Mechanistic rate laws, sulfidation product phases, and unexplained kinetic anomalies for Cu-modified cerium oxide H₂S sorbents

Kobayashi and Flytzani-Stephanopoulos (Ind. Eng. Chem. Res. 2002) showed that copper additions (5–40 atom %) strongly promote the reducibility and H₂S-sulfidation kinetics of bulk cerium oxide over 623–923 K, and explicitly left open several issues: an unexplained reduction anomaly of the 15 atom % Cu sample; sulfidation product speciation (Ce₂O₂S versus Ce₂S₃) inferred only from weight changes; rate laws resting on single-temperature reaction orders ($n=0.78$ for H₂, $n=0.38$ for H₂S, both at 823 K); and confounding of sulfidation kinetics by incomplete prereduction. A further item — an unexplained slow approach to final sulfur uptake at 550 °C (823 K) — is noted in the source work but is retained here only as a background lead, not as a required component of the problem. Later work has substantially resolved the speciation question on adjacent systems — Ce₂O₂S is the established cerium product under reducing H₂S/H₂, and a 500–700 °C product map exists for supported CuO–CeO₂ — and has partially established the prereduction–sulfidation link on ceria alone. The surviving open problem is a mechanistically validated, product-phase-resolved kinetic description of the bulk Cu–CeOₓ system itself: rate laws as functions of temperature, H₂/H₂S partial pressures, and prereduction state; phase-specific identification of the sulfidation products of the bulk composite; quantified prereduction effects on sulfidation kinetics; and an explanation of the 15 atom % reduction anomaly.

Background

Hot-gas desulfurization — the removal of hydrogen sulfide (H₂S) from coal- or biomass-derived fuel gas (syngas) at process temperature — is required in integrated gasification combined cycle (IGCC) power plants and in fuel-cell power systems, because H₂S corrodes turbines and poisons catalysts and fuel-cell anodes. The removal medium is a solid metal-oxide sorbent that reacts with or adsorbs H₂S; a practical sorbent needs favorable sulfidation equilibria (how completely it converts to sulfide), fast kinetics, high sulfur capacity, structural stability, and regenerability over a wide temperature window. The incumbent oxides have well-known weaknesses: ZnO loses structural integrity because Zn is reduced and vaporizes above roughly 673 K, and CuO is readily reduced to metallic Cu, whose sulfidation equilibrium constant is far smaller than that of CuO or Cu₂O.

Ceria (CeO₂, a fluorite-structure oxide) is attractive because cerium shuttles between the +4 and +3 oxidation states: partially reduced ceria (CeO₂₋ₓ, ultimately Ce₂O₃, containing Ce(III) sites and oxygen vacancies) has much more favorable sulfidation equilibria than fully oxidized CeO₂, and the oxide can in principle be regenerated by oxidation. The obstacle is that bulk ceria reduces only at high temperature, so generating Ce(III) sites at accessible temperatures became the central design problem. Prior work in the authors' laboratory had shown that composites of CuO with CeO₂ take up more sulfur than copper alone can account for, suggesting a ceria contribution — but quantitative kinetics were lacking.

Kobayashi and Flytzani-Stephanopoulos (2002) supplied that kinetics for bulk Cu-modified ceria (Cu–CeOₓ) sorbents containing 5–40 atom % Cu, prepared by urea coprecipitation/gelation and tested at 0.1 MPa over 623–923 K. The methods are standard gas–solid kinetics tools: temperature-programmed reduction (TPR — heating in dilute H₂ while monitoring weight loss, to characterize reducibility), thermogravimetric analysis (TGA — continuous weighing to follow reaction rates), in situ X-ray diffraction (XRD — identifying crystal phases during reaction), and Brunauer–Emmett–Teller (BET) nitrogen physisorption for surface area. They found that CuO in the composites reduces to metallic Cu below 453 K; as little as 5 atom % Cu induces low-temperature surface reduction of ceria near 423 K; and prereduced Cu–CeOₓ sulfidizes much faster than pure ceria in H₂S–H₂–N₂ mixtures. Initial rates were fitted as power laws, $R_{r,0}=k_r C_{H_2}^{n}$ with $n=0.78$ for reduction and $R_{s,0}=k_s C_{H_2S}^{n}$ with $n=0.38$ for sulfidation (both orders measured only at 823 K), giving apparent activation energies of about 14.2 and 21.9 kJ/mol for reduction (low-Cu and 40 atom % samples) and 16 ± 3 kJ/mol for sulfidation of the 5 and 15 atom % samples versus roughly 34 kJ/mol for the 40 atom % sample. The fast sulfidation was attributed to Ce(III) sites whose density copper promotes.

