BUSCADOR RECOLECTA

Digital.CSIC. Repositorio Institucional del CSIC

oai:digital.csic.es:10261/357787

2. Experimental runs An exemplary for an entire isothermal experiment is given in Fig. 2. It shows the initial conditioning, consisting of ten redox cycles, followed by the actual experiment, consisting of seven isothermal plateaus. These isothermal plateaus each consist of four redox cylces employing the following H2 volume fractions: 50, 50, 30, 16.66 vol%. The 50 vol% run was repeated to possibly detect inconsistencies introduced through the preceding temperature change. Between reduction and oxidation or when changing from one temperature to another, the system was purged with nitrogen. 3. Reproducibility In order to reach the kinetic regime small sample masses (< 2 mg) and high gas flow rates (> 300 ml · min−1) were required. The reproducibility of the measurements was assessed at these conditions. All isothermal runs were repeated six times, all on different days using fresh solid samples and the same TGA programs. Fig. 3 shows the conversion rate dX/dt. for the individual runs during reduction at T = 900 ◦C. Even though, there were noticeable deviations between individual sets, the overall reproducibility of results was convenient. 4. Gas conversion The hydrogen conversion XH2 during the TGA experiments was estimated to ensure sufficient gas supply and negligible impact of steam produced. Low gas conversion signifies that enough gas was present and the atmosphere did not change significantly during the experiments. The mass balance ˙V | in · ci{nz· yH2,i}n n˙ H2,in − (m0 − m∞) MO · (dX/dt)max = ˙V| out · cou{tz· yH2,ou}t n˙ H2,out (1) was used calculate the molar flow n˙ H2,out of unreacted H2 leaving the reactor at maximum reactivity. This gives the highest possible gas conversion for a specific case, i.e. the worst case scenario for reaching the kinetic regime. The n˙ H2,out was used to calculate the conversion XH2 = n˙ H2,in − n˙ H2,out n˙ H2 in at different temperatures and H2 contents. The results are summarized in Table 1. As can be seen, the hydrogen conversion XH2 was low (< 2 %) for all cases, which means that neither gas starvation nor limitations due to steam generation should have played a significant role. 5. Isoconversional methods The additional plots for the isoconversional methods (i.e. evaluation of linear regression and R2) which were referenced in our main work are given in this section. Figure 4 shows the differential isoconversional method for one exemplary experiment with higher mass at 50 vol%H2. Except for the first point (X=0.1) a reasonably high R2 was again achieved. Figure 5 depicts the analysis of the nonisothermal reductions with the differential isoconversional method. Figure 6 shows the results for the integral isoconversional KAS method [1]. For both methods high R2 values were achieved.-- Under a Creative Commons license CC-BY 4.0 Deed., 1. TGA setup.-- 2. Experimental runs.-- 3. Reproducibility.-- 4. Gas conversion.-- 5. Isoconversional methods.-- 6. CFD study.-- 1. TGA setup: A schematic of the experimental setup is given in Fig. 1. The TGA experiments were performed with a NETZSCH STA 449 F3 Jupiter. It consisted of an inner reactor and an outer shell, both made of Al2O3. Different sample holders (crucible, plate, basket) could be mounted on top of a vertical sample carrier at the center of the inner reactor. If not stated differently, the experiments were conducted with the plate (NETZSCH alumina slipon plate, diameter 17 mm). The solid sample was placed on this sample holder. A type S thermocouple within the vertical sample carrier was used for temperature measurements. It had been calibrated with calcium oxalate and pure metal melting points (In, Sn, Bi, Zn, Al, Ag) prior to the experiments. Two different gas flows were relevant in the TGA setup. Firstly, a protective gas flow (20 ml · min−1 N2), which was always switched on, entered the reactor below the sample carrier. Secondly, the main gas flow (H2, O2, N2 and mixtures thereof) entered a water vapor generator, where a predefined mass flow of H2O could be added. The gas mixture was preheated to 180 ◦C and then entered the TGA system from the side. It was heated to the desired temperature during the upwards flow in the outer shell. After reaching the top of the reactor, the gases flowed downwards into the inner reaction chamber, which contained the solid sample, before leaving the reactor via the gas outlet. Both, the main and protective gas flows, were controlled by Bronkhorst mass flow controllers (MFC, 0-250 ml · min−1). 6. CFD study A small, transient CFD study of the TGA was conducted to gauge the temporal evolution of the hydrogen concentration above the solid sample. The TGA geometry of Fig. 1 was simplified (mostly at the inlet and the outlet). Simulations were performed in ANSYS Fluent 2023 R1 using the lamiar flow model and a fixed hydrogen diffusivity of 5·10−4 m2 · s−1. Heat transfer was considered via conduction, convection and radiation (Discrete Ordinate model). A homogeneous velocity distribution was assumed at the inlet (T = 200 ◦C, yH2 = 0.5). The heating was simulated via isothermal walls (T = 800 ◦C) which quickly heated the gases to the operating temperature. At the outlet, a simple pressure outlet condition was chosen. The discretization methods for momentum, mass and heat transfer were set to second order (only Discrete Ordinate was first order), and a first order implicit time solver (time step of 0.1 s) was chosen. Fig. 7 (a) shows the simplified TGA model, Fig. 7 (b) shows the velocity profile in steadystate and Fig. 7 (c) shows the H2 concentration profile at t = 7 s for one exemplary case. As can be seen, the flow around the plate leads to nearly zero velocity at the sample, which could possibly lead to diffusion limitations. More importantly, Fig. 7 (c) suggests an appreciable spatial gradient of hydrogen due to axial dispersion. The temporal evolution of the hydrogen concentration above the plate was given in our main work., Peer reviewed