Cellulosic Ethanol: Improving Cost E ﬃ ciency by Coupling Semi-Continuous Fermentation and Simultaneous Sacchariﬁcation Strategies

: A novel approach to improve ethanol production from sugarcane bagasse is proposed. Biomass was pretreated with sodium hydroxide, sulfuric, oxalic, and maleic acids (1% w / v ) at di ﬀ erent temperatures (130–170 ◦ C) and times (10–30 min). The pretreatment with NaOH at 160 ◦ C for 20 min was found to be the most e ﬃ cient for further enzymatic sacchariﬁcation. A semi-continuous fermentation system coupled with a simultaneous sacchariﬁcation and fermentation strategy was used, attaining fermented liquor every 24 h. The amount of enzymes needed for sacchariﬁcation was optimized, as well as the production time and ethanol concentration. The process occurred with near to complete depletion of glucose, obtaining ethanol concentrations ranging from 8.36 to 10.79% ( v / v ). The whole system, at bench scale, showed stability over 30 days, and ease of management and control. This strategy may improve cost e ﬃ ciency in the production of cellulosic ethanol at industrial scale. SSJcF: Simultaneous saccharification and juice co-fermentation; SSF: Simultaneous saccharification and fermentation; mSSF: Modified simultaneous saccharificationandfermentation;SSCF:Simultaneoussaccharificationandco-fermentation;PSSSF:Pre-saccharificationandsimultaneoussaccharificationandfermentation;PSSF:Prehydrolysisandsimultaneoussaccharificationandfermentation;HCDCandSSF:Highcell-densitycultureandsimultaneoussaccharificationandfermentation;dSSF:Simultaneoussaccharificationandfermentationwithdelayedyeastinoculation.


Introduction
Fossil fuel depletion and climate change are problems that have encouraged the research and development of technologies aimed at obtaining possible substitutes for fossil fuel sources [1]. Biofuels that come from a renewable source, such as biomass, are regarded as a promising replacement for petroleum-based fuels [2]. Plant biomass is a raw material of interest because it is readily available, both as crops or as waste from industrial activities, and can be used for energy purposes [3].
Lignocellulosic ethanol refers to the product of fermentation of both glucose from cellulose and pentoses from hemicellulose, while cellulosic ethanol is obtained from cellulose as the sole source of glucose [4]. The leftovers and/or waste obtained from agriculture and industry, such as sugarcane bagasse (SCB), corn stover, wheat and rice straw, wood chips, and the like, can be regarded as suitable lignocellulosic materials [5,6]. The production of ethanol from such lignocellulosic biomass (primarily composed of cellulose, hemicellulose, and lignin) requires four steps: (i) Pretreatment of biomass, (ii) enzymatic saccharification to sugar monomers, (iii) fermentation of hydrolyzed sugars to ethanol, and (iv) distillation and dehydration [7].
A pretreatment is essential for the fractionation of lignocellulosic biomass. It contributes to removal of lignin, reduction of cellulose crystallinity, and rise in the material's porosity, thus resulting

Materials and Yeast Strain
SCB was provided by a local sugar mill (Ingenio Azucarero del Norte, Imbabura, Ecuador). Cellic ® CTec2 and HTec2 enzymes for cellulosic ethanol were supplied by Novozymes (Franklinton, NC, USA). The Saccharomyces cerevisiae strain used in the study (CLQCA-INT-005) was supplied by the Catholic University Yeast Collection in Quito (CLQCA). This strain was selected for its high ethanol yield (0.49 G·g −1 , in grams of ethanol produced per gram of glucose) after screening 150 different S. cerevisiae yeasts (data not shown).

