Carbon Dioxide in Biomass Processing: Contributions to the Green Biorefinery Concept

Transcription

1 This is an open access article published under an ACS AuthorChoice License, which permits copying and redistribution of the article or any adaptations for non-commercial purposes. pubs.acs.org/cr Carbon Dioxide in Biomass Processing: Contributions to the Green Biorefinery Concept Ana R. C. Morais, Andre M. da Costa Lopes, and Rafał Bogel-Łukasik* Downloaded via on September 7, 2018 at 23:48:11 (UTC). See for options on how to legitimately share published articles. Unidade de Bioenergia, Laboratoŕio Nacional de Energia e Geologia, I.P., Estrada do Pac o do Lumiar 22, Lisboa, Portugal CONTENTS 1. Introduction Biomass as Feedstock for Biorefinery Supercritical Fluid Technology scco 2 and Its Place in Green Chemistry High-Pressure CO 2 /H 2 O Technology 6 2. Scope of the 7 3. Supercritical CO 2 Pretreatment Process Condition Effects Water Influence Biomass Sources and Recalcitrance CO 2 :Biomass Ratio Effect of High-Pressure CO 2 on the Morphology of Lignocellulosic Biomass Enzymatic Hydrolysis in Supercritical CO High-Pressure CO 2 /H 2 O Conversion Technologies Hydrolysis of Biomass-Derived Carbohydrates Hemicellulose Cellulose Starch Hexose-Derived Sugar Conversion in High- Pressure CO 2 and CO 2 /H 2 O Mixtures Conversion of Biomass-Derived Compounds in High-Pressure CO 2 and CO 2 /H 2 O Mixtures Hydrogenation Conversion of Proteins Delignification and Lignin Extraction under scco Lignin Depolymerization Overview of CO 2 Applications within the Biorefinery Concept Economic Aspects of CO 2 Processing of Biomass Perspectives Conclusions 23 Author Information 23 Corresponding Author 23 Notes 23 Biographies 23 Acknowledgments 24 References INTRODUCTION The 21st century is witnessing a huge demand of fossil reserves coupled with a rapid reduction in readily and economically reachable oil feedstocks. 1,2 The present energy demand is not fulfilled from fossil fuel sources, making the world exposed to geopolitical risk. Furthermore, concerns regarding the security of the supply chain and the environmental impacts have resulted in an ever-increasing shift of global energy policies to seek alternative technologies and sustainable sources of energy, materials, chemicals, and value-added products. 1 Recently, the need for development of an economy based on renewable resources has been recognized by society, and diverse R&D activities have started to be funded to accomplish this aim. 3 However, generation of bioproducts based on sustainable supply chains poses vast challenges for an eco-based economy. The simplest way to provide a supportable supply chain is through the employment of renewable biomass feedstocks, which is the only sustainable option to substitute for fossil fuel resources, as sources of organic compounds over a relatively short time scale and with limitless supply. All these factors have reinforced the need for research on production of biomassderived commodities produced in a sustainable manner. 4 The biorefinery concept considers the use of biomass as a low-cost feedstock for the chemical and biological industries. The most widely used description of biorefinery is a definition adopted by International Energy Agency Bioenergy Task 42. It states that biorefining is the sustainable processing of biomass into a spectrum of marketable products and energy. 5 In other words, the biorefinery is a term used to define industrial facilities that cover an extensive range of combined technologies in which biomass is transformed and converted, in a sustainable manner, into a wide range of value-added products, leading to direct similarities to today s petrorefineries. Following this idea, the aim of future biorefineries is the extraction of high-value chemicals present in biomass, such as flavoring agents, fragrances, and nutraceuticals and, in the next step, processing of biomass-derived polysaccharides, lignin, and proteins toward bioderived materials, fuels, and other commodities. 6 Received: June 23, 2014 Published: November 20, American Chemical Society 3

2 Chemical s 1.1. Biomass as Feedstock for Biorefinery The worldwide production of lignocellulosic biomass is estimated as million tons. 7 The lignocellulosic materials (LCMs) include hardwoods, softwoods, and residues from agricultural and forest activity as well as energy crops. LCMs demonstrate several advantages, namely, their lower price in comparison to that of traditional agricultural feedstocks, not requiring arable land and fertilizers to grow, and not competing with food and feed sectors. 8,9 The primary metabolites of LCMs are polysaccharides, for example, cellulose and hemicellulose, plus lignin as depicted in Figure 1. Secondary Figure 1. Schematic presentation of the lignocellulosic biomass structure. Adapted with permission from refs 12 and 13. Copyright 1995 Elsevier and 2013 InTech, respectively. Chapter 16 of ref 13 was distributed under the terms of the Creative Commons Attribution License. compounds such as waxes, terpenoids, proteins, and phenolic compounds such as lignans account for up to 10% of the dry weight of LCMs. 10 All these fractions could be extracted and later converted into valuable products, making the valorization of biomass more sustainable. In addition, the starchy lignocellulosic biomasses, such as waste from starch processing factories, e.g., potato processing, are important sources of another carbohydrate for valorization. 11 Cellulose is one of the main fractions present in biomass and is the world s most abundant biopolymer. It is a homopolysaccharide composed of glucose units linked by β-1,4-glycosidic bonds in which hydroxyl groups are oriented to arrange strong intra- and intermolecular hydrogen bonds. This internal organization results in ordered micro- and macrofibrils, which are present in crystalline and amorphous structures. 14 The presence of intramolecular bonds inside glucose monomers and intermolecular bonds between various chains contributes to a complex supramolecular structure. The cellulose linearity, crystallinity, and fibrous structure are responsible for the mechanical strength of the plant cell. Hemicellulose in turn can be described as a heteropolymer located in the primary and secondary plant cell walls. It is composed of various monosaccharides, including D-glucose, D- xylose, and D-mannose, and displays both branched and amorphous structures. The rate of hydration depends on the sugar type and decreases in the following order: xylose, mannose, and glucose. Thus, hemicellulose is more susceptible to hydrolysis than cellulose, and dehydration of hexoses produces 5-hydroxymethylfurfural (5-HMF), while furfural is produced from the pentose fraction. 15 Lignin is a complex network of three different phenylpropane units, namely, p-hydroxyphenyl, guaiacyl, and syringyl. Additionally, lignin is attached to hemicellulose by covalent linkages, helping to establish the rigid matrix of LCMs. Starch is an α-liked polysaccharide composed of two components with various molecular masses: 20 30% linear amylose and 70 80% 1,6-α-linked branched amylopectin. The efficiency of starch hydrolysis depends on the potential for acids to break down a starch granule and the physical distribution and structure of each starch component. As with cellulose, starch hydrolysis gives D-glucose and its degradation products, such as 5-HMF, levulinic acid, and formic acid. 16 When starch is partially hydrolyzed, it produces maltodextrins, nutritive saccharide polymers, and nonsweet products used as texturizers, fat substitutes, freezing aids, and filming control agents among other applications. 17 The diverse composition of biomass and the interactions between fractions make its structure very complex and resistant to deconstruction. The molecular interactions define a complex and recalcitrant structure, 18 which is difficult to hydrolyze and thus leads to harsh and expensive fractionation methods. 19 Thus, processing of this feedstock into reactive intermediates, which can be used as building blocks for fuels and chemicals, is very difficult. 20 The production of biofuels or other value-added products from these kinds of biomasses comprises several steps, such as hydrolysis of immobilized polysaccharides (mostly hemicellulose) and their further conversion, separation of lignin from the remaining residue, and purification of the final products. 1 Despite dozens of years of research, the hydrolysis of polysaccharides into fermentable sugars is still the major obstacle, since it depends on many structural, compositional, and physicochemical factors. These factors have a significant effect on the entire biomass valorization process. 1 The current valorization technologies are still mostly characterized by low yields and high costs, which burden the production of chemicals and fuels at competitive costs. Thus, the choice of the technology employed as a biomass processing method should be based on the characteristics of each biomass and the impact of the process on the further valorization pathway by, for instance, production of toxic compounds for metabolism by microorganisms, energy demand, and waste production. 21,22 Because of this, pretreatment is one of the most challenging steps in biorefinery since it is among the most expensive and has the most influence on subsequent stages such as hydrolysis, fermentation, and downstream processing Several pretreatment technologies have been under exhaustive research for many years. Among them are conventional (e.g., dilute acid hydrolysis, alkaline, 20 hydrothermal (steam explosion, 26 and autohydrolysis 27 )) and novel (ionic liquids and sub/supercritical fluids ) processes. The conventional pretreatments lead generally to low sugar yields, are unselective, and normally involve more severe conditions. Thus, they favor the formation of degradation products, resulting in a decrease of the efficiency of the process, higher investment needs, and elevated processing costs. 35,36 Additionally, there is a continuous need for newer and cleaner methods of biomass processing with less energy demand and lower waste generation in which the product of interest can be formed, isolated, and separated from the biomass at competitive costs. Zhu and Pan published a review on conventional pretreatment processes (supercritical CO 2 (scco 2 ) is not included) in which the aspects related to energy demands of each technology discussed are presented. They showed serious concerns 4

3 Chemical s Figure 2. Temperature-provoked changes in propane behavior from a liquid gas mixture to supercritical conditions (from left to right). Adapted with permission from C. M. Rayner, A. A. Clifford, and K. D. Bartle, University of Leeds, U.K. associated with the conversion of biomass-derived compounds at competitive costs. 37 The development of sustainable technologies that lead to innovative separation and depolymerization processes is one of the solutions for this problem Supercritical Fluid Technology A fluid is termed supercritical when it is above the critical temperature (T c ) and pressure (p c ) and below the pressure needed for condensation. As defined by Himmelblau, the critical state for the gas-liquid transition is the set of physical conditions at which the density and other properties of the liquid and vapour become identical. 38 As shown in Figure 2, when the gas and liquid phases at supercritical conditions coexist and become indistinguishable, the critical point is then reached. The unique physicochemical properties of supercritical fluids (SCFs), such as liquidlike density as well as gaslike diffusivity and viscosity, are intermediate between those of both gas and liquid. The solvating power of an SCF is much lower than that of ordinary fluids, which are expected to change the reactivity and selectivity of reaction. These and many other properties are tunable; for example, the solubility can be easily increased by slight changes in temperature and pressure near to the critical point. 39 Table 1 summarizes the critical temperatures and Table 1. Critical Temperatures, Pressures, and Solubility Parameters for Pure Chemicals critical conditions chemical T ( C) P (bar) solubility param (cal/cm 3 ) 0.5 carbon dioxide ammonia hexene acetone methanol ethanol propanol toluene water pressures as well as the solubility parameters for some chemicals commonly used as SCFs. Some of the listed SCFs found an application in biorefinery processes mostly because high reaction and conversion rates by variation of temperature, pressure, and residence time are easily achievable with these fluids, in addition to a fast and clean production and elimination of the need for a catalyst. 40 Water is one of the fluids considered as potentially interesting in high-pressure biomass treatment Hydrothermal pretreatments were applied in the pulp industry, where biomass can be impregnated with water under high pressure, thus improving the removal of most of the hemicellulose and leading to a higher enzymatic saccharification rate. 45 This technology shows several advantages, such as particle size reduction and lower production of residues requiring neutralization. Furthermore, after high-pressure hydrothermal pretreatment, the pressure can be quickly reduced by simple discharging, leading to explosive decompression of the feedstock and thus enhancing the cellulose accessibility for enzymatic attack. However, it is important to underline that these technologies face several engineering challenges, which should be taken into consideration prior to considering them as alternatives for conventional biomass processing methods. One of these limitations is the requirement of specialized equipment capable of withstanding high temperatures and pressures plus corrosive media, due to the high acidity and elevated level of dissolved oxygen, especially with processes using supercritical water. The energy demand to reach supercritical conditions might be another important drawback in the case of some fluids. Other limitations are the presence of ionic inorganic species plus unspecific kinetics and reaction pathways. Despite these limitations, a recent evaluation of biofuel production reported that supercritical water may be able to increase the conversion of lignocellulosic residues to biofuels scco 2 and Its Place in Green Chemistry. Nowadays, there is a lot of interest in the principles of green chemistry and, in particular, the replacement of common solvents by safer solvents/chemicals such as CO 2 and the use of renewable feedstocks that help to make processes more green. 47 scco 2 fulfills these requirements and has been used recently in biomass processing. 1,32,40,43,48 51 The number of scientific publications related to the use of scco 2 as a green technology medium for biomass processing is increasing in proportion to the interest in the development of sustainable technologies for biomass conversions. scco 2 is a readily available, nontoxic, nonflammable, inexpensive solvent that does not contaminate the product. 1 Carbon dioxide has a low critical temperature (31.1 C) and pressure (73.8 bar); hence, the energy demands required to bring CO 2 to supercritical fluid conditions are relatively low in comparison to those of other solvents, as depicted in Table At the present, the movement toward the use of solvents that can help to achieve better performance and replace hazardous solvents in organic reactions is strongly promoted. However, CO 2 even at supercritical conditions is a bad solvent. This is mostly due to the poor solvent capacity toward polar compounds. Conversely, CO 2 as a nonpolar compound is a good solvent for nonpolar materials, 53 and this clear difference in solubility and its easily adjustable solvent power make CO 2 a selective solvent in many applications. Furthermore, CO 2 has a strong solvent power due to its low polarizability ( cm 3 ) and large quadrupole moment ( cm 2 ), 54 which makes the comparison to conventional solvents difficult because of its ambivalent 5

