Dynamic imaging of diffusion by ESR

Abstract

A comprehensive analysis of the accuracy and reliability of dynamic imaging of diffusion by ESR is presented. The importance of analyzing the data in Fourier space is emphasized, and a new method which enables the determination of the important Fourier modes while also providing a test of the reliability of the measurement of the diffusion coefficient, D x , is presented. It is shown that values of D x ∼ 10 −9 cm −2 s −1 can be measured in about one hour to 10–20% accuracy, whereas for D x ∼ 10 −7 cm 2 s −1 the error should be below 1%. These statements are applicable for experiments in which there is unrestricted diffusion, and an initial Gaussian concentration profile. Systematic error resulting from non-Gaussian concentration profiles is shown to be relatively unimportant. However, the finite sweep time through a spectrum is found to yield a systematic error analogous to a Doppler shift, which tends to cancel only for the case of unrestricted diffusion. The experimental geometry in which there is diffusion from a reflective boundary requires special adjustments to align the gradient-on and -off spectra. A convenient and reliable method to accomplish this is presented. Nevertheless, there are inherently greater sources of error for this geometry, and this is confirmed by the experimental results. Utilizing the reflective boundary geometry, the longitudinal and transverse diffusion coefficients for PD-TEMPONE ( 15N labeled) in the nematic phase of MBBA at 20°C are 3.7 × 10 −7 and 2.5 × 10 −7 cm 2 s −1, respectively. Their ratio of 1.5 is consistent with results for other probe molecules in MBBA.