Physical and Chemical Pretreatments of SCB
SCB was washed, dried, and ground to obtain particles smaller than 100 µm. Several solutions containing different catalyzers such as oxalic acid, maleic acid, sulfuric acid, and sodium hydroxide (all at a concentration of 1% w/v) were tested with SCB at 10% (w/v) to assess different thermal pretreatments. The trials were performed in triplicate in a pressure reactor (Parr 4848; Parr Instrument Co., Moline, IL, USA) according to the following variables: Temperature-130, 140, 150, 160, and 170 • C-and reaction time-10, 20, and 30 min. The pretreated samples of SCB were centrifuged at 7000 rpm for 5 min, and the solid fraction was washed with distilled water by centrifugation until a neutral pH was reached and then dried at 105 • C for 24 h. The solid fraction recovery was quantified according to Equation (1): where Pf is the weight of the insoluble part of pretreated SCB (g) and Pi is the biomass weight of SCB (g). A vibrational analysis of the recovered solid fraction was conducted using a Nicolet iS50 FTIR spectrometer (Thermo Scientific, Waltham, MA, USA), equipped with an in-built diamond attenuated total reflectance (ATR) system. The spectra were collected in the 400-4000 cm −1 region with a 1 cm −1 spectral resolution; 128 scans were co-added, and the resulting interferogram was averaged.

Enzymatic Saccharification of the Holocellulosic Fraction of SCB
Enzymatic saccharification was performed in triplicate in 100 mL flasks at 15% solids (w/v) (dry pretreated biomass of SCB) suspended in 0.05 M citrate buffer (pH 4.5), with the addition of 10 FPU (filter paper unit) of Cellic ® CTec2 per gram of dry biomass and Cellic ® HTec2 at 20% of the volume of Cellic ® CTec2.
The activity of Cellic ® CTec2 was determined in accordance with the standard protocols of the National Renewable Energy Laboratory [34]. The saccharification conditions were: 55 • C in a stirring incubator at 200 rpm for 72 h. The samples were centrifuged at 7000 rpm for 5 min, the liquid fraction containing dissolved hydrolysis products was analyzed by high-performance liquid chromatography (HPLC) with an Agilent (Santa Clara, CA, USA) 1200 Series apparatus, and the hydrolysate was used in the batch fermentation.
The yield of saccharide production from pretreated SCB hydrolysate was calculated according to Equation (2), adapted from Li et al. [35]: Saccharide yield (%) = C glucose ×0.9 Cellulose content + C xylose ×0.88 where C glucose and C xylose represent the concentrations of glucose and xylose in the enzymatic hydrolysates (in g·L −1 ), respectively; 0.9 is the glucan-to-glucose content conversion factor; 0.88 is the xylan-to-xylose content conversion factor; and C substrate is the substrate loading (in g·L −1 ). Cellulose and xylan contents correspond to the composition of SCB (in g·g −1 ).

Yeast Strain Acclimation
The chosen S. cerevisiae strain (CLQCA-INT-005) was acclimated by a series of successive liquid cultures consisting of pretreated SCB hydrolysate and yeast extract peptone dextrose in proportions of 0:1; 0.25:0.75; 0.5:0.05; 0.75:0.25; and 1:0 v/v. The acclimated yeast culture was used in fermentation tests, both in batch and semi-continuous fermentation experiments.

Batch Fermentations (15% Solids w/w)
Batch fermentations were performed in tubes with 5 mL of hydrolyzed SCB as the sole carbon source, 0.67% (w/v) yeast nitrogen base, and 1 × 10 8 cells·mL −1 of acclimated S. cerevisiae. All fermentations were carried out in triplicate. The incubation was conducted at 30 • C, under stirring at 100 rpm, for 72 h. The fermented product was centrifuged at 7000 rpm for 10 min to separate yeast cells and remnant non-hydrolyzed biomass. The supernatant was then collected and filtered with a 0.22 µm membrane filter and stored at −20 • C.

Measurement of Sugar and Ethanol Yield
The sugar and ethanol concentrations were analyzed by HPLC at 79 • C with 5 mM H 2 SO 4 as an eluent at a flow rate of 0.6 mL·min −1 and an injection volume of 10 µL. An organic acid H + ion exchange column (Rezex ROA; Phenomenex, Torrance, CA, USA) was used; sugars and ethanol were detected by a refractive index detector. The compounds were identified by their relative retention times and quantified based on a calibration curve prepared with standard sugars and fermentation products.
The theoretical yield of ethanol (%) was calculated with Equation (3) [36]: where C ethanol corresponds to the ethanol concentration at the end of the fermentation (in g·L −1 ); C glucose corresponds to the glucose concentration at the beginning of the fermentation (in g·L −1 ); and 0.51 represents the conversion factor for glucose-to-ethanol based on the stoichiometry of fermentation.

