This work proposed non-dispersive solvent extraction to recover acrylonitrile from acrylonitrile-containing wastewater from ABS production by using styrene as the extractant. The effects of temperature, aqueous phase and organic phase flow rate on the outlet acrylonitrile concentration were systematically studied. The results showed that lower flow rate of the aqueous phase, and higher flow rate of the organic phase and temperature favored acrylonitrile recovery. The overall mass transfer coefficient was determined, ranging from 6.97×10-6 to 2.02×10-5 m s-1, and continuous 100 h operation confirmed long‐term stability. After optimization, the outlet concentration was reduced to below 10 g m-3, which can meet the subsequent biochemical requirements. Four machine learning models (ANN, GPR, XGBoost, RF) were established; and the GPR precision was the highest (R2=0.9997). SHAP analysis identified the aqueous phase flow rate as the most important factor. The NSGA-II algorithm was then employed to calculate the Pareto front of the optimal combination of the minimum outlet concentration and the maximum efficiency, providing a quantitative optimization scheme for industrial operation. Process intensification evaluation showed NDSX achieved ~594-fold higher volumetric capacity than acid hydrolysis pretreatment currently used in industry. The styrene‐based extractant has the potential to be directly recycled to the ABS polymerization unit, offering a promising route for resource recovery and circular utilization.

Constructing strong metal-support interaction (SMSI) between gold nanoparticles and conventionally nonreducible oxide supports presents a significant challenge, and the mechanistic understanding of SMSI in selective catalytic reactions remains limited. Herein, we demonstrate a carbon induced SMSI between the irreducible oxide alumina and Au nanoparticles (Au/Al2O3@C-SMSI). The resulting system exhibits electronic and geometric features analogous to classical SMSI, with the driving force inherently rooted in the nature of Au-Al and Au-O-Al interfacial bonding. The Au/Al2O3@C-SMSI catalyst exhibits exceptional selectivity toward the partial oxidation product 5-hydroxymethyl-2-furancarboxylic acid (HMFCA) in the oxidation of 5-hydroxymethylfurfural (HMF), achieving a remarkable HMFCA yield of 98.5%. Advanced characterization combined with density functional theory calculations reveals that the high selectivity stems from the synergistic differential activation of C─H bonds and competitive adsorption of HMFCA on Au and Al sites. Furthermore, the Au/Al2O3@C-SMSI catalyst maintain high activity without significant deactivation after six consecutive catalytic cycles using industrial-grade HMF (85% purity) as the substrate, demonstrating strong antipoisoning capability and promising potential for industrial application. This study offers an innovative approach for establishing SMSI on nonreducible oxide supports and opens new avenues for designing highly active and selective heterogeneous catalysts.

Chemical recycling of plastics is typically hampered by energy-intensive processes requiring harsh conditions. Here, we report an electrified non-thermal plasma (NTP) catalysis strategy for the selective hydrogenolysis of polycaprolactone (PCL) into valuable caprolactones at ambient pressure and without solvents. By replacing high-pressure H2 and noble-metal catalysts with an atmospheric methane plasma over a Ni/HY catalyst, this approach achieves complete PCL conversion and a remarkable 93.1% yield of caprolactones at 150°C utilizing solely plasma-generated heat. In situ hydrogen radicals provided by the plasma and Ni sites regenerate Brønsted acid sites, driving the selective dissociation of the PCL alkoxy bond and enabling the preferential formation of γ-caprolactone (γ-CL) both kinetically and thermodynamically at low temperatures. Simultaneously, plasma polarization promotes electron redistribution between HY and PCL, significantly lowering the dissociation energy barrier of PCL. Compared with thermal catalysis at 200°C, synergistic plasma catalysis reduces the PCL activation energy barrier from 2.29 to 1.71 eV and increases the PCL conversion by 8.77-fold. We further demonstrate the scalability of this process by operating a 100 g batch system powered by photovoltaics. This work establishes a synergistic pathway integrating methane cracking with waste plastic upcycling, offering a practical route for closed-loop polymer upcycling.

