Universal Quasichemical Functional Group Activity Coefficient Research Articles

Multi-objective optimization for inherently selective atomic layer deposition of zirconia on silicon and not on copper

Atomic layer deposition (ALD) deposits uniform, pinhole free metal or metal oxide films via a self-limiting surface reaction. Area-selective ALD allows selective deposition of these films on targeted areas. Other than using masking to prevent deposition on undesired areas of the substrate, the process parameters can be altered to achieve selectivity. However, determination of whether an ALD process is selective under the chosen reaction conditions requires rigorous experiments to be performed. ALD is a slow process and requires expensive organometallic complexes known as metal precursors. Simulation models built on the experimental data can eliminate the need for performing multiple experiments by predicting reaction conditions that promote selective growth thus saving time and cost. A little or no information was found in literature on models that can predict selectivity window of an ALD process by looking at process parameters. Our inhouse experimental data obtained from the inherently selective atomic layer deposition (ISALD) of zirconia on silicon and not on copper were used to build a simulation model that predicts selectivity of the ISALD process over a wide range of temperature. Multi-objective optimization algorithm was developed to find optimal process condition that can maximize the deposition rate on silicon while minimizing the same on copper. The theoretical framework is based on the equilibrium reached at the substrate surface after chemisorption of the precursor molecules. The chemisorption process is influenced by the interactions of the groups present in the precursor molecules and on the substrate surface. Universal Quasi-chemical Functional-group Activity Coefficient method for group contribution was used to quantify the interactions between different groups. Further, we show that the same group interaction parameters can be used with a different zirconium precursor having same functional groups to predict growth characteristics of zirconia on silicon under similar reaction conditions.

Surface tension of binary and ternary mixtures mapping with ASP and UNIFAC models based on machine learning

Modeling predictions of surface tension for binary and ternary liquid mixtures is difficult. In this work, we propose a machine learning model to accurately predict the surface tension of binary mixtures of organic solvents-ionic liquids and ternary mixtures of organic solvents-ionic liquids–water and analytically characterize the proposed model. In total, 1593 binary mixture data points and 216 ternary mixture data points were collected to develop the machine learning model. The model was developed by combining machine learning algorithms, UNIFAC (UNIversal quasi-chemical Functional group Activity Coefficient) and ASP (Abraham solvation parameter). UNIFAC parameters are used to describe ionic liquids, and ASP is used to describe organic solvents. The effect of each parameter on the surface tension is characterized by SHAP (SHapley Additive exPlanation). We considered support vector regression, artificial neural network, K nearest neighbor regression, random forest regression, LightGBM (light gradient boosting machine), and CatBoost (categorical boosting) algorithms. The results show that the CatBoost algorithm works best, MAE = 0.3338, RMSE = 0.7565, and R2 = 0.9946. The SHAP results show that the surface tension of the liquid decreases as the volume and surface area of the anion increase. This work not only accurately predicts the surface tension of binary and ternary mixtures, but also provides illuminating insight into the microscopic interactions between physical empirical models and physical and chemical properties.

Predicting viscosity of ionic liquids - water mixtures by bridging UNIFAC modeling with interpretable machine learning

Predicting the viscosity of ionic liquids (ILs)-water mixtures precisely is considerable for diversity applications in chemical industries. In this work, interpretable machine learning models incorporate physics information developed to link the UNIversal quasi-chemical Functional group Activity Coefficient (UNIFAC) models as new predictive tools for the viscosity of ILs-water mixtures. Machine learning algorithms including Adaptive Boosting (AdaBoost), Categorical Boosting (CatBoost), Extreme Gradient Boosting (XGBoost), K-Nearest-Neighbour Regression (KNNR) and Random Forest (RF) are compared, with the Catboost model achieving the best accuracies (RMSE = 0.0084, MSE = 0.0001, MAE = 0.0020, R2 = 0.9941). We demonstrate that the UNIFAC model and the Stokes-Einstein relation assisted to reduce the feature dimensionality, and improve the predictive power of the viscosity of ILs-water mixtures. The describe of the Shapley’s additive explanations (SHAP) and Partial dependence plots (PDP) are giveing expression to the features importance of the UNIFAC model. Our work created a new paradigm to hyperlink the machine learning models with ILs-related properties, the UNIFAC model contains prior physical information to optimize the features of the experiment, and fill the gap of the input features and the viscosity of ILs-water mixtures.

