UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Cold flow improvements to biodiesel through the use of heterogeneous catalytic skeletal isomerization Reaume, Stephen John 2013

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
24-ubc_2013_fall_reaume_stephen.pdf [ 2.23MB ]
Metadata
JSON: 24-1.0103370.json
JSON-LD: 24-1.0103370-ld.json
RDF/XML (Pretty): 24-1.0103370-rdf.xml
RDF/JSON: 24-1.0103370-rdf.json
Turtle: 24-1.0103370-turtle.txt
N-Triples: 24-1.0103370-rdf-ntriples.txt
Original Record: 24-1.0103370-source.json
Full Text
24-1.0103370-fulltext.txt
Citation
24-1.0103370.ris

Full Text

     Cold Flow Improvements to Biodiesel through the use of Heterogeneous Catalytic Skeletal Isomerization   by STEPHEN JOHN REAUME    A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY  in  THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES   (Chemical and Biological Engineering)   THE UNIVERSITY OF BRITISH COLUMBIA  (Vancouver)    October 2013  ? Stephen John Reaume, 2013  ii  Abstract  Biodiesel is a promising alternative to petroleum diesel with the potential to reduce overall net CO2 emissions. However, the high cloud point of biodiesel must be reduced when used in cold climates. Cloud point is the temperature at which solid crystals first start to appear. Skeletal isomerization of biodiesel and/or its feedstocks was investigated to reduce the high cloud point. Catalytic isomerization and hydroisomerization reactions were carried out on pure unsaturated fatty acid (UFA) and saturated fatty acid (SFA) samples, respectively. The catalyst used for the experiments was a beta zeolite and 0.5 wt% Pt-doped beta zeolite for the isomerization and hydroisomerization reactions, respectively. Reaction conditions of temperature, pressure, co-catalyst and time were varied to find an optimal reduction in the cloud point of the products. It was concluded that isomerization was unsuccessful at reducing the cloud point; in contrast, hydroisomerization was successful at reducing cloud point. A 10?C reduction was achieved at 285?C and 4.0 MPa H2 pressure. The next stage of the research studied the combined effects of isomerization and hydroisomerization on a mixture of UFAs and SFAs, namely oleic and palmitic acids. It was shown that the combination of the reaction gave a cloud point reduction of 7.5?C on a 55/45 mass ratio of oleic to palmitic acids. These results led to the conclusion that SFAs and UFAs through skeletal isomerization can reduce the cloud point of a mixture of fatty acids. Thus, vegetable oil feedstocks can be improved for their biodiesel cloud point. A study of ten different oils was conducted with varying contents of fatty acids. Results have shown that high unsaturated fatty acid biodiesels increased in cloud point, due to the hydrogenation side reaction. In contrast, low unsaturated fatty acid biodiesels yielded cloud point reductions, and overall improvement in the flow properties. A maximum cloud point reduction of 16.5?C was observed with coconut oil as the starting material. These results led to the design of an optimal cloud point improvement process of for vegetable oil biodiesel of: hydrolysis (of vegetable oil) ? hydroisomerization (300?C, 4.0 MPa H2 pressure) ? esterification.     iii  Preface  During my initial coursework I was part of a case study for educational research that led to the publication #1. I carried out the case study and all other writing was performed by Dr. Andreas Mahecha-Botero and Dr. Naoko Ellis and edited for publication by Dr. John Grace.  1. Mahecha-Botero, A., Reaume, S., Grace, J.R., Ellis, N. ?Independent research as a teaching tool in graduate chemical reaction engineering. Case study: Modelling isomerization of unsaturated fatty acids with catalyst deactivation? Education for Chemical Engineers. 2011; 6(1): e1-e9.  Selection of the initial idea of isomerization of unsaturated fatty acids as a method of cloud point reduction was proposed by Dr. Ellis. The experimental design, methods and analysis of the results were carried out by myself. This initial research led to the publication #2.  2. Reaume, S.J. and Ellis, N. ?Optimizing reaction conditions for the isomerization of fatty acids and fatty acid methyl esters to their branch chain products? Journal of American Oil Chemists Society. 2011; 88(5): 661-671.  The experimental analysis and writing was carried out by myself and edited for publication by Dr. Ellis. The other research areas that compose Chapters 3, 4 and 5 were designed and carried out by myself with advice and guidance from Dr. Ellis, Dr. Kevin Smith and Dr. Glenn Sammis. This work led to publications 3, 4 and 5.  3. Reaume, S.J. and Ellis, N. ?Synergistic effects of skeletal isomerization on oleic acid and palmitic acid mixtures for the reduction in cloud point of their methyl esters? Energy and Fuels. 2012; 26(7): 4514-4520.  4. Reaume, S.J. and Ellis, N. ?Use of hydroisomerization to reduce the cloud point of saturated fatty acids and methyl esters used in biodiesel production? Biomass and Bioenergy. DOI: 10.1016/j.biombioe.2012.12.008.  5. Reaume, S.J. and Ellis, N. "Cold flow improvements to vegetable oil based biodiesel using isomerization and hydroisomerization reactions" Energies. 2013; 6: 619-633. For publications 3, 4 and 5, all methods, analysis and writing were carried out by myself and edited for publication by Dr. Ellis. Publication 3 has a molecular simulation that was performed with the aid of Dr. Farnaz Sootodeh.    iv  Table of Contents  ABSTRACT ?????????????????????????????.........ii PREFACE ?????????????????????????????...........iii TABLE OF CONTENTS  ???????????????????????...........iv LIST OF TABLES  ?????????????????????????...........viii LIST OF FIGURES ??????????????????....??????..............x LIST OF ABBREVIATIONS ?????????????????...........................xiii NOMENCLATURE ????????????????????????...............xv ACKNOWLEDGEMENTS ??????????????????????.........xvii DEDICATION ???????????????????????????........xviii CHAPTER 1  1.1 Global Warming and CO2 Rise ??????????????................1 1.2 Vegetable Oils and Biodiesel  1.2.1 Food Production ????????????????.............4 1.2.2 Other Uses for Fatty Acids ????????????.............6 1.2.3 Biodiesel Fuel Quality ??????????????............8 1.3 Improvement Methods  1.3.1 Additives ???????????????????...........11 1.3.2 Winterization ?????????????????.............11 1.3.3 Branched Higher Alcohols ????????????............12 1.3.4 Epoxidation/Alkoxylation ????????????.............12 1.3.5 Addition/Substitution Reactions ??????????... ........13 1.3.6 Isomerization/Hydroisomerization  ?????????...........16 1.4 Isomerization and Hydroisomerization Reactions  1.4.1 Background ??????????????????...........16 1.4.2 Mechanisms  ??????????????????..........18 1.4.3 Kinetics and Equilibrium  ?????????????..........21 1.4.4 Catalysts  1.4.4.1 Solid Acid Catalysts ?????????.???.........23 1.4.4.2 Zeolites  ?????????????????.........24 1.4.4.3 Catalyst Deactivation ???????????............25 1.5 Objectives/Research Questions and Objectives ?????????..........27  v   CHAPTER 2   2.1 Introduction ?????????????????????.?............30 2.2 Experimental Methods  2.2.1 Isomerization ??? ??????????????............30 2.2.2 Esterification ?????????????????.............32 2.2.3 Transesterification ??????????.?????............32 2.2.4 Analytical Techniques  2.2.4.1 Cloud Point Analysis  ??????????................33 2.2.4.2 FTIR Quantification  ???????????.............34 2.2.4.3 GC/MS Method  ?????????????............35 2.2.4.4 Carbon Deposition and Surface Area  ?????...........35 2.3 Results and Discussion  2.3.1 FTIR Analysis ?????????????????...........36 2.3.2 Methyl Oleate Study  ??????????????.............37 2.3.3 Oleic Acid Study  ????????????????..........38 2.3.4 GC/MS Analysis  ????????????????...........40 2.3.5 Melting Point and Mechanism  ???????????..........45 2.3.6 Additional Isomerization/Esterification Reactions  ???...........46 2.3.7 Kinetic Study  ?????????????????............48 2.3.8 Catalyst Reusability  ???????????????..........53 2.4 Conclusions  ??????????????????????............56  CHAPTER 3   3.1 Introduction  ??????????????????????............57 3.2 Experimental Methods  3.2.1 Experimental Design  ??????????????............57 3.2.2 Hydroisomerization  ???????????????..........58 3.2.3 Esterification ?????????????????.............58 3.2.4 Catalyst Preparation ??????????????..............59 3.2.5 Analytical Techniques  3.2.5.1 Cloud Point Analysis  ???????????............60 3.2.5.2 Gas Chromatography/Mass Spectrometry (GC/MS)  ?....60 3.2.5.3 Fourier Transform Infrared Spectroscopy (FTIR)  ??....60 3.2.5.4 Carbon Deposition, Surface Area and CO Adsorption ......60 3.3 Results and Discussion  3.3.1 Catalyst Characterization ?????????????..........61 3.3.2 MP and PA Results ???????????????...........61  vi  3.3.3 Kinetic Study ???? ?????????????............68 3.3.4 FTIR Results ??? ???????????????.........71 3.3.5 Reactions Performed of Vegetable Oils  ???????...........73 3.3.6 Catalyst Recycle  ????????????????...........75 3.3.7 Reaction By-Products  ??????????????...........78 3.4 Conclusions  ??????????????????????............79  CHAPTER 4    4.1 Introduction  ??????????????????????............80 4.2 Experimental Methods  4.2.1 Materials  ???????????????????...........83 4.2.2 Catalyst Preparation ???????????????..........83 4.2.3 Isomerization  ?????????????????............83 4.2.4 Hydroisomerization  ???????????????..........84 4.2.5 Centrifugation ?????????????????...........84 4.2.6 Esterification ?????????????????.............84 4.2.7 Simulation Calculations ?????????????............85 4.2.8 Experimental Error ???????? ???????...........85 4.2.9 Product Analysis  4.2.9.1 FTIR ?????????????????..............86 4.2.9.2 GC/MS ????????????????..............86 4.2.9.3 Cloud Point  ???????????????...........86 4.3 Results and Discussion  ??????????????????...........86 4.4 Cloud Point Model Based on Fatty Acid Content  ???????..............96 4.5 Conclusions  ??????????????????????............98  CHAPTER 5   5.1 Introduction  ??????????????????????............99 5.2 Experimental Methods  5.2.1 Catalyst Preparation ???????????????........100 5.2.2 Hydrolysis ??????????????????...........100 5.2.3 Isomerization  ?????????????????..........101 5.2.4 Hydroisomerization  ??????????????............101 5.2.5 Esterification ?????????????????...........101 5.2.6 Analytical Techniques  5.2.6.1 GC/MS ???????????????................102 5.2.6.2 Cloud and Pour Point  ??????????..............102 5.2.6.3 Viscosity Measurements  ???????.....................102  vii  5.2.6.4 Acid Number  ?????????????..............102 5.2.6.5 Oligomer Determination  ??????????.........103 5.2.7 HYSYS Set-Up Details  .?????????????.........104    5.3 Results and Discussion  5.3.1 Cloud Point Analysis  ??????????????..........104   5.3.2 Reaction By-Products  ?????????????.............109 5.3.3 Optimization of Low Unsaturated Fatty Acid Oils (Coconut    /Palm Kernel) ????????????????..............110 5.3.4 Energy Use  ??????????????????.........111 5.3.5 Economic Analysis  ?????????????...?..........114 5.4 Conclusions  ??????????????????????..........120  CHAPTER 6   6.1 General Findings ????????????????????...........122 6.2 Strengths and Weaknesses of Research  ???????????............123 6.3 Potential Applications ?????????????????...............123 6.4 Future Work ??????????????????????..........124  REFERENCES ??????????????????????????.............125  APENDICIES ???????????????????????????...........134           