1: %%
2: %% $Modified: Tue Jun 29 13:45:47 2004 by puwer $
3: %%
4: \documentclass[fleqn,twoside]{article}
5: \usepackage{espcrc2}
6:
7: \usepackage{pslatex}
8: \usepackage{graphicx}
9: %%\usepackage{hyperref}
10: \usepackage[latin1]{inputenc}
11: \usepackage[T1]{fontenc}
12:
13:
14: %% put your own definitions here:
15: \def\mhiggs{{m_H}}
16: \def\Fig#1{Fig.~\ref{#1}}
17: \def\Tab#1{Tab.~\ref{#1}}
18: \def\Eq#1{Eq.~(\ref{#1})}
19: \def\Ref#1{Ref.~\cite{#1}}
20: \def\Refs#1{Refs.~\cite{#1}}
21: \def\ttbarjet{\ensuremath{t \bar t + \mbox{jet}}}
22: \def\nn{\nonumber}
23:
24:
25: %% declarations for front matter
26: \title{Top quark pair + jet production at next-to-leading order:\\
27: NLO QCD corrections to $gg\to t\bar t g$}
28:
29: \author{
30: A.~Brandenburg\address{DESY-Theorie, 22603 Hamburg, Germany}%
31: \thanks{Heisenberg Fellow of the
32: Deutsche Forschungsgemeinschaft},
33: S.~Dittmaier\address[MPI]{Max-Planck-Institut für Physik
34: (Werner-Heisenberg-Institut), Föhringer
35: Ring 6, D-80805 Munich, Germany},
36: P.~Uwer\address{Department of Physics,
37: TH Division, CERN, CH-1211 Geneva 23,
38: Switzerland
39: },
40: S.~Weinzierl\addressmark[MPI]\footnotemark[1]
41: }
42:
43: \begin{document}
44: \setcounter{page}{0}
45: \thispagestyle{empty}
46: \begin{flushright}
47: \large
48: CERN-PH-TH/2004-122\\
49: DESY 04-111\\
50: MPP-2004-74
51: \end{flushright}
52: \vspace*{4cm}
53:
54: \begin{center}
55: \Large
56: {\bf Top quark pair + jet production at next-to-leading order:\\
57: NLO QCD corrections to $gg\to t\bar t g$}
58: \vspace*{2cm}
59:
60: A.~Brandenburg$^a$,
61: S.~Dittmaier$^b$, P.~Uwer$^c$, S.~Weinzierl$^b$\\[0.5cm]
62:
63: {\large $^a$DESY-Theorie, 22603 Hamburg, Germany\\[0.2cm]
64:
65: $^b$Max-Planck-Institut für Physik
66: (Werner-Heisenberg-Institut),\\ Föhringer
67: Ring 6, D-80805 Munich, Germany\\[0.2cm]
68:
69: $^c$Department of Physics,
70: TH Division, CERN, CH-1211 Geneva 23,
71: Switzerland}\\[0.5cm]
72:
73: {\large \bf Abstract}\\[0.1cm]
74:
75: \parbox{16cm}{\large The reaction $pp/p\bar p \to \ttbarjet+X$ is an important background
76: process for Higgs boson searches in the mass range below 200~GeV. Apart
77: from that it is also an ideal laboratory for precision measurements in
78: the top quark sector. Both applications require a solid theoretical
79: prediction, which can be achieved only through a full next-to-leading
80: order (NLO) calculation.
81: In this work we describe the NLO computation of the
82: subprocess $gg\to t \bar t g$.}
83: \end{center}
84:
85:
86: \begin{abstract}
87: The reaction $pp/p\bar p \to \ttbarjet+X$ is an important background
88: process for Higgs boson searches in the mass range below 200~GeV. Apart
89: from that it is also an ideal laboratory for precision measurements in
90: the top quark sector. Both applications require a solid theoretical
91: prediction, which can be achieved only through a full next-to-leading
92: order (NLO) calculation.
