1: \documentclass[10pt,draft]{dis03}
2: \usepackage{epsf,amsmath}
3:
4: \renewcommand{\baselinestretch}{1.15}
5: \textwidth 12cm \textheight 17cm
6: \pagestyle{myheadings}
7:
8: \begin{document}
9:
10: \title{NNLO parton distributions from deep-inelastic-scattering data}
11:
12: \author{S.~Alekhin \\IHEP, Protvino \\E-mail: alekhin@sirius.ihep.su}
13:
14: \maketitle
15:
16: \begin{abstract}
17: \noindent
18: The parton distributions functions (PDFs) derived
19: from the NNLO QCD analysis of existing light-targets
20: deep-inelastic-scattering data are presented. The NLO and NNLO PDFs
21: are compared in order to analyze perturbative stability
22: of the analysis and estimate impact of the higher-order QCD
23: corrections. The main theoretical uncertainties and experimental uncertainties
24: in PDFs due to all sources of experimental errors in data
25: are estimated and used to assess corresponding
26: uncertainties in the cross sections of other hadronic processes.
27: \end{abstract}
28:
29: A study of the hard processes in the hadron collisions
30: requires knowledge of the parton distribution functions (PDFs). The
31: PDFs determine the normalization of the cross sections
32: and therefore very often the uncertainty in the PDFs puts limit on
33: the precision of measurements. For this reason the estimation
34: of the PDFs uncertainties and improvement of our knowledge of the PDFs
35: was the scope of intensive studies in recent years~\cite{Giele:2002hx}.
36: Estimation of the different uncertainties in PDFs shows that the
37: uncertainties in the ansatz of analysis (theoretical ones)
38: dominate over the uncertainties
39: propagated from the experimental errors in data used to constrain
40: the PDFs (experimental ones) for wide kinematics~\cite{Botje:1999dj}.
41: One of the potentially
42: dangerous theoretical uncertainty in the PDFs extracted in the NLO QCD
43: comes from the higher orders (NNLO) corrections. The NNLO
44: corrections are unknown for the most of hard processes, but
45: have been almost completely known for the DIS
46: with calculation of the splitting function moments up to
47: the 12-th~\cite{Retey:2000nq}.
48: The remaining uncertainty in the DIS structure functions due to
49: missing terms in the NNLO correction
50: was estimated at the level of few percent
51: for the realistic kinematics~\cite{vanNeerven:2000wp}.
52:
53:
54: \begin{figure}[h]
55: \vspace*{7.0cm}
56: \begin{center}
57: \special{psfile=pert.ps voffset=-20 vscale=45 hscale=65 hoffset=-15 angle=0}
58: \vskip 0.5cm
59: \caption[*]{The $1\sigma$ experimental error bands
60: for the A02 gluon distributions obtained
61: in the NNLO (solid lines) and the NLO (dashes) at different values of $Q$.
62: The NNLO gluon distribution for Set 1 of MRST2001 at
63: $Q^2=2~{\rm GeV}^2$ is given for comparison (dots).}
64: \end{center}
65: \end{figure}
66:
67: The approximate NNLO corrections of Ref.~\cite{vanNeerven:2000wp} were
68: applied
69: in the QCD analysis of the existing charged-leptons DIS data including
70: data from the SLAC, BCDMS, NMC, H1, and ZEUS
71: experiments~\cite{Alekhin:2002fv}.
72: The NNLO and the NLO gluon distributions obtained in this analysis (A02) are
73: compared in Fig.1. Variation of the gluon distribution
74: under account of the NNLO correction is generally within the experimental
75: errors that indicates perturbative stability of NNLO PDFs.
76: In the same plot we give the NNLO gluon distribution extracted from the
77: MRST global fit of ~Ref.\cite{Martin:2002dr}.
78: At low $Q$ the MRST gluon distribution is well below the A02 one and gets
79: negative at $x \sim 0.0001$. The origin of this difference
80: has not been clarified entirely. One of the possible reasons
81: is certain inconsistency of the NNLO MRST fit due to
82: the NNLO corrections have been applied for the DIS only, while the
83: MRST fit
84: includes the Drell-Yan (DY) and the jet production data as well.
85: This inconsistency was explicitly demonstrated in study of impact of the
86: NNLO correction on the DY cross sections~\cite{Anastasiou:2003yy}:
87: The NLO DY cross sections calculated using MRST PDFs are in good
88: agreement with experimental data, while the NNLO predictions are
89: well above the data. Note for the jet data used in the MRST analysis the
90: NNLO corrections are presumably even more important than for the DY process.
91: Therefore, for the NNLO
92: PDFs extracted from the DIS data the higher-order QCD
93: theoretical
94: errors are under much better control than for PDFs from the global fits.
95:
96: \begin{figure}[h]
97: \vspace*{7.0cm}
98: \begin{center}
99: \special{psfile=fl.ps voffset=-30 vscale=40 hscale=65 hoffset=-10 angle=0}
100: \vskip 1cm
101: \caption[*]{The measurements of $F_{\rm L}$
102: by the H1 collaboration~\protect\cite{H1}
103: compared to the NNLO QCD predictions (solid lines:
104: A02(leading twist+high twist); dashes: A02(leading twist); dots: MRST(leading
105: twist)).}
106: \end{center}
107: \end{figure}
108:
109: The NNLO QCD predictions for the structure function $F_{\rm L}$ based
110: on the different sets of PDFs are compared to the H1 measurements
111: of Ref.\cite{H1} in Fig.2. The MRST predictions are everywhere
112: below the data.