viii  List of Tables  Table 1.1 SFA, UFA and BCFA effects on biodiesel quality ????????.................11 Table 1.2 Various methods of cloud point improvement ????????????.......13 Table 1.3 Description of selected solid acid catalysts  ??????...................................24 Table 1.4 Zeolite characteristics .??????????? ?????????...........25 Table 2.1 Experimental design for testing different isomerization reaction  conditions?............................................................................................................31 Table 2.2 Conversion to methyl oleate of isomerized oleic acid samples after   esterification with sulphuric acid in methanol ........................................................33 Table 2.3 FTIR results of methyl oleate and oleic acid isomerization ?????.?..........39 Table 2.4 GC/MS analysis of isomerized methyl oleate products  ??.??????........43 Table 2.5 GC/MS analysis of isomerized esterified oleic acid products ???.??...........44 Table 2.6 Additional isomerization/esterification reactions performed using beta   zeolite catalyst  ........................................................................................................47 Table 2.7 Catalyst properties of original and recycled catalyst ?????????.?.....54 Table 3.1 Acid values of palmitic acid hydroisomerization reaction products after   undergoing esterification ........................................................................................59 Table 3.2 FTIR results for the hydroisomerization of the C16 species MP and PA ?...........72  Table 3.3 Vegetable oil hydroisomerization reactions performed using Pt zeolite  catalysts ??????????????????????????..........74  Table 3.4 Physical properties of isomerization and hydroisomerization catalysts before  and after the reactions ?????????????????? ??.?........76 Table 4.1 GC/MS analysis of the methyl esters of the isomerization/hydroisomerization   reactions on oleic acid (OA) and palmitic acid (PA) ????............??...........87 Table 4.2 FTIR wavelengths of isomerized and hydroisomerized oleic and palmitic acids at various group frequencies ?????????????????......95 Table 4.3 Experimentally determined cloud points of different fatty acid content                   biodiesels  ??????????????????????????.......97  ix  Table 5.1 Acid number results for hydrolysis and esterification reactions ???????103 Table 5.2 GC/MS compositional analysis of reactant and product stream from                  isomerization and hydroisomerization reactions   ??????????.........106 Table 5.3 Effect of isomerization/hydroisomerization reactions on fuel quality  ????..108 Table 5.4 Comparison of I/HI improvement against hydroisomerization       improvement ??????????????????????????110 Table 5.5 Energy use comparison for the hydroisomerization improvement  biodiesel production process versus the standard biodiesel production ...............113 Table 5.6 Composition data for process flow diagram of standard production        process  ?????????????????????????............115 Table 5.7 Composition data for process flow diagram of improvement production  process ??????????????????????????..........117 Table 5.8 Economic comparison for the hydroisomerization improvement  biodiesel production process versus the standard biodiesel production  ?..........119              x  List of Figures  Figure 1.1 Atmospheric CO2 increases from 100 AD to present  ??????.?......... ........2  Figure 1.2 Temperature and sea level increase from 1840 to present  .....................................2  Figure 1.3 Current land usage and potential for extra cropland around the  world  ????????????????????????......................4  Figure 1.4 Current and potential land available for cropland in Canada (Total  current and potential land is 67.2 million ha)  ????? ????....................5  Figure 1.5 Amount of land needed to produce one tonne of oil from various        sources  ????????????????????????....................6  Figure 1.6 (A) Representative structures of biodiesel molecule; (B) UFA oleic         Acid; (C) SFA palmitic acid; and (D) representative structure of a         BCFA ????????????????????????......................8  Figure 1.7 Mechanism for the radical addition of hydrocarbons  ??????...................15  Figure 1.8 Electrophilic substitution of methylene group on an alkene  ???....................15  Figure 1.9 (A) Methyl branch formation through alkyl shift route; and (B)  methyl branch formation through carbocation ring formation  route. (molecules shown are representative isomers of the mixture of  isomers that are created)  ......................................................................................20  Figure 1.10 Schematic of research plan  ???????????????...?..............28  Figure 2.1 FTIR absorption spectra of isomerized methyl oleate  ??????..................36 Figure 2.2 GC analysis of reaction products of isomerized ester at reaction  condition 250?C, 0.1 MPa N2 pressure and 6 h reaction time  ...???..............41  Figure 2.3 Kinetic analysis of isomerization of OA studying the yield of  branched chain (BC) fatty acids versus reaction time  ????????.........52  Figure 2.4 Arrhenius plot of ln(k) vs inverse temperature for the determination        of activation energy  ..............................................................................................53   xi  Figure 2.5 Correlation of surface area and carbon deposits of beta zeolite catalyst ................55  Figure 3.1 TEM images of zeolites: (A) with platinum; and  (B) without platinum ?..........62        Figure 3.2 (A) Effect of temperature on cloud point; and (B) effect of  temperature on mass conversion for C16 species methyl palmitate (MP) and palmitic acid (PA) and C6 species hexanoic acid (HA) at  conditions of 4.0 MPa H2 pressure, 16 h reaction time and 5 wt% catalyst  ??????????????????????????.........64  Figure 3.3 (A) Effect pressure on cloud point; and (B) effect of pressure  on mass conversion for C16 species methyl palmitate (MP)  and palmitic acid (PA) and C6 species hexanoic acid (HA). Reaction  conditions: 285?C, 16 h reaction time and 5 wt% catalyst  ??..........................65  Figure 3.4 (A) Effect of time on cloud point; and (B) effect of time on  mass conversion for the hydroisomerization of palmitic acid (PA)  and hexanoic acid (HA) at reactions conditions of 285?C, 4.0 MPa and 5 wt% catalyst  ??????????????????????.......66  Figure 3.5 Kinetic analysis of hydroisomerization of OA studying the yield of  branched chain (BC) fatty acids versus reaction time  ????????.........70  Figure 3.6 Arrhenius plot of ln(k) vs. inverse temperature for the determination        of activation energy  ..............................................................................................71  Figure 3.7 FTIR plot for the peak range of 3050 to 2735 cm-1 for 3 samples,   (purple) unreacted PA, (red) PA reacted at conditions of 200?C   and 1.0 MPa H2 pressure, and (blue) PA reacted at conditions   of 285?C and 4.0 MPa H2 Pressure  ??????????????..............73  Figure 3.8 Correlation between catalyst carbon deposition and CO adsorption   of Pt-zeolite catalyst  ............................................................................................77  Figure 3.9 Correlation between catalyst carbon deposition and surface area   of Pt-zeolite catalyst  ............................................................................................78  Figure 4.1 Energy levels for the reaction of A ? B or C  ???????????..........81  Figure 4.2 Desired and undesired products from isomerization/hydroisomerization  of oleic acid (OA) based on cloud point  ????????????...............82  Figure 4.3 Hydroisomerization of palmitic acid (PA)  ???????????...............82    xii  Figure 4.4 (A) Effect of isomerization of OA on the cloud point of the  Ester; (B) Effect of hydroisomerization of OA on the cloud point of  the ester; and (C) Effect of hydroisomerization of PA on cloud point of  the ester   ????????????????????????...............89   Figure 4.5 Effect of increasing iso-fatty acid methyl esters (branched compounds)        on the cloud point of a fatty acid methyl ester mixture (Mixture starts        with saturated fatty acid methyl esters)  ???????????????....91  Figure 4.6 Effect of increasing cis-unsaturated fatty acid methyl esters on the        cloud point of fatty acid methyl ester mixture (Mixture starts        with saturated fatty acid methyl esters)  ???????????????....91  Figure 4.7 Synergistic effects of methyl branching on improving the  cloud point of a mixture of fatty acid methyl esters (arrows indicate  reference axis, area% for bar graph and cloud point for line graph)  ?...............93  Figure 4.8 Simulated stable configurations of: (A) methyl palmitate and    methyl palmitate; and (B) methyl palmitate and branched methyl    isopalmitate ???????????????????????..............94  Figure 4.9 Experimental cloud point vs. predicted cloud point  ????????.............98 Figure 5.1 Cloud point changes of the various vegetable oil biodiesels  before and after the isomerization and hydroisomerization reactions  (error bars represent standard deviation of triplicate tests) ?????............107  Figure 5.2 Process flow diagram for standard biodiesel production process  ??..............115 Figure 5.3 Process flow diagram for cloud point improvement biodiesel         production process  ????????????????????.............116         xiii  List of Abbreviations  BC  Branched Chain BCFA  Branched Chain Fatty Acid BD  Biodiesel BET  Brunauer Emmett Teller CP  Cloud Point DFT  Density Functional Theory DSC  Differential Scanning Calorimetry  FA  Fatty Acid FAME  Fatty Acid Methyl Ester FTIR  Fourier Transform Infrared Spectroscopy GC/MS Gas Chromatograph/Mass Spectrometer GPS  Global Positioning System HA  Hexanoic acid HC  Hydrocarbon HI  Hydroizomerization HO  Hydroizomerized Oleic Acid HP  Hydroizomerized Palmitic Acid I  Isomerization IO  Isomerized Oleic Acid IP  Isomerized Palmitic Acid IR  Infrared LA  Linoleic Acid LAU  Lauric Acid  xiv  LC   Long Chain LCA  Life Cycle Assessment MC  Medium Chain MO  Methyl Oleate MP  Methyl Palmitate OA  Oleic Acid OAME Oleic Acid Methyl Ester PA  Palmitic Acid PAME  Palmitic Acid Methyl Ester SAME  Stearic Acid Methyl Ester SFA  Saturated Fatty Acid SC  Short Chain TEM  Transmitting Electron Microscope TGA  Thermogravametric Analyzer UFA  Unsaturated Fatty Acid           xv  Nomenclature  (?1??????)   Observed rate of reaction (mol/s*cm3)  ?    Order of reaction ?T    Difference in temperature between reaction and initial temperature (?C) ?G  Change in Gibbs free energy (J) ?    Void fraction ?   Effectiveness factor ?    Particle density (g/cm3) ?    Tortuosity A  Group contribution to Gibbs free energy As    Surface area of vessel (m) B(T)  Group contribution to Gibbs free energy based on temperature Ca    Concentration of species a (mol L-1) Cs    Reactant concentration at catalyst surface (g/cm3) Cv    Specific heat capacity of sample (J/g?C) Cv?   Specific heat capacity of Safire standard (J/g?C) Deff    Diffusion (cm2/s) E   Activation energy (J/mol) F   Frequency factor (mol L-1)1- ? h-1 HP    Horsepower (W) Keq  Equilibrium constant  k     Ratio of specific heats  k(T)    Reaction rate constant (mol L-1)1- ? h-1 M    Molecular weight (g/mol) m    Mass of sample (g) m?   