93: In this work we describe the NLO computation of the
94: subprocess $gg\to t \bar t g$.
95: \end{abstract}
96: \maketitle
97:
98: \section{Introduction}
99: The main objective of the Large Hadron Collider (LHC) at CERN is
100: the discovery of the Higgs boson and the measurement of its mass and couplings.
101: To achieve this important goal a solid knowledge of the production
102: mechanisms and the corresponding backgrounds is mandatory. In the
103: Standard Model a light Higgs boson is currently favoured by the
104: available data.
105: Using the recently
106: updated top mass of $m_t = 178.0 \pm 4.3$~GeV \cite{Azzi:2004rc}
107: the electroweak fits yield an upper bound of 251~GeV (at~95\%~C.L.)
108: and a central
109: value of $\mhiggs = 117$~GeV \cite{SRoth-Moriond}.
110: To achieve a high signal significance in the Higgs searches
111: in general, different production and decay mechanisms are combined.
112: In the range up to 200~GeV the so-called weak boson
113: fusion (WBF) process with the subsequent decay of the Higgs into a W-boson
114: pair plays a dominant r\^ole. The most important background for the
115: WBF process comes from the \ttbarjet\ process \cite{Alves:2003vp}.
116: A very precise knowledge of this process is thus
117: mandatory for the discovery of the Higgs boson. It is obvious that for precise
118: measurements of the couplings a precise background
119: determination is equally important. For example,
120: it has been shown in \Ref{Rainwater:2002hm}
121: that even if one assumes only a 10\% uncertainty of the
122: \ttbarjet\ cross section it is still the dominant theoretical
123: uncertainty in the measurement of
124: $\sigma_{\rm H}=\sigma_{\rm WBF} \times B(H\to WW)$.
125: As also pointed out in \Ref{Rainwater:2002hm}
126: this accuracy might be achievable only
127: through a full next-to-leading order (NLO) calculation. In a recent
128: analysis \cite{Cavalli:2002vs,Kauer:2004fg}, the possibility
129: to extract the background from
130: extrapolation of experimental data has been studied. In this analysis
131: it was found that a background determination with 5--10\% accuracy
132: might be possible. This is a very promising result. On the other hand
133: --- given the significance of the precise background determination ---
134: we believe that a cross-check with a full NLO QCD prediction is
135: important. At the very end --- having a good understanding of both
136: results --- both methods/results could and should be used as complementary.
137:
138: In fact the \ttbarjet\ reaction is not only important as background
139: for Higgs searches, it is also an important signal process on its own.
140: It is well known that top quark physics allows a test of the Standard Model at
141: high scale. In particular one can search for possible extensions
142: of the Standard Model at the scale of the top quark mass.
143: As far as top quark production at hadron colliders is concerned,
144: the state of the art is as follows. The
145: differential cross section for top quark pair production is known to
146: next-to-leading order accuracy in QCD
147: \cite{Nason:1988xz,Nason:1989zy,Beenakker:1989bq,Beenakker:1991ma}.
148: In addition the resummation of logarithmically enhanced contributions has
149: been studied in detail in
150: \Refs{Laenen:1992af,Laenen:1994xr,Kidonakis:1995wz,Berger:1996ad,Catani:1996yz,Kidonakis:1997gm,Berger:1998gz,Bonciani:1998vc}.
151: Recently also the spin correlations between top quark and antitop quark
152: were calculated at NLO in QCD \cite{Bernreuther:2001rq,Bernreuther:2004jv}.
153: Since single top quark production provides an excellent
154: opportunity to test the charged-current weak interaction of the top
155: quark, it has also attracted a lot of interest in the past. In
156: particular NLO corrections were studied in
157: \Refs{Bordes:1995ki,Smith:1996ij,Stelzer:1997ns,Harris:2002md}. In
158: \Ref{Harris:2002md} the NLO corrections for the fully
159: differential cross section are given, keeping also the spin information
160: of the top quark.
