Capacity-loss mechanisms in cycled zinc ferrite–silica hot-gas desulfurization sorbents: capacity retention beyond the 20-cycle window

The zinc ferrite–silica composite sorbent developed for high-temperature coal-gas desulfurization loses sulfur-uptake capacity over repeated sulfidation–regeneration–reduction cycles. A 1997 study of two $\mathrm{ZnFe_2O_4}$–$\mathrm{SiO_2}$/$\mathrm{TiO_2}$ sorbents prepared by urea and ammonia precipitation attributed the rapid durability loss of the ammonia-precipitated sorbent to sintering of its roughly 10 nm zinc ferrite primary particles, with a possible additional unconfirmed contribution from zinc or iron silicate formation at the silica interface, and showed that residual sulfur — left after oxidative regeneration and incompletely removed in the subsequent reduction step, consistent with sulfate formation — gradually increased over its 20-cycle test window while both sorbents maintained over 75% of their initial sulfur capacity. A 2002 sequel by the same group extrapolated the 20-cycle data with a sulfidation model, predicting that zinc sulfur capacity stabilizes near 40% of its initial value within about 50 cycles, and stated that verifying this stabilization requires multiple-cycle testing over more than 50 cycles. Later literature identified sulfate residues ($\mathrm{ZnSO_4}$ after regeneration, $\mathrm{Fe_2(SO_4)_3}$ in sulfided material) in related $\mathrm{ZnFe_2O_4}$–$\mathrm{SiO_2}$ and iron-oxide sorbents, and confirmed zinc silicate formation in a different Ni/$\mathrm{ZnO}$-on-silica system, but silicate formation in the cycled composite has not been directly phase-confirmed, the sulfate formation-and-removal mechanism under the source regeneration–reduction protocol has not been quantified, and the required extended-cycle verification has not been reported. The problem is to establish the best-available estimate of the capacity evolution beyond the 20-cycle window and the uncertainty and epistemic limits of that estimate, since the multiple-cycle testing over more than 50 cycles that the sequel study stated is necessary to confirm the stabilization has not been reported.

Background

In integrated gasification combined cycle (IGCC) power plants, coal is gasified to a fuel gas whose sulfur impurities (mainly hydrogen sulfide, $\mathrm{H_2S}$, with some carbonyl sulfide, $\mathrm{COS}$) poison downstream catalysts and corrode turbines, so the gas must be cleaned while hot and pressurized to preserve plant efficiency. Regenerable metal-oxide sorbents accomplish this: the oxide reacts with $\mathrm{H_2S}$ to form metal sulfides (sulfidation), and the sulfides are converted back to oxides in an oxidative regeneration step that releases the sulfur as $\mathrm{SO_2}$ for downstream recovery. Zinc ferrite ($\mathrm{ZnFe_2O_4}$, a mixed zinc–iron oxide with the spinel — a common cubic mixed-metal-oxide — crystal structure) is attractive because it combines the very low equilibrium $\mathrm{H_2S}$ concentrations of zinc oxide with the fast reaction of iron oxide; below about 723 K (450 °C) it is thermodynamically capable of sub-ppm sulfur removal. Before sulfidation the ferrite is reduced by the fuel gas ($\mathrm{H_2}$, $\mathrm{CO}$) to a $\mathrm{ZnO}$/iron-oxide mixture, so each operating cycle consists of reduction, sulfidation, oxidative regeneration, and another reduction; the practical sorbent is a fixed bed of pellets, and its useful life per cycle is read from the breakthrough curve — the outlet sulfur concentration versus time, whose integration gives the sulfur absorbing capacity.

The sorbent system studied here is a Central Research Institute of Electric Power Industry (CRIEPI, Japan) design: a composite powder of zinc ferrite and silica (Zn:Fe:Si molar ratio 1:2:2) supported on $\mathrm{TiO_2}$ (30:70 by weight). Silica is added to suppress carbon deposition from CO-rich gas, and $\mathrm{TiO_2}$ provides pellet strength. Two precipitation routes give profoundly different microstructures: homogeneous precipitation with urea yields zinc ferrite crystallites around 100 nm surrounded by 10–20 nm amorphous silica particles, whereas coprecipitation with ammonia yields roughly 10 nm zinc ferrite primary particles intimately mixed with similarly sized silica. In fixed-bed cyclic tests at 723 K and 0.98 MPa, both sorbents lost 22–25% of their sulfur absorbing capacity over 20 cycles: the loss in the first few cycles was quantitatively matched by residual sulfur remaining in the sorbent after regeneration and the following reduction step (measured by total-sulfur analysis, incinerating the solid in oxygen and detecting evolved $\mathrm{SO_2}$), while the additional loss by cycle 20 required both residual sulfur and physical sintering — the thermally driven growth and coalescence of particles that destroys the 3–30 nm mesopores and reduces reactive surface. Bulk Zn and Fe contents stayed constant, so metal volatilization was excluded. Sintering hit the ammonia-precipitated sorbent hardest, but even after its zinc ferrite particles had grown to sizes comparable to the urea-precipitated sorbent's, its spent-state sulfidation rate remained inferior, and the authors attributed the deficit to probable chemical interaction between silica and zinc or iron — i.e., formation of zinc or iron silicates (such as $\mathrm{Zn_2SiO_4}$), which have no desulfurization reactivity — while noting they had no direct phase-analytical confirmation of this. The residual sulfur itself is understood as sulfate (e.g., $\mathrm{ZnSO_4}$ or $\mathrm{Fe_2(SO_4)_3}$) formed when the oxidative regeneration step over-oxidizes sulfides, only part of which is decomposed or reduced away in the next reduction step; U.S. DOE work a decade earlier had shown explicitly that sulfate formation during zinc ferrite regeneration is strongly temperature-dependent, devising a two-stage regeneration that deliberately forms sulfates at about 454 °C and decomposes them at about 760 °C.