Mass of Safire (g) N     Number of compression stages P1    Initial pressure (kPa)  xvi  P2    Pressure required (kPa) Pf     Pressure drop between inlet and outlet (kPa) Q    Energy (J) R   Universal gas constant (8.314 J/g*K) -ra    Rate of reaction of species a (mol L-1) h-1 Ri    Insulation thermal resistance (m2K/Wm) Rp     Radius of catalyst particle (cm) Sa    Surface area (cm2/g) T  Temperature (K) V    Volume (m3) W  Symbol to represent a reactant species in the general reaction term X  Symbol to represent a reactant species in the general reaction term Y  Symbol to represent a product species in the general reaction y    Distance to baseline of sample (mm) y?   Distance to baseline of Safire (mm) Z    Symbol to represent a product species in the general reaction term            xvii  Acknowledgements   I wish to truly thank my advisor, Dr. Naoko Ellis for her leadership, wisdom and guidance throughout this thesis. Without her advice, support and selfless contributions this study would not have been able to be conducted or completed. I would like to extend my gratitude to Dr. Kevin Smith and Dr. Glenn Sammis for their valuable suggestions as my committee members. Thanks to all my close friends who have helped me along in my studies, including special thanks to Dr. Jidon Janaun for his invaluable advice and assistance, Amir Dehkhoda, Soojin Lee, Jordon Cheung, Saad Dara and Joylenne Yu for their help in the lab, together all their help throughout this thesis was instrumental. I would like to thank Dr. Farnaz Sootodeh for her valuable assistance in molecular simulations. Lastly I would like to thank NSERC, and the Chemical and Biological Engineering Department at UBC for the financial support throughout this study.             xviii   Dedication  I would like to dedicate this work to my loving and supportive wife Ashley Reaume.                1  Chapter 1  Introduction  1.1 Global Warming and CO2 Rise  There are many problems affecting civilization due to the overuse of fossil-derived fuels. One problem is the severe increase in atmospheric CO2 levels. In 2009 alone, 84,000 MT of carbon in the form of CO2 were released from fossil fuel burning and related uses (Friedlingstein et al., 2010). The 2010 atmospheric CO2 level was 385 ppm, up from 300 ppm in the year 1900 (Lammertsma et al., 2011), i.e., a 28% increase over the 1900 value. Figure 1.1 shows the increase in atmospheric CO2 levels from 1,000 A.D. to the present. The world?s oceans have absorbed most of the CO2, allowing for only a slight rise (Friedlingstein et al., 2010). However, emissions need to be slowed before the oceans reach their capacity for CO2 absorption. A second issue facing modern society is the depletion of reserves of a resource that society has become extremely dependent upon. Fossil fuels are a far more valuable resource than just fuels. They are used to make plastics and materials, pharmaceuticals, roads, coatings and various other items that are essential to modern society.  Due to this fact, fossil fuels are far too integrated into society to outright replace. However, partial replacement is a must in order to combat the effects of CO2 increases, in addition to diminishing conventional oil supplies. While unconventional supplies have shown growth, these reserves are harder to extract and much more carbon intensive (Crude oil and commodity prices, 2011). The negative effects of increased CO2 include rising global temperatures, increased sea levels and severe weather, i.e., floods, hurricanes, drought and tornados (Pachauri and Reisinger, 2007).  The significant rise in CO2 emissions started at the beginning of the industrial revolution in the 18th century. Two effects that are parallel to the significant rise in atmospheric CO2, Figure 1.2 shows the increases in average global temperature and sea levels over the past 150 years. Alternatively, with increases in average temperature there is more energy available for intense storm activity, leading to increased severe weather. By contrast higher temperatures prevent water vapour in the air to form condensing forming light precipitation, leading to drought (Pachauri and Reisinger, 2007).    2   Figure 1.1 Atmospheric CO2 increases from 100 AD to present (adapted from Revelle and Suess, 2010; Tamino word press, 2012)  Figure 1.2 Temperature and sea level increase from 1840 to present (adapted from Pachauri and Reisinger, 2007)   3  The environmental issues are compounded by the limited supply of conventional fossil fuels. The age of cheap, easily accessible oil is in the past. Reserves are becoming more remote and require higher energy input in order to extract (Crude oil and commodity prices, 2011). This causes increased releases of CO2 into the atmosphere and further accelerates global warming. Biofuels such as ethanol, biodiesel and pyrolysis oil can help alleviate increased CO2 additions to the atmosphere (Beer et al., 2002). Fuels such as biodiesel help alleviate emissions due to the fact that the biomass material used to produce biodiesel is made from plant material. Biodiesel is burned releasing the CO2 to the atmosphere. Next the plants uptake CO2 from the air and produce the raw material needed for biodiesel, which is then converted to fuel. While this happens over the course of years or decades, it is a CO2 cycle instead of a drill and release addition, as in the case of fossil fuels (Tyson, 2001). The goal of increasing biofuels production is to partially replace petroleum fuel with the more environmentally friendly CO2 cycle biofuel. This will reduce net CO2 additions to the environment. Various life cycle assessment (LCA) studies have been performed on biodiesel that show a clear reduction in net CO2 emissions to the environment. For example, when canola, palm, waste and soybean oils are used for biodiesel production, LCAs show a net CO2 reduction of 57-78% (Kaltschmitt et al., 1997; Beer et al., 2002), taking into account all fossil fuels used in production. However, biofuels also have negative aspects. These include, but are not limited to: use of food products for fuel; higher cost and lower availability of feedstocks; fuel quality issues; and requires fossil fuel input for production. The vast majority of biofuels production to date is produced from food based feedstocks. This has caused food supplies to drop and prices to rise (Collins, 2008). As of 2008 the U.S. cultivates approximately 40% of the world?s corn crop; of which, 23% is used for ethanol production while 19% is exported. This is a reduction from the 23% that was used for export in 1987/88 before major ethanol production (O'Brien, 2012). It is estimated that biofuel production was responsible for 10-15% of retail food price increase in the years 2007-2008 (Gecan et al., 2009), and 23-25% of the increase from 2007-2009 (Collins, 2008). While there are negatives to the use of biofuels, if current energy sources are not supplemented away from fossil fuels, the destruction to the environment could become irreversible, if not already.    4  1.2 Vegetable Oils and Biodiesel 1.2.1 Food Production  Biofuels have ethical and quality issues that have to be overcome before wider market acceptance. Poor fuel qualities in addition to using food products and cropland for fuel production are a few. The use of food products creates ethical issues for biofuels. Therefore, a balance must be reached between food, animal feed and fuel (Amonette, 2012). Crops such as soybean and canola can be used for oil production with the leftover material used as a high protein animal feed, hence, not removing the feedstock entirely from the food cycle. With soybean and canola crops being grown for fuel and feed purposes, extra land could be utilized for these crops, which would yield an increase in animal feed with minimal loss of vegetable oils. Figure 1.3 shows global unused cropland that could be devoted to this endeavour. Figure 1.4 gives Canadian values for potential cropland. Other methods of alleviating food versus fuel could be the use of inedible oils for biodiesel production. Oils such as castor, jatropha and waste oils and fats could be used for biodiesel production. With proper production balances between inedible oils, food, feed and fuel, biodiesel can gain wider market acceptance (Amonette, 2012).  Figure 1.3 Current land usage and potential for extra cropland around the world (adapted from USDA (AC21), 2012)  5   LUC ? land use changes Figure 1.4 Current and potential land available for cropland in Canada (Total current and potential land is 67.2 million ha) (adapted from StatsCan, 2012)    An additional method for the alleviation of food disruption by fuel production is advances in crop loadings and yields.  Better management of land allows for higher amounts of crops to be planted per land area, increasing farm production without increasing land use (StatsCan, 2012). This can be accomplished through fertilizer and pesticide management, and global positioning system (GPS) monitoring of crop areas. These methods pinpoint under producing areas of land to improve crop yields (USDA (AC21), 2012). This has the added benefit of reducing energy consumption. With a reduction in fertilizer wasted on high production areas, energy and money is saved through fewer applications and the need for less. Furthermore, crop yields in addition are improved through the use of biotechnology. Genetically engineered crops are designed to withstand drought, pests and have lower fertilizer needs than regular crops. There are potential drawbacks to the use of biotech crops. These range from allergic reactions created by splicing genes, in addition to cross pollination with other crops (McHughen and Smyth, 2012). Cross pollination can occur when the biotech crops are planted close to a regular crop field. One other major concern is the monopolization of the biotech crop industry, i.e., one company controls over 85% of the seeds in addition to controlling the resistant pesticides and herbicides brands that are needed for the crops (McHughen and Smyth, 2012). A source of oil feedstock that has much higher yields allowing  6  for better utilization of land use is algae. Studies have shown that algae has the potential to produce 250 times more oil than soybeans and up to 30 times more oil than palm based on land area (Hossain et al., 2008). The yields of oil from algae currently are between 8,200 and 34,000 L/ha/y (Scott et al., 2010). This significantly reduces the land needed to produce the same amount of oil. Figure 1.5 shows the amount of land need to grow one tonne/y of oil through various crops. However, algae is not without challenges, as it cannot be simply planted and harvested. It has to be grown in either photo bioreactors or ponds. These require energy intensive methods of harvesting, providing CO2 and nutrient addition. LCAs of current production show low net energy gains from producing oil from algae versus current terrestrial plants, i.e., canola production (Lardon et al., 2009).   Figure 1.5 Amount of land area required to produce one tonne of oil from various sources (adapted from Mittelbach and Remschmidt, 2004; Lardon et al., 2009)  1.2.2 Other Uses for Oils and Fatty Acids  The potential for fats and oils does not stop at fuel use. Oils and fatty acids, more specifically, have many uses in industry other than biodiesel. The three most common uses include (Ruston, 1952):   7  Food Products This can range from straight vegetable oils used in food, oils used as preservatives and different fatty acid emulsifying agents. Soaps and Detergents This includes the use of multiple oils with sodium or potassium hydroxide to create a multitude of soaps. Detergents on the other hand, are composed of anionic and non-ionic fatty acid sulphate or phosphate salts. Examples of these are sodium lauryl sulphate and ethyoxylated oleic acid. In addition, fatty acids are used in fabric softeners and water and mildew repellants for clothing. Cosmetics  There are fatty acid salts contained in shampoos, gels, lotions and moisturizers.  