Deep inelastic scattering in the dipole picture at next-to-leading order

This thesis studies gluon saturation in hadronic matter at high energy by calculating next-to-leading order (NLO) corrections to inclusive and diffractive deep inelastic scattering cross sections in the Color Glass Condensate (CGC) effective field theory. We demonstrate that the large soft gluon logarithm is correctly factorized into the Balitsky–Kovchegov (BK) renormalization group equation by accurately connecting the NLO scattering kinematics to the rapidity scale of the dipole amplitude in the scattering. This brings the perturbative expansion under control and enables us to do precision comparisons between theory and data. We fit the initial condition of the BK evolution equation to HERA inclusive deep inelastic scattering data by combining of the NLO accuracy inclusive cross sections with beyond leading order BK evolution prescriptions. This results in the state-of-the- art accuracy comparison between CGC theory and HERA data, and determination of the dipole amplitude initial shape which is a necessary input for all NLO CGC phenomenology. In the introductory part of this thesis, the effect of the NLO BK equation on the fits is assessed, and an alternative form for the NLO loop correction to the inclusive cross sections is derived which enables the consistent setting of the dipole amplitude rapidity scale in the NLO corrections. The underlying mechanism of diffraction in particle scattering is still unknown, with multiple competing pictures. Diffraction is studied in this thesis in the CGC formalism, and we calculate the tree-level qqg NLO contribution to the diffractive deep inelastic scattering structure functions where the qqg Fock state scatters off the target and becomes the diffractively produced system. This contribution has previously been known in the literature only in leading log(Q2) accuracy valid at large Q2, and only for the structure function FD T . The qqg contribution to both structure functions FD T and FD L are presented in full NLO accuracy.