Nanocomposites result from combinations of materials with vastly different properties in the nanometer scale. These materials exhibit many unique properties such as improved thermal stability, reduced flammability, and improved mechanical properties. Many of the properties associated with polymer-clay nanocomposites are a function of the extent of exfoliation of the individual clay sheets or the quality of the nanodispersion. This work demonstrates that solid-state NMR can be used to characterize, quantitatively, the nanodispersion of variously modified montmorillonite (MMT) clays in polystyrene (PS) matrices. The direct influence of the paramagnetic Fe3+, embedded in the aluminosilicate layers of MMT, on polymer protons within about 1 nm from the clay surfaces creates relaxation sources, which, via spin diffusion, significantly shorten the overall proton longitudinal relaxation time (T-1(H),). Deoxygenated samples were used to avoid the particularly strong contribution to the T-1(H) of PS from paramagnetic molecular oxygen. We used T-1(H) as an indicator of the nanodispersion of the clay in PS. This approach correlated reasonably well with X-ray diffraction and transmission electron microscopy (TEM) data. A model for interpreting the saturation-recovery data is proposed such that two parameters relating to the dispersion can be extracted. The first parameter, f, is the fraction of the potentially available clay surface that has been transformed into polymer-clay interfaces. The second parameter, epsilon, is a relative measure of the homogeneity of the dispersion of these actual polymer-clay interfaces. Finally, a quick assay of T-1(H) is reported for samples equilibrated with atmospheric oxygen. Included are these samples as well as 28 PS/MMT nanocomposite samples prepared by extrusion. These measurements are related to the development of high-throughput characterization techniques. This approach gives qualitative indications about dispersion; however, the more time-consuming analysis, of a few deoxygenated samples from this latter set, offers significantly greater insight into the clay dispersion. A second, probably superior, rapid-analysis method, applicable to oxygen-containing samples, is also demonstrated that should yield a reasonable estimate of the f parameter. Thus, for PS/MMT nanocomposites, one has the choice of a less complete NMR assay of dispersion that is significantly faster than TEM analysis, versus a slower and more complete NMR analysis with sample times comparable to TEM, information rivaling that of TEM, and a substantial advantage that this is a bulk characterization method. (C) 2003 Wiley Periodicals, Inc.