200 likes | 212 Views
This talk discusses QCD models and the production of inclusive jets, forward jets, and forward pions in collisions between electrons and protons. It highlights the challenges and progress in understanding the partonic structure of matter.
E N D
Forward jet and particle production at low xLeif JönssonrepresentingH1 and ZEUS Outline of the talk: ● Introduction ● QCD models ● Inclusive jet production ● Forward jet production ● Forward pion production ● Summary and conclusion
Introduction • Collisions between electrons and protons offer the ideal possibility to study the partonic structure of matter. • The big advantage is the possibility to vary the ’hardness’ of the interaction by changing the energy and virtuality of the probing photon • The dependence of the parton distributions on the energy fraction, x, and photon virtuality, Q2, is one of the most interesting and challenging problems in QCD. • At small enough x the DGLAP parton evolution description is expected to become inadequate since ln(1/x) terms in the evolution equation will become increasingly important. Significant progress has been made in the theoretical description of the parton evolution but there are still a number of problems to be solved. • Nevertheless the DGLAP scheme has been succesful in describing F2(x,Q2) down to small x and thus signatures for new parton dynamics is expected to be more visible in hadronic final states. • At low Q2 the partonic content of the photon can be probed by the interacting parton of the proton. • HERA offers the highest energies available today for such studies.
Matrix elements and higher order corrections Boson-gluon fusion: ● Dominates in the low x region ● Probes the gluon content of the proton ● Calculated in perturbation theory ● Most calculations have been performed up to NLO (DISENT, DISASTER, MEPJET, JetViP) Use evolution equations to simulate higher order emission fj = fo + fi·Pij,k ●fi and fj are the parton density functions for the mother and daughter propagator ●f0 is the input gluon density function ●Pij,k is the probability for the splitting ij,k
QCD models DGLAP (collinear approximation) ● QCD expansion by resumming terms of the type (αSlnQ2)n ●Strict ordering in virtuality of the propagators m2 >> kn2>> …>>k12 >> k02 which means strict ordering in transverse momenta of the propagators m2 >> ktn2>>… >> kt12 >> kt02 ● Good approximation at high Q2 values Resolved photons ●The photon can interact via its partonic content two DGLAP chains
QCD models contd. BFKL (kt-factorization) ● Evolution equation includes terms of the type (αSln1/x)n ● These terms become important at small x-values ● Strict ordering in the longitudinal momentum, ln(1/x), of the propagators, ln(1/x) x02 >> x12 >> ... >> xn2 >> xBj2 ●no ordering in transverse momenta CCFM ●CCFM combines in a consistent manner the properties of DGLAP and BFKL since it resums terms of both the form (αSln1/x)n and (αSln1/(1-x))n ●CCFM evolution is valid both at large and small x ●The CCFM evolution is based on angular ordering of the emitted partons: Ξ >> ξn >> ξn-1 ... >> ξo ●Virtual corrections in the gluon vertex are automatically taken into account (resummed to all orders)
QCD models contd. The Colour Dipole Model (CDM) ●Gluon emission originates from colour dipoles which radiate independently.
Single inclusive jet measurements Jet search is performed with the longitudinally invariantkT-algorithm requiring: ● Et,jet > 6 GeV ● -1 < hjet < 3 in the laboratory frame Restricted to events with: ● Q2 > 25 GeV2 and y > 0.04 The analysis is performed for two cases: ● The full phase space as defined above ● The ’dijet phase space’, cos gh < 0, hjet < 0 Data are compared to predictions from: ● LEPTO (LO QCD Matrix elements + parton showers) ●Ariadne (colour dipole model) ●DISENT (NLO matrix elements + CTEQ6 pdf)
Single inclusive jet cross sections, full phase space ● The CDM model (ARIADNE) gives a very good description of the data, both as a function of hjet and xBj. ● The LO MC model (LEPTO) exhibits slight deviations from data in the very forward rapidity region and in the small x region ● The NLO calculation (DISENT O(aS) deviates rapidly from data as one moves toward the forward region and can also not describe the small x behaviour ● RAPGAP without parton cascade gives similar results as DISENT NLO is not enough; higher order corrections are necessary
Single inclusive jet cross sections, dijet phase space ●The CDM model (ARIADNE) again gives the best description of the data ● The LO MC model (LEPTO) clearly undershoots data over the full rapidity range and in the small x region ● The NLO calculation (DISENT) does not reproduce the shape of the rapidity distribution but is in good agreement with the xBj distribution ● Notice: the scale uncertainties are very big which excludes firm conclusions
Forward jet selection ● Ptjet > 3.5 GeV (5 GeV) ● 7.0o < Θjet < 20.0o ● Suppress the DGLAP evolution by choosing the momentum transferred by the virtual photon equal to the transverse momentum of the first propagator in the ladder (forward jet; 0.5< pt2/Q2<2) ● Enhance BFKL by choosing Bjorken-x much smaller than the momentum fraction of the first propagator in the ladder (xjet>0.035 where xjet=Ejet/Ep)
’Typical’ forward jet event Experimental difficulties: ● Acceptance limitations ● Interference with the proton remnant angular cut necessary
Forward jet data ● DGLAP direct too low ● DGLAP direct + resolved and CDM in agreement with data ● CASCADE too high Pgg(z,q,kt) = aS(q2)/(1-z) + aS(kt2)/z · Dns(z,q2,kt) Fit s = dkt2dxgA(xg,kt2,q)s(g*g*qq) to data on F2(x,Q2) Fit performed to F2(x,Q2) from H1 1994
New fits of the unintegrated gluon density • ● Adjust the input pdf’s to fit data • on F2(x,Q2) from H1 and ZEUS taken • in 1994 and 1996/97, in the region • Q2 > 4.5 GeV2 and x < 5·10-3 • ● CCFM is based on angular ordering • but requires no ordering in kt of the • propagator, which means that kt can • take any kinematically allowed value • need for a cut, ktcut • ● If kt < ktcut no real emission is allowed in CASCADE and the evolution in angle is continued until kt > ktcut • ● What is the effect of the ktcut? ● JS2001 has a soft cut ktcut> 0.25 GeV, and Q0 = 1.4 GeV, where Q0 is the collinear cut for real emissions ● In the new fits ktcut = Q0 .
New fits of the unintegrated gluon density J2003 set 1: ktcut = Q0 = 1.33 GeV Only singular terms in the splitting function J2003 set 2: ktcut = Q0 = 1.18 GeV Includes also non-singular terms in the splitting function ● It is clearly seen that the behaviour of the unintegrated gluon density, in the low x region for small values of kt, depends on how the soft region, kt < ktcut, is treated.
Forward jet data contd. J2003-1: Pgg(z,q,kt) = aS(q2)/(1-z) + aS(kt2)/z · Dns(z,q2,kt) Q0 = ktcut = 1.33 GeV J2003-2: Pgg(z,q,kt) = aS(kt2)[(1-z)/z + z(1-z)/2]·Dns + aS(q2)[z/(1-z)+z(1-z)/2] Q0 = ktcut = 1.18 GeV (fits to F2(x,Q2) from H1 1994&1996/97and ZEUS 1994&1996/97)
Forward pion production Select p0 mesons via their two-photon decayusing the calorimeter Require: ● 5o < Θp < 25o ● xp = Ep/Ep > 0.01 ● p*T,p > 2.5 GeV Advantage: ● Extends the forward region to smaller angles and to smaller x Disadvantage: ● Smaller rates compared to jets ● Fragmentation effects more significant
x dependence of the forward po cross section ● DGLAP direct undershoots the data ● DGLAP direct + resolved gives the best description using m2 = Q2+4pt2 ● LO BFKL modified by KMO also reproduces the data ● CCFM too low at small x ● Agreement with forward jet data in the overlapping region
Transverse energy flow around forward po ●DGLAP direct + resolved gives the best description of the data compared to DGLAP direct and CASCADE ●The flat distribution of the data points outside the pion peak indicates that energy compensation occurs over the full phase space
Summary Inclusive jet production: ● NLO calculations (DISENT) give a reasonable description of the data except in the high h and low x regions. This is probably due to missing higher order corrections ● The LO DGLAP model (LEPTO) gives a similar description of the data with deviations at high h and low x values ● The colour dipole model (ARIADNE) gives an overall good agreement with data Forward jet production: ● NLO calculations (DISENT) are not able to reproduce the data ● The LO DGLAP model (RAPGAP DIR) does also not describe the data ● Inclusion of resolved photon contributions to the LO DGLAP model (RAPGAP DIR+RES) leads to good overall agreement ● The colour dipole model (ARIADNE) ● The CCFM model (CASCADE) gives a good description of data with the parameter settings from the fits to the improved data on F2
Summary contd. Forward pion production: ● The LO DGAP model (RAPGAP DIR) does not reproduce data ●The LO DGLAP model with resolved photons (RAPGAP DIR+RES) is in good agreement with data ● A modified LO BFKL calculation also describes the data ●The CCFM model with the splitting function restricted to singular terms and Q0 = 1.40 GeV and ktcut = 0.25 GeV undershoots data at low x but is consistent with fwd jet data in the overlapping x-region Conclusions ●Although the simple DGLAP model fails to describe data the inclusion of resolved photons leads to good agreement ●The CCFM model is able to reproduce most data but several problems still to be solved ● No clear evidence for new parton dynamics so far