Development of a lab-scale auger reactor for biomass fast pyrolysis ( PDFDrive ).pdf

Graduate Theses and Dissertations
Iowa State University Capstones, Theses and
Dissertations
2009
Development of a lab-scale auger reactor for
biomass fast pyrolysis and process optimization
using response surface methodology
Jared Nathaniel Brown
Iowa State University
Follow this and additional works at: https://lib.dr.iastate.edu/etd
Part of the Mechanical Engineering Commons
This Thesis is brought to you for free and open access by the Iowa State University Capstones, Theses and Dissertations at Iowa State University Digital
Repository. It has been accepted for inclusion in Graduate Theses and Dissertations by an authorized administrator of Iowa State University Digital
Repository. For more information, please contact digirep@iastate.edu.
Recommended Citation
Brown, Jared Nathaniel, "Development of a lab-scale auger reactor for biomass fast pyrolysis and process optimization using response
surface methodology" (2009). Graduate Theses and Dissertations. 10996.
https://lib.dr.iastate.edu/etd/10996

Development of a lab-scale auger reactor for biomass fast pyrolysis and process
optimization using response surface methodology
by
Jared Nathaniel Brown
A thesis submitted to the graduate faculty
in partial fulfillment of the requirements for the degree of
MASTER OF SCIENCE
Co-Majors: Mechanical Engineering; Biorenewable Resources and Technology
Program of Study Committee:
Robert C. Brown, Major Professor
Theodore J. Heindel
D. Raj Raman
Iowa State University
Ames, Iowa
2009
Copyright © Jared Nathaniel Brown, 2009. All rights reserved.

ii
DEDICATION
With gratitude and appreciation, this thesis is dedicated to my wife, Pari. Her love, support
and understanding helped make this project possible.

iii
TABLE OF CONTENTS
LIST OF TABLES v
LIST OF FIGURES vii
ABSTRACT xi
CHAPTER 1. INTRODUCTION 1
1.1 Biomass fast pyrolysis 1
1.2 Thesis overview 2
CHAPTER 2. TECHNICAL LITERATURE REVIEW 3
2.1 Introduction 3
2.2 Fast pyrolysis fundamentals 5
2.2.1 Operating conditions 6
2.2.2 End products description 8
2.2.3 End product utilization 10
2.2.4 Systems technology 12
2.3 State of the art for auger type reactors 20
2.3.1 Fossil fuel processing 21
2.3.2 Biomass processing 26
CHAPTER 3. EXPERIMENTAL APPARATUS 39
3.1 Lab-scale system design 40
3.2 Lab-scale system components 49
3.3 Lab-scale system development 62
CHAPTER 4. EXPERIMENTAL METHODS AND MATERIALS 67
4.1 Introduction 67
4.2 Experimental design 67
4.3 Experimental materials 72
4.4 Testing procedures 75
4.4.1 Product distribution 76
4.4.2 Product analysis 85
4.4.1 Data analysis and hypothesis testing 91
CHAPTER 5. RESULTS AND DISCUSSION 99
5.1 Introduction 99
5.2 Product distribution results 99
5.3 Product analysis results 123
CHAPTER 6. CONCLUSIONS 168
6.1 Research conclusions 168
6.1.1 Regression models 169
6.1.2 Product analysis 170
6.2 Recommendations for future work 172
APPENDIX A. DESIGN AND DEVELOPMENT 175

iv
APPENDIX B. MIXING STUDY 188
APPENDIX C. AUXILLARY EQUIPMENT AND INSTRUMENTS 194
APPENDIX D. EXPERIMENTAL DATA 200
REFERENCES 242
ACKNOWLEDGEMENTS 248

v
LIST OF TABLES
Table 1. Typical physical properties for bio-oil 9
Table 2. Comparison of auger reactor published data 35
Table 3. Biomass feeding system descriptions 50
Table 4. Heat carrier system descriptions 53
Table 5. Reactor system descriptions 56
Table 6. Product recovery system descriptions 59
Table 7. Shakedown trials operating conditions 65
Table 8. Factor considerations for experimental design procedure 68
Table 9. Selected factors and levels for experimental design 70
Table 10. Final central composite design, coded experiments 71
Table 11. Final central composite design, actual experiments 72
Table 12. Red oak biomass composition 73
Table 13. Red oak biomass ultimate and proximate analyses 74
Table 14. Steel shot composition and select properties 75
Table 15. Experimental operating conditions 76
Table 16. Description of symbols used in mass balance procedure 78
Table 17. Gas rotometer settings for experiments 79
Table 18. Non-condensable gas analysis for Run #24 83
Table 19. Thermal gravimetric analysis program method 88
Table 20. Chemical compounds quantified by GC/MS analysis 90
Table 21. ANOVA table 92
Table 22. Critical F-values for ANOVA F-test 93
Table 23. Lack of fit table 93
Table 24. Critical F-values for lack of fit F-test 94
Table 25. Critical t-values for t-test 96
Table 26. Summary of hypothesis tests 97
Table 27. Regression model coefficients and terms 98
Table 28. Sample experimental conditions for 8 selected tests 100
Table 29. Sample mass balance data for 8 selected tests 100
Table 30. Bio-oil yield model, statistics summary 103
Table 31. Biochar yield model, statistics summary 111
Table 32. Carbon monoxide yield model, statistics summary 119
Table 33. Carbon dioxide yield model, statistics summary 121
Table 34. Bio-oil moisture content model, statistics summary 127
Table 35. Water insoluble content model, statistics summary 132
Table 36. Ultimate analysis for biochar at center point tests 138
Table 37. Bio-oil carbon content model, statistics summary 144
Table 38. Bio-oil hydrogen content model, statistics summary 145
Table 39. Bio-oil oxygen content model, statistics summary 147
Table 40. GC/MS characterized compound comparison, whole bio-oil 156
Table 41. Reaction temperature model, statistics summary 162
Table 42. Regression models, summary of statistics 165
Table 43. Regression models, summary of significant terms 166
Table 44. Bio-oil analysis summary and comparison 167
Table 45. Motor power requirements analysis 181

vi
Table 46. Shakedown trial operating conditions 186
Table 47. Shakedown trial yield and operating condition results 187
Table 48. Baseline biomass and sand mixture densities analytical data 192
Table 49. Biomass and sand mixture densities analytical data 193
Table 50. Feedstock and experimental condition data 200
Table 51. Product distribution and mass balance data 201
Table 52. Heat carrier system temperature data and other operating conditions 202
Table 53. Reactor system temperature data 203
Table 54. Product recovery system temperature data 204
Table 55. Bio-oil fraction mass balance data 205
Table 56. Bio-oil yield model statistical data 206
Table 57. Coded levels for model equations 207
Table 58. Biochar yield model statistical data 208
Table 59. Non-condensable gas yield model, statistics summary 209
Table 60. Non-condensable gas yield model statistical data 210
Table 61. Non-condensable gas data, composition 211
Table 62. Non-condensable gas data, molar analysis 212
Table 63. Non-condensable gas data, mass analysis 213
Table 64. Non-condensable gas data, volume meter properties 214
Table 65. Carbon monoxide yield model statistical data 215
Table 66. Carbon dioxide yield model statistical data 216
Table 67. Moisture content analytical data 217
Table 68. Moisture content model statistical data 218
Table 69. Water insoluble content analytical data 219
Table 70. Water insoluble content model statistical data 220
Table 71. Solids content analytical data 221
Table 72. Higher heating value analytical data 221
Table 73. Thermal Gravimetric Analysis data, bio-oil 221
Table 74. Thermal Gravimetric Analysis data, biochar 222
Table 75. Elemental analysis data, biochar 223
Table 76. Elemental analysis data, SF1 bio-oil 224
Table 80. Elemental analysis data, whole bio-oil 228
Table 81. Bio-oil carbon content model statistical data 229
Table 82. Bio-oil hydrogen content model statistical data 230
Table 83. Bio-oil oxygen content model statistical data 232
Table 84. Total acid number analytical data for center point tests 232
Table 85. GC/MS sample analytical data, Run #20 (maximum bio-oil yield) 234
Table 86. GC/MS analytical data, SF1 summary 235
Table 90. GC/MS analytical data, whole bio-oil summary 239
Table 91. Viscosity analytical data 240
Table 92. Reaction temperature model statistical data 241

vii
LIST OF FIGURES
Figure 1. Biomass fast pyrolysis schematic 2
Figure 2. Thermochemical processes 5
Figure 3. Fast pyrolysis product applications 12
Figure 4. Fast pyrolysis subsystem schematic 13
Figure 5. Biomass pretreatment schematic 13
Figure 6. Bio-oil recovery schematic 14
Figure 7. Bubbling fluidized bed reactor schematic 15
Figure 8. Circulating fluidized bed reactor schematic 16
Figure 9. Rotating cone reactor schematic 17
Figure 10. Auger reactor schematic, configuration 1 18
Figure 11. Auger reactor schematic, configuration 2 19
Figure 12. Ablative reactor concept 20
Figure 13. Hayes Process reactor 22
Figure 14. Lugi-Ruhrgas process schematic 23
Figure 15. Screw reactor concept 27
Figure 16. Twin screw mixer-reactor schematic 29
Figure 17. FZK twin screw mixer-reactor 29
Figure 18. Mississippi State University lab-scale auger reactor 32
Figure 19. University of Georgia auger reactor schematic 35
Figure 20. Lab-scale auger reactor system 39
Figure 21. Reactor design schematic 40
Figure 22. Heat carrier mass feed rates as a function of temperature change 42
Figure 23. Various auger flighting designs 43
Figure 24. Volumetric feed rate as a function of screw size and speed 44
Figure 25. Volumetric feed rate as a function of screw speed and configuration 45
Figure 26. Reactor lid drawing with dimensions in inches 46
Figure 27. Auger reactor rendering with lid removed 47
Figure 28. Auger reactor system rendering 48
Figure 29. Biomass feeding system schematic 49
Figure 30. Heat carrier auger drawing with dimensions in inches 51
Figure 31. Heat carrier system schematic 52
Figure 32. Reactor auger drawing with dimensions in inches 54
Figure 33. Reactor system schematic 55
Figure 34. Product recovery system schematic 58
Figure 35. LabVIEW program screenshot for data acquisition and process monitoring 61
Figure 36. Cold flow mixing images of cornstover biomass and silica sand 64
Figure 37. Corn stover (1.0 mm), corn fiber (1.0 mm) and red oak biomass (0.75 mm) 65
Figure 38. Sand, silicon carbide, alumina ceramic and steel shot heat carrier examples 65
Figure 39. Simplified reactor schematic with operational parameters shown 66
Figure 40. Central Composite Design schematic for two factors 69
Figure 41. Red oak biomass samples of three different grind sizes 73
Figure 42. SAE J827 steel shot size distribution 75
Figure 43. Reactor system schematic showing mass balance 77
Figure 44. Mass balance worksheet for experiments 80
Figure 45. Micro-GC gas analysis profile for Run #24 82

viii
Figure 46. Temperature profile example for Run #20 84
Figure 47. Typical bio-oil recovery system temperatures 84
Figure 48. Product distribution results for the 30 fast pyrolysis tests 100
Figure 49. Pyrolysis product distribution range 101
Figure 50. Average operating temperature schematic for 6 center point runs 102
Figure 51. Bio-oil fraction distributions for 6 center point tests and for all tests 102
Figure 52. Absolute values for t-test statistics for bio-oil yield model 105
Figure 53. Actual vs. predicted bio-oil yield 105
Figure 54. Three response surfaces for modeled bio-oil yield 107
Figure 55. Modeled bio-oil yield as a function of heat carrier temperature and auger speed 109
Figure 56. Modeled bio-oil yield as a function of heat carrier temperature and feed rate 109
Figure 57. Absolute values for t-test statistics for biochar yield model 112
Figure 58. Actual vs. predicted biochar yield 112
Figure 59. Two response surfaces for modeled biochar yield 113
Figure 60. Modeled biochar yield as a function of heat carrier temperature and feed rate 115
Figure 61. Modeled biochar yield as a function of heat carrier temperature and auger speed 115
Figure 62. Average non-condensable gas composition at center points 117
Figure 63. Carbon monoxide and carbon dioxide yields vs. bio-oil yield for all tests 118
Figure 64. Actual vs. predicted carbon monoxide yield 120
Figure 65. Gas yields for 4 different species vs. bio-oil yield for all tests 122
Figure 66. Total non-condensable gas yield vs. bio-oil yield for 29 tests 122
Figure 67. Typical appearance of bio-oil fractions 123
Figure 68. Bio-oil moisture content at center points 124
Figure 69. Bio-oil moisture content range 124
Figure 70. Bio-oil moisture content vs. bio-oil yield for all tests 125
Figure 71. Absolute values for t-test statistics for moisture content model 127
Figure 72. Actual vs. predicted moisture content 128
Figure 73. Response surface for modeled moisture content 128
Figure 74. Modeled moisture content as a function of heat carrier temperature and auger speed 129
Figure 75. Water insoluble content for center points 130
Figure 76. Water insoluble content range 131
Figure 77. Modeled H2O insoluble content as a function of heat carrier temperature and feed rate 133
Figure 78. Water insoluble content vs. bio-oil yield for all tests 133
Figure 79. Actual vs. predicted water insoluble content 134
Figure 80. Solids content for center point tests 135
Figure 81. Higher heating value range 136
Figure 82. Biochar proximate analysis for center point tests 137
Figure 83. Bio-oil carbon content for center points 139
Figure 84. Bio-oil nitrogen content for center points 140
Figure 85. Bio-oil hydrogen content for center points 140
Figure 86. Bio-oil sulfur content for center points 141
Figure 87. Bio-oil ash content for center points 142
Figure 88. Bio-oil oxygen content for center points 143
Figure 89. Modeled bio-oil H content as a function of heat carrier temperature and feed rate 146
Figure 90. Actual vs. predicted oxygen content 147
Figure 91. Biochar and non-condensable gas yield vs. bio-oil yield for 29 tests 149
Figure 92. Bio-oil C, O, H, H2O and water insoluble contents as a function of yield for 30 tests 149
Figure 93. Bio-oil H:C ratio vs. O:C ratio (Van Krevelen diagram) for all 30 tests 150
Figure 94. C, O, H, H2O and H2O insoluble contents as a function of yield for 30 tests, dry basis 151

ix
Figure 95. Bio-oil H:C ratio vs. O:C ratio for all 30 tests, including dry basis analysis 152
Figure 96. Total acid number for center points 153
Figure 97. GC/MS chromatogram for SF1, Run #20 (bio-oil max yield) 154
Figure 98. GC/MS chromatogram for SF4, Run #20 (bio-oil max yield) 155
Figure 99. GC/MS quantified volatile compounds 157
Figure 100. GC/MS quantified volatile compounds by fraction for center points 158
Figure 101. Viscosity measurements for Run #20 vs. time 159
Figure 102. Bio-oil viscosity range 160
Figure 103. Reaction temperature schematic 161
Figure 104. Vapor temperatures vs. heat carrier temperatures 161
Figure 105. Actual vs. predicted reaction temperature 163
Figure 106. Absolute values for t-test statistics for vapor temperature model 164
Figure 107. Modeled vapor temperature vs. heat carrier temperature 165
Figure 108. Recommended system design modifications 173
Figure 109. Heat carrier residence time as a function of auger speed 180
Figure 110. Biomass feeding system 182
Figure 111. Close-up of reactor augers 182
Figure 112. Reactor mounted on frame 182
Figure 113. Reactor lid and thermocouple detail 183
Figure 114. Gas cyclone 183
Figure 115. Condensers 1 and 2 (SF1 and SF2) 183
Figure 116. Electrostatic precipitator (SF3) 184
Figure 117. Condenser 3 in ice bath (SF4) 184
Figure 118. Reactor system detail 185
Figure 119. Biomass and sand mixture densities 188
Figure 120. Mixture density (L) vs. auger speed at three axial locations, Run 1 190
Figure 121. Mixture density (L) vs. auger speed at three axial locations, Run 2 190
Figure 122. Mixture density (C) vs. auger speed at four axial locations, Run 1 191
Figure 123. Pentapycnometer instrument 191
Figure 124. Mixture density (C) vs. auger speed at four axial locations, Run 2 192
Figure 125. Hammer mill 194
Figure 126. Knife mill 194
Figure 127. CHN/O/S analyzers 195
Figure 128. Thermal gravimetric analyzer 195
Figure 129. Bomb calorimeter 196
Figure 130. Moisture analyzer 196
Figure 131. Micro-GC cart 197
Figure 132. Gas volume meter and pressure gauge 197
Figure 133. Moisture titrator 198
Figure 134. Total acid number titrator 198
Figure 135. GC/MS 199
Figure 136. Viscometer 199
Figure 137. Residuals for bio-oil yield full model 206
Figure 138. Residuals for biochar yield full model 208
Figure 139. Residuals for non-condensable gas yield full model 209
Figure 140. Residuals for carbon monoxide yield full model 215
Figure 141. Residuals for carbon dioxide yield full model 216
Figure 142. Residuals for moisture content full model 218
Figure 143. Residuals for water insoluble content full model 220

x
Figure 144. Residuals for bio-oil carbon content full model 229
Figure 145. Residuals for bio-oil hydrogen content full model 230
Figure 146. Predicted vs. actual hydrogen content 231
Figure 147. Residuals for bio-oil oxygen content full model 231
Figure 148. GC/MS chromatogram for SF2, Run #20 (bio-oil maximum yield) 233
Figure 149. GC/MS chromatogram for SF3, Run #20 (bio-oil maximum yield) 233
Figure 150. Quantified mass for all runs 240
Figure 151. Residuals for reaction temperature full model 241

xi
ABSTRACT
A lab-scale biomass fast pyrolysis system was designed and constructed based on an auger
reactor concept. The design features two intermeshing augers that mix biomass with a heated bulk
solid material that serves as a heat transfer medium. A literature review, engineering design process,
and shake-down testing procedure was included as part of the system development.
A response surface methodology was carried out by performing 30 experiments based on a
four factor, five level central composite design to evaluate and optimize the system. The factors
investigated were: (1) heat carrier inlet temperature, (2) heat carrier mass feed rate, (3) rotational
speed of the reactor augers, and (4) volumetric flow rate of nitrogen used as a carrier gas. Red oak
(Quercus Rubra L.) was used as the biomass feedstock, and S-280 cast steel shot was used as a heat
carrier. Gravimetric methods were used to determine the mass yields of the fast pyrolysis products.
Linear regression methods were used to develop statistically significant quadratic models to estimate
and investigate the bio-oil and biochar yield. The optimal conditions that were found to maximize
bio-oil yield and minimize biochar yield are high nitrogen flow rates (3.5 sL/min), high heat carrier
temperatures (625°C), high auger speeds (63 RPM) and high heat carrier feed rates (18 kg/hr).
The produced bio-oil, biochar and gas samples were subjected to multiple analytical tests to
characterize the physical properties and chemical composition. These included determination of bio-
oil moisture content, solid particulate matter, water insoluble content, higher heating value, viscosity,
total acid number, proximate and ultimate analyses and GC/MS characterization. Statistically
significant linear regression models were developed to predict the yield of gaseous carbon monoxide,
the hydrogen content, moisture content and water-insoluble content of the bio-oil, and the vapor
reaction temperature at the reactor outlet. A significant result is that with increasing bio-oil yield, the
oxygen to carbon ratio and the hydrogen to carbon ratio of the wet bio-oil both decrease, largely due
to a reduction in water content.
The auger type reactor is currently less researched than other systems, and the results from
this study suggest the design is well suited for fast pyrolysis processing. The reactor as designed and
operated is able to achieve high liquid yields (greater than 70%-wt.), and produces bio-oil and biochar
products that are physically and chemically similar to products from other fast pyrolysis reactors.

1
CHAPTER 1. INTRODUCTION
A fundamental branch of the mechanical engineering discipline is energy conversion,
transforming naturally occurring resources into forms that are more usable by society. Energy
conversion is an application of engineering principles from thermodynamics, fluid mechanics, and
heat transfer, as well as machine design and mechanics of materials. A classic example of an energy
conversion process is coal combustion to provide heat for raising steam that runs turbines and
generators for producing electricity. More recently, however, biomass has been recognized as a viable
and abundant resource that can be used for the production of renewable fuels, energy, chemicals and
other bioproducts [1].
According to The Global Summit on the Future of Mechanical Engineering 2028, “One of
the most critical challenges facing mechanical engineers…is to develop solutions that foster a cleaner,
healthier, safer and sustainable world [2].” Biomass fast pyrolysis is an energy conversion process
that can be considered one such solution to these challenges.
The objective of this research study is to design and develop a novel lab-scale auger reactor
for biomass fast pyrolysis processing, determine its optimal operating conditions, and relate the
product yields and composition to these conditions. This will allow for the reactor design to be
evaluated and compared to existing, published data.
This reactor type is relatively new in the field of biomass fast pyrolysis, and can be currently
considered as an “alternative reactor.” Though there are potential economic and processing
advantages of utilizing this reactor technology for bio-oil production, there is little published data
relating the pyrolysis product yields and composition to the operating conditions of the reactor.
1.1 Biomass fast pyrolysis
Fast pyrolysis is a thermochemical process used to produce primarily a liquid product known
as pyrolysis oil or bio-oil [3], and is considered a promising route for biomass conversion. When
biomass is rapidly heated in a controlled, oxygen depleted environment at atmospheric pressure to a
final temperature of approximately 500°C, it is decomposed and converted within seconds to liquid
bio-oil, solid biochar, and non-condensable gases [4]. Fast pyrolysis can collect over 70% of the
starting material mass as liquid bio-oil, with the balance formed by approximately equal portions of
biochar and gases.

2
Bio-oil can be used as a renewable industrial fuel to generate heat and electrical power, or can
be upgraded into transportation fuels and specialty chemicals. Biochar can be used as a solid fuel
source, and has more recently found applications as an agricultural soil amendment. The non-
condensable gases are typically recycled into the process to provide process heat.
A general schematic of this thermal process is shown in Figure 1, noting the relationship
between the fast pyrolysis reactor and the system that separates and collects the reaction products.
Also note the energy input to the reactor in the form of heat, which is required to carry out the
endothermic fast pyrolysis reactions.
Figure 1. Biomass fast pyrolysis schematic
1.2 Thesis overview
This thesis consists of five remaining sections to systematically explain and support the
research effort. The next section, Chapter 2, will summarize the literature review performed to
determine the general state-of-the-art of the science and technology of biomass fast pyrolysis and
review previous research efforts related to auger reactors. Chapter 3 will review the R&D efforts
required to construct the laboratory reactor system, including a detailed description of the apparatus.
Chapter 4 will detail the methodology and materials used for the experimental phase of the research.
The results of the experiments will be presented in Chapter 5 along with a discussion, and Chapter 6
includes the conclusions of the research and recommendations for future work. Supplemental
information is located in Appendices and will be referred to as necessary.

3
CHAPTER 2. TECHNICAL LITERATURE REVIEW
2.1 Introduction
Lignocellulosic biomass is an abundant and geographically diverse natural resource. The
USDA estimates that over one billion tons of dry matter may be available annually in the United
States [5]. Examples of biomass resources include: agricultural crop residues such as corn stover,
wood residues from the forest and milling industries, municipal solid waste (MSW) from urban areas,
herbaceous energy crops such as switchgrass, and short-rotation woody crops [6].
Through photosynthesis, plants convert sunlight and CO2 into stored chemical energy,
therefore biomass can be considered an indirect form of solar energy and a renewable source of
carbon [7]. The stored chemical energy in biomass can be converted into bioenergy (heat and
electricity), liquid biofuels for transportation, chemicals, and other biobased products. This utilization
of biomass can contribute to a net reduction in greenhouse gas emissions which may impact global
climate change, and provide other benefits such as reducing foreign energy imports [8].
There are many biomass conversion pathways in various stages of development, and these
pathways are commonly grouped into two major technology platforms: biochemical and
thermochemical. These platforms are not exclusive, though, and opportunities exist to combine
technologies into so-called “hybrid processes” [9]. Biochemical technologies, such as fermentation to
produce alcohol fuels and anaerobic digestion to produce methane gas, are outside the scope of this
research and will not be discussed.
Thermochemical conversion techniques utilize heat to decompose biomass, and include four
main processes (in order of increasing temperature): direct liquefaction, pyrolysis, gasification and
combustion. Though pyrolysis and liquefaction are sometime grouped into one process, they will be
discussed here separately.
Direct liquefaction. Direct liquefaction, or often just “liquefaction”, is a mild temperature,
high pressure conversion process (around 300°C and up to 240 bar, respectively) with the primary
goal of producing a liquid product [1]. Liquefaction is often a catalytic process, and requires that the
feedstock material be slurried in an aqueous solution, usually with water as a solvent. Because of this
requirement, liquefaction may be well suited for resources that naturally have particularly high
moisture contents, such as animal manure. Huber et al. [10] note that while bio-oils from liquefaction
(often referred to as bio-crude) are of a high quality due to the low oxygen content, this comes at the

4
expense of a lower liquid yield. In general, direct liquefaction has been less investigated than other
thermochemical processes; however refer to Behrendt et al. [11] for a recent review of direct
liquefaction. Applications of bio-crude are similar to applications for bio-oil, and will be discussed
later.
Pyrolysis. Pyrolysis is the thermal decomposition of organic matter without oxygen present
[3]. The origins of pyrolysis date back as far as ancient Egypt [4]. Upon heating, moisture is first
driven off from a material, and then pyrolysis reactions occur before any remaining thermal processes
occur. Depending on the conditions, varying amounts of solid, liquid and gas will be produced [12].
Pyrolysis occurs over a range of temperatures from 400°C – 600°C, and usually at atmospheric
pressure. Fast pyrolysis is marked by high heating rates, short vapor residence times (seconds) and
rapid cooling of the reaction products, which favors maximum formation of liquids around 500°C
[13]. Slow pyrolysis, alternatively, is marked by slower heating rates, longer vapor residence times
(minutes), and high yields of solid char material [4]. Slow pyrolysis – also known as conventional
pyrolysis – has basically been applied for many years as a carbonization type process for converting
wood into charcoal [4, 14]. As slow pyrolysis yields minimal bio-oil, it will not be reviewed further.
In addition to fast and slow pyrolysis (which are not always clearly delineated), several other types of
pyrolysis are reviewed by Mohan et al. [4]. The fast pyrolysis process and technology, including
applications for the end-products, are discussed in more depth in the next section.
A benefit of direct liquefaction and pyrolysis over gasification and combustion is the ability
to produce a liquid product, which can be more readily stored and transported compared to gaseous
fuels. This implies that bio-oil can be produced in a separate location from the end-use application,
and this “distributed processing” scheme may be advantageous as biomass transportation costs can be
minimized for small scale regional facilities [15].
Gasification. Gasification is an endothermic process occurring around process temperatures
of 750°C – 1000°C to produce primarily a combustible fuel gas commonly referred to as producer gas
or syngas [1]. Depending on the process conditions and the fluidizing gas, the syngas composition
will contain varying amounts of CO, H2, CH4, CO2, N2 and other organic species. The heat required
for gasification is often provided by partially oxidizing a portion of the feedstock material. Syngas
can be combusted for heat and power applications, or upgraded into transportation fuels and
chemicals using the Fischer-Tropsch process or other techniques [16]. For more information on
gasification technology refer to Ciferno et al [17].
Combustion. Combustion is the highest temperature thermochemical conversion process (in
excess of 1500°C), and it is well understood and commonly used in many industries. With

5
stochiometric or excess air present to fully oxidize the feedstock fuel, combustion produces heat with
water and CO2 as byproducts. Heat from combustion is used for various processes, including steam
production and electricity generation. This process will not be reviewed further, however refer to
Jenkins et al. for more information on biomass combustion [18].
Refer to Demirbas for a thorough review of thermal conversion of biomass [19], and
Olofsson et al. for a review of applicable technologies and reactor configurations [20]. An overview
of these processes and their general applications is shown in Figure 2. Liquefaction is shown offset in
Figure 2 because it is not as well researched as the other thermochemical processes, and is sometimes
not even mentioned as a thermal process and lumped together with pyrolysis as a means for
producing primarily liquid fuels.
COMBUSTION
LIQUEFACTION
LIQUID
SOLID
GAS POWER
HEAT
CHEMICALS
FUELS
GASIFICATION
PYROLYSIS UPGRADING
Figure 2. Thermochemical processes
2.2 Fast pyrolysis fundamentals
Fast pyrolysis is a complex process, and though much research has been performed over the
past few decades, it is still developing at a rapid rate. This process has shown great promise for being
flexible and diverse, and is prized for the ability to produce a high yielding liquid fuel from almost
any type of biomass feedstock. Minimal biomass pretreatments are required for fast pyrolysis [3], and

6
depending on the desired outputs the process can be carried out such that no outside energy inputs are
required. Furthermore, depending on the product applications, fast pyrolysis can be a carbon neutral
or even carbon negative process.
As previously noted, fast pyrolysis is a rapid heating process in the lack of oxygen to
decompose biomass into a liquid fuel, with solid and gaseous by-products. It is generally accepted
that there are four main process characteristics for fast pyrolysis [4, 21], and will be discussed in more
depth in the next section:
• Very high heat transfer rates
• Controlled reaction temperature
• Short vapor residence times
• Rapid separation and cooling of reaction products
As the fast pyrolysis process occurs in a few seconds or less, heat transfer and mass transfer
effects as well as reaction kinetics are all important phenomenon, however these considerations will
not be reviewed here and can be found elsewhere [22-27]. Blasi [28] and Babu [29] discuss several
different kinetic models in separate reviews of biomass pyrolysis.
2.2.1 Operating conditions
There is much literature reported on the operating conditions for biomass fast pyrolysis, and
biomass properties are the first consideration in maximizing liquid yield. To ensure rapid heating and
complete devolatilization, small biomass particles are required. Though the particle size requirement
is somewhat dependent on the specific reactor technology used, the general particle size requirement
is agreed to be around 2.0 mm [10, 21, 30]. The only other pretreatment required prior to fast
pyrolysis is reduction of the biomass moisture content. Typical requirements are around 10%-wt. or
less, which minimizes the amount of water that is collected in the final bio-oil [21] and decreases the
overall reaction heat energy requirements.
Overall fast pyrolysis is an endothermic process, with sensible heat required to bring the
biomass from ambient conditions to the reaction temperature regime. Above the sensible heat
requirements, though, the fast pyrolysis reactions require a minimal heat addition. Daugaard et al.
estimate a total heat for pyrolysis ranging from 1.0 – 1.8 MJ/kg depending on the feedstock [31]. The
condition that this heat be rapidly transferred to the biomass is critical for fast pyrolysis, and many

7
mechanisms and reactor configurations have been researched and developed to accomplish this.
Heating rates on the order of 103
°C/sec have been claimed [4]. If biomass is slowly heated, secondary
reactions occur and more solid products are formed as liquid yields are decreased [32]. Rapid heating
in a fast pyrolysis reactor typically occurs by means of a hot carrier gas or solid heat carrier material,
or a heated reactor wall, or a combination of these [21, 33]. Though the addition of air or oxygen can
provide heat by oxidizing a portion of the feedstock (as is performed for biomass gasification), this
approach decreases the yield of bio-oil. Therefore, many reactors (especially lab-scale systems for
research purposes) utilize a flow of nitrogen gas to provide an inert reaction environment. Depending
on the reactor configuration, the heat transfer mode of conduction, convection or radiation may
dominate, however they will each contribute to some degree.
The reaction temperature is also critical for fast pyrolysis and has effects on the product
yields and qualities. Higher char formations occur at temperatures less than approximately 425°C,
and non-condensable gas production increases for temperatures above 600°C. Several sources report
bio-oil yields are maximized around temperatures of 500°C ± 25°C [4, 13, 21, 30]. The heat transfer
rate and the reaction temperature are both important: rapid heating to a reaction temperature that is
too low or too high will adversely affect the products, as will a slow heating rate to the optimal
reaction temperature. The reaction pressure for fast pyrolysis is typically near atmospheric, as higher
pressures favor the formation of biochar [34].
In addition to rapid heating and controlled temperatures, a short residence time for the
pyrolysis products is important to maximize liquid yields. As biomass is pyrolyzed, the reaction
products evolve in the form of condensable vapors, tiny aerosol droplets, non-condensable gases and
biochar. From the time biomass enters the reactor, the vapor residence time (where “vapors” here are
considered to be all the reaction products other than solid biochar particles) is traditionally less than 2
seconds for fast pyrolysis. This consideration is extended to include the cooling of the vapor and
aerosol products to collect them as bio-oil. For instance if the reaction products are formed in the
reactor within the first second, within the next second they should be rapidly cooled to condense and
recover as much vapors and aerosols as possible. The cooling process during the collection of bio-oil
effectively minimizes further reactions that occur at high temperatures, so fast pyrolysis can not be
considered an equilibrium process [4]. As with the “optimal” reaction temperature for fast pyrolysis,
the 2 second vapor residence time is currently a well accepted and documented operating condition
[4, 13, 30]. Longer residence times allow for secondary reactions to occur which form either
additional gases or char, both of which are undesired and reduce the liquid yield.

8
2.2.2 End products description
The chemical and physical characteristics of the fast pyrolysis products are dependent on
many factors, including: the biomass composition and the operating conditions (as discussed
previously), as well as the reactor and product recovery technology used for the processing (as
discussed in Section 2.2.4). A brief introduction to the products of fast pyrolysis is presented next.
Bio-oil. The primary product from fast pyrolysis is a dark brown liquid known as pyrolysis
oil, bio-oil, liquid smoke, wood distillate, or a number of many other terms. Bio-oil has a distinct odor
similar to smoke from a wood fire, and is often quite pungent. As discussed, the yield of bio-oil will
vary depending on the operating conditions and feedstock properties, but yields in excess of 70%-wt.
are common for wood biomass and well documented in the literature. In general, bio-oil yields for
biomass fast pyrolysis range from 65%-wt. – 75%-wt.
Bio-oil is a complex mixture of more than 300 organic compounds formed during pyrolysis
reactions that are essentially “trapped” in a liquid form [35]. Bio-oil is very different from traditional
fossil-fuel based liquids, and indeed some researchers prefer not to refer to it as oil at all (“pyrolysis
liquid”, for example). Many of these differences and the unique properties of bio-oil are attributed to
its high oxygen content (over 40%-wt.), which originates from the oxygen contained in the biomass
feedstock. It is often noted that bio-oil elemental composition is very similar to that of the original
feedstock, just in a more convenient liquid form [4, 36]. Bio-oil also contains significant amounts of
water, which results from condensing any moisture contained in the feedstock as well as significant
“reaction water” formed during the process. A common value for bio-oil moisture content is 25%-wt.
An important aspect of bio-oil is that it can not be directly mixed with hydrocarbon fuels because
phase separation occurs due to the high moisture content. Also due to the high oxygen and water
contents, bio-oil has a lower heating value than petroleum based fuel-oils, often reported around 40%
– 50% less [4, 36]. In addition to higher oxygen and water contents, bio-oil is more acidic than
petroleum based fuel-oils, with a common pH value of 2.5. Common physical properties of bio-oil are
shown in Table 1, as reported by the reviewed literature.
Note that bio-oil cannot be readily heated for distillation purposes. Due to its unique nature, a
residue of up to 50%-wt. may remain. This has implications for bio-oil upgrading operations, which
are discussed in the next section.
The chemical composition of bio-oil is dependent on many factors, and includes many classes
of oxygenated species. In addition to water, Bridgwater et al. describe the major chemical constituents
of bio-oil as: aldehydes (15%-wt.), carboxylic acids (12%-wt.), carbohydrates (8%-wt.), phenols (3%-
wt.), furfurals (2%-wt.), alcohols (3%-wt.) and ketones (3%-wt.) [13]. Alternatively, Mohan et al. list

9
five more general categories of chemical compounds: hydroxyaldehydes, hydroxyketones, sugars and
dehydrogugars, and phenolic compounds [4]. Another major constituent of bio-oil (15%-wt. – 30%-
wt.) is a water-insoluble fraction thought to originate from the lignin portion of the biomass, and is
therefore often referred to as “pyrolytic lignin” [4, 13]. Some of the interesting properties of bio-oil
are based on the pyrolytic lignin fraction, as are the processing challenges and opportunities
associated with bio-oil.
Table 1. Typical physical properties for bio-oil
Unit Value Notes
%-wt. 15 - 35 Wet basis
- 1.1 - 1.3
cP 40 - 100 @ 40°C
- 2 - 3.7
MJ/kg 16 - 19 HHV
%-wt. 0.1 - 1.0 Wet basis
%-wt. Wet basis
Carbon 32 - 58
Hydrogen 5 - 8.6
Oxygen 35 - 60
Nitrogen 0 - 0.3
Ash 0 - 0.2
%-wt. ~50 Wet basis
Property
Water content
Specific gravity
Viscosity
Distillation residue
pH
Heating value
Solids content
Elemental analysis
Note: Adapted from [10, 21, 36-38]. Refer to these sources
for more in-depth reviews of bio-oil physical properties.
Though the chemical and physical characterization of bio-oil has been researched for
decades, methodologies and standards are still being developed. Specific methodologies and practices
important to this research will be discussed as necessary in later sections. Refer to Oasmaa et al. for
several studies of commonly used procedures and recommendations for bio-oil testing [38-40].
Bio-oil has unique aging and stability issues, and as it is not an equilibrium reaction product it
is known to change over time. Low temperature storage is a commonly used practice to minimize
these changes. Bio-oil stability will not be discussed here, but is documented in the literature and is
currently a research topic of great interest.
Non-condensable gas. The gaseous products from fast pyrolysis will be referred to as non-
condensable gas (NCG), rather than syngas or producer gas which is reserved for the reaction
products of gasification. The NCG fraction from fast pyrolysis is a combustible mixture, and contains
many species including: large amounts of carbon monoxide (CO) and carbon dioxide (CO2), with

10
lesser amounts of hydrogen (H2), methane (CH4), ethylene (C2H4), ethane (C2H6), propane (C3H8),
and other light hydrocarbons. The NCG stream will also contain any un-reactive gases that were used
in the process for fluidization, such as nitrogen. As with bio-oil and biochar, the NCG yield and
composition will be dependent on many factors including the process conditions and feedstock. NCG
yield is in the range of 10%-wt. to 20%-wt, and commonly has a yield similar to that of biochar.
Biochar. The solid product from fast pyrolysis is a black, powdery substance known as
biochar, char, agri-char or just charcoal [41]. Biochar yields from fast pyrolysis range from
approximately 11%-wt. to around 25%-wt., with 13%-wt. to 15%-wt. being common values for fast
pyrolysis of wood biomass. Elementally, biochar is composed mostly of carbon (> 60%-wt.), with
smaller amounts of hydrogen, oxygen, nitrogen and sulfur depending on the biomass composition. In
2007, Mohan et al. reported biochar with fixed carbon values up to 78%-wt. and higher heating values
up to 31.7 MJ/kg [42]. Typically the majority of the ash component in the biomass feedstock ends up
concentrated in the biochar. The physical, chemical and biological properties of biochar vary widely,
and are reviewed in a recent and comprehensive book by Lehmann & Joseph [41].
2.2.3 End product utilization
Bio-oil. There are many applications for bio-oil, in varying stages of research and commercial
implementation. As produced, bio-oil can be considered a fuel source for standard industrial
equipment such as boilers, furnaces, burners, stationary diesel engines, gas turbines and stirling
engines [4, 36, 43]. Bio-oil used for generating heat or electricity in these applications displaces the
use of light fuel oil, heavy fuel-oil or even diesel fuel; however modifications are often required to
accommodate the unique properties of bio-oil as discussed. In these applications, options exist to
potentially emulsify or co-fire traditional fuels with bio-oil. Bio-oil used for heat and power
applications have been demonstrated with documented decreases in certain emissions. Refer to
Bridgwater et al. [13], Czernik et al. [36], and Oasmaa et al. [38] for more information. Gust et al.
also review potential standards for bio-oil properties used in heat and power applications [44].
In addition to utilizing bio-oil directly, there are various options for bio-oil utilization that
require an intermediate upgrading step. Recently there has been considerable research and
commercial interest in upgrading bio-oil into synthetic hydrocarbon fuels for transportation
applications. For this type of application, the high oxygen content of bio-oil is reduced through
“deoxygenation” processes commonly used in the petrochemical industry: hydrotreating and catalytic
cracking [10, 13, 21, 36]. These processes upgrade bio-oil at high temperatures and pressures with

11
hydrogen present. Refer to Jones et al. for a design study considering bio-oil as a feedstock for
upgrading to diesel and gasoline [45], Huber et al. for a study of integrating bio-oil into petroleum
refineries [46], and Elliot for a review of bio-oil upgrading [47].
A different approach to synthesizing transportation fuels from bio-oil is using the pyrolysis
liquid as a feedstock for gasification, rather than raw biomass. By gasifying a slurry of bio-oil and
biochar, it is possible to produce a clean syngas which is then upgraded to transportation fuels using
Fischer-Tropsch processing [48].
A final upgrading consideration for bio-oil is using steam reforming techniques for the
production of hydrogen [10, 35, 49]. Hydrogen is required for many industrial processes, is frequently
used in the petrochemical industry, and can be used in fuel cells to generate electricity.
Even though bio-oil is a complex liquid, is contains specific compounds such as acetic acid,
levoglucosan, and hydroxyacetaldehyde that have been researched for potential extraction [36, 43].
There are many other “specialty products” originating from bio-oil that are already in commercial use
or have been identified, including: wood preservatives, insecticides and fungicides, fertilizers, resins,
adhesives, road de-icers and numerous food flavorings and additives [21, 36].
Non-condensable gas. The gaseous by-products of fast pyrolysis are of relatively low value;
hence their main application is direct combustion to provide heat as part of the fast pyrolysis process.
This use of the gas, rather than flaring or reserving for a different application, makes the process more
thermally efficient and more greenhouse gas neutral because auxiliary fossil-fuel based sources are
minimized. In lab-scale applications, the non-condensable gas is typically vented.
Biochar. Until recently, biochar was often considered a fast pyrolysis by-product similar to
the non-condensable gas in that its best use is as a fuel source to provide energy for the process. As
biochar has a high carbon content, it is a relatively energetic material that can contribute heat energy
for the reactions or biomass drying, thus minimizing external fuel inputs. More recently, however,
there has been research interest in utilizing biochar as a soil amendment [41, 50]. In this approach,
biochar is incorporated into the soil where the biomass was harvested from, which provides benefits
to the soil, the crop and the environment – including a net reduction in atmospheric carbon [51]. In
this sense the biochar is referred to as a “carbon sequestration agent”, as shown in the schematic of
fast pyrolysis (FP) applications in Figure 3.

12
Figure 3. Fast pyrolysis product applications
Image adapted from Bridgwater et al. [21]
2.2.4 Systems technology
As with any advanced process, fast pyrolysis systems are composed of multiple subsystems.
The generalized subsystems that will be discussed are: pretreatment, reactor, and product recovery
with a relationship as shown in Figure 4.
Pretreatment. Compared to other biomass conversion technologies, the so called
“pretreatment” required for fast pyrolysis processing is minimal [30]. Typically no catalysts are used,
and no chemical treatments are required. Typically only size reduction and drying are required. As
shown in Figure 5, raw biomass from a storage and handling facility is passed through a chopping
device to reduce the particle size of the biomass and homogenize it such that it can easily be
transported through a dryer. Heat for drying purposes can be provided with either flue gas from a
direct use combustor as shown, or with process heat originating elsewhere. Finally, a grinder or
milling device is used to reduce the biomass particles to the desired size range, typically around 2 mm
as discussed previously. This is a general pretreatment subsystem, and specific technologies related to
drying and size reduction will not be discussed and can be found elsewhere.

13
Figure 4. Fast pyrolysis subsystem schematic
Figure 5. Biomass pretreatment schematic
Before specific reactor technologies are discussed, common biochar and bio-oil recovery
technologies will be discussed. The biochar recovery and bio-oil recovery technologies combined
represent the product recovery subsystem shown in Figure 4. These technologies are discussed first
because very similar product recovery technologies are utilized on fast pyrolysis systems, largely
independent of the reactor type.
Biochar recovery. It is important that biochar is separated from the remaining pyrolysis
products quickly because interactions with char may cause unwanted secondary reactions. For biochar
collection and separation, gas cyclones are common and used frequently because of their simple
design and operation [21]. Gas cyclones have no moving parts, are well understood and used
successfully in many industrial applications. Though well researched and able to provide high
collection efficiencies (above 99%), cyclone separators are not able to collect very fine particulate
matter. This is true even when multiple cyclones are used in series [13]. Other biochar collection
equipment such as hot vapor filtration and moving bed filters [52] have been researched and are still
under development. As such, some biochar particles inevitably bypass biochar collection equipment
and end up as fine particulate matter suspended in the collected bio-oil. This can be a problematic for

14
certain bio-oil utilization equipment or processes, whereas in other applications this may be
advantageous because of the increased energy value associated with the biochar. The amount of
biochar in bio-oil (reported on %-wt. basis) can be determined by addition of a solvent such as
methanol because the biochar particles will not dissolve with the liquids and can be filtered out.
Bio-oil recovery. As discussed, bio-oil recovery technology is crucial to quickly cool the
reaction products so as high yields can be realized. The associated technology can be complex and
varies greatly between systems, with major differences between lab-scale and commercial reactors.
Many research sized fast pyrolysis systems use a staged approach to cool and collect the reaction
products sequentially. Small systems typically use water and ice cooled condensers or impingers, and
a 2002 study by Gerdes et al. reviews the design and construction of a common lab-scale setup [53].
In contrast to simple heat exchangers, a more traditional technology for larger scale systems is a
“quenching” type device in which collected bio-oil is re-circulated and cooled before being sprayed
onto a stream of hot vapors and aerosols exiting the reactor. This type of quenching process
minimizes potential blockages in heat exchanger type condensers [13]. Bio-oil aerosols are
particularly difficult to collect, and a secondary device in addition to the quench system is often
required. Denoted as a “filter” in Figure 6, an electrostatic precipitator (ESP) is the preferred
secondary collection technology [13, 21]. Non-condensable gas is effectively separated from the
collected bio-oil in the quench system as shown.
Figure 6. Bio-oil recovery schematic

15
Bubbling fluidized bed. Bubbling fluidized bed (BFB) reactors, or more simply fluidized
beds or even just fluid beds, are commonly used for bio-oil production and data from these systems is
widely published and available. Refer to Boateng et al. [54] for a recent representative study, and
Bridgwater et al. [30] for a detailed review. These systems used for fast pyrolysis have been
developed over decades, based on similar technology used for combustors in industries including
petrochemical and manufacturing. A well recognized company in fast pyrolysis processing is
Dynamotive Energy Systems (Canada), and several bubbling fluidized bed reactors operate
commercially using their patented BioTherm process [55]. Refer to a review of “short residence time
cracking processes” by Hulet et al. for details on the BioTherm process [56].
Referring to Figure 7, a feeding system is used to mechanically (or pneumatically) convey
biomass into a vertical reactor vessel featuring a bed of hot sand. A large flow of inert gas is used to
fluidize the sand, providing a well-mixed volume with excellent heat transfer characteristics in which
the reactions occur. The reactor, in this example, is heated indirectly by combustion flue gas in an
annulus around the reactor, where other heating provisions such as tubes through the reactor are
possible [21]. Pyrolysis products, including condensable bio-oil vapors and aerosols, biochar and
non-condensable gases exit the top of the reactor with the fluidizing gas. Biochar and bio-oil are then
collected as discussed. Resulting non-condensable gases are recirculated as a fluidizing gas or can be
combusted for process heat.
Figure 7. Bubbling fluidized bed reactor schematic

16
Note that Figure 7 is only one representation of this reactor type, where modifications to the
gas handling and reactor heating configurations are common. For instance, biochar is shown here to
be a by-product, where instead it could be used as a fuel source for the combustor to limit the
auxiliary fuel requirement. Details of the biomass pretreatment and bio-oil recovery operations can be
found in Figure 5 and Figure 6, respectively, as discussed previously.
Though fluidized beds have been demonstrated commercially and provide high liquid yields,
heat transfer problems can be significant and significant energy can be required to handle the
fluidizing gas.
Circulating fluidized beds. Circulating fluidized bed (CFB) reactors, sometimes referred to
as transport beds, are similar to bubbling fluidized beds. However rather than having bed material
remain suspended in one reactor, CFBs have a separate combustion reactor used to re-heat the sand
which is continuously recirculated. As with fluidized beds, the CFB reactor is well understood and is
currently used in several industries on commercial scales. One configuration of a CFB reactor for fast
pyrolysis is shown in Figure 8, noting that biochar entrained with the bed material is combusted in the
presence of air to provide heat for the re-circulated sand.
Another Canadian company, Ensyn, utilizes circulating fluidized bed reactors as part of their
proprietary Rapid Thermal Processing (RTP) technique used at several commercial fast pyrolysis
plants [57]. Through a partnership with Red Arrow, the RTP technique is used to produce a consumer
grade food flavoring, which is often referred to as “liquid smoke” [58].
Figure 8. Circulating fluidized bed reactor schematic

17
The CFB design has similar advantages as the BFB and may have fewer problems scaling up,
but the sand recirculation loop requires significant complexity. As such, the CFB is not a common
reactor design used for lab-scale fast pyrolysis studies. Refer to Hulet et al. for a review of several
configurations of the Ensyn RTP design [56].
Rotating Cone. The rotating cone reactor is quite different than the bubbling fluidized or
circulating fluidized bed reactors [3]. Rather than a vertical reaction vessel with bed material that
remains well-mixed due to flowing fluidization gas, biomass is mechanically mixed in a rotating cone
with a bulk solid heat transfer medium. Sand is used as the heat transfer medium, and is referred to as
a “heat carrier”. Though sand is used as a heat carrier material in the fluidized bed reactors, hot
fluidizing gas is also used to promote heat transfer and mixing effects. Therefore, one benefit of the
rotating cone reactor is minimizing the amount of gas required for the process. However, as shown in
Figure 9, one configuration of the rotating cone reactor includes a separate fluidized bed reactor to
combust the biochar to provide heat for the recirculated sand. This aspect of the rotating cone design
is very similar to the operation of the CFB reactor.
The rotating cone reactor concept has been commercialized through work by Biomass
Technology Group (BTG) in the Netherlands, which has developed a 50 ton per day facility in
Malaysia [59]. In this design, sand and biomass are driven up the wall of the cone due to fast rotation
speeds from 300 – 600 RPM, and pyrolysis products exit from the top of the cone [13, 21, 56].
HOT
SAND
BIOCHAR
CYCLONES
PRODUCTS
COMBUSTOR
&
HX
R
E
A
C
T
O
R
Figure 9. Rotating cone reactor schematic

18
Though the BTG rotating cone concept has claimed high liquid yields from a physically
compact system [59], it has not been proven at large scales or operated for significant time periods.
Auger reactor. The auger reactor concept also features mechanical mixing of biomass and a
bulk solid heat transfer medium. However instead of the reactor vessel itself rotating, there are mixing
devices that rotate inside a stationary horizontal reaction vessel. Typically the biomass and heat
carrier are independently metered into the reactor, and the heat carrier is heated prior before entering
the reactor. Figure 10 shows a reactor with two augers (or screws); however a single auger or similar
mechanical mixing implement may also be used. As vapor products evolve they exit the reactor due
to pressure differences, and the solid materials including biochar and the heat carrier exit at the end of
the reactor. Similar to the previous designs, some biochar does leave the auger reactor with the vapor
products and is removed with cyclones as discussed previously. A solid separator device can be used
to remove biochar from the heat carrier material based on differences in particle size or density.
Similar to CFB and rotating cone reactors, a combined heat exchanger and combustion reactor then
reheats the heat carrier before it is recirculated into the auger reactor.
Figure 10. Auger reactor schematic, configuration 1
Alternatively, Figure 11 shows an auger reactor that does not separate the biochar from the
heat carrier. Similar to the CFB reactor, biochar is combusted to reheat the recirculated heat carrier.

19
BIO-OIL
RECOVERY
PRE
TREATMENT
BIOMASS
AUGER
COMBUSTOR
&
HX
AIR FUEL
ASH
FLUE
GAS
REACTOR
PRODUCTS
BIO-OIL
NON-
CONDENSABLE
GAS
HEAT
CARRIER
B
Figure 11. Auger reactor schematic, configuration 2
The auger reactor has similar advantages and disadvantages to the rotating cone reactor. As
no fluidization gas is necessary, a smaller reactor volume can be realized which has the potential to
decrease capital costs. Mechanical wear is a potential problem with this reactor. This concept has not
been demonstrated on large scales, and there is no known commercial system in operation. This
technology is still in the research phase and will be reviewed in depth in the next section.
Ablative reactors. Rather than heat transfer to biomass through contact with hot solid
material or hot gas, ablative pyrolysis is a completely different approach that has been researched.
Biomass is pyrolyzed be being brought into contact with a hot surface, either under the influence of
mechanical pressure or high gas flow rates. One version of an ablative reactor as shown in Figure 12
is a spinning disk or plate, and biomass is pressed against the hot surface to produce biochar and
vapors. The influence of pressure for this reaction mechanism is often likened to melting butter on a
hot frying pan by pressing down on it [4, 13].
The major benefit of this design is that much larger biomass particles can be used, and no
carrier gas is required. However it is clearly a complex mechanical design which complicates the
scale up.
An alternative to the spinning disk is a vortex type reactor that uses high gas velocities rather
than pressure to force biomass against a hot cylindrical surface. The National Renewable Energy
Laboratory (NREL) operated a vortex reactor for some time with high bio-oil yields, but it required a
“gas ejector” to provide extremely high gas velocities [13, 56]. There have been some

20
commercialization efforts for these types of fast pyrolysis reactors, but there has been much less
research performed compared to BFBs and CFBs.
BIOCHAR
PRODUCTS
BIOMASS
HOT
ROTATING
DISC
PRESSURE
Figure 12. Ablative reactor concept
Other types. There are other reactor concepts that have been researched; however they will
not be reviewed here. These include using vacuum pressure to quickly remove pyrolysis vapors,
entraining biomass in a flow of hot gas, and cyclonic type reactors similar to the vortex reactor
previously mentioned. These reactors typically either have low liquid yields or are complicated, but
they have had some commercialization efforts and are reviewed by Bridgwater [13, 21], Mohan et al.
[4] and Hulet et al. [56], among others.
Refer to recent pyrolysis reviews by Bridgwater [21] and Mohan et al. [4] for comparisons of
reactor technologies, and Bridgwater & Peacocke for a particularly in-depth review of many fast
pyrolysis reactor technologies and configurations [30].
2.3 State of the art for auger type reactors
Reported literature was reviewed to determine past and present research efforts related to
auger reactors for processing biomass. It was quickly determined that there is a long history of augers
being used to mechanically convey and mix materials in a reaction vessel, beginning as far back as
the 1920s with coal as a feedstock. Therefore, auger type reactors for fossil fuel processing will be
reviewed first, followed by research on biomass processing.

21
2.3.1 Fossil fuel processing
In 1927, Laucks investigated a simple device used to process coal for “smokeless fuel
production [60].” Though not described as such, this system was essentially a slow pyrolysis auger
reactor used to produce a coke-like product from coal. The reactor was a heated tube with a screw
installed, where coal was introduced at one end, and the carbonized product exited the other. A 6 in
(15.2 cm) diameter tube with a length of 12 ft (3.7 m) was situated vertically, and eventually scaled
up to 12 in (30.5 cm) diameter and a length of 18 ft (5.5 m). While theoretically simple, many
problems were noted during operation of the system, and were attributed to the difficulties in
handling coal and conveying bulk solid type materials with a screw. The screw would often bind up
upon coal decomposition, and residues would adhere to the screw. Modifications to the geometry of
the screw, as well the feed direction did not remedy the clogging problems. Eventually it was
determined that the reactor wall was at a much higher temperature than the screw surface, so that the
coal adhered to the screw during the reaction. Design modifications included heating the hollow shaft
of the screw, which allowed scaling up to a 36 inch (91.4 cm) diameter. The paper presents an
interesting discussion on coal decomposition and the effect of temperature and pressure. It was
concluded that the reactor system is favorable based on low power requirements and simple
operation, the ability for continuous processing, high heat transfer, and the ability to heat different
zones independently. It can be said that these types of considerations are all still important.
Later, in 1941, Woody investigated the commercial viability of the Hayes Process for
producing a residential fuel from petroleum coke or coal [61]. A 40 ton per day plant was operated in
West Virginia, based on a 17 in (43.2 cm) ID, 20 ft (6.1 m) long reactor installed in a furnace. Similar
to Laucks’ work, this system was an early auger reactor for slow pyrolysis of coal for solid fuel
production (to be used as a heating or cooking fuel source). The reactor tube itself rotated slowly at
1.5 – 4.0 RPM, and the auger inside was mated to a gear system that allowed for forward and
backward rotation resulting in an “apparent rotational speed” of 13.5 RPM. The feed had a residence
time of 20 minutes, and the product exited at the end of the reactor into another screw system where a
water quench was used for cooling. Gas and tar also exited at the end of the reactor and were passed
through a cooling and collection system. Using a coal combustion system, the reactor was operated at
593°C to 704°C. Brief analyses of the products are given, including production costs. A schematic of
the reactor used in the Hayes Process is shown in Figure 13.

22
Figure 13. Hayes Process reactor
Image source: Woody [61]
Hulet et al. review the Lurgi-Ruhrgas (LR) process, developed in the 1950s to upgrade
various carbonaceous feedstocks [56]. Developed in Germany to produce town gas from oil shale, the
LR reactor is sometime referred to as a “sand cracker” because sand was used as a heat carrier to
decompose (or crack) feedstock materials into higher value products such as fuel gases and
hydrocarbon liquids. A more common heat carrier material used in the process was coke particles.
The reactor in this system is also referred to as a “mixer-reactor”, as intermeshing screws are used to
quickly combine the feedstock and the heat carrier material. The vapor products quickly exit the
reactor (as low as 0.3 second residence times) and travel through cyclones and a product recovery
section, whereas the solids exit the reactor and can be separated and recycled. A commercial plant
utilizing the LR process was built in 1958 (Germany) to process naphtha for ethylene production on
the order of 1.5 x 107
kg/year. In this review there was no mention of the mixing characteristics inside
the reactor with regards to screw speed, ratio of heat carrier to feedstock, or other conditions. A
schematic of the LR process is shown in Figure 14.
By the 1980s, the LR process had begun limited operation in the United States. Schmalfeld
favorably reviews the LR process by highlighting its versatility in the ability to utilize various
feedstocks for generating of a wide variety of products [62]. He states that this flexible and efficient
process has responded to “changes in the energy market, as well as to environmental concerns.” In
addition to oil shale, feedstocks listed include: tar sands, asphaltic rock, heavy oil and diatomaceous

23
earth. Entrained particulate matter is removed from the vapor products in cyclones, and condensers
are used to collect products. A liftpipe section is used to reheat (via combustion of carbon residues)
and convey the heat carrier material back into the reactor. The process is noted to operate at
temperatures and pressures (704°C and 13.8 kPa, respectively) such that specialty equipment is not
required. Schmalfeld suggests sulfur dioxide emissions could be controlled with the addition of lime
or dolomite in the heat carrier. A pilot scale operation was referenced to be operating by 1981 in
McKittrick, California, near the McKittrick tar pits. One conclusion of this conference proceeding is
that the LR process is superior to other similar methods and that products are of high enough quality
for traditional refining methods.
Figure 14. Lugi-Ruhrgas process schematic
Adapted from Probstein et al. [63]
Daniels et al. review another LR pilot plant operation in California to process tar sands for the
production of 20,000 barrels per day of hydrogenated oil [64]. Similar to the LR process, TOSCO II
is a commercial process to convert shale to fuels which utilizes a rotating drum reactor with
recirculated ceramic balls as a heat transfer medium, and is reviewed by Probstein and Hicks [63].
During the 1990s, the auger type reactor was researched for pyrolysis of coal. Lin et al.
investigated a dual-auger to lower the sulfur content in coal prior to combustion [65]. Coal pyrolysis
was deemed an inexpensive alternative to post-combustion cleaning methods such as wet flue gas
desulfurization and dry injection processes. A ‘dual screw coal feeder reactor’ was employed in the
study to simultaneously carry out two steps: desulfurization of coal via mild pyrolysis, and the

24
reaction/separation of the resulting H2S gas using a sorbent. At temperatures less than 550°C, the coal
structure was maintained, while still allowing for sulfur to be removed as H2S. The unique design of
this system features concentric augers, operated by independent motors. The inner tube (2.54 cm) was
where the coal was fed and pyrolyzed, whereas the outer tube (5.08 cm) conveyed limestone pellets in
the opposite direction to react with the H2S gas produced. The two motors were used to control the
respective particle residence times. The reactor was heated for a length of 0.521 m via three electric
heaters, and featured collecting tanks on either end (one for char opposite the coal feed, and one for
the spent sorbent on the opposite side of the CaO feed). The cleaned gas exited the reactor and passed
through a volume meter before entering three condensers to collect liquid products. The tar and char
yields were determined gravimetrically and the gas was analyzed via gas chromatography. Variable
parameters included the process temperature (400°C – 475°C), coal residence time in the reactor (2
min – 6 min) and coal particle size (4 – 35 mesh). The resulting parameters of interest were the
‘extent of devolatilization’, product distribution, gas composition (especially H2S concentration), and
desulfurization yield. It was concluded that both devolatilization and desulfurization increased with
both residence time and temperature. The H2S was determined to be mostly from organic sulfur in the
coal and was found to be released more readily than organic volatiles due to lower activation energy
values. Also, CaO pellets were deemed to be an acceptable sorbent for this application.
In a descriptive and useful report, Camp discusses various aspects of the Lawrence Livermore
National Laboratory’s involvement in assisting the DOE and the Coal Technology Corporation with
several screw reactors for coal pyrolysis [66]. Here pyrolysis (termed “mild gasification” or “low
temperature carbonization”) is understood to be slow pyrolysis based on the low liquid yields and
high char yields. However, producing liquid fuels and chemicals from the coal feedstock was the
major aim of their research and development efforts. Caking and agglomerating coals were used in
the study, and are mentioned to be problematic during processing. Screw pyrolyzers heated externally
with combustion gas were deemed appropriate for this type of coal. Advantages of the externally
heated reactor include no separation of recirculated solids or carrier gas. Disadvantages include
mechanical maintenance and low liquid yields (which are likely attributed to the low gas flow rate
and the low heat transfer rates). Three types of screw configurations are listed as design candidates –
single screw and two types of twin screw configurations: weld fabricated or machined type (as used
in twin-screw extruders). The single screw design is the most simple and least inexpensive, but is
prone to deposit formation likely similar to that described by Laucks [60]. The twin-screw extruder
type system is the least prone to forming carbon build-up as the screws are fully meshing; however
this results in the highest cost.

