Cellulose is a ubiquitous polysaccharide component of plant cell walls. The cellulose content of cell walls of vegetative organs varies from ~20 to 40%. The lower end of the range applies to non-lignified primary walls, whereas the upper end refers to lignified secondary walls.

Cellulose molecules are long, unbranched chains of (3-D-glucopyranose (p-D-Glcp) residues joined by (1^4)-glucosidic linkages (Figure 4.1). The average degree of polymer­ization is about 6000 glucose units in primary walls and up to 14 000 in secondary walls. The molecules have an extended, ribbon-like conformation that allows parallel packing of the chains into three-dimensional microfibrillar aggregates stabilized by extensive intermolecu­lar hydrogen bonding and van der Waals interactions. The microfibrils have diameters of ~2- 4 nm inboth primary walls and lignified secondary walls of softwoods and hardwoods, but in the secondary walls ofboth types ofwoods the microfibrils apparently form aggregates ~ 14— 23 nm in diameter (10-13). Although the microfibrils of primary walls have often been as­sumed to consist of 36 individual cellulose molecules, solid-state 13C NMR studies have indi­cated fewer molecules (~20-25), mostly packed into highly ordered arrays (14,15). However, the molecular conformation and hydrogen bonding arrangement is different on the surfaces (16). (The structure of plant cell wall cellulose and the organization of the molecular chains in microfibrils are discussed in detail in Chapter 6, Structures of Plant Cell Wall Celluloses.)

Biochemical and thermochemical conversion technologies can be integrated into a large- scale biorefinery for additional efficiency, yield, and cost improvements. For example, an ad­vanced state-of-technology biorefinery optimized for maximum ethanol production could use an approach as follows: biochemical conversion extracts the carbohydrate portion of the feedstock and then converts it to fermentable sugars and, ultimately, ethanol. The remain­ing residue, primarily lignin, cannot be fermented, but it is a valuable organic feedstock. By directing this byproduct to a thermochemical process, it can be converted to syngas and, ultimately, ethanol. Integrating these technologies will improve the energy efficiency of the process, lower costs, and produce more ethanol than a standalone biochemical or thermochemical process.

Figure 2.13 depicts the advanced, integrated biochemical and thermochemical alcohol production scheme analyzed. Some of the lignin-rich residue is used to provide steam and electricity to the biochemical process, and the remainder is processed in the thermochem­ical process. The larger biochemical processes (8000-10 000 tonnes/day) expected in the 2020-2030 time frame will be needed to feed a reasonably sized gasification plant (1500­2000 tonnes/day) with only lignin-rich residues. The scale of the biochemical processing plant is five times larger than that targeted for near-term biorefineries, but the scale of the thermochemical conversion plant is the same.

This combined process can maximize feedstock handling efficiencies and heat and power integration. Integrated biorefineries can also process feedstocks with both high and low carbohydrate contents. A steady supply of low-carbohydrate feedstock could be fed directly into the thermochemical process, which allows increased size and some benefit to the capital cost. Integrated biochemical-thermochemical biorefineries also capitalize on the process improvements identified in the independent developments of the two technologies.

Given the variability of both the physical as well as chemical characteristics of the different feedstocks shown in Figure 2.2, robust conversion technologies must be developed that can economically accommodate this resource diversity. Additionally, feedstocks tend to be geographically diverse, i. e., softwoods in the Southeastern United States, and corn stover in the Midwest. Accommodating this diversity implies that the conversion technologies of future integrated biorefineries will be a function of the locally available feedstock resources. For this reason, cellulosic ethanol production technology must be sufficiently robust to optimize the conversion of multiple biomass resources to fuel.

Although there are a multitude of conversion technology routes under development for converting biomass to fuels and chemicals (15, 16), the predominant differentiation is the primary catalysis system (Figure 2.2).

Biochemical conversion uses biocatalysts (such as enzymes and microbial cells), heat, and chemicals to convert biomass first to an intermediate sugar stream and then to ethanol or other fermentation-produced fuel and co-products such as heat, power, and chemicals (17).

On the other hand, thermochemical conversion technologies use heat and/or physical catalysts to convert biomass to an intermediate product, and then a chemical transforma­tion step to convert that intermediate product into fuels and chemicals. Thermochemical conversion technologies tend to be grouped into two distinct categories for fuel production (18): gasification and pyrolysis. Gasification conversion reduces biomass to a fundamental chemical building block, syngas (carbon monoxide and hydrogen), that canbe reconstructed into ethanol and other fuel products through catalytic fuel synthesis processes (19). There are several gasification technologies capable of converting biomass to syngas via a network of chemical reactions, which can include partial oxidation, pyrolysis, and steam reforming, among others. There are several biomass gasification technologies under development that offer various pros and cons. A good recent review of the different types of biomass gasification technologies and their current state of development is provided by Spath and Dayton (19).

Pyrolysis is another thermochemical conversion technology but, unlike gasification that converts biomass to a syngas, pyrolysis converts the biomass to a liquid intermediate. Fast pyrolysis produces a pyrolysis oil or “bio-oil” in a short residence time process (0.1-2 s) in the absence of air at intermediate reaction temperatures typically in the range of 400-650° C. A good recent review on fast pyrolysis technologies and their current status is provided by Bridgwater and Peacocke (20, 21). In its produced form, the bio-oil is unacceptable for use directly as a transportation fuel because of its instability, high viscosity, and highly corrosive nature. Unfortunately, the highly heterogeneous nature of bio-oil makes the economical conversion to a fuel a very difficult challenge. However, many groups are currently research­ing the process of converting bio-oil to transportation fuels inside a petroleum refinery (22), and preliminary indications are that this work looks promising for the economical conversion of bio-oils to transportation fuel.

In addition to the base biological and thermochemical conversion routes of biomass to fuels listed above, there are a number ofhybrid processes that take advantage ofthe synergies of both biochemical and thermochemical technologies for innovative and economically promising options for biorefineries. Some examples of such innovative approaches are syngas fermentation (23) and aqueous phase processes for converting sugars, sugar alcohols, and polyols into alkanes ranging from C to C15 (24-26). In fact, it can be argued that even the biological approach described in the next section is a synergistic biochemical and thermochemical approach in the fact that a thermochemical pretreatment process is the first step required in the process. Huber and coworkers (16) provide a good recent comprehensive recent review of biorefinery conversion technology options and their current status.

In the context of individual biorefineries there really is no clear single technology choice. Biorefinery developers will need to best match the conversion technology with the charac­teristics of the locally available feedstock(s) as well as match co-product options with local chemical markets to develop the best match for their particular set of conditions. Therefore, to realize the ultimate potential of biofuels for supplying the largest possible percentage of transportation fuels, the suite of conversion technologies must be capable of accommodating the diversity of feedstocks. Hence, it is necessary to look closely at the feedstock conversion technology interface.

Thermochemical conversion technologies, particularly gasification approaches that re­duce the biomass to a syngas, are robust to feedstock physical and chemical diversity (27). Forest residues — from small-wood forest thinnings or residues such as “hog fuel” from the forest products industry-are considered primarily for thermochemical conversion because of their compositional variability and lack of control over this diversity. Additionally, these types of feedstocks tend to have higher lignin content than herbaceous feedstocks, making them more suitable for thermochemical gasification conversion. Agricultural residues, in contrast, can be considered better suited for biochemical conversion technologies. These resources are expected to have a more uniform chemical composition because they are de­rived from cultivated crops that can be genetically engineered or selected for properties more amenable to biochemical conversion technologies (such as low recalcitrance or high cellulose or xylan content); this also holds for energy crops. The more uniform chemical composi­tion of the feedstocks is in the macro sense, recognizing that in the micro sense there can be considerable variability. Biomass grown specifically for transportation fuel production can be engineered or selected to have the most desirable chemical and physical properties for a conversion technology. In addition, increasing the biomass resource base that can be biochemically converted to fuels provides an additional resource: lignin-rich fermentation residues that can be used for combined heat and power production or converted to biofuel in advanced, integrated biochemical-thermochemical biorefineries.

Although the long-term attractiveness of both biochemical and thermochemical ligno — cellulosic biomass conversion technologies looks very good, they are not yet economically competitive with either petroleum-derived gasoline or starch-based ethanol. To achieve economic viability of Phase III biorefineries, parallel efforts need to be undertaken to re­duce both the feedstock cost component as well as the conversion cost component. The next three sections describe the R&D and technical challenges of achieving near-term economical competitiveness of Phase III biorefineries. The final section discusses long-term technology and R&D needs to realize the ultimate potential of biorefineries in supplying a significant portion of transportation fuel needs.

Cellulose is often considered to be one of the most abundant biopolymers on earth. Although the composition of cellulose, the [3-1,4 linked linear glucose polymer (glucan), is relatively simple compared to other plant cell wall polysaccharides and the physical structure of cellulose is complicated. Celluloses can be crystalline, sub-crystalline, and even amorphous, depending on their tissue source in native plant, or the way that cellulose is isolated. The structural integrity of cellulose is believed to be one of the major causes of resistance to chemical and enzymatic hydrolysis.

Until recently, the crystal structures of native cellulose (cellulose I) were elucidated using algae Glaucocystis (Ia, triclinic) (15) and the tunicate Halocynthia roretzi (Ip, monoclinic) (16) celluloses. It has been suggested that the Ia and Ip allomorphs naturally coexist in various proportions in different organisms (17). These celluloses obtained from algae or tunicate; however, do not necessarily represent the cellulose in plant cell walls. The cellulose crystallite in higher plants is reported to be much smaller (3-5 nm in diameter) (9, 17), whereas they can be approximately 20 nm in diameter in some algae. Plant cell wall cellulose is embedded in a complex polymer matrix forming the microfibrils. Indeed, the interactions between cellulose and other polysaccharides are ubiquitous. The plant cell walls are dynamic and their compositions and properties may differ for different tissues, cell types, as well as different developmental stages. While the molecular structure of plant cell wall cellulose remains unclear, some scientists believe that cellulose in higher plant cell walls exists pri­marily in the Ip form with a small proportion of Ia form (13, 18-20). Other researchers have suggested that there is only the Ip form, with some disorder on the crystallite surface, based on solid-state 13C NMR spectroscopy studies (21). The challenges for characterizing plant cell wall celluloses stem not only from the limited resolution of available measurement techniques, but also from the processes commonly used to isolate celluloses, as well as sam­ple preparation methods. Typically, sequential extraction processes using acid and alkaline incubations are used, sometimes at high temperature. Such extensive sample processing can result in fiber aggregation (22). The isolated “microfibrils” are thought to be cellulose crys­tallites surrounded by a small pocket of partially hydrolyzed non-crystalline polysaccharides (13, 23). The question remains, how does this extensively treated “microfibril” relate to its native form in the plant cell wall? The plant cell wall cellulose crystallites can be summarized as being too small for traditional microscopy (3-5 nm in cross section) and containing a large proportion of disordered surface glucan chains (20, 24, 25). The detailed molecular structure of plant cell wall cellulose remains unknown.

2.3.1 Introduction

Thermochemical conversion technology options include gasification and pyrolysis. Although both processes show long-term promise, gasification approaches show consid­erable promise for near-term economically competitive liquid fuels production. As stated

Ethanol target price is competitive

2.50 2.00

1.50 1.00 0.50 0.00

Figure 2.8 Ethanol and gasoline price comparison.

above, the thermochemical process to liquid transportation fuels adds technology robustness to a scenario for producing a significant portion of transportation fuels from biomass. It can convert low-carbohydrate, or “non-fermentable,” biomass materials such as forest and wood residues to fuels at lower technical challenge levels than the biochemical conversion process route. This section describes the R&D needed to achieve economical competitiveness for a stand-alone biomass gasification/mixed alcohol process.

Biomass gasification converts heterogeneous feedstock supplies into a consistent gaseous intermediate that can be converted to liquid fuels. The product gas called “synthesis gas” or “syngas” has a low-to-medium energy content (depending on the gasifying agent) and consists mainly of carbon monoxide, hydrogen, carbon dioxide, water, nitrogen, and hy­drocarbons. Minor components, also referred to as contaminants, include tars, sulfur and nitrogen oxides, alkali metals, and particulates. These contaminants threaten the success of downstream syngas to liquid fuels conversion and must either be reformed or removed. Commercially available and near-commercial syngas conversion processes were evaluated on technological, environmental, and economic bases by Spath and Dayton (19). Their re­port provides the basis for identifying promising, cost-effective fuel synthesis technologies that maximize the impact of biomass gasification.

Figure 2.9 shows a representative block process flow diagram of a thermochemical process that produces ethanol from lignocellulosic biomass. The process also includes ancillary supporting operations such as feedstock interface handling and storage, product recovery, and product storage not shown in the figure. Phillips and coworkers (39) provide a detailed description of this process, which is being capable of producing economically viable ethanol from a plant processing 2000 tonnes/day of biomass. Although syngas-to-liquid processes are capable of producing a variety of transportation fuels, ethanol was selected here to provide synergy with the biochemical conversion route. A brief overview of the process developed by Phillips and coworkers is described below.

The feedstock interface addresses the main biomass properties that affect the long-term technical and economic success of a thermochemical conversion process: moisture content, fixed carbon and volatiles content, impurity concentrations, and ash content. High moisture

S, N, Cl mitigation CO2 removal H2/CO adjustment

Figure 2.9 Process flow diagram with research barriers for economical thermochemical ethanol production.

and ash content reduce the usable fraction of delivered biomass. Therefore, maximum system efficiencies are possible with dry, low-ash biomass.

Gasification is a thermochemical process that involves the thermal decomposition of biomass at temperatures that maximize syngas yield. Tar and char produced during decom­position may also react with steam, CO2, and hydrogen in the gasifier to produce additional gas. This is followed by partial oxidation of the fuel with a gasifying agent — usually air, oxy­gen, or steam — to yield raw syngas. The raw gas composition and quality are dependent on a range of factors, including feedstock composition, type of gasification reactor, gasification agents, stoichiometry, temperature, pressure, and the presence or lack of catalysts.

Gas cleanup removes contaminants from biomass gasification product gas. It generally involves an integrated, multi-step approach, which varies depending on the intended end use of the product gas. However, gas cleanup normally entails removing or reforming tars, acid gas removal, ammonia scrubbing, capturing alkali metal, and removing particulates. Gas conditioning is the final modification to gas composition to make it suitable for a fuel synthesis process. Typical gas conditioning steps include sulfur polishing (to reduce hydrogen sulfide to acceptable levels for fuel synthesis) and water-gas shift (to adjust the final hydrogen-carbon monoxide ratio for optimized fuel synthesis).

Comprehensive cleanup and conditioning of the raw biomass gasification product gas yields a “clean” syngas composed of carbon monoxide and hydrogen in a given ratio. This gas can be converted to a mixed-alcohol product. Separation of the ethanol and higher molecular weight alcohols from this product yields a methanol-rich stream that can be recycled with unconverted syngas to improve process yield. The higher-alcohol-rich stream yields byproduct chemical alcohols. The fuel synthesis step is exothermic, so heat recovery is essential to maximizing process efficiency.

4.2.1.2.1 (1^3,1^4)-$-d-GLUCANS

(1^3,1^4)-p-D-Glucans in seed plants (angiosperms and gymnosperms) are found ex­clusively in the walls of the grasses and certain related families (3, 17). Although the largest concentrations, up to 75%, of these polysaccharides occur in walls of the starchy endosperm of cereal grains such as barley (Hordeum vulgare) and oats (Avena sativa), they also occur, in lower concentrations in the primary walls of their vegetative organs (18, 19). For exam­ple, Smith and Harris (19) found 8.8% in primary walls from perennial ryegrass (Lolium perenne) stems.

The (1^3,1^4)-p-D-glucans are linear, unbranched polymers in which the (3-D-Glcp residues are joined by both (1^3)- and (1^4)-glucosidic linkages. Single (1^3)-linkages are separated by two or more (1^4)-linkages (Figure 4.1) and regions of two or three adjacent (1^4)-linkages predominate with longer (1^4)-p-D-glucosides (DP 5-14) ac­counting for <10% by weight of the molecules.

Shi-You Ding and Michael E. Himmel

3.1 Introduction

Lignocellulosic biomass has long been recognized as a potential sustainable source of mixed sugars for fermentation to fuels and other bio-based products. However, the chemical and enzymatic conversion processes developed during the past 80 years are inefficient and ex­pensive. The inefficiency of these processes is, in part, due to the lack of knowledge about the structure of biomass itself; the plant cell wall is indeed a complex nanocomposite material at the molecular and nanoscales. Current processing strategies have been derived empiri­cally, with little knowledge of the nanostructure of the feedstocks, and even less informa­tion about the molecular processes involved in biomass conversion. Substantial progress toward the cost-effective conversion of biomass to fuels is contingent upon fundamental breakthroughs in our current understanding of the chemical and structural properties that have evolved in biomass which prevent its disassembly, collectively known as “biomass recalcitrance.”

This chapter is not a strict review of plant anatomy. It deals only with those aspects of plant structure that are believed important to the availability and digestion of cell walls and the breakdown of biomass materials to fermentable sugars through chemical and bio­logical processes. The anatomy and ultrastructure of plant cell walls will be reviewed and emphasis will be given to recent progress made toward gaining an understanding of cell wall biosynthesis as well as characterization of cell walls at the molecular level using atomic force microscopy (AFM) and fluorescence labeling techniques. Future work and new techniques needed for characterization of the molecular architecture of the plant cell walls are also dis­cussed. In this context, plant cell walls from maize (Zea mays L.) stem are used as a model to cover what is currently known about cell wall structures related to biomass recalcitrance and subsequent conversion to biofuels.