161:
162: In that context the natural next step is the calculation of the NLO
163: corrections for \ttbarjet\ production. As far as top quark physics
164: is concerned, interesting observables to study are those that vanish
165: if there is no additional jet. Such observables allow for a direct test
166: of the dynamics in the top quark sector. For example the
167: asymmetry \cite{Rainwater}
168: \begin{equation}
169: A(y) = { N(\ell,{\rm forward}) -N(\ell,{\rm backward})\over
170: N(\ell,{\rm forward}) +N(\ell,{\rm backward})}
171: \end{equation}
172: is such an observable. Here $ N(\ell,{\rm forward/backward})$
173: denotes the number of forward/backward-going leptons as a function
174: of the rapidity $y$ of the additional
175: jet. The measurement of this asymmetry allows for a direct test of the
176: production and decay mechanisms.
177: Using similar observables one can search for example
178: for anomalous top--gluon couplings. A more precise understanding of
179: the cross sections for $pp/p\bar p \to t \bar t + \mbox{jets} +X$
180: is also important for measurements of the top quark mass.
181:
182:
183: \section{Outline of the calculation}
184: In this section we briefly summarize the calculation of the NLO corrections
185: for the subprocess $gg\to t\bar t g$.
186: In view of the number of external legs
187: and the top mass as additional parameter it is obvious that even
188: partial results are in general quite lengthy.
189: In the following we will restrict our attention only to those parts
190: of the calculation
191: where special care is needed to construct a numerically stable program.
192:
193: \subsection{Virtual corrections}
194: The calculation of the virtual corrections proceeds via the following
195: steps:
196: \begin{enumerate}
197: \item Generation of the Feynman diagrams using for example
198: Feynarts \cite{Kublbeck:1990xc} or QGRAF \cite{Nogueira:1993ex}.
199: \item Reduction of the tensor integrals to scalar one-loop integrals.
200: \item Reduction of the amplitudes to standard matrix elements.
201: \item Numerical phase-space integration of the squared
202: matrix elements, including appropriate phase-space cuts.
203: \end{enumerate}
204: Technically the most complicated part is the evaluation of the
205: pentagon-diagrams.
206: \begin{figure}[htbp]
207: \begin{center}
208: \includegraphics[width=3.cm]{d263}\hfill
209: \includegraphics[width=3.cm]{d266}
210: \caption{Sample Feynman diagrams contributing to the virtual corrections.
211: \label{fig:Virtual}}
212: \end{center}
213: \end{figure}
214: Two sample diagrams are shown in \Fig{fig:Virtual}. Let us first
215: address the evaluation of the scalar 5-point integrals. To calculate
216: these, we use two different methods. One
217: calculation is based on the method given in
218: \Refs{Beenakker:2001rj,Beenakker:2002nc}.
219: The basic idea of this method is that finite
220: 5-point integrals can be expressed in terms of 4-point integrals (see
221: for example \cite{Melrose:1965kb,vanNeerven:1984vr,Bern:1994kr}). To
222: apply this observation also to soft- and mass-singular integrals they
223: are rewritten according to \Refs{Beenakker:2001rj,Beenakker:2002nc}
224: in the following way:
225: \begin{equation}
226: E^d = E^{d}_{\rm sing.} + \left[ E^{({\rm mass},d=4)}
227: -E^{({\rm mass},d=4)}_{\rm sing.}\right].
228: \end{equation}
229: Here $E^d$ denotes the original 5-point integral in $d$ dimensions
230: while $E^{({\rm mass},d=4)}$ is obtained from the original integral
231: by dressing the massless propagators with a small mass $\lambda$.
