Bio-ethanol ? the fuel of tomorrow

Bio-ethanol ? the fuel of tomorrow

(Parte 1 de 2)

Bio-ethanol – the fuel of tomorrow from the residues of today

B. Hahn-Hagerdal, M. Galbe, M.F. Gorwa-Grauslund, G. Liden and G. Zacchi Lund University, PO Box 124, S-221 0 Lund, Sweden, Getingevagen 60

The increased concern for the security of the oil supply and the negative impact of fossil fuels on the environment, particularly greenhouse gas emissions, has put pressure on society to find renewable fuel alternatives. The most common renewable fuel today is ethanol produced from sugar or grain (starch); however, this raw material base will not be sufficient. Consequently, future large-scale use of ethanol will most certainly have to be based on production from lignocellulosic materials. This review gives an overview of the new technologiesrequiredandtheadvancesachievedinrecentyears to bring lignocellulosic ethanol towards industrial production. One of the major challenges is to optimize the integration of process engineering, fermentation technology, enzyme engineering and metabolic engineering.

Introduction One of the greatest challenges for society in the 21st century is to meet the growing demand for energy for transportation, heating and industrial processes, and to provide raw material for the industry in a sustainable way. An increasing concern for the security of the oil supply has been evidenced by increasing oil prices, which during 2006 approached US$80 per barrel. More importantly, the future energy supply must be met with a simultaneous substantial reduction of green house gas emissions. Actions towards this aim have been initiated. The EuropeanCommissionplanstosubstituteprogressively20%of conventional fossil fuels with alternative fuels in the transport sector by 2020, with an intermittent goal set at 5.75% in 2010. In the USA, the Energy Policy Act of 2005 requires blending of 7.5 billion gallons of alternative fuels by 2012 [1], and recently the US President, in his state of the union address, set the goal to replace more than 75% of imported oil with alternative fuels by the year 2025 [2]. Liquid biofuels from renewable resources, particularly from lignocellulose materials, will have a substantial role in meeting these goals (Box 1).

Ethanol has already been introduced on a large scale in

Brazil, the US and some European countries, and we expect it to be one of the dominating renewable biofuels in the transport sector within the coming 20 years. Ethanol can be blended with petrol or used as neat alcohol in dedicated engines, taking advantage of the higher octane number and higher heat of vaporization; furthermore, it is an excellent fuel for future advanced flexi-fuel hybrid vehicles.Currently, ethanolfor the fuelmarket is produced from sugar (Brazil) or starch (USA) at competitive prices. However, this raw material base, which also has to be used for animal feed and human needs, will not be sufficient to meet the increasing demand for fuel ethanol; and the reduction of greenhouse gases resulting from use of sugaror starch-based ethanol is not as high as desirable [3]. Both these factors call for the exploitation of lignocellulose feedstocks, such as agricultural and forest residues as well as dedicated crops, for the production of ethanol. This review summarizes recent developments in the bioconversion processes aimed at fuel ethanol production, with emphasis on process integration. In particular, the concept that each individual unit operation has to be developed and optimized in relation to the preceding and subsequent process steps will be discussed.

Overview of the conversion process With so many advantages, why are there still no production facilities using lignocellulosic materials? Ethanol is currently produced from sugar cane and starch-containing materials, where the conversion of starch to ethanol includes a liquefaction step (to make the starch soluble) and a hydrolysis step (to produce glucose). The resulting glucose is then readily fermented. Although there are similarities between the lignocellulosic and the starch process, the techno-economic challenges facing the former are large. There are several options for a lignocellulose-to-ethanol process but, regardless of which is chosen, the following features must be assessed in comparison with established sugar- or starch-based ethanol production. (i) Efficient de-polymerization of cellulose and hemicellulose to soluble sugars. (i) Efficient fermentation of a mixed-sugar hydrolysate containing six-carbon (hexoses) and five-carbon (pentoses) sugars as well as fermentation inhibitory compounds. (i) Advanced process integration to minimize process energy demand. (iv) Cost-efficient use of lignin.

