We present NeuralIL, a model for the potential energy of an ionic liquid that accurately reproduces first-principles results with orders-of-magnitude savings in computational cost. Built on the basis ...of a multilayer perceptron and spherical Bessel descriptors of the atomic environments, NeuralIL is implemented in such a way as to be fully automatically differentiable. It can thus be trained on ab initio forces instead of just energies, to make the most out of the available data, and can efficiently predict arbitrary derivatives of the potential energy. Using ethylammonium nitrate as the test system, we obtain out-of-sample accuracies better than 2 meV atom–1 (<0.05 kcal mol–1) in the energies and 70 meV Å–1 in the forces. We show that encoding the element-specific density in the spherical Bessel descriptors is key to achieving this. Harnessing the information provided by the forces drastically reduces the amount of atomic configurations required to train a neural network force field based on atom-centered descriptors. We choose the Swish-1 activation function and discuss the role of this choice in keeping the neural network differentiable. Furthermore, the possibility of training on small data sets allows for an ensemble-learning approach to the detection of extrapolation. Finally, we find that a separate treatment of long-range interactions is not required to achieve a high-quality representation of the potential energy surface of these dense ionic systems.
Full text
Available for:
IJS, KILJ, NUK, PNG, UL, UM
The thermodynamic properties of ionic liquids (ILs) bearing alkylsilane and alkylsiloxane chains, as well as their carbon-based analogs, were investigated. Effects such as the replacement of carbon ...atoms by silicon atoms, the introduction of a siloxane linkage, and the length of the alkylsilane chain were explored. Differential scanning calorimetry (DSC) and thermogravimetric analysis (TGA) were used to study the thermal and phase behavior (glass transition temperature, melting point, enthalpy and entropy of fusion, and thermal stability). Heat capacity was obtained by high-precision drop calorimetry and differential scanning microcalorimetry. The volatility and cohesive energy of these ILs were investigated via the Knudsen effusion method coupled with a quartz crystal microbalance (KEQCM). Gas phase energetics and structure were also studied to obtain the gas phase heat capacity as well as the energy profile associated with the rotation of the IL side chain. The computational study suggested the existence of an intramolecular interaction in the alkylsiloxane-based IL. The obtained glass transition temperatures seem to follow the trend of chain flexibility. An increase of the alkylsilane chain leads to a seemingly linear increase in molar heat capacity. A regular increment of 30 J·K–1·mol–1 in the molar heat capacity was found for the replacement of carbon by silicon in the IL alkyl chain. The alkylsilane series was revealed to be slightly more volatile than its carbon-based analogs. A further increase in volatility was found for the alkylsiloxane-based IL, which is likely related to the decrease of the cohesive energy due to the existence of an intramolecular interaction between the siloxane linkage and the imidazolium headgroup. The use of Si in the IL structure is a suitable way to significantly reduce the IL’s viscosity while preserving its large liquid range (low melting point and high thermal stability) and low volatilities.
Full text
Available for:
IJS, KILJ, NUK, PNG, UL, UM
This work has been performed to study about the thermophysical properties of the binary mixtures of 1-ethyl-3-methylimidazolium dicyanamide (EMIMN(CN)2) ionic liquids (ILs) with 1-butanol and ...2-butanol molecular organic solvents. With the help of high precision vibrating tube density meter, density (ρ), and speed of sound (u) while using an automated falling ball microviscometer, viscosities (η) have been experimentally evaluated for pure as well as binary mixtures over whole mole fraction (x 1) range of IL at T = 298.15–323.15 K and p = 0.1 MPa. Using these data, the excess/deviation parameters have been determined. Fitting of evaluated excess/deviation parameters was performed using the well-known Redlich–Kister polynomial equation. Prigogine–Flory–Patterson equation was also used to correlate the excess molar volume V E data of each binary mixture and found that both V E values are very close to each other. The solute-solvent interactions were discussed in terms of various forces of attraction, and the details are described in the Results and Discussion section. Classical MD simulations at room temperature for the chosen concentrations confirmed that the molecular environment is very similar, regardless of whether 1-butanol or 2-butanol is present in the binary mixtures. However, 1-butanol showed slightly greater interaction with the IL than 2-butanol, which seems to have more freedom to accommodate itself within the interstitial sites of the IL.
