Fuel ethanol industry has advanced in SSF technology by incorporating the yeast propagation (from active dry yeasts) in the fermenter during initial saccharifi­cation, a process called simultaneous saccharification, yeast propagation, and fermentation (SSYPF), as indicated in Figure 9.1. High sugar concentrations are not achieved in the fermenter avoiding the inhibition of the enzymatic hydro­lysis that is characteristic for the amylases. Due to this, the bacterial growth is inhibited because of the lack of substrate caused by the immediate conversion of glucose into ethanol. Maintaining the pH, nutrients, and sterility in relation to bacteria, the complete conversion of available starch into ethanol is accom­plished. SSYPF technology has been utilized in several plants in North America mainly employing corn, milo (a variety of sorghum), and wheat (Madson and Monceaux, 1995).

The objective of this configuration is to increase the contact time between the mash and yeasts in order to reduce the bacterial growth. This allows reaching higher yields, an earlier ethanol production (that also reduces the contamination), and reducing the need of handling large amounts of yeasts (Novozymes & BBI International, 2005). At the beginning of an SSYPF process, pH is adjusted to 5.2 for favoring the growth of microorganisms and, as the cultivation goes forward, pH is diminished to 4.5 at the end of fermentation (Madson and Monceaux, 1995).

11.2.1 Configuration Involving the Separate Hydrolysis

and Fermentation (SHF) of Corn Starch

Starch is a high yield feedstock for ethanol production. Glucose is obtained by the hydrolysis of starch. Then, the glucose solution undergoes fermentation toward ethanol. From each 100 g of starch, 111 g of glucose theoretically can be obtained, which implies a stoichiometric ratio of 9:10. The output/input ratio of energy for

corn ethanol is in the range of 1.1 to 1.2 (Prakash et al., 1998; Sanchez and Cardona, 2005). Fuel ethanol production from materials with a high content of starch needs some additional process steps compared to the process that employs sucrose-based materials as the feedstock. The process flowsheet comprises a pretreatment step of the starchy materials to make starch more susceptible to enzymatic hydrolysis. In this step, known as liquefaction, starch is partially hydrolyzed at a high tempera­ture. In the following step of saccharification, liquefied starch undergoes a deeper hydrolysis where fermentable sugars (glucose) are obtained (Figure 11.5). After glucose fermentation, the process does not significantly differ from that one that employs materials with a high content of sucrose. Nevertheless, depending on the specific type of employed starchy feedstock, certain co-products used for animal feed can be produced during the evaporation of stillage.

When cereals are used as the feedstock for producing fuel ethanol, the raw material enters the process as grains, which should undergo cleaning and milling. Either wet-milling or dry-milling processes can carry out grain milling of such cereals as corn, wheat, and barley. The wet-milling process implies that only the starch enters fuel ethanol production process, as discussed in Chapter 4, Section 4.2. In this process, all the components of the kernel should be separated prior to the cooking step. These components represent value-added products that are used for food and feed. Moreover, part of the produced starch can be deviated to the production of sweeteners, such as the high fructose corn syrup (HFCS). During the dry milling of grains, the whole kernel enters the ethanol production line, which means that all its components are processed along with starch. Nonutilized components of the kernel are built up in the bottoms of the first distillation col­umn and are concentrated to form a product utilized as animal feed. In general, the liquefaction, saccharification, and fermentation steps are the same for both types of technologies.

The overall production process of bioethanol from corn by the dry-milling technology includes the breakdown of this polysaccharide to obtain an appropriate

concentration of fermentable sugars, which are transformed into ethanol by yeasts (Figure 11.6). After washing, crushing, and milling the corn grains, the starchy material is gelatinized in order to dissolve the amylose and amylopectin (cooking step). This process is accomplished with the help of a jet cooker working at 110°C. In dissolved form, starch is accessible for enzymatic attack in the following step, the liquefaction that is carried out at 88°C. This step is considered a pretreatment process because of the partial hydrolysis of the starch chains using thermostable bacterial a-amylase that yields a starch hydrolyzate with a hydrolysis degree of approximately 10%. The obtained hydrolyzate has a reduced viscosity and con­tains oligomers, such as dextrins.

Then, this liquefied starch enters the saccharification process where it is hydro­lyzed by microbial glucoamylase to produce glucose. This process operates at 60°C. The saccharified starch is cooled and sent to the next step where it is fer­mented by the yeast S. cerevisiae and converted into ethanol at 30°C. Fermentation gases, mostly CO2, are washed in an absorption column to recover more than the 98% of the volatilized ethanol from the fermenter and sent to the first distillation column. The culture broth containing 11% (by weight) ethanol is recovered in a separation step consisting of two distillation columns. In the first (concentra­tion) column, aqueous solutions of ethanol are concentrated to 50%. In the second (rectification) column, the concentration of the ethanolic stream reaches a com­position near the azeotrope (95.6%). The dehydration of the ethanol is achieved through adsorption in a vapor phase with molecular sieves. The stream obtained during the regeneration of molecular sieves containing 70% ethanol is recycled to the rectification column.