The same paper acknowledged unresolved issues that define this problem. First, the 15 atom % Cu sample reduced less than the 40 atom % sample at every temperature although its texture (BET surface area 32.27 versus 19.97 m²/g; ceria crystallite size 8.8 versus 9.6 nm) should favor reduction — an acknowledged anomaly attributed speculatively to preparation or dispersion peculiarities. Second, the sulfidation product speciation was inferred from TGA weight changes alone, computed under an assumed Ce₂S₃ stoichiometry; mass change cannot by itself distinguish cerium oxysulfide (Ce₂O₂S — the phase in which oxide anions are partly replaced by sulfide, stable for cerium under reducing sulfur-bearing gas) from Ce₂S₃ or mixed oxy-sulfides, because the candidate stoichiometries imply different weight gains. Third, the reaction orders were determined at a single temperature and assumed temperature-independent, so the reported activation energies conflate concentration and temperature dependence, and no mechanistic rate expressions exist. Fourth, prereduction of the 40 atom % sample is incomplete below about 823 K, so its measured low-temperature sulfidation rates are depressed — and the apparent activation energy derived from them inflated — by unachieved Ce(III) site density rather than by intrinsic kinetics. In addition, the source work reports a fifth issue — a reproducibly slow approach to final sulfur uptake at 550 °C (823 K), left unexplained, with transport limitations, phase-change kinetics, and competing adsorption/desorption among the candidate explanations — retained here only as a background lead, not as a required component of the problem below.

Later work has redrawn the boundaries of these questions without closing them. The speciation question has been substantially answered for ceria and adjacent composites: the Ce₂O₂S–CeOₓ phase boundary was determined by Ferrizz, Gorte and Vohs (2003); thermodynamic analyses (Lohsoontorn et al. 2008; Svoboda et al. 2017) and numerous identifications by XRD and by X-ray photoelectron spectroscopy (XPS, which measures surface elemental composition and oxidation states) establish Ce₂O₂S as the cerium product of ceria sulfidation under reducing H₂S/H₂, with Ce₂S₃ appearing only under more severe conditions; and Liu, Zhou and Wu (2016) mapped the products of a CuO–CeO₂ composite supported on ZSM-5 zeolite (Ce₂O₂S plus copper sulfides at 600 °C; Ce₂S₃ appearing at 700 °C). The same group's follow-ups (Wang and Flytzani-Stephanopoulos 2005 on regenerative reformate desulfurization; the 2006 Science paper on regenerative H₂S adsorption on rare-earth oxides) established that Cu additions sinter ceria severely above 650 °C and that fast regenerative surface adsorption and slower bulk sulfidation are distinct regimes. None of this determined the rate laws for the bulk Cu–CeOₓ system, mapped its products against prereduction state, or revisited the two anomalies.

Problem Statement

Starting from the empirical kinetic description of Kobayashi and Flytzani-Stephanopoulos (2002) for bulk Cu-modified cerium oxide H₂S sorbents (urea coprecipitation/gelation materials with 5–40 atom % Cu, tested at 0.1 MPa in H₂S–H₂–N₂ mixtures over 623–923 K), establish a mechanistically validated, product-phase-resolved quantitative description of the reduction and sulfidation of this sorbent class. This requires:

(i) determining the rate laws — reaction orders and mechanistic rate expressions — for ceria reduction by H₂ and for sulfidation by H₂S as functions of temperature (623–923 K), H₂ and H₂S partial pressures, and the prereduction state of the sorbent, replacing the single-temperature orders ($n=0.78$ for H₂ and $n=0.38$ for H₂S, both measured at 823 K) that underlie the reported apparent activation energies;

