- Email: [email protected]
Carbonic acid enhancement of hydrolysis in aqueous pretreatment of corn stover
Bioresource Technology 93 (2004) 217–226
Carbonic acid enhancement of hydrolysis in aqueous pretreatment of corn stover G. Peter van Walsum *, Helen Shi
1
Department of Environmental Studies and Glasscock Energy Research Center, Baylor University, PO Box 97266, Waco, TX 76798-7266, USA Received 9 May 2003; received in revised form 23 October 2003; accepted 9 November 2003
Abstract Carbonic acid and liquid hot water pretreatments were applied to corn stover. Temperatures ranged from 180 to 220 °C; reaction times varied between 2 and 32 min and prereaction carbon dioxide pressure was either 0 or 800 psig. Over the range of reaction conditions tested, it was found that the presence of carbonic acid had an effect of increasing the concentrations of xylose and furan compounds in the hydrolysate that was significant at above the 99% confidence level. Thus there appears to be an increase in the severity of the pretreatment conditions with the presence of carbonic acid. These results are contrary to previously reported results on aspen wood, where the presence of carbonic acid was not found to have an effect on either the xylose or furan concentrations. Although pretreatment conditions were more severe with the addition of carbonic acid, the presence of carbonic acid resulted in a hydrolysate with a higher final pH. Thus it appears that the higher severity conditions reduce the accumulation of organic acids in the hydrolysate. This result was consistent with previously reported work on carbonic acid pretreatment of aspen wood. Ó 2003 Elsevier Ltd. All rights reserved. Keywords: Carbonic acid pretreatment; Liquid hot water pretreatment; Autohydrolysis; Corn stover; Aspen wood; pH
1. Introduction Conversion of lignocellulosic material to ethanol requires hydrolysis of carbohydrate polymers to their constituent sugars. The use of cellulase is a common approach to hydrolysis and offers the benefits of mild reaction conditions and selective hydrolysis. In order to achieve useful rates of enzymatic hydrolysis, the lignocellulose must first be pretreated to reduce the recalcitrance of the substrate. Pretreatment accomplishes a variety of alterations to the biomass, typically including, to varying degrees: hydrolysis of the hemicellulose, solubilization of lignin and carbohydrate oligomers, and increased accessibility of the cellulose to cellulase enzymes (McMillan, 1994). Several pretreatment methods have been explored to varying degrees. The most com* Corresponding author. Tel.: +1-254-710-3405; fax: +1-254-7103409. E-mail address: [email protected] (G. Peter van Walsum). 1 Present address: Department of Chemical Engineering, Stauffer III, 381 North-South Mall, Stanford University, Stanford, CA 943055025, USA.
0960-8524/$ - see front matter Ó 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2003.11.009
monly reported technologies include dilute-acid pretreatment, in which sulfuric acid is used in low concentrations (on the order of 1%) and at temperatures usually less than 200 °C (Bouchard et al., 1992; Quang et al., 2000; Torget et al., 1996, 2000; Wooley et al., 1999); and steam explosion, which exposes the substrate to steam at elevated temperature and then explosive decompression to physically break apart the plant fibers (Boussaid et al., 2000; Brownell and Saddler, 1987; Montane et al., 1998). Often, steam explosion is coupled with acid catalysis by impregnating the substrate with sulfur dioxide prior to steam treatment (Clark et al., 1989; Mackie et al., 1985; Schell et al., 1998). Other techniques include ammonia fiber explosion (AFEX) that breaks down the lignin using ammonia and explosive decompression (Dale et al., 1999; Holtzapple et al., 1992; Wang et al., 1998), treatment with organic solvents (Avgerinos and Wang, 1983; Bouchard et al., 1993), and treatment with liquid hot water (Allen et al., 1996, 2001; Mok and Antal, 1992; van Walsum et al., 1996). Some methods that have been examined less thoroughly include treatment with supercritical fluids (Muratov and Kim, 2002; Sasaki et al.,
218
G. Peter van Walsum, H. Shi / Bioresource Technology 93 (2004) 217–226
2000) and carbonic acid (McWilliams and van Walsum, 2002; Puri and Mamers, 1983; van Walsum, 2001). Steam-explosion and dilute-acid pretreatment have undergone research and development for many years. Much of this research has been devoted to fuel production from biomass. Dilute-acid pretreatment offers good performance in terms of recovering hemicellulose sugars, but suffers from its use of sulfuric acid. Sulfuric acid is highly corrosive and its neutralization results in copious production of solid wastes (van Walsum, 2001). For applications where landfill costs are high, gypsum disposal costs can detract from the use of dilute sulfuric acid. The calcium sulfate resulting from neutralization also has problematic solubility characteristics in that it becomes less soluble at higher temperatures, such as those encountered in a reboiler (van Walsum, 2001). Compared to dilute-acid pretreatment, steam explosion makes no use of sulfuric acid and is less corrosive to equipment. However, steam explosion yields lower amounts of hemicellulose sugars. Studies suggest an 80% recovery of 5-carbon sugars with dilute-acid pretreatment compared with 65% using steam explosion (Heitz et al., 1991; Quang et al., 2000). SO2 catalyzed steam explosion can achieve recovery of hemicellulose sugars similar to that of sulfuric acid (Tengborg et al., 1998), but this again introduces a reliance on sulfur oxide acids and the requisite neutralization. One process that may offer benefits of acid catalysis without the drawbacks of sulfuric acid is the use of carbonic acid. The pH of carbonic acid is determined by the partial pressure of carbon dioxide in contact with water, and thus it can be neutralized by releasing the reactor pressure. Carbonic acid is relatively mild and hence does not offer the same hydrolytic capability of sulfuric acid. However, van Walsum (2001) has demonstrated that at temperatures on the order of 200 °C, carbonic acid does exhibit a catalytic effect on hydrolysis of xylan. van Walsum observed enhanced release of xylose and low degree of polymerization (DP) xylan oligomers compared to pretreatment using hot water alone. Puri and Mamers (1983) compared steam explosion of biomass with and without carbon dioxide pressurization and reported enhanced enzymatic degradation with the carbonic acid-enhanced steam. Dale and Moreira (1982) found that low temperature (25 °C) carbon dioxide explosion enhanced enzymatic hydrolysis rates of alfalfa, but not to the same extent as did steam pretreatment or AFEX. McWilliams and van Walsum (2002) reported that compared to liquid hot water, high temperature carbonic acid offered no improvement in xylose yields for pretreatment of aspen wood. It was proposed that this may have been due to the high level of endogenous acid produced by the highly acetylated substrate, and suggested that a less acidic substrate may show benefits from employing carbonic acid.
This study seeks to characterize the qualities of hydrolysates produced by liquid hot water and carbonic acid catalyzed pretreatment of corn stover. Metrics of pretreatment effectiveness will include recovery of xylose and low DP oligomers, production of furan compounds and the final pH of the hydrolysate after pretreatment.
2. Methods 2.1. Materials Corn stover and its compositional analysis (Table 1) was kindly supplied by Mark Ruth (2003) at the National Renewable Energy Laboratory in Golden, CO. Aspen wood chips were kindly supplied by the USDA Forest Products Laboratory in Madison WI. Prior to pretreatment, the biomass was ground in a domestic brand coffee grinder and sifted to a particle size of between 0.5 and 1 mm. Dry weight was determined by oven drying. Ash content of the substrates was determined by ashing dry samples in a muffle furnace. Carbon dioxide was standard laboratory grade, and H2 O was of standard laboratory de-ionized quality. 2.2. Hydrolysis Reactions were performed in a simple reactor constructed of 316 stainless steel tubing, 1/2 in. in diameter. The reactor had a volumetric capacity of 15 ml and was filled and emptied by removing a swage connection on one end. For reactions using CO2 , a 1/8 in. stainless steel tubing connection and valve was fitted to the reactor to allow introduction of CO2 from a gas cylinder. Initial pressure of the CO2 was regulated using a high-pressure regulator on the CO2 cylinder. Once charged, the valve on the reactor was closed and the contents of the reactor were left to equilibrate for 15 min, allowing the CO2 gas to go into solution forming carbonic acid. Upon heating to pretreatment temperature, the pressure in the sealed reactor rose. The reaction temperature was controlled Table 1 Compositional analysis of corn stover used in this study (Ruth, 2003) Component
Mass fraction
Cellulose Xylan Arabinan Galactan Mannan Lignin Ash Protein Extractives Acetate Unknown
0.371 0.192 0.025 0.016 0.013 0.207 0.052 0.038 0.026 0.024 0.036
G. Peter van Walsum, H. Shi / Bioresource Technology 93 (2004) 217–226
by immersing the reactor in a fluidized sand bath (Techne, Oxford UK, model SBL 2D) with temperature controller (Techne model TC-8D) that maintained temperatures in the bath to ±1 °C. Reaction temperatures ranged from 180 to 220 °C. Reactions were carried out for durations from 2 to 32 min and were quenched in a cold-water bath. Reactors were brought up to reaction temperature by first immersing them in a preliminary sand bath, maintained at a temperature 40 °C hotter than the intended reaction temperature. After preheating the reactor for 45 s, it was transferred to a sand bath held at reaction temperature and timing for this reaction was started. Further description of the hydrolysis procedure has been reported previously (van Walsum, 2001; McWilliams and van Walsum, 2002). 2.3. Experimental design In each of the reactions, 0.100 g of ground corn stover and 8.0 ml of de-ionized water were used. In the samples reacted with carbonic acid, the reactor was also charged with 800 psi of CO2 at room temperature. The experimental conditions included time intervals of 2, 4, 8, 16 and 32 min and temperatures of 180, 190, 200, 210 and 220 °C. Triplicate experiments were conducted at each severity. Table 2 outlines the conditions tested. 2.4. Severity functions Time and temperature combinations were chosen as outlined in Table 2 to give log severity values in the range of 3.54–4.44 across the experimental space. The log severity values are calculated as first defined by Overend and Chornet (1987). Also listed in Table 2, in bold numerals, are the combined severity values for the experimental conditions. The combined severity function was defined by Chum et al. (1990). In applying the combined severity factor to the carbonic acid system, van Walsum (2001) suggested the following equation, Table 2 Experimental design space and severity values for the time and temperature variables investigated in this study Time (min)
Temperature 180 °C
190 °C
200 °C
32
3.86 0.32 3.56 0.02
4.16 0.58 3.85 0.27 3.55 )0.03
4.45 0.83 4.15 0.53 3.85 0.22 3.55 )0.08
16 8 4 2
210 °C
4.44 0.78 4.14 0.48 3.84 0.17 3.54 )0.13
220 °C
4.44 0.72 4.14 0.42 3.83 0.12
Plain numbers represent log severity, logðR0 Þ, values, bold numbers represent combined severity, logðR0 Þ pH, where the pH is estimated from the fugacity of CO2 in the reactor as represented in Eq. (1).
