Elsevier

Bioresource Technology

Volume 81, Issue 1, January 2002, Pages 33-44
Bioresource Technology

A comparison of liquid hot water and steam pretreatments of sugar cane bagasse for bioconversion to ethanol

https://doi.org/10.1016/S0960-8524(01)00103-1Get rights and content

Abstract

Sugar cane bagasse was pretreated with either liquid hot water (LHW) or steam using the same 25 l reactor. Solids concentration ranged from 1% to 8% for LHW pretreatment and was ⩾50% for steam pretreatment. Reaction temperature and time ranged from 170 to 230 °C and 1 to 46 min, respectively. Key performance metrics included fiber reactivity, xylan recovery, and the extent to which pretreatment hydrolyzate inhibited glucose fermentation. In four cases, LHW pretreatment achieved ⩾80% conversion by simultaneous saccharification and fermentation (SSF), ⩾80% xylan recovery, and no hydrolyzate inhibition of glucose fermentation yield. Combined effectiveness was not as good for steam pretreatment due to low xylan recovery. SSF conversion increased and xylan recovery decreased as xylan dissolution increased for both modes. SSF conversion, xylan dissolution, hydrolyzate furfural concentration, and hydrolyzate inhibition increased, while xylan recovery and hydrolyzate pH decreased, as a function of increasing LHW pretreatment solids concentration (1–8%). These results are consistent with the notion that autohydrolysis plays an important, if not exclusive, role in batch hydrothermal pretreatment. Achieving concurrently high (greater than 90%) SSF conversion and xylan recovery will likely require a modified reactor configuration (e.g. continuous percolation or base addition) that better preserves dissolved xylan.

Introduction

Cellulosic biomass represents the only foreseeable sustainable source of organic fuels, chemicals, and materials (Lynd et al., 1999). A primary technological challenge in biologically processing cellulosic biomass into fuels and chemicals is overcoming the recalcitrance of cellulose to hydrolysis. Cellulose hydrolysis processes are typically categorized into those that use strong mineral acids and those that use cellulase enzymes. Although processes using acids are more technologically mature, enzymatic processes have comparable projected costs and are expected to enjoy an increasing cost advantage as the technology improves (Lynd et al., 1999, NREL, 1999). Due to its resistance to enzymatic attack, however, naturally occurring cellulosic biomass must be pretreated before it can be enzymatically hydrolyzed. Pretreatment is one of the most expensive and least technologically mature unit operations in lignocellulosic conversion processes using enzymatic hydrolysis (Lynd, 1996).

To be effective, a pretreatment process must produce reactive fiber, preserve the utility of the pentosan (hemicellulose) fraction, and limit the extent to which the pretreated material inhibits growth of the fermenting microorganism(s). To be economical, the process should minimize energy demands and limit costs associated with feedstock size reduction, materials of construction, and treatment of process residues. These performance metrics are discussed in van Walsum et al. (1996), and Lynd (1996).

In naturally occurring cellulosic substrates, carbohydrate-rich microfibrils are surrounded by a lignin seal forming a complex structural matrix that is resistant to enzymatic attack (Holtzapple, 1993a, Holtzapple, 1993b, Holtzapple, 1993c). Principal substrate factors that have been correlated with pretreatment effectiveness include cellulose pore volume (Grethlein, 1985, Weimer and Weston, 1985), hemicellulose and lignin removal (Weimer and Weston, 1985), and cellulose crystallinity (Converse, 1993). Process conditions – temperature, reaction time, pH, and biomass concentration – affect these substrate factors, and thus influence pretreatment performance.

Pretreatment processes can be loosely grouped into three categories: physical, chemical, and hydrothermal. Physical pretreatments, which typically demand large amounts of energy and are expensive, employ purely mechanical means to reduce feedstock particle size, thus increasing surface area. A variety of chemicals – acids, alkalis, organic solvents, oxidizing agents, supercritical fluids, and ligninase enzymes – have been considered as pretreatment agents (Weil et al., 1994). Dilute acid pretreatment, ammonia fiber explosion (AFEX), and lime pretreatment have emerged as particularly effective chemical methods (Himmel et al., 1997, Dale et al., 1996, Kaar and Holtzapple, 2000).

