Development and characterization of a high-solids deacetylation process
© The Author(s) 2016
Received: 30 December 2015
Accepted: 6 May 2016
Published: 20 May 2016
Dilute-acid pretreatment has proven to be a robust means of converting herbaceous feedstock to fermentable sugars. However, it also releases acetic acid, a known fermentation inhibitor, from acetyl groups present in the biomass. A mild, dilute alkaline extraction stage was implemented prior to acid pretreatment to separate acetic acid from the hydrolysate sugar stream. This step, termed deacetylation, improved the glucose and xylose yields from enzymatic hydrolysis and ethanol yields from fermentation of the sugars relative to the control experiments using dilute-acid pretreatment of native corn stover without deacetylation. While promising, deacetylation as it was historically practiced is conducted at low solids loadings, and at fixed conditions. Thus, many questions have been left unanswered, including the relationship between sodium hydroxide and solids loading, and acetate and xylan solubilization, as well as the impact of temperature and residence time on the process efficacy.
A central composite experiment was designed to evaluate the impact of solids loading, sodium hydroxide loading, reaction time and temperature during deacetylation on the acetate and xylan solubilization of corn stover. Using the ANOVA test, it became apparent that neither of the responses was significantly impacted by the solids loading, while the reaction time was a minor factor—the responses were largely driven by reaction temperature and the sodium hydroxide loading. Based on the results, we successfully demonstrated the ability to transition the low-solids (10 % w/w) deacetylation process to a higher-solids process (30 % w/w) with minimal impact on the ability to extract acetate from biomass. Conditions were selected to minimize xylose loss during deacetylation, while also removing 70 % of acetyl groups. The impact of selected conditions on the enzymatic hydrolysis and fermentation was further investigated.
Evaluation of the whole-process impact demonstrated that despite the upfront reduction in carbohydrate loss during deacetylation, the overall process sugar yields were depressed by the high-solids, low alkali process relative to the historical control. Consequently, ethanol titers were reduced, though strong fermentation performance was still observed, indicating that 70 % acetate removal is sufficient to depress acetic acid concentrations to a level that does not substantially inhibit ethanol fermentation by rZymomo nas.
Renewable liquid transportation fuels present a challenge distinct from other energy sources, as the tradable commodity and fuel type has yet to be defined as it is in the case of electrical power. Thus, many potential fuel products are being considered, ranging from alcohol fuels, such as ethanol and butanol, to low-oxygen hydrocarbon products that may be directly substituted for currently available fuels. Further complicating the issue, each of these potential solutions can be either blended into the current fuel supply or used as a standalone replacement, although the latter scenario would require potentially significant infrastructure changes in many cases. While this lack of a standardized product specification has presented challenges to the industry, many biological conversion processes utilize sugars to synthesize their respective target products, and require solutions with high concentrations of monomeric sugars and low concentrations of inhibitory compounds, including furfural, 5-hydroxymethylfurfural (HMF), and acetic acid.
Dilute-acid pretreatment has proven to be a robust means of converting herbaceous lignocellulosic feedstock to fermentable 5- and 6-carbon sugars . However, it is also prone to producing potent fermentation inhibitors (i.e., furfural), and also releases acetic acid from the acetyl groups present in biomass, commingled with the desired sugars. Furfural concentrations can be reduced by implementing an effective flash system at the conclusion of the pretreatment process, which has been demonstrated to remove approximately 50 % of furfural from the hydrolysate and enable furfural concentrations less than 2 g/L (in a 20 % total solids whole slurry) . Similarly, acetate, in the form of acetic acid, can be separated from the hydrolysate sugar stream through a mild, dilute alkaline extraction stage implemented prior to pretreatment . This step, termed deacetylation, has been demonstrated to solubilize approximately 80 % of acetyl groups present in native corn stover, which can then be separated from the biomass prior to pretreatment though a wash process. Deacetylating corn stover prior to a typical dilute sulfuric acid pretreatment improved both glucose and xylose yields during enzymatic hydrolysis and ethanol yields relative to the control experiments using dilute-acid pretreatment of native corn stover .
