proton to $n$ reggeized gluons.}
460: \label{Nn}
461: \end{center}
462: \end{figure}
463: %figure--------------------------------------------------
464:
465: Let us now turn to the main goal of this paper, to the use of the pQCD
466: cutting rules in a nonperturbative environment (e.g.multi-ladder exchanges
467: in $pp$ scattering). Basic ingredients are the
468: nonperturbative couplings of $n$ reggeized gluons to the proton.
469: In order to justify the use of pQCD we need a hard scattering subprocess:
470: we will assume that all reggeized gluons are connected to some hard
471: scattering subprocess; consequently, each gluon line will have its
472: transverse momentum in the kinematic region where the use of pQCD can be
473: justified. Since AGK applies to the high energy limit
474: (i.e. the small-$x$ region), all $t$-channel gluons are reggeized.
475: Based upon the analysis in pQCD,
476: we formulate a few general conditions which the nonperturbative
477: couplings have to satisfy.
478: Following the original AGK paper, we will denote these couplings
479: by $\cN_n(\bk_1,a_1;\ldots;\bk_n,a_n;\o)$,
480: where $\bk_i$ and $a_i$ refer to transverse momenta
481: and color of the $i$th gluon, and require that
482: \begin{quote} \em
483: (i) they are symmetric under the simultaneous exchange of momenta
484: and color, e.g. $(k_i,a_i)\leftrightarrow (k_j,a_j)$;\\
485: (ii) cut and uncut vertices are identical, independently where the
486: cut line enters.
487: \end{quote}
488: In the following sections we will work out a few results which follow from
489: these conditions.
490:
491: We conclude this section with a remark on the reggeizing pieces, $D_4^R$.
492: The analysis of
493: $\cD_2$, $\cD_3$, and $\cD_4$ in ~\cite{BW},
494: and of $\cD_5$ and $\cD_6$ in ~\cite{BE}
495: lead to the following conjecture about the general coupling of $n$
496: reggeized gluons to virtual photons. Functions with an even number of lines,
497: $\cD_{2n}$, contain
498: \begin{itemize}
499: \item a piece which is completely symmetric, $\cD_{2n}^I$;
500: \item a second piece which can be expressed in terms of $\cD_{2n-2}$,
501: by either having two of the outgoing reggeized gluons each split into two gluons
502: or one of the outgoing gluons split into three gluons;
503: \item a third piece which is reduced to $\cD_{2n-4}$ with further splittings
504: at the lower end etc.
505: \end{itemize}
506: Finally, functions with an odd number, $\cD_{2n+1}$,
507: \begin{itemize}
508: \item can always be reduced to $\cD_{2n}$, with one splitting at the lower
509: end.
510: \end{itemize}
511: As a consequence, also these reggeizing pieces can be reduced to the
512: couplings $\cN_{2n}$.
513: Since the reggeization of the gluon is a general
514: feature of QCD, valid to all orders in perturbation theory, the appearance
515: of these reggeizing pieces can be expected to happen also in a
516: nonperturbative coupling; in other words, in addition to the functions
517: $\cN_{2n}$ mentioned before, we expect also contributions with the same
518: functions $\cN_{2n}$, but one (or more) of the reggeized gluons split into
519: two gluons etc.
520:
521: In the following we will apply the AGK analysis to $pp$ scattering,
522: assuming that the couplings of $n$ gluons to the proton satisfy the
523: requirements listed above. In this paper we will not analyse yet
524: the reggeizing pieces; their analysis will be taken up in a future
525: paper. Another application of the AGK analysis is DIS, in particular
526: the production of diffractive final states. Such an analysis cannot avoid to
527: include the reggeizing pieces: to lowest order in $\alpha_s$
528: the coupling of $n$ gluons to the virtual photon consists of reggeizing
529: pieces only and hence cannot be
530: ignored.
531:
532: %Recently the case of one and two gluon inclusive production has started to be
533: %analysed~\cite{Kovchegov1_2, Braun2005} in relation to the dipole picture.
534: %There is still a debate and some contraddictory results but the
535: %emerging picture is that there may be a contribution from the transition
536: %vertices (cutted) which originate a jet to the counting related to AGK rules
537: %(a generaled version of).
538: %Nevertheless any kind of extra contributions seems subleading in the high
539: %energy limit. We shall therefore ignore them in our analysis.
540:
541: %%%%%%%%%%%%%%%%%%%%%%%%% Inclusive C.S. %%%%%%%%%%%%%%%%%%%%%%%%%
542: \section{Inclusive cross section}
543: In this section we discuss the AGK rules for inclusive cross sections.
544: Since we are not asking for a final state which provides a hard scale,
545: such a discussion might seem somewhat academic: rescattering in $pp$
546: collisions is described by multiple exchanges of nonperturbative Pomerons, not
547: simply of bound states of reggeized gluons. For consistency, however, it is
548: important to show that BFKL exchanges lead to the same counting rules as the
549: nonpertubative Pomerons. Also, for the modelling of events with multiple
550: scattering events, it will be useful to illustrate the pattern of
551: AGK cancellations.
552:
553: \subsection{The nonperturbative case}
554:
555: Let us begin by recapitulating the counting rules of the original AGK paper.
556: The contribution of the $n$ Pomeron cut to the scattering amplitude
557: is given by
558: \beq
559: \cT_{AB}^{\textrm{n-cut}}(s,t) =
560: \int \frac{d \o}{2 i} \xi(\o) s^{1+\o}~
561: \disc_{\o}^{(n)}[\cF(\o,t)],
562: \label{ampdecomposition}
563: \eeq
564: and the discontinuity of the partial wave has the form given in
565: (\ref{regunitarity}). Doing the $\omega$ integral we arrive at:
566: \bea
567: \cT_{AB}^{\textrm{n-cut}}(s,t) &=&
568: \pi \int \frac{d\Omega_n}{n!} \gamma_{\{\beta_j\}}~
569: \xi(\beta)s^{1+\beta}~
570: \cN_n^{A}~
571: \cN_n^{B}
572: \label{ampexp}
573: \eea
574: where we have defined $\beta \equiv \beta(\{\bk_j\}) = \sum_{j=1}^n \beta_j$
575: and $\cN_n^{A,B} \equiv \cN_n^{A,B}(\{\bk_j\}; \beta )$.
576: The energy discontinuity (imaginary part) is obtained by simply
577: replacing the signature factor (\ref{signature}) by $1$.
578: As we have explained before, the aim of AGK is to obtain the same result for
579: the discontinuity from the unitarity equation, i.e. from products of
580: production amplitudes. Each such contribution can graphically be illustrated
581: by a cutting line (Fig.3). One of the main results of AGK states that
582: the $s$-discontinuity of the $n$-cut contribution to the amplitude,
583: $\disc_s[\cT_{AB}^{\textrm{n-cut}}(s,t)]$,
584: can be written as a sum over the number
585: $k$ of cut pomerons ($k=1,...,n$),
586: \beq
587: \cA^n(s,t) \stackrel{\textrm{def}}{=}
588: \disc_s[\cT_{AB}^{\textrm{n-cut}}(s,t)] = \sum_{k=0}^n \cA_k^n(s,t),
589: \label{muldecomp}
590: \eeq
591: and the terms in the sum are
592: \beq
593: \cA_k^n(s,t) = 2 \pi i \int \frac{d\Omega_n}{n!}~ \cF_k^n~
594: s^{1+\beta}~ \cN_n^{A}~ \cN_n^{B},
595: \eeq
596: where we have introduced the AGK factors
597: \beq
598: \cF_k^n =
599: \Bigg\{ \begin{array}{ll}
600: (-1)^n~2^{n-1}+\gamma_{\{\beta_j\}} & \textrm{if $k=0$} \\
601: (-1)^{n-k}~2^{n-1}\binom{n}{k} & \textrm{if $k>0$}
602: \end{array}.
603: \label{agk}
604: \eeq
605: If a (nonperturbative) pomeron is viewed as a multiperipheral
606: chain of secondary particles, the cut of each pomeron
607: gives a uniform distribution in rapidity, and the sum in (\ref{muldecomp})
608: leads to density fluctuations.
609:
610: The simplest case, the two-Pomeron
611: exchange, has the three contributions illustrated in Fig.8:
612: %figure-------irreducible diagrams------------------------
613: \begin{figure}[ht]
614: \begin{center}
615: \includegraphics[width=15cm]{fig/2pomeronAGK.eps}
616: \caption{There are three different ways to cut the two-pomerons diagram:
617: a) diffractive cut, b) single multiplicity cut and c) double multiplicity cut.}
618: \label{fig:agk2p}
619: \end{center}
620: \end{figure}
621: %figure--------------------------------------------------
622: Figure \ref{fig:agk2p}.a shows the diffractive cut: all the pomerons are left
623: uncut, and there is a rapidity gap between the fragmentation regions of the
624: two particles. Figure \ref{fig:agk2p}.b represents a single multiplicity
625: cut in which only one pomeron has been cut. Eventually, the situation
626: shown in figure \ref{fig:agk2p}.c corresponds to two cut pomerons, and the
627: multiplicity of particles is doubled with respect to the previous case.
628: Neglecting the real part of the pomeron signature factor \eqref{signature}
629: it reduces to the imaginary unit $i$, and the $\gamma$ factor
630: \eqref{gammafactor} is just $(-1)^{n-1}$.
631: From \eqref{agk} we obtain for the weight factors the well-known
632: results:
633: \bea
634: 1: & \textrm{diffractive} & (k=0) \nonumber \\
635: -4: & \textrm{single multiplicity} & (k=1) \\
636: 2: & \textrm{double multiplicity} & (k=2).\nonumber
637: \label{eq:agk4}
638: \eea
639: In other words, the different contributions are in the proportion
640: \beq
641: \cA^2_0:\cA^2_1:\cA^2_2 = 1:-4:2.
642: \label{prop}
643: \eeq
644: This result can be summarized as follows: final states with a rapidity
645: gap ($k=0$) are accompanied by other final states with double density ($k=2$),
646: and their respective cross sections come with the
647: relative weight given by \eqref{prop}.
648: At the same time, these final states are connected with corrections
649: to the cross section of final states with normal density ($k=1$),
650: as contained in \eqref{prop}.
651:
652: For later use we note an important generalization: suppose we
653: substitute one of the soft pomerons by a different Regge pole, and we
654: concentrate on those contributions where this Regge pole is cut
655: (later on, in the context of inclusive jet production, we shall apply this
656: argument to a hard cut gluon ladder). Following the previous argument
657: used to derive the usual AGK rules, we simply sum over
658: cut and uncut soft pomerons, where the uncut soft Pomerons appear on both the
659: left and the right side of the cut reggeon. As a result,
660: the $\gamma$ factor in the first line of \eqref{agk} does not appear, and
661: for the contribution of $k$ cut Pomerons we simply obtain:
662: \beq
663: \cF^n_k \propto (-1)^{n-1-k} \binom{n-1}{k},~~~~~~~k=0,...,n-1
664: \eeq
665: Clearly the sum over $k$ vanishes identically: soft multi-Pomeron
666: corrections to a single cut Regge pole cancel.
667: This argument is easily generalized to two or more singled out cut
668: Regge poles.
669:
670: So far we have discussed the AGK counting only at one rapidity value.
671: The AGK paper also addresses the question of how to continue
672: the $s$-discontinuity cutting lines inside the upper or lower vertex
673: function $\cN_n$. When following, for example, the cut Pomerons
674: inside $\cN_n$ (or cutting lines running between Pomerons)
675: one faces the question of how these cutting lines pass through
676: Pomeron interaction vertices ('cut vertices').
677: AGK constraints can be formulated only for a very restricted class
678: of interaction vertices (in particular, for the $1 \to n$ Pomeron vertex). For
679: the general case (for example, for the $2 \to 2$ vertex) this is not the case;
680: only explicit models, e.g. calculations in pQCD, can provide further
681: information.
682:
683:
684: \subsection{The pQCD case}
685: Let us now derive the counting rule \eqref{agk} in pQCD.
686: Since a Pomeron appears as a bound state of two reggeized gluons,
687: the discussion of $n$ Pomerons has to start from $2n$ reggeized gluons.
688: Moreover, in LLA the signature factor of the reggeized gluon is real,
689: $\xi_\mathds{G}(\bk) = -2/\pi \bk^2$; therefore, working in LLA, the
690: diagrams with any cut reggeized gluon are suppressed and will not be
691: considered.
692: The derivation of \eqref{agk} follows from straightforward combinatorics.
693: The simplest case, $n=2$, has been discussed in ~\cite{BR}, and we
694: can use these results (Fig.9) for illustrating the general proof.
