A flexible model for water based on TIP4P/2005

A new flexible water model, TIP4P/2005f, is developed. The idea was to add intramolecular degrees of freedom to the successful rigid model TIP4P/2005 in order to try to improve the predictions for some properties, and to enable the calculation of new ones. The new model incorporates flexibility by means of a Morse potential for the bond stretching and a harmonic term for the angle bending. The parameters have been fitted to account for the peaks of the infrared spectrum of liquid water and to produce an averaged geometry close to that of TIP4P/2005. As for the intermolecular interactions, only a small change in the σ parameter of the Lennard-Jones potential has been introduced. The overall predictions are very close to those of TIP4P/2005. This ensures that the new model may be used with the same confidence as its predecessor in studies where a flexible model is advisable.     

Study of multipole contributions to the structure of water around ions in solution using the soft sticky dipole-quadrupole-octupole (SSDQO) model of water

The solvation of ions in the soft sticky dipole-quadrupole-octupole (SSDQO) model for liquid water is presented here. This new potential energy function for liquid water describes water-water interactions by a Lennard-Jones term plus a sticky potential consisting of an approximate moment expansion with point dipole, quadrupole, and octupole moments. The SSDQO potential energy function using the moments from extended simple point charge (SPC/E), TIP3P, or TIP5P reproduces the pair potential energy functions and radial distribution functions of the respective multipoint model but it is much faster than even the three-point models. Here, the solvation of ions in SSDQO water is studied using ion-water potential energy functions consisting of moment expansions up to the charge-quadrupole term, up to the charge-octupole term, and up to an approximate charge-hexadecapole term using the moments of SPC/E water. The radial distributions from Monte Carlo simulations show the best agreement with the results for ions in SPC/E water for the expansion up to the charge-hexadecapole term. Thus, the best results are obtained when the water-water and ion-water potentials are exact up to the 1r(4) term and also contain an approximate 1r(5) term. Overall, the simplicity, efficiency, and accuracy of the SSDQO potential make it potentially very useful for computer simulations of aqueous solvation.     

A long-range electrostatic potential based on the Wolf method charge-neutral condition

Molecular simulations rely heavily on a long range electrostatic Coulomb interaction. The Coulomb potential decays inversely with distance, indicating infinite effective range. In practice, molecular simulations do not directly take into account such an infinite interaction. Therefore, the Ewald, fast multipole, and cutoff methods are frequently used. Although cutoff methods are implemented easily and the calculations are fast, it has been pointed out that they produce serious artifacts. Wolf and coworkers recently discovered one source of the artifacts. They found that when the total charge in a cutoff sphere disappeared, the cutoff error is dramatically suppressed. The Wolf method uses the charge-neutral principle combined with a potential damping that is realized using a complementary error function. To date, many molecular simulation studies have demonstrated the accuracy and reliability of the Wolf method. We propose a novel long-range potential that is constructed only from the charge-neutral condition of the Wolf method without potential damping. We also show that three simulation systems, in which involve liquid sodium-chloride, TIP3P water, and a charged protein in explicit waters with neutralized ions using the new potential, provide accurate statistical and dielectric properties when compared with the particle mesh Ewald method.     

Long-range Lennard-Jones and electrostatic interactions in interfaces: application of the isotropic periodic sum method