4 Chemical s properties. These characteristics make CO 2 a special solvent for monomers but a poor solvent for polymers, leading to easy separation as well as faster impregnation in the complex structure of LCMs, in contrast to other solvents. Although the use of scco 2 brings the environmental benefits depicted above, the economic viability should be demonstrated, to make this technology appropriate for use in industrial processes within the biorefinery concept High-Pressure CO 2 /H 2 O Technology. Highpressure CO 2 /H 2 O mixtures are broadly used in acid-catalyzed reactions. 55,56 This is consistent with Sheldon s statement in his recently published critical review,...the use of water and supercritical CO 2 as reaction media is also consistent with the current trend towards the use of renewable, biomass-based raw materials His statement clearly shows that water is an important medium and together with CO 2 will play an important role in novel methodologies of biomass processing. Thus, knowledge concerning the phase behavior (Figure 3) of Figure 3. Temperature pressure composition phase diagram of the binary system CO 2 /H 2 O. Reprinted with permission from ref 58. Copyright 1964 American Journal of Science. the mixture of CO 2 and H 2 O is crucial for high efficiency of the high-pressure CO 2 /H 2 O processes. This knowledge could be useful to tune the reaction conditions to the process requirements for maximal efficiency. The updated research on the importance of water in biomass processing reveals that the presence of H 2 O is an advantage, since it allows for high sugar yields, increasing the effectiveness of the process. 59 At pressures 200 bar and temperatures between 160 and 250 C, the CO 2 -rich phase has a density very similar to that of a liquid containing at least 30 mol % water, while a water-rich phase contains up to 2 mol % CO Another important aspect of this binary system is that CO 2 / H 2 O leads to in situ formation of carbonic acid, which dissociates in two stages according to the following equation: + CO2 + 2H2O HCO3 + H3O ; 2 + HCO3 + H2O CO3 + H3 O (1) Thus, CO 2 dissolved in water promotes acid-catalyzed dissolution of the biomass by formation of carbonic acid. When formation of carbonic acid occurs, an increase in hydronium ion concentration is observed due to the dissociation of the unstable acid. This leads to a decrease in the ph value (slightly above 3), sufficient to promote dissolution and hydrolysis of hemicellulose 32 into its corresponding sugars and leading, at the same time, to high cellulose enzymatic digestibility. 60 Furthermore, the acidity of the medium does not constitute an environmental problem because the depressurization step removes CO 2, thus increasing the ph of the solution. Apart from the chemical hydrolysis based on the in situ acidic environment, the physical effect of CO 2 is also noticeable. This physical phenomenon is based on the fact that polysaccharides react with liquid hot water and the CO 2 acts as a catalyst, which under supercritical conditions leads to high diffusivities and promotes the swelling effect of biomass. 61,62 At the depressurization step, CO 2 can also act as a detoxification medium, and due to the solubility of some compounds in high-pressure CO 2, it can help to remove degradation products such as furans from the hydrolysate. 32,63 Furthermore, another complementary advantage of SCFs, including scco 2, is that the product streams are completely sterilized in regard to pathogens, viruses, and biotoxins Summarizing, scco 2 pretreatment with H 2 Ooffers the same benefits as acid hydrolysis without the typical drawbacks observed for sulfuric acid application Reaction Severity. To evaluate the effect of CO 2 on the reaction severity, where the ph of the system is indispensable, a combined severity (CS) factor was proposed: 60 CS = log( R0 ) ph (2) where R 0 is the severity factor. Both the solubility of CO 2 in water and the dissociation of H 2 CO 3 determine the hydronium ion concentration in the system. The solubility of CO 2 in water can be estimated using Henry s constant and depends on the temperature. 60 Furthermore, the dissociation of H 2 CO 3 in water also depends on the temperature, and its constant can be expressed by the following equation: pka = T T (3) where the pk a is the negative decimal logarithm value of the acid dissociation constant (K a ) and T is the temperature expressed in degrees Celsius. Taking into account the effect of the temperature, partial pressure of CO 2, solubility of CO 2 in water, and solubility of H 2 CO 3 in water, van Walsum et al. 60 proposed an equation which estimates the value of the ph in the binary CO 2 /water system: = 6 2 ph ( ) T T ln( p ) CO (4) 6

5 Chemical s where the temperature was in the range of C and the partial pressure of CO 2 was up to bar. To assess the combined severity factor of the reaction at a specific temperature, reaction time, and partial pressure of CO 2, the combined severity factor is depicted by the following equation: 6 2 CS = log( R ) ( T T p 0 CO ln( p ) 3.92 CO2 (5) where CS pco2 indicates the CS for reaction with CO 2, T is the temperature in degrees Celsius, and p CO2 is the partial pressure of CO 2 expressed in atmospheres. 60 The proposed combined severity factor reflects all the most important parameters affecting the sugar production yields in the high-pressure CO 2 / H 2 O process. On the other hand, it is important to highlight that the use of a severity factor should facilitate the comparison of results between process conditions. However, very often a lack of information about the ph value, process parameters such as the heating profile, and differences between substrates make the comparison of literature data a difficult task Corrosion in High-Pressure CO 2 /H 2 O Systems. Considering the high compressibility of high-density fluids, special attention should be given to the possible corrosion issues caused by the reaction environment. Corrosion has a significant influence on the safety of the high-pressure installation as well as on the economics of the process. There is no universal material which could be applied and which would withstand the reaction conditions across the existing diverse technologies. 67 The stainless steels are the most commonly used materials for reactor manufacturing due to the relative resistance against reaction corrosive conditions and because of their much lower price than that of other alloys such as Hastelloy, zirconium, or titanium. Considering the reactions carried out in high-pressure CO 2 /H 2 O systems, it is worth emphasizing that the rate of material corrosion is predominantly affected by the solubility of the protective oxides. However, the oxide solubility is highly influenced by the dissociation of acids and bases and also by the ph of the solutions thus formed. In high-pressure CO 2 /H 2 O systems one of the main issues is an in situ formation of carbonic acid according to eq The formation of carbonic acid depends on the CO 2 partial pressure and temperature, and consequently, the ph of the solution changes, affecting the corrosion rate. 69 It must not be forgotten that, during biomass processing, corrosion induced by addition of CO 2 is not a sole function of the CO 2 partial pressure and temperature but also of other chemicals formed during reaction, e.g., organic acids such as acetic or formic acids. Both acids are known as strong organic acids and are recognized as strong corrosive agents with a corrosion power superior to that of carbonic acid. 70 At certain conditions the corrosion force of acetic and formic acids can be compared even to the corrosive environment of sulfuric acid used often in acid hydrolysis SCOPE OF THE REVIEW The use of carbon dioxide under sub- or supercritical conditions in biomass processing is a field of research still in its infancy but under rapid development in recent years. The growing interest in the use of sub- or supercritical CO 2 as a versatile tool in biomass processing for the hydrolysis of biomass fractions, such as cellulose, starch, hemicellulose, lignin, and proteins, to produce various bio-based materials and chemicals is reflected in the number of publications available. The unique properties of CO 2 as a supercritical fluid open a wide variety of opportunities in the field of biorefinery, and it is expected that the use of scco 2 is going to grow even more in the near future. This review presents the current state of the art of biomass processing using carbon dioxide and its mixture with water for a wide range of raw materials. This work focuses on the role of supercritical CO 2 as a tool for pretreatment to enhance the enzymatic digestibility to convert lignocellulosic/starch biomass into biofuels and biochemicals. Especially, in the field of enzymatic hydrolysis, the saccharification of biomass in highpressure CO 2 to improve bioconversion to ethanol is addressed. It also addresses the effect of high-pressure CO 2 in the conversion of polysaccharides and proteins for the production of various biochemicals. Additionally, this review depicts the role of CO 2 in both the delignification process and conversion of lignin. Finally, critical issues and prospectives in the field of CO 2 -assisted hydrothermal technologies are also enhanced. The present review summarizes the current achievements in the aforementioned areas published in scientific periodicals up to the end of September SUPERCRITICAL CO 2 PRETREATMENT Supercritical carbon dioxide is mostly used as an extraction medium but is considered for other nonextractive applications, such as pretreatment of biomass. 40,71,72 This is mostly due to the size of CO 2 molecules being similar to that of both H 2 O and ammonia. Also, at high pressures CO 2 easily penetrates into the small pores of recalcitrant LCMs, resulting in structural changes and enhancing the glucan and xylan accessibility for enzymatic hydrolysis. In the case of CO 2 explosion, the quick release of CO 2 facilitates the disruption of cellulose structure, decreasing the degree of crystallinity, enhancing the permeability of cellulose, and increasing the accessibility of enzymes to an extended surface area. 73 Liu et al. performed a study on the capacity of high-pressure CO 2 to pretreat various lignocellulosic residues such as corncob, cornstalk, and rice straw. 74 It was shown that, to reach the maximal reducing sugar yield, the most important factors were the temperature and moisture content. However, they also highlighted the importance of CO 2 presence in biomass pretreatment for the aforementioned reasons. The interest in using enzymes to hydrolyze polysaccharides arises because the overall cost of enzymatic hydrolysis is lower than that of acid hydrolysis. This is mostly because enzymes are usually used at milder conditions (ph 4.8 and temperatures of C), thus causing less corrosion problems. 75 Furthermore, enzymes are considered as green catalysts since they show exceptional advantages, such as high chemo- and regioselectivity, minimizing the formation of coproducts, and can be reused without loss of activity. Therefore, in the past several years, a lot of works have been dedicated to demonstrating the ability of carbohydrate-degrading enzymes. However, the quick and efficient enzymatic hydrolysis of biomass using a low enzyme loading has proven to be one of the major technical and economical drawbacks in the overall bioconversion process for biofuel production and continues to be a challenge for biofuel commercialization. 76 Although the cellulose crystallinity and degree of polymerization (DP) are considered as the most important factors which determine the final rates of enzymatic hydrolysis, several literature reports showed that these factors alone do not explain 7

6 Chemical s the low rates of enzymatic hydrolysis. 77 For instance, for LCMs, a decrease of crystallinity accompanied by changes in the biomass morphology, such as particle size reduction and an increase of the available surface area, was observed. In addition, the removal of acetyl groups from hemicellulose reduces the steric hindrance of hydrolytic enzymes and greatly enhances polysaccharide digestibility. 78 The use of enzymes to catalyze cellulose conversion is considered as a greener alternative in comparison to acid hydrolysis, which mostly requires two-step hydrolysis. 79 Because of these benefits, the use of enzymes to convert cellulose into glucose is the subject of numerous reviews. 22,80 The high-pressure CO 2 /H 2 O process coupled with enzymatic hydrolysis is a very attractive approach for the development of green chemistry into a biorefinery concept. 51 Moreover, high-pressure CO 2 in the presence of water provides environmentally friendly pretreatment conditions for depolymerization of hemicellulose and enables the enzymatic process without the necessity for ph adjustment of the medium for the enzyme-catalyzed hydrolysis. 81 Less severe pretreatment is more economically favorable, though the conditions must be optimized since the mildness of the process leads to lower sugar release, and consequently, elevated amounts of enzymes are required to achieve high yields and thus increase the costs. Because of these benefits, a number of research groups have been actively studying green solvents, including ionic liquids and high-pressure CO 2, as innovative tools for lignocellulosic biomass processing. 36,82,83 For example, Zheng et al. explored the effect of high-pressure CO 2 on the rate and degree of cellulose hydrolysis in cellulose-rich materials such as Avicel (a commercial form of pure cellulose). 73 They concluded that the pretreatment with high-pressure CO 2 followed by quick release of CO 2 leads to disruption of cellulosic structures, thus increasing the accessible area of cellulose. Their results showed that high-pressure CO 2 is an effective pretreatment method for cellulose during which a pressure increase facilitates the penetration of CO 2 molecules into the crystalline cellulose and increases the rate of enzymatic hydrolysis, resulting in a glucose yield of approximately 50%. However, if pretreatment is conducted under subcritical conditions at 207 bar, then there are no significant changes observed in the digestibility of cellulose. The same Avicel sample treated with high-pressure CO 2 at 35 C showed a significant increase of glucose yield with increased pressure (76, 138, 207, and 276 bar). It is important to highlight that Avicel is pure cellulose; thus, it is much more susceptible to hydrolysis than recalcitrant lignocellulosic materials. To face this challenge, Zheng and co-workers also investigated the effect of high-pressure CO 2 on cellulosic-rich materials, such as recycled paper mix and sugarcane bagasse. They found maximal glucan and glucose yields of 75% and 22%. 73 Furthermore, an increase of glucose yield was found by changing the reaction conditions from 35 C at 69 bar to 80 C at 207 bar. However, in this case the biomass was subject to dilute-acid hydrolysis after high-pressure CO 2, enhancing the cellulose susceptibility to enzyme attack: therefore, the effect is not straightforward. 73 Dale and Moreira reported a great effect of CO 2 explosion at low temperature (25 C) and 56.2 bar on enzymatic hydrolysis of alfalfa (75% sugar yield). 84 Although a high glucose yield was obtained, this technology did not show a similar efficiency when compared to steam explosion or ammonia fiber expansion (AFEX) pretreatment. 84 However, in this case, the reduced effectiveness of enzymatic hydrolysis can be counterbalanced by the fact that in a process with carbonated water the need for an enzyme is 33% lower than when using dilute sulfuric acid pretreatment. 85 Additionally, it was found that high-pressure CO 2 can be compared to ionic liquid pretreatments, 28 30,36 being more effective than ammonia explosion without the formation of inhibitory compounds, which can be found after a steam explosion. 84 They proved that these fluids lead to an increased efficiency of the enzymatic hydrolysis, though still less than with CO 2. Hohlberg et al. reported an increase of digestibility of pine sawdust after CO 2 pretreatment, obtaining 32.5 g of reducing sugars/100 g of substrate in comparison to 26 g 100 g 1 for steam explosion after the substrate impregnation by SO However, the opposite effect was found for hemicellulose hydrolysis in which the pretreatment with SO 2 as the impregnating agent was more efficient than with CO 2. Better results were obtained with SO 2, although severer conditions were needed. 87 At these conditions, 68.3% hemicellulose removal with SO 2 was attained, in comparison to 40.5% attained from CO 2 pretreatment. However, it is important to stress that the pretreatment with CO 2 led to the production of a xylose-rich fraction, mainly in oligomer form, in contrast to pretreatment with SO 2, which acted more as an impregnating agent. Zhang and Wu demonstrated that biomass pretreatment with CO 2 at subcritical conditions has a positive influence on subsequent enzymatic hydrolysis of both eucalyptus chips and sugarcane bagasse. 88,89 They reported maximum glucan conversions of 92.2% and 93% (at 180 C and 50 bar for 80 min), respectively, which are higher than for native eucalyptus feedstock and sugarcane bagasse (46.8% and less than 10%, respectively). Matsushita et al. also found that the addition of CO 2 to hydrothermal technologies improved the saccharification rate of eucalyptus. 90 Kim and Hong examined the effect of CO 2 on aspen (hardwood) and yellow pine (softwood) at different temperatures ( C) and reaction pressures (214 and 276 bar) for min processes with different moisture contents (0 73%, w/w). 91 They showed that, for temperatures below 120 C and pressure up to 214 bar for 60 min, no significant effect on the sugar yield was observed. Even a pressure increase to 276 bar did not affect the yields of the sugars released. However, the same study demonstrated that temperature seems to have a more pronounced effect, because at 165 C bar a higher glucose yield was achieved. 91 It was also determined that this treatment is much more effective for aspen than for yellow pine, possibly because of structural differences between both raw materials 91 discussed in section Alinia et al. also investigated the effect of the residence pressure on wheat straw at constant temperature (190 C) after 30 min. 31 They concluded that the changes in pressure from 80 to 120 bar caused an increase of the yield of reducing sugars. However, a pressure above 120 bar did not change the final reducing sugar concentration. Also Gurgel et al. examined the effect of hot water coupled with high-pressure CO 2 pretreatment on sugarcane bagasse. 92 They used a flow-through supercritical extraction stainless steel reactor pressurized with CO 2 (68 bar and initial temperature of 25 C, liquid:solid ratio of 12:1). After high-pressure CO 2 /H 2 O pretreatment at 115 C for 60 min, a maximum cellulose conversion of 41.2% was achieved. It is important to underpin that all the studies presented above were performed in externally heated reactors without any stirring system. Thus, hot spots, especially near the reactor walls, and overheating may lead to local biomass pyrolysis. This incorrect processing may lead to the formation of degradation products from biomass subjected to high temperatures while at the same time not being pretreated homogeneously, affecting 8