The stillage from the concentration column is evaporated and the obtained sol­ids are separated by centrifugation. These solids are dried for producing the dis­tiller’s dried grains with solubles (DDGS), a co-product used for animal feed due to its high content of proteins and vitamins. The remaining liquid or thin stillage is evaporated in a double effect evaporator. The obtained syrup is combined with the DDGS and dried. The condensed water from the evaporators is recirculated to the liquefaction stage, while the bottoms of the rectification column and one frac­tion of the thin stillage (backset) are recycled back to the saccharification step.

The absolute majority of methods for conceptual process design have been devel­oped for processes of basic chemical and petro-chemical industries. Most of these approaches are especially oriented to the design of separation schemes, particularly distillation trains. For this reason, there is a paramount interest in the application of these methodologies to biotechnological processes, which are of a complex and highly multicomponent nature. None of the described approaches can be applied in a generic way to any type of synthesis problem, even more so in the case of biological processes. Considering this, the application of a minimum of two strate­gies or synthesis procedures is required to undertake the design of processes for fuel ethanol production with a high technoeconomic and environmental perfor­mance. In this way, a higher amount of possibilities can be covered allowing the creativity during the design process that is the source of innovation. Therefore, the employment of synthesis strategies that systematically compile and utilize the accumulated knowledge around a research subject, such as fuel ethanol, directly contributes to screening and filtering the best alternative process configurations.

Further, some of the approaches mentioned above were applied to the genera­tion of different process flowsheets with a high technical, economic, and envi­ronmental performance. Thus, both the hierarchical decomposition method and optimization-based strategies were used for the main process steps needed for ethanol production. For specific process steps, such as the synthesis of separation and ethanol dehydration scheme, the analysis of the statics and the exergy balance were also employed. These issues will be discussed in following chapters.

The chemical methods of detoxification are based on the addition of certain chemical compounds that vary the conditions of the aqueous medium provok­ing changes in the pH, formation of precipitates, or the direct transformation of the toxic compounds. Among these methods, the ionic interaction of the ionic exchange resins can be included in this group of detoxification methods. The most employed chemical methods of detoxification are shown in Table 4.7. By neutral­ization, the solubility of many inhibitory substances is changed. This allows their removal by a later filtration or adsorption. However, the addition of alkali up to very high pH values (alkaline detoxification) leads to the formation of a significant amount of precipitate composed by calcium salts (if lime is used), which entrains the inhibitory compounds or causes them to settle. In addition, many inhibitors are unstable at pH higher than 9. Alkaline treatment is considered one of the best detoxification methods since a high percentage of substances such as furalde- hydes and phenolic compounds can be removed by this method, improving the fermentability of the resulting liquid medium especially when biomass hydrolyz- ates pretreated with dilute acid are employed (Persson et al., 2002a). The addition of calcium hydroxide (overliming) or ammonium has shown better results than the use of sodium or potassium hydroxide. Some methods to determine the optimal

Physical Methods for Detoxification of Pretreated Biomass

Remarks Reduction of acetic acid and phenolic compounds in nonvolatile fraction; roto — evaporation 93% yield of ref. fermn.; removal: 54% acetic acid, 100% furfural, 29% vanillin; roto-evaporation Solv: diethyl ether (solv.); yield comparable to ref. fermn.; removal of acetic, formic, and levulinic acids, furfural, HMF Solv.: ethyl acetate; 93% yield of ref. fermn.; removal: 56% acetic acid, 100% furfural, 100% vanillin, 100% hydroxybenzoic acid Solv.: ethyl acetate; removal of low molecular phenolic compounds Solv.: ethyl acetate; EtOH yield (SSF): detoxified hz. 0.51 g/g, undetox, hz. 0 g/g; high degree of phenolic removal	References Palmqvist and Hahn-Hagerdal (2000a)
Palmqvist and Hahn-Hagerdal (2000a) Palmqvist and Hahn-Hagerdal (2000a)
Palmqvist and Hahn-Hagerdal (2000a)
Palmqvist and Hahn-Hagerdal (2000a) Cantarella et al. (2004)
Continued

Chemical Methods for Detoxification of Pretreated Biomass

References

Precipitation or removal of toxic compounds; 10% lower yield for Pichia sp.

EtOH yield (SSF): detoxified hz. 0.86 g/g, undetox, hz. 0 g/g Yield comparable to ref. fermn.; 20% removal of furfural and HMF

Removal of acid acetic, furfural and part of phenolic compounds

7.5% lower yield for Pichia sp.

39% reduction in fermentation time Saha et al. (2005a)

Reduction in fermn. time: SSF Saha et al. (2005b)

-18%, SHF-67%

Removal: 51% furfural, 51% HMF, Martinez et al. (2000, 2001) 41% phenolic compounds, 0% acetic acid; overliming at 60°C or 25°C, at high temperature, the required amounts of lime and acid are reduced

Continued

Chemical Methods for Detoxification of Pretreated Biomass

Source: Adapted from Sanchez, O. J., and C. A. Cardona. 2008. Bioresource Technology 99:5270-5295. Elsevier Ltd.

