Scholarly article on topic 'Techno-economic evaluation of conditioning with sodium sulfite for bioethanol production from softwood'

Techno-economic evaluation of conditioning with sodium sulfite for bioethanol production from softwood Academic research paper on "Chemical engineering"

CC BY-NC-ND
0
0
Share paper
Academic journal
Bioresource Technology
OECD Field of science
Keywords
{"Cellulosic ethanol" / Conditioning / "Sodium sulfite" / "Simultaneous saccharification and fermentation" / "Techno-economic evaluation"}

Abstract of research paper on Chemical engineering, author of scientific article — Adnan Cavka, Carlos Martín, Björn Alriksson, Marlene Mörtsell, Leif J. Jönsson

Abstract Conditioning with reducing agents allows alleviation of inhibition of biocatalytic processes by toxic by-products generated during biomass pretreatment, without necessitating the introduction of a separate process step. In this work, conditioning of steam-pretreated spruce with sodium sulfite made it possible to lower the yeast and enzyme dosages in simultaneous saccharification and fermentation (SSF) to 1g/L and 5FPU/g WIS, respectively. Techno-economic evaluation indicates that the cost of sodium sulfite can be offset by benefits resulting from a reduction of either the yeast load by 0.68g/L or the enzyme load by 1FPU/g WIS. As those thresholds were surpassed, inclusion of conditioning can be justified. Another potential benefit results from shortening the SSF time, which would allow reducing the bioreactor volume and result in capital savings. Sodium sulfite conditioning emerges as an opportunity to lower the financial uncertainty and compensate the overall investment risk for commercializing a softwood-to-ethanol process.

Academic research paper on topic "Techno-economic evaluation of conditioning with sodium sulfite for bioethanol production from softwood"

ELSEVIER

Contents lists available at ScienceDirect

Bioresource Technology

journal homepage: www.elsevier.com/locate/biortech

Techno-economic evaluation of conditioning with sodium sulfite for bioethanol production from softwood

Adnan Cavka a, Carlos Martín a'*, Björn Alriksson b, Marlene Mörtsellc, Leif J. Jönssona

a Department of Chemistry, Umea University, SE-901 87 Umea, Sweden b SP Processum AB, SE-891 22 Örnsköldsvik, Sweden c SEKAB E-Technology AB, SE-891 26 Örnsköldsvik, Sweden

HIGHLIGHTS

• Inhibition problems can be alleviated in situ in bioreactors using reducing agents.

• Conditioning of pretreated spruce with sodium sulfite was evaluated.

• Reductions of yeast load or enzyme load compensate for cost of sodium sulfite.

• Estimation of required reduction levels: yeast, P0.68 g/L; enzyme, pi FPU/g WIS.

ABSTRACT

Conditioning with reducing agents allows alleviation of inhibition of biocatalytic processes by toxic by-products generated during biomass pretreatment, without necessitating the introduction of a separate process step. In this work, conditioning of steam-pretreated spruce with sodium sulfite made it possible to lower the yeast and enzyme dosages in simultaneous saccharification and fermentation (SSF) to 1 g/L and 5 FPU/g WIS, respectively. Techno-economic evaluation indicates that the cost of sodium sulfite can be offset by benefits resulting from a reduction of either the yeast load by 0.68 g/L or the enzyme load by 1 FPU/g WIS. As those thresholds were surpassed, inclusion of conditioning can be justified. Another potential benefit results from shortening the SSF time, which would allow reducing the bioreactor volume and result in capital savings. Sodium sulfite conditioning emerges as an opportunity to lower the financial uncertainty and compensate the overall investment risk for commercializing a softwood-to-ethanol process.

© 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CCBY-NC-ND license

(http://creativecommons.org/licenses/by-nc-nd/4XI/).

CrossMark

ARTICLE INFO

Article history:

Received 25 May 2015

Received in revised form 15 July 2015

Accepted 16 July 2015

Available online 21 July 2015

Keywords: Cellulosic ethanol Conditioning Sodium sulfite

Simultaneous saccharification and fermentation

Techno-economic evaluation

1. Introduction

Energy security and environmental concerns favor energy carriers from renewables, such as plant biomass, compared to the utilization of fossil resources, such as oil. Large-scale utilization of sugarcane-based first generation bioethanol as a transportation fuel started in 1975 in Brazil (Goldemberg et al., 2004), which remained the world leader until 2005, when the United States became the largest ethanol producer using corn starch as the main feedstock. Currently USA and Brazil produce, respectively, around 50 and 26 billion liters annually, and they provide around 87% of the world's fuel ethanol market (REN21, 2013; McMillan et al., 2014). Cellulosic ethanol produced from lignocellulosic biomass

* Corresponding author. Tel.: +46 90 7866099. E-mail address: carlos.martin@chem.umu.se (C. Martín).

does not affect the food sector and can serve as a useful complement to ethanol from cane sucrose and corn starch (Ho et al., 2014). A lignocellulose-to-ethanol biorefining process also has potential to generate other products including energy carriers based on lignin and on digestion of parts of hemicelluloses to biogas.

