A classical polarizable model for simulations of water and ice

We develop a classical rigid polarizable model of water for molecular simulations of water and ice. The model uses the Rowlinson five-site geometry: oxygen bearing the Lennard-Jones interaction and linearly polarizable point dipole, two positively charged hydrogens, and two massless negative charges placed symmetrically off oxygen so that the experimental dipole moment is reproduced. The target properties are the densities of water and ice Ih, diffusivity, enthalpies of fusion and vaporization, and the ice Ih melting point. The surface tension at lower temperatures is by 7% underestimated whereas the dielectric constant by 6% overestimated. Diffusivity and viscosity worsen at higher temperatures, although the Stokes radius is overestimated only by 2-7%. The ice Ih melting temperature is 260 K and the temperature of maximum density is 269 K. Rescaling the charges by a factor of 1.01 and Lennard-Jones energy by 1.0201 improves the melting point and energy-related quantities but shifts the agreement of kinetic properties to higher temperatures. For the model we propose abbreviation POL4D.     

A non-polarizable model of water that yields the dielectric constant and the density anomalies of the liquid: TIP4Q

A four-site rigid water model is presented, whose parameters are fitted to reproduce the experimental static dielectric constant at 298 K, the maximum density of liquid water and the equation of state at low pressures. The model has a positive charge on each of the three atomic nuclei and a negative charge located at the bisector of the HOH bending angle. This charge distribution allows increasing the molecular dipole moment relative to four-site models with only three charges and improves the liquid dielectric constant at different temperatures. Several other properties of the liquid and of ice Ih resulting from numerical simulations with the model are in good agreement with experimental values over a wide range of temperatures and pressures. Moreover, the model yields the minimum density of supercooled water at 190 K and the minimum thermal compressibility at 310 K, close to the experimental values. A discussion is presented on the structural changes of liquid water in the supercooled region where the derivative of density with respect to temperature is a maximum.     

Simulations of proton order and disorder in ice Ih

Computer simulations of ice Ih with different proton orientations are presented. Simulations of proton disordered ice are carried out using a Monte Carlo method which samples over proton degree of freedom, allowing for the calculation of the dielectric constant and for the examination of the degree of proton disorder. Simulations are also presented for two proton ordered structures of ice Ih, the ferroelectric Cmc2(1) structure or ice XI and the antiferroelectric Pna2(1) structure. These simulations indicate that a transition to a proton ordered phase occurs at low temperatures (below 80 K). The symmetry of the ordered phase is found to be dependent on the water potential. The stability of the two proton ordered structures is due to a balance of short-ranged interactions which tend to stabilize the Pna2(1) structure and longer-range interactions which stabilize the Cmc2(1) structure.     

Homogeneous ice nucleation from supercooled water

Homogeneous ice nucleation from supercooled water was studied in the temperature range of 220-240 K through combining the forward flux sampling method (Allen et al., J. Chem. Phys., 2006, 124, 024102) with molecular dynamics simulations (FFS/MD), based on a recently developed coarse-grained water model (mW) (Molinero et al., J. Phys. Chem. B, 2009, 113, 4008). The calculated ice nucleation rates display a strong temperature dependence, ranging from 2.148 ± 0.635 × 10(25) m(-3) s(-1) at 220 K to 1.672 ± 0.970 × 10(-7) m(-3) s(-1) at 240 K. These rates can be fitted according to the classical nucleation theory, yielding an estimate of the effective ice-water interface energy γ(ls) of 31.01 ± 0.21 mJ m(-2) for the mW water model. Compared to experiments, our calculation underestimates the homogeneous ice nucleation rate by a few orders of magnitude. Possible reasons for the discrepancy are discussed. The nucleating ice embryo contains both cubic ice Ic and hexagonal ice Ih, with the fraction of each structure being roughly 50% when the critical size is reached. In particular, a novel defect structure containing nearly five-fold twin boundaries is identified in the ice clusters formed during nucleation. The way such defect structure is formed is found to be different from mechanisms proposed for the formation of the same defect in metallic nanoparticles and thin film. The quasi five-fold twin boundary structure found here is expected to occur in the crystallization of a wide range of materials with the diamond cubic structure, including ice.     