Note: Reference fermentation (ref. fermn.) refers to fermentation carried out in a glucose-based medium without inhibitors; hz = hydrolyzate.

amount of lime to be added in dependence on the acid content of the hydrolyzate have been developed (Martinez et al., 2001). The positive effects of alkaline treat­ment on the hydrolyzate fermentability cannot be explained only by the removal of inhibitors. It has been postulated that this detoxification method may have possible stimulating effects on ethanol-producing microorganisms (Persson et al., 2002a).

Ionic exchange also has been studied as a detoxification method showing a high efficiency for removing inhibitors. This method can be considered as a special case of adsorption because ionized groups of the ionic exchange resin (the adsor­bent) interact electrostatically with the charged molecules of inhibitors. In par­ticular, some anionic exchange resins are used to eliminate phenolic compounds as a consequence of the strong bonds formed between quaternary ammonium groups of the resin (positively charged) and phenols (negatively charged). The rest of substances that do not interact with the resin pass through the adsorbent leading to a detoxified hydrolyzate. Besides the high resin cost, one drawback of this method lies in the fact that the content of fermentable sugars in the hydrolyz — ate can be reduced (Oliva, 2003). In the model process developed for NREL, the ionic exchange was proposed as a detoxification method for ethanol production process using poplar wood as the feedstock (Wooley et al., 1999). The biomass pretreated with dilute sulfuric acid at 190°C and high pressure is cooled by flash­ing, which removes 61% furfural and HMF as well as 6.5% of the acetic acid released from hemicellulose. The liquid fraction of the pretreated biomass is sent to an ionic exchange column whose effluent undergoes overliming to enhance the detoxification efficiency. Then, sulfuric acid is added to remove the calcium and suspended solids by forming a precipitate of calcium sulfate (gypsum). This design was based on data obtained at pilot scale level using a column with a diam­eter of 20 mm and a length of 1 m containing Amberlyst A20, a weak base resin. The regeneration of the resin is accomplished by passing the eluent (ammonium) through the column. Xie et al. (2005), in turn, demonstrated the successful detox­ification of corn stover hydrolyzate using a polymeric adsorbent without the need of a subsequent alkalinization. In contrast, the newer model process for ethanol production for corn stover designed for NREL only suggested the overliming as the detoxification method (Aden et al., 2002).

Different methods of detoxification that combine physical and chemical prin­ciples have been proposed, such as the neutralization with CaO or Ca(OH)2 fol­lowed by the addition of activated carbon and filtration to remove the acetic acid (Olsson and Hahn-Hagerdal, 1996). For lignocellulosic materials pretreated by pyrolysis and hydrolyzed with dilute acid, the utilization of several adsorbents, such as activated carbon, diatomite, bentonite, and zeolites, after the treatment by neutralization has been also studied (Yu and Zhang, 2003).

As mentioned above, the liquid fraction resulting from lignocellulosic biomass pretreatment can be unified with the cellulose hydrolyzate obtained after the enzymatic treatment of the cellulose contained in the solid fraction coming from the mentioned pretreatment process. The liquid stream produced contains all the soluble sugars derived from the biomass, mostly glucose and xylose. This stream can be fermented by microorganisms able to convert these two sugars into etha­nol. For this, recombinant bacteria can be used. Leksawasdi et al. (2001) employed an engineered strain of Z. mobilis able to co-ferment hexoses and pentoses to
process a solution containing a mixture of glucose and pentose. In addition, they developed and experimentally tested an accurate mathematical description that considers substrate limitation by the two sugars, substrate inhibition by both sug­ars, and ethanol inhibition. For this, it employs the concepts of threshold ethanol concentration for which inhibition of growth begins, and maximum ethanol con­centration for which biomass growth becomes zero. Inhibition constants for the substrates are taken into account for considering glucose and xylose concentra­tions inhibiting both cell growth and ethanol biosynthesis as a result of catabolic repression. Kinetic expressions can be found in the work of Leksawasdi et al. (2001). This description was used by the authors of this book to consider co­fermentation processes during process synthesis procedures. The equations of the model are as follows:

r = [ar, 1 + (1 — a )rx, 2]X

where rx is the overall cell growth rate, rxj and rxj are the cell growth rate from glucose and xylose, respectively; rsj and rsj are the glucose and xylose consump­tion rates, respectively; rp is the overall ethanol production rate; rp1 and rp1 are the ethanol production rates from glucose and xylose, respectively; X, Sj, S2, and P are the concentrations (in g/L) of cell biomass, glucose, xylose, and ethanol,

respectively; a, ^max, j, Amax,2, qs, max, j, qs, max,2, qp, max, j, qp, max,2, Ksx, j, Ksx,2, Kss, j, Kss,2,

K K P P P P P P K K K K K K P

sp 1 sp 2 mx,1 mx,2 ms,1 ms,2 mp,1 mp,2 ix,1 ix,2’ ■ is,1 is,2 ip,1 ip,2 ix,1’

Pix, Pisj, Pis,:2, Ppj, Pip,2 are the kinetic parameters.

