N.m.r. of solid polymers: a review

Abstract

The application of nuclear magnetic resonance (n.m.r.) to the study of solid polymers has developed significantly in recent years. The emergence of more refined mathematical procedures for data interpretation, improvements in spectrometer design, the availability of well characterized samples and, most important, the introduction of the rotating frame experiment have been largely responsible. The rotating frame relaxation time, Tip, is sensitive to motions with correlation frequencies, vc, in the kHz range whereas spin-lattice relaxation times, T1, respond to much more rapid motions in the MHz region. As a result, Tip measurements extend the dynamic range of molecular motions, sensitive to n.m.r., to the ultraslow region. The use of fibre materials for n.m.r, investigation has enhanced the value of the technique considerably because of the additional information contained in the anisotropy of the n.m.r, data, recorded as a function of fibre orientation in the laboratory magnetic field. Theoretical predictions, based upon plausible molecular models, are tested more stringently when compared with anisotropic fibre data. These magnitude comparisons have been most successful for spin-spin relaxation time, T~, measurements. Second, M2, and fourth, M4, moment measurements of the resonance absorption envelope have been comparably informative. Comparison with 71 and Tip data, on the other hand, is often severely hampered by spin diffusion and correlation frequency distribution effects; theories are usually based upon the description of molecular motion in terms of a single correlation frequency which results in predicted magnitudes which are low compared with experimental values. The translation of raw n.m.r, data into molecular information is not, as yet, totally unambiguous. Complications principally arise from the non-exponentiai character of the various magnetization decays which are often observed. For example, at low temperatures, the decays are more Gaussian than Lorentzian. Two and sometimes three discrete components may be observed which are usually ascribed to morphological inhomogeneities in the polymer in which the resonant nuclei experience appreciably different molecular environments. Typically, nuclei in the amorphous regions of a semi-crystalline polymer which is above its glass transition temperature, Tg, are undergoing rapid motions, evidenced by a long 72 value (or narrow line in the absorption spectrum) while the crystalline regions may still be well below their melting point for which a much shorter T2 (or broad line) is appropriate. Non-exponential decays interpreted in this way have been used to determine the crystallinity of several polymers by measuring the relative intensities of the T2 or second moment components i-5. Such measurements in fact provide the 'mobile fraction' which has obvious temperature dependence. To equate this measurement, in a general way, with crystaUinity is clearly inappropriate. Rather, the mobile fraction must be viewed as complementing the wider range of experimental estimates available 6. There has been the suggestion that the intensities of component Tip decays may be used as a measure of crystallinity 7. As discussed in detail in a subsequent section, both the magnitudes and intensities of Tip components, attributed to different regions in the polymer, may be affected appreciably by spin diffusion s . In such cases the relative intensities in no way reflect the amounts of material contributing to each component and therefore may not be used as an estimate of either mobile fraction or crystallinity. Less obvious composite spectra have been analysed rigorously in an attempt to separate out the overlapping components 9-ii. Theoretical lineshapes have been synthesized from up to four components for comparison with experimental traces from drawn polyethylene (PE) 11. Similar studies have been undertaken on the spectra from polymers rotated at the magic angle i2, i3 It is important to realize that other reasons may be responsible for non-exponential behaviour. Molecular weight distributions may be responsible or indeed relaxations which are dominated by defect diffusive mechanisms will lead to 71 and Tip decays which are non-exponential. While specific motions in polymers such as side group or main chain re-orientations, main chain translations or motions of a more general nature (glass transition phenomena) may be readily detected and identified by n.m.r, methods, precise information about the motional mechanism is generally not forthcoming. T2, for example, is insensitive to motions below ,,~ 104Hz, while changes in T2 due to the onset of rotation about an axis at frequencies in excess of 104Hz cannot distinguish between classical rotation and rotation over a threefold or higher potential. In some cases combined /1, Tip, and Tz measurements can be helpful as, for example, for methyl group motion in poly(vinyl acetate) (PVAc) where quantum mechanical tunnelling is indicated i4. The fact that the three relaxation times sample widely different frequency ranges in the spectrum of molecular motions allows activation energies, AE, to be determined. /'1 and Tip minima and Tz transitions are translated