x I
5 .2. 4'. .2
9 l 23. an.

um.) .r .

1..
«

 

Huh
Bunny“:
5 .’.V ,
> 4 ‘I
m 31,... r

 

 

m LIBRARY
306 Michigan State
4" University

 

 

 

This is to certify that the
dissertation entitled

SPECTROSCOPIC AND CHEMICAL CHARACTERIZATION OF
BIOMASS

presented by
LIZBETH LAUREANO—PEREZ

has been accepted towards fulfillment
of the requirements for the

Doctoral degree in Chemical Engineering

_7?rw\7c M

U Major Professor’s Signature

 

 

 

 

Date

I
.
c
.
a
.
.
.
v
-
o
u
n
n
u
u
n
o
.
o
o
9
.
.
.
.
u
u
o
o
o
o
o
.
'
n
n
.

MSU is an Afiinnative Action/Equal Opportunity Institution

 

 

PLACE IN RETURN Box to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

Wm
JAE “10153339”? 5: ‘35; 72 J.

 

 

20 72

T9
Irv

 

 

 

 

 

 

 

 

 

 

 

 

 

15

 

SPECTROSCOPIC AND CHEMICAL CHARACTERIZATION OF BIOMASS

By

Lizbeth Laureano-Pérez

A DISSERTATION
Submitted to
Michigan State University
In partial fulfillment of the requirements
For the degree of
DOCTOR OF PHILOSOPHY

Department of Chemical Engineering and Materials Science

2005

ABSTRACT

SPECTROSCOPICAND CHEMICAL cram CTERIZA TION 0F BIOMASS
By

Lizbeth Laureano-Pérez

Spectroscopic characterization of both untreated and treated material is being
performed in order to determine changes in the biomass and the effects of pretreatment on
crystallinity, lignin content, selected chemical bonds and depolymerization of
hemicellulose and lignin. The methods used are X-Ray diffraction for determination of
cellulose crystallinity (CrI); diffuse reflectance Fourier transform infrared (DRIFT) for
changes in C-C and C-0 bonds; and fluorescence to determine lignin content. Changes in
spectral characteristics and crystallinity are statistically correlated with enzymatic
hydrolysis results to identify and better understand the fundamental features of biomass
that govern its enzymatic conversion to monomeric sugars. Raman spectroscopy was also
used to create a statistical model that relates the spectral characteristics of poplar to its
enzymatic hydrolysis results. DRIFT can be used to compare various pretreatments and
their effect on the biomass along with their parameters. The PCR model gives not only
better correlation, but also better prediction for initial rate and 72-hr conversion for poplar
samples. On the other hand, MLR gives a better correlation and prediction for the AF EX
pretreated corn stover. This difference in model applicability is due to the nature of the
sarnples. The models for 72-hr conversion give a better correlation and prediction than
the initial rate models indicating that factors, other than lignin, biomass crystallinity and

acetyl content, affect enzymatic hydrolysis. The pretreated corn stover MLR model

indicates that the factor that most affects the initial rate is the aldehyde content or the
bonds between the lignin and hemicellulose. However, for the 72 hr conversion the
model indicates that lignin is the primary factor affecting hydrolysis. The PCR model

indicates that both initial rate and 72-hr conversion is more affected by O-H content.

DEDICATION

Dedicated to my mom, Lourdes Perez, who continually encourages and supports
me and whose personal sacrifices helped me to succeed in my endeavors. To my best
friends, Alexei J. Lozada-Ruiz and Ayrin A. Pefialoza Arroyo, whose love, support, and

encouragement were indispensable in the achievement of my goals.

ACKNOWLEDGEMENTS

I would like to thank Dr. Bruce Dale, for excellent guidance, encouragement,
understanding and support throughout my project. The best advisor anyone can have.

I am very grateful for the advice of my guidance committee members: Dr. R. M.
Worden, Dr. Dennis Miller and Dr. Merlin Bruening. In addition, I would like to thank
Dr. Mariam Sticklen for let me use her laboratory and equipment.

I am very thankful to Dr. Hasan Alizadeh, and Dr. Farzaneh Teymouri for their
suggestions, comments and technical assistance.

I am very grateful to Dr. Barbara O’Kelly and Linda Leon for their help in
financial and emotional support. Furthermore, I must thank the Chemical Engineering
and Crop and Soil Sciences office staff for all of their help.

Last but not least, I would like to thank all my colleagues and collaborators at
Auburn University, Dartmouth University, Purdue University, NREL and Texas A&M

University.

TABLE OF CONTENTS

Page
List of Figures ........................................................................ xi
List of Tables ......................................................................... xiii
Chapter 1: Introduction ......................................................... 1
Chapter 2: Literature Review .................................................. 5
2.1 Bioethanol ............................................................... 5
2.2 Cell Wall ................................................................. 6
2.2.1 Cellulose ...................................................... 7
2.2.2 Hemicellulose ................................................ 9
2.2.3 Lignin ......................................................... 10
2.3 Pretreatments ........................................................... 13
2.3.1 Ammonia Fiber Explosion (AF EX) ...................... 14
2.3.2 Ammonia Recycle Percolation (ARP) ................... 14
2.3.3 Uncatalyzed Explosion/ Dilute Acid ..................... 15
2.3.4 Controlled pH ............................................... 16
2.3.5 Lime .......................................................... 16
2.4 Pretreatment Effectiveness ............................................ 17
2.5 Enzymatic Hydrolysis Hindrance .................................... 17
2.5.1 Cellulose Crystallinity ...................................... 19
2.5.2 Lignin Content .............................................. 19
2.5.3 Changes in C-O, C-H bonds and deacetylation ........ 20
2.6 Analytical Analyses .................................................... 21
2.6.1 X-Ray Diffraction .......................................... 21
2.6.2 Fluorescence ................................................ 22
2.6.3 Diffuse Reflectance FT-IR(DRIFT) ...................... 23
2.6.4 Raman ........................................................ 24
2.7 Statistical Analysis ..................................................... 25
2.7.1 Multilinear Regression ..................................... 26
2.7.2 Principal Component Regression ......................... 28
Chapter 3: Effect of pretreatment on corn stover as measured
by DRIFT ........................................................... 33
3.1 Background ............................................................. 33
3.2 Materials and Methods ................................................. 37
3.2.1 Materials ...................................................... 37
3.2.2 Experimental Equipment and Procedures ................ 39

vi

3.2.2.1 APEX Treatment ................................ 39

3.2.2.2 ARP Treatment '. ................................. 42
3.2.2.3 Uncatalyzed Explosion/
Dilute Acid Treatment .......................... 42
3.2.2.4 Controlled pH Pretreatment .................... 43
3.2.2.5 Lime Pretreatment ............................... 44
3.2.3 Analytical Methods .......................................... 44
3.2.3.1 Diffuse Reflectance FT-IR(DRIFT) ........... 44
3.2.3.2 Enzymatic Hydrolysis ........................... 45
3.2.3.3 Simultaneous Saccharification
and Fermentation (SSF) ........................ 46
3.2.3.4 Acid Hydrolysis .................................. 46
3.3 Results and Discussion ................................................ 47
3.4 Conclusions ............................................................. 63

Chapter 4: Statistical correlation of spectroscopic analysis

and enzymatic hydrolysis of corn stover ........................ 65

4.1 Background ............................................................. 65
4.2 Materials and Methods ................................................ 69
4.2.1 Materials ...................................................... 69
4.2.2 Experimental Equipment and Procedures ................ 69
4.2.3 Analytical Methods .......................................... 69

4.2.3.1 Fluorescence ...................................... 70

4.2.3.2 Diffuse Reflectance FT -IR(DRIF T) ........... 70

4.2.3.3 X-Ray Diffraction ................................ 70

4.2.3.4 Enzymatic Hydrolysis ........................... 70

4.2.3.5 Simultaneous Saccharification

and Fermentation (SSF) ....................... 70

4.2.3.6 Acid Hydrolysis .................................. 70
4.2.4 Statistical Analysis .......................................... 70

4.3 Results and Discussion ................................................ 73
4.4 Conclusions ............................................................. 87

Chapter 5: Statistical correlation of spectroscopic analysis and

hydrolysis of poplar samples ..................................... 89
5.1Background .............................................................. 89
5.2 Materials and Methods ................................................ 92

5.2.1 Materials ...................................................... 92

5.2.2 Analytical Methods .......................................... 92

5.2.2.1 Fluorescence ...................................... 92
5.2.2.2 Diffuse Reflectance F T-IR(DRIFT) ........... 92
5.2.2.3 X-Ray Diffraction ................................ 92
5.2.2.4 Raman Spectroscopy ............................. 92
5.2.2.5 Enzymatic Hydrolysis ........................... 92
5.2.3 Statistical Analysis .......................................... 93

vii

5.3 Results and Discussion ................................................. 93

5.4 Conclusions ............................................................. 114
Chapter 6: Conclusions and Recommendations ............................. 116
Appendices ........................................................................... 119

Appendix 1: Ammonia MSDS ............................................. 120

Appendix 2: Safety Requirements and Equipment ..................... 125

Appendix 3: DRIFT Results (Raw Data) ................................. 126

Appendix 4: Corn Stover Characterization Raw Data .................. 138

Appendix 5: Poplar Characterization Raw Data ........................ 168
Bibliography ......................................................................... 211

viii

LIST OF TABLES

Page

Table 3.]: Corn stover main peaks in DRIFT spectra ........................... 37
Table 3.2. Composition of the corn stover ....................................... 38
Table 3.3: Changes in DRIFT results from changes in

AF EX temperature ..................................................... 51
Table 3.4: Changes in DRIFT results from changes in

ammonia loading ....................................................... 53
Table 3.5: Changes in DRIFT results from corn stover

moisture content ....................................................... 56
Table 3.6: Changes in DRIFT results from AF EX

pretreatment time ...................................................... 58
Table 3.7: Changes in DRIFT results from the different

Pretreatments .......................................................... 63
Table 4.1: AFEX Crystallinity Index .............................................. 74
Table 4.2: Different Pretreatment Crystallinity Indices ......................... 75
Table 4.3: Prediction using MLR model for corn stover ...................... 81
Table 4.4: Corn stover MLR data for hypothesis testing ....................... 82
Table 4.5: Correlation using various spectroscopic techniques ............... 83
Table 4.6: Prediction using PCR model for corn stover ........................ 86
Table 4.7: Corn stover PCR data for hypothesis testing ....................... 87
Table 5.1: Poplar crystallinity index .............................................. 97
Table 5.2: AN OVA correlations obtained using various

Techniques for MLR .................................................. 103

ix

Table 5.3: Correlations obtained using various techniques

for PCR ................................................................. 106
Table 5.4: Predictions using MLR model for poplar ............................ 110
Table 5.5: Predictions using PCR model for poplar ............................. 1 12
Table 5.6: Poplar MLR and PCR data for hypothesis testing .................. 1 13
Table A4.1: Crystallinity Index for AF EX pretreated corn stover ............ 138
Table A4.2: Crystallinity Index for pretreated corn stover ..................... 140

Table A4.3: MLR model coefficients for corn stover using all
spectroscopic data ................................................... 141

Table A4.4: PCR model coefficients for corn stover using DRIFT data. . . 142

Table A4.5: MLR Raw data for AF EX treated corn stover Unscrambler

modeling ............................................................. 143
Table A4.6: MLR Raw data for pretreated corn stover Unscrambler

modeling ............................................................. 144
Table A5.1: Crystallinity Index for TAMU poplar samples .................... 168
Table A5.2: MLR model coefficients for poplar using Raman, DRIFT

and XRD data ......................................................... 173

Table A5.3: PCR model coefficients for poplar using DRIFT data ........... 174
Table ASA: MLR Raw Data for poplar Unscrambler Modeling .............. 175

LIST OF FIGURES

Page

Figure 2.1: A three—dimensional molecular model of a cell wall ............... 8
Figure 2.2: Cellulose microfibril .................................................... 9
Figure 2.3: Most common lignin monomers ..................................... 11
Figure 2.4: Bonds in lignin ......................................................... 12
Figure 2.5: Enzymatic hydrolysis scheme ....................................... 18
Figure 2.6: Matrix R plotted in column space with the

first eigenvector ...................................................... 30
Figure 3.1: AF EX equipment setup ............................................. 41
Figure 3.2: DRIFT spectrum of potassium nitrate ............................. 48
Figure 3.3: DRIFT reproducibility .............................................. 49
Figure 3.4: Effect of AF EX pretreatment temperature on the

DRIFT spectra of corn stover ....................................... 51
Figure 3.5: Effect of ammonia loading on DRIFT spectra of

corn stover ............................................................ 53
Figure 3.6: Effect of moisture content of corn stover on the

DRIFT spectra ........................................................ 55
Figure 3.7 : Effect of pretreatment time on the DRIFT spectra

of com stover ......................................................... 57
Figure3.8: Effect of ARP in biomass as determined by DRIFT ............... 59
Figure 3.9: Effect of dilute acid pretreatment on DRIFT results ............. 60

xi

Figure 3.10: Effect of controlled pH pretreatment on the DRIFT

Results .................................................................... 61

Figure 3.1 1: Effect of various pretreatments on corn stover as
determined by DRIFT .............................................. 62
Figure 4.1: Fluorescence results for the different pretreatments ............... 76

Figure 4.2: Initial Rate MLR Model using a) DRIFT data only

and; b) all the data ..................................................... 78
Figure 4.3: MLR model for the 72 hrs glucan conversion using

a) DRIFT data only and; b) all the data ............................ 79
Figure 4.4: PCR model for initial rate using only DRIFT data ................ 84
Figure 4.5: PCR model for glucan conversion at 72 hrs using

only DRIFT data ....................................................... 84
Figure 5.1: Effect of acetyl content on DRIFT spectra ......................... 94
Figure 5.2: Effect of lignin content on the DRIFT spectra of

Poplar ................................................................... 95
Figure 5.3: Effect of crystallinity on the DRIFT spectra ...................... 96
Figure 5.4: Effect of acetyl content on fluorescence results ................... 98
Figure 5.5: Effect of lignin content on fluorescence results .................... 99
Figure 5.6: Effect of crystallinity on the fluorescent results .................... 99
Figure 5.7: Effect of acetyl content on the Raman spectra .................... 100
Figure 5.8: Effect of lignin on the Raman spectra .............................. 101
Figure 5.9: Effect of crystallinity on the Raman spectra ........................ 102
Figure 5.10: MLR model for the initial rate using DRIFT,

Raman and XRD ................................................... 104
Figure 5.1 l: MLR model for the 72 hr conversion using DRIFT,
Raman and XRD ................................................... 105

Figure 5.12: PCR model for initial rate using DRIFT data .................... 107

xii

Figure 5.13: PCR model for the 72 hr conversion using DRIFT

data only ............................................................. 108
Figure A3.1: AF EX DRIFT results (20%, 0.7:], 60-90C) .................... 126
Figure A3.2: AF EX DRIFT results (20%, 1.3:1, 60-90C) .................... 127
Figure A33: AF EX DRIFT results (40%, 0.5:1, 60-90C) .................... 128
Figure A3.4: AFEX DRIFT results (40%, 1:1, 60-90C) ...................... 129
Figure A3.5: AF EX DRIFT results (40%, 1:3, 70-90C) ...................... 130
Figure A3.6: AFEX DRIFT results (60%, 0.5:], 70-90C) .................... 131
Figure A3.7: AF EX DRIFT results (60%, 0.7:1, 60-90C) .................... 132
Figure A3.8: AFEX DRIFT results (60%, 1.3: 1, 60-90C) .................... 133
Figure A3.9: AFEX DRIFT results (80%, 1:1, 90C, 5-15 min) .............. 134
Figure A3.l0: AFEX DRIFT results (80%, 1:1, 100C, 5-15 min) ........... 135
Figure A3.ll: NREL samples DRIFT results ................................... 136
Figure A3.12: Lime treated sample DRIFT results ............................ 137
Figure A4.1: AF EX fluorescence results (20%, 0.7 :1, 60-90C) .............. 145
Figure A4.2: AF EX fluorescence results (20%, 1.3:], 60-90C) .............. 146
Figure A43: AF EX fluorescence results (40%, 0.5:], 60-90C) .............. 147
Figure A4.4: AF EX fluorescence results (40%, 1:1, 60-90C) ................ 148
Figure A4.5: AF EX fluorescence results (40%, 1.3:], 60-90C) .............. 149
Figure A4.6: AF EX fluorescence results (60%, 0.5:1, 60-90C) .............. 150
Figure A4.7: AF EX fluorescence results (60%, 0.7:1, 60-90C) .............. 151
Figure A4.8: AF EX fluorescence results (60%, 1:1, 60-90C) ................ 152
Figure A4.9: AF EX fluorescence results (60%, 1.321, 60-90C) .............. 153

xiii

Figure A4.10:

Figure A4.1l:

Figure A4.12:

Figure A4.l3
Figure A4.14
Figure A4.15
Figure A4.16
Figure A4.l7
Figure A4.l8

Figure A4.l9

Figure A4.20:

Figure A4.21

Figure A4.22:

Figure A4.23:

Figure A4.24:

Figure A4.25:

Figure A4.26:

Figure A4.27:

AF EX fluorescence results (80%, 1:1, 90C, 5-

15min) .............................................................. 154
AF EX fluorescence results (80%, 1:1, 100C, 5-

15min) .............................................................. 155
AF EX fluorescence results (60%, 1:1, 110C, 5-

15min) .............................................................. 156
ARP fluorescence results .......................................... 157
Controlled pH fluorescence results ............................. 158
Dilute acid fluorescence results ................................. 159
NREL samples fluorescence results ............................ 160
Lime treated samples fluorescence results ..................... 161
PCR initial rate model using fluorescence data ................ 162
PCR 72-hr model using fluorescence data ..................... 162
PCR initial rate model using XRD data ........................ 163
PCR 72-hr model using XRD data .............................. 163

PCR initial rate model using DRIFT and
fluorescence data ................................................ 164

PCR 72-hr model using DRIFT and fluorescence

data ................................................................. 164
PCR initial rate model using DRIFT and XRD

data ................................................................. 165
PCR 72-hr model using DRIFT and XRD data ............... 165

PCR initial rate model using fluorescence and
XRD data .......................................................... 166

PCR 72-hr model using fluorescence and XRD
data ................................................................. 166

xiv

Figure A4.28: PCR initial rate model using DRIFT, fluorescence

and XRD data ..................................................... 167
Figure A4.29: PCR 72-hr model using DRIFT, fluorescence and
XRD data .......................................................... 167
Figure A5.1: MLR initial rate model using DRIFT data ....................... 187
Figure A5.2: MLR 72-hr model using DRIFT data ............................. 187
Figure A53: MLR initial rate model using Raman data ....................... 188
Figure A5.4: MLR 72-hr model using Raman data ............................. 188
Figure A5.5: MLR initial rate model using DRIFT and Raman
data ................................................................... 189
Figure A5.6: MLR 72-hr model using DRIFT and Raman data ............... 189
Figure A5.7: MLR initial rate model using DRIFT and XRD
data ................................................................... 190
Figure A5.8: MLR 72-hr model using DRIFT and XRD data ................. 190

Figure A5.9: MLR initial rate model using DRIFT and
fluorescence data ................................................... 191

Figure A5.10: MLR 72-hr model using DRIFT and fluorescence

data ................................................................. 191
Figure A5.11: MLR initial rate model using Raman and XRD

data ................................................................. 192
Figure A5.12: MLR 72-hr model using Raman and XRD data ............... 192

Figure A5.13: MLR initial rate model using Raman and
fluorescence data ................................................. 193

Figure A5.14: MLR 72-hr model using Raman and fluorescence
data ................................................................. 193

Figure A5.15: MLR initial rate model using DRIFT, Raman
and fluorescence data ............................................ 194

XV

Figure A5.16:

Figure A5.17:

Figure A5.18:

Figure A5.19:

Figure A5.20:

Figure A5.21:
Figure A5.22:
Figure A5.23:
Figure A5.24:
Figure A5.25:
Figure A5.26:

Figure A5.27:

Figure A5.28:

Figure A5.29:

Figure A530:

Figure A531:

Figure A532:

Figure A533:

MLR 72-hr model using DRIFT, Raman and

fluorescence data .................................................. 194
MLR initial rate model using XRD, Raman

and fluorescence data ............................................. 195
MLR 72-hr model using XRD, Raman and

fluorescence data .................................................. 195
MLR initial rate model using DRIFT, XRD,

Raman and fluorescence data .................................... 196
MLR 72-hr model using DRIFT, XRD,

Raman and fluorescence data .................................... 196
PCR initial rate model using Raman data ........................ 197
PCR 72-hr model using Raman data ............................. 197
PCR initial rate model using XRD data ......................... 198
PCR 72-hr model using XRD data .............................. 198
PCR initial rate model using fluorescence data ................ 199
PCR 72-hr model using fluorescence data ..................... 199
PCR initial rate model using Raman and DRIFT

data ................................................................. 200
PCR 72-hr model using Raman and DRIFT data .............. 200
PCR initial rate model using XRD and DRIFT

data ................................................................. 201
PCR 72-hr model using XRD and DRIFT data ............... 201

PCR initial rate model using DRIFT and
fluorescence data ................................................. 202

PCR 72-hr model using DRIFT and fluorescence
data ................................................................. 202

PCR initial rate model using Raman and XRD
data ................................................................. 203

xvi

Figure A534:

Figure A535:

Figure A536:

Figure A537:

Figure A538:

Figure A539:

Figure A5.40:

Figure A5.4l:

Figure A5.42:

Figure A5.43:

Figure A5.44:

Figure A5.45:

Figure A5.46:

Figure A5.47:

Figure A5.48:

PCR 72-hr model using Rarnan and XRD data ................ 203
PCR initial rate model using Raman and

fluorescence data .................................................. 204
PCR 72-hr model using Raman and fluorescence

data .................................................................. 204
PCR initial rate model using XRD and fluorescence

data ................................................................. 205
PCR 72-hr model using XRD and fluorescence

data ................................................................. 205
PCR initial rate model using DRIFT, XRD and

Raman data ........................................................ 206
PCR 72-hr model using DRIFT, XRD and Raman

data ................................................................. 206
PCR initial model using DRIFT, fluorescence and

Rarnan data ........................................................ 207
PCR 72-hr using DRIFT, fluorescence and Raman

data ................................................................. 207
PCR initial rate using XRD, fluorescence and

Raman data ........................................................ 208
PCR 72-hr using XRD, fluorescence and Raman

data ................................................................. 208
PCR initial rate model using DRIFT, XRD and

fluorescence data .................................................. 209
PCR 72-hr model using DRIFT, XRD and

fluorescence data .................................................. 209

PCR initial rate model using DRIFT, Raman,
XRD and fluorescence data ..................................... 210

PCR 72-hr model using DRIFT, Raman,
XRD and fluorescence data ..................................... 210

xvii

Chapter 1
Introduction

Production of ethanol from lignocellulosic biomass such as corn stover could
improve energy security, reduce trade deficits, decrease urban pollution and contribute
little, if any to the net atmospheric carbon dioxide accumulation [124]. Biomass can be
transformed into liquid transportation fuels that have inherent convenience, cost and
efficiency. In this process, pretreatment is necessary (key) to achieve high glucose yields
from cellulose in enzyme-catalyzed processes. Pretreatment will break or affect the
complex hemicellulose-lignin shield that surrounds cellulose and limits its accessibility to
enzymes, making digestibility easier. Pretreatment alters many characteristics of the plant
material that impede digestion including: 1) cellulose crystallinity, 2) lignin content, 3)
acetyl linkages and 4) the complex hemicellulose-lignin shield that surrounds cellulose in
the plant cell wall.

The diverse composition of biomass lends itself to manufacture a variety of
products. Cellulose (40-50%) and hemicellulose (25-30%) can be broken down to sugars
for fermentation or chemical reaction to a wide range of fuels and chemicals. Lignin (15-
20%) can be converted into aromatic compounds or burned to provide heat and electricity
to the process. In addition, a significant amount of protein in some biomass can be
recovered for food and feed [123].

Pretreatment technology development has been pursued for decades, but lack of
fundamental understanding of pretreatment effectiveness has limited technological
applications. For example, several studies have tried to explain the roles of lignin content

and crystallinity on the hydrolysis rate, along with the effect of acetylation, pore volume

and surface area accessibility but contradictory results have emerged. No widely accepted
models for predicting the effects of pretreatment on plant materials exist. If we are to
unlock the energy and nutrients in lignocellulosic materials for food, feed, fuel and
chemical uses, we must better understand the fundamental factors that affect
lignocellulose conversion.

Better understanding and application of pretreatment would allow to recycle
biomass that otherwise is discarded and will make its transformation to fuel easier
allowing a better enzymatic hydrolysis and recovery of sugars for production of useful
products, such as ethanol. Pretreatment information will also be useful in the creation of
a model for product recovery from biomass. The optimization of biomass conversion will
reduce waste and decrease environmental pollution by utilizing discarded material and/or
replacing the oil used nowadays for fuel and chemical production

This research involves the spectroscopic characterization of lignocellulose
(initially corn residue) prepared using different pretreatments to determine their
effectiveness in reducing biomass processing cost by improving hydrolysis. Analytical
techniques used are X-Ray Diffraction to determine cellulose crystallinity; Diffuse
Reflectance Infra Red (DRIFT) for changes in C-C and C-0 bonds; and Fluorescence for
the determination of lignin content. Raman spectroscopy is also used in an effort to
compare the information obtained with DRIFT and x-ray diffraction and/or complement
it.

Pretreatment of lignocellulosic biomass is necessary to obtain high sugar yields by
enzyme catalysis. However, the fimdamental characteristics of biomass that limit its

enzymatic conversion are not clearly understood. A better fundamental understanding of

these factors would help improve pretreatment/hydrolysis systems. Toward this end of
improved fundamental understanding, leading biomass pretreatment techniques are being
studied in an integrated multi-university research project funded by the U. S. Department
of Agriculture’s Initiative for Future Agriculture and Food Systems (IFAF S). As part of
this joint research effort, Michigan State University (MSU) is using spectroscopy and
other methods to characterize corn stover pretreated by a variety of approaches including
aqueous ammonia recycle percolation (ARP) performed at Auburn University,
uncatalyzed hydrolysis and dilute acid hydrolysis, performed at Dartmouth University. In
addition, controlled pH treatment was performed by Purdue University, lime preteatrnent
was done at Texas A&M University and ammonia fiber explosion (AFEX) was
performed at MSU. The overall objective of the IFAF S research is to develop
comparative information on five different pretreatment operations, (e. g AF EX) for
production of sugars from hemicellulose and cellulose fractions of biomass for

fermentation or chemical reaction to a wide range of commodity products.

The current research objectives are:

/ Apply AF EX pretreatment to corn stover and determine optimum conditions
based on sugar and ethanol yields.

/ Monitor recovery, reactions and fate of lignin, hemicellulose, cellulose and
protein for each pretreatment by X-Ray Diffraction, DRIFT and Fluorescence
analysis.

/ Develop accurate material balances for the AF EX process.

/ Hydrolyze AF EX pretreated solids and evaluate their ferrnentability.

V Compare performance of the different pretreatment systems by X-Ray
Diffraction, DRIFT and Fluorescence results.

/ Develop a statistical model that would predict the ethanol/sugar yield based on
changes in the chemical structure, lignin content and crystallinity of the
biomass.

/ Study the biomass structure information provided by Raman Spectroscopy as

compared to X-Ray Diffraction and DRIFT for poplar.

Chapter 2
Literature Review

2.1 Bioethanol

Biomass is defined as all non-fossil organic materials that have an intrinsic
chemical energy content. It includes all water- and land-based vegetation and trees, or
virgin biomass, and all waste biomass such as; municipal solid waste, municipal bio-solid
(sewage) and animal waste (manure), forestry and agricultural residues, and certain types
of industrial waste [71]. The conversion of biomass to liquid fuels, such as ethanol, has
been the focus of much interest during the 20th century. The process of making alcohol
fiom cellulose, in principle, is relatively simple: after hydrolysis of cellulose to glucose
and a subsequent fermentation, the ethanol can be recovered by distillation. Much of the
United States’ (USA) energy use is derived from petroleum, over half of which is
imported, rendering us dependent on resources from unstable regions of the world. An
economically feasible biomass conversion technology would reduce crude oil
dependence. The attractiveness of bioethanol as a potential substance for replacement of
conventional fossil fuel lies in its low carbon dioxide release compared to when fossil
fuels are burned. The conversion of lignocellulosic materials to ethanol does not require
net energy input from fossil fuels [123]. Lignin, which is a by-product, can be burned to
provide the energy required for bioethanol production. The carbon dioxide released
during the production and the use of bioethanol can be converted back to biomass in the
cultivation of energy crops to provide new raw material for the production [123]. As a
result, the contribution of biomass ethanol to the greenhouse gas content of the

atmosphere is negligible.

In addition, domestically abundant sources of biomass including agricultural and
forestry residues are available for bioethanol production. Biomass feedstocks for energy
can be provided by short-rotation intensive-culture plantations of trees or plantations of
herbaceous plants such as sugar cane, switch grass and corn stover [51]. The biomass can
be converted by acid or enzymatic based approaches. Enzymes are used to break apart or
hydrolyze hemicellulose and cellulose chains to form their component sugars. The sugars
are fermented to bioethanol by adding yeast, bacteria or other suitable organisms. The
ethanol is then recovered and used as fuel. A very detailed study on bioethanol

production, possible uses, characteristics and demand is presented elsewhere [123].

2.2 Cell Wall

Plant cell walls directly affect the raw material quality of human and animal food,
textiles, wood and paper and may play a role in medicine [17]. Modification of various
cell wall constituents is a goal in the food processing, agriculture and biotechnology
industries. Successful achievement of this goal depends on understanding the molecular
basis for mechanical and structural properties of plant-derived materials. A better
understanding of plant structures and the effect pretreatment has on these structures, will
help identify of specific variables (e.g. crystallinity, acetyl content, types of bond) that can
be used to tune the pretreatment parameters (e. g pretreatment time, moisture content, etc.)
to obtain the desired product and/or manipulate the system in such a way as to obtain the

optimum yield.

Plants use complex polymers of arabinose, mannose, glucose and xylose to

intertwine with, and thus strengthen and stabilize homogeneous polymers of cellulose

[17]. These polymers are mainly in the plant cell wall. Currently, no definitive model of
the cell wall exists, particularly one that relates the cell wall composition to its
mechanical properties. However, the architectural features of the primary cell wall are
the following. The primary cell wall is made of two, sometimes three, structurally
independent but interactive networks. The fundamental framework of cellulose and
cross-linking glucans lies embedded in a second matrix of pectic polysaccharides. The
third independent network consists of the structural proteins or a phenylpropanoid
network. Cellulose and pectin networks are largely independent or only interact weakly
through hydrogen bonding. Pectin facilitates the realignment of cellulose microfibrils in
systems under strain, but herrricellulose interacts much more strongly with cellulose and
makes the network more rigid. Plant cell walls and structural tissues are primarily

composed of cellulose, a polymer of B(1,4)-linked cellobiose residues, hemicellulose and

lignins (Figure 2.1).
2.2.1 Cellulose

Cellulose is the most abundant polysaccharide on earth, accounting for 15% to
30% of the dry mass of all primary cell walls and an even larger percentage of secondary
walls [17]. Cellulose in ligrrocellulosics is composed of crystalline and amorphous
components. The amorphous component is digested more easily by enzymes than the
crystalline component. The crystalline cellulose exists in the form of microfibrils, which
are paracrystalline assemblies of several dozen (1-)4) B-D— glucan chains hydrogen-
bonded to one another along their length. Each (1 -)4) B-D- glucan may be several

thousands units long but individual chains begin and end at different places within the

microfibril to allow the microfibril to reach the length of thousands of micrometers and to
contain thousands of individuals glucan chains. The (194) B—D- glucan chains are
tightly linked by numerous hydrogen bonds, both side-to-side and top-to-bottom in a
lattice like manner. The glucan chains in the core of the microfibril have a precise
spacing (Figure 2.2). The arrangement of atoms in the unit structure of the microfibril

core has been determined by X-Ray Diffraction [108].

------------------------------------------

 

Cellulose microfian g x 1 f 5

E Junction Zone ‘ i

: Ca+ linked I

Xyloglucans fMA ‘ sax :
3 RG 1 RG 1 E

, - with arabinogalactans with arabinans I
Extensrn __. 1 ~

Arabinosides

Figure 2.1: A three-dimensional molecular model of a cell wall. Reprinted with

permission.

 

Figure 2.2: Cellulose microfibril. Reprinted with permission from Buchanan,

Bob 8.. Wilhelm Gruissem and Russell L. Jones. Biochemistg & Molecular

Biology of Plants, 3rd edition Courier Companies, Inc., 2001

2.2.2 Hemicellulose

Cross-linking glycans are a class of polysaccharides that can hydrogen—bond to

cellulose microfibrils. They may coat microfibrils but are also long enough to span the

aldopentoses (arabinose, xylose, galactose), which are in either pyranose or a furanose
form. The principal pentose sugar in hemicellulose is B-D-Xylopyranose. The other
common five-carbon sugar is arabinose, which is distinct in that it forms a furanose ring
structure. Arabinose is usually linked to the 2-and-3 carbons of xylose in arabinoxylan
and at the 3-and 6- carbons of galactose in arabinogalactan. Hemicelluloses also link the
polyphenolic portion of the plant cell in the three-dimensional structures, known as

lignin-carbohydrate complexes [17].

2.2.3 Lignin

The most distinguishing feature of secondary walls is the incorporation of lignins,
complex networks of aromatic compounds called phenylpropanoids. Afier cellulose,
lignins are the most abundant organic natural products known and account for as much as
20% to 30% of all vascular plant tissue. Lignins are irregular phenylpropane polymers
with different linkages and substitutions on the primary branch [123]. Lignins are
thought to be racemic (optically inactive). Practically, no lignin exists in primary walls.
The phenylpropanoids, hydroxycinnamoyl alcohol and “monolignols” (p-coumaryl,
coniferyl and sinapyl alcohols (Figure 3 [16])) account for most of the lignin networks
[17]. Non-woody plants contain lignins that appear to be formed from mixtures of
monolignols and hydroxycinnamic acids. The monolignols are linked by way of ester,
ether or carbon-carbon bonds. Monolignols form lignin. Lignin is covalently linked to
cellulose and xylans in ways that indicate the orientations of polysaccharides may serve
as a template for the lignin patterning. A range of cross-linking possibilities exists

including hydrogen-bonding, ionic bonding with Ca+ ions, covalent ester linkages, ether

10

linkages and van der Waals interactions (Figure 4 [17]). Lignin-carbohydrate interactions

exert a great influence on digestibility of forage crops by animals.

 

Figure 23: Most common lignin monomers

ll

 

 

 

 

 

 

 

. 0 Direct ester linkage.

3 9 Direct ether linkage

I o l-lydmxyclnnamln arid ester

_' o Hydrnxyclrlnamlc acid ether

1 a Fertilii: acid bridge

- o Demvrlmdlrerullc acld diester bridge

I o Dem-'drndlferullr: arid dicsterrther bridge

 

 

 

Figure 2.4: Bonds in lignin. Reprinted with permission from Buchanan, Bob 3..
Wilhelm Gruissem and Russell L. Jones. Biochemistg & Molecular Biology of Plants,

3rd edition Courier Companies, Inc., 2001

23 Pretreatments

Lignocellulosic biomass feedstocks typically contain 55%—7_5% by dry weight
carbohydrates that are polymers of five-and-six carbon sugar units [123]. These
carbohydrate polymers must be broken down to their respective low-molecular weight
sugar components before microorganisms can complete the conversion to ethanol. Due to
the location of the cellulose fraction within the cell wall, enzymatic access is restricted by

the lignin and hemicelullose interference. As a result, pretreatment of the biomass is

necessary.

Many pretreatments have been studied through the years [23, 57, 58, 125], each
one having their advantages and disadvantages. Pretreatments can be categorized as
chemical, physical or physicochemical treatments by the effect they have on the biomass.
The pretreatments can be acidic, alkaline or neutral. Strong acids can break glycosidic
linkages of polysaccharides, freeing the individual monosaccharide components. Some
alkaline pretreatments yield highly digestible cellulose and produce liquid streams rich in
extracted lignins and polymeric hemicellulose. For industrial applications, a pretreatment
must be effective, economical, safe, environmentally friendly and easy to use. This study
emphasizes the use of AF EX. A comparison between aqueous ammonia recycle
percolation (ARP), uncatalyzed hydrolysis, dilute acid hydrolysis, controlled pH, lime and

ammonia fiber explosion (AF EX) is also presented.

13

 

C3:
lg.
llSt
mi

blc

bicl
pri
ac

lea

sul

 

 

Pie

 

1th.

315.

 

23.1 Ammonia Fiber Explosion (AFEX)

The ammonia fiber explosion treats lignocellulosic biomass with moderate
pressure liquid ammonia, and an explosively release of the pressure [58]. The ammonia
can then be recovered and recycled. The small amount of ammonia that remains in the
biomass (~1% by weight of the biomass) serves as a nitrogen source for the microbes that
use the sugars enzymatically hydrolyzed from the lignocellulose [34]. AFEX uses
moderate pressures (up to 280 psi) and moderate temperatures (60-100°C) to treat the

biomass.

AF EX is thought to affect both chemical and physical characteristics of the
biomass. The chemical effects include cellulose decrystallization, henricellulose
prehydrolysis and lignin alterations. The physical effects include the increase of
accessible surface area and decrease in bulk density by disrupting the fiber. AF EX also
leaves behind small amounts of ammonia that can serve as a nitrogen source in
subsequent ferrnentations. These effects increase the susceptibility of the biomass to
enzymatic hydrolysis. However, AF EX does not change significantly the macroscopic

appearance of the substrate.

23.2 Ammonia Recycle Percolation (ARP)

In Ammonia Recycle Percolation (ARP) aqueous ammonia is used as a
pretreatment reagent in a packed bed flowthrough-type reactor in the recirculation mode
where the ammonia is continuously recycled. ARP is a delignification pretreatment that

also solubilizes significant amounts of xylan into the pretreatment effluent. The process

14

 

 

 

ho

 

181‘.

alll

all

su;
oil;
Ac:
Pie

1mm

 

involves treating biomass with an ammonium hydroxide solution at temperatures above
150°C and pressures around 325psi. Most of the delignification and hemicellulose
removal occur within 30 minutes of beginning the pretreatment. A detailed description of

the experimental setup is presented elsewhere [125].

ARP enhances the enzymatic digestibility of cellulose by removing lignin and
hemicelluloses and leaving the remaining solids containing nearly pure celluloses [60].

Physically, ARP increases the pore size and porosity of biomass.

233 Uncatalyzed Explosion/Dilute Acid

The uncatalyzed explosion is based on heating biomass rapidly with steam,
holding the material for a time and rapidly discharging to flash cool the product. The
temperature used is about 220°C for a few minutes. This process removes hemicellulose

and produces digestible cellulose.

The dilute acid pretreatment uses acidic hot water through a biomass bed, which
allows a fairly selective removal of hemicellulose fi'om biomass but the remaining solids
are high in lignin [60]. The acid breaks down the hemicellulose to form xylose and other
sugars. The hemicellulose or xylan component is converted into arabinose, xylose, plus
oligomeric sugar products that constitute 30% or more of the total recovered products.
Acid also catalyzes hydrolysis of the cellulose fraction to produce glucose. The
pretreatment uses high temperature (MO-180°C) and pressures from 50-200 psi for 1—40

minutes and an acid concentration of 0-1.5%, which may cause formation of undesirable

15

 

 

SI.

fr?

m

ELK

‘
.3

 

 

 

 

sugar degradation products such as furfural and hydroxymethyl furfural, which in turn

could be degraded to form tars that might inhibit the hydrolysis.

The main advantage of the dilute acid pretreatment includes the production of a
soluble pentose stream that can be physically separated from the particulate residue.
Secondly, a substantially increased reaction rate on enzymatic hydrolysis of the residual

cellulose portion results presumably due to the acid induced increased fiber porosity [56].

23.4 Controlled pH Pretreatment

The controlled pH pretreatment is carried out at a pH between 4 and 7 to result in
greater susceptibility of the cellulose to enzymes and also to minimize formation of the
monosaccharide degradation products, furfural and hydroxymethyl furfural, which
otherwise interfere with subsequent cellulose hydrolysis or ethanol fermentation. The
process uses high temperatures (180 to 200°C) and pressures of about 150-250 psi for 5-
15 minutes. The run is carried about at 200 g of corn stover per liter of deionized water;
in other words a mass ratio of 1:5 solid to liquid.

Physical changes include increase in pore size. There is also an increase in
accessible cellulose by decreasing its crystallinity and association with lignin.
2.3.5 Lime

Lime is used as a pretreatment reagent because it is inexpensive, safe and can be
recovered with carbonating wash water. The biomass is pretreated with lime and air at 25
to 55°C for 1-5 months. The lime loading is 0.1 g Ca(OH)2 lg biomass and a water

loading of 5 to 15 g of HZO/g biomass.

l6

 

 

 

If

Ell

at

exl

 

The treatment effect on biomass is some lignin removal and complete acetate
removal. Lime has a selective effect on hemicellulose. The likely mechanism is that lime
removes acetate groups fi'om henricellulose rendering it more accessible to hydrolytic
enzymes [23, 25]. This pretreatment produces calcium carbonate that needs to be
regenerated in order to recycle the lime, and other product that might be formed during
the pretreatment is calcium acetate, which also inhibits the hydrolysis.

2.4 Pretreatment Effectiveness

The effectiveness of a pretreatment is measured by its success in increasing the
susceptibility of the biomass to enzymatic hydrolysis. Enzymatic hydrolysis is
accomplished by cellulolytic enzymes. A mixture of different enzymes is normally used
to obtain efficient hydrolysis of the cellulose. The mixture should contain
endoglucanases, exoglucanases, and B-glucosidases. The endoglucanases randomly
attack cellulose chains to produce polysaccharides of shorter chain length, whereas
exoglucanases attach to the nonreducing ends of those shorter chains and remove
cellobiose molecules. B-glucosidases act on cellobiose and oligosaccharides to produce
glucose for fermentation into ethanol. A scheme of this process is presented in Figure 5
[124]. This research only focuses on cellulases and not in xylanases.

2.5 Enzymatic Hydrolysis Hindrance

Several structural and compositional factors affect the enzymatic digestibility of
lignocellulosic materials. The most generally cited factors are cellulose crystallinity,
cellulose protection by lignin, accessible surface area, hemicellulose sheathing and degree

of hemicellulose acetylation [23]. Analytical methods have been developed through the

17

years to measure these biomass properties in an effort to identify the effect these factors

have on the enzymatic hydrolysis.

endo-glucanase
Cellulose chain

exo-glucanase 1 'ns
Celloboose
MO W W “wants
Glucose
beta-glucosidase 1

Figure 2.5: Enzymatic hydrolysis scheme

18

2.5.1 Cellulose Crystallinity

It is well known that cellulose is composed of crystalline and amorphous
components [1, 17, 102, 123]. It is also known that the amorphous component is digested
more easily than the crystalline component [1]. Hence, anything that will increase the
amorphous component in cellulose will also increase the digestibility. Several
pretreatments have been used to study the decrystallization of cellulose. Pretreatment of
cellulose opens up the cellulose structure and reduces the interaction between glucose
chains.

The degree of crystallinity of cellulose is expressed in terms of the crystallinity
index (CrI) as defined by Segal et al.[102]. This index is determined by the ratio of the
crystalline peak to valley (amorphous region) in the diffractogram based on a monoclinic
structure of cellulose.

2.5.2 Lignin Content

Lignin is covalently bonded to polysaccharides in the intact plant cell wall.
Relatively little is known about the structure of these lignin-polysaccharides complexes.
It is known, however, that the composition is polymeric with a great variety of bonds.
The mechanism that explains the protective effect of lignin against polysaccharides
hydrolysis remains uncertain although a number of factors; like the degree and type of
cross-linkage to polysaccharide, the diversity of structures found in the lignin component
and the distribution of phenolic polymers through the cell wall are important [114]. Since
cell walls are solubilized by alkali it can be supposed that much of the lignin in those

walls is bound through ester linkages. Also, since reducing sugars were found in the low

19

molecular weight fiactions from alkaline solubilization, ether linkages of the lignin to the
phenolics are suggested [114].