219
which estimates the value of the solution pH from the temperature and the partial pressure of the CO2 : CSPCO2 ¼ logðRO Þ 8:00 T 2 þ 0:00209 T 0:216 lnðPCO2 Þ þ 3:92
ð1Þ
where CSPCO2 is the combined severity determined from the partial pressure of CO2 , RO is the severity, T is the temperature expressed in degrees Celsius and PCO2 is the partial pressure of CO2 in atmospheres. At pretreatment reaction conditions, CO2 does not behave as an ideal gas but is supercritical, and thus the use of partial pressure in Eq. (1) is an oversimplification that requires justification. Eq. (1) should more precisely employ the fugacity of CO2 in the non-liquid phase. For a binary i, j gas mixture, the multicomponent Virial equation for the fugacity coefficient simplifies to: P ðBii þ yj2 dij Þ with dij ¼ 2Bij Bii Bjj ln ;i ¼ ð2Þ RT where ;i is the fugacity coefficient of component i, R is the universal gas constant, T is the temperature in degrees Kelvin, yj is the molar fraction of the solvent in the gas phase. Bij is the second virial coefficient, which can be expressed using mixing rules proposed by Smith and Van Ness (1975) as: 3 1=3 1=3 Vci þ Vcj 0:422 0:083 1:6 Bij ¼ Tr 4ðZci þ Zcj Þ x þ x 0:172 i j þ ð3Þ 0:139 4:2 Tr 2 where Vc is the critical volume, Zc the critical compressibility, Tr the reduced temperature and x the accentric factor. Eq. (3) can be solved iteratively starting with known variables and estimates for molar fractions obtained by assuming ideal gas behavior. Results from these calculations are summarized in Table 3. Also shown in Table 3 are results for pH calculations made assuming ideal gas behavior for the CO2 /H2 O system. It is seen that at the temperature and pressure ranges investigated for carbonic acid pretreatment, the idealized assumption of partial pressures yields essentially the same pH results as the fugacity-based calculations. 2.5. Relative furan concentration Relative furan concentrations were inferred from hydrolysate absorbance, measured using an UV–visual spectrophotometer (Beckman Corporation, Fullerton CA, model DU 500). The hydrolysates were diluted 10fold with water to improve the spectral resolution. Spectrophotometric scans were conducted from a wavelength of 190 up to 1100 nm in 5 nm increments. The ‘‘Delta Absorbance’’ value was calculated as the difference between the absorbance at 285 nm and the absorbance at 320 nm. The difference between the absorbance at these two wavelengths has been correlated
220
G. Peter van Walsum, H. Shi / Bioresource Technology 93 (2004) 217–226
Table 3 Results of thermodynamic determination of carbonic acid pH at elevated temperatures and pressures Temperature (°C)
Total pressure (psia)
YCO2
;CO2
fCO2 (psia)
pH (fCO2 )
PCO2 (psia)
pH (PCO2 )
180 190 200 210 220
1941
0.925
0.861
2165
0.896
0.864
2416
0.86
0.880
1606 1686 1766 1845 1924
3.54 3.58 3.62 3.67 3.71
1797 1869 1940 2011 2079
3.53 3.58 3.63 3.67 3.72
Results shown for fugacity and idealized partial pressure calculations.
to furan concentration, which in turn has been correlated to microbial toxicity (Martinez et al., 2000).
of the corn stover indicates an ash content of 5.2% (Table 1), thus the corn stover has much higher ash content than the aspen wood.
2.6. pH 3.2. Xylose yield Analysis of pH used a pH meter (Fisher Scientific, Acumet research model AR15) and was done with the samples de-pressurized and at room temperature. 2.7. Soluble carbohydrates Soluble carbohydrate concentrations were measured using high performance anion exchange chromatography with pulsed amperometric detection (HPAE-PAD), a Dionex Carbopac PAX-100 column and a DS3 conductivity cell (Dionex, Sunnyvale, CA) (Anonymous, 1993). An eluent of sodium hydroxide was used with a concentration gradient to elute the sample peaks. Oligomers detected with the HPAE-PAD were quantified by comparing their peak area, which indicates molar concentration, to a standard xylose peak area and recalibrating for the molecular weight of the oligomers. This method had been tested on purified xylan and was found to close a mass balance on partially hydrolysed xylan under a range of severity conditions. 2.8. Data analysis All experiments were carried out in triplicate. Values reported represent mean and standard deviation values of these triplicate experiments. Analysis of variance to determine the effect of carbonic acid on the measured output variables was carried out by using the severity function to collapse reaction time and temperature into one input variable. Then, the level of significance was determined for the effect of carbonic acid on the variables of delta absorbance, pH and xylose concentration.
3. Results 3.1. Ash content Samples of aspen wood had an experimentally determined ash content of about 0.8%. The composition
Mean values of the output variables obtained at each severity condition are listed in Table 4. Plots for the concentration of xylose produced under different reaction conditions are given in Fig. 1. Error bars on the graph represent ± one standard deviation. Comparing the carbonic acid system to the water system, there is an apparent enhancement of xylose production with carbonic acid. Analysis of variance showed this effect to be significant at a level of confidence greater than 99%. Mean xylose concentrations ranged between close to 0 mg/l for the least severe conditions tested (log R0 ¼ 3:54) to 490 mg/l for higher severity conditions (log R0 ¼ 4:45). With an initial xylan concentration of approximately 2.2 g/l, the high xylose value represents about 22.5% of the initial xylan released as monomers. 3.3. Release of oligomers For each sample, release of xylan oligomers was measured using the HPAE. Results from oligomer quantification for corn stover pretreated at 210 °C for various reaction times, with and without carbonic acid, are shown in Fig. 2. It can be seen that as the hydrolysis conditions increased in severity (longer reaction time), the accumulation of smaller chain oligomers increased, peaking with monomer yields being the greatest at highest severity. Comparing the results with and without CO2 , it can be seen that the CO2 enhanced the hydrolysis and increased the accumulation of small chain oligomers. This result was observed at all temperatures investigated. A mass balance was carried out on the xylose and xylan oligomers recovered from the pretreatment. Closure of the xylan mass balance at the different severity conditions was calculated and results are tabulated in Table 4 and plotted versus combined severity in Fig. 3. It can be seen from Fig. 3 that increased severity has a generally positive effect on closure of the xylan balance, but that the highest mass balance numbers were outliers
G. Peter van Walsum, H. Shi / Bioresource Technology 93 (2004) 217–226
221
Table 4 Experimental results of delta absorbance 185–230 nm, final pH, xylose concentration and xylan mass balance closure at different reaction conditions Sample conditions
180 180 190 190 190 200 200 200 200 210 210 210 210 220 220 212
°C, °C, °C, °C, °C, °C, °C, °C, °C, °C, °C, °C, °C, °C, °C, °C,
32 min 16 min 32 min 16 min 8 min 32 min 16 min 8 min 4 min 16 min 8 min 4 min 2 min 8 min 4 min 2 min
D Absorbance 285–320 nm
Final pH
H2 O
H2 CO3
H2 O
H2 CO3
H2 O
H2 CO3
H2 O
H2 CO3
0.099 0.077 0.232 0.132 0.098 0.505 0.209 0.174 0.115 0.271 0.111 0.085 0.065 0.197 0.148 0.080
0.124 0.076 0.296 0.161 0.120 0.595 0.277 0.188 0.137 0.346 0.113 0.085 0.048 0.276 0.154 0.075
3.89 4.35 3.93 4.17 4.61 3.65 4.01 4.39 4.72 3.95 4.20 4.70 5.31 3.80 4.09 4.77
4.87 5.70 4.43 4.85 5.66 3.86 4.37 4.94 5.87 4.23 4.47 5.50 5.96 4.09 4.78 5.40
67.0 19.1 216.8 51.8 9.6 409.2 140.1 29.2 16.2 290.6 86.7 11.0 3.0 280.1 49.7 6.8
175.0 48.9 425.1 155.1 40.4 454.9 260.0 101.3 26.4 483.5 183.2 29.5 4.6 409.4 107.1 32.8
0.44 0.36 0.44 0.24 0.19 0.47 0.40 0.16 0.08 0.35 0.22 0.07 0.05 0.45 0.25 0.16
0.68 0.35 0.59 0.28 0.28 0.46 0.44 0.27 0.11 0.46 0.33 0.13 0.06 0.46 0.33 0.17
Fig. 1. Xylose concentration (mg/l) in corn stover hydrolysates versus temperature of reaction. Filled symbols ¼ carbonic acid. Hollow symbols ¼ water-only. ( ) ¼ 2 min, ðr; }Þ ¼ 4 min, ðj; Þ ¼ 8 min, ðd; sÞ ¼ 16 min and ðN; MÞ ¼ 32 min. Error bars represent ± one standard deviation.
Xylose (mg/l)
Xylan mass balance closure
Fig. 2. Xylan oligomers accumulated in corn stover hydrolysates at 210 °C and various reaction durations. Filled symbols ¼ carbonic acid. Hollow symbols ¼ water-only. ( ) ¼ 2 min, ðr; }Þ ¼ 4 min, ðj; Þ ¼ 8 min and ðd; sÞ ¼ 16 min.
3.4. Delta absorption generated at mid severity and at low temperature. For example, the highest xylan recovery in this study is recorded at 32 min reaction time and 180 °C, the lowest temperature investigated.