Other common uses for fatty acids are paints, varnishes, rubbers, plasticizers, candles, inks and coatings. Oils are made up of two types of fatty acids: saturated fatty acids (SFAs) and unsaturated fatty acids (UFAs). There is a third type of fatty acid that does not occur naturally in large quantities, which are branched chain fatty acids (BCFAs). Lubricants and cosmetics benefit from branched fatty acid precursors over traditional fatty acids (Heidbreder et al., 1999). The BCFAs have lower melting points combined with higher oxidative stabilities (Moser, 2009), which are the two very important factors in making quality lubricants. The low melting points ensure proper flow for lubricants with no crystal formation, and proper coverage with cosmetics. The high oxidative stabilities ensure that the product quality will last through storage. Structures of the different fatty acids (UFAs, SFAs and BCFAs) along with biodiesel are shown in Figure 1.6.  8   Figure 1.6 (A) Representative structure of biodiesel molecule, (B) UFA oleic acid, (C) SFA plamitic acid, and (D) representative structure of a BCFA  1.2.3 Biodiesel Fuel Quality  Biodiesel is not refined to the extent petroleum based diesel fuels are; therefore, there are problems associated with biodiesel fuel quality. Refining processes for diesel fuels include cracking reactions, which remove high molecular weight hydrocarbons lowering the cloud point. In addition to distillations and hydroprocessing reactions that remove contaminants, such as sulphur. These issues include: high cloud point, low fuel stability and contaminants (Knothe, 2008). Cloud point is defined as the temperature at which solid crystals first start to appear and can cause the fuel to become ?cloudy? (Mittelbach and Remschmidt, 2004). Fuel stability is a fuel?s ability to resist oxidation while being stored (Moser, 2009). Compounds such as hydroperoxides are formed in the oxidation of biodiesel. The mechanism for the formation of hydroperoxides are shown in Equations 1.1 to 1.4).   Initiation RH   ?   R?  +  H?            1.1 Propagation R?  +  O2   ?    ROO?          1.2 ROO?  RH   ?    ROOH(hydroperoxide)  +  R?      1.3 Termination ROO?   +   H?   ?   ROOH(hydroperoxide)       1.4  9   The hydroperoxide can polymerize to form oligomers, causing major fuel problems. The greater the degree of unsaturation the more susceptible the molecule is to oxidation (Knothe, 2007). Low stability fuels have shown in multiple tests to cause fouling and/or coking to fuel injectors in addition to piston ring fouling. To combat the problem of fuel oxidation, the use of antioxidants has been widely studied. Antioxidants commonly used in biodiesel are butylated hydroxytoluene (BHT) and butylated hydroxyanisole (BHA), which are phenolic type compounds that that release a hydrogen free radical through homolytic fusion. The free radical hydrogen terminates the oxidation of the FAME and the free radical BHT or BHA is stable due to the butyl groups. (Knothe, 2007). The hydrogen attaches to the radical species and leaves a stable molecule that does not contribute to further oxidation. Antioxidants are found naturally in vegetable oils as tocopherols. However, vegetable oil processing along with biodiesel production removes much of the antioxidants and substitutes are necessary to store the fuel for long periods of time (Knothe, 2007). Other methods for oxidation control are storing in cool, dark and/or low oxygen environments. This is because heat and light are the two main drivers of the oxidation of biodiesel. Heat and light provide energy to start the release of the hydrogen atom adjacent to the double bond, creating the radical and subsequent hydroperoxide (Moser, 2009).   Biodiesel can contain various contaminants within the fuel that cause it to fail specifications. Contaminants such as fatty acids, glycerol, water, methanol, metals and soaps can have detrimental effects on the entire fuel system (Schmidt and Van Gerpen, 1996). Methods used to combat the contaminants are mainly extra washing steps. However, extra washing steps reduce yields and create additional waste streams (Kiss et al., 2007). Ion exchange resins have proven to be successful at removing soap and metal ions present in the biodiesel. These resins replace the K+ ion with an H+ ion re-creating the fatty acid, which can be removed in distillation steps (Berrios and Skelton, 2008).   Biodiesel is produced through the transesterification of vegetable oils and fats. Alcohols, most commonly methanol, are mixed with an alkali base, usually KOH. This creates a methoxide ion, which attacks the glycerol fatty acid ester bond, creating three methyl esters and a glycerol. The methoxide is formed according to the following acid/base reaction, Equation 1.5. The ratio of methoxide to hydroxide can be predicted from the reaction constants  10  of the methanol and water dissociation reactions. The reaction constants are calculated from the pKa values of methanol and water, which are 15.5 and 15.7, respectively.  OH-   +   CH3OH   ?    CH3O-    +   H2O      1.5 ??(????????)??(?????)=  [CH3O? ][H2O][CH3OH][OH?]        1.6  Vegetable oils and fats contain two classes of fatty acids: saturated and unsaturated. The ratio, size and extent of unsaturation of the fatty acids affect fuel qualities in the biodiesel (Knothe, 2008). Unsaturated fatty acids (UFAs) cause low cloud points (<4?C), good flow properties (viscosity, filter plugging) and poor fuel stabilities (<3 h). Saturated fatty acids (SFAs) yield higher cloud points (>15?C), poor flow properties (viscosity, filter plugging) and high fuel stabilities (>40 h). If free fatty acids are present in the vegetable oil, they react with the KOH and form soaps. This can negatively affect fuel quality and reduce production yields. Several methods have been proposed to avoid soap formation. The use of acid pre-treatment (Kiss et al., 2006) and solid acid/base catalysts (Dimian et al., 2009) are two examples. In large scale productions repeated water washing is often used (Jacobson et al., 2008). Other methods of post-production treatment used in industry are: distillation, drying, acid esterification and cold filtering (Souza et al., 2009).   UFAs and SFAs provide both negative and positive qualities to the biodiesel mixture. There is a class of fatty acid not present in vegetable oils that has the positive qualities of both SFAs and UFAs. Branch chain fatty acids (BCFAs) are saturated fatty acids which yield esters with high fuel stability, low cloud points and good flow properties (viscosity, filter plugging). The difference with BCFAs is a methyl or ethyl side chain on the hydrocarbon (HC) chain of the fatty acid. This forces individual molecules apart and prevents organized stacking. This effect lowers the melting point and subsequently the cloud point of a mixture of fatty acid methyl esters (Cirey and Sundburg, 2007). While branched molecules have the positive effect of lowering the cloud point of a fuel, they negatively affect the cetane number, which is a measure of the ability of a fuel to combust in the engine. A study by Santana et al. (2006) shows that creating a methyl branch on an alkane can lower the cetane number by up to 30 points. This decrease in cetane number can be reduced by increasing the saturation of the  11  molecule. Knothe (2008) shows that as the degree of saturation increases from methyl linoleate to methyl oleate to methyl stearate, the cetane number increases by up to 63 points. This is significant as the benchmark for ASTM approved biodiesel is a cetane number of 51. The effect of different fatty acids on biodiesel qualities is shown in Table 1.1.  Table 1.1 SFA, UFA and BCFA effects on biodiesel quality Fatty Acid Cloud Point  Fuel Stability (Storability) Cetane (ignition quality) Flow Quality   LC-SFA (-) (+) (+) (-) MC-SFA (+) (+) (-) (+) UFA (+) (-) (-) (+) BC-SFA (+) (+) (-) (+) SFA ? saturated fatty acid  (+) ? yields biodiesel with a positive quality LC ? long chain    (-) ? yields biodiesel with a negative quality MC ? medium chain   BC ? branch chain UFA ? unsaturated fatty acid  1.3 Improvement Methods 1.3.1 Additives  Traditional fuel additives are either wax inhibitors or pour point depressors. Wax formation inhibitors work by preventing the formation of large wax crystals; however, many small crystals are formed. The smaller crystals tend not to clog filters and injectors as larger ones would (Smith et al., 2010). Pour point depressors work in much the same way in attaching to wax crystals and preventing them from growing. Additives are mostly copolymers including, Tween 80, dihydroxy fatty acids, palm based polyol, and polyester propolymer (Ming et al., 2004). 1.3.2 Winterization  Winterization is the process of removing the higher melting point compounds from the biodiesel mixture. This is usually accomplished through repeated heating and cooling cycles. When the biodiesel is cooled to where solids start to form, the liquid fraction is removed and heated up to room temperature, and the solid fraction is discarded. This process is repeated until the desired cloud point is achieved (Knothe, 2008). This method drastically reduces the yield of the final product with biodiesel losses of 30-100% being reported (Gonz?lez et al.,  12  2002). Another method of winterization is through the use of solvents, called solvent fractionation. This technique uses solvents such as hexane to separate high melting point fatty acid methyl esters (FAMEs). The biodiesel is mixed with hexane and cooled, the crystal particles then fall out at low temperatures and the liquid fraction is removed and solvent vaporized off.  This process decreases losses, however, has extra costs and energy requirements.  1.3.3 Branched Higher Alcohols  Another method for cloud point reduction is the use of higher alcohols (ethanol, isopropanol, propanol, butanol and tert-butanol) as reactants for the transesterification and esterification reactions. These alcohols are larger and create branched structures at the ester bond site, effectively increasing the fluidity of the molecule lowering the melting point of the ester molecule versus the use of methyl alcohol (Lee et al., 1995). These alcohols have been proven to reduce the cloud point of a biodiesel sample by up to 10?C over methanol (Lee et al., 1995; Foglia et al., 1997). The main drawback to this method is the increased cost of the higher alcohol. As the level of alcohol increases (ethanol, propanol, butanol), which increasingly lowers the cloud point, the cost of production increases. This raises the already relatively high cost of the biodiesel. Currently, methanol is used most frequently in commercialized processes due to its relatively high availability, low cost and higher reaction rates over higher alcohols (Wang et al., 2011). A second drawback to the use of higher alcohols is the higher reaction temperatures, stronger catalysts (i.e. potassium methoxide as opposed to potassium hydroxide) and high alcohol to oil ratios needed for high reaction conversions to the methyl ester (Lee et al., 1995; Foglia et al., 1997).  