In other studies cellulase inhibition by lignin was reported to be due to lignin-
enzyme adsorption [104]. Lignin also reduces fiber digestibility, presumably by
interlinkages with carbohydrates. Morrison [87] suggested that the phenolic acids may
act as cross-linking agents between lignin and a carbohydrate component consisting of a
backbone of [54,4 linked xylose with 1,3 linked arabinose side chains and a significant
amount of b—1,4 linked D-glucose which may have originated from a cellulose-like
polymer (probably xyloglucan).

2.53 Changes in GD, GE bonds and Deacetylation

Pretreatments that have a chemical effect on the biomass may do so by affecting
the types of bonds between the lignin-polysaccharides complexes, the hydrogen bonds
that keep the molecular chains of cellulose in a highly ordered arrangement (microfibrils)
[12] as well as the bonds of the acetyl groups in hemicellulose.

In lignin the monomer units produce a range of highly complex compounds,
which may contain some twenty different bond types. Aryl ether linkages, which
predominate in the uncondensed part of the molecule, and carbon-carbon bonds
contributing to the condensed portion, are the most numerous [114]. The bonds between
lignin and carbohydrates are predominantly ester-linked to arabinose side chains of
arabinoxylans. Xylans are extensively acetylated. It has been shown that an increase in

the degree of deacetylation increases the yield of sugars from enzymatic hydrolysis [7 5].

20

2.6 Analytical Analyses

Several analytical techniques have been developed in order to monitor the changes
in the biomass cell wall structure and composition and to determine the effect these
changes have on the enzymatic hydrolysis. A brief explanation of the different
techniques for biomass characterization is presented here. Cellulose crystallinity is
measured with X-ray diffraction. The changes in OH, C-O bonds and acetylation are
measured with DRIFT. The amount of lignin in the biomass is measured with
fluorescence spectroscopy. Raman spectroscopy is also used as a way to compare and/or
complement the data obtained from the other spectroscopic analyses.
2.6.1 X-Ray Diffraction

X-ray diffraction is the use of electromagnetic radiation to determine the
interplanar spacings and crystal structure. The pattern of diffraction obtained is directly
related to the unit cell size and shape. X-ray diffraction measurements (XRD) of relative
crystallinity showed that the crystalline structure of cellulose is affected by pretreatment.
Cellulose crystallinity decreases in extent and changes its molecular arrangement to an
amorphous form that is highly susceptible to enzymatic hydrolysis.

The crystallinity as defined by Segal et al. was calculated using:

(1200 — Iain) *100

Crystallinity Index (CrI) = (2.1)

 

I200

° 1200is the intensity of the peak at 22.30°

' Iam peak is the intensity of the peak at 18°

21

The planes were calculated using the equation 2.2 obtained from Cullity and Stock

[30] for a monoclinic unit cell.

 

 

2 2 ° 2 2
__ h It srn fl+l__hlcos,6] (2.2)

1 l
= ——+—
d2 sin2 ,6(a2 b2 c2 ac

Where:
h, k, l = Miller Indices
d = interplanar spacing
l '= Copper wavelength (target) = 1.54 A
a, b, c = axial lengths; a = 8.01 A, b = 8.17 A; c = 10.36 A [46]
b = angle between axes a and c ; b = 97.3°
Each spectrum is collected using the 0-20 method in a Rigaku Rotaflex 200B
diffractometer at 45 kV and lOOmA with slits size 0.5°,0.5°, 0.3° and 045°, respectively.
The sample is placed vertically in the slide using double-sided tape to hold it in place and

analyzed using the horizontal goneometer.

2.6.2 Fluorescence

The determination of Klason lignin in biomass (the more accurate estimate of
plant cell wall lignin reference) is part of two-stage sulfuric acid hydrolysis that is
commonly used to determine the neutral sugar components of cell wall polysaccharides
[66]. However, lignin behaves as if it contained one single chromophore, which makes it

a good candidate for fluorescence analysis [82]. Fluorescence is a form of luminescence

22

in which light is emitted by a molecule following excitation with radiation of a shorter
wavelength. Complex conjugated and/or aromatic compounds generally absorb in the
ultraviolet range and are fluorescent, since excited states may be stabilized through
delocalization over the aromatic system [14]. Fluorescence has been attributed to
structures in lignin such as aromatic carbonyl groups; biphenyls, phenylcoumarins and
stilbenes, and changes in emission are thought to be due to the formation or destruction of
these fluorophores. The intensity of the peaks in the fluorescence spectra is directly
related to the lignin concentration in the sample. Fluorescence is a surface technique. It
does not penetrate the solid but measures what is present at the surface.

Fluorescence spectra are recorded using a SPEX-3 Fluorolog instrument. Auto
emission spectra were obtained at an excitation wavelength of 350 nm with an interval of
0.5 nm. The excitation and emission slits were set at 3 and 5 nm, respectively. The solid
sample holder is filled with powdered sample and is held in place with a quartz cover
slip. Sample was placed 45 ° from the incident beam. The mode of detection was front
face.

2.63 Diffuse Reflectance Fourier Transform Infrared (DRIFT)

A molecule absorbs infrared (IR) radiation only when the permanent electric
dipole changes during molecular vibration. In IR a more polar bond gives greater peak
intensity and nonsymmenic vibrations are stronger. For example, carbonyl groups are
strong in IR [107]. IR spectroscopy is a very useful tool for obtaining rapid information
about the structure of biomass constituents and chemical changes taking place in biomass
due to pretreatments. Diffuse reflectance infrared (DRIFT) has been used to study the

differences in the chemical structures of biomass and to estimate the relative amount of

23

lignin and cellulosic polymer [94]. It is a quick, easy, and nondestructive method (the
structure of the biomass is maintained as compared to other analyses such as Klason
lignin and wet chemistry). In addition, changes in relative proportions of crystalline and
amorphous cellulose accompanying chemical treatments are reflected in DRIFT. These
results, however, have not been proven to be precise quantitatively but are useful
qualitatively.

The DRIFT results presented were obtained in a Nicolet Protege 460 Magna IR
instrument with an auxiliary experiment module for diffuse reflectance. Spectra were
obtained using 100 scans of the sample, triangular apodization, a resolution of 16 cm‘1
and an interval of 1 cm". The sample is loaded in the holder and analyzed. The
equipment was calibrated using KBr for water and air background. Peaks are identified
following Stewart et al [107].

2.6.4 Raman Spectroscopy

Raman is a light scattering process. Raman spectra are obtained by irradiating a
sample with a powerful laser source of a visible or infrared monochromatic radiation
[105]. In Raman the wavelength of a small fraction of the radiation scattered by certain
molecules differs from that of the incident beam and the shifts in wavelength depend
upon the chemical structure of the molecules responsible for the scattering. Raman
intensity is usually directly proportional to the concentration of the active species. In
general, in Raman less polar bonds give greater scattering, for example, C-C double
bonds are strong in Raman.

The advantages of this non-destructive technique include small sample

requirements, minimal sensitivity toward interference by water, the spectra detail and the

24

conformational and environmental sensitivity [105]. Raman spectra have regions that are
useful for functional group detection and fingerprints region that permit the identification
of specific compounds. Raman studies are potentially useful sources of information
concerning the composition, structure and stability of coordination compounds. A
drawback to the use of Raman is the interference by fluorescence of the sample or
impurities in the sample.

The Raman analysis is performed with a Hololab Series 1000 from Kaiser Optical
Systems, Inc. The sample is analyzed without removing it from a clear plastic bag. The
laser probe (Model HFPH-632.8nm) is placed in close contact to the sample and exposed
to the laser 10 times for 5 seconds each time and the spectra collected. The bag spectrum
is “invisible” to Raman.

2.7 Statistical Analysis

The analysis of the different spectra (DRIFT, XRD, Fluorescence, Raman) for the
different samples create an enormous amount of data that must be analyzed and correctly
interpreted in order for it to be useful. Multivariate analysis allows the scientist to relate
and model different variables simultaneously. In other words, multivariate statistics is a
collection of powerful mathematical tools that can be applied to chemical analysis when
more than one measurement is acquired for each sample [10].

Multivariate analysis is used for different purposes of which the following are
important: (1) Structural simplification, by transforming a set of interdependent variables
to dependence, or reducing the dimensionality of a complex (matrix). (2) Classification,
whether the objects fall into groups or clusters, or are randomly scattered. (3) Grouping

variables, whereas classification is concerned with the grouping of the objects. (4)

25

Analysis of interdependence, whether a variable is a linear or nonlinear function of the
others. (5) Analysis of dependence, one or more variables are singled out to examine
their dependence on the others, as in regression analysis. (6) Hypothesis construction and
testing [68].

Multivariate calibrations are useful in spectral analyses and can greatly improve
the precision and applicability of quantitative spectral analysis [114]. With multivariate
calibrations, empirical models are developed that relate the spectral data for multiple
samples to the known concentration of the samples. These empirical relationships can
then be used in multivariate prediction analyses of spectra of unknown samples to predict
their concentrations.

The analysis presented here consists of two parts: calibration and prediction.
Analysis begins with the construction of a data matrix (R) obtained from the instrument
(variables) for a given set of calibration samples. A matrix of concentration values (C)
using independent methods such as rate of enzymatic hydrolysis is then constructed. The
goal of the calibration is to produce a model that relates the data from the instrument to
the results by an independent method. The prediction then uses the model to predict the
value for an unknown sample.

2.7.1 Multiple Linear Regression (MLR)

The goal of MLR in this study is to find a linear combination of the variables such
that the rate of enzymatic hydrolysis value estimated by the model is as close to the
experimental value as possible. The criterion of closeness for MLR is defined as
minimizing the sum of the squares of the deviations of the predicted values from the true

values.

26

The procedure followed for the coefficient calculation for MLR is presented

below. As a way to demonstrate the process let us use a spectral matrix (R) and a

concentration vector (c)
39 29 30
R = 18 15 ; c = 20
11 6 10

The rows represent the samples and the columns the variables. In matrix R columns are
spectral data and in the c column is the sample concentration as measured by an
independent method. The two columns in R define a subspace in a row space. To
perform MLR, one can plot the c in the same row space and project c onto the plane

formed by the r. and r2 (columns of R). The regression coefficients are in this case.

0.20
s =
0.85
From S one can estimate c;
39 29 32.5
, 0.20
pr01c= 18 15 = 16.4

11 6 7.3

To predict the concentrations of an unknown sample with a given vector (e.g. runk=

[10,5]) one multiplies the vector by S
0.20
cpred=[10 5 =6.25
0.85

In statistical language MLR is the regression of the columns of c onto the space defined

by the columns ofR [110].

27

The MLR is a straightforward matrix algebra, which makes the analysis simple
and fast. Since there is direct relationship between the model coefficients and the
parameters included in the regression, an easier identification of the factors affecting the
model is obtained by simple observation of the absolute value of these coefficients. A
limiting factor for this regression is the need for more samples than variables, making it
necessary to identify the important parameters to consider in the model in anticipation
and eliminating data points that might be considered as noise. This regression utilizes the
samples characteristics (parameters) to make the prediction.

2.7.2 Principal Component Regression (PCR)

This model building procedure has two steps. The first step is the determination
of the eigenvectors of factors for the matrix R; eigenvectors are used to redefine the
variables using a smaller number of factors. This approach is used to convert R into a
score matrix U by projecting the R matrix onto the space defined by the eigenvectors.
The first principal component is the linear combination of the original variables that
points in a direction that is more correlated to all the columns in row space. The principal
component is also the direction that best describes the variation of the sample. At the
moment of factor building PCR ignores the matrix C, which is only used in the second
part of the model. The second step of the PCR method uses MLR to regress the C matrix
onto the score matrix as US=C, where S is a matrix of regression coefficients. The
eigenvectors and the matrix of regression coefficients (S) form the PCR model. One way
of viewing this procedure is that MLR uses all the space described by the columns of R,
whereas PCR determines the subspace (plane formed in the space formed with the rows

of R as the axes (row space)) by possibly ignoring some of the eigenvectors.

28

it

_._-. .

 

 

I 4‘—
7....~.‘..r ._.u..
“‘ “TEL—1.- ‘-
m

5P

10

Pr.

C01

 

 

 

For principal component analysis (PCA), the first step in data analysis is the
calculation of the average spectra. This average is then subtracted from the individual
spectra, in order to study the deviations of the data (mean-centered). The columns are
then scaled, dividing each entry by the variance of the column. This gives equal weight

to each column. A mean-centered and scaled matrix R is presented below.

2 4
1 2
R: 0 0
-1 —2
b-2 —4.l

 

 

Principal component analysis is then performed on the covariance matrix RTR formed
from the mean centered and scaled matrix. Plotting the matrix in column space (Figure
2.6) gives the data distribution and shows the eigenvalue. In this case the eigenvector

covers all the data.

29

fav _‘—. — —_.v H_— ._._ -i» _ Q * A. ———“r'- .__-._ ‘4 _k i 7 7 7 V .v V _ i

l l
Column 2

k i __ __ —_ _.__—_ ____ ~_. _ ._. .._. .W 9.. _-..___. __ __. __ .- _ a _~_..__ .__ . ..__ . - __ ‘u.

 

 

 

Figure 2.6: Matrix R plotted on column space with the first eigenvector

The eigenvector from the figure is vT = [0.447, 0.894]. The first factor of the principal
component can be shown to equal a linear combination.
Rv = u (23)
Where:
R = original matrix response
v = first eigenvector of RTR

u = score vector (projection of R onto the first eigenvector)

30

 

hi;

lSl

 

 

When more than one eigenvalue is calculated and use to form an I x J matrix of U scores
the original variables in R can be expressed:
R = UV‘"(VV")'l ‘ (2.4)
Where:
VT(VVT)'l = generalized inverse of V

In PCA notation the elements in VT(VVT)'l are called principal component
“loadings”, they range from —1 to +1 and are the cosine of the angles between the
eigenvectors and the variable axes. High loadings correspond to high correlation and
small loadings (orthogonality) to low correlation.

In more complicated situations, like the one analyzed here, the object can be of
higher dimensionality than two, and eigen values are calculated until the sample variation
is explained. In the calibration step R is re-expressed as a score matrix U, by projecting
R into the eignevectors V:

U = lRV (2.5)
(U is composed of the original data in a new coordinate system described by the
eigenvectors) then regressing the C matrix as:

C = US (2.6)
The prediction of an unknown sample uses the V and S derived in the calibration. To
predict the concentration of an unknown sample (Cu...) the spectral data matrix of the
unknown R.“ is multiplied by the eigenvector matrix:

Um = R.“ V (2.7)

and then multiplied by the regression matrix

c = SRm (2.8)

unk

31

PCR is a more complex matrix algebra that uses orthogonality of the data to create
components that are independent from one another in order to reduce the dimensionality
of the data. An advantage of this regression is the fact that due to this processing of the
data, all the data available can be used and, the analysis extract the relevant information
and removes the noise from the data. This in turn causes a variable reduction and data
compression giving a more clear view of the important factors or relation between factors
that affects the model. It uses the relationship between the parameters to make the

predictions.

32

 

 

 

 

cc?

b0

“‘0

 

COL..-

 

Chapter 3
Effect of pretreatment on corn stover as measured by DRIFT

3.1 Background

Corn stover is a crop residue whose major constituents include cellulose,
hemicellulose, lignin, protein, and structural inorganics [52]. The henricellulose in corn
plants is a complex network consisting of an acetylated xylan backbone with branches
incorporating galactose, arabinose and uronic acids. The lignin in corn stover contains
both syringyl and guaiacyl aromatic rings and the side chains are enriched in ester
groups. Cellulose is composed of both crystalline and amorphous regions. These
cellulose polymers are held tightly together in the microfibril structure by hydrogen
bonding

Cellulose is interconnected to the hemicellulose in the plant cell wall while lignin
works as glue that keeps the structure together. The hemicellulose, xyloglucan, is
hydrogen bonded to the surface of the cellulose microfibrils. The pectic polymers
arabinogalactan and rhamngalacturonan interconnect the cellulose rrricrofibrils through
the hemicellulose. Lignin is a three dimensional polyphenol and can be visualized filling
the space between microfibrils, and also penetrating the spaces between elementary
fibrils in noncrystalline regions [48]. Cell wall polymeric lignin is covalently bound to
herrricelluloses in the plant cell wall. Ferulic and para-coumaric acids are esterified to
cell wall polysaccharides and appear to be the primary means of lignin attachment to cell
wall polysaccharides [69].

These substituted cinnamic acids, mainly ferulic and p-coumaric acids, are present
in the cell walls in different bonding arrangements and different forms. F erulic and p-

coumaric acid can be linked by ester bonds alone to arabinose residues or arabinoxylans

33

 

lr.

pl

of

slr

 

Ill;

for

 

in secondary walls and to arabinans and galactans in some primary walls. A high
proportion of ferulic acid residues is also linked to secondary walls by ether bonds and
may act as ester/ether bridges [6]. In addition 5-5’-dihydrodiferulic acid has been found
as a diester bridge between two polysaccharide chains as well as being further joined by
an ether link to lignin [88]. Homogeneous and heterogeneous cyclobutane-type dimers
of both ferulic and p-coumaric acid have also been identified and may be linked in
similar ways.

It is believed that the cross-linkage effect on cell wall digestibility may be more
important than concentration of the lignin or phenolic acids. We usually are not aware of
these influences when studying normal forage samples because the cross-linkages are
formed in conjunction with the deposition of lignin and phenolic acids in the developing
plant cell wall [64].

The types of bonds present in the cell wall molecules can be used to identify the
cell wall components. Infrared (IR) spectrometry has made the greatest contribution to
the knowledge of structures because of its sampling versatility, high resolution and
relative freedom from environmental pressure and temperature limitations [112]. Diffuse
reflectance F ourier-transform infrared (DRIFT) spectroscopy measures the changes in the
amount of bonds present based on the intensity of the stretching, bending or deformation
vibrations of the different bonds. DRIFT has been identified as a reliable method for
monitoring structural changes in biomass [7,12,14,39,94,107]. The diffuse reflectance
technique measures the spectrum, which, when converted into Kubelka-Munk (K-M)
format, is proportional to the sample concentration [29]. The diffuse reflection

spectroscopic technique was developed to facilitate analysis of materials such as papers

34

and powders in their neat state. When a material is illuminated, some of the radiation
penetrates the sample and some is reflected from the surface. The portion that penetrates
the sample undergoes scattering at a large number of points in its path. The fraction of
this radiation that comes back out of the sample is the diffusely reflected component.
This component is theoretically described by Kubelka-Munk model.

The K-M theory relates sample concentration and scattered radiation intensity

[42]. The K-M equation is as follows:

(I-R.)” _£
2R _

co

f(R..)= (3.1)

Where R... is the absolute reflectance of the layer, s is a scattering coefficient, and k is the
molar absorption coefficient.

The K-M theory predicts a linear relationship between the molar absorption
coefficient, k, and the peak value of flRQ) for each band, provided 3 remains constant.
Since s depends on particle size and range, these parameters should be made as consistent
as possible if quantitative data are needed. The K-M equation is only expected to hold
for moderately absorbing species at controlled particle size. Therefore the samples
analyzed were ground to the same particle size and loaded without pressure in the sample
cup. The greatest contribution to flRm) originates in the first few layers of sample. Area
or peak intensity of infrared absorption bands in the spectra can be used as a measure of a
functional group concentration in the analyte.

Several peaks have been identified in the DRIFT spectra of corn stover samples,

both treated and untreated. Seven main peaks are identified and presented in table 3.1.

35

 

 

 

(Ill

inc-

 

Each peak can be related to a cell wall component. The peak present at 3400 cm1 is due
to the O-H stretch. The C-H stretch peak is observed around 2900 cm". An increase of
these peaks signifies an increase in these kinds of bonds, which are more abundant in
smaller molecules. Hence, the concentration of these bonds is directly related to
breakage of biomass into smaller molecules. The hydrogen bonds in the polymers have
been broken down to produce carbohydrates. Another peak can be found around 1720

1 related to the ester carbonyl. Such bonds are present in hemicellulose. The

cm'
aldehyde peak is found at 1640 cm". These types of bonds are found in the
hemicellulose-lignin complex. A decrease in these peaks will be directly related to
delignification and hydrolysis of hemicellulose. The peak at 1595 cm'1 is due to aromatic
ring stretch that is strongly associated with 00 stretching mode. The peak at 1510 cm'1
is due to ring stretch vibration. Both peaks are associated with the monoligols bonds that
form lignin such as p-coumaryl, coniferyl alcohol and sinapyl alcohol. The peak at 900
cm'1 is due to the antisymmetric, out of phase ring stretch of amorphous cellulose.

Since the plant cell wall is a strongly interrelated matrix of structures rather than
merely a complex of isolated fractions the effect that pretreatment has on the cell wall
will affect all parts of the cell wall. Spectra in which all peaks vary can not be used for
quantification purposes. When quantification of the component concentrations is needed,
a reference peak is desired, a peak that does not change from one sample to the other.
This peak will provide a reference point on which all the peak changes are based. To
quantify precisely the effect that the pretreatment has on the biomass and in an effort to

increase the precision of the analytical method an internal standard was selected. This

standard, potassium nitrate, was chosen because it has a relatively simple spectrum. The

36

peak in the standard sample that did not interfere with the biomass main peaks was the
one chosen. By using this technique the differences in particle size and physical

properties would not affect the quantitative results [47, 52, 88].

Table 3.1: Corn stover main peaks in DRIFT spectra

 

 

 

 

 

 

 

 

 

 

 

Functional Group Absorption (cm'T
Alcohol O-H stretch 3400
Alkyl C-H stretch 2900
Ester C=O stretch 1720
Aldehyde C=O stretch 1640
Aromatic 00 stretch 1595
Aromatic C=C stretch with 1510
CEC stretch contribution
Ring C=C stretch 900
3.2 Materials and Methods
3.2.1 Materials

The feedstock used by all the collaborating institutions, corn stover, is milled (to
pass a 6-mm screen) and dried (about 10% moisture content). Corn stover and its
composition were provided by National Renewable Energy Laboratory (N REL) Golden,

CO. The composition is presented in Table 3.2.

37

 

Table 3.2: Composition of the corn stover

 

 

 

 

 

 

 

 

 

 

 

 

 

Component Percentage (based on dry basis)
Glucan 36.1
Xylan 21.4

Arabinan 3.5

Mannan 1.8

Galactan 2.5

Lignin 17.2

Protein 4.0

Acetyl 3.2

Ash 7.1

Uronic Acid (est) 3.6

Non-structural Sugars 1.2
Total 101.6

 

Auburn University, Dartmouth University, Michigan State University, Purdue
University and Texas A&M University pretreated the corn stover with their treatment of
expertise (ARP, Steam, AF EX, Controlled pH and Lime, respectively). The biomass is
then provided to the other universities to be analyzed. B-glucosidase (Novozyrnesl 88, lot
number 11K1088) was obtained from Sigma, St Louis, MO. Cellulase enzyme (Spezyme
cp) was provided by NREL, CAS 9012-548, activity: 28 F PU/ml; the activity was
measured based on NREL standard filter paper unit assay protocol, LAP-006. A Sigma

(rt-cellulose (catalog # C-8002, lot number 11K0246) provided by Auburn University was

38

 

[LA

3

 

 

used as a standard in the analyses. Anhydrous ammonia was obtained from AGA
(Lansing, MI).
The internal standard used was potassium nitrate (Baker, ) and ground to the same

particle size as the powder samples.

3.2.2 Experimental Equipment and Procedures

3.2.2.1 AFEX Treatment

The AF EX process treats lignocellulosic materials with liquid ammonia under pressure
and then the pressure is rapidly released. The moisture of the material treated ranges
from 20% to 80 % moisture (dry weight basis). The ammonia to biomass ratio used
ranges from 0.5:] to 1.3:] (masszmass) and pressure are generally up to 300 psia. Fifieen
grams (15 g) of previously chopped and cleaned corn stover provided by NREL in
Golden, CO. is prewetted in order to obtain the desired moisture content, and loaded in a
300 mL stainless steel vessel (PARR Instrument Co., IL). The vessel is topped with
stainless steel pellets (approximately 1mm diameter) to occupy the void space and thus
minimize transformation of the ammonia from liquid to gas during loading, and then the
lid is bolted shut. The predetermined amount of liquid ammonia is delivered to the vessel
using precalibrated ammonia cylinders to reach the desired ammonia to biomass ratio.
The temperature of the vessel is increased using a 400W PARR heating mantle until the
target temperature is obtained. After reaching the target temperature/pressure the reactor
is held at these conditions for 5 minutes and then the pressure is suddenly released. A

picture of the AF EX experimental apparatus is presented below (Figure 3.1). The

39

 

”"3

 

 

pretreated corn stover is then removed from the vessel and left under a fume hood until
the remaining liquid ammonia evaporates (approximately 24 hr). The treated samples are
kept in plastic bags at 4°C for further analysis. In AF EX the rapid pressure release
literally blows the fiber apart, greatly increasing the surface area available for enzymatic
and microbial attack by splitting fiber bundles axially (across the fiber radius). The
cellulose swelling or decrystallizing effect of liquid ammonia also distends the unit cell
of crystalline cellulose, opening cellulose up for hydrolysis on the molecular level.
Essentially complete conversion of plant cellulose and hemicellulose to fermentable

sugars is achieved by enzymatic hydrolysis of many AFEX treated materials [37].

40

cylinder

 

 

Figure 3.1: AF EX experimental setting

 

 

 

3C

l0;

3.2

hO:

 

 

 

3.2.2.2 ARP Treatment

In Ammonia Recycle Percolation (ARP), aqueous ammonia solution (5—15 wt.%)
is used as a pretreatment reagent in a packed bed flowthrough-type reactor in the
recirculation mode where the ammonia is continuously recycled. The process involves
treating biomass with an ammonium hydroxide solution at temperatures ranging from 80
—180°C and pressure around 325psi. A detailed description of the experimental setup is

presented elsewhere [125].

Aqueous ammonia depolyrnerizes lignin including breaking lignin-hemicellulose
bonds but degrades little cellulose. ARP achieves high and adjustable degrees of
delignification for hardwood and agricultural residues but is somewhat less effective for
softwood-based pulp mill sludge. Removing lignin helps by increasing cellulose
accessibility to cellulase and by reducing nonproductive binding of cellulase to lignin.
The digestibility of ARP treated corn stover was 90% with 10 FPU/g-glucan of enzyme
loading, far better yields than possible when more enzyme is used with al.-cellulose [18].
The challenge for ARP is to reduce liquid loadings to keep energy costs low.

3.2.23 Uncatalyzed Explosion/Dilute Acid Treatment

The uncatalyzed explosion is based on heating biomass rapidly with steam,
holding the material for a time and rapidly discharging to flash cool the product. The
temperature used is about 220°C for a few nrinutes. Pretreatment performed without
addition of dilute acid (i.e. using autohydrolysis) produces hemicellulose sugars yields

lower than when dilute acid is added.

42

In the flowthrough dilute acid pretreatment hot water flows through the biomass
bed. Percolation reactors are used because they reduce the treatment time causing fewer
sugars to be degraded. Forcing liquid through a packed biomass bed enhances
hemicellulose and lignin removal and gives high yields of hemicellulose and cellulose
sugars even without acid addition. However, percolation or flowthrough reactor are
challenging to implement commercially, and the high amounts of water used result in
high-energy requirements for pretreatment and product recovery. This pretreatment uses
high temperature (140-1 80°C) and pressures from 50-200 psi for 1-40 minutes and an

acid concentration of 0-1 .5%.

The dilute-acid pretreatment (~0.5-l .0% sulfuric) at moderate temperatures
(~140-190°C) is effective in removing and recovering most of the hemicellulose as
dissolved sugars, and glucose yields from cellulose increase with hemicellulose removal
to almost 100% for complete hemicellulose hydrolysis [83]. Lignin is disrupted,
increasing cellulose susceptibility to enzymes [124]. Sulfuric acid is cheap, up to 90%
hemicellulose yields are achieved, and enzymatic hydrolysis yields of glucose can be
over 90% [124]. Nonetheless, dilute acid pretreatment results in costly materials of
construction, high pressures, neutralization and conditioning of hydrolyzate prior to

biological steps, slow cellulose digestion by enzymes, and non-productive binding of
enzymes to lignin [32,39]
3.2.2.4 Controlled pH Pretreatment

The controlled pH pretreatment is carried out at a pH between 4 and 7. The

process uses high temperatures (180 to 200°C) and pressures of about 150-250 psi for 5-

15 minutes. The run is carried about at 200 g of corn stover per liter of deionized water;

43

in other words a mass ratio of 1:5 solid to liquid. The goal is to stop hemicellulose
hydrolysis with formation of soluble oligomers and minimize break down to sugar
monomers that are subject to subsequent degradation reactions, hurting yields. This
system is being applied to release hemicellulose sugars.
3.2.2.5 Lime Pretreatment

Lime is used as a pretreatment reagent. The biomass is pretreated with lime and
air at 25 to 55°C for 1-5 months. The lime loading is 0.1 g Ca(OH)2 /g biomass and a
water loading of 5 to 15 g of H20/g biomass. Lime also removes acetyl groups that have
been shown to affect hydrolysis rates. Lignin removal again improves cellulose digestion
by enzymes through opening up the structure and reducing non-productive cellulase
adsorption. The action of lime is slower than for ammonia or more expensive bases, but
its simplicity may make it applicable for pretreatment in piles [25].
3.23 Analytical Methods

In order to perform the analytical procedures both the untreated and treated
samples must be prepared. The samples treated with ARP, controlled pH and dilute acid
were washed before we received them. The wet samples received were dried at 45°C
overnight. Once the samples were dry, they were ground using a mortar and pestle and
sieved through a 140 —mesh screen with 106um openings. The fine powder obtained is
analyzed by the following techniques.
3.23.1 Diffuse Reflectance FT-IR (DRIFT)

The DRIFT results presented are performed in a Nicolet Protege 460 Magna IR
Technology equipment with an auxiliary experiment module for diffuse reflectance. The

signal was calibrated using a mirror. Spectra were obtained using 100 scans of the

powdered sample diluted with potassium nitrate (1:0.75), triangular apodization, a
resolution of 16 cm'1 and an interval of 1 cm]. The compartment was filled with
nitrogen to create an inert atmosphere. The sample is loaded in the cup without pressure,
the surface smoothed with a spatula and analyzed. The standard was ground to the same
size of the sample using a mortar and pestle and was mixed 0.75:1 just before analysis.
The equipment was calibrated using KBr for water and air background. Peaks are
identified following Stewart et al [58,59].
3.23.2 Enzymatic Hydrolysis

The digestibility of the treated and untreated corn stover was determined using
NREL Laboratory Analytical Procedure (LAP) - 009.. All the samples are hydrolyzed in
a pH 4.8 citrate buffer with the desired cellulase enzyme (provided by NREL, CAS 9012-
548, activity: 28 FPU/ml) at loadings of (60, 15, and 7.5 FPU/g of glucan) and a [3-
glucosidase from Sigma (St. Louis, M0) at 40 IU/g of glucan. All the samples are
hydrolyzed at 50°C with gentle rotation (75 RPM) for a period of 168 hrs. At
predetermined time intervals (0, 3, 6, 24, 48, 72, and 168 hrs), 1 mL of hydrolyzate is
taken for sugar analysis. Sugar analysis is performed using a BioRad (Richmond, CA)
High Performance Liquid Chromatograph (HPLC) unit equipped with an Aminex
carbohydrate analysis column HPX87P and a BioRad Deashing Cartridge guard column.

The mobile phase used is degassed HPLC water at a flow rate of 0.6mL/min at 85°C.

The injection volume used is 20uL with run time of 20 minutes.

 

° h_ttp://www.ott.doe.gov/biofuels/a_nalytical methodshtml

45

 

 

CL.
.. ,

I)

 

 

 

3.23.3 Simultaneous Saccharification and Fermentation (SSF)

SSF experiments were conducted according to NREL standard protocol LAP-
008'. Each ssr flask is loaded with 6% w/w glucan, 1% w/v yeast extract, 2% w/v
peptone, 0.05 M citrate buffer (pH 4.8), the appropriate amount of cellulase enzyme for
15 F PU/ g of glucan and the appropriate amount of Saccharomyces cerevisiae D5A
(provided by NREL) inocula (starting 0D. 0.5). The SSF flasks are equipped with water
traps to maintain the anaerobic condition and are incubated at 37°C with gentle rotation
(130 RPM) for 168 hrs.

At time intervals of 0, 3, 6, 24, 48, 72, 96 and 168 hrs, a 2 mL aliquot is removed
aseptically from each flask. The samples are centrifuged; and the supematants filtered
for sugar analysis by HPLC and ethanol analysis on a gas chromatograph (GC). At the
last time point, a sample from each SSF flask is streaked on an YPD (yeast-peptone-
dextrose) plate to check for any contamination.

The samples taken from fermentation at different time intervals are analyzed for
ethanol yield by a GC 17 Shirnadzu (Maryland, USA) unit. The injection temperature is
240°C and the detector temperature is 255°C. The carbowax column is first maintained at
80°C for 3 min then ramped up to 125°C in 6 min.

3.23.4 Acid Hydrolysis

The lignin content (soluble and insoluble) and carbohydrates in biomass were
obtained by following the NREL LAP - 002, 003 and 004. procedure. The treated
biomass was passed through a 40-mesh sieve. Both pretreated and untreated corn stover
along with the high purity sugars and the method verification standard are loaded in a test

tube to be hydrolyzed. Sulfirric acid (H2804) at a concentration of 72% is added. The

46

samples are hydrolyzed for exactly two hours and then diluted with water to obtain a final
concentration of 4% H2804. The diluted samples are autoclaved for 1hr at 121°C.

Once the samples reach room temperature they are vacuum filtered using a
filtering crucible. The supernatant is neutralized with calcium carbonate and filtered for
sugar analysis by HPLC. In addition, the supernatant is diluted and its absorbance
obtained at A=205nm using a Perkin Elmer Lambda 3A UV/V IS Spectrophotometer. The
solids left in the crucible are dried at 105°C, then weighed and bumed at 575°C to
determine the amount of insoluble lignin and ash in the sample, respectively.

3.3 Results and Discussion

The standard selected was potassium nitrate (KNO3). The selection was based on
its low toxicity and no reaction with the biomass samples and/or any element of biomass
pretreatment (i.e. acid, ammonia, etc). Another aspect to consider was the alteration of
the biomass spectra. This was modified by the dilution ratio. The final dilution (1g of
sample: 0.75 g of potassium nitrate) was selected because an increase in KNO3 would
offset the biomass signal and vice versa. Figure 3.2 shows the DRIFT spectrum of
KNO3. When analyzing this spectrum and comparing the standard peaks to the important
peaks found in biomass (Table 3.1) it was found that the KN03 peak at 2400 cm'1 does
not interfere with the key biomass peaks. The spectra of the pretreated and untreated
samples were normalized to this peak.

Figure 3.3 depicts the reproducibility of DRIFT analysis. The sample mixed with
the standard was divided in three parts. One part was loaded, analyzed by DRIFT and
spectrum recorded. The same procedure was repeated with the other two parts. The

spectra of the parts were recorded and are presented in figure 3.3. DRIFT shows a good

47

reproducibility with :l:12, $13 and i9 % difference in the aldehyde, C-H and lignin peaks

intensity, respectively. The whole spectra have a standard deviation (s) of 0.187.

Intensity (K-M)
W
I l
l
l
I l
I l
l l
l l
l
l
l
l
I
l
l
l
l
l
l

 

 

 

 

   

 

"7— l T I I I -r—T_-*_"T-’ TflI I _V—$——r [N

640 840 1040 1240 1440 1640 1840 2040 2240 2440 2640 2840 3040 3240 3440 3640

Wavenumber (cm!)

Figure 3.2: DRIFT spectrum of potassium nitrate

48

     

 

 

     

 

  
   

 

 

 

 

 

 

3.5 T— —— —f-—_._ ..____....-_. --_- an _- _- -, _ I ._
| ___ _.
l 1—60%, 1:1, 90°C-1I
3 i——i —-. — __|_60%,131,90.C-2 ,, .7 :7.7 ~
l L— 60%, 1:1, 90°03!
I
$2.5 i" ~-—
5 l
a? 2 i
5 l
.5
3 1.5 ~ “"
i
2 1 i
0.5
o J _ l f r ~-

 

  

 

 

640 8401040 1240144016401840204022402440264028403040324034403640
Wavenumber (cm-1)

Figure 33: DRIFT reproducibility

The analysis of the changes in the peak intensities was performed taking into
consideration the variation of the DRIFT technique. Figure 3.4 shows the effect of the
AF EX treatment temperature on the corn stover DRIFT spectrum. As seen in the plot the
treatment at 60°C, as compared to the untreated sample, shows a decrease in ester
carbonyl, OH and O-H peaks, while there is an increase in the aldehydic peak. In
addition, there is negligible change in the lignin and amorphous cellulose peaks. These
results imply hydrolysis of the hemicellulose.

When the pretreatment temperature is increased to 70°C it can be seen, as
compared to 60°C, that C-H, O-H and aldehydic peaks are increased. On the other hand,
ester carbonyl, lignin and amorphous cellulose peaks show negligible changes. In this

case increased temperature improve the breakage into smaller molecules. When

49

compared to the untreated sample, the 70°C treatment shows a decrease in ester carbonyl,
lignin and amorphous cellulose and an increase in the aldehyde peak, indicating some
delignification and hydrolysis of herrricellulose.

A pretreatment temperature of 80°C, as compared to 70°C, shows an increase in
amorphous cellulose, and O-H peaks. This implies that an increase in temperature affects
positively the breakage into smaller molecules and decrystallization. When compared
with the untreated sample, the 80°C treatment shows an increase in aldehydic and O-H
bonds and a decrease in ester carbonyl indicating depolymerization and hydrolysis of the
hemicellulose.

The pretreatment temperature of 90°C, as compared to 80°C, shows a decrease in
amorphous cellulose, aldehyde, ester carbonyl and O-H peaks. However, when
compared to the untreated sample it shows a decrease in amorphous cellulose, lignin,
ester carbonyl and O-H peaks, implying delignification and hydrolysis of hemicellulose.

Table 3.3 is a summary of the comparative changes in DRIFT results between the
different pretreatment temperatures and the untreated corn stover. The enzymatic
hydrolysis data indicate that the sample treated at 90°C has the highest glucan
conversion. According to this result a decrease in depolymerization of corn stover
macromolecules is observed by the decrease in the OH and O-H peaks intensity.
However, it seems that with an increase in temperature two factors have influence in the
glucan conversion. These two factors are the hydrolysis of the hemicellulose and
delignification of corn stover. In other words, certain degree of depolymerization of
cellulose, hemicellulose and lignin, is enough to improve conversion, beyond, which

further depolymerization is not effective. Apparently, other factors, perhaps surface area

50

or pore size distribution are affected enough by the 90°C treatment vs. 80°C to more than

compensate for the less favorable trends in these peak intensities.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

'4 T ‘ on "__——{—60%, 1:1,60CTTT‘T ”Tiff!“ ”T

4 -—60%, 1:1,70c 00c . i

12 TTT "w—‘i 60%,1:1,80C 1
_.60%,1,1’90C untm ' A l

. 10 T” UNTREATED,__soc__ .4
E l ‘aldehydic , l
2’ . l
'g 3 i— _______fl_rl 4”. '
3 l l
5 l
E 6 I eater ”“Ti
'2 cell carbonyl l
2 4 ‘ “AL__—‘_~ — __—___ ‘

 

 

 

 

2" “T" w M V
o 1 $1 1 r l r r

I I I l T

640 840 10401240 14401640 1840 2040 2240 2440 2640 2840 3040 3240 3440 3640
Wavenumber (cm!)

Figure 3.4: Effect of AF EX pretreatment temperature on the DRIFT spectra of corn
stover

Table 33: Changes in DRIFT results from changes in AF EX temperature

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Predomimnt 60°C 70°C 80°C 90°C
Compared to Compared to Compared to Compared to Compared to Compared to Compared to
Peaks untreated untreated 60°C untreated 70°C untreated 80°C
amorphous 0 - 0 0 + - -
cellulose
gain 0 - 0 0 0 - 0
aldehyde + + + + 0 +
ester carbonyl - - 0 - 0 - -
C-H - 0 + 0 0 0 0
O-H - 0 + + + -
Legend: + : increase
- : decrease
0 : no change

51

Figure 3.5 depicts the effect of ammonia loading during the AF EX pretreatment
of corn stover on the DRIFT spectra. When compared to untreated corn stover, a loading
of 0.5 g of ammonia per g of dry biomass (0.5:1) causes a decrease in amorphous
cellulose, ester carbonyl, C-H and O-H peaks at the same time that it shows an increase in
the aldehyde peak. This amount of ammonia in the AF EX pretreatment causes some
hydrolysis of the hemicellulose, but does not affect much the breakage into smaller
molecules, delignification and decrystallization of the biomass. Using more ammonia
(0.7: 1) decreases the amorphous cellulose and O-H peaks and has negligible effect on the
lignin, aldehyde, ester carbonyl, and OH peaks.

The use of a higher amount of ammonia (1:1), when compared to 0.7:1, gives an
increase in amorphous cellulose, aldehyde C-H and O-H peaks. In addition the spectrum
shows a decrease in the ester carbonyl peak. These results indicate hydrolysis of the
hemicellulose accompanied by some breakage into smaller molecules and some
decrystallization. An additional increase in ammonia loading (1 .3:1) decreases aldehyde,
C-H and O-H peaks. By enzymatic hydrolysis, the amount of ammonia used in the
AF EX pretreatment reaches an optimum at 1:1. Using 0.7:1 ammonia loading gives the
highest degree of delignification. However, at the 1:1 biomass to ammonia condition
there is some delignification and decrystallization (enough to improve hydrolysis) along
with depolymerization of the corn stover and hydrolysis of hemicellulose as compared to
the other ammonia loadings. According to the hydrolysis data the optimum ammonia
loading is 1:1 where highest depolymerization is observed and this agrees with the
previously stated optimum AF EX conditions [109]. The DRIFT changes caused by a

change in ammonia loading are summarized in table 3.4.

52

 

 

 

 

 

 

 

 

 

 

 

 

 

 

14 - —— M— - ——»—--~ __,_,. he 5“, ,__ _———_ if.
r Kori l‘ 60%, 05:1, 90 c on T.
i [{ —60%,0.7:l,90C I T
l l 60%,13:1, 90c A g
lignin ;— UNTREATED ,.\ "0 '
E 10 rh—-r—— - —* —- , ~ - ~ I!
9'5
a? s a» ——s —.--—
fl
3
fl
l— 4/
3 ‘1 amor
I
E ‘_ cell
0
Z 4 _
2 ‘ T ' w
«J I
0 f r 1 l k I 1 h r f I

 

 

640 840 IMIZNIWIMIWZMZZNMMMM3MMM
Wavenumbcr (cm")

Figure 3.5: Effect of ammonia loading on DRIFT spectra of corn stover.

Table 3.4: Changes in DRIFT results from changes in ammonia loading

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Predominant 0.5:] 0.7:] 1.0:1 1.3:1
Compared to Compared to Compm to Compared to Compared to Compared to Compared to
Peaks untreated untreated 0.5:1 untreated 0.7:1 untreated 1.0:1
amorphous - - - - + - -
cellulose
lignin 0 0 0 0 0 - -
aldehyde + + 0 + + + -
ester carbonyl - - 0 - - - +
011 - - 0 - + - -
O—H - - - - + - -
Legend: + : increase
- : decrease
0 : no change

The effect of moisture content on the DRIFT spectra of corn stover is presented in
figure 3.6. Pretreating corn stover with 20% moisture decreases the amorphous cellulose,

lignin, ester carbonyl C-H and O-H peaks as compared to the untreated sample, while

53

increasing the aldehyde peak. Thus, there is some delignification and hydrolysis of
hemicellulose.