695:
696: %figure-------irreducible diagrams------------------------
697: \begin{figure}[ht]
698: \begin{center}
699: \includegraphics[width=8cm]{fig/agk4.eps}
700: \caption{The three $s$-channel cuts contributing to the four
701: reggeized gluon exchange.}
702: \label{fig:agk4}
703: \end{center}
704: \end{figure}
705: %figure--------------------------------------------------
706:
707: The situation is depicted in figure \ref{fig:agk4}: each term denotes
708: the product of production amplitudes, and it is understood that
709: we sum and integrate over the produced gluons. We again take the
710: discontinuity in $\omega$ across the four $t$-channel gluons
711: (denoted by crosses),
712: and all the produced gluons are absorbed inside the blobs above and below.
713: In the first diagram, the production amplitude on the left of the cutting line
714: (vertical dashed line) contains the exchange of one gluon,
715: the amplitude on the rhs three gluons.
716: Three gluon exchange comes with a negative sign (from the $\gamma$-factor
717: in (\ref{regunitarity}) and from the signature factor
718: in (\ref{scat})); moreover, there is a symmetry factor $1/3!$.
719: Similarly, the second term in Fig.\ref{fig:agk4} denotes the square of two
720: production amplitudes: two-gluon exchange is purely imaginary and has the
721: symmetry factor $1/2!$. Note that, in contrast to the 2 pomeron exchange
722: discussed above, in the case of reggeized gluons we do not need to consider
723: cutting lines inside the reggeized gluons: compared to an uncut gluon,
724: a cut gluon line is suppressed in order $\alpha_s$.
725:
726: Since we are looking for the contributions of BFKL Pomerons which are
727: bound states of two
728: gluons, we must consider all possible pairings among the reggeized gluons;
729: using the notation $(i_1j_1)(i_2j_2)$ to indicate that gluon $i_1$
730: forms a bound state with the gluon $j_1$, and gluon $i_2$ with $j_2$,
731: the three possibilities are: $(12)(34)$, $(13)(24)$ and $(14)(23)$.\\
732: Let us first consider the `diagonal' configuration, in which the
733: pairings in the upper and in the lower blob are identical.
734: One easily sees that these different configurations
735: contribute with different multiplicities.
736: Starting with the first graph in Fig.\ref{fig:agk4}, in all three
737: possible pairings one of the gluon pairs is cut, and the other one is not.
738: This means that all the three configurations contribute
739: with single multiplicity ($k=1$); the weight factor is $- 3 \times 1/3!$.
740: The same argument holds for the third graph in Fig.\ref{fig:agk4}.
741: In the second graph, the configuration $(12)(34)$ does not have any cut pair;
742: this contributes to the diffractive term ($k=0$), and the weight is
743: $(1/2!) \times (1/2!) = 1/4$. The other two pairings have
744: both gluon pairs cut, and therefore they contribute to the
745: double multiplicity term ($k=2$); the weight factor is
746: $2 \times (1/2!) \times (1/2!) = 1/2$. This agrees with the AGK result
747: (\ref{prop}).
748:
749: It should be clear how to generalize this counting to the exchange of $2n$
750: reggeized gluons. The general result for the
751: contribution of $k$ cut Pomerons assumes the form
752: (for the explicit computation see Appendix A):
753: \beq
754: \tA_k^n(s,t) = 2 \pi i~ \tilde{\cF}_k^n
755: \int \frac{d\Omega_{2n}}{n!}~ s^{1+\tilde{\beta_n}}~
756: \cN_{2n}~ \cN_{2n}~
757: \tilde{\gamma}_{\{\bk_j\}},
758: \label{agkpQCD}
759: \eeq
760: where
761: \beq
762: \tilde{\cF}_k^n =
763: \Bigg\{ \begin{array}{ll}
764: (-1)^n~2^{n-1}+(-1)^{n-1} & \textrm{if $k=0$} \\
765: (-1)^{n-k}~2^{n-1}\binom{n}{k} & \textrm{if $k>0$}
766: \end{array}.
767: \label{tildeF}
768: \eeq
769: and
770: \beq
771: \tilde{\gamma}_{\{\bk_j\}} \stackrel{\textrm{def}}{=}
772: \frac{2^{1-n}}{\pi} \prod_{j=1}^{2n} \xi_\mathds{G}(\bk_j) =
773: \frac{2^{1-n}}{\pi} \prod_{j=1}^{2n} \frac{2}{\pi \bk_j^2}.
774: \label{pQCDgamma}
775: \eeq
776: Here we use the tilde symbol to indicate that the result is obtained in pQCD.
777: Note that the weight factors $\tilde{\cF}^n_k$ coincide with those defined
778: in \eqref{agk} for $k\ne0$, while for $k=0$ they coincide when the
779: real part of the soft pomeron signature factor is neglected
780: ($\gamma_{\{\beta_j\}}=(-1)^{n-1}$).
781:
782: It should be stressed that this discussion has made use only of the
783: very general symmetry properties of the $4$-gluon vertex functions above
784: and below the four gluon state. Each vertex function can very well have a
785: complicated internal structure of reggeized gluons; for example, it
786: may consist of a single BFKL ladder which splits into four gluons.
787: In this case, our analysis has demonstrated the AGK rules inside a closed
788: Pomeron loop. The crucial ingredient is the symmetry structure, and in this
789: example it is garanteed by the form of the $2\to4$ gluon vertex.
790: In an anologous way one can generalize the
791: validity of the AGK analysis to general reggeon diagrams in pQCD.
792:
793: What remains is the discussion \cite{BR} of the case where,
794: in Fig.\ref{fig:agk4}, the pairings above and below the four gluon state
795: do not match (non-diagonal configurations): an example is given in
796: Fig.\ref{fig:4reg-nondiag}a. The corresponding final states are depictured in
797: Fig.\ref{fig:4reg-nondiag}b. Here the situation is the following:
798: in the upper rapidity interval we have multiplicity $k$ which can take
799: one of the values $0$, $1$, or $2$ ($k=0$ in our
800: example). Each multiplicity $k$ comes with a relative weight, described by
801: the three-component vector $(1,-4,2)$. In the lower rapidity interval
802: we have multiplicity $k'$ (in our example $k'=2$).
803: Contributions to this $k'$ can come from different
804: $k$ in the upper interval, so the transition from the upper to the lower
805: interval - named a 'switch' - can be described by a $3 \times 3$
806: matrix, whose elements are defined to the fraction of configurations
807: that lead from $k$ to $k'$:
808: %figure-------irreducible diagrams-------- ----------------
809: \begin{figure}[ht]
810: \begin{center}
811: \includegraphics[width=6cm]{fig/4-reg-gluon-AGK.eps}
812: \caption{An example of non-diagonal configuration: a double wavy line
813: represent a two reggeized gluons bound state (they can be freely moved
814: due to the symmetry of the impact factors). The bound state configuration
815: is not the same above and below the cut. In a) the cut forward amplitude
816: is shown, while in b) we depict an example of a final state coming from the cut.}
817: \label{fig:4reg-nondiag}
818: \end{center}
819: \end{figure}
820: %figure--------------------------------------------------
821: \bea
822: \cM = \left( \begin{array}{ccc}
823: 0 & 0 & \frac{1}{2}\\
824: 0 & 1 & 0 \\
825: 1 & 0 & \frac{1}{2}
826: \end{array}
827: \right)
828: \eea
829: (here $k$ labels the columns, $k'$ the rows). Now one easily verifies that the
830: vector $\mbf{\cF}^t = (1, -4, 2)$ is eigenvector of this matrix $\cM$:
831: this implies that the ratios: diffraction : single density : double density
832: holds for both
833: the upper and the lower rapidity interval, i.e. it is invariant.
834:
835: This pattern holds for an arbitrary number, $2n$, of exchanged gluons.
836: The number $k$ of cut BFKL Pomerons
837: denotes the density of gluons in the final state, and for diagonal
838: configurations the AGK factors given in \eqref{tildeF} measure
839: the relative weight of states with density $k$. If we move to the neighbouring
840: rapidity interval, the density changes to $k'$: the transitions from
841: $k$ to $k'$ define a matrix $\cal M$, and the eigenvectors of this
842: matrix are formed by the AGK factors in \eqref{tildeF}.
843: They define, for each rapidity interval, the relative weight of density $k$.
844: Some details are given in the Appendix.
845: %%%%%%%%%%%%%%%%%%%%%%%%% Eikonal part %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
846: \subsection{An example}
847: More detailed results can be derived if one assumes a specific model
848: for the couplings $\cN_{2n}(k_1,a_1;\ldots;k_{2n},a_{2n};\omega)$ to
849: hadrons $A$ and $B$. A popular
850: choice is the eikonal model. In order to satisfy our symmetry requirements,
851: we have to start from the ansatz:
852: \bea
853: \cN_{2n}^A(\bk_1,a_1;\ldots;\bk_{2n},a_{2n};\omega)=
854: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
855: \nonumber\\
856: \frac{1}{\sqrt{(N_c^2-1)^n}}
857: \Biggl( \phi^A(\bk_1,\bk_2;\omega_{12})
858: \delta_{a_1a_2}
859: \cdot
860: ... \cdot\phi^A(\bk_{2n-1},\bk_{2n};\omega_{{2n-1},{2n}})
861: \delta_{a_{2n-1}a_{2n}}
862: +\sum_{Permutations}
863: \Biggr)\;.
864: \eea
865: Here $\omega_{12}= \omega_1 + \omega_2$, and each factor
866: $\phi^A(k_1,k_2;\omega_{12})\delta_{a_1a_2}$ represents a BFKL amplitude
867: in the color singlet state, convoluted with an impact factor
868: $\phi_0^A(\bk_1,\bk_2)$.
869: The sum has to extend over all possible
870: pairings of the gluons $1,...,2n$: $(i_1,j_1),...,(i_n,j_n)$
871: (alltogether, there are $C^n = (2n-1)!!$ such possibilities). When inserting
872: this ansatz for $\cN^A_{2n}$ and $\cN^B_{2n}$ into (\ref{discspQCD}),
873: (\ref{discspQCD1}) and performing the counting described in the first
874: subsection of the Appendix we proceed in the following way:\\
875: (a) we rewrite (\ref{discspQCD}) in the same way as (\ref{regunitarity}),
876: i.e. we define the Sommerfeld-Watson transformation $\tF^n(\o,\bq)$
877: of the scattering amplitude $\tT^n(s,\bq)$
878: and take the discontinuity across the $2n$-reggeon cut:
879: \beq
880: \disc_{\o} \tF^n(\o,\bq) =
881: 4 i \sum_{j=1}^{2n-1} S^n_j~
882: \int d\Omega_{2n}~ \cN^A_{2n}~ \delta(\o-\Sigma_j \beta_j)~ \cN_{2n}^B~
883: \prod_j \xi_\mathds{G}(\bk_j)\;.
884: \eeq
885: (b) when combining the $\cN_{2n}$ from
886: above and from below, we retain only the diagonal combinations,
887: i.e. those combinations where the pairings above and below match.
888: In this way we sum over all bound states.
889: Formally, this coincides with the large-$N_c$ limit.
890: Each term obtained in this way is a product of $n$ BFKL exchanges.
891: \\
892: (c) in each of these products of $n$ BFKL exchanges we have to apply the
893: discussion given after eq.(\ref{gammafactor}):
894: for each pair of gluons $(ij)$ we define the variable
895: $\omega_{ij}$ and we substitute
896: $\delta(\omega - \sum \beta_j) \to [\prod_{(ij)} \int d\omega_{ij}
897: \delta(\omega_{ij} - \beta_i - \beta_j)] \delta(\omega -
898: \sum \omega_{ij})$. In order to take care of momentum conservation we
899: write the measure as
900: \beq
901: d\Omega_{2n} = \int d^2\bb e^{i \bb \cdot \bq}
902: \prod_{j=1}^{2n}\frac{d^2\bk_j}{(2 \pi)^2}e^{-i \bb \cdot \bk_j }.
903: \eeq
904: The discontinuity then assumes the form:
905: \bea
906: \disc_\o \tF^n(\o,t) =
907: 4 i \sum_{j=1}^{2n-1} S^n_j C^n~
908: \int d^2\bb e^{i \bb \cdot \bq}
909: \prod_{(ij)} \Bigg[
910: \int \frac{d^2\bk_{i}}{(2 \pi)^2}
911: \int \frac{d^2\bk_{j}}{(2 \pi)^2}
912: e^{-i \bb \cdot (\bk_{i}+\bk_{j})} \nonumber \\
913: \label{discFna}
914: \cdot \int d\o_{ij}
915: \xi_\mathds{G}(\bk_{i}) \xi_\mathds{G}(\bk_{j})
916: \phi^A(\bk_{i},\bk_{j},\o_{ij})
917: \delta(\o_{ij} -\beta_{i}-\beta_{j})
918: \phi^B(\bk_{i},\bk_{j},\o_{ij})
919: \Bigg]
920: \delta(\o - \Sigma_{(ij)} \o_{ij} )\;.
