The many-body expansion (MBE) method is the basis of many fragment-based methods and is widely applied to the computation of large molecular systems. To reach linear-scaling computation, a cutoff must be used to discard those subsystems with long interfragment distances. However, this leads to a discontinuous potential energy surface (PES) that would cause various problems in geometry optimizations and molecular dynamics simulations. To solve this problem, we present a generalized-switch-function (GSF) approach to smooth the PES computed by the MBE method with the use of a cutoff distance. The GSFs are naturally normalized and are permutation invariant. This approach can be applied to adaptively computing any order of many-body correction energies with multilevel computational methods and is a dynamic subsystem approach. We have applied the two versions of our method, GSF-MBE(m)/L1 and GSF-MBE(m)/(L1:L2:L3), to water clusters. Thorough tests show that our method can indeed give smooth potential-energy surface and is linear scaling but without losing much accuracy for very large water clusters with appropriately chosen cutoff distances.

•Formic acid firstly adsorbs as monodendate configuration.•Heating surface converts monodendate configuration to bridging configuration.•At low temperature formic acid decomposes into CO2 and surface hydroxyl groups.•At high temperature formic acid decomposes into CO and H2O.•Catalytic centers at high temperature are surface oxygen defects.The adsorption and decomposition of formic acid (FA) on the Ga2O3(100) surface was studied with density functional theory. On the perfect Ga2O3(100) surface, the preferred adsorption state of FA is a monodentate configuration while the most stable adsorption state is a bridging configuration. Heating the surface would convert FA from monodentate to bridging configuration and further heating would decompose FA into CO2 and two surface hydroxyl groups. On the other hand, on the O(2)-defect Ga2O3(100) surface the preferred adsorption state of FA is a bridging formate with one O atom of formate filling the O(2) vacancy. Heating the surface would generate CO and two surface hydroxyl groups. If the Ga2O3(100) surface is used as decomposition catalyst, then at low temperature the formation of a small amount of CO2 can be observed. On the other hand, at high temperature continuous formation of CO and H2O can be observed. The active sites for FA decomposition are the O(2) defects on the surface formed in situ from the removal of water from surface hydroxyl groups. The strong dependence of mechanism on experimental conditions explains why no consensus has been reached in the previous experimental studies regarding the adsorption and decomposition mechanism of FA.First principle calculations on the decomposition mechanism of formic acid on the Ga2O3(100) surface indicate that the mechanism strongly depends on experimental conditions. At low temperature the decomposition would form CO2 and surface hydroxyl groups while at high temperature the catalytic decomposition is through dehydration pathway to CO + H2O with surface oxygen defects as activation center.

A strategy for the reduction of CO2 to CO by or catalyzed by metal-free silylboranes has been proposed with the aid of density functional theory (DFT) computations. We showed that one oxygen atom of CO2 can be abstracted by silylboranes without catalysts or by diboranes in the presence of silylborane catalysts with surprisingly low free-energy barriers so that the reaction can be realized under mild experimental conditions. To achieve this, the reduction mechanism of CO2 by a hierarchy of silylboranes (R1)2BSi(R2)3 was systematically investigated. Several rules of thumb were obtained to guide the design of silylboranes with high activity toward CO2 reduction. After considering many factors, such as side reactions, the stability of the silylboranes, and the solvent effect, two silylboranes, (PFP)2BSi(CH2F)3 and Me2BSi(CH2F)3, suitable for the reduction of CO2 under mild experimental conditions were designed. The overall free-energy barriers for the reduction of CO2 by the two silylboranes are just 26.1–27.0 kcal mol−1 and 28.1–28.9 kcal mol−1, respectively, at 298.15 K in solution. We further showed that CO2 can be reduced to CO by diborane Me2BBMe2 using Me2BSi(CH2F)3 as the catalyst. The overall free-energy barrier for this catalytic reaction is just 30.6–30.7 kcal mol−1 at 298.15 K in solution.

