Reduced Order Machine Learning Models for Accurate Prediction of CO2 Capture in Physical Solvents.

CO2 sorption in physical solvents is one of the promising approaches for carbon capture from highly concentrated CO2 streams at high pressures. Identifying an efficient solvent and evaluating its solubility data at different operating conditions are highly essential for effective capture, which generally involves expensive and time-consuming experimental procedures. This work presents a machine learning based ultrafast alternative for accurate prediction of CO2 solubility in physical solvents using their physical, thermodynamic, and structural properties data. First, a database is established with which several linear, nonlinear, and ensemble models were trained through a systematic cross-validation and grid search method and found that kernel ridge regression (KRR) is the optimum model. Second, the descriptors are ranked based on their complete decomposition contributions derived using principal component analysis. Further, optimum key descriptors (KDs) are evaluated through an iterative sequential addition method with the objective of maximizing the prediction accuracy of the reduced order KRR (r-KRR) model. Finally, the study resulted in the r-KRR model with nine KDs exhibiting the highest prediction accuracy with a minimum root-mean-square error (0.0023), mean absolute error (0.0016), and maximum R2 (0.999). Also, the validity of the database created and ML models developed is ensured through detailed statistical analysis.

Correction: Metal-organic-framework derived Co-Pd bond is preferred over Fe-Pd for reductive upgrading of furfural to tetrahydrofurfuryl alcohol.

Correction for 'Metal-organic-framework derived Co-Pd bond is preferred over Fe-Pd for reductive upgrading of furfural to tetrahydrofurfuryl alcohol' by Saikiran Pendem et al., Dalton Trans., 2019, 48, 8791-8802, https://doi.org/10.1039/C9DT01190K.

Leveraging Cu/CuFe2O4-Catalyzed Biomass-Derived Furfural Hydrodeoxygenation: A Nanoscale Metal-Organic-Framework Template Is the Prime Key.

Enormous efforts have been initiated in the production of biobased fuels and value-added chemicals via biorefinery owing to the scarcity of fossil resources and huge environmental synchronization. Herein, non-noble metal-based metal/mixed metal oxide supported on carbon employing a metal-organic framework as a sacrificial template is demonstrated for the first time in the selective hydrodeoxygenation (HDO) of biomass-derived furfural (FFR) to 2-methyl furan (MF). The aforementioned catalyst (referred to as Cu/CuFe2O4@C-A) exhibited extraordinary catalytic proficiency (100% selectivity toward MF) compared with the conventional Cu/CuFe2O4@C-B catalyst which was prepared by the wet impregnation method. High-resolution transmission electron microscopy and synchrotron X-ray diffraction studies evidenced the existence of both metal (Cu) and mixed metal oxide (CuFe2O4) phases, in which the metal could help in hydrogenation to alcohol and metal oxide could assist in the hydroxyl group removal step during HDO reaction. The stabilization of encapsulated metal/metal oxide nanoparticles in the carbon matrix, modulation of the electronic structure, and regulation of geometric effects in the Cu/CuFe2O4@C-A are thought to play an important role in its excellent catalytic performance, confirmed by X-ray photoelectron spectroscopy and X-ray absorption spectroscopy investigations. Furthermore, the structure and activity interconnection was confirmed by in situ attenuated total reflection-IR studies, which manifested the strong interfacial interaction between FFR and the Cu/CuFe2O4@C-A catalyst. This finding was further supported by NH3 temperature-programmed desorption analysis, which suggested that the presence of more Lewis/weak acidic sites in this catalyst was beneficial for the hydrogenolysis step in HDO reaction. Additionally, H2 temperature-programmed reduction studies revealed that the adsorption of H2 was stronger on the Cu/CuFe2O4@C-A than that over the conventional Cu/CuFe2O4@C-B catalyst; thus, the former catalyst promoted activation of H2. A detailed kinetic analysis which demonstrated the lower activation energy barrier along with dual active sites attributed for the activation of the two separate reactions in the HDO process on the Cu/CuFe2O4@C-A catalyst. This work has great implication in developing a highly stable catalyst for the selective upgradation of biomass without deactivation of metal sites in extended catalytic cycles and opens the door of opportunity for developing a sustainably viable catalyst in biomass refinery industries.

Porous Organic Polymer-Driven Evolution of High-Performance Cobalt Phosphide Hybrid Nanosheets as Vanillin Hydrodeoxygenation Catalyst.

Hydrodeoxygenation (HDO) is a promising route for the upgrading of bio-oils to eco-friendly biofuel produced from lignocellulose. Herein, we report the sequential synthesis of a hybrid nanocatalyst CoxP@POP, where substoichiometric CoxP nanoparticles are distributed in a porous organic polymer (POP) via solid-state phosphidation of the Co3O4@POP nanohybrid system. We also explored the catalytic activity of the above two nanohybrids toward the HDO of vanillin, a typical compound of lignin-derived bio-oil to 2-methoxy-4-methylphenol, which is a promising future biofuel. The CoxP@POP exhibited superior catalytic activity and selectivity toward desired product with improved stability compared to the Co3O4@POP. Based on advanced sample characterization results, the extraordinary selectivity of CoxP@POP is attributed to the strong interaction of the cation of the CoxP nanoparticle with the POP matrix and the consequent modifications of the electronic states. Through attenuated total reflectance-infrared spectroscopy, we have also observed different interaction strengths between vanillin and the two catalysts. The decreased catalytic activity of Co3O4@POP compared to CoxP@POP catalyst could be attributed to the stronger adsorption of vanillin over the Co3O4@POP catalyst. Also from kinetic investigation, it is clearly demonstrated that the Co3O4@POP has higher activation energy barrier than the CoxP@POP, which also reflects to the reduction of the overall efficiency of the Co3O4@POP catalyst. To the best of our knowledge, this is the first approach in POP-encapsulated cobalt phosphide catalyst synthesis and comprehensive study in establishing the structure-activity relationship in significant step-forwarding in promoting biomass refining.