232: The subtraction term $E^{({\rm mass},d=4)}_{\rm sing.}$, which
233: has the same singular structure as the 5-point integral
234: $E^{({\rm mass},d=4)}$ in the limit $\lambda \to 0$, is
235: obtained by studying the soft and collinear behaviour of
236: $E^{({\rm mass},d=4)}$
237: and can be expressed in terms of 3-point functions \cite{Dittmaier:2003bc}.
238: Rewriting now the finite integral $E^{({\rm mass},d=4)}$ in terms
239: of 4-point integrals we thus succeeded
240: in expressing the original 5-point integral in terms of
241: 3- and 4-point functions.
242: A more detailed discussion can be found in \Ref{Beenakker:2002nc}.
243: The second method we used to calculate the five-point integrals is
244: based on the fact that, even for divergent integrals, it is possible to
245: obtain a representation as linear combination of 4-point
246: integrals (see for example \Ref{Bern:1994kr}).
247: Expressing the 4-point function for $d=4-2\epsilon$ in terms of the
248: finite 4-point function
249: in 6 dimensions plus a combination of 3-point integrals allows us also
250: to
251: shift all the divergences to the 3-point integrals.
252: Defining the 5-point functions through
253: \begin{eqnarray}
254: &&E^d(p_0,p_1,p_2,p_3,p_4,m_0,m_1,m_2,m_3,m_4)\nn\\
255: &=&
256: {1\over i\*\pi^2}
257: \int d^d\ell \prod_{j=0}^{4}{1\over (\ell+p_j)^2-m_j^2+i\epsilon},
258: \end{eqnarray}
259: we obtain for example
260: \begin{eqnarray}
261: &&\hspace{-1cm}
262: \left.E_0(0,p_1,p_1-p_3,p_4-p_2,-p_2,m_t,m_t,0,0,m_t)\right|_{\rm
263: sing.} \nn \\
264: &=&
265: P(t_{13})P(s_{45})\*C_0(p_1-p_3,p_4-p_2,p_1,0,0,m_t) \nn \\
266: &+&
267: P(t_{24})P(s_{35})\*C_0(p_4-p_2,-p_2,p_1-p_3,0,m_t,0)\nn \\
268: &-&
269: (t_{13}-t_{24})^2 P(t_{13})P(t_{24})P(s_{35})P(s_{45})\nn\\
270: &&\times C_0(0,p_1-p_3,p_4-p_2,m_t,0,0),
271: \end{eqnarray}
272: with $P(x) = 1/(x-m_t^2)$ and $s_{ij} = (p_i+p_j)^2$,
273: $t_{ij} = (p_i-p_j)^2$. The parton momenta are assigned according to
274: $g(p_1)g(p_2) \to t(p_3)\bar t(p_4) g(p_5)$.
275: For the cases at hand it is possible to solve all the required
276: box-integrals in 6 dimensions.
277: We checked that the two methods yield the same results for the 5-point
278: integrals $E^d$.
279:
280: Having solved the scalar integrals, the next step is the reduction of
281: the 5-point tensor integrals to scalar one-loop integrals.
282: In principle one could attack
283: this problem using the standard Passarino--Veltman approach
284: \cite{Passarino:1979jh}.
285: This method leads to spurious singularities in
286: individual terms at the
287: phase-space boundary due to vanishing Gram determinants in the
288: denominator.
289: These spurious singularities create
290: numerical instabilities when doing the phase-space integration.
291: Note that the spurious singularities cancel if one combines the
292: individual terms analytically before doing the numerical integration.
293: One solution of this problem is a time-consuming
294: extrapolation technique, as was used for example in
295: \Ref{Beenakker:2002nc}. As an alternative to the extrapolation
296: technique a different reduction procedure \cite{Denner:2002ii} was also
297: used in \Ref{Beenakker:2002nc}.
298: In this work we
299: follow the method developed in \Ref{Denner:2002ii}.
