Alejandro Montesinos-Castellanos

This study experimentally characterizes the spatial thermal response and resulting mechanical properties during air-assisted spray quenching of a full-scale 23 kg A356 aluminum cylinder head. Because internal thermal gradients are non-negligible at the component scale, subsurface temperature measurements were corrected to estimate the corresponding surface thermal response for local heat transfer coefficient (HTC) determination. Four spray conditions were evaluated by combining two spray heights (350 and 500 mm) and two nozzle configurations (3 and 5 nozzles) at constant total water flow. Reducing the spray height from 500 to 350 mm approximately doubled the spatial mean cooling severity over the 500–200 °C interval, increasing mean HTC by 106–108% and mean cooling rate by 103–109%, whereas changing from five to three nozzles produced smaller increases of approximately 11–15% in these spatial mean thermal descriptors. Local analysis at the tensile-specimen locations showed a nonlinear, concave-down relationship between cooling rate and yield strength, indicating diminishing strength gains as local cooling severity increased. At 350 mm, the three- and five-nozzle conditions exhibited similar mean local cooling rates at these locations (4.41 and 4.13 °C s⁻¹, respectively), but markedly different local cooling-rate ranges (2.17–7.86 and 3.47–4.82 °C s⁻¹, respectively). The more spatially consistent five-nozzle condition achieved a mean yield strength of 242.8 MPa, comparable to the 242.4 MPa obtained with the water-immersion reference (REF-IMM). These results indicate that component-scale spray-quench design should consider both sufficient local cooling severity and its spatial distribution, with spray height providing strong control of cooling intensity and nozzle configuration providing an additional means of modifying spatial cooling behavior in complex aluminum castings.

The accurate detection of carbon dioxide (CO2) is essential for industrial safety and environmental monitoring; however, conventional metal-oxide sensors are hindered by high power consumption and thermal instability. To address these limitations, this work introduces a Joule-heating-assisted strategy to fabricate a fully suspended carbon nanotube/carbon fiber (CNT/CF) hybrid sensor architecture. In this design, a single glassy carbon fiber serves simultaneously as the synthesis platform, conductive channel, and self-heating element. Driving electrical current directly through the suspended fiber generates localized temperatures exceeding 1000°C within seconds, enabling in situ iron nanoparticle (Fe NP) activation and hierarchical CNT growth with minimal thermal loss to the substrate. Deposition temperature was identified as the critical parameter governing nanoparticle evolution: lower temperatures (740°C) favor fine catalyst dispersion, whereas elevated temperatures (1240°C) promote substantial particle growth and Fe–C hybridization. The resulting hybrid material exhibits an electrical conductivity of approximately 476 S.cm-1. When deployed as a single suspended fiber CO2 sensor, the architecture demonstrates repeatable performance with a near-linear baseline drift (9.6 Ω/°C) across an operating window of 160 – 470°C. The sensor achieves a normalized sensitivity of 1.56%/vol% at elevated temperatures with rapid response kinetics (t90 < 1.2 s), while requiring less than 0.4 mW at peak operating conditions. This performance represents a two- to three-order-of-magnitude power reduction compared to conventional commercial sensors. This suspended hybrid architecture provides a thermally robust, low-power platform for high-concentration CO2 monitoring in industrial safety and flue-gas analysis.

The drive toward Process Intensification 4.0 demands a fundamental shift in how chemical processes are designed and optimized. We discuss the quantum computing potential across key areas of Process Intensification such as accelerating the discovery of novel materials and solving logistics problems. To illustrate this potential concretely, we present three unit-commitment case studies as a proof-of-concept formulation study, showing how power-planning problems can be expressed as QUBO/Ising Hamiltonians and solved with the Quantum Approximate Optimization Algorithm (QAOA) on a quantum simulator. The unit-commitment problem is an NP-hard problem. We draw a clear two-tier distinction throughout: near-term \emph{Noisy Intermediate-Scale Quantum} (NISQ) hardware supports shallow-circuit hybrid algorithms such as QAOA and VQE that are executable today but offer no guaranteed exponential speedup; the prospect of exponential computational advantage for certain problem classes is conditional on \emph{Fault-Tolerant Quantum Computing} (FTQC), which requires error-corrected logical qubits and remains a medium-to-long-term research goal.

