1: \documentclass[twocolumn,superscriptaddress,showpacs,prl]{revtex4}
2: \usepackage{graphicx,epsfig,psfrag}
3:
4: \newcommand{\be}{\begin{equation}}
5: \newcommand{\ee}{\end{equation}}
6: \newcommand{\al}{\alpha}
7: \newcommand{\bt}{\beta}
8: \newcommand{\lm}{\lambda}
9: \newcommand{\bea}{\begin{eqnarray}}
10: \newcommand{\eea}{\end{eqnarray}}
11: \newcommand{\gm}{\gamma}
12: \newcommand{\Gm}{\Gamma}
13: \newcommand{\dl}{\delta}
14: \newcommand{\Dl}{\Delta}
15: \newcommand{\ep}{\epsilon}
16: \newcommand{\kp}{\kappa}
17: \newcommand{\Lm}{\Lambda}
18: \newcommand{\om}{\omega}
19: \newcommand{\pa}{\partial}
20: \newcommand{\nn}{\nonumber}
21: \newcommand{\dd}{\mbox{d}}
22:
23: \begin{document}
24:
25: \noindent
26: TTP04-09\hfill
27: \title{
28: \boldmath
29: Two-Loop Sudakov Form Factor in a Theory with Mass Gap
30: \unboldmath}
31: \author{Bernd Feucht}
32: \affiliation{Bernd Jantzen in later publications}
33: \affiliation{Institut f\"ur Theoretische Teilchenphysik,
34: Universit\"at Karlsruhe, 76128 Karlsruhe, Germany}
35: \author{Johann H. K\"uhn}
36: \affiliation{Institut f\"ur Theoretische Teilchenphysik,
37: Universit\"at Karlsruhe, 76128 Karlsruhe, Germany}
38: \affiliation{Kavli Institute for Theoretical Physics,
39: University of California, Santa Barbara, CA 93106 USA}
40: \author{Alexander A. Penin}
41: \affiliation{Institut f\"ur Theoretische Teilchenphysik,
42: Universit\"at Karlsruhe, 76128 Karlsruhe, Germany}
43: \affiliation{Institute for Nuclear Research,
44: Russian Academy of Sciences, 117312 Moscow, Russia}
45: \author{Vladimir A. Smirnov}
46: \affiliation{Institute for Nuclear Physics,
47: Moscow State University, 119992 Moscow, Russia}
48: \affiliation{II. Institut f{\"u}r Theoretische Physik,
49: Universit{\"a}t Hamburg, 22761 Hamburg, Germany}
50:
51: %\date{}
52:
53: \begin{abstract}
54: The two-loop Sudakov form factor is computed in a $U(1)$ model with a massive gauge
55: boson and a $U(1)\times U(1)$ model with mass gap. We analyze the result
56: in the context of hard and infrared evolution equations and establish
57: a matching procedure which relates the theories with and without
58: mass gap setting the stage for the complete calculation of the dominant
59: two-loop corrections to electroweak processes at high energy.
60: \end{abstract}
61: \pacs{12.15.Lk, 13.40.Ks, 12.38.Bx, 12.38.Cy}
62:
63: \maketitle
64:
65: Since the pioneering works by Sudakov \cite{Sud} and Jackiw \cite{Jac}
66: the high energy asymptotics of the electromagnetic form factor
67: has been the subject of numerous investigations. The problem is relevant
68: for a wide class of phenomenological applications from Drell-Yan
69: processes to deep inelastic scattering. Recently
70: a new wave of interest to the Sudakov asymptotic regime
71: has been risen in connection with higher-order corrections
72: to electroweak processes at high energies
73: \cite{Kur,Bec1,Bec2,Fad,KPS,CCC,KMPS,HKK,FKM}.
74: Experimental and theoretical studies of electroweak interactions have
75: traditionally explored the range from very low energies, e.g. through
76: parity violation in atoms, up to energies comparable to the masses of the
77: $W$- and $Z$-bosons, e.g. at the LEP or the Tevatron.
