AIChE Journal

The direct and selective conversion of syngas into C2+ oxygenates is challenging due to the complex reaction network. Here, we report a robust relay system for the direct synthesis of methyl acetate (MA) from syngas, which combines CuZnAlOx/H‐ZSM‐5 for syngas to dimethyl ether (DME) with modified H‐MOR for DME carbonylation. The dehydration of methanol to DME on H‐ZSM‐5 significantly enhanced the hydrogenation of CO on CuZnAlOx, because of high DME equilibrium yields. Blocking of Brönsted acid sites with basic molecules or selective dealumination of 12‐membered rings in H‐MOR suppressed the zeolite coking. Besides, reaction temperatures above 240°C avoided H2O poisoning of carbonylation sites inside 8‐MR side pockets of H‐MOR, further benefiting the catalytic stability. Eventually, this relay system provided a high MA selectivity of 75% and an acetic acid selectivity of 13% at a CO conversion of 65%, outperforming reported catalysts.

Large language models (LLMs) are often criticized for lacking true “understanding” and the ability to “reason” with their knowledge, being seen merely as autocomplete engines. I suggest that this assessment might be missing a nuanced insight. LLMs do develop a kind of empirical “understanding” that is “geometry”‐like, which is adequate for many applications. However, this “geometric” understanding, built from incomplete and noisy data, makes them unreliable, difficult to generalize, and lacking in inference capabilities and explanations. To overcome these limitations, LLMs should be integrated with an “algebraic” representation of knowledge that includes symbolic AI elements used in expert systems. This integration aims to create large knowledge models (LKMs) grounded in first principles that can reason and explain, mimicking human expert capabilities. Furthermore, we need a conceptual breakthrough, such as the transformation from Newtonian mechanics to statistical mechanics, to create a new science of LLMs.

Microbubbles have been widely applied in various fields. Here, an oscillating feedback microreactor (OFM) was designed to produce microbubbles at high throughput (5–80 mL/min), where the hydrodynamics and mass transfer performance of gas–liquid two‐phase system were investigated. The hydrodynamics results showed that three secondary flows (oscillation, vortex, and feedback) could be effectively generated for inducing chaotic flow in the OFM, and the gas phase could be effectively broken up into small microbubbles. The bubble size was more sensitive to the liquid phase flow rate than the gas phase. Two dimensionless prediction formulas for bubble Sauter size were proposed based on gas–liquid flow ratio and Reynolds number at different liquid flow rates. The mass transfer experiments showed that the volumetric average mass transfer coefficient kLa was 1–3 orders of magnitude higher than those of conventional reactors.

Supported metal catalysts have been widely applied and commonly fabricated through the H2 reduction process. Herein, we develop a H2‐free room‐temperature discharge‐driven reduction (RT‐DR) reactor for fabricating supported metal catalysts at room temperature without H2. By RT‐DR reactor, a catalyst with pseudo‐boehmite (PB) as support (CdS/Pt/PB) is fabricated. In visible‐light‐driven photocatalytic H2O splitting to H2, CdS/Pt/PB shows a H2 evolution rate of 1132 μmol h⁻¹, which is greatly enhanced than that on catalyst prepared by traditional H2‐reduction (633 μmol h⁻¹). RT‐DR reactor is also used to prepare a catalyst with low sodium PB (LSPB) as support (CdS/Pt/LSPB). In visible‐light‐driven photocatalytic H2O splitting to H2, CdS/Pt/LSPB shows a H2 evolution rate of 2554 μmol h⁻¹, which is 2.5 times higher than that on catalyst prepared by traditional H2‐reduction (1029 μmol h⁻¹). Thus, RT‐DR reactor has high efficiency and universality in preparing catalysts, thus offering a great potential for commercialization.

