1: \subsection{\boldmath Jets at small-$x$}
2: \label{sec:xjets}
3:
4: \textit{Main authors L.~Jönsson and A.~Knutsson}\\
5:
6: In the region of low $x$-values the interacting parton frequently
7: produces a cascade of emissions before it interacts with the virtual
8: photon. Due to the strong ordering in virtuality, the emissions of
9: the DGLAP evolution are very soft close to the proton direction,
10: whereas BFKL emissions can produce large transverse momenta in this
11: region. Thus, deviations from the DGLAP parton evolution scheme are
12: expected to be most visible in a region close to the direction of
13: the proton beam.
14:
15: HERA has extended the available region in the Bjorken scaling
16: variable, $\xbj$, down to values of $\xbj \simeq 10^{-4} $, for
17: values of the four momentum transfer squared, $Q^2$, larger than a
18: few GeV$^2$, where perturbative calculations in QCD are expected to
19: be valid.
20:
21: A measurement of the forward jet production cross section at small
22: $\xbj$, as proposed by Mueller and
23: Navelet~\cite{Mueller:1990er,Mueller:1990gy,Mueller:1986ey}, has
24: long been regarded as the most promising test of perturbative parton
25: dynamics. The idea is to select events with a jet close to the proton
26: direction having the virtuality of the propagator closest to the
27: proton approximately equal to the virtuality of the exchanged photon.
28: This will suppress an evolution with strong ordering in virtuality as
29: is the case in the DGLAP evolution. The additional requirement that
30: the forward jet takes a large fraction of the proton momentum,
31: $x_{jet} = E_{jet}/E_p$, such that $x_{jet} \gg x_{Bj}$ opens up for
32: an evolution where the propagators are strongly ordered in the
33: longitudinal momentum fraction like in the BFKL scheme.
34: Experimentally this is realized by demanding the squared transverse
35: momentum of the forward jet to be of the same order as $Q^2$ and
36: $x_{jet}$ to be larger than a preselected value which still gives
37: reasonable statistics. More exclusive final states, like those
38: containing a di-jet system in addition to the forward jet (called
39: `2+forward jet'), provide an additional handle to control the parton
40: dynamics.
41:
42: \paragraph{Production of forward jets in DIS}
43:
44:
45: The H1 experiment has measured the forward jet cross
46: section~\cite{Aktas:2005up}
47: using data collected in 1997, comprising an
48: integrated luminosity of 13.7 pb$^{-1}$.
49: The proton energy is 820~GeV and the positron energy is
50: 27.6~GeV which correspond to a center-of-mass-energy of
51: $\sqrt{s}\approx$300~GeV.
52:
53: DIS events are obtained by applying the cuts $E_{e'} > 10$~GeV, $ 156
54: \dg < \theta_e < 175\dg$, $0.1 < y < 0.7$ and $5$~GeV$^2 < Q^2 <
55: 85$~GeV$^2$, where $E'_e$ is the energy of the scattered electron,
56: $\theta_e$ the polar angle, and $y$ is the inelasticity of the
57: exchanged photon. Jets are defined using the inclusive $k_t$-jet
58: algorithm~\cite{Catani:1993hr, Catani:1992zp} applied in the
59: Breit-frame. A forward jet is defined in the laboratory system as having
60: $p_{t,jet} > 3.5 $~GeV and being in the angular range $7\dg <
61: \theta_{jet} < 20\dg$. In order to enhance BFKL evolution it is
62: required that $x_{jet} > 0.035$ whereas DGLAP evolution was suppressed
63: in the single differential cross section measurement by introducing
64: the requirement $0.5 < p_t^2/Q^2 < 5$.
65:
66: Another event
67: sample, called the '2+forward jet' sample, is selected by requiring
68: that, in addition to the forward jet, at least two more jets are found,
69: all of them having $p_{t,jet}$ larger than 6~GeV. In
70: this scenario the $p_t^2/Q^2$-cut is not applied, due to the limited
71: statistics.