Other processes have been explored that use neutral, ionic liquids or alkaline chemistry, including hydrothermal pretreatment , ionic liquid pretreatment using imidazole , peroxide pretreatment, oxidative lime pretreatment, and pulping-like sodium hydroxide pretreatments [7–15]. However, these processes are typically intended to function as a stand-alone pretreatment, and the severity of the process is increased to enable effective enzymatic hydrolysis with no further processing. The increased severity, while effective in achieving high cellulose conversion yields through lignin extraction, also results in the solubilization and degradation of a significant fraction of hemicellulosic sugars; in many cases, greater than 20 % is found in the alkaline liquor stream [8, 11]. Due to the low concentrations of sugars and comparatively high concentrations of undesirable components, including salts and acetic acid, this stream is typically treated as waste. Coupling deacetylation, which solubilizes less than 10 % of the hemicellulose, with dilute-acid pretreatment has proved to be a cost-effective, process-relevant means to achieve high sugar yields, while also obtaining one of the primary benefits of alkaline pretreatment (i.e., low inhibitor concentrations).
While promising, previous work on the deacetylation process was conducted at low (~10 %) total solids loadings, and at fixed conditions (0.4 % w/w NaOH, 80 °C, 120 min) . Thus, many questions have been left unanswered, including the relationship between sodium hydroxide loading and acetate and xylan solubilization, as well as the impact of temperature and residence time on the process efficacy. Further, process economics favor reductions in chemical and water utilization. In particular, increasing the solids loading during deacetylation would offer significant cost savings by reducing reactor volume requirements, decreasing energy requirements for heating during deacetylation, and reducing wastewater generation. Other significant cost drivers include the process residence time and temperature.
In this work, we characterized low-severity alkaline extraction of corn stover, termed deacetylation, using response surface methodology. Experimental conditions range from sodium hydroxide loadings of 0.1–0.5 g/g of dry biomass, temperatures between 60 and 100 °C and extraction times of 45–105 min. The characterization experiments were evaluated to select a promising condition reflecting more favorable economics (e.g., decreased xylan loss, increased solids loading, decreased chemical usage) while maintaining >60 % acetyl group removal. This condition was replicated to generate more deacetylated biomass used in further experiments to evaluate the impact of the new conditions on enzymatic hydrolysis and fermentation at low and moderate pretreatment severities compared to those pretreated at control conditions.
Results and discussion
Parameters and test levels selected for development of the central composite design
Center point (0)
Solids loading (wt%)
NaOH loading (g/g acetate)
Xylan balance closure was consistently between 90 and 100 %, with only a few instances of closure outside this range, indicating that we can account for most of the xylan-derived species solubilized in this low-severity alkaline process. However, the remainder of the xylan that is unaccounted for is likely degraded due to the peeling reactions under alkaline conditions and referred to as xylan loss in this paper. The percent of xylan loss is calculated by dividing the part of unaccountable xylan after deacetylation process by the total amount xylan in the initial biomass.
In addition, the solubilization of acetate was calculated based on the acetate concentration in the liquor phase divided by the total amount of acetate groups in the initial biomass.
These data suggest that solids concentration is not a significant factor within the range of conditions tested, and also provides more confidence in the scalability and ultimate economic viability of the process. The CAPEX and OPEX of the unit operation will be potentially reduced through decreased process volumes at the higher solids operation. The effluent stream from the deacetylation process has the potential for producing valuable side products because high titers are achieved from high-solids operation.
Relative to previous low-solids deacetylation work (control), less xylose was measured in the post-extraction liquor, indicating a significant reduction in xylan loss. The amount of acetate removed from the corn stover was also decreased, with 70 % recovered in the post-extraction liquor in the high-solids process, and 80 % solubilized in the low-solids control process.
Dilute acid pretreatment
As shown in Figs. 5 and 6, monomeric xylose yield was improved by roughly 10 yield points in the three high-solids deacetylation samples in both sets of pretreatment experiments. This is likely due to the decreased xylan loss in deacetylation, and thus the preservation of easily-extractible xylose in the feedstock that was solubilized during acid hydrolysis. Interestingly, the soluble xylo-oligosaccharides yield from xylan was comparable across all samples at a given pretreatment condition, despite the difference in total and monomeric xylose yield. Conversion to furfural was minimal (1–2 %), as expected, given the relatively mild pretreatment conditions and modest acid loadings (0.5 % w/w during pre-impregnation). Xylan component balance closure ranged between 97 and 100 % for all samples.