A raw hydrogen mixture frequently results in a reduction in conversion efficiency and the generation of undesired by‐products. The application of advanced membrane technology has the potential to offer an economically viable solution for the recovery of hydrogen from such mixtures. BaZr1−x−yCexYyO3−δ is increasingly regarded as an optimal perovskite hydrogen permeable membrane. Nevertheless, the main drawback to its use in a larger scale is the extremely low hydrogen permeability and stability. An original perovskite material is proposed in this study, BaZr0.7Ce0.2Y0.1O3−δ‐Fx. A thermodynamic‐controlled sintering strategy (TCS) has been employed to inhibit the evaporation of metals from ceramic solids. The TCS directly caused the hydrogen permeation flux to reach 1.07 ml·min⁻¹ cm⁻², representing a fourfold improvement. Furthermore, F‐doping demonstrated enhanced performance at low and medium temperatures. The aforementioned successful strategy provides an effective path for the tailoring of perovskite materials and promotes its application for the industrial‐scale separation of hydrogen.

The development of a high‐performance electrocatalytic acetylene semi‐hydrogenation catalyst is the key to the selective removal of acetylene from industrial ethylene gas and non‐oil route to ethylene production. However, it is still hampered by the deactivation of the catalyst and hydrogen evolution interference. Here, we proposed an interface engineering strategy involving the Cu and cupric oxide nanoparticles dispersed on amorphous SiO2 (Cu/CuOx/SiO2) by a simple stöber method. x‐ray photoelectron spectroscopy demonstrated the strong interfacial interaction between cupric oxide nanoparticles and SiO2. The formed Cu‐O‐Si interface stabilized the Cuσ+ at high reduction potentials, thus improving the activity and stability of the acetylene reduction reaction, as confirmed by in situ Raman spectroscopy. Consequently, the electrochemical test results showed that at 0.5 M KHCO3, the maximum Faraday efficiency (FE) of ethylene on the optimized Cu/CuOx/SiO2 reached 96%. And ethylene FE remains above 85% at −100 mA cm⁻² for 40 h.

Biomass gasification for syngas production is a key operating unit in the biomass utilization process. However, its overall efficiency and stability are often restricted by the presence of complex impurities, including particulate matters (PMs) and tars. In this study, a highly integrated ceramic membrane‐based reactor was developed for high‐temperature syngas cleaning, enabling the efficient in situ removal of PMs and tars from bio‐vapors produced by biomass gasification. Specifically, a silicon carbide (SiC) membrane could separate PMs from biomass volatiles in situ, while a structured Ni15La5/S1‐SiC catalyst (nickel and lanthanum‐laden silicalite‐1 zeolite supported on SiC foam) facilitated the catalytic reforming of tars. Compared to other control reactors (i.e., those containing either a membrane or catalyst alone), the integrated reactor showed synergistic intensification in producing clean syngas from biomass gasification, achieving PM and tar removal efficiencies of up to ~97% and ~90%, and exhibited excellent stability in five‐cycle evaluations at 800°C.

A combined simulated moving bed (SMB) and distillation separation scheme is developed to recover 2,3‐butanediol (BDO) from a dilute fermentation broth. The scheme was integrated into a lignocellulosic biorefinery that produces hydrocarbon fuels from corn stover with BDO as an intermediate. BDO recovery is one of the most challenging processes in this biorefinery; and given the high associated energy duties, direct distillation is considered cost‐prohibitive. An alternative separation is SMB adsorption in nanoporous materials, which can reject 90% of the water and reduce subsequent distillation costs. Rigorous models were used to optimize the SMB and distillation simultaneously. The separation can be added to the biorefinery while keeping the projected minimum fuel selling price (MFSP) below $0.66 USD (US dollars) per liter gasoline‐equivalent ($2.50/GGE, gallon gasoline equivalent). Finally, sensitivity analyses were conducted to assess the effects of cost and lifetime of the adsorbent, titer concentration, and BDO purity.