Separation of methanol and ethanol from azeotropic MTBE mixtures using choline chloride-based deep eutectic solvents

In this study, the performance of five different choline chloride – based deep eutectic solvents (DESs), including ChCl – urea, ChCl – ethylene glycol, ChCl – glycerol, ChCl – 1,2- propanediol, and ChCl – glucose, has been experimentally and computationally investigated for the purpose of liquid – liquid extraction at 298.15 K and 1 atm. In this regard, the effect of them on the extraction efficiency to separate methanol and ethanol from azeotropic systems, methyl tert-butyl ether (MTBE) + ethanol and MTBE + methanol, is studied. To synthesize these DESs, chlorine chloride as a hydrogen bond acceptor (HBA) is blended with urea, ethylene glycol, glycerol, 1,2- propanediol, and glucose as hydrogen bond donors (HBDs). Through the investigation, it is revealed that all the studied DESs deliver a distribution coefficient, β, that is higher than unity (β > 1). Similarly, the measured selectivities for these DESs have presented values that are far greater than unity (S ≫ 1). This is advantageous as it demonstrates their industrial applications. Furthermore, to have a better view of how effective they are in terms of the extraction capacity, the experimental data have been compared to the performance of ionic liquids used for the liquid – liquid extraction of MTBE + methanol. The reliability of the tie-line compositions was checked by applying the Othmer–Tobias equation. Additionally, to show the consistency, UNIQUAC Functional-group Activity Coefficients (UNIFAC) as a thermodynamic model has been employed to predict the equilibrium tie-lines. The predicted data is compared with the experimentally measured data through the calculation of the root-mean-square deviation (RMSD), showing that both predicted and experimental data are in a good agreement together.

Thermodynamic analysis, synthesis, characterization, and evaluation of 1-ethyl-3-methylimidazolium chloride: Study of its effect on pretreated rice husk

This work is focused on the determination of the optimal reaction conditions to synthesize the ionic liquid 1-ethyl-3-methylimidazolium chloride ([EMIM][Cl]) and assess its suitability for the pretreatment of rice husk. The modified UNIFAC (UNIversal quasi-chemical Functional-group Activity Coefficients) approach for ionic liquids is used to develop a thermodynamic model that describes the reactive system methylimidazole (MIM), chloroethane (C2H5Cl) and [EMIM][Cl]. The model allows to study the phase equilibria coexistence (vapor–liquid equilibria and solid–liquid equilibria) and yields the theoretically optimal conditions to synthesize the ionic liquid. The model predictions are validated with the available experimental and reported data. By implementing the developed model, a simple way to synthesize ionic liquid [EMIM][Cl] was found allowing to study its influence on the structure and morphology of pretreated rice husk. The lignocellulosic materials involved in this study are characterized by their composition, enzymatic digestibility, scanning electron microscopy, and crystallinity. Compared to untreated material, [EMIM][Cl]-pretreated rice husk produces cellulose that can be efficiently enzymatic hydrolyzed with high sugar yields. This work offers a suitable methodology to include the synthesis and thermodynamics of the solvent media within the design of low-cost ionic liquids for lignocellulosic biomass pretreatment.

Flash point study of ternary mixtures comprising binary constituents that exhibit maximum flash point behavior and minimum flash point behavior

The flash point (FP) behaviors of two ternary blends were investigated in this study: propyl butyrate + acetic acid + 1-butanol and propyl butyrate + acetic acid + 2-pentanol. The mentioned blends contain one binary constituent with maximum FP behavior (MaxFPB) and one binary constituent with minimum FP behavior (MinFPB). The maximum and minimum FP values are not observed in the ternary constituents of the aforementioned blends. Therefore, these mixtures are the first to be classified into the fp 3.2.0−1 FP class. The Wilson, nonrandom two-liquid (NRTL), NRTL+Ideal, original universal quasi-chemical functional-group activity coefficients (UNIFAC), and UNIFAC-Dortmund models were employed to anticipate the activity coefficients of the blends. The two employed UNIFAC-based models exhibited poor performance in predicting the FPs of binary blends containing acetic acid. The findings of this study are applicable for fire and explosion hazard identification, as well as fuel design, particularly biofuel design.

