Session Q36: Water and Ice

Ionic force field optimization and modified ion-pair mixing rules

We propose an optimization scheme to obtain good ionic force fields for classical simulations of salt solutions, also of biological relevance. Our work is based on Molecular Dynamics simulations with explicit (SPC/E) water for different halide and alkali ions forming salt solutions at finite ion concentration. The force field derivation technique we propose is based on a simultaneous optimization of single-ion and ion-pair properties and the determination of the cation-anion interaction parameters (traditionally given by the mixing rules). From the finite-concentration simulations, thermodynamic properties of the salt solutions are derived, using the Kirkwood-Buff theory of solutions, and compared to relevant experimental data. For the rather size-symmetric salt solutions involving bromide and chloride ions, this scheme using the standard mixing rules works fine. For the iodide and fluoride solutions, corresponding to the largest and smallest anion we have considered, a rescaling of the mixing rules was necessary. In this respect, we have introduced scaling factors for the cation-anion Lennard-Jones interaction that quantify deviations from the standard mixing rules. We discuss the efficiency and complications of the proposed ionic force field optimization scheme.   

Effects of Applied Electric Field on the Dynamics of Nano-Confined Water

We present quasi-elastic neutron scattering measurements of the proton diffusion in water confined in silica nanopores (FSM), with average pore diameters of 16 {\AA} and 39 {\AA}. The measurements were performed on the high resolution backscattering silicon spectrometer (BaSiS) at the Spallation Neutron Source (SNS). From the data, we determine the self diffusion constants, and the translational and rotational relaxation times, as a function of temperature from 300 K down to 200 K. We observe a significant slowing down of the proton diffusion as the temperature is lowered, and a remarkable effect of confinement of the translational motion. Recent results on the effects of applied electrical field on these dynamical processes will be reported.   

Competing Nuclear Quantum Effects and van der Waals Interactions in Water

Water plays a central role in many scientific disciplines, and a number of studies have been performed to understand its properties. However, providing an accurate ab initio description is a significant challenge, and because of this, many of water's properties remain elusive. In particular, the description of hydrogen bonding and the importance of van der Waals (vdW) interactions and nuclear quantum effects are still matters of debate. Recent computational advancements have been made that allow for the accurate and efficient modeling of such effects. We present results from path integral molecular dynamics simulations based on density functional theory employing exchange and correlation functionals capable of accounting for vdW interactions (so-called vdW-DF, vdW-DF2, and optB88-vdW). We demonstrate that, contrary to expectation, the interaction between nuclear quantum effects and vdW interactions hardens the structure of water. These results suggest that ad hoc methods to account for these effects, such as temperature rescaling of simulations employing classical nuclei, are insufficient to describe water, and that fully ab initio calculations must be performed. We discuss the implications of these results for understanding the local structure and hydrogen bonding in water.   

BSE/GW calculations of liquid and solid H$_2$O

We have calculated both the UV/VIS and Oxygen K-edge x-ray spectra of model ice and water systems within many-body perturbation theory using state-of-the-art Bethe-Salpeter equation (BSE) and GW self-energy approximations [1], as implemented in the valence- and core-excitation codes OCEAN and AI2NBSE [2]. While the various phases of crystalline ice have well-characterized structures, the local environment and fluctuations of liquid water remain subjects of debate. Due in part to limitations of previous theoretical models, the interpretation of experimental probes has been controversial. We find that the BSE approach, which provides an accurate treatment of core-hole interactions, is vital for a quantitative agreement between experiment and theory. Likewise the effects of self-energy corrections within the GW approximation are needed to explain the observed band-stretching and damping of the spectra. Prospects for further improvements are briefly discussed. \\[4pt] [1] J. Vinson, J. J. Kas, F. D. Vila, J. J. Rehr, and E. L. Shirley, arXiv:1010.0025 (2011).\\[0pt] [2] J. Vinson et al., Phys. Rev. B 83, 115106 (2011); H. M. Lawler et al., Phys. Rev. B 78, 205108 (2008).   

Effect of hydrogen bond cooperativity on the phase behavior of water

Four scenarios have been proposed for the low--temperature phase behavior of liquid water, each predicting different thermodynamics. The physical mechanism which leads to each is debated. Moreover, it is still unclear which of the scenarios best describes water, as there is no definitive experimental test. Here we address both open issues by analyzing a microscopic cell model within a mean--field limit. We show that a common physical mechanism underlies each of the four scenarios, and that two key physical quantities determine which of the four scenarios describes water: (i) the strength of the directional component of the hydrogen bond and (ii) the strength of the cooperative component of the hydrogen bond. The four scenarios may be mapped in the space of these two quantities. Using estimates from experimental data for H bond properties, the model predicts that the low-temperature phase diagram of water exhibits a liquid--liquid critical point at positive pressure.   

