Glycerin Conversion and its Impact on Biodiesel · The Alkoxylation of Biodiesel and its Impact on Fuel Properties Paul C. Smith Thesis submitted for the degree of Doctor of Philosophy

The Alkoxylation of Biodiesel and its

Impact on Fuel Properties

Paul C. Smith

Thesis submitted for the degree of Doctor of Philosophy

in

School of Chemical Engineering

Faculty of Engineering, Computer and Mathematical Sciences

The University of Adelaide

August 2009

Table of Contents _________________________________________________________________________

i

TABLE OF CONTENTS _________________________________________________________________________

TABLE OF CONTENTS ..................................................................................................... i

LIST OF TABLES ............................................................................................................... v

LIST OF FIGURES ........................................................................................................... vii

ABSTRACT ......................................................................................................................... xi

DECLARATION .............................................................................................................. xiii

ACKNOWLEDGEMENTS ............................................................................................. xiv

LIST OF PUBLICATIONS .............................................................................................. xv

GENERAL INTRODUCTION ........................................................................................... 2

1.1 Context ...................................................................................................................... 2

1.2 Purpose ..................................................................................................................... 2

1.3 Process of Investigation and Thesis Structure ...................................................... 3

BIODIESEL LOW-TEMPERATURE PROPERTIES .................................................... 6

2.1 The Problem ............................................................................................................. 6

2.2 Factors Influencing Biodiesel Cloud Point ............................................................ 7

2.2.1 Fatty Acid Profile ............................................................................................ 7 2.2.2 Alcoholic Adduct ........................................................................................... 12 2.3 Low-Temperature Property Improvement ......................................................... 13

2.3.1 Additives ........................................................................................................ 13 2.3.2 Feedstock Modification ................................................................................ 15 2.3.3 Biodiesel Modification .................................................................................. 16 2.4 Impact on Other Properties of Biodiesel ............................................................. 19

2.5 Proposed Solution .................................................................................................. 20

MATERIALS AND METHODS ...................................................................................... 25

3.1 Materials ................................................................................................................. 25

3.2 Equipment .............................................................................................................. 25

3.2.1 Chemical synthesis ........................................................................................ 25 3.2.2 Analytical Equipment ................................................................................... 26 3.3 Biodiesel Synthesis ................................................................................................. 28

3.3.1 Methyl and Ethyl Biodiesel .......................................................................... 28 3.3.2 Butyl Biodiesel ............................................................................................... 28


ii

3.4 Epoxidation ............................................................................................................. 30

3.5 Alkoxylation ........................................................................................................... 31

3.6 Analytical Methods ................................................................................................ 33

3.6.1 Biodiesel Purity ............................................................................................. 33 3.6.2 Biodiesel Fatty Acid Profile .......................................................................... 33 3.6.3 Epoxy and Alkoxy Biodiesel ......................................................................... 34 3.6.4 Cloud Point .................................................................................................... 35 3.6.5 Pour Point ...................................................................................................... 35 3.6.6 Viscosity ......................................................................................................... 35 3.6.7 Free Fatty Acid Content of Canola Oil ....................................................... 35

INITIAL SYNTHESIS AND CHARACTERISATION ................................................. 37

4.1 Biodiesel Synthesis ................................................................................................. 38

4.1.1 Methyl Biodiesel ............................................................................................ 38 4.1.2 Ethyl Biodiesel ............................................................................................... 43 4.1.3 Butyl Biodiesel ............................................................................................... 44

4.2 Epoxidation of Methyl Biodiesel ........................................................................... 48

4.2.1 24 Hour Epoxidation of Methyl Biodiesel ................................................... 48 4.2.2 GC-FID Analysis ........................................................................................... 49

4.3 Alkoxylation of Epoxy Methyl Biodiesel .............................................................. 51

4.3.1 Procedure ....................................................................................................... 51 4.3.2 GC-FID Analysis ........................................................................................... 52

4.4 Fourier Transform Infrared Analysis.................................................................. 53

4.4.1 Synthesis of Epoxy and Alkoxy Ethyl Oleate ............................................. 54 4.4.2 FTIR Analysis ................................................................................................ 54

4.5 GC-MS Method Development .............................................................................. 58

4.5.1 GC-MS Analysis of Biodiesel ....................................................................... 59 4.5.2 GC-MS Analysis of Epoxy Biodiesel ........................................................... 61 4.5.3 GC-MS Analysis of Alkoxy Biodiesel .......................................................... 63

4.6 Summary ................................................................................................................. 65

PRELIMINARY STUDIES .............................................................................................. 67

5.1 Epoxidation of Methyl Biodiesel ........................................................................... 68

5.1.1 Method ........................................................................................................... 68 5.1.2 Results and Discussion .................................................................................. 70

5.2 Optimisation of the Epoxidation Step .................................................................. 71

5.2.1 Results and Discussion .................................................................................. 71 5.3 Alkoxylation of Epoxy Methyl Biodiesel .............................................................. 73

5.3.1 Method ........................................................................................................... 73 5.3.2 Results and Discussion .................................................................................. 74

5.4 Extension to Ethyl and Butyl Biodiesel ................................................................ 77

5.4.1 Epoxidation .................................................................................................... 77


iii

5.4.2 Alkoxylation ................................................................................................... 78 5.5 Cloud Point Assessment ........................................................................................ 79

5.5.1 Impact of Higher Alkoxy Content ............................................................... 80 5.6 Summary ................................................................................................................. 81

BUTOXYLATION OF BUTYL BIODIESEL ................................................................ 83

6.1 Method .................................................................................................................... 84

6.1.1 Synthesis of Epoxy Butyl Biodiesel .............................................................. 84 6.1.2 Optimisation of Alkoxylation Reaction Conditions ................................... 86

6.2 Results and Discussion........................................................................................... 90

6.2.1 Effect of Temperature .................................................................................. 90 6.2.2 Effect of Catalyst Concentration ................................................................. 91 6.2.3 Effect of Molar Ratio of Alcohol .................................................................. 92 6.2.4 Further Optimisation .................................................................................... 93

6.3 Reaction Kinetics ................................................................................................... 95

6.4 Cloud Point Impact ................................................................................................ 99

6.4.1 Impact of Higher Conversion .................................................................... 100 6.4.2 Impact of Linearisation of Ester ................................................................ 102

6.5 Summary ............................................................................................................... 104

OTHER ADDUCTS OF BUTYL BIODIESEL ............................................................ 106

7.1 Synthesis of Alkoxy Butyl Biodiesel ................................................................... 107

7.1.1 Method ......................................................................................................... 107 7.1.2 Results and Discussion ................................................................................ 110

7.2 Impact on Cloud Point ........................................................................................ 120 7.3 Impact on Pour Point .......................................................................................... 122 7.4 Impact on Viscosity .............................................................................................. 123

7.5 Summary ............................................................................................................... 125

DISCUSSION AND CONCLUSIONS ........................................................................... 127

8.1 Discussion ............................................................................................................. 127

8.2 The Economics of Alkoxylated Biodiesel ........................................................... 132

8.3 Further Work ....................................................................................................... 136

8.4 Conclusions ........................................................................................................... 137

APPENDIX A ................................................................................................................... 140

Chromatograms ............................................................................................................... 140


iv

APPENDIX B ................................................................................................................... 150

Mass Spectra .................................................................................................................... 150

NOMENCLATURE ......................................................................................................... 158

BIBLIOGRAPHY ............................................................................................................ 159

List of Tables _________________________________________________________________________

v

LIST OF TABLES _________________________________________________________________________

Table 2.1: Properties of Fatty acids and Esters. Source: (Knothe 2005) ........................................................... 9 Table 2.2: Properties of Fats/oils and their esters. Source: (Knothe et al. 2004) ............................................ 11 Table 2.3: Fatty Acid Compositions of Naturally Occurring Fats and Oils. ................................................... 12

Table 4.1: Comparison of GC-MS conditions for Wilson et al. (1997) and this work. .................................. 58 Table 4.2: Fatty acid profile of biodiesel derived from canola oil. ................................................................. 59

Table 5.1: Optimisation of epoxy selectivity: reaction conditions. ................................................................. 71 Table 5.2: Selectivity for epoxy methyl biodiesel. .......................................................................................... 72 Table 5.3: Epoxy conversion, selectivity and glycol content for all alkyl esters. ........................................... 77 Table 5.4: Alkoxy selectivity and by-product content for all alkyl esters. ...................................................... 78 Table 5.5: Cloud point for all alkyl esters. ...................................................................................................... 79 Table 6.1: Effect of temperature on selectivity for butoxy butyl biodiesel and by-product content after 6h of reaction time. ................................................................................................................................................... 91 Table 6.2: Effect of catalyst concentration on selectivity for butoxy butyl biodiesel and by-product content after 6h of reaction time. .................................................................................................................................. 92 Table 6.3: Effect of molar ratio of butanol on selectivity for butoxy butyl biodiesel and by-product content after 6h of reaction time. .................................................................................................................................. 93 Table 6.4: Effect of higher temperature and catalyst concentration on selectivity for butoxy butyl biodiesel and by-product content after 6h of reaction time. ............................................................................................ 94 Table 6.5: Specific reaction rate and corresponding cefficient of determination (R2) value for the four experiments at different molar ratios of alcohol. ............................................................................................. 96 Table 6.6: Specific reaction rate and corresponding cefficient of determination (R2) value for the four experiments at different molar ratios of alcohol, excluding the initial 10 minutes. ......................................... 97 Table 6.7: Impurity profile and cloud point results for various butoxy butyl biodiesel batches. .................. 100

Table 7.1: Alkoxylation of butyl biodiesel: retention time of major alkoxy peaks, transesterification fractions before and after purification of the product. .................................................................................................. 112 Table 7.2: Fragmentation pattern for alkoxy butyl biodiesel for ions a to d in Figure 7.4. ........................... 114 Table 7.3: Alkoxy selectivity and by-product content for all alkoxy esters. ................................................. 115 Table 7.4: Transesterified alkoxy oleate content for all alkoxy esters. ......................................................... 120

List of Tables _________________________________________________________________________

vi

Table 7.5: Flow properties of butyl biodiesel and modified biodiesel, including: cloud point, pour point and kinematic viscosity.. ...................................................................................................................................... 121

Table 8.1: Estimates of some critical biodiesel raw material costs. .............................................................. 134

List of Figures _________________________________________________________________________

vii

LIST OF FIGURES _________________________________________________________________________

Figure 1.1: Flow chart of the process of investigation. ..................................................................................... 4 Figure 3.1: Kettle reactor with stirrer/hotplate for synthesis .......................................................................... 26 Figure 3.2: Test apparatus for cloud point determination ............................................................................... 27 Figure 3.3: Reaction scheme for the epoxidation of biodiesel (methyl) including the main product and possible by-products: a – methyl oleate, b – 9,10-epoxy methyl stearate, c – 9,10-dihydroxy methyl stearate, d – 9(10)-keto methyl stearate. ........................................................................................................................ 31 Figure 3.4: Reaction scheme for the alkoxylation of epoxy biodiesel (methyl) including the main product and possible by-products: b – 9,10-epoxy methyl stearate, c – 9,10-dihydroxy methyl stearate, d – 9(10)-keto methyl stearate, e – 9(10)-hydroxy,10(9)-methoxy methyl stearate. ............................................................... 32

Figure 4.1: Chromatogram of standard solutions for contaminants associated with EN 14105. .................... 39 Figure 4.2: GC-FID Chromatogram of methyl biodiesel with high monoglyceride. ...................................... 40 Figure 4.3: Transesterification reaction mechanism. ...................................................................................... 41 Figure 4.4: GC-FID Chromatogram of methyl biodiesel with low monoglyceride. ....................................... 42 Figure 4.5: Chromatogram of butyl biodiesel with high mono- and diglycerides. ......................................... 45 Figure 4.6: Chromatogram of butyl biodiesel with low monoglyceride. ........................................................ 47 Figure 4.7: GC-FID chromatogram of methyl biodiesel analysed according to EN 14103. ........................... 49 Figure 4.8: GC-FID chromatogram of methyl biodiesel epoxidised for 24h analysed according to EN 14103. ......................................................................................................................................................................... 50 Figure 4.9: Conversion of methyl biodiesel to epoxy methyl biodiesel over time.......................................... 50 Figure 4.10: GC-FID chromatogram of methoxy methyl biodiesel alkoxylated for 24h, analysed according to EN 14103. ........................................................................................................................................................ 52 Figure 4.11: GC-FID chromatogram of methoxy methyl biodiesel alkoxylated for 24h, analysed according to EN 14105. ........................................................................................................................................................ 53 Figure 4. 12: Overlay of FTIR spectra of epoxy ethyl oleate, ethoxy ethyl oleate and glycol ethyl oleate. ... 55 Figure 4.13: Overlay of FTIR spectra of epoxy butyl biodiesel and butoxy butyl biodiesel. ......................... 56 Figure 4.14: Residue spectrum of the subtraction result of butoxy butyl biodiesel produced 22/7/08 and butoxy butyl biodiesel produced 7/7/08. .......................................................................................................... 57 Figure 4.15: Residue spectrum of the subtraction result of butoxy butyl biodiesel produced 22/7/08 and butoxy butyl biodiesel produced 7/7/08. .......................................................................................................... 57 Figure 4.16: Chromatogram of methyl biodiesel generated on GC-MS. ........................................................ 60 Figure 4.17: Chromatogram of methyl biodiesel generated on GC-MS showing the C18 peaks. The C18:1/3 peaks are those labelled 22.17 and 22.24 minutes. .......................................................................................... 60

List of Figures _________________________________________________________________________

viii

Figure 4.18: Chromatogram of epoxy methyl biodiesel. Major peaks: methyl palmitate (16.31 min.), methyl oleate (22.07 min.), 9,10-epoxy oleate (28.90 min.), 9,10-dihydroxy methyl stearate (35.88 min.). .............. 61 Figure 4.19: Mass spectra of methyl biodiesel derivatives: 9,10-epoxy methyl stearate. ............................... 62 Figure 4.20: Mass spectra of methyl biodiesel derivatives: 9,10-dihydroxy methyl stearate. ........................ 62 Figure 4.21: Chromatogram of methoxy methyl biodiesel. Major peaks: methyl palmitate (16.33 min.), methyl oleate/linolenate (22.08 - 22.16 min.), 9(10)-keto methyl stearate (29.09 min.), 9(10)-hydroxy,10(9)-methoxy stearate (33.30 min.), 9,10-dihydroxy methyl stearate (35.88 min.). ................................................ 63 Figure 4.22: Mass spectrum of 9(10)-hydroxy,10(9)-methoxy methyl stearate. ............................................ 64 Figure 4.23: Mass spectrum of 9(10)-keto methyl stearate. ............................................................................ 64

Figure 5.1: Conversion of the unsaturated portion of methyl biodiesel to epoxy methyl biodiesel. Reaction conditions: molar ratio of 0.5 and 2 for formic acid and hydrogen peroxide to biodiesel, respectively; temperature of 60°C. Selectivity for epoxy methyl biodiesel (inset). .............................................................. 70 Figure 5.2: Graphical representation of the 5 trial batches for the optimisation of the epoxidation of methyl biodiesel. .......................................................................................................................................................... 72 Figure 5.3: Selectivity for methoxy biodiesel and fractions of by-product (glycol and ketone), including fractionated samples. ....................................................................................................................................... 74 Figure 5.4: Chromatogram of methoxy methyl biodiesel. Major peaks: methyl palmitate (16.33 min.), methyl oleate/linolenate (22.08 - 22.16 min.), 9(10)-keto methyl stearate (29.09 min.), 9(10)-hydroxy,10(9)-methoxy stearate (33.30 min.), 9,10-dihydroxy methyl stearate (35.88 min.). ................................................ 75 Figure 5.5: Methoxy methyl biodiesel at room temperature with precipitate. ................................................ 76 Figure 5.6: Chromatogram of methoxy methyl biodiesel: (I) supernatant, (II) precipitate. ............................ 76 Figure 6.1: Chromatogram of butyl biodiesel. ................................................................................................ 84 Figure 6.2: Chromatogram of epoxy butyl biodiesel. Main peaks: butyl oleate/linolenate (31.38-31.56 min.), epoxy butyl stearate (38.45 min.). ................................................................................................................... 86 Figure 6.3: Chromatogram of epoxy butyl biodiesel 20 min. in to the butoxylation. Main peaks: butyl oleate/linolenate (31.36 - 31.56 min.), epoxy butyl stearate (38.39 min.), butoxy butyl stearate (49.02 min.). ......................................................................................................................................................................... 87 Figure 6.4: Chromatogram of butoxy butyl biodiesel at completion of butoxylation showing the absence of the epoxy butyl stearate peak but the presence of the keto butyl stearate/butyl eicosenoate (C20:1) at 38.73 min. .................................................................................................................................................................. 88 Figure 6.5: Mass spectrum of peak at 49.0 min. identified as 9-butoxy,10-hydroxy butyl stearate. .............. 89 Figure 6.6: Mass spectrum of peak at 38.8 min. identified as 9(10)-keto butyl stearate. ............................... 89 Figure 6.7: Effect of temperature on conversion of epoxy butyl biodiesel. .................................................... 90 Figure 6.8: Effect of catalyst concentration on conversion of epoxy butyl biodiesel. .................................... 91 Figure 6.9: Effect of molar ratio of butanol on conversion of epoxy butyl biodiesel. .................................... 93 Figure 6.10: Results of further optimisation work for the reaction conditions for the epoxidation of epoxy butyl biodiesel. ................................................................................................................................................. 94

List of Figures _________________________________________________________________________

ix

Figure 6.11: Kinetic plots for the various molar ratios of alcohol: 5:1 (a), 10:1 (b), 20:1 (c), 40:1 (d). ......... 95 Figure 6.12: Specific reaction rate versus molar ratio of alcohol. .................................................................. 97 Figure 6.13: Plot of Ln k versus 1/T for the 3 temperatures of 40ºC, 60 ºC and 80ºC. .................................. 98 Figure 6.14: Cloud point of butoxy butyl biodiesel from 2 wt% to 74 wt%. ................................................ 101

Figure 7.1: Reaction scheme for the case of ethylhexoxy butyl biodiesel: a – butyl oleate; b – epoxy butyl oleate ; c – ethylhexoxy butyl oleate; d – ethylhexoxy ethylhexyl oleate. ..................................................... 109 Figure 7.2: Chromatogram of epoxy butyl biodiesel. Major peaks: I - C16 butyl biodiesel; II - C18 butyl biodiesel (31.0-32.3 min.) – oleate/linolenate fraction is 31.5-31.6 min.; III - epoxy butyl biodiesel (37.5-38.4 min.). .............................................................................................................................................................. 111 Figure 7.3: Chromatogram of octoxy butyl biodiesel. Major peaks: I - C18 octyl biodiesel (45.1-45.6 min.); II - octoxy butyl biodiesel (64.2-66.1 min.). .................................................................................................. 112 Figure 7.4: Schematic of the molecular fragmentation pattern of the alkoxy butyl oleates. ......................... 113 Figure 7.5: Octoxy butyl oleate purified without dichloromethane with particulates. .................................. 117 Figure 7.6: Ethylhexoxy butyl biodiesel purified without dichloromethane (left) and with dichloromethane (right). ............................................................................................................................................................ 117 Figure 7.7: Chromatogram of methoxy butyl biodiesel. Major peaks are: methyl oleate/linolenate (22.00 min.), butyl oleate/linolenate (31.45 min.), methoxy methyl oleate/linolenate (33.29 min.) and methoxy butyl oleate/linolenate (42.72 min.). ....................................................................................................................... 118 Figure 7.8: Mass spectrum of ethoxy ethyl oleate. ....................................................................................... 118 Figure 7.9: Mass spectrum of ethoxy butyl oleate. ....................................................................................... 119

Figure 8.1: Conceptual process flow for ethylhexoxy butyl biodiesel production........................................ 133

Figure A.1: Chromatogram of methoxy butyl biodiesel. Major peaks are: methyl oleate/linolenate (22.00 min.), butyl oleate/linolenate (31.45 min.), methoxy methyl oleate/linolenate (33.29 min.) and methoxy butyl oleate/linolenate (42.72 min.). ....................................................................................................................... 140 Figure A.2: Chromatogram of ethoxy butyl biodiesel. Major peaks are: ethyl oleate/linolenate (24.24 min.), butyl oleate/linolenate (31.51 min.), ethoxy ethyl oleate/linolenate (36.61 min.) and ethoxy butyl oleate/linolenate (43.63 min.). ....................................................................................................................... 141 Figure A.3: Chromatogram of propoxy butyl biodiesel. Major peaks are: propyl oleate/linolenate (26.9 min.), butyl oleate/linolenate (31.51 min.) and propoxy butyl oleate/linolenate (46.30 min.). ................................ 142 Figure A.4: Chromatogram of butoxy butyl biodiesel. Major peaks are: butyl oleate/linolenate (31.52 min.) and butoxy butyl oleate/linolenate (49.25 min.). ........................................................................................... 143 Figure A.5: Chromatogram of tert-butoxy butyl biodiesel. Major peaks are: butyl oleate/linolenate (31.51 min.), keto butyl oleate (38.81 min.) and dihydroxy butyl oleate/linolenate (44 - 46 min.). ......................... 144 Figure A.6: Chromatogram of pentoxy butyl biodiesel. Major peaks are: pentyl oleate/linolenate (34.84 min.), butyl oleate/linolenate (31.55 min.) and pentoxy butyl oleate/linolenate (52.28 min.). ...................... 145

List of Figures _________________________________________________________________________

x

Figure A.7: Chromatogram of hexoxy butyl biodiesel. Major peaks are: hexyl oleate/linolenate (38.41 min.), butyl oleate/linolenate (31.51 min.) and hexoxy butyl oleate/linolenate (55.98 min.)................................... 146 Figure A.8: Chromatogram of octoxy butyl biodiesel. Major peaks are: octyl oleate/linolenate (45.34 min.), butyl oleate/linolenate (31.46 min.) and octoxy butyl oleate/linolenate (66.21 min.). .................................. 147 Figure A.9: Chromatogram of ethylhexoxy butyl biodiesel. Major peaks are: ethylhexyl oleate/linolenate (42.67 min.), butyl oleate/linolenate (31.56 min.) and ethylhexoxy butyl oleate/linolenate (59.91 min.). .... 148

Figure B.1: Mass spectrum of methoxy butyl oleate. ................................................................................... 150 Figure B.2: Mass spectrum of ethoxy butyl oleate. ...................................................................................... 151 Figure B.3: Mass spectrum of ethoxy ethyl oleate. ....................................................................................... 151 Figure B.4: Mass spectrum of propoxy butyl oleate. .................................................................................... 152 Figure B.5: Mass spectrum of butoxy butyl oleate. ...................................................................................... 153 Figure B.6: Mass spectrum of pentoxy butyl oleate. .................................................................................... 154 Figure B.7: Mass spectrum of hexoxy butyl oleate. ..................................................................................... 155 Figure B.8: Mass spectrum of octoxy butyl oleate. ...................................................................................... 156 Figure B.9: Mass spectrum of ethylhexoxy butyl oleate............................................................................... 157

Abstract _________________________________________________________________________

xi

ABSTRACT _________________________________________________________________________

A property of biodiesel that currently inhibits its use is its relatively poor low-temperature

properties, most commonly expressed as cloud point. Improving the low-temperature

properties of biodiesel to those for petroleum based diesel will remove one of the few

physicochemical barriers to its more widespread application. Improvement of biodiesel

low-temperature properties by alkoxylation is a potential method that is investigated in this

thesis. While previous work has been performed with model compounds and synthetic

laboratory conditions, this work investigates the likely success of a commercial process to

produce alkoxylated biodiesel. Process parameters were constrained to atmospheric

pressure, low temperatures and reasonable reaction times, while avoiding the use of

organic solvents.

Epoxidation and alkoxylation of methyl biodiesel produced from canola oil was studied to

determine the best conditions while simultaneously developing the analytical methods. A

gas chromatography-mass spectrometry method was developed to determine conversion

and selectivity for epoxy and alkoxy biodiesel. The best reaction conditions for the

epoxidation step, based on conversion and selectivity, and the option of either in-situ

generated peroxyformic acid or peroxyacetic acid as the oxygen carrier were determined.

Optimal conditions were H2O2 / biodiesel molar ratio of 2:1, acetic acid / biodiesel molar

ratio of 0.2:1, acid catalyst to acetic acid / peroxide of 2 wt% and a 6h reaction time at

60°C. The optimal reaction conditions for methyl biodiesel were then transferred to ethyl

and butyl biodiesel. An acid catalysed alkoxylation with the same alcohol as the ester

head-group was then performed and the cloud point impact was assessed. Alkoxylation of

methyl and ethyl biodiesel resulted in reduced low-temperature tolerance while alkoxy

butyl biodiesel displayed a slightly improved tolerance.

Since butoxylated butyl biodiesel was the most promising in terms of cloud point

improvement, the next phase of work was concerned with maximising selectivity for

butoxy biodiesel. A range of conditions including reaction time, temperature, catalyst

concentration and molar ratio of alcohol were studied. Optimal conditions for the

butoxylation of epoxy butyl biodiesel were: 80°C, 2 wt% sulfuric acid and a 40:1 molar

ratio of butanol over a period of 1h. Conversion of epoxy butyl biodiesel was 100% and

Abstract _________________________________________________________________________

xii

selectivity for butoxy biodiesel was 87.0%. The cloud point of butoxy butyl biodiesel (46%

conversion of unsaturated fraction) was identical to that for butyl biodiesel. To determine

the impact of higher conversion of unsaturated ester to butoxy ester, a batch of butyl

biodiesel was subjected to 30h of epoxidation resulting in a conversion of 93%,

corresponding to a butoxy content of 74 wt%. The cloud point of this material was 2°C,

representing an increase of 5K over that for butyl biodiesel. Blends of the high conversion

batch of butoxy biodiesel showed that cloud point was virtually unchanged at

concentrations below 35 wt% and then increased 1K every 8 wt% to approximately

70 wt % butoxy biodiesel.

The last phase involved the investigation of the impact of longer and branched side-chains

on the properties of butyl biodiesel. Longer straight-chain alcohols were added at the

epoxidised double bonds, as were some branched isomers under the optimal conditions

determined in phase two. Alcohols included: methanol, ethanol, n-propanol, n-butanol,

tert-butanol, n-pentanol, n-hexanol, n-octanol and 2-ethylhexanol. Alkoxylation of butyl

biodiesel with methanol, ethanol and propanol increased the cloud and pour point of butyl

biodiesel. Alkoxylation with alcohols larger than butanol produced significant

improvements in low-temperature properties as indicated by lower cloud and pour points.

The lowest cloud point achieved was for ethylhexoxy butyl biodiesel at -6°C, a 6K

reduction in cloud point over conventional methyl biodiesel. Alkoxylation also resulted in

significant increases in kinematic viscosity, with the viscosity of ethylhexoxy butyl

biodiesel being 9.76 mm2.s-1, more than double that for methyl biodiesel.

Declaration _________________________________________________________________________

xiii

DECLARATION _________________________________________________________________________

This work contains no material which has been accepted for the award of any other degree

or diploma in any university or other tertiary institution to Paul Smith and, to the best of

my knowledge and belief, contains no material previously published or written by another

person, except where due reference has been made in the text.

I give consent to this copy of my thesis when deposited in the University Library, being

available for loan and photocopying, subject to the provisions of the Copyright Act 1968. I

also give permission for the digital version of my thesis to be made available on the web,

via the University’s digital research repository, the Library catalogue, the Australian

Digital Thesis Program (ADTP) and also through web search engines, unless permission

has been granted by the University to restrict access for a period of time.

SIGNED:……………………………… DATE:……………..

Acknowledgements _________________________________________________________________________

xiv

ACKNOWLEDGEMENTS _________________________________________________________________________

This work has been undertaken with financial assistance from a Faculty of Engineering,

Computer and Mathematical Sciences Divisional Scholarship, for which I am grateful.

I received invaluable assistance from laboratory managers in the school of Chemical

Engineering (Mr. Andrew Wright), Petroleum Engineering (Ms. Delise Hollands) and

Chemistry and Physics (Mr. Gino Farese). I am also grateful for the use of the school of

Petroleum Engineering’s GC-MS and the school of Chemistry’s distillation equipment.

I am indebted to the team of supervisors that have provided advice and guidance

throughout this process: Associate Professor Brian O’Neill, Associate Professor Q. Dzuy

Nguyen and Doctor Yung Ngothai. I am especially thankful to my primary supervisor

Associate Professor Brian O’Neill not only for his support and wisdom, but for facilitating

my return to study. Without his assistance and belief in my abilities, I would not have had

the opportunity to move into chemical engineering.

I am grateful to my parents Raymond and Faye, for providing me with the opportunity to

gain a good education and their encouragement to excel. They are also largely responsible

for the qualities that have allowed me to reach this goal including the tenacity, focus,

patience, attention to detail and critical thinking skills.

Lastly, but by no means least, I am indebted to my wife, Tracey. I would not have

contemplated this work without her support and encouragement. She has provided

unwavering support throughout the whole process.

List of Publications _________________________________________________________________________

xv

LIST OF PUBLICATIONS _________________________________________________________________________

Smith, P. C., O'Neill, B. K., Nguyen, Q. D., Colby, C. and Ngothai, Y. (2008).

"Alkoxylation of Biodiesel and its Impact on Low-Temperature Properties."

Chemeca 2008, Newcastle, Australia.

Smith, P. C., Ngothai, Y., Nguyen, Q. D., and O'Neill, B. K. (2009). "Alkoxylation of

Biodiesel and its Impact on Low-Temperature Properties." Fuel, 88, 605-612.