Mathematical Analysis of Discontinuous Rectification Columns at Pilot Scale Based on the Continuous Stable States Concept and MESH Equations

Mathematical analysis and simulation of a discontinuous rectification column was performed using an operational strategy during the start-up before reaching a pseudo-stable state in discontinuous operation. The mathematical model was formulated focusing on the equilibrium state (ES) and implementing MESH equations (M: Mass balance, E: Equilibrium thermodynamics, S: Stoichiometry relations, H: Enthalpy or heat balance) to provide solutions using the Thomas method and the Wang-Henke algorithms internally coupled to the Fourth Order Runge-Kutta method. The results were validated with experimental data from a distillation column at a pilot scale using an ethanol-water system with an equilibrium behavior described by the UNIQUAC Functional-group Activity Coefficients (UNIFAC) and Predictive Soave-Redlich-Kwong (PSRK) thermodynamic models with a global error of 1.84%. The molar ethanol concentrations presented deviations from the mathematical model predictions from 1.51% to 0.02%, with a global mean error of 0.48%. A mean error of 0.055% was obtained for the temperature profile of the column, thus demonstrating the effectiveness of the solution and its convergence capacity. The solution based on the Thomas method and the Wang-Henke algorithms coupled to the Runge-Kutta method made it possible to describe the behavior and variables of all stages of the distillation column. Operation at total reflux from start-up avoids wasting product and allows for the stabilization of the state variables, such as temperature and molar composition.

Evaluation of the adequacy of phase equilibria modeling based on various sets of experimental data

Objectives. The purpose of the paper is to compare the adequacy of mathematical models of vapor–liquid equilibrium (VLE) and their ability to reproduce the phase behavior of the ternary system benzene–cyclohexane–chlorobenzene using different experimental data sets to evaluate binary interaction parameters.Methods. The research methodologies were mathematical modeling of VLE in the Aspen Plus V.10.0 software package using activity coefficient models (Non-Random Two-Liquid (NRTL), Wilson) and the Universal quasichemical Functional-group Activity Coefficients (UNIFAC) group model, which allows for independent information. For the benzene–cyclohexane–chlorobenzene ternary system, the use of the NRTL equation is warranted because it provides a better description of the VLE experimental data.Results. The diagram construction of the constant volatility of cyclohexane relative to benzene lines revealed three topological structures. Only one of them can be considered reliable because it corresponds to the experimental data and coincides with the UNIFAC model diagram constructed based on independent UNIFAC model data. The results indicate that to study systems containing components with similar properties, it is necessary to improve the description quality of the available data sets (the relative error should not exceed 1.5%).Conclusions. The reproduction of the thermodynamic features of various manifolds in the composition simplex obtained by processing direct VLE data can be used to supplement the adequacy of the model. For the cyclohexane–benzene–chlorobenzene system, the best NRTL equation parameters are those regressed from the extensive experimental VLE data available in the literature for the ternary system as a whole.

Vapor-liquid equilibrium in methanol + water system and modeling from 298.15 to 373.15 K

Vapor-liquid equilibrium (VLE) relationships are needed in the solution of many engineering problems. The required data can be found by performing an experiment. But such measurements are not so easy, even for binary systems. The fundamental idea of a solution- of- groups model is to utilize existing phase equilibrium data for predicting phase equilibria of systems for which no experimental data are available. The popular predictive method of industries called the UNIFAC (Universal Quasi-Chemical Functional Group Activity Coefficient) group contribution model has been found to be successful to represent the available experimental data of VLE for wide variety of mixtures at various values of temperatures and pressures. In this work, the vapour-liquid equilibrium (VLE) for the methanol (1) + water (2) system has been investigated at temperatures ranging from 298.15 K to 373.15 K. The VLE data of methanol (1) + water (2) mixture for this temperature range have been also regressed using UNIFAC model available in data pro ii. The thermodynamic consistency test of experimental data such as point test has been also performed. The phase equilibrium data generated in this work are relatively good in agreement with previous data reported in the literature.

Flash point investigation of ternary mixtures of 1-butanol/2-pentanol + acetic acid + ethylbenzene

Flash point information is crucial for evaluating the fire and explosion (F&E) hazards of flammable liquids. Most flash point studies have investigated binary mixtures; thus, their results are inapplicable to commercial mixtures of several liquids. In this study, two ternary mixtures, namely 1-butanol − acetic acid − ethylbenzene and 2-pentanol − acetic acid − ethylbenzene, were investigated. These mixtures comprise two binary constituents with minimum flash point behavior (MinFPB) and one binary constituent with maximum flash point behavior (MaxFPB). The ternary constituents do not exhibit MinFPB or MaxFPB; therefore, the aforementioned ternary mixtures can be considered the first known members of the fp 3.3.0−1b class (Da Cunha et al., 2018). A model developed for ternary miscible mixtures was used to predict the flash point behaviors of the mixtures. The nonrandom two-liquid (NRTL), Wilson, original universal quasi-chemical functional-group activity coefficients (UNIFAC) and UNIFAC-Dortmund models were used to predict the activity coefficients of the binary and ternary constituents. The mixture of 1-butanol or 2-pentanol with acetic acid exhibited MaxFPB, indicating that acetic acid can substantially reduce the F&E risks of these alcohols. The findings can be applied in hazard identification, F&E hazard reduction, process safety design, fuel design, and the risk management of ignitable waste liquids.

Modeling, simulation, and optimization for producing ultra-low sulfur and high-octane number gasoline by separation and conversion of fluid catalytic cracking naphtha

Contradiction between the desulfurization and loss of octane number (ON) for fluid catalytic cracking (FCC) naphtha is the long-term problem of clean gasoline production. Moreover, ON loss is substantially aggravated during deep hydrodesulfurization because of the global limited supply of sulfur for gasoline. This paper introduced a novel process for ultra-clean gasoline production that solves the above problem by using a unique separation approach. The components were separated by the coupling of distillation and solvent extraction, which the separation efficiency depends on the relationship between solvent and feedstock components. However, the compositions of FCC naphtha are complicated and always changing due to different processing technologies. Therefore, changes in the material composition must be considered during production. In this study, ternary liquid–liquid equilibrium (LLE) was explored to investigate the interaction of olefin, sulfide, and sulfolane solvent. A universal quasi-chemical functional group activity coefficient (UNIFAC) group contribution model was selected to predict the product composition when the FCC naphtha was treated by this process. The lack and necessity of group interaction parameters for the prediction were obtained through the regressive calculation of LLE results. A simulation method was established by Aspen Plus, and the reliability was well illustrated. Finally, the high olefin content of FCC naphtha critical components was analyzed by this simulation method. The sulfur and olefin levels of the product were 9.0 mg/kg and 18.6 wt%, respectively. The operation conditions simulated using the proposed technique met the requirements for the latest commercial gasoline worldwide.

Dispersions of Zirconia Nanoparticles Close to the Phase Boundary of Surfactant-Free Ternary Mixtures.

The achievement of a homogeneous dispersion of nanoparticles is of paramount importance in supporting their technological application. In wet processing, stable dispersions were largely obtained via surfactant or surface functionalization: although effective, the use of dispersant can alter, or even impair, the functional properties of the resulting nanostructured systems. Herein, we report a novel integrated modeling and experimental approach to obtain stable ZrO2 nanoparticle (NP) dispersions at native dimensions (about 5 nm) in homogeneous ternary mixtures of solvents (i.e., water, ethanol, and 1,2-dichlorobenzene) without any further surface functionalization. A miscibility ternary diagram was computed exploiting the universal quasi-chemical functional-group activity coefficient (UNIFAC) model, which was then experimentally validated. Dynamic light scattering (DLS) on these mixtures highlights that nanometric structures, resembling nanoemulsion droplets, form close to the mixture two-phase boundary, with a size that depends on the ternary mixture composition. ZrO2–NPs were then synthesized following a classic sol–gel approach and characterized by XRD and Raman spectroscopy. ZrO2–NPs were dispersed in HCl and mixed with different mixtures of ethanol and 1,2-dichlorobenzene (DCB), obtaining homogeneous and stable dispersions. These dispersions were then studied by means of DLS as a function of DCB concentration, observing that the nanoparticles can be dispersed at their native dimensions when the mass fraction of DCB was lower than 60%, whereas the increase of the hydrophobic solvent leads to the NPs’ agglomeration and sedimentation. The proposed approach not only offers specific guidelines for the design of ZrO2–NPs dispersions in a ternary solvent mixture but can also be extended to other complex solvent mixtures in order to achieve stable dispersions of nanoparticles with no functionalization.