Formic acid decomposition (FAD) reaction has been an innovative way for hydrogen energy. Noble metal catalysts, especially palladium-containing nanoparticles, supported or unsupported, perform well in this reaction. Herein, we considered the simplest model, wherein one Pd atom is used as the FAD catalyst. With high-level theoretical calculations of CCSD(T)/CBS quality, we investigated all possible FAD pathways. The results show that FAD catalyzed by one Pd atom follows a different mechanism compared with that catalyzed by surfaces or larger clusters. At the initial stage of the reaction, FAD follows a dehydration route and is quickly poisoned by CO due to the formation of very stable PdCO. PdCO then becomes the actual catalyst for FAD at temperatures approximately below 1050 K. Beyond 1050 K, there is a switch of catalyst from PdCO to Pd atom. The results also show that dehydration is always favoured over dehydrogenation on either the Pd-atom or PdCO catalyst. On the Pd-atom catalyst, neither dehydrogenation nor dehydration follows the formate mechanism. In contrast, on the PdCO catalyst, dehydrogenation follows the formate mechanism, whereas dehydration does not. We also systematically investigated the performance of 24 density functional theory methods. We found that the performance of the double hybrid mPW2PLYP functional is the best, followed by the B3LYP, B3PW91, N12SX, M11, and B2PLYP functionals.

In the present computational study by using the density functional theory (DFT) method, we found that silylboranes, which have metal-free Lewis acid centers only, can break the C–H bond of the exceedingly unreactive methane. The study shows that, unlike the activation mechanism of small molecules by the frustrated Lewis pairs (FLPs), the Lewis acidic boron center plays a key role in breaking the C–H bond of methane. Detailed analyses indicate that in the transition state the C–H bond is substantially activated by the empty 2p orbital of boron (2pB) primarily due to the orbital interaction between the C–H σ-bonding orbital and 2pB. On the other hand, the orbital interaction between the C–H σ-anti-bonding orbital and the B–Si σ-bonding orbital also contributes to the activation but plays a minor role. A statistical method was used to find the relationship between the reactivity of 57 silylboranes and their electronic properties. The results indicate that the boron center does have more prominent effect on the reactivity, especially the occupancy (nB2p) and energy (εB2p) of 2pB, where lowering nB2p and εB2p will increase the reactivity of the silylboranes. Based on the activation mechanism and taking kinetic and thermodynamic possibilities, as well as the possible side reactions, into consideration, three silylboranes suitable for methane activation under mild experimental conditions were designed. The analogous line of thought can be used as a hint for further experimental realizations, even under ambient conditions. This strategy can also be expected to be transplanted to more extensive C–H activation of hydrocarbons.

A bulky organoborane ArF2BMe (ArF = 2,4,6-tris(trifluoromethyl)phenyl, 1) has been synthesized. In C6D6 solution this organoborane and pyridine form a frustrated Lewis pair. Under mild conditions, 1 can efficiently catalyze 1,4-hydroboration of a series of pyridines. This reaction is highly chemo- and regioselective. The reaction intermediate, a boronium complex [Py2Bpin][ArF2B(H)Me] (3), was characterized in solution by NMR spectroscopy, which was also confirmed by DFT calculation.

The infrared spectra of mass-selected mononuclear copper nitrosyl cation complexes [Cu(NO)n]+ with n = 1–5 and their argon tagged complexes are measured via infrared photodissociation spectroscopy in the nitrosyl stretching frequency region in the gas phase. The experimental spectra provide distinctive patterns allowing the determination of the geometries and electronic structures of these complexes by comparison with the predicted spectra from density functional theory computations. The argon tagged [Cu(NO)2Ar2]+ and [Cu(NO)3Ar]+ complexes as well as the higher n = 4 and 5 complexes each involve a bidentate (NO)2 dimer ligand, suggesting that ligand–ligand coupling plays an important role in the bonding of these cation systems. The results also show that argon tagging has a strong influence on the geometric and electronic structures of the n = 2 and 3 complexes. The [Cu(NO)4]+ cation is the most intense peak in the mass spectrum, which is characterized to be the fully coordinated ion with a D2d structure involving two (NO)2 units but with only 14-valence electrons on Cu. The [Cu(NO)5]+ cation complex is determined to involve a [Cu(NO)4]+ core ion that is coordinated by an external NO ligand.

The highly selective isomerization and dehydration of various carbohydrates (glucose, xylose, cellobiose and cellulose) are one-pot conversions conducted in a simple borate-containing phosphate buffer solution (PBS) under microwave irradiation. It is demonstrated that the key to glucose converting into 5-hydromethylfurfural (5-HMF) with a high selectivity is matching the isomerization and dehydration processes of glucose at the appropriate boron/glucose mole ratio (B/G) and pH of PBS with the aid of microwave irradiation. Moreover, the interaction between borate and glucose in the isomerization process has been demonstrated by both Raman spectra and theoretical calculations. Furthermore, the reusability of such a PBS system with borate has been accomplished by the successive addition of glucose and by the continual removal of 5-HMF in a biphasic system. These results not only deepen the understanding of the isomerization and dehydration behavior of glucose, but also provide the possibility for practical applications, owing to the appealing green, inexpensive and sustainable characteristics of such a catalytic system.