Molecular dynamics (MD) simulations of heptane/vapor, hexadecane/vapor, water/vapor, hexadecane/water, and dipalmitoylphosphatidylcholine (DPPC) bilayers and monolayers are analyzed to determine the accuracy of treating long-range interactions in interfaces with the isotropic periodic sum (IPS) method. The method and cutoff (rc) dependences of surface tensions, density profiles, water dipole orientation, and electrostatic potential profiles are used as metrics. The water/vapor, heptane/vapor, and hexadecane/vapor interfaces are accurately and efficiently calculated with 2D IPS (rc=10 A). It is demonstrated that 3D IPS is not practical for any of the interfacial systems studied. However, the hybrid method PME/IPS [Particle Mesh Ewald for electrostatics and 3D IPS for Lennard-Jones (LJ) interactions] provides an efficient way to include both types of long-range forces in simulations of large liquid/vacuum and all liquid/liquid interfaces, including lipid monolayers and bilayers. A previously published pressure-based long-range LJ correction yields results similar to those of PME/IPS for liquid/liquid interfaces. The contributions to surface tension of LJ terms arising from interactions beyond 10 A range from 13 dyn/cm for the hexadecane/vapor interface to approximately 3 dyn/cm for hexadecane/water and DPPC bilayers and monolayers. Surface tensions of alkane/vapor, hexadecane/water, and DPPC monolayers based on the CHARMM lipid force fields agree very well with experiment, whereas surface tensions of the TIP3P and TIP4P-Ew water models underestimate experiment by 16 and 11 dyn/cm, respectively. Dipole potential drops (DeltaPsi) are less sensitive to long-range LJ interactions than surface tensions. However, DeltaPsi for the DPPC bilayer (845+/-3 mV proceeding from water to lipid) and water (547+/-2 mV for TIP4P-Ew and 521+/-3 mV for TIP3P) overestimate experiment by factors of 3 and 5, respectively, and represent expected deficiencies in nonpolarizable force fields.     

Characterization of the TIP4P-Ew water model: vapor pressure and boiling point

The liquid-vapor-phase equilibrium properties of the previously developed TIP4P-Ew water model have been studied using thermodynamic integration free-energy simulation techniques in the temperature range of 274-400 K. We stress that free-energy results from simulations need to be corrected in order to be compared to the experiment. This is due to the fact that the thermodynamic end states accessible through simulations correspond to fictitious substances (classical rigid liquids and classical rigid ideal gases) while experiments operate on real substances (liquids and real gases, with quantum effects). After applying analytical corrections the vapor pressure curve obtained from simulated free-energy changes is in excellent agreement with the experimental vapor pressure curve. The boiling point of TIP4P-Ew water under ambient pressure is found to be at 370.3+/-1.9 K, about 7 K higher than the boiling point of TIP4P water (363.7+/-5.1 K; from simulations that employ finite range treatment of electrostatic and Lennard-Jones interactions). This is in contrast to the approximately +15 K by which the temperature of the density maximum and the melting temperature of TIP4P-Ew are shifted relative to TIP4P, indicating that the temperature range over which the liquid phase of TIP4P-Ew is stable is narrower than that of TIP4P and resembles more that of real water. The quality of the vapor pressure results highlights the success of TIP4P-Ew in describing the energetic and entropic aspects of intermolecular interactions in liquid water.     

Vapor+liquid equilibrium of water, carbon dioxide, and the binary system, water+carbon dioxide, from molecular simulation

NVT- and NpT-Gibbs ensemble Monte Carlo (GEMC) simulations were applied to describe the vapor–liquid equilibrium of water (between 323 and 573 K), carbon dioxide (between 230 and 290 K) and their binary mixtures (between 348 and 393 K). The properties of supercritical carbon dioxide were determined between 310 and 520 K by NpT-MC simulations. Literature data for the effective pair potentials (for water: the SPC-, SPC/E-, and TIP4P potential models; for carbon dioxide: the EPM2 potential model) were used to describe the properties of the pure substances. The vapor pressures of water and carbon dioxide are calculated. For water, the SPC- and TIP4P models give superior results for the vapor pressure when compared to the SPC/E model. The vapor–liquid equilibrium of the binary mixture, carbon dioxide–water, was predicted using the SPC- as well as the TIP4P model for water and the EPM2 model for carbon dioxide. The interactions between carbon dioxide and water were estimated from the pair potentials of the pure components using common mixing rules without any adjustable binary parameter. Agreement of the predicted data for the compositions of the coexisting phases in vapor–liquid equilibrium and experimental results is observed within the statistical uncertainties of the simulation results in the investigated range of state, i.e. at pressures up to about 20 MPa.     