(ii) identifying the sulfur-containing solid products of sulfidation of the bulk Cu–CeOₓ system as a function of Cu content (5–40 atom %), the H₂S/H₂ ratio, and the prereduction state over 623–923 K, distinguishing cerium oxysulfide (Ce₂O₂S), cerium sesquisulfide (Ce₂S₃), and copper sulfide phases and resolving their phase fractions — the source work inferred speciation only from TGA weight changes under an assumed Ce₂S₃ stoichiometry;

(iii) quantifying how the prereduction state achieved before sulfidation (extents of CuO-to-Cu and ceria reduction) affects subsequent sulfidation rates and apparent activation energies, since incomplete prereduction of high-Cu-load samples below about 823 K (notably at 623–723 K) artificially depresses measured low-temperature sulfidation rates and inflates apparent activation energies (e.g., the roughly 34 kJ/mol reported for the 40 atom % sample);

(iv) explaining a reproducible but unexplained observation from the source work: the lower extent of ceria reduction of the 15 atom % Cu–CeOₓ sample compared with the 40 atom % sample despite its more favorable texture (BET surface area 32.27 versus 19.97 m²/g; ceria crystallites 8.8 versus 9.6 nm).

Items (i)–(iv) are components of one objective: a kinetic and phase-behavior model of the same gas–solid reaction system, whose predictive reliability is currently limited by single-temperature empirical orders, unverified product stoichiometry, uncontrolled prereduction state, and the unexplained 15 atom % reduction anomaly. Later literature has already fixed the general phase identity of ceria sulfidation products on adjacent systems (Ce₂O₂S dominant under moderately reducing H₂S/H₂; Ce₂S₃ under more severe conditions), so the speciation component is targeted at the bulk Cu–CeOₓ system itself.

References

1. Makoto Kobayashi and Maria Flytzani-Stephanopoulos (2002), 'Reduction and Sulfidation Kinetics of Cerium Oxide and Cu-Modified Cerium Oxide', Industrial & Engineering Chemistry Research 41, 3115–3123. DOI: 10.1021/ie010815w. https://doi.org/10.1021/ie010815w
2. R. M. Ferrizz, R. J. Gorte, and J. M. Vohs (2003), 'Determining the Ce2O2S–CeOx phase boundary for conditions relevant to adsorption and catalysis', Applied Catalysis B: Environmental 43, 273–280. DOI: 10.1016/S0926-3373(02)00323-5. https://doi.org/10.1016/S0926-3373(02)00323-5
3. Zheng Wang and Maria Flytzani-Stephanopoulos (2005), 'Cerium Oxide-Based Sorbents for Regenerative Hot Reformate Gas Desulfurization', Energy & Fuels 19, 2089–2097. DOI: 10.1021/ef049664a. https://doi.org/10.1021/ef049664a
4. Maria Flytzani-Stephanopoulos, Mann Sakbodin, and Zheng Wang (2006), 'Regenerative Adsorption and Removal of H2S from Hot Fuel Gas Streams by Rare Earth Oxides', Science 312, 1508–1510. DOI: 10.1126/science.1125684. https://doi.org/10.1126/science.1125684
5. Bing Liu, Hengyong Xu, and Zehui Zhang (2012), 'Temperature-Programmed Surface Reaction Study of Adsorption and Reaction of H2S on Ceria', Chinese Journal of Catalysis 33, 1631–1635. DOI: 10.1016/S1872-2067(11)60416-6. https://doi.org/10.1016/S1872-2067(11)60416-6
6. Singfoong Cheah, Yves O. Parent, Whitney S. Jablonski, Todd Vinzant, and Jessica L. Olstad (2012), 'Manganese and ceria sorbents for high temperature sulfur removal from biomass-derived syngas – The impact of steam on capacity and sorption mode', Fuel 97, 612–620. DOI: 10.1016/j.fuel.2012.03.007. https://doi.org/10.1016/j.fuel.2012.03.007
7. Dongjing Liu, Weiguo Zhou, and Jiang Wu (2016), 'CuO-CeO2/ZSM-5 composites for reactive adsorption of hydrogen sulphide at high temperature', The Canadian Journal of Chemical Engineering 94, 2276–2281. DOI: 10.1002/cjce.22613. https://doi.org/10.1002/cjce.22613
8. P. Lohsoontorn, D. J. L. Brett, and N. P. Brandon (2008), 'Thermodynamic predictions of the impact of fuel composition on the propensity of sulphur to interact with Ni and ceria-based anodes for solid oxide fuel cells', Journal of Power Sources 175, 60–67. DOI: 10.1016/j.jpowsour.2007.09.065. https://doi.org/10.1016/j.jpowsour.2007.09.065
9. Karel Svoboda, Jindřich Leitner, Jaromír Havlica, Miloslav Hartman, Michael Pohořelý, Jiří Brynda, Michal Šyc, Yau-Pin Chyou, and Po-Chuang Chen (2017), 'Thermodynamic aspects of gasification derived syngas desulfurization, removal of hydrogen halides and regeneration of spent sorbents based on La2O3/La2O2CO3 and cerium oxides', Fuel 197, 277–289. DOI: 10.1016/j.fuel.2016.12.035. https://doi.org/10.1016/j.fuel.2016.12.035
10. Lei Yang, Shang Guan Ju, Hui Qing Wang, Jia An, Jun Ming Yu, Jie Mi, and Mao Qian Miao (2017), 'The Effect of Regeneration Conditions on the Properties of CeO2 Desulfurization Sorbent in SO2 Atmosphere', Materials Science Forum 896, 155–161. DOI: 10.4028/www.scientific.net/msf.896.155. https://doi.org/10.4028/www.scientific.net/msf.896.155
11. Jiang Wu, Dongjing Liu, Weiguo Zhou, Qizhen Liu, and Yaji Huang (2018), 'Desulfurization Kinetics and Thermodynamics', chapter in 'High-Temperature H2S Removal from IGCC Coarse Gas' (Energy and Environment Research in China series), Springer, pp. 121–151. DOI: 10.1007/978-981-10-6817-1_5 (print 2018; online November 2017). https://doi.org/10.1007/978-981-10-6817-1_5
12. Dongjing Liu, Qian Wang, Jiang Wu, and Yangxian Liu (2019), 'A review of sorbents for high-temperature hydrogen sulfide removal from hot coal gas', Environmental Chemistry Letters 17, 259–276. DOI: 10.1007/s10311-018-0792-x (online September 2018). https://doi.org/10.1007/s10311-018-0792-x
13. Bo Guo, Sai Li, and Yu Tian (2019), 'Performance of Cerium Oxides from Different Preparation Methods for Desulfurizing Hot Coal Gas', ACS Omega 4, 9301–9305. DOI: 10.1021/acsomega.9b00247. https://doi.org/10.1021/acsomega.9b00247
14. Ramiar Sadegh-Vaziri and Matthäus U. Bäbler (2019), 'Removal of Hydrogen Sulfide with Metal Oxides in Packed Bed Reactors—A Review from a Modeling Perspective with Practical Implications', Applied Sciences 9, 5316. DOI: 10.3390/app9245316. https://doi.org/10.3390/app9245316
15. Qiang Liu, Bingsi Liu, Qinze Liu, Qiuyue Zhang, Rongnian Xu, Xuanyue Wu, and Hong Xia (2020), 'Walnut wood-derived hierarchically 3D self-assembly of (a%Ce–Mn)yAl2−yOx and rapid diffusion character of straight micron channel during hot coal gas desulfurization', Chemical Engineering Journal 393, 124761. DOI: 10.1016/j.cej.2020.124761. https://doi.org/10.1016/j.cej.2020.124761
16. 'Study on Removal of H2S with Cerium Oxide at High Temperature' (氧化铈脱硫剂的高温煤气脱硫研究; in Chinese), Coal Conversion (煤炭转化), 2008, no. 2. DOI: 10.3969/j.issn.1004-4248.2008.02.006. https://doi.org/10.3969/j.issn.1004-4248.2008.02.006