This work presents an extended Pitzer model for the simultaneous representation of ionic speciation and thermodynamic properties in the quaternary reciprocal Cu-Na-Cl-SO 4 -H 2 O system and its binary and ternary subsystems. The model was calibrated using literature water activity, osmotic coefficient, vapor pressure, and solid-liquid equilibrium data over 273.15-373.15 K for the binary and ternary subsystems, and at 298.15 K for the quaternary system. Explicit association equilibria were considered for CuSO 4 (aq), CuCl + , CuCl 2 (aq), NaSO 4- , and NaCl(aq), allowing the description of sulfate/chloride complexation effects in concentrated aqueous electrolytes. New Pitzer interaction parameters were obtained for the sodium-containing binary systems, selected ternary systems, and the quaternary reciprocal system. The temperature dependence of the solubility product of CuSO 4 & sdot;Na 2 SO 4 & sdot;2H 2 O(cr) was also determined from 273.15 to 333.15 K. The model reproduced the experimental data with low combined relative standard deviations and yielded lower AIC, AIC c , and BIC values than selected literature Pitzer parameterizations. Speciation calculations showed that increasing chloride concentration strongly promotes Cu(II) chloro-complexation. In the CuSO 4 + NaCl + H 2 O reciprocal composition at 298.15 K, CuCl + reached a maximum fraction of 87.8 % at m(NaCl) = 2.22 mol & sdot;kg -1 , followed by increasing formation of CuCl 2 (aq). The proposed formulation provides a consistent thermodynamic framework for concentrated sulfate/ chloride media relevant to copper hydrometallurgy.
Ferric phosphate for the production of LiFePO4 cathodes in lithium ion batteries is commonly obtained by a precipitation reaction between water-soluble iron species and orthophosphates. The kinetics and purity of the precipitate are dependent on its solubility. For this reason and inconsistencies in the solubility data, the solubility of ferric phosphate was currently investigated. The solid dispersed in water was characterized by TGA/DTGA, SEM, EDX and XRD analyses, revealing that it was amorphous FePO4·~1.6H2O. The solubility experiments were carried out in the pH ~1 to 10 at approximately 278, 298, 318 and 338 K and ionic strength lower than approximately 0.5 mol L−1. The solubility was determined on the basis of measurements of equilibrium concentration of soluble phosphorus by spectrophotometry. pH-solubility diagrams were made by solving a system of nonlinear algebraic equations involving a set of eight reactions at equilibrium. By combining this equilibrium model and an optimization method, the negative logarithm of the solubility product and the enthalpy of formation of amorphous FePO4·~1.6H2O were found to be 22.89 ± 0.81 and − 1918.90 ± 22.92 kJ mol−1 (α = 0.05 and ν ≥ 16). The results demonstrate no evidence of a statistically significant difference between the solubility of amorphous hydrated and amorphous anhydrous FePO4 (p = 0.501), but the solubility of both were greater than that of crystalline FePO4·2H2O (p < 0.001). The difference between the enthalpy of formation of amorphous FePO4·~1.6H2O and crystalline FePO4·2H2O was statistically significant (p = 0.019).
The solubilities of furosemide and pentoxifylline in aqueous low alcohol solutions, such as ethanol, isopropanol, methanol, and n-propanol, were experimentally obtained with the use of an isothermal shake-flask process. As stated by an XRD analysis, the crystals of furosemide and pentoxifylline did not exhibit any solvation or transition. Solubility data correlation was performed through using two models, namely modified van't Hoff-Jouyban-Acree and Jouyban-Acree. This resulted in average relative deviations of ≤6.02% and average-root-square deviations of ≤16.5 × 10−3. The approach of extended Hildebrand solubility (EHSA) was also worked to analyze the equilibration solubility data at 298.15 K, resulting in average relative deviation values of no higher than 6.66%. An inversion of the Kirkwood-Buff theory was further involved to investigate the preferential solvation of the two drugs. Blends with water concentrated compositions exhibited positive magnitudes for the pentoxifylline preferential solvation parameter, indicating that alcohol preferentially solvated pentoxifylline. Alcohols exhibited preferential solvation of furosemide within the entire alcohol composition ranges. The intermolecular interactions of pentoxifylline/furosemide with solvents were examined using molecular dynamics simulation, e.g. radial distribution function and solvation free energy, indicating the leading contribution of hydrogen bonding interactions of solute-solvent and solvent-solvent to the pentoxifylline and furosemide solubility in blended solvents. An examination of the entropy-enthalpy compensation revealed that the pentoxifylline crystal dissolution in blends was an entropy-driven and enthalpy process within alcohol-rich region. The enthalpy-driven compensation mechanism was demonstrated by the positive slopes of ΔsolH3,1+2o versus ΔsolG3,1+2o curves for furosemide, which encompassed the entire composition range of alcohols.
A physics-informed model for the dynamic viscosity of deep eutectic solvents (DES) is developed, requiring only SMILES molecular representations as input. The model is grounded in Eyring's transition-state theory for viscous flow and decomposes the activation Gibbs energy into five physically distinct contributions: hole formation (Fürth theory), hydrogen-bond network reorganization, electrostatic cage escape, dispersion interactions, and molecular shape effects. Each contribution is quantified through molecular descriptors computed from SMILES using the open-source RDKit library, with composition-dependent mixing rules derived from the underlying physics. The 45 universal parameters are calibrated against a curated database of 3440 experimental viscosity data points spanning 237 chemically diverse DES systems (84 hydrogen-bond acceptors, 94 hydrogen-bond donors), covering temperatures from 278 to 378 K and viscosities from 1.3 to 85,000 mPa·s. The model achieves R2 = 0.77 and RMSE = 0.35 log₁₀(mPa·s), with 88% of predictions within ±0.5 log units. Leave-one-component-out cross-validation demonstrates robust extrapolation to completely unseen molecular species (74% with RMSE <0.5). Physical soundness is confirmed: calibrated parameters reproduce experimental activation energies (R2 = 0.63), coefficient signs match theoretical expectations, and the hydrogen-bond contribution correctly ranks known DES families. An open-access predictive tool is provided.
