Applied Microbiology and Biotechnology

, 83:447

Identification and characterization of fermentation inhibitors formed during hydrothermal treatment and following SSF of wheat straw

Authors

    • Biosystems Division, Risø DTUTechnical University of Denmark
  • Anders Thygesen
    • Biosystems Division, Risø DTUTechnical University of Denmark
  • Anne Belinda Thomsen
    • Biosystems Division, Risø DTUTechnical University of Denmark
Biotechnological Products and Process Engineering

DOI: 10.1007/s00253-009-1867-1

Cite this article as:
Thomsen, M.H., Thygesen, A. & Thomsen, A.B. Appl Microbiol Biotechnol (2009) 83: 447. doi:10.1007/s00253-009-1867-1

Abstract

A pilot plant for hydrothermal treatment of wheat straw was compared in reactor systems of two steps (first, 80°C; second, 190–205°C) and of three steps (first, 80°C; second, 170–180°C; third, 195°C). Fermentation (SSF) with Sacharomyces cerevisiae of the pretreated fibers and hydrolysate from the two-step system gave higher ethanol yield (64–75%) than that obtained from the three-step system (61–65%), due to higher enzymatic cellulose convertibility. At the optimal conditions (two steps, 195°C for 6 min), 69% of available C6-sugar could be fermented into ethanol with a high hemicellulose recovery (65%). The concentration of furfural obtained during the pretreatment process increased versus temperature from 50 mg/l at 190°C to 1,200 mg/l at 205°C as a result of xylose degradation. S. cerevisiae detoxified the hydrolysates by degradation of several toxic compounds such as 90–99% furfural and 80–100% phenolic aldehydes, which extended the lag phase to 5 h. Acetic acid concentration increased by 0.2–1 g/l during enzymatic hydrolysis and 0–3.4 g/l during fermentation due to hydrolysis of acetyl groups and minor xylose degradation. Formic acid concentration increased by 0.5–1.5 g/l probably due to degradation of furfural. Phenolic aldehydes were oxidized to the corresponding acids during fermentation reducing the inhibition level.

Keywords

Pilot scale pretreatmentLignocelluloseFuransPhenolsSimultaneous saccharification and fermentationSacharomyces cerevisiae

Introduction

Ethanol derived from biomass has great potential to be a renewable transportation fuel that can replace gasoline. The largest potential feedstock for ethanol is by far lignocellulosic biomass, which includes straw residues (e.g., straw, corn stover, and sugar cane bagasse; Varga et al. 2004), herbaceous plants (e.g., alfalfa and clover grass; Thomsen and Hauggaard-Nielsen 2008), and wood fiber residues (Söderström et al. 2004). The cellulose and hemicellulose fractions in this lignocellulosic biomass must be enzymatically hydrolyzed to monomer sugars after a pretreatment for hydrolytic cleavage of its partially crystalline structure (Puls and Schuseil 1993).

During hydrothermal treatment of wheat straw, a substantial part of the hemicellulose sugars are extracted to be contained in a liquid fraction. The obtained fiber fraction of mainly cellulose and lignin can be used for simultaneous saccharification and fermentation (SSF; Thomsen et al. 2008). The liquid stream from the process can be utilized as feedstock for ethanol fermentation with pentose-fermenting thermophiles (Klinke et al. 2001) and with gene modified yeast (Wahlbom et al. 2003; Rudolf et al. 2008). A hydrothermal treatment pilot plant has been tested as a two-step process, which includes soaking and pretreatment (Thomsen et al. 2006). These trials showed a high enzymatic cellulose convertibility (90%), but at temperatures of 200°C, a significant part of the hemicellulose was degraded (50%) resulting in formation of fermentation inhibitors (Thomsen et al. 2006). Therefore, the plant was changed to a three-step system. The pretreatment step was divided into a hemicellulose extraction step at lower temperature to avoid monomer degradation (170–180°C) followed by a step at higher temperature (195°) to increase enzymatic convertibility in the cellulose that remained solid.