7 Chemical s Table 2. scco 2 Coupled with Conventional Pretreatment Technologies reaction conditions glucose yield/reducing sugar yield (%) conditions of conventional technology moisture content (%) CO 2 technology technology raw material T ( C) P (bar) t (min) conventional technology both technologies untreated biomass ref Two-Step Pretreatment (scco 2 Followed by Conventional Technologies) ultrasound corncob khz, 600 W, / / / C, 6 h cornstalk khz, 600 W, a / / / C, 8 h sugarcane khz, 154 W, 65 /380 ± 9 b /350 b /300 b /127 ± 16 b 100 bagasse 30 C, 8 h sugarcane C, 4 h / 20.2/ 97.8/ 13.4/ 101 bagasse steam c wheat straw C, 15 min 23 / alkaline sugarcane % H 2 O 2, / 22.9/ 65.8/ 13.4/ 101 (H 2 O 2 /NaOH) bagasse 60 C, 9 h CO 2 -Assisted Conventional Technologies AFEX rice straw % NH d / e / 99.04/ 102 a CO 2 pretreatment at 170 C and 200 bar of CO 2 pressure for 1 h. b Units of g kg of dry biomass 1. c CO 2 explosion was performed after the steam explosion. d Conditions: 160 C, 50 min, 15 bar. e Conditions: 160 C, 50 min, 10% ammonia concentration. the process results. Additionally, the usefulness of the aforementioned works is reduced because the particle size of biomass used in these experiments, which was in the range of 0.5 and 2 mm, would require an additional mechanical treatment step. Recently, Morais et al. demonstrated that pressure changes from 0 (autohydrolysis) to 54 bar of initial CO 2 (using a liquid:solid ratio of 10:1) at constant temperature and nonisothermal conditions in wheat straw pretreatment promoted an increase of sugar yields. 51 The glucose yield rose to 82.21% at the aforementioned conditions, which corresponds to an increase of 26% versus the yield from the autohydrolysis reaction at the same conditions. Some experiments related to biomass pretreatment in a twostep temperature system were published. Luterbacher et al. used a biphasic CO 2 /H 2 O system in which two temperature stages were used to pretreat hardwood and switchgrass. 93 In this process, a short high-temperature stage at 210 C for 16 and 1 min for hardwood and switchgrass, respectively, followed by a long low-temperature stage at 160 C for 60 min was examined. 93 At a high solids loading (40 wt %), maximal glucose yields of 83% and 80% were achieved for hardwood and switchgrass and overall sugar yields were as high as 65% and 55%, respectively, for these materials. Compared with the pretreatment performed with only one temperature stage at 170 C for 60 min for which the maximal glucose yield was 73% for hardwood, 59 it can be concluded that, for a given material, either in an acidic or a neutral environment, the use of a two-temperature pretreatment (which prevents formation of unwanted products from hemicellulose, such as furfural) can be advantageous. 93 The hemicellulose depolymerization to produce oligosaccharides occurs quickly during the pretreatment. These oligosaccharides are hydrolyzed to monomers, which can degrade to undesirable products such as furfural. 94 With a single temperature stage it is difficult to simultaneously obtain both high glucan and high hemicellulose sugar yields. For hardwood, the hemicellulose sugar yields decreased from 180 to 190 C, with a contrary trend for the glucose yield. This tendency was also observed with a two-temperature stage, but the effect was less pronounced. 95 Thus, in the biorefinery concept, it may be advantageous to use high temperatures for a short period of time during the initial lag in formation of degradation products and then continue with lower temperatures for a longer time, which prevents monomer degradation. 93 This procedure was applied in other types of pretreatments, such as dilute-acid hydrolysis (190 C for 150 s with 1.1% acid and 210 C for 115 s with 2.5% acid) for barkrich biomass and aspen wood pretreated by autohydrolysis at 175 C The results confirmed that a two-stage process gives better results than a typical single-stage pretreatment. Another interesting application is the use of high-pressure CO 2 with water/ethanol as a cosolvent system in enzymatic hydrolysis of corn stover. 99 The use of water/ethanol as a cosolvent plays an important role in the yields of enzymatic hydrolysis for several reasons: (i) high-pressure CO 2 promotes a physical effect, which contributes to improvement of lignin cleavage; (ii) carbonic acid formation helps hemicellulose hydrolysis; (iii) water as a polar solvent can break down the bonds between hemicellulose and lignin structures, increasing the cellulose digestibility; (iv) CO 2 enhances the capacity of ethanol to dissolve lignin. All these factors produce an advantageous pretreatment in which potential inhibitors of cellulose hydrolysis are eliminated. For instance, an improvement of 14% in glucose yield was achieved when 300 ml of ethanol was added to the CO 2 /water system, and after a pretreatment at 150 bar and 180 C for 60 min, a maximal glucose yield of 77.8% after enzymatic hydrolysis was obtained. 99 Ferreira-Leitaõ et al. studied the steam pretreatment of sugarcane bagasse using CO 2 as an impregnating agent in terms of glucose and xylose yields after enzymatic hydrolysis and formation of degradation products. 87 The highest glucose yields of 86.6% and 97.2% (theoretical) were obtained for sugarcane bagasse and cane leaves, respectively. In some cases, the pretreatment of LCMs, such as southern yellow pine, even at very severe conditions shows glucose yields very similar to that for untreated biomass. 91 Thus, it can be stated that the glucose yield following enzymatic hydrolysis, apart from the aforementioned factors, is also dependent on the composition and structure of biomass as presented in section It is important to remember that other parameters, discussed in sections , such as the moisture of the raw material, biomass loading and particle size, and CO 2 / biomass loading, have a great impact on glucose yields. 9

8 Chemical s Table 3. Examples of the CO 2 Effect on Biomass Presented as Glucan Conversion Yields after Enzymatic Hydrolysis reaction conditions raw material T ( C) P (bar) t (min) moisture content (%) CO 2 :biomass ratio (w/w) solid loading (w/w) particle size (mm) glucan conversion yield (%) glucan conversion yield (%) from untreated biomass enzymatic hydrolysis conditions a ref rice straw :1 1:1 < ± 0.5 b 27.7 ± 0.5 b cellulase, 30 FPU g 1 (DS); β-glucosidase, 15 CBU g 1 (DS); 50 C; 150 rpm; 48 h eucalyptus chips :10 < cellulase, 20 FPU g 1 (DS); 50 C; 150 rpm; 60 h 88 wheat straw c 10:1 1:10 < cellulase, 60 FPU g 1 ; β-glucosidase, 64 p-npgu g 1 ;50 C; 200 rpm; 96 h guayule bagasse : cellulase, 0.92 FPU ml 1 ; β-glucosidase, 0.76 U ml 1 ; d 50 C; 200 rpm; 72 h switchgrass :1 40 wt % ± ± 0.4 cellulase, 15 FPU g 1 (DS); β-glucosidase, 30 CBU g 1 (DS); 50 C; 72 h / /16 11:1 40 wt % cellulase, 15 FPU g 1 (DS); β-glucosidase, 30 CBU g 1 (DS); 50 C; 72 h :1 < e 12 e cellulase, 50 U 100 mg 1 (DS); β-glucosidase, 20 U 100 mg 1 (DS); C; 24 h big bluestem :1 40 wt % ± 2 17 ± 1 cellulase, 15 FPU g 1 (DS); β-glucosidase, 30 CBU g 1 (DS); 50 C; 72 h 59 grass corn stover :1 < e 12 e cellulase, 50 U mg 1 (DS); β-glucosidase, 20 U mg 1 (DS); 47 C; 24 h 34 sugarcane bagasse sugarcane (leaves) sugarcane bagasse (pulp) sugarcane bagasse (skin) mixed hardwood mixed perennial grasses : ± 2 36 ± 1 cellulase, 15 FPU g 1 (DS); β-glucosidase, 30 CBU g 1 (DS); 50 C; 72 h > cellulase, 15 FPU g 1 (DS); 50 C; 72 h :1 1:12 < cellulase, 10 FPU g 1 (DS); β-glucosidase, 20 U g 1 (DS); 50 C; 150 rpm; 72 h 205 f cellulase, 65 FPU g 1 (DS); β-glucosidase, 376 U g 1 (DS); 40 C; 180 rpm; 96 h :15 < <10 cellulase, 20 FPU g 1 (DS); 50 C; 150 rpm; 72 h f cellulase, 15 FPU g 1 (DS); β-glucosidase, 18 U g 1 (DS); 40 C; rpm; 96 h cellulase, 70 FPU L 1 (CC); β-glucosidase, 250 CBU ml 1 (CC); 50 C; rpm; 8 h cellulase, 70 FPU L 1 (CC); β-glucosidase, 250 CBU ml 1 (CC); 50 C; 130 rpm; 8 h :1 20 wt % ± ± 0.3 cellulase, 15 FPU g 1 (DS); β-glucosidase, 30 CBU g 1 (DS); 50 C; 72 h :1 40 wt % ± 5 cellulase, 15 FPU g 1 (DS); β-glucosidase, 30 CBU g 1 (DS); 50 C; 72 h / /60 11:1 40 wt % cellulase, 15 FPU g 1 (DS); β-glucosidase, 30 CBU g 1 (DS); 50 C; 72 h :1 40 wt % ± ± 0.4 cellulase, 15 FPU g 1 (DS); β-glucosidase, 30 CBU g 1 (DS); 50 C; 72 h 59 a Abbreviations: FPU, filter paper unit; CBU, cellobiase unit; p-npgu, p-nitrophenyl glucoside unit; U, international unit (μmol (mg s) 1 ); DS, dry substrate; CC, cellulose content. b Glucose yield. c Nonisothermal conditions. d No data on the enzyme loading. e Glucose yield (g of glucose/100 g of dry biomass). f CO2 added to the pretreatment corresponds to 3% by weight on the basis of the water content of the material (30% of the dry matter)

9 Chemical s Table 4. Overview of the High-Pressure CO 2 Pretreatment Effect on Either Enzymatic Hydrolysis or Dilute-Acid Hydrolysis raw material reaction conditions T ( C) P (bar) t (min) moisture content (%) CO 2 :biomass ratio (w/w) solid loading (w/w) particle size (mm) reducing sugar yield (%) reducing sugar yield (%) from untreated biomass enzymatic hydrolysis conditions a High-Pressure CO 2 Pretreatment Followed by Enzymatic Hydrolysis aspen < ± ± 2.3 cellulase, 1000 U g 1 (DS); 50 C; 100 rpm; 72 h southern yellow pine sugarcane bagasse < ± ± 2.7 cellulase, 1000 U g 1 (DS); 50 C; 100 rpm; 72 h b; 50 C; 100 rpm; 8 h 100 cornstalk cellulase, 4000 U g 1 (DS); C; 100 rpm; 72 h corncob cellulase, 4000 U g 1 (DS); 50 C; 100 rpm; 72 h 62 wheat straw : cellulase, 15 FPU g 1 (DS); 50 C; 240 rpm; 72 h 31 High-Pressure CO 2 Pretreatment Followed by Dilute-Acid Hydrolysis corncob :1 < wt % H 2 SO 4 ; 160 C; 40 min 74 cornstalk :1 < wt % H 2 SO 4 ; 160 C; 40 min 74 rice straw :1 < wt % H 2 SO 4 ; 160 C; 40 min 74 a Abbreviations: U, international unit (μmol (mg s) 1 ); DS, dry substrate; FPU, filter paper unit. b No data on enzyme activity and enzyme loading. ref Combined pretreatments exhibited a synergistic effect on carbohydrate recovery and enzymatic conversion of sugars. This could be considered as an option in the field of biomass pretreatments, as shown in Table 2. Additional research into the influence of CO 2, coupled with other technologies aiming to improve the biomass pretreatment efficiency, was performed by Benazzi et al. 100 The Plackett Burmann design was employed to describe the influence of each process variable in CO 2 pretreatment on enzymatic hydrolysis. The results showed that, when high-pressure CO 2 is used, the amount of fermentable sugars obtained was increased by 280%, which corresponds to a hydrolysis efficiency as high as 74.2% compared to that of nontreated bagasse. The addition of high-pressure CO 2 to ultrasound technology enhanced by about 16% the amount of fermentable sugars compared to the ultrasound process alone. 100 Another example of CO 2 combined with ultrasound, for corncob and cornstalk pretreatment, was proposed by Yin et al. 62 This group reported that the application of high-pressure CO 2 enhanced the enzymatic hydrolysis of corncob and cornstalk by 50% and 29.8% compared to that of the untreated materials, respectively. At the same time, high-pressure CO 2 and ultrasound pretreatment increased the enzymatic hydrolysis yield by 75% and 13.4%, respectively. Recently, Phan and Tan compared the effect of a sequential combination of CO 2 explosion and alkaline hydrogen peroxide (H 2 O 2 ) with that of CO 2 explosion followed by ultrasound to pretreat sugarcane bagasse. 101 They demonstrated that CO 2 explosion (187 C, 156 bar, and 40 min retention time) followed by 1% H 2 O 2 pretreatment resulted in an enhancement of the maximal glucose yield by almost 50% (from 65.8% to 97.8%) compared to high-pressure CO 2 followed by ultrasound (the same conditions referred to above plus 4 h of ultrasound). High-pressure CO 2 promoted an increase of the accessible biomass surface that was susceptible to H 2 O 2 contact. At alkaline conditions there was formation of hydroxyl radicals and superoxide anion radicals from H 2 O 2, which could oxidize and degrade the lignin. This enhanced the hydrolysis of hemicellulose and lignin dissolution, increasing the digestibility of the remaining cellulose. When each pretreatment was individually applied, high-pressure CO 2 was demonstrated to be the best method in which a maximal glucose yield of 61.3% was obtained, contrary to the low glucose yield obtained individually by H 2 O 2 (22.9%) and ultrasound (20.2%). 101 Cha et al. investigated the effect of CO 2 - added ammonia explosion on the glucan conversion from rice straw in comparison to that of AFEX. 102 Both methods when used separately demonstrated an increase in the sugar conversion yield. Ammonia promoted both swelling of lignocellulosic biomass and removal of lignin, while highpressure CO 2 promoted pore opening in the lignocellulosic complex. The addition of high-pressure CO 2 to ammonia pretreatment demonstrated an increase of the maximal glucose yield to 99.04%; however, this increase was insignificant mostly due to the very low CO 2 pressure used (7.5 bar). Alinia et al. combined steam explosion and CO 2 explosion. The best conditions (200 C and 15 min for the first stage, steam injection and supercritical CO 2 explosion at 120 bar and 190 C for a 60 min retention time in the second stage) gave a reducing sugar yield as high as 60%. Another approach was examined by Moscon et al., who integrated enzymatic hydrolysis under high-pressure CO 2. They examined rice hulls subjected to this treatment in comparison to those subjected to integrated enzymatic hydrolysis in an ultrasound bath. 103 They found that the addition of high-pressure CO 2 as a cosolvent does not have a positive effect since the maximum yield of fermentable sugars obtained from traditional enzymatic hydrolysis was about 16 g kg 1 versus 4.2 g kg 1 obtained when high-pressure CO 2 was added to the saccharification process. Thus, observing data presented in Tables 3 and 4, it can be stated that the promising results obtained for CO 2 -mediated enzymatic conversions presented in the literature should encourage research in this field, mostly due to lower temperature requirements and lack of hazardous solvents Process Condition Effects Water Influence. One of the most important requirements for the efficient utilization of lignocellulosic materials is removal of hemicellulose to provide a feasible cellulose-rich cake for further enzymatic hydrolysis and fermentation to get, for example, ethanol. 20,107 To obtain 11