The observed delta absorption generally increased with increasing reaction time and temperature (Fig. 4). At low severity, the delta absorption values were on the
222
G. Peter van Walsum, H. Shi / Bioresource Technology 93 (2004) 217–226
furan degradation products are responsive to the severity of the reaction. As with the xylose measurements, there was an effect indicating an increase in the absorption with the addition of CO2 to the reactor. The effect of CO2 increasing the absorption measurements was significant at above the 99% confidence level. 3.5. Hydrolysate pH
Fig. 3. Mass balance closure on xylan versus combined severity. Filled symbols ¼ carbonic acid. Hollow symbols ¼ water-only. ( ) ¼ 2 min, ðr; }Þ ¼ 4 min, ðj; Þ ¼ 8 min, ðd; sÞ ¼ 16 min and ðN; MÞ ¼ 32 min.
Final pH values are plotted in Fig. 5. Consistent with the xylose and absorbance results, the final (depressurized, room temperature) pH of the hydrolysate indicated a difference between reactions carried out with and without carbonic acid in the reactor. However, the fact that the final pH of the hydrolysate of the carbonic acid reactions was consistently higher than the water-only reactions across the full range of reaction conditions seems counter intuitive. It appears that as severity increases, the difference in final pH values decreases. At low severities, the increase in pH was on the order of 1.0, at higher severity the difference between the two pretreatments narrowed to around a range of 0.25 pH units. Across the whole range, the effect of CO2 addition on
Fig. 4. DUV–vis absorbance (285–320 nm) in corn stover hydrolysate versus temperature of reaction. Filled symbols ¼ carbonic acid. Hollow symbols ¼ water-only. ( ) ¼ 2 min, ðr; }Þ ¼ 4 min, ðj; Þ ¼ 8 min, ðd; sÞ ¼ 16 min and ðN; MÞ ¼ 32 min. Error bars represent ± one standard deviation.
order of 0.06–0.1 and increased to 0.5–0.6 at higher severity. This follows the logical assumption that the
Fig. 5. Final (depressurized, room temperature) pH of corn stover hydrolysate versus temperature of reaction. Filled symbols ¼ carbonic acid. Hollow symbols ¼ water-only. ( ) ¼ 2 min, ðr; }Þ ¼ 4 min, ðj; Þ ¼ 8 min, ðd; sÞ ¼ 16 min and ðN; MÞ ¼ 32 min. Error bars represent ± one standard deviation.
G. Peter van Walsum, H. Shi / Bioresource Technology 93 (2004) 217–226
final pH was found to be statistically significant at above the 99% confidence level. 3.6. Response to reaction severity Xylose data for carbonic acid-pretreated corn fiber were plotted against different calculations of reactor severity (data not shown), with a linear fit made to the data. The severity functions compared were: the combined severity function described by Eq. (1), which uses a theoretically predicted pH at reaction conditions; the combined severity function using the pH that was measured post-reaction; and logðR0 Þ. The R2 values were 0.872, 0.812 and 0.792 respectively, indicating that Eq. (1) offered the best fit to the data. The collected results of the corn stover pretreatment are plotted versus the combined severity in Fig. 6. In this figure, the ordinate axis is directly relevant to the carbonic acid results, but has only nominal relevance to the water pretreatment data, since the conditions for the
223
water pretreatment experiments did not include high CO2 pressures. Nevertheless, it can be seen that the addition of carbonic acid results in discernable differences in all three measured variables for most of the pretreatment severities tested.
4. Discussion 4.1. Xylose recovery The effect of carbonic acid to enhance the hydrolysis of hemicellulose fractions in corn stover appears to be significant at a confidence level above 99%. Release of xylose increased with reaction severity, as expected. Consistent with these results, previous work with pure xylan (van Walsum, 2001) has shown that carbonic acid substantially increases hydrolysis activity in comparison with water alone. Contrary to these results, however, a similar study on aspen wood (McWilliams and van Walsum, 2002) suggested that there was no increased hydrolysis activity with the addition of carbonic acid. A possible explanation for these different results is that autocatalysis of the aspen wood is taking place to a greater extent than it is in corn stover or pure xylan. A high level of endogenous acid production would obscure the action of the carbonic acid. Autocatalysis is thought to result primarily from the release of acetyl groups, which are present at higher concentration in aspen wood (on the order of 5% (Wooley et al., 1999)) than in corn stover (2.4% in this study). Corn stover may also have a higher acid buffering capacity, since it contains more ash than aspen wood (5.2% versus 0.8%). 4.2. Xylan oligomers
Fig. 6. Xylose concentration, delta absorbance and final pH of pretreated corn stover. Filled symbols ¼ carbonic acid. Hollow symbols ¼ water-only. ðj; Þ ¼ xylose concentration, ðd; sÞ ¼ DUV–vis absorbance (AU@285 nm–AU@320 nm) and ðN; MÞ ¼ final pH.
Enhancement of hydrolysis due to the presence of carbonic acid is also evident from the accumulation of xylan oligomers in the reactor. In Fig. 2, it is shown that as reaction severity increases, the average DP of the oligomers decreases and monomeric xylose increases. This effect is amplified with the presence of carbonic acid, with both the accumulation of monomeric xylose and total oligomers exceeding that found in the pretreatment in water alone. This indicates the potential for improved pretreatment with carbonic acid compared to water alone, as this would reduce the requirement for additional hydrolysis of the oligomers prior to fermentation. Closure of the mass balance around xylan is low for both the water and carbonic acid pretreatment systems. The trend in the mass balance data appears to indicate that higher severities than those tested in this study may result in further recovery of xylan as detectable oligomers, though the increasing furan concentrations indicate that higher severity will also result in more
224
G. Peter van Walsum, H. Shi / Bioresource Technology 93 (2004) 217–226
degradation, limiting this possible improvement. In this study, the best conditions for high xylan recovery occurred at low temperature (180 °C) and long residence time (32 min). At these conditions, xylose monomer accumulation was relatively moderate (175 mg/l), indicating that the comparatively high closure on the mass balance was associated with high production and retention of oligomers. This suggests that lower yield of xylan products at higher severity is due in part to degradation of xylose monomers. 4.3. Delta absorption of 285–320 nm Results from measurements of absorption were similar to those found on accumulation of xylose and xylan oligomers: there was an increase in the absorption of the hydrolysates produced with carbonic acid that was significant at above the 99% confidence level. This again suggests that the severity in the carbonic acid reactor is greater than in the water reactor. As with the results on xylose, these results are contrary to previously reported results on aspen wood, where there was no significant effect of carbonic acid detected on accumulation of absorptive compounds. 4.4. Hydrolysate pH Given that reaction severity appears to be enhanced with the addition of carbonic acid to the pretreatment system, it is surprising that the final pH of the carbonic acid treated hydrolysate is consistently higher than the liquid hot water hydrolysate, across all reaction conditions. The difference in pH is in the range of 0.25–1.0, which indicates that the acid concentration in the carbonic acid hydrolysate may be as little as one tenth, and is at most 56% that of the liquid hot water hydrolysate. Presumably the pH at reaction temperature is more acidic in the carbonic acid system––increasing the combined severity––but when the CO2 is released after the reaction, the pH of the solution rises to a level above that of the water-pretreated product. This implies that there is more acid ultimately remaining in the watergenerated hydrolysate. Similar results were reported by McWilliams and van Walsum (2002) where higher final pH resulted in the carbonic acid system but severity, as judged by concentrations of hydrolysis products, appeared in that case to be unaffected by the carbonic acid. A mechanism to explain this phenomenon is unknown, but the result may be important. Since inhibitory byproducts of pretreatment are often acidic, this may indicate lower levels of microbial toxicity in the carbonic acid pretreated hydrolysates. This hypothesis of microbial toxicity has been tested in a recent study (Yourchisin and van Walsum, in press). It was found that carbonic acid- and water-pretreated hydrolysates showed no difference in their inhibition of glucose-fed
Fig. 7. Comparison of final pH values for aspen and corn stover pretreated with carbonic acid and liquid hot water. Filled symbols ¼ carbonic acid. Hollow symbols ¼ water-only. ðj; Þ ¼ Corn stover and ðd; sÞ ¼ Aspen wood.
yeast. Thus, accurate determination of the organic acid concentrations in the final hydrolysates is needed to shed light on this pH phenomenon. Results from an experiment comparing final pH for aspen wood and corn stover, pretreated with both liquid hot water and carbonic acid, are shown in Fig. 7. It can be seen that the final pH for aspen wood is far lower than for corn stover, indicating corn stover has either lower concentrations of endogenously produced organic acids, a higher acid buffering capacity, or a combination of the two. This difference in final pH supports the explanation for why carbonic acid is capable of producing a discernible hydrolysis effect on corn stover but not on aspen wood. Further work should be done to quantify the amount of organic acids present in the hydrolysate so as to close an ‘‘acid balance’’ and expand the explanation for the differing extents of hydrolysis. 4.5. Severity function The increased reaction severity of the carbonic acid system should be predictable through use of the combined severity function, which incorporates the pH into the severity expression. Eq. (1) calculates a combined severity function based on the partial pressure (or fugacity) of CO2 in the reactor. It was found that this function fits the xylose data better than the simple severity function or the combined severity function using the pH measured post-reaction. This result makes sense, as it is the pH during reaction, not post-reaction, that should have the greater affect on hydrolysis rates. It also demonstrates the value of the combined severity function in moderately acidic pH ranges.
G. Peter van Walsum, H. Shi / Bioresource Technology 93 (2004) 217–226
225
The enhanced severity in the presence of carbonic acid and the higher ultimate pH of these systems implies that the carbonic acid is acting primarily as a direct hydrolysis catalyst, rather than through a more complex mechanism of increasing the release of acid-forming groups. This is consistent with earlier work on purified xylan, which showed increased hydrolysis with carbonic acid in a system free of any endogenous acid production (van Walsum, 2001).