This work deals with the problem of developing reliable data-driven real-time dynamic models orientedtowards chemical processes. To this end, we propose a novel optimization-based training formulation for liquid neural network models whose purpose is to track the dynamic transition behaviour of a mildly nonlinear flash tank system. Another research contribution refers to the training algorithm conventionally done by the back-propagation through time approach. For the data-driven training of small- or medium-scale systems, the underlying dynamic optimization problem is discretized and solved deploying the full transcription approachcombined with a multi-trajectory training scheme. For the dynamic flash tank case study, the purely data-driven liquid neural network model with 2 hidden states successfully tracks unseen dynamic transitions with low errorachieving a root mean square error below 1.2 $^\circ$C even with modest training data.Overall, both strategies, the transcription-based training and multi-trajectory scheme, are deployed to build a digital twin version of the dynamic flash tank system whose purpose is to support simulation, optimization, control and fault-detection tasks in real-time environments.

Biomass gasification has the potential to be a reliable technology for renewable hydrogen production once its technoeconomic efficiency is increased. Dolomite is a material commonly used to improve gasification performance after calcination (CaMgO₂), which depends on its composition (primarily calcium and magnesium) and the interactions among its components. However, its effectiveness is highly variable due to compositional differences based on the extraction site. Therefore, the composition and properties of dolomite can vary significantly between samples, complicating our understanding of their effect on biomass gasification. This research aims to elucidate the impact of the proportions of calcium and magnesium oxide (CaO and MgO) in dolomite and mechanical mixtures, along with the addition of potassium carbonate (K₂CO₃), on hydrogen production through biomass steam gasification in a fluidized bed reactor. Remarkably, the mechanical mixtures showed a similar performance to dolomite. Even the highest H₂ yield (441 mL/g-of-biomass-feed) was achieved with a mechanical mixture of CaO, MgO, and K₂CO₃, which was 11% higher than the yield from the dolomite–K₂CO₃ mixture and a 4-fold increase compared to no additives. These findings suggested a synergistic interaction between calcium and magnesium in mechanical mixtures that was previously attributed solely to dolomite. Additionally, CaO could be related to the decreased concentration of acidic tar, while K₂CO₃ diminished basic tar compounds. This kind of study is crucial for the development of cost-effective catalysts using these abundant and inexpensive materials, advancing biomass gasification as a truly viable and sustainable method for large-scale H₂ production.

Ultrasound-assisted synthesis of MOF-235(Fe) and PVP accelerates its phase transition to MIL-53(Fe).

This work presents a methodology integrating Non-Linear Programming (NLP) for multi-objective and multi-period optimization, addressing sustainable waste management and energy conversion challenges. It integrates waste-to-energy (WtE) technologies such as Anaerobic Digestion (AD), Incineration (Inc), Gasification (Gsf), and Pyrolysis (Py), and considers thermochemical, technical, economic, and environmental considerations through rigorous non-linear functions. Using Mexico City as a case study, the model develops waste management strategies that balance environmental and economic aims, considering social impacts. A trade-off solution is proposed to address the conflict between objectives. The economical optimal solution generates 1.79M$ with 954 tons of CO₂ emissions while the environmental one generates 0.91M$ and reduces emissions by 54%, where 40% is due to gasification technology. Moreover, the environmentally optimal solution, with incineration and gasification generates 9500 MWh/day and 5960 MWh/day, respectively, demonstrates the capacity of the model to support sustainable energy strategies. Finally, this work presents an adaptable framework for sustainable waste management decision-making.