When the pretreated corn stover with 40 % moisture is analyzed by DRIFT, a
decrease in amorphous cellulose and O-H peaks is observed as compared to 20%
moisture pretreated corn stover. An increase in moisture causes an increase in hydrolysis
of hemicellulose, and has little effect on the depolymerization, delignification and
decrystallization.

An increase in corn stover moisture content to 60% causes a decrease in O-H and
no effect in the OH, ester carbonyl, aldehyde, lignin and amorphous cellulose peaks, as
compared to 40 % moisture. This indicates that the presence of more water in the system
has negligible effect on the chemical structure of biomass. The hydrolysis data indicate
that the moisture content that gives the highest conversion is 60%. Apparently, other
factors, perhaps surface area or pore size distribution are affected enough by the 60%
moisture as compared to 40% moisture to more than compensate for the less favorable
trends in these peak intensities. The changes in DRIFT as a result of a change in corn

stover moisture content are summarized in table 3.5.

54

 

14 G“

. ~ ._2°%’ 1.3:1, 90 c on

. l l_ 40%, 1.3:1, 90 c $ .
'2 '“'— "‘ ““— " "l— ‘60%,13:1,90CTT"_T"T'T—T “r

 

 

   
  

l — UNTREATED

 

 

 

,E‘ T ‘ aldehydic / ,
E . ____,, ” * ' ' ' ,/ r l
l *' ll ~
.. . l __ __g J,“ _,________,g‘
i ______ll __ - _ _, r-_ * ~
Z ester
carbonyl

 

T fl I

l
0 v u
640 84010401240144016401840204022402440264028403040324034403640
Wavenumber (cm")

Figure 3.6: Effect of moisture content of corn stover on the DRIFT spectra.

55

Table 3.5: Changes in DRIFT results from corn stover moisture content

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Predominant 20% 40% 60%
Compared to Compared to Compared to Compared to Compared to
Peaks untreated untreated 20% untreated 40%
amorphous - - - - 0
cellulose
_li_gnin - - 0 - 0
aldehyde + + 0 + 0
ester carbonyl - - 0 - 0
OH - - 0 - 0
O-H - - - - -
Legend: + : increase
- : decrease
0 : no change

The effect of AF EX pretreatment time on the DRIFT spectra of corn stover is
presented in figure 3.7. The corn stover was heated to the target temperature and held at
that temperature for various times. When comparing the treatment for five (5) minutes

with the untreated sample it is found that there is a decrease in amorphous cellulose,

lignin, ester carbonyl, OH and O-H peaks. On the other hand there is an increase in the

aldehyde peak. This implies delignification of the biomass and hydrolysis of the

hemicellulose. The increase of the pretreatment time to 10 minutes has negligible effect

on the lignin, ester carbonyl and C-H peaks and increases the amorphous cellulose and O-

H peaks. This result agrees with the hydrolysis data that an increase in pretreatment time
beyond 5 minutes has little effect on the biomass conversion.

When the pretreatment time is increased to 15 minutes there is a decrease in

amorphous cellulose, lignin, C-H and O-H as compared to 10 minutes indicating increase

in delignification while also apparently hindering the breakage of the polymer into

56

smaller molecules and biomass decrystallization. Table 3.6 shows the changes in the

DRIFT as a result of a change in the AF EX pretreatment time.

 

 

Tl—60%,l:l,110C,5min * (m
l—6o°/., 1:1,110 C,10min
"——"— — *i—so°/.,1:r,rroc,rsmrni "—‘ “‘WL _. — —

 

 

 

 

 

 

 

 

 

 

 

 

 

Y I I j I

640 840 1040 1240 1440 1640 1840 2040 2240 2440 2640 2840 3040 3240 3440 3640
Wavenumber (cm")

Figure 3.7: Effect of pretreatment time on the DRIFT spectra of corn stover

57

Table 3.6: Changes in DRIFT results from AF EX pretreatment time

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Predominant 5 min 10 min 15 min
Compared to Compared to Compared to Compared to Compared to
Peaks untreated untreated 5 min untreated 10 min
amorphous - - + - -
cellulose
lignin - - 0 - -
aldehyde + + + - +
ester carbonyl - - 0 - 0
GB - - 0 - -
O-H - - + - -
Legend: + : increase
- : decrease

O : no change

Figure 3.8 shows the DRIFT results of ammonia recycle percolation pretreated

corn stover. When comparing the material pretreated for 10 minutes with the untreated
sample a decrease in O-H, ester carbonyl and lignin peaks is observed while there is an
increase in amorphous cellulose. As the pretreatment time increases from 10 to 20
minutes an increase in O-H, C-H, aldehyde and amorphous cellulose is observed, along
with a decrease in the lignin peak. A further increase in pretreatment time from 20 to 30
minutes causes a decrease in the O—H and OH, and an increase in the amorphous
rellulose and lignin. This indicates that ARP is effective in delignifying the biomass,
xdrolyzing hemicellulose and provides some decrystallization. According to these
:ults the optimum pretreatment time is 20 minutes. However, the hydrolysis data
icates that at these conditions a highest initial rate and the highest glucan conversion

)btained for this set of samples at 10 minutes.

58

 

 

 

 

y A 1—10%NH3,170C,10min - i 0'"
a A l—rov. NH3, 170 C, 20 min l ”min;
12 7"“ s—f * ”—1r “ii—10% NH3,170 (2,30 min h— "
1 —UNTREATED
L lignin I “h F“ k“ __‘—__—
0 +_ h_h ~ ._i , W W._ _ _ __ , v

 

 

e a

   
 

 

 

aldehydic

 

gter
carbonyl

 

 

Normalized Intensity (K-M)
as an
—> T
a I

1
l
i
I
l
‘P
=

 

 

 

 

 

N A
. . _ _+ .f

!

_1 r
«—

P :

l

l

I

l

I

0 s
640 84010401240144016401840204022402440264028403040324034403640
Wavennmber (cm")

Figure3.8: Effect of ARP in biomass as determined by DRIFT

In figure 3.9 the effect of dilute acid pretreatment on the DRIFT spectra is
observed. The spectra show a decrease in the amorphous cellulose, ester carbonyl and
aldehyde peaks for all pretreatment conditions. An increase in GR and OH peaks is
observed for all the conditions. These results indicate that dilute acid pretreatment is
effective in depolymerization and delignification of the biomass and hydrolysis of the
hemicellulose. It appears that the pretreatment conditions of time and temperature are
interchangeable, in other words it seems that an increase in temperature and decrease in
time is equivalent to a decrease in temperature and an increase in pretreatment time. This

tradeofi‘ (the basis of the severity factor) [124] has been independently proved by the

DRIFT results presented here.

59

 

 

18T77_i#~#’7ivT-TT '-’#_‘_'””‘"

 

  
 

 

 

 

    

 

 

 

r on I —140C,40min,l°/o H2804 ‘ a“
16 f— *_7‘\‘“——‘_1 ——l60 c, 10 min, 1% st04 l —*-- ~--—;
14 i f \ j 130 C,1min,l% ms04 *—2min ‘
her ”my. -.__ _-__‘ _____ rm. _ n -__._
5 fl 1 ~ ~rsoc,2min,1-/. ms04 S
l j J' ;.
512 a A ;~ Win—{fumnm'rnn _ _ - r— —— —v _
m _~~r*_*—f++s
g aldehydic
%
0
Z

 

 

 

 

1040 1240 1440 1640 1840 2040 2240 2440 2640 2840 3040 3240 3440 3640
Wavennmber (enf')

640840

Figure 3.9: Effect of dilute acid pretreatment on DRIFT results.

Figure 3.10 shows the DRIFT results for the controlled pH pretreatment. A
pretreatment at 160°C and 5 min and 200°C and 20 min both show an increase in the O-
H, C-H, amorphous cellulose and lignin peaks as compared to the untreated sample,
indicating breakage into smaller molecules, decrystallization and some delignification.
However, based on the enzymatic hydrolysis data, the conditions of 190°C and 15 min of
pretreatment are optimal, indicating that other factors not measured by DRIFT also
influenced the results. At these conditions a decrease in ester carbonyl and lignin peaks
are observed as compared to the untreated sample. These results agree with previous
knowledge about the pretreatment. For controlled pH pretreatment, the main effect on

the biomass to improve hydrolysis is the hemicellulose hydrolysis [123].

 

307 __ _______..-__ ._.___ _._.A , _.__ _ h. _ .h __ _ _a_ _

 
 

 

  

 

 

‘ C-H ;— 160 C, 5min y

f A g— 190 c, 15min

' __ , __;; —— 200 c, 20min a
[— mung}—

 

 

 

Normalized Intensity (K-M)

 

 

 

640 840 1040 1240 1440 1640 1840 2040 2240 2440 2640 2840 3040 3240 3440 3640
Wavennmber (cm' )

Figure 3.10: Effect of controlled pH pretreatment on the DRIFT results.

Figure 3.11 shows the DRIFT results of the different pretreatments as a way to
compare them. All the pretreatments show a decrease in the ester carbonyl peak
indicating that they are all effective in the hydrolysis of hemicellulose. Dilute acid and
controlled pH show an increase in the O-H peaks while AFEX, ARP and Lime
pretreatment show a decrease, as compared to the untreated corn stover. Dilute acid
pretreatment shows an increase in the C-H peak while the other pretreatment show
negligible effect. Thus, all the pretreatments show certain degree of depolymerization. It
is interesting to note that all the basic pretreatments show decrease in the O-H peak as
compared to the acidic pretreatments. All the pretreatments show a decrease in the ester

carbonyl peak as compared to the untreated sample, indicating different degrees of

61

hemicellulose hydrolysis. All the basic pretreatments in addition show a decrease in the
lignin peak while the acidic pretreatments show either negligible or apparent increase in
lignin. This indicates that the basic treatments affect the bonds present in the lignin
molecules more effectively while the acids are more effective in the breakage of bonds

present in the hemicellulose. The changes observed in the DRIFT are summarized in

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

table 3.7.
18 -,_h A * ~ -* ~—r . * “Wm—~— --— — ea
16 L”. ___...__ j —10% N113, 170°C,10min(ARP) I _ “WW
'1 1 180°C, 2 min, 1% st04 (Dilute Acid) l— l 1
14 ,____ _-.__ fi__ 4 ~ 190°C, 15min (Controlled pH) w _1\_____!
E l — Air, 55°C 4 months (Lime) . Dun“ Acid!
:3}; - _.. ”.1 —UNTREATED
€101 1 “gm # e
g l aldehydic
E 8 J ‘ _ #_____ __
.3 .
e 6~ '
° _. ,- . (ester
Z 4 _ :1 ll carbonyl
,_ ll
0 ‘F T r T j l A V T I I T

 

 

640 840 104012401440 16401840 2040 2240 2440 2640 2840 3040 3240 3440 3640
Wavenumber (cm")

Figure 3.11: Effect of various pretreatments on corn stover as determined by DRIFT

In general, DRIFT results tend to support previous understanding of these
pretreatments’ effects on biomass. Previous literature has related the changes in biomass

to the respective peaks in the DRIFT spectra [2,5,6,17,38,45] and these results confirm it.

62

Table 3.7 : Changes in DRIFT results from the different pretreatments

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Predominant Pretreatments compared to untreated
Peaks AF EX ARP Dilute Acid Controlled pH Lime
amorphous 0 + 0 - -
cellulose
lignin - - + 0 -
aldehyde + 0 0 0 0
ester carbonyl - - - - -
C-H 0 0 + 0 0
O-H - - + + -
Legend: + : increase
- : decrease
0 : no change

3.4 Conclusions

From the results presented here it can be concluded that each parameter (i.e.,
temperature, moisture content, time, etc) of the AF EX pretreatment affects the DRIFT
spectra. An increase in temperature causes a decrease in lignin and ester carbonyl; this
agrees with the hydrolysis data and previous knowledge that an increase in temperature
causes some dissolution of the lignin.

An increase in ammonia to biomass ratio caused an increase in OH and O-H
peaks. Ammonia in the AF EX pretreatment breaks ester bonds and generates smaller
molecules. There is a relation between the amount of ammonia and the bonds breakage.
As the amount of ammonia increases more bonds are broken until an optimum amount is
obtained and a further increase in ammonia causes no additional effect. An increase in
moisture causes a decrease in lignin and ester carbonyl peaks. The moisture in the

biomass allows formation of ammonium hydroxide that alters the hemicellulose and

63

lignin linkages. An increase in pretreatment time has little effect on the DRIFT spectra as
is observed in the hydrolysis data.

According to the DRIFT spectra ARP is effective in delignifying the biomass,
hydrolyzing hemicellulose and provides some decrystallization. Dilute acid pretreatment
is effective in depolymerization of the biomass and hydrolysis of the hemicellulose.
Lime pretreatment is effective in hydrolysis of hemicellulose and some delignification.
Controlled pH pretreatment’s main effect on the biomass is the hydrolysis of the
hemicellulose. All these observations agree with previous knowledge of the
pretreatments [25,47,58,70,79, 111, 123-125].

In conclusion, DRIFT spectra can be used as a powerful tool to identify the
effects of the pretreatments on biomass. DRIFT also can be used to monitor how the
pretreatment parameters (i.e., temperature, time) influence/affect the biomass and how
these effects are related to the hydrolysis data. In addition, these results confirm that a
change in one parameter (i.e temperature, pretreatment time) affects the entire spectrum

indicating an interrelation between the different components of the cell wall.

Chapter 4
Statistical correlation of spectroscopic analysis and enzymatic
hydrolysis of corn stover

4.1 Background

It is in the best interests of the biomass processing industry to understand biomass
structure and the effects of pretreatments on structure in order to minimize the cost of
both enzyme and pretreatment. The creation of an appropriate model may provide an
easy, fast and inexpensive way to predict ethanol/sugar yield based on structural changes
in biomass (corn stover) as a result of pretreatment. In addition, modeling will broaden
the knowledge of biomass structures and properties and the effects that pretreatment has
on these properties. This knowledge may help identify parameters thatcan be tailored in
order to obtain an optimum sugar conversion.

Modification of the structural features of the substrate, in this case corn stover,
enhance the hydrolysis rate [44]. The most generally cited factors affecting enzymatic
hydrolysis are: 1) cellulose crystallinity [1,7,10,30]. The degree of crystallinity of
cellulose is expressed in terms of the crystallinity index (CrI) as defrned by Segal et al.
[30], this is determined by the ratio of the crystalline peak to valley (amorphous region)
in the diffractogram based on a monoclinic structure of cellulose. 2) cellulose protection
by lignin [31,32,34,39]. Lignin is covalently bonded to polysaccharides in the intact plant
cell wall, thus reducing accessible surface area of cellulose. The mechanisms that
explain the protective effect of lignin against polysaccharides hydrolysis remain uncertain
although a number of factors; such as the degree and type of cross-linkage to
polysaccharide, the diversity of structures found in the lignin component and the

distribution of phenolic polymers through the cell wall are important. 3) hemicellulose

65

sheathing and degree of hemicellulose acetylation [12, 45, 46]. The bonds between lignin
and carbohydrates are predominantly ester-linked to arabinose side chains of
arabinoxylans. Xylans are extensively acetylated.

Many previous publications have attributed the limitation of enzyme hydrolysis to
be due mainly to one aspect (single variable correlation). Some investigators studied
lignin as the limiting component [4, 15, 31, 32, 44, 88, 100, 104, 108, 114, 120] while
others focused on the crystalline structure of cellulose as the main barrier for enzymatic
hydrolysis [1, 12, 30, 38, 51, 64, 79, 96, 97, 101, 102]. In addition, some investigators
considered the surface area or pore size [27, 45, 47, 75, 111] of the biomass the limiting
factor. It is important to note that the cell wall is a complete system and not a complex of
isolated fi‘actions. Hence, an analysis that studies and relates the change in biomass
composition to sugar yield taking into account all of the cell wall structures may be more
effective than single variable correlations.

Most of the procedures used nowadays for biomass analysis, such as enzymatic
hydrolysis and acid hydrolysis and Kjehdahl method are empirical in nature and make
determinations of the glucose conversion, lignin content, ethanol yield, etc difficult. Also
these analyses are time consuming and expensive, and therefore impractical for an
industrial setting. What is needed is a fast, inexpensive and easy to use analysis that can
provide detailed qualitative and quantitative information.

The analytical methods used in this study are easy to use, fast and inexpensive as
compared to enzyme hydrolysis. Diffuse reflectance Fourier transform infrared (DRIFT)
has been identified as a reliable method for monitoring structural changes in biomass [7].

Sample preparation is faster and sensitivity is often higher than is found with normal

66

Fourier transform infiared (FT -IR). Quantitative precision is good, so that the
composition of the sample can be determined in less time than is currently required for
standard analytical techniques. X-ray diffraction has also been used in the field for
decades to determine the biomass crystallinity. The determination of total lignin is
performed using fluorescence spectroscopy. When combined with increased knowledge
of cell wall composition and construction, autofluorescence is likely to make an
increasing contribution to cell wall analysis [118].

A statistical analysis will mathematically correlate these spectroscopic data to
experimental data (e. g. hydrolysis initial rate and yield). Multivariate analysis is widely
used in spectral analyses because it can be used to identify variable interactions and
reveal variables that strongly influence a process, thus providing a framework for
optimizing the process or function [22]. Multivariate analysis also allows the scientist to
relate and model different variables simultaneously.

Multivariate calibrations are useful in spectral analyses and can greatly improve
the precision and applicability of quantitative spectral analysis [18]. With multivariate
calibrations, empirical models are developed that relate the spectral data for multiple
samples to the known concentration of the samples. These empirical relationships can
then be used in multivariate prediction analyses of spectra of unknown samples to predict
their concentrations. Good pattern recognition and detection of relationships can help
generalize rules for predicting future values and outcomes based upon current known
patterns and relationships in the data

The multivariate methods are many and various, nearly all of them are used in

order to simplify and reduce the complexity of a problem. Several models could be

67

developed based on the frnal use of the model and the precision desired, however, in this
study two models are made using two different multivariate techniques; multiple linear
regression (MLR) and principal component regression (PCR). The best model will be
chosen based on accuracy and variance analysis. In other words, the best model would
be the one that predicts a sugar yield closest to the one determined by wet chemistry
(within the desired degree of precision).

The multiple linear regression (MLR) model assumes that the best way to
estimate the enzymatic hydrolysis yield of sugars in the sample is by finding a linear
combination of the variables that minimizes the errors in reproducing the concentration.
This model takes into account all the information available without discerning the
importance of each item of information. In other words, MLR is incapable to discern
between data in the spectra or noise. MLR is a general technique through which one can
analyze the relationship between a dependent variable and a set of independent or
predictor variables. The most important uses of the technique are: (1) to find the best
linear prediction equation and evaluate its accuracy; (2) to control for other confounding
variations in order to evaluate the contribution of a specific variable or set of variables;
(3) to find relations between variables and provide explanations for them; and (4) to
estimate population parameters and test hypotheses about the population.

Principal components are uncorrelated linear functions of the original variables.
They are not only not correlated but in fact independent. The pattern of vectors reveals
the nature of the relationship among the variables. The first principal component is the
normalized linear combination with maximum variance, also known as an eigenvector.

When all variations cannot be accounted for using one eigenvector, a second eigenvector

68

can be found that is perpendicular to the first and describes the maximum amount of
residual variation and so on. The principal components turn out to be the characteristic
vectors of the covariance matrix and give a new set of linearly combined measurements.
A detailed mathematical explanation of both models is presented elsewhere in chapter 2.

The Unscrambler® software helps in the statistical analysis with visual
capabilities. It presents a way to change the multivariate relationship from a complicated
set of equations into graphical presentation. Unscrambler® provides exploratory analysis,
multivariate regression analysis, prediction and validation.
4.2 Materials and Methods
4.2.1 Materials

The description of the materials used is presented in section 3.2. 1.
4.2.2 Experimental Equipment and Procedures

AF EX, ARP, Uncatalyzed Explosion/Dilute Acid, Controlled pH and Lime
pretreatments are explained in detail in section 3.2.2
4.2.3 Analytical Methods

In order to perform the analytical procedures both the untreated and treated
samples must be prepared. The samples treated with ARP, controlled pH and dilute acid
were washed before we received them. The wet samples received were dried at 45°C
overnight. Once the samples were dry, they were ground using a mortar and pestle and
sieved through a 140 -mesh screen with 106um openings. The fine powder obtained is

analyzed by the following techniques.

69

4.2.3.1 Fluorescence

Fluorescence spectra are recorded using a SPEX-3 F luorolog. Auto emission
spectra were obtained at an excitation wavelength of 350 nm with an interval of 0.5 nm.
The excitation and emission slits were set at 3 and 5 nm, respectively. The solid sample
holder is filled with powdered sample and is held in place with a quartz cover slip.
Sample was placed 45 ° from the incident beam. The mode of detection was front face.

4.2.3.2 Diffuse Reflectance FT-IR (DRIFT)

A detailed explanation is presented in section 3.2.3.1.

4.2.3.3 X-Ray Diffraction
Each spectrum is collected using 0-20 method in a Rigaku Rotaflex ZOOB

diffractometer at 45 kV and 100mA with slits size 0.5°,0.5°, 0.3° and O.45°, respectively.
Double-sided tape is put on a quartz slide, the powdered sample is loaded by pressing and
is smoothed over with another slide. The slide sample is placed vertically in the holder
and analyzed using the horizontal goneometer.

4.2.3.4 Enzymatic Hydrolysis

A detailed explanation is presented in section 3.2.3.2.
4.2.3.5 Simultaneous Saccharification and Fermentation (SSF)
A detailed explanation is presented in section 3.2.3.3.

4.2.3.6 Acid Hydrolysis

A detailed explanation is presented in section 3.2.3.4.

4.2.4 Statistical Analysis
The statistical analysis was performed using the software Unscrambler v8.0.

(CAMO Process, New Jersey) Spectroscopic data were entered along with their

70

 

respective hydrolysis data at 3 and 72 hrs in order to obtain both MLR and the principal
components regression coefficients. A model is created and from this a prediction for the
unknown concentration was obtained.

The steps to follow in multivariate modeling are to first study the raw data, then
decide if reprocessing is needed. Principal component analysis (PCA), can help identify
possible outliers and the quality of the spectral data along with its useful areas (peaks of
interest), and to examine the distribution of the samples (whether or not the samples can
be explained with a model). The next step is to calibrate the model. The different
calibration models used are MLR and PCR. Once the model is calibrated the outliers can
be identified. If any outlier is found and eliminated from the model, then a refinement of
such model is needed. The model is then validated using different methods available
including leverage correction, cross validation or test set method. After the validation,
regressions using the developed models are performed and the final model interpreted.
The statistics of the models are analyzed. Part of these statistics is analysis of variance
(ANOVA), which includes the standard deviation and degrees of freedom. The root
mean square error of prediction (RMSEP), bias and square error of prediction (SEP) are
the error to be expected in future predictions, the mean difference between the reference
and the data and the measure of the distribution of residuals, respectively. If the statistics
show a satisfactory value (bias, RMSEP and SEP ~ 0) the models are used to predict

values of the samples with unknown sugar yields.
The predictions of the unknown values were compared to the measured values.

The percentage of difference between them was calculated using the following equation:

71

predicted - measured
predicted + measured
2

%d1flerence =

 

x100 (4.1)

 

 

 

In order to evaluate the effectiveness of the prediction a hypothesis was made and
tested statistically. The Student’s t test or t test was used in order to test the hypothesis.
The hypothesis to be tested was whether the average predicted conversion value was
equal to the average measured conversion value. The two-sided hypotheses are:
H0: [3:110 vs. H1: 11in

The null hypothesis (Ho) is rejected if the predicted average conversion value (p) is too
far away fi'om the measured average conversion value (no) in either direction; that is if
the t is too large or too small. This is done with certain level of confidence (or). In this

case the selected confidence level was 95%. The rejection region is then defined as:

R:| t] 2 tan

Here to); is obtained from the t distribution table [126] with a degree of freedom equal to

N-l. The 1 value is defined as:

 

 

 

t = (l:- flo) (4.2)
An
where:
p predicted average conversion value
[lo measured average conversion value
N — 2
2 (X1 - X )
s standard deviation; 3 = '='

N

72

X predicted/measured glucan conversion mean

N number of samples

4.3 Results and Discussion

In order to characterize the samples several analytical methods were used that
have been proven to measure important characteristics of the biomass. These methods
are crystallinity, lignin content and changes in selected bonds present in biomass. The
results of these spectroscopic techniques are presented below. In chapter 3 a detailed
explanation of the DRIFT technique along with the resulting spectra were presented. The
data obtained from these spectra were processed by the Unscrambler® software as part of
the model.

The sample crystallinity was determined using X-Ray Diffraction (XRD) by the
Sega] et al. method presented earlier in chapter 2. Table 4.1 shows AF EX crystallinity
index (CrI) results. A complete list of AF EX results at different conditions is presented
in Appendix 4.

In general, AFEX treated samples show decrease in crystallinity as compared to
untreated corn stover. However, it is important to notice that the optimal AF EX
condition (highest sugar yield in enzymatic hydrolysis) shows the highest crystallinity
index among the treated samples. This implies that the crystallinity is not the only factor
affecting hydrolysis. It is also possible that ammonia-treated cellulose is assuming a

different crystalline form that is more enzyme accessible than native cellulose.

73

Table 4.]: AF EX Crystallinity Index

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Treatment Conditions Crystallinity
Index (%)

60%, 13:1, 60°C 25.95
60%, 1.3:], 70°C 29.82
60%, 1.3:], 80°C 26.50
60%, 1.3:], 90°C 26.98
60%, 1:1, 60°C 27.00
60%, 1:1, 70°C 23.15
60%, 1:1, 80°C 22.96
60%, 1:1, 90°C 36.29
60%, 0.7:], 60°C 20.09
60%, 0.7 :1, 70°C 24.00
60%, 0.7 :1, 80°C 22.81
60%, 0.7:], 90°C 31.94
40%, 13:1, 60°C 12.40
40%, 1.3:], 70°C 19.25
40%, 1.3:1,80°C 19.25
40%, 1.3:], 90°C 22.30
40%, 1:1, 60°C 23.45
40%, 1:1, 70°C 25.09
40%, 1:1, 80°C 13.71
40%, 1:1, 90°C 23.48
20%, 0.7 :1, 60°C 20.21
20%, 0.7:1, 70°C 23.07
20%, 0.7:], 80°C 23.61
20%, 0.7:], 90°C 16.77
Untreated 50.30
a-cellulose 66.53

 

 

Table 4.2 shows the CrI results for different pretreatments at their optimal
conditions. A complete list of the samples analyzed is in appendix 4. AFEX and ARP
are effective in decrystallizing cellulose. Controlled pH pretreatment also shows a
decreased CrI. Lime and dilute acid give an apparent increase in CrI as compared to the

untreated corn stover. These indices were also included in the models for the different

conditions and pretreatments.

74

Table 4.2: Crystallinity Index of Different Pretreatments

 

 

 

 

 

 

 

 

 

Treatments and Conditions Crystallinity Index (%)
Dilute Acid (180°C, 2 min, 1% HZSO4) 52.51
Controlled pH (190°C, 15 min) 44.52
Lime (Air, 55°C, 4 months) 56.17
AFEX (60%, 1:1, 90°C) 36.29
ARP (10% NH3, 170°C, 10 min) 25.98
Untreated 50.30
a-cellulose 66. 53

 

 

 

 

Fluorescence spectra show a direct relationship between the lignin content and the
area under the curve at 425nm. Figure 4.1 shows the fluorescence spectra for different
pretreatments. When the powdered sample is excited at 350nm, a peak at 425nm related
to lignin will appear. All the pretreatments except ARP show a decrease in this peak as

compared to that for untreated corn stover. However, it is well known that ARP is a
delignification process so this result for ARP may not be particularly noteworthy. It is
believed that since ARP extracts lignin some lignin might be re-deposited on the surface.
Fluorescence, a surface analysis method, might thereby show a higher lignin content.

The fluorescence spectra of all the samples were included in the models.

75

 

 

 

   
  

 

 

 

 

 

 

 

 

 

 

l— -ARP (10% N113, 170'c, 10 min) i I
450000 —— m —w 1w I_LI ~
/ I me ( Air, 55°C, 4 months)
400000 r , ~l - - - Untreated ~
Lune I
I -— - 0 0
350000 - I I Untreated fl “fl AFEX (60 /., 1.1,90‘C) a
A g I — - Dilute Acid (180°C, 2 min, 1%n2s04)
$00000 - l "“5 "f—M'ri C trolled H 190°C lSmin H-
Dilute Acid _ i:_ 0" p ( ’ ) ___l
Controlled pH 1
1
/ A __ 1
l

 

 

420

 

 

520 570

470 -1
Wavennmber (cm )

Figure 4.1: Fluorescence results for the different pretreatments

The data obtained fiom DRIFT, Fluorescence and XRD were input into the

statistical analysis software Unscrambler ® in order to obtain a model using multiple

linear regression. Because of the limitation that more samples than variables are needed

to create a MLR model, the intensity of selected peaks (presented in table 3.1) were used

from the DRIFT spectra along with the intensity of the fluorescence peak related to lignin

and the crystallinity index. These intensities were determined using a baseline correction.

In other words a new baseline was used for every peak in order to determine its intensity.

The results for both initial rate and conversion at 72 hrs are presented below. Figure 4.2

shows the MLR model for initial rate using only DRIFT data and using DRIFT,

Fluorescence and XRD data. The model using only the DRIFT data (a) gives a

76

correlation of R2 = 0.666 while the model using all the data gives a correlation of R2 =
0.807. It is evident that an increase in the amount of data in the model gives a better fit of
the data. When analyzing the coefficients of the models it is seen that the model using
the DRIFT data is more affected by the ester carbonyl content, related to the degree of
hemicellulose hydrolysis. A decrease in the ester carbonyl bonds will increase the initial
rate. The model using all the data is more affected by the aldehyde. A decrease in the
aldehyde bonds will cause an increase in the initial rate.

The model for the conversion at 72 hrs is presented in figure 4.3 using only the
DRIFT data and also using all the data. The model with only DRIFT data gives a
correlation of R2 = 0.700. Analyzing the coefficients, it is found that the parameter that
most influences the model using only DRIFT data is the lignin content. The model using
all the data gives a correlation of R2 = 0.757. In this case an increase in the amount of
data has little effect on the data fit. When using all the data the parameter that influences
most the model is the OH bond, a result of hydrolysis of hemicellulose. An increase in
this parameter will increase the glucan conversion at 72 hrs. The coefficients (B) for the
initial rate and 72 hrs conversion MLR model using all the spectroscopic data are
presented in the appendix 4; Table A4.3.

The differences observed between the initial rate model and the 72 hrs model may
be due to the fact that some parameters that have not been taken into account in this
model, (i.e. surface area, pore size) are more influential at the beginning of the hydrolysis

than at 72 hrs.

77

 

“1 ”5.“ rec “tn Y“

TX; —.C_m..9>:$.u CSU-imv 1.0.0.3051

 

 

i e... E. 3.

. at

1...va 9.6.9.0) CGrU =3":sz Daub-DOLE

 

Predicted Glucan Conversion (%)

b)

Predicted Glucan Conversion (%)

 

80 Predicted Y
Elements 43
Slope 0 666397
Offset 9 748908

Correlation 08 16332
RMSEC 11 79507
SEC 1 1.93466
8165 3 616e-06

q
C

8

 

 

8

8 8

..n
O

[11111111111111]ll[IlllIllllIIllIlllllllllllllllllllllllllll lllllllllllllllllll

 

....... I .I .. ..60_1-.100-15mn..

. I dander—40m
- , aUDUIl'IUTI-l)

...... -“60.1;90-'15mm /

. -80§Q§0%30mm , /

/,

 
   
 
 

601-1

‘ 80-1-100-10mtn
801-1005m1n ~

  

untreated-6, “97f ' We _ eat-120,5. . ..

 

 

 

 

 

 

 

0
Measured Y
rf"'_fi" 'I' 'I "I' ' I 'ffifi' I "' ' I I ' I" """ I ""'fifi'l
0 10 20 30 40 so so 70 so
Measured Glucan Conversion (%)
80 Predicted Y
Elements 43
Slope 0 809558
Offset 5565307
70 Correlation 0 898321 ........................ . ‘ . . . . _ . . . . .
PMSEC. 8972123 - Dal-‘W'mfigmflafim
SEC 9078305 _ . . - .
Bias -1 2756-07 ' . '
60 ................... ........... 3.8%911fl5mwmfl aubumlm— ......... . ......
50 ..................... owmcfimm . 50110.... .......................

8 8

'3

lllllllIllllllllllllllllfljlllllllllllllllllllllllllllllllll lllllllllllllllllll

 

_ . 3.... “ : ”1'0- .nwfl
- 601-80 2 013.7.4nggggeo

     
  
 
   
  
 

. _8°'lid°f.hlff59un° .
-90-10mln ;

.....6:0_.1._1.10_10mm3 .. .. . . .. .. ,.

   

 

 

 

 

1 . 40:” I’mgo
° aromas" '
0 untreated -
MeesutedY
Ifi'v 'I """"" I """ ' 'I‘ """" I """"" I'T ' T'I"" 'ITT ‘ "T'I
0 10 20 30 40 50 60 70 80

Measured Glucan Conversion (%)

Figure 4.2: Initial Rate MLR Model using a) DRIFT data only and; b) all the data.

78

ll

Aexev r..v_:..0>23.u :.~u..:v 39.9.3.1...»—

b)

 

 

v.1.
Ae\ev CC..WLO>:C.U Cuba-.0 30.0.32l

 

 

a)

Predicted Glucan Conversion (%)

b)

Predicted Glucan Conversion (%)

 

 

 

 

 

  

     

 

 

 

Predicted Y
160
d Elements 601-10015m1n
3 Slope 0 700206
1 Offset 66851
140 —. Correlanon 0136783
4 RMSEC 1807801
5 - 1839196 /
‘ 81215 -9 204efl7 .
120 . .
1 601.901ng
100 _I .............
‘ 80—1-90—10mln
80 —
_ . '602030030
60 _. . /_ 5041 33010430 .. I . . . . .........
7 . - 40.1.70 201370
41M 390 . .
j . 0‘5 so. Slidfieo
40 L ntreat <1CL 2 $38 3
a ' -~2 0
4 1:71:1th
q
20 g
o — :
Measured Y
I I l l I l I l l
o 20 40 so 80 100 120 140 160
Measured Glucan Conversion (%)
15° 23:13.? Y .3
l Slope 0756325 ‘ 691‘10915m'”
Offset 15 09258
_ Correlauon 0870228
16 26589
_ SEC 15 45839
Bras -4 957906

 

 

 

 

 

  

 
   
  

 

 

100 — 60‘1'11osm'r601419116tfi1'n
' “me RI‘LLSELlOmIn
' 10095W‘119015mm
- 201.190
_ / - edits—0030
o 7
- %‘1s%rt1.go
50 —l 01 390' $17.70 ~
n 61112171611.
'1 . -. 73331.80
.I untreated
OJ
Measured V
0 5'0 100 150

Measured Glucan Conversion (%)

Figure 4.3: MLR model for the 72 hrs glucan conversion using a) DRIFT data only and;

b) all

the data.

79

 

 

 

P160621

influen

 

 

 

Table 4.3 shows the predicted values for both the initial rate and the 72 hr glucan
conversion using the MLR model. When using the MLR model the initial rate
differences between measured and predicted values for AF EX are 16-73 %, while for the
72 hr data the differences are 10-26%. Thus, the model has a better fit to the 72 hr
conversion. Once again, the model does not include factors such as surface area and pore
size, which may have greater effect on the initial rate.

As can be seen by the results presented in the table this model could also be used
to predict the glucan conversion obtained using pretreatments other than AF EX. For the
different pretreatments the differences between measured and predicted values vary by
approximately 2-40% for the initial rate and 7-57% for the glucan conversion at 72 hrs.
This model predicts the initial rate obtained from the different pretreatments slightly
better than the conversion at 72 hrs. Apparently the parameters that are not taken into
consideration in this model have more influence at the 72 hrs conversion than on the
initial rate in ARP, dilute acid and controlled pH pretreatments.

It appears that for the AFEX process the initial rate of glucan conversion is also
influenced somewhat by the parameters not included in the model like surface area. This
result agrees with previous knowledge that AF EX creates more enzyme accessible
surface area in order to obtain higher yields [58]. For ARP, dilute acid and controlled pH
pretreatment it seems that the other parameters (i.e. surface area, pore size) have more

influence at the 72 hrs conversions.

80

 

 

 

 

 

 

 

 

 

 

 

 

 

«I. file 8.3 36 made amdm EEom doom :a eo=otcoo

EN— Nmen 5.3 we; 36m 36m EEfl doo— Into—.2280

edd— ooen 94% ~02 cod» 33% EN .vOmN: .x; dow— Eo< 835

2.3 8.3 mwde a; 8.3 Kg. En: .vOmN: .x: .0 2: 23.. 25.5

3.8 cod» node 5.2 2.0m node 5E on .8: £22 $2 my?

3.0m 3.8 mud: 8.0m mode mode 5.: on do: .mIZ $3 5?.

”I: 5.3 ”mew made 3.9 2.: EEoTer_-_ ” _-c\cow Km“?

8.: 5.2. 2.1:: ”mm: $.Nv 8.3 55903 T: Tease Xmm<

Sew deem meme 3.2 e52 3.2 Def—”Tees Xmm<

3.: 8.3. mmdn 2.3” 5.2 Seam Doe; 5.7958 can?

QWNN 3.5 have Suom am: me. am Don; ”M. Texan Xmm<
ooeocoEn was; 2: NB 3 «.2382 ex; me: Q. «a @2335 oococomme ex; e238: ex; 3862““

.x. :o_m._o>:ou 5620 2295250 5020 Ax. 23— 3:5 23. BEE 295m

 

 

 

 

 

 

 

 

28m 58 com .32: ~52 mama: sonogram and. 03.5.

81

 

 

Here 11
llhen
than 2
level.

11163511

Spec
model
the co
for 1h
alone
infom
only

repres

 

 

Table 4.4 shows the results of the hypothesis testing with a 95% confidence level.
Here the rejection region for both initial rate and 72 hrs conversion would be It] 2 2.306.
When observing the calculated M for both initial rate and 72 hrs conversion is smaller
than 2.306, hence the null hypothesis (Ho: para) is not rejected with a 95% confidence
level. In other words, there is no statistical difference between the predicted and the
measured conversions.

Table 4.4: Corn stover MLR data for hypothesis testing

 

MLR t tut,2
initial rate -00741 2.306
72 hr -o.4219 2.306

 

 

 

 

 

 

 

Table 4.5 shows the correlation (R) obtained using PCR regression and different
spectral inputs. In PCR every single data point from every spectrum was input into the
model. In table 4.4 we see that including all the analytical techniques actually decreases
the correlation for the initial rate, indicating that the model is over specified. However,
for the 72 hr conversion the same correlation is obtained using either the DRIFT data
alone or all the spectroscopic data available. Therefore DRIFT contains all the
information necessary to make an adequate predictive model. Hence, the model using
only DRIFT data will be used to make subsequent predictions. A graphical
representation of the models using the various spectroscopic techniques is presented in

Appendix 4.

82

 

 

 

 

 

 

I“. "T

f—p—I’Jf ‘— l"

daul
initial

hemjc

break;

90601

 

0111.1 21

the d5

C.

 

bonds
the p
belie

 

 

Table 4.5: Correlation using various spectroscopic techniques

 

 

 

 

 

 

 

 

Analytical Technique (s) Initial Rate 72 hrs Glucan Conversion
DRIFT R = 0.860 R = 0.903
Fluorescence R = 0.664 R = 0.906
XRD R = 0.691 R = 0.530
DRIFT + Fluorescence R = 0.374 R = 0.033
DRIFT + XRD R = 0.691 R = 0.529
Fluorescence + XRD R = 0.374 R = 0.033
DRIFT + XRD +Fluorescence R = 0.697 R = 0.903

 

 

 

 

 

Figure 4.4 shows the PCR model for predicting initial rate using only DRIFT
data. When analyzing the coefficients it is found that the parameter that most affects the
initial rate is the aldehyde content, representing the bonds between the lignin and
hemicellulose.

Figure 4.5 shows the 72 hrs glucan conversion model using the DRIFT data. The
Parameter that affects the 72 hr glucan conversion is the C-H content reflecting the
breakage of structural carbohydrate into smaller molecules. The most influential
Coefficients for the initial rate and 72-hrs conversion PCR models using the DRIFT data

Only are presented in appendix 4, table A4.4. It is very important to realize that given all
the data points input into the model, PCR is able to identify the wave number of those
bonds that have previously been identified as influential in the biomass hydrolysis. Thus
the PCR method has independently verified the importance of those bonds already

believed to most influence biomass hydrolysis.

83

nukes r-temflhoesdfuuu. h :EC..—u \ cynu-zmvsruhn-

 

 

 

 

Predicted Y

 

80 Elements. 4.3
Slope 0739750 ‘ ’ ' ‘ ‘ '
01'5“ 7505323 60110015m'"

N
O

C orrelatlon 0 860087

RMSEC 1041793 801 1011 tfiirnqurmem

 

 

SEC 1054122 _' I i ' I
60 Blas _75416_07 ......... ~aubquJV/
'6001-13011571011110/
50 80190-10m1 .
. 50,130 i - 80m5mln 30110610111'7'
40 ' °. 80190m060111015mm
601110onn

80140050010 .

 

8

llllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllll lllllllllllllllllllll

- 40- L800 3-90
1, M ~ 200 to..-

 

Predicted Glucan Conversion (%)

 

 

 

 

 

 

  
    

 

   

 

 

 

 

lag-L 07—90
untreated 5 111 £60
- - swm-.- ' go 11,, : - : 4
: 40121191134000 ; : r :
10 I ..... BUW'EB'G
/ 600780 ,
0 C
60—07-80
-10 .I ., .'
MeasuredY
I """"" I """"" I """"" I """"" III'IWII """""" I """"" I """"" I """"" I
.10 0 10 20 30 40 50 50 70 80
Measured Glucan Conversion (%)
Figure 4.4: PCR model for initial rate usrng only DRIFT data.
PredictedY
140 _ Elements 43 ' ' i ‘ i
3 3.009 [13155.79 ' ' ‘ j - 601-100-15m1n
d Offset 11.41635 ' ' ' - 80—LlOO-15mm
2 Correlation 0903150 : ' ‘ ' ' /‘
10: RMSEC 1417509 ............... ..............
SEC 14 34284 g j , ~ 0. 1110111111
3 ‘ 8'85 53319417 : I - - 601-110-10m1n ° 80731.41
83 ~ ~ 4 . 00190-50111
1:: 100 T -‘ 601-90
o ‘ 1-100 10311.3. 901%
o— d ' In
g "1 , : j . _50_.1_ aim/"500141015m1n I
E 30.4 ..... rl4Ckl-2Q’tjlr11005mln ................
8 ‘ ' ' I . . ' '
.1 . 60—1- - 20-13—80
a 60 _. . )tyg0/204370c 6013-90. . . I
2 d éw3 570911187 81:1
0 ‘ BOIWMO ~ 401-90
B I ' 501“" 40180 1
' A
a 7 441.1451?
0 ‘ . 009160 ' ““"98‘9‘3 " 60'1 3’80
E 20— ...... .‘ (7’_90
- ~ 60-07-80
0 —1 ‘
MeesuredY
T I r I I I I I I I I I I I T I T I I T I
0 20 4o 60 80 100 120 140

Measured Glucan Conversion (%)

Figure 4.5: PCR model for glucan conversion at 72 hrs using only DRIFT data

84

-1

 

 

 

AFEX
late. 11
Both 1

and pr

 

used
differ
from
hrs. 1
the ini
the P
acid a

better

 

 

 

The prediction obtained using the PCR model is presented in table 4.6. For the
AF EX pretreatment, the difference between predicted and measured values of the initial
rate, ranges fiom 3-74%. The difference for the 72 hr conversion is between 8-71%.
Both the initial rate and 72 hr conversion therefore show a wide range between measured
and predicted values.