Artificial neural network approach for predicting blood brain barrier permeability based on a group contribution method

Background and ObjectiveThe purpose of this study was to develop a quantitative structure-activity relationship (QSAR) model for the prediction of blood brain barrier (BBB) permeability by using artificial neural networks (ANN) in combination with molecular structure and property descriptors. MethodsUsing a database composed of 300 compounds, 52 structure descriptors obtained based on the universal quasichemical functional group activity coefficients (UNIFAC) group contribution method and the selected 8 molecular property descriptors were used as the network inputs, whereas logBB values of compounds constituted its output. ResultsThe correlation coefficient R of the constructed prediction model, the relative error (RE) and the root mean square error (RMSE) was 0.956, 0.857, and 0.171, respectively. These indicators reflected the feasibility, robustness and accuracy of the prediction model. Compared with the previously published results, a significant improvement in the predictions of the proposed ANN model was observed. ConclusionsANN model based on the group contribution method could achieve a satisfactory performance for logBB prediction.

Isolation of potential α-glucosidase inhibitor from Inonotus obliquus by combining ultrafiltration-liquid chromatography and consecutive high-speed countercurrent chromatography.

Inonotus obliquus is a rare medicinal fungus that contains several potential therapeutic ingredients. In this study, the α-glucosidase inhibitory activity of I. obliquus was examined, and a potential α-glucosidase inhibitor, (E)-4-(3,4-dihydroxyphenyl)but-3-en-2-one, was isolated from the I. obliquus extract through ultrafiltration-liquid chromatography (UF-LC). Consecutive high-speed countercurrent chromatography (HSCCC) was used for separation to obtain large quantities of the target compound. The universal quasi-chemical functional group activity coefficient (UNIFAC) model was utilized to prepare a two-phase solvent system, n-hexane/ethyl acetate/ethanol/water (4 : 4.5 : 3.5 : 5, v/v/v/v), wherein the proportions of n-hexane/ethyl acetate/ethanol/water in the stationary and mobile phases were 19.8 : 19.7 : 7.9 : 2.2 (v/v/v/v) and 1 : 16.4 : 57.5 : 136.6 (v/v/v/v), respectively. A flow rate of 2.5 mL min-1 and a column speed of 860 rpm were maintained. Consequently, 10.3 mg of the target compound (95.9% purity) was obtained from 900 mg (6 × 150 mg) of the I. obliquus extract. The use of the UNIFAC model, in combination with consecutive HSCCC separations, allows the purification of large quantities of samples over a short time. Furthermore, the volume of the organic solvent required is reduced. Thus, UF-LC is an effective technique for screening potential α-glucosidase inhibitors isolated from I. obliquus. This can ultimately aid in the discovery of bioactive compounds for the prevention and treatment of diabetes.

Machine Learning Assisted Discrimination of 20 Volatile Organic Compounds with Response Prediction Via UNIFAC Modeling