72:
73: The forward jet cross sections for single and triple differential cross sections
74: are compared to LO ($\alpha_s$) and NLO
75: ($\alpha_s^2$) calculations of direct photon interactions as obtained
76: from the DISENT program. Comparisons of the inclusive forward jet
77: cross sections with the DISENT predictions for a di-jet final state
78: are adequate, since the forward jet events always contain at least
79: one additional jet due to the kinematics. The renormalization scale
80: $(\mu_r^2)$ is given by the average $p_t^2$ of the di-jets from the
81: hard scattering process, while the factorization scale $(\mu_f^2)$ is given by
82: the average $p_t^2$ of all forward jets in the selected sample.
83:
84: In the analysis of events with two jets in addition to the
85: forward jet, the measured cross sections are compared to the predictions of
86: NLOJET++. This program provides perturbative calculations of cross
87: sections for three-jet production in DIS at NLO ($\alpha^3_s$) accuracy.
88: In this case the scales $\mu_r=\mu_f$ are set to the
89: average $p_t^2$ of the three selected jets in the calculated event.
90:
91: The NLO calculations by DISENT~\cite{Catani:1996jh, Catani:1996vz} and
92: NLOJET++~\cite{Nagy:2001xb} are performed using the
93: CTEQ6M~\cite{Pumplin:2002vw} parameterization of the parton distributions in the proton.
94:
95: \paragraph{Single Differential Cross Section}
96: \label{sec:inclusive}
97:
98: The measured single differential forward jet cross sections on hadron
99: level are compared with LO ($\alpha_s$) and NLO ($\alpha^2_s$)
100: calculations from DISENT in Fig.~\ref{xfj_djcorr}a. In
101: Fig.~\ref{xfj_djcorr}b and c the data are compared to the various QCD
102: models.
103: \begin{figure}[htb]
104: \begin{center}
105: \vspace*{1mm}
106: \vspace*{1cm}
107: \epsfig{figure=fig/fig3a.eps,height=0.24\textwidth}
108: \epsfig{figure=fig/fig3b.eps,height=0.24\textwidth}
109: \epsfig{figure=fig/fig3c.eps,height=0.24\textwidth}
110: \caption{
111: The hadron level cross section for forward jet production
112: as a
113: function of $\xbj$ as measured by H1~\protect\cite{Aktas:2005up}
114: compared to NLO predictions from DISENT (a)
115: and to QCD Monte Carlo models (b and c). The shaded band around
116: the data points shows the error from the uncertainties in the
117: energy scales of the liquid argon calorimeter and the SpaCal
118: electromagnetic calorimeter. The hatched band around the NLO
119: calculations illustrates the theoretical uncertainties in the
120: calculations, estimated as described in the text. The dashed
121: line in (a) shows the LO contribution.
122: \label{xfj_djcorr}}
123: \end{center}
124: \end{figure}
125:
126: In Fig.~\ref{xfj_djcorr}a it can be observed that, at small $x_{Bj}$,
127: the NLO di-jet calculations from DISENT are significantly larger than
128: the LO contribution. This reflects the fact that the contribution
129: from forward jets in the LO scenario is suppressed by kinematics. For
130: small $x_{Bj}$ the NLO contribution is an order of magnitude larger
131: than the LO contribution. The NLO contribution opens up the phase
132: space for forward jets and improves the description of the data
133: considerably. However, the NLO di-jet predictions are still a factor
134: of 2 below the data at low $\xbj$. The somewhat improved agreement at
135: higher $x_{Bj}$ can be understood from the fact that the range in
136: the longitudinal momentum fraction which is available for higher
137: order emissions decreases.
138:
139: From Fig.~\ref{xfj_djcorr}b it is seen that the CCFM model (both set-1
140: and set-2) predicts a somewhat harder $x_{Bj}$ distribution, which
141: results in a comparatively poor description of the data.