Use of molecular dynamics simulations with ab initio SCF calculations for the determination of the deuterium quadrupole coupling constant in liquid water and bond lengths in ice

The deuterium quadrupole coupling constant and asymmetry parameter in heavy water were determined using ab initio SCF calculations. Snapshots from a molecular dynamics simulation were used to give liquid water cluster configurations and the influence of simulation parameters on the quadrupole coupling constant was investigated. The electronic potential model and the number of molecules in the molecular dynamics simulation and the pressure of the system were found to have only a small influence on the quadrupole coupling constant. The average value of the quadrupole coupling constant at room temperature, corrected for the known deficiency of the ab initio calculation in the gas phase, yields a quadrupole coupling constant of 253 kHz, in perfect agreement with the most recent experiments. The oxygen—deuterium bond lengths in ice Ih, ice II, and ice IX were determined using experimental quadrupole coupling constants and a model equation. An averaged bond length of 98.9 pm was obtained for the Ih form, which is approximately 2 pm shorter than that determined by neutron diffraction studies, whereas the bond lengths for the four deuterium sites in ice II and the three sites in ice IX are in fair agreement with experiment. © John Wiley & Sons, Inc.     

Isotope effects in ice Ih: a path-integral simulation

Ice Ih has been studied by path-integral molecular dynamics simulations, using the effective q-TIP4P/F potential model for flexible water. This has allowed us to analyze finite-temperature quantum effects in this solid phase from 25 to 300 K at ambient pressure. Among these effects we find a negative thermal expansion of ice at low temperatures, which does not appear in classical molecular dynamics simulations. The compressibility derived from volume fluctuations gives results in line with experimental data. We have analyzed isotope effects in ice Ih by considering normal, heavy, and tritiated water. In particular, we studied the effect of changing the isotopic mass of hydrogen on the kinetic energy and atomic delocalization in the crystal as well as on structural properties such as interatomic distances and molar volume. For D(2)O ice Ih at 100 K we obtained a decrease in molar volume and intramolecular O-H distance of 0.6% and 0.4%, respectively, as compared to H(2)O ice.     

Temperature dependence of solvation dynamics of probe molecules in methanol-doped ice and in liquid ethanol

We have studied the solvation statics and dynamics of coumarin 343 and a strong photoacid (pK* approximately 0.7) 2-naphthol-6, 8-disulfonate (2N68DS) in methanol-doped ice (1% molar concentration of methanol) and in cold liquid ethanol in the temperature range of 160-270 K. Both probe molecules show a relatively fast solvation dynamics in ice, ranging from a few tens of picoseconds at about 240 K to nanoseconds at about 160 K. At about 160 K in doped ice, we observe a sharp decrease of the dynamic Stokes shift of both coumarin 343 and 2N68DS. Its value is approximately only 200 cm-1 at approximately 160 K compared to about 1100 cm-1 at T >/= 200 K (at times longer than t > 10 ps). We find a good correlation between the inefficient and slow excited-state proton-transfer rate at low-temperature ice, T < 180 K, and the dramatic decrease of the solvation energy, as measured by the dynamic band shift, at these low temperatures. We find that the average solvation rate in ice is similar to its value in liquid ethanol at all given temperatures in the range of 200-250 K. The surprisingly fast solvation rate in ice is explained by the relatively large freedom of the water hydrogen rotation in ice Ih.     

Investigation of the hydrogen bonding in ice Ih by first-principles density function methods