A companion CRIEPI study using in situ X-ray diffraction (XRD) and Mössbauer spectroscopy (a nuclear-spectroscopy technique sensitive to iron valence and local environments) established the legitimate phase-analysis tools for this system, identifying the reduction products (zincite, i.e., $\mathrm{ZnO}$, and zinc-dislocated franklinite — franklinite being the mineral name of zinc ferrite) and sulfidation products ($\mathrm{ZnS}$ and $\mathrm{FeS}$). A 2002 modeling study by the same group then fitted the measured cycle-by-cycle change of the zinc and iron sulfur capacities during the first 20 cycles and extrapolated that the zinc sulfur capacity may level off at about 40% of its initial value within about 50 cycles, which would hold the breakthrough time at roughly 70% of the fresh sorbent over more than 500 cycles — while stating that verifying this stabilization requires multiple-cycle testing over more than 50 cycles. That verification remains outstanding: no reported study has cycled this sorbent family beyond the 20-cycle calibration window, and the mechanistic questions it raises — phase-level confirmation (or exclusion) of silicate formation and quantitative treatment of sulfate accumulation across the regeneration–reduction sequence — frame the interpretation of the long-term capacity trajectory.

Problem Statement

Establish the best-available estimate of the long-term capacity retention of the silica-composited zinc ferrite desulfurization sorbent ($\mathrm{ZnFe_2O_4}$–$\mathrm{SiO_2}$ composite powder, Zn:Fe:Si = 1:2:2, supported on $\mathrm{TiO_2}$) under repeated sulfidation–regeneration–reduction cycling at process-relevant conditions (about 723 K, about 1 MPa, coal-gas-like atmosphere). The 2002 modeling study by the source group fitted a sulfidation reaction model to the measured cycle-by-cycle change of the zinc and iron sulfur capacities during the first 20 cycles and extrapolated that the zinc sulfur capacity may level off at about 40% of its initial value within about 50 cycles, holding the breakthrough time (the operating time before outlet sulfur exceeds the process limit) at roughly 70% of that of the fresh sorbent over more than 500 repetitions of the multiple desulfurization cycles. The same study stated that testing this stabilization effect requires multiple-cycle testing over more than 50 cycles; no such test has been reported, and no measured trajectory beyond the 20-cycle calibration window exists for this sorbent system. Because the extrapolation cannot be checked against measured extended-cycle data, the problem is to establish the best-available estimate of that capacity evolution — the capacity-versus-cycle trajectory over more than 50 cycles, grounded in the measured 20-cycle data — together with its uncertainty and epistemic limits.

The mechanistic questions developed in the Background — how much of the observed capacity loss is attributable to sintering of the zinc ferrite primary particles versus possible zinc or iron silicate formation at the silica interface, and how residual sulfate formed during oxidative regeneration accumulates across cycles and is partially removed in the reduction step — frame the interpretation of the measured capacity trajectory, but the question is the best-available extended-cycle capacity evolution of this sorbent system under a faithful cycle protocol: a sulfur-capacity-versus-cycle trajectory over more than 50 cycles, grounded in the measured 20-cycle data, together with its uncertainty and epistemic limits and with how far the extrapolated stabilization near 40% of the initial zinc sulfur capacity is supported on that basis. The system under study is the zinc ferrite–silica composite powder ($\mathrm{ZnFe_2O_4}$–$\mathrm{SiO_2}$, Zn:Fe:Si = 1:2:2, supported on $\mathrm{TiO_2}$), together with the honeycomb-form successor of the same composition (the same zinc ferrite–silica formulation extruded as a honeycomb monolith) and other $\mathrm{ZnFe_2O_4}$–$\mathrm{SiO_2}$ composites whose formulation, cycle protocol (including the reduction step), and conditions are faithful to the source system.