78: The advent of multi-TeV colliders like the LHC during the present decade
79: or a future linear electron-positron
80: collider will give access to a completely new
81: energy domain. Once the characteristic energies $\sqrt{s}$ are far larger than the masses of the
82: $W$- and $Z$-bosons, $M_{W,Z}$, exclusive
83: reactions like electron-positron (or quark-antiquark) annihilation into a
84: pair of fermions or gauge bosons will receive virtual corrections
85: enhanced by powers of the large
86: {\it electroweak} logarithm $\ln\bigl({s/ M_{W,Z}^2}\bigr)$.
87: The leading double-logarithmic corrections
88: may well amount to ten or even twenty percent in one-loop
89: approximation and reach a few percent in two-loop approximation.
90: Moreover,
91: in the TeV region, the subleading logarithms turn out to be equally
92: important \cite{KPS,KMPS} and a percent accuracy of the theoretical estimates
93: for the cross sections necessary for the search of new physics
94: beyond the standard model can be guaranteed only by including {\it all} the
95: logarithmic two-loop corrections.
96:
97: The calculation of the two-loop electroweak corrections even in the high
98: energy limit is an extremely challenging theoretical problem.
99: It is complicated in particular by
100: the presence of the mass gap and mixing in the
101: gauge sector. However, the logarithmic corrections are quite
102: insensitive to fine details of the spontaneous symmetry breaking.
103: The calculation of the
104: leading logarithmic (LL) electroweak corrections can be performed
105: using the fields of the unbroken symmetry phase and how the infrared singular
106: virtual photon contribution can be separated within the infrared
107: evolution equation approach ~\cite{Fad}. This scheme has been extended
108: to the next-to-leading (NLL) and next-to-next-to-leading
109: logarithmic (N$^2$LL) approximation in Refs.~\cite{KPS,KMPS}.
110:
111: In the study of Sudakov corrections the analysis of the
112: form factor plays a special role since it is the simplest
113: quantity which includes the
114: complete information about the universal {\it collinear} logarithms
115: \cite{Fre} directly applicable to a process with an arbitrary
116: number of fermions.
117: In this Letter we formulate a general matching procedure
118: which relates the logarithmic corrections in the
119: theories with and without mass gap by combining the hard and infrared
120: evolution equation approach with the explicit two-loop results for
121: the form factor in an Abelian gauge model. This
122: reduces the calculation of the dominant two-loop corrections
123: to electroweak processes at high energy to a
124: single-mass problem without mixing.
125:
126: The structure of the Letter is as follows. First,
127: we present the explicit two-loop results for the form factor
128: in a $U(1)$ model with a massive gauge boson. Then we introduce
129: the evolution equations, compute the
130: two-loop corrections to the form factor in a $U(1)\times U(1)$ model
131: with mass gap, and establish the matching procedure.
132: Finally we outline how the approach can be applied to
133: the calculation of the two-loop electroweak corrections to
134: neutral current four-fermion processes.
135:
136: The vector form factor ${\cal F}$ determines the
137: fermion scattering amplitude in an external Abelian field.
138: It is a function of the Euclidean momentum transfer
139: $Q^2=-(p_1-p_2)^2$ where $p_{1,2}$ is the incoming/outgoing fermion
140: momentum and we consider on-shell massless fermions, $p_1^2=p_2^2=0$.
141: Let us write the perturbative expansion for the form factor
142: as ${\cal F}_\al(M,Q)=\sum_n\left(\al\over 4\pi\right)^nf^{(n)}{\cal F}_B$
143: where ${\cal F}_B$ corresponds to the Born approximation, $f^{(0)}=1$.