Infrared spectra of mass-selected homoleptic cobalt carbonyl cluster cations including dinuclear Co2(CO)8+ and Co2(CO)9+, trinuclear Co3(CO)10+ and Co3(CO)11+, as well as tetranuclear Co4(CO)12+ are measured via infrared photodissociation spectroscopy in the carbonyl stretching frequency region. The geometric structures of these complexes are determined by comparison of the experimental spectra with those calculated by density functional theory. The Co2(CO)8+ cation is characterized to have a Co–Co bonded structure with Cs symmetry involving a bridging CO ligand. The Co2(CO)9+ cation is determined to be a mixture of the CO-tagged Co2(CO)8+–CO complex and the Co(CO)5+–Co(CO)4 ion–molecular complex. The Co3(CO)10+ cation is the coordination-saturated trinuclear cluster, which is characterized to have a triangle Co3 core with C2 symmetry involving two edge-bridging and eight terminal CO ligands. The Co3(CO)11+ cation is a weakly bound complex involving a Co3(CO)10+ core ion. The Co4(CO)12+ cluster cation is deduced to have a tetrahedral Co4+ core structure with three edge-bridging and nine terminal carbonyls.

Infrared spectra of mass-selected oxygen-rich iron dioxygen complexes Fe(O2)n+ with n = 3–5 are measured via infrared photodissociation spectroscopy in the gas phase. These cation complexes are produced via a laser vaporization supersonic ion source. The structures are established by comparison of the experimental spectra with the simulated spectra derived from density functional calculations. All of the Fe(O2)n+ complexes studied have a single IR-active band in the 1050–1100 cm–1 region, arising from the O–O stretching vibration of the superoxo ligand(s). These complexes are determined to have structures with a chemically bound Fe(O2)2+ core ion that is weakly coordinated by neutral O2 molecules. The Fe(O2)2+ core ion has a planar D2h symmetry with two equivalent side-on superoxo ligands bound to an Fe3+ cation center.

The divide-and-conquer (DC) scheme, the most popular linear-scaling method, is very important in the quantum mechanics computation of large systems. However, when a chemical system is divided into subsystems, its covalent bonds are often broken and then capped by complementary atoms/groups. In this paper, we show that the charge transfer between subsystems and the complementary atoms/groups causes the nonconservation of the total charge of the whole system, and this is the main source of error for the computed total energy. On the basis of this finding, an extension of the many-body expansion method (energy-based divide-and-conquer, EDC) utilizing charge conservation (E-EDC) is proposed. In the E-EDC method, initially the total energies of the whole system at different many-body correction levels are computed according to the EDC scheme. The total charges of the whole system, that is, the sum of the charges of the subsystems without cap atoms/groups at different many-body correction levels, are also computed. Then the total energy is extrapolated to the value at which the net charge of the whole system equals to the real value. Other properties such as atomic forces can also be extrapolated in a similar way. In the test of 24 and 32 glycine oligomers, this scheme reduces the error of the total energy by about 40–70%, but the computational cost is almost the same as that of the EDC scheme.

Infrared spectra of mass-selected homoleptic nickel carbonyl cluster cations including dinuclear Ni2(CO)7+ and Ni2(CO)8+, trinuclear Ni3(CO)9+ and tetranuclear Ni4(CO)11+ are measured via infrared photodissociation spectroscopy in the carbonyl stretching frequency region. The structures are established by comparison of the experimental spectra with simulated spectra derived from density functional calculations. The Ni2(CO)7+ cation is characterized to have an unbridged asymmetric (OC)4Ni–Ni(CO)3+ structure with a Ni–Ni single bond. The Ni2(CO)8+ cation has a Ni–Ni half-bonded D3d structure with both nickel centers exhibiting an 18-electron configuration. The trinuclear Ni3(CO)9+ cluster cation is determined to have an open chain like (OC)4Ni–NiCO–Ni(CO)4 structure. The tetranuclear Ni4(CO)11+ cluster cation is determined to have a tetrahedral structure with two-center and three-center bridge-bonded carbonyl units. These nickel carbonyl cluster cations all involve trigonal pyramid like Ni(CO)4 building blocks that satisfy the 18-electron configuration of the nickel centers.