300: Essentially the same technique to reduce scalar 5-point integrals to scalar
301: 4-point integrals is also applied to the tensor integrals. In this way the
302: 5-point tensor integrals are directly reduced to 4-point tensor
303: integrals. The explicit calculation shows that in this way the
304: spurious singularities in individual terms, due to vanishing Gram
305: determinants depending on 4 external momenta, are avoided \cite{Denner:2002ii}.
306:
307: Let us just mention at the end that there are also other methods to solve
308: the scalar 5-point integrals and perform the reduction of the tensor
309: integrals. For example one could also use the methods developed in
310: \Refs{Giele:2004iy,Binoth:1999sp,Duplancic:2003tv}.
311: (For \Ref{Giele:2004iy} see also Walter Giele's talk in these proceedings.)
312:
313: \subsection{Real corrections}
314: The calculation of the required matrix elements is straightforward.
315: \begin{figure}[htbp]
316: \begin{center}
317: \includegraphics[width=3.cm]{d20}
318: \caption{Sample Feynman diagram contributing to the real corrections.
319: \label{fig:Real}}
320: \end{center}
321: \end{figure}
322: A sample diagram for the reaction $gg \to t \bar t gg$ is shown in
323: \Fig{fig:Real}. We used two different methods to obtain the required
324: colour-ordered helicity amplitudes:
325: \begin{enumerate}
326: \item A Feynman-diagram-based approach where we evaluate all the
327: diagrams contributing to one specific colour-ordered subamplitude.
328: \item Using the recurrence relations à la Berends and Giele
329: \cite{Berends:1988me}.
330: \end{enumerate}
331: We find complete agreement in the results of the two methods.
332: Furthermore we also checked
333: that our results agree with the ones obtained using
334: Madgraph \cite{Stelzer:1994ta}.
335: To extract the singularities from collinear or soft partons we use the
336: dipole subtraction method \cite{Catani:1997vz,Phaf:2001gc,Catani:2002hc}.
337: The idea of the subtraction method is to add and subtract a term which,
338: on the one hand, cancels pointwise the singularities of the
339: matrix elements in the singular regions of the phase-space and is, on the
340: other hand, easy enough to be integrated analytically. Schematically
341: the NLO contribution is then obtained from the following formula:
342: \begin{eqnarray}
343: \label{eq:SubtractionMethod}
344: \sigma_{\rm NLO}
345: &=&
346: \underbrace{\int_{m+1}\left[\sigma_{\rm real}-\sigma_{\rm sub}\right]}_{\rm
347: finite}
348: +
349: \underbrace{\int_{m}\left[\sigma_{\rm virt.}+\int_1 \bar\sigma^1_{\rm
350: sub}\right]}_{\rm finite}\nn\\
351: &+&
352: \underbrace{\int dx\int_{m}\left[\sigma_{\rm fact.}(x)
353: +\bar\sigma_{\rm sub}(x)\right]}_{\rm finite}.
354: \end{eqnarray}
355: Here $\sigma_{\rm fact.}(x)$ denotes the contribution from the
356: factorization of initial-state singularities due to the presence of
357: coloured partons in the initial
358: state. The contributions $\bar\sigma^1_{\rm
359: sub},\bar\sigma_{\rm sub}$ are obtained from $\sigma_{\rm sub}$
360: by integrating out the `unresolved' parton. The result is split
361: into the two terms $\bar\sigma^1_{\rm
362: sub},\bar\sigma_{\rm sub}$ to render the last two integrals
363: individually finite. A remarkable feature of the subtraction method is
364: that the analytic integration of the subtraction has to be done only
365: once and that in the whole procedure no approximation is made.
366: This is made possible by the universality of soft and collinear
367: factorization in QCD. The explicit expressions for $\sigma_{\rm sub},
368: \sigma^1_{\rm sub},$ and $\bar\sigma_{\rm sub}$ can be obtained from
369: the colour-ordered subamplitudes using the formulae given in
370: \Ref{Catani:2002hc}. In particular $\sigma_{\rm sub}$ is obtained
371: from a sum over individual {\it dipole contributions}. In the case at hand we
372: have to include the contribution from 36 individual dipoles. We do not
373: consider the splitting $g\to t\bar t$ because the divergence is
374: regulated by the quark masses. (For light quarks one could consider the
375: corresponding dipoles to render the integration numerically more stable.)
