We employed deep neural networks (NNs) as an efficient and intelligent surrogate of molecular simulations for complex sorption equilibria using probabilistic modeling. Canonical (N1N2VT) Gibbs ensemble Monte Carlo simulations were performed to model a single-stage equilibrium desorptive drying process for (1,4-butanediol or 1,5-pentanediol)/water and 1,5-pentanediol/ethanol from all-silica MFI zeolite and 1,5-pentanediol/water from all-silica LTA zeolite. A multi-task deep NN was trained on the simulation data to predict equilibrium loadings as a function of thermodynamic state variables. The NN accurately reproduces simulation results and is able to obtain a continuous isotherm function. Its predictions can be therefore utilized to facilitate optimization of desorption conditions, which requires a laborious iterative search if undertaken by simulation alone. Furthermore, it learns information about the binary sorption equilibria as hidden layer representations. This allows for application of transfer learning with limited data by fine-tuning a pretrained NN for a different alkanediol/solvent/zeolite system.
3 October 2019 in Journal of Chemical & Engineering Data
Anaerobic fermentation can transform carbohydrates to yield a multicomponent mixture comprising mainly of acetone, 1-butanol, and ethanol (ABE) in a typical weight ratio of 3:6:1. Compared to ethanol, 1-butanol, the main product of ABE fermentation, offers significant advantages as a biofuel or a fuel additive. However, the toxicity of 1-butanol for cell cultures requires broth concentrations to be low in 1-butanol (≈1–2 wt %). An energy-efficient recovery method that performs well even at low 1-butanol concentrations is therefore necessary to ensure economic feasibility of the ABE fermentation process. In this work, configurational-bias Monte Carlo simulations in the Gibbs ensemble are performed to probe the adsorption of 1-butanol/water solutions onto all-siliceous zeolites with the framework types MFI and FER. At low solution concentration, the selectivity and capacity for 1-butanol in MFI are larger than those in FER, while the opposite is true for concentrations at or above those of ABE broths. Structural analysis at various loadings sheds light on the different sorbate–sorbate and sorbate–sorbent interactions that govern trends in adsorption in each zeolite.