Smith, P. C., O'Neill, B. K., Ngothai, Y. and Nguyen, Q. D. (2009). "Butoxylation of Butyl

Biodiesel: Reaction Conditions and Cloud Point Impact." Energy and Fuels, 23,

3798-3803.

Smith, P. C., O'Neill, B. K., Ngothai, Y. and Nguyen, Q. D. (2009). "Butoxylation of Butyl

Biodiesel and its Impact on Cloud Point." 8th World Congress of Chemical

Engineering, Montreal, Quebec, Canada.

Smith, P.C., O'Neill, B.K., Ngothai, Y., Nguyen, Q.D. (2010). “Improving the Low-

Temperature Properties of Biodiesel: Methods and Consequences.” Renewable

Energy 35, 1145-1151.

Smith, P.C., O'Neill, B.K., Ngothai, Y., Nguyen, Q.D. (2009). “The Flow Properties of

Alkoxylated Biodiesel” Submitted to Fuel in June 2009, still under review.

Chapter 1

General Introduction

Chapter 1 General Introduction _________________________________________________________________________

2

GENERAL INTRODUCTION _________________________________________________________________________

1.1 Context Biodiesel has been widely accepted as an additive for petroleum derived diesel (hereafter

referred to as “diesel”) and is typically added at proportions of 2 - 20 vol%. Biodiesel has

similar physical properties to diesel and can be substituted without major modification to

the engine or fuel systems. A property that currently inhibits its use is the relatively poor

low-temperature properties, most commonly expressed as cloud point (CP), pour point

(PP) or cold-filter plugging point (CFPP). Diesel has a CP of -16°C and requires

‘winterisation’ during the colder months in high latitudes. Biodiesel typically has a CP of

around 0°C, thereby limiting its use to ambient temperatures above freezing. Blends of

20% methyl soyate with diesel can reduce the ambient temperature cut-off to -10°C.

Improving the low-temperature properties of biodiesel to those for diesel, or even better,

will remove one of the few physicochemical barriers to its more widespread application.

1.2 Purpose The simplest description of the purpose of this study is the determination of whether

alkoxylation can produce an improvement in the low-temperature properties of biodiesel.

The ultimate aim was to develop a commercially viable process for improving the low-

temperature performance of biodiesel that did not add significantly to the final cost of the

product and one which used renewable feedstocks. n-Butanol was chosen because it can be

produced biologically from biomass in a similar process to that for bio-ethanol. The

epoxidation and alkoxylation steps required for the conversion of the biodiesel can be

performed with readily available materials, under mild conditions and with substantially

the same simple equipment as is currently used for biodiesel.

The knowledge that was not previously available and which was discovered during this

work includes the following:

� the optimal conditions and conversion rate for the epoxidation of biodiesel within

the bounds of commercial viability; atmospheric pressure, the absence of additional

organic solvents either during conversion or purification, and <12h reaction time,

� the optimal conditions for the alkoxylation of epoxy biodiesel,


3

� the selectivity for alkoxy biodiesel under the optimal conditions,

� the impact of alkoxylation, under the optimal conditions, on the low-temperature

properties and viscosity of biodiesel,

� the impact of side-chain length and branching on the low-temperature properties

and viscosity of biodiesel.

1.3 Process of Investigation and Thesis Structure The process of investigation is described here with the aid of a flow chart (Figure 1.1). The

investigation contained in this thesis could not begin until a reliable method for the

characterisation of reaction products was developed. This was a time consuming exercise

and is described in chapter 4. Initial experimental work for the synthesis of alkoxy

biodiesel centred on the epoxidation and alkoxylation of methyl, ethyl and butyl biodiesel

produced from canola oil (chapter 5). Methyl biodiesel was chosen as a control since the

vast majority of biodiesel is produced with fossil derived methanol. Ethanol and butanol

were chosen because they can be produced from renewable sources such as sugar based

crops. The first trials were aimed at determining the best reaction conditions for the

epoxidation step, based on conversion and selectivity, and the option of either in-situ

generated peroxyformic acid or peroxyacetic acid as the oxygen carrier. An acid catalysed

alkoxylation with the same alcohol as the ester head-group was then performed and the

cloud point impact was assessed. During this process the analytical methods were

developed.

Since butoxylated butyl biodiesel was the most promising in terms of cloud point

improvement, the next phase of work was concerned with maximising selectivity for

butoxy biodiesel (chapter 6). A range of conditions including reaction time, temperature,

catalyst concentration and molar ratio of alcohol were studied. The impact on cloud point

of the alkoxy biodiesel was determined at the optimal conditions and at a higher

conversion rate.

The third phase involved the investigation of the impact of longer and branched side-

chains on the properties of biodiesel (chapter 7). Longer straight-chain alcohols were

added at the epoxidised double bonds, as were some branched isomers under the optimal

conditions determined in phase two. Other properties that were likely to be altered due to


4

the chemical modification of biodiesel were also assessed, including pour point and

viscosity. The final section (chapter 8) of the thesis is a discussion of the findings and

suggestions for further work. Chapter 8 also contains a discussion of the economic

implications of a process to produce the best candidate fuel, in terms of low-temperature

tolerance.

Figure 1.1: Flow chart of the process of investigation.

Biodiesel synthesis and

characterisation

Initial epoxidation and alkoxylation of methyl biodiesel

Optimisation of epoxidation of methyl biodiesel to improve selectivity

Extension to ethyl and butyl biodiesel

Alkoxylation of methyl, ethyl, butyl biodiesel

Test method development

Decision to concentrate on butyl biodiesel

Assessment of physical properties

Extension to higher order alkoxy groups

Optimisation of butoxylation of butyl biodiesel

Assessment of cloud point impact

Assessment of cloud point impact

Test methods

GC-FID, GC-MS

GC-FID, GC-MS, FTIR

GC-MS, cloud point

GC-MS

GC-MS

Cloud point

GC-MS

Cloud point

GC-MS

Cloud point, Pour point, viscosity

Chapter 2

Biodiesel Low-Temperature Properties

Chapter 2 Biodiesel Low-Temperature Properties _________________________________________________________________________

6

BIODIESEL LOW-TEMPERATURE PROPERTIES _________________________________________________________________________

This chapter discusses the background to the content and process of investigation

undertaken for this thesis. This is not a comprehensive literature review encompassing all

of the peripheral concepts contained within this work. It is assumed that the reader is well

versed in biodiesel processing, production and use. Section 2.1 outlines the problem of

poor low-temperature flow properties of biodiesel. Section 2.2 discusses the issue of

biodiesel cloud point in more detail while section 2.3 discusses the possibilities for the

improvement of cloud point. Section 2.4 briefly outlines the potential impact of cloud point

improvement techniques on other properties of biodiesel. Finally, section 2.5 proposes a

solution to the problem and provides the rationale for the chosen process of investigation.

2.1 The Problem Biodiesel has been widely accepted as an additive for fossil-derived diesel fuel in

compression ignition engines. Production of biodiesel in the EU reached 5.7 million tonnes

in 2007, up from 4.9 million tonnes in 2006. Biodiesel constituted 76% of the EU’s

biofuels production and the EU produced 68% of biodiesel worldwide. The United States

is the second largest producer at 1.5 million tonnes in 2007 (EBB 2008). The vast majority

of biodiesel is used as a transport fuel and is typically blended with diesel in proportions

from 5 to 20 vol%.

B20 (20% biodiesel/diesel blend) reduces life cycle petroleum consumption by 19% and

lifecycle CO2 emissions by 16% (Harrow 2007). Other benefits of biodiesel have been

widely recognised and include; high cetane number (CN), high lubricity (even in blends of

1 - 2%) (Van Gerpen 2005) and significant reductions in emissions of sulfur oxides,

hydrocarbons, particulates, polycyclic aromatic hydrocarbons and carbon monoxide

(Lapuerta et al. 2008). With a relatively high flash point of 154°C (for methyl soyate), high

biodegradability and low toxicity, biodiesel is regarded as a much safer fuel than petroleum

diesel (Demirbas 2005). Biodiesel is also recognised as a renewable fuel and requires 0.31

MJ of fossil energy to produce 1 MJ of fuel product (Harrow 2007).


7

A key property of biodiesel currently restricting its application to blends of 20% or less is

its relatively poor low-temperature properties. Petroleum diesel fuels are plagued by the

growth and agglomeration of paraffin wax crystals when ambient temperatures fall below

the fuel’s cloud point (CP). These solid crystals may cause start-up problems such as filter

clogging when ambient temperatures drop to around -10 to -15°C (Chandler et al. 1992).

Whilst the CP of petroleum diesel is reported as -16°C, biodiesel typically has a CP of

around 0°C, thereby limiting its use to ambient temperatures above freezing (Chandler et

al. 1992), (Knothe et al. 2004), (Dunn et al. 1996). Limiting blends to 20% methyl soyate

can reduce the ambient temperature cut-off to -10°C (Dunn et al. 1996).

Waste cooking oils and tallows have the potential to significantly reduce the final price of

biodiesel due to their low raw material cost when compared to virgin vegetable oils.

Unfortunately, the highly saturated nature of these oils means that the CP of biodiesel

derived from such feedstocks can rise to 17°C. Oils such as palm oil are also beneficial

because they do not directly compete with food crops and therefore place pressure on food

prices. This must however, be tempered with recent concerns about deforestation and the

associated reduction in carbon sequestration, as well as the loss of habitat in regions

suitable for palm plantations. The cloud point of palm oil methyl esters is typically 13°C,

severely limiting their use at high blend ratios and in cooler climates (Knothe et al. 2004).

Hence, the use of highly unsaturated virgin vegetable oils as feedstock is favoured in terms

of cold-temperature properties. Unfortunately, these feedstocks incur increased cost, and

the resulting biodiesel has a reduced cetane number and reduced oxidation stability.

2.2 Factors Influencing Biodiesel Cloud Point

2.2.1 Fatty Acid Profile

The dominant factor influencing the low-temperature properties of biodiesel is the number

of double bonds in the alkyl chain of the fatty acid ester. Oils commonly used as a feed for

biodiesel production such as rapeseed, canola and soybean are highly unsaturated with

unsaturated fatty acid levels typically 85 wt% or greater. Cloud points for these fuels are

normally around 0°C. Tallow and palm oil esters exhibit similar saturation levels of around

50 wt% and CP’s lie in the range from 8 to 17°C.


8

A thermodynamic study of binary mixtures of various fatty acid methyl esters (FAME)

was performed to develop a predictive model for estimating the cloud point of biodiesel

fuels produced from various feedstocks (Imahara et al. 2006). The conclusion was that the

cloud point of biodiesel could be determined solely from quantitation of the amount of

saturated fatty acid methyl esters, regardless of the chemical nature of unsaturated esters.

Binary mixtures of a saturated and an unsaturated FAME such as methyl palmitate (C16:0)

and methyl oleate (C18:1) produced a monotonic increase in CP with increasing fraction of

C16:0 from the CP of pure C18:1 to that of pure C16:0. For mixtures containing only

saturated esters (C16:0 and C18:0 for instance), a lower CP at a eutectic point was

observed at a particular composition below the CP of either pure component. The

comparison of the CP of the binary mixtures with the so called CP of the pure components

is however not a valid one since the ‘CP’ of pure compounds is actually better described as

the melting point. Extreme caution should be used when comparing the so called cloud

point of pure compounds with that of mixtures. In a more realistic scenario that is more

representative of an actual FAME, multi-component mixtures of four different esters were

prepared and the CP of those was determined to elucidate trends. When the amount of

saturated esters was fixed and only the fractions of unsaturated esters was changed, that is,

the ratio of mono-and di-unsaturated ester (C18:1 and C18:2), the CP remained practically

unchanged. By contrast, if the total amount of saturated esters changed, then a dramatic

change in CP was observed. Finally, the thermodynamic model was compared with the CP

of FAME produced from real feedstocks. The discrepancy between the predicted CP for

the 8 different fats/oils and the actual ranged from 0K and 5K. Clearly, further work is

required to improve the accuracy of the model.

Park et al. (2007) blended palm, rapeseed and soybean oil methyl esters to create 21

different fatty acid (FA) profiles and measured the cold filter plugging point (CFPP) and

oxidation stability of the resulting blends (Park et al. 2007). The authors were unable to

discover a direct correlation between CFPP and oleic acid (C18:1) content but they found a

strong positive relationship with palmitic acid (C16:0) content and a strong negative

relationship with the total fraction of unsaturated fatty acid content. As well, oxidation

stability, as would be expected, decreased with increasing fractions of unsaturated FA.

The remaining key feature of the fatty acid moiety is the length of the fatty chain. A strong

correlation is observed between chain length and melting point of the corresponding FA.


9

Table 2.1 summarizes literature values for the melting points of some saturated fatty acids

and their esters. A clear and pronounced trend of increasing melting point with increasing

chain length is evident for both the fatty acid and the ester. This trend can also be

elucidated from the multi-component data presented by Imahara et al. (2006). Three of the

blends (samples 13-15) with an identical fraction of C18:1, progressively larger

proportions of C18:0 and reducing proportions of C16:0 exhibited increasing cloud points

from 4°C to 18°C. Two blends (samples 15 and 12) possessed an identical fraction of

C18:1 of 0.79. In the case of sample 15, the remaining 29% was C18:0 resulting in a CP of

18°C, while sample 12 contained 29% C16:0 and possessed a CP of 7°C (Imahara et al.

2006).

Table 2.1: Properties of Fatty acids and Esters. Source: (Knothe 2005)

Trivial (systematic) name; acronyma Melting Point (°C) Cetane Number

Caprylic (Octanoic) acid; 8:0 16.5 -

Ethyl ester -43.1 -

Capric (Decanoic) acid; 10:0 31.5 47.6

Ethyl ester -20 51.2

Lauric (Dodecanoic) acid; 12:0 44 -

Methyl ester 5 61.4

Ethyl ester -1.8 -

Myristic (Tetradecanoic) acid; 14:0 58 -

Methyl ester 18.5 66.2

Ethyl ester 12.3 66.9

Palmitic (Hexadecanoic) acid; 16:0 63 -

Methyl ester 30.5 74.5

Ethyl ester 19.3/24 93.1

Stearic (Octadecanoic) acid; 18:0 71 61.7

Methyl ester 39 86.9

Ethyl ester 31-33.4 76.8 a Numbers denote the number of carbons and double bonds

The complex nature of natural fats and oils ensures that any elucidation of such trends

from the fatty acid profiles is difficult. However, certain trends are evident from the data

presented in Tables 2.2 and 2.3. Coconut oil contains very high levels of saturated, low

molecular weight fatty acids and relatively little unsaturated fatty acids. As a consequence,

its ethyl ester possesses a relatively high CP of 5°C. By contrast, safflower oil has a very

high proportion of unsaturated FA, with little saturated FA and its CP is -6°C (ethyl ester).


10

Palm oil is composed of virtually identical proportions of saturated FA (C16:0, C18:0) and

unsaturated FA (C18:1, C18:2) and its CP is 8°C. Safflower and sunflower oil have similar

proportions of saturation but safflower ethyl ester has a CP of -6°C, whilst sunflower ethyl

ester has a CP of -1°C. The nature of the saturated and unsaturated fractions has created a

much higher CP for sunflower ethyl ester. Firstly, the fraction of C18:0 and C20:0 in

sunflower is almost double that for safflower, hence the longer chain length of the

saturated fraction results in a higher CP for sunflower ethyl ester. Secondly, the

unsaturated portion of sunflower contains less of the highly non-linear C18:2 and more of

the C18:1 than safflower. This large diversion in CP demonstrates the complex inter-

relationship between chain length and saturation level. It is therefore evident that very

small differences in the proportions of fatty acids can have a large impact on cold-flow

properties.


11

Table 2.2: Properties of Fats/oils and their esters. Source: (Knothe et al. 2004)

NOTE: This table is included on page 11 of the print copy of the thesis held in the University of Adelaide Library.


12

Table 2.3: Fatty Acid Compositions of Naturally Occurring Fats and Oils. Source: (Hasenhuetti 2005)

Fat or Oil Fatty Acid (%) Saturated

(%)

Unsaturated

(%) 6:0 8:0 10:0 12:0 14:0 16:0 16:1 18:0 18:1 18:2 18:3 20:0 22:0 22:1

Canola 3.9 0.2 1.9 64.1 18.7 9.2 0.6 0.2 6.6 92.2

Coconut 0.5 8.0 6.4 48.5 17.6 8.4 2.5 6.5 1.5 0.1 92 8

Corn 12.2 0.1 2.2 27.5 57.0 0.9 0.1 14.5 85.5

Olive 13.7 1.2 2.5 71.1 10.0 0.6 0.9 17.1 82.9

Palm 0.3 1.1 45.1 0.1 4.7 38.5 9.4 0.3 0.2 51.4 48.3

Rapeseed 0.1 2.8 0.2 1.3 21.8 14.6 7.3 0.7 0.4 34.8 5.3 78.7

Safflower 0.1 6.5 2.4 13.1 77.7 0.2 9.2 90.8

Soybeana 0.1 10.9 0.1 4.2 25.0 52.7 6.2 0.3 0.1 15.6 84.0

Sunflower 0.5 0.2 6.8 0.1 4.7 18.6 68.2 0.5 0.4 12.6 87.4

Beef

tallow 0.1 0.1 3.3 25.5 3.4 21.6 38.7 2.2 0.6 0.1 50.7 44.9

a (Zlatanic et al. 2004)

2.2.2 Alcoholic Adduct

The nature of the alkyl head-group of the ester produces an equally clear effect on the low-

temperature properties of biodiesel. The use of long (3 - 8 carbon) normal alcohols or

branched alcohols to manufacture biodiesel reduces the CP compared to those for

conventional methyl esters. Methyl fatty acids possess sufficient polarity in their head-

groups to provide an amphiphilic nature that results in the head to head alignment of

molecules. Ethyl or larger alkyl esters include non-polar head-groups that are sufficiently

large to shield the forces between more polar portions of the head-group. These esters

therefore orient themselves in a head to tail arrangement with much larger molecular

spacing. Bulky head-groups disrupt the spacing between individual molecules in the crystal

structure causing rotational disorder in the hydrocarbon tail-group (Knothe et al. 2004).

The net result is that melting points for alkyl palmitate and stearate esters exhibit a decline

for head-group lengths up to and including n-butyl, but then increase for n-pentyl and

larger. Esters with branched headgroups show reductions in CP of around 3K for canola oil

and up to 9K for soybean oil (Table 2.2).

Lee et al. (1995) discovered that isopropyl and 2-butyl esters of normal soybean oil

crystallised 7K to 11K and 12K to 14K lower, respectively, than the corresponding methyl

esters. Isopropyl esters of lard and tallow had crystallisation temperatures similar to those


13

for methyl esters of soybean oil despite their increased saturation level (Lee et al. 1995). In

addition to improvements in the low-temperature flow properties of biodiesel with longer

head-groups, the ignition quality may also be improved. Increasing carbon chain length in

saturated esters increases cetane number. While branching of aliphatic hydrocarbons

results in reduced CN, branching of the ester at the head-group produces a CN that is

similar to that for methyl esters (Knothe et al. 2004).

2.3 Low-Temperature Property Improvement Several approaches have been proposed to improve the low-temperature properties of

biodiesel, including; blending with petroleum diesel; the use of additives; and the chemical

or physical modification of either the oil feedstock or the biodiesel product. Blending with

diesel is only effective at low biodiesel proportions (up to 30 vol%) with cloud points to

around -10°C (Dunn et al. 1996). Clearly, blends with petroleum diesel do not change the

chemical nature and therefore properties of biodiesel and will not facilitate their use at

higher concentrations. The use of additives can be further classified into traditional diesel

additives and emerging new technologies developed specifically for biodiesel.

2.3.1 Additives

Traditional petroleum diesel additives can be described as either pour point (PP)

depressants or wax crystalline modifiers. Pour point depressants were developed to

improve pumpability of crude oil and do not affect nucleation habit (Knothe et al. 2004).

Instead, these additives inhibit crystalline growth thereby eliminating agglomeration. They

are typically composed of low molecular weight copolymers similar in structure to

aliphatic alkane molecules, the most widely applied group being copolymers of ethylene

vinyl ester. Wax crystalline modifiers, as the name suggests, are copolymers that disrupt

part of the crystallisation process to produce a larger number of smaller, more compact

wax crystals (Knothe et al. 2004). Several studies have been undertaken on both types of

additives in both petroleum blends and pure biodiesel. For example, the pour point of neat

soybean methyl ester was lowered by as much as 6K (Dunn et al. 1996). Similar

improvements in cold filter plugging point (CFPP) were achieved but no discernable

improvement in CP was reported, as may be expected when taking into account their mode

of action.


14

As previously stated, a potential mechanism for reducing the CP of biodiesel is the use of

bulky moieties that disrupt the orderly stacking of ester molecules during crystal

nucleation. To this end, Knothe et al. (2000) synthesised fatty diesters via the p-toluene

sulfonic acid-catalysed esterification of both mono- and bi-functional fatty acids and

alcohols (Knothe et al. 2000). However, CP values for blends with soybean methyl ester

were reduced by 1K at most. A second method is the modification of block copolymers of

long-chain alkyl methacrylates and acrylates specifically for the purpose of addition to

biodiesel to modify low-temperature performance. Patent applications by Scherer et al.

(2001) have been lodged claiming reductions in PP and CFPP for lubricant oils and

biodiesel (Knothe et al. 2004).

Ming et al. (2005) studied the effect of a variety of additives, either synthesised themselves

or commercially available, including: Tween-80, dihydroxy fatty acid (DHFA), acrylated

polyester pre-polymer (APP), palm-based polyol (PP), a blend of 1:1 DHFA and ethyl

hexanol (DHFAPP), an additive synthesised using DHFA and ethyl hexanol (DHFAEH)

and castor oil ricinoleate (Ming et al. 2005). They reported an average reduction in CP

values of 5.5K, with the largest reduction being 10.5K for the addition of 1% DHFA and

1% PP to palm oil methyl ester. They speculated that the effectiveness in particular of the

polyhydroxy compounds was due to the interaction between the hydroxy groups of the

additives and the samples. Unfortunately, a large increase in the viscosity of the blends was

reported. The addition of 1.0% PP to palm oil methyl ester increased its viscosity from

29.5 cP to 42.2 cP. Some effort has also been made to utilise the major by-product of

biodiesel manufacture, glycerol, by reacting it with isobutylene to produce butyl ethers of

glycerol (Noureddini 2002). A CP of -5°C for 12% butyl ether and methyl ester was

claimed.

Soriano Jr. et al. (2006) suggested that the addition of fatty acid moieties whose chain

length was similar to the corresponding fatty acid ester, but with protruding polar groups

would have a positive (reduced PP) impact on low-temperature properties (Soriano Jr. et

al. 2006). Ozonised vegetable oils of sunflower (SFO), soybean (SBO) and rapeseed

(RSO) oil were prepared by passing ozone through the oil in a bubbling bed reactor. First,

ozonised SFO was added to SFO, SBO and RSO derived biodiesel at rates of 1 or 1.5 wt%.

The pour points of these blends of SFO, SBO and RSO were reduced, however cloud

points remained unchanged. In the case of similar blends of ozonised SFO with palm oil


15

derived biodiesel, PP remained unchanged, but CP was significantly reduced (from 18 to

-12°C). A further set of blends were prepared by adding the ozonised oils to biodiesel

derived from the same oil. This resulted in the greatest reductions in PP. The pour points of

these blends of SFO, SBO and RSO were lowered to -24, -12 and -30°C, respectively.

Ozonolysis on a large scale however, would undoubtedly add significantly to the cost of an

already marginal commodity, along with the obvious safety issues for such a process.

2.3.2 Feedstock Modification

Winterisation is a method for separating that fraction of oils with a solidification

temperature below a specific cut-off. One technique involves refrigeration of the oils for a

prescribed period at a specific temperature followed by decanting of the remaining liquid.

Another more energy efficient method is to allow tanks of oil to stand outside in cold

temperatures for extended periods of time. In either case, the fraction that remains molten

is separated from the solid producing an oil with improved pour and handling qualities.

Dunn et al. (1996) employed a stepwise winterisation technique with soybean methyl ester

(SME) until the oil could withstand 3 hours at a bath temperature of -10°C without

clouding (Dunn et al. 1996). Typically, five to six iterations over a period of a week were

required to produce SME with a PP that did not exceed -16°C, but resulted in a loss of

approximately 75% of the starting quantity of SME. As expected, the largest loss of

material (in percentage terms) was C16:0 which decreased by a factor of 3, while the

proportion of C18:3 increased by almost half. They found that a small fraction of saturated

long-chain methyl esters in the winterised SME had a dramatic effect on the cold-flow

properties of the transesterified winterised SME. A mixture of 26.0 wt% methyl oleate

(C18:1), 68.3% linoleate (C18:2) and 5.7% methyl linolenate (C18:3) produced a CP of

-23°C and PP of -48°C. These results were 7K to 32K below that of winterised

transesterified SME containing only 5.6 wt% of saturated methyl esters. Removal of such a

large proportion of product would be prohibitive on a large scale, as would the period of a

week to complete the process.

A more direct method for altering biodiesel feedstocks is to genetically modify the fatty

acid profile of oilseeds as suggested by Duffield et al (1998). They suggested that soybean

could be modified to produce an oil profile with elevated oleic acid (C18:1) and with

reduced polyunsaturated and saturated fatty acids. Such an oil would produce biodiesel

with increased oxidative stability and would be useful (in terms of low-temperature


16

properties) in most climates (Duffield et al. 1998). A second possibility would be to couple

elevated oleic acid (C18:1) with an increase in stearic acid (C18:0), which would have

enhanced burn qualities (CN, reduced NOx emissions) and oxidative stability but with

relatively poor low-temperature properties (Kinney and Clemente 2005). To this date little

or no progress has been made in genetic modification for the purpose of creating an oil

feedstock tailored for biodiesel production.

2.3.3 Biodiesel Modification

An obvious physical method for modifying the low-temperature properties of biodiesel is

by crystallisation fractionation, a similar process as that applied in the winterisation of oil

feedstocks. Both dry fractionation and solvent fractionation have been applied to biodiesel.

Dry fractionation involves crystallisation from the melt without the addition of a solvent

and is the simplest and least expensive method (Knothe et al. 2004). Lee et al. (1996)

applied bench-scale dry fractionation to methyl soyate in a ten step process spanning 84h

to produce a product with 5.5 wt% saturated ester and a crystallisation temperature 10.8K

lower than the equivalent non-fractionated product (Lee et al. 1996). However, the liquid

yield was only 25.5% of the starting quantity of methyl soyate. By contrast, the dry

fractionation of waste cooking oil methyl esters by Gonzalez et al.(2002) resulted in a

modest reduction in CFPP from -1 to -5°C (Gonzalez Gomez et al. 2002). This was

accompanied by a yield loss of 30 to 100% (no product) depending on the fractionation

conditions.

Solvent fractionation has significant advantages over dry fractionation, including reduced

crystallisation times and improved yields, but suffers from reduced safety and increased

costs. Hexane extraction was employed by Dunn et al. (1997) in a single-step process with

a residence time of 3.5 - 6.5h for SME (Dunn et al. 1997). The yield was 78.4% and the

resultant CP was -10°C. A similar process using iso-propanol produced a similar CP but

the relatively small reduction in saturated methyl ester content suggested that the reduced

CP was more a result of residual solvent in the SME. Other workers have attempted

solvent fractionation with various solvents, including methanol, acetone, chloroform (Lee

et al. 1996), and ethanol (Hanna et al. 1996), with limited success. Fractionation of SME

with acetone failed to provide any benefit in CP and chloroform failed to generate any

crystallisation.


17

Fractionation of biodiesel via any means results in a reduction in the proportion of

saturated esters and therefore an increase in the fraction of unsaturated esters. This has a

significant effect on other properties of the biodiesel, especially oxidative stability and

ignition quality. Fractionation will also have a large impact on the cost of production and

therefore price competitiveness of biodiesel compared to petroleum diesel. Fractionation

not only introduces a new set of unit operations including solvent addition, crystallisation

and solvent recovery, but it also results in significantly reduced yields, in some cases by

75%.

An approach for chemically converting biodiesel to triesters has been reported by Yunus et

al. (2005). They synthesised palm oil polyol esters (POTE) from fractionated palm oil

methyl esters and trimethylolpropane (TMP) using sodium methoxide as a catalyst. Pour

points of -30°C were achieved for product that was greater than 90 wt% triester. However,

the pour point of the final product was strongly dependent on the fraction of palmitate

(C16:0) remaining after fractionation of the palm oil methyl esters. The PP of POTE with a

C16:0 fraction of 9.8 wt% was -11°C, whereas the PP of POTE with 8.7% C16:0 was

-29 °C. Hence, chemical modification of palm oil based biodiesel was only successful

when combined with fractionation to significantly reduce the C16:0 content (Yunus et al.

2005).

Another approach exploits the π-bonds of the unsaturated fatty acids via electrophilic

addition to produce branched or bulky esters. Isopropyl oleate (IPO) was first epoxidised

via the in-situ peroxyformic acid method to produce around 95% epoxidised IPO (EIPO)

(Moser and Erhan 2006). �-hydroxy ethers of IPO were then prepared by the reaction of a

1M solution of EIPO with a variety of alcohols in the presence of an acid catalyst (H2SO4).

