OPTIMIZATION OF CHEMICAL-SHIFT-BASED POLARIZATION GRADIENTS IN H-1-NMR SPIN-DIFFUSION EXPERIMENTS ON POLYMER BLENDS WITH CHEMICALLY SIMILAR CONSTITUENTS

Proton CRAMPS (combined rotation and multiple-pulse spectroscopy) spectra of glassy partially aromatic polymers seldom display more than two bands (aromatic and aliphatic). This paper addresses the question of how, using the few spectral differences in the CRAMPS spectra of two polymers, one can optimize the preparation of a chemical-shift-based polarization gradient in a proton spin-diffusion experiment designed to probe domain size and stoichiometry in a blend of these materials. The criterion adopted for optimizing the gradient is that maximal changes in lineshape, rather than intensity, shoulld accompany the spin-diffusion process. Large lineshape changes are associated with those initial conditions where, in the absence of spin diffusion between homopolymers but after spin diffusion within each monomer, the signs of polarizations associated with each polymer are opposite. This condition often forces one to work with small spin-diffusion integrals relative to the Boltzmann-equilibrium signal. Model calculations are used to illustrate the method and selected data for a blend of two dominantly aromatic polymers are presented. The dissipation of the initially imposed aromatic/aliphatic polarization gradient can be followed in the blend over 2.5 orders of magnitude, and this is enough to obtain information on both the size of domains and their stoichiometry, although extraction of the latter information requires further assumptions. The sensitivity of this method is also considered in the case where the contrast in lineshapes between phases diminishes. Also examined are other phenomena encountered when one works with small signals such as magnetic-susceptibility and -field inhomogeneity effects and background signals. Finally, a comparison is made of the relative merits of using a proton versus a 13C spectral readout in these experiments. © 1992.