On facilitated computation of mesoscopic behavior of reaction‐diffusion systems

Various cellular and subcellular biological systems occur in the conditions where both reactions and diffusion take place. Since the concentration of species varies spatially, application of reaction‐diffusion master equation has served as an effective method to handle these complicated systems; yet solving these equation incurs a large CPU time penalty. Counter to the traditional technique of generating many sample paths, this article introduces a method which combines Grima's effective rate equation approach (Grima, J Chem Phys. 2010;133:3) with a linear operator formalism for diffusion to capture averaged species behaviors. The formulation also shows correct results at various choices of compartment sizes, which have been found to be an important factor that can affect accuracy of the final predictions (Erban, Chapman, Phys Biol. 2009;6:4). It is shown that the method presented allows the computation of the mesoscopic average of reaction‐diffusion systems at considerably accelerated rates (exceeding a thousand fold) over those based on sample path averages. © 2017 American Institute of Chemical Engineers AIChE J, 63: 5258–5266, 2017     

Dynamic modeling and validation of a biomass hydrothermal pretreatment process—a demonstration scale study

Hydrothermal pretreatment of lignocellulosic biomass is a cost effective technology for second generation biorefineries. The process occurs in large horizontal and pressurized thermal reactors where the biomatrix is opened under the action of steam pressure and temperature to expose cellulose for the enzymatic hydrolysis process. Several by‐products are also formed, which disturb and act as inhibitors downstream. The objective of this study is to formulate and validate a large scale hydrothermal pretreatment dynamic model based on mass and energy balances, together with a complex conversion mechanism and kinetics. The study includes a comprehensive sensitivity and uncertainty analysis, with parameter estimation from real‐data in the 178–185°C range. To highlight the application utility of the model, a state estimator for biomass composition is developed. The predictions capture well the dynamic trends of the process, outlining the value of the model for simulation, control design, and optimization for full‐scale applications. © 2015 American Institute of Chemical Engineers AIChE J, 61: 4235–4250, 2015     

Modeling of aggregation kernel using Monte Carlo simulations of spray fluidized bed agglomeration

The present work attempts to consider the microscopic mechanisms of spray fluidized bed agglomeration while modeling the macroscopic kinetics of the process. A microscale approach, constant volume Monte‐Carlo simulation, is used to analyze the effects of micro‐processes on the aggregation behavior and identify the influencing parameters. The identified variables, namely the number of wet particles, the total number of particles, and the number of droplets are modeled and combined in the form of an aggregation kernel. The proposed kernel is then used in a one‐dimensional population balance equation for predicting the particle number density distribution. The only fitting parameters remaining in the population balance system are the collision frequency per particle and a success fraction accounting for the dissipation of kinetic energy. Predictions of the population balance model are compared with the results of Monte‐Carlo simulations for a variation of significant operating parameters and found to be in good agreement. © 2014 American Institute of Chemical Engineers AIChE J, 60: 855–868, 2014     

Adsorption of glucose into zeolite beta from aqueous solution

Hydrophobic zeolites, including Ti‐ and Sn‐beta, have been found to adsorb and isomerize glucose into fructose. An experimental question has been the significance of the entropic contribution to the free energy of transfer of glucose from solution to zeolite. We here perform Gibbs ensemble Monte Carlo calculations to quantify the enthalpy, entropy, and free energy of transfer of glucose from the aqueous phase to the zeolite phase. We find that the entropic contribution is large and positive, nearly compensating for an unfavorable enthalpy of transfer in all‐silica zeolite beta. A significant component of the positive entropy of transfer from the aqueous phase to zeolite is the unstructuring of first coordination shell waters around glucose as it leaves the solution. © 2013 American Institute of Chemical Engineers AIChE J, 59: 3523–3529, 2013     

Simulations of crystal aggregation and growth: Towards correct crystal area

Simulating crystal growth and aggregation can provide insight into factors that control the final product properties. Classical models involve formation of a volume-equivalent single crystal upon aggregation. This approach does not preserve particle area, resulting in an inadequate model of supersaturation depletion. Alternatively, crystal area can be computed accurately by a Monte Carlo method where each primary particle of an aggregate is described in its full geometric complexity. However, the drawback of this method is its computational cost. Thus, a third method is introduced as a compromise, describing particles by their volume and area and preserving both upon aggregation. The so-obtained two-dimensional model requires growth rate expressions in volume and area. We provide a method for parametrizing these expressions so that total volume and area closely match the values obtained with the method based on full geometric information. The parameters depend on primary particle geometry and the amount of growth. © 2019 American Institute of Chemical Engineers AIChE J, 65: e16525, 2019     

Learning from the past: Are catalyst design principles transferrable between hydrodesulfurization and deoxygenation?

Molybdenum‐oxide (MoO₃) is a promising catalyst candidate for hydrodeoxygenation (HDO) of pyrolysis vapor or liquefaction products to renewable fuels or value‐added chemicals. Density functional theory is used to study the mechanism and active site requirements for HDO of furan over the MoO₃(010) facet and contrast our results with prior work on hydrodesulfurization (HDS) of thiophene over MoS₂ model catalysts. The potential energy diagram for HDO over a realistically terminated MoO₃(010) surface facet reveals that the elementary reaction steps for deoxygenation are facile, but the formation of oxygen‐vacancies is slow and endothermic. In general, HDO over MoO₃ and HDS over MoS₂ exhibit mechanistic similarities, which suggests that knowledge transfer from the mature HDS system to the emerging field of HDO catalysis is possible. For example, transition metal promotion of MoO₃ resulted in an improvement of the kinetics and thermodynamics of oxygen vacancy formation, similar to Co and Ni promotion of MoS₂. © 2018 American Institute of Chemical Engineers AIChE J, 64: 3121–3133, 2018     

Adsorption of nitrogen‐ and sulfur‐containing compounds on NiMoS for hydrotreating reactions: A DFT and vdW‐corrected study

Adsorption of 35 molecules, comprising of organonitrogen and organosulfur compounds and hydrocarbons relevant to hydrotreating, was studied on the nickel‐promoted metal edge of molybdenum sulfide catalysts using periodic DFT, accounting for van der Waals interactions in several cases. Basic molecules tend to adsorb via their nitrogen atoms directly on top of nickel atoms while nonbasic molecules adsorb via carbon atoms relatively weakly. Molecular size, electron density, and alkyl substitution affects binding at the GGA‐PW91 level of theory. van der Waals corrections influence adsorption geometry and lead to significant additional stabilization of adsorbates. The differential binding energy of nitrogen‐containing compounds decreases by 0.2–0.3 eV for each additional molecule added on the edge and their presence destabilizes the binding of organosulfur compounds by more than 0.2 eV. The inhibition of hydrodesulfurization is suggested to arise from site blocking and destabilization of reaction intermediates and transition states by organonitrogen compounds. © 2015 American Institute of Chemical Engineers AIChE J, 61: 4036–4050, 2015     

Computational screening of porous carbons,zeolites, and metal organic frameworks for desulfurization and decarburization of biogas,natural gas,and flue gas

Eighteen kinds of porous materials from carbons, zeolites, and metal organic frameworks (MOFs) have been extensively investigated for desulfurization and decarburization of the biogas, natural gas, and flue gas by using a molecular modeling approach. By considering not only the selectivity but also capacity, Na‐5A, zeolite‐like MOF (zMOF), and Na‐13X, MIL‐47 are screened as the most promising candidates for removal of sulfide in the CH₄? CO₂? H₂S and N₂? CO₂? SO₂ systems, respectively. However, for simultaneous removal of sulfide and CO₂, the best candidates are zMOF for the natural gas and biogas (i.e., CH₄? CO₂? H₂S system) and MOF‐74‐Zn for the flue gas (i.e., N₂? CO₂? SO₂ system). Moreover, the regeneration ability of the recommended adsorbents is further assessed by studying the effect of temperature on adsorption. It is found that compared to the zMOF and MIL‐47 MOFs, the Na‐5A and Na‐13X zeolites are not easily regenerated due to the difficulty in desorption of sulfide at high temperature, which results from the stronger adsorbent–adsorbate interactions in zeolites. The effect of sulfide concentration on the adsorption properties of the recommended adsorbents is also explored. We observe that the zMOF and MIL‐47 are also superior to the Na‐5A and Na‐13X for desulfurization of gas mixtures containing high sulfide concentration. This is because MOFs with larger pore volume lead to a greater sulfide uptake. The effects of porosity, framework density, pore volume, and accessible surface area on the separation performance are analyzed. The optimum porosity is about 0.5–0.6, to meet the requirements of both high selectivity and uptake. It is expected this work provides a useful guidance for the practical applications of desulfurization and decarburization. © 2013 American Institute of Chemical Engineers AIChE J, 59: 2928–2942, 2013     

Bilayer-mediated assembly of cationic nanoparticles adsorbed to lipid bilayers: Insights from molecular simulations

Understanding interactions between functionalized gold nanoparticles (NPs) and lipid bilayers is essential for biomedical applications. Experiments have shown that NPs that are stable in solution can assemble into clusters when adsorbed to a lipid bilayer, suggesting that bilayer-mediated interactions facilitate assembly. In this work, we use coarse-grained molecular dynamics simulations to study bilayer-mediated interactions between NPs adsorbed to single- and multicomponent lipid bilayers. We perform unbiased simulations and umbrella sampling calculations using an implicit solvent force field to determine the thermodynamic contributions to assembly. We show that bilayer-mediated interactions drive the assembly of NPs into linear aggregates on liquid-disordered bilayers, which we attribute to a reduction in bilayer curvature. Similar bilayer-mediated interactions induce the alignment of NP clusters with phase boundaries in phase-separated bilayers. Together, these simulation results provide new physical insight into the balance of forces that dictate the assembly of charged NPs at multicomponent lipid bilayer interfaces.     