Case Study 7.1 Modeling of Co-Fermentation Fermentation

Based on the model of Leksawasdi et al. (2001), the simulation of alcoholic fer­mentation from biomass was performed with initial glucose concentration of 100 g/L and initial xylose concentration of 50 g/L. These concentrations approximately correspond to those of lignocellulosic hydrolyzates. Inhibition of growth rate can be observed from Figure 7.9 due to relatively high amounts of ethanol in the broth, as can be seen after 25 h. The use of more concentrated culture media leads to the underutilization of expensive feedstocks, which cannot be transformed into ethanol despite their availability in the broth. According to the model, when a medium con­taining up to 400 g/L of fermentable sugars is employed, an ethanol concentration of about 71.2 g/L is reached only after 80.5 h of cultivation. An ethanol concentra­tion of 24.7 g/L is attained at 48 h remaining more than 347 g/L of substrate (com­pared with data of Figure 7.9 for a medium with a lower substrate concentration). Therefore, this model proved its suitability for describing the complex inherent phenomena of the co-fermentation of lignocellulosic hydrolyzates.

Modeling of extractive fermentation processes for fuel ethanol production plays a crucial role when different process alternatives are being analyzed in the framework of conceptual process design, especially when process synthe­sis procedures are applied. Considering that technologies for ethanol production from lignocellulosic biomass are not currently mature, the analysis of different options intended to reduce lignocellulosic ethanol production costs is a task of great significance. In a previous work (Sanchez et al., 2006), the co-fermentation process integrated by means of reaction-separation approach was assessed. The objective of that work was to model the extractive fermentation process for fuel ethanol production from lignocellulosic biomass analyzing cultivation kinetics coupled with liquid extraction.

To describe the continuous process of extractive fermentation for fuel ethanol production, и-dodecanol was selected as an extracting agent (solvent). A feed aque­ous stream containing sugars and nutritive components is added to a CSTR where a solvent stream is continuously fed as well. Fed sugars are generated during the pretreatment of lignocellulosic biomass in which major polysaccharides are broken down into elementary sugars like hexoses (glucose) and pentoses (mainly xylose). Formed sugars are converted into ethanol in the reactor. Ethanol is distributed between aqueous and organic (solvent) phases, diminishing its concentration in the culture aqueous broth and allowing reduction of the product inhibition effect on the microorganisms. The ethanol-enriched solvent phase is continuously removed from the reactor through a decanting unit. This stream is sent to a flash unit in order to recover the obtained ethanol and to regenerate the solvent, which can be recycled to the CSTR.

With the aim of developing a rigorous model that describes both the fermen­tation and liquid-liquid extraction processes, the kinetics cultivation is coupled with an extraction model. The liquid-liquid equilibrium was described through an algorithm based on the mass balance equations developed for the isothermal flash in the case of two liquid immiscible phases. Activities of components in each phase were calculated by means of the UNIFAC model, since this model has demonstrated to be the most appropriate for description of equilibrium when two or more liquid phases are present for this case. This algorithm was integrated into the ModELL-R software, which was especially designed by the research group to which the authors of this book belong. The software couples two convergence algorithms (Newton-Raphson and False Position Method) in order to calculate the liquid fraction of each phase. ModELL-R was developed in Delphi package v7.0 (Borland Software Corp., Austin, TX, USA).

The kinetic model of alcoholic fermentation was taken from Leksawasdi et al. (2001; see Chapter 7, Case Study 7.1). This model describes the simultane­ous consumption by a recombinant strain of Z. mobilis of two main substrates contained in the lignocellulosic hydrolyzates: glucose and xylose. The following assumptions were considered for the development of the overall model of extractive fermentation:

• The substrate uptake, biomass formation, and product biosynthesis are car­ried out only in the aqueous phase; hence, no reactions occur in the organic (solvent) phase.

• Ethanol is the main component migrating to the solvent phase; small amounts of water can migrate to the organic phase depending on the solvent.

• No migration of substrates and biomass to the solvent phase takes place.

• Solvent is biocompatible with the microorganisms and does not have effect on the fermentation process.

• Stirring of bioreactor ensures total mixing between liquid phases and does not produce damage to the growing cells.

