Hexamethyldisiloxane adsorption prediction in an Ordered Mesoporous Carbon

Abstract Driven by the continuous search to substitute fossil fuels, biomethane has gained significant relevance. However, its production requires biogas upgrading to remove siloxane compounds, notably hexamethyldisiloxane (L2). Adsorption is a cost-effective separation method for this application, with ordered mesoporous carbons (OMCs) proving to be highly effective adsorbents. This study proposes a methodology to predict L2 adsorption on a CMK-3 OMC at 298 K via molecular simulation, validating the results against experimental data. An OMC model was characterized using representative pore kernels (diameters of 20, 25, 30, 38, 45, and 55 Å) to simulate N₂ isotherms at 77 K and derive pore size distributions (PSDs). While the full six-pore kernel yielded the highest fit to experimental data (R² = 0.989), a reduced five-pore kernel (up to 45 Å pore) was defined as the optimal minimum kernel, achieving comparable accuracy (R² = 0.983) with one fewer pore. Local L2 isotherms at 298 K were subsequently simulated for these pores utilizing two literature-adapted force fields. Combining these local isotherms with the N₂ PSDs enabled the prediction of total L2 isotherms, where the highest predictive accuracy (R² = 0.88) was achieved using the kernel restricted to 45 Å. Finally, simulations of conventional materials, granular activated carbon (GAC) and activated carbon fiber (ACF), successfully reproduced the experimentally observed adsorption capacity hierarchy (OMC > GAC > ACF).