10 Chemical s high glucose yields from enzymatic hydrolysis, a sufficient humidity of LCMs is required. Indeed, in the presence of water, high-pressure CO 2 facilitates a swelling effect on biomass. Stamenic et al. showed that the structure of biomass and operating conditions such as the pressure, temperature, and time influence the degree of biomass swelling. 61 Narayanaswamy et al. found that for a wet corn stover (75%, w/w) at 240 bar and 120 C for 60 min, the glucose yield was doubled in comparison to that obtained for dry biomass. 34 These results are consistent with the findings for high-pressure CO 2 pretreatment for dry Avicel. 73 The intramolecular hydrogen bonds present in biomass decreased either the solubility in water or the reactivity of the functional groups of the solvents. As a result, the development of any effective and environmentally benign recycling process for cellulose did not appear to be possible as long as it could be swollen or dissolved only in those solvents. On the other hand, as previously described, the presence of CO 2 acidifies the aqueous environment, making the process similar to dilute-acid hydrolysis. In this context, Kim et al. reported no differences in sugar yields between dry pretreated biomass and an untreated (control) aspen sample. 91 The increase of the moisture content in aspen resulted in a general increase of the sugar yield obtained in enzymatic hydrolysis. 91 Alinia et al. investigated the effect of CO 2 pretreatment of dry and wet wheat straw by varying the temperature ( C), pressure (80, 100, 120, 150, 170, and 200 bar), and residence time (10, 30, 60, and 70 min). 31 They found that the high-pressure CO 2 pretreatment with water led to the best overall yield for sugar (208.4 g kg of wheat straw 1 ), which is more than 50 g kg 1 higher than for dry wheat straw (149.1 g kg 1 ). The high solid loadings of substrates are desirable since the sugar-rich hydrolysate is essential for the profitability of the process for a wide variety of products, including bioethanol. Nevertheless, due to the high viscosity of most lignocellulosic materials, it is difficult to pretreat with a solid content higher than 10%. 108 The lignocellulosic substrates have a high viscosity because of the presence of considerable amounts of insoluble compounds and the water binding capacity of hemicellulose in the material. Luterbacher et al. tested a biphasic CO 2 /H 2 O mixture in several biomass sources with a high solids content (40 wt %) over a range of temperatures between 150 and 250 C and residence times from 20 s to 60 min at 200 bar of CO Hardwood with a solids content of 20% (by weight) at 170 C, 60 min, and 200 bar of CO 2 resulted in a 73% glucose yield compared to 77% achieved for the same biomass with 40% solids under equivalent conditions. This negligible increase of the glucose yield shows that there is no clear evidence for a biomass loading effect on the sugar release. On the other hand, they found that a high solids content is less sensitive to temperature changes, which could be explained by a higher inertia of biomass as well as heat and mass transfer limitations caused by the high solids content. Additionally, for a higher biomass loading, a significant decrease of the hemicellulose sugar yield in the hydrolysate was accompanied by increased hemicellulose degradation product yields Biomass Sources and Recalcitrance. The composition and concentration of the fractions vary, depending on the biomass source, e.g., grass, hardwood, or softwood. Furthermore, these factors are dependent on external parameters. The most important ones are climacteric conditions, origin, age, harvesting method, drying, and storage of biomass. In addition, the recalcitrant structure of biomass is influenced by inherent properties, such as the lignin content, the cellulose crystallinity and DP, cellulose accessibility to enzymes, sheathing of cellulose by hemicellulose and lignin, and the fiber strength. All these factors influence the enzymatic digestibility of a given biomass feedstock. It cannot be forgotten that a major aim of biomass pretreatment with high-pressure CO 2 is not to achieve a maximal glucose yield but rather to obtain satisfactory glucose yields at energy inputs as low as possible. Thus, all these aforementioned parameters must be assessed before selection of the operation pretreatment conditions. In addition, the physical properties of biomass are important; for example, softwoods are generally characterized by a lower density than that of hardwoods, which facilitates high-pressure CO 2 pretreatment. Different untreated biomass species, when subjected to enzymatic hydrolysis under the same conditions, give substantially different saccharification yields. For instance, yields on enzymatic digestion of untreated switchgrass and corn stover were 10.4% and 36%, respectively. 59 On the other hand, some biomasses are more susceptible to processing and, independently of the process conditions, give higher sugar yields. As an example, the glucose yields of switchgrass at 160 or 170 C and 200 bar after 60 min in binary CO 2 /H 2 O pretreatment were always higher than for corn stover, big bluestem, and mixed perennial grasses. 59 Kim et al. compared the enzymatic digestibility of two different biomasses (aspen wood and southern yellow pine) pretreated at 165 C and 214 bar for 30 min with a moisture content of 75%. 91 The highest glucose yields were 84.7% and 27.3% for aspen and southern yellow pine, respectively. The more rigid and denser structure of yellow pine leads to more difficult access of CO 2, and additionally, the large moisture content negatively influences the pretreatment, which results in much lower enzymatic digestibility of this biomass. Thus, biomass species characterized by a low density are desirable to achieve good sugar yields while avoiding pretreatments at high temperature and pressure. Santos et al. studied the highpressure CO 2 explosion of two different samples of sugarcane (bagasse pulp and skin), which were compared to Avicel (pure cellulose). 106 With these sugarcane samples, different glucose yields were obtained; for instance, at 60 C and 140 bar for 5 min the best glucose yields were 76.10% and 28.39% for sugarcane skin and bagasse, respectively. However, the best glucose yield (72.0%) for sugarcane bagasse was obtained at 60 C and 140 bar for 60 min. This result shows that the residence time of pretreatment is an important variable. 106 Puri et al. studied the effect of steam and CO 2 under supercritical conditions on cellulose hydrolysis before release of CO The operating conditions used were 200 C and between 34.5 and 138 bar of CO 2 pressure. The highest glucose yields were 81%, 78%, and 75% for wheat straw, bagasse, and Eucalyptus regnans woodchips, respectively. Due to the structural and chemical composition differences between feedstocks, the times required to obtain the maximal digestibility rate were 5 min for wheat straw and bagasse and 15 min for E. regnans CO 2 :Biomass Ratio. One of the most important factors influencing the effectiveness of the pretreatment with high-pressure CO 2 is the number of moles of CO 2 and mass of biomass present in the reactor. The quantity of CO 2 for a known volume of reactor containing biomass is greatly dependent on the thermodynamics at the set point plus the pretreatment pressure and temperature. Unfortunately, few research reports provide sufficient data for comparison. Narayanaswamy et al. used a CO 2 :dry biomass ratio close to 12

11 Chemical s treated solids suggests that hemicellulose removal has a greater impact than lignin removal. 14:1 (w/w) for high-pressure CO2 pretreatment and obtained a high glucose yield of 85% for corn stover.34 Also Walker and co-workers obtained the same glucose yield under the same conditions but at lower CO2:dry biomass ratio of 8:1 (w/w).59 The ratio of the treatment medium (in this case CO2) to biomass is lower than those for other pretreatment methods such as dilute-acid hydrolysis, autohydrolysis, or aqueous ammonia. This shows that, even with a relatively low quantity of CO2, a satisfactory glucose yield can be achieved due to the impregnation of the biomass with CO2 at supercritical conditions. 4. ENZYMATIC HYDROLYSIS IN SUPERCRITICAL CO2 Supercritical carbon dioxide is a promising green solvent for many enzyme-catalyzed chemical reactions. The finding of enhanced enzyme stability and activity in alternative solvents, e.g., with lipases,111 suggested a wide range of high-pressure CO2 applications as selective biocatalysts for synthesis and/or hydrolysis reactions. Recently, the number of studies involving enzymatic catalysis under CO2 at supercritical conditions as a solvent have increased, since it has been demonstrated that enzymes are stable and active in high-pressure CO2 and that CO2 favors transport properties, accelerating mass transfer in enzymatic reactions.112 At the same time, high-pressure CO2 was also successfully used as a medium for the recovery of products and reactants.113 The most important factor in achieving high enzymatic hydrolysis rates is the stability of the enzymes. This in turn depends on the pressure and temperature exerted, the ph, the number of depressurization steps, the water content, carbonic acid formation, the solid content, the enzyme species, and the substrates and types of end products due to the inhibition of enzymatic processes by blocking the active sites of the enzyme.114 However, the use of CO2 at supercritical conditions is not challenging from the enzyme point of view, because either the temperature or pressure is within the operational range of most of the enzymes. However, enzyme activity is more difficult to maintain than for other typical solvents, mostly caused by lowering of the ph at higher CO2 pressures. The CO2 pressure itself, even in the range of 200 bar, does not have an effect on enzyme denaturation, but the ph generated by this exerted pressure may inactivate the enzyme. Additionally, the high substrate and product concentrations contribute to lower enzyme effectiveness, which can be solved by a continuous mode of substrate feeding and removal of products by CO2 extraction. In this way, the inhibitory effect can be minimized or even eliminated. Xiao et al. investigated the effect of sugar inhibition of glucose on cellulase and β-glucosidase hydrolysis of both Avicel and acid-pretreated softwood.115 They found that the increase of the glucose concentration in the hydrolysate resulted in a dramatic decrease of both cellulase and glucosidase activities.115 Under the severe pretreatment conditions examined, the three-dimensional structure of the enzymes may be irreversibly changed by denaturation, resulting in loss of enzyme activity. With less severe conditions, the proteins may recover their structures and only minor changes are observed, which may lead to altered enzyme activity and stability. The water content in the system is also an important factor that influences the activity of the enzyme. To maintain the enzyme activity, a specific water activity is required, affecting diffusion and thus influencing both the reaction equilibrium and enzyme structure by noncovalent binding or disruption of hydrogen bonds.116 However, in a CO2/H2O system, the enzymes can change their activity and stability with formation of carbonic acid due to a decrease of the ph in the environmental media. Zheng et al. reported that the presence of small amounts of water improves the reaction rate.117 However, the addition of water in quantities higher than necessary to form a microaqueous environment led to a decrease in rate (Figure 5). Additionally, a high quantity of water results in the formation of an aqueous layer close to the active site, which inhibits the 3.2. Effect of High-Pressure CO2 on the Morphology of Lignocellulosic Biomass The morphological changes in biomass after pretreatment with high-pressure CO2 can be assessed by scanning electron microscopy (SEM).34,51,104 Figure 4 shows the structure of Figure 4. Scanning electron microscopy pictures of rice straw samples untreated (a) and treated at 110 C and 300 bar for 30 min with a liquid to solid ratio of 1 g g 1 (b). Reprinted with permission from ref 104. Copyright 2010 Elsevier. SEM images of corn stover samples with 75% moisture before (c) and after (d) pretreatment at 150 C and bar for 60 min. Reprinted with permission from ref 34. Copyright 2011 Elsevier. untreated rice straw biomass (a) and morphological changes observed after pretreatment (b) at 110 C and 300 bar for 30 min at a liquid:solid ratio of 1 g g 1, plus samples of corn stover with 75% moisture before (c) and after (d) pretreatment at 150 C and bar for 60 min. The untreated samples exhibited a rigid, tight, and contiguous surface, while pretreated samples had fibrous characteristics with anomalous porosity and the lamellar structures becoming fleecy. The fibers seem to be separated from the initial structure and totally exposed, making the surface more susceptible to cellulase enzyme attack.96 Some pretreatments facilitated hydrolysis, but at the same time, an increase in cellulose crystallinity was observed. This may be explained by the reduction or even total removal of amorphous cellulose after pretreatments.110 Unlike the results with crystalline cellulose, the decrease of polymerization in pre13

