Activated carbon adsorption of heteroatom components from pyrolysis oil for improved chemical recycling
Thien Nguyen Luu Minh,Dave Manhaeghe,Gwendoline Bernaert,Joël Hogie,Michel Clarembeau,Kevin M. Van Geem,Steven De Meester
DOI: https://doi.org/10.1016/j.cej.2023.148394
IF: 15.1
2024-02-01
Chemical Engineering Journal
Abstract:Pyrolysis is a promising technique for managing mixed plastic waste. However, its industrial implementation faces significant challenges due to the presence of heteroatom components in the pyrolysis oil. These impurities are the major hurdle to use these feeds in steam crackers. To address this issue, we evaluated the potential of adsorption on activated carbon as a purification method by measuring the adsorption capacity of nine of the most important heteroatom components in pyrolysis oils. Through single-isotherm experiments and modeling, our aim was to understand the interactions between the adsorbate and the adsorbent as well as the influence of the component-pyrolysis oil interaction. Furthermore, multicomponent isotherm experiments further explored cooperative and competitive interactions among different heteroatom components, providing valuable insights into their behavior in complex systems. Our findings highlight benzoic acid as the most effective adsorbate (qm = 193.4 ± 12.8 mg g−1), while less polar components such as cyclopentanone (qm = 15.9 ± 1.79 mg g−1), hexamethylcyclotrisiloxane (qm = 7.11 ± 1.27 mg g−1), and 1-chloropentane (no detectable removal) pose challenges for pyrolysis oil purification without excessive dosing of the adsorbent, in our case activated carbon. Moreover, multicomponent experiments revealed that pyridine exhibits cooperative adsorption behavior, whereas other components engage in competitive adsorption dynamics. The analysis of six multicomponent models emphasized the importance of considering adsorbate–adsorbate interactions in describing the adsorption behavior in complex mixtures. Models equipped with interaction parameters (correlative model) outperformed those without such parameters (predictive model), as evidenced by higher R2 values (R2 = 0.95–0.99) for the former compared to the latter (R2 = 0.84–0.86).
engineering, chemical, environmental