921: \eea
922: (d)
923: Defining the variable
924: \bea
925: \bq_{ij} = \bk_{i} + \bk_{j},
926: \eea
927: using the result (\ref{cutcounting}) of the Appendix
928: and observing that, since $\xi_\mathds{G}(\bk)$ is real,
929: $\xi_\mathds{G}(\bk_{i})\xi_\mathds{G}(\bk_{j})=
930: \Im[-i(i\xi_\mathds{G}(\bk_{i}))(i\xi_\mathds{G}(\bk_{j}))]=
931: \gamma_2(\bk_{i},\bk_{j})$,
932: we can write the multiplicity k (for $k > 0$) contribution to
933: \eqref{discFna} in the following form:
934: $$
935: \disc_\o \tF^n_k(\o,t) =
936: 4 i \frac{(-1)^{n-k}}{k!(n-k)!}
937: \int d^2\bb e^{i \bb \cdot \bq}
938: \prod_{(ij)} \Bigg[
939: \int \frac{d^2\bq_{ij}}{(2 \pi)^2} ~ e^{-i \bb \cdot \bq_{ij}}
940: \int d\o_{ij}\; \cdot
941: $$
942: \beq
943: \label{discFnb}
944: \int \frac{d^2 \bk_i}{(2 \pi)^2}
945: \gamma_2(\bk_i,\bk_j)
946: \phi^A(\bk_{i},\bq_{ij}-\bk_{i},\o_{ij})
947: \delta(\o_{ij}-\beta_i - \beta_j)
948: \phi^B(\bk_i,\bq_{ij}-\bk_i,\o_{ij})
949: \Bigg]
950: \delta(\o - \Sigma_{(ij)} \o_{ij})\; .\\
951: \eeq
952: (e)
953: Comparing the integrand of the $\omega_{ij}$ integral
954: with formula \eqref{regunitarity}, we identify it as the discontinuity in
955: $\omega_{ij}$ of the partial wave of the BFKL pomeron across the two-reggeon
956: cut:
957: \bea
958: \label{discP}
959: \int \frac{d^2\bk_{i}}{(2 \pi)^2}
960: \gamma_2(\bk_{i},\bk_{j})
961: \phi^A(\bk_{i},\bq_{ij}-\bk_{i},\o_{ij})
962: \delta(\o_{ij}-\beta_i - \beta_j)
963: \phi^B(\bk_{i},\bq_{ij}-\bk_i;\o_{ij})= \nonumber \\
964: \frac{1}{\pi i} \disc_{\o_{ij}} [\cF_{BFKL}(\o_{ij},\bq_{ij})]
965: \eea
966: where
967: \beq
968: \cF_{BFKL}(\o_{ij},\bq_{ij}) =
969: \int \frac{d^2 \bk d^2 \bk'}{(2 \pi)^4} \phi_0^A(\bk, \bq_{ij}-\bk)
970: \cG_2(\bk,\bq_{ij}-\bk,\bk',\bq_{ij}-\bk';\omega_{ij})
971: \phi_0^B(\bk', \bq_{ij}-\bk').
972: \eeq
973: Therefore \eqref{discFnb} becomes
974: \bea
975: \label{discFc}
976: \disc_\o \tF^n_k(\o,t) =
977: 4 i \frac{(-1)^{n-k}}{k!(n-k)!}
978: \int d^2\bb e^{i \bb \cdot \bq} \nonumber \\
979: \cdot \prod_{(ij)}\Bigg[
980: \frac{1}{\pi i}
981: \int \frac{d^2\bq_{ij}}{(2 \pi)^2} ~ e^{-i \bb \cdot \bq_{ij}}
982: \int d\o_{ij}
983: \disc_{\o_{ij}} [\cF_{BFKL}(\o_{ij},\bq_{ij})]\Bigg]
984: \delta(\o - \Sigma_{(ij)} \o_{ij}).
985: \eea
986: (f) Returning to the energy representation:
987: \bea
988: \tA_k^n(s,t) =
989: 4 i s \frac{(-1)^{n-k}}{k!(n-k)!}\!
990: \int \!d^2\bb e^{i \bb \cdot \bq}
991: \Bigg[
992: \frac{1}{\pi i}
993: \int \frac{d^2\bq'}{(2 \pi)^2} ~ e^{-i \bb \cdot \bq'}
994: \!\int \!d\o'
995: \disc_{\o'} [\cF_{BFKL}(\o',\bq')] s^{\o'}
996: \Bigg]^n ,
997: \eea
998: and defining:
999: \beq
1000: \label{Omega}
1001: \Omega(s,\bb)=
1002: \frac{1}{\pi i}
1003: \int \frac{d^2\bq}{(2 \pi)^2} ~ e^{-i \bb \cdot \bq}
1004: \int d\o~
1005: \disc_{\o} [\tF^n_{BFKL}(\o,\bq)] s^{\o},
1006: \eeq
1007: we obtain, after summation over $n \ge k$,
1008: \beq
1009: \label{Ak}
1010: \tA_k(s,t) =
1011: 4 i s \int d^2\bb e^{i \bb \cdot \bq} P(s, \bb),
1012: \eeq
1013: where
1014: \beq
1015: P(s, \bb) = \frac{[\Omega(s,\bb)]^k}{k!}
1016: e^{-\Omega(s,\bb)}
1017: \eeq
1018: is the probability of having $k$ cut Pomerons at fixed impact parameter $b$.\\
1019: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1020:
1021: %%%%%%%%%%%%%%%%%%%%% n-jets inclusive C.S. %%%%%%%%%%%%%%%%%%%%%%
1022: \section{Single and double inclusive cross sections}
1023: In this section we turn to the realistic case, where one ore more
1024: hard final states (e.g. jets or heavy flavors) are produced. Again we begin
1025: with a brief repeat of the AGK results and then turn to pQCD.
1026:
1027: In the AGK paper it has been shown that, for the single (or double) particle
1028: inclusive cross section, large classes of multi-pomeron corrections cancel.
1029: %figure------------------- 1 jet ------------------------
1030: \begin{figure}[ht]
1031: \begin{center}
1032: \includegraphics[height=4cm]{fig/1jetAGK.eps}
1033: \caption{The pattern of cancellation is shown for the single jet inclusive
1034: case: the first graph survives, the three other cancel each other.}
1035: \label{fig:1jetAGK}
1036: \end{center}
1037: \end{figure}
1038: %figure--------------------------------------------------
1039: We illustrate this result in Fig.\ref{fig:1jetAGK}:
1040: for the single inclusive case (Fig.\ref{fig:1jetAGK}a)
1041: all multi-Pomeron exchanges across the produced particle cancel
1042: (Fig.\ref{fig:1jetAGK}b),
1043: and the same is true for the double inclusive case
1044: (Fig.\ref{fig:2jetAGK}a, \ref{fig:2jetAGK}b).
1045: For the latter case, however, there is a new contribution:
1046: the produced particles originate from different Pomerons
1047: (Fig.\ref{fig:2jetAGK}c).
1048: %figure------------------- 1 jet ------------------------
1049: \begin{figure}[ht]
1050: \begin{center}
1051: \includegraphics[height=4cm]{fig/2jetAGK.eps}
1052: \caption{The pattern of cancellation is shown for the double jet inclusive
1053: case: the graph a and c contribute, the other interfere destructively
1054: giving a vanishing contribution.}
1055: \label{fig:2jetAGK}
1056: \end{center}
1057: \end{figure}
1058: %figure--------------------------------------------------
1059: This term is of particular
1060: interest, since it introduces longe range correlations in the
1061: rapidity difference $y_1 - y_2$:
1062: \beq
1063: \rho(y_1, y_2) \sim \frac{1}{\sigma_{tot}}\frac{d^2 \sigma}{dy_1
1064: dy_2} -
1065: \frac{1}{\sigma_{tot}^2}
1066: \frac{d\sigma}{dy_1} \frac{d\sigma}{dy_2}
1067: \eeq
1068: where, for simplicity, we have suppressed all variables other than the
1069: rapidities.
1070: The multi-Pomeron corrections to this term, again, cancel
1071: (Fig.\ref{fig:2jetAGK}d).
1072: In all cases, however, there remain multipomeron corrections between
1073: the production vertices and the projectiles. Examples are
1074: illustrated in Fig.\ref{surviving-terms}.
1075: %figure------------------- 1 jet ------------------------
1076: \begin{figure}[ht]
1077: \begin{center}
1078: \includegraphics[height=4cm]{fig/surviving-terms.eps}
1079: \caption{Some examples of the terms that survive to the AGK cancellation for
1080: the single jet production a) and for the double jet production b).}
1081: \label{surviving-terms}
1082: \end{center}
1083: \end{figure}
1084: %figure--------------------------------------------------
1085:
1086: \subsection{Inclusive single jet production in pQCD}
1087: We now turn to the corresponding final states in pQCD.
1088: As usual, the presence of the hard scale in the final state
1089: justifies, as far as the cut ladder with the produced jet or
1090: heavy flavor state is concerned, the use of perturbation theory.
1091: %figure------------------- 1 jet ------------------------
1092: \begin{figure}[ht]
1093: \begin{center}
1094: \includegraphics[height=4cm]{fig/1jetpQCD.eps}
1095: \caption{Pictorial representation of the basic process for the single
1096: inclusive jet production.}
1097: \label{fig:1jetpQCD}
1098: \end{center}
1099: \end{figure}
1100: %figure--------------------------------------------------
1101: The basic process is illustrated in Fig.\ref{fig:1jetpQCD}: in the context of
1102: LL BFKL ladders (or $k_t$ factorization),
1103: the cross section has the form:
1104: \bea
1105: \frac{d\sigma}{dy d^2\bp}(y,\bp) =
1106: \frac{1}{4}
1107: \int \frac{d^2\bk_1}{(2 \pi)^2}
1108: \int \frac{d^2\bk_2}{(2 \pi)^2}
1109: (2 \pi)^2\delta^{(2)}(\bp-\bk_1-\bk_2) \nonumber \\
1110: \frac{\Phi_2(x_1; \bk_1, -\bk_1)}{\bk_1^2 \bk_1^2}
1111: \cK_{\textrm{real}}(\bk_1,-\bk_1;-\bk_2,\bk_2)
1112: \frac{\Phi_2(x_2; \bk_2, -\bk_2)}{\bk_2^2 \bk_2^2}
1113: \label{1jetfirst}
1114: \eea
1115: with $x_1$ and $x_2$ being the momentum fractions
1116: of the incoming gluons with momenta $\bk_1$ and $\bk_2$, the rapidity
1117: of the emitted jet given by $y=1/2 \log (x_1/x_2)$, and $\cK_{\textrm{real}}$
1118: is the real emission BFKL kernel (Lipatov vertex) in the forward direction,
1119: \beq
1120: \cK_{\textrm{real}}(\bk_1,-\bk_1;-\bk_2,\bk_2) =
1121: \frac{4 \alpha_s N_c}{(2 \pi)^2} \frac{\bk_1^2 \bk_2^2}{(\bk_1+\bk_2)^2}.
1122: \label{realkernel}
1123: \eeq
1124: The connection between the unintegrated gluon densities
1125: $\Phi_2(x,\bk,-\bk)$ and the usual gluon density $g(x,Q^2)$ is given by:
1126: \beq
1127: x g(x,Q^2) = \int_{Q_0^2}^{Q^2} \frac{d^2\bk}{\bk^2}
1128: \Bigg[ \frac{2}{(2 \pi)^4} \Phi_2(x,\bk,-\bk) \Bigg].
1129: \label{unint}
1130: \eeq
1131: In $k_t$ factorization, the kernel \eqref{realkernel}
1132: builds up the jet production subprocess in the approximation where
1133: two reggeized gluons with momenta $\bk_{1,2}$ merge into a single
1134: gluon with momentum $\bk_1+\bk_2$ which subsequently
1135: originates the observed jet:
1136: \beq
1137: \frac{d \hat{\sigma}_{gg}(x_1, \bk_1, x_2, \bk_2 , y, \bp)}{dyd^2\bp} =
1138: \frac{1}{2 s x_1 x_2}
1139: \cK_{\textrm{real}}(\bk_1,-\bk_1;-\bk_2,\bk_2)
1140: (2 \pi)^4 \delta^{(4)}(p-k_1-k_2) = \nonumber
1141: \eeq
1142: \beq
1143: =\frac{(2 \pi)^4}{s^2 x_1 x_2}
1144: \cK_{\textrm{real}}(\bk_1,-\bk_1;-\bk_2,\bk_2)
1145: \delta^{(2)}(\bp-\bk_1-\bk_2)
1146: \delta \Big( x_1-\frac{E_\perp(\bp^2)}{\sqrt{s}} e^y \Big)
1147: \delta \Big( x_2-\frac{E_\perp(\bp^2)}{\sqrt{s}} e^{-y} \Big),
1148: \label{1jethard}
1149: \eeq
1150: where $E_\perp(\bp^2)=\sqrt{m^2+\bp^2}$ is the transverse energy
1151: of the jet and $m$ its invariant mass (in our case of a single gluon
1152: we have $m=0$).