Acknowledgements
4.6. High pressure conditions
References
The high pressure conditions used for the carbonic acid systems in this work would present design challenges for large scale pretreatment reactors. Industrial processes in the range of 2000 psi are not unheard of; hydrogenation often operates under similar pressure conditions, for example. And the power requirement for CO2 compression has been found to be a relatively minor consideration (Puri and Mamers, 1983; Jayawardhana and van Walsum, submitted). However, handling solids at these pressures could be problematic. Design configurations that minimize movement of solids and move only liquids against strong pressure gradients would simplify sealing and heat recovery issues. The draw back to such designs would be the likely necessity of batch or semi-batch processing.
Allen, S., Kam, L., Zemann, A., Antal, M., 1996. Fractionation of sugar cane with hot compressed liquid water. Ind. Eng. Chem. Res. 35 (8), 2709–2715. Allen, S.G., Schulman, D., Lichwa, J., Antal, M.J., Laser, M., Lynd, L.R., 2001. A comparison between hot liquid water and steam fractionation of corn fiber. Ind. Eng. Chem. Res. 40, 2934–2941. Anonymous, 1993. Analysis of carbohydrates by high performance anion exchange chromatography with pulsed amperometric detection (HPAE-PAD). Dionex Technical note #20, Dionex Corp. 1228 Titan Way, P.O. Box 3603, Sunnyvale, CA 94088-3603. Avgerinos, G.C., Wang, D.I.C., 1983. Selective solvent delignification for fermentation enhancement. Biotechnol. Bioeng. 15, 67–83. Bouchard, J., Overend, R.P., Chornet, E., 1992. Mechanism of dilute acid hydrolysis of cellulose accounting for its degradation in the solid state. J. Wood Chem. Technol. 12 (3), 335–354. Bouchard, J., Lacelle, S., Chornet, E., Vidal, P.F., Overend, R.P., 1993. Mechanism of depolymerization of cellulose by ethylene glycol solvolysis. Holzforschung 47 (4), 291–296. Boussaid, A.-L., Esteghlalian, A.R., Gregg, D.J., Lee, K.H., Saddler, J.N., 2000. Steam pretreatment of douglas-fir wood chips. Appl. Biochem. Biotechnol. 84–86, 693–705. Brownell, H.H., Saddler, J.N., 1987. Steam pretreatment of lignocellulosic material for enhanced enzymatic hydrolysis. Biotechnol. Bioeng. 29, 228–235. Chum, H.L., Johnson, D.K., Black, S.K., Overend, R.P., 1990. Pretreatment-catalyst effects and the combined severity parameter. Appl. Biochem. Biotechnol. 24/25, 1–14. Clark, T.A., Mackie, K.L., Dare, P.H., McDonald, A.G., 1989. Steam explosion of the softwood Pinus radiata with sulphur dioxide addition. J. Wood Chem. Technol. 9 (2), 135–166. Dale, B.E., Moreira, M.J., 1982. A freeze-explosion technique for increasing cellulose hydrolysis. Biotechnol. Bioeng. Symp. 12, 31–43. Dale, B.E., Weaver, J., Byers, F.M., 1999. Extrusion processing for ammonia fiber explosion (AFEX). Appl. Biochem. Biotechnol. 77– 79, 35–45. Heitz, M., Capek-Menard, E., Koeberle, P.G., Gagne, J., Chornet, E., Overend, R.P., Taylor, J.D., Yu, E., 1991. Fractionation of Populus tremuloides at the pilot plant scale: optimization of steam pretreatment conditions using the STAKE II technology. Bioresour. Technol. 35, 23–32. Holtzapple, M.T., Lundeen, J.E., Sturgis, R., Lewis, J.E., Dale, B.E., 1992. Pretreatment of lignocellulosic municipal solid waste by ammonia fiber explosion (AFEX). Appl. Biochem. Biotechnol. 34/ 35, 5–21. Jayawardhana, K., van Walsum, G.P., Modeling of carbonic acid pretreatment process using ASPEN-Plus. Appl. Biochem. Biotechnol., in press. Mackie, K.L., Brownell, H.H., West, K.L., Saddler, J.N., 1985. Effect of sulphur dioxide and sulfuric acid on steam explosion of aspenwood. J. Wood Chem. Technol. 5 (3), 405–425. Martinez, A., Rodriguez, M.E., York, S.W., Preston, J.F., Ingram, L.O., 2000. Use of UV absorbance to monitor furans in dilute acid hydrolyzates of biomass. Biotechnol. Prog. 16, 637–641.
5. Conclusions Carbonic acid was found to enhance the concentrations of xylose and furan compounds in corn stover hydrolysate as compared to water-only pretreatment. This result is contrary to previously reported results on aspen wood, where no pretreatment enhancement was observed for carbonic acid pretreatment. It was found that the combined severity function, wherein the pH value was determined by the partial pressure of CO2 in the reactor, most adequately described the accumulation of xylose in the carbonic acid system, supporting the notion that the carbonic acid was contributing to the acid catalysis of hydrolysis. In a seemingly contradictory mode, hydrolysates generated with carbonic acid exhibit a higher final pH than hydrolysates generated under identical severity conditions using water alone. These results seem to suggest reduced acid concentrations in the hydrolysate and are in agreement with previous results that investigated carbonic acid pretreatment of aspen wood. This result may indicate that the presence of carbonic acid reduces, or at least slows, the accumulation of organic acids from corn stover. Future research will aim to quantify the accumulation of organic acids in reactions with and without carbonic acid that may explain the unexpected pH phenomenon.
This research was supported by the Department of Environmental Studies and Glasscock Energy Research Center, the University Research Council, and the High School Summer Student Research Program, all at Baylor University. We extend special thanks to Vanessa Castleberry for assistance in sample analysis.
226
G. Peter van Walsum, H. Shi / Bioresource Technology 93 (2004) 217–226
McMillan, J.D., 1994. Pretreatment of Lignocellulosic Biomass. In: Himmel, M.E., Baker, J.O., Overend, R.P. (Eds.), ACS Symposium Series 566, pp. 292–324. McWilliams, R.C., van Walsum, G.P., 2002. Comparison of aspen wood hydrolsates produced by pretreatment with liquid hot water and carbonic acid. Appl. Biochem. Biotechnol. 98–100, 109–134. Mok, W.S.-L., Antal Jr., M.J., 1992. Uncatalyzed solvolysis of whole biomass hemicellulose by hot compressed liquid water. Ind. Eng. Chem. Res. 31 (4), 1157–1161. Montane, D., Farriol, X., Salvado, J., Jollez, P., Chornet, E., 1998. Fractionation of wheat straw by steam-explosion pretreatment and alkali delignification. cellulose pulp and byproducts from hemicellulose and lignin. J. Wood Chem. Technol. 18 (2), 171–191. Muratov, G., Kim, C., 2002. Enzymatic hydrolysis of cotton fibers in supercritical CO2 . Biotechnol. Bioprocess Eng. 7 (2), 85–88. Overend, R.P., Chornet, E., 1987. Fractionation of lignocellulosics by steam-aqueous pretreatments. Phil. Trans. R. Soc. Lond. A321, 523–536. Puri, V.P., Mamers, H., 1983. Explosive pretreatment of lignocellulosic residues with high-pressure carbon dioxide for the production of fermentation substrates. Biotechnol. Bioeng. 25, 3149–3161. Quang, A.N., Tucker, M.P., Keller, F.A., Eddy, F.P., 2000. Two stage dilute-acid pretreatment of softwoods. Appl. Biochem. Biotechnol. 84–86, 561–576. Ruth, M., 2003. National Renewable Energy Laboratory, Golden, CO. Personal communication. Sasaki, M., Fang, Z., Fukushima, Y., Adschiri, T., Arai, K., 2000. Dissolution and hydrolysis of cellulose in subcritical and supercritical water. Ind. Eng. Chem. Res. 39, 2882–2890. Schell, D., Nguyen, Q., Tucker, M., Boynton, B., 1998. Pretreatment of softwood by acid-catalyzed steam explosion followed by alkali extraction. Appl. Biochem. Biotechnol. 70–72, 17–24.
Smith, J.M., Van Ness, H.C., 1975. Introduction to chemical engineering thermodynamics. In: Clark, B.J., Maisel, J.W. (Eds.), Chemical Engineering Series. McGraw-Hill, New York, pp. 271– 272. Tengborg, C., Stenberg, K., Galbe, M., Zacchi, G., Larrson, S., Palmqvist, E., Hahn-Hagerdal, B., 1998. Comparison of SO2 and H2 SO4 impregnation of softwood prior to steam pretreatment on ethanol production. Appl. Biochem. Biotechnol. 70–72, 3–15. Torget, R., Hatzis, C., Hayward, T.K., Hsu, T., Phillipidis, G.P., 1996. Optimization of reverse flow, two-temperature, dilute acid pretreatment to enhance biomass conversion to ethanol. Appl. Biochem. Biotechnol. 57/58, 85–101. Torget, R.W., Kim, J.S., Lee, Y.Y., 2000. Fundamental aspects of dilute acid hydrolysis/fractionation kinetics of hardwood carbohydrates. 1. cellulose hydrolysis. Ind. Eng. Chem. Res. 39, 2817–2825. van Walsum, G.P., 2001. Severity function describing the hydrolysis of xylan using carbonic acid. Appl. Biochem. Biotechnol. 91–93, 317– 329. van Walsum, G.P., Allen, S.G., Spencer, M.J., Laser, M.S., Antal, M.J., Lynd, L.R., 1996. Conversion of lignocellulosics pretreated with liquid hot water to ethanol. Appl. Biochem. Biotechnol. 54/55, 157–170. Wang, L., Dale, B.E., Yurttas, L., Goldwasser, I., 1998. Cost estimates and sensitivity analyses for the ammonia fiber explosion process. Appl. Biochem. Biotechnol. 70–72, 51–66. Wooley, R., Ruth, M., Glassner, D., Sheehan, J., 1999. Process design and costing of bioethanol technology: a tool for determining the status and direction of research and development. Biotechnol. Prog. 15, 794–803. Yourchisin, D., van Walsum, G.P., Comparison of the microbial inhibition and enzymatic hydrolysis rates of liquid and solid fractions produced from pretreatment of biomass with carbonic acid and liquid hot water. Appl. Biochem. Biotechnol., in press.