Concentrations of acetic acid were modestly higher in pretreated hydrolysates generated in the high-solids process than those from the control. In low-severity pretreatment experiments, they were 2–3 g/L in the high-solids process and <1 g/L in the low-solids process, respectively. At moderate severity, they were 3–4 g/L in the high-solids process and <1 g/L in the low-solids process. The higher concentrations of acetic acid resulting from the moderate severity pretreatment were reflective of the more severe pretreatment in which more acetic acid will be released from residue acetyl groups in the biomass.
The solids compositional analysis on the acid-pretreated corn stover substrates reveals that the low-solids deacetylation process has lower lignin content (14 %) and higher glucan and xylan content (44 and 29 %, respectively), while the high-solids deacetylation has more lignin (18 %) and less glucan and xylan (39 and 26 %, respectively). This is mainly caused by lower sodium hydroxide loading (0.24 w/w%) in the high-solids deacetylation, resulting in less severe delignification compared to the low-solids deacetylation process.
The fermentability of each hydrolysate was tested using the glucose-xylose fermenting microorganism Zymomonas mobilis 8b. Fermentation was run for 72 h, though glucose was consumed (<2 g/L) in all samples within the first 24 h. The utilization of xylose was slightly higher in the low-solids control process samples, though all were within the same range (83–88 %). Ethanol titers, though, were significantly higher (70 vs 60 g/L) in the control-process-derived samples, largely due to the increased glucose yield from enzymatic hydrolysis making a greater quantity of fermentable sugars available.
We successfully demonstrated the ability to transition the low-solids (10 % w/w) deacetylation process to a higher solids process with minimal impact on the ability to extract acetate from biomass. Conditions were selected to minimize xylan loss during deacetylation, while also removing 70 % of the acetyl groups. Evaluation of the impacts on enzymatic hydrolysis and fermentation demonstrated that, despite the upfront reduction in carbohydrate loss in deacetylation, the overall process sugar yields were depressed by the high-solids, low-alkali process relative to the low solids control. Consequently, ethanol titers were reduced, though strong fermentation performance was still observed, indicating that 70 % acetate removal is sufficient to depress acetic acid concentrations to a level that does not substantially inhibit ethanol fermentation by Z. mobilis.
The corn stover used in this study was harvested in Emmetsburg, IA, USA in October 2010 and shipped to Idaho National Laboratory (INL, Idaho Falls, ID, USA). There, it was processed using the pilot-scale feedstock-handling unit operated by INL as described by Yancey et al.  and hammer milled to pass through a 1.9 cm (0.75 in.) round rejection screen. The raw feedstock contains approximately 14.6 % lignin, 34.0 % glucan, 22.8 % xylan, 1.9 % galactan, 3.3 % arabinan, 2.3 % acetate, and 8.4 % ash.
Deacetylation was performed in a jacketed 90-L paddle mixing reactor. Dry corn stover (3–5 kg) was first loaded into the vessel and the reactor sealed. With the agitator mixing at moderate speed (approximately 100 rpm), water and sodium hydroxide were transferred into the reactor to achieve the desired sodium hydroxide and solids loading. The water and sodium hydroxide were previously mixed in a 60-L stainless steel pressurized dispensing tank. After adding the liquids, a low pressure steam jacket (35 psig supply) was used to heat the contents of the reactor to the target temperature. After a 15-min heat-up period, the steam flow was controlled by throttling the globe and ball valves upstream of the tank and adjusting periodically to maintain the headspace temperature at the target set point. The corn stover was mixed and maintained at temperature for the target amount of time before being discharged from the vessel.
Samples were taken of the deacetylated biomass and the resultant liquor for subsequent compositional analysis and characterization.
