Off-campus South Dakota State University users: To download campus access theses, please use the following link to log into our proxy server with your South Dakota State University ID and password.

Non-South Dakota State University users: Please talk to your librarian about requesting this thesis through interlibrary loan.

Document Type

Dissertation - University Access Only

Award Date

2012

Degree Name

Doctor of Philosophy (PhD)

Department

Agricultural and Biosystems Engineering

First Advisor

James L. Julson

Abstract

As global energy demands grow and as the environmental and economic issues of fossil fuel use arise, lignocellulosic biomass is starting to attract increased attention as a potential source of energy and chemicals. Being an abundant, accessible and cost effective feedstock for a wide variety of products, ranging from transportation fuel to pharmaceuticals, lignocellulose shows great promise for the future. However, in order to utilize its potential, an efficient pretreatment method has to be applied.
This dissertation investigates organosolv and hydrothermal treatment, as well as the combination of the two, applied to several lignocellulosic feedstocks. The main purpose of the organosolv treatment (clean fractionation) was to extract lignin, hemicellulose and digestible cellulose, while the main purpose of the hydrothermal treatment was solely to improve cellulose's enzymatic digestibility. Response surface methodology was used for optimizations of all the processes.
In the evaluation and optimization of the hydrothermal pretreatment, when applied alone, three lingocellulosic materials were examined: prairie cordgrass, switchgrass and com stover. The primary response variables used to optimize process conditions (temperature and time) were glucose yields obtained by enzymatic hydrolysis. Optimal conditions for prairie cordgrass switchgrass and com stover were found to be at 210.00°C and 10.00 min, 200.13''C and 12.81 min, and 197.20°C and 16.50 min, respectively. Samples pretreated at optimal conditions produced hydrolysis glucose yield results of 88.20 ± 2.16 % for prairie cordgrass, 87.67 ± 4.14 % for switchgrass and 90.18 ± 1.77 % for com stover.
Simultaneous saccharification and fermentation (SSF) applied to the pretreated material resulted in relative ethanol yield of 94.13 ± 3.75 % for prairie cordgrass, 83.89 ± 4.04 % for switchgrass and 86.41 ± 0.82 % for com stover. These results show that the yeast species Saccharomyces cerevisiae are able to ferment hydrothermally pretreated cellulosic material efficiently, without signs of major stress or inhibition.
Catalyzed modified clean fractionation was applied to switchgrass as a feedstock. The original procedure developed by National Renewable Energy Laboratory (Black et al., 1998) was modified by replacing one of the solvent mixture components - MIBK - with less toxic and less expensive ethyl acetate. In order to improve lignin recovery, the process was catalyzed with sulfuric acid. The processing conditions (temperature, catalyst concentration and solvent composition) were optimized for maximum lignin recovery as primary response variable and enzymatic hydrolysis glucose yields as secondary response variable. Optimal conditions were validated and resulted in 59.03 ± 7.01 % lignin recovery, 84.85 ± 1.34 % hydrolysis glucose yield, and 44.11 ± 3.44 % xylose yield. The cellulose fraction was fermented (by SSF) and resulted in a high ethanol yield (89.60 ± 2.10 %), revealing no inhibition to the yeast's performance.
An integrated process of catalyzed modified clean fractionation and hydrothermal post-treatment was used to fractionate and pretreat prairie cordgrass and com stover, since application of clean fractionation alone resulted in low cellulose's enzymatic digestibility. The modified clean fractionation process (using ethyl acetate-ethanol-water mixture) was employed to extract lignin and hemicellulose fractions. Sulfuric acid was used as a catalyst to improve lignin removal. Optimization was performed for fractionation processing conditions using lignin recovery and glucose yields as response variables. Optimal conditions resulted in 51.33 % lignin recovery, 53.82 % hydrolysis glucose yield and 33.73 % xylose yield for prairie cordgrass and 66.77 ± 4.20 % lignin recovery, 73.58 ± 4.26 % hydrolysis glucose yield and 22.06 ± 2.70 % xylose yield for corn stover. To improve cellulose digestibility, the solid fraction was subjected to hydrothermal post-treatment. The hydrothermal post-treatment optimal processing conditions (198.28 °C and 16min for prairie cordgrass and 210.00 °C and 10.00 min for com stover), resulted in 78.93 % hydrolysis glucose yield for prairie eordgrass and 89.88 ± 4.70 % hydrolysis glucose yield for com stover. Simultaneous saeeharifieation and fermentation resulted in high ethanol yield (95.15 ± 4.94 % and 94.76 ± 1.07 % for prairie cordgrass and com stover, respectively), proving that the pretreated materials show high potential for ethanol production.
