Metal–organic frameworks (MOFs) have potential applications for efficient filtration of toxic gases from ambient air. We have used computational methods to examine the efficacy of functionalized UiO-66 with a wide range of functional groups to identify materials suitable for selective adsorption of NH3, H2S, or CO2 under humid conditions. To this end, adsorption energies at various favorable positions in the structures are obtained from both cluster-based and periodic models. Our cluster calculations show that DFT calculations using the PBE-D2 functional can reliably predict the ranking of materials obtained at the MP2 level. Performing PBE-D2 calculations using periodic models gives rankings of materials that are significantly different from those of cluster calculations, showing that confinement effects are important in these materials. On the basis of these calculations, recommendations for high performing materials are made using PBE-D2 calculations from periodic models that use the full structure of each MOF.

To establish a model of metal–organic framework (MOF) surfaces and build an understanding of surface-specific ligand adsorption phenomena in MOFs, we present a computational study exploring multiple models of a series of MOF-2 nanosheet materials, “M-BDCs”, with M = Zn, Cu, and Co and BDC = benzene-1,4-dicarboxylate. We study and assess the appropriateness of a series of models ranging from small clusters (18 atoms) to fully periodic sheet models. We additionally study the interactions of these models with acid gases and energy-relevant small molecules (CO, CO2, H2O, SO2, NO2, and H2S). We employ computational methods ranging from DFT with various exchange–correlation functionals to perturbative and coupled-cluster methods. For these systems, we present binding energies and enthalpies with the various ligands studied as well as IR frequency shifts for the normal modes of these ligands upon complexation with the open-metal sites of these materials. Our calculations lead to an understanding of phenomena unique to MOF surfaces and the importance of the periodicity in these materials in capturing surface-specific adsorption behaviors.

The rates of adsorbate diffusion in zeolitic imidazolate frameworks (ZIFs) can be varied by several orders of magnitude by incorporating two different imidazolate linkers in the ZIF crystals. Although some prior measurements of short-range order in these mixed-linker materials have been reported, it is unclear how this short-range order impacts the net diffusion of adsorbates. We introduce a lattice diffusion model that treats diffusion in ZIF-8x-90100–x crystals as a series of activated hops between cages, allowing us to assess the effects of short-range imidazolate order on molecular diffusion.

The development of adsorbed natural gas (ANG) technology creates opportunities for use of pipeline natural gas as clean fuel in vehicles. Metal–organic frameworks (MOFs) are one class of materials that have received considerable attention as possible adsorbents in ANG applications. We examine how accumulation of trace components from pipeline natural gas will impact the performance of MOFs in ANG during long-term cycling. Our approach combines information from grand canonical Monte Carlo (GCMC) simulations of single-component adsorption, ideal adsorbed solution theory (IAST) of multicomponent adsorption, and an isothermal model of tank cycling to assess accumulation of heavy hydrocarbons and tert-butyl mercaptan (TBM). In a series of MOFs, a reduction in deliverable energy up to 50% is observed after 200 cycles. These results highlight the importance of considering multicomponent effects during consideration of adsorbents for ANG applications.

We present a joint computational and experimental study of Mg–Ni-MOF-74 and Mg–Cd-MOF-74 to gain insight into the mixing of metals and understand how metal mixing affects the structure of the undercoordinated open-metal sites. Our calculations predict that metal mixing is energetically preferred in these materials. Recent experimental work has demonstrated that Mg–Ni-MOF-74 shows a much greater surface area retention in the presence of water than Mg-MOF-74. To probe this effect, we study H2O adsorption in Mg–Ni-MOF-74, finding that the adsorption energetics and electronic structure do not change significantly at the metal sites when compared to Mg-MOF-74 and Ni-MOF-74, respectively. We conclude that the increased stability of Mg–Ni-MOF-74 is a result of a M–O bond length distortion in mixed-metal MOF-74, consistent with recent work on the stability of MOF-74 under water exposure.

Swing capacity is a key performance metric for processes designed to capture CO2 by pressure swing adsorption (PSA). Sub-ambient operation of PSA units enables large changes in CO2 swing capacity, and can be economically viable when coupled with heat integration and power recovery. Here, we examine what upper bounds on CO2 swing capacity exist via molecular simulation of a large collection of metal–organic frameworks (MOFs). As has been observed previously for zeolites, the materials with the largest swing capacity at a given temperature have large pore volumes and heats of adsorption within a narrow range of optimal values. A number of materials are identified with swing capacities up to 40 mol kg−1 using a pressure swing from 0.1 bar to 2.0 bar.

Solid porous materials such as cationic zeolites have shown great potential in energy-efficient separation processes. Conventional adsorbent design involves ad-hoc and inefficient experimental evaluation of a large structural and compositional space. We developed a computational methodology to screen cationic zeolites for CO2 separation processes with quantitative accuracy, and identified a number of novel high-performing materials. This study enabled us to develop an intuitive design workflow for selecting optimal materials and dramatically accelerate the development of industrially relevant separation processes.

Most classical simulations of metal–organic frameworks model electrostatic interactions using point charges on each atom in the structure. We report atomic point charges derived from periodic density functional theory (DFT) electronic structure calculations for more than 2000 unique experimentally synthesized metal–organic frameworks (MOFs). These charges are publicly available as a supplement to the Computation-Ready Experimental MOF database. These DFT-derived atomic point charges are compared to semiempirical group contribution and charge equilibration methods for assigning charges. As an example of using these charges, we examined each MOF for usefulness in the adsorptive removal of tert-butyl mercaptan (TBM) from natural gas. Monte Carlo simulations revealed many candidate MOF structures with high selectivity for TBM over CH4 and high TBM capacity. We anticipate that this public data set of atomic point charges for MOFs will facilitate high-throughput screening for a wide variety of applications in which electrostatic interactions must be considered.

Direct adsorption of CO2 from ambient air, also known as direct air capture (DAC), is gaining attention as a complementary approach to processes that capture CO2 from more concentrated sources such as flue gas. Oxide-supported amine materials are effective materials for CO2 capture from dilute gases, but less work has been done on metal organic framework (MOF) supported amine materials. Use of MOFs as supports for amines could be a versatile approach to the creation of effective amine adsorbents because of the tunability of MOF structures. In the present work, MIL-101(Cr) materials functionalized with amine species are evaluated for CO2 capture from simulated air. Two amines are loaded into the MOF pores, tris (2-amino ethyl) (TREN) and low molecular weight, branched poly(ethylene imine) (PEI-800), at different amine loadings. The MIL-101(Cr)-TREN composites showed high CO2 capacities for high loadings of TREN, but a significant loss of amines is observed over multicycle temperature swing adsorption experiments. In contrast, MIL-101(Cr)-PEI-800 shows better cyclic stability. The amine efficiency (mol CO2/mol amine) as a function of amine loading is used as a metric to characterize the adsorbents. The amine efficiency in MIL-101(Cr)-PEI-800 showed a strong dependence on the amine loading, with a step change to high amine efficiencies occurring at ∼0.8 mmol PEI/g MOF. The kinetics of CO2 capture, which have important implications for the working capacity of the adsorbent, are also examined, demonstrating that a MIL-101(Cr)-PEI-800 sample with a 1–1.1 mmol PEI/g MOF loading has an excellent balance of CO2 capacity and CO2 adsorption kinetics.Keywords: Amine efficiency; Cyclic stability; Metal organic frameworks; Open metal sites; Poly(ethylene imine)

Transition state theory (TST) methods are useful for predicting adsorbate diffusivities in nanoporous materials at time scales inaccessible to molecular dynamics (MD). Most TST applications treat the nanoporous framework as rigid, which is inaccurate in highly flexible materials or where adsorbate dimensions are comparable to the size of pore apertures. In this study, we demonstrate two computationally efficient TST methods for simulating adsorbate diffusion in nanoporous materials where framework flexibility has a significant influence on diffusion. These methods are applied to light gas diffusion in porous organic cage crystal 3 (CC3), a highly flexible molecular crystal that has shown promise in gas separation applications. Diffusion in CC3 is modeled as a series of uncorrelated adsorbate hops between cage molecules and the voids between adjacent cages. The first method we applied to compute adsorbate hopping rates in CC3 is implicit ligand sampling (ILS). In ILS TST, hopping rates are calculated in an ensemble of rigid framework snapshots captured from a fully flexible MD trajectory of the empty CC3 structure. The second TST method we applied is umbrella sampling (US), where hopping rates are computed from a series of biased MD simulations. Our ILS and US TST calculations are shown to agree well with direct MD simulation of adsorbate diffusion in CC3. We anticipate that the efficient TST methods detailed here will be broadly applicable to other classes of flexible nanoporous materials such as metal–organic frameworks.

Degradation of metal–organic frameworks (MOFs) in aqueous, humid, and acid gas environments likely begins at defect sites. Until now, however, theoretical studies of MOFs have widely assumed an ideal defect-free structure. Here we present a computational model for low-energy extended defects in bulk zeolitic imidazolate frameworks (ZIFs) that are analogous to stacking faults in zeolites. We demonstrate the thermodynamic accessibility of stacking faults in ZIFs and examine the impact of these defects on pore diffusion and accessible surface area. We identify strong correlations between the defect density of a structure and its X-ray diffraction spectra. By examining a topologically isomorphic ZIF that has been reported experimentally we find characteristic defect-induced peak broadening and splitting in the reported powder patterns, giving strong evidence for the existence for stacking faults in this material.

Efficient processes for adsorptive separation of light olefin/paraffin mixtures are likely to have many advantages over traditional separation techniques for these commodity chemicals. Although some metal organic frameworks (MOFs) have been studied experimentally for this process, a large-scale computational screening study has not been reported due to the inherent difficulty in describing the critical role of interactions of olefins with open metal sites (OMS). In this paper, we introduce new density functional theory (DFT) derived force fields (FFs) that accurately describe adsorption of C2 and C3 olefins and paraffins in CuBTC. Using detailed DFT calculations for MOF-505 and PCN-16, we show that the energetics predicted by our FFs are transferable to other related MOFs that contain Cu OMS. Next, we evaluate the performance of 94 distinct Cu–OMS MOFs for the industrially important propylene/propane separation and identify 18 MOFs predicted to have attractive properties as adsorbents. Finally, we show that the ideal adsorbed solution theory is inaccurate for inhomogeneous olefin/MOF systems and present extensive binary propane/propylene adsorption isotherms for the top-performing MOFs identified in our calculations.