376: We have checked that the combination of the 36 dipoles indeed
377: reproduces all the singular limits arising from single unresolved
378: configurations.
379:
380: \section{Status and results}
381: The current status of the project is as follows. Most of the separate
382: contributions are implemented in the form of computer
383: programs allowing the numerical evaluation of the cross sections.
384: \begin{figure}[htbp]
385: \begin{center}
386: \includegraphics[width=7.5cm]{parton}
387: \caption{ Result for the virtual corrections for the
388: subprocess $gg\to t\bar t g$
389: as defined by the
390: second term in \Eq{eq:SubtractionMethod}\ ($k_\perp>$~20~GeV).
391: \label{fig:ResultVirt}}
392: \end{center}
393: \end{figure}
394: In \Fig{fig:ResultVirt} we show as an example, the result at the parton level
395: for the virtual
396: corrections (defined as the second term in \Eq{eq:SubtractionMethod})
397: for different centre-of-mass energies.
398: (Given that the separation shown in \Eq{eq:SubtractionMethod}
399: involves some freedom, this individual contribution does not have a
400: direct physical interpretation
401: unless the remaining contributions are added --- we just show it for
402: illustrative purposes.) As can be seen from \Fig{fig:ResultVirt} the
403: method we used for the treatment of the tensor integrals gives indeed
404: numerically stable results.
405: Furthermore we note that the inclusion of $d\bar \sigma^1_{\rm sub}$
406: together with the renormalization of the coupling and the quark mass
407: renders the second term in \Eq{eq:SubtractionMethod}
408: finite, as it must be. This is an important cross-check.
409: As mentioned earlier we also checked that the integrand for the
410: first contribution in \Eq{eq:SubtractionMethod} is also finite for all
411: single unresolved phase-space configurations.
412: Given the complexity of the project we think it is very important to
413: have independent cross-checks for every individual contribution.
414: While most of the calculation is already cross-checked, we still
415: work to finish also the remaining checks. Complete results will
416: be presented elsewhere.
417:
418: \section{Conclusions}
419: In this work we discuss the NLO calculation for the partonic
420: reaction $gg\to t \bar t g$. Up to remaining cross-checks, the
421: calculation is almost finished. In particular we have shown that the
422: virtual corrections are stable using the reduction procedure discussed
423: in \Ref{Denner:2002ii}.
424:
425:
426: {\bf Acknowledgements:} We would like to thank the organizers for the
427: pleasant atmosphere at Loops and Legs 2004.
428:
429: %%\bibliographystyle{h-elsevier2}
430: %%\bibliography{literatur}
431: \newcommand{\zp}{Z. Phys. }\def\as{\alpha_s }\newcommand{\prd}{Phys. Rev.
432: }\newcommand{\pr}{Phys. Rev. }\newcommand{\prl}{Phys. Rev. Lett.
433: }\newcommand{\npb}{Nucl. Phys. }\newcommand{\psnp}{Nucl. Phys. B (Proc.
434: Suppl.) }\newcommand{\pl}{Phys. Lett. }\newcommand{\ap}{Ann. Phys.
435: }\newcommand{\cmp}{Commun. Math. Phys. }\newcommand{\prep}{Phys. Rep.
436: }\newcommand{\jmp}{J. Math. Phys. }\newcommand{\rmp}{Rev. Mod. Phys. }
437: \begin{thebibliography}{10}
438:
439: \bibitem{Azzi:2004rc}
440: CDF Collaborattion, P. Azzi et~al.
441: \newblock (2004), hep-ex/0404010.