12 Chemical s Figure 5. Effect of the water content on the relative initial rate of Avicel hydrolysis in supercritical CO 2. Reprinted with permission from ref 117. Copyright 1996 Springer Science and Business Media. Springer Science + Business Media makes no representations or warranties with respect to the licensed material. transport of reactants and products from the CO 2 -rich phase to the enzyme-containing phase. 117 The enzyme inactivation is predominantly attributed to the low ph formed during hydrolysis in the presence of CO 2. Russel and co-workers determined that the ph of water in CO 2 in the absence of buffer is around 3.0 at 100 bar. 118 As the concentration of hydronium ions in water is dependent on dissociation of carbonic acid, a simple method to control the environmental ph to avoid this dissociation is an addition of sodium acetate buffer. The addition of this buffer shifts the dissociation equilibrium toward the nondissociated form, resulting in a decrease of the concentration of acidic hydronium ions. 119 Also, Ziegler et al. 120 found that even with buffered CO 2 /H 2 O systems a change in the temperature and pressure of the system resulted in a change of the aqueous solution ph. They reported that the addition of NaOH together with an adjustment of the CO 2 pressure over a range of 400 bar may change the ph value by more than 1.5. Enzymes have attracted special attention as green catalysts due to their high turnover numbers and acceleration of reaction rates. Enzymes are usually used as a specialized form of heterogeneous catalyst since they are amenable to a wide variety of consecutive processes, after which they can be recovered and reused. Furthermore, immobilized enzymes can enable the formation of products with high selectivity in both stereochemical and structural terms under relatively mild conditions. Because of the relatively low activity of crude enzymes in scco 2, many attempts have been made to stabilize the enzyme by modifying the form in which they are used. Enzymes in several forms (powder, reversed micelles, etc.) or immobilized in a carrier such as a sol gel matrix or a resin 121 can be used in the presence of supercritical fluids without significant losses of their properties. Immobilized enzymes are expected to be more stable in sub- and supercritical media and retain their activities. Park et al. reported that the reaction rate was improved and reached 100% at 160 bar and 50 C after 90 min in CO Unlike that at supercritical conditions, hydrolysis at ambient pressure increased the glucose yield at temperatures up to 50 C and then started to decrease on further heating. The glucose yield showed a maximum at 50 C, at which the cellulase activity was the highest. Encapsulation is another method of enzyme stabilization. Cellulase from Humicola insolens was encapsulated in hydrophilic silica aerogels by a sol gel procedure and scco 2 drying. Immobilized cellulase was used for the hydrolysis of carboxymethyl cellulose at atmospheric pressure in a binary CO 2 /water system at 100 bar. 123 The activity of immobilized cellulase at atmospheric pressure was higher than the activity of the nonimmobilized enzyme. Under CO 2 at supercritical conditions, the activity of the immobilized enzyme increased by 461%, while the residual activity of crude enzyme at the same conditions was only 101.7%. Furthermore, the immobilized enzyme was used more than 20 times without loss of activity. The explanation for the high activity of the immobilized enzyme under scco 2 is the high dispersion of enzyme in the silica aerogel matrix, enhancing the access of the bulky substrate to the enzyme. Thermal stability studies for both crude and immobilized cellulase in scco 2 were performed by Paljevac et al. They found that the activity of preincubated immobilized cellulase at 110 C was 172% higher than the activity of crude cellulase. 123 Muratov et al. examined the effect of high-pressure CO 2 on hydrolysis of boll fibers of cotton with cellulase from Trichoderma viride, Trichoderma reesei, and Aspergillus niger. 124 Cellulase activities were examined in scco 2, and the obtained results showed that all enzymes are stable up to 120 bar and 50 C after 48 h. Additionally, the yields of glucose in scco 2 were found to be always higher (approximately 1.2 times) than at atmospheric pressure. The degree of conversion under these conditions varied from 33.9% (A. niger) to 92.3% (T. viride). 124 Starch is a polysaccharide which is also present in the lignocellulosic biomass in considerable quantities and has great potential as a carbon source that can be converted analogously to cellulose. The hydrolysis of starch produces D-glucose and D- glucose degradation products. 16 The enzymatic hydrolysis of starch under high-pressure CO 2 proved to be a great technique to produce glucose. 126 Nevertheless, the macromolecular structure of starch cannot be directly converted to ethanol through conventional technologies such as fermentation processes, due to the need to first release the glucose monomers. Starch enzymatic hydrolysis starts with the action of amylases that liberates maltodextrin, oligosaccharides that are further hydrolyzed by enzymes such as pullulanases and glucoamylases. 127 The schematic representation of the enzymatic hydrolysis of starch is shown in Figure 6. As with cellulases, amylases and glucoamylases are also active and stable under scco 2. For instance, Lee et al. reported that the stability of amylase was sustained for 48 h at pressure up to 93 bar and at 50 C inco 2, showing a low denaturation effect. 126 For comparison, the enzyme activity decreases by over 30% under scco 2 compared to supercritical nitrogen. 126 Moreover, at low ph, structural changes to the enzyme and consequently changes to its activity and stability were observed. The activity of certain enzymes showed the maximum yield of glucose at close to 50 C, since the pressure is the most important factor in application of supercritical fluids. They also found that starch hydrolysis is influenced not only by enzyme activity, but also by the mass transfer limitation due to the heterogeneous nature of the solid liquid system. Senyay-Oncel and Yesil-Celiktas investigated the ability of CO 2 to change the activity of immobilized α-amylase after consecutive reactions. 128 The α-amylase from Aspergillus oryzae treated with scco 2 (at optimal conditions of 41 C, 240 bar, and 4 g min 1 of CO 2 flow for 150 min) and immobilized in 14

13 Chemical s Figure 6. Schematic presentation of starch hydrolysis. Reprinted with permission from ref 125. Copyright 1997 Biochemical Society. calcium alginate beads and NaY zeolite yielded 67.7% higher activity (29.7 mmol ml 1 min 1 ) than the untreated enzyme (maximum of 17.7 mmol ml 1 min 1 ). 129 The immobilized enzyme was reused in starch hydrolysis by consecutive activation by scco 2. This procedure was repeated until the activity of the engaged enzyme was no longer re-established. The best results were achieved when α-amylase was immobilized in NaY zeolite and the enzyme was used in 17 reactions and reactivated 4 times HIGH-PRESSURE CO 2 /H 2 O CONVERSION TECHNOLOGIES 5.1. Hydrolysis of Biomass-Derived Carbohydrates Hydrothermal treatments were studied for the production of hemicellulose-derived sugars in either oligomeric or monomeric form. The monosaccharides thus produced can be converted into a wide variety of products such as biofuels (either cellulosic ethanol or other furanic fuels), 20,107 chemicals, and biomaterials. Degradation products are also produced during hydrolysis, and they may inhibit the fermentative processes. ph lowering of the medium to slightly above 3 but with enough CO 2 to facilitate biomass hydrolysis was investigated as an alternative technology for hydrolysis of biomass. 119,130 The solubility of CO 2 in H 2 O determines the ph of the medium, and unlike acid-hydrolysis pretreatment, the acidity of the medium produced by the high-pressure CO 2 process does not represent an environmental problem, as after depressurization the ph becomes solely dependent on the other compounds in the solution. Therefore, the ph of the solution mixture can be controlled by the degree of dissolution of CO 2 in water, which, on the other hand, is dependent on the temperature and pressure. The CO 2 /H 2 O approach is a methodology that offers the benefits of acid catalysis without the typical drawbacks of sulfuric acid such as corrosion problems and/or the need for employment of separation methods Hemicellulose. Due to a lack of repeating β-1,4- glycoside bonds, hemicellulose does not present the crystalline nor the resistant structure of cellulose. 131 Thus, hemicellulose is a polymer most susceptible to hydrothermal treatments. Liu et al. reported a maximal hemicellulose yield of 99% as monomeric sugars at 220 C after a 15 min process. 132 The addition of CO 2 to the hydrothermal reaction led to the use of lower temperatures and shorter reaction times. van Walsum et al. observed that at elevated temperatures (above 200 C) carbonic acid has a catalytic effect on pure xylan hydrolysis, allowing an increase of pentose release and a decrease of the DP of xylan oligomers when compared to those with autohydrolysis pretreatment. 60 However, McWilliams et al. compared the hydrolysis of aspen wood under high-pressure CO 2 at C to a water-only reaction and found that addition of CO 2 and consequent formation of carbonic acid improved neither the xylose yield nor the additional formation of furan compounds. 133 These results are not surprising because aspen wood contains highly acetylated hemicellulose; therefore, for water-only treatments at temperatures 170 C hydrolysis of these bonds occurs by hydronium ion formation originating from water autoionization, leading to a sufficiently strong catalytic effect without the need for CO 2 -generated carbonic acid. 135 McWilliams et al. also found that biomass processing with CO 2 produces a hydrolysate with a higher ph in comparison to the autohydrolysis experiments. 133 In fact, these data can be elucidated by either a reduced CO 2 pressure (55 bar) used in the experiments or the low solubility of CO 2 in the aqueous phase at the examined conditions. The literature studies report that, up to 100 C, the CO 2 solubility in water decreases with temperature and increases with pressure. 136 Therefore, experiments should be done at as a low temperature as possible and at relatively high CO 2 pressures to obtain a high solubility of CO 2 in water and, consequently, a more pronounced effect of CO 2 on xylan hydrolysis. Chuang and Johanssen reported that carbonic acid has a great effect on the ph of aqueous solutions. 134 Figure 7 presents the effect of both the temperature and pressure of CO 2 on the ph of carbonated water reported by Chuang et al. On the other hand, the data reported by McWilliam et al. 133 can also be explained by the fact that ph measurement was done after the experiments were finished. The depressurization of the reaction mixture removes the greater part of CO 2 ; however, the produced mixture is under nonequilibrium conditions, and a prolonged time is needed to attain equilibrium, which is achieved when the concentration of CO 2 in the liquor is equal to the partial pressure in the surrounding atmosphere. 32 Other studies showed the hydrolytic effect of CO 2 on a wide range of polysaccharides under hydrothermal conditions. 137 Miyazawa and Funazukuri found that high-pressure CO 2 hydrolyzed the polysaccharides into low molecular weight sugars. In the absence of CO 2, the final monosaccharide yields from xylan were less than 5%, while, in the presence of CO 2, a significant yield improvement was reached. Additionally, they reported that the CO 2 -assisted hydrothermal reaction did not produce degradation products in such amounts as acid-catalyzed hydrolysis. Recently, Gurgel et al. reported the benefits of liquid hot water associated with high-pressure CO 2 instead of a water-only reaction. 92 They 15

14 Chemical s Figure 7. ph of water in equilibrium with CO 2 for different temperatures measured by UV/vis spectrophotometry (open symbols represent the predicted values, while solid symbols represent the experimental values). Reprinted with permission from ref 134. Copyright 2009 International Society for Advancement of Supercritical Fluids (ISASF). studied the extraction of hemicellulose and, at the same time, the production of D-xylose from sugarcane bagasse. The maximal xylose concentration obtained at 115 C and an initial CO 2 of 68 bar (initial temperature of 25 C) for 60 min was 9.8 wt %. King et al. reported that the employment of high-pressure CO 2 /H 2 O (up to 450 bar) to pretreat both corn stover and switchgrass yielded between 9% and 13% less sugars than dilute-acid hydrolysis and from 6% to 10% less cellulosederived sugars after enzymatic hydrolysis of oligomers initially present in the hydrolysates. 85 Zhang and co-workers processed raw cornstalk using various methodologies, namely, ammonia and H 2 O 2, butanediol, NaOH, H 2 O 2, ammonia, 50% ethanol, hot water, hot limewater, and high-pressure CO 2. Among the methods examined, high-pressure CO 2 and butanediol led to 77% total hemicellulose removal from cornstalk, with a minimal coextraction of lignin in the case of high-pressure CO When hydrothermal technologies were used, a substantial part of hemicellulose was converted to oligomers and a negligible quantity was converted into monomers. 139 Nevertheless, Zhang and Wu reported superior production of xylose monomers during treatment of eucalyptus chips with subcritical CO 2 (9.8 g 100 g of feedstock 1, which represents 69.1% conversion of xylan in raw material to xylose). 88 The subcritical CO 2 pretreatments were conducted with 30 g of feedstock and 300 ml of deionized water, which actually means that the experiments were CO 2 -assisted water reactions. Therefore, it can be expected that this technology results in liquors rich in hemicellulose-derived sugars, mainly in oligomer form since conversion of hemicellulose to oligomers is much faster than their further conversion to xylose. 40,50 To optimize the production of oligomers, moderate temperatures should be applied. Relvas et al. developed models for modeling the hydrolysis kinetics for hemicellulose-derived wheat straw under high-pressure CO 2 /H 2 O technology. 50 Modeling demonstrated that addition of CO 2 to a water-only reaction, as well as high pressures of CO 2, plays an important role in the kinetics of hemicellulose hydrolysis, being the fastest step in hydrolysis to oligomers. Even a small addition of CO 2 (20 bar of initial CO 2 pressure) improves the kinetic constant of this reaction by almost 40% in comparison to that of the water-only process. 50 Other work from this group also showed that the CO 2 /H 2 O system at 210 C and an initial pressure of 60 bar (2.62 mol of CO 2 ) gave a xylose yield as high as 94%. 32 This condition also allowed a noteworthy improvement in the yield of xylose oligomers in comparison to autohydrolysis. 27 Zhang and Wu studied the effects of subcritical CO 2 pretreatment of sugarcane bagasse with 15% (w/v) solids on the pentose yield. The maximum xylose yield (160 C and 50 bar for 80 min of residence time) obtained was g 100 g of raw material 1, in which 45.15% of xylose could be recovered as xylooligosaccharides. It is important to highlight that temperature is the principal factor influencing the production of organic acids. In this respect, Van Walsum et al. reported that the CO 2 addition inhibits the formation of organic acids. 140 In addition, the depressurization step can act as a detoxification method, due to the solubility of some organic compounds in high-pressure CO In this way, some toxic compounds for microbial metabolism, such as furan derivates from the hydrolysate, can be removed from the post-pretreatment medium. Bogel- Łukasik and co-workers found that volatile compounds from hemicellulose hydrolysis at 210 C and 60 bar (2.62 mol of CO 2 ) remain in the gas phase, and during depressurization 68% of total furfural was extracted. 32 Following this approach, it can be proposed that the highpressure CO 2 technology would consist of dissolution of hemicellulose and the simultaneous extraction of furfural and other volatile compounds. Gairola and Smirnova studied this process, examining the production of furfural from D-xylose and hemicellulose and the simultaneous furfural extraction with high-pressure CO The maximal furfural yield obtained from xylose at 230 C and 120 bar of CO 2 was 68%. Additionally, they found that kinetic models for production of furfural from conversion of L-arabinose were similar to those for furfural production from D-xylose, and high-pressure CO 2 extraction led to high furfural yields when both the xylose and furfural concentrations were kept low in the reactor. Supercritical CO 2 as a tool for extraction of furfural from the reaction mixture proved to greatly improve the selectivity and the yield of furfural, owing to the suppression of its side reactions. 143 Sangarunlert et al. optimized the production of furfural from rice husks by acid hydrolysis coupled with highpressure CO 2 extraction and reported an almost 90% yield Cellulose. Cellulose, in the presence of water, is a polymer very resistant to swelling as well as to enzymatic attack. Water at high pressure and temperature is able to hydrolyze the hydrogen-bond-linked structure of cellulose and its glycosidic bonds into glucose monomers. However, due to the severe conditions required, glucose itself can be a subject of undesired further hydrolysis into degradation products such as 5-HMF. 145 Many researchers have taken an interest in the hydrolysis of cellulose to glucose as a model with intriguing decomposition behavior in high-pressure technology. Celluloses from different biomass sources have different chemical and physical properties that can affect their hydrolysis behavior. Several literature reports about the cellulose conversion in sub- and supercritical water indicate the decomposition of glucose occurring rapidly with an increase of the severity of the reaction conditions. Therefore, processes with water under subcritical conditions are more appropriate to produce glucose monomers, due to the higher density of the medium. 146 Sasaki et al. suggested that the cellulose hydrolysis rate in subcritical water is much slower than in supercritical water, hindering a high glucose yield. 147 Studies 16