Hydrothermal pretreatment refers to the use of water – as liquid or vapor or both – and heat to pretreat biomass. Relative to dilute acid pretreatment, hydrothermal pretreatment processes offer several potential advantages: there is no requirement for purchased acid, for special non-corrosive reactor materials or for preliminary feedstock size reduction (van Walsum et al., 1996). Furthermore, hydrothermal processes produce much lower quantities of hydrolyzate neutralization residues. Steam explosion, where biomass exposed to pressurized steam is rapidly depressurized or “exploded”, can produce reactive fiber (90% conversion, e.g. Heitz et al., 1991) at high solids concentrations (⩾50%), but is often found to generate inhibitory hydrolyzates (Forsberg et al., 1986, Mes-Hartree and Saddler, 1983) and results in low pentosan recoveries (less than 80%; Heitz et al., 1991). Liquid hot water (LHW) pretreatment, where biomass is exposed to pressurized hot water (solids concentration  20%), appears to have the potential to generate reactive fiber (⩾90% conversion; van Walsum et al., 1996), recover most of the pentosans (⩾90%; Mok and Antal, 1992), and produce hydrolyzate that results in little or no inhibition of glucose fermentation (van Walsum et al., 1996). van Walsum et al. and Allen et al. (1997) attempted to examine all three performance metrics but did not present results for pentosan recovery due to difficulties in obtaining reproducible data at 10% solids. Poor reproducibility resulted from high transient temperatures (up to 260 °C) that were necessary to achieve 220 °C at these solids concentrations (the upper design limit for the reactor). Allen et al. (1996) did not have this problem when operating the same reactor at lower solids concentrations. Allen et al. (2001) have examined the three metrics for LHW and steam pretreatment of corn fiber. At a solids concentration of 5%, they obtained 86% SSF conversion, 82% xylan recovery, and no hydrolyzate inhibition of fermentation yield (fermentation rate, however, was inhibited).

Using the same reactor for both processes, this study compared LHW pretreatment where biomass was immersed in liquid water, and steam pretreatment where biomass was not immersed, by examining fiber reactivity, pentosan recovery and hydrolyzate inhibition over a range of LHW pretreatment solids concentrations.

Section snippets

Feed materials

Sugar cane bagasse generously supplied by The Hawaii Commercial & Sugar Company (Puunene, HI) was used in all experiments. On average, the bagasse contained 44% glucose, 26% xylose, 2% arabinose, and 23% acid insoluble lignin (dry weight basis). The bagasse was stored at ⩽15 °C and screened (+14 mesh retained) prior to use. Note: In previous LHW pretreatment studies using a 250 ml reactor (van Walsum et al., 1996, Allen et al., 1997), the authors also screened bagasse to a +14 mesh. We did the

Overall mass recovery and dissolution

Overall mass recovery ranged from 74% to 107% and 84% to 101% for LHW and steam pretreatment, respectively, with both modes averaging 90% overall mass recovery (Table 1, Table 2). Overall dissolution ranged from 17% to 50% and 4% to 35% for LHW and steam pretreatment, respectively (Table 1, Table 2).

SSF conversion

Both steam pretreatment and LHW pretreatment achieved ⩾90% SSF conversion based on both residual cellulose and theoretical ethanol production. SSF conversion increased as a function of increasing

Discussion

This study compared LHW and steam pretreatment of sugar cane bagasse by measuring fiber reactivity, pentosan recovery, and hydrolyzate inhibition over a range of solids concentrations. Both methods were capable of preserving glucan (approximately 97% recovery) and producing highly reactive fiber (⩾90% SSF conversion; Fig. 1). These results are consistent with previous studies (glucan recovery: Bobleter, 1994, Mok and Antal, 1992, Heitz et al., 1991; fiber reactivity: van Walsum et al., 1996,

Acknowledgements

This project was funded by the Consortium for Plant Biotechnology Research, Inc. (CPBR), Michigan Biotechnology Institute (MBI), and BC International, Inc. (BCI). The National Renewable Energy Laboratory (NREL) and the Coral Industries Endowment (University of Hawaii) provided additional support. This support does not constitute an endorsement by CPBR, MBI, BCI, NREL, or the Coral Industries Endowment of the views expressed in this article. We thank Sam Friedlander, Mary Bigelow, and Viktoria

References (37)

  • O. Bobleter et al.

    Steam explosion–hydrothermolysis–organosolv. A comparison

  • G. Bonn et al.

    Hydrothermal and organosolv pretreatments of poplar wood and wheat straw for saccharification by a Trichoderma viride cellulase

    Wood Sci. Technol.

    (1987)
  • A.O. Converse

    Substrate factors limiting enzymatic hydrolysis

  • A.O. Converse et al.

    Kinetics of thermochemical pretreatment of lignocellulosic materials

    Appl. Biochem. Biotechnol.

    (1989)
  • C.W. Forsberg et al.

    Release of fermentable carbohydrate from peat by steam explosion and its use in the microbial production of solvents

    Biotechnol. Bioeng.

    (1986)
  • H.E. Grethlein

    The effect of pore size distribution on the rate of enzymatic hydrolysis of cellulosic substrates

    Bio/Technol.

    (1985)
  • K. Grohmann et al.

    Optimization of dilute acid pretreatment of biomass

    Biotechnol. Bioeng. Symp.

    (1986)
  • Himmel, M., Adney, W., Baker, J., Elander, R., McMillan, J., Nieves, R., Sheehan, J., Thomas, S., Vinzant, T., Zhang,...
  • Cited by (525)

    View all citing articles on Scopus
    View full text