Carbonic acid enhancement of hydrolysis in aqueous pretreatment of corn stover G. Peter van Walsum *, Helen Shi
1
Department of Environmental Studies and Glasscock Energy Research Center, Baylor University, PO Box 97266, Waco, TX 76798-7266, USA Received 9 May 2003; received in revised form 23 October 2003; accepted 9 November 2003
Abstract Carbonic acid and liquid hot water pretreatments were applied to corn stover. Temperatures ranged from 180 to 220 °C; reaction times varied between 2 and 32 min and prereaction carbon dioxide pressure was either 0 or 800 psig. Over the range of reaction conditions tested, it was found that the presence of carbonic acid had an effect of increasing the concentrations of xylose and furan compounds in the hydrolysate that was significant at above the 99% confidence level. Thus there appears to be an increase in the severity of the pretreatment conditions with the presence of carbonic acid. These results are contrary to previously reported results on aspen wood, where the presence of carbonic acid was not found to have an effect on either the xylose or furan concentrations. Although pretreatment conditions were more severe with the addition of carbonic acid, the presence of carbonic acid resulted in a hydrolysate with a higher final pH. Thus it appears that the higher severity conditions reduce the accumulation of organic acids in the hydrolysate. This result was consistent with previously reported work on carbonic acid pretreatment of aspen wood. Ó 2003 Elsevier Ltd. All rights reserved. Keywords: Carbonic acid pretreatment; Liquid hot water pretreatment; Autohydrolysis; Corn stover; Aspen wood; pH
1. Introduction Conversion of lignocellulosic material to ethanol requires hydrolysis of carbohydrate polymers to their constituent sugars. The use of cellulase is a common approach to hydrolysis and offers the benefits of mild reaction conditions and selective hydrolysis. In order to achieve useful rates of enzymatic hydrolysis, the lignocellulose must first be pretreated to reduce the recalcitrance of the substrate. Pretreatment accomplishes a variety of alterations to the biomass, typically including, to varying degrees: hydrolysis of the hemicellulose, solubilization of lignin and carbohydrate oligomers, and increased accessibility of the cellulose to cellulase enzymes (McMillan, 1994). Several pretreatment methods have been explored to varying degrees. The most com* Corresponding author. Tel.: +1-254-710-3405; fax: +1-254-7103409. E-mail address: [email protected] (G. Peter van Walsum). 1 Present address: Department of Chemical Engineering, Stauffer III, 381 North-South Mall, Stanford University, Stanford, CA 943055025, USA.
0960-8524/$ - see front matter Ó 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.biortech.2003.11.009
monly reported technologies include dilute-acid pretreatment, in which sulfuric acid is used in low concentrations (on the order of 1%) and at temperatures usually less than 200 °C (Bouchard et al., 1992; Quang et al., 2000; Torget et al., 1996, 2000; Wooley et al., 1999); and steam explosion, which exposes the substrate to steam at elevated temperature and then explosive decompression to physically break apart the plant fibers (Boussaid et al., 2000; Brownell and Saddler, 1987; Montane et al., 1998). Often, steam explosion is coupled with acid catalysis by impregnating the substrate with sulfur dioxide prior to steam treatment (Clark et al., 1989; Mackie et al., 1985; Schell et al., 1998). Other techniques include ammonia fiber explosion (AFEX) that breaks down the lignin using ammonia and explosive decompression (Dale et al., 1999; Holtzapple et al., 1992; Wang et al., 1998), treatment with organic solvents (Avgerinos and Wang, 1983; Bouchard et al., 1993), and treatment with liquid hot water (Allen et al., 1996, 2001; Mok and Antal, 1992; van Walsum et al., 1996). Some methods that have been examined less thoroughly include treatment with supercritical fluids (Muratov and Kim, 2002; Sasaki et al.,
218
G. Peter van Walsum, H. Shi / Bioresource Technology 93 (2004) 217–226
2000) and carbonic acid (McWilliams and van Walsum, 2002; Puri and Mamers, 1983; van Walsum, 2001). Steam-explosion and dilute-acid pretreatment have undergone research and development for many years. Much of this research has been devoted to fuel production from biomass. Dilute-acid pretreatment offers good performance in terms of recovering hemicellulose sugars, but suffers from its use of sulfuric acid. Sulfuric acid is highly corrosive and its neutralization results in copious production of solid wastes (van Walsum, 2001). For applications where landfill costs are high, gypsum disposal costs can detract from the use of dilute sulfuric acid. The calcium sulfate resulting from neutralization also has problematic solubility characteristics in that it becomes less soluble at higher temperatures, such as those encountered in a reboiler (van Walsum, 2001). Compared to dilute-acid pretreatment, steam explosion makes no use of sulfuric acid and is less corrosive to equipment. However, steam explosion yields lower amounts of hemicellulose sugars. Studies suggest an 80% recovery of 5-carbon sugars with dilute-acid pretreatment compared with 65% using steam explosion (Heitz et al., 1991; Quang et al., 2000). SO2 catalyzed steam explosion can achieve recovery of hemicellulose sugars similar to that of sulfuric acid (Tengborg et al., 1998), but this again introduces a reliance on sulfur oxide acids and the requisite neutralization. One process that may offer benefits of acid catalysis without the drawbacks of sulfuric acid is the use of carbonic acid. The pH of carbonic acid is determined by the partial pressure of carbon dioxide in contact with water, and thus it can be neutralized by releasing the reactor pressure. Carbonic acid is relatively mild and hence does not offer the same hydrolytic capability of sulfuric acid. However, van Walsum (2001) has demonstrated that at temperatures on the order of 200 °C, carbonic acid does exhibit a catalytic effect on hydrolysis of xylan. van Walsum observed enhanced release of xylose and low degree of polymerization (DP) xylan oligomers compared to pretreatment using hot water alone. Puri and Mamers (1983) compared steam explosion of biomass with and without carbon dioxide pressurization and reported enhanced enzymatic degradation with the carbonic acid-enhanced steam. Dale and Moreira (1982) found that low temperature (25 °C) carbon dioxide explosion enhanced enzymatic hydrolysis rates of alfalfa, but not to the same extent as did steam pretreatment or AFEX. McWilliams and van Walsum (2002) reported that compared to liquid hot water, high temperature carbonic acid offered no improvement in xylose yields for pretreatment of aspen wood. It was proposed that this may have been due to the high level of endogenous acid produced by the highly acetylated substrate, and suggested that a less acidic substrate may show benefits from employing carbonic acid.
This study seeks to characterize the qualities of hydrolysates produced by liquid hot water and carbonic acid catalyzed pretreatment of corn stover. Metrics of pretreatment effectiveness will include recovery of xylose and low DP oligomers, production of furan compounds and the final pH of the hydrolysate after pretreatment.
2. Methods 2.1. Materials Corn stover and its compositional analysis (Table 1) was kindly supplied by Mark Ruth (2003) at the National Renewable Energy Laboratory in Golden, CO. Aspen wood chips were kindly supplied by the USDA Forest Products Laboratory in Madison WI. Prior to pretreatment, the biomass was ground in a domestic brand coffee grinder and sifted to a particle size of between 0.5 and 1 mm. Dry weight was determined by oven drying. Ash content of the substrates was determined by ashing dry samples in a muffle furnace. Carbon dioxide was standard laboratory grade, and H2 O was of standard laboratory de-ionized quality. 2.2. Hydrolysis Reactions were performed in a simple reactor constructed of 316 stainless steel tubing, 1/2 in. in diameter. The reactor had a volumetric capacity of 15 ml and was filled and emptied by removing a swage connection on one end. For reactions using CO2 , a 1/8 in. stainless steel tubing connection and valve was fitted to the reactor to allow introduction of CO2 from a gas cylinder. Initial pressure of the CO2 was regulated using a high-pressure regulator on the CO2 cylinder. Once charged, the valve on the reactor was closed and the contents of the reactor were left to equilibrate for 15 min, allowing the CO2 gas to go into solution forming carbonic acid. Upon heating to pretreatment temperature, the pressure in the sealed reactor rose. The reaction temperature was controlled Table 1 Compositional analysis of corn stover used in this study (Ruth, 2003) Component
Mass fraction
Cellulose Xylan Arabinan Galactan Mannan Lignin Ash Protein Extractives Acetate Unknown
0.371 0.192 0.025 0.016 0.013 0.207 0.052 0.038 0.026 0.024 0.036
G. Peter van Walsum, H. Shi / Bioresource Technology 93 (2004) 217–226
by immersing the reactor in a fluidized sand bath (Techne, Oxford UK, model SBL 2D) with temperature controller (Techne model TC-8D) that maintained temperatures in the bath to ±1 °C. Reaction temperatures ranged from 180 to 220 °C. Reactions were carried out for durations from 2 to 32 min and were quenched in a cold-water bath. Reactors were brought up to reaction temperature by first immersing them in a preliminary sand bath, maintained at a temperature 40 °C hotter than the intended reaction temperature. After preheating the reactor for 45 s, it was transferred to a sand bath held at reaction temperature and timing for this reaction was started. Further description of the hydrolysis procedure has been reported previously (van Walsum, 2001; McWilliams and van Walsum, 2002). 2.3. Experimental design In each of the reactions, 0.100 g of ground corn stover and 8.0 ml of de-ionized water were used. In the samples reacted with carbonic acid, the reactor was also charged with 800 psi of CO2 at room temperature. The experimental conditions included time intervals of 2, 4, 8, 16 and 32 min and temperatures of 180, 190, 200, 210 and 220 °C. Triplicate experiments were conducted at each severity. Table 2 outlines the conditions tested. 2.4. Severity functions Time and temperature combinations were chosen as outlined in Table 2 to give log severity values in the range of 3.54–4.44 across the experimental space. The log severity values are calculated as first defined by Overend and Chornet (1987). Also listed in Table 2, in bold numerals, are the combined severity values for the experimental conditions. The combined severity function was defined by Chum et al. (1990). In applying the combined severity factor to the carbonic acid system, van Walsum (2001) suggested the following equation, Table 2 Experimental design space and severity values for the time and temperature variables investigated in this study Time (min)
Temperature 180 °C
190 °C
200 °C
32
3.86 0.32 3.56 0.02
4.16 0.58 3.85 0.27 3.55 )0.03
4.45 0.83 4.15 0.53 3.85 0.22 3.55 )0.08
16 8 4 2
210 °C
4.44 0.78 4.14 0.48 3.84 0.17 3.54 )0.13
220 °C
4.44 0.72 4.14 0.42 3.83 0.12
Plain numbers represent log severity, logðR0 Þ, values, bold numbers represent combined severity, logðR0 Þ pH, where the pH is estimated from the fugacity of CO2 in the reactor as represented in Eq. (1).
219
which estimates the value of the solution pH from the temperature and the partial pressure of the CO2 : CSPCO2 ¼ logðRO Þ 8:00 T 2 þ 0:00209 T 0:216 lnðPCO2 Þ þ 3:92
ð1Þ
where CSPCO2 is the combined severity determined from the partial pressure of CO2 , RO is the severity, T is the temperature expressed in degrees Celsius and PCO2 is the partial pressure of CO2 in atmospheres. At pretreatment reaction conditions, CO2 does not behave as an ideal gas but is supercritical, and thus the use of partial pressure in Eq. (1) is an oversimplification that requires justification. Eq. (1) should more precisely employ the fugacity of CO2 in the non-liquid phase. For a binary i, j gas mixture, the multicomponent Virial equation for the fugacity coefficient simplifies to: P ðBii þ yj2 dij Þ with dij ¼ 2Bij Bii Bjj ln ;i ¼ ð2Þ RT where ;i is the fugacity coefficient of component i, R is the universal gas constant, T is the temperature in degrees Kelvin, yj is the molar fraction of the solvent in the gas phase. Bij is the second virial coefficient, which can be expressed using mixing rules proposed by Smith and Van Ness (1975) as: 3 1=3 1=3 Vci þ Vcj 0:422 0:083 1:6 Bij ¼ Tr 4ðZci þ Zcj Þ x þ x 0:172 i j þ ð3Þ 0:139 4:2 Tr 2 where Vc is the critical volume, Zc the critical compressibility, Tr the reduced temperature and x the accentric factor. Eq. (3) can be solved iteratively starting with known variables and estimates for molar fractions obtained by assuming ideal gas behavior. Results from these calculations are summarized in Table 3. Also shown in Table 3 are results for pH calculations made assuming ideal gas behavior for the CO2 /H2 O system. It is seen that at the temperature and pressure ranges investigated for carbonic acid pretreatment, the idealized assumption of partial pressures yields essentially the same pH results as the fugacity-based calculations. 2.5. Relative furan concentration Relative furan concentrations were inferred from hydrolysate absorbance, measured using an UV–visual spectrophotometer (Beckman Corporation, Fullerton CA, model DU 500). The hydrolysates were diluted 10fold with water to improve the spectral resolution. Spectrophotometric scans were conducted from a wavelength of 190 up to 1100 nm in 5 nm increments. The ‘‘Delta Absorbance’’ value was calculated as the difference between the absorbance at 285 nm and the absorbance at 320 nm. The difference between the absorbance at these two wavelengths has been correlated
220
G. Peter van Walsum, H. Shi / Bioresource Technology 93 (2004) 217–226
Table 3 Results of thermodynamic determination of carbonic acid pH at elevated temperatures and pressures Temperature (°C)
Total pressure (psia)
YCO2
;CO2
fCO2 (psia)
pH (fCO2 )
PCO2 (psia)
pH (PCO2 )
180 190 200 210 220
1941
0.925
0.861
2165
0.896
0.864
2416
0.86
0.880
1606 1686 1766 1845 1924
3.54 3.58 3.62 3.67 3.71
1797 1869 1940 2011 2079
3.53 3.58 3.63 3.67 3.72
Results shown for fugacity and idealized partial pressure calculations.
to furan concentration, which in turn has been correlated to microbial toxicity (Martinez et al., 2000).
of the corn stover indicates an ash content of 5.2% (Table 1), thus the corn stover has much higher ash content than the aspen wood.
2.6. pH 3.2. Xylose yield Analysis of pH used a pH meter (Fisher Scientific, Acumet research model AR15) and was done with the samples de-pressurized and at room temperature. 2.7. Soluble carbohydrates Soluble carbohydrate concentrations were measured using high performance anion exchange chromatography with pulsed amperometric detection (HPAE-PAD), a Dionex Carbopac PAX-100 column and a DS3 conductivity cell (Dionex, Sunnyvale, CA) (Anonymous, 1993). An eluent of sodium hydroxide was used with a concentration gradient to elute the sample peaks. Oligomers detected with the HPAE-PAD were quantified by comparing their peak area, which indicates molar concentration, to a standard xylose peak area and recalibrating for the molecular weight of the oligomers. This method had been tested on purified xylan and was found to close a mass balance on partially hydrolysed xylan under a range of severity conditions. 2.8. Data analysis All experiments were carried out in triplicate. Values reported represent mean and standard deviation values of these triplicate experiments. Analysis of variance to determine the effect of carbonic acid on the measured output variables was carried out by using the severity function to collapse reaction time and temperature into one input variable. Then, the level of significance was determined for the effect of carbonic acid on the variables of delta absorbance, pH and xylose concentration.
3. Results 3.1. Ash content Samples of aspen wood had an experimentally determined ash content of about 0.8%. The composition
Mean values of the output variables obtained at each severity condition are listed in Table 4. Plots for the concentration of xylose produced under different reaction conditions are given in Fig. 1. Error bars on the graph represent ± one standard deviation. Comparing the carbonic acid system to the water system, there is an apparent enhancement of xylose production with carbonic acid. Analysis of variance showed this effect to be significant at a level of confidence greater than 99%. Mean xylose concentrations ranged between close to 0 mg/l for the least severe conditions tested (log R0 ¼ 3:54) to 490 mg/l for higher severity conditions (log R0 ¼ 4:45). With an initial xylan concentration of approximately 2.2 g/l, the high xylose value represents about 22.5% of the initial xylan released as monomers. 3.3. Release of oligomers For each sample, release of xylan oligomers was measured using the HPAE. Results from oligomer quantification for corn stover pretreated at 210 °C for various reaction times, with and without carbonic acid, are shown in Fig. 2. It can be seen that as the hydrolysis conditions increased in severity (longer reaction time), the accumulation of smaller chain oligomers increased, peaking with monomer yields being the greatest at highest severity. Comparing the results with and without CO2 , it can be seen that the CO2 enhanced the hydrolysis and increased the accumulation of small chain oligomers. This result was observed at all temperatures investigated. A mass balance was carried out on the xylose and xylan oligomers recovered from the pretreatment. Closure of the xylan mass balance at the different severity conditions was calculated and results are tabulated in Table 4 and plotted versus combined severity in Fig. 3. It can be seen from Fig. 3 that increased severity has a generally positive effect on closure of the xylan balance, but that the highest mass balance numbers were outliers
G. Peter van Walsum, H. Shi / Bioresource Technology 93 (2004) 217–226
221
Table 4 Experimental results of delta absorbance 185–230 nm, final pH, xylose concentration and xylan mass balance closure at different reaction conditions Sample conditions
180 180 190 190 190 200 200 200 200 210 210 210 210 220 220 212
°C, °C, °C, °C, °C, °C, °C, °C, °C, °C, °C, °C, °C, °C, °C, °C,
32 min 16 min 32 min 16 min 8 min 32 min 16 min 8 min 4 min 16 min 8 min 4 min 2 min 8 min 4 min 2 min
D Absorbance 285–320 nm
Final pH
H2 O
H2 CO3
H2 O
H2 CO3
H2 O
H2 CO3
H2 O
H2 CO3
0.099 0.077 0.232 0.132 0.098 0.505 0.209 0.174 0.115 0.271 0.111 0.085 0.065 0.197 0.148 0.080
0.124 0.076 0.296 0.161 0.120 0.595 0.277 0.188 0.137 0.346 0.113 0.085 0.048 0.276 0.154 0.075
3.89 4.35 3.93 4.17 4.61 3.65 4.01 4.39 4.72 3.95 4.20 4.70 5.31 3.80 4.09 4.77
4.87 5.70 4.43 4.85 5.66 3.86 4.37 4.94 5.87 4.23 4.47 5.50 5.96 4.09 4.78 5.40
67.0 19.1 216.8 51.8 9.6 409.2 140.1 29.2 16.2 290.6 86.7 11.0 3.0 280.1 49.7 6.8
175.0 48.9 425.1 155.1 40.4 454.9 260.0 101.3 26.4 483.5 183.2 29.5 4.6 409.4 107.1 32.8
0.44 0.36 0.44 0.24 0.19 0.47 0.40 0.16 0.08 0.35 0.22 0.07 0.05 0.45 0.25 0.16
0.68 0.35 0.59 0.28 0.28 0.46 0.44 0.27 0.11 0.46 0.33 0.13 0.06 0.46 0.33 0.17
Fig. 1. Xylose concentration (mg/l) in corn stover hydrolysates versus temperature of reaction. Filled symbols ¼ carbonic acid. Hollow symbols ¼ water-only. ( ) ¼ 2 min, ðr; }Þ ¼ 4 min, ðj; Þ ¼ 8 min, ðd; sÞ ¼ 16 min and ðN; MÞ ¼ 32 min. Error bars represent ± one standard deviation.
Xylose (mg/l)
Xylan mass balance closure
Fig. 2. Xylan oligomers accumulated in corn stover hydrolysates at 210 °C and various reaction durations. Filled symbols ¼ carbonic acid. Hollow symbols ¼ water-only. ( ) ¼ 2 min, ðr; }Þ ¼ 4 min, ðj; Þ ¼ 8 min and ðd; sÞ ¼ 16 min.
3.4. Delta absorption generated at mid severity and at low temperature. For example, the highest xylan recovery in this study is recorded at 32 min reaction time and 180 °C, the lowest temperature investigated.