1.3.4 Epoxidation/Alkoxylation  Epoxidation is another method in which to create a branched ester. In epoxidation an oxirane ring is formed at the carbon carbon double bond site, accomplished through the reaction with H2O2 using formic acid as a catalyst (Smith et al., 2010). This allows for the ring to be opened and an alcohol to be attached. Sulphuric acid opens the ring and alcohol is attached to the hydrocarbon chain with an ether linkage. This creates a branched molecule; however, in addition to the branched molecule, a ketone and dihydroxide can also be created from the oxirane ring destruction. Due to these side products, this method does not  13  significantly lower the cloud point of the biodiesel. Studies show only a 1?C reduction in cloud point using a butyl alcohol (Smith et al., 2009). All these methods and potential drawbacks are summarized in Table 1.2.   Table 1.2 Various methods of cloud point improvement  Method of Improvement Process definition Drawbacks Source Mixing with petroleum diesel Involves mixing biodiesel with petroleum diesel in various ratios - Still dependant on fossil sources for fuel (Kalam and Masjuki, 2002) Fractionation Cycles of cooling and heating where solid fractions are removed in each cooling cycle - Removing fractions of the biodiesel significantly reduces the yield (Smith et al., 2010) Fatty acid modification Genetic modification of the oil seeds to increase unsaturated oils - Expensive  - Political issues in genetic modification of crops - Reduces the oxidative stability of the fuel (Knothe, 2008) Various alcohols Using longer chain or branched alcohols for the esterification instead of methanol - Expensive to acquire alcohols (Marchetti and Errazu, 2008) Fuel additives Use of pour and cloud point suppressors in the biodiesel itself - These can be toxic and add to the cost of the fuel (Smith et al., 2010) Epoxidation Adding of an alcohol through an ether bond to the double bond site of the fatty acid chain - Only works if there is a high unsaturated fraction - Can be expensive to obtain materials (Smith et al., 2010)  1.3.5 Addition/Substitution Reactions  Other methods of creating branched chain fatty acids (BCFA) are through the use of reactions involving chemical intermediates. Unsaturated fatty acids are in essence carboxylic acid alkenes with a carbon carbon pi bond that is electron rich and can be subjected to various addition and substitution reactions. The two main branching reactions investigated were  14  radical addition and electrophilic substitution (Biermann and Metzger, 2008). Addition reactions studied for biodiesel cloud point improvement are the addition of compounds to fatty acids through the use of radical intermediates. The radical compounds are created when a pi bond is subjected to high temperatures and pressures. The radical compound then bonds with a subsequent hydrocarbon to form a branched structure. These reactions, although successful, produce low yields when the pi bond is not on a terminal carbon (Cirey and Sundburg, 2007). Additionally there are problems with polymerization and oligomerization as the radical compound is not specific in targeting the hydrocarbon that is intended to be branched (Cirey and Sundburg, 2007). Electrophilic substitution in regards to biodiesel improvement is where an electrophile or ?electron liking? compound is substituted for a hydrogen ion on an electron rich area (i.e., pi bond) of the unsaturated fatty acid. An electrophile is usually a positive ion or slightly positive compound. Once the electrophile breaks the pi bond, a carbocation is created, at which point the negative ion or slightly negative compound removes the hydrogen ion, leaving the alkyl group and reforming the pi bond (Schreiner et al., 1993). For substitution of alkyl branches, alkyl halides are the most commonly used reactants. The catalyst AlCl3 is used to initiate the reaction. The mechanisms of both the addition and substitution reactions are shown in Figures 1.6 and 1.7, respectively. These show the difference between the two reactions and illustrate some of the problems associated with them. Figure 1.6 shows that once the radical is created, it can be non-specific in attacking a pi bond leading to polymerization. Methods to prevent polymerization are limiting the reactant species or the use of antioxidants Figure 1.7 shows the necessity of the aluminum chloride which will be very difficult to separate out, possibly contaminating the biodiesel.   15   Figure 1.7 Mechanism for the radical addition of hydrocarbons  Figure 1.8 Electrophilic substitution of methylene group on an alkene    16  1.3.6 Isomerization/Hydroisomerization As stated previously, branched chain fatty acids (BCFAs) improve biodiesel fuel qualities in addition to benefiting other industries that use fats and oils. BCFAs do not exist in nature in very high volumes, therefore, must be created through catalytic reactions. A promising method of BCFA production is catalytic isomerization (I), and hydroisomerization (HI). I/HI reactions use solid acid catalysts to create a methyl or ethyl side chain on the hydrocarbon tail of the biodiesel molecule. Studies have shown varying levels of improvement using isomerization reactions (Yori et al., 2006; Ngo et al., 2007; Ha et al., 2009). Isomerization needs UFAs with the carbon carbon double bond; whereas, the hydroisomerization reaction can use both SFAs and UFAs because it creates a carbon carbon double bond (Ono, 2003). A carbocation is formed at the double bond site and re-arranged into a methyl or ethyl side chain. This method has the potential to reduce the cloud point of a biodiesel mixture by reducing the melting points of individual fatty acid methyl esters. The improvement methods discussed up to now either have issues that are improbable to solve, or do not significantly improve the cloud point. In reference to the I/HI reactions, there are large gaps in the literature about its use with vegetable oils, saturated fatty acids and the full effect the reactions have on cloud point. While there are studies on UFAs and already improved biodiesel samples, these do little to definitively show if the overall cloud point of biodiesel can be improved. Further analysis on the current knowledge and knowledge gaps about the use of I/HI reactions for the improvement of biodiesel cloud point are summarized in the following.  1.4 Isomerization and Hydroisomerization Reactions 1.4.1 Background  A class of fatty acids not found in vegetable oils are branched chain fatty acids. These fatty acids have lower melting points than SFAs and higher stability than UFAs (Moser, 2009). BCFA esters added to biodiesel have the ability to lower the cloud point and raise the stability, yielding a higher value product with superior fuel qualities. Since vegetable oils do not naturally contain BCFAs, they have to be synthesized through catalytic reactions. Isomerization and hydroisomerization are two reactions that have shown success at creating branched compounds (Claude and Martens, 2000; Biermann and Metzger, 2008). The two reactions differ only by the extra hydrogenation/dehydrogenation step associated with the  17  hydroisomerization reaction (Ono, 2003). The main purpose of these reactions is to create an isomer with a methyl side chain on the hydrocarbon tail of the fatty acid. This lowers the melting point of the ester by decreasing the interactive dispersion and Van der Waals forces, forcing the molecules further apart, thus, increasing the fluidity of the mixture (Cirey and Sundburg, 2007). Ideal conditions as stated by Koivusalmi and Jakkula (2006), for the isomerization of fatty acids are temperatures of 220-300?C, pressures of 0.1-2 MPa, and with water co-catalyst of 1-3 wt%. The batch process reaction is ideally carried out for approximately 5-6 h with a zeolite catalyst loading of 1-10%. Ngo et al. (2007) studied the effects of different co-catalysts (water and methanol). The researchers performed the isomerization reaction on oleic acid (C18:1) at 250?C for 6 h and 2.5 wt% zeolite loading. A maximum conversion of 99% oleic acid to products was achieved in the presence of 2 wt% H2O based on the amount of oleic acid used. Under the conditions utilized, the desired product, methyl isostearate, was formed with a selectivity of 82%. While current literature data covers the final cloud point after isomerization and a winterized sample, what is lacking is a complete before and after analysis of the cloud point. Only this way will a conclusion be reached on whether isomerization is a suitable cloud point reduction method for biodiesel. Literature data covers the isomerization of UFAs fairly well; however, what is lacking from the data is the effect of SFAs on isomerization. The SFAs themselves are not suitable for isomerization as they do not contain a carbon carbon double bond; however, they are present in biodiesel feedstocks and their effect on the reaction must be looked at for a complete study on isomerization. Hydroisomerization has had wide success branching long chain hydrocarbons in the size range of C10-C20 (Claude and Martens, 2000; Park and Ihm, 2000; Claude et al., 2001); however, there are limited studies performed on the fatty acids (FAs) or fatty acid methyl esters (FAME) of the C10 to C20 size range. The problem when applying I/HI to the fatty acid or subsequent methyl ester is the effect of the COOH/COOCH3 group on the isomerization and hydroisomerization reactions (Ha et al., 2009). It has been shown that a yield of 6% isostearic acid is possible from hydroisomerization of stearic acid using reaction conditions of 0.4 wt% Pt doped zeolite catalyst, 340?C and 689 kPa H2 for 6 h (Kenneally and Connor, 2001). This  18  yield is low and will have little effect on the overall cloud point of biodiesel; therefore, a more structured study is needed to determine if greater yields are possible. Hydroisomerization of alkanes has shown higher conversions and yields of branched chain hydrocarbons in various studies. A study by Claude et al. (2001) shows that a 95% conversion of hydrocarbons is possible with an iso-species yield of 79% under the conditions of 233?C, H2:hydrocarbon ratio of 13:1 and total pressure of 450 kPa. A study by Modhera et al. (2009) showed that varying the reaction conditions for n-hexane hydroisomerization significantly affected the conversion and yield of the reaction. Hence, it is feasible to suggest that adjusting the reaction parameters might increase the yield of iso-species for the fatty acid/methyl ester hydroisomerization. With studies examining isomerization of UFAs and very few studying hydroisomerization of SFAs, there is a gap in the knowledge on the interaction the UFAs and SFAs have when reacted together in either isomerization or hydroisomerization.  Biodiesel production is a commercial process and adding an improvement reaction raises the question of how such an improvement will be integrated into the production process. The main feedstock of biodiesel is vegetable oils and fats which are made up of triglycerides. Improvement could start with the triglycerides or end with the fatty acid methyl esters (FAMEs). If these are the optimal reactants for the improvement reactions, then little is needed in terms of changing the production process, only to simply add in the improvement process either before or after transesterification. However, there is another potential reactant for the improvement process which is the fatty acids. If these turn out to be the optimal reactant for cloud point improvement, then a hydrolysis step is required to release them from the triglyceride.   