Pretreatment and Prehydrolysis of SCB
The biomass (SCB) was pretreated with NaOH (1%, w/v) at 160 • C for 20 min and at a 1:10 (w/v, solids:NaOH) rate. The process was carried out in a 400 L-capacity pressure reactor heated by thermal oil. The pretreated SCB was filtered, washed, dried, and eventually ground. The resultant biomass was liquefied (prehydrolysis) in flasks by suspending the biomass in 0.05 M citrate buffer (pH 4.5). Finally, 6 FPU of Cellic ® CTec2 per gram of dry biomass and Cellic ® HTec2 were incorporated into the mix. Following this, 60% of the prehydrolyzed SCB (15 g) was incubated for 6 h at 55 • C at 200 rpm; once liquefied, the remaining 40% of the prehydrolyzed SCB (10 g) was added to the suspension. A second incubation period of 4 h was needed to reach full liquefaction of the SCB.

Semi-Continuous Fermentation System Coupled with Simultaneous Saccharification
The SFSS experiments were carried out in a 2 L Bioflo/Celligen 310 jar bioreactor (New Brunswick Scientific, Edison, NJ, USA) at 100 rpm, pH 4.5, 30 • C for 30 days. As a starter culture, 1 L of enzymatic liquefied SBC was added with acclimated S. cerevisiae (1 × 10 8 cells·mL −1 ). Every 24 h, 100 mL of the product was harvested from the jar, and 100 mL of enzymatically prehydrolyzed SCB was added. The concentration of fermentation products and residual glucose (as well as other saccharides) was determined by high-performance liquid chromatography.

Statistical Analyses
The enzymatic saccharification, fermentation, and SFSS data were analyzed using IBM SPSS statistics software (IBM, Armonk, NY, USA). An analysis of variance was performed on the results, and the means were compared by Tukey's honest significant difference test with a significance level of 0.05.

Characterization of SCB
SCB was characterized to determine the lignocellulosic fraction, obtaining 42.78% of cellulose, 32.07% of hemicellulose, and 14.16% of lignin. The overall content of polysaccharides (holocellulose) in SCB was 74.85%, which represents the biomass fraction that can be utilized for bioethanol production processes. Other studies show a variation of ±10% in the composition of cellulose, hemicellulose, and lignin in SCB [37]. Such variation in SCB composition can be ascribed to the sugarcane variety, age, type of crop, and method of harvesting [38].

Physical and Chemical Pretreatments of SCB
The effects of the chemical pretreatments on the recovery of the solid fraction of SCB are summarized in Table 1 (see Table S1 for all the different process conditions). The solid fraction left after NaOH pretreatment was always higher both in mild and strong conditions. In mild conditions, the remnant biomass after alkaline pretreatment was 10.51, 4.7, and 7.4% higher than the ones left after sulfuric, oxalic, and maleic acids pretreatments, respectively, whereas, after the strong alkaline pretreatment, the remnant biomass was 5.5, 0.6, and 4.8% higher than the ones left after sulfuric, oxalic, and maleic acids pretreatments. This result was in accordance with other studies, which have shown that, in comparison to acid pretreatments, alkaline pretreatments can remove lignin more efficiently with a relatively mild attack to the holocellulose fraction. In consequence, the hydrolytic enzymes' accessibility to the polysaccharide fraction is highly increased [39]. The ATR-FTIR spectra of the solid fractions recovered showed the higher removal of lignin by the pretreatment with NaOH ( Figure S1). On the other hand, it could be observed that the H 2 SO 4 pretreatment led to the lowest solid fraction recovery, indicating a greater solubilization of lignin and hemicellulose and even some degradation of the cellulose fraction [40]. Similar studies have reported that the solid fraction of lignocellulosic materials decreases as the pretreatment's severity increases, causing a loss of sugars, which are converted into non-fermentable by-products such as fermentation inhibitors or saccharinic acids [41].
These results coincide with those reported by Qing et al. [40], who found that alkaline pretreatments produce strong delignification on lignocellulosic materials, whereas acid pretreatments can mainly hydrolyze and remove hemicellulose from the solid fraction of pretreated SCB, resulting in a higher concentration of sugar-degradation by-products.
The aim of a pretreatment is to facilitate the accessibility of cellulase enzymes to convert cellulose into fermentable sugars via hydrolysis. Since an adequate pretreatment is closely related to the digestibility of the material, pretreatment choice has a noticeable impact on ethanol yield after fermentation. Apart from the type of chemical catalyst, other pretreatment conditions, such as temperature and reaction time, also have a significant influence on the overall process [12].
At this point, it should be clarified that the drying of pretreated cellulose fibers (performed at 105 • C in the present study) may cause hornification, meaning that the accessibility of enzymes may be hindered by the agglomeration of cellulose fibers [42]. However, for industrial ethanol production, the drying step would not be required, so the enzymatic accessibility of cellulose should not be compromised.