Rapid and robust activation of iron-based catalysts is critical for industrial Fischer-Tropsch synthesis. This work introduces a multi-stage, counter-current continuous fluidized-bed system that enables uninterrupted catalyst activation, addressing key limitations of conventional batch fluidized beds. Compared with previous studies that mainly addressed batch activation or conventional fluidized-bed activation, the novelty of this work lies in the combined cold-flow, CFD, and hot-flow validation of a multi-stage counter-current continuous fluidized-bed reactor for iron-based catalyst activation. Cold-flow tracer-based RTD experiments were conducted in a transparent acrylic reactor, and hot-flow activation tests were performed in a stainless-steel reactor under syngas at 265 degrees C and 0.2 MPa. Cold-flow experiments show that the inter-stage overflow design ensures unidirectional particle transfer, markedly suppresses back-mixing, and improves plug-flow behavior. The dimensionless variance of the residence time distribution decreases from 1.0 to 0.475, while the fitted tanks-in-series number increases from 1.0 to 2.34. CFD simulations accurately reproduce the observed hydrodynamics and extend the analysis to operating limits, predicting an upper gas-velocity threshold of 0.29 m/s under cold-flow conditions. Hot-flow simulations based on the validated model identify a slightly lower limit of 0.27 m/s, which provides practical guidance for reactor design and scale-up. Hot-flow experiments confirm stable fluidization at 265 degrees C and 0.2 MPa with continuous catalyst feeding and withdrawal. The catalysts activated in the multi-stage continuous system exhibit CO conversions of 51 to 53% and CH4 selectivities of approximately 1% in a stirred-tank evaluation reactor, achieving performance comparable to that obtained with our batch fluidized bed activation method.

Complete oxidation of propane has attracted widespread attention in the treatment of volatile organic compounds (VOCs) from Fischer-Tropsch synthesis tail gas. Pt-based catalysts have been considered highly promising catalysts for the catalytic oxidation of propane due to their high-activity. However, despite the recognized importance of the synergistic effect between Pt nanoparticles (NPs) and support materials, the detailed reaction mechanism investigation remains limited. In this work, we demonstrate that the Pt/Al2O3 catalyst exhibits improved propane catalytic performance, which is attributed to the synergistic interaction between Pt NPs and surface hydroxyl groups on the support. The improved Pt dispersion in Pt/Al2O3 catalysts with smaller Pt NPs facilitates the spillover of gaseous oxygen species and the stabilization of active sites for propane oxidation. Furthermore, in-situ DRIFTS characterization and density functional theory (DFT) calculation confirm that surface hydroxyl groups on Al2O3 effectively reduce the activation barrier for C-H bond cleavage, thereby improving the reactivity of propane molecules. This work provides a novel insight into metal-support synergy for propane oxidation, which offers strategic guidance for the rational design of efficient catalysts in hydrocarbon conversion processes.

The catalytic hydrogenation-rearrangement of furfural (FFA) to cyclopentanone (CPO) and cyclopentanol (CPL) holds significant promise for biomass upgrading and the valorization of renewable resources. However, the complexity of the reaction network makes product selectivity highly sensitive to catalyst concentration. Under high catalyst loading, basic sites inhibit the protonation step of furfuryl alcohol (FA), which limits the feasibility of continuous production in fixed-bed reactors. In this study, we constructed a Mo/CuZnZr catalyst via a controlled MoO x modification strategy for catalyst-intensive systems, which efficiently catalyzed the hydrogenation-rearrangement of FFA to CPO and CPL in a continuous-flow reactor, achieving a total yield of 82.6% and demonstrating excellent stability over a 300-h long-term operation. Detailed experimental investigations revealed that atomically dispersed Mo species preferentially occupy basic hydroxyl sites on ZrO2, forming Zr-O-Mo bonds that reconstruct the acidity of the catalyst and enhance the protonation of FA under high catalyst loading. This study presents a feasible strategy for continuously upgrading FFA and highlights the potential of rational catalyst design in the selective transformation of biomass-derived platform molecules.