It is a well recognized difficult task to simulate the vibrational dynamics of ices using the density functional theory (DFT), and there has thus been rather limited success in modelling the inelastic neutron scattering (INS) spectra for even the simplest structure of ice, ice Ih, particularly in the translational region below 400 cm(-1). The reason is partly due to the complex nature of hydrogen bonding (H-bond) among water-water molecules which require considerable improvement of the quantum mechanical simulation methods, and partly owing to the randomness of protons in ice structures which often requires simulation of large super-lattices. In this report, we present the first series of successful simulation results for ice Ih using DFT methods. On the basis of the recent advancement in the DFT programs, we have achieved for the first time theoretical outcomes that not only reproduce the rotational frequencies between 500 to 1200 cm(-1) for ice Ih, but also the two optic peaks at ～240 and 320 cm(-1) in the translational region of the INS spectra [J. C. Li, J. Chem. Phys 105, 6733 (1996)]. Besides, we have also investigated the impact of pairwise configurations of H(2)O molecules on the H-bond and found that different proton arrangements of pairwise H(2)O in the ice Ih crystal lattice could not alter the nature of H-bond as significantly as suggested in an early paper [J. C. Li and D. K. Ross, Nature (London) 365, 327 (1993)], i.e., reproducing the two experimental optic peaks do not need to invoke the two H-bonds as proposed in the previous model which led to considerable debates. The results of this work suggest that the observed optic peaks may be attributed to the coupling between the two bands of H-O stretching modes in H(2)O. The current computational work is expected to shed new light on the nature of the H-bonds in water, and in addition to offer a new approach towards probing the interaction between water and biomaterials for which H-bond is essential.     

An application of flexible constraints in Monte Carlo simulations of the isobaric--isothermal ensemble of liquid water and ice Ih with the polarizable and flexible mobile charge densities in harmonic oscillators model

The method of flexible constraints was implemented in a Monte Carlo code to perform numerical simulations of liquid water and ice Ih in the constant number of molecules, volume, and temperature and constant pressure, instead of volume ensembles, using the polarizable and flexible mobile charge densities in harmonic oscillators (MCDHO) model. The structural and energetic results for the liquid at T=298 K and rho=997 kg m(-3) were in good agreement with those obtained from molecular dynamics. The density obtained at P=1 atm with flexible constraints, rho=1008 kg m(-3), was slightly lower than with the classical sampling of the intramolecular vibrations, rho=1010 kg m(-3). The comparison of the structures and energies found for water hexamers and for ice Ih with six standard empirical models to those obtained with MCDHO, show this latter to perform better in describing water far from ambient conditions: the MCDHO minimum lattice energy, density, and lattice constants were in good agreement with experiment. The average angle HOH of the water molecule in ice was predicted to be slightly larger than in the liquid, yet 1.2% smaller than the experimental value.     

Adsorption and structure of water on kaolinite surfaces: possible insight into ice nucleation from grand canonical monte carlo calculations

Grand canonical Monte Carlo calculations are used to determine water adsorption and structure on defect-free kaolinite surfaces as a function of relative humidity at 235 K. This information is then used to gain insight into ice nucleation on kaolinite surfaces. Results for both the SPC/E and TIP5P-E water models are compared and demonstrate that the Al-surface [(001) plane] and both protonated and unprotonated edges [(100) plane] strongly adsorb at atmospherically relevant relative humidities. Adsorption on the Al-surface exhibits properties of a first-order process with evidence of collective behavior, whereas adsorption on the edges is essentially continuous and appears dominated by strong water lattice interactions. For the protonated and unprotonated edges no structure that matches hexagonal ice is observed. For the Al-surface some of the water molecules formed hexagonal rings. However, the a o lattice parameter for these rings is significantly different from the corresponding constant for hexagonal ice ( Ih). A misfit strain of 14.0% is calculated between the hexagonal pattern of water adsorbed on the Al-surface and the basal plane of ice Ih. Hence, the ring structures that form on the Al-surface are not expected to be good building-blocks for ice nucleation due to the large misfit strain.     