The alcohols included: ethanol, n-propanol, isopropanol, n-butanol, isobutanol, 2-methoxy

ethanol, hexanol, octanol, 2-ethyl hexanol and decanol. The authors report that the ethyl �-

hydroxy ether was the sole derivative exhibiting reduced low-temperature tolerance in

comparison with IPO. The authors reported that the three carbon ethers had similar CP and

PP to IPO even though they were determining the ‘CP’ of pure compounds. The reported

CPs were actually the melting points of the pure compounds and therefore serve only as a

comparison of similar pure compounds. As the chain length of the ether increased beyond

three carbons, a large improvement in low-temperature properties was claimed. For


18

instance, the decyl ether had a reported CP of -23°C and a PP of -24°C compared with a

CP of -9°C and a PP of -12°C for unmodified isopropyl soyate. However, this again is not

a valid comparison since the properties of a pure compound (decyl ether of isopropyl

oleate) cannot be compared with those of a mixture such as isopropyl soyate. It should be

noted also that these compounds are not representative of a commercial biodiesel fuel and

the results for CP recorded for the pure compounds can only be used as a comparison of

the impact of chemical changes to a similarly pure compound. Instead, the impact on cloud

point due to the addition of these compounds to biodiesel would be more useful.

A subsequent study by Moser and Erhan went some way towards addressing this issue by

blending four different ethers of IPO with methyl soyate at up to 2.0 wt%. At blend levels

below 0.5 wt%, no cloud point improvement was evident but the authors reported a

reduction in CP of 3K when 2 wt% 2-ethyl-hexyl ether was added to methyl soyate (Moser

and Erhan 2008). Specific gravity, surface tension and lubricity were virtually identical to

those for neat methyl soyate. However, kinematic viscosity for two of the ether blends was

higher and even at 2.0 wt% was out of specification according to EN 14214.

Yet another approach produced diesters of oleic acid via a similar technique to the

aforementioned �-hydroxy ether synthesis. Epoxy alkyl oleates (propyl, isopropyl, octyl,

2-ethylhexyl) were esterified with propionic and octanoic acid without a catalyst or

solvent. An elevated temperature of 100°C was required, coupled with a relatively long

residence time of 14 - 15h to produce approximately 95% diester following purification.

The authors again reported the cloud and pour points of these compounds, but as they were

pure substances, the results cannot be directly compared with the CP of biodiesel. The

longer mid-chain length created by the addition of octanoic acid produced better low-

temperature properties. However, the shorter mid-chain length created with propionic acid

had superior oxidation stability than the longer variant (Moser et al. 2007).

Estolides of oleic acid and tallow fatty acids, at varying ratios, were prepared by Cermack

et al. (2007) by treating the fatty acids with perchloric acid at 60°C under vacuum

(Cermack et al. 2007). After 24h, 2-ethylhexanol was added, vacuum restored and reaction

continued for a further 4h. The result was estolides of fatty acids capped with

2-ethylhexanol at yields of between 65% and 72%. As the amount of tallow in the starting


19

material was decreased, the pour and cloud points improved (lowered). At a 1 to 1 ratio of

tallow to oleic, the observed cloud point was -21°C. However, while the cloud point seems

to be a great improvement over standard biodiesel, it is suggested that most of the

improvement came from the branched head-group provided by the 2-ethylhexyl cap.

Estolides are large molecules that would undoubtedly exhibit very high viscosities and

since one of the main reasons for transesterification of triglycerides is to reduce viscosity,

it is suggested that these compounds could only be used at low concentrations in biodiesel.

Unless estolides produce a dramatic improvement in cold-flow properties at very low

addition rates, they are unlikely to be a viable solution.

2.4 Impact on Other Properties of Biodiesel Methods for the lowering of the cloud point of biodiesel that reduce the proportion of

saturated esters, thereby increasing the proportion of unsaturated esters, impact directly on

the oxidative stability and cetane number of the fuel. Oxidative stability refers to the

autoxidation of the double bonds in the tail-group of the fatty acid chains of biodiesel.

Hence, the long-term storage stability of biodiesel can be correlated with the number and

position of double bonds. The positions allylic to double bonds are especially susceptible

to autoxidation under extended storage conditions. The bis-allylic positions such as those

present in linoleic (C18:2) and linolenic (C18:3) acids are even more prone to oxidation.

Literature values for relative rates of oxidation are 1 for oleates, 41 for linoleates and 98

for linolenates (Knothe 2005).

The cetane number of a fuel is a measure of the ignition delay, that is, the time that passes

between injection of the fuel into the cylinder and ignition. The shorter the ignition delay,

the higher the CN. Higher CN’s in petroleum diesel have been associated with lower NOx

emissions. Despite the fact that biodiesels generally have higher CN’s, emissions of NOx

from these fuels are slightly higher than for petroleum diesel. The higher CN of biodiesel is

still considered as a positive. Studies have shown that as levels of unsaturation increase,

CN decreases, and as the chain length increases the CN increases (Ribeiro et al. 2007).

Generally, branching of aliphatic molecules reduces the cetane number of diesel fuels.

However, it has been shown that branched esters are competitive with methyl esters and


20

straight-chain analogues in terms of CN but possess improved low-temperature properties

(Knothe 2005).

High fuel viscosity can affect fuel atomisation upon injection into the cylinder and

ultimately result in the formation of engine deposits. The viscosity of transesterified oil

(biodiesel) is roughly one order of magnitude lower than the parent oil (Knothe et al.

2004). Biodiesel has a higher viscosity than petroleum diesel as reflected in the relevant

standards, which are 3.5 - 5.0 mm2.s-1 for biodiesel (EN 14214) and 1.9 - 4.1 mm2.s-1 for

petroleum diesel (EN 590). Therefore, any further increase in the viscosity of biodiesel is

highly undesirable. Unfortunately, many additive and chemical modification techniques

proposed for improving the cold-flow properties of biodiesel can result in an increase in

viscosity. Any method that introduces bulky branched molecules, even at low levels, may

have a dramatic effect on viscosity, as demonstrated by Ming et al. (2005). Therefore, even

though the reports of the use of estolides, triesters or polyols as cloud point suppressants

did not mention specifically their impact on viscosity, the likelihood of an undesirable

effect is high.

2.5 Proposed Solution The relatively poor low-temperature properties of biodiesel remain as one of the major

impediments to its more widespread application as a fuel for compression ignition engines.

Many methods for improvement have been assessed, unfortunately most have had limited

success. The impact of each method on other properties of biodiesel as a fuel must also be

thoroughly assessed to ensure that the benefits that biodiesel currently offer are not

excessively eroded. In addition, the impact on production costs of any method applied for

the alteration of low-temperature performance must be assessed carefully so that biodiesel

can continue to compete with petroleum and other renewable fuels for market share.

Clearly, a trade-off exists between the use of lower cost feedstocks, that generally have

poorer low-temperature properties, and the increased cost of measures to improve low-

temperature performance.

Alteration of the fatty acid profile of seed oils by means of genetic modification has been

successfully performed in the past to create canola oil for instance. However, such attempts


21

will require substantial effort and time, and will most likely only deliver mild

improvements in the low-temperature performance of highly saturated oils such as palm

oil. Dramatic improvements in cloud point cannot be expected when the narrow range of

cloud point for existing oils, including those with favourable fatty acid profiles is noted.

Likewise, genetic alteration of oil seed crops will not have a dramatic effect on the lower-

cost biodiesel feedstocks such as used cooking oils, waste grease or animal based tallows.

These feedstocks promise large reductions in the raw material costs which are estimated to

account for up to 88% of the total cost of biodiesel (Haas et al. 2006). Limited information

is available on the impact of higher proportions of lower molecular weight fatty acids (C12

or less) on properties but anecdotal evidence suggests that a major shift to these smaller

molecules could result in a significant reduction in cloud point. Methods for cleavage of

unsaturated C16 and C18 fatty acids at the labile double bonds therefore warrant

investigation. The lower average molecular weight would also provide a benefit in terms of

reduced viscosity of the final product but, at the expense of a lower heating value and

cetane number.

Additives designed for petroleum diesel have proven ineffective in reducing the cloud

point of biodiesel, largely because they were designed to disrupt the crystallisation process

of alkane based mixtures. Other additives have been more successful but depending on the

chosen additive require the use of expensive materials or require relatively high

percentages and therefore may detrimentally alter other properties of the fuel. As well, the

introduction of any additive mandates the incorporation of a blending step in the process

which adds cost and complexity.

Winterisation and fractionation of either the oil feedstock or the finished product can

provide large improvements in low-temperature properties by altering the fatty acid profile.

Available methods involve a multi-step, long duration process to remove higher melting

point components. Normally these steps consume large amounts of energy and result in

unacceptably high reductions in yield. Such methods are the crudest methods available and

should only be considered as a last resort.

Incorporation of longer and/or branched chain alcohols as head-groups on the esters that

constitute biodiesel offers significant improvements in low-temperature performance with

a reduction in CP for soybean oil of up to 9K. Currently the vast majority of biodiesel is


22

produced from methanol derived from fossil sources, mainly because of its low cost.

Ethanol provides only minimal improvement in low-temperature performance and is more

costly to produce, but has the advantage of being produced from renewable sources. Propyl

and butyl alcohols are currently obtained from fossil sources also and would almost

certainly carry a cost penalty. Recently, bio-butanol has gained some attention and similar

to ethanol, can be produced from renewable sources, but is currently costly. Bio-butanol is

also n-butanol which is much less effective than iso-butanol at improving low-temperature

properties.

Addition of side-chains at the double bonds on the tail-group of fatty acid esters may

produce some improvement in the low-temperature performance of biodiesel. Furthermore,

elimination of the labile double bonds results in a direct improvement in oxidation

stability, while maintaining a high cetane number. A significant advantage of this method

is that the process uses materials and methods very similar to those already employed for

the conversion of oils to biodiesel. The same alcohol as that used for transesterification of

the oil can be used as the adduct to the alkyl chain of the ester and the conditions,

including the catalyst for etherification are the same as those currently employed for

esterification. The similarity in materials and conditions could also lead to a process that

uses substantially the same equipment but with an altered process sequence. Therefore,

depending on the percentage of alkoxy biodiesel required to sufficiently lower the cloud

point, the additional processing penalties may be relatively minor, especially if low-cost

feedstocks are also utilised. A potential negative impact of alkoxylation on biodiesel

properties may be changes to viscosity. Depending on the proportion of ��-hydroxy ethers

(alkoxy biodiesel) required to sufficiently lower cloud point, viscosity may be materially

affected since hydroxyl groups are generally associated with higher viscosities.

However, this method of chemical modification only modifies the available unsaturated

portion and leaves the saturated portion unchanged. The branched moieties must therefore

have the ability to suppress the normal crystallisation behaviour of the saturated portion if

they are to be successful at improving the low-temperature flow properties of biodiesel. To

date only heterogeneous ester/ether model compounds, at high concentrations, have been

assessed for their low-temperature properties. Clearly, further work in this area is

warranted, particularly with real feedstocks to produce modified biodiesel with the same


23

head-group and ether group. The impact of lower proportions of alkoxy biodiesel on the

properties of biodiesel should also be assessed. If it is shown that low concentrations of

these modified esters delivers a large improvement in low-temperature flow properties then

a large impact on other important qualities of biodiesel including; viscosity, oxidative

stability and cetane number may be avoided.

Chapter 3

Materials and Methods

Chapter 3 Materials and Methods _________________________________________________________________________

25

MATERIALS AND METHODS _________________________________________________________________________

3.1 Materials Chemicals for synthesis and analysis were purchased from laboratory chemical suppliers

and were of a grade appropriate for their intended purpose. The vegetable oil feedstock

was canola oil (Crisco, Meadow Lea Foods), purchased from a supermarket. Free fatty acid

(FFA) content was determined according to ASTM D 5555-95 and was found to be 0.2 mg

KOH/g oil, equivalent to 0.1 wt% FFA.

The following chemicals were used without further purification: n-Butanol (ACS, ISO,

Reag. Ph Eur, Merck KGaA); potassium hydroxide (pellets GR for analysis, Merck Pty.

Limited); sulfuric acid 96% (Merck KGaA); formic acid (ACS, Sigma-Aldrich); acetic

acid (ACS, Sigma-Aldrich); sodium sulfate (Analytical reagent, Chem-Supply); sodium

bicarbonate (ACS reagent, Sigma-Aldrich); pyridine (ACS reagent, Ph Eur, Merck

KGaA); n-heptane (Reag. Ph Eur, Merck KGaA); hydrogen peroxide (solution 30%,

puriss., stabilised, Riedel-de Haen, Sigma-Aldrich); ethyl oleate (Sigma-Aldrich); oleic

acid (Technical grade, 90%, Sigma-Aldrich); N-methyl-N-(trimethylsilyl)

trifluoroacetamide (for synthesis, Merck KGaA); methyl margarate (reference substance

for GC, Merck KGaA); tricaprin (MP Bimedicals, Inc.); methanol (GR for analysis, Merck

Pty. Limited, Kilsyth, Vic); ethanol (absolute, ACS reagent, Sigma-Aldrich, Sheboygan

Falls WI); 1-propanol (laboratory reagent, Ajax Chemicals); 1-pentanol (laboratory

reagent, Ajax Chemicals); 1-hexanol (purum, GC, Fluka); 1-octanol (puriss, GC, Fluka);

tert-butanol (analytical reagent, Fisons); 2-ethyl hexanol (puriss, GC, Fluka); glycerol (GR

for analysis, Merck Pty. Limited).

3.2 Equipment

3.2.1 Chemical synthesis

Large scale synthesis (>200mL) was performed in a 1L glass kettle with 4 Quickfit style

ports (Figure 3.1). One port was fitted with a Friedrichs condenser to return any vapour to

the reactor. A Heidolph stirrer/hotplate (Part No. MR 3001 K) provided heat and agitation


26

via a Teflon coated stirrer bar. A Heidolph temperature controller (Part No. EKT 3001)

with a Pt 100 stainless steel temperature probe was used to control temperature. Small

batches or syntheses that did not produce significant quantities of vapour were produced in

a conical flask of the appropriate size using the aforementioned hotplate/stirrer and

controller.

Figure 3.1: Kettle reactor with stirrer/hotplate for synthesis

Purification steps including glycerol and alcohol removal via water washes, neutralisation,

drying and filtration. These operations were performed with standard laboratory glassware

including: separating funnels, beakers and conical funnels.

3.2.2 Analytical Equipment

Analysis of biodiesel and the associated derivatives was largely performed by gas

chromatography (GC). Two different gas chromatographs with different detectors were

used depending on the type of analysis required. Analysis of biodiesel content and


27

contaminants such as glycerol, mono-, di- and triglycerides was performed on the School

of Chemical Engineering Perkin Elmer Clarus GC 500 with flame ionisation detector

(FID). Analysis of epoxy and alkoxy biodiesel and associated compounds was performed

on the School of Petroleum Engineering Perkin Elmer Clarus GC 500 coupled with a mass

spectrometer, as was fatty acid content. Associated hardware including columns and

conditions for each analysis is detailed in section 3.6.

Cloud point and pour point determinations were performed according to ASTM D 2500

using the test jar shown in Figure 3.2. A refrigerated water bath (Julabo model F34)

containing a 72 vol% propylene glycol solution set at -18°C was used for the controlled

cooling of the test sample as outlined in ASTM D 2500. Replicates of cloud point tests on

the same material were performed to determine the repeatability of the method, with

measurements generally accurate to within 1K of each other, allowing for meaningful

comparisons to be made. The significance of a 1K difference in CP may therefore be

considered marginal. Pour point (PP) was reported to the nearest 3°C as required by the

ASTM D 97.

Figure 3.2: Test apparatus for cloud point determination

Dynamic viscosity was determined with a Haake VT 550 concentric cylinder viscometer at

a temperature of 40°C as recommended in ASTM D 445. A type MV rotor was used with a

1.0 mm gap and the determination was conducted in controlled shear rate mode from 60 to


28

150 s-1, with a dwell time of 200 s. Density of the samples was determined at 40°C with a

25 mL pycnometer in order to convert dynamic viscosity to kinematic viscosity.

Infrared analysis of samples for qualitative purposes was performed on a Thermo Scientific

Nicolet 6700 with Thermo Electron’s OMNIC spectroscopy software. Analysis was

performed with the Smart Orbit ATR sampling unit.

3.3 Biodiesel Synthesis

3.3.1 Methyl and Ethyl Biodiesel

The synthesis route for methyl and ethyl biodiesel was identical except for the temperature,

which was 60°C and 70°C, respectively. Canola oil was transesterified via the well

established homogeneous base catalytic method using potassium hydroxide as catalyst (2

wt%). The molar ratio of alcohol to oil was 7:1 and the residence time was 1h. The

reaction products were cooled before transfer to a separating funnel. A small quantity

(typically 100mL) of deionised water was added to facilitate glycerol settling and phase

separation. After allowing time to settle, the lower aqueous layer was drained, taking most

of the glycerol and alcohol with it. The remaining solution was neutralised with 96%

sulfuric acid as confirmed by universal pH indicator papers. Several water washes

followed to remove residual glycerol and alcohol, until no alcohol odour was evident. The

top organic layer was dried with anhydrous sodium sulfate and filtered under gravity

through Whatman No. 54 filter paper.

3.3.2 Butyl Biodiesel

Butyl biodiesel was synthesised in a similar fashion and in identical equipment to that used

for methyl and ethyl biodiesel but employed a two stage transesterification. Potassium

hydroxide at 2 wt% and butanol at a molar ratio of 7:1 was used for both stages. At the

completion of the first stage, the reactor contents were cooled then transferred to a

separating funnel. A quantity of deionised water equivalent to approximately half of the

total reactor contents was added to facilitate phase separation. The lower layer was

removed and discarded after a few hours settling time, taking most of the glycerol with it.

Another aliquot of water was added and the mixture was left to settle overnight. Again, the

lower phase was drained and discarded. The top phase was transferred back to the reactor


29

for the second transesterification with the same quantities of alcohol and catalyst as was

added during the first stage.

In the preliminary work described in section 4.0 a temperature of 110°C was chosen to

mirror the methyl and ethyl reaction temperatures which were a few degrees below the

normal boiling point of the respective alcohol. For the later work described in section 5.0, a

temperature of 80°C was chosen as adequate considering the long residence time and the

two-stage nature of the process. In all cases, the product was analysed and found to be

>99% butyl biodiesel, that is, <1% mono-, di- and triglyceride.

Purification of the product following the second stage was identical to that for methyl and

ethyl biodiesel, except that the catalyst was not neutralised with acid. Instead, repeated

water washes were employed to remove the catalyst and alcohol. Water washes were

repeated until the aqueous lower phase was clear and the odour of butanol was absent

(typically 14 - 16 aliquots of water). The reasons for removing catalyst and alcohol with

water washes only are described below.

Early attempts to speed up the purification process by neutralising the catalyst and

evaporating the excess solvent were unsuccessful. The temperatures required to remove the

majority of butanol were in the order of 140 – 150°C as the fraction of butanol reduced.

Any residual acid from the crude neutralisation step (not perfectly neutral) resulted in the

formation of dibutyl ether at the elevated temperature. On some occasions gel-like material

was also formed due to other undesirable side-reactions. It was therefore decided that

heating the product with an uncertain quantity of either acid or alkali catalyst was not wise.

Due to the partial solubility of butanol in both water and biodiesel, a large number of water

washes were required to reduce the butanol level to an acceptable value.

Butanol also has an affinity for glycerol and can therefore act as a co-solvent making it

more difficult to wash glycerol from the biodiesel phase. This is one of the contributing

factors to the relatively high content of mono- and di-glycerides encountered in butyl

biodiesel when only a single step transesterification is undertaken. Removal of the majority

of the glycerol produced during the first transesterification step significantly reduces the

opportunity for reversal of the reaction.


30

3.4 Epoxidation Epoxidation was performed on a temperature controlled hotplate/stirrer via the in-situ

peroxyacid method with either formic acid or acetic acid. Residual acid and peroxide were

neutralised with sodium bicarbonate solutions, followed by several water washes and

drying over anhydrous sodium sulfate. Formic acid does not require an acid catalyst for the

formation of the peroxyformic acid intermediate whereas acetic acid does. Peroxyformic

acid therefore formed from the addition of formic acid and dropwise addition of 30 %v/v

hydrogen peroxide (Equation 3.1).

CH2O2 + H2O2 CH2O3 + H2O (3.1)

Formation of peroxyacetic acid was catalysed by the addition of 2 wt% sulfuric acid

(Equations 3.2 and 3.3).

CH3CO2H + H+ CH3CO2H2+ (3.2)

CH3CO2H2+ + H2O2 CH3CO3H2

+ + H2O (3.3)

In both cases, hydrogen peroxide was added dropwise with the reaction mixture at room

temperature to minimise temperature rise (negative reaction enthalpy). Residence time

began when the first drop of hydrogen peroxide was added. Once all peroxide was added,

the temperature set point was raised to the final desired value.

The main product was epoxy biodiesel with oxirane rings at the positions of double bonds,

of which there are many, including esters with multiple points of unsaturation. Potential

by-products included keto compounds as a result of re-arrangement of the oxirane group or

vicinal di-hydroxy compounds as a result of oxirane hydrolysis (Figure 3.3).


31

Figure 3.3: Reaction scheme for the epoxidation of biodiesel (methyl) including the main product and

possible by-products: a – methyl oleate, b – 9,10-epoxy methyl stearate (product), c – 9,10-dihydroxy methyl

stearate (by-product), d – 9(10)-keto methyl stearate (by-product).

3.5 Alkoxylation Oxirane ring opening and subsequent addition of alkyl groups was performed in a glass

reaction kettle on a hotplate/stirrer as detailed in section 3.2.1. Sulfuric acid catalyst at a

molar ratio of 2.5 wt%, an alcohol molar ratio of 14:1 and a reaction time of 1h were the

conditions chosen after the initial batch to determine optimal residence time. These

conditions were used throughout the preliminary studies but were varied as indicated in

section 5.0 for the optimisation exercise. Residual catalyst and alcohol were removed by

repeated water washes and phase separation. Initial water washes included sodium

bicarbonate to neutralise the residual acid. The quantity of bicarbonate was gradually

+

+

a

b

c

d


32

reduced to zero over the first 4 - 5 washes. The remaining washes were bicarbonate free to

ensure that no bicarbonate remained. The final product was dried with anhydrous sodium

sulfate and was filtered to remove the sodium sulfate.

The main product was alkoxy biodiesel due to the addition of the alcohol at the oxirane

ring. Because of the mild conditions, only one alkoxy group was added with the other

position � to the alkoxy being a hydroxy group in the trans- conformation (e in figure 3.4).

As for epoxidation, potential by-products included keto compounds as a result of re-

arrangement of the oxirane group or vicinal dihydroxy compounds as a result of oxirane

hydrolysis (Figure 3.4).

Figure 3.4: Reaction scheme for the alkoxylation of epoxy biodiesel (methyl) including the main product and

possible by-products: b – 9,10-epoxy methyl stearate, c – 9,10-dihydroxy methyl stearate (by-product), d –

9(10)-keto methyl stearate (by-product), e – 9(10)-hydroxy,10(9)-methoxy methyl stearate (product).

H2SO4 ROH Heat

+

+

b

e

c

d


33

3.6 Analytical Methods

3.6.1 Biodiesel Purity

Biodiesel purity was confirmed by gas chromatographic analysis, in accordance with The

European Standard EN14105 (Fat and Oil Derivatives – Fatty Acid Methyl Esters (FAME)

– Determination of Free and Total Glycerol and Mono-, Di-, Triglyceride Contents

(Reference Method)), using a Perkin Elmer Clarus 500 with flame ionisation detector

(FID). Glycerol, mono-, di- and triglyceride concentrations were determined following

derivatisation with N-methyl-N-(trimethylsilyl) trifluoroacetamide (MSTFA). A 15 m

Perkin Elmer Elite - 5HT capillary column was used with H2 as the carrier gas. Aliquots of

1 �L were injected onto the column via a split/splitless injector at 300°C. Hydrogen flow

rate was 2 mL.min-1 and the detector temperature was 380°C. The oven program was 1 min

at 50°C, 15°C.min-1 to 180°C, 7°C.min-1 to 230°C, 20°C.min-1 to 380°C, maintained for 10

minutes. Mono-, di- and triglycerides were quantified with reference to a known quantity

of tricaprin added as an internal standard.

3.6.2 Biodiesel Fatty Acid Profile

The fatty acid profile of biodiesel produced from canola oil was determined by gas

chromatography-mass spectrometry (GC-MS) on a Perkin Elmer Clarus 500 with an

electron ionisation mass spectrographic detector. The capillary column was a 30 m Perkin

Elmer Elite - 5MS with 0.25 mm i.d. and the carrier gas was helium at a flow rate of 0.9

mL.min-1. The oven temperature program was based on the program reported by Wilson et

al. (1997) and was: 100°C for 3 min, 25°C.min-1 to 170°C, 2°C.min-1 to 260°C, maintained

for 5 min (Wilson et al. 1997).

Representative samples of the biodiesel produced by the transesterification of canola oil

with methanol to form methyl esters of the triglycerides were analysed. The relative

fractions of each fatty acid were calculated from the fraction of total peak area attributed to

each by the mass spectrum libraries supplied with the GC-MS software. The fatty acid

profile was compared to literature values to validate the determination.


34

3.6.3 Epoxy and Alkoxy Biodiesel

3.6.3.1 Gas Chromatography-Mass Spectrometry (GC-MS) Epoxy and alkoxy biodiesel were analysed by gas chromatography-mass spectrometry

(GC-MS) on a Perkin Elmer Clarus 500 with an electron ionisation mass spectrographic

detector. The capillary column was a 30 m Perkin Elmer Elite - 5MS with 0.25 mm i.d. and

the carrier gas was helium at a flow rate of 0.9 mL.min-1. The oven temperature program

was identical to that used for the fatty acid profiling detailed in section 3.6.3 and was:

100°C for 3 min, 25°C.min-1 to 170°C, 2°C.min-1 to 260°C, maintained for 5 min (Wilson

et al. 1997). Relative quantities were determined from the relative peak area normalised to

the alkyl palmitate peak. This proved to be a reliable method since alkyl palmitate is a

saturated ester and does not take part in either reaction.

3.6.3.2 Gas Chromatography-Flame Ionisation Detector (GC-FID)

Epoxy and alkoxy biodiesel were analysed by gas chromatography-flame ionisation

detector (GC-FID) on the same Perkin Elmer Clarus 500 as used for biodiesel purity in

section 3.6.1. The method employed was based on European Standard EN 14103 (Fat and

Oil Derivatives – Fatty Acid Methyl Esters (FAME) – Determination of Ester and

Linolenic Acid Methyl Ester Contents) which is designed for biodiesel content

determination. A 30 m Perkin Elmer Elite – WAX polyethylene glycol capillary column

was used with H2 as the carrier gas. Aliquots of 1 �L were injected onto the column via a

split/splitless injector at 250°C. Hydrogen flow rate was 1.75 mL.min-1, the split flow was

80 mL.min-1 and the detector temperature was 250°C. The oven program was 1 min at

180°C, 8°C.min-1 to 210°C, maintained for 19 minutes.

3.6.3.3 Fourier Transform Infrared Spectroscopy (FTIR) A Thermo Scientific Nicolet 6700 Fourier Transform Infrared (FTIR) spectrometer

instrument was used for qualitative identification of functional groups associated with the

epoxidation and alkoxylation products. A drop of the neat sample was placed on the Smart

Orbit ATR crystal and 32 scans were performed to produce a spectrum from 4000 to 500

cm-1. The OMNIC software was used to subtract two or more spectra or simply to overlay

two or more spectra for comparison.


35

3.6.4 Cloud Point

Cloud point was determined visually as the appearance of a distinct clouding of a

30 - 50 mL sample in a 150 mL test jar (Figure 3.2). A cloud formed at the base of the test

jar was indicative of a legitimate cloud point of the sample. Homogeneous clouding of the

entire sample indicated that moisture (water) may be present in the sample. In that case the

sample was re-dried with anhydrous sodium sulfate and subjected to a repeat test. The

equipment and method were according to ASTM D 2500 using a Julabo F34 water bath

containing a propylene glycol solution set at -18°C.

3.6.5 Pour Point

Pour point was an extension of the cloud point determination. The sample was cooled

further, removing the jar at temperatures that were multiples of 3°C, beginning at 0°C. The

temperature at which the sample failed to flow when the test jar was held in a horizontal

position for 5 seconds was recorded. The reported pour point was that temperature plus

3°C, as stated by the standard test method for cloud point (ASTM D 2500).

3.6.6 Viscosity

The dynamic viscosity of a 30 mL sample of biodiesel was determined in the apparatus

described in section 3.2.2. The temperature of the apparatus was maintained at the test

temperature of 40°C by flowing water through the jacket of the test cup. An equilibration

time of at least 15 minutes for the sample to reach steady-state was employed and the

temperature was confirmed with a PT-100 thermocouple. Kinematic viscosity was

calculated from the dynamic viscosity and the density also measured at 40°C with a 25 mL

pycnometer.

3.6.7 Free Fatty Acid Content of Canola Oil

The free fatty acid (FFA) content of the canola oil feedstock for biodiesel was determined

using the titrimetric method specified by ASTM D 5555-95 (Standard Test Method for

Determination of Free Fatty Acids Contained in Animal, Marine, and Vegetable Fats and

Oils used in Fat Liquors and Stuffing Compounds). The method performed for canola oil

was identical to the standard except that potassium hydroxide was substituted for sodium

hydroxide.

Chapter 4

Initial Synthesis and Characterisation

Chapter 4 Initial Synthesis and Characterisation _________________________________________________________________________

37

INITIAL SYNTHESIS AND CHARACTERISATION _________________________________________________________________________

The initial stages of the process of investigation centred on the conversion of refined

canola oil which represents the prime raw material for biodiesel production in the region.

