(634e) Predicting the Morphological Properties of Activated Carbons Produced from Lignocellulosic Materials

Every year approximately 200 billion tons of biomass waste is produced. Identifying value added products that can be produced from waste biomass is important for both economic viability and sustainability metrics for the forest products, agricultural, and other biomass industries. Numerous researchers have explored producing activated carbon from various biomass sources ranging from palm fronds to leftover food. However, few studies have explored how key constituents of lignocellulosic biomass affect the properties of activated carbon. In this work, an augmented second-order simplex lattice mixture design was used to explore how the relative concentrations of cellulose nanofibers (CNFs), cellulose nanocrystals (CNCs), and alkali lignin (LIG) affected the properties of activated carbons produced from a two stage carbonization and activation process. The response variables included yield, specific surface area and micropore fraction. The specific surface area and micropore fraction were determined by BET analysis and the t-plot method. The mixture regression models revealed clear relationships between all three activated carbon characteristics and biomass precursor composition. The highest yield of 43% was obtained from pure lignin, the most thermally stable material. Pure CNF resulted in both the highest micropore fraction, ~80%, and the lowest specific surface area, 350 m²/g. The highest observed specific surface area of 1940 m²/g was obtained from a precursor composition of 17/17/66 % (by mass) CNF/CNC/LIG blend. The models resulting from the experimental design were validated by producing activated carbon from a 15/35/50 % (by mass) CNF/CNC/LIG blend. This composition was chosen by equally weighting the desirability of maximizing both surface area and micropore fraction. The models predicted that the activated carbon would have a specific surface area of 1320 m²/g and a micropore fraction of 60%. The experimentally measured values were 1240 ± 120 m²/g and 58 ± 1%. The percent differences amongst the predicted and experimental values were 6% and 3% for surface area and micropore fraction respectively. This result highlights that the mixture design was effective in enabling the development of a robust model that can be used to design activated carbon precursor mixtures to maximize the most desirable properties for a given application. Moreover, understanding the role of individual biomass components in activated carbon precursors will aide in determining what types of biomass waste will be most effective for producing activated carbons for applications requiring different properties such as adsorbents for environmental contaminants and electrochemical devices.