Degradation products from pretreatment of lignocellulosic materials can be divided into the following classes: carboxylic acids, furan derivatives, and phenolic compounds. A recent review about inhibitors produced from wood fibers and straw fibers was made by Almeida et al. (2007). Carboxylic acids such as formic acid and acetic acid are the degradation products formed at highest concentration (0.5–5 g/l; Klinke et al. 2002; Rudolf et al. 2008). Acetic acid is formed by the initial hydrolysis of acetyl groups in hemicellulose, but is also the end product from many oxidation reactions (Foussard et al. 1989; Mishra et al. 1995). Aromatic degradation products from sugar degradation are predominantly furans: 2-furfural (referred to as furfural in the following text) from pentose degradation, 5-hydroxy-2-methylfurfural (5-HMF) from hexose degradation (Fengel and Wegener 1989), and 2-furoic acid from degradation of furans (Klinke et al. 2002). These have been found after pretreatment of lignocellulose in the range of 0.007–11 g/l depending on conditions and biomass (Almeida et al. 2007). Phenolic monomers are produced by solubilization, hydrolysis, and oxidation of lignin (Klinke et al. 2002). These compounds include alcohols, aldehydes, ketones, and acids (Almeida et al. 2007).

The aim of this study was to identify the inhibitors produced during hydrothermal treatment of wheat straw and to examine the effect of these inhibitors on ethanol fermentation of the whole slurry (liquid/hemicellulose and solid/cellulose fraction) using fermentation with Saccharomyces cerevisiae. The effects of pretreatment temperature and oxidative conditions were investigated to get low fermentation inhibition, high ethanol yield, and high hemicellulose recovery.

Materials and methods

Pretreatment of wheat straw

In the two-step experiments, wheat straw was fed to the presoaking vessel (reactor 1) continuously at a rate of 50 kg straw per hour (dry matter content = 80–90%). The temperature was 80°C, and the residence time was 6 min. After presoaking, the wet straw with a dry matter content of 25–35% was transported to the reactor 2 inlet. Water was introduced at the top of reactor 2 giving a countercurrent flow. The straw and water were heated by addition of steam in the reactors. The pretreated biomass was weighed and collected in containers. Process water was recycled in reactor 1 and let to a drain. Samples of the liquid fraction were collected from a drain in the bottom of reactor 1. The temperatures tested in reactor 2 were 190°C, 195°C, 200°C, and 205°C and the reaction time 3 and 6 min. Furthermore, addition of 0.5% H2O2 (wet oxidation) was tested.

The experiments in the three-step reactor system are described in detail by Thomsen et al. (2008). In this setup, the straw passed three reactor steps. The straw passed reactor 1 at 80°C for 20 min. The straw from reactor 1 was pumped into reactor 2 and treated at 170/180°C for 7.5 min or 180°C for 15 min. The temperature in the final reactor 3 was 195°C, and the residence time was 3 min. Two different water flows were tested (4 and 5 kg water/kg DM corresponding to 420 and 600 l water/h). Water was introduced at the end (top) of reactor 3 and passed in countercurrent flow through the three steps to reactor 1. The straw and water were heated by addition of steam in the reactors. Samples of liquid from the extraction process were taken in the bottom of reactor 2 from a drain. Samples of fiber fraction were taken at the end of reactor 3 (Table 1; Thomsen et al. 2008).
Table 1

Pretreatment conditions divided into group A, B, and C for the 2-step experiments and group D for the 3-step experiments

Trial

Reactor 2

Reactor 3

Water flow

Catalyst

Temperature (°C)

Time (min)

Temperature (°C)

Time (min)

l/h

Percent H2O2

A1

200

6

250

A2

195

6

250

A3

190

6

250

B1

200

6

250

0.5

B2

195

6

250

0.5

B3

190

6

250

0.5

C1

205

3

250

C2

200

3

250

C3

195

3

250

D1

180

15

195

3

600

D2

180

7.5

195

3

420

D3

170

7.5

195

3

420

The conditions in reactor 1 (presoaking) was in groups A, B, C, 80°C and 6 min and in group D, 80°C and 20 min

Table 1 shows the parameter settings in the experiments. The trials are divided into four groups: group A is experiments performed with a residence time (RT) of 6 min and no addition of chemicals; in group B, the residence time is 6 min with addition of H2O2; in group C, the residence time is 3 min with no addition of chemicals; and group D is the three-step experiments. In each group, the experiments are listed with the highest severity (temperature) first.