15 Chemical s Scheme 1. Formation of 5-HMF in an Acid-Catalyzed Reaction of glucose hydrolysis over a wide range of temperatures, including for the use of supercritical water, depict the formation of various products according to the temperature used. In addition, the subcritical water technology is not able to separate lignin from the cellulose, which prevents a high production yield of fermentable sugars. This problem can be resolved by the introduction of CO 2 into water-only processes, as this leads to higher acidity of the reaction medium, creating a greater concentration of glucose monomers at lower temperatures. To understand the behavior of glucose production in CO 2 -assisted hydrothermal technology, kinetic studies were performed for the decomposition of cellulose into oligomers and monomers. For example, Schacht et al. demonstrated the enhancement of cellulose dissolution in CO 2 -saturated water in comparison to pure water over a temperature range of C. 40 For temperatures higher than 260 C, no further enhancement of the hydrolysis rate was obtained due to the ph drop in the medium, which leads to a less pronounced hydrolysis. Furthermore, the same authors stated that the CO 2 solubility at set conditions is more important than the amount added to the water reaction, since undissolved CO 2 does not have any effect on hydrolysis. Thus, CO 2 has a great contribution to the acidity of the medium, but for a given CO 2 pressure, the ph depends mostly on the conditions applied. It is important to add that the presence of CO 2 in a hydrothermal process has not only a chemical but also a physical effect. For instance, Morais et al. found that, besides the CO 2 pressure influencing the solubility of CO 2 in water, the pressure of CO 2 exerts a force on the CO 2 /water mixture, helping CO 2 molecules to penetrate the porous biomass structure, and at the same time, due to CO 2 s presence in the headspace of the reaction, it reduces the partial pressure of water and keeps the hydronium ions, formed during autohydrolysis of water, in the liquid phase. 51 This also leads to higher glucose formation within shorter residence times, due to faster cleavage of glycosidic bonds than in the case of a pure water reaction. 41 Contrary to findings obtained by McWilliam et al. 133 and van Walsum et al., 60 a relationship between the amount of CO 2 loaded and the ph of the CO 2 - enriched water was attained. 51 Therefore, various reactions within the pathways are sensitive to the ph, especially if compared to different pretreatment methods Starch. Each type of polysaccharide needs different reaction conditions. Many biomass sources contain starch as the main polysaccharide, and the hydrolysis conditions require optimization to obtain a high yield of glucose, maltose, and fructose in a sustainable manner. Actually, the research on hydrolysis of starch started at the end of the 1990s. 40 Unlike cellulose, starch is easily hydrolyzed under hydrothermal conditions without the addition of organic acids or enzymes. However, the starch to glucose yield is lower than that achieved by enzymatic technologies. It is supposed that starch is much more susceptible to hydrolysis, leading to a higher concentration of degradation products from glucose. 148 Thangavelu et al. found that microwave hydrothermal hydrolysis of sago pith waste (starchy lignocellulosic biomass) was improved when CO 2 (dry ice) was added. A maximum glucose yield of 43.8% was obtained when CO 2 was applied at 900 W and 2 min of irradiation. 11 Nagamori and Funazukuri studied the hydrolysis of starch in water as a function of the glucose, fructose, maltose, and 5-HMF yields at C for various reaction times. They reported that high amounts of glucose were obtained early in the reaction, and with continuation of the reaction, glucose was degraded, mostly to 5-HMF. 149 Miyazawa and Funazukuri reported significantly lower yields of glucose from starch (3.7%) at 200 C after 15 min at unknown pressures. However, they found that the production of glucose increases significantly (14-fold) when CO 2 is added (CO 2 :starch ratio of 9, w/w) under similar conditions. 137 They found that high CO 2 pressure promotes hydrolysis of polysaccharides into low molecular weight sugars, as mentioned earlier for xylose production. The same tendency was found for glucose obtained from starch, where the yield was increased from 3.7% to 53.0% by increasing the CO 2 :biomass loading from 0 to 10.7 g g 1 when the reaction was conducted at 200 C for 15 min. Moreschi et al. found that a CO 2 /H 2 O binary system under subcritical conditions (200 C and 150 bar of CO 2 ) had a great effect on the hydrolysis of starch-rich materials into a mixture of gluco-oligosaccharides and glucose (above 97.1% after 15 min of residence time). 150 Also, Liu et al. reported a higher yield of starch-derived sugars when CO 2 was used as a soluble catalyst, which easily dissociates under the reaction conditions. 151 They observed an increase of the glucose yield by a factor of 50 quite independently of the initial starch concentration. Elevated glucose yields were obtained when CO 2 was used as a catalyst. An increased glucose yield of around 50% was obtained when starch was hydrolyzed at 230 C and 240 bar (96% CO 2 saturation) in comparison to that of a water-only reaction under the same conditions. 40 In addition, they found that, at a higher initial starch concentration (10%), hydrolysis with a CO 2 -saturated aqueous mixture resulted in a similar glucose yield and selectivity toward glucose formation compared to that of a lower starch loading (0.2%). 40 Orozco et al. reported a maximum glucose yield of 548 g kg of starch 1 from hydrolysis in hot water (200 C) with 30 bar of CO 2, accompanied by generation of 5-HMF HEXOSE-DERIVED SUGAR CONVERSION IN HIGH-PRESSURE CO 2 AND CO 2 /H 2 O MIXTURES The dehydration of biomass-derived cellulose and starch was studied in many systems using a large variety of catalysts, including organic acids or solid acids. 156,157 Many solvents, including supercritical water, 158,159 acetone, and methanol plus their mixtures, were employed in dehydration of glucose and fructose. 155,160 The decomposition of hexoses was widely studied under both sub- and supercritical water conditions for various reaction times. 161 In high-pressure water systems, the addition of CO 2 has a large role in the decomposition of hexoses due to the high concentration of 17

16 Chemical s Scheme 2. Chemistry and Applications of 5-HMF and Its Derivatives a a A solid arrow indicates direct transformation and a broken arrow a multistep reaction. Abbreviations: 5-HMF, 5-(hydroxymethyl)furfuran; LEVA, levulinic acid; LEVE, levulinic ester; FA, formic acid; HFCA, 5-(hydroxymethyl)furoic acid; FDC, 2,5-furandicarbaldehyde; FDCA, 2,5- furandicarboxylic acid; DHMF, 2,5-bis(hydroxymethyl)furan; DHM-THF, 2,5-bis(hydroxymethyl)tetrahydrofuran; HMTHFA, 5-(hydroxymethyl)- tetrahydrofuran-2-carbaldehyde. Compound numbers: (1) 2-(hydroxy(5-(hydroxymethyl)tetrahydrofuran-2-yl)methyl)-5-(hydroxymethyl)- tetrahydrofuran-2-carbaldehyde; (2) (E)-4-(5-(hydroxymethyl)furan-2-yl)but-3-en-2-one; (3) (1E,4E)-1,5-bis(5-(hydroxymethyl)furan-2-yl)penta- 1,4-dien-3-one; (4) tetrahydrofurfuryl alcohol; (5) 2,5-dimethyltetrahydrofuran; (6) furan; (7) 2-(hydroxymethyl)-5-vinylfuran; (8) furfuryl alcohol; (9) 2,5-bis(aminomethyl)furan; (10) 2-methyltetrahydrofuran; (11) 2,5-dimethylfuran; (12) 2-methylfuran. Reprinted from ref 83. Copyright 2011 American Chemical Society. hydronium and hydroxide ions from autoionization of water and from interaction between CO 2 and water, which promotes the dehydration reaction of glucose to a furan derivative (Scheme 1). Furthermore, in acid-catalyzed synthesis of 5- HMF from D-fructose, two different reactions can occur depending on the ph of the medium. The first one, at low ph (1.5), is 5-HMF rehydration into levulinic and formic acids, and the second is a polymerization reaction between 5-HMF and fructose taking place at ph > According to a report of the U.S. Department of Energy, 5- HMF is considered one of the most versatile platform chemicals derived from biomass. 162 The presence of two functional groups, combined with a furan ring, makes it a versatile precursor for a wide variety of chemicals relevant to the fuel, polymer, and pharmaceutical industries as depicted in Scheme For this reason, diverse raw materials were used to produce 5-HMF. For instance, lignocellulosic biomass, polysaccharides with a high molecular weight (cellulose, 163 starch, 164 and inulin 165 ), disaccharides (sucrose, maltose, and cellobiose 166,167 ), and monosaccharides (glucose and fructose 158,161 ) were transformed to 5-HMF, and the latter was converted to a wide variety of value-added products using supercritical fluids. Water under supercritical conditions is the most often used SCF to produce 5-HMF. However, highpressure water technologies require temperatures in the range of C so that, on an industrial scale, a potential reduction of the dehydration temperature means energy savings. The way to achieve this is the use of a CO 2 /water mixture. However, there is still little research in this field, but the outcomes indicate promising results. For example, some groups demonstrated that the production of 5-HMF from hexoses can be achieved at lower temperatures in the presence of CO 2. Wu et al. reported the conversion of inulin, an oligosaccharide consisting of glucose fructose or fructose polymers, to 5-HMF using a CO 2 /water mixture. 165 Such a system led to a higher conversion of inulin to 5-HMF with a yield of 53%, an increase of 15% compared to that of reactions without CO 2. Therefore, it can be assumed that the yield enhancement is a consequence of the in situ carbonic acid formation. The same authors described that conversion of inulin to 5-HMF is independent of the temperature used, since 18

17 Chemical s Scheme 3. γ-valerolactone Production Pathways a a A double-crossed arrow means several reaction steps. the CO 2 pressure is optimized to obtain a maximum 5-HMF yield. The maximum yield of 5-HMF was obtained for 60 bar of CO 2 at 160, 180, and 200 C. The main reason for this is that there is an optimal ph that can be achieved and can be adjusted by the CO 2 pressure. Similar results were obtained in an aqueous solution using iron vanadyl phosphate as the catalyst (around 40% yield) 168 and with ionic liquids/oxalic acid (56% yield). 169 Recently, the influence of scco 2, in an ionic liquid system, on the formation of 5-HMF from hexoses was reported by Machida et al. 170 In a series of reactions with an ionic liquid (1-butyl-3-methylimidazolium chloride) as a solvent plus scco 2 as a cosolvent and D-fructose as the starting material, a strong effect of scco 2 on the reaction results was shown. scco 2 demonstrated the ability to decrease the viscosity of ionic liquids, which is favorable for a high-yielding conversion of D-fructose to 5-HMF (above 90%). In this case, CO 2 acted as a viscosity-reducing agent for the ionic liquid reactions. 170 After dehydration of hexoses to 5-HMF, the CO 2 employed in the reaction can be used as a medium for extraction and recovery of 5-HMF, analogous to the previously discussed furfural recovery CONVERSION OF BIOMASS-DERIVED COMPOUNDS IN HIGH-PRESSURE CO 2 AND CO 2 /H 2 O MIXTURES 7.1. Hydrogenation Most of the processes where gases react with liquids are characterized by low reaction rates. This is mostly caused by the mass transfer limitations at the active sites of the catalysts. The solubility of some gases (for example, H 2 ) in water or other typical organic solvents is very low. scco 2 is an alternative solvent that is feasible for carrying out these reactions with either homogeneous or heterogeneous catalysts. CO 2 under supercritical conditions is still a liquid and is completely miscible with H 2. At the same time, scco 2 having a liquidlike density is a good solvent for nonpolar organics. Thus, taking these dual properties of CO 2 under supercritical conditions, it acts as a carrier for gas to the liquid phase, minimizing the mass transfer limitations An additional benefit is the reduction of both flammability and catalyst fouling risks. Thus, CO 2 can enhance the sustainable and economic viability of a hydrogenation process. 176 In this section we aim to demonstrate hydrogenations in scco 2 with chemicals obtained directly from biomass in the frame of the biorefinery concept. An example of application of scco 2 is a conversion of levulinic acid to γ-valerolactone (GVL) from glucose-derived biomass in which water was used as a cosolvent (Scheme 3). 177 Poliakoff s group developed an approach where a high GVL yield was achieved (>99%) even at a lower pressure (around 100 bar) with a high concentration of levulinic acid. 177 Furthermore, as GVL is soluble in water, the product accumulated in the aqueous phase can be easily extracted by scco 2 after the reaction. Levulinic acid is one of the most important platform molecules produced from biomass, and the conversion to GVL represents an important step in its further conversion to a liquid transport fuel. However, this is very challenging because of the relatively low solubility of levulinic acid in CO 2 (x levulinic acid = at bar at C), so a cosolvent (e.g., ethanol) is needed to improve this, giving a 3-fold increase of the levulinic solubility at the same conditions (x ethanol = 0.051). 178 Outstanding results for the hydrogenation of 5-HMF under a high-pressure CO 2 /H 2 O mixture were presented by Chatterjee et al., who designed a high-pressure CO 2 /water approach for the conversion of 5-HMF into 2,5-dimethylfuran (DMF) with a total yield of 100%. 179 Various products can be obtained, 19