The configuration corresponding to continuous extractive fermentation involves the continuous feeding of culture medium and solvent to the reactor and the continuous removal of the liquid aqueous phase and solvent phase from the reactor in a separate way with the help of a decanter (see Figure 9.14). In this case, the flowrate (in L/h) of influent aqueous stream (FA) is greater than the flowrate of effluent aqueous stream (QA) because of the migration of ethanol to the sol­vent phase. Similarly, the flowrate of influent solvent stream (FE) is less than the flowrate of effluent solvent stream (QE). Mass balance equations representing this process are as follows:

where rX is the cell growth rate (in g/(L x h)), rS1 and rS2 are the glucose and xylose consumption rates, respectively (in g/(L x h)), and rP is the ethanol formation rate (in g/(L x h)); X, S1, S2, and P are the concentrations of cell biomass, glucose, xylose, and ethanol in the aqueous effluent from the bioreactor (in g/L), and X0, S10, S20, and P0 are the corresponding concentrations in the aqueous feed stream (in g/L); P* and P0* are the ethanol concentrations in the solvent effluent and in the solvent feed streams, respectively (in g/L). The balance can be applied for the case when the solvent contains small amounts of ethanol as a result of noncomplete regeneration of the extracting agent. This system of equations is nonlinear due to the equations describing the process kinetics and is solved through multivariate Newton-Raphson algorithm. For this, a constant solvent volume/aqueous volume ratio is assumed. The relationship between both effluent flow rates is fixed. The ethanol concentration in the solvent phase (P*) that is in equilibrium with ethanol concentration in the aqueous phase is determined using the distribution coefficient &EtOH as shown by equation (9.18), which is calculated by the algorithm for liquid — liquid equilibrium. The determination of all variables involved in the model is per­formed using the algorithm shown in Figure 9.15 incorporated into the software ModELL-R. For specified inlet aqueous dilution rate (DAi = FA/VA) and solvent feed flow rate/aqueous feed flow rate ratio (R = FE/FA), the program requires concentra­tions of cell biomass, substrates, and ethanol in feed streams.

The simulation of alcoholic extractive fermentation from biomass was performed with a glucose concentration of 100 g/L and an xylose concentration of 50 g/L in the feed aqueous stream. These concentrations correspond to those of lignocel — lulosic hydrolyzates. The behavior of this process for R = 2 in dependence of inlet dilution rate (DAi) is shown in Figure 9.16. The cell washout occurs at dilution rates near 0.33 h-1. These results were obtained for a solvent volume/aqueous volume ratio (VE/VA) of 2. Higher ethanol productivities are found in the range 0.25 to 0.30 h-1. In order to elucidate the best operating value of DAi, GAMS software (General Algebraic Modeling System, GAMS Development Corp., Washington, DC, USA) was used for maximizing total ethanol productivity. For this, a simple linear rela­tionship for Equation (9.18) was considered. The optimal DAi was 0.265 h-1.

The coupled algorithm was used for process simulation varying the solvent feed flow rate/aqueous feed flow rate ratio (R) for an inlet dilution rate of 0.265 h-1. Best results were obtained for values greater than 4 that correspond to an increased

amount of consumed substrates. Both total productivity and productivity for etha­nol recovered from solvent phase are approaching constant values. For R higher than 8, the model predicts the formation of homogeneous mixture without extrac­tion. The simulation of this process modifying both R and DAi shows that the zone of manipulating variables with higher productivities (inlet concentrations of glu­cose and xylose in the aqueous stream of 100 g/L and 50 g/L, respectively) corre­sponded to dilution rates near washout conditions, and to higher values of R within the range 1.29 to 7.9. If the concentration of both substrates in the feed stream is varied, the problem becomes more complex. Physically, the increase in substrate can be achieved as a result of evaporation of initial hydrolyzate obtained from bio­mass pretreatment. For this reason, the proportion of glucose and xylose in the inlet aqueous stream should be constant and equal to 2:1. Best values of total productiv­ity and ethanol productivity recovered from solvent phase correspond to an inlet concentration of total sugars of about 600 g/L. The simulation was carried out until the concentration of sugars was less than or equal to 600 g/L, which corresponds to maximum solubility of these sugars in water.

In this way, a model describing extractive co-fermentation was developed. This model allows for doing the analysis of this reaction-separation integration process in order to consider it in subsequent process synthesis methodologies. This is espe­cially valuable for such process synthesis approaches as the hierarchical decompo­sition (see Chapter 2) that requires the development of proper models for simulation of alternative configurations during each hierarchical level of analysis.

DA(i) I1’/"]

(b)

FIGURE 9.16 Continuous extractive co-fermentation using я-dodecanol. Effect of inlet aqueous dilution rate (DAi) on (a) effluent concentrations of glucose (51), xylose (S2), etha­nol in aqueous phase (P), and ethanol in solvent phase (P*); (b) total ethanol produc­tivity (PrT), productivity for ethanol recovered from aqueous phase (PrA), productivity for ethanol recovered from solvent phase (PrE), and effluent concentration of cells (X). Concentration of sugars in feed aqueous stream: glucose, 100 g/L; xylose, 50 g/L.