25
Camp concluded that the welded flight screws (intermeshing, but not fully) combine the
benefit of preventing deposits from forming while remaining relatively inexpensive. A single screw
pyrolyzer was first developed to help determine design and scale up equations. The 89 cm long screw
had a diameter of 38 mm, with a pitch equal to the diameter. Various screw materials were
investigated, and the reactor was heated electrically. Various types of coal and experimental
conditions were investigated, and over 51 hours of operation were accomplished. Problems with the
single screw design included clogging of vapor ports and binding of the screw. Coal would become
packed in the reactor, and could bind the augers. To remedy these problems, the screw could be either
turned off and on, or operated in reverse. Depending on the screw construction, the feed rate ranged
from 3.7 kg/hr to 7.6 kg/hr. Rotational speeds of the auger ranged from 12 RPM to 36 RPM. It was
determined that the single screw pyrolyzer was an unattractive option based on the torque
requirements and the low feed rates that were achievable. Interestingly, Camp found that feed rates
did not appear to increase with increasing screw speed.
Therefore, Camp recommends a twin screw pyrolyzer to help free the char deposits that may
form, as well as aid in mixing and heat transfer within the system. Welded flights are an inexpensive
option compared to fully intermeshing screws. Recommendations for screw design include a hollow
shaft to introduce a heat transfer fluid, as well as modifying the profile of the screw flighting to
increase the intermeshing effect. Many design type equations and relationships were presented for
externally heated screw pyrolyzers. For instance: the feed rate as a function of screw speed, geometry
and fill conditions, as well as a heat transfer correlation also based on the same parameters. The feed
rate and the heat transfer coefficient were related by the heat transfer area and a log mean temperature
difference. The solid material residence time is shown to be a function of the feed rate, screw
geometry and the fill characteristics. Each of these equations is combined into a final design equation
to solve for the maximum feed rate of coal, which is claimed to be limited by heat transfer and not
“conveyance problems”. To increase the heat transfer coefficient, Camp notes that radial mixing must
be improved by flight design modifications, recommending a non-standard pitch of 0.25 to 0.5 times
the diameter (in standard auger construction, the pitch is equal to the flight diameter).
As a final recommendation, Camp recommends pre-heating the coal before the entering the
pyrolyzer. The benefit of this pre-heating is the ability to condense water separate from the oil
fractions. With a reported outlet temperature of 280°C, this pre-heating process is at higher
temperatures than standard drying practices (approximately 100°C), and therefore appears to be a
torrefaction chemical conversion process. Torrefaction is a mild thermal treatment process, and is

26
reviewed by Bergman et al [67]. For a follow up report on this research investigating twin screw heat
transfer and other topics, refer to a later report by Camp et al [68].
2.3.2 Biomass processing
The following literature sources detail various aspects of pyrolysis carried out in reactors that
feature one or more augers (or screws), or a similar rotating mechanical element. It is important to
note that the operating conditions for these reactors varies widely, and important conditions such as
auger speed are often not reported. A heat transfer medium is sometimes used that is mixed with the
feedstock, whereas other reactors have heated walls that induce the pyrolysis reactions. There has
been no finding of research relating to the mixing mechanisms of biomass and a heat transfer
medium. Also, the reported research lacks clear relationships between product yields and composition
with the reactor operating conditions. There is a wide range of system sizes, stages of development,
biomass feedstocks and product distributions.
The first known reference to the auger type reactor for biomass pyrolysis is from 1969, when
Lakshmanan et al. investigated pyrolysis of starch and cellulose for the production of levoglucosan
[69]. This reference includes a detailed account of the chemistry involved in the pyrolysis process to
produce levoglucosan. Among two other reactor schemes, a screw conveyor was investigated because
the researchers perceived this design would be useful for continuous handling the biomass as it
underwent chemical and physical changes throughout the length of the reactor. The screw reactor was
comprised of a feed hopper (batch feeding), a 1” ID steel tube with a similar diameter screw to allow
for scraping of the reactor wall. No heat transfer medium was used as the reactor walls were heated
via electrical means. The biomass feed rate was 200 g/hr, and the reactions were carried out at
temperatures ranging from 340°C to 500°C. A heated vessel was installed at the end of the reactor to
collect solid products, from which stemmed a tube that carried pyrolysis vapors to a product receiver
and traps. Surprisingly, the screw had to be rotated by hand via a simple handle mechanism. The
results of the experiments indicate that the screw reactor had slightly lower yields of levoglucosan
than the batch reactor investigated, possibly due to further decomposition of the vapors as they
traveled through the reactor. The study also found that char residues occasionally bound the screw
inside the reactor. This was remedied by occasionally adding oxygen into the heated reactor to “burn-
out” any deposits and free the screw. The authors state that this type of reactor would be problematic
at a large scale due to heat transfer issues. Also, the shaft seals were noted for areas to be concerned
with mechanical wear.

27
Yongrong et al. discuss an auger reactor concept for pyrolyzing tire waste in a conference
proceeding from 2000 [70]. They note that worldwide generation of used tire rubber exceeds 9
million tons annually, a sizeable amount when considered as a feedstock high in carbon. There is a
brief review on the mechanisms and kinetics of fast pyrolysis, as well as a literature review on the
reactor schemes currently being used for tire pyrolysis in China. At the Zhejiang University, two
reactors have been developed that were briefly described (but will not be discussed here): an
externally heated rotary kiln, and an internally heated cascade moving bed (CMB) reactor. There was
also a description for the design of a screw reactor that can be either externally or internally heated.
Perceived benefits include lower costs for construction and operation. Figure 15 shows a conceptual
schematic of this reactor. The end of the paper lists 21 Chinese patents related to pyrolyzing tire
rubber, including 4 patents on ‘screw reactors’ and 4 patents on ‘agitator reactors’ (including
impellers and mechanical scrapers).
Figure 15. Screw reactor concept
Image source: Yongrong et al. [70]
Around 2002, researchers at the Forschungszentrum Karlsruhe (FZK) center in Germany
began investigating a two step biomass to liquid (BTL2) processing scheme consisting of
decentralized fast pyrolysis followed by centralized gasification of bio-oil and biochar mixtures [71,
72]. The regional (also known as distributed or decentralized) processing includes: drying and
grinding of biomass, fast pyrolysis in a twin-screw reactor, and recombination of the bio-oil and
biochar into a slurry mixture. This slurry is formed to reclaim most of the energy from the biomass,
but in a form that is more easily transported to a central facility where it can be pumped into a

28
pressurized entrained flow gasification reactor. Clean syngas is produced and subsequently upgraded
into fuel using the Fischer Tropsch process.
Operational in 2003, a 10 kg/hr fast pyrolysis reactor at FZK was the first known system to
intentionally and directly utilize the mixer-reactor concept from the Lurgi-Ruhrgas process as
discussed previously [73]. A schematic of the FZK mixer-reactor is shown in Figure 16, and the
reactor system is shown in Figure 17. Note the long vertical pipe seen in the right side of Figure 17 is
the hot sand recirculation loop as shown in Figure 16.
A 2006 publication describes the reactor, the process and some preliminary results [74]. The
twin screw reactor was selected for the fast pyrolysis system because of the experience and operation
related to commercial sized Lurgi-Ruhrgas plants, and the fact that carrier gas which dilutes the
product stream is not required for this type of reactor. The 10 – 15 kg/hr reactor has a length of 1.5 m,
with intermeshing screws with inner and outer diameters of 20 mm and 40 mm, respectively. Sand is
heated indirectly to 500°C - 550°C in a vertical tube, surrounded by a shell with fluidized sand that is
heated with flue gases from combusting the non-condensable pyrolysis gas. In the axial direction,
straw biomass enters the reactor before the hot sand enters. There is no mention of mixing
characteristics or mechanisms of the heat carrier and the biomass, other than a “mechanically
fluidized” state is achieved. This is most likely based on the high rotational speed of the screws, up to
240 - 300 RPM. To provide heat for the reactions, the sand to biomass feed ratio was originally 20:1
(mass basis), but ultimately reduced to 6:1. The vapor products leave the reactor due to pressure
differences, coarse char is transported to the end of the reactor, and fine char is separated with two
cyclones. It is not clear how or if the coarse char is separated from the sand before recirculating. If it
is not separated, the coarse char will enter the reactor with the sand, as there appears to be no direct
combustion process to burn off residual char (only the NCG is combusted).
Bio-oil is collected in two condensers: one is mostly organics with low water content, and the
other is an aqueous fraction with high water content. The mass yields of rice and wheat straw were:
50 – 55%-wt. bio-oil, 20%-wt. non-condensable gas, and 25 – 30%-wt. char. The mass yields of
wood sawdust were: 70%-wt. bio-oil, 15%-wt. non-condensable gas, and 14 – 18% char. The heat
carrier is recycled into the system via a mechanical type bucket elevator. The slurry is produced using
a colloid mixer, and has 25 – 40%-wt. biochar solids with an energy density value of 17 – 33 GJ/m3
,
compared to 0.7 – 2.6 GJ/m3
for the raw biomass.

29
Figure 16. Twin screw mixer-reactor schematic
Image source: Henrich [75]
Figure 17. FZK twin screw mixer-reactor
Image source: Henrich [75]

30
Other than the current system as described in this thesis, this is the only known reactor for
fast pyrolysis of biomass utilizing two co-rotating, intermeshing screws and an independently metered
heat carrier material. However there is no published information relating the product yields to process
conditions such as sand temperature, heat carrier to biomass feed rate, screw speed, or others. As
such, it is unclear why the particular operating conditions were selected and if the system or process
is considered to be optimized. Furthermore, as the produced bio-oil has a specific intended end-use
application, chemical analysis and composition is unfortunately not provided. The only analysis
appears to be on the “gasification feedstocks”, which are understood to be the bio-oil and char
slurries.
In 2007, Plass discussed a partnership between FZK and Lurgi AG to commercialize the two
step biofuel process [76]. This process, termed Bioliq, is reviewed in detail by Henrich, et al. in a
2009 publication documenting the cost estimates and energy balance [48]. Dinjus, et al. [77] and
Leible, et al. [78] have also had opportunities to describe the process and the economics. In the Bioliq
processing scheme, biomass is transported from a 25 km radius to a decentralized 0.1 GW fast
pyrolysis plant, where approximately 90 plants across Germany supply bio-oil slurries to one of three
3.5 GW centralized gasification facilities for synthetic fuel production.
Similarly in the U.S., commercialization efforts related to the auger reactor for biomass fast
pyrolysis can be dated to the early 2000s. Renewable Oil International, ROI (Florence, AL) was
formed in 2001 by Phillip Badger who describes the concept of having a small scale bio-oil plants to
supply bio-oil to multiple end-users, or multiple plants supply bio-oil to one end-user [79]. ROI
developed a 5 ton per day auger reactor system for use on a poultry farm to convert animal wastes to
bio-oil, which is used for on-farm heating purposes. In a 2006 conference, Badger further describes
the technology as simple and inexpensive to implement, with claims of liquid yields up to 60% [80].
The ROI commercialization strategy includes scaling up to a 125 ton per day plant located at a
Massachusetts saw mill, though the construction or operation of this plant can not be confirmed as no
information is currently available. The ROI system features a reactor with a single auger and uses
steel shot as a heat carrier, but no known operational, yield or product composition data has been
published. In a 2008 article, Badger discusses plans to have auger reactor systems on portable trailers
that will be transported to various sites to process energy crops such as switchgrass [81].
The ROI technology was developed in conjunction with Peter Fransham, who in a 2006
article describes how scale-up limitations with fluidized bed reactors in the 1990s led to the auger
reactor design [82]. Fransham claims his work with a “heated auger reactor” for processing treated
wood dates back to the early 1990s, through Encon Enterprises. The concept of using a horizontal

31
reactor with a heat carrier material allowed for rapid vapor removal from the reactor and from the
char, and high liquid yields around 60% are claimed at temperatures around 400°C. At some point,
Encon Enterprises became Advanced BioRefinery, Inc., ABRI (Ontario, Canada), which is
simultaneously commercializing the same reactor technology as ROI. ABRI has developed a 1 ton per
day unit built for on-farm use, and also makes claims to the 5 ton per day unit operating on the
Alabama chicken farm as described by Badger. A 50 ton per day unit was slated for operation at a
logging site in Canada, though no information is currently available.
In 2006, Badger and Fransham published an article describing fast pyrolysis technology that
could be applied to modular, possibly transportable systems for bio-oil production [83]. As described
above, ROI is developing small scale pyrolysis plants to place them in close proximity to a given
biomass source, however there are no known commercial operating systems developed by ROI. The
article notes underbrush material cleared by the U.S. Forestry Industry to minimize fires is expensive
to transport due to its low density, and as an alternative Badger and Fransham suggest bio-oil
production to simply handling, transportation and storage issues. A comparison of the energy density
of bio-oil to various types of raw biomass and current densification techniques is presented. Based on
various types of biomass and their moisture contents, bio-oil exhibits an energy content increase from
1.5 to 15 times on a volumetric basis (MJ/m3
). Another comparison is conducted for transporting
solid biomass in a standard tractor trailer van versus transporting liquid bio-oil in a standard tanker
trailer. Hauling solid wood chips results in approximately 24.5 tons maximum per trailer load, with an
energy storage capacity of 220 GJ. However if transporting bio-oil in a tanker capable of hauling
9500 gallons of No. 2 fuel oil, the energy storage increase to 558 GJ. The authors note that gross
vehicle weight regulations limit the amount of bio-oil that can be transported in this method, not the
volume of the tanker. As a final comparison, a bulk solids handling system is compared to a liquid
handling system for a 50 MW power plant concept. The solid fuel system incorporates a complicated
array of many operations whereas the liquid system is simply composed of a few operations. Though
both systems have a comparable capital cost, no analysis was conducted for operations and
maintenance costs, and it’s likely they would be much lower for the liquid system due to the lower
number of unit operations. Another noted advantage of the bio-oil fuel system over the solid fuel
system is the area requirement on site: 4.5 versus 9.6 acres, respectively. This study does not present
any experimental data from an auger reactor system, or biomass fast pyrolysis in general.
The only known published data specifically using a system constructed by ABRI or ROI is
from a 2007 study by Schnitzer et al. that characterizes the composition of bio-oils and chars
produced by fast pyrolysis of chicken manure [84]. Animal wastes are noted as a threat to the

32
environment, as well as posing health risks to humans and animals. These wastes, however, have the
potential to be a feedstock for thermal conversion processes as opposed to alternative disposal
methods. The “reactor screw conveyor” used for this study was developed by ABRI as discussed
above. Steel shot heated to a mild temperature of 330°C was used as a heat transfer medium, where
the size of the steel shot and the mixing of the shot with the feedstock were not described. There was
also no mention of reactor design or crucial operating characteristics such as: feedstock or steel shot
feed rate, or auger rotational speed. The vapors exited the reactor and were cooled to less than 100°C
within 1 to 2 seconds. The product distribution was described as: 10% of the initial mass was
converted to gas, 63% of the mass exited as hot vapor, and 27% left as solid char. Of the 63% vapor
however, 13% was non-condensable, and no distinction was made between how the initial gas
fraction was delineated from the final “non-condensable” gas fraction and how or if they exited the
reactor separately. The bio-oil yield of 50% was split into two fractions by gravity via a separatory
funnel. Several analytical methods were employed to characterize the products, including:
combustion, NMR (both CP-MAS and C), and FTIR. Results indicated the heavier bio-oil fraction
was higher in carbon and hydrogen, and lower in nitrogen and oxygen than the light bio-oil fraction.
Utilizing a design from ROI, Mississippi State University (MSU) has been researching the
auger reactor concept for bio-oil production since at least 2004, and has published multiple studies. A
lab-scale auger reactor system has been developed at MSU as shown in Figure 18.
Figure 18. Mississippi State University lab-scale auger reactor
Image source: Steele [85]

33
In 2007, Mohan et al. published a paper documenting the biochar produced by the Mississippi
State University auger reactor as a means for adsorbing heavy metals [42]. Lead, cadmium, arsenic
and zinc can be toxic to plants and animals, and can be released into the environment by many
industries. Though several methods to adsorb these materials currently exist, using biochar may be
advantageous. Oak and pine samples (wood and bark) were pyrolyzed in a 1 kg/hr reactor at 400°C
and 450°C. The 40 in (101.6 cm) reactor is externally heated in four separate zones, and no heat
carrier is mixed with the biomass feed. The four heated zones are marked by an isothermal
temperature and length in brackets, respectively: a “pre-heat” section [130°C, 4 in (10.2 cm)], an
initial pyrolysis zone [either 400°C or 450°C, 10 in (25.4 cm)], a secondary pyrolysis zone [100°C
less than the previous section, 8 in (20.3 cm)], and a cooling zone [300°C, 8 in (20.3 cm)]. The final 3
in (7.6 cm) is left unheated, leaving 7 in (17.8 cm) unaccounted for in the description. The reactor has
a simple pipe configuration, and features a single auger with a diameter of 3 in (7.6 cm), and a pitch
equal to the diameter (standard flight construction). The rotational speed was said to be highly
changeable, but 12 RPM was used for this study. A descriptive schematic is provided that clearly
illustrates the temperature profile down the length of the reactor, which also shows the residence time
in each section. The char residence time is 30 seconds in the pyrolysis zones, and around 60 seconds
in the whole reactor (linear speed of 91.4 cm/min). A wealth of characterization studies were
performed on the char products, including proximate and ultimate analyses, as well as kinetic,
equilibrium and adsorption studies. Oak bark was found to be the best adsorbent due to the high
surface area and pore volume of the char it produced. The results indicated that the biochar has less
specific surface area than activated carbon, but the researchers concluded that biochar may still be
more valuable as an adsorbent than a source of solid fuel.
In 2008, Ingram et al. published data on bio-oil produced from the previous study [86]. Oak
and pine samples (both wood and bark) were pyrolyzed in the 1 kg/hr electrically heated reactor at
450°C, with no carrier gas or heat carrier material, and at a low auger speed of 12 RPM. The authors
note that the system operation has lower heat transfer rates and longer vapor residence times than
prescribed for traditional fast pyrolysis, but that these characteristics are not inherent to the ROI
design or the auger design in general. Though not described explicitly, the inclusion of a heat carrier
is what provides the increased heat transfer in the ROI design. As such, the authors state that this
system is a first generation design and a second generation system is under development. It is
assumed the new system will include the capability of adding heat carrier material into the reactor.
The bio-oil yields were relatively low for fast pyrolysis of wood biomass (44%-wt. - 56%-wt.), which
can be attributed to the low heat transfer rates. A number of bio-oil characterization studies are

34
performed to conclude that the lab-scale auger reactor system produces bio-oil that is very similar to
other fast pyrolysis reactors that have higher heat transfer rates. The authors note that the auger
reactor may be a suitable technology for small scale, distributed fast pyrolysis systems as described
by Badger and Fransham [83].
In 2009, Bhattacharya et al. published a study investigating fast pyrolysis of wood and plastic
mixtures using the MSU auger reactor [87]. As a means to recycle the 30 millions tons of plastic
produced in the U.S. annually, the authors consider fast “co-pyrolysis” of plastic and wood. Three
different types common plastics were mixed with yellow pine wood at 50:50 mixtures by weight:
polystyrene, polypropylene, and high density polyethylene. The feedstocks were pyrolyzed at 1 kg/hr
at 450°C (the polystyrene mixture was pyrolyzed at 525°C), and the bio-oil vapors were collected in a
series of three water cooled condensers. As the authors refer to the previous studies for operation of
the reactor, the auger speed is assumed to be the same at 12 RPM. As previously, there is no mention
of heat carrier or purge gas used in this system. After detailed chemical and physical analyses, the
authors conclude that the bio-oil from wood and plastic is upgraded relative to bio-oil from wood
alone. As the plastic materials are hydrocarbons, the bio-oil from the mixed feed has a lower oxygen
and water content, which increases the heating value. The bio-oil was also found to be less acidic and
less dense, which are important considerations for storage and handling. For the various feedstock
combinations, the bio-oil yields ranged from 38%-wt. to 64%-wt.
Around this same time, a lab-scale auger reactor for slow pyrolysis was under development at
the University of Georgia (UG). Garcia-Perez et al. published a 2007 report documenting the
properties of bio-oil produced from pine wood in an indirectly heated reactor system [88]. The reactor
is an electrically heated 100 mm diameter tube, and biomass is fed with a rotary valve at 1.5 kg/hr.
The auger speed is very low at 2.2 RPM, which correlates to a solid residence time of almost 6
minutes in the heated zone. Biochar exits at the end of the reactor into a char trap, and vapors exit into
a vertical heat exchanger and a set of five ice traps. The reactor operates at a slight negative pressure
using a vacuum pump, and is purged with 3 L/min of nitrogen. The pine wood resulted in a bio-oil
yield of almost 58%-wt., and a char yield of 30%-wt. The collected bio-oil was separated into two
fractions before analysis. Bio-oil was blended into biodiesel at various mass fractions from 10% to
50%, with additional analyses performed. The authors concluded the bio-oil addition to biodiesel is
feasible, and results in minimal changes in the fuel properties. There is minimal discussion on the
reactor design, and it is not clear why the temperature or auger speed conditions were selected. No
heat carrier is used in this system, which is shown schematically in Figure 19.

35
Table 2 summarizes the product yields and selected operating conditions from published data
on auger type reactors used for biomass fast pyrolysis.
Figure 19. University of Georgia auger reactor schematic
Image source: Garcia-Perez et al. [88]
Table 2. Comparison of auger reactor published data
Bio-oil Biochar NCG
44 - 56 17 - 28 nr Oak
43 - 52 10 - 24 nr Pine
UGb
58 30 12 Pine 2 500 None
70 14 - 18 15 Wood
50 - 55 25 - 30 20 Wheat straw
ABRId
50 27 23 Chicken manure nr 330 Steel shot
12 450 None
MSUa
FZKc
60 - 300 500 Sand
Heat
carrier
Temperature
(°C)
Reactor
Product yields (%-wt.)
Feedstock
Auger
speed
(RPM)
Notes: All product yields are on a wet biomass basis (as reported or assumed)
nr – Not reported
a – Mississippi State University reactor, data from Ingram et al. [86]
b – University of Georgia reactor, data from Garcia-Perez et al. [88]
c – Forschungszentrum Karlsruhe reactor, data from Raffelt et al. [74]
d – AdvancedBioRefinery Inc. reactor, data from Schnitzer et al. [84]
Of special interest in Table 2 is the large variation in auger speed among the different reactor
systems. Note that the MSU system [86] uses an auger speed very similar to that used in the Hayes
Process reactor [61], and within the range of auger speeds as reported by Camp for the twin-screw
coal pyrolyzer [66], and none of these three systems have a heat carrier material. These auger speeds
are much lower than reported by Raffelt et al. for the twin screw reactor using sand as a heat carrier

36
[74]. Also it is noteworthy that over a period of six decades, researchers investigating auger reactors
for coal and biomass processing repeatedly reported difficulties with mechanical binding and feed
clogging.
Hornung, et al., describe a system of two rotary type kilns used to process scrap electronic
material [89]. This material, known as Waste Electronic and Electrical Equipment (WEEE), contains
environmentally harmful components such as: dioxins, furans, lead, cadmium, and bromine.
Discarded WEEE is typically either combusted or landfilled with traditional trash; however this
allows the toxins to enter the atmosphere and ground water supply. In an effort to address this
problem, 12 European entities have developed the Haloclean pyrolysis process. The feedstock is
reduced to a size of 25 mm, where it is then mixed with steel spheres in a rotary kiln to promote both
heat transfer and grinding of the material. This first kiln (Haloclean) operates around 350°C by means
of external electric heaters, and features an axial screw to convey and mix the WEEE and metallic
spheres as a heat carrier material. Volatile products exit the reactor, and the remaining material and
heat carrier enter a second rotary kiln (PYDRA) operating at 500°C. The literature states that the
rotary kiln is able to provide good heat transfer rates and “short” residences times to prevent
secondary reactions, however these residence times are on the order of hours rather than seconds.
There is no distinction mentioned between solid and gaseous residence times. The “thermal chemical
treatment pilot plant” has a range of capabilities in regards to product utilization in terms of
combusting gas and oil for process heat, or cooling and cleaning the products for other end uses. The
oil formed was found to be composed mostly of phenols, and had a bromine content too high for
subsequent processing. The system is considered to be successful in that it is able to both recover and
separate precious metals from the electronic waste.
Kodera et al. developed a small scale reactor based on a screw conveyor to process waste
plastics for fuel gas production in Japan [90]. This gas production is envisioned as a way to recycle
plastic, and as an energy source for various industries. When considering the pyrolysis of polyolefins,
the paper mentions such difficulties as controlling the residence time and the formation of waxy
products and coke. Also, when using a fluidized bed reactor, the reaction products require separation
from the inert fluidizing gas. This led to the development of a horizontal, tubular reactor referred to as
a moving bed reactor. This reactor features a screw conveyor used to mix a feedstock and sand that
was used as a heat carrier. The reactor was heated electrically for 900 mm (500 mm at a constant
temperature), and has dimensions of 1200 mm length by a diameter of 70 mm. The processing occurs
at atmospheric pressure, and the system is purged with nitrogen. The gaseous products travel through

37
a gas meter and are analyzed via gas chromatography. The sand and any liquid products are collected
in a receiver at the end of the reactor where a screen traps the sand and allows liquids to pass through.
Two main experiments were performed: pyrolysis using polypropylene pellets (3 mm
diameter) and sand (0.3 mm diameter), and catalysis using the same products in conjunction with a
silica aluminum catalyst mixed in with the sand. The results of the experiments carried out at 700°C
were a gas yield on the order of 82%-wt. and 94%-wt., with the main constituents being methane,
ethylene, propylene and some C4 – C6 species. The sand in the 500 mm isothermal section had a
residence time of 10 minutes, which was considered to be the reaction time. Important trends were
that the mass yield of gas increased with reaction time and temperature, while the opposite was true
for the oil yield (which decreased with both reaction time and process temperature). The yields were
found to be linear with reaction time. Temperatures ranged from 500°C to 700°C, and reaction times
ranged from 5 to 25 minutes. The authors concluded that the rotation rate of the screw “effectively
controlled residence time of the polymer and liquid products”, but the gas composition was
independent of reaction time implying that the gas residence time is independent of the screw speed.
Also sand was deemed as an adequate heat transfer medium. There was no mention of screw speed,
screw geometry, or mixing of the sand and feedstock. The authors suggest that the fuel gas could have
applications for residential cooking, heating and even transportation. Based on the development of
this bench-scale reactor, a similar demonstration scale reactor was conceptually designed. The 3 m
long unit is sized for 100 kg/hr, and is heated via combustion of the oil by-product formed. The
unique design features six screws: twin screws in the main reactor, which is flanked on either side by
reactors each with two more screws rotating in opposite directions.
Oudhuis et al. discuss a unique screw reactor that is part of a two-stage gasification process as
part of the “Waste to Energy” research platform at the ECN of the Netherlands [91]. The Pyromaat
facility is a 25 kWth two-stage gasifier concept that has processed such waste feedstocks as: scrap
metal and plastics (including electronic equipment as discussed for the Haloclean reactor), tire rubber,
construction and demolition material, carpets, and biomass. The pyrolysis system is electrically
heated and features a horizontal screw reactor with a diameter and length of 10 cm and 150 cm,
respectively. The screw is said to have “open flighting” and helps to ensure the feedstock contacts the
hot reactor wall, similar to a rotary kiln. The feed rate ranges from 1 kg/hr – 10 kg/hr, and the
operating temperature and pressure is 500°C and atmospheric, respectively. Approximately 28% char
is formed by this step, and the volatile products are then gasified in a reactor with a diameter and
height of 15 cm and 150 cm, respectively, at 1200°C. A gas cooler and scrubber follows the gasifier,
which prepares the gas for sampling and various end-uses such as a Solid Oxide Fuel Cell. An

38
interesting aspect of this system is the feedstock composition and the potentially toxic chemicals
contained therein. The metal constituents in the feedstock are contained in the char by-product, along
with carbon and ash. This is a potential way to keep these environmentally harmful products out of
the environment, for this byproduct has been envisioned by the ECN to be smelted for re-utilization
rather than the alternative of landfill storage or incineration.
Similarly, Brandt et al. investigate unique gasification system aimed at minimizing the tar
yield in the producer gas stream [92]. A 100 kWth gasifer was preceded by an externally heated
pyrolysis unit featuring a screw conveyor. The pyrolysis system, operating at 400 – 600°C, was fed
with wood chips at a rate of approximately 28 kg/hr. The gaseous and char components were then
directly fed into a gasifer where steam and air were introduced to oxidize the products at 1050 –
1100°C. Before exiting the reactor and being analyzed, the gaseous products were also passed
through a bed of char at the bottom of the reactor for further reactions to occur. This unique system
results in documented decreases in tar production.

39
CHAPTER 3. EXPERIMENTAL APPARATUS
The lab-scale fast pyrolysis system designed and developed for this research is shown below
in Figure 20, and the design and description of each sub-system will be described. The system
consists of the following main components: biomass feeding sub-system, heat carrier sub-system,
auger reactor sub-system, product recovery sub-system, data acquisition and control sub-systems.
Figure 20. Lab-scale auger reactor system

40
3.1 Lab-scale system design
The engineering design procedure for the lab-scale system centered primarily on the reactor
and heat carrier sub-systems which will be described briefly. For a “lab-scale” biomass fast pyrolysis
system, biomass feed rates around 0.5 kg/hr – 2.0 kg/hr are common. Therefore, early in the design
phase a nominal biomass feed rate of 1.0 kg/hr was selected and became a fixed parameter. The initial
design calculations were based largely on thermodynamics, and were used to calculate the required
heat carrier mass feed rate. After the heat carrier feed rate was determined, then the geometry of the
reactor and other sub-systems could was considered. The Mathcad 2001i Professional and Interactive
Thermodynamics v1.5 software packages were used extensively during the design phase for
simultaneous equation solving purposes.
For the discussion of the reactor design, the generalized schematic shown in Figure 21 will be
useful. Note that the parameters and variables shown will be discussed as necessary.
b
m

HC
m

c
b
P m
m
m 

 −
=
C
HC
S m
m
m 

 +
=
Figure 21. Reactor design schematic
Regarding the system mass balance, the reactor was considered an open system with biomass
and heat carrier as entering flows, and solids and pyrolysis products as exiting flows. The mass
balance for steady state conditions is described by Equation 1.
( ) 0
m
m
m
m
m
m
m
m
m
dt
dm
HC
C
p
HC
b
S
p
HC
b =
+
−
−
+
=
−
−
+
= 







 Equation 1

41
Where , , , , and are the mass flow rates of biomass, heat carrier,
pyrolysis products, solids, and biochar respectively, in units of kg/hr. This analysis assumes all
biochar exits with the heat carrier material, and the pyrolysis products include condensable vapors,
aerosols and permanent gases. Considering the reactor system as a heat exchanger in which flowing
heat carrier material transfers heat to flowing biomass material, the heat carrier mass feed rate is
determined by Equation 2.
b
m
 HC
m
 p
m
 S
m
 C
m

( )
f
HC,
HC,i
HC
p,
b
P
HC
T
T
C
m
Q
m
−
⋅
⋅
=

 Equation 2
Where QP (J/kg) is the heat required for pyrolysis of biomass, Cp,HC (J/kg-K) is the specific
heat capacity for the heat carrier material on a mass basis, THC,i (K) and THC,f (K) are the inlet and exit
temperatures of the heat carrier material, respectively, and the feed rates are as discussed. Biomass
and heat carrier inlet properties and assumptions can be found in Appendix A.
The heat for pyrolysis, QP, includes the sensible heat energy required to bring the biomass to
the reaction temperature, plus the energy required to initiate and complete the pyrolysis reactions
[31]. A value of 1.61 MJ/kg was selected for QP, which is slightly above the value required for
pyrolysis of corn stover [31]. The system was originally designed to process corn stover biomass.
Though fast pyrolysis is an endothermic process, it is noteworthy that the majority of the heat
required is simply for the sensible heat input to raise the biomass temperature. For instance, assuming
an average specific heat of 2.27 kJ/kg-K for biomass [93], a temperature increase from 25°C to 500°C
requires 1.08 MJ/kg, or over 2/3 of the total heat required for pyrolysis. The details of the heat for
pyrolysis analysis can be found in Appendix A.
Referring again to Equation 2, the specific heat capacity of the heat carrier was selected to be
815.2 J/kg-K for sand [94, 95]. Therefore, to determine the heat carrier mass feed rate, the only
unknown variables are the inlet and outlet temperatures of the material. However to ensure sufficient
heat is available for pyrolysis, the outlet temperature should remain above a threshold near the
minimum pyrolysis reaction temperature. For suitable outlet temperatures between 400°C – 500°C,
the required heat carrier feed rates are shown in Figure 22 as a function of inlet temperatures ranging
from 475°C – 700°C. These are considered to be reasonable and achievable inlet temperatures, and
the heat carrier feed rate results are in agreement with information regarding the FZK twin screw
mixer-reactor [74, 75] and the CFB reactor as part of the RTP design [56].

42
5
7
9
11
13
15
17
19
21
23
25
475 500 525 550 575 600 625 650 675 700
Heat carrier inlet temperature, THC,i (°C)
Heat
carrier
mass
feed
rate,
(kg/hr) 400
450
500
Heat carrier exit
temperature, THC,f (°C)
400 450 500
HC
m

Figure 22. Heat carrier mass feed rates as a function of temperature change
Similar theoretical analyses were performed to determine the feed rate requirements for heat
carrier materials other than sand, including steel and aluminum shot.
After a suitable range of heat carrier feed rates was known, the design procedure was
extended to consider the flow of biochar and pyrolysis products through the reactor. The analysis and
assumptions for properties of biochar and pyrolysis products are found in Appendix A. To determine
the volumetric flow rate of gaseous pyrolysis products through the reactor, the average molecular
weight for bio-oil vapors and non-condensable gas was adapted from a 2006 design study utilizing
ASPEN Plus software to analyze large-scale bio-oil production [96].
With feed rates and flow rates of reactants and products as determined, the design procedure
continued by considering the reactor as a mechanical conveying system. This system was first
designed to convey and mix biomass and heat carrier, at room temperature conditions. The volumetric
“fill” of the solids in the reactor cross section, τfeed, was assumed to be 0.5 as is common for screw
conveyors [97]. This assumption also allows for gaseous products to occupy volume in the reactor
above the solids. The detailed analysis of the reactor fill specifications for biomass, heat carrier,
biochar and vapors can be found in Appendix A.

43
Before any reactor dimensions were considered, the auger geometry and configuration were
developed. To help ensure sufficient mixing in the reactor, a twin-auger design resembling the FZK
system [74] and the LR process [56] was favored over the single auger design used by MSU [86] and
UG [88]. The twin auger design was also chosen to limit the potential for feedstock clogging in single
auger pyrolyzers as described by Camp [66, 68] and Laucks [60]. Literature on mechanical conveying
of bulk solids and industrial mixing was reviewed to determine standard practices and design
parameters [97-102]. A 2005 study by Al-Kassier et al. investigating a screw dryer for biomass was
also referenced [103]. Many technologies for mixing of particulate solids were reviewed, and these
sources helped verify that a mixer based on co-rotating, intermeshing augers with standard flighting is
a suitable design for the system of interest. It was found that special auger flighting designs such as
those shown in Figure 23 are often preferred for mixing applications [97], however these were not
considered due to the corrosive, abrasive and high temperature environment inside the reactor.
Figure 23. Various auger flighting designs
Image adapted from Screw Conveyor Corporation [104]
To assist in selection of the auger diameter, dA, manufacturer data was referenced to relate
volumetric capacity to diameter and rotational speed [105]. For example, a 1.5 in (2.81 cm) OD auger
has a nominal capacity of 57.9 cm3
/revolution. For a requirement to convey a given volumetric flow
of material, there is a tradeoff between auger size and speed: a small auger rotating quickly can
theoretically convey the same volume of material as a larger auger rotating slower. For the range of
volumetric feed rates determined for biomass and heat carrier mixtures (175 cm3
/min – 475 cm3
/min),
a 1.0 in (2.54 cm) OD auger was found to be sufficient for reasonable rotational speeds that are within
the range of similar pyrolysis reactor systems. Figure 24 shows the resulting volumetric feed rate of
various size screws and screw speeds.

44
0
100
200
300
400
500
600
700
800
900
1000
1.25 1.50 1.75 2.00 2.25 2.50 2.75 3.00 3.25 3.50 3.75 4.00 4.25 4.50 4.75 5.00
Screw diameter (cm)
Volumetric
feed
rate
(cm
3
/min) 10 RPM
30 RPM
60 RPM
180 RPM
DESIGN MAX
DESIGN MIN
Figure 24. Volumetric feed rate as a function of screw size and speed
Note that the volumetric feed rates shown in Figure 24, however, are in reference to a single
auger. As the design calls for two augers, the volumetric carrying capacity is greater than for a single
auger. The fact that the augers are intermeshing, though, implies the capacity will be inherently
greater than for a single auger, but will not be doubled. It is therefore assumed that the capacity of
twin intermeshing augers is 1.5 times that of a single auger with the same outer diameter.
For a single #16 auger [1 in (2.54 cm) OD, 1.25 in (3.175 cm) pitch], the volumetric capacity
is 16 cm3
/revolution [105], so twin #16 intermeshing augers are assumed to have a capacity of 24
cm3
/revolution. These two cases represent the first two lines in Figure 25, respectively, as a function
of screw speed. However as the system is designed not to operate completely full to allow for
thorough mixing and efficient vapor removal, an additional case is shown for 50% volumetric fill.
Recall the previous design assumption for the level of fill, τfeed, is 50%, implying faster auger speeds
are required for a given volumetric feed rate.

45
0
100
200
300
400
500
600
700
800
900
1000
0 10 20 30 40 50 60 70 8
Screw speed (RPM)
Volumetric
feed
rate
(cm
3
/min)
0
2.54 cm OD
2 x 2.54 cm OD
2 x 2.54 cm OD, 50% fill
DESIGN MAX
DESIGN MIN
2 x 2.54 cm OD, 50% fill
2.54 cm OD
2 x 2.54 cm OD
Figure 25. Volumetric feed rate as a function of screw speed and configuration
After 1.0 in (2.54 cm) was selected as a suitable auger size, the reactor dimensions were
developed. Based on suitable auger speeds as shown in Figure 25, the superficial linear velocity of
heat carrier and biomass could be determined based on the pitch of the auger and the rotational speed.
Based on velocity and volumetric feed rate, the minimum required cross-sectional area to transport
the materials can also be determined. To ensure sufficient volume for mixing operations, a factor of
1.3 was included to increase the minimum required cross-sectional area. Refer to Appendix A for
details on the cross-sectional area requirements analysis.
With the auger dimensions specified, and the minimum area and volume requirements
known, the reactor dimensions and geometry were drafted in the computer aided design package
SolidWorks 2005. To eliminate any potential “dead space” between auger flighting where biomass
and heat carrier are not able to mix, the reactor cross-section is omega-shaped (ω) rather than the
rectangular design used by FZK. For a single auger design the reactor cross-section can be circular.
After the reactor cross-section was designed, the reactor length was determined based on an
iterative procedure to analyze the vapor residence time. Based on the known cross-section of the
reactor, the auger geometry, and the fill specifications for the solids, the volume for vapors to occupy
is known. Recall that for this analysis, the term “vapor” is used to describe the pyrolysis products

46
exiting the reactor: condensable vapors, aerosols and non-condensable gases. Based on the vapor
properties analysis mentioned previously, the vapor velocity can now be determined, and with an
assumption for reactor length the residence time can be determined. As discussed previously, the
“optimal” residence time for fast pyrolysis is well documented, and typically around 2 seconds or
less. Note that the residence time for vapors in the auger reactor is largely independent of the auger
speed, and hence also independent of the solids residence time which is directly dependent on the
auger speed. The length consideration for the vapor residence time analysis was the center-to-center
distance from the biomass inlet to the vapor outlet. This length was chosen such that the residence
time in the reactor was less than 1 second. So as to provide a mechanism for varying the residence
time, however, five vapor outlet ports were incorporated into the reactor lid design. The residence
time at the first outlet port was calculated to be less than 0.4 seconds. The first vapor outlet port is
located an axial distance of 4.25 in (10.795 cm) from the heat carrier inlet, and the reaming ports are
each spaced 2 in (5.08 cm) apart. The top view of the final reactor lid is shown in Figure 26, with
dimensions in inches.
Vapor outlets
Biomass inlet
Heat carrier inlet
Figure 26. Reactor lid drawing with dimensions in inches
It is worth noting that the vapor residence time calculations are based on many assumptions
and are likely to be accurate within ± 30%. The analysis is especially difficult for this type of reactor
configuration (compared to fluidized beds, for instance) given the limited amount of design
references available. The internal volume of the reactor that is occupied by the vapor products is
difficult to calculate, based on the unknown level of solids inside the reactor as a function of axial
length. Heat transfer and reaction effects were also not taken into account for this residence time

47
analysis, implying that biomass is basically instantly converted to reaction products upon entering the
reactor. Nonetheless, refer to Appendix A for detailed calculations as part of the residence time
analysis.
The heat carrier residence time analysis is also shown in Appendix A, noting that this
residence time is based only on reactor length and linear velocity, which is only a function of screw
speed and geometry. For typical operating speeds as predicted by Figure 25, the heat carrier residence
time is between 8 and 15 seconds as shown in Figure 109 in Appendix A. As the biomass entering the
reactor is converted into various reaction products, no biomass residence time is given. Furthermore,
as the reaction time and mechanism for biochar is unknown, it is difficult to calculate the solid
residence time but it will be similar to that of the heat carrier.
A digital rendering of the reactor design is shown in Figure 27, where the lid is shown
removed to so the augers can be seen.
Solids
canister
Biomass
inlet
Heat
carrier
inlet
Reactor
heater
Vapor
outlet port
Reactor
(Lid removed)
Figure 27. Auger reactor rendering with lid removed

48
A single 1/8 HP (92.3 W) motor was selected for the augers, and the procedure based on
power requirements for conveying bulk solid materials [97] is saved for Table 45 of Appendix A. The
motor transfers power to both augers through a system of three spur gears.
The heat carrier system was designed as a vertical electrically heated pipe, with a storage
hopper on top. The heat input requirements are known based on the heat for pyrolysis analysis as
discussed. The hopper volume and pipe dimensions were based on geometric considerations to
provide enough material for up to 3 hours of run time (depending largely on the heat carrier feed
rate). The heat carrier material is volumetrically conveyed from the vertical pipe by a horizontal 1-1/8
in (2.858 cm) OD auger into the reactor. Similar to the augers in the reactor, this auger is powered by
a 1/8 horsepower motor as determined by the analysis shown in Table 45 of Appendix A. The heat
carrier and biochar is conveyed out of the reactor into a cylindrical storage vessel sized to hold a
greater volume than the vertical heat carrier assembly. Biomass is volumetrically conveyed into the
reactor using a screw feeder. All components are housed on a portable aluminum frame. A digital
rendering of the heat carrier system as designed is shown in Figure 28.
Heat carrier
hopper
Solids
canister Reactor
Heat
carrier
heater
Aluminum
frame
heater
Biomass
feeder
Figure 28. Auger reactor system rendering

49
3.2 Lab-scale system components
Biomass feeding system. As shown in Figure 29, the main component of the biomass
feeding system is a simple, off-the-shelf volumetric screw feeder (Tecweigh Flex-Feed 05 Series
Volumetric Feeder). The descriptions for the symbols shown in Figure 29 are provided in Table 3.
The unit has a 0.5 ft3
(14.16 dm3
) capacity hopper that stores biomass, and a 0.5 in (1.27 cm) OD
auger that serves to both meter and inject biomass into the reactor. The biomass is “agitated” and
encouraged to exit the feeder by the walls of the hopper which flex alternately at the same speed as
the metering auger. The feeder has a clear polycarbonate lid to view the condition of the biomass
during a test, and is fitted with a nitrogen purge inlet. By purging a small amount of nitrogen through
the hopper, a slight positive pressure is provided to discourage the back flow of pyrolysis vapors into
the hopper. The injection auger feed tube [0.75 in (1.905 cm) OD] is wrapped with a water cooled
copper coil, which serves to remove heat conducted from the reactor to ensure that the biomass does
not begin decomposing prematurely. The cooling water for the injection auger is room temperature,
and the flow rate is manually controlled with a 22 GPH (83.3 L/hr) rotometer.
Figure 29. Biomass feeding system schematic
Refer to Table 3 for descriptions

50
Table 3. Biomass feeding system descriptions
Symbol designation Description
Temperature measurement
B Biomass inlet
Material flow
A Biomass
B Compressed nitrogen
C Nitrogen purge - Biomass hopper
D Nitrogen purge - Reactor inlet
E Cooling water (from tap)
F Cooling water return
Component
1 Gas rotometer (4.5 sL/min N2 max)
2 Feeder
3 Feeder motor, 90 VDC
4 Feeder controller
5 Metering auger, 0.5 in (1.27 cm) OD
6 Metering auger cooling coil
7 Reacor inlet cooling coil
8 Biomass exit (to reactor)
9 Liquid rotometer (1.39 L/min H2O max)
The biomass feed tube connects to the reactor with a ¾ in (1.905 cm) bored-through
compression fitting, so as a quick connection and disconnection can be accomplished. At this 90°
connection, biomass enters the reactor through a 1.5 in (3.81 cm) OD stainless steel tube, and a quick-
clamp cap on top of the inlet features an additional nitrogen purge inlet to prevent any back-flow of
vapors. The quick-clamp cap allows for easy removal to visually inspect the biomass inlet area. The
total volumetric flow rate of nitrogen to the biomass hopper and the biomass inlet is manually
controlled with a 4.5 sL/min rotometer, and the flow rate between the two is equalized as necessary.
The total volumetric flow rate of nitrogen to the system is controlled with an Alicat 20 sL/min mass
flow controller.
Directly above the biomass injection auger is a type-K thermocouple to measure the
temperature at the biomass inlet. The feeding system is positioned on aluminum rails and can slide
back and forth to allow for ease of separating the feeder from the reactor during biomass calibration
procedures. The biomass feeding system is shown in Figure 110 of Appendix A.
Heat carrier system. The heat carrier feeding system is marked by a vertical 2 in (5.08 cm)
Schedule 40 storage pipe and a 0.4 ft3
(11.33 dm3
) conical feed hopper. The 2 in (5.08 cm) heat
carrier storage pipe transitions to a 1 in (2.54 cm) Schedule 40 pipe at the bottom and mates to a
perpendicular 1-1/4 in (3.175 cm) Schedule 40 pipe. Inside the horizontal 1-1/4 in (3.175 cm) pipe is
a #18 standard size metering auger [1-1/8 in (2.858 cm) OD, 1.5 in (3.81 cm) pitch] fabricated by

51
Auger Manufacturing Specialists (Frazer, Pennsylvania). This stainless steel auger features one piece
construction with right-hand flighting and dimensions in inches as shown in Figure 30.
Figure 30. Heat carrier auger drawing with dimensions in inches
A schematic of the heat carrier system is shown in Figure 31, with a listing of descriptions in
Table 4. As the metering auger rotates, it draws material from the vertical storage pipe and conveys it
at a certain volumetric flow rate. The metering auger extends into a 90° bend, which reduces to a
vertical 1 in (2.54 cm) Schedule 40 pipe that allows heat carrier material to drop directly into the
reactor vessel. The entire heat carrier feeding system is constructed from stainless steel. A Dayton
3XA80 1/8 HP (93.21 W), 90 VDC gearmotor (60 RPM max) powers the heat carrier metering auger,
and is mounted on an adjustable bracket.
The heat carrier feeding system is electrically heated by three sets of Watlow ceramic fiber
heaters. Each set of cylindrical heaters forms a “clamshell” that wraps around the pipe and heat is
transferred radiantly from the heater surface through an air gap and the pipe wall into the interior of
the pipe. Below the hopper are two sets of 6 in (15.24 cm) x 3 in (7.62 cm) x 7.5 in (19.05 cm) [L x
ID x OD] 450W/90V “pre-heaters”, followed by a 24 in (60.96 cm) x 3.5 in (8.89) x 7.5 in (19.05) [L
x ID x OD] 1800W/240V heater. In-between the pre-heaters and the main heater is a flanged section
of pipe where the assembly attaches to the reactor frame. This section of pipe is heated with an
HTS/Amptek electrical heat tape. The horizontal feed pipe also features and electrical heat tape to
maintain the desired temperature of the heat carrier material in-between the vertical pipe outlet and
the reactor inlet. All exposed pipes and metal surfaces of the heat carrier system are insulated with
ceramic insulation material to minimize heat losses.

52
Figure 31. Heat carrier system schematic

53
Table 4. Heat carrier system descriptions
PH Pre-heater section, upper vertical pipe
HC1 Heat carrier 1, midway vertical pipe
HC2 Heat carrier 2, vertical pipe outlet
HC3 Heat carrier 3, reactor inlet (gas phase)
Control temperature
PH1C Pre-heater 1 control (air gap)
PH2C Pre-heater 2 control (air gap)
HT1C Heating tape 1 control (wall)
HC1C Main heater control (air gap)
Material flow
A Hea
B Compressed nitrogen
C Nitrogen purge - Heat carrier hopper
D Nitrogen purge - Heat carrier auger
Component
1 Heat carrier hopper
2 Pre-heater 1 (15.24 cm, 450W x 2)
2a Pre-heater 1 controller
3 Pre-heater 2 (15.24 cm, 450W x 2)
3a Pre-heater 2 controller
4 Heating tape 1
4a Heating tape 1 controller
5 Heat carrier pipe
6 Main heater (60.96 cm, 1800W x 2)
6a Main heater contoller
7 Meter auger motor, 90 VDC
7a Metering auger motor controller
8 Metering auger, 2.858 cm OD
9 Heating tape 2
10 Heat carrier outlet (reactor inlet)
t carrier
The heat carrier feed hopper has a nitrogen purge inlet, again, to provide a positive pressure
to ensure there is no back flow of pyrolysis vapors. As the heat carrier material empties from the
hopper and storage pipe, this flow of nitrogen becomes especially important to fill the displaced
volume that would otherwise create a low pressure zone that would encourage the flow of pyrolysis
vapors into the heat carrier system. There is an additional nitrogen purge inlet where the metering
auger shaft enters the horizontal feed tube, which provides a positive pressure to eliminate any air
entering the system or any pyrolysis vapors exiting the system. A 4.5 sL/min rotometer manually
controls the total volumetric flow of nitrogen to these inlets, and the flow rate between the two is
equalized as necessary.

54
There are four temperature measurements associated with the heat carrier feeding system, all
with Type-K thermocouples to measure process conditions inside the respective pipes. The first
temperature is in-between the two pre-heaters, followed by a temperature measurement halfway down
the length of the main heater, and then another location directly above the metering auger. As there is
a pipe reducer directly above this thermocouple, the heat carrier material becomes more ‘packed’ and
well mixed, giving the best indication of the entering heat carrier material temperature. This
temperature (HC2) will be referred to often. The final temperature measurement location is in the
vertical heat carrier inlet pipe; however this temperature measurement does not adequately measure
the heat carrier material temperature, and instead provides a “gas phase” temperature. All four
temperature measurement locations are in the middle of the respective pipes.
Reactor system. The reactor system is completely constructed from stainless steel. The
reactor outer dimensions are approximately 22 in (55.88 cm) x 2.5 in (6.35 cm) x 1.5 in (3.81 cm) [L
x W x H], however the cross section is “omega shaped” (ω) rather than rectangular. The two #16
standard augers [1 in (2.54 cm) OD, 1.25 in (3.175 cm) pitch], manufactured by Auger Manufacturing
Specialists (Frazer, Pennsylvania), are identical and feature one piece 309 stainless steel construction,
with left-hand flighting. The general auger dimensions in inches are shown below in Figure 32.
Figure 32. Reactor auger drawing with dimensions in inches
The augers in the reactor rotate in the same direction, and intermesh slightly (no contact). A
detail of the augers is shown in Figure 111 of Appendix A. A Dayton 3XA78 1/8 HP (93.21 W)
90VDC gearmotor (180 RPM max) drives the augers in the reactor through a solid stainless steel 5/16
in (0.794 cm) power shaft. The motor is mounted on an adjustable bracket on the opposite end of the
heat carrier inlet. In a custom housing at the motor end of the reactor, the power shaft terminates with
a spur gear that transfers power to two identical gears so as the augers rotate at the same rotational

55
speed. As with the metering auger for the heat carrier material, there is a nitrogen purge on the reactor
where the power shaft enters, which eliminates unwanted air from entering the system. Similarly, on
the opposite end of the reactor, the terminating bearings on the auger shafts are purged with nitrogen
as well. The volumetric flow rate of nitrogen on the biomass inlet side of the reactor is manually
controlled with a 8.0 sL/min (max) rotometer, while the flow rate on the opposite end of the reactor
can not be controlled [but can be inspected with a 5.0 sL/min (max) flow meter]. A ¼ in (0.635 cm)
stainless steel lid is connected to the reactor with 24 bolts, and a custom ceramic gasket is used for
sealing. In axial terms, the heat carrier material enters the reactor 2.25 in (5.715 cm) after the biomass
inlet (center-to-center). Similarly, the first product outlet port is 4.25 in (10.795 cm) from the heat
carrier inlet, or 6.5 in (16.51 cm) from the biomass inlet, center-to-center. There are 4 more product
outlet ports, each spaced 2 in (5.08) axially from one another. Each of the 5 outlet ports are 0.75 in
(1.905 cm) OD stainless steel tubes, with a height of 4 in (10.16 cm) above the reactor lid. The
reaction products can exit only one port at a time, and the remaining ports are capped off. These
features are seen in Figure 112 of Appendix A. A schematic of the reactor system is shown in Figure
33, with associated descriptions provided in Table 5.
5a
7a
5
7
6
R1
A
B
R2
R3
R4
R5
RHC
SO
F
G
E
8
HT3C
4a
4
C
9
10
11
1
3
1
2
A
1
A
Temperature
measurement
Material flow
Component
Electrical
Ac
Control
temperature
Pressure
measurement
1
D
Figure 33. Reactor system schematic

56
Table 5. Reactor system descriptions
R1 Reactor 1, gas phase, 5.398 cm
SO Solids outlet, gas phase
Control temperature
RHC Reactor heater control (air gap)
Pressure measurement
1 Reactor, gage pressure
Material flow
A Bi
B Heat carrier
C Pyrolysis products
D Compressed nitrogen
E Nitrogen purge - Reactor end
F Nitrogen purge - Reactor main 1
G Nitrogen purge - Reactor main 2
Component
1 Reactor vessel
2 Reactor augers, 2.54 cm OD
3 Vapor outlet port (5)
4 Heating tape 3
5 Reactor heater (30.48 cm, 900W x 2)
5a Reactor heater controller
6 Solids canister
7 Reactor augers motor, 90 VDC
7a Reactor augers motor controller
8 Compressed nitrogen cylinder
9 Nitrogen mass flow controller
omass
A Type-K thermocouple measures the temperature in-between each product outlet port (5
axial temperatures), with the measurement location just below the inside surface of the reactor lid as
shown in Figure 113 of Appendix A. There is an additional outlet port at the end of the reactor lid
which serves as the high pressure measurement location for a pressure transducer.
Solid materials (heat carrier and bio-char) exit the reactor at the opposite end of the biomass
feeder through a rectangular opening with approximate dimensions of 3.5 in (8.89 cm) x 1.45 in
(3.683 cm) [L x W]. These solids fall into a cylindrical stainless steel canister with dimensions of 10

57
in (25.4) x 12 in (30.48) [OD x H]. A Type-K thermocouple measures the gas phase temperature at
the solids outlet location.
Heat losses from the reactor are minimized by the use of an additional 900W/120V Watlow
ceramic fiber heater. This heater, positioned in-between the heat carrier inlet and the solids outlet, has
dimensions of 12 in (30.48 cm) x 3.5 in (8.89) x 7.5 in (19.05) [L x ID x OD]. All major exposed
metal surfaces of the reactor are insulated with ceramic insulation material to minimize additional
heat losses. The major reactor equipment is housed on a heavy-duty 80/20 aluminum frame with
casters.
Product recovery system. Downstream of the product outlet tube from the reactor, 0.5 in
(1.27 cm) OD stainless steel tubing is used with additional electric heat tapes to ensure sufficient
process temperatures are maintained. In addition to the electric heat tape, ceramic insulation is used to
insulate the tube. As part of the product recovery system, a gas cyclone separator is used to remove
fine biochar particles entrained in the process stream exiting the reactor. Biochar is collected in a 1-
1/2 in (3.81 cm) OD stainless steel canister [L = 6 in (15.24 cm)], connected to the cyclone with a
quick-clamp fitting. There are outlet ports before and after the gas cyclone which are used to measure
the pressure drop across the device. There are also Type-K thermocouples before and after the
cyclone to measure the process temperature at these locations. The cyclone is shown in Figure 114 of
Appendix A. A schematic of the product recovery system is shown in Figure 34, with descriptions
provided in Table 6.
After the gas cyclone, the product stream enters a set of water cooled condensers. The
condensers are single tube heat exchangers, and feature 1-1/2 in (3.81 cm) OD 304 stainless steel
quick-clamp tubing wrapped with copper cooling coils. The quick-clamp tubing style allows for easy
disassembly between runs to allow for thorough cleaning of the condensers. The first two condenser
stages are each 17.5 in (44.45 cm) L, connected by a 4 in (10.16 cm) horizontal tee section. The vapor
stream travels down through the first stage (co-current with the cooling water flow), and up through
the first stage (counter-current with the cooling water flows), where condensed bio-oil collects on the
walls of each condenser stage and drips down into separate 250 mL Nalgene bottles. The first
condenser stage is cooled with room temperature water, and the flow rate is manually controlled with
a 22 GPH (83.27 L/hr) rotometer. The second stage is cooled with chilled water using an Elkay TR2-
10 water chiller, and the flow rate is manually controlled using a 22 GPH (83.27 L/hr) rotometer.
Condensers 1 and 2 are referred to as stage fraction 1 (SF1) and stage fraction 2 (SF2), respectively.

58
2
2
D1
D2
C1
C2
D3
E
D
C
HT4C
A
D4
VM
F
2
3
6
5
4
TO 3
B
A
1
A
Temperature
measurement
Material flow
Component
Electrical
Ac
Control
temperature
Pressure
measurement
1
1
2
3
3a
7
8
4
5
11
9
10
6
18
12
13
14
15
16
17
19
20
21
22
Figure 34. Product recovery system schematic
After the first two condenser stages, the process stream exits though a 0.5 in (1.27 cm) OD
stainless steel tube and enters an electrostatic precipitator (ESP) collection device [106]. The ESP is
constructed of 2 in (5.08 cm) OD quick-clamp tube fittings, and is approximately 22 in (55.88 cm) L
(inlet center-outlet center). A 1/8 in (0.3175 cm) stainless steel rod (or electrode) hangs down through
the center of the ESP and is charged with approximately -15kV by using a Glassman Series ER high
voltage power supply (30 kV max). The outer body of the ESP is grounded through the power supply,
and the 15kV voltage difference encourages liquid bio-oil aerosol droplets to be attracted to the ESP
walls. The ESP device basically serves to dis-entrain and collect any liquid aerosols (“bio-oil mist”)

59
in the process stream. Bio-oil that collects along the inner walls of the ESP drips down and is
collected in a 250 mL Nalgene bottle. In-between the second condenser stage and the inlet to the ESP
is a temperature measurement with a Type-K thermocouple, as well as a port that serves to determine
the pressure drop across various components. The ESP is referred to as stage fraction 3 (SF3).
After the ESP, the process stream flows through a flexible tube into another condenser, this
one a 3/8 in (0.9525 cm) OD stainless steel coil placed in an ice bath container. This fourth and final
bio-oil collection device serves to drop the process temperature to below atmospheric, and remove as
much moisture and as many condensable products as possible. Condensed bio-oil is collected in a
“tee section” at the end of the coil, before the coil exit. At the exit of the coil condenser, there is a
final temperature measurement and pressure port, as well as a 0-5 in-H2O pressure gauge to ensure
there is a slight positive pressure within the system. The third condenser is referred to as stage
fraction 4 (SF4).
Table 6. Product recovery system descriptions
D1 Downstream 1, cyclone inlet
D2 Downstream 2, SF1 inlet
C1 Condenser 1 (wall)
C2 Condenser 2 (wall)
D3 Downstream 3, SF3 inlet
D4 Downstream 4, SF4 outlet
VM Volume meter inlet
Control temperature
Pressure measurement
2 Cyclone, differential
3 SF1 - SF2, differential
4 SF3 - SF4, differential
5 SF4 outlet, gage
6 Volume meter inlet, gage
Material flow
A Pyrolysis products
B Cooling water (from tap)
C Condenser 1 cooling water
D Condenser 2 cooling water
E Cooling water return
F Non-condensable gas (to vent)

60
Table 6. (Continued)
Component
1 Gas cyclone
2 Biochar collection canister
3 Heating tape 4
6 Chiller
7 Condenser 1
8 Cooling coil
9 SF1 collection bottle
11 Condenser 2
12 Electrostatic Precipitator (ESP)
13 SF3 collection botttle
14 ESP electrode
15 Power supply, 30 kV
16 Condenser 3
18 Ice bath
19 Gas drier tube
20 Vacuum pump
21 Micro GC
22 Volume meter
The remaining fast pyrolysis products in the permanent gas phase are passed through a
packed bed of desiccant to remove any further moisture or particulate matter that might be remaining.
A Gast vacuum pump aids in overcoming the pressure drop associated with flowing the product
stream through the packed bed of desiccant, and a loop in the process stream around the vacuum
pump helps to maintain a slight positive pressure throughout the entire system. After the vacuum
pump, the gas is analyzed in-situ with a Varian CP-4900 Micro-Gas Chromatograph (Micro-GC).
Before venting, the total gas volume is measured in an Excel TY-LNM-1.6 diaphragm meter (2.5
m3
/hr max). The gage pressure and temperature are measured at the volume meter inlet with a 0-5 in-
H2O pressure gauge and Type-K thermocouple, respectively. Refer to Figure 115 in Appendix A for
an image of condensers 1 and 2, and Figure 116 and Figure 117 for images of the ESP and condenser
3, respectively. Refer to Figure 118 of Appendix A for an overview picture of the reactor system.