442:
443: \bibitem{SRoth-Moriond}
444: S. Roth,
445: \newblock {Talk given at ``Rencontres de Moriond: QCD and Hadronic
446: interactions'', La~Thuile (Italy), March 28 - April 4 2004}.
447:
448: \bibitem{Alves:2003vp}
449: A. Alves et~al.,
450: \newblock Phys. Rev. D69 (2004) 075005, hep-ph/0309042.
451:
452: \bibitem{Rainwater:2002hm}
453: D.L. Rainwater, M. Spira and D. Zeppenfeld
454: \newblock (2002), hep-ph/0203187.
455:
456: \bibitem{Cavalli:2002vs}
457: D. Cavalli et~al.
458: \newblock (2002), hep-ph/0203056.
459:
460: \bibitem{Kauer:2004fg}
461: N. Kauer
462: \newblock (2004), hep-ph/0404045.
463:
464: \bibitem{Nason:1988xz}
465: P. Nason, S. Dawson and R.K. Ellis,
466: \newblock Nucl. Phys. B303 (1988) 607.
467:
468: \bibitem{Nason:1989zy}
469: P. Nason, S. Dawson and R.K. Ellis,
470: \newblock Nucl. Phys. B327 (1989) 49.
471:
472: \bibitem{Beenakker:1989bq}
473: W. Beenakker et~al.,
474: \newblock Phys. Rev. D40 (1989) 54.
475:
476: \bibitem{Beenakker:1991ma}
477: W. Beenakker et~al.,
478: \newblock Nucl. Phys. B351 (1991) 507.
479:
480: \bibitem{Laenen:1992af}
481: E. Laenen, J. Smith and W.L. van Neerven,
482: \newblock Nucl. Phys. B369 (1992) 543.
483:
484: \bibitem{Laenen:1994xr}
485: E. Laenen, J. Smith and W.L. van Neerven,
486: \newblock Phys. Lett. B321 (1994) 254, hep-ph/9310233.
487:
488: \bibitem{Kidonakis:1995wz}
489: N. Kidonakis and J. Smith,
490: \newblock Phys. Rev. D51 (1995) 6092, hep-ph/9502341.
491:
492: \bibitem{Berger:1996ad}
493: E.L. Berger and H. Contopanagos,
494: \newblock Phys. Rev. D54 (1996) 3085, hep-ph/9603326.
495:
496: \bibitem{Catani:1996yz}
497: S. Catani et~al.,
498: \newblock Nucl. Phys. B478 (1996) 273, hep-ph/9604351.
499:
500: \bibitem{Kidonakis:1997gm}
501: N. Kidonakis and G. Sterman,
502: \newblock Nucl. Phys. B505 (1997) 321, hep-ph/9705234.
503:
504: \bibitem{Berger:1998gz}
505: E.L. Berger and H. Contopanagos,
506: \newblock Phys. Rev. D57 (1998) 253, hep-ph/9706206.
507:
508: \bibitem{Bonciani:1998vc}
509: R. Bonciani et~al.,
510: \newblock Nucl. Phys. B529 (1998) 424, hep-ph/9801375.
511:
512: \bibitem{Bernreuther:2001rq}
513: W. Bernreuther et~al.,
514: \newblock Phys. Rev. Lett. 87 (2001) 242002, hep-ph/0107086.
515:
516: \bibitem{Bernreuther:2004jv}
517: W. Bernreuther et~al.,
518: \newblock Nucl. Phys. B690 (2004) 81, hep-ph/0403035.
519:
520: \bibitem{Bordes:1995ki}
521: G. Bordes and B. van Eijk,
522: \newblock Nucl. Phys. B435 (1995) 23.
523:
524: \bibitem{Smith:1996ij}
525: M.C. Smith and S. Willenbrock,
526: \newblock Phys. Rev. D54 (1996) 6696, hep-ph/9604223.