We report the computational discovery and experimental evaluation of nanoporous materials targeted at the adsorptive separation of p-xylene from a C8 aromatics mixture. We first introduce a computational method that is capable of efficiently predicting the p-xylene selectivities and capacities for a large database of porous materials. We then demonstrate the application of this method to screen a database of several thousand metal–organic framework (MOF) structures. Our computational screening methodology predicted that two MOFs with good solvothermal stability and commercially available linkers give comparable performance to the state-of-the-art zeolite BaX currently used in industrial p-xylene separations. The best-performing MOFs are then synthesized, and their xylene separation characteristics are evaluated in detail through breakthrough adsorption experiments and modeling. We find that the selectivities obtained in these materials are higher than that of any MOF previously reported in the literature and in some cases exceed the measured performance of zeolite BaX. In the case of the p-xylene selective material MOF-48, we use calculated free energy profiles to show how the presence of methyl substituents on the linkers allows the inversion of selectivity from the equivalent MOF with no methyl substituents (MIL-47, which is o-xylene selective). This combined computational and experimental methodology is a useful step in the development of MOFs for separation of aromatic hydrocarbons and can also be applied to other chemical separations and other classes of porous materials as long as the appropriate intermolecular force fields are available.

Many simulations of adsorption in nanoporous materials such as metal–organic frameworks (MOFs) treat the adsorbing materials as rigid. We show that the use of this approximation to describe the multicomponent adsorption of C8 aromatics in MOFs under industrial conditions gives results that differ dramatically from descriptions that include local framework flexibility. To address this issue, we develop an efficient method for capturing the effect of framework flexibility on adsorption in nanoporous materials. Our “flexible snapshot” method uses GCMC simulations to model adsorption in snapshots collected using fully flexible MD simulations and can be applied to any framework-adsorbate system for which reliable force fields are available. Our method gives considerably better agreement with experiments for multicomponent C8 aromatic selectivities in multiple MOFs than more traditional calculations using a single rigid framework. The rotation of organic linkers in the MOFs has a strong influence on selectivities in these systems. Because many MOFs contain this structural feature, we expect that using simulations that incorporate this kind of internal flexibility will be important in obtaining accurate adsorption predictions in a range of circumstances. This is especially true for many industrially relevant separations in MOFs, in particular, those that exploit high pore loadings of adsorbed species.

Zeolitic imidazolate frameworks (ZIFs) are a set of nanoporous metal–organic frameworks (MOFs) with tunable porosity and functionality. Among MOFs, they also show relatively good stability with respect to temperature and humidity. These characteristics lead to their possible applications in separation processes. In many practical separation processes, adsorbents are exposed to a variety of molecular species including acid gases. However, there is little knowledge of the effects of such acid gas exposure on the adsorption and separation properties of ZIFs. Here, the stability of a model ZIF material (ZIF-8) under SO2 exposure in dry, humid, and aqueous environments has been investigated in detail. Combined characterization by several techniques (PXRD, N2 physisorption, EDX, XPS, and FTIR) allowed us to track the structural and compositional properties of ZIF-8 before and after SO2 exposure. ZIF-8 is stable after prolonged exposure in dry SO2 and in humid air without SO2. However, exposure to 10–20 ppm concentrations of SO2 in the presence of high relative humidity led to its irreversible structural degradation over time as evidenced by substantial losses in crystallinity and textural properties. Exposure to similar concentrations of aqueous SO2 did not lead to bulk degradation. Humid SO2 exposed ZIF-8 showed a significant presence of sulfur (S) even after reactivation, with vibrational characteristics corresponding to (bi)sulfite and (bi)sulfate groups. A mechanism of ZIF-8 degradation combining the synergistic effects of SO2 and humidity is proposed. Attack by sulfuric and sulfurous acid species (generated in humid SO2) leads to protonation of nitrogen in the imidazole ring, resulting in cleavage of metal–linker (Zn–N) bonds. Our detailed experimental findings serve as a starting point for developing a generalized mechanism of acid gas interactions with ZIF materials.

We have introduced a simple modification of the well-known Hill–Sauer force field for silica. The modified force field improves the accuracy with which pore sizes and framework flexibility in small pore zeolites are described. The modification focused on the Si–O–Si and O–Si–O angles in these materials, which are instrumental in controlling vibrations of the framework such as twisting of the near-rigid SiO4 units. The accuracy of the modified Hill–Sauer force field was compared with data from extensive density functional theory calculations of zeolite structures and dynamics. The transferability of the force field was tested on 13 experimentally known silica 8-ring zeolites. Additional tests examining the thermal expansion behavior of selected zeolites were also performed.

Accurate and efficient predictions of hydrocarbon diffusivities in zeolitic imidazolate frameworks (ZIFs) are challenging, due to the small pore size of materials such as ZIF-8 and the wide range of diffusion time scales of hydrocarbon molecules in ZIFs. Here we have computationally measured the hopping rates of 15 different molecules (kinetic diameters of 2.66–5.10 Å) in ZIF-8 via dynamically corrected transition state theory (dcTST). Umbrella sampling combined with the one-dimensional weighted histogram analysis method (WHAM) was used to calculate the diffusion free energy barriers. Both the umbrella sampling and dynamical correction calculations included ZIF-8 flexibility, which is found to be critical in accurately describing molecular diffusion in this material. Comparison of the computed diffusivities to extant experimental results shows remarkable agreement within an order of magnitude for all the molecules. The dcTST method was also applied to study the effect of hydrocarbon loadings. Self and transport diffusion coefficients of methane, ethane, ethylene, propane, propylene, n-butane, and 1-butene in ZIF-8 are reported over a temperature range of 0–150 °C and loadings from infinite dilution to liquid-like loadings.

A microkinetic model containing 53 elementary steps based on extensive Density Functional Theory calculations is developed to describe syngas reactions on a Mo2C catalyst under high temperature and pressure conditions, with the aim of determining the elementary steps that control reaction selectivity. The effects of adsorbate–adsorbate interactions are found to be strong, so these interactions are described using the quasi-chemical approximation. Agreement with experimental observations of selectivity for syngas reactions at P = 30 bar and T = 573 K was found to be good without parametrizing the model in any way to the experimental reaction data. The activation energies of the elementary steps in the model were estimated using a Bronsted–Evans–Polanyi relation, and sensitivity analysis is used to examine the impact of uncertainties in this relation on the selectivity-determining steps of the reaction network. Our results are a useful example of identification of key elementary steps in a complex reaction network for the reactions available with syngas over a heterogeneous catalysis.Keywords: alcohol synthesis; Bronsted−Evans−Polanyi relation; density functional theory; molybdenum carbides; quasi-chemical approximation; syngas reactions

Density Functional Theory (DFT) based force fields (FFs) for Ar and Xe adsorption in six metal–organic frameworks were developed using three DFT functionals (PBE-D2, vdW-DF, vdW-DF2) in periodic systems. These force fields include van der Waals (vdW) and polarization terms, and the effect of the latter was shown to be small. Using our DFT-derived and standard (UFF) FFs in grand canonical Monte Carlo simulations, adsorption isotherms and heats of adsorption were calculated and compared with experiment. In most of the cases, it was possible to accurately predict adsorption isotherms using one of the DFT-derived FFs. Still, among the DFT functionals investigated, no single DFT functional could accurately describe all of the adsorbate-framework interactions. On average, performance of UFF and PBE-D2 based FFs to predict experimental isotherms were at a similar quality, still, UFF was slightly superior. Although vdW-DF2 based FFs predicted experimental isotherms almost perfectly for ZIF-8 and HKUST-1 up to 20 bar, their average performance was less than that of PBE-D2 based FFs. Nevertheless, the overall performance of UFF, PBE-D2 and vDW-DF2 FFs was similar. Lastly, vdW-DF based FFs always over-predicted experiments.

A test set of chemically and topologically diverse Metal–Organic Frameworks (MOFs) with high accuracy experimentally derived crystallographic structure data was compiled. The test set was used to benchmark the performance of Density Functional Theory (DFT) functionals (M06L, PBE, PW91, PBE-D2, PBE-D3, and vdW-DF2) for predicting lattice parameters, unit cell volume, bonded parameters and pore descriptors. On average PBE-D2, PBE-D3, and vdW-DF2 predict more accurate structures, but all functionals predicted pore diameters within 0.5 Å of the experimental diameter for every MOF in the test set. The test set was also used to assess the variance in performance of DFT functionals for elastic properties and atomic partial charges. The DFT predicted elastic properties such as minimum shear modulus and Young's modulus can differ by an average of 3 and 9 GPa for rigid MOFs such as those in the test set. The partial charges calculated by vdW-DF2 deviate the most from other functionals while there is no significant difference between the partial charges calculated by M06L, PBE, PW91, PBE-D2 and PBE-D3 for the MOFs in the test set. We find that while there are differences in the magnitude of the properties predicted by the various functionals, these discrepancies are small compared to the accuracy necessary for most practical applications.

Generic force fields such as UFF and DREIDING are widely used for predicting molecular adsorption and diffusion in metal–organic frameworks (MOFs), but the accuracy of these force fields is unclear. We describe a general framework for developing transferable force fields for modeling the adsorption of alkanes in a nonflexible MIL-47(V) MOF using periodic density functional theory (DFT) calculations. By calculating the interaction energies for a large number of energetically favorable adsorbate configurations using DFT, we obtain a force field that gives good predictions of adsorption isotherms, heats of adsorption, and diffusion properties for a wide range of alkanes and alkenes in MIL-47(V). The force field is shown to be transferable to related materials such as MIL-53(Cr) and is used to calculate the free-energy differences for the experimentally observed phases of MIL-53(Fe).

We use two methods, the changing snapshot method and transition state theory (TST)/snapshot method, to characterize the effects of zeolite framework flexibility on diffusion of spherical molecules in 8MR zeolites. These methods are applied to noble gases (Ne, Ar, Kr, Xe, and Rn) and CF4. We demonstrate the effect of the zeolite framework flexibility on diffusion by considering the size and loading of adsorbates and temperature. In both the methods, we approximate flexible structures as a set of discrete rigid snapshots obtained from molecular dynamics simulations of an empty framework. We show that the diffusivities predicted with these two efficient methods agree with direct MD simulations in the fully flexible structures. We studied in detail how the framework flexibility affects the loading dependence of diffusion. By looking at the computational costs, we demonstrated that both the methods are orders of magnitude more efficient than the fully flexible simulations. We then apply the changing snapshot method to binary mixtures of adsorbates to obtain accurate binary diffusivities and binary selectivities.

The diffusivities of linear hydrocarbons (CH4, C2H6, C2H4, C3H8, C3H6, and C4H10) in pure silica zeolite LTA (ITQ-29) are computed at 300 K and infinite dilution. To overcome the time scale problem arising from the slow diffusion process at room temperature, we used transition path sampling (TPS). The influence of framework flexibility on diffusion is investigated by combining TPS simulations with fully flexible molecular dynamics performed in the NpT ensemble. The ensemble of the collected reactive trajectories was used to characterize sets of transition states, and the corresponding configurations were analyzed to construct window size distributions during the molecular hopping events. The diffusion process is affected by framework flexibility, and the influence of framework flexibility on diffusion of propane and butane is much larger than for methane and ethane.