144: For the $U(1)$ model with a gauge boson of mass $M$
145: in the Sudakov limit $M/Q\to 0$ the one-loop correction is well known
146: \be
147: f^{(1)}=-{\cal L}^2
148: +3{\cal L}
149: -{7\over 2}-{2\over 3}\pi^2\,,
150: \label{1loopf}
151: \ee
152: where ${\cal L}=\ln\left({Q^2/M^2}\right)$ and all the power-suppressed
153: terms are neglected. For the two-loop term we find
154: \bea
155: f^{(2)}&=&{1\over 2}{\cal L}^4-{3}{\cal L}^3
156: +\left(8+{2\over 3}\pi^2\right){\cal L}^2
157: -\big(9+4\pi^2
158: \nn\\
159: &&
160: -24\zeta(3)\big){\cal L}
161: +{25\over 2}+{52\over 3}\pi^2
162: +80\zeta(3)-{52\over 15}\pi^4
163: \nn\\
164: &&
165: -{32\over 3}\pi^2\ln^22+{32\over 3}\ln^42
166: +256\,{\rm Li}_4\left({1\over 2}\right)\,,
167: \label{2loopf}
168: \eea
169: where
170: $\zeta(3)=1.202057\ldots$ and
171: ${\rm Li}_4\left({1\over 2}\right)=0.517479\ldots$
172: are the values of the Riemann's $\zeta$-function
173: and the polylogarithm, respectively. In Eq.~(\ref{2loopf}) we
174: do not include the contribution
175: due to the virtual fermion loop computed in \cite{FKM}.
176: For the calculation of the leading power behavior of
177: the two-loop on-shell vertex diagrams
178: with two massive propagators in the Sudakov limit
179: we used the expansion by regions approach
180: \cite{BenSmi} (for the application to the Sudakov form factor see
181: also \cite{KPS}). The method is based on the factorization of the
182: contributions of the dynamical modes characteristic for the
183: Sudakov limit \cite{Ste} in dimensional regularization.
184: Our result for the contribution of the hard modes agrees with
185: the dimensionally regularized massless result of Ref.~\cite{KraLam}.
186: Details of our calculation will be published elsewhere.
187:
188: \begin{figure}
189: \psfrag{Q [GeV]}{$Q$ [GeV]}
190: %\epsfig{figure=F2R-logs.eps,height=6cm}
191: \epsfig{figure=F2R-logs-color.eps,height=6cm}
192: \caption{\label{fig1} The two-loop correction to the form factor ${\cal F}_\al(M,Q)$
193: in LL (including $\al^2{\cal L}^4$), NLL (including $\al^2{\cal L}^3$),
194: N$^2$LL (including $\al^2{\cal L}^2$), N$^3$LL (including $\al^2{\cal L}^1$)
195: approximations and the complete two-loop correction
196: as functions of the momentum transfer for $M=80$~GeV, $\al/(4\pi)=3\cdot 10^{-3}$.
197: }
198: \end{figure}
199:
200: In Fig.~\ref{fig1} the numerical results for the two-loop correction
201: to the form factor in the different logarithmic approximations
202: are plotted as functions of the momentum for
203: the values of $M$ and $\al$ typical for electroweak interactions.
204: The two-loop logarithmic terms have a sign-alternating
205: structure resulting in significant cancellations.
206: In the region of a few TeV the form factor does not reach the
207: double-logarithmic asymptotics. The quartic, cubic and quadratic
208: logarithms are comparable in magnitude and dominate the two-loop corrections.
209: Then the logarithmic expansion starts
210: to converge and, after including the linear-logarithmic contribution,
211: provides a very accurate approximation of the total two-loop correction.
212: Such a behavior is typical for the Sudakov limit and holds
213: for the non-Abelian corrections as well \cite{KMPS,FKM}.
214: Note that by rescaling $M\to e^{3/4}M$ in the argument of the logarithm
215: the NLL contribution can be made to disappear. That improves
216: significantly the
217: convergence of the logarithmic expansion and prevents the strong
218: cancellation between the logarithmic terms (see Fig.~\ref{fig2}).
219: Still, the N$^3$LL contribution
220: is a must for the quantitative approximation.
221:
222: \begin{figure}
223: \psfrag{Q [GeV]}{$Q$ [GeV]}
224: %\epsfig{figure=F2Rxi-logs.eps,height=6cm}
225: \epsfig{figure=F2Rxi-logs-color.eps,height=6cm}
226: \caption{\label{fig2} The same as Fig.~\ref{fig1}
227: after changing the argument of the logarithm.
228: }
229: \end{figure}
230:
231: The asymptotic dependence of the form factor on $Q$ is governed by the
232: linear {\it hard} evolution equation \cite{Mue}.
233: As a consequence, the logarithmic corrections exponentiate.
