Experimental investigation of a packed bed membrane reactor for the direct conversion of CO2 to dimethyl ether

In this study, the performance of a packed bed membrane reactor (PBMR) based on carbon molecular sieve membranes for the one-step CO2 conversion to dimethyl ether (DME) is experimentally compared to that of a conventional packed bed reactor (PBR) using a CuO-ZnO-Al2O3/HZSM-5 bifunctional catalyst. The PBMR outperforms the PBR in most of the experimental conditions. The benefits were greater at lower GHSV (i.e., conditions that approach thermodynamic equilibrium and water formation is more severe), with both XCO2 and YDME improvements of +35–40 % and +16–27 %, respectively. Larger sweep gas-to-feed (SW) ratios increase the extent of water removal (ca. 80 % at SW=5), and thus the performance of the PBMR. Nevertheless, alongside the removal of water, a considerably amount of all products are removed as well, leading to a greater improvement in the CO yield (+122 %) than the DME yield (+66 %). Higher temperatures selectively improve the rWGS reaction, leading to a lower YDME with respect to the PBR at 260 °C, due to the significant loss of methanol. Furthermore, larger transmembrane pressures (∆P) were not beneficial for the performance of the PBMR due to the excess reactant loss (i.e., 98–99 % at ∆P = 3 bar). Finally, the reactor models developed in our previous studies accurately describe the performance of both the PBR and PBMR in the range of tested conditions. This result is of high relevance, since the reactor models could be used for further optimization studies and to simulate conditions which were not explored experimentally. •A packed bed membrane reactor tested based on carbon molecular sieve membranes.•CuO-ZnO-Al2O3/HZSM-5 bifunctional catalyst result in high DME yields.•PBMR outperforms the PBR in XCO2 and YDME improvements of +35–40 % and +16–27 %.•Higher temperatures selectively improve the rWGS reaction, leading to a lower YDME.