113: Since for the kinematics of Fig.2 $F_{\rm L}$ is defined mainly
114: by the gluon distribution this might mean that the
115: MRST underscore the gluon distribution at small $x$ and $Q$.
116: For the A02 prediction agreement with the data is much better.
117: At $x \sim 0.0001$ the agreement is achieved mainly due to the leading-twist
118: (LT) term, which was calculated using the NNLO PDFs.
119: For $x \sim 0.00001$ the main contribution comes from the high-twist (HT)
120: term in $F_{\rm L}$, which was fitted in the A02 analysis
121: simultaneously with the leading-twist term.
122: This HT term might be manifestation of a new phenomena
123: expected at small $x$ or can be caused by the conventional
124: effects like missing resummation of large logs
125: or, eventually, by a disagreement of
126: data from separate experiments used in the combined fit.
127: In any case the unusual behavior of $F_{\rm L}$ at small $x$ and $Q$
128: must stimulate more precise measurements in this region.
129:
130: \begin{figure}[t]
131: \vspace*{7.0cm}
132: \begin{center}
133: \special{psfile=fnal.ps voffset=-75 vscale=55 hscale=65 hoffset=-10 angle=0}
134: \vskip 1.7cm
135: \caption[*]{The NNLO $W/Z$ production rates in
136: the $\overline{p}p$ collisions at $\sqrt{s}=1.96~{\rm TeV}$ compared to the
137: preliminary results for Run II \protect\cite{Evans:2002wj}.
138: The area between dashes gives $1\sigma$ band
139: uncertainty in the calculations due to theoretical and experimental
140: uncertainties in the PDFs.}
141: \end{center}
142: \end{figure}
143:
144: With a good control of theoretical errors the NNLO PDFs
145: extracted from the inclusive DIS data provide good
146: tool for the precision measurements on the hadron colliders.
147: The NNLO predictions for the $W/Z$ cross sections
148: obtained using the code of Ref.\cite{Hamberg:1990np}
149: with the A02 PDFs are compared to the experimental data
150: in Fig.3. The prediction
151: error bands given in Fig.3 includes both theoretical and
152: experimental errors in PDFs. The accuracy of predictions is much better
153: than the errors in data, which are mostly due to uncertainty in
154: the luminosity. Therefore these prediction combined with the measurements
155: of the $W/Z$ rates can be used for better monitoring
156: of the rates of other processes including exotic ones.
157:
158: \section*{Acknowledgments} The work was supported by the RFBR grant
159: 03-02-17177.
160:
161: \begin{thebibliography}{0}
162:
163: \bibitem{Giele:2002hx}
164: W.~Giele {\it et al.}, arXiv:hep-ph/0204316.
165: %%CITATION = HEP-PH 0204316;%%
166:
167: \bibitem{Botje:1999dj}
168: M.~Botje, Eur.\ Phys.\ J.\ C {\bf 14} (2000) 285;
169: S.~I.~Alekhin, Phys.\ Rev.\ D {\bf 63} (2001) 094022;
170: C.~Adloff {\it et al.} [H1 Collaboration],
171: Eur.\ Phys.\ J.\ C {\bf 21} (2001) 33.
172:
173: \bibitem{Retey:2000nq}
174: A.~Retey and J.~A.~M.~Vermaseren,
175: Nucl.\ Phys.\ B {\bf 604} (2001) 281.
176: %%CITATION = HEP-PH 0007294;%%
177:
178: \bibitem{vanNeerven:2000wp}
179: W.~L.~van Neerven and A.~Vogt, Phys.\ Lett.\ B {\bf 490} (2000) 111.
180: %%CITATION = HEP-PH 0007362;%%
181:
182: \bibitem{Alekhin:2002fv}
183: S.~Alekhin, Phys.\ Rev.\ D {\bf 68} (2003) 014002.
184: %%CITATION = HEP-PH 0211096;%%
185:
186: \bibitem{Martin:2002dr}
187: A.~D.~Martin, R.~G.~Roberts, W.~J.~Stirling and R.~S.~Thorne,
188: Phys.\ Lett.\ B {\bf 531} (2002) 216.
189: %%CITATION = HEP-PH 0201127;%%
190:
191: \bibitem{Anastasiou:2003yy}
192: C.~Anastasiou, L.~Dixon, K.~Melnikov and F.~Petriello, arXiv:hep-ph/0306192.
193: %%CITATION = HEP-PH 0306192;%%
194:
195: \bibitem{H1}
196: E.~Lobodzinska, these proceedings.
197:
198: \bibitem{Hamberg:1990np}
199: R.~Hamberg, W.~L.~van Neerven and T.~Matsuura,
200: Nucl.\ Phys.\ B {\bf 359} (1991) 343
201: [Erratum-ibid.\ B {\bf 644} (2002) 403];
202: R.~V.~Harlander and W.~B.~Kilgore,
203: Phys.\ Rev.\ Lett.\ {\bf 88} (2002) 201801.
204:
205: \bibitem{Evans:2002wj}
206: H.~G.~Evans,
207: [arXiv:hep-ex/0211061].
208: %%CITATION = HEP-EX 0211061;%%
209: \end{thebibliography}
210:
211: \end{document}
212:
213:
214:
215:
216: