Search

Activity of water and electrolytes in aqueous solutions is of utmost importance for multiple industrial applications. However, experimental determination of such values is time-consuming, while calculation of activity coefficients using numerical methods is challenging. By training neural networks models on literature data, one could predict activity of water and electrolytes easily, without requiring any experiment. In this paper, multiple descriptors (or features) are compared to predict activity coefficients of electrolytes and activity of water in electrolyte solutions. A neural network based on the Levenberg-Marquardt algorithm (LM-NN) showed high accuracy to calculate values, despite the small size of the training datasets. Both activity coefficients of electrolytes and activity of water in electrolyte solutions can be predicted accurately even on unseen data, using simple descriptors such as electrolyte concentration, ion sizes and charges. However, some discrepancies were observed due to the lack of representativeness of the training dataset. This could be solved by selecting training data sets that are similar (e.g. same group of the periodic table) to the unknown values, or by including available experimental data for the salt considered. The ability of the LM-NN to solve non-linear least square curve fitting problems makes it a good candidate to fit experimental activity coefficient data, with the advantage of simplicity as compared to e-NRTL or UNIQUAC methods. This method paves the way for accurate and quick determination of thermodynamic data for electrolyte solutions (and beyond) using machine learning, without necessitating large training datasets.

The ionic association constants of sodium with carbonate ion (K1C′) and acidic carbonate ions (K2C′) were measured in NaCl-NaHCO3-H2O ternary systems to determine the distribution of sodium among the chemical species present in the growth medium of Chlorella homosphaera 424 algae. The mean activity coefficients of sodium chloride (in pure sodium chloride and in a mixture of electrolytes) were determined experimentally using two electrochemical cells, namely Ag, AgCl| KCl (3 M)|| NH4NO3 (1 M)| NaCl (mNaCl)| Na+-ISE and Ag, AgCl|KCl (3 M)|| NH4NO3 (1 M)| NaCl (mNaCl)| Cl−-ISE. The studies carried out show that the values of the association constants of K1C′ and K2C′ do not depend on the composition of the medium, but only on the effective ionic strength. The experimentally obtained γNaCl0 values in the binary system are comparable to the mean activity coefficients values for NaCl, calculated using data from the literature, with −0.9 to 0.1% relative standard deviation. The obtained results show that the experimentally determined mean activity coefficient in the ternary system, γNaCl, is smaller than γNaCl0 in the binary system over the entire field of ionic strengths studied. The ternary system NaCl-NaHCO3-H2O obeys Harned’s rule.

In previous work we developed a new statistical method for calculating the individual activities of ions including the association of ions. Here we study multi-particle electrostatic interactions connected within higher cluster integrals and identify the ionization constants of the mass action law of associating ion clusters. In contrast to Bjerrum and Fuoss, our concept of association is not based on spatial criteria, but instead on the strength of interaction measured in powers of the Bjerrum parameter (e^2/(D_0)(k_B)Ta; a is contact) and defined by asymptotic properties of the cluster integrals. For ion pair formation our mass action constant is the classical counterpart of Planck's famous hydrogenic partition function. As a rule, the new association constants are smaller than traditional expressions, e.g., by Fuoss and Kraus, in the interesting regions of interaction parameters about fifty percent. Several examples including CaCl_2, MgCl_2, Na_2SO_4, K_2SO_4, LaCl_3 and a model of seawater are studied. For several associating electrolytes and seawater, reasonable agreement with experiments and Monte Carlo results is achieved.

Abstract Precursor solubility is a crucial factor in industrial applications, dominating the outcome of reactions and purification steps. The outcome and success of thermodynamic modelling of this industrially important property with equations of states, such as Perturbed‐Chain Statistical Associating Fluid Theory (PC‐SAFT), vastly depends on the quality of the pure‐component parameters. The pure‐component parameters for low‐volatile compounds such as ionic liquids (ILs) have been commonly estimated using mixture properties, e. g. the osmotic pressure of aqueous solutions. This leads to parameters that depend on the solvent, and transferability to other mixtures often causes poor modeling results. Mixture‐independent experimental properties would be a more suitable basis for the parameter estimation offering a way to universal parameter sets. Model parameters for ILs are available in the literature [10.1016/j.fluid.2012.05.029], but they were estimated using pure‐IL density data. The present work focuses on a step towards a more universal estimation strategy that includes new experimental vapor‐pressure data of the pure IL. ILs exhibit an almost negligible vapor pressure in magnitude of usually 10−5 Pa even at elevated temperatures. In this work, such vapor‐pressure data of a series of 1‐ethyl‐3‐methyl‐imidazolium‐based [C2mim]‐ILs with various IL‐anions (e. g. tetrafluoroborate [BF4]−, hexafluorophosphate [PF6]−, bis(trifluoromethylsulfonyl)imide [NTf2]−) were experimentally determined and subsequently used for PC‐SAFT parameter estimation. The so‐determined parameters were used to predict experimental molecular precursor solubility in ILs and infinitely diluted activity coefficients of various solvents in ILs. The parameters were further compared to modeling results using classical parametrization methods (use of liquid‐density data only for the molecular PC‐SAFT and the ion‐based electrolyte PC‐SAFT). As a result, the modeled precursor solubilities using the new approach are much more precise than using the classical parametrization methods, and required binary parameters were found to be much smaller (if needed). In sum, including the pure‐component vapor‐pressure data of ILs opens the door towards parameter estimation that is not biased by mixture data. This procedure might be suitable also for polymers and for all kind of ionic species but needs extension to ion‐specific parametrization in the long term.

Abstract The hydrogen bond network of different small alcohols is investigated via cluster analysis. Methanol/alcohol mixtures are studied with increasing chain length and branching of the molecule. Those changes can play an important role in different fields, including solvent and metal extraction. The extended tight binding method GFN2‐xTB allows the evaluation and geometry optimization of thousands of clusters built via a genetic algorithm. Interaction energies and geometries are evaluated and discussed for the neat systems. Thermodynamic properties, such as vaporization enthalpies and activity coefficients, are calculated with the binary quantum cluster equilibrium (bQCE) approach using our in‐house code Peacemaker 2.8. Combined distribution functions of the distances against the angles of the hydrogen bonds are evaluated for neat and mixed clusters and weighted by the equilibrium populations achieved from bQCE calculations.

The separation of volatile organic solvents (herewith, solutes) were tested in an imidazolium-based ionic liquid (1,3-dimethylimidazolium dimethylphosphate), [MMIM] [DMP] as a solvent at four different temperatures, T = (313 .15 to 343.15) K and at atmospheric pressure. The solutes include alcohols, aromatic hydrocarbons, thiophene, ketones, alkanes, alkynes, alkenes, acetonitrile, THF, as well as water. The separation potential of the investigated ionic liquid was based on the activity coefficients at infinite dilution, γ13∞, of organic solutes. This was based on the retention data acquired by gas liquid chromatography. Excess thermodynamic parameters at infinite dilution such as Gibbs free energy, ΔG1E,∞, entropy term, Tref ΔS1E,∞ and partial molar enthalpy, ΔH1E,∞ were determined from the γ13∞ data. The results are discussed in terms of intermolecular interactions occurring between the solutes under investigation and the ionic liquid. The effectiveness of 1,3-dimethylimidazolium dimethylphosphate to separate some of the separation problems discussed in this work is determined from the selectivity and capacity values calculated from the limiting activity coefficients at infinite dilution. These thermodynamic functions can be used to develop, design or optimise industrial extraction processes.

In this work, the solubility of vanillic, gallic, syringic, p-coumaric, ferulic and caffeic acids was determined at 37 °C under different conditions, namely pure water and two different ionic media, NaCl(aq) and NaClO4(aq), at different ionic strengths (i.e., 0.16, 0.50, 1.0, 2.0 and 3.0 M). The solubility in water of all the acids was found to be higher than that in both of the ionic media. Moreover, the solubility of hydroxycinnamic acids was lower than that of hydroxybenzoic acids. The activity coefficients of neutral species were calculated from these data; this knowledge is necessary when modeling the dependence of equilibrium constants on the ionic strength. Results obtained in this work can be useful for further studies regarding complex formation equilibria between these ligands and bioavailable metal cations.

The esterification of nopol with acetic acid to produce nopyl acetate using sulfuric acid as homogeneous catalyst was studied. The reactions were carried out in a batch reactor, at different temperatures (50, 60, 70 and 80 °C), catalyst concentrations (0.0184, 0.0275, 0.0367 and 0.0480 molL−1 and molar ratio of acetic acid to nopol (1: 1, 1:2, 1:3 and 1: 4); the chemical equilibrium composition was measured at those reaction conditions. It was found that the equilibrium composition is a weak function of temperature, equilibrium conversion was 63, 68, 71 and 75% at 50, 60, 70 and 80°C, respectively. The reaction was described with a simple power-law model with a second-order kinetic model for both the forward and the backward reaction, using concentration and activities which were predicted by the Universal Functional group Contribution (UNIFAC) method for considering non-ideal behavior of the liquid phase. The forward reaction rate and the equilibrium constants increased with temperature; the relation of the pre-exponential factor with the catalyst amount was evaluated. The activation energy and pre-exponential factor estimated for the forward reaction using the kinetic model based on concentration, were respectively, 28.08 kJmol−1 and 11126 Lmol−1h −1 with a concentration of catalyst of 0.0275 M. Using the kinetic model based on activities, the forward reaction rate constant was 28.56 kJmol−1 and the kfo,act was 33860 Lmol−1h −1 . The enthalpy (34.90 kJmol−1 ) and the entropy (0.12 kJmol−1K−1 ) of reaction were determined using van’t Hoff equation.

A new method using the headspace solid phase microextraction (HS-SPME) technique was used to evaluate the infinite dilute activity coefficient (γ1∞) in an alcohol/water/salt system. The studied systems were ethanol and water with NaCl and NH4Cl at salt concentrations of 5, 10, 15, and 30% m/v and temperatures of 303.15 and 313.15 K. The method was used to investigate the salt effect on vapor/liquid equilibrium in an ethanol/water system, yielding satisfactory results. The study focused on the rich side in ethanol. The data were compared with the literature infinite dilution data determined by other methods such as differential ebulliometry (EBUL), differential static cell equilibrium (STAT), and gas-liquid chromatography with no gas phase correction (GC). In this study, NaCl showed better separation rates than NH4Cl.

The aldol condensation of benzaldehyde and heptanal is taken as an example of reversible liquid phase organic reactions to show that inclusion of activity coefficients reveal distinct differences in conversion and product distribution when different solvents methanol, ethanol, n-propanol, or n-butanol are used. The purpose of this work is to show a pronounced solvent effect for a given set of identical kinetic parameters, i.e., the same liquid phase kinetics can result in different conversion and yield values, depending on the choice of solvent. It was shown that subsequent parameter estimation without inclusion of the activity coefficients resulted in a pronounced deviation from the ‘true’ kinetics, up to a factor of 30. It is proposed that the usage of average activity coefficients gives already a significant improvement, resulting in acceptable parameter estimates.

In this work a new method (SPME-GC/FID) was developed to analyze the activity of binary liquid mixtures. The purpose is to demonstrate that SPME is capable to be used to determinate activity coefficients at infinite dilution knowing the fiber properties, with a lower cost than the conventional methods encountered in literature such as GLC. The activity coefficients at infinite dilution in furfural for n-hexane, n-heptane and cyclohexane at 298.15 K was determined using SPME and deviations of literature data was about 7%.

Based on van't Hoff and Gibbs equations the thermodynamic functions Gibbs energy, enthalpy, and entropy of solution, mixing and solvation of acetaminophen in propylene glycol + water (PG + W) cosolvent mixtures, were evaluated from solubility data determined at several temperatures. The solubility was greater at 100% of PG at all temperatures studied. The solvation of this drug in the mixtures increases as the PG proportion is also increased founding a maximum at 70% of PG. From 10% up to 20% of PG and from 70% up to 100% of PG, entropy driving was found, while from pure water up to 10% of PG and from 20% up to 70% of PG, enthalpy driving was found. These facts can be explained in terms of water-structure loss, and a diminishing in the energy required for cavity formation in the solvent, for mixtures from 30% up to 70% of PG.

A model on a feed forward back propagation neural network was employed to calculate the isobaric vapourliquid equilibrium (VLE) data at 40, 66.67, and 101.32 �� 0.02 kPa for the methylcyclohexane toluene and isopropanol methyl isobutyl ketone binary systems, which are composed of different chemical structures (cyclic, aromatic, alcohol and ketone) and do not show azeotrope behaviour. Half of the experimental VLE data only were assigned into the designed framework as training patterns in order to estimate the VLE data over the whole composition range at the mentioned pressures. The results were compared with the data calculated by the two classical models used in this field, the UNIFAC and Margules models. In all cases the deviations the experimental activity coefficients and those calculated by the neural network model (NNET) were lower than those obtained using the Margules and UNIFAC models.

A model on a feed forward back propagation neural network was employed to calculate the isobaric vapour–liquid equilibrium (VLE) data at 40, 66.67, and 101.32 ±0.02 kPa for the methylcyclohexane – toluene and isopropanol – methyl isobutyl ketone binary systems, which are composed of different chemical structures (cyclic, aromatic, alcohol and ketone) and do not show azeotrope behaviour. Half of the experimental VLE data only were assigned into the designed framework as training patterns in order to estimate the VLE data over the whole composition range at the mentioned pressures. The results were compared with the data calculated by the two classical models used in this field, the UNIFAC and Margules models. In all cases the deviations the experimental activity coefficients and those calculated by the neural network model (NNET) were lower than those obtained using the Margules and UNIFAC models.

We report surface tension of two weakly segregating alloys Al-Ga and Cd-In in molten state at temperatures of 1023 K and 800 K respectively using different approaches. Our analysis based on different assumptions reveal that the metal with lower surface tension tends to segregate on the surface of molten alloy and the metal with higher surface tension tends to segregate in the bulk. Different approaches predict consistency in the values of the surface tension of Al-Ga liquid alloy that increases with increase in bulk concentration of aluminium in the alloy with all values smaller than the ideal values. In Cd-In alloy the models reveal no such regularity in the surface tension; it varies slightly from the ideality. DOI: http://dx.doi.org/10.3126/bibechana.v9i0.7183 BIBECHANA 9 (2013) 103-112

We report surface tension of two weakly segregating alloys Al-Ga and Cd-In in molten state at temperatures of 1023 K and 800 K respectively using different approaches. Our analysis based on different assumptions reveal that the metal with lower surface tension tends to segregate on the surface of molten alloy and the metal with higher surface tension tends to segregate in the bulk. Different approaches predict consistency in the values of the surface tension of Al-Ga liquid alloy that increases with increase in bulk concentration of aluminium in the alloy with all values smaller than the ideal values. In Cd-In alloy the models reveal no such regularity in the surface tension; it varies slightly from the ideality. DOI: http://dx.doi.org/10.3126/bibechana.v9i0.7183 BIBECHANA 9 (2013) 103-112