The first step in the conversion of biomass to ethanol is size reduction and pretreatment. In this review, only the enzymatic process (Figure 1) will be discussed because it is considered to be the most promising technology [4–6]. The hemicellulose and cellulose polymers are hydrolyzed with enzymes or acids to release monomeric sugars. The sugars from the pretreatment and enzymatic hydrolysis steps are fermented by bacteria, yeast or filamentous fungi,

Review TRENDS in Biotechnology Vol.24 No.12

Corresponding author: Zacchi, G. ( Available online 16 October 2006. 0167-79/$ – see front matter 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.tibtech.2006.10.004 although the enzymatic hydrolysis and fermentation can also be performed in a combined step – a so-called simultaneous saccharification and fermentation (SSF). After final purification (by distillation and molecular sieves or other separation techniques), the ethanol is ready to be used as a fuel, either neat or blended with petrol. A part of the lignin, the principal solid part of the biomass remaining, can be burnt to provide heat and electricity for the process, whereas the rest is retained as a valuable coproduct. The most probable use today would be as an ash-free solid fuel, but various technologies are under development to convert it to a higher-value product, which could form the basis for a new branch of industrial chemistry [7].

First, the biomass is converted to sugars

It is only 40 years ago that the biodegradation of lignocellulosics was first discussed [8]. Enzyme conversion is substrate-specific without by-product formation, which reduces inhibition of the following process steps. However, enzyme-catalysed conversion of cellulose to glucose is slow unless the biomass has been subjected to pretreatment, which is also required to reach high yields and to make the process commercially successful [9]:thepretreatment aims to increase pore size and reduce cellulose crystallinity. In acid-catalyzed pre-treatment, the hemicellulose layer is hydrolyzed, whereas in alkali-catalyzed pretreatment, mainly, a part of the lignin is removed and hemicellulose has to be hydrolysed by the use of hemicellulases. Hence, pretreatment is necessary to expose the cellulose fibres to the enzymes or to at least make the cellulose more accessible to the enzymes. An efficient pretreatment can substantially reduce the enzyme requirements, which make up a large part of the production cost.

Pretreatment is usually assessed in a number of ways: by enzymatic hydrolysis (EH) of the solid material to determine the digestibility; by fermentation of the liquid to assess the effect of potential inhibitors towards the fermenting microorganism; and/or by simultaneous saccharification and fermentation (SSF) of the pretreated material. The conditions for the assessment can vary, particularly in terms of washed or non-washed solids and the concentration of solids and enzymes in the EH and SSF, making comparison of results from different investigations difficult.

InanextensivestudyundertakenintheUSA,wherethe same batch of corn stover was pretreated using various methods (e.g. dilute acid, AFEX, hot water treatment) and then subjected to standard evaluation techniques, the yields of sugars were found to be more or less the same [10]. Total sugar yields – after pretreatment followed by enzymatic hydrolysis – of around 90% or more were reached, demonstrating that corn stover is an easily degradable material. When corn stover was steam pre- treated with small amounts of SO2, overall sugar yields close to the theoretical value were obtained; steam pre- treatment without a catalyst also resulted in 90% glucose yield [1].

In countries such as Sweden, Canada and the USA, much of the available biomass is softwood, which is more difficult to hydrolyse than corn stover. For softwood, steam pretreatment with the addition of an acid catalyst such as

Box 1. Advantages of lignocellulose-based liquid biofuels

Biofuel sources are geographically more evenly distributed than the fossil fuels; thus, the sources of energy will, to a larger extent, be domestic and provide security of supply. Lignocellulosic raw materials minimize the potential conflict between land use for food (and feed) production and energy feedstock production. The raw material is less expensive than conventional agricultural feedstock and can be produced with lower input of fertilizers, pesticides, and energy. Biofuels from lignocellulose generate low net greenhouse gas emissions, reducing environmental impacts, particularly climate change. Biofuels might also provide employment in rural areas

Figure 1. Schematic flowsheet for the conversion of biomass to ethanol.

550 Review TRENDS in Biotechnology Vol.24 No.12

H2SO4 or SO2 is a prerequisite to reach high sugar yields. Acid increases the recovery of hemicellulose sugars and it improves the enzymatic hydrolysis of the solid fraction; the acid catalyst in steam pretreatment functions similar to acid pulp cooking but with less liquid.

Steam pretreatment of SO2 impregnated spruce chips yields a material that is relatively easy to hydrolyze and ferment, whereas dilute acid impregnation results in a material harder to ferment because of the generation of inhibitory compounds [12]. With present bench-scale technology, it is possible to obtain around 300 litres of ethanol per metric ton spruce, which is 70% of the theoretical overall yield based on hexose sugars [12,13].

Steam pretreatment with the addition of a catalyst for hydrolysis and improved enzymatic digestibility is the closest to commercialisation. It has been widely tested in pilot-scale equipment, for example, in the Iogen pilot plant (Canada) [14], the Souston pilot plant (France) [15] and in the pilot plant in O rnskoldsvik (Sweden), and is used in a demonstration-scale ethanol plant at Iogen (Canada). It is also to be used in Salamanca (Spain), in a plant constructed by the company Abengoa.