18 Chemical s Scheme 4. Schematic Acid-Hydrolysis Reaction of a Peptide Bond Table 5. scco 2 -Mediated Delignification from Lignocellulosic Materials reaction conditions lignin/model compd solvent T ( C) P (bar) t (min) pulp yield (%) delignification yield (%) ref Pinus taeda wood CO 2 /C 2 H 6 O/H 2 O, 1:1, v/v sugarcane bagasse CO 2 /C 2 H 6 O/H 2 O, 1:1, v/v butanol/H 2 O, 60:40, v/v aspen chips CO 2 /SO 2, 98:2, mol/mol depending on the CO 2 pressure. At 80 C for 2 h with pressures below 100 bar of CO 2, tetrahydro-5-methyl-2- furanmethanol was produced, while pressures close to or above 100 bar led to the production of DMF and 2,5- dimethyltetrahydrofuran, with high selectivity. Another interesting note is that, in the absence of CO 2 or water, the selectivity toward DMF is negligible. 179 The same approach was tested in the hydrogenation of furfural, which was quantitatively converted into 2-methylfuran at a very short reaction time (10 min) Conversion of Proteins Amino acids, the building blocks for proteins, have a higher commercial value than other biomass constituents. Amino acids are interesting products for diverse industrial sectors such as the food industry, the agroindustry, and pharmaceutical manufacturing. Therefore, protein conversion into amino acids has been extensively studied using hydrothermal technologies. 180,181 Surprisingly, the hydrolysis of proteins with CO 2 has rarely been studied. Rogalinski et al. demonstrated that hydrolysis of proteins had a behavior similar to the hydrolysis of polysaccharides. 182 Scheme 4 illustrates the reaction mechanism of acid-catalyzed hydrolysis of peptides. The addition of CO 2 accelerates the hydrolysis reaction because proton addition to the peptide is a determining step for the reaction. 182 However, the amino acid yields are fairly low even under optimized conditions since the peptide bond in the protein molecule is much more stable than β-1,4- and β-1,6- glycosidic linkages in polysaccharides Delignification and Lignin Extraction under scco 2 Sustainable treatments of lignocellulosic materials become particularly attractive when employed into a complete cycle of all components. Biomass is capable of being entirely valorized, enabling its components to be converted into value-added products, followed by green treatment of any wastes generated with renewed buildup of biomass. 40 The development of cost-effective pretreatment technologies to separate polymers from biomass is needed. 183 Most of the technologies allow for a high sugar yield, but the full recovery of all components for maximal exploitation, including nonsugar fractions, should be one of the most important goals of biobased processes from biomass. In this section, special attention is given to the use of high-density fluids to remove lignin from the biomass matrix as well as prevent the degradation of wood polysaccharides. The valorization of lignin plays an important key role in the development of biorefinery processes for the production of biofuels, biomaterials, and biochemicals. The use of CO 2 as a solvent for wood pulping demonstrated several benefits, such as the generation of anomalous porosity and a lamellar structure of the biomass when subjected to the pretreatment. These effects make the organic matter more accessible to the solvents and/or to the pulping reagents, improving the pulping efficiency. 104,184 The results of scco 2 applications as a solvent for delignification and lignin extraction are compiled in Table 5. Schrems et al. investigated the effect of CO 2 presence on an organic solvent pulping process (ethanol/water) using model lignin substances. 187 Carbon dioxide was found to increase the efficiency of the delignification process by a decrease of the activation energy (42.0 and kj mol 1 for processes with and without scco 2, respectively) and by accelerating the breakdown. While CO 2 led to accelerated consumption of the model compounds, the formation of stable end products was slowed. scco 2 changed the chemical pathways and activation parameters and therefore the kinetics of the delignification process. A high CO 2 pressure was essential to promote the reaction between solvent and lignin. Pasquini et al. studied the pulp produced in an scco 2 process using the pulp of Pinus taeda wood chips and sugarcane bagasse with an ethanol/water mixture as a cosolvent. 184 In this work, different ethanol/water mixtures (50 100% ethanol for sugarcane bagasse and % ethanol for P. taeda wood chips), temperatures ( C), pressures ( bar), and residence times ( and min for sugarcane bagasse and P. taeda wood chips, respectively) were evaluated. They found that higher pressures led to an elevated residual lignin content, while pulping yields were found to be similar. The process with 160 bar of pressure at 190 C gave residual Klason lignin contents of 4.9% and 8.7% for P. taeda wood chips and sugarcane bagasse, respectively. The pulping yields were 43.7% and 32.7% for these materials, and delignification was in the range of 93% and 88%, respectively. They also observed that higher pressures led to the same yields but lower delignification. Virtually complete delignification occurred after 30 min of cooking time, mostly due to the high pressure and high acidity resulting from the in situ carbonic acid generation. 184 Curvelo and co-workers studied the delignification of sugarcane bagasse with sub- and supercritical CO 2 using a 1-butanol/water mixture as a cosolvent. 185 The highest delignification content (94.5%) was obtained at 190 C and 70 bar for 105 min using a mixture of 1- butanol/water (60%:40%, v/v). However, it is important to 20

19 Chemical s Table 6. Engagement of CO 2 in the Diverse Fields of Biomass Processing Described in This primary components applications catalyst outcome ref Main Fractions cellulose (dried) pretreatment scco 2 cellulose with low crystallinity index; morphology changes; increase of surface area 73 cellulose (wet) hydrolysis CO 2,CO 2 /H 2 O cellulose oligomers (low concentration); oligomers with low molecular weight; cellulose with low crystallinity index; particle size reduction and increase of surface area; lower reaction temperature and shorter residence times; low ph of medium cellulose (specific a w ) biocatalysis scco 2 + enzymes production of D-glucose monomers; increase of activity and stability of cellulases and β-glucosidases; better transport properties and accelerated mass transfer; increase of reaction rate 41, 51, 60, 133, 146 starch biocatalysis scco 2 + enzymes production of D-glucose monomers and its degradation products; increase of activity 16, 126 and stability of amylases and glucoamylases hydrolysis CO 2 /H 2 O higher production of glucose and fructose (in both oligomer and monomer forms) 149 hemicellulose hydrolysis CO 2 /H 2 O hemicellulose depolymerization and hydrolysis; high concentration of hemicellulose oligomers (with low DP); slow conversion of oligomer to monomers; acetic acid release; low production of degradation products at mild conditions; low ph of medium; lower reaction temperatures; shorter residence times; extraction of degradation compounds (mainly furfural) with CO 2 32, 132, 142 lignin delignification and lignin extraction scco 2 +H 2 O/C 2 H 6 O; scco butanol/H 2 O; scco 2 /SO 2 ; CH 3 COOH/H 2 O; CH 3 COOH/CO 2 ; CH 3 COOH/H 2 O/CO 2 ; dioxane and CO 2 promotion of the reaction between solvent and lignin; high yields of lignin extraction; decrease of activation energy; CO 2 changes the chemistry of the process; degradation of polysaccharides proteins hydrolysis CO 2 /H 2 O production of amino acids; accelerated hydrolysis rate 41 Intermediates hexoses (glucose and fructose) , , 188, 190 dehydration CO 2 /H 2 O; scco 2 /IL production of 5-HMF; decreases the viscosity of ionic liquids; 5-HMF extraction with 165, 170, CO levulinic acid hydrogenation CO 2 /H 2 O production of γ-valerolactone and its extraction with CO HMF hydrogenation CO 2 ;CO 2 /H 2 O production of furanic compounds; increase of the reaction selectivity; CO 2 pressure 179 drives the products obtained furfural hydrogenation CO 2 /H 2 O production of furanic compounds 179 lignin depolymerization CO 2 /C 3 H 6 O/H 2 O recovery of phenolic and aromatic compounds 193, 194 point out that high delignification under these conditions also led to significant loss of polysaccharides. Nevertheless, the optimum conditions for lignin separation and polysaccharide preservation were obtained with high pressures and a low content of 1-butanol in the cosolvent mixture. Shah et al. studied the effect of scco 2 /SO 2 (98%:2%, mol/mol) treatment on hardwood by separation of lignin from the polysaccharide fraction. 186 CO 2 /SO 2 pretreatment at 140 C for 4 h removed 84.0% of the lignin while retaining 78.3% of the polysaccharides. They also observed the highest loss of carbohydrates at temperatures above 130 C. Kiran et al. investigated high-pressure extraction and delignification of red spruce using acetic acid/water, acetic acid/carbon dioxide, and acetic acid/ water/carbon dioxide binary and ternary systems. 188 The best result was obtained at 180 C, 250 bar, and 3 h of extraction time, employing a 0.73 mole fraction of acetic acid. Under these conditions, 53% mass loss and 93% delignification were obtained. In other work, a binary mixture of ethanol and carbon dioxide (0.76 mole fraction of CO 2 ) under supercritical conditions (T = 167 C and p = 325 bar) was employed, and 4.8% dissolution was achieved after 2 h of treatment. A water/ carbon dioxide mixture (0.39 and 0.88 mole fractions of CO 2 ) at 190 C and 290 bar led to mass losses of 17.5% and 35.9% for lower and higher water contents, respectively. The system ethanol/water/carbon dioxide with molar fractions of CO 2 and H 2 O of 0.91 and 0.022, respectively, under the same conditions as the binary water/co 2 system, resulted in a dissolution yield of 19.3% of the wood present in the reactor. 189 One of the most important aspects was that the physical properties of the pulp obtained with this mixture were comparable to those of conventional pulps. The delignification of Eucalyptus globulus wood using a mixture of dioxane and CO 2 in the temperature range from 160 to 180 C at 170 bar and 5 h of residence time was studied by Machado et al. 190 Within the range of temperature used, only negligible amounts of lignin and hemicellulose were recovered. These data confirm that the composition of the reactive mixture influences the selectivity of the extraction. For instance, using a CO 2 -rich mixture, it was possible to remove all hemicellulose, while the maximum amount of lignin was obtained with pure dioxane. Ivanhov et al. used oxidative delignification in scco 2 to decompose lignin by demethoxylation Lignin Depolymerization The development of biorefinery processes will result in large quantities of lignin, in addition to the lignin produced in pulp and paper factories, and this opens opportunities for the recovery of phenolic compounds and other aromatic chemicals for the production of biofuels, bio-based materials, and chemicals. 30,46,192 It is known that scco 2 decreases the temperature of the solvent stream and allows the condensation of aromatic intermediate products, with no solvent present in the final product mixture. Conversions of lignin to phenolic compounds using a CO 2 /acetone/water mixture were studied and are presented in the literature discussed below. The maximal yield of oligomeric fragments and aromatic compounds in monomeric form produced was 10 12%, based on lignin reacted at 300 C and 100 bar. This yield is similar to those obtained in base-catalyzed thermal processes for lignin depolymerization, and the approach also led to the production 21

20 Chemical s of a wide variety of monomeric compounds, giving maximum yields of 2% and 3.6% for syringic acid and syringol, respectively. 193 Eckert et al. also demonstrated that CO 2 is a gas expander liquid to extract phenolic compounds such as vanillin, syringaldehyde, and syringol OVERVIEW OF CO 2 APPLICATIONS WITHIN THE BIOREFINERY CONCEPT This review has shown that high-pressure CO 2 is an interesting solvent, which can be used in a great variety of reactions within the biorefinery concept starting from its use in direct biomass processing to its use as a catalyst for dehydration of biomassderived carbohydrates and its use in specific applications for biomass intermediary compounds. The application of highpressure CO 2 technology in such a wide variety of processes is still in its infancy. However, due to the extraordinary properties of CO 2, either acting individually or as a cosolvent in mixtures, CO 2 (under both sub- and supercritical conditions) may contribute to tackling the aforementioned challenges. Table 6 shows the most representative examples of CO 2 application in biomass processing and the effects of sub/supercritical CO 2 and mixtures of CO 2 with other solvents (water, ethanol, etc.) on each biomass fraction (either principal compounds or intermediary products). 9. ECONOMIC ASPECTS OF CO 2 PROCESSING OF BIOMASS A number of reports of catalytic reactions in high-pressure CO 2 confirm the commercial potential of these processes. In spite of the requirement for sustainable reactions, CO 2 can be used as a helpful chemical in green technologies. Furthermore, CO 2 -based technologies can be designed to deliver optimal performance with maximal energy savings. Furthermore, in terms of biomass processing, CO 2 -assisted technologies have benefits over conventional processes, as presented above. Considering the economic aspects of CO 2 use in biomass processing, it is important to mention that the biorefinery concept assumes the integrated valorization of biomass toward energy, materials, and chemicals. The production of bulky but low-value energy sources such as ethanol can be counterbalanced by coproduction of small quantities of highly priced chemicals, for instance, polyphenols and oligosaccharides. This joint on-site valorization gives economic benefits, which are already observed in classical petrorefineries. The production of bulky low-price fuels is subsidized by production of low amounts of high-value polymers. It is also very important to point out that the price of feedstock is a large proportion of the final product price. In the case of fossil feedstock, the prices are sometimes far above $100 per barrel, 195 while biomass feedstock can be obtained at low or no cost. Furthermore, the great variety and versatility of existing biomass make future biorefineries less dependent on geopolitical turmoil. One of the important aspects that should be taken into consideration when comparing classical petrochemical refineries and new biorefineries is the maturity of both technologies. The classical petrochemical refineries were developed and optimized over more than a century, 196 while biorefineries are still in their infancy. Even nowadays, existing biorefineries do not valorize all production streams, or valorize them in a way that is not the most economically feasible. Thus, because of these aspects, both classical petrochemical refineries and biorefineries should not be compared straightaway. The use of CO 2 as a potential reactant in biorefineries might be especially beneficial, as CO 2 is often a waste product from many industrial processes (cement factories or power plant factories) and, as such, can be obtained with little cost. At the same time, it is considered as a green solvent, which definitively improves the environmental aspects of the future biorefinery. Independently of the manner that CO 2 pretreatment is employed, current efforts to develop new methods of biomass valorization with CO 2 do not yet guarantee its economic and technical viability. These technologies require large investment costs, mostly related to the high-pressure equipment. On the other hand, these technologies demonstrated some advantages, e.g., in downstream processing for various applications, especially in comparison to other classical processes such as acid hydrolysis. Additionally, in the case of high-pressure technologies, higher yields and easier control of process selectivity can be obtained normally, which are beneficial for the process economics. Therefore, to find the most adequate application of CO 2 in biorefineries, a large investment in R&D and later in pilot and demonstration projects is needed. 3 In addition, each process must be analyzed separately to answer the question related to the economic feasibility of the processes with CO 2, considering the state of the art of the technology applied. 10. PERSPECTIVES The present review shows the most important aspects of CO 2 use as a green solvent in biomass valorization. As mentioned before, the scope of this review does not include the extraction part, as this was tackled by dozens of investigations during the previous decades. The conclusion for CO 2 use in biomass processing is that, in general, CO 2 is still rarely used and that its special benefits as a solvent under supercritical conditions would justify its broader application. Thus, there is still place for improvements, and the use of CO 2 should be explored further in several areas. Some of these are (i) direct hydrolysis of untreated biomass in the presence of CO 2, (ii) use of other potentially beneficial cosolvents in mixtures with CO 2 instead of the CO 2 /H 2 O system, (iii) development of CO 2 -formed value-added products from all biomass fractions, (iv) process analysis from a sustainability point of view, and (v) integrated valorization of each biomass fraction. (i) A direct hydrolysis of untreated biomass in the presence of CO 2 would allow the integration and intensification of the biomass valorization. In this way, biomass would not require a cost-demanding pretreatment step, allowing the use of biomass directly in glucose production, which later could be converted to energy sources such as biofuels. This challenging task has barely been tackled up to now, and only pure cellulose samples (cotton 124 or Avicel 117 ) have been examined. It is important to stress that this is one of the most important aspects in the economic and technological development of biomass processing, which also would require advances in enzyme engineering to develop more resistant and efficient biological catalysts. 127 (ii) Water is one of the best-known solvents; however, most organic compounds are insoluble in water. In addition, the mutual solubility of water and CO 2 is limited; thus, to take advantage of sub- or supercritical conditions, high pressures and temperatures are required. Therefore, looking for alternatives for water used in mixtures with CO 2 is an interesting field of research, which has not yet been properly explored. The latent solution might be the use of carbonate-type solvents. This new class of solvents is an alternative for many existing problems. 22