As can be seen by the results presented in the table, the PCR model can also be
used to predict the glucan conversion obtained using other pretreatments. For the
different pretreatments the difference between predicted and measured values, ranges
from approximately 7-77 % for the initial rate and 4-75% for the glucan conversion at 72
hrs. However, following closer examination we see that the model fails to predict either
the initial rate or the 72 hr conversion of the ARP pretreated material. On the other hand,
the PCR model seems to predict the initial rate and 72 hr conversion of both the dilute
acid and controlled pH pretreatments. According to these results the MLR model gives a

better prediction for the AF EX, ARP, dilute acid and controlled pH pretreatments.

85

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8s :14 4:4 :3. as. 3.8 58 .88 3.0358
340 NS» on; :8 8mm 8% sea .8: ma 8:880
NS. 8.8 8R 8% 8.? as: sea «88.x: .8: 22 225
89. 8:. £8 a: 8.9. 3% as .Emfié do: 3438
88. 8.3 Mae 8.: 8.9. on: €538§§0§ 22625
3% 83 SM: was” 2.8 New E8 .8: .mmzsé e2
:fi SQ 8.3: mes 8.8 :4: as 8 .8: .82 :2 ea.
:9 $8 and Rm MEN 8% Bernice;
anon 8:. .88 mg 8.8 EN 038.38%

5: N333 .88 em: 3:. e24 §m_8:-_u_e\8xma<
8.: ans can 8% on: $8 083.28%?

as .88 :fi 80 2.: a? vacancxemg
casemegwefieeoségmefieaeeon eefléeosaaiéeeeaa 6:88

so 5828596 8:880; .x. came? as: 35

 

 

 

.558 88 com .088 mom wig cowomeoem and 2:3.

86

 

 

Here t

When

lex e1.

 

H1835

4.4 C(

 

 

Table 4.7 shows the results of the hypothesis testing with a 95% confidence level.
Here the rejection region for both initial rate and 72 hrs conversion would be M 2 2.62.
When observing the calculated |t| for both initial rate and 72 hrs conversion is smaller
than 2.262, hence the null hypothesis (Ho: u=uo) is not rejected with a 95% confidence
level. In other words, there is no statistical difference between the predicted and the
measured conversions.

Table 4.7: Corn stover PCR data for hypothesis testing

 

PCR t tan
initial rate 0.03711 2.262
72 hr 0.4255 2.262

 

 

 

 

 

 

 

4.4 Conclusion

Changes in plant cell wall components can be monitored by spectroscopic
techniques and analyzed by statistical methods including MLR and PCR. In MLR an
increase in the amount of data used improves the correlation between spectral
information and enzymatic hydrolysis, both initial rate and 72 hr conversion. It is
important to notice that for the MLR modeling only certain peaks were selected for the
model due to the restriction that in order to create the model there must be more samples
than variables. A model can be developed using spectroscopic data only to predict the
glucan conversion. For the PCR model the DRIFT data seems to have all the information
necessary for the modeling since increasing the amount of data actually causes a decrease

in the correlation

87

 

 

rate 0
the li;
indica

main ‘

predic
PCR ;

invest

 

 

Both models (MLR and PCR) indicate that the factor that affects most the initial
rate of hydrolysis is the aldehyde content, which is directly related to the bonds between
the lignin and hemicellulose. However, for the 72-hr conversion the MLR model
indicates that lignin is the main factor affecting hydrolysis, while PCR indicates that the
main factor is C-H bond content, representing breakage into smaller molecules.

It was found that, in general, the MLR model gives better correlation, and better
prediction for initial rate and 72-hr conversion for the AF EX samples. In addition, the
PCR approach independently verified that the peaks thought to be important by previous

investigation were in fact the crucial ones affecting enzymatic hydrolysis.

88

 

 

 

 

Chapter 5
Statistical correlation of spectroscopic analysis and hydrolysis of poplar samples

5.1Background

Pretreatment of lignocellulosic biomass is necessary to obtain high sugar yields
by enzyme catalysis. However, the fundamental characteristics of biomass that limit its
enzymatic conversion are not clearly understood. A better fundamental understanding of
these factors would help improve pretreatment/hydrolysis systems.

In an effort to better understand these factors a series of hybrid poplar (Populus)
samples was prepared varying the characteristics that are thought to affect the hydrolysis
of biomass; e. g acetyl content, lignin content and crystallinity. Poplar has been studied as
a biomass feedstock in order to improve hydrolysis and tailor pretreatment conditions.
Poplar has been studied for effect of pore size [45], lignin content [24, 66], chemical
structure [7, 94] and crystallinity [24]

In addition to the spectroscopic techniques discussed previously, Raman
spectroscopy has been suggested [2] as a complementary technique in the
characterization of biomass. In Raman spectroscopy less polar bonds give greater
scattering. Hence, bonds that cannot be observed in DRIFT will be present in the Raman
spectra. Stewart et al. believe that there are enough differences between the kinds of
groups that are infrared active and those that are Raman active to make the techniques
complementary to each other [2]. However, these same differences might give a specific
enough Raman spectrum for the biomass samples that could be related to the changes in
the sample characteristics (e.g. lignin content, cellulose crystallinity, etc.)

An important advantage of Raman is that water does not cause interference and

the magnitude of the Raman shifts is independent of the wavelengths of excitation [8].

89

 

 

The
mean

COfch'

relatit
hydrc
dime]
keepi
functi
that v
relatit

not in

Costa

 

The analysis presented here attempts to relate the changes in the biomass structure as
measured by Raman, DRIFT, XRD and Fluorescence spectroscopy to the glucan
conversion by enzymatic hydrolysis using statistical analysis.

A previous statistical analysis was performed by Texas A&M University [24]
relating the crystallinity, acetylation and lignin content of the poplar samples to extent of
hydrolysis by a model. They used Table Curve3D TM software that only fits three-
dimensional data (two x-variables related to one y-variable) and created a model by
keeping the acetyl content as a constant. From this model they obtained an empirical
function that was inserted into Sigma Plot software and created a polynomial equation
that was then substituted in the four dimensional equation and obtained the acetyl content
relationship to the other biomass parameters. This model was complicated since it did
not involve a model that directly relates all the properties measured (e. g lignin, content,
crystallinity and acetyl content) to the glucan conversion but the ratio of variables was
used instead.

In the analysis presented here two models were developed based on multivariate
analysis. The Unscrambler ® software finds the relationship between the parameters
based on the variance of the parameters measured. An advantage of the software is the
ability to correlate multiple variables at the same time. The entire spectrum can be used
and analyzed for patterns recognition and correlation with hydrolysis data.

Multivariate analysis allows the scientist to relate and model different variables
simultaneously. In other words, multivariate statistics is a collection of powerfitl
mathematical tools that can be applied to chemical analysis when more than one

measurement is acquired for each sample [10]. Multivariate calibrations are useful in

90

 

 

spect
spect
that

anal}

desir.
gluca
\‘arial
into a

of inf

PTOjec
redllCt

Small

malhe

capab

lfllO g]

 

medic

 

 

spectral analyses and can greatly improve the precision and applicability of quantitative
spectral analysis [68]. With multivariate calibrations, empirical models are developed
that relate the spectral data for multiple samples to the known concentration of the
samples. These empirical relationships can then be used in multivariate prediction
analyses of spectra of unknown samples to predict their glucan conversions.

The selection of the model is based on the final use intended and the precision
desired. The multiple linear regression models assume that the best way to estimate the
glucan conversion (hydrolysis) of the sample is finding a linear combination of the
variables that minimizes the errors in reproducing the concentration. This model takes
into account all the information available without discerning the importance of each item
of information.

The principal component model uses the data and projects it into planes. These
projections enable study of the deviations in the data. Principal component regression
reduces the number of variables to be analyzed, discards the linear combinations with
small variances and studies only those combinations with large variances. A detailed
mathematical explanation of the models is presented elsewhere [40, 68, 110].

The Unscrambler® software helps in the statistical analysis with visual
capabilities. Unscrambler® provides the meanings to translate multivariate relationships
into graphical displays and permits exploratory analysis, multivariate regression analysis,

prediction and validation.

91

 

 

 

 

SLZEV

SJLI

were

perfia

tray

thrnc

SJLZ

SJLZ

SJL2

SJLZ

SJLI

SFSII

The

5.2.2

 

5.2 Materials and Methods

5.2.1 Materials

Hybrid poplar samples were obtained from Texas A&M University. The samples
were selectively delignified, deacetylated and decrystallized. Delignification was
performed using peracetic acid. Deacetylation was achieved using dilute KOH.
Decrystallization was obtained with ball milling. The samples were prepared in such a
way as to obtain a broad spectrum of crystallinities, acetyl contents and lignin contents.
A more detailed explanation of sample preparation is presented elsewhere [24].

5.2.2 Analytical Methods
5.2.2.1 Fluorescence

A detailed explanation is presented in section 4.2.3.1
5.2.2.2 Diffuse Reflectance FT-IR (DRIFT)

A detailed explanation is presented in section 3.2.3.1
5.2.2.3 X-Ray Diffraction

A detailed explanation is presented in section 4.2.3.3
5.2.2.4 Raman Spectroscopy

The Raman analysis is performed in a Hololab Series 1000 from Kaiser Optical
Systems, Inc. The sample is analyzed without removing it from the clear plastic bag.
The laser probe (Model HFPH-632.8nm) is placed in close contact to the sample and the
8amPle exposed to the laser 10 times for 5 seconds each time and the spectra collected.
5.2.2.5 Enzymatic Hydrolysis

A detailed explanation is presented in section 3.2.3.2

92

 

 

5.2.3

53R

recei‘
cryst;
var)~
into c
5.] s
Obti
indie
Chang

andt

 

5.2.3 Statistical Analysis

A detailed explanation is presented in section 4.2.4
5.3 Results and Discussion

Poplar samples with a range in crystallinity, acetyl and lignin content were
received from Texas A & M University dry and ground for 0, 3 or 6 days according to the
crystallinity desired. Presumably these samples were prepared in such a way as to only
vary one of these three characteristics at the time [24]. The spectra were analyzed taking
into consideration the variation obtained from the reproducibility of the spectrum. Figure
5 .1 shows the DRIFT results for samples with constant lignin content and crystallinity.
Obviously, a change in acetyl content causes changes in the whole DRIFT spectrum
indicating that all the characteristics in the biomass are correlated. There is a marked
change in the 1500-1800 cm'1 area. This area is related to the hydrolysis of hemicellulose

and the peaks related to the acetyl links are also prominent in this region.

93

 

 

Intensity (K-l“)

Const

area

ACCQ

Visib

 

 

 

   

 

 

: A - 2.9%

16 111 '—“— ' " " ’“"‘—-A-2.8%
: 1—A-2.5-/.
‘4 t—m—— m—‘“[:_A -o.4%_ 9““ ——

 

Intensit-y (K-M)
l
l
l
l
l

 

8 -.t____ ___._ __n_ ,V
6 — ~-aldchydic-~—“
4 d 1/
ester
2 / carbonyl

 

 

 

0 f T T f f T f T f

640 84010401240144016401840204022402440264028403040324034403640
Wavenumbcr (cm")

Figure 5.1: Effect of acetyl content on DRIFT spectra

Figure 5.2 shows the DRIFT spectra of samples with constant acetyl content and
constant crystallinity but varying lignin. Once again, the graph shows that a supposed
change in lignin content only, affects the whole spectrum. More drastic changes in the
area around 1500-1620 cm'1 are noted as compared to the rest of the spectrum.
According to Stewart et al. the peaks related to lignin in DRIFT are the ones at 1510 and

1595 cm", which are in this area. Noticeable changes in the OH and O-H peak are also

visible.

94

 

.26: 5.3:...

 

 

Intensity (K-M)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

20 7 ————- --— — — - —-—-—— --~~~7 __._____ — — —- ~--~— A F m -— -——— - ,_
l A t—L-26.3% O-H
,. l___ . . _. T . l—L-zss'I-nwrwaw M1 ___-,~‘
’ i—L-21.5% ‘
,6 ,_ _____L—L- 18.7% _ _~_;
‘ [ 1am w t
1. if __ . - t
I
l
12-—— i~ v??? v~ i—Vfié——— — ——W
10«»— _ - -__- #—— —~- a—AA -—— “—14
A aldehydic 263% j
8 UH 23.9% 1
5 amor lignin _ _*,,,A, 0'“ 21.5.4
[ ester 1
2 _. .____r__ __ _
o WEE—fl" w”? T T r : T 1 T 1 r r Y fi
640 840 1040 1240 1440 1640 1840 2040 2240 2440 2640 2840 3040 3240 3440 3640
Wave-umber (cm")

Figure 5.2: Effect of lignin content on the DRIFT spectra of poplar.

95

bcfor

Chan;

\.

cellul

0b" 1’

thata

l

l I

A24: 5.3—.8:—

 

 

 

Figure 5.3 shows the effect of crystallinity index on the DRIFT spectrum. As
before, a change in this one parameter affects the whole spectrum. The most notable
change is observed in the C-H and O-H peaks, indicating that decrystallization of the
cellulose is accompanied by breakage into smaller molecules. Some changes can also be

observed in the 1540-1640 cm'1 area. This area is noted for aromatic C-O stretch bonds

 

 

 

 

 

 

 

that are related to lignin.
on

24 T l—c-sa9% ’ m; 1

22 ..... _ —_——-_——i—-—-C-46.0% . __+__-__,

I :5: - 13.6% 1

20 T“ T ' l " ' W‘_7fli 1‘

18 L 4;

 

 

 

 

 

 

 

 

Intensity (K-M)
3 :3 '3 '5:
i T
l
1 z
‘_ t I
fi-rh
; l
l
l
l
(p '63
= 5": E-

 

 

 

 

'\ cell Its-i- __ ‘ _ .
q l __ , i aldehydic // _ ___. f;
,. ____ @711“ “W _____ // .. _ 1 ;
o e . . . W . . M

H

640 84010401240144016401840204022402440264028403040324034403640
Wavenumbcr (cm")

 

 

 

 

 

 

 

 

 

Figure 5.3: Effect of crystallinity on the DRIFT spectra

96

Michigan State University (MSU) also analyzed the samples for crystallinity in
order to corroborate the Texas A & M University (TAMU) results. Table 5.1 shows the
crystallinity index for some of the samples. A complete list of the crystallinity results
determined by MSU is in Appendix 5. These samples are supposed to have a constant
crystallinity index. The results show some variation from sample to sample but in
general the CrI measured at MSU are in the same range as those CrI values provided by

TAMU and the trend of decreasing CrI with increased milling time is also corroborated.

Table 5.1: Poplar crystallinity index

 

_ CrI (%) Measured by MSU CrI (%) Measured b TAMU
Sampb 0 days 3 days 6 days 0 days 3 days 6 days

 

 

DIJOl-DAOOO 47.7 24.5 15.6 53.9 34.8 24.8

 

DLOl-DAOIS 55.7 26.8 18.1 52.2 36.2 21.2

 

DLOl-DA035 52.7 22.5 18.0 55.3 37.0 20.0

 

DLOl-DA055 50.2 20.7 23.2 56.6 35.9 26.9

 

DLOl-DA150 55.5 21.2 15.1 62.6 37.2 21.4

 

 

 

 

 

 

 

 

 

In figure 5.4 the effect of acetyl content on the fluorescence results is presented.
These samples had constant lignin content and crystallinity. The peak at 425 cm'1 is
directly related to lignin content. As can be seen, a change in acetyl content causes little

change in this peak. There is, in addition, little change in the rest of the spectra.

97

 

600000:— — , #8 ———~ __u-uu 8* e —A-29‘/or ~ ~ a
l ' 1—A-2.s%
.. l—A-2.5% *
5000001“ ; “‘“*"‘“— u‘*—”*"‘ i A-0.4-/_.f‘”“"*t
! |
3400000— ~~—i—— —————#v— ———— ——~
l
smm:__m:._w_:_fifif:_-_--
5

 

 

 

 

0 P‘WT—r—T‘ “DP—T“ '_*_T— ”—7?" '—“—T "' #7— __.__.Y "'fi —_ ‘ '_——7—_“~—T'——

360380400420440460480500520540560580600
Wavenumbcr (cm")

Figure 5.4: Effect of acetyl content on fluorescence results

Figure 5.5 shows the effect of lignin content on the fluorescence. There are
notable differences between the samples at the 425 cm'1 peak. In addition there are more
differences between the samples in the rest of the spectra

Figure 5.6 shows the effect of crystallinity on the fluorescence spectra. At the
smaller CrI, the spectra do not change. However, an increase in crystallinity causes

decreased absorbance up to 450 cm'1 and then an increase after 450 cm".

98

Intensity (cps)

Intensity (cps)

 

1200000 I “—mmumL2: .__ -_..: _n-m

 

 

 

 

 

[ii-25.5%!
] (”L-23.17” 1
1000000 7'» *~ *___ —e n.» 8* #8 ~ L-17s%~-—~~w— a
23.1% .
‘ / l:—_.__ 43.4% *
17.3%
300000 , — - /

1 13.4%
A 25.5% 1
.l 7

 

 

 

 

370 390 410 430 450 470 490 510 530 550 570 590

Wavcnumber (cm")

Figure 5.5: Effect of lignin content on fluorescence results.

 

 

 

 

 

 

 

 

 

900000 -— M Ma
g—ZIAV 1
800m ._. fl. . .- _#__.r"r - ;-_- __-..:.~
_37,2-/. 1_C'62'6%1 [
i—C-37.2%' j
7000““ H ”'m' "—i ' "——‘ ‘7—C-21.4%i" *‘ ”‘1
T”—‘”—_” 1

600000 —

 

 

 

..... [Al _
ml/‘M

300000- -
2000001

100000 1—
I

 

 

 

 

 

 

 

 

 

 

 

 

 

o +__ wefmfiw f .

360 380 400 420 440 460 480 500 520 540 560 580 600

Wavenunher (cm")
Figure 5.6: Effect of crystallinity on the fluorescent results

99

It is believed that Raman spectroscopy might be used as a complement to the
DRIFT results [2]. The effect of acetyl content on Raman results is presented in figure
5.7. The peaks around 1400 cm'1 are directly related to crystalline cellulose. The peaks
at 2900 cm'1 are related to OH bonds in the biomass. The spectra are presented here to

demonstrate the information that can be obtained from the different analytical methods.

 

{—11-28%7 7
450000.»— #-.# *w .__.# +.._ .._—--_ .1—A-2.5%E:

/ 2g# 1 A - 1.9%

400000 14‘ 1- —————- ~~— - ~~ 1 ~-1 _A-0-9%_

 

 

Raman Shift (cm")
1
1
1
1
1
1
1
1
1
1
1

 

 

 

 

 

 

 

 

 

 

0 t"--"————_—r— -' ——r—"—" " r T

—'T_‘ . _ ___—.. . _,_ .

200 600 1000 1400 1800 2200 2600 3000 3400 3800
Wavcnumbcr (cm")

Figure 5.7 : Effect of acetyl content on the Raman spectra
Figure 5.8 shows the effect of lignin content on the Raman spectra. It can be seen

that a change in lignin content causes only a slight difference in the spectra. However,

the difference is not as pronounced as the one observed in figure 5.7.

100

  
      

{—73 - 25.5%‘_ -1 f 2
1— L - 23.1%! ‘
1.. L ' 1781/.

 

Raman Shifl (0111")

 

 

 

200 600 1000 1400 1800 2200 2600 3000 3400 3800
Wavcnnumbcr (cm")

Figure 5.8: Effect of lignin on the" Raman spectra.

Figure 5.9 shows the effect of crystallinity on the Raman spectra. The

features are not easily distinguished, but a change in the crystallinity does change the

intensity of the spectra.

101

25000~ In 111 ~74 1 J - , 1,- 1,1- ,_ 11
—C-6l.5%,

;—" C ' 50.5%l
. _ egg-9:41

 

 

Ramon sum (cm")

———- -fi —-- -—————Y — -—_—'-T --—-————v -' ——-——J -- w—- — ‘

200 600 1000 1400 1800 2200 2600 3000 3400 3800

Wave-umber (c.")

Figure 5.9: Effect of crystallinity on the Raman spectra.

The analyses of variance (AN OVA) correlations for the models of the initial rate
using multi linear regression are presented in table 5.2. The initial rate is considered as
the glucan conversion obtained at 1 hr by enzymatic hydrolysis using 5 FPU/ g glucan.
Different spectral information was used to create various models in order to obtain the
best model with the least effort in terms of sample analysis. Because of the limitation
that more samples than variables are needed to create a MLR model, the intensity of
selected peaks (presented in table 3.1) was used from the DRIFT spectia along with the
intensity of the fluorescence peak related to lignin, crystallinity index and selected peaks
0f the Raman spectra. These intensities were determined using a baseline correction. In

Other words a new baseline was used for every peak in order to determine its intensity.

102

The AN OVA correlations for the 72-hr hydrolysis MLR model are presented in
table 5.2. The 72-hr hydrolysis was taken as the glucan conversion in enzymatic
hydrolysis at 72 hrs using 5 F PU/g glucan.

Table 5.2: AN OVA correlations obtained using various techniques

 

 

 

 

 

 

 

 

 

 

 

 

 

Analytical Technique (8) Initial Rate 72 hr Glucan Conversion
DRIFT R2 = 0.746 R2 = 0.733
Raman R7 = 0.365 R2 = 0.697
DRIFT + Raman R2 = 0.778 R2 = 0.847
DRIFT + XRD R2 = 0.775 R2 = 0.740
DRIFT + Fluorescence R2 = 0.747 RT: 0.736
Raman + XRD R2 = 0.659 R2 = 0.702
Rarnan + Fluorescence R2 = 0.397 R2 = 0.698
DRIFT + Raman +XRD R2 = 0.806 R2 = 0.849
DRIFT + Raman + Fluorescence R2 = 0.778 R2 = 0.847
Raman + Fluorescence + XRD R7: 0.659 R2 = 0.704
DRIFT + Raman + Fluorescence + XRD R2 = 0.806 R2 = 0.842

 

 

 

 

The best correlation is obtained using the data obtained from DRIFT, Raman and
XRD techniques. Addition of the fluorescence data has little effect on the model for the
initial rate and decreases the correlation for the 72 hr conversion model. According to
these results, adding the fluorescence data seems to over fit the 72 hr model. In addition,
this result indicates that Raman can be used as a complement to the DRIFT spectra to

inlprove the model predictions, as previously surmised.

103

The MLR model for initial rate using DRIFT, Raman and XRD techniques is
presented in figure 5.10. When evaluating the regression coefficients it is found that the
factor that affects most the initial rate is the aldehyde bonds, which are the bonds between

lignin and hemicellulose.

 

 

 

   

 

 

 

 

PrsdlcthY
50 : Elements 115
: S . 11806760
: ffse 2888417
: Cormlallon [18117997
40.: MSE‘N 1.157176
1; : 9E1: 4074932 //
g) : 8155 51629.05 ,/"'//
.5 2 4/
E 30 1 DLU‘S-DAS‘S-DCS _Ig-1§ga_t5@rpcc~
> 5 0103.0 " . . - J-DCE
S ; - . 07.ch . DLIUDMSDCE.
U 3 ‘ DLOSDAOT-DCE.
5 20 _: ' pee DLIODASSDCE-
o _ .
3 :
0 E
'U :
8 1o —
.9. I
1: :
E :
0 —I
: o '
: EILOUITLC1A1fui17-DC[I
.10 _‘
MeosuredY
I I I T I I l I I I I
.5 0 5 10 15 20 25 30 35 40 45

Measured Glucan Conversion (%)
Figure 5.10: MLR model for the initial rate using DRIFT, Raman and XRD.

The MLR model for the 72hr glucan conversion using DRIFT, Raman and XRD
is given in figure 5.11. The coefficients analysis indicate that the parameter that has the
most influence on the 72 hr conversion is the aldehyde bonds, representing the bonds
bean lignin and hemicellulose. The coefficients (B) for the initial rate and 72 hrs
conversion MLR model using all the spectroscopic data are presented in the appendix 5;

Table A5.2.

 

 

 

 

 

     

 

 

 

PradmledY
120 Elements l 16
" Slope 0 849459 /
- Offset 1095303 ./
— Correlanon 0921650
100 _ RMSEC 9475786
r; _ SEC 9517255
0\ BlaS -8 4889-05
V
t:
.2 ‘
80 —
E
E _
o I
0 ~ . /
g 60 1 DUB-0506000110—
3 T // LO$DA035 D‘CI‘
. . l
"' 'l m -0 DLUMOSWCE? '
O ‘ DLO"DAOU‘DL‘ DLO’NSAEQEQ‘fiWb‘CQCU ‘ [lLUE—DAU730CO
B DLOZQAUEDQO " ‘ . DLUEADAI‘SODCU
._. 40 fl - v
.2 . » om: _ 50-0
E J flat/‘EELAOQDAlscLEUJJH
. ,1 1 , V n:
j /,/ ~ObLEI2E.bEAu9.IIJD79d?:At9 E 0
20 ‘ ,
, ,7/ . DLWW
,/
- ’ . 1 Al ‘ . "1
- DL‘BhRgAQJaiBtgéU
0 _ Dmmmnmocn
MeasuredY
I I I I I I I
0 20 40 60 80 100 120

Measured Glucan Conversion (%)

Figure 5.11: MLR model for the 72 hr conversion using DRIFT, Raman and XRD.

The model correlations for both initial rate and 72 hr conversion using principal

component regression (PCR) are presented in table 5.3. As with the MLR models

different spectroscopic data were used to relate the changes in the cell wall structure

(changes in bonds) to the enzymatic hydrolysis yield. However, in this case every single

data point from every spectrum was input into the model. The initial rate used here is

defined as the glucan conversion at 1 hr of the enzymatic hydrolysis using 5 FPU/g

glucan. The 72-hr hydrolysis was taken as the glucan conversion in enzymatic hydrolysis

at 72 hrs using 5 FPU/g glucan. It is important to note that the correlation analyzed here

105

is not the relation between the data and the initial rate but the relationship between the

variables in the model.

Table 5.3: Correlations obtained using various techniques for PCR

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Analytical Technique (3) Initial Rate 72 hr Glucan Conversion
DRIFT R = 0.887 R = 0.914
Raman R = 0.518 R = 0.830
XRD R=0.812 R= 0.742
Fluorescence R = 0.403 R = 0.523
DRIFT + Raman R = 0.518 R = 0.830
DRIFT+XRD R=0.813 R=0.844
DRIFT + Fluorescence R = 0.424 R = 0.692
Raman + XRD R = 0.740 R = 0.853
Raman + Fluorescence R = 0.648 R = 0.866
Fluorescence + XRD R = 0.812 R = 0.480
DRIFT + Raman +XRD R = 0.591 R = 0.853
DRIFT + Raman + Fluorescence R = 0.648 R = 0.834
Raman + Fluorescence + XRD R = 0.648 R = 0.834
DRIFT + XRD + Fluorescence R = 0.403 R = 0.692
DRIFT + Raman + Fluorescence + XRD R = 0.648 R = 0.872

 

 

 

 

The data presented above indicate that the model with only the DRIFT data gives

the best correlation for both initial rate and 72 hr hydrolysis. This is a useful finding

106

because it decreases the amount of analysis necessary in order to obtain a good
prediction.

The PCR model for the initial rate using DRIFT data only is presented in figure
5.12. When analyzing the coefficients of the PCR models, it is found that the factor that
most affects both the initial rate and the 72 hr conversion is the O-H bonds content. It is
important to mention that from all the data points (over 6000 points) input into the model
PCR recognizes the wavenumber of the O-H bond as influential to the model. The
content of O-H bonds is directly related to the breakage into smaller molecules. The
model for the 72 hr conversion is presented in figure 5.13. The coefficients for the initial

rate and 72 hrs conversion PCR model using DRIFT data are provided in the appendix 5;

 

 

 

 

  
 
 

 

 

 

 

 

Table A53.
50 Pradlcledv
: Elements
: Slope
: Offset .
: Forrelamm ;/ D
40 _‘ RMSEC
Q : sec /.
9, : Blas 711299.07 //.,,»/
I: — /~/
.2 : /.—
“ DLOlDAO-DCEI . -
E 30 1 ' Iflfifirfiturfi
> : 3. b
C 2 . '
8 : DL». 5 - DLIL’LDAISDCE-
: " gIQEOSDAW-Enifi
g 20 __ ." 7 DLIllDAE’JICE
o - _
3 :
0 E
'U :
2 10 —
.2 :
E i
o 3.
.10 _:
Measuredv
I I I I I I I T | I I
—5 0 5 10 15 20 25 30 35 40 45

Measured Glucan Conversion (%)

Figure 5.12: PCR model for initial rate using DRIFT data

107

Piadlcled Y

 

120 Elements, 1 13 .....
‘ Slope 0835099
‘ Offset 1206715

- Correlation 0.913837
100 ._ RMSEC 9694792
SEC. 9 737977
8185 1 6790-06

 

 

  
     
  
 

     

  
    

 

 

 

’3‘ 80 d ...................................................................... ' I,- ' ' H L '.“ ' '
E —1 .3715." 1'“; 1178.“. 8"”...‘43 “:0? 3”:
a —1 ‘N 2" .".-'!'_ ”ATS-0C3
o q . . 0,; :__‘! 0
g 7 I "- ‘3‘" J v DL03-DAOI50-DCO
60 _ . '31:: "HR-W! 33,.“ .. .... .' A0740
g . - DLoz- ' “"1 "a" o - DLo'- co
DL
0 1-LosDA007Dca
g - i - D - .OiAOSS-DCO - DL03-0A035-DCO : : ' .
2 40__ Ill! ‘DLOUI—DATSO-UCO ...........
O _. I . I Z I I '
"‘83 .
._ 20 _‘ ............................................................................
B -« ‘
a: I ;
0 — L
Moasude
r‘ 'W"'I"****"'I""""' ""7""I""""' "‘""'*I""""'1'*'fi—*' I ' ""1 """" I
10 20 30 40 50 60 70 80 90 100 110

Measured Glucan Conversion (%)

Figure 5.13: PCR model for the 72 hr conversion using DRIFT data only.

There are clear differences in the factors that most affect both the initial rate and
72 hr conversion according to the different models (MLR, PCR). The MLR model
indicates that both initial and 72 hr conversion are more affected by the aldehyde bonds
and the PCR models indicate that is O-H the factor that affects both conversions the most.
However, both agree that it is through depolymerization of the biomass that the
hydrolysis can be improved. Finding a way to break the bonds between various cell wall

components (lignin, cellulose and hemicellulose) will facilitate the enzymatic digestion.

108

The predictions using all the models are presented below. The predictions for the
MLR initial rate and 72 hr conversion using DRIFT, Raman and XRD data are presented
in table 5.4. The results show that the models vary in their predictive capacity, they
predict well some samples, but do not predict others. For the initial rate a difference from
5 — 68% in the predicted vs. measured value is found. On the other hand the predictions
for the 72 hr conversion show differences between 3 — 22%. In this case, as with the
AFEX treated sample model, it is observed that the model predicts better the conversion
at 72 hr than it does the initial rate. Once again, this model does not include parameters
that probably affect the initial rate, including surface area, pore size, etc. The predicted
72-hr conversion is over 100% indicating that the model must be modified in some way
so as to constrain prediction between the established limits (0-100%). Another possible
reason for observed discrepancies is that the MLR model may include some noise from

the spectra that is assumed to be meaningful data.

109

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

mom 8.8 8.8 8.2 8.8 8.8 eon-885-885
8.: 8.8 2.8 8.2. 8.2 8.8. 85-88555
8.0 8.8 8% 8.8 8.8 8.2 85885-235
8.8 $5 2.2 8.2 one 3.: m52o<5¢m5
:2 8.8 2:: 2.8 2.8 2.: 85525835
Se 8.8 8.8 8.2 8.2 8.: 85885.85
8. a 8.82. cm. s 8.8 8.2 8.: 85-88585
2: 8.8 8.8 8.8 8.2 2 .2 8582525
5.2 E .8 8.8 8.8 2. 8.8 85-88525
8.2 8.8 8.8 2.8m w? 88 85.85.55
8.: 8.8 :8 2.8 88 28 85885-85
856% g a; 8 a 85822 g E 8 a 8282 82.20% g 85.82 g 82.28
.X. £282.50 G826 £28250 5820 x 3mm 3%: 3mm REE 285m

 

 

.8305 8m E58 ~52 win: 533325 32m 03:.

110

The PCR prediction for the initial rate and 72 hr conversion using the DRIFT data
only models is presented in table 5.5. The same pattern is found here in that the models
predict very well some samples (< 5 % difference) but do not well predict others. The
initial rate informstion shows differences between measured and predicted values of
approximately 5 — 33%, while the prediction for the 72 hr hydrolysis differs from the
measured values by approximately 4 - 18%. The PCR model predicts the 72 hr
hydrolysis conversion better than the initial rate (based on correlation and prediction

values) confirming that there are some parameters that affect the initial rate that are not

included in this model.

111

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

88 8.8 882 8.8 8.8 8.8 85-85-25
8.8 8.8 8.8 2.8 8.2 8.8 85-85-85

8.0 8.8 8.8 8.2 8.8 8.2 8585-85

8.8 3.8 8.8 8.8 8.2 2.2 85-2585

88 8.8 8.8 8.” 2 .8 8.8 858 25-85
8.8 8.8 8.8 $8 8.2 8.: 8585-885

8.: 8.8 8.8 8.2 8.2 8.2 85-85885
8.: 8.8 8.8 8.8 8.2 8.2 858 2525
8.2 2.8 :8 8.8 88 28 85-8525

8.8 8.8 8.8 8.8 88 8.8 85-85885

8.: 8.8 8.8 :8 8.8 8.8 85-85885

080528 :88 E 8 a 85802 88 E 8 a 8888 855% A88 8882 A88 8825: 28am
o\o comm—OZHOU 5320 comm-5250 8020 X. 3mm 35%: 05 E2

 

 

838?»: 385885 .3 3.5308 98 232: M05 05 3 683825 88 355 5953 82.85580 "Wm e.g—ah

112

 

f0

CC

re.

31'

 

 

In general, PCR shows higher correlation for both the initial rate and 72hrs
conversion compared to MLR. This higher correlation could be due to the fact that the
PCR model does not take into consideration the spectral noise while MLR does not
discern between important data and spectral noise. In addition, the predicted values show
a smaller percent difference as compared to those measured by enzymatic hydrolysis
when predicted by PCR than by MLR.

Table 5.6 shows the results of the hypothesis testing with a 95% confidence level
for both MLR and PCR models. Here the rejection region for both initial rate and 72 hrs
conversion would be III 2 2.62. When observing the calculated |t| for both initial rate and
72 hrs conversion is smaller than 2.262, hence the null hypothesis (Ho: 11:1,10) is not
rejected with a 95% confidence level.

The same results are observed for both MLR and PCR model. The M for both
initial rate and 72 hrs conversion is smaller than 2.262. In other words, there is no

statistical difference between the predicted and the measured conversions.

Table 5.6: Poplar MLR and PCR data for hypothesis testing

 

MLR t 19,, PCR t 5,,
initial rate 0.55619 2.262 initial rate 0.33395 2.262
721:: 0.8218 2.262 72hr -0.1859 2.262

 

 

 

 

 

 

 

 

 

 

113

5.4 Conclusion

Statistical models predicting hydrolysis of poplar samples with a wide range of
crystallinity values, lignin content and acetyl content were developed. Different
spectroscopic techniques (DRIFT, XRD, Fluorescence and Raman), were used to track
structural changes in the cell wall as reflected in bond intensities. The spectroscopic data
were related to both the initial rate and 72-hr conversion obtained by enzymatic
hydrolysis.

Different spectroscopic data were selected to develop the models in order to
determine a simple (i.e., data from only one spectroscopic method) way to predict
conversion and determine the factor that most affects hydrolysis. Then, correlations for
each model containing different sets of spectroscopic data (DRIFT, XRD, DRIFT +XRD,
etc) were determined and predictions using the models were made in order to examine the
model accuracy.

This study has shown that it is possible to predict with some accuracy the
enzymatic hydrolysis conversion using only spectroscopic data. In general, the PCR
model gives not only better correlation, but also better predictions for initial rate and 72-
hr conversion. In addition, using DRIFT data only gives accurate enough results so as to
provide a reasonable predictive model. The MLR models overestimate the 72-hr
conversion, which might be due to the fact that MLR uses all the data available without
discerning whether it is relevant data for the model or just spectral noise.

When comparing the models for initial rate to the 72-hr conversion it is observed
in both MLR and PCR that the models for 72-hr conversion give better correlations and

predictions than the initial rate models. This may be due to the fact that only some of the

114

 

factors that are known to affect hydrolysis have been studied here. Effects of variations
in sample characteristics such as particle size, or surface area were not studied and are
also likely to be most pronounced at early stages of the hydrolysis. Further study is

needed to test this idea.

115

Chapter 6
Conclusion and Recommendations

AF EX has been widely studied using different feedstocks and ample literature is
available [32-34, 56, 58-59, 86, 109]. AFEX is the basis for the model because of our
expertise in this process. It has been shown that the changes produced by AF EX in
chemical bonds can be easily identified by DRIFT. Each parameter (i.e., temperature,
moisture content, time, etc) of the AFEX pretreatment affects the DRIFT spectra. It has
also been proven that for native lignocellulosics, altering one structural feature results in
substantial changes on other features.

DRIFT can be used to compare various pretreatments and their effect on
the biomass. According to the DRIFT spectra ARP is effective in delignifying the
biomass, hydrolyzing hemicellulose and provides some decrystallization. Lime is
effective in hydrolysis of hemicellulose (breaking ester carbonyl bonds). Dilute acid
pretreatment is effective in depolymerization and delignification of the biomass and
hydrolysis of the hemicellulose. DRIFT thereby independently proved validity of the
severity factor which indicates that time and temperature are interchangeable in acid
pretreatment. Controlled pH pretreatments main effect on the biomass is hydrolysis of
the hemicellulose. According to the DRIFT results AFEX is effective in hydrolysis of
the hemicellulose. AF EX also shows (based on DRIFT spectra) some delignification,
depolymerization and decrystallization.

Changes in the cell wall components can be monitored by spectroscopic
techniques. Models were developed using spectroscopic data only to predict the glucan

conversion. In this research the techniques, analysis and data handling have been widely

116

studied before in various biomass areas. However, this is the first time that all of these
factors were combined for corn stover. This combination of factors allowed us to obtain
the most information fi'om each analytical method and hydrolysis result.

Multivariate analysis was used to create models that relate the spectroscopic data
to enzymatic hydrolysis results. The pretreated corn stover models (MLR and PCR)
indicate that the factor that most affects the initial rate is the aldehyde content, i.e., the
bonds between the lignin and hemicellulose. However, for the 72 hr conversion the MLR
model indicates that lignin is the primary factor affecting hydrolysis, while PCR indicates
that C-H bonds, representing breaking into smaller molecules are what affect hydrolysis
the most. It is important to notice that the models based on AF EX pretreated materials
are able to predict with some degree of certainty the glucan conversion of other
pretreatments like dilute acid and controlled pH, indicating that these are broadly
applicable phenomena that we are studying here.

This study has shown that it is possible to predict with some degree of accuracy
the hydrolysis conversion with only spectroscopic data. It was found that, in general, the
PCR model gives not only better correlation, but also better predictions for initial rate and
72-hr conversion for poplar samples.) In addition, the PCR method has independently
verified the importance of those bonds (amorphous cellulose, lignin, aldehyde, ester
carbonyl, C-H and O-H) already believed to most influence biomass hydrolysis. On the
other hand, MLR gives a better correlation and prediction for the AFEX pretreated corn
stover. This difference in model applicability may be partly due to the nature of the
samples. Corn stover includes different parts of the plant (leaves, stalks, etc) while for

the poplar samples only the trunk of the tree is considered. In addition, poplar does not

117

contain the protein that is present in corn stover. This result indicates that for these
models to be more generally applied, more closely related biomass substrates may be
required. However, according to hypothesis testing the predicted value is not statistically
different fiom the measured value for both poplar and corn stover; MLR and PCR models
and initial rate and 72 hr conversion.

Comparing the initial rate to 72-hr conversion within the model it is observed in
both MLR and PCR that the models for 72-hr conversion give a better correlation and
prediction than the initial rate models. This is probably due to the fact that factors other
than lignin, biomass crystallinity and acetyl content affect enzymatic hydrolysis. Such
factors might include degree of polymerization, particle and pore size and surface area.
Hence a model including these parameters is probably necessary to provide greater

accuracy in prediction.

118

APPENDICES

119

APPENDIX 1
Ammonia MSDS

Ingredients

Cas: 7664-41-7

RTECS #: B00875000

Name: AMMONIA; (ANHYDROUS AMMONIA)
% low Wt: 99.5

% high Wt: 100.

OSHA PEL: 35 MG/M3;50 PPM

ACGIH TLV: 17 MG/M3;25 PPM

ACGIH STEL: 24 MG/M3;35 PPM

EPA Rpt Qty: 100 LBS

DOT Rpt Qty: 100 LBS

 

 

 

Health Hazards Data

Route Of Entry Inds - Inhalation: YES

Skin: YES

Ingestion: YES

Carcinogenicity Inds - NTP: NO

IARC: NO

OSHA: NO

Effects of Exposure: ACUTE: EYES: CONTACT MAY CAUSE CORROSION, PAIN,
REDNESS AND ULCERATION OF THE CORNEA, LENS AND CONJUNCTIVA.
SKIN: CONTACT CAN CAUSE FROSTBITE, FREEZE BURNS AND/OR
CHEMICAL BURNS, RESULTING IN SEVERE DERMAL DAMAGE.
INHALATION: GAS IS EXTREMELY IRRITATING TO MUCOUS MEMBRANES
AND LUNG TISSUE. COUGHING, CHEST PAIN, AND DIFFICULTY IN
BREATING MAY RESULT. PROLONGED EXPOSURE MAY RESULT IN
BRONCHITIS, PUL MONARY EDEMA, AND CHEMICAL PNEUMONITIS.
BREATHMG HIGH CONCENTRATIONS MAY RESULT IN DEATH.

INGESTION: EXTREMELY IRRITATING TO MUCOUS MEMBRANES CAUSING
VOMITING, NAUSEA AND BURNS.

CHRONIC: NO CHRONIC HEALTH EFFECTS HAVE (EFTS OF OVEREXP)

Signs And Symptoms Of Overexposure: HLTH HAZ: BEEN FOUND TO DATE.
Medical Cond Aggravated By Exposure: ADDITIONAL MEDICAL AND
TOXICOLOGICAL

INFORMATION: MAY AGGRAVATE PREXISTING PULMONARY, LUNG, OR
EYE CONDITIONS.

First Aid:

120

 

....lll 1| l

 

EYES: IMMED FLUSH W/LRG AMTS OF H*ZO FOR AT LEAST 15 MIN,
INCLUDING UNDER EYE LIDS. SEEK MED ATTN IMMED, PREFERABLY AN
OPHTHALMOLOGIST. SPEED & THOROUGHNESS IN RINSING EYES ARE
IMPORTANT TO AVOID PERM INJURY.

SKIN: IMMED FLUSH W/LRG AMT S OF TEPID WATER WHILE REMOVING
CLOTHING. THAW FROZEN CLOTHING BEFORE REMOVAL. IF A FREEZE
BURN HAS OCCURED, GET MED ATTN.

INHAL: REMOVE PROMPTLY TO FRESH AIR. IF BR EATHING HAS STOPPED,
APPLY ARTF RESP. APPLY OXYGEN AS SOON AS POSSIBLE. SEEK MED
ATTN IMMED. INGEST: DO NOT INDUCE VOMIT. RINSE MOUTH OUT
W/WATER. DRINK LARGE AMTS OF WATER/MILK. SEEK MED ATTN IMMED.

