Breaking the polar‐nonpolar division in solvation free energy prediction

Implicit solvent models divide solvation free energies into polar and nonpolar additive contributions, whereas polar and nonpolar interactions are inseparable and nonadditive. We present a feature functional theory (FFT) framework to break this ad hoc division. The essential ideas of FFT are as follows: (i) representability assumption: there exists a microscopic feature vector that can uniquely characterize and distinguish one molecule from another; (ii) feature‐function relationship assumption: the macroscopic features, including solvation free energy, of a molecule is a functional of microscopic feature vectors; and (iii) similarity assumption: molecules with similar microscopic features have similar macroscopic properties, such as solvation free energies. Based on these assumptions, solvation free energy prediction is carried out in the following protocol. First, we construct a molecular microscopic feature vector that is efficient in characterizing the solvation process using quantum mechanics and Poisson–Boltzmann theory. Microscopic feature vectors are combined with macroscopic features, that is, physical observable, to form extended feature vectors. Additionally, we partition a solvation dataset into queries according to molecular compositions. Moreover, for each target molecule, we adopt a machine learning algorithm for its nearest neighbor search, based on the selected microscopic feature vectors. Finally, from the extended feature vectors of obtained nearest neighbors, we construct a functional of solvation free energy, which is employed to predict the solvation free energy of the target molecule. The proposed FFT model has been extensively validated via a large dataset of 668 molecules. The leave‐one‐out test gives an optimal root‐mean‐square error (RMSE) of 1.05 kcal/mol. FFT predictions of SAMPL0, SAMPL1, SAMPL2, SAMPL3, and SAMPL4 challenge sets deliver the RMSEs of 0.61, 1.86, 1.64, 0.86, and 1.14 kcal/mol, respectively. Using a test set of 94 molecules and its associated training set, the present approach was carefully compared with a classic solvation model based on weighted solvent accessible surface area. © 2017 Wiley Periodicals, Inc.     

Density functional theory of solvation and its relation to implicit solvent models

We describe a density functional theory approach to solvation in molecular solvents. The solvation free energy of a complex solute can be obtained by direct minimization of a density functional, instead of the thermodynamic integration scheme necessary when using atomistic simulations. In the homogeneous reference fluid approximation, the expression of the free-energy functional relies on the knowledge of the direct correlation function of the pure solvent. After discussing general molecular solvents, we present a generic density functional describing a dipolar solvent and we show how it can be reduced to the conventional implicit solvent models when the solvent microscopic structure is neglected. With respect to those models, the functional includes additional effects such as the microscopic structure of the solvent, the dipolar saturation effect, and the nonlocal character of the dielectric constant. We also show how this functional can be minimized numerically on a three-dimensional grid around a solute of complex shape to provide, in a single shot, both the average solvent structure and the absolute solvation free energy.     

Solvation forces between molecularly rough surfaces

Surface heterogeneity affects significantly wetting and adhesion properties. However, most of the theories and simulation methods of calculating solid-fluid interactions assume a standard thermodynamic model of the Gibbs' dividing solid-fluid interface, which is molecularly smooth. This assumption gives rise to a layering of the fluid phase near the surface that is displayed in oscillating density profiles in any theories and simulation models, which account for the hard core intermolecular repulsion. This layering brings about oscillations of the solvation (or disjoining) pressure as a function of the gap distance, which are rarely observed in experiments, except for ideal monocrystal surfaces. We present a detailed study of the effects of surface roughness on the solvation pressure of Lennard-Jones (LJ) fluids confined by LJ walls based on the quenched solid density functional theory (QSDFT). In QSDFT, the surface roughness is quantified by the roughness parameter, which represents the thickness of the surface "corona" - the region of varying solid density. We show that the surface roughness of the amplitude comparable with the fluid molecular diameter effectively damps the oscillations of solvation pressure that would be observed for molecularly smooth surfaces. The calculations were done for the LJ model of nitrogen sorption at 74.4 K in slit-shaped carbon nanopores to provide an opportunity of comparing with standard adsorption experiments. In addition to a better understanding of the fundamentals of fluid adsorption on heterogeneous surfaces and inter-particle interactions, an important practical outcome is envisioned in modeling of adsorption-induced deformation of compliant porous substrates.     

Complete resolution of the ionization equilibria of 5-deoxypyridoxal in water-dioxane mixtures

We have used a potentiometric method to determine the thermodynamic equilibrium constants for the macroscopic ionization processes of 5-deoxypyridoxal (DPL) in water-dioxane mixtures (0-70% weight fraction in dioxane) at temperatures ranging from 10°C to 50°C. These data, together with previously published equilibrium constants for the tautomerism and hydration processes, have allowed us to resolve the complete microconstant system. We have also calculated the microscopic ionization equilibrium constants under all the experimental conditions. The changes of standard thermodynamic function for the macroscopic and microscopic ionization processes were obtained in various water-dioxane mixtures at 25°C. The values of a given microscopic pK with different solvents and temperatures fit very well to an equation which relates this magnitude with the thermodynamic parameters, the solvation of the components of the reaction, and a solvent parameter. We have obtained an interesting linear correlation between the thermodynamic parameters corresponding to all the microscopic ionizations of DPL and the net change of the solvation during the process: enthalpies correlate linearly for all the microscopic ionizations, while entropies do so for the phenols and pyridinium ions separately.     

Capillary condensation of a binary mixture in slit-like pores

We investigate the capillary condensation of two model fluid mixtures in slit-like pores, which exhibit different demixing properties in the bulk phase. The interactions between adsorbate particles are modeled by using Lennard-Jones (12,6) potentials and the adsorbing potentials are of the Lennard-Jones (9,3) type. The calculations are performed for different pore widths and at different concentrations of the bulk gas, by means of density functional theory. We evaluate the capillary phase diagrams and discuss their dependence on the parameters of the model. Our calculations indicate that a binary mixture confined to a slit-like pore may exhibit rich phase behavior.     

Deformability of adsorbents during adsorption and principles of the thermodynamics of solid-phase systems

A microscopic theory of adsorption, based on a discrete continuum lattice gas model for noninert (including deformable) adsorbents that change their lattice parameters during adsorption, is presented. Cases of the complete and partial equilibrium states of the adsorbent are considered. In the former, the adsorbent consists of coexisting solid and vapor phases of adsorbent components, and the adsorbate is a mobile component of the vapor phase with an arbitrary density (up to that of the liquid adsorbate phase). The adsorptive transitioning to the bound state changes the state of the near-surface region of the adsorbent. In the latter, there are no equilibrium components of the adsorbent between the solid and vapor phases. The adsorbent state is shown to be determined by its prehistory, rather than set by chemical potentials of vapor of its components. Relations between the microscopic theory and thermodynamic interpretations are discussed: (1) adsorption on an open surface, (2) two-dimensional stratification of the adsorbate mobile phase on an open homogeneous surface, (3) small microcrystals in vacuum and the gas phase, and (4) adsorption in porous systems.     

Mean-field kinetic nucleation theory

A new semiphenomenological model of homogeneous vapor-liquid nucleation is proposed in which the cluster kinetics follows the "kinetic approach to nucleation" and the thermodynamic part is based on the revised Fisher droplet model with the mean-field argument for the cluster configuration integral. The theory is nonperturbative in a cluster size and as such is valid for all clusters down to monomers. It contains two surface tensions: macroscopic (planar) and microscopic. The latter is a temperature dependent quantity related to the vapor compressibility factor at saturation. For Lennard-Jones fluids the microscopic surface tension possesses a universal behavior with the parameters found from the mean-field density functional calculations. The theory is verified against nucleation experiments for argon, nitrogen, water, and mercury, demonstrating very good agreement with experimental data. Classical nucleation theory fails to predict experimental results when a critical cluster becomes small.     

Solvation effects for polymers at an interface: a hybrid self-consistent field-density functional theory approach

Using polyatomic density functional theory of Kierlik and Rosinberg, we show that Wertheim's thermodynamic perturbation theory (TPT) incorporates solvation effects in a systematic, although simplified form. We derive two approximate solvation potentials, which require the knowledge of the correlation function in the reference unbonded fluid only. The theoretical predictions are tested against many-chain Monte Carlo simulations for moderate chain lengths. The predictions of the end-to-end distance in the bulk are in a reasonable agreement with simulations for the TPT(M-1) approximation, while the simpler TPT2_e approximation leads to the solvation potential that is shorter ranged and considerably less accurate. The resulting conformations are used in the subsequent self-consistent field theory calculations of hard-sphere polymers at a hard wall. While the incorporation of the solvation effects has little impact on the density profiles, the predictions of the components of the end-to-end distance vector as a function of the distance to the wall are much improved.     