61
Data acquisition and control system. The data acquisition system is based off a National
Instruments cDAQ-9172 system, which has a simple 8-slot “plug and play” chassis and USB
interface. Two NI9263 analog output modules (4-channel, ± 10V) are used to provide signals for
controlling various devices such as the biomass feeder or nitrogen mass flow controller. One NI9205
analog input module (32-channel, ± 10V) is used to monitor voltage inputs from pressure transducers
and other devices. Five NI9211thermocouple input modules (4-channel, ± 80mV) are used to measure
up to 20 process temperatures. The NI hardware communicates with a Dell Optiplex 755 PC through
a single high speed USB cable and LabVIEW 8.2 software. A LabVIEW program was developed to
both monitor process conditions during a test, as well as record important data during the test. A
screenshot from this program is shown in Figure 35.
Figure 35. LabVIEW program screenshot for data acquisition and process monitoring
The control system is split into various components on the reactor frame. On the feeder side
of the reactor, the pre-heaters and heat tapes associated with the heat carrier feed system are

62
controlled with a stand-alone 6-channel ‘heater control box’. Both of the 6 in (15.24 cm) ceramic pre-
heaters and both electrical heat tapes associated with the heat carrier feed system are controlled with
Series 16A PID temperature controllers from Dwyer Instruments/Love Controls. Each of these
controllers has feedback temperatures from Type-K thermocouples attached to the surface of the heat
carrier pipe in their respective locations. The electrical heat tapes on the downstream product section,
in-between the reactor exit and the condenser inlet are controlled with Series 32A PID temperature
controllers from Dwyer Instruments/Love Controls located on the face of an electrical enclosure on
the far end of the reactor. Similarly, each of these controllers has a feedback temperature from Type-
K thermocouples that measure the surface temperature of the 0.5 in (1.27 cm) OD process tube.
The main 24 in (60.96 cm) heat carrier heater and the 12 in (30.48 cm) heater around the
reactor are controlled with EZ-ZONE PID controllers from Watlow, located on the opposite end of
the reactor frame in a dedicated control box. The heat carrier heater feedback temperature is from a
Type-K thermocouple measuring the surface temperature of the pipe, and the reactor heater feedback
temperature is from a Type-K thermocouple measuring the air gap temperature in-between the heater
surface and the reactor surface.
On the same side as this heater control box are the motor controllers for the heat carrier
metering auger and the reactor augers. The heat carrier metering auger and the reactor augers are
manually controlled by Dayton 4Z527 and Dayton 2M171 DC motor controllers (potentiometer
based), respectively. The biomass feeder is manually controlled by an off-the-shelf potentiometer
based controller from Techweigh.
3.3 Lab-scale system development
After the complete system was designed, constructed and all components were installed, a
development effort was undertaken to understand and refine the operation. Before high temperature
pyrolysis experiments were performed, an informal “cold-flow” mixing study was completed. The
goal of this study was to determine if the degree of mixing for solid particulate matter was a function
of auger speed or location (axial and radial) within the reactor. A gas pycnometer was used to
measure the particle density of sand and biomass mixtures sampled from different axial positions of
the reactor for various auger speeds. Details of this study can be found in Appendix B.
The novel use of a gas pycnometer to determine the particle density of solid mixtures was
found to be a poor method of characterizing solid mixtures. Subsequently, a literature review on
experimental apparatus for mixing studies confirmed that complex and specialized analytical

63
equipment is required for proper characterization of solid mixtures [107-111]. For instance Ziegler et
al. used a special color meter to determine how well two different types of chocolate were mixed in a
co-rotating twin-screw device [107]. In 2007, Jones et al. used positron emission particle tracking
(PEPT) to follow tracer material in a ploughshare mixer to study axial mixing behavior [110]. Paul et
al. describe other characterization techniques such as: diffusing wave spectroscopy, positron emission
topography, magnetic resonance imaging and X-ray tomography in a comprehensive handbook [99].
Therefore, a more qualitative approach was undertaken to study the mixing behavior of the
system. A clear polycarbonate lid for the reactor was designed so the mixing interaction of biomass
and heat carrier materials could be viewed in real time. Biomass particles and sand were fed into the
reactor at various feed rates while the auger speed was varied. In time order from left to right and top
to bottom, still images of sand and corn stover biomass (dyed green to enhance the contrast) are
shown in Figure 36 to illustrate the mixing behavior. Note the black dot on the polycarbonate lid
designates the location of the first vapor outlet port. The feed rates of biomass and sand are 1.0 kg/hr
and 20 kg/hr, respectively, in Figure 36.
By visual inspection, the “degree of mixing” between biomass and heat carrier was
considered to be excellent. The mixing mechanisms could be described as a “bulk mixing process”, in
which the materials would fold on top of the other by way of the screw flighting design. Mixing of
material between the two augers was also noted to occur. More complete mixing was observed at
lower auger speeds, with approximately 45 RPM being the ideal rotational speed for the design feed
rates. At higher auger speeds (> 70 RPM), the material is quickly conveyed through the reactor with
minimal mixing, and at lower speeds (< 35 RPM) the augers are not able to convey the materials
through the reactor before clogging problems occur. At these low speeds, material begins to build up
within the reactor and is not conveyed out quickly enough. Mechanical fluidization of the materials
was not observed.
As a result of these quantitative and qualitative mixing studies, general rotational auger
speeds were known such that actual fast pyrolysis shakedown trials commenced. In all, over 19
shakedown trials were performed to investigate the system operation and performance with various
feedstocks and conditions. Different size particles of corn stover, corn fiber and two types of wood
were tested as biomass feedstocks. Shown from left to right in Figure 37 are 1.0 mm corn stover, 1.0
mm corn fiber, and 0.75 mm oak wood.

64
Figure 36. Cold flow mixing images of cornstover biomass and silica sand

65
Figure 37. Corn stover (1.0 mm), corn fiber (1.0 mm) and red oak biomass (0.75 mm)
Different types and sizes of heat carrier materials were tested as well, as shown from left to
right in Figure 38: sand, silicon carbide, alumina ceramic, 1.0 mm steel shot and 0.7 mm steel shot. A
sample of operating conditions used during the shakedown trial phase is shown in Table 7.
Figure 38. Sand, silicon carbide, alumina ceramic and steel shot heat carrier examples
Table 7. Shakedown trials operating conditions
THC TR db dHC QN2 ωA
(°C) (°C) (kg/hr) (kg/hr) (μm) (μm) (SLPM) (RPM)
Low 425 450 0.5 12 500 400 1.0 36
High 825 750 1.0 24 1000 1200 4.0 70
b
m
 HC
m

Calibration procedures are performed to determine the proper heater temperatures to maintain
the desired heat carrier inlet temperature, THC (°C). This corresponds to temperature THC2 in Figure
31. The electrical guard heater around the reactor is set to a sufficient temperature, TR (°C), to
minimize heat losses. This corresponds to temperature TRHc in Figure 33. The biomass mass feed rate,
(kg/hr), is achieved by setting the auger speed rate on the biomass feeder, based on calibration
procedures. Similarly, the heat carrier mass feed rate, (kg/hr), is achieved by setting the
metering auger speed rate based on calibration procedures. The nominal biomass particle size, db
b
m

HC
m


66
(mm), is achieved by the screen size used in a cutting mill or based on a standard sieving procedure.
The nominal heat carrier particle size, dHC (mm), is as received by the manufacturer or based on a
standard sieving procedure. The total volumetric flow rate of nitrogen into the system, QN2 (SLPM), is
controlled by a mass flow controller and is constant for the duration of a test. The rotational speed of
the augers in the reactor, ωA (RPM), is achieved by setting the desired speed rate on the motor
controller and is constant for the duration of the test. These operating conditions correspond to the
simplified schematic of the reactor shown in Figure 39.
, db
A
, THC, dHC
QN2
TR
PRODUCTS
HEAT CARRIER
BIOMASS
NITROGEN
SOLIDS
REACTOR
HEATER
MOTOR
HC
m

b
m

Figure 39. Simplified reactor schematic with operational parameters shown
In addition to testing various feedstocks and operating conditions, the shakedown trials were
useful in refining the configuration and operation of the laboratory apparatus and finalizing the
experimental procedures. Use of the different vapor outlet ports was investigated, as were different
gas cyclones are condenser configurations. A final benefit of performing numerous shakedown trials
was to demonstrate proof-of-concept of the reactor design, including steady-state operation.
Several challenges and solutions were realized during the shakedown trial testing phase, and
will not be discussed. Refer to Table 46 in Appendix A for details of the operating conditions
performed for the shakedown trials, and Table 47 for the product distribution results. These tables
illustrate that the procedures and the lab-scale system produce respectable mass balances and
repeatable bio-oil yields within the range of reported literature for fast pyrolysis. This provided
evidence and confidence to proceed with the experimentation phase of the research which will be
described next.

67
CHAPTER 4. EXPERIMENTAL METHODS AND MATERIALS
4.1 Introduction
For this research, a Response Surface Methodology (RSM) was selected to optimize the
auger reactor design. This is a systematic methodology that allows for statistical investigation of
responses that are a function of multiple factors (or variables), including any interaction effects
between factors. For instance as discussed previously, the yields from biomass fast pyrolysis are
known to be dependent on several conditions, thus these conditions need to be investigated
simultaneously. As there is minimal data available on the auger reactor operating conditions, it is
worthy to study the effects of these conditions on responses such as bio-oil yield and composition.
RSM is a common experimental methodology used in the optimization of industrial processes [112,
113]. By constructing a theoretical model to estimate a given response, useful visual representations
and equations can be developed to maximize or minimize the response. A thorough procedure was
followed to determine a specific design of experiments to carry out the RSM, and will be discussed.
4.2 Experimental design
The first step in a RSM is the selection of an appropriate experimental design. This selection
is dependent not only on the number of factors of interest, but also the availability of resources. At
least thirteen factors associated with the reactor system were identified that have possible effects on
the product distribution and composition, as summarized in Table 8 along with the fast pyrolysis
consideration affected by each factor. As the reactor system is a first generation design, factors that
were assumed to have minimal effects were eliminated, as were factors that were not continuous
(“categorical” factors). As the lab-scale system was designed for a certain biomass feed rate (1 kg/hr),
this was also eliminated as a factor. Furthermore, changing the biomass feed rate will alter the heat
removal requirements for the bio-oil recovery system, potentially causing inconsistent system
operation. In regards to the bio-oil recovery system, the design and operation of these components
will affect the pyrolysis products; however these considerations are outside the scope of this research.
Finally, the system was designed to provide heat for pyrolysis by means of the heat carrier material,
so the reactor heater temperature was eliminated as a variable.
According to the literature review performed on biomass fast pyrolysis and solids mixing, the
remaining factors were all considered to be important enough to warrant further study. The heat

68
carrier feed rate, , is easily adjustable by means of controlling the metering auger speed, and
will intuitively effect the yields and composition because of the heat transfer effects. Similarly, the
temperature of the heat carrier material, THC, can be controlled by setting the electric heaters and the
importance of reaction temperature is well documented. The rotational speed of the augers in the
reactor, ωA, will affect the mixing behavior of the biomass and heat carrier, as discussed previously,
which is assumed to then affect the heat transfer and devolatilization of the biomass. Finally, the flow
rate of nitrogen, QN2, effects the vapor residence time in the reactor and is easily controlled.
HC
m

Table 8. Factor considerations for experimental design procedure
Factor
No.
Factor
category Factor
Fast pyrolysis
considerationa
Concern for selection
1 Type 5 Not continuous
2 Feed rate 1 Small turndown, system design
3 Particle size 1 Minimal effect compared to other factors
4 Moisture content 1 Minimal effect compared to other factors
5 Type 1,5 Minimal effect, not continuous
6 Feed rate 1,2 None
7 Particle size 1 Minimal effect, system capabilities
8 Temperature 1,2,~3 None
9 Auger rotational speed 1,2,~3 System capabilities
10 Vapor outlet port 3 Not continuous
11 Total nitrogen flow rate 3 None
12 Reactor heater temperature 1,2 Control, system design
13
Product
recovery
Condenser temperature(s) 4 Outside scope of research
Biomass
properties
Heat carrier
properties
Reactor
configuration
Note: a - Fast pyrolysis considerations: (1) Rapid heat transfer, (2) Controlled reaction temperature,
(3) Short vapor residence times, (4) Rapid cooling of reaction products, (5) Catalytic effects
With four factors, n, selected (n = 4), the experimental design selection process was
continued. As mentioned previously, a design was required to study possible interactions between
factors and develop a response surface, so a 2n
factorial design was eliminated [113]. A 3n
factorial
design could be used to develop this response surface; however 81 experiments are required for four
factors which were deemed impractical to implement. Therefore, a Central Composite Design (CCD)
was selected as a suitable design for the response surface methodology, and is often used in place of
3n
factorials to minimize experimental time and expenses [113]. Out of the possible CCD options, the
commonly used “circumscribed option” was selected (can be referred to as CCC) as it allows for
investigation of a large design space [112]. Other CCD options such as the “inscribed” (CCI), “face
centered” (CCF), and Box-Behnken designs may require fewer runs than the CCC, but have a more
restricted experimental space [112, 113]. The circumscribed central composite design is called such

69
because it has a 2n
factorial design imbedded within “axial points” as shown in Figure 40. Note this
diagram only shows two factors, as all four factors can not be shown conveniently in two (or even
three) dimensions. The axial points, also called “star points” [112], test conditions outside the main
design space to help generate the curvature of the quadratic model. Note that typically all the points
are given coded coordinates, with the so-called “center-points” having coordinates of (0, 0), and axial
points at a distance “α” from the center point. Center point replicates are performed to help establish
the experimental error [113].
+1,+1
+1,-1
+ , 0
0, -
0, +
-1,-1
-1,+1
0, 0
, 0
CENTER
POINT
FACTOR X1
COORDINATE
FACTOR
X
2
COORDINATE
Figure 40. Central Composite Design schematic for two factors
Image adapted from Kuehl [113]
The number of experiments, N, and the α value required for a CCD with n factors and m
center point replicates are calculated by Equations 3 and 4 respectively.
m
n
2
2
N n
+
⋅
+
= Equation 3
1/4
n
)
(2
=
α Equation 4

70
For 6 center point tests (m = 6), this results in a four factor, five level central composite
design requiring 30 experiments and an α value of 2.0. This α value implies that, for a given level,
the “step” from the factorial point (1) to the axial point (α) is the same as from the center point (0) to
the factorial point (1). The levels for the design were chosen based on information gathered or
determined during the literature review, engineering design and shakedown trial portions of the
project. In selecting levels for this type of design, there is a tradeoff between what the experimental
apparatus can physically achieve and what will allow for a suitable response surface to be developed.
The final factors and levels chosen for the design are shown in Table 9, with notation as discussed
previously.
Table 9. Selected factors and levels for experimental design
`
THC
(°C)
QN2
(SLPM)
ωA
(RPM) (kg/hr)
−α 425.0 1.5 45.0 9.0
- 1 475.0 2.0 49.5 12.0
0 525.0 2.5 54.0 15.0
+1 575.0 3.0 58.5 18.0
+α 625.0 3.5 63.0 21.0
Level
Factor
HC
m

As noted, the resulting model from this experimental design procedure is quadratic (second
order) with 15 coefficients as shown in Equation 5, and serves to estimate the response surface [113].
There is an intercept term, βo, and 14 remaining coefficients associated with each factor (4), each of
the interaction terms between factors (6), and each factor squared (4). Note that the response in
Equation 5, Yi, is general and different models can be developed for any number of responses.
2
HC
44
2
A
33
2
N2
22
2
HC
11
HC
A
34
HC
N2
24
HC
HC
14
A
N2
23
A
HC
13
N2
HC
12
HC
4
A
3
N2
2
HC
1
o
i
m
β
ω
β
Q
β
T
β
m
ω
β
m
ω
β
m
T
β
ω
Q
β
ω
T
β
Q
T
β
m
β
ω
β
Q
β
T
β
β
Y





⋅
+
⋅
+
⋅
+
⋅
+
⋅
⋅
+
⋅
⋅
+
⋅
⋅
+
⋅
⋅
+
⋅
⋅
+
⋅
⋅
+
⋅
+
⋅
+
⋅
+
⋅
+
=
Equation 5
s an expansion of Table 9, a list of all the experiments performed is shown in Table 10 in
the coded format. Note the three sections shown: 16 factorial design experiments, 8 axial point
experiments, and 6 center point experiments with the same conditions.
A

71
Table 10. Final central composite design, coded experiments
Factor
DOE
#
THC
(°C)
QN2
(SLPM)
ωA
(RPM) (kg/hr)
2 +1 +1 +1 - 1
3 +1 +1 - 1 +1
4 +1 +1 - 1 - 1
5 +1 - 1 +1 +1
6 +1 - 1 +1 - 1
7 +1 - 1 - 1 +1
8 +1 - 1 - 1 - 1
9 - 1 +1 +1 +1
10 - 1 +1 +1 - 1
11 - 1 +1 - 1 +1
12 - 1 +1 - 1 - 1
13 - 1 - 1 +1 +1
14 - 1 - 1 +1 - 1
15 - 1 - 1 - 1 +1
16 - 1 - 1 - 1 - 1
17 0 0 0 −
1 +1 +1 +1 +1
α
18 0 0 0 +α
19 0 0 −α 0
20 0 0 +α 0
21 0 −α 0 0
22 0 +α 0 0
23 −α 0 0 0
24 +α 0 0 0
25 0 0 0 0
26 0 0 0 0
27 0 0 0 0
28 0 0 0 0
29 0 0 0 0
30 0 0 0 0
2
n
factorial
treatment
design
(n
=
4)
2n
axial
points
(n
=
4)
Center
points
(m
=
6)
HC
m

Similarly, Table 11 is shown for the actual experimental conditions. Note the second column
shows the order the experiments were performed in. Due to lengthy and comp
procedures for the heat carrier mass feed rate, , and inlet temperature, THC, the experiments were
random
erati
lex calibration
HC
m

ized within blocks of heat carrier feed rates that were grouped together. While it is often
preferred to completely randomize the experiments including the center point tests, minimization of
experimental error is also an important consid on. Calibrating the system for one group of feed
rates and completing that block of experiments was determined to be the best option for maintaining
and repeating the operating conditions of the system.

72
Table 11. Final central composite design, actual experiments
DOE
#
Run
#
THC
Factor
(°C)
VN2
(SLPM)
ωA
(RPM) (kg/hr)
2 24 575 3.0 58.5 12
3 10 575 3.0 49.5 18
4 29 575 3.0 49.5 12
5 8 575 2.0 58.5 18
6 26 575 2.0 58.5 12
7 4 575 2.0 49.5 18
8 23 575 2.0 49.5 12
9 5 475 3.0 58.5 18
10 27 475 3.0 58.5 12
11 7 475 3.0 49.5 18
12 25 475 3.0 49.5 12
13 6 475 2.0 58.5 18
14 28 475 2.0 58.5 12
15 3 475 2.0 49.5 18
16 30 475 2.0 49.5 12
17 1 525 2.5 54.0 9
18 2 525 2.5 54.0 21
19 14 525 2.5 45.0 15
20 16 525 2.5 63.0 15
21 11 525 1.5 54.0 15
22 18 525 3.5 54.0 15
23 13 425 2.5 54.0 15
24 20 625 2.5 54.0 15
25 19 525 2.5 54.0 15
26 21 525 2.5 54.0 15
27 17 525 2.5 54.0 15
28 12 525 2.5 54.0 15
29 15 525 2.5 54.0 15
30 22 525 2.5 54.0 15
2
n
factorial
treatment
design
(n
=
4)
Center
points
(m
=
6)
2n
axial
points
(n
=
4)
1 9 575 3.0 58.5 18
HC
m

4.3 Experimental materials
he biomass used for this research was northern red oak (Quercus Rubra L.) obtained from
Wood Residuals Solutions (Montello, WI). Red oak is a hardwood species in the Beech family used
e eastern United States [114]. This biomass was chosen based
on two
T
for lumber, and is native to most of th
factors: superior performance as determined during shakedown testing, and the ability to
compare the results to other pyrolysis studies using oak wood. Often used as animal bedding, this oak
wood was kiln dried before delivery in a ‘super-sack’ to the BECON facility in Nevada, IA. Here it
was first processed into more homogenous sized particles in an Art’s Way Manufacturing stationary

73
hammer-mill with a 1/8” screen size, as shown in Figure 125 of Appendix C. Further size reduction
was accomplished using a Retsch SM 200 heavy duty cutting mill with a 0.75 mm screen size, as
shown in Figure 126 of Appendix C. This size was selected to minimize heat transfer limitations.
Other than size reduction, no further drying or pre-treatment steps were carried out before testing.
After grinding, the biomass was stored at ambient conditions in 5 gallon plastic buckets with sealed
lids. The red oak biomass is shown in Figure 41 from left to right: as received, after hammer mill
processing with a 1/8” screen, and after knife mill processing with a 0.75 mm screen.
Figure 41. Red oak biomass samples of three different grind sizes
Soil Control Lab (Watsonville, CA) analyzed the composition of the red oak biomass on
April 21, 2009 with icellulose
nd lignin account for over 93% of the mass, and that the biomass has a low ash content.
results as shown in Table 12. These results shown that cellulose, hem
a
Table 12. Red oak biomass composition
1
Component Results Notes on method
Fats, Waxes and Oils 0.1 Ether extract
Resins cohol extraction
Water soluble polysacchardies 1.7 Hot water extraction
Hemicellulose 20.0 Hydrolysis with 2% HCl
Cellulose 29.8 Hydrolysis with 80% H2SO4
Protein 0.5 Total Nitrogen X 6.25
Lignin-humus 43.3 Total carbon X 1.724
Ash 0.3 550 deg. C
Total 97.3
Other or missing components 2.7
Percent Moisture 4.8
1 - Percent dry weight, unless otherwise noted
1.5 Al

74
The elemental composition of the biomass was determined with a LECO TruSpec CHNOS
analyzer as shown in Figure 127 of Appendix C. Carbon, hydrogen and nitrogen were analyzed based
on the ASTM D5373 standard, and ASTM D4239 was referenced for the sulfur analysis. Thermal
gravimetric analysis methods and ASTM D5142 were used to determine the ash content and perform
the proximate analysis of the biomass using a Mettler Toledo Stare
System as shown in Figure 128 of
Appendix C. The higher heating value of the biomass was determined using standard calorimetric
methods with a Parr 1341EB oxygen bomb calorimeter as shown in Figure 129 of Appendix C.
These analyses, performed in triplicate, are summarized in Table 13.
Table 13. Red oak biomass ultimate and proximate analyses
Carbon Nitrogen Hydrogen Sulfur Ash Oxygena
Average 48.70 0.072 6.80 0.0016 0.395 44.03
Standard
deviation
3.15 0.011 0.35 0.0013 0.162 3.42
HHVb
Moisture Volatiles Fixed carbon Ash Total (MJ/kg)
Average 3.86 81.90 12.56 0.395 98.72 18.05
Standard
deviation 1.11 0.39 0.45 0.162
Proximate Analysis (%-wt., ar)
Ultimate Analysis (%-wt., ar)
1.12 0.87
Notes: a - Oxygen by difference. b - Higher heating value. ar - As received
The heat carrier used for this research was AMASTEEL cast steel shot from Irvin Industries
(Ann Arbor, MI), and is typically used for abrasive or shot peening applications [115]. Steel shot was
selected as a heat carrier based on superior performance as determined during shakedown testing.
Compared to sand, steel shot is denser and more thermally conductive, and is less likely to clog upon
becoming moist. Though not important for this study, a potential downside to steel shot compared to
sand is its inability to be conveyed pneumatically. The steel shot size used was S-280, and though the
“280” indicates a nominal diameter of 0.028 in (0.71 mm), the official designation is a distribution
based on SAE J827 standards as shown in Figure 42.
The composition of the steel shot and select properties (as provided by Irivin Industries and
not tested) is shown in Table 14 [115].

75
To ensure similar composition of steel shot between tests, 1500 lbs (680.4 kg) was obtained
from a single manufactured lot through LS Industries (Wichita, KS). The steel shot was stored at
ambient conditions in sealed 50 pound bags, and fresh steel shot was used for each experiment.
Figure 42. SAE J827 steel shot size distribution
Image source: Marco U.S.A. [116]
Table 14. Steel shot composition and select properties
Element %-wt.
Iron > 96.0
Carbon < 1.20
Manganese < 1.30
Silicon < 1.20
Chromium < 0.25
Copper < 0.20
Meltin 1583
Nickel < 0.20
Specific gravity (@ 15.6°C) > 7.6
g point (°C) 1371 -
4.4 Testing procedures
The RSM was carried out by performing three major types of testing: reactor operation to
determine the fast pyrolysis product distribution, analytical testing to determine the composition of
the bio-oil and biochar that was produced, and statistical methods to analyze and evaluate the data.

76
Extensive graphical methods were also carried out to interpret and analyze the results. These three
procedures will be discussed independently next.
4.4.1 Product distribution
The product distribution for each of the CCD runs was determined by performing
experiments with the lab-scale reactor system previously described. The product yields are
determined gravimetrically in the case of bio-oil and biochar while the mass of non-condensable gas
calculated from its volumetric yield. With notation as discussed previously, the operating
conditions for the experiments are shown Table 15.
is
Table 15. Experimental operating conditions
THC TR db dHC QN2 ωA
(°C) (°C) (kg/hr) (kg/hr) (μm) (μm) (SLPM) (RPM)
425 - 625 550 1.0 9.0 - 21.0 750 711 1.5 - 3.5 45 - 63
b
m
 HC
m

The biomass feed rate is controlled by setting the motor speed on the biomass feeder, and is
calibrated to feed a relatively constant mass rate. As the feeder conveys material volumetrically, small
vary the mass feed rate slightly. Therefore, the rate is given as an
average
mass is then weighed with a 2100 x 0.01g Ohaus
xplorer scale and placed into the feed hopper. This mass is denoted as mb in the mass balance
schematic shown in Figure 43, and recorded on the mass balance worksheet as shown in Figure 44.
Note the biomass feed is not begun until the eratures.
Similarly, the steel shot feed rate is controlled by setting the motor speed on the heat carrier
metering auger motor controller. The calibration procedure includes recording the time required to
feed a known amount of heat carrier through the reactor, and is performed for a minimum of one hour
of feed
23 kg of steel shot is then weighed and placed in the feed system. A 64 kg x 0.1g Sartorius FBG-64
fluctuations in bulk density will
over the duration of an experiment. Calibration tests are a minimum of ten minutes each and
performed for at least five speed settings.
Before each pyrolysis run, the moisture content of the biomass is determined by heating a 4 g
sample to 105°C using an Omnimark Mark 2 Standard moisture scale as shown in Figure 130 of
Appendix C. Approximately 1200 g of prepared bio
E
system has reached operating temp
time for three different speed settings. Typical rotating speeds for the heat carrier metering
auger are 15 - 30 RPM. Depending on the feed rate required for a specific experiment, approximately

77
EDE-H scale is used to determine the mass of the heat carrier. The steel shot mass is denoted as mHC
in Figure 43, and is recorded on the mass balance worksheet shown in Figure 44. The important mass
balance symbols used in Figure 44 are listed in Table 16.
Figure 43. Reactor system schematic showing mass balance
Refer to Table 16 for nomenclature
Important components are then cleaned, weighed and installed on the system. These include
the solids canister, the cyclone catch, condenser 1 (SF1) and condenser 2 (SF2) , the ESP collection
bottle (SF3), and the third condenser coil (SF4). The masses are all recorded on the mass balance
worksheet. Each of these components is weighed on the Ohaus scale, except the condensers which are
each weighed on the Sartorius scale separately.
The electric heaters associated with the heat carrier system and reactor, including heat tapes,
are then initiated to began the warm-up phase of the procedure. The down-stream heat tapes in-
between the reactor and the condensers are set to 485°C. The reactor heater set point is constant for

78
all the tests, and the set point temperatures for the remaining heaters are determined based on suitable
calibration procedures. These procedures are performed to determine the correct heater temperatures
to maintain a steady heat carrier inlet temperature, THC, as a function of the heat carrier feed rate and
the final desired temperature. The set points for the heat carrier heaters range from approximately
40°C - 100°C above the required heat carrier inlet temperature, with 60°C - 70°C being the most
common range. The warm-up phase takes approximately two hours.
Table 16. Description of symbols used in mass balance procedure
Symbol Description
mNCG Mass of non-condensable gas
mHC Mass of heat carrier
mb Mass of wet biomass
mS Mass of solids (heat carrier and biochar)
mcy Mass of biochar collected in cyclone
mC Total mass of biochar
mSF1 Mass of stage fraction 1 bio-oil
m
mbio-oil Mass of total bio-oil
b,H2O Mass of moisure in wet biomass
g the warm up phase, and the total volumetric flow
rate is
Early in the warm-up phase, cooling water flow is initiated to the biomass injection auger at
12 GPH (0.757 L/min), and condensers 1 and 2 at 20 GPH (1.26 L/min) each. The chiller is started to
provide cold water to condenser 2. Around 5 gallons of Ice is added to the container where the third
condenser is located.
The nitrogen flow is also initiated durin
controlled with a mass flow controller based on the desired CCD value. As described
previously, four gas rotometers are used to split the total flow between various other components on
the system. As shown in Table 17, the volume fraction of flow through each rotometer remains
constant for each flow rate. Before the pyrolysis phase of an experiment the nitrogen gas is vented.
After sufficient heat carrier temperatures are attained, the augers in the reactor are initiated
and set to the desired CCD value using the motor controller. The controller is set to a percentage of
180 RPM (the maximum speed). For instance the center point setting is 30%, corresponding to 54

79
RPM. The heat carrier feed rate is begun and a lab timer is started. The LabVIEW program is then
started to observe system temperatures and pressures during the heat carrier feeding phase. The time
quired for the heat carrier feeding phase is dependent on the feed rate, and ranges from
approximately 25 m eat carrier
inlet temperatures have been attained, the pyrolysis phase can be initiated.
Table 17. Gas rotometer settings for experiments
re
inutes to over one hour to reach steady state conditions. Once steady h
Total flow rate
QN2
(sL/min)
Reactor
(end)
Heat carrier
system
Reactor
(main)
Biomass
feed system
1.5 214.0 475.9 233.5 576.5
2.0 285.4 634.6 311.3 768.7
2.5 356.7 793.2 389.2 960.9
3.0 428.1 951.9 467.0 1153.0
3.5 499.4 1110.5 544.8 1345.2
%-vol. of total 14.3 31.7 15.6 38.4
Purge flow rate through rotometers (smL/min)
The fast pyrolysis phase of the experiment is begun by switching the flow of purge nitrogen
om the vent to the Micro-GC and gas volume meter. The ESP is then energized to -15 kV and the
LabVIE
rksheet as shown in the lower right portion of Figure 44.
he
tempera
fr
W program is set to begin collecting temperature and pressure data. The Micro-GC program
and the biomass feed are now initiated, while a lab timer is started and the volume reading on the gas
meter is recorded. Gage pressure readings at the volume meter are observed and recorded periodically
on the mass balance wo
The pyrolysis phase is continued until the biomass or heat carrier material is depleted, or until
the bio-oil collection bottles become full, and typically lasts around one hour. The shutdown
procedure begins with stopping the biomass feed and the lab timer, and recording the final volume
reading on the gas meter. This is followed by stopping the heat carrier feed and lab timer. The heaters
are then shutdown, and the water and nitrogen flows continue to cool the system while t
tures are observed. After the water and nitrogen are shut down, the condensers, the ESP bottle
and third condenser coil are removed. The final masses are determined and recorded on the mass
balance worksheet. After cooling to room temperature, the char catch and the solids canister are
removed and the masses are determined and recorded. Any biomass and heat carrier material
remaining in the system is also removed and the masses are determined and recorded.

80
Run date 550
Run ID
Run No./
Reactor heater set point temperature (°C)
Heat carrier inlet temperature (°C)
Heat carrier heater set point temperature (°C)
DOE No.
Vapor port
Note Value 1 Value 2 Unit Note Value 1 Value 2 Units Note Value Units
Moisture
content
- %-wt.
Canister
mass
- g
Initial
volume
m3
Final
volume m3
Bucket 1 g Bucket 1 g Note
Bucket 2 g Bucket 2 g
Bucket 3 g Bucket 3 g
Hopper 1 g Canister - g
Hopper 2 g In reactor g
Below auger g Bucket 1 g
Feed tube 1 g Bucket 2 g
Feed tube 2 g Bucket 3 g
Vaccum g Bucket 4 g
Auger rotational speed (% of 180 RPM)
Initial mass Initial mass
Final mass Final mass
Jared Brown
Biomass Heat Carrier NCG
ume
meter
Values
Run
operators
N2 volutermic flow rate (SLPM)
1 Heat carrier feed rate (kg/hr)
Misc. g Bucket 5 g
Start/stop Time Start/stop Time
Elapsed time - min Elapsed time - min
Note Initial Final Units Note Initial Final Units
Condenser 1 g Catch g
Condenser 2 g In cyclone g
ESP (SF3) g Misc. 1 g
Tube (SF2-3) g
Coil (SF4) g NOTES
Feed rate Feed rate
Bio-oil Biochar (cyclone)
NCG:
Pressure
readings
(in-H
2
O)
at
vol
Condensers g
SF1 bottle g
SF2 bottle g
SF3 bottle g
Misc. 1 g
Misc. 2 g
Rotometer settings
Figure 44. Mass balance worksheet for experiments
This procedure is repeated for all the central composite design experiments. Based on the data
collected during the mass balance procedures, the product distribution can be completed as follows.
As shown in Equation 6, the bio-oil yield on a “wet basis” (wb) is given as a weight
percentage of the original wet biomass mass, mb. The total collected bio-oil mass, mbio-oil, is a sum of
the individually collected fractions, SF1-SF4, as shown in Figure 43.
b
oil
bio
b
SF4
SF3
SF2
SF1
wet
oil,
bio
m
m
m
m
m
m
m
wb)
wt.,
(%
Y −
− =
+
+
+
=
− Equation 6

81
However the weight of moisture carried in by the biomass, mb,H2O, varies slightly between
experiments, so it is often appropriate to normalize the bio-oil yields to a “dry basis” (db). This is
done by calculating the yield on a dry biomass basis, and with the biomass moisture content removed
from the bio-oil mass, as shown in Equation 7. Note that the biomass moisture content is determined
with the moisture scale as discussed previously. Also note that this calculation does not “remove” any
reaction water contained in the bio-oil, only the original biomass moisture mass.
H2O
b,
b
H2O
b,
oil
bio
dry
oil,
bio
m
m
m
m
db)
wt.,
(%
Y
−
−
=
−
−
− Equation 7
The calculation of the biochar yield on a wet biomass basis is shown in Equation 8, with
notation as discussed previously and shown in Figure 44.
( )
b
C
b
cy
HC
S
C
m
m
m
m
m
m
wb)
wt.,
(%
Y =
+
−
=
− Equation 8
For notation used in Equations 6 – 8, refer to Table 16. As noted previously, the non-
condensable gas stream is analyzed with a Varian CP-4900 Micro-GC, connected to Galaxie
Chromatogrpahy 1.9 software on a Dell D630 laptop. A Varian Molsieve 5A column is used for
detecting hydrogen, oxygen, nitrogen, methane and carbon monoxide (110°C injector temperature,
100°C oven temperature, with argon carrier gas at 151.7 kPa). A Varian Pora Plot Q column is used
to detect carbon dioxide, ethylene, acetylene and ethane (110°C injector temperature, 58
temperature, with helium carrier 3 and 4
minutes, and approximately 15 analysis points are collected during the steady state portion of an
experim
n
with ga
state region
here the pyrolysis reactions are occurring.
°C oven
gas at 117.2 kPa). Each gas sampling program lasts between
ent. The Micro-GC is shown in Figure 131 of Appendix C.
The non-condensable gas yield is determined by applying the ideal gas law, in conjunctio
s analysis data from the Micro-GC and gas property data collected at the gas meter
(temperature, pressure and volume) as shown in Figure 132 of Appendix C. A characteristic sample
output from the Micro-GC (Run #24/DOE #2) is shown in Figure 45, noting the steady
w

82
0
10
20
30
40
50
60
70
80
90
100
tration
(%-vol.)
N2
H2
CO
CH4
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Analysis number
Gas
concen
C2H6
C2H4
CO2
Steady state region
Note: Run #24/DOE #2. Heat carrier temperature = 575°C, Heat carrier feed rate = 12 kg/hr,
N2 flow rate = 3.5 sL/min, Auger speed = 58.5 RPM.
Figure 45. Micro-GC gas ana sis pr
The numerical results that correspond to this graphical representation are shown in the third
column
gas produced, and the total gage
pressure
ly ofile for Run #24
of Table 18, noting the Micro-GC analyzer was able to detect approximately 96.4%-vol. of
the gas. Therefore, the concentration of gas species needs to be normalized to account for the
unknown portion by assuming the known composition of gas sums to 100% of the volume. Then,
based on molecular weights of each gas species and the normalized concentration, the “weighted
molecular weight” can be determined for each species’ contribution. The sum of these, shown in the
last column of Table 18, is assumed to be the apparent molecular weight of the non-condensable gas
mixture, MNCG (kg/kmol). During the steady-state operation as shown in Figure 45, the volume of
nitrogen passing though the meter is also known based on the Micro-GC gas composition results.
This allows for calculation of the total volume of non-condensable
and gas temperature at the meter inlet are known, so the mass (and then the yield) of the
NCG can be estimated by applying the ideal gas law. As shown in Table 18, 34 kg/kmol is a common
value for the apparent molecular weight of the non-condensable gas mixture.

83
Table 18. Non-condensable gas analysis for Run #24
Compound, i Formula
Known
concentration
(%-vol)
Mi
(kg/kmol)
Normalized
concentration
(%-vol)
yi
Nitrogen free
(kmol/kmol)
yi·Mi
Nitrogen free
(kg/kmol)
Nitrogen N2 68.68 28.01 71.22 0 -
Hydrogen H2 0.77 2.02 0.80 0.0277 0.06
Carbon monoxide CO 11.83 28.01 12.27 0.4262 11.94
Methane CH4 1.40 16.04 1.45 0.0505 0.81
Ethane C2H6 0.13 30.07 0.13 0.0045 0.14
Etheylene C2H4 0.18 28.05 0.18 0.0064 0.18
Carbon dioxide CO2 13.45 44.01 13.95 0.4847 21.33
Unknown - 3.57 - 0 - -
Sum 100 100 1.00 34.45
Note: Data from Run #24/DOE #2. Heat carrier inlet temperature = 625°C, Heat carrier feed rate = 12 kg/hr, N2 flow
rate = 3.5 sL/min, Auger speed = 58.5 RPM
A typical temperature profile of an experiment is shown in Figure 46, with data presented
from Run #20/DOE #24. rease during the
warm-up phase, and then level out to the desired value (625°C for this particular experiment). The gas
phase te
The heat carrier inlet temperature, THC, is shown to inc
mperatures inside the reactor are seen to increase with time as the heat carrier is fed, and then
decrease once the biomass feeding begins. This happens because the cold biomass enters and absorbs
heat from the reactor. However the reactor temperatures quickly steady out to a temperature ranging
from approximately 450 – 515°C, depending on the axial location in the reactor. The condenser inlet
temperature is maintained above approximately 430°C to prevent preliminary condensation of
pyrolysis products, and is seen to quickly increase once biomass feeding begins and hot vapors leave
the reactor. The downstream temperatures associated with the bio-oil recovery system are shown in
Figure 47, and the ranges are based on the temperature of the vapor products entering the reactor,
which is a function of the heat carrier temperature, THC. The wall temperature of the first condenser
quickly increases once biomass is fed into the reactor and hot vapors evolve. Note the non-
condensable gas leaving the final condenser is typically less than 15°C, however it increases to above
ambient (approximately 30°C) by the time it reaches the volume gas meter after passing through the
vacuum pump and Micro-GC.

84
0
50
100
150
200
250
300
350
400
450
500
550
600
650
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7
Temperature
(°C)
Time (hours)
HC (Inlet)
Reactor 1
Reactor 2
Reactor 3
Reactor 4
SF1 (Inlet)
SF1 (Wall)
SF3 (Inlet)
Start heat
carrier feed
Start biomass
fee
Approximate steady state region
d
Note: For Run #20/DOE #24. Heat carrier inlet temperature = 625°C, Heat carrier feed rate = 15
kg/hr, N2 flow rate = 2.5 sL/min, Auger speed = 54 RPM
Figure 46. Temperature profile example for Run #20
Figure 47. Typical bio-oil recovery system temperatures

85
4.4.2 Product analysis
After a given pyrolysis experiment was complete, the bio-oil and biochar collected were
subjected to a number of analytical tests to characterize the physical properties and chemical
composition. As applicable, the analysis methods closely follow the recommendations for testing as
commonly reported in related literature - refer to Oasmaa et al. for one such study [117]. Note that
the methods will only be described briefly here, and complete laboratory standard operating
procedures (SOPs) were referenced during the analysis. Also, not all analyses were performed for all
the bio-oil fractions or all of the biochar samples.
Sample preparation. After cooling, biochar samples from the cyclone were stored in sealed,
labeled plastic bags. Bio-oil fractions were immediately stored in sealed, separate and labeled plastic
bottles (250 mL HDPE bottles for SF1 – SF3, and 50 mL polypropylene bottles for SF4) in dark,
refrigerated conditions around 5°C. Prior to any testing procedures, all bio-oil samples were removed
from the refrigerator, and homogenized by vigorously shaking the sample bottle by hand for a
minimum of one minute, followed by stirring at 1700 RPM (Eastern Mixers 5VB-C) for a minimum
of one additional minute. Some tests require additional homogenization techniques and will be
discussed.
Various lab balanc h the numbers in
parenthesis corresponding to the test method(s) that are described below.
a. Cole-Parmer Symmetry PA220, 220 g x 0.1 mg (1, 4, 8)
b. Sartorius ME 254S, 250 g x 0.1 mg (2, 3)
c. Mettler Toledo MX5, 5 g x 1 μg (5, 7)
d. Mettler AE 100, 110 g x 0.1 mg (6)
1. Moisture content. Moisture content is an important fuel property because it affects
combustion behavior; however it is also used as an indication of bio-oil quality and has implications
for stability. After the bio-oil preparation techniques were performed, moisture content of bio-oil was
determined by the common Karl-Fischer (KF) titration method. This was accomplished by using a
MK5 KF Moisture Titrator as shown in Figure 133 of Appendix C, and referencing ASTM E203.
This is an accepted method for determining the moisture content for pyrolysis liquids.
A 20 – 30 μg sample is injected into the instrument, and is dissolved in a solvent (Hydranal
Working Medium K) and a reagent (Hydranal Composite 5K) that reacts with and consumes the
water present. A syringe e sample mass.
es are used as part of the analytical test procedures, wit
is weighed before and after sample injection to determine th

86
Calibration standards are performed with D.I. water prior to testing. Moisture content is reported on a
percent
ple in a bottle and homogenizing in an ultrasonic
water b
.
is filter paper, assisted by a
vacuum pump (Fischer Scientific MaximaDry), where the water soluble components pass through the
filter an aper, sample bottle and lid
are drie esicca vesse 5 minutes. The masses of the
filter pa d. The water insoluble
content is then determ weight of the fil ple bottle and lid from the
mass of the total bio-oil sample. Insoluble content is reported on a
percent weig
content is used, except methanol is used
weight of the wet bio-oil, denoted by (%-wt., wb).
2. Water insoluble content. Bio-oil can generally be separated into water soluble and water
insoluble fractions, though some components are partially soluble which leads to inconsistencies in
definitions used in related literature. Nonetheless, the water insoluble fraction is often referred to as
the “pyrolytic lignin” portion of the bio-oil, and is an important property for bio-oil upgrading
considerations and may reveal insight on the pyrolysis reactions.
The method starts by placing 20 mL of sam
ath (Branson B-52) for 30 minutes, followed by additional homogenization on a laboratory
shaker table (Thermo-Scientific Max Q 2500) for 30 more minutes. From this bottle, 2 g bio-oil
samples are retrieved and 10 g of D.I. water is added to each. The mixtures are homogenized using a
vortex mixer for 1 minute, after which 10 additional grams of water is added. This procedure is
repeated twice more so the final mass ratio of water to bio-oil is 20:1. The final mixture is sonicated
for an additional 30 minutes, placed on the shaker table for one hour, and rotated in a centrifuge
(Fischer Scientific accuSpin 1) at 2500 RPM for 20 minutes to fully solubulize any miscible
components
Filter paper (Whatman No. 42) is weighed after drying at 105°C for 15 minutes and cooled in
a desiccant vessel for 15 minutes. The prepared sample is poured over th
d the water insoluble components are left on the filter. The filter p
d at 50°C for 20 hours and cooled in a d nt l for 1
per, sample bottle and the sample lid are then determined and recorde
ined by subtracting the final ter, sam
initial weights, divided by the
ht of the wet bio-oil, denoted by (%-wt., wb).
3. Solids content. The solids content of the bio-oil is an important fuel property because it
affects combustion behavior and particulate emissions. As noted, the solids suspended in bio-oil are
typically fine biochar particles that were not removed by the gas cyclone, but could include sand
particles in the case of fluidized beds.
To determine the solids content, a procedure similar to that for determining water insoluble
as a solvent rather than water. This allows for all compounds
to be solubulized, except particulate matter. A 1 g sample of bio-oil is added to 12 g of ACS grade
methanol, and homogenized using a vortex mixer. Filter paper (Whatman No. 42) is weighed after

87
drying at 105°C for 15 minutes and cooled in a desiccant vessel for 15 minutes. The prepared sample
is poured over this filter paper, assisted by a vacuum pump, where the methanol soluble components
pass thr
eat of combustion, of bio-oil is a
major fu
for liquid hydrocarbon fuels. The bio-oil sample mass is typically 0.7 g, and
approxi
g, with 80 mg being a common value, and
biochar
tains many compounds other than water that volatilize at
tempera
ough and the methanol insoluble components are left on the filter. The filter paper is dried
under a fume hood for 15 minutes, dried at 105°C for 30 minutes, and cooled in a desiccant vessel for
15 minutes. The final mass of the filter paper is determined, and the change in mass divided by the
bio-oil sample mass is the solids content, reported on a percent weight of the wet bio-oil (%-wt., wb).
4. Higher heating value. The higher heating value, or h
el property of interest. Using a calorimeter, the chemical energy stored in a fuel sample (solid
or liquid) is released during combustion, and is quantified by measuring the temperature change of
2000 g of water surrounding the combustion vessel. As the instrument is well-insulated (assumed to
be adiabatic), the energy released from the combustion reaction is completely absorbed by the water,
reflected as in increase in temperature that is precisely measured.
A Parr 1341EB oxygen bomb calorimeter was used as shown in Figure 129 of Appendix C.
The instrument includes a stainless steel vessel that is pressurized to 30 atmospheres with oxygen to
ensure complete combustion. The procedure for determining the heating value is modified from
ASTM D240
mately 0.2 g of mineral oil is often added to the sample to aid in complete combustion. This is
especially required for high water content samples, and is accounted for in the combustion value
calculations. The higher heating value is typically reported in units of (MJ/kg) on a wet bio-oil basis.
5. Thermal gravimetric analysis. Thermal gravimetric analysis, or TGA, is used to
determine the mass change of a sample with increasing temperature and time. This data is used for the
proximate analysis, which gives the percent weight of moisture, volatiles, fixed carbon and ash. As
noted previously, a TGA/DSC Mettler Toledo Stare
System is used (see Figure 128 in Appendix C),
and ASTM D5142 is referenced for analyzing biochar. Calcium carbonate is used as a reference
standard. Sample masses for bio-oil range from 60 – 100 m
sample masses range from 11 – 20 mg, with 15 mg being a common value. The program
method for the TGA is as shown in Table 19.
Note that the moisture value as determined by TGA is much higher than as determined by KF
titration methods, because bio-oil con
tures less than 105°C. As such, for this study, the main property of interest as determined by
TGA is the ash content of the bio-oil and biochar.

88
Table 19. Thermal gravimetric analysis program method
Step Start Stop
Ramp - 25 105 10 N2 100
Hold 40 105 105 - N2 100
Ramp - 105 900 10 N2 100
Hold 20 900 900 - N2 100
Hold time
(min)
Temperature (°C) Heating rate
(°C/min) Purge gas
Flow rate
(mL/min)
Hold 30 900 900 - Air 100
6. Elemental composition. The elemental composition of interest for bio-oil and biochar
includes the percent weight amount of carbon (C), hydrogen (H), nitrogen (N), oxygen (O) and sulfur
(S) present. In combination with the ash content, this represents the ultimate analysis of the product.
As discussed previously, the elemental composition is determined with LECO TruSpec CHN/O/S
analyze
2.5 mgKOH/g) is added to the solvent and analyzed to
verify t
Varian CP-3800 GC and Saturn 220 GC/MS are
used as shown in Figure 135. The capillary column is a CP-19CB/CP 8722 (86%
dimethy ysiloxane phase, 14% cyanopropyl-phenyl), with dimensions of 60 m x 0.25 mm x 0.25
rs (see Figure 127 in Appendix C). This system completely combusts fuel samples (solid or
liquid), and analyzes the evolved gas products to determine the composition.
Typical sample weights are 0.1g and 0.2g for the C/H/N analyses and the S analysis,
respectively, for both bio-oil and biochar. ASTM D5291 and ASTM D1552 are referenced for
analyzing the C/H/N and S content in the bio-oil, respectively, and ASTM D5373 and ASTM D4239
are referenced for analyzing the C/H/N and S in the biochar, respectively. With C, H, N, S and ash
known, the oxygen content is determined by difference for this study.
7. Total acid number. The total acid number, or TAN, is a valuable property of interest
when comparing bio-oil to petroleum based fuels. In general, this test determines the amount of
potassium hydroxide (KOH) required to neutralize a 1 gram quantity of sample, given in units of
milligrams KOH per gram of sample (mg/g). A Metrohm 798 MPT Titrino titrator is used for the
TAN analyses, as shown in Figure 134 of Appendix C, and ASTM D664 is referenced for the
procedure. A solvent of 50%-wt. toluene, 49.5%-wt. 2-propanol and 0.5%-wt. D.I. water is prepared
at a volume of 100 mL and analyzed as a “blank” to calibrate the instrument, after which 5.0 g of
TAN standard (Fischer Scientific ST112-500,
he instrument operation. Then a 0.2 g sample of bio-oil is dissolved in 5 mL
dimethylformamide (DMF), and added to 75 mL of methanol before analysis.
8. Gas chromatography/Mass spectrometry (GC/MS). GC/MS methods are used to help
characterize the chemical composition of bio-oil. A
lpol

89
μm (length x ID x film njected sample
(99.999% helium is used as a carrier gas at 1 mL/min) and separates compounds based on the column
selection described. The compounds are then analyzed and detected in the MS portion of the
instrument using the electron ionization mode. Ideally, specific compounds are detected which
produce specific signals at a corresponding retention times. A m/z range from 30 to 300 is scanned,
and standard mass spectra with 70 –eV ionization energy is recorded. The Varian GC/MS software
package includes a NIST library that is used to match the resulting mass spectra r peak
entification if necessary.
as an internal GC/MS
standard
mpounds are grouped into the
broad ch
in an adequate torque
reading
thickness). The GC portion of the instrument vaporizes the i
fo
id
The injector temperature on the instrument is maintained at 250°C, and the GC/MS interface
is maintained at 235°C. The initial oven heating begins at 45°C for four minutes and is brought to the
GC/MS interface temperature at a heating rate of 3°C/min (63.3 minutes). The GC/MS interface
temperature is then maintained for an additional 13 minutes.
Bio-oil samples on the order of 0.25 g are diluted in HPLC grade methanol at 4.5%-wt
(95.5%-wt. methanol). The methanol solution is prepared with phenanthrene
at a concentration of 0.02%-wt. The bio-oil is homogenized with the methanol by mixing
with a vortex mixer, and then filtered with a 0.2 μm filter before placing into a GC/MS sample vial.
In addition to the phenatnthrene standard, the GC/MS instrument is calibrated to quantify the
concentration of 32 additional compounds as shown in Table 20. In addition to acetic acid and
levoglucosan (two common bio-oil constituents), the 30 remaining co
emical families of furans, phenols, guiacols, syringols, and “other GC/MS” as shown.
9. Viscosity. Viscosity of bio-oil is an important property because it affects the fluid flow
characteristics in pipes, pumps and injection nozzles on utilization equipment. Dynamic (absolute)
viscosity measurements are made with a Brookfield LV-DV-II+ Pro viscometer as shown in Figure
136 of Appendix C. This instrument determines the viscosity of a fluid by sensing the torque required
to rotate a shaft spinning at a constant rotational speed within the fluid.
Depending on the composition of a given bio-oil sample, different shaft attachments
(spindles) are used to attain a minimum amount of torque required by the instrument. Depending on
the spindle used, 3 – 16 mL of sample is required for analysis. In general, more viscous samples
require a smaller diameter spindle. The spindle speed is adjusted to mainta
(in-between 10% and 90% of the maximum), and a Thermo-Haake B7 water heater is used to
maintain the temperature of a water jacket around the sample vessel. For this study, viscosity
measurements were made at 40°C, which is a commonly reported value.

90
There are many other bio-oil and biochar analysis methods available for determination of
other properties; however these are some of the most commonly reported methods and will indicate a
broad overview of the product composition.
Table 20. Chemical compounds quantified by GC/MS analysis
Chemical compound Chemical formula
Acetic (ethanoic) acid C2H4O2
1,6-Anhydro-β-D-glucopyranose (Levoglucosan) C6H10O5
Furans
C H O
C6H8O2
2H-Pyran-2-one C6H10O3
2-furancarboxaldehyde (Furfural) C5H4O2
2-Furanmethanol (Furfuryl alcohol) C5H6O2
3-Methyl-2(5H)-furanone C5H6O2
2-Furancarboxaldehyde, 5-methyl- C6H6O2
Phenols
Phenol C6H6O
Benzene-1,4-diol (Hydroquinone) C6H6O2
Phenol, 2-methyl- (o-cresol) C7H8O
Phenol, 3-methyl- (m-cresol) C7H8O
Phenol, 4-methyl- (p-cresol) C7H8O
Phenol, 2,4-dimethyl- C8H10O
Phenol, 2-ethyl- C8H10O
Phenol, 3-ethyl- C8H10O
Guaiacols
enol, 2-methoxy- C H O
Ph 7 8 2
Phenol, 2-methoxy-4-methyl- C8 H10 O2
4-OH-3-methoxybenzaldehyde (Vanillin) C8H8O3
Phenol, 4-ethyl-2-methoxy- C9H12O2
2-Methoxy-4-(2-propenyl)phenol (Eugenol) C10H12O2
Phenol, 2-methoxy-4-(1-propenyl)-, (E)- C10H12O2
Syringols
Phenol, 2,6-dimethoxy- 9 12 3
4 methyl 2,6 dimethoxy phenol C9H12O3
Ethanone, 1-(4-hydroxy-3,5-dimethoxyphenyl) C14H12O2
Other GC/MS compounds
1-Hydroxy-2-Propanone C3H6O2
propane-1,2,3-triol (Glycerin) C3H8O
3-Hydroxy-2-butanone C4H8O2
2-Furancarboxaldehyde, 5-(hydroxymethyl) C6H6O3
2-methyl-2-cyclopenten-1-one C6H8O
1,2-Cyclopentanedione, 3-methyl-

91
4.4.1 D
es of biomass and heat carrier. The Micro-GC data was
analyzed, and the average temperature at the gas meter was determined from the LabVIEW data to
help calculate the NC
volume basis to a mass basis composition. With the mass of NCG calculated as discussed previsouly,
as well as an apparent molecular weight based on the normalized and weighted gas composition, the
number of moles of NCG produced can be calculated. Then, based on the molar concentration as
determined by the Micro-GC and each gas species molecular weight, the mass of each species could
be determined. Finally, temperature data was also analyzed to determine the average heat carrier inlet
temperature over the duration of the biomass feed time.
The SAS-JMP 6.0 statistical software package was utilized to perform the regression
modeling procedures. The parameters of the experimental design (type, factors, levels, and number of
center points) were input into the program, along with the raw data values for a given response. The
standard least squares method was selected to run the model, first with all coefficients present (“full
model”) as show previously in Equation 5. The resulting model data was then analyzed graphically
and statistically.
The residuals (distance of actual experimental data from the predicted values) were first
observed to ensure the experimental measurements were not related to each other in some way, which
would decrease the validity of the model. The assumptions required to perform a linear regression
model will not be discussed, but were reviewed by Kuhel [113] and Levine et al. [118]. The
assumptions required to perform a linear regression are assumed to hold true for this study unless
determined otherwise by analysis of the residuals.
The overall fit of the data to the model was correlated through the coefficient of
determination (R2
value), which gives the percentage of variation that can be explained by the model.
A high R2
value does not imply the fit of the model to the data is “significant”, though, and for this
purpose a simple F-test is carried out by reviewing the analysis of variance (ANOVA) table. This is
common practice for validating linear regression models. A standard ANOVA table is provided by
the JMP software, and provides the F-test statistic as the ratio of mean squares for the regression
model (MSR) and the error (MSE). For this reason the F-statistic is often referred to as the “F-ratio”.
The mean squares are determined based on the degrees of freedom (number of estimated parameters
in the model and the number of observations), and the sum of squares based on the regression model.
A sample ANOVA table is shown in Table 21, with standard notation that will not be discussed.
ata analysis and hypothesis testing
As the product distribution tests were completed, the yields of bio-oil and biochar were
calculated, as were the resulting mass feed rat
G yield. The analysis of the NCG was extended to convert the composition on a

92
Table 21. ANOVA table
Degrees of
freedom (DOF)
Sum of
squares Mean sqaure FANOVA
Regression (model) k SSR MSR = SSR / k MSR / MSE
Error ν = N-k-1 SSE MSE = SSE / ν
Total k + ν = N - 1 SST = SSR + SSE
Recall that for this study the number of experiments, N, is 30, and k represents the number of
parameters (besides the intercept term) estimated by the model. Also, note the R2
statistic is computed
as the ratio of SSR over SST, and the root mean square error (RMSE or σ) is the square root of MSE.
The RMSE approximates standard deviation of residual error, and is an important value to evaluate
the model. The F-statistic calculated by the ANOVA table can be compared to a “critical F-value”
based on the
esis in Equation 10 states that at least one coefficient is not equal to zero and
implies
degrees of freedom and a desired confidence level. For this study, the confidence level
for all analyses was selected to be 95% (α = 0.05). If the F-value from the ANOVA table is greater
than the critical F-value, then the model is considered to be significant at a 95% confidence level.
More formally, a null hypothesis, Ho, is stated such that each coefficient of the model is equal to zero
as shown in Equation 9, implying that the full regression model is insignificant and is not useful. The
alternative hypoth
that the model is significant and therefore useful for further analysis.
Ho1: β1 = β2 = … = βi = 0 Equation 9
Ha1: βi ≠ 0 (for at least one i) Equation 10
The null hypothesis is rejected if the F-value from the ANOVA table, FANOVA, is greater than
the critical F-value, Fα,k,ν, evaluated at the confidence level α, and degrees of freedom of k and ν, as
denoted in Equation 11. Refer to Table 21 for descriptions of each value.
Ho1 rejection region: FANOVA > Fα,k,ν Equation 11
The critical F-values for the F-test to determine if the model is useful are shown in the last
column of Table 22, and are based on the degrees of freedom as shown. Note that this is a general
table and for certain situations the critical value of interest is not shown. Examples of this include
reduced models that may not contain one or more of the main effects, and will be discussed.