1.4.2 Mechanisms Isomerization and hydroisomerization reactions share the same branching reaction, i.e., the creation of a methyl or ethyl side chain. The only difference between the two is the extra dehydrogenation/hydrogenation involved in the hydroisomerization reaction. The general mechanism for the hydroisomerization reaction is (Ono, 2003).  CnH2n+2     CnH2n  +  H2      (on Pt)        (1.7)  19  CnH2n  +  H+    CnH2n+1+    (on solid acid)      (1.8) CnH2n+1+  ?  isoCnH2n+1+    (on solid acid)      (1.9) isoCnH2n+1+    isoCnH2n   +  H+    (on solid acid)     (1.10) isoCnH2n  +H2    iso CnH2n+2   (on Pt)      (1.11) (For isomerization omit parts with Pt)  The mechanism describes the general isomerization/hydroisomerization reaction steps. The rate limiting step (Eq 1.3) is the carbocation rearrangement, which creates the branch in the hydrocarbon chain.  The rearrangement step is the rate limiting step due to the fact that it has to overcome the steric interaction in formation of an alkyl branch. The exact cause of the shift is under debate; however, two theories are suggested. First the cation shifts to a terminal carbon and undergoes an alkyl shift with a methyl group (Rigby and Frash, 1997). This is a high energy shift; however, the reactions are run at high temperatures, which allows for the reaction to take place. The second proposed mechanism is through a cycloproponium ion (Olah, 1972). Once the carbocation is formed, it rearranges to form a methyl bridge involving a carbon carbon bond delocalization. The methyl bridge forms into a methyl branch on the chain. The two processes are described in the Figure 1.8, which is a representation of one product that can form from the processes.    20   Figure 1.9 (A) Methyl branch formation through alkyl shift route; and (B) methyl branch formation through carbocation ring formation route. (molecules shown are representative isomers of the mixture of isomers that are created)   The effect of branching on the fatty acid methyl esters is significant. Methyl stearate, a straight chain saturated ester has a melting point of 37?C; whereas, the branched chain methyl  21  isostearate has a melting point of 26?C. Two oils used for biodiesel production are canola and palm oils, which generate biodiesels with cloud points of -1.0 and 16?C, respectively. The difference in cloud point is caused by the different UFA to SFA mass ratios which are 90/10 for canola and 50/50 for palm oil (Karmakar et al., 2010). This demonstrates that with improvements to the individual melting points of the fatty acid esters, the overall cloud point of biodiesel can be reduced (Knothe, 2008).  This effect is shown in the cloud points of biodiesels made from soybean (5.0?C), olive (7.8?C) and corn oil (10.1?C) (Reaume and Ellis, 2012). As the cloud point increases, the saturated fatty acid content of the oil/biodiesel increases as well. BCFAs created through isomerization have been shown to reduce the cloud point of biodiesel. A study by Yori et al. (2006) shows a 6?C reduction in cloud point of a winterized biodiesel sample, using sulphated zirconium solid acid catalyst; while, a study by Ngo et al. (2007) produces a product with a -13?C cloud point for an isomerized methyl oleate sample using a zeolite catalyst at 250?C and 6 h reaction time. Many studies show improvements to ester samples from isomerization using solid acid catalysts. Hydroisomerization has been shown to have minimal BCFA yields (<6%) using stearic acid and platinum doped zeolites at 350?C and 689 kPa hydrogen pressure (Kenneally and Connor, 2001). High yields using hydroisomerization have been shown on hydrocarbon species of various sizes (C5-C20) (Claude and Martens, 2000; Claude et al., 2001). A study by Claude and Martens (2000) shows branched chain yields of 98% using 0.5 wt% Pt doped beta zeolites at 233?C and a ratio of 13:1 H2:HC.  1.4.3 Kinetics and Equilibrium Both isomerization (I) and hydroisomerization (HI) reactions are equilibrium reactions. The two main reactions UFA ? BC-UFA and SFA ? BC-SFA are under different thermodynamic constraints due to the presence or absence of the carbon carbon double bond (Carey and Sundburg, 2007). The difference between the two reactions is the need for a Pt hydrogenation/dehydrogenation catalyst. This creates additional constraints on the hydroisomerization reaction due to the dehydrogenated compound existing for a fraction of a second in a hydrogen atmosphere (Zaera and Somorjai, 1984). However, hydrogen is important for the initiation of the dehydrogenation step (Sinfelt et al., 1962; Ono, 2003). A study by Biloen et al. (1977) shows that the preferred mechanism of dehydrogenation initiates with a  22  Pt-H complex versus the empty Pt surface. The short existence of the dehydrogenated compound requires the fatty acid to contact the Pt, then immediately contact the acid site to undergo isomerization. This requires higher reaction temperatures to increase fluidity and mass transfer, as well as longer reaction times to allow for more reaction attempts (Soualah et al., 2008). The equilibrium can be predicted from the Gibbs free energy of formation using the Van Krevelen Chermin equation (Eqs 1.6, 1.7) to estimate the ?G values (Wilde, 1990). The ?G is then applied to Eq 1.8, which calculates the equilibrium constant. From the constant the equilibrium concentrations can be calculated using Eq 1.9,0 based on the general reaction equation Eq 1.10.  ?G = ?Group contributions + ?Structural corrections + RTln?         (1.12) ?Ggroup = A + B(T)             (1.13) Keq = e (-?G/RT)             (1.14) Keq = [Product Y]y[Product Z]z/[Reactant W]w[Reactant X]x       (1.15) wW  +  xX  ?  yY  +  zZ        (1.16)   Different thermodynamic energy states play a key role in UFAs. The C=C double bond can occur in two different configurations: cis and trans. The cis configuration, methyl oleate is by far the most common in vegetable oils (>99%) (Karmakar et al., 2010). However, the trans configuration is the more thermodynamically stable configuration (Berstein, 1962; Garcia-Pino et al., 2006). While undergoing isomerization or hydroisomerization the H+ ion from the zeolite generates a carbocation on the hydrocarbon chain. This carbocation is highly mobile along the chain length. This is due to the high temperatures required for the reactions, which provides more than enough energy for the migration of the ion. This makes predicting the exact location of the branch impossible, as the cation is free to move along the chain. Additionally, if the cation moves close to the carboxylic acid group (C3-C5) the production of a stereolactone will occur. This is caused by the electron rich oxygen in the acid group attracting the carbocation and bonding with the carbocation.   23   Mass transfer will affect the reaction conversion which then affects the cloud point of the final products. There are two types of mass transfer: external and internal. External mass transfer is the ability of the reactants to adsorb onto the catalyst surface. It is defined as the diffusion of the reactant from the bulk fluid to the external surface of the catalyst. External mass transfer can be improved by using smaller catalyst particles, increasing temperature and fluid velocity. In a batch process increasing fluid velocity is accomplished by increasing stir rates. Internal mass transfer is the diffusion of the reactant from the external surface of the catalyst to the interior of the catalyst (Scatterfield, 1981). In heterogeneous catalysis it is quantified as the effectiveness factor; which is defined as the overall rate of the reaction over the intrinsic rate of reaction. Internal mass transfer can be improved by reducing catalyst pellet size and creating a less tortuous path within the catalyst particle (Soualah et al., 2008).  1.4.4 Catalysts 1.4.4.1 Solid Acid Catalysts  Solid acid catalysts are defined as solid support structures which contain acid sites. These catalysts do not homogenously mix with the reactants or dissolve into a solution. Instead the reactant is usually adsorbed onto the catalyst structure where the reaction takes place. The acid sites are made up of either Bronsted or Lewis acids (Tanabe, 1989). Bronsted acids are proton donors and contain H+ ions that catalyze a reaction. Lewis acids are electron pair acceptors which can bond with Lewis bases to form an adduct. Bronsted acids can change to Lewis acids and vice versa through dehydration/hydration reactions (Ha et al., 2009). Solid acids have many advantages over the use of homogeneous acids (eg., sulphuric and hydrochloric acids) the most notable of which being ease of separation after the reaction is complete. Homogeneous acids either require neutralization or washing to remove; whereas, solid acids if not built into the reactor itself (i.e., packed bed reactor) can be filtered out (Kiss et al., 2006). Examples of different solid acids are described below in Table 1.3 along with reactions commonly catalyzed.    24  Table 1.3 Descriptions of selected solid acid catalysts (adapted from Tanabe, 1989) Solid Acid Reaction Commonly Catalyzed Description Zeolites Cracking/dewaxing Framework of SiO4/AlO4, which leaves a -1 charge on the joining of the Si/Al open for a cation SAPOs (silicoaluminophosphate) Isomerizations Network of silica and alumina with phosphate attached creating acid sites Clays Oligomerizations Charged silica/alumina mixtures Ion exchange resins Hydration/pre-esterifications Resins with an exchangeable ion Metal oxides Hydro-treating Oxides of different metals that hold a negative charge allowing for a positive ion Heteropoly acids Hydrations A combination of metals linked by oxygen atoms containing a acidic hydrogen in the centre  Solid acids all have advantages in catalysis, and zeolites are the optimal choice for isomerization and hydroisomerization reactions involving fatty acids (Zhang et al., 2004). They are robust, commercially available and other catalytic sites can be easily attached for bi-functional use.  1.4.4.2 Zeolites A solid acid catalyst that is proven for the isomerization of fatty acids, biodiesel and alkenes is zeolites (Zhang et al., 2004). The structure of zeolites is a SiO4/AlO4 structure that occurs in different ratios. The combination of the silica/alumina yields an overall negative charge on the structure that is neutralized with a positive ion. When the embedded cation is an H+ ion, the zeolite becomes a Bronsted acid (H+) donor. There are a variety of zeolite structures with different pore spaces. Different zeolites with pores spaces, Si/Al ratios and characteristics are shown below in Table 1.4. As with all catalysts, there are many factors that affect catalyst activity. When it comes to solid acid catalysis, surface area, total acidity, pore size, pore volume and reusability are all key factors to consider. This is where zeolites have an advantage over other solid acid catalysts. Zeolites have high surface areas, acidity, and specifically designed pore spaces (Park and Ihm, 2000).   25  Table 1.4 Zeolite characteristics (adapted from L?nyi and Valyon, 2001; Zhang et al., 2004) Zeolite Pore Size (?)  Acidity (mmol NH3/g)  Surface Area (m2/g) A 3 Not reported 480 Beta 5.5 to 7.6 0.54 680 Y 7.4 0.42 750 ZSM 5.1 to 5.6 0.9 425 Mordendite 6.5 to 7.0 2.08 500   Solid acids catalyze many other reactions including biodiesel esterification, etherification and catalytic cracking (Janaun and Ellis, 2010). Biodiesel esterification can be advantageous to I/HI reactions because it is the reaction that creates biodiesel from fatty acids. Solid acids have been shown to produce biodiesel esters from both vegetable oils and fatty acids (Kiss et al., 2006; Kiss et al., 2007). The advantage to I/HI is that the same catalyst can be used for both the I/HI reactions and the biodiesel production.  