•ODP reaction on four VOx/TiO2 catalysts has been studied by DFT methods.•Key factors affecting activity: support surface, vanadia loading and active sites.•Activity is controlled by the oxidizability and bonding ability of active sites.•The coexistence of OV= and OV–Ti sites enhances the catalytic activity.The oxidative dehydrogenation of propane (ODP) on the anatase supported vanadia catalysts (VOx/TiO2) have been investigated using periodic DFT calculations. Free energy profiles indicate that the first C–H activation step is the rate-determining (RD) step and the transition state (TS) of the propene formation step is the RD–TS. ODP activity can be tuned by vanadia dispersion and support surface via the modification of the electronic structure of the active oxygen sites. For the RD step, on both dimer VOx/TiO2 catalysts terminal sites have higher activity. On monomer VOx/TiO2 (1 0 0) terminal and interface sites exhibit similar activity, while on monomer VOx/TiO2 (0 0 1) interface sites have higher activity. With increasing vanadia loading, the formation of propene changes from propyl radical mechanism to a concerted propoxide one. The results suggest that TiO2 (1 0 0) is a better support surface. Terminal and interface oxygen sites act cooperatively as the first and second C–H bond activation centers, respectively.First principle calculations on the ODP reaction catalyzed by VOx/TiO2 indicate that the catalytic activity is related with support surface, vanadia loading and active sites by affecting the oxidizability and bonding ability of active sites. The coexistence of OV= and OV–Ti sites enhances the catalytic activity.

Infrared spectra of mass-selected homoleptic iron carbonyl cluster cations including mononuclear Fe(CO)5+ and Fe(CO)6+, dinuclear Fe2(CO)8+ and Fe2(CO)9+, and trinuclear Fe3(CO)12+ are measured via infrared photodissociation spectroscopy in the carbonyl stretching frequency region. The structures are established by comparison of the experimental spectra with simulated spectra derived from density functional calculations. Only one IR band is observed for the Fe(CO)5+ cation, which is predicted to have a C4v structure. The Fe(CO)6+ cation is determined to be a weakly bound complex involving a Fe(CO)5+ core ion. In contrast to neutral clusters which have symmetric structures with two and three bridging carbonyl ligands, the dinuclear Fe2(CO)8+ and Fe2(CO)9+ cations are characterized to have unbridged asymmetric (OC)5Fe–Fe(CO)n+ (n = 3 and 4) structures. The trinuclear Fe3(CO)12+ cluster cation is determined to have an open chain like (OC)5Fe–Fe(CO)2–Fe(CO)5 structure instead of the triangular structure with two bridging CO groups for the Fe3(CO)12 neutral. The di- and trinuclear cluster cations all involve a square pyramid like Fe(CO)5 building block that satisfies the 18-electron configuration of this iron center. The Fe(CO)5 building block is isolobal to the CH3 fragment in hydrocarbon chemistry, the Fe2(CO)9+ and Fe3(CO)12+ cluster cations may be considered through isolobality to be metal carbonyl analogues of the ethyl and isopropyl cations.

Infrared spectra of mass-selected homoleptic dinuclear iron carbonyl cluster anions Fe2(CO)n− (n = 4–9) are measured via infrared photodissociation spectroscopy in the carbonyl stretching frequency region. The cluster anions are produced via a laser vaporization supersonic cluster source. Density functional calculations have been performed and the calculated vibrational spectra are compared to the experimental data to identify the gas-phase structures of the cluster anions. The experimentally observed Fe2(CO)n− (n = 4–7) cluster anions are characterized to have unusual asymmetric (OC)4Fe–Fe(CO)n−4 structures, which also correspond to the computed lowest energy structures. The experimentally observed Fe2(CO)8− cluster anion is determined to have an unbridged structure instead of the previously reported dibridged structure. The Fe2(CO)9− cluster anion is determined to involve a Fe2(CO)8− core anion that is solvated by an external CO molecule. Bonding analysis indicates that these anions each have a Fe–Fe single bond to satisfy the 18-electron configuration of one iron center. The results provide important new insight into the structure and bonding mechanisms of transition-metal carbonyl clusters.

Periodic density functional theory has been utilized to investigate the structure and stability of monomeric HVOx species on anatase support. The three most stable surfaces of anatase were investigated, namely the (001), (100) and (101) surfaces. Unlike previous theoretical studies it was found that on the (001) surface vanadia species with five-coordinated vanadium atom are more stable than those with tetrahedrally coordinated vanadium atom. On the other hand, on the (100) and (101) surfaces, the vanadium atom in the vanadia species is still tetrahedrally coordinated. The stability of different VOx/TiO2 structures which are not fully dehydrated has been systematically studied and the results show that the vanadia species on the three surfaces follow an order of TiO2 (001) > TiO2 (100) > TiO2 (101). This can be understood from the acidity and basicity of the three anatase surfaces. The results suggest that monomeric VOx species may be better stabilized if the support exposes more (001) surfaces. Our analyses on electronic structure of the most stable VOx/TiO2 structure (D001) suggest that its bridging V–O–Ti oxygen atoms may have higher reactivity than the terminal vanadyl oxygen atoms.Highlights► Adsorption of HVOx species on anatase TiO2 surfaces was modeled by periodic DFT method. ► Five-coordinated vanadia species was found to be more stable on the TiO2(001) surface. ► Monomeric VOx species was found to be stabilized better on the (001) surface. ► Bridging V–O–Ti oxygens may have higher reactivity than terminal vanadyl oxygens.

The reactions of titanium monoxide and dioxide molecules with carbon dioxide were investigated by matrix isolation infrared spectroscopy. It was found that the titanium monoxide molecule is able to activate carbon dioxide to form the titanium dioxide–carbon monoxide complex upon visible light excitation via a weakly bound TiO(η1-OCO) intermediate in solid neon. In contrast, the titanium dioxide molecule reacted with carbon dioxide to form the titanium monoxide–carbonate complex spontaneously on annealing. Theoretical calculations predicted that both activation processes are thermodynamically exothermic and kinetically facile.

Infrared spectra of mass-selected homoleptic dinuclear chromium carbonyl cluster cations Cr2(CO)n+ with n = 7–9 are measured via infrared photodissociation spectroscopy in the carbonyl stretching frequency region in the gas phase. The structures are established by comparison of the experimental spectra with the simulated spectra derived from density functional calculations. The Cr2(CO)n+ cluster cations are characterized to have the (OC)5Cr–C–O–Cr(CO)n−6+ structures with a linear bridging carbonyl group bonded to one chromium atom through its carbon atom and to the other chromium atom through its oxygen atom. The cluster cations all have a sextet ground state with the positive charge and the unpaired electrons located on the Cr(CO)n−6 moiety. The formation of the linear bridging structures without Cr–Cr bonding can be rationalized that chromium forms strong Cr–CO bonds but weak Cr–Cr bonds.

Quantum chemical calculations were carried out on CO oxidation catalyzed by a single gold atom. To investigate the performance of density functional theory (DFT) methods, 42 DFT functionals have been evaluated and compared with high-level wavefunction based methods. It was found that in order to obtain accurate results the functionals used must treat long range interaction well. The double-hybrid mPW2PLYP and B2PLYP functionals are the two functionals with best overall performance. CAM-B3LYP, a long range corrected hybrid GGA functional, also performs well. On the other hand, the popular B3LYP, PW91, and PBE functionals do not show good performance and the performance of the latter two are even at the bottom of the 42 functionals. Our accurate results calculated at the CCSD(T)/aug-cc-pVTZ//mPW2PLYP/aug-cc-pVTZ level of theory indicate that Au atom is a good catalysis for CO oxidation. The reaction follows the following mechanism where CO and O2 adsorb on Au atom forming an Au(OCOO) intermediate and subsequently O2 transfer one oxygen atom to CO to form CO2 and AuO. Then AuO reacts with CO to form another CO2 to complete the catalytic cycle. The overall energy barrier at 0 K for the first CO oxidation step (Au + CO + O2 → AuO + CO2) is just 4.8 kcal mol−1, and that for the second CO oxidation step (AuO + CO → Au + CO2) is just 1.6 kcal mol−1.

Carbon dioxide coordination and activation by niobium oxide molecules were studied by matrix isolation infrared spectroscopy. It was found that the niobium monoxide molecule reacted with carbon dioxide to form the niobium dioxide carbonyl complex NbO2(η1-CO) spontaneously on annealing in solid neon. The observation of the spontaneous reaction is consistent with theoretical predictions that this carbon dioxide activation process is both thermodynamically exothermic and kinetically facile. In contrast, four niobium dioxide–carbon dioxide complexes exhibiting three different coordination modes of CO2 were formed from the reactions between niobium dioxide and carbon dioxide, which proceeded with the initial formation of the η1-O bound NbO2(η1-OCO) and NbO2(η1-OCO)2 complexes on annealing. The NbO2(η1-OCO) complex rearranged to the η2-O,O bound NbO2(η2-O2C) isomer under visible light irradiation, while the NbO2(η1-OCO)2 complex isomerized to the NbO2(η1-OCO)(η2-OC)O structure involving an η2-C,O ligand under IR excitation. In these niobium dioxide carbon dioxide complexes, the η1-O coordinated CO2 ligand serves as an electron donor, whereas both the η2-C,O and η2-O,O coordinated CO2 ligands act as electron acceptors.

