Measurement Of The Absolute Drell-Yan Dimuon Cross Section In 800 Gev/C Proton-Proton And Proton-Deuterium Collisions

The Fermilab E866/NuSea Collaboration has measured the Drell-Yan dimuon cross sections in 800 GeV/c proton-proton and proton-deuterium collisions. This is the first measurement of the absolute Drell-Yan cross section in proton-proton collisions over a broad kinematic region and the most extensive study to date of the Drell-Yan cross section in proton-deuterium collisions. The Drell-Yan mechanism is sensitive to both the beam and target parton distributions, In particular, with the kinematics of the E866/NuSea data, the Drell-Yan mechanism is sensitive to the target antiquark distributions at low and intermediate Bjorken-x and to the beam quark distributions at high-x. Approximately 55K proton-proton and 121K proton-deuterium Drell-Yan events over the longitudinal momentum fraction (Feynman-x) range -0.05 < x(F) < 0.8 and the mass ranges 4.2 < Mmu+mu- < 8.7 GeV and 10.85 < M <16.85 GeV are included. The data analysis will be described, and the doubly-differential M(3)d(2)sigma/dMdx(F), and triply-differential cross sections Ed(3)sigma/dp(3) will be presented. These results will be compared with previous measurements by E605 and E772 and to predictions based upon next-to-leading order calculations utilizing the MRST2001 and CTEQ6 global parton distribution function fits. The results indicate that recent global parton distribution fits provide a good description of the light antiquark sea in the nucleon over the Bjorken-x range 0.03 < x < 0.15. In contrast, the valence quark distributions appear to be overestimated by the current parton distribution fits as x --> 1; a region in which, prior to this data, there was very little proton data to constrain the global fits.While the overall normalization is well reproduced, there are systematic deviations between the measurements and the predictions that are reflected in the large X 2 values. To elucidate these deviations, it is useful to examine the experimental cross sections separately as functions of x(1) and x(2) As noted, most of the events have x(1) much greater than x(2) which implies that the xi dependence is primarily sensitive to the valence quarks in the proton beam while the x2 dependence measures the antiquarks in the target. Figure 2 shows the ratios of the measured cross sections to NLO calculations using the MRST2001 PDFs, separately as functions of x(2) and x(1). The uncertainties in the NLO calculations from the PDF fit are also shown. he MRST2001 partons provide a good description of the x2 dependence of both the pp and pd cross sections. The CTEQ6 fits describe the x2 dependence equally well.As a function of x(1), MRST2001 and CTEQ6 have similar behavior: both overestimate the valence quarks by 15-20% at large x(1). The MRST2001 and CTEQ6 partons show essentially the same x(1) dependence, and the discrepancy between the data and current PDFs appears to be larger for the pd cross sections than the pp cross sections. Some of this discrepancy may be accounted for by radiative corrections, which are presently being studied. The pp and pd cross sections at large x, constrain two slightly different linear combinations of u(V) and d(V). The results imply that the u quark distributions in CTEQ6 and MRST2001 are overestimated as x --> 1. The Drell-Yan cross sections may also point to problems with the d/u ratio as x --> 1 [11]. However, determination of the d/u ratio at large x from these data will require a global fit to the full two-dimensional invariant cross sections together with the rest of the current world's data.It is important to recognize that this discrepancy is nearly within the quoted uncertainties on the valence quark distributions in the current PDFs [1, 12] and the +/-6.5% normalization uncertainty on the cross sections. Therefore, future global PDF fits that include these new results should provide a much better description of the E866 Drell-Yan cross sections, together with improved determinations of the antiquark distributions for 0.03 less than or equal to x < 0.15 and of the valence quark distributions for x --> 1, without significant degradation to the quality of the fit for the rest of the world's current data. This work was supported in part by the U.S. Department of Energy.