Studies by Kiss et al. (2007) have shown improved ester yields with virtually no need for post-production processing, including washing, fatty acid and ion removal. If a simultaneous improvement/production process can be developed this would save material and energy costs.  1.4.4.3 Catalyst Deactivation  The deactivation of zeolite catalysts by coking is either caused by poisoning of the acid sites or pore blockages (Cerqueira et al., 2005). Poisoning of the catalyst is much less severe as the carbon material only blocks the active site, leaving the pores or channels open. Pore blocking coke causes more severe deactivation due to the fact that the whole pore or channel is blocked, which significantly reduces the surface area of the catalyst (Guisnet, 2002). Coke itself can consist of many different compounds, many of which are usually hydrocarbon based (Li et al., 2000). Coke formation can form through various different pathways. One pathway is through radical addition of hydrocarbon molecules. The high temperatures of the reactions cause hydrocarbons to form terminal radicals which attach and branch other hydrocarbons forming long chain, highly branched and cyclic hydrocarbons. A second pathway of coke formation is through acid/base catalysis. On acidic catalysts, coke at low temperatures is formed from olefinic compounds with electron rich pi bonds. The olefinic compounds undergo various reactions to form coke, through oligomerization, branching, hydrogen transfer, and  26  cycilization. These reactions create high molecular weight, branched, and cyclic molecules (Guisnet and Ribeiro, 2011). A study by Villegas et al. (2006) found that the reactant molecule influences the composition of the coke material deposited on the catalyst. Smaller alkanes tend to produce long chain alkenes and alkylaromatics. While longer chain alkanes tend to form highly branched and cyclic large chain compounds, e.g., C24H12 (Cerqueira et al., 2005). The solution for coke removal is normally thermal degradation. This consists of heating the catalyst under flow of four main gases (O2, CO2, H2O, and H2). The heating causes the carbon in the coke to form various compounds (Li et al., 2000). C  +  O2  ?  CO2          (1.17) C  +  CO2  ?  2CO          (1.18) C  +  H2O  ?  CO  +  H2         (1.19) C  +  2H2  ?  CH4          (1.20) Thermal removal of coke deposits does not restore catalyst activity to the same level prior to fouling (Li et al., 2000). This is caused by two main reasons: first, coke is deposited unevenly and is inconsistently formed; second, the high temperatures can cause sintering, irreversibly damaging the catalyst. The former reason stems from the fact that not all coke deposits are the same and require different temperatures and pressures to thermally remove. In addition some catalysts contain metal deposits which form coke through different mechanisms than the support material (Park and Ihm, 2000; Guisnet, 2002). The latter reason, sintering can irreversibly fuse the support material closing pores and channels permanently (Guisnet, 2002). Other causes of catalyst deactivation not related to coking are poisoning, thermal degradation (sintering), attrition, and loss of active phase. Poisoning of the catalyst can be caused by many different effects. A few of the causes are poisons blocking surface area, converting the catalyst species and physically or chemically degrading the catalyst. The deactivation by poisoning can be caused by chemisorption of an impurity on the active site, adsorption of a compound that blocks the active site, or modifying the active site irreversibly. Catalytic inhibitors are considered weak poisons as they do not usually cause permanent or severe damage to the catalyst. Poisons can also be categorized as reversible and irreversible. Reversible poisoning is usually caused by a weak bond and can normally be resolved through  27  removing the poison from the feed stream for an extended period of time (Forzatti and Lietti, 1999). Sintering of the catalyst is the physical change to the catalyst structure that results in a reduction of surface area. Sintering is caused by the production and coalescence of small metal crystallites into larger ones, thereby reducing the surface area to volume ratio. Attrition is the loss of catalyst material through the physical breakdown and/or breakup of the catalyst itself. This is common in fluidized beds, as the catalyst is contacting itself or the vessel wall at high rates of speed causing break up. Another cause of attrition is a high linear velocity through a thin support structure which can cause breakup and flow through of the catalyst. Lastly the loss of active phase can occur through volatilization or dissolution of the active phase into the feed stream (Forzatti and Lietti, 1999). 1.5 Research Questions and Objectives The three main research questions to be answered are: 1. Will isomerization through the use of isomerization and/or hydroisomerization reactions of biodiesel and/or its feedstocks improve the cold flow properties of the final product? 2. Can and if so where in the production process could the improvement process be integrated? 3. What is the extra energy and economic requirements for the cloud point improvement process?   The main objective of the work is to significantly improve the cold flow properties of biodiesel. The principle method for the improvement will be through isomerization, carried out utilizing both isomerization and hydroisomerization reactions. If the isomerization does not produce solid improvements, then other methods are further investigated. These methods are described in Sections 1.3.2 to 1.3.5. If, however, a successful method cannot be achieved, a review and testing of different additives will be examined to find the optimal improvement. If a solution can be found before additives are needed, it would be optimal as additives only prevent wax crystals from growing, not forming. This is satisfactory for fuel filters but can clog injectors and cause ignition problems. Once a reduction in cloud point is achieved the integration of the improvement process will be investigated to find the optimal improvement/biodiesel production process. This is relevant in that the improvement process could be reacted on the oil/fatty acids/esters, each with their own advantages and disadvantages.  28  This part will examine the best feedstock for the improvement and then apply necessary reaction steps for biodiesel production. The last step is to investigate the extra energy and costs associated with the improvement to the biodiesel. Since there is bound to be extra energy and costs, an engineering project would not be complete without an examination of the requirements. A process simulator will be used to assess the energy and material costs. These objectives will be investigated as described in the schematic plan in Figure 1.9.    Figure 1.10 Schematic of research plan  Testing a Variety of Reaction Conditions on Skeletal Isomerization Using Beta Zeolites No Literature Search on Fuel Additives for Cloud Point Improvement Testing of Various Catalysts for Skeletal Isomerization Investigate the Energy and Economic Demands of the Improvement Process Integrating Improvement into Biodiesel Production Testing of Other Methods of Cloud Point Improvement Improvement of Cloud Point  Yes No Yes Yes No Improvement of Cloud Point  Improvement of Cloud Point  Yes  29   The research plan simultaneously investigates the improvement of cloud point as well as the integration of the process. However, the biodiesel cloud point is the primary objective, therefore, will be the driving force that dictates the direction of the study. Once cloud point shows an improvement the integration will be studied followed by the energy and economic analyses to complete the study.                       30  Chapter 2 Isomerization of Unsaturated Fatty Acid Oleic Acid Using Solid Acid beta-Zeolite Catalyst for the Reduction in Cloud Point of the Methyl Ester  2.1 Introduction The objective of this study is to find the optimal reaction conditions for the isomerization of the model compounds methyl oleate and oleic acid to their branched chain isomers. In addition to optimizing reaction conditions, an investigation as to whether the isomerization step should occur on the acid or the ester will be performed. With limited studies done on optimizing reaction conditions for the zeolite catalyzed branch chain formation, a wide range of temperatures, pressures, amounts of co-catalysts and starting materials are examined. These conditions are evaluated and optimized based on cloud point and conversion to their respective branched chain isomers. Once these conditions are optimized for the model compounds, i.e., oleic acid as the unsaturated acid/ester and palmitic acid for the saturated acid/ester, future studies can be applied to mixtures generated from vegetable oils.  2.2 Experimental Methods 2.2.1 Isomerization The isomerization reactions were carried out in an Autoclave series mini-reactor purchased from Autoclave Engineers (Div. of Snap-Title Inc, Erie, PA) under the various reaction conditions described in Table 2.1. The mixture of methyl oleate (Acros Organics Fair Lawn, NJ), methanol (Acros Organics Fair Lawn, NJ) (if called for) and 4 wt% beta zeolite (Zeolyst International, Valley Forge, PA) were charged into the autoclave for the methyl oleate study.  For the oleic acid (Fisher Scientific, Fair Lawn, NJ) experiments water was used in place of methanol as the co-catalyst.  The beta zeolite catalyst used had a average particle size of 24 ?m, two different pore channels with pore channel sizes of 5.5 x 5.5 ? and 7.6 x 6.4 ?, a Si/Al molar ratio of 25 and surface area of 620 m2/g with a nominal cation ammonium. The catalyst was pre-calcined prior to experiments at 500?C for 3 h to change the ammonium ion to a hydrogen ion and create a Bronsted acid site. The catalyst was then stored at room  31  temperature until it was used for the experiments. The reaction mixture was mixed at a speed of approximately 600 rpm and purged with >99% N2 for 15 min to remove all traces of oxygen prior to heating. The temperature was ramped up to the desired temperature within 10 min while purging with N2.  The temperature, pressure and co-catalyst were set to the desired values as stated in Table 2.1.  