Sticking of CO to crystalline and amorphous ice surfaces

We present results of classical trajectory calculations on the sticking of hyperthermal CO to the basal plane (0001) face of crystalline ice Ih and to the surface of amorphous ice Ia. The calculations were performed for normal incidence at a surface temperature Ts = 90 K for ice Ia, and at Ts = 90 and 150 K for ice Ih. For both surfaces, the sticking probability can be fitted to a simple exponentially decaying function of the incidence energy, Ei: Ps = 1.0e(-Ei(kJ/mol)/90(kJ/mol)) at Ts = 90 K. The energy transfer from the impinging molecule to the crystalline and the amorphous surface is found to be quite efficient, in agreement with the results of molecular beam experiments on the scattering of the similar molecule, N2, from crystalline and amorphous ice. However, the energy transfer is less efficient for amorphous than for crystalline ice. Our calculations predict that the sticking probability decreases with Ts for CO scattering from crystalline ice, as the energy transfer from the impinging molecule to the warmer surfaces becomes less efficient. At high Ei (up to 193 kJ/mol), no surface penetration occurs in the case of crystalline ice. However, for CO colliding with the amorphous surface, a penetrating trajectory was observed to occur into a large water pore. The molecular dynamics calculations predict that the average potential energy of CO adsorbed to ice Ih is -10.1 +/- 0.2 and -8.4 +/- 0.2 kJ/mol for CO adsorbed to ice Ia. These values are in agreement with previous experimental and theoretical data. The distribution of the potential energy of CO adsorbed to ice Ia was found to be wider (with a standard deviation sigma of 2.4 kJ/mol) than that of CO interacting with ice Ih (sigma = 2.0 kJ/mol). In collisions with ice Ia, the CO molecules scatter at larger angles and over a wider distribution of angles than in collisions with ice Ih.     

Melting points and thermal expansivities of proton-disordered hexagonal ice with several model potentials

A method of free energy calculation is proposed, which enables to cover a wide range of pressure and temperature. The free energies of proton-disordered hexagonal ice (ice Ih) and liquid water are calculated for the TIP4P [J. Chem. Phys. 79, 926 (1983)] model and the TIP5P [J. Chem. Phys. 112, 8910 (2000)] model. From the calculated free energy curves, we determine the melting point of the proton-disordered hexagonal ice at 0.1 MPa (atmospheric pressure), 50 MPa, 100 MPa, and 200 MPa. The melting temperatures at atmospheric pressure for the TIP4P ice and the TIP5P ice are found to be about T(m)=229 K and T(m)=268 K, respectively. The melting temperatures decrease as the pressure is increased, a feature consistent with the pressure dependence of the melting point for realistic proton-disordered hexagonal ice. We also calculate the thermal expansivity of the model ices. Negative thermal expansivity is observed at the low temperature region for the TIP4P ice, but not for the TIP5P ice at the ambient pressure.     

Unusual Gruneisen and Bridgman parameters of low-density amorphous ice and their implications on pressure induced amorphization

The low-temperature limiting value of the Grüneisen parameter for low-frequency phonons and the density dependence of the thermal conductivity (Bridgman parameter) of low-density amorphous (LDA) ice, high-density amorphous (HDA) ice, hexagonal ice Ih, and cubic ice Ic were calculated from high-pressure sound velocity and thermal conductivity measurements, yielding negative values for all states except HDA ice. LDA ice is the first amorphous state to exhibit a negative Bridgman parameter, and negative Grüneisen parameters are relatively unusual. Since Ih, Ic, and LDA ice all transform to HDA upon pressurization at low temperatures and share the unusual feature of negative Grüneisen parameters, this seems to be a prerequisite for pressure induced amorphization. We estimate that the Grüneisen parameter increases at the ice Ih to XI transition, and may become positive in ice XI, which indicates that proton-ordered ice XI does not amorphize like ice Ih on pressurization.     

The bend angle of water in ice Ih and liquid water: The significance of implementing the nonlinear monomer dipole moment surface in classical interaction potentials

The implementation of the physically accurate nonlinear dipole moment surface of the water monomer in the context of the Thole-type, polarizable, flexible interaction potential results in the only classical potential, which, starting from the gas phase value for the bend angle (104.52 degrees), reproduces its experimentally observed increase in the ice Ih lattice and in liquid water. This is in contrast to all other classical potentials to date, which predict a decrease of the monomer bend angle in ice Ih and in liquid water with respect to the gas phase monomer value. Simulations under periodic boundary conditions of several supercells consisting of up to 288 molecules of water used to sample the proton disorder in the ice Ih lattice yield an average value of vartheta(HOH)(I(h))=108.4 degrees +/-0.2 degrees for the minimized structures (T=0 K) and 108.1 degrees +/-2.8 degrees at T=100 K. Analogous simulations for liquid water predict an average value of vartheta(HOH)(liquid)=106.3 degrees +/-4.9 degrees at T=300 K. The increase of the monomer bend angle of water in condensed environments is attributed to the use of geometry-dependent charges that are used to describe the nonlinear character of the monomer's dipole moment surface. Our results suggest a new paradigm in the development of classical interaction potential models of water that can be used to describe condensed aqueous environments.     