The emerging biorefining industry is dependent on a large and sustainable supply of biomass resources provided at an effective cost and quality. In general, the feedstock cost can be sub­divided into two components: the grower payment (or “stumpage fee” for forest resources) to cover the value of the biomass and the feedstock supply system costs. Feedstock supply system costs include harvesting, collecting, storing, handling, transporting, and any prepro­cessing required. Ultimately, market dynamics will control the grower payment component of the costs, and feedstock supply logistics will dictate supply system costs. Foust and cowork­ers (4) set a target of $35/dry ton (2002 dollars) for the initial deployment of economically viable Phase III biorefineries in the United States, of which $10 was for the grower payment. This is much too low to cover production costs for perennial energy crops. However, consid­ering the 1.3-billion-ton potential, it is estimated that as much as 130 million tons could be accessed for a grower payment of less than or equal to $10/dry ton, primarily from existing agricultural and forestry residues.

As the industry expands from grain ethanol to cellulosic ethanol, it is expected that agri­cultural crop and forest logging residues will be the first resources developed for biorefining. Energy crops will be integrated into the agricultural cropping system as the biorefining in­dustry matures and creates a demand for them. An increase in energy crop production will likely occur as land managers (e. g., farmers and plantation foresters) use the additional crop options provided by the biomass energy market to maximize the productive capacity and economic returns of the land they manage.

The expanding use of lignocellulosic biomass resources will also create a demand for them, resulting in biorefineries paying more to access larger tonnages of the more expensive feedstocks (i. e., resources that require more than a $10/dry ton grower payment). However, feedstock demand will always be limited by the price the biorefining industry can pay while remaining competitive in the ethanol fuel market. Initially, government policies and programs maybe the means to access higher-value feedstocks. Up to and beyond the 2030 time frame, technology advancements will reduce feedstock supply system costs, which will then provide increased purchasing power for biorefineries to access higher-value biomass feedstocks. This strategy of improving supply and conversion technologies to purchase higher-value feedstocks is well established in other processing and refining industries (28). This combination of policy and technology advancement will help develop a biomass resource large enough to support the long-term goals of producing economical biofuels on a large scale.

Matrix polymers in plant cell walls are polysaccharides (hemicelluloses and pectins) and lignins. In contrast to cellulose, the matrix polysaccharides are synthesized in the Golgi ap­paratus, and deposited into the cell wall network while cellulose is synthesized (26). These polymers form a matrix network directly or indirectly associated with the surface of cel­lulose elementary fibrils. Matrix polysaccharides are non-crystalline structures and vary in glycosidic linkages, branching chemistry, and sugar residues. Lignins are polymers of mono — lignols (phenylpropane derivatives) providing mechanical strength and water resistance for the plant cell wall, as well as resistance of microbial attack (27-29). Lignin deposition is the final differentiation stage of plant cells that have thick secondary cell walls and proceeds via several phases, starting at the cell corners in the region ofthe middle lamella and the primary wall after the S1 layer formation has initiated. When the formation of the polysaccharide matrix of the S2 layer is completed, lignification proceeds through the secondary wall. The bulk of lignins are deposited after cellulose and hemicellulose have been deposited in the S3 layer. This is why lignin concentration is higher in the middle lamella and cell corners than in the S2 secondary wall (29-31).

Techno-economic analysis (40) has shown that removing chemical contaminants such as tar, ammonia, chlorine, sulfur, alkali metals, and particulates has the single greatest effect on the cost of liquid fuel synthesis. To date, gas cleanup and conditioning technologies are unproven in integrated biorefinery applications. Although water quench is an effective approach for removing tars and other particulates from the syngas, it is highly problematic from efficiency and waste disposal perspectives. Therefore, developing catalytic consolidated tar and light hydrocarbon reforming is desirable to eliminate the need for water quench.

An appropriate target for tar and light hydrocarbon reforming is to convert all tars and light hydrocarbons to syngas to a sufficient level as to not require an additional steam methane-reforming unit operation. Specific research needed to accomplish these objectives is as follows:

Perform tar deactivation/regeneration cycle tests to determine activity profiles to maintain the required long-term tar-reforming catalyst activity

• Perform catalyst studies to determine deactivation kinetics and mechanisms by probing catalyst surfaces to uncover molecular-level details

• Determine optimized catalyst formulations and materials at the pilot scale to demonstrate catalyst performance and lifetime as a function of process conditions and feedstock

• Design catalysts with higher tolerances for sulfur and chlorine poisons to enable further process intensification

• Lower or eliminate the sulfur and chlorine removal cost prior to reforming to achieve further reductions in gas cleanup costs

• Optimize the water gas shift activity of reforming catalysts to reduce or eliminate the need for an additional downstream shift reactor.