In lignocellulose-to-ethanol processes, cellulose is hydrolyzed with either acids or enzymes, and the released sugars are converted to ethanol by a fermenting microorganism, usually the yeast Saccharomyces cerevisiae. These two steps can be performed either separately as a separate hydrolysis and fermentation (SHF) or combined in a simultaneous saccharification and fermentation (SSF) (Ohgren et al., 2007). If the hydrolysis is to be performed enzymat-ically it should be preceded by a pretreatment step that should ensure the reactivity of cellulose towards cellulolytic enzymes (Chandra et al., 2007; Hu and Ragauskas, 2012; Galbe and Zacchi, 2012; Behera et al., 2014).

http://dx.doi.org/10.1016/j.biortech.2015.07.051 0960-8524/© 2015 The Authors. Published by Elsevier Ltd.

This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

During acid pretreatment carbohydrates and lignin are partially degraded leading to formation of by-products, some of which have an inhibitory effect on fermenting microorganisms and cellulolytic enzymes (Pienkos and Zhang, 2009; Jonsson et al., 2013). Recalcitrant forms of biomass, such as softwood, require harsh pre-treament conditions that increase problems with inhibitors. Recycling of process water would also lead to increasing problems with inhibitors. Furthermore, the trend towards using high solids concentrations to gain higher ethanol titre (Kristensen et al., 2009) also results in higher inhibitor concentrations.

Detoxification by different chemical, biological and physical means, also known as conditioning, is one strategy for minimizing inhibition problems (Pienkos and Zhang, 2009; Jonsson et al., 2013). Results achieved so far suggest that potent detoxification methods give good results also for strongly inhibitory lignocellu-losic hydrolysates, while other measures, such as using more resistant microbial strains, tend to have a more limited effect (Jonsson et al., 2013). This becomes obvious in studies where the fermentation of inhibitory lignocellulosic hydrolysates is benchmarked against reference fermentations without inhibitors.

A common weakness of most detoxification methods is the requirement of an additional process step, which adversely affects the process cost. This is true for example with regard to treatment with alkali, which is otherwise known as a very potent detoxification method (Jonsson et al., 2013). That drawback is overcome by detoxification with reducing agents, such as sodium dithionite, sodium sulfite and sodium borohydride, an approach that was recently developed by our group (Alriksson et al., 2011; Cavka et al., 2011; Cavka and Jonsson, 2013). Since conditioning with reducing agents can be performed at commonly used fermentation pH and temperature in the presence of microorganisms and enzymes, and since the reaction with the inhibitors is rapid, no separate process step is required. Additionally, this novel detoxification method neither results in sugar degradation nor in formation of precipitates, and it can be applied ad hoc if inhibition signs are observed during the fermentation (Cavka, 2013).

The most studied inhibitors of fermenting microorganisms include aliphatic carboxylic acids (such as acetic acid, formic acid, and levulinic acid), furan aldehydes [such as furfural and 5-hydroxymethylfurfural (HMF)], and phenolic compounds (for example coniferyl aldehyde and ferulic acid) (Jonsson et al., 2013). Sulfur oxyanions, such as sulfite and dithionite, sulfonate inhibitors which renders them less reactive, charged at process-relevant pH values, and highly hydrophilic (Cavka et al., 2011). Sodium borohydride reduces inhibitors, which become less reactive but not as hydrophilic as the corresponding sulfonated substances as no charge is introduced (Cavka and Jonsson, 2013). Previous results indicate that sulfur oxyanions are effective against inhibitors of both microbes and enzymes, while sodium borohy-dride is effective against inhibitors of microbes, but not inhibitors of enzymes (Cavka and Jonsson, 2013). Neither sulfur oxyanions nor sodium borohydride react with sugars (Alriksson et al., 2011; Cavka and Jonsson, 2013), and therefore inhibitory effects of sugars on cellulolytic enzymes are not affected by treatments with these substances. Inhibitory effects of sugars on cellulolytic enzymes can instead be decreased by using SSF (Ohgren et al., 2007) or by using enzymes that are less susceptible to sugar inhibition.

The current work was aimed to clarify the economic feasibility of sodium sulfite conditioning of spruce slurries prior to SSF for ethanol production. Sodium sulfite was used for conditioning as it has a favorable effect on both microbial and enzymatic conversion and as it is an industrial chemical that is well suited for process up-scaling. Selected experimental options were tested in order to demonstrate the importance of conditioning of the slurries for running SSF at lower yeast and enzyme loads. Based on the experimental results, a techno-economic evaluation was

performed to elucidate whether the resulting economic benefits can offset the cost of the addition of sodium sulfite.

2. Methods

2.1. Raw material and pretreatment

Debarked wood chips of Norway spruce (Picea abies) were pre-treated thermo-chemically by SEKAB E-Technology in the Biorefinery Demonstration Plant (BDP) in Ornskoldsvik, Sweden. The pretreatment was performed in a 30-L reactor, loaded to approximately 50% during operation. Spruce wood chips were treated in continuous mode at an overpressure of 20 bar (corresponding to 210 °C). Sulfur dioxide was added at a rate of 1.2 kg/h, which corresponds to approximately 1% of spruce dry weight (DW). The pretreatment lasted 7 min, and finished with a sudden release of pressure. The resulting slurry had a water-insoluble solids (WIS) content of around 18.5% and its pH was around 1.5. The slurry was cooled directly after pretreatment and stored at 4 °C until further use.