234: For the purely Abelian contribution the exponent has a
235: particularly simple form
236: \bea
237: {\cal F}_\al(M,Q)&=&\exp\Big\{{\al\over 4\pi}\Big[-{\cal L}^2
238: +\Big(3-{\al\over 4\pi}\Big(-{3\over 2}+2\pi^2
239: %\right.\right.\right.\right.
240: \nn\\
241: &&
242: %\left.\left.\left.\left.
243: -24\zeta(3)\Big)+{\cal O}(\al^2)\Big){\cal L}\Big]\Big\}{\cal F}_\al(M,M)\,.
244: \label{expf}
245: \eea
246: The double-logarithmic term
247: in the exponent is protected against the Abelian multiloop corrections
248: by the properties of the light-cone Wilson loop \cite{KorRad}.
249: Our two-loop result determines the next-to-next-to-next-to-next-to-leading
250: logarithmic (N$^4$LL) approximation of
251: the form factor which includes the $\al^n{\cal L}^{m}$ corrections
252: with $m=2n-4,\ldots,2n$ to all orders in $\al$.
253:
254: Let us now turn to the second example with two
255: Abelian gauge bosons of the masses $\lm$ and $M$, $\lm\ll M$,
256: and couplings $\al'$ and $\al$, respectively.
257: We can introduce the {\it infrared} evolution equation which
258: governs the dependence of the form factor ${\cal F}(\lm,M,Q)$
259: on $\lm$ \cite{Fad}. The virtual corrections
260: become divergent in the limit $\lm\to 0$.
261: According to the Kinoshita-Lee-Nauenberg theorem \cite{KLN}, these divergences are
262: cancelled against the ones of the corrections due to
263: the emission of real light gauge bosons of vanishing energy and/or
264: collinear to one of the on-shell fermion lines.
265: The singular behavior of the form factor
266: must be the same in the full
267: $U_{\al'}(1)\times U_{\al}(1)$ theory and the
268: effective $U_{\al'}(1)$ model with only the
269: light gauge boson. Thus for $\lm\ll M\ll Q$
270: the solution of the infrared evolution equation
271: is given by the exponent of Eq.~(\ref{expf}) with $M$, $\al$ replaced by
272: $\lm$, $\al'$, and the form factor can be written in a factorized form
273: \be
274: {\cal F}(\lm,M,Q)=
275: \tilde{F}(M,Q){\cal F}_{\al'}(\lm,Q)+{\cal O}(\lm/M)\,,
276: \label{fac}
277: \ee
278: where the function $\tilde{F}(M,Q)$ depends on $\al$ and $\al'$
279: and incorporates all the
280: logarithms of the form $\ln\left({Q^2/M^2}\right)$.
281: It can be obtained directly by calculating the ratio
282: \be
283: \tilde{F}(M,Q)=\left[{{\cal F}(\lm,M,Q)\over
284: {\cal F}_{\al'}(\lm,Q)}\right]_{\lm\to 0} \,.
285: \label{tf}
286: \ee
287: Since the function $\tilde{F}(M,Q)$ does not depend on the
288: infrared regularization, we compute the ratio in Eq.~(\ref{tf})
289: with $\lm=0$ using dimensional regularization
290: for the infrared divergences. The method of calculation of the
291: two-loop diagrams with both massive and massless gauge bosons
292: is similar to the purely massive case. We obtain the two-parameter
293: perturbative expansion
294: $\tilde{F}(M,Q)=\sum_{n,m}{\al'{}^n\al^m\over (4\pi)^{n+m}}\tilde
295: f^{(n,m)}$,
296: where $\tilde f^{(0,0)}=1$, $\tilde f^{(n,0)}=0$,
297: $\tilde f^{(0,m)}=f^{(m)}$, and the two-loop interference term reads
298: \bea
299: \tilde f^{(1,1)}&=&
300: \big(3-4\pi^2+48\zeta(3)\big){\cal L}
301: -2+{20\over 3}\pi^2
302: \nn\\
303: &&
304: -84\zeta(3)+{7\over 45}\pi^4\,.
305: \label{2looptf}
306: \eea
307: The numerical structure of the corrections to $\tilde{F}(M,Q)$
308: is very similar to the one of ${\cal F}_{\al}(M,Q)$ (see Fig.~\ref{fig1}).