In a previous work (Sanchez et al., 2006), a preliminary approximation to the pro­cess synthesis of ethanol production from lignocellulosic materials employing the optimization-based approach was presented. The analyzed system comprised the step of biological transformation of the pretreated feedstock through different tech­nological options (separate hydrolysis and fermentation, simultaneous saccharifi­cation and fermentation, simultaneous saccharification and co-fermentation) and the step of ethanol dehydration using distillation. The initial stream entering the system contains the main components that are formed after the pretreatment using dilute acid, i. e., cellulose, pentoses (mainly xylose), glucose, lignin, and water. The system should process this stream in such a way that the final product stream has an ethanol content greater than 99.5% wt. For this preliminary study, the wastewater treatment step was not considered.

For tackling such a complex process as the bioethanol production, the Jacaranda synthesis package was employed, which has been described elsewhere (Fraga, 1998; Fraga et al., 2000). It has been successfully applied to the preliminary design of a hydrofluoric acid plant (Laing and Fraga, 1997), the generation of optimal downstream processing flowsheets of bioprocesses (Steffens et al., 1999b), and the process synthesis for the microbial production of penicillin (Steffens et al., 1999a). Any optimization-based strategy for process synthesis requires the implementa­tion of models of process units. In this case, the biological transformations were described by kinetic models considering the cellulose hydrolysis, glucose forma­tion and consumption, cell growth, and ethanol biosynthesis (South et al., 1995; see Chapter 7, Case Study 7.2). When co-fermentation using recombinant bacteria was

taken into account, the model reported by Leksawasdi et al. (2001) was used (see Section 7.2.2). For the calculation of distillation units, the Fenske-Underwood — Gilliland (FUG) method was utilized considering the presence of binary azeo­tropes in the system ethanol-water. The separation section of the process, used to generate 99.5% pure ethanol, consisted of distillation units alone. As there was an azeotrope formed by water and ethanol, an extractive distillation step was used with ethylene glycol as the solvent.

The objective function used, in this case, is net revenue defined as the value of the ethanol produced minus the annualized cost of the process, which is a function of both capital and operating costs. For the continuous bioreactors, operating and capital costs are directly related to the residence time. The capital costs of distilla­tion units are related to the vapor velocity inside the columns and to the number of stages, and the operating costs are linked to the energy consumption (mainly heat duty). The problem posed to Jacaranda consists of a superstructure, which is shown in Figure 11.14. The reaction section has a choice of three paths, the SSF reactor, the SSCF reactor, and the combination of cellulose hydrolysis followed by hexose fermentation (SHF). The separation section consists of three distillation steps: con­centration column, extractive column, and recovery column with a recycle of the solvent to the extractive column. The superstructure is the basis for an MINLP model. This model has some characteristics that make it difficult to solve, such as the physical properties models used (NRTL [nonrandom two-liquid model] in this case) and the equations for the concentrations in the reactors as a function of resi­dence time. These are present in the optimization problem as equality constraints and are difficult to satisfy. Furthermore, the use of different hot and cold utilities was allowed to meet the heating and cooling demands of any process alternative. Using discrete utilities means that the objective function is discontinuous even as a function of only the real valued variables. Furthermore, the capital cost function for the distillation units uses integer values for the number of stages determined by the FUG procedure, also leading to discontinuities in the objective function. The result is that the overall optimization problem is not solvable using standard mathematical programming approaches.

Jacaranda provides access to a number of optimization procedures including direct search methods (Kelley, 1999) and stochastic methods, such as genetic algo­rithms (Goldberg, 1989) and simulated annealing (van Laarhoven and Aarts, 1987).

In this case study, it was decided to use the genetic algorithm (GA) approach. The GA uses a replacement policy for the population at each generation, with an elite size of 1, a mutation rate of 10%, a crossover rate of 70%, and a roulette wheel selection procedure. The fitness function is based on the objective function value directly with infeasible solutions discarded if they arise (which they do with a fre­quency of approximately 5 to 6%).

For this first attempt at automated design for the production of ethanol from biomass, the number of degrees of freedom was limited. Specifically, the residence times of each reactor and the top and bottom key component recoveries in each distillation column were selected as the manipulating variables. Therefore, four residence time variables and six recovery variables were manipulated. The super­structure makes use of two binary variables for identifying the path taken through the reaction section of the process.

The results obtained identify the SSCF configuration as the best performing for the given process. This is reasonable given the high degree of integration achieved with this configuration, which makes possible the immediate consumption of the glucose formed during the cellulose hydrolysis. In this way, the inhibition of cellulose-degrading enzymes (cellulases) is avoided. In addition, the utilization of xylose allows an increase in the content of fermentation sugars and, therefore, in the overall amount of produced ethanol. This enhanced utilization of the feedstock is not characteristic for the SSF process. The SHF option implies the utilization of an additional bioreactor (the enzymatic hydrolysis and the fermentation are car­ried out in different units), which involves the increase in the capital costs for this configuration. Jacaranda allowed the determination of the values of the operating parameters corresponding to the separation section. In particular, the make-up of ethylene glycol and the recycle stream flow rate are determined automatically.