Enzymatic Saccharification of the Holocellulosic Fraction of SCB
The pretreated SCB hydrolysates showed the presence of saccharides such as glucose, xylose, cellotriose, and cellobiose (detailed for all process conditions in Table S2). The results in this study showed that the type of chemical catalyst exerted a more pronounced effect on the saccharide composition and concentration than temperatures or reaction times. For instance, a 33% higher glucose production was observed for NaOH pretreatments compared to the acid pretreatments (Table 2). Likewise, higher xylose contents were obtained from the NaOH pretreatment than for those consisting of diluted acid; this result is attributed to the lack of a degradation effect from NaOH on hemicellulose. Conversely, the acid pretreatments resulted in lower concentrations of xylose [16]. In all reactions, partial digestion products of cellulose, i.e., cellobiose and cellotriose, were detected (at concentrations between 2.22-10.41 and 0.84-4.43 g·L −1 , respectively). Xylose released from hemicellulose hydrolysis can be fermented to produce ethanol by certain yeast strains and bacteria. A great number of genetically modified yeast capable of fermenting both hexoses and pentoses have been developed in recent years [43]. However, in this study, we tested a non-genetically modified Saccharomyces cerevisiae strain that can only convert glucose to ethanol. Thus, the calculations of ethanol yields in this study were only focused on glucose fermentation (although it should be noted that the hemicellulose fraction that yields predominantly pentose sugars in the context of a biorefinery may be converted into ethanol by using recombinant yeasts or bacteria).
The impact of the pretreatment temperature and reaction times was also studied. The results showed that an increase in temperature and reaction time resulted in an increase in the amount of glucose and xylose released from cellulose and hemicellulose, respectively. Glucose and xylose release was higher when using biomass pretreated at a high temperature and extended time. The increase factor was 3.86 and 69 times higher in the case of glucose and xylose, respectively. These figures were obtained taking into account the maximum and minimum values experimentally registered from all the chemical pretreatments. However, the statistical analysis showed no significant differences (p > 0.05) as a function of the reaction time (in the two-way and three-way interactions), showing a maximum variation of 0.2% for glucose and xylose.
Conversely, there were significant differences (p < 0.05) depending on the temperature of the pretreatments: The highest saccharide yields were attained with NaOH at 160 and 170 • C. According to Tukey's HSD tests (based on saccharide yield), the combination of NaOH (1% w/v), 160 • C, and 20 min may be considered the most suitable conditions for SCB pretreatment.
The saccharide yield was determined after 72 h of enzymatic saccharification, showing higher values obtained from NaOH pretreatment (data are summarized in Table 2). The results were similar to those reported by Qing et al. [40], who found that soybean hull and straw pretreated with alkali showed better yields as compared to acid-pretreated samples (86.9 and 75.6%, respectively).