Handling and Disposal

 

 

Spill Release Procedures: REMOVE SOURCES OF HEAT OR IGNITION,
INCLUDING INTERNAL COMBUSTION ENGINES AND POWER TOOLS. KEEP
PEOPLE AWAY. STAY UPme AND WARN PEOPLE DOWNWIND OF
POSSIBLE EXPOSURE. WEAR NIOSH APPROVED SELF-CONTAIN ED
BREATHING APPARATUS IF CONDITION WARRANTS. CONSULT DOT
"EMERGENCY RESPONSE GUIDEBOOK"-GUIDE 15.

Waste Disposal Methods: ANHYDROUS AMMONIA WILL NOT LEAVE RESIDUE
WHEN SPILLED; NO CHEMICAL CLEAN-UP WILL BE REQUIRED.
VEGETATION, INSECTS, REPTILES, FISH AND SMALL MAMMALS
CONTACTED BY LIQUID AMMONIA AND/OR THE VAPOR CLOUD WILL
LIKELY DIE; POST-SPILL CONSERVATION MEASURES MAY BE REQUIRED.
Handling And Storage Precautions: STORE CYLINDERS & TANKS IN A WELL
VENTILATED AREA, AWAY FROM INCOMPATIBLE MATERIALS (I.E.
CHLORINE), SOURCES OF HEAT & IGNITION. EMPTY CONTAINERS MAY
CONTAIN RESIDUAL GAS AND CAN BE DANGEROUS. GROUND/ BOND ALL
LINES & EQUIPMENT USED FOR THE TRANSFER & STORAGE OF AMMONIA
GAS TO PREVENT STATIC SPARKS.

Other Precautions: DO NOT PRESS, CUT, WELD, BRAZE, SOLDER, DRILL,
GRIND/EXPOSE SUCH CONTRS TO HEAT, FLAMES, SPKS/OTHER SOURCES
OF IGNIT; THEY MAY EXPLODE & CAUSE INJURY/DEATH. CONSULT
COMPRESSED GAS ASSOC PUBLICATIONS: (G-2) "ANHYDROUS
AMMONIA"; (G-2.1) "AMERICAN NATIONAL STD SFTY REQS FOR STOR &
HNDLG OF ANHYDROUS AM MONIA. ANSI K61.l".

Fire and Explosion Hazard Information

 

 

Autoignition Temp: =651.1C, 1204.F

Lower Limits: 16.0%

Upper Limits: 25.0%

Extinguishing Media: WATER FOG IS BEST. (AMMONIA WILL REACT WITH
CARBON DIOXIDE TO FORM A DENSE WHITE CLOUD).

121

Fire Fighting Procedures: USE NIOSH APPROVED SCBA & FULL PROTECTIVE
EQUIPMENT (FPN). USE WATER SPRAY OR FOG TO KEEP FIRE-EXPLOSION
CONTAINERS COOL. DO NOT COMPLETELY EXTINGUISH FLAME UNLESS
GAS FLOW IS SHUT OFF! AMMONIA BURNS TO FORM OXIDES OF
NITROGEN.

Unusual Fire/Explosion Hazard: FLASH POINT: NOT FLAMMABLE UNDER
CONDITIONS TYPICALLY ENCOUNTERED. ALTHOUGH CLASSIFIED
NONFLAMMABLE, AMMONIA DOES HAVE AN EXPLOSIVE RANGE.
AMMONIA CAN BE A DANGEROUS FIRE AND EXPLOSION HAZARD WHEN
MIXED WITH AIR. NFPA CODE: H - 3; F - 1; R - 0.

 

 

Control Measures

 

Respiratory Protection: USE NIOSH APPROVED FULL FACE RESPIRATORY
PROTECTIVE EQUIPMENT WHEN CONCENTRATIONS OF GASEOUS
AMMONIA ARE GREATER THAN STEL. SCBA IS REQUIRED TO CONTAIN A
LIQUID LEAK, UPON ENTRY ITO BUILDINGS AND ENTR Y INTO
DESIGNATED CONFINED SPACE AREAS, OR IN ANY SITUATIONS WHERE
AIRBORNE CONENTRATIONS MAY EXCEED OCCUPATIONAL EXPOSURE
LIMITS.

Ventilation: PROVIDE ADEQ GEN & LOC EXHST VENT TO ATTAIN OCCUP
EXPOS LIMITS, TO PVNT FORMATION OF EXPLOSIVE ATM; & TO PVNT
FORMATION OF AN OXYGEN DEF ICIENT ATM, (SUPDAT)

Protective Gloves: NONPOROUS GLOVES.

Eye Protection: ANSI APPRVD CHEM WORK GOGGS & FULL LENGTH
F ACESHLD (FP N). AMMONIA IS (SUPDAT)

Other Protective Equipment: EYE WASH & DELUGE SHWR MTG ANSI DESIGN
CRIT (FP N). AMMONIA IS SEVERELY CORROSIVE TO EPIDERMAL TISSUE.
WEARING NONPOROUS CLOTHING: PANTS, SLEEVES & FOOTWEAR IS
RECOMMENDED PROTECTION AGAINST SKIN CONT.

Supplemental Safety and Health: VENT: PATICULARLY IN CONFINED SPACE
AREA. EYE PROT: EXTREMELY CORROSIVE TO MUCOSAL MEMBRANES
(EYES, NOSE, THROAT). REMOVE CONTACT LENSES AND WEAR CHEMICAL
GOGGLES. A FACE SHIELD IS ALSO ADVISED FOR ADDITIONAL SKIN
PROTECTION WHERE CONTACT WITH LIQUID OR VAPOR MAY OCCUR.

Physical/Chemical Properties

Boiling Point: =-33.3C, -28.F

B.P. Text: @ 1 ATMOSPHERE

Vapor Pres: 124 @ 68F

Vapor Density: 0.6(AIR=1)

Solubility in Water: 5 LG/ 100G @ 68F

Appearance and Odor: COLORLESS LIQUEFIED GAS; PUNGENT AND
EXTREMELY IRRITATING ODOR.

Percent Volatiles by Volume: 100%

122

Reactivity Data

 

 

Stability Indicator: YES

Materials To Avoid: ACIDS, STRONG OXIDIZING AGENTS, CHLORINE,
BROMINE, PENTAFLUORIDE, NITROGEN TRIFLUORIDE, MERCURY, SILVER
OXIDE, CALCIUM, AND CHLORIDES OF IRON. DO NOT USE COPPER, BRASS,
BRONZE, OR GALVANIZED STEEL IN AMMONIA SERVICE.

Hazardous Decomposition Products: AMMONIA AND OXIDES OF NITROGEN
(NITROGEN DIOXIDE, NITRIC OXIDE).

Hazardous Polymerization Indicator: NO

Conditions To Avoid Polymerization: WILL NOT OCCUR.

Toxicological Information

 

Ecological Information

 

 

MSDS Transport Information

 

 

Transport Information: DOT HAZARD CLASS: 2.2.

Regulatory Information

 

Sara Title 111 Information: ANHYDROUS AMMONIA: EPA SARA TITLE III
INFORMATION: SECTION 311/312 HAZARD CATEGORIZATION: ACUTE, FIRE,
PRESSURE. SARA HAZARDOUS SUBSTANCES: INGREDIENT: ANHYDROUS
AMONIA. CAS NO: 7664-41-7, % WT: 99.5- 100, SEC 313, SEC 302, RQ LB: 100
TPQ LB: 500. SEC 313 = TOXIC CHEMICAL, SECTION 313. SEC 302 =
EXTREMELY HAZARDOUS SUBSTANCES (EHS), SECTION 302. RQ:
REPORTABLE QUANTITY OF EHS. TPQ = THRESHOLD PLANN ING QUANTITY
OF EHS.

Other Information

 

Other Information: CHEMICAL FORMULA: NH*3. ANHYDROUS AMMONIA:
NFPA CODE: 3: HEALTH HAZARD (BLUE): CAN CAUSE INJURY DESPITE
MEDICAL TREATMENT. I: FLAMMABLILITY HAZARD (RED): IGNITES
AFTER CONSIDERABLE PREHEATING. O: REACTIVITY HAZARD (YELLOW):
NORMALLY STABLE. NOT REACTIVE WITH WATER. NONE: SPECIAL
NOTICE (WHITE): NONE LISTED.

HAZCOM Label

 

123

Product ID: ANHYDROUS AMMONIA, STCC# 4904210, UN 1005

Cage: 1F8L6

Company Name: COASTAL ST HELENS CHEMICAL

Street: 63149 COLUMBIA RIVER HWY

City: ST HELENS OR

Zipcode: 97051 .

Health Emergency Phone: 800-424-9300 (CHEMTREC)

Label Required IND: Y

Date Of Label Review: 09/28/1999

Status Code: A

Origination Code: F

Eye Protection IND: YES

Skin Protection IND: YES

Signal Word: DANGER

Respiratory Protection IND: YES

Health Hazard: Severe

Contact Hazard: Severe

Fire Hazard: Slight

Reactivity Hazard: None

Hazard And Precautions: CORROSIVE. ACUTE:

EYES: CONTACT MAY CAUSE CORROSION, PAIN, REDNESS AND
ULCERATION OF CORNEA, LENS AND CONJUNCTIVA.

SKIN: CONTACT CAN CAUSE FROSTBITE, FREEZE BURNS AND/OR
CHEMICAL BURNS, RESULTING IN SEVERE DERMAL DAMAGE.
INHALATION: GAS IS EXTREMELY IRRITATING TO MUCOUS MEMBRANES
AND LUNG TISSUE. COUGHING, CHEST PAIN, AND DIFFICULTY IN
BREATHING MAY RESULT. PROLONGED EXPOSURE MAY RESULT IN
BRONCHITIS, PULMONARY EDEMA, AND CHEMICAL PNEUMONITIS.
BREATHING HIGH CONCENTRATIONS MAY RESULT IN DEATH.

INGESTION: EXTREMELY IRRITATING TO MUCOUS MEMBRANES CAUSING
VOMITING, NAUSEA AND BURNS. CHRONIC: NO CHRONIC HEALTH
EFFECTS HAVE BEEN FOUND TO DATE.

124

APPENDIX 2
Safety Requirements and Equipment

Safety requirements are continuously reviewed. Maj or areas of concern are the
location of safety equipment, toxicity of the chemicals, high temperature, electrical and
mechanical equipment.

The characterization techniques in this research include high temperatures or
electrical apparatus. When working with the electrical equipment it is important to avoid
water near the plugs and connections and to not touch any hot surface. Whenever
needed, we use tweezers to work with hot objects. Work in the hood at all times is
necessary as well as protective gloves when handling these chemicals. One always wears
protective glasses, lab coat and gloves at all times when working in the lab. These are to
maintain safety and to avoid contact of chemicals with skin and eyes.

The laboratory is equipped with chemical hoods, fire extinguishers, chemical
safety showers, chemical spill kit and several eye wash stations. Other available
equipment include gloves, organic vapor masks, dust masks, face shields, protective
clothing and eye protection. Material safety data sheets (MSDS) have been obtained for
all chemicals used in the laboratory are placed at the entrance to a rapid access in case of
emergency. A list of people to contact and the protective equipment needed in the area
also are posted at the entrance to the laboratory. Specific emergency procedures are

placed near all the equipment to ease the shut down in case of emergency.

125

APPENDIX 3

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

DRIFT Results (Raw Data)
1% . +‘*C“7
: on — 20%, 0.7:1, 60 C
16 1* -- — —-— elk—- —---- *9 —20-/., 07:1, 70 C .— ‘w-rlfl- *---—- ~ 4
1 f ‘ \ — 20%, 07:1, 80 C
14 1' —— — —20%, 07:1, 90 C
5 li in
an oz ______. l_‘ g!___
E“
B 10 +— A - , 1a , z I’m“
3 MW
,5, .
3 8 z , . 71 ,1 1‘1 ..__
= m1 10 111
£111 * .. "11 #
4 _ \W .... V
2 1 Elsi: { “mm 1 V I
0 fi— T— i I r \1‘ %M I f n V l

 

640 840 1040 1240 1440 1640 1840 2040 2240 2440 2640 2840 3040 3240 3440 3640
Wavenumber (cm")

Figure A3.1: AF EX DRIFT results (20%, 0.7:], 60-90C)

126

Normalized Intensity (K-M)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 
  

I on ._ 20%, 1.3:1, «DC 0.11
16 1‘ I _\T “ll—20%, 1.3:], 70C *— Vi fl—_ '—
‘ _ O
4 lignin 1 20 /e, 1.3:1,80C
14 . ——— | 1—20'/., 1.3: 1,90C
12 T ‘ I K “aldehydic
10 '1— ———— __ ‘ T ’ T TX
1 V
8 _ _ ii *fz_—C-u—_
' m0
6 — ___ i v
I
{\w w"
2 \J J I carbonyl
0 ‘IL I 1? f $ T j 7 f

 

640 84010401240144016”1840204022402440364028403040324034403640

Wavenumber (cm )

127

Figure A3.2: AF EX DRIFT results (20%, 1.3:], 60-90C)

 

 

 

 

    
 

 

 

 

 

   

 

 

 

 

CH ‘ 3—40-/.,0.5:1,60C
, \ l— 40%, 0.5:1, 70C
12 47* ——[-- k —40%,0.5:l,80C
an / lignin 1—40°/.,0.s:1,9oc
aw —~— .
E l ‘aldehydic
2.1L“. __zw m z_2-
0
E
36
C
E
z"
2

 

 

 

 

 

 

 

640 840 1040 1240 I440 1640 1840 2040 2240 2440 2640 2840 3040 3240 3440 3640

Wavenumber (cm")

Figure A3.3: AF EX DRIFT results (40%, 0.5:], 60-90C)

128

lull in I— N
G

H

. ‘1“

$1

Normalized Intensity (K-M)

 

A c-..

  

 

¥—M%Jmam

 

 

 

  

 

 

 

 

 

 

 

 

8 { \ 1—4070, 1:1, 70C
[ j—40./e, 1:], 80C
6 I *}—w%ummm
4 lignin
l/l

 

 

 

”‘q
,5"

11

 

 

 

 

1

 

0‘

 

 

 

, 11,1

 

 

 

 

 

0

T I T— T T FF T T

640 840 1040 1240 I440 1640 1840 2040 2240 2440 2640 2840 3040 3240 3440 3640

Wavenumber (cm")

Figure A3.4: AF EX DRIFT results (40%, 1:1, 60-90C)

129

 

 

”1—40-/.,1.3:1,70C “—-

 

 
   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1 1—40-/.,1.3:1,80C 22* ____ _
lignin 1— 40%,13:1, 90 C
“fl 1, aldehydic
E 1111f ._ ____‘z_zz
E.
,5. - ___ mm *m‘ _
3
3
.5
'3 4J 11 fl 7— T T TIA T—
cc

2. --m_zw._.z~+-
z V

new —~ f—fl —— a"-

‘ carbonyl
1 T -———Q ~4# ‘/
0 i; f T r I 1 A

 

640 840 10401240144016401840 2040 2240 2440 2640 2840 3040 3240 3440 3640
Wave-umber (cm")

Figure A3.5: AF EX DRIFT results (40%, 1.3:], 70-9OC)

130

Normalized Intensity (K-M)

 

 

 

 

 

 

 

 

i ‘ I —60%, 05:1, 60C

1 A G" — 60%, 05:1, 70C 0"

f _ . . , | J
I l—60-/., 0.5:1, 90 C A 1
., - . 1

 

 

 

 

 

 

 

 

 

of—fi I I I I WWI? Hi I I I I r j—fi—

640 840 10401240144016401840 2040 2240 2440 2640 2840 3040 3240 3440 3640
Wavenumber (cm")

Figure A3.6: AFEX DRIFT results (60%, 05:1, 60-90C)

131

 

 

 

 

‘4 1 2—60%,0.7:l,60C ” on ‘
1

 

 

 

 

 

 

 

 

 

 

 

I C-H
‘ l— 60%, 0.7:], 70C
12 '—— - ~—-——-— --60‘/,0.7:l 80C flfi __ 23
i 7” 1 _ .' .’
1 60/o,0.7.l,90C
E 10 I § _ m Ii nin
5 l 1 ldhd'
3‘ . , a e y K
E 8+-— A v :1 K '4"
3 V
E
6 f 7 ifi____ Wfi—_Wfi
E Illfll'
E eel
2° 4 — —
ester
2 ‘— *#‘ ca
0 L I 1 T A?

 

 

640 840 10401240144016401840 2040 2240 2440 2640 2840 3040 3240 3440 3640
Wavenumber (cm")

Figure A3.7: AFEX DRIFT results (60%, 0.7:1, 60-90C)

I32

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

16 T—~ -—— ——— -—w —~- 3 --
,- 0" i— 60%, 1.3:1, 60C (:9
- 1—60'/. 1.3:1 70C ‘1
1_. [ 11 ‘ ,z__z_ _. 9 3 a A ,,
”T 1 3 1—60%,1.3:1,80c F— I
A 12 4! A - 1—60%,1.3:l,90C A _
5. 1 .
E Kaldehydic !
g . ___. ___ z._ _; fl
8
8
.5 .3.
g 1
e l i
Z 1
1
1

 

640 840 10401240144016401840 2040 2240 2440 2640 2840 3040 3240 3440 3640
Wavenumber (cm")

Figure A3.8: AFEX DRIFT results (60%, 1.3:], 60-90C)

133

 

Normalized Intensity (K-M)

12

10

0

 

 

 

 

 

 

 

    

 

 

 

 

 

 

 

 

 

 

A

#1 W’fi—Ffifi I 1 T f 1 7 r T r

640 840 1040 1240 1440 1640 1840 2040 2240 2440 2640 2840 3040 3240 3440 3640
Wavenumber (cm")

Figure A3.9: AFEX DRIFT results (80%, 1:1, 90C, 5-15 min)

134

A C'“ — 80%, 1:1, 90C, 5min o-n 1
f \ — 30%. 1:1, 90C, 10min l
— 80%, 1:1. 90C, 15min A
xaldehydic
-1___ _ _ r __ k“

 

'4 1"“ "wig—T” "T—sova 1:1, 100C, 5min ‘ 1
L A — 80%, 1:1, 100C, 10min 0.11 |
{ ~ { \ -— —80%, 1:1, 100C, 15min e—~——l——J#a
l

1:

BIIIOI’

A 1

.\\\1 _

W

 

 

12

 

i—a
e

  
   

 

 

 

 

GK
1

 

 

 

A

Normalized Intensity (K-M)

 

 

 

 

 

 

 

 

 

 

 

1

f I I I r

640 84010401240144016401840204022402440264028403040324034403640

Wavenumber (cm")

I I I I 7 I I "

Figure A3.l0: AFEX DRIFT results (80%, 1:1, 100C, 5-15 min)

135

Normalized Intensity (K-

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   

 

l6 7— ———«—e .
on 1— N30117A

14 4 — N30124C

I ’ ' N30128B

1 11 J1 - N30128C
12 l ,‘v " 1—N30312

{—NCONTROL
10 — 1 4— _. #m—
lignin
8 z__-... f w_i _27m__ __. __
111 1
6 _ 3»
cell
4 1 1 1 / -_
2 c} — w . IL..— _.___
0 I I I I I I I I I I f I I I f I
640 840 1040 1240 1440 640 1840 2040 2240 2440,2640 2840 3040 3240 3440 3640

Wave-umber (cm' )

Figure A3.ll: NREL samples DRIFT results

136

 

 

— Air, 55 C 4 monthsl

I 3" TAMU M-l IL

16 +— ———--— - 1|

1 CCAZ -— TAMU M-2 1

14 r~ - ———— —~~-—TAMU M—3 * ~m~t #1

—TAMU 1111-4 ‘

* fl .1 lignin — UNTREATED
. .. n 5

___.____ __ ___;' .1
{/41 .

 

 

 

 

 

 

:—
N

3
_____+,, _

1

q

 

 

 

@
.—.__1_+‘

'5 Os
._ H— g

 

Normalized Intensity (K-M)

 

 

 

 

 

 

 

 

 

640 840 10401240144016401840 2040 2240 2440 2640 2840 3040 3240 3440 3640
Wavenumber (em'l)

Figure A3.12: Lime treated samples DRIFT results

137

APPENDIX 4
Corn Stover Characterization Raw Data

Table A4.1: Crystallinity Index for AF EX pretreated corn stover

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Samples Treatment Crystallinity index

and conditions (%)
20%, 0.7:], 60°C 20.21
20%, 0.7:], 70°C 23.07
20%, 0.7:], 80°C 23.61
20%, 0.7:], 90°C 16.77
20%,1.3:1, 60°C 34.83
20%, 1.3:], 70°C 33.59
20%, 1.3:1,80°C 35.86
20%, 1.3:], 90°C 36.16
40%,0.5:1, 60°C 36.25
40%, 0.5:], 70°C 32.72
40%, 0.5:], 80°C 42.29
40%, 0.5:], 90°C 35.89

40%, 1:1, 60°C 23.45

40%, 1:1, 70°C 25.09

40%, 1:1, 80°C 13.71

40%, 1:1, 90°C 23.48
40%, 1.3:], 60°C 12.40
40%, 1.321, 70°C 19.25
40%, 1.3:1,80°C 19.25
40%, 1.3:], 90°C 22.30
60%, 0.5:], 60°C 48.83
60%, 0.521, 70°C 60.43
60%, 0.5:], 80°C 55.63
60%, 0.5:], 90°C 52.04
60%, 0.7:], 60°C 20.09
60%, 0.7:], 70°C 24.00
60%, 0.7:], 80°C 22.81
60%, 0.7:], 90°C 31.94

 

I38

 

Table A4.1 (cont’d)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Samples Treatment Crystallinity index
and conditions (%)
60%, 1:1, 60°C 27.00
60%, 1:1, 70°C 23.15
60%, 1:1, 80°C 22.96
60%, 1:1, 90°C 36.29

60%, 1.3:], 60°C 25.95
60%, 1.3:], 70°C 29.82
60%, 1.321, 80°C 26.50
60%, 1.3:], 90°C 26.98
60%, 1:1, 90°C, 15min 63.66
60%, 1:], 110°C, 5min 68.68
60%, 1:1, 110°C, 10min 48.93
60%,1:1,110°C, 15min 61.30
60%, 1:1, 100°C, 15min 50.67
80%, 1:1, 90°C, 5min 55.89
80%, 1:1, 90°C, 10min 47.84
80%, 1:1, 90°C, 15min 50.07
80%, 1:1, 100°C, 5min 57.18
80%, 1:1, 100°C, 10min 50.78
80%, 1:1, 100°C, 15min 69.73
Untreated 50.3

 

 

139

 

Table A4.2: Crystallinity Index for pretreated corn stover

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Samples Treatment Crystallinity index

and conditions (%)

ARP #1(3D)7-25-03 56.17
ARP #2(5D)7-25-03 48.40

ARP 10%NH3,170°C, 10 min 25.98
ARP 10%NH3,170°C, 20 min 42.07
ARP 10%NH3,170°C, 30 min 37.27
Dilute Acid, 140°C, 40 min, 1% H2804 39.94
Dilute Acid 160°C, 10 min, 1% H2804 45.79
Dilute Acid, 180°C, 1 min, 1% H2804 48.00
Dilute Acid, 180°C, 2 min, 1% H2804 52.51
Dilute Acid 190°C, 20ml/min, 0.05% H2804 64.09
Dilute Acid, 200°C, 20ml/min, H20 51.33
Controlled pH, 160°C, 5min 53.84
Controlled pH, 190°C,15 min 44.52
Controllede, 200°C, 20min 54.12
NREL-P030312 44.84
NREL-0301 17A 48.88
NREL-030124C 41 .86
NREL-030128B 51.73
NREL-030128C 49.44
NREL-CONTROL1&2 48.25

Lime (M-l) 43.09

Lime (M-2) 38.99

Lime (M-3) 54.41

Lime (M-4) 44.90

Lime-02 56.17

Lime-03 51.76

Untreated 50.3

a-cellulose 66.53

 

 

140

 

 

Table A4.3: MLR model coefficients for corn stover using all spectroscopic data

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Initial Rate 72-hr Conversion
B-coefficients B-coefficients
Intercept 5.077 37.823
CrI 0.648 0.645
Amorphous Cell (902.537) 5.431 -4.232
Ligning(1511.94) 3.085 26.34
Ligfii (1604.51) 6.501 -22.37
Aldehyde (1658.51) -12.995 2.94
Ester Carbonyl (1712.51) -7.038 -14.372
C-H (2915.89) 12.638 35.478
O-H (3417.3) -2.655 -11.032
Fluorescence (425.5) -5.79E-06 -5.64E-06

 

141

 

 

 

Table A4.4: PCR model coefficients for corn stover using DRIFT data

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I initial Rate

Amorphous Lignin Lignin Aldehyde Ester Carbonyl C-H O-H
Cell (902.54) 1511.94 1604.51 1658.51 1712.51 2915.89 3417.3

PC 01 (X-Vars + Interactions) 0.015 0.016 0.020 0.018 0.009 0.019 0.033
PC_02 (X-Vars + interactions) 0.030 0.049 0.094 0.087 -0.030 0.079 -0.004
PC_03 (X-Vars + interactions) 0.045 0.049 0.096 0.088 -0.016 0.073 -0.042

PC_O4 (X-Vars + interactions) 0.166 -0.032 -0.354 -0.314 0.001 0.215 0.028

PC_O5 (X-Vars + interactions) 0.149 0.107 -0.267 -0.291 0.131 0.229 0.042

PC_06 (X-Vars+ interactions) 0.029 0.140 —0.380 -0.514 0.227 0.325 0.135
pc_07 (X-Vars + Interactions) 0.045 .0290 —0.570 -0.874 -0.603 0.565 -0.275
PC_08 (X-Vars + interactions) 0.031 -0.258 -0.552 -0.877 -0.649 0.557 -0.237
PC_09(X-Vars+ interactions) 0.031 -0.116 -0.681 -1.259 -l.012 0.781 -0.133
PC_I 0 (X-Vars + interactions) 0.029 -0.001 -0.580 -1.3 16 -1.013 0.859 -0224
PC_II (X-Vars + interactions) 0.101 -0.003 -0.575 -l.354 -l.093 0.913 -0.103
PC_12 (X-Vars + Interactions) 0.165 —0.936 -0.158 -I .347 -1.052 0.624 -0.383
PC_l3 (X-Vars + interactions) 0.191 -0.953 -0. 170 -l.314 -l.078 0.579 -O.363
U’C_l4 (X-Vars + interactions) 0.204 -0.936 -0.165 -l.320 -1.103 0.581 -0417
PC_15 (X-Vars + interactions) 0.180 -0.932 -0. 172 -1.320 -1.094 0.551 -0.477
PC_I6 (X-Vars + interactions) 0.762 -0.550 -l.112 -l.83l -0.918 0.800 -2. 166
PC_17 (X-Vars + interactions) 0.858 -0.569 -l.l32 -l.897 -0.929 0.746 -2.395
WJ 8 (X-Vars + interactions) 0.676 0.282 -2.082 -2.000 -0924 1.282 -2.418
I PC J9 (x-Vars + interactions) 0.711 0.286 -2.118 -2.008 .0937 1.283 -2.366
[PC_20 (x-Vais + interactions) 0.697 0.026 -l.465 -1 .910 -0.894 1.597 -2505

72 hr conversion

Amorphous Lignin Lignin Aldehyde Ester Carbonyl C-H O-H
Cell (902.54) 1511.94 1604.51 1658.51 1712.51 2915.89 3417.3

[ Pc_01 (X-Vars + interactions) 0.020 0.021 0.027 0.025 0.012 0.025 0.044
[PC_02 (X-Vars + interactions) 0.066 0.124 0.282 0.292 -0.111 0.214 -0.072
[PC_03 (X-Vars + interactions) 0.058 0.124 0.281 0.294 -0.1 19 0.217 0052
[E34 (X-Vars + interactions) 0.222 0.014 .0294 -0.407 -0.095 0.410 0.043
QCAOS (X-Vars + interactions) 0.185 0.315 -O.108 -0.354 0.187 0.439 0.073
[P—C 06 (X-Vars + interactions) 0.118 0.333 0.170 -0.511 0.240 0.493 0.125
EC_07 (X-Vars + interactions) 0.144 -0.406 -0.327 -i.037 -1.187 0.906 ~0.580
[PC_08 (X-Vars + Interactions) 0.036 -0.147 -0.129 -0.852 -1.547 0.846 -0.279
I PC_09 (X-Vars + interactions) 0.036 -0.204 -0.097 -0.670 -1.402 0.756 -0.320
I PC J 0 (X-Vars + interactions) 0.032 0.055 0.124 -0.829 -1.405 0.933 -0.525
[Pc_i i (X-Vars + interactions) 0.197 0.049 0.083 -0.989 -l.589 1.055 -0.246

I PC_12 (X-Vars + interactions) 0.110 1.307 -0.611 -0.735 -l.644 1.444 0.131
I PC_I3 (X-Vars + interactions) -0.069 1.426 -0.569 -1.012 -1.462 1.763 -0.007
[Pc_i4 (X-Vars + interactions) 0.015 1.528 .0572 -1.022 -1.612 1.775 -0.338
IPC_15 (X-Vars + interactions) 0.327 1.467 -0.535 -0.880 4.719 2.165 0.444
[Pc_i6 (X-Vars + interactions) 0.630 1.666 -0.846 -l.098 -1627 2.294 -0.434
[P917 (X-Vars + interactions) 0.403 1.710 -0.695 -l.165 -l.601 2.421 0.106
[P—c_i 8 (X-Vars + interactions) 0.212 2.604 -1.541 -1021 -l.596 2.984 0.083
I PC_I9 (x-Vars + Interactions) -0.385 2.528 -l.574 -0.940 -1379 2.971 -0.811
[PC_20 (X-Vars + interactions) ~0.388 2.458 -I.297 -l.086 -1.367 3.055 -O.848

 

 

142

L..-

1
s.

 

Table A4.5: MLR Raw data for AF EX treated corn stover Unscrambler Modeling

 

AFEX
C I"

I"l
Conversion

72hr|oi

Corwusm

Aimqiious
Cellulose

Lisn'm

AW

Ester
Carbonyl

C—H

O-H

Fluoreaomcel

 

2007-60

14.582122

47.00513

20.2138

2.2658451

2.93079

4.3175

4.65591

1.87229

3.03679

7.10034

1 745705

 

2007-70

33.126671

56.58599

23.0699

3.4506159

4.01417

5.7517

6.027483

2.581855

4.57958

9.79861

1.19E+05

 

2007-80

25.445362

54.94281

23.6107

5.051744

6.20674

9.0665

9.436663

3.977695

6.38625

13.5359

I.32E+05

 

2007-90

21.086733

54.8974

16.7683

2.916136

3.74261

5.5083

5.934785

2.179224

4.03377

9.54688

1 .09E+05

 

20-1.3-60

11.683777

43.19607

34.8297

1.431939

2.53624

3.3974

3.745461

1 .332364

2.24579

6.30807

1 .04E-t05

 

20-I.3-70

17.91564

63.40086

33.5856

1.375549

2.59712

3.5547

4.032147

1 .242687

2.55593

6.54226

7.38E+04

 

20-l.3-80

17.388397

81.47572

35.8584

1.55826

2.8041

3.7548

4.420683

1.150971

2.8374

6.89432

7.38E+04

 

20-1.3-90

15.91172

79.0871

36. 1607

2.358652]

4. 18245

5.4158

6.222168

1.764136

4.32729

10.4928

8.88E+04

 

400.5-60

15.51636

36.2491

1.887974

3.55736

4.5403

4. 528491

1.813123

3.31089

8.71527

9. 75E+04

 

4005-70

15.41342

32. 7249

1.757002

3.33203

4.3656

4.450801

1 .967598

3.21083

8.42273

1 .32E+05

 

4005-80

14.21958

42.2892

1.584417

2.8802

3.8374

4.054025

1 .523967

2.66417

6.88091

1 .49E705

 

4005-90

22.00139

35.8913

2.021435

3.60647

4.8204

5.109474

1.860871

3.5486

9.44428

1.21E705

 

40-1-60

18.405565

39.18448

23.4486

3.81321

4.39733

6.1697

6.19516

3.179807

4.59913

9.22244

2.(X)E+05

 

40-1-70

15.125252

52.1017

25.0927

3.1541641

3.79291

5.1422

5.366601

2.391274

3.69382

7.72074

1.51E+05

 

40-1-80

18.83864

62.92634

13.7069

2.6958101

3.33414

4.9192

5.270827

1.88684

3.51647

7.97909

1.36E+05

 

40-1-90

24.230513

73.87222

23.4755

2.6025441

3.19674

4.2868

4.77509

1.66312

3.03787

6.64144

1.57E+05

 

40-1.3—60

12.071383

10.84921

12.4041

2.503684

3.31315

4.5856

4.873533

2.030616

3.28505

7.92866

1.50E'+05

 

40-I.3-70

12.145831

17.06903

19.245

3.0821011

3.79087

5.6071

6.218038

2.178183

4.2089

10.0236

1 .40E+05

 

40-l.3-80

16.553133

8.678007

[9.2458

1.507985

2.22815

3.4115

3.804906

1.119988

2.3075

6113884

1.55E+05

 

40-1.3-90

19.374275

18.80384

22.3003

2.0357831

2.98521

4.0513

4.619017

1.557951

2.64697

6.33175

1.51E-105

 

6005-60

in

15.41443

48.8292

3. 76339 1

4.52083

6. 1745

6.382354

3.026507

4.89144

9.61027

1.12E+05

 

6005-70

19.38564

60.4291

5.702043 1

7.07757

9.8659

9.826607

4.728756

7.3452

13.861

1 .20E+05

 

6005-80

17.43585

55.6256

4.4436698

5.30183

7.4482

7.315739

3.571665

5.88884

11.1434

1. 13E+05

 

6005-90

17.26768

52.0409

5.0191932

5.79123

7.6157

7.849838

3.900492

6.03569

12.0298

1.11E705

 

6007-60

9.2901602

8.766775

20.0911

2.137181

2.70186

3.8234

4.049492

1.721558

2.6471

5.8362

180137105

 

6007-70

1 1.342645

10.31331

23.9958

2179749

2.78636

3.9806

4.249293

1 . 732426

2.75594

6.5335

1.60E+05

 

6007-80

12.421634

10.05193

3.8063

2676317

3.36465

5.0199

5.166181

2.023615

3.29783

8.37356

1 .94E+05

 

6007-90

12.65335

4.874799

31.9423

4.4497809

5.74406

8.1“”

8.653449

3.7234

5.62681

13.3266

1 46805

 

60-1-60

5.4927187

50.57309

27.001

2.5393 159

3.28002

5.164

5.270546

2.031191

3.53181

8.15014

1.755105

 

60-1-70

13.307735

50.83123

23.1546

1.825696

2.12525

3.1521

3.306175

1.234235

2.45901

5.67974

1.54E+05

 

60-1-80

1 5. 763963

54.08817

2.9628

0.367737 1

0.94378

1.5698

1.774695

0.379584

1 .04779

3.14785

1 .56E+05

 

60-1-90

 

20.790194

89.73552

36.2858

3.904192

5.(X)832

6.7902

7.295038

2.778517

5.57527

10.8115

1.15E+05

 

60-1-90-15 °

46.519829

115.1998

63.6594

5. 1673241

6.09841

8.2859

8.565197

3.39686

7.17253

12.3832

1 .74E+06

 

i-iooismiLn960358

12.4265

50.6732

6.52811

8.28626

1 1.274

11.71844

4.304543

9.72407

16.5127

4.03E+04

 

60-1-1 iosmiri 42.672844

77.86687

68.6754

2.3673639

3.0177]

4.141

4. 532546

1.453167

3.58483

6.40021

1.20E+06

 

sol-iioiornil 47.429924

87.31892

489.339

3.6313729

4.67816

6.491 I

6. @345

2.271889

5.69845

10.2574

2.34E+06

 

50-1-110-15

47.950237

94.52048

61.2992

3.3744509

4.07211

5.3512

5.674253

2.002612

4.63488

8.70196

1.26E+06

 

60-1.3-60

20.084051

49.75323

25.9468

3.444277

4. 13033

5.8907

6. 152644

2.784384

4.48941

10.1888

1.75E+05

 

60-1.3-70

[6.509945

52. 19736

29.817

3.7347341

4.55295

6.7135

6.986624

2.858991

4.83853

10.7316

1.38E+05

 

60-1.3-80

21.381002

62.30127

26.5018

4.3410592

5.18761

7.5187

7.901489

3.123646

6.1008

12.871

1 .27E+05

 

60-l.3-90

23.682922

78.9688

26.9776

3.1078329

4.06281

5.3766

5.967174

2.320657

4.31855

9.23333

1.2015405

 

 

 

80-1-90-5min 45.923897

109.1964

55. 8939

4.8832259

5.93628

7.9672

8.230081

3.368195

6.53693

11.6453

7.26E+04

 

[80-1-90-10miq 43.437111

117.1937

47.8409

4.1162901

5.33559

7.0567

7.571356

2.980752

5.41611

10.1421

8.80E+04

 

@1-90-15nrir136777916

1 09.0775

50.0729

3.841012

4.77474

6.5254

6.841831

2.694047

4.90846

9.44807

8.49E+04

 

[80-1-100-5mirI 74.531647

94.97057

57.1772

2.34239 1

3.21754

4.4953

4.693123

1 .725395

3.43866

6. 80248

8.5913704

 

Bo-1-100-1an72971 161

95.87077

50. 7768

4.1628089

4.92708

6.8669

7.260736

2.871164

5.61197

10.9943

7.79E+04

 

lio-i-loo-ismii 77.267792

107.1232

69.7343

5.3936749

6. 16658

8.3967

8. 808738

3.42447

7.2015

13.9098

6.68E+04

 

[ merited [6.9943924

 

29.70606

 

25.112

 

2.9457941

 

 

3.67382

4.497

 

3.944827

 

3.069079

 

3.4408

 

7.20473

 

4.88E+06

 

143

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8882 23.: 882 888; 388 882 8:3. 883; 88.3. E E 21.6 fig
838.“. 8:88 8888 888.8 s86 88:. $82 :88; «88.3 E E 9-2825
8+8; 383 83; :84; 888 :88 88.38 8888; 58.8 E E 9.8%..
8+8? 88.: 88... 88.. RE: 88% 28: 3.88.” 888.9. E E 2.235
88:; En: 888 8388 3.2: :88 8:3 83.88 33.8 E E 8-2::
38:3. 2.8.8 888 8:8 888 8888 ”888 88.2 8.2.3 881.. 2888.2. EE88~ :8 8888
388.8 88.2 88% 83mm 888 N33.“ 888 888.... 5:3... 8.8 8882.8 E858 8888
8+m3._ 85.8 88.. 833.. 383 883 88:8 83.2 88.8 888.9. 8885.8 2808.38 8.388
3+8; E88 838 838.2 883 383 83m :88? 838.8. E E 1.98282
$82... 83.8 3.8.. 88:: 883 82.8; 8:8 328: 8.3.3. E E 282
3+8? 88: 883 8888 888 :88 8884. 583 8.88.8 E E 08.82
3.884. 88.» :88 :52 888 :23” 8:4. 582 2858 E E 8282
3+8? 88% 32mm 888.. 833 383m 88: 8882 88:. E E 9.882
3388 83.8.8 8:.“ ”383 8:3 N388 883. 883.. 88.8 E E <25?
3.83 88.: 88: .833. 883 :88 888 58% 88.8 E E E8~8~dN= 22 2:5
3+8? 8:8. 682 988; :83 $8.” :8; 388.. 888.3 E E EthdéaSsisE
388.8 88.: 888 5.88.“. 888... 3:2 $88 8883 988.8 8 8 8878222 2:5
8+m8._ 32% 2.888 888; 3:: 388 828.” 882.. 88.5. t. 9. 877836355
3+8; 88.2 88.8 588.8 88.8 28.8 88.8 8885. 88.8 3 8 E8778€w< 2:5
3+8; R85 883 5838 :58 28; 283 888.8 838.8 8 88 88-332285
8888 83.: 838 3.5.8 :88 38% 888. 522.... 88.8 3 888.8 988...
8+8; 835: 88.2 8288 8:8 238 88»... 8883 E88 883.8 888.8 @887.
838.. 3.2 888.... 888.. 89.3. 88+ 883 8283 58.8 388.3 E838 2.8%
838.» 88.: 322 8:3; 3.83. 883.2. 8:2 338 833.8. E E 8388...
83:8 282 52; 8;: 882 8:2 88% :83 8.28 E E €958?
tar—0980 823:an comm—0250 comm—02.50 mCOEUGOU
38828: :d :6 228 8882 E8: E8: 89.85 :0 E 8 =85 E8888

 

mam—oven. 83889.83 80>on 88 688.505 8m 83 3am ~52 6.3. 03:.

 

Intensity (cps)

_ fl
6 g g
1 J 1

 

200000 ~———* ff V- v~ if ~—— —-——+ e a

l I

160000 ~——- - “—~—*** 1“~--—~{—20°/.,0:7:1,90c~-— ~ 4
i
140000 y -— ———— w_—-*——- ———-—— M «w ~

! — 20%,0.7:1,60C
f” If #V___ —20'/o,0.7:l,70C _._H A .%

 

 

 

 

 
 

 

 

 

 

 

 

 

 

400 420 440 460 480 500 520 540 560 580 600

Wavenumber (and)

Figure A4.1: AF EX fluorescence results (20%, 0.7:1, 60-9OC)

145

Intensity (cps)

   
   

400 420 440 460 480 500 520 540 560 580 600
Wavenumber (cm'l)

-r__— _ _._

 

* ‘7—20°/.,0.7:1,60c
i 1—20%,0.7:1,70c - Hi _ l
[—20%,0.7:l,80C .

 

* —j— 20%,0.7:1,90Cr~ -—---**

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure A4.1: AFEX fluorescence results (20%, 0.7:1, 60-90C)

145

Intensity (cps)

 

 

 

 

  
 

 

— 20%,].3:1,60C

_ H ,‘rv v "i ’fi—20%9103:197w

 

— 20%,].3:1,80C

 

 

{— 20%,1.3a,900

 

 

 

 

 

 

 

 

 

 

400 420 440 460 480 500

520 540 560
Wavenumber (cm")

 

Figure A4.2: AF EX fluorescence results (20%, 1.3:], 60-90C)

146

Intensity (cps)

180000 ——— #~*~ A~AA~~~—~—rgi~ ~—~ *—~ + ——‘—»— —‘-

160000~

140000 ~*

120000 ~

 

 

 

 
 
 
 

.____ - *A 7;— — ‘—“ ‘ —— !‘ “"i _ 40%,0-5:1970C ‘1

f — 40%,0.5:l,80C }

 

 

 

 

 

 

 

 

 

 

 

 

0 ‘l—“‘" fi-q f T

 

l— 40%,0.5:l,60C L

A A** ‘ i_— 40%,0.5:1,90(:f““* ;

— 774

 

 

 

400 420 440 460 480 500 520 540 560 580

Wavenumber (cm")

Figure A4.3: AF EX fluorescence results (40%, 0.5: 1, 60-90C)

147

Intensity (cps)

 

 

 

1140%,1:1,60_E!
l P—40-/.,1:1,70c I
200000 l —— w #_+___ ——»— 1—40°/.,1:1,80c —-~ ~-—* l
l [7—40°/.,1:1,90c *
l \
150000 — —— —— ———— N~e ~ ‘ 4
100000 ——~ e~ # -————~———r~——— ———

 

t ‘\‘_
l
0 + , s - . T f 1 - , , r T
400 420 440 460 480 500 520 540 560 580 600
Wavenumber (cm")

Figure A4.4: AF EX fluorescence results (40%, 1:1, 60-9OC)

148

250000 T” —— m— —-~————— ——» —-— ’F —————--— ———w ~

 

| i—40%,l:l,60C
| i—40°/.,1:1,70c .
2000001 —‘ — ——'* ——' *‘— ____ ‘fi—40%,1:1,30c _'“'"_“T

l [fit/041$“ l

150000

 

A
M
G-
9
v
a?
Q
8
a
E
:—

 

 

 

 

1

l
400 420 440 460 480 500 520 540 560 580 600
Wavenumber (cm")

Figure A4.4: AFEX fluorescence results (40%, 1:1, 60-9OC)

148

Intensity (cps)

 

E— 40%,].3:l,60C

 

 

 

 

 

1600““ F—fi “”"_” ’—‘"‘ — ”“—”4—40-/.,13:1,70c *
i—40°/.,1.3:1,80C'

110000 L— 40%,13:1,90gr

120000 « -

100000

80000 «

60000 ~»

40000 1 ——————

20000 K
0 ‘l’-*—‘ l.’ '— — ——" fin_-—‘ ' "—"'_—T" "_—"'" _ fi—‘_'“ T ’*~_ H
400 420 440 460 480 500 520 540 560 580 600

Waveulmber (cm")

Figure A4.5: AF EX fluorescence results (40%, 13:1, 60-9OC)

149

Intensity (cps)

 

14m _________-.__ _.__.___ _____.. ___.__.._ ___, __.,,_. 77_
13— 60%,0.5:l,60C l

_ l—-— 60%,0.5:l,70C
‘— 60%,0.5:l,80C

i— 60%,0.5:l,90C

 

120000 4 , —~— ”~-~-—— VF

  
     
  

l

 

 

 

 

 

 

0 _ fl ' _ T — T T ' ' 7 7 l f ’ 1
400 420 440 460 480 520 540 560 580
Wavenumber (cm'l)

Figure A4.6: AF EX fluorescence results (60%, 0.5:], 60-90C)

150

Intensity (cps)

 

 

 

 

 

l '—60°/.,0.7:1,60c

 
 

 

 

 

 

 

 

 

200000 l -——»~» ___# —.~ #‘_— __ __-—60-/.,0.7:1,70c___fiA A
—60%,0.7:l,80C 1
-60%,0.7:1,90C :
150000 « gens
100000 -~ fl 1»
1
50000 4 _ f

0 ' T 7 T 1 1 . - "v r a!

400 420 440 460 480 500 520 540 560 580 600

Wavenumber (cm’1)

Figure A4.7: AF EX fluorescence results (60%, 0.7:], 60-90C)

151

200000 *—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

i—80%,1:1,60€ !
180000 —~m» —60°/.,1:1,70c Wm
'—60% 1:1,80Cl
l ___.___ 1 __1 ______~ ___ ______ ._ __ l ’ ,_ 2~_,m,
fillllllfl l ,~\ :—60%,1:1ch
/ 1
140000 +—'--4 —«— A, —— A—i—A—— —-———~———-— 1
/ l.
A __ ‘\ . _ _ _
3% 120000 ‘1 M“. l
:3 1000001? -- ~—-——— - -. ———-~——«* #V¥ # w a
= l
§ 80000 — — —— —— i f v 4
60000 1 — — — — e ‘
40000 «~- W#_ _______ —— ~ — — l
20000 ~~ —— —— =~—- l
0 JH— l ‘_— 1' _" __T_—' T T l T 7 _ T fill
400 420 440 460 480 500 520 540 560 580 600

Wavenumber (cm'l)

Figure A4.8: AF EX fluorescence results (60%, 1:1, 60-9OC)

152

Intensity (cps)

 

200000

180000

 

1600001

 
  

P—'———‘ + -___7 fl. ... ‘— "W ___._

 

 

'— 60%,18:1,606 1

 

 

 

"m” — 60%,].3:1,70C ”‘"*“l
\— 60%,l.3zl,80C _ _A_ _4
f— 60%,13:1,90C

 

 

 

 

 

 

 

 

 

 

 

 

 

 

400 420 440 460 480 500 520
Waveuumber (cm")

Figure A4.9: AF EX fluorescence results (60%, 13:1, 60-9OC)

153

Intensity (cps)

 

 

 

 

 

, 1
*WV s—80-/.,1:1,90C,5min l“,

f— 80%,] :1,90C,10min l

 

 

 

 

1__1:_ so%,1:1,90C.15min 1F?

 

520 540 560 580 600

Waveuumber (cm")

Figure A4.10: AFEX fluorescence results (80%, 1:1, 90C, 5-15min)

154

Intensity (cps)

 

 

   
 
 

 

 

 

 

— —————% — 80%,]:1,100C,15min

 

l—80%,1:1,100C,5m1n
— 80%,]:1,100C,10miu

 

 

 

 

 

 

 

 

 

 

 

 

T

 

460 480 500

520 540 560 580

Wavcnumber (cm")

Figure A4.11: AFEX fluorescence results (80%, 1:1, 100C, 5-15min)

155

 

 

Intensity (cps)

 

  
 

 

25000001

7.

l—l — 60%,]:1,110C,5min
1— 60%,]:1,110C,10min l
_j —— 60%,]:1,110C,15min| ___*_

 

 

 

 

 

 

 

 

 

 

 

400 420 440 460 480 500 520 540 560 580 600
Wavcnumber (cm")

Figure A4.12: AFEX fluorescence results (60%, 1:1, 110C, 5-15min)

156

Intensity (cps)

 

l

,_ _ ‘ 1— Auburn 81 (30) 11-? ;

7— Auburn #2 (5d) 1

i ,1 —10°/. N113, 170, 10 min ..... 7
.— 10% NH3, 170, 20 min

1—10°/. NH3, 170, 30 min kw 1

   
 
   

 

 

 

 

 

 

 

 

 

 

 

 

400 420 440 460 480 500 520 540 560 580 600
Wavenumber (cm")

Figure A4.13: ARP fluorescence results

157

Intensity (cps)

1600001e ~————-———- —--——- - —-— ;-—— -—+ g * ~~ 1 f 1

    
  

 

140000 ——— __,,__1__ ,1________ «~—
=——P190c,15mi;l

g1 — P 160C,5min ~__f
1: r 200C, 20min

120000

1

 

 

 

 

—7a--— *'_"_"H__—'Y_ ———-—————- 7— If

0 +_‘—-—_-' "_—7_."'_'_ _T——_—‘ "T— "'— "' 'T‘_"'_

400 420 440 460 480 500 520 540 560 580 600
Wavenumber (cm")

Figure A4.14: Controlled pH fluorescence results

158

Intensity (cps)

 

 

 

 

 

”00°” f 1—1) 140c,40m1n, 17.112804 1?

 

  

 

1 '—D160c,10m1n,1°/.n2s04 : i

10000” L 11 FF— ““1 ——n 180c,1m1.,1%u2s04 1 g

1 — D 180C, 2min, 1%H2s04 3 ’

80000 A / F; ‘V ”F —D 190C, 20mllmin, 0.5 11
°/.n2s04

 

 

 

 

 

 

20000 f —#—-—~m

 

g
0 1 Y 1 r T 1 I r 1 T
400 420 440 460 480 500 520 540 560 580 600
Wavenumbcr (cm'l)

 

 

Figure A4.15: Dilute acid fluorescence results

159

g

1 i 1
5 1

1

1

1

1

1

1

 

 

T— N CONTROL‘
1— 013011711 1

  
 
  

— N301288
1— N30128C

 

 

 

é

Intensity (cps)

g

 

 

é

 

1

100 200 300 400 500 600
Wavenumbcr (cm")

Figure A4.16: NREL samples fluorescence results

160

— N30124c + —~ i

Intensity (cps)

 

6000000 1* 11, 111 A~ 1 —-- __1 4—f 11 +——~ +‘——-----~ -~~—1
1 1

500m 1_____A_A____A -1 1 1 11 -1 A—TAMUM-HL 1

f—TAMU(M-2)|
i—TAMU (M-3)
——~4—TAMU(M-4)[— #1

 

 

 

 

 

 

 

 

l
0 -1 1 _ fl 1—'_-' v —’._ fl _ T

400 420 440 460 480 500 520 540 560 580 600
Waveuumber (cm")

Figure A4.17: Lime treated samples fluorescence results

161

 

  

 

 

 
    

 

 

 

 

 

  

 

 

 
    

 

 

 

 

BO premed Y .......................................................................
‘ Elements 39 i
J Slope: 0.440263
Offset 15.38567
‘ Correlation; 0.663523
80" RMSEC 1545975 ........................................................................
1 sec; 1555195 601 100-5rn1n .
. Bias: 4.646e-07 .
' T 11061111116 . .
40- .................................. 20.4.3501} 90 ........ 834- .90—1 :1)?” arMUll-‘iflrhin' .......... Z
‘ 20-1 3-30 20—0 7- 90 :
1 2 ° aubum(m1)
20..4 . ............................... 13160-5 .................................
1 : g_ :
. / “1.31.1060
0 ' . 60-0 7-80
. 40—1-60
40.. ..........................................................................................
Measured Y
I l l T l I ' Y ‘ I
-20 0 20 40 60 60
Figure A4.18: PCR initial rate model using fluorescence data.
80 pm” Y ........................................................................
‘ Elements: 39
Slope: 0.440263
‘ Offset 15.36567
‘ Correlation: 0.663523
80.1 RMSEC? 15.45975 ........................................................................
1315c; 15.66165 801 1005mm
. Bias: 46458-07 .
. r : 912161901; .
40. .................................. QU..1.W_ go ........ 801-1305315119 ar1'40ll-40rhiri3'””'””
‘ 20-1 3-60 20-0 7 90 *
1 , - aubum(ml)
20... : ............................... 13180-5 .................................
* / 60- i . .
‘ untrea ed -
‘ ~ .7-80
0‘ .
1 40—1-60
_ 0.1 ..........................................................................................
2 Measured Y
I I t 1 1 1 ' I ' f f 1 r r I
-20 0 20 40 60 60

Figure A4.19: PCR 72-hr model using fluorescence data.

162

Predicted Y
60

 

 

 

   
    
  

Elements: 43
‘ Slope; 0.476342
1 Offset 15.24446
Correlation: 0.691623 .
' RMSEC: 14.74955 3 f f f ' '
30.1350; 14192410....1...........1................. 15..
Bias: 2.440e-07 ' - E D 181% min
‘ ' ' . 60-
3 I i Woo—1- 91113 “5mm $931111
‘ . 2 I 1- 90-10min I 80.1-1011.
1 ' . 60-1- 110—15min '
40- '

'un'

' 60—1- lllJ-lUmIn

 

 

 

 

 

 

 

    
 

 

 

 

 

.. aubum(m-l) .
20.. .......... “20.8. .1... B?- a .. ... 11811434401711”
. 604-60 . 406.2% 380
0d '
MeasmedY
1 ........................... 1 ......... 1...-.-...1....-.-..1 .-.....1 ......... 1
0 10 20 30 40 50 60 70 60
Figure A4 20 PCR mltlal rate model usmg XRD data
Pred'cledY
14015131119115; 43 ............ ............ ............
« Slope; 0.260574 ‘ ' ' ‘
‘Offset 44.55940 11111111
120.1. Correlation: 0.529693 .................................................................
1 RMSEC: 2600476
100 ‘ SEC: 29.33819 ; ; ; ; 3 ;
481331 11203-06 ....... ............ ....................... 1...”... ..... -
1 : 1 E sol—1105mm? . lifi-iifi'nfimm 5
4 1 j ; : .mjflm‘ _ - 10min ;
80.. ............ 1 ............ ........ 1
0.1 .... _...1.:_ .... ;. '.
8 ‘ . 5':,'2 ’ WWW 4U19Q aubum(m-1) i 1
‘ '- 5011151133t31bo'.80da69%11411_%m ‘ :
40.. ................................. DTI-. .7 .......... g.‘20;13'80...3 ..........................
; 6%15911
J
20... ...........................................................................................
0_ .
MeasuredY
f. 1-1.1...1.1.1.1+1...1..Tfi_
0 20 40 60 60 100 120 140

Figure A4.21: PCR 72-hr model using XRD data

163

Predicted Y

 

60 Elements: 43 ; j j i
1 Slope: 0.140241 ‘ *
Offset 25112483
l Correlation: 0.374467
1 RMSEC: 18.93539 . . . .
BU-SEC. 19.15948--~-E ...................... ......................
1 Blas. 2.863e-07 . 60-1-90—15m1n

 

 

- 60—1-110-10m1nf

‘ E 3 . 881-1160511111 0-15m1nf

 

 

 

 

 

 

 

 

- ‘—-—-1:1ntreated

0 Measured Y

I fi I ' ' ' I ' ' ' I ' F f I

0 2o 40 60 60

Figure A4.22: PCR 1nlt1al rate model usmg DRIFT and fluorescence data
Predicted Y

140 1 Elements: 43 : : 2 : : ;
« Slope. 01001062 ' ' '

12 ‘ Offset. 61 167167 I . . . . .
DjCorrelation; 01032591.“..“1 ............ ............ ............
1 RMSEC: 32.99955 i 3 - i - -

1 SEC: 33.39009 ; : : : : :
100: 8.83 14758-06 . . . . . . . ............ ............ : ............ : ............ . . . . . . . . . . . . .

 

 

 

 

 

 

 

oi ‘
Measured Y
I r I ' l T fir f 1 I v v r f
0 20 40 80 80 100 120 140

Figure A4.23: PCR 72-hr model using DRIFT and fluorescence data

164

Predicted Y

Elements: 43

‘ Slope: 0.479107
1 Offset 15.25133
Correlation. 0.691453
‘ RMSEC: 14.75267
60- SEC: 14.92746 --- ....

Bias: 24403.07 .1 %15mm
1 1 - - .eo-.1-gafigfif11tb91mmun . 33% “$11,
' -90—:10m1n _ --

160-1-110-15min?

 

80

 

 

       
 
 
 

‘ ' “Qt? - g i
40-1. 35040-130 7- 90 '

   

60—1 110-10m1n

E - ‘ 99905930 99 zoo. 7 7o” : a“””’”‘"‘*” 3 E
20.. .......... 1‘20 ...dar148‘1}_40m'n
1 ; 6W? : : : : I

~ 60-1-60 . 411%311‘5380

 

 

a
I ''''''' fl ''''''''' l VVVVVVVVV l' """""" T VVVVVVVVV I'V'V'7'V'l'v'VVVWY'Ir‘rYrTVVV—r

0 1 0 20 30 40 50 80 70 60

 

 

Figure A4.24: PCR initial rate model using DRIFT and XRD data

Predicted Y
140 1 Elements: 43 : : : : : :
1 Slope: 0.260392
120 ‘ Offset 44.57067

‘1 Correlation: 0.529521
1 RMSEC' 28.00830

1 SEC: 26.33977

100- Bias: 17418-06 ............ ............ 601110-5rm% ........ 1-51
1 I : ; . ,mjflm '-1 - 10min ;

80‘ ---------------------------------------------------------------------------------------
. - . . . . -
. . 5 - . . .

 

.........................................................................

 

 

  

80; ...._. 3.1.1::3 .......... 1 ....... . ..... ‘ z. ......... ............
3 ' 1911133111qu 31% 10119393511911" 2 '
40.1 ............ ..................... D—l .......... ‘20-13-60‘“; ............

20; ............ ............ g ............ g ............ ;

q

 

 

 

 

_ T Measured Y
. 1 . . . 1 1 . . 1 . 1 . . . 1 1 1 1 1 f . 1
40 60 60 100 120 140

Q-
to
D

Figure A4.25: PCR 72-hr model using DRIFT and XRD data

165

Predlded Y

 

80 Elements 113 ............................................. 1 ...................... 1
* Slepe; 0.140241
1 Offset 25.12481

Correlation; 0.374488

 

 

   

 

 

 

 

 

 

 

 

 

 

 

 

‘ RMSEC‘ 18.83538 3 1 1 .
BU‘SEC. 1915943 ......................
1 8135. 15528-07 - - 1 80-1-80-15m1n '
-80-1-110-10m1n;
‘ f 3 ~ 60.1-r1ms1n1ri015mm}
40.1 ..............................................
1 . z .
1':'.'11j. _.11-1-1!80.20-05%819-‘1331mmm . éMWafiimllUmin
1
o- ~————enereated -
MeasuredY
I V 1 I l T v v I y f f r
0 20 40 60 80
Figure A4.26: PCR initial rate model using fluorescence and XRD data
140 pmedy .........................................................................
1Elements: 43 : 3 : I : 3
1 81008: 0001082 ‘ ' ' '
* Offset 81.87187
12031 Correlation 0032591 .........................................................................
1 RMSEC: 32 88855
1 SEC; 33.38008 : ; : ; : ;
100‘8185'14758-05 ............
90- ............ ............ ............ ............ ............
80-
40.. ............ 1 ............ ............ ......... ............ 1
20.. ..........................................................................................
l 3
11 .
Measude
1 11f1111111111111111111111
0 20 40 60 80 100 120 140

Figure A4.27: PCR 72-hr model using fluorescence and XRD data

166

Predicted Y

 

30188111811181 38 ................. .................
Slope: 0.485232 1 1 1 1
‘ Offset 14.42738 1 1 180-1-100-5m1n
‘ Correlation: 0.688588 - - .

80‘ RMSEC‘ 14.82389

 

 

 

 

 

 

1 SEC: 15.02288
1 Bias: -7 528e—08
404 .................
1
1 Z
1 s
20.1 1 ................ .
0; - °60—0.7-80
' ' 40—1-60
_20 I ................... 1....”.1....”31...“..1...H...E.H......H.....i......H......H.:
T Measured Y
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
-20 0 20 40 80 80

Figure A4.28: PCR initial rate model using DRIFT, fluorescence and XRD data

 

 

 

 

 

 

 

140 pmedy .........................................................................
1 Elements: 43 : : ; : : :
Slope: 0.815483
120 Offset 1 1 42864 t : : .
-1 COI'TEIatiOf'l: 01903041 ....... ............ ............ ............... -. ................. . . .
1 RMSEC: 14.18265 3 3 g 60 "10911-38339 m'"
‘SEC: 14.35050 ; ; ; _ 11 ' 1-"910mm .
100... 81390000101 ....... ............ ............ .......... . 1.T111‘1..1 . '11mélja3&1?g1nl'fln ........
1 E E . darw l 1 T? 1min '
30- .....H.....l.H.........I ............ 1 ........... 1.: 1H ...............................
1 i i ~ 201.31 - 0293311 '3'"
1 : : “2901'.“0 .40190 : : :
80.1 ............ 1 ............ “120136. 1...}..‘iwd ............... 1 ............
600 7-90 I E ‘ “7‘“ " 1130-13-80 ’ i i
1 ' : HBO-1 50 : : .
40- ....... 11601071711 ......... 1 ..... 29.07.3q........1...1 ............
4.01110.- '1“ 1- untreated 404-370 1 1 1 1
20- ..... 2:71:31 ....... EEO-1.960...........:............: ..........................
1 80-0 711:- ' 1 39° : : : ; .
0; 3 .34011-an -
Measude
1 4 1 1 1 1 1 1 1 1 1 f 1 1 1 1 1 1 1 1 1 1 1 1 1 1 f
0 20 40 80 BO 100 120 140

Figure A4.29: PCR 72-hr model using DRIFT, fluorescence and XRD data

167

Appendix 5

Table A5.1: Crystallinity Index for TAMU poplar samples

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Crystallinity Index
Sample (%)

DLO-DAOO-DCO 35.91
DLO-DA007-DCO 48.12

DLO-DAlS-DCO 50.55

DLO-DA35-DCO 44.12

DLO-DASS-DCO 12.82

DLO-DA75-DCO 52.05
DLO-DAl 50-DCO 59. 17
DLOl-DAOO-DCO 47.70
DLOl-DA07-DCO 39.58
DLOl-DA15-DCO 55.72
DLOl-DA35-DCO 52.73
DLOl-DASS-DCO 50.24
DLOl-DA75-DCO 36.08
DLOl-DAI 50-DCO 55.47

DL02-DAO-DCO 37.81
DL02-DA07-DCO 26.25
DL02-DA15-DCO 54.73
DL02-DA35-DCO 50.39
DL02-DA55-DCO 56.05
DL02-DA75-DCO 52.33
DL02-DA1 50-DCO 58.76

DL03-DAO-DCO 53.85
DL03-DA07-DCO 42.17
DL03-DA15-DCO 47.26
DL03-DA35-DCO 48.01
DL03-DA55-DCO 39.78
DL03-DA75-DCO 56.92
DL03-DA150-DCO 58.80

 

168

 

Table A5.1 (cont’d)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Crystallinity Index

Sample (%)

DL5 -DAO-DCO 3 1 .66
DL05-DA007-DCO 41 .78
DL5-DA15-DCO 46.51
DLOS-DAO35-DCO 44.31
DL5-DA55-DCO 45.85
DL05-DA075-DCO 38.90
DL5-DA150-DCO 60.10
DLlO-DAO-DCO 29.88
DL10-DA7-DC-0 45.45
DL10-DA15-DCO 35.92
DLlO-DASS-DCO 39.31
DLlO-DA75-DCO 34.33
DLlO-DAlSO—DCO 48.69
DL50-DAO-DCO 50.60
DL50-DA07-DCO 50.59
DLSO-DAIS-DCO 54.32
DL50-DA35-DCO 55.00
DL50-DA55-DCO 52.98
DL50-DA75-DCO 51.21
DL50-DA150-DCO 55.56
DLO-DAO-DC3 22.95
DLO-DA7-DC3 24.62
DLO-DAIS-DC3 21.22
DLO-DA150-DC3 22.15

 

169

 

Table A5.1 (eont’d)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Crystallinity Index
Sample (%)

DL1-DAO-DC3 24.48

DL1-DA7-DC3 25.67
DL1-DA15-DC3 26.75
DL1-DA35-DC3 22.51
DL1-DA55-DC3 20.68
DL1-DA75-DC3 16.03
DL1-DA150-DC3 21.16

DL2-DAO—DC3 1 1.07

DL2-DA7-DC3 22.39
DL2-DA15-DC3 23 .34
DL2-DA35-DC3 23.53
DL2-DA55-DC3 22.17
DL2-DA75-DC3 21 .76
DL2-DA150-DC3 23.52

DL3 -DAO-DC3 l 8.30

DL3-DA7-DC3 24.96
DL3-DA15-DC3 24.77
DL3-DA35-DC3 24.80
DL3-DA55-DC3 24.73
DL3-DA75-DC3 28.99
DL3 -DA1 50-DC3 20. 10

DL5-DAO-DC3 19.47
DL5-DA07-DC3 26.14
DL5-DA15-DC3 21.61
DL5-DA35-DC3 23.13
DL5-DA55-DC3 16.99
DL5-DA75-DC3 24.85
DL5-DA150-DC3 17.62

 

170

 

Table A5.1 (cont’d)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Crystallinity Index

Sample Q/o)
DL10-DAO-DC3 19.85
DL10-DA07-DC3 16.75
DL10-DA15-DC3 17.13
DL10-DA35-DC3 26.67
DL10-DA55-DC3 23.75
DL10-DA75-DC3 14.76
DL10-DA150-DC3 19.26
DL50—DAO—DC3 25.77
DL50-DA7-DC3 37.04
DL50-DA15-DC3 41.24
DL50-DA35-DC3 35.87
DL50-DA55-DC3 37.13
DL50-DA75-DC3 31.26
DLO-DA7-DC6 19.39
DLO-DAlS-DC6 67.83
DLO-DA35-DC6 21 .46
DLO-DA55-DC6 14.57
DLO-DA75-DC6 14.53
DLO-DAlSO-DC6 20.12
DL1-DAO-DC6 15.63
DL1-DA15-DC6 18.10
DL1-DA35-DC6 18.05
DL1-DA55-DC6 23.16
DL1-DA150-DC6 15.07

 

 

171

 

Table AS.1 (eont’d)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Crystallinity Index
Sample (%)
DL2-DA7-DC6 12.06
DL2-DA35-DC6 24.68
DL2-DA75-DC6 23.33
DL2-DA150-DC6 10.62
DL3 -DAO-DC6 1 7.1 1
DL3-DA07-DC6 23.36
DL3-DA15-DC6 13.71
DL3-DA35-DC6 14.12
DL3 -DA55-DC6 1 8.45
DL3-DA75-DC6 15.22
DL3-DA150-DC6 26.00
DL5-DAO-DC6 8.47
DL5-DA7-DC6 22.32
DL5-DA15-DC6 13.89
DL5-DA35-DC6 17.89
DL5-DA75-DC6 15.08
DL10-DAO-DC6 1 8.23
DL10-DA15-DC6 13.01
DL10-DA55-DC6 21.68
DL10-DA150-DC6 15.25
DL50-DAO-DC6 10.37
DL50—DA07-DC6 13.35
DL50-DA35-DC6 1 1.84
DL50—DA75-DC6 17.58
DL50-DA150-DC6 32.13

 

 

 

172

 

Table A5.2: MLR model coefficients for poplar usingRaman, DRIFT and XRD

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

data
Initial Rate 72 hr Conversion
B-coefficients B-coefficients

Intercept 16.9030 81.2270

CrI -0.1930 -0.1360

Raman (1038.9) -0.0005 -0.0016
Raman (1293.3) 0.0039 0.0094
Raman (1369.5) -0.0097 -0.0161
Raman (1399.5) 0.0055 0.0068
Raman (1445.1) 0.0009 0.0018
Raman (2888.7) -0.0022 -0.0038
Raman (2919.6) 0.0020 0.0032
Amorphous cellulose (902.54) 0.0000 0.0000
Lignin 1(1504.23) 2.8130 7.9840

Lignign 2(1596.8) -0.9670 -13.7100
Aldehyde (1650.79) -2.3570 0.7890
Ester carbonyl (1735.65) 7.1900 14.0580
C-H (2900.46) -2.0340 -2.8540

O-H (3394.16) -6.0690 -5.4140

 

 

173

 

 

Table A5.3: PCR model coefficients for poplar using DRIFT data

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Initial Rate
Amorphous Lignin Lignin Aldehyde Ester Carbonyl C-II O-II

Cell (902.54) 1511.94 1604.51 1658.51 1712.51 2915.89 3417.3

PC_OI (X-Vars + Interactions) 0.0009 0.0004 0.0005 0.0008 0.0013 0.0017 0.0043
PC_02 (X-Vars + Interactions) -0.0247 -0.0453 -0.0400 -0.0324 ~0.0233 -0.0643 0.0276
PC_03 (X-Vars + Interactions) -0.0275 -0.0575 -0.0462 -0.0357 ~0.0372 -0.0596 0.0384
PC_04 (X-Vars + Interactions) -0.0253 -0.0170 0.0178 -0.0128 -0. l 190 -0.0628 0.0309
PC_05 (X-Vars + Interactions) -0.0008 0.1290 0.1280 0.0456 -0.0605 -0. 1480 -0.0004
PC_06 (X-Vars + Interactions) 0.0339 0.0361 0.0283 0.0105 -0. 1560 -0. 1640 -0. 1050
PC_07 (X-Vars + Interactions) 0.0990 -0.0612 -0.0330 0.0376 -0.0824 -0.2160 -0.1690
PC_08 (X-Vars + Interactions) 0.1110 -0.0554 -0.0414 0.0272 -0.0993 -0.2230 -0. 1590
PQW (X—Vars + Interactions) 0.1260 -0.l710 -0.0137 0.0662 -0.0684 -0.2160 -0.2950
PC_IO (X-Vars + Interactions) 0.1270 01880 0.0011 0.0847 -0.0267 -0.2330 -0.2160
PC_II (X-Vars + Interactions) 0.1280 -0.l900 0.0019 0.0854 -0.0254 -0.2340 -0.2140
PC_IZ (X-Vars + Interactions) 0.1250 .0. I 720 0.0059 0.0918 -0.0231 -0.2430 -0.2570
PC_13 (X-Vars + Interactions) 0.1240 -0.2120 0.0489 0.1750 0.0132 -0.2580 -0.1490
PC_14 (X-Vars + Interactions) 0.1210 -0.2020 0.0175 0.1960 0.0357 -0.2830 -0. I630
PC_15 (X-Vars + Interactions) 0.1300 -0.1 190 -0.0349 0.1790 0.0307 -0.2800 -0.l300
PC_l6 (X-Vars + Interactions) 0.1260 -0.1100 -0.0180 0.1800 0.0310 «02940 -O. 1440
PC’I 7 (X-Vars + Interactions) 0.1480 -0.0949 -0.0892 0.1060 0.0285 -0.3220 -0.0875
PC_18 (X-Vars + Interactions) 0.1410 -0.3590 -0. 1630 0.3110 —0.2580 -0.4450 0.1290
PC_I9 (X-Vars + Interactions) 0.1580 -0.3900 -0.1530 0.3310 -0.2140 -0.4070 0.2140
PC_20 (X-Vars + Interactions) 0.1140 -0.3850 -0.2200 0.2710 -0.3840 -0.5290 0.2920

72 hr conversion
Amorphous Lignin Lignin Aldehyde Ester Carbonyl C-H O-H

Cell (902.54) 1511.94 1604.51 1658.51 1712.51 2915.89 3417.3

PC_OI (X-Vars + Interactions) 0.0048 0.0021 0.0028 0.0040 0.0057 0.0087 0.0241
PC_02 (X-Vars + Interactions) -0.0362 -0.0697 -0.0614 -0.0492 -0.0403 -0.0975 0.0983
PC_03 (X-Vars + Interactions) -0.0612 -0.l730 -0. I 130 -0.0775 -0.1240 -0.0632 0.2080
PC_04 (X-Vars + Interactions) -0.0657 -0.2480 -0.2310 -0. 1200 -0.0157 -0.0579 0.2160
PC_05 (X-Vars + Interactions) -0.0516 -0.1610 -0. 1670 -0.0864 0.0077 -0.1050 0.2180
PC_06 (X-Vars + Interactions) 0.0413 -0.4100 -0.4380 -0.1850 -0.1580 -0. 1460 0.0130
PC_07 (X-Vars + Interactions) 0.2880 -0.7700 -0.6790 -0.0934 0.1370 -0.3520 -0.0806
PC_08 (X-Vars + Interactions) 0.3840 0.7250 -0.7460 -0. 1750 0.0523 -0.4050 -0.0029
PC_09 (X-Vars + Interactions) 0.4550 -l.3040 -0.5410 0.0974 0.1810 -0.5250 0.0271
PC_IO (X-Vars + Interactions) 0.6430 -2.7840 -0.3450 0.3550 0.1640 -0.5400 0.3020
PC_II (X-Vars + Interactions) 0.6860 -2.8640 -0.5780 0.1550 0.0088 -0.I480 -0.0100
PC_IZ (X-Vars + Interactions) 0.6620 -2.8480 -0.4200 0.5660 0.3210 -0.3080 -0.2350
PC_13 (X-Vars + Interactions) 0.6660 -2.8800 -0.3650 0.5420 0.2770 -0.2870 -0.4740
PC_I4 (X-Vars + Interactions) 0.6660 -2.8790 —0.3660 0.5420 0.2780 -0.2860 -0.4750
PC_15 (X-Vars + Interactions) 0.7110 -2.9870 -0.5350 0.5300 0.2860 -0.1500 -0.5050
PC_I6 (X-Vars + Interactions) 0.6590 -3.0190 -0.3660 0.7030 0.2960 -0. 1330 0.2450
PC_I 7 (X-Vars + Interactions) 0.6210 4.5210 -0.7970 1.8630 -0.3850 -0.5920 -1.6050
PC_I8 (X-Vars + Interactions) 0.6810 -4.6240 -0.7500 1.9350 -0.3490 -0.4230 -I.2980
PC_I9 (X-Vars + Interactions) 0.6470 4.6340 -0.8020 1.9050 -0.3490 -0.4860 -0.9610
PC_20 (X-Vars + Interactions) 0.6960 4.5340 -0.7960 1.7500 -0.5910 -0.5610 -I.0650

 

174

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

wait—cod 3+mnmg. nc+mch n¢+mNNN 21m: .N mc+mm _ .N m¢+MmNN Shanda Named” mmmomonfi oUQéW ~ <Q-NOAD

v¢+mwod m¢+mmc._ nc+mmc.m wo+mm _ .m n¢+mN _ .m n¢+m3 .m warm—Fwd ommmmdm mvmcmdm Nowcvwed oUDfinc<Q-~cAD

warm—No.— mc+mo _ ._ m¢+mm_ .m nfmamd n¢+mnm.m m¢+mom.m mc+m§wm Nagén .ecmcKm Eb _ heed o8£w¢<n7mo40

vc+mo06 3+mONS new: .N mo+mwmd n¢+mVNN moim VNN no+mov.m v0 _ amdn orchvdm abhomé c8-n mc<Q-No‘_D

vo+mM _ .n v¢+mMo.m mo+mwm._ n¢+mwo._ moth—no.— no+mv©._ no+mmb._ mocmhén :meéN gmvma _ .v cuntn ~o<Q-NoI—D

v¢+mmm. m vo+mmo.v v¢+mm©.w nc+m _ c._ §Mcvd vo+m _ ad Kim—ow.” nNdN hvmno. _ N gram _ .m c00$¢c<n¥~c40

v¢+m _ a.” 215%.” warm—Nod moth—”EN mc+m MB.N mGimNhN warm—wad Shcwfin mvnowd_ 8N _ mmmd oUQéG¢<DtNoAD

mo+moo; 21mg.— m¢+mwm.m n¢+Mwm.m mimbnd him—mod mc+mha6 nonbvdm ©3003. cosmgwfi oUQén _<D- _ cox—Q
¢o+m _ m.” vc+m©od n¢+mhm.m mam—SN mimhvd mc+mnvN no+m_wm.m mombcém «.memafim mnwowhcfi cUD-mbc<Qu_81—D
mo+va._ m¢+m5m._ nfmaod mime. .v moim—N— .v warm—N— .v mimaNé bacmmdn awgmdm on _ covaé ODD-nmo<D- _ cox—D
vc+mmbd £1me _ no+m©¢.m n¢+mvmd nc+ma _ .m m¢+m2 .m no+mhmd bNbNth magi: wvw aged cUO-nmc<n—-_oou_0
vo+mchd vc+mcmfi nc+mN_ .N warm—mud m¢+mvNN nO+mNN.N no+mch an _ Nbdn mom — ”.3 shawwhd GSA _ c<Qt~cox~Q
Kim—mm“. vc+mbwfi mc+mhmdu n¢+mnnd firm—OWN n¢+mwv.m no+mwcd mmmwndm aw _ Qum— o—cavaN OUQ$8<DAocAQ
vo+m G.» vo+mmmd mafia _ .m moth—am. m n¢+momd nfmwm. m no+mon. m 2 85.5% @3656— _Nccow _ .N oUQéco<Du _ GAD

mo+m_mm.N m¢+mbvN m¢+Mwm6 Rim—end m¢+mmmd warm—end mc+MNm6 o _ h _ .am wmcmmdm whmmommd GOA—1cm _ (5-250

mint—hm; m¢+mv.m._ memcmé m¢+mhv.m m¢+mmv.m m¢+mcv.m n¢+M¢©.m mm _ m¢.~m ac _ ”Num— Nooce QM cvfltmbo<flucoqfl

VATm—vnd nc+m8._ mimbbd mO+m—V¢.N m¢+mcqm n¢+m 3N warm—mm...” ~m¢mw.m_ mw§._ _ ammowosg cgnmnc<aéo40

wimp—cc; m¢+m§2 no+mmad no+mm_ .m mimaod m¢+m_n _ .m m¢+mwe.m no: <3. moaomé; vmvncm Wm o8-nmc<01cc.._fl

aim—ab; nimww; mc+m£ .v mo+m2 m4. mo+m~mmé nc+mNNé no+momé wcmméw Sccmmd wwwof m._ con—.6 84378.3—

v¢+mnmd $1.95. _ n¢+mm _ .N warm—ANN n¢+mmmd nc+m3 .N mc+mmm.m v”: ..wv Nag: 39.52 .N con—-haoxxfltoocu—D
8+9. _ .0 mc+mcm6 v¢+mNN._ v¢+mmm._ v¢+mm~._ vim—om; 317m— m. _ 325.3 a. 35.2 ”vague; o8éc<fléooqa

@638 9&wa A _ 635 $63 5 3.8». S Amman 3 8&8: comm—gnaw =o_&o>:oo

583. 55 58am 585— 583— 555— 585— 10 p... Q. 88 RES 838%

 

 

 

 

 

 

 

 

 

 

 

3:082 338885 338 é 55 33. 32 3.3. as:

175

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

95.53 «.2 2......

 

 

.888. 8888 8.8:... 88888 8:88 883. 83888 8988 89-883983
a 83.888 .88.. . N8 888.... 888.8 8888.8 3.. :2... 8992... 8983983

.888... 8.888 :88. 83.8.. 88...... 888.. 888.8... 8.988 89883-83

88. .88 8.88... 888.88 88888 8888 8888 8888.. 89.8.8 89883-83

8888. 888.8 88.88 8888 8.8... 8:8... 8888.8 8.8.88 898.3983

888:. .828. 8 $88888 8888 8. 88 8888 8.8888 8.98. 8 89-88.8983

88888 88888 8.... ..v 4.8.88 888.8 88.8 2.5.88 8.92.8 89-83983

58.8. 8.88.8... 8883... .8888... 883 88.8 88888 8.98... 89-8....9-.8-.9
83.88. 88888 .88... 8.88.... .888 :88 8888 8.98... 89839.83
88.... 82.5.... 888.8 88.88 8888 8888 88. 8. .8 8.9.3. 89889-83
88888 8. .88... 88888 88888 888 2.8.8.8 .8888 8.98.... 898839.83
8888 88:88 . ...8888 88.8.8 8. .8... «8.88 88.88.. 8.8.8... 89-8.3983
82.... . 888.88 88.888 8. ...88 888 88888... .8888 8.9:... 89839.83
88 . 8.8. .8888 8388.8. 8888 88 . 8... 8888 .888 . 8.8 8.93.... 89.839. . o3

8888. 8828 8:8... 8888... .88. 8 85.88 88848 8.988 89.882983

8.. . 8.. . 8. 888 88...... 8:88 :8... N88... 88. .8... 8.988 8983983

8888. 88. 88 8.. 8888 E. . 88 8.888 88. 8.8 8:808 8988 89883-83

.88.... 8888... 3.888 883.88 882.8... 888.8 888.8 8.8.2.8 89883-83

8.888 8.8. 88 88888 8. .o. .N .8: 8 3.8.8 8. .88.. 8.8.8... 898.3983

8. .88 a8 . 3. .888 8.2.8.8 888. ... .88... a. 89.88 88.88 89-83.9839
882.... 8.88.8 8888 8888... 888.8 4888 8888... 8+8; 8983-839
9:889 @188. 8.8.... 888: 9.88: 88.88: 988.88 98.

:-0 2-0 3.59.8 .28. 09.2%? N 5:»: . 59m: :8 82.9.8.5 8:88.93..— 8.958
I

 

 

176

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8.2.8. .84. 2...;

 

 

 

 

 

8+8... 8+8... 8+9... 8+8... 8+8... 8+8... 8+9... .8888 888.8 88.2.... 89.29.83
318% 8+8... 8+8... 8+8: 8+8: 8+8... 888.... 88.8 888.88. 89.829839
8+8... 8+8; 8+8... 8+8... 8+8... 8+8... 8+8... 8.88.... 888.8. .888... 89829-...3
3%.. 8+8... 88.8.. 8+8... 8+8... 8+8... 8888.8 88.8. 888.... 89-88<9-...3
8+8... 8+8... 8+8... 8+8... 8+8... 8+8... 8+8? 88888 88...... 88:88... 89829-...3
88.88 8+8... 8+8... 8+8... 8+8... 8+8... 8+8... 8.8.8.. 888.... 888...... 89-8.2923
8+8... 8+9... 8+8... 8+8... 8+8? 8+8? 8+8... 88.8 888.... 8.888... 893.923
Helm-an 8+8... 8+8... 8+8... 8+8... 8+8... 8+8... 888...... 8.888. 88888 89.82983
88:... 8+8... 8+8... 8+8... 8+8... 8+8... 8+9? 8.8.... 8...... 8888.. 89-8.2983
88.88 8.1888 8+8... 8+8... 8+8... 8+8... 88.888 8.8.8.. 8.88.8 888...... 8988.983
8.188... 8+8... 8+8... 8+8... 8+8... 8+8... 8+8... 888.... .88.: 8.8.8... 89-88.2983
8+8... 8+8... 88:88 8+8... 8+8... 88.8... 8+m8. 8888.. .888 8888... 89.2.2983
8+m88 88.88 8+9? 8+8... 8+8? 8+8? 8+8? 888.... .8888 8.888 89-8.2983
8+8... 88.88 8+8... 8+8... 8+8... 8+8... 8+8... .88... 8.8.8 88.8... 89-8.2983
8+8... 8+9... 8+8... 8+8... 8+8... 8+8... 8+...8... .8888 888... 888.88 8988.39-83
#8.. 8+8... 8+8... 8.18.... 8+8... 8+9... 8+8... 8.888 88...... 8888... 898.29%
8+8... 8+8... 8+8... 88.888 8+888 8+888 8+8... 8.8.8 88.8 ...-888.8 89%
8+8... 8+8... 8+8... 88.88 8+8... 8+8... 8+8... 88...... 88.8 8.8.88 89-88.2983
8+8... 8+8... 8+8... 8+8... 8+8... 8+8..8 8+8... 8888.. 8888.. .888... 898.8983
8+8... 8+9... 8+8... 8+8... 8+9... 3+8... 8+8... 8.8.8.. 8.8.8 88.888 898.2983
8+8... 8+8... 8+8... 8+8... 8+8... 8+8... 8+988 8.8.88 888.. .888... 898.2983
8...... 8...... 98...... .88... 8.88.. .88... 8.8.. 8.8.289 g.

E .889— 9~Eam .5an. 999.5. 9895. 5.9.5. to 8.... Q. 88 1.2:— ...—988m

 

177

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

88.8 88:: 82 :2 _ 2888 383 $83 8888.» 8389 89-83-83
88”.: 888.0 .2889 88m.” 888 988.. :88: 8389 89823-83
885.: 3.33 888 2 88¢ 88.“ 8888 $823- 838; 89-83-23
288.8 ~88.» .823 8:83- :82 £25 2883. 833.8 89-83-23
8:: .8 w _ ES- 888... 883 8888 o: 2 .N 88: c _ .m 8389 89% _ 3-23
2 83: 88...» 8888.» m _ o: :- ms-wma .888 :88? 8389 89-83-23
4»: 3- 683 8833 8389 883 e E ._ 2889 2:82 89-83-23
288.: 8888 828: 888+ 82 ..m 85% 2 888 8+8 2. 89823.83
8 _ RE «888 :52 £889 883 88: 8838 8+9. 2 89.83.83
882 38: 882 2 888 8:3 882 £882 838.,- 89-83-83
N833. 8888 823 t 82 882 823 88.3 838.» 89-83-83
8892 888.8 888.8 5888 883- 888 .888."- 3-99 .9. 89-83-83
”:8.“ 8382 3888 88m: $82 $32 828.. 83%.» 89.83983
888.9 8838 28338 E 5: 8:3 882 88:2 8+8? 89-88383
#182: £888 _ _ 3%; 889m 883- 883- 882 3388 89-8 _ 83-83
28; 888+ 38. E 5.83 8888 883 £23 8388 89-83-83
T8102; 888+ 838$ _ $83 823 889-.” 8888 8+9 _ 2 89-83-83
83-3 «283 3883- 82 EN 882 :8 _ .m 83.88 8+8? 89-83-83
882 .833 888% 88:: $82 8:3 £88.. 8358 89.23383
88: 3828 8883 S82: :3: 823 as: 8.9 8388 89.83983
8888 882 :88.” 882 889 «88.. eve 2 “$988 89883-83
82-80 @885 8.85 888: 3.83 8.83 98.83 98$
2-0 1-0 3:398 98mm 3:332 N 5&5 _ 59w: :3 Bonn-5:8 accuomESE 833nm

 

 

8.233 «.8- oz:-

178

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

833.8 838.8 3+8? 338.. 333.. 338.. 333.. .3888 3.8 .8. 83-83-83
838.8 838.8 338.. 338.. 338.. 333.. 3388.. 8888 .8888 8388.... 89-23-83
838... 838.8 338.. 338.. 338.. 338.. 338.. 8.888 88.8 888.8. 898383
88888 8388 338.. 3388 3388 3328 3388 .88... 888.8 888.2 89-83-83
838.... 838... 3388 3388 338.8 3328 33.88 8828 8.8 883.3 89323.83
8.388 8388 338.. 333.. 338.. 338.. 3328 888.8. 888.8 8888.... 89-83-83
888.8 8388 338.. 338.. 338.. 338.. 338.. 888.88 3.8.- 8.: 89-83-83
838... 83...... 338? 338.. 338.. 338.. 338.. 88888 888.8 8832 89-83-83
8388 8388 338.. 338.. 33.? 333.. 333.. 2888 8.3 888.... 89-23-83
838... 888....- 338. 333.. 333.. 338.. 3+8... 838.8 883.: 38:88 89-83-83
838... 8388 338.. 338.. 338.. 338.. 3328 8.8.3 8.3 8... 89-83-83
338.. 338.. 33.88 33888 3388 33.8.8 333.8 88288 888.8 2888. 898.383
838.8 8388 338.. 338.. 338.. 3388.. 338.. .3888 8.8.8 888.... 89-23-83
8388 83.8 3388 3388 3388 3388 3388 8838.3 88.8 88. 89-83-83
3+8... 3388.. 3388 3388 3338 3388 33888 88888 888.8 828.... 89-83-83
8+8... 8388.. 8388 8388 8388 8388 8388 888.8 888.8 88.88... 89323-83
838.. 833.. 838... 838.8 838..- 8388 83:8 888.8 888.8 .8838 89-83-83
838.8 83888 838... 8.38... 838.8 838.8 838..- .888 888.8 8888..- 89-83-83
8328 8388 838.8 838.8 838.8 838.8 8388 8 888.8 8288 89-83-83
8838.8 8328 8388 338.. 338.. 3+8... 332.. 8.88.3 .8888 83838 89-23-83
8.338 8388 8388 8388 8388 8388 338.. .888 8.8 88888- 89-83-83
@8388 88.8 9.8:. 8.88: 8.88: 88.88: 8883 8.22:8 8.888

ES... 585. 5&5. 8.3. 5:5. 585. 583 .6 Q. 88 22 .88 8.8.53

 

 

 

 

 

 

 

 

 

 

 

 

25:85 v.2 038,—.

179

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. ENS «830... an. .3... mums... .m 8...... .v 38.; 3338 33......V mom-mm<o-mo+.o
38mm... vaomocd .e. ~88” 8.33... 8... 2888 8. . 3.... 83.3... non-n . «5-8.5
mummm. . 3.38.8. .8383... 8338.8 8.98 3......“ 3.... . Nd 3+3. ... moo-8.8983.
$38.8 3.0on on . 8.3m 3. no“... «33.. «88.8 .332... 33.8.. moo-c<n.-mo-.Q
.83»... 902-8... than. .. Niko . .m 3. mm... 330... 333.8 33.8...‘ mom-ow . <0- . Sn.
gong... 8.82.8.8 8.33.. 8.33. .N 83.38 3338 «9.38. 3+3... MUD-m8<n.-.o+.n.

#29... 3838 .8838“ 82.8...” 33... Sam . ... .3838 33.8.3 moo-mn<o-...+.n.
053.... Q... . .....m .33....” «sown-m .33.». 338... .3858 33:-m... mum-3.5-8.5
«3. ...N. ... 838... .3386 . 89$... momma.» 8838 2.8.3:. 33.3... moo-n 2.5.53.
52...... 3.8.88 N333... Emma-82m «~33. «......n a. . $8.... 33.3.8 MUD-..<n.-.o1.n
38m...» 3388 .8380». 88.3.. .338 8.338 .88 . 88.. 33.de muo-o<n.-.e-.n.
3030... Sue . ... m o. .3.-mm... 3.88.8 888.6 033.. m 383.. 33.8.8 MUD-ow . «5-39
3.3%... mmm . «Om $88.08 $388 338.8 ”388 .333. 3+3...” MUD-m.<o-o-.Q
8.8.8.2 3.3.3... ~39. ... mmnwowd NW8 . Nd 8N8 . 3... 238.8 33.38 moo-8.80-30

wmmomd 8.338 onmmmmmd «338 0.868 .338 85.38 33.88 MUD-35-30
832.3. .3.: .o 38$... No.88. .9.va a . 03”.. 3...... . 2m 33.3.. coo-omEQém-E
mm .08.. . N 533.8 8.535 2.8.338 .836 88.58 38.308 8+3... coo-3.85-33.
o. .3.... ...-8.8.3.8 ~58... ... 8388.8 38 . .m 3888 .8838“ 33.3.. con-3.3.685
8.333 38...... ._ .v. m . ..- mnmum. .v «3&3. 3.38.8 m .33... 33.3.. coo-3.5-81.0
3% . .8.” moved. 38.93.... momcomw 3. . .... 3338 3308.3 33.3.. coo-n . .3.-3+5
3.02m.” $2.0m. v8.2.8. wow-35m 33.m «NE... «3038... 33.3.. coo-$3.685
G. .38.. .388. .882. 82.8.. 9.88 .. .832. .8888. .88...

...-O :6 .33-.8 .88.. 3.302... N 583 . 9.55 :8 2.9.8.9938 8:88.32“. 87.53
III.

 

 

 

8.2.8. «.8. 2......-

180

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

838.8 838.8 S32N S32N S32NN S38N S328 288.8 88.8 888.2 83-83-23
838.8 838.8 S32NN S38N S38N S32.»N S38N 222.2 8.8 288N.8 83-23-23
8382- 8388 S32N S38N S3»NN S3»NN S33N 828.2 888.8 882.2 83-83-23
838.8 S382 S38.» S38.» S38.» S32.» S38.» 8882 2.8 8.8 83-83-23
838.8 $32.» S38N S32N S32N S32N S38N 828.2 828.8 2282.8 m3-823-83
838.» 838.» S38N S32N S32N S32N S38N 88.8 82-» 2.8 83-83-83
838.8 832.» S38N S3:N S32N S32N S38N 288.2 8S».8 882.2 83-83-83
838.8 838.8 S3SN S38N S38N S32N S38N 882.8 8.8 8888 83-83-83
838.8 838.8 S3SN S38N S32N S32N S38N S888 888.8 882 83-23-83
838.» 838.». S382 S382 S382 S382 S382 8328 8.8 8.2 83-83-83
838.» 838.» S38N S328 S328 S3$N S38N 288.2 888.» 888.22 83.83.83
838.8 838.» S382 S38N 2328 S38.» S32N 888.8 8.8 8.8 83-82383
838.» 838.8 S32N S+mNNN S38N S38N S38N 288.8 288.8 888.2 83-83-83
832.8 8382- S382 S382 S382 S382 S382 888.8 8.8 8.8 83-83-83
838.» 838.» S382 S382 S382 S382 S382 8.88.8 282.8 888.: 83-83-83
838.8 838.8 S382 S382 S382 S382 S38N 288.8 8.8 888.2 83-23-83
838.8 838.» S382 S352 S382 S382 S382 888.8 8.0.82» 888.2 83-83-83
838.8 838.» S32N S32N S32N S32NN S38N 8»8.»2 N8 8888 83-83-83
838.» 838.» S38N S32N S322N S32N S38N 8:88 8288» 2.22.8.8 83823-83
838.8 838.8 S382 S382 S382 S382 S382 S888 8.8 8888 83-83-83
838.» 838.» S382 S382 S382 S382 S382 882.8 8828» »88».2 83-83-83
8.288 3.888 3.822 G8»: G8»: 8.828 8.88: 8233.28 8232.58

5:52 8:52 5:52 5:52 5:52 5:52 as 28 e2 8 38 282.2 8283»

 

 

 

 

 

 

 

 

 

 

 

 

Anvznouv v.2 033.!

181

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

398.32 2332.26 83203.3 32322.2.” 82mg” $39.3 22338.3 3323.22 Beam-420-3222
.8332 3332-: 333386 3mm 23:8. 323. 2 3332 289203 33232 .3 moo-m 245-3 2 22
853.2 333.2. 33333 58- 2 3.m N33». 2 N332 .N mmwwemd 3322-.» mun-823232-232
N38. 2 2 3m: 3.m 32 2.89.6 9532 .N 32 3.2 :- 23m .2 852828 3322-22 mun-22.32-22 2 29
v2 womdm $223.2. no 2 SNN 2333.». 323.32% we» 2 m.m N328- 2 3.2V 332388 moo-on 2 «5-3-2Q
2238-62 $32.6 233322222. 3833 2323.3 2033.3 mnmnqm 3323.6 806532-3120
0323228 230303 223323.2- om33.m 38.23 NS 2226 Neon-83.3 333.3. moo-mn<Q-3-2n2
flow. 2 2 353.3 v2 398 5N2 nvd «$22.3- 3333 32: EN 33836 moo-3232-322
«33.02 280 2 v.3 NEAR-mad @3836 men» 2 .m 2 80:” 2 $3336 33223.2. moo-n 2 232-332
2285362 33.3856 $333.3 wmoowwd 2 23.333 2 2323." 238936 3323222V mom-2.932-322
32238.22 $332.8 2822.235 23323.2 2 23222 2.332 3323.2 3323.3 MUD-35-320
22.2 Edn 3532». 3322.2 @3330.»- m23~.w 3332.” 33322” 33:-Nd moo-222232-332
33232 8.9808 3 2 2.33.3 SE-omd 0333 228.2 .N 33333 3+”: 2 .m MUD-2.232-332
2332.22 Smmmwé 33223.2» 33283.3 2233.2» 2.8323 33% 2 22 2 .m 3323.2.- MUD-3225-815
2 22.383 383.32 3:22.3n .3 «2323.22 0229.32 22823.22 New-2230.32 33232-2 moo-325-3232
3393.32 @3336 3332-32226 33333.3 2 2mm 2 .m 2.2 222m.m 333N336 3328.2. moo-n 2432-8-22
333.3 2 83332.»- 3wm 2 336 228-323 38. 2 ed 32 3.3 28.3 2 $8 33222:. moo-8.32-315
333.2 M3832. 2.282286 2.82 m: 336 8335 2 2828.3 3+3»; mom-332-8232
N332 N2 3332 5. M3322 ES. 2 .3 3333.2. 533.3 2 mm 2 3.3 33233.. moo-ow 2 «82-8-20
3832.02 2832206 328233 2 2 28v.» 3332; 33m 23 23333.3 332221. MUD-2.45-820
3 2 35.32 23382” 22 2303 2 332. 2 333 2333-8 2223: 3+3». 2 mom-3285-822
22 .882 22889 28.82- 22 22.822 88»: 8S»: 288.82 282
:-O 22.0 22928228 202mm 2222220222 N 322322 2 323322 :8 2232232052» 8220820322 82353

 

 

$533812 033,—.

182

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3% 8&8; 8&8; 8&8; 8&8; 8&8; 82.8.8 883.8 888.8 8983-8-29
E 8+8: 8&8; 8&8; 8&8; 8&8; 8&8; 288.2 882;; 8.88.8 89-8<9-8;9
8&8.» 8&88 8&2? 8&8.N 8&8.” 8&8.~ 8&8.~ 3:8,“; 5% 8888.:- 898298-29
8&23 8&83 8&8; 8&8; 8&8; 8&8; 8&8; 28.8 888.8 358.8 898398-29
a 8&38 8&8; 8&8; 8&8; 8&8; 8&8.N 838.3; 2.8 888.2 89829-83
l8+|m~2 8&me 8&8; 8&8; 8&8; 8&8; 8&8; 838.3; 822;: :832 89-2<9-8;9
8&8.“ 42:82 8&8; 8&8; 8&8; 8&8; 8&8.~ «28.2 88.2 288.8 89-39939
3; 8&8; 8&3.m 8+8?” 8&Nnm 8&8m 8+8; 88:8 8.2. 8.: 89829.39
8&8; 8&8; 8&85 88288 8&5; 8&«3 8&8: 288.: 888; 888.8 89829-39
8&8; 8&8; 8&:.m 8&mmm 8&8.m 8&8? 2:82 883.; 8.8 2.2 89829-39
8&3: 8&8; 8&8.” 8&2? 8&88 8&8.~ 8&8.” 25.8 888.8 888.2 898298-29
8&8; 8&:; 8+qu 8&32 8+8? 8&8.~ 8&8.~ 88.8 8.8 8.: 89829-39
8&8? 8&8; 8&8.N 8&8.N 8&8.N 8&88 8+8? 8822 822.8 $.88; 89-85-39
8&8.m 8&88 8&8; 8&8; 8&8; 8&8; 8&2; 88.: 8.8 2.2 89829839
8&8: 8&8.m 8&2; 8&2; 8&2; 8&2; 8&8; 88.8 838.8 88.: 89829-839
8&88 8&88 8&3; 8&8; 8&8; 8&8; 8&8; 88.8 8.8 8d 89829839
8&88 8&2; 8&8; 8&8; 8&8; 8&:; 8&8; «38.:- 8832. 828;: 89829839
33%; 8&8; 8&8; 8&8; 8&8; 88.8 3.8 82 89-8<9-8;9
8&88 8828.8 8&2.” 8&3; 8&8.N 8&8.~ 8&8.~ 88.8 888.8 88:8; 89-39839
8&8.m 8&86 8&8; 8&8; 8&8; 8&8; 8&8; 888.2 8.8 8.8 89-829839
8&88 8+8: 8&8; 8&8; 8&8.N 8&8.~ 8&:.~ “82.: :88 288.8 89829-839
8+8: 8+8? 8&8.N flea 8&8: 8&2.” 8&3? 888.8 8.8 888.: 89829-839
123.283 28.383 228: 8.83 2383 98m: 238: 8% 82528

was EMF—QM Ems 52.-Ham 2.2-NE; Eds 2.2-weav— —.~U mu: Nb Due 33:: mO—QEQW

 

 

225.592 8.3; as;

183

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Sam: .3 _ Rvgd wamocmmd Nmonhmé Nommvd 938d wmnwnfiwv 2:98.— oUD-mm<D-NSO
«$3.2 momhhwd ow _ an _ 0.x. womnvwé v _ 89m «336 N38 _ N.m 21mg.— 000-535-840
«Seq—N 58356 33: mm; $336 2936 gag. on 5 _ .v 8+Mwo; eUQén _ «RT—30
manned _ Cmnoné 23366 anemmad mmzzé $88... 2 :hRd 8+muv; coo-mm<Q-_oqQ
S _ 33: 3886. 29.0.36 _vnwov.m aghod Send .vavN 21mg.— oUQ-nm<Q-_o‘_Q
a mom—EM «3336 N. SeoN 3cm: Ewmfim Second eo+mofim 009m _ <0- _ can
383: vhwommd ma _ ”mm... Evand :thoN. $551” _ KEVQN 8+mvwé con—-35- _ can
can _ .N_ 3886 h _ named 2.8 _ .m mnmnwé n _ homé ovmvomwd 21mm m6 000-835-93—
_onmo._ _ $3 $4” 834%.— oamgwd 532 .v Emmad ammo—Nd mc+meo.v eon—-mfixn—AEQ
02 3..». _ n _ moved 2 Sand Seammé quagd 0256 N. _ wmvod nc+mmm.m oUQ-mn<Q-c1_D
«$5.8 om _ omvé oak—$4” 35%. m «mm _ o6 mmovmd avwvwo. m Sign. m con—A $5-3m
Evian _ onmveofi @5336 3:2 3» mmwcvd mmmwwfi gonad m¢+mflnm won—$2930
33:: Such. m 3886 vafl .m 3086 831v $8 _ and mo+m 36 oofl$o<néxa
a _ o8 .5 2.38.0 $5336 vmhwmmd NNVNQN .89: 252.86 03mg; moo-m\.<0-cmua
335.2 m _ $36 cwommoé ”885d N393 m3 3; Sbmmomd Rim—or.— mUD-nn<Q-omAD
oaanN— 38— .m wwmmmé _vvmmn.m vwnmvfl Sam»; awvmvo _ .m 21mg.— m0fl-mm<0-ogo
Gum”: _ Semi; :: 3mm...” compo; S _ 5.. wZVN: onvawam.m Rim—bod nonzm 35-315
825.: :5 _ aqm 333$ 82 we; 2 Emd owns.— oow 52 .m 8+Mam; aux—$35639
nomom.o_ wmowomw Swmfibé SmnONN Emu: «Soc; 3336 mc+mflmm mom—54.9340
038.2 339mm mmaommd mange; 33w; wet. _ ._ 332..— £19.: non—6w _ <D-o:Q
macs: 533% 3.303;. 3936. Embed SSNN mt‘oovd m¢+Mmmd mogmfixaéso
28M:— mmmomwé 32.9.6 ooamfim Km an 38a; Svewvfim mo+mood mUO-mn<D-o:D
§ .38 $683 $39 a 838; 9.0%: 92%: @383 :3
:.0 1-0 795.18 53m ooze—024x N 5&5 _ 5:34 :8 30:90:; 358202”. 838%

 

 

€3.53 «.2 2.3.

184

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8+m..».~ 888.» 8+8... 8+8... 8+8...» 8+m».... 8+8... 8.2.8 8.8 8.». 8382983
888.» 888.» 8+m8. 888.. 888.. 8+8... 8+8». 8...»... .888 888.2 8.3-8.983
88.8.» 88:.» 8+m8. 8+8... 88.8.. 8+8... 8+8... 258.». £83» 888.: 8983-83
888... 8+8... 8+8.» 888.» 888.» 888.» 888.» N88... ...8 888.8 898.983
8+8... 8+8»... 8+m... 88.... 8+8... 8+8... 888. 8.8.2 N828 5.23» 8382923
8+8»... 888.8 88:. 8+8: 8+8... 8+8». 88.8.. 88....» 888.. 888...» 89838.3
88:8 8+8»... 8+8... 888.. 888.. 888.. 888.. 8.8.». 8.8.»... ..88».»» 83-23-23
888... 888.8 8+8...» 888.» 888.» 8+8...» 888.» 8.8.». 82...» 888.8» 83-3923
8+8»... 88:8 8+2». 8+8». 8+8». 8+8». 8+8...» «88.2 888.8 »»8....»~ 8982983
8+8.» 888.» 8+8.» 888.» 888.» 8+m»~.~ 8+8...» ..»8».... 888.8 288.8 88-82983
8+8...» 8+m8.» 888.» 888.» 888.» 888.» 8+8...» 883. 8.8 8...» 83-23-83
88.2.8 88.8.8 88...... 88»... 8+8... 888.. 8+2... :83» ...»...».»» 8...»...»» 898.983
888.» 8+8...» 8+8...» 8+8...» 8+8...» 8+8...» 8+8...» 88...» ....8 888.8 898.983
88.8.. 8+8»... 8+m»». 888.. 8+8... 8+8... 8888.» 8 »»»»~.8 888.8 8982983
8+8»... 888... 8+2». 88...». 8+8». 8+8». 8+8... 8882 8.8 2.8 8982983
8+8»... 8+is 8+8». 8+3». 8+8». 8+8». 88»... 88...». 8.8.8 888....» 898.983
88:8 888.8 8+8». 8+m8. 8+8». 8+8». 8+8... 8.»... 8.8 .88»..» 83-82983
888.8 888.8 888.. 8+m8. 888.. 8+8». 8+8... 88.». 888.8 88m...» 23-28-83
88:8 888.8 8+8... 888.. 8+8... 888.. 8+8... 28.8 8.8 .88»... 83-88-83
88...»... 88...... 88...... 8+8». 8+8». 8+8». 8+8... :8... .82» 882.8 893.983
8+8...» 888.» 888.» 888.» 888.» 888.» 8+8...» 58.8. 88 ..».N» 8982983
888.» 8+8...» 888.» 8+8...» 8+8...» 8888.» 8+8.» 8.8.8 8.8.8 882....» 8982983
3.28. .28». 28...... 8.8»: 8.88.. 9.82. 438.. 8.23:8 8.23:8

5:5. 5:5. 5:5. 5:5. 8:5. 5:5. SS... .8 a. R 38 .8... 8.95»

 

 

 

 

 

 

 

 

 

 

 

 

3.23312 035—.

185

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5 _ 523.6 mi nmvd NV5N _ nd aoawwm; $89.— _o_5c._ 58m _ 5.. eo+m8N oUOén _ 45-815
@5303 53356 SN— :3” 52%— .N can»; 235”. .332: wo+mmo._ oUD-nn<Q-cm‘—Q
5593.3 «589.? vamoMNé n85 _ a; ecNoo; 8— m _ ._ o5owMN5N mo+mc5.w eUQ-5o<Q-om1_Q
NomoN.m_ 585;V N332. awaoNoN 5nNNo._ m _ Nam; 336de mo+m5m.w con—545695

N312 ”5326 385 _ ._ wNmtvN 03amN 59%: 258M5N gigs.» 08-3 55-215
SN. ..N. 3335 3963.6 cwnwONN 85m.N «0N3; _m°NV5N.N m¢+Mnn.n cUQfiQxQéSQ
N. 503 @5336. 2082 .n c5wVNnN @250.— wN5ww._ _mnm5o.N mo+mvw.v ooo-m_<fl-o3n_
$35.2 35mg; 53356 353: .N N5owc._ 5 Sam.— _ .mmEVN n¢+m5_ .m eon—-356 _ AD
ewMNv.5_ wgmwmd _noNowN.v mom _ mN.m 3nNad 3N¢m.N new—ed mc+mNN.m eon—655-340
Same. _ N 333.6 8333.5 330v :6 _ mi 59355 oawwcmé mo+mcm.w oUD-mm<Q-nSQ
nmvww.m_ 338$ wwon5N.m m _ memN _ 5 _ oNN woNovN vwmovaN nc+m8mé con—6 ~ «.0695
new _ 3: 26°36 N835 aNN_ N5.m NomwoN 353% aNo _ 35m nc+mmmé oUQ5o<D-nSQ
5m _ NN.m_ mcw5No.v NEW—356. SEEN n33.— oognN aoNow _ o.m 9358.0 oUQé<Q-ncAQ
585mg 58356 .538.— 383.". on _ Nvd $N5e.m «$386 Rim—ma.» eonén _<Q-mo‘_o
$53.2 vo§N5é Ragga Nmo5o¢N 5Nwocé 0886 E vogN 8+me5 coo-m5<o-moqa
Nme5.5_ aonmm.n 0532 .m oomwwvd ”530v 089.6 253?“ oc+mO5d won-mm<a-moqo
5350.2 8536 NemSNod m _ m5¢n.m Gm: .v Senna _an5vm.m 31mg.” oUD-mm<D-moAQ
895.: ova—ad 3528M; ¢ceN5mN NNvomN vanNmN mmmmc5vN mo+mNo.w 000% 35-815
”33.2 cNaovvé manned Novww5N EEHN mwmvod cNmmoaN 3+9: .0 000-535-8‘5
N _ vN_ 6N NoemN _ .w 03 _ momd $386 a _ 83. 5N5om.n Nmmow _ N.m 23$qu oUQ-o<Q-moAQ
«No. 3: GVNwod 05332 com—”.56 5$N.w gamma Nag—QM 5¢+m5._ oUQ-o2<Q-No1—Q
Nwwwo._N omoM _ ..5 $330.». gown“: $306 Manama wNvamv.v 53mg; cUQ-m5<Q-No1_Q
Giana $.8an $0.25 c 838: 86%: 5.33 92.8% 53

2-0 1-0 1:880 56m onset? N 5:»: _ 5&5 =3 mac—ESE“ 858202.; moan—um

 

 

 

 

 

 

 

 

 

 

5.83 ...? as:

186

 

 
 
 

 

 

   

 

 

 

 

 

 

 

  
 
  

 

 

 
 
     
 
 

   

 
 
   

 

 

 

 

En pIEd’CIEd Y ...............
“U Elements 115
Slope. 0746422
Offset 3.790370 '
40- Correlation 0863957 rrrrrrrrrr 1 . . ...........
RMSEC' 4.843273 l _
SEC 4.863594 3 3 f
Bias, -2 [Ilse-06 : ' L05-DA55-DC3 :
30 ' ..... r ..... .............. ........ .. UL . . ................
i ' DL_0l
20- ...............
10.. ..........................................................
0- I74 .
- - DLsooAv-{oco
gm- ..............
Measured Y
...... ,,, ..........[ W,
10 0 10 20 30 40 50
Figure A5.1: MLR initial rate model using DRIFT data
nn pred’CIed Y . . . . l , . . . . , , . , ................................................
‘u Elements: 115 1 ’ . DLUS-DA55-DC3
Slope, 0.732920 V ‘ '
Offset 19 44055
1004 Correlation 0.858112 . , .......................... “I.
R 12 82035 ,
SEC 12.87558
Bias -5 598e-06
80 . ...........................................
5|]- .................................... DLEl—l ‘ .................
I .
l . ' I 0‘ U I , :
' * 7 ' ' DM7-D‘éfiooi-DAHE09CGA‘5QDCD :
4UJ ........... 3g; 9 ' l0 ...................... ...............
20- .......... ; Ulfi‘l'—UAUUU—UCO . . .............. .............. .............. ..............
ml -
Measured Y
,...,...,...,...,..,,Y",
0 20 4D 60 80 100 120

Figure A5.2: MLR 72-hr model using DRIFT data

187

Predicted Y

 

:n
"” Elements 115
Slope 0 384517
Offset 9 498907
40. Correlation 0 603753 . . , . . . . . .....
RMSEC 7.358554 -
SEC 7 382723
Bias -1 393e-05

 

 

 

 

 

 

 

-1U‘ .............
Measude

........ I IIII I......I .rrj—rr

—10 0 10 20 30 40 0

 

  
  
 
 
  
 
 
     

 

 

 

 

 

 

1n" p’wuw Y ........................................................................
'W Elements: 115
Slope: 0.697512
Offset: 2201345 . _
Correlation. 0.835164 j '
ao~ RMSEC: 13.43133 000C“ ----------- ; ------- 151035.151 . - 1
SEC. 13.49017 ' '
Bias. 88909-05 DLOS—DADDD—DCU DLOB—DAIJlSD—DCU .
30- ............... : ..... DLOB—DA055-DC0..,.,H..H..l.;
: . co om3owasoco DL001-DA150-DCU :
i ' : - - ECU :
4U_ ................. : ................ I ................. : ...... DL&B%?&EE§:CU ............... I
, -"_ '- ‘ oco - '
' DL001- DA035-DCO ~
20 ............. oLuoinreoco....,.,.....;.................;...._............;
1 “BLDG DA055-DCO
0 0 I
Measured Y
['1 2'0 4'0 5'0 30 100

Figure A5.4: MLR 72—hr model using Raman data

188

Predicted Y

 

  
 
 

" Elements 115
Slope 0.777990
Offset. 3.318470 : ‘ '
40_Correlation 0862048
' 4 344473 1 - L
SEC 4.363487
Bias -4 559e-05

 

 

 

 

......................................................................

 

 

 

 

Figure A5.5: MLR initial rate model using DRIFT and Raman data

 

 
 
 
 
 
  

 

 

 

     

 
 

 

 

 

1"]fl prEdmedY .....................................
‘" Elements 115
Slope 0.847398
Offset 11 10806 ‘ : .
100- Correlation 0.920559 .. ,,,,, j,, ........ . .. ..... ..
RMSEC 9 538782 ? 1 - .
SEC. 9.580528
Bias -7.516e-05
80 -----
80.. [j .....................
4'55 DmUS_DA035_DCIDL0501-DA150—DCU
@1369 1:5-(5C0 ‘ DL03— DA075- DCO
DL02- DA15ELDCO
4U- ....................
20.. .......................................................................
ICO
DL00-DA035DCO .
0— ‘ DL000—DA007- DCO .
Measured Y
I...I..I...I...I...I...I
0 20 40 60 80 100 120

Figure A5.6: MLR 72-hr model using DRIFT and Raman data

189

 

 

 

 

 

  
  
 
 

 

 

 

 

 

 

 
 
 
 
 
 

 

:n Predicted Y IIIIIII
W Elements 115
Slope. 0.774790
Offset 3 386339 ‘
4g- Correlatlon. 0 860222 .... . , , . . . .‘
RMSEC. 4 375347
SEC 4394990
Bras -9 412e-07
30
2D _. ......................................
10.. ......................................................
0-
_ 1 0 _ ..........................................................................................
Measured Y
I ......... I ......... I ......... I ......... I ........ rIfi I .41 . I . . I
-10 0 10 20 3D 40 50
Figure A5.7: MLR initial rate model using DRIFT and XRD data
1"“ pm!” Y ...............................
‘” Elements 115 ;
Slope 0.740230 ‘ DL05DA55-0C3
Offset 19 90908 -. I’ .
100_ Correlatlon 0 860366 . ..........
RMSEC’ 12 44555
SEC 12 501 12
Bias -7 878e-07
BU .............................................
30 _ .................

   

 

 

 

 

I0 ' .
20.1 .......... iDLIUT-DAUUU-DCU"; .......... ........
0 _
Measude
I I Li, I f A I I I . I . . . I I . . I
0 20 40 60 80 100 120

Figure A5.8: MLR 72-hr model using DRIFT and XRD data

190

Predicted Y

 

 
 
 

" Elements 115
Slope. 0 746730
Offset 3 785764 : : :
40_Correlation 0.864136 ...........
RMSEC' 4.640451 - - - -
SEC 4.660759
8135‘ -2 3638-06

 

 

 

 

 

 

 

 

Figure A5.9: MLR initial rate model using DRIFT and fluorescence data

 

 

 
 
 
 
 
 

 

 

 

 
    
 
 
 
     

 

 

 

m PredictedY
‘" Elements 115
Slope. 0.736311
Offset” 19 19433 I . I
1gg_Correlation. 8858885.. . ,, .....
RMSEC 1254018 . ~ 3 .
SEC 12.59506
BIBS 61958-06
an I .....................................
30- .................................... DL1].I
I: ' 4 5-3 'I E
,. a “ . out—DA'IQLQRfiAlSO-DCD
40- ........... 0. . ‘ _ .. _ 0 ....... I .............. Z .............. I .....
20 —DA00—DCO ? : : : ;
_ .......... .- ULIUT—DAUUU-UCU . . E .............. I .............. 3 .............. . .............. I
0_ .
Measde
I I..III..I...I.III.I,I
0 20 40 60 BO 100 120

Figure A5.10: MLR 72-hr model using DRIFT and fluorescence data

191

 

 

 

 

 

 
 
 

 

 

 

 

  
 
 

 
  

 

 

   
 
   

 

 

 

 

m Predicted Y _
”U Elements 115
Slope 0.659338
Offset 5 092035
40- Correlatnon 0 8118-48
R 5 302453
SEC 5 408010
BIaS -3 70213.05
30 . . I . ..................................................
20.. ...........
10_'II,, ..
0- 74 ‘
_10_ ..........................................................................................
Measured Y
IIII II .I. I ......... I ......... I he- I IIIIIIIII I
.10 0 10 20 30 40 50
Figure A5.11: MLR initial rate model using Raman and XRD data
1““ pmed Y ........................................................................
'W Elements 115
Slope 0.701787
Offset 21 70727
Correlation; 0.837720
00— RMSEC' 13.33014 000c0 ----------------- QM
SEC: 13.39450
Blast -6.859e—05 ‘
DL03—DA0150DCO ~
50- ......................................................... t DL03DAQ§5CJCO ................
~ DL03+DA035-DCO :
- . DL001-QA150000
0_ ............................................................ '.D1-1].3TDAQ.7.5TDC.U ............
4 omz-omsooco ~
DL001 DA035—DCO Z 3 5
20_ ............. DL001WW§WI5DCQ.....,....,:.....,...........: ..................
DL00-DA055—DCO ' ' ’
-DLmommmo' 0
0 _—'D.L[1EEDAD75-DCD
Measured Y
0 2'0 4'0 8'0 0'0 100

Figure A5.12: MLR 72-hr model using Raman and XRD data

192

Predicted Y

 

Elements 115
Slope 0.387086
Offset 8 01 1835

Correlation 0 630145
40- RMSEC 7 158738
SEC 7 181074
Blas -1 061e-05

 

 

 

 

 

 

Figure A5.13: MLR initial rate model using Raman and fluorescence data

 

 
 
 
  
 
   
  

 

 

4 nn pmw Y ........................................................................
'W Elements: 115

Slope: 0.686039

Offset 21.88010 _ ,

Correlation: 0.835513 j '
80- RMSEC' 1341846 98000 ----------- ; ------ {QM :; _

sec, 13.47710 ~ , :

3'35 '6'5876'05 . 01.03074000000 DL03-DA0150-DCO ~

00- ................. ................ 1...,:DL03-DA055-E1C0...............;

c. DLU3EDA035DC“ - mom—014150000

§ "..I .- [ecu . ; §
40_ ................. ; ..... ................. DLWQJQEB’EBCU

- DL001-DA035—DCO

   

 

 

 

 

204 ............ amounmmimco...,..,.....3 ................. .................
. [own—011055000

I] ' . 0 .

Measured Y

I . . . I . I . I f . I I . I . I . . . I

0 20 40 80 00 100

Figure A5.14: MLR 72-hr model using Raman and fluorescence data

193

Prembfed Y

 

   

"° Elements 115
Slope. 0.776068
Offset 3 316878

40- Correlation. 0.662107 . ........
4 343455
SEC 4.362463
Bias 41 621e-05

  

 

 

 

 

 

 

 

Figure A5.15: MLR initial rate model using DRIFT, Raman and fluorescence

data

 

 
 
 
 
 
 

 

 

 

     
 

 

 

 

1'1“ pmfwy . ........................
"" Elements 115

Slope 0647655

Offset 1108833 '
100- Correlation 0820572 . . I I. .. .‘ ................

RMSEC 8538028 1

SEC 8.578770

Bias -7 4839-05
80 ......
60‘ ........... I0 . i55—DCO .........................

n 55 DmDS-DA035-DCE'LOU" -DA150- DCO
619 (I300 DL03- DA075 DCO
--DL02 DA150DCO
40- ........................................................................................
20- ......................................................................
ICU
-DL00— DA035—DCO
0- L -DL000—DA007 DCO .
MeasuredY
IIIIIIIII..I.I.IIfiIIfiIT
0 20 40 60 60 100 120

Figure A5.16: MLR 72-hr model using DRIFT, Raman and fluorescence data

194

Predicted Y

 

W Elements 115
Slope 0.658162
Offset 5 084360

40. Correlatxon 0 811867 .
R 5 382208
SEC 5 405765
8135 -3 8130-05

 

 

 

 

 

 

 

 

 
 
 
 
 
 
 

 

 

   
  

 

 

 

 

 

 

_1 U _ .........................................................................................
Measured Y
.10 0 1'0 2'0 3'0 4'0 5'0
Figure A5.17: MLR initial rate model using XRD, Raman and fluorescence data
4 nn pmed Y ........................................................................
'W Elements: 115
Slope 0.704212
Offset. 21 .53077
Correlation: 0.838180
00- RMSEC: 13.20120 D-DCD ------------------ 41889628
sec: 13.33940
Bias -7.631e-05
DL03-DA0150 DOD
and ......................................................... PLOW/595.5? F199 ................
lIDL032— DA035—DCD .
DL001-DA150—DCO
40- ............................................................. DL03DADEDCD ............
DL02-DA150—DCO
OL001 DA035—DCO i f :
205 ............. mommmzwuefimww
f-DL00 04055000 3 3 3
We!)
a _ 4.1—ll mumnm
Measured Y
I fi" ' T ' ' l I I '
0 20 40 00 00 100

Figure A5.18: MLR 72-hr model using XRD, Raman and fluorescence data

195

 

  
 
 
 
  

 

 

 

 

 

 

 

 
 
 
 
 
  
 

 

 

  
   

 

 

 

 

m PredicfedY I” VVVVVVV
“U Elements 115
Slope 0.806200
Offset 2 888778 . .
40_Correlat|on, 0.898031 ., ..i . .. ................................
R 4.056516 i . : I
SEC. 4.074271
Has -5 207e-05
30
20- ..............
1U- ............................................................
0.1
MeasuredY
-10 0 4'0 5'0
Figure A5.19: MLR initial rate model using DRIFT, XRD, Raman and
fluorescence data
1'1“ p(ed£{9dy .........................................................
"" Elements 115 '
Slope 0.848416
Offset 10 88114
100- Correlation 0821660 ......
R 9475164
SEC 8516831
Bras -6 066e-05
BU .........................................
80.7 ..........................
03-DA035—DCI
DL03 DA075 .DCO
40-
20* .
0001196888768?”
0" , DL000-DA007-DCO - -
MeasuredY
IIIIII I.II.IIIII.IIIII
0 20 40 80 60 100 120

Figure A5.20: MLR 72-hr model using DRIFT, XRD, Raman and fluorescence
data

196

Predr'cfed Y

 

30 Elements 114
Slope. 0.266436
Offset 1083272 ‘ 1 '
4g_Correlanon. 0518110....... ......... . ..........
R : 7821141 1 - i I
SEC: 7.856112
Bras 3 6810-07

 

 

 

 

 

 

 

 

 
 
  
 
   
  
 
 
 

 

 

OWNWMDCU
[J—
. DL001£DA150-DCO
: - DL00—DA150—DCO _ . ,
40- .............. ...............
MeasuredY
......... I IIe. IfiIII..4 III
—10 0 10 20 30 40 50
Figure A5.21: PCR initial rate model using Raman data
1““ prEd’C(8dY . ..............................................
"’u Elements 114
Slope 0.686340
Offset: 22 63646 I _ I
30- Correlanon 8 828653 . . ,-.DL080-DA08.DC8 . . . . . . . . 737:
RMSEC 13 05053 g : mm; .
SEC 13.71004 . ;
Blas -2 1338-06 . I DL03-DA075 ch ‘
30 " ' ‘ ' -| 000' DL03-DA01‘50-DCO" """" .
0C0 OL03 BMW—E? C0
- ............... 0......
40 DL03~D A01 socfiDL03-DA055DCU :

0L0 . I
. . , :DL0000—DA007- DCD ..... : .............................................

DL001W60-DCO : . .
~ DL00-DA055—DCO

n. An n.nqp nan

0‘ , ' uLuu-UAU73~LJ\.U

~ DL00-DA035—DCO

 

Measured Y
r . . . I I . I . . I
-20 0 2D 48 60 80 100

 

 

Figure A5.22: PCR 72-hr model using Raman data

197

 

   

 

    
  

 

 

 

 

 

 

 

 

 
 

 

 

 

 

En pwd’CIedY . . . ...................
" Elements 114
Slope 0.880275
Offset 5 078980
Correlation. 0.812573 ' - '
4U-RMSEC 5397908 ................. .......... ......
SEC 5 421741 ' ' '
8105 5 647e-07
30"
20—1 .............
10- .......
0-1 1
Measured Y
IIfi—rII. II .Ifi—firw IIIII4—firII—I—IIIfIIrII
0 10 20 30 40 50
Figure A5.23: PCR initial rate model using XRD data
4 4" PIEd’CYw Y ....................................................
'7" Elements 114
Slope 0.551411 : . ‘, ;
Offset 32 58195 i i j ‘ DL50—DAO—DC8
1207C0N6|atl0n 0742570. .......... I. II. '.51_B ................
v 10 30770 f ; .- g
SEC 18 48013
100- 8105 -5 1110-06 ~~~~~~~
80.. ..........................................
L
8
60 e' H I!" 0
@WDCU
4D- ....... . u __ . ................................. .........................
DL000-DA007-DCO
20.. ................................
0‘ .
Measured Y
' 2'0 40 0'0 0'0 100 120 140

Figure A5.24: PCR 72-hr model using XRD data

198

 

 

 

 

 

    

 

 

 

 

 
 
 
 
 
 
 
   

 

 

 

 

 

 

En pred'CIed Y ...........................
W Elements 114 VVVVVV
Slope. 0 162267
Offset 12 51007
Correlation 0 402648
40~ RMSEC 0470307
SEC 0.513700
Bias 4 141e-07
30.. ................. ................. ................. C"Dt01'-DA1'5U—DC8I
0L0? WQ§A75 oca
DL02.O_A55DC
20— - .
10—
0“ I
Measured Y
0 1'0 2'0 3'0 4'0 5'0
Figure A5 25 PCR 1mt1al rate model usmg fluorescence data
"In pmted Y .....................................................................
"’ Elements 114
Slope 0.273005
Offset 5270050
100- Correlation. 0.522585 I . . ............... _‘ .1
MSEC 20 00000 1153
SEC 20.05300 .-
8135 -1 4640-06
80 ................................
DL00Il-1DIIAI7-III,
80.. ..........
411 ..........................................................................................
20., ..........................................................................................
0 .
Measured Y
0 2'0 4'0 0'0 00 100 120

Figure A5.26: PCR 72-hr model using fluorescence data

Predicted Y

 

"” Elements 114
Slope. 0.268438
Offset 10 83272 .
40-Correlatlon. 0.518110...,...... ............ .........................
RMSEC 7 821 141 . . .
SEC 7.856112
Bias 3 681e-07

 

 

 

 

 

 

 

 

 
 
  
 
  

  

 

 

   
  
 

 

 

 

 

0 DL001EQW90—DCO
. OL001iDA150—DC0
; . DL00—DA150-DCO . _ .
-1U" ? ........................................... ...............
Measured Y
.10 ['1 1'0 2'0 3'0 4'0 5'0
Figure A5.27: PCR initial rate model using Raman and DRIFT data
1nn prEdmted Y I ......................................
'W Elements 114
Slope: 0.666340
Offset 22 03040 .
3']. Correlation. 0.828663 . . . .~.DL000-DA00-DCU I . . I . . . . ..., II'It-"I .
RMSEC 13 05053 g . 11111113111101: ‘ ‘
SEC 13.71004 _ . .
Bias -2 0660-06 I l u _ -_ A87 0 .
00 : - . ., - . - . - . DL02 OA01“:’:"i‘a' 105:0 - :- DCfl owgggogogctéc' ' ' ‘ ' ' ' ' I
. . . DCO- omjsoflaggdfif 50900
-3 ....................... . .1 , n-0A075OC0 ........... ,
4U : . : DL03 DA015DC0'3L'33 DABSWC" :
DC '
DLfl1- WE . :
.....I...-.DL0UDU-DAUU7.DCU ..... ..............................
m101100350600C0 : : -
. - DL00—DA055-DCO
0' 'DLUfi-DHG75-BLU
- DL00—DA035-DCO
1 l fi—l f 'I—'i I ' '
-20 0 20 40 00 00 100

Figure A5.28: PCR 72-hr model using Raman and DRIFT data

200

Predicted Y

 

 
 
 
  
   

 

 

 

 

 

 

Elements. 114
Slope 0660380
Offset: 5075248
Correlation 0.012644 ‘ - ‘ ~
4D‘RMSEC 5396998 ........ .................
SEC 5.420627 ‘ - -
Bias 4.788e-07
30.1 'DL50-Dma-DAfl-DC8
. ll.‘ . 3
-- . SE
20_ .............. DL-Ul DA15U— DCB
D-LlU-DAIS-DCB
10‘ DL10-DA550C8 .......
U-
MeasuredY
, ......... I 3 I ...... , ......... 1 Y . ,
0 10 20 30 40 50

Figure A5.29: PCR initial rate model using XRD and DRIFT data

120 ........ l . ‘DtSD‘DAO'DCB ......

‘ DL50§DA150006

    
 

  

. 0 712813
Offset 20 85887

 
 
     
 
   

 

100 Correlanon 0844282. .
1311228 - . 1 ' 1

SEC 13.17017 ‘
-1281e-05 ‘ : . ‘ »
30 _ ’ I g. . ‘ . ....... I

sou _ ......... .