2.2. Detoxification

Prior to detoxification, 1.4 kg of the pretreated slurry was diluted with Milli-Qwater to a WIS content of 12.5% in a 4-L plastic container, and its pH was adjusted to 5.5 with 10 M sodium hydroxide. Then sodium sulfite powder was added to the diluted slurry for reaching a concentration of 12.5 mM. The suspension was mixed manually and allowed to react for 10 min at room temperature (20 °C).

2.3. Simultaneous saccharification and fermentation (SSF) at lab scale

The effect of detoxification with reducing agents on fermentation parameters was investigated in a set of SSF experiments using 250 mL Erlenmeyer flasks filled with 100 g of spruce slurry, either detoxified or non-detoxified, with a pH of 5.5 and a WIS content of 12.5%. Since it has previously been shown that nutrient supplementation is not required for SSF of pretreated Norway spruce (Alriksson et al., 2011), no extra nutrients were added in order to simplify the subsequent techno-economic evaluation. Freeze-dried yeast (S. cerevisiae Ethanol Red, Fermentis Ltd., Marcq-en-Baroeul, France) and a state-of-the-art preparation of cellulolytic enzymes from a leading enzyme manufacturer were added directly to the fermentation flasks according to the experimental design (see Section 2.4), and the SSF was run in batch mode of operation. The flasks were sealed with Parafilm (Pechiney Plastic Packaging Company, Chicago, IL, USA) to prevent evaporation losses, and they were incubated for 96 h at 35 °C and 120 rpm in an orbital shaker (Ecotron, Infors AG, Bottmingen, Switzerland). Samples for sugar and ethanol analysis were withdrawn at 0, 24, 48, 60, 84 and 96 h of fermentation. The 48-h ethanol concentrations were used for calculating the volumetric productivity (Q) and the specific productivities on basis of either initial yeast inoculum (qx) or enzyme dosage (qz).

2.4. Experimental design

Two series of SSF experiments were performed. In the first series, the yeast concentration was varied between 1 and 2 g/L, and the enzyme load between 5 and 15 FPU/g WIS, while sodium sulfite was either added (12.5 mM) or not added (Table 1). Using the Modde 8.0 statistical software (Umetrics, Umea, Sweden), a second series of experiments was performed for further evaluation of the yeast concentration, which was varied between 0.5 and

Table 1

Experimental conditions of the SSF of pretreated spruce slurry with and without sodium sulfite detoxification.

Experimental series A

Combination Yeast (g/L) Enzyme (FPU/g WIS) Na2SOs (mM)

A1 1.0 5 0/12.5

A2 1.0 10 0/12.5

A3 1.0 15 0/12.5

A4 1.5 10 0/12.5

A5 2.0 10 0/12.5

A6 2.0 15 0/12.5

5 g/L, while the enzyme load was still in the range 5-15 FPU/g WIS (Table 2). The second experimental series consisted of two 22-blocks, each of them augmented by two replicate centre points. The first block included yeast loads below 2 g/L and it was applied to the detoxified slurry, whereas the second one involved yeast concentrations above 2 g/L, and it was applied to the non-detoxified material (Table 2).

2.5. Chemical analysis

Analyses of glucose and ethanol were performed by using high-performance liquid chromatography (HPLC). A Shodex SH-1011 column (6 im, 8 x 300 mm) (Showa Denko, Kawasaki, Japan) was used in a YoungLin YL9100 series system (YoungLin, Anyang, Korea) equipped with a YL9170 series refractive index (RI) detector for analysis of glucose and ethanol. Elution was performed with isocratic flow of a 0.01 M aqueous solution of H2SO4. The flow rate was 1.0 mL/min and the column temperature was set to 50 °C. YLClarity software (YoungLin, Anyang, Korea) was used for data analysis.

2.6. Techno-economic estimates

The estimated costs of the yeast were based on a 1-g/L inoculum in the SSF process. Taking into account the assumed plant capacity (60,000 m3 of fuel-grade bioethanol per year), the required load is 130kg/h (corresponding to 1040 tons/year with an annual operation of 8000 h). Two different alternatives, external supply and in situ production, were considered for the yeast. The costs for externally supplied yeast were provided by two major European yeast producers, while for the internally-produced yeast an estimate was made by performing a sensitivity analysis on overall production costs per liter of fuel-grade bioethanol. The sensitivity analysis was performed by calculating the difference in production costs at a yeast load of 2 g/L and subsequently lowering it to 0.01 g/L while keeping all other parameters constant. After

Table 2

Experimental conditions for the evaluation of threshold yeast concentrations for the SSF of pretreated spruce slurry with and without sodium sulfite detoxification.

