Water in confined geometries has obvious relevance in biology, geology, and other areas where the material properties are strongly dependent on the amount and behavior of water in these types of materials. Another reason to restrict the size of water domains by different types of geometrical confinements has been the possibility to study the structural and dynamical behavior of water in the deeply supercooled regime (e.g., 150-230 K at ambient pressure), where bulk water immediately crystallizes to ice. In this paper we give a short review of studies with this particular goal. However, from these studies it is also clear that the interpretations of the experimental data are far from evident. Therefore, we present three main interpretations to explain the experimental data, and we discuss their advantages and disadvantages. Unfortunately, none of the proposed scenarios is able to predict all the observations for supercooled and glassy bulk water, indicating that either the structural and dynamical alterations of confined water are too severe to make predictions for bulk water or the differences in how the studied water has been prepared (applied cooling rate, resulting density of the water, etc.) are too large for direct and quantitative comparisons.

No abstract

2H NMR is used to study the mechanisms for the reorientation of protein hydration water. In the past, crossovers in temperature-dependent correlation times were reported at Tx1 approximately 225 K (X1) and Tx2 approximately 200 K (X2). We show that neither X1 nor X2 are related to a fragile-to-strong transition. Our results rule out an existence of X1. Also, they indicate that water performs thermally activated and distorted tetrahedral jumps at T < Tx2, implying that X2 originates in an onset of this motion, which may be related to a universal defect diffusion in materials with defined hydrogen-bond networks.

The microscopic details of local particle dynamics is studied in a glass-forming one component supercooled liquid modeled by a Dzugutov potential developed for simple metallic glass formers. Our main goal is to investigate particle motion in the supercooled liquid state, and to ascertain the extent to which this motion is cooperative and occurring in quasi-one-dimesional, string-like paths. To this end we investigate in detail the mechanism by which particles move along these paths. In particular, we show that the degree of coherence--that is, simultaneous motion by consecutive particles along a string--depends on the length of the string. For short strings, the motion is highly coherent. For longer strings, the motion is highly coherent only within shorter segments of the string, which we call "microstrings." Very large strings may contain several microstrings within which particles move simultaneously, but individual microstrings within a given string are temporally uncorrelated with each other. We discuss possible underlying mechanism for this complex dynamical behavior, and examine our results in the context of recent work by Garrahan and Chandler [Phys. Rev. Lett. 89, 035704 (2002)] in which dynamic facilitation plays a central role in the glass transition.

(2)H NMR reveals two dynamic crossovers of supercooled water in nanoscopic (∼2 nm) confinement. At ∼225 K, a dynamic crossover of liquid water is accompanied by formation of a fraction of solid water. Therefore, we do not attribute the effect to a liquid-liquid phase transition but rather to a change from bulk-like to interface-dominated dynamics. Moreover, we argue that the α process and β process are observed in experiments above and below this temperature, respectively. Upon cooling through a dynamic crossover at ∼175 K, the dynamics of the liquid fraction becomes anisotropic and localized, implying solidification of the corresponding water network, most probably, during a confinement-affected glass transition.