Analysis methods

The composition of the raw and pretreated straw fibers was measured by strong acid hydrolysis of the carbohydrates: cellulose and hemicellulose. This resulted in determination of glucose, xylose and arabinose, and lignin. In the liquid fractions (hydrolysates), carbohydrates, organic acids, and ethanol were determined by high-performance liquid chromatography (HPLC; Shimadzu, Kyoto, Japan) as described previously (Thygesen et al. 2005). The overall composition of the raw straw and the pretreated straw are presented by Thomsen et al. (2008).

The liquids were filtered (0.45 μm), and the furans 5-hydroy-2-methylfurfural (5-HMF) and 2-furfural (furfural) were separated on a Luna SU C18 250 × 4.6 mm column (Phenomenex) using a linear eluent gradient of methanol (10–90%) at pH 3 with a flow of 0.7 ml/min. The furans were quantified by HPLC (Shimadzu) with UV detection at 280 nm using authentic compounds as calibration standards.

The phenolic compounds were selectively extracted from the liquids by solid phase extraction on polystyrene divinylbenzene polymer columns: IST Isolute ENV+ 100 mg/l ml (International Sorbent Technology, UK) according to the method described in Klinke et al. (2002). GC was performed on a HP 6890 Series GC-system with flame ionization detection (heater, 300°C; H2 flow, 30 ml/min; air flow, 400 ml/min; makeup flow, helium 25 ml/min). The GC was fitted with a 30 × 0.25-00 Restec Corp xti-5 column. Helium was used as a carrier gas at a flow rate of 0.7 ml/min. The temperature profile of the GC-method was 3 min at 80°C and was then increased by 6°C min−1 to 220°C and by 40°C min−1 to 280°C. The GC inlet temperature was kept at 250°C, and 1 μl sample was injected splitless.

Simultaneous saccharification and fermentation

Prehydrolysis (liquefaction) and simultaneous saccharification and fermentation (SSF) was performed in 200 ml fermentation flasks. Eight grams of the dried solid fiber fraction were mixed with 80 ml of pH adjusted (pH 4.8) filtrate/liquid fraction originated from the same pretreatment (10% DM). Prehydrolysis of the solid fraction was performed at 50°C for 24 h at an enzyme loading of 10 FPU/g DM filter cake, using Cellubrix L [Novozymes A/S]. After liquefaction, the fermentation flasks were supplemented with a second batch of Cellubrix L at an enzyme loading of 10 FPU/g DM, added 0.2 ml of a sterile filtered urea (24%), and inoculated with 0.2 g dried baker’s yeast (S. cerevisiae) [V&S Denmark A/S] after having cooled down to room temperature. The flasks were sealed with a yeast lock filled with glycerol and incubated at 32°C for 6 days. The ethanol concentration was determined by HPLC analysis as described above. Samples were taken of the raw liquids, after hydrolysis, and at 24, 75, and 145 h of fermentation.

Calculations

Recoveries and yields of cellulose and hemicellulose were calculated as outlined by Thomsen et al. (2006) and shown in Table 2. The theoretical ethanol production was based on the pretreatment and the enzymatic hydrolysis yields. The ethanol yield in SSF experiments was calculated as percentage of theoretical based on cellulose content of the fiber fraction and the glucose in the filtrates (Eq. 1).
$${\text{EtOH}}\;{\text{yield}} = \frac{{{\text{EtOH}}_{{{{\text{Gravimetric}}} \mathord{\left/ {\vphantom {{{\text{Gravimetric}}} {{\text{HPLC}}}}} \right. \kern-\nulldelimiterspace} {{\text{HPLC}}}}} \times 162{\text{g}} \mathord{\left/ {\vphantom {{\text{g}} {{\text{mol}}}}} \right. \kern-\nulldelimiterspace} {{\text{mol}}}}}{{{\left( {{\text{glucan}}\;{\text{in}}\;{\text{solid}}\;{\text{ + }}\;{\text{glucan}}\;{\text{in}}\;{\text{hydrolysate}}} \right)} \times 46{\text{g}} \mathord{\left/ {\vphantom {{\text{g}} {{\text{mol}} \times 2}}} \right. \kern-\nulldelimiterspace} {{\text{mol}} \times 2}}}$$
(1)
Table 2

Total amount of glucose (including enzymatic hydrolyzed cellulose) and produced ethanol in SSF fermentation of pretreated fiber—and liquid fraction

Trial

Cellulose in fibers (g/l)

Cellulose enz. conv. (%)

