Some perspectives on the interpretation of proton NMR spin diffusion data in terms of polymer morphologies

Proton spin diffusion data yield morphological information over dimensions covering approximately the 2-50 nm range. In this article, the interpretation of such data for polymers is emphasized, recognizing that the mathematical framework for much of this interpretation already exists in the literature. Practical issues are considered, for example, a useful scaling of plotted data is suggested, key attributes of the data are identified and ambiguities in the mapping of data into morphological models are spelled out. Discussion is limited to two-phase systems, where it is assumed that, by employing multiple-pulse methods polarization gradients can be generated, whose spunal sharpness is limited solety by the morphological definition of the interfaces. Interpretation of data in terms of morphology and stoichiometry is emphasized, where stoichiometric issues pertain only to chemically heterogeneous systems. Extraction of stoichiometric information from spin diffusion data is not commonly attempted; the discussion included herein allows for the possibility that the composition of phases may be chemically mixed. Methods for generating gradients are discussed only briefly. A standardized spin diffusion plot is proposed and the initial slope of this plot is tocussed on for providing information about morphology and stoichiometry. Ambiguities of interpretation considered include the dimensionality of the deduced morphology and, for systems with chemical heterogeneity the uniqueness of the compositional characterization of each phase. In addition, funite difference methods are used to simulate entire spin diffusion curves for idealized lamellar and hexagonal rod/matrix morphologies. Comparisons of these curves show that distinguishing 1-D and 2-D morphologies on the basis of experimental data is unlikely to be successful over the range of stoichiometrics where such morphologies are expected. Several examples of spin diffusion data are presented. Brief treatments of the following topics are included: finite interface width, estimation of spin diffusion constants, and incorporation of longitudinal relaxation effects. Finally, a short experimental discussion on the preparation of polarization gradients is given including those preparations which make use of differences in the multiple-pulse relaxation time, T1xz. It is noted that T1xz decays may be strongly perturbed in the presence of magic angle spinning, therefore, strategies are also outlined for minimizing these effects.