We present an extension of previous methods that derive transferable force fields to describe the adsorption of small molecules in zeolites based on density functional theory (DFT) calculations to examine the adsorption of C8 cyclic hydrocarbons in metal–organic frameworks (MOFs). We use our DFT-based force field to predict the adsorption properties of these molecules in MOFs where dispersion governs adsorption properties using grand canonical Monte Carlo (GCMC) simulations. We observe that DFT-derived force fields provide moderately more accurate predictions compared to generic force fields for single-component adsorption in these systems. We find that generic force fields can give qualitative agreement with experiments for binary selectivities, which could eventually be useful for materials screening purposes. We also assess the influence of factors such as framework relaxation due to guest adsorption on these calculations and find that these effects can produce significant changes in the simulated binary selectivities at high loadings. Our methodology will eventually be useful for developing force fields for systems in which generic force fields are known to fail and represent a useful step in understanding and predicting adsorption properties of C8 hydrocarbons in MOFs.

Hydrogen separation using metal membranes offers significant energetic, technological, and economic advantages over conventional separation processes, resulting in extensive efforts to develop favorable membrane materials. Intermetallics are stoichiometric compounds of two or more metals that form an ordered structure. This work exhibits a systematic search for intermetallic membrane materials for hydrogen separation from potential candidates using density functional theory (DFT)-based methods to quantitatively predict solubility, diffusivity, and permeability. Geometric correlations were used to significantly decrease the number of calculations performed without compromising on the accuracy of the predictions. In this work, 1001 intermetallic structures were screened, and eight materials, MnTi, MgZn2, PtTl2, FeHf2, HfTa, NiTi, TiV, and Fe2Y, were identified as potential candidates for hydrogen separation membranes based on the calculated hydrogen permeability. This work, in addition to identifying novel membrane materials, highlights the significance of computational tools for screening large libraries of materials for specific applications.

The development of metal organic frameworks (MOFs) with high porosity, large surface area, and good electrical properties would offer opportunities for producing functionalized porous materials suitable for energy storage, conversion, and utilization. Realizing these applications remains challenging because of the limited numbers of electrically conductive porous MOFs that are known. We apply density functional theory (DFT) to assess a large number of potentially electrically conductive MOFs generated by infiltrating known materials with conjugated and redox-active 7,7,8,8-tetracyanquinododimethane (TCNQ) molecules. DFT results demonstrate that TCNQ coordinating with dimeric Cu paddlewheels can create molecular chains in a variety of MOFs. Several of these materials feature the formation of multiple dimensional conducting chains, making the materials promising for electrical conductivity.

The “surface explosion” associated with decomposition of acetic acid on Rh surfaces is studied by density functional theory. The surface configuration of adsorbed acetate and the reaction paths available to the adsorbed species are determined. The reactivity of adsorbed acetate is found to be dependent on the local surface coverage, which allows a model explaining the experimentally observed surface explosion to be developed. Comparison of the reactivity of different surfaces shows that Rh(111) is more active than the Rh(110) surface for both clean and oxygen precovered surfaces, and the oxygen precovered surfaces are less active than the clean surfaces.Keywords: density functional theory (DFT); acetic acid; clean surface; oxygen precovered surface; Rh(110); Rh(111); surface explosion

Metal hydrides with high thermodynamic stability are desirable for high-temperature applications, such as those that require high hydrogen release temperatures or low hydrogen overpressures. First-principles calculations have been used previously to identify complex transition metal hydrides (CTMHs) for high temperature use by screening materials with experimentally known structures. Here, we extend our previous screening of CTMHs with a library of 149 proposed materials based on known prototype structures and charge balancing rules. These proposed materials are typically related to known materials by cation substitution. Our semiautomated, high-throughput screening uses density functional theory (DFT) and grand canonical linear programming (GCLP) methods to compute thermodynamic properties and phase diagrams: 81 of the 149 materials are found to be thermodynamically stable. We identified seven proposed materials that release hydrogen at higher temperatures than the associated binary hydrides and at high temperature, T > 1000 K, for 1 bar H2 overpressure. Our results indicate that there are many novel CTMH compounds that are thermodynamically stable, and the computed thermodynamic data and phase diagrams should be useful for selecting materials and operating parameters for high temperature metal hydride applications.

•Physisorption on low and high Miller index Cu surfaces was modeled with DFT.•Dispersion corrected DFT shows tendency to overbind the adsorbate.•Calculated results are qualitatively consistent with experimental TPD data.•Calculation indicates possible adsorbate induced surface reconstruction on Cu(110).The physisorption of R-3-methylcyclohexanone on low and high Miller index Cu surfaces is studied with temperature programmed desorption (TPD) and density functional theory (DFT). The DFT calculations are performed with D2, vdW-optB86b, and vdW-optB88 dispersion corrected methods. The adsorption energies calculated by the dispersion corrected methods are more comparable to the TPD results than those calculated without dispersion corrections, although, the former methods have a tendency to overbind the surface adsorbates. The implementation of dispersion corrected methods also indicates a possible adsorbate induced surface reconstruction on Cu(110).

•We study metal nanoparticles on graphene/Ru(0001) using DFT.•The adsorption and diffusion of Rh and Au dimers and trimers was examined.•Rh cluster mobility increases with cluster size.•Au2 shows higher mobility than Au1 and Au3.Periodic density functional theory (DFT) calculations have been performed to study metal nanoparticles on graphene/Ru(0001). Specifically, we examined adsorption and diffusion of monomers, dimers, and trimers of Rh and Au on graphene/Ru(0001). These two metal species were chosen for their distinct behaviors in cluster formation on the graphene/Ru(0001) moiré. The fcc region of this surface was predicted to be where the nucleation of metal nanoclusters occurs. The diffusion mechanisms and energy barriers for metal dimers and trimers were calculated. It was shown that the mobility of Rh clusters decreases with the increase of cluster size. For Au, however, dimers and trimers diffuse faster than monomers on the moiré surface. These calculations give insights into the nucleation process of metal clusters on the graphene moiré.

•Au8 on graphene/Ru(0001) was studied in Genetic Algorithm-based DFT calculations.•Au8 adsorbed in the fcc region forms a double-layer wall.•Au aggregates via Ostwald ripening with Au2 being the major diffusion intermediate.Gold nanoparticles have been extensively studied for their catalytic activity both theoretically and experimentally. The moiré pattern formed by graphene supported on single crystal substrates creates a useful environment where the properties of Au nanoclusters can be studied, provided the structure and evolution of these clusters can be controlled. We used a genetic algorithm combined with DFT calculations to predict the lowest energy structures of a Au8 cluster on graphene/Ru(0001). The most stable cluster forms a double-layer Au wall structure for Au8 in the fcc region of the moiré pattern, where the Au8 cluster is most strongly adsorbed. Further calculations give estimates for the net diffusion barrier of Au8 as an intact cluster on the surface. Our results are consistent with the Au island structure experimentally observed on graphene/Ru(0001), and support the hypothesis that Au clusters aggregate through Oswald ripening with Au dimers being the most important diffusing species.

Computational methods that characterize the thermodynamic properties of metal hydrides that operate at high temperatures, i.e., T > 800 K, are desirable for a variety of applications, including nuclear fuels and energy storage. Ternary hydrides tend to be less thermodynamically stable than the strongest binary hydride that forms from the metals. In this paper we use first-principles methods based on density functional theory, phonon calculations, and grand potential minimization to predict the isobaric phase diagram for 0 K ≤ T ≤ 2000 K for the Th–Zr–H element space, which is of interest given that ThZr2Hx ternary hydrides have been reported with an enhanced stability relative to the binary hydrides. We compute free energies including vibrational contributions for Th, Zr, ThH2, Th4H15, ZrH2, ThZr2, ThZr2H6, and ThZr2H7. We develop a cluster analysis method that efficiently computes the configurational entropy for ThZr2H6 interstitial hydride and conclude that the configurational entropy is not a major driver for the enhanced stability of ThZr2H6 relative to the binary hydrides. Density functional theory (DFT) predicted thermodynamic stabilities for the hydride phases are in reasonable agreement with experimental values. ThZr2H6 is stabilized by finite temperature vibrational effects, and ThZr2H7 is not predicted to be stable at any studied temperature or pressure.

Li-oxide garnet-related structures are promising solid-state Li-ion electrolytes in Li-ion batteries. However, garnet-type structures are susceptible to carbonate and hydroxide formation in environments containing gaseous CO2 and H2O. Therefore, in considering garnets for Li-ion conducting applications, chemical stability is an important issue. We examine the chemical stability of Li7La3Zr2O12, Li7La3Sn2O12, and Li7La3Hf2O12 with respect to carbonate and hydroxide formation reactions using density functional theory (DFT) calculations. From these studies, we rank the chemical stability of Li-oxide garnet-related structures against CO2 and H2O. The ranking of these materials by their chemical stability with respect to carbonate and hydroxide formation changes at higher partial pressures of CO2 and H2O.

Efficient purification of hydrogen from high temperature mixed gas streams using dense metal membranes can potentially play a critical role in the large-scale production of hydrogen from gasification of coal or biomaterials. We use first-principles calculations together with statistical methods to systematically predict hydrogen permeability through amorphous ternary Zr–Cu–T films (T = 17 elements) and other selected amorphous materials. These results greatly expand the range of amorphous materials that have been considered as hydrogen purification membranes. More importantly, we demonstrate that relatively simple descriptions of the site binding energies in these amorphous materials can account for the key observations from our detailed first-principles calculations. This outcome significantly reduces the computational effort required in future screening of materials in this application and also places bounds on the ultimate performance of these materials.

The development of accurate force fields is vital for predicting adsorption in porous materials. Previously, we introduced a first principles-based transferable force field for CO2 adsorption in siliceous zeolites (Fang et al., J. Phys. Chem. C, 2012, 116, 10692). In this study, we extend our approach to CO2 adsorption in cationic zeolites which possess more complex structures. Na-exchanged zeolites are chosen for demonstrating the approach. These methods account for several structural complexities including Al distribution, cation positions and cation mobility, all of which are important for predicting adsorption. The simulation results are validated with high-resolution experimental measurements of isotherms and microcalorimetric heats of adsorption on well-characterized materials. The choice of first-principles method has a significant influence on the ability of force fields to accurately describe CO2–zeolite interactions. The PBE-D2 derived force field, which performed well for CO2 adsorption in siliceous zeolites, does not do so for Na-exchanged zeolites; the PBE-D2 method overestimates CO2 adsorption energies on multi-cation sites that are common in cationic zeolites with low Si/Al ratios. In contrast, a force field derived from the DFT/CC method performed well. Agreement was obtained between simulation and experiment not only for LTA-4A on which the force field fitting is based, but for other two common adsorbents, NaX and NaY.