Most probably, there will not be one general method because different types of raw material require different pretreatments. So far, methods such as ammonia fiber explosion (AFEX), wet oxidation and liquid hot water (LHW) treatment seem to be more successful for agricultural residues [16–18], whereas steam pretreatment has resulted in high sugar yields for both forestry and agricultural residues. Glucose yields >90% and xylose yields >80% were obtained after enzymatic hydrolysis, both with and without the addition of an acid catalyst [12,19–21].

Acid-catalyzed pretreatment primarily solubilizes the hemicellulose fraction into the liquid phase. For softwood, the liquid mainly contains solubilized mannose in addition tosmallamountsofxylose,arabinose,galactoseandglucose. The solid phase comprises lignin and cellulose, the latter of which is subjected to enzymatic hydrolysis. The maximum cellulase activity of most fungal-derived cellulases and bglucosidases is observed at 50 8C and at a pH of 4.0–5.0; however, the optimal conditions vary with the hydrolysis time and are dependent on the source of the enzymes [2]. Cellulases belong to two groups of enzymes known as endoglucanases (EG) and cellobiohydrolases (CBH), respectively. EG randomly attack the cellulose chain, creating free ends for CBH to cleave dimers of glucose (cellobiose) off [23]. A third type of enzyme, b-glucosidase, which hydrolyzescellobioseintotwoglucosemolecules,isalsonecessary: in the absence ofb-glucosidase, end-product inhibition from cellobiose will occur. Furthermore, compounds generated during pretreatment might have an adverse effect on enzymatic hydrolysis. The enzymatic hydrolysis of spruce was greatlyimprovedwhentheliquidfractionfromthepretreatment step was replaced with a buffer solution [24].T his could not be entirely ascribed to the reduction in end-product inhibition, suggesting that inhibitory compounds had also been removed.

In an initiative of the US Department of Energy, two companies, Genencor International (http://w.genencor. com/) and Novozymes Inc. (, were awarded $17 million each, with the goal to reduce the enzyme cost 10-fold ( However, the production cost of enzymes is still too high and requires further reduction. One way to achieve this is to use a fraction of the feedstock and/or the hydrolysate for in situ enzyme production by fungi or other microorganisms [25].

and then the sugars are fermented to ethanol

Contrary to sucrose- and starch-based ethanol production, lignocellulose-based production is a mixed-sugar fermentation in the presence of inhibiting compounds – low molecular weight organic acids, furan derivatives, phenolicsandinorganiccompounds–releasedandformedduring pretreatment and/or hydrolysis of the raw material [26]. Lignocellulosic raw materials, in particular hardwood and agricultural raw materials, can contain 5–20% (or more) of the pentose sugars xylose and arabinose, which are not fermented to ethanol by the most commonly used industrial fermentation microorganism, the yeast Saccharomyces cerevisiae. Xylose is by far the most abundant pentose sugar, whereas arabinose can constitute as much as 14–15% in corncob hulls and wheat bran, respectively. Consequently, most research efforts have been devoted to thedevelopmentofefficientxylose-fermentingmicroorganisms [27,28].

Xylose-fermenting microorganisms are found among bacteria, yeast and filamentous fungi [29].A naerobic bacteria ferment pentoses, but are inhibited already at low sugar and ethanol concentrations. In addition, the ethanolic fermentation occurs with considerable by-product formation, which reduces the ethanol yield [30]. Natural xylose-fermenting yeast, notably Pichia stipitis CBS 6054, ferment xylose to ethanol with reasonable yield and productivity; however, these yeast strains are inhibited by compounds generated during pretreatment and hydrolysis of the lignocellulose material [31]. Filamentous fungi tolerate inhibitors but are too slow for a competitive industrial process. Therefore, efforts have predominantly been made to obtain recombinant strains of bacteria and yeast able to meet the requirements of industrial lignocellulose fermentation (Figure 2).

Figure 2. Strains that are metabolically engineered for ethanol production from pentoses. In essence, either the tail end, as in Escherichia coli and Klebsiella oxytoca, or the front end of metabolism, as for Saccharomyces cerevisiae and Zymomonas mobilis, have been engineered. Abbreviation: rec, recombinant.