The direct esterification reaction of CO2 with methanol to produce dimethyl carbonate (DMC) provides a sustainable pathway for the resourceful utilization of CO2. However, the chemical stability of CO2 makes the reaction kinetically difficult with low methanol conversion and DMC yields. In this study, a series of catalysts (denoted as M-CeO2, M = Fe, Cu, Co, La, Zr, Ni, and Al) with oxygen vacancy defects were constructed by a doping strategy, which provided critical sites for the adsorption and activation of methanol and CO2. Additionally, 2-cyanopyridine (2-CP) was employed as a dehydrating agent to facilitate the esterification reaction through dehydration. Under the optimized experimental conditions, the conversion of methanol was 61.9% and the DMC yield was 61.6%. The experimental and characterization results show that the extensive oxygen vacancies and base sites on the surface of the 5% Co-CeO2 catalyst are crucial for activating methanol and CO2, which is the main reason for the high catalytic activity of the catalyst. In situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) effectively demonstrated the efficient activation of methanol and CO2 on the catalyst and further confirmed the formation of monomethyl carbonate species during the reaction process. The high catalytic performance was also demonstrated in the direct esterification of other monohydric alcohols, thus confirming the broad applicability of the catalyst. This study presents a viable approach for the efficient utilization of CO2 resources, offering a promising solution for the sustainable management of greenhouse gas.

Polyesters, especially polyethylene terephthalate (PET), are widely used in plastic bottles and clothing fibers because of their stability and cost-effectiveness. Upcycling waste polyesters into value-added materials not only solves the environmental crisis but also realizes significant economic interests. Here, we report a step-economic two-step catalytic process for the upcycling of waste polyester materials, specifically PET, into 1,4-cyclohexanedimethanol (CHDM), an essential monomer for functional polyesters and key feedstock for the liquid crystal industry. The combination of PET methanolysis and hydrogenation of aromatic rings significantly reduces the reaction temperature and energy consumption of the depolymerization of PET, which introduces remarkable engineering benefits. By developing the CO-resistant bifunctional Ru/MnO2 and CuZnZr mixed oxide catalyst system, PET is demonstrated to be converted completely with the final yield of CHDM up to 78%. The implementation of the two-step catalytic process for transforming PET into CHDM holds significant implications for advancing the value chain of polyesters and contributing to waste material utilization for a renewable future.

Due to the depletion of traditional fossil resources and the increasing environmental problems, the utilization of sustainable biomass resources has become imperative. The selective oxidation of biomass‐derived 5‐hydroxymethylfurfural (HMF) to 2,5‐diformylfuran (DFF) has attracted much attention in recent years. However, the large‐scale production of DFF remains a challenge given the low dosage of HMF in previous research. In this contribution, a desirable DFF yield of 90.5% is obtained from concentrated HMF (130 °C, 2.0 MPa oxygen, 8 h) over the prepared Ru/MnO 2 catalyst with abundant oxygen vacancies ( O V ). The Ru/MnO 2 catalyst with high O V exhibits a weak MnO bond intensity resulting in high lattice oxygen ( O L ) reactivity, which promotes a high catalytic activity of HMF oxidation. Reusability studies show that the Ru/MnO 2 catalyst is stable and reusable. Furthermore, the direct production of DFF from fructose is also successfully realized via two consecutive steps, leading to a cost‐efficient oxidation of HMF solution and facilitating the industrial production of DFF. This research provides profound insights into the selective oxidation of HMF to DFF, thereby fostering further investigations into the conversion of fructose into high‐value chemicals.