The first exploratory batches converted canola oil directly to epoxy biodiesel followed by

acid-catalysed transesterification to produce epoxy ethyl biodiesel. Reaction conditions

were similar to those proposed by La Scala and Wool (La Scala and Wool 2002) and Pages

and Alfos (Pages and Alfos 2001). Preliminary GC results indicated approximately 40%

conversion of the C18:1/3 fraction, based on GC-FID data. This initial work provided a

platform to further develop the processes and procedures for synthesis and the associated

analytical methods. Later trials were preceded by the conversion of canola oil directly to

biodiesel and the conversion of biodiesel to epoxy and subsequently alkoxy biodiesel. This

shifted the focus to the direct modification of biodiesel, where it needed to be.

While GC-FID was able to accurately determine the extent of conversion of the C18:1/3

fraction by the reduction in area of that peak(s), it was not able to adequately resolve the

product peak and more importantly, could not identify the product peak as epoxy or alkoxy

biodiesel. There are also no commercially available reference standards for epoxy or

alkoxy biodiesel. To complicate matters further, biodiesel derived from canola oil is a

complex mixture of many different fatty esters with multiple double bonds (reaction sites).

This means that there are many different products possible that may not be completely

resolved by the GC method. A rigorous program to positively identify the reaction

products (epoxy and alkoxy biodiesel) was therefore undertaken.

Outlined here is the procedural method development for the synthesis and analysis of

biodiesel, including methyl, ethyl and butyl biodiesel. This is followed by the initial

synthesis and analysis by GC-FID of epoxy methyl biodiesel and alkoxy methyl biodiesel.

The development of the GC-MS analytical method for the characterisation of epoxy and

alkoxy biodiesel is outlined and finally the use of FTIR to provide some further assurance

of the identity of reaction product is discussed.


38

4.1 Biodiesel Synthesis

4.1.1 Methyl Biodiesel

Before the well-established base-catalysed transesterification of vegetable oil (canola oil)

was performed, the oil had to be confirmed as low in free fatty acid (FFA) since it is

known that excessive FFA can cause problems during purification. Large quantities of

FFA (acid value of >1 wt%) present during base-catalysed transesterification can result in

the formation of soap. The presence of soap can cause an increase in viscosity, the

formation of gels and foams and make the separation of glycerol difficult (Demirbas 2008).

Free fatty acid (FFA) content was determined according to ASTM D 5555-95 and was

found to be 0.2 mg KOH/g oil, equivalent to 0.1 wt% FFA.

4.1.1.1 Initial Procedure A catalyst (KOH) concentration of 2 wt%, a molar ratio of methanol to oil of 7:1 and a

reaction time of 1h at 65°C was employed following the procedure outlined below:

1. Pre-heated the canola oil to 60°C in the reactor described in section 3.2.1, including

the condenser,

2. Pre-weighed and mixed the potassium hydroxide and methanol,

3. Added the KOH/methanol to the stirred reactor and set the temperature set-point to

65°C,

4. Began timing for the reaction once the KOH/methanol were added,

5. Removed the reactor from the hotplate/stirrer after 1h and began to cool under

running tap water,

6. Returned reactor to hotplate once cooled and added sulfuric acid to neutralise the

catalyst using universal indicator paper,

7. Transferred to a 1L separating funnel and added a small quantity of RO water

equivalent to approximately one half the total volume to aid phase separation,

8. Once settling was complete (approximately 2h), drained and discarded the lower

phase,

9. Repeated the addition of water amounting to approximately the same volume as the

unrefined biodiesel, mixed and allowed to settle,


39

10. Drained the lower phase and discarded. Continued to wash as in 9. until the odour

of methanol was absent and the pH was neutral,

11. The top phase was dried with excess anhydrous sodium sulfate, then filtered

through Whatman No. 54 filter paper into a storage bottle.

4.1.1.2 Analysis The GC-FID method of analysis was according to EN 14105 as outlined in section 3.6.1,

which provides a sample chromatogram including the likely elution times for the various

impurities. However, there are natural variables between instruments, columns and

samples that necessitate the identification of the elution times for the various species of

interest. To that end, standard solutions of glycerol, monoolein, diolein and triolein were

injected onto the column at the specified conditions to provide the approximate ranges for

elution of glycerol, mono-, di- and triglycerides in a real biodiesel sample (Figure 4.1). The

approximate elution times for the contaminants were: Glycerol elutes at 4.6 minutes,

monoglycerides at 15.5 minutes, diglycerides at 21.5 minutes and triglycerides at 24.5

minutes.

Figure 4.1: Chromatogram of standard solutions for contaminants associated with EN 14105. Retention

times are shown in minutes at the top of the figure.

Glycerol Monoglycerides

Diglycerides

Triglycerides


40

A 200 �L sample of the biodiesel product was removed and precisely weighed into a

sample vial. A 100 �L aliquot of tricaprin standard was added, as was 100�L of MSTFA.

The sample was mixed and allowed to stand at room temperature to react for 15 to 20

minutes. The sample was diluted with 5 mL of n-heptane, injected into the GC-FID and

analysed for purity as detailed in section 3.6.1.

As is evident from Figure 4.2, the purification process was successful in that glycerol is

virtually absent. However, the peak at 15.79 min. is very large and represents a

monoglyceride content of 1.7%. Another feature of the chromatogram in Figure 4.2 is the

low di- and triglyceride content, which elute at approximately 21.5 and 24.5 mins.,

respectively.

Figure 4.2: GC-FID Chromatogram of methyl biodiesel with low glycerol and high monoglyceride.

Glycerol

Monoglycerides


41

Transesterification is a stepwise, reversible reaction (Figure 4.3). Since the di- and

triglyceride contents were very low, conversion was clearly high but some reversal of the

equilibrium reaction was suspected. The purification procedure was therefore altered to

remove most of the glycerol prior to neutralisation of the catalyst.

Triglyceride + ROH � Diglyceride + RCOOR1 Diglyceride + ROH � Monoglyceride + RCOOR2 Monoglyceride + ROH � Glycerol + RCOOR3

Figure 4.3: Transesterification reaction mechanism.

4.1.1.3 Modified Procedure The same reaction conditions as for 4.1.1.1 were employed but with a modified

purification procedure:


the condenser,

2. Pre-weighed and mixed the potassium hydroxide and methanol,

3. Added the KOH/methanol to the stirred reactor and set the temperature set-point to

65°C,

4. Began timing for the reaction once the KOH/methanol were added,


running tap water,




phase including most of the glycerol,

8. Repeated the addition of water as in 7., mixed and allowed to settle,

9. Drained the lower phase and discarded,

10. The top layer was neutralised with sulfuric acid,




42

12. Continued to wash as in 11. until the odour of methanol was absent and the pH was

neutral,

13. The top phase was dried with excess anhydrous sodium sulfate then filtered through

Whatman No. 54 filter paper into a storage bottle.

4.1.1.4 Analysis A 200 �L sample of the product was removed and treated exactly as detailed in section

4.1.1.2. The monoglyceride content was 0.5% and glycerol, di- and triglycerides were not

detected (Figure 4.4). The procedure outlined in section 4.1.1.3 was adopted for the

synthesis of methyl biodiesel via the base-catalysed transesterification of canola oil.

Figure 4.4: GC-FID Chromatogram of methyl biodiesel with low monoglyceride.


43

4.1.2 Ethyl Biodiesel

The same reaction conditions as for methyl biodiesel were used, except that the reaction

temperature was increased to 70°C.

4.1.2.1 Detailed Procedure A catalyst (KOH) concentration of 2 wt%, a molar ratio of methanol to oil of 7:1 and a



the condenser,

2. Pre-weighed and mixed the potassium hydroxide and ethanol,

3. Added the KOH/ethanol to the stirred reactor,

4. Began timing for the reaction once the KOH/ethanol were added,


running tap water,







10. The top phase was neutralised with sulfuric acid,



12. Continued to wash as in 11. until the odour of ethanol was absent and the pH was

neutral,




4.1.1.2. The monoglyceride content was <1.0% and glycerol, di- and triglycerides were not

detected (data not shown). The same procedure as for methyl biodiesel except for the


44

increased reaction temperature was therefore adopted for the synthesis of ethyl biodiesel

via the base-catalysed transesterification of canola oil.

4.1.3 Butyl Biodiesel

The same reaction conditions as for methyl and ethyl biodiesel were attempted, except that

the temperature was increased to 110°C and was maintained for 4h. A catalyst (KOH)

concentration of 2 wt%, a molar ratio of butanol to oil of 7:1 and a reaction time of 4h at

110°C was employed.

4.1.3.1 Initial Procedure A similar procedure as for ethanol was employed but due to the much lower solubility of

butanol in water, residual butanol was evaporated:

1. Pre-heated the canola oil to 110°C in the reactor described in section 3.2.1,

including the condenser,

2. Pre-weighed and mixed the potassium hydroxide and butanol,

3. Added the KOH/butanol to the stirred reactor,

4. Began timing for the reaction once the KOH/butanol were added,


running tap water,







10. The top phase was neutralised with sulfuric acid,

11. Transferred the top phase back to the hot-plate without the lid or condenser and

heated to drive-off the residual butanol,

12. Cooled, then transferred to a 1L separating funnel and added an equal quantity of

RO water mixed and settled,


phase,



45


16. Continued to wash as in 14 - 15. until the odour of butanol was absent and the pH

was neutral,




4.1.1.2. The monoglyceride content was 9.8% and diglyceride was 1.4% while glycerol

and triglyceride were absent (Figure 4.5). There was also evidence of polymerisation

products due to the high temperatures experienced during the butanol evaporation step.

Figure 4.5: Chromatogram of butyl biodiesel with high mono- and diglycerides.

Monoglycerides

Diglycerides


46

4.1.3.3 Modified Procedure Clearly the procedure in 4.1.3.2 was not effective. Several other techniques were attempted

including a two-stage transesterification, the omission of the catalyst neutralisation step

and the replacement of the evaporation step for the removal of residual butanol with

repeated water washes. The procedure settled upon that produced satisfactory results for

contaminant levels, in-line with methyl and ethyl biodiesel, equating to a biodiesel content

of >99% is detailed below. The procedure included two 4h transesterification steps with a

glycerol removal step between them. Repeated water washes (approximately 15) were then

used after the second transesterification to remove the residual catalyst and butanol. It was

also found that the modified procedure allowed for the reaction temperature to be lowered

to 80°C with acceptable results.


the condenser,

2. Pre-weighed and mixed the potassium hydroxide and butanol,

3. Added the KOH/butanol to the stirred reactor,

4. Began timing for the reaction once the KOH/butanol were added,


running tap water,







10. Transferred the top phase back to the hot-plate and pre-heated to 80°C,

11. Added the same quantities of KOH/butanol as in step 3.,


running tap water,

13. Transferred to a 1L separating funnel and added an equal quantity of RO water

mixed and settled,


phase,



47


17. Continued to wash as in 15 - 16. until the odour of butanol was absent and the pH

was neutral,




4.1.1.2. The monoglyceride content was typically 0.7% and glycerol, di- and triglycerides

were not detected (Figure 4.6). The procedure outlined in section 4.1.3.3 was adopted for

the synthesis of butyl biodiesel via the base-catalysed transesterification of canola oil.

Figure 4.6: Chromatogram of butyl biodiesel with low monoglyceride.


48

4.2 Epoxidation of Methyl Biodiesel The first batches of epoxy methyl biodiesel were produced from methyl biodiesel via the

in-situ peroxyformic acid method described in section 3.4. Analysis of samples was

performed by GC-FID based on EN 14103 for biodiesel quantitation. The method was able

to verify the conversion of biodiesel but was not able to confirm the identity of the reaction

products.

4.2.1 24 Hour Epoxidation of Methyl Biodiesel

A molar ratio of formic acid to biodiesel of 0.3:1, a molar ratio of hydrogen peroxide to oil

of 2.6:1 and a reaction time of 24h at 60°C was employed following the procedure outlined

below:

1. Pre-heated the combined biodiesel and formic acid to 50°C in an Erlenmeyer flask

on the hot-plate/stirrer,

2. Added the 30 vol% hydrogen peroxide dropwise while monitoring the temperature

to guard against a temperature overshoot or runaway reaction. The reaction time

began once the first drop of H2O2 was added,

3. All H2O2 was added over a period of approximately 10 minutes,

4. The temperature set point was then raised to 60°C,

5. Samples were taken regularly throughout the reaction for GC analysis,


running tap water,


saturated with sodium bicarbonate,


phase,



10. Drained the lower phase and discarded, continued to wash as in 9. until the pH was

neutral,




49

4.2.2 GC-FID Analysis

The GC-FID method of analysis was according to EN 14103 as outlined in section 3.6.3.2.

Samples were prepared as follows:

1. Sample vials (10 mL) were prepared by adding approximately 2mg of NaHCO3 and

2 mL of RO water,

2. Approximately 1 mL of sample was removed directly from the reactor, added to the

vial and immediately shaken to neutralise the reaction,

3. Samples were allowed to settle while the reaction continued,

4. 150 �L was weighed accurately into another 10 mL vial,

5. 5 mL of a known concentration of methyl heptadecanoate (IS) solution in heptane

was added to each vial,

6. Approximately 1 mL was transferred to a 2 mL GC vial for analysis.

Figure 4.7 is the chromatogram of methyl biodiesel, including a large peak for the methyl

margarate internal standard (IS) at 4.5 minutes. The peaks attributed to C18 methyl

biodiesel are at 5.43 to 6.30 minutes.

Figure 4.7: GC-FID chromatogram of methyl biodiesel analysed according to EN 14103.

C17 (IS)

C18 biodiesel


50

Conversion of biodiesel was estimated from the reduction in area of the C18 peaks,

referenced to the IS peak and the known masses (refer to Figure 4.8). The large peak at

10.05 minutes can be attributed to epoxy methyl biodiesel. Conversion was estimated to be

92% after 24h (Figure 4.9).

Figure 4.8: GC-FID chromatogram of methyl biodiesel epoxidised for 24h analysed according to EN 14103.

Figure 4.9: Conversion of methyl biodiesel to epoxy methyl biodiesel over time.

92.39

0

10

20

30

40

50

60

70

80

90

100

0 3 6 9 12 15 18 21 24

Conv

ersi

on o

f Uns

atur

ated

Bio

dies

el (%

)

ReactionTime (h)

C17 (IS)

C18 biodiesel

Epoxy methyl biodiesel


51

4.3 Alkoxylation of Epoxy Methyl Biodiesel The first batches of alkoxy methyl biodiesel were produced from epoxy methyl biodiesel in

the presence of an excess of methanol and a sulfuric acid catalyst as described in section

3.5. Analysis of samples was performed by GC-FID based on EN 14105, described in

section 3.6 and EN 14103 for biodiesel quantitation. These methods were able to verify the

conversion of epoxy methyl biodiesel but again, were not able to confirm the identity of

the reaction products.

4.3.1 Procedure

A molar ratio of methanol to biodiesel of 9:1, a catalyst concentration of 5 wt% and a


1. Pre-heated the combined epoxy methyl biodiesel and methanol to 50°C in the

reactor described in section 3.2.1, including the condenser,

2. Added the sulfuric acid and began timing,




running tap water,

6. Returned the kettle to the hotplate, added a small quantity of RO water and

neutralised with KOH pellets,

7. Transferred to a 1L separating funnel and added an equal quantity of RO water,


phase,



10. Drained the lower phase and discarded, continued to wash as in 9. until the odour

of methanol was absent,




52

4.3.2 GC-FID Analysis

4.3.2.1 Use of EN 14103 to Determine extent of reaction A 200 �L sample was diluted with 5 mL of heptane and analysed according to EN 14103

as outlined in section 3.6.3.2. A large, broad peak at 16.7 minutes was evident (Figure

4.10) which is indicative of methoxy methyl biodiesel.

Figure 4.10: GC-FID chromatogram of methoxy methyl biodiesel alkoxylated for 24h analysed according to

EN 14103.

4.3.2.2 Determination of Extent of Reaction (EN 14105) A 200 �L sample of the methoxy methyl biodiesel product was removed and precisely

weighed into a sample vial. A 100 �L aliquot of tricaprin standard was added, as was

100�L of MSTFA. The mixture was mixed and allowed to stand at room temperature to

react for 15 to 20 minutes. The sample was diluted with 5 mL of n-heptane, injected into

the GC-FID and analysed for purity as detailed in section 3.6.1.

A large, broad peak at 12.6 minutes was evident, next to the original peak for C18 methyl

biodiesel at 10.1 minutes (Figure 4.11). This demonstrates the conversion of biodiesel but

Remaining C18 methyl biodiesel

Methoxy methyl biodiesel


53

again is only indicative of the presence of methoxy methyl biodiesel. Many other relatively

large peaks exist that are not present for a standard methyl biodiesel batch (cf. Figure 4.4).

Figure 4.11: GC-FID chromatogram of methoxy methyl biodiesel alkoxylated for 24h analysed according to

EN 14105.

4.4 Fourier Transform Infrared Analysis Fourier Transform Infrared (FTIR) analysis of reaction products was performed to aid in

the identification of the reaction products. Several authors have used FTIR qualitatively

and/or quantitatively to monitor the reaction progress of both epoxidation and alkoxylation

reactions (Mungroo et al. 2008), (Hwang and Erham 2001), (Lathi and Mattiasson 2007),

(Lin et al. 2008), (Sharma et al. 2006), (Doll et al. 2007), (Dalhlke et al. 1995), (Moser and

Erhan 2006), (Wang and Zhang 2006). The epoxidation and alkoxylation reaction products

are a complex mixture due to the mixed fatty acid nature of canola oil, making the

elucidation of composition difficult. The reaction products of the epoxidation and

Remaining C18 methyl biodiesel

Methoxy methyl biodiesel


54

alkoxylation of ethyl biodiesel were therefore compared with those for epoxidised and

alkoxylated ethyl oleate.

4.4.1 Synthesis of Epoxy and Alkoxy Ethyl Oleate

Laboratory grade ethyl oleate was subjected to epoxidation at 60°C for 3h. The hydrogen

peroxide to ethyl oleate ratio was 2:1 (molar) with an acetic acid to ethyl oleate ratio of

0.2:1 and the sulfuric acid concentration was 2 wt%. Residual peroxide and acid were

neutralised with repeated bicarbonate washes followed by water only washes, drying over

anhydrous sodium sulfate and filtration.

In order to create the expected reaction by-products, half of the refined material (25 mL)

was added to 80 mL of water and 10 drops of 96% sulfuric acid. The solution was heated

to 60°C for 3h to hydrolyse the epoxy to produce a reference by-product (glycol).

Purification entailed repeated water washes to remove residual catalyst, drying with

sodium sulfate and filtration.

The remaining half was alkoxylated with ethanol at a 6:1 molar ratio and 2.5 wt% sulfuric

acid, at 60°C for 2h. The cooled solution was neutralised with potassium hydroxide prior

to repeated water washes, drying and filtration.

4.4.2 FTIR Analysis

Various samples were analysed including actual biodiesel derivatives and the

aforementioned reference compounds. In some cases spectra of compounds displayed

peaks which were easily attributed to the specific functional groups of interest. For others,

some of the features provided by the software for comparison of the spectra of two or more

compounds were used to identify more subtle features.

Figure 4.12 is an overlay of spectra for the three reference materials synthesised as above

i.e. epoxy ethyl oleate, ethoxy ethyl oleate and glycol ethyl oleate. The dominant feature is

the much larger absorbance band between 3500 and 3000 cm-1 which is typically

associated with hydroxyl groups. There is a much larger absorbance in the case of glycol

ethyl oleate, confirming conversion of the oxirane to glycol under the conditions of acid

hydrolysis.


55

Figure 4.12: Overlay of FTIR spectra of epoxy ethyl oleate, ethoxy ethyl oleate and glycol ethyl oleate.

Hydroxyl groups


56

Figure 4.13 is an overlay of epoxy butyl biodiesel and butoxy butyl biodiesel indicating a

much more prominent absorbance at approximately 3500 cm-1 for butoxy butyl biodiesel,

which is indicative of the additional hydroxyl � to the butoxy group. In addition, there is a

much larger absorbance evident at 1090 cm-1 which has been attributed to a C-O-C

asymmetric stretch by Socrates for ethers (Socrates 2004). A doublet is also observed for

branched ethers, which also correlates with the secondary nature of the alkoxy group

expected for the butoxy butyl biodiesel.

Figure 4.13: Overlay of FTIR spectra of epoxy butyl biodiesel and butoxy butyl biodiesel.

Figure 4.14 is a spectrum representing a batch produced on 7/7/08 with a lower alkoxy

content (44%) subtracted from butoxy biodiesel produced on 22/7/08 with a much higher

conversion (67%). As is evident, the absorbance at 1090 cm-1 is much greater, confirming

the greater content of butoxy and confirming the assignment of that peak to alkoxy.

C-O-C Asymmetric stretch

Hydroxyl groups


57

Figure 4.14: Residue spectrum of the subtraction result of butoxy butyl biodiesel produced 22/7/08 and

butoxy butyl biodiesel produced 7/7/08.

Figure 4.15 is a spectrum representing the residue of butyl biodiesel subtracted from epoxy

butyl biodiesel. A feature is the peak at 822.7 cm-1 which was identified as belonging to the

epoxy group by Mungroo et al. (Mungroo et al. 2008) and Hwang and Erhan (Hwang and

Erhan 2001). The peak at 1717.2 cm-1 may also be attributed to the keto group of the by-

product as suggested by Socrates (2004).

Figure 4.15: Residue spectrum of the subtraction result of butoxy butyl biodiesel produced 22/7/08 and

butoxy butyl biodiesel produced 7/7/08.

Butoxy

Keto group Epoxy


58

4.5 GC-MS Method Development A method using gas chromatography coupled with electron impact mass spectrometry was

chosen for the analysis of epoxy biodiesel because of the ability to simultaneously resolve

and identify the constituent compounds. Most workers in the field of the epoxidation of

fatty acids have used a titrimetric method to determine the loss of double bonds and a

second titrimetric method for the generation of oxirane. Since the ultimate aim was to

produce alkoxy fatty acid esters (biodiesel) and gas chromatography-mass spectrometry

(GC-MS) had been used by other authors for alkoxy fatty acids, GC-MS was chosen for

the epoxy biodiesel as well. Wilson et al. (Wilson et al. 1997) developed a method that

would resolve positional isomers of monohydroxy fatty acids derived from, amongst

others, linoleic acid. Their GC conditions, in particular the oven program, were used as a

basis for this work but required some adjustment due to analyte and equipment differences

(Table 4.1).

Table 4.1: Comparison of GC-MS conditions for Wilson et al. (1997) and this work.

Equipment/Conditions Wilson et al. Current Work

Capillary column CP Sil 19 PE Elite-5MS

dimensions 25 m X 0.25 mm i.d. 30 m X 0.25 mm i.d.

packing 14% cyanopropylmethyl, 86%

dimethyl

5% diphenyl, 95% dimethyl

Sample volume 1 �L 1 �L

Sample preparation Derivatisation of hydroxy groups

with tetramethylammonium

hydroxide

Diluted with heptane

Injector mode Splitless 50:1 split

Injector condition 300°C 250°C

Carrier gas Helium at 0.5 mL.min-1 Helium at 0.9 mL.min-1

Oven program 100°C for 3 min, 25°C.min-1, to

170°C, 2°C.min-1 to 230°C,

20°C.min-1 to 250°C, maintained

for 10 min.

100°C for 3 min, 25°C. min-1, to

170°C, 2°C. min-1 to 260°C,

maintained for 5 min.


59

Wilson et al. (1997) deemed it necessary to derivatise the hydroxy groups of the analyte in

order to achieve adequate resolution of positional isomers. Derivatisation of the

epoxy/alkoxy material was initially performed in this work with N-methyl-N-

(trimethylsilyl) trifluoroacetamide (MSTFA) but was found to be unnecessary for adequate

resolution. The addition of the MSTFA also complicated the mass spectra and increased

the already very large m/z of the fragments.

4.5.1 GC-MS Analysis of Biodiesel

Methyl biodiesel produced by the transesterification of canola oil was analysed by GC-MS

according to the conditions described above and allowed for the elucidation of the fatty

acid profile of canola oil and therefore the biodiesel product (Figure 4.16, Table 4.2).

Table 4.2: Fatty acid profile of biodiesel derived from canola oil (ND = Not Detected).

Fatty Acid Symbol Weight Percentage

Palmitic C16:0 4.0

Palmitoleic C16:1 0.2

Stearic C18:0 2.1

Oleic/Linolenic C18:1/3 73.5

Linoleic C18:2 18.4

Arachidic C20:0 0.5

Eicosanoic C20:1 1.0

Behenic C22:0 0.2

Erucic C22:1 ND

Lignoceric C24:0 0.1

Nervonic C24:1 0.1

The most relevant information here is the mass fraction of unsaturated material at 93.2% or

approximately 93%. Another feature is the fraction of oleic/linolenic acid at 73.5% (Figure

4.17). Oleic and linolenic acids have been combined here because the chromatographic

method does not adequately resolve the two. Mass spectra for the peaks at 22.17 and 22.24

minutes indicate that both peaks are a mixture of C18:1 and C18:3. Because those two

peaks represent the vast majority of the unsaturated material in biodiesel, the reduction in

area of those two peaks was later used and found to be a reliable measure of the conversion

of biodiesel.


60

Figure 4.16: Chromatogram of methyl biodiesel generated on GC-MS.

Figure 4.17: Chromatogram of methyl biodiesel generated on GC-MS showing the C18 peaks. The C18:1/3

peaks are those labelled 22.17 and 22.24 minutes.


61

4.5.2 GC-MS Analysis of Epoxy Biodiesel

Epoxy methyl biodiesel was analysed by GC-MS according to the conditions described

above. As is evident from Figure 4.18, the method developed produced excellent resolution

of unmodified biodiesel and epoxy biodiesel.

Figure 4.18: Chromatogram of epoxy methyl biodiesel. Major peaks: methyl palmitate (16.31 min.), methyl

oleate (22.07 min.), 9,10-epoxy oleate (28.90 min.), 9,10-dihydroxy methyl stearate (35.88 min.).

Mass spectra for the main reaction product (epoxy methyl stearate) and the by-product

(9,10-dihydroxy methyl stearate) are provided in Figures 4.19 and 4.20, including

diagrammatic representations of the fragmentation pattern. Cleavage of epoxy methyl

stearate occurs either side of the oxirane group as identified by Wilson and Lyall (Wilson

and Lyall 2002). The mass spectrum for epoxy methyl stearate (Figure 4.19) was

characterised by a large peak at zm = 155 representing cleavage fragments either side of

carbon 9 on the fatty chain and minor peaks at 171 and 199 for cleavage either side of the

oxirane group. The mass spectrum of the glycol by-product (Figure 4.20) is characterised

by a base peak at zm = 155 as a result of cleavage between carbons 9 and 10 to produce a

Time (min.)


62

secondary peak at 187, and the base peak after the loss of methanol. The large peak at zm

= 138 represents the loss of a hydroxyl group from the fragment between C1 and C9 of the

fatty chain and is easily distinguished from the fragment at zm = 137 for the alkoxy ester

variant in section 4.5.3.

Figure 4.19: Mass spectra of methyl biodiesel derivatives: 9,10-epoxy methyl stearate.

Figure 4.20: Mass spectra of methyl biodiesel derivatives: 9,10-dihydroxy methyl stearate.

- O = 155 171

155

199

m/z

m/z

155 -OH = 138

-CH3OH = 155 187


63

4.5.3 GC-MS Analysis of Alkoxy Biodiesel

Alkoxylation with methanol at a molar ratio of 6:1 in the presence of 10 wt% sulfuric acid

was monitored over a 24h period. The reaction product was analysed by GC-MS as

detailed above. As is evident from Figures 4.21 and 4.22, the method developed produced

excellent resolution of all 3 groups of possible components, that is, unmodified biodiesel,

epoxy biodiesel and methoxy biodiesel.

Figure 4.21: Chromatogram of methoxy methyl biodiesel. Major peaks: methyl palmitate (16.33 min.),

methyl oleate/linolenate (22.08 - 22.16 min.), 9(10)-keto methyl stearate (29.09 min.), 9(10)-hydroxy,10(9)-

methoxy stearate (33.30 min.), 9,10-dihydroxy methyl stearate (35.88 min.).

Mass spectra of the major peaks confirmed these identities. The spectrum for methoxy-

hydroxy methyl stearate (Figure 4.21) included peaks for fragments between carbons 9 and

10 creating ions with zm = 187 and 201 representing the respective positional isomers of

hydroxyl and methoxy at C9. An equally large peak at zm = 155 represents the loss of

methanol from the hydroxyl isomer, while a smaller peak at zm = 169 represents the loss

Time (min.)


64

of a hydroxyl. A large peak at zm = 137 results from the further loss of methanol (Wilson

et al. 1997). The mass spectrum of 9-oxomethyl stearate is characterised by peaks

representing cleavage between C1 and C9 to create fragments at zm = 156 and 125

(Figure 4.23). There is also a distinctive peak at zm = 281 for the loss of methanol from

the ester group.

Figure 4.22: Mass spectrum of 9(10)-hydroxy,10(9)-methoxy methyl stearate.

Figure 4.23: Mass spectrum of 9(10)-keto methyl stearate.

156

281

-CH3 = 141 -CH3OH = 125

m/z

m/z

187

+ CH2 = 215 + CH2 = 229

201

-CH3 = 155

187 187 – OH = 170


65

4.6 Summary Methods developed for the transesterification of canola oil were successful for the

synthesis of methyl, ethyl and butyl biodiesel. This was confirmed by the European

Standard methods of analysis for biodiesel. Epoxidation of biodiesel under the conditions

used for the preliminary experiments resulted in the conversion of a large portion of the

unsaturated fraction. This was confirmed by GC-FID, via two different methods, as the

reduction in area of the peaks identified as oleate, linoleate and linolenate.