Correlations in liquid water for the TIP3P-Ewald, TIP4P-2005, TIP5P-Ewald, and SWM4-NDP models

Water is one of the simplest molecules in existence, but also one of the most important in biological and engineered systems. However, understanding the structure and dynamics of liquid water remains a major scientific challenge. Molecular dynamics simulations of liquid water were performed using the water models TIP3P-Ewald, TIP4P-2005, TIP5P-Ewald, and SWM4-NDP to calculate the radial distribution functions (RDFs), the relative angular distributions, and the excess enthalpies, entropies, and free energies. In addition, lower-order approximations to the entropy were considered, identifying the fourth-order approximation as an excellent estimate of the full entropy. The second-order and third-order approximations are ~20% larger and smaller than the true entropy, respectively. All four models perform very well in predicting the radial distribution functions, with the TIP5P-Ewald model providing the best match to the experimental data. The models also perform well in predicting the excess entropy, enthalpy, and free energy of liquid water. The TIP4P-2005 and SWM4-NDP models are more accurate than the TIP3P-Ewald and TIP5P-Ewald models in this respect. However, the relative angular distribution functions of the four water models reveal notable differences. The TIP5P-Ewald model demonstrates an increased preference for water molecules to act both as tetrahedral hydrogen bond donors and acceptors, whereas the SWM4-NDP model demonstrates an increased preference for water molecules to act as planar hydrogen bond acceptors. These differences are not uncovered by analysis of the RDFs or the commonly employed tetrahedral order parameter. However, they are expected to be very important when considering water molecules around solutes and are thus a key consideration in modelling solvent entropy.     

The melting temperature of the most common models of water

The melting temperature of ice I(h) for several commonly used models of water (SPC, SPC/E,TIP3P,TIP4P, TIP4P/Ew, and TIP5P) is obtained from computer simulations at p = 1 bar. Since the melting temperature of ice I(h) for the TIP4P model is now known [E. Sanz, C. Vega, J. L. F. Abascal, and L. G. MacDowell, Phys. Rev. Lett. 92, 255701 (2004)], it is possible to use the Gibbs-Duhem methodology [D. Kofke, J. Chem. Phys. 98, 4149 (1993)] to evaluate the melting temperature of ice I(h) for other potential models of water. We have found that the melting temperatures of ice I(h) for SPC, SPC/E, TIP3P, TIP4P, TIP4P/Ew, and TIP5P models are T = 190 K, 215 K, 146 K, 232 K, 245 K, and 274 K, respectively. The relative stability of ice I(h) with respect to ice II for these models has also been considered. It turns out that for SPC, SPC/E, TIP3P, and TIP5P the stable phase at the normal melting point is ice II (so that ice I(h) is not a thermodynamically stable phase for these models). For TIP4P and TIP4P/Ew, ice I(h) is the stable solid phase at the standard melting point. The location of the negative charge along the H-O-H bisector appears as a critical factor in the determination of the relative stability between the I(h) and II ice forms. The methodology proposed in this paper can be used to investigate the effect upon a coexistence line due to a change in the potential parameters.     

Structure and dynamics of liquid water with different long-range interaction truncation and temperature control methods in molecular dynamics simulations

We have used molecular dynamics simulations to study the physical properties of modified TIP3P water model included in the CHARMM program, using four different methods-the Ewald summation technique, and three different spherical truncation methods-for the treatment of the long-range interactions. Both the structure and dynamics of the liquid water model were affected by the methods used to truncate the long-range interactions. For some of the methods artificial structuring of the model liquid was observed around the cutoff radius. The model liquid properties were also affected by the commonly applied temperature control methods. Four different methods for controlling the temperature of the system were studied, and the effects of these methods on the bulk properties for liquid water were analyzed. The system size was also found to change the dynamics of the model liquid water. Two control simulations with the SPC/E water model were carried out. The self-diffusion coefficient (D), the radial distribution function (g(OO)), the distance dependent Kirkwood G-factor [G(k)(r)] and the intermolecular potential energy (E(pot)) were determined from the different trajectories and compared with the experimental data.     

