Date of Award
Open Access Dissertation
During the past decades, heterogenous catalyzed conversion of biomass to hydrocarbons with similar or identical properties to conventional fossil fuels has gained significantly academic and industrial interest. However, the conventional heterogeneous catalysts such as sulfided NiMo/Al2O3 and CoMo/Al2O3 used have various drawbacks, such as short catalyst lifetime and high sulfur content of product. To overcome the limitations of the conventional sulfided catalysts, new catalysts must be developed, which requires a better understanding of the reaction mechanism of the biomass conversion. Based on density functional theory, in this thesis, we reported a computational calculation study of the reaction mechanism of the hydrodeoxygenation of propionic acid (our choice of model biomass molecules). For most of the times, however, biomass conversion is a very complicated process which usually have a very large reaction network, and thousands of intermediates and reaction steps are involved. Therefore, to theoretically identify the energy of those species by computational calculations is not realistic. We therefore also reported the potential of a combination of computational calculations and machine learning method to address the large reaction network of biomass conversion.
Based on first principles calculations, a full microkinetic model have been developed for the vapor and liquid phase hydrodeoxygenation of propionic acid over a Pt (111) surface. Calculations suggest that decarboxylation does not occur at an appreciable rate. In the vapor phase, decarbonylation products, propionaldehyde and propanol are all produced at similar rates. However, in both liquid water and 1,4-dioxane, propanol and propionaldehyde are favored over decarbonylation products. While a condensed phase can shift the reaction rate and selectivity significantly, the dominant pathways towards the various products are hardly affected. Only for propionaldehyde production do we observe a shift in mechanism. A similar study was also conducted on the vapor and liquid phase hydrodeoxygenation of propionic acid on Rh (111) surface. Calculations suggest that both decarboxylation and decarbonylation do not occur at an appreciable rate in all reaction environments. Propanol and propionaldehyde are the main products and produced at similar rates in both vapor and liquid phases. Although a condensed phase can shift the reaction rate, the dominant pathways and selectivity towards the various products are hardly affected.
In a combination of density functional theory calculations and machine learning method, we proposed a retraining cycle to predict the reaction mechanism of the hydrodeoxygenation of propionic acid on transition metal surfaces and the catalyst activity. With proper metal descriptors and species descriptors being used, our model predicts almost the same rate controlling species and reaction rates as that from models based on DFT calculations. We conclude that the approach we proposed can be readily used to address the complicated biomass conversion chemistry at a DFT accuracy without the need to do full DFT calculations for the large reaction network involved.
Yang, W.(2020). Theoretical Investigation of the Biomass Conversion on Transition Metal Surfaces Based on Density Functional Theory Calculations and Machine Learning. (Doctoral dissertation). Retrieved from https://scholarcommons.sc.edu/etd/6073