93
Table 22. C VA F-test
ritical F-values for ANO
Model Error Total
k ν = n-k-1 N-1 F0.05,k,ν
14 (Full) 14 15 29 2.42
13 13 16 29 2.40
12 12 17 29 2.38
11 11 18 29 2.37
10 10 19 29 2.38
No. terms in
model
Degrees of freedom
9 9 20 29 2.39
8 8 21 29 2.42
7 7 22 29 2.46
6 6 23 29 2.53
5 5 24 29 2.62
4 4 25 29 2.76
3 3 26 29 2.98
2 2 27 29 3.35
1 1 28 29 4.20
In addition to the ANOVA table to evaluate the variance in the model, a “lack of fit” (LOF)
analysis is also provided by the JMP software program and is reviewed. This analysis is only possible
because of the replications performed at the center point conditions, and compares the error from the
model to that originating from the replicated experimental data. The latter is called “pure error”, and
originates from the realties of experimental apparatus and test procedures, and can not be explained
by any type of model regardless of complexity.
The “lack of fit table” is very similar to the ANOVA table, except that the first row describes
the “lac
DOF for the pure error is based on the number of center point replicates, m, as discussed previously,
and the e is based on the error DOF from NO
k of fit”, and the second row describes the “pure error” as shown in Table 23. Note that the
total DOF for the lack of fit tabl the A VA table.
Table 23. Lack of fit table
Degrees of
freedom (DOF)
Sum of
squares Mean sqaure FLOF
Lack of Fit λ = ν - (m-1) SSR MSR = SSR / λ MSR / MSE
SS SSE / (
Pure error m-1 E MSE = m-1)
Total ν = N-k-1 SST = SSR + SSE
The F-test is used again to determine if the lack of fit is considered significant, with the null
hypothesis as stated in Equation 12, the alternative hypothesis in Equation 13, and the null hypothesis

94
rejection region shown in Eq e rejected, the model
usefulness must be carefully scrutinized.
Ho2 = Lack of fit is significant Equation 12
Ha2 = Lack of fit is insignificant Equation 13
Ho2 rejection region: FLOF < Fα,λ,m-1 Equation 14
The critical F values for the lack of fit test, FLOF, are shown in the last column of Table 24,
with the degrees of freedom as shown. As with the table of critical FANOV values, Table 24 is
generalized and considers most but not all possible modeling situations.
uation 14. If the lack of fit hypothesis can not b
A
Table 24. Critical F-values for lack of fit F-test
Lack of fit Pure error Total
λ = ν - (m-1) m-1 ν F0.05,λ,m-1
14 (Full) 10 5 15 4.74
13 11
No. terms in
model
Degrees of freedom
5 16 4.70
7 17 5 22 4.59
6 18 5 23 4.58
5 4 4.57
4 20 5 25 4.56
3 21 5 26 4.55
2 22 5 27 4.54
1 23 5 28 4.53
12 12 5 17 4.68
11 13 5 18 4.66
10 14 5 19 4.64
9 15 5 20 4.62
8 16 5 21 4.60
19 5 2
Also, note that for the ANOVA F-test, a high FANOVA is desired because this implies the null
hypothe
rejection region as shown. This form of the lack of fit test is chosen based on common convention.
sis Ho1 will likely be rejected and the model can be considered significant. To reject the null
hypothesis Ho2 for the lack of fit F-test (accept Ha2), however, a low FLOF is desired based on the

95
If visual analysis of the residuals from the model verifies the assumptions to use a regression
model, and the null hypotheses for the significance of the model and the lack of fit are rejected, then
the model can be used as an approximation of the response surface [118]. If the whole model is found
to be sig i ply that all the terms in the model are
the significance of each term is also reviewed to determine if the model can be reduced by removing
terms. R lways decrease the R2
ay
value and decrease the FLOF value which implies the “reduced model” may be more significant and
less like full odel. The JMP vi
for each coefficient estimate, βi, and the null hypothesis shown in Equation 15 is rejected and the
alternati
o3,i = βi is insignificant Equation 15
Ha3,i = βi is significant Equation 16
Ho3,i rejection region: |t|i > t0.05,ν Equation 17
The critical t-values to evaluate the significance of each estimate are shown in Table 25, as a
function of the degrees of freedom, ν, as discussed previously.
After the t-test is used to determine which coefficients are significant, the regression
procedure is duplicated with insignificant terms removed and the new model is re-evaluated as
discussed. In other words hypotheses 1, 2 and 3 are tested again for the new model. To determine if
the reduced model is significant compared to the full model, a so-called “Model utility test” (MUT) is
performed. The MUT also uses the F-statistic as a means to evaluate the significance of one model
compared to another, as calculated by Equation 18.
nificant, though, that does not m significant. As such,
emoving terms from the model a s value, but m increase the FANOVA
ly to occur by chance compared the m software pro des the t-test statistic
ve hypothesis in Equation 16 is accepted if the absolute value of the t-statistic is greater than
the critical t-value as shown in Equation 17.
H







 −
=
ν
k
MUT SSE
r
k
F Equation 18



 − k
r SSE
SSE

96
Where r is the degrees of freedom in the reduced model and other notation is as described
previously. After the FMUT value has been calculated, the null hypothesis shown in Equation 19 is
either rejected to accept the alternative hypothesis shown in Equation 20, or accepted based on the
rejection region shown in Equation 21.
Table 25. Critical t-values for t-test
t0.05,ν
14 (Full) 15 2.131
13 16 2.120
12 17 2.110
11
No. terms in
model
Degrees of freedom
ν
18 2.101
10 19 2.093
9 20 2.086
8 21 2.08
7 22 2.074
6 23 2.069
5 24 2.06
4 25 2.060
3 26 2.056
27 2.052
1 28 2.048
0
4
2
Ho4 = Reduced model is less significant than full model Equation 19
erous graphical representations can be prepared
for further analysis. A three-dimensional response surface can be generated to observe the influence
of multiple factors on a given response, or two dimensional plots can be generated to observe the
effect of a single factor while the others are held constant.
Ha4 = Reduced model is more significant than full model Equation 20
Ho4 rejection region: FMUT > F0.05,k-r,ν Equation 21
After statistical analysis of the models, num

97
As a summary, the hypothesis tests are listed below in Table 26, noting that if the null
hypothesis can be rejected based on the region and test statistic shown, then the alternative hypothesis
can be accepted. As such, in general it is desired that the null hypotheses 1 – 3 are rejected and the
null hypothesis 4 is not rejected, if applicable.
Table 26. Summary of hypothesis tests
Null Alternative
Rejection region
Ho1
Regression model
is NOT significant
FANOVA > F0.05, k, ν Ha1
Regression model
IS significant
Ho2
Model Lack of Fit
IS significant
FLack Of Fit < F0.05, λ, m-1 Ha2
Model Lack of Fit
is NOT significant
Ho3,i
Parameter estimate, i,
is NOT significant
|t|i > t0.05,ν Ha3,i
Parameter estimate, i,
IS significant
Ho4
Reduced model is
LESS significant
FMUT >
F0.05, k-r, ν
Ha4
Reduced model is
MORE significant
Hypotheses Hypotheses
Refer to Table 27 for descriptions of the coefficients in the full regression model, as well as
the terms, symbols and coded symbols associated with each coefficient. Note the three main
horizontal sections correspond to: (1) the four “main effects” based on the factors selected, (2) six
interaction or “cross-terms”, and (3) four higher order terms, all as discussed previously. The coded
symbols r value (see Tab
and the difference between levels. It is important the coded values are always used when analyzing
the regr
shown in Appendix D, Equations D1 – D5. The full regression model with the coded symbols is
shown i ha values for the coded symbols are not the p
onditio s assoc ted wi that t m.
s a final and important note concerning data analysis, most bio-oil properties of interest are
extensiv , implying that the total value for a given property is the sum of property values for a number
a e used to normalize each term based on the “0” level le 9) for each factor
ession model equations. More information and sample calculations for the coded symbols are
n Equation 22, noting t t the hysical properties or
c n ia th er
HC
A
34
HC
N2
24
HC
HC
14
A
N2
23
A
HC
13
N2
HC
12
HC
4
A
3
N2
2
HC
1
o
i
μ
Ω
β
μ
θ
β
μ
τ
β
Ω
θ
β
Ω
τ
β
θ
τ
β
μ
β
Ω
β
θ
β
τ
β
β
Y
⋅
⋅
+
⋅
⋅
+
⋅
⋅
+
⋅
⋅
+
⋅
⋅
+
⋅
⋅
+
⋅
+
⋅
+
⋅
+
⋅
+
=
Equation 22
2
HC
44
2
A
33
2
N2
22
2
HC
11 μ
β
Ω
β
θ
β
τ
β ⋅
+
⋅
+
⋅
+
⋅
+
A
e

98
of “sub-
y Equation 23, where ySFi (i = 1,2,3,4) is the property
for each bio-oil fraction, and other notation is as previously discussed. It is advantageous to perform
this procedure so the resulting wh le bio-oils referenced in
the literature.
total values” [95]. In other words, if a given property is analyzed for each fraction (SF1-SF4),
the resulting property can be determined for the “whole bio-oil” (equivalent of all fractions mixed
together) by adding the weighted values of the property for each fraction. A given property for the
whole bio-oil, ybio-oil, is frequently determined b
ole bio-oil can be compared to other who








⋅
+








⋅
+








⋅
+








⋅
=
−
−
−
− oil
bio
SF4
SF4
oil
bio
SF3
SF3
oil
bio
SF2
SF2
oil
bio
SF1
SF1
oil
-
bio
m
m
y
m
m
y
m
m
y
m
m
y
y
Equation 23
Table 27. Regression model coefficients and terms
Coefficient
number
Coefficient
symbol Associated term Symbol
Coded
Symbol
1 β0 (Intercept) - -
2 β1
Heat carrier inlet temperature, THC
(Temperature)
X1 τHC
3 β2
N2 volumetric flow rate, QN2
(N2 flow rate)
X2 θN2
4 β3
Auger rotational speed, ωA
(Auger speed)
X3 ΩA
5 β4
Heat carrier feed rate,
(HC feed rate)
X4 μHC
6 β12 Temperature · N2 flow rate X1 ·X2 τHC ·θN2
7 β13 Temperature · Auger speed X1 ·X3 τHC ·ΩA
8 β23 N2 flow rate · Auger speed X2 ·X3 θN2 ·ΩA
9 β14 Temperature · HC feed rate X1 ·X4 τHC ·μHC
10 β24 N2 flow rate · HC feed rate X2 ·X4 θN2 ·μHC
11 β34 Auger spee · HC feed rate X3 ·X4 ΩA ·
d μHC
12 β11 Temperature · Temperature X1
2
τHC
2
13 β22 N2 flow rate · N2 flow rate X2
2
θN2
2
14 β33 Auger speed · Auger speed X3
2
ΩA
2
15 β44 HC feed rate · HC fe d r te X4
2
e a μHC
2
HC
m


99
CHAPTER 5. RESULTS AND DISCUSSION
5.1 Introduction
After the testing and analysis procedures were completed, numerous results were available
that will be discussed. First the results of the product distribution testing will be presented, which
cludes the regression models for the yields of bio-oil, biochar and NCG. Next, the results for the
product analysis testing will be presented, which includes general results and regression m els for
certain properties of interest.
.2 Product distribution results
The product distrib which allowed for
examining the spectrum of pyrolysis. For instance Figure 48 shows the product distribution results for
all the experiments, noting that the “carried water” present in the bio-oil is simply the moisture
content of the biomass. This figure shows that the bio-oil yields on a wet basis ranged from just over
42%-wt. to almost 74%-wt. Also seen is that in general, the mass balance closures were excellent, and
for the 30 runs averaged 98.4 ± 1.08%-wt. Only one run required measurement of the non-
condensable gas yield by difference. The feedstock data including feed times and masses for biomass
and heat carrier can be found summarized in Table 50 of Appendix D. This table shows that the red
oak biomass moisture content averaged 5.8 ± 0.25%-wt. The biomass feed rate averaged 1.0 ± 0.04
kg/hr for the 30 runs, and the absolute heat carrier temperature difference between the desired value
and the value averaged over the steady state feeding time averaged 4.7 ± 3.7°C. Heat carrier feed
rates were also shown to be consistent. A subset of Table 50 is shown in Table 28 for all 6 center
point tests, plus the maximum bio-oil yield test (Run 20) and the minimum bio-oil yield test (Run 13).
These specific tests will be referred to frequently.
The yield data including masses of bio-oil and biochar collected, as well as the calculated
mass of NCG and the totals can be found in Table 51 of Appendix D. Also shown in this table is the
“dry basis” bio-oil yield as discussed (refer to Equation 7). A subset of Table 51 is shown in Table 29,
which shows the yield data for the test conditions presented in Table 28. A graphical representation of
the data in Table 29 is shown in Figure 49.
in
od
5
ution tests resulted in a wide range of product yields

100
0
10
20
30
40
50
60
100
70
-wt.,
80
90
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Run No.
(%
wb
Product
yield
)
Unaccounted
NCG
Biochar
Bio-oil
Carried water
Figure 48. Product distribution results for the 30 fast pyrolysis tests
Table 28. Sample experimental conditions for 8 selected tests
Run
No.
DOE
No.
Run
Date
Moisture
content
(%-wt.)
Mass fed
(g)
Feed time
(min)
Feed rate
(kg/hr)
Mass fed
(g)
Feed time
(min)
Feed rate
(kg/hr)
Average
temperature
(°C)
12 28 14-Apr 5.88 910.4 55.7 0.982 23361 94.9 14.8 528.8
15 29 24-Apr 6.01 867.5 53.5 0.973 22796 90.0 15.2 536.5
17 27 30-Apr 6.04 882.8 53.9 0.983 22806 93.5 14.6 527.7
19 25 4-May 5.94 925.1 55.9 0.994 22987 90.5 15.2 529.5
21 26 6-May 5.64 964.1 60.0 0.964 23197 94.6 14.7 538.7
22 30 7-May 5.72 919.0 56.9 0.969 22443 91.0 14.8 535.6
20 24 5-May 5.93 1026.7 59.8 1.031 24873 106.1 14.1 630.5
13 23 21-Apr 5.64 994.8 57.9 1.030 22816 90.3 15.2 427.8
Biomass Heat carrier
Table 29. Sample mass balance data for 8 selected tests
TOTAL
Run DOE Run Mass Yield Yiel
No. No. Date (g) (%-wt., wb) (%-wt., db) (g) (%-wt., wb) (g) (%-wt., wb) (%-wt., wb)
12 28 14-Apr 604.7 66.4 64.3 211.6 23.2 101.4 11.1 100.8
15 29 24-Apr 586.6 67.6 65.6 173.3 20.0 100.8 11.6 99.2
17 27 30-Apr 586.8 66.5 64.3 180.9 20.5 99.5 11.3 98.2
19 25 4-May 615.3 66.5 64.4 190.1 20.6 105.1 11.4 98.4
21 26 6-May 662.4 68.7 66.8 182.3 18.9 110.0 11.4 99.0
22 30 7-May 629.9 68.5 66.6 165.3 18.0 103.6 11.3 97.8
20 24 5-May 755.7 73.6
a
23 21-Apr 419.6 42.2
d Mass Yield Mass Yield
71.9 113.1 11.0 132.2 12.9 97.5
13 38.7 355.0 35.7 - 22.1 100.0
Note: a - NCG yield measured by difference for Run No. 13
Bio-oil Biochar NCG

101
73.6
42.2
67.4
35.7
20.2
11.0
22.1
11.3
12.9
2.49
1.08
0
10
20
30
40
50
60
70
80
Bio-oil minimum yield
(Run 13)
Center point average
(6 runs, same conditions)
Bio-oil maximum yield
(Run 20)
Product
yield
(%-wt.,
wb)
Bio-oil
Biochar
NCG
Unaccounted
*
* By difference
Figure 49. Pyrolysis product distribution range
Extensive exp ta acquisition
hardware and LabVIEW software as discussed. Temperature profiles for the duration of each
experiment were plotted, and a suitable steady state region was determined as shown in Figure 46.
The collected data was then averaged over this region for each experiment. The average pressure in
the reactor for the steady state operating region, averaged for all experiments, was negligibly above
atmospheric at 2.0 ± 0.58 in-H2Og. Refer to Table 52 of Appendix D for reactor pressure data, heat
carrier system temperature data and biomass inlet temperature data. Similarly, refer to Table 53 in
Appendix D for reactor temperature data, recalling that the thermocouples at these locations measure
vapor phase temperatures (see Figure 113 of Appendix C). Product recovery system temperature data
can also be found for all experiments in Table 54 f Appendix D. The steady state operating
temperatures averaged fo locations on the
reactor schematic in Figure 50. Though the heat carrier system temperatures (422°C, 506°C and
533°C as shown circled in Figure 50) vary for each experiment, the remaining values shown in Figure
50 are highly characteristic of the overall operation of the system.
As discussed previously, the bio-oil was collected in four sequential stage fractions as
follows: warm condenser (SF1), cool condenser (SF2), electrostatic precipitator (SF3), and an ice-
cooled condenser coil (SF4). It was found that the mass distribution among stage fractions was
largely independent of test conditions. As shown in Figure 51, the average distribution at the six
erimental testing data was collected using National Instruments da
o
r the six center point tests are shown at their respective

102
center point tests (same conditions) was very similar to the overall average distribution for 30 tests
(varying conditions). The stage fraction mass distribution data is shown in Table 55 of Appendix D.
On average, over 98% of the mass of bio-oil was collected in stage fractions 1 – 3.
Figure 50. Average operating temperature schematic for 6 center point runs
0
Overall average
(30 runs, different conditions)
F
10
20
30
40
50
60
raction
of
total
(%-wt)
SF1
SF2
SF3
SF4
Figure 51. Bio-oil fraction distributions for 6 center point tests and for all tests

103
Bio-oil yield. The full regression model for total bio-oil yield (Equation 6) resulted in the
statistical analysis summarized in Table 30 and additional details are saved for Table 56 of Appendix
D. The residuals of the model, as shown in Figure 137 of Appendix D, were reviewed and determined
to be sufficient for satisfying the assumptions to use a linear regression model. A high R2
value (>
98%) and a low RMSE value compared to the response (< 1.2 %-wt., wb) indicated an excellent fit of
the data to the model. The null hypothesis Ho1 is rejected for the full model according to the F-test,
implying the alternative hypothesis Ha1 is accepted and the model is considered significant at a 95%
confidence level. In other words, with a p-value (area to the right of the critical F-value on the F-
distribution) less than 0.0001, there is basically zero probability of obtaining a higher FANOVA value by
chance if Ho1 were true. The null hypothesis Ho2 is also rejected, and the alternative hypothesis Ha2 is
accepted to imply there is no significant lack of fit. This implies the regression model is adequate.
Table 30. Bio-oil yield model, statistics summary
Statistic Value Significant Value Significant Hypothesis tests
R2
0.988 - 0.984 - -
FANOVA 91.22 √ 163.1 √ FANOVA > F0.05,k,ν *
F0.05,k,ν 2.424 - 2.420 - Reject Ho1
FLOF 1.19 X 1.13 X FLOF < F0.05,λ,m-1 *
F0.05,λ,m-1 4.74 - 4.60 - Reject Ho2
t0.05,ν 2.13 - 2.08 - -
|t| statistics
for model terms Value Significant Value Significant Hypothesis tests
β0 145.56 √ 206. √ |t| > t0.05,ν Reject Ho3
β1 31.79 √ 32.20 √ |t| > t0.05,ν Reject Ho3
β2 2.73 √ 2.76 √ |t| > t0.05,ν Reject Ho3
β3 2.26 √ 2.29 √ |t| > t0.05,ν Reject Ho3
β4 9.84 √ 9.97 √ |t| > t0.05,ν Reject Ho3
β12 1.02 X - - |t| < t0.05,ν Don't reject Ho3
β13 4.37 √ 4.42 √ |t| > t0.05,ν Reject Ho3
β14 2.26 √ 2.29 √ |t| > t0.05,ν Reject Ho3
β11 11.17 √ 11.22 √ |t| > t0.05,ν Reject Ho3
β44 3.64 √ 3.46 √ |t| > t0.05,ν Reject Ho3
FMUT FMUT < F0.05,r-k,ν
F0.05,r-k,ν Don't reject Ho4
0.91
2.79
Full Reduced
99
Note: * The null hypotheses Ho1 and Ho2 are rejected for the full model and the reduced model

104
After the full model was considered significant, the t-test was performed for each term to
accept or reject Ho3, and 6 terms were found to be insignificant as shown in Table 30. The reduced
model was also found to be significant (F-test to reject Ho1), with no significant lack of fit (F-test to
reject Ho2), and was determined to be more significant than the full model (using F-test to accept Ho4)
with results also shown in Table 30. These results imply the reduced model provides an adequate
estimate of the response surface and can be investigated further.
As shown in Table 30, the four main factors of the experimental design were all found to be
significant, as were two interaction effects and two higher order effects. Identification of significant
interaction terms and higher order terms justifies the use of the experimental design selected. The
relative significance of each of the model coefficients is shown graphically in Figure 52, noting the
vertical line of the critical t-test statistic for significance at a 95% confidence level. If the absolute
value of the t-test statistic for a given term is greater than the critical value shown, it is significant.
However by reviewing the t- ared to another can
also be determined. For instance it is easily seen that heat carrier temperature is more significant than
nitrogen flow rate in terms of bio-oil yield. Also, according to the t-tests, the terms shown in Figure
52 are the only ones to affect bio-oil yield.
The response surface form of the bio-oil yield is shown in Equation 24 below, noting the
factor coefficients are associated with the coded levels and not the physical quantity. For instance the
temperature value in Equation 24, THC, must range from -2 (–α) to +2 (+α), which correlates to the
physical quantities of 425°C and 625°C, respectively. All other values of interest can be interpolated.
Equation 24
More information on the model equation is provided in Table 57 and Equations D1 – D5 of
Appendix D. A common way to represent the fit of the model is to plot the expected values versus the
actual experimental values, as shown in Figure 53. The narrow shaded band is the 95% confidence
interval, and the broader shaded band is the 95% prediction interval associated with the fit of the data.
If one product distribution experiment was conducted as described (with known conditions that need
not be the same as those used to develop the model), the model “predicts” that the resulting bio-oil
yield will fall within the broader range of values. However if several such experiments were
statistics, the relative significance of one term comp
2
HC
2
HC
HC
HC
A
HC
HC
A
N2
HC
oil
bio
μ
0.73
τ
2.36
μ
τ
0.64
Ω
τ
1.24
μ
2.28
Ω
0.52
θ
0.63
τ
7.36
66.9
wb)
wt.,
(%
Y
⋅
−
⋅
−
⋅
⋅
−
⋅
⋅
+
⋅
+
⋅
−
⋅
+
⋅
+
=
−
−

105
conducted with the same conditions, similar to the procedure used for the center point runs, the
resulting average is expected (with 95% confidence) to fall within the narrow band of values.
0 5 10 15 20 25 30 35
HC temperature
t-test statistic absolute value
N2 flow rate
Auger speed
HC feed rate
HC temperature ·
Auger speed
HC temperature ·
HC feed rate
HC temperature ·
HC temperature
HC feed rate · HC
feed rate
Model
term
Interaction
effects
t
0.05,
21
Higher order
effects
Main
effects
Figure 52. Absolute values for t-test statistics for bio-oil yield model
70
65
60
75
o-oil
yield
,
wb)
40
45
50
55
Actual
bi
(%-wt.
40 45 50 55 60 65 70 75
Predicted bio-oil yield
(%-wt., wb)
Figure 53. Actual vs. predicted bio-oil yield

106
In general, the model reveals several insights to the operation of the reactor regarding bio-oil
yield. The three dimensional surface nature of the model response, however, presents both challenges
and unique opportunities to display and discuss these insights. For instance Figure 54 shows typical
response surface representations of bio-oil yield as a function of heat carrier temperature and each
remaining factor separately. In these plots two factors are held constant.
In Figure 54 (a), the heat carrier feed rate and the auger speed are kept constant at the center
point conditions of 15 kg/hr and 54 RPM, respectively. Though temperature is a much more
influential factor, the nitrogen flow rate is a significant factor and bio-oil yield is shown to increase
for increasing nitrogen flow rate. This is an expected result as the residence time is decreased for
increasing carrier gas flow rate. This is in accordance with Gronli & Antal [32] who discuss the effect
of low gas flow rates increasing charcoal production at the expense of bio-oil yield. This simple
correlation is also evident by inspection of Equation 24 and Figure 52 (no interaction or higher order
terms with nitrogen flow rate).
In Figure 54 (b), the heat carrier feed rate and the nitrogen flow rate are kept constant at the
center point conditions of 15 kg/hr and 2.5 sL/min, respectively. Though not immediately apparent,
this graphic shows that at lower heat carrier temperatures the yield increases for lower auger speeds,
however the rear corner of the response surface shows that at higher temperatures low auger speeds
begin to decrease the yield. This interaction effect between heat carrier temperature and auger speed
will be discussed shortl
In Figure 54 (c), the auger speed and the nitrogen flow rate are kept constant at the center
point conditions of 54 RPM and 2.5 sL/min, respectively. This response surface shows that, in
general, the bio-oil yield increases with increasing heat carrier feed rate. This may be explained by
the increased heat transfer effects associated when more heat carrier material is present.
Note the similarity between Figure 54 (a), (b), and (c) – the bio-oil yield tends to increase
continuously and quickly with increasing heat carrier temperature, and then begin to level out and
plateau at the high heat carrier temperature conditions. No apparent “maximum” point is shown after
which the yields begin to decrease, which was unexpected given the high heat carrier temperature of
625°C (recall the review of literature stating the “optimal” fast pyrolysis temperature is
approximately 500°C). This ‘anomaly’ begins to reveal interesting effects between the transfer of heat
between the hot heat carrier and the cool biomass, as well as the relationship between the heat carrier
temperature and the pyrolysis vapor reaction temperature.
y.

107
1
.
5
2
.
0
2
.
5
3
.
0
3
.
5
4
2
5
4
7
5
5
2
5
5
7
5
6
2
5
40
45
50
55
60
65
70
75
l
yiel
Bio-oi
d
N2 volumetric
flow rate
(sL/min)
Heat carrier
temperature (°C)
70-75
65-70
60-65
55-60
50-55
45-50
40-45
%-wt., wb
(a) Bio-oil yield as a function of heat carrier temperature and N2 flow rate
4
5
.
0
4
9
.
5
5
4
.
0
.
5
5
8
6
3
.
0
4
2
5
5
4
7
5
2
5
5
7
5
6
2
5
35
40
45
50
55
60
65
70
75
Bio-oil
yield
Auger speed
Heat carrier
temperature
(RPM) (°C)
70-75
65-70
60-65
55-60
50-55
45-50
40-45
35-40
%-wt., wb
(b) Bio-oil yield as a function of heat carrier temperature and auger speed
9
1
2
1
5
1
8
2
1
4
2
5
4
7
5
5
2
5
5
7
5
6
2
5
35
40
45
50
55
60
65
70
75
Bio-oil
yield
Heat carrier
feed rate
(kg/hr)
Heat carrier
temperature (°C)
70-75
65-70
60-65
55-60
50-55
45-50
40-45
35-40
%-wt., wb
(c) Bio-oil yield as a function of heat carrier temperature and heat carrier feed rate
Figure 54. Three response surfaces for modeled bio-oil yield

108
To further investigate the interaction effects between heat carrier temperature and auger
speed, and heat carrier temperature and feed rate, Figure 55 and Figure 56 were developed,
respectively. In Figure 55 the feed rate and nitrogen flow rate are kept constant at the center point
conditions (15 kg/hr and 2.5 sL/min, respectively), while the reduced model equation is used to plot
the bio-oil yield as a function of temperature and the 5 levels of constant auger speeds. Similarly, in
Figure 56, the nitrogen flow rate and auger speed are kept constant at the center point conditions (2.5
sL/min and 54 RPM, respectively), while the reduced model equation is used to plot the bio-oil yield
as a function of heat carrier temperature and the 5 levels of heat carrier feed rates.
As shown in Figure 55, the model shows a clear interaction between auger speed and heat
carrier temperature in relation to bio-oil yield. As the auger speeds and temperature are continuously
changing in the respo in Figure 54
(b). Nonetheless, the bio-oil yield prediction equation suggests that for heat carrier temperatures
below 550°C a low auger speed is preferred to achieve high bio-oil yields. This may be explained by
the increased mixing of biomass and heat carrier that was observed for low auger speeds during the
development of the reactor as described previously. However for temperatures above 550°C, higher
auger speeds are desired to increase the yield, which suggests that additional mixing time between
heat carrier material and biomass is not required and provides minimal benefit. The hot temperature
of the material at these conditions may adequately pyrolyze biomass quickly without the additional
solids residence time afforded by slow auger speeds. At the apparent “intersection point” shown, the
auger speed is of little importance in predicting the bio-oil yields. As the general response shows that
heat carrier temperatures above 550°C are desired for increasing bio-oil yield, the result from this
interaction effect implies that high auger speed are also desired to maximize liquid yield.
As shown in Figure 56, a higher heat carrier feed rate is preferred up to a temperature of near
500°C, however for higher temperatures the next two lowest feed rates shown are desired. As with the
auger speed and temperature interaction effect, this interaction is not clear in Figure 54 (b), but is
revealed by the negative coefficients for the feed rate higher order term and the temperature and feed
rate interaction term in Equation 24. The reason for this slight decrease in yield for high heat carrier
feed rates at high temperatures is unclear, and could perhaps be due to the increased volume of the
reactor occupied by heat carrier material at these conditions. Recall that the auger speed is constant
for all the yield curves shown in Figure 56, so for more heat carrier material in a fixed volume, the
bio-oil vapors have more surface area to interact with biochar. This interaction may decrease the bio-
oil yield by prom ribed by
Gronli & Antal [32] and Babu [
nse surface representation, this interaction effect is not readily seen
oting undesired reactions that convert hot vapors into secondary char as desc
29].

109
25
30
35
40
45
400 425 450 475 500 525 550 575 600 625 650
Modeled
bio-
50
55
60
65
70
75
80
oil
yield
(%-wt.,
wb)
45.0
49.5
54.0
58.5
63.0
Constant conditions:
Heat carrier feed rate = 15 kg/hr
N2 flow rate = 2.5 sL/min
Auger speed
(RPM)
High auger speeds desired to
increase bio-oil yield
Low auger speeds desired to
increase bio-oil yield
Figure 55. Modeled bio-oil yield as a function of heat carrier temperature and auger
speed
20
25
30
35
40
45
50
55
60
400 425 450 475 500 525 550 575 600 625 650
Modeled
bio-oil
yield
(%-wt.,
65
70
75
wb)
9
12
15
18
21
N2 flow rate = 2.5 sL/min, Auger speed = 54 RPM
Heat carrier
feed rate
(kg/hr)
Figure 56. Modeled bio-oil yield as a function of heat carrier temperature and feed rate

110
However for less reactor volume for the vapor products to occupy, the residence time is
decreased because the vapor velocity increases, which does not agree with this interpretation of the
interaction effect. Therefore, an alternate explanation may simply be that there can be “too high” of a
heat carrier feed rate where excess heat is available and may actually decrease bio-oil yield and favor
char or NCG production. This effect may be described as a high temperature cracking phenomenon.
In general and with all conditions considered simultaneously, the regression model for bio-oil
yield suggests that within the range of levels tested, the yield would be maximized at the highest
nitrogen flow rate (3.5 sL/min), the highest auger speed (63 RPM), the highest heat carrier
temperature (625°C) and a heat carrier feed rate of 18 kg/hr. Note that high auger speeds to promote
high bio-oil yield is in agreement with the twin-screw reactor (up to 300 RPM) reported by Raffelt et
al. [74], and the rotating cone reactor (600 RPM) as reported by Bridgwater [13]
Biochar yield. After reviewing the experimental residuals for the biochar yield (Equation 8)
as shown in Figure 138 of Appendix D, it was determined the same regression technique could be
applied. The statistical summary of the model analysis is shown in Table 31, and more detailed results
are saved for Table 58 in Appendix D. A high R2
value of 96.5% and low RMSE value (< 2%-wt., wb
biochar yi o1
and accept Ha1 to validate the significance of the model. The F-test for the lack of fit was used to
reject Ho2, and the t-test was used to reject Ho3 for 6 significant terms as shown below.
The same tests were used to analyze the reduced model, which was found to be significant
with no lack of fit, and the t-test was used again to reject Ho3 for all included terms. Finally, the model
utility F-test was used to verify that the reduced model is more significant than the full model. As
shown in Figure 57, the reduced biochar model also contained an interaction effect and two higher
order effects, further validating the experimental design selection. Recall that the parameter estimates
shown in Figure 57 are all significant (|t|i > t0.05,ν), however the magnitude of the test statistic shows
the relative significance of one parameter compared to another. It is clear that heat carrier temperature
and heat carrier feed rate are both influential terms, much more compared to the other terms in the
model.
The response surface equation for biochar yield is shown in Equation 25, noting that
coefficients greater than one increase the response value and coefficients less than one decrease the
value. Also recall that detailed parameter estimate information is included in Table 58 in Appendix D.
The predicted biochar yield versus the actual experimental data is shown in Figure 5
95% confiden
eld) suggest the full model fit the data well, which was confirmed by the F-test to reject H
8, with a
ce interval (thin band) and 95% prediction interval (thick band) shown as discussed.

111
Table 31. Biochar yield model, statistics summary
Full Reduced
R2
0.965 - 0.948 - -
F0.05,k,ν 2.424 - 2.528 - Reject Ho1
FLOF 1.32 X 1.21 X FLOF < F0.05,λ,m-1 *
F0.05,λ,m-1 4.74 - 4.58 - Reject Ho2
t0.05,ν 2.13 - 2.07 - -
|t| statistics
β0 25.00 √ 36.81 √ |t| > t0.05,ν Reject Ho3
β1 18.05 √ 18.47 √ |t| > t0.05,ν Reject Ho3
β2 0.05,ν Reject Ho3
2.20 √ 2.25 √ |t| > t
3 1.43 X - - |t| < t0.05,ν Don't reject Ho3
β
β4 6.87 √ 7.03 √ |t| > t0.05,ν Reject Ho3
β13 2.66 √ 2.72 √ |t| > t0.05,ν Reject Ho3
β11 2.84 √ 2.83 √ |t| > t0.05,ν Reject Ho3
β44 3.22 √ 3.23 √ |t| > t0.05,ν Reject Ho3
1.16
2.64
Note: * The null hypotheses Ho1 and Ho2 are rejected for the full model and the reduced model
2
HC
2
HC
A
HC
HC
N2
HC
biochar
μ
1.17
τ
1.03
Ω
τ
1.31
μ
2.77
θ
0.89
τ
7.29
20.55
wb)
wt.,
(%
Y
⋅
+
⋅
+
⋅
⋅
−
⋅
−
⋅
−
⋅
−
=
−
Equation 25
The biochar yield response surface plots are shown in Figure 59, where Figure 59 (a) holds
the heat carrier feed rate and auger speed constant at the center point conditions and Figure 59 (b)
holds the auger speed and nitrogen flow rate constant at the center point conditions as discussed
previously.

112
0 2.5 5 7.5 10 12.5 15 17.5 20
HC temperature
N2 flow rate
HC feed rate
HC temperature ·
Auger speed
HC temperature ·
HC temperature
HC feed rate · HC
feed rate
Model
term
t
0.05,
23
Interaction
effect
Main
effects
Higher order
effects
Figure 57. Absolute values for t-test statistics for biochar yield model
10
15
20
25
30
35
40
Actual
biochar
yield
(%-wt.,
wb)
10 15 20 25 30 35 40
Predicted biochar yield
(%-wt., wb)
Figure 58. Actual vs. predicted biochar yield

113
Note that the biochar yield as a function of auger speed is not displayed because this was
found to be an insignificant term by itself in the reduced model, even though it is one of the main
factors and present in other model terms. Temperature is shown to be more influential than nitrogen
flow rate in Figure 59 (a), but this plot shows that biochar yield is decreased for increasing nitrogen
flow rate for all other parameters held constant. This result is in accordance with the results from
Figure 54 (a) where bio-oil yield increases with increasing nitrogen flow rate. This is due to decreased
residence time associated with higher gas flow rates which favors bio-oil production.
1
.
5
2
.
0
2
.
5
3
.
0
3
.
5
4
2
5
4
7
5
5
2
5
5
7
6
2
5
5
10
15
20
25
30
35
40
Biochar
yield
flow rate
(SLPM)
Heat carrier temperature
(°C)
35-40
30-35
25-30
20-25
15-20
10-15
N2
volumetric
%-wt., wb
(a) Biochar yield as a function of heat carrier temperature and N 2
2
2 flow rate
9
1
2
1
5
1
8
2
1
4
2
5
4
7
5
5
2
5
5
7
5
6
2
5
10
15
20
25
30
35
40
45
Biochar
yield
Heat carrier
feed rate
(kg/hr)
Heat carrier temperature
(°C)
40-45
35-40
30-35
25-30
20-25
15-20
10-15
%-wt., wb
(b) Biochar yield as a function of heat carrier temperature and feed rate
Figure 59. Two response surfaces for modeled biochar yield

114
As shown in Figure 59 (b), the biochar yield is shown (in general) to decrease for increasing
heat carrier feed rate. This result is in accordance with the bio-oil yield that increased (in general) for
increasing heat carrier feed rate. It is theorized that this because more rapid heat transfer occurs for
higher carrier feed rates, which minimizes secondary char forming reactions as reported by Ball et al
[119] and Di Blasi [120]. However, as with the bio-oil yield, the biochar yield model has higher order
and interaction effects that are not readily apparent in the response surfaces. For instance, in Figure
59 (b), note that at a contour of constant temperature of 560°C, the biochar yield is less for 18 kg/hr
than for 21 kg/hr heat carrier feed rate (shown in the lower right hand corner). The biochar yield as a
function of heat carrier temperature and feed rate, and heat carrier temperature and auger speed are
shown in Figure 60 and Figure 61, respectively.
Note that the nitrogen flow rate and auger speed rate are kept constant at the center point
conditions in Figure 60, and the nitrogen flow rate and heat carrier feed rate are kept constant in
Figure 61 as shown. An interesting result is shown in Figure 60 – regardless of heat carrier
temperature, a heat carrier feed rate of 18 kg/hr results in slightly lower char yields than a feed rate of
21 kg/hr. This result seems to agree with the result from the bio-oil yield model in that (for higher
temperatures) the 18 kg/hr feed rate results in the highest liquid yields. This result also supports the
hypothesis that because of the higher feed rate of heat carrier material, there is less internal reactor
volume for the vapor products to occupy as they are produced, hence more chance to react with
biochar and decrease the liquid yield. However, in comparison to the standard deviation of the center
point test, this effect
The significant interaction effect between heat carrier temperature and auger speed is also
apparent in regards to the biochar yield response as shown in Figure 61. For a heat carrier temperature
near 525°C, the auger speed is seen to have little consequence on the product yield. At temperatures
below this point, the lowest possible auger speed results in the lowest biochar yield. This result is in
accordance with this interaction effect on the bio-oil yield response: low auger speeds may promote
mixing and more complete pyrolysis to occur. However, at temperatures above 525°C, faster auger
speeds are desired to decrease the biochar yield. This may be explained by the rapid pyrolysis of
biomass as it comes into contact with hot heat carrier material – too much contact time results in
increased solids yield (recall that solids residence time is directly related to auger speed). This is akin
to stating that longer solid residence times are desired at less than 525°C to minimize biochar yield,
and shorter solid residence times are desired at temperatures above 525°C to minimi biochar yield.
This interesting inter actor design:
depending on the opera modify.
is rather minor.
ze
action effect also speaks to the potential versatility of this re
ting conditions, the pyrolysis product distribution may be easy to

115
5
10
400 425 450 475 500 525 550 575 600 625 650
M
15
20
25
30
35
40
45
50
55
60
odeled
biochar
yield
(%-wt.,
wb)
9
12
15
18
21
Heat carrier
feed rate (kg/hr)
N2 flow rate = 2.5 sL/min, Auger speed = 54 RPM
Figure 60. Modeled biochar yield as a function of heat carrier temperature and feed
rate
45
50
55
wb)
0
5
10
15
20
25
30
35
40
400 425 450 475 500 525 550 575 600 625 650
Modeled
biochar
yield
(%-wt.,
45.0
49.5
54.0
58.5
63.0
Heat carrier feed rate = 15 kg/hr
Auger speed
(RPM)
decrease biochar yield
decrease biochar yield
Figure 61. Modeled biochar yield as a function of heat carrier temperature and aug
speed
er

116
In general, to minimize biochar yield in this reactor system, a heat carrier feed rate of 18
kg/hr is preferred, with high nitrogen flow rates (3.5 sL/min) and high auger speeds (63 RPM), which
also correlates to the conditions that favor maximizing bio-oil yield.
Note that the results from plotting the modeled biochar yield as a function of auger speed
(which is analogous to solids residence time) is in general agreement with a kinetic model of wood
fast pyrolysis as reported by Di Blasi [27]. This model (2002) showed that for constant temperature,
increased residence time resulted in increased solid char yields due to secondary reactions. It is
theorized that this is what is occurring in the right half of Figure 61. At high temperatures, low auger
speeds may promote secondary reactions that convert condensable vapors into biochar.
Non-condensable gas yield. The analysis of the residuals for the NCG yield model showed
that the assumptions required for performing a linear regression model could not be satisfied. As
shown in Figure 139 of Appendix D, a clear relationship was seen between the residuals, and this
relationship was not observed for the residual bio-oil and biochar experimental data. The relationship
shown is e
experiments were performed in. Recall as noted previously that in an effort to minimize experimental
error and ensure consistent heat carrier feed rates and heat carrier inlet temperatures as a function of
feed rate, the experiments were randomized within groupings according to heat carrier feed rates.
This is shown in Table 11 and Table 50 in Appendix D. Nonetheless, the regression modeling
procedure was performed for the NCG yield data for discussion purposes only and not for further
investigation. Refer to Table 59 and Table 60 in Appendix D for a summary and detail of the
statistical analyses, respectively. These tables show that even if the residuals were acceptable, the
non-condensable gas yield model would not be significant at a 95% confidence level.
Despite the inability to evaluate a regression model for the overall yield of non-condensable
gases, the yield of individual species was investigated. As discussed, the non-condensable gas
mixture was analyzed with a Micro-GC, with gas concentration data (including nitrogen) as shown in
Table 61 of Appendix D. Recall that these are the averaged values taken over the steady state region
of an experiment (as shown in Figure 45), typically around 15 sample points. The mole fraction of
each gas species on a nitrogen free basis is shown in Table 62 of Appendix D, calculated as
previously discussed. The average gas composition for the center point runs is shown in Figure 62.
time based and is directly related to the grouping of heat carrier feed rates in which th

117
2.45
41.44
50.73
H2
CO
CH4
0.42
0.60
4.36
C2H6
C2H4
CO2
Values in %-vol.,
nitrogen free basis
Figure 62. Average non-condensable gas composition at center points
Also note in Table 62 of Appendix D that the estimated mass of NCG is provided, which
allows for calculating the number of total moles of gas. Determining the apparent molecular weight
and using the ideal gas law to calculate the mass was discussed previously. The total number of moles
is used to convert the mole fraction of each gas species into the number of moles of each species,
which is finally used to calculate the mass of each gas species as shown in Table 63 of Appendix D
based on individual molecular weights. The gas property data found in Table 64 of Appendix D
(pressure, temperature and total volume) is also required for this analysis. With the mass of a gas
species known, the mass yield based on the biomass input can be determined as is performed for bio-
oil and biochar. The mass yields of carbon monoxide and carbon dioxide are of the most interest, but
the yield of any species, i, on a percent weight of the original wet biomass is calculated as shown in
Equation 26 with standard notation and as already discussed.
b
i
i
NCG
NCG
b
i
i
NCG
b
i
i
NCG
m
M
y
M
m
m
M
y
n
m
M
n
wb)
wt.,
(%
Y
⋅
⋅








=
⋅
⋅
=
⋅
=
− Equation 26
The regression models for gas yields were chosen to be performed on a mass basis rather than
on a volume basis because it results in a more interesting comparison with bio-oil and biochar yields

118
which are both on a mass basis. For instance, the carbon monoxide and carbon dioxide yields
averaged 3.77%-wt. and 7.24%-wt., respectively for the center point tests. As the center point average
gas yield was 11.35%-wt., CO and CO2 accounted for over 97% of the gas on a mass basis. The
carbon dioxide and carbon monoxide yields are shown for all 30 tests in Figure 63 as a function of
bio-oil yield. This result shows some type of relationship between gas yield and bio-oil yield, and
prompted further study of the gas yields of independent species.
4.5
6.5
7.5
8.5
s
yield
(%-wt.,
wb)
CO
CO2
5.5
2.5
3.5
40 45 50 55 60 65 70 75
Bio-oil yield (%-wt., wb)
Ga
Figure 63. Carbon monoxide and carbon dioxide yields vs. bio-oil yield for all tests
Carbon monoxide yield. The same regression modeling procedure for CO yield was
performed after the residuals, shown in Figure 140 of Appendix D, were analyzed and deemed
adequate. The full and reduced model were both found to be significant with no lack of fit (Ho1 and
Ho2 rejected for both models) as shown in Table 32. The details of the analysis, including the
ANOVA and Lack of Fit data, are shown in Table 65 of Appendix D. The reduced model was found
to be more significant than the full model, and includes 7 significant terms as shown in Equation 27.
2
HC
HC
A
A
N2
A
HC
HC
A
HC
CO
μ
0.07
-
μ
Ω
0.05
Ω
θ
0.07
Ω
τ
0.08
⋅
⋅
⋅
+
⋅
⋅
+
⋅
⋅
+ Equation 27
μ
0.21
Ω
0.05
τ
0.50
3.75
wb)
wt.,
(%
Y ⋅
+
⋅
−
⋅
+
=
−

119
Table 32. Carbon monoxide yield model, statistics summary
R
2
0.985 - 0.980 - -
F0.05,k,ν 2.424 - 2.464 - Reject Ho1
FLOF 1.51 X 1.27 X
F 4.74 - 4.59 -
FLOF < F0.05,λ,m-1 *
0.05,λ,m-1 Reject Ho2
t0.05,ν 2.13 - 2.07 - -
|t| statistics
β0 105.83 √ 192.05 √ |t| > t0.05,ν Reject Ho3
1 28.32 √ 29.83 √ |t| > t0.05,ν Reject Ho3
β
β
3 2.91 √ 3.07 √ |t| > t0.05,ν Reject Ho3
β
4 11.60 √ 12.22 √ |t| >
β t0.05,ν Reject Ho3
β
13 3.69 √ 3.88 √ |t| > t0.05,ν Reject Ho3
β
23 3.15 √ 3.32 √ |t| > t0.05,ν Reject Ho3
β
β
β
34 2.15 √ 2.26 √ |t| >
β
β
β
44 4.13 √ 4.30 √ |t| >
Note: * The null hypotheses Ho1 and Ho2 are rejected the full model and the reduced model
0.80
2.71
Full Reduced
As before, the predicted vs. actual carbon monoxide values are shown in Figure 64 with the
95% co
for high uger speeds.
nfidence and prediction intervals. With a high R2
and low RMSE, the model fit the data well.
The model for carbon monoxide predicts a yield behavior that is similar to that for bio-oil, which is to
be expected based on Figure 63. The CO yield increases with temperature and heat carrier feed rate,
and the interaction effect between auger speed and heat carrier temperature is significant. In other
words at low temperatures where low bio-oil yields are favored, CO yields are maximized for low
auger speeds. However at higher temperatures favoring high liquid yields, the CO yield is maximized
a

120
2.5
2.75
3
3.25
3.5
3.75
4
4.25
4.5
4.75
5
Actual
CO
yield
(%-wt.,
wb)
2.5 2.75 3 3.25 3.5 3.75 4 4.25 4.5 4.75 5
Predicted CO yield
(%-wt., wb)
Figure 64. Actual vs. predicted carbon monoxide yield
Carbon dioxide yield. As CO2 is also a major constituent in the NCG mixture, a regression
model was developed to analyze the CO2 yield as a function of all test conditions. The residuals for
the experimental data compared to the full model are shown in Figure 141 of Appendix D, and are
appropriate for performing a linear regression. The full model was found to have a have an R value
f 90% and a high FANOVA value (indicating significance) as shown in Table 33, however the lack of
fit F-tes
t was not constructed and the model was not further analyzed.
2
o
t indicates a marginally significant lack of fit. This implies that the null hypothesis Ho2 can not
be rejected, and that the lack of fit is considered significant. The t-tests were performed as before to
remove insignificant terms, however the reduced model (though significant as a whole) also exhibited
a significant lack of fit. The detailed statistical analysis of the CO2 yield model is shown in Table 66
of Appendix D. The lack of fit implies that the linear regression model as fitted is not sufficient,
despite including interaction and squared terms. Though a higher order (cubic) model may reduce the
lack of fit, this was not attempted. As this form of the model is considered inadequate, the predicted
vs. actual CO2 yield plo

121
Table 33. Carbon dioxide yield model, statistics summary
R2
0.902 - 0.833 - -
F0.05,k,ν 2.424 - 2.464 - Reject Ho1
FLOF 5.04 √ 5.28 √ FLOF > F0.05,λ,m-1
F0.05,λ,m-1 4.74 - 4.59 - Don't reject Ho2
t0.05,ν 2.13 - 2.07 - -
|t| statistics
β0 121.36 √ 157.64 √ |t| > t0.05,ν Reject Ho3
β1 6.13 √ 5.67 √ |t| > t0.05,ν Reject Ho3
β3 4.93 √ 4.56 √ |t| > t0.05,ν Reject Ho3
β4 5.57 √ 5.16 √ |t| > t0.05,ν Reject Ho3
β13 2.80 √ 2.60 √ |t| > t0.05,ν Reject Ho3
β23 2.66 √ 2.46 √ |t| > t0.05,ν Reject Ho3
β11 2.63 √ 2.24 √ |t| > t0.05,ν Reject Ho3
β44 3.14 √ 3.19 √ |t| > t0.05,ν Reject Ho3
Full Reduced
1.78
2.71
Note: * The null hypothesis Ho1 is rejected the full model and the reduced model
Gas yield of other species. To further investigate the relationship between bio-oil and non-
condensable gas production, the calculated yields of other species were also plotted as a function of
bio-oil yield. As shown in Figure 65, the yields of CH4, C2H6, C2H4 and H2 are all shown to increase
with bio-oil yield. For C2H6, C2H4 and H2 this occurs slowly and then more rapidly past
approximately 70%-wt. bio-oil yield. The yield of gaseous methane is not linear and resembles the
trend for carbon monoxide as shown in Figure 63. These trends result in the total non-condensable
gas yield trend as shown in Figure 66: gas yields tend to increase slightly as bio-oil yields increase. A
simple linear fit is shown to illustrate this correlation. This phenomenon will be discussed shortly
after discussions of the physical properties and chemical composition of the bio-oil. Note that based
on the trends shown in Figure 65, regression models were not developed for the four gas species
shown. It is expected that these models would reveal that the conditions that favor high bio-oil yields
would also favor higher gas yields of each species.

122
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
40 45 50 55 60 65 70 75
Gas
yield
(%-wt.,
wb)
CH4
C2H6
C2H4
H2
Figure 65. Gas yields for 4 different species vs. bio-oil yield for all tests
y = 0.119x + 3.522
R
2
= 0.828
8
9
10
11
12
45 50 55 60 65 70 75
Non-condensable
gas
yield
(%
13
14
-wt.,
wb)
Note: Does not include Run #13 (gas yield measure by difference)
Figure 66. Total non-condensable gas yield vs. bio-oil yield for 29 tests

123
5.3 Product analysis results
The biochar and bio-oil fractions collected were subjected to various chemical and physical
tests as described. The appearance of the bio-oil fractions were all quite different from each other, but
were similar from one experiment to the next. A characteristic picture of the bio-oil fractions is shown
in Figure 67, with SF1 – SF4 pictured from left to right.
Figure 67. Typical appearance of bio-oil fractions
Visually, SF1 and SF3 were relatively viscous oils, while SF 2 and SF 4 appeared much less
viscous. Stage fraction 4 was typically translucent, with an orange tint. SF2 and SF4 have a much
stronger acidic odor than SF1 or SF3. SF3, the fraction from the ESP, was the most “syrup-like”
fraction.
Moisture content. Karl-Fischer (KF) moisture content analyses were performed on all four
fractions for all experiments in triplicate. Average values and standard deviations are shown in Table
67 of Appendix D. Moisture values were found to vary by stage fraction as shown in Figure 68 for the
six center point tests. Standard deviations are shown among the triplicate moisture analyses. The
whole bio-oil moisture content was calculated as discussed (Equation 22). The average moisture
content for the center point tests is shown in Figure 69 along with the moisture contents from the
minimum bio-oil yield and maximum bio-oil yield tests. The whole bio-oil moisture content is shown
to be in agreemen 8].
SF1 SF2 SF3 SF4
t with recently published data from two other auger reactors as shown [86, 8

124
0
10
20
SF1
80
SF2
30
40
50
60
e
content
(%-w
70
12 15 17 19 21 22
Center Point tests
Moistur
t.)
SF3
SF4
Whole
Moisture content range
(15 - 30%-wt.)
As reported by Czernik & Bridgwater for wood pyrolyis
bio-oil. Source: Energy & Fuels 2004, 18 , 590-598
Figure 68. Bio-oil moisture content at center points
0
5
(Run 13)
(Run 20)
10
15
20
55
60
65
70
75
80
85
Mois
t
(%-wt.
SF1
)
SF2
SF3
SF4
25
30
35
40
45
50
ture
conten
Whole
Reported by Ingram et al. for whole bio-oil from 1.0 kg/hr auger
reactor. Source: Energy & Fuels 2008, 22 , 614-625
Oak wood (22.5%-wt.)
Oak bark (22.0%-wt.)
Reported by Garcia-Perez et al. for whole bio-oil from 1.5 kg/hr auger
Pine wood (23.2%-wt.)
Figure 69. Bio-oil moisture content range

125
Note in Figure 69 the standard deviations for the center point averages are presented as
pooled standard deviations, which takes into account the variance among standard deviations
between the six tests. The pooled standard deviation is larger than the average of the standard
deviations. In effect, this captures the deviations attributed to the analytical procedures, the in-
homogeneity of the bio-oil sample, and the variation among experimental testing. See Equations D6
and D7 in Appendix D for a sample calculation of pooled standard deviation for the whole bio-oil.
Also note in Figure 69 that the maximum yield bio-oil (73.6%-wt., wb) correlates to the
lowest whole bio-oil moisture content (22 ± 2.4%-wt., wb), and the minimum yield sample (38.7%-
wt., wb) has the highest moisture content (35 ± 1.3%-wt., wb). So this image effectively shows the
range of moisture contents for all samples. It is interesting to note that the moisture content for SF1
and SF2 are seen to vary drastically among the three situations presented, whereas the water contents
in SF3 and SF4 do not vary as much. As SF1 and SF2 typically represent over 75% of the total bio-oil
mass collected, the whole bio-oil moisture content is seen to closely follow the trends for SF1 and
SF2. The moisture conten al whole bio-oils,
hereas the typical moisture contents for SF2 and SF4 are typically more than traditional oils.
To extend the concept presented in Figure 69, the moisture content (whole bio-oil) for each
experiment was plotted as a function of bio-oil yield with interesting results as shown in Figure 70.
This result suggests that as bio-oil yield increases, secondary reactions that may increase water
content are minimized. A simple linear regression fit is shown to indicate this correlation.
ts for SF1 and SF3 are typically less than tradition
w
y = -0.37x + 50.27
R
2
= 0.88
17.5
20.0
22.5
25.0
27.5
30.0
32.5
35.0
37.5
40 45 50 55 60 65 70 75
Bio-oil
moisture
content
(%-wt.)
Figure 70. Bio-oil moisture content vs. bio-oil yield for all tests

126
The results from Figure 70 are shown to be in agreement with those shown in Figure 63,
Figure 65 and Figure 66. The explanation for the decrease in moisture content as the bio-oil yield
increases may be attributed to the increase in yields of the gas species. For high bio-oil yield
situations, the hydrogen in the biomass appears to be converted into higher levels of gaseous H2, CH4,
C2H4 and C2H6 rather than liquid H2O. Similarly, the oxygen from the biomass appears to be
converted into gaseous CO and CO2 rather than liquid H2O for high bio-oil yield situations.
As the bio-oil yield is a function of multiple factors, the relationship between bio-oil yield
and moisture content provided evidence that a regression model was necessary for further
investigation. The modeling procedure was performed as before with residuals shown in Figure 142
of Appendix D, results summarized as shown in Table 34 and detailed results in Table 68 of
Appendix D. The model was significant (Ho1 rejected) with no significant lack of fit (Ho2 rejected),
and the reduced model was found to be more significant (don’t reject Ho4) and only includes four
significant parameters as shown in Equation 28 (Ho3 rejected for these terms). Only two of the
original four factors are significant (heat carrier temperature and feed rate), as well as one interaction
term (heat carrier temperature and auger speed) and one higher order term (temperature squared).
2
HC
A
HC
HC
HC
τ
0.696
Ω
τ
0.684
μ
0.535
τ
2.96
25.67
wt.)
(%
content
Moisture
⋅
+
⋅
⋅
−
⋅
−
⋅
−
=
−
Equation 28
As shown in Figure 71, heat carrier temperature is the most influential term in the moisture
content model. The predicted moisture content values (whole bio-oil) versus the actual experimental
values are shown in Figure 72 with the 95% confidence and prediction intervals. The response surface
for the moisture content is shown in Figure 73, where the auger speed and nitrogen flow rate are kept
constant at the center point conditions. This surface shows that as feed rate and temperature are
increased, moisture content is decreased (for a constant auger speed).
However, as before, this response surface does not reveal the interaction effect between auger
speed and heat carrier temperature. As shown in Figure 74, there is a distinct heat carrier temperature
value where the auger speed has little influence on the moisture content. This phenomenon was also
seen in the models for bio-oil yield and biochar yield. Below heat carrier temperatures of 525°C, low
auger speeds are desired to minimize bio-oil moisture content, whereas above 525°C higher auger
speeds are desired. Based on Figure 55 and Figure 70, this result is not unexpected due to the
established relationship between bio-oil yield and bio-oil moisture content.

