Steam Regeneration of Polyethylenimine-Impregnated Silica Sorbent for Postcombustion CO₂ Capture: A Multicyclic Study

Abstract

Steam regeneration of polyethylenimine (PEI)-impregnated commercial grade silica was investigated in a packed bed reactor. Adsorption was performed at 75 °C under 10% CO2/N2, and desorption was carried out under steam at 110 °C for 20 consecutive cycles. CO2 adsorption capacity was found to decrease by 9 mol % over the period of 20 cycles. No evident signs of sorbent degradation due to PEI leaching or changes in surface morphology and amine functionalities were observed upon characterization of the sorbent after the cyclic study. Most of the loss in adsorption capacity was associated with thermal degradation of the sorbent during drying under N2 after steam stripping at 110 °C. The desorption kinetics during steam stripping was found to be much faster than during N2 stripping. Over 80% of the total CO2 was released within the first 3 min of steam injection into the reactor. A separate packed bed study was conducted to investigate the influence of moisture content (5.3–14.7 vol %) in flue gas on the CO2 adsorption capacity of PEI-impregnated silica. The presence of moisture had a positive impact on CO2 uptake of the sorbent; a 4–9 mol % increase in CO2 uptake was observed in comparison to the adsorption under dry conditions. However, the presence of moisture increased the heat of regeneration of the sorbent significantly. It was calculated that the energy demand increased approximately 2-fold on introduction of 14.7% moisture compared to that of dry flue gas.

Insights into the Hydrothermal Stability of Triamine‐Functionalized SBA‐15 Silica for CO2 Adsorption

“…Following two 3h steam strippingc ycles, the wet-grafted samples showedu ptake losses of 11-27 %, 5-15 %, and À7-8 %, for adsorption at 25, 50, and 75 8C, respectively (Figure 2a). Given the outstanding performance of TRI-SBA-15-0.2-48Sa nd the much shorter duration needed in actual cycling, [18,19] the current adsorbentsc ould withstand am uch highern umber of cycles. The sample exposed to 48 ho fs team regeneration (TRI-SBA-15-0.2-48S) also preserved much of its CO 2 adsorptionc apacity,s howingu ptake losses of 23, 12, and 8% relative to the correspondingf resh grafteds ample (TRI-SBA-15-0.2) for adsorption temperatures of 25, 50, and 75 8C, respectively.T hree noteworthy comments can be made here:( i) contrary to the fresh samples (TRI-SBA-15-x), in whicht he adsorption was thermodynamically controlled (i.e.,h ighest uptakeo ccurring at 25 8C), adsorption on wet-grafted samples after steam stripping seemed to be diffusion-limited as adsorption capacities at 50 8Cw ere similar to those at 25 8C, (ii)consistent with the occurrence of diffusion resistance upon exposure to steam, the CO 2 uptake was least affected for adsorptiont emperatures of 50 and 75 8C, as opposed to 25 8C, and (iii)following the initial loss of adsorption capacity after the first 3h steam treatmentc ycle, the materials remained stable.…”

“…Therefore, thermald egradation shouldb ed ismissed as ap otentialc ontributor to the observed behavior. [19,22] The BET surfacea reas were measured for the fresh TRI-SBA-15-0 and all the wet-grafted samples before and after steam regeneration (Figure 7). [22] FTIR measurementsi nt he range 1400-3500 cm À1 werec arried out on TRI-SBA-15-0.2 ( Figure 6) before the first and after the second steam stripping cycles.…”

“…[9a, 17] However,s uch at echnique would not be applicable in ar eal-worlds cenario because, eventually,C O 2 would again be diluted into the inert gas. [19] This strategy is particularly interesting because low-temperature steam is typicallya vailable in power plants as waste heat and may be used for CO 2 desorption. Such atechnique is advantageous as it is relativelyeasy to separate high purity CO 2 through steam condensation, following the desorption step.…”

“…[21] The authors attributed the poor performance of the impregnatedS BA-15 following exposure to steam to structural collapse or restructuring of the support and amine leaching. [22] Sandhu et al [19] explored the stabilityo fP EI-impregnated commercial grade Q-10 silica (Fuji Silysia ChemicalL td.) Although not ex-plicitly mentioned, the much smaller pore size of the SBA-15 support (6.8 nm) relative to the g-alumina support (17.3 nm) [21] is consistentw ith the vast differencei nt he performance of the adsorbentsb ecause of enhanced diffusion resistance.…”

“…As no significant change in amine loading was observed following exposure to steam, the authors suggested that the amine might have leachedo ut of the pores, coveringt he external surface of the support. [19] Provided that the support is hydrothermally stable, aminegraftedm aterials are expected to show good steam stability as ar esult of covalentb onding between the amine and the support, minimizing potential amine leaching. The larger differencei nt he performance of the adsorbents during direct air capture is also in line with mass transfer limitations at low CO 2 partial pressure.…”