Combination Yeast (g/L) Enzyme (FPU/g WIS) Na2SOs (mM)

B1 0.5 5 12.5

B2 0.5 15 12.5

B3 0.75 10 12.5

B4 0.75 10 12.5

B5 1 5 12.5

B6 1 15 12.5

B7 3 5 0

B8 3 15 0

B9 4 10 0

B10 4 10 0

B11 5 5 0

B12 5 15 0

that, the estimates were recalculated to include the capital cost for the yeast propagation unit. Enzyme costs were set to the industrial target cost of 0.10 USD per liter of fuel-grade bioethanol using an enzyme load of 5 FPU/g WIS (Olson et al., 2012). The costs of sodium sulfite were obtained from an international online business-to-business trading portal. The input costs in Swedish kronor (SEK) were converted to US dollars using the official 10-years average rate of 7.0317 SEK per USD, which resulted from averaging the Swedish National Bank (www.riksbanken.se) daily closing rates over the period from July 1, 2005, to June 30, 2015. For facilitating the calculations the rate was rounded to 7.0 SEK per USD throughout this paper.

3. Results and discussion

3.1. SSF of detoxified and non-detoxified spruce slurries

Previous studies of conditioning with reducing agents have shown promising improvements in fermentability, ethanol productivity and balanced ethanol yields as well as positive effects on enzymatic hydrolysis (Alriksson et al., 2011; Cavka and Jonsson, 2013). In this study, in order to perform a techno-economic evaluation of conditioning with reducing agents, pretreated Norway spruce, either detoxified with sodium sulfite or undetoxified, was hydrolyzed and fermented following an SSF scheme using different yeast concentrations and enzyme loads.

The first set of SSF experiments (Table 1, Fig. 1) clearly showed that all slurries that had been conditioned with 12.5 mM sodium sulfite had high volumetric ethanol productivity (Q) compared to slurries to which no sodium sulfite was added (Fig. 1A). Although the ethanol productivity of cultures without sodium sulfite was always low, using higher yeast inoculum than 1 g/L (Table 1) had a clear beneficial effect, while increased enzyme dosage was of no avail (Fig. 1A). For cultures to which sodium sulfite was added, there seemed to be a beneficial effect of raising the enzyme dosage, at least when the inoculum size was as low as 1 g/L (Table 1, Fig. 1A).

The specific ethanol productivities on basis of yeast inoculum size (qx in Fig. 1B) and enzyme dosage (qz in Fig. 1C) were evaluated. Fig. 1B shows that without sodium sulfite addition more ethanol was produced in relation to the yeast added, i.e. qx was higher, when the inoculum was large (1.5-2.0 g/L). This agrees with previous observations that using a large yeast inoculum alleviates inhibition problems (Pienkos and Zhang, 2009; Jonsson et al., 2013; Chung and Lee, 1985), evidently as the amount of inhibiting substance per yeast cell decreases. When sodium sulfite was added to the cultures, the situation was reversed so that qx was higher for a small inoculum size (1 g/L, A1-A3 in Fig. 1B) than for larger ones (1.5-2 g/L, A4-A6 in Fig. 1B). This indicates that the sodium sulfite treatment was so potent that with a larger inoculum the full capacity of the yeast cells was not utilized. With sodium sulfite, there was always a positive effect of increasing the enzyme dosage for a given size of the inoculum (qx of A1 < A2 < A3 and A5<A6) (Fig. 1B). This indicates that with sodium sulfite enzymatic glucose production was rate-limiting regardless of whether the inoculum was small (1 g/L) or large (2 g/L). The same trend was not observed for cultures with no sodium sulfite addition (Fig. 1B). This is in agreement with Fig. 1D, which shows that the concentrations of accumulated glucose in cultures to which no sodium sulfite was added were at least 25 times higher than in those treated with sodium sulfite.

Fig. 1C shows that treatment with sodium sulfite resulted in large quantities of ethanol in relation to the amount of enzyme added, especially for lower enzyme dosages, as qz of A1 (5 FPU/g) > A2, A4, A5 (10 FPU/g) > A3, A6 (15 FPU/g), and

D 0.80 0.70 0.60 0.50 0.40 0.30 0.20 0.10 0.00

Res. Glu [g/L]

Fig. 1. Volumetric ethanol productivity (g L-1 h-1 ) (A), specific ethanol productivity on basis of initial yeast inoculum (gg-1 h-1) (B), specific ethanol productivity on basis of enzyme dosage (mgFPU-1 h-1) (C), and concentration of residual glucose (g L-1) (D) in samples taken after 48 h of an SSF of spruce slurry. For (B-D), values of sodium-sulfite-treated (detoxified) cultures are given on the left axis and values of cultures to which no sodium sulfite was added (w/o detox) are given on the right axis. Labels below graphs (A1-A6) refer to the six combinations indicated in Table 1.

disregarding the inoculum size. As could be expected from the amounts of accumulated glucose (Fig. 1D), the qz values of cultures without sodium sulfite were higher when the inoculum was large (1.5-2 g/L) than when it was small (1 g/L).