1153: Using \eqref{1jethard}, eq. \eqref{1jetfirst} can be cast in the form
1154: \bea
1155: \frac{d\sigma}{dy d^2\bp}(y,\bp) =
1156: \frac{s^2}{4(2\pi)^2}
1157: \int \frac{d^2\bk_1}{(2 \pi)^2}
1158: \int \frac{d^2\bk_2}{(2 \pi)^2}
1159: \int dx_1 \int dx_2 \nonumber \\
1160: \frac{x_1 \Phi_2(x_1; \bk_1, -\bk_1)}{\bk_1^2 \bk_1^2}
1161: \frac{d\hat{\sigma}_{gg}}{dy d^2\bp}(x_1, \bk_1, x_2, \bk_2 , y, \bp)
1162: \frac{x_2 \Phi_2(x_2; \bk_2, -\bk_2)}{\bk_2^2 \bk_2^2}
1163: \label{eq1jet}
1164: \eea
1165: which, by means of \eqref{unint}, strongly resembles the analogous formula
1166: emerging in collinear factorization.
1167:
1168: Next we consider the exchange of four reggeized gluons (Fig.\ref{fig:1jet});
1169: the corresponding production processes are illustrated in
1170: Fig.\ref{1jet-example}. The coupling of the gluons to the proton goes via
1171: the functions $\cN_4$; compared with (\ref{eq1jet}) we restrict ourselves to
1172: the limit of small $x_1$, $x_2$. We do not write down formulae,
1173: but restrict ourselves to a qualitative discussion.
1174: %figure--------------------------------------------------
1175: \begin{figure}[ht]
1176: \begin{center}
1177: \includegraphics[width=12cm]{fig/1jet.eps}
1178: \caption{Graphical representation of the process describing the
1179: single inclusive cross section.}
1180: \label{fig:1jet}
1181: \end{center}
1182: \end{figure}
1183: %figure--------------------------------------------------
1184: %figure--------------------------------------------------
1185: \begin{figure}[ht]
1186: \begin{center}
1187: \includegraphics[width=15cm]{fig/1jet-events.eps}
1188: \caption{An example of the interference among different processes that
1189: produce the cancellation. The line with the arrow in the final state
1190: correspond to the emission of the jet.}
1191: \label{1jet-example}
1192: \end{center}
1193: \end{figure}
1194: %figure--------------------------------------------------
1195: Applying the counting arguments of the previous section to Fig.\ref{fig:1jet}
1196: we will show that all contributions sum up to zero. The symmetry factors
1197: $1/2!$ etc. are the same as before; the counting of pairings is
1198: slightly different. In the previous case we have asked for
1199: multiplicities which has led us to count bound states above and below
1200: the $t$-channel intermediate state. Now we compute a one jet inclusive cross
1201: section and count the number of possibilities to attach the produced
1202: parton (jet). According to our assumptions, the couplings above and below
1203: are assumed to be symmetric under the exchange of gluon lines.
1204: In the first diagram of Fig.\ref{fig:1jet}, the number of ways of
1205: attaching the produced parton to the reggeons in the left is $1$,
1206: on the right $3$. Therefore, the symmetry factor $-1/3!$
1207: must be multiplied by $3$. The same counting applies to the third diagram
1208: of Fig.\ref{fig:1jet}. For the diagram in the center we obtain $(1/2!)^2
1209: \times 2^2$ which cancels against the other two diagrams. In summary,
1210: the four gluon corrections to the inclusive cross section
1211: Fig.\ref{fig:1jetpQCD} cancel.
1212:
1213: Several remarks have to made about this result. First, all arguments
1214: given above apply to the inclusive cross section: in Fig.16 we have
1215: illustrated, as an example, a few final states which contribute.
1216: These final states alone will not sum up to zero: the cancellations
1217: are valid only {\it after summation and integration over all final states
1218: partons other than the parton singled out by the jet (marked by an arrow)}.
1219: This means that, for individual events, these AGK cancellations are not
1220: visible. It is only after the summation over many events that the
1221: cancellations work. A necessary ingredient for this are the
1222: rescattering contributions (first and second term in Fig.\ref{1jet-example}):
1223: if they would be left out, AGK would not work.
1224:
1225: Second, the observed jet has introduced the hard scale which is necessary
1226: for justifying the use of gluon ladders. Such a hard scale is not present
1227: in the other ladders (e.g. in the uncut ladder in the first diagram of Fig.16,
1228: or in the cut ladder in the last graph of Fig.16):
1229: in the inclusive cross section we sum over final states which might include a
1230: large fraction of soft final states. So, strictly
1231: speaking, we have demonstrated only the cancellations between hard final
1232: states. This is certainly important for the modelling of multiple parton
1233: interactions which lead to the production of partons in the final states.
1234: However, it is important to include also soft rescattering,
1235: i.e. the additional exchange of nonperturbative Pomerons. This can be done
1236: by combining the perturbative discussion of this subsection
1237: with the nonperturbative one given in subsection 3.1 There
1238: we have noted that, when adding soft Pomerons to a single cut
1239: Regge pole, the sum over all cuts across the soft Pomerons cancels.
1240: If we simply substitute our hard cut Pomeron (containing the
1241: produced jet) for this single cut Regge pole,
1242: we conclude that, in our pQCD inclusive cross section,
1243: also all additional soft Pomeron exchanges cancel. This coincides
1244: with the well known result in the collinear
1245: factorization which follows from the QCD factorization theorems~\cite{CSS}.
1246:
1247: A final comment applies to multiple exchanges between the produced jet
1248: and one of the hadron projectiles. If one of the two momentum fractions,
1249: say $x_1$, becomes very small, saturation effects are expected to
1250: become important: first corrections of this type are shown in
1251: Fig.\ref{surviving-terms}.
1252: Obviously, they require higher order jet production vertices which
1253: have not yet been calculated. These vertices
1254: are somewhat analogous to the lowest order coupling of two BFKL ladders to the
1255: photon, i.e. to the process: photon + BFKL $\to$ quark-antiquark. Namely,
1256: if we open, in the first figure of Fig.\ref{surviving-terms},
1257: the cut Pomeron below the jet vertex, we can view this vertex
1258: as the square of the subprocess: gluon + BFKL $\to$ jet. An important
1259: difference between the two cases lies in the fact that
1260: the incoming virtual photon is replaced by a colored gluon with transverse
1261: momentum $\bk_2$. In this context it is to be expected that the reggeization
1262: of the gluon will be an issue: similar to our remarks on $\cD_4$ (second part
1263: of section 2, eq.(\ref{D4decomp})), there will be pieces which belong to
1264: antisymmetric two-gluon states and require a separate discussion.
1265: Another question of particular
1266: interest is the applicability of the dipole picture: in contrast to the
1267: color singlet photon, the incoming gluon carries color and might lead to
1268: changes of the impact factor.
1269:
1270: We conclude this section with the generalization of our discussion to an
1271: arbitrary number of reggeized gluons: in the appendix we show that
1272: the cancellation works for a general (even) number of additional
1273: reggeized gluons. This leads to the remarkable conclusion that
1274: there are no multi-Pomeron corrections to the basic process illustrated
1275: in Fig.14: all soft or hard exchanges between the upper and lower
1276: projectiles cancel. What remains are only the multiple exchanges between
1277: the produced gluon and the upper (or lower) proton.
1278:
1279: \subsection{Inclusive double jet production}\label{2jets}
1280: We now discuss the inclusive production of two jets, the Mueller-Navelet
1281: cross section.
1282: %figure--------------------------------------------------
1283: \begin{figure}[ht]
1284: \begin{center}
1285: \includegraphics[height=5cm]{fig/2jet-2reggeons.eps}
1286: \caption{a) Leading term for the inclusive production of two jets with a
1287: large rapidity interval between them. b) Definition of the two-to-two
1288: reggeized gluons BFKL Green function.}
1289: \label{2jets-2reggeons}
1290: \end{center}
1291: \end{figure}
1292: %figure--------------------------------------------------
1293: The leading contribution is shown in Fig.\ref{2jets-2reggeons}a: when the
1294: rapidity interval between the two observed jets is large the process
1295: is described by Regge kinematics, and the large rapidity interval is due
1296: to the BFKL Green function $\cG_2$, which contains the exchanges of
1297: gluons between the reggeized gluons (Fig. \ref{2jets-2reggeons}b);
1298: note that, in our convention, $\cG_2$ does not contain the propagators
1299: for the external gluons.
1300:
1301: The formula for the cross section associated with Fig.\ref{2jets-2reggeons}a is
1302: \beq
1303: \frac{x_1 x_2 d^2\sigma}{dx_1 dx_2 d^2\bp_1 d^2\bp_2}
1304: (x_1,x_2,y,\bp_1,\bp_2) =
1305: \frac{1}{4}
1306: \int \frac{d^2\bk}{(2 \pi)^2}~ \frac{d^2\bk'}{(2 \pi)^2} \nonumber
1307: \eeq
1308: \beq
1309: \frac{\Phi_2( x_1; \bk, -\bk )}{\bk^2\bk^2}
1310: \cK_{\textrm{real}}(\bk,-\bk;\bk-\bp_1,-\bk+\bp_1) \nonumber
1311: \eeq
1312: \beq
1313: \frac{{\cal G}_2
1314: ( y ; \bk-\bp_1 , -\bk+\bp_1 ; -\bk'+\bp_2 , \bk'-\bp_2 )}
1315: {(\bk-\bp_1)^2 (\bk-\bp_1)^2 (\bk'-\bp_2)^2 (\bk'-\bp_2)^2}
1316: \label{MNnf}
1317: \eeq
1318: \beq
1319: \cK_{\textrm{real}}(\bk',-\bk';\bk'-\bp_2,-\bk'+\bp_2)
1320: \frac{\Phi_2( x_2; \bk', -\bk' )}{\bk'^2\bk'^2}. \nonumber
1321: \eeq
1322: Defining $\tilde{\cal G}_2$ as the usual Green function but with the
1323: propagators for the gluons on the left side of the cut,
1324: \beq
1325: \tilde{\cal G}_2(y; \bk_1,\bk_2; \bk'_1,\bk'_2 ) =
1326: \frac{{\cal G}_2(y; \bk_1,\bk_2; \bk'_1,\bk'_2 )}{\bk_1^2 \bk'^2_1}
1327: \label{green_scaled}
1328: \eeq
1329: and using \eqref{realkernel}, \eqref{MNnf} can be rewritten as
1330: \beq
1331: \frac{x_1 x_2 d^2\sigma}{dx_1 dx_2 d^2\bp_1 d^2\bp_2}
1332: (x_1,x_2,y,\bp_1,\bp_2) = \nonumber
1333: \eeq
1334: \beq
1335: \frac{1}{4 (2 \pi)^4}
1336: \frac{1}{\bp_1^2 \bp_2^2}\bigg[\frac{4 \alpha_s N_c}{(2 \pi)^2}\bigg]^2
1337: \int d^2\bk
1338: \frac{\Phi_2( x_1; \bk, -\bk )}{\bk^2}
1339: \int d^2\bk'
1340: \frac{\Phi_2( x_2; \bk', -\bk' )}{\bk'^2}
1341: \label{MNnfb}
1342: \eeq
1343: \beq
1344: \tilde{\cal G}_2
1345: ( y ; \bk-\bp_1 , -\bk+\bp_1 ; -\bk'+\bp_2 , \bk'-\bp_2 ). \nonumber
1346: \eeq
1347: Note that in practice, because of the limited energy of the hadron
1348: collider (e.g. the Tevatron), the kinematics of the Mueller-Navelet is
1349: chosen such that it maximizes the rapidity gap
1350: between the two jets; this implies that the momentum fractions $x_1$ and
1351: $x_2$ are not necessarily small, i.e. the separations in rapidity between
1352: the jets and the projectiles $A$ and $B$ are not particularly large.