A longstanding aim in the development of electrolytes has been to find dopants that give high proton conductivity coupled with good chemical stability. Perovskite-type oxides are useful materials for proton conduction. We used first-principles calculations to address this topic in doped BaZrO3 with efficient methods to examine a wide range of possible dopants (Y, In, Ga, Sc, Nd, Al, Tl, Sm, Dy, La, Pm, Er, and Ho). These calculations correctly identify the doped BaZrO3 materials that are already known to have favorable properties, but also identify a number of promising materials that have not been examined previously. We investigated the physical origins of the trends in chemical stability and proton mobility among the different dopants. Our data allows us to consider several possible physical descriptors for characterizing doped perovskites as proton conductors.

We examine the adsorption and diffusion of small alcohols in ZIF-8 and ZIF-90 with a combined experimental and modeling approach. Our Grand Canonical Monte Carlo (GCMC) simulations predict that both ZIFs exhibit a slight adsorption selectivity for ethanol over methanol, in good agreement with previous experimental data. The adsorption uptake of the alcohols at low pressures is found to be significantly higher in ZIF-90 than ZIF-8. Our simulations indicate that this is due to hydrogen bonding between the alcohols and the carbonyl group of ZIF-90 but that this effect is not strong enough to cause appreciable flexibility of the ZIF-90 framework during adsorption. We also report alcohol self-diffusivities and Arrhenius parameters measured using pulsed field gradient NMR (PFG-NMR) and molecular dynamics (MD) simulations. The diffusivities measured using PFG-NMR indicate that the diffusion selectivity of methanol over ethanol is significantly higher in ZIF-8 (S = 229) than in ZIF-90 (S = 6) at T = 25 °C. Qualitative agreement is obtained between experimental and simulated diffusivities using the generalized AMBER (GAFF) force field including framework flexibility.

Catalytic materials to dissociate hydrogen molecules play an important role in both the large-scale production of hydrogen from gasification of coal and efficient storage of hydrogen by metal hydrides. We use first-principles calculations together with statistical mechanics methods to systematically investigate hydrogen transport properties through Mo2C-coated vanadium membranes. We evaluate hydrogen fluxes to determine the rate-limiting steps during hydrogen purification with these membranes. The existence of a high desorption resistance explains the non-Sieverts’ behavior observed in previous experiments with these membranes and suggests a general rule to select ideal catalysts to facilitate hydrogen transport through metal films by spillover.

We develop transferable force fields describing water adsorption in copper-based metal–organic frameworks (MOFs) by combining dispersion-corrected density functional theory (DFT) calculations and classical atomistic simulations. The DFT-D2 approach was found to give reasonable agreement with high level quantum chemistry results for the interaction between water and CuBTC. A classical force field for water adsorption in CuBTC including Lennard-Jones (LJ), Coulombic interactions, and a water–copper distance-dependent correction term was then developed on the basis of 1200 DFT-D2 calculations that probed the full range of accessible volume in CuBTC via random sampling. Good agreement was obtained between adsorption isotherms predicted with our first-principles-derived force field and experiments. Other commonly used models such as simple combinations of LJ and Coulomb potentials cannot adequately describe the interaction between water and CuBTC due to the chemical bonding between water oxygen and copper atoms. The transferability of this force field was examined using available adsorption isotherms for water in two structurally distinct copper-based MOFs, CuMBTC (methyl-1,3,5-benzenetricarboxylate) and CuEBTC (ethyl-1,3,5-benzenetricarboxylate), again with reasonably good agreement between calculated and experimental results.

Molecular dynamics (MD) and transition state theory (TST) methods are becoming efficient tools for predicting diffusion of molecules in nanoporous materials. The accuracy of predictions, however, often depends upon a major assumption that the framework of the material is rigid. This saves a considerable amount of computational time and is often the only method applicable to materials for which accurate force fields to model framework flexibility are not available. In this study, we systematically characterize the effect of framework flexibility on diffusion in four model zeolites (LTA, CHA, ERI, and BIK) that exhibit different patterns of window flexibility. We show that for molecules with kinetic diameters comparable to (or larger than) the size of the window the rigid framework approximation can produce order(s) of magnitude difference in diffusivities as compared to the simulations performed with a fully flexible framework. We also show that simple recipes to include the effect of framework flexibility are not generally accurate. To account for framework flexibility effects efficiently and reliably, we introduce two new methods in which the flexible structure is approximated as a set of discrete rigid snapshots obtained from simulations of dynamics of an empty framework, using either classical or, in principle, ab initio methods. In the first method, we perform MD simulations of diffusion in a usual manner but replace the rigid structure with a new random snapshot at a certain characteristic frequency corresponding to the breathing motion of the window, while keeping positions of adsorbate molecules constant. In the second method, we directly compute cage to cage hopping rates in each rigid snapshot using TST and average over a distribution of snapshots. Excellent agreement is obtained between diffusivities predicted with these two new methods and direct MD simulations using fully flexible structures. Both methods are orders of magnitude more efficient than the simulations with the fully flexible structure. The new methods are broadly applicable for fast and accurate predictions of both infinite dilution and finite loading diffusivities of simple molecules in zeolites and other nanoporous materials, generally without the need for an accurate flexible force field.

Understanding the factors that govern the thermodynamic stability of metal–organic frameworks (MOFs) is critical to the development of these materials. In this work we present the first theoretical investigation of the stability of MOF polymorphs under solvothermal conditions. We combine thermodynamic integration (TI) and osmotic framework adsorbed solution theory (OFAST) calculations to reliably predict the free energy of immersion, ΔGimm, of MOFs. We also demonstrate that the configurational free energy can be estimated in the harmonic approximation from the vibrational density of states (VDOS) of the framework. This approach is applied to a set of hypothetical Zn(mIm)2 polymorphs under ambient conditions using methanol as the solvent to mimic the actual solvothermal synthesis conditions for ZIF-8. Our simulations predict that the relative contribution of ΔGimm to the framework stability is minor compared to the configurational free energy relative to the most stable structure for all polymorphs. However, we also find that solvation affects the energetic ordering of metastable polymorphs. We observe good agreement between predictions of ΔGimm using the OFAST method and those using the more rigorous TI calculations. The methods developed in this work can be applied to study the effect of the interplay between structural topology, solvent, and temperature on the thermodynamic stability of the framework for a wide range of nanoporous materials.

Electrostatic interactions are a critical factor in the adsorption of quadrupolar species such as CO2 and N2 in metal–organic frameworks (MOFs) and other nanoporous materials. We show how a version of the semiempirical charge equilibration method suitable for periodic materials can be used to efficiently assign charges and allow molecular simulations for a large number of MOFs. This approach is illustrated by simulating CO2 and N2 adsorption in ∼500 MOFs; this is the largest set of structures for which this information has been reported to date. For materials predicted by our calculations to have promising adsorption selectivities, we performed more detailed calculations in which accurate quantum chemistry methods were used to assign atomic point charges, and molecular simulations were used to assess molecular diffusivities and binary adsorption isotherms. Our results identify two MOFs, experimentally known to be stable upon solvent removal, that are predicted to show no diffusion limitations for adsorbed molecules and extremely high CO2/N2 adsorption selectivities for CO2 adsorption from dry air and from gas mixtures typical of dry flue gas.

Understanding the adsorption and mobility of metal–organic framework (MOF)-supported metal nanoclusters is critical to the development of these catalytic materials. We present the first theoretical investigation of Au-, Pd-, and AuPd-supported clusters in a MOF, namely MOF-74. We combine density functional theory (DFT) calculations with a genetic algorithm (GA) to reliably predict the structure of the adsorbed clusters. This approach allows comparison of hundreds of adsorbed configurations for each cluster. From the investigation of Au8, Pd8, and Au4Pd4 we find that the organic part of the MOF is just as important for nanocluster adsorption as open Zn or Mg metal sites. Using the large number of clusters generated by the GA, we developed a systematic method for predicting the mobility of adsorbed clusters. Through the investigation of diffusion paths a relationship between the cluster’s adsorption energy and diffusion barrier is established, confirming that Au clusters are highly mobile in the MOF-74 framework and Pd clusters are less mobile.

Organic–inorganic hybrid (mixed matrix) membranes can potentially extend the separation performance of traditional polymeric materials while maintaining processing convenience. Although many dense films studies have been reported, there have been few reported cases of these materials being successfully extended to asymmetric hollow fibers. In this work we report the first successful production of mixed matrix asymmetric hollow fiber membranes containing metal-organic-framework (MOF) ZIF-8 fillers. Specifically, we have incorporated ZIF-8 into a polyetherimide (Ultem® 1000) matrix and produced dual-layer asymmetric hollow fiber membranes via the dry jet-wet quench method. The outer separating layer of these composite fibers contains 13 wt% (17 vol%) of ZIF-8 filler. These membranes have been tested over a range of temperatures and pressures for a variety of gas pairs. An increase in separation performance for the CO2/N2 gas pairs was observed for both pure gas and mixed gas feeds.Highlights► Formation of mixed matrix hollow fiber membranes containing metal organic framework (MOF) fillers. ► Incorporation of a MOF sub-class, zeolitic imidazolate framework (ZIF) fillers, into polymer matrices. ► Improved separation performance compared to pure polymer membranes for CO2/N2 separations. ► Platform development for future polymer/MOF mixed matrix membranes.

We develop a nonempirical atoms-in-molecules (AIM) method for computing net atomic charges that simultaneously reproduce chemical states of atoms in a material and the electrostatic potential V(r) outside its electron distribution. This method gives accurate results for a variety of periodic and nonperiodic materials including molecular systems, solid surfaces, porous solids, and nonporous solids. This method, called DDEC/c3, improves upon our previously published DDEC/c2 method (Manz, T. A.; Sholl, D. S. J. Chem. Theory Comput. 2010, 6, 2455–2468) by accurately treating nonporous solids with short bond lengths. Starting with the theory all AIM charge partitioning functionals with spherically symmetric atomic weights must satisfy, the form of the DDEC/c3 functional is derived from first principles. The method is designed to converge robustly by avoiding conditions that lead to nearly flat optimization landscapes. In addition to net atomic charges, the method can also compute atomic multipoles and atomic spin moments. Calculations performed on a variety of systems demonstrate the method’s accuracy, computational efficiency, and good agreement with available experimental data. Comparisons to a variety of other charge assignment methods (Bader, natural population analysis, electrostatic potential fitting, Hirshfeld, iterative Hirshfeld, and iterative stockholder atoms) show that the DDEC/c3 net atomic charges are well-suited for constructing flexible force-fields for atomistic simulations.

