The Nancy Grace Roman Space Telescope will have a wide field imaging instrument, the WFI, with detectors sensitive to ~2.5 microns but no filter beyond 2 microns. The telescope will be passively cooled to 260K which results in the thermal background dominating the noise at the long wavelength end of the observable wavelength range. A carefully chosen bandpass centered at 2.1 microns with R~6-10 will however, enable surveys that will run at least ten times faster than any ground-based survey to similar depth at corresponding wavelengths. Furthermore, it will have a factor of ~2 better spatial resolution and higher Strehl ratios than a ground-level adaptive optics (GLAO) system in typical conditions. The potential depths achievable with this wavelength coverage over wide areas, compared to Euclid and Subaru/ULTIMATE will reveal pristine stellar populations at the end of the cosmic dark ages at z~17, and characterize the earliest epochs of dust production by measuring the rest-frame ultraviolet slopes of galaxies at z~15. This will allow for timely spectroscopic follow-up with the James Webb Space Telescope. At lower redshifts, by sampling longer wavelengths where the nebular emission increasingly dominates over the stellar photospheric emission, it will provide better constraints on photometric redshifts for samples of galaxies used for weak lensing, and provide a bandpass with a better-sampled point spread function and smaller spectral variations. This will allow us to independently measure weak lensing shear more precisely, with improved galactic shape noise at z > 1. Galactic surveys will also benefit from a Ks bandpass due to the reduced dust extinction and the improved sensitivity to cooler stars as discussed in Stauffer et al. (2018; arXiv:1806.00554).