Development, experimental investigation and numerical simulation of a demonstration reactor system for the indirect solar-thermochemical reduction of redox particles

Kurzfassung

In light of the great urgency for climate change mitigation, an important measure is to transform the production of chemical energy carriers towards low‑emission technologies. Solar‑thermochemical two‑step redox cycles yield fuels by absorbing highly concentrated sunlight to drive solid‑gas reactions. Their prospects are accompanied by demanding process conditions and potential operational compromises: Decoupling the absorption of radiation from the chemical reaction is, due to their different corresponding time scales, deemed to enable further optimization. An indirect redox concept can accommodate this – it uses a Particle Mix Reactor (PMR) to perform thermal reduction of redox particles through heat provided by inert particles. Its principle is demonstrated and investigated in this work. To this, the perovskite redox material $\mathrm{SrFeO}_{3-\delta}$ was synthesized via a solid‑state route and found to show high redox activity. Using the mixing granulation and spheronization techniques, particles were processed in the size range of 1.5$\,$mm and characterized along with inert $\mathrm{Al_2O_3}$ particles. A substantial improvement in particle shape and size distribution was achieved, resulting in superior yield. An experimental setup including a vacuum batch reactor was designed for temperatures of up to 1100$\,$°C and 1$\,$kg of redox and inert particles each at sufficiently low heat losses (<$\,$35$\,$K). Particle motion and simultaneous heat transfer was simulated via the Discrete Element Method (DEM). An existing program considering thermal radiation was adapted for dissimilar particle types by modeling particle‑particle conductive heat transfer and chemical reaction at transient pressure. Particle contact parameters were determined in a dedicated calibration procedure. PMR experiments and simulations were performed on three temperature levels (up to 1100$\,$°C, 700$\,$°C and 400$\,$°C), evaluated for temperature, released amount of oxygen, and mixture homogeneity, as well as the effect of uncertainty in model input quantities. Pressure measurements proved an increasingly effective thermal reduction at higher temperature levels. Results for mixture homogeneity are qualitatively consistent between experiment and simulation, exhibiting gradients in both axial and circumferential directions. Temperature results reflect this, while their deviations cannot exclusively be explained that way: Simulated initial mean bed temperatures are overestimated by about 19$\,$K due to the neglected gas phase, sensor inertia, and uncertainty in valve opening delay. Throughout 30 minutes of the reaction phase, their deviation is less than 5$\,$% from experimental results, which is below the uncertainty propagated from input parameters. Within this margin, qualitative discrepancies are hypothesized to arise from conduction along thermocouple sheaths and heat losses by radiation. Released oxygen amounts of up to 0.328$\,$mol exhibit quantitative deviations of 25$\,$% due to high measurement uncertainty and model simplification. Time‑dependent temperature distribution in the particle bed indicates quick heat transfer on the particle scale. In conjunction with fast reaction kinetics, particle residence in the reactor can be limited to the order of a minute, which promotes the decoupling motive. The validated model facilitates further simulations at good accuracy to analyze future concepts using different redox materials, and may thus contribute to the advancement of solar fuel production technologies.