Nonequilibrium equation of state for Lennard‐Jones fluids and the calculation of strain‐rate dependent shear viscosity

Nonequilibrium molecular dynamics simulation data for a 12‐6 Lennard‐Jones fluid are obtained over a wide range of temperatures, densities and strain‐rates. The data, which cover 660 different state points, are used to deduce the nonequilibrium contributions to the energy and pressure of the fluid under steady‐state conditions. These contributions are analysed and used in conjunction with an equilibrium equation of state to obtain an accurate nonequilibrium steady‐state equation of state for the 12‐6 Lennard‐Jones fluid. Comparison with simulation data indicates that the nonequilibrium contributions can be obtained with a similar accuracy to the equilibrium contributions. Relationships for the shear viscosity as functions of density and pressure are obtained, which adequately reproduce the shear viscosity simulation data. The isochoric shear viscosity as a function of pressure is shown to be independent of strain‐rate at sufficiently high strain‐rates. © 2010 American Institute of Chemical Engineers AIChE J, 2011     

Understanding interfacial behaviors of isobutane alkylation with C4 olefin catalyzed by sulfuric acid or ionic liquids

The interfacial properties between the hydrocarbon phase including isobutane and 2‐butene and the catalyst phase including H₂SO₄ or ionic liquids (ILs) with various alkyl chain length on their imidazolium cations have been investigated using molecular dynamics (MD) simulations. Compared to H₂SO₄, ILs can obviously improve the interfacial width, solubility and diffusion of reactants at the interface. The ILs with longer chains on cations exhibit a significant density enrichment of alkyl chains at the interface and tend to orient themselves with alkyl chains perpendicular to the interface and protruding into the reactant phase, which is in good agreement with the van der Waals energy between the reactants and cations of the ILs. The ILs with longer chains can improve the interfacial width and facilitate the dissolution of isobutane in catalyst phase, and thus exhibit a better catalytic performance, which agrees well with alkylation experiments in this work. © 2017 American Institute of Chemical Engineers AIChE J, 64: 950–960, 2018     

Liquid li structure and dynamics: A comparison between OFDFT and second nearest‐neighbor embedded‐atom method

The structure and dynamics of liquid lithium are studied using two simulation methods: orbital‐free (OF) first‐principles molecular dynamics (MD), which employs OF density functional theory (DFT), and classical MD utilizing a second nearest‐neighbor embedded‐atom method potential. The properties studied include the dynamic structure factor, the self‐diffusion coefficient, the dispersion relation, the viscosity, and the bond angle distribution function. Simulation results were compared to available experimental data when possible. Each method has distinct advantages and disadvantages. For example, OFDFT gives better agreement with experimental dynamic structure factors, yet is more computationally demanding than classical simulations. Classical simulations can access a broader temperature range and longer time scales. The combination of first‐principles and classical simulations is a powerful tool for studying properties of liquid lithium. © 2015 American Institute of Chemical Engineers AIChE J, 61: 2841–2853, 2015     

Predicting the Solubility of Solid Phenanthrene: A Combined Molecular Simulation and Group Contribution Approach

A simple correction to the infinite dilution activity coefficient computed via molecular simulation for a nonelectrolyte solid solute in solution is proposed. The methodology adopts the concept that the activity coefficient may be fundamentally interpreted as a product of a residual and combinatorial term. The residual contribution is assumed to be insensitive to concentration, and the combinatorial term is modeled using the athermal Flory–Huggins theory. The proposed method uses only properties for the solute computed at infinite dilution to estimate solution‐phase properties at finite concentrations. Properties of the pure solid solute are estimated using the group contribution method of Gani and coworkers, allowing for efficient blind solubility predictions to be made. The method is applied to predict the solubility of solid phenanthrene in 17 different solvents. For all cases, the combinatorial correction lowers the predicted solubility relative to the infinite dilution approximation, and in general, improves agreement with experiment. © 2013 American Institute of Chemical Engineers AIChE J, 59: 2647–2661, 2013     

Modeling and simulation of gas separations with spiral-wound membranes

Models for gas separations with spiral-wound membranes are developed and found to exhibit good agreement with experiments performed on N₂/O₂ mixtures. The two-dimensional (2D) model can be accurately approximated by a one-dimensional (1D) surrogate model when the spacer widths are chosen to make the channel pressure drops small. Subsequently, the separation of propane/propylene mixtures from the recycle purge stream of a polypropylene reactor is investigated. Assuming ideal gas is found to lead to significant overestimations in membrane stage cuts (sometimes more than 10%), an extent comparable to that associated with extrapolating constant olefin permeance from a low-pressure condition. While olefin permeance can change significantly with pressure, using a constant-permeance formulation can result in a small (< 2.5%) underprediction in stage cut if the value for the permeance is taken from the feed condition. Finally, membrane properties and costs necessary for a viable separation process are discussed.