Full text
Available for:
IJS, KILJ, NUK, PNG, UL, UM
Ionic Liquids (ILs) show remarkable properties which make them environmental friendly. This work focused on sophisticated measurements of thermophysical properties of IL, 1-butyl-3-methylimidazolium ...trifluoromethanesulfonate (BMIMCF3SO3) with molecular solvents, dimethylsulfoxide (DMSO) and ethylene glycol (EG). The densities, ρ, speeds of sounds, u, and dynamic viscosities, η, of pure IL and its binary/ternary mixtures with DMSO or/and EG over the entire mole fractions range at various temperatures from 298.15 K to 323.15 K with an interval of 5 K and at pressure 0.1 MPa have been measured experimentally. Excess molar volumes, VE, excess molar isentropic compressions, ΚS, mE, viscosity deviations, ∆η, and Gibbs free energy of activations, ∆G⁎ have been evaluated from the experimental data at all the studied compositions and temperatures. The experimental studies of the focused systems reveal that the interaction found in IL + DMSO system is comparatively more pronounced as compared to IL + EG systems. But for DMSO + EG system, the interactions are strongest in contrast to IL-solvent systems. The VE, ΚS, mE and ∆η are fitted to Redlich-Kister polynomial equation and it fits well for all the studied systems. The experimental and derived parameters, varied with mole fraction and temperature, discussed in terms of ion-ion, ion-dipole and dipole-dipole interactions. Some semi-empirical models for the correlation of experimental dynamic viscosity data and to evaluate interactional parameters prevailing in the studied binary mixtures are also employed. The excess molar volumes of all the studied binary systems have been correlated with the Prigogine-Flory-Patterson (PFP) theory which shows good agreement with experimental one. The study of ternary system shows the strong interaction among the three components i.e., IL + DMSO + EG. The MD simulations studies have also been performed for pure components and their binary mixtures to determine the radial distribution function (RDF), the dynamic properties, mean square displacement (MSD), velocity auto-correlation function (vacf), Fourier transform and vibrational density of states (vDos) terms. The RDFs suggest that both the solvents, DMSO/EG substantially disrupt the long-range-anion-cation network characteristic of IL. Herein, DMSO shows their preference towards BMIM+ whereas, the EG towards CF3SO3− and it is attributed to the presence of H-bonding. The SDF studies reveal that BMIM+ has strong coordination sites for DMSO while, the EG shows weaker coordination.
•ρ, u and η are measured for pure and mixtures, IL + DMSO/EG.•VE, Κs, mE, ∆η parameters of mixtures are correlated with Redlich-Kister equation.•Properties are explained by ion-ion, ion-dipole & dipole-dipole interactions.•PFP theory is applied to VE for binary systems.•Experimental parameters are further interpreted by MD simulations.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP
In this paper we report classical Molecular Dynamics simulations of the interfacial structure of two ternary mixtures, based on both a protic (ethylammonium nitrate, EAN) and an aprotic ...(1-ethyl-3-methylimidazolium tetrafluoroborate, EMIMBF4) ionic liquid. A molecular solvent (dimethylsulfoxide, DMSO) and a lithium salt with common anion are the rest of the components of the mixtures. The simulations were performed with implicit graphene walls, both neutral and charged with a surface charge density of ±1 e/nm2. Density, charge and electrostatic potential profiles as well as integral capacitances were calculated for all systems. For both liquids, the evolution of the density profiles throughout the DMSO concentration range are in accordance with previously characterized bulk properties of such systems. In the case of charged interfaces, the adsorption of Li+ cations into the negative electrode was found to be possible for the protic liquid, but unfavourable in the case of the aprotic one. Moreover, the probability distribution functions for the orientation of all molecular species near the interfaces were computed, and they indicate a tendency of the solvent molecules to form a dense layer at the interface separating the ionic liquids from the electrodes. The influence of hydrogen bonds in determining the dissimilitudes between protic and aprotic mixtures is highlighted.
•Ternary ionic liquid mixtures of ionic liquids with DMSO and lithium salts with interfaces.•Comparison of polar and apolar ionic liquids.•Direct relation between structure and dynamics.•Enhanced adsorption of lithium in polar ionic liquid.•Structure and potential drop studied for the whole concentration range.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP
•Binary and ternary mixtures of ionic liquids with DMSO and lithium salts.•Structure and ionic conductivity in the whole concentration range.•Direct relation between structure and dynamics.•Molecular ...dynamics and density functional theory calculations.•Lithium cage correlation as a function of solvent concentration.
