To be effectively practiced in a commercially relevant (i.e., economical) process context, enzymatic hydrolysis must be conducted with consideration of an entire lignocellulosic biomass conversion process. The choice of biomass feedstock and its specific compositional and structural attributes define the various types and amounts of structural carbohydrates that must be hydrolyzed, while the amounts and prop- erties of noncarbohydrate components (lignin, ash, protein, etc.) also determine the overall recalcitrance that must be overcome to effectively deconstruct the biomass feedstock and produce sugars in high yields. Biomass pretreatment processes have been developed to help overcome the natural recalcitrance of lignocellulosic biomass by rendering it much more susceptible to rapid and complete enzymatic deconstruc- tion. Additionally, the process context in which the resulting soluble carbohydrates are used to produce a chemical and/or fuel product of interest can also define the specific enzyme system and process configuration in which enzymatic hydrolysis is performed.
4.5.1 Feedstock
There is a wide variety of biomass feedstocks that are potentially viable for utilization in commercial-scale biochemical conversion processes. While biomass feedstocks all consist of three primary compounds (cellulose, hemicellulose, lignin), the relative amounts of these compounds can substantially vary across different feedstock classes and the structural features and associated cross-linking cause wide differences in biomass recalcitrance across feedstock classes (McMillan, 1997; Perez et al., 2010;
Zhao et al., 2012). In general, woody feedstocks (particularly softwoods) contain more lignin than herbaceous agricultural residues and perennial energy crops. The types of covalent linkages between lignin and hemicellulose also vary across feed- stock classes and can impact the requirement for accessory enzyme activities if such
linkages survive pretreatment in significant amounts. The levels of the minor sug- ars (arabinose, galactose, and mannose) also vary considerably with biomass type.
Softwoods typically contain more galactose and mannose than hardwoods, whereas hardwoods, herbaceous plants, and agricultural residues generally contain higher levels of arabinose and xylose. In some herbaceous energy crops and agricultural residues, arabinose levels are high enough that conversion of arabinose (in addi- tion to glucose and xylose) is required to achieve overall economic viability, which may impact the required amounts of additional accessory enzyme activities if levels of residual arabinan (or arabinose that has not fully been depolymerized) are still substantial after pretreatment of these feedstock types.
The presence of other relatively minor components in feedstocks (or extrane- ous substances harvested with the feedstock) can either directly influence the feed- stock reactivity and resulting enzyme activity and processing requirements or require inclusion of feedstock “pre-processing” steps to mitigate the effects of these compo- nents. While ash components (either from extraneous soil collected during the feed- stock harvesting process or from intrinsic ash components within the feedstock itself (Vassilev et al., 2010)) are not widely believed to directly impact enzymatic hydrol- ysis performance, their presence can impact pretreatment performance by causing a buffering effect on added pretreatment chemicals, which can indirectly impact enzymatic hydrolysis performance. Acetyl groups, which are often present on side chains of the primary hemicellulosic xylan backbone, have also been shown to affect pretreatment and enzymatic hydrolysis performance (Kong et al., 1992; Chang and Holtzapple, 2000). Research is being conducted to alter plant physiological pro- cesses to reduce the acetyl content in potential biomass feedstocks (Lee et al., 2011).
Chen et al. (2012) have recently developed a feedstock deacetylation process for a lignocellulosic ethanol process using corn stover that has been shown to lower the optimal pretreatment severity, improve overall enzymatic saccharification, reduce the requirement for AXE enzyme activities, and improve the fermentability of high con- centration sugar hydrolysates. As shown in Figure 4.4, a simple feedstock soaking deacetylation process can significantly improve the enzymatic hydrolysis of cellulose in pretreated corn stover, even when using whole pretreatment slurry at an enzymatic hydrolysis total solids loading of 25% (Chen et al., 2012). In this case, the combined (glucose + xylose) liquid-phase sugar concentration exceeds 150 g/L, although little additional xylose is enzymatically generated due to apparent end-product inhibition at such high sugar concentrations.
4.5.2 Pretreatment
Naturally occurring cellulolytic bacteria and fungi produce distinct enzyme compo- nents that work synergistically to degrade lignocellulosic biomass structural carbo- hydrates to sugars, as discussed in section 4.2. The enzymatic hydrolysis of native biomass is typically very slow, requiring that a pretreatment of the biomass be con- ducted in order to increase the rate and extent of enzymatic hydrolysis reactions to be economically viable for producing a commodity chemical and/or fuel product.
The pretreatment operation can directly influence the amount and types of enzymes
10%
20%
40%
60%
80%
Native
Deacetylated
Native
Deacetylated Glucose yield (%) Xylose yield (%)
Yield (%)
FIGURE 4.4 Whole slurry enzymatic hydrolysis yields of dilute acid pretreated corn stover, with and without feedstock deacetylation. Pretreatment conditions: 150◦C, 10 minutes, 8 mg H2SO4/g dry biomass, 45% solids in pretreatment reactor prior to steam injection. Enzymatic hydrolysis conditions: 25% total solids, no detoxification/conditioning (other than pH adjust- ment to 4.8), 50◦C, 168 hours, Novozymes Cellic Ctec2 (20 mg/g cellulose in pretreated slurry) plus Novozymes Cellic Htec2 (20 mg/g cellulose in pretreated slurry). The three bars in each group represent different corn stover varieties as described in Chen et al. (2012).
needed to saccharify all plant cell wall structural carbohydrates to monomeric sugars, as some pretreatment approaches can hydrolyze virtually all of the hemicellulose directly to monomeric xylose, while other pretreatment approaches largely leave the hemicellulose intact or only achieve partial hydrolysis to oligomeric sugars (Sun and Cheng, 2002; Mosier et al., 2005b; Elander et al., 2009). Therefore, the pretreatment process directly impacts the types and relative amounts of enzyme activities that are required in the subsequent enzymatic hydrolysis step, including hemicellulases and other accessory enzymes needed to deconstruct and hydrolyze hemicellulose that was not saccharified during pretreatment, in addition to hydrolyzing cellulose that is not typically converted at appreciable extents in leading pretreatment approaches.