Accurate detection of volatile organic compounds (VOCs) in complex background plays an important role in both natural and human-related activities, such as environmental protection, medical screening, food safety, and homeland security. Reliable monitoring of atmospheric VOCs is vital because vapor exposure can accelerate air-borne toxic generation and lead to tropospheric air pollution. While conventional analytical instruments such as gas chromatography-mass spectrometry (GC-MS) are largely used for identification of unknown VOCs, it is desirable for sensing technology to transition into a low cost, energy-efficient, and miniaturized media that can be implemented into personal devices.As one of the fastest growing types of chemical sensors, chemiresistors possess many desirable characteristics such as straightforward fabrication and easy to integrate. Sensing responses for nanoparticle-based chemiresistor are based on the sorbent-phase interaction of an analyte with the sensing materials, which can be explained by a thermally activated electron tunneling mechanism. The morphology of the gold-nanoparticle (AuNP) assembly (i.e. the interparticle distance) between the electrodes determines the baseline resistance of that individual sensor element.In this work, we demonstrate the detection and classification of 20 VOC vapors at room temperature using a multi-AuNP chemiresistor array fabricated with gold nanoparticles with chemically diverse ligands, including 4-(dimethylamino)pyridine (DMAP), tetradecylamine (TDA), and 4-aminothiophenol (4-ATP). The castellated electrode structures were fabricated using standard photolithography and liftoff processes. Classification accuracy was determined using a machine learning algorithm on a matrix of sensing response data. Supervised classifiers applied in this work include linear discriminant analysis (LDA), support vector machine (SVM), k-nearest neighbors (KNN), and random forest (RF). The results demonstrated that pattern recognition accuracy increases when a combination of ligands was used, as compared to sensors with one type of ligand. This proof-of-concept study illustrated that we were able to improve classification accuracy by choosing ideal sensing elements with optimized built-in redundancy. Meanwhile, a vapor sorption model was established to predict the chemiresistive sensor responses. One of the biggest challenges of AuNP-based chemiresistive sensor technology is finding the best candidates of materials that provide sufficient resolving power to differentiate chemicals in complex surroundings for molecular pattern recognition. To evaluate the sensor performance of ligand-analyte interactions, a UNIFAC (UNIQUAC Functional-group Activity Coefficients) model was established in this study to predict the sorption between a certain ligand material and VOC vapors. Therefore, a more performance-efficient sensing element library can be established for different types of analytes. Specifically, partition coefficients were determined from the UNIFAC method, and the sorbent phase concentration was obtained from modified Raoult’s Law. To validate our model, simulated UNIFAC partition coefficients were compared with experimental results and data reported from literature. Besides sensor responses, other characteristic parameters of chemiresistor performance are also studied, including sensitivity, sensitivity ratio, volumetric sensitivity, and molar sensitivity. This work illustrated the versatility of gold nanoparticle-based chemiresistor array systems for VOC detection, discrimination, and sensor signal prediction. Figure 1

Choline chloride-based deep eutectic solvents as green extractant for the efficient extraction of 1-butanol or 2-butanol from azeotropic n-heptane + butanol mixtures

In this study, choline chloride-based deep eutectic solvents are used as green extractant for the efficient extraction of 1-butanol or 2-butanol from azeotropic n-heptane + butanol mixtures at 298.15 K and atmospheric pressure. Choline chloride as hydrogen bond acceptor (HBA) is mixed with urea, ethylene glycol, glycerol, and 1,2-propanediol (propylene glycol) as hydrogen bond donors (HBDs) to study the effect of different choline chloride-based deep eutectic solvents on the extraction efficiency. The obtained results show that, although all the DESs provide solute distribution ratios smaller than unity (β < 1), the measured experimental selectivity values are much higher than unity (S> > 1), which is of interest for the economic feasibility of industrial extraction operations. The extraction capacity of the studied deep eutectic solvents for the separation of n-heptane +1-butanol mixture is compared with seven commonly used imidazolium-based ionic liquids. The experimental results are compared to UNIQUAC Functional-group Activity Coefficients (UNIFAC) model. It is visually observed that UNIFAC model predicts satisfactorily the equilibrium tie-lines, where the root-mean-square deviation (RMSD) for extraction of 1-butanol by ChCl-urea, ChCl-ethylene glycol, ChCl-glycerol and ChCl-1,2-propanediol are 2.99, 4.38, 1.12, and 1.56, respectively. In the case of 2-butanol extraction using the mentioned DESs, the RMSD is 0.69, 3.13, 1.47, and 1.11, respectively.

Assessment of the SM12, SM8, and SMD Solvation Models for Predicting Limiting Activity Coefficients at 298.15 K

The SMx (x = 12, 8, or D) universal solvent models are implicit solvent models which using electronic structure calculations can compute solvation free energies at 298.15 K. While solvation free energy is an important thermophysical property, within the thermodynamic modeling of phase equilibrium, limiting (or infinite dilution) activity coefficients are preferred since they may be used to parameterize excess Gibbs free energy models to model phase equilibrium. Conveniently, the two quantities are related. Therefore the present study was performed to assess the ability to use the SMx universal solvent models to predict limiting activity coefficients. Two methods of calculating the limiting activity coefficient where compared: (1) the solvation free energy and self-solvation free energy were both predicted and (2) the self-solvation free energy was computed using readily available vapor pressure data. Overall the first method is preferred as it results in a cancellation of errors, specifically for the case in which water is a solute. The SM12 model was compared to both the Universal Quasichemical Functional-group Activity Coefficients (UNIFAC) and modified separation of cohesive energy density (MOSCED) models. MOSCED was the highest performer, yet had the smallest available compound inventory. UNIFAC and SM12 exhibited comparable performance. Therefore further exploration and research should be conducted into the viability of using the SMx models for phase equilibrium calculations.

Transesterification of triolein and methanol with Novozym 435 using co-solvents

To enhance the reaction rate and yield during biodiesel synthesis with Novozym 435 (an immobilized lipase), a co-solvent method was applied, after which the relationships between reaction time and reaction yield during the single-phase enzymatic transesterification of triolein and methanol in the presence of co-solvents such as tetrahydrofuran, acetone, and hexane were investigated. The addition of hexane led to inactivation of Novozym 435, whereas in the presence of acetone or tetrahydrofuran, the reaction was accelerated and the yield was higher than with a conventional solvent-free enzymatic reaction. There was a clear difference in the dispersion of Novozym 435 particles between with and without co-solvents. The co-solvents prevented particle aggregation from occurring during the reaction without co-solvent. This is the likely reason for the enhanced reaction yield in the presence of co-solvents. To better understand the difference among the co-solvents employed, we estimated phase compositions during transesterification using the LLE (liquid-liquid equilibrium)-UNIFAC (UNIversal quasichemical Functional group Activity Coefficients) model, and examined the effect of these compositions on the reaction and dispersion of the glycerin by-product phase. Using the model, we also accounted for why hexane led to inactivation by estimating the transfer free energy of methanol, glycerin, and water between the initial bulk reaction solution and the water layer on the enzyme surface.

Prediction of liquid-liquid equilibria of multicomponent fatty systems using the ASOG method

Purification steps, as solvent extraction, play an important role in biodiesel and edible oil processing. Their continuous improvement relies on predictive tools for computing liquid-liquid equilibria (LLE), and among these tools, group contribution methods, such as the UNIQUAC Functional-group Activity Coefficients (UNIFAC) method and its variants, and the Analytical Solutions of Groups (ASOG) method, are commonly used. In this work the ASOG method was evaluated for predicting LLE of 372 multicomponent fatty systems from the literature. The ASOG method, although quite unexplored, in particular for fatty systems, performed satisfactorily well, resulting in a global deviation of 9.27%. The Solution Structure of Isoactivity Equations (SSIE), a recently published procedure for computing LLE based on the solution of the isoactivities was adopted as the calculation strategy. The predictive capacity of the ASOG method was also improved by means of readjustment of interaction parameters related to functional groups of fatty acid (COOH), ester (COO), aldehydes (COH), water (H2O), and unsaturation (CC). The new parameters allowed prediction of solubility of fatty acid, alcohol and aldehyde in both fatty and solvent phases more precisely, diminishing deviations between calculated and experimental compositions.

Flexible heuristic algorithm for automatic molecule fragmentation: application to the UNIFAC group contribution model

A priori calculation of thermophysical properties and predictive thermodynamic models can be very helpful for developing new industrial processes. Group contribution methods link the target property to contributions based on chemical groups or other molecular subunits of a given molecule. However, the fragmentation of the molecule into its subunits is usually done manually impeding the fast testing and development of new group contribution methods based on large databases of molecules. The aim of this work is to develop strategies to overcome the challenges that arise when attempting to fragment molecules automatically while keeping the definition of the groups as simple as possible. Furthermore, these strategies are implemented in two fragmentation algorithms. The first algorithm finds only one solution while the second algorithm finds all possible fragmentations. Both algorithms are tested to fragment a database of 20,000+ molecules for use with the group contribution model Universal Quasichemical Functional Group Activity Coefficients (UNIFAC). Comparison of the results with a reference database shows that both algorithms are capable of successfully fragmenting all the molecules automatically. Furthermore, when applying them on a larger database it is shown, that the newly developed algorithms are capable of fragmenting structures previously thought not possible to fragment.