Washing and acid impregnation
If the deacetylated biomass was to be carried forward into pretreatment and enzymatic hydrolysis, the solids from deacetylation were placed in a Hastelloy C-276 20-mesh wire basket and exhaustively washed with water to achieve a wash-water pH of 8.0. A sample of the washed solids was taken for moisture analysis.
The deacetylated solids were then loaded back into the 90-L paddle reactor, and the required amount of water to achieve a 10 % total solids loading and 93 % sulfuric acid to achieve the target acid concentration were added to the pressurized transfer tank. After the reactor was sealed, the agitator was started and set to approximately 100 rpm, and the contents of the pressurized transfer tank were added to the reactor. The contents were allowed to mix at room temperature for 2 h and then discharged.
The solids were dewatered to 50 % total solids using a 25.4-cm (10-in.) hydraulic press operated at pressures of up to 440 psig (3 MPa). Samples of the impregnated and dewatered solids were taken for solids compositional analysis and moisture analysis. The prepared feedstock was stored at 4 °C for no longer than 1 week prior to use in pretreatment experiments.
Dilute-acid pretreatment experiments were carried out in a 4-L steam explosion reactor, the construction and operation of which has been previously described . Prepared deacetylated and acid-impregnated corn stover feedstock were loaded into the reactor (500 g) and heated to the target temperature by direct steam injection. The internal temperature and pressure of the reactor were monitored and recorded. After the desired time elapsed, the contents were discharged from the vessel through a fast-actuating ball valve and collected in a 55-gal drum liner (CDF Corporation, Plymouth, MA, USA).
Pretreatment was performed at two different conditions—150 °C for 20 min and 165 °C for 10 min. All experiments were performed in triplicate. Samples from each experiment were taken for complete compositional analysis. The remaining portions of the triplicate slurries were then pooled to create eight batches prior to enzymatic hydrolysis (three high-solids deacetylation and one low-solids deacetylation, for each pretreatment condition).
Each of these eight hydrolysates was neutralized with 28 % ammonium hydroxide to a pH of 5.2–5.4 and tested in duplicate in enzymatic hydrolysis and fermentation. Enzymatic hydrolysis was performed in 1-L bottles using a total mass of 450 g at 20 % total solids using CTec2 at a loading of 40 mg of protein per gram of cellulose. The bottles were placed in a shaker box incubator (150 rpm) and held at 48 °C for 120 h, after which samples were taken for analysis and the remainder of the hydrolysate carried forth to fermentation.
Fermentations were performed in batch mode in Biostat-Q plus (Sartorius, Goettingen, Germany) fermenters with a 400 mL working volume. Fermentation conditions were: temperature 33 °C, pH 5.8 controlled using 4 N KOH, and 300 rpm. The microorganism used was the NREL strain, Zymomonas mobilis 8b. The fermentations were completed in 72 h.
The media used in the fermentation was rich media (RM) with the composition of 10 g/L yeast extract and 2 g/L KH2PO4 supplemented with desired sugar level. RM was prepared as 10× stock solution and sugars, glucose, and xylose were prepared as 50× stock solution; all were filter sterilized. To start the inoculum, first one frozen vial of strain 8b was thawed and revived in 9 mL RMGX (8:2 %) in a 15 mL Falcon tube and incubated for 8 h at 33 °C. The revived culture was used to start the main seed fermenter. The seed fermenter was a Biostat-Q plus fermenter with 400 mL media RMGX (15:2 %) plus 1 g/L sorbitol under similar conditions to those mentioned above. When the optical density (OD @600 nm) of about 10 was reached, usually in 18 h, the remaining glucose was usually between 40–50 g/L. Then the seed culture was used to inoculate the main fermenters at a 10 % (v/v) ratio to achieve an initial OD of 1 (@600 nm).
