Our results confirm that AHP is an effective pretreatment of corn stover for subsequent enzymatic saccharification to monomeric Glc and Xyl, and that it can be further improved in performance and economics by increasing the biomass loading, reducing the H2O2 loading, extending the pretreatment time, and maintaining the pH.
The advantages of AHP over other existing effective pretreatments are several. First, it uses readily available and environmentally benign chemicals, namely, hydrogen peroxide and sodium hydroxide. Second, it operates at low temperature (21°C to 50°C) and atmospheric pressure, so no expensive specialized reactors are necessary. It can be performed in any standard scientific laboratory, which means that researchers interested in studying AHP do not have to rely on outside collaborators with specialized equipment. Third, it is less expensive than two other highly effective pretreatments, ionic liquids and phosphoric acid/organic solvent. Phosphoric acid/ethanol was reported to give a monomeric Glc yield of > 89% with a very low enzyme loading (0.66 mg cellulase/g glucan), but it requires a phosphoric acid:biomass loading of approximately 13:1 (w/w) [6, 7]. Ionic liquids gave 96% digestibility of switchgrass but use a reagent: biomass ratio of 32:1 . Practical implementation of these two effective pretreatments will require highly efficient recycling of the respective reagents.
The main potential disadvantages of AHP pretreatment are the cost of H2O2 and the presence of significant residual salt. At an H2O2 price of US$0.80 to $1.00/kg (on a 100% basis) (Dr Philip Block, FMC Corporation, personal communication), the current cost for processing 1 metric ton of biomass would be US$102 to US$125 at an H2O2 loading of 0.125 g/g biomass, plus US$36 for NaOH (at $300/ton). However, AHP can probably be improved to operate at even lower H2O2 loadings, for example, by H2O2 stabilization, improved pH control, or H2O2 recycling . The use of lower H2O2 loadings would also have the advantage of reducing the consumption of NaOH and acid. In addition, it might be possible to reduce the intrinsic cost of H2O2 by on-site production or by the development of new methods of production .
In these and earlier studies, we used an enzyme loading of 15 mg total protein/g glucan as our standard hydrolysis condition. Comparison of this enzyme loading to other studies in which enzyme loadings are expressed in filter paper units (FPU) is difficult because FPU is a measure of cellulase activity only and does not take into account the contribution of the many other enzymes in most cellulase preparations, such as BG and xylanases, to the release of monomeric Glc and Xyl from lignocellulose . However, a point of comparison can be made to Spezyme CP, a widely used commercial cellulase cocktail. It has been reported to contain between 0.3 and 0.48 FPU/mg protein [35, 39], and therefore our standard loading of 15 mg/g glucan corresponds to a Spezyme CP loading of between 4.5 and 7.2 FPU/g glucan.
In the current experiments, the biomass was not washed or otherwise separated into multiple processing streams after pretreatment. Many pretreatment protocols involve post-washing the biomass to remove inhibitors and salts. Depending on the type and severity, some pretreatments also remove some of the lignin and hemicelluloses [8, 19, 23, 24, 26–28]. However, washing uses a large amount of water (exactly how much is rarely indicated), and, depending on the severity of the pretreatment, removes a proportion of the cellulose and hemicellulose. This lowers the energy content of the biomass and, for laboratory-scale studies, necessitates reanalysis of the glucan and xylan content post pretreatment in order to calculate final sugar yields . In our experiments, all of the original material was still present regardless of the pretreatment severity. Since no cellulose or hemicellulose was lost, reanalysis of the biomass composition was unnecessary. The main disadvantage of not washing after AHP is that the final product contains salt from the NaOH and HCl. With AHP conditions of 2% biomass loading and 0.5 g H2O2/g biomass, the final NaCl concentration was approximately 200 mM (Additional file 1, Table S1). At the highest biomass loading (20%) and highest H2O2 loading, the NaCl concentration was approximately 1.5 M (Additional file 1, Table S1). This high salt concentration was not a problem in our experiments because all enzyme digestions were performed at 0.2% glucan loading, which lowered the salt concentration to approximately 20 mM, but it might become a problem if enzymatic hydrolysis (and fermentation) were performed at industrially relevant high-solids loading.
In addition to lowering the consumption of H2O2, reducing the H2O2 loading results in less residual salt in the pretreated material. For example, at 10% biomass loading and 0.125 g H2O2/g biomass, the NaCl concentration in the biomass after neutralization was approximately 350 mM (Additional file 1, Table S1). Saha and Cotta [21, 22] used AHP-pretreatment on barley and wheat straw at approximately 10% biomass loading and approximately 2.5% H2O2 and successfully digested and fermented the resulting material without post-washing. Residual salt did, however, inhibit butanol fermentation by Clostridum beijerinckii . Possible routes to alleviating the problem of salt from AHP include further decreasing the H2O2 loading, removing the salt after pretreatment, or performing the fermentation with salt-tolerant organisms. Saccharomyces cerevisiae strains able to tolerate NaCl concentrations of greater than 0.5 M NaCl have been developed under laboratory conditions [41–43].
Previous studies with AHP have shown that it solubilizes lignin but its effect on cellulose crystallinity is unclear [17, 19, 36]. One or both of these modifications could be contributing to the effectiveness of AHP as a pretreatment. Additional studies on the mechanisms of AHP are needed, for example, on the relative composition of the insoluble and soluble fractions after pretreatment. In the experiments reported here, the soluble and insoluble portions of the biomass were not separated after pretreatment, and therefore the total (free, oligomeric, and polymeric) Glc and Xyl composition of the biomass did not change between the original biomass and the hydrolyzed material.
AHP satisfies many criteria of an ideal pretreatment, in particular high yields of fermentable sugars at low enzyme loadings, low cost of pretreatment reactors, minimal post-pretreatment conditioning, low disposal challenges, and low heat and power demands . Further improvements will be necessary, however, before AHP can be considered an economically relevant pretreatment, in particular a reduction in the cost of hydrogen peroxide.