Recent measurement of ∆G/G at COMPASS

We present a preliminary measurement of the gluon polarization ∆G/G in the nucleon, based on the spin asymmetry of quasi-real photoproduction events for which a pair of large transverse momentum hadrons is produced. The data were obtained by the COMPASS experiment at CERN using a 160 GeV polarized muon beam scattered on a large polarized 6LiD target. The preliminary helicity asymmetry for the selected events is A‖/D = 0.002± 0.019(stat.)± 0.003(exp.syst.). From this value, a leading order analysis based on the model of PYTHIA leads to the gluon polarization in the nucleon ∆G/G(xg = 0.095,μ2 = 3 GeV2) = 0.024±0.089(stat.)± 0.057(syst.). This value is consistent with parameterizations obtained from QCD fits to the g1 data, with a first moment ∆G ≡ ∫ 1 0 ∆G(x)dx . 1, at the same scale. The decomposition of the nucleon spin in the contributions from its constituents has been a central topic of investigation in polarized lepton-nucleon scattering in the last 20 years. The EMC measurement of the proton spin structure [1] has shown that only 20 to 30 % of the proton spin could be attributed to the total quark spin ∆Σ, in contrast to the 60 % expected in the naive quark-parton model. In inclusive lepton-nucleon scattering the contribution of the gluon spin ∆G to the nucleon spin can only be measured indirectly by studying the Q2 dependence of the polarized spin-structure functions in QCD. Fits at next-to-leading order (NLO) provide evaluations of ∆G which are of the order of 0.50. The precision of these fits is strongly limited by the small Q2 range covered by the data at any value of x, a situation resulting from the lack of a polarized lepton-nucleon collider. In addition, the shape of ∆G(x) at a fixed Q2 used as reference, has to be provided as an input parameterization. It varies considerably between the different analyses and is only poorly constrained by the results of the fits. A direct measurement of the gluon polarization ∆G(x)/G(x) can be obtained from the helicity asymmetry of the photongluon fusion (PGF, γ∗g → qq̄) cross-section, which constitutes an important part of the experimental program of COMPASS The COMPASS experiment [2] is located at the M2 beam line of the CERN SPS, which provides a 160 GeV μ+ beam, with a natural polarization of −76±5%. The target consists in an upstream and a downstream cell, longitudinally polarized in opposite directions. Typical target polarizations of 50.0±2.5% are obtained. The forward spectrometer is divided in two stages allowing the reconstruction of the scattered muon and of the produced hadrons in broad momentum and angular ranges. The trigger system provides efficient tagging down to Q2 = 0.002 GeV2. The present analysis focuses on the data collected during the 2002 and 2003 runs. We only consider quasi-real photoproduction events (Q2 < 1 GeV2), in which at least two charged hadrons are associated to the primary vertex in addition to the incident and the scattered muons. The fraction of PGF is enhanced by asking the two leading hadrons to have a large transverse momentum: ph1 T > 0.7 GeV, p h2 T > 0.7 GeV and (ph1 T ) 2 +(ph2 T ) 2 > 2.5 GeV2. In total, around 350,000 events are selected. For this high pT sample, the measured helicity asymmetry (defined as in [3]) is