The reactions of titanium oxide molecules with dinitrogen have been studied by matrix isolation infrared spectroscopy. The titanium monoxide molecule reacts with dinitrogen to form the TiO(N2)x (x = 1–4) complexes spontaneously on annealing in solid neon. The TiO(η1-NN) complex is end-on bonded and was predicted to have a 3A′′ ground state arising from the 3Δ ground state of TiO. Argon doping experiments indicate that TiO(η1-NN) is able to form complexes with one or more argon atoms. Argon atom coordination induces a large red-shift of the N–N stretching frequency. The TiO(η2-N2)2 complex was characterized to have C2v symmetry, in which both the N2 ligands are side-on bonded to the titanium metal center. The tridinitrogen complex TiO(η1-NN)3 most likely has C3v symmetry with three end-on bonded N2 ligands. The TiO(η1-NN)4 complex was determined to have a C4v structure with four equivalent end-on bonded N2 ligands. In addition, evidence is also presented for the formation of the TiO2(η1-NN)x (x = 1–4) complexes, which were predicted to be end-on bonded.

The performance of various density functional theory methods on the geometries and energetics of Au2, Au3, Au4, and Au5 has been systematically evaluated. The results were compared with those from experiments or high-level wave function theory methods. In the present study, spin−orbit (SO) coupling was considered. It was found that SO coupling plays a very important role in the calculation of both the atomization energies and relative stability of the isomers of gold clusters. Functionals including SO coupling effect will overestimate the atomization energies of gold clusters compared with those just including the scalar relativistic (SC) effect. On the other hand, hybrid functionals will underestimate the atomization energies compared with those of the corresponding pure functionals. For the calculation of the relative stability of the different isomers, many functionals not including SO coupling will predict the wrong stability order. In addition, SO correction to the atomization energy of the cluster (ΔESO) has a weak dependence on the choice of functional. A linear relationship was established between ΔESO and the number of Au atoms and Au−Au bonds in the cluster. The relationship indicates that inclusion of SO coupling will favor the isomer with more Au−Au bonds. Among all of the functionals evaluated, the SO TPSSh method has the best overall performance, and SC M06-L also performs well, although it predicts that the two isomers of Au3 are almost degenerate in energy.

Propane dehydrogenation over perfect Ga2O3(100) was studied in detail by periodic density functional theory (DFT) calculations. It was found that the initial C−H bond activation mainly follows a radical mechanism that the two-coordinated surface oxygen site (O(2)) abstracts a hydrogen atom from propane with the formation of propyl radical and hydroxyl group (O(2)H). Physically adsorbed propyl radical can easily form propoxide or propylgallium intermediate. Subsequently, propene is formed by a second H abstraction from propyl, propoxide, or propylgallium by surface oxygen and Ga sites. H abstraction by O(2) site always has low energy barrier. However, it is difficult for the hydrogen atoms in the hydroxyl groups to leave the surface in the form of either H2 or H2O. In addition, propene formed through H abstraction by oxygen site has high adsorption energy and is prone to further dehydrogenation or oligomerization, leading to fast deactivation of the catalyst. On the other hand, the formation of H2 from GaH and hydroxyl group is much easier, although the formation of GaH has to overcome high energy barrier. Thus, there is a shift of rate-determining step for propane dehydrogenation: at the initial stage of the reaction, the rate-determining step is H abstraction by oxygen sites and then it shifts to H abstraction from various propyl sources by Ga sites to form gallium hydrides after the surface oxygen sites are consumed. Our results also indicate that dehydrogenation of propane mainly follows a direct dehydrogenation mechanism (DDH), whereas oxidative dehydrogenation (ODH) is energetically less feasible but cannot be ruled out in the presence of mild oxidant such as CO2.