For cultures without sodium sulfite, the concentration of accumulated glucose was directly related to the enzyme dosage when the inoculum size was the same (A1 < A2 < A3 and A5 < A6)

(Fig. 1D), as the enzymes generated glucose more quickly than the inhibited yeast was able to consume it. When the enzyme dosage was the same, larger inoculum was inversely related to residual glucose (A2 > A4 > A5 and A3 > A6) (Fig. 1D), evidently as a larger portion of the released glucose was metabolized to ethanol by the yeast (Fig. 1A). Cultures with sodium sulfite contained only low concentrations of glucose, of which the highest concentration (0.7 g/L) was found when the enzyme load was high and the yeast load was low (i.e. A3 in Fig. 1D) indicating that the yeast was struggling to keep up with the pace of the cellulolytic enzymes. At low enzyme load (A1) (Fig. 1D) the residual glucose was almost nil indicating that the yeast, even at a low load, was able to consume practically all available glucose. At higher inoculum levels (A4, A5 and A6), the detected residual glucose was comparable (0.40.5 g/L) independently of the enzyme load. These results agree with beneficial effects of in situ detoxification with sodium sulfite on ethanolic fermentation of lignocellulosic hydrolysates as previously demonstrated (Alriksson et al., 2011; Cavka et al., 2011 ).

As the first experiment did not reveal the threshold levels of yeast concentrations that could be utilized with and without detoxification, a second SSF experiment was performed. In the new experiment the range of yeast concentrations was expanded down to 0.5 g/L and up to 5 g/L, while the enzyme loads were kept at the same levels as previously (Table 2). The results indicate that in spite of the detoxification, yeast loads of 0.5 and 0.75 g/L were insufficient for achieving a good SSF conversion (Fig. 2). With sodium sulfite addition, the lowest yeast concentration that gave good fermentability was 1 g/L, which gave ethanol yields in the range 40.8-43.6 g/L and low concentrations of residual glucose (Fig. 2). While the difference between small (0.5-0.75 g/L) and large (1 g/L) yeast inoculum size was sharp, varying the enzyme dosage between 5 and 15 FPU/g had only small effects on the etha-nol yield (Fig. 2). Without sodium sulfite addition (i.e. B7-B12 in Fig. 2), increased inoculum size in the range 3-5 g/L did result in higher ethanol yield, but the highest yield was no more than 10.2 g/L, and in all cases there were still large amounts of residual glucose. The poor fermentability without sodium sulfite but with 5 g/L inoculum differs from the results achieved by Wingren et al. (2005), who reported ethanol concentrations close to 25 g/L in SSF of spruce slurries with a yeast load of 5 g/L. The better fer-mentability observed by Wingren et al. (2005) might be due to the fact that they used a considerably lower load of solids (5%) than what was used in the current work, and therefore the inhibitors in that slurry were more diluted, which led to a better fermentation.

The results of the second experimental series showed that the lowest yeast load that may be used for SSF of sodium-sulfite-detoxified pretreated spruce slurries was 1 g/L, and that yeast loads of up to 5 g/L were insufficient for fermenting non-detoxified slurries. Yeast loads higher than 5 g/L were not tested in this study, as they were not considered to be economically viable for large-scale bioethanol production. The results of the SSF experiments also indicated that enzyme loads of 5 FPU/g WIS were sufficient to reach at least 4% (40 g/L) of ethanol provided that detoxification with sodium sulfite was employed.

3.2. Techno-economic evaluation

The results obtained at laboratory scale were used as starting point for a cost analysis and a techno-economic evaluation of the implementation of sodium sulfite conditioning for bioethanol production from softwood. The calculations were based on a full-scale plant scenario with a capacity of 60,000 m3 of fuel-grade bioethanol per year. The plant processes 55 tons of air-dry feedstock, corresponding to around 25 tons dry weight (DW), per hour. The process consists of feedstock handling, steam pretreatment, detoxification with sodium sulfite, SSF, distillation,

■ Final ethanol ■ Residual glucose

1: lllllll.lHl-1

B1 B2 B3 B4 B5 B6 B7 B8 B9 B10 B11 B12 Fig. 2. Yields of ethanol and residual glucose concentrations after 96 h of SSF of spruce slurry with (B1-B6) or without (B7-B12) addition of sodium sulfite.

and lignin dewatering. The plant has seven SSF reactors, each of them with a working volume of 1500 m3. The time between SSF batches is assumed to be 8 h. A simplified process flowchart is shown in Fig. 3.

Around 1700 tons of sodium sulfite, either supplied externally or produced internally from NaOH and SO2, will be required for performing the detoxification operation at the same concentration used in the experimental trials (12.5 mM). Internal production of sodium sulfite may be favorable, especially if SO2 is used for other purposes, such as pretreatment, but assuming that it is acquired externally (at 715 USD per ton) and that the investment required in its handling is negligible, the total annual cost for detoxification chemical would be:

17001 x 715 USD/t = 1,215, 500 USD,

which corresponds to:

1, 215, 500 USD ^ 60,000,000 L = 0.02 USD per liter of ethanol.