Glucose in liquid (g/l)

Total glucose (g/l)

Ethanol produced (percent theor.)

pH after SSF

log(R′0)

Hemicel. Recovery (%)

A1

49.2

67.4

3.8

53.0

74

3.9

−0.19

55.1

A2

48.1

64.6

3.1

51.2

69

3.7

−0.15

65.4

A3

46.3

57.6

2.3

48.6

51

3.4

0.03

78.5

B1

51.8

67.9

1.8

53.6

75

3.6

0.10

48.4

B2

51.3

56.6

1.8

53.1

64

3.4

0.14

57.6

B3

44.3

57.9

3.5

47.8

50

3.3

0.10

80.4

C1

52.6

72.9

1.3

53.9

67

3.8

−0.27

30.2

C2

50.6

63.5

1.3

51.9

64

3.9

−0.48

43.3

C3

47.7

59.1

1.3

49.0

64

3.7

−0.46

52.7

D1

56.3

51.1

1.5

57.8

65

3.6

−0.21

83.3

D2

59.3

55.7

1.5

60.8

63

3.7

-0.18

55.6

D3

55.8

53.4

0.9

56.7

61

3.7

-0.36

66.6

The pretreatment severity factor R0 is a function of temperature, time, and pH. Finally, the hemicellulose recovery is presented (Thomsen et al. 2006 (A1–C3); group D—Thomsen et al. 2008).

The pretreatment severity (R0) was calculated as outlined by Kabel et al. (2007) in cases of one hot reactor step and as outlined by Thomsen et al. (2008) in cases of two hot reactor steps.

Results

Formation of inhibitors in general

During pretreatment, the main degradation products were, acetic acid ranging from 2 to 3 g/L and formic acid ranging from 1 to 2 g/L. The amount of degradation products from carbohydrates were for furfural 0.03–1.2 g/L, for 5-HMF 0.02–0.3 g/L, and for 2-furoic acid 0.01–0.07 g/L. The amount of lignin degradation products such as phenols was in the range of 0.03–0.17 g/L (Table 3). In general, S. cerevisiae showed high ability to detoxify the hydrolysates assimilating several toxic compounds such as furfural, 5-HMF, vanillin, syringaldehyde, and hydroxybenzaldehyde during fermentation. However, the concentration of several important fermentation inhibitors (e.g., acetic acid, formic acid, vanillic acid, ferulic acid, and acetovanillone) was increased during fermentation. Of available glucose polymers, 25–50% were not utilized due to low enzymatic cellulose convertibility.
Table 3

Development in concentration of sugar degradation products (furfural and 5-HMF), degradation products from furfural and 5-HMF (acetic acid + formic acid + 2-furoic acid), and phenols during enzymatic hydrolysis and fermentation

Experiment

After step

Furfural (mg/l)

5-HMF (mg/l)

Acetic acid (g/l)

Formic acid (g/l)

2-furoic acid(mg/l)

Phenols (mg/l)