In this paper we report classical Molecular Dynamics and ab initio Density Functional Theory simulations of the structure and single particle dynamics of ternary mixtures of a protic (ethylammonium nitrate, EAN) and an aprotic (1-Ethyl-3-methylimidazolium tetrafluoroborate, EMIMBF4) ionic liquid, lithium salts with common anion and a molecular cosolvent, dimethyl sulfoxide (DMSO). The coordination numbers and electrical conductivities of the different mixtures have been calculated throughout the whole concentration range, and the differences between mixtures with the protic and aprotic ionic liquids were analyzed. In both cases, the evolution of the electrical conductivity is seen to correlate well with the formation of mixed tetrahedral solvation complexes of lithium cations with ionic liquid anions and cosolvent molecules. The differences in the solvation and charge transport mechanisms in hydrogen bonded mixtures and those based in the aprotic ionic liquid are analyzed. Our conclusions indicate that the major feature behind the electrical conductivity of the ternary mixtures is the composition of the solvation shell of the metal cations in the mixtures and the rate at which anions in it are replaced by DMSO molecules.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP
Display omitted
•Maximum occupancy of the hydroquinone clathrate by hydrogen guest molecules.•Dynamics of the filling process of an empty by molecular dynamics simulations.•Slow diffusion of the ...guests and formation of a high-density adsorbed hydrogen layer.•Occupancy by hybrid GCMC/MD simulations matches closely experimental results.
Hydroquinone clathrates have been proposed as potential gas separation and storage media. Experimental results have demonstrated enhanced preferential adsorption for certain guest molecules, and also stability over temperature and pressure ranges that make them promising candidates to be employed in applications as hydrogen storage. Despite this, the characterization of these inclusion solids from thermodynamic and kinetic perspectives is still poor. In this work, we have tried to estimate the hydrogen storage ability of these clathrates using molecular simulations. The process of diffusion of hydrogen guest molecules from an external reservoir has been simulated using molecular dynamics, and the thermodynamic occupancy limit at different (T,p) conditions has been computed using hydrid Grand-Canonical Monte Carlo/Molecular Dynamics. The results show that hydrogen diffusion from an external reservoir is limited by interfacial phenomena in the clathrate surface, and also that multiple guest occupancy and its distribution can be computed using the described approach.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP
•Simulations and conductivity measurements of ionic liquid-ethanol mixtures.•Differences compared with aqueous mixtures in both structure and dynamics.•Agreement between computational and ...experimental ionic.•Data accurately described by the pseudo-lattice random-alloy model.•Consistency between microscopic jumping frequencies and macroscopic conductivity.
We study the microscopic structure and transport properties of ions in mixtures of 1-butyl-3-methylimidazolium and 1-butyl-3-ethylimidazolium iodide with ethanol using atomistic molecular dynamics simulations and conductivity measurements. Compared with the same ionic liquids in water, we reveal essential differences in ionic structure that are closely related to the differences in the solubility mechanisms of both types of solvents. In particular, unlike for aqueous solutions, we find a homogeneous distribution of solvent molecules in the system, i.e., we observe no cluster formation, which agrees with the nano-structured solvation paradigm. In addition, we calculate the conductivities of these systems in the whole concentration range and compare them with experimental data. Although the simulated values slightly underestimate the experimental ones, they reproduce the shape of the experimental conductivity dome reasonably well. We also show that the pseudo-lattice random-alloy model, which is based on microscopic ion jumping frequencies, describes the conductivity data accurately. We compute the average jumping frequencies directly from simulations and find that they agree well with those obtained by fitting the simulation conductivity data. These results show that the pseudo-lattice random-alloy model provides a valuable tool to describe the conductivities of ionic liquid–solvent mixtures and particularly their concentration dependence. It shall also apply to other systems, e.g., inorganic electrolytes and dispersed ionic conductors.
Full text
Available for:
GEOZS, IJS, IMTLJ, KILJ, KISLJ, NLZOH, NUK, OILJ, PNG, SAZU, SBCE, SBJE, UILJ, UL, UM, UPCLJ, UPUK, ZAGLJ, ZRSKP
PDF