References

1. Makoto Kobayashi, Hiromi Shirai, Makoto Nunokawa, 'Investigation on Desulfurization Performance and Pore Structure of Sorbents Containing Zinc Ferrite', Energy & Fuels 11 (1997) 887–896, DOI 10.1021/ef960231k, https://doi.org/10.1021/ef960231k
2. Makoto Kobayashi, Hiromi Shirai, Makoto Nunokawa, 'Estimation of Multiple-Cycle Desulfurization Performance for Extremely Low-Concentration Sulfur Removal with Sorbent Containing Zinc Ferrite−Silicon Dioxide Composite Powder', Energy & Fuels 16 (2002) 1378–1386, DOI 10.1021/ef020024m, https://doi.org/10.1021/ef020024m
3. Makoto Kobayashi, Hiromi Shirai, Makoto Nunokawa, 'High-Temperature Sulfidation Behavior of Reduced Zinc Ferrite in Simulated Coal Gas Revealed by in Situ X-ray Diffraction Analysis and Mössbauer Spectroscopy', Energy & Fuels 16 (2002) 601–607, DOI 10.1021/ef010178m, https://doi.org/10.1021/ef010178m
4. Lu Li, Hongbo Zhang, Pin Zhou, Xianglong Meng, Lizhong Liu, Jinping Jia, Tonghua Sun, 'Three dimensional ordered macroporous zinc ferrite composited silica sorbents with promotional desulfurization and regeneration activity at mid-high temperature', Applied Surface Science 470 (2019) 177–186, DOI 10.1016/j.apsusc.2018.11.033, https://doi.org/10.1016/j.apsusc.2018.11.033
5. Makoto Kobayashi, Hiroyuki Akiho, 'Performance evaluation of honeycomb shaped sorbent for sulfur removal in advanced oxy-fuel IGCC power generation plant', Fuel 203 (2017) 1015–1025, DOI 10.1016/j.fuel.2017.02.087, https://doi.org/10.1016/j.fuel.2017.02.087
6. Makoto Kobayashi, Hiroyuki Akiho, Yasushi Ozawa, Akira Nakajima, 'Evaluation of dry acid gas removal process on bench scale test facility coupled with syngas produced by O2/CO2-blown gasifier', Fuel 250 (2019) 144–153, DOI 10.1016/j.fuel.2019.03.136, https://doi.org/10.1016/j.fuel.2019.03.136
7. Bo Peng, Kang Zou, Ye Song, Mudi Xin, Xue Yang, Lijun Lu, Jun Liu, Wei Lin, 'Development of Ni/ZnO desulfurization adsorbent with high stability: Formation of Zn2SiO4 and the impact from substrate', Chemical Engineering Journal 409 (2021) 127374, DOI 10.1016/j.cej.2020.127374, https://doi.org/10.1016/j.cej.2020.127374
8. Yu Feng, John Wen, Yongfeng Hu, Bin Wu, Mengmeng Wu, Jie Mi, 'Evaluation of the cycling performance of a sorbent for H2S removal and simulation of desulfurization-regeneration processes', Chemical Engineering Journal 326 (2017) 1255–1265, DOI 10.1016/j.cej.2017.05.098, https://doi.org/10.1016/j.cej.2017.05.098
9. Mayank Behl, Junghoon Yeom, Quentin Lineberry, Prashant K. Jain, Mark A. Shannon, 'A regenerable oxide-based H2S adsorbent with nanofibrous morphology', Nature Nanotechnology 7 (2012) 810–815, DOI 10.1038/nnano.2012.194, https://doi.org/10.1038/nnano.2012.194
10. M. C. Jha, M. H. Berggren, 'Two-stage regeneration of zinc ferrite desulfurization sorbent', U.S. DOE Morgantown Energy Technology Center Technical Report, June 1988, DOI 10.2172/5066440, https://doi.org/10.2172/5066440
11. Makoto Kobayashi, 'Development of sorbents used in dry syngas purification', chapter in 'Dry Syngas Purification Processes for Coal Gasification Systems', Elsevier (2021), ISBN 9780128188668, DOI 10.1016/B978-0-12-818866-8.00004-5, https://doi.org/10.1016/B978-0-12-818866-8.00004-5
12. Dicka Ar Rahim, Wei Fang, Haryo Wibowo, Yuswan Muharam, Herri Susanto, Kunio Yoshikawa, Yingjie Zhong, Mi Yan, 'Review of high temperature H2S removal from syngas: perspectives on downstream process integration', Chemical Engineering and Processing - Process Intensification 183 (2023) 109258, DOI 10.1016/j.cep.2022.109258, https://doi.org/10.1016/j.cep.2022.109258