3.2.1. Effect of reducing the yeast load

In order to justify the inclusion of the detoxification in a full-scale production unit, the cost of the sodium sulfite must at least be offset by the savings resulting from lowering the yeast and enzyme loadings and shortening the SSF time. The required yeast load in the full-scale production unit was estimated to be 130 kg per hour (1040 tons yeast annually) if fermentation is performed at a yeast concentration of 1 g/L. The cost of yeast was provided by two major European producers, which quoted rather close values, 7.286 and 6.714 USD per kg of dry yeast (Table 3), for their products. The cost of internally-produced yeast was also considered, and it was estimated at 1.714 USD per kg (DW) when capital costs for a propagation unit were included.

Using the yeast acquired at the lowest cost would give narrower margins and a more realistic economic evaluation. Therefore, the internally-produced yeast was chosen for the calculations. The annual cost for yeast using an inoculum of 1 g/L will be:

1040 t x 1714 USD/t = 1, 782,560 USD

The cost of the detoxification chemical is equivalent to:

1,215,500 USD ^ 1, 782, 560 USD x 100

= 68% of the yeast cost at 1 g/L (Fig.4).

That also means that the expenditure of the sodium sulfite corresponds to the annual cost of a 0.68-g/L yeast load. The inclusion of detoxification must thus enable the yeast concentration to be lowered by at least 0.68 g/L to offset the cost of the sodium sulfite provided that it will be supplied externally at a price of no more than 715 USD/t and used at a concentration of 12.5 mM for detoxification, and that the same ethanol yield and productivity will be reached. The experimental trials performed in this study show that the yeast load may be lowered by far more than the 0.68 g/L required to make sodium sulfite conditioning economically viable in a production unit. Indeed, the yeast concentrations may be lowered by at least 4 g/L, from 5 to 1 g/L, when sodium sulfite is included. That lowering of the yeast loadings would correspond to an annual saving on operating costs of:

4 x 1, 782,560 USD = 7,130,240 USD

The net saving, taking into account the cost of the sodium sulfite addition, will be:

7,130,240 USD - 1, 215, 500 USD = 5,914,740 USD

Although using larger yeast inoculum could theoretically be an alternative to detoxification (Pienkos and Zhang, 2009; Jonsson et al., 2013; Chung and Lee, 1985), it is only advisable if the

Spruce chips

55ton/h 25 ton DM/h

Steam Flash steam 22 t/h 15 t/h

Pretreatment

Slurry

Water Na2SO3 (s)

Start: 12.5% WIS

SO2(g)

1% of spruce DM

1 kg/m3 yeast 5,000 FPU/kg WIS

Distillation

116ton/h

TS 11% WIS 8%

Fig. 3. Basic flowchart of pretreatment and SSF in the modeled full-scale bioethanol plant (25 ton/h).

Table 3

Cost estimates of yeast, enzymes and sodium sulfite for production of bioethanol from Norway spruce.3

Input Input Annual cost Contribution to the unit

cost (Million USD/ cost (USD/L ethanol)

(USD/t) year)

Yeast - Supplier A 7286 7.6b 0.13b

Yeast - Supplier B 6714 7.0b 0.12b

Yeast - Internal 1714 1.7b 0.03b

production

Sodium sulfite - 715 1.2 0.02

External

supplier

Enzymesd 6.0cd 0.10cd

a The estimates are based on a plant capacity of 25 t/h, as described in Section 3.2. b Costs corresponding to a yeast load of 1 g/L.

c Based on a load of 5 FPU/g WIS and on a concentration of pretreated solids of 12.5% in the SSF.

d The enzyme costs were set to the industrial target that corresponds to 0.1 USD/L of ethanol.

50 В 40

1 30 !Л

Jj 20 га 3

Enzymes

Sodium sulfite

Fig. 4. Annual cost of the inputs. The calculations are based on yeast loads of 1 g/L (internal production) and enzyme loads of 5 FPU/g. The yellow areas show the share of yeast and enzyme costs corresponding to the sodium sulfite expenditure. USD 1 million would equal SEK 7 million. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

fermenting microorganism can be recycled at low cost. That is not the case in an industrial SSF process, where the high content of solids in the fermentation broth makes it problematic to separate yeast from biomass, and the use of fresh inocula is preferred instead of recycling the microorganism (Wingren et al., 2005). On the other hand, the cost of a larger inoculum would be higher than the cost of sodium sulfite conditioning. For instance a 5-g/L yeast inoculum, the highest load tested in the experimental trials and the base-case load in a published study on the effect of yeast and enzyme concentrations in SSF (Wingren et al., 2005), would amount to a total annual cost of:

The contribution of that expenditure to the unit cost, provided that the same ethanol yield is achieved in the fermentation of both non-detoxified slurries with high inoculum and detoxified slurries with low yeast load, will be:

8,912, 800 USD ^ 60,000,000 L = 0.15 USD/L of ethanol

The cost of a so high inoculum would be prohibitive and the produced ethanol would not be competitive.