Batch Fermentation
To establish which process performed the best in terms of yield, it was necessary to determine the ethanol production efficiency. The influence of the different pretreatments on ethanol yield is summarized in Table 3 (detailed in Table S3 for all process conditions). All the maximum fermentation yield values are very similar as seen in Table 3, which demonstrates the satisfactory performance of the yeast strain. Nevertheless, the ethanol concentration is a function not only of the yeast's fermentation yield, but also of the fermentable sugar concentration in the wort. According to our experiments, the biomass that was enzymatically hydrolyzed and previously pretreated with NaOH, yielded more fermentable sugar concentration compared to the three different acid pretreatment experiments above reported in Table 2. The ethanol yield of this experiment was calculated based on the theoretical 0.51 g ethanol produced per g glucose consumed. According to Tyagi et al. [44], alkaline pretreatments of SCB can result in a cellulose recovery of 81% and delignification of 68.5%. The comparison of the means of the treatments used in this study showed significant differences (p < 0.05) in ethanol concentration influenced by different chemical pretreatments. The results showed that the NaOH-pretreated SCB hydrolysates exhibited the highest glucose concentrations, yielding up to 43.91 g·L −1 of ethanol, as expected.
Regarding the theoretical ethanol yields, a range between 97.92 and 99.85% was reached, and no significant differences among pretreatments were found (p > 0.05). The attainment of these high yields may be ascribed to: (i) Selection of an adequate pretreatment, which does not release by-products that may act as fermentation inhibitors for S. cerevisiae [45]; (ii) selection of yeast strains that can efficiently convert sugars to ethanol [46]; (iii) acclimation of yeasts in pretreated biomass hydrolysates as culture medium; and (iv) proper carbon:nitrogen proportion in the wort. The combination of the aforementioned factors made it possible to eliminate the inhibitory effect, resulting in a satisfactory yeast performance.

The SFSS System
The kinetics of glucose and ethanol production for the SFSS system of the prehydrolyzed SCB with high solids load (25% w/v) during a period of 30 days are shown in Figure 1. The yeast strain (CLQCA-INT-005) was added to the system after an acclimation process. A lag phase of 5 days occurred before the fermentation attained the steady state. Then, during the following 25 days of fermentation, glucose consumption was almost complete, remaining less than 1% in the medium.