A1

Treatment

780

289

2.35

0.52

52

65

Enz. hydr

429

308

2.58

0.57

37

81

SSF

5

8

2.91

1.43

69

75

A2

Treatment

239

188

1.70

0.51

30

66

Enz. hydr

443

308

2.19

0.34

41

82

SSF

6

7

2.95

1.92

60

60

A3

Treatment

27

7

0.69

0.14

14

45

Enz. hydr

34

24

1.69

0.18

17

58

SSF

2

2

4.13

1.79

22

120

B1

Treatment

429

88

1.30

0.51

20

67

Enz. hydr

382

121

2.17

0.58

25

78

SSF

3

2

3.46

1.71

49

167

B2

Treatment

241

43

1.06

0.42

15

60

Enz. hydr

223

76

2.15

0.43

20

78

SSF

2

2

3.57

2.17

36

184

B3

Treatment

95

18

0.77

0.26

NA

NA

Enz. hydr

96

37

1.59

0.30

14

67

SSF

1

2

4.94

2.03

19

160

C1

Treatment

1212

186

1.12

0.91

22

40

Enz. hydr

1091

224

1.96

0.43

27

83

SSF

4

5

1.95

1.40

51

137

C2

Treatment

909

187

0.98

1.93

21

36

Enz. hydr

666

185

1.56

0.38

24

66

SSF

4

8

2.33

1.17

65

133

C3

Treatment

436

69

0.51

0.00

18

32

Enz. hydr

657

167

2.11

0.49

25

64

SSF

3

5

2.42

1.37

41

135

D1

Treatment

1155

74

1.42

0.42

16

63

Enz. hydr

986

83

1.89

0.00

23

98

SSF

4

2

2.90

1.42

49

127

D2

Treatment

847

71

1.28

0.45

4

42

Enz. hydr

721

82

1.98

0.23

24

72

SSF

3

2

2.48

1.68

50

138

D3

Treatment

802

55

1.17

0.42

16

67

Enz. hydr

662

66

2.17

0.48

23

82

SSF

3

2

3.11

1.02

50

131

For each experiment, row 1 mentions concentrations in the hydrolysate after pretreatment; row 2, concentrations after enzymatic hydrolysis; and row 3, the concentrations after 145 h of SSF fermentation. NA means not measured.

Simultaneous saccharification and fermentation

In Table 2, the enzymatic cellulose convertibility is presented for the 12 experiments. The cellulose convertibility increased versus temperature. The highest cellulose convertibility was found at 205°C and 3 min (73%). However, a tendency is seen that a prolonged reaction time increases the cellulose convertibility (comparing A1 and A2 with C2 and C3). No significant effect was seen from addition of H2O2 on cellulose convertibility comparing the A group with the B group.

Figure 1 shows the ethanol production during SSF fermentation of the pretreated straw samples. Comparing all 12 experiments, the highest ethanol production (presented as g/l in Fig. 1 and as percent of theoretical in Table 2) was obtained in trial A1 and B1 (75%) in which high cellulose convertibility (68%) was obtained. In trial A3 at 6 min reaction time, low inhibition was illustrated as a short lag-phase but with low final yield (51%) due to the low cellulose convertibility (58%). At higher temperature, the lag-phase was longer and the fermentation rate slower due to inhibitors like furfural (Table 3). The lag-phase is used by the yeast to adjust to and/or detoxify the substrate. The yeast was capable of metabolizing some of the inhibitors. However, it cannot produce ethanol from furfural and 5-HMF (Palmqvist and Hähn-Hägerdal 2000).
https://static-content.springer.com/image/art%3A10.1007%2Fs00253-009-1867-1/MediaObjects/253_2009_1867_Fig1_HTML.gif
Fig. 1

ad Ethanol production during SSF fermentation of the pretreated wheat straw (10% DM) using the 12 different sets of parameters

Comparing group B with A shows that H2O2 addition possessed no significant effect on the fermentation rate. Furthermore, all the glucose became consumed, so inhibition was not obtained as shown in Fig. 2. Since the concentrations of furfural, 5-HMF, and acetate were higher in group A (no chemicals) than group B (H2O2 addition; Table 2), it appeared that some inhibitors were oxidized to less inhibitory compounds with H2O2. In group C at 3 min reaction time, the best experiment was found at the highest temperature 205°C. As seen in Table 3, the concentration of formic acid was reduced presumably due to decomposition at this temperature (Bjerre and Sørensen 1992), which decreases the inhibition.
https://static-content.springer.com/image/art%3A10.1007%2Fs00253-009-1867-1/MediaObjects/253_2009_1867_Fig2_HTML.gif
Fig. 2

Monomer sugar and carboxylic acid concentrations in the fermentation broths during the enzymatic hydrolysis period from -24 h until 0 h (inoculation time) and during the fermentation period from 0 h until 145 h after inoculation. The A3 sample was pretreated at 190°C and A1 at 200°C (Table 1). The sample B1 is similar to A1 but with H2O2 addition

In SSF of material pretreated in the three-step process, the two trials performed at 180°C in the second step (D1, RT = 15 min and D2, RT = 7.5 min) gave very similar results with final ethanol yields of 63–65%. It should be noted that a higher water flow (600 l/h) was used in the most severe three-step experiment (180°C, 15 min—Table 1), which is probably the reason why a high hemicellulose recovery was obtained in this experiment (83%—Table 2).

