(453h) Study of COS Removal from Gases on Doped or Undoped Aluminas

Gaseous sulfur-containing molecules are considered as harmful contaminants from several points of view in most industrial refining or energy production processes. Their toxicity, strong ability to cause catalyst poisoning and corrosion are the main negative issues induced by these molecules.

The need for light hydrocarbon feedstocks (C1-C4) ultrapurification is presently becoming more and more acute in order to increase the lifetime of catalysts in refineries or petrochemical plants and to achieve the increasingly stringent sulfur content specifications.. Besides the removal of H₂S, which is the main sulfur contaminant, this supposes that carbonyl sulfide (COS), the second most abundant sulfur contaminant in light hydrocarbons, has to be also removed. COS is mainly formed during the thermal or catalytic conversion process of natural gas, oil or biomass, either by reaction of H₂S with CO₂, or CH₄ with SO₂. In refineries, COS is most likely to be found in LPG (liquefied petroleum gases) streams such as C3 cuts (propane-propylene) or C3-C4 cuts (such as propylene/butenes streams issued from FCC); Moreover, in certain cases, COS removal has to be totally removed form process streams as its presence can be very detrimental for downstream processes. For instance, COS is a poison for propylene polymerization catalysts.

Existing desulfurization processes (guard bed sorbents, solvents-based Acid-Gas Removal processes) are often dedicated to the removal of H₂S which generally represents the major sulfur impurity. But yet, COS being less reactive than H₂S, these processes are not or little effective for the removal of COS. For example, ZnO (H₂S capture mass) does not react with COS (Aboulayt et al. 1996), COS removal kinetics using chemical solvents such as amines are not favorable (about 100 times lower than CO₂) (Vaidya et Kenig 2009), and COS removal using physical solvents is effective at very low temperatures (around -40°C for washing with methanol-based solvents) which would make it very costly process. Thus, the preferred industrial COS removal processes are based on the catalytic hydrolysis (COS + H₂O + CO₂ â?? H₂S + CO₂) and/or the adsorption of COS (P. K. T. Liu 1988).

In both cases, the adsorbent or catalyst generally consists of a basic solid (Bulushev et Ross 2011; A. Aboulayt et al. 1996). The most used materials are simple or mixed metal oxides eventually doped by alkali or alkaline earth metals, by transition metals or even by rare-earth elements. According to the literature, adsorption of COS on metal oxides occurs through interactions with the surface hydroxyl groups (OH), and is more favored at low temperatures (T < 100°C) and without water. Hydrolysis of COS is thermodynamically favored but its kinetic is quite slow (Rhodes et al. 2000). Thus it is rather favored at high temperatures (T > 100°C) and requires the presence of water (H. M. Huang et al. 2006). Industrially, the choice between adsorption or hydrolysis based method thus depends on the operating conditions imposed in practice. The interaction of COS with basic oxides, such as alumina, has been studied in the literature for widely varying operating conditions. Nevertheless, a number of points are not described in the literature and published data are scarce, especially regarding interaction mechanisms of COS with the oxide surfaces and adsorption competitions with species also involved in the hydrolysis reaction (either CO₂ or H₂S).

The present work has been focused on the study of the adsorption of COS on doped or undoped oxides (Al₂O₃, TiO₂ and ZrO₂). For all of them, we showed that COS is partially hydrolyzed when adsorbed, but only CO₂ is adsorbed on the same sites than COS and may thus compete with it for its adsorption. The formation of CO₂ is less pronounced for more basic oxides that suggests a role of Lewis acid sites. In parallel, the regeneration of the adsorbent and the stability of their capacities have also been investigated. It appears than the most basic oxides show stronger interactions with COS. COS is not fully desorbed during regeneration, explaining a decrease of adsorption capacities over cycling. Moreover, in the case of less basic oxides, some elemental sulfur is observed, whose formation has been assigned to the presence of Lewis acid sites. However, this phenomenon does not seem to be responsible for the decrease of COS adsorption capacity.

In brief, this study gives us a better comprehension of the overall system showing the specific role of Lewis acid sites and also the benefic effect of increasing the basicity of the adsorbent.

Bibliography

A. Aboulayt; F. Mauge; P. E. Hoggan; J. C. Lavalley (1996) Combined FTIR, reactivity and quantum chemistry investigation of COS hydrolysis at metal oxide surfaces used to compare hydroxyl group basicity. In : Catalysis Letters, vol. 39, n° 3-4, p. 213â??218.

Aboulayt, A.; Mauge, F.; Hoggan, P. E.; Lavalley, J. C. (1996) Combined FTIR, reactivity and quantum chemistry investigation of COS hydrolysis at metal oxide surfaces used to compare hydroxyl group basicity. In : Catalysis Letters, vol. 39, n° 3-4, p. 213â??218.

Bulushev, D. A.; Ross, J. R. H. (2011) Catalysis for conversion of biomass to fuels via pyrolysis and gasification: A review. In : Recent Developments in Operando Spectroscopy, vol. 171, n° 1, p. 1â??13.

H. M. Huang; N. Young; B. P. Williams; S. H. Taylor; G. J. Hutchings (2006) High temperature COS hydrolysis catalysed by gamma-Al2O3. In : Catalysis Letters, vol. 110, n° 3-4, p. 243â??246.

P. K. T. Liu (1988) Adsorption of carbonyl sulfide from liquid hydrocarbons with activated alumina and other adsorbents. In : AIChE Symposium Series, vol. 84, n° 264, p. 109â??118.

Rhodes, C.; Riddel, S. A.; West, J.; Willams, B. P.; Hutchings, G. J. (2000) The low-temperature hydrolysis of carbonyl sulfide and carbon disulfide: a review. In : Recent Developments in Operando Spectroscopy, vol. 59, n° 3-4, p. 443â??464.

Vaidya, P. D.; Kenig, E. Y. (2009) Kinetics of carbonyl sulfide reaction with alkanolamines: A review. In : Chemical Engineering Journal, vol. 148, n° 2-3, p. 207â??211.