21 Chemical s For example, the synthesis of 1,2-glycerolcarbonate 197,198 uses the overproduced glycerol. In addition, the synthesis of this class of compounds also involves CO 2 insertion, making the synthesis greener and more sustainable. It cannot be forgotten that other solvents can be used, although they have to fulfill the requirements of green chemistry and green chemical engineering. 47, (iii) Lignin was often considered as a less valuable fraction of biomass. Nowadays, however, it serves as a source of polyphenols. 10,30,46,192,193 These references show the great potential in the CO 2 -assisted production of phenolic compounds directly from biomass. It should also be mentioned that this approach is different from the already well-known and examined extraction of value-added phenolic-like compounds from specific plants for pharmaceutical or food purposes. 203,204 Hemicellulose is also very often considered as a fraction with lower value. The reason for this is its diverse composition, which is strongly dependent on the biomass type and other external factors (e.g., climacteric and storage conditions). This diversity makes the hemicellulose fraction valorization challenging and less versatile, but it gives room for numerous products that can be obtained, depending on the technique applied. 20,107 Among them are oligosaccharides, 32,205 xylitol, 206 and many others, including simple sugars, 107 which can be produced with the assistance of high-pressure CO 2. Cellulose is mostly used for ethanol production, but there are many other chemicals that can be produced from this feedstock. Some examples (5-HMF, levulinic acid, formic acid, GVL) presented in this work show the unrevealed potential of CO 2 adjunctive application in cellulose processing. 16 (iv) SWOT (strengths, weaknesses, opportunities, and threats) analysis of any process is crucial to find the pros and cons of each technology. This can be done in many ways considering various aspects. One of the easiest methods to do so is via the use of green metric techniques. 207,208 Although the proposed manner of process analysis is much simpler than, for example, life cycle analysis, it is important to add that it can be used successfully for very complex processes. It considers important metrics such as economic feasibility amd energy and material efficiencies, 209 depending on the particular needs of the examined process. In the case of processes with CO 2 in the biorefinery, it would be especially beneficial because the complexity of the processes and lack of many data would be compensated by the simplicity that the use of green metrics brings. 210 (v) As was already mentioned, LCMs are constituted by three main fractions: cellulose, hemicellulose, and lignin. Up to now the integrated valorization of these three fractions at the industrial scale has not existed. The need for more research in this direction is crucial to provide a sufficient knowledge in this area and to gain sufficient maturity to advance toward pilot or demonstration scales of LCM biorefineries. The major problem is that, even at the bench-scale level, these kinds of processes are rarely presented in the literature; thus, academia needs to make more effort to perform research that tackles the various aspects of biomass processing. 11. CONCLUSIONS Awareness of today s energy crisis and the climatic changes, which are threatening the world economy and sustainability, have increased the drive to substitute the current hazardous technologies by more sustainable, greener, and environmentally friendly processes. Moreover, the wide-world production, high abundance, and great variety of LCMs makes them part of the solution for the dependence on fossil fuel processes. Supercritical carbon dioxide as well as other nowadays broadly examined solvents and processes fits into this trend, recognizing CO 2 as a nontoxic, green, and more sustainable solvent facilitating biomass processing. When in contact with water, which might be the moisture in biomass, this technology gets additional benefits from the acidity introduced by the formation of carbonic acid. High-pressure CO 2 has been demonstrated to be very effective in fractionation and hydrolysis of lignocellulosic fractions, showing its utility to achieve a range of goals in biomass processing. Furthermore, after pretreatment, highpressure CO 2 can act as an extraction medium, removing products of interest or inhibitors of further processing, e.g., fermentation. Thus, supercritical CO 2 has proved to be an attractive alternative medium for integrated biomass processing within the biorefinery concept. AUTHOR INFORMATION Corresponding Author * rafal.lukasik@lneg.pt. Fax: Phone: , ext Notes The authors declare no competing financial interest. Biographies Ana R. C. Morais received her M.Sc. and B.Sc. degrees in food engineering and in biosciences microbiology from the School of Agronomy of the Technical University of Lisbon and Portuguese Catholic University. Currently, she is a Ph.D. researcher in sustainable chemistry in the Bioenergy Unit at the National Laboratory of Energy and Geology (LNEG), Portugal. She is experienced in high-pressure fluids, namely, CO 2, working under the supervision of Dr. Rafał Bogel- Łukasik and Prof. Manuel Nunes da Ponte. Her research is focused on the application of supercritical CO 2 as a green solvent in processing biomass in the frame of the biorefinery concept. She has given a series of lectures on her work to international congresses. Recently, she was awarded the Innovation for Sustainability Award (2014) in the category processes for the RefinOlea Integrated Valorization of Residues and Sub-Products of Olive Oil Extraction Project. In 2012 she was awarded the Poster Award at the third UBIOCHEM COST Action Conference for the work entitled Biological Pathways of Isoprene Production. 23

22 Chemical s ACKNOWLEDGMENTS This work was supported by the Fundac ao para a Cie ncia e a Tecnologia (FCT, Portugal) through Bilateral Cooperation Project FCT/CAPES 2014/2015 (Grant FCT/1909/27/2/ 2014/S) and Grants SFRH/BD/94297/2013 (A.R.C.M.), SFRH/BD/90282/2012 (A.M.d.C.L.), and IF/00424/2013 (R.B.-Ł.) and by CAPES (Brazil) through the project Pesquisador Visitante Especial 155/2012. Special thanks are given to Mr. R. Andrew Wilbey for proof-reading this work. REFERENCES (1) Clark, J. H.; Deswarte, F. E. I.; Farmer, T. J. Biofuels, Bioprod. Biorefin. 2009, 3, 72. (2) FitzPatrick, M.; Champagne, P.; Cunningham, M. F.; Whitney, R. A. Bioresour. Technol. 2010, 101, (3) HORIZON2020 Proposal for a Council Decision Establishing the Specific Programme Implementing Horizon 2020 The Framework Programme for Research & Innovation ( ); European Commission: Brussels, Belgium, (4) Morais, A. R. C.; Bogel-Lukasik, R. Sustainable Chem. Processes 2013, 1, 18. (5) Van Ree, R. IEA Bioenergy Task 42 Biorefining; International Energy Agency IEA Bioenergy: Paris, (6) Elvnert, J.; Kleinschmit von Lengefeld, A.; Ringman-Beck, J.; Kerckow, B.; Sternberg, K.; Carrez, D.; Laurmaa, J.; Fehrmann, J.; Metzlaff, K.; Travella, S.; Perimenis, A.; Majer, S.; Meisel, K.; MullerLanger, F.; Makinen, T.; Clark, J. H.; Kazmi, A.; Thomas, D.; van Ree, R. Joint European Biorefinery Vision for 2030, Star-COLIBRI, Strategic Targets for 2020 Collaboration Initiative on Biorefineries; European Union: Brussels, Belgium, (7) Roper, H. Starch/Staerke 2002, 54, 89. (8) Budarin, V. L.; Shuttleworth, P. S.; Dodson, J. R.; Hunt, A. J.; Lanigan, B.; Marriott, R.; Milkowski, K. J.; Wilson, A. J.; Breeden, S. W.; Fan, J. J.; Sin, E. H. K.; Clark, J. H. Energy Environ. Sci. 2011, 4, 471. (9) Bourissou, D.; Guerret, O.; Gabbai, F. P.; Bertrand, G. Chem. Rev. 2000, 100, 39. (10) Buranov, A. U.; Mazza, G. Ind. Crops Prod. 2008, 28, 237. (11) Thangavelu, S. K.; Ahmed, A. S.; Ani, F. N. Appl. Energy 2014, 128, 277. (12) Tomme, P.; Warren, R. A. J.; Gilkes, N. R. Adv. Microb. Physiol. 1995, 37, 1. (13) Ratanakhanokchai, K.; Waeonukul, R.; Pason, P.; Tachaapaikoon, C.; Kyu, K. L.; Sakka, K.; Kosugi, A.; Mori, Y. In Biomass Now Cultivation and Utilization; Matovic, M. D., Ed.; InTech: Rijeka, Croatia, (14) Hsu, T. C.; Guo, G. L.; Chen, W. H.; Hwang, W. S. Bioresour. Technol. 2010, 101, (15) Saeman, J. F. Ind. Eng. Chem. 1945, 37, 43. (16) Moreschi, S. R. M.; Petenate, A. J.; Meireles, M. A. A. J. Agric. Food Chem. 2004, 52, (17) Theander, O.; Nelson, D. A. Adv. Carbohydr. Chem. Biochem. 1988, 46, 273. (18) Garrote, G.; Dominguez, H.; Parajo, J. C. Holz Roh- Werkst. 1999, 57, 191. (19) Lynd, L. R.; Weime, P. J.; Van Zyl, W. H.; Pretorius, I. S. Microbiol. Mol. Biol. Rev. 2002, 66, 506. (20) Girio, F. M.; Fonseca, C.; Carvalheiro, F.; Duarte, L. C.; Marques, S.; Bogel-Lukasik, R. Bioresour. Technol. 2010, 101, (21) Galbe, M.; Zacchi, G. Biofuels 2007, 108, 41. (22) Sun, Y.; Cheng, J. Y. Bioresour. Technol. 2002, 83, 1. (23) Wyman, C. E.; Dale, B. E.; Elander, R. T.; Holtzapple, M.; Ladisch, M. R.; Lee, Y. Bioresour. Technol. 2005, 96, (24) Wyman, C. E.; Dale, B. E.; Elander, R. T.; Holtzapple, M.; Ladisch, M. R.; Lee, Y. Y. Bioresour. Technol. 2005, 96, (25) Wooley, R.; Ruth, M.; Glassner, D.; Sheehan, J. Biotechnol. Prog. 1999, 15, 794. Andre M. da Costa Lopes received both B.Sc. and M.Sc. degrees in biotechnology from the University of Aveiro, Portugal. He gained experience in using alternative solvents for the valorization of agricultural and industrial wastes, particularly pretreatment of lignocellulosic biomass using ionic liquids and hydrogenation of animal fat residues in supercritical CO2. Currently, he is a Ph.D. student in sustainable chemistry working under the supervision of Dr. Rafał Bogel-Łukasik and Prof. Manuel Nunes da Ponte. His research is focused on valorization of lignocellulosic biomass using alternative solvents in the frame of green and sustainable chemistry toward integration in biorefineries. Rafał Bogel-Łukasik graduated from the Warsaw University of Technology, Poland, with M.Sc. and B.Sc. degrees (2002). He received his Ph.D. (2007) in chemical engineering from the New University of Lisbon, Portugal. Since 2009 he has held the position of Researcher in the Bioenergy Unit at the National Laboratory of Energy and Geology (LNEG), Portugal. He was Marie Curie Fellow at The Queen s University Ionic Liquid Laboratories Research Centre, United Kingdom, at the Institute of Experimental Biology and Technology, Portugal, and at the Technical University of Hamburg, Germany. In 2008 he was awarded The International Association of Chemical Thermodynamics Junior Award for Excellence in Thermodynamics. Recently, he was awarded the Innovation for Sustainability Award in the category processes and the Green Project Award for the RefinOlea Integrated Valorization of Residues and Sub-Products of Olive Oil Extraction Project. Overall, he is the coauthor of more than 60 peer-reviewed papers and a few book chapters and an invited editor of a book of the Royal Society of Chemistry Series, entitled Ionic Liquids in the Bioref inery Concept. He is also a guest editor of the White Biotechnology Sustainability in Industry thematic issue of Sustainable Chemical Processes. His work is focused on biomass processing using the integrated approach of green chemistry in the biorefinery concept. 24