While GC-FID was able to accurately determine the extent of conversion of the C18:1/3

fraction by the reduction in area of that peak(s), it was not able to adequately resolve the

product peak and more importantly, could not identify the product peak as epoxy or alkoxy

biodiesel. FTIR was able to positively identify the reaction products as epoxy compounds.

A GC-MS method was successfully developed that was able to resolve the unreacted

biodiesel from the epoxy biodiesel so that reaction progress could be monitored. GC-MS

also provided the facility to positively identify the alkoxylation reaction products,

including those of the reaction by-products.

GC-MS proved to be an accurate and reliable quantitative method for the determination of

conversion, selectivity and the impurity profile of both epoxy and alkoxy biodiesel. This

was only possible however once some confidence in the qualitative identity of the species

was confirmed. FTIR was used as a secondary qualitative measure to provide some initial

certainty that the reaction schemes were successful. FTIR was able for example to confirm

that alkoxy compounds were present due to the existence of bands for secondary ethers and

was able to confirm the presence of glycol by-product.

Chapter 5

Preliminary Studies

Chapter 5 Preliminary Studies _________________________________________________________________________

67

PRELIMINARY STUDIES _________________________________________________________________________

Preliminary studies were conducted to assess the feasibility of the two-step process of

converting biodiesel to alkoxy biodiesel without the use of harsh reaction conditions or

organic solvents. Most of the prior work in this area included conditions of high

temperature and/or pressure, long reaction times (up to 10 days) or organic solvents to

minimise by-product formation. The GC-MS method developed in Chapter 4 was used to

confirm conversion and selectivity for both epoxy and alkoxy biodiesel.

This chapter contains the early work conducted to optimise the epoxidation of methyl

biodiesel, specifically the trade-off between reaction rate and epoxy selectivity. The

optimal conditions were then applied to ethyl and butyl biodiesel. Crude alkoxylation of

the resulting epoxy biodiesel was then performed in order to determine the impact of

alkoxy biodiesel on cloud point. A further extension of this work was to explore the impact

of alkoxy content on cloud point.


68

5.1 Epoxidation of Methyl Biodiesel Epoxidation of fatty acids via the in-situ peroxyacid method is commonly performed with

either formic acid or acetic acid, the latter requiring an acid catalyst to increase the rate of

peroxyacetic acid formation. The formic acid method is known to proceed at a faster rate

than the acetic acid method but the former provides an environment more conducive to the

hydrolysis of oxirane to form by-product, especially in the presence of aqueous hydrogen

peroxide. To maximise selectivity for epoxy biodiesel the molar ratio of organic acid was

kept low. Normally, the reaction is performed in an organic solvent to limit by-product

formation but in this case a concerted effort was made to limit the potential increase in

processing cost given the already marginal unit cost of biodiesel. Epoxidation was

therefore performed without the addition of an organic solvent. A further consideration for

a commercial process is hazard minimisation. To that end, the hydrogen peroxide

concentration was restricted to 30 %v/v.

5.1.1 Method

Initially, epoxidation of methyl biodiesel was performed via the in-situ peroxyformic acid

method. Reaction conditions were: molar ratios of 0.5 and 2 for formic acid and hydrogen

peroxide to biodiesel, respectively at a temperature of 60°C (Trial 1, Table 5.1). The

procedure employed is outlined below:

1. Pre-heated the combined biodiesel and formic acid to 50°C in an Erlenmeyer flask

on the hot-plate/stirrer,

2. Added the 30 vol% hydrogen peroxide dropwise while monitoring the temperature

to guard against a temperature overshoot or runaway reaction. The reaction time

began once the first drop of H2O2 was added,

3. All H2O2 was added over a period of approximately 10 minutes,




running tap water,


saturated with sodium bicarbonate,


69


phase,


unrefined epoxy biodiesel, mixed and allowed to settle,

10. Drained the lower phase and discarded, continued to wash as in 9. until the pH was

neutral,



As the bulk of the unsaturated fraction of biodiesel derived from canola oil is alkyl

oleate/linolenate, conversion of alkyl oleate/linolenate ( BX ) to epoxy biodiesel was

monitored by the reduction in area of the alkyl oleate/linolenate peaks referenced to the

alkyl palmitate peak (Eq. 5.1). Because alkyl palmitate is saturated and does not participate

in the reaction, this method proved to be reliable.

1001(%)10

01 ��

��

��

��

��

��PB

PBB AA

AAX (5.1)

Epoxy selectivity ( ES ) was calculated from the relative areas of the peak representing

9,10-oxirane-alkyl stearate versus 9,10-dihydroxy-alkyl stearate (glycol) (Equation 5.2).

100��

��

�

�GE

EE AA

AS (5.2)

Glycol content ( Gm ) in the epoxy biodiesel was calculated from the selectivity for epoxy

biodiesel and the fraction of unsaturated fatty acid (Equation 5.3).

10093.01001(%) ��

��

��

� �� E

GSm (5.3)


70

5.1.2 Results and Discussion

Conversion of the unsaturated content was almost complete after 24h at 91.3%, however

the selectivity for epoxy biodiesel continued to decline over time and was only 91% after

24h (Figure 5.1). Vicinal dihydroxy compounds (glycol) were identified as the main by-

product, as discussed in chapter 4. The use of formic acid as the oxygen carrier for

epoxidation has previously been linked with hydrolysis of the oxirane group but is not

normally an issue where the end product is polyols. In this case, hydrolysis should be

minimised in order to maximise the selectivity for alkoxy biodiesel.

A reaction time of approximately 6 – 9h appears to result in a reasonably high selectivity

of approximately 98% and a conversion of 85 – 89%.

Figure 5.1: Conversion of the unsaturated portion of methyl biodiesel to epoxy methyl biodiesel. Reaction

conditions: molar ratio of 0.5 and 2 for formic acid and hydrogen peroxide to biodiesel, respectively;

temperature of 60°C. Selectivity for epoxy methyl biodiesel (red curve).

0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

100.0

0 4 8 12 16 20 24

Conv

ersi

on (%

)

Reaction Time (h)


71

5.2 Optimisation of the Epoxidation Step An optimisation exercise was conducted to maximise epoxy selectivity whilst maintaining

high conversion. Temperature was eliminated as a variable by fixing it at 60°C and an

optimal residence time of 6 - 9h was chosen based on the previously observed trends for

conversion and selectivity (Figure 5.1). Three different molar ratios of formic acid were

trialled, followed by two batches using acetic acid as the oxygen carrier, requiring the use

of sulfuric acid as a catalyst (Table 5.1).

Table 5.1: Optimisation of epoxy selectivity: reaction conditions.

Trial

Number

Oxygen

Carrier

Molar Ratio

Oxygen Carrier to

Biodiesel

Temperature

(°C)

Residence

Time (h)

Catalyst

1 Formic acid 0.5 60 9 -



4 Acetic acid 0.3 60 6 2% H2SO4

5 Acetic acid 0.2 60 6 3% H2SO4


A reduction in the formic acid ratio from 0.5 to 0.3 (Trial 2) vastly improved selectivity but

resulted in a slight reduction in conversion (Table 5.2). A further reduction in molar ratio

of formic acid to 0.1 (Trial 3) improved selectivity to almost 100.0% but reduced

conversion to an unacceptably low value of 36.2%. The use of formic acid as oxygen

carrier has been linked to excessive oxirane ring opening (La Scala and Wool 2002) in the

presence of the aqueous hydrogen peroxide. Hence the switch to acetic acid as oxygen

carrier was undertaken. Acetic acid was used initially at a molar ratio of 0.3, with 2 wt%

(of combined hydrogen peroxide and acetic acid) sulfuric acid (Trial 4). These conditions

produced a selectivity between that for trials 1 and 2 with a reduced conversion partly due

to the reduced residence time. A molar ratio of acetic acid of 0.2 and sulfuric acid at 3 wt%

(Trial 5) further reduced conversion but increased selectivity to 99.8%.


72

Table 5.2: Selectivity for epoxy methyl biodiesel.

Trial

Number

Conversion

(wt%)

Di-hydroxy

Biodiesel (wt%)

Epoxy Biodiesel

Selectivity (wt%)

1 89.1 2.1 97.5

2 81.4 0.3 99.6

3 36.2 0.0 99.9

4 67.7 1.2 98.2

5 44.8 0.1 99.8

The conditions of trial 5 were selected as the optimal compromise in terms of the trade-off

between conversion and selectivity and were standardised for subsequent epoxidation

exercises (Figure 5.2). The end result is the conversion of approximately 45% of the

unsaturated fraction of biodiesel to epoxide, equating to an epoxy content of approximately

42%.

Figure 5.2: Graphical representation of the 5 trial batches for the optimisation of the epoxidation of methyl

biodiesel.

0

10

20

30

40

50

60

70

80

90

100

1 2 3 4 5

Con

vers

ion/

Sel

ectiv

ity (%

)

Trial

Conversion (%)

Epoxy Selectivity

Glycol Biodiesel (%)


73

5.3 Alkoxylation of Epoxy Methyl Biodiesel Initially, alkoxylation of epoxy methyl biodiesel was performed with a nominal excess of

methanol and a relatively high mass ratio of catalyst (section 4.3.1). The reaction was

monitored over a 24h period by GC-MS as detailed in chapter 4.

5.3.1 Method

A molar ratio of methanol to biodiesel of 9:1, a catalyst concentration of 5 wt% and a


1. Pre-heated the combined epoxy methyl biodiesel and methanol to 50°C in the

reactor described in section 3.2.1, including the condenser,

2. Added the sulfuric acid,




running tap water,

6. Returned the kettle to the hotplate, added a small quantity of RO water and

neutralised with KOH pellets,

7. Transferred to a 1L separating funnel and added an equal quantity of RO water,


phase,


unrefined alkoxy biodiesel, mixed and allowed to settle,

10. Drained the lower phase and discarded, continued to wash as in 9. until the odour

of methanol was absent,




74

Alkoxy selectivity ( AS ) was calculated from the areas of the alkoxy and by-product peaks.

100��

��

��

�KGA

AA AAA

AS (5.4)

Glycol content ( Gm ) and ketone content ( Km ) in the alkoxy biodiesel were calculated in a

similar fashion whilst accounting for the proportion of glycol with respect to ketone:

� � BGK

KAK XAA

ASm ��

��

��

�

�� 93.011(%) (5.5)


Complete conversion of epoxide was achieved within 30 minutes with a selectivity for

methoxy of 75.2%, rising slightly to 75.9% after 24h (Figure 5.3). The remaining

component was the aforementioned by-product of glycol and another not seen in the

epoxide; a ketone (Figure 5.4) formed by re-arrangement of oxirane, as reported by (Rios

et al. 2005). Given the results of the 24h trial batch, alkoxylation of epoxy methyl biodiesel

was performed with only 2.5 wt% sulfuric acid and a reduced residence time of 1h to

improve selectivity for methoxy biodiesel. Selectivity rose to 89.0% with a glycol content

of 3.0% and keto content of 1.7%.

Figure 5.3: Selectivity for methoxy biodiesel and fractions of by-product (glycol and ketone), including

fractionated samples.

0

10

20

30

40

50

60

70

80

0 4 8 12 16 20 24

Sele

ctiv

ity (%

)

Reaction time (h)

Methoxy selectivityGlycolGlycol - fractionatedKetoneKetone - fractionated


75

Figure 5.4: Chromatogram of methoxy methyl biodiesel. Major peaks: methyl palmitate (16.33 min.),

methyl oleate/linolenate (22.08 - 22.16 min.), 9(10)-keto methyl stearate (29.09 min.), 9(10)-hydroxy,10(9)-

methoxy stearate (33.30 min.), 9,10-dihydroxy methyl stearate (35.88 min.).

A precipitate was observed in the product following overnight storage in the laboratory

(Figure 5.5). GC analysis of the supernatant and lower fraction containing the precipitate

confirmed that the solids were principally glycol methyl stearate. The supernatant

contained 4.2% glycol compared to 11.2% in the portion containing the solids. The two

chromatograms were superimposable except for the peak identified as glycol methyl

stearate (Figure 5.6). The fraction of keto methyl stearate in the supernatant was identical

to that in the settled phase. It is suggested that the bulk of the glycol production occurs

during oxirane ring opening in the alkoxylation step under acid conditions. Although

alkoxylation proceeds in the absence of water (in contrast to the epoxidation step), oxirane

ring opening is a significantly faster reaction and occurs via the nucleophilic attack of

hydronium ions generated by the acid catalyst. The ability for these vicinal dihydroxy alkyl

esters to form hydrogen bonds with each other facilitated the crystallisation of these

compounds.

Time (min.)


76

A two-stage crystallisation fractionation at 0°C and a single fractionation at -20°C

substantially reduced the glycol content from 17.6% to 5.9% and 5.4%, respectively

(Figure 5.3). However, such levels of contaminant would be unacceptable in a commercial

biodiesel and would almost certainly adversely affect the cloud point. Clearly, it is critical

to minimise the formation of this by-product to prevent an increase in CP over

conventional biodiesel.

Figure 5.5: Methoxy methyl biodiesel at room temperature with precipitate.

Figure 5.6: Chromatogram of methoxy methyl biodiesel: (I) supernatant, (II) precipitate.

I

II

Time (min.)


77

5.4 Extension to Ethyl and Butyl Biodiesel Ethyl and butyl biodiesel were subjected to epoxidation using the identical optimal

conditions to those identified in sections 5.1 and 5.2. This was followed by alkoxylation

with the respective alcohol under the same conditions as for epoxy methyl biodiesel to

compare the reaction rate and selectivity for the three.

5.4.1 Epoxidation

Ethyl and butyl biodiesel were subjected to epoxidation using the identical optimal

conditions to those identified in sections 5.1 and 5.2: H2O2 / biodiesel molar ratio of 2:1,

acetic acid / biodiesel molar ratio of 0.2:1, acid catalyst to acetic acid / peroxide of 2 wt%,

6h reaction at 60°C. The procedure was identical to that described in section 5.1.1. GC-MS

was used to determine conversion and selectivity as previously described. Interpretation of

the respective mass spectra involved the assumption of an identical fragmentation

(ionisation) pattern as seen for methyl biodiesel but with the concordant increase in m/z

due to the higher molecular mass.

A slightly lower conversion was achieved for ethyl biodiesel whereas conversion of butyl

biodiesel was reduced by as much as 1.5% (Table 5.3). The reduction in conversion

reflects the increase in the hydrophobic nature afforded by the longer ester head group and

the associated reduction in accessibility to the double bonds for peroxyacetic acid.

Selectivity for epoxy alkyl esters was similar for methyl, ethyl and butyl biodiesel at

virtually 100%.

Table 5.3: Epoxy conversion, selectivity and glycol content for all alkyl esters.

Alkyl Ester Conversion

(wt%)

Epoxy Selectivity

(wt%)

Glycol Biodiesel

(wt%)

Methyl 44.8 99.8 0.2

Ethyl 44.6 99.8 0.2

Butyl 43.3 100.0 0.0


78

5.4.2 Alkoxylation

Alkoxylation with the respective alcohol (ethanol for epoxy ethyl biodiesel and butanol for

epoxy butyl biodiesel) was performed as detailed in section 5.3. Analysis of reaction

products was also identical to that for the methyl variant while taking into account the

increase in molecular mass.

Alkoxylation of epoxy methyl biodiesel resulted in a selectivity of 89.0% and a

corresponding glycol content of 3.0% (Table 5.4). The respective ethoxy and butoxy

selectivities were 6.3% and 7.3% lower, resulting in increased glycol contents of 4.3% and

3.8%. However, keto alkyl stearate content for ethyl and butyl biodiesel was much greater

at 2.9% and 3.6% respectively, compared with 1.7% for methyl biodiesel. It is evident that

the ratio of ketone to glycol by-product formation increased as the alcoholic adduct

increased in chain-length, almost reaching parity for the butyl biodiesel. This may be due

to the increasing difficulty for the large alcohol to reach the oxirane group or more likely a

result of the reduced nucleophilic strength of the longer alcohol.

Table 5.4: Alkoxy selectivity and by-product content for all alkyl esters.

Alkyl Ester Alkoxy

Selectivity

(wt%)

Alkoxy

Content (wt%)

Glycol

(wt%)

Ketone

(wt%)

Methyl 89.0 37.1 3.0 1.7

Ethyl 82.7 34.3 4.3 2.9

Butyl 81.7 32.9 3.8 3.6

Butyl (high X) 86.3 53.6 2.8 5.7


79

5.5 Cloud Point Assessment Methyl biodiesel of canola oil had a cloud point of -2°C which is in line with the findings

of other researchers (Knothe et al. 2004). Methyl biodiesel with methoxy groups

substituted for approximately 45% of the available double bonds exhibited a cloud point of

4°C, representing an increase of 6K (Table 5.5). Ethyl biodiesel, substituted to

approximately the same extent, displayed a CP increase of 3K. Butyl biodiesel, also

substituted to approximately the same extent, had a CP of -4°C representing a CP

improvement of 1K.

Table 5.5: Cloud point for all alkyl esters.

Alkoxy Ester Alkoxy

Content

(wt%)

CP Raw

Biodiesel (°C)

CP Modified

Biodiesel (°C)

CP

Improvement

(°C)

Methyl 37.1 -2 4 -6

Ethyl 34.3 -3 0 -3

Butyl 32.9 -3 -4 1

Butyl (high X) 53.6 -3 0 -3

The increase in cloud point for the methyl and ethyl variants was not entirely unexpected

because of the short length of the side-chain. This was also observed by Moser and Erhan

(2006) who found that an n-butyl side-chain or longer resulted in a reduction in cloud point

for isopropyl oleate while an ethyl side-chain resulted in a higher CP. Unsaturated fatty

acids of canola oil are primarily cis-isomers resulting in a chain conformation that is

rotated at the carbon-carbon double bond by 30° (Figure 3.3). This results in a fatty acid

chain with a bend at the double bond, unlike the saturated or the trans-isomer unsaturated

fatty acids. Molecules are therefore unable to pack as closely together resulting in a

reduced melting point. Epoxidation of the double bonds and subsequent ring opening

transforms the unsaturated fraction to a straight chain allowing closer packing and

potentially a higher melting point. Substitution at the oxirane group of a linear alkyl chain

can disrupt the side-by-side packing of ester molecules resulting in a decrease in melting

point. Clearly, the length of the alkyl side-chain is important and in terms of molecule

spacing, the longer the side-chain the greater the disruption to crystal nucleation.


80

Transformation of the double bonds of unsaturated biodiesel will therefore produce an

increase in cloud point until a point is reached where the length of the alkyl side-chain is

such that it offers a benefit over the cis-unsaturated esters in terms of molecular spacing.

On the basis of the crystallisation theory and past work, the cloud point for methyl and

ethyl biodiesel was not expected to improve, however the improvement for butyl biodiesel

was significantly smaller than anticipated. Two factors may have influenced the

crystallisation temperature and cloud point: first, the low conversion rate and second, the

presence of by-product. The reduction in unsaturated content acts to increase cloud point

so the alkoxylated fraction must reach a point where it is able to inhibit crystal nucleation

of the remaining mixture of saturated and unsaturated alkyl esters. The presence of both

cis-unsaturated esters and butoxy substituted biodiesel may actually facilitate a closer

packing arrangement between the ‘bent’ unsaturated molecules and those with a side-chain

that can reach in to gain improved access to the double bond. This would promote the side-

by-side parallel packing of the molecules and result in an increase in the cloud point. As

well, the presence of a significant quantity of glycol and ketone (3.8% and 3.6%

respectively in the case of butyl biodiesel) could mean that these molecules are able to

approach each other to form hydrogen bonds and initiate crystallisation sooner than for the

remaining mixture.

5.5.1 Impact of Higher Alkoxy Content

In an attempt to determine the impact of higher conversion to butoxy biodiesel, butyl

biodiesel was subjected to epoxidation with peroxyformic acid for 9h at 60°C with a

resultant conversion of 66.8% and a selectivity of 99.9%. Alkoxylation under identical

conditions as prior batches resulted in a butoxy biodiesel content of 53.6% (Table 5.4).

Unfortunately, the cloud point of this product was 0°C, an increase of 3K over butyl

biodiesel (Table 5.5). Interestingly, the glycol content of the product was lower at 2.8% but

the keto content was increased to 5.7%. Since the glycol content was lower than for the

previous batch which had a much lower conversion with a higher proportion of ketone, it is

postulated that there was some dehydration of the glycol to ketone. As well, the relatively

high content of contaminant may have resulted in hydrogen bonds forming between the

keto and hydroxyl groups thereby causing a higher cloud point at a higher butoxy content.

Clearly, the focus of further work must be on the alkoxylation step to determine the

optimal conditions for high conversion rates to alkoxy biodiesel with high selectivity.


81

5.6 Summary Optimal conditions for the epoxidation of canola derived biodiesel were H2O2 to biodiesel

molar ratio of 2:1, acetic acid to biodiesel molar ratio of 0.2:1, acid catalyst to acetic acid /

peroxide of 2 wt%, 6h reaction at 60°C. Selectivity for epoxy biodiesel over vicinal

dihydroxy biodiesel was 99.8%, 99.8% and 100.0% for methyl, ethyl and butyl biodiesel,

respectively. Alkoxylation for 1h at 60°C, alcohol to epoxy biodiesel molar ratio of 14:1

and 2.5 wt% sulfuric acid resulted in alkoxy substitution rates of 37.1% (methyl), 34.3%

(ethyl) and 32.9% (butyl). Selectivity for alkoxy biodiesel was 89.0%, 82.7% and 81.7%

for methoxy, ethoxy and butoxy biodiesel, respectively. Cloud point for methyl and ethyl

biodiesel increased from -2°C to 4°C, and from -3°C to 0°C, respectively. The cloud point

of butyl biodiesel reduced from -3°C to -4°C.

Optimisation of the epoxidation conditions to minimise vicinal dihydroxy biodiesel

production resulted in low conversions in the commercially reasonable residence time of

6h. In addition, significant quantities of vicinal dihydroxy and keto biodiesel were

produced during the alkoxylation step that have a much higher melting point than both

standard and alkoxy biodiesel. The presence of the by-product negated much of the

expected improvement in cloud point for butoxy butyl biodiesel. Further optimisation work

was required to improve both conversion and selectivity to butoxy butyl biodiesel.

Chapter 6

Butoxylation of Butyl Biodiesel

Chapter 6 Butoxylation of Butyl Biodiesel _________________________________________________________________________

83

BUTOXYLATION OF BUTYL BIODIESEL _________________________________________________________________________

Significant quantities of vicinal dihydroxy and keto biodiesel were produced during the

alkoxylation step described in the previous chapter. The presence of the by-product was

suspected of negating much of the expected improvement in cloud point. Butoxy butyl

biodiesel provided the most promising results in terms of cloud point improvement.

Therefore, further work concentrated on the optimisation of the butoxylation of epoxy

butyl biodiesel to improve both conversion and selectivity.

This chapter describes the process of investigation for the butoxylation of epoxy butyl

biodiesel produced from canola oil to determine the optimal reaction conditions for

maximising conversion while limiting by-product formation. Parameters examined were

temperature, reaction time, catalyst concentration and molar ratio of alcohol to epoxy

biodiesel. Conversion and selectivity were monitored by GC-MS. The impact of

butoxylation on cloud point of biodiesel was determined. A discussion of the kinetic

parameters of the reaction is also included.


84

6.1 Method

6.1.1 Synthesis of Epoxy Butyl Biodiesel

The well established base-catalysed transesterifaction of canola oil was performed in a

glass reaction kettle on a hotplate/stirrer with feedback temperature control. Canola oil was

pre-heated to 80°C. The combined n-butanol (molar ratio of 7:1) and catalyst (2 wt%

potassium hydroxide) were added at room temperature. A two-stage process, each with 4h

batch time and the purification method outlined in section 4.1.3.3, was performed to ensure

>99% conversion. Purification steps included a first phase separation to remove the bulk of

the glycerol followed by repeated water washes to remove residual glycerol and alcohol.

The top organic phase was dried over anhydrous sodium sulfate followed by filtration to

produce butyl biodiesel with a monoglyceride content of <1.0 wt% (Figure 6.1).

Figure 6.1: Chromatogram of butyl biodiesel.

Epoxidation was performed at 60°C for 6h on a temperature controlled hotplate/stirrer with

in-situ generated peroxyacetic acid (oxygen carrier) and sulfuric acid catalyst. The molar

ratio of hydrogen peroxide to biodiesel was 2:1, the molar ratio of acetic acid to biodiesel

was 0.2:1 and the sulfuric acid concentration was 2 wt%. Residual acid and peroxide were


85

neutralised with sodium bicarbonate solutions, followed by several water washes and

drying over anhydrous sodium sulfate, as detailed in section 5.2.

As for methyl biodiesel, the mixed fatty acid nature of canola oil derived butyl biodiesel

ensures that a number of different epoxy compounds are possible. These included multi-

epoxy esters and epoxy esters ranging from palmitoleic to nervonic. The most likely by-

product is vicinal di-hydroxy compounds (glycol). Epoxides and glycol may also be

present at one or more of the available double bonds for polyunsaturated esters. As the

bulk of the unsaturated portion of biodiesel derived from canola oil is butyl

oleate/linolenate, conversion of butyl oleate/linolenate to epoxy biodiesel was monitored

by the reduction in area of the butyl oleate/linolenate peaks, referenced to the butyl

palmitate peak (Figure 6.2). Selectivity for epoxy butyl biodiesel was determined to be

100% for all batches. Epoxidation of butyl biodiesel for 6h resulted in a conversion of

approximately 46% of the available unsaturated portion. This was determined from the

extent of conversion of the oleate/linolenate (C18:1/3) fraction of biodiesel. Conversion of

C18:1/3 was multiplied by the percentage of unsaturated ester (93 wt%) in the original

biodiesel to arrive at an epoxy biodiesel content of 43% for the bulk epoxy biodiesel used

for subsequent butoxylation. While conversion of unsaturated ester was far from complete,

a batch time of 6h was deemed as the maximum for a commercially viable process and

would provide a satisfactory basis for the butoxylation optimisation exercise. In addition,

the preliminary work found that alkoxylation at higher proportions actually resulted in

reduced low-temperature tolerance. Three batches of epoxy butyl biodiesel were produced

and then combined to produce a pool for the optimisation experiments.


86

Figure 6.2: Chromatogram of epoxy butyl biodiesel. Main peaks: butyl oleate/linolenate (31.38-31.56 min.),

epoxy butyl stearate (38.45 min.).

6.1.2 Optimisation of Alkoxylation Reaction Conditions

Oxirane ring opening and subsequent addition of n-butanol was performed in a glass

reaction kettle on a hotplate/stirrer with feedback temperature control (section 3.2.1).

Sulfuric acid was chosen as the catalyst and the reaction was monitored for the batch time

of 6h. For the initial temperature experiments, the molar ratio of butanol was fixed at 10:1

as was the acid concentration at 2 wt%. The optimal reaction temperature was then chosen

based on the conversion rate and butoxy selectivity. Catalyst concentration was then varied

at the optimal temperature. Finally, the alcohol molar ratio was varied at the optimal

temperature and catalyst concentration.

Samples were taken periodically and placed into vials containing 1 mL of saturated sodium

bicarbonate solution and 3 mL of heptane. Samples were shaken to neutralize the acid and

allowed to stand at room temperature before transferring an aliquot of the top heptane layer

to vials for GC analysis.


87

Conversion of epoxy butyl biodiesel was monitored via GC-MS over the 6h period for the

optimisation batches by disappearance of the epoxy butyl stearate peak (Figure 6.3). The

main by-products of this reaction are the aforementioned glycol and the keto form as a

result of re-arrangement of oxirane, as reported by Rios et al. (Rios et al. 2005).

Figure 6.3: Chromatogram of epoxy butyl biodiesel 20 min. in to the butoxylation. Main peaks: butyl

oleate/linolenate (31.36 - 31.56 min.), epoxy butyl stearate (38.39 min.), butoxy butyl stearate (49.02 min.).

It was identified that 9,10-keto butyl stearate, the main by-product, co-eluted with butyl

eicosenoate (C20:1) at approximately 38.8 minutes (Figures 6.1 and 6.4). The proportion

of the aforementioned peak attributed to the butyl eicosenoate was calculated by

subtracting the area for butyl eicosenoate found in the bulk biodiesel normalised to the

butyl palmitate peak in the respective chromatogram.


88

Figure 6.4: Chromatogram of butoxy butyl biodiesel at completion of butoxylation showing the absence of

the epoxy butyl stearate peak but the presence of the keto butyl stearate/butyl eicosenoate (C20:1) at 38.73

min.

Mass spectra of the major peaks confirmed these identities, as detailed in Figures 6.5 and

6.6. Butoxy selectivity ( AS ) was calculated from the relative areas of the butoxy ( AA ),

glycol ( GA ) and ketone ( KA ) peaks.

100��

��

��

�KGA

AA AAA

AS (6.1)

Glycol and keto content in the butoxy biodiesel were calculated from the selectivity and

conversion of unsaturated biodiesel ( BX ) as demonstrated for ketone:

� � BGK

KAK XAA

ASm ��

��

��

�

�� 93.011(%) (6.2)


89

Figure 6.5: Mass spectrum of peak at 49.0 min. identified as 9-butoxy,10-hydroxy butyl stearate.

Figure 6.6: Mass spectrum of peak at 38.8 min. identified as 9(10)-keto butyl stearate.

285

229 – OC4H10 = 155

– CH – OC4H10 = 199

229

– OC4H10 = 155

199

143

m/z

281

141 - O = 125

- CH2 = 185 199

242

+CH2 = 256

213


90

6.2 Results and Discussion

6.2.1 Effect of Temperature

Temperature had a dramatic effect on conversion (Figure 6.7). At 100°C, all epoxy was

converted within 10 minutes at the initial conditions of 2 wt% sulfuric acid and 10:1 molar

ratio of butanol to biodiesel. At the lowest temperature of 40°C, conversion was not

complete after 6h but at 60°C was complete within 4h.

Figure 6.7: Effect of temperature on conversion of epoxy butyl biodiesel.

Selectivity was dependent on temperature and increased sharply from a low of 79.4% at

40°C to a maximum of 86.8% at 60°C (Table 6.1). At 80°C, selectivity was slightly lower

and at 100°C was much lower, due mainly to the increase in keto formation at this elevated

temperature. Glycol formation was influenced by temperature and was non-existent at

100°C. Keto formation however mirrored the overall selectivity and was at a minimum at

60°C. An optimal reaction temperature of 60°C was chosen based on overall selectivity

with the possibility of further reducing glycol and improving overall selectivity for

subsequent batches.

0

10

20

30

40

50

60

70

80

90

100

0 1 2 3 4 5 6

Reaction Time (h)

Con

vers

ion

(%)

100

40C

60C

80C


91

Table 6.1: Effect of temperature on selectivity for butoxy butyl biodiesel and by-product content after 6h of

reaction time.

Temperature

(°C)

Glycol

(wt%)

Ketone

(wt%)

Selectivity

(%)

40 1.3 7.5 79.4

60 1.3 4.3 86.8

80 0.6 5.7 85.1

100 0.0 7.1 83.3

6.2.2 Effect of Catalyst Concentration

A clear positive trend is evident for catalyst concentration on conversion of epoxy to

butoxy (Figure 6.8). Even at the lowest catalyst concentration of 1 wt%, 100% conversion

was achieved within 4h, while at 10 wt% conversion was complete in 1h.

Figure 6.8: Effect of catalyst concentration on conversion of epoxy butyl biodiesel.

Selectivity for butoxy butyl biodiesel displayed a similar trend to the temperature data with

a slight increase from 1 wt% to 2 wt%, then a rapid decline as catalyst concentration

increased (Table 6.2). The glycol content was strongly dependant on catalyst concentration

0

10

20

30

40

50

60

70

80

90

100

0 1 2 3 4 5 6

Reaction Time (h)

Con

vers

ion

(%)

1%

2%

5%

10%


92

and increased significantly at >2 wt%. Keto content exhibited an inverse relationship with

catalyst concentration but plateaued at � 5 wt%. The catalyst optimisation exercise

therefore confirmed that a concentration of 2 wt% was optimal and was maintained for the

alcohol ratio batches.

Table 6.2: Effect of catalyst concentration on selectivity for butoxy butyl biodiesel and by-product content

after 6h of reaction time.

Catalyst

Concentration

(wt%)

Glycol

(wt%)

Ketone

(wt%)

Selectivity

(%)

1 0.9 5.1 85.8

2 1.3 4.3 86.8

5 6.9 2.6 77.5

10 9.7 2.6 71.2

6.2.3 Effect of Molar Ratio of Alcohol

Molar ratio of n-butanol had a significant positive effect on conversion of epoxy to butoxy

(Figure 6.9). Butoxylation was complete within 1h at a molar ratio of 40:1 and was barely

complete after 6h at a molar ratio of 5:1.

Selectivity was relatively unchanged for all molar ratios of butanol but was highest at 40:1

(Table 6.3). Glycol content reduced significantly with increasing molar ratio of butanol

while keto content increased slightly. A molar ratio of 40:1 for butanol to biodiesel was

chosen as optimal, producing a selectivity of 87.9%. Given the marginal impact of alcohol

ratio on selectivity, a lower ratio could be chosen for an industrial-scale process to reduce

the costs associated with a larger working volume and the corresponding cost increases due

to the requirement to recover the excess alcohol. As the required residence time increases

with the reduction in alcohol ratio, an economic trade-off would need to be considered.


93

Figure 6.9: Effect of molar ratio of butanol on conversion of epoxy butyl biodiesel.

Table 6.3: Effect of molar ratio of butanol on selectivity for butoxy butyl biodiesel and by-product content

after 6h of reaction time.

Butanol

Molar

ratio

Glycol

(wt%)

Ketone

(wt%)

Selectivity

(%)

5 1.3 4.1 87.3

10 1.3 4.3 86.8

20 1.0 4.3 87.7

40 0.7 4.5 87.9

6.2.4 Further Optimisation

In an attempt to further optimise the reaction conditions, four batches were produced at a

40:1 molar ratio of n-butanol, but at the two higher temperatures, and at catalyst

concentrations of 2 wt% and 5 wt% (Table 6.4). The higher temperature of 100°C resulted

in reduced glycol content but increased keto content, as seen earlier (Figure 6.10).

0

10

20

30

40

50

60

70

80

90

100

0 1 2 3 4 5 6

Reaction Time (h)

Con

vers

ion

(%)

5

10

20

40


94

Figure 6.10: Results of further optimisation work for the reaction conditions for the epoxidation of epoxy

butyl biodiesel.

The higher catalyst concentration did not alter keto content but appeared to have a large

impact on glycol content in the case of the 80°C batch, which is consistent with the data

presented in Table 6.2.

Table 6.4: Effect of higher temperature and catalyst concentration on selectivity for butoxy butyl biodiesel

and by-product content after 6h of reaction time.

Temperature

(°C)

Catalyst

Concentration

(wt%)

Glycol

(wt%)

Ketone

(wt%)

Selectivity

(%)

80 2 0.5 4.0 89.4

80 5 2.4 3.8 85.4

100 2 0.2 4.9 87.8

100 5 0.0 5.0 88.3

0

10

20

30

40

50

60

70

80

90

100

0 1 2 3 4 5 6

Reaction Time (h)

Con

vers

ion

(%)

80C 5%

100C 2%

100C 5%

80C 2%


95

6.3 Reaction Kinetics Although there was initially no intent to determine reaction kinetics, the quality of the data

collected in section 6.2 does allow for a first approximation of some kinetic parameters.

The specific reaction rate ( k ) was determined based on an assumed first order reaction and

the integrated form of the specific reaction rate equation:

� � � � kteAA �� 0 (6.3)

A plot of � �Aln against time t gives a straight line with slope k� . In this case since the

data available is conversion ( X ) versus t , a plot of � �X�1ln versus t is presented and

yields the specific reaction rate constant k (Figure 6.11).

a b

c d Figure 6.11: Kinetic plots for the various molar ratios of alcohol: 5:1 (a), 10:1 (b), 20:1 (c), 40:1 (d).

-3

-2.5

-2

-1.5

-1

-0.5

0

0 1 2 3

ln (1

-X)

Time (h)-4

-3.5

-3

-2.5

-2

-1.5

-1

-0.5

0

0 1 2 3

ln (1

-X)

Time (h)

-5

-4

-3

-2

-1

0

0 1 2 3

ln (1

-X)

Time (h)-5

-4

-3

-2

-1

0

0 0.5 1 1.5

ln (1

-X)

Time (h)


96

Table 6.5: Specific reaction rate and corresponding coefficient of determination (R2) for the four experiments at different molar ratios of alcohol.

Molar Ratio

Alcohol

Rate, k (h-1) R2

5 1.219 0.983

10 1.751 0.988

20 2.183 0.984

40 4.215 0.978

As is evident from Table 6.5 the coefficients of determination (R2) are reasonable but, all

of the smoothed curves of Figure 6.11 display an inflection at around 0.17 h. A review of

the data presented in Figure 6.9 also reveals that the reaction rate is much faster in the

initial 10 minutes of the reaction. It was noted that there was an immediate temperature rise

above the set-point when the acid catalyst was added to the reactor. The reaction is clearly

an exothermic one and the temperature control of the reactor was not adequate to control

the initial jump in temperature. This undoubtedly affected the reaction rate during the first

10 minutes of the reaction and caused the relatively fast initial reaction rate.

If the initial rate data (0 to 10 minutes) are excluded from the kinetic plots, the R2 values

are greatly improved (Table 6.6). If the specific reaction rate constants are plotted against

alcohol ratio, it cannot be conclusively said that the reaction is first or even second order

with respect to alcohol concentration given the relatively poor R2 value (Figure 6.12). The

quality of the data collected is not sufficient to accurately assign a reaction order.


97

Table 6.6: Specific reaction rate and corresponding cefficient of determination (R2) value for the four experiments at different molar ratios of alcohol, excluding the initial 10 minutes.

Molar Ratio

Alcohol

Rate, k (h-1) R2

5 1.144 0.998

10 1.649 1.000

20 2.012 0.999

40 3.782 1.000

Figure 6.12: Specific reaction rate versus molar ratio of alcohol.

The conversion over time at various temperatures (Figure 6.7) provides the opportunity to

estimate the activation energy (E) from an Arrhenius plot, where:

RTEAeTk /)( �� (6.4)

Taking the natural logarithm of Equation 6.4:

�

��

TREAk 1)ln()ln( (6.5)

y = 0.0008x2 + 0.038x + 1.0287R² = 0.9892

0

0.5

1

1.5

2

2.5

3

3.5

4

0 10 20 30 40 50

Spec

ific

Reac

tion

Rtae

Con

stan

t(h-1

)

Molar ratio of Alcohol


98

A plot of )ln(k versus T1 should be a straight line with slope

RE� (Figure 6.13).

Figure 6.13: Plot of ln k versus 1/T for the 3 temperatures of 40ºC, 60 ºC and 80ºC.

Accordingly, the apparent activation energy is 53.7 kJ/mol, which seems reasonable but

there are no other validating data for comparison. It is clear from Figure 6.13 that the fit for

the data is not linear and that again, the quality of the data collected was not sufficient to

accurately determine the activation energy of the reaction. The main reason for the low

quality of the data is the relatively poor temperature control of the apparatus used for

synthesis. A much better reactor system and procedure would be necessary for a serious

investigation of the reaction kinetics, this was however, not the main aim of the body of

work presented here. Another reason for the poor fit of the data to even a second order

model may be an inhibitory effect of the major by-product of the reaction, the water from

the etherification. n-Butanol and water are somewhat miscible and it is suggested that the

build-up of water after the initial fast reaction period may inhibit the further progress of the

reaction. A method for testing this postulate would be to conduct the reaction in a suitable

organic solvent.

y = -6.4537x + 0.0205R² = 0.84

-0.0005

0

0.0005

0.001

0.0015

0.002

0.0025

0.003

0.0028 0.0029 0.003 0.0031 0.0032 0.0033

ln k

(s-1

)

1/T (k-1)


99

6.4 Cloud Point Impact From the results presented in Table 6.4, a larger batch at the optimal conditions of 80°C

and a catalyst concentration of 2 wt% was produced in order to determine the impact on

biodiesel cloud point. The procedure used for purification was similar to that described in

section 5.3.1 but used sodium bicarbonate to neutralise the catalyst instead of potassium

hydroxide. The buffering effect of the bicarbonate minimised pH swings and by-product

formation. The procedure was as follows:

1. Pre-heated the combined epoxy butyl biodiesel and butanol to 80°C in the reactor

described in section 3.2.1, including the condenser,

2. Added the sulfuric acid, began timing,


running tap water,

4. Transferred to a 1L separating funnel and added an equal quantity of RO water and

two large spatulas of sodium bicarbonate, mixed and allowed to settle,


phase,


unrefined biodiesel plus one spatula of sodium bicarbonate, mixed and allowed to

settle,

7. Drained the lower phase and discarded, continued as above but with successively

reducing quantities of bicarbonate,

8. Finished the purification process with pure water washes until the odour of butanol

was absent,



The impurity profile of the larger batch was slightly worse than that for the smaller one

with a selectivity of 87.0%. The cloud point of butyl biodiesel was -3°C, as was the cloud

point of butoxy butyl biodiesel produced under the aforementioned optimal conditions.

Butoxylation of butyl biodiesel, at the conversion rate of 46% therefore had no impact on

cloud point.


100

One may suspect that the relatively high content of keto by-product may have influenced

the cloud point of the butoxy butyl biodiesel. Data presented in Table 6.7 are for butoxy

batches produced under various conditions from the same 43% epoxy butyl biodiesel. All

batches had similar overall selectivities but a range of by-product contents, from 0% glycol

and 7.1 % keto to approximately equal amounts of both. There is no discernable trend in

cloud point with respect to either glycol, keto content or both. A review of the literature for

the relative melting points for the methyl esters reveals that methyl-9,10-dihydroxystearate

has a melting point of 103°C compared to methyl-9(10)-keto-stearate at 46 to 48°C

(Markley 1967). This can be contrasted with the melting point of methyl stearate at 39°C.

This suggests that the dihydroxy esters will have a much greater impact on cloud point

than the keto esters. Even the batches in Table 6.7 with significant quantities of glycol had

essentially the same cloud point as for the original butyl biodiesel. It is therefore unlikely

that the cloud point of the butoxy butyl biodiesel batches were significantly affected by the

by-product content since the glycol content was always very low.

Table 6.7: Impurity profile and cloud point results for various butoxy butyl biodiesel batches.

Butoxy

Selectivity (%)

Glycol

(wt%)

Ketone

(wt%)

Cloud Point

(°C)

83.6 3.8 2.8 -4

80.6 3.1 5.2 -3

83.0 1.6 5.6 -3

87.0 0.6 4.9 -3

83.4 0.0 7.1 -5

6.4.1 Impact of Higher Conversion

To determine the impact of a higher conversion of unsaturated ester to butoxy ester, a

batch of butyl biodiesel was subjected to 30h of epoxidation resulting in the conversion of

93% of the oleic/linolenic portion. Subsequent butoxylation at the optimal conditions

resulted in 100% conversion of epoxy and a selectivity for butoxy of 84.6%. Since the

fraction of canola oil that is unsaturated is 93%, the butoxy content was calculated to be

74%. The glycol content was very low as for the lower conversion batches at 0.5%, but


101

keto content was significant at almost 13%. The cloud point of this material was 2°C,

representing an increase of 5K over the original butyl biodiesel.

Blends of the high conversion batch of butoxy biodiesel with butyl biodiesel were prepared

to determine the cloud point trend across a range of fractions from 2 wt% to 74 wt%

(Figure 6.14). Cloud point was virtually unchanged at concentrations below 35% and then

increased 1K every 8 wt% to approximately 70 wt % butoxy biodiesel.

Figure 6.14: Cloud point of butoxy butyl biodiesel from 2 wt% to 74 wt%.

The loss of unsaturated ester, due to conversion to butoxy ester, appears to have a

significant effect on cloud point only after approximately one third of the unsaturated ester

is converted. Butoxy biodiesel therefore is able to prevent the earlier onset of

crystallisation due to the decrease in unsaturated content but only at lower concentrations.

Once the proportion of unsaturated material had decreased by approximately one third,

crystallisation temperature rose linearly.

The major fatty acid contained in canola oil is oleic acid at approximately 64 wt%, which

has a cis- double bond at the 9,10 carbons of the fatty chain. The result is a fatty acid with

a 30° bend in the middle of the tail-group. The non-linear conformation of the unsaturated

esters is what ordinarily restricts the molecules from approaching each other close enough

-4

-3

-2

-1

0

1

2

3

0 10 20 30 40 50 60 70 80

Clo

ud P

oint

(°C

)

Butoxy Biodiesel Content (%w/w)


102

for crystals to nucleate, resulting in a lower melting point than the saturated equivalent

(stearate). Epoxidation of a cis-unsaturated fatty acid ester is known to produce a cis-epoxy

fatty acid ester (Swern 1967). Butoxylation under these relatively mild conditions results in

the addition of only one molecule of n-butanol at either the 9 or 10 position and a hydroxyl

group on either the 10 or 9 position in a trans- conformation. Thus, opening of the oxirane

ring results in the conversion of a cis-isomer to a straight-chain trans-isomer, in this case a

trans-�-hydroxy ether of the fatty acid ester. Therefore butoxylation results in the

linearisation of molecules that were originally non-linear.

6.4.2 Impact of Linearisation of Ester

To investigate the impact of linearisation of the ester molecules, without the complication

of saturated fatty esters, an identical treatment to that for canola oil derived ester of oleic

acid derived ester was performed. Technical grade oleic acid was subjected to

esterification with n-butanol followed by epoxidation and alkoxylation under the

aforementioned optimal conditions for the canola oil based biodiesel. The conversion of

butyl oleate to epoxy stearate was identical to that for biodiesel at 46%. The conversion to

butoxy butyl stearate was complete and the selectivity was 88.9%.

The cloud point for butyl oleate was -18°C while the cloud point for butoxy butyl stearate

was -12°C, demonstrating that butoxylation results in a net increase in cloud point, even in

the absence of saturated ester. To determine whether the increase in cloud point for

butoxylated oleic acid was due only to the linearisation of the molecule, butyl stearate was

substituted for butoxy butyl stearate. Stearic acid was esterified with n-butanol and was

blended with butyl oleate to produce the same relative fractions as that for the butoxylated

butyl oleate (including the by-products which are also linear molecules). If the cloud point

impact of butoxylation was simply due to the linearisation of the unsaturated portion of

biodiesel then the cloud point of a mixture of 46% butyl stearate and 54% butyl oleate

would be similar to that for the 46% butoxy butyl oleate. The cloud point of the

aforementioned blend of butyl oleate and butyl stearate was actually +18°C. This

represents an increase in CP of 30K compared to the increase in CP of 4K for the butoxy

butyl oleate. The oleate/stearate blend also did not contain any of the keto or glycol

product that is found in the butoxy butyl oleate, providing further evidence that the by-

product does not adversely influence CP. The reason for the failure to significantly alter

the crystallisation habit of biodiesel is likely to be the short length of the butoxy side-chain.


103

Substitution of a longer or branched alkoxy group may be more effective at interrupting the

orderly alignment of the ester head-groups by providing more effective steric hindrance

between the ester functional groups.


104

6.5 Summary Butyl biodiesel derived from canola oil was epoxidised via the in-situ peroxyacetic acid

method for 6h resulting in a conversion of 46% of the unsaturated esters. Epoxy butyl

biodiesel was butoxylated with n-butanol with sulfuric acid catalyst without the use of

solvents. Conversion and selectivity for butoxy butyl biodiesel were optimised by

examining reaction conditions including; temperature, reaction time, catalyst concentration

and molar ratio of alcohol to epoxy biodiesel.

Optimal conditions for the butoxylation of epoxy butyl biodiesel were: 80°C, 2 wt%

sulfuric acid and a 40:1 molar ratio of butanol over a period of 1h. Conversion of epoxy

butyl biodiesel was 100% and selectivity for butoxy biodiesel was 87.0%. An attempt at

the elucidation of the reaction kinetics indicated that the reaction order is second or higher.

There appears to be an inhibitory process occurring after an initially fast reaction rate

which is probably due to the production of water as a by-product of the alkoxylation.

The cloud point of butoxy butyl biodiesel (conversion of 46% of unsaturated portion) was

identical to that for butyl biodiesel. Butoxy butyl biodiesel at a conversion of 93% of

unsaturated portion had a cloud point 5K higher than that for butyl biodiesel. Blends of the

high conversion batch of butoxy biodiesel showed that cloud point was virtually

unchanged at concentrations below 35% and then increased 1K every 8 wt% to

approximately 70 wt % butoxy biodiesel.

The loss of unsaturated ester, due to conversion to butoxy ester, appears to have a

significant effect on cloud point only after approximately one third of the unsaturated ester

is converted. Butoxy biodiesel is therefore able to prevent the earlier onset of

crystallisation due to the decrease in unsaturated content but only at lower concentrations.

Once the proportion of unsaturated material was lowered by approximately one third,

crystallisation temperature rose linearly. Butyl biodiesel can be butoxylated under mild

conditions and reasonable reaction times but this process does not offer an improvement in

biodiesel cloud point. The length of the butoxy side-chain is not sufficient to cause

disruption to the normal crystallisation mechanism for fatty esters.

Chapter 7

Other Adducts of Butyl Biodiesel

Chapter 7 Other Adducts of Butyl Biodiesel _________________________________________________________________________

106

OTHER ADDUCTS OF BUTYL BIODIESEL _________________________________________________________________________

The work presented in the previous chapter focussed on finding the optimal conditions for

the alkoxylation of butyl biodiesel after determining that butoxy butyl biodiesel was the

best candidate (so far) for lowering the cloud point. Optimisation of the reaction conditions

for the butoxylation of epoxy butyl biodiesel resulted in 100% conversion and a selectivity

for butoxy biodiesel of 87.0%. The cloud point of butoxy butyl biodiesel (conversion of

46% of unsaturated portion) was however, identical to that for butyl biodiesel. After ruling

out the by-product content as the reason for the lack of improvement in cold-flow

properties, the remaining question was the impact of the length of the alkoxy group.

To that end, alkoxy side-chains of various lengths were added to butyl biodiesel derived

from canola oil, using the optimal conditions determined previously. The alkoxy groups

included: methyl, ethyl, n-propyl, n-butyl, tert-butyl, n-pentyl, n-hexyl, n-octyl and 2-

ethylhexyl (2-EH). The conversion of unsaturated fraction to epoxy and subsequently to

alkoxy was approximately 46% as was the case in previous work. Conversion of epoxy and

rates of transesterification were determined by gas chromatography–mass spectrometry

(GC-MS). The impact on cloud point, pour point and viscosity were determined according

to the relevant ASTM for biodiesel and discussed.


107

7.1 Synthesis of Alkoxy Butyl Biodiesel

7.1.1 Method

Canola oil was pre-heated to 80°C prior to the addition of the combined n-butanol (molar

ratio of 7:1) and catalyst (2 wt% potassium hydroxide), at room temperature. Residence

time was 4h, followed by a water wash to remove the bulk of the glycerol. A second

transesterification under identical conditions was performed for 4h to ensure complete

conversion. Purification steps included a first phase separation to remove the bulk of the

glycerol followed by neutralisation with sulfuric acid and repeated water washes to remove

residual glycerol and alcohol. The top organic phase was dried over anhydrous sodium

sulfate followed by filtration (Refer to section 4.1.3.3). Methyl biodiesel was synthesised

as above but the reaction was performed at 60°C for 1h, once only (Refer to section

4.1.1.3). This material served as a reference for conventional biodiesel that makes up the

vast bulk of the commercial biodiesel.

Epoxidation was performed at 60°C for 6h on a temperature controlled hotplate/stirrer with

in-situ generated peroxyacetic acid (oxygen carrier) at a 0.2:1 molar ratio, with a 2 wt%

sulfuric acid catalyst (Refer to section 5.2). These conditions had previously produced a

conversion of approximately 46% of the unsaturated fraction of biodiesel with 100%

selectivity for epoxy biodiesel. Residual acid and peroxide were neutralised with sodium

bicarbonate, followed by several water washes and drying over anhydrous sodium sulfate.

Oxirane ring opening and subsequent addition of alkyl groups was performed in a glass

reaction kettle on a hotplate/stirrer with feedback temperature control. Sulfuric acid

catalyst at 2 wt%, an alcohol molar ratio of 40:1 and a residence time of 1h were the

optimal conditions determined previously and utilised here (Chapter 6). Reaction

temperature was 80°C for those alcohols with normal boiling points exceeding 80°C but

this was lowered to 60°C and 70°C for methanol and ethanol, respectively. Residual

catalyst was neutralised with an aqueous solution of sodium bicarbonate. For the alcohols

soluble in water (methyl to butyl), the residual alcohol was removed with repeated water

washes and phase separation, as previously.


108

For the remaining alcohols which were not soluble in water, the residual was removed by

vacuum distillation. Distillation under vacuum was necessary because initial attempts to

evaporate the excess alcohol at atmospheric pressure resulted in significant browning of

the product (refer to Figures 7.5 and 7.6). A vacuum of 20 – 80 �m Hg and boiling points

ranging from 26°C to 52°C, depending on the alcohol, were achieved but in some cases

significant browning of the material was still observed. The purification procedure for

alcohols from pentyl to octyl therefore included the addition of dichloromethane as a

solvent to minimise the impact of water on the product, especially during distillation. The

dichloromethane was introduced prior to neutralisation and was removed by rotary vacuum

evaporation prior to distillation of the residual alcohol. The entire procedure for pentyl to

octyl batches was as follows:

1. Pre-heated the combined epoxy butyl biodiesel and alcohol to 80°C in the reactor

described in section 3.2.1, including the condenser,

2. Added the sulfuric acid, began timing,


running tap water,

4. Transferred to a 1L separating funnel and added an equal quantity of

dichloromethane,

5. Added approximately 200 mL of RO water and 1.5 large spatulas of sodium

bicarbonate, mixed and allowed to settle,

6. Once settling was complete (approximately 1h), drained and discarded the aqueous

phase,

7. Continued as above but with successively reducing quantities of bicarbonate,

8. Finished the purification process with two pure RO water washes to remove

residual bicarbonate,

9. The organic phase was dried with excess anhydrous sodium sulfate then filtered

through Whatman No. 54 filter paper into a storage bottle,

10. The organic phase was transferred to a rotary evaporator to remove the

dichloromethane,

11. The remaining alkoxy biodiesel and residual alcohol were transferred to a vacuum

distillation rig to remove the residual alcohol,

12. The residue was transferred to a storage bottle with no further purification.


109

The extent of epoxidation of butyl biodiesel and conversion of epoxy biodiesel during the

alkoxylation step were determined using the same GC-MS parameters as previously, but

with an extended period at the maximum temperature of up to 40 min. depending on the

alkoxy chain length. Peaks attributed to the various alkoxy butyl biodiesel compounds

were identified and assigned in a similar manner as previously described.

Figure 7.1: Reaction scheme for the case of ethylhexoxy butyl biodiesel: a – butyl oleate; b – epoxy butyl

oleate ; c – ethylhexoxy butyl oleate (product); d – ethylhexoxy ethylhexyl oleate (by-product).

+

a

b

c

d


110

Alkoxy selectivity ( AS ) was calculated from the relative areas of the alkoxy ( AA ), glycol

( GA ) and ketone ( KA ) peaks.

100��

��

��

�KGB

BA AAA

AS (7.1)

Glycol and keto content in the alkoxy biodiesel were calculated from the selectivity and

conversion of unsaturated biodiesel ( BX ) as demonstrated for ketone:

� � BGK

KAK XAA

ASm ��

��

��

�

�� 93.011(%) (7.2)

Cloud point and pour point were determined with the same equipment as that described by

ASTM D 2500, using a Julabo F34 water bath containing a propylene glycol solution set at

-18°C. Replicates of cloud point tests on the same material were performed to determine

the repeatability of the method, with measurements generally accurate to within 1K of each

other, allowing for meaningful comparisons to be made. The significance of a 1K

difference in CP may therefore be considered marginal. Pour point (PP) was reported to the

nearest 3°C as required by the ASTM D 97. Dynamic viscosity was determined with a

Haake VT 550 concentric cylinder viscometer at a temperature of 40°C as recommended in

ASTM D 445. A type MV rotor was used with a 1.0mm gap and the determination was

conducted in controlled shear rate mode from 60 to 150 s-1, with a dwell time of 200 s.

Density of the samples was determined at 40°C with a 25 mL pycnometer in order to

convert dynamic viscosity to kinematic viscosity.


As the bulk of the unsaturated portion of butyl biodiesel derived from canola oil is butyl

oleate/linolenate, conversion of butyl oleate/linolenate to epoxy biodiesel was monitored

by the reduction in area of the butyl oleate/linolenate peaks (C18:1/3), referenced to the

butyl palmitate peak (C16:0). Epoxidation of butyl biodiesel for 6h resulted in a conversion

of 46% of the available unsaturated portion, and selectivity for epoxy butyl biodiesel was

100% (Figures 7.1 and 7.2). This was determined from the extent of conversion of the

oleate/linolenate fraction of biodiesel. Conversion of C18:1/3 was multiplied by the


111

percentage of unsaturated ester (93 wt%) in the original biodiesel to arrive at an epoxy

biodiesel content of 43% for the bulk epoxy biodiesel used for subsequent butoxylation.

Figure 7.2: Chromatogram of epoxy butyl biodiesel. Major peaks: I - C16 butyl biodiesel; II - C18 butyl

biodiesel (31.0-32.3 min.) – oleate/linolenate fraction is 31.5-31.6 min.; III - epoxy butyl biodiesel (37.5-38.4

min.).

Alkoxylation proceeded for 1h and was confirmed as complete by GC-MS from the

absence of the epoxy butyl oleate peak and the appearance of a group of peaks identified as

alkoxy butyl oleate (Figure 7.3) from their mass spectra.

I

III

II

Time (min.)


112

Figure 7.3: Chromatogram of octoxy butyl biodiesel. Major peaks: I - C18 octyl biodiesel (45.1-45.6 min.);

II - octoxy butyl biodiesel (64.2-66.1 min.).

Appendix A contains the chromatograms for all of the alkoxy butyl biodiesel following the

final purification step. Retention times progressively increased with alkoxy chain length as

expected, with 2-ethylhexoxy eluting between n-hexoxy and n-octoxy (Table 7.1).

Table 7.1: Alkoxylation of butyl biodiesel: retention time of major alkoxy peaks, transesterification fractions

before and after purification of the product. (NA = not applicable)

Material Retention

time (min.)

Transesterification

– initial (%)

Transesterification

– final (%)

Methoxy butyl biodiesel 41.7-42.7 42.6 46.4

Ethoxy butyl biodiesel 42.8-43.6 9.9 12.1

Propoxy butyl biodiesel 45.4-46.3 8.9 10.0

Butoxy butyl biodiesel 48.3-49.3 NA NA

tert-Butoxy butyl biodiesel NA NA NA

Pentoxy butyl biodiesel 51.1-52.0 6.3 6.7

Hexoxy butyl biodiesel 54.6-55.9 6.4 7.7

Octoxy butyl biodiesel 64.2-66.1 6.1 12.6

Ethyl Hexoxy butyl biodiesel 58.3-59.8 2.2 4.9

I

II

Time (min.)


113

The mass spectra for the peaks identified as alkoxy butyl oleate for all alkoxy butyl

biodiesels are included in Appendix B, Figures B.1 - 2 and B.4 - 9. In contrast with the

alkoxy biodiesel discussed in chapters 5 and 6, the head-group of the ester (butyl) was

different to the alkoxy side-chain, except for butoxy butyl biodiesel. This required the re-

interpretation of fragmentation patterns produced by the mass spectrometer. As previously,

the major ions of interest were those associated with the fragmentation between the 9(10)-

hydroxy and the 10(9)-alkoxy group (Figure 7.4).

Figure 7.4: Schematic of the molecular fragmentation pattern of the alkoxy butyl oleates.

Fragment (c) with a m/z of 229 for 9-hydroxy,10-alkoxy butyl oleate was always present

and prominent (Table 7.2). Fragment (a) was only present for methoxy to propoxy because

from butoxy on the fragment became too large. From butoxy on, fragment (b) became

more prominent. Fragment (d) for 9-hydroxy, 10-alkoxy butyl oleate was always

prominent and the best determinant of the species identity.

c

d

R

a

b R


114

Table 7.2: Fragmentation pattern for alkoxy butyl biodiesel for ions a to d in Figure 7.4.

Material m/z

a b c d

Methoxy butyl biodiesel 243 - 229 157

Ethoxy butyl biodiesel 257 - 229 171

Propoxy butyl biodiesel 271 - 229 185

Butoxy butyl biodiesel - 143 229 199

Pentoxy butyl biodiesel - 143 229 213

Hexoxy butyl biodiesel - 143 229 227

Octoxy butyl biodiesel - 143 229 255

Ethyl Hexoxy butyl biodiesel - 143 229 255

Selectivity for the respective alkoxy butyl biodiesel ranged from approximately 85% for

pentoxy butyl biodiesel to 98% for methoxy butyl biodiesel (Table 7.3). The selectivity for

butoxy butyl biodiesel is very similar to that experienced in previous work, with pentoxy

being very close also. In all cases, the glycol content was very low when compared to the

keto content, again as experienced in previous work. The selectivity and therefore glycol

and keto fractions could not be determined for propoxy and octoxy because in both cases

the glycol butyl biodiesel peak was coincident with another much larger peak. In the case

of propoxy, the propoxy butyl oleate peak eluted at a similar time obscuring the glycol

peak. For octoxy butyl biodiesel, the octyl oleate peak eluted at a similar time.


115

Table 7.3: Alkoxy selectivity and by-product content for all alkoxy esters. (NA = not applicable, ND = not

determined)

Material Selectivity (%) Glycol (%) Keto (%)

Methoxy butyl biodiesel 97.9 0.0 0.9

Ethoxy butyl biodiesel 92.7 0.3 2.8

Propoxy butyl biodiesel ND ND ND

Butoxy butyl biodiesel 86.2 1.2 4.7

tert-Butoxy butyl biodiesel NA ND ND

Pentoxy butyl biodiesel 84.7 1.6 4.9

Hexoxy butyl biodiesel 92.6 0.0 3.1

Octoxy butyl biodiesel ND ND ND

Ethyl Hexoxy butyl biodiesel 94.4 0.0 2.4

In the case of tert-butoxy butyl biodiesel, complete conversion of epoxy butyl biodiesel

was achieved but no tert-butoxy butyl biodiesel was produced. Instead, hydrolysis of the

oxirane group under the acidic conditions produced a mixture of vicinal dihydroxy- and

keto- butyl biodiesel, in approximately equal fractions based on relative area. The low

reactivity of tert-butanol may be attributed to the tertiary position of the hydroxyl group.

Transesterification of butyl biodiesel with the respective alcohol, present for the

alkoxylation reaction, occurred to varying degrees for all. This was due to the presence of

the acid catalyst, the high molar ratio of alcohol and the elevated temperature. The values

for transesterification in Table 7.1 were calculated from the relative peak areas for the

alkyl oleate/linolenate produced from the alcohol used for alkoxylation and the butyl

oleate/linolenate peaks. The ‘initial’ value was determined immediately following

completion of the alkoxylation reaction (1h sample) and prior to neutralisation and

purification. The ‘final’ value was determined after purification, which included distillation

at elevated temperature for alcohols larger than butyl or repeated water washes for those

that are shorter. Transesterification in the case of methanol was greater than 40%,

reflecting the greater nucleophilic strength of the methoxy group. The initial

transesterification values generally reduced as the alcohol chain length increased. The

transesterification rates (percentage) generally increased by 1 to 2% during the purification

process, except for octoxy which more than doubled to 12.6%. This was due to the


116

relatively high temperature, and time at that temperature, required to remove the residual

octanol. While the transesterification rate more than doubled for 2-EH, the initial value

was very low due to the lower reactivity of the secondary alcohol.

Transesterification occurs during alkoxylation due to hydrolysis in the presence of the acid

catalyst creating a protonated carbonyl group. The percentage (rate) of transesterification

with the large excess of alcohol present for the alkoxylation is therefore a function of the

nucleophilic strength of the alcohol in comparison with butanol. As the chain length of the

alcohol increases the nucleophilic strength reduces and hence the very large percentage for

methanol and the gradual reduction in initial transesterification rates as chain length

increased. It appears that the butyl esters are susceptible to further hydrolysis during the

purification in the presence of water and catalyst for the butyl and smaller esters and due to

the application of heat in the presence of small amounts of water and/or catalyst in the case

of higher esters. This is most evident for octoxy butyl biodiesel where the conditions

during the evaporation of dichloromethane in the presence of some water and the relatively

high temperature experienced during distillation of the octanol. The length of exposure of

octoxy butyl biodiesel to boiling octanol was also considerably greater than for the others.

A large increase in transesterification during the purification process is therefore an

indicator of possible degradation of the product. The experience with the higher alcohols

(pentanol to octanol) prior to the introduction of the dichloromethane solvent step provides

further evidence of the potential for product degradation during purification (Figures 7.5

and 7.6). As the product was reduced (vacuum distillation of residual alcohol) undissolved

solids became evident and the colour of the remaining liquid darkened considerably. The

Octoxy butyl biodiesel in Figure 7.5 illustrates the solid material that was evident adhering

to the inside of the bottle above the liquid level. The two samples in Figure 7.6 illustrate

the effect that the addition of dichloromethane had to the ethylhexoxy butyl biodiesel

batch. The dichloromethane was able to minimise the contact of the residual catalyst and

water with the product, thus minimising further hydrolysis. In addition, the enhanced

purification technique ensured that the vast majority of water had been removed from the

product prior to vacuum distillation to remove the excess alcohol. While the GC-MS failed

to detect any possible degradation product, it is suspected that polymers of esters

(estolides) or fragments of esters produced the solids and darkened the product.


117

Figure 7.5: Octoxy butyl oleate purified without dichloromethane with particulates.

Figure 7.6: Ethylhexoxy butyl biodiesel purified without dichloromethane (left) and with dichloromethane

(right).

A further complication introduced by transesterification was the production of alkoxy alkyl

biodiesel. That is, the alkoxylated fraction was transesterified with the alcohol present for

the alkoxylation reaction also. This is evident from the chromatogram for ethoxy butyl

biodiesel (Figure A.1 and reproduced here as Figure 7.7) where a relatively large peak is

present at 36.61 minutes for ethoxy ethyl oleate (Figure B.3 and reproduced here as Figure

7.8) which has a mass spectrum that differs from that for ethoxy butyl biodiesel (Figure

B.2 and reproduced here as Figure 7.9).


118

Figure 7.7: Chromatogram of methoxy butyl biodiesel. Major peaks are: methyl oleate/linolenate (22.00

min.), butyl oleate/linolenate (31.45 min.), methoxy methyl oleate/linolenate (33.29 min.) and methoxy butyl

oleate/linolenate (42.72 min.).

Figure 7.8: Mass spectrum of ethoxy ethyl oleate.

m/z


119

Figure 7.9: Mass spectrum of ethoxy butyl oleate.

The (a) fragment at m/z 257 is absent since the ion mass is lower by 28 (C2H4) which

coincides with the (c) fragment at m/z 229. In addition, the m/z 201 ion is more prominent

in Figure 7.8 because of the reduction in (c) by 28. Transesterified alkoxy oleate was

detected by GC-MS for alkoxy biodiesels up to octoxy (Table 7.4). However,

transesterified alkoxy oleate may have been present at low levels for octoxy and

ethylhexoxy biodiesel but were not detected due to their low volatility.

m/z


120

Table 7.4: Transesterified alkoxy oleate content for all alkoxy esters. (NA = not applicable, ND = not

determined)

Material Alkoxy alkyl oleate

(%)

Methoxy butyl biodiesel 41.4

Ethoxy butyl biodiesel 9.0

Propoxy butyl biodiesel 8.0

Butoxy butyl biodiesel NA

Pentoxy butyl biodiesel 5.5

Hexoxy butyl biodiesel 5.0

Octoxy butyl biodiesel ND

Ethyl Hexoxy butyl biodiesel ND

7.2 Impact on Cloud Point Both butyl biodiesel and epoxy butyl biodiesel had relatively low cloud points at -3°C,

when compared to the corresponding methyl biodiesel (Table 7.5). The cloud point of

methoxy, ethoxy and propoxy butyl biodiesel was 0°C, a 3K increase over butyl biodiesel.

The cloud point for butoxy butyl biodiesel was -2°C, while the CP for pentoxy, hexoxy and

octoxy butyl biodiesel was -5°C. The CP for 2-EH butyl biodiesel was the lowest at -6°C.

The general trend in cloud point for the straight-chain alkoxy groups is therefore an

increase for chain lengths shorter than butyl, then a step decrease to -5°C, or 2K below that

for butyl biodiesel, for chain lengths greater than butyl. The CP reported in Table 7.5 for

tert-butoxy biodiesel was 7°C and is clearly an anomaly. GC-MS analysis showed that no

tert-butoxy biodiesel was actually produced. This is due to the tertiary nature of the alcohol

and resulted in the acid hydrolysis of the epoxy butyl biodiesel to produce vicinal

dihydroxy and keto butyl biodiesel. The much higher CP was probably due to the high

levels of vicinal dihydroxy groups as identified by Smith et al. (Smith et al. 2009).

Electrostatic attraction between the electropositive H of the hydroxy groups and the

carbonyl oxygen of an ester group, or a corresponding keto group, may facilitate the

formation of ordered crystals at a higher temperature.


121

Table 7.5: Flow properties of butyl biodiesel and modified biodiesel, including: cloud point, pour point and

kinematic viscosity. (ND = Not Determined).

Material Cloud point

(°C)

Pour point

(°C)

Viscosity

(mm2.s-1)

Methyl biodiesel 0 -12 4.08

Butyl biodiesel -3 -12 4.88

Epoxy butyl biodiesel -3 -12 6.80

Methoxy butyl biodiesel 0 -9 6.67

Ethoxy butyl biodiesel 0 -9 7.19

Propoxy butyl biodiesel 0 -9 7.32

Butoxy butyl biodiesel -2 -9 7.13

tert-Butoxy butyl biodiesel 7 > 20 ND

Pentoxy butyl biodiesel -5 -12 6.72

Hexoxy butyl biodiesel -5 -12 8.07

Octoxy butyl biodiesel -5 -12 7.92

Ethyl hexoxy butyl

biodiesel

-6 -12 9.76

The initial increase in CP is likely to be due to the straightening of the tail-group of the

esters while the reduction in CP for longer alkoxy groups is due to the physical hindrance

provided by the alkoxy chain. Chain lengths greater than butyl are therefore able to

interfere with the orderly packing of molecules as the temperature reduction brings them

closer. The branched nature of the 2-EH group appears to provide a greater steric

hindrance than either the 6 carbon hexoxy or the 8 carbon octoxy group. The position of

the branched 2-EH group at the alkyl head-group of the chain would result in a long chain

oriented away from the main tail-group (Figure 7.1, d). This would restrict access to the

carbonyl group of the ester. However, it should be noted that the accuracy of the cloud

point method and the presence of minor contaminants means that the difference may not be

significant.

A factor that makes the reduction in CP due to the longer-chain alkoxy groups even more

significant is the competing increase in CP due to the presence of transesterified biodiesel.

It has been reported that the melting point of saturated fatty esters decreases with

increasing chain length of the alkyl head-group up to butyl, then increases with a chain


122

length of 5 or greater (Knothe et al. 2004). Thus, the presence of 12.6% octyl biodiesel in

the case of octoxy butyl biodiesel would tend to increase the CP of the mixture. Yet

another factor influencing the cloud point of the alkoxy esters is the presence of

transesterified alkoxy esters. This is especially significant for methoxy butyl biodiesel

since the fraction of alkoxy biodiesel attributed to methoxy methyl biodiesel was over

41%. As identified in chapter 5, significant quantities of methoxy methyl biodiesel tended

to increase the cloud point of biodiesel. The same can be said for ethoxy ethyl and propoxy

propyl biodiesel because it was found that alkoxylation of alkyl biodiesel increased the

cloud point of the fuel up to butoxy butyl biodiesel. The impact of the presence of pentoxy

pentyl and hexoxy hexyl biodiesel on cloud point is unknown but because of the low

proportions, is not expected to have a major impact here.

7.3 Impact on Pour Point Pour point (PP) measurement is a relatively crude method since it can only be reported to

the nearest 3°C. The method also requires that 3°C is added to the temperature at which the

material fails to flow. Since the PP of all samples was well below 0°C, observations began

at that temperature and continued at 3°C intervals until the material failed to flow when the

test jar was held in a horizontal position for 5 s. The PP for those samples that had a CP of

-2°C or greater was -9°C, except for methyl biodiesel that had a reported PP of -12°C

(Table 7.5). However, it should be noted that methyl biodiesel was only just fluid at -12°C

as it only began to flow after approximately 3 s in the horizontal position. The lower PP for

methyl biodiesel than methoxy, ethoxy, propoxy and butoxy butyl biodiesel is therefore not

considered significant.

The pour point for pentoxy, hexoxy, octoxy and ethylhexoxy butyl biodiesel was -12°C.

This again is a reflection of the lower cloud point for those samples. It can be concluded

therefore that the alkoxylation of butyl biodiesel does not significantly improve (lower) the

PP of butyl biodiesel or conventional methyl biodiesel. The unusually high pour point for

the tert-butoxy biodiesel sample occurred when an attempt was made to re-heat the sample

in the test jar to confirm the surprisingly high CP. Agitation of the sample when the

temperature was approximately 10°C caused the whole sample to rapidly solidify. The

sample remained in a solid state when stored at a room temperature of approximately


123

20°C. This further supports the suggestion that electrostatic attraction associated with the

dihydroxy fraction of the biodiesel caused a much higher CP and once the molecules

approached closely enough via agitation, H-bonding occurred resulting in solidification.

7.4 Impact on Viscosity The kinematic viscosity of methyl biodiesel was 4.08 mm2.s-1 (Table 7.5) which is well

within the European Standard (EN 14214) of 3.5 to 5.0 mm2.s-1 and is very close to the

kinematic viscosity reported elsewhere for methyl biodiesel derived from canola oil (4.53

mm2.s-1) (Knothe et al. 2004). The viscosity of butyl biodiesel was slightly higher at 4.88

mm2.s-1 due to the added length of the alkyl head-group, as reported by Knothe and

Steidley (Knothe and Steidley 2005). The viscosity of epoxy butyl biodiesel was

significantly higher than butyl biodiesel at 6.80 mm2.s-1. The trend for the series of

methoxy through to propoxy butyl biodiesel is for viscosity to increase slightly. A reversal

of the trend is seen for butoxy and pentoxy butyl biodiesel while the viscosity of hexoxy,

octoxy and 2-EH butyl biodiesel trends higher again. The value for ethylhexoxy butyl

biodiesel is almost double the upper limit for biodiesel and is twice that for butyl biodiesel.

Several structural factors at the molecular level have affected the viscosity of biodiesel

including the loss of double bonds, protruding alkoxy groups on the ester tail and the

introduction of hydroxyl groups. Double bonds in fatty compounds are known to reduce

the kinematic viscosity and trans- double bonds impart a higher viscosity than cis- double

bonds (Knothe and Steidley 2005). It follows then that the elimination of 46% of the

unsaturated fraction of biodiesel will increase the viscosity of biodiesel. Butyl tallowate

has a kinematic viscosity of 6.90 mm2.s-1 (Foglia et al. 1997) and a saturated ester content

of 53%, reported separately by Zlatanic et al. (Zlatanic et al. 2004). The viscosity increase

for methoxy, ethoxy and propoxy butyl biodiesel may therefore be predominantly due to

the removal of the double bonds and the linearisation of the esters. The protruding alkoxy

groups may be likened with branching of the alkyl head-group. However, branching of

esters has less of an effect on viscosity than double bonds and Knothe and Steidley

(Knothe and Steidley 2005) found that the viscosity of branched esters did not differ

significantly from that of the straight-chain analogues with the same number of carbons.

Butyl myristate had a kinematic viscosity of 4.47 mm2.s-1 whereas iso-butyl myristate had


124

a viscosity of 4.65 mm2.s-1. In contrast, the viscosities of butyl palmitate and iso-butyl

palmitate were 6.49 and 6.02 mm2.s-1, respectively. Unfortunately, only short-chain alkyl

groups up to and including butyl were reported.

The branched nature of the 2-EH group may contribute to the greater viscosity of the 2-EH

butyl biodiesel over the equivalent hexoxy and octoxy biodiesel in two ways: the alkoxy

group in the tail and the alkyl group in the head of the transesterified portion. Interaction of

these two moieties would certainly result in hindered flow. Hydroxyl groups and in

particular, compounds with multiple hydroxy groups, are known to have a dramatic effect

on viscosity and may add further to the increased viscosity of the alkoxy biodiesel. The

effect is not expected to be highly significant in this case however because the hydroxy

groups are singular.


125

7.5 Summary Epoxidation of butyl biodiesel for 6h resulted in a conversion of 46% of the available

unsaturated portion, and selectivity for epoxy butyl biodiesel was 100%. Alkoxylation

proceeded for 1h and was confirmed as complete by GC-MS from the absence of the

epoxy butyl oleate peak and the appearance of a group of peaks identified as alkoxy butyl

oleate from their mass spectrum. Retention times progressively increased with alkoxy

chain length as expected, with 2-ethylhexoxy eluting between hexoxy and n-octoxy.

Selectivity for the respective alkoxy biodiesel was greater than 97%. Complete conversion

of epoxy biodiesel was achieved in the case of tert-butyl alcohol but no tert-butoxy

biodiesel was produced. Instead, hydrolysis of the oxirane group under the acid conditions

produced a mixture of vicinal dihydroxy and keto butyl biodiesel, in approximately equal

fractions. Transesterification of butyl biodiesel with the respective alcohol, present for the

alkoxylation reaction, occurred to varying degrees for all due to the presence of the acid

catalyst, the high molar ratio of alcohol and the elevated temperature.

Alkoxylation of butyl biodiesel with methanol, ethanol and propanol increases the cloud

and pour point of butyl biodiesel. Alkoxylation with alcohols larger than butanol can

produce significant improvements in low-temperature properties as indicated by lower

cloud and pour points. The lowest CP achieved was for ethylhexoxy butyl biodiesel at

-6°C, a 6K reduction in CP over conventional methyl biodiesel. Alkoxylation also results

in significant increases in kinematic viscosity with the viscosity of ethylhexoxy butyl

biodiesel being 9.76 mm2.s-1, more than double that for methyl biodiesel. The improved

low-temperature properties of the longer chain alkoxy biodiesel were most likely due to the

protruding alkoxy chain, which also resulted in an increase in viscosity. The use of

alcohols larger than pentanol does not provide significant benefit in terms of low-

temperature properties and results in an undesirable increase in viscosity.

Chapter 8

Discussion and Conclusions

Chapter 8 Discussion and Conclusions _________________________________________________________________________

127

DISCUSSION AND CONCLUSIONS _________________________________________________________________________

8.1 Discussion The problem of poor low-temperature properties of biodiesel is a significant one. The

cloud point of methyl biodiesel derived from highly unsaturated virgin vegetable oil is

approximately 0°C. With most commercial operations looking to move towards lower-

grade oil feedstocks, which produce biodiesel with much higher cloud points, a viable

method for improving cloud point is overdue. The already marginal cost of biodiesel

mandates that any process developed to improve cloud point must not add significantly to

the final fuel cost and therefore must use substantially the same equipment and conditions.

Demonstrated here is a method for improving the low-temperature properties of biodiesel

under mild conditions, with reasonable reaction times.

Initially, an assessment was made of whether biodiesel derived from a highly unsaturated

virgin vegetable oil could be epoxidised with 30% hydrogen peroxide under mild reaction

conditions without the use of organic solvents. Either formic acid or acetic acid can be

employed as the oxygen carrier. Acetic acid necessitates the use of sulfuric acid as a

catalyst for peroxyacetic acid production and results in a reduced overall reaction rate.

Initial epoxidation trials with formic acid as oxygen carrier, chosen for the higher reaction

rate produced unacceptably high levels of vicinal dihydroxy by-product which was

exacerbated by the relatively harsh conditions for alkoxylation. The result was methoxy

biodiesel that solidified at room temperature. This highlighted the need to minimise by-

product formation and was the impetus for an optimisation study to determine the optimal

conditions for epoxidation. Trials indicated that the slower reaction with acetic acid was

beneficial since the production of by-product was greatly reduced. Epoxidation under the

optimal conditions of H2O2 / biodiesel molar ratio of 2:1, acetic acid / biodiesel molar ratio

of 0.2:1, acid catalyst to acetic acid / peroxide of 2 wt%, 6 h reaction at 60°C, resulted in

the conversion of approximately 45% of the unsaturated fraction. Under those mild

conditions selectivity for epoxy biodiesel was 100%.


128

Nominal reaction conditions of 2.5% sulfuric acid, a molar ratio of alcohol to epoxy

biodiesel of 14:1 and a reaction time of 1h were chosen based on the literature for the

initial alkoxylation. The cloud point of methyl biodiesel with methoxy groups substituted

for 45% of the available double bonds exhibited a cloud point of 4°C, representing an

increase of 6K. Ethoxy ethyl biodiesel, substituted to approximately the same extent,

displayed a CP increase of 3K. Butoxy butyl biodiesel, also substituted to approximately

the same extent, had a CP of -4°C representing a CP improvement of 1K. This indicates a

clear trend towards a reduction in cloud point as the chain-length of the alkoxy group

increased. It must also be noted however that the cloud point of the alkyl esters also

increases in this range with increasing chain-length. The cloud point of methoxy methyl

biodiesel is well above that for methyl biodiesel though and is most likely due to the

reduction in non-linear molecules (unsaturated ester) that act to restrict the orderly packing

during temperature reduction. The evidence for this phenomenon was enhanced when butyl

biodiesel was epoxidised to a greater extent (67%) and the subsequently alkoxylated

product produced a cloud point that was 4K greater than the 45% substituted product. The

optimal substitution fraction and therefore epoxidation conditions were confirmed.

It was hoped that butoxy butyl biodiesel would have significantly better low-temperature

performance than conventional biodiesel since butanol can be produced from renewable

feedstocks in a process similar to bio-ethanol production, thus improving the renewable

credentials of the fuel. The initial work with butoxy butyl biodiesel demonstrated promise

and initially the impact of the relatively high levels of by-product on the potential for cloud

point reduction was not known. Therefore, an extensive optimisation regime to minimise

by-product content was undertaken. The exercise demonstrated that the by-product content

could be minimised without resorting to reaction conditions that are unreasonable from a

commercial processing standpoint. That is, the reaction conditions were mild, keeping the

energy costs and therefore the unit cost of the fuel low. Complete conversion was achieved

within one hour and there was no need to resort to the use of organic solvents to minimise

glycol production.

As the temperature of the reaction was increased, the reaction rate increased. An optimal

temperature of 60 to 80°C was necessary to maximise selectivity. A higher reaction

temperature reduced the glycol content but increased the keto content. As catalyst

concentration increased the reaction rate increased but so did the fraction of glycol. Once


129

again a compromise was required to minimise glycol and keto content. The molar ratio of

alcohol had a minimal effect on reaction rate once the temperature and catalyst

concentration had been optimised. Selectivity did however increase with increasing molar

ratio of alcohol and a high molar ratio of 40:1 was adopted to minimise the glycol content

while sacrificing the ketone content to some degree. The very high molar ratio of alcohol

increases the physical size requirements for a commercial plant and would potentially

result in an expensive process for removal and recycling of the excess. Selectivity for

alkoxy biodiesel was only approximately 87% for the optimised conditions but it was

shown that the by-product content achieved did not significantly impact low-temperature

properties. A commercial process could opt for a lower molar ratio of alcohol (10 to 20:1)

to keep the processing costs down.

Subsequent investigations revealed that of the two major by-products, glycol had the

largest negative impact on cloud point. In addition, increased substitution of butoxy groups

at the optimal conditions of low by-product content had a negative impact on cloud point.

This provided further evidence that a substitution fraction of 45% was optimal and that an

alkoxy side-chain of only four carbons does not sufficiently disrupt the crystallisation

mechanism. The final result in terms of cloud point was that even with the by-product

content minimised, the cloud point of butoxy butyl biodiesel was not an improvement over

unmodified butyl biodiesel.

Alkoxylation with alcohols with a longer chain was identified as a possible method for

increasing the disruption to crystallisation at low temperatures. Because of the renewable

credentials of butyl biodiesel and its superior low-temperature tolerance, the alkoxylation

of butyl biodiesel with various straight-chain and branched-chain alcohols was

investigated. As expected the low-temperature tolerance (cloud point) of the methoxy,

ethoxy and propoxy variants was 3K higher than for unmodified butyl biodiesel. Butoxy

butyl biodiesel had effectively the same cloud point as butyl biodiesel, as demonstrated

previously. The normal alcohols including pentanol, hexanol and octanol when added as an

alkoxy side-chain to butyl biodiesel produced a reduction in cloud point of 2K over butyl

biodiesel or 5K over that for conventional methyl biodiesel. 2-Ethylhexoxy butyl biodiesel

had a cloud point of -6°C, a reduction in cloud point of 3K and 6K over butyl biodiesel and

conventional biodiesel, respectively. Pour point for the longer alkoxy biodiesels was also


130

improved. The clear result here is that alkoxy side-chains with a length greater than four

carbons significantly improves the low-temperature properties of biodiesel.

The viscosity of alkoxy biodiesel was significantly increased over that for conventional

methyl biodiesel and butyl biodiesel. The trend for the series of methoxy through to

propoxy butyl biodiesel was for viscosity to increase slightly, while the viscosity of

hexoxy, octoxy and 2-EH butyl biodiesel was much higher. Several structural factors at the

molecular level have affected the viscosity of biodiesel including the loss of double bonds,

protruding alkoxy groups on the ester tail and the introduction of hydroxyl groups. The

viscosity increase for methoxy, ethoxy and propoxy butyl biodiesel may be predominantly

due to the removal of the double bonds and the linearisation of the esters. The branched

nature of the 2-EH group may contribute to the greater viscosity of the 2-EH butyl

biodiesel over the equivalent hexoxy and octoxy biodiesel due to the alkoxy group in the

tail and the alkyl group in the head of the transesterified portion.

Alkoxylation of butyl biodiesel can be performed under mild conditions to produce a fuel

that has significantly improved low-temperature tolerance over conventional biodiesel.

However, a negative impact is an increase in viscosity and an increase in production cost.

Low-grade feedstocks such as tallows and waste oils have the potential to offset the likely

increase in production costs due to their substantially lower cost. This was highlighted by

Haas et al. when they estimated that the contribution of the feedstock price to the overall

fuel price was 88% (Haas et al. 2006). These oils have higher fractions of saturated fatty

acids at approximately 50% compared with 10% for virgin vegetable oils such as canola

oil. Since the optimal fraction of alkoxy biodiesel was discovered to be � 43 wt%, the

epoxidation and alkoxylation of these lower-grade oils could result in a similar proportion

of alkoxy biodiesel. This may produce the desired improvement in low-temperature

tolerance without a large increase in unit cost for the fuel. However, the large increase in

viscosity remains as a serious problem for alkoxylated biodiesel.

Further positive impacts of the alkoxylation of biodiesel are the expected improvement in

oxidation stability, cetane number and potentially NOx emissions. Alkoxylation reduces the

proportion of unsaturated esters, which directly impacts on the oxidative stability and

cetane number of the fuel. Oxidative stability refers to the autoxidation of the double bonds

in the tail-group of the fatty acid chains of biodiesel. The positions allylic to double bonds


131

are especially susceptible to autoxidation under extended storage conditions. The bis-

allylic positions such as those present in linoleic (C18:2) and linolenic (C18:3) acids are

even more prone to oxidation. Hence, the long-term storage stability of biodiesel can be

correlated with the number and position of double bonds. The alkoxylation process

demonstrated here reduces the number of double bonds by almost half. Moser and Erhan

demonstrated that �-hydroxy ethers such as those produced here have significantly higher

oxidation onset and signal maximum temperatures than normal soybean methyl esters,

indicating that oxidation stability may be significantly improved (Moser and Erhan 2008).

