The large diffusion coefficients of gases result in significant spin motion during the application of gradient pulses that typically last a few milliseconds in most NMR experiments. In restricted environments, such as the lung, this rapid gas diffusion can lead to violations of the narrow pulse approximation, a basic assumption of the standard Stejskal-Tanner NMR method of diffusion measurement. We therefore investigated the effect of a common, biologically inert buffer gas, sulfur hexafluoride (SF(6)), on (129)Xe NMR and diffusion. We found that the contribution of SF(6) to (129)Xe T(1) relaxation in a 1:1 xenon/oxygen mixture is negligible up to 2 bar of SF(6) at standard temperature. We also measured the contribution of SF(6) gas to (129)Xe T(2) relaxation, and found it to scale inversely with pressure, with this contribution approximately equal to 1 s for 1 bar SF(6) pressure and standard temperature. Finally, we found the coefficient of (129)Xe diffusion through SF(6) to be approximately 4.6 x 10(-6) m(2)s(-1) for 1 bar pressure of SF(6) and standard temperature, which is only 1.2 times smaller than the (129)Xe self diffusion coefficient for 1 bar (129)Xe pressure and standard temperature. From these measurements we conclude that SF(6) will not sufficiently reduce (129)Xe diffusion to allow accurate surface-area/volume ratio measurements in human alveoli using time-dependent gas diffusion NMR.