Sugar concentrations were measured using an Agilent 1100 series high performance liquid chromatograph (HPLC) (Santa Clara, CA, USA) with a Shodex SP0810 carbohydrate column (Shawa Denko K.K., Kawasaki, Japan) and a de-ashing guard cartridge (BioRad Laboratories, Hercules, CA, USA). The HPLC was operated with a column temperature of 85 °C and 0.6 mL/min flow rate of the mobile phase (ultra-pure water). Furfural, HMF, and acetic acid concentrations were measured by HPLC using a Phenomenex Rezex RFQ Fast Fruit H + organic acid column and Cation H + guard cartridge (BioRad Laboratories) at 55 °C. The mobile phase was dilute sulfuric acid (0.01 N) at a flow rate of 0.6 mL/min. A refractive index detector was used for compound detection for both columns. Mixed component standards were periodically run with the HPLC samples to verify calibration accuracy. The density of liquid samples was measured using an Anton-Paar model DMA-500 density meter (Anton Paar USA, Inc., Ashland, VA, USA).
The composition of pretreated solids was determined using a two-stage acid digestion procedure . Concentrations of total soluble sugars in pretreated liquor samples were determined using a mild dilute-acid hydrolysis procedure and HPLC analysis . The oligomeric sugar concentration was the difference between total and monomeric sugar concentration. Slurry TS concentrations were determined by drying samples at 45 °C in a vacuum oven (0.6 bar) until repeated weight measurements were constant. Slurry insoluble solid concentrations were determined by a six-step washing and centrifugation procedure . Triplicate measurements were performed on each sample.
The enzyme was desalted prior to measuring the protein content. Desalting was performed on a HiPrep 26/10 desalting column (GE Healthcare, Uppsala, Sweden) with a Sephadex G-25 column matrix using a 2-mL sample. The mobile phase was a 50 mM Tris, 150 mM NaCl pH 5 buffer at a flow rate of 10 mL/min. Protein concentration was measured using the Pierce BCA (BCA Protein Assay Kit, Pierce, IL, USA) assay following the manufacturer’s protocols with bovine serum albumin as the protein standard. The measurement was performed in triplicate.
Response surface methodology
A four-factor central composite design was used to characterize the reaction space for deacetylation as a function of temperature, reaction time, sodium hydroxide loading on a gram per gram of acetate basis, and total solids loading. The design used in this experiment was a five-level, four-factor factorial with the center point replicated six times and six axial points with the axial points specified to maintain full rotatability. The calculation of the axial points, specification of the design, and randomization of the experimental design was done using Design-Expert 8.0 (Stat-Ease, Minneapolis, MN, USA), but was consistent with the methodology described by Meyers, Montgomery and Cook . The selection of experimental conditions was based on previously published results [3, 11]. The range of conditions tested is displayed in Table 1.
Regression analysis was completed using Design-Expert 8.0. The responses modeled included the overall solubilization of acetic acid and xylose during the deacetylation stage, with the objective of maximizing the former while minimizing the latter. The software was used to determine which factors from a linear, quadratic, or two-factor interaction model were significant beyond a p value of 0.05 using one-way ANOVA analysis. Factors with the largest p values were progressively removed from each respective model until only significant (p < 0.05) terms remained.
JS executed the deacetylation experiments, analyzed data, and prepared the manuscript along with XC. HS designed and conducted the enzymatic hydrolysis and fermentation experiments. MPT, XC and JS conceived the work and developed the experimental design. All authors read and revised. All authors read and approved the final manuscript.
Pilot plant operation support was provided by Casey Gunther and Eric Nelson. Analytical sample analysis was performed by Deborah Hyman, Darren Peterson, Michelle Reed, David Templeton, and Jeffery Wolfe.
This work was supported by the US Department of Energy under Contract No. DE-AC36-08GO28308 with the National Renewable Energy Laboratory. Funding provided by U.S. DOE Office of Energy Efficiency and Bioenergy Technologies Program.
The US Government retains and the publisher, by accepting the article for publication, acknowledges that the US Government retains a nonexclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this work, or allow others to do so, for US Government purposes.
