https://doi.org/10.1016/j.ccst.2021.100011
“A typical untreated flue gas also contains several impurities including 100 ppm HCl, 800 ppm SO2, 10 ppm SO3 and 500 ppm NOx. Tolerance to these acid impurities is critically important for potassium-based adsorbents. Unfortunately, CO2 adsorption capacities of dry potassium-based adsorbents had been reported to be negatively affected by the presence of acid impurities, since they were easily poisoned by H2S, HCl, SOx and NOx.
Kim et al. reported the deactivation of K2CO3/Al2O3 in post-combustion flue gas containing SOx in a pilot-scale plant. The deactivation pathways were disclosed as two aspects (Kim et al., 2012). On the one hand, K2CO3 would react with trace SO2 (10-40 ppm) in flue gas irreversibly to form stable K2SO3. On the other hand, the carbonation product KHCO3 would also interact with SO2 in regeneration tower. The formed byproduct K2SO3 could be further oxidized into stable K2SO4 with excessive oxygen (Kim et al., 2012). Chae et al. confirmed the formation of stable K2SO4 via the sulphuration K2CO3 in a low temperature and humidity environment. K2CO3/Al2O3 with 30 wt.% K2CO3 loading was capable of capturing 310 mg SO2/g adsorbent under the conditions of 50°C, 5000 ppm SO2 and 11%H2O (Chae et al., 2016). The increased SO2 sorption capacity with increasing RH should indicate that K2CO3/Al2O3 might suffer more severe deactivation in the presence of SO2 under a higher RH. Kim et al. elucidated that the active component in K2CO3/Al2O3 would react with 175-700 ppm HCl at 200-500°C and 1-20 bar (Kim et al., 2016). K2CO3 supported on Al2O3 nanowire had been employed to capture trace HCl from industrial coal gas. Thermodynamic calculation results showed that the Gibbs free energy of the reaction between K2CO3 and HCl kept at -190∼-200 kJ/mol in the temperature range of 100-300°C. This indicated that K2CO3 would preferentially capture HCl in the ternary CO2-HCl-H2S mixtures. In other words, K2CO3/Al2O3 would suffer significant deactivation when used for capturing CO2 from flue gas containing trace HCl (Liang et al., 2020). Jin’s group had also confirmed the deactivation of potassium-based adsorbents in flue gas containing SO2 and HCl. The K2CO3/AC adsorbent showed a slow failure rate within the initial 10 cycles, while the failure rate increased remarkably to 21.8% after 23 cycles. The failure mechanisms were attributed to the reaction between K2CO3 and trace impurities of SO2 and HCl to form stable byproducts (Fan et al., 2018).
Wu et al. attributed the deactivation of K2CO3/Al2O3 to the fact that H2O accelerated the sulphuration reaction of K2CO3 to generate K4H2(CO3)3•1.5H2O and K2SO3. Whereas, the presence of NO exerted no obvious effect on the carbonation reactivity (Wu et al., 2013; Wu et al., 2011a; Wu et al., 2011b). Guo et al. illustrated that both SO2 and NO2 adversely affected the carbonation performance of K2CO3/AC. The present SO2 and NO2 prevailed over CO2 in the dynamic competition for the surface adsorption sites, as they reacted with both the active K2CO3 and the K4H2(CO3)3•1.5H2O intermediate to form byproducts of K2SO3 and KNO3 in a humid environment (Guo et al., 2016). The formed K2SO3 and KNO3 were stable and could not be regenerated, thus reducing the K2CO3 utilization efficiency. Beyond that, the formed byproducts retained on the surface and accumulated in the pores in repeated cycles, and they further covered the surface active sites and blocked pore structures (Guo et al., 2016).
Effective strategies had been proposed to alleviate the deactivation of dry potassium-based adsorbents in the presence of acid impurities. One feasible approach involved doping potassium-based adsorbents with alkaline constituents such as Ca(OH)2, CaO, CaCO3, KOH and PEI. These species would provide considerable alkaline and amine groups that showed stronger affinity towards SO2 and NO2 than CO2. The doped alkaline species possessed a double functionality, namely, to inhibit adsorbent deactivation via preferentially trapping SO2 and NO2 in the dynamic competitive sorption process and to promote the carbonation process by offering more alkaline active sites (Fig. 8) (Guo et al., 2017; Wu et al., 2014). Wu et al. reported that Ca(OH)2 worked satisfactorily in inhibiting the deactivation of K2CO3/Al2O3 as well as improving its CO2 sorption capacity. This had been evidenced by the increase of CO2 sorption capacity from 1.39 to 2.05 mmol CO2/g with the increasing Ca(OH)2 loading from 0 to 20 wt.% (Wu et al., 2014). As had been noted, K4H2(CO3)3•1.5H2O was prone to suffer sulphuration in the carbonation process. In contrast, the hydrated K2CO3•1.5H2O intermediate was stable and exhibited better resistance towards SO2 and NO2 (Wu et al., 2011a). Another suggested strategy involved water vapor pretreatment and steam regeneration to convert K2CO3 or KHCO3 to K2CO3•1.5H2O to improve the reactivity and stability of potassium-based adsorbents in impurities (Fig. 8) (Guo et al., 2018b; Wu and Chen, 2015). Guo et al. reported that K2CO3/AC pretreated with water vapor retained stable CO2 sorption capacity in trace SO2 and NO2, with about 4.62% reduction in its CO2 uptake after 10 repeated cycles (Guo et al., 2017). Potassium-based adsorbents had been reported to be capable of removing multiple components in flue gas stream. The simultaneous adsorption of various components including SO2, NO and CO2 on K2CO3/Al2O3 had been illustrated by Yi et al.. Under the simulated flue gas condition (50-200°C, 82 kPa, 12%CO2, 0-10%O2, 0-2%H2O, 1000 ppm NO and 2000 ppm), NO was firstly oxidized to NO2 and the ternary acid gases were co-adsorbed on the alkaline surface by physical adsorption and chemisorption (Yi et al., 2014). More efficient measures for inhibiting sorbent deactivation in multiple impurities are still critically necessary to stabilize the CO2 adsorption performance of potassium-based adsorbents in cyclic operations.”
“Fig. 8. Water vapor pretreatment, KOH addition and PEI modification proposed as efficient strategies for inhibiting deactivation of potassium-based adsorbent in trace SO2 and NO2. Reproduced with permission from (Guo et al., 2017). Copyright 2016 Elsevier B. V.”