Second-harmonic generation and effect of adsorbates for free-electron metal and semiconductor surfaces

We discuss aspects of a developing microscopic theory of SHG from simple metal and semiconductor surfaces. For semiconductors calculations of the dynamical nonlinear susceptibility on the basis of realistic tight-binding parametrizations of the electronic Hamiltonian provide a practical scheme. In the resulting spectra the effect of the dangling bonds on SHG is clearly seen together with a strong decrease upon saturation with H atoms. In the metal case the adsorbate induced changes of the static nonlinear electron density can be calculated self-consistently by applying density functional theory to the jellium model. The second-order dipole moment determines the effect of adsorbates on the SHG intensity in the adiabatic limit. Quite general a correlation with the nature of the adsorbate expressed by its electronegativity and the characteristic charge transfer, adsorption dipole and polarizabilities in first and second order is found.     

Mechanochemical phenomena in microporous bodies

Mechanochemical phenomena in microporous bodies, including initial compression in a vacuum, additional compression in the initial gas adsorption stage, and expansion during subsequent adsorption, were analyzed under the constraint of mechanical equilibrium using adsorbate pressure tensor in pores of various shapes. Precise calculation of deformation was performed for cylindrical and spherical pores. Pore surface tension was shown to be responsible for initial compression of adsorbent in a vacuum and its expansion during gas adsorption. Insignificant additional compression of adsorbent in the initial adsorption stage was attributed to specificity of the adsorbate pressure tensor in a pore, effect of temperature, and chemisorption (provided that active centers are present).     

First-principles prediction of the pK(a)s of anti-inflammatory oxicams

The gas- and aqueous-phase acidities of a series of oxicams have been computed by combining M05-2X/6-311+G(3df,2p) gas-phase free energies with solvation free energies from the CPCM-UAKS, COSMO-RS, and SMD solvent models. To facilitate accurate gas-phase calculations, a benchmarking study was further carried out to assess the performance of various density functional theory methods against the high-level composite method G3MP2(+). Oxicams are typically diprotic acids, and several tautomers are possible in each protonation state. The direct thermodynamic cycle and the proton exchange scheme have been employed to compute the microscopic pK(a)s on both solution- and gas-phase equilibrium conformers, and these were combined to yield the macroscopic pK(a) values. Using the direct cycle of pK(a) calculation, the CPCM-UAKS model delivered reasonably accurate results with MAD ~ 1, whereas the SMD and COSMO-RS models' performance was less satisfactory with MAD ~ 3. Comparison with experiment also indicates that direct cycle calculations based on solution conformers generally deliver better accuracy. The proton exchange cycle affords further improvement for all solvent models through systematic error cancellation and therefore provides better reliability for the pK(a) prediction of compounds of these types. The latter approach has been applied to predict the pK(a)s of several recently synthesized oxicam derivatives.     

Investigation of excess adsorption, solvation force, and plate-fluid interfacial tension for Lennard-Jones fluid confined in slit pores

The excess Helmholtz free energy functional is formulated in terms of a modified fundamental measure theory [Y. X. Yu and J. Z. Wu, J. Chem. Phys. 117, 10156 (2002)] for a short ranged repulsion and a first-order mean-spherical approximation theory [Y. P. Tang, J. Chem. Phys. 118, 4140 (2003)] for a long ranged attraction. Within the framework of the density functional theory, the density profile, excess adsorption, solvation force, and plate-fluid interfacial tension of a Lennard-Jones fluid confined in slit pores are predicted, and the results agree well with the simulation data. The phase equilibria inside the slit pores are determined according to the requirement that temperature, chemical potential, and grand potential in coexistence phases should be equal, and the plate-fluid interfacial tensions at equilibrium states are predicted consequently.     