The authors declare that they have no competing interests.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
- Weiss ND, Nagle NJ, Tucker MP, Elander RT (2009) High xylose yields from dilute acid pretreatment of corn stover under process-relevant conditions. Appl Biochem Biotechnol 155:418–428View ArticleGoogle Scholar
- Shekiro Iii J, Kuhn EM, Nagle NJ, Tucker MP, Elander RT, Schell DJ (2014) Characterization of pilot-scale dilute acid pretreatment performance using deacetylated corn stover. Biotechnol Biofuels 7:23View ArticleGoogle Scholar
- Chen X, Shekiro J, Elander R, Tucker MP (2012) Improved xylan hydrolysis of corn stover by deacetylation with high solids dilute acid pretreatment. Ind Eng Chem Res 51:6Google Scholar
- Chen X, Shekiro J, Franden M, Wang W, Zhang M, Kuhn E, Johnson D, Tucker M (2012) The impacts of deacetylation prior to dilute acid pretreatment on the bioethanol process. Biotechnol Biofuels 5:2–14View ArticleGoogle Scholar
- Morais ARC, Mata AC, Bogel-Lukasik R (2014) Integrated conversion of agroindustrial residue with high pressure CO2 within the biorefinery concept. Green Chem 16:4312–4322View ArticleGoogle Scholar
- Morais ARC, Pinto JV, Nunes D, Roseiro LB, Oliveira MC, Fortunato E, Bogel-Łukasik R (2016) Imidazole: prospect solvent for lignocellulosic biomass fractionation and delignification. ACS Sustain Chem Eng 4:1643–1652View ArticleGoogle Scholar
- Alvira P, Tomas-Pejo E, Ballesteros M, Negro MJ (2010) Pretreatment technologies for an efficient bioethanol production process based on enzymatic hydrolysis: a review. Bioresour Technol 101:4851–4861View ArticleGoogle Scholar
- Banerjee G, Car S, Liu T, Williams DL, Meza SL, Walton JD, Hodge DB (2012) Scale-up and integration of alkaline hydrogen peroxide pretreatment, enzymatic hydrolysis, and ethanolic fermentation. Biotechnol Bioeng 109:922–931View ArticleGoogle Scholar
- Chang VS, Nagwani M, Holtzapple MT (1998) Lime pretreatment of crop residues bagasse and wheat straw. Appl Biochem Biotechnol 74:135–159View ArticleGoogle Scholar
- Elander RT, Dale BE, Holtzapple M, Ladisch MR, Lee YY, Mitchinson C, Saddler JN, Wyman CE (2009) Summary of findings from the biomass refining consortium for applied fundamentals and innovation (CAFI): corn stover pretreatment. Cellulose 16:649–659View ArticleGoogle Scholar
- Karp EM, Donohoe BS, O’Brien MH, Ciesielski PN, Mittal A, Biddy MJ, Beckham GT (2014) Alkaline pretreatment of corn stover: bench-scale fractionation and stream characterization. ACS Sustain Chem Eng 2:1481–1491View ArticleGoogle Scholar
- Kumar P, Barrett DM, Delwiche MJ, Stroeve P (2009) Methods for pretreatment of lignocellulosic biomass for efficient hydrolysis and biofuel production. Ind Eng Chem Res 48:3713–3729View ArticleGoogle Scholar
- Liu T, Williams DL, Pattathil S, Li M, Hahn MG, Hodge DB (2014) Coupling alkaline pre-extraction with alkaline-oxidative post-treatment of corn stover to enhance enzymatic hydrolysis and fermentability. Biotechnol Biofuels 7:48View ArticleGoogle Scholar
- Tsaih ML, Chen RH (2003) The effect of reaction time and temperature during heterogenous alkali deacetylation on degree of deacetylation and molecular weight of resulting chitosan. J Appl Polym Sci 88:2917–2923View ArticleGoogle Scholar
- Xu J, Cheng JJ, Sharma-Shivappa RR, Burns JC (2010) Sodium hydroxide pretreatment of switchgrass for ethanol production. Energy Fuels 24:2113–2119View ArticleGoogle Scholar
- Yancey N, Wright CT, Westover TL (2013) Optimizing hammer mill performance through screen selection and hammer design. Biofuels 4:85–94View ArticleGoogle Scholar
- Raymond H, Myers DCM, Christine M (2009) Anderson-cook: response surface methodology: process and product optimization using designed experiments, 3rd edn. John Wiley & Sons, HobokenGoogle Scholar