Monday, April 30, 2007

Scale-up studies for determining biomass reactivity during pretreatment

Noah D. Weiss, Nicholas J. Nagle, and Richard T. Elander. National Bioenergy Center, National Renewable Energy Laboratory, 1617 Cole Blvd., Golden, CO 80401

Achieving the U.S. Department of Energy goal to displace 30% of 2004 gasoline demand with biofuels by 2030 requires the ability to efficiently convert a wide array of biomass feedstocks.  These feedstocks exhibit a broad range of structural and chemical properties that affect the yield and efficiency of biomass conversion processes. The authors have previously developed a robust, higher-throughput method for screening the susceptibility of biomass feedstocks to saccharification using a two-stage process based on pretreatment followed by enzymatic hydrolysis. This screening methodology enables the susceptibility of biomass feedstocks to biochemical conversion to be assessed over a wide range of pretreatment chemistries and severities. The yields of glucose and xylose solubilized during the two-stage saccharification process can be combined to provide a measure of feedstock reactivity, which can be used to compare the relative reactivity of different feedstocks to biochemical conversion and as a function of pretreatment conditions.  While this method is useful for comparing reactivity between different feedstocks across a range of pretreatments, it is performed under laboratory conditions that do not fully mimic conditions anticipated for industrial processing. In particular, in the higher-throughput screening methodology pretreatments are performed at low solids concentrations (5%) with indirect heating, whereas scaled up pretreatments are expected to be performed at high solids concentrations (³30%) with direct heating.  We here report on experiments performed to assess how predictive feedstock reactivity results obtained in bench scale screening studies are of saccharification performance achieved when pretreatment is carried out in a small pilot scale reactor (Zipperclave) under more process-relevant conditions. Results are presented for bench and pilot scale experiments run with switchgrass feedstock using a similar range of pretreatment severities and acid loadings, and the impact of higher solids loading in the pilot scale system is discussed. Ultimately, understanding how the results of smaller-scale, higher-throughput screening methods can be extrapolated to predict performance in larger-scale systems will enable conditions appropriate for biochemical processing at an industrial scale to be more efficiently identified.