Structure and dynamics of the liquid water-ZnO$(10\bar{1}0)$ interface from first principles

Liquid water-metal oxide interfaces are of fundamental and technological interest. In this context, the water-ZnO$(10\bar{1}0)$ interface is an extensively studied system, which is also relevant for instance to the field of heterogeneous catalysis and photocatalysis. Yet, whether or not water dissociates at this surface at coverages exceeding one monolayer is still a matter of debate. Likewise questions about proton transfer to the surface, water diffusion and rearrangement within the hydrogen-bonded network remain unanswered. Here we report the first density functional theory (DFT) molecular dynamics study of a liquid water film on ZnO$(10\bar{1}0)$ and of water at monolayer coverages. The water structure obtained for the first layer in the liquid simulation differs quite significantly from that at monolayer coverage. Hydrogen bonding between the first layer and the water overlayer plays a crucial role in the stabilisation of the new adsorption structure. Rapid proton transfer and rattling within the hydrogen bonding network at the interface is observed and analysed in detail. On the whole, this study provides considerable new insight into water structure and dynamics and proton transfer at ZnO$(10\bar{1}0)$ and in the field of liquid water-metal oxides interfaces in general.   

Density fluctuations and dielectric constant of water in low and high density liquid states

The hypothesis of a liquid-liquid critical point (LLCP) in the phase diagram of water, though first published many years ago, still remains the subject of a heated debate. According to this hypothesis there exists a critical point near $T \approx 244$ K, and $P \approx 215$ MPa, located at the end of a coexistence line between a high density liquid (HDL) and a low density liquid state (LDL). The LLCP lies below the homogenous nucleation temperature of water and it has so far remained inaccessible to experiments. We study a model of water exhibiting a liquid-liquid phase transition (that is a liquid interacting through the ST2 potential) and investigate the properties of dipolar fluctuations as a function of density, in the HDL and LDL. We find an interesting correlation between the macroscopic dielectric constants and the densities of the two liquids in the vicinity of the critical point, and we discuss possible implications for measurements close to the region where the LLCP may be located.   

The role of quantum nuclear effects in hydrogen bonded crystals and their calculated NMR shielding constants

Because of its ubiquity in nature the hydrogen atom plays a very important role in computational materials science. As the lightest element, hydrogen nuclei are also the most strongly affected by quantum nuclear effects (QNEs). The path integral (PI) formalism provides a rigorous approach to obtain equilibrium quantum static properties, but PI simulations in conjunction with electronic structure calculations are rarely used due to high computational requirements. This contribution will discuss ab initio PI simulations aimed at elucidating the role of QNEs in hydrogen bonded crystals and how these impact upon experimental observables such as nuclear magnetic resonance (NMR) shielding constants and chemical shifts. We find that ab initio PI simulations improve the agreement with experimental chemical shifts compared to simulations with classical nuclei and that the influence of QNEs is very sensitive to the strength of the hydrogen bonds in the crystal.   

Selective mode analysis of nuclear quantum effects for liquid water using non-Markovian thermostats

Simulating nuclear quantum effects in liquid water using both DFT and force fields has been an active area of research in recent years. Recently, Ceriotti et. al [1] introduced a comprehensive framework to use a custom-tailored Langevin equation with correlated-noise in the context of molecular-dynamics simulations. One of the interesting applications of these thermostats is that, such a framework can be used to selectively damp normal modes whose frequency falls within a prescribed range. In this work we study how the flexible force field models respond to the selective mode thermostating using the delta-like memory kernels. We apply this delta thermostat to the molecular dynamics of TIP4P/F water force field [2], a model explicitly fitted with the lack of zero point ionic vibrations. We address the question of whether thermostating each mode to its zero point temperature is enough to generate the nuclear quantum effects in water and similar systems. This work also provides a way to identify the dominant modes for which the quantum effects are important. [1] M. Ceriotti, G. Bussi, and M. Parrinello, Phys. Rev. Lett. 103, 030603 (2009). [2] S. Habershon, T. E. Markland, and D. E. Manolopoulos, J. Chem. Phys 131, 024501 (2009 ).   

