https://doi.org/10.1016/j.seppur.2022.120786
”
A polypropylene (PP) hollow-fibre membrane was investigated within the same conditions to establish if a tighter pore size and sharper geometrical pore shape factor (B, from 0.15 to 0.56) could eliminate wetting following shell-side crystallisation, through providing 700% increase in breakthrough pressure (Table 2). Similar CO2 flux profiles were initially observed for both membranes (Fig. 8). For the PTFE membrane, shell-side crystallisation proceeded once 2.36 M CO2 L−1 were absorbed, equivalent to a supersaturation index (C/C*) of 1.34 based on bicarbonate speciation (2.06 M HCO3− L−1; Fig. 9). For reference, shell-side crystallisation occurred soon after at a supersaturation index (C/C*) of between 1.34 and 1.48 for the PP membrane. Classical nucleation theory describes how nucleation rate is dependent upon the level of supersaturation [25]. Despite nucleation commencing at comparable levels of supersaturation, a higher nucleation rate was identified for the PTFE membrane, which recorded 118 # cm−3 (crystals/vol. of solvent) at C/C* of 1.48, compared to only 1.4 # cm−3 for the PP membrane at equivalent conditions (Fig. 10). This difference is not clearly explained by the interfacial energy of the two membrane substrates (ΔGhet/ΔGhom, ∼90% for PTFE and ∼ 60% for PP), where the PP polymer should foster greater reduction in activation energy to limit the supersaturation level required to initiate nucleation [13], [6]. While the role of surface roughness cannot be discounted, a critical difference between the membranes is the coarse pore structure of the PTFE membrane, which we suggest can induce higher ‘local fluxes’, thereby raising the local supersaturation profile in the boundary layer, which can correlate to an increase in both nucleation and crystal growth [13]. While wetting soon followed shell-side crystallisation in the PTFE membrane, a stable CO2 flux was observed in the PP membrane following primary nucleation, leading to the continuation of nucleation and growth (Fig. 10), which suggests that with the correct selection of pore size to mitigate wetting, solids formation can be promoted in membrane contactors to reduce the energy penalty for CO2 separation.
Fig. 8. Impact of membrane properties (PTFE, dmax 3.4 µm; PP, dmax 0.36 µm) on reactive membrane crystallisation. Absorbent HCO3− concentration and relative supersaturation (C/C*, based on HCO3− concentration; solubility, 1.53 mol NH3 L−1 at 5 °C and 1 atm) illustrated for PTFE membrane. Conditions: Recirculating ammonia absorbent, 3 M; G/L 5. Error bars indicate standard deviation. Red line indicates ammonium bicarbonate solubility at 5 °C.
Fig. 9. Dynamic equilibria following progressive CO2 absorption into 3 M aqueous ammonia in recirculation. Absorbent temperature 5 °C. Blue line indicates ammonium bicarbonate solubility at 5 °C.
Fig. 10. Impact of membrane properties (PTFE, 3.4 µm; PP, 0.36 µm) on reactive membrane crystallisation: (a) crystal number; (b) median crystal size versus crystal number. Conditions: 3 M aqueous ammonia in recirculation; absorbent temperature 5 °C. Ammonium bicarbonate peak corresponds to minima in HCO3− and CO32−, and is coincident with HCO3− concentration needed for induction (C/C* 1, 1.53 mol NH3 L−1 at 5 °C and 1 atm). Error bars indicate standard deviation obtained from sacrificial experiments carried out in triplicate.
Fig. A1. Distribution of carbonic acid, bicarbonate and carbonate (H2CO3, HCO3−, CO32−) (a) and of ammonium-ammonia (b) as functions of pH and solvent temperature.
Fig. B1. (a) Ammonium bicarbonate (NH4HCO3) solubility in water. Symbols correspond to the mean calculated between experimental data from PubChem, [33] and Perry and Green, [31]; (b) Saturated vapour pressure of NH3 over aqueous ammonia as a function of solution temperature. An increase in ammonia concentration or temperature increases the probability for ammonia slip.
“