Chemical reaction networks (CRNs) serve to describe the behavior of complex chemical reaction systems. Analyzing CRNs of a reactive system requires kinetic data that are typically obtained by time-consuming experiments or computational chemistry. Machine learning (ML) has emerged as a promising approach for rapid property prediction based on historical data. However, the accuracy of ML model predictions in kinetics remains a limitation for their application in CRN analysis. In this study, we integrate the cross-attention mechanisms in neural networks and CRN sensitivity and uncertainty analysis to enable the practical application of the ML models in reliable gas-phase CRN analysis. Specifically, a message-passing neural network (MPNN) architecture along with a cross-attention mechanism (CA-MPNN) was developed for accurate prediction of the reaction rate constants with prediction uncertainty. CA-MPNN model outperformed the conventional deep neural network architectures on most of the reaction property data sets. We combined reaction network sensitivity analysis and ML prediction uncertainty analysis to identify influential reactions with high-level uncertainty of the predicted rate constant, which are further calibrated using high-accuracy quantum chemistry methods to mitigate the problem of inaccurate machine learning predictions. Compared with the traditional workflow, this framework significantly reduces up to 80% computational cost to construct a reliable CRN in the demonstrated gas-phase pyrolysis and combustion applications.

The chemical conversion of 5-hydroxymethylfurfural (HMF) into high-value-added products offers a potential substitute for fossil-based products, which has garnered significant attention in light of energy depletion and environmental concerns. Among the various HMF derivatives, 2,5-bis(hydroxymethyl)furan (BHMF), applicable in polyester manufacturing, pharmaceutical intermediates, biodiesel, and other sectors, can be synthesized through the C=O hydrogenation of HMF. However, the inevitable humins formation and the over-hydrogenation/hydrogenolysis constrain the further application of this process. In this study, a series of hydroxyl-enriched Pt-based catalysts have been prepared and applied in the hydrogenation of HMF. For 1% Pt/CeO2, the selectivity for BHMF over 85% can be consistently maintained at 30 degrees C or at a substrate concentration of 100 g/L. Experiments and characterization studies indicate that hydroxyl groups presented on the catalyst surface could protect BHMF, inhibiting side reactions and thereby increasing carbon balance and main product selectivity. This superior selective catalytic performance was further demonstrated in the C=O hydrogenation over various aromatic aldehyde substrates, providing a novel approach for biomass compounds conversion to value-added chemicals and providing a perspective for catalyst design.

Understanding the fluidization behavior of binary mixtures containing non-spherical particles is crucial for optimizing and designing thermal conversion processes in fluidized beds. This study investigated the effect of non-spherical particle shape and mass fraction on plastic-sand binary mixture fluidization behavior. Fluidization characteristics including pressure drop, particle height distribution, expanded bed height, and bubble properties, were investigated and compared. A machine learning-aided image processing method was employed to segment non-spherical plastic particles from sand particles. The coarse-grained SuperDEM-CFD method was adapted to simulate the fluidization behavior of binary mixtures, and the results were compared with experimental data. The particle mixing process was also analyzed based on SuperDEM simulation results. This study found that particle shape and mass fraction significantly affect binary fluidization behavior and provide a detailed experimental dataset for validating CFD-DEM non-spherical particle simulation.

Cyclopentanone (CPO) and cyclopentanol (CPL), derived from renewable biomass resource hemicellulose, are considered sustainable chemicals that can substitute fossil-based products, which could be obtained through the hydrogenation-rearrangement reaction of furfural (FFA). However, the inevitable condensation reactions of FFA and furfuryl alcohol (FA) in aqueous solutions generate humins and result in carbon loss, limiting the industrial application of this process. Herein, a series of Cu-based catalysts were synthesized using the reverse co-precipitation method with various metal promoters and applied in the FFA hydrogenation-rearrangement reaction. The highest yield of 97.1 % CPO/CPL achieved in a typical reaction condition at 200 degrees C, 4 MPa, in 20 wt % isopropanol-water mixed solvent over CuZnZr catalyst and the recyclability test confirmed that the catalyst had excellent stability. Experimental studies and characterizations showed that the enhanced performance could be attributed to the adequate Lewis acid sites, regulated through the synergistic interaction between Cu, Zn, and Zr compounds. The simultaneous introduction of Zn and Zr promotes the Cu reduction, and results in more dispersed Cu0 and Cu+, providing an ample and reliable active site. This study outlines a viable method for the catalyst designing and valorizing biomass into high-value-added products.