Furans from xylose and glucose degradation

In the experiments performed at 190°C (A3 and B3), very low concentrations of furfural and 5-HMF were found in the liquid fractions, and the concentrations increased versus temperature (Table 3). The highest concentration of furfural (1200 mg/l) was thereby found in the liquid from experiment C1 at the highest temperature (205°C). Since furfural is a product of pentose degradation, this is in agreement with the low hemicellulose recovery (Table 2; 30%). Also, the three-step experiments caused high formation of furfural, probably because the material was heated twice in this process. In experiment D1, a furfural concentration of 1,100 mg/l was observed, even though the hemicellulose recovery was high (Table 3; 83%). This indicates that furfural is more stable at the higher water flow used in D1 compared with D2. Furfural inhibition begins in the range 2–5 g/l with Baker’s yeast (Zhao et al. 2005).

The highest concentration of 5-HMF was found at 195–200°C with a residence time of 6 min. Reducing the residence time to 3 min significantly decreased the formation of 5-HMF, but the lowest 5-HMF concentrations were found with H2O2 and in the three-step experiments. In general, the concentration of 5-HMF was lower than that of furfural, but this is also expected since 5-HMF is a degradation product of hexose sugars, and the glucose recovery was high in all the experiments (95–105%; Thomsen et al. 2006). The concentration of 5-HMF increased during enzymatic hydrolysis, whereas furfural concentration decreased slightly. Both furfural and 5-HMF were degraded during the fermentation (Table 3).

Formation of acetic acid and formic acid

Acetic acid concentration in the pretreatment liquids was ranging between 0.5 and 1.5 g/l (Table 3). The highest concentrations were found in the most severe pretreatments (A1 and A2), and the lowest were found in experiments performed at 190°C (A3 and B3). The concentration of acetic acid was significantly increased during enzymatic hydrolysis (0.3–1 g/l), since acetic acid is formed by hydrolysis of acetyl groups in hemicellulose (Table 3). The concentration was further increased during fermentation, since it is produced as a by-product. Part of the acetate was produced by the inoculum, since the commercial Baker’s yeast contains a minor content of lactic acid bacteria which can degrade xylose (Fig. 2). The pH value after fermentation was 3–4 (Table 2), and it was lowest in trial A3 and B3 due to large formation/release of acetic acid resulting in a final concentration of 4–5 g/l. This is due to the low inhibition level obtained at 190°C, since yeast is more tolerant to inhibitors than many bacteria that can degrade xylose. Succinic acid was also formed during the fermentation (Fig. 2). This can be the result of reactions between acetyl radicals forming the C-4 compound (Thomsen 1998).

Formic acid is produced from both furfural and 5-HMF (Almeida et al. 2007; Dunlop 1948; Larsson et al. 1999). Formic acid has a dissociation constant of 3.75 in water and can inhibit fermentation at a lower concentration than acetic acid due to the small molecule size (Larsson et al. 1999). The formic acid concentrations in the pretreatment liquids, and the hydrolysates were around 0.5 g/l (Table 3). It increased during the first 72 h of fermentation to 1.5–2 g/l (Fig. 2).

Phenols from lignin degradation

The concentration of the free phenolics analyzed for ranged between 30 and 65 mg/l in the pretreatment liquids (Table 3). The concentration in the liquid fractions prior to enzymatic hydrolysis increased versus pretreatment severity (Fig. 3). Especially lowering the retention time from 6 min (group A) to 3 min (group C) appear to have a positive effect, reducing the concentration of free phenolics with almost 50% (Table 3). During enzymatic hydrolysis, free phenolics were released from the fibers, which caused the concentration after hydrolysis to be similar in most experiments (around 70–80 mg/l; and less dependent on pretreatment severity; Fig. 3). During SSF, even more free phenolics were liberated especially toward the end of fermentation, possibly as a result of increased solubility due to the produced ethanol and hydrolysis of suspended particles.
https://static-content.springer.com/image/art%3A10.1007%2Fs00253-009-1867-1/MediaObjects/253_2009_1867_Fig3_HTML.gif
Fig. 3

Concentration of phenols in the liquid fractions versus the pretreatment severity factor log(R0) including all the experiments. Data are presented after pretreatment and after enzymatic hydrolysis