127
Table 34. Bio-oil moisture content model, statistics summary
Reduced
Full
R2
0.942 - 0.907 - -
F0.05,k,ν 2.424 - 2.759 - Reject Ho1
FLOF 0.84 X 0.48 X FLOF < F0.05,λ,m-1 *
F0.05,λ,m-1 4.74 - 2.54 - Reject Ho2
t0.05,ν 2.13 - 2.06 - -
|t| statistics
β0 62.64 √ 110.35 √ |t| > t0.05,ν Reject Ho3
β1 14.39 √ 14.67 √ |t| > t0.05,ν Reject Ho3
β4 2.61 √ 2.66 √ |t| > t0.05,ν Reject Ho3
β13 2.72 √ 2.77 √ |t| > t0.05,ν Reject Ho3
11 3.58 √ 3.78 √ |t| > t0.05,ν Reject Ho3
β
β
β
β
1.51
2.54
0 2.5 5 7.5 10 12.5 15
HC temperature
HC feed rate
term
HC temperature ·
Auger speed
HC temperature · HC
temperature
Model
Interaction effect
t
0.05,
25
Main
effects
e content model
Higher order
effect
Figure 71. Absolute values for t-test statistics for moistur

128
20
22.5
25
27.5
30
32.5
35
37.5
Actual
KF
moisture
content
(%-wt.,
wb)
20 22.5 25 27.5 30 32.5 35
Predicted KF moisture content
(%-wt., wb)
Figure 72. Actual vs. predicted moisture content
These results suggest that the conditions that favor high bio-oil yield and low biochar yield
also favor low moisture content in the produced whole bio-oil. These conditions, in regards to the
moisture content, include high auger speeds and high heat carrier feed rates to quickly transfer heat.
9
12
15
18
21
4
2
5
4
7
5
5
2
5
5
7
5
6
2
5
20
23
26
29
32
35
38
KF
moisture
content
Heat carrier
feed rate
(kg/hr)
Heat carrier temperature (°C)
35-38
32-35
29-32
26-29
23-26
20-23
(%-wt., wb)
Figure 73. Response surface for modeled moisture content

129
17.5
20.0
22.5
25.0
27.5
30.0
32.5
35.0
37.5
40.0
42.5
400 425 450 475 500 525 550 575 600 625 650
Modeled
bio-oil
H
2
O
content
(%-wt.) 45.0
49.5
54.0
58.5
63.0
N2 flow rate = 2.5 sL/min Heat carrier feed rate = 15 kg/hr
Auger speed
(RPM)
decrease moisture content
decrease moisture content
Figure 74. Modeled moisture content as a function of heat carrier temperature and
auger speed
Water insoluble content. The water insoluble content was determined for stage fractions 1, 2
and 3 for each experiment. The water insoluble content was not performed for the SF4 fractions due
to the low mass collected to help ensure there was adequate sample to test moisture content and to
perform the ultimate and proximate analyses and the GC/MS characterization. Furthermore, as the
SF4 sample is highly aqueous, it is likely to contribute a negligible amount of water insoluble
material to the whole bio-oil as it represents such as small portion of the total bio-oil mass. This
assumption of minimal insoluble content is also based on the physical design of the reactor system
and by visual inspection of the SF4 oil. Finally, this provides a conservative estimate for the water
insoluble content, as testing the SF4 sample would only increase the total, albeit only slightly if at all.
Refer to Table 69 of Appendix D for analytical data collected for water insoluble content.
As shown in Figure 75, the water insoluble content varied among fractions SF1, SF2 and
SF3, but was fairly con has the highest
Note that the relationship between bio-oil yield, non-condensable gas yield and bio-oil
moisture content will again be discussed after review of the elemental analysis of the bio-oil.
sistent among each center point experiment. The SF3 fraction

130
average water insoluble content (26.6%-wt., wb), followed by SF1 (17.3%-wt., wb), and SF2 had the
lowest water insoluble content (7.6%-wt., wb). Recall the water content for SF2 was significantly
higher than in SF1 or SF3. As shown, the whole bio-oil has a water insoluble content (15.6%-wt., wb)
within the range for bio-oil as reported by Bridgwater [13], but it is on the low end of the range.
0
5
O
in
10
15
20
25
30
35
H
2
soluble
content
(%-wt.,
wb)
SF1
SF2
SF3
Whole
12 15 17 19 21 22
Center Point tests
Source: A.V. Brigdwater, et al., The status of biomass
fast pyrolysis. In Fast pyrolysis of biomass: A
handbook, CPL Press: Newbury, UK, 2002; Vol. 2.
Insoluble pyrolytic lignin
bio-oil. For
each fra
Typical range: 15 - 30 %-wt.,wb
Figure 75. Water insoluble content for center points
The average water insoluble content for the center point tests are compared to the results from
the maximum and minimum bio-oil yield tests in Figure 76. The standard deviations from triplicate
analyses are shown, and the center point averages are shown with pooled standard deviations among
the six runs as discussed previously.
This figure is of interest because is reveals that there is a relationship between the reaction
conditions that favor high bio-oil yield and the amount of water insoluble content in the
ction and the resulting whole bio-oil, the amount of water insoluble material increases with

131
liquid yield. This phenomenon provides sufficient evidence that a model for water insoluble content is
necessary to investigate the relationship.
22.5
15.6
9.6
0
5
10
15
20
25
30
35
40
(Run 13)
(Run 20)
H
2
O
insoluble
content
(%-wt.,
wb)
SF1
SF2
SF3
Whole
Source: A.V. Brigdwater, et al., The status of biomass
fast pyrolysis. In Fast pyrolysis of biomass: A
handbook,CPL Press: Newbury, UK, 2002; Vol. 2.
Insoluble pyrolytic lignin
Typical range: 15 - 30 %-wt.,wb
Figure 76. Water insoluble content range
A regression mod l water insoluble
content as discussed previously, and the resulting residuals are shown in Figure 143 of Appendix D.
Visual
nificant lack of fit was found (Ho2 rejected). In addition to the
intercep
model (don’t reject HO4). The resulting regression model is described by Equation 29.
eling procedure was performed for the whole bio-oi
analysis of the residuals indicated that a linear regression model could be developed. The
statistical results for the water insoluble content model are shown in Table 35, and more detailed
results are saved for Table 70 of Appendix D. Both the full and reduced model were found to be
significant (Ho1 rejected), and no sig
t, only two significant terms were found to affect the water insoluble response: heat carrier
temperature and feed rate. The t-test was used to reject Ho3 for these terms as shown in Table 35.
Finally, the model utility test also confirmed that the reduced model is more significant than the full

132
Table 35. Water insoluble content model, statistics summary
Fu
R2
0.951 - 0.912 - -
F0.05,k,ν 2.424 - 3.354 - Reject Ho1
FLOF 2.43 X 1.85 X FLOF < F0.05,λ,m-1 *
F0.05,λ,m-1 4.74 - 2.59 - Reject Ho2
t0.05,ν 2.13 - 2.57 - -
|t| statistics
β0 49.19 √ 2.0 √ |t| > t0.05,ν Reject Ho3
β1 16.49 √ 2.50 √ |t| > t0.05,ν Reject Ho3
ll Reduced
1.97
2.48
Equation 29
ues are shown
HC
HC μ
0.374
τ
2.61
16.15
wb)
wt.,
(%
content
insoluble ⋅
+
⋅
+
=
−
Water
The water insoluble content model is quite simple, and predicts that the water insoluble
material will increase with both temperature and heat carrier feed rate; statistically independent of all
the other operating conditions. The modeled water insoluble content as a function of heat carrier
temperature and feed rate is shown in Figure 77. As the bio-oil yield tends to increase with
temperature and heat carrier feed rate as well, a relationship exists between water insoluble content in
the bio-oil and the yield of bio-oil as shown in Figure 78. This suggests that conditions that favor high
bio-oil yields may decompose lignin into water insoluble compounds in the bio-oil rather than
conversion of lignin to biochar. One such condition may be higher heat carrier temperatures, which
are required to decompose lignin [4, 23]. The predicted water insoluble content val

133
plotted with the actu nfidence and
prediction intervals.
al experimental values in Figure 79, along with the 95% co
15.3
15.5
15.7
15.9
16.1
16.3
16.5
16.7
16.9
17.1
400 425 450 475 500 525 550 575 600 625 650
Modeled
H
2
O
insoluble
content
(%-wt.,
wb)
9
12
15
18
21
Auger speed = 54 RPM
Heat carrier feed rate
(kg/hr)
Figure 77. Modeled H2O insoluble content as a function of heat carrier temperature and
feed rate
0.0
2.5
5.0
7.5
10.0
12.5
40 45 50 55 60 65 70 75
Water
insoluble
cont
15.0
17.5
20.0
22.5
25.0
ent
(%-wt.,
wb)
Figure 78. Water insoluble content vs. bio-oil yield for all tests

134
7.5
10
12.5
15
17.5
20
22.5
25
Actual
water
insoluble
content
(%-wt.,
wb)
7.5 10 12.5 15 17.5 20 22.5 25
Predicted water insoluble content (%-wt., wb)
Figure 79. Actual vs. predicted water insoluble content
Solids content. The solids content analysis was performed in triplicate for the center point
tests on
were not performed for SF4 to preserve the mass for other kinds of analysis. Since the percent of total
bio-oil mass collected in SF4 was less than 2%-wt., it contributes an insignificant amount to the
overall solids content. Also, based on visual inspection and the physical design of the system, it is
likely that the solids content in SF4 is negligible. The analytical data for the solids content is shown
in Table 71 of Appendix D.
Recall that in general, the amount of solid material suspended in the bio-oil is a reflection of
the biochar separation efficiency. In this sense, the solids content will be particularly dependent on
the size of the biomass particles. In this study, the biomass particle size was kept constant for all test,
and all the biomass was prepared in the same manner with the same equipment. Therefore, it is not
expected that the solids content will vary as a function of the test parameters, and a regression model
of this data would not be of much interest.
The solids content was not found to vary significantly between fractions SF1 – SF3, and the
average value in each fraction varied from 0.7%-wt., wb to 1.07%-wt., wb which is within the range
of commonly reported values for bio-oil. The overall average for the whole bio-oil was 0.94 ± 0.22
SF1, SF2 and SF3 to determine the magnitude of value for the samples. Solids content tests

135
%-wt., wb for the center point tests shown in Figure 80, which is in agreement for the typical range of
wood pyrolysis bio-oil as reported by Czernik & Bridgwater [36]. Note that these values are also
within a general range that agrees with recently published literature on bio-oil produced from wood
biomass in a 1 kg/hr auger reactor as shown and discussed previously [86].
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0 SF1
SF2
)
Solids
content
(%-wt.,
wb
SF3
Whole
12 15 17 19 21 22
Center point test
Reported by Ingram et al. for whole bio-oil from 1.0 kg/hr
auger reactor. Source: Energy & Fuels 2008, 22 , 614-625
Oak wood (0.80%-wt., wb)
Oak bark (1.83%-wt., wb)
Wood pyrolysis bio-oil range (0.2 - 1.0 %-wt., wb)
As reported by Czrenik et al. Source:
Energy & Fuels 2004, 18 , 590-598
Figure 80. Solids content for center point tests
heating values for all of the bio-oil samples, largely based
on the w
Higher heating value. The higher heating value was investigated for three of the center point
tests, the minimum bio-oil yield sample (Run 13), and the maximum bio-oil yield sample (Run 20).
Based on the low pooled standard deviation among the 3 center point runs, it is likely that the other 3
center point runs would have similar higher heating values. This range of samples is believed to give
a representation of the magnitude of higher
ater content as previously discussed. The higher heating value collected data is shown in
Table 72 of Appendix D, which is presented graphically in Figure 81. When comparing this image to
Figure 69, it is evident there is an inverse relationship between bio-oil moisture content and heating

136
value. The average higher heating value for the whole bio-oil for three of the center point runs was
16.4 ± 0.16 MJ/kg, which although is on the low end is within the range of typical pyrolysis oils [36].
The higher heating value for SF3 is seen to be within the range of two other recent studies as shown
[86, 88]. Recall the averaged center point heating values are presented with a pooled standard
deviation, which takes into account the standard deviations among the different runs.
0
2
4
6
8
10
12
14
16
18
20
22
(Run 13)
3 run center point average
(Runs 12, 17, 21)
(Run 20)
Higher
heating
value
(MJ/kg)
SF1
SF2
SF3
SF4
Whole
Reported by Ingramet al. for whole bio-oil from1.0 kg/hr auger
Pine wood (21.5 MJ/kg)
Reported by Garcia-Perez et al. for polar bio-oil from1.5 kg/hr auger
Pine wood (19.5 MJ/kg)
Low range for typical properties as reported by Czernik &
Bridgwater. Source: Energy & Fuels 2004
is bio-oil (16MJ/kg) , 18 , 590-598
Wood pyrolys
Figure 81. Higher heating value range
Note that the highest yielding bio-oil sample has the highest energy value (whole bio-oil)
shown in Figure 81, which may be attributed to this sample having the lowest moisture content.
Similarly, the lowest bio-oil yield sample has the lowest energy value among tested samples, which is
attributed to it having the highest moisture content. This is a common and documented relationship
between bio-oil heating value and moisture content. The SF3 energy content varies the least.

137
Thermal gravimetric analysis. TGA tests were performed for all biochar samples and for
many of the bio-oil samples to develop the proximate analysis.
This analytical procedure is more useful for evaluating biochar than pyrolysis liquids. This is
because, as discussed, the “moisture” and “volatiles” determined by this method are not directly
applicable to bio-oil because of the many compounds that volatilize over a wide range of
temperatures. For instance, as shown in Table 73 of Appendix D, the combined moisture and volatiles
(as determined by TGA) of each bio-oil fraction exceeds 85 %-wt. The TGA analysis for bio-oil will
be discussed again as part of the elemental analysis.
The TGA data for biochar samples originating from the cyclone is shown in Table 74 of
Appendix D. The data was plotted to determine if any visible trends warranted further study by
developing a regression model, however no trends were observable. Therefore, no biochar regression
models were developed for the proximate analysis results. As shown in Figure 82 for the center point
tests, however, the proximate analysis results are in general agreement with recently published data
on biochar originating from oak bark processed in an auger reactor [42].
0
10
20
30
40
50
60
70
80
90
Red oak
Moisture
Volatiles
Fixed carbon
Ash
12 15 17 19 21 22 Mohan et
%-wt.
biomass al.*
Center Point test
* As reported by Mohan et al. for oak bark biochar from a 1.0 kg/hr auger reactor.
Source: J. Colloid and Interface Science 2007, 310 , 57-73.
Figure 82. Biochar proximate analysis for center point tests

138
Compared to this study, the proximate analysis results from the analyzed biochar show
slightly lower fixed carbon content, higher volatiles content, as well as lower ash content. The center
point av
for biochar at center point tests
erage for ash was 5.5%-wt., however it varied from 3.3 – 12 %-wt. among all samples. Also
shown in Figure 82 is the average red oak biomass proximate analysis as presented previously. Note
that compared to the biomass, fixed carbon and ash contents are concentrated in the biochar, while the
volatile matter is markedly lower due to much of this mass being converted into liquid bio-oil.
Table 36. Ultimate analysis
Averagea
St. Dev.b
Oak wood Oak bark
Moisture 4.30 0.557 3.17 1.56
Carbon 70.85 1.919 82.83 71.25
Nitrogen 0.11 0.046 0.31 0.46
Hydrogen 3.64 0.218 2.70 2.63
Sulfur 0.012 0.005 0.02
Ash 5.51 0.625 2.92
c
This study
0.02
11.09
Oxygend
19.88 1.919 11.22 14.55
Notes: All values in %-wt. a - Average of center point tests.
b - Standard deviation among runs (not replicates).
c - As reported by Mohan et al. for a 1 kg/hr auger reactor.
Source: J. Colloid & Interface Science 2007, 310 , 57-73.
d - Oxygen calculated by difference
Mohan et al.
Elemental analysis. The elemental composition of all bio-oil and biochar samples was
determined by analyzing the carbon, nitrogen, hydrogen, and sulfur contents. Assuming these are the
major constituents present, in combination with the ash content as determined by the TGA methods,
the oxygen content is determined by difference.
The analytical data for the elemental analysis of the biochar is shown in Table 75 of
Appendix D, noting triplicate analyses were performed for the center point tests. It was found that the
elemental analysis of the biochar did not vary significantly for different operating conditions, which
implies regression modeling would be of little interest. A summary of the elemental analysis for the
biochar is shown in Table 36 for the center point tests, noting the comparison to another study. Also,
note that almost 82% of the carbon content as determined by the ultimate analysis (70.9 %-wt.)
remains as fixed carbon during the proximate analysis (58.0 %-wt.).
The data for th ollows: Table
76 for SF1, Table 77 for SF2, Table 78 for SF3, Table 79 for SF4 and Table 80 for the resulting
whole bio-oil as calculated. The carbon content for each of the fractions and the whole bio-oil is
e elemental composition of bio-oil is found in Appendix D as f

139
shown in Figure 83 for the center point tests with standard deviations shown from triplicate analyses.
The fractions that are high in water content (SF2 and especially SF4) are shown to have carbon
contents less than pyrolysis liquids as reported by Oasmaa & Meier [38]. However the remaining
fractions and the whole bio-oil have carbon contents within the expected range. Also shown in the
figure below is the repeatability among center points, and the small instrument error.
0
5
10
15
20
25
30
35
40
45
50
12 15 17 19 21 22
Center Point tests
Carbon
content
(%-wt.,
wb)
SF1
SF2
SF3
SF4
Whole
Source: A. Oasmaa et al., Analysis, characterisation and test
methods of fast pyrolysis liquids. In Fast pyrolysis of biomass:
A handbook ,CPL Press: Newbury, UK, 2002; Vol. 2.
Carbon content range
(Wood pyrolysis liquids):
tions and the whole bio-oil is shown in Figure 84 for
the cen
and shown in Figure 84.
32 - 49 %-wt., wb
Figure 83. Bio-oil carbon content for center points
The nitrogen content for each of the frac
ter point tests with standard deviations shown for triplicate analyses. Clearly nitrogen is a
more difficult element to analyze because of its low levels in the bio-oil samples. In fact often the
nitrogen content was below the detection level of the instrument (80 PPM), and for these cases it was
then assumed that the nitrogen content in the sample was 80 PPM. These cases can be clearly
identified in Figure 84. Out of the 120 samples (tested in triplicate), 61.7% had nitrogen contents that
were below the detection limit. The nitrogen values that were above the detection limit, however, are
shown to be less than the upper limit of 0.30% as reported by Oasmaa

140
0.000
0.025
0.050
0.075
0.100
0.125
0.150
0.175
0.200
0.225
0.250
0.275
0.300
ent
(%-wt.,
wb
12 15 17 19 21 22
Center Point tests
Nitrogen
cont
)
SF1
SF2
SF3
SF4
Whole
Detection limit
80 PPM
Nitrogen content range
0.0 - 0.30 %-wt., wb
Figure 84. Bio-oil nitrogen content for center points
The hydrogen content for each of the fractions and the whole bio-oil is shown in Figure 85
for the center point tests with standard deviations shown for triplicate analyses.
0
1
2
3
4
5
6
7
12 15 17 19 21 22
Center Point tests
Hydrogen
content
8
9
10
11
(%-wt.,
wb
SF1
)
SF2
SF3
SF4
Whole
Hydrogen content range
6.9 - 8.6 %-wt., wb
Figure 85. Bio-oil hydrogen content for center points

141
The hydrogen content for each of the first three fractions (SF1, SF2, SF3) and the whole bio-
oil was found to be within the range of hydrogen for wood pyrolysis liquids reported by Oasmaa &
Meier [38]. However the last fraction, SF4, had particularly high hydrogen content which may be
attributed to the high water content.
The sulfur content for each of the fractions and the whole bio-oil is shown in Figure 86 for
the center point tests with standard deviations shown for triplicate analyses. It is shown the SF4
typically exhibited the highest sulfur content, but no other clear trends were observed. It is shown that
most of the center point runs produced bio-oil with sulfur contents on the lower end of the expected
range for pyrolysis liquids from wood.
0.000
0.005
0.010
0.015
0.020
ur
conte
0.025
0.030
0.035
0.040
0.045
0.050
12 15 17 19 21 22
Center Point tests
Sulf
nt
(%-wt.,
wb
SF1
)
SF2
SF3
SF4
Whole
Sulfur content range
0.006 - 0.05 %-wt., wb A handbook ,CPL Press: Newbury, UK, 2002; Vol. 2.
Figure 86. Bio-oil sulfur content for center points
The ash content is required for the elemental analysis to estimate the oxygen content of the
bio-oil, but recall the ash content is determined using the TGA analysis previously discussed. The ash
content for the first three fractions is shown in Figure 87 for all center points with standard deviations
for duplicate tests. The ash analysis for SF4 was not performed for all runs as shown. Therefore, the
ash content for the whole bio-oil shown in Figure 87 is calculated based on the assumption that the
contribution from SF4 is negligible. For four separate tests to determine the ash content of SF4 from

142
differen
ered negligible. In general, the ash content for the center point
tests is w
t tests (as shown in Table 79 of Appendix D), the average ash content was found to be 0.028
%-wt. For these four tests, the average mass fraction of SF4 was 1.0 %-wt. of the total bio-oil.
Therefore, the ash contribution from SF4 to the whole bio-oil ash content for these tests is only
0.00028 %-wt., which can be consid
ithin the range for pyrolysis liquids as reported by Oasmaa & Meier [38] as shown below.
0.000
0.025
0.050
0.075
0.100
0.125
0.150
0.175
0.200
t.,
wb)
12 15 17 19 21 22
Center Point tests
Ash
content
(%-w
SF1
SF2
SF3
SF4
Whole
Source: A. Oasmaa et al., Analysis, characterisation and test methods
of fast pyrolysis liquids. In Fast pyrolysis of biomass: A
handbook ,CPL Press: Newbury, UK, 2002; Vol. 2.
Ash content range
0.01 - 0.20 %-wt., wb
Figure 87. Bio-oil ash content for center points
The elemental oxygen content in the bio-oil fractions was then calculated by subtracting the
contributions of carbon, nitrogen, hydrogen, sulfur and ash from 100%. This calculation assumes no
other elements have major contributions to the composition. As just discussed, also recall that the ash
content was not determined ered negligible for
the oxygen calculation for SF4 and for the whole bio-oil. The calculated oxygen content for the center
point te
for all SF4 samples, so the ash contribution is consid
sts for each fraction and the whole bio-oil is shown in Figure 88. Note that the fractions with
higher water content (SF2 and SF4) have oxygen contents that are above the range for pyrolysis
liquids as reported by Oasmaa & Meier [38] as shown. The remaining fractions and the whole bio-oil
fraction have oxygen contents that are within the range for pyrolysis liquids from wood.

143
0
10
20
30
40
50
60
70
80
90
12 15 17 19 21 22
Center Point tests
Oxygen
content*
(%-wt.,
wb)
SF1
SF2
SF3
SF4
Whole
Oxygen content range
44 - 60 %-wt., wb
* Note: By difference
Figure 88. Bio-oil oxygen content for center points
A regression model was performed for each of the main elements (carbon, hydrogen and
oxygen content in the whole bio-oil) with data obtained as discussed. Given that many of the bio-oil
samples had nitrogen values below the detection limit, a model for nitrogen content would provid
little if any insight. Similarly, among different test
conditions, it was assumed a model for sulfur content would also be of little value.
del could be
improved by adding complexity to the model such as cubed terms or more interaction terms, but this
was also not investigated further.
e
as the sulfur values were not found to vary greatly
The residuals for the carbon content data are shown in Figure 144 of Appendix D, and appear
satisfactory for performing a linear regression model. The resulting model was found to be significant
(reject Ho1) with a high R2
value of 97%, however the lack of fit was found to be significant as shown
in Table 37. As with the carbon dioxide yield model, the significance of lack of fit for carbon content
was marginal but still considered significant at the 95% confidence level. The details of the carbon
content model are shown in Table 81 of Appendix D. A reduced model was developed by removing
the 10 insignificant terms for which Ho3 could not be rejected; however this did not improve the lack
of fit of for the reduced model. As the lack of fit was found to be significant and Ho2 could not be
rejected, the model was not investigated further. It is possible that the carbon content mo

144
Table 37. Bio-oil carbon content model, statistics summary
R2
0.970 - 0.951 - -
F0.05,k,ν 2.424 - 2.579 - Reject Ho1
FLOF 6.10 √ 6.33 √ FLOF > F0.05,λ,m-1
F0.05,λ,m-1 4.74 - 2.84 - Don't reject Ho2
t0.05,ν 2.13 - 2.06 - -
|t| statistics
β0 224.22 √ 390.63 √ |t| > t0.05,ν Reject Ho3
β1 18.98 √ 19.03 √ |t| > t0.05,ν Reject Ho3
β4 8.26 √ 8.28 √ |t| > t0.05,ν Reject Ho3
β14 2.13 √ 2.14 √ |t| > t0.05,ν Reject Ho3
11 6.70 √ 7.0 √ |t| > t0.05,ν Reject Ho3
7
β
β22 0.05,ν o3
0.20 X - - |t| < t Don't reject H
Note: * The null hypothesis Ho1 is rejected the full model and the reduced model
1.64
2.54
Full Reduced
The resulting residuals for the hydrogen content in the whole bio-oil compared to the values
predicted by the full model are shown in Figure 145 of Appendix D, and suggest a regression model
is appropriate. The resulting full model for hydrogen content was not found to have a particularly
high R2
value (85.7%), however the F-test was used to reject Ho1 which shows the model is still
significant at 95% confidence as seen in Table 38. As compared to the carbon content model, there
was clearly no lack of fit in the hydrogen content model, so Ho2 was also rejected. A reduced model
was developed by eliminating 11 insignificant terms, and the reduced model was also found to be
significant with no significant lack of fit, and was more significant than the full model (use the MUT
F-test to accept HO4). The details of this model are shown in Table 82 of Appendix D.
The resulting form of the hydrogen content in the whole bio-oil is represented by Equation
30, noting that it is only a function of heat carrier temperature, feed rate and feed rate squared.

145
Table 38. Bio-oil hydrogen content model, statistics summary
R2
0.857 - 0.773 - -
F0.05,k,ν 2.424 - 2.975 - Reject Ho1
FLOF 0.22 X 0.53 X FLOF < F0.05,λ,m-1 *
F0.05,λ,m-1 4.74 - 2.68 - Reject Ho2
t0.05,ν 2.13 - 2.06 - -
|t| statistics
β0 244.20 √ 445.81 √ |t| > t0.05,ν Reject Ho3
β1 7.98 √ 8.34 √ |t| > t0.05,ν Reject Ho3
β4 3.36 √ 3.52 √ |t| > t0.05,ν Reject Ho3
β11 2.98 √ 2.59 √ |t| > t0.05,ν Reject Ho3
Full Reduced
1.47
2.51
n in Figure 89
as a fu
actual
2
HC
HC
HC
μ
0.034
-
μ
0.051
τ
0.122
7.52
wb)
wt.,
(%
content
Hydrogen
⋅
⋅
−
⋅
−
=
−
Equation 30
The model for hydrogen content implies that as temperature and heat carrier feed rate are
increased, the hydrogen content decreases. This effect is likely related to the effect determined by the
moisture content model. The moisture content model showed that for a constant auger speed and
nitrogen flow rate, the moisture content in the bio-oil decreased with increasing temperature and
increasing heat carrier feed rate. The modeled hydrogen content in the bio-oil is show
nction of heat carrier feed rate and temperature. Note that the overall decrease, though
apparent, is relatively minor in terms of the overall percentage of the bio-oil. The predicted and

146
hydrogen content valu nd prediction
intervals.
es are shown in Figure 146 of Appendix D with 95% confidence a
6.8
6.9
7.0
7.1
7.2
7.3
7.4
7.5
7.6
7.7
7.8
400 425 450 475 500 525 550 575 600 625 650
Modeled
bio-oil
H
content
(%-wt.,
wb)
9
12
15
18
21
N2 flow rate = 2.5 sL/min, Auger
speed = 54 RPM
Heat carried
feed rate
(kg/hr)
Figure 89. Modeled bio-oil H content as a function of heat carrier temperature and feed
rate
The bio-oil oxygen content model was also investigated after the residuals as shown in Figure
47 of Appendix D were considered adequate for regression modeling. The full model was found to
be significant with an R2
value of 0.969 (Ho1 rejected), and no significant lack of fit was found (reject
Ho2). Besides the intercept term, there were three additional significant ter
rejected. These details are summarized in Table 39 below. The detailed statistical analysis is saved for
Table 8 Appendix D. In an attempt to reduce the model to fewer significant terms, it was found
that the
1
ms for which Ho3 could be
3 of
lack of fit became significant (Ho2 can not be rejected for reduced model). As with previous
cases where the lack of fit was determined to be significant, there is a potential to develop a more
complex model to decrease the lack of fit, however this was not investigated. However the full model
is significant, and the resulting predicted versus actual oxygen values are shown in Figure 90 with the
95% confidence and prediction intervals.

147
Table 39. Bio-oil oxygen content model, statistics summary
R
2
0.969 - 0.945 - -
FANOVA 33.09 √ 106.4 √ FANOVA > F0.05,k,ν
a
F0.05,k,ν 2.424 - 2.759 - Reject Ho1
FLOF 2.10 X 5.85 √ FLOF < F0.05,λ,m-1
b
F0.05,λ,m-1 4.74 - 2.84 - Reject Ho2
t0.05,ν 2.13 - 2.06 - -
|t| statistics
β0 319.10 √ - - |t| > t0.05,ν Reject Ho3
β1 18.09 √ - - |t| > t0.05,ν Reject Ho3
β11 6.70 √ - - |t| > t0.05,ν Reject Ho3
β44 2.46 √ - - |t| > t0.05,ν Reject Ho3
Full Reduced
Notes: a - The null hypotheses Ho1 is rejected the full model and the reduced model
b - The null hypotheses Ho2 is rejected the full model, but not rejected for the reduced model
60
51
52
53
54
55
56
57
58
59
Actual
oxygen
contnent
(%-wt.,
whole
bio-oil)
51 52 53 54 55 56 57 58 59 60
Predicted oxygen content (%-wt., whole bio-oil)
Figure 90. Actual vs. predicted oxygen content

148
The full model educed model
equation is not presented because it has a significant lack of fit). Though difficult to interpret, the
terms that were found to be significant (as shown in Table 39) dominate the equation.
Equation 31
The model basically predicts that with increasing temperature and heat carrier feed rate, the
oxygen content in the bio-oil will decrease. This is in general agreement with some of the other
findings: higher temperatures and heat carrier feed rates tend to increase the liquid bio-oil yield. With
higher bio-oil yields the moisture content in the bio-oil was found to decrease, which will have an
effect in decreasing the total oxygen content.
With the amounts of elemental carbon, hydrogen and oxygen known in the bio-oil, the
interesting concepts revealed in Figure 63, Figure 65, Figure 66 and Figure 70 can be extended to
offer a possible ‘unifying’ explanation. In Figure 63, Figure 65, Figure 66, it was shown that the gas
yields all increase with bio-oil yield. Due to gas species with hydrogen and oxygen, it was theorized
that this helps explain why the moisture content in the bio-oil also decreases with yield as shown in
Figure 70. This concept is extended further in Figure 91 to show that biochar yield decreases with
increasing bio-oil yields. As discussed previously, this is likely attributed to high heat transfer rates
and short residence times that limit secondary reactions which can increase char formation [32, 119].
The decrease in char yield as a function of bio-oil yield can also be used to help explain the
fascinating results shown in Figure 92. Note that each response is fit with a linear regression line to
indicate a correlation between product yield and bio-oil yield. When the biochar yield decreases, there
is more available carbon in the original biomass available for conversion into liquid and gases. As
Figure 92 shows that bio-oil total carbon content increases with yield, it is clear that although the
formation of carbon containing gases is a competing reaction [27], the formation does not result in
significant carbon losses from the liquid. The total oxygen content in the bio-oil is shown to decrease
with yield, which may be attributed to oxygen containing gases (CO and CO2) being formed. It is
interesting to note that the slope of the regression lines for the carbon content and o
the bio-oil have the same mag
for oxygen content is described by Equation 31 below (the r
2
HC
2
A
2
N2
2
HC
HC
A
HC
N2
HC
HC
A
N2
A
HC
N2
HC
HC
A
N2
HC
μ
0.193
Ω
0.02
θ
0.001
τ
0.526
μ
Ω
0.11
μ
θ
0.13
μ
τ
0.22
Ω
θ
0.11
Ω
τ
0.170
θ
τ
0.010
μ
0.689
Ω
0.004
θ
0.039
τ
1.52
53.64
wb)
wt.,
(%
content
Oxygen
⋅
−
⋅
+
⋅
−
⋅
+
⋅
⋅
−
⋅
⋅
+
⋅
⋅
+
⋅
⋅
−
⋅
⋅
−
⋅
⋅
+
⋅
−
⋅
+
⋅
+
⋅
−
=
−
xygen content in
nitude, just opposite signs.

149
y = 0.12x + 3.52
y = -1.12x + 94.80
R
2
= 0.98
38
43
b)
Non-condensable gas
Biochar
R
2
= 0.83
8
13
45 50 55 60 65 70 75
Pr
18
23
28
33
oduct
yield
(%-wt.,
w
Note: Does not include Run #13 (gas yield measure by difference)
Figure 91. Biochar and non-condensable gas yield vs. bio-oil yield for 29 tests
y = 0.22x + 24.05
R
2
= 0.91
y = -0.22x + 67.99
R
2
= 0.87
y = -0.37x + 50.27
R
2
= 0.88
y = -0.02x + 8.60
R
2
= 0.76
y = 0.29x - 2.70
R
2
= 0.78
0
5
10
15
20
25
30
40 45 50 55 60 65 70 75
%-wt.
bio-o
35
40
45
50
55
60
65
il,
wb
C content
O content
H content
H2O content
H2O insolubles
Total oxygen
Total carbon
Water (H2O)
Total hydrogen
Water insolubles
Figure 92. Bio-oil C, O, H, H2O and water insoluble contents as a function of yield for 30
tests

150
As presented previously (Figure 70), Figure 92 shows that moisture content decreases with
bio-oil yield, with a similar slope to the decrease in total oxygen content. With decreasing water
content as a function of bio-oil yield, a subtle result is a decrease in overall hydrogen content in the
bio-oil. Though this result is not readily apparent in Figure 92, it is shown in the negative slope of the
regression line. This decrease in hydrogen content for increasing bio-oil yield may also be attributed
to the increasing yields of hydrogen containing gas species such as CH4 (Figure 65). Also shown in
Figure 92 for comparison purposes is the water insoluble content of the bio-oil, which is seen to
increase with bio-oil yield as discussed previously (Figure 78).
To summarize and simply some of the underlying concepts resulting from interpretation of
Figure 92, a so-called Van Krevelen diagram [121] was prepared as shown in Figure 93 for all 30
tests. This plot compares the atomic oxygen:carbon ratio vs. the atomic hydrogen:carbon ratio, and a
linear regression fit is shown to indicate a correlation between the ratios. This image clearly illustrates
that in general, as bio-oil yield increases, both the H:C ratio and the O:C ratio decrease.
y = 0.113x + 0.038
R
2
= 0.962
0.17
0.18
0.19
0.20
0.21
0.22
0.23
0.24
0.25
1.20 1.25 1.30 1.35 1.40 1.45 1.50 1.55 1.60 1.65 1.70 1.75 1.80 1.85
Bio-oil O:C ratio
Bio-oil
H:C
ratio
Increasing bio-oil yield
Minimum bio-oil
yield (Run 13)
Maximum bio-oil
yield (Run 20)
Increasing heating value
Figure 93. Bio-oil H:C ratio vs. O:C ratio (Van Krevelen diagram) for all 30 tests
oxygen:carbon ratio all decrease with increasing bio-oil yield, it was suspected this was largely due
Though these results show that total oxygen content, the hydrogen:carbon ratio and the

151
to the re
ng bio-oil yield, as was the case with
for the w
pproximately
15%-wt. to 30%-wt, and has a significant increasing linear relationship with increasing bio-oil yield.
duction in water content (see Figure 70). To investigate this further, the analysis as presented
in Figure 92 and Figure 93 was extended to consider the bio-oil on a dry basis. With known amounts
of elemental carbon, hydrogen and oxygen in the wet bio-oil (ultimate analysis), as well as the
moisture content of the bio-oil (Karl-Fischer titration), the elemental composition can be calculated
for theoretically moisture-free bio-oil. The extenuation of Figure 92 is shown in Figure 94 for the
elemental contents of carbon, oxygen, and hydrogen on a dry bio-oil basis as a function of dry bio-oil
yield, as well as the water insoluble content on a dry bio-oil basis. It is shown here that there is not a
significant linear relationship of increasing carbon with increasi
et bio-oil analysis. However with increasing yield, the organic oxygen content in the bio-oil
is still shown to decrease slightly, independent of oxygen in the bio-oil water content. The results
from this analysis therefore support the previous theory that a portion of the oxygen from the original
biomass is converted to oxygen containing gases at higher bio-oil yield conditions (see Figure 63). On
a dry bio-oil basis, the water insoluble portion of the bio-oil is shown to range from a
y = 0.0371x + 50.378
R
2
= 0.0821
y = -0.056x + 45.443
R
2
= 0.2252
y = 0.0279x + 3.3036
R
2
= 0.7021
y = 0.3165x + 6.6928
R
2
= 0.7319
0
5
10
15
20
25
30
35
40
45
50
55
25 30 35 40 45 50 55 60
Bio-oil yield (%-wt., db)
%-wt.
bio-oil,
db
C content
O content
H content
H2O insolubles
Total oxygen
Total carbon
Total hydrogen
Water insolubles
p-value = 0.1248
p-value = 0.008
p-value = < 0.0001
p-value = < 0.0001
Figure 94. C, O, H, H2O and H2O insoluble contents as a function of yield for 30 tests,
dry basis

152
Whereas on a wet bio-oil basis the hydrogen content was shown to decrease with yield (due
to decreasing water content), on a dry basis the hydrogen content is shown to increase slightly with
increasing yield. The analysis presented in Figure 93 was then extended to consider the whole bio-oil
hydrogen:carbon ratio and oxygen:carbon ratio on a dry basis as shown in Figure 95. Similar to the
wet basis analysis, the oxygen:carbon ratio on a dry basis also decreases with increasing bio-oil yield,
though less dramatically. This result is shown by comparing Figure 92 and Figure 94. However,
unlike the wet basis analysis, the hydrogen:carbon ratio increases with increasing bio-oil yield on a
dry basis. The dry basis analysis in Figure 95 is seen to have a much closer grouping of elemental
ratios on the Van Krevelen diagram compared to the wet basis analysis. Based on these results, the
reduction in the hydrogen:carbon ratio and the oxygen:carbon ratio (on a wet basis) with increasing
bio-oil yields is largely due to decreasing water content. However it is important to note the results
show there is still a reduction in the oxygen content of the organic portion of the bio-oil as yield
increases.
0.070
0.095
0.120
0.145
0.170
0.195
0.220
0.245
0.70 0.80 0.90 1.00 1.10 1.20 1.30 1.40 1.50 1.60 1.70 1.80 1.90
Bio-oil O:C ratio
Bio-oil
H:C
ratio
Dry basis
Wet basis
Figure 95. Bio-oil H:C ratio vs. O:C ratio for all 30 tests, including dry basis analysis
Total acid number. The total acid number was determined for six center points for all four
fractions, except the test was not performed for SF4 for Run 12. The analytical data for these test
performed in duplicate, is shown in
Figure 96. The whole bio-oil for five of t e t tests averaged a TAN value of 108 ± 0.64
s,
Table 84 of Appendix D. These results are shown graphically in
h center poin

153
mg/g, where SF2 has the highest TAN, followed by SF1, SF3 and SF4. As shown in Figure 96, the
SF1, SF3 and whole bio-oil values are within the range of recently published values from bio-oil
produced from oak wood and pine wood in the MSU auger reactor system [86].
0
10
20
30
40
50
60
70
80
90
100
110
120
130
tal
Acid
Number
(mg/g)
12 15 17 19 21 22
Center Point tests
To
SF1
SF2
SF3
SF4
Whole
Oak wood/bark (120 mg/g)
Pine bark (84 mg/g)
Figure 96. Total acid number for center points
Gas chromatography/Mass spectrometry (GC/MS). The GC/MS analysis was used to
understand and interpret the chemical composition of the bio-oils from the auger rector, which may
be useful to compare to other studies. Each bio-oil fraction was analyzed as discussed previously, and
the concentrations of 32 compounds were quantified (Table 20). However for initial qualitative
analysis, the chromatogram output from the GC/MS is often instructive. For instance t
chromato ,
and the SF4 chromatogram for the same test is shown in Figure 98. Visual comparison of these two
figures
he
gram from SF1 for the test with the highest bio-oil yield (Run 20) is shown in Figure 97
reveals that there are indeed variations in chemical composition between SF1 and SF4. The
peaks for certain compounds that were among those quantified are labeled on the figures. The
chromatograms for SF2 and SF3 for the same test can be found in Figure 148 and Figure 149 of
Appendix D, respectively.

154
1
7
6
5
4
3
2
(1) Acetic acid (2) 2-Propanone, 1-hydroxy- (3) Phenol, 2-methoxy-4-methyl (4) Phenol, 2,6-dimethoxy-
(5) 4 methyl 2,6 dimethoxy phenol (6) Levoglucosan (7) Ehtanone, 1-(4-hydroxy-3,5-dimethoxyphenyl)
Figure 97. GC/MS chromatogram for SF1, Run #20 (bio-oil max yield)
A sample of the complete quantified GC/MS data for run 20 as shown in Figure 97 and
Figure 98 can be found in Ta ot presented for each
run however, and instead the compounds will be grouped together as discussed previously.
ble 85 of Appendix D. This complete analysis is n
The GC/MS data for each run, with compounds grouped by chemical families, can be found
summarized in Appendix D as follows: Table 86 for SF1, Table 87 for SF2, Table 88 for SF3, Table
89 for SF4, and Table 90 for the resulting whole bio-oil. Inspection of the results indicated that
though there was some difference in certain compounds among the fractions, overall the values were
similar among different runs. Regression modeling procedures were attempted for various compounds
and grouping of compounds, but the resulting models were insignificant and did not warrant further
investigation. Instead, the data was compared to known information on bio-oil chemical composition,
and organized to compare the composition among bio-oil fractions.

155
1
2
(1) Acetic acid (2) 2-Butanone, 3-hydroxy (3) Furfual (4) Phenanthrene - internal standard
3
4
Figure 98. GC/MS chromatogram for SF4, Run #20 (bio-oil max yield)
As the reactor for this project is a first gene tion design, it is instructive to compare the
compou
pared to
the valu
so found to be within the range,
and though levoglucosan is shown to be slightly higher than the range reported by Diebold, other
references such as Mohan et al. [4] would consider this within range or even on the low end. It is
interesting to note that most of the phenolic compounds were found to be lower than the common
values for pyrolysis oil.
ra
nds in the bio-oil to known information. For this purpose, a comprehensive list of common
chemicals and their concentration in bio-oil was referenced by Diebold [122]. The “low” and “high”
values common for bio-oil were tabulated for 27 of the 32 quantified compounds, and com
es averaged from the whole bio-oil for the 6 center points runs and the average for all 30 runs.
The results of this comparison are shown in Table 40. Note that the last column indicates if the values
from this study are in agreement with the values according to Diebold (denoted by a “√”), higher than
the values reported (denoted by a “+”), or lower than the values reported (denoted by a “-”).
In this study, 13 of the quantified chemical compounds (48%) were found to be in agreement
within the range as reported by Diebold [122]. These compounds, highlighted in gray, also represent
at least one compound from each of the five major groupings (furans, phenols, guaiacols, syringols,
and other oxygenates) as described in Table 20. Acetic acid was al

156
Table 40. GC/MS characterized compound comparison, whole bio-oil
Chemical compound Low High
30 run
average
Center point
average
Comparison
to typical b
Acetic acid 0.50 12.00 3.01 2.95 √
2-Propanone, 1-hydroxy- 0.70 7.40 2.16 2.49 √
2-Butanone, 3-hydroxy- - - 0.19 0.19
Furfural 0.10 1.10 0.21 0.27 √
2-Furanmethanol 0.10 5.20 0.22 0.22 √
2-Cyclopenten-1-one, 2-methyl- 0.10 1.90 0.03 0.03 -
2-Furancarboxaldehyde, 5-methyl- 0.10 0.60 0.07 0.07 -
2H-Pyran-2-one - - 0.12 0.10
1,2-Cyclopentanedione, 3-methyl- 0.10 0.50 0.59 0.59 +
2(5H)-Furanone, 3-methyl- 0.10 0.60 0.19 0.24 √
Phenol 0.10 3.80 0.04 0.04 -
Phenol, 2-methoxy- 0.10 1.10 0.52 0.51 √
Glycerin - - 0.18 0.30
Phenol, 2-methyl- 0.10 0.60 0.04 0.04 -
Phenol, 4-methyl- 0.10 0.50 0.06 0.07 -
Phenol, 3-methyl- 0.10 0.40 0.05 0.05 -
Phenol, 2-methoxy-4-methyl- 0.10 1.90 0.23 0.24 √
Phenol, 2,5-dimethyl- 0.10 0.40 0.04 0.04 -
2,4-Dimethylphenol 0.10 0.30 0.04 0.04 -
Phenol, 2-ethyl- 0.10 1.30 0.04 0.04 -
Phenol, 3-ethyl- 0.10 0.30 0.04 0.04 -
Phenol, 3,4-dimethyl- 0.10 1.90 0.04 0.04 -
Phenol, 4-ethyl-2-methoxy- - - 0.11 0.11
ugenol 0.10 2.30 0.15 0.14 √
Typical bio-oil rangea
This study
E
2-Furancarboxaldehyde, 5-(hydroxymethyl) 0.30 2.20 0.33 0.34 √
Phenol, 2,6-dimethoxy- 0.70 4.80 1.00 1.03 √
Phenol, 2-methoxy-4-(1-propenyl)-, (E)- 0.10 7.20 0.33 0.34 √
4 methyl 2,6 dimethoxy phenol - - 0.75 0.80
Vanillin 0.10 1.10 0.42 0.41 √
Hydroquinone 0.10 1.90 0.10 0.11 √
1,6-Anhydro-β-D-glucopyranose 0.40 1.40 1.92 2.07 +
Ethanone, 1-(4-hydroxy-3,5-dimethoxyphenyl) 0.10 0.30 1.21 1.21 +
Sum 4.80 63.00 14.46 15.18
Notes: All values in %-wt. a - Reference: Diebold, J.P. A review of the chemical and physical mechanisms of
the storage stability of fast pyrolysis bio-oils. In Fast pyrolysis of biomass: A handbook ,CPL Press: Newbury,
UK, 2005; Vol. 2. b - √ = Values within range for typical bio-oils, - = Values less than typical range for
typical bio-oils, + = Values greater than range for typical bio-oils.
resented as the average values for the 6 center point runs, and the standard
The quantified compounds were also cross-checked with those as reported by Ingram et al.
[86] and Garica-Perez et al. [88] for bio-oil produced from wood in two different lab-scale auger
reactors. Ingram et al. identified 17 of the 32 quantified compounds in this study, and Garcia-Perez et
al. identified 8 of the 32 compounds.
The concentration of acetic acid, levoglucosan, and the remaining groups of chemical
families are shown as a function of bio-oil fraction and whole bio-oils in Figure 99 below. The
fractions SF1 – SF4 are p

157
deviations are sh io-oil is
shown for the center point averages (Whole – CP), and can be compared to the resulting whole bio-oil
data averaged from all runs (Whole – 30 tests). When comparing the two whole bio-oils, this image
illustrates that there is minimal difference among chemical speciation as a function of test conditions.
When considering the deviation among runs, the composition of the bio-oil is basically identical for
the center point average (same conditions) and the total experimental average (many different
conditions). Figure 99 shows that the average acetic acid does not vary greatly among fractions, but
that the instrument and procedure may cause some difficulty in quantifying acetic acid. This is noted
because many of the other quantified compounds have much lower deviations among the tests, and
because all of the other bio-oil tests for the center point runs have shown these runs are very similar in
composition. It is also shown in Figure 99 that fraction SF1 and SF3 have very similar chemical
compositions. Besides phenols (and furans to a less extent), SF2 was found to have lower levels of
levoglucosan, guaiacols, and syringols compared to SF1 and SF3. Fraction SF4 had low levels of
furans, guaiacols, syringols, virtually no phenols and no levoglucosan.
own as deviations among runs and not replicates. The resulting whole b
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
5.5
6.0
SF1 - CP
SF2 - CP
SF3 - CP
A
c
L
e
v
o
Chemical compound or family
e
t
i
c
a
c
i
d
g
l
u
c
o
s
a
n
F
u
r
a
n
s
P
h
e
n
o
l
s
G
u
a
i
a
c
o
l
s
S
y
r
i
n
g
o
l
s
O
t
h
e
r
Concentration
(%-wt.)
SF4 - CP
Whole - CP
Whole - 30 tests
Note:
CP = Center Point
average
(same conditions)
Figure 99. GC/MS quantified volatile compounds
The GC/MS data presented in Figure 99 can alternatively be presented by fraction rather than
by compound as shown in Figure 100. This image clearly shows the similarities between SF1 and
SF3, as well as the general similarity of these two fractions to the whole bio-oil. Also note the total

158
amount of mass quantified by the procedure varies among fraction: from less than 5 %-wt. for SF4 to
almost 18 %-wt. for SF1. As discussed, there are many different compounds in bio-oil, and though
the GC/MS instrument detects many of them, it is only calibrated to quantify the concentration of
certain, common compounds. Also, a significant portion of bio-oil is non-volatile, implying that many
of the compounds present can not be quantified with GC/MS analysis. Oasmaa & Meier [38] estimate
that only 35%-wt. of the bio-oil mass is volatile, with the balance made up of water, water insolubles,
and non-volatile compounds. Based on the low amount of volatiles quantified (< 18%-wt.), this
implies that there is an opportunity to identify and quantify more of the compounds in the bio-oil
form the auger reactor.
0
2
4
6
8
10
12
14
16
18
(%-wt.
SF1 SF2 SF3 SF4 Whole
Bio-oil fraction (Center point average)
Concentration
)
Other GC/MS
Syringols
Guaiacols
Phenols
Furans
Levoglucosan
Acetic acid
Figure 100. GC/MS quantified volatile compounds by fraction for center points
The data presented in Figure 100 can be extended to consider the total mass of bio-oil that
was quantified. Recall that the KF moisture content test determines what percentage of the bio-oil is
water, and the water insoluble test determines a certain percentage of the bio-oil mass as well. When
considering the mass that is quantified by f bio-oil
that remains unidentified. Refer to Figure 150 in Appendix D for an image of the quantified mass for
all runs.
GC/MS analysis, there is some additional mass o

159
Viscosity. The viscosity of some representative bio-oil samples was investigated for
comparison purposes. Similar to the higher heating value of bio-oil, viscosity is a strong function of
moisture content. Therefore, in addition to testing the six center point tests (SF1 – SF3), the minimum
and maximum water content samples were also investigated. The SF4 fraction was not tested for
viscosity because it’s high water content implies the viscosity will be very similar to that of water and
is therefore of little interest. Also, the small volume available from this sample precludes viscosity
testing.
Viscosity measurements were taken every 30 seconds for five minutes at a constant shear
rate, and a minor shear thinning effect was observed with time. Typically this effect was not observed
past five minutes as shown in Figure 101. This figure shows data from the maximum bio-oil yield test
(run 20). An attempt was made to analyze each fraction at the same shear rate, however based on the
large difference in viscosity among fractions; different spindles were required for analysis. This
resulted in shear rates that ranged from 38 s-1
to 98 s-1
, which is actually a very close range compared
to the whole possible range. The standard deviation among these 11 measurements is shown in Figure
102, and the viscosities at the center points are averages of the six center points and are shown with
pooled standard deviations. The analytical data for these tests is shown in Table 91 of Appendix D.
0
50
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Time (min)
V
100
osit
150
200
250
300
isc
y
(cP)
@
40°C
SF1
SF2
SF3
Constat shear rates (s
-1
): SF1, SF3 = 38.4, SF2 = 48.9
Figure 101. Viscosity measurements for Run #20 vs. time

160
As shown in Figure 102, the viscosity range is in general agreement with published literature,
though it is often difficult to compare viscosity measurements due to differences in temperatures,
shear rates, testing methods and other inconsistencies. The viscosity measurements for this study were
taken at 40°C as recommended by Oasmaa et al. in a 2005 report on the norms and standards for
pyrolysis liquids [117]. Note that two of the comparisons from Ingram et al. as shown in Figure 102
are at 50°C and different shear rates. However the “typical range” for wood derived bio-oil as
reported by Bridgwater is 2007 is 40 – 100 cP [21], which is shown in Figure 102 as well, and it is
likely the mixed bio-oil from this study would fall within that range. Unlike other properties,
however, viscosity was not mass averaged to determine the whole bio-oil value.
0
25
50
75
100
125
150
175
200
225
250
275
cP)
@
40°C
(Run 13)
(Run 20)
Viscosity
(
SF1
SF2
SF3
Oak wood (171 cP @ 50°C, 0.05s
-1
)
Oak wood (36 cP @ 50°C, 300s
-1
)
*a b c c b c c d c
* Shear rate (s
-1
): a = 30.6, b = 97.8, c = 38.4, d = 48.9
Wood derived bio-oil (40 - 100 cP @ 40°C)
Range as reported by Bridgwater. Source: Int. J.
Global Energy Issues 2007, 27 , 160 - 203.
Figure 102. Bio-oil viscosity range
Vapor temperature. To conclude the results section, a brief discussion is given on the
temperature of the vapor exiting the reactor. This temperature is useful because it gives insight to the
actual “reaction” temperature rather than the heat carrier inlet temperature. To aid in future discussion
and comparison efforts, the average of
the “vapor phase reaction temperature” was estimated as

161
reactor temperatures 1 and 2 as shown schematically in Figure 103. Recall that for all experiments,
these temperatures were averaged over the steady state duration as shown in Table 53 of Appendix D.
The dimensions related to Figure 103 are shown in Figure 26, and the thermocouple configuration is
shown in Figure 113 of Appendix A.
Figure 103. Reaction temperature schematic
As the heated solid heat carrier reacts with biomass in the reactor, the vapor products leave at
a certain temperature that is a function of the heat carrier temperature, but also other a function of
other parameters. This is shown by the variance in the data points plotted in Figure 104 (all tests) as a
function of heat carrier temperature only. For instance it has been shown that auger speed and heat
carrier feed rate are significant factors for many responses, so it is likely these factors will also
influence the reaction temperature as defined as the average of reactor temperatures 1 and 2.
Reactor 1 (R1)
425
450
475
500
525
Vapor
temperature
(°C
R1, R2 Average ("RXN temperature")
Reactor 2 (R2)
)
400
400 425 450 475 500 525 550 575 600 625 650
Figure 104. Vapor temperatures vs. heat carrier temperatures

162
Therefore, a regression model was performed as described previously, to determine the
reaction temperature as a response to multiple factors. The details will not be discussed, but the
residuals were found to be acceptable as shown in Figure 151 of Appendix D, and the subsequent full
model fit the data very well with an R2
value of 0.981. The statistical summary is shown in Table 41
below, and the details are shown in Table 92 of Appendix D.
Table 41. Reaction temperature model, statistics summary
R2
0.981 - 0.971 - -
F0.05,k,ν ject Ho1
FLOF 3.58 X 0.66 X FLOF < F0.05,λ,m-1 *
F 4.74 - 2.59 - Reject H
2.424 - 2.621 - Re
0.05,λ,m-1 o2
t0.05,ν 2.13 - 2.06 - -
|t| statistics
β0 641.54 √ 1136.5 √ |t| > t0.05,ν Reject Ho3
β1 25.12 √ 25.71 √ |t| > t0.05,ν Reject Ho3
β3 4.29 √ 4.39 √ |t| > t0.05,ν Reject Ho3
β4 9.65 √ 9.88 √ |t| > t0.05,ν Reject Ho3
β
β
β
14 4.41 √ 4.51 √ |t| > t0.05,ν Reject Ho3
β
β
β
β
β
β
β44 2.43 √ 2.43 √ |t| > t0.05,ν Reject Ho3
Full Reduced
1.32
2.59
Both the full model and the reduced model were found to be significant based on the F-test
(reject Ho1), and there was no significant lack of fit for either model (reject Ho2). The model utility test
showed that the reduced model was more significant than the full model (don’t reject Ho4), and the t-
tests eliminated 9 pa perature as
rameters that were insignificant. The expected vapor reaction tem

163
measure
and not the physical quantities.
Equation 32
d compared to the temperature predicted by the model is shown in Figure 105, along with
95% confidence and prediction intervals. The reduced model has a very low RMSE of 1.7°C.
The resulting equation from the reduced model for the reaction temperature is shown in
Equation 32, noting there is a significant interaction term and a significant higher order term. Also
recall that the coefficients are associated with the coded factors
2
HC
HC
HC
RXN
μ
0.789
-
μ
τ
1.96
46
C)
(
T
⋅
⋅
⋅
+
=
 HC
A
HC μ
3.51
Ω
1.56
τ
9.13
5.9 ⋅
+
⋅
+
⋅
+
445
450
455
460
465
470
475
480
485
Actual
reaction
temperature
(Deg
C)
445 450 455 460 465 470 475 480 485
Predicted reaction temperature (Deg C)
Figure 105. Actual vs. predicted reaction temperature
The absolute values of the t-statistics from Table 92 are shown graphically in Figure 106. As
fully expected, heat carrier temperature is the most influential term in modeling the reaction
temperature. However heat carrier feed rate is also shown to be significant, which is based the effect
it has on heat transfer as previously discussed.

164
0 2.5 5 7.5 10 12.5 15 17.5 20 22.5 25 27.5
HC temperature
Auger speed
HC feed rate
HC temperature · HC
feed rate
HC feed rate · HC feed
rate
Model
term
t
0.05,
24
Interaction effect
Higher order effect
Main
effects
Figure 106. Absolute values for t-test statistics for vapor temperature model
In general, the models shows that the reaction temperature increases for increasing heat
carrier temperature, but it also increases as feed rate and auger speed increase. This result provides
some insight into the heat transfer mechanisms and suggests higher heat carrier feed rates and auger
speeds provide higher heat transfer rates. This result is in agreement with the bio-oil yield model that
shows yield increases for high heat carrier feed rates and high auger speeds (at high temperatures).
This also agrees with general fast pyrolysis knowledge that yield is increased with high heat transfer
rates.
The modeled vapor reaction temperature is shown in Figure 107 as a function of heat carrier
temperature and feed rate while holding the nitrogen flow rate and auger speed constant at the center
point conditions. This representation strengthens the understanding of the relationship between the
heat carrier temperature and the “optimal” fast pyrolysis temperature as reported by the literat re. For
instance the bio-oil yield rrier temperature
around C, which is shown to correspond to a vapor temperature of 490°C for 18 kg/hr. This is an
expecte
u
model predicts the highest liquid yields occur at a heat ca
625°
d optimal vapor temperature value for biomass fast pyrolysis [21].

165
440
450
460
470
480
490
500
425 450 475 500 525 550 575 600 625
Modeled
vapor
temperature
(°C) 9
12
15
18
21
Auger speed = 54 RPM
(kg/hr)
ater. Similarly, the critical F-value for lack of fit is different for each model, but typically it
become
Figure 107. Modeled vapor temperature vs. heat carrier temperature
Summary. The statistical results from the regression models are shown summarized in Table
42 for the reduced models. Also recall that the root mean square error (RMSE) has the same units as
the model response. The critical FANOVA value to indicate significance is a function of the model
parameters, but for this study typical significance (95% confidence) occurs around F-values of 2.4
and gre
s significant (95% confidence) for values of approximately 2.5 and greater.
Table 42. Regression models, summary of statistics
Other
Stat.
Bio-oil
yield
(%-wt.)
Biochar
yield
(%-wt.)
CO
yield
(%-wt.)
CO2
yield
(%-wt.)
KF
moisture
(%-wt.)
H2O
insolubles
(%-wt.)
C
content
(%-wt.)
H
content
(%-wt.)
O
content
(%-wt.)
Vapor
RXN T.
(°C)
R2
0.984 0.948 0.980 0.833 0.907 0.912 0.951
RMSE 1.12 1.93 0.08 0.16 0.99 0.77 0.42
Yield models Bio-oil properties models
0.773 0.945 0.971
0.07 0.42 1.74
FANOVA 163.1 70.0 156.9 15.6 61.1 139.7 121.4 29.5 106.4 160.8
P-value <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001
FLOF 1.13 1.21 1.27 5.28 0.48 1.85 6.33 0.53 5.85 0.66

166
The significance of each term for all models is summarized below in Table 43, noting the
grouping of main effects, interaction and higher order terms. Refer to Table 27 for the description of
each term. Note that β1 (heat carrier temperature) and β4 (heat carrier feed rate) were significant terms
for all the models, and β24, β22 and β33 were insignificant for all models. It is also noted that in
general, the yield models tended to have more significant terms than the models for bio-oil properties.
Table 43. Regression models, summary of significant terms
Other
Term
Bio-oil
yield
(%-wt.)
Biochar
yield
(%-wt.)
CO
yield
(%-wt.)
CO2
yield
(%-wt.)
KF
moisture
(%-wt.)
H2O
insolubles
(%-wt.)
C
content
(%-wt.)
H
content
(%-wt.)
O
content
(%-wt.)
Vapor
RXN T.
(°C)
β0 √ √ √ √ √ √ √ √ √ √
β1 √ √ √ √ √ √ √ √ √ √
β2 √ √ X X X X X X X X
β3 √ X √ √ X X X X X √
β4 √ √ √ √ √ √ √ √ √ √
β12 X X X X X X X X X X
β13 √ √ √ √ √ X X X X X
β23 X X √ √ X X X X X X
β14 √ X X X X X √ X X √
24 X
X X X X X X X X X
β
34 X X √ X X X X X X X
β
β11 √ √ X √ √ X √ √ √ X
β44 √ √ √ X X X X √ √
Note:
X Term is not significant at 95% confidence level
√ Term is significant at 95% confidence level
Higher
order
effects
√
Yield models Bio-oil properties models
Main
effects
Interaction
effects
A summary of the analyzed bio-oil physical properties for the center point average and the
maximum bio-oil yield sample (Run #20) is shown in Table 44, which compares the properties to
typical bio-oil. Compared ample has a lower
water content and oxygen content, and a higher carbon content all of which help to increase its
heating value. This result is also reflected in Figure 92 and Figure 93.
to the center point average, the highest bio-oil yield s

167
Table 44. Bio-oil analysis summary and comparison
Typical
propertiesa
Bio-oil fraction > SF1 SF2 SF3 SF4 Whole SF1 SF2 SF3 SF4 Whole Whole
Mass yield
(%-wt. bio-oil) 47.5 31.9 18.8 1.8 100.0 45.4 31.6 20.8 2.1 100.0 100.0
Moisture content
Center point test average
(67.4 %-wt. bio-oil)
Maximum bio-oil yield test
(73.6%-wt. bio-oil)
(%-wt.) 16.5 41.3 17.8 65.9 25.7 10.7 37.5 18.2 71.0 22.0 20 - 35
Solids content
(%-wt.) 0.01 - 1.0
HHV (MJ/k 16 - 19
Viscosity
(cP @ 40°C) 115.6 5.3 146.0 - - 234.5 9.7 255.0 - - 40 - 100
Water insolubles
(%-wt., wb) 17.3 7.6 26.6 0.21 15.6 24.4 12.1 36.5 - 22.5 15 - 30
Ultimate analysis
C (%-wt., wb) 44.5 28.0 45.7 11.7 38.8 46.0 30.8 46.0 13.8 40.5 32 - 49
N (%-wt., wb) 0.045 0.035 0.072 0.070 0.047 0.035 0.008 0.028 0.008 0.025 0.0 - 0.3
H (%-wt., wb) 7.0 8.2 7.1 9.0 7.5 7.0 8.0 7.2 9.7 7.4 6.9 - 8.6
S (%-wt., wb) 0.006 0.006 0.004 0.011 0.006 0.000 0.006 0.007 0.012 0.004 0.006 - 0.05
Ash (%-wt., wb) 0.037 0.056 0.064 0.032 0.048 - - - - - 0.01 - 0.2
O - By diff. (%-wt., wb) 48.4 63.7 47.1 79.2 53.6 47.0 61.1 46.8 76.4 52.1 44 - 60
1.07 0.96 0.70 0.32 0.94 - - - - -
g) 18.7 12.1 19.2 6.5 16.4 19.2 13.3 19.5 7.0 17.1
Notes: a – References: Bridgwater et al. [13], Czernik et al. [36], Oasmaa et al. [38]

168
CHAPTER 6. CONCLUSIONS
Based on the development of the auger reactor system, experimental testing procedures, and
interpretation of the analyzed data as discussed, several conclusions can be made. These conclusions,
along with future recommendations are discussed next.
6.1 Research conclusions
An operational lab-scale auger reactor for biomass fast pyrolysis was researched, designed,
constructed, demonstrated and extensively tested. There is minimally published data on auger reactors
for bio-oil production, and the results from this study contribute to the body of knowledge for fast
pyrolysis.
The engineering design and operational procedures are validated by the product yields well
within the fast pyrolysis regime. This system achieved higher bio-oil yields than any published results
from similar lab-scale reactors using similar feedstocks. The product composition of the bio-oil
produced was very similar to accepted values as reported by published literature.
Design. The auger reactor design is concluded to be suitable for fast pyrolysis processing.
The tested heat carrier feed rates provide sufficient reaction heat and heat transfer rates. In practice,
the auger reactor design holds promise for being a robust system capable of continuous processing
with minimal carrier gas compared to other designs. For industrial sized systems, this may lead to
lower operating costs due to minimal gas handling and compression equipment. The requirement for
minimal carrier gas also suggests that the auger design may be more compact than other reactor types.
Operation. The auger reactor system can be operated to obtain repeatable results with
excellent mass closures near 100%. The mass balance procedure is adequate, including the estimation
of non-condensable gas mass yield using Micro-GC data and a dry volume meter. The operating
conditions to achieve high bio-oil yields have been established.
Design of experiments. The results from this study indicate that the experimental design
selected was not only adequate, but necessary to discover the existence of interaction effects and
higher order terms. These significant terms, namely the interaction between auger speed and heat
carrier feed rate, have not yet been discussed in the literature. The four factors and five levels of the
design were carefully selected and allowed for a wide range of responses to be investigated. In

169
addition to the operating procedures, these experimental factors and levels allowed for the collection
of data that was used to deve ll be discussed next.
he bio-oil yield model is the interaction effect
etween heat carrier temperature and auger speed. For heat carrier temperatures less than 550°C,
higher auger speeds increase bio-oil yield whereas at temperatures below 550°C, low auger speeds
el conclusion because it helps explain why lab-scale auger reactors
that do
e heat carrier material for a short time period. The hypothesis is that when no heat carrier
material
econdary reactions,
whereas
increasing temperature and heat carrier feed rate. This implies that as bio-oil yield increases, the
loped several linear regression models that wi
6.1.1 Regression models
Bio-oil yield. The most notable conclusion for t
b
increase bio-oil yield. This is a nov
not use a heat carrier material operate with such low auger speeds compared to those reactor
systems that do use a heat carrier material. The conclusion from this study is that the introduction of
heat carrier material can provide high liquid yields by improving heat transfer, but only if the biomass
contacts th
is used, longer solid residence times are required (via slow auger speeds) to provide
sufficient reaction heat and time.
Biochar yield. Biochar yield is minimized for similar conditions that favor bio-oil yield. To
minimize biochar yield, high auger speeds are desired above 525°C to minimize s
low auger speeds minimize biochar yield at temperature below 525°C by encouraging
mixing. This conclusion provides insight into the flexibility of the auger reactor design, and the
ability to easily shift the product distribution based on auger speed.
Carbon monoxide yield. The conditions that favor carbon monoxide yield are similar to
those that increase bio-oil yield. As the carbon monoxide yield increases it may contribute to a
reduction in oxygen content in the organic portion of the bio-oil.
Bio-oil moisture content. Moisture content was found to decrease for increasing heat carrier
temperature and feed rate, which are conditions that favor high bio-oil yield and low biochar yield.
For heat carrier temperatures above 525°C, high auger speeds are desired to decrease bio-oil moisture
content.
Bio-oil hydrogen content. Total elemental hydrogen in the bio-oil was found to decrease for
increasing heat carrier temperature and feed rate. The decrease bio-oil hydrogen content for
increasing bio-oil yield is attributed to gas formation of hydrogen containing species, and is related to
the reduction of moisture content as bio-oil yield increases.
Bio-oil water insoluble content. Water insoluble content was found to increase with

170
amount of water insoluble material in the pyrolysis oil will also increase because the higher
temperatures help to decompose lignin.
Vapor reaction temperature. Based on measured temperature data, it is concluded that the
optimal vapor temperature for the auger reactor is similar to other systems, around 490°C. The higher
equired, though, provide evidence of substantial temperature gradients
and infl
ts were performed to characterize the physical and chemical
compos
conditions. The resulting whole
bio-oil
in the bio-oil samples was found to
vary am
pyrolysis oil.
he bio-oil samples is inversely related the
moistur
heat carrier inlet temperatures r
uential solid-solid heat transfer effects.
The development and interpretation of these models is important to satisfy the optimization
objective of this project. The resulting equations for each of these models can be used to estimate the
resulting response for any number of different operating conditions, and the output value will lie
within a known interval. This then provides a powerful tool to estimate the operating conditions of the
reactor required to give a desired result.
6.1.2 Product analysis
Extensive analytical tes
ition of the pyrolysis products from the auger reactor.
Moisture content. The moisture content of the bio-oil varied among fractions consistently,
and the experimental testing procedure was acceptable. Though the moisture content between SF1
and SF3 varied (neither was consistently higher than the other), their magnitudes were similar and the
SF3 moisture content varied much less among tests with different
moisture content was within the accepted range for fast pyrolysis liquids and was similar to
reported literature.
Water insoluble content. The water insoluble content
ong fractions, in decreasing order: SF3, SF1, SF2. The water insoluble content for SF1 and
SF3 was well within the range for typical pyrolysis oil from wood, and the insoluble portion in SF2 is
less than typical bio-oil. For the whole bio-oil, the water insoluble content is on the low end of the
range for
Solids content. The solids content of the center point tests did not vary among fractions, and
was found to be within the range for pyrolysis liquids. This implies the gas cyclone used for this
research was adequate for separating biochar from the vapor product stream.
Higher heating value. The higher heating value of t
e content of the liquid, and increases for increasing bio-oil yield. The higher heating value of

171
SF1, SF3 and the whole bio-oil was found to be within the expected range for pyrolysis liquids. The
heating value for SF2 and SF4 was found to be less than expected for pyrolysis liquids.
the moisture and volatiles.
rolysis oil. Though there was not great variation among
actions, SF4 was found to have the highest nitrogen levels. The hydrogen content was found to be
higher in the bio-oil fractions with higher water levels, SF2 and SF4. The reaming fractions, SF1 and
o-oil had hydrogen content values within the expected range for bio-oil.
The sul
-oil samples was not found to vary
greatly
the range for typical bio-oil, but on the high end of the range. This result is in
accorda
ecreasing with yield largely due to a reduction in moisture content, however
on a dry
by SF1, SF3 and SF4 in
order. T
Thermal gravimetric analysis. The TGA methodology allowed for a complete proximate
analysis of the biochar, which was found to be in agreement with published data on biochar from a
similar lab-scale auger reactor. The methodology also allows for determining the fixed carbon content
and ash content of bio-oil, with less emphasis on
Elemental analysis. The carbon content for SF1, SF3 and the whole bio-oil was found to be
within the range for pyrolysis liquids, and slightly on the lower end of the range. The carbon content
for SF2 and SF4 was below the typical range. The nitrogen was found to be very low in all fractions
and was often below the detection limit of the instrument. The nitrogen levels that were detected were
within the expected range for wood py
fr
SF3, as well as the whole bi
fur content was found to be very low in all fractions, and well within the range expected for
fast pyrolysis oils. Similar to nitrogen, the sulfur content in the bio
between fractions, but overall SF4 was found to have the highest sulfur levels. The ash
content in the bio-oils was found to be within the range for typical bio-oils, and did not appear to vary
among fractions. The oxygen content in the bio-oil, calculated as discussed, was found to be very
similar for SF1 and SF3. Due partly to the high water contents, SF2 and SF4 were found to have
oxygen contents outside the range for typical bio-oils from wood. The resulting whole bio-oil oxygen
content was within
nce with the average carbon content being on the low end of the range for pyrolysis oils.
On a wet bio-oil basis, the hydrogen:carbon and oxygen:carbon ratios decrease with
increasing bio-oil yield. On a dry bio-oil basis, the hydrogen:carbon ratio increases with increasing
yield, and the oxygen:carbon ratio still decreases with increasing yield. On a wet basis the
oxygen:carbon ratio is d
basis the ratio decreases with yield in part to an increase in oxygen containing gases (namely
carbon monoxide).
Total acid number. SF2 had the highest total acid number, followed
hough SF4 had the lowest TAN, it also has the highest water content. The TAN values are
similar to those reported in a recently published study on bio-oil from a lab-scale auger reactor.