309:
310: In the equal mass case, $\lm=M$,
311: we have an additional reparameterization symmetry, and
312: the form factor is determined by Eq.~(\ref{expf})
313: with the effective coupling $\bar\al= \al'+\al$ so that
314: ${\cal F}(M,M,Q)={\cal F}_{\bar\al}(M,Q)$.
315: We can now write down the matching relation
316: \be
317: {\cal F}(M,M,Q)=
318: C(M,Q)\tilde{F}(M,Q)
319: {\cal F}_{\al'}(M,Q)\,,
320: \label{matf}
321: \ee
322: where the matching coefficient $C(M,Q)$ represents the effect of
323: the power-suppressed terms neglected in Eq.~(\ref{fac}).
324: By combining the explicit results for ${\cal F}_{\al'}(M,Q)$ and
325: $\tilde{F}(M,Q)$ we find the two-loop matching coefficient
326: \bea
327: C(M,Q)&=&1+{\al'\al\over(4\pi)^2}
328: \left[{59\over 4}+{70\over 3}\pi^2+244\zeta(3)-{113\over 15}\pi^4\right.
329: \nn\\
330: &&
331: \left.
332: -{64\over 3}\pi^2\ln^22
333: +{64\over 3}\ln^42+512\,{\rm Li}_4\left({1\over 2}
334: \right)\right]\!.
335: \label{2loopc}
336: \eea
337: Eq.~(\ref{2loopc}) does not contain logarithmic terms,
338: and up to the N$^3$LL accuracy the product $\tilde{F}(M,Q){\cal F}_{\al'}(\lm,Q)$
339: continuously approaches ${\cal F}(M,M,Q)$ as $\lm$ goes to
340: $M$. Therefore, to get {\it all} the
341: logarithms of the heavy gauge boson mass
342: in two-loop approximation for the theory with mass gap,
343: it is sufficient to divide the form factor
344: ${\cal F}_{\bar\al}(M,Q)$ of the symmetric phase
345: by the form factor ${\cal F}_{\al'}(\lm,Q)$ of the effective
346: $U_{\al'}(1)$ theory taken at the symmetric point $\lm=M$.
347: Thus we have reduced the calculation in the theory with
348: mass gap to the one in the symmetric theory with a single mass parameter.
349: Note that the absence of the linear-logarithmic term in Eq.~(\ref{2loopc})
350: is an exceptional feature of the Abelian corrections.
351: The general analysis of the evolution equation \cite{KMPS}
352: shows the terms neglected in Eq.~(\ref{fac}) to contribute
353: starting from the N$^3$LL approximation.
354: This implies the absence of the second and higher powers
355: of the logarithm in the matching coefficient of Eq.~(\ref{matf}),
356: irrespectively of the gauge group and the mass generation mechanism.
357: Moreover, in the approximately equal mass case, $(M-\lm)/M\equiv\delta\ll 1$,
358: one can compute the form factor as an expansion around the equal mass result.
359: Up to N$^2$LL accuracy only the leading term of Eq.~(\ref{fac})
360: contributes and the expansion takes the form
361: \bea
362: \left.{\cal F}(\lm,M,Q)\right|_{\lm\to M}&=&
363: \left[1-\delta{\al'\over \pi}\left({\cal L}-{3\over 2}\right)
364: +{\cal O}(\delta^2)\right]
365: \nn\\
366: &&\times{\cal F}_{\tilde\al}(M,Q)+{\cal O}(\delta\al'\al{\cal L})\,.
367: \label{delexp}
368: \eea
369: Let us show how the above procedure applies to the
370: calculation of two-loop electroweak corrections.
371: To be specific, we consider a four-fermion neutral
372: current process, which is of primary phenomenological
373: importance, with light fermions. The four-fermion amplitude can be decomposed into
374: (the square of) the form factor and a {\it reduced} amplitude \cite{KPS,KMPS}.
375: The latter carries all the Lorentz and isospin indices
376: and does not contain collinear logarithms in perturbative expansion.
377: The logarithmic corrections to the reduced amplitude
378: are obtained by solving a renormalization group like equation \cite{Sen}.