The separation of methane (CH4) and nitrogen (N2) in coalbed methane (CBM) is a crucial process to produce methane as a clean energy source. Given the similar physicochemical properties of the two gases, efficient adsorptive separation of CH4 and N2 has stringent requirements on the pore chemistry and geometry of the adsorbents. In this study, we report two Zn‐pyrazolates frameworks, HIAM‐213 and HIAM‐214, which possess Lewis basic sites that exhibit precise recognition of CH4 over N2. The capability of these compounds in extracting CH4 from binary CH4/N2 mixtures has been confirmed through column breakthrough measurements, demonstrating a high recovery rate under mild regeneration conditions. Grand canonical Monte Carlo simulations and in situ Fourier transform infrared spectroscopy experiments provide insights into the selective adsorption mechanism by revealing stronger interactions between CH4 and Lewis acid sites.

Aminoxyl radicals electrocatalysis presents a sustainable method for oxidizing alcohols into high‐value products. Nonetheless, the requirement for high doses of aminoxyl radicals diminishes product purity and economic viability. This study synthesized methylimidazole‐functionalized 4‐acetylamino‐2,2,6,6‐tetramethylpiperidine‐N‐oxyl derivative (MIAcNH‐TEMPO) with a strongly electron‐withdrawing imidazole group and combined it with bioinspired nickel‐supported carbonaceous octopus tentacles for effective electrooxidation of alcohols, achieving high current density of 200 mA cm⁻², selectivity of 99%, and turnover frequency of 26,490 h⁻¹. In situ experiments and theoretical calculations indicated that the synergistic effect of Ni‐3dxz orbitals on the tentacle surface interacting with the π orbitals of MIAcNH‐TEMPO creates a strong d–π interaction, which effectively facilitating the creation of a locally intermediate‐enriched microenvironment, decreased the required quantity of aminoxyl radicals. Moreover, the high aqueous solubility of MIAcNH‐TEMPO reduces the difficulty of separation process. Scale‐up experiments conducted in a continuous flow electrolyzer showcased the potential of this strategy for practical applications.

Separating water‐in‐heavy oil (W/HO) emulsions at low (room) temperature is challenging when exploiting heavy oil. We propose an adaptable strategy for constructing Si/O‐doped demulsifiers. A nonionic demulsifier (APBMP) has been synthesized based on polysiloxane modified by allyl polyether and butyl acrylate. APBMP achieves 95.97% dehydration within 5 min for W/HO emulsions at 288.15 K and complete dehydration in 15 min at 323.15 K. Mechanistic studies found that doping Si/O into the demulsifier molecules increases the number of hydrogen bond sites, which enables the demulsifiers to quickly disperse natural stabilizers (e.g., asphaltenes) and replace them at the oil–water interfacial film. The demulsifiers prefer to occupy the interfacial sites rather than dissolve into the bulk oil or water phases. Driven by hydrogen‐bond‐dominated noncovalent interactions, the oil–water interfacial film is softened, reconstructed, and broken. These findings provide insights into developing novel materials for oil–water separations in a low‐carbon way.

We present a techno‐economic optimization model for the design and dynamic operation of proton exchange membrane (PEM) electrolyzers, for enabling cost‐effective hydrogen production. This model integrates a 0‐D model of the electrolyzer stack, process‐wide mass and energy balances, operational constraints, and an empirical relation to characterize degradation as a function of operating current density. Utilizing a decomposition‐based solution approach, the model predicts optimal electrolyzer size, operation, and necessary hydrogen storage to satisfy hydrogen demand across various technology and electricity price scenarios. Analysis for 2022 electricity prices and technology costs shows that including use‐dependent degradation raises the levelized cost of hydrogen (LCOH) from $4.56/kg to$6.60/kg and increases frequency of stack replacement (2 vs. 7 years). However, by 2030, we anticipate a significant reduction in LCOH to $2.50/kg due to lower capital expenses, leading to longer stack lifetimes and less hydrogen storage. The proposed modeling framework is adaptable to study other electrochemical systems relevant for decarbonization.

