The native, oxidation layer of Nb is complex and includes magnetic and conducting sub-oxides that make it unsuitable as a tunnel barrier in junction devices. Recently it has been shown in various experiments that the Nb oxide layer also exhibits significant RF losses that limit the performance of 2-d and 3-d qubits. For tunnel junctions the oxide problem has been mitigated using an ultra-thin Al capping layer on the Nb that prevents the formation of Nb oxide. Tunnel junctions on the superconducting/normal metal (S/N) bilayers are fully described by the Arnold model of the proximity effect. This capping layer approach is now being tested in 2-d transmon qubits that utilize large area Nb thin film capacitors. Significant improvements in T1 decay time are found using Nb layers capped by Al, Ta and TiN. The Arnold model will be presented and the impact of relevant parameters such as the quasiparticle scattering in the N layer will be discussed. An important result is that the N layer need not be superconducting in bulk as the effective gap for quasiparticle excitations is the Andreev bound state E0 ~ ΔS formed in the superconducting potential well ΔS – ΔN. This opens the door to using capping layers that have little or no oxide layer.