379: The corresponding two-loop anomalous dimensions can be extracted
380: from the existing massless QCD calculations \cite{AGOT}
381: (see \cite{KMPS,SteTej}). Thus, the problem of the calculation of the
382: two-loop electroweak logarithms in the four-fermion
383: processes reduces to the analysis of the form factor.
384:
385: In Ref.~\cite{KMPS} by analyzing the
386: hard evolution equation
387: %within the expansion by regions approach
388: it has been found that the two-loop electroweak corrections up to
389: the next-to-next-to-leading (quadratic) logarithms are not sensitive to
390: the structure of the theory at the electroweak symmetry
391: breaking scale. The prediction of Ref.~\cite{KPS,KMPS} for the two-loop
392: logarithmic corrections fully agrees with the available explicit results
393: for the light fermion contribution \cite{FKM} and the Abelian contribution
394: obtained in this Letter. The only trace of the
395: Higgs mechanism of the gauge boson mass generation in N$^2$LL
396: approximation is the $Z-W$ boson mass splitting which can be
397: systematically taken into account within an expansion
398: around the equal mass approximation similar to Eq.~(\ref{delexp}).
399: Thus, the calculation of the two-loop electroweak corrections up to
400: the quadratic logarithms can be performed in two steps
401: outlined above:
402: (i) the corrections are evaluated using the fields of unbroken
403: symmetry phase with all the gauge bosons of the same mass $M\approx
404: M_{Z,W}$ introduced by hand;
405: (ii) the QED contribution with an auxiliary photon mass $M$
406: is factorized as in Eq.~(\ref{matf}) leaving the pure electroweak logarithms.
407: The separated virtual QED corrections accompanied by the real photon radiation
408: in the limit of vanishing photon mass result in the universal infrared safe factor
409: independent of $M_{Z,W}$.
410:
411: By contrast, the N$^3$LL approximation is sensitive to
412: fine details of the gauge boson mass generation and
413: the coefficient of the linear two-loop electroweak
414: logarithm depends {\it e.g.} on the Higgs boson mass.
415: For the full calculation of this coefficient
416: one has to use the true mass eigenstates of the standard model.
417: Our result, Eqs.~(\ref{2loopf},~\ref{2looptf}), is an example of such a
418: calculation when applied to the two-loop diagrams with photon and
419: $Z$ boson exchanges. We can, however, make a reasonable approximation
420: which dramatically simplifies the analysis. Namely, consider
421: a simplified model with a Higgs boson of zero hypercharge.
422: Then the mixing is absent and the hypercharge gauge boson
423: remains massless. The interference diagrams including the heavy $SU_L(2)$
424: and the light hypercharge $U(1)$ gauge bosons are identical with the ones of the
425: purely Abelian model discussed in this Letter, where the above two-step procedure
426: can be applied to get all the two-loop logarithms including the linear term.
427: In the standard model the mixing of the gauge bosons
428: results in a linear-logarithmic contribution, which is not accounted
429: for within this procedure. It is, however, suppressed by a small factor
430: $\sin^2{\theta_W}\approx 0.2$, with $\theta_W$ being the Weinberg angle.
431: Therefore, the above simplified model gives an estimate of the coefficient
432: in front of the linear electroweak logarithm with $20\%$ accuracy.
433: From the numerical result of Fig.~\ref{fig1},
434: which represents the typical structure of the two-loop
435: corrections, we see that a $20\%$ error in this coefficient
436: leads to an uncertainty comparable to the nonlogarithmic
437: contribution and is practically negligible. Thus we are able to get an accurate
438: estimate of the two-loop correction, which is sufficient
439: for practical applications to the future
440: collider physics, by performing the calculations
441: in the model without mass gap and mixing of the gauge bosons.
442: The last ingredient necessary to
443: complete the calculation of the dominant two-loop electroweak
444: corrections is the generalization of Eq.~(\ref{2loopf})
445: up to the linear-logarithmic term to
446: the pure $SU_L(2)$ gauge model with the Higgs mechanism
447: of mass generation. Note that up to the quadratic logarithm
448: the two-loop corrections are predicted by the
449: evolution equation \cite{KMPS}.