Figure 4 shows the different phenolic compounds versus fermentation time for the material from the pretreatments with H2O2 addition (group B2) grouped as alcohols, aldehydes, ketones, and acids. No single phenolic compound was found in concentrations higher than 30 mg/l at any stage during fermentation (Fig. 4). The levels of the phenolic monomers were similar at 190, 195, and 200°C (data not shown). Vanillin and syringaldehyde were present in highest concentration in the hydrolysate—but were degraded in the beginning of fermentation. Vanillyl alcohol was not present after hydrolysis but was formed during fermentation, since yeast can reduce vanillin to vanillyl alcohol (Zaldivar et al. 2000). Also acetovanillone, vanillic acid, homovanillic acid, and ferulic acid were formed during fermentation. These six phenolic compounds became predominant in all SSF broths.
https://static-content.springer.com/image/art%3A10.1007%2Fs00253-009-1867-1/MediaObjects/253_2009_1867_Fig4_HTML.gif
Fig. 4

Free phenols in the fermentation broth versus time. The sample was produced by pretreatment at 195°C for 6 min with 0.5% H2O2 (B2) and enzymatic hydrolysis. The graphs are grouped as a alcohols, b aldehydes, c ketones, and d acids

Discussion

The most important inhibitors for ethanol fermentation with Baker’s yeast (S. cerevisiae) were identified using hydrothermal-treated wheat straw as substrate. These inhibitors were found to be sugar degradation products like furfurals and fatty acids and lignin degradation products like phenols. The determined ethanol yields (50–75%) were only on glucose basis since Baker’s yeast cannot ferment pentose sugars. If a pentose-fermenting strain was used, the resulting yields would change, since the hemicellulose recovery was lowest in the experiments with high cellulose convertibility (Thomsen et al. 2008). Therefore, you have to compromise between a high ethanol yield from glucose and a high yield of xylose.

Aldehydes like furfural are easily oxidized in wet oxidation to carboxylic acids (Thomsen 1998). S. cerevisiae has been found capable of utilizing both furfural and 5-HMF (Taherzadeh et al. 2000), which was in agreement with this study showing that detoxification takes place during the first 24 h of fermentation. This explains why the lag-phase was short in ethanol fermentation (<5 h—Fig. 1) in the experiments A3 and B3 with low furfural—and 5-HMF concentrations (Table 3). In general, the effect of furans can be explained as a re-direction of yeast energy to fixing damage, which result in reduced ATP and NAD(P)H levels (Almeida et al. 2007) and lower fermentation rate.

Yeast can grow at pH as low as 2.5 if acetic acid is not present in the medium, but at an acetic acid concentration of 3 g/l, growth will stop at pH 3–3.5 (Taherzadeh et al. 1997; Palmqvist et al. 1999). Experiments A3, B2, and B3 all had end-pH of around 3.5 and acetic acid concentrations of more than 3 g/l. However, the low ethanol yield obtained in these experiments was caused by the low cellulose convertibility (Table 2; 57–58%), since all the glucose was consumed in the fermentation. Phenolics have been found to inhibit fermentation (Clark and Mackie 1984; Sierra-Alvarez and Lettinga 1991; Klinke et al. 2004), since these partition into cell membranes and cause loss of integrity (Palmqvist and Hähn-Hägerdal 2000). S. cerevisiae could assimilate the aromatic aldehydes: vanillin, syringaldehyde, and hydrolybenzaldehyde (Fig. 4). The inhibitory effect is increased versus the hydrophobicity of the compound (Delgenes et al. 1996). Phenolic aldehydes and ketones are thereby more inhibitory than phenolic acids (Zaldivar et al. 2000), so the observed formation of phenol acids decreased the inhibition level.

The ethanol yield was, in general, limited by the enzymatic cellulose convertibility, which increased versus temperature, while the hemicellulose recovery decreased due to monomer degradation. The best treatment is thereby the two-step process with 200°C, 6 min, and H2O2 (high ethanol yield from glucose (75 g/100 g)). However, if the xylose should be used, the optimal conditions are 195°C and 6 min due to the higher hemicellulose recovery.

Acknowledgments

The work was financially supported by EU-contract ENK6-CT-2002–00650 and Danish project: PSO-F&U-project 2006-1-6412. Børge Holm Christensen (Sicco K/S, Denmark), Jan Larsen, Mai Østergaard Petersen, and Erik Hedahl-Frank (DONG Energy, Denmark) are thanked for collaboration on pretreatment experiments on the pilot plant, and Tomas Fernqvist and Ingelis Larsen (Technical University of Denmark) are thanked for technical assistance on analytical work.

Copyright information

© Springer-Verlag 2009