Early results demonstrate that the genetic algorithm used by Jacaranda handles the complexity of the problem design robustly with respect to the numerical diffi­culties that may arise. The solutions obtained show variability in the technological option. From 10 different runs, three of the solutions corresponded to SSCF con­figurations (two of them with the best values of the objective function), six solutions to the SSF process, and one solution to the SHF configuration.

Undoubtedly, the development of this approach will make possible the synthesis of technological flowsheets considering the structure of the system on a mathemati­cal programming basis. The complementation with tools of a knowledge-based approach will allow gaining a deeper insight of the overall process needed for the synthesis of technological configurations with increased performance.

In the past, and as a consequence of the concerns generated by the atmospheric pollution in cities, the better combustion of gasoline through the incorporation of oxygen into its composition has been pursued. Again, the United States boosted the development of new additives for the world automobile industry. In this sense, the signing of the Clean Air Act in 1970 and the Clean Air Amendments in 1990 introduced reformulated gasoline as an answer to the need for diminishing the
air pollution levels in the biggest cities of the United States. This type of gasoline should have a minimum 2% (by weight) oxygenate and benzene levels less than 1% (by volume; Nadim et al., 2001). The employing of oxygenates is aimed at reducing the atmospheric contamination (smog during summer, CO in winter, and toxic emissions throughout the year) because of the better combustion toward CO2 due to the involvement of one atom of oxygen in their molecules. In addi­tion, these additives present significant antiknocking properties. In this way, the oxygenates play two important roles for elevating the gasoline quality. Precisely, the most employed oxygenates as antiknocking additive in gasoline and has a branched carbon chain like TEL (see Figure 1.2) that suggests the increase of the octane number in its fuel blends. In this way, the oxygenates have allowed the replacement of lead in the gasoline.

3.2.1 Starch

Starch is a polymeric carbohydrate made up of glucose units linked by glycosidic bonds. The starch is the most abundant carbohydrate in nature after the lignocel — lulosic complex and is present in high amounts in very important crops for human food, such as corn, wheat, potato, cassava, rye, oats, rice, sorghum, and barley. About 54 million tons per year of starch are produced for industrial purposes, from which 55% comes from the United States. From the total produced starch,

44.1 million tons come from corn, 2.6 million from cassava and rice, and 2.8 mil­lion from potato (Messias de Braganga and Fowler, 2004)

Starch is composed of two types of a-glucans, i. e., polymers based on glu­cose monomers linked by a-type glycosidic bonds, amylose and amylopectin, which represent 98 to 99% of the dry weight of starch kernels. The properties of these two polysaccharides along with the ratio between them in the starch kernels

FIGURE 3.3 Structure of amylose: n « 1,000.

determine the properties of the starch obtained from different plant sources. This ratio ranges from less than 15% amylose in waxy starch and 20 to 35% amylose in normal starch to greater than 40% amylose in amylo-starch. In addition, the water content of the starch kernels in equilibrium with the air goes from 10 to 12% in grains to 14 to 18% in tubers (Tester et al., 2004).

The amylose is a linear polysaccharide made up of D-glucopyranose units linked by a(1,4) glycosidic bonds, although it has been established that some molecules present several branching points due to the presence of a(1,6) (Buleon et al., 1998). These branching points are present in an amount of 9 to 20 per mole­cule (Sajilata and Singhal, 2005). The molecular weight of amylose is in the range 1×105-1×106 Da with a polymerization degree of 324 to 4,920 (1,000 on average). These variations depend on the starch origin. The basic structure of amylose is depicted in Figure 3.3. Two ends can be distinguished: one end with intact gly — cosidic hydroxyl (corresponding to the glucose residue of the right-hand side of Figure 3.3) called the reducing end, and the end with the glucose residue whose glycosidic hydroxyl is participating in the glycosidic bond with the following glu­cose residue (the left-hand side of Figure 3.3) called the nonreducing end. These ends are very important during the process of enzymatic hydrolysis of starch.

The amylopectin is a highly branched polymer made up of chains of D-glucopyranose units linked by a(1,4) bonds, but with 5 to 6% a(1,6) bonds leading to branching points (Buleon et al., 1998). These branching points occur every 15 to 30 glucose units on average. The amylopectin has a greater molecular weight than amylose (1×107-1×109 Da) with a higher polymerization degree of 9,600-15,900 (Tester et al., 2004). Because of the branches, the amylopectin pres­ents an elevated amount of chains that are differentiated by their inner or outer character. As in the case of amylose, the nonreducing ends of amylopectin can be distinguished (see the left-hand side of Figure 3.4a). The single reducing end of the amylopectin molecule can be observed on the right-hand side of Figure 3.4b.

