In a large-scale process, pre-impregnation of catalyst into large pieces of biomass (>1 cm) is often overlooked; however, milling biomass to reduce this problem can incur large energy and equipment costs [1, 14, 15]. This problem is compounded by the widespread use of process irrelevant biomass sizes for laboratory exper­iments. Most laboratory studies on biomass to ethanol conversion processes use finely milled materials (20-80 mesh is standard) where the effects of macroscopic transport processes are not easily observed or are masked altogether [43-45]. In larger pilot studies using compression screw feeders, these transport effects can be further masked by the high-shear feeder causing biomass size reduction [6, 8]. Often this size reduction occurs after catalyst impregnation, limiting catalyst effectiveness on pretreatment. A further complication is that compression of the feed stock may cause biomass pore structure collapse, leading to uneven heat and mass transfer dur­ing pretreatment [10,13] as well as limitation of catalyst access to the interior of the biomass.

Before larger biomass particles containing intact tissues are used in processing, it is essential to understand the catalyst transport processes and pathways and the limitations associated with them (Fig. 1). In living plants, vascular tissues such as xylem and phloem are the primary routes for transport of water and nutrients along the length of the plant stem and leaves. Additional transport within tissues and between adjacent cells is carried out through (1) the pits, areas of thin primary cell wall devoid of secondary cell wall between adjacent cells and (2) the apoplast, the contiguous intercellular space exterior to the cell membranes [46]. In dry senesced plants, studies with dyes to visualize fluid movement through tissues showed that the apoplastic space is the major catalyst carrier route, with limited fluid movement occurring through the vascular tissue [11]. In untreated biomass, the pits do not appear to support significant transport. It is probable that these pits disintegrate and open up during pretreatment allowing fluid to flow through [40]. Thus, new path­ways for catalyst penetration are formed either during the drying process that creates fractures in plant tissues or after some degree of biomass degradation.

The primary major barrier to fluid transport into native dry plant tissue appears to be air entrained in the cell lumen. Simple exposure of tissues to high tempera­ture fluids is insufficient to achieve catalyst distribution to all parts of the biomass [11]. The primary escape route for the intracellular air is most likely through pits. However, the small pit openings (approx 20 nm) could be blocked due to cell wall drying and water surface tension may prevent movement through these narrow open­ings. Forced air removal by vacuum provides additional driving force for the bulk fluid mobility necessary to enhance liquid and catalyst penetration into tissues as demonstrated by Viamajala and coworkers [11]. Heating dry biomass can minimize the amount of entrained air (due to expansion of air by heat) and assist in drawing liquid into the cells by contraction of the entrained air when cooled by immersion in catalyst-carrying liquid. Thus, bulk transport, rather than diffusive penetration, is the dominant mass transfer mechanism into dry biomass.

Although movement of fluids is associated with catalyst transport, the primary goal of catalyst distribution is to deliver the catalyst to cell wall surfaces con­taining fuel-yielding carbohydrates, rather than to empty cytoplasmic space in dry tissues. In fact, entrainment of fluids in the biomass bulk can be detrimen­tal to small time-scale dilute acid or hot water pretreatments, as the presence of excess water increases the net heat capacity of the material, increasing the heating time needed to achieve desired pretreatment temperatures. Data shown in Fig. 2 support this hypothesis. In this set of experiments, un-milled sections of corn stems

(~ 1 inch long) were saturated to various degrees with dilute sulfuric acid (2% w/w) and pretreated in 15 mL of the same acid solution at 150°C for 20 min. Milled corn stems (-20 mesh) pretreated under identical conditions served as controls. All pretreatments were performed in 22 mL gold coated Swage-Lok (Cleveland, OH) pipe-reactors, heated in an air-fluidized sand bath [42]. After pretreatment, whole stem sections were air-dried, milled and enzymatically digested for 120 h with a 25 mg/g of cellulose loading of a commercial T. reesei cellulase preparation (Spezyme CP, Genencor International, Copenhagen, Denmark) supplemented with an excess loading (90 mg/g of cellulose) of commercial Aspergillus niger cellobiase preparation (Novo 188, Novozymes Ltd., Bagsvaerd, Denmark) using procedures described previously [47]. Milled stover pretreated as controls in this experiment was dried and digested similarly, but without any further comminution.

In Fig. 2a, dry internodes pretreated without pre-impregnation of catalyst were poorly pretreated as evidenced by the high amounts of xylan remaining in the biomass after reaction. Stem sections pre-impregnated to achieve 20% satura­tion showed better reactivity and xylan removal and this trend continued when stem sections pre-impregnated to 50% saturation were pretreated. However, when completely saturated (100%) stem sections were pretreated, xylan conversion was observed to be lower. Milled materials with and without pre-impregnation of catalyst — conditions that would have lowest mass transfer limitations, showed com­parable pretreatment performance with each other as well as with the 50% saturated stem sections. These results confirm that only limited catalyst penetration and pre­treatment is achieved when air remains entrapped in cytoplasmic spaces such as in dry internodes. Enhanced catalyst distribution and transport dramatically enhances pretreatability up to a certain point, after which excess fluid impedes pretreatment. Similar conclusions on the negative impacts of poor bulk transfer on biomass pre­treatability can be inferred from other reported studies also. Tucker and coworkers

[10] observed poor pretreatability of biomass during steam explosion of corn stover when materials were not pre-wetted with dilute acid and ascribed their results to mass transport limitations. In another study Kim and coworkers [13] observed poor pretreatment of biomass when the biomass was pressed prior to pretreatment and hypothesized that the mechanical compression of biomass caused pore structure collapse resulting in formation of material that was relatively impervious to heat and mass transfer.

Enzymatic digestion results corresponding to pretreatments shown in Fig. 2a, are presented in Fig. 2b. As expected, release of monomeric sugars from pretreated whole stem sections was proportional to the degree of pretreatment they experi­enced. Unmilled biomass that was 50% saturated with acid before pretreatment showed better digestibility than the sections that were pre-saturated to lower or higher levels. Milled biomass, however, digested best, demonstrating the importance of enhanced enzyme transport — an outcome of the more thorough and uniform pretreatment of milled materials. With woody feedstocks, milling to fine parti­cle sizes may be impractical and pre-impregnation of biomass with catalyst, as practiced in the pulp and paper industry [48], might need to be utilized to improve conversion efficiencies.