A modified TIP3P water potential for simulation with Ewald summation

The charges and Lennard-Jones parameters of the TIP3P water potential have been modified to improve its performance under the common condition for molecular dynamics simulations of using Ewald summation in lieu of relatively short nonbonded truncation schemes. These parameters were optimized under the condition that the hydrogen atoms do not have Lennard-Jones parameters, thus making the model independent of the combining rules used for the calculation of nonbonded, heteroatomic interaction energies, and limiting the number of Lennard-Jones calculations required. Under these conditions, this model provides accurate density (rho = 0.997 g/ml) and heat of vaporization (DeltaH(vap) = 10.53 kcal/mol) at 25 degrees C and 1 atm, but also provides improved structure in the second peak of the O-O radial distribution function and improved values for the dielectric constant (epsilon(0) = 89) and the diffusion coefficient (D = 4.0 x 10(-5) cm(2)/s) relative to the original parametrization. Like the original parameterization, however, this model does not show a temperature density maximum. Several similar models are considered with the additional constraint of trying to match the performance of the optimized potentials for liquid simulation atom force field to that obtained when using the simulation conditions under which it was originally designed, but no model was entirely satisfactory in reproducing the relative difference in free energies of hydration between the model compounds, phenol and benzene. Finally, a model that incorporates a long-range correction for truncated Lennard-Jones interactions is presented, which provides a very accurate dielectric constant (epsilon(0) = 76), however, the improvement in this estimate is on the same order as the uncertainty in the calculation.     

Relation between the melting temperature and the temperature of maximum density for the most common models of water

Water exhibits a maximum in density at normal pressure at 4 degrees above its melting point. The reproduction of this maximum is a stringent test for potential models used commonly in simulations of water. The relation between the melting temperature and the temperature of maximum density for these potential models is unknown mainly due to our ignorance about the melting temperature of these models. Recently we have determined the melting temperature of ice I(h) for several commonly used models of water (SPC, SPC/E, TIP3P, TIP4P, TIP4P/Ew, and TIP5P). In this work we locate the temperature of maximum density for these models. In this way the relative location of the temperature of maximum density with respect to the melting temperature is established. For SPC, SPC/E, TIP3P, TIP4P, and TIP4P/Ew the maximum in density occurs at about 21-37 K above the melting temperature. In all these models the negative charge is located either on the oxygen itself or on a point along the H-O-H bisector. For the TIP5P and TIP5P-E models the maximum in density occurs at about 11 K above the melting temperature. The location of the negative charge appears as a geometrical crucial factor to the relative position of the temperature of maximum density with respect to the melting temperature.     

A polarizable force field for water using an artificial neural network

A force field for liquid water including polarization effects has been constructed using an artificial neural network (ANN). It is essential to include a many-body polarization effect explicitly into a potential energy function in order to treat liquid water which is dense and highly polar. The new potential energy function is a combination of empirical and nonempirical potentials. The TIP4P model was used for the empirical part of the potential. For the nonempirical part, an ANN with a back-propagation of error algorithm (BPNN) was introduced to reproduce the complicated many-body interaction energy surface from ab initio quantum mechanical calculations. BPNN, described in terms of a matrix, provides enough flexibility to describe the complex potential energy surface (PES). The structural and thermodynamic properties, calculated by isobaric-isothermal (constant-NPT) Monte Carlo simulations with the new polarizable force field for water, are compatible with experimental results. Thus, the simulation establishes the validity of using our estimated PES with a polarization effect for accurate predictions of liquid state properties. Applications of this approach are simple and systematic so that it can easily be applied to the development of other force fields besides the water-water system.     

Computer simulation study of metastable ice VII and amorphous phases obtained by its melting