Contamination of metal films by sulfur-containing compounds presents a major challenge to using metal membranes for H2 purification in processes involving large-scale gasification of coal or biomass. Formation of bulk sulfide phases in these applications is typically associated with irreversible loss of performance and ultimately membrane failure. The concept of using metal alloys to reduce sensitivity to sulfide formation has been explored in a variety of experiments, but development of alloys for this purpose has been hampered by a lack of thermodynamic data on formation of the relevant sulfide phases. We show that first principles calculations using density functional theory can be used to predict the formation of bulk sulfide phases from metal alloys under conditions that are relevant for operation of metal membranes. Our methods are illustrated by assessing the formation of sulfides for PdAg, PdAu, and PdCu alloys of all possible binary compositions.

Direct capture of CO2 from air is a concept that, if successfully implemented, could lead to capture of CO2 from disperse sources. We have developed process models to consider the viability of adsorption-based air capture technologies. Our models focus on using an amino-modified silica adsorbent, TRI-PE-MCM-41, and a structured monolithic contactor unit. We have studied several different temperature swing adsorption processes using the purity of CO2 and annual product throughput as metrics for comparing process performance. This analysis identifies some of the operational parameters, adsorbent characteristics, and other factors that have a significant effect on the performance of the process. Using the total energy requirement of the process and available sources of energy, such as low pressure steam and electricity, we carry out an economic analysis to obtain a net operating cost for air capture of CO2. We identify a process with a daily throughput of ∼1.1 t CO2 at 88.5% purity using standard shipping container sized air capture units. The total energy required (6745 MJ/t CO2) is dominated by the parasitic losses—sensible heat requirements of the contactor (40%) and the adsorbent (28%) and not by the mechanical energy associated with air flow (∼5%). On the basis of our analysis of factors such as source of electricity, availability of low pressure steam, and geographic location, the net operating cost of capture is estimated to be ∼$100/t CO2. These cost estimates do not include capital expenses necessary to construct or maintain the air capture units. Potential strategies for further reducing the energy and monetary cost of these processes are identified. Our analysis supports continued work to establish the technological and economic feasibility of adsorption-based air capture.

MgH2 is a prototypical light metal hydride, but it is too stable for use in practical hydrogen storage applications. Experiments with ball-milled MgH2–0.1TiH2 have shown that the enthalpy of reaction of MgH2 in this mixture is lower than for bulk MgH2 (Lu, J. et al. J. Am. Chem. Soc.2009,131, 15843). We have explored the mechanism for this phenomenon using first-principles density functional theory calculations of MgH2/TiH2 interfaces. We found that epitaxial contact between MgH2 and TiH2 can be formed with strongly favorable interface formation energies. Similar epitaxial interfaces also exist for Mg and TiH2. The strain induced by TiH2(111) on MgH2 and Mg in these interfaces can reduce the enthalpy for H2 release from MgH2.

We demonstrate a new approach to develop transferable force fields describing molecular adsorption in zeolites by combining dispersion-corrected density functional theory (DFT) calculations and classical atomistic simulations. This approach is illustrated with the adsorption of CO2 in zeolites. Multiple dispersion-corrected DFT methods were tested for describing CO2 adsorption in sodium-exchanged ferrierite. The DFT-D2 approach was found to give the best agreement with high level quantum chemistry results and experimental data. A classical force field for CO2 adsorption in siliceous zeolites was then developed on the basis of hundreds of DFT-D2 calculations that probed the full range of accessible volume in purely siliceous chabazite (Si-CHA) via random sampling. We independently performed experiments with Si-CHA measuring CO2 isotherms and heats of adsorption by microcalorimetry. Excellent agreement was obtained between adsorption isotherms predicted with our first-principles-derived force field and our experiments. The transferability of this force field was examined using available adsorption isotherms for CO2 in siliceous MFI and DDR zeolites, again with reasonably good agreement between calculated and experimental results. The methods demonstrated by these calculations will be broadly applicable in using molecular simulations to predict properties of adsorbed molecules in zeolites and other nanoporous materials.

A collection of >3000 MOFs with experimentally confirmed structures were screened for performance in three binary separations: Ar/Kr, Kr/Xe, and Xe/Rn. 70 materials were selected for further analysis, and calculations were performed to account for inaccessible regions. Single component GCMC calculations were performed to parametrize IAST calculations on these 70 materials. An approach that avoids possible imprecision in IAST due to curve-fitting of single component isotherms is introduced. The precision of IAST for these gas pairs was confirmed with extensive binary GCMC calculations. For each binary separation, materials were identified with predicted performance that surpasses the state of the art. A significant number of materials were found to be “reverse selective” in the sense that a smaller gas species is preferably adsorbed over a larger species. The physical origin of this phenomenon is explained. The effect of temperature on separation performance was also examined.

Metal organic frameworks (MOFs) have experimentally been demonstrated to be capable of supporting isolated transition-metal clusters, but the stability of these clusters with respect to aggregation is unclear. In this letter we use a genetic algorithm together with density functional theory calculations to predict the structure of Pd clusters in UiO-66. The cluster sizes examined are far larger than those in any previous modeling studies of metal clusters in MOFs and allow us to test the hypothesis that the physically separated cavities in UiO-66 could stabilize isolated Pd clusters. Our calculations show that Pd clusters in UiO-66 are, at best, metastable and will aggregate into connected pore filling structures at equilibrium.Keywords: agglomeration; cluster stability; density functional theory; genetic algorithm; heterogeneous catalysis; support structures;

The MOF Mg/DOBDC has one of the highest known CO2 adsorption capacities at the low to moderate CO2 partial pressures relevant for CO2 capture from flue gas but is difficult to regenerate for use in cyclic operation. In this work, Mg/DOBDC is modified by functionalization of its open metal coordination sites with ethylene diamine (ED) to introduce pendent amines into the MOF micropores. DFT calculations and experimental nitrogen physisorption and thermogravimetric analysis suggest that 1 ED molecule is added to each unit cell, on average. This modification both increases the material’s CO2 adsorption capacity at ultradilute CO2 partial pressures and increases the regenerability of the material, allowing for cyclic adsorption–desorption cycles with identical adsorption capacities. This is one of the first MOF materials demonstrated to yield significant adsorption capacities from simulated ambient air (400 ppm CO2), and its capacity is competitive with the best-known adsorbents based on amine–oxide composites.Keywords: adsorption; air capture; carbon capture; CCS; metal−organic framework; MOF; MOF-74;

The selection of metal–organic frameworks (MOFs) for gas adsorption and separation has become a significant challenge over the past decade because of the large number of new structures reported every year. We applied a multiscale computational approach to screen existing MOFs for CO2/N2 separation. Pore characteristics of 1163 MOFs were analyzed by the method developed by Haldoupis, Nair, and Sholl (Haldoupis, E.; Nair, S.; Sholl, D. S. J. Am. Chem. Soc.2010, 132, 7528) using a simple steric model. On the basis of the pore size analysis, 359 MOFs were examined by classical molecular simulations. Adsorption and diffusion properties were computed using grand canonical Monte Carlo (GCMC) and molecular dynamics (MD) simulations, respectively. These molecular simulations were used to assess which materials hold the greatest promise as membrane materials for CO2/N2 separations. Finally, density functional theory (DFT) calculations were performed to provide preliminary information on the dynamic framework motion of selected MOFs.

The interaction of Rhn (n = 1–5) clusters with nonhydrated γ-Al2O3(100), hydrated γ-Al2O3(100), and hydrated γ-Al2O3(110) surfaces has been investigated using density functional theory methods. On these surfaces, Rh3 prefers a triangular geometry, while Rh4 and Rh5 adopt 3D structures. On the (100) surfaces, Rhn binds considerably more strongly on the nonhydrated surface than on the hydrated surface. On the hydrated (110) surface, Rhn binds to surface hydroxyl groups, which is consistent with experimental observations. Characterizing the structure of Rhn clusters makes it possible to identify the critical cluster size for nucleation on each surface.

The partitioning of electron spin density among atoms in a material gives atomic spin moments (ASMs), which are important for understanding magnetic properties. We compare ASMs computed using different population analysis methods and introduce a method for computing density derived electrostatic and chemical (DDEC) ASMs. Bader and DDEC ASMs can be computed for periodic and nonperiodic materials with either collinear or noncollinear magnetism, while natural population analysis (NPA) ASMs can be computed for nonperiodic materials with collinear magnetism. Our results show Bader, DDEC, and (where applicable) NPA methods give similar ASMs, but different net atomic charges. Because they are optimized to reproduce both the magnetic field and the chemical states of atoms in a material, DDEC ASMs are especially suitable for constructing interaction potentials for atomistic simulations. We describe the computation of accurate ASMs for (a) a variety of systems using collinear and noncollinear spin DFT, (b) highly correlated materials (e.g., magnetite) using DFT+U, and (c) various spin states of ozone using coupled cluster expansions. The computed ASMs are in good agreement with available experimental results for a variety of periodic and nonperiodic materials. Examples considered include the antiferromagnetic metal organic framework Cu3(BTC)2, several ozone spin states, mono- and binuclear transition metal complexes, ferri- and ferro-magnetic solids (e.g., Fe3O4, Fe3Si), and simple molecular systems. We briefly discuss the theory of exchange-correlation functionals for studying noncollinear magnetism. A method for finding the ground state of systems with highly noncollinear magnetism is introduced. We use these methods to study the spin–orbit coupling potential energy surface of the single molecule magnet Fe4C40H52N4O12, which has highly noncollinear magnetism, and find that it contains unusual features that give a new interpretation to experimental data.

First principles calculations have played a useful role in screening mixtures of complex metal hydrides to find systems suitable for H2storage applications. Standard methods for this task efficiently identify the lowest energy reaction mechanisms among all possible reactions involving collections of materials for which DFT calculations have been performed. The resulting mechanism can potentially differ from physical reality due to inaccuracies in the DFT functionals used, or due to other approximations made in estimating reaction free energies. We introduce an efficient method to probe the robustness of DFT-based predictions that relies on identifying reactions that are metastable relative to the lowest energy reaction path predicted with DFT. An important conclusion of our calculations is that in many examples DFT cannot unambiguously predict a single reaction mechanism for a well defined metal hydride mixture because two or more mechanisms have reaction energies that differ by a small amount. Our approach is illustrated by analyzing a series of single step reactions identified in our recent work that examined reactions with a large database of solids [Kim et al., Phys. Chem. Chem. Phys. 2011, 13, 7218].

The metal–organic framework (MOF) Cu(4,4′-hexafluoroisopropylidene-bis-benzoate)1.5 (Cu–hfipbb) has been projected as an important new nanoporous material for fabricating membranes with applications in gas separation and CO2 capture, among others. Synthesis of submicrometer crystals of Cu–hfipbb, however, is impeded by several factors, including the extreme hydrophobicity of the hfipbb ligand. We report a fast synthesis of submicrometer particles of Cu–hfipbb via a sonochemical technique, at temperatures as low as 0 °C, and with the addition of 2-propanol to control the particle morphology. The particles were characterized by powder X-ray diffraction, thermogravimetry, light scattering, and electron microscopy to ascertain the effects of synthesis parameters on the size distribution, structure, and morphology. The presence of a small amount of 2-propanol substantially alters the particle morphology from needles to a more isotropic shape. The Cu–hfipbb particles produced by this approach are suitable for use in applications involving fabrication of membranes and thin films.

Systematic thermodynamics calculations based on density functional theory-calculated energies for crystalline solids have been a useful complement to experimental studies of hydrogen storage in metal hydrides. We report the most comprehensive set of thermodynamics calculations for mixtures of light metal hydrides to date by performing grand canonical linear programming screening on a database of 359 compounds, including 147 compounds not previously examined by us. This database is used to categorize the reaction thermodynamics of all mixtures containing any four non-H elements among Al, B, C, Ca, K, Li, Mg, N, Na, Sc, Si, Ti, and V. Reactions are categorized according to the amount of H2 that is released and the reaction's enthalpy. This approach identifies 74 distinct single step reactions having that a storage capacity >6 wt.% and zero temperature heats of reaction 15 ≤ ΔU0 ≤ 75 kJ mol−1 H2. Many of these reactions, however, are likely to be problematic experimentally because of the role of refractory compounds, B12H12-containing compounds, or carbon. The single most promising reaction identified in this way involves LiNH2/LiH/KBH4, storing 7.48 wt.% H2 and having ΔU0 = 43.6 kJ mol−1 H2. We also examined the complete range of reaction mixtures to identify multi-step reactions with useful properties; this yielded 23 multi-step reactions of potential interest.

Computational methods have been used in the past to generate large libraries of hypothetical zeolite structures, but to date analysis of these structures has typically been limited to relatively simple physical properties such as density. We use efficient methods to analyze the adsorption and diffusion properties of simple adsorbate molecules in a library of >250000 hypothetical silica zeolites that was generated previously by Deem and co-workers (J. Phys. Chem. C, 2009, 113, 21353). The properties of this library of materials are compared to the complete set of ∼190 zeolites that have been identified experimentally. Our calculations provide information on the largest cavities available in each material for adsorption, and the size of the largest molecules that can diffuse through each material. For a subset of ∼8000 materials, we computed the Henry's constant and diffusion activation energy for adsorbed CH4 and H2. We show that these calculations provide a useful screening tool for considering large collections of nanocrystalline materials and choosing materials with particular promise for more detailed modeling.

Removal of toxic species such as As, Se, and S is critical to the successful implementation of high efficiency Integrated Gasification Combined Cycle (IGCC) processes for coal utilization. In this work we study the initial low-coverage surface reactions of AsH3, H2Se and H2S with a regenerable sorbent, zinc orthotitanate (Zn2TiO4), using first principles density functional theory. AsH3 adsorbs more preferentially on oxygen-rich (010) surfaces, while H2Se and H2S are more favorably bound to metal-rich (010) surfaces. We calculated the dissociation pathways and rates for each adsorbed species, finding that dehydrogenation of AsH3, H2Se, and H2S should be facile on these surfaces at the high temperatures relevant for IGCC processes.Research Highlights► Dissociative adsorption of AsH3, H2Se and H2S on Zn2TiO4(010) is studied by DFT. ► Fundamental insight into surface chemistry of trace contaminant removal is provided.

Dense metal membranes are one useful approach for purifying H2 from mixed gas streams. A longstanding challenge in the development of metal membranes has been to find film compositions that give high permeability for H2 relative to well-known materials such as elemental Pd. We used first-principles calculations to predict the H2 permeability of all disordered alloys of composition Pd96M4. These calculations not only identify the Pd-based alloys that are already known to have favorable membrane properties but also identify as promising a number of materials that have not been previously examined. We tested our predictions by fabricating and testing PdTm films, which, in agreement with our calculations, were found to have good permeability properties for pure H2.Keywords: density functional theory; hydrogen purification; permeability;

Molecular simulations have become an important complement to experiments for studying gas adsorption and separation in crystalline nanoporous materials. Conventionally, these simulations use force fields that model adsorbate−pore interactions by assigning point charges to the atoms of the adsorbent. The assignment of framework charges always introduces ambiguity because there are many different choices for defining point charges, even when the true electron density of a material is known. We show how to completely avoid such ambiguity by using the electrostatic potential energy surface (EPES) calculated from plane wave density functional theory (DFT). We illustrate this approach by simulating CO2 adsorption in four metal−organic frameworks (MOFs): IRMOF-1, ZIF-8, ZIF-90, and Zn(nicotinate)2. The resulting CO2 adsorption isotherms are insensitive to the exchange-correlation functional used in the DFT calculation of the EPES but are sensitive to changes in the crystal structure and lattice parameters. Isotherms computed from the DFT EPES are compared to those computed from several point charge models. This comparison makes possible, for the first time, an unbiased assessment of the accuracy of these point charge models for describing adsorption in MOFs. We find an unusually high Henry’s constant (109 mmol/g·bar) and intermediate isosteric heat of adsorption (34.9 kJ/mol) for Zn(nicotinate)2, which makes it a potentially attractive material for CO2 adsorption applications.

Using alkali metals as promoters is known to strongly influence the selectivity of catalytic reactions on Mo2C catalysts. To provide fundamental information about this observation, density functional theory calculations have been performed to study K and Rb adsorption of seven low-index α-Mo2C surfaces and the equilibrium crystal shape of Mo2C has been predicted using the Wulff construction. K and Rb are found to bind most strongly on Mo2C (001). This surface is also shown to favor a reconstruction in the absence of adsorbates. This reconstruction strengthens the adsorption of K and H while it weakens the adsorption of CO on the surface. A small number of calculations of H/K and CO/K coadsorption were performed to probe the effect of alkali promoters on these adsorbed species.

The addition of nonpolymeric particles to polymer films represents an important avenue for enhancing the performance of polymeric membranes for gas separations. We examine the challenge of selecting metal organic frameworks (MOFs) for use in high performance mixed matrix membranes using a combination of atomistic and continuum modeling. We validate our models by comparing with experimental data for IRMOF-1/Matrimid membranes. We then identify a highly selective MOF that is predicted to greatly enhance the performance of Matrimid and a range of other polymers for CO2/CH4 separations. The methods we introduce will create many opportunities for selecting MOF/polymer combinations for mixed matrix membranes with useful properties for large volume separations.

The very large number of distinct structures that are known for metal−organic frameworks (MOFs) and related materials presents both an opportunity and a challenge for identifying materials with useful properties for targeted applications. We show that efficient computational models can be used to evaluate large numbers of MOFs for kinetic separations of light gases based on finding materials with large differences between the diffusion coefficients of adsorbed gas species. We introduce a geometric approach that rapidly identifies the key features of a pore structure that control molecular diffusion and couple this with efficient molecular modeling calculations that predict the Henry’s constant and diffusion activation energy for a range of spherical adsorbates. We demonstrate our approach for >500 MOFs and >160 silica zeolites. Our results indicate that many large pore MOFs will be of limited interest for separations based on kinetic effects, but we identify a significant number of materials that are predicted to have extraordinary properties for separation of gases such as CO2, CH4, and H2.

First-principles calculations offer a useful complement to experiments by characterizing hydrogen permeance through dense metal membranes. We report calculations that combine quantum chemistry calculations and cluster expansion methods to describe the solubility, diffusivity, and permeation of interstitial H in fcc Pd-based binary and PdCu-based ternary alloys. Specifically, we examine Pd96M4 and Pd70Cu26M4 where M = Ag, Au, Pt, Rh, Cu, Pd, and Ni. We analyze Pd-based binary alloys to demonstrate the capability of the cluster expansion approach, which we then extended to the PdCu-based ternary alloys. Our results make predictions about the properties of these alloys as membranes at moderate hydrogen pressures over the temperature range 600 ≤ T ≤ 1200 K.

Net atomic charges (NACs) can be used both to understand the chemical states of atoms in a material as well as to represent the electrostatic potential, V, of the material outside its electron distribution. However, many existing definitions of NACs have limitations that prevent them from adequately fulfilling this dual purpose. Some charge methods are not applicable to periodic materials or are inaccurate for systems containing buried atoms, while others work for both periodic and nonperiodic materials containing buried atoms but give NACS that do not accurately reproduce V. We present a new approach, density derived electrostatic and chemical (DDEC) charges, that overcomes these limitations by simultaneously optimizing the NACs to be chemically meaningful and to reproduce V outside the electron distribution. This atoms-in-molecule method partitions the total electron density among atoms and uses a distributed multipole expansion to formally reproduce V exactly outside the electron distribution. We compare different methods for computing NACs for a broad range of materials that are periodic in zero, one, two, and three dimensions. The DDEC method consistently performs well for systems with and without buried atoms, including molecules, nonporous solids, solid surfaces, and porous solids like metal organic frameworks.

Density functional theory calculations have been used to study the adsorption of glycine, alanine, serine, cysteine, aspartic acid and asparagine on the hydroxylated (100) surface of α-quartz. We found measurable energy differences between the two enantiomers of serine, aspartic acid and asparagine in their most stable states on this surface, while negligible differences in adsorption energies for the most stable minima of enantiomers of alanine and cysteine were observed. Our results provide fundamental information on how amino acids can exhibit enantiospecific adsorption on hydroxylated quartz surfaces.

First-principles calculations represent a potent tool for screening metal hydride mixtures that can reversibly store hydrogen. A number of promising new hydride systems with high hydrogen capacity and favorable thermodynamics have been predicted this way. An important limitation of these studies, however, is the assumption that H2 is the only gas-phase product of the reaction, which is not always the case. This paper summarizes new theoretical and numerical approaches that can be used to predict thermodynamic equilibria in complex metal hydride systems with competing reaction pathways. We report thermochemical equilibrium calculations using data obtained from density functional theory (DFT) computations to describe the possible occurrence of gas-phase products other than H2 in three complex hydrides, LiNH2, LiBH4, and Mg(BH4)2, and mixtures of these with the destabilizing compounds LiH, MgH2, and C. The systems under investigation contain N, C, and/or B and thus have the potential to evolve N2, NH3, hydrocarbons, and/or boranes as well as H2. Equilibria as a function of both temperature and total pressure are predicted. The results indicate that significant amounts of these species can form under some conditions. In particular, the thermodynamic model predicts formation of N2 and NH3 as products of LiNH2 decomposition. Comparison with published experimental data indicates that N2 formation must be kinetically limited. Our examination of C-containing systems indicates that methane is the stable gas-phase species at low temperatures, not H2. On the other hand, very low amounts of boranes (primarily BH3) are predicted to form in B-containing systems.

A combination of in situ synchrotron powder diffraction, energy minimization (DFT), and Raman and infrared spectroscopy confirmed porous interpenetrated 3D-framework structures of recently discovered alkali-metal−zinc borohydrides, AZn2(BH4)5 (A = Li, Na). In the less zinc rich NaZn(BH4)3 the 3D-framework structural model has been confirmed but with a slightly modified description giving an isolated triangular anion, [Zn(BH4)3]−, rather than a 1D anionic chain, {[Zn(BH4)3]n}n−. Another polymorph of NaZn(BH4)3, isostructural to a new compound, LiZn(BH4)3, is proposed by energy minimization. Both compounds, the new NaZn(BH4)3 polymorph and LiZn(BH4)3, are, however, not observed experimentally at ambient pressure and in the temperature range of 100−400 K. The alkali-metal−zinc borohydride NaZn(BH4)3 containing the triangular anion [Zn(BH4)3]− is an equivalent of recently characterized alkali-metal−scandium borohydrides NaSc(BH4)4 and LiSc(BH4)4 based on the tetrahedral [Sc(BH4)4]− complex anion.

The diffusion of hydrogen in metal hydrides is crucial to the kinetics of H2 storage in these materials. Although the role of charged H-related defects in several metal hydrides has been studied previously, little is known about the possible role of charged defects associated with metal sites. We describe a comprehensive set of first-principles calculations examining Schottky, Frenkel, and anti-Frenkel defects in NaH, MgH2, and NaMgH3 to address this issue. Our calculations give information on the diffusivities of H and metal atoms in these materials and show that defects associated with metal sites can play a crucial role in net atomic transport. We also consider doping of NaH by heterovalent MgH2, finding that this doping can significantly enhance the diffusivity of H.Keywords (keywords): defect diffusivity; first-principles calculation; hydrogen storage; metal hydride; native defect; reaction kinetics;

Single-walled aluminosilicate nanotubes (NTs) are attractive for molecular separation applications because of their highly ordered structure, tunable dimensions, as well as their hydrophilic and functionalizable interiors. These NTs possess a pore surface consisting of an ordered array of silanol groups with flexible hydroxyls. We show that the flexibility of these hydroxyl groups is critical in the adsorption of hydrogen-bonding molecules. Specifically, we study the adsorption of water, methanol, CO2, and CH4 in the NT via grand canonical Monte Carlo (GCMC) simulations. The experimentally observed hydrophilicity of the surface can be captured in adsorption calculations only if the structural and orientational flexibility of the surface hydroxyls is incorporated. The adsorption selectivity of water over methanol is predicted to be larger than 100, which makes aluminosilicate NTs promising for dehydration of alcohols. Flexibility effects are less significant for the adsorption of non-hydrogen-bonding molecules.Keywords (KEYWORDS): adsorption; aluminosilicate; flexibility; inorganic nanotubes; nanoporous; water;

The bialkali borohydride LiK(BH4)2 has a total H2 content of 10.65 wt % and has been suggested as a potential candidate for hydrogen storage applications. We report first-principles calculations of the structure and reaction thermodynamics of LiK(BH4)2. We have also examined KBH4 and its structural analog NaBH4. Both KBH4 and NaBH4 have crystal structures involving disordered H atoms on lattice sites, and we show that the configurational entropy associated with these sites plays an important role in determining the stable phases of these materials. Our calculations indicate that LiK(BH4)2 is unstable with respect to decomposition into LiBH4 and KBH4 at almost all temperatures. We have also searched for reactant compositions that can destabilize LiK(BH4)2, KBH4, or NaBH4. This effort identified several coreactants in mixtures with KBH4 and NaBH4 that are predicted to have reaction thermodynamics for hydrogen release within the range of other borohydride-based systems being actively considered for hydrogen storage applications.

Degradation of metal membranes by sulfur in high temperature hydrogen purification is an important challenge in practical use of these membranes. If metal sulfides could be identified with high hydrogen permeabilities, they could potentially sidestep the difficulties faced by metal membranes. We have used first principles calculations to predict the permeability of hydrogen through eight metal sulfides in their defect-free, crystalline form. Although the solubility of H in some of these materials is considerable, H diffusion is extremely slow in every case. Our calculations indicate that films of any of the sulfides we considered will have extremely low permeabilities for hydrogen.

The identification of membrane materials with high selectivity for CO2/CH4 mixtures could revolutionize this industrially important separation. We predict using computational methods that a metal organic framework (MOF), Cu(hfipbb)(H2hfipbb)0.5, has unprecedented selectivity for membrane-based separation of CO2/CH4 mixtures. Our calculations combine molecular dynamics, transition state theory, and plane wave DFT calculations to assess the importance of framework flexibility in the MOF during molecular diffusion. This combination of methods should also make it possible to identify other MOFs with attractive properties for kinetic separations.

First principles calculations suggest that mobility of H in solid borohydrides is dominated by neutral interstitial H2, not charged defects.

Metal–organic frameworks are intriguing crystalline nanoporous materials that have potential applications in adsorption-based and membrane-based gas separations. We describe atomically detailed simulations of gas adsorption and diffusion in CuBTC that have been used to predict the performance of CuBTC membranes for separation of H2/CH4, CO2/CH4 and CO2/H2 mixtures. CuBTC membranes are predicted to have higher selectivities for all three mixtures than MOF-5 membranes, the only other metal–organic framework material for which detailed predictions of membrane selectivities have been made. Our results give insight into the physical properties that will be desirable in tuning the pore structure of MOFs for specific membrane-based separations.

Metal−organic frameworks (MOFs) have emerged as a fascinating alternative to more traditional nanoporous materials. Although hundreds of different MOF structures have been synthesized in powder form, little is currently known about the potential performance of MOFs for membrane-based separations. We have used atomistic calculations to predict the performance of a MOF membrane for separation of various gas mixtures in order to provide information for material selection in membrane design. Specifically, we investigated the performance of MOF-5 as a membrane for separation of CO2/CH4, CO2/H2, CO2/N2, CH4/H2, N2/H2, and N2/CH4 mixtures at room temperature. In every case, mixture effects play a crucial role in determining the membrane performance. Although the membrane selectivities predicted for MOF-5 are not large for the mixtures we studied, our result suggest that atomistic simulations will be a useful tool for considering the large number of MOF crystal structures that are known in order to seek membrane materials with more desirable characteristics.

Metal−organic framework (MOF) materials are a class of nanoporous materials that have many potential advantages over traditional nanoporous materials for adsorption and other chemical separation technologies. Because of the large number of different MOFs that exist, efforts to predict the performance of MOFs using molecular modeling can potentially play an important role in selecting materials for specific applications. We review the current state-of-the-art in the molecular modeling and quantum mechanical modeling of MOFs. Quantum mechanical calculations have been used to date to examine structural and electronic properties of MOFs and the calculation of MOF−guest interactions. Molecular modeling calculations using empirical classical potential calculations have been used to study pure and mixed fluid adsorption in MOFs. Similar calculations have recently provided initial information about the diffusive transport of adsorbed fluids in MOFs.

Metal organic frameworks (MOFs) define a diverse class of nanoporous materials having potential applications in adsorption-based and membrane-based gas separations. We have previously used atomically detailed models to predict the performance of MOFs for membrane-based separations of gases, but these calculations require considerable computational resources and time. Here, we introduce an efficient approximate method for screening MOFs based on atomistic models that will accelerate the modeling of membrane applications. The validity of this approximate method is examined by comparison with detailed calculations for CH4/H2, CO2/CH4, and CO2/H2 mixtures at room temperature permeating through IRMOF-1 and CuBTC membranes. These results allow us to hypothesize a connection between two computationally efficient correlations predicting mixture adsorption and mixture self-diffusion properties and the validity of our approximate screening method. We then apply our model to six additional MOFs, IRMOF-8, -9, -10, and -14, Zn(bdc)(ted)0.5, and COF-102, to examine the effect of chemical diversity and interpenetration on the performance of metal organic framework membranes for light gas separations.

Density functional theory calculations have been used to study the adsorption of glycine, alanine, serine, and cysteine on the hydroxylated (0001) surface of α-quartz. We found negligible differences in adsorption energies for the most stable minima of enantiomers of alanine on this surface. There are, however, measurable energy differences between the two enantiomers of both serine and cysteine in their most stable states. The source of this enantiospecificity is mainly the difference in the strength of hydrogen bonds between the surface and the two enantiomers. Our results provide initial information on how amino acids can exhibit enantiospecific adsorption on hydroxylated quartz surfaces.

Understanding transport phenomena of fluids through nanotubes (NTs) is of great interest in order to enable potential application of NTs as separation devices, encapsulation media for molecule storage and delivery, and sensors. Single-walled metal oxide NTs are interesting materials because they present a well-defined solid-state structure, precisely tunable diameter and length, as well as a hydrophilic and functionalizable interior for tuning transport and adsorption selectivity. Here, we study the transport properties of hydrogen-bonding liquids (water, methanol, and ethanol) through a single-walled aluminosilicate NT to investigate the influence of liquid−surface and liquid−liquid interactions and the effects of competitive transport of different chemical species using molecular dynamics (MD) simulations. The self-diffusivities (Ds) for all the three species decrease with increasing loading and are comparable to bulk liquid diffusivities at low molecular loadings. We show that the hydrogen-bond network associated with water makes its diffusion behavior different from methanol and ethanol. Mixtures of water and methanol show segregation in the NT, with water located closer to the tube wall and the alcohol molecules localized near the center of the NT. Ds values of water in an analogous aluminogermanate NT are larger than those in the aluminosilicate NT due to a larger pore diameter.Keywords: aluminosilicate; ethanol; inorganic nanotubes; methanol; self-diffusion; water

Efficient purification of hydrogen from high temperature mixed gas streams can potentially play a critical role in the large-scale production of hydrogen from gasification of coal or biomass. Dense metal membranes have many favorable properties for this kind of purification, but existing membranes based on crystalline metal alloys have a number of limitations. The use of amorphous metal films as membranes has potential to overcome at least some of the disadvantages of crystalline metal membranes. We present new modeling methods that make it possible for the first time to quantitatively predict the performance of amorphous metal films as hydrogen purification membranes. These methods are introduced by examining amorphous Fe3B, a material where comparisons can be made to a crystalline material with the same composition. A membrane made from the amorphous material is predicted to have a hydrogen permeability 1.5–2 orders of magnitude higher than a crystalline membrane. The methods we introduce here will be useful in accelerating the development of amorphous membranes for practical applications.

Density functional theory calculations have been used to study the adsorption of methylamine and methanol on the hydroxylated (0 0 0 1) surface of α-quartz. The surface structure for these calculations was based on the structure reported recently by Goumans et al. [T.P.M. Goumans, A. Wander, W.A. Brown, C.R.A. Catlow, Phys. Chem. Chem. Phys. 9 (2007) 2146]. Adsorption of methylamine or methanol in their most energetically preferred sites occurs by breaking one of the hydrogen bonds that exists on the bare surface and creating two hydrogen bonds between the surface and the adsorbed molecule. We report the adsorption energy and the vibrational frequencies associated with adsorption of these two species. Understanding the adsorption of these species on α-quartz (0 0 0 1) will be useful in future consideration of the adsorption of chiral molecules such as amino acids on this surface, which is intrinsically chiral.