The SFSS System
The kinetics of glucose and ethanol production for the SFSS system of the prehydrolyzed SCB with high solids load (25% w/v) during a period of 30 days are shown in Figure 1. The yeast strain (CLQCA-INT-005) was added to the system after an acclimation process. A lag phase of 5 days occurred before the fermentation attained the steady state. Then, during the following 25 days of fermentation, glucose consumption was almost complete, remaining less than 1% in the medium. Consequently, with glucose consumption, the production of ethanol was constant, showing an average concentration of 9.07% v/v (which corresponded to 71.57 g·L −1 ), with a standard deviation of 0.85, during the whole experiment. Ethanol concentration did not show significant differences (p > 0.05) in production until the end of the experiment.
At an industrial scale, a compromise between concentration and fermenter residence time for ethanol production needs to be found. A minimum of 40 g·L −1 of ethanol is necessary for the process to be profitable. Xu et al. [29] and Dahnum et al. [22] reported that SSF was more efficient in terms of ethanol productivity, exhibiting advantages-such as shorter production time and better operational and financial feasibility-compared to SHF systems.
Some studies have reported that the gradual addition of substrate and/or enzymes improves enzymatic hydrolysis, which, in turn, allows the increase in substrate load (>10% solids); thus, the increase in sugar concentration and ethanol production was possible [28]. According to Sotaniemi et al. [47], a 10% increase in enzymatic hydrolysis can be reached in SSF systems in fed-batch compared to batch systems. However, previous studies have shown an unfavorable correlation between the initial solids loading and cellulose conversion, which mainly results in high viscosity of the hydrolyzed sludge in the reactor [48]. To overcome these limitations, the prehydrolysis of biomass prior to SSF allows high loads of solids to be handled with lower sludge viscosity, resulting in higher ethanol productivity.
Over the past two years, other approaches have been developed by modifying SSF in different manners. Aiming at increasing ethanol yields, the mix of first and second generation sources, such as sweet sorghum with its stalks or corn cobs with tea seed cake [49,50], has been proposed. Other variations to the SSF method have included pre-saccharification, fed batch, and a high-density yeast inoculum [51][52][53]. A summary of the raw materials, main process parameters, and ethanol yields in these studies and in other recent reports on SSF modification strategies is presented in Table 4. Reported fermentation times varied from 36 to 216 h; the amounts of enzymes used ranged from 6 to 30 FPU; and the ethanol percentages attained were in the 1.9 to 15.26% interval.
It is worth noting that the ethanol concentration obtained in this work was among the highest Consequently, with glucose consumption, the production of ethanol was constant, showing an average concentration of 9.07% v/v (which corresponded to 71.57 g·L −1 ), with a standard deviation of 0.85, during the whole experiment. Ethanol concentration did not show significant differences (p > 0.05) in production until the end of the experiment.
At an industrial scale, a compromise between concentration and fermenter residence time for ethanol production needs to be found. A minimum of 40 g·L −1 of ethanol is necessary for the process to be profitable. Xu et al. [29] and Dahnum et al. [22] reported that SSF was more efficient in terms of ethanol productivity, exhibiting advantages-such as shorter production time and better operational and financial feasibility-compared to SHF systems.
Some studies have reported that the gradual addition of substrate and/or enzymes improves enzymatic hydrolysis, which, in turn, allows the increase in substrate load (>10% solids); thus, the increase in sugar concentration and ethanol production was possible [28]. According to Sotaniemi et al. [47], a 10% increase in enzymatic hydrolysis can be reached in SSF systems in fed-batch compared to batch systems. However, previous studies have shown an unfavorable correlation between the initial solids loading and cellulose conversion, which mainly results in high viscosity of the hydrolyzed sludge in the reactor [48]. To overcome these limitations, the prehydrolysis of biomass prior to SSF allows high loads of solids to be handled with lower sludge viscosity, resulting in higher ethanol productivity.
Over the past two years, other approaches have been developed by modifying SSF in different manners. Aiming at increasing ethanol yields, the mix of first and second generation sources, such as sweet sorghum with its stalks or corn cobs with tea seed cake [49,50], has been proposed. Other variations to the SSF method have included pre-saccharification, fed batch, and a high-density yeast inoculum [51][52][53]. A summary of the raw materials, main process parameters, and ethanol yields in these studies and in other recent reports on SSF modification strategies is presented in Table 4. Reported fermentation times varied from 36 to 216 h; the amounts of enzymes used ranged from 6 to 30 FPU; and the ethanol percentages attained were in the 1.9 to 15.26% interval. It is worth noting that the ethanol concentration obtained in this work was among the highest ones. If operational factors (viz., enzyme concentration and time) are weighted in, the SFSS strategy (6 FPU cellulases, 24 h process) has an unmatched performance (9.07% ethanol yield). In the works in which higher ethanol yields were attained (Table 4), fermentation times were at least three times longer, and enzyme concentration was at least 67% higher. In view of its efficiency, the proposed strategy holds enough promise for cellulosic ethanol production to deserve further examination in a pilot-scale.

Conclusions
The pretreatment with NaOH (1%, w/v) at 160 • C for 20 min was the most efficient for further conversion of cellulose into fermentable sugars. The SFSS system was used for enzymatic saccharification and fermentation of SCB, using a high proportion of solid feedstock to liquid (25% solids), and harvesting the fermented product every 24 h. Monitoring of the SFSS system over 30-days showed an average ethanol concentration of 9.07% (v/v) and <1% of residual glucose, demonstrating that the system can reach a steady state with no significant variations during the whole process. The proposed SFSS system would have potential advantages in terms of: (i) Reduction of the cellulase/xylanase enzymes use; (ii) the addition of high solids load; (iii) continuous production of ethanol in a stable system; (iv) ability for daily processing of ethanol; and (v) a more practical and simplified operational system. All these improvements would have a highly positive impact on the economy of cellulosic ethanol production at the industrial scale, and call for further studies.
Supplementary Materials: The following are available online at http://www.mdpi.com/2227-9717/8/11/1459/s1, Figure S1: ATR-FTIR spectra of the solid fraction recovered after alkali and acid pretreatments at 160 • C for 20 min; Table S1: Solid fraction recovery of pretreated SCB as a function of pretreatment conditions; Table S2: Saccharide composition of SCB hydrolysate as a function of pretreatment conditions; Table S3: Ethanol produced by fermentation of pretreated SCB hydrolysate as a function of pretreatment conditions.