Molecular dynamics simulations of metastable ice VII and cubic ice Ic are carried out in order to examine (1) the ability of commonly used water interaction potentials to reproduce the properties of ices, and (2) the possibility of generating low-density amorphous (LDA) structures by heating ice VII, which is known to transform to LDA at approximately 135 K at normal pressure [S. Klotz, J. M. Besson, G. Hamel, R. J. Nelmes, J. S. Loveday, and W. G. Marshall, Nature (London) 398, 681 (1999)]. We test four simple empirical interaction potentials of water: TIP4P [W. L. Jorgensen, J. Chandrasekhar, J. D. Madura, R. W. Impey, and M. L. Klein, J. Chem. Phys. 79, 926 (1983)], SPC/E [H. J. C. Berendsen, J. R. Grigera, and T. P. Straatsma, J. Phys. Chem. B 91, 6269 (1987)], TIP5P [M. W. Mahoney and W. L. Jorgensen, J. Chem. Phys. 112, 8910 (2000)], and ST2 [F. H. Stillinger and A. Rahman, J. Chem. Phys. 60, 1545 (1974)]. We have found that TIP5P ice VII melts at 210 K, TIP4P at 90 K, and SPC/E at 70 K. Only TIP5P water after transition has a structure similar to that of LDA. TIP4P and SPC/E have almost identical structures, dissimilar to any known water or amorphous phases, but upon heating both slowly evolve towards LDA-like structure. ST2 ice VII is remarkably stable up to 430 K. TIP4P and SPC/E predict correctly the cubic ice collapse into a high-density amorphous ice (HDA) at approximately 1 GPa whereas TIP5P remains stable up to approximately 5 GPa. The densities of the simulated ice phases differ significantly, depending on the potential used, and are generally higher than experimental values. The importance of proper treatment of long-range electrostatic interactions is also discussed.     

A general purpose model for the condensed phases of water: TIP4P/2005

A potential model intended to be a general purpose model for the condensed phases of water is presented. TIP4P/2005 is a rigid four site model which consists of three fixed point charges and one Lennard-Jones center. The parametrization has been based on a fit of the temperature of maximum density (indirectly estimated from the melting point of hexagonal ice), the stability of several ice polymorphs and other commonly used target quantities. The calculated properties include a variety of thermodynamic properties of the liquid and solid phases, the phase diagram involving condensed phases, properties at melting and vaporization, dielectric constant, pair distribution function, and self-diffusion coefficient. These properties cover a temperature range from 123 to 573 K and pressures up to 40,000 bar. The model gives an impressive performance for this variety of properties and thermodynamic conditions. For example, it gives excellent predictions for the densities at 1 bar with a maximum density at 278 K and an averaged difference with experiment of 7 x 10(-4) g/cm3.     

A simple, efficient polarizable coarse-grained water model for molecular dynamics simulations

The development of coarse-grained (CG) models that correctly represent the important features of compounds is essential to overcome the limitations in time scale and system size currently encountered in atomistic molecular dynamics simulations. Most approaches reported in the literature model one or several molecules into a single uncharged CG bead. For water, this implicit treatment of the electrostatic interactions, however, fails to mimic important properties, e.g., the dielectric screening. Therefore, a coarse-grained model for water is proposed which treats the electrostatic interactions between clusters of water molecules explicitly. Five water molecules are embedded in a spherical CG bead consisting of two oppositely charged particles which represent a dipole. The bond connecting the two particles in a bead is unconstrained, which makes the model polarizable. Experimental and all-atom simulated data of liquid water at room temperature are used for parametrization of the model. The experimental density and the relative static dielectric permittivity were chosen as primary target properties. The model properties are compared with those obtained from experiment, from clusters of simple-point-charge water molecules of appropriate size in the liquid phase, and for other CG water models if available. The comparison shows that not all atomistic properties can be reproduced by a CG model, so properties of key importance have to be selected when coarse graining is applied. Yet, the CG model reproduces the key characteristics of liquid water while being computationally 1-2 orders of magnitude more efficient than standard fine-grained atomistic water models.     

Monte Carlo calculation of the thermodynamic properties of water

The Monte Carlo method and parallel computing are used to calculate the thermodynamic properties of water (density, heat capacity, compressibility, thermal expansion coefficient, and static dielectric constant) in a wide range of temperatures (from 70 K to 530 K) at constant (atmospheric) pressure. Four groups of computational experiments are carried out, each for its own model of the water molecule: TIP3P (Jorgensen et al., 1983), SPC/E (Berendsen et al., 1987), TIP4P/2005 (Abascal&Vega, 2005), and TIP5P-E (Rick, 2004). An additional calculation based on the replica exchange method is conducted for the TIP4P/2005 model. A comparison of the calculated properties of water with experimental data suggests that the TIP4P/2005 model can provide highly realistic computer simulation results for water and aqueous solutions.     

