We investigated the glass former triphenyl phosphite by calorimetry, Brillouin scattering, and dielectric and nuclear magnetic resonance spectroscopy. Time-resolved experiments demonstrate the conversion between three distinctly different phases, namely, the supercooled liquid, a recently discovered apparently amorphous phase (glacial phase), and the crystalline phase. The temperature dependencies of the properties provided by the different methods are compared. We find significant molecular motion in the glacial phase. From this we hypothesize that the glacial phase is a different highly viscous liquid or a solid phase with some kind of cooperative relaxation.

We investigated binary low-molecular-weight glass formers as model systems for mixtures of small and large molecules. Tricresyl phosphate (TCP) in oligomeric styrenes (OS), benzene in OS and polystyrene (PS), and benzene in TCP were studied by applying dielectric spectroscopy as well as 1H, 2H, and 31P NMR spectroscopy. Temperatures above and below the glass-transition temperature (T G) are covered. The dielectric loss of the small component appears broader the higher the molecular ratio M/m is, and the lower the TCP concentration and the lower the temperature are chosen. Close to T G, extremely broad distributions of correlation times G(log τ) result, which are similar to those reported in the cases of polymer−plasticizer systems, although for our systems the motional heterogeneities are already established at similar M and m. By applying 1H and 2H (1D and 2D) NMR on benzene in OS and PS, we can demonstrate that the large molecules basically behave as in neat glass formers. However, the small molecules exhibit an isotropic reorientation also well below T G, and the dynamics is rather characterized by a random jump process than by rotational diffusion, the latter being found in neat systems. Furthermore, we can prove that within G(log τ), exchange processes take place, even below T G, which essentially occur on the same time scale as reorientation.

We present an analysis of 2H spin−lattice relaxation of deuterated benzene in the glass former tricresyl phosphate. Whereas pronounced nonexponential relaxation is observed well below the glass transition temperature T g (181 K), a crossover to exponential relaxation is found at T ≅ T g. By assuming a certain distribution of barriers for the rotational jump process of benzene in the glassy matrix, we are able to describe all relaxation curves in the temperature range 20 K < T < 160 K. The crossover from nonexponential to exponential relaxation is described by additionally assuming translational exchange processes of benzene molecules between sites with different barriers for the rotational jumps. Within a simple model we are able to determine exchange rates over more than two decades. Comparing the results with diffusion data, we conclude that the obtained rates correspond to diffusion coefficients in the range 10-18 < D/m2 s-1 < 10-16 and that translational rotational coupling does not hold near T g. The further potential of the method in order to determine very slow exchange processes is discussed.

The supercooled plastically crystalline phase (glassy crystal) of cyanoadamantane was investigated by multidimensional 2H NMR (T>Tg). Although the orientationally disordered crystalline phase always coexisted with the orientationally ordered crystalline phase, we were able to single out the signal from the glassy crystal by selective excitation and it was possible to carry out line shape measurements and two-dimensional (2D) experiments (in frequency and time domain). The latter directly reveal sixfold jumps with an reorientation of the molecular C3 axis via 90 degrees angles, thus reflecting the symmetry of the lattice. The motion around the C3 axis is found to be always fast. We can reproduce the line shape by random walk simulations properly taking into account these molecular motions. Both methods (line shape and 2D experiments) yield time constants which agree with those reported by other techniques. Refining the analysis a narrow distribution of correlation times is introduced to account for a weak stretching of the correlation function. We did not find any indication of a small angle process usually found in structural glasses. Thus, the motional process in the glassy crystal appears to be simple and quite different from that in structural glasses.

No abstract