The observed delta absorption generally increased with increasing reaction time and temperature (Fig. 4). At low severity, the delta absorption values were on the
222
G. Peter van Walsum, H. Shi / Bioresource Technology 93 (2004) 217–226
furan degradation products are responsive to the severity of the reaction. As with the xylose measurements, there was an effect indicating an increase in the absorption with the addition of CO2 to the reactor. The effect of CO2 increasing the absorption measurements was significant at above the 99% confidence level. 3.5. Hydrolysate pH
Fig. 3. Mass balance closure on xylan versus combined severity. Filled symbols ¼ carbonic acid. Hollow symbols ¼ water-only. ( ) ¼ 2 min, ðr; }Þ ¼ 4 min, ðj; Þ ¼ 8 min, ðd; sÞ ¼ 16 min and ðN; MÞ ¼ 32 min.
Final pH values are plotted in Fig. 5. Consistent with the xylose and absorbance results, the final (depressurized, room temperature) pH of the hydrolysate indicated a difference between reactions carried out with and without carbonic acid in the reactor. However, the fact that the final pH of the hydrolysate of the carbonic acid reactions was consistently higher than the water-only reactions across the full range of reaction conditions seems counter intuitive. It appears that as severity increases, the difference in final pH values decreases. At low severities, the increase in pH was on the order of 1.0, at higher severity the difference between the two pretreatments narrowed to around a range of 0.25 pH units. Across the whole range, the effect of CO2 addition on
Fig. 4. DUV–vis absorbance (285–320 nm) in corn stover hydrolysate versus temperature of reaction. Filled symbols ¼ carbonic acid. Hollow symbols ¼ water-only. ( ) ¼ 2 min, ðr; }Þ ¼ 4 min, ðj; Þ ¼ 8 min, ðd; sÞ ¼ 16 min and ðN; MÞ ¼ 32 min. Error bars represent ± one standard deviation.
order of 0.06–0.1 and increased to 0.5–0.6 at higher severity. This follows the logical assumption that the
Fig. 5. Final (depressurized, room temperature) pH of corn stover hydrolysate versus temperature of reaction. Filled symbols ¼ carbonic acid. Hollow symbols ¼ water-only. ( ) ¼ 2 min, ðr; }Þ ¼ 4 min, ðj; Þ ¼ 8 min, ðd; sÞ ¼ 16 min and ðN; MÞ ¼ 32 min. Error bars represent ± one standard deviation.
G. Peter van Walsum, H. Shi / Bioresource Technology 93 (2004) 217–226
final pH was found to be statistically significant at above the 99% confidence level. 3.6. Response to reaction severity Xylose data for carbonic acid-pretreated corn fiber were plotted against different calculations of reactor severity (data not shown), with a linear fit made to the data. The severity functions compared were: the combined severity function described by Eq. (1), which uses a theoretically predicted pH at reaction conditions; the combined severity function using the pH that was measured post-reaction; and logðR0 Þ. The R2 values were 0.872, 0.812 and 0.792 respectively, indicating that Eq. (1) offered the best fit to the data. The collected results of the corn stover pretreatment are plotted versus the combined severity in Fig. 6. In this figure, the ordinate axis is directly relevant to the carbonic acid results, but has only nominal relevance to the water pretreatment data, since the conditions for the
223
water pretreatment experiments did not include high CO2 pressures. Nevertheless, it can be seen that the addition of carbonic acid results in discernable differences in all three measured variables for most of the pretreatment severities tested.
4. Discussion 4.1. Xylose recovery The effect of carbonic acid to enhance the hydrolysis of hemicellulose fractions in corn stover appears to be significant at a confidence level above 99%. Release of xylose increased with reaction severity, as expected. Consistent with these results, previous work with pure xylan (van Walsum, 2001) has shown that carbonic acid substantially increases hydrolysis activity in comparison with water alone. Contrary to these results, however, a similar study on aspen wood (McWilliams and van Walsum, 2002) suggested that there was no increased hydrolysis activity with the addition of carbonic acid. A possible explanation for these different results is that autocatalysis of the aspen wood is taking place to a greater extent than it is in corn stover or pure xylan. A high level of endogenous acid production would obscure the action of the carbonic acid. Autocatalysis is thought to result primarily from the release of acetyl groups, which are present at higher concentration in aspen wood (on the order of 5% (Wooley et al., 1999)) than in corn stover (2.4% in this study). Corn stover may also have a higher acid buffering capacity, since it contains more ash than aspen wood (5.2% versus 0.8%). 4.2. Xylan oligomers
Fig. 6. Xylose concentration, delta absorbance and final pH of pretreated corn stover. Filled symbols ¼ carbonic acid. Hollow symbols ¼ water-only. ðj; Þ ¼ xylose concentration, ðd; sÞ ¼ DUV–vis absorbance (AU@285 nm–AU@320 nm) and ðN; MÞ ¼ final pH.
Enhancement of hydrolysis due to the presence of carbonic acid is also evident from the accumulation of xylan oligomers in the reactor. In Fig. 2, it is shown that as reaction severity increases, the average DP of the oligomers decreases and monomeric xylose increases. This effect is amplified with the presence of carbonic acid, with both the accumulation of monomeric xylose and total oligomers exceeding that found in the pretreatment in water alone. This indicates the potential for improved pretreatment with carbonic acid compared to water alone, as this would reduce the requirement for additional hydrolysis of the oligomers prior to fermentation. Closure of the mass balance around xylan is low for both the water and carbonic acid pretreatment systems. The trend in the mass balance data appears to indicate that higher severities than those tested in this study may result in further recovery of xylan as detectable oligomers, though the increasing furan concentrations indicate that higher severity will also result in more
224
G. Peter van Walsum, H. Shi / Bioresource Technology 93 (2004) 217–226
degradation, limiting this possible improvement. In this study, the best conditions for high xylan recovery occurred at low temperature (180 °C) and long residence time (32 min). At these conditions, xylose monomer accumulation was relatively moderate (175 mg/l), indicating that the comparatively high closure on the mass balance was associated with high production and retention of oligomers. This suggests that lower yield of xylan products at higher severity is due in part to degradation of xylose monomers. 4.3. Delta absorption of 285–320 nm Results from measurements of absorption were similar to those found on accumulation of xylose and xylan oligomers: there was an increase in the absorption of the hydrolysates produced with carbonic acid that was significant at above the 99% confidence level. This again suggests that the severity in the carbonic acid reactor is greater than in the water reactor. As with the results on xylose, these results are contrary to previously reported results on aspen wood, where there was no significant effect of carbonic acid detected on accumulation of absorptive compounds. 4.4. Hydrolysate pH Given that reaction severity appears to be enhanced with the addition of carbonic acid to the pretreatment system, it is surprising that the final pH of the carbonic acid treated hydrolysate is consistently higher than the liquid hot water hydrolysate, across all reaction conditions. The difference in pH is in the range of 0.25–1.0, which indicates that the acid concentration in the carbonic acid hydrolysate may be as little as one tenth, and is at most 56% that of the liquid hot water hydrolysate. Presumably the pH at reaction temperature is more acidic in the carbonic acid system––increasing the combined severity––but when the CO2 is released after the reaction, the pH of the solution rises to a level above that of the water-pretreated product. This implies that there is more acid ultimately remaining in the watergenerated hydrolysate. Similar results were reported by McWilliams and van Walsum (2002) where higher final pH resulted in the carbonic acid system but severity, as judged by concentrations of hydrolysis products, appeared in that case to be unaffected by the carbonic acid. A mechanism to explain this phenomenon is unknown, but the result may be important. Since inhibitory byproducts of pretreatment are often acidic, this may indicate lower levels of microbial toxicity in the carbonic acid pretreated hydrolysates. This hypothesis of microbial toxicity has been tested in a recent study (Yourchisin and van Walsum, in press). It was found that carbonic acid- and water-pretreated hydrolysates showed no difference in their inhibition of glucose-fed
Fig. 7. Comparison of final pH values for aspen and corn stover pretreated with carbonic acid and liquid hot water. Filled symbols ¼ carbonic acid. Hollow symbols ¼ water-only. ðj; Þ ¼ Corn stover and ðd; sÞ ¼ Aspen wood.
yeast. Thus, accurate determination of the organic acid concentrations in the final hydrolysates is needed to shed light on this pH phenomenon. Results from an experiment comparing final pH for aspen wood and corn stover, pretreated with both liquid hot water and carbonic acid, are shown in Fig. 7. It can be seen that the final pH for aspen wood is far lower than for corn stover, indicating corn stover has either lower concentrations of endogenously produced organic acids, a higher acid buffering capacity, or a combination of the two. This difference in final pH supports the explanation for why carbonic acid is capable of producing a discernible hydrolysis effect on corn stover but not on aspen wood. Further work should be done to quantify the amount of organic acids present in the hydrolysate so as to close an ‘‘acid balance’’ and expand the explanation for the differing extents of hydrolysis. 4.5. Severity function The increased reaction severity of the carbonic acid system should be predictable through use of the combined severity function, which incorporates the pH into the severity expression. Eq. (1) calculates a combined severity function based on the partial pressure (or fugacity) of CO2 in the reactor. It was found that this function fits the xylose data better than the simple severity function or the combined severity function using the pH measured post-reaction. This result makes sense, as it is the pH during reaction, not post-reaction, that should have the greater affect on hydrolysis rates. It also demonstrates the value of the combined severity function in moderately acidic pH ranges.
G. Peter van Walsum, H. Shi / Bioresource Technology 93 (2004) 217–226
225
The enhanced severity in the presence of carbonic acid and the higher ultimate pH of these systems implies that the carbonic acid is acting primarily as a direct hydrolysis catalyst, rather than through a more complex mechanism of increasing the release of acid-forming groups. This is consistent with earlier work on purified xylan, which showed increased hydrolysis with carbonic acid in a system free of any endogenous acid production (van Walsum, 2001).
Acknowledgements
4.6. High pressure conditions
References
The high pressure conditions used for the carbonic acid systems in this work would present design challenges for large scale pretreatment reactors. Industrial processes in the range of 2000 psi are not unheard of; hydrogenation often operates under similar pressure conditions, for example. And the power requirement for CO2 compression has been found to be a relatively minor consideration (Puri and Mamers, 1983; Jayawardhana and van Walsum, submitted). However, handling solids at these pressures could be problematic. Design configurations that minimize movement of solids and move only liquids against strong pressure gradients would simplify sealing and heat recovery issues. The draw back to such designs would be the likely necessity of batch or semi-batch processing.
Allen, S., Kam, L., Zemann, A., Antal, M., 1996. Fractionation of sugar cane with hot compressed liquid water. Ind. Eng. Chem. Res. 35 (8), 2709–2715. Allen, S.G., Schulman, D., Lichwa, J., Antal, M.J., Laser, M., Lynd, L.R., 2001. A comparison between hot liquid water and steam fractionation of corn fiber. Ind. Eng. Chem. Res. 40, 2934–2941. Anonymous, 1993. Analysis of carbohydrates by high performance anion exchange chromatography with pulsed amperometric detection (HPAE-PAD). Dionex Technical note #20, Dionex Corp. 1228 Titan Way, P.O. Box 3603, Sunnyvale, CA 94088-3603. Avgerinos, G.C., Wang, D.I.C., 1983. Selective solvent delignification for fermentation enhancement. Biotechnol. Bioeng. 15, 67–83. Bouchard, J., Overend, R.P., Chornet, E., 1992. Mechanism of dilute acid hydrolysis of cellulose accounting for its degradation in the solid state. J. Wood Chem. Technol. 12 (3), 335–354. Bouchard, J., Lacelle, S., Chornet, E., Vidal, P.F., Overend, R.P., 1993. Mechanism of depolymerization of cellulose by ethylene glycol solvolysis. Holzforschung 47 (4), 291–296. Boussaid, A.-L., Esteghlalian, A.R., Gregg, D.J., Lee, K.H., Saddler, J.N., 2000. Steam pretreatment of douglas-fir wood chips. Appl. Biochem. Biotechnol. 84–86, 693–705. Brownell, H.H., Saddler, J.N., 1987. Steam pretreatment of lignocellulosic material for enhanced enzymatic hydrolysis. Biotechnol. Bioeng. 29, 228–235. Chum, H.L., Johnson, D.K., Black, S.K., Overend, R.P., 1990. Pretreatment-catalyst effects and the combined severity parameter. Appl. Biochem. Biotechnol. 24/25, 1–14. Clark, T.A., Mackie, K.L., Dare, P.H., McDonald, A.G., 1989. Steam explosion of the softwood Pinus radiata with sulphur dioxide addition. J. Wood Chem. Technol. 9 (2), 135–166. Dale, B.E., Moreira, M.J., 1982. A freeze-explosion technique for increasing cellulose hydrolysis. Biotechnol. Bioeng. Symp. 12, 31–43. Dale, B.E., Weaver, J., Byers, F.M., 1999. Extrusion processing for ammonia fiber explosion (AFEX). Appl. Biochem. Biotechnol. 77– 79, 35–45. Heitz, M., Capek-Menard, E., Koeberle, P.G., Gagne, J., Chornet, E., Overend, R.P., Taylor, J.D., Yu, E., 1991. Fractionation of Populus tremuloides at the pilot plant scale: optimization of steam pretreatment conditions using the STAKE II technology. Bioresour. Technol. 35, 23–32. Holtzapple, M.T., Lundeen, J.E., Sturgis, R., Lewis, J.E., Dale, B.E., 1992. Pretreatment of lignocellulosic municipal solid waste by ammonia fiber explosion (AFEX). Appl. Biochem. Biotechnol. 34/ 35, 5–21. Jayawardhana, K., van Walsum, G.P., Modeling of carbonic acid pretreatment process using ASPEN-Plus. Appl. Biochem. Biotechnol., in press. Mackie, K.L., Brownell, H.H., West, K.L., Saddler, J.N., 1985. Effect of sulphur dioxide and sulfuric acid on steam explosion of aspenwood. J. Wood Chem. Technol. 5 (3), 405–425. Martinez, A., Rodriguez, M.E., York, S.W., Preston, J.F., Ingram, L.O., 2000. Use of UV absorbance to monitor furans in dilute acid hydrolyzates of biomass. Biotechnol. Prog. 16, 637–641.
5. Conclusions Carbonic acid was found to enhance the concentrations of xylose and furan compounds in corn stover hydrolysate as compared to water-only pretreatment. This result is contrary to previously reported results on aspen wood, where no pretreatment enhancement was observed for carbonic acid pretreatment. It was found that the combined severity function, wherein the pH value was determined by the partial pressure of CO2 in the reactor, most adequately described the accumulation of xylose in the carbonic acid system, supporting the notion that the carbonic acid was contributing to the acid catalysis of hydrolysis. In a seemingly contradictory mode, hydrolysates generated with carbonic acid exhibit a higher final pH than hydrolysates generated under identical severity conditions using water alone. These results seem to suggest reduced acid concentrations in the hydrolysate and are in agreement with previous results that investigated carbonic acid pretreatment of aspen wood. This result may indicate that the presence of carbonic acid reduces, or at least slows, the accumulation of organic acids from corn stover. Future research will aim to quantify the accumulation of organic acids in reactions with and without carbonic acid that may explain the unexpected pH phenomenon.
This research was supported by the Department of Environmental Studies and Glasscock Energy Research Center, the University Research Council, and the High School Summer Student Research Program, all at Baylor University. We extend special thanks to Vanessa Castleberry for assistance in sample analysis.
226
G. Peter van Walsum, H. Shi / Bioresource Technology 93 (2004) 217–226
McMillan, J.D., 1994. Pretreatment of Lignocellulosic Biomass. In: Himmel, M.E., Baker, J.O., Overend, R.P. (Eds.), ACS Symposium Series 566, pp. 292–324. McWilliams, R.C., van Walsum, G.P., 2002. Comparison of aspen wood hydrolsates produced by pretreatment with liquid hot water and carbonic acid. Appl. Biochem. Biotechnol. 98–100, 109–134. Mok, W.S.-L., Antal Jr., M.J., 1992. Uncatalyzed solvolysis of whole biomass hemicellulose by hot compressed liquid water. Ind. Eng. Chem. Res. 31 (4), 1157–1161. Montane, D., Farriol, X., Salvado, J., Jollez, P., Chornet, E., 1998. Fractionation of wheat straw by steam-explosion pretreatment and alkali delignification. cellulose pulp and byproducts from hemicellulose and lignin. J. Wood Chem. Technol. 18 (2), 171–191. Muratov, G., Kim, C., 2002. Enzymatic hydrolysis of cotton fibers in supercritical CO2 . Biotechnol. Bioprocess Eng. 7 (2), 85–88. Overend, R.P., Chornet, E., 1987. Fractionation of lignocellulosics by steam-aqueous pretreatments. Phil. Trans. R. Soc. Lond. A321, 523–536. Puri, V.P., Mamers, H., 1983. Explosive pretreatment of lignocellulosic residues with high-pressure carbon dioxide for the production of fermentation substrates. Biotechnol. Bioeng. 25, 3149–3161. Quang, A.N., Tucker, M.P., Keller, F.A., Eddy, F.P., 2000. Two stage dilute-acid pretreatment of softwoods. Appl. Biochem. Biotechnol. 84–86, 561–576. Ruth, M., 2003. National Renewable Energy Laboratory, Golden, CO. Personal communication. Sasaki, M., Fang, Z., Fukushima, Y., Adschiri, T., Arai, K., 2000. Dissolution and hydrolysis of cellulose in subcritical and supercritical water. Ind. Eng. Chem. Res. 39, 2882–2890. Schell, D., Nguyen, Q., Tucker, M., Boynton, B., 1998. Pretreatment of softwood by acid-catalyzed steam explosion followed by alkali extraction. Appl. Biochem. Biotechnol. 70–72, 17–24.
Smith, J.M., Van Ness, H.C., 1975. Introduction to chemical engineering thermodynamics. In: Clark, B.J., Maisel, J.W. (Eds.), Chemical Engineering Series. McGraw-Hill, New York, pp. 271– 272. Tengborg, C., Stenberg, K., Galbe, M., Zacchi, G., Larrson, S., Palmqvist, E., Hahn-Hagerdal, B., 1998. Comparison of SO2 and H2 SO4 impregnation of softwood prior to steam pretreatment on ethanol production. Appl. Biochem. Biotechnol. 70–72, 3–15. Torget, R., Hatzis, C., Hayward, T.K., Hsu, T., Phillipidis, G.P., 1996. Optimization of reverse flow, two-temperature, dilute acid pretreatment to enhance biomass conversion to ethanol. Appl. Biochem. Biotechnol. 57/58, 85–101. Torget, R.W., Kim, J.S., Lee, Y.Y., 2000. Fundamental aspects of dilute acid hydrolysis/fractionation kinetics of hardwood carbohydrates. 1. cellulose hydrolysis. Ind. Eng. Chem. Res. 39, 2817–2825. van Walsum, G.P., 2001. Severity function describing the hydrolysis of xylan using carbonic acid. Appl. Biochem. Biotechnol. 91–93, 317– 329. van Walsum, G.P., Allen, S.G., Spencer, M.J., Laser, M.S., Antal, M.J., Lynd, L.R., 1996. Conversion of lignocellulosics pretreated with liquid hot water to ethanol. Appl. Biochem. Biotechnol. 54/55, 157–170. Wang, L., Dale, B.E., Yurttas, L., Goldwasser, I., 1998. Cost estimates and sensitivity analyses for the ammonia fiber explosion process. Appl. Biochem. Biotechnol. 70–72, 51–66. Wooley, R., Ruth, M., Glassner, D., Sheehan, J., 1999. Process design and costing of bioethanol technology: a tool for determining the status and direction of research and development. Biotechnol. Prog. 15, 794–803. Yourchisin, D., van Walsum, G.P., Comparison of the microbial inhibition and enzymatic hydrolysis rates of liquid and solid fractions produced from pretreatment of biomass with carbonic acid and liquid hot water. Appl. Biochem. Biotechnol., in press.