1353: As a result, we have strong $k_t$ ordering between the jets and the
1354: projectiles, and we are led to the usual integrated gluon densities
1355: (eq.(\ref{unint})). The cross section \eqref{MNnf} can be cast in the form
1356: \beq
1357: \frac{x_1 x_2 d^2\sigma}{dx_1 dx_2 d^2\bp_1 d^2\bp_2}
1358: (y,x_1,x_2,\bp_1,\bp_2) = \nonumber
1359: \eeq
1360: \beq
1361: =
1362: x_1 g(x_1,M^2)~
1363: \bigg[\frac{\alpha_s N_c}{\bp_1^2}\bigg]
1364: \tilde{\cal G}_2(y;-\bp_1,\bp_1;\bp_2,-\bp_2)
1365: \bigg[\frac{\alpha_s N_c}{\bp_2^2}\bigg]~
1366: x_2 g(x_2,M^2)
1367: \label{eq:MN}
1368: \eeq
1369: which is the well known Mueller-Navelet formula for dijet production in
1370: hadron-hadron scattering \cite{MN}. Also, the incoming gluons with momenta
1371: $\bk_1$ and $\bk_2$ may be replaced by quark lines, and we have additional
1372: contributions from the quark densities. Throughout our discussion of the
1373: AGK counting, however, we restrict ourselves to the region of small
1374: $x_1$ and $x_2$.
1375:
1376: The first correction to \eqref{eq:MN} comes from the exchange of four
1377: reggeized gluons. Our discussion requires small $x_1$ and $x_2$, and the
1378: coupling of the four gluons proceeds via the function $\cN_4$.
1379: The situation is akin to that shown in Fig.\ref{fig:1jet},
1380: but now there are two vertices for the jet emission on each side of the cut.
1381: Between the two vertices, the four reggeized gluons interact exchanging gluons;
1382: this interaction is symmetric under the exchange of two gluon lines.
1383: We denote this interaction by the Green's function ${\cal G}_4$ (which
1384: technically is obtained from the corresponding BKP
1385: kernel \cite{BKP,Jaroszewicz} describing the evolution of $n$ interacting
1386: reggeized gluon states in the $t$-channel).
1387: Following the same line of counting as before, we will find that only a
1388: particular subset of the terms gives a non vanishing contribution to the
1389: inclusive observable.
1390: %figure--------------------------------------------------
1391: \begin{figure}[ht]
1392: \begin{center}
1393: \includegraphics[width=10cm]{fig/2jet.eps}
1394: \caption{We show the central cut as an example of the classification
1395: of the different terms which a priori should contribute to the cross section:
1396: a) both vertices attached to the same gluon on both side, b) to the same
1397: gluon on one side but to different gluons to the other, and c) to different
1398: gluons on both sides. In d) we show the definition of the four-to-four
1399: reggeized gluons Green function.}
1400: \label{fig:2jets}
1401: \end{center}
1402: \end{figure}
1403: %figure--------------------------------------------------
1404: We treat three different cases, which are illustrated in
1405: Fig.\ref{fig:2jets} (where only the central cut case is shown):\\
1406: (a) Both production vertices are attached to the same reggeized gluons
1407: on each side of the cut; in this case the counting is exactly the same as
1408: in the previous section, and the sum of the correction terms vanishes.\\
1409: (b) Both vertices are attached to the same reggeized gluon on one side
1410: of the cut, but to different reggeized gluons on the other side.
1411: For example, suppose that the vertices
1412: are connected to the same reggeized gluon on the right side
1413: (the other case is identical).
1414: The first cut (between $1$ and $2$) does not contribute since there is only
1415: one reggeized gluon on the left side.
1416: The combinatorial factor for the central cut (between $2$ and $3$) is $2$
1417: (the reggeized gluon whom to attach the right sides of the vertices in the
1418: right side of the cut can be chosen among two) times $2$
1419: (the left side of the first vertex can be attached to one of the two
1420: reggeized gluon on the left, but the other vertex can only be attached
1421: to the one left); together with
1422: the symmetry factor $1/4$, the contribution of the diagram is $1$.
1423: Finally, the combinatorial factor of the third
1424: cut (between $3$ and $4$) is $3 \times 2$ because one of the vertices
1425: can be attached to one of the three reggeized gluons on the left but
1426: the other one to one of the two not connected to the first one.
1427: Combining this with the symmetry factor $-1/6$, the contribution
1428: of the diagram becomes $-1$,
1429: which cancels exactly the $+1$ of the central cut.\\
1430: (c) The vertices are attached to different reggeized gluons on both sides of
1431: the cut: only one diagram contributes, and there are no cancellations.
1432:
1433: What emerges from this analysis of the four gluon exchange is that those
1434: diagrams, in which there are reggeized gluons without a production vertex,
1435: give a vanishing contribution. This statement holds for an arbitrary number of
1436: reggeized gluons and any number of jets; in particular, diagrams
1437: involving the exchange of more that four reggeized gluons do not contribute
1438: to the double inclusive jet production. A general proof is given in the
1439: Appendix.
1440:
1441: %figure--------------------------------------------------
1442: \begin{figure}[ht]
1443: \begin{center}
1444: \includegraphics[height=5cm]{fig/2jet-4reggeons.eps}
1445: \caption{a) Non vanishing correction to the two jets inclusive jet production
1446: due to the exchange of four reggeized gluons.}
1447: \label{2jets-4reggeons}
1448: \end{center}
1449: \end{figure}
1450: %figure--------------------------------------------------
1451: For the case of the Mueller-Navelet jets we are thus left with only one
1452: type of corrections (Fig.\ref{2jets-4reggeons}): the jets are emitted
1453: from different reggeized gluons on both sides of the cut.
1454: We write the explicit formula for the correction represented in
1455: Fig.\ref{2jets-4reggeons} in the following form:
1456: \beq
1457: \frac{x_1 x_2 d^2\sigma}{dx_1 dx_2 d^2\bp_1 d^2\bp_2}(y,x_1,x_2,\bp_1,\bp_2) =
1458: \frac{1}{4}
1459: \int \Bigg[ \prod_{i=1}^4
1460: \frac{d^2\bk_{i}}{(2 \pi)^2}
1461: \frac{d^2\bk'_{i}}{(2 \pi)^2} \Bigg]
1462: (2 \pi)^2\delta^{(2)}(\Sigma_i \bk_i)
1463: (2 \pi)^2\delta^{(2)}(\Sigma_i \bk'_i) \nonumber
1464: \eeq
1465: \beq
1466: \frac{\Phi_4(x_1;\bk_1,\bk_2,\bk_3,\bk_4)}
1467: {\bk_1^2 \bk_2^2 \bk_3^2 \bk_4^2}~
1468: \frac{\cK_{\textrm{real}}(\bk_1,\bk_4;\bk_1-\bp_1,\bk_4+\bp_1)}
1469: {(\bk_1-\bp_1)^2 (\bk_4+\bp_1)^2} \nonumber
1470: \eeq
1471: \beq
1472: {\cal G}_4 \bigg(y;
1473: \begin{array}{c}
1474: \bk_1-\bp_1 , \bk_2 , \bk_3 , \bk_4+\bp_1 \\
1475: \bk'_1 , \bk'_2-\bp_2 , \bk'_3+\bp_2 , \bk'_4
1476: \end{array}
1477: \bigg)~
1478: \label{MN4g}
1479: \eeq
1480: \beq
1481: \frac{\cK_{\textrm{real}}(\bk'_2,\bk'_3;\bk'_2-\bp_2,\bk'_3+\bp_2)}
1482: {(\bk'_2-\bp_2)^2 (\bk'_3+\bp_2)^2}~
1483: \frac{\Phi_4(x_2;\bk'_1,\bk'_2,\bk'_3,\bk'_4)}
1484: {\bk'^2_1 \bk'^2_2 \bk'^2_3 \bk'^2_4} \nonumber
1485: \eeq
1486: Here we have used another notation for the coupling of the reggeons
1487: to the proton, $\Phi_4$ rather than $\cN_4$: in the limit of small
1488: $x_i$, $\Phi_4$ coincides with $\cN_4$. In a more realistic
1489: situation we might allow for strong ordering of transverse momenta
1490: in those reggeized gluons which connect the jet vertex (cf.our discussion
1491: after \eqref{MNnfb}). As a consequence, the pattern of integrations
1492: in \eqref{MN4g} is more involved than it was in \eqref{MNnfb}.
1493: In particular, the relation between the coupling $\Phi_4$
1494: and the vertex functions $\cN_4$ requires a more detailed discussion.
1495:
1496: We finally turn our attention to the Green function $\cG_4$; since
1497: only the central cut survives, there are only diagrams corresponding
1498: to multiplicity $k=0,2$. The diffractive case (i.e. the first diagram
1499: drawn in Fig.\ref{fig:2jets}d) is known in the literature as ``hard color
1500: singlet'' \cite{HCS}: there is a rapidity gap between the two observed jets.
1501: Performing the usual counting it is trivially verified
1502: that the weight factors of the double multiplicity part of
1503: the cut Green function ($k=2$) is twice the one of the diffractive part
1504: ($k=0$) and, as before, a switch of the pairing structure
1505: inside $\cG_4$ does not spoil this relation.
1506:
1507: Specific models ~\cite{Kovchegov1_2,Braun2005} require the discussion
1508: of reggeon diagrams which are more complicated than those discussed in this
1509: paper; in particular, this includes diagrams which change the number
1510: reggeized gluon in the $t$-channel. For any $t$-channel state with
1511: a fixed number of gluons, $2n$, our discussion applies. However, when
1512: considering, for example, double inclusive jet cross sections where
1513: between the two emissions the number of gluons changes, one expects to see
1514: new contributions where one of the produced jet originates from the
1515: gluon number changing vertex. As we have said above, the application of
1516: of AGK cutting rules to such vertices requires a separate discussion.
1517:
1518: \subsection{Remarks on the relation to phenomenological models}
1519:
1520: The topic of multiple interactions has been addressed for many years
1521: (see, for example, \cite{SS} and references therein, or \cite{RT}).
1522: Key elements are multi-parton distributions, which are interpreted as
1523: probabilities of finding, inside the hadron, a number
1524: $n_p \ge 1$ of partons with longitudinal momentum fractions
1525: $\{x_i\},i=1,...,n_p$. In the framework of $k_t$-factorization
1526: the partons also carry transverse momenta $\{k_i\}$.
1527: These partons then interact through hard subprocesses
1528: and produce partonic final states. In the final step, color strings between
1529: the produced partons and the remnants of the hadrons describe the
1530: hadronization.
1531:
1532: The AGK analysis of this paper has mostly been formulated in terms of angular
1533: momenta, which are conjugate to energy variables. When translating our
1534: results, one first has to emphasize that all the AGK analysis applies to
1535: the limit of small momentum fractions $x_i$. In particular, in the coupling
1536: functions $\cN_n$ the gluons have longitudinal momentum
1537: fractions $x_i$ which are all small and defined to be all of the same order.
1538: Hence, in our notation we have not specified the dependence upon the
1539: $x_i$ but rather used the conjugate variable of total angular momentum
1540: $\omega$.
1541: However, as we have discussed after eq.(\ref{gammafactor}) and
1542: illustrated by our example of the eikonal model,
1543: it is possible to define more general coupling functions $\cN$ by introducing
1544: a dependence upon angular momenta of subsystems of gluons, e.g. $\omega_{ij}$.
1545: This corresponds to momentum fractions which are small but of different
1546: order of magnitude. But, as usual
1547: in the small-$x$ approximations, conservation of longitudinal momenta
1548: is not observed, and therefore a detailed assignment of longitudinal
1549: momenta is not easy. Therefore, as far as the modelling of multiple
1550: interactions is concerned,
1551: the AGK analysis presented in this paper should mainly be
1552: viewed as providing constraints for the limit of small-$x$ values.
1553: For a hadron collider which operates at energies as high as the LHC,
1554: these constraints should be quite essential.
1555: %figure--------------------------------------------------
1556: \begin{figure}[ht]
1557: \begin{center}
1558: \includegraphics[height=5cm]{fig/3jet.eps}
1559: \caption{Some diagrams contributing to the three jet production
1560: notwithstanding the AGK cancellations.}
1561: \label{3jet}
1562: \end{center}
1563: \end{figure}
1564: %figure--------------------------------------------------
1565:
1566: From the AGK analysis it becomes quite clear that
1567: there exist contributions to the cross section which cannot
1568: simply be interpreted as probability distributions.
1569: Examples have been illustrated in Fig.16: whereas the
1570: third contribution represents a square of production amplitudes, and
1571: the couplings of the gluons to the upper (or lower) proton might be
1572: interpreted as a 'probability of finding two gluons inside a proton',
1573: the first and the second contributions are interference terms.
1574: Another example of a probabilistic contribution is shown in
1575: Fig.\ref{2jets-4reggeons}. The third diagram in Fig.20, on the other hand,
1576: illustrates an another interference term which survives after all AGK
1577: cancellations have been worked out.
1578:
1579: One is therefore lead to a more general concept of {\it multiparton
1580: correlators}. Even in the framework of collinear
1581: factorization at the leading twist level, there is already the well-known
1582: example of
1583: the \emph{Generalized Parton Distributions} (see \cite{GPD} for a review),
1584: which represent correlations between two partons inside the proton.
1585: In this case it is possible to obtain direct experimental information
1586: of these correlations by exploiting, for
1587: example, the interference between the DVCS
1588: (Virtual Compton Scattering) and the BH (\emph{Bethe-Heitler})
1589: process \cite{DVCS}. Beyond these two-parton correlators
1590: very little has been worked out in the literature.
1591: %%%%%%%%%%%%%%%%%%%%%%%%% conclusions %%%%%%%%%%%%%%%%%%%%%%%%%%%%
1592: \section*{Conclusions}
1593: In this paper we have discussed the origin and a few consequences of the
1594: AGK cutting rules in pQCD where the Pomeron is described by the
1595: BFKL gluon ladders. We have identified the basic ingredient for the
1596: validity of the AGK cutting rules, the symmetry properties of
1597: couplings of $n$ reggeized gluons to the projectile; they are in
1598: agreement with general properties of reggeon unitarity equations,
1599: and they have explicitly been verified in pQCD in $\gamma^* \gamma^*$
1600: (or onium-onium scattering).
1601:
1602: As to consequences of the AGK counting rules, we have considered
1603: single and double inclusive jet production. In both cases,
1604: multipomeron exchanges across the produced jets cancel; this holds
1605: for both soft and hard Pomerons. In the two-jet inclusive case, there
1606: exists an extra contribution in which the jets are emitted from two
1607: different parton chains. As an example, we have analysed how this
1608: contribution looks like in in the Mueller-Navelet jet final states;
1609: this contribution turns out to be related also to the hard color singlet
1610: exchange cross section.
1611:
1612: At very small $x$ values, QCD saturation effects are expected to become
1613: important. Our AGK analysis leads to the conclusion that such nonlinear
1614: corrections should be modelled as multiple exchanges between the jet vertices
1615: or between jet vertices and either of the projectiles,
1616: but not between the two projectiles.
1617:
1618: Several important questions have not been addressed in this paper,
1619: and our analysis of AGK counting rules in proton proton scattering
1620: remains incomplete.
1621: First, QCD allows two-gluon states in the $t$-channel that are
1622: antisymmetric under
1623: momentum and color exchange: in pQCD, bootstrap equations have been shown
1624: to be valid, as a result of which the such two-gluon states are identical
1625: to a single reggeized gluon. We conjecture that this property holds also
1626: beyond pQCD: in this case, as far as the reggeization of gluons is
1627: concerned, we do not need to introduce new couplings to the hadron, and our
1628: discussion can be reduced to the symmetric functions which ensure the validity
1629: of the AGK counting rules. Nevertheless, the existence of these
1630: reggeizing pieces require further studies: the most prominent example
1631: is $\gamma^*$-proton scattering (i.e. the deep inelastic structure
1632: function), where in leading order the coupling of four reggeized
1633: gluons to the virtual photon consist of reggeized pieces only. This
1634: process will be studied in a forthcoming paper.
1635:
1636: In this paper we also have not yet included the Odderon. Perturbative
1637: QCD contains an Odderon state, built from (at least) three interacting
1638: reggeized gluons. It requires couplings to the proton which satisfy
1639: special symmetry properties.
1640: Since the intercept of the highest Odderon state known so far
1641: ~\cite{BLV} lies at unity, at high energies it will be less important;
1642: a complete analysis, however, will have to include Odderon exchange.
1643:
1644: Another aspect which requires further studies is the limitation
1645: due to finite energies. Strictly speaking, the AGK analysis in terms
1646: of ladder diagrams requires infinite energies: at any finite energy,
1647: the number of produced gluons inside a cut ladder is limited, and,
1648: therefore, also the number of different 'cut ladders', $k$, cannot be
1649: arbitrarily large. When modelling multiple parton interactions,
1650: this may have consequences (e.g. incomplete cancellations) which have to
1651: worked out in detail.
1652: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1653: \section*{Acknowledgements}
1654: M.S. is grateful for the kind hospitality to the \emph{II Institut
1655: f\"ur Theoretische Physik} of the Hamburg University where part
1656: of this work was done and the \emph{Programma Marco Polo} for
1657: the financial support. G.P.V. is very grateful to the Hamburg
1658: University for the warm hospitality.
1659: We are grateful to Markus Diehl for his numerous helpful comments.
1660: %%%%%%%%%%%%%%%%%%%%%%%%% Appendix %%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1661: \newpage
1662: \section*{Appendix}
1663: In this appendix we present a few details of the
1664: generalization of our results to $2n$ gluons.
1665: We start with a few notations.
1666: Let $\tA^n(s,t)$ denote the energy discontinuity of the contribution
1667: of $n$-reggeized gluon exchange in the elastic amplitude for the process
1668: $AB \to AB$, $\tF^n(\o,\bq)$.The energy discontinuity can be written as the
1669: sum over
1670: products of production amplitudes $\tilde{\cT}_{AB \to X}^{j}(s,t)$
1671: for the process $AB \to X$, where the index $j$ stands for the number
1672: of exchanged reggeized gluons. Each such term is of the form
1673: $\tilde{\cT}_{AB \to X}^{j}(s,t) (\tilde{\cT}_{AB \to X}^{2n-j}(s,t))^*$,
1674: i.e. $j$ gluons are exchanged inside the ``left''
1675: amplitude, $2n-j$ in the ``right'' complex conjugate amplitude,
1676: and the total number of exchanged gluons is $2n$.
1677: In order to construct $\tA^n(s,t)$,
1678: we have to sum over $j=1,...,2n-1$.
1679: Furthermore, we also have to sum (denoted by
1680: $\sum_\textrm{B.S.}$) over all the possibilities to form bound states
1681: of two-reggeized gluons bound states (pomerons) inside the amplitudes
1682: $\cN_{2n}$:
1683: \beq
1684: \tA^n(s,t) \stackrel{\textrm{def}}{=}
1685: 2 i \sum_{\textrm{B.S.}}\sum_{j=1}^{2n-1}
1686: \int d \Omega_X~
1687: \tilde{\cT}_{AB \to X}^{j}(s,t_1)~
1688: \big(\tilde{\cT}_{AB \to X}^{2n-j}(s,t_2)\big)^*
1689: = \nonumber
1690: \eeq
1691: \beq
1692: =2 i \sum_{\textrm{B.S.}}\sum_{j=1}^{2n-1}
1693: \int d\Omega_{2n}~s^{1+\tilde{\beta}}~ \cN_{2n}~ \cN_{2n}~
1694: \frac{i\xi_\mathds{G}(\bk_1)...i\xi_\mathds{G}(\bk_j)}
1695: {j!}
1696: \frac{[i\xi_\mathds{G}(\bk_{j+1})...i\xi_\mathds{G}(\bk_{2n})]^*}
1697: {(2n-j)!}
1698: \label{discspQCD}
1699: \eeq
1700: Here we have made used of the fact that $\cN_{2n}$ does not depend on the
1701: position of the cut. The tilde symbol indicates that we are working in pQCD.
1702: Since the signature
1703: factor $\xi_\mathds{G}(\bk)= - 2/\pi \bk^2$ is real
1704: (cf. the discussion after eq.(13)),
1705: \eqref{discspQCD} can be cast into the form
1706: \beq
1707: \tA^n(s,t) = 2^n \pi i \sum_{\textrm{B.S.}} \sum_{j=1}^{2n-1} S^n_j~
1708: \int d\Omega_{2n}~s^{1+\tilde{\beta}_n}~ \cN_{2n}~ \cN_{2n}~
1709: \tilde{\gamma}_{\{\bk_l\}}
1710: \label{discspQCD1}
1711: \eeq
1712: where we have used \eqref{pQCDgamma}, and we have introduced the symmetry
1713: factor
1714: \beq
1715: S^n_j=\frac{(-1)^{n-j}}{j!(2n-j)!}.
1716: \label{symfac}
1717: \eeq
1718:
1719: \subsection*{Inclusive case}
1720: We begin with the fully inclusive case of section 3.
1721: Starting from \eqref{discspQCD} we want to find a decomposition in terms
1722: of cut Pomerons.
1723: For each position of the cut (denoted by $j$) we first classify all the
1724: different possibilities of forming $2$-gluon bound states out of the $2n$
1725: reggeized gluons; at the same time we have to keep track of
1726: how many of these bound states are cut by the energy cutting line.
1727: Let $k$ denote the number of such cut bound states: then $k$ also
1728: labels the multiplicity of $s$-channel gluons intersected by the
1729: cutting lines. We then must count how many different
1730: configurations $(i_1j_1)...(i_nj_n)$ contribute to a term belonging to
1731: the multiplicity $k$.
1732:
1733: Denoting by $C^n$ the number of different pairings among $2n$ reggeized
1734: gluons, is clear that $C^n=(2n-1)!!$:
1735: starting with the first reggeized gluon to the left of the cutting line,
1736: its partner can be chosen among $2n-1$ other $t$-channel gluons;
1737: the partner of the next unpaired reggeized gluon can be chosen among $2n-3$
1738: other gluons, and so forth until all the reggeized gluons have been put into
1739: pairs. Next we will decompose $C^n$ into contributions with
1740: fixed-multiplicity $k$:
1741: \beq
1742: C^n = \sum_k C^n_{jk},
1743: \label{cnsum}
1744: \eeq
1745: where $C^n_{jk}$ is the number of configurations that
1746: contain exactly $k$ cut pairs. The subscript $j$ indicates that this
1747: number will depend upon the position of the cutting line.
1748: It is clear that $C^n_{jk}$ is $0$ unless $k$ and $j$ are both even or both
1749: odd, i.e. $C^n_{jk} \propto 1/2(1+(-1)^{j+k})$. Also, we need $k\le j$ and
1750: $k\le 2n-j$, i.e. the range of $k$ values in \ref{cnsum} depends upon $j$.
1751: If $j$ and $k$ satisfy these conditions
1752: we can chose $k$ reggeized gluons in each side of the cut in $\binom{j}{k}
1753: \binom{2n-j}{k}$ different ways and couple each reggeized gluon in the left
1754: side with a reggeized gluon in the right side (this can be done in $k!$ ways).
1755: We are left with $j-k$ reggeized gluons on the left side and $2n-j-k$
1756: on the right side, which must be coupled in pairs without crossing the cut; there are
1757: $(j-k-1)!!(2n-j-k-1)!!$ ways to do that. Eventually the expression obtained is
1758: \beq
1759: C^n_{jk} = \frac{1+(-1)^{j+k}}{2}\frac{j! (2n-j)!}{k!(j-k)!!(2n-j-k)!!},
1760: \label{cmjk}
1761: \eeq
1762: and it is easy to verify that \eqref{cmjk} satisfies \eqref{cnsum}.
1763:
1764: With this expression for $C^n_{jk}$ we can rewrite
1765: \eqref{discspQCD} as a sum over the multiplicity index $k$
1766: of fixed-multipicity contributions $\tA_k^n(s,t)$:
1767: \beq
1768: \tA^n(s,t) = \sum_{k=0}^n \tA_k^n(s,t)
1769: \label{pQCDAGKdec}
1770: \eeq
1771: where
1772: \beq
1773: \tA_k^n(s,t) = 2^n \pi i \sum_{j=k}^{2n-k} S^n_j C^n_{jk}~
1774: \int d\Omega_{2n}~s^{1+\tilde{\beta}}~
1775: \cN_{2n}~ \cN_{2n}~
1776: \tilde{\gamma}_{\{\bk_l\}}
1777: \label{pQCDAk}
1778: \eeq
1779: The reader should note that the integral in \eqref{pQCDAk} does
1780: not depend upon $j$ or $k$; therefore there is just a combinatorial factor
1781: in front of the integral. Performing the sum in \eqref{pQCDAk} we arrive
1782: at:
1783: \bea
1784: \sum_{j=k}^{2n-k} S^n_j C^n_{jk} &=&
1785: \frac{(-1)^n}{2~k!}\sum_{j=k}^{2n-k}
1786: \frac{(-1)^j+(-1)^k}{(j-k)!!(2n-j-k)!!} \nonumber \\
1787: &=& \Bigg\{ \begin{array}{ll}
1788: \frac{(-1)^n}{n!} ( 1-2^{1-n} ) & \textrm{if $k=0$} \\
1789: \frac{(-1)^{n-k}}{k!(n-k)!} & \textrm{if $k>0$}
1790: \end{array}.
1791: \label{cutcounting}
1792: \eea
1793: With the definition \eqref{tildeF}, we have shown that
1794: \eqref{pQCDAk} can be written as in \eqref{agkpQCD}.
1795:
1796:
1797: \subsection*{Switches of Pairings}
1798: Now we want to show that a switch from a configuration
1799: $(i_1j_1) ... (i_nj_n)$ to $(i'_1j'_1) ... (i'_nj'_n)$
1800: preserves the relative weight between different multiplicities $k$.
1801: The situation is the following: at some point in rapidity we have the
1802: configuration $(i_1j_1) \dots (i_nj_n)$ with muliplicity
1803: $k$, and the weight factor of this multiplicity $k$ is the usual $\tF^n_k$
1804: given in \eqref{tildeF}; moving now to a neighbouring rapidity interval
1805: we switch to a different configuration $(i'_1j'_1) \dots (i'_nj'_n)$,
1806: and we want to compute the weight of the term with $k'$ cut rungs.
1807: Contributions to this $k'$ can come from different $k$ terms before the
1808: switch; we therefore must sum over $k$ the $\tF^n_k$ muliplied by the
1809: number of ways of obtaining a configuration with $k'$ cut rungs
1810: (normalized by the total number).
1811:
1812: Put in other words, we must build an $(n+1) \times (n+1)$ matrix, where the
1813: initial multiplicity $k$ labels the columns, and the final multiplicity
1814: $k'$ labels the rows; the elements ${\cal M}^n_{k'k}$ of this
1815: matrix are defined to be the fraction of configurations
1816: which, after the switch, lead to
1817: $k'$ cut rungs. We then want to show that the vector $\tF^n_k$ is
1818: an eigenvector of this matrix ${\cal M}^n_{k'k}$.
1819: Using a vector notation and dropping the index $n$ we want to show that:
1820: \beq
1821: \mbf{\cF} \propto {\cal M} \mbf{\cF}.
1822: \label{eq:ev}
1823: \eeq
1824: But $\mbf{\cF}$ is an eigenvector of ${\cal M}$ iff it is an eigenvector
1825: of ${\cal M}+c \mathds{1}$ where $c$ is an arbitrary constant.
1826: Note that, by definition, ${\cal M}$ is computed by considering only
1827: transitions between \emph{different pairwise configurations}, and a
1828: transition to the same configuration, in our matrix notation,
1829: is proportional to the identity (the nature of the cut does not change).
1830: The proportionality constant is just $1/C^n$, where $C^n$ is the total number
1831: of configurations defined before.
1832: We normalize the matrix ${\cal M}$ by dividing by $C^n-1$
1833: instead of $C^n$, because
1834: we are considering only transitions to different configurations; of course,
1835: this normalization does not affect the form of the eigenvectors (it changes
1836: only the eigenvalues). So we can change its normalization multiplying by
1837: $(C^n-1)/C^n$. Eventually, proving \eqref{eq:ev} is equivalent to prove
1838: \beq
1839: \mbf{\cF} \propto \overline {\cal M} \mbf{\cF},
1840: \label{eq:ev2}
1841: \eeq
1842: where we have introduced the new matrix $\overline {\cal M}$ defined by
1843: \beq
1844: \overline {\cal M} = \frac{(C^n-1) {\cal M} + \mathds{1}}{C^n}.
1845: \eeq
1846: This new matrix $\overline {\cal M}$ now contains \emph{all possible switches}
1847: to a new configuration, not only to the different ones.
1848:
1849: Instead of computing explicitely the coefficients of $\overline {\cal M}$,
1850: it is easier to compute directly the RHS of \eqref{eq:ev2}, i.e.
1851: write the sum over the various contributions due to different positions of
1852: the cut before the switch. For fixed $j$ and $k$ it is trivial to obtain
1853: the fraction of configurations with $k'$ cut rungs: it is $C^n_{jk'}/C^n$
1854: (note that this is true only because we are no longer restricting ourselves
1855: to the new configuration being different from
1856: the previous one). Therefore, using
1857: \beq
1858: \tF^n_k = 2^{n-1} n!~
1859: \sum_{j=k}^{2n-k} S^n_j C^n_{jk},
1860: \eeq
1861: the rhs of \eqref{eq:ev2} can be written as
1862: \beq
1863: \sum_{k=0}^{n}
1864: 2^{n-1} n!~
1865: \sum_{j=k}^{2n-k} \frac{C^n_{jk'}}{C^n} S^n_j C^n_{jk}.
1866: \eeq
1867: After using \eqref{cnsum}, this gives exactly $\tF^n_{k'}$, i.e.
1868: the $\tF^n_{k}$ form an eigenvector.
1869:
1870:
1871: \subsection*{$1$-Jet inclusive}
1872: In this subsection we generalize the cancellation of diagrams with
1873: $n>2$ gluon lines in the single inclusive cross section. The discontinuity
1874: of the amplitude for single jet production is
1875: given by
1876: \beq
1877: \tA^n_{1\textrm{-jet}}(s,t) =
1878: 2 i \sum_{\textrm{B.S.}}\sum_{j=1}^{2n-1} j(2n-j) S^n_j~
1879: \int d\Omega_{2n}~s^{1+\tilde{\beta}}~ \cN_{2n}~ \cN_{2n}~
1880: \tilde{\gamma}_{\{\bk_l\}},
1881: \label{discspQCD1jet}
1882: \eeq
1883: where the factor $j(2n-j)$ counts the number of ways in which the jet can be
1884: connected to the reggeized gluons on the lhs and rhs of the cutting line.
1885: Since now we are not interested in counting the number of cut pomerons,
1886: the sum over different bound state configuration gives just a global factor
1887: $C^n$.
1888: Performing the summation over the position of the cut $j$ we get
1889: \bea
1890: \sum_{j=1}^{2n-1} j (2n-j) S^n_j &=&
1891: (-1)^n \sum_{1}^{2n-1} \frac{(-1)^j}{(j-1)!(2n-j-1)!} \nonumber \\
1892: &=& (-1)^{n-1} \sum_{0}^{2n-2} \frac{(-1)^j}{j!(2n-j-2)!} \\
1893: &=& \Bigg\{ \begin{array}{ll}
1894: 1 & \textrm{if $n=1$} \\
1895: 0 & \textrm{if $n>1$}
1896: \end{array}\;, \nonumber
1897: \eea
1898: and we have the result
1899: \beq
1900: \tA^n_{1\textrm{-jet}}(s,t) = 0~~~~~~~ \textrm{if $n>1$}.
1901: \eeq
1902:
1903: \subsection*{$2$-Jet inclusive}
1904: The situation is similar to the previous one: the combinatoric
1905: factor $j(2n-j)$ is replaced by $j^2(2n-j)^2$. The separation in the
1906: three cases analyzed in section \ref{2jets} can formally be
1907: obtained by writing one of the factors $j$ as $(j-1)+1$ and one of the factors
1908: $2n-j$ as $(2n-j-1)+1$; the combinatorial factor $j^2 (2n-j)^2$ is therefore
1909: written as a sum of four terms:
1910: \beq
1911: j^2(2n-j)^2=j(j-1)(2n-j)(2n-j-1)+j(2n-j)(2n-j-1)+j(j-1)(2n-j)+j(2n-j)\,.
1912: \label{2jetdecomp}
1913: \eeq
1914: Let us also classify these terms counting the numbers ($n_l$,$n_r$)
1915: of reggeized gluons on the (left, right)
1916: side of the cut which emit at least one gluon (from the jet vertex).
1917: If $n=1$ clearly only the last term in eq. \eqref{2jetdecomp} is
1918: present (since $j=1$) and this correspond to the case $(n_l,n_r)=(1,1)$.
1919: For $n\ge 2$ all the terms are not trivial:
1920: \begin{itemize}
1921: \item $j(j-1)(2n-j)(2n-j-1)$, corresponding to the jets being connected to
1922: different reggeized gluons on both sides, $(n_l,n_r)=(2,2)$;
1923: \item $j(2n-j)(2n-j-1)$, corresponding to the jets being connected to
1924: different reggeized gluons on the right hand side but to the same gluon
1925: on the left hand side, $(n_l,n_r)=(1,2)$;
1926: \item $j(j-1)(2n-j)$, corresponding to the jets being connected to different
1927: reggeized gluons on the left hand side but to the same gluon on the right
1928: hand side, $(n_l,n_r)=(2,1)$;
1929: \item $j(2n-j)$, corresponding to the jets being connected to the same
1930: reggeized gluons on both sides, $(n_l,n_r)=(1,1)$.
1931: \end{itemize}
1932: It is easy to verify that after multiplication with the usual symmetry factor
1933: and summation over $j$, the last three term vanishes independently
1934: and the first one vanishes unless $n=2$, so that the only surving
1935: contributions appear when $n=(n_l+n_r)/2$.
1936:
1937: \subsection*{$m$-jet inclusive}
1938: We are now ready to generalize this result to the emission
1939: of $m$ jets. The contribution of the $2n$-reggeized gluon diagram is
1940: \beq
1941: \tA^n_{m\textrm{-jet}}(s,t) =
1942: 2 i \sum_{\textrm{B.S.}}\sum_{j=1}^{2n-1} S^n_j j^m (2n-j)^m~
1943: \int d\Omega_{2n}~s^{1+\tilde{\beta}}~ \cN_{2n}~ \cN_{2n}~
1944: \tilde{\gamma}_{\{\bk_l\}}.
1945: \label{discspQCDmjet}
1946: \eeq
1947: With the manipulations described in the previous subsection, we can
1948: reduce this expression to a sum of terms, each of which corresponds to
1949: a diagram in which a certain number $n_r$ of reggeized gluons on the right
1950: side of the cut emit at least one gluon, and $n_l$ reggeized gluons
1951: on the left hand side emit one or more gluons.
1952: Each term is of the form
1953:
1954: $$(-1)^n \sum_{j=n_l}^{2n-n_r} \frac{(-1)^j}{j!(2n-j)!}
1955: j(j-1)\ldots(j-n_l+1) \cdot
1956: (2n-j)(2n-j-1)\ldots(2n-j-n_r+1)$$
1957: \bea
1958: =& (-1)^n \sum_{n_l}^{2n-n_r} \frac{(-1)^j}{(j-n_l)!(2n-j-n_r)!}\nonumber \\
1959: =& (-1)^n \sum_{0}^{2n-n_r-n_l} \frac{(-1)^{j+n_l}}{j!(2n-j-n_r-n_l)!} \\
1960: =& \frac{(-1)^{n_l+n_r}}{(2n-n_l-n_r)!} \sum_{0}^{2n-n_r-n_l}
1961: \binom{2n-n_l-n_r}{j}(-1)^j \nonumber \\
1962: =& \Bigg\{ \begin{array}{ll}
1963: (-1)^{\frac{n_l-n_r}{2}} & \textrm{if $n=\frac{n_l+n_r}{2}$} \\
1964: 0 & \textrm{if $n>\frac{n_l+n_r}{2}$}
1965: \end{array}\;. \nonumber
1966: \eea
1967: Since $n_{l,r} \le m$, the inequality $(n_l+n_r)/2 \le m$ holds,
1968: and all diagrams with $n>m$ vanish. Among the others, the non-vanishing ones
1969: are those which satisfy the condition $n=(n_l+n_r)/2$. Briefly the
1970: condition is
1971: \beq
1972: 2n=n_l+n_r \, , \quad m>n
1973: \label{jetcondition}
1974: \eeq
1975: and it fixes also the position of the cut to $j=n_l$.