5-Hydroxymethylfurfural(HMF) is a versatile chemical synthesized from glucose dehydration catalyzed by metal chloride (MClx) in deep eutectic solvents (DESs). However, the low glucose concentration and high catalyst dosage hinder large-scale HMF production. Herein, we report an aqueous DES of tetraethylammonium bromide(TEAB)-glucose for converting concentrated glucose (40 wt %, relative to TEAB) using ultra-dilute SnCl4 (0.25 mol %), achieving a 62 % yield of HMF. Ultra-dilute MClx-catalyzed selective conversion of glucose is feasible only when combining SnCl4 with Br-based DES, which is elucidated by density functional theory and molecular dynamic calculations. Using SnCl4 is essential due to its higher glucose isomerization activity than AlCl3 and CrCl3, which can be attributed to its low-barrier coordination with glucose and its barrier-free separation from fructose. Halide anions in DESs strongly interact with glucose, hindering the MClx-glucose coordination and thereby reducing MClx's activity for glucose isomerization. Consequently, Br-based DESs facilitate higher activity of MClx than Cl-based DESs, due to the weaker interaction between halide anion and glucose. In addition, we elucidated the side reactions including condensation, polymerization, and isomerization, and proposed a reaction network. Our findings clarify the differential activity of MClx and the impact of halide anions in DESs on MClx's activity.

The use of cellulose as a substitute for fossil resources has gained significant attention due to the depletion of nonrenewable energy sources and the increasing potential environmental crisis. The selective hydrogenolysis of cellulose to C-3 polyols, namely, 1,2-propanediol and glycerol, is one of the attractive biomass depolymerization and utilization pathways. However, coordinating sugar isomerization and retro-aldol condensation remains a challenge. In this paper, we manipulate the hydrogen spillover behavior of the Pd-Mo/TiO2 catalyst by tuning the density and aggregation states of Mo/Pd species to promote the yield of C-3 polyols in the one-pot cellulose hydrogenolysis reaction. The optimal Pd-Mo/TiO2 catalyst realizes a similar to 50% C-3 polyol yield with the total polyol yield approaching 70%, surpassing the performances of known heterogeneous catalysts. We demonstrate that the strong hydrogen spillover on the Pd/TiO2 facilitates the reduction of MoOX and enhances the formation of continuous oxygen vacancies on Ti-O-Mo sites as adjacent Lewis acid pairs that serve as the adsorption sites for the tridentate complex of hexose and efficiently catalyze the isomerization and subsequent retro-aldol condensation steps. The acceleration of glucose -> fructose isomerization and inhibition of undesirable condensation account for the high product yield.

Humins, an unavoidable by-product of the acid-catalyzed conversion of carbohydrates, significantly limits the yield of platform molecules and accounts for the majority of carbon loss in the biomass utilization processes. Therefore, the utilization and recovery of humins is highly importance to enhance the atom efficiency of biomass valorization. In this work, using potassium ferrate (K 2 FeO 4 ) to realize synchronous carbonization and graphitization, humins obtained from the dehydration of fructose is transformed into a novel porous biochar material and further applied to prepare noble metal based hydrogenation catalysts. By tuning the carbonization temperature, we demonstrate biochar calcined at 600 degrees C has large mass specific surface area while retaining abundant furan groups in the framework. With the high density of furan groups in structure, the Pd/biochar catalysts enhances the preferential adsorption capability of furanyl functional groups, thereby promoting the hydrogenation rate of furan compounds and improving the hydrogenation selectivity of furan rings. The demonstration of this novel application of humins will not only enhance the carbon efficiency of biomass conversion processes but also provide functionalized materials for catalytic applications.