Spacer meshes in electrodialysis (ED) play a crucial role in influencing fluid and electric field distributions during mass transfer. This study employed finite element analysis using real‐world parameters to explore how spacer mesh geometry and material affect mass transport. Comparisons among different wire mesh configurations revealed increased fluid velocity near mesh wires and reduced electric field intensity nearby, enhancing overall transport efficiency. Increasing mesh count or wire diameter notably improves transport, with optimal results achieved when wire orientation aligns with fluid flow. Additionally, the study showed that spacer mesh conductivity significantly influences ED transport, particularly when it exceeds the conductivity of the solution. These findings advance the design and application of spacer meshes, offering valuable insights for future developments in ED technology.

Hydrogen, the simplest and most abundant element in the universe, is poised today to play a critical role in the transition toward a clean energy future. This perspective article explores the potential of hydrogen as a clean and sustainable energy carrier. First, the historical and modern hydrogen applications across various sectors are examined. Then, the article delves into the different methods for hydrogen production. Since conventional fossil fuel‐based production methods raise environmental concerns, the article explores promising alternatives for cleaner hydrogen generation, including electrolysis powered by renewable energy, biomass gasification, and photocatalytic reforming. The technical details and advantages/disadvantages of hydrogen production methods are discussed, as well as the ongoing research and development efforts to improve the existing techniques and to make them more sustainable, efficient, and cost‐effective. As the world strives for a cleaner energy future, hydrogen holds immense potential to play a transformative role.

Blending hydrogen into existing natural gas pipelines is considered the most feasible choice for long‐distance, large‐scale hydrogen transportation in the early stage of hydrogen economy development. To integrate the optimization of hydrogen‐blended natural gas pipeline network and subsequent hydrogen/natural gas separation process, this article presents a mixed‐integer nonlinear programming model, aiming to minimize the total annual project net cost. To tackle the computational complexity resulting from the large‐scale and nonlinear nature of practical design problems, a decomposition algorithm is tailored to the proposed model. Two case studies demonstrate that compared to stepwise model, the proposed pipeline‐separation integrated model offers economic benefits and practical value, incorporating separation processes and satisfying constraints of hydrogen demand, pressure and blending ratio requirements, which achieves an economically optimal design for both pipeline transportation and separation systems, and provides a viable solution for the broader application of hydrogen‐blended natural gas networks.

In this article, we propose a physics‐informed learning‐based Koopman modeling approach and present a Koopman‐based self‐tuning moving horizon estimation design for a class of nonlinear systems. Specifically, we train Koopman operators and two neural networks—the state lifting network and the noise characterization network—using both data and available physical information. The first network accounts for the nonlinear lifting functions for the Koopman model, while the second network characterizes the system noise distributions. Accordingly, a stochastic linear Koopman model is established in the lifted space to forecast the dynamic behaviors of the nonlinear system. Based on the Koopman model, a self‐tuning linear moving horizon estimation (MHE) scheme is developed. The weighting matrices of the MHE design are updated using the pretrained noise characterization network at each sampling instant. The proposed estimation scheme is computationally efficient, as only convex optimization needs to be solved during online implementation, and updating the weighting matrices of the MHE scheme does not require re‐training the neural networks. We verify the effectiveness and evaluate the performance of the proposed method via the application to a simulated chemical process.

Packed bed microreactors offer a promising platform for intensifying heterogeneously catalyzed reactions. To understand hydrodynamics therein, N2 or water flow was investigated experimentally through microreactors packed with glass beads in this work, corresponding to a microreactor to particle diameter ratio (D/d) of 1.29–25.12. The porosity of a single pellet string microreactor (D/d < 1.866) agrees with the literature's theoretical equation. For microreactors with larger D/d ratios, an empirical porosity correlation is proposed to address the dense packing nature of the bed. The existing correlations are inadequate to describe the pressure drop data in microreactors within the entire D/d ratios and modified Reynolds numbers (Rem < 291). At D/d ≥ 3, the measured pressure drop is described by the modified Ergun equation using properties of the bulk bed zone to exclude the wall effect. At D/d < 3, it can be predicted by introducing a correction term for the wall effect into the Ergun equation.

