Paper detail

Resolving the controversy on the glass transition temperature of water?

We consider experimental data on the dynamics of water (1) in glass-forming aqueous mixtures with glass transition temperature Tg approaching the putative Tg=136 K of water from above and below, (2) in confined spaces of nanometer in size and (3) in the bulk at temperatures above the homogeneous nucleation temperature. Altogether, the considered relaxation times from the data range nearly over 15 decades from 10-12 to 103 s. Assisted by the various features in the isothermal spectra and theoretical interpretation, these considerations enable us to conclude that relaxation of un-crystallized water is highly non-cooperative. The exponent Beta_K of its Kohlrausch stretched exponential correlation function is not far from having the value of one, and hence the deviation from exponential time decay is slight. Albeit the temperature dependence of its alpha-relaxation time being non-Arrhenius, the corresponding Tg-scaled temperature dependence has small steepness index m, likely less than 44 at Tg, and hence water is not 'fragile' as a glassformer. The separation in time scale of the alpha- and the beta-relaxations is small at Tg, becomes smaller at higher temperatures, and they merge together shortly above Tg. From all these properties and by inference, water is highly non-cooperative as a glass-former, it has short cooperative length-scale, and possibly minimal configurational entropy and small change of heat capacity at Tg compared with other organic glass-formers. This conclusion is perhaps unsurprising because water is the smallest molecule. Our deductions from the data rule out that the Tg of water is higher than 160 K, and suggest that it is close to the traditional value of 136 K.