Review TRENDS in Biotechnology Vol.24 No.12 551

Pentose-fermenting Escherichia coli [32] and Klebsiella oxytoca [3] have been generated by introducing ethanologenic genes from Zymomonas mobilis (Figure 2). At the same time, the first xylose-fermenting S. cerevisiae strain wasgeneratedthroughtheintroductionofgenesforxylosemetabolizing enzymes from P. stipitis [34] (Figure 2). Later xylose-fermenting strains of S. cerevisiae were constructed by introducing the genes encoding xylose isomerase from the bacterium Thermus thermophilus [35] and the anaerobic fungus Piromyces sp. [36], respectively. For xyloseusing S. cerevisiae, high ethanol yields from xylose also require metabolic engineering strategies to enhance the xylose flux. This was independently demonstrated in an SSF set-up for corn stover, where glucose and xylose were fermented simultaneously [37], and in a recombinant strain with increased carbon flux [38]. Z. mobilis also efficiently produces ethanol from the hexose sugars glucose andfructosebutnotfrompentosesugars,althoughaxylose fermenting Z. mobilis was generated by introducing a xylose-metabolizing pathway from E. coli [39]. More recently, the obligatory anaerobic bacterium Thermoanaerobacteriumsaccharolyticum hasbeengeneticallyengineered for improved ethanolic fermentation (Joe Shaw et al., oral presentation, Nashville, 2006).

E. coli and K. oxytoca naturally metabolize arabinose, such that the ethanologenic strains ferment all lignocellulose-derived sugars [40]; furthermore, xylose- and arabinose-fermenting strains of Z. mobilis have been constructed [41]. Because yeast only ferment arabinose to ethanol in rich media [42,43], S. cerevisiae has been engineered for arabinose use by introducing both bacterial [4] and fungal genes [45] encoding arabinose-metabolizing enzymes, where the fungal approach did not result in appreciable arabinose fermentation. The functional arabinose-metabolizing pathway has recently been integrated into the diploid xylose-fermenting S. cerevisiae strainTMB 3400, and co-usage of xylose and arabinose has been demonstrated [46].

In addition to being able to ferment both hexose and pentose sugars, the fermenting microorganisms must do this in the presence of inhibiting compounds such as weak acids, furan derivatives and phenolics [47]. Detoxification by chemical or physical methods before fermentation reduces the concentration of inhibitors and improves the performance in the fermentation step. Treatment with

Ca(OH)2 (so-called ‘overliming’) and ion-exchange treatment increased the fermentability significantly [48]; how- ever, this was at the expense of increased process cost and sugar loss [49]. Furthermore, Ca(OH)2 is unlikely to be accepted in a full-scale ethanol plant, where precipitation of calcium salts might foul distillation columns, evaporators and heat-exchanger surfaces [50,51].

In ethanolic yeast fermentation, in-situ biological detoxification occurs when carbonyl compounds – furans and phenolics – are reduced to the corresponding alcohols [52,53], which are less inhibitory to yeast [54]. This phenomenon has been successfully exploited in online probing process control in fed-batch fermentation, by adjusting the feed rate of hydrolysate to the intrinsic capacity for inhibitor conversion of the yeast [5]. In such cases, the fermentation productivity will be a function of the inhibitor concentration, the conversion capacity of the yeast and the quality of the process control. Only recently it was demonstrated that the overexpression of a gene encoding alcohol dehydrogenase activity improved the fermentation of a 5-hydroxymethyl furfural-containing medium [56].

The competitiveness of fuel ethanol research makes strain comparisons difficult. Literature data are hampered by incomplete reporting, differences in fermentation conditions (including media composition and oxygenation),

Table 1. Latest fermentation results on hydrolysate from various recombinant and native xylose-fermenting organisms.

Organism Hydrolysate Detoxification Fermentation mode Yield g/g initial sugar

Fermentation time(hours) Sp. productivity g/g cells.h


Escherichia coli KO11 Bagasse hemicellulose1 + n/a 0.49 n/a 0.37 [6]

E. coli FBR5 Corn stover + n/a 0.46 n/a 0.214 [6]

Zymomonas mobilis 8b6 Corn stover7 +3 Batch 0.42 n/a n/a [6] [69]

Pichia stipitis8 Wheat straw +3 n/a 0.41 n/a 0.4 [6,70]9 P. stipitis CBS 5773 Spent sulfite liquor +10 Continuous 0.35 - 0.05 [6] Saccharomyces cerevisiae 424A (LNFST)

S. cerevisiae TMB3006 Spruce Fed batch 0.37 n/a 0.6 [6] S. cerevisiae TMB3400 Spruce Fed batch 0.43 n/a 0.25 [6] Corn stover14 SSF, Batch15 0.3316 96 0.036 [37]