Structured catalysts exhibit the advantages of high diffusion efficiency and low heat transfer resistance, which have attracted increasing attention to non‐adiabatic gas–solid process. However, the metal‐supported coating catalysts face the problems of weaker bond strength and severe sintering, especially under the conditions of large flow rate and high temperature. Herein, metal@Silicalite‐1 structured catalysts with high adhesion and thermal stability were successfully prepared by hydroxylating the substrate with anatase. Rich surface Ti‐OH significantly strengthened the adhesion stability of the zeolite coating. In propane dehydrogenation reaction, the optimized PtZn@S‐1‐15Ti showed a high specific activity of 49.6 molC3H6·molPt⁻¹·s⁻¹ with propylene selectivity above 99% at 600°C. The introduction of anatase accelerated the aggregation of silicon sources and induced nucleation with growth content of zeolite increased by 3.6 times. It breaks the inherent contradiction between high loading amount and strong binding ability of coated catalysts, which broadens the avenues for industrial applications.

Solar‐driven photocatalysis is a promising strategy for clean hydrogen (H2) generation cooperated with selective organic synthesis. Lignin, rich in aromatic units and functional groups, serves as an ideal hole sacrificial agent and substrate, facilitating H2 evolution and yielding high‐value chemicals/fuels. To boost overall photocatalytic redox efficiency, thermal catalysis was further combined to enhance the transfer and activity of photo‐generated carriers. And a highly controllable Cu‐based catalyst was developed using technical lignin‐carbon as an electron buffer. The active‐pyrolyzed lignin‐carbon layer precisely regulated the crystal dispersion of Cu species on Cu/SiO2, simultaneously dynamically constructing active electron‐rich Cu⁰ and electron‐deficient Cuσ+ (1 < σ ≤ 2) sites. Excellent thermo‐photo redox performances were achieved, with an H2 evolution rate up to 1313.2 μmol·gcat⁻¹·h⁻¹ and a yield of 45.2% for C13–C16 aromatic dimers from lignin monomers. This study reveals the highly utilization of lignin in functional catalysts, as well as the efficient production of H2 and jet fuel precursors.

Purifying ultra‐high purity propylene (>99.995%) with an energy‐efficient adsorptive separation method is a promising yet challenging technology that remains unfulfilled. Instead of solely considering the effect of adsorbents on guest molecules, we propose a synergistic adsorption mechanism for the deep removal of propane and propyne, utilizing supramolecular interactions in both “host‐guest” and “guest‐guest” systems. Through modulation of the pore environment, Ni‐DMOF‐DM exhibits exceptionally high adsorption capacities for propane and propyne (171 and 197 cm³/g at ambient temperature and pressure, respectively), and unprecedented propane/propylene separation selectivity (2.74). Theoretical calculations confirm the geometric interactions of C‐H···π bonds and C‐H···O hydrogen bonds resulting from host‐guest interactions, alongside C‐H···H guest‐guest interactions within the confined pore space. Breakthrough experiments demonstrated that ultra‐high purity propylene (propane < 0.005% and propyne < 1.0 ppm) can be directly collected from ternary mixtures on Ni‐DMOF‐DM, achieving a productivity of up to 152.14 L/kg.

