A still unsettled issue in the nucleon spin decomposition problem: On the role of surface terms and gluon topology

In almost all the past analyses of the decomposition of the nucleon spin into its constituents, surface terms are simply assumed to vanish and not to affect the integrated sum rule of the nucleon spin. However, several authors claim that neglect of surface terms is not necessarily justified, especially owing to the possible nontrivial topological configuration of the gluon field in the QCD vacuum. There also exist some arguments indicating that the nontrivial gluon topology would bring about a delta-function type singularity at zero Bjorken variable into the longitudinally polarized gluon distribution function, thereby invalidating a naive partonic sum rule for the total nucleon spin. In the present paper, we carefully examine the role of surface terms in the nucleon spin decomposition problem. We shall argue that surface terms do not prevent us from obtaining a physically meaningful decomposition of the nucleon spin. In particular, we demonstrate that nontrivial topology of the gluon field would not bring about a delta-function type singularity into the longitudinally polarized gluon distribution functions. We also make some critical comments on the recent analyses of the role of surface terms in the density level decomposition of the total nucleon angular momentum as well as that of the total photon angular momentum.