Melting behavior of water in cylindrical pores: carbon nanotubes and silica glasses

We report a study of the effects of confinement in multi-walled carbon nanotubes and mesoporous silica glasses (SBA-15) on the solid structure and melting of both H(2)O and D(2)O ice, using differential scanning calorimetry, dielectric relaxation spectroscopy, and neutron diffraction. Multi-walled nanotubes of 2.4, 3.9 and 10 nm are studied, and the SBA-15 studied has pores of mean diameter 3.9 nm; temperatures ranging from approximately 110 to 290 K were studied. We find that the melting point is depressed relative to the bulk water for all systems studied, with the depression being greater in the case of the silica mesopores. These results are shown to be consistent with molecular simulation studies of freezing in silica and carbon materials. The neutron diffraction data show that the cubic phase of ice is stabilized by the confinement in carbon nanotubes, as well as in silica mesopores, and persists up to temperatures of about 240 K, above which there is a transition to the hexagonal ice structure.     

On the Debye-Waller factor of hexagonal ice: a computer simulation study

We investigate by molecular dynamics (MD) simulations the temperature dependence of the Debye-Waller (DW) factor of hexagonal ice with 25 different proton-disordered configurations. Each initial configuration is composed of 288 water molecules with no net dipole moment. The intermolecular interaction of water is described by TIP4P potential. Each production run of the simulation is 15 ns or longer. We observe a change in slope of the DW factor around 200 K, which cannot be explained within the framework of either classical or quantum harmonic approximation. Configurations generated by MD simulations are subjected to the steepest descent energy minimization. Analysis of the local energy minimum structures reveals that water molecules above 200 K jump to other lattice sites via some local energy minimum structures which contain some water molecules sitting on the locations other than the lattice sites. As time evolves, these defect molecules move back and forth to the lattice sites yielding defect-free structures. Those motions are responsible for the unusual increase in the DW factor at high temperatures. In making a transition from an energy-minimum structure to another one, a small number of water molecules are involved in a highly cooperative fashion. The larger DW factor at higher temperature arises from jump-like motions of water molecules among these locally stable configurations which may or may not be a family of the proton-disordered ice forms satisfying the "ice rule".     

Quasi-harmonic approximation of thermodynamic properties of ice Ih, II, and III

Several thermodynamic properties of ice Ih, II, and III are studied by a quasi-harmonic approximation and compared to results of quantum path integral and classical simulations. This approximation allows to obtain thermodynamic information at a fraction of the computational cost of standard simulation methods, and at the same time permits studying quantum effects related to zero-point vibrations of the atoms. Specifically, we have studied the crystal volume, bulk modulus, kinetic energy, enthalpy, and heat capacity of the three ice phases as a function of temperature and pressure. The flexible q-TIP4P/F model of water was employed for this study, although the results concerning the capability of the quasi-harmonic approximation are expected to be valid independently of the employed water model. The quasi-harmonic approximation reproduces with reasonable accuracy the results of quantum and classical simulations showing an improved agreement at low temperatures (T< 100 K). This agreement does not deteriorate as a function of pressure as long as it is not too close to the limit of mechanical stability of the ice phases.     

Raman spectra of proton ordered phase XI of ICE I. Translational vibrations below 350 cm(-1)

Polarized Raman spectra from single crystals of ice XI (proton ordered phase of ice Ih) were measured and assigned for the modes below 350 cm(-1) in the translational vibration region. In contrast to the proton disordered ice Ih, the spectra in ice XI show clear polarization dependence and several new peaks are observed. Most of the vibrational modes were successfully assigned by the simplified point mass model with the symmetry C(2v) (12)(Cmc2(1)) and by the depolarization effect. In particular, LO-TO splitting of the mode near 240 cm(-1) was experimentally confirmed for the first time, which indicates that the long range force effect appears distinctly in ice XI.