3.2.2. Effect of reducing the enzyme load

The experimental results indicate that 5 FPU/g is enough to reach at least 4% (40 g/L) of bioethanol with the used WIS content and using sodium sulfite detoxification (Fig. 2). Using that load in a large-scale production unit, and considering the input cost to be

0.10 USD per liter of fuel-grade bioethanol, the cost of enzymes for fulfilling an annual production of 60,000 m3 fuel-grade bioetha-nol would be:

0.10 USD/L x 60,000,000 L = 6,000,000 USD

The cost of the sodium sulfite would be equivalent to:

1,215,500 USD ^ 6,000,000 USD x 100

= 20% of the enzyme cost (Fig.4).

In other words, the cost of the sodium sulfite corresponds to the annual cost of:

5 FPU/g WIS x 20 ^ 100 = 1.0 FPU/g WIS

Therefore, at a cost of 0.10 USD per 5 FPU/g WIS, the enzyme loads have to be lowered by around 1 FPU/g in order to offset the cost of detoxification assuming that sodium sulfite can be supplied externally at a cost of 715 USD/t and that the glucose yield from hydrolysis will not be affected by using a lower enzyme load.

Inhibitory effects on enzymes by sugars, alleviation of sugar inhibition by the fermenting microorganism, and ethanol produced from hemicellulosic sugars, primarily mannose, contribute to making quantification of positive effects of conditioning on the actions of enzymes difficult. Previous studies based on separate hydrolysis and fermentation (SHF) indicate that a gain of 30% in enzyme activity is a reasonable assumption (Cavka and Jonsson, 2013). The annual enzyme cost would then be reduced by:

When the cost for sodium sulfite is taken into account, the annual net saving on operating costs would amount to:

3.2.3. Effect of the reduction of both the yeast and the enzyme loads If the yeast and the enzyme loading are simultaneously reduced from 5 to 1 g/L yeast and from 6.5 to 5 FPU/g WIS enzyme while including sodium sulfite conditioning the annual cost reduction could potentially be:

7,130,240 USD + 1, 800,000 USD - 1, 215,500 USD

3.2.4. Effect on the capital investment cost

Verification of sodium sulfite conditioning in demonstration scale in the BDP indicated that the same overall ethanol yield can be reached with shorter residence time (24 h) compared to a conventional SSF with no sodium sulfite (72 h) using the same yeast and enzyme loads. Assuming that the same reduction of the SSF time can be achieved in a full-scale plant and considering the volume of the SSF reactors as 1500 m3, the time interval between batches was set to 8 h and the annuity factor to 0.12. The number of bioreactors required for running the SSF of detoxified slurries can be reduced from 7 to 3 when the total needed bioreactor volume in the full-scale plant decreases as a result of reduced residence time. The reduction of the number of SSF biore-actors leads to a decrease of the capital investment cost equivalent to 10-11.4 M USD, depending on the localization and logistics of the full-scale plant. Since the detoxification is performed in situ in the SSF bioreactor, no additional expenses in equipment are required, and the investment savings are considerably higher than the cost of the addition of sodium sulfite. Therefore, even if neither the enzyme load nor the inoculum are decreased, the sulfite addition leads to a shortening of the SSF time, which results in a reduction of the investment costs that justifies the inclusion of sodium sulfite conditioning.

6,000,000 USD

0.30 = 1,800,000 USD

1, 800,000 USD - 1,215, 500 USD = 584,500 USD

= 7,714, 700 USD

5 x 1040 t x 1714 USD = 8,912,800 USD

3.2.5. Potential effects on investment risk

The projected capital costs of the emerging cellulosic ethanol industry are significantly influenced by the over-design of initial projects to compensate for lack of large-scale experience and by the high returns expected by the financial institutions for rewarding the perceived risk of a not yet proven technology (Wyman, 2007). Inclusion of sodium sulfite conditioning would lower the financial uncertainty and hedge the overall investment risk.

The savings derived from lowering the yeast and enzyme loads are on the annual operating costs which in turn may result in higher profit margins or operating margins, provided that all other factors remain equal. The higher operating or profit margins may create a financial cushion that possibly can decrease the effects of external risks, such as political uncertainty, regulation policies, exchange rates or interest rate risks, which are outside of the investor's control. The inclusion of detoxification could thus provide investment evaluations with higher financial certainty and lower overall investment risk. The lower operating margins or savings achieved with the inclusion of detoxification may thus decrease the required return on investment (RRI) to the level of industries, such as the first generation ethanol companies, which are considered less risky. The RRI for advanced biofuels is currently considered to be around 15% annually (Miller et al., 2013), suggesting that the investment should be returned in less than seven years. During those seven years the inclusion of detoxification may result in an operating cost saving that amounts to around 58 million USD if the annual cost saving is considered to be around 7.7 million USD (see Section 3.2.3), and an average risk-free interest rate over the period is set to 2%. The saving is calculated with the use of a compound interest formula with the operating cost saving as the annually deposited amount.