Soft sticky dipole-quadrupole-octupole potential energy function for liquid water: an approximate moment expansion

A new, efficient potential energy function for liquid water is presented here. The new model, which is referred here as the soft sticky dipole-quadrupole-octupole (SSDQO) model, describes a water molecule as a Lennard-Jones sphere with point dipole, quadrupole, and octupole moments. It is a single-point model and resembles the hard-sphere sticky dipole potential model for water by Bratko et al. [J. Chem. Phys. 83, 6367 (1985)] and the soft sticky dipole model by Ichiye and Liu [J. Phys. Chem. 100, 2723 (1996)] except now the sticky potential consists of an approximate moment expansion for the dimer interaction potential, which is much faster than the true moment expansion. The object here is to demonstrate that the SSDQO potential energy function can accurately mimic the potential energy function of a multipoint model using the moments of that model. First, the SSDQO potential energy function using the dipole, quadruple, and octupole moments from SPC/E, TIP3P, or TIP5P is shown to reproduce the dimer potential energy functions of the respective multipoint model. In addition, in Monte Carlo simulations of the pure liquid at room temperature, SSDQO reproduces radial distribution functions of the respective model. However, the Monte Carlo simulations using the SSDQO model are about three times faster than those using the three-point models and the long-range interactions decay faster for SSDQO (1/r(3) and faster) than for multipoint models (1/r). Moreover, the contribution of each moment to the energetics and other properties can be determined. Overall, the simplicity, efficiency, and accuracy of the SSDQO potential energy function make it potentially very useful for studies of aqueous solvation by computer simulations.     

Dynamics of TIP5P and TIP4P/ice potentials

The dynamics of a thin film of ice Ih deposited on MgO (001) is studied through molecular dynamics simulations performed with two new potential models of ice. This system is chosen because it is possible to compare the results of the simulations to incoherent neutron quasielastic scattering experiments performed few years ago and to previous molecular dynamics simulations using the TIP4P potential model. The present simulations are performed to determine the evolution of the translational and orientational order parameters of the ice film upon temperature increase in the 250-280 K range. They are also used to calculate the translational and orientational diffusion coefficients of the water molecules in the supported film as a function of the temperature. When using the TIP5P potential, the present results show a better agreement with experimental data than those calculated with the TIP4P potential, especially regarding the temperature above which significant changes are obtained in the dynamics of the water film. Similar conclusions are obtained when using the TIP4P/ice potential, although this latter potential clearly underestimates the translational diffusion coefficients.     

Molecular dynamics simulations of the aqueous interface with the [BMI][PF6] ionic liquid: Comparison of different solvent models

We report a Molecular Dynamics (MD) study of the interface between water and the hygroscopic room temperature Ionic Liquid "IL" [BMI][PF6] (1-butyl-3-methyl-imidazolium hexafluorophosphate), comparing the TIP3P, SPC/E and TIP5P models for water and two IL models where the ions are +/-1 or +/-0.9 charged. A recent MD study (A. Chaumont, R. Schurhammer and G. Wipff, J. Phys. Chem. B, 2005, 109, 18964) showed that using TIP3P water in conjunction with the IL(+/-1) model led to water-IL mixing without forming an interface, whereas a biphasic system could be obtained with the IL(+/-0.9) model. With the TIP5P and SPC/E models, the juxtaposed aqueous and IL phases are found to remain distinct for at least 20 ns. The resulting IL humidity, exaggerated with the IL(+/-1) model, is in better agreement with experiment using the IL(+/-0.9) model. We also report demixing simulations on the "randomly mixed" liquids, using the IL(+/-0.9) model for the ionic liquid. With the three tested water models, the phases separate very slowly ( approximately 20 ns or more) compared to "classical" chloroform-water mixtures (less than 1 ns), leading to biphasic systems similar to those obtained after equilibration of the juxtaposed liquids. The characteristics of the interface (size, polarity, ion orientation, electrostatic potential) are compared with the different models. Possible reasons why, among the three tested water models, the widely-used TIP3P model exaggerates the inter-solvent mixing, are analyzed. The difficulty in computationally and experimentally equilibrating water-IL mixtures is attributed to the slow dynamics and micro-heterogeneity of the IL and to the different states of water in the IL phase.