Optimization of Cosolvent Enhanced Lignocellulosic Fractionation for Isolating Switchgrass Lignin with Distinct Structural Features Using Response Surface Methodology

Pretreatment and fractionation technologies have been used to separate and isolate biomass polymers for conversion into fuels, chemicals, and other products. A great deal of work has focused on dialing in reaction conditions (e.g., time, temperature, acid concentration, etc.) that are amenable to isolating an uncondensed lignin product that could be converted into high value aromatic platform molecules. Pretreatment severity emerged as a term that combines time, temperature, and acid concentration into a single value that can be used to compare various pretreatment technologies. However, combining the effects of these conditions into a single term, while convenient, confounds the effects that these conditions have on lignin quality, both individually and when combined with each other. Moreover, pretreatment and fractionation reactors do not have a severity “knob,” and several different sets of conditions could mathematically achieve the same severity but have different effects on the resulting lignin product slate. In this study, we set out to model the effects of time (10-30 min), temperature (140-180 °C), and acid concentration (0.025-0.1 M H2SO4) on lignin yield (up to quantitative), molecular weight (Mw = 700-2000 g/mol), and hydroxyl group content (3.55-6.06 mmol OH/g) using the co-solvent enhanced lignocellulosic fractionation (CELF) process on switchgrass. Our results show that lignin yield is most sensitive to acid concentration, with an additional 4.96% yield per 10 mM of acid. In addition, molecular weight is sensitive to acid concentration and temperature, with a decrease of 77.9 g/mol per 10 mM of acid and a decrease of 19.3 g/mol per °C. Moreover, total hydroxyl group content decreases at a rate of 89 μmol total OH per g lignin per min at short time (t = 12 min, T = 160 °C) and is increases at a rate of 125 μmol total OH per g lignin per min at long time (t = 28 min, T = 160 °C). Finally, our results demonstrate that residence time does not have a statistically significant effect on yield or molecular weight within the studied ranges, which could have implications for continuous and flow-through processes, where short residence times could lead to substantial cost savings. Overall, these results demonstrate that practitioners can design a process that maximizes one or more of industrially relevant lignin properties by exerting careful control of fractionation conditions, which could ultimately lead to greater utilization of lignin for fuels, chemicals, and other products.