Metal-organic-framework derived Co-Pd bond is preferred over Fe-Pd for reductive upgrading of furfural to tetrahydrofurfuryl alcohol.

Combined noble-transition metal catalysts have been used to produce a wide range of important non-petroleum-based chemicals from biomass-derived furfural (as a platform molecule) and have garnered colossal research interest due to the urgent demand for sustainable and clean fuels. Herein, we report the palladium-modified metal-organic-framework (MOF) assisted preparation of PdCo3O4 and PdFe3O4 nanoparticles encapsulated in a graphitic N-doped carbon (NC) matrix via facile in situ thermolysis. This provides a change in selectivity with superior catalytic activity for the reductive upgrading of biomass-derived furfural (FA). Under the optimized reaction conditions, the newly designed PdCo3O4@NC catalyst exhibited highly efficient catalytic performance in the hydrogenation of furfural, providing 100% furfural conversion with 95% yield of tetrahydrofurfuryl alcohol (THFAL). In contrast, the as-synthesized Pd-Fe3O4@NC afforded a THFAL yield of 70% after an 8 h reaction with four consecutive recycling tests. Based on different characterization data (XPS, H2-TPR) for nanohybrids, we can conclude that the presence of PdCo-Nx active sites, and the multiple synergistic effects between Co3O4 and Pd(ii), Co3O4 and Pd0, as well as the presence of N in the carbonaceous matrix, are responsible for the superior catalytic activity and improved catalyst stability. Our strategy provides a facile design and synthesis process for a noble-transition metal alloy as a superior biomass refining, robust catalyst via noble metal modified MOFs as precursors.

Surface and bulk aspects of mixed oxide catalytic nanoparticles: oxidation and dehydration of CH(3)OH by polyoxometallates.

The molecular structures and surface chemistry of mixed metal oxide heteropolyoxo vanadium tungstate (H(3+x)PW(12-x)V(x)O(40) with x = 0, 1, 2, and 3) Keggin nanoparticles (NPs), where vanadium is incorporated into the primary Keggin structure, and supported VO(x) on tungstophosphoric acid (TPA, H(3)PW(12)O(40)), where vanadium is present on the surface of the Keggin unit, were investigated with solid-state magic angle spinning (51)V NMR, FT-IR, in situ Raman, in situ UV-vis, CH(3)OH temperature-programmed surface reaction (TPSR), and steady-state methanol oxidation. The incorporated VO(x) unit possesses one terminal V horizontal lineO bond, four bridging V-O-W/V bonds, and one long V-O-P bond in the primary Keggin structure, and the supported VO(x) unit possesses a similar coordination in the secondary structure under ambient conditions. The specific redox reaction rate for VO(x) in the Keggin primary structure is comparable to that of bulk V(2)O(5) and the more active supported vanadium oxide catalysts. The specific acidic reaction rate for the WO(x) in the TPA Keggin, however, is orders of magnitude greater than found for bulk WO(3), supported tungsten oxide catalysts, and even the highly acidic WO(3)-ZrO(2) catalyst synthesized by coprecipitation of ammonium metatungstate and ZrO(OH)(2). From CH(3)OH-TPSR and in situ Raman spectroscopy it was found that incorporation of vanadium oxide into the primary Keggin structure is also accompanied by the formation of surface VO(x) species at secondary sites on the Keggin outer surface. Both CH(3)OH-TPSR and steady-state methanol oxidation studies demonstrated that the surface VO(x) species on the Keggin outer surface are significantly less active than the VO(x) species incorporated into the primary Keggin structure. The presence of the less active surface VO(x) sites in the Keggins, thus, decreases the specific reaction rates for both methanol oxidation and methanol dehydration. During methanol oxidation/dehydration (O(2)/CH(3)OH = 2.17, T = 225 degrees C), in situ UV-vis diffuse reflectance spectroscopy revealed that vanadium oxide is primarily present as the V(+5) cation, which reflects the Mars-van Krevelen redox mechanism and rapid reoxidation by molecular O(2). The bulk TPA Keggin structure becomes more disordered and less thermally stable as the vanadium content increases. Although surface polyaromatic carbon forms during methanol oxidation on the Keggin surfaces, its influence on the reaction kinetics seems minimal as the carbon content diminishes as the vanadium oxide content increases and the reaction temperature is raised. No relationships were found between the electronic structure (UV-vis E(g) values) and TOF(redox) and TOF(acid) (TOF = turnover frequency) kinetics, which reflect the complexity of H(3+x)PW(12-x)V(x)O(40) Keggins. The overall catalytic performance of the H(3+x)PW(12-x)V(x)O(40) Keggin materials results from a complex interplay among the presence of redox vanadium (as secondary surface VO(x) species and substituted VO(x) sites in the primary Keggin NP structure), structural disorder of the Keggin NPs, exposed surface acid and redox sites, and coke deposition. These new insights reveal that the Keggin heteropolyoxometallates are much more complex than originally thought and that care must be taken in using Keggins as model mixed metal oxide NPs in catalytic kinetic and theoretical studies because their surface and bulk structures are dynamic under the reaction conditions.