Anomalous nuclear quantum effects in ice

The lattice parameters of light (H$_2$O) and heavy (D$_2$O) Ih ice at 10 K differ by 0.09\%.$[1]$ The larger lattice constant is that of the heavier isotope. This isotope shift with anomalous sign is linked to the zero point point energy of phonons in ice. To determine the origin of this anomaly, we use \textit{ab initio} density functional theory to compute the free energy of ice within the quasiharmonic approximation. As expected, the frozen lattice constant at T = 0 K is smaller than the quantum lattice constant, independent of the isotopic substitution. We find that, the heavy isotope D gives more zero point expansion than H, whereas the heavy isotope $^{18}$O gives normal zero point expansion, i.e smaller than $^{16}$O. Relative to the the classical result, the net effect of quantum nuclei (H and O) on volume has the conventional (positive) sign at T = 0 but it becomes negative above 70 K, indicating that it may be also relevant for liquid water. These results are not reproduced by state of art polarizable empirical potentials.$[2]$ [1] B. K. R\"{o}ttger \textit{et. al.}, Acta Cryst. B {\bf 50}, 644-648 (1994). [2] C. P. Herrero and R. Ram\'{i}rez, J. Chem. Phys. {\bf 134}, 094510 (2011).   

Mechanism of sessile water droplet evaporation

The energy transport mechanisms during the evaporation of sessile water droplets have been investigated. Steady-state evaporation experiments were conducted on substrates of Cu, Au(111) and PDMS. The buoyancy-driven convection was suppressed by maintaining the droplets' base temperature just less than 4$^{\circ}$C while evaporation cooled the liquid-vapor interface to lower temperatures. The temperature fields were measured in solid (only in Au(111) experiments), liquid and vapor phases. On all three substrates, the energy balance at the liquid-vapor interface showed that thermocapillary convection transported the major portion of the energy required for the evaporation. It transported up to 98\% for Cu, up to 87\% for Au(111) and up to 72\% for PDMS. The role of thermocapillary convection is dominant close to three-phase contact line where most of the evaporation occurs. The experiment on Au (111) showed that of the energy supplied by the solid substrate, only a small portion is transported perpendicular to the solid-liquid interface to the bulk liquid phase. A much larger proportion is conducted through the adsorbed layer at the solid-liquid interface to the three-phase contact line where it is distributed by thermocapillary convection over the liquid-vapor interface and consumed by the phase change process.   

Theory and Simulation of Droplet Wetting on Patterned Surfaces

Liquid droplets can have multiple wetting modes on physically patterned surfaces, each corresponding to a (meta)stable state. For example, in the Cassie mode, the droplet resides on top of the pattern, while in the Wenzel mode, the droplet penetrates into the pattern. In this work, we study the wetting of patterned surfaces on two different length scales: on the nano-scale using molecular dynamics (MD) and on the macro-scale by minimizing free-energy expressions for various droplet wetting modes. We find that surface topography, size, and initial position of the droplet strongly affect the wetting states and contact angles. In the small Bond-number (small droplet) regime, the surface topography can be scaled by the droplet size, such that the preferred wetting modes and contact angles become independent of droplet size for surfaces with the same scaled topography. MD simulations and theory are in good agreement for small Bond numbers. For moderate to large Bond numbers, gravity plays an important role and MD simulations cannot accurately describe wetting. We create wetting phase diagrams and find that our predictions are in good agreement with experiment. The resulting wetting phase diagrams may serve as a guideline in creating surfaces with desired wettability.   

Freezing of Water next to Solid Surfaces Probed Using Sum-Frequency Generation Spectroscopy

The control of ice formation next to solid surfaces is important in many technological applications such as de-icing for aircrafts and generation of power using wind turbines. We have studied the water-ice transition next to sapphire surface to understand the freezing transition and nucleation of ice. The infrared-visible sum frequency generation spectroscopy is sensitive to the structure and orientation of water molecules next to the solid interface and provides direct information on transition kinetics at the interface. The differences in the nucleation kinetics will be discussed for water in contact with hydrophilic and hydrophobic surfaces.   

Charging and transmission of low energy particles through Amorphous Solid Water films

The interaction of charged particles with condensed water films has drawn significant attention in recent years due to its importance in biological and atmospheric processes. We have studied low energy electrons (3-25 eV) and positive argon ions (55 eV) charging and transmission effects while striking Amorphous Solid Water (ASW) films, 240-1080 ML thick, deposited on ruthenium single crystal substrate, utilizing contact potential difference (CPD) measurements. Charging by both species has shown a plate capacitor-like behaviour. L-defects energetically located just below the conduction band of ice, are likely to stabilize them. The incoming electrons kinetic energy dictates the maximal CPD by retardation of any further electrons from adding up to the already accumulated charges. Electron transmission measurements (0.5-1.5 microamps) have shown that the maximal and stable CPD values were obtained only following a relatively slow change that has developed within the ASW structure. Upon film stabilization, the spontaneous discharge was measured over a period of up to three hours. UV laser photo-emission study of the charged films has suggested that the negative charges tend to reside primarily at the ASW-vacuum interface, in good agreement with a study of charged water nano-clusters.