4. Conclusions

Experimental trials indicate that sodium sulfite conditioning of steam-exploded spruce allows for lowering the enzyme and yeast loads or for shortening the SSF time without affecting the final ethanol concentration. The techno-economic evaluation showed that conditioning with sodium sulfite can be economically justified by reduction of either the enzyme load by 1 FPU/g or the yeast load by 0.68 g/L, and the estimations indicate that both thresholds would be surpassed. Alternatively, the shortening of the SSF time would result in a cost reduction higher than the sodium sulfite cost. Conditioning with sodium sulfite was validated in a biorefin-ery demonstration plant.

Acknowledgements

SEKAB E-Technology (Ornskoldsvik, Sweden) is acknowledged for supplying the pretreated materials and for providing facilities for running the demonstration-scale verification. The financial

support received from the Swedish Energy Agency (35367-1 and 35844-1), the Kempe Foundations, the Swedish Research Council (621-2011-4388), and the Bio4Energy research environment (www.bio4energy.se) is gratefully appreciated.

References

Alriksson, B., Cavka, A., Jonsson, L.J., 2011. Improving the fermentability of enzymatic hydrolysates of lignocellulose through chemical in-situ detoxification with reducing agents. Bioresour. Technol. 102,1254-1263.

Behera, S., Arora, R., Nandhagopal, N., Kumar, S., 2014. Importance of chemical pretreatment for bioconversion of lignocellulosic biomass. Renew. Sust. Energy Rev. 36, 91-106.

Cavka, A., Alriksson, B., Ahnlund, M., Jonsson, L.J., 2011. Effect of sulfur oxyanions on lignocellulose-derived fermentation inhibitors. Biotechnol. Bioeng. 108, 25922599.

Cavka, A., Jonsson, L.J., 2013. Detoxification of lignocellulosic hydrolysates using sodium borohydride. Bioresour. Technol. 136, 368-376.

Cavka, A., 2013. Biorefining of lignocellulose - Detoxification of inhibitory hydrolysates and potential utilization of residual streams for production of enzymes. PhD thesis. Umeà University, Sweden.

Chandra, R.P., Bura, R., Mabee, W.E., Berlin, A., Pan, X., Saddler, J.N., 2007. Substrate pretreatment: the key to effective enzymatic hydrolysis of lignocellulosics? Adv. Biochem. Eng. Biotechnol. 108, 67-93.

Chung, I.S., Lee, Y.Y., 1985. Ethanol fermentation of crude acid hydrolysate of cellulose using high-level yeast inocula. Biotech. Bioeng. 27, 308-315.

Galbe, M., Zacchi, G., 2012. Pretreatment: the key to efficient utilization of lignocellulosic materials. Biomass Bioenergy 46, 70-78.

Goldemberg, J., Coelho, S.T., Nastari, P.M., Lucon, O., 2004. Ethanol learning curve— the Brazilian experience. Biomass Bioenergy 26, 301-304.

Ho, D.P., Ngo, H.H., Guo, W., 2014. A mini review on renewable sources for biofuel. Bioresour. Technol. 169, 742-749.

Hu, F., Ragauskas, A., 2012. Pretreatment and lignocellulosic chemistry. Bioenergy Res. 5, 1043-1066.

Jonsson, L.J., Alriksson, B., Nilvebrant, N.-O., 2013. Bioconversion of lignocellulose: inhibitors and detoxification. Biotechnol. Biofuels 6, 16.

Kristensen, J.B., Felby, C., Jorgensen, H., 2009. Yield-determining factors in high-solids enzymatic hydrolysis of lignocellulose. Biotechnol. Biofuels 2,11.

McMillan, J.D., Saddler, J.N., van Dyk J.S., 2014. Commercializing conventional and advanced liquid biofuels from biomass. IEA Bioenergy Task 39 Newsletter. Issue 36, April 2014.

Miller, N., Christensen, A., Park, J.E., Baral, A., Malins, C., Searle, S., 2013. Measuring and addressing investment risk in the second-generation biofuels industry. ICCT-The International Council on Clean Transportation. Washington, D.C., 46 pp. Available at <www.theicct.org/sites/default/files/publications/ICCT_ AdvancedBiofuelsInvestmentRisk_Dec2013.pdf>.

Ohgren, K., Bura, R., Lesnicki, G., Saddler, J., Zacchi, G., 2007. A comparison between simultaneous saccharification and fermentation and separate hydrolysis and fermentation using steam-pretreated corn stover. Process Biochem. 42, 834839.

Olson, D.G., McBride, J.E., Shaw, J., Lynd, L.R., 2012. Recent progress in consolidated bioprocessing. Curr. Opin. Biotechnol. 23, 396-405.

Pienkos, P.T., Zhang, M., 2009. Role of pretreatment and conditioning processes on toxicity of lignocellulosic biomass hydrolysates. Cellulose 16, 743-762.

REN21 Renewables, 2013. Global Status Report. REN21 Secretariat, Paris, France.

Wingren, A., Galbe, M., Roslander, C., Rudolf, A., Zacchi, G., 2005. Effect of reduction in yeast and enzyme concentrations in a simultaneous-saccharification and fermentation-based bioethanol process. Appl. Biotechnol. Biochem. 122, 485500.

Wyman, C.E., 2007. What is (and is not) vital to advancing cellulosic ethanol. Trends Biotechnol. 25, 153-157.