FeO2− anions were produced by co-condensation of laser-ablated iron atoms and electrons with dioxygen in excess argon at 6 K. A photosensitive absorption at 870.6 cm−1 is assigned to the antisymmetric OFeO stretching vibration (ν3) of the inserted FeO2− anion trapped in solid argon. On the basis of the observed ν3 vibrational frequencies for Fe16O2 and Fe18O2, the anion is estimated to be linear. Due to the severe symmetry-breaking problems of the reference wave function, calculations with single-reference methods, including various DFT and post-HF methods, are unreliable for this molecule. However, the state-averaged multireference MRCI method, which incorporates both dynamical and nondynamical correlations, predicted that the anion has a linear doublet ground state, consistent with the experimental observations.

CdSO4 and ZnSO4 as co-modifiers of RuLa/SBA-15 lead to improved catalysts for the partial hydrogenation of benzene to cyclohexene. Based on the experimental results and theoretical calculations, it is shown that CdSO4 acts as surface modification, suppressing more the adsorption of cyclohexene than that of benzene, while the function of ZnSO4 is mainly the stabilization of cyclohexene in the liquid phase, accelerating the desorption as well as hindering the re-adsorption of cyclohexene.CdSO4 and ZnSO4 act as co-modifiers for RuLa/SBA-15 catalyst to enhance the selectivity in the partial hydrogenation of benzene to cyclohexene suppressing re-adsorption and stabilizing cyclohexene, respectively.Download high-res image (75KB)Download full-size image

In the present computational study by using the density functional theory (DFT) method, we found that silylboranes, which have metal-free Lewis acid centers only, can break the C–H bond of the exceedingly unreactive methane. The study shows that, unlike the activation mechanism of small molecules by the frustrated Lewis pairs (FLPs), the Lewis acidic boron center plays a key role in breaking the C–H bond of methane. Detailed analyses indicate that in the transition state the C–H bond is substantially activated by the empty 2p orbital of boron (2pB) primarily due to the orbital interaction between the C–H σ-bonding orbital and 2pB. On the other hand, the orbital interaction between the C–H σ-anti-bonding orbital and the B–Si σ-bonding orbital also contributes to the activation but plays a minor role. A statistical method was used to find the relationship between the reactivity of 57 silylboranes and their electronic properties. The results indicate that the boron center does have more prominent effect on the reactivity, especially the occupancy (nB2p) and energy (εB2p) of 2pB, where lowering nB2p and εB2p will increase the reactivity of the silylboranes. Based on the activation mechanism and taking kinetic and thermodynamic possibilities, as well as the possible side reactions, into consideration, three silylboranes suitable for methane activation under mild experimental conditions were designed. The analogous line of thought can be used as a hint for further experimental realizations, even under ambient conditions. This strategy can also be expected to be transplanted to more extensive C–H activation of hydrocarbons.

Formic acid decomposition (FAD) reaction has been an innovative way for hydrogen energy. Noble metal catalysts, especially palladium-containing nanoparticles, supported or unsupported, perform well in this reaction. Herein, we considered the simplest model, wherein one Pd atom is used as the FAD catalyst. With high-level theoretical calculations of CCSD(T)/CBS quality, we investigated all possible FAD pathways. The results show that FAD catalyzed by one Pd atom follows a different mechanism compared with that catalyzed by surfaces or larger clusters. At the initial stage of the reaction, FAD follows a dehydration route and is quickly poisoned by CO due to the formation of very stable PdCO. PdCO then becomes the actual catalyst for FAD at temperatures approximately below 1050 K. Beyond 1050 K, there is a switch of catalyst from PdCO to Pd atom. The results also show that dehydration is always favoured over dehydrogenation on either the Pd-atom or PdCO catalyst. On the Pd-atom catalyst, neither dehydrogenation nor dehydration follows the formate mechanism. In contrast, on the PdCO catalyst, dehydrogenation follows the formate mechanism, whereas dehydration does not. We also systematically investigated the performance of 24 density functional theory methods. We found that the performance of the double hybrid mPW2PLYP functional is the best, followed by the B3LYP, B3PW91, N12SX, M11, and B2PLYP functionals.