172
Gas chromatography/Mass spectrometry (GC/MS). The chemical speciation among
fractions was found to be different, though SF1 and SF3 were very similar. SF4 was found to be the
most ch
ase temperature. This thermocouple configuration is also recommended for the
reactor
lt to
emically different from the other fractions, and mostly low molecular weight compounds
were identified in this fraction. The chemical composition of the whole bio-oil was found to vary little
as a function of operating conditions. Many of the quantified compounds were within the range of
values for typical pyrolysis oil.
Viscosity. The viscosity of the bio-oil samples was found to be related to the moisture
content of the sample, as expected. SF3 has the highest viscosity, followed by SF1 and SF2.
6.2 Recommendations for future work
As the auger reactor system for this project is a first generation design, there are several
recommendations to improve the performance and operation of the system. In general, the system can
be greatly improved by modifying the design for the heat carrier heating and feeding system as shown
in Figure 108. Rather than having a tall vertical pipe in which the heat carrier material is heated, a
horizontal design with the heating occurring in the metering auger section offers several benefits.
With the horizontal design, the mass feed rate of heat carrier material will be more constant over time,
and the heat transfer from the heaters to the steel shot will be increased (because of the agitation
offered by the metering auger). Whereas the vertical design had significant wall effects because the
inner heat carrier material never came into contact with the heated wall, the recommended design will
allow for bulk mixing of the heat carrier and thus more straightforward calibrations will be possible.
An additional nitrogen purge inlet at the junction where the heat carrier metering auger ends may help
to prevent back-flow of pyrolysis vapors into the heat carrier hopper system.
Additional temperature measurement locations are suggested to improve the understanding of
the heat transfer associated with the heat carrier material. To effectively monitor the temperature of
the heat carrier material as it enters the reactor, a thermocouple can be fitted such that it only
protrudes slightly past the “inner surface” of the metering pipe. This thermocouple can then measure
the heat carrier temperature as it falls into the reactor. If it is placed further into the pipe, it will only
measure the gas ph
outlet to measure the heat carrier exit temperature. By referring back to Equation 2, this
temperature is clearly an important value for understanding the thermodynamic behavior and heat
transfer mechanisms of the reactor. With the presence of rotating augers, however, it is difficu

173
accurate
the relatively low flow rate of nitrogen used in the operation of the
auger r
for the
successful operation of cyclone separators; however this becomes more difficult at the lab-scale given
the low volumetric flow rate available.
ly measure the exiting temperature of the solids (heat carrier and biochar). The current
configuration only measures the gas phase temperature at the reactor outlet.
Another design recommendation is a modified cyclone and perhaps two cyclones in series.
Though the current cyclone on the reactor was able to remove biochar such that the solids content in
the bio-oil was within the range of reported literature, it is believed a modified cyclone can improve
the collection efficiency. Given
eactor system, cyclone design for a lab-scale system is particularly difficult. For larger
systems, the flow of pyrolysis vapors would provide an adequate volumetric flow rate
Figure 108. Recommended system design modifications
The reliability of the system may be improved by upgrading the DC motor that drives the
augers in the reactor to a unit with additional power and torque. The current motor provides marginal
torque and sometimes had problems with material binding inside the reactor. Also, the motor
controllers for both the augers in the reactor and the heat carrier metering auger could be upgraded to
provide improved speed control.
The range of operating conditions may be extended by improving the design of the seal
between the reactor housing and main auger shaft. Moderate gas flow rates and pressures have the

174
potential to create leaks from the system, which causes pyrolysis vapors to escape at undesired
locations. This design improvement can also be applied to the seal where the shaft for the heat carrier
meterin
ntrol of the cooling water. The current configuration tends to
result in high wall temperatures near the vapor outlet, which may adversely effect the collection of
bio-oil. Also the heat transfer can be improved by soldering the cooling coils to the condenser wall.
In terms of continued research and testing, one recommendation is to immediately begin
testing and characterizing the heat carrier materials used. There is a high likelihood that commercial,
off-the-shelf heat carrier type materials (such as the ones used in this study) have catalytic properties
that may adversely affect product yields and composition. However this also suggests an opportunity
to deliberately introduce catalysts into the reactor, either combined with the heat carrier material or as
the heat carrier directly. There is also an opportunity to study the effects of different particle sizes and
shapes of heat carrier material, as well as biomass particle sizes.
g auger enters the system.
Another possible design modification or research topic is improvement of the vapor outlet
port configuration. As multiple outlet ports currently exist, they could potentially be connected into
one outlet tube that leads to the bio-oil recovery system. This would allow for more vapor products to
exit the reactor as they produced further downstream (in the axial direction) from the initial vapor
outlet port. This design change is also shown schematically in Figure 108.
In terms of the bio-oil recovery system, the condenser design could be improved by providing
more functionality in the temperature co

175
APPENDIX A. DESIGN AND DEVELOPMENT
Biomass inlet properties
Type: Cornstover
Mb 1
kg
hr
:= Mass flow rate
kg
ρb 225
m
3
:= Bulk density, measured
Qb
Mb
ρb
:= Volumetric feed rate Qb 74.074
cm
3
min
= Equation A1
Tb1 25 273.15
+
( )K
:= Initial temperature
Tb2 500 273.15
+
( )K
:= Final temperature
Cpb 2273
J
kg K
⋅
⋅
:= Specific heat capacity (mass basis)
Heat carrier inlet properties
Type: Sand
ρHC 1631.3
kg
m
3
:= Bulk density, measured
CpHC 815.2
J
kg K
⋅
⋅
:= Specific heat capacity (mass basis)
THCi 550 273.15
+
( )K
:= Initial temperature
THCf 450 273.15
+
( )K
:= Final temperature

176
Heat for pyrolysis analysis
Equation A6
QHC 201.341
cm
3
min
=
Heat carrier volumetric
feed rate
QHC
MHC
ρHC
:=
Equation A5
MHC 19.707
kg
hr
=
required to provide
heat for pyrolysis
MHC
QdotP
CpHC THCi THCf
−
( )
⋅
:=
Equation A4
Qrxn 0.527
MJ
kg
=
Reaction heat
required for pyrolysis
Qrxn QP Qsens
−
:=
Equation A3
Qsens 1.08
MJ
kg
=
Sensible heat input
required
Qsens Cpb Tb2 Tb1
−
( )
⋅
:=
Equation A2
QdotP 446.25W
=
Heat transfer rate
required for pyrolysis
QdotP QP Mb
⋅
:=
Heat required
for pyrolysis
QP 1.6065 10
6
×
J
:=
kg
Biochar properties analysis
Yc .18
:= Ch
bio
ar yield (%-wt., wet
mass basis)
ρc 400
kg
m
3
:= Char bulk density
Mc Yc Mb
⋅
:= Char mass flow rate Equation A7
Qc
Mc
ρc
:= Char volumetric
flow rate
Qc 7.5
cm
3
min
= Equation A8
Ycvol
Qc
Qb
:= Char yield (%-vol., wet
biomass basis)
Ycvol 0.101
= Equation A9

177
Vapor properties analysis
Qp 9.777
L
min
=
Equation A13
Product stream
volumetric flow rate
Qp
Mp
ρp
:=
Equation A12
ρp 1.398
kg
m
3
=
Product stream mass density,
assuming Ideal Gas Law
ρp
pp
Rp Tp
⋅
( )
:=
Equation A11
Rp 93.753
J
kg K
⋅
=
Product stream gas
constant
Rp
Rbar
MWp
:=
Universal gas
constant
Rbar 8314
J
kmol K
⋅
:=
Product stream
pressure (atmospheric)
pp 101325
Pa
:=
MWp 88.68
kg
kmol
:=
Product streamt
temperature
Tp 500 273.15
+
( )K
:=
Molecular
weight of
products
Equation A10
Mp 0.82
kg
hr
=
Mass flow rate of pyrolysis
products without char. Includes
bio-oil vapors,aerosols, and NCG
Mp Mb Mc
−
( )
:=
Reactor fill specifications analysis
Equation A18
τp 0.621
=
Volume percent fill of
pyrolysis vapor products
τp 1 τHC
− τc
−
:=
Equation A17
τc 0.014
=
Volume percent fill of char, final
τc Ycvol τb
⋅
:=
Equation A16
τN2 0.5
=
Volume percent fill of nitrogen
(or excavated space), Initial
τN2 1 τfeed
−
:=
Equation A15
τb 0.134
=
Volume percent fill of
biomass, Initial
τb τfeed τHC
−
:=
Equation A14
τHC 0.366
=
Volume percent fill of heat carrier
τHC τfeed
QHC
QHC Qb
+
⋅
:=
Volumetric percent fill of
biomass and heat carrier
(common for screw conveyors)
τfeed 0.5
:=

178
Reactor cross-sectional area requirement analysis
Equation A23
AcsREQ 5.397cm
2
=
otal required cross
sectional area for biomass
and heat carrier.
τfeed
T
AcsREQ
Acsb AcsHC
+






Material factor, assumed
to allow for mixing volume
FM 1.4
:=
Equation A22
AcsHC 1.409cm
2
=
Required cross sectional
area for heat carrier
AcsHC
MHC
ρHC vHC
⋅
FM
⋅
:=
:=
Equation A21
Acsb 0.518cm
2
=
Required cross sectional
area for biomass
Acsb
Mb
ρb vb
⋅
:=
Equation A20
Heat carrier initial linear
velocity (superficial)
vHC vb
:=
Equation A19
vb 2.381
cm
s
=
Biomass initial linear
velocity (superficial)
vb na Pa
⋅
:=
Auger rotation frequency
na
Na
60
1
s
⋅
:=
Auger rotation speed (RPM)
Na 45
:=
Auger pitch
Pa 1.25in
=
Reactor dimension specifications
w 1.75 in
⋅
:= Equivalent reactor width (for rectangular cross section)
h 1.396 in
⋅
:= Equivalent reactor height (for rectangular cross section)
da 1in
:= Auger outer diameter (#16 auger)
ds .3125in
:= Axle shaft diameter (5/16 in.)
Pa 1.25in
:= Auger pitch

179
Reactor vapor residence time analysis
trv Lr
( )
0.513
0.671
0.829
0.987
1.145














s
=
Vapor residence
time
trv Lr
( )
ρv Av
⋅ Lr
⋅
Mv






:=
Equation A27
Lr
6.5in
8.5in
10.5in
12.5in
14.5in














:=
Reactor length, biomass
inlet to vapor exit ports
(1 - 5)
Equation A26
vv 32.161
cm
s
=
Vapor velocity
vv
Mv
ρv Av
⋅
:=
Equation A25
Av 5.067cm
2
=
Cross sectional area
for vapor to occupy
Av Acs τv
⋅
:=
Equation A24
Acs 8.161cm
2
=
Average cross sectional
area for reactants and
products to occupy
Acs w h
⋅
( ) 1.5
π
4






Ds
2
⋅






−
:=
Downstream and total vapor residence time analysis
Equation A32
ttotal Lr
( )
0.741
0.899
1.057
1.215
1.373














s
=
Total resience time: Biomass
inlet to condenser inlet
ttotal Lr
( ) trp Lr
( ) te
+
:=
Equation A31
te 0.228s
=
Residence time from reactor
outlet to condenser inlet
te
Lc
ve
:=
Exiting vapor velocity
Equation A30
ve 2.227
m
s
=
ve
Mp
ρp Ae
⋅






:=
Total tube length from
reactor outlet to condenser
inlet, assumed initially
Lc 20in
:=
Equation A29
Ae 0.732cm
2
=
Cross sectional area
for exit port and tube
Ae
π
4






de
2
⋅
:=
Equation A28
Vapor exit port and
tube diameter
de 0.380in
:=
(1/2" OD)

180
Reactor heat carrier residence time analysis
Lexit 12in
:= Length from heat carrier
inlet to solids exit
tHC
Lexit
vHC
:= Heat carrier residence
time in reactor tHC 12.8s
= Equation A33
0
5
10
15
20
25
30
35
40
45
50
0 20 40 60 80 100 120 140 160 180
Screw speed (RPM)
Heat
carrier
residence
time
(s)
Typical operating regime
Figure 109. Heat carrier residence time as a function of auger speed

181
Table 45. Motor power requirements analysis
Reactor
augers
Heat carrier
metering
auger
ρ kg/m3
Bulk density 1563 1631
Weighted average of biomass and
heat carrier for reactor augers.
Sand bulk density for heat
carrier metering auger
kg/hr Mass feed rate 25 24
Maximum feed rate
(biomass and heat carrier)
Q m
3
/hr Volumetric feed rate 0.016 0.015 Q = /ρ
C ft
3
/hr Volumetric feed rate 0.56 0.52 C = Q x 35.314
e - Drive efficiency 0.50 0.75
Overall mechanical efficiency.
Estimated low to be conservative
Fb - Hanger bearing factor 4.4 4.4
Fb = 4.4 for Group D
hard surfaced bearings
Fd - Conveyor diameter factor 13.57 13.34
Fd = .508x
2
- 2.89x + 15.95
(x = Auger diameter, inches)
Ff - Flight factor 1.0 1.0 Ff = 1.0 for standard helicoid screws
Fm - Material factor 3.0 3.0
Fm = 3.0 for class III materials
(abrasive, poor flowing, etc.)
Fo - Overload factor 3.0 3.0 Fo = 3.0 max for small motors
Fp - Paddle factor 1.0 1.0 Fp = 1.0 for no paddles
L ft Length of conveyor 1.83 1.00 As designed
H ft Lift 0.0 0.0 H = 0 for no lift (horizontal conveying)
N RPM Operating speed 180 60 N = max speed to be conservative
W lbs/ft3
Bulk density 97.6 101.8 W = ρ x 0.06243
Pf HP
Power required to overcome
converyor friction
0.0197 0.0035
Pf = (L x N x Fd x Fb) /
1000000
Pl HP
Power required to
lift the material
0.000 0.000
Pl = (0.5 x C x W x H) /
1000000
Pm HP
Power required to transport Pm = (C x L x W x Ff x Fp x Fm) /
PT HP Total power requirement 0.120 0.015 PT = [(Pf + Pl + Pm) x Fo] / e
Notes
Symbol Units Description
lue
Va
material at specified rate
0.0003 0.0002
1000000
m

m


182
182
Figure 110. Biomass feeding system
Figure 110. Biomass feeding system
Feed direction
Figure 111. Close-up of reactor augers
Feed direction
Figure 111. Close-up of reactor augers
Figure 112. Reactor mounted on frame
Vapors Heat carrier
Figure 112. Reactor mounted on frame
Solids
Biomass

183
Flow
Figure 113. Reactor lid and thermocouple detail
Figure 114. Gas cyclone
Fig 2)
ure 115. Condensers 1 and 2 (SF1 and SF
Flow
Thermocouple
SF2
outlet
SF1 inlet
SF1
oil
SF2
l
oi

184
SF3
outlet
SF3
inlet
Figure 116. Electrostatic precipitator (SF3)
SF3 oil
SF4 inlet
SF4 outlet
Figure 117. Condenser 3 in ice bath (SF4)

185
Figure 118. Reactor system detail
Heat carrier
Biomass
Cyclone
SF1
inlet
Vapors
Vapors
Biochar

186
Table 46. Shakedown trial operating conditions
Shake-
down
trial
Trial
date Type Type
1 5.13.08
Corn
stover
1.00 1.00 Sand 24.0 ~650 1.0 45
2 5.16.08
Corn
stover
1.00 1.00 Sand 24.0 ~775 1.5 45
3 5.18.08
Corn
stover
1.00 0.50 Sand 12.0 650 2.0 38
4 6.03.08
Corn
stover
0.50 1.00 Sand 16.5 ~700 3.0 40
5 6.10.08
C
stover
0.50 1.00 Steel shot 12.5 650 2.0 38
6 .08
Corn
stover
0.50 0.50 Steel shot 22.0 1.5 40
7 6.17.08
Co
stover
0 0.75 Sand 16.5 675 2.5 40
8 7.09.08
Corn
stover
0.50 0.50
Silicon
carbide
12.5 500 2.0 40
9 7.14.08
Wood
chips
0.79 0.50
Silicon
ca
12.5 450 2.0 40
10 7.23.08
Wood
chi
0.50
Silicon
carbide
12.5 500 2.0 40
11 8.22.08
Corn
stover
0.75 1.00 Steel shot 22.0 450 2.0 40
12 8.27.08
Corn
stover
0.75 1.00
Steel shot/
Al ceramic
20.0 460 2.5 40
13 9.03.08
Corn
stover
0.75 1.00 Steel shot 20.0 425 2.5
14 9.10.08
Corn
fiber
1.00 1.00 Steel shot 20.0 475 2.5 40
15 9.16.08
Corn
fiber
1.00 1.00 Steel shot 20.0 525 4.0 50
16 9.26.08
Wood
chips
550 2.5 45
17 10.02.08
Wood
chips
1.00 1.00 Steel shot 20.0 550 2.5 45
18 10.07.08
Wood
chips
1.00 1.00 Steel shot 20.0 550 2.5 45
19 10.17.08
Red
Oak
0.75 1.00 Steel shot 15.0 550 2.5 45
Auger
speed
(RPM)
N2
volume
flow
rate
(SLPM)
Heat
carrier
heater
temp.
(°C)
Biomass Heat carrier
Particle
size
(mm)
Nominal
feed rate
(kg/hr)
Nominal
feed rate
(kg/hr)
orn
6.13 650
rn
0.5
rbide
ps
0.79
40
1.00 1.00 Steel shot 20.0

187
Table 4 results
7. Shakedown trial yield and operating condition
Shake-
down
trial Biomass
Heat
carrier Bio-oil Biochar NCG
1 0.99 nd 32.5 22.5 45.0
2 nd nd nd nd nd
3 0.49 nd 43.4 nd nd
4 1.10 4.49 39.9 31.7 28.4
5 1.18 4.39 35.7 31.3 33.0
6 0.57 21.15 33.6 nd nd
7 0.73 15.62 39.1 nd nd
8 0.59 13.21 40.9 42.9 16.2
9 0.54 14.14 23.2 23.9 52.9
10 nd nd 43.3 37.6 19.1
11 1.13 nd 24.9 60.1 15.0
12 nd nd 38.5 42.7 18.7
13 1.20 15.06 23.1 58.2 18.7
14 1.02 19.05 54.5 26.3 19.2
15 1.03 16.66 56.4 26.5 17.1
16 0.89 nd 61.6 21.4 17.0
17 0.99 nd 60.9 28.8 10.4
18 1.03 22.76 62.8 26.1 13.0
19 1.03 16.55 70.8 15.9 13.0
Notes:
nd - Not determined. wb - Wet basis
NCG yields are by difference except Nos. 18 and 19
Product yields
(%-wt., wb)
Feed rate
(kg/hr)

188
APPENDIX B. MIXING STUDY
A series of four tests was performed to collect 22 samples of biomass-sand mixtures for
particle density measurements using a gas pycnometer as shown in Figure 123. Each test and
subsequent analysis was repeated once. The goal of characterizing these mixtures was to determine
the optimal operating speed of the augers, with respect to maximizing the “degree of mixing”.
Initially 12 samples of sand and biomass were manually mixed at various mass fractions to
develop a “calibration curve” for comparison to collected samples. The biomass tested was corn
stover, ground to 1.0 mm particles using a Retsch SM 200 knife mill, and the sand was No. 35 to 45
mesh. The baseline data plot, as seen in Figure 119, shows that the particle density of the biomass
tested was 1.53 g/cm3
(100% biomass – right side), and the sand has a particle density of 2.65 g/cm3
(0% biomass – left side). Samples were approximately all the same in volume, with masses ranging
from 4g to 40g depending on the composition. Each of the samples was analyzed three times, and the
raw data is presented in Table 48.
y = -1.12x + 2.60
R
2
= 0.99
0.00
0.25
0.50
0.75
1.00
1.25
1.50
1.75
2.00
2.25
2.50
2.75
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Biomass mass fraction
Density
(g/cm
3
)
100% sand
particle density
100% biomass
particle density
Sand
bulk density
Linear particle density
Biomass
bulk density
Figure 119. Biomass and sand mixture densities
In comparing the mixture densities, the variables were: the axial position and radial positions
of the sample and the auger speed. Four screw speeds were selected based on visual tests of feeding
solid mixtures at various rates: N = 45, 50.4, 63 and 90 RPM. The lowest speed case represents the
approximate lower operating end, and highest speed case represents the approximate higher operating
end. It is especially important to note that these speeds seem suitable for cold flow feeding of this
particular set of feed rates (the original design case: 1 kg/hr biomass and 24 kg/hr sand). At auger
speeds less than 25% of the full speed (45 RPM), the auger flights become “full” of solid material and

189
a potential exists for many operational problems such as auger binding. This is noted to be a
‘conservative’ low end spee ible. At speeds above 50%
of the maximum, the augers convey minimal mixing.
Also, at these higher speed conditions, solid material only fills the bottom portion of the mixer and no
material exist
that a
fifth va
ents show that material does travel between both augers, and
several phenomena have been observed. As expected, fine biomass particles readily segregate by
“falling” to the bottom, while other low density biomass particles “float” on the top of the mixture.
Also of interest is the “pulsing” effect characteristic of the screw feeders: small chunks of biomass
and sand are fed in during each successive auger flight.
The biomass volumetric screw feeder and the heat carrier feeding system were independently
calibrated, and the mass feed rates were found to be linearly proportional to screw speeds. The mixer
was cleaned out, and biomass and sand feeding was begun (1.0 and 24.0 kg/hr, respectively) as the
main augers were started at a specified speed. After a short time period required before reaching an
apparent “steady state” condition, samples were taken from a moving stream of solid material exiting
at the end of the mixer. The 3 motors were then shut down simultaneously and the lid removed.
Samples were taken at each position by scooping out material with a small spatula, and placed in
plastic containers. This procedure was repeated for each of the four auger speeds, and then duplicated
once. The particle densities of the mixtures were then analyzed with a Pentapycnometer from
Quantachrome Instruments. Each sample was analyzed a minimum of three times, resulting in
standard deviations less than ± 0.8%. The collected data is presented in Table 49.
Based on the particle density results, the data was analyzed and several plots were
constructed to reveal any trends between operating conditions and mixture density. Figure 120 shows
the mixture density (sampled from the left radial position) resulting from the experiments compared
to the “expected density” based on the calibration curve as discussed. There are no clear trends
observed between mixture density and speed or position, except that the density measurements e
slightly most consistent for t rom the heat carrier
entrance). Data points above the expected density higher than
expected, meaning more sand was sampled. It is un lear whether this indicates less mixing was
achieve
d, as speeds as low as 20% (36 RPM) are poss
the material so quickly that there appears to be
s for easy sampling.
The four axial sampling positions correspond to four of the vapor outlet ports, where the
distance represents the length from the center of the heat carrier inlet (X = 4.25, 6.25, 8.25, 10.25 in.).
The fourth position at 10.25 inches is at the mixer exit (the entrance to the solids catch). Note
por port exists at the end of the reactor, at the end of the solids exit.
Samples were taken from two radial positions: the center, C (in-between the augers), and the
left, L (facing auger motor: left edge in-between auger and mixer wall) of the mixer. For higher
speeds, material is continuously moved to the left auger and it was difficult to obtain a sample from
the right side. At these conditions it is not to say that mixing doesn’t occur, though, as there are
mixing processes at the bottom of the screw and as material is conveyed from one auger to the next.
ualitative, visual experim
Q
ar
he 8.25 in position (longest mixing time/furthest f
line indicate a measured density
c
d, or whether the fine biomass particles were able to be sampled from the top of the augers
(due to the segregation and settling effect mentioned previously). Conversely, points below the
expected density line indicate lower density, meaning more biomass was sampled than expected.

190
2.7
2.9
3.1
(g/cm
3
)
1.5
1.7
1.9
Mixture
m
2.1
2.3
2.5
ass
densit
40 50 60 70 80 90 100
Augers rotational speed (RPM)
y
8.25 in
6.25 in
4.25 in
Expected
Axial distance
Figure 120. Mixture density (L) vs. auger speed at three axial locations, Run 1
The duplicated test, as shown in Figure 121, found that each sample was denser than
expected. Notable differences in density with respect to auger speed were not found.
2.7
2.8
2.9
ity
(g/cm
3
)
2.3
2.4
2.5
2.6
40 50 60 70 80 90 100
Mixture
mass
dens
8.25 in
6.25 in
4.25 in
Expected
Axial distance
Figure 121. Mixture density (L) vs. auger speed at three axial locations, Run 2

191
When comparing the densities obtained from the center of the mixer, an interesting result is
shown in Figure 122. Noting the exit mixture (axial distance of 10.25 in), the density was extremely
consistent and not a function of speed. Though this indicates an intuitive result (the materials that
enter the mixer are the same materials that exit the mixer), it also speaks to the preferred method of
sampling from a moving stream. Note that at the higher speeds, material was only sampled at the end
of the mixer, and these results slightly validate the overall procedure.
1.2
1.5
M
1.8
2.4
2.7
3.0
3.3
3.6
3.9
4.2
4.5
4.8
5.1
40 50 60 70 80 90 100
ix
e
mass
density
(g/cm
3
)
10.25 in
8.25 in
6.25 in
4.25 in
2.1
tur
Expected
Axial distance
Figure 122. Mixture density (C) vs. auger speed at four axial locations, Run 1
Figure 124 shows the results from the material sampled from in-between the augers (the
center position) on the duplicated run. Similar results were observed in regards to the consistency
sampled from the end of the mixer, and the general result of the sampled densities being higher than
expected.
Figure 123. Pentapycnometer instrument

192
2.4
2.5
2.6
2.7
2.8
2.9
3.0
3.1
3.2
nsity
(g/cm
3
)
40 50 60 70 80 90 100
Mixture
mass
de
8.25 in
6.25 in
4.25 in
10.25 in
Axial distance
Expected
Figure 124. Mixture density (C) vs. auger speed at four axial locations, Run 2
These results indicate that, for the materials tested at their respective feed rates, there are no
clear trends for mixer performance as a function of auger speed or position. The method of using a
gas Pycnometer for characterizing density to predict mixer performance, though unique, poses some
challenges for this specific system (mainly in sampling).
Table 48. Baseline biomass and sand mixture densities analytical data
18
Biom
mass
fraction
0.0 0.05 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
2.6530 2.5705 2.5008 2.3648 2.2184 2.0974 2.0440 1.8847 1.8593 1.7257 1.5757 1.5337
2.6547 2.5723 2.5080 2.3566 2.2138 2.0806 2.0508 1.8912 1.8117 1.7270 1.5665 1.5303
2.6533 2.5723 2.5159 2.3532 2.2133 2.0906 2.0558 1.8948 1.8105 1.7266 1.5702 1.5403
- - - - - - - - - - - 1.53
- - - - - - - - - - - 1.5349
Averge 2.6537 2.5717 2.5082 2.3582 2.2152 2.0895 2.0502 1.8902 1.8272 1.7265 1.5708 1.5342
St. Dev. 0.0009 0.0010 0.0075 0.0060 0.0028 0.0085 0.0059 0.0052 0.0278 0.0006 0.0046 0.0038
Mass
density
(g/cm3
)
ass
Sand
mass
fraction
1.0 0.95 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0

193
Table 49. Biomass and sand mixture densities analytical data
45 50.4 63 90 45 50.4 63 90
(25) (28) (35) (50) (25) (28) (35) (50)
2.9691 1.7570 2.6741 2.7341 2.5885 2.5883 2.5815 2.8023
2.9093 1.7653 2.6675 2.7865 2.5861 2.6084 2.5976 2.8754
2.8816 1.7395 2.7020 2.7798 2.5790 2.5825 2.6370 2.8563
2.9033 1.7285 2.7118 2.7967 2.5846 2.5982 2.6100 2.8533
Avg. 2.9158 1.7476 2.6888 2.7743 2.5846 2.5944 2.6065 2.8468
St.Dev. 0.0375 0.0167 0.0214 0.0277 0.0040 0.0114 0.0235 0.0312
2.9107 1.9375 - - 2.7205 2.7733 - -
2.9739 2.1280 - - 2.8401 2.8578 - -
2.9853 2.0304 - - 2.7812 2.8398 - -
2.8810 1.9703 - - 2.8175 2.8616 - -
Avg. 2.9377 2.0166 - - 2.7898 2.8331 - -
St.Dev. 0.0501 0.0836 - - 0.0522 0.0410 - -
2.6883 2.3722 2.5358 2.8656 2.7792 2.6612 2.6712 2.6977
2.7152 2.3638 2.5079 2.9306 2.8711 2.7090 2.6855 2.7486
2.7406 2.3964 2.5216 3.0488 2.8442 2.7049 2.7058 2.7424
Avg. 2.7286 2.3782 2.5300 2.9351 2.8336 2.6942 2.6883 2.7300
St.Dev. 0.0350 0.0139 0.0200 0.0803 0.0388 0.0222 0.0143 0.0227
- 2.5297 3.1008 - -
Avg. 2.4827 1.7045 - - 2.5431 3.0178 - -
St.Dev. 0.0353 0.3011 - - 0.0447 0.0960 - -
52
2.8124 2.6924 2.8275 2.5549 2.7815 2.6003 2.7016 2.7439
2.7876 2.7129 2.8293 2.5180 2.7868 2.6063 2.7102 2.7898
- - - - 2.7846 2.6047 2.7226 2.7616
Avg. 2.7904 2.7273 2.8147 2.5302 2.7786 2.6002 2.7054 2.7576
St.Dev. 0.0207 0.0440 0.0238 0.0213 0.0116 0.0076 0.0148 0.0241
2.5706 4.7228 - - 2.5457 2.6418 - -
2.5484 4.1649 - - 2.5594 2.7418 - -
2.5744 4.7576 - - 2.5916 2.6988 - -
2.5596 - - - 2.5776 2.7620 - -
Avg. 2.5632 4.5484 - - 2.5686 2.7111 - -
St.Dev. 0.0117 0.3326 - - 0.0202 0.0532 - -
2.6096 2.6280 2.6119 2.6183 2.5751 2.6205 2.6130 2.6073
2.6102 2.6319 2.6092 2.6107 2.5797 2.6229 2.6162 2.6131
2.6140 2.6246 2.6103 2.6080 2.5785 2.6282 2.6164 2.6091
- - - - 2.5764 2.6253 2.6144 2.6096
- - - - 2.5783 2.5825 2.5754 2.5361
- - - - 2.5805 2.5835 2.5786 2.5352
- - - - 2.5824 2.5863 2.5771 2.5361
- - - - 2.5794 2.5853 2.5761 2.5362
Avg. 2.6113 2.6282 2.6105 2.6123 2.5788 2.6043 2.5959 2.5728
St.Dev. 0.0024 0.0037 0.0013 0.0053 0.0023 0.0214 0.0205 0.0396
Left
Screw speed [RPM, (% of max)]
Center
Axial
position
(in)
Radial
position
8.25
1st Run 2nd Run
10.25
Center
Left
Center
Left
4.25
Screw speed [RPM, (% of max)]
2.7702 2.3806 2.5545 2.8956 2.8398 2.7016 2.6909 2.7315
2.4597 1.9715 - - 2.5302 2.8792 - -
2.4838 1.2722 - - 2.5046 3.0437 - -
2.4551 1.7763 - - 2.6077 3.0476 - -
2.5321 1.7980 -
6.25
Center
2.7713 2.7767 2.7873 2.5178 2.7615 2.5895 2.6874 2.73

194
APPENDIX C ENTS
. AUXILLARY EQUIPMENT AND INSTRUM
Figure 125. Hammer mill
Figure 126. Knife mill

195
Figure 127. CHN/O/S analyzers
Figure 128. Thermal gravimetric analyzer

196
Figure 129. Bomb calorimeter
Figure 130. Moisture analyzer

197
Figure 131. Micro-GC cart
Figure 13 re gauge
2. Gas volume meter and pressu

198
Figure 133. Moisture titrator
Figure 134. Total acid number titrator

199
Figure 135. GC/MS
Figure 136. Viscometer

200
APPENDIX D. EXPERIMENTAL DATA
Table 50. Feedstock and experimental condition data
Run
No.
DOE
No.
Run
Date
(2009)
Moisture
content
(%-wt.)
Mass fed
(g)
Feed time
(min)
Feed rate
(kg/hr)
Mass fed
(g)
Feed time
(min)
Feed rate
(kg/hr)
Average
temperature,
THC (°C)
1 17 9-Feb 5.83 1000.4 60.1 0.999 16984 108.3 9.4 523.1
2 18 24-Feb 5.77 721.7 45.1 0.959 21948 60.2 21.9 529.3
3 15 10-Mar 5.55 765.2 48.5 0.946 20517 65.1 18.9 479.8
4 7 11-Mar 5.63 792.3 49.1 0.969 22321 71.5 18.7 586.3
5 9 13-Mar 5.83 855.4 51.0 1.007 23804 73.9 19.3 480.6
6 13 16-Mar 5.56 894.6 50.2 1.069 23434 72.1 19.5 482.2
7 11 18-Mar 5.70 940.3 53.4 1.056 23401 70.2 20.0 478.3
8 5 20-Mar 5.97 462.5 27.6 1.005 16340 49.0 20.0 586.9
9 1 30-Mar 5.86 931.4 55.1 1.015 24445 78.1 18.8 576.5
10 3 1-Apr 5.95 902.8 52.8 1.027 25349 81.1 18.8 577.8
11 21 11-Apr 5.93 860.9 53.7 0.962 21983 91.3 14.4 527.6
12 28 14-Apr 5.88 910.4 55.7 0.982 23361 94.9 14.8 528.8
13 23 21-Apr 5.64 994.8 57.9 1.030 22816 90.3 15.2 427.8
14 19 22-Apr 6.11 788.0 47.9 0.987 21002 85.5 14.7 527.1
15 29 24-Apr 6.01 867.5 53.5 0.973 22796 90.0 15.2 536.5
16 20 28-Apr 6.02 789.5 50.4 0.941 22740 88.4 15.4 526.7
17 27 30-Apr 6.04 882.8 53.9 0.983 22806 93.5 14.6 527.7
18 22 3-May 5.98 934.9 54.3 1.034 23692 92.0 15.4 526.2
19 25 4-May 5.94 925.1 55.9 0.994 22987 90.5 15.2 529.5
20 24 5-May 5.93 1026.7 59.8 1.031 24873 106.1 14.1 630.5
21 26 6-May 5.64 964.1 60.0 0.964 23197 94.6 14.7 538.7
22 30 7-May 5.72 919.0 56.9 0.969 22443 91.0 14.8 535.6
23 8 16-May 5.81 959.3 58.3 0.987 21168 114.6 11.1 576.1
24 2 18-May 5.91 922.4 56.5 0.980 23149 107.8 12.9 582.7
25 12 21-May 6.14 896.5 55.3 0.972 24979 112.8 13.3 481.3
26 6 22-May 6.24 833.5 50.1 0.998 19911 108.1 11.1 571.5
27 10 24-May 6.29 926.3 55.0 1.011 23225 108.6 12.8 475.9
28 14 31-May 5.66 1059.3 57.9 1.098 22993 114.0 12.1 475.3
29 4 2-Jun 5.05 1005.8 54.1 1.115 21001 109.8 11.5 576.0
30 16 4-Jun 5.56 995.6 57.9 1.032 24544 108.4 13.6 477.8
5.84 891.0 53.3 1.00 - - - -
0.247 114.4 6.2 0.042 - - - -
5.87 911.5 56.0 0.977 22931 92.4 14.9 532.8
0.161 34.0 2.4 0.011 326 2.2 0.3 4.7
6.29 1059.3 60.1 1.115 25349 114.6 21.9 630.5
5.05 462.5 27.6 0.941 16340 49.0 9.4 427.8
Heat carrier
Cntr. Pt. Avg.
Biomass
Cntr. Pt. St. Dev.
MAX
MIN
Overall Avg.
Overall St. Dev.

201
Table 51. Product distribution and mass balance data
TOTAL
Run
No.
DOE
No.
Mass
(g) (%-wt., wb) (%-wt., db) MC (g) (%-wt., wb) MCNG (g) (%-w
.52 59.23 56.70 299.59 29.95 107.89 10
Yield Yield Mass Yield Mass, Yield
t., wb) (%-wt., wb)
1 17 592 .78 99.96
2 18 502.32 69.60 67.74 129.29 17.91 88.87 12.31 99.83
3 15 460.88 60.23 57.89 209.57 27.39 84.68 11.07 98.68
4 7 565.35 71.35 69.64 119.95 15.14 99.87 12.60 99.10
5 9 506.71 59.23 56.71 240.25 28.09 92.47 10.81 98.13
6 13 506.67 56.64 54.08 262.42 29.33 92.50 10.34 96.31
7 11 572.22 60.86 58.49 250.29 26.62 102.66 10.92 98.39
8 5 334.15 72.26 70.50 64.96 14.05 56.05 12.12 98.42
9 1 675.87 72.57 70.86 107.28 11.52 117.14 12.58 96.67
10 3 654.66 72.51 70.77 124.61 13.80 110.36 12.22 98.54
11 21 563.16 65.41 63.23 184.85 21.47 97.70 11.35 98.23
12 28 604.66 66.42 64.32 211.61 23.24 101.43 11.14 100.80
13
a
23 419.59 42.18 38.72 354.99 35.68 - 22.14 100.00
14 19 530.46 67.32 65.19 136.63 17.34 91.03 11.55 96.21
15 29 586.62 67.63 65.55 173.28 19.98 100.75 11.61 99.22
16 20 528.55 66.95 64.83 161.87 20.50 87.12 11.03 98.48
17 27 586.81 66.47 64.31 180.91 20.49 99.54 11.28 98.24
18 22 622.79 66.61 64.49 184.79 19.77 106.27 11.37 97.75
19 25 615.26 66.51 64.39 190.12 20.55 105.12 11.36 98.42
20 24 755.73 73.61 71.95 113.12 11.02 132.22 12.88 97.51
21 26 662.43 68.71 66.84 182.32 18.91 110.01 11.41 99.03
22 30 629.89 68.54 66.63 165.26 17.98 103.64 11.28 97.80
23 8 653.96 68.17 66.18 181.16 18.88 110.86 11.56 98.61
24 2 654.72 70.98 69.16 139.76 15.15 106.98 11.60 97.73
25 12 520.10 58.02 55.27 262.51 29.28 92.93 10.37 97.66
26 6 574.80 68.96 66.90 150.71 18.08 92.80 11.13 98.18
27 10 483.07 52.15 48.94 333.78 36.03 87.63 9.46 97.64
28 14 529.57 49.99 46.99 409.88 38.69 96.55 9.11 97.80
29 4 691.34 68.74 67.08 178.23 17.72 115.44 11.48 97.94
30 16 544.83 54.73 52.06 344.27 34.58 108.02 10.85 100.16
- - - - - - - 98.38
- - - - - - - 1.08
- 67.38 65.34 - 20.19 - 11.35 98.92
- 1.07 1.18 - 1.79 - 0.16 1.06
755.73 73.61 71.95 409.88 38.69 132.22 22.14 100.80
334.15 42.18 38.72 64.96 11.02 56.05 9.11 96.21
Note: a - NCG yield for Run 13 was calculated by difference.
Cntr. Pt. Avg.
Cntr. Pt. St. Dev.
MAX
MIN
NCG
Overall Avg.
Overall St. Dev.
Biochar
Bio-oil

202
Table 52. He conditions
at carrier system temperature data and other operating
Run
No.
DOE
No.
Nominal
heat carrier
temperature (°C)
No. data
pointsa
Average reactor
pressure
(in-H2Og) PH HC1 HC2 HC3
Average biomass
inlet temperature
(°C)
1 17 525 1306 1.71 425.4 529.7 523.1 340.1 41.3
2 18 525 855 1.26 147.4 479.2 529.3 351.8 40.8
3 15 475 904 1.14 158.5 434.1 479.8 329.7 42.4
4 7 575 993 1.67 238.2 551.4 586.3 377.7 42.6
5 9 475 1324 1.60 279.8 432.0 480.6 338.4 39.0
6 13 475 1076 1.35 224.7 425.5 482.2 338.2 40.2
7 11 475 1296 1.58 250.3 420.1 478.3 334.5 37.4
8 5 575 494 1.52 184.4 514.5 586.9 384.5 43.1
9 1 575 1275 1.91 390.6 507.6 576.5 385.4 36.9
10 3 575 1199 2.55 354.4 490.2 577.8 388.2 38.9
11 21 525 1379 3.06 361.4 504.5 527.6 349.6 44.0
12 28 525 1411 2.99 401.8 505.8 528.8 352.3 37.9
13 23 425 1525 2.40 336.9 434.5 427.8 297.7 36.5
14 19 525 1183 2.94 410.3 502.4 527.1 350.4 37.6
15 29 525 1192 2.17 449.3 518.8 536.5 351.6 38.3
16 20 525 1306 1.79 433.1 504.6 526.7 358.8 39.9
17 27 525 1284 2.03 442.6 505.4 527.7 359.4 39.5
18 22 525 1323 2.46 435.5 500.0 526.2 362.1 38.8
19 25 525 1404 2.19 435.2 507.3 529.5 360.5 39.6
20 24 625 1575 2.30 466.5 587.0 630.5 409.6 40.6
21 26 525 1298 1.39 417.5 502.2 538.7 359.8 38.5
22 30 525 1408 2.11 386.3 497.3 535.6 356.0 39.0
23 8 575 1493 2.19 521.9 597.8 576.1 373.7 41.8
24 2 575 1485 1.99 502.5 583.1 582.7 379.8 38.0
25 12 475 1365 2.09 427.9 492.6 481.3 334.0 36.9
26 6 575 1306 1.50 460.0 604.5 571.5 370.8 40.0
27 10 475 1466 2.68 411.3 493.9 475.9 330.8 36.8
28 14 475 1531 1.05 416.5 505.0 475.3 329.2 38.0
29 4 575 1278 3.00 448.6 600.2 576.0 373.5 39.1
30 16 475 1306 2.56 439.1 502.8 477.8 327.9 37.9
1275 2.04 - - - - 39.4
227 0.58 - - - - 2.0
1333 2.15 422.1 506.1 532.8 356.6 38.8
90 0.51 24.7 7.2 4.7 3.9 0.7
1575 3.06 521.9 604.5 630.5 409.6 44.0
494 1.05 147.4 420.1 427.8 297.7 36.5
Note: a - Number of data points collected for steady state operation. Data collection rate = 0.5 Hz
Heat carrier system
average temperatures (°C)
Cntr. Pt. Avg.
Cntr. Pt. St. Dev.
Overall Avg.
Overall St. Dev.
MAX
MIN

203
Table 53. Reactor system temperature data
Run
No.
DOE
No.
Nominal
heat carrier
temperature (°C)
R1 R2 R3 R4 R5
Average solids
outlet temperature
(°C)
1 17 525 427.5 486.2 478.8 434.3 344.9 237.1
2 18 525 443.0 496.0 486.4 438.2 341.3 238.8
3 15 475 427.4 484.9 476.6 432.0 339.9 230.6
4 7 575 455.9 502.8 492.2 443.4 346.1 247.1
5 9 475 432.1 486.6 477.5 430.6 340.1 233.6
6 13 475 433.2 486.2 477.0 430.4 339.6 232.1
7 11 475 425.4 483.3 475.0 429.9 338.6 230.1
8 5 575 457.4 507.3 498.3 449.6 353.0 245.6
9 1 575 452.8 504.6 497.0 449.5 354.9 251.7
10 3 575 452.8 503.8 497.6 451.6 357.2 258.5
11 21 525 439.5 496.1 490.1 446.0 355.2 243.7
12 28 525 438.2 495.3 489.7 445.5 355.1 244.4
13 23 425 417.1 479.0 476.4 435.2 349.0 229.7
14 19 525 432.2 493.6 489.4 446.4 354.8 240.0
15 29 525 439.2 496.8 490.7 446.1 355.0 242.9
16 20 525 441.8 499.5 491.5 444.5 352.8 239.7
17 27 525 435.4 495.1 489.5 445.1 354.2 241.0
18 22 525 431.4 493.8 489.2 444.7 353.8 240.2
19 25 525 437.0 495.5 489.6 445.1 354.0 240.6
20 24 625 458.6 509.6 502.0 454.5 360.4 257.7
21 26 525 435.3 495.7 490.4 445.7 354.0 242.1
22 30 525 434.3 496.1 490.8 446.3 354.8 243.1
23 8 575 438.1 497.6 493.4 450.0 359.7 250.0
24 2 575 443.5 501.8 496.3 450.3 358.5 253.7
25 12 475 423.6 487.9 485.3 444.0 356.2 239.6
26 6 575 438.4 499.4 493.8 448.5 357.8 245.0
27 10 475 423.4 486.5 482.2 439.1 352.7 237.6
28 14 475 421.8 484.4 480.7 438.7 353.4 237.0
29 4 575 433.6 494.8 491.1 447.7 357.8 245.1
30 16 475 419.9 484.1 482.5 442.9 357.4 240.6
436.6 495.7 490.1 445.6 354.5 242.3
1.9 0.6 0.6 0.5 0.5 1.4
458.6 509.6 502.0 454.5 360.4 258.5
417.1 479.0 475.0 429.9 338.6 229.7
Average reactor temperatures (°C)
MAX
MIN
Cntr. Pt. Avg.
Cntr. Pt. St. Dev.

204
Tab ata
le 54. Product recovery system temperature d
Run
No.
DOE
No.
Nominal
heat carrier
temperature (°C)
Nominal heat
carrier feed rate
(kg/hr)
SF1
inlet
SF1
wall
SF2
wall
SF3
inlet
SF4
outlet
1 17 525 9 455.0 117.1 12.2 56.3 11.5
2 18 525 21 463.3 113.1 12.9 61.3 9.9
3 15 475 18 461.1 106.9 12.8 53.9 11.9
4 7 575 18 462.7 117.8 14.5 60.2 13.1
5 9 475 18 459.4 106.5 13.2 59.2 11.8
6 13 475 18 458.8 109.7 13.0 55.8 13.9
7 11 475 18 466.1 108.8 13.3 64.6 12.3
8 5 575 18 471.7 117.9 13.8 59.5 13.8
9 1 575 18 468.1 138.9 20.4 72.1 8.1
10 3 575 18 469.4 105.0 17.2 74.0 9.9
11 21 525 15 464.8 124.4 12.0 50.2 14.8
12 28 525 15 463.2 121.6 13.7 58.8 14.5
13 23 425 15 464.3 92.7 10.3 38.2 15.0
14 19 525 15 466.9 113.5 12.9 57.7 11.5
15 29 525 15 465.6 124.0 14.3 52.4 13.1
16 20 525 15 469.1 118.0 12.1 56.3 13.1
17 27 525 15 465.7 127.6 12.4 63.1 11.4
18 22 525 15 466.0 128.7 14.1 70.5 10.0
19 25 525 15 465.9 119.6 14.3 64.0 13.3
20 24 625 15 468.7 121.9 11.5 66.7 10.7
21 26 525 15 464.3 118.6 12.6 63.9 14.5
22 30 525 15 466.4 113.6 14.0 62.5 13.3
23 8 575 12 466.5 117.6 14.5 64.0 15.8
24 2 575 12 466.6 112.7 14.7 68.1 13.3
25 12 475 12 463.3 96.5 12.4 56.7 10.7
26 6 575 12 466.7 118.6 13.4 60.0 13.6
27 10 475 12 465.1 112.4 15.2 58.1 9.7
28 14 475 12 463.9 99.4 14.3 56.5 12.1
29 4 575 12 466.9 105.8 16.2 73.0 10.4
30 16 475 12 464.0 95.9 13.6 55.2 13.3
465.0 114.2 13.7 60.4 12.3
3.4 10.3 1.9 7.4 1.8
465.2 120.8 13.5 60.8 13.3
1.2 4.8 0.8 4.5 1.1
471.7 138.9 20.4 74.0 15.8
455.0 92.7 10.3 38.2 8.1
Product recovery system temperatures (°C)
MAX
MIN
Overall Avg.
Overall St. Dev.
Cntr. Pt. Avg.
Cntr. Pt. St. Dev.

205
Table 55. Bio-oil fraction mass balance data
Run
No.
DOE
No. SF1 SF2 SF3 SF4
1 17 48.96 30.45 18.57 2.03
2 18 45.23 32.83 20.25 1.69
3 15 64.36 17.75 16.64 1.26
4 7 49.00 30.67 18.69 1.64
5 9 60.17 16.66 20.41 2.76
6 13 63.55 17.84 16.82 1.78
7 11 42.08 34.43 21.11 2.38
8 5 58.12 22.71 17.34 1.83
9 1 40.54 33.32 24.27 1.87
10 3 38.98 32.64 25.64 2.73
11 21 56.38 28.64 14.11 0.87
12 28 53.98 28.07 16.78 1.17
13 23 65.40 21.26 12.03 1.31
14 19 62.51 18.89 17.32 1.28
15 29 53.13 30.04 15.55 1.28
16 20 52.35 26.73 19.23 1.68
17 27 43.52 34.73 19.69 2.05
18 22 43.84 29.34 23.94 2.89
19 25 42.91 35.01 19.98 2.10
20 24 45.40 31.63 20.83 2.15
21 26 41.62 35.04 21.14 2.20
22 30 49.85 28.32 19.78 2.05
23 8 43.64 36.81 18.07 1.48
24 2 51.55 23.48 22.82 2.16
25 12 58.49 20.23 19.22 2.06
26 6 58.54 23.05 16.49 1.91
27 10 45.98 32.02 19.49 2.51
28 14 47.85 34.42 16.24 1.48
29 4 48.27 27.19 22.18 2.35
30 16 54.73 28.16 15.36 1.75
51.03 28.08 19.00 1.89
7.76 5.95 3.07 0.50
47.50 31.87 18.82 1.81
5.49 3.42 2.16 0.46
65.40 36.81 25.64 2.89
38.98 16.66 12.03 0.87
Mass fraction of bio-oil collected (%-wt., wb)
MAX
MIN
Overall Avg.
Overall St. Dev.
Cntr. Pt. Avg.
Cntr. Pt. St. Dev.