This work employed a hybrid approach combining experiments, numerical simulation, and genetic algorithm (GA) to optimize the non-dispersive solvent extraction (NDSX) process of H2O2. 2 O 2 . A numerical simulation model with high accuracy was developed. The relative error between the simulation results and experimental data was controlled within 2.2 %. The results indicated the velocity of lumen side and temperature are the key factors affecting the extraction efficiency. The non-dominated sorting GA-II algorithm coupled with the numerical simulation was used to optimize the NDSX process. The Pareto optimal design of the NDSX process can be achieved within the water outlet concentration range of 13.44 to 550.2 kg m-3.-3 . Intensification factors of NDSX versus extraction towers are evaluated to be 17.9 and 13.2. In summary, this work provides a design framework for NDSX of H2O2, 2 O 2 , which is expected to simultaneously promote the mass transfer efficiency and safety of the H2O2 2 O 2 separation process.

The prediction of the thermodynamic and kinetic properties of elementary reactions has shown rapid improvement due to the implementation of deep learning (DL) methods. While various studies have reported the success in predicting reaction properties, the quantification of prediction uncertainty has seldom been investigated, thus compromising the confidence in using these predicted properties in practical applications. Here, we integrated graph convolutional neural networks (GCNN) with three uncertainty prediction techniques, including deep ensemble, Monte Carlo (MC)-dropout, and evidential learning, to provide insights into the uncertainty quantification and utility. The deep ensemble model outperforms others in accuracy and shows the highest reliability in estimating prediction uncertainty across all elementary reaction property data sets. We also verified that the deep ensemble model showed a satisfactory capability in recognizing epistemic and aleatoric uncertainties. Additionally, we adopted a Monte Carlo Tree Search method for extracting the explainable reaction substructures, providing a chemical explanation for DL predicted properties and corresponding uncertainties. Finally, to demonstrate the utility of uncertainty qualification in practical applications, we performed an uncertainty-guided calibration of the DL-constructed kinetic model, which achieved a 25% higher hit ratio in identifying dominant reaction pathways compared to that of the calibration without uncertainty guidance.

Filtered drag models have been extensively developed and applied in coarse grid simulation of fluidized beds. However, existing filtered drag models were mainly developed using the particle properties of Fluid Catalytic Cracking (FCC), and these models were found to perform less satisfactorily in predicting flow fields for low-density particles. To broaden the applicability of filtered drag models, this study developed a novel filtered drag model tailored explicitly for low-density Geldart A particles. Additionally, this study investigated the differences in filtering outcomes using various filter types. Statistical analyses revealed significant differences between filters, particularly between the median and other filters. The statistical skewness distribution of the filtered quantity datasets can lead to differences in results between median and mean filters. After analyzing one-marker and two-marker models, it was found that filter type has a more pronounced influence on model performance prediction differences at small filter sizes, while it becomes insignificant at larger filter sizes. Validation studies in fluidized bed simulations found that the two-marker model established based on the Median filter performs best in predicting performance. It is also found that more substantial drag corrections are required in the extremely dense phase domain for low-density particle fluidization systems.

This work employed a hybrid approach combining computational fluid dynamics (CFD), artificial neural network (ANN), and genetic algorithm (GA) to optimize the non-dispersed solvent extraction (NDSX) process of H2O2. A CFD model with high accuracy was developed. The relative error between the simulation results and experimental data was controlled within 2.2%. The results indicated the velocity of lumen side and temperature are the key factors affecting the extraction efficiency. A feed-forward back-propagation ANN model with 20 hidden layer neurons was trained to predict the NDSX simulation results. The determination coefficient R2 of the 1230 test set data reached 1.000. Compared with CFD simulation, the computational efficiency of ANN model was improved by three orders of magnitude. The non-dominated sorting GA-II algorithm coupled with the ANN model was used to optimize the NDSX process. The Pareto optimal design of the NDSX process can be achieved within the water outlet concentration range of 13.47 to 549.9 kg m-3. In summary, this work provides a design framework for NDSX of H2O2, which is expected to simultaneously promote the mass transfer efficiency and safety of the H2O2 separation process.