527:
528: \bibitem{Stelzer:1997ns}
529: T. Stelzer, Z. Sullivan and S. Willenbrock,
530: \newblock Phys. Rev. D56 (1997) 5919, hep-ph/9705398.
531:
532: \bibitem{Harris:2002md}
533: B.W. Harris et~al.,
534: \newblock Phys. Rev. D66 (2002) 054024, hep-ph/0207055.
535:
536: \bibitem{Rainwater}
537: D. Rainwater,
538: \newblock {private communication}.
539:
540: \bibitem{Kublbeck:1990xc}
541: J. Küblbeck, M. Böhm and A. Denner,
542: \newblock Comput. Phys. Commun. 60 (1990) 165.
543:
544: \bibitem{Nogueira:1993ex}
545: P. Nogueira,
546: \newblock J. Comput. Phys. 105 (1993) 279.
547:
548: \bibitem{Beenakker:2001rj}
549: W. Beenakker et~al.,
550: \newblock Phys. Rev. Lett. 87 (2001) 201805, hep-ph/0107081.
551:
552: \bibitem{Beenakker:2002nc}
553: W. Beenakker et~al.,
554: \newblock Nucl. Phys. B653 (2003) 151, hep-ph/0211352.
555:
556: \bibitem{Melrose:1965kb}
557: D.B. Melrose,
558: \newblock Nuovo Cim. 40 (1965) 181.
559:
560: \bibitem{vanNeerven:1984vr}
561: W.L. van Neerven and J.A.M. Vermaseren,
562: \newblock Phys. Lett. B137 (1984) 241.
563:
564: \bibitem{Bern:1994kr}
565: Z. Bern, L.J. Dixon and D.A. Kosower,
566: \newblock Nucl. Phys. B412 (1994) 751, hep-ph/9306240.
567:
568: \bibitem{Dittmaier:2003bc}
569: S. Dittmaier,
570: \newblock Nucl. Phys. B675 (2003) 447, hep-ph/0308246.
571:
572: \bibitem{Passarino:1979jh}
573: G. Passarino and M.J.G. Veltman,
574: \newblock Nucl. Phys. B160 (1979) 151.
575:
576: \bibitem{Denner:2002ii}
577: A. Denner and S. Dittmaier,
578: \newblock Nucl. Phys. B658 (2003) 175, hep-ph/0212259.
579:
580: \bibitem{Giele:2004iy}
581: W.T. Giele and E.W.N. Glover,
582: \newblock JHEP 04 (2004) 029, hep-ph/0402152.
583:
584: \bibitem{Binoth:1999sp}
585: T. Binoth, J.P. Guillet and G. Heinrich,
586: \newblock Nucl. Phys. B572 (2000) 361, hep-ph/9911342.
587:
588: \bibitem{Duplancic:2003tv}
589: G. Duplancic and B. Nizic,
590: \newblock Eur. Phys. J. C35 (2004) 105, hep-ph/0303184.
591:
592: \bibitem{Berends:1988me}
593: F.A. Berends and W.T. Giele,
594: \newblock Nucl. Phys. B306 (1988) 759.
595:
596: \bibitem{Stelzer:1994ta}
597: T. Stelzer and W.F. Long,
598: \newblock Comput. Phys. Commun. 81 (1994) 357, hep-ph/9401258.
599:
600: \bibitem{Catani:1997vz}
601: S. Catani and M.H. Seymour,
602: \newblock Nucl. Phys. B485 (1997) 291, hep-ph/9605323.
603:
604: \bibitem{Phaf:2001gc}
605: L. Phaf and S. Weinzierl,
606: \newblock JHEP 04 (2001) 006, hep-ph/0102207.
607:
608: \bibitem{Catani:2002hc}
609: S. Catani et~al.,
610: \newblock Nucl. Phys. B627 (2002) 189, hep-ph/0201036.
611:
612: \end{thebibliography}
613:
614: \end{document}
615: