1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2: \chapter{The Deep Inelastic Compton process in NLO QCD}\label{nlo}
3: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
4:
5: Our calculation bases on the standard leading twist perturbative QCD
6: description of hard hadronic processes.
7: This means that we consider such diagrams where
8: one active parton from each initial particle
9: is involved in the hard process. However, in some Monte Carlo
10: simulations performed
11: %\PP
12: by experimental groups recently (2004)
13: a large effect of multiple interactions was found
14: for the prompt photon production at HERA. This effect is
15: included in experimental analyses by introducing some models.
16: We present
17: %\PP
18: predictions based on our calculations
19: and ``corrected'' by an experimental group
20: for the multiple interactions as well as hadronization processes
21: in Chapters \ref{isol} and \ref{jet}.
22:
23: There are several calculations
24: for the photoproduction of prompt (isolated) photons at the HERA collider,
25: namely: the calculation of Gordon and Vogelsang (GV)
26: \cite{Gordon:1995km}, Gordon (LG) \cite{Gordon:1998yt},
27: Krawczyk and Zembrzuski (K\&Z)
28: \cite{Krawczyk:1998it,Krawczyk:2001tz,Zembrzuski:2003nu}
29: and Fontannaz, Guillet and Heinrich (FGH)
30: \cite{Fontannaz:2001ek,Fontannaz:2001nq}.
31: %,Fontannaz:2003yn,Heinrich:2003vx,Fontannaz:2004qv}.
32: All these calculations differ from one another by set of diagrams
33: included in next-to-leading order (NLO).
34: Below we discuss our choice of diagrams (Sec. \ref{Snlo:diagrams}),
35: as well as some calculation details (Secs. \ref{det}, \ref{epa}).
36:
37: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
38: \section{Contributing processes}\label{Snlo:diagrams}
39: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
40:
41: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
42: \subsection{Born process and $\mathcal{O}$$(\alpha_S)$ corrections,
43: resolved $\gamma$ {\underline{or}} fragmentation into $\gamma$}\label{CnloB}
44: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
45:
46: We wish to study the hard electron-proton scattering
47: leading to a production of a photon or a photon plus a jet
48: and anything else, $ep\ra e\gamma ~(jet) ~X$
49: (Figs. \ref{figcompton}-\ref{figcomptonborn}).
50: In this reaction the mediating photon arising from the electron
51: interacts with a partonic constituent (quark or gluon) inside the
52: proton \cite{Bjorken:1969ja}.
53: In the lowest order (Born) process, $\gamma q\ra\gamma q$,
54: the photon is scattered from a quark yielding the final photon and the quark,
55: see Fig.~\ref{figborn}.
56: This subprocess is of order $\mathcal{O}$$(\alpha^2)$ and has
57: a pure electromagnetic nature. To obtain the NLO QCD predictions,
58: the corrections of order $\mathcal{O}$$(\alpha_S)$ to the Born process
59: have to be taken into account. These corrections lead to partonic processes
60: of order $\mathcal{O}$$(\alpha^2\alpha_S)$. They
61: include the virtual gluon exchange (Fig. \ref{figvirt}), the real gluon
62: emission (Fig. \ref{figreal}) and the process $\gamma g\ra\gamma q\bar{q}$
63: (Fig. \ref{fig23})~\cite{Duke:1982bj,mkcorr,Aurenche:1984hc,jan}.
64:
65: \begin{figure}[b]
66: \vskip 3cm\relax\noindent\hskip 0cm
67: \relax{\special{psfile=fig.born.ps}}
68: \begin{picture}(0,0)
69: \put(300,40){+ crossed diagram}
70: \end{picture}
71: \vspace{0cm}
72: \caption{\small\sl The Born process (the Compton scattering on the quark).}
73: \label{figborn}
74: %\end{figure}
75: %\begin{figure}[ht]
76: \vskip 3.3cm\relax\noindent\hskip -0.2cm
77: \relax{\special{psfile=fig.virt.ps}}
78: \vspace{0cm}
79: \caption{\small\sl An example of the virtual gluon
80: corrections to the Born process.}
81: \label{figvirt}
82: %\end{figure}
83: %\begin{figure}[ht]
84: \vskip 3.3cm\relax\noindent\hskip -0.2cm
85: \relax{\special{psfile=fig.real2.ps}}
86: \vspace{0cm}
87: \caption{\small\sl An example of the real gluon
88: corrections to the Born process,
89: $\gamma q\ra\gamma qg$.}
90: \label{figreal}
91: %\end{figure}
92: %\begin{figure}[ht]
93: \vskip 3.5cm\relax\noindent\hskip 0.7cm
94: \relax{\special{psfile=fig.23.ps}}
95: \vspace{1.cm}
96: \caption{\small\sl An example diagram for
97: the process $\gamma g\ra\gamma q\bar{q}$.}
98: \label{fig23}
99: \end{figure}
100:
101: The corrections due to the virtual gluon
102: exchange and the real gluon emission contain {\sl infrared} singularities which
103: cancel when both contributions are added up properly in the scattering
104: amplitude squared (see Secs. \ref{Snon:x},
105: \ref{Sapprox}, \ref{pss}).
106: The contribution due to the processes $\gamma q\ra\gamma qg$
107: and $\gamma g\ra\gamma q\bar{q}$ contains another type of singularities,
108: so called {\sl mass} or {\sl collinear} singularities, which do not
109: cancel. In order to remove them from the cross section the
110: {\sl factorization} procedure is applied: the singularities are subtracted
111: and shifted into corresponding parton densities in the proton or
112: photon or into parton-to-photon fragmentation functions
113: \cite{Aurenche:1984hc,jan}.
114: At this stage the {\sl factorization} scale appears.
115: %As a consequence of the factorization procedure,
116: The {\sl bare} ({\sl scale invariant}) parton densities in the proton are
117: replaced by renormalized {\sl scale dependent} densities. Moreover,
118: the {\sl scale dependent} parton densities in the photon and
119: the {\sl scale dependent} fragmentation functions appear in the
120: calculation. They are necessary ingredients in the NLO calculation
121: to absorb the
122: %\PP
123: %unphysical
124: mass singularities.
125:
126: \begin{figure}[b]
127: \vskip 5cm\relax\noindent\hskip -2cm
128: \relax{\special{psfile=fact-phot.ps}}
129: \vspace{1.5cm}
130: \caption{\small\sl A sample of diagrams illustrating the factorization
131: of the collinear singularities from the process $\gamma g\ra\gamma q\bar{q}$
132: into the parton density in the photon, $f_{q/\gamma}$.}
133: \label{fact-phot}
134: %\end{figure}
135: %\begin{figure}[ht]
136: \vskip 4.5cm\relax\noindent\hskip -2cm
137: \relax{\special{psfile=fact-frag.ps}}
138: \vspace{0.7cm}
139: \caption{\small\sl A sample of diagrams illustrating the factorization
140: of the collinear singularities from the process $\gamma q\ra\gamma qg$
141: into the fragmentation function, $D_{\gamma/q}$.}
142: \label{fact-frag}
143: \end{figure}
144:
145: The factorization procedure is illustrated in Figs. \ref{fact-phot}
146: and \ref{fact-frag}: the singularities due to the collinear configurations
147: in the vertexes $\gamma\ra q\bar{q}$ and $q\ra\gamma q$ are shifted
148: and absorbed by the corresponding quark density in the photon ($f_{q/\gamma}$)
149: or by the quark-to-photon fragmentation function ($D_{\gamma/q}$), respectively
150: (for precise definitions and details see \cite{Aurenche:1984hc,jan}).
151:
152: It is worth mentioning
153: that there was a discussion whether
154: for the photoproduction of isolated photons in $e^+e^-$ collisions
155: the conventional factorization breaks down, and whether the cross section
156: is an infrared safe quantity
157: \cite{Berger:1995cc,Aurenche:1996ng,Catani:1998yh}.
158: In principle these questions could as well
159: occur for the photoproduction of isolated photons in $ep$ collisions.
160: However we do not deal with this problem
161: because it arises from $2\rightarrow 3$ processes
162: involving the parton-to-photon fragmentation, which are absent
163: in our calculation.
164: We checked this explicitly and found that all the singularities in
165: our calculations for the isolated photon production
166: are canceled or factorized, as in the case of non-isolated photon
167: production, and the cross sections are well defined
168: \cite{Krawczyk:1998it,Krawczyk:2001tz,Zembrzuski:2003nu} (see
169: also~\cite{Gordon:1995km,Fontannaz:2001ek,Gordon:1998yt,Fontannaz:2001nq},
170: \cite{Fontannaz:2003yn}-\cite{Gehrmann-DeRidder:2000ce}).
171:
172: The processes involving the parton densities in the {\sl resolved}
173: initial photon are shown in Fig.
174: \ref{figsingi}. In these processes a parton (quark or gluon) from the photon
175: interacts with a parton arising from the proton yielding the {\sl direct}
176: final photon and a parton. Fig. \ref{figsingf} shows the processes
177: with the fragmentation of a parton into the final photon. Here
178: the {\sl direct} initial photon interacts with a parton from the proton
179: leading to a production of two partons;
180: one of these partons produces the final photon
181: in the
182: %\PP
183: %long-distance
184: {\sl fragmentation} process. The name {\sl direct} photons
185: stand for the photons participating directly in the hard partonic
186: process.
187:
188: \begin{figure}[h]
189: \vskip 4cm\relax\noindent\hskip -2cm
190: \relax{\special{psfile=fig.res1.ps}}
191: \vskip 1cm\relax\noindent\hskip 4cm
192: \relax{\special{psfile=fig.res2.ps}}
193: \vskip 2cm\relax\noindent\hskip -2cm
194: \relax{\special{psfile=fig.res3.ps}}
195: \begin{picture}(0,0)
196: \put(340,10){+ crossed diagrams}
197: \end{picture}
198: \vspace{0cm}
199: \caption{\small\sl The processes with the resolved initial photon.}
200: \label{figsingi}
201: \end{figure}
202: \begin{figure}[ht]
203: \vskip 4cm\relax\noindent\hskip -2.5cm
204: \relax{\special{psfile=fig.frag1.ps}}
205: \vskip 1cm\relax\noindent\hskip 3.5cm
206: \relax{\special{psfile=fig.frag2.ps}}
207: \vskip 2cm\relax\noindent\hskip -2.5cm
208: \relax{\special{psfile=fig.frag3.ps}}
209: \begin{picture}(0,0)
210: \put(340,10){+ crossed diagrams}
211: \end{picture}
212: \vspace{0cm}
213: \caption{\small\sl The processes with the parton-to-photon
214: fragmentation.}
215: \label{figsingf}
216: \end{figure}
217:
218: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
219: \subsection{Counting of orders. Alternative set of diagrams}\label{Cnlo:disc}
220: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
221:
222: The hard partonic processes in Figs. \ref{figsingi} and \ref{figsingf}
223: are of order $\mathcal{O}$$(\alpha\alpha_S)$. They are convoluted
224: with the corresponding parton densities or fragmentation functions
225: which are proportional to the electromagnetic coupling constant $\alpha$.
226: These distributions contain the logarithmic dependence
227: on the factorization/renormalization scale $\mu$:
228: $f_{\gamma}\sim\alpha\ln\mu^2/\Lambda_{QCD}^2$ and
229: $D_{\gamma}\sim\alpha\ln\mu^2/\Lambda_{QCD}^2$.
230: The logarithm behaves like the inverse
231: of the strong coupling, $\ln\mu^2/\Lambda_{QCD}^2\sim 1/\alpha_S$ and
232: many authors (e.g.
233: \cite{Gordon:1995km,Fontannaz:2001ek,Gordon:1998yt,Fontannaz:2001nq,
234: Fontannaz:2003yn,Heinrich:2003vx,Fontannaz:2004qv})
235: treat $f_{\gamma}$ and $D_{\gamma}$ as being of order
236: $\mathcal{O}$$(\alpha/\alpha_S)$. This conclusion arises
237: from the {\sl evolution equation} and the
238: {\sl renormalization group equation}, as it shown in \cite{Vogt:1999mu}.
239: If so, then the processes shown in Figs. \ref{figsingi}, \ref{figsingf}
240: are of order $\mathcal{O}$$(\alpha\alpha_S\cdot\alpha/\alpha_S)$=
241: $\mathcal{O}$$(\alpha^2)$, i.e. identical with the order of the Born process
242: (Fig. \ref{figborn}).
243: Note however, that the logarithmic behavior of the parton densities
244: in the photon or the parton-to-photon fragmentation functions
245: arises from the pure electromagnetic coupling $\gamma\ra q\bar{q}$
246: and $q\ra\gamma q$, respectively \cite{Walsh:1973mz},
247: %\PP
248: as already discussed in Introduction.
249: From this point of view the considered distributions
250: should be rather treated as the quantities of order
251: $\mathcal{O}$$(\alpha)$ since the logarithm $\ln\mu^2/\Lambda_{QCD}^2$
252: coming from the electromagnetic interaction is not related to the
253: strong coupling constant,
254: see also \cite{Krawczyk:1990nq,Krawczyk:1998it,Krawczyk:2001tz}
255: and for more detailed discussion in \cite{Chyla:1999mw}.
256: Such a counting, $f_{\gamma}$ and $D_{\gamma}\sim$
257: $\mathcal{O}$$(\alpha)$, leads to an assignment
258: of the $\mathcal{O}$$(\alpha^2\alpha_S)$ order to the processes
259: shown in Figs. \ref{figsingi}, \ref{figsingf}.
260:
261: The different counting of orders in the strong coupling leads
262: to the different sets of processes included in our NLO calculation
263: for the prompt photon production at HERA
264: \cite{Krawczyk:1998it,Krawczyk:2001tz,Zembrzuski:2003nu}
265: in comparison with other NLO calculations.
266: They include in addition (beside the $\mathcal{O}$$(\alpha_S)$ corrections to
267: the Born cross section)
268: the $\mathcal{O}$$(\alpha_S)$ corrections to the processes
269: with the resolved initial photon \underline{or}
270: parton-to-photon fragmentation
271: \cite{Gordon:1995km,Fontannaz:2001ek,Gordon:1998yt,Fontannaz:2001nq},
272: and the $\mathcal{O}$$(\alpha_S)$ corrections to the processes
273: with both the resolved initial photon \underline{and}
274: parton-to-photon fragmentation
275: \cite{Gordon:1995km,Fontannaz:2001ek,Fontannaz:2001nq}~\footnote{All these
276: QCD corrections have been calculated (most of them twice) in
277: \cite{AurenchePLB140}-\cite{Gordon:1994wu}}.
278: These corrections are not included in our NLO calculation
279: since in our approach they should be taken into account
280: together with the next-to-next-to-leading (NNLO) order corrections to
281: the Born process.
282:
283: It should be emphasized that the $\mathcal{O}$$(\alpha_S)$ corrections
284: to the Born process contain (after the factorization procedure)
285: an explicit logarithmic dependence
286: on the scale $\mu$, $\ln s/\mu^2$. This dependence is
287: compensated by $\ln\mu^2/\Lambda_{QCD}^2$ from $f_{\gamma}$ and
288: $D_{\gamma}$ if the $\mathcal{O}$$(\alpha_S)$ corrections to the
289: Born process are
290: included in the cross section consistently with the corresponding
291: contributions involving parton densities in the photon or parton-to-photon
292: fragmentation functions. Similarly, the
293: $\mathcal{O}$$(\alpha_S)$ corrections to the processes with the
294: resolved initial photon or the fragmentation into the final photon
295: should be, from our point of view,
296: taken into account \underline{together} with the
297: $\mathcal{O}$$(\alpha_S^2)$ corrections to the Born process
298: in order to compensate the dependence on the $\mu^2$ in the hadronic
299: cross section.
300:
301: One can argue that, since the other authors
302: \cite{Gordon:1995km,Fontannaz:2001ek,Gordon:1998yt,Fontannaz:2001nq}
303: include more diagrams, their calculation
304: is more accurate. It is true that including more diagrams usually
305: improves the quality of QCD predictions
306: but it is not obvious that it is always the case
307: %\PP
308: (see also the discussion in Secs. \ref{results3} and \ref{results5}).
309:
310: It is worth mentioning that there are also other authors who
311: treat the processes involving photons in a similar way as we do.
312: The authors of \cite{Glover:1993xc}
313: (see also \cite{Glover:1992sf,Glover:1994th})
314: present the NLO calculation for the prompt photon production
315: in the $e^+e^-$ collision with the quark-to-photon fragmentation.
316: They include the fragmentation
317: function at the same order as the $\mathcal{O}$$(\alpha_S)$ corrections,
318: since the fragmentation is in a close relation with the corresponding
319: collinear configuration in $\mathcal{O}$$(\alpha_S)$ corrections
320: (compare Fig. \ref{fact-frag}). It allows to cancel
321: an explicit dependence on the scale $\mu^2$.
322: A similar NLO calculation for the prompt photon production
323: in the deep inelastic $ep$ events (with $Q^2> 10$ GeV$^2$)
324: at the HERA collider is presented in
325: \cite{Gehrmann-DeRidder:2000ce,Gehrmann-DeRidder:1999wy,
326: Gehrmann-DeRidder:1999yu}.
327: In this calculation the
328: diagrams shown in Figs. \ref{figborn}-\ref{fig23} and
329: \ref{figsingf} are included without the $\mathcal{O}$$(\alpha_S)$ corrections
330: to the processes involving the fragmentation
331: (the contributions due to the resolved initial photon are negligible
332: for large $Q^2$). The cancellation (to a large extent) of the
333: $\mu^2$ dependence is therein also stressed.
334:
335: Another kind of arguments is presented in \cite{Kunszt:1992ab}, where
336: the NLO calculation for the prompt photon production in the
337: $e^+e^-$ collision is investigated. The authors of \cite{Kunszt:1992ab}
338: claim that the
339: parton-to-photon fragmentation, despite being of order
340: $\mathcal{O}$$(\alpha/\alpha_S)$ for the non-isolated photon production,
341: should be counted as the quantity of order
342: $\mathcal{O}$$(\alpha)$ for the isolated
343: final photon, since the isolation itself is a correction
344: of order $\mathcal{O}$$(\alpha_S)$.
345:
346: Counting $f_{\gamma}$ and $D_{\gamma}$ as
347: being of order $\mathcal{O}$$(\alpha/\alpha_S)$ allows for a self-consistent
348: expansion of physical quantities (structure functions, cross sections)
349: in powers of $1/\ln\mu^2/\Lambda_{QCD}^2$ (with
350: $\alpha_S\sim 1/\ln\mu^2/\Lambda_{QCD}^2+...$) \cite{Vogt:1999mu}.
351: On the other hand, our approach,
352: where $f_{\gamma}$, $D_{\gamma}\sim$$\mathcal{O}$$(\alpha)$,
353: allows for a consistent counting of powers of $\alpha_S$
354: (since the logarithm $\ln\mu^2/\Lambda_{QCD}^2$ in $f_{\gamma}$ and
355: $D_{\gamma}$ it is not $1/\alpha_S$,
356: see also \cite{Krawczyk:1990nq,Chyla:1999mw}).
357:
358: The photon is a very special particle: being the point-like object
359: it sometimes exhibits hadronic-like ``structure''.
360: This double nature of the photon leads to more complicated description
361: within the QCD than for purely hadronic processes,
362: and further study is needed to clarify what is the
363: proper organization of the QCD perturbative
364: series for processes involving photons.
365: %\PP
366: This should include among others also the calculation of the
367: $\mathcal O$$(\alpha_s^2)$ corrections to the Born process.
368:
369: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
370: \subsection{Full set of diagrams included in analysis.
371: Resolved $\gamma$ {\underline{and}} fragmentation into $\gamma$,
372: box diagram}\label{Cnlo:other}
373: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
374: %\PP
375:
376: Besides the diagrams discussed in Sec. \ref{CnloB}, we include
377: in our numerical analysis also other diagrams,
378: %\PP these are
379: namely the diagrams with the
380: resolved $\gamma$ {\underline{and}} the
381: fragmentation into $\gamma$ and the box diagram.
382: In Figs. \ref{figdoub}, \ref{figdoub2} the processes with the resolved
383: initial photon \underline{and}
384: the fragmentation into the final photon are shown.
385: \begin{figure}[t]
386: \vskip 4cm\relax\noindent\hskip -3.5cm
387: \relax{\special{psfile=fig.doub1.ps}}
388: \vskip -0.5cm\relax\noindent\hskip 3cm
389: \relax{\special{psfile=fig.doub2.ps}}
390: \vskip 3.5cm\relax\noindent\hskip -3.5cm
391: \relax{\special{psfile=fig.doub3.ps}}
392: \vskip -0.5cm\relax\noindent\hskip 3cm
393: \relax{\special{psfile=fig.doub4.ps}}
394: \vskip 3.5cm\relax\noindent\hskip -3.5cm
395: \relax{\special{psfile=fig.doub5.ps}}
396: \vskip -1.cm\relax\noindent\hskip 3cm
397: \relax{\special{psfile=fig.doub6.ps}}
398: \vskip 4cm\relax\noindent\hskip 0.5cm
399: \relax{\special{psfile=fig.doub7.ps}}
400: \vskip 3.5cm\relax\noindent\hskip -3.5cm
401: \relax{\special{psfile=fig.doub8.ps}}
402: \vskip -0.5cm\relax\noindent\hskip 3cm
403: \relax{\special{psfile=fig.doub9.ps}}
404: \vspace{0cm}
405: \caption{\small\sl The processes involving the resolved initial photon
406: and the parton-to-photon fragmentation.}
407: \label{figdoub}
408: \end{figure}
409: \begin{figure}[t]
410: \vskip 4cm\relax\noindent\hskip -4.5cm
411: \relax{\special{psfile=fig.doub13.ps}}
412: \vskip -0.5cm\relax\noindent\hskip 0.4cm
413: \relax{\special{psfile=fig.doub14.ps}}
414: \vskip -0.5cm\relax\noindent\hskip 5.4cm
415: \relax{\special{psfile=fig.doub12.ps}}
416: \vskip 5cm\relax\noindent\hskip -1.cm
417: \relax{\special{psfile=fig.doub11.ps}}
418: \vskip -2cm\relax\noindent\hskip 5.4cm
419: \relax{\special{psfile=fig.doub10.ps}}
420: \vspace{1.5cm}
421: \caption{\small\sl Continuation of Fig.~\ref{figdoub}.}
422: \label{figdoub2}
423: \end{figure}
424: If we take $f_{\gamma}$, $D_{\gamma}\sim$$\mathcal{O}$$(\alpha)$,
425: all these processes are of order $\mathcal{O}$$(\alpha^2\alpha_S^2)$
426: and strictly speaking
427: go beyond the NLO accuracy of our calculations. Nevertheless
428: we include them, since they were taken into account in most of
429: existing NLO calculations for the (non-isolated or isolated)
430: prompt photon production at HERA and are found to be important
431: \cite{Duke:1982bj,Aurenche:1984hc},
432: \cite{Krawczyk:1990nq}-\cite{Aurenche:1992sb},
433: \cite{Gordon:1994sm}-\cite{Fontannaz:2003yn}.
434:
435: Finally, we take into account the photon-by-gluon scattering,
436: so called box process,
437: $\gamma g\ra\gamma g$, shown in Fig. \ref{figbox} \cite{Combridge:1980sx}.
438: \begin{figure}[h]
439: \vskip 5.3cm\relax\noindent\hskip 0.2cm
440: \relax{\special{psfile=fig.box.ps}}
441: \begin{picture}(0,0)
442: \put(290,65){+ crossed diagrams}
443: \end{picture}
444: \vspace{0cm}
445: \caption{\small\sl The box diagram.}
446: \label{figbox}
447: \end{figure}
448: The box process is also of order $\mathcal{O}$$(\alpha^2\alpha_S^2)$,
449: i.e. beyond the accuracy of our calculation.
450: However it is a very special process as it is
451: the lowest order contribution to the photon-by-gluon
452: scattering, it contains no
453: singularities, and no parton densities in the photon or fragmentation
454: functions are involved. We think that including the box process
455: does not introduce any additional ambiguities in the
456: summation of partonic cross sections. The box contribution
457: to the reaction $ep\ra e\gamma X$ or $\gamma p\ra\gamma X$ was
458: considered in previous calculations and it is known to be large
459: \cite{Fontannaz:1982et,Aurenche:1984hc},
460: \cite{Bawa:1988qs}-\cite{Aurenche:1992sb}, \cite{Gordon:1994sm}.
461: It is included in the calculations for the isolated photon
462: production at HERA
463: by authors K\&Z \cite{Krawczyk:1998it,Krawczyk:2001tz,Zembrzuski:2003nu}
464: and FGH \cite{Fontannaz:2001ek,Fontannaz:2001nq}
465: but it is omitted by GV \cite{Gordon:1995km}
466: and LG \cite{Gordon:1998yt}.
467:
468: To summarize, in our NLO calculation we take into account the
469: following contributions:
470: \begin{itemize}
471: \item the Born contribution (Fig. \ref{figborn}),
472: \item the $\mathcal{O}$$(\alpha_S)$ corrections to the Born diagram,
473: %\PP
474: including the $\gamma g\ra \gamma q \bar{q}$ process
475: (Figs. \ref{figvirt}-\ref{fig23}),
476: \item the processes with the resolved initial photon (Fig. \ref{figsingi})
477: \underline{or} with the fragmentation into the final
478: photon (Fig. \ref{figsingf}),
479: \item the processes with the resolved initial photon \underline{and}
480: the fragmentation into the final photon (Figs. \ref{figdoub}, \ref{figdoub2}),
481: \item the box process (Fig. \ref{figbox}).
482: \end{itemize}
483:
484: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
485: \section{Calculation details}\label{det}
486: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
487:
488: We perform NLO QCD calculations in the modified Minimal Subtraction
489: ($\overline{\rm MS}$) renormalization scheme \cite{Bardeen:1978yd}.
490: The fac\-tor\-ization/renor\-malization scales in parton densities and
491: fragmentation functions are assumed being equal to the
492: renormalization scale in the strong coupling constant and are denoted
493: as $\mu$. As a reference we take $\mu$ equal to the transverse
494: momentum~\footnote{By the transverse momentum, $p_T$,
495: we mean the component of the momentum perpendicular to
496: momenta of the initial particles.} (transverse energy) of the final photon,
497: $\mu=p_T^{\gamma}=E_T^{\gamma}$. For a comparison $\mu=E_T^{\gamma}/2$
498: and $\mu=2E_T^{\gamma}$ will be also considered.
499:
500: The quark masses are neglected in the calculation
501: and the number of active flavors
502: is assumed to be $N_f$=4 or, for a comparison, $N_f$=3 or 5.
503: The cross section is proportional to the electric
504: charge of the quark in the fourth power, so
505: the contribution of the bottom quark ($e_b$=-1/3) is expected
506: to be much smaller than the contributions of the up and charm quarks
507: ($e_u$=$e_c$=2/3), and the predictions for $N_f$=4 and 5 should not
508: differ considerably. On the other hand, the differences between
509: the results obtained using $N_f$=4 and 3 can be large.
510:
511: The two-loop coupling constant $\alpha_s$ is applied in the form
512: \be
513: \alpha_S(\mu^2)={{12 \pi}\over {(33-2N_f)\ln(\mu^2/\Lambda_{QCD}^2)}}
514: [1-{{6(153-19N_f)}\over (33-2N_f)^2} {{\ln[\ln(\mu^2/\Lambda_{QCD}^2)]}
515: \over{\ln(\mu^2/\Lambda_{QCD}^2)}}].
516: \label{alfas}
517: \ee
518: To obtain the QCD parameter $\Lambda_{QCD}$ appearing in the strong coupling
519: constant (\ref{alfas}), we use the world average of $\alpha_S$ at the scale
520: $M_Z$ (the mass of the $Z^0$ gauge boson) \cite{Bethke:2002rv}:
521: \be
522: \label{1183}
523: \overline{\alpha_S} (M_Z^2) = 0.1183 \pm 0.0027.
524: \ee
525: In order to
526: minimize theoretical and experimental uncertainties the above
527: $\overline{\alpha_S} (M_Z^2)$ value was determined in \cite{Bethke:2002rv}
528: from precise data based on NNLO analyses only;
529: the data given at scales different than $M_Z$ were
530: extrapolated to the $M_Z$ scale using the four-loop coupling.
531: Although we use the two-loop expression (\ref{alfas}),
532: we apply (\ref{1183})
533: as the best estimation of the true value of $\alpha_S (M_Z)$.
534: We take the number of active flavours $N_f$=3, 4 or 5
535: at scales $\mu < m_c$, $m_c<\mu < m_b$ and $\mu < m_b$, respectively,
536: with the following charm, bottom and $Z$ masses: $m_c=1.5$ GeV, $m_b=4.7$ GeV,
537: $M_Z=91.2$ GeV \cite{Bethke:2002rv}.
538: Assuming that $\alpha_S (\mu^2)$ is a continuous function
539: at $\mu=m_c$ and $\mu=m_b$ the obtained $\Lambda_{QCD}$ parameters are:
540: $\Lambda_{QCD}$=0.386, 0.332 and 0.230 GeV
541: for $N_f$=3, 4 and 5, respectively.
542: %The corresponding LO parameter for $N_f$=4 is $\Lambda_{QCD}$=0.123 GeV.
543: The above $\Lambda_{QCD}$ values are used in numerical calculations
544: discussed in the next chapters (Chapters~\ref{non},
545: \ref{isol}, \ref{jet})~\footnote{Note
546: that in Chapters~\ref{small} and \ref{Sglu}
547: %slightly
548: different numbers are used for
549: consistency with the published results.}.
550:
551: In the calculations we apply the Gl\"uck-Reya-Vogt (GRV)
552: parton densities in the proton \cite{Gluck:1995uf}
553: and photon \cite{Gluck:1992ee}, and the GRV
554: fragmentation functions \cite{Gluck:1993zx}. For a comparison
555: we also use other parametrizations, namely
556: Martin-Ro\-berts-Stir\-ling\--Thorne (MRST98) \cite{Martin:1998sq},
557: (MRST99) \cite{Martin:1999ww}, (MRST2002) \cite{Martin:2002aw},
558: CTEQ4M \cite{Lai:1996mg}, CTEQ6M \cite{Pumplin:2002vw},
559: Au\-ren\-che\--Chiap\-pet\-ta\--Fon\-tan\-naz-Guillet-Pilon
560: (ACFGP) \cite{Aurenche:1992sb},
561: Aurenche-Guillet-Fon\-tan\-naz
562: (AFG) \cite{Aurenche:1994in} and (AFG02) \cite{Fontannaz:2002nu},
563: Gordon-Storrow (GS) \cite{Gordon:1997pm},
564: Cornet-Jankowski-Kraw\-czyk\--Lor\-ca (CJ\-KL) \cite{Cornet:2002iy},
565: Duke-Owens (DO) \cite{Duke:1982bj},
566: Bourhis-Fontannaz-Guillet (BFG) \cite{Bourhis:1997yu} and
567: Gl\"uck-Reya-Stratmann (GRS) \cite{Gluck:1994tv}.
568:
569: Following experimental analysis \cite{Breitweg:1997pa}-\cite{unknown:2004uv},
570: we consider the photoproduction of the photon at HERA with transverse
571: momentum (transverse energy) higher than 5 GeV. In such processes
572: the emission of the final large-$p_T$ photon directly from the electron
573: (Bethe-Heitler process) is negligible \cite{ula} and we omit
574: it~\footnote{For transverse energy higher than 5 GeV the momentum transfer,
575: -$t$ (see Eq. (\ref{mand})), is higher than 25 GeV$^2$.
576: Note that for $Q^2_{max}$ = 1 GeV$^2$ and lower values of
577: the momentum transfer, 4$\le -t \le 10$ GeV$^2$,
578: the contribution of the Bethe-Heitler process may be no-negligible
579: in some kinematic regions, see Fig. 6 in Ref. \cite{Hoyer:2000mb}.}
580: (this process is neglected in all existing calculations for the
581: photoproduction of isolated photons at HERA
582: \cite{Gordon:1995km}-\cite{Fontannaz:2003yn}).
583:
584: The results presented in next sections are obtained in NLO QCD
585: with use of the GRV set of parametrizations
586: with $\mu =E_T^{\gamma}$, $N_f=4$ and
587: $\Lambda_{QCD}$=0.332 GeV
588: unless stated otherwise.
589:
590: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
591: \section{Equivalent photon approximation}\label{Cnlo:epa}
592: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
593:
594: Our aim is to consider the production of photons with large transverse momentum
595: in the electron-proton scattering in processes in which the electron
596: is scattered at a small angle. In such events the mediating photon is almost
597: on-shell, $Q^2\approx 0$ (photoproduction events), and
598: the cross section can be calculated using the
599: equivalent photon (Weizs\"{a}cker-Williams)
600: approximation~\cite{vonWeizsacker:1934sx,Williams:1934ad}
601: (see also e.g.~\cite{Budnev:1974de,Frixione:1993yw,Nisius:2000cv}).
602: In this approximation the differential cross section for the $ep$ collision
603: is related
604: to the corresponding differential cross section for the $\gamma p$ collision;
605: in case of the considered herein reaction, $ep\ra e\gamma X$
606: (or $ep\ra e\gamma ~jet ~X$),
607: we get (see Fig.~\ref{figepa.slide}):
608: \be\label{epa}
609: d\sigma^{ep\ra e\gamma (jet) X} = \int G_{\gamma /e}(y)
610: d\sigma^{\gamma p\ra \gamma (jet) X} dy ~,
611: \ee
612: \begin{figure}[b]
613: \vskip 4.5cm\relax\noindent\hskip 0cm
614: \relax{\special{psfile=epa2.slide.ps}}
615: \vspace{0cm}
616: \caption{\small\sl The factorization of the $ep\ra e\gamma X$ reaction.}
617: \label{figepa.slide}
618: \end{figure}
619: where $y$ is the fraction of the initial electron momentum carried
620: by the exchanged photon,
621: and $G_{\gamma /e}$ stands for the flux of the real photons emitted from
622: the electron.
623: \\
624: We use the photon spectrum in the form \cite{Frixione:1993yw}:
625: \be
626: G_{\gamma/e}(y)={\alpha\over 2\pi} \{ {1+(1-y)^2\over y}
627: \ln [ {Q^2_{max}(1-y)\over m_e^2 y^2}]
628: - ~ {2\over y}(1-y-{m_e^2y^2\over Q^2_{max}}) \} ~,
629: \ee
630: with $m_e$ being the electron mass. In the numerical calculations
631: the maximal photon virtuality $Q^2_{max}$ = 1 GeV$^2$ is assumed,
632: what is a typical value for the
633: recent photoproduction measurements at the HERA
634: collider~\cite{Breitweg:1997pa}-\cite{unknown:2004uv}.
635: The above formula describes the spectrum of equivalent real
636: (transversally polarized) photons. We do not take into account
637: longitudinally polarized
638: photons and the interference between longitudinally and transversally
639: polarized photons, since they give a very small contribution
640: %\PP
641: in the kinematic regions which we study~\cite{ula,Jezuita-Dabrowska:bp}.
642:
643: