Leading nucleon production at HERA and the structure of the pion

Experimental results from HERA on the production of leading neutrons in neutral current e+p collisions are reviewed. The data cover a large kinematic range from photoproduction to deep inelastic scattering. The neutrons are produced with low transverse momentum (pT<0.7 GeV) and carry a large fraction of the incoming proton's energy (xL>0.2). Neutron production is studied relative to the inclusive cross section, reducing considerably the systematic error. The rate of neutrons in photoproduction is about half that expected from hadroproduction experiments. The photoproduction and hadroproduction data are otherwise in agreement and broadly in accord with particle-exchange models. There is a 20-30% rise in the rate of neutrons between photoproduction and deep inelastic scattering attributable to absorptive rescattering in the γ*p system. Aside from absorptive effects, the rate of neutron production depends only logarithmically on Q2, the virtuality of the exchanged photon, and x, the momentum fraction of the incoming proton carried by the struck quark. Factorization breaking is observed: for 0.64<xL<0.82 the rate of neutrons is almost independent of the leptonic variables; however, at low and high xL there is a weak dependence. Over the whole kinematic range from photoproduction to deep inelastic scattering, the structure function for leading neutron production is approximately given by F2LN(3)(x, Q2, xL), = A(Q2)F2(x, Q2)f(xL), where F2 is the structure function of the proton, f(xL) is the flux of neutrons and A(Q2) is an absorptive factor. The total γ π cross section in photoproduction, σ(γ π), and the structure function of pion in deep inelastic scattering, F2π, can be extracted from the data on the assumption that pion exchange is the dominant mechanism for the production of leading neutrons. An effective flux method is introduced which avoids many theoretical uncertainties. σ(γ π) is found to be approximately one third σ(γ p), violating quark counting rules which predict two thirds. The x and Q2 dependence of F2π is similar to F2. (author)