1976:
1977: The vanishing of \eqref{discspQCDmjet} for $n>m$ can also be seen in
1978: another way:
1979: \bea
1980: \sum_{j=1}^{2n-1} S^n_j j^m (2n-j)^m &=&
1981: \sum_{1}^{2n-1} \frac{(-1)^{n-j}}{j!(2n-j)!} j^m (2n-j)^m \nonumber \\
1982: &=& \frac{1}{(2n)!} \sum_{0}^{2n}
1983: \binom{2n}{j} (-1)^{n-j} j^m (2n-j)^m \nonumber \\
1984: &=& \frac{(-1)^n}{(2n)!}
1985: \frac{\partial^{2m}}{\partial \alpha^m \partial \beta^m} e^{2n\beta}
1986: \underbrace{\sum_{0}^{2n} \binom{2n}{j} (-1)^{2n-j} e^{(\alpha - \beta) j}}_{
1987: =(e^{\alpha-\beta}-1)^{2n}}|_{\alpha,\beta=0} \label{eq:espcomp} \nonumber \\
1988: &=& \frac{(-1)^n}{(2n)!}
1989: \frac{\partial^{2m}}{\partial \alpha^m \partial \beta^m}
1990: (e^{\alpha}-e^{\beta})^{2n}|_{\alpha,\beta=0} \\
1991: &=& \frac{(-1)^n}{(2n)!} \Big( x \frac{\partial}{\partial x} \Big)^m
1992: \Big( y \frac{\partial}{\partial y} \Big)^m (x-y)^{2n}|_{x,y=1} \nonumber
1993: \eea
1994: The operator $(x\partial_x)^m$ can be written as
1995: \beq
1996: \Big( x \frac{\partial}{\partial x} \Big)^m =
1997: \sum_{k=1}^{m} a_k^m x^k \frac{\partial^k}{\partial x^k},
1998: \label{eq:opdec}
1999: \eeq
2000: where the coefficient $a_k^m$ are positive integer numbers whose explicit
2001: expression is not needed. Using
2002: \eqref{eq:opdec} in \eqref{eq:espcomp} we obtain
2003: \bea
2004: \frac{(-1)^n}{(2n)!} \sum_{k,k'=1}^{m} a_k^m a_{k'}^m x^k y^{k'}
2005: \frac{\partial^k}{\partial x^k} \frac{\partial^{k'}}{\partial y^{k'}}
2006: (x-y)^{2n}|_{x,y=1} = \nonumber \\
2007: (-1)^n \sum_{k,k'=1}^{m} \delta_{k+k',2n} a_k^m a_{k'}^m (-1)^{k'}
2008: \eea
2009: which vanishes if $n>m$.
2010:
2011: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2012:
2013: \begin{thebibliography}{99}
2014:
2015: \bibitem{AGK}
2016: V.~A.~Abramovsky, V.~N.~Gribov and O.~V.~Kancheli,
2017: %``Character Of Inclusive Spectra And Fluctuations Produced In Inelastic
2018: %Processes By Multi - Pomeron Exchange,''
2019: Yad.\ Fiz.\ {\bf 18} (1973) 595
2020: [Sov.\ J.\ Nucl.\ Phys.\ {\bf 18} (1974) 308].
2021: %%CITATION = YAFIA,18,595;%%
2022:
2023: \bibitem{BFKL}
2024: L.~N.~Lipatov,
2025: %``Reggeization Of The Vector Meson And The Vacuum Singularity In Nonabelian
2026: %Gauge Theories,''
2027: Sov.\ J.\ Nucl.\ Phys.\ {\bf 23} (1976) 338
2028: [Yad.\ Fiz.\ {\bf 23} (1976) 642].
2029: %%CITATION = SJNCA,23,338;%%
2030: \\
2031: E.~A.~Kuraev, L.~N.~Lipatov and V.~S.~Fadin,
2032: %``The Pomeranchuk Singularity In Nonabelian Gauge Theories,''
2033: Sov.\ Phys.\ JETP {\bf 45} (1977) 199
2034: [Zh.\ Eksp.\ Teor.\ Fiz.\ {\bf 72} (1977) 377].
2035: %%CITATION = SPHJA,45,199;%%
2036: \\
2037: I.~I.~Balitsky and L.~N.~Lipatov,
2038: %``The Pomeranchuk Singularity In Quantum Chromodynamics,''
2039: Sov.\ J.\ Nucl.\ Phys.\ {\bf 28} (1978) 822
2040: [Yad.\ Fiz.\ {\bf 28} (1978) 1597].
2041: %%CITATION = SJNCA,28,822;%%
2042:
2043: \bibitem{SS}
2044: T.~Sjostrand and P.~Z.~Skands,
2045: %``Multiple interactions and the structure of beam remnants,''
2046: JHEP {\bf 0403} (2004) 053
2047: [arXiv:hep-ph/0402078].
2048: %%CITATION = HEP-PH 0402078;%%
2049: Eur.Phys.J.C39:129-154,2005
2050: e-Print Archive: hep-ph/0408302
2051:
2052: \bibitem{Braun}
2053: M.~A.~Braun,
2054: %``Inclusive jet production on the nucleus in the perturbative QCD with N(c)
2055: %$\to$ infinity,''
2056: Phys.\ Lett.\ B {\bf 483} (2000) 105
2057: [arXiv:hep-ph/0003003].
2058: %%CITATION = HEP-PH 0003003;%%
2059:
2060: \bibitem{Kovchegov}
2061: Y.~V.~Kovchegov and E.~Levin,
2062: %``Diffractive dissociation including multiple pomeron exchanges in high
2063: %parton density QCD,''
2064: Nucl.\ Phys.\ B {\bf 577} (2000) 221
2065: [arXiv:hep-ph/9911523].
2066: %%CITATION = HEP-PH 9911523;%%
2067:
2068: \bibitem{BR}
2069: J.~Bartels and M.~G.~Ryskin,
2070: %``The space-time picture of the wee partons and the AGK cutting rules in
2071: %perturbative QCD,''
2072: Z.\ Phys.\ C {\bf 76} (1997) 241
2073: [arXiv:hep-ph/9612226].
2074: %%CITATION = HEP-PH 9612226;%%
2075:
2076: \bibitem{BW}
2077: J.~Bartels and M.~Wusthoff,
2078: %``The Triple Regge limit of diffractive dissociation in deep inelastic
2079: %scattering,''
2080: Z.\ Phys.\ C {\bf 66} (1995) 157.
2081: %%CITATION = ZEPYA,C66,157;%%
2082:
2083: \bibitem{BE}
2084: J.~Bartels and C.~Ewerz,
2085: %``Unitarity corrections in high-energy QCD,''
2086: JHEP {\bf 9909} (1999) 026
2087: [arXiv:hep-ph/9908454].
2088: %%CITATION = HEP-PH 9908454;%%
2089:
2090: \bibitem{RT}
2091: R.~Ragazzon and D.~Treleani,
2092: %``Multiparton interactions and production of mini - jets in high-energy
2093: %hadronic collisions,''
2094: Phys.\ Rev.\ D {\bf 53} (1996) 55
2095: [arXiv:hep-ph/9508286].
2096: %%CITATION = HEP-PH 9508286;%%
2097:
2098: \bibitem{MN}
2099: A.~H.~Mueller and H.~Navelet,
2100: %``An Inclusive Minijet Cross-Section And The Bare Pomeron In QCD,''
2101: Nucl.\ Phys.\ B {\bf 282} (1987) 727.
2102: %%CITATION = NUPHA,B282,727;%%
2103:
2104: \bibitem{ABSW}
2105: H.~D.~I.~Abarbanel, J.~B.~Bronzan, R.~L.~Sugar and A.~R.~White,
2106: Phys.\ Rept.\ {\bf 21} (1975) 119.
2107: %%CITATION = PRPLC,21,119;%%
2108:
2109: \bibitem{GPT}
2110: V.~N.~Gribov, I.~Y.~Pomeranchuk and K.~A.~Ter-Martirosian,
2111: Phys.\ Rev.\ {\bf 139} (1965) B184.\\
2112: V.~N.~Gribov, I.~Y.~Pomeranchuk and K.~A.~Ter-Martirosian,
2113: Sov.\ J.\ Nucl.\ Phys.\ {\bf 2} (1965) 258
2114: [Yad.\ Fiz.\ {\bf 2} (1965) 361].
2115:
2116: \bibitem{CSS}
2117: J.~C.~Collins, D.~E.~Soper and G.~Sterman,
2118: %``Factorization Of Hard Processes In QCD,''
2119: Adv.\ Ser.\ Direct.\ High Energy Phys.\ {\bf 5} (1988) 1
2120: [arXiv:hep-ph/0409313].
2121: %%CITATION = HEP-PH 0409313;%%
2122:
2123: \bibitem{BKP}
2124: J.~Bartels,
2125: %``High-Energy Behavior In A Nonabelian Gauge Theory. 1. T (N--->M) In The
2126: %Leading Log Normal S Approximation,''
2127: Nucl.\ Phys.\ B {\bf 151} (1979) 293.
2128: %%CITATION = NUPHA,B151,293;%%
2129: \\
2130: J.~Bartels,
2131: %``High-Energy Behavior In A Nonabelian Gauge Theory. 2. First Corrections To
2132: %T(N--->M) Beyond The Leading Lns Approximation,''
2133: Nucl.\ Phys.\ B {\bf 175} (1980) 365.
2134: %%CITATION = NUPHA,B175,365;%%
2135: \\
2136: J.~Kwiecinski and M.~Praszalowicz,
2137: %``Three Gluon Integral Equation And Odd C Singlet Regge Singularities In
2138: %QCD,''
2139: Phys.\ Lett.\ B {\bf 94} (1980) 413.
2140: %%CITATION = PHLTA,B94,413;%%
2141:
2142: \bibitem{Jaroszewicz}
2143: T.~Jaroszewicz,
2144: %``Infrared Divergences And Regge Behavior In QCD,''
2145: Acta Phys.\ Polon.\ B {\bf 11} (1980) 965.
2146: %%CITATION = APPOA,B11,965;%%
2147:
2148: \bibitem{HCS}
2149: A.~H.~Mueller and W.~K.~Tang,
2150: %``High-energy parton-parton elastic scattering in QCD,''
2151: Phys.\ Lett.\ B {\bf 284} (1992) 123.
2152: %%CITATION = PHLTA,B284,123;%%
2153: \\
2154: B.~Cox, J.~Forshaw and L.~Lonnblad,
2155: %``Hard colour singlet exchange at the Tevatron,''
2156: JHEP {\bf 9910} (1999) 023
2157: [arXiv:hep-ph/9908464].
2158: %%CITATION = HEP-PH 9908464;%%
2159: \\
2160: R.~Enberg, G.~Ingelman and L.~Motyka,
2161: %``Hard colour singlet exchange and gaps between jets at the Tevatron,''
2162: Phys.\ Lett.\ B {\bf 524} (2002) 273
2163: [arXiv:hep-ph/0111090].
2164: %%CITATION = HEP-PH 0111090;%%
2165:
2166: \bibitem{Kovchegov1_2}
2167: Y.~V.~Kovchegov and K.~Tuchin,
2168: %``Inclusive gluon production in DIS at high parton density,''
2169: Phys.\ Rev.\ D {\bf 65} (2002) 074026
2170: [arXiv:hep-ph/0111362];\\
2171: %%CITATION = HEP-PH 0111362;%%
2172: J.~Jalilian-Marian and Y.~V.~Kovchegov,
2173: %``Inclusive two-gluon and valence quark-gluon production in DIS and p A,''
2174: Phys.\ Rev.\ D {\bf 70} (2004) 114017
2175: [arXiv:hep-ph/0405266].
2176: %%CITATION = HEP-PH 0405266;%%
2177:
2178: \bibitem{Braun2005}
2179: M.~A.~Braun,
2180: %``On the inclusive gluon jet production off the nucleus in the perturbative
2181: %QCD,''
2182: arXiv:hep-ph/0502184.
2183: %%CITATION = HEP-PH 0502184;%%
2184:
2185: \bibitem{GPD}
2186: M.~Diehl,
2187: %``Generalized parton distributions,''
2188: Phys.\ Rept.\ {\bf 388} (2003) 41
2189: [arXiv:hep-ph/0307382].
2190: %%CITATION = HEP-PH 0307382;%%
2191:
2192: \bibitem{DVCS}
2193: M.~Diehl, T.~Gousset, B.~Pire and J.~P.~Ralston,
2194: %``Testing the handbag contribution to exclusive virtual Compton scattering,''
2195: Phys.\ Lett.\ B {\bf 411} (1997) 193
2196: [arXiv:hep-ph/9706344].
2197: %%CITATION = HEP-PH 9706344;%%
2198: \\
2199: A.~V.~Belitsky, D.~Muller, L.~Niedermeier and A.~Schafer,
2200: %``Leading twist asymmetries in deeply virtual Compton scattering,''
2201: Nucl.\ Phys.\ B {\bf 593} (2001) 289
2202: [arXiv:hep-ph/0004059].
2203: %%CITATION = HEP-PH 0004059;%%
2204: \\
2205: A.~V.~Belitsky, D.~Muller and A.~Kirchner,
2206: %``Theory of deeply virtual Compton scattering on the nucleon,''
2207: Nucl.\ Phys.\ B {\bf 629} (2002) 323
2208: [arXiv:hep-ph/0112108].
2209: %%CITATION = HEP-PH 0112108;%%
2210:
2211: \bibitem{BLV}
2212: J.~Bartels, L.~N.~Lipatov and G.~P.~Vacca,
2213: %``A new odderon solution in perturbative QCD,''
2214: Phys.\ Lett.\ B {\bf 477} (2000) 178
2215: [arXiv:hep-ph/9912423].
2216: %%CITATION = HEP-PH 9912423;%%
2217:
2218: \end{thebibliography}
2219:
2220: \end{document}
2221: