A Dissertation Entitled Pyrolysis Strategies for Effective Utilization of Lignocellulosic and Algal Biomass by Balakrishna Madd

A Dissertation

entitled

Pyrolysis Strategies for Effective Utilization of Lignocellulosic and Algal Biomass

Balakrishna Maddi

Submitted to the Graduate Faculty as partial fulfillment of the requirements for the

Doctor of Philosophy Degree in Engineering

______Dr. Sasidhar Varanasi, Committee Chair

______Dr. Sridhar Viamajala, Committee Member

______Dr. Glenn Lipscomb, Committee Member

______Dr. Arunan Nadarajah, Committee Member

______Dr. Thomas Bridgeman, Committee Member

______Dr. Patricia R. Komuniecki, Dean College of Graduate Studies

The University of Toledo December 2014

This document is copyrighted material. Under copyright law, no parts of this document may be reproduced without the expressed permission of the author. An Abstract of

Pyrolysis Strategies for Effective Utilization of Lignocellulosic and Algal Biomass

Balakrishna Maddi

Submitted to the Graduate Faculty as partial fulfillment of the requirements for the Doctor of Philosophy Degree in Engineering

The University of Toledo

December 2014

Pyrolysis is a processing technique involving thermal degradation of biomass in the absence of oxygen. The bio-oils obtained following the condensation of the pyrolysis vapors form a convenient starting point for valorizing the major components of lignocellulosic as well as algal biomass feedstocks for the production of fuels and value- added chemicals. Pyrolysis can be implemented on whole biomass or on residues left behind following standard fractionation methods. Microalgae and oil seeds predominantly consist of protein, carbohydrate and triglycerides, whereas lignocellulose is composed of carbohydrates (cellulose and hemicellulose) and lignin. The differences in the major components of these two types of biomass will necessitate different pyrolysis strategies to derive the optimal benefits from the resulting bio-oils. In this thesis, novel pyrolysis strategies were developed that enable efficient utilization of the bio-oils (and/or their vapors) from lignocellulose, algae, as well as oil seed feed-stocks.

With lignocellulosic feedstocks, pyrolysis of whole biomass (chapter 2)as well as the lignin residue left behind following well-established pretreatment and saccharification

(i.e., depolymerization of cellulose and hemicellulose to their monomeric-sugars) of the

iii biomass was studied with and without catalysts(Appendix 1). Following this, pyrolysis of (lipid-deficient) algae and lignocellulosic feedstocks, under similar reactor conditions, was performed for comparison of product (bio-oil, gas and bio-char) yields and composition (Chapter 2). In spite of major differences in component bio-polymers, feedstock properties relevant to thermo-chemical conversions, such as overall C, H and

O-content, C/O and H/C molar ratio as well as calorific values, were found to be similar for algae and lignocellulosic material. Bio-oil yields from algae and some lignocellulosic materials were similar; however, algal bio-oils were compositionally different and contained several N-compounds (most likely from protein degradation). Algal bio-char also had a significantly higher N-content. Overall, our results suggest that it is feasible to convert algal cultures deficient in lipids, such as nuisance algae obtained from natural blooms, into liquid fuels by thermochemical methods.

Next, pyrolysis characteristics of each of the major components present in lignocellulosic as well as algal biomass were studied independently in a thermogravimetric analyzer, using model compounds (Chapter 3). From those studies, we have established that, with algae and oil seed feedstocks, triglycerides degrade at distinctly higher temperatures (T>380 °C) compared to both protein and carbohydrate fractions (T

~ 250-350 °C). Similar trend was not seen for lignocellulosic biomass, where degradation temperature interval of lignin overlapped with that of carbohydrates. This unique trend observed for algal biomass (and oil seeds) can be exploited in multiple ways. First, it permits to separately collect high value triglyceride degradation products not contaminated with N-compounds from protein and oxygenates from carbohydrates; this observation formed the basis of a novel “pyrolytic fractionation technique” developed in

iv this thesis (Chapter 3). Second, it led to the development of a new and simple analytical method for rapid estimation of the lipid content of oleaginous feed-stocks (Chapter 4)

Pyrolytic fractionation is a two-step pyrolysis approach that can be implemented for oleaginous feed stocks (algae and oil-seeds) to separately recover triglyceride degradation products as a “high-quality” bio-oil fraction. The first step is a low- temperature pyrolysis (T ~ 300-320 °C) to produce bio-oils from degradation of protein and carbohydrate fractions. Solid residues left behind can subsequently be subjected to a second higher temperature pyrolysis (T ~ 420-430 °C) to volatilize and/or degrade triglycerides to produce fatty acids and their derivatives (such as mono-, di- and triglycerides) and long chain hydrocarbons. Proof-of-concept micro-pyrolyser

(PyroprobeTM) and lab-scale fixed-bed experiments were performed using oleaginous algae (Chlorella Sp.) to establish pyrolytic fractionation technique and also to determine the yields of triglyceride-specific bio-oils (Chapter 3). As expected, triglyceride-specific bio-oils have hydrocarbons and free fatty acids that were nearly free of water, organic acids and carbohydrate degradation products. Another unique feature of the fractional pyrolysis method is that it allows upgrading of the triglyceride-specific bio-oil vapors via in situ gas-phase hydrodeoygenation to drop-in fuels (hydrocarbons), without the need to condense the vapors. Similarly, these vapors can also be converted to other value-added products such as fatty acid methyl esters and amides though efficient catalytic and non- catalytic in-situ gas-phase conversion methods. Energy requirements for this new pyrolytic fractionation method were also assessed, using energy estimates for the individual steps obtained via differential scanning calorimetry experiments. A comparison of these energy needs against those of alternative thermal processing

v methods of algae (hydro-thermal processing) proposed in the literature established the viability of this new method (Chapter 3).

Finally, the last chapter (Chapter 4) describes a new TGA-based analytical method developed in this thesis for rapid quantitation of the triglyceride content of oleaginous feedstocks, by exploiting the non-overlapping thermal degradation range of triglycerides and the other major components.

Acknowledgements

I would like to acknowledge the patient guidance of my supervisors, Dr. Sasidhar

Varanasi and Dr. Sridhar Viamajala, throughout the course of this doctorate degree. Their careful direction and consistent challenge is the key in the completion of this work. I would also thank them for their suggestions in professional development and skills that have made my learning experience much more wholesome.

I would like to thank Dr. Glenn Lipscomb for offering me with the opportunity of pursuing my graduate studies at the Chemical & Environmental Engineering Department at The University of Toledo. I would also like to sincerely thank my dissertation committee members (Dr. Glenn Lipscomb, Dr. Arunan Nadarajah and Dr. Thomas

Bridgeman) for their constructive comments during my proposal and final defense.

This work would not have been possible without the support and encouragement of my colleagues and labmates Ananth Dadi, Thehazhnan (Thihal) Ponnaiyan, Satish

Lakhapatri, Rahul Patil and Indira Priya Samayam under whose supervision I began working toward my dissertation. I would like to thank my other labmates Brook Urban,

Agasteswar Vadlamani, Sricharan Nanduri, Patrick Bollin and Yasir Sherazi.

I would like to thank my parents and my brother for their loving support and words of encouragement. I am forever indebted to my family for their understanding, endless patience and words of encouragement when it was most required. Above all, I thank God through whom all things are possible.

vii

Table of Contents

Abstract ...... iii

Acknowledgements ...... vii

Table of Contents ...... viii

List of Tables ...... xiii

List of Figures ...... xvi

1. Introduction ...... 1

1.1 Production of liquid fuels and high-value chemicals from lignocellulosic

biomass: ...... 1

1.2 Advantages of algal biomass: ...... 3

1.3 Research Outline: ...... 8

1.3.1 Comparative study of pyrolysis products of algal and lignocellulosic

feedstocks: ...... 8

1.3.2 Pyrolytic fractionation of non-ligneous biomass especially oleaginous

and algal feedstocks: ...... 9

1.3.3 Triglyceride quantification of oleaginous algal biomass using thermo-

gravimetry: ...... 10

viii

Comparative study of pyrolysis of algal biomass from natural lake blooms with

lignocellulosic biomass ...... 11

1.4 Introduction ...... 11

1.5 Methods and materials ...... 13

1.5.1 Raw materials and sample preparation ...... 13

1.5.2 Thermo-gravimetric analysis (TGA) of feed stocks ...... 14

1.5.3 Experimental set-up ...... 14

1.6 Analytical methods ...... 16

1.6.1 Compositional analysis: ...... 16

1.6.2 Proximate analysis ...... 17

1.6.3 Ultimate analysis ...... 17

1.6.4 Analysis of bio-oil...... 18

1.7 Results & Discussion ...... 18

1.7.1 Feedstock characterization ...... 18

1.7.2 Thermo-gravimetric analysis of feedstocks ...... 21

1.7.3 Pyrolysis experiments ...... 25

1.7.4 Energy implications of processing nuisance algae ...... 31

1.8 Conclusions ...... 32

1.9 Acknowledgements ...... 32

Pyrolytic fractionation: A thermo-chemical technique for processing oleaginous (algal)

biomass ...... 34

1.10 Introduction: ...... 34

1.11 Experimental ...... 36

1.11.1 Feedstocks and chemicals ...... 36

1.11.2 Triglyceride quantification: ...... 37

1.11.3 Thermo-gravimetric (TG) analysis and differential scanning

calorimetry (DSC) ...... 37

1.11.4 Pyrolysis experiments ...... 38

1.11.5 Experimental set-up: ...... 40

1.12 Results and Discussion ...... 42

1.12.1 TG studies to identify biopolymer degradation temperatures: ...... 42

1.12.2 Simulation of pyrolytic fractionation by thermo-gravimetry: ...... 45

1.12.3 Bench-scale micro-pyrolysis experiments: ...... 48

1.12.4 Lab-scale pyrolysis experiments: ...... 59

1.12.5 Conceptual process design of pyrolytic fractionation: ...... 67

1.12.6 Estimates of energy required for pyrolytic fractionation: ...... 69

1.12.7 ChemcadTM simulation of hydrothermal liquefaction process: ...... 77

1.13 Conclusions ...... 80

Triglyceride quantification of oleaginous algal biomass using thermo-gravimetry (TG) .. 82

1.14 Introduction ...... 82

1.15 Methods and methods: ...... 84

1.15.1 Materials: ...... 84

1.15.2 Thermo-gravimetric Analysis (TGA): ...... 85

1.15.3 Solvent extraction: ...... 86

1.15.4 In situ transesterification method: ...... 86

1.15.5 Quantification of triglycerides and FAMEs using GC-FID: ...... 87

1.15.6 Statistical analysis: ...... 87

1.16 Results and discussion: ...... 88

1.16.1 Pyrolysis of triglycerides: ...... 88

1.16.2 Pyrolysis of triglycerides at 420 °C: ...... 90

1.16.3 Pyrolysis of triglycerides-rich feedstocks: ...... 91

1.16.4 Loss of triglycerides at low temperatures: ...... 92

1.17 Conclusions: ...... 101

References ...... 102

A. Characterization of pyrolysis products of residual lignin obtained via IL

pretreatment followed by enzymatic hydrolysis ...... 123

B. Pyrolytic fractionation of triglyceride-lean Chlorella sp...... 148

C. Co-pyrolysis of triglyceride-lean Chlorella sp. and tristearate performed at 320 °C

and then at 420 °C ...... 152

D. Supplementary data ...... 156

xii

List of Tables

1.1: Chemical compounds formed during pyrolysis of carbohydrates, lignin, protein and triglycerides...... 6

2.1: Proximate analysis and compositional characteristics of feedstocks used in the present study. Carbohydrates include cellulose and hemicellulose in the lignocellulosic materials and starch in algal biomass. Lignin is not present in the algal feedstocks. The values reported are mean of three samples and expressed as mass fractions (%) of the original samples. Standard deviations were < 8% for all reported values...... 20

2.2: Elemental analyses of all feedstocks "As-recieved" values indicate measurements directly obtained from the CHN analyzer. "Dry basis" and "dry-ash free basis" values were calculated from "as-recieved" values after discounting moisture and ash content, respectively. Calorific values (HHV) were calculated using equation 2, 3 and 4. Algae samples were dried (after dewatering) at 45 C for 24 h prior to analysis. The values reported are average of two samples and expressed as mass fractions (%). Mass fraction of oxygen was calculated by difference. Standard deviations were <2 % for all reported values...... 20

2.3: Elemental analyses of bio-char obtained by pyrolysis are expressed in dry and dry- ash free basis. “Dry-basis” values were obtained by CHN analyzer. “dry-ash free basis” values were calculated by using “dry-basis” values, and ash content. Calorific values

(HHV) were calculated using equations 2, 3 and 4. All values are reported as mass

xiii fractions (%). Mass fraction of oxygen was calculated by difference. The values reported are average of two identical samples. Standard deviations were < 4% for all reported values...... 27

2.4: Instantaneous composition of gases (mole %) measured during the course of pyrolysis experiments. Measurements were made at 510 °C (15 min) and 600 °C (35 min)...... 29

2.5: Chemical compounds identified in the bio-oils by GC-MS. The identified compounds are classified based on their most-likely source biopolymer - polysaccharide

(cellulose, hemicellulose or starch) lignin and protein. “X” marks indicate the presence of the compound in the bio-oil from the corresponding feedstock. Blanks indicate that the compound was not identified in the bio-oil from the corresponding feedstock...... 31

3.1: Confidence value of chemical compounds present in protein as well as carbohydrate- based bio-oils of Chlorella sp. (Refer to figure 3-6)...... 53

3.2: Confidence value of chemical compounds present in protein as well as carbohydrate- based bio-oils of Scenedesmus sp. (Refer to figure 3-6)...... 54

3.3: Confidence value of chemical compounds present in triglyceride-based bio-oils of

Chlorella sp. (Refer to figure 3-7)...... 58

3.4: Confidence value of chemical compounds present in triglyceride-based bio-oils of

Scenedesmus sp. (Refer to figure 3-9)...... 59

3.5: Confidence values of chemical compounds present in bio-oil collected during Step 1

(320 °C) of bench-scale fixed-bed experiments with oleaginous Chlorella sp. at 320 °C.

...... 62

xiv

3.6: Confidence values of chemical compounds present in bio-oil collected during Step 2 of bench-scale fixed-bed experiments with oleaginous Chlorella sp. at 420 °C...... 64

3.7: Elemental analysis of bio-oil collected during Step 2 of bench-scale fixed-bed pyrolytic fractionation experiments with Chlorella sp. at 420 °C...... 66

3.8: Comparison of energy requirements for pyrolytic fractionation and hydrothermal liquefaction...... 73

4.1: Moisture content of oleaginous algal feedstocks estimated using ASTM D7582-12 protocol...... 85

4.2: Comparison of triglyceride content of oleaginous algae estimated via TG method and in situ transesterification method. Triglyceride content determined using TG method is statistically similar to estimates obtained from in situ transesterification method...... 95

4.3: Triglyceride content of oleaginous algal feedstocks estimated via TG method and solvent extraction method...... 96

4.4: Summary of triglyceride mass balance around solvent extraction method for algal samples. Sum of triglyceride content in solvent extract and solid residue is statistically similar to values obtained for whole biomass. Triglyceride content in whole biomass was estimated using in situ transesterification method. Triglyceride content in solid residue was estimated by TG methods as the results obtained from TG methods and in situ transesterification procedure is statistically similar...... 98

4.5: Comparison of triglyceride content of wet and dry sample of Chlorella sp. estimated via TG method. Triglyceride content (on dry-basis) of wet sample is statistically similar to estimates obtained for dry sample estimated using TG method and In situ transesterification...... 100

List of Figures

1-1: Overview of lignocellulosic-based bio-refinery...... 3

1-2: Reactions occur during biomass pyrolysis ...... 5

2-1: Schematic diagram of lab-scale pyrolysis set-up...... 15

2-2: Weight loss (dashed lines) and its temperature derivative (solid line) of (a) Corncobs

(b) Wood chips (c) Rice husk (d) Lyngbya sp. (e) Cladophora sp. (f) Bovine serum albumin (g) Lysozyme (h) Corn starch...... 23

2-3: Yields of bio-oil, ash-free bio-char, ash and gases obtained after pyrolysis at 600 °C for all the feedstocks tested. Ash was assumed to stay associated with the solid residue and subtracted from residue weight to obtain ash-free bio-char values. All reported values are an average of two experiments. Error bars indicate one standard deviation from mean values...... 25

2-4: Yields of bio-oil, ash-free bio-char, ash and gases obtained after pyrolysis of

Lyngbya sp. at different temperatures. Ash was assumed to stay associated with the solid residue and subtracted from residue weight to obtain ash-free bio-char values...... 28

3-1: Residual weight (dashed) and derivative weight loss curves (solid lines) obtained during thermal degradation of (a) Chlorella sp. and (b) Scenedesmus sp...... 43

3-2: Derivative weight loss (solid) and residual weight (dashed) curves obtained during thermal degradation of soy oil...... 44

xvi

3-3: TG profiles resulting from pyrolytic fractionation of the protein as well as carbohydrate portion of (a) Chlorella sp. and (b) Scenedesmus sp.. Yellow residual weight (dashed) and derivative weight loss curves (solid lines) show pyrolysis of the protein and carbohydrate fractions of the biomass (T<.320 °C). Green curves show thermograms of the samples obtained after removal of pyrolyzable protein and carbohydrate which indicate that prolonged exposure to lower temperatures does not negatively impact the thermal decomposition characteristics of the constituent triglycerides. The arrows indicate the temperature path for followed for these experiments...... 45

3-4: Kinetics of thermal degradation of the carbohydrate fractions of (a) Chlorella sp. and

(b) Scenesdesmus sp.. The non-isothermal zone shows weight loss during temperature ramp from 250 to 320 °C. The isothermal zone shows weight loss during incubation at

320 °C...... 47

3-5: Kinetics of thermal degradation of the lipid fractions of (a) tripalmitate, (b) tristearate and (c) triolein. The non-isothermal zone shows weight loss during temperature ramp from room temperature to 420 °C. The isothermal zone shows weight loss during incubation at 420 °C...... 48

3-6: GC-MS chromatogram of protein as well as carbohydrate derived bio-oils from (a)

Chlorella sp. and (b) Scenedesmus sp... The chemical compounds were identified using the NIST2008 library. Refer to table 3.1 and table 3.2 for full list of chemical compounds identified in this chromatogram...... 52

3-7: Derivative weight loss (solid) and residual weight (dashed) curves obtained during thermal degradation of (a) Myristic acid (C14) and (b) Stearic acid (C18)...... 55

xvii

3-8: GC-MS chromatogram of products from pyrolysis of 1,3-diolein at 320 °C...... 55

3-9: GC-MS chromatogram of products from pyrolysis of soy oil at 320 °C...... 56

3-10: GC-MS chromatogram of products from pyrolysis of soy oil at 420 °C...... 56

3-11: GC-MS chromatogram of triglyceride derived bio-oils from (a) Chlorella sp. and

(b) Scenedesmus sp.. The chemical compounds were identified using the NIST2008 library. Refer to table 3.3 and table 3.4 for full list of chemical compounds identified in this chromatogram ...... 57

3-12: Total- and component- mass balance data obtained from bench-scale fixed-bed experiments for Step 1 pyrolysis of oleaginous Chlorella sp. at 320 °C. †indicates lipid is quantified as FAMEs; ‡ indicates lipid is sum of hydrocarbons, free fatty acids, fatty amides and fatty nitrile; * indicates lipid is triglycerides...... 60

3-13: GC-MS chromatogram of bio-oil collected from Step 1 of fixed-bed pyrolytic fractionation performed on oleaginous Chlorella sp. at 320 C. 2.6 mg bio-oil was dissolved in 1mL of methanol for analyses. Confidence levels of products identified in the GC-MS chromatogram are shown in table 3.5...... 61

3-14: Total- and component- mass balance data obtained from bench-scale fixed-bed experiments for Step 2 pyrolysis of oleaginous Chlorella sp. at 420 °C. Step 2 was performed on residues from pyrolysis Step 1. †indicates lipid is quantified as FAMEs; ‡ indicates lipid is sum of hydrocarbons, free fatty acids, fatty amides and fatty nitrile. ... 62

3-15: GC-MS chromatogram of bio-oil collected from Step 2 of fixed-bed pyrolytic fractionation performed on oleaginous Chlorella sp. at 420 C. 5.0 mg of bio-oil was dissolved in 1 mL of chloroform for analyses. Confidence levels of products identified in the GC-MS chromatogram are shown in table 3.6...... 63

xviii

3-16: GC-FID chromatogram of bio-oil collected from Step 2 of fixed-bed pyrolytic fractionation performed on oleaginous Chlorella sp. at 420 C...... 63

3-17: Mass fraction of bio-oil collected during Step 2 of bench-scale fixed-bed pyrolytic fractionation experiments with oleaginous Chlorella sp. at 420 °C...... 65

3-18: Conceptual process design of the pyrolytic fractionation system showing mass balances and potential product pathways to fuels and co-products...... 67

3-19: (a) Differential heat flux and (b) cumulative heat flux data for Chlorella sp...... 69

3-20: ChemcadTM flowchart simulating hydrothermal liquefaction process without heat recovery system...... 78

3-21: ChemcadTM flowchart simulating hydrothermal liquefaction process with heat recovery system...... 79

4-1: (a) Residual weight (dashed lines) profiles and (b) derivative weight loss (solid lines) of triolein, tripalmitate and tristearate obtained during heating under N2 atmosphere to determine their volatilization temperatures...... 88

4-2: Residual weight (solid lines) and temperature (dashed lines) of (a) tripalmitate, (b) tristearate and (c) triolein with time obtained during dynamic heating from room temperature to 420 °C and maintained isothermally at 420 °C for 30 min under N2 atmosphere to determine the time required for complete volatilization of substrates...... 89

4-3: Derivative weight loss (solid lines) and residual weight (dashed lines) profiles represents the pyrolysis of (a) Sunflower seeds, (b) Chlorella sp., (c) Scenedesmus sp., and (d) Schizochytrium sp.. Derivative weight loss peak in the temperature interval of

370-480 °C indicates volatilization of triglyceride fraction present in these feedstocks. . 91

xix

4-4: Residual weight (dashed lines) and temperature (solid lines) of (a) tripalmitate, (b) tristearate and (c) triolein with time obtained while employing temperature protocol to determine their triglyceride content...... 93

4-5: Residual weight (dashed lines) and temperature (solid lines) of (a) Sunflower seeds,

(b) Chlorella sp., (c) Scenedesmus sp., and (d) Schizochytrium sp. with time obtained while employing temperature protocol to determine their triglyceride content...... 94

4-6: Derivative weight loss (solid lines) and residual weight (dashed lines) profiles of (a)

Sunflower seeds, (b) Chlorella sp., (c) Scenedesmus sp., and (d) Schizochytrium sp. solid residues obtained after solvent extraction...... 97

4-7: Residual weight (dashed lines) and temperature (solid lines) of wet Chlorella sp. sample with time obtained while employing temperature protocol to determine their triglyceride content...... 99

4-8: Operations or steps involved in different analytical methods for quantifying triglyceride content in oleaginous feedstocks...... 101

Chapter 1

1. Introduction

Steady growth of world population, finite fossil fuel resources, uncertainty of fossil fuel supplies and concerns over the rise of carbon dioxide levels in the atmosphere have increased a global awareness for energy sustainability [1-3]. Solar energy (including photovoltaics and solar-thermal), wind energy, hydropower, geothermal and biomass are the primary renewable sources that could potentially replace fossil-derived energy [4]. Of these, biomass is the only carbon based alternative energy resource [5]. Broadly, biomass includes any organic material such as forest resources, agricultural residue, algae, oilseeds, municipal solid waste, and as well as carbon-rich industrial wastes (e.g. from pulp and paper industry or from food processing) [6]. Based on the constituent biopolymers biomass can also be classified into two broad categories - lignocellulosic biomass and non-ligneous feedstocks. While all biomass feedstocks can potentially be utilized to produce carbon-based liquid transportation fuels and high value chemicals, the technologies and challenges for conversion of lignocellulosic and non-ligneous feedstocks are substantially different.

1.1 Production of liquid fuels and high-value chemicals from

lignocellulosic biomass:

Lignocellulosic biomass is derived from terrestrial plants (e.g. forest resources and agricultural residues) and consists of cellulose, hemicellulose and lignin and the major constituent biopolymers [7]. In native lignocellulosic biomass, cellulose is highly crystalline and embedded in an amorphous hemicellulose matrix. Together, hemicellulose and cellulose are further enveloped by an outer shield of lignin [8]. Cellulose and hemicellulose are carbohydrate polymers that can be hydrolyzed to their monomeric sugars with the assistance of heat and catalysts such as acids and enzymes [9]. To make lignocellulosic biomass amenable for enzymatic hydrolysis, a thermal chemical pretreatment is necessary. Pretreatment disrupts the lignin and/or hemicellulose sheath, opens the biomass structure to allow enzyme penetration and decreases cellulose crystallinity for fast enzymatic hydrolysis. Several pretreatment technologies such as dilute acid pretreatment, ammonia fiber explosion, ionic liquid pretreatment, steam explosion, torrefaction have been reported in the literature [10, 11]. The pretreated biomass has proven to produce high yields of sugar monomer units at low process time

[9-11]. These monomeric sugars units obtained via hydrolysis process can be converted to produce wide range chemicals such as ethanol, succinic acid, levulinic acid, lactic acid, furans by various catalytic processes [12-18]. The steps involved in the production of liquid fuels and chemicals from lignocellulosic biomass are shown in Figure 1-1.

Sugars Ethanol Biomass Pretreatment Hydrolysis Catalytic process Furans Ethylene glycol

Lignin

Figure 1-1: Overview of lignocellulosic-based bio-refinery. Biomass residue left behind after carbohydrate hydrolysis primarily consists of lignin

(usually termed “residual lignin”) and accounts up to 30 wt% of dry weight and 40 % of the energy content of biomass [5, 19]. At present, most process designs use residual lignin for production of low-value process heat through combustion. However, utilization of residual lignin for production of high-value chemicals or transportation fuels would generate additional revenue and could thus significantly improve the overall economics of lignocellulosic-based bio-refineries [19, 20].

1.2 Advantages of algal biomass:

While abundantly available, terrestrial plant derived lignocellulosic materials available in the U.S. can, at best, satisfy only 30 % of the national gasoline consumption, if sustainable cultivated and harvested without adverse environmental impacts [21, 22].

Microalgae (also often referred to as algae) are non-ligneous aquatic photosynthetic microorganisms that can be cultivated as feedstocks to supplement lignocellulosic resources. As feedstocks for energy production, microalgae have several advantages including high areal productivity and ability to grow using low-quality -water (brackish

3 water or sea water) and –nutrients (e.g. from animal or municipal waste). Also, since microalgae are aquatic species, their cultivation does not require fertile land and these feedstocks can be grown in low quality arid/desert lands. As such, microalgae would not compete with resources required for traditional agriculture. Further, several microalgae strains are capable of producing and accumulating large amounts of lipids such as triglycerides or hydrocarbons (up to 50% of cell mass) [23-28] that can be converted to hydrocarbon fuels. However, commercialization of algal bio-refineries faces scientific and commercial challenges such as maintaining high biomass and lipid productivity in over prolonged periods, efficient harvesting methods and feedstock-relevant downstream conversion processes to produce fuels and value-added chemicals [29].

Conventional methods employed for converting algae to fuels involves extraction of triglyceride in organic solvents followed thermo-chemical/thermo-catalytic processes to produce fuels and oleo-chemicals [30-33]. In some cases, additional cells disruption is required through mechanicals operations such as sonication or ultra-pressing [34-36].

Alternatively, triglycerides in microalgae can be converted to fatty acid methyl esters

(FAMEs) or biodiesel through in situ acid/base-catalyzed transesterification [37, 38].

Overall, existing methods for require chemicals or organic solvents and chemicals such as hexane, methanol and sulfuric acid as well as involve multiple energy-intense steps.

While most of the current research focus in production of algal biofuel is focused on utilizing lipid-rich algae and producing biodiesel [39], it is important to recognize that oleaginous algal species have slow growth rates and also need expensive infrastructure to prevent contamination with lipid-lean and fast growing “rogue algae.” Some genetic engineering approaches to enhance productivities of lipid content of otherwise non-

4 oleaginous have recently been reported [40, 41], but the GMO strains could face regulatory challenges for large commercial application [39]. An alternate approach to generate fuels from lipid-lean algae could be through application of thermochemical methods such as pyrolysis [42]. Even though there could some differences in the cell structure and relative mass of component biopolymers (starch, protein and lipid), pyrolysis conditions could be adjusted to generate bio-oils from most, if not all, feedstocks.

Gases Gases Biomass Bio-oil Bio-char Bio-char

Figure 1-2: Reactions occur during biomass pyrolysis Pyrolysis is the thermal degradation of organic matter by means of heat with or without the aid of a catalyst in the absence of air or oxygen[43]. It involves cleavage of chemical bonds to yield shorter carbon chain molecules primarily from the depolymerization and fragmentation reactions of biomass components [44]. Biomass pyrolysis in an inert atmosphere produces liquid products, non-condensable gases (such as CO2, H2, CO, CH4), and solid residue (biochar) [45]. Two types of reactions occur during pyrolysis – primary and secondary (see Figure 1-2). During primary reactions, biomass thermally degrades to produce bio-oils, biochar and non-condensable gases.

During secondary reactions, bio-oils formed from primary reactions further undergo degradation to produce biochar and gases. The yields and composition of pyrolysis

5 products depends on type of feedstocks used, hot vapor residence time in the pyrolysis reactor, particle size of feed, heating rate, pressure and temperature of reactor [6].

Pyrolysis of lignocellulosic feedstocks has been extensively studied [46].

Table 1.1: Chemical compounds formed during pyrolysis of carbohydrates, lignin, protein and triglycerides.

Compounds formed in Source Biopolymer the pyrolysis oil Aldehydes, ketones and Carbohydrate alcohols

Protein N- compounds

Triglycerides Fatty acids, alkanes

Lignin Phenolic compounds

Bio-oil, the liquid product from pyrolysis, is a multi-component mixture comprised of different size molecules [47]. Chemical composition of bio-oils depends upon the biopolymers present in biomass (see Table 1.1). Carbohydrates upon pyrolysis produce aldehydes, ketones, organic acids, alcohols while proteins produce N- compounds. Pyrolysis of lignin produces phenolic compounds whereas triglyceride pyrolysis produces fatty acids and hydrocarbons. These bio-oils can be upgraded through hydro-treating and hydro-cracking to liquid hydrocarbon fuels [48]. Bio-oils can also be used to produce heat and power through boilers and can serve as feed for the production of hydrogen by steam reforming [5]. A variety of important applications can be envisioned for bio-char as well. Addition of bio-char to soil could potentially replace the underground injection of CO2 for carbon sequestration. It can also be used to improve the

6 soil quality by increasing the retention time of nutrients and agrochemicals [49].

Moreover, high ash bio-chars have been shown to adsorb heavy metals [50]. Bio-char also forms a precursor for producing valuable materials: it can be used for the production of carbon fibers, activated carbon and carbon nano-tubes [50]. Bio-char can be further gasified at high temperature to produce hydrogen or syn-gas (CO+ H2) by steam reformation and gasification.

Pyrolysis of algal biomass has been demonstrated previously in fixed-bed and fluidized bed reactors [42, 51-61] and these studies show the feasibility of production of bio-oils from both lipid-rich and lipid-lean algae. While bio-oil yields from oleaginous algae were higher, pyrolysis of lipid-lean algae resulted in a higher fraction of char.

Based on these studies, it could be inferred that algal biomass from natural blooms might also result in useful products via pyrolysis. Other studies have simulated the pyrolysis process on analytical-scale pyrolysis instruments. Ross et al., 2008 analyzed brown algae pyrolysis oils by pyroprobe-gas chromatography (GC) equipped with mass spectrometry

(MS) and reported N-compounds in addition to carbohydrate degradation products.

Similar conclusions were reported for aquatic species and/or macro algae like duck weed

[62, 63]. Further studies concluded that bio-oils produced from pyrolysis of macro-algae and microalgae may not be attractive for the production of fuels due to the high nitrogen content. N-containing fuels are of low quality and might not pass fuel standards due to formation of NOx upon combustion [64, 65]. In addition, N-compounds of algal bio-oils may also poison the catalyst used for hydrogenation or de-oxygenation processes. This disadvantage with regards to pyrolysis of non-ligneous biomass, in particular algal

7 feedstocks, encourages us to propose pyrolytic fractionation process by exploiting its thermal/pyrolytic characteristics.

1.3 Research Outline:

The goal of this dissertation is develop strategies for production of fuels and chemicals through pyrolysis of algal feedstocks. The research objectives involved in this dissertation are as follows:

1.3.1 Comparative study of pyrolysis products of algal and lignocellulosic

feedstocks:

Pyrolysis experiments were performed with algae, lignocellulosic feedstocks and residual lignin (from ionic liquid pretreatment followed by enzymatic digestion; see appendix 1) under similar reactor conditions for comparison of product (bio-oil, gas and bio-char) yields and composition. In spite of major differences in component biopolymers, feedstock properties relevant to thermochemical conversions, such as overall

C, H and O-content, C/O and H/C molar ratio as well as calorific values, were found to be similar for algae and lignocellulosic material. Bio-oils obtained were analyzed by GC-

MS and TG analysis. Bio-oil yields from algae and some lignocellulosic materials were similar; however, algal bio-oils were compositionally different and contained several N- compounds (most likely from protein degradation). Algal bio-char also had a significantly higher N-content. Overall, our results suggest that algal cultures deficient in lipids such as nuisance algae obtained from natural blooms can be thermo-chemically converted into fuels with efficiencies and yields similar to lignocellulosic materials. As

8 such, pyrolysis technologies being developed for lignocellulosic biomass may be directly applicable to algal feedstocks as well.

1.3.2 Pyrolytic fractionation of non-ligneous biomass especially oleaginous and

algal feedstocks:

Conventional pyrolysis of oleaginous microalgal biomass results in simultaneous thermal degradation of constitutive biopolymers – protein, carbohydrate and triglycerides

– to produce bio-oils that contain a heterogeneous mixture of water, acetic acid, N- compounds, oxygenated carbohydrate degradation products and fatty acids. Due to the low pH of the bio-oils and presence of water, these bio-oils are not directly usable as fuels. In this paper, we describe a “pyrolytic fractionation” technique to produce two homogenous bio-oils from microalgae. One bio-oil fraction contains water, acetic acid,

N-compounds and oxygenated carbohydrate degradation products while other fraction has fatty acids, fatty nitriles, fatty amides, long chain hydrocarbons, di-glycerides and triglycerides. This method is based on the observation that triglyceride degrades at higher temperature than that of protein and carbohydrate. Our studies show that at relatively low temperatures of 320 °C, the protein as well as carbohydrate fraction of microalgal biomass is preferentially degraded to produce water, acetic acid, N-compounds (which include high-value indoles and pyrimidines) and oxygenated compounds (such as levoglucosans, furans and other sugar degradation products). Triglyceride pyrolysis occurs when sample temperatures exceed 350 °C. Interestingly, our results showed that thermal degradation of triglycerides produced hydrocarbons and free fatty acids that were nearly free water, acetic acid and carbohydrate degradation products suggesting that the triglyceride pyrolysis products could directly be converted to drop-in fuels

(hydrocarbons) if combined with gas-phase hydrodeoxygenation. Alternately, reaction of the free fatty acids vapors with methanol in an integrated gas-phase process could result in direct production of fatty acid methyl esters (biodiesel). Energy requirements for pyrolytic fractionation, estimated through differential scanning calorimetry measurements, are also presented.

1.3.3 Triglyceride quantification of oleaginous algal biomass using thermo-

gravimetry:

Laboratory analytical techniques employed for quantification of triglyceride content in oleaginous biomass samples, especially microalgae, are time consuming, require volatile organic solvents, and involve multiple steps. We have observed that, when heated in an inert atmosphere, triglycerides volatilize over a narrow temperature interval of 370-460 °C, with negligible solid residue. Degradation and volatilization of the other constituents of oleaginous biomass (protein and carbohydrates) is largely complete at temperatures below 350 °C. Based on these observations, we have estimated the triglyceride content of microalgae by measuring net mass loss of samples during dynamic heating from 322 °C to 420 °C followed by isothermal heating at 420 °C under an inert atmosphere. Our results indicate that the triglyceride content determined using thermogravimetric (TG) method is statistically similar to estimates obtained from conventional gas-chromatography based quantification techniques.

Chapter 2

Comparative study of pyrolysis of algal biomass from natural lake blooms with lignocellulosic biomass

1.4 Introduction

Thermo-chemical conversion of lignocellulosic biomass is an area of intense research, development and commercialization activity with several pilot scale facilities currently engaged in scale-up and demonstration [66]. Bio-oils from biomass pyrolysis could be upgraded to liquid hydrocarbon fuels through well-established hydro-treating and/or hydro-cracking processes [48, 67, 68]. Alternatively, bio-oil could be converted to hydrogen by steam reforming or directly combusted in boilers for heat and electricity

[69]. Syngas (CO+ H2) from biomass pyrolysis could also be converted to liquid fuels by

Fisher-Tropsch processes or combusted directly for power generation [68]. The solid residue remaining after biomass pyrolysis, bio-char, is useful as a soil amendment and is known to improve soil quality by increasing the retention time of nutrients and agrochemicals [50]. Since bio-char is virtually non-biodegradable, such applications also result in carbon sequestration. Bio-char could also be used as a feedstock for producing

11 valuable materials such as carbon fibers, activated carbon and carbon nano-tubes [50].

Steam reformation/gasification of bio-char into hydrogen or syngas is also possible [70].

Algae hold promise as a biofuel feedstock that can complement lignocellulosic biomass [71, 72]. Traditionally, algae have been considered for their lipid-production potentials. However, high-lipid containing algae grow slowly and require stringent and controlled cultivation. In open systems, slow growing lipid-rich cultures are likely to be contaminated and rapidly overtaken by fast-growing “rogue algae” that are generally deficient in lipids. Pyrolysis is an especially attractive option for fuel production from such lipid-lean algae [73] and could also be applied to post-extraction residues from lipid-rich algae.

Algal blooms, commonplace in several water bodies, are one potential source of algal biomass. Ordinarily, algal blooms are detrimental to local ecosystems as well as economy [74]. It is estimated that in the US, annual economic losses of nearly $2.2 billion and $100 million, respectively, are incurred due to eutrophication of freshwater systems and coastal algal blooms [74, 75]. Although the overall concentrations of algae in such large water bodies is relatively low, some promising low cost technologies for recovery of algal biomass from sources such as lakes have recently been reported [76].

Thus, it might be possible to recover nuisance algae and profitably use the biomass as feedstock for renewable bio-fuels and products. An additional benefit would be mitigation of eutrophication [77].

The few studies that have previously reported on pyrolysis of microalgae show the feasibility of production of bio-oils and bio-char but have focused on treatment of

12 laboratory-grown cultures [42, 51, 65]. However, pyrolytic behavior of naturally- occurring lipid-lean microalgal assemblages is not known. Some studies have reported pyrolysis of other aquatic flora, especially macroalgae [62, 64, 78-81]. However most of these higher plant species contain larger amounts of more thermostable structural carbohydrates unlike microalgae. Additionally, the studies that report results of pyrolysis of microalgae or other aquatic species do not include comparisons of results with lignocellulosic biomass for which a much larger body of literature and economic feasibility studies are available [46]. In this study we report the results of pyrolysis of microalgae from freshwater blooms and select lignocellulosic feedstocks (corncobs, woodchips and rice husk). A comparative assessment of the product yields from all biomass types, pyrolyzed under similar conditions, is presented.

1.5 Methods and materials

1.5.1 Raw materials and sample preparation

Lignocellulosic biomass feedstocks (corncobs, woodchips and rice husk) were supplied by Red-Lion Bio-energy (Toledo, OH). Prior to use in pyrolysis experiments, these feedstocks were milled to a particle size of -80 mesh using a laboratory Wiley mill

(Model 4, Thomas Scientific, Swedesboro, NJ) and dried in an oven at 45 °C for 24 h.

Samples of algal biomass were collected from two distinct blooms in Maumee bay of Lake Erie on 31st July 2009. It was determined that the Lyngbya sp. dominated in one bloom while the other bloom primarily consisted of Cladophora sp. Both algal samples were filamentous in nature. The collected algal samples were thoroughly washed with distilled water and then dried in an oven at 45 °C for 24 h. Dry algal biomass was

13 also milled down to a particle size of -80 mesh. All biomass samples were stored in desiccators until they were used.

Bovine serum albumin, lysozyme and corn starch were obtained from Sigma-

Aldrich (St. Louis, MO).

1.5.2 Thermo-gravimetric analysis (TGA) of feed stocks

TGA experiments were performed in a TA Instruments Q 50 series Thermo- gravimetric Analyzer (Schaumburg, IL). Approximately 10 mg of sample was heated from room temperature to 600 °C at a heating rate of 10 °C/min under 40 mL/min N2 flow.

1.5.3 Experimental set-up

A schematic of the experimental set-up is shown in Figure 2-1. Pyrolysis experiments were conducted in a stainless steel tubular reactor (L = 43 cm, OD = 2.54 cm) placed in a vertical split shell electric furnace (Applied Test Systems Inc., Butler,

PA). A K-type thermocouple remained in contact with the biomass during the experiments to directly measure the temperature inside the pyrolysis chamber. The outlet of the reactor was connected to a glass condenser that had a continuous flow of refrigerated water (4 °C) as coolant. Non-condensable vapors exiting from the reactor were routed to a gas-chromatography (GC) system (Shimadzu 2010 series, Columbia,

MD) equipped with a pulsed discharge helium ionization detector (PDHID) and Supelco

Carboxen 1010 plot column (Sigma-Aldrich, Allentown, PA) for online measurement of gas composition. Ar and He were continuously passed through the reactor during pyrolysis to maintain oxygen-free conditions. The flow rates of He and Ar were

14 maintained at 200 and 20 cm3/min, respectively, using mass flow controllers (model

316L MCS, Alicat Scientific, Tucson, AZ). Helium also served as the carrier gas for vapors passing through the GC-PDHID while addition of Ar served as an internal standard and enabled quantitative analysis of the gaseous products from pyrolysis reactor.

The reactor and glass condenser were connected by ¼″ stainless steel tubing that was maintained at pyrolysis temperature using heating tape to prevent in-line condensation.

⅛″ stainless steel tubing was used for all other connecting lines. Lines carrying He and

Ar were first routed through the pyrolysis furnace to preheat the gases before entering the reactor.

Fixed bed pyrolysis reactor

T1 - thermocouple

Figure 5 Schematic diagram of lab-scale pyrolysis set-up

Figure 2-1: Schematic diagram of lab-scale pyrolysis set-up. Before the start of each experiment, 6-10 g of biomass was placed in the tubular reactor using quartz wool as a support and the reactor was purged with He (200 cm3/min)

15 for 15 min to remove air from the system. Thereafter, the pyrolysis furnace was heated to set-point temperature at a ramp rate of 30 °C/min (verified through monitoring the thermocouple readout). We observed that all biomass samples in the pyrolysis reactor reached reaction temperatures within 20 min. After reaching set-point, the reactor was maintained at that temperature for 20 min. Our preliminary experiments showed that, within our experimental system, this duration was sufficient for the pyrolysis reactions to be complete. Based on the dimensions of pyrolysis reactor and gas flow rates used (He and Ar) in our studies, the vapor residence time (reactor to condenser) was calculated to be 33.3 s.

At the end of the experiment, the pyrolysis reactor was cooled to room temperature and the solid char as well as bio-oils (collected in the condenser) were weighed (to calculate corresponding product yields) and stored at -20 °C for subsequent analyses as described below. The gas yields were determined by subtracting the mass of bio-oil and bio-char collected from the initial mass of feedstock added to the reactor.

1.6 Analytical methods

1.6.1 Compositional analysis:

The carbohydrate fraction of algal samples was determined by 2-stage acid hydrolysis according to the standard laboratory analytical procedures developed by the

National Renewable Energy Laboratory (NREL)[82]. Bligh & Dyer method was used to obtain the lipid content of algal samples [30]. Protein fraction was estimated based on the nitrogen content (wt%) of the biomass using a nitrogen factor (NF) of 6.25 [83].

Protein fraction of all the feedstocks was calculated as follows:

푤푡% 표푓 푝푟표푡푒𝑖푛 = 푤푡% 표푓 푛𝑖푡푟표푔푒푛 ∗ 푁퐹 (1)

1.6.2 Proximate analysis

Proximate analysis included measurement of moisture content, volatile matter, fixed carbon and ash. For moisture content measurement, biomass samples were dried in a convection oven at 105 °C until constant weight was recorded. Moisture content was calculated from the weight loss and represents water that may be physically present or chemically bound in the biomass with the exception of mineral hydrates [84]. Volatile matter measurements were performed in the TGA analyzer. Volatile matter content was determined by measuring weight loss after heating biomass samples to 575 °C in an Al crucible for 7 minutes under N2 atmosphere [84]. Ash content was measured by heating samples at 575 ± 25 °C for 24 ± 6 h to constant weight in a muffle furnace [85]. The fixed carbon fraction was calculated by subtracting the percentages of volatile matter, moisture content and ash from 100%.

1.6.3 Ultimate analysis

Elemental analysis was performed using Perkin-Elmer 2400 Series II CHN

Elemental Analyzer (Waltham, MA). Based on the elemental composition, high heating values (HHV) were calculated using the following well-established correlations [86]

퐻퐻푉 (푂퐿푆) = 1.87 퐶2 – 144 퐶 – 2802 퐻 + 63.8 퐶 ∗ 퐻 + 129 푁 + 20147

(2)

퐻퐻푉 (푃퐿푆) = 5.22 퐶2 – 319 퐶 − 1647 퐻 + 38.6 퐶 ∗ 퐻 + 133 푁 + 21028

(3)

Where, C, H and N denote the mass fractions of carbon, hydrogen and nitrogen within the sample. Eq. (2) and (3) provide empirical estimates based on ordinary least square (OLS) and partial least square (PLS) regression analysis of heating value data available in the literature for a wide variety of substrates. Per the suggestion of Friedl et al. (2005)[87], average values of HHV were calculated and are reported here.

HHV(OLS) + HHV(PLS) 퐻퐻푉 = (4) 2

1.6.4 Analysis of bio-oil

Bio-oils produced from pyrolysis experiments were analyzed using a Hewlett

Packard 6890 Gas Chromatograph equipped with 5973 series mass selective detector

(GC-MS) (Agilent Technologies, Santa Clara, CA). A PTE-5 (30m ×0.25mm×0.25μm film thickness) fused silica capillary column supplied by Supelco (Sigma-Aldrich,

Allentown, PA) was used for the analyses. GC carrier gas was He (1.4 mL/min). The temperature program was as follows: constant temperature of 40 °C for 4 min followed by temperature ramp to 280 °C at 5 °C/min and finally a constant temperature of 280 °C for 15 min. The injector and MS temperatures were maintained at 270 °C. 200 mg of bio- oil samples were dissolved in 1 mL of methanol for the GC-MS analysis. Chemical compounds corresponding to each peak in bio-oil chromatograms were identified by using NIST98 mass spectral database.

1.7 Results & Discussion

1.7.1 Feedstock characterization

The results of proximate analysis of feedstocks are shown in Table 2.1. The volatile matter and fixed carbon content values of the lignocellulosic feedstocks reported here (corncobs, woodchips and rice husk) compare well with values previously reported in the literature [1]. The ash content of algae samples was observed to be higher than that of lignocellulosic biomass (except rice husk) and is consistent with previous observations of high ash content in aquatic flora from natural ecosystems, including micro- and macroalgae [78, 79, 88]. High ash content in rice husk is likely due to the presence of silica

[89]. Based on similarities in volatile matter and fixed carbon, the yields of pyrolyzed products can be expected to be similar between algae and lignocellulosic biomass.

However, the presence of high amounts of ash in algae (as well as in rice husk) could lead to more char formation since inorganic elements in the ash are known to catalyze the formation of char during pyrolysis[1].

Table 2.1: Proximate analysis and compositional characteristics of feedstocks used in the present study. Carbohydrates include cellulose and hemicellulose in the lignocellulosic materials and starch in algal biomass. Lignin is not present in the algal feedstocksTable 1 . Proximate. The values analysis reported and compositional are mean characteristics of three ofsamples feedstocks and used expressed in the as present study. Carbohydrates include cellulose and hemicellulose in the lignocellulosic mass fractionsmaterials (%) and starchof the in originalalgal biomass. samples. Lignin is Standardnot present in deviations the algal feedstocks. were < The 8% for all reportedvalues values. reported are mean of three samples and expressed as mass fractions (%) of original samples. Standard deviations were < 8% for all reported values.

Proximate analysis Compositional analysis Feedstock Ash Volatile Fixed Carbohydrate Moisture Lignin Lipid Protein matter carbon s Corncob 6.5 73.5 15.7 4.3 81.1a 10.9a - 4.3

Woodchips 15.1 65.9 17.5 1.5 62.2b 23.7b - 0.6

Rice husk 7.7 57.3 15.3 19.7 56.1c 13.3c - 3.1

Lyngbya sp. 2.4 55.6 16.3 25.7 13.25 - 1.39 29.9

Cladophora.sp 5.91 64.1 16.7 13.3 24.8 - 5.8d 24.6d

a Pan et al, 2010 b DOE Biomass feedstock composition and property database c Zaid and Ganiyat, 2008 d Gottumukala, 2010

Table 2.2 shows the elemental composition of the feedstocks used in pyrolysis experiments. On an ash-free basis, overall C and H fractions, molar C/O and H/C ratios as well high heating values (HHV) of all the feedstocks are statistically similar. However, the N content of the algal feedstocks was higher in comparison to lignocellulosic biomass, most likely due to higher protein content in algae. The N values for the algae species tested in this study are in the same range of those previously reported in literature

[25, 42, 78, 79, 88]. Hence, presence of nitrogenous compounds (formed due to thermal degradation of proteins) could be expected in pyrolysis products of algal biomass. Table 2 . Elemental analyses of all feedstocks. “As-received” values indicate measurements directly obtained from the CHN analyzer. “Dry basis” and “dry-ash free basis” values were calculated from “as-received” values after discounting moisture and ash content, respectively. Calorific values (HHV) were calculated using equations 2, 3 and 4. Algae samples were dried (after dewatering) at 45 C for 24 h prior to analysis. The values reported are average of two samples and expressed as mass fractions (%). MassTable fraction of 2 oxygen.2: Elemental was calculated analyses by difference. of Standard all feedstocks deviations were "As< 2% for-recieved" all reported values. values indicate

Corncob Woodchips Ricehusk Lyngbya sp. Cladophora sp.

Dry, As- As- Dry As- Dry Dry, As- Dry Dry, Dry Dry, ash- As- Dry Dry, ash- ash- receive received basis received basis ash-free received basis ash-free basis free received basis free free d Carbon 43.1 46.0 48.2 44.9 52.9 53.9 39.3 42.6 54.2 36.4 37.3 50.6 36.6 38.9 45.3

Hydrogen 6.1 6.5 6.9 6.0 7.1 7.2 5.3 5.7 7.3 5.7 5.8 7.9 5.8 6.1 7.1

Nitrogen 0.7 0.8 0.9 0.1 0.1 0.1 0.5 0.5 0.6 4.1 4.2 5.8 2.5 2.7 3.1 Oxygen 39.2 41.9 43.9 32.2 37.9 38.6 27.5 2029.8 37.8 25.6 26.2 35.6 35.9 38.1 44.4 C/O 1.5 1.5 1.5 1.8 1.8 1.8 1.9 1.9 1.9 1.9 1.9 1.9 1.4 1.4 1.4 H/C 1.7 1.7 1.7 1.6 1.6 1.6 1.6 1.6 1.6 1.9 1.9 1.9 1.9 1.9 1.9

HHV (MJ/kg) 17.1 18.4 19.5 17.8 21.7 22.3 15.9 17.0 22.5 15.3 15.6 21.6 15.1 15.8 18.5 measurements directly obtained from the CHN analyzer. "Dry basis" and "dry-ash free basis" values were calculated from "as-recieved" values after discounting moisture and ash content, respectively. Calorific values (HHV) were calculated using equation 2, 3 and 4. Algae samples were dried (after dewatering) at 45 C for 24 h prior to analysis. The values reported are average of two samples and expressed as mass fractions (%). Mass fraction of oxygen was calculated by difference. Standard deviations were <2 % for all reported values.

Our estimates of calorific values of feedstocks used in this study are also consistent with previously reported data [1, 42, 78, 79, 88]. To verify the validity of HHV calculations for algal biomass, we substituted the C, H and N values of Chlorella protothecoides reported by Miao et al. (2004)[51] in equations 2,3 and 4 and calculated the calorific value to be 29.8 MJ/kg. This corresponds closely to the experimentally measured calorific value of 30 MJ/kg reported by Miao et al. (2004), thus validating the correlations used in this study. In addition, our HHV estimates (using Eqs. 2, 3, and 4) for the feedstocks studied by Demibras (2010)[81] also closely correspond to the experimentally measured values reported in that study.

1.7.2 Thermo-gravimetric analysis of feedstocks

Thermo-gravimetric analysis (TGA) was performed to obtain an a priori estimate of pyrolysis behavior of feedstocks being tested and the thermograms of all feedstocks are shown in Figure 2-2. For lignocellulosic feedstocks (corncobs, woodchips and rice husk), the thermal degradation profiles showed similar patterns and consisted of one major derivative weight loss peak (dw/dT, where w is the sample weight and T is the temperature) at ~350-370 °C preceded by a shoulder (Figure 2-2(a), (b) and (c)). The primary classes of thermally degradable biopolymers in lignocellulosic biomass are cellulose, hemicelluloses and lignin (Table 2.1). Upon heating in an oxygen-free 21 atmosphere, cellulose and hemi-cellulose volatilize due to the thermal cleavage of the carbohydrate monomers units. Hemi-cellulose is more thermally labile and decomposes first in the temperature range of 220-315 °C (seen as the shoulder that precedes the derivative weight loss peak in Figure 2-2(a), (b) and (c)). Cellulose, more thermally stable due to its crystalline structure [90], decomposes at a higher temperature of 315-400

°C (see peaks shown with arrows in Figure 2-2(a), (b) and (c)). In contrast, lignin is a significantly more heterogeneous polymer (relative to cellulose or hemicelluloses) and decomposes over a much wider temperature window of 190-900 °C. Therefore, a distinct weight loss derivate peak is not discernible for lignin [44, 91]. Moreover, a large portion of the thermal degradation temperature interval of lignin overlaps with that of cellulose and hemicellulose.

Thermograms of the two algal feedstocks also showed one major derivative weight loss peak (Figure 2-2(d) and (e)); however, the characteristics of the peak are distinctly different between the two algal samples. The derivative weight loss peak for

Cladophora sp. (Figure 2-2(e)) is much broader than that of Lyngbya sp. (Figure 2-2(d)) suggesting that the thermal degradation of Cladophora sp. occurs over a much wider

22 temperature range than Lyngbya sp..

Figure 2-2: Weight loss (dashed lines) and its temperature derivative (solid line) of (a) Corncobs (b) Wood chips (c) Rice husk (d) Lyngbya sp. (e) Cladophora sp. (f) Bovine serum albumin (g) Lysozyme (h) Corn starch. Comparing Figure 2-2(d) and (e), it can be seen that the derivative weight loss peak of Cladophora sp. is shorter than that of Lyngbya sp., thus suggesting that

Cladophora sp. has a lower rate of thermal degradation. (Note: The derivative weight

dw  dw   dT  loss (dw/dT) can be expressed as     , and thus the derivative peaks of dt  dT   dt 

23 the thermogram also reflect a kinetic rate of degradation for a constant rate of change of temperature (dT/dt), as in our experiments).

We hypothesized that these differences in the thermograms between Cladophora sp. and Lyngbya sp. could be due to differences in the thermal degradation characteristics of proteins – a major biopolymer in these microalgal samples (Table 2.1). To verify if thermal degradation behavior was protein-specific, we performed TGA analyses on two pure proteins – BSA and lysozyme. Thermograms of BSA and lysozyme show clear differences in the weight loss profiles (shown in Figure 2-2(f) and (g), respectively). BSA degrades over a smaller temperature range (150-350 °C) and at a faster rate (1.1wt%/ °C) than lysozyme (200-500 °C, 0.5 wt%/ °C). In addition, a greater fraction of lysozyme mass remains non-degradable (~40%) than BSA (~20%). Although the proteins studied here are not directly derived from microalgae, the differences observed with these pure proteins clearly illustrate that thermal degradation characteristics of polypeptides are protein-specific. As such, different microalgal species might be expected to show variances in pyrolysis behavior depending upon the composition of constituent proteins

(Figure 2-2(d) and (e)).

We also performed TGA experiments with pure starch (Figure 2-2(h)) and consistent with previous reports [92], observed that most of the degradation occurred between 310-360 °C with a single derivative weight loss peak. Comparing thermograms of pure proteins and starch (Figure 2-2(f), (g) and (h)), it can be concluded that at least some degree of overlap exists between thermal degradation temperatures of these biopolymers. The simultaneous degradation of proteins and carbohydrates was most likely the reason for only single peaks observed on thermograms of Lyngbya sp. and

Cladophora sp. (Figure 2-2(d) and (e)) at temperatures <400 °C. Cladophora sp. also appears to undergo additional thermal degradation at high temperatures (505-600 °C). We speculate that this second degradation region might be due to presence of some thermally stable proteins (or other unknown biopolymers) in Cladophora sp.

Overall, while the TGA profile of Lyngbya sp. is similar to that of other cyanobacterial species such as Synechococcus sp., thermal decomposition of Cladophora sp. did not follow the same trend as reported for other green microalgae [93] due to additional degradation at high temperatures.

1.7.3 Pyrolysis experiments

50 Bio-char Bio-oil 40 Gases Ash

Product yields (wt %) 10

0 Corn Cobs Wood Chips Rice Husk Lyngbya sp. Cladophora sp.

Figure 7 Yields of bio-oil, ash-free bio-char, ash and gases obtained after Figure 2-3: Yields of bio-oil, ash-free bio-char, ash and gases obtained after pyrolysispyrolysis at at 600 600 °C Cfor for all all the the feedstocks feedstocks tested. tested. Ash Ashwas assumedwas assumed to stay to stay associatedassociated with with the the solid solid residue residue and and subtracted subtracted from from residue residue weight weight to obta toin ash- freeobtain bio- ashchar- freevalues. bio All-char reported values. values All reported are an averagevalues areof two an averageexperiments. of two Error barsexperiments. indicate one Error standard bars indicate deviation one from standard mean values deviation. from mean values

Since thermal decomposition continued at high temperatures for most of our samples, we performed pyrolysis experiments at 550 and 600 °C. The primary goal of these experiments was to assess the quality and relative yields of gas, liquid and solid products to enable a side-by-side comparison between product characteristics and establish the viability of algal biomass as a pyrolysis feedstock vis-a-vis lignocellulosic materials.

Figure 2-3 shows the yields of bio-char, bio-oil and bio-gases of all feed stocks formed during pyrolysis at a final temperature of 600 °C. The bio-oil yields obtained from lignocellulosic biomass in the present study are close to values reported in the literature under similar fixed bed reactor conditions [94, 95]. Even on an ash-free basis, highest char yields were observed for the feedstocks with highest ash content - rice husk,

Lyngbya sp. and Cladophora sp. It is possible that the presence of high amounts of ash in the rice husk and algae samples led to greater char formation under our experimental conditions since inorganic elements in the ash are known to catalyze char forming reactions during pyrolysis [1]. In the case of lignocellulosic biomass, hemicellulose and lignin are expected to be the biopolymers that contributed most to char formation while cellulose was likely more completely volatilized [91]. For the algal samples, based on our thermograms (Figure 2-2(f), (g) and (h)), it appears likely that both proteins and carbohydrates could have degraded into char.

The elemental analysis of bio-char produced from all feedstocks at 600 °C is shown in Table 2.3. The calorific values of lignocellulosic bio-char are generally higher than the algae bio-char (except rice husk) due to higher carbon content, most likely derived from lignin. Since most of the biopolymers in algae (with the exception of small

26 amounts of lipids in our samples) are oxygen-containing, incomplete de-oxygenation was likely the cause of lower energy content of algal chars. However, nitrogen-rich algal bio- chars could be used as soil amendment agent [49].

Table 2.3: Elemental analyses of bio-char obtained by pyrolysis are expressed in dry and dry-ash free basis. “Dry-basis” values were obtained by CHN analyzer. “dry-ash free basis” values were calculated by using “dry-basis” values, and ash content. Calorific values (HHV) were calculated using equations 2, 3 and 4. All values are reported as mass fractions (%). Mass fraction of oxygen was calculated by difference. The values reported are average of two identical samples. Standard deviations were < 4% for all reported values.

Table 3. Elemental analyses of bio-char obtained by pyrolysis are expressed in dry and dry-ash free basis. “Dry-basis” values were obtained by CHN analyzer. “dry-ash free basis” values were calculated by using “dry-basis” values, and ash content. Calorific values (HHV) were calculated using equations 2, 3 and 4. All values are reported as mass fractions (%). Mass fraction of oxygen was calculated by difference. The values reported are average of two identical samples. Standard deviations were < 4% for all reported values.

Bio-char Corncob Woodchips Ricehusk Lyngbya sp. Cladophora sp.

Dry Dry, Dry Dry, Dry Dry, Dry Dry, ash- Dry Dry, ash- basis ash-free basis ash-free basis ash-free basis free basis free Carbon 73.1 86.4 83.9 89.5 37.4 62.7 28.5 68.8 42.3 62.7

Hydrogen 2.1 2.4 2.8 2.9 1.4 2.33 1.0 2.5 1.5 2.2

Nitrogen 1.3 1.6 0.4 0.4 0.4 0.7 2.8 6.7 2.7 4.0

Oxygen 8.1 9.6 6.6 7.1 20.4 34.3 9.1 21.9 20.9 31.0

C/O 12.0 12.0 16.8 16.8 2.4 2.4 4.2 4.2 2.7 2.7

H/C 0.3 0.3 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.4

HHV (MJ/kg) 25.9 32.6 31.9 35.3 16.5 22.4 16.4 25.6 17.4 22.7

On an ash-free basis, lowest bio-oil yields were obtained from Lyngbya sp.

Interestingly, Lyngbya sp. also had the highest amounts of protein of all the feedstocks tested in this study. In addition, biochar obtained from Lyngbya sp. contained the highest amount of N (Table 2.3). Taken together, these observations suggest that proteins in

Lyngbya sp. are more prone to charring. To test if other pyrolysis temperatures would improve bio-oil production, additional pyrolysis experiments were performed at 530 °C and 650 °C. From Figure 2-4 it can be seen that pyrolysis at lower temperatures resulted in even lower oil yields and greater char formation. At 600 and 650 °C, oil yields were similar; however, lower amounts of char obtained at the higher temperature suggest that cracking of char into volatile products occurred above 600 °C. Overall, the product yields from pyrolysis of algal biomass obtained in the present study are similar to those previously observed for other lipid-lean algae [42, 51] that employed a much higher heating rate (600 °C/s). Thus, product yields of algae pyrolysis appear to primarily depend on the component bio-polymers present in the biomass.

60 Bio-char Bio-oil 50 Gases Ash

Product yields (wt %)

0 530 °C 600 °C 650 °C

Figure 8 Yields of bio-oil, ash-free bio-char, ash and gases obtained after pyrolysis of Lyngbya sp. at different temperatures. Ash was assumed to stay associated with the solid residue and subtracted Figure 2-4: Yieldsfrom residueof bio weight-oil, ashto obtain-free ash bio-free- char,bio-char ash values. and gases obtained after pyrolysis of Lyngbya sp. at different temperatures. Ash was assumed to stay associated with the solid residue and subtracted from residue weight to obtain ash- free bio-char values. The instantaneous gas composition at temperatures of 510 and 600 °C

(corresponding to time points of 15 and 35 min, respectively, after the start of the pyrolysis experiments) are shown in Table 2.4. Higher amounts of H2 and CH4 along with

28 lower concentrations of CO and CO2 at high temperatures was observed for all feed stocks including algal samples suggesting greater extent of deoxygenation at higher temperatures. Our observations are consistent with similar trends previously observed with lignocellulosic biomass [44, 91] as well as algae [81].

Table 2.4: Instantaneous composition of gases (mole %) measured during the course of pyrolysis experiments. Measurements were made at 510 °C (15 min) and 600 °C (35 min).

Bio-oils produced from pyrolysis of feedstocks were analyzed using GC-MS. For all feedstocks, except Cladophora sp., most of the bio-oil components appear to be of relatively high volatility that eluted from the GC column within 35 min. However, several additional peaks were observed in the chromatogram of Cladophora sp. bio-oil beyond a retention time of 35 min suggesting presence of significant quantities of high molecular weight components that were not identifiable using our mass spectrometry method (confidence levels of the IDs provided by the spectral analysis software was less than 60%). Visually, the bio-oil from Cladophora sp. was also more viscous. It is

29 possible that these higher molecular weight compounds might correspond to products obtained from the volatilization of thermally stable protein(s) or bio-polymer(s) that degrade above 510°C (see Figure 2-2(e)).

The identifiable bio-oil components for all feedstocks were broadly classified into three groups corresponding to the source biopolymer (Table 2.5). These classifications were based on literature information about pure biopolymer pyrolysis products [7]. N- containing compounds were assumed to be derived from protein degradation. The first group presented in Table 2.5 consists of compounds most-likely derived from thermal degradation of polysaccharides (cellulose and hemi-cellulose in lignocellulosic biomass; starch in algal feedstocks). Since, both lignocellulosic and algal feedstocks contain polysaccharides, carbohydrate degradation products were observed in all bio-oils. The second group consists of chemical compounds derived by the thermal decomposition of lignin - mostly aromatic compounds and their derivatives. Since algae do not contain lignin, these compounds were not present in the algal bio-oils. Lastly, chemical compounds in the third group consist of the volatilization products of proteins – mostly various amines and their derivatives. Algal bio-oils contained several of these N- compounds, most of which were not present in the bio-oils from lignocellulosic biomass.

Interestingly, protein degradation products from the two algal feedstocks tested here were different suggesting that Lyngbya sp. and Cladophora sp. were composed of different type of proteins.

Table 2.5: Chemical compounds identified in the bio-oils by GC-MS. The identified compounds are classified based on their most-likely source biopolymer - polysaccharide (cellulose, hemicellulose or starch) lignin and protein. “X” marks indicate the presence of the compound in the bio-oil from the corresponding feedstock. Blanks indicate that the compound was not identified in the bio-oil from the corresponding feedstock.

1.7.4 Energy implications of processing nuisance algae

Recent studies have developed energy estimates for lignocellulosic biomass pyrolysis using thermal analysis methods (differential scanning calorimetry and differential thermo-gravimetry). These studies have reported that the energy requirements of diverse biomass feedstocks fall within the relatively narrow range of 200-400 kJ/kg

[96, 97]. The calorific value of the algae used in this study was estimated to be approximately 15 MJ/kg (Table 2-2). Assuming that the energy needs for algae pyrolysis are approximately in the same range as other biomass feedstocks, only a small fraction of energy contained in the feedstock is anticipated to be expended for pyrolysis. However, rigorous energy estimates in industrially-relevant systems will need to be performed to accurately establish the true energy costs of algae pyrolysis.

1.8 Conclusions

Pyrolysis could be a viable alternative for processing nuisance algae, such as those obtained from eutrophic water bodies. N-rich bio-char could be a valuable product from such processes, although presence of N-compounds in bio-oils would likely diminish their fuel value. Alternatively, pyrolysis could also be applied to algae residues after oil extraction from lipid-rich biomass. Since algae pyrolysis can also produce liquid fuels with appreciable yields, technologies being developed for thermochemical conversion of lignocellulosic materials could be applied to algae as well.

1.9 Acknowledgements

The authors would like to thank Center for Innovative Food Technology (CIFT) for financial support through a subcontract from US-AFRO. The authors would also

32 thank Red Lion Bio-energy for providing lignocellulosic biomass and Dr. Thomas

Bridgeman for algal feedstocks.

Chapter 3

Pyrolytic fractionation: A thermo-chemical technique for processing oleaginous (algal) biomass

1.10 Introduction:

There is a growing interest towards development of renewable fuels as a result of increasing global energy consumption, finite petroleum resources and global warming concerns [1, 2, 98]. Non-food biomass materials, such as microalgae, could be viable feedstocks for environmentally sustainable biofuels. In general, microalgae have greater areal productivity than terrestrial plants, can be grown on non-agricultural and marginal lands and can use low quality water and nutrients from waste streams [29, 72, 73].

Several strains are known to accumulate triglycerides – a platform chemical that is in current use for production of biodiesel as well as high value oleochemicals [24, 33, 72,

99]. However, a key bottleneck in the commercial development of algal bio-refineries is a lack of scalable and viable conversion processes that can produce fuels as well as value- added chemicals [24, 29, 100].

The most common downstream processing approach suggested in the literature involves extraction of triglycerides from algal cells using organic solvents such as

34 chloroform and hexane [30, 34, 35, 100-102]. However, due to the microscopic cell size and robust cell walls, additional mechanical disruption is necessary [35, 103]. After extraction, the solvent must be separated, usually through evaporation, to recover the triglycerides. The recovered triglycerides may then be further converted to hydrocarbon fuels via thermo-catalytic de-carboxylation or hydro-cracking [32]. Alternately, if biodiesel is the desired product, fatty acid methyl esters (FAMEs) may be more easily obtained from cellular triglycerides through in situ transesterification where oleaginous biomass is directly reacted with a mixture of methanol and catalyst without prior solvent extraction [104-106]. However, FAME recovery from the reaction mixture would still require solvent extraction (e.g. with chloroform or hexane) followed by solvent evaporation. Solvent extraction methods have, so far, proven effective only with dry biomass and in situ transesterification is likely more sensitive to even small amounts of moisture in the biomass [105]. In methods involving solvent use, the post-extraction solid residues, generally rich in protein, may also need extensive treatment for solvent removal before use as animal feed or fertilizer [107].

As an alternative to solvent extraction, thermochemical conversion processes such as pyrolysis and hydrothermal liquefaction can be employed to obtain bio-oil or bio-crude for subsequent conversion to liquid fuels and value-added chemicals [42, 51, 52, 59, 64,

65, 98, 108-116]. Thermo-chemical methods are expected to be less species-sensitive than solvent extraction. In addition, these processes can produce fuel/chemical precursors from even the non-triglyceride portions of algal cells (e.g. carbohydrates, other lipids and proteins). However, thermo-chemical processes, as traditionally applied, produce bio- oils/bio-crude that contains a complex and highly heterogeneous mixture of chemical

35 compounds – long chain fatty acids from degradation of triglycerides and other cellular lipids, short chain oxygenates (e.g. aldehydes, ketones, organic acids, water and alcohols) from degradation of carbohydrates and N-compounds from protein degradation [33, 65,

113, 117-121]. Oxygenates in bio-oil lower its heating value and degrade/polymerize over time to produce humins or char [122]. In addition, algal bio-oil/bio-crude obtained from traditional thermochemical processes would consist of a broad molecular weight distribution of chemical species – longer chain products from triglyceride degradation and lower molecular weight compounds from degradation of carbohydrate and protein – that would necessitate further distillation into suitable fuel fractions and result in additional energy inputs for fuel production [43].

In this study, we have developed a “pyrolytic fractionation” approach whereby products from pyrolysis of triglycerides - the highest energy component of oleaginous microalgae - are obtained as more homogenous bio-oils. The validity of pyrolytic fractionation was demonstrated with the oleaginous feedstocks – Chlorella sp. and

Scenedesmus sp. Further, energy requirements for the process are estimated and a conceptual design for integrated production of drop-in fuels and oleochemicals through pyrolytic fractionation has been developed.

1.11 Experimental

1.11.1 Feedstocks and chemicals

Chlorella sp. (a natural isolate) and Scenedesmus sp. were heterotrophically grown on glucose using previously described culture conditions [101, 123]. Stationary phase cultures that were rich in triglycerides were centrifuged (2500×g), washed with de- ionized water and freeze-dried (Labconco Freezone 2.5 L bench-top freeze drying

36 system, Kansas city, MO) to obtain feedstocks used in this study. Soy oil was purchased from Spectrum Naturals (Boulder, CO) and used as received. GC-grade 1,3-diolein was purchased from Sigma-Aldrich (St. Louis, MO).

1.11.2 Triglyceride quantification:

10-15 mg of freeze dried algae was added to 1 mL of chloroform in a 2 mL stainless steel bead beating vials with polypropylene plug cap (BioSpec Products,

Bartlesville, OK). Beads used in this analytical method was described in the literature[124]. A Mini-Beadbeater-1 (BioSpec Products, Bartlesville, OK) was used to agitate the stainless steel vials. Each vial was agitated for 20 s at 2500 oscillations per minute and then cooled in an ice bath for 1 min. Total bead beating time was varied from

2 to 45 min. The organic phase in stainless steel vials was transferred to 5 mL glass vial.

Stainless steel vials were then rinsed with 1 mL of chloroform and added to the organic phase collected. This organic phase was filtered and transferred to GC vials for quantification of triglycerides, mono-glycerides, di-glycerides and fatty acids using gas chromatography equipped with flame ionization detector (GC-FID).

1.11.3 Thermo-gravimetric (TG) analysis and differential scanning calorimetry

(DSC)

TG and DSC analyses were performed on a TA Instruments SDT Q600 series analyzer (Schaumburg, IL) that provides simultaneous measurement of weight change and differential heat flow on a single sample. For these measurements, 10-15 mg of biomass was loaded into one alumina crucible while a second identical crucible served as

37 a reference. N2 was used as the carrier gas and also to maintain inert atmosphere. The

-1 flow rate of N2 was kept at 100 mL min .

Overall thermal degradation behavior of biomass feedstocks was determined by heating the samples from room temperature to 600 °C at a constant ramp rate of 20 °C

-1 min under N2 atmosphere.

To simulate pyrolytic fractionation, a two-stage heating protocol was used where samples were successively heated to 320 °C and 420 °C and maintained isothermal for 10 min at each of these temperatures. The 10 min isothermal incubation time at each stage was chosen since little, if any, weight loss was detected after this period. Inter-stage heating rate was 20 °C min-1.

To estimate energy requirements for pyrolytic fractionation, differential heat flux data for biomass samples (normalized by subtracting differential flux values of empty pans) were integrated over time (Eq.1).

푑푄 푄 = ∫ 푑푡 (1) 푑푡

Data obtained over the temperature interval of 150 °C to 420 °C were integrated to discount heat of vaporization of bound-moisture associated with biomass samples at lower temperatures. Numerical integration was performed by the trapezoidal rule using the in-built “cumtrapz” function in MatlabTM.

1.11.4 Pyrolysis experiments

Pyrolysis experiments were performed on a CDS PyroprobeTM 5200 unit (CDS

Analytical, Oxford, PA) connected to a Bruker 450 gas chromatograph (GC) equipped with a 300 series mass spectrometer (MS) (Billerica, MA). An open-ended quartz tube

(1" long) served as a micro-pyrolysis reactor in the PyroprobeTM system. The reactor

38 temperature was set and maintained using a resistively heated a platinum element coiled around the tube. Vapors from pyrolysis were routed through a gas trap packed with

Tenax® adsorbent material. After pyrolysis, the volatiles from the trap were desorbed and sent to the GC-MS for analysis. A heated transfer line connected the trap to the GC injector.

Before the start of the experiment, approximately 1 mg of biomass sample was placed into the quartz tube for pyrolysis. During the experiment, the system environment

(including reactor and trap) was kept inert by applying a continuous helium purge (50 mL min-1). To simulate pyrolytic fractionation, a three stage heating protocol was used similar to the TG-DSC experiments described above. At each stage, the pyrolysis reactor was heated to the desired set point (320 and 420 °C) and maintained isothermal for 15 min. Similar to the TG-DSC experiments, inter-stage heating rate was kept at 20 °C min-

1. The vapors generated during pyrolysis were adsorbed in the gas trap that was held at a much lower temperature of 50 °C to facilitate better retention of the volatiles. Following the completion of each pyrolysis stage, the reactor was allowed to cool down to room temperature while the trap was heated for 7 min to desorb the volatiles for GC analysis.

Helium was used as the purge gas (100 mL min-1). For the first two desorption stages, the trap temperature matched the pyrolysis temperature. However, the trap was heated only up to 350 °C in the third desorption stage since Tenax® degrades above this temperature.

Desorbed volatiles were routed to the GC injector via a transfer line that was also maintained at the same temperature as the trap.

GC-MS analysis was synchronized with the desorption steps. An Agilent DB-

5MS fused silica capillary column (30 m × 0.25 mm × 0.25 μm film thickness, Agilent

Technologies, Santa Clara, CA) was used in the GC. The injector temperature was 300

°C and a split ratio of 1:100 was maintained. Helium, used to purge the trap, also served as the carrier gas (1.0 mL min-1) in the GC column. The temperature program of GC column was as follows: constant temperature of 50 °C for 7 min (to match the time for desorption of volatiles from the trap) followed by temperature ramp to 300 °C at 10 °C min-1 and finally a constant temperature of 300 °C for 3 min. The MS source was maintained at 150 °C. The transfer line (between GC and MS) stayed at 300 °C.

Chemical compounds corresponding to chromatogram peaks were identified using the

NIST2008 mass spectral database. Only compounds with a “confidence” value above 600 were reported.

1.11.5 Experimental set-up:

A schematic of the experimental set-up was previously shown in the literature

[113] (see Figure 2-1). Pyrolytic fractionation experiments were conducted in a quartz tubular reactor (L = 43 cm, OD = 2.54 cm) placed in a horizontal split shell electric furnace (Applied Test Systems Inc., Butler, PA). A K-type thermocouple remained in contact with the biomass during the experiments to directly measure the temperature inside the pyrolysis chamber. The outlet of the reactor was connected to a glass condenser that had a continuous flow of ethylene glycol (-15 °C) as coolant. N2 was continuously passed through the reactor during pyrolysis to maintain oxygen-free conditions. The flow rate of N2 was maintained at 100 mL/min or 1 L/min using mass flow controllers (model 316L MCS, Alicat Scientific, Tucson, AZ). The reactor and glass condenser were connected by ¼″ stainless steel tubing that was maintained at pyrolysis temperature using heating tape to prevent in-line condensation. ⅛″ stainless

40 steel tubing was used for all other connecting lines. Lines carrying N2 was first routed through the pyrolysis furnace to preheat the gases before entering the reactor.

Before the start of each experiment, 5-10 g of biomass was placed in the tubular reactor using quartz wool as a support and the reactor was purged with N2 (100mL/min or

1 L/min) for 15 min to remove air from the system. Thereafter, the pyrolysis furnace was heated to set-point temperature (i.e. 320 and 420 °C) at a ramp rate of 30 °C/min (verified through monitoring the thermocouple readout). We observed that all biomass samples in the pyrolysis reactor reached reaction temperatures within 20 min. After reaching set- point, the reactor was maintained at that temperature for 10 min. Our pyroprobe experiments showed that, this duration was sufficient for the pyrolysis reactions to be complete. Based on the dimensions of pyrolysis reactor and gas flow rates used (N2) in our studies, the vapor residence time (reactor to condenser) was calculated to be 20 s or 5 s at room temperature.

At the end of the experiment, the pyrolysis reactor was cooled to room temperature and bio-oils (collected in the condenser) were weighed (to calculate corresponding product yields) and stored at -20 °C for subsequent analyses as described below. The solid char/ residues remained in the pyrolysis reactor was also collected and stored to perform CHN analysis. Solid residue obtained from pyrolysis at 320 °C was also collected to perform CHN analysis as well as pyrolysis experiments. The gas yields were determined by subtracting the mass of bio-oil and bio-char collected from the initial mass of feedstock added to the reactor.

First set of experiments involve using oleaginous Chlorella sp. (3.1 g) as feed to pyrolytic fractionation experiments. The flow rate of the inert carrier gas was 100 mL/min. During these experiments, the liquid products obtained after pyrolysis at 320 °C were collected while solid residues were not collected. Solids residues obtained at 320 °C were then pyrolyzed again at 420 °C to produce char and triglyceride-based bio-oil. The primary aim of this experiment is to determine the mass balance of triglyceride.

Second set of experiments involve using oleaginous Chlorella sp. (OA) and non- oleaginous Chlorella sp. as feed to pyrolytic fractionation experiments. The flow rate of the inert carrier gas was 1000 mL/min. During these experiments, the liquid products obtained as well as solids residue obtained after pyrolysis at 320 °C were collected to perform further GC, GC/MS, CHN and FAME analysis. Solids residues obtained at 320

°C were then pyrolyzed again at 420 °C to produce char and triglyceride-based bio-oil.

The primary aim of this experiment is to determine the overall mass balance, carbon balance, hydrogen balance, nitrogen balance and triglyceride balance.

1.12 Results and Discussion

1.12.1 TG studies to identify biopolymer degradation temperatures:

420 °C 100 0.7 0.6 80 320 °C

0.5 C)

60 (a) 0.4 40 0.3

0.2 (wt%/ 20

0.1 Residual weight%

0 0 Derivative weightloss 0 100 200 300 400 500 600 Temperature ( C)

100 420 °C 0.9 0.8 80 (b) 0.7

320 °C 0.6 C) 60 0.5 40 0.4

0.3 (wt%/ 20 0.2

Residual weight% 0.1

0 0 Derivative weightloss 0 100 200 300 400 500 600 Temperature ( C)

Figure 3-1: Residual weight (dashed) and derivative weight loss curves (solid lines) obtained during thermal degradation of (a) Chlorella sp. and (b) Scenedesmus sp..

TG analysis was first performed to obtain an a priori estimate of pyrolysis behavior of the Chlorella sp. and Scenedesmus sp. feedstocks being tested. Figure 3-1 show two well-separated derivative weight loss peaks at 320 °C and 420 °C for both of these oleaginous algal samples. Based on previous studies, the derivative weight loss peak at 320 °C likely resulted from degradation of algal proteins and carbohydrates [92,

113, 125, 126], while the derivative weight loss peak at 420 °C could be attributed to volatilization of the triglyceride fraction of algal samples. TG analysis of soy oil (Figure

3-2) confirmed that the peak at 420 °C was associated with pyrolysis of triglycerides.

Figure 3-2: Derivative weight loss (solid) and residual weight (dashed) curves obtained during thermal degradation of soy oil. Since the protein and carbohydrate fractions degrade at lower temperatures (280-

350 °C) than triglycerides in oleaginous feedstocks (370-450 °C), it is expected that a sequential exposure of biomass to these temperature intervals would first result in pyrolysis of proteins and carbohydrates followed by pyrolysis of triglycerides. Further, when condensed separately, this “pyrolytic fractionation” method would produce triglyceride-specific bio-oils with low, if any, contamination by small molecular weight

N- and O- compounds produced from carbohydrate and protein pyrolysis. One additional

44 observation from these experiments is that triglyceride pyrolysis did not produce any measurable residues (Figure 3-2) suggesting that high product yields from triglycerides could be expected upon implementation of pyrolytic fractionation.

1.12.2 Simulation of pyrolytic fractionation by thermo-gravimetry:

100 3.5 3 80 (a)

2.5 )

1 -

60 2 C

40 1.5

1 (wt% 20

Residual weight % weight Residual 0.5 Derivative weight lossweight Derivative 0 0 0 100 200 300 400 500 600 Temperature ( C) 100 3.5 3 80 (b)

2.5 )

1 -

60 2 C

40 1.5

1 (wt% 20

Residual weight % weight Residual 0.5 Derivative weight lossweight Derivative 0 0 0 100 200 300 400 500 600 Temperature ( C)

Figure 3-3: TG profiles resulting from pyrolytic fractionation of the protein as well as carbohydrate portion of (a) Chlorella sp. and (b) Scenedesmus sp.. Yellow residual weight (dashed) and derivative weight loss curves (solid lines) show pyrolysis of the protein and carbohydrate fractions of the biomass (T<.320 °C). Green curves show thermograms of the samples obtained after removal of pyrolyzable protein and carbohydrate which indicate that prolonged exposure to lower temperatures does not negatively impact the thermal decomposition characteristics of the constituent triglycerides. The arrows indicate the temperature path for followed for these experiments.

In these studies, pyrolytic fractionation was simulated on the TG analyzer by step- wise heating. The dotted arrows in Figure 3-3(a) indicate the temperature path implemented for pyrolysis. Samples were first heated to 320 °C and then incubated isothermally (yellow arrows) to pyrolyze the protein and carbohydrate fractions. After 10 min at this temperature, further weight loss was not significant (Figure 3-4) suggesting that pyrolysis of the protein and carbohydrate fractions were substantially complete during this period. Thereafter, when the samples were cooled back to 100 °C and reheated (green arrows in Figure 3-3a), the peak at 320 °C was absent from the thermogram (green differential weight loss profile in Figure 3-3a) confirming the removal of thermally labile protein and carbohydrate fractions. Further heating resulted in triglycerides pyrolysis at 420 °C indicating that the removal of protein and carbohydrate did not have observable effects on the thermal characteristics of triglyceride fractions of

Chlorella sp. (green derivative weight loss curves in Figure 3-3a). Similar results were obtained with Scenedesmus sp. (Figure 3-3b), when subjected to an identical thermal treatment for volatilization of protein and carbohydrate fractions. These observations suggest that prolonged exposure at 320 °C during pyrolytic fractionation did not result in triglyceride degradation or alter the thermal degradation characteristics of triglyceride fraction in the tested biomass samples. Overall, our results suggest that application of pyrolytic fractionation would result in separate recovery of triglyceride-based bio-oils, with net product yields that would be comparable with traditional pyrolysis. Further, the removal of water, organic acids, nitrogenous and oxygenated compounds during pyrolysis at 320 °C would result in low N- and O- content in triglyceride-based bio-oils.

Finally, 10 min of isothermal incubation at 420 °C was sufficient to pyrolyze triglyceride fraction from biomass (Figure 3-5) that had been previously subjected to pyrolytic fractionation for thermal degradation of proteins and carbohydrates.

100 90 Region Region II 80 I (a) 70 60 50

Residula weight % weight Residula 40 30 -6 -4 -2 0 2 4 6 8 10 Time, min

100

80 (b)

60 Region 40 I Region II

20 Residual weight % weight Residual

0 -5 0 5 10 15 Time, min

Figure 3-4: Kinetics of thermal degradation of the carbohydrate fractions of (a) Chlorella sp. and (b) Scenesdesmus sp.. The non-isothermal zone shows weight loss during temperature ramp from 250 to 320 °C. The isothermal zone shows weight loss during incubation at 320 °C.

100 450 400

80 C) 350 (a) 300 60 250 40 200 150

20 100 Temperature ( Temperature

Residual weight% 50 0 0 0 10 20 30 40 50 Time (min)

100 450 400

80 C) 350 (b) 300 60 250 40 200 150

20 100 Temperature ( Temperature

Residual weight% 50 0 0 0 10 20 30 40 50 Time (min)

100 450 400

80 C) 350 (c) 300 60 250 40 200 150

20 100 Temperature ( Temperature

Residual weight% 50 0 0 0 10 20 30 40 50 Time (min)

Figure 3-5: Kinetics of thermal degradation of the lipid fractions of (a) tripalmitate, (b) tristearate and (c) triolein. The non-isothermal zone shows weight loss during temperature ramp from room temperature to 420 °C. The isothermal zone shows weight loss during incubation at 420 °C. 1.12.3 Bench-scale micro-pyrolysis experiments:

To assess the degradation products formed during pyrolytic fractionation of algal biomass within the temperature intervals predicted during TG studies, further experiments were performed on a PyroprobeTM micro-pyrolysis system. Products

48 obtained were analyzed using GC-MS and the product identities were used to verify the source biopolymer. While py-GC-MS (pyrolysis probe coupled with GC-MS) rapidly provides reliable information on the class of compounds produced, it is not a convenient tool for quantification [127, 128]. However, the py-GC-MS is commonly practiced as a screening method to determine optimal operating conditions for pyrolysis and preliminary data obtained through py-GC-MS can be corroborated through larger-scale fixed/fluidized bed experiments [129].

The first step of pyrolytic fractionation was implemented by heating and incubating the samples at 320 °C to degrade the protein and carbohydrate fractions of algal biomass and the GC chromatograms for the resulting products are shown in Figure

3-6 (confidence levels of identified products are shown in Table 3.1 and Table 3.2). At this temperature, pyrolysis products obtained from both Chlorella sp. and Scenedesmus sp. contained oxygenated compounds such as acetic acid, furfural, 2-furanmethanol, 2-

(5H)-furanone, 2-hydroxy-2-cyclopenten-1-one, cyclopropyl carbinol, glucopyranose, levoglucosans and maltols, which are typically form from volatilization of carbohydrate

[43]. N-compounds were also present in pyrolysis products from both algae species.

However, N-products were more diverse from Chlorella sp. pyrolysis (such as amino-2- oxazolidinone, 3-butoxy propanenitrile, 1-amino-2,6-dimethylpiperidine and indole)

(Figure 3-6a) than Scenedesmus sp. (only 3-amino-2-oxazolidinone was identified)

(Figure 3-6b), possibly due to the lower protein content of Scenedesmus sp. It is also possible that some N-compounds (e.g. NH3 and NOx) might not have adsorbed on the

Tenax® trap of PyroprobeTM.

The products associated with protein and carbohydrate degradation products are extensively described in literature [7, 90, 130-132], relatively little is known about products of lipid pyrolysis. During the pyrolysis step at 320 °C intended for protein and carbohydrate pyrolysis, some C14-C18 hydrocarbons and fatty acids (e.g. octadecanoic acid) were also identified among the products. While cellular free fatty acids could have volatilized at this temperature (normal boiling points of C14-C18 fatty acids are in the range 250-300 °C [133]; see Figure 3-7), limited breakdown of other lipids including glycerides or phospholipids could also have occurred. Indeed, when di- and triglycerides were pyrolyzed at 320 °C, small amounts of fatty acids and fatty acid anhydrides were observed (Figure 3-8 and Figure 3-9). Diglycerides underwent more thermal degradation than triglycerides, as is evident by the significantly higher concentrations of products obtained from diglyceride pyrolysis at 320 °C (note the order of magnitude difference in the y-axis scales between Figure 3-8 and Figure 3-9) suggesting that smaller molecular weight lipids (possibly including monoglycerides and other short-chain lipids) are more thermally labile than triglycerides.

Most triglyceride pyrolysis however occurred at much higher temperatures (>380

°C, Figure 3-1and Figure 3-2). When pyrolyzed at 420 °C, soy oil produced large quantities of C16 and C18 fatty acids as well as hydrocarbons such as tetradecane, pentadecane, 8-heptadecene, heptadecane, 9,12-octadecadien-1-ol (Figure 3-10) confirming the extensive cleavage of the glyceride ester bonds as well as partial decarboxylation during the thermal degradation process [33].

Accordingly, in the second step of pyrolytic fractionation, solid residues from the algae samples left behind from previous step (at 320 °C) were heated at 420 °C. Similar

50 to soy oil pyrolysis products, the primary products for all the biomass samples were observed to be C16-C18 fatty acids and alkanes such as tetradecane, pentadecane, 8- heptadecene, heptadecane (Figure 3-11, Table 3.3 and Table 3.4). Only a small C16 fatty nitrile peak was seen in the Chlorella sp. chromatogram but no N-compounds were discernible in the products from Scenedesmus sp. samples pyrolyzed at 420 °C. It is possible that N-derivatives of fatty acids were formed as a result of reactions of the free fatty acids (produced in this step due to breakdown of triglycerides) with polymerized

(charred) proteins from the previous pyrolysis steps [134]. Since, fatty nitriles have been considered as fuel additives to improve lubricating properties [130, 134-137], the presence of these compounds at low concentrations may, in fact, enhance the fuel value of triglyceride-derived bio-oil from pyrolytic fractionation. Trace amounts of toluene and

4-methyl-phenol were also observed in triglyceride-based bio-oils from both the oleaginous algal feedstocks and could have formed from degradation of recalcitrant or charred proteins.

3.50E+09 Fatty acids 3.00E+09 3-buten-2-ol Propanenitrile, 3-butoxyl 2.50E+09 Furfural (a) 2.00E+09 2-furanmethanol 1-amino-2,6- 1.50E+09 dimethylpiperidine

Abundance 1.00E+09 5.00E+08 0.00E+00 0 5 10 15 20 25 30 35 40 Time, min

3.00E+09 Fatty acids Acetic acid 2.50E+09 3-amino-2-oxazolidinone 2-cyclopenten-1-one, 2- 2.00E+09 hydroxy Cyclopropyl (b) 1.50E+09 carbinol

1.00E+09 Abundance 5.00E+08

0.00E+00 0 5 10 15 20 25 30 35 40 Time, min

Figure 3-6: GC-MS chromatogram of protein as well as carbohydrate derived bio-oils from (a) Chlorella sp. and (b) Scenedesmus sp... The chemical compounds were identified using the NIST2008 library. Refer to table 3.1 and table 3.2 for full list of chemical compounds identified in this chromatogram.

Table 3.1: Confidence value of chemical compounds present in protein as well as carbohydrate-based bio-oils of Chlorella sp. (Refer to figure 3-6).

Retention Compound Confidence time 1.96 3-buten-2-ol 625 2.29 2-Propanone, 1-hydroxy 864 3.75 3-amino-2-oxazolidinone 695 4.00 Propanenitrile, 3-butoxy 687 4.83 Furfural 799 5.52 2-furanmethanol 879 7.77 2(5H)-furanone 714 8.00 2-furancarboxaldehyde, 5-methyl 803 8.34 2-cyclopenten-1-one, 2-hydroxy 893 9.61 2,3-pentanedione 660 9.70 2-Butanone, 1-(acetyloxy) 808 10.18 4(1H)-Pyrimidinone, 6-methyl 646 11.24 1,2-Cyclopentanedione, 3-methyl 796 11.71 1-amino-2,6-dimethylpiperidine 648 12.57 2,5-dimethyl-4-hydroxy-3(2H)- 770 furanone 12.80 Cyclopropyl carbinol 752 13.10 Maltol 716 14.21 1,4-dioxane-2,5-dione,3,6-dimethyl 763 14.58 Metyl isobutyl ketone 668 14.68 1,4:3,6-dianhydro-alpha-d- 661 glucopyranose 15.17 3,4-anhydro-d-galactosan 651 16.19 Indole 714 19.05 Beta-D-glucopyranose 824 19.74 Dodecanoic acid 639 21.4 Heptadecane 735 22.04 Tetradecanoic acid 628 23.34 3,7,11,15-tetramethyl-2-hexadecen-1- 714 ol 24.23 Hexadecanoic acid 904 25.83 9,12-octadecadienoic acid 797 25.91 9-octadecenoic acid 793 26.10 Octadecanoic acid 790

Table 3.2: Confidence value of chemical compounds present in protein as well as carbohydrate-based bio-oils of Scenedesmus sp. (Refer to figure 3-6).

Retention Compound Confidence time 1.93 Acetic acid 617

2.34 2-propanone, 1-hydroxy 865

3.40 Acetic acid anhydride 630

3.73 3-Amino-2-oxazolidinone 716

5.53 2-furanmethanol 901

7.87 2(5H) Furanone 772

8.41 2-Cyclopenten-1-one, 2-hydroxy 830

12.48 3-furancarboxylic acid, methyl ester 820

12.82 Cyclopropyl carbinol 771

13.97 1,4-dioxaspiro[2.4]hepta-5-one, 7-methyl 666

14.10 1,3-dioxolane-4-methanol 623

14.25 Undecanoic acid 644

14.96 1,4:3,6-dianhydro-alpha-d-glucopyranose 735

15.15 2-furancarboxaldehyde, 5-(hydroxymethyl) 712

15.78 Nonanoic acid 673

19.02 Beta-d-glucopyranose, 1,6-anhydro 739

19.73 Dodecanoic acid 634

21.41 Heptadecane 822

22.90 3,7,11,15-tetramethyl-2-hexadecen-1-ol 661

24.25 Hexadecanoic acid 842

25.91 Oleic acid 818

26.11 Octadecanoic acid 793

30.48 1-docosene 689

100 2.5

80 2

) )

1 -

60 1.5 C

(a)

40 1 %

wt (

20 0.5

Residual weight % weight Residual Derivative weight lossweight Derivative 0 0 0 100 200 300 400 500 600 Temperature ( C) 100 2.5

80 2 )

(b) 1 -

60 1.5 C

40 1 %

wt (

20 0.5

Residual weight % weight Residual Derivative weight lossweight Derivative 0 0 0 100 200 300 400 500 600 Temperature ( C)

Figure 3-7: Derivative weight loss (solid) and residual weight (dashed) curves obtained during thermal degradation of (a) Myristic acid (C14) and (b) Stearic acid (C18).

9-octadecenoic acid 8

7 ) 9 6 Oleic anhydride 5 4 3

2 Abundance (10 Abundance 1 0 0 10 20 30 40 Retention time (min)

Figure 3-8: GC-MS chromatogram of products from pyrolysis of 1,3-diolein at 320 °C.

7 C 18 fatty acids

) 8 5 C 16 fatty acids 4 8- Heptadecene 3 2

Abundance (10 Abundance 1 0 0 10 20 30 40 Retention time (min)

Figure 3-9: GC-MS chromatogram of products from pyrolysis of soy oil at 320 °C.

C 18 fatty acids

2 ) 10 1.5

1 C 16 fatty C 18 fatty acids 0.5 2-Propenal acids

derivatives Abundance(10 0 0 10 20 30 40 Retention time (min)

Figure 3-10: GC-MS chromatogram of products from pyrolysis of soy oil at 420 °C.

Fatty acids 7.00E+09 6.00E+09 5.00E+09 (a) 4.00E+09 3.00E+09 Hydrocarbons

Abundance 2.00E+09 Toluene Phenol, 3-methyl 1.00E+09 0.00E+00 0 5 10 15 20 25 30 35 40 Time, min

4.50E+09 Fatty acids 4.00E+09 3.50E+09 3.00E+09 2.50E+09 (b) 2.00E+09

1.50E+09 Hydrocarbons Abundance 1.00E+09 Toluene 5.00E+08 0.00E+00 0 10 20 30 40 Time, min

Figure 3-11: GC-MS chromatogram of triglyceride derived bio-oils from (a) Chlorella sp. and (b) Scenedesmus sp.. The chemical compounds were identified using the NIST2008 library. Refer to table 3.3 and table 3.4 for full list of chemical compounds identified in this chromatogram

Table 3.3: Confidence value of chemical compounds present in triglyceride- based bio-oils of Chlorella sp. (Refer to figure 3-7).

Retention Compound Confidence time 1.92 1-Hexene 750 2.48 2,3-dimethyl pentane 761 3.40 Toluene 854 3.80 2-octene 780 4.00 Octane 863 7.50 Nonane 786 10.50 Hexanoic acid 714 10.60 Decane 659 12.34 Phenol, 3-methyl 749 12.68 Heptanoic acid 742 12.90 Undecane 833 13.0 2-Undecene 716 14.00 Benzene, pentyl 658 14.40 Octanoic acid 829 15.90 Nonanoic acid 779 16.15 1-Tridecene 755 16.30 Tridecane 768 17.26 Decanoic acid 725 17.58 1-Tetradecanol 740 17.70 Tetradecane 813 19.0 Pentadecane 881 19.74 Dodecanoic acid 639 21.16 8-Heptadecene 823 21.4 Heptadecane 735 22.04 Tetradecanoic acid 712 23.58 Hexadecanenitrile 705 24.43 Hexadecanoic acid 886 24.89 Heptadecanoic acid 844 25.31 Octadecanoic acid,2-propenyl ester 714 26.10 9-octadecenoic acid 887 26.26 Octadecanoic acid 856 26.90 Erucic acid 679

Table 3.4: Confidence value of chemical compounds present in triglyceride- based bio-oils of Scenedesmus sp. (Refer to figure 3-9).

Ret. Time Compound confidence

2.49 3-methyl Hexane 757 3.40 Toluene 784 19.0 Pentadecane 813 19.74 Dodecanoic acid 639 21.16 8-Heptadecene 778 21.4 Heptadecane 663 22.04 Tetradecanoic acid 712 23.98 9-hexadecenoic acid 792 24.26 Hexadecanoic acid 881 25.31 Octadecanoic acid,2-propenyl ester 632 26.00 9-octadecenoic acid 889 26.16 Octadecanoic acid 840 26.90 Erucic acid 679

1.12.4 Lab-scale pyrolysis experiments:

Uncondensed vapors

Pyrolysis step 1 Step 1 Feed 1a 1d (320 °C) residue

Step 1 biooil

Stream number Summative Mass 1a 1b 1d Mass closure Total mass (g) 9.33 0.82 6.06 74% Lipid and lipid 2.51† 0.18‡ 1.94† 84% derivatives (g) (TAG* = 2.14) Nitrogen (g) 0.22 0.02 0.17 86%

Figure 3-12: Total- and component- mass balance data obtained from bench- scale fixed-bed experiments for Step 1 pyrolysis of oleaginous Chlorella sp. at 320 °C. †indicates lipid is quantified as FAMEs; ‡ indicates lipid is sum of hydrocarbons, free fatty acids, fatty amides and fatty nitrile; * indicates lipid is triglycerides. To validate the scalability of observations made from micropyrolysis experiments for production of triglyceride-specific bio-oils, larger scale fixed bed pyrolytic fractionation tests were performed using oleaginous Chlorella sp. These experiments were performed under “fast pyrolysis” mode with a vapor residence time approximately

2s [45, 122]. Biomass was first pyrolyzed at 320°C (Step 1) using 9.33g of biomass with a total lipid content of 27% (g-lipid/g-biomass) (measured as total fatty acid methyl ester

(FAME)[102]). The majority of the lipids were measured to be triglycerides (TAG) (23%

(g-TAG/g-biomass) and the remaining small fraction (4% (g-lipid/g-biomass)) was likely composed of membrane lipids and cellular free fatty acids. Mass balance data for Step 1 pyrolysis (at 320 °C) is shown in Figure 3-12. As expected, the bio-oils produced in this

60 step contained several protein and carbohydrate derived products including levoglucosans and N-compounds (see Figure 3-6 and Table 3.5). In addition, a small amount of fatty acids were also produced, likely from the more thermally labile cellular lipids. However, the vast majority of lipids, likely triglycerides, were not pyrolyzed during Step 1 and remained associated with the residue from this step (see lipid balance data in Figure 3-12 quantified using GC-FID analysis of bio-oil samples and solid residues). Nearly 25% of the biomass N was also removed in Step 1 (see nitrogen balance data in Figure 3-12).

Levoglucosanes C16 fatty acid 1.20E+11

1.00E+11 C18 fatty acids 1-amino- 8.00E+10 2,6- dimethyl- 6.00E+10 piperidine

4.00E+10 Abundance 2.00E+10

0.00E+00 0 5 10 15 20 25 30 35 40 Time, min

Figure 3-13: GC-MS chromatogram of bio-oil collected from Step 1 of fixed- bed pyrolytic fractionation performed on oleaginous Chlorella sp. at 320 C. 2.6 mg bio-oil was dissolved in 1mL of methanol for analyses. Confidence levels of products identified in the GC-MS chromatogram are shown in table 3.5.

Table 3.5: Confidence values of chemical compounds present in bio-oil collected during Step 1 (320 °C) of bench-scale fixed-bed experiments with oleaginous Chlorella sp. at 320 °C.

Retention Compound Confidence time 6.70 Methamphetamine, propiony 607 7.10 1,2-cyclopentanedione, 3-methyl 763 7.44 1-amino-2,6-dimethylpiperidine 625 8.01 Hexanoic acid, 2-propenyl ester 666 8.12 2,5-dimethyl-4-hydroxy- 731 3(2H)furanone 8.54 Levoglucosenone 849 9.63 Pentanoic acid, methyl ester 643 10.31 1,4:3,6-dianhydro-alpha-d- 773 glucopyranose 10.83 4-pyridinol 609 11.51 2-butenedioic acid, 2-methyl 661 12.38 Alpha-D-glucopyranoside, methyl 674 14.69 Beta-D-glucopyranose, 1,6-anhydro 731

Uncondensed vapors

Pyrolysis step 2 Step 2 feed 2a 2d Char (420 °C)

Step 2 biooil

Stream number Summative Mass 2a 2b 2d Mass closure Total mass (g) 5.16 1.63 2.04 71% Lipid and lipid 1.65† 1.55‡ 0.0 94% derivatives (g) Nitrogen (g) 0.14 0.03 0.09 86%

Figure 3-14: Total- and component- mass balance data obtained from bench- scale fixed-bed experiments for Step 2 pyrolysis of oleaginous Chlorella sp. at 420 °C. Step 2 was performed on residues from pyrolysis Step 1. †indicates lipid is

62 quantified as FAMEs; ‡ indicates lipid is sum of hydrocarbons, free fatty acids, fatty amides and fatty nitrile.

C16 fatty acid C18 fatty acids 1.40E+11 1.20E+11 C5-C14 fatty acids and C10-C17 1.00E+11 hydrocarbons 8.00E+10

6.00E+10 C18 fatty amides

Abundance 4.00E+10 2.00E+10 0.00E+00 0 5 10 15 20 25 30 35 40 Time, min

Figure 3-15: GC-MS chromatogram of bio-oil collected from Step 2 of fixed- bed pyrolytic fractionation performed on oleaginous Chlorella sp. at 420 C. 5.0 mg of bio-oil was dissolved in 1 mL of chloroform for analyses. Confidence levels of products identified in the GC-MS chromatogram are shown in table 3.6.

Palmitic acid C 18 fatty acids

Triglycerides Mono-glycerides Di-glycerides

Figure 3-16: GC-FID chromatogram of bio-oil collected from Step 2 of fixed- bed pyrolytic fractionation performed on oleaginous Chlorella sp. at 420 C.

Table 3.6: Confidence values of chemical compounds present in bio-oil collected during Step 2 of bench-scale fixed-bed experiments with oleaginous Chlorella sp. at 420 °C.

Retention Compound Confidence time 6.2 Pentanoic acid 704 7.6 Butyl-benzene 759 7.95 Heptanoic acid 682 8.15 1-decene 752 8.3 Undecane 859 8.40 4-undecene 835 8.53 5-undecene 744 8.76 Cyclodecene 704 9.3 Pentyl-benzene 696 9.52 Octanoic acid 802 9.80 3-dodecene 759 9.92 Dodecane 826 10.12 Undecane, 2,6-dimethyl 766 10.32 Cyclododecene 820 10.86 Hexyl benzene 690 10.97 Nonanoic acid 827 11.10 Octanoic acid, 2-propenyl ester 707 11.30 1-tridecene 840 11.41 Tridecane 818 12.37 Decanoic acid 758 12.70 1-tetradecene 869 12.81 Tetradecane 797 13.50 Cyclopentane, nonyl 806 13.65 Undecanoic acid 812 14.02 1-pentadecene 864 14.12 Pentadecane 666 14.84 Nonylcyclohexane 715 14.90 Pentadecane, 2-methyl 757 15.17 1-hexadecene 850 15.26 1-hexadecene 821 15.30 Hexadecane 871 16.18 6,9-heptadecadiene 765 16.27 8-heptadecene 723 16.52 Heptadecane 680 16.62 Benzene, 1-methyldecyl 707 17.20 2-hexadecenoic acid 649 18.70 Hexadecanenitrile 814 19.40 Hexadecanoic acid 671 20.00 Heptadecanoic acid 788 20.52 Oleanitrile 749 21.11 Oleic acid 768 21.33 Octadecanoic acid 787 22.03 Oleic anhydrde 639 23.02 9-octadecenamide 798 23.30 Octadecanamide 605

Unknown, 5% Hydrocarbons, Fatty amides 6% and nitriles, 4%

Mono-, di- and tri-glyceride, 14% C16 fatty acid, 24%

C18 fatty acids, 47%

Figure 3-17: Mass fraction of bio-oil collected during Step 2 of bench-scale fixed-bed pyrolytic fractionation experiments with oleaginous Chlorella sp. at 420 °C. The residue from Step 1 was subjected to a second pyrolysis reaction at 420 °C

(Step 2). As expected, the second step at 420°C, targeted towards triglyceride pyrolysis, produced bio-oil largely composed of fatty acids and glycerides. In addition, small amounts of hydrocarbons and fatty amides were also produced (see GC chromatograms in Figure 3-15 and Figure 3-16and identified products in Table 3.6). Quantification of bio-oil constituents through correlation of GC-FID peak areas with corresponding calibration standards for all detected compounds shows that nearly 95% of the recovered mass could be attributed to lipid-derived products, most likely obtained from thermal degradation of triglycerides (see total mass and lipid-derivative mass data for stream 2b in Figure 3-14 and mass fraction of bio-oil components in Figure 3-17). The lipid mass balance data for step 2 (Figure 3-14) also shows that nearly 94% of the lipid mass fed to step 2 (steam 2a) was recovered in the bio-oil (stream 2b). Although some N was also present in Step 2 bio-oil, N-balance analysis across both Steps 1 and 2 indicates that

<15% of N present in biomass was recovered in liquid products in Step 2. In contrast,

65 other studies have reported that nearly 60% of biomass N is accumulated in bio-oil after single-step pyrolysis[138].

Table 3.7: Elemental analysis of bio-oil collected during Step 2 of bench-scale fixed-bed pyrolytic fractionation experiments with Chlorella sp. at 420 °C.

Triglyceride-specific biooils from oleaginous algae C (wt%) 74.56 H (wt%) 11.11 N (wt%) 1.87 O (wt%) 12.46 HHV 41 (MJ/kg)

The calorific value of triglyceride-specific bio-oil from step 2 was calculated

[139] to be approximately 41 MJ/kg (Table 3.7) and is similar to higher heating values of petro diesel (42 MJ/kg). Overall, lab-scale pyrolytic fractionation experiments demonstrated that triglyceride-specific bio-oils with low N-content and high calorific value can be produced from oleaginous biomass via pyrolytic fractionation. Also, since pyrolytic fractionation produces fatty acid vapors, it should be possible to synthesize fuels (e.g. biodiesel) and chemicals (oleo-chemicals such as fatty amides and fatty nitriles) through gas-phase reactions of fatty acids. A more detailed conceptual design that incorporates the attributes of pyrolytic fractionation is described below.

1.12.5 Conceptual process design of pyrolytic fractionation:

Hydro-deoxygenation/ Feed In Combustion

Total mass: 1.0 kg Total mass: 0.35 kg OR Lipid: 0.27 kg Lipid: 0.06 kg Bio-oil for furan N: 0.024 kg N: 0.006 kg recovery

310-330 C Vapor-phase Conversion (e.g. Esterification, Inert Total mass: 0.40 kg * gas Lipid: 0.21 kg Hydrotreatment of FFAs) N: 0.008 kg OR Total mass: 0.65 kg FFAs for Lipid: 0.21 kg oleochemicals N: 0.018 kg 400-430 C Inert gas*

Total mass: 0.25 kg Lipid: 0 kg N: 0.010 kg Bio-char for combustion * For control of vapor residence time through reactor(s), and process heat if needed

Figure 3-18: Conceptual process design of the pyrolytic fractionation system showing mass balances and potential product pathways to fuels and co-products. Figure 3-18 shows a pyrolytic fractionation process flow diagram with two reactors that are sequentially maintained at temperature of 320 and 420 °C to volatilize protein as well as carbohydrate, and triglycerides. Based on results from the TG and

PyroprobeTM experiments described above, the solids in each reactor would likely require a residence time of approximately 5 min to achieve near-complete pyrolysis. The vapor residence time could, however, be maintained to be much lower (a few seconds) to prevent secondary gas-phase degradation reactions [45, 122] by flow of additional inert gases (to sweep out the vapors) thus accomplishing independent control of solid and vapor residence times. Red Lion Bio-Energy, LLC. (Toledo, OH), has been operating a pilot-scale (20 ton/day) pyrolyzer of similar design [140] with independent control of

67 solids and vapor residence times for the past 2 years with good success. The commercially operational Pyrocycling® process is also based on a similar reactor design

[45, 122].

This process design offers several advantages relative to fluidized bed reactors.

Low (or no) flow of inert gas would be needed in this system since fluidization of solids is not needed. Also, since the pyrolysis steps in our process are carried out at much lower temperatures (<450 °C) than more traditional single-step pyrolysis (>550 °C), secondary gas-phase reactions are likely to be much slower. As a result, maintaining very short vapor residence times (<1s, typical of fast pyrolysis) by using a high flow rates of inert gas might not be necessary to achieve high yields of bio-oil. In fact, when reactors are operated at high solid, the pressure of the generated vapors might itself be sufficient to impart gas velocities required to prevent secondary reactions. Thus, the costs associated with heating the carrier gas and subsequent energy recovery could be minimized or eliminated. Additionally, if solids are not fluidized, cyclones would not be needed for solids recovery and vapor product condensation would also be more efficient since concentrations of condensable vapors would be higher in the absence of non-condensable inert gases. Char blow-out and subsequent contamination of bio-oils would also be avoided [45].

In the process scheme shown in Figure 3-18, bio-oils from the first reactor

(protein as well as carbohydrate pyrolysis) could be processed to recover high value N- compounds such as indole and pyrimidine and oxygenated compounds (furan compounds). The bio-oil vapors from the first reactor could be directly integrated with a hydro-treatment system to produce drop-in green gasoline [141]. Furans (and derivatives)

68 could also be recovered as higher value products. The triglyceride-rich residue would be pyrolyzed in a second reactor to produce and recover fatty acids. Alternately, the fatty acid vapors could be directly converted to fatty acid methyl esters (FAMEs) or alkanes via vapor-phase reactions. Finally, char could be combusted for process heat or used as soil amendment or to recover nutrients for re-use in algal cultivation [49].

1.12.6 Estimates of energy required for pyrolytic fractionation:

4.5 450 300

4 400

)

C) -

3.5 350 3 300 200 2.5 250 2 200 1.5 150 100

1 100

(mW), (mW), Endoup Energy (kJ Energy(kJ kg

0.5 50 ( Temperature Differential heat flowDifferential 0 0 0 0 5 10 15 20 0 5 10 15 20 Time (min) Time (min)

(a) (b)

Figure 3-19: (a) Differential heat flux and (b) cumulative heat flux data for Chlorella sp. To estimate energy requirements, normalized differential heat flux data were obtained from DSC measurements during simulation of pyrolytic fractionation. Data was collected over the temperature range of 150-420 °C to discount the enthalpy of vaporization of residual moisture associated with the biomass samples. Similar to the derivative weight loss peaks observed in TG profiles (Figure 3-1), differential heat flux peaks, likely corresponding to thermal degradation of constituent biopolymers [142], was also observed for algal sample (Figure 3-19a). Numerical integration of the derivative weight loss curves was performed to obtain cumulative heat input profiles that are shown in Figure 3-19b. The monotonic increase in cumulative energy supply indicates that

69 pyrolytic fractionation is overall endothermic and includes the energy supplied for biopolymer decomposition as well as sensible heat to the biomass samples. Sensible heat to sweep-gas and sample holder is discounted through subtraction of the heat input to the empty reference pan.

The energy consumed in the pyrolytic fractionation reactors, starting with dry biomass at ambient temperature (20 °C), can be calculated by summing the cumulative heat input values from Figure 3-19b with the sensible heat associated with raising the biomass temperature from 20 °C to 150 °C. Using an average biomass heat capacity value of 1.7 kJ kg-1 °C-1 [143], this sensible heat amounts to 221 kJ kg-1. When added to the cumulative heat input values at 420 °C from Figure 8b, the net energy requirement for pyrolytic fractionation of dry oleaginous biomass samples tested here is estimated to be in the range of 600-700 kJ/kg and is consistent with previously reported values for biomass pyrolysis[97, 144]. Also, energy required for heating the algal biomass from

-1 ambient temperature to 420 °C is 680 kJ/kg ( = 푚푏𝑖표푚푎푠푠푐푝,푏𝑖표푚푎푠푠∆푇 = 1 kg×1.7 kJ kg

°C-1×400 °C). This implies that energy for biomass pyrolysis is equivalent to sensible heat of the algae from ambient temperature to 420 °C.

One concern with algae pyrolysis is the energy requirement for drying.

Hydrothermal liquefaction, where wet biomass is thermally processed, has been proposed in the literature as an energy-efficient thermochemical alternative [98, 108-111]. To compare the energy inputs, we calculated the energy requirements for both processes starting from a wet paste with a moisture content of 80 wt %. These estimates are shown in Table 3.8 and detailed calculations are shown below.

Basis: 1 kg algae + 4 kg water (=wet slurry with moisture content of 80% (w/w))

1. Energy for slurry drying:

= (sensible heat to water from 20 °C to 100 °C) + (energy for vaporization of

water at 100 °C)

= 푚푤푎푡푒푟푐푝,푤푎푡푒푟∆푇 + 푚푤푎푡푒푟∆퐻푣,푤푎푡푒푟

= (4 kg×4.2 kJ kg-1 °C-1×80 °C) + (4 kg×2260 kJ kg-1)

= 10384 kJ

2. Energy consumed in the pyrolysis reactor:

= (sensible heat to raise biomass temperature from 20 °C to 150 °C) +

(sensible heat to raise biomass temperature from 150 °C to 420 °C) +

(enthalpy of reaction of biomass during pyrolytic fractionation)

푚푏𝑖표푚푎푠푠푐푝,푏𝑖표푚푎푠푠∆푇 + (푐푢푚푢푙푎푡𝑖푣푒 푒푛푒푟푔푦 푓푙푢푥 푓푟표푚 퐹𝑖푔푢푟푒 8푏)

= (1 kg×1.7 kJ kg-1 °C-1×130 °C) + (400 to 480 kJ)

= 620 to 700 kJ [average value of 660 kJ]

3. Energy recovered by combustion of biochar:

= (푚푏𝑖표푐ℎ푎푟 − 푎푠ℎ 푐표푛푡푒푛푡)∆퐻푐표푚푏,푏𝑖표푐ℎ푎푟

=(0.20 kg – 0.05kg)×25 MJ kg-1 [mass of biochar is based on average weight

of residue observed from Figure 3-1; ash content is estimated to be 5%

(w/w) of algae feed; enthalpy of combustion values are based on previous

estimates from[113]]

= 3750 kJ

4. Energy consumed in the hydropyrolsis reactor:

= (sensible heat of biomass from 20 °C to 350 °C) + (sensible heat of water

from 20 °C to 350 °C) [assuming reaction carried out at 350 °C; heat of

reaction is not included since the values are not known]

= 푚푏𝑖표푚푎푠푠푐푝,푏𝑖표푚푎푠푠∆푇 + 푚푤푎푡푒푟(퐻푤푎푡푒푟,350 °퐶 − 퐻푤푎푡푒푟,20 °퐶)

= (1 kg×1.7 kJ kg-1 °C-1×330 °C) + 4 kg×(1671.8 kJ kg-1 – 83.9 kJ kg-1)

[enthalpy values for saturated liquid are from steam tables; specific heat

capacity of biomass is assumed constant over the temperature range due

to non-availability of temperature dependent correlation]

=6913 kJ

5. Energy consumed for solvent recovery in hydrothermal liquefaction:

= energy for vaporization of solvent

= 휌푑𝑖푐ℎ푙표푟표푚푒푡ℎ푎푛푒푉푑𝑖푐ℎ푙표푟표푚푒푡ℎ푎푛푒∆퐻푣,푑𝑖푐ℎ푙표푟표푚푒푡ℎ푎푛푒 [dichloromethane is

assumed as solvent]

= (1.33 kg L-1×7.5 L×341.2 kJ kg-1) [dichloromethane is assumed as solvent;

1.5L of solvent is assumed to be required per L of slurry; properties of

dichloromethane are from the NIST chemistry webbook]

= 3412 kJ

6. Energy recovered through heat exchange with hydrothermal liquefaction effluent

= 50 % of energy input into the reactor [from estimates assumed by [25, 145,

146]]

= 0.5×6913 kJ

=3457 kJ

Table 3.8: Comparison of energy requirements for pyrolytic fractionation and hydrothermal liquefaction.

Basis: 1 kg dry algae mixed with 4 kg of water (i.e. 20% (w/w) slurry) initially at 20°C and atmospheric pressure Thermo-chemical processes → Pyrolytic fractionation Hydrothermal liquefaction Energy components ↓ (KJ/kg dry-algae) (KJ/kg dry-algae)a

Drying 10,400 n/a Energy consumed in reactor 660b 6913c Solvent recovery n/a 2269d Total Energy consumption 11,060 9,182

Energy that can be recovered -5000e 3457f

Net energy consumed 6,060 5,725 aslurry reaction carried out at 350°C and 165bar (sat. steam pressure), Brown et al., 2010. bthis value includes sensible heat of biomass (from 20 to 420 °C) and heat of reaction. cthis value includes sensible heat for water (from standard steam tables) and biomass (1500 J/kg-K). Does not include heat of reaction due to unavailability of literature values. dAfter reaction, products are assumed to be extracted from the slurry using dichloromethane (Brown et al.). This value represents the enthalpy of vaporization (341.2 kJ/kg) of the solvent to recover biooil and facilitate solvent re-use. 1L of solvent was assumed to be used per kg of slurry. e0.20 kg ash-free-char/kg-biomass was assumed with a calorific value of 25 MJ/kg. f 50% of energy input into the hydrothermal liquefaction reactor was assumed to be recoverable based on estimates from Du et al., 2012 and Minowa et al., 1998.

The basis for these calculations was 1 kg dry algae (equivalent to a wet paste mass of 5 kg). To produce bio-oils from this feed using pyrolytic fractionation, energy inputs would be needed to – (i) evaporate 4 kg of water, and (ii) perform pyrolytic fractionation of the dried material. Based on enthalpy of vaporization of water and the

DSC measurements (described in the previous paragraph), the net energy requirement for pyrolytic fractionation was estimated to be approximately 11,060 kJ (kg-dry-algae)-1.

However, part of the energy needs could be offset by recovering energy through combustion of the bio-char produced during pyrolytic fractionation. Assuming that (i) the

73 calorific value of the bio-char is 25 MJ/kg (on an ash-free basis) [113], (ii) the biomass has a 5% ash content (w/w – dry basis) and (iii) one kg of dry-algae produces 0.20 kg of bio-char (final residual weight observed during TG analysis of Chlorella sp. and

Scenedesmus sp., Figure 3-1), we calculated that approximately 35% of total energy required for pyrolytic fractionation (including drying) could be recovered from combustion of bio-char produced during the process (Table 3.8).

For hydrothermal liquefaction, energy would be required for heating the algal slurry from ambient conditions to the high temperature and pressure in the reactor and for supplying enthalpy of the reaction. Savage and co-workers [108, 110] have reported that best hydrothermal liquefaction yields are obtained at a slurry temperature of 350 °C.

Since the slurry consists of a mixture of solid algal biomass and water, both components would have to undergo transition to the operating state in the reactor. Assuming a thermodynamically closed system, the energy input associated with the state change of water was estimated from saturated steam tables. Sensible heat transferred to the biomass

-1 -1 was estimated assuming a constant biomass heat capacity (cp,biomass = 1.7 kJ kg °C

[143]) over the entire temperature range (due to unavailability of literature values for its temperature dependence). This calculation further assumed that biomass breakdown was not significant during the heating period – a reasonable possibility given the long reaction times (10-60 min) reported for hydrothermal liquefaction [108-110]. Since enthalpy of reaction values for hydrothermal liquefaction of algae are, to our knowledge, not available in the literature, this energy component (endothermic) was not included in our calculations.

After reaction, organic solvents are likely to be used to extract fuel components

(bio-oils) from the wet hydrothermal liquefaction product [108, 110]. After extraction, the solvents are further expected to be recovered through evaporative (distillation) methods to facilitate solvent re-use. Therefore, energy required for recovery of organic solvents should also be considered in energy estimates for the hydrothermal liquefaction process. In previous studies, 15mL of dichloromethane was used as the extraction solvent per 5mL of hydrothermally liquefied algal slurry [108]. Assuming that at least 1.5L

(=2kg) of dichloromethane would be required per L of hydrothermally treated slurry in a process-scale system, the amount of solvent needed would equal nearly 10 kg per kg of dry algal biomass. Since the enthalpy of vaporization of dichloromethane is 341.2 kJ kg-1, the energy required to recover dichloromethane was estimated to equal 3,412 kJ per kg of dry algae (= m×∆Hvap = 10×341.2). Therefore, the total energy required for processing 5 kg of wet algal biomass with 80 wt% water content (equal to processing 1 kg dry algae) via hydrothermal liquefaction is estimated to be approximately 10,325 kJ (Table 3.8).

This value includes the sensible heat to biomass and water as well as enthalpy of vaporization associated with the extraction solvent.

Although wet solid residue left behind after hydrothermal liquefaction of algal slurry could also be combusted, it would incur an energy penalty for drying or evaporation of water/moisture and was therefore not considered as an option for energy recovery. However, energy could be recovered through heat exchange with the hot hydrothermally treated slurries as they are cooled down after reaction. While direct heat exchange between hot and cold streams would be most desirable [98, 110], such heat recovery through conventional designs (e.g. shell and tube heat exchangers) is likely

75 infeasible for slurries with high solids loadings due to unpredictable flows, low heat transfer coefficients and propensity for fouling, especially on the shell side. Instead, a heat carrier such as steam could be used to facilitate transfer of heat between the hot and cold slurries in an efficient manner, given the low cost and favorable heat transfer properties of the water/steam system [147]. Simulation of such a heat recovery system

(performed using the process simulator ChemcadTM and section 3.3.7) suggest that, at best, approximately 50% of the energy supplied for hydrothermal liquefaction might be recovered. Other previous studies have also assumed a value of 50% for energy recovery during hydrothermal liquefaction [145, 146] and this estimate is also included in Table

3.8.

The calculations in Table 3.8 suggest that the net energy requirements for both of the thermochemical processes compared are somewhat similar with energy requirements for hydrothermal liquefaction calculated to be approximately 6% lower than the energy needs for pyrolytic fractionation. However, appropriate caution must be employed while interpreting these results since the calculations were made using several assumptions for processes that do not currently exist at commercial scales. In addition, other aspects of the process not considered in these calculations must also be taken into account. For example, in the case of hydrothermal liquefaction, the recovered bio-oils would have to undergo re-heating as they are upgraded (through processes such as hydrotreatment) and distilled into appropriate fuel fractions. For pyrolytic fractionation, this additional energy penalty could be avoided by appropriate integration of downstream upgrading steps through gas phase conversion(s) of the pyrolysis vapors (Figure 3-18). The fate of the unconverted residues is also not clear in the case of hydrothermal liquefaction. While

76 energy recovery through combustion of the residues is possible, there would also be an energy penalty for removal of moisture (drying) from the wet residues. Capital costs for hydrothermal liquefaction are also expected to be higher due to the high pressure systems

(vapor pressure of water at 350 °C is approximately 165 bar) with residence times in the order of several minutes. Solvent use also adds additional unit operations as well as increases capital cost. While it might be possible to avoid solvent use through processing of more dilute slurries, this option would also result in increase in the capital and operating costs of the hydrothermal treatment reactor.

One final advantage of using a process based on dry material is that such a design allows for storage of the feedstock. Although algae can potentially be grown throughout the year, the biomass productivity can suffer from significant seasonal fluctuations due to changes in solar irradiation and temperatures. Since downstream operations are designed for fixed capacity, the equipment processing wet biomass may be under-utilized during the less productive seasons and might not be able to process all of the feedstock during peak productivity. These fluctuations can impact process economics more significantly in the case of capital intense conversion systems. For a dry conversion process, the excess feedstock produced during the warmer seasons could be stored and used in the cooler seasons to “average out” variations in the seasonal productivity. The energy for drying could be supplied (at least partially) from low grade heat sources (e.g. hot flue gas or low grade steam) from other industries if algae facilities are co-located appropriately. Solar drying might also be possible, especially in the warmer and more productive months.

1.12.7 ChemcadTM simulation of hydrothermal liquefaction process:

To determine the amount of the energy that can be recovered from effluents of hydrothermal liquefaction reactor, process with and without heat recovery system was simulated in ChemcadTM using the same feed.

Energy required for hydrothermal liquefaction without heat recovery system:

Figure 3-20: ChemcadTM flowchart simulating hydrothermal liquefaction process without heat recovery system. Figure 3-20 shows the ChemcadTM flowchart developed to simulate the hydrothermal liquefaction process without any heat recovery using 4 kg h-1 of water at 25 oC and 1 bar pressure as feed. Water was used as feed (instead of algal slurry) to simplify calculations and since the energy contributions of the biomass are much smaller relative to the water component of the slurry (see calculations in previous section). The pump was used to increase pressure from ambient (1 bar) conditions to the operating pressure (170 bar). A heat exchanger was used to simulate the reactor since the enthalpies of reaction are likely

78 much smaller relative to the energy requirements for heating the slurry to operating conditions (350 °C and 170 bar; see calculations in previous section). Also included in

Figure 3-20are the stream data (P, T, mass flow rate, vapor mass fraction and enthalpy) and the specifications of the heat exchanger used to simulate the reactor. The following energy inputs were computed from the simulation:

 Power required for pumping = 0.0969 MJ h-1 (assuming pump efficiency of 0.7)

 Heat duty of reactor = 6.11 MJ h-1.

Therefore, total energy required without heat recovery

= Power supplied to pump + Heat duty of reactor

= 0.1 MJ h-1 + 6.11 MJ h-1

= 6.2 MJ h-1

Energy required for hydrothermal liquefaction with heat recovery system:

Figure 3-21: ChemcadTM flowchart simulating hydrothermal liquefaction process with heat recovery system.

Figure 3-21 shows the ChemcadTM flowchart developed to simulate the hydrothermal liquefaction process without additional heat recovery using high pressure steam (240 °C,

35 bar). In this simulation, water was converted to high pressure steam in the heat recovery boiler (“Ht rec boil”) at the reactor effluent. This steam is then used to pre-heat the feed in the “Feed heater”. The pre-heater and the heat recovery boiler were simulated with a temperature difference of at least 30 °C between the hot and cold fluids at all locations in the heat exchangers.

Also included in Figure 3-21 are the stream data (P, T, mass flow rate, vapor mass fraction and enthalpy) and the specifications of the heat exchanger used to simulate the reactor. The following energy inputs were computed from the simulation:

 Power required for pumping = 0.1 MJ h-1 (assuming pump efficiency of 0.7)

 Heat duty of reactor = 3.3 MJ h-1.

Therefore, total energy required without heat recovery

= Power supplied to pump + Heat duty of reactor

= 0.1 MJ h-1 + 3.3 MJ h-1

= 3.4 MJ h-1

Since the feed used for both process simulations is same, energy recovered can be calculated as the difference in energy input to the reactor and pump in the “with” and

“without” scenarios:

Energy recovered = 6.2 - 3.4 = 2.8 MJ h-1

% energy that can be recovered from hydrothermal liquefaction process using heat recovery system = (2.8/6.2) × 100 = 45 %

1.13 Conclusions

In this chapter, we have demonstrated a pyrolytic fractionation process for producing high calorific triglyceride-specific bio-oil from oleaginous algal biomass.

Further, these products have little, if any, contamination of N-compounds and oxygenates from carbohydrate degradation. We have also developed a conceptual process design that would integrate product upgradation steps with the pyrolysis reactors to directly produce drop-in fuels. Our preliminary estimates also indicate that the net energy requirements are similar to those for alternate hydrothermal liquefaction processes and could be further lowered if appropriately integrated with other industries that produce waste low-grade heat.

Chapter 4

Triglyceride quantification of oleaginous algal biomass using thermo-gravimetry (TG)

1.14 Introduction

The working principle of pyrolytic fractionation (i.e. thermal degradation of triglyceride doesn’t overlap with that of other constituent biopolymers of oleaginous algal feedstocks - protein and carbohydrate) can be employed to determine the triglyceride content in oleaginous feedstocks. The objective of this chapter is to describe a TG-based analytical method for quantifying total triglycerides. This method does not require chemicals/organic solvents; instead this method involves only a single-step use of thermo-gravimetric analysis. The triglyceride content in sunflower seed and three microalgae (Chlorella sp., Scenedesmus sp. and Schizochytrium sp.) samples had been quantified using TG and results are compared with estimates obtained from conventional solvent extraction and in situ transesterification methods.

Microalgal feedstocks have the potential to sustainably produce liquid fuels and oleo-chemicals [29] without competing with food crops for resources such as agricultural land, irrigation water and fertilizers [24, 29, 72, 73, 99]. Cultivation of oleaginous

82 microalgae is preferable since triglycerides can be converted into drop-in biodiesel or hydrocarbons. Laboratory studies have shown that triglycerides can constitute up to 50 wt% of the cell mass of some microalgal species. [23-28]. During laboratory- or large- scale cultivation of oleaginous cultures, it is necessary to monitor triglyceride content of cultures to determine optimal harvesting periods. Rapid and accurate methods for triglyceride analysis are therefore essential.

Conventional methods employed for quantification of triglycerides in biological materials involve extraction into organic solvents followed by gravimetric or chromatography measurements [30, 31]. In the case of microalgae, additional cell- disruption (e.g. though sonication or by application of high pressure) is also often required [34-36]. Estimates obtained from some solvent extraction techniques for algal feedstocks have also been reported to be inconsistent and potentially inaccurate [102,

148]. In situ acid-catalyzed transesterification can be performed on algal or other oleaginous biomass and subsequent quantification of the fatty acid methyl esters

(FAMEs) obtained gives an estimate of the sum of the mono-, di-and tri-glycerides as well as other fatty acid components (e.g. free fatty acids and phospholipids) present in the samples [37, 38, 102].

Regardless of their accuracy, these conventional analytical methods for quantifying triglyceride content require chemicals or organic solvents (such as hexane, methanol, sulfuric acid) and involve multiple time- and labor- intense steps such as thermal and acid treatment, liquid-solid separation, liquid-liquid extraction and gravimetry or liquid analysis using chromatography. In manufacturing scale operations, triglyceride analysis by these methods would require a dedicated laboratory for analysis

83 of samples and appropriate quality control/quality assurance standards and well-trained technicians.

Thermo-gravimetry (TG) is a thermal analysis technique by which change in mass of volatile or thermally-degradable materials is measured as a result of change in temperature or time. In dynamic measurements, samples are heated over a temperature gradient, whereas in isothermal measurements, sample weight loss is measured over time at a constant temperature [149]. TG-based techniques are widely employed to perform proximate analyses of carbonaceous materials (such as coal, biomass, and fossil fuels)(ASTM D-7582-12) [84, 113, 150, 151] and to evaluate the thermal stability of materials (ASTM E2550-11)[7, 90, 152]. In addition, TG has also been used to determine the boiling point distribution of liquid fuels [65, 153] and to measure reaction kinetics for design of high-temperature processes [154-159]. TG also has the potential to couple with other analytical instruments such as Fourier transform infrared spectrometry (FTIR), mass spectrometry (MS), and gas chromatography (GC) for analysis of volatiles obtained from the sample [150, 160-162]. However, to our knowledge, use of TG to perform composition analysis of biomass has not been previously reported. In this article, a TG based analytical method has been described to estimate the triglyceride content of oleaginous biomass.

1.15 Methods and methods:

1.15.1 Materials:

Table 4.1: Moisture content of oleaginous algal feedstocks estimated using ASTM D7582-12 protocol.

Tristerarte, tripalmitate, triolein, hexane, methanol, tetrahydrofuran, chloroform, sulfuric acid were purchased from Sigma-Aldrich (St. Louis, MO). Chlorella sp. and

Schizochytrium sp. were cultivated in our lab using previously described methods.

Cultures were harvested by centrifugation, washed using distilled water and then freeze dried for use in this study. Freeze-dried Scenedesmus sp. was obtained from Arizona

State University (Phoenix, AZ). Sunflower seeds were obtained from Bassett’s health food store (Toledo, OH). The oil seeds were ground to -80 mesh size using a laboratory

Wiley mill (Model 4, Thomas Scientific, Swedesboro, NJ). Moisture content was estimated using ASTM D7582-12 [163] and is shown in Table 4.1.

1.15.2 Thermo-gravimetric Analysis (TGA):

TG analyses were performed on a TA Instruments SDT Q600 series analyzer

(Schaumburg, IL) that provides simultaneous measurement of weight change and differential heat flow on a single sample. Each weight loss measurement was performed using approximately 10.0 ± 1.0 mg of biomass sample that were loaded into an alumina crucible while a second identical crucible served as a reference. N2 was used as the

85 carrier gas and also to maintain inert atmosphere. The flow rate of N2 was kept at 100 mL min-1.

Overall thermal degradation behavior of biomass feedstocks was determined by heating the samples from room temperature to 600 °C at a constant ramp rate of 20 °C min-1.

To quantify triglyceride content, a two-stage heating protocol was used. Samples were first heated to 320 °C and kept isothermal for 15 min. Thereafter, the samples were re-heated to 420 °C and again maintained isothermal for 15 min. The non-isothermal heating rates were 20 °C min-1.

1.15.3 Solvent extraction:

Triglyceride extraction was carried out using an equal-volume mixture of hexane, chloroform and tetrahydrofuran. 50 mg of dry biomass sample was transferred to 15 mL glass tubes containing 5 mL of solvent mixture. The mixture was placed in an ice cold bath and sonicated for 3 min using Fisher Scientific Model 120 sonic dis-membrator

(Pittsburgh, PA) to break algal cell walls and extract the cellular triglycerides. The tube was then centrifuged to separate the extract and the biomass residue. The liquid extract was analyzed using gas chromatography (GC). Triolein was used as standard to quantify triglycerides. The solid residue was retained for subsequent TG analyses.

1.15.4 In situ transesterification method:

Fatty acid methyl esters (FAMEs) were synthesized by a one-step transesterification method [164, 165]. Dry biomass samples (30 mg) were weighed into

86 clean, 2 mL crimp-top GC vials to which 1 mL of a solution of 5 % sulfuric acid in methanol (v/v) was added. These vials were sealed with teflon-lined caps and placed in a heating block set at 90 °C for 2 h. The vials were removed every 10 min and vortexed to ensure adequate mixing. After 2 h, the vials were cooled to room temperature and then centrifuged to separate the biomass residue. The liquid phase was transferred to 15 mL serum bottles and 5 mL of hexane was added to perform liquid-liquid extraction for recovery of FAMEs from the acidified methanol solution into hexane. The serum bottles were crimp-sealed with Teflon-lined aluminum foil caps to avoid leakage and placed in a

90 °C water bath for 20 min. Thereafter, the serum bottles were removed and cooled to room temperature and an aliquot was recovered from the hexane phase and analyzed using GC. FAME standards purchased from Sigma-Aldrich were used to obtain calibration curves. [101, 102, 148].

1.15.5 Quantification of triglycerides and FAMEs using GC-FID:

A Shimadzu 2010 GC equipped with a Restek (Bellefonte, PA) Rtx Bio-diesel column – 15 m x 0.32 mm ID x 0.1 µm and flame ionization detector (FID) was used to quantify triglycerides and FAMEs. The GC oven was programmed to initially hold a temperature of 60 °C for 1 min followed by a steady increase to 370 °C at a ramp rate of

10 °C/min. Finally, the oven temperature was maintained at 370 °C for 6 min. Helium was used as a carrier gas with a constant linear velocity of 50 cm/s in the column. The injector and FID temperatures were both maintained at 370 °C.

1.15.6 Statistical analysis:

Single factor analysis of variance (ANOVA) was performed with confidence level of 95% in Microsoft Office ExcelTM to compare the estimates obtained from TG method to other conventional triglyceride quantification techniques.

1.16 Results and discussion:

1.16.1 Pyrolysis of triglycerides:

In the first set of experiments, the volatilization temperature intervals of triglycerides were determined using glyceryl trioleate (triolein), glyceryl tristearate

(tristearate) and glyceryl tripalmitate (tripalmitate) as model compounds.

Figure 4-1: (a) Residual weight (dashed lines) profiles and (b) derivative weight loss (solid lines) of triolein, tripalmitate and tristearate obtained during heating under N2 atmosphere to determine their volatilization temperatures. Figure 4-1 shows that triolein, tristearate and tripalmitate completely volatilizes in the narrow temperature range of 360-480 °C with a single derivative weight loss peak leaving behind small amount of char (i.e. < 2 wt % of initial weight triglyceride).

Previous studies also showed that vegetable oils which is a mixture of triglycerides and saturated long chain triglycerides volatilizes in the temperature interval of 370-480 °C

[125, 166]. Therefore, TG analysis of triglyceride model compounds clearly shows that triglycerides in general volatilize in the temperature interval of 370-480 °C.

Figure 4-2: Residual weight (solid lines) and temperature (dashed lines) of (a) tripalmitate, (b) tristearate and (c) triolein with time obtained during dynamic heating from room temperature to 420 °C and maintained isothermally at 420 °C for 30 min under N2 atmosphere to determine the time required for complete volatilization of substrates. Since the derivative weight loss (dw/dT ) peaks of thermograms reflect the kinetics of degradation for a constant heating rate (i.e. dT/dt= constant), the derivative weight loss profiles obtained for triolein, tristearate and tripalmitate indicate that the maximum rate of volatilization was in the temperature range of 420-430 °C (see Figure

4-1(b)). To determine time required for complete volatilization of triglycerides at 420 °C, a set of isothermal experiments were performed. Samples of model the triglyceride compounds were heated to 420 °C and maintained at this temperature for 30 min.

1.16.2 Pyrolysis of triglycerides at 420 °C:

TG profiles shown in Figure 4-2 were obtained from isothermal experiments at

420 °C. Tripalmitate (Figure 4-2(a)) and tristearate (Figure 4-2(b)) volatilize completely within 5 min of isothermal heating while 15 min was required for complete volatilization of triolein (Figure 4-2(c)). Therefore, these isothermal experiments showed that upto 15 min may be required to volatilize triglycerides completely at 420 °C.

Figure 4-3: Derivative weight loss (solid lines) and residual weight (dashed lines) profiles represents the pyrolysis of (a) Sunflower seeds, (b) Chlorella sp., (c) Scenedesmus sp., and (d) Schizochytrium sp.. Derivative weight loss peak in the temperature interval of 370-480 °C indicates volatilization of triglyceride fraction present in these feedstocks. 1.16.3 Pyrolysis of triglycerides-rich feedstocks:

To determine thermal degradation behavior of oleaginous feedstocks (here

Sunflower seeds, Chlorella sp., Scenedesmus sp., and Schizochytrium sp.), TG profiles

91 was obtained. Figure 4-3 shows the thermal degradation of these samples over temperature range of 150-480 °C. Three distinct derivative weight loss peaks were observed in these thermograms, including a peak observed in the temperature interval of

370-460 °C (Figure 4-3) most likely from volatilization of biomass triglycerides.

Previous TG studies of lipid-lean biomass feedstocks have not observed derivative weight loss peaks in this temperature region, suggesting that a peak at 420 C would result from triglyceride degradation alone, without significant mass loss contribution from other constituent biopolymers [79, 113]. Indeed other biopolymers in oleaginous algae, proteins and carbohydrates, have been shown to decompose below 350 °C [92, 113].

These results suggest that heating oleaginous samples were at low temperature (< 350 °C for sufficient period to allow selective degradation of constituent protein and carbohydrate fractions, would result in residues with triglycerides as the principal thermally labile constituents.

1.16.4 Loss of triglycerides at low temperatures:

A two-step thermal degradation protocol was implemented with the intention of thermally-degrade triglycerides alone, without interference with other biomass constituents. Biomass samples were first heated to 320 °C and maintained isothermal for

15 min to allow degradation of protein and carbohydrate. In the second step, sample temperatures were increased to 420 °C and again kept isothermal to volatilize the triglyceride fraction. During these experiments, a small, but significant, mass loss was also observed during the process of heating the sample from 322 °C to 420 °C (Figure

4-5). To determine if this (small) mass loss could be attributed to triglyceride

92 volatilization, a similar temperature program was applied to triglyceride model compounds. Weight loss profiles from these experiments are shown in Figure 4-4.

Figure 4-4: Residual weight (dashed lines) and temperature (solid lines) of (a) tripalmitate, (b) tristearate and (c) triolein with time obtained while employing temperature protocol to determine their triglyceride content.

Indeed, small mass loss was observed for triglyceride model compounds during isothermal heating at 320 °C for 15 min.

Figure 4-5: Residual weight (dashed lines) and temperature (solid lines) of (a) Sunflower seeds, (b) Chlorella sp., (c) Scenedesmus sp., and (d) Schizochytrium sp. with time obtained while employing temperature protocol to determine their triglyceride content. This loss could be due to triglyceride volatilization as a result of continuous purge with N2 for prolonged periods which would allow evaporation of the triglycerides similar to evaporation of water observed at room temperature when in contact with a constant flow of dry air.

It was determined (from data associated with Figure 4-4) that 92%, 95%, and

89% of tripalmitate, tristearate and triolein, respectively, were volatilized during heating from 322 °C to 420 °C followed by isothermal condition at 420 °C. These results show that, on an average, nearly 8% of the triglycerides could evaporate during the 15 min

94 incubation of samples at 320 °C. This mass loss correction factor would therefore have to be applied when estimating triglyceride content as mass loss between 320 – 420 °C. For our TG set-up, this correction factor was equal to 0.92, but could be different for other instrument configurations that used alternate carrier gases or other flow rates.

Above proposed temperature protocol was employed to quantify the triglyceride content in sunflower seeds, Chlorella sp, Scenedesmus sp., and Schizochytrium sp. and the TG profiles shown in Figure 4-5 were obtained. Based on known initial weight of samples, mass fraction during temperature ramp from 320 to 420 °C, isothermal incubation at 420 °C and the triglyceride mass correction factor, triglyceride contents of the samples were estimated and are shown in Table 4.2.

Table 4.2: Comparison of triglyceride content of oleaginous algae estimated via TG method and in situ transesterification method. Triglyceride content determined using TG method is statistically similar to estimates obtained from in situ transesterification method.

The calculations of this method are illustrated using Chlorella sp. as an example:

Initial mass of Chlorella sp. sample recorded by TG instrument = 10 mg.

Weight loss of Chlorella sp. sample in during dynamic heating from 321 °C to

420 °C and isothermal heating at 420 °C recorded by TG instrument = 3.324 mg.

According to TG-method described earlier, triglyceride content of Chlorella sp. sample = (3.324 mg / 10 mg) × [100/(100-2.4)] × (1/correction factor) = 37.84 wt% dry- basis.

Lipid content of these feedstocks was also determined using in situ transesterification method (see Table 4.2). Statistical (single factor ANOVA) analysis indicated that triglyceride content estimated using TG method and the in situ transesterification method were statistically similar.

During in situ transesterification method, in addition to triglycerides, protein and carbohydrates can also degrade upon exposure to heat and acid.

Table 4.3: Triglyceride content of oleaginous algal feedstocks estimated via TG method and solvent extraction method.

To selectively extract only triglycerides into organic solvents, solvent extraction method was also employed. This organic layer was analyzed using GC to estimate triglyceride content in these feedstocks.

Triglyceride content estimated for sunflower seeds using TG method is statistically similar to values obtained for solvent extraction method (see Table 4.3 and

ANOVA analysis). The solid residues left behind after solvent extraction of oleaginous

96 algal feedstocks were collected and dried for subsequent TG analysis. Absence of derivative weight loss peak in the temperature interval of 370-480 °C in TG profiles

(Figure 4-6(a)) of sunflower solid residue indicates complete extraction of triglycerides.

Figure 4-6: Derivative weight loss (solid lines) and residual weight (dashed lines) profiles of (a) Sunflower seeds, (b) Chlorella sp., (c) Scenedesmus sp., and (d) Schizochytrium sp. solid residues obtained after solvent extraction. It also indicates that the inverse of our hypothesis is also correct i.e. absence of derivative weight loss peak over temperature interval of 370-460 °C in the TG profiles implies the lack of triglycerides in the sample. Therefore, TG method can also be used as analytical tool to determine the performance of solvent extraction techniques to extract triglyceride from algal biomass. This can be measured by simply performing TG analysis of resulted solid residues of oleaginous biomass.

In contrast to sunflower seeds, triglyceride content estimated for algal feedstocks using solvent extraction method was significantly low when compared to the values

97 obtained from in situ transesterification procedure (see Table 4.3). This result may be due to incomplete extraction of triglycerides from algal cells into solvent mixture during solvent extraction procedure. Inefficiency of solvent extraction procedure was reported previously in the literature [102]. The solid residues left behind after solvent extraction of oleaginous algal feedstocks were collected and were subjected to TG analysis and showed derivative weight loss peaks in the temperature interval of 370-480 °C that corresponds to the triglyceride volatilization (Figure 4-6(b)(c)(d)) which indicates incomplete extraction of triglycerides from algal cells into solvent mixture. To demonstrate the accuracy of TG method, triglyceride mass balance was performed around solvent extraction procedure for algal feedstocks. This mass balance involves measurements of triglyceride content in (i) original algal sample, (ii) organic solvent extract and (iii) algal residue left behind after solvent extraction. Triglyceride content in the algal residue obtained after solvent extraction was estimated using TG method since the results obtained from TG method and in situ transesterification procedure are similar

(see Table 4.4). It shows that triglyceride content in algal feedstocks is similar to sum of triglyceride content in solvent extract and their solid residues. Triglyceride mass balance demonstrates the accuracy of TG method to quantify triglycerides content in oleaginous biomass.

TG method for quantification of triglycerides in algal biomass was demonstrated with dry oleaginous biomass. This method may also be extended to quantify triglycerides content for wet algal slurry obtained after dewatering the algal culture.

Table 4.4: Summary of triglyceride mass balance around solvent extraction method for algal samples. Sum of triglyceride content in solvent extract and solid residue is statistically similar to values obtained for whole biomass. Triglyceride

98 content in whole biomass was estimated using in situ transesterification method. Triglyceride content in solid residue was estimated by TG methods as the results obtained from TG methods and in situ transesterification procedure is statistically similar.

Before measuring triglyceride content (on dry-basis) in wet algal sample by TG method, moisture content of wet sample has to be estimated using TG-based ASTM protocol (ASTM D-7582-12) to determine dry weight of the sample. The temperature program of TG for wet algal slurry involves heating to 105 ± 3 °C for 60 min. It was followed by heating to 320 °C at ramp rate of 20 °C/min and maintained at that temperature for 15 min.(see Figure 4-7).

Figure 4-7: Residual weight (dashed lines) and temperature (solid lines) of wet Chlorella sp. sample with time obtained while employing temperature protocol to determine their triglyceride content. Again, it was heated to 320 to 420 °C and maintained at 420 °C for 15 min. Mass loss in the region I of Figure 4-7was accounted to the moisture content of the sample.

This moisture content determined was used to determine triglyceride content of the

99 sample in dry-basis. After calculating moisture content, triglyceride content of the wet sample was estimated as described earlier for dry algae sample. The results are tabulated in Table 4.5.

Table 4.5: Comparison of triglyceride content of wet and dry sample of Chlorella sp. estimated via TG method. Triglyceride content (on dry-basis) of wet sample is statistically similar to estimates obtained for dry sample estimated using TG method and In situ transesterification.

Triglyceride content of Chlorella sp. estimated for dry sample using TG method is statistically similar to values obtained for wet sample. Thereby, TG method can be extended to quantify triglycerides content for wet algal slurry obtained after dewatering the algal culture.

Traditional analytical methods for quantifying triglycerides content in the oleaginous algae or biomass require chemicals or organic solvents, involve multiple steps which are time consuming and have to use gravimetry and/or chromatography analytical equipment. In contrast, TG method is simple and does not need chemicals as shown in

Figure 4-8. This method needs small amount of dry or wet algal sample, 10 mg, (solvent extraction technique and in situ transesterification method requires 50 and 30 mg, respectively) and can be performed in shorter time when compared to other conventional

100 methods.

Figure 4-8: Operations or steps involved in different analytical methods for quantifying triglyceride content in oleaginous feedstocks.

1.17 Conclusions:

In this study, we developed methodology of thermo-gravimetry based analytical method for triglyceride quantification in oleaginous and/or algal biomass. Results obtained from this TG method were statistically similar to those estimated from in situ transesterification method for oleaginous and/or algal biomass. This method may be employed to quantify triglycerides content in wet samples and also to verify the presence of triglycerides in biomass.

101

References

1. Cheng, J., ed. Biomass to Renewable Energy Processes. 2010, CRC press,

Taylor& Francis Group.

2. Mousdale, D.M., Introduction to biofuels. 2010: Taylor & Francis Group.

3. Petroleum, B., BP Statistical Review of World Energy June 2014. 2014, British

Petroleum.

4. Demirbas, A., ed. Biofuels. 2009, Springer.

5. Huber, G.W., S. Iborra, and A. Corma, Synthesis of Transportation Fuels from

Biomass:Chemistry, Catalysts, and Engineering. Chemical Reviews, 2006.

106(9): p. 4044-4098.

6. Bridgwater, A.V., D. Meier, and D. Radlein, An overview of fast pyrolysis of

biomass. Organic Geochemistry, 1999. 30(12): p. 1479-1493.

7. Yang, H., et al., Characteristics of hemicellulose, cellulose and lignin pyrolysis.

Fuel, 2007. 86(12-13): p. 1781-1788.

8. Viamajala, S., et al., Heat and Mass Transport in Processing of Lignocellulosic

Biomass for Fuels and Chemicals, in Sustainable Biotechnology, O.V. Singh and

S.P. Harvey, Editors. 2010, Springer Netherlands. p. 1-18.

102

9. Samayam, I.P. and C.A. Schall, Saccharification of ionic liquid pretreated

biomass with commercial enzyme mixtures. Bioresource Technology, 2010.

101(10): p. 3561-3566.

10. Dadi, A.P., S. Varanasi, and C.A. Schall, Enhancement of cellulose

saccharification kinetics using an ionic liquid pretreatment step. Biotechnology

and Bioengineering, 2006. 95(5): p. 904-910.

11. Dadi, A., C. Schall, and S. Varanasi, Mitigation of cellulose recalcitrance to

enzymatic hydrolysis by ionic liquid pretreatment. Applied Biochemistry and

Biotechnology, 2007. 137-140(1): p. 407-421.

12. Chheda, J.N., Y. Roman-Leshkov, and J.A. Dumesic, Production of 5-

hydroxymethylfurfural and furfural by dehydration of biomass-derived mono- and

poly-saccharides. Green Chemistry, 2007. 9(4): p. 342-350.

13. Bozell, J.J., et al., Production of levulinic acid and use as a platform chemical for

derived products. Resources, Conservation and Recycling, 2000. 28(3–4): p. 227-

239.

14. Chung, D., et al., Direct conversion of plant biomass to ethanol by engineered

Caldicellulosiruptor bescii. Proceedings of the National Academy of Sciences of

the United States of America, 2014. 111(24): p. 8931-6.

15. Luque, L., et al., Pyrolysis based bio-refinery for the production of bioethanol

from demineralized ligno-cellulosic biomass. Bioresource Technology, 2014. 161:

p. 20-28.

16. Park, Y.K., et al., Catalytic Conversion of Cellulose Over Mesoporous Y Zeolite.

Journal of Nanoscience and Nanotechnology, 2014. 14(7): p. 5120-5123.

103

17. Hu, F.S., et al., Synthesis and Characterization of Thermosetting Furan-Based

Epoxy Systems. Macromolecules, 2014. 47(10): p. 3332-3342.

18. Dornath, P. and W. Fan, Dehydration of fructose into furans over zeolite catalyst

using carbon black as adsorbent. Microporous and Mesoporous Materials, 2014.

191: p. 10-17.

19. Holladay, E.J., et al., Top Value-Added Chemicals from Biomass Volume II -

Results of Screening for Potential Candidates from Biorefinery Lignin, P.N.N.

Laboratory, Editor. 2007.

20. Klein-Marcuschamer, D., B.A. Simmons, and H.W. Blanch, Techno-economic

analysis of a lignocellulosic ethanol biorefinery with ionic liquid pre-treatment.

Biofuels, Bioproducts and Biorefining, 2011. 5(5): p. 562-569.

21. Kim, S. and B.E. Dale, Global potential bioethanol production from wasted crops

and crop residues. Biomass and Bioenergy, 2004. 26(4): p. 361-375.

22. von Blottnitz, H. and M.A. Curran, A review of assessments conducted on bio-

ethanol as a transportation fuel from a net energy, greenhouse gas, and

environmental life cycle perspective. Journal of Cleaner Production, 2007. 15(7):

p. 607-619.

23. Gardner, R., et al., Medium pH and nitrate concentration effects on accumulation

of triacylglycerol in two members of the chlorophyta. Journal of Applied

Phycology, 2011. 23(6): p. 1005-1016.

24. Hu, Q., et al., Microalgal triacylglycerols as feedstocks for biofuel production:

perspectives and advances. The Plant Journal, 2008. 54(4): p. 621-639.

104

25. Minowa, T., et al., Oil production from algal cells of Dunaliella tertiolecta by

direect thermochemical liquefaction. Fuel, 1995. 74(12): p. 1735-1738.

26. Miao, X. and Q. Wu, Biodiesel production from heterotrophic microalgal oil.

Bioresource Technology, 2006. 97(6): p. 841-846.

27. Gardner, R., et al., Use of sodium bicarbonate to stimulate triacylglycerol

accumulation in the chlorophyte Scenedesmus sp. and the diatom Phaeodactylum

tricornutum. Journal of Applied Phycology, 2012: p. 1-10.

28. Griffiths, M. and S. Harrison, Lipid productivity as a key characteristic for

choosing algal species for biodiesel production. Journal of Applied Phycology,

2009. 21(5): p. 493-507.

29. Georgianna, D.R. and S.P. Mayfield, Exploiting diversity and synthetic biology

for the production of algal biofuels. Nature, 2012. 488(7411): p. 329-335.

30. Bligh, E.G. and W.J. Dyer, A Rapid Method of Total Lipid Extraction and

Purification. Canadian Journal of Biochemistry and Physiology, 1959. 37(8): p.

911-917.

31. Kumari, P., C.R.K. Reddy, and B. Jha, Comparative evaluation and selection of a

method for lipid and fatty acid extraction from macroalgae. Analytical

Biochemistry, 2011. 415(2): p. 134-144.

32. Behr, A., A. Westfechtel, and J. Pérez Gomes, Catalytic Processes for the

Technical Use of Natural Fats and Oils. Chemical Engineering & Technology,

2008. 31(5): p. 700-714.

105

33. Maher, K.D. and D.C. Bressler, Pyrolysis of triglyceride materials for the

production of renewable fuels and chemicals. Bioresource Technology, 2007.

98(12): p. 2351-2368.

34. Mercer, P. and R.E. Armenta, Developments in oil extraction from microalgae.

European Journal of Lipid Science and Technology, 2011. 113(5): p. 539-547.

35. Amaro, H.M., A.C. Guedes, and F.X. Malcata, Advances and perspectives in

using microalgae to produce biodiesel. Applied Energy, 2011. 88(10): p. 3402-

3410.

36. Lee, S., B.-D. Yoon, and H.-M. Oh, Rapid method for the determination of lipid

from the green alga Botryococcus braunii. Biotechnology Techniques, 1998.

12(7): p. 553-556.

37. D'Oca, M.G.M., et al., Production of FAMEs from several microalgal lipidic

extracts and direct transesterification of the Chlorella pyrenoidosa. Biomass and

Bioenergy, 2011. 35(4): p. 1533-1538.

38. Mendes, R.L., et al., Supercritical carbon dioxide extraction of compounds with

pharmaceutical importance from microalgae. Inorganica Chimica Acta, 2003.

356(0): p. 328-334.

39. Al-Zuhair, S., Production of biodiesel: possibilities and challenges. Biofuels,

Bioproducts and Biorefining, 2007. 1(1): p. 57-66.

40. Rosenberg, J.N., et al., A green light for engineered algae: redirecting

metabolism to fuel a biotechnology revolution. Current Opinion in Biotechnology,

2008. 19(5): p. 430-436.

106

41. Chen, P., et al., Review of biological and engineering aspects of algae to fuels

approach. 2010.

42. Miao, X., Q. Wu, and C. Yang, Fast pyrolysis of microalge to produce renewable

fuels. Journal of analytical and applied pyrolysis, 2004(71): p. 855-863.

43. Mohan, D., U.C. Pittman, and H.P. Steele, Pyrolysis of wood/biomass for bio-oil:

A critical review. Energy & Fuels, 2006. 20(3): p. 848-889.

44. Demirbas, A., Pyrolysis Mechanism of Biomass Materials. Energy sources, 2009.

Part a(31): p. 1186-1193.

45. Bridgwater, V.A. and C.V.G. Peacocke, Fast pyrolysis processes for biomass.

Renewable & Sustainable Energy Review, 2000. 4(1): p. 1-73.

46. Wright, M.M., et al., Techno-Economic Analysis of Biomass Fast Pyrolysis to

Transportation Fuels. 2010, National Renewable Energy Laboratory.

47. Ingram, L., et al., Pyrolysis of wood and bark in an Augar Reactor: Physical

properties and chemical analysis of the produced bio-oil. Energy & Fuels, 2008.

22(1): p. 614-625.

48. Czernik, S. and V.A. Bridgwater, Overview of application of biomass fast

pyrolysis oil. Energy & Fuels, 2004. 18(2): p. 590-598.

49. Lehmann, J., J. Gaunt, and M. Rondon, Bio-char sequetration in terrestrial

ecosystem - a review. Mitigation and adaptation strategies for global change,

2006. 11(2): p. 395-419.

50. Lehmann, J. and S. Joseph, eds. Biochar for Environmental Management Science

and Technology. 2009, Earthscan: London, UK.

107

51. Miao, X. and Q. Wu, High yield bio-oil production from fast pyrolysis by

metabolic controlling of chlorella protothecoides. Journal of Biotechnology,

2004(110): p. 85-93.

52. Yanik, J., et al., Pyrolysis of algal biomass. Journal of Analytical and Applied

Pyrolysis, 2013. 103(0): p. 134-141.

53. Li, R., et al., Selection of Temperature for Bio-oil Production from Pyrolysis of

Algae from Lake Blooms. Energy & Fuels, 2012. 26(5): p. 2996-3002.

54. Sanchez-Silva, L., et al., Pyrolysis, combustion and gasification characteristics of

Nannochloropsis gaditana microalgae. Bioresource Technology, 2013. 130(0): p.

321-331.

55. Kim, S.-S., et al., Pyrolysis characteristics and kinetics of the alga Saccharina

japonica. Bioresource Technology, 2012. 123(0): p. 445-451.

56. Trinh, T.N., et al., Comparison of Lignin, Macroalgae, Wood, and Straw Fast

Pyrolysis. Energy & Fuels, 2013. 27(3): p. 1399-1409.

57. Hu, Z., et al., Bio-oil production through pyrolysis of blue-green algae blooms

(BGAB): Product distribution and bio-oil characterization. Energy, 2013. 52(0):

p. 119-125.

58. Chaiwong, K., et al., Study of bio-oil and bio-char production from algae by slow

pyrolysis. Biomass and Bioenergy, 2013. 56(0): p. 600-606.

59. Kim, S.W., B.S. Koo, and D.H. Lee, A comparative study of bio-oils from

pyrolysis of microalgae and oil seed waste in a fluidized bed. Bioresource

Technology, 2014. 162(0): p. 96-102.

108

60. Yuan, T., A. Tahmasebi, and J. Yu, Comparative study on pyrolysis of

lignocellulosic and algal biomass using a thermogravimetric and a fixed-bed

reactor. Bioresource Technology, (0).

61. Peng, W., Q. Wu, and P. Tu, Effects of temperature and holding time on

production of renewable fuels from pyrolysis of Chlorella protothecoides. Journal

of Applied Phycology, 2000. 12(2): p. 147-152.

62. Muradov, N., et al., Pyrolysis of fast growing aquatic biomass- Lemna minor

(duckweed): Characterization of pyrolysis products. Bioresource Technology,

2010. 101(21): p. 8424-8428.

63. Wang, S., et al., Compositional analysis of bio-oil derived from pyrolysis of

seaweed. Energy Conversion and Management, 2013. 68(0): p. 273-280.

64. Bae, Y.J., et al., The characteristics of bio-oil produced from the pyrolysis of

three marine macroalgae. Bioresource Technology, 2011. 102(3): p. 3512-3520.

65. Du, Z., et al., Microwave-assisted pyrolysis of microalgae for biofuel production.

Bioresource Technology, 2011. 102(7): p. 4890-4896.

66. Vispute, T.P., et al., Renewable Chemical Commodity Feedstocks from Integrated

Catalytic Processing of Pyrolysis Oils. Science, 2010. 330(6008): p. 1222-1227.

67. Elliot, C.D., et al., Catalytic hydroprocessing of biomass fast pyrolysis bio-oil to

produce hydrocarbon products. Environmental Progress & Sustainable Energy,

2009. 28(3): p. 441-449.

68. Lange, J.-P., Lignocellulosic conversion: an introduction to chemistry, process

and economics. Biofuels, Bioproducts and Biorefining, 2007. 1(1): p. 39-48.

109

69. Garcia-Perez, M., et al., Production and Fuel Properties of Pine Chip Bio-

oil/Biodiesel Blends. Energy & Fuels, 2007(21): p. 2363-2372.

70. Chaudhari, T.S., K.A. Dalai, and N.N. Bakhshi, Production of hydrogen and /or

syngas via steam gasification of biomass-derived chars. Energy & Fuels, 2003.

17(4): p. 1062-1067.

71. Smith, H.V., et al., The ecology of algal biodiesel production. Trends in Ecology

& Evolution, 2010. 25(5): p. 301-309.

72. Posten, C. and G. Schaub, Microalgae and terrestrial biomass as source of fuels -

a process view. Journal of Biotechnology, 2009. 142(1): p. 64-69.

73. DOE, U.S., National Algal Biofuels Technology Roadmap, in U.S Department of

Energy Office of Energy Efficiency and Renewable Energy Biomass Program

2010.

74. Hoagland, P. and S. Scatasta, The Economic Effects of Harmful Algal Blooms, in

Ecology of Harmful Algae, E. Granéli, et al., Editors. 2006, Springer Berlin

Heidelberg. p. 391-402.

75. Dodds, W.K., et al., Eutrophication of U.S. Freshwaters: Analysis of Potential

Economic Damages. Environmental Science & Technology, 2008. 43(1): p. 12-

19.

76. Smith, A.S., Harmful Algal Blooms: Action Plans for Scientific Solutions, H.

Congressional Testimony, Editor. 2011.

77. Hudnell, H.K., Cyanobacterial harmful algal blooms: state of the science and

research needs. 2008: Springer.

110

78. Ross, A.B., et al., Classification of maroalgae as fuel and its thermochemical

behaviour. Bioresource Technology, 2008. 99: p. 6494-6504.

79. Ross, A.B., et al., Investigation of the pyrolysis behaviour of brown algae before

and after pre-treatment using PY-GC/MS and TGA. Journal of analytical and

applied pyrolysis, 2009(85): p. 3-10.

80. Li, D., et al., Pyrolytic characteristics and kinetic studies of three kinds of red

algae. Biomass and Bioenergy, 2011. 35(5): p. 1765-1772.

81. Demirbas, A., Hydrogen from mosses and algae via pyrolysis and steam

gasification. Energy sources, 2010. Part A(32): p. 172-179.

82. Sluiter, A., et al., Determination of Structural Carbohyddrates and Lignin in

Biomass. 2008, National Renewable Energy Laboratory.

83. Hames, B., C. Scarlata, and A. Sluiter, Determination of protein content in

biomass. 2008, National Renewable Energy Laboratory.

84. Petrakis, L. and D.W. Grandy, Coal Analysis, Characterization and petrography.

Journal of chemical education, 1980. 57(10): p. 689-694.

85. Sluiter, A., et al., Determination of Ash in Biomass. 2008, National Renewable

Energy Laboratory.

86. Mahinpey, N., et al., Analysis of Bio-oil, Bio-gas and Biochar from Pressurized

Pyrolysis of Wheat Straw Using a Tubular Reactor. Energy & Fuels, 2009(23): p.

2736-2742.

87. Friedl, A., et al., Prediction of heating values of biomass fuel from elemental

composition. Analytica Chimica Acta, 2005. 544(1–2): p. 191-198.

111

88. Shuping, Z., et al., Pyrolysis characterisrtics and kinetics of the marine

microalgae dunaliella tertiolecta using thermogravimetric analyzer. Bioresource

Technology, 2010. 101(21): p. 359-365.

89. Zaid, A.A. and G.O. Taiwo, Comparitive Utilization of Biodegraded Ricehusk.

Journal of Fisheries and Aquatic Science, 2008. 3(5): p. 312-319.

90. Yang, H., et al., In-Depthinvestigation of biomass pyrolysis based on three major

components: Hemicellulose, Cellulose and Lignin. Energy & Fuels, 2006. 20: p.

388-393.

91. Yang, H., et al., Characteristics of hemicellolose, cellolose and lignin pyrolysis.

Fuel, 2007(86): p. 1781-1788.

92. Guinesi, S.L., et al., Kinetics of thermal degradation applied to starches from

different botanical origins by non-isothermal procedures. Thermochimica Acta,

2006. 447(2): p. 190-196.

93. Grierson, S., et al., Thermal characterisation of microalgae under slow pyrolysis

conditions. Journal of analytical and applied pyrolysis, 2009(85): p. 118-123.

94. Ates, F. and M.A. IsIkdag, Influence of temperature and alumina catalyst on

pyrolysis of corncob. Fuel, 2009. 88(10): p. 1991-1997.

95. Natarajan, E. and E. Ganapathy Sundaram, Pyrolysis of Rice Husk in a Fixed Bed

Reactor. World Academy of Science, Engineering and Technology, 2009. 56: p.

504.

96. Van de Velden, M., et al., Fundamentals, kinetics and endothermicity of the

biomass pyrolysis reaction. Renewable Energy, 2010. 35(1): p. 232-242.

112

97. He, F., W. Yi, and X. Bai, Investigation on caloric requirement of biomass

pyrolysis using TGA/DSC analyzer. Energy Conversion and Management, 2006.

47(15-16): p. 2461-2469.

98. Frank, E., et al., Life cycle comparison of hydrothermal liquefaction and lipid

extraction pathways to renewable diesel from algae. Mitigation and adaptation

strategies for global change, 2013. 18(1): p. 137-158.

99. Scott, S.A., et al., Biodiesel from algae: challenges and prospects. Current

Opinion in Biotechnology, 2010. 21(3): p. 277-286.

100. Pienkos, P.T. and A. Darzins, The promise and challenges of microalgal-derived

biofuels. Biofuels, Bioproducts and Biorefining, 2009. 3(4): p. 431-440.

101. Nelson, R.D., Transesterification and recovery of intracelluar lipids using a

single step reactive extraction., in Biological Engineering. 2010, Utah State

University: Logan, Utah, US. p. 71.

102. Laurens, L.L., et al., Accurate and reliable quantification of total microalgal fuel

potential as fatty acid methyl esters by in situ transesterification. Analytical and

Bioanalytical Chemistry, 2012. 403(1): p. 167-178.

103. Singh, A., P.S. Nigam, and J.D. Murphy, Mechanism and challenges in

commercialisation of algal biofuels. Bioresource Technology, 2011. 102(1): p.

26-34.

104. Lewis, T., P.D. Nichols, and T.A. McMeekin, Evaluation of extraction methods

for recovery of fatty acids from lipid-producing microheterotrophs. Journal of

Microbiological Methods, 2000. 43(2): p. 107-116.

113

105. Haas, M.J. and K. Wagner, Simplifying biodiesel production: The direct or in situ

transesterification of algal biomass. European Journal of Lipid Science and

Technology, 2011. 113(10): p. 1219-1229.

106. Johnson, M.B. and Z. Wen, Production of Biodiesel Fuel from the Microalga

Schizochytrium limacinum by Direct Transesterification of Algal Biomass. Energy

& Fuels, 2009. 23(10): p. 5179-5183.

107. Bryant, H.L., et al., The value of post-extracted algae residue. Algal Research,

2012. 1(2): p. 185-193.

108. Brown, T.M., P. Duan, and P.E. Savage, Hydrothermal Liquefaction and

Gasification of Nannochloropsis sp. Energy & Fuels, 2010. 24(6): p. 3639-3646.

109. Jena, U. and K.C. Das, Comparative Evaluation of Thermochemical Liquefaction

and Pyrolysis for Bio-Oil Production from Microalgae. Energy & Fuels, 2013.

25(11): p. 5472-5482.

110. Valdez, P.J., et al., Hydrothermal liquefaction of Nannochloropsis sp.: Systematic

study of process variables and analysis of the product fractions. Biomass and

Bioenergy, 2013. 46(0): p. 317-331.

111. Chow, M.C., et al., Thermal Treatment of Algae for Production of Biofuel. Energy

& Fuels, 2013. 27(4): p. 1926-1950.

112. Toor, S.S., L. Rosendahl, and A. Rudolf, Hydrothermal liquefaction of biomass:

A review of subcritical water technologies. Energy, 2011. 36(5): p. 2328-2342.

113. Maddi, B., S. Viamajala, and S. Varanasi, Comparative study of pyrolysis of algal

biomass from natural lake blooms with lignocellulosic biomass. Bioresource

Technology, 2011. 102(23): p. 11018-11026.

114

114. Brennan, L. and P. Owende, Biofuels from microalgae - A review of technologies

for production, processing, and extractions of biofuels and co-products.

Renewable and Sustainable Energy Reviews, 2010. 14(2): p. 557-577.

115. Ferrera-Lorenzo, N., et al., Conventional and microwave pyrolysis of a

macroalgae waste from the Agar–Agar industry. Prospects for bio-fuel

production. Bioresource Technology, 2014. 151(0): p. 199-206.

116. Ferrera-Lorenzo, N., et al., Pyrolysis characteristics of a macroalgae solid waste

generated by the industrial production of Agar–Agar. Journal of Analytical and

Applied Pyrolysis, 2014. 105(0): p. 209-216.

117. Patwardhan, P.R., et al., Product distribution from fast pyrolysis of glucose-based

carbohydrates. Journal of analytical and applied pyrolysis, 2009. 86(2): p. 323-

330.

118. Douda, J. and V.A. Basiuk, Pyrolysis of amino acids: recovery of starting

materials and yields of condensation products. Journal of analytical and applied

pyrolysis, 2000. 56(1): p. 113-121.

119. Solina, M., et al., Volatile aroma components of soy protein isolate and acid-

hydrolysed vegetable protein. Food Chemistry, 2005. 90(4): p. 861-873.

120. Biller, P. and A.B. Ross, Potential yields and properties of oil from the

hydrothermal liquefaction of microalgae with different biochemical content.

Bioresource Technology, 2011. 102(1): p. 215-225.

121. Duan, P. and P.E. Savage, Hydrothermal Liquefaction of a Microalga with

Heterogeneous Catalysts. Industrial & Engineering Chemistry Research, 2011.

50(1): p. 52-61.

115

122. Bridgwater, A.V., Review of fast pyrolysis of biomass and product upgrading.

Biomass and Bioenergy, 2012. 38(0): p. 68-94.

123. Chowdhury, R. and S. Viamajala, Alkaliphilic Microalgae for Biodiesel

Production, in 33rd Symposium on Biotechnology for Fuels and Chemicals. 2011:

Seattle, WA.

124. Lohman, E.J., et al., An efficient and scalable extraction and quantification

method for algal derived biofuel. Journal of Microbiological Methods, 2013.

94(3): p. 235-244.

125. Goodrum, J.W. and D.P. Geller, Rapid thermogravimetric measurements of

boiling points and vapor pressure of saturated medium- and long-chain

triglycerides. Bioresource Technology, 2002. 84(1): p. 75-80.

126. Perry, E.S., W.H. Weber, and B.F. Daubert, Vapor Pressures of Phlegmatic

Liquids. I. Simple and Mixed Triglycerides. Journal of the American Chemical

Society, 1949. 71(11): p. 3720-3726.

127. Boateng, A.A., K.B. Hicks, and K.P. Vogel, Pyrolysis of switchgrass (Panicum

virgatum) harvested at several stages of maturity. Journal of analytical and

applied pyrolysis, 2006. 75(2): p. 55-64.

128. Boateng, A.A., H.G. Jung, and P.R. Adler, Pyrolysis of energy crops including

alfalfa stems, reed canarygrass, and eastern gamagrass. Fuel, 2006. 85(17â€“18):

p. 2450-2457.

129. Mullen, C.A. and A.A. Boateng, Chemical Composition of Bio-oils Produced by

Fast Pyrolysis of Two Energy Crops Energy & Fuels, 2008. 22(3): p. 2104-2109.

116

130. Gunstone, F.D., J.L. Harwood, and F.B. Padley, The Lipid Handbook. 1994:

Chapman and Hall.

131. Lin, Y.-C., et al., Kinetics and Mechanism of Cellulose Pyrolysis. The Journal of

Physical Chemistry C, 2009. 113(46): p. 20097-20107.

132. Smith, W.T., T.B. Harris, and J.M. Patterson, Pyrolysis of soybean protein and an

amino acid mixture having the same amino acid composition. Journal of

Agricultural and Food Chemistry, 1974. 22(3): p. 480-483.

133. Linstrom, P.J. and W.G. Mallard, NIST Chemistry WebBook, NIST Standard

Reference Database Number 69. 2005: National Institute of Standards and

Technology.

134. Reck, R., Nitrogen derivatives (Amides, diamides, nitriles, primary amines and

oxides). Journal of the American Oil Chemists' Society, 1979. 56(11): p. 796A-

801A.

135. Muth, S.A., Diesel fuel, diesel fuel additive, and associated method for using the

same. 2008, The Penray Companies, Inc.

136. Maag, H., Fatty acid derivatives: Important surfactants for household, cosmetic

and industrial purposes. Journal of the American Oil Chemists' Society, 1984.

61(2): p. 259-267.

137. DeJarlais, W. and H. Teeter, Nitrogenous derivatives of cyclic fatty acids. Journal

of the American Oil Chemists' Society, 1962. 39(10): p. 421-424.

138. Wang, K., et al., Fast pyrolysis of microalgae remnants in a fluidized bed reactor

for bio-oil and biochar production. Bioresource Technology, 2013. 127(0): p.

494-499.

117

139. Friedl, A., et al., Prediction of heating values of biomass fuel from elemental

composition. Analytica Chimica Acta, 2005. 544: p. 191-198.

140. Jorgenson, R. and D. Struble, Tapered Drum Pyrolysis (USPTO Application

number 13/237,331). 2011, Red Lion Bio-Energy Technologies: USA.

141. Singh, N.R., et al., Estimation of Liquid Fuel Yields from Biomass. Environmental

Science & Technology, 2010. 44(13): p. 5298-5305.

142. Park, W.C., A study of pyrolysis of charring materials and its applicationt to fire

safety and biomass utilization, in Mechanical Engineering. 2008, The University

of Michigan: Ann Arbor. p. 140.

143. Collazo, J., et al., Determination of the specific heat of biomass materials and the

combustion energy of coke by DSC analysis. Energy, 2012. 45(1): p. 746-752.

144. Daugaard, D.E. and R.C. Brown, Enthalpy for Pyrolysis for Several Types of

Biomass. Energy & Fuels, 2003. 17(4): p. 934-939.

145. Du, Z., et al., Hydrothermal pretreatment of microalgae for production of

pyrolytic bio-oil with a low nitrogen content. Bioresource Technology, 2012.

120(0): p. 13-18.

146. Minowa, T., T. Kondo, and S.T. Sudirjo, Thermochemical liquefaction of

indonesian biomass residues. Biomass and Bioenergy, 1998. 14(5-6): p. 517-524.

147. McAdams, W.H., Heat Transmission. 3rd ed. 1954: Mc-Graw Hill.

148. Lohman, E.J., et al., An efficient and scalable extraction and quantification

method for algal derived biofuel. Journal of Microbiological Methods, 2013(0).

149. Brown, E., Introduction to Thermal Analysis: Techniques and Applications. 2001:

Springer.

118

150. Kök, M., Recent developments in the application of thermal analysis techniques

in fossil fuels. Journal of Thermal Analysis and Calorimetry, 2008. 91(3): p. 763-

773.

151. Elder, J.P., Proximate analysis by automated thermogravimetry. Fuel, 1983.

62(5): p. 580-584.

152. Canetti, M., et al., Thermal degradation behaviour of isotactic polypropylene

blended with lignin. Polymer Degradation and Stability, 2006. 91(3): p. 494-498.

153. Laresgoiti, M.F., et al., Characterization of the liquid products obtained in tyre

pyrolysis. Journal of analytical and applied pyrolysis, 2004. 71(2): p. 917-934.

154. Friedman, H.L., Kinetics of thermal degradation of char-forming plastics from

thermogravimetry. Application to a phenolic plastic. Journal of Polymer Science

Part C: Polymer Symposia, 1964. 6(1): p. 183-195.

155. Kok, M.V. and M.R. Pamir, Pyrolysis and combustion studies of fossil fuels by

thermal analysis methods. Journal of analytical and applied pyrolysis, 1995.

35(2): p. 145-156.

156. Murugan, P., et al., Kinetics of the Pyrolysis of Lignin Using Thermogravimetric

and Differential Scanning Calorimetry Methods. Energy & Fuels, 2008. 22(4): p.

2720-2724.

157. Damartzis, T., et al., Thermal degradation studies and kinetic modeling of

cardoon (Cynara cardunculus) pyrolysis using thermogravimetric analysis

(TGA). Bioresource Technology, 2011. 102(10): p. 6230-6238.

158. Doyle, C.D., Kinetic analysis of thermogravimetric data. Journal of Applied

Polymer Science, 1961. 5(15): p. 285-292.

119

159. Harlick, P.J.E. and A. Sayari, Applications of Pore-Expanded Mesoporous Silicas.

3. Triamine Silane Grafting for Enhanced CO2 Adsorption. Industrial &

Engineering Chemistry Research, 2006. 45(9): p. 3248-3255.

160. Liu, Q., et al., Mechanism study of wood lignin pyrolysis by using TG-FTIR

analysis. Journal of analytical and applied pyrolysis, 2008. 82(1): p. 170-177.

161. Worasuwannarak, N., T. Sonobe, and W. Tanthapanichakoon, Pyrolysis

behaviors of rice straw, rice husk, and corncob by TG-MS technique. Journal of

analytical and applied pyrolysis, 2007. 78(2): p. 265-271.

162. Arii, T., T. Senda, and N. Fujii, A combined thermogravimetric-gas

chromatographic/mass spectrometric analysis (TG-GC/MS) using a high

resolution TG technique. Thermochimica Acta, 1995. 267(0): p. 209-221.

163. ASTMStandard, Standard Test Methods for Proximate Analysis of Coal and Coke

by Macro Thermogravimetric Analysis. 2010, ASTM International.

164. Wahlen, B.D., R.M. Willis, and L.C. Seefeldt, Biodiesel production by

simultaneous extraction and conversion of total lipids from microalgae,

cyanobacteria, and wild mixed-cultures. Bioresource Technology, 2011. 102(3):

p. 2724-2730.

165. Indarti, E., et al., Direct FAME synthesis for rapid total lipid analysis from fish oil

and cod liver oil. Journal of Food Composition and Analysis, 2005. 18(2–3): p.

161-170.

166. Cayli, G. and S. Kusefoglu, Increased yields in biodiesel production from used

cooking oils by a two step process: Comparison with one step process by using

TGA. Fuel Processing Technology, 2008. 89(2): p. 118-122.

120

167. Zakzeski, J., et al., The Catalytic Valorization of Lignin for the Production of

Renewable Chemicals. Chemical Reviews, 2010. 110(6): p. 3552-3599.

168. Hatakeyama, T. and H. Hatakeyama, Thermal properties of green polymers and

biocomposites. 2004: Kluwer Academic Publishers.

169. Doherty, W.O.S., P. Mousavioun, and C.M. Fellows, Value-adding to cellulosic

ethanol: Lignin polymers. Industrial Crops and Products, 2010. 33(2): p. 259-276.

170. Pandey, M.P. and C.S. Kim, Lignin Depolymerization and Conversion: A Review

of Thermochemical Methods. Chemical Engineering & Technology, 2010. 34(1):

p. 29-41.

171. Scholze, B. and D. Meier, Characterization of the water-insoluble fraction from

pyrolysis oil (pyrolytic lignin). Part I. PY-GC/MS, FTIR, and functional groups.

Journal of analytical and applied pyrolysis, 2001. 60(1): p. 41-54.

172. Scholze, B., C. Hanser, and D. Meier, Characterization of the water-insoluble

fraction from fast pyrolysis liquids (pyrolytic lignin): Part II. GPC, carbonyl

goups, and 13C-NMR. Journal of analytical and applied pyrolysis, 2001. 58-59: p.

387-400.

173. Nowakowski, D.J., et al., Lignin fast pyrolysis: Results from an international

collaboration. Journal of analytical and applied pyrolysis, 2010. 88(1): p. 53-72.

174. Beis, S.H., et al., Fast pyrolysis of lignins. Bioresources, 2010. 5(3): p. 1408-

1424.

175. Mahinpey, N., et al., Analysis of Bio-Oil, Biogas, and Biochar from Pressurized

Pyrolysis of Wheat Straw Using a Tubular Reactor. Energy & Fuels, 2009. 23(5):

p. 2736-2742.

121

176. Sazanov, Y., et al., Thermochemical transformations of hydrolysis lignin. Russian

Journal of Applied Chemistry, 2010. 83(9): p. 1607-1614.

177. Ferdous, D., et al., Pyrolysis of Lignins: Experimental and Kinetics Studies.

Energy & Fuels, 2002. 16(6): p. 1405-1412.

178. Kadla, J.F. and S. Kubo, Lignin-based polymer blends: analysis of intermolecular

interactions in lignin-synthetic polymer blends. Composites Part A: Applied

Science and Manufacturing, 2004. 35(3): p. 395-400.

122

Appendix A

A. Characterization of pyrolysis products of residual lignin

obtained via IL pretreatment followed by enzymatic

hydrolysis

Introduction:

Ionic liquid (IL) pretreatment of lignocellulose is a promising technology that mitigates the cellulose recalcitrance and does not produce any degradation products that would inhibit or interfere with the subsequent enzymatic hydrolysis. Also, IL-pretreated biomass has proven to produce high yields of sugar monomer units at low process time

[9-11]. These monomeric sugars units obtained could be converted to produce wide range of fuels and chemicals such as ethanol, succinic acid, lactic acid, furfural, hydroxyl- methyl-furfural by various catalytic processes[12]. Biomass residue left behind primarily consists of lignin and hence it is called “residual lignin” which accounts up to 30 wt% of

123 dry weight as well as 40 % of the energy content of biomass (see figure 1) [5, 19].

Figure 1 Flowchart of production of residual lignin from IL-based bio-refinery.

At present, residual lignin produced on industrial-scale in pulp and paper industry fed to combustion unit for low-value process heat. [167] The factors that stands on the way of commercialization of IL- and cellulosic- based bio-refinery such as high cost of

ILs, speculative ethanol and feedstocks process can be negotiated when high-value added products are produced from residual lignin [19, 20].

Lignin is complex amorphous polymer composed of aliphatic propyl side chain and aromatic component which includes Para-hydroxyphenyl, guaiacyl, syringyl groups.

The relative proportion of aromatic monomer units primarily depends on the biomass source [168]. Lignin itself can be used as raw materials in various industrial processes as a substitute for fillers in polymers, carbon fibers, binders, rubber, polyurethane foams, epoxy resins and bio-dispersants to produce low-value products [169, 170]. To produce high-value products such as liquid hydrocarbons and/or aromatic chemicals, lignin should be de-polymerized and then followed by series of catalytic oxidation and/or reduction reactions [167, 170]. Selective catalytic de-polymerization of lignin occurs slowly and 124 requires expensive catalysts as well as harsh operating conditions which may not be economically feasible for large-scale production. In addition, extensively-studied pyrolysis technique could be employed to de-polymerize lignin rapidly on large-scale. It was determined that bio-oils (organic phase or heavy oils in particular) produced from lignin pyrolysis have unique structural characteristics that would provide an opportunity to produce specific aromatic hydrocarbons [167, 171, 172]. The non-biodegradable, heavy metal-free, and energy-dense solid residue left behind after lignin pyrolysis could be fed to combustion unit to produce process heat for biorefinery [50]. Bio-char itself can also be used as a feedstock for producing high-valuable materials such as carbon fibers and carbon nano-tubes [50].

Pyrolysis of kraft lignin, organosolv lignin and other industrial lignin were studied and reported the formation of solid residue or char and aromatic-rich bio-oils [170, 173,

174]. Since lignin isolation processes and pretreatment techniques would drastically affect its nature and structure [19, 170], lignin obtained from novel IL pretreatment technology may modify its physical &thermal properties and as well as its pyrolysis behavior. In addition, there is no existing literature on pyrolysis characteristics of IL pretreated residual lignin which motivated us to pursue it.

Materials and methods

Sample preparation and chemicals used:

Poplar (provided by National Renewable Energy Laboratory) was pretreated with ionic liquid, EmimOAc (Sigma Aldrich, MO, USA) and then hydrolyzed by Novozyme

(New York, NY) C-tech and H-tech enzymes. The process parameters for both

125 pretreatment and enzymatic hydrolysis were extensively described previously in the literature [9, 10]. Biomass residue was washed with water and then centrifuged to remove any trace amount of sugars. The precipitated solids were separated, freeze dried and stored in a desiccator.

2-methoxy phenol, 2,6-dimethoxy phenol and deuterated dimethylsulfoxide

(DMSO) were purchased from Sigma-Aldrich (St.Louis, MO).

Carbohydrate analysis:

The carbohydrate fraction of residual lignin samples was determined by 2-stage acid hydrolysis according to the standard laboratory analytical procedures developed by the National Renewable Energy Laboratory (NREL) [82].

Ultimate analysis:

The elemental analysis was performed by Perkin-Elmer 2400 Series II CHN

Elemental Analyzer. Based on the elemental composition, heating values were calculated using the following well established correlations [175].

퐻퐻푉 (푂퐿푆) = 1.87퐶2 – 144퐶 – 2802퐻 + 63.8퐶 ∗ 퐻 + 129푁 + 20147

(1)

퐻퐻푉 (푃퐿푆) = 5.22퐶2 − 319퐶 − 1647퐻 + 38.6퐶 ∗ 퐻 + 133푁 + 21028

(2)

Where C, H, N denotes the mass fractions of carbon, hydrogen and nitrogen, respectively. The acronym HHV represents High Heating Value. Above correlations were

126 obtained by using two different regression methods i.e ordinary least square (OLS) and partial least square (PLS) method, respectively. The HHV computed from both these empirical equations has the units of MJ/Kg. The reported HHV values were average of

HHV(OLS) and HHV(PLS) as suggested by [139].

HHV(OLS) + HHV(PLS) 퐻퐻푉 = (3) 2

Thermo-gravimetric analysis and Differential scanning calorimetry:

About 10 mg of sample was used to perform thermo-gravimetric analysis by using

TA instrument Q50 series thermo-gravimetry (TG). Sample was placed in a platinum crucible and heated to 600 °C at constant heating rates (10 °C/min, 20 °C/min and 30

°C/min). Nitrogen flow of about 40 mL/min was continuously purged over sample and reference pan of TG to achieve inert atmosphere for pyrolysis experiments. TG universal analysis software provided us with two profiles for every experiment – residual weight percent vs temperature (TGA) and temperature derivative of weight loss (wt%/ °C) vs temperature (DTG).

Heat flow into 7 mg lignin samples sealed in a 30 µL aluminum pans was monitored over an temperature interval of 50- 500 °C by Perkin Elmer Diamond

Differential Scanning Calorimeter (Shelton, CT). Empty 30 µL aluminum pan was considered as reference. The samples were heated to 50 °C, held at this temperature for 1 minute and then heated to 500 °C at a constant ramp rate of 5 °C/min. DSC experiments were performed under inert atmosphere by purging with N2 gas.

Characterization of bio-oils by gas chromatograph:

127

Bio-oils produced from lignin pyrolysis were analyzed by a Hewlett Packard 6890

Gas Chromatograph (GC) equipped with 5973 series mass selective detector (MS)

(Agilent Technologies, Santa Clara, CA) and Rtx-5 (30m ×0.25mm×0.25μm film thickness) fused silica capillary column (Restek, Bellefonte, PA). The following GC oven temperature program was adopted: the initial temperature of column oven was set to 60

°C, after 2 min, it was to 270 °C at a rate of 5 °C/min and maintained at 270 °C for 15 min. Helium was used as a carrier gas with a constant linear velocity of 25 cm/s in the column. The injector and mass spec temperatures were maintained at 270 °C. 200 µL of light and heavy fraction of lignin bio-oils were dissolved in 1 mL of methanol and hexane respectively. Chemical compounds corresponding to a peak observed in lignin bio-oil chromatograms were identified by GC-MS NIST98 library embedded in the chem.- station software.

Some chemical compounds of interest identified by GC-MS were quantified by

Shimadzu 2010 Gas chromatograph equipped with Restek MXT Bio-diesel column – 15 m x 0.32 mm ID x 0.1 µm and flame ionization detector (FID). The GC oven temperature program was held for 1 min at 60 °C and then heated to a temperature of 300 °C at a ramp rate of 5 °C/min and maintained at 300 °C for 6 min. Helium was used as a carrier gas with a constant linear velocity of 25 cm/s in the column. The injector temperature and

FID temperature was maintained at 300 °C and 370 °C, respectively.

Characterization of pyrolysis oil by 13C nuclear magnetic resonance (NMR) spectroscopy

Solution-state 13C NMR spectra were recorded using a Varian 600MHz Unity

Inova NMR spectrometer (Agilent Technologies, Santa Clara, CA). 13C NMR of

128 pyrolysis heavy oils were acquired using 70 mg of sample in 1 mL DMSO-d6 employing an inverse gated decoupling pulse sequence, 60 pulse angle, a total pulse delay of 5 s and

10000 scans. The sample was maintained at the temperature of 30 °C throughout the experiments.

Pyrolysis experiments:

Pyrolysis experiments were conducted in a vertical tubular reactor (L = 43 cm,

OD = 2.54 cm), made of stainless steel and placed in a vertical split shell electric furnace

(Applied Test Systems. Inc, Butler, PA)(see Figure 2.1). A K-type thermocouple would be in contact with the biomass during the experiments to directly measure the temperature inside the pyrolysis chamber. He was used to maintain inert atmosphere in the pyrolysis reactor and it also facilitates as carrier gas for volatile products. 316L MCS series mass flow controllers (Alicat Scientific, Tucson, AZ) were used to control the flow of He into the pyrolysis reactor. The outlet of the reactor was connected to a glass condenser that had a continuous flow of refrigerated water (4 °C). Non condensable vapors exiting from the reactor were routed to online-gas flow meter. The reactor and glass condenser were connected by ¼″ stainless steel tubing that was maintained at pyrolysis temperature using heating tape to prevent in-line condensation. ⅛″ stainless steel tubing was used for all other connecting lines.

10 g of lignin was packed in the tubular reactor using quartz wool as a support on either side. Before the start of each experiment, the reactor was purged with He at a flow rate of 500 cm3/min for 15 min to remove the air from the system. Pyrolysis reactor was heated to desired temperature at a ramp rate of 30 °C/min. Volatile pyrolysis products

129 formed were then routed to glass condenser. Condensable pyrolysis products were collected in a glass condenser with water (4 °C) as a heat-transfer fluid. The reactor was then maintained at this desired temperature for 10 min so that the sample completely reacts. At the end of the experiment, the pyrolysis reactor was cooled to room temperature and the solid char as well as bio-oils were weighed and stored in -20 °C refrigerator for subsequent analyses as described below. The gas yield was determined by difference of bio-oil and bio-char yield. All the yields are expressed on the basis of the dry weight of sample.

Results and Discussion

Ultimate & carbohydrate analysis:

Table 1 Ultimate analysis of IL pretreated residual lignin. Mass fraction (wt%) of carbon, hydrogen and nitrogen of IL pretreated lignin were obtained by CHN analyzer. Bone-dry lignin samples were used. Calorific values (HHV) were calculated using equations 1, 2 and 3. C/O and C/H were expressed in mol/mol ratio.

130

The weight fraction of carbon, hydrogen and nitrogen were obtained by CHN elemental analyzer and tabulated in Table 1. Since the ash content of IL pretreated residual lignin determined was 0.5 wt%, oxygen weight fraction was calculated by difference. Ultimate analysis values obtained are consistent with other types of lignin although their biomass sources, pretreatment and isolation techniques are different [156,

173, 176, 177]. Carbohydrate analysis revealed that this IL-pretreated residual lignin has only small fraction (~ 5 wt%) of carbohydrates. These carbohydrates may be impregnated on the residual lignin particles during freeze drying of hydrolysate broth.

Thermo-gravimetric analysis of ILPSF lignin:

Generally, lignin pyrolytically degrades in a wide temperature interval of 160-900

°C due to heavy cross linkages [7]. However, TG profiles of lignin were obtained within

131 the temperature window of 150-600 °C as the main motive of the chapter is to determine or understand its pyrolysis behavior.

Figure 2: Weight loss (dashed lines) and their derivatives (solid line) of IL pretreated poplar lignin at heating rate of (a) 10 °C/min (b) 20 °C/min (c) 30 °C/min. IL pretreated poplar lignin degrade in four stages and classified based on its rate of decomposition irrespective of heating rate. TGA and DTG profiles (shown in Figure 2) indicates that volatilization of residual lignin started at 150 °C in inert atmosphere. It was observed that lignin degrades in four stages classified on the basis of its rate of degradation. The temperature intervals corresponds to each stage are independent of heating rates of TG experiments. In the first stage, small amount of mass loss was observed which takes place in the temperature range of 150-230 °C due to the degradation of easily-volatilizable fraction of lignin. In the second stage, about 40 wt% of mass loss was observed in the temperature range of

132

230-340 °C and also reached maximum rate of decomposition at a temperature of 300 °C.

In the third stage, weight loss observed in the temperature window of 340-550 °C could be referred to the volatilization of lignin fraction that may have higher thermal stability than those degraded in second stage. Finally, the weight loss observed in the temperature above 550 °C could be attributed to gasification of both high molecular weight portion of lignin and bio-char produced during previous three stages. Moreover, it could be concluded that optimum temperature for lignin pyrolysis is 550 °C to achieve optimum carbon conversion which results in high yields of condensable volatile fraction.

DSC profile of ILPSF lignin:

The DSC profile of IL pretreated residual lignin is shown in Figure 3. Although, physical properties, especially glass transition temperature, of lignin primarily depends upon its isolation techniques and biomass source, pure amorphous lignin, in general, exhibits glass transition temperature at 108 °C [178]. For IL pretreated residual lignin, we observed a sharp endothermic peak in DSC profile at 99 °C which could be glass transition temperature. Heat flow into lignin remains constant upto 250 °C. However, it increased rapidly with increase in temperature from 250 °C to 500 °C due to formation of higher heat capacity product, bio-char, during pyrolysis.

133

Figure 3: DSC profile of IL pretreated poplar lignin and depicts the outset temperature of pyrolysis reaction at 250 °C. It was also observed from TG experiments that lignin degradation rate increased rapidly from 250 °C and most of the weight loss volatilized in the temperature range of

250-500 °C (which corresponds to the second and third stage of lignin degradation; see

Figure 2(a)(b)&(c)). Hence it could be concluded from both TGA and DSC profiles that

250 °C was the onset temperature of lignin pyrolysis reactions. Thermal degradation temperature of IL pretreated lignin started at 250 °C and which is consistent with the literature value of 270 °C [178].

Evaluation of activation energy by KAS method:

A thorough knowledge of thermal behavior and activation energies of lignin pyrolysis could help in the design and operation of the robust large-scale pyrolysis system that would be integrated with bio-refinery. TGA data obtained at different heating

134 rates was used to calculate the activation energy of pyrolysis of IL pretreated residual lignin. The rate of pyrolysis of lignin under non-isothermal conditions can be summarized by the general equation [156]:

푑훼 푟 = = 푘 푓(훼) (4) 푑푡

Here, k is the rate constant of the reaction, n is the order of the reaction and α is the relative weight loss of the sample and defined as follows [160]

(푚 − 푚) 훼 = 표 (5) (푚표− 푚∞)

Where m refers to the weight of the sample, m0 and m∞ represents the initial and final mass of the sample respectively. The temperature dependence of the rate constant can be assumed to follow the Arrhenius equation:

−퐸푎/푅푇 푘 = 푘표푒 (6)

Where T is the absolute temperature, ko is the pre-exponential factor, Ea is the activation energy and R is the universal gas constant. The heating rate was assumed to be constant and linear. Hence, the rate of the reaction can be expressed follows:

푑훼 푑푇 푑훼 푟 = = 훽 (7) 푑푇 푑푡 푑푇

The rate of the reaction can be expressed as integral equation as follows:

훼 푑훼 푘 푇 푘 푇 퐺(훼) = ∫ = 0 ∫ 푒−퐸푎/푅푇 푑푇 ≈ 0 ∫ 푒−퐸푎/푅푇 푑푇 = 0 푓(훼) 훽 푇표 훽 표

푘 퐸 ∞ 푘 퐸 0 푎 ∫ 푢−2 푒−푢푑푢 = 0 푎 푃(푥) (8) 훽푅 푥 훽푅

135

Where x = E/RT. By introducing the approximation [157],

푃 = 푥−2푒−x (9)

into equ (5) the relation between heating rate and inverse temperature becomes:

훽 푘표푅 퐸푎 ln [ 2] = 푙푛 [ ] − (10) 푇 퐸 푎퐺(훼) 푅푇

The activation energy of lignin volatilization for a specified value of conversion

훽 1 was estimated from the slope of the plot ln [ ] vs (see Figure 4). 푇2 푇

Figure 4: Estimation of activation energy of IL pretreated lignin using KAS method at different degree of conversion (indicated as “a”).

136

Activation energies calculated from the slope of are tabulated in Table 2 and these values are consistent with other types of lignin studied in the literature [160, 170].

Table 2: Activation energies of pyrolysis of IL pretreated lignin calculated at different conversions by KAS model. These values would help in the design and operation of the robust large-scale pyrolysis systems.

Pyrolysis experiments:

Residence time of volatile pyrolysis products in the reactor, rapid quenching of hot volatile products and heating rate of the furnace are the main parameters that would determine the relative yields of pyrolysis products [45]. Based on the dimensions of pyrolysis reactor and gas flow rates used (both He and Ar) in our studies, the vapor residence time (reactor to condenser) was calculated to be 20 s. Heating rate of the

137 reactor or split-shell furnace was also calculated as 30 °C/min by monitoring output temperature recorded by thermocouple inserted into pyrolysis reactor over time. Thus, from the above calculated parameters, lignin pyrolysis experiments performed could be considered as slow-pyrolysis. The product yields from lignin pyrolysis obtained in the present study are similar to those previously observed for other types of lignin with fast- pyrolysis conditions [173, 174]. This may be due to inherent property of lignin, in general, towards pyrolytic degradation.

The yields of bio-oils, bio-char and gases on dry basis of lignin used during pyrolysis are depicted in Figure 5.

Figure 5: Yield of bio-char, bio-oil and gases obtained by lignin pyrolysis. Bio- char yields were calculated by calculating the fraction solid-residue (bio-char) collected at the end of the experiment. Bio-oils yields were obtained by measuring the fraction of liquid product collected by glass condenser. Gases yields were calculated by difference.

138

It was observed that lignin bio-oils collected clearly have two phases, aqueous phase (light oils) and organic phase (heavy oils). These two phases were separated using a separating funnel and stored at -20 °C for further compositional and elemental analysis.

Elemental analysis of bio-oils and bio-char obtained:

The ultimate analysis of lignin pyrolysis products are tabulated in Table 3.

Table 3: Mass fraction (wt%) of carbon, hydrogen and nitrogen of lignin pyrolysis products were obtained by CHN analyzer. Mass fraction of oxygen was calculated by difference. Calorific values (HHV) were calculated using equations 1, 2 and 3. C/O and C/H are expressed in mol/mol ratio.

Since the ash content of IL pretreated residual lignin determined was 0.5 wt%, oxygen weight fraction of bio-char and bio-oils was calculated by difference. It is interesting to note that ultimate analysis values of lignin and heavy oils are similar to each other.

Hence, it could be inferred that quality of pyrolysis liquid product would primarily 139 depend upon its feedstock. The low carbon fraction and high hydrogen and oxygen fraction observed in light oils indicates the presence of large amount of water. The water in light oils was produced either by dehydration reactions during pyrolysis or by vaporization and then subsequent condensation of bound moisture associated to residual lignin itself.

Table 4: Chemical compounds identified in heavy oils produced from IL pretreated poplar lignin by GC-MS equipped with NIST98 library.

140

The calorific value (HHV value) of bio-char obtained was higher than that of lignin itself. Moreover, bio-char obtained has very small fraction of ash or inorganic elements. Hence, energy-dense, ash-free bio-char could be fed to a conventional coal-

141 gasifier to produce syn-gas (CO+H2). Therefore, bio-char obtained by lignin pyrolysis could be referred as “green coal.”

Characterization of bio-oils using GC-MS and 13C-NMR:

The GC-MS chromatograms obtained for lignin pyrolysis oils. All the compounds identified by NIST98 library are tabulated in Table 4 and Table 5: Chemical compounds identified in light oils produced from IL pretreated poplar lignin by GC-MS equipped with NIST98 library. 5.

Table 5: Chemical compounds identified in light oils produced from IL pretreated poplar lignin by GC-MS equipped with NIST98 library.

As expected, upon lignin de-polymerization via pyrolysis, liquid product, both heavy and light oils, collected has aromatic compounds such as toluene, phenol and alkyl-, methoxy- and carbonyl- substituted phenolic compounds – 2-methyl phenol, guaiacol or

2-methoxyphenol, syringol or 2,6-dimethoxy phenol, m-eugenol or 3-allyl-6- methoxyphenol, isovanillic acid or 3-hydroxy-4-methoxybenzoic acid, etc. In addition to these phenolic compounds, we observed small amount of polycyclic aromatic hydrocarbons (PAHs) in heavy oils such as anthracene, pyrene, benzo[ghi]perylene,

142 benzo[k]fluroanthene, etc formed via the polymerization reactions during lignin pyrolysis. Light oils also contain some sugar degradation compounds - acetic acid, 3- furanmethanol, 1-hydroxy-2-butanone formed during pyrolysis of sugars monomers units impregnated on lignin particles.

A typical 13C NMR spectrum for heavy oil and chemical shifts corresponds to different carbon functional groups depicted in Figure 6.

Figure 6: 13C-NMR spectra for the heavy oils produced by IL pretreated poplar lignin and depict chemical shifts of different carbon functional groups fingerprints based on literature. C=O linkage observed in the spectra was due to the presence of aromatic carbonyl acid (i.e from GC-MS analysis 3-Hydroxy-4-methoxybenzoic acid). Two peaks observed in chemical shift range of 55-60 ppm (one at 55.6-56 and another at 56-56.6 ppm) indicate the predominance of aromatic-methoxyl bond which justifies our GC-MS characterization of heavy oils. Aliphatic C-C fingerprint observed in the NMR spectra could be due to presence of methyl- substitute of phenolic compounds. Therefore, from

143

GC-MS and 13C-NMR analysis, bio-oils retained most of functional groups such as methoxy- and carbonyl- substitutes of phenol present in lignin.

From GC-MS analysis, two major peaks at the retention time of 7.5 and 15 min were identified as 2-methoxy phenol and 2,6-dimethoxy phenol, respectively. Calibration curves for these two compounds were obtained using GC-FID. It was calculated that the weight fraction of 2-methoxy phenol and 2,6-dimethoxy phenol in heavy oils is 2.5 wt% and 10 wt%, respectively.

Thermo-gravimetric analysis of lignin pyrolysis oils:

TG and DTG profiles of light oils and heavy oils in nitrogen atmosphere

(shown in Figure 7) were obtained to estimate their boiling point distribution.

144

Figure 7: Residual weight fraction (dashed line) and derivative weight loss (solid lines) profiles of (a) light oil and (b) heavy oil obtained by lignin pyrolysis. TG analysis of light oils indicates the presence of high fraction of water. Lignin heavy oil corresponds to the naphtha fraction of crude oil with respective to TG analysis. These studies performed for heavy oils indicates that [65] [153] -

(i) 39 wt% volatilizes below 160 °C (corresponds to the boiling point limit of

light naphtha of petroleum crude oil),

(ii) 29 wt% in the temperature range of 160-200 °C, (corresponds to heavy

naphtha of petroleum crude oil) and

(iii) 23 wt% in the temperature interval of 200-350 °C (corresponds to middle

distillate of petroleum crude oils).

In total, about 68 wt% of aromatic-rich heavy oil fraction (evident from GC-MS and 13C-NMR) has similar boiling point distribution to naphtha fraction of crude oil.

Therefore, heavy oils could be directly fed to a traditional catalytic reforming unit used in petroleum refinery to produce liquid hydrocarbon. However, other parameters must be studied to validate this comparison. It was also observed that about 80 wt% of heavy oil volatilizes below 270 °C which indicates that only small fraction of heavy oils (≈20 wt%) could not be analyzed by GC-MS.

We speculated from the CHN analysis that light oils mostly consists of water. TG and DTG profiles of light oil supported our speculation since 90% of weight loss was observed below 100 °C. Therefore, it could be concluded that light oils consists of water, sugar degradation products, if any and light aromatic compounds.

145

3.50E+09 2,6-dimethoxy- phenol 3.00E+09 Ethanone, 1-(2-hydroxy-5- 2.50E+09 methylphenyl) 1,2,4-trimethoxy- benzene 1,2,3-trimethoxy-4-methyl- 2.00E+09 2-methoxy-4-methyl- phenol benzene 2-methoxyphenol Ethanone, 1-(3,4-

1.50E+09 dimethoxyphenyl) Intensity 1.00E+09 5.00E+08 0.00E+00 0 5 10 15 20 25 Retention time, min

Figure 8: GC-MS chromatogram of bio-oil obtained from pyrolysis of residual poplar lignin. Chemical compounds identified by GC-MS NIST08 library embedded in the chem-station software are depicted.

Figure 9: GC-MS chromatogram of upgraded bio-oil obtained from pyrolysis followed by deoxygenation of residual poplar lignin using (a) HZM-5 and (b) Y. Chemical compounds identified by GC-MS NIST08 library embedded in the chem- station software are depicted.

146

Conclusion:

Thermal characteristics (TG, DSC profiles and calculated activation energies) of

IL pretreated residual lignin revealed its similarity with other type of commercial available lignins produced by different techniques. Hence, upon commercialization of IL- based bio-refinery, this residual lignin itself could serve as a valuable by-product. GC-

MS and 13C-NMR analysis of lignin pyrolysis oils indicates the predominance of oxygenated aromatic compounds.

147

Appendix B

B. Pyrolytic fractionation of triglyceride-lean Chlorella sp.

1 g Pyrolysis at 0.54 g Pyrolysis at 0.40 g (Feed) 300-320 °C (Triglyceride- 420-450 °C (Biochar) rich solid)

0.23 g 0.086 g (Biooil) (Triglyceride- specific biooil)

Figure 1: Overall mass balance obtained for pyrolytic fractionation of non-oleaginous Chlorella sp.. The carrier gas flow rate of these experiments is 1000 mL/min.

0.369 g Pyrolysis at 0.245 g Pyrolysis at 0.158 g (Feed) 300-320 °C (Triglyceride- 420-450 °C (Biochar) rich solid)

0.077(org) + 0.012(aq) g 0.056 g (Biooil) (Triglyceride-specific biooil)

Figure 2: Elemental carbon mass balance determined for pyrolytic fractionation of non- oleaginous Chlorella sp.. The carrier gas flow rate of these experiments is 1000 ml/min.

148

0.06 g Pyrolysis at 0.037 g Pyrolysis at 0.023 g (Feed) 300-320 °C (Triglyceride- 420-450 °C (Biochar) rich solid)

0.0027(aq) + 0.01123 (org) g 0.0075 g (Triglyceride-specific biooil) (Biooil)

Figure 3: Elemental nitrogen mass balance determined for pyrolytic fractionation of non- oleaginous Chlorella sp.. The carrier gas flow rate of these experiments is 1000 ml/min.

0.05 g Pyrolysis at Pyrolysis at 0 g (Feed) 300-320 °C 420-450 °C (Biochar)

0.025 g 0 g (Biooil) (Triglyceride-specific biooil)

Figure 4: Lipid balance determined for pyrolytic fractionation of non-oleaginous Chlorella sp.. The carrier gas flow rate of these experiments is 1000 ml/min.

Palmitic acid

C 18 fatty acids

Mono-glycerides

Figure 5: GC-FID chromatogram of aqueous fraction of bio-oil collected during non- oleaginous algae pyrolysis at 320 °C. The chemical compounds present in this bio-oil fraction were identified using GC-MS. The carrier gas flow rate of these experiments is 1000 ml/min.

149

Palmitic acid

C 18 fatty acids

Mono-glycerides

Figure 6: GC-FID chromatogram of organic fraction of bio-oil collected during non- oleaginous algae pyrolysis at 320 °C. The chemical compounds present in this bio-oil fraction were identified using GC-MS. The carrier gas flow rate of these experiments is 1000 ml/min.

Figure 7: GC-FID chromatogram of bio-oil collected during non-oleaginous algae pyrolysis at 420 °C. The chemical compounds present in this bio-oil fraction were identified using GC-MS. The carrier gas flow rate of these experiments is 1000 ml/min.

150

Table 1: Elemental analysis of bio-oil and solid residues collected during pyrolytic fractionation of triglyceride-lean Chlorella sp. sample.

Solid residue Biochar collected obtained after after 2-step Biooil pyrolysis of non- pyrolytic collected at oleaginous fractionation of 420 °C Chlorella sp. at non-oleaginous 320 °C Chlorella sp. C 45.35 ± 0.11 38.95 ± 0.02 65.40 ± 0.18 H 4.37 ± 0.01 2.90 ± 0.01 8.91 ± 0.14 N 6.85 ± 0.01 5.65 ± 0.04 8.66 ± 0.16 O 43.43 ± 0.11 52.5 ± 0.05 17.03 ± 0.20 HHV 18.7 ± 0.06 17.0 ± 0.23 31.76 ± 0.42 (MJ/kg)

151

Appendix C

C. Co-pyrolysis of triglyceride-lean Chlorella sp. and

tristearate performed at 320 °C and then at 420 °C

152

1000 mg tristearate + Pyrolysis at Pyrolysis at 185 mg of membrane 300-320 °C 420-430 °C lipids (Feed)

92.4 mg 731 mg (Biooil) (Triglyceride-specific biooil)

Figure 1: Triglyceride balance determined during co-pyrolysis of non-oleaginous algae and tristearate at 320 and 420 °C in a fixed-bed pyrolysis reactor. The carrier gas flow rate of these experiments is 100 ml/min.

1.20E+11

1.00E+11

8.00E+10

6.00E+10

4.00E+10 Abundance 2.00E+10

0.00E+00 0 10 20 30 40 Time, min

Figure 2: GC-MS chromatogram of bio-oils recovered during co-pyrolysis of non- oleaginous algae and tristearate at 320 °C. The carrier gas flow rate of these experiments is 100 ml/min.

153

Nitrile of Stearic acid Stearic acid 1.40E+11 Stearic acid, 2-propenyl ester 1.20E+11 Amide of Hydrocarbons 1.00E+11 Stearic acid 8.00E+10 6.00E+10

4.00E+10 Abundance 2.00E+10 0.00E+00 0 10 20 30 40 Time, min

Figure 3: GC-MS chromatogram of triglyceride-specific bio-oils from feed which is a mixture of non-oleaginous algae and tristearate. The carrier gas flow rate of these experiments is 100 ml/min.

Stearic acids

Palmitic acid

Mono-glycerides

Figure 4: GC-FID chromatogram bio-oil collected during pyrolysis of tristearate-non- oleaginous algae mixture at 320 °C. The chemical compounds present in this bio-oil

154 fraction were identified using GC-MS. The carrier gas flow rate of these experiments is 100 ml/min.

Nitrile of stearic acids

Hydrocarbons and short chain fatty Stearic acid acids

Di-stearate Tristearate

Figure 5: GC-FID chromatogram bio-oil collected during pyrolysis of tristearate-non- oleaginous algae mixture at 420 °C. The chemical compounds present in this bio-oil fraction were identified using GC-MS. The carrier gas flow rate of these experiments is 100 ml/min.

C6-C14 fatty acids and C8-C17 hydrocarbons, 16% Fatty amides and nitriles , 22%

Monoglycerides, diglycerides and triglycerides, 10%

Fatty acids, 52%

Figure 6: Pie-chart indicates the composition of triglyceride-specific bio-oils collected during pyrolytic fractionation of feed which is a mixture of non-oleaginous algae and tristearate. Values indicate the weight fraction of total triglyceride-specific bio-oils.

155

Appendix D

D. Supplementary data

2.5

C) °

1.5

Organosolv Lignin 1 Cellulose

0.5

0 Derivative loss(%wt/ weightDerivative 0 200 400 600 Temperature (°C)

Figure 1: DTG profile of organosolv lignin and cellulose.

156

100 3 320 °C 80 2.5 420 °C 2

60 C) ° Soy oil 1.5 Corn starch 40 Starch 1 BSA BSA (wt%/ Soy oil

20 0.5 Residual weight%

0 Derivative weightloss 0 0 100 200 300 400 500 600 0 100 200 300 400 500 600 Temperature (°C) Temperature (°C)

(a) (b)

Figure 2:TG profiles that simulate pyrolysis of bovine serum albumin (BSA), corn starch and soy oil.

80 70 60 50

Yield % 40 %Bio-char(ash free basis) 30 %Bio-oil(ash free basis) 20 %Gases (ash free basis) 10 0

Figure 3: Pyrolysis product yields of biomass feedstocks. Final temperature of the pyrolysis experiments is 600 °C. Hot vapor residence time of these experiments is 33.3s. † indicates the final temperature of the experiment is 550 °C. ‡ indicates the final temperature of the experiment is 430 °C. †,‡ indicates the hot vapor residence time is 20s.

157

Table 1: Summary of bio-oil obtained from different biomass feedstocks.

Chemical compounds Source present in Corn Wood Rice Lyngbya Cladophor Soy Lignin Biopolymer pyrolysis oils of cobs chips husk sp. a sp. flakes corresponding substrates Aldehydes, Carbohydrate ketones and X X X - X X X alcohols Phenolic Lignin X X X X - - - compounds Protein N-compounds - - - - X X X Lipids Fatty acids ------X

“X” marks indicate the presence of these compounds in the bio-oil from the corresponding feedstock. Blanks indicate that these compound was not identified in the bio-oil from the corresponding feedstock.

Triglyceride N2 flow

Cups Trays (occurs in fixed bed reactor)

Figure 4: Design of alumina crucibles used in thermos-gravimetry.

100

Heating 60 from 25 to Isothermal at 320 °C 320 °C 100 mL/min in trays 40 100 mL/min in cups

Residual weight% 20

0 0 10 20 30 Time, min

Figure 5: TG profile of soy oil performed with different crucibles.

158

100

60 Heating 900 mL/min in Isothermal 40 from 25 to cups at 320 °C 320 °C 100 mL/min in

20 cups Residual weight% 0 0 10 20 30 Time, min

Figure 6: TG profile of soy oil performed at different N2 flow rates.

100

60 Heating 900 mL/min in Isothermal 40 from 25 to cups at 320 °C 320 °C 900 mL/min in

20 trays Residual Residual weight% 0 0 10 20 30 Time, min

Figure 7: TG profile of soy oil performed at flow rate of 900 ml/min with different crucibles.

159

Protein and carbohydrate Triglyceride degradation degradation products products

100 Isothermal Isothermal 80 at 320 °C at 420 °C 321 to 420 °C 60 230 to 320 °C 40

20 Residual weight% 0 0 5 10 15 20 25 30 Time (min)

Figure 8: TG profile that simulates pyrolytic fractionation of oleaginous Chlorella sp. to estimate solid residence time in the reactor.

• Basis: 1 kg of dry algae mixed with 4 kg of water (i.e. 20% w/w slurry) initially at 20 °C and 1 atm. Items Energy consumed in KJ/(kg of dry algae) Drying 10,400 Energy consumed in the reactor 680 Total energy consumed 11,080

Energy that can be recovered from -5000 combustion of char Net energy required 6080 • Energy content of triglyceride-specific biooil from 1 kg of dry algae = (0.20 kg × 41,000 KJ/kg = 8200 KJ/(kg of dry algae). • Energy content of organic fraction of biooil collected at 320 °C from 1 kg of dry algae = (0.05 kg × 23,000 KJ/kg = 1150 KJ/(kg of dry algae). • Net energy gain = energy in the biooils – energy required for this process = (8200 +1150 )KJ – 6080 KJ = 3270 KJ

Figure 9: Energy required for pyrolytic fractionation.

160

Table 2: Elemental analysis of non-oleaginous and oleaginous Chlorella sp.. The calorific value of the feedstocks was estimated using correlations.

Table 3: Triglyceride and lipid content of oleaginous Chlorella sp. samples.

Sample name Wt% of Wt% of Wt% of FAMEs in Wt% of FAMEs in the triglyceride using lipids using the supernatant solid residue collected bead beating FAME collected after bead after bead beating analysis beating Oleaginous algae 23.05 0.4 27.22 0.2 25.5 0.04 2.0 0.13 batch 1 Oleaginous algae 21.50 0.9 23.8 0.03 21.4 0.4 1.3 0.06 batch 2

161

0.534 g Pyrolysis at 0.407 g Pyrolysis at 0.135 g (Feed) 300-320 °C (Triglyceride- 420-450 °C (Biochar) rich solid)

0.0047(aq.) g + 0.149 g 0.026 (or.)g (Triglyceride- (Biooil) specific biooil)

Figure 10: C balance obtained for pyrolytic fractionation of oleaginous Chlorella sp.. The carrier gas flow rate of these experiments is 1000 ml/min.

Figure 11: GC-FID chromatogram of aqueous fraction of bio-oil collected during oleaginous algae pyrolysis at 320 °C. The chemical compounds present in this bio-oil fraction were identified using GC-MS. The carrier gas flow rate of these experiments is 1000 ml/min.

162

Oleic acid

Palmitic acid

Mono-glycerides Di-glycerides

Figure 12: GC-FID chromatogram of organic fraction of bio-oil collected during oleaginous algae pyrolysis at 320 °C. The chemical compounds present in this bio-oil fraction were identified using GC-MS. The carrier gas flow rate of these experiments is 1000 ml/min.

Monoglyceri Fatty amides C6-C14 fatty des, and nitriles, acids and C8- diglycerides 3% C17 and hydrocarbons triglycerides, , 11% 5%

C16 fatty C18 fatty acids, 35% acids, 46%

Figure 13: Pie-chart indicates the composition of triglyceride-specific bio-oils collected during pyrolytic fractionation of oleaginous algae a residence time of ~20s. Values indicate the weight fraction of total triglyceride-specific bio-oils.

163

Table 4: Elemental analysis of bio-oil fractions collected during pyrolytic fractionation of oleaginous Chlorella sp. sample.

Biooil collected during 1st step Triglyceride-specific of pyrolytic fractionation of biooil collected from Weight oleaginous algae pyrolytic % Aqueous fractionation of Organic phase phase oleaginous algae C 12.08 ± 0.1 53.66 ± 0.08 74.56 ± 0.07 H 10.57 ± 0.05 7.73 ± 0.03 11.11 ± 0.05 N 0.65 ± 0.01 3.79 ± 0.06 1.87 ± 0.01 O 76.68 ± 0.05 34.81 ± 0.12 12.46 ± 0.1 HHV 5.5 ± 0.1 22.9 ± 0.26 41.0 ± 1.1 (MJ/kg)

Table 5: Enthalpy and sensible heat of triolein as a function of temperature.

Latent heat of Heat capacity at Enthalpy of vaporization Temperature Temperature constant pressure of triolein (kJ/mole) (K) (°C) triolein (kJ/mole) (kJ/mole) 167.5 298.15 25 1.68 165.0 323.15 50 1.72 42.4 162.5 348.15 75 1.76 85.9 159.9 373.15 100 1.80 130.3 157.3 398.15 125 1.84 175.8 154.6 423.15 150 1.88 222.3 151.8 448.15 175 1.92 269.8

164

Table 6: Elemental analysis of solid residues collected during pyrolytic fractionation of oleaginous Chlorella sp. sample.

Solid residue Biochar collected obtained after after 2-step pyrolysis of pyrolytic oleaginous fractionation of Chlorella sp. at oleaginous 320 °C Chlorella sp. C 62.66 ± 0.04 53.86 ± 0.13 H 7.82 ± 0.05 3.73 ± 0.06 N 2.74 ± 0.04 4.20 ± 0.08 O 26.79 ± 0.13 38.22 ± 0.16 HHV 28.0 ± 0.16 21.0 ± 0.31 (MJ/kg)

165

8 600 350 600 7 300 500 500 C)

6 C)

250 400 5 400 4 200 300 3 300 150 200 2 100

200 mW;Endoup 1 KJ/kg Energy, 50 100 0 ( Temperature 100 ( Temperature Differential heatDifferential flow, -1 0 5 10 15 20 25 0 0 0 5 10 15 20 25 -2 0 Time, min Time, min

(a) (b)

Figure 14: (a) DSC profile obtained while simulating pyrolysis of soy, and (b) Cumulative energy calculated from DSC profile during soy pyrolysis.

10 600 450 600 400 8 500 500 C)

350 C)

6 300 400 400 4 250 300 2 300 200 150 200 0 200 100

0 5 10 15 20 25 100 mW;Endo up

-2 ( Temperature Energy, KJ/kg Energy, 50

100 ( Temperature -4 0 0 Differentialheat flow, 0 5 10 15 20 25 -6 0 Time, min Time, min

(a) (b)

Figure 15: (a) DSC profile and (b) Cumulative energy calculated from DSC profile during Chlorella sp. pyrolysis.

738/666 mg Pyrolysis at Pyrolysis at (Lipids/triglycerides 300-320 °C 420-430 °C in feed)

78.9 mg 603 mg (Biooil) (Triglyceride-specific biooil)

Figure 16: FAME balance determined for pyrolytic fractionation oleaginous algae performed in bench-scale fixed-bed reactor at vapor residence time of ~ 20s. .

166

Palmitic acid C 18 fatty acids

Figure 17: GC-FID chromatogram of bio-oil collected during Step 1 of pyrolytic fractionation performed with oleaginous Chlorella sp. algae in bench-scale fixed-bed reactor at vapor residence time of ~ 20s.

Palmitic acid C 18 fatty acids

Hydrocarbons and short chain fatty acids

Mono-glycerides Di-glycerides Triglycerides

Figure 18: GC-FID chromatogram of bio-oil collected during Step 2 of pyrolytic fractionation performed with oleaginous Chlorella sp. algae in bench-scale fixed-bed reactor at vapor residence time of ~ 20s.

167

Table 7: Confidence value of chemical compounds present in bio-oil collected during co- pyrolysis of non-oleaginous Chlorella sp. and tristearate at 320 °C.

Retention Compound Confidence time 4.79 Pentanenitrile, 4-methyl 820 5.22 2-Furanmethanol 663 5.96 2-cyclopenten-1-one, 2-methyl 604 6.11 Pyrazine, 2,5-dimethyl 743 7.07 1-hepten-3-ol 650 7.32 Cylopentanone, 2,4,4-trimethyl 654 7.71 Pyrazine, trimethyl 762 8.11 2-piperidinemethanamine 691 8.57 1H-pyrrole, 1-pentyl 684 9.00 Pyrazine, 2-ethyl-,3,5dimetyl 622 9.28 2-pyrrolidinone 708 12.50 Indole 826 15.02 1-pentadecene 776 15.12 Pentadecane 815 16.26 1-hexadecene 806 16.35 Hexadecane 776 17.33 8-heptadecene 852 17.53 Heptadecane 703 18.00 Heptadecane, 7-methyl 766 20.43 Hexadecanoic acid 796 21.75 Octadecanenitrile 791 21.90 Octadecanoic acid, methyl ester 876 22.10 9,12-octadecadienoic acid 842 22.36 Octadecanoic acid 748 23.25 Octadecanoic acid, 2-propenyl ester 841 24.26 Octadecanamide 746

168

Table 8: Confidence value of chemical compounds present in bio-oil collected during co- pyrolysis of non-oleaginous Chlorella sp and tristearate at 420 °C.

Retention Compound Confidence time 9.79 1-dodecene 878 9.92 Dodecane 883 11.30 1-tridecene 896 11.42 Tridecane 871 12.71 1-tetradecene 880 12.82 Tetradecane 784 14.03 1-pentadecene 843 14.13 Pentadecane 757 14.18 Cyclopentadecane 886 14.32 1-pentadecene 888 15.27 1-hexadecene 680 15.38 Hexadecane 730 16.45 8-heptadecene 788 16.56 Heptadecane 566 16.72 Heptadecanol 879 17.56 1-octadecene 847 17.64 Octadecane 788 18.70 Nonadecane 767 19.32 Hexadecanoic acid 839 20.76 Octadecanenitrile 755 21.35 Octadecanoic acid 752 22.26 Octadecanoic acid, 2-propenyl ester 810 23.25 Octadecaamide 658

169