SSF, Fed batch15 0.3016 96 0.076Supplemented with 2.5% (w/v) corn steep liquor; supplemented with 5% (w/v) corn steep liquor; overliming; neutralization by anion exchange; membrane pervaporation; acid-tolerantderivativeofZM4/Ac (pZB5); 80%hydrolysatesupplementedat100 g/lglucose; adaptedstrainofNRRLY-7124; calculatedfromNigam2001 [70]; rotoevaporated spent sulfite liquor; optimized pH-controlled liquid hot water pretreatment; Semi-anaerobic fermentations with 10 g/l yeast extract added; Calculatedwith the giveninitial celldry weightof 8.5 g/l; 0.05% waterinsoluble solid; 1 g/l yeastextract added; %of theoreticalbased on the glucoseandxylose content in the raw material, (i.e. taking into account both hydrolysis and fermentation yields). Abbreviations. n/a, not available; SHF, separate hydrolysis and fermentation; SSF, simultaneous saccharification and fermentation.

552 Review TRENDS in Biotechnology Vol.24 No.12 vastly different raw materials, and pretreatment and hydrolysis conditions [29]. Table 1 summarizes the most recently published results on the fermentation of lignocellulose hydrolysate with natural and recombinant bacteria and yeast. In essence, so far only recombinant S. cerevisiae strains have been able to ferment xylose in non-detoxified hydrolysates, where the fed-batch technology also permits the fermentation of extremely inhibitory softwood hydrolysates.

Moving on from research to a commercial process The estimated cost of producing ethanol from cellulosic materials varies widely between investigations, as shown both in some earlier review papers [57,58], with production costs in the range of 0.28 to 1.0 US$/l ethanol, and in more recent techno-economic evaluations [13,59,60]. However, most cost estimations are based on laboratory-scale and, to some extent, pilot-scale data for individual process steps and should be treated with caution. The cost of raw material, which varies considerably between different studies (US$2–US$61 per metric ton dry matter), and the capital costs, which makes the total cost dependent on plant capacity, contribute most to the total production cost. The cost for hydrolysis, particularly for the enzymatic process, is also a major cost contributor.

Raw material cost is reduced by using the whole crop for products and co-products. High ethanol yield requires complete hydrolysis of both cellulose and hemicellulose with a minimum of sugar degradation, followed by efficient fermentation of all sugars in the biomass. In the shortterm, co-products are likely to be used for the production of fuel, heat and electricity; however, in the long term, bioethanol technology will form the basis for the sustainable production of commodity chemicals and materials in future biorefineries. High co-product yield requires reduced energy demand for ethanol production. This is achieved when high solids concentrations (Figure 3) are combined with integration of energy-intensive process steps (e.g. pretreatment, distillation, evaporation and drying). In SSF, 12 % WIS (water-insoluble solids) result in an ethanol concentration >4 wt-% (weight-%; kg ethanol per 100 kg solution), which is necessary to reduce the energy demand in the distillation steps (A. Wingren, PhD thesis, Lund University, 2005). Further reductions in the energy demand can be obtained by recycling certain process streams, to minimise the amount of fresh water used [61]. However, high solids concentrations and recycling of process streams increase the concentration of compounds that are inhibitory to enzymatic hydrolysis and fermentation, necessitating detoxification or fed-batch technology, as described previously. It also results in high viscosity, which limits mixing and pumping.

Process integration reduces the capital costs. In the separate hydrolysis and fermentation (SHF) process, cellulose is first hydrolyzed to glucose and then glucose is fermented to ethanol. The primary advantage of SHF is that hydrolysis and fermentation occur at optimum conditions; the disadvantage is that cellulolytic enzymes are end-product inhibited so that the rate of hydrolysis is progressively reduced when glucose and cellobiose accumulate [24]. Product inhibition was the rationale for the first report on simultaneous saccharification and fermentation (SSF) of cellulose [62]: in SSF, hydrolysis and fermentation occur simultaneously in the same vessel, and the end-product inhibition of the enzymes is relieved because the fermenting organism immediately consumes the released sugars. Furthermore, the fermentation seems to decrease the inhibition of the enzymes by converting some of the toxic compounds present in the hydrolysate [24]. This increases the overall ethanol productivity, the ethanol concentration and the final ethanol yield [12,63] (Figure 4).

More recently, the SSF technology has proved advantageous for the simultaneous fermentation of hexose and pentose sugars (so called SSCF). In SSCF, the enzymatic hydrolysis continuously releases hexose sugars, which increases the rate of glycolysis such that the pentose sugars are fermented faster and with higher yield [37].

(Parte 1 de 2)