The possibility of eliminating the catalyst from the modified clean fractionation was also evaluated. For this purpose, a sequential process including uncatalyzed clean fractionation and hydrothermal post-treatment was developed to fractionate and pretreat prairie cord grass prior to ethanol fermentation. Both steps of the process were optimized using Response Surface Methodology. The main purpose of hydrothermal post-treatment application was to improve digestibility of delignified cellulose, and thus reduce the harshness of the clean fractionation process. Clean fractionation process efficiency was determined by lignin recovery. Optimization was performed for solvent composition as well as time and temperature for clean fractionation and temperature and time only for hydrothermal post-treatment. Hydrothermal post-treatment was optimized using glucose yields as response variables. The recovery of organic solvent-soluble lignin was found to be 19.7%at optimal conditions (125°C, 37 min, with the solvent composition of ethyl acetate:ethanol:water = 32.5:22.5:45). This result proves that the catalyst caimot be removed from the fractionation process without compromising its efficiency. Cellulose enzymatic digestibility, even though lower after the uncatalyzed clean fractionation step (38.0 %), reached high values after the hydrothermal post treatment (predicted to be 84.0%at optimal conditions, which were 191°C and 10 min). Dueto low lignin recovery, this process was concluded to be inefficient, since high glucose yields can be achieved by using hydrothermal pretreatment alone.
Lignins extracted from prairie cordgrass, switchgrass and com stover (using catalyzed organosolv pretreatment) were analyzed and characterized by several methods. These methods included analysis of purity (by determination of Klason lignin, carbohydrates and ash contents), solubility test (with several organic solvents) phenolic groups' analysis (ultraviolet ionization difference spectra method and nitrobenzene oxidation), and general functional groups analysis (^H NMR). The analyses showed that all the lignins were relatively pure (contained over 50 % Klason lignin, less than 10 % carbohydrates contamination and less than 3 % ash). Switchgrass-derived lignin was observed to be the purest. All the lignins were found to contain a high number of phenolic groups, while switchgrass-derived lignin was the most phenolic according to the ionization difference spectra method. Nitrobenzene oxidation revealed that all the lignins contain high amounts of available guaiacyl units, which were oxidized to vanillin. Com stover-originated lignin produced the maximum vanillin yield (59.63 % w/w). Proton NMR analysis confirmed the presence of guaiacyl units in all lignins. Dioxane and methanol were observed to be the most efficient solvents for examined lignin types.
Mass balances were performed for the all the examined processes (except for uncatalyzed clean fractionation integrated with hydrothermal post-treatment, which was concluded inefficient). All biomass types treated by catalyzed modified clean fractionation presented similar trends. The analysis revealed that the catalyzed clean fractionation process resulted in low cross-contamination of the phases, generating a relatively clean solid fraction and removing most of the lignin to the organic fraction. Most of ethyl acetate was extracted to the organic fraction, while almost all of the ethanol used and produced during hydrolysis of ethyl acetate remained in the aqueous fraction. High losses of xylose were observed in each step of the process, especially when hydrothermal treatment was applied. Further delignification of the cellulose-rich solid was observed during the hydrothermal treatment.
Energy balance revealed that all analyzed processes resulted in positive net energy gain, and energy recovery efficiencies were reasonable (40-70 %). The highest energy demands in the clean fractionation and hydrothermal treatment processes were associated with initial heating of the feed. It was found that if the heat streams were to be recycled in a commercial scale, the rejected heat would supply almost the entire heating needs of the process.
To illustrate possible energy and mass interactions on a commercial scale, a theoretical upscaling was performed based on the laboratory scale data. A theoretical facility producing 50 million gallons of ethanol per year was considered as an example. Clean fractionation upscaled mass and energy balances showed that ethanol produced during ethyl acetate hydrolysis was more than enough to supply the ethanol needs allowing the excess, which could be used in generating energy for the process. Due to losses during hydrolysis to ethanol and acetic acid, large amounts of ethyl acetate were predicted to be lost in a commercial scale facility. In the case of hydrothermal pretreatment, about 50 % of the ethanol product streams would have to be used in order to supply the energy requirements.
Economic analyses were based on the upscaling results and many general assumptions. It was found that the highest cost of the catalyzed clean fractionation process (with or without the hydrothermal post-treatment) was associated with ethyl acetate due to the large losses of this solvent in the sulfuric acid-catalyzed hydrolysis. Ethanol generated in the same reaction could partly compensate for ethyl acetate losses as an extra source of revenue (priced at $2.69/gal) and energy. Lignin was assumed to be sold as a value-added product, and would have to be priced at about $20/kg in order to achieve the assumed simple payback period (10 years). Hydrothermal pretreatment-based facility alternatives all showed similar trends. In this case, in order for the facility to make a profit, biomass feedstock and fermentation supplies would have to be free of charge.
Ethanol produced in the SSF of the biomass pretreated by catalyzed clean fractionation (with or without hydrothermal post-treatment) would have to be priced at $2.36/gal. Ethanol produced by SSF of the hydrothermally pretreated biomass would have to be priced at $15.25-22.96/gal depending on the feedstock used.

Library of Congress Subject Headings

Lignocellulose
biotechnology
lignin
feedstock
ethanol as fuel
cellulosic ethanol
plant biomass

Description

Includes bibliographical references (342-354)

Format

application/pdf

Number of Pages

389

Publisher

South Dakota State University

Rights

Copyright © 2012 Iwona Cybulska. All rights reserved

Share

COinS