Liquid holdup is a crucial factor in the study of hydrodynamic behaviors in the micro‐packed bed reactor (μPBR). In this work, the values of liquid holdup are studied with the weighing method with good accuracy. The packed bed is a tube made of stainless steel with a length of 20 cm and an inner diameter of 4 mm, packed with 177–250 μm or 350–500 μm microbeads. The gas and liquid flow rates vary from 5 to 20 mL/min and 0.25 to 2 mL/min, respectively. A new hypothesis of the flow regions is proposed based on the experimental results. Furthermore, a new set of empirical correlation is built with great agreement, particularly for viscous liquids, whose viscosity ranges from 0.99 to 5.98 mPa·s, showing an atypical tendency.

The integrated design of the heat exchanger network (HEN) and organic Rankine cycle (ORC) system with new working fluids is a complex optimization problem. It involves navigating a vast design space across working fluid molecules, ORC processes, and networks. In this article, a new two‐stage reverse strategy is developed. The optimal HEN‐ORC configurations and operating conditions, and the thermodynamic properties of the hypothetical working fluid are identified by an equation of state (EOS) free HEN‐ORC model in the first stage. With two developed group contribution‐artificial neural network thermodynamic property prediction models, working fluid molecules are screened out in the second stage from a database containing more than 430,000 hydrofluoroolefins (HFOs). The presented method is employed in two cases, where new working fluids are found. The total annual cost of Case 1 is 12%–22% lower than the literature, and the power output of Case 2 is 5%–8% higher than the literature.

Despite the practical importance, 3D measurements of gas–solid distribution in fluidized beds calls for further breakthroughs. Here an approach combing a recently developed mobile electrical capacitance tomography (ECT) sensor with Fourier Neural Operator (FNO) is developed, in which the fluidized bed is divided into a series of cross‐sectional slices along axial direction. At any given instant, the gas–solid distribution in one slice is measured by mobile ECT and the others, meantime, are predicted by FNO pre‐trained using experimental data. We verified this approach via computational fluid dynamics (CFD) simulations and experimental measurement of static object (i.e., cone, cylinder, and sphere) in fluidized bed. Following we applied this approach to direct measure 3D gas–solid distribution in a bubbling fluidized bed, and found that satisfactory image correlation coefficients and solid concentration average absolute deviation could be obtained, which indicates the proposed approach is promising for 3D fluidized bed measurements.

Conventional testing of novel contrast agents for magnetic resonance imaging (MRI) involves cell and animal studies. However, 2D cultures lack dynamic flow and in vivo MRI is limited by regulatory approval of long‐term anesthesia use. Microfluidic tumor models (MTMs) offer a cost‐effective, reproducible, and high throughput platform for bridging cell and animal models. Yet, MRI of microfluidic devices is challenging, due to small fluid volumes generating low sensitivity. For the first time, an MRI of MTMs was performed at low field strength (1 T) using conventional imaging equipment without microcoils. To enable longitudinal MRI, we developed (1) CHAMP‐3 (controlled heating apparatus for microfluidics and portability) which heats MTMs during MRI scans and (2) an MRI‐compatible temperature monitoring system. CHAMP‐3 maintained chip surface temperature at ~37°C and the media inside at ~35.5°C. Enhanced T1‐weighted MRI contrast was achieved in 3D MTMs with free manganese (Mn²⁺) solutions and Mn²⁺ labeled tumor cells.

Vapor–liquid phase equilibrium (VLE) plays a crucial role in chemical process design, process equipment control, and experimental process simulation. However, experimental acquisition of VLE data is a challenging and complex task. As an alternative to experimentation, VLE data prediction offers great convenience and utility. In this article, an artificial intelligence network is proposed to predict the temperature and the vapor phase composition of binary mixtures. We constructed a graph neural network (GNN) and designed an uncertainty‐aware learning and inference mechanism (UALF) in the prediction process. The model was tested on both a self‐constructed dataset and a publicly available dataset. The results demonstrate that the proposed method effectively reveals the phase equilibrium properties of the target data. This work presents a novel approach for predicting vapor–liquid phase equilibrium in binary systems and proposes innovative ideas for investigating phase equilibrium mechanisms and principles.