Quantum chemical calculations were carried out on CO oxidation catalyzed by a single gold atom. To investigate the performance of density functional theory (DFT) methods, 42 DFT functionals have been evaluated and compared with high-level wavefunction based methods. It was found that in order to obtain accurate results the functionals used must treat long range interaction well. The double-hybrid mPW2PLYP and B2PLYP functionals are the two functionals with best overall performance. CAM-B3LYP, a long range corrected hybrid GGA functional, also performs well. On the other hand, the popular B3LYP, PW91, and PBE functionals do not show good performance and the performance of the latter two are even at the bottom of the 42 functionals. Our accurate results calculated at the CCSD(T)/aug-cc-pVTZ//mPW2PLYP/aug-cc-pVTZ level of theory indicate that Au atom is a good catalysis for CO oxidation. The reaction follows the following mechanism where CO and O2 adsorb on Au atom forming an Au(OCOO) intermediate and subsequently O2 transfer one oxygen atom to CO to form CO2 and AuO. Then AuO reacts with CO to form another CO2 to complete the catalytic cycle. The overall energy barrier at 0 K for the first CO oxidation step (Au + CO + O2 → AuO + CO2) is just 4.8 kcal mol−1, and that for the second CO oxidation step (AuO + CO → Au + CO2) is just 1.6 kcal mol−1.

Infrared spectra of mass-selected homoleptic nickel carbonyl cluster cations including dinuclear Ni2(CO)7+ and Ni2(CO)8+, trinuclear Ni3(CO)9+ and tetranuclear Ni4(CO)11+ are measured via infrared photodissociation spectroscopy in the carbonyl stretching frequency region. The structures are established by comparison of the experimental spectra with simulated spectra derived from density functional calculations. The Ni2(CO)7+ cation is characterized to have an unbridged asymmetric (OC)4Ni–Ni(CO)3+ structure with a Ni–Ni single bond. The Ni2(CO)8+ cation has a Ni–Ni half-bonded D3d structure with both nickel centers exhibiting an 18-electron configuration. The trinuclear Ni3(CO)9+ cluster cation is determined to have an open chain like (OC)4Ni–NiCO–Ni(CO)4 structure. The tetranuclear Ni4(CO)11+ cluster cation is determined to have a tetrahedral structure with two-center and three-center bridge-bonded carbonyl units. These nickel carbonyl cluster cations all involve trigonal pyramid like Ni(CO)4 building blocks that satisfy the 18-electron configuration of the nickel centers.

Infrared spectra of mass-selected homoleptic iron carbonyl cluster cations including mononuclear Fe(CO)5+ and Fe(CO)6+, dinuclear Fe2(CO)8+ and Fe2(CO)9+, and trinuclear Fe3(CO)12+ are measured via infrared photodissociation spectroscopy in the carbonyl stretching frequency region. The structures are established by comparison of the experimental spectra with simulated spectra derived from density functional calculations. Only one IR band is observed for the Fe(CO)5+ cation, which is predicted to have a C4v structure. The Fe(CO)6+ cation is determined to be a weakly bound complex involving a Fe(CO)5+ core ion. In contrast to neutral clusters which have symmetric structures with two and three bridging carbonyl ligands, the dinuclear Fe2(CO)8+ and Fe2(CO)9+ cations are characterized to have unbridged asymmetric (OC)5Fe–Fe(CO)n+ (n = 3 and 4) structures. The trinuclear Fe3(CO)12+ cluster cation is determined to have an open chain like (OC)5Fe–Fe(CO)2–Fe(CO)5 structure instead of the triangular structure with two bridging CO groups for the Fe3(CO)12 neutral. The di- and trinuclear cluster cations all involve a square pyramid like Fe(CO)5 building block that satisfies the 18-electron configuration of this iron center. The Fe(CO)5 building block is isolobal to the CH3 fragment in hydrocarbon chemistry, the Fe2(CO)9+ and Fe3(CO)12+ cluster cations may be considered through isolobality to be metal carbonyl analogues of the ethyl and isopropyl cations.

Infrared spectra of mass-selected homoleptic dinuclear iron carbonyl cluster anions Fe2(CO)n− (n = 4–9) are measured via infrared photodissociation spectroscopy in the carbonyl stretching frequency region. The cluster anions are produced via a laser vaporization supersonic cluster source. Density functional calculations have been performed and the calculated vibrational spectra are compared to the experimental data to identify the gas-phase structures of the cluster anions. The experimentally observed Fe2(CO)n− (n = 4–7) cluster anions are characterized to have unusual asymmetric (OC)4Fe–Fe(CO)n−4 structures, which also correspond to the computed lowest energy structures. The experimentally observed Fe2(CO)8− cluster anion is determined to have an unbridged structure instead of the previously reported dibridged structure. The Fe2(CO)9− cluster anion is determined to involve a Fe2(CO)8− core anion that is solvated by an external CO molecule. Bonding analysis indicates that these anions each have a Fe–Fe single bond to satisfy the 18-electron configuration of one iron center. The results provide important new insight into the structure and bonding mechanisms of transition-metal carbonyl clusters.