The cetane number of a fuel is a measure of the ignition delay, that is, the time that passes

between injection of the fuel into the cylinder and ignition. The shorter the ignition delay,

the higher the cetane number. Higher cetane numbers in petroleum diesel have also been

associated with lower NOx emissions. Despite the fact that biodiesels generally have higher

cetane numbers, emissions of NOx from these fuels are slightly higher than for petroleum

diesel. However, the increase in NOx for biodiesel is a complex phenomenon and is

generally not considered to be connected to the cetane number. The higher cetane number

of biodiesel is considered as a positive. Studies have shown that as levels of unsaturation

increase, cetane number decreases, and as the chain length increases the cetane number

increases. As the fraction of unsaturated esters has reduced significantly here for

alkoxylated biodiesel and the chain length has increased (butyl compared to methyl

biodiesel), the cetane number would more than likely be higher.

Alkoxylation of most of the unsaturated fraction of a biodiesel produced from a low-grade

feedstock would reduce the fraction of unsaturated ester to virtually zero. It would

therefore follow that the cetane number and oxidation stability would be enhanced. What is

not clear however is whether the cloud point of such a fuel would be better (lower) than

conventional biodiesel. As was seen with butoxy butyl biodiesel, alkoxylation at higher

rates of up to 93% of the unsaturated fraction, resulted in impaired low-temperature

tolerance. This was most likely due to the very low fraction of unsaturated ester, as would

be the case with the aforementioned alkoxylated biodiesel produced from low-grade

feedstock. There is therefore a real potential for such a fuel to have a high cloud point and

maybe even higher than the conventional methyl ester of the feedstock.


132

8.2 The Economics of Alkoxylated Biodiesel A full process synthesis and economic analysis is outside the scope of this work but it may

be useful to discuss further the likely cost implications of this method for improving the

low-temperature tolerance of biodiesel. The addition of 2-ethylhexanol produced the

greatest reduction in cloud point (6K) and the use of waste canola oil, rather than virgin oil

would significantly reduce the final price of the fuel. In addition, it would be desirable to

utilise bio-butanol produced from a renewable source such as corn. The work conducted by

Zhang et al. (2003) has been used to roughly calculate the total manufacturing cost of

ethylhexoxy butyl biodiesel and to compare that value with their finding for methyl

biodiesel. An outline of the proposed process is provided in Figure 8.1 and Table 8.1

contains some estimated material costs.

The process would begin with the butanolysis of waste cooking oil to produce butyl

biodiesel. As discussed in chapter 4, unlike methanolysis, butanolysis would be a two-

stage process to ensure that conversion and selectivity were high enough to meet the

biodiesel standard. Process III discussed by Zhang et al. (2003) is a two-step acid catalysed

process from waste cooking oil and would therefore be a reasonable analogue for the two-

stage butanolysis of waste cooking oil, at least in terms of the capital and operating costs.

The reaction temperature would be much higher for n-butanol to reflect the reduced

volatility of butanol over methanol so the manufacturing cost determined by Zhang et al.

would be conservative here. The cost of the oil feedstock and methanol accounted for 38%

of the total manufacturing cost, after glycerol credit, at $1.65M (million US dollars) and

$0.31M, respectively for a 8000 tonne/yr plant. Based on a butanol price of $550/tonne

(Qureshi and Blaschek, 2000), the alcohol cost would rise to $2.19M. Therefore, the total

manufacturing cost of the biodiesel would rise from $5.15M to $7.03M.


133

Figure 8.1: Conceptual process flow for ethylhexoxy butyl biodiesel production.

Butanolysis (1st)

Butanolysis (2nd)

Epoxidation

Alkoxylation

Vacuum distillation

Crude butyl biodiesel

Vacuum distillation

Neutralisation and water washing

Epoxy butyl biodiesel

Vacuum Distillation

Butyl biodiesel

Ethylhexoxy butyl

Waste canola oil

butanol, sodium hydroxide

glycerol

butanol, sodium hydroxide

glycerol

Water, sodium bicarbonate

Water, sodium salts

2-ethylhxanol, sulphuric acid,

hexane

2-ethyl hexanol,

Hydrogen peroxide, acetic acid, sulphuric acid


134

Epoxidation and alkoxylation of the butyl biodiesel would follow and while the process

conditions are mild, more materials and equipment would be required. In order to make an

approximate estimation of the total manufacturing cost for this part of the process, one

could assume that the capital and operating costs are the same as for the original biodiesel

manufacturing process. The processing steps are similar in nature to for the two processes

with mixing, two reaction steps and purification by distillation. It should also be

remembered that the majority of the total manufacturing cost has been attributed to the cost

of materials by authors including Hass et al. (2006) and Zhang et al. (2003). The major

material cost is likely to be for 2-ethylhexanol. The mass of 2-ethylhexanol required based

on a 0.5:1 molar ratio with biodiesel (approximately 50% conversion of unsaturated

fraction of biodiesel as in this work) would be 1750 tonne. The price listed in Table 8.1

was adjusted for the likely price in 2003 by comparing the price for methanol listed by

Zhang et al. (2003) and that listed by the same source as that for the 2-ethylhexanol. The

cost of the 2-ethylhexanol alone would be $2.29M. The total manufacturing cost for the

epoxidation and alkoxylation steps would therefore be $3.61M.

Table 8.1: Estimates of some critical biodiesel raw material costs.

Material Price (USD/tonne)

Methanol 275 a

Methanol 180b

n-Butanol 550 c

2-Ethyl hexanol 1991 a

Waste cooking oil 200 b

Hexane 410 b

Virgin cooking oil 500 b

Sodium hydroxide 4000 b

Sulfuric acid 60 b

Biodiesel 600 b

aSource: ICIS Pricing, http://www.icispricing.com

bSource: Zhang et al. (2003) Multiplied by 1.5 to estimate current pricing based on the methanol price rise.

C Source: Qureshi and Blaschek (2000)


135

The total manufacturing cost for ethylhexoxy butyl biodiesel would therefore be

approximately $10.65M compared to $5.15M for a 8000 tonne/yr plant. The value of

$5.15M calculated by Zhang et al. (2003) for methyl biodiesel was less than the revenue

based on a biodiesel price of $600/tonne making methyl biodiesel an unviable process. The

real manufacturing cost of ethylhexoxoy butyl biodiesel is likely to be even higher than the

value calculated here, especially given the expected additional energy costs of this process.

2-Ethylhexanol has a normal boiling point of 185°C and is not soluble in water. Removal

of the excess 2-ethylhexanol will therefore most likely be achieved by distillation as

employed in this work. Even under high vacuum, serious product degradation can occur,

hence the necessity for an organic solvent during purification, potentially hexane. The

introduction of an organic solvent in a commercial process is undesirable for reasons of

safety, process complexity and cost. Unit operations for both the addition and removal of

the solvent would be required, greatly adding to capital and running costs. Similarly, the

evaporative removal of the excess 2-ethylhexanol would be costly, especially given the

high molar ratio of 40:1. The epoxidation and neutralisation steps require the introduction

of further materials such as hydrogen peroxide, acetic acid, sulphuric acid and sodium

bicarbonate which have not been taken into account here.

The development of an alternative process for thoroughly drying the biodiesel product

prior to alcohol removal would render the use of an organic solvent redundant. However,

the evaporation and recycling of the excess 2-ethylhexanol would remain as a serious

challenge for both capital and operating cost minimisation. The use of a shorter alcohol

such as pentanol that provides substantially the same improvement in low-temperature

properties but higher volatility and lower volume may be advantageous. The higher

volatility would reduce the energy requirements for recovery of the excess and the reduced

volume would lower the capital cost of equipment. It is clear however that the 6K gain in

low-temperature tolerance would not justify the likely increase in the price of the fuel. It is

suggested that fuel distributors would continue to blend petroleum diesel with biodiesel

and use the existing techniques for “winterising” the fuel in cold climates.


136

8.3 Further Work Suggestions for further work must include the scenario mentioned above of the

alkoxylation of a biodiesel produced from a low-grade feedstock under the optimised

conditions discovered in this investigation. The impact of alkoxylation on other

characteristics of biodiesel such as oxidation stability, cetane number and exhaust

emissions should also be assessed. The next logical step in this line of investigation is the

use of the position � to the alkoxy group to attach another branch to the chain opposing the

alkoxy chain to determine whether that has a more significant impact on cloud point. The

low reactivity of this site would however necessitate more severe reaction conditions or

much more aggressive reactants. The likely impact of such a transformation on other fuel

properties would similarly need to be determined. One suspects that the viscosity of this

fuel would be even higher than that for the fuel developed here. However, such a highly

branched version of biodiesel may more efficiently impede the close packing of molecules

at low temperatures and may only need to be added at low concentrations. This may lessen

the negative impact on the other properties such as viscosity. If only small quantities at

additive levels were required, only a small fraction of the product would require diversion

for modification. The process could therefore be performed on a much smaller scale and

could employ the much harsher conditions necessary without greatly impacting the final

fuel price.


137

8.4 Conclusions

� Methods developed for the transesterification of canola oil were successful for the

synthesis of methyl, ethyl and butyl biodiesel.

� FTIR was able to positively identify the reaction products as epoxy compounds.

� FTIR was able to confirm that alkoxy compounds were present due to the existence

of bands for secondary ethers and was able to confirm the presence of glycol by-

product.

� GC-MS proved to be an accurate and reliable quantitative method for the

determination of conversion, selectivity and the impurity profile of both epoxy and

alkoxy biodiesel.

� Optimal conditions for the epoxidation of canola derived biodiesel were H2O2 to

biodiesel molar ratio of 2:1, acetic acid to biodiesel molar ratio of 0.2:1, acid

catalyst to acetic acid / peroxide of 2 wt%, 6h reaction at 60°C.

� Selectivity for epoxy biodiesel over vicinal dihydroxy biodiesel was 99.8%, 99.8%

and 100.0% for methyl, ethyl and butyl biodiesel, respectively.

� Alkoxylation for 1h at 60°C, alcohol to epoxy biodiesel molar ratio of 14:1 and 2.5

wt% sulfuric acid resulted in alkoxy substitution rates of 37.1% (methyl), 34.3%

(ethyl) and 32.9% (butyl). Selectivity for alkoxy biodiesel was 89.0%, 82.7% and

81.7% for methoxy, ethoxy and butoxy biodiesel, respectively.

� Cloud point for methyl and ethyl biodiesel increased from -2°C to 4°C, and from

-3°C to 0°C, respectively. The cloud point of butyl biodiesel reduced from -3°C to

-4°C.

� Butyl biodiesel derived from canola oil was epoxidised via the in-situ peroxyacetic

acid method for 6h resulting in a conversion of 46% of the unsaturated esters.

� Optimal conditions for the butoxylation of epoxy butyl biodiesel were: 80°C, 2wt%

sulfuric acid and a 40:1 molar ratio of butanol over a period of 1h. Conversion of

epoxy butyl biodiesel was 100% and selectivity for butoxy biodiesel was 87.0%.

� The cloud point of butoxy butyl biodiesel (conversion of 46% of unsaturated

portion) was identical to that for butyl biodiesel.

� Butoxy butyl biodiesel at a conversion of 93% of unsaturated portion had a cloud

point 5K higher than that for butyl biodiesel.


138

� Blends of the high conversion batch of butoxy biodiesel showed that cloud point

was virtually unchanged at concentrations below 35% and then increased 1K every

8 wt% to approximately 70 wt % butoxy biodiesel.

� Butoxy biodiesel is able to prevent the earlier onset of crystallisation due to the

decrease in unsaturated content but only at concentrations below 35%.

� Butyl biodiesel can be butoxylated under mild conditions and reasonable reaction

times but this process does not offer an improvement in biodiesel cloud point.

� Epoxidation of butyl biodiesel for 6h resulted in a conversion of 46% of the

available unsaturated portion, and selectivity for epoxy butyl biodiesel was 100%.

� Alkoxylation of butyl biodiesel with methanol, ethanol and propanol increases the

cloud and pour point of butyl biodiesel.

� Alkoxylation of butyl biodiesel with alcohols larger than butanol can produce

significant improvements in low-temperature properties as indicated by lower cloud

and pour points.

� The lowest CP achieved was for ethylhexoxy butyl biodiesel at -6°C, a 6K

reduction in CP over conventional methyl biodiesel.

� Alkoxylation also results in significant increases in kinematic viscosity with the

viscosity of ethylhexoxy butyl biodiesel being 9.76 mm2.s-1, more than double that

for methyl biodiesel.

� The improved low-temperature properties of the longer chain alkoxy biodiesel were

most likely due to the protruding alkoxy chain, which also resulted in an increase in

viscosity.

� The use of alcohols larger than pentanol do not provide significant benefit in terms

of low-temperature properties and results in an undesirable increase in viscosity.

� A process that produces ethyhexoxy butyl biodiesel would not be an economically

viable option for achieving a 6K improvement in cloud point.

Appendix A

Appendix A Chromatograms _________________________________________________________________________

140

APPENDIX A _________________________________________________________________________

Chromatograms Included here are the chromatograms for the alkoxy butyl biodiesel batches referred to in

chapter 7. They are samples taken from the final product following purification and

represent the product as it was tested for flow-properties. The major peaks are annotated

with the retention time in minutes and the relative area. The time scale is in minutes and

the area scale is normalised to the tallest peak.

Figure A.1: Chromatogram of methoxy butyl biodiesel. Major peaks are: methyl oleate/linolenate (22.00

min.), butyl oleate/linolenate (31.45 min.), methoxy methyl oleate/linolenate (33.29 min.) and methoxy butyl



141

Figure A.2: Chromatogram of ethoxy butyl biodiesel. Major peaks are: ethyl oleate/linolenate (24.24 min.),

butyl oleate/linolenate (31.51 min.), ethoxy ethyl oleate/linolenate (36.61 min.) and ethoxy butyl



142

Figure A.3: Chromatogram of propoxy butyl biodiesel. Major peaks are: propyl oleate/linolenate (26.9 min.),

butyl oleate/linolenate (31.51 min.) and propoxy butyl oleate/linolenate (46.30 min.).


143

Figure A.4: Chromatogram of butoxy butyl biodiesel. Major peaks are: butyl oleate/linolenate (31.52 min.)

and butoxy butyl oleate/linolenate (49.25 min.).


144

Figure A.5: Chromatogram of tert-butoxy butyl biodiesel. Major peaks are: butyl oleate/linolenate (31.51

min.), keto butyl oleate (38.81 min.) and dihydroxy butyl oleate/linolenate (44 - 46 min.).


145

Figure A.6: Chromatogram of pentoxy butyl biodiesel. Major peaks are: pentyl oleate/linolenate (34.84

min.), butyl oleate/linolenate (31.55 min.) and pentoxy butyl oleate/linolenate (52.28 min.).


146

Figure A.7: Chromatogram of hexoxy butyl biodiesel. Major peaks are: hexyl oleate/linolenate (38.41 min.),

butyl oleate/linolenate (31.51 min.) and hexoxy butyl oleate/linolenate (55.98 min.).


147

Figure A.8: Chromatogram of octoxy butyl biodiesel. Major peaks are: octyl oleate/linolenate (45.34 min.),

butyl oleate/linolenate (31.46 min.) and octoxy butyl oleate/linolenate (66.21 min.).


148

Figure A.9: Chromatogram of ethylhexoxy butyl biodiesel. Major peaks are: ethylhexyl oleate/linolenate

(42.67 min.), butyl oleate/linolenate (31.56 min.) and ethylhexoxy butyl oleate/linolenate (59.91 min.).

Appendix B

Appendix B Mass Spectra _________________________________________________________________________

150

APPENDIX B _________________________________________________________________________

Mass Spectra Included here are the mass spectra for the alkoxy butyl biodiesel batches referred to in

chapter 7. Each mass spectrum represents the fragmentation pattern for the peak identified

as the alkoxy butyl oleate peak.

Figure B.1: Mass spectrum of methoxy butyl oleate.

m/z


151

Figure B.2: Mass spectrum of ethoxy butyl oleate.

Figure B.3: Mass spectrum of ethoxy ethyl oleate.

m/z

m/z


152

Figure B.4: Mass spectrum of propoxy butyl oleate.

m/z


153

Figure B.5: Mass spectrum of butoxy butyl oleate.

m/z


154

Figure B.6: Mass spectrum of pentoxy butyl oleate.

m/z


155

Figure B.7: Mass spectrum of hexoxy butyl oleate.

m/z


156

Figure B.8: Mass spectrum of octoxy butyl oleate.

m/z


157

Figure B.9: Mass spectrum of ethylhexoxy butyl oleate.

m/z

Nomenclature _________________________________________________________________________

158

NOMENCLATURE _________________________________________________________________________

Symbols A Area of peak

A Reactant A

E Activation energy

k Specific reaction rate constant

m Mass fraction

R Universal gas constant

S Selectivity

T Temperature

X Conversion

Subscripts

A Alkoxy

B Biodiesel (oleate/linolenate)

E Epoxy biodiesel

G Glycol biodiesel

K Keto biodiesel

P Palmitate

0 Initial

1 Final

Bibliography _________________________________________________________________________

159

BIBLIOGRAPHY _________________________________________________________________________

Cermack, S. C., Skender, A. L., Deppe, A. B., and Isbell, T. A. (2007). "Synthesis and

Physical Properties of Tallow-Oleic Estolide 2-Ethylhexyl Esters." Journal of the

American Oil Chemist's Society, 84, 449-456.

Chandler, J. E., Horneck, F. G., and Brown, G. I. (1992). "The Effect of Cold Flow

Additives on Low-Temperature Operability of Diesel Fuels." In: SAE Technical

Paper Series, Warrendale.

Dalhlke, B., Hellbardt, S., Paetow, M., and H, Z. W. (1995). "Polyhydroxy Fatty Acids and

Their Derivatives from Plant Oils." Journal of the American Oil Chemist's Society,

72(3), 349-353.

Demirbas, A. (2005). "Biodiesel Production from Vegetable Oils via Catalytic and Non-

Catalytic Supercritical Methanol Transesterification Methods." Progress in Energy

and Combustion Science, 31, 466-487.

Demirbas, A. (2008). "Progress and Recent Trends in Biodiesel Fuels." Energy Conversion

and Management, 50, 14-34.

Doll, K. M., Sharma, B. K., and Erham, S. Z. (2007). "Synthesis of Branched Methyl

Hydroxy Stearates Including an Ester from Bio-Based Levulinic Acid." Industrial

& Engineering Chemistry Research, 46, 3513-3519.

Duffield, J., Shapouri, H., Graboski, M., McCormick, R., and Wilson, R. (1998). "U.S.

Biodiesel Development: New Markets for Conventional and Genetically Modified

Agricultural Products." 770, USDA.

Dunn, R. O., Shockley, M. W., and Bagby, M. O. (1996). "Improving the Low-

Temperature Properties of Alternative Diesel Fuels: Vegetable Oil-Derived Methyl

Esters." Journal of the American Oil Chemists' Society, 73(12), 1719-1728.

Dunn, R. O., Shockley, M. W., and Bagby, M. O. (1997). "Winterized Methyl Esters form

Soybean Oil: An Alternative Diesel Fuel with Improved Low-Temperature Floe

Properties." In: State of Alternative Fuel Technologies, SAE Special Publications,

Warrendale, P A, 133-142.

EBB. (2008). "2007-2008 Production Statistics Show Restrained Growth in the EU Due to

Market Conditions and Competition from US B99 Imports."

Bibliography _________________________________________________________________________

160

Foglia, T. A., Nelson, L. A., Dunn, R. O., and Marmer, W. N. (1997). "Low-Temperature

Properties of Alkyl Esters of Tallow and Grease." Journal of the American Oil

Chemist's Society, 74(8), 951-955.

Gonzalez Gomez, M. E., Howard-Hildige, R., Leahy, J. J., and Rice, B. (2002).

"Winterization of Waste Cooking Oil Methyl Ester to Improve Cold Flow

Temperature Fuel Properties." Fuel, 81, 33-39.

Haas, M. J., McAloon, A. J., Yee, W. C., and Foglia, T. A. (2006). "A Process Model to

Estimate Biodiesel Production costs." Bioresource Technology, 97, 671-678.

Hanna, M. A., Ali, Y., Cuppett, S. L., and Zheng, D. (1996). "Crystallization

Characteristics of Methyl Tallowate and Its Blends with Ethanol and Diesel Fuel."

Journal of the American Oil Chemists' Society, 73, 759-763.

Harrow, G. (2007). "E85 and Biodiesel Deployment, NREL/PR-540-42312." In: Clean

Cities Activity, National Renewable Energy Laboratory, Golden, CO, USA.

Hasenhuetti, G. L. (2005). "Fats and Fatty Oils." In: Kirk-Othmer Encyclopedia of

Chemical Technology, John Wiley & Sons, Inc., New York, 816.

Hwang, H.-S., Adhvaryu, A., and Erhan, S. Z. (2003). "Preparation and Properties of

Lubricant Basestocks from Epoxidized Oil and 2-Ethylhexanol." Journal of the

American Oil Chemist's Society, 80(8), 811-815.

Hwang, H.-S., and Erhan, S. Z. (2001). "Modification of Epoxidized Soybean Oil for

Lubricant Formulations with Improved Oxidative Stability and Low Pour Point."

Journal of the American Oil Chemist's Society, 78(12), 1179-1184.

Imahara, H., Minami, E., and Saka, S. (2006). "Thermodynamic study on cloud point of

biodiesel with its fatty acid composition." Fuel, 85, 1666-1670.

Kinney, A. J., and Clemente, T. E. (2005). "Modifying Soybean Oil for Enhanced

Performance in Biodiesel Blends." Fuel Processing Technology, 86, 1137-1147.

Knothe, G. (2005). "Dependence of Biodiesel Fuel Properties on the Structure of Fatty

Alkyl Esters." Fuel Processing Technology, 86, 1059-1070.

Knothe, G., Dunn, R. O., Shockley, M. W., and Bagby, M. O. (2000). "Synthesis and

Characterization of Some Long-Chain Diesters with Branched or Bulky Moieties."

Journal of the American Oil Chemists' Society, 77(8), 865-871.

Knothe, G., Krahl, J., and Van Gerpen, J. (2004). The Biodiesel Handbook, First Ed.,

AOCS Press, Champaign, Illinois.

Bibliography _________________________________________________________________________

161

Knothe, G., and Steidley, K. A. (2005). "Kinematic viscosity of bodiesel fuel components

and related compounds. Influence of compound structure and comparison to

petrodiesel fuel components." Fuel, 84, 1059-1065.

La Scala, J., and Wool, R. P. (2002). "The Effect of Fatty Acid Composition on the

Acrylation Kinetics of Epoxidized Triglycerols." Journal of the American Oil

Chemist's Society, 79(1), 59-63.

Lapuerta, M., Herreros, J. M., Lyons, L. L., Garcia-Contreras, R., and Bricerio, Y. (2008)

"Effect of the Alcohol Type Used in the Production of Waste Cooking Oil

Biodiesel on Diesel Performance and Emissions." Fuel, 87, 3161-3169.

Lathi, P. S., and Mattiasson, B. (2007). "Green Approach for the Preparation of

Biodegradable Lubricant Base Stock from Epoxidized Vegetable Oil." Applied

Catalysis B: Environmental, 69, 207-212.

Lee, I., Johnson, L. A., and Hammond, E. G. (1995). "Use of Branched-Chain Esters to

Reduce the Crystallization Temperature of Biodiesel." Journal of the American Oil

Chemists' Society, 72(10), 1155-1160.

Lee, I., Johnson, L. A., and Hammond, E. G. (1996). "Reducing the Crystallization

Temperature of Biodiesel by Winterizing Methyl Soyate." Journal of the American

Oil Chemists' Society, 73, 631-636.

Lin, B., Yang, L., Dai, H., and Yi, A. (2008). "Kinetic Studies of Oxirane Cleavage of

Epoxidized Soybean Oil by Methanol and Characterization of Polyols." Journal of

the American Oil Chemist's Society, 85, 113-117.

Markley, K. (1967). Fatty Acids: Their chemistry, properties, production, and uses,

Second Ed., Interscience Publishers, New York.

Ming, T. C., Ramli, N., Lye, O. T., Said, M., and Kasim, Z. (2005). "Strategies for

Decreasing the Pour Point and Cloud Point of Palm Oil Products." European

Journal of Lipid Science and Technology, 107, 505-512.

Moser, B. R., and Erhan, S. Z. (2006). "�Synthesis and Evaluation of a Series of �-

Hydroxy Ethers Derived from Isopropyl Oleate." Journal of the American Oil

Chemists' Society, 83(11), 959-963.

Moser, B. R., and Erhan, S. Z. (2008). "Branched Chain Derivatives of Alkyl Oleates:

Tribological, Rheological, Oxidation, and Low Temperature Properties." Fuel, 87,

2253-2257.

Bibliography _________________________________________________________________________

162

Moser, B. R., Sharma, B. K., Doll, K. M., and Erhan, S. Z. (2007). "Diesters From Oleic

Acid: Synthesis, Low Temperature Properties, and Oxidation Stability." Journal of

the American Oil Chemist's Society, 84, 675-680.

Mungroo, R., Pradhan, N. C., Goud, V. V., and Dalai, A. K. (2008). "Epoxidation of

Canola Oil with Hydrogen Peroxide Catalyzed by Acidic Ion Exchange Resin."

Journal of the American Oil Chemist's Society, 85, 887-896.

Noureddini, H. (2002). "System and process for producing biodiesel fuel with reduced

viscosity and a cloud-point below thirty two (32) degree Fahrenheit." In: 6,174,501,

United States of America.

Pages, X., and Alfos, C. (Year). "Synthesis of New Derivatives from Vegetable Sunflower

Oil Methyl Esers via Epoxydation and Oxirane Opening." 15th International

Sunflower Conference, Oleaginex, Corps Gras, Lipides, La filiere.

Park, J. Y., Kim, D. K., Lee, J. P., Park, S. C., Kim, Y. J., and Lee, J.-S. (2007). "Blending

Effects of Biodiesels on Oxidation Stability and Low Temperature Flow

Properties." Bioresource Technology.

Qureshi, N. and Blaschek, H. P. (2000). "Economics of Butanol Fermentation Using

Hyper-Butanol Producing Clostridium Beijerinckii BA101" Transaction of

IChemE, 78, Part C, 139-144.

Ribeiro, N. M., Pinto, A. C., Quintella, C. M., da Rocha, G. O., Teixeira, L. S. G.,

Guarierio, L. L. N., do Carmo Rangal, M., Veloso, M. C. C., Rezende, M. J. C., da

Cruz, R. S., de Oliveira, A. M., Torres, E. A., and de Andrade, J. B. (2007). "The

Role of Additives for Diesel and Diesel Blended (Ethanol or Biodiesel) Fuels: A

Review." Energy & Fuels, 21, 2433-2445.

Rios, L. A., Weckes, P. P., Schuster, H., and Hoelerich, W. F. (2005). "Resin catalyzed

alcoholysis of epoxidized fatty esters: Effect of the alcohol and the resin

structures." Applied Catalysis A: General, 284, 155-161.

Sharma, B. K., Adhvaryu, A., Liu, Y., and Erhan, S. Z. (2006). "Chemical Modification of

Vegetable Oils for Lubricant Applications." Journal of the American Oil Chemist's

Society, 83(2), 129-136.

Smith, P. C., Ngothai, Y., Nguyen, Q. D., and O'Neill, B. K. (2009). "Alkoxylation of

Biodiesel and its Impact on Low-Temperature Properties." Fuel, 88, 605-612.

Socrates, G. (2004). Infrared and Raman Characteristic Group Frequencies, Tables and

Charts, Third Ed., John Wiley & Sons, Ltd.

Bibliography _________________________________________________________________________

163

Soriano Jr., N. U., Migo, V. P., and Matsumura, M. (2006). "Ozonized vegetable oil as

pour point depressant for neat biodiesel." Fuel, 85, 25-31.

Swern, D. (1967). Fatty Acids: Their chemistry, properties, production, and uses, Second

Ed., Interscience Publishers, New York.

Van Gerpen, J (2005). "Biodiesel Processing and Production" Fuel Processing

Technology, 86, 1097-1107.

Wang, X., and Zhang, Z. (2006). "Preparation of Polystyrene Latex Particles by �-Rays-

induced Emulsifier-free Emulsion Polymerization." Radiation Physics and

Chemistry, 75, 1001-1005.

Wilson, R., and Lyall, K. (2002). "Simultaneous Determination by GC-MS of Epoxy and

Hydroxy FA as Their Methoxy Derivatives." Lipids, 37(9), 917-924.

Wilson, R., Smith, R., Wilson, P., Shephard, M. J., and Reimersma, R. A. (1997).

"Quantitative Gas Chromatography-Mass Spectrometry Isomer-Specific

Measurement of Hydroxy Fatty Acids in Biological Samples and Food as a Marker

of Lipid Peroxidation." Analytical Biochemistry, 248, 76-85.

Yunus, R., Fakhru'l-Razi, A., Ooi, T. L., Omar, R., and Idris, A. (2005). "Synthesis of

Palm Oil Based Trimethylolpropane Esters with Improved Pour Points." Industrial

& Engineering Chemistry Research, 44, 8178-8183.

Zhang, Y., Dube, M. A., McLean, D. D., Kates, M. (2003), "Biodiesel Production From

Waste Cooking Oil: 2. Economic Assessment and Sensitivity Analysis."

Bioresource Technology, 90, 229-240.

Zlatanic, A., Lava, C., Zhang, W., and Petrovic, Z. S. (2004). "Effect of Structure on

Properties of Polyols and Polyurethanes Based on Different Vegetable Oils."

Journal of Polymer Science, 42, 809-819.

Documents

Glycerin Conversion and its Impact on Biodiesel · The Alkoxylation of Biodiesel and its Impact on Fuel Properties Paul C. Smith Thesis submitted for the degree of Doctor of Philosophy