40- ...... . .. . ,ii'rggg ......................... .............. g

. 01.039151885861150 DC”

 
 

DL00—DA150- DCU

 

 

Measured Y
20 40 60 00 100 1:50

 

1
‘1

Figure A530: PCR 72-hr model using XRD and DRIFT data

201

 

 

 

    
  

 

 

 

 

 

      
 
  

 

 

 

 

 

 

m PredictedY
“U Elements: 114

Slope 0179545

Offset 12.26114 _

Correlation. 0 423728 ‘ - ' -
40-RMSEC 8388583"""“"""":' AAAAAAAAA 4

SEC 8425629 ‘ ‘ ‘ ‘

Blas 3.832e-07 . ; : ' DLD1—DA1SU-DCB:
30- .................................. ................. ................. .
20- . .. : I‘iIi, I' ,. ... ' .. . .......

‘ . :- Ww-Dcs i
8 I i
9001100? ...... 5 ............... 1
0- ,
MeasuredY
, ......... I ......... , ......... l ......... I . v fir I
0 1U 2U 30 40 50
Figure A5.31: PCR initial rate model using DRIFT and fluorescence data
1‘" prEd’CfedY ............................................
H” Elements 114 .
Slope 0.478839 1 I ‘ I
Offset” 37 85299 i I 7' DL50—DA150—DCU .
120- Conelatlon 0691982 . .. . . ............................

1122113 W200 . .
1001 3138 -1 214305 .... .,: ............ 3 ............ : ..... -.DL50.DA7SD.C3 ..... E .......... 3
80“ . . .. ............... " .......
60- . 4 _ ..... . _' ‘ ' . ....... '. .............
40‘ . . .. ............ ............
QUT . . . . ................................

0- .
MeasuredY
0 120 1710

Figure A532: 4PCR 72-hr 6model using UDRIFT and] fluorescence data

202

 

   
 
 
 
 
 
   
   

 

 

 

 

 

 

 

 

 

 

 

 

 

m Predicted Y
‘N Elements 114
Slope. 0548042
Offset 3 754221
40. Correlation 0.740290
RMSEC 8.228031
SEC 5 253513
Blas 1 2382-08
30 ...............................................................
20.1 ,,,,,,,
10.4 ...................................
U- ;__
_1u_ ............... ............. ...............
Measured'Y
.1'0 ['1 1'0 2'0 3'0 4'0 5'0
Figure A533: PCR initial rate model using Raman and XRD data
40“ pmedy ........................................................................
'W Elements 114 : j t i I
Slope. 0.728943 ~ ‘
Offset: 13 83268 : : . _
304 Correlation. 0.852610 . . _ "-DLUODAEJAOUDCO , . . . . . L .............. ' “a. ,_.. _
RMSEC‘ 12.73553 3 g 561% '- .- " ‘
SECI 1284203 : : I
B' : 22137 -05 , ‘ - -
60 1as E ................. fDLU3—IIIIICU ..........................
, . - I‘ '
. . OLD? “0.15105 , »-..,. - Wm 011153000.
: : DL03-DA055-DCD
40- .......... ?t9599?99°..DL.U&DAoa&DCU
DL03-DA075DCU ;
. 0 ...........................................
_ DL00—DA035-DCO . . ,
I l I l l l ' ' l
-20 0 20 40 80 30 100

Figure A534: PCR 72-hr model using Raman and XRD data

203

 

 

 

 

      
  

 

 

 

 

 

 
 
 
 
 
 
 

 

 

 

    

 

 

 

 

:n prEdCIed Y .......................
"u Elements. 114
Slope: 01420040
Offset' 8.687123 . 1 :
4geCorrelat1on 0.848105.” 1 ..............
RM 7052793 7 . . 1 .
SEC 7 033930
Blas‘ 3 38813-07
30 ......................................... .............................
5 DLD1-DA150-DC -
20' ...... Wm" ' ". BS-DCB
. IE): 3
; . 9901—131135ch 13 - ;
10.. .............. , . ........... 8g ...................
/ DL01 DALEJngéh gag: :
U ' 'DLfino-DAG IS-DLU
°-DL001 DA—150 000
-—DL00 DA15D-DCD
40- .....................................................................
Measured Y
, ......... , ......... , 1 , . ..... , . . , ......... ,
-10 0 10 20 30 40 50
Figure A535: PCR initial rate model using Raman and fluorescence data
1nn Predwfed Y . ......
'W Elements 114
Slope 0 749953 3
Offset 1316140 - - . A
30- Corretatmn 0 885998 ....................... , , .............. ‘ J 7 . 1. 83
R ' 12235013 ' . . .. .
SEC 12.20909 BLOOD QA‘JE’DCU
Blas 31828-08
an . ........................ ‘ .....
E -DL02-DA015§D 3'
: ’ 7'A035 000- -DL03-DA0350CU
4U .. ............................................................... DL03.DA055_DCU ...........
20.. ........... >"DLUUUDMJRWDT‘DCOU'”(H-"”1 .............. ..............
/ DL001 DA035 DCU . : .
0- ' ‘ DLD%WWDCU
' - DLDQ-DADBS—DCU
_20_ .........................................................................................
Measured Y
I r 1 ' l I I l ' ' I
20 0 20 40 60 90 100

Figure A536: PCR 72-hr model using Raman and fluorescence data

204

Pris-dialed Y

 

U1
C3

Elements: 114 I I I
Slope 0.859966
Offset 5.081582 , . 2
Correlation, 0012383 - - j

40.. RMSEC 5 400365 . . V . , , V . . . . V . . . . . . . , .............. ................. I
SEC 5.424200 ' ' -
Blas: 52298-07

 

 

,,,,,,,,,,,,, EH.71....rhr.,..:.,DL50._DW°_DAD_DCB.
U

 

 

 

 

 

Figure A5.37: PCR initial rate model using XRD and fluorescence data

 

   
   

 

 

 

 

 

 

1flfl prEdhfed Y , . i . . . . , i ........................
‘” Elements 114 : .
Slope 0 230840 ,
Offset 5598587 . A
100- Correlatlon, 0490459 .................... P" ' .l $006
R 21 45373 : : , -
SEC 21 55347 I ‘. ‘ DLU2II','.
8135 -l 885e-06 - ~
90 , .....
BU _ .......
4U .. ..................
20- ........................................................................................
0.
Measured Y
. . . , . . . 1 . . . , . , . , . . . , . . . 1
0 20 40 60 90 100 120

Figure A5.38: PCR 72-hr model using XRD and fluorescence data

205

Predicted Y

(.71
C3

 

Elements

Slope

Offset

0 _ Correlaoon
R

A

SEC
8135

114

0 349390

9 722937

0 591092

7 470031

7 503012
4 1418-07

 

 

 

 

 

 

 

 

 

 

 
 
 
 
 
 
  
 

40.7 ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, V , V V .
Measured Y
, +51, V .,fi, 'fi TELL . , ......... , ......... , 111111111 ,
—10 0 10 20 30 40 50
Figure A539: PCR initial rate model using DRIFT, XRD and Raman data
‘ fin pmed Y ........................................................................
'W Elements 114
Slope. 0.726943
Offset 19 93269
30.. Correlation 0.052310 . t . a.DLUDQ DAOO-DCO .....................
R 12 79563 315069631651
sec, 12 94209
13‘ . -2 234 -06
50 IaS e .............................. IDLU3-‘HI ICU ...........................
, ,_» '.-.- - ' .mozomsooco
: DL03—DA055- 000
40- 270M35DC0- DL03 DA035—DCO

    

 

 

 

 

0
I DL00—DA035—DCO
' I l I ' I I ' I
-20 0 20 40 80 80 100

Figure A5.40: PCR 72-hr model using DRIFT, XRD and Raman data

206

 

 

 

      
 
 
 
  
 
  
   
 
 

 

 

 

 

 

 

 

En predmed Y . ...............................
“u Elements 114 ...................
Slope. 0.420040
Offset El 887123 Y i 1
40_Correlanon, 08-48105.., ..............
RMSEC 7 052793 i » : t .
SEC. 7.003930 I
BIBS 3 3888-07 I : : I I
30 I .......................... ............................. ..............
' ' A f DL01-DA150—DC - ‘
20- I ....................................... 85<DCO , , .
10" / , . . ...........
0' 7 ' DLUS-Ulwnls-DCU
‘ DL001 :DA150-DCO
' DL00—DA150-DCO
40- ..........................................................................................
Measured Y
i 0 0 1'0 2'0 3'0 40 50
Figure A5.41: PCR initial model using DRIFT, fluorescence and Raman data
4““ prEd’C‘ed Y ...........................................................
'W Elements 114
Slope 0 702107
Offset 21.53859 . ,
305 Correlation 0.037910 . . V V . . r : ,,,,,,,,,,,,, 1 ............
R 13 35445 - oLooo-oA00-oc0
SEC 13 41341 _ ,
BIBS 4.442e-06 ; ' , ' ‘. 3
50 . . . . ' 7‘ ."'."" _ ;:.' _ .. ,. - ........ i
40- ............................. 'I-"I .._

DL001— DMDOB055DCU

 

 

 

 

18839995370300
0‘ .
~ om0-oA075-oco
; ~DL00-DA035—DCO I : I :
-20 cl 2'0 4'0 6'0 90 100

Figure A5.42: PCR 72-hr using DRIFT, fluorescence and Raman data

207

-a

Predicted Y

 

 

C3
C3

 
 
 
 
 
 
  

 

 

Elements: 1 14

Slope. 0.702136

Offset" 21 63445
30 _ Correlatlon 0.837936 ..................

R 13 35379 DL000— voooco

SEC 13 41275

Bias 40748-06
60 ........
4U .. ........................................................................................

   
 
 

DL001- ammooaoéeoco

 

 

 

 

 

 
 
 
 
 
 
 

 

 

: 61660636996350
0 - .
- DLO0DA075-DCO
- DL00-DA035—DCU
an - ..........................................................................................
Measured Y
1 ' l V 7 1 ' 1 1 1 I
-20 0 20 40 60 00 100
Figure A5.43: PCR initial rate using XRD, fluorescence and Raman data
1 M1 prEd’CfEd Y . ........................................................
'w Elements 114
Slope. 0 702136
Offset 21 63445
30- Correlanon. 0 037936 ......................
RMSEC‘ 13 35379 DL000- DADQDCU
SEC 13 41275
Blas -4 0746-08
60 ..........
40- ...................................................................................

 

 

     

DL001 omooaoésoco

2 BLBBWQR‘BU

 

 

. oL000A075-oc0
. DL00—DA035—DCO

 

Figure A5.44: PCR 72-hr using XRD, fluorescence and Raman data

208

 

 

 

 

 

 

 

 

 

   
  

   
 

 

 

  
    

 

 

 

 

E prEd’Cfed Y ...................................
J Elements 114
Slope 0 182287
Offset 12 51907
Correlatvon 0 402849 '
4U- RMSEC 8475385 . . . . . . ..... . . . ........ ...............
SEC 9.513798 ' '
Bias 4 2588—07
30- ................ i ................. I ................. : ................. : ~1 DL'Ul-DATSU-DCBE
DLUQ-Wfitfifimeoca
20-
10-
0—
Measured Y
, ......... , ......... , ......... , ......... , ......... ,
0 10 20 30 40 50
Figure A5.45: PCR initial rate model using DRIFT, XRD and fluorescence data
4 An prEdwted Y , 1 .........................
'W Elements 114
Slope 0 479069 ‘ ‘ i 2
Offset 37 93838 T I 3‘ DL50—DA150vDCO
120- Cone|afion 0 692147 .................... ........... ..........................
RMSEC 17 55982 ' 5 DL50—DA55—DCU
SEC 17 73779 , _ . '
1003 8188‘ -1.274e-05 ': .............. . . .-.DL50.-DA75.—D .............
80.4 . V . V . V . , V .............
501 ...... nDLOUO;DAUD.DCU ......................................
w’01_50—oc0
40.. ......................................................................
20_ ......................................
04 '
Measured Y
1'] 2'0 4'0 6'0 6'0 100 120 140

Figure A5.46: PCR 72-hr model using DRIFT, XRD and fluorescence data

209

 

 

 

    

 

 

 

 

 

  
 
 
  
   
  

m Predr'cledY
“U Elements 114
Slope, 0420040
Offset 3987129
40- Correlatlon. 0.648105 . 1 1 . . . ............................. 1 ........... .
R 7052792 3 ' 1
SEC 7.083930
Blas 30539-07
30 .............,.........V....,,,,.V....,..I ..............................
foL01—oA150—oc -
10- / ' ..............................
0' : UUDi#WIWUCU
‘ DL001 jomsooco
. - DL00-DA150—DCO , .
-10— ...............
MeasuredY
....... W. ,,,Y,
-10 20 50
Figure A5. 47: PCR0 initial rate model using DRIFT, Raman, 4XRD and
fluorescence data
inn predmtwy , .................................
'W Elements 1213 I I
Slope 0760109
Offset 17 29721

0- Correlation 0.971941 . .
RMSEC 12 09953

III

 

D-L000 DA00—DCO

 

 

 

 

 

 

SEC 12.14607
8135 3 0036-06
60 .................................................
3 ~ . I - -DA035- 000 3 1
40_ ........................ DL09—DA035-VDCUHUQH.H......V.:
I' ' DL03-DA055—DCO
205 i .............. . . . . . |.—D.C9 .....................................
1 “HELENE? 999cc
; DL00- DA015- DCO
0- ' ‘ 095951979981an
' - DL000A03SDCO
_20_ i..,,,,........ ,..V.V......,...,........,...3..,....,,.,...l..............i ...............
Measured Y
1fi.r,V..,.Vfi¢j.tf..vl
-20 0 20 40 60 80 100

Figure A5.48: PCR 72-hr model using DRIFT, Raman, XRD and fluorescence
data

210

BIBLIOGRAPHY

. Abraham, Mariamma and G. Muraleedhara Kurup. (1997), “Pretreatment Studies of
Cellulose Wastes for Optimization of Cellulase Enzyme Activity”, Applied
Biochemistry and Biotechnology, Vol. 62, 201-211.

. Adar, Fran, Gwenaelle le Bourbon, John Reffner, and Andrew Whitley. “FT-IR and
Raman Microscopy on a United Platform”, Spectroscopy 18(2) February 2003, 34-40.

. Akin, DE, A. Sethuraman, W.H. Morrison III, S.A. Martin and K.E.L. Eriksson,

(1993) “Microbial Delignification with White Rot Fungi Improves Forage
Digestibility”, Applied and Environmental Microbiology, Vol.59, No.12, p. 4274-
4282.

. Akin, D. E., L.L. Rigsby, Anand Sethuraman, W.H. Morrison 111, GR. Gamble and
K.-E.L. Eriksson, (1995), “Alteration in Structure, Chemistry and Biodegradability of
Grass Lignocellulose Treated with the White Rot Fungi”, Applied and Environmental
Microbiology, Vol. 61, No. 4, pp. 1591-1598.

. Akin, DE. and LL Rigsby, (1992), “Scanning Electron Microscopy and Ultraviolet
Absorption Microspectrophotometry of plant cell types of different
Biodegradabilities”, Food Structure, Vol. 11, pp. 259-271.

. Akin, Danny B. and Roy D. Hartley, (1992) “Microspectrophotometry and
Digestibility of Alkali-Treated Walls in Bermudagrass Cell Types”, Crop Science,
Vol. 32,1116-1122.

. Backa, Stefan and Anders Brolin “ Determination of pulp characteristics by diffuse
reflectance FTIR”, T appi Journal, May 1991, 218-226.

. Ball, David .W., “Theory of Raman Spectroscopy” Spectroscopy 16(11) November
2001, 32-34.

. Ball, David W., “Raman Spectroscopy” Spectroscopy 17(2) February 2002, 50-52.

211

10.

11.

12.

. 13.

14.

15.

16.

17.

18.

19.

Beebe, Kenneth R. and Bruce R. Kowalski. (1987) “An Introduction to Multivariate
Calibration and Analysis”, Analytical Chemistry, Vol. 59, No. 17, September 1, 1987,
1007A-1017A.

Belkacemi, Khaled, Ginette Turcotte, Damien de Halleux and Phillipe Savoie. (1998)
“Ethanol Production fi'om AF EX-Treated Forages and Agricultural Residues”,
Applied Biochem and Biotech, Vol. 70/72, 441-462.

Bertran, Maria Silvia and Bruce E. Dale. (1985) “Enzymatic Hydrolysis and
Recrystallization Behavior of initially Amorphous Cellulose”, Biotechnology and
Bioengineering, Vol. XXVII, Pp 177-181.

Beyer, M., D. Steger and K. Fischer ((1993) “The luminescence of lignin-containing
pulps - A comparison with the fluorescence of model compounds in several media” J.
Photochem. Photobiol. A: Chem, 76, 217-224.

Billa, E., E. Koutsoula and E.G. Koukios (1999). “Fluorescence analysis of paper
pulps”, Bioresource Technology, 67, 25-33.

Billa E. and B Monties. (1995) “Molecular Variability of lignin fractions isolated
from wheat straw” Res. Chem Intermed, Vol. 21, Nos. 3-5, pp303-311.

Boudet, Alain-M.. “A new view of lignification”, Trend in Plant Science, February
1998, Vol. 3, No. 2, 67-71.

Buchanan, Bob B.. Wilhelm Gruissem and Russell L. Jones. Biochemistry &
Molecular Biology of Plants, 3rd edition Courier Companies, Inc., 2001, 1367p

Burns, Donald B and Emil W Ciurczak, Handbook of Near-Infrared Analysis 1992,
Marcel Dekker Inc, New York, USA, 681p.

Castellan, Alain, Hasneen Choudhury, R Stephen Davidson and Stephane Grelier.
(1994), “ Comparative study of stone-ground wood pulp and native wood 2.
Comparison of the fluorescence of stone-ground wood pulp and native wood.” J.
Photochem. Photobiol. A.: Chem, 81, Pp. 117-122.

212

20. Castellan, Alain and R. Stephen Davidson (1994) “Steady-state and dynamic

21.

22.

23.

24.

25.

fluorescence emission from Abies wood” J. Photochem. Photobiol. A: Chem., 78, Pp.
275-279.

Castellan , A., V. Trichet, J .-C. Pommier, A. Siohan and S. Armancq, (1995) “Photo
and Thermal Stability of Totally Chlorine Free Softwood Pulps Studied by UV/V IS

Diffuse Reflectance and Fluorescence Spectroscopy”, Journal of Pulp and Paper
Science, Vol. 21, No. 9, J291-J295.

Chan, Christina, Daehee Hwang, Gregory N. Stephanopoulos, Martin L. Yarmush
and George Stephanopoulos, (2003), “Application of Multivariate Analysis to
Function of Cultured Hepatocytes.” Biotechnol. Prog., 19, p. 580-598.

Chang, Vincent 8., Barry Burr and Mark T. Holtzapple. (1997) “Lime Pretreatment of
Switchgrass”, Applied Biochemistry and Biotechnology, Vol. 63-65, 3-19.

Chang, Vincent 8., and Mark T. Holtzapple. (2000) “Fundamental Factors Affecting
Biomass Enzymatic Reactivity”, Applied Biochemistry and Biotechnology, Vol. 84-
86, 5-35.

Chang, Vincent 8., Murlidhar Nagwani and Mark T. Holtzapple. (1998) “Lime
Pretreatment of Crop Residues Bagasse and Wheat Straw” Applied Biochemistry and
Biotechnology, Vol. 74, l35-159.Charles, Nichola, Shawn D. Mansfield, Olga
Mirochnik and Sheldon J.B. Duff, (2003) “Effect of Oxygen Delignification
Operating Parameters on Downstream Enzymatic Hydrolysis of Softwood Substrates”
Biotechnology Progress, 19, 1606-1611.Chesson, Andrew, Peter T. Gardner and
Timothy J. Wood, (1997) “Cell Wall Porosity and Available Surface Area of Wheat
Straw and Wheat Grain Fractions”, J. Sci. Food Agric., 75, p. 289-295.Choudhury,
Hasneen, Stephen Collins and R. Stephen Davidson (1992), “The colour reversion of
papers made fiom high yield pulp-a photochromic process?” J. Photochem.
Photobiol. A.: Chem., 69, Pp. 109-119.Christy, Alfred A., Egil Nodland, Alan K.
Burnham, Olav M. Kvalheim and Birger Dahl, (1994), “Determination of Kinetic
Parameters for the Dehydration of Calcium Oxalate Monohydrate by Diffuse
Reflectance FT -IR Spectroscopy”, Applied Spectroscopy, Vol. 48, No. 5, p. 561-
568.Cullitty, B.D and S. R. Stock. Elements of X-Ray Diffrajgtion. 3rd ed., Upper
Saddle River, NJ: Prentice Hall, c2001Dale, B.E., C.K. Leong, T.K. Pham, V.M.
Esquivel, I. Rios and V.M. Latimer. (1996) “Hydrolysis of Lignocellulosics at Low
Enzyme Levels: Application of the AF EX Process”, Bioresource T echnology, 56, 11-
116.

213

32.

33.

34.

35.

36.

37.

38.

39.

40.

Dale, Bruce E., Justin Weaver and F. Michael Byers. (1999) “Extrusion processing
for Ammonia Fiber Explosion (AF EX)”, Applied Biochemistry and Biotechnology,
Vol. 77-79, 35-45.

Dale, Bruce E, Lisa E. Gollapalli and Douglas M. Rivers. “Predicting Digestibility of
Ammonia Fiber Explosion (AF EX) treated Rice Straw”, Applied Biochemistry and
Biotechnology, 2001

Dale, BE. and MJ. Moreira, (1983), 1982. "A Freeze-Explosion Technique for
Increasing Cellulose Hydrolysis." Biotech. and Bioengr. Symp. #12, "Biotechnology
in Energy Production and Conversation," p. 13.

Davis, Mark W. (1998), “ A rapid modified method for compositional carbohydrate
analysis of lignocellulosics by high pH anion-exchange chromatography with pulsed
amperometric detection (HPAEC/PAD)”, Journal of Wood Chemistry and
Technology, 18(2), pp. 235-252.

Davidson, R. Stephen and Linda A Dunn, (1991) “A study of the phhotobleaching
and photoyellowing of paper containing lignin using fluorescence spectroscopy” J.
Photochem. Photobiol. A.: Chem., 58, Pp. 349-359.

De La Rosa, Luis B., Sultan Reshamwala, Vivian Latimer, Bahha T. Shawky, Bruce
E. Dale and Earnest D. Stuart, (1994), “Integrated Production of Ethanol Fuel and
Protein from Coastal Bermudagrass”, Applied Biochemistry and Biotechnology, Vol.
45/46, p.483-497.

Fan, L.T., Yong-Hyun Lee, and David H. Beardmore. (1980) “Mechanism of the
Enzymatic Hydrolysis of Cellulose: Effects of Major Structural Features of Cellulose

on Enzymatic Hydrolysis”, Biotechnology and Bioengineering, Vol. XXIII, Pp 177-
199.

Ferraz, Andre, Jaime Baeza, Jaime Rodriguez and Juanita Freer. (2000) “Estimating

the chemical composition of biodegraded pine and eucalyptus wood by DRIFT
spectroscopy and multivariate analysis”, Bioresource Technology, 74, 201-212.

Fuller, M.P., G.L. Ritter and CS. Draper. “Partial least squares quantitative analysis
of infrared spectroscopy data Part 1: Algorithm Implementation” Applied
Spectroscopy submitted.

214

41.

42.

43.

44.

45.

46.

47.

48.

49.

Fuller, Michael P., Issam M. Hamadeh, Peter Griffiths and Douglas E. Lowenhaupt,
(1982), “Diffuse reflectance infrared spectrometry of powdered coals”, Fuel, Vol. 61,
529-536.

Fuller, Michael P. and Peter R. Griffiths, (1978), “Diffuse Reflectance Measurements
by Infiared Fourier Transform Spectrometry”, Analytical Chemistry, Vol. 50, No. 13,
pp. 1906-1910.

Gardner, Peter T., Timothy J. Wood, Andrew Chesson and Trevor Stuchbury, (1999),
“Effect of degradation on the porosity and surface area of forage cell walls of
differing lignin content”, Journal of the Science of Food and Agriculture, 79: 1 1-18.

Gharpuray, M.M., Young-Hyun Lee and LT. Fan. (1983) “Structural Modification of
Lignocellulosics by Pretreatment to Enhance Enzymatic Hydrolysis” Biotechnology
and Bioengineering, Vol. XXV, Pp 157-172.

Grethlein, Hans. (1985) “The Effect of Pore Size Distribution on the Rate of
Enzymatic Hydrolysis of Cellulosic Substrates”, Bio/Technology, February 1985,
155-160.

Grethlein, Hans E. “Pretreatment of Biomass for Enzymatic Hydrolysis”, 939-953.

Grethlein, Hans E. and Alvin O. Converse, “Common Aspects of Acid Prehydrolysis
and Steam Explosion for Pretreating Wood.” Bioresource Technologi 36 (1991), 77-
82.

Guettler, Michael V., (1984) “Bacteria and Cellulose”, Manuscript

Guilardi Ruggiero, Silvana, Alain Castellan, Stephan Grelier, Aziz Nourmamode and
Michel Cotrait, (1999) “Luminescence properties and spatial arrangement of lignin
model in the solid 11. Crystal structures of 2,2’-dibenzyloxy-5,5’-di(l-hydroxyethyl)-
3.3’dimethoxybiphenyl and 2,2’-dibenzyloxy-5,5’-dihydroxymethyl-3,3’-
dimethoxybiphenyl and their fluoescence emission.” Journal of Molecular Structure
484, 235-248.

' 215

50.

51.

52.

53.

54.

55.

Guilardi Ruggiero, Silvana, Alain Castellan, Michel Cotrait, Stephan Grelier, Mariza
Guimaraes Drumond and Dorila Pilo Veloso, (1997) “Luminescence properties and
spatial arrangement of lignin model in the solid. Part 1. Crystal structures of 5,5’-
diacetyl-Z, 2’-dibenzyloxy-3. 3’ dimethoxybiphenyl and 2,2’-dibenzyloxy-5,5’-
diformyl-3,3 ’-dimethoxybiphenyl and their phosphorescence emission.” Journal of
Molecular Structure 435, 77-87.

Hall, D.O.; F. Rosillo-Calle; R.H. Williams; J. Woods. 1993 “Biomass for Energy:
Supply Prospects” in Renewable Energy-Sources for Fuels and Electricity. L.
Burham,ed.

Harnes, Bonnie R., Steven R. Thomas, Amie D Sluiter, Christine J. Roth and David
W. Templeton, “Rapid Biomass Analysis: New Tools for Compositional Analysis of
Com Stover F eedstocks and Process Intermediates from Ethanol Production” Applied
Biochemistry and Biotechnology, 2002

Hattori, Tadashi, Kenji Shirai, Miki Niwa and Yuichi Murakami, (1981),
“Performance of an Emisionless Infrared Diffuse Reflectance Spectrometer”, Anal.
Chem., 53, 1129-1130.

Hecht, Harry G., (1983), “A Comparison of the Kubelka-Munk, Rozenberg, and
Pitss-Giovanelli Methods of Analysis of Diffuse Reflectance for Several Model
Systems”, Applied Spectroscopy, Vol. 37, No. 4, p. 348-354.

Henderson, D.O., E. Silberman, N. Chen and F. Snyder, (1993), “Adsorption Kinetics
of EGDN on ZnO by Diffuse Reflectance Infrared Fourier Transform Spectroscopy”
Applied Spectroscopy, Vol. 47, No. 4, p. 528-531.

56. Hespell, Robert. B, Patricia J. O’Bryan, Mohammed Moniruzzaman and Rodney J.

Bothast. (1997) “Hydrolysis by Commercial Enzyme Mixtures of AF EX-Treated
Corn Fiber and Isolated Xylans”, Applied Biochemistry and Biotechnology, Vol. 62,
87-97.

57. Hogan, Charlene M. and Mary Mes-Hartree. (1990) “ Recycle of cellulases and the

use of lignocellulosic residue for enzyme production after hydrolysis of steam-
pretreated aspenwood”, Journal of Industrial Microbiology, 6, 253-262.

216

58. Holtzapple, Mark T, Jae-Hoon Jun, Ganesh Ashok, Srinivas L. Patibandla, and Bruce
E. Dale. (1991) “The Ammonia Freeze Explosion (AF EX) Process”, Applied
Biochem and Biotech, Vol. 28/29, 59-74.

59. Holtzapple, Mark T., Joseph E. Lundeen, Russell Sturgis, John E. Lewis, and Bruce
E. Dale. (1992) “Pretreatment of Lignocellulosic municipal Solid Waste by Ammonia
Fiber Explosion (AF EX)”, Applied Biochemistry and Biotechnology, Vol. 34/35, 5-
21.

60. Holtzapple, Mark T., and Robert Torget. (1997) “Thermal and Biological
Processing”, Applied Biochemistry and Biotechnology, Vol. 63-65, 1-2.

61. Itagaki, Hideyuki and Sanami Kato (1999), “Dynamic process of recrystallization of
poly(ethylene terephthalate) solids revealed by fluorescence spectroscopy”, Polymer,
40, Pp. 3501-3504.

62. Iyer, Prashant V., Zhang-Wen Wu, Sung Bae Kim and Yoon Y. Lee. (1996)
“Ammonia Recycled Percolation Process for Pretreatment of Herbaceous Biomass”,
Applied Biochemistry and Biotechnology, Vol. 57/58, 121-132.

63. Juang, Ring-Hwei and David E. Storey, (1998), “Quantitative Determination of the
Extent of Neutralization of Carboxylic Acid Functionality in Carbopol® 974P NF by
Diffuse Reflectance Fourier Transform Infrared Spectrometry Using Kubelka-Munk
Function”, Pharmaceutical Research, Vol. 15, No. 11, p. 1714-1720.

64. Jung, H.G. and MS. Allen (1995), “Characteristics of Plant Cell Walls Affecting
Intake and Digestibility of Forages by Ruminants”, Journal of Animal Science 73:
2774-2790.

65. Jung, Hans-Joachim G.. (1997) “Analysis of Forage Fiber and Cell Walls in Rurniant
Nutrition” Conference: New developments in Forage Science Contributing to
Enhance Fiber Utilization by Rumiants, 810S-813S.

66. Jung, H.G., D.R. Mertens and A.J. Payne. (1997) “Correlation of Acid Detergent
Lignin and Klason Lignin with Digestibility of Forage Dry Matter and Neutral
Detergent Fiber”, J. Dairy Science, Vol. 80, 1622-1628.

217

67. Kabner, Petra and Manfred Baems, (1996), “Comparative characterization of basicity
and acidity of metal oxide catalysts for the oxidative coupling of methane by different
methods”, Applied Catalysis A: General 139, pp107-129.

68. Kendall, Sir Maurice, Multivariate Analysis 2“d edition, 1980, Macmillan Publishing,
Co., New York, USA, 210p

69. Kerley, M.S., G.C. Fahey, Jr., J.M.Gould and EL. Iannotti (1988), “Effects of
Lignfication, Cellulose Crystallinity and Enzyme Accessible Space on the
Digestibility of Plant Cell Wall Carbohydrates by the Rumiant”. Food
Microstructure, Vol. 7, pp. 59-65.

70. Kim, Tae Hyun, Jun Seok Kim, Changshin Sunwoo and Y.Y. Lee, “Pretreatment of
corn stover by aqueous ammonia” Bioresource Technology 90 (2003 ), 39-47

71. Klass, Donald L. “An Introduction to Biomass Energy:A renewable resource” from
http://wwwberal .org/about.html

72. Klier, K. “Reflectance Spectroscopy as a Tool for investigating Dispersed Solids and
Their Surfaces” Catal. Rev., 1, (1967), 207-232.

73. Klier, K. (1980) “Investigations of Adsorption Centers, Molecules, Surface
Complexes, and Interactions Among Catalyst Components by Diffuse Reflectance

Spectroscopy”, Vibrational Spectroscopies for Adsorbed Species, ACS Symposium
Series #137, p. 141-162.

74. Knappert, Diane, Hans Grethlein and Alvin Converse. (1980) “Partial Acid
Hydrolysis of Cellulosic Materials as a Pretreatment for Enzymatic Hydrolysis”,
Biotechnology and Bioengineering, Vol. XXII, Pp. 1449-1463.

75. Kong, F anran, Cady R. Engler and E(1992) “Effects of Cell-Wall Acetate, Xylan
Backbone, and Lignin on Enzymatic Hydrolysis of Aspen Wood”, Applied
Biochemistry and Biotechnology, Vol. (1 J. Soltes. 34/35, 23-35.

76. Kosflcova, B., L. Zékutné and D. Joniak. (1978) “Investigation of the Lignin-
Saccharidic Complex by Electron Microscopy”, Holzforschung, Bd. 32, 15-18.

218

77. Krishnan, K., S.L. Hill and RH. Brown, (1980), “FTIR spectroscopy using diffuse
reflectance and a diamond cell” American Laboratory, p.106-114.

78. Krivacsy, Z. and J. Hlavay, (1995), “Comparison of Calibration methods in
Quantitative Diffuse Reflectance Infrared Spectroscopy”, Talanta, Vol. 42, No. 4, pp.
613-620.

79. Ladisch, M.R., K.W. Lin, M. Voloch and GT. Tsao, (1983) “Process considerations
in the enzymatic hydrolysis of biomass”, Enzyme Microb. T echnol., Vol. 5, 82-102.

80. Lakowicz, Joseph R. Principles of FlLorescence Spectroscopy. Plenum Press, New
York.

81. Laureano-Perez, Lizbeth, Farzaneh Teymouri, Hasan Alizadeh and Bruce Dale (2004)
“Understanding Factors that Limit Enzymatic Hydrolysis of Biomass:
Characterization of Pretreated Corn Store”, Applied Biochemistry and Biotechnology
(In Press)

82. Lundquist, Knut, Bjorn Josefsson and Gunnar Nyquist.(1978) “Analysis of Lignin
Products by Fluorescence Spectroscopy”, Holzforschung, Bd32, 27-32.

83. Lynd, Lee R., Charles E. Wyman and Tillman U. Gemgross. (1999) “Biocommodity
Engineering”, Biotechnol. Prog, 15, 777-793.

84. Mandels, Mary, (1982) “Cellulases” Annual Reports on Fermentation Processes,
Vol.5, Pp. 35-77.

85. Molecular Fluorescence Spectrometry Sec 15.3

86. Moniruzzaman, M., BR. Dale, R.B. Hespell and RJ. Bothast (1997) “Enzymatic
Hydrolysis of High-Moisture Corn Fiber Pretreated by AFEX and Recovery and

Recycling of the Enzyme Complex”, Applied Biochemistry and Biotechnology, Vol.
67, 113-126.

87. Morrison, I.M. (1974) “Structural investigations on the lignin-carbohydrate
complexes from Lolium perenne”, Biochem J., 139:197-204.

219

88. Morrison, Ian M. and Derek Stewart, “ Determination of Lignin in the Presence of
Ester-Bound Substituted Cinnamic Acids by a Modified Acetyl Bromide Procedure”
JSci FoodAgric 1995, 69, 151-157.

89. Naummann, D. “Infrared and NIR Raman Spectroscopy in Medical Microbiology ” In
press

90. Noutary, C., P. Fornier de Violet, J. Vercauteren and A. Castellan (1995),
“Photochemical studies of phenolic phenylcoumarone lignin model molecule in
relation to the photodegradation of lignocellulosics materials. Part 2. Photophysical
and photochemical studies”, Res. Chem. Intermed. Vol. 21, Nos. 3-5, pp. 247-262.

91. Nunes Pereira, Eduardo J ., Mario N. Berberan-Santos, Aleksandre Fedorov, Michel
Vincent, Jacques Gallay and Jose M. G. Martinho, (1999), “Molecular radiative
transport. III. Experimental intensity decays”, Journal of Chemical Physics, Vol. 110,
No. 3, pp. 1600-1610.

92. Olmstead, J .A. and D.G. Gray (1997) “Fluorescence Spectroscopy of Cellulose,
Lignin and Mechanical Pulps: A Review”, Journal of Pulp and Paper Science, Vol.
23, No. 12 December 1997, J571-J579.

93. Orskov, ER and G.W. Reid, (1988) “Comparison of Chemical and Biological
Methods for Predicting Feed Intakes and Animal Performance” Animal Production,
158-165.

94. Pandey, K.K. (1998), “ A study of Chemical Structure of Soft and Hardwood and
Wood Polymers by F TIR Spectroscopy”, Journal of Applied Polymer Science, Vol.
71, 1969-1975.

95. Pandley, Krishna K. and Dwarika P. Khali, (1998), “ Accelerated Weathering of
Wood Surface Modified by Chromium Trioxide”, Holzforschung, Vol.52, pp. 467-
471.

96. Paralikar, KM. and S. Aravindanath. (1988) “Crystallization of Cellulose”, Journal
of Applied Polymer Science, Vol. 35, 2085-2089.

220

97. Puri, Vinod P., (1984) “Effect of Crystallinity and Degree of Polymerization of
Cellulose Enzymatic Saccharification”, Biotechnology and Bioengineering, Vol.
XXVI, Pp 1219-1222.

98. Reeves, James B. and Barry A. Francis (1997), “Pyrolysis-gas-chromatography-mass-
spectrometry for the analysis of forages and by-products”, Journal of Analytical and
Applied Pyrolysis, 40-41, Pp. 243-265.

99. Russell, J.D., 1. Murray and AR. Fraser “Near and Mid-Infrared studies of the cell
wall structure of cereal straw in relation to its rumen degradability.

100. Samac, Deborah A., Lynn Litterer, Glena Temple, Hans-Joachim G. Jung and
David A. Somers, “Expression of UDP-Glucose Dehydrogenase Reduces Cell-Wall
Polysaccharide Concentration and Increases Xylose Content in Alfalfa Stems”,
Applied Biochemistry and Biotechnology, Vol. 113-116, p. 1167-1182.

101. Sasaki, Takashi, Takashi Tanaka, Noriko Nanbu, Yohko Sato and Keiji Kainuma.
(1979) “Correlation between X-Ray Diffraction Measurements of Cellulose
Crystalline Structure and the Susceptibility to Microbial Cellulase”, Biotechnology
and Bioengineering, Vol. XXI, Pp 1031-1042.

102. Segal, L., J.J. Creely, A.E. Martin, Jr., and C. M. Conrad (1959) “An Empirical
Method for Estimating the Degree of Crystallinity of Native Cellulose Using the X-
Ray Diffi‘actometer”, Textile Res. J. 29, 786 —794.

103. Sewalt, Vincent J.H., Joseph P. Fontenot, Vivien G. Allen and Wolfgang
G.Glasser. (1996) “Fiber Composition and in vitro Digestibility of Com Stover
Fractions in Response to Ammonia Treatment”, J.Agric. Food Chem., 44, 3136-3142.

104. Sewalt, V.J.H., W.G. Glasser and K.A Beauchemin. (1997) “Lignin Impact on
Fiber Degradation. 3. Reversal of Inhibition of Enzymatic Hydrolysis by Chemical
Modification of Lignin and by Additives”, J. Agric. Food Chem., Vol. 45, 1823-1828.

105. Skoog, Douglas A and James J Leary, Principles of In_strumental Analysis 4th
edition, 1992, Saunders College Publishing, Fort Worth, USA, 196-208

221

106. Smyrl, N.R., E.L. Fuller, Jr., and G.L. Powell, (1983), “Monitoring the
Heterogeneous Reaction of LiH and LiOH with H20 and C02 by Diffuse Reflectance

Infrared Fourier Transform Spectroscopy”, Applied Spectroscopy, VoL 37, No. l, p
38-44

107. Stewart, D., H.M. Wilson, P.J. Hendra and I.M Morrison (1995). “Fourier-
Transform Infrared Spectroscopic Study of Biochemical and Chemical Treatments of
Oak Wood (Quercus rubra) and Barley (Hordeum vulgare) Straw”, J. Agric Food
Chem, 43, 2219-2225.

108. Sugiyarna, J ., T. Okano, H. Yamamoto, F. Horii. (1990) “Macromolecules
transformation of Valonia cellulose crystals by an alkaline hydrothermal treatment,
23, 2461-2498

109. Teymouri, Farzaneh, Lizbeth Laureano-Perez, Hasan Alizadeh and Bruce E. Dale
(2004) “Ammonia Fiber Explosion Treatment of Com Stover” Applied Biochemistry
and Biotechnology, Vol. 113-116, 951-963.

110. Thomas, Howard V. and David Haaland. (1990) “Comparison of Multivariate
Calibration Methods for Quantitative Spectra Analysis”, Analytical Chemistry, Vol.
62, No. 10, May 15, 1990, 1091-1099.

111. Thompson David N., Hsin-Chih Chen and Hans Grethlein. (1992) “ Comparison
of Pretreatment Methods on the Basis of Available Surface Area”, Bioresource
Technology, 39, 155-163.

112. Van Every, Kenneth W. and Peter R. Griffiths, (1991), “Characterization of
Diffuse Reflectance FT-IR Spectrometry for Heterogeneous Catalyst Studies”,
Applied Spectroscopy, Vol. 45, No. 3, 347-359.

113. Van Soest, P.J. and RH. Wine (1968), “Determination of Lignin and Cellulose in
Acid-Detergent Fiber with Permanganate”, Journal of the A.0.A.C., Vol. 51, No. 4,
Pp. 780-785.

114. Wallace,G. A. Chesson, J.A. Lomax and M.C. Jarvis. (1991) “Lignin-
carbohydrate complexes in graminaceous cell walls in relation to digestibility”,
Animal Feed Science and Technology, 32, 193-199.

222

115. Wang, Lin, Bruce E. Dale, Lale Yuritas and I. Golwasser. (1998) “Cost estimates
and Sensitivity Analyses for the Ammonia Fiber Explosion Process”, Applied
Biochemistry and Biotechnology, Vol. 70-72, 51-66.

116. Wei, Chia-Jenn and Chung-Yueh Cheng, (1985) “Effect of Hydrogen Peroxide
Pretreatment on the Structural Features nad the Enzymatic Hydrolysis of Rice Straw”.
Biotechnology and Bioengineering, Vol. XXVII, Pp. 1418-1426.Windham, W.R.,
D.R. Mertens and RE. Barton 11. “Protocol for NIRS Calibration: Sample Selection
and Equation Development and Validation” 96-103.

117. Wielage, B., Th. Lampke, G. Marx, K. Nestler and D. Starke, (1999),
“Thermogravimetric and differential scanning calorimetric analysis of natural fibers
and polypropylene”, Thermochimica Acta, 337, pp. 169-177.

118. Willemse, M.T.M., (1988)“Cell Wall Autofluorescence” Acta Botanical
Neerlandica, p.50-57.

119. Willey, RR, (1976), “Fourier Transform Infrared Spectrophotometer for
Transmittance and Diffuse Reflectance Measurements”, Applied Spectroscopy, Vol.
30, No. 6, p. 593-601.

120. Wilson, J.R. and DR. Mertens (1995) “Cell Wall Accessibility and Cell Wall
Structure Limitations to Microbial Digestion of Forage”, Crop Science, Vol. 35, 251-
259.

121. Windham, W.R., D.R. Mertens and F .E. Barton 11. “Protocol for NIRS
Calibration: Sample Selection and Equation Development and Validation” 96—103.

122. Wu, Zhangwen and Y.Y. Lee. (1997) “Ammonia Recycled Percolation as a
Complementary Pretreatment to the Dilute-Acid Process”, Applied Biochemistry and
Biotechnology, Vol. 63-65, 21-34.

123. Wyman, Charles (editor). Handbook on Bioethanol: Production and Utilization,
Taylor and Francis, Washington, DC, USA 1996, 424p

124. Wyman, Charles E. (1994) “Ethanol from Lignocellulosic Biomass: technology,
Economics and Opportunities”, Bioresource Technology, 50, 3-16.

223

125.Yoon, H.H., Z.W. Wu and Y.Y. Lee. (1995) “Ammonia-Recycled Percolation
Process for Pretreatment of Biomass Feedstock”, Applied Biochem and Biotech, Vol.
51/52, 5-19.

126.Bhattacharyya, Gouri K. and Richard A. Johnson Statistical Concepts argl Methods
John Wiley and Sons Inc., New York, USA 1977, p.599

224

IJIKI

   

I I

  
 

“II I III IIII

293 02736 5497