142:
143: Fig.~\ref{xfj_djcorr}c shows that the DGLAP model
144: with direct photon interactions alone (RG-DIR) gives results similar
145: to the NLO di-jet calculations and falls below the data, particularly
146: in the low $x_{Bj}$ region. The description of the data by the
147: DGLAP model is significantly improved if contributions from resolved
148: virtual photon interactions are included (RG-DIR+RES). However, there
149: is still a discrepancy in the lowest $x_{Bj}$-bin, where a possible
150: BFKL signal would be expected to show up most prominently. The CDM
151: model, which gives emissions that are non-ordered in transverse
152: momentum, shows a behavior similar to the RG DIR+RES model.
153:
154:
155: \paragraph{Events with Reconstructed Di-jets in Addition to the Forward Jet}
156:
157:
158: By requiring the reconstruction of the two hardest
159: jets in the event in addition to the forward jet, different
160: kinematic regions can be investigated by applying cuts on the jet momenta and their
161: rapidity separation.
162:
163: In this scenario it is demanded that all jets have transverse momenta
164: larger than 6~GeV. By applying the same $p_{t,jet}$ cut to all three
165: jets, evolution with strong $k_{t}$-ordering is not favored.
166: The jets are ordered in rapidity according to $\eta_{fwdjet} >
167: \eta_{jet_2} > \eta_{jet_1} > \eta_e$ with $\eta_e$ being the
168: rapidity of the scattered electron. The
169: cross section is measured by H1~\cite{Aktas:2005up}
170: in two intervals of $\Delta \eta_1 =
171: \eta_{jet_2} - \eta_{jet_1} $. If
172: the di-jet system originates from the quarks $q_1$ and $q_2$
173: (see Fig.~\ref{fwdjet-dijet}), the phase space for evolution in $x$ between
174: the di-jet system and the forward jet is increased by requiring that
175: $\Delta\eta_1$ is small and that
176: $\Delta\eta_2 = \eta_{fwdjet} - \eta_{jet_2}$ is large. $\Delta\eta_1 < 1$
177: favors small invariant masses of the di-jet system and thereby small values
178: of $x_g$ (see Fig.~\ref{fwdjet-dijet}). With $\Delta\eta_2$ large,
179: $x_g$ carries only a small fraction
180: of the total propagating momentum, leaving the rest for additional radiation.
181:
182: The directions of the other jets are related to the forward
183: jet through the $\Delta\eta$ requirements. When $\Delta\eta_2$ is small, it is
184: therefore possible that one or both of the additional jets originate
185: from gluon radiation close in rapidity space to the forward jet.
186: With $\Delta\eta_1$ large, BFKL-like evolution may then occur between the
187: two jets from the di-jet system, or, with both $\Delta\eta_1$ and $\Delta\eta_2$ small, even between
188: the di-jet system and the hard scattering vertex. By studying the cross
189: section for different $\Delta\eta$ values one can test theory and models
190: for event topologies where the $\kt$ ordering is broken at varying
191: locations along the evolution chain.
192:
193: \begin{figure}[htb]
194: \begin{center}
195: \vspace*{1mm}
196: \vspace*{1cm}
197: \epsfig{figure=fig/fig7.eps,height=5cm}
198: \caption{
199: A schematic diagram of an event giving a forward jet and two
200: additional hard jets. These may stem from the quarks ($q_1$ and
201: $q_2$) in the hard scattering vertex or gluons in the parton
202: ladder. $x_g$ is the longitudinal momentum fraction carried by
203: the gluon, connecting to the hard di-jet system (in this case
204: $q_1$ and $q_2$) .
205: \label{fwdjet-dijet}}
206: \end{center}
207: \end{figure}
208:
209: In this investigation the same settings of the QCD models are used as
210: in sections~\ref{sec:inclusive}, while the NLO three-jet
211: cross sections are calculated using NLOJET++.
212:
213: From Fig.~\ref{2+fwdnlo} it is observed that NLO three-jet gives good
214: agreement with the data if the two additional hard jets are emitted
215: in the central region ($\Delta\eta_2$ large). It is interesting to
216: note that a fixed order calculation ($\alpha_s^3$), including the
217: $\log (1/x)$-term to the first order in $\alpha_s$, is able to
218: describe these data well. However, the more the additional hard jets
219: are shifted to the forward region ($\Delta\eta_2$ small), the less
220: well are the data described by NLO three-jet. A possible
221: explanation is that the more forward the additional jets go, the
222: higher the probability is that one of them, or even both, do not
223: actually originate from quarks but from additional radiated gluons.
224: NLO three-jet calculates the NLO contribution to final states containing
225: one forward jet and two jets from the di-quarks, i.e. it accounts for
226: the emission of one gluon in addition to the three jets. Since the
227: radiated gluon is predominantly soft it has a small probability to
228: produce a jet that fulfills the transverse momentum requirement
229: applied in this analysis. This results in a depletion of the
230: theoretical cross section in the small $\Delta\eta_2$ region, which
231: is more pronounced when $\Delta\eta_1$ is also small, i.e. when all
232: three jets are in the forward region. Consequently a significant
233: deviation between data and NLOJET++ can be observed for such events
234: (see the lowest bin in Fig.~\ref{2+fwdnlo}b). Accounting for still
235: higher orders in $\alpha_s$ might improve the description of the data
236: in this domain, since an increased number of gluon emissions would
237: enhance the probability that one of the radiated gluons produces a
238: jet which is above the threshold on the transverse momentum.
239:
240:
241: As explained above, evolution with strong $\kt$-ordering is
242: disfavored in this study. Radiation that is non-ordered in $\kt$
243: may occur at different locations along the evolution chain,
244: depending on the values of $\Delta\eta_1$ and $\Delta\eta_2$. In a
245: comparsion to QCD models (these figures are not shown, for details
246: see~\cite{Aktas:2005up}) the following observations
247: where made. The colour dipole model gives good agreement in all
248: cases, whereas the LO DGLAP models give cross sections that are too
249: low except when both $\Delta\eta_1$ and $\Delta\eta_2$ are large. For
250: this last topology all models and the NLO calculation agree with the
251: data, indicating that the available phase space is exhausted and that
252: little freedom is left for dynamical variations.
253:
254: Furthermore it was seen that the `2+forward jet' sample
255: differentiates between the CDM and the DGLAP-resolved model, in contrast to the
256: more inclusive samples where CDM and RG-DIR+RES give the same
257: predictions. The conclusion is that additional breaking of
258: the $\kt$ ordering is needed compared to what is included in the
259: resolved photon model (see Ref.~\cite{Aktas:2005up}).
260:
261: \begin{figure}[htb]
262: \begin{center}
263: \vspace*{1mm}
264: \vspace*{1cm}
265: \epsfig{figure=fig/fig8a.eps,width=0.24\textwidth}
266: \epsfig{figure=fig/fig8b.eps,width=0.24\textwidth}
267: \epsfig{figure=fig/fig8c.eps,width=0.24\textwidth}
268: \caption{
269: The cross section for events with a reconstructed high
270: transverse momentum di-jet system and a forward jet as a
271: function of the rapidity separation between the forward jet and
272: the most forward-going additional jet, $\Delta\eta_2$ as measured by
273: H1~\protect\cite{Aktas:2005up}.
274: Results
275: are shown for the full sample and for two ranges of the
276: separation between the two additional jets, $\Delta\eta_1<1$ and
277: $\Delta\eta_1>1$. The data are compared to the predictions of a
278: three-jet NLO calculation from NLOJET++. The band around the
279: data points illustrates the error due to the uncertainties in
280: the calorimetric energy scales. The band around the NLO
281: calculations illustrates the theoretical uncertainties in the
282: calculations.
283: \label{2+fwdnlo}}
284: \end{center}
285: \end{figure}
286:
287:
288: % $Revision: 1.14 $
289: % Local Variables:
290: % mode: LaTeX
291: % TeX-master: "summary04"
292: % End:
293: