Obtained by rapidly quenching a liquid to slow atomic motion sufficiently to prevent crystallization, glasses with their amorphous, noncrystalline structures are, in principle, subject to some instability associated with later relaxation toward a more stable state. For transparent and insulating silica glasses, this aging is far too slow to be of concern at normal temperatures and time scale. But metallic glasses in which aging occurs more rapidly are a different story. In particular, studying relaxation dynamics in metallic glasses opens an important window for looking into the inner workings of these materials, says Wei Hua Wang of the Institute of Physics, Chinese Academy of Sciences, in Beijing, who argues that it may one day be possible to control their properties by means of models constructed from understanding their relaxation dynamics. An important step in this direction has been taken by Wang’s group, which reported in a recent issue of Physical Review Letters that it has discovered a new two-step decoupling of relaxation processes at temperatures below the glass transition temperature.

Owing to the lack of a crystalline structure and the consequent lack of crystalline defects such as dislocations and grain boundaries, the mechanical and physical properties of metallic glasses have pluses and minuses relative to their crystalline cousins. For example, they can be very strong and corrosion-resistant but also brittle. But understanding the origin of their wide spectrum of properties and hence routes to improving them suffers because the relationship between structure and properties is difficult to characterize. For example, how does deformation of an amorphous material occur at the atomic level? Relaxation dynamics offers a way into the problem since both it and deformation involve local structural rearrangements. Still, says Wang, in-depth understanding of the intrinsic relaxation mechanism, especially the extremely slow dynamics and the relationship to the properties of metallic glasses remain elusive, owing to the lack of coverage of broad time and temperature ranges in previous studies.

One of the advantages of metallic glasses is that the structural units are atoms that can be modeled as simple hard spheres, in contrast to, for example, silica glass with covalently bonded SiO₂ units. Another is that there is enough atomic motion below the glass transition temperature to make experiments feasible. Relaxation rates in a metallic glass drop dramatically as the temperature drops, first below the melting temperature and later below the glass transition temperature, and the mechanisms throughout this range are turning out to be similarly diverse. Above the glass transition temperature in metallic glasses, there is a two-step decoupling of relaxation processes: a fast (β) process resembles the rattling and subsequent escape of particles in the cages formed by their nearest neighbors, and a slower (α) process associated with long-range translational motion. The two-step decoupling observed by Wang and his co-workers, however, takes place below the glass transition temperature and involves different processes.

Earlier work in Wang’s group and elsewhere had already provided evidence for new structural rearrangements at work at low temperatures. A key addition to the mechanical relaxation dynamics technique, Wang says, was to cover a wide range of temperatures near and below the glass transition temperature and times spanning five orders of magnitude to allow time for them to emerge. A convenient way to measure relaxation processes is to apply a fixed strain to a sample and monitor the decay of the stress required. Stress relaxation profiles were done for three materials: Zr₄₄Ti₁₁Cu₁₀Ni₁₀Be₂₅, Zr₅₀Cu₄₀Al₁₀, and La₅₅Ni₂₀Al₂₅ with glass transition temperatures of 621 K, 693 K, and 471 K, respectively. Samples were in the form of melt-spun ribbons 25-µm thick that had been pre-annealed at 90% of the glass transition temperature for 48 hours prior to measurements in order to minimize aging effects.

Plotting the decaying stress for times longer than 30 hours revealed a two-step “slow-flow” relaxation. Near the glass transition temperature, there was a single-step decay, but 20–30 K below, a shoulder appeared in the decay curve that lasted for about 30 minutes before giving way to a second decay process. Fitting the curves with a standard expression accordingly yielded two processes, one with a fast relaxation rate and a low activation energy that changed less than an order of magnitude as the temperature dropped to about two-thirds of the glass transition temperature, and one with a higher activation energy and a rate that plunged by as much as 10 orders of magnitude over the same temperature range. With this finding, the experimenters added annealing time and strain as variables and found that the second, slow process only occurred with a sufficiently long aging time.

From these and other detailed considerations, including comparing reaction rates and activation energies, Wang’s group concluded that interpreting the findings in terms of previous models was ruled out. In general, they ascribed the fast process to microscopic internal stress relaxation, probably associated with ballistic particle motion. Evidence for such a process had been found recently in x-ray photon correlation spectroscopy (XPCS) by researchers including co-author Beatrice Ruta of the European Synchrotron Radiation Facility in Grenoble. The XPCS provides indirect information on density fluctuations on a length scale of a few Ångstroms. This correlation between microscopic and macroscopic techniques may open a new avenue of research. The slower mode appears associated with a broad distribution of relaxation times and larger length scales.

Overall, says Kia Ngai of the Italian Consiglio Nazionale di Recerche in Pisa, “With their finding by macroscopic mechanical relaxation, this two-step decoupling comprising a fast relaxation, for which there is previous microscopic evidence from x-ray photon correlation spectroscopy, and a slower relaxation not seen earlier, Wang’s group has provided new insight into the dynamics of the glassy state of metallic glasses.” However, many questions remain to be answered.

Read the abstract in Physical Review Letters.