Many pretreatment processes that do solubilize hemicellulose to a significant extent often do not fully convert soluble oligosaccharides to monomeric sugars, thus requir- ing oligomer-hydrolyzing enzyme activities if high yields of monomer sugars are ultimately desired (Qing et al., 2010). In order to achieve high yields of soluble sugars from hemicellulose, there are also covalent linkages between hemicellulose and lignin residues that require cleavage by appropriate enzymes if such linkages survive pretreatment (Kong et al., 1992; Chang and Holtzapple, 2000). In general, acidic (Lloyd and Wyman, 2005; Weiss et al., 2009; Humbird et al., 2011) and hot water/steam (Brownell and Saddler, 1987; Mok and Antal, 1992; Liu and Wyman, 2003; Mosier et al., 2005a) pretreatments solubilize greater amounts of hemicellulose
directly in the pretreatment step, while alkaline (Iyer et al., 1996; Chang et al., 2001;
Kim and Lee, 2005; Chundawat et al., 2010) and oxidative (Klinke et al., 2002;
Varga et al., 2003; Saha and Cotta, 2006) pretreatments primarily depolymerize and disrupt the lignin structure and/or the cellulose crystalline structure to enhance enzy- matic digestibility. Pretreatment approaches that solubilize hemicellulose to a large extent but leave much of the solubilized hemicellulose in oligomeric form, including less severe dilute-acid pretreatment conditions, may require additional mild thermo- chemical and/or enzymatic steps to convert oligosaccharides to monomeric sugars.
The enzymatic hydrolysis process configuration, along with the concentration of solubilized pretreatment products and the extent to which the pretreated slurry has been separated or washed, can impact the ability of enzyme systems to effectively hydrolyze hemicellulose-derived oligosaccharides (Shekiro et al., 2012).
While various pretreatment approaches have been extensively studied in laboratory-scale reactor systems, economic viability can only be achieved when pre- treatment is conducted in a commercially relevant manner. For large-scale processes, this generally involves the use of high solids reactor systems designed for continu- ous operation. Pretreatment performance can change significantly when scaling from batch, low solids, laboratory-scale systems to continuous, high solids, pilot-scale (or larger) systems (Schell et al., 2003; Shekiro et al., 2012). It is important to characterize enzymatic hydrolysis of pretreated solids from commercially relevant pretreatment systems, as factors such as grinding/compression and residence time distribution will cause changes in enzymatic hydrolysis performance as compared with similarly pretreated biomass generated in small laboratory-scale reactor systems.
4.5.3 Downstream Conversion
The simplest downstream conversion process configuration would utilize all biomass sugars derived from pretreatment and enzymatic hydrolysis in a combined, high con- centration hydrolysate. The 2011 National Renewable Energy Laboratory corn stover lignocellulosic ethanol design report (Humbird et al., 2011) describes such a scenario.
In this case, enzymatic hydrolysis is performed using the whole slurry obtained from a high solids dilute acid pretreatment, where all soluble compounds generated dur- ing pretreatment (including monomeric and oligomeric sugars, acetic acid, soluble extractives, and some solubilized lignin) are present in relatively high concentra- tions. Such a configuration is possible for lignocellulosic ethanol, as cofermenting microorganisms capable of utilizing five- and six-carbon sugars exist and have been shown to perform well under process-relevant conditions. Other enzymatic hydrol- ysis process configurations, including SSF and washed-solids enzymatic hydrolysis, both of which reduce the background and in-process sugar concentrations in order to alleviate end-product sugar inhibition of hydrolytic enzymes, are no longer found to be necessary, as improved commercially ready enzyme preparations can generate combined sugar concentrations greater than 150 g/L at reasonable enzyme loadings and reaction times. However, the presence of high concentrations of sugars and other compounds generated or liberated in the pretreatment step (primarily hemicellu- losic sugars, acetate, and sugar degradation products) can negatively impact enzyme
performance compared with the situation where the background concentration of such compounds is lower. Background concentrations of xylose from high solids dilute-acid pretreatment of corn stover can cause inhibition of enzymatic release of additional xylose from remaining xylan and/or oligomeric xylose (Shekiro et al., 2012). Additionally, acetyl groups present in the feedstock or when released as acetic acid into hydrolysates can impact performance of various cellulolytic enzymes (Kong et al., 1992; Chang and Holtzapple, 2000; Selig et al., 2009; Agger et al., 2010), unless selectively removed as described in section 4.5.1.
Enzymatic hydrolysis of pretreated biomass has generally been studied within the context of a process that requires released sugars to be in monomer form for subsequent microbial conversion to fuel and chemical products. Enzyme preparations with sufficient beta-glucosidase (and possibly beta-xylosidase) activities to produce high yields of monomeric sugars are typically required to prevent accumulation of dimeric sugars. However, there are some microorganisms that can utilize some sugar oligomers like cellobiose and would therefore not require an enzyme preparation rich in beta-glucosidase activity (Weimer and Zeikus, 1977; Yang et al., 2010; Ha et al., 2011). Additionally, processes are now being developed that utilize chemical catalysts to upgrade sugars (monomeric and oligomeric) and other forms of soluble carbon, such as sugar degradation products from pretreatment, acetic acid, and certain soluble lignin compounds. Such processes would not require all of the enzyme activities needed for hydrolysis processes optimized for producing monomeric sugars.