Structure Functions and Parton Distributions

Much of the predictive content of perturbative QCD (pQCD) treatment of the hadronic hard scattering is contained in factorization theorems [ 1]. Their purpose is to separate longfrom short-distance effects in scattering amplitudes. They supply perturbatively uncalculable long-distance effects with physics content in terms of well-defined matrix elements, which allows them to be measured experimentally or by numerical simulation. They also define the normalization of short-distance factors, which allows them to be calculated perturbatively. Predictions follow when processes with different hard scatterings but the same nonperturbative matrix elements are compared. Thus, quark and gluon distributions measured in deep inelastic scattering may be used to normalize the Drell-Yan or jet cross section. Deep inelastic lepton scattering has long been regarded as the cleanest probe of constituent substructure. Considerable interest, therefore, greeted the observation [ 2] by European Muon Collaboration (EMC) that the structure function F2(xB, Q ) of an iron nucleus differs in significant ways as a function of xB from that for deuterium. It was the discovery of the EMC effect that opened a door for systematic study of QCD dynamics in a nuclear environment, which has led to many new QCD phenomena, e.g., shadowing, saturation, and color glass condensate. In this talk, I review our abilities and limitations to generalize the pQCD factorization theorems to the hard scattering involving nuclei. Nuclear parton distributions are heavily used in phenomenological description of hard processes in heavy ion reactions at SPS and RHIC energies and in calculating predictions at the LHC energies. According to the factorization theorems, nuclear dependence of parton distributions should be universal or process independent. I discuss the constraints on nuclear parton distributions from existing data and the global QCD analysis.