450:
451: To conclude, we have obtained the complete results for the
452: two-loop corrections to the vector form factor in the Sudakov limit
453: in Abelian theories with one massive gauge boson or with two
454: gauge bosons of essentially different masses. The results are in
455: full agreement with the predictions of the evolution equation approach.
456: We have formulated a systematic procedure
457: of factorizing the infrared singular virtual corrections
458: and reducing the calculation in the theory with mass gap to the
459: single-mass problem.
460: The analysis can directly be generalized to the
461: standard model with spontaneous breaking of the $SU_L(2)\times U(1)$
462: theory to low-energy QED.
463: This solves the principal problems of the calculation of the
464: dominant two-loop electroweak corrections
465: to the neutral current four-fermion processes
466: which are mandatory for the high-precision physics at
467: the LHC and the next generation of linear colliders.
468:
469:
470: \begin{acknowledgments}
471: We thank S. Pozzorini for useful comments on the manuscript.
472: J.H.K. acknowledges the hospitality of Kavli Institute for
473: Theoretical Physics and partial support by the NSF under
474: Grant No. PHY99-0794.
475: The work of J.H.K and A.A.P. was supported in part by BMBF Grant No.\
476: 05HT4VKA/3 and Sonderforschungsbereich Transregio 9. The work of V.A.S. was
477: supported in part by Volkswagen Foundation
478: Contract No. I/77788, and DFG Mercator Visiting Professorship No. Ha
479: 202/110-1.
480: \end{acknowledgments}
481:
482:
483:
484: \begin{thebibliography}{99}
485:
486: \bibitem{Sud} V.V. Sudakov, {Zh.\ Eksp.\ Teor.\ Fiz.\ } {\bf 30}, 87 (1956).
487:
488: \bibitem{Jac} R. Jackiw, {Ann.\ Phys.\ } {\bf 48}, 292 (1968); {\bf 51}, 575 (1969).
489:
490: \bibitem{Kur} M. Kuroda, G. Moultaka and D. Schildknecht, {Nucl.\ Phys.\ }
491: {\bf B350}, 25 (1991);
492: G. Degrassi and A. Sirlin, {Phys.\ Rev.\ } D {\bf 46}, 3104 (1992).
493:
494: \bibitem{Bec1} M. Beccaria {\it et al.}, {Phys.\ Rev.\ } D {\bf 58}, 093014 (1998);
495: P. Ciafaloni and D. Comelli, {Phys.\ Lett.\ } B {\bf 446},
496: 278 (1999); J.H. K\"uhn and A.A. Penin, Report TTP/99-28,
497: hep-ph/9906545.
498:
499: \bibitem{Fad} V.S. Fadin, L.N. Lipatov, A.D. Martin, and M. Melles,
500: {Phys.\ Rev.\ } D {\bf 61}, 094002 (2000).
501:
502: \bibitem{KPS} J.H. K\"uhn, A.A. Penin, and V.A. Smirnov,
503: {Eur.\ Phys.\ J.\ } C {\bf 17}, 97 (2000);
504: {Nucl.\ Phys.\ B (Proc. Suppl.)} {\bf 89}, 94 (2000).
505:
506: \bibitem{Bec2} M. Beccaria {\it et al.}, {Phys.\ Rev.\ } D {\bf 61},
507: 011301 (2000); D {\bf 61}, 073005 (2000);
508: M. Beccaria, F.M. Renard, and C. Verzegnassi,
509: {Phys.\ Rev.\ } D {\bf 63}, 053013 (2001);
510: D {\bf 64}, 073008 (2001); A. Denner and S. Pozzorini,
511: {Eur.\ Phys.\ J.\ } C {\bf 18}, 461 (2001); C {\bf
512: 21}, 63 (2001).
513:
514: \bibitem{HKK} M. Hori, H. Kawamura, and J. Kodaira,
515: {Phys.\ Lett.\ } B {\bf 491}, 275 (2000);
516: W. Beenakker and A. Werthenbach, {Nucl.\ Phys.\ } {\bf B630}, 3 (2002);
517: A. Denner, M. Melles, and S. Pozzorini, {Nucl.\ Phys.\ }
518: {\bf B662}, 299 (2003); S. Pozzorini, Report TTP04-01 and hep-ph/0401087.
519:
520: \bibitem{CCC} M. Ciafaloni, P. Ciafaloni, and D. Comelli,
521: {Phys.\ Rev.\ Lett.\ } {\bf 84}, 4810 (2000);
522: {\bf 87}, 211802 (2001);
523: {Nucl.\ Phys.\ } {\bf B589}, 359 (2000).
524:
525: \bibitem{KMPS} J.H. K\"uhn, S. Moch, A.A. Penin, and V.A. Smirnov,
526: {Nucl.\ Phys.\ } {\bf B616}, 286 (2001);
527: {\bf B648}, 455(E) (2002).
528:
529: \bibitem{FKM} B. Feucht, J.H. K\"uhn, and S. Moch,
530: {Phys.\ Lett.\ } B {\bf 561}, 111 (2003).
531:
532: \bibitem{Fre} J. Frenkel and J.C. Taylor, {Nucl.\ Phys.\ }
533: {\bf B116}, 185 (1976);
534: D. Amati, R.~Petronzio and G.~Veneziano,
535: {Nucl.\ Phys.\ } {\bf B146}, 29 (1978).
536:
537: \bibitem{BenSmi} M. Beneke and V.A. Smirnov, {Nucl.\ Phys.\ } {\bf B522}, 321 (1998);
538: V.A. Smirnov and E.R. Rakhmetov, {Teor.\ Mat.\ Fiz.\ } {\bf 120}, 64 (1999);
539: V.A. Smirnov, {Phys.\ Lett.\ } B {\bf 465}, 226 (1999);
540: V.A. Smirnov, {\it Applied Asymptotic Expansions in Momenta and Masses}
541: (Springer-Verlag, Heidelberg, 2001).
542:
543: \bibitem{Ste} G. Sterman, {Phys.\ Rev.\ } D {\bf 17} 2773 (1978);
544: S. Libby and G. Sterman {Phys.\ Rev.\ } D {\bf 18}, 3252 (1978);
545: A.H. Mueller, {Phys. Rev.} D {\bf 18}, 3705 (1978);
546: {Phys.\ Rep.\ } {\bf 73}, 237 (1981).
547:
548: \bibitem{KraLam} G. Kramer and B. Lampe, {Z. Phys.} C {\bf 34}, 497 (1987);
549: C {\bf 42}, 504(E) (1989);
550: T. Matsuura, S.C. van der Marck, and W.L. van Neerven,
551: {Nucl.\ Phys.\ } {\bf B319} (1989) 570.
552:
553: \bibitem{Mue} A.H. Mueller {Phys. Rev.} D {\bf 20}, 2037 (1979);
554: J.C. Collins, {Phys. Rev.} D {\bf 22}, 1478 (1980);
555: in {\it Perturbative QCD}, ed. A.H. Mueller, 1989,
556: p. 573; A. Sen, {Phys.\ Rev.\ } D {\bf 24}, 3281 (1981).
557:
558: \bibitem{KorRad} G.P. Korchemsky and A.V. Radyushkin,
559: {Nucl.\ Phys.\ } {\bf B283}, 342 (1987).
560:
561: \bibitem{KLN} T. Kinoshita, {J.\ Math.\ Phys.\ } {\bf 3}, 650, (1962);
562: T.D. Lee and M. Nauenberg, {Phys.\ Rev.\ }
563: D {\bf 133}, 1549 (1964).
564:
565: \bibitem{Sen} A. Sen, {Phys.\ Rev.\ } D {\bf 28}, 860 (1983).
566:
567: \bibitem{AGOT} C. Anastasiou, E.W.N. Glover, C. Oleari, and
568: M.E. Tejeda-Yeomans, {Nucl.\ Phys.\ } {\bf B601}, 341
569: (2001); E.W.N. Glover, Report DCTP-04-04 and hep-ph/0401119.
570:
571: \bibitem{SteTej} G.~Sterman and M.E.~Tejeda-Yeomans,
572: {Phys.\ Lett.\ } B {\bf 552}, 48 (2003).
573:
574: \end{thebibliography}
575:
576: \end{document}
577:
578:
579:
580:
581:
582:
583:
584:
585: