8.3.1 Dilute Acid Pretreatment (DAP)

Among the numerous pretreatment techniques, dilute acid pretreatment (DAP) has been shown as a leading pretreatment process that is currently under commercial de­velopment. DAP can significantly reduce lignocellulosic recalcitrance by disrupting the composite material linkage, such as the covalent bonds, hydrogen bonds, and van der Waals forces [44]. The most widely used and tested approaches in DAP are based on dilute sulfuric acid (H2SO4) since it is inexpensive and effective [45, 46].

However, nitric acid [47], hydrochloric acid (HCl) [48], and phosphoric acid [49] have also been examined. In addition, it was shown that sulfur dioxide (SO2) was also an efficient acid catalyst in the DAP, especially for softwood [50-52]. How­ever, there are certain drawbacks with such an approach. It is difficult to handle SO2 (gas) at large scales, as safety issues may constitute a concern, and it is also a more expensive option as compared to similar alternatives such as using H2SO4.

Naturally, all wood materials and residues contain manganese elements, which are present sometime in high concentration depending upon the type of the wood ma­terials, varying from 10mg/kg to100mg/kg of dry wood. The importance of Mn2+ can clearly be found during the fungal decay on woody materials as it accumulates in the form of MnO2 precipitates. Indeed, the insoluble Mn4+ species deposits at the tip of new fungal hyphae in the early stages of infestation and growth [125]. As men­tioned earlier, Mn2+ stimulates the production of MnP and enhances the degradation of lignin components during oxidation reaction, where Mn3+ is generated by MnP and acts as a mediator for the oxidation of various phenolic compounds. Therefore, addition of Mn2+ increases the biological oxidation rate in biological pretreatment of lignocellulose. On the other hand, addition of Mn2+ inhibits the action of LiPs and its production [98, 125]. Hence, it is very essential to optimize the concentra­tion of Mn2+ in order to achieve better biological pretreatment. Indeed, in decaying wood, naturally a manganese concentration gradient is established, allowing soluble forms of manganese (Mn(I1) and Mn(III)) to diffuse into regions of low manganese concentration [126].

Hongzhang Chen and Junying Zhao

Abstract This chapter presents several integrated refining and fractionation tech­nologies for multiple products platform based on the understanding of the hetero­geneous property of corn stalk. This heterogeneous property was found at the levels of tissue, cell, and chemical composition. This property can be advantageous when proper fractionation technologies are adopted to pro-duce different products. Based on low pressure steam explosion technology, steam explosion integrated me-chanical carding, steam explosion integrated super grinding and steam explosion integrated washing and alkali extraction are designed to realize stalk fractiona-tion at different levels. Industry implementation of these technologies was also presented.

Keywords Lignocelluloses ■ Inhomogeneity ■ Steam explosion ■ Integrated ■ Fractional refining ■ Multi-production

4.1 Introduction

At present, stalk-based products from stalk refining has been regarded as perspective both in research and industrialization [1,2]. In that, the world is rich in stalk [3, 4] without application completely. However, stalk refining is hard to be industrialized for lack of cost-competitiveness.

The main reason for lack of cost-competitive is that stalk is converted into one product as a whole with single linear technology. For example, corn stalk is con­verted into ethanol with dilute acid pretreatment and simultaneous saccharification and fermentation technology [5]. The whole corn stalk is regarded as raw material with single conversion property. As a result, only about 30 % cellulose and a little hemicellulose are converted into ethanol. Other components including lignin are dis­charged as wastes. Therefore, this increases cost and causes pollution. Lignin from stalk could also be converted into various products [6]. However, if lignin in stalk

H. Chen (H) ■ J. Zhao

National Key Laboratory of Biochemical Engineering, Institute of Process Engineering, Chinese Academy of Sciences, Beijing 100190, China e-mail: hzchen@home. ipe. ac. cn

Z. Fang (ed.), Pretreatment Techniques for Biofuels and Biorefineries,

Green Energy and Technology,

DOI 10.1007/978-3-642-32735-3_4, © Springer-Verlag Berlin Heidelberg 2013 is converted into products as the only effective component, there would be the same problems including high cost and pollution. Hemicellulose in corn cob is usually converted into furfural [7]. There is report that microorganism could also convert pentose from hemicellulose [8]. Therefore, cellulose, hemicelluloses, and lignin in stalk are all potential resources. If only one component of stalk is converted with single linear technique, high cost and pollution would be unavoidable.

Lignocellulose materials including stalk are not only heterogeneous in the level of component but also in the levels of tissue and cell [9, 10]. Therefore, the conversion properties of stalk are also different. If the whole stalk is converted into one product, the product quality is hard to be controlled.

So it is necessary to fractionate stalk according to its heterogeneous characteristics in levels of component, cell, and tissue; then different fractions are converted into op­timum products. The whole process from stalk to products includes pretreatment, sac­charification, and fermentation. Fractionation should be realized in pretreatment pro­cess. Pretreatment methods reported focus on separation of one component including steam explosion [11], dilute acid or alkali [12], and hot water [13], and the whole stalk could not be fractionated completely into different fractions with only one of these technologies. Though there is report about integrated pretreatment technology, it is also amid at one product [14]. The intrinsic property of heterogeneity is ignored.

For example, steam explosion method is integrated with many other pretreatment technologies to improve ethanol yield including methanol, hydrogen dioxide, sodium hydroxide [15], ammonia, and sulfur dioxide. Sodium hydroxide and 1,4-dihydroxy — anthraquinone pretreatment are integrated for corn stalk to improve the methane yield [16]. Chemical and ultrasonic techniques are integrated to remove lignin and hemicelluloses in wood and get 5-10 nm fiber [17]. Therefore, integrated technology system should be set up to realize fractionation of stalk.

The author team have found heterogeneous characteristics of stalk and its different conversion properties [1, 9]. Moreover, steam explosion integrated pretreatment technology is set up to fractionate stalk into fractions and converted into multiple products. Different fractionation technologies are designed for different levels of fractionations.

This paper analyzes the heterogeneous characteristic of stalk and its require­ment for conversion technology, and then steam explosion integrated pretreatment technologies are introduced as a multiple products model. Finally, industrial demon­stration is given and advantages of fractionation refining technology oriented by multiple products are presented.

Armando T. Quitain, Mitsuru Sasaki and Motonobu Goto

Abstract Pretreatment has been considered an important step for efficient and effec­tive biomass-to-biofuel conversion. One of many promising methods of pretreatment includes the use of microwave (MW). MW-based pre-treatment approach utilizes both thermal and non-thermal effects generated by an extensive intermolecular col­lision as a result of realignment of polar molecules with MW oscillations. Compared to conventional heating, electromagnetic field generated by MW has the ability to directly interact with the material to produce heat, thereby accelerating chemical, physical, and biological processes. The advantages of employing MW rather than the conventional heating include reduction of process energy requirements, selective processing, and capability for instantaneous start and ceasing of a process. This also offers enormous benefits such as energy efficiency due to rapid and selective heating, and the possibility for developing a compact process. This chapter reviews recent advances on the utilization of MW irradiation for pretreatment of biomass for more efficient biofuel (bioethanol, biogas (methane), and biodiesel) production.

Keywords Microwave ■ Biomass ■ Sludge ■ Bioethanol ■ Biogas ■ Methane ■ Biodiesel ■ Free ■ fatty acids

6.1 Introduction

Pretreatment is an extremely important step in the synthesis of biofuels from ligno- cellulosic biomass or plant oils containing high fraction of free fatty acids (FFAs). Thorough knowledge of the fundamentals underlying various processes is necessary

M. Goto (H)

Bioelectrics Research Center, Faculty of Engineering, Kumamoto University, 2-39-1 Kurokami, 860-8555 Kumamoto, Japan e-mail: mgoto@kumamoto-u. ac. jp

A. T. Quitain ■ M. Sasaki

Graduate School of Science and Technology, Faculty of Engineering, Kumamoto University, 2-39-1 Kurokami, 860-8555 Kumamoto, Japan

Z. Fang (ed.), Pretreatment Techniques for Biofuels and Biorefineries,

Green Energy and Technology,

DOI 10.1007/978-3-642-32735-3_6, © Springer-Verlag Berlin Heidelberg 2013 in choosing a suitable pretreatment method appropriate for a particular structure of the biomass substrate and the hydrolysis agent [1].

The pretreatment process for the lignocellulosic biomass targets the breakdown of the lignin structure and the disruption of the crystalline structure of cellulose for an easy access of acids or enzymes into the matrix resulting into more efficient hydrol­ysis of the cellulose [2]. Comprehensive review of several methods of pretreatment including physical, pyrolysis, physicochemical, and biological has been reported by Kumar et al. [1], while pretreatment technologies based on the use of enzymes have been extensively summarized by Alvira et al. [3]. Mtui [4] also highlighted recent advances in the treatment of lignocellulosic wastes focusing on domestic and agro-industrial wastes. Mechanical, physical, and biological treatment systems were discussed, and among the systems mentioned, physicochemical and biological treatments seem to be most favored options for production of biofuel or bio-based products. In these published reviews, limited information has been reported regarding the use of microwave (MW), which is considered to be a very promising pretreatment technique.

Research on the development of pretreatment methods utilizing MW irradiation has reached a significant level during the past few years. Literature search using Scopus resulted into heavy turn outs. This dramatic increase in the number of pub­lished research literatures is due to the introduction of scientific MW equipment to the market in the 1980s, the search for a more energy efficient methods for bio­fuel production, and the increasing interests of the academe, research institutes, and industry on its utilization for biomass-to-biofuel conversion.

This chapter summarizes recent advances on the utilization of MW irradiation for pretreatment of biomass for more efficient biofuel (bioethanol, biogas (or methane), and biodiesel) production. The fundamentals of MW technology and its benefits are also discussed, while introducing some challenges of the technique, prospects, and future outlook.

Sugar beet pulp (SBP), a by-product of the sugar beet industry, is the fibrous material left over after the sugar is extracted from sugar beets and is mainly composed of cellulose, hemicellulose, and pectin. Beet pulp is used in countries with an intensive cattle raising industry, as livestock feed. In other countries, it is dumped in landfills. However, SBP can be an important renewable resource and its bioconversion appears to be of great biotechnological importance [69].

There are several reports on the pretreatments applied to SBP. Autoclaving of SBP at 122 °C and 136 °C for 1 h was reported to change its composition and physico­chemical properties causing increased swelling and improved solubility of pectins and arabinans [70]. Ammonia pressurization depressurization (APD) pretreatment where SBP is exploded by the sudden evaporation of ammonia, was found to sub­stantially increase hydrolysis efficiency of the cellulose component [71]. Recently, Kuhnel et al. [72] examined the influence of six mild sulfuric acid or hydrother­mal pretreatments at different temperatures on the enzymatic degradability of SBP and they found that optimal pretreatment at 140 °C of 15 min in water was able to solubilize 60 % w/w of the total carbohydrates present, mainly pectins.

Taking into consideration that SBP is carbohydrate-rich with a high carbon-to — nitrogen ratio (C/N, 35-40), that sugar beet farming is a widespread and already mature industry, and that beet pulp is abundant and cheap, this coproduct has potential for use as a renewable biomass feedstock for microbial fermentations for biopolymer production [71]. On the other hand, there are very limited numbers of reports on the use of SBP as a resource for microbial polysaccharide production. Yoo and Harcum investigated the feasibility of using autoclaved SBP as a supplemental substrate for xanthan gum production from X. campestris and they reported a production yield of

0. 89 g xanthan per gram of SBP in 4 days of fermentation time [54]. Sogutgu et al. [63] investigated the effects of autoclaving, reducing the particle size by milling and accessibility of SBP for EPS production by halophilic Halomonas sp. AAD6 cultures. In this study, milling of dried SBP in a mortar grinder, supplying SBP in dialysis tubes rather than directly in culture media and autoclaving SBP separately, and then adding to the fermentation media were all found to increase the EPS yields.

Stalk is heterogeneous, the same as its conversion property. According to the het­erogeneous characteristics of stalk, three platforms have been set up including low pressure, non-pollution steam explosion platform, solid-state fermentation platform and solid enzyme hydrolysis, fermentation, and separation coupling platform. For each platform, pilot test apparatus are designed, as well as industrial equipment. By process integration, multi-layered stalk refining has been established and indus­trialized in plant about ethanol, sheet, spinning. Especially, in September, 2010, Laihe Chemcial Co., Ltd., located in Jilin Province, China introduced this integrated process (Fig. 4.4 and 4.5). Steam explosion-alkali extraction-mechanical carding integrated pretreatment is applied. Its capacity is 300 thousand tons corn stalk annu­ally. The products from this technology are 50 thousand tons butanol, acetone, and

ethanol, 30 thousand tons pure lignin which could be converted into 20 thousand tons phenol formaldehyde resin adhesive, 120 thousand tons cellulose which could be converted into 50 thousand tons biological polyether polyol. The cost of solvent products reduce more than 50 % after cost apportion by lignin and cellulose. Many other stalk refining plants are also on the way.

Cost apportion Cost Multiple products oriented model Single product oriented model

Biomass, as a renewable energy source, refers to living or to recently dead biological material that can be used as fuel or for industrial production. Biomass comes in many different types, which may be grouped into five basic categories of material:

• Virgin wood, from forestry, arboricultural activities or from wood processing.

• Energy crops: high yield crops grown specifically for energy applications, for example, energycane, switchgrass, and miscanthus.

• Agricultural residues: residues from agriculture harvesting or processing.

• Food waste, from food and drink manufacture, preparation, and processing, and post-consumer waste.

• Industrial waste and co-products from manufacturing and industrial processes.

Biomass is a feedstock for chemicals, electricity, and natural gas, not just liquid fuels. It is essential to successfully grow energy crops at large scale. Biotechnology improvements and agronomics are both key to improving yields and low cost biomass will be critical to future energy production.

The US Department of Energy (DOE) and the US Department of Agriculture (USDA) are both strongly committed to expanding the role of biomass as an energy source. In particular, they support biomass fuels and products as a way to reduce the need for oil and gas imports; to support the growth of agriculture, forestry, and rural economies; and to foster major new domestic industries—biorefineries—making a variety of fuels, chemicals, and other products. As part of this effort, the biomass R&D Technical Advisory Committee, a panel established by the US Congress to guide the future direction of federally funded biomass R&D, envisioned a 30 % replacement of the current US petroleum consumption with biofuels by 2030 [4].

Primary forest resources are logging residues from conventional harvest opera­tions and residues from forest management and land clearing operations, removal of excess biomass from timberlands and other forestlands, and fuelwood extracted from forestlands. Secondary forest resources are primary wood processing mill residues, secondary wood processing mill residues, and pulping liquors (black liquors). Ter­tiary resources are urban wood residues—construction and demolition debris, tree trimmings, packaging wastes, and consumer durables.

Primary agricultural resources are crop residues from major crops, for example, corn stover, small grain straw and sugarcane baggase, grains of corn and soybeans used for ethanol, biodiesel, and bioproducts, perennial grasses, and perennial woody crops; secondary agricultural residues are animal manures and food/feed processing residues, and tertiary are municipal solid wastes, post-consumer residues, and landfill gases.