Mass transport of chemical mixtures in nanoporous materials is important in applications such as membrane separations, but measuring diffusion of mixtures experimentally is challenging. Methods that can predict multicomponent diffusion coefficients from single-component data can be extremely useful if these methods are known to be accurate. We present the first test of a method of this kind for molecules adsorbed in a metal−organic framework (MOF). Specifically, we examine the method proposed by Skoulidas, Sholl, and Krishna (SSK) (Langmuir, 2003, 19, 7977) by comparing predictions made with this method to molecular simulations of mixture transport of H2/CH4 mixtures in CuBTC. These calculations provide the first direct information on mixture transport of any species in a MOF. The predictions of the SSK approach are in good agreement with our direct simulations of binary diffusion, suggesting that this approach may be a powerful one for examining multicomponent diffusion in MOFs. We also use our molecular simulation data to test the ideal adsorbed solution theory method for predicting binary adsorption isotherms and a method for predicting mixture self-diffusion coefficients.

A test set of chemically and topologically diverse Metal–Organic Frameworks (MOFs) with high accuracy experimentally derived crystallographic structure data was compiled. The test set was used to benchmark the performance of Density Functional Theory (DFT) functionals (M06L, PBE, PW91, PBE-D2, PBE-D3, and vdW-DF2) for predicting lattice parameters, unit cell volume, bonded parameters and pore descriptors. On average PBE-D2, PBE-D3, and vdW-DF2 predict more accurate structures, but all functionals predicted pore diameters within 0.5 Å of the experimental diameter for every MOF in the test set. The test set was also used to assess the variance in performance of DFT functionals for elastic properties and atomic partial charges. The DFT predicted elastic properties such as minimum shear modulus and Young's modulus can differ by an average of 3 and 9 GPa for rigid MOFs such as those in the test set. The partial charges calculated by vdW-DF2 deviate the most from other functionals while there is no significant difference between the partial charges calculated by M06L, PBE, PW91, PBE-D2 and PBE-D3 for the MOFs in the test set. We find that while there are differences in the magnitude of the properties predicted by the various functionals, these discrepancies are small compared to the accuracy necessary for most practical applications.

Density Functional Theory (DFT) based force fields (FFs) for Ar and Xe adsorption in six metal–organic frameworks were developed using three DFT functionals (PBE-D2, vdW-DF, vdW-DF2) in periodic systems. These force fields include van der Waals (vdW) and polarization terms, and the effect of the latter was shown to be small. Using our DFT-derived and standard (UFF) FFs in grand canonical Monte Carlo simulations, adsorption isotherms and heats of adsorption were calculated and compared with experiment. In most of the cases, it was possible to accurately predict adsorption isotherms using one of the DFT-derived FFs. Still, among the DFT functionals investigated, no single DFT functional could accurately describe all of the adsorbate-framework interactions. On average, performance of UFF and PBE-D2 based FFs to predict experimental isotherms were at a similar quality, still, UFF was slightly superior. Although vdW-DF2 based FFs predicted experimental isotherms almost perfectly for ZIF-8 and HKUST-1 up to 20 bar, their average performance was less than that of PBE-D2 based FFs. Nevertheless, the overall performance of UFF, PBE-D2 and vDW-DF2 FFs was similar. Lastly, vdW-DF based FFs always over-predicted experiments.

The development of accurate force fields is vital for predicting adsorption in porous materials. Previously, we introduced a first principles-based transferable force field for CO2 adsorption in siliceous zeolites (Fang et al., J. Phys. Chem. C, 2012, 116, 10692). In this study, we extend our approach to CO2 adsorption in cationic zeolites which possess more complex structures. Na-exchanged zeolites are chosen for demonstrating the approach. These methods account for several structural complexities including Al distribution, cation positions and cation mobility, all of which are important for predicting adsorption. The simulation results are validated with high-resolution experimental measurements of isotherms and microcalorimetric heats of adsorption on well-characterized materials. The choice of first-principles method has a significant influence on the ability of force fields to accurately describe CO2–zeolite interactions. The PBE-D2 derived force field, which performed well for CO2 adsorption in siliceous zeolites, does not do so for Na-exchanged zeolites; the PBE-D2 method overestimates CO2 adsorption energies on multi-cation sites that are common in cationic zeolites with low Si/Al ratios. In contrast, a force field derived from the DFT/CC method performed well. Agreement was obtained between simulation and experiment not only for LTA-4A on which the force field fitting is based, but for other two common adsorbents, NaX and NaY.

First principles calculations suggest that mobility of H in solid borohydrides is dominated by neutral interstitial H2, not charged defects.

Density functional theory calculations have been used to study the adsorption of glycine, alanine, serine, cysteine, aspartic acid and asparagine on the hydroxylated (100) surface of α-quartz. We found measurable energy differences between the two enantiomers of serine, aspartic acid and asparagine in their most stable states on this surface, while negligible differences in adsorption energies for the most stable minima of enantiomers of alanine and cysteine were observed. Our results provide fundamental information on how amino acids can exhibit enantiospecific adsorption on hydroxylated quartz surfaces.

Computational methods have been used in the past to generate large libraries of hypothetical zeolite structures, but to date analysis of these structures has typically been limited to relatively simple physical properties such as density. We use efficient methods to analyze the adsorption and diffusion properties of simple adsorbate molecules in a library of >250000 hypothetical silica zeolites that was generated previously by Deem and co-workers (J. Phys. Chem. C, 2009, 113, 21353). The properties of this library of materials are compared to the complete set of ∼190 zeolites that have been identified experimentally. Our calculations provide information on the largest cavities available in each material for adsorption, and the size of the largest molecules that can diffuse through each material. For a subset of ∼8000 materials, we computed the Henry's constant and diffusion activation energy for adsorbed CH4 and H2. We show that these calculations provide a useful screening tool for considering large collections of nanocrystalline materials and choosing materials with particular promise for more detailed modeling.

The identification of membrane materials with high selectivity for CO2/CH4 mixtures could revolutionize this industrially important separation. We predict using computational methods that a metal organic framework (MOF), Cu(hfipbb)(H2hfipbb)0.5, has unprecedented selectivity for membrane-based separation of CO2/CH4 mixtures. Our calculations combine molecular dynamics, transition state theory, and plane wave DFT calculations to assess the importance of framework flexibility in the MOF during molecular diffusion. This combination of methods should also make it possible to identify other MOFs with attractive properties for kinetic separations.

Swing capacity is a key performance metric for processes designed to capture CO2 by pressure swing adsorption (PSA). Sub-ambient operation of PSA units enables large changes in CO2 swing capacity, and can be economically viable when coupled with heat integration and power recovery. Here, we examine what upper bounds on CO2 swing capacity exist via molecular simulation of a large collection of metal–organic frameworks (MOFs). As has been observed previously for zeolites, the materials with the largest swing capacity at a given temperature have large pore volumes and heats of adsorption within a narrow range of optimal values. A number of materials are identified with swing capacities up to 40 mol kg−1 using a pressure swing from 0.1 bar to 2.0 bar.

First principles calculations have played a useful role in screening mixtures of complex metal hydrides to find systems suitable for H2storage applications. Standard methods for this task efficiently identify the lowest energy reaction mechanisms among all possible reactions involving collections of materials for which DFT calculations have been performed. The resulting mechanism can potentially differ from physical reality due to inaccuracies in the DFT functionals used, or due to other approximations made in estimating reaction free energies. We introduce an efficient method to probe the robustness of DFT-based predictions that relies on identifying reactions that are metastable relative to the lowest energy reaction path predicted with DFT. An important conclusion of our calculations is that in many examples DFT cannot unambiguously predict a single reaction mechanism for a well defined metal hydride mixture because two or more mechanisms have reaction energies that differ by a small amount. Our approach is illustrated by analyzing a series of single step reactions identified in our recent work that examined reactions with a large database of solids [Kim et al., Phys. Chem. Chem. Phys. 2011, 13, 7218].

First-principles calculations represent a potent tool for screening metal hydride mixtures that can reversibly store hydrogen. A number of promising new hydride systems with high hydrogen capacity and favorable thermodynamics have been predicted this way. An important limitation of these studies, however, is the assumption that H2 is the only gas-phase product of the reaction, which is not always the case. This paper summarizes new theoretical and numerical approaches that can be used to predict thermodynamic equilibria in complex metal hydride systems with competing reaction pathways. We report thermochemical equilibrium calculations using data obtained from density functional theory (DFT) computations to describe the possible occurrence of gas-phase products other than H2 in three complex hydrides, LiNH2, LiBH4, and Mg(BH4)2, and mixtures of these with the destabilizing compounds LiH, MgH2, and C. The systems under investigation contain N, C, and/or B and thus have the potential to evolve N2, NH3, hydrocarbons, and/or boranes as well as H2. Equilibria as a function of both temperature and total pressure are predicted. The results indicate that significant amounts of these species can form under some conditions. In particular, the thermodynamic model predicts formation of N2 and NH3 as products of LiNH2 decomposition. Comparison with published experimental data indicates that N2 formation must be kinetically limited. Our examination of C-containing systems indicates that methane is the stable gas-phase species at low temperatures, not H2. On the other hand, very low amounts of boranes (primarily BH3) are predicted to form in B-containing systems.

Systematic thermodynamics calculations based on density functional theory-calculated energies for crystalline solids have been a useful complement to experimental studies of hydrogen storage in metal hydrides. We report the most comprehensive set of thermodynamics calculations for mixtures of light metal hydrides to date by performing grand canonical linear programming screening on a database of 359 compounds, including 147 compounds not previously examined by us. This database is used to categorize the reaction thermodynamics of all mixtures containing any four non-H elements among Al, B, C, Ca, K, Li, Mg, N, Na, Sc, Si, Ti, and V. Reactions are categorized according to the amount of H2 that is released and the reaction's enthalpy. This approach identifies 74 distinct single step reactions having that a storage capacity >6 wt.% and zero temperature heats of reaction 15 ≤ ΔU0 ≤ 75 kJ mol−1 H2. Many of these reactions, however, are likely to be problematic experimentally because of the role of refractory compounds, B12H12-containing compounds, or carbon. The single most promising reaction identified in this way involves LiNH2/LiH/KBH4, storing 7.48 wt.% H2 and having ΔU0 = 43.6 kJ mol−1 H2. We also examined the complete range of reaction mixtures to identify multi-step reactions with useful properties; this yielded 23 multi-step reactions of potential interest.