206
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
Residual
(Bio-oil
yield
%-wt.,
wb)
0 5 10 15 20 25 30
Run number
Figure 137. Residuals for bio-oil yield full model
Table 56. Bio-oil yield model statistical data
Term Estimate
Standard
error t-ratio
Prob > |t|
(p-value) Estimate
Standard
error t-ratio
Prob > |t|
(p-value)
67.379 0.4629 145.56 <0.0001 66.884 0.3231 206.99 <0.0001
7.357 0.2314 31.79 <0.0001 7.357 0.2285 32.20 <0.0001
0.631 0.2314 2.73 0.0156 0.631 0.2285 2.76 0.0117
-0.524 0.2314 -2.26 0.0389 -0.524 0.2285 -2.29 0.0324
2.278 0.2314 9.84 <0.0001 2.278 0.2285 9.97 <0.0001
-0.288 0.2835 -1.02 0.3254 - - - -
1.238 0.2835 4.37 0.0006 1.238 0.2798 4.42 0.0002
0.090 0.2835 0.32 0.7544 - - - -
-0.639 0.2835 -2.26 0.0394 -0.639 0.2798 -2.29 0.0328
-0.209 0.2835 -0.74 0.4732 - - - -
0.207 0.2835 0.73 0.4774 - - - -
-2.417 0.2165 -11.17 <0.0001 -2.356 0.2099 -11.22 <0.0001
0.387 0.2165 1.79 0.0938 - - - -
-0.108 0.2165 -0.50 0.6266 - - - -
-0.787 0.2165 -3.64 0.0024 -0.725 0.2099 -3.46 0.0024
R2
R2
adj RMSE Mean R2
R2
adj RMSE Mean
0.9884 0.9776 1.13 64.42 0.9842 0.9781 1.12 64.42
ANOVA DOF
Sum of
Squares
(SS-)
Mean
Square
(MS-) FANOVA
Prob > F
(p-value) DOF
Sum of
Squares
(SS-)
Mean
Square
(MS-) FANOVA
Prob > F
(p-value)
Regression (R) 14 1641.83 117.27 91.22 <0.0001 8 1634.80 204.35 163.09 <0.0001
Error (E) 15 19.28 1.29 21 26.31 1.25
Total (T) 29 1661.11 29 1661.11
Lack of fit analysis DOF
Sum of
Squares
Mean
Square FLOF
Prob > F
(p-value) DOF
Sum of
Squares
Mean
Square FLOF
Prob > F
(p-value)
Lack of fit 10 13.58 1.36 1.19 0.4489 16 20.61 1.29 1.13 0.4865
Pure error 5 5.70 1.14 5 5.70 1.14
Total 15 19.28 21 26.31
HC temperature · HC temperature
Reduced model
Full model
Intercept
HC temperature
HC feed rate
Auger speed
N2 flow rate
HC temperature · N2 flow rate
HC temperature · Auger speed
Auger speed · HC feed rate
N2 flow rate · HC feed rate
HC temperature · HC feed rate
N2 flow rate · Auger speed
Summary of model fit
HC feed rate · HC feed rate
Auger speed · Auger speed
N2 flow rate · N2 flow rate

207
Table 57. Coded levels for model equations
THC
(°C)
QN2
(SLPM)
ωA
(% of 180 RPM) (kg/hr) Coded level
425 1.5 25.0 9 -2
475 2.0 27.5 12 -1
525 2.5 30.0 15 0
575 3.0 32.5 18 1
625 3.5 35.0 21 2
Factor
HC
m

The parameter values in the resulting model equations must be substituted according to Table
57, which only lists the five levels associated with the experimental design. To investigate values
other than these levels, the normalized Equations D1 – D4 are used to interpolate and find the correct
value for the model based on an experimental level of interest. This form is used based on the
software package selected to perform the regression procedures. Note the equations can be solved by
using values beyond the ran ng must be closely
scrutinized. Equation D5 is used as an example calculation for bio-oil yield (see Equation 24).
ge of -2 to +2, but results obtained by extrapolati





 −
°
=
50
525
C)
(
T
τ HC
HC
Equation D1





 −
=
0.5
2.5
(sL/min)
Q
θ N2
N2
Equation D2





 −
=
2.5
30
RPM)
180
of
(%
ω
Ω A
A
Equation D3





 −
=
3
12
(kg/hr)
m
μ HC
HC

 Equation D4
2
HC
2
HC
HC
HC
A
HC
HC
A
N2
HC
oil
bio
3
15
m
0.73
50
525
T
2.36
3
15
m
50
525
T
0.64
2.5
30
ω
50
525
T
1.24
3
15
m
2.28
2.5
30
ω
0.52
0.5
2.5
Q
0.63
50
525
T
7.36
66.9
wb)
wt.,
(%
Y





 −
⋅
−





 −
⋅
−











 −
⋅





 −
⋅
−











 −
⋅





 −
⋅
+





 −
⋅
+





 −
⋅
−





 −
⋅
+





 −
⋅
+
=
−
−



Equation D5

208
3
2
1
0
-1
-2
4
(Biochar
yield
%-wt.,
wb)
-4
-3
Residual
0 5 10 15 20 25 30
Run number
Figure 138. Residuals for biochar yield full model
Table 58. Biochar yield model statistical data
Term Estimate
Standard
error t-ratio
Prob > |t|
(p-value) Estimate error t-ratio
ob > |t|
(p-value)
20.193 0.8078 25.00 <0.0001 20.545 0.5582 36.81 <0.0001
-7 292 0.4039 -18.05 <0.0001 -7 .0001
-0.889 0.4039 -2.20 0.0438 -0.889 0.3947 -2.25 0.0341
0.577 0.4039 1.43 0.1734 - - - -
-2.773 0.4039 -6.87 <0.0001 -2.773 0.3947 -7.03 <0.0001
0.126 0.4947 0.25 0.8025 -
-1.314 0.4947 -2.66 0.0180 -1.314 0.4834 -2.72 0.0123
-0.050 0.4947 -0.10 0.9211 - - - -
0.740 0.4947 1.49 0.1557 - - - -
0.386 0.4947 0.78 0.4479 -
-0.466 0.4947 -0.94 0.3612 - - - -
1.072 0.3778 2.84 0.0125 1.028 0.3626 2.83 0.0094
0.388 0.3778 1.03 0.3202 - - - -
-0.036 0.3778 -0.10 0.9255 - - - -
1.216 0.3778 3.22 0.0057 1.172 0.3626 3.23 0.0037
R2
R2
adj RMSE Mean R2
R2
adj RMSE Mean
0.9645 0.9314 1.98 22.31 0.9481 0.9345 1.93 22.31
ANOVA DOF
Sum of
Squares
(SS-)
Mean
Square
(MS-) FANOVA
Prob > F
(p-value) DOF
Sum of
Squares
(SS-)
Mean
Regression (R) 14 1596.80 114.06 29.13 <0.0001 6 1569.53 261.59 69.96 <0.0001
Error (E) 15 58.73 3.92 23 86.00 3.74
Total (T) 29 1655.53 29 1655.53
Sum of
Squares
Mean
Square FLOF
Prob > F
(p-value) DOF
Sum of
Squares
Mean
Square FLOF
Prob > F
(p-value)
Lack of fit 10 42.64 4.26 1.32 0.3985 18 69.91 3.88 1.21 0.4543
Pure error 5 16.09 3.22 5 16.09 3.22
Total 15 58.73 23 86.00
Full mod
HC feed rate
Intercept
HC temperature
N2 flow rate
Auger speed
Standard Pr
el Reduced model
. .292 0.3947 -18.47 <0
- - -
- - -
Square
(MS-) FANOVA
Prob > F
(p-value)

209
-4
-2
0
2
4
6
Residual
(NCG
yield
%-wt.,
wb)
0 5 10 15 20 25 30
Run number
Figure odel
Table 59. Non-condensable gas yield mmary
139. Residuals for non-condensable gas yield full m
model, statistics su
Statistic Value Significant Hypothesis tests
R2
0.494 - -
FANOVA 1.04 X FANOVA < F0.05,k,ν
F0.05,k,ν 2.424 - Don't reject Ho1
FLOF 261.8 √ FLOF > F0.05,λ,m-1
F0.05,λ,m-1 4.74 - Don' reject Ho2
t0.05,ν 2.13 - -
|t| statistics
for model terms Value Significant Hypothesis tests
β0 13.12 √ |t| > t0.05,ν Reject Ho3
β1 0.76 X |t| < t0.05,ν Don't reject Ho3
β11 3.03 √ |t| > t0.05,ν Reject Ho3

210
Table 60. Non-condensable gas yield model statistical data
Term Estimate
Standard
error t-ratio
Prob > |t|
(p-value)
11.347 0.8647 13.12 <0.0001
-0.326 0.4324 -0.76 0.4619
0.099 0.4324 0.23 0.8225
-0.136 0.4324 -0.31 0.7581
0.353 0.4324 0.82 0.4265
0.123 0.5295 0.23 0.8196
0.296 0.5295 0.56 0.5847
0.071 0.5295 0.13 0.8945
-0.079 0.5295 -0.15 0.8829
0.115 0.5295 0.22 0.8316
0.229 0.5295 0.43 0.6716
1.225 0.4044 3.03 0.0085
-0.313 0.4044 -0.77 0.4509
87
-0.265 0.4044 -0.66 0.5221
R2
R2
adjusted RMSE Mean
.60
ANOVA DOF
Sum of
Squares
(SS-)
Mean
Square
(MS-) FANOVA
Prob > F
(p-value)
Regression (R) 14 65.61 4.69 1.04 0.4650
Error (E) 15 67.30 4.49
Total (T) 29 132.91
Sum of
Squares
Mean
Square FLOF
Prob > F
(p-value)
Lack of fit 10 67.17 6.72 261.77 <0.0001
Pure error 5 0.13 0.03
Total 15 67.30
Full model
Auger sp
HC feed rate
Intercept
HC temperature
N2 flow rate
Auger speed
-0.329 0.4044 -0.81 0.42
eed · Auger speed
0.4937 0.0211 2.1181 11

211
Table 61. Non-condensable gas data, composition
Run
No.
DOE
No.
N2 H2 CO CH4 C2H6 C2H4 CO2
1 17 68.26 0.021 10.51 0.886 0.082 0.152 15.82
2 18 65.14 0.012 12.28 1.569 0.139 0.206 16.06
3 15 65.54 0.593 11.30 1.038 0.106 0.142 15.01
4 7 59.13 1.547 14.03 2.062 0.187 0.266 15.11
5 9 72.32 0.413 8.31 0.702 0.072 0.100 11.25
6 13 63.82 0.713 11.24 0.961 0.095 0.144 15.35
7 11 71.16 0.450 8.55 0.749 0.075 0.092 11.91
8 5 59.20 1.310 15.05 2.110 0.181 0.297 15.81
9 1 69.11 0.724 10.11 1.361 0.119 0.174 11.01
10 3 67.38 0.820 11.33 1.617 0.143 0.196 12.54
11 21 49.26 1.185 19.58 2.093 0.193 0.311 23.56
12 28 58.12 0.844 15.91 1.666 0.158 0.233 19.50
13 23 63.04 0.783 12.55 0.844 0.090 0.152 19.90
14 19 61.24 0.757 14.66 1.598 0.153 0.199 18.42
15 29 56.91 1.220 16.58 1.803 0.169 0.262 19.85
16 20 67.93 0.779 12.28 1.259 0.116 0.187 14.77
17 27 66.56 0.693 12.58 1.269 0.123 0.175 15.63
18 22 72.83 0.474 10.09 0.995 0.098 0.127 12.76
19 25 66.50 0.698 12.72 1.298 0.127 0.181 15.65
20 24 60.87 1.409 16.26 2.204 0.192 0.317 15.72
21 26 67.06 0.743 12.40 1.341 0.129 0.178 15.15
22 30 66.25 0.736 12.71 1.351 0.127 0.180 15.60
23 8 58.94 0.942 15.55 1.756 0.161 0.222 18.28
24 2 68.68 0.769 11.83 1.403 0.125 0.178 13.45
25 12 71.06 0.502 10.11 0.857 0.089 0.119 13.85
26 6 58.78 0.946 15.73 1.600 0.144 0.233 18.66
27 10 72.54 0.434 9.40 0.666 0.072 0.109 13.60
28 14 63.89 0.660 12.53 0.814 0.094 0.148 18.41
29 4 63.54 0.801 13.81 1.359 0.136 0.177 16.67
30 16 63.16 0.455 10.73 0.765 0.088 0.118 15.42
63.57 0.822 13. 2 1.455 0.139 0.201 16.90
4.71 0.20 1. 0.22 0.02 0.04 2.16
72.83 1.55 19. 8 2.20 0.19 0.32 23.56
49.26 0.01 8. 0.67 0.07 0.09 11.01
MAX
MIN
Cntr. Pt. Avg.
Cntr. Pt. St.Dev.
Gas composition (%-vol., kmoli/100 kmolNCG)
8
89
5
31

212
Table 62. Non-condensable gas data, molar analysis
Mass Molecular weight No. mols
Run
No.
DOE
No.
m
(gNCG)
M
(kg/kmolNCG)
n = m/M
(molNCG)
H2 CO CH4 C2H6 C2H4 CO2
1 17 107.89 36.84 2.93 0.001 0.382 0.032 0.003 0.006 0.575
2 18 88.87 35.88 2.48 0.000 0.405 0.052 0.005 0.007 0.530
3 15 84.68 35.55 2.38 0.021 0.401 0.037 0.004 0.005 0.532
4 7 99.87 33.35 2.99 0.047 0.423 0.062 0.006 0.008 0.455
5 9 92.47 35.73 2.59 0.020 0.399 0.034 0.003 0.005 0.540
6 13 92.50 35.58 2.60 0.025 0.394 0.034 0.003 0.005 0.539
7 11 102.66 35.80 2.87 0.021 0.392 0.034 0.003 0.004 0.546
8 5 56.05 33.59 1.67 0.038 0.433 0.061 0.005 0.009 0.455
9 1 117.14 34.03 3.44 0.031 0.430 0.058 0.005 0.007 0.469
10 3 110.36 34.02 3.24 0.031 0.425 0.061 0.005 0.007 0.470
11 21 97.70 34.86 2.80 0.025 0.417 0.045 0.004 0.007 0.502
12 28 101.43 35.07 2.89 0.022 0.415 0.043 0.004 0.006 0.509
13 23 95.17 36.40 2.61 0.023 0.366 0.025 0.003 0.004 0.580
14 19 91.03 35.17 2.59 0.021 0.410 0.045 0.004 0.006 0.515
15 29 100.75 34.65 2.91 0.031 0.416 0.045 0.004 0.007 0.498
16 20 87.12 34.86 2.50 0.026 0.418 0.043 0.004 0.006 0.503
17 27 99.54 35.14 2.83 0.023 0.413 0.042 0.004 0.006 0.513
18 22 106.27 35.35 3.01 0.019 0.411 0.041 0.004 0.005 0.520
19 25 105.12 35.08 3.00 0.023 0.415 0.042 0.004 0.006 0.510
20 24 132.22 33.24 3.98 0.039 0.450 0.061 0.005 0.009 0.435
21 26 110.01 34.93 3.15 0.025 0.414 0.045 0.004 0.006 0.506
22 30 103.64 35.00 2.96 0.024 0.414 0.044 0.004 0.006 0.508
23 8 110.86 34.71 3.19 0.026 0.421 0.048 0.004 0.006 0.495
24 2 106.98 34.45 3.11 0.028 0.426 0.051 0.005 0.006 0.485
25 12 92.93 35.79 2.60 0.020 0.396 0.034 0.003 0.005 0.543
26 6 92.80 34.85 2.66 0.025 0.421 0.043 0.004 0.006 0.500
27 10 87.63 36.19 2.42 0.018 0.387 0.027 0.003 0.004 0.560
28 14 96.55 36.21 2.67 0.020 0.384 0.025 0.003 0.005 0.564
29 4 115.44 34.99 3.30 0.024 0.419 0.041 0.004 0.005 0.506
30 16 108.02 36.20 2.98 0.016 0.389 0.028 0.003 0.004 0.559
34.98 2.957 0.024 0.414 0.044 0.004 0.006 0.507
0.178 0.110 0.003 0.001 0.001 0.000 0.000 0.005
36.84 3.977 0.047 0.450 0.062 0.006 0.009 0.580
33.24 1.669 0.000 0.366 0.025 0.003 0.004 0.435
Note: a - Nitrogen free basis
Gas mol fraction, yi (kmoli/kmolNCG)a
Cntr. Pt. Avg.
Cntr. Pt. St.Dev.
MAX
MIN

213
Table 63. Non-condensable gas data, mass analysis
Run
No.
DOE
No.
H2 CO CH4 C2H6 C2H4 CO2 H2 CO CH4 C2H6 C2H4 CO2
1 17 0.00 1.12 0.09 0.01 0.02 1.68 0.00 31.34 1.51 0.26 0.46 74.11 3.13 7.41
2 18 0.00 1.00 0.13 0.01 0.02 1.31 0.00 28.12 2.06 0.34 0.47 57.79 3.90 8.01
3 15 0.05 0.96 0.09 0.01 0.01 1.27 0.10 26.76 1.41 0.27 0.34 55.81 3.50 7.29
4 7 0.14 1.27 0.19 0.02 0.02 1.36 0.28 35.45 2.98 0.51 0.67 59.97 4.47 7.57
5 9 0.05 1.03 0.09 0.01 0.01 1.40 0.10 28.90 1.40 0.27 0.35 61.46 3.38 7.18
6 13 0.07 1.02 0.09 0.01 0.01 1.40 0.13 28.71 1.41 0.26 0.37 61.63 3.21 6.89
7 11 0.06 1.12 0.10 0.01 0.01 1.56 0.12 31.46 1.58 0.30 0.34 68.86 3.35 7.32
8 5 0.06 0.72 0.10 0.01 0.01 0.76 0.13 20.24 1.62 0.26 0.40 33.40 4.38 7.22
9 1 0.11 1.48 0.20 0.02 0.03 1.61 0.21 41.48 3.20 0.53 0.72 71.01 4.45 7.62
10 3 0.10 1.38 0.20 0.02 0.02 1.53 0.20 38.64 3.16 0.52 0.67 67.17 4.28 7.44
11 21 0.07 1.17 0.13 0.01 0.02 1.41 0.14 32.75 2.01 0.35 0.52 61.93 3.80 7.19
12 28 0.06 1.20 0.13 0.01 0.02 1.47 0.13 33.63 2.02 0.36 0.49 64.80 3.69 7.12
13 23 0.06 0.96 0.06 0.01 0.01 1.52 0.12 26.78 1.03 0.21 0.32 66.71 2.69 6.71
14 19 0.05 1.06 0.12 0.01 0.01 1.33 0.11 29.70 1.85 0.33 0.40 58.63 3.77 7.44
15 29 0.09 1.21 0.13 0.01 0.02 1.45 0.18 33.86 2.11 0.37 0.54 63.70 3.90 7.34
16 20 0.07 1.04 0.11 0.01 0.02 1.26 0.13 29.24 1.72 0.30 0.45 55.28 3.70 7.00
17 27 0.06 1.17 0.12 0.01 0.02 1.45 0.13 32.76 1.89 0.34 0.46 63.96 3.71 7.25
18 22 0.06 1.24 0.12 0.01 0.02 1.56 0.12 34.62 1.96 0.36 0.44 68.77 3.70 7.36
19 25 0.07 1.24 0.13 0.01 0.02 1.53 0.14 34.80 2.03 0.37 0.49 67.28 3.76 7.27
20 24 0.16 1.79 0.24 0.02 0.03 1.73 0.31 50.16 3.89 0.63 0.98 76.21 4.89 7.42
21 26 0.08 1.30 0.14 0.01 0.02 1.59 0.16 36.53 2.26 0.41 0.53 70.13 3.79 7.27
22 30 0.07 1.23 0.13 0.01 0.02 1.50 0.14 34.34 2.09 0.37 0.49 66.21 3.74 7.20
23 8 0.08 1.35 0.15 0.01 0.02 1.58 0.16 37.68 2.44 0.42 0.54 69.61 3.93 7.26
24 2 0.09 1.32 0.16 0.01 0.02 1.51 0.17 37.07 2.52 0.42 0.56 66.24 4.02 7.18
25 12 0.05 1.03 0.09 0.01 0.01 1.41 0.10 28.81 1.40 0.27 0.34 62.01 3.21 6.92
26 6 0.07 1.12 0.11 0.01 0.02 1.33 0.14 31.44 1.83 0.31 0.47 58.62 3.77 7.03
27 10 0.04 0.94 0.07 0.01 0.01 1.36 0.09 26.25 1.07 0.22 0.30 59.71 2.83 6.45
28 14 0.05 1.02 0.07 0.01 0.01 1.50 0.11 28.65 1.07 0.23 0.34 66.15 2.70 6.24
29 4 0.08 1.38 0.14 0.01 0.02 1.67 0.16 38.73 2.18 0.41 0.50 73.46 3.85 7.30
30 16 0.05 1.16 0.08 0.01 0.01 1.67 0.10 32.52 1.33 0.29 0.36 73.43 3.27 7.38
0.07 1.23 0.13 0.01 0.02 1.50 0.15 34.32 2.07 0.37 0.50 66.01 3.77 7.24
0.01 0.05 0.01 0.00 0.00 0.06 0.02 1.28 0.12 0.02 0.03 2.43 0.075 0.076
0.16 1.79 0.24 0.02 0.03 1.73 0.31 50.16 3.89 0.63 0.98 76.21 4.89 8.01
0.00 0.72 0.06 0.01 0.01 0.76 0.00 20.24 1.03 0.21 0.30 33.40 2.69 6.24
Note: a - Percent weight yield on a wet biomass basis (gramsi/gram biomass)
CO
yielda
(%-wt.,
wb)
CO2
yielda
(%-wt.,
wb)
Cntr. Pt. St. Dev.
MAX
MIN
Massi (gi), mi = ni · Mi
Number of moli (moli), ni = n · yi
Cntr. Pt. Avg.

214
Table rties
64. Non-condensable gas data, volume meter prope
Run
No.
DOE
No.
Volume meter
average
temperature
(°C)
Volume meter
average
pressure
(in-H2Og)
Total
elapsed
volume
(m3
)
1 17 27.46 1.19 0.228
2 18 29.84 1.30 0.179
3 15 25.01 0.82 0.183
4 7 25.32 1.20 0.189
5 9 25.87 1.89 0.257
6 13 27.29 1.26 0.192
7 11 26.70 1.79 0.275
8 5 25.66 1.02 0.105
9 1 23.63 1.83 0.304
10 3 26.17 1.61 0.260
11 21 24.10 0.48 0.134
12 28 23.27 0.62 0.166
13 23 22.75 0.50 nd
14 19 23.26 0.84 0.161
15 29 24.06 0.72 0.163
16 20 24.46 1.03 0.189
17 27 24.85 1.02 0.206
18 22 23.57 1.57 0.266
19 25 24.89 1.07 0.218
20 24 24.26 1.17 0.246
21 26 28.12 1.11 0.238
22 30 28.06 1.05 0.218
23 8 31.20 0.82 0.197
24 2 28.96 1.26 0.247
25 12 28.84 1.13 0.224
26 6 30.19 0.75 0.163
27 10 26.97 1.04 0.218
28 14 27.51 0.73 0.183
29 4 29.68 1.23 0.227
30 16 27.91 1.08 0.236
Overall Avg. 26.33 1.10 0.21
Overall St. Dev. 2.35 0.37 0.04
Cntr. Pt. Avg. 25.54 0.93 0.20
Cntr. Pt. St. Dev. 2.06 0.21 0.03
Notes: nd - Not determined

215
-0.10
-0.05
0.00
0.05
0.10
0.15
Residual
(CO
yield
%-wt.,
wb)
0 10 20 30
Run number
Figure 140. Residuals for carbon monoxide yield full model
Table 65. Carbon monoxide yield model statistical data
Term Estimate
Standard
error t-ratio
Prob > |t|
(p-value) Estimate
Standard
error t-ratio
Prob > |t|
(p-value)
3.766 0.0356 105.83 <0.0001 3.746 0.0195 192.05 <0.0001
0.504 0.0178 28.32 <0.0001 0.504 0.0169 29.83 <0.0001
-0.002 0.0178 -0.12 0.9029 - - - -
-0.052 0.0178 -2.91 0.0108 -0.052 0.0169 -3.07 0.0057
0.206 0.0178 11.60 <0.0001 0.206 0.0169 12.22 <0.0001
-0.003 0.0218 -0.12 0.908 - - - -
0.080 0.0218 3.69 0.0022 0.080 0.0207 3.88 0.0008
0.069 0.0218 3.15 0.0066 0.069 0.0207 3.32 0.0031
0.038 0.0218 1.72 0.1053 - - - -
-0.022 0.0218 -0.99 0.3381 - - - -
0.047 0.0218 2.15 0.0484 0.047 0.0207 2.26 0.0339
0.000 0.0166 -0.02 0.9867 - - - -
-0.009 0.0166 -0.54 0.5954 - - - -
-0.013 0.0166 -0.80 0.4373 - - - -
-0.069 0.0166 -4.13 0.0009 -0.066 0.01542 -4.3 0.0003
R2
R2
adjusted RMSE Mean R2
R2
adjusted RMSE Mean
0.9851 0.9713 0.087 3.69 0.9804 0.9741 0.0828 3.69
ANOVA DOF
Sum of
Squares
(SS-)
Mean
Square
(MS-) FANOVA
Prob > F
(p-value) DOF
Sum of
Squares
(SS-)
Mean
Square
(MS-) FANOVA
Prob > F
(p-value)
Regression (R) 14 7.557 0.540 71.04 <0.0001 7 7.520 1.0 156.88 <0.0001
Error (E) 15 0.114 0.008 22 0.151 0.0
Total (T) 29 7.671 29 7.671
Sum of
Squares
Mean
Square FLOF
Prob > F
(p-value) DOF
Sum of
Squares
Mean
Square FLOF
Prob > F
(p-value)
Lack of fit 10 0.086 0.009 1.51 0.3398 17 0.122 0.007 1.27 0.4278
Pure error 5 0.028 0.006 5 0.028 0.006
Total 15 0.114 22 0.151
Full model Reduced model
HC feed rate
Intercept
HC temperature
N2 flow rate
Auger speed
74
06

216
-0.25
-0.20
-0.15
-0.10
-0.05
0.00
0.05
0.10
0.15
0.20
Residual
(CO2
yield
%-wt.,
wb)
0 10 20 30
Run number
Figure 141. Residuals for carbon dioxide yield full model
Table 66. Carbon dioxide yield model statistical data
Term Estimate
Standard
error t-ratio
Prob > |t|
(p-value) Estimate
Standard
error t-ratio
Prob > |t|
(p-value)
7.243 0.0597 121.36 <0.0001 7.188 0.0456 157.64 <0.0001
0.183 0.0298 6.13 <0.0001 0.183 0.0322 5.67 <0.0001
0.036 0.0298 1.20 0.2474 - - - -
-0.147 0.0298 -4.93 0.0002 -0.147 0.0322 -4.56 0.0002
0.166 0.0298 5.57 <0.0001 0.166 0.0322 5.16 <0.0001
0.025 0.0365 0.69 0.5023 - - - -
0.103 0.0365 2.80 0.0133 0.103 0.0395 2.60 0.0165
0.097 0.0365 2.66 0.0178 0.097 0.0395 2.46 0.0221
-0.039 0.0365 -1.07 0.3028 - - - -
0.041 0.0365 1.13 0.2768 - - - -
0.078 0.0365 2.12 0.0507 - - - -
-0.073 0.0279 -2.63 0.019 -0.066 0.0296 -2.24 0.0352
-0.021 0.0279 -0.75 0.4665 - - - -
-0.034 0.0279 -1.23 0.239 - - - -
0.088 0.0279 3.14 0.0068 0.094 0.0296 3.19 0.0043
R2
R2
R2
adjusted RMSE Mean
0.9022 0.8109 0.1462 7.210 0.8326 0.7793 0.1580 7.210
ANOVA DOF
Sum of
Squares
(SS-)
Mean
Square
(MS-) FANOVA
Prob > F
(p-value) DOF
Sum of
Squares
(SS-)
Mean
Square
(MS-) FANOVA
Prob > F
(p-value)
Regression (R) 14 2.957 0.211 9.88 <0.0001 7 2.729 0.390 15.63 <0.0001
Error (E) 15 0.321 0.021 22 0.549 0.025
Total (T) 29 3.278 29 3.278
Sum of
Squares
Mean
Square FLOF
Prob > F
(p-value) DOF
Sum of
Squares
Mean
Square FLOF
Prob > F
(p-value)
Lack of fit 10 0.292 0.029 5.04 0.044 17 0.520 0.031 5.28 0.0374
Pure error 5 0.029 0.006 5 0.029 0.006
Total 15 0.321 22 0.549
HC feed rate
Intercept
HC temperature
N2 flow rate
Auger speed

217
Table 67. Moisture content analytical data
Run
No.
DOE
No. Avg. St. Dev Avg. St. Dev Avg. St. Dev Avg. St. Dev Avg. St. Dev
1 17 26.29 1.40 17.37 0.38 43.77 3.52 16.04 0.39 73.29 3.32
2 18 25.01 1.15 15.04 0.41 40.86 2.29 17.88 0.86 69.49 2.19
3 15 28.37 1.92 25.66 1.60 46.62 1.11 16.47 3.93 66.62 3.24
4 7 23.48 2.55 15.93 0.96 37.86 4.96 15.76 2.52 67.80 5.43
5 9 27.84 1.18 24.56 1.35 44.35 0.85 17.38 0.70 77.07 2.92
6 13 28.84 0.77 25.38 0.68 45.60 0.41 15.93 0.73 67.51 4.36
7 11 27.78 1.19 17.18 1.23 42.25 1.50 20.46 0.62 70.83 1.16
8 5 23.00 0.99 17.67 1.05 39.17 1.05 14.51 0.54 72.15 2.70
9 1 22.14 0.76 11.50 1.00 34.58 0.86 19.20 0.15 69.37 1.56
10 3 24.09 1.18 10.85 0.35 36.91 2.11 23.40 1.17 66.28 2.15
11 21 25.90 1.68 18.43 1.78 45.13 1.80 13.80 0.89 73.36 4.18
12 28 26.13 1.46 20.26 1.02 41.13 1.43 17.44 2.70 61.75 4.75
13 23 35.00 1.29 31.02 0.56 54.97 2.80 17.11 1.85 74.05 7.79
14 19 25.82 1.61 22.43 1.62 43.57 2.12 15.62 0.98 67.52 2.70
15 4.61 4.52
16 20 24.84 1.32 16.89 0.86 44.10 1.93 15.86 1.78 68.88 0.99
17 27 27.38 1.17 16.37 0.97 42.89 0.88 20.11 1.60 68.03 6.25
18 22 68.08 4.61
19 25 67.17 5.71
20 24 22.04 2.45 10.73 2.59 37.50 2.50 18.17 1.95 71.04 3.67
21 26 24.86 1.95 13.22 2.55 38.78 1.51 19.87 1.14 71.09 5.56
22 30 24.39 2.59 17.20 2.27 37.63 3.45 19.57 1.83 62.89 5.65
23 8 22.79 1.90 11.54 2.24 37.39 1.63 17.06 1.55 61.23 2.92
24 2 23.03 1.79 16.23 1.70 37.98 2.43 18.56 1.09 70.15 4.36
25 12 28.46 0.40 24.36 0.23 46.13 0.58 18.12 0.37 67.94 3.78
26 6 22.51 0.61 14.63 0.24 42.12 0.33 17.31 1.84 72.46 4.63
27 10 31.58 1.49 22.85 0.81 48.71 2.53 18.77 1.22 72.50 2.82
28 14 30.39 1.63 20.18 2.55 49.20 0.88 17.20 0.57 67.86 0.55
29 4 27.14 1.97 16.96 0.87 49.15 4.86 17.72 0.79 70.34 2.26
30 16 29.92 1.17 23.04 1.08 48.13 1.25 16.66 1.02 68.24 3.59
25.74 1.79 16.52 1.60 41.25 1.76 17.82 2.14 65.92 5.41
- 1.71 - 1.96 - 2.37 - 5.44
35.00 31.02 2.59 54.97 4.96 23.40 4.13 77.07 7.79
22.04 0.40 10.73 0.23 34.58 0.33 11.77 0.15 61.23 0.55
Notes: All values in %-wt., wb. Each analysis performed in triplicate. a- Pooled standard deviation
MIN
Whole bio-oil SF1
Cnt. Pt. Avg.
SF2 SF3 SF4
MAX
Cntr. Pt. St. Dev.a
29 25.50 1.75 18.52 1.30 43.28 1.20 11.77 4.13 6
26.15 2.17 14.59 2.42 43.01 2.67 21.59 0.82
26.17 1.80 13.52 1.50 43.82 2.13 18.13 1.44
- 1.84
2.59
( ) ( ) ( )
( )
1/2
k
2
1
2
k
k
2
2
2
2
1
1
p
k
n
...
n
n
s
1
n
...
s
1
n
s
1
n
s








−
+
+
+
⋅
−
+
+
⋅
−
+
⋅
−
= Equation D6
Where: nk = Number of tests performed for sample k, sk = Standard deviation for sample k
1.84
=
( ) 6
3
3
3
3
3
3
2.59
2
1.95
2
1.80
2
1.17
2
1.75
2
1.46
2
s
1/2
2
2
2
2
2
2
p 







−
+
+
+
+
+
⋅
+
⋅
+
⋅
+
⋅
+
⋅
+
⋅
= Equation D7

218
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
Residual
(KF
moisture
content
%-wt,
wb.)
0 5 10 15 20 25 30
Run number
Figure 142. Residuals for moisture content full model
Table 68. Moisture content model statistical data
Term Estimate
Standard
error t-ratio
Prob > |t|
(p-value) Estimate
Standard
error t-ratio
Prob > |t|
(p-value)
25.738 0.4109 62.64 <0.0001 25.671 0.2326 110.35 <0.0001
-2.955 0.2054 -14.39 <0.0001 -2.955 0.2015 -14.67 <0.0001
0.136 0.2054 0.66 0.5174 - - - -
-0.193 0.2054 -0.94 0.3619 - - - -
-0.535 0.2054 -2.61 0.0199 -0.535 0.2015 -2.66 0.0135
0.406 0.2516 1.61 0.1278 - - - -
-0.684 0.2516 -2.72 0.0159 -0.684 0.2468 -2.77 0.0104
-0.193 0.2516 -0.77 0.456 - - - -
0.298 0.2516 1.18 0.2554 - - - -
-0.402 0.2516 -1.60 0.1312 - - - -
-0.069 0.2516 -0.27 0.7875 - - - -
0.688 0.1922 3.58 0.0027 0.696 0.1839 3.78 0.0009
0.063 0.1922 0.33 0.7457 - - - -
-0.110 0.1922 -0.57 0.5772 - - - -
-0.029 0.1922 -0.15 0.8815 - - - -
R
2
R2
R2
adjusted RMSE Mean
0.9421 0.8880 1.01 26.23 0.9071 0.8923 0.9870 26.23
ANOVA DOF
Sum of
Squares
(SS-)
Mean
Square
(MS-) FANOVA
Prob > F
(p-value) DOF
Sum of
Squares
(SS-)
M
Square
(MS-) FANOVA
Prob > F
(p-value)
gression (R) 14 247.11 17.65 17.42 <0.0001 4 237.95 59.49 61.06 <0.0001
Error (E) 15 15.19 1.01 25 24.36 0.97
Lack of fit nalysis DOF
Sum of
Squares
Mean
Square FLOF
Prob > F
(p-value) DOF
Sum of
Squares
Mean
Square FLOF
Prob > F
(p-value)
Lack of fit 10 9.53 0.95 0.840 0.6203 10 5.93 0.59 0.483 0.8762
Pure error 5 5.67 1.13 15 18.42 1.23
Total 15 15.19 25 24.36
Intercept
HC temperature
N2 flow rate
Auger speed
HC feed rate
C feed rate · HC feed rate
H
ean
Re
Total (T) 29 262.30 29 262.30
a

219
Table 69. Water insoluble content analytical data
SF4
Run
No.
DOE
No. Avg. St. Dev Avg. St. Dev Avg. St. Dev Avg. St. Dev
1 17 15.0 0.53 15.29 0.09 6.41 0.11 29.92 2.5 -
2 18 17.4 1.33 18.95 1.23 9.80 0.97 27.62 2.3 -
3 15 14.2 0.31 13.10 0.17 5.91 0.04 28.17 1.2 -
4 7 18.3 0.83 20.10 0.95 8.64 0.44 30.92 1.2 -
5 9 14.3 0.30 12.94 0.33 7.86 0.04 25.47 0.5 -
6 13 14.3 0.24 13.03 0.19 7.64 0.15 27.58 0.5 -
7 11 14.0 0.22 15.50 0.34 7.36 0.13 23.61 0.1 -
8 5 18.9 0.49 19.48 0.70 10.35 0.10 30.36 0.3 -
9 1 19.6 0.47 24.65 0.80 9.74 0.15 26.41 0.4 -
10 3 18.5 0.20 22.31 0.23 9.43 0.28 26.07 0.1 -
11 21 15.7 0.61 16.82 0.87 8.80 0.30 26.53 0.2 -
12 28 15.4 0.45 15.86 0.06 6.79 0.15 29.24 2.2 -
13 23 9.6 0.20 9.81 0.20 2.12 0.08 22.76 0.4 -
14 19 16.1 0.40 15.7 0.37 7.93 0.30 27.68 0.6 -
15 29 15.9 0.27 17.36 0.20 7.21 0.07 29.02 0.9 -
16 20 17.1 0.23 17.9 .21 10.29 0.23 28.90 0.3 -
17 27 15.3 0.47 17.71 .34 7.62 0.08 25.37 1.5 0.10
18 22 0.8 -
19 25 16.4 0.42 19.15 0.55 8.67 0.31 25.95 0.4 -
20 24 22.5 0.61 24.4 0.64 12.08 0.44 36.52 0.9 0.26
21 26 14.8 0.52 17.10 0.30 7.55 0.55 23.83 0.9 0.32
22 30 15.6 0.64 16.58 0.36 7.9 0.26 26.07 2.0 -
23 8 17.62 0.14 20.34 0.13 10.27 0.09 27.47 0.25 -
24 2 18.34 0.51 18.62 0.35 10.31 0.26 27.73 1.16
25 12 14.47 2.31 14.27 3.58 6.00 0.82 25.55 0.29 -
26 6 19.50 0.42 19.75 0.19 11.03 0.13 32.69 1.70 -
27 10 13.75 - 15.83 - 4.63 - 25.64 - -
28 14 12.99 - 14.47 - 4.89 - 27.00 - -
29 4 17.65 - 19.15 - 9.72 - 26.00 - -
30 16 13.57 - 14.81 - 4.80 - 26.76 - -
15.59 0.46 17.29 0.30 7.62 0.24 26.58 1.31 -
- 0.47 - 0.34 - 0.29 - 1.47 -
22.51 2.31 24.65 3.58 12.08 0.97 36.52 2.47 -
9.61 0.14 9.81 0.06 2.12 0.04 22.76 0.06 -
Notes: All values in %-wt., wb. Analysis performed in triplicate for samples
with standard deviations shown. a- Pooled standard deviation
Whole bio-oil SF1 SF2 SF3
MIN
Cnt. Pt. Avg.
Cntr. Pt. St. Dev.a
MAX
0
0
17.4 0.44 19.4 0.27 9.92 0.44 25.23

220
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
Residual
(Water
insolubles
%-wt.,wb)
0 5 10 15 20 25 30
Run number
Figure 143. Residuals for water insoluble content full model
Table 70. Water insoluble content model statistical data
Term Estimate
Standard
error t-ratio
Prob > |t|
(p-value) Estimate
Standard
error t-ratio
Prob > |t|
(p-value)
15.585 0.3169 49.19 <0.0001 16.146 0.1412 114.38 <0.0001
2.612 0.1584 16.49 <0.0001 2.612 0.1578 16.55 <0.0001
0.197 0.1584 1.25 0.2320 - - - -
0.231 0.1584 1.46 0.1648 - - - -
0.374 0.1584 2.36 0.0320 0.374 0.1578 2.37 0.025
-0.111 0.1940 -0.57 0.5762 - - - -
0.333 0.1940 1.72 0.1064 - - - -
-0.042 0.1940 -0.22 0.8322 - - - -
0.013 0.1940 0.07 0.9463 - - - -
0.014 0.1940 0.07 0.9419 - - - -
0.060 0.1940 0.31 0.7624 - - - -
0.101 0.1482 0.68 0.5070 - - - -
0.235 0.1482 1.58 0.1339 - - - -
0.234 0.1482 1.58 0.1358 - - - -
0.132 0.1482 0.89 0.3856 - - - -
R2
R2
R2
adjusted RMSE Mean
0.9507 0.9047 0.776 16.15 0.9119 0.9054 0.773 16.15
ANOVA DOF
Sum of
Squares
(SS-)
Mean
Square
(MS-) FANOVA
Prob > F
(p-value) DOF
Sum of
Squares
(SS-)
Mean
Square
(MS-) FANOVA
Prob > F
(p-value)
Regression (R) 14 174.18 12.44 20.65 <0.0001 2 167.07 83.53 139.72 <0.0001
Error (E) 15 9.04 0.60 27 16.14 0.598
Total (T) 29 183.21 29 183.21
Sum of
Squares
Mean
Square FLOF
Prob > F
(p-value) DOF
Sum of
Squares
Mean
Square FLOF
Prob > F
(p-value)
Lack of fit 10 7.491 0.749 2.43 0.170 6 5.57 0.93 1.85 0.1384
Pure error 5 1.545 0.309 21 10.57 0.50
Total 15 9.036 27 16.14
Intercept
HC temperature
N2 flow rate
Auger speed
HC feed rate

221
Table 71. Solids content analytical data
SF4
Run
No.
DOE
No. Avg. St. dev. Avg. St. dev. Avg. St. dev. Avg. St. dev.
12 28 1.142 0.075 1.0378 0.073 1.198 0.077 1.465 0.080 nd
15 29 1.237 0.249 1.3653 0.230 1.370 0.257 0.6425 0.320 nd
17 27 0.838 0.255 1.164 0.103 0.846 0.569 0.1695 0.063 0.2071
19 25 0.810 0.312 1.0359 0.295 0.715 0.380 0.5662 0.262 0.1099
21 26 0.655 0.196 0.821 0.120 0.561 0.226 0.4868 0.316 0.6328
22 30 0.957 0.168 0.9764 0.189 1.073 0.181 0.8431 0.117 nd
0.940 0.209 1.0667 0.168 0.9605 0.282 0.6955 0.193 -
- 0.222 - 0.185 - 0.323 - 0.222 -
SF3
Average
Whole bio-oil SF1 SF2
St. Dev. a
Notes: All values in %-wt., wb. Each analysis performed in triplicate except for SF4.
a- Pooled standard deviation
Table 72. Higher heating value analytical data
Run
No.
DOE
No. Avg. St. dev. Avg. St. dev. Avg. St. dev. Avg. St. dev. Avg. St. dev.
12 28 16.4 nd -
17 27 16.17 0.23 18.80 0.27 11.95 0.13 18.86 0.30 5.98 0.23
21 26 16.57 0.10 19.16 0.11 12.63 0.00 19.03 0.25 6.94 0.09
16.41 0.15 18.72 0.15 12.14 0.14 19.23 0.19 6.46 0.16
- 0.16 - 0.18 - 0.19 - 0.23 - 0.18
20 24 17.12 - 19.22 - 13.25 - 19.45 - 6.97 -
13 23 13.64 - 14.50 - 8.11 - 19.67 - 5.38 -
SF4
Average
Notes: nd - Not determined. All values in (MJ/kg) on a wet basis. Analyses with standard deviations
performed in duplicate. Run No. 12 whole bio-oil average calculated without HHV contribution from SF4.
a - Pooled standard deviation
St. Dev.a
SF2 SF3
Whole bio-oil SF1
7 0.13 18.19 0.08 11.85 0.30 19.82 0.02
Table 73. Thermal Gravimetric Analysis data, bio-oil
Run
No.
DOE
No. M V FC A M V FC A M V FC A M V FC A
12 28 36.9 50.8 13.3 0.045 68.2 25.0 6.9 0.065 28.0 57.9 14.2 0.059 98.0 1.55 0.36 0.043
15 29 35.2 50.6 14.1 0.056 67.8 25.1 7.1 0.028 24.1 59.8 16.2 0.144 98.1 1.53 0.37 0.011
17 27 30.8 54.6 14.6 0.032 67.5 25.5 7.1 0.051 33.7 52.8 13.5 0.001 98.2 1.56 0.20 0.040
19 25 29.3 55.5 15.2 0.029 68.1 24.8 7.2 0.060 31.1 54.7 14.0 0.142 98.6 1.43 0.04 0.063
21 26 27.3 57.6 15.2 0.028 66.8 26.3 6.9 0.072 30.4 53.4 16.2 0.010 98.6 1.24 0.11 0.021
22 30 32.4 53.2 14.4 0.034 67.5 25.6 6.9 0.062 31.0 53.6 15.4 0.031 98.9 1.02 0.12 0.011
32.0 53.7 14.5 0.037 67.7 25.4 7.0 0.056 29.7 55.4 14.9 0.064 98.4 1.39 0.20 0.032
3.63 2.73 0.71 0.011 0.52 0.55 0.12 0.015 3.29 2.81 1.18 0.064 0.34 0.22 0.14 0.021
Avg.
St. Dev.
Notes: All values in %-wt., wb. M - Moisture, V - Volatiles, FC - Fixed Carbon, A - Ash.
Stadard deviation shown among runs, not replicates.
SF1 SF2 SF3 SF4

222
Table 7 iochar
4. Thermal Gravimetric Analysis data, b
Run
No.
DOE
No. Moisture Volatiles Fixed Carbon Ash
1 17 4.80 46.92 45.00 3.28
2 18 4.77 27.01 63.82 4.42
3 15 4.33 30.78 57.75 7.15
4 7 4.76 27.38 60.62 7.25
5 9 4.15 38.01 49.09 8.77
6 13 4.50 34.18 51.67 9.64
7 11 4.68 40.28 48.06 6.98
8 5 4.85 29.13 59.97 6.04
9 1 4.74 31.30 58.81 5.16
10 3 4.82 29.83 60.15 5.19
11 21 4.80 28.80 62.86 3.53
12 28 3.75 31.53 58.33 6.40
13 23 4.23 26.70 56.90 12.17
14 19 5.34 33.74 53.90 7.03
15 29 4.39 31.71 59.21 4.70
16 20 4.13 33.23 54.70 7.93
17 27 4.43 35.49 53.86 6.07
18 22 4.51 38.16 52.22 5.11
19 25 4.43 31.20 58.93 5.45
20 24 5.10 27.62 62.22 5.07
21 26 5.18 26.82 62.80 5.21
22 30 3.65 36.24 54.90 5.21
23 8 5.24 33.01 57.86 3.89
24 2 4.93 31.14 58.47 5.46
25 12 4.72 28.79 57.56 8.94
26 6 4.24 36.85 54.36 4.56
27 10 4.53 31.50 52.71 11.26
28 14 4.52 29.90 55.18 10.40
29 4 4.58 29.70 60.59 5.11
30 16 4.35 24.47 63.01 8.20
4.30 32.16 58.00 5.51
0.56 3.40 3.23 0.63
MAX 5.34 46.92 63.82 12.17
MIN 3.65 24.47 45.00 3.28
Ov
Ov
Note: All values in %-wt., wb
Cntr. Pt. Avg.
Cntr. Pt. St. Dev.
4.58 32.05 56.85 6.52
0.39 4.75 4.68 2.28
erall Avg.
erall St. Dev.

223
Table 75. Elemental analysis data, biochar
Run
No.
DOE
No. C
St.
Dev. N
St.
Dev. H
St.
Dev. S
St.
Dev. Ash Oa
1 17 67.01 - 0.261 - 4.72 - 0.023 - 3.28 24.71
2 18 75.45 - 0.247 - 3.94 - 0.012 - 4.42 15.93
3 15 72.22 - 0.302 - 3.82 - 0.012 - 7.15 16.50
4 7 70.88 - 0.177 - 3.55 - 0.014 - 7.25 18.13
5 9 65.81 - 0.206 - 3.99 - 0.012 - 8.77 21.22
6 13 69.56 - 0.326 - 3.65 - 0.015 - 9.64 16.81
7 11 65.47 - 0.160 - 4.23 - 0.014 - 6.98 23.15
8 5 71.67 - 0.432 - 3.71 - 0.018 - 6.04 18.12
9 1 70.62 - 0.456 - 3.78 - 0.018 - 5.16 19.97
10 3 73.70 - 0.194 - 3.55 - 0.014 - 5.19 17.35
11 21 73.74 - 0.300 - 3.75 - 0.016 - 3.53 18.67
12 28 73.27 0.121 0.149 0.038 3.47 0.058 0.007 0.004 6.40 16.70
13 23 68.98 - 0.199 - 3.02 - 0.027 - 12.17 15.61
14 19 70.42 - 0.280 - 3.83 - 0.018 - 7.03 18.43
15 29 71.41 0.273 0.063 0.057 3.75 0.019 0.012 0.005 4.70 20.07
16 20 68.44 - 0.140 - 3.93 - 0.016 - 7.93 19.55
17 27 68.74 0.138 0.162 0.106 3.78 0.316 0.007 9E-04 6.07 21.24
18 22 67.04 - 0.404 - 4.31 - 0.014 - 5.11 23.12
19 25 70.03 0.174 0.134 0.032 3.54 0.054 0.022 0.0197 5.45 20.83
20 24 73.58 - 0.182 - 3.55 - 0.015 - 5.07 17.60
21 26 72.70 1.230 0.114 0.056 3.36 0.065 0.011 0.003 5.21 18.60
22 30 68.93 0.021 0.051 0.042 3.94 0.011 0.012 0.005 5.21 21.85
23 8 70.83 - 0.386 - 4.08 - 0.017 - 3.89 20.80
24 2 71.33 - 0.172 - 3.67 - 0.015 - 5.46 19.35
25 12 72.34 - 0.382 - 3.15 - 0.015 - 8.94 15.17
26 6 69.19 - 0.150 - 4.15 - 0.013 - 4.56 21.93
27 10 66.57 - 0.362 - 3.42 - 0.021 - 11.26 18.37
28 14 68.97 - 0.187 - 3.17 - 0.019 - 10.40 17.26
29 4 71.98 - 0.229 - 3.76 - 0.017 - 5.11 18.90
30 16 74.38 - 0.264 - 3.13 - 0.016 - 8.20 14.01
70.51 - 0.24 - 3.72 - 0.015 - 6.52 19.00
2.61 - 0.11 - 0.37 - 0.004 - 2.28 2.56
70.85 0.326 0.11 0.055 3.64 0.087 0.0117 0.006 5.51 19.88
- 0.734 - 0.082 - 0.192 - 0.012 - -
MAX 75.45 - 0.46 - 4.72 - 0.03 - 12.17 24.71
MIN 65.47 - 0.05 - 3.02 - 0.01 - 3.28 14.01
Overall Avg.
Overall St. Dev.
Cntr. Pt. Avg
Cntr. Pt. St. Dev.b
Notes: All values in %-wt., wb. a - Oxygen by difference. b- Pooled standard deviation

224
Table 76. Elemental analysis data, SF1 bio-oil
Run
No.
DOE
No. C
St.
Dev. Na
St.
Dev. H
St.
Dev. S
St.
Dev. Ash Ob
1 17 43.51 0.155 0.048 0.0091 7.26 0.098 0.007 - 0.029 49.14
2 18 45.97 0.152 0.008 0 7.01 0.041 0.004 - 0.303 46.71
3 15 39.30 0.141 0.008 0 7.49 0.015 0.006 - 0.011 53.19
4 7 45.51 0.315 0.008 0 7.02 0.031 0.003 - 0.008 47.45
5 9 39.36 0.053 0.008 0 7.48 0.030 0.003 - 0.252 52.90
6 13 39.35 0.159 0.008 0 7.50 0.031 0.007 - 0.075 53.07
7 11 44.84 0.113 0.008 0 6.99 0.031 0.004 - 0.094 48.07
8 5 43.48 0.338 0.008 0 7.03 0.060 0.002 - 0.131 49.35
9 1 47.54 - 0.046 - 6.87 - 0.003 - 0.131 45.41
10 3 47.96 - 0.038 - 6.86 - 0.001 - 0.23 44.92
11 21 43.20 - 0.030 - 7.17 - 0.006 - 0.143 49.45
12 28 42.79 0.141 0.008 0.0003 7.02 0.055 0.010 0.0024 0.045 50.13
13 23 34.71 - 0.023 - 7.82 - 0.001 - 1.089 56.35
14 19 39.76 - 0.008 - 7.52 - 0.002 - 0.640 52.07
15 29 43.10 0.385 0.008 0 7.19 0.063 0.008 0.0013 0.056 49.64
16 20 43.42 - 0.009 - 7.18 - 0.002 - 0.715 48.67
17 27 45.22 0.319 0.008 0 6.94 0.016 0.007 0.0010 0.032 47.79
18 22 44.93 - 0.047 - 6.99 - 0.001 - - 48.03
19 25 45.82 0.201 0.126 0.0743 7.03 0.127 0.003 0.0021 0.029 47.00
20 24 45.98 - 0.035 - 6.97 - 0.000 - - 47.02
21 26 46.08 0.384 0.008 0 6.78 0.069 0.004 0.0017 0.028 47.10
22 30 44.29 0.342 0.111 0.0659 7.07 0.052 0.003 0.0005 0.034 48.49
23 8 46.81 - 0.028 - 6.84 - 0.006 - 0.992 45.33
24 2 43.52 - 0.023 - 7.11 - 0.004 - 0.487 48.86
25 12 38.09 - 0.044 - 7.59 - 0.004 - 0.966 53.30
26 6 41.82 - 0.107 - 7.32 - 0.007 - 0.900 49.85
27 10 41.30 - 0.054 - 7.28 - 0.006 - 0.887 50.48
28 14 40.36 - 0.116 - 7.14 - 0.005 - 0.996 51.39
29 4 43.77 - 0.137 - 7.17 - 0.008 - 0.894 48.02
30 16 40.33 - 0.008 - 7.39 - 0.007 - 0.897 51.37
43.07 - 0.038 - 7.17 - 0.005 - 0.396 49.35
3.10 - 0.040 - 0.25 - 0.002 - 0.401 2.70
44.55 0.295 0.045 0.0234 7.01 0.064 0.006 0.0015 0.04 48.36
- 0.409 - 0.0574 - 0.098 - 0.0023 0.01 1.31
MAX 47.96 - 0.137 - 7.82 - 0.010 - 1.09 56.35
MIN 34.71 - 0.008 - 6.78 - 0.000 - 0.01 44.92
Cntr. Pt. Avg
Cntr. Pt. St. Dev.c
Notes: All values in %-wt., wb. a - Minimum detection level = 80 PPM (if St. Dev = 0, triplicate
samples were all below detection limit). b - Oxygen by difference. c - Pooled standard deviation,
except for ash and O which are shown as STDEV among runs
Overall Avg.
Overall St. Dev.

225
Run
No.
DOE
No. C
St.
Dev. Na
St.
Dev. H
St.
Dev. S
St.
Dev. Ash Ob
1 17 27.22 0.126 0.011 0.0094 8.55 0.043 0.007 - 0.044 64.17
2 18 29.99 0.025 0.008 0 8.20 0.082 0.005 - 0.879 60.92
3 15 27.21 0.477 0.008 0 8.47 0.059 0.004 - 0.082 64.23
4 7 30.14 0.017 0.008 0 8.19 0.015 0.004 - - 61.65
5 9 27.65 0.068 0.008 0 8.43 0.034 0.006 - 0.020 63.89
6 13 27.24 0.133 0.008 0 8.38 0.040 0.006 - 0.152 64.21
7 11 27.54 0.026 0.008 0 8.34 0.042 0.007 - 0.471 63.64
8 5 30.57 0.164 0.008 0 8.22 0.048 0.009 - 0.178 61.02
9 1 31.13 - 0.008 - 8.05 - 0.003 - 0.051 60.76
10 3 30.41 - 0.008 - 8.10 - 0.009 - 0.143 61.33
11 21 27.42 - 0.008 - 8.48 - 0.003 - - 64.10
12 28 27.30 0.057 0.008 0 8.23 0.047 0.008 0.0012 0.065 64.39
13 23 20.18 - 0.008 - 8.56 - 0.004 - 0.176 71.07
14 19 27.78 - 0.008 - 8.38 - 0.008 - 0.395 63.43
15 29 27.59 0.067 0.008 0 8.40 0.090 0.007 0.0007 0.028 63.97
16 20 26.81 - 0.008 - 8.38 - 0.006 - 0.040 64.76
17 27 28.01 0.064 0.008 0 8.05 0.039 0.008 0.0011 0.051 63.88
18 22 28.54 - 0.008 - 8.32 - 0.006 - 0.672 62.46
19 25 28.27 0.148 0.063 0.0598 8.36 0.045 0.003 0.0023 0.060 63.25
20 24 30.79 - 0.008 - 8.04 - 0.006 - - 61.15
21 26 28.76 0.180 0.020 0.0357 8.07 0.051 0.004 0.0027 0.072 63.08
22 30 27.80 0.112 0.102 0.0484 8.35 0.057 0.005 0.0009 0.062 63.67
23 8 28.79 - 0.008 - 8.24 - 0.007 - 0.907 62.05
24 2 29.25 - 0.008 - 8.20 - 0.005 - - 62.54
25 12 25.05 - 0.008 - 8.58 - 0.006 - 0.645 65.72
26 6 28.11 - 0.008 - 8.26 - 0.008 - 0.956 62.66
27 10 23.85 - 0.008 - 8.63 - 0.010 - - 67.50
28 14 24.16 - 0.008 - 8.62 - 0.009 - - 67.20
29 4 29.46 - 0.012 - 7.75 - 0.005 - - 62.77
30 16 24.87 - 0.008 - 8.60 - 0.009 - 0.821 65.68
27.73 - 0.014 - 8.31 - 0.006 - 0.303 63.70
2.33 - 0.020 - 0.21 - 0.002 - 0.334 2.20
27.95 0.105 0.035 0.0240 8.24 0.055 0.006 0.0015 0.056 63.71
- 0.160 - 0.0489 - 0.072 - 0.0023 0.015 0.485
MAX 31.13 - 0.102 - 8.63 - 0.010 - 0.96 71.07
MIN 20.18 - 0.008 - 7.75 - 0.003 - 0.02 60.76
Overall Avg.
Overall St. Dev.
Cntr. Pt. Avg
Cntr. Pt. St. Dev.c

226
Run
No.
DOE
No. C
St.
Dev. Na
St.
Dev. H
St.
Dev. S
St.
Dev. Ash Ob
1 17 46.31 0.292 0.048 0.0041 7.14 0.130 0.002 - 0.035 46.47
2 18 46.16 0.197 0.008 0 7.07 0.102 0.004 - 0.064 46.70
3 15 47.21 0.249 0.008 0 7.18 0.046 0.005 - 0.394 45.21
4 7 47.11 0.324 0.008 0 7.13 0.040 0.005 - 0.691 45.06
5 9 45.36 0.117 0.008 0 7.20 0.021 0.004 - 0.012 47.41
6 13 46.85 0.149 0.008 0 7.13 0.008 0.007 - 0.134 45.87
7 11 43.71 0.045 0.008 0 7.33 0.030 0.001 - 0.134 48.81
8 5 48.24 0.028 0.008 0 7.04 0.019 0.001 - 0.138 44.58
9 1 44.01 - 0.041 - 7.25 - 0.001 - 0.185 48.51
10 3 43.48 - 0.008 - 7.25 - 0.003 - - 49.26
11 21 48.14 - 0.015 - 7.10 - 0.002 - 0.203 44.54
12 28 46.03 0.176 0.008 0.0008 7.05 0.069 0.005 0.0018 0.059 46.84
13 23 47.11 - 0.038 - 7.19 - 0.003 - - 45.66
14 19 46.85 - 0.031 - 7.15 - 0.007 - 0.762 45.19
15 29 48.27 0.121 0.008 0 7.02 0.033 0.006 0.0009 0.144 44.55
16 20 46.57 - 0.030 - 7.12 - 0.001 - 0.820 45.46
17 27 45.31 0.249 0.008 0 7.07 0.078 0.007 0.0006 0.001 47.60
18 22 42.49 - 0.013 - 7.34 - 0.001 - 0.881 49.28
19 25 44.94 0.201 0.207 0.0683 7.14 0.141 0.002 0.0017 0.142 47.57
20 24 45.99 - 0.028 - 7.17 - 0.007 - - 46.80
21 26 44.60 0.296 0.042 0.0206 7.15 0.031 0.001 0.0004 0.010 48.20
22 30 44.89 0.032 0.162 0.1118 7.24 0.098 0.003 0.0025 0.031 47.67
23 8 44.75 - 0.028 - 7.22 - 0.001 - 0.531 47.47
24 2 44.39 - 0.013 - 7.25 - 0.004 - 0.363 47.98
25 12 45.37 - 0.009 - 7.28 - 0.002 - 0.861 46.48
26 6 46.73 - 0.065 - 7.10 - 0.003 - 0.642 45.46
27 10 44.71 - 0.020 - 7.34 - 0.002 - 0.827 47.10
28 14 45.57 - 0.039 - 7.23 - 0.002 - - 47.16
29 4 43.76 - 0.038 - 7.32 - 0.014 - - 48.87
30 16 45.26 - 0.008 - 7.19 - 0.003 - 0.828 46.71
45.67 - 0.032 - 7.18 - 0.004 - 0.356 46.82
1.47 - 0.045 - 0.09 - 0.003 - 0.330 1.43
45.67 0.179 0.072 0.0336 7.11 0.075 0.004 0.0013 0.064 47.07
- 0.277 - 0.0765 - 0.118 - 0.0021 0.064 1.3
MAX 48.27 - 0.207 - 7.34 - 0.01 - 0.88 49.28
MIN 42.49 - 0.008 - 7.02 - 0.00 - 0.00 44.54
Overall Avg.
Overall St. Dev.
Cntr. Pt. Avg.
Cntr. Pt. St. Dev.c

227
Run
No.
DOE
No. C
St.
Dev. Na
St.
Dev. H
St.
Dev. S
St.
Dev. Ash Ob
1 17 12.46 0.050 0.008 0 10.04 0.039 0.014 - 0.043 77.43
2 18 13.62 0.143 0.008 0 9.96 0.082 0.016 - 0.051 76.34
3 15 13.22 0.342 0.008 0 9.77 0.341 0.021 - 0.008 76.97
4 7 13.82 0.328 0.008 0 9.68 0.305 0.011 - 0.702 75.78
5 9 11.98 0.220 0.008 0 9.88 0.117 0.014 - 0.368 77.75
6 13 14.06 0.171 0.008 0 9.52 0.143 0.013 - 0.332 76.07
7 11 14.02 0.082 0.008 0 9.73 0.117 0.013 - - 76.23
8 5 11.83 0.979 0.008 0 8.96 0.831 0.012 - - 79.18
9 1 13.86 - 0.008 - 9.75 - 0.012 - 0.566 75.81
10 3 13.10 - 0.008 - 9.58 - 0.010 - - 77.30
11 21 11.63 - 0.008 - 9.96 - 0.010 - - 78.39
12 28 12.43 0.499 0.008 0 8.64 0.153 0.017 0.0013 0.011 78.90
13 23 6.71 - 0.008 - 5.85 - 0.002 - - 87.42
14 19 12.96 - 0.008 - 9.95 - 0.014 - - 77.07
15 29 12.03 0.127 0.104 0.0977 10.00 0.132 0.013 0.0008 - 77.85
16 20 10.60 - 0.008 - 9.19 - 0.004 - - 80.19
17 27 9.73 0.750 0.009 0.0082 7.66 0.683 0.015 0.0008 0.040 82.54
18 22 12.51 - 0.008 - 9.57 - 0.017 - 0.029 77.87
19 25 12.89 0.197 0.158 0.1303 9.94 0.175 0.009 0.0009 0.063 76.94
20 24 13.84 - 0.008 - 9.73 - 0.012 - - 76.41
21 26 10.32 2.073 0.008 0 7.91 1.104 0.008 0.0015 0.021 81.74
22 30 12.88 0.136 0.131 0.1264 9.92 0.153 0.008 0.0016 0.011 77.05
23 8 13.97 - 0.008 - 9.81 - 0.007 - - 76.22
24 2 13.18 - 0.008 - 9.92 - 0.011 - - 76.88
25 12 11.26 - 0.008 - 10.20 - 0.012 - - 78.53
26 6 11.74 - 0.008 - 10.03 - 0.012 - - 78.21
27 10 10.64 - 0.008 - 9.58 - 0.011 - - 79.76
28 14 10.71 - 0.008 - 9.82 - 0.013 - - 79.45
29 4 13.25 - 0.008 - 9.93 - 0.013 - - 76.80
30 16 11.05 - 0.008 - 9.77 - 0.014 - - 79.16
12.21 - 0.02 - 9.47 - 0.012 - 0.173 78.21
1.61 - 0.04 - 0.90 - 0.004 - 0.24 2.41
11.71 0.631 0.07 0.0604 9.01 0.400 0.011 0.0011 - 79.17
- 1.314 - 0.1123 - 0.768 - 0.0016 - 2.42
MAX 14.06 - 0.158 - 10.20 - 0.021 - 0.70 87.42
MIN 6.71 - 0.008 - 5.85 - 0.002 - 0.01 75.78
except for O which is shown as STDEV among runs
Overall Avg.
Overall St. Dev.
Cntr. Pt. Avg.
Cntr. Pt. St. Dev.c

228
Table 80. Elemental analysis data, whole bio-oil
Run
No.
DOE
No. C
St.
Dev. Na
St.
Dev. H
St.
Dev. S
St.
Dev. Ash Ob
1 17 38.44 0.169 0.036 0.0081 7.69 0.086 0.006 - 0.035 53.79
2 18 40.22 0.119 0.008 0 7.46 0.067 0.005 - 0.440 51.87
3 15 38.14 0.221 0.008 0 7.64 0.032 0.006 - 0.087 54.12
4 7 40.57 0.225 0.008 0 7.45 0.032 0.004 - 0.145 51.82
5 9 37.87 0.073 0.008 0 7.65 0.031 0.004 - 0.168 54.30
6 13 38.00 0.153 0.008 0 7.63 0.031 0.007 - 0.103 54.26
7 11 37.91 0.068 0.008 0 7.59 0.037 0.005 - 0.230 54.25
8 5 40.79 0.256 0.008 0 7.34 0.064 0.003 - 0.140 51.72
9 1 40.59 - 0.031 - 7.41 - 0.003 - 0.126 51.85
10 3 40.13 0.020 - 7.44 - 0.005 - 0.135 52.27
11 21 39.10 - 0.022 - 7.56 - 0.005 - 0.109 53.20
12 28 38.63 0.127 0.008 0.0003 7.39 0.056 0.009 0.0018 0.059 53.91
13 23 32.75 - 0.022 - 7.88 - 0.002 - 0.037 59.31
14 19 38.38 - 0.012 - 7.65 - 0.004 - 0.606 53.35
15 29 38.85 0.245 0.009 0.0012 7.56 0.068 0.007 0.0010 0.041 53.53
16 20 39.03 - 0.013 - 7.53 - 0.003 - 0.543 52.88
17 27 38.53 0.226 0.008 0.0002 7.37 0.050 0.007 0.0009 0.032 54.05
18 22 38.60 - 0.026 - 7.54 - 0.003 - 0.409 53.42
19 25 38.80 0.182 0.121 0.0692 7.58 0.102 0.003 0.0021 0.042 53.45
20 24 40.49 - 0.025 - 7.41 - 0.004 - 0.000 52.08
21 26 38.91 0.331 0.019 0.0093 7.34 0.077 0.003 0.0018 0.038 53.69
22 30 39.10 0.211 0.119 0.0712 7.53 0.065 0.004 0.0010 0.029 53.23
23 8 39.32 - 0.020 - 7.47 - 0.006 - 0.863 52.32
24 2 39.71 - 0.017 - 7.46 - 0.004 - 0.334 52.47
25 12 36.30 - 0.029 - 7.79 - 0.004 - 0.861 55.02
26 6 38.89 - 0.075 - 7.55 - 0.007 - 0.853 52.62
27 10 35.60 - 0.031 - 7.78 - 0.007 - 0.569 56.01
28 14 35.19 - 0.065 - 7.70 - 0.006 - 0.477 56.56
29 4 39.16 - 0.078 - 7.43 - 0.009 - 0.432 52.90
30 16 36.22 - 0.008 - 7.74 - 0.007 - 0.849 55.17
38.47 - 0.03 - 7.55 - 0.005 - 0.29 53.65
1.77 - 0.03 - 0.14 - 0.002 - 0.29 1.62
38.80 0.221 0.05 0.0252 7.46 0.070 0.006 0.0014 0.040 53.64
- 0.308 - 0.0576 - 0.098 - 0.0021 0.010 0.305
MAX 40.79 - 0.121 - 7.88 - 0.009 - 0.86 59.31
MIN 32.75 - 0.008 - 7.34 - 0.002 - 0.00 51.72
Overall Avg.
Overall St. Dev.
Cntr. Pt. Avg.
Cntr. Pt. St. Dev.c

229
-0.8
-0.6
-0.4
-0.2
0.0
0.2
0.4
0.6
Residual
(C
content
%-wt.,
wb)
0 10 20 30
Run number
Figure 144. Residuals for bio-oil carbon content full model
Table 81. Bio-oil carbon content model statistical data
Term Estimate
Standard
error t-ratio
Prob > |t|
(p-value) Estimate
Standard
error t-ratio
Prob > |t|
(p-value)
38.804 0.1731 224.22 <0.0001 38.920 0.0996 390.63 <0.0001
1.642 0.0865 18.98 <0.0001 1.642 0.0863 19.03 <0.0001
-0.036 0.0865 -0.41 0.686 - - - -
0.008 0.0865 0.10 0.9238 - - - -
0.715 0.0865 8.26 <0.0001 0.715 0.0863 8.28 <0.0001
-0.008 0.1060 -0.08 0.939 - - - -
0.170 0.1060 1.60 0.1306 - - - -
0.104 0.1060 0.98 0.3425 - - - -
-0.226 0.1060 -2.13 0.0499 -0.226 0.1057 -2.14 0.0424
-0.134 0.1060 -1.27 0.225 - - - -
0.130 0.1060 1.23 0.2376 - - - -
-0.543 0.0809 -6.70 <0.0001 -0.557 0.0788 -7.07 <0.0001
0.016 0.0809 0.20 0.846 - - - -
-0.020 0.0809 -0.25 0.807 - - - -
0.135 0.0809 1.67 0.1151 - - - -
R2
R2
R2
adjusted RMSE Mean
0.9704 0.9429 0.4293 38.47 0.9510 0.9432 0.4227 38.47
ANOVA DOF
Sum of
Squares
(SS-)
Mean
Square
(MS-) FANOVA
Prob > F
(p-value) DOF
Sum of
Squares
(SS-)
Mean
Square
(MS-) FANOVA
Prob > F
(p-value)
Regression (R) 14 88.513 0.211 35.18 <0.0001 4 86.741 21.685 121.36 <0.0001
Error (E) 15 2.695 0.021 25 4.467 0.179
Total (T) 29 91.208 29 91.208
Sum of
Squares
Mean
Square FLOF
Prob > F
(p-value) DOF
Sum of
Squares
Mean
Square FLOF
Prob > F
(p-value)
Lack of fit 10 2.491 0.249 6.10 0.030 4 2.442 0.611 6.33 0.0017
Pure error 5 0.204 0.041 21 2.025 0.096
Total 15 2.695 25 4.467
Intercept
HC temperature
N2 flow rate
Auger speed
HC feed rate

230
-0.10
-0.05
0.00
0.05
0.10
Residual
(H
content,
%-wt.,
wb)
0 10 20 30
Run number
F
T
igure 145. Residuals for bio-oil hydrogen content full model
able 82. Bio-oil hydrogen content model statistical data
Term Estimate
Standard
error t-ratio
Prob > |t|
(p-value) Estimate
Standard
error t-ratio
Prob > |t|
(p-value)
7.459 0.0305 244.20 <0.0001 7.523 0.0169 445.81 <0.0001
-0.122 0.0153 -7.98 <0.0001 -0.122 0.0146 -8.34 <0.0001
-0.001 0.0153 -0.06 0.9508 - - - -
-0.011 0.0153 -0.73 0.4743 - - - -
-0.051 0.0153 -3.36 0.0043 -0.051 0.0146 -3.52 0.0016
-0.011 0.0187 -0.56 0.5806 - - - -
-0.001 0.0187 -0.06 0.9502 - - - -
0.008 0.0187 0.42 0.6773 - - - -
0.014 0.0187 0.76 0.4599 - - - -
0.003 0.0187 0.15 0.8825 - - - -
-0.010 0.0187 -0.54 0.5985 - - - -
0.043 0.0143 2.98 0.0094 -0.035 0.0133 -2.59 0.0156
0.019 0.0143 1.33 0.2027 - - - -
0.028 0.0143 1.98 0.0664 - - - -
0.029 0.0143 2.03 0.1017 - - - -
R2
R2
R2
adjusted RMSE Mean
0.8571 0.7236 0.0748 7.55 0.7731 0.7469 0.0716 7.55
ANOVA DOF
Sum of
Squares
(SS-)
Mean
Square
(MS-) FANOVA
Prob > F
(p-value) DOF
Sum of
Squares
(SS-)
Mean
Square
(MS-) FANOVA
Prob > F
(p-value)
Regression (R) 14 0.503 0.036 6.42 <0.0005 3 0.454 0.151 29.53 <0.0001
Error (E) 15 0.084 0.006 26 0.133 0.005
Total (T) 29 0.587 29 0.587
Sum of
Squares
Mean
Square FLOF
Prob > F
(p-value) DOF
Sum of
Squares
Mean
Square FLOF
Prob > F
(p-value)
Lack of fit 10 0.026 0.003 0.22 0.980 5 0.015 0.003 0.53 0.7516
Pure error 5 0.058 0.012 21 0.118 0.006
Total 15 0.084 26 0.133
Intercept
HC temperature
N2 flow rate
Auger speed
HC feed rate

231
7.3
7.4
7.5
7.6
7.7
7.8
7.9
Actual
hydrogen
content
(%-wt.,
whole
bio-oil)
7.3 7.4 7.5 7.6 7.7 7.8 7.9 8
Predicted hydrogen content (%-wt., whole bio-oil)
Figure 146. Predicted vs. actual hydrogen content
-0.75
-0.50
-0.25
0.00
0.25
0.50
0.75
Residual
(O
content,
%-wt.
wb)
0 5 10 15 20 25 30
Run number
Figure 147. Residuals for bio-oil oxygen content full model

232
Table 83. Bio-oil oxygen content model statistical data
Term Estimate
Standard
error t-ratio
Prob > |t|
(p-value) Estimate
Standard
error t-ratio
Prob > |t|
(p-value)
53.636 0.1681 319.10 <0.0001 53.648 0.1225 438.09 <0.0001
-1.521 0.0840 -18.09 <0.0001 -1.521 0.0866 -17.56 <0.0001
0.039 0.0840 0.47 0.6472 - - - -
0.004 0.0840 0.05 0.9617 - - - -
-0.688 0.0840 -8.19 <0.001 -0.688 0.0866 -7.95 <0.0001
0.010 0.1029 0.10 0.9235 - - - -
-0.170 0.1029 -1.65 0.1200 - - - -
-0.107 0.1029 -1.04 0.3147 - - - -
0.219 0.1029 2.13 0.0502 - - - -
0.134 0.1029 1.30 0.2133 - - - -
-0.112 0.1029 -1.09 0.2923 - - - -
0.527 0.0786 6.70 <0.0001 0.525 0.080 6.600 <0.0001
-0.009 0.0786 -0.11 0.9131 - - - -
0.021 0.0786 0.27 0.7935 - - - -
-0.193 0.0786 -2.46 0.0266 -0.195 0.080 -2.450 0.022
R2
R2
R2
adjusted RMSE Mean
0.9686 0.9394 0.41 53.91 0.9445 0.9356 0.42 53.91
ANOVA DOF
Sum of
Squares
(SS-)
Mean
Square
(MS-) FANOVA
Prob > F
(p-value) DOF
Sum of
Squares
(SS-)
Mean
Square
(MS-) FANOVA
Prob > F
(p-value)
Regression (R) 14 78.52 5.61 33.09 <0.0001 4 76.56 19.14 106.37 <0.0001
Error (E) 15 2.54 0.17 25 4.50 0.180
Total (T) 29 81.06 29 81.062
Lack of fit analysis DOF FLOF
n
Square FLOF
Prob > F
(p-value)
Lack of fit 10 2.054 0.205 2.10 0.213 4 2.37 0.59 5.85 0.0025
Pure error 5 0.488 0.098 21 2.13 0.10
Total 15 2.543 25 4.50
HC feed rate
Intercept
HC temperature
N2 flow rate
Auger speed
Sum of
Squares
Mean
Square
Prob > F
(p-value) DOF
Sum of
Squares
Mea
Table 84. Total acid number analytical data for center point tests
Run
No.
DOE
No. Avg. St. dev. Avg. St. dev. Avg. St. dev. Avg. St. dev. Avg. St. dev.
12 28 nd - 112.6 0.10 119.9 0.77 84.2 0.42 nd -
15 29 106.1 0.74 109.1 0.64 113.0 0.64 85.9 1.30 63.3 0.50
17 27 108.4 0.86 107.0 0.50 124.9 1.19 88.2 1.13 54.9 0.20
19 25 109.3 0.26 107.5 0.40 125.7 0.03 91.0 0.28 48.8 1.32
21 26 107.9 0.43 104.5 0.62 125.9 0.29 90.9 0.29 50.1 0.20
22 30 108.4 0.74 111.1 0.86 120.4 0.93 90.6 0.22 48.9 0.09
108.0 0.60 108.6 0.52 121.6 0.64 88.5 0.61 53.2 0.46
- 0.64 - 0.57 - 0.75 - 0.75 - 0.55
Average
Notes: nd - Not determined. All values in (mgKOH/gbio-oil). Analyses with standard deviations performed in
duplicate. a - Pooled standard deviation
St. Dev.a
Whole bio-oil SF1 SF2 SF3 SF4

233
1
5
2
(1) Acetic acid (2) 2-Butanone, 3-hydroxy (3) Furfual (4) 2H-Pyran-2-one (5) Phenol, 2-methoxy-4-methyl-
(6) Phenol, 2,6-dimethoxy- (7) 4 methyl 2,6 dimethoxy phenol (8) Levoglucosan
3
8
4
6
7
Figure 148. GC/MS chromatogram for SF2, Run #20 (bio-oil maximum yield)
1
7
6
5
4
2
(1) Acetic acid (2) 2-Butanone, 3-hydroxy (3) 1,2-Cyclopentanedione, 3-methyl
(4) Phenol, 2-methoxy-4-methyl (5) Phenol, 2,6-dimethoxy- (6) 4 methyl 2,6 dimethoxy phenol
(7) Levoglucosan (8) Ehtanone, 1-(4-hydroxy-3,5-dimethoxyphenyl)
8
3
Figure 149. GC/MS chromatogram for SF3, Run #20 (bio-oil maximum yield)

234
Table 85. GC/MS sample analytical data, Run #20 (maximum bio-oil yield)
Chemical compound SF1 SF2 SF3 SF4 Whole
Acetic acid 1.090 2.756 1.533 3.510 1.723
2-Propanone, 1-hydroxy- 1.468 1.578 1.933 1.200 1.584
2-Butanone, 3-hydroxy- 0.178 0.200 0.200 0.289 0.191
Furfural 0.044 0.111 0.044 0.156 0.067
2-Furanmethanol 0.200 0.156 0.178 0.000 0.179
2-Cyclopenten-1-one, 2-methyl- 0.022 0.044 0.022 0.022 0.029
2-Furancarboxaldehyde, 5-methyl- 0.067 0.067 0.067 0.000 0.066
2H-Pyran-2-one 0.000 0.156 0.133 0.000 0.073
1,2-Cyclopentanedione, 3-methyl- 0.645 0.444 0.533 0.000 0.552
2(5H)-Furanone, 3-methyl- 0.111 0.089 0.089 0.000 0.098
Phenol 0.044 0.044 0.044 0.000 0.044
Phenol, 2-methoxy- 0.556 0.444 0.489 0.111 0.502
Glycerin 0.000 0.000 1.133 0.000 0.212
Phenol, 2-methyl- 0.044 0.044 0.044 0.000 0.044
Phenol, 4-methyl- 0.067 0.067 0.067 0.000 0.066
Phenol, 3-methyl- 0.067 0.044 0.067 0.000 0.059
Phenol, 2-methoxy-4-methyl- 0.267 0.178 0.244 0.000 0.231
Phenol, 2,5-dimethyl- 0.044 0.044 0.044 0.000 0.0
2,4-Dimethylphenol 0.044 0.044 0.044 0.000 0.0
Phen
Phen
Phenol, 3,4-dimethyl- 0.0 4 0.044 0.044 0.000 0.044
Phenol, 4-ethyl-2-methoxy- 0.111 0.089 0.111 0.000 0.103
Eugenol 0.178 0.133 0.156 0.000 0.157
2-Furancarboxaldehyde, 5-(hydroxymethyl) 0.356 0.000 0.333 0.000 0.237
Phenol, 2,6-dimethoxy- 1.134 0.511 1.067 0.000 0.912
Phenol, 2-methoxy-4-(1-propenyl)-, (E)- 0.400 0.133 0.356 0.000 0.303
4 methyl 2,6 dimethoxy phenol 0.912 0.356 0.889 0.111 0.724
Vanillin 0.489 0.356 0.489 0.000 0.440
Hydroquinone 0.133 0.067 0.111 0.000 0.107
1,6-Anhydro-β-D-glucopyranose 2.246 1.333 2.244 0.000 1.929
Ethanone, 1-(4-hydroxy-3,5-dimethoxyphenyl) 1.357 1.022 1.378 0.000 1.236
Sum 12.41 10.65 14.18 5.40 12.08
Bio-oil fraction
44
44
ol, 2-ethyl- 0.044 0.044 0.044 0.000 0.044
ol, 3-ethyl- 0.044 0.044 0.044 0.000 0.044
4

235
Table 86. GC/MS analytical data, SF1 summary
Run
No.
DOE
No.
Acetic
acid
Levoglucosan
Furans
Phenols
Guaiacols
Syringols
Other
GS/MS
Total
1 17 1.50 2.87 0.72 0.54 2.00 3.44 3.88 14.95
2 18 1.06 3.32 0.49 0.58 2.10 3.67 2.74 13.96
3 15 3.69 2.38 0.55 0.53 1.84 3.15 3.53 15.67
4 7 3.38 3.31 0.49 0.60 2.09 3.47 3.98 17.32
5 9 3.69 2.29 0.76 0.53 1.91 3.18 3.29 15.65
6 13 1.21 1.57 0.71 0.57 1.92 3.18 3.11 12.28
7 11 6.11 3.02 0.81 0.58 2.28 4.02 3.16 19.97
8 5 2.70 2.10 0.60 0.60 2.01 3.20 4.25 15.45
9 1 4.28 4.14 0.47 0.60 2.07 3.58 2.81 17.95
10 3 1.25 3.64 0.47 0.61 2.19 3.78 3.05 14.99
11 21 2.31 3.14 1.00 0.53 1.91 3.18 3.25 15.34
12 28 4.42 3.27 0.94 0.56 1.93 3.32 5.63 20.07
13 23 1.51 0.00 1.09 0.51 1.78 2.71 5.11 12.72
14 19 4.30 3.21 1.76 0.53 1.98 3.30 5.71 20.79
15 29 3.91 3.53 1.24 0.58 1.82 3.49 5.78 20.35
16 20 1.20 2.02 0.73 0.56 1.87 3.33 5.16 14.87
17 27 4.70 2.25 0.76 0.56 2.09 3.70 3.72 17.76
18 22 3.42 2.02 0.71 0.55 2.09 3.71 3.11 15.60
19 25 1.67 2.16 1.07 0.56 2.00 3.76 3.62 14.83
20 24 1.09 2.25 0.42 0.58 2.00 3.40 2.67 12.41
21 26 1.20 2.35 0.71 0.56 2.20 3.98 4.13 15.13
22 30 2.42 2.22 0.87 0.56 2.07 3.64 5.02 16.80
23 8 3.05 2.43 0.87 0.58 2.18 4.06 4.19 17.36
24 2 3.76 2.30 1.16 0.56 2.02 3.44 2.58 15.81
25 12 1.25 1.85 0.80 0.53 1.87 3.30 4.55 14.15
26 6 3.42 2.21 0.72 0.56 1.93 3.38 4.12 16.35
27 10 3.16 1.74 0.73 0.53 2.09 3.52 3.67 15.44
28 14 1.22 1.71 0.80 0.53 2.13 3.53 4.84 14.78
29 4 1.64 2.40 0.49 0.56 1.98 3.62 4.20 14.89
30 16 3.40 1.89 0.76 0.53 2.02 3.49 4.29 16.38
2.73 2.45 0. 9 0.56 2.01 3.48 3.97 16.00
1.37 0.80 0. 8 0.02 0.12 0.29 0.93 2.20
3.05 2.63 0. 3 0.56 2.02 3.65 4.65 17.49
1.49 0.61 0. 0 0.01 0.13 0.23 0.95 2.37
MAX 6.11 4.14 1 6 0.61 2.28 4.06 5.78 20.79
MIN 1.06 0.00 0. 2 0.51 1.78 2.71 2.58 12.28
Overall Avg.
Overall St. Dev.
Cntr. Pt. Avg.
Cntr. Pt. St. Dev.
7
2
9
2
.7
4

236
Run
No.
DOE
No.
Acetic
acid
Levoglucosan
Furans
Phenols
Guaiacols
Syringols
Other
GS/MS
Total
1 17 1.86 0.00 0.45 0.34 1.17 1.66 4.12 9.59
2 18 6.27 0.00 0.46 0.37 1.32 1.82 4.43 14.67
3 15 5.50 0.00 0.51 0.35 1.24 1.74 3.87 13.21
4 7 7.00 0.00 0.47 0.49 1.31 1.84 3.69 14.80
5 9 8.45 0.00 0.49 0.38 1.25 1.76 2.92 15.25
6 13 3.21 0.00 0.51 0.39 1.14 1.81 3.39 10.45
7 11 3.78 0.00 0.52 0.50 1.32 1.84 2.01 9.96
8 5 6.82 0.00 0.44 0.44 1.28 1.86 3.41 14.24
9 1 7.69 0.00 0.49 0.51 1.49 1.95 4.53 16.65
10 3 1.22 0.00 0.44 0.52 1.40 1.92 3.21 8.71
11 21 0.98 0.00 0.85 0.42 1.22 1.80 5.05 10.32
12 28 4.58 0.00 0.73 0.38 1.18 1.80 5.00 13.67
13 23 0.98 0.00 0.73 0.27 0.98 1.49 3.39 7.83
14 19 4.31 0.00 1.00 0.47 1.31 1.98 5.69 14.75
15 29 1.87 0.00 0.87 0.47 1.29 1.91 4.95 11.35
16 20 1.07 1.15 0.47 0.36 1.13 1.71 3.00 8.88
17 27 1.44 1.29 0.47 0.47 1.27 1.84 2.27 9.04
18 22 1.09 0.98 0.53 0.42 1.36 1.91 2.58 8.87
19 25 1.00 1.29 0.51 0.47 1.27 1.82 1.27 7.61
20 24 2.76 1.33 0.42 0.49 1.33 1.89 2.42 10.65
21 26 2.33 1.36 0.44 0.42 1.20 1.91 2.82 10.49
22 30 4.69 1.38 0.58 0.42 1.24 1.84 4.77 14.92
23 8 3.15 1.47 0.58 0.47 1.29 1.90 2.59 11.45
24 2 2.78 1.50 0.48 0.48 1.27 1.89 4.37 12.77
25 12 4.02 1.32 0.51 0.33 1.12 1.74 0.94 9.97
26 6 4.30 1.47 0.49 0.42 1.18 1.85 1.27 10.98
27 10 2.18 1.04 0.47 0.38 1.11 1.64 2.80 9.62
28 14 3.58 1.16 0.47 0.36 1.09 1.71 2.76 11.12
29 4 2.65 1.47 0.47 0.47 1.42 2.05 1.47 9.99
30 16 1.04 1.31 0.49 0.38 1.09 1.78 2.87 8.96
3.42 0.65 0.54 0.42 1.24 1.82 3.26 11.36
2.18 0.67 0.14 0.06 0.11 0.11 1.24 2.50
2.65 0.88 0.60 0.44 1.24 1.85 3.51 11.18
1.60 0.69 0.17 0.04 0.04 0.05 1.61 2.76
MAX 8.45 1.50 1.00 0.52 1.49 2.05 5.69 16.65
MIN 0.98 0.00 0.42 0.27 0.98 1.49 0.94 7.61
Overall Avg.
Overall St. Dev.
Cntr. Pt. Avg.
Cntr. Pt. St. Dev.

237
Run
No.
DOE
No.
Acetic
acid
Levoglucosan
Furans
Phenols
Guaiacols
Syringols
Other
GS/MS
Total
1 17 5.91 3.18 0.54 0.52 2.11 3.68 2.99 18.94
2 18 4.12 4.10 0.47 0.54 1.90 3.51 3.67 18.31
3 15 1.32 3.54 0.58 0.54 2.19 3.78 2.87 14.82
4 7 4.33 3.38 0.65 0.56 2.08 3.50 3.00 17.50
5 9 0.97 3.12 0.75 0.53 2.13 3.59 2.71 13.80
6 13 4.37 2.83 0.58 0.56 2.20 3.74 2.80 17.07
7 11 4.87 0.00 0.74 0.56 2.09 3.69 2.97 14.93
8 5 3.13 3.90 0.47 0.62 2.11 3.53 2.36 16.12
9 1 1.15 3.36 0.42 0.56 1.88 3.27 2.63 13.29
10 3 1.80 3.11 0.44 0.54 1.89 3.27 2.95 14.01
11 21 4.73 4.27 1.27 0.53 2.18 3.67 3.69 20.33
12 28 4.09 3.47 1.27 0.53 1.96 3.51 3.11 17.93
13 23 1.74 2.69 1.62 0.53 2.20 3.41 3.47 15.67
14 19 4.68 3.85 1.22 0.56 2.27 3.99 4.70 21.27
15 29 3.54 3.56 1.20 0.53 2.18 3.80 4.23 19.04
16 20 2.27 2.33 0.76 0.53 2.04 3.62 2.33 13.88
17 27 4.44 2.40 0.67 0.53 2.02 3.71 4.00 17.78
18 22 1.47 2.11 0.64 0.53 2.11 3.95 2.75 13.57
19 25 4.02 2.13 0.62 0.51 1.98 3.62 1.64 14.53
20 24 1.53 2.24 0.38 0.56 1.84 3.33 4.29 14.18
21 26 1.91 2.33 0.47 0.53 2.07 3.78 3.69 14.78
22 30 1.11 2.42 0.82 0.51 1.98 3.64 4.75 15.24
23 8 4.07 2.45 0.80 0.53 2.05 3.67 4.54 18.10
24 2 3.13 2.38 0.56 0.51 1.91 3.53 4.47 16.49
25 12 2.96 2.29 0.91 0.53 2.18 3.97 4.54 17.40
26 6 4.86 2.41 0.76 0.53 2.03 3.70 4.34 18.62
27 10 4.00 2.07 0.58 0.51 2.16 3.69 2.58 15.58
28 14 1.49 2.11 0.62 0.51 2.20 3.74 3.54 14.21
29 4 1.20 2.44 0.47 0.51 1.96 3.76 2.84 13.18
30 16 3.36 2.36 0.76 0.53 2.27 3.96 3.45 16.67
3.09 2.76 0.73 0.54 2.07 3.65 3.40 16.24
1.46 0.84 0.30 0.02 0.12 0.19 0.82 2.20
3.19 2.72 0.84 0.53 2.03 3.68 3.57 16.55
1.35 0.62 0.33 0.01 0.08 0.11 1.09 1.93
MAX 5.91 4.27 1.62 0.62 2.27 3.99 4.75 21.27
MIN 0.97 0.00 0.38 0.51 1.84 3.27 1.64 13.18
Overall Avg.
Overall St. Dev.
Cntr. Pt. Avg.
Cntr. Pt. St. Dev.

238
Run
No.
DOE
No.
Acetic
acid
Levoglucosan
Furans
Phenols
Guaiacols
Syringols
Other
GS/MS
Total
1 17 1.39 0.00 0.34 0.00 0.13 0.11 2.02 3.99
2 18 3.48 0.00 0.29 0.02 0.24 0.24 1.73 6.01
3 15 3.67 0.00 0.24 0.00 0.13 0.00 2.36 6.41
4 7 3.64 0.00 0.40 0.16 0.24 0.11 1.24 5.79
5 9 3.98 0.00 0.33 0.00 0.20 0.11 0.29 4.91
6 13 3.89 0.00 0.21 0.00 0.14 0.12 2.00 6.36
7 11 2.90 0.00 0.28 0.00 0.14 0.12 0.44 3.88
8 5 3.77 0.00 0.16 0.00 0.21 0.12 1.65 5.91
9 1 1.51 0.00 0.28 0.00 0.37 1.00 0.46 3.63
10 3 1.39 0.00 0.26 0.00 0.26 1.11 0.53 3.55
11 21 1.99 0.00 0.40 0.00 0.20 0.00 1.41 4.00
12 28 2.87 0.00 0.36 0.00 0.13 0.11 1.66 5.13
13 23 2.69 0.00 0.62 0.00 0.11 0.00 0.24 3.67
14 19 1.53 0.00 0.49 0.00 0.13 0.11 1.85 4.11
15 29 3.60 0.00 0.40 0.00 0.11 0.11 0.58 4.80
16 20 1.13 0.00 0.27 0.00 0.11 0.11 2.27 3.89
17 27 1.65 0.00 0.27 0.00 0.00 0.11 2.54 4.56
18 22 4.22 0.00 0.38 0.00 0.18 0.00 1.65 6.43
19 25 3.46 0.00 0.11 0.00 0.00 0.00 2.33 5.90
20 24 3.51 0.00 0.16 0.00 0.11 0.11 1.51 5.40
21 26 3.53 0.00 0.13 0.00 0.00 0.00 1.42 5.09
22 30 2.20 0.00 0.22 0.00 0.13 0.00 2.18 4.74
23 8 3.01 0.00 0.22 0.00 0.11 0.11 2.07 5.53
24 2 3.00 0.00 0.22 0.02 0.20 0.11 1.89 5.45
25 12 1.58 0.00 0.22 0.00 0.20 0.11 2.04 4.15
26 6 3.69 0.00 0.13 0.00 0.11 0.00 2.02 5.96
27 10 3.40 0.00 0.00 0.00 0.00 0.00 2.82 6.22
28 14 3.18 0.00 0.11 0.00 0.11 0.00 1.84 5.25
29 4 3.22 0.00 0.22 0.00 0.24 0.00 1.69 5.38
30 16 1.87 0.00 0.09 0.00 0.11 0.00 2.02 4.09
2.83 0.00 0.26 0.01 0.15 0.13 1.63 5.01
0.94 0.00 0.13 0.03 0.08 0.26 0.70 0.94
2.88 0.00 0.25 0.00 0.06 0.06 1.78 5.04
0.81 0.00 0.12 0.00 0.07 0.06 0.72 0.48
MAX 4.22 0.00 0.62 0.16 0.37 1.11 2.82 6.43
MIN 1.13 0.00 0.00 0.00 0.00 0.00 0.24 3.55
Overall Avg.
Overall St. Dev.
Cntr. Pt. Avg.
Cntr. Pt. St. Dev.

239
Tab ary
le 90. GC/MS analytical data, whole bio-oil summ
Run
No.
DOE
No.
Acetic
acid
Levoglucosan
Furans
Phenols
Guaiacols
Syringols
Other
GS/MS
Total
1 17 2.43 2.00 0.60 0.46 1.73 2.87 3.75 13.84
2 18 3.43 2.33 0.47 0.49 1.77 2.97 3.47 14.94
3 15 3.61 2.12 0.55 0.49 1.77 2.97 3.46 14.98
4 7 4.67 2.26 0.51 0.55 1.82 2.92 3.66 16.39
5 9 4.65 1.71 0.67 0.48 1.72 2.77 3.02 15.00
6 13 2.46 1.30 0.62 0.51 1.70 2.82 3.12 12.52
7 11 5.11 1.48 0.70 0.54 1.91 3.22 2.72 15.69
8 5 4.06 1.76 0.52 0.54 1.77 2.80 3.60 15.05
9 1 4.69 2.66 0.46 0.56 1.83 2.98 3.27 16.45
10 3 1.34 2.37 0.45 0.56 1.86 3.07 3.04 12.69
11 21 2.35 2.33 0.99 0.49 1.72 2.80 3.85 14.55
12 28 4.38 2.25 0.93 0.49 1.67 2.84 4.90 17.46
13 23 1.41 0.50 1.07 0.43 1.58 2.42 4.20 11.62
14 19 4.33 2.29 1.41 0.51 1.80 2.97 5.45 18.75
15 29 3.21 2.40 1.11 0.53 1.70 3.01 5.15 17.09
16 20 1.36 1.78 0.65 0.48 1.65 2.84 3.92 12.67
17 27 3.60 1.95 0.64 0.52 1.79 3.07 3.31 14.87
18 22 2.35 1.68 0.64 0.50 1.84 3.14 2.85 13.01
19 25 1.93 1.85 0.80 0.51 1.74 3.08 2.51 12.41
20 24 1.72 1.93 0.41 0.54 1.74 2.87 2.88 12.08
21 26 1.72 2.01 0.57 0.50 1.83 3.24 3.60 13.47
22 30 2.87 1.96 0.76 0.50 1.77 3.03 4.85 15.73
23 8 3.27 2.10 0.76 0.53 1.85 3.26 3.73 15.49
24 2 3.33 2.03 0.82 0.52 1.74 2.92 3.47 14.83
25 12 2.42 1.74 0.72 0.46 1.67 2.89 3.40 13.31
26 6 3.96 1.98 0.65 0.50 1.69 2.91 3.25 14.96
27 10 3.02 1.56 0.61 0.47 1.77 2.92 3.19 13.53
28 14 2.03 1.59 0.65 0.47 1.79 2.95 3.91 13.39
29 4 1.90 2.08 0.47 0.51 1.78 3.10 3.07 12.91
30 16 2.64 1.77 0.66 0.48 1.75 2.99 3.66 13.96
3.01 1.92 0.70 0.50 1.76 2.96 3.61 14.46
1.12 0.41 0.22 0.03 0.07 0.17 0.71 1.72
2.95 2.07 0.80 0.51 1.75 3.04 4.05 15.18
1.01 0.21 0.19 0.01 0.06 0.13 1.07 1.99
MAX 5.11 2.66 1.41 0.56 1.91 3.26 5.45 18.75
MIN 1.34 0.50 0.41 0.43 1.58 2.42 2.51 11.62
Overall Avg.
Overall St. Dev.
Cntr. Pt. Avg.
Cntr. Pt. St. Dev.

240
0
10
20
30
40
50
60
70
80
90
100
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Run number
%-wt.
Unidentified
GC/MS quantified
Moisture
Water insoluble
Figure 150. Quantified mass for all runs
Table 91. Viscosity analytical data
Run
No.
DOE
No. Avg. St. dev. Avg. St. dev. Avg. St. dev.
12 28 75.0 1.9 4.7 0.08 147.4 13.2
15 29 99.3 4.9 5.1 0.06 239.8 11.7
17 27 115.4 3.4 5.3 0.10 111.6 9.7
19 25 157.2 6.0 5.5 0.15 135.6 8.4
21 26 124.0 3.8 5.9 0.11 98.7 8.8
22 30 122.8 4.7 5.3 0.14 142.6 8.6
115.6 4.09 5.3 0.11 146.0 10.07
- 4.29 - 0.11 - 10.23
20 24 234.5 2.8 9.7 0.4 255.0 15.1
13 23 75.0 1.9 4.7 0.1 147.4 13.2
Notes: All values in (cP) @ 40°C. a - Shear rates for center points and
Run 20 = 38.4 s-1
, shear rate for run 13 = 30.6 s-1.
b - Shear rates for
center points and Run 13 = 97.8 s-1
, shear rate for run 20 = 48.9 s-1
.
c - All shear rates = 38.4 s-1
.
d - Pooled standard deviation
Average
St. Dev.d
SF1a
SF2b
SF3c

241
-3
-2
-1
0
1
2
3
Residual
(Reaction
temperature,
C)
0 5 10 15 20 25 30
Run number
Figure 151. Residuals for reaction temperature full model
Table l data
92. Reaction temperature model statistica
Term Es
Standard Prob > |t|
mate
Standard
error t-ratio
Prob > |t|
(p-value)
466.163 0.7266 641.54 <0.0001 465.864 0.4099 1136.52 <0.0001
9.127 0.3633 25.12 <0.0001 9.127 0.3550 25.71 <0.0001
-0.489 0.3633 -1.35 0.1975 - - - -
1.558 0.3633 4.29 0.0006 1.558 0.3550 4.39 0.0002
3.506 0.3633 9.65 <0.0001 3.506 0.3550 9.88 <0.0001
-0.492 0.4450 -1.11 0.2863 - - - -
0.253 0.4450 0.57 0.5787 - - - -
0.269 0.4450 0.60 0.5550 - - - -
1.962 0.4450 4.41 0.0005 1.962 0.4348 4.51 0.0001
-0.780 0.4450 -1.75 0.1002 - - - -
0.132 0.4450 0.30 0.7705 - - - -
-0.097 0.3399 -0.29 0.7791 - - - -
-0.315 0.3399 -0.93 0.3693 - - - -
0.076 0.3399 0.22 0.8256 - - - -
-0.826 0.3399 -2.43 0.0281 -0.789 0.3241 -2.43 0.0228
R
2
timate error t-ratio (p-value) Esti
R2
R2
adjusted RMSE Mean
0.9810 0.9633 1.78 465.23 0.9710 0.9650 1.74 465.23
ANOVA DOF
Sum of
Squares
(SS-)
Mean
Square
(MS-) FANOVA
Prob > F
(p-value) DOF
Sum of
Squares
(SS-)
Mean
Square
(MS-) FANOVA
Prob > F
(p-value)
Regression (R) 14 2457.03 175.50 55.40 <0.0001 5 2431.96 486.39 160.84 <0.0001
Error (E) 15 47.52 3.17 24 72.59 3.024
Total (T) 29 2504.55 29 2504.550
Sum of
Squares
Mean
Square FLOF
Prob > F
(p-value) DOF
Sum of
Squares
Mean
Square FLOF
Prob > F
(p-value)
Lack of fit 10 41.693 4.169 3.58 0.086 9 20.57 2.29 0 66 0.7323
Pure error 5 5.827 1.165 15 52.01 3.47
Total 15 47.520 24 72.59
Intercept
HC temperature
N2 flow rate
Auger speed
HC feed rate
.

242
REFERENCES
1. Brown, R.C., Biorenewable Resources: Engineering New Products from Agriculture. 1st ed.;
Iowa State Press, Blackwell Publishing: Ames, IA, 2003.
2. Zamrik, S.Y.; Carter, V.R. ASME Annual Report, 2007-2008; American Society of
Mechanical Engineers: New York, NY, 2008; p 9.
3. Bridgwater, A.V., An introduction to fast pyrolysis of biomass for fuels and chemicals. In
Fast Pyrolysis of Biomass: A Handbook Volume 1, Bridgwater, A. V., Ed. CPL Press:
Newbury, UK, 1999; Vol. 1.
4. Mohan, D.; Pittman, C.U.; Steele, P.H. Pyrolysis of wood/biomass for bio-oil: A critical
review. Energy & Fuels 2006, 20, 848-889.
5. Perlack, R.D.; Wright, L.L.; Turhllow, A.F.; Graham, R.L. Biomass as feedstock for a
bioenergy and bioproducts industry: The technical feasability of a billion-ton annual supply;
ORNL/TM-2005/66; Oak Ridge National Laboratory: Oak Ridge, TN, 2005.
6. Brown, R.C., Resource availability. In Handbook of Conservation and Renewable Energy,
Kreith, F.; Goswami, D. Y., Eds. CRC Press: Boca Raton, FL, 2006.
7. Nag, A.; Manchikanti, P., Photosynthetic plants as renewable energy sources. In iofuels
Refining and Performance, Nag, A., Ed. McGraw-Hill: New York, NY, 2008.
8. Godfrey, B., Re ord University
Press: New York, NY, 2004.
. Brown, R.C. Hybrid thermochemical/biological processing of biomass. Applied Biochemistry
and Biotechnology
10. Huber, G.W.; Iborra, S.; C
Chemistry, catalysts, and engineeirng. Chemical Reviews 2006, 106, 4044-4098.
11. Behrendt, F.; Neubauer, Y.; Oevermann, M.; Wilmes, B., et al. Direct liquefaction of
biomass. Chemical Engineering Technology 2008, 31, (5), 667-677.
12. Bridgwater, A.V. Biomass fast pyrolysis. Thermal Science 2004, 8, (2), 21-49.
13. Bridgwater, A.V.; Czernik, S., The status of biomass fast pyrolysis. In Fast Pyrolysis of
Biomass: A Handbook Volume 2, Bridgwater, A. V., Ed. CPL Press: Newbury, UK, 2002;
Vol. 2.
14. Brown, R.C., Biochar production technology. In Biochar for Environmental Management:
Science and Technology, Lehmann, J.; Joseph, S., Eds. Earthscan: London, UK, 2009.
15. Wright, M.M.; Brown, R.C.; Boateng, A.A. Distributed processing of biomass to bio-oil for
the subsequent production of Fischer-Tropsch liquids Biofuels, Bioproducts, Biorefining
2008, 2, (3), 229-238.
16. Tijmensen, M.J.A.; Faaij, A.P.C.; Hamelinck, C.N.; VanHardeveld, M.R.M. Exploration of
the possibilities for production of Fischer-Tropsch liquids and power via biomass
gasification. Biomass & Bioenergy 2002, 23, 129-152.
17. Ciferno, J.P.; Marano, J.J. Benchmarking biomass gasification technologies for fuels,
chemicals and hydrogen production; National Energy Technology Laboratory: Pittsburgh,
PA, 2002.
18. Jenkins, B.M.; Baxter, L.L.; Jr., T.R. Miles; Miles, T.R. Combustion properties of biomass.
Fuel Processing Technology 1998, 54, 17-46.
19. Demirbas, A. Current technologies for the thermo-conversion of biomass into fuels and
chemicals. Energy Sources 2004, 26, 715-730.
20. Olofsson, I.; Nordin, A.; Soderlind, U. Initial review and evaluation of process technologies
and systems suitable for cost-efficient medium-scale gasification for biomass to liquid fuels;
B
newable Energy: Power for a Sustainable Future. 2nd ed.; Oxf
9
2007, 137-140, 947-956.
orma, A. Synthesis of transportation fuels from biomass:

243
Energy Technology & Thermal iversity of Umea: Umea, Sweden,
2005.
1. Bridgwater, A.V. The production of biofuels and renewable chemicals by fast pyrolysis of
22. les in fluid-bed reactors. Chemical
23. lysis kinetics of lignocellulosic materials
24. on
t wood species. J. Analytical and Applied Pyrolysis 2003, 68-
25. olomon, P.R. Measurement and modeling of
26.
27. ysis. AIChe
28. lysis of Biomass: A
29. , Biorefining
30. .V.C. Fast pyrolysis processes for biomass. Renewable and
31. rolysis for several types of biomass. Energy &
32.
d. CPL
roduction of
.
34. .J.; Mochidzuki, K.; Paredes, L.S. Flash carbonization of biomass. Industrial and
35. rading
36. rgy
598.
ing. ACS
merican Chemical Society:
38. .
olysis of Biomass: A Handbook Volume 2, Bridgwater, A. V., Ed. CPL Press:
39. ical property characterisation of biomass-
d,
Vol. 3.
Process Chemistry, Un
2
biomass. International J. Global Energy Issues 2007, 27, (2), 160-203.
Blasi, C.D. Modelling the fast pyrolysis of cellulosic partic
Engineering Science 2000, 55, 5999-6013.
Orfao, J.J.M.; Antunes, F.J.A.; Figueiredo, J.L. Pyro
- three independent reaction models. Fuel 1999, 78, 349-358.
Muller-Hagedorn, M.; Bockhorn, H.; Krebs, L.; Muller, U. A comparitive kinetic study
the pyrolysis of three differen
69, 231-249.
Serio, M.A.; Charpenay, S.; Bassilakis, R.; S
lignin pyrolysis. Biomass & Bioenergy 1994, 7, 107-124.
Rao, T.R.; Sharma, A. Pyrolysis rates of biomass materials. Energy 1998, 23, (11), 973-978.
Blasi, C.D. Modeling of intra- and extra- particle processes of wood fast pyrol
2002, 48, (10), 2386-2397.
Blasi, C.D., Kinetics and modeling of biomass pyrolysis. In Fast Pyro
Handbook Volume 3, Bridgwater, A. V., Ed. CPL Press: Newbury, UK, 2005; Vol. 3.
Babu, B.V. Biomass pyrolysis: A state-of-the-art review. Biofuels, Bioproducts
2008, 2, 393-414.
Bridgwater, A.V.; Peacocke, G
Sustainable Energy Reviews 2000, 4, 1-73.
Daugaard, D.E.; Brown, R.C. Enthalpy for py
Fuels 2003, 17, 934-939.
Gronli, M.; Antal, M.J.; Schenkel, Y.; Crehay, R., The science and technology of charcoal
production. In Fast Pyrolysis of Biomass: A Handbook Volume 3, Bridgwater, A. V., E
Press: Newbury, UK, 2005; Vol. 3.
33. Diebold, J.P.; Bridgwater, A.V., Overview of fast pyrolysis of biomass for the p
liquid fuels. In Fast Pyrolysis of Biomass: A Handbook Volume 1, Bridgwater, A. V., Ed
CPL Press: Newbury, UK, 1999; Vol. 1.
Antal, M
Engineering Chemistry Research 2003, 42, 3690-3699.
Qi, Z.; Jie, C.; Tiejun, W.; Ying, X. Review of biomass pyrolysis oil properties and upg
research. Energy Conversion and Management 2007, 48, 87-92.
Czernik, S.; Bridgwater, A.V. Overview of applications of biomass fast pyrolysis oil. Ene
& Fuels 2004, 18, 590-
37. Piskorz, J.; Scott, D.S.; Radlein, D., Compostion of bio-oils obtained by fast pyrolysis of
different woods. In Pyrolysis oils from biomass: Producing, analyzing and upgrad
Symposium series 376, Soltes, E. J.; Milne, T. A., Eds. A
Washington, D.C., 1988.
Oasmaa, A.; Meier, D., Analysis, characterisation and test methods of fast pyrolysis liquids
In Fast Pyr
Newbury, UK, 2002; Vol. 2.
Oasmaa, A.; Peacocke, G.V.C. A guide to phys
derived fast pyrolysis liquids; VTT Technical Research Centre of Finland: Espoo, Finlan
2001.
40. Oasmaa, A.; Meier, D., Characterisation, analysis, norms & standards. In Fast Pyrolysis of

244
41. y.
an: London, UK, 2009.
-oil produciton.
43. of biomass fast pyrolysis oil. In Fast Pyrolysis of
s
f
andbook Volume 3, Bridgwater, A. V., Ed. CPL Press: Newburk, UK, 2005;
45. t al. Production of gasoline and
46. nergies between bio- and oil refineries for the produciton of fuels
47. 07,
48. en, N.; Dinjus, E. Cost estimate for biosynfuel production via biosyncrude
49. onidou, A.A.; Vasalos, I.A. Hydrogen
63.
51.
53. ier, D., et al. Design, construction, and operation of
54. K.B. Bench-scale fluidized bed
Accessed July 20, 2009.
57. uly 20,
59. e: <http://www.btgworld.com/index.php?id=22&rid=8&r=rd>
60.
9, 8-11.
.
Lehmann, J.; Joseph, S., Biochar for Environmental Management: Science and Technolog
Earthsc
42. Mohan, D.; Pittman, C.U.; Bricka, M.; Smith, F., et al. Sorption of arsenic, cadmium, and
lead by chars produced from fast pyrolysis of wood and bark during bio
Colloid and Interface Science 2007, 310, 57-73.
Czernik, S.; Bridgwater, A.V., Applications
Vol. 3.
44. Gust, S.; Mclellan, R.J.; Meier, D.; Oasmaa, A., et al., Determination of norms and standard
for bio-oil as alternate renewable fuel for electricity and heat production. In Fast Pyrolysis o
Biomass: A H
Vol. 3.
Jones, S.B.; Valkenburg, C.; Walton, C.W.; Elliot, D.C., e
diesel from biomass via fast pyrolysis, hydrotreating and hydrocracking: A design case;
PNNL-18284; Pacific Northwest National Laboratory: Richland, WA, 2009.
Huber, G.W.; Corma, A. Sy
from biomass. Angewandte Chemie International Edition 2007, 46, 7184-7201.
Elliot, D.C. Historical developments in hydroprocessing of bio-oils. Energy & Fuels 20
21, 1792-1815.
Henrich, E.; Dahm
gasificiaton. Biofuels, Bioproducts, Biorefining 2009, 3, 28-41.
Kechagiopoulos, P.N.; Voutetakis, S.S.; Lem
produciton via steam reforming of the aqeous phase of bio-oil in a fixed bed reactor. Energy
& Fuels 2006, 20, 2155-21
50. Marris, E. Black is the new green. Nature 2006, 442, (10), 624-626.
Lehmann, J. Bio-energy in the black. Frontiers in Ecology and the Environment 2007, 5, (7),
381-387.
52. Ritzert, J.A.; Brown, R.C.; Smeenk, J. In Filtration efficiency of a moving bed granular filter,
Science in thermal and chemical biomass conversion, Victoria, Canada, 2004.
Gerdes, C.; Simon, C.M.; Ollesch, T.; Me
a fast pyrolysis plant for biomass. Eng. Life Sci. 2002, 6, 167-174.
Boateng, A.A.; Daugaard, D.E.; Goldberg, N.M.; Hicks,
pyrolysis of switchgrass for bio-oil production. Industrial and Engineering Chemistry
Research 2007, 46, 1891-1897.
55. Dynamotive, Corporate history. Web page: <http://www.dynamotive.com/about-
us/corporate-history/> 2009.
56. Hulet, C.; Briens, C.; Berruti, F.; Chan, E.W. A review of short residence time cracking
processes. International Journal of Chemical Reactor Engineering 2005, 3, Review R1.
Ensyn, Technology. Web page: <http://www.ensyn.com/tech.htm/> 2009. Accessed J
2009.
58. Ensyn, Products: Bio-chemicals. Web page: <http://www.ensyn.com/food.htm/> 2009.
Accessed July 20, 2009.
BTG, Technologies. Web pag
2009. Accessed July 20, 2009.
Laucks, I.F. The screw as a carbonizing machine. Industrial and Engineering Chemistry
1927, 1
61. Woody, G.V. Production of coke as a domestic fuel source: The Hayes process for
carbonizing coal at low temperatures. Industrial and Engineering Chemistry 1941, 33, (7),
841-846

245
62. Schmalfeld, J. In The Lurgi-Ruhrgas (LR) process applied to tar sand and other bituminous
materials, 17th Intersociety Energy Conversion Engineeirng (IECEC), Los Angeles, CA,
63.
64. ology assessment:Environmental, health
65. Khang, S.J.; Keener, T.C. Coal desulfurization by mild pyrolysis in a dual-auger coal
66.
67. H.A. In Torrefaction for biomass upgrading, 14th European
68. ilot-
Lawrence
69. yrolysis of
70. zhong, C.; Guibin, Z. In Technical advance on the pyrolysis of used tires in
ndai, Japan, 2000.
ional
72. ree, high pressure synthesis gas from biomass, Expert meeting
etwork (PyNE) newsletter 17 2004, pp 6-7.
and
75. te to technical
.
76. urope
arket conference, Growth and security, Washington, D.C.,
77. om biomass - The bioliq process,
ia fast pyrolysis and gasification, 15th European biomass conference & exhibition,
79. world's growing energy source, Southern Alliance for the
80. lue-added forestry, National bioenergy conference, Denver,
81. a company to produce biofuel. Web page:
ly 21,
1982.
Probstein, R.F.; Hicks, R.E., Synthetic Fuels. Dover Publications: Mineola, NY, 2006.
Daniels, J.I.; Anspaugh, L.R.; Ricker, Y.E. Techn
and safety impacts associatd with oil recovery from U.S. tar-sand deposits; UCRL-53210;
Lawrence Livermore National Laboratory: Livermore, CA, 1981.
Lin, L.;
feeder. Fuel Processing Technology 1997, 53, 15-29.
Camp, D.W. Mild coal gasification: Screw pyrolyzer development and design; UCRL-JC-
104514; Lawrence Livermore National Laboratory: Livermore, CA, 1990.
Bergman, P.C.A.; Kiel, J.
biomass conference & exhibition, Paris, France, 2005.
Camp, D.W.; Coburn, T.T.; Wallman, P.H. Mild coal gasification - Product separation, p
unit support, twin screw heat transfer, and H2S evolution; UCRL-JC-107903;
Livermore National Laboratory: Livermore, CA, 1991.
Lakshmanan, C.M.; Gal-or, B.; Hoelscher, H.E. Production of levoglucosan by p
carbohydrates. I&EC Product Research and Development 1969, 8, (3), 261-267.
Yongrong, Y.; Ji
China, China-Japan international academic symposium: Environmental problem in Chinese
iron-steelmaking industries and effective technology transfer, Se
71. Henrich, E.; Weirich, F. In Pressurized entrained flow gasifiers for biomass, Internat
Thermal Treatment Technologies (IT3), New Orleans, LA, 2002.
Henrich, E.; Dinjus, E. In Tar-f
on pyrolysis and gasification of biomass, Strassbourg, France, 2002.
73. Henrich, E., Fast pyrolysis of biomass with a twin screw reactor - A first BTL step. Biomass
Pyrolysis N
74. Raffelt, K.; Henrich, E.; Koegel, A.; Stahl, R., et al. The BTL2 process of biomass utilization
entrained flow gasification of pyrolyzed biomass slurries. Applied Biochemistry
Biotechnology 2006, 129, (1), 153-164.
Henrich, E. In Syngas from biomass: Problems and solutions en rou
realization, International conference on synthesis gas chemistry, Dresden, Germany, 2006
Plass, I.L. In Energy and transportation status and future developments of biofuels in E
and the U.S., 3rd Transatlantic m
2007.
Dinjus, E.; Dahmen, N.; Koerber, R. In Synthetic fuel fr
International trade fair for water, sewage, refuse, recycling and natural energy sources
(IFAT), Shanghai, China, 2008.
78. Leible, L.; Kalber, S.; Kapplier, G.; Lange, S., et al. In Fischer-Tropsch synfuels from cereal
straw v
Berlin, Germany, 2007.
Badger, P. In BioOil - The
Utilization of Biomass Resources (SAUBR), Tuscaloosa, AL, 2004.
Badger, P. In A new vision for va
CO, 2006.
Sims, B., Alabam
<http://www.biomassmagazine.com/article.jsp?article_id=1781> 2008. Accessed Ju
2009.

246
82. Fransham, P.B., Advances in dry distillation technology. Biomass Pyrolysis Network (PyNE)
newsletter 18 2006, pp 11-12.
Badger,
83. P.C.; Fransham, P.B. Use of mobile fast pyrolysis plants to densify biomass and
30,
84.
iooils and char by 13C
uger
o-oils. Energy & Fuels
87. sis
e products. Fuel 2009, 88, 1251-
88. d fuel
89.
ITC-
90. to fuel gas using a moving
91. fficiency electricity
chnology exhibition on biomass for
92. y &
f the
94.
.; McGraw-Hill: New York, NY, 1996.
96. . Large-scale pyrolysis oil production: A technology
97. onveyors for Bulk Solids. Elsevier Science Publishers B.V.: New
98. tional heat transfer
99. tiemo-Obeng, V.A.; Kresta, S.M., Handbook of Industrial Mixing: Science and
100.
y Inc.: New York, NY, 1983.
reduce biomass handling costs - A preliminary assessment. Biomass & Bioenergy 2006,
321-325.
Schnitzer, M.I.; Monreal, C.M.; Facey, G.A.; Fransham, P.B. The conversion of chicken
manure to biooil by fast pyrolysis I. Analyses of chicken manure, b
and 1H NMR and FTIR spectrophotometry. Environmental Science and Health, Part B 2006,
42, 71-77.
85. Steele, P.H. Bioenergy research in the Department of Forest Products; The Forest and
Wildlife Research Center, Mississippi State University: Year unknown.
86. Ingram, L.; Mohan, D.; Bricka, M.; Steele, P., et al. Pyrolysis of wood and bark in an a
reactor: Physical properties and chemical analysis of the produced bi
2008, 22, (1), 614-625.
Bhattacharya, P.; Steele, P.H.; Hassan, E.B.M.; Mitchell, B., et al. Wood/plastic copyroly
in an auger reactor: Chemical and physical analysis of th
1260.
Garcia-Perez, M.; Adams, T.T.; Goodrum, J.W.; Gellar, D.P., et al. Production an
properties of pine chip bio-oil/biodiesel blends. Energy & Fuels 2007, 21, 2363-2372.
Hornung, A.; Koch, W.; Seifert, H. HALOCLEAN and PYDRA - A dual staged pyrolysis
plant for the recycling of Waste Electronic and Electrical Equipment (WEEE); FZK,
TAB: Karlsruhe, Germany, Year unknown.
Kodera, Y.; Ishihara, Y. Novel process for recycling waste plastics
bed reactor. Energy & Fuels 2006, 20, 155-158.
Oudhuis, A.B.J.; Bos, A.; Ouweltjes, J.P.; Rietveld, G., et al. In High e
and products from biomass and waste; experimental results of proof of principle staged
gasification and fuel cells, 2nd World conference and te
energy, industry, and climate protection, Rome, Italy, 2004.
Brandt, P.; Larsen, E.; Henriksen, U. High tar reduction in a two-stage gasifier. Energ
Fuels 2000, 14, 816-819.
93. Strezov, V.; Moghtaderi, B.; Lucas, J.A. Computational calorimetric investigation o
reactions during thermal conversion of wood biomass. Biomass & Bioenergy 2004, 27, 459-
465.
Avallone, E.A.; III, T. Baumeister, Mark's Standard Handbook for Mechaical Engineers.
10th ed
95. Moran, M.J.; Shapiro, H.N., Fundamentals of Engineering Thermodynamics. 5th ed.; John
Wiley & Sons, Inc.: Hoboken, NJ, 2004.
Ringer, M.; Putsche, V.; Scahill, J
assessment and economic analysis; NREL/TP-560-39181; National Renewable Energy
Laboratory: Golden, CO, 2006.
Colijn, H., Mechanical C
York, NY, 1985; Vol. 4.
Kern, D.Q.; Karakas, H.J. In Mechanically aided heat transfer, 2nd na
conference (AIChE and ASME), Chicago, IL, 1958.
Paul, E.L.; A
Practice. John Wiley & Sons, Inc.: Hoboken, NJ, 2004.
Martelli, F.G., Twin-Screw Extruders: A Basic Understanding. Van Nostrand Reinhold
Compan

247
101. Shamlou, P.A., Handling of bulk solids: Theory and Practice. Butterworth & Co.: London,
UK, 1988.
103. , A.; Ganan, J.; Tinaut, F.V. Theoretical and experimental study of a direct contact
104.
mond, IN, 1992.
ctober 25, 2007.
107.
step change method. J. of Food Engineering 2003, 59, (161-167).
logy 2005, 157, 138-143.
111. , C. The performance of continuous
112. dbook for Statistical Methods. Ch. 5,
113.
r
erica Program: 2003.
116. , S.A.E. Specification for steel shot and grit. Web page:
117. fast pyrolysis liquids. J. Analytical and
118. Applied Statistics for Engineers and Scientists.
119.
. Chem. Chem. Phys. 1999, 1, 5035-5043.
plied Pyrolysis 1998, 47, 43-64.
andbook Volume 2, Bridgwater, A.
102. Weinekotter, R.; Gericke, H., Particle Technology Series: Mixing of Solids. Kluwer
Academic Publishers: Dordrecht, Netherlands, 2000.
Al-Kassier
thermal screw dryer for biomass residues. Applied Thermal Engineeirng 2005, 25, 2816-
2826.
Screw-Conveyor-Corporation Quality Screw Conveyor Systems and Components, Catalog
991A; Ham
105. Auger-Manufacturing-Specialists, Standard auger sizes and capacities. Web page:
<http://www.augermanufacturing.com/standard.html> 1998. Accessed O
106. Reist, P.C., Aerosol Science and Technology. 2nd ed.; McGraw-Hill: New York, NY, 1993.
Ziegler, G.R.; Aguilar, C.A. Residence time distribution in a co-rotating, twin screw
continuous mixer by the
108. Marikha, K.; Berthiaux, H.; Mizonov, V.; Barantseva, E. Experimental study of the stirring
conditions taking place in a pilot plant continuous mixer of particulate solids. Powder
Techno
109. Pernenkil, L.; Cooney, C.L. A review on the continuous blending of powders. Chemical
Engineering Science 2006, 61, 720-742.
110. Jones, J.R.; Parker, D.J.; Bridgwater, J. Axial mixing in a ploughshare mixer. Powder
Technology 2007, 178, 73-86.
Harwood, C.F.; Walanski, K.; Luebcke, E.; Swanstrom
mixers for dry powders. Powder Technology 1975, 11, 289-296.
Tobias, P.; Trutna, L., NIST/SEMATECH e-Han
Process Imporvement. Web page: <http://www.itl.nist.gov/div898/handbook> 2005.
Accessed August 3, 2009.
Kuehl, R.O., Design of Experiments: Statistical Principles of Research Design and Analysis.
2nd ed.; Brooks/Cole: Pacific Grove, CA, 2000.
114. Nesom, G. Northern Red Oak (Quercus rubra L.); USDA NRCS National Plant Data Cente
and the Biota of North Am
115. Ervin-Industries, Steel shot and grit applications. Web page:
<http://www.ervinindustries.com/application.htm> 2009. Accessed June 17, 2009.
Marco
<http://www.marcousa.com/select/datasheets/93740.pdf> 2005. Accessed June 17, 2009.
Oasmaa, A.; Meier, D. Norms and standards for
Applied Pyrolysis 2005, 73, (2), 323-334.
Levine, D.M.; Ramsey, P.P.; Smidt, R.K.,
Prentice Hall: Upper Saddle River, NJ, 2001.
Ball, R.; McIntosh, A.C.; Brindley, J. The role of char-forming processes in the thermal
decomposition of cellulose. Phys
120. Blasi, C.D. Comparison of semi-global mechanisms for primary pyrolysis of lignocellulosic
fuels. J. Analytical and Ap
121. McKendry, P. Energy production from biomass (part I): overview of biomass. Bioresource
Technology 2002, 83, 37-46.
122. Diebold, J.P., A review of the chemical and physical mechanisms of the storage stability of
fast pyrolysis bio-oils. In Fast Pyrolysis of Biomass: A H
V., Ed. CPL Press: Newbury, UK, 2002; Vol. 2.

248
ACKNOWLEDGEMENTS
the individuals who contributed to the success of this research effort.
on
career has
chanical
,
provide
for
this res ty biofuels research program (Project 2007-P-02,
nat
),
in port in
various nce and
t and camaraderie, especially: Cody Ellens,
hnicians at CSET that provided
Heithoff,
John Ho
and
ce.
d me
to learn ergy and thermal
s
.
Brown
ncoura up to an undergraduate engineering degree,
d continue to show interest and support in my current endeavors.
The author would like to take this opportunity to formally acknowledge and thank some of
My major professor, Dr. Robert C. Brown, provided helpful guidance, support and inspirati
for the duration of this research. Dr. Brown’s consideration during my graduate school
provided me with phenomenal opportunities for which I will always be grateful.
My committee members, Dr. Theodore Heindel from the Department of Me
Engineering and Dr. D. Raj Raman from the Department of Agricultural and Biosystems Engineering
d helpful feedback on the technical content of this work.
ConocoPhillips Company provided generous financial support and technical assistance
earch, through the Iowa State Universi
Alter ive Fast Pyrolysis Reactor Design).
The technical staff at the Center for Sustainable Environmental Technologies (CSET
includ g Dr. Justinus Satrio, Marjorie Rover and Patrick Johnson provided advice and sup
capacities. In particular, Dr. Samuel Jones provided exceptional engineering assista
invaluable guidance for this project, including a helpful review of this thesis.
My graduate student colleagues provided suppor
Anthony Pollard, Patrick Meehan, David Chipman, Mark Wright, Pedro Ortiz and Raj Pathwardan.
Much gratitude is given to the undergraduate Laboratory Tec
obbie Hable, Trevor
technical and analytical assistance, especially: Stephen Laskowski, R
yt, Ben Franzen, Ben Peterson, Brad Williams, and Guy Lasley.
The staff at Country Plastics, Ames Laboratory, and the Chemistry Department
Mechanical Engineering Department machine shops provided technical and manufacturing assistan
Special recognition is given to Dr. Daren Daugaard who initially challenged and inspire
engineering thermodynamics, and later introduced me to renewable en
conversion of biomass. His positive influence on my career continued with his involvement on thi
project through ConocoPhillips Company.
Many thanks and much appreciation are due to my parents: J.M. Doerr, M.S.N. and F.W
III, M.D. They first helped me to be inquisitive and appreciate science at a young age, then
ged and supported me throughout my education
e
an

249
BIOGRAPHICAL SKETCH
ared Nathaniel Brown was born on January 6, 1983 in San Antonio, Texas. As a young
adult, J
engineer.
m UTSA, majoring
in Mech
analytical experience with fluidized
bed reac
yndmoor, Pennsylvania.
e
the auge
ogy.
J
ared found great interest in mechanics, internal combustion engines and computer-aided
drawing programs, which led to aspirations of becoming an automotive
In 2005, after learning about biomass resources in an Alternative Energy elective course at
the University of Texas at San Antonio (UTSA), Jared’s interests and career goals began to shift
towards bioenergy and biofuels. In December 2006 he earned his B.S. degree fro
anical Engineering with a specialization in Thermal and Fluid Systems.
Supervised by Dr. Daren Daugaard at UTSA and Dr. Akwasi Boateng at the Agricultural
Research Service branch of the USDA, Jared gained practical and
tors and biomass fast pyrolysis beginning around 2005. At the USDA during the summer of
2006, he served as an Engineering Technician for the Crop Conversion Science and Engineering unit
at the Eastern Regional Research Center in W
In the spring of 2007 Jared moved to Ames, Iowa to join the Center for Sustainable
Environmental Technologies (CSET) at Iowa State University, directed by Dr. Robert C. Brown. As a
Lab Technician, he performed numerous fast pyrolysis experiments with a fluidized bed reactor at the
Biomass Energy Conversion facility in Nevada, Iowa.
By May 2007 Jared had began work as a Graduate Research Assistant for the Department of
Mechanical Engineering, on a project sponsored by ConocoPhillips Company to develop and evaluat
r reactor as an alternative fast pyrolysis reactor design.
In early 2009, with several other colleagues from CSET, Jared co-founded Avello Bioenergy,
Inc. for the purpose of commercializing proprietary fast pyrolysis technol

Development of a lab-scale auger reactor for biomass fast pyrolysis ( PDFDrive ).pdf

More Related Content

Similar to Development of a lab-scale auger reactor for biomass fast pyrolysis ( PDFDrive ).pdf (20)

More from DktrSaimYiitBayrak (11)

Recently uploaded (20)

Development of a lab-scale auger reactor for biomass fast pyrolysis ( PDFDrive ).pdf