Forest resources (US) 368 million dry tones per year

Agricultural resources (US) 998 million dry tonnes per year

Total (US) 1,366 million dry tones per year

In biological pretreatment process, bacteria and actinomycetes are not as efficient as white — and brown-rot fungi. Very few bacteria, such as filamentous bacteria be­longing to the genus Streptomycetes are well known degraders of lignin, have been studied for pretreatment. These bacteria have been found to have some role in final mineralization of lignin. Non-filamentous bacteria Pseudomonas degrade very little amount of lignin. Since these bacteria do not have extra cellular oxidoreductase, which is one of the very essential enzymes for delignification and cannot be utilized in biological pretreatment. Actinomycetes are bacteria which form multicellular fila­ments; thus, they resemble fungi, also produce extracellular peroxidase as white-rot and brown-rot fungi, for example LiP-type enzyme. Streptomyces sp. EC1 produces peroxidase and cell-bound demethylase requiring H2O2 and Mn2+, both have been produced at relatively high levels in the presence of Kraft lignin or wheat straw [49]. Bacteria actinomycetes Streptomyces viridosporus have also been studied up to some extent [95]. Godden et al. [96] studied activity of peroxidase and catalase in six actinomycetes strains.

Thermophilic actinomycetes have been isolated from a wide range of natural sub­strates, for example from desert sand and compost. The genera of the thermophilic actinomycetes isolated from compost include Nocardia, Streptomyces, Thermoacti — nomyces, and Micromonospora. Actinomycetes degrade lignin as their primary metabolic activity and at high nitrogen levels compared to white-rot fungi, most of which degrade lignin via their secondary metabolism [97].

The lignin-degrading actinomycete species examined till date have been shown to oxidatively de-polymerize lignin. The primary degradative activity of actinomycetes is solubilization of lignin, with low levels of mineralization compared with the white — rot fungi. The depolymerization reactions produce a modified water-soluble, acid precipitable polymeric lignin as the principal lignin degradation product. The range of actinomycete species capable of metabolizing lignin is still unknown. Moreover, the strains examined thus far solubilize lignin to an acid-precipitable polymeric lignin-like product.

The major factors influencing the un-catalysed steam explosion processes are res­idence time, temperature, moisture content and particle size [20]. The process conditions which result in the best substrates for hydrolysis and the least amount of soluble sugars lost to side reactions (i. e. sugar dehydration) usually considered to be the optimum conditions [13]. The use of steam temperatures ranging from 140 to 240 °C have been investigated in the literature and its influence on the overall process efficiencies is usually associated with the period in which the biomass is exposed to the steam pre-treatment. An optimum solubilisation (and hydrolysis) of the hemi- cellulosic component of lignocellulosic biomass was reported to be realised either by using a combination of high temperatures and short residence times (i. e. 270 °C, 1 min) or lower temperatures with longer residence times (i. e. 190 °C, 10 min) [24]. This is due to the fact that at lower steam pre-treatment temperatures (i. e. 190 °C) the recovery of the obtained hemisellulosic sugars in the lignocellulosic hydrolysate is maximised, with the acid-labile biomass polysaccharides partially converted to water soluble sugars [25]. On the other hand, the drastic conditions provided by the use of increased steam temperatures would most likely facilitate an enhanced accessibility of the macromolecules of the biomass material, but with inevitable sugar losses [13]. The employment of high steam temperatures has also been demonstrated to lead to an increase in the relative amount of acid-insoluble lignin in the post-treated materials [23]. With the use of steam temperatures (i. e. 220-240 °C) and residence times, con­densation reactions involving the by-products are derived from the biomass lignin and hemicelluloses. The acid-soluble lignin was observed to increasingly occur, re­sulting in the formation and accumulation of acid-insoluble polymeric materials [26]. This knowledge of the extensive condensation driven modification of the biomass lignin during steam pre-treatment has two important implications: That the formed polymeric materials could cause an apparent increase in the overall lignin yields (even potentially higher than the theoretical lignin yield, based on the content of the starting biomass), and the fact that part of these by-products are likely to remain in the steam pre-treated material even after employing additional washing steps (i. e. alkaline washing) [13]. A careful compromise must therefore be considered with the use of these two important conditions with opposite trends. The selection of the best temperature and time conditions could therefore be dependent on the other pa­rameters such as the subsequent processing and conversion steps to be applied after the biomass pre-treatment and the targeted fuel or chemical which is aimed to be produced. In general, an increase in the steam temperatures would correspond to a decrease in the carbohydrate yields, while longer reaction times have been seen to favour an increased lignin condensation and a degradation of pentosan, with acid hy­drolysis observed to predominate over degradation reactions with the use of shorter exposure times [13]. Furthermore, the use of an energy-material assessment would prove useful to ascertain the benefits (if any) of compromising high-energy inputs with the potential substrate conversion efficiencies.

The extent of biomass drying or water content is an important economic and technical parameter with the processing of lignocellulosic biomass. This is since the biomass costs could be substantially increased depending on the methods used in achieving a reduction in the biomass moisture content. To minimise processing costs, single fuel and chemical production systems as well as biorefineries would therefore ideally prefer the utilisation of cheap unprocessed biomass (containing high mois­ture content) or naturally dry biomass (containing ~5-15 % moisture). Investigations have been carried out on the influence of the use of lignocellulosic biomass with vary­ing moisture contents on the steam pre-treatment process efficiencies. Using ‘green’ freshly harvested and air-dried Aspen wood chips (3.2 mm, in the direction of the fibres), with moisture contents (oven dried basis) of 108.2 and 7.16 %, respectively, investigations carried out in [21] showed that statistically comparable reducing sugar yields (after enzymatic hydrolysis) were obtainable with the use of the different lev­els of water in the biomass materials. However, with an increase in the chip size, the samples with lower moisture contents (i. e. the air died chips) were observed to attain the steam temperatures within a shorter time, thus facilitating improved and quicker pre-treatment conversions. The steam requirements for the auto-hydrolysis process as examined by that study were therefore observed to increase with an increase in the lignocellulosic biomass size and moisture content.

Regarding the influence of particle sizes on the biomass steam pre-treatment, the application of mechanical size reduction schemes, that is, chipping and milling is usually carried out before the pre-treatment stage mainly to improve handling and to ease the biomass transportation from the acquisition site. The biomass size has also been described to be a critical parameter to be considered for the pre-treatment and conversion reactor designs [21]. The use of small biomass sizes were discussed to be preferable for most batch and continuous process operations due to the enhanced heat transfer facilitated with the use of such sizes for the biomass treatment [13]. However, much finer materials with smaller particle sizes (i. e. sawdust) have been seen to be difficult to utilise in batch units, with the use of plug flow reactors em­ployed to increase the pre-treatment conversion efficiencies [13]. An investigation on the influence of a range of lignocellulosic biomass (Brassica sp.) with particle sizes of 2-5, 5-8 and 8-12 mm respectively using process steam temperatures of 190 and 270 °C and a residence time of 4 and 8 min as carried out in [16] showed that an extensive size reduction was not desirable for optimum pre-treatment of the lignocellulosic biomass using steam for all the different temperature and residence time conditions examined. This was demonstrated by the higher cellulose concentra­tions and subsequent enzymatic digestibilities exhibited by the largest particle sizes (i. e. 8-12 mm) studied.

Another important aspect with the application of the non-catalytic steam pre­treatment route is the consideration of the need to employ the explosive decompres­sion step or not. Studies carried out in [22] using Aspen chips concluded that the explosive decompression step of the steam explosion process was not important and contributed little or nothing to hydrolysis of the lignocellulosic biomass and eventual accessibility of the biomass cellulose contents. Similarly, the use of steam explo­sion for the pre-treatment of green Eucalyptus chips was observed not to yield any considerable hydrolytic improvements; however, the use of high-pressured steam without an abrupt decompression step for the pre-treatment of air-dried Eucalyptus chips was seen to result in the production of poor substrates for further hydrolysis, thus suggesting that the use of the explosive decompression scheme is most suited for hardwood chips with a low moisture content [27]. The results of that study were however in somewhat in contrast to that presented in the patent proposed in [28], which discussed that the explosion step was essential for the production of hydrolytic substrates with improved macromolecular accessibility.

5.2.3.1 Torrefaction

Torrefaction is used as a pretreatment prior to pelletization for upgrading the woody biomass primarily for energy production [36]. Torrefaction is a thermochemical treatment that subjects the biomass under heat at the reaction temperature between 200 and 300 °C in an inert medium (e. g., nitrogen) for a certain period of time (ranging between several seconds and an hour) depending on the particle size.

Reaction Chemistry

A detailed review of torrefaction chemistry and its process conditions is summa­rized in the literature [37, 38]. The reactions that take place for torrefaction are mainly decomposition and which can further divide into (1) drying, (2) depoly­merisation and recondensation, (3) limited devolatilization and carbonisation, (4) extensive devolatilization and carbonisation throughout the reaction temperature regime (Fig. 5.3). Hemicelluloses soften at temperatures between 150 and 200 °C and undergo dehydration, deacetylization, and depolymerization reactions at processing temperatures in the 200-300 °C [ 39,40]. Xylan is the predominant form of hemicel — lulose for hardwood while glucomannan is the predominant form of hemicellulose of softwood. Xylan tends to break down more quickly than glucomannan at lower tem­peratures. Therefore, hardwood has a higher breakdown of hemicellulose (or higher mass loss) than that of softwood when treated at the same temperature. This suggests

that different species of wood have different torrefaction kinetics which is worth for further in-depth investigation. A simple one-step (single stage) kinetic model with the first-order reaction of chemical reactions of cellulose, hemicellulose, and lignin during torrefaction of British Columbia (BC) softwoods was reported for slow resi­dence time reaction while a two-component and one-step first-order reaction yielded a better prediction of chemical components for the torrefaction with short residence time [41].

A small degree of degradation of cellulose and/or lignin also occurs during tor — refaction [43]. When the reaction temperature is high (>270 °C), a greater proportion of cellulose degradation was reported. In contrast, lignin is relatively stable and does not undergo significant chemical changes during torrefaction even at high temperature.

Particle Size and Residence Time

Particles with small size distribution would be recommended for some reactor types to achieve an optimized torrefaction efficiency and product quality. Note that different biomass species have different physical properties (porosity, specific heat capacity, thermal conductivity, particle size distribution of grinds under the same size reduction process, etc.) that further introduces a non-homogenous reaction due to heat and mass transfer limitation and subsequently results in a non-homogenous torrefied product.

In general, large particles within a wider range of particle size distributions may not be completely torrefied, and small particles may be over-carbonized under the same torrefaction condition. Therefore, a narrow or mono-disperse particle size dis­tribution of a certain specific wood particles may be ideal for a certain type of torrefaction reactor operating at a specific temperature and time. If the particle size distribution of raw materials is a bi-modal distribution, sieving and separation of mixtures would be recommended instead of torrefying the whole mixture of par­ticles. Large particles may require more than one pass milling in order to achieve smaller particle size suitable for homogenous torrefaction. The particle size distri­bution of ground particles depends on the biomass species, the types of mill, the MC of the biomass, and other factors, etc. Pretreatments (e. g., drying, size reduction) prior to torrefaction are necessary and critical to control the feedstock properties for producing a homogenous product quality.

Residence time also affects the mechanism of torrefaction decomposition. Resi­dence time is related to the reactor design (i. e., size) and operation condition (e. g., feeding velocity). Long residence time allows a greater degree of devolatilization from the biomass. Particle size and MC of wood particle also affect the heat and mass transfer. Small particles require less residence time to be treated in order to achieve the same degree of chemical reaction as the longer residence time is required for the large particles to allow nitrogen diffusion to initiate the decomposition reac­tion. A slow heating rate is critical for product homogeneity. Slow heating facilitates a uniform temperature gradient across the particle during torrefaction. Less volatile with low heating value is vaporized and thus results in a higher yield of solid residue with high heating value for pellet production.