The amylose and amylopectin are located in the starch kernel in a radial way in the form of concentric layers where there exist crystalline and amorphous zones of these two polysaccharides. This implies that kernel degradation, the first stage in the solubilization of starch in water, is very difficult at low temperatures. To achieve this goal, it is necessary to break the starch kernels through a hydrother­mal treatment involving the adsorption of hot water by the kernels and their swell­ing and destruction with the corresponding release of amylose and amylopectin in a soluble form. This process is known as starch gelatinization and has a crucial

importance during the industrial processing of this carbohydrate, especially dur­ing the production of starch hydrolyzates with sweetening properties (starch syr­ups), glucose, and ethanol. It is worth emphasizing that the solubilization of starch is required for the enzymatic attack of starch for producing glucose.

The most employed microorganisms for fuel ethanol production are the yeasts of the Saccharomyces cerevisiae species that convert hexoses, such as glucose and fructose, into pyruvate through glycolysis, which is finally reduced to etha­nol generating two moles of ATP for each molecule of consumed hexose under anaerobic conditions (Claassen et al., 1999) as shown in Figure 6.1. This micro­organism also has the ability to convert hexoses into CO2 by aerobic respiration. One of the two processes may be favored depending on the oxygen concentration in the culture medium and the carbon source. In the latter case, mainly biomass is formed and it is the base for large-scale production of baker’s yeast. In addition to their ability to be grown under anaerobic conditions, yeasts have the advantage of tolerating relatively high concentrations of ethanol (up to 150 g/L). Below a concentration of 30 g/L the inhibition is negligible (Kargupta et al., 1998). The yeast Schizosaccharomyces pombe has the additional advantage of tolerating high osmotic pressures (high amounts of salts) and high solids content (Bullock, 2002; Goyes and Bolanos, 2005). In fact, a fermentation process using a wild strain of this yeast has been patented (Carrascosa, 2006).

Other yeasts having the capability of growing under thermophilic conditions have been evaluated from an industrial viewpoint. Increased fermentation temper­ature accelerates metabolic processes and lowers the refrigeration requirements. For this reason, one of the yeasts that is most studied for ethanol production is Kluyveromyces marxianus, which can be cultivated at temperatures higher than 40°C (Ballesteros et al., 2001). This condition makes this yeast very promising in the case of cellulose conversion schemes for ethanol production by the simul­taneous accomplishment of hydrolysis and fermentation (Ballesteros et al., 2004) because the cellulases have greater activity at temperatures much higher (50 to 60°C) than those of conventional fermentations.

One of the main problems during ethanol production from lignocellulosic mate­rials is that S. cerevisiae can ferment only certain mono — and disaccharides, such as glucose, fructose, maltose, and sucrose. Nevertheless, this microorganism cannot
assimilate either cellulose or hemicellulose directly. Furthermore, this yeast does not assimilate the pentoses obtained during the pretreatment of lignocellulosic biomass when hemicellulose is hydrolyzed at a higher degree. This hemicellulose hydrolyzate contains pentoses (mostly xylose, though also arabinose) as well as other hexoses (glucose, mannose, and galactose). For this reason, the utilization of pentose-utilizing microorganisms has been proposed similarly to some species of yeasts. Yeasts, such as Pichia stipitis, Candida shehatae, and Pachysolen tan — nophilus can assimilate both pentoses and hexoses (Olsson and Hahn-Hagerdal, 1996) as shown in Table 6.2. One key aspect in the metabolism of xylose is its conversion into xylulose that is integrated to the metabolic pathways for pyruvate synthesis (final product of glycolysis) from which ethanol is derived. Pyruvate is also the starting point for the cycle of tricarboxylic acids (Krebs cycle; Figure 6.2). The cultivation of these yeasts requires a thorough control to ensure low levels of oxygen in the medium needed for the oxidative respiratory metabolism.

For pentose-assimilating yeasts, the hexoses are, however, the most readily and rapidly assimilable substrate during ethanol production. This implies a diauxic growth, which means that the hexoses are consumed firstly before the pentoses if the fermentation is extended enough. After a relatively short lag-phase in which the enzymes necessary for pentose metabolism are synthesized, the pentoses are consumed until the end of fermentation. This means that the microorganisms do not utilize the two types of sugar at the same time, which causes a decrease in the

biomass utilization rate. As a rule, the microorganisms prefer the glucose over the galactose, followed by the xylose and arabinose (Gong et al., 1999), which is explained by the catabolic repression the glucose exerts on the consumption rate of xylose and other pentoses as in the case of C. shehatae. In addition, ethanol productivity achieved using xylose-assimilating yeasts is lower than that of micro­organisms fermenting only hexoses. Thus, their ethanol production rate from glu­cose is at least five times lower than that observed in S. cerevisiae. Moreover, their culture requires oxygen and ethanol tolerance that is two to four times lower (Claassen et al., 1999). Most xylose-utilizing yeasts are mesophiles, i. e., they are cultivated at temperatures near 30°C; likewise S. cerevisiae, though there exist reports about the methylotrophic yeast Hansenula polymorpha cultivated at 37°C in a xylose-containing medium (Ryabova et al., 2003).