Adsorption-induced deformation of microporous carbons: pore size distribution effect

We present a thermodynamic model of adsorption-induced deformation of microporous carbons. The model represents the carbon structure as a macroscopically isotropic disordered three-dimensional medium composed of stacks of slit-shaped pores of different sizes embedded in an incompressible amorphous matrix. Adsorption stress in pores is calculated by means of Monte Carlo simulations. The proposed model reproduces qualitatively the experimental nonmonotonic dilatometric deformation curve for argon adsorption on carbide-derived activated carbon at 243 K and pressure up to 1.2 MPa. The elastic deformation (contraction at low pressures and swelling at higher pressures) results from the adsorption stress that depends strongly on the pore size. The pore size distribution determines the shape of the deformation curve, whereas the bulk modulus controls the extent of the sample deformation.     

Application of reference‐modified density functional theory: Temperature and pressure dependences of solvation free energy

Recently, we proposed a reference‐modified density functional theory (RMDFT) to calculate solvation free energy (SFE), in which a hard‐sphere fluid was introduced as the reference system instead of an ideal molecular gas. Through the RMDFT, using an optimal diameter for the hard‐sphere reference system, the values of the SFE calculated at room temperature and normal pressure were in good agreement with those for more than 500 small organic molecules in water as determined by experiments. In this study, we present an application of the RMDFT for calculating the temperature and pressure dependences of the SFE for solute molecules in water. We demonstrate that the RMDFT has high predictive ability for the temperature and pressure dependences of the SFE for small solute molecules in water when the optimal reference hard‐sphere diameter determined for each thermodynamic condition is used. We also apply the RMDFT to investigate the temperature and pressure dependences of the thermodynamic stability of an artificial small protein, chignolin, and discuss the mechanism of high‐temperature and high‐pressure unfolding of the protein. © 2017 Wiley Periodicals, Inc.     

Calculation of the characteristics of ethanol and methanol adsorbed in a pore of activated carbon by the density functional method

Density functional theory was used to perform quantum-chemical calculations of changes in the energy and structure characteristics of methanol and ethanol molecules caused by their adsorption in model slitlike pores of activated carbon. The conclusion was made that changes in these characteristics (bond lengths, angles, charges on atoms, and harmonic vibration frequencies) is additional evidence of the validity of the Tolmachev thermodynamic model, in which adsorption is treated as a quasi-chemical reaction of the addition of adsorbate molecules to adsorption centers of an adsorbent. It was shown that the arrangement of alcohol molecules, when the C-O and C-C bonds were approximately parallel to pore walls and the hydrogen atom was directed toward a nearer pore wall, was most favorable energetically. Two adsorbed alcohol molecules are also arranged parallel to pore walls and form a hydrogen bond.     

Analytical theory for permeation of condensable gases in a nanopore

The permeation of a condensable gas mixture in a pressure gradient is examined within a dynamic density functional theory (DDFT). The non-equilibrium density and flux profiles of gas molecules trapped within a nanopore are calculated for each species as a function of feed gas density. Because of important fluid–fluid interaction close to condensation the flux and density gradients are not related by constant transport diffusivities with the thermodynamic correction of uniform density. For long narrow pores the relation of the selectivity to the equilibrium adsorption isotherms is validated. Improved separation is achieved by combining preferential wall interaction and advantageous attraction between gas molecules of different species and examples are discussed. Results from experiments and simulations of permeation in binary mixtures near condensation are still rare and the theory provides a simple basis to study qualitative trends using known parameters.     

Redox entropy of plastocyanin: developing a microscopic view of mesoscopic polar solvation

We report applications of analytical formalisms and molecular dynamics (MD) simulations to the calculation of redox entropy of plastocyanin metalloprotein in aqueous solution. The goal of our analysis is to establish critical components of the theory required to describe polar solvation at the mesoscopic scale. The analytical techniques include a microscopic formalism based on structure factors of the solvent dipolar orientations and density and continuum dielectric theories. The microscopic theory employs the atomistic structure of the protein with force-field atomic charges and solvent structure factors obtained from separate MD simulations of the homogeneous solvent. The MD simulations provide linear response solvation free energies and reorganization energies of electron transfer in the temperature range of 280-310 K. We found that continuum models universally underestimate solvation entropies, and a more favorable agreement is reported between the microscopic calculations and MD simulations. The analysis of simulations also suggests that difficulties of extending standard formalisms to protein solvation are related to the inhomogeneous structure of the solvation shell at the protein-water interface combining islands of highly structured water around ionized residues along with partial dewetting of hydrophobic patches. Quantitative theories of electrostatic protein hydration need to incorporate realistic density profile of water at the protein-water interface.