The reaction was carried out for 5 h based on literature studies (Zhang et al., 2004; Zhang and Zhang, 2007); after which point, the reaction mixture was cooled to room temperature by cooling the reaction vessel using recirculating water, while the pressure was released before the sample was removed from the autoclave. The sample and catalyst were transferred to a 50 mL centrifuge tube. The separation of catalyst from the sample was carried out in an Allegra 25R Centrifuge at 12,000 rpm for 20 min at room temperature. The ester samples were then transferred to a 20 mL glass container and analyzed for cloud point, fourier transform infrared (FTIR) absorbance and gas chromatograh/ mass spectrometer (GC/MS) analysis, while the fatty acid samples were esterified according to the esterification conditions. After analysis the samples were stored at 4?C and shaded from direct light. Table 2.1 Experimental design for testing different isomerization reaction conditions Run # Temperature (?C) Pressure (N2) (MPa) H2O or Methanol (wt%) C1 23 0.0 0 C2 250 1.5 0 2.1 (triplicate) 200 0.1 0 2.2 225 0.1 2 2.3 (triplicate) 250 0.1 1 2.4 275 0.1 0 2.5 300 0.1 2 2.6 200 1.5 1 2.7 225 1.5 0 2.8 250 1.5 2 2.9 275 1.5 2 2.10 300 1.5 0 2.11 200 3.0 2 2.12 225 3.0 1 2.13 250 3.0 0 2.14 275 3.0 1 2.15 300 3.0 1 C1 ? control #1 C2 ? control #2   32  2.2.2 Esterification  Methanol and fatty acid were used at a molar ratio of 10:1. Sulphuric acid catalyst was mixed with methanol in a 100 mL glass reaction chamber until fully dissolved to achieve 2 wt% based on the amount of fatty acid. Next the fatty acid was added and the reaction mixture stirred at 300 rpm and heated under reflux for 1.5 h. Once the reaction was complete, the mixture was washed with distilled water 3 times to separate the ester from the methanol, acid and other impurities. Lastly, the ester was dried with a desiccant pack (calcium chloride) to remove any traces of water emulsified in the ester. Acid value of the ester was taken according to ASTM D974-08 method in order to ensure adequate conversion of fatty acid to fatty acid methyl esters (FAME). Table 2.2 shows that the esterification reaction achieved yields of no less than 99.91% and acid values between 0.049 and 0.185 mg KOH/g ester.  For the combined isomerisation/esterification reactions, 30 g oleic acid, 1.2 g or 4 wt% beta zeolite catalyst and 34 g of methanol, yielding a 10:1 alcohol to oil molar ratio, were charged into the reactor. Reaction conditions of 100-200?C, 1.2-4.0 MPa and 2-24 h reaction time were used. Once the reaction was complete, methanol was boiled off and the oil and catalyst mixture cooled to <60?C. The catalyst was removed by centrifugation at 12,000 rpm for 5 min. Acid number was then taken to examine the extent of reaction conversion and esterification. Lastly, the products were esterified by sulphuric acid and tested for cloud point, GC/MS and FTIR. 2.2.3 Transesterification  The reactor unit was charged with 30 g vegetable oil, 1.5 g of catalyst and 11.34 g methanol giving a catalyst weight of 4 wt% and a methanol to oil molar ratio of 8:1. Reaction conditions were set to 200?C and 4.0 MPa for 2, 4 and 6 h. Once the reaction was complete, methanol was flashed off and the products were cooled and centrifuged to remove the catalyst. The product was run under FTIR to examine conversion to ester.    33  Table 2.2 Conversion to methyl oleate of isomerized oleic acid samples after esterification with sulphuric acid in methanol Run # Acid Value  (mg KOH/g ester) Conversion to Methyl Oleate* (%) 2.1 0.156 99.93 2.2 0.128 99.94 2.3 0.096 99.95 2.4 0.185 99.91 2.5 0.146 99.93 2.6 0.163 99.92 2.7 0.056 99.97 2.8 0.142 99.93 2.9 0.187 99.91 2.10 0.151 99.93 2.11 0.123 99.94 2.12 0.151 99.93 2.13 0.049 99.97 2.14 0.135 99.93 2.15 0.131 99.94 * Remaining difference is oleic acid  2.2.4 Analytical Techniques 2.2.4.1 Cloud Point Analysis The cloud point analysis was carried out on the Cloud, Pour and Freeze Point Analyzer, model PSA-70X, acquired from Phase Technologies (Richmond, BC) with an accuracy of ?1?C.  At the start of a run a standard sample of known cloud point was tested to be within 1?C of its cloud point in order to ensure proper operation of the instrument. The sample cup was cleaned with heptane, and flushed with 150 ?L of sample twice. Next the sample of unknown cloud point was tested. A 150 ?L sample was placed inside the analyzer and the temperature was lowered at a rate of 1.5?C/min until crystals first appeared. The cloud point analyzer meets and complies with ASTM D5773 method for testing of cloud point.      34  2.2.4.2 FTIR Quantification A Varian 3100 Excalibur FTIR (fourier transform infrared spectroscopy) was used for the detection of bond frequencies, which makes it useful in studying whether a certain bond was created or removed from a compound. Two bonds were of importance in this study: the methyl group (CH3) in order to determine branching; and the methylene (CH2) group to see whether the hydrocarbon chain was shortened. A Varian 3100 FTIR was equipped with an Attenuated Total Reflectance (ATR). Before a sample was run, a background scan of the crystal was taken for a baseline to be subtracted and get the absorbance of the sample. The absorbance of each sample was taken at the wavelength ranges of 4500 to 650 cm-1. A 50 ?L sample, was placed on a horizontal ZnSe crystal at a resolution of 4 cm-1. The sample was scanned 64 times, subtracted from the background and the scans were co-added together to produce a chromatograph. The signal was analyzed using Varian Resolutions Pro software which employed baseline and ATR corrections to the chromatographs.    The compound methyl 16-methyl heptadecanoate (Sigma-aldrich, Oakville, ON) was used as a branched chain isomer for quantification. Standards were made in methyl oleate at concentrations of 75, 50, 25, 10 and 0%. Peaks of interest in the absorbance profile are 2955 and 2864 cm-1 which corresponds to the ?CH3 and ?CH2- groups, respectively (Doumenq et al., 1990; Ripoche and Guillard, 2001). The ratio of the 2864/2955 cm-1 peaks was calculated and plotted against the concentration of branched chain isomer. A standard curve was generated with an R2 value of 0.99 in order to calculate the concentration of branched chain fatty acid methyl ester in the unknown samples. The ratio of peaks was linear to the wt% of branch chain fatty acid methyl ester in the sample analyzed. This allows for a linear equation to be generated and fast accurate analysis of conversion to branch chain isomers.  The calibration of trans double bond formation was similar to the branched chain fatty acid except that the compound methyl elaidate (Sigma-Aldrich, Oakville, ON) was used as the standard. The absorbance of the trans double bond was at 968 cm-1 and a peak at 1744 cm-1 (C=O) was used to stabilize the reading by taking the ratio of the two peaks (Doumenq et al., 1990).    35   The heights of the peaks are utilized due to poor peak resolution between 3100 cm-1 to 2750 cm-1. Taking the heights simplifies the analysis and removes errors of area integration due to poor peak resolution (Doumenq et al., 1990).  The second safeguard against error was taking the ratio of the two peaks, i.e., 2864/2955 cm-1 for branch chain isomers, and 968/1744 cm-1 for trans isomers. This minimizes the error due to variations from scan to scan. 2.2.4.3 GC/MS Method GC/MS analysis and characterization was carried out on a Varian CP 3800 gas chromatograph (Varian, Mississauga, ON) equipped with a capillary WCOT fused silica 60 m x 0.32 mm wax 52 CP column with split ratio 10:1. A Varian 4000-8 mass spectrometer set to scan from 30 to 350 m/z at a rate of 1.5 scans/s. Helium was the carrier gas through the column at a rate of 2 mL/min. The injector temperature was set at 280?C and the column was initially set at 150?C for 0.5 min, ramped at 5?C/min to 200?C and held for 2 min then ramped to 250?C at 5?C/min and held for 5 min for a total time of 27 min. 2.2.4.4 Carbon Deposition and Surface Area    The amount of carbon deposited on the catalyst was determined through the use of thermal gravimetric analysis (TGA) on a TA instruments SDT 600. Catalyst samples of approximately 20 mg were degassed overnight at 120?C in a N2 flow of 10 mL/min prior to loading into an alumina sample cup and into the TGA furnace. The purging gas was air at a flow rate of 100 mL/min. The furnace was set to isothermal for 5 min at 30?C and then ramped up to 600?C at 20?C/min for a total time of 33.5 min. Weight loss of the material was taken as carbonaceous material that burned off in the O2 rich air environment.  Surface area analysis was taken with the use of single point BET analyzer. Catalyst samples were degassed overnight at 120?C in a N2 flow of 10 mL/min. A 5 mg sample of catalyst was loaded into the glass sample tube and set into the unit. The sample tube was then immersed in liquid N2 and an adsorption reading was taken. Liquid N2 was then removed and a subsequent desorption value was taken. The two values were compared to assess accuracy of the results and the surface area was calculated (Appendix A).    36  2.3 Results and Discussion 2.3.1 FTIR Analysis  Methyl oleate was taken as a baseline sample for absorbances and all samples after undergoing isomerization were compared for any change in absorbance peaks. The peaks of interest that are altered after the isomerization reactions take place are 2955 cm-1 (-CH3 group), 2864 cm-1 (-CH2- group) and 968 cm-1 (trans ?HC=CH- group), as shown in Figure 2.1. There was no other significant change to the absorbance spectrum of the samples after each run.  The FTIR method is used as a fast quantification of branched and trans fatty acid methyl esters, further backed up with a GC/MS analysis. The FTIR concentration results have ?5% accuracy based on the FTIR sensitivity. This provides a general trend for the reaction conversions that can be compared to the cloud point results.  Figure 2.1 FTIR absorption spectra of isomerized methyl oleate    37  2.3.2 Methyl Oleate Study  Results from both the methyl oleate and oleic acid analyses are summarized and compared in Table 2.3. The table shows the results of the FTIR quantification of branched chain (BC) isomers and trans configuration isomers. For clarity in analyzing the results the term ?branched? refers to an alkyl side chain on the unsaturated hydrocarbon chain of the fatty acid or methyl ester. Due to the overlap of these results (branched chain esters can be either cis or trans and subsequently trans isomers can be both branched and unbranched) simplification by grouping all branched chain isomers, cis and trans were included under  the BC isomers heading; thus, only unbranched trans isomers were included in the trans isomer column. In the process of compiling data, run numbers were assigned to include both the methyl oleate (MO) study and oleic acid (OA) study results, as the same conditions were studied for both. When referring to particular numbers in Section 2.3.2 results are under the heading ?MO study? and in Section 2.3.3 the results are under the ?OA study? heading. As shown in Table 2.3, there is an increase in cloud point as reaction temperature increases, with the pressure having little effect, except for the 300?C and 0.1 MPa condition having a sharp increase in cloud point. The cloud points of the different experimental conditions range from -15.2?C for 200?C and 3.0 MPa (Run 2.11) to 16.2?C for 300?C and 0.1 MPa (Run 2.5 in Table 2.3). These results are consistent with the study by Yori et al. (2006), where a low temperature of 150?C was the optimal reaction temperature for decreasing the cloud point of biodiesel, using a similar zeolite based catalyst and a temperature range of 125 to 275?C in nitrogen gas at atmospheric pressure.  Experimental reproducibility was tested by running triplicates of two experimental conditions chosen at random. The standard deviation of the results were calculated based on the variance of the BC yield data, as 3.2% and 2.1% for Runs 2.1 (200?C and 0.1 MPa) and 2.3 (250?C and 0.1 MPa), respectively. Therefore, the difference in results had to be >4% for the difference to be significant. The cloud point analysis was also repeated and the deviation is within the ?1?C accuracy of the instrument which confirmed the reproducibility of the results.   38  The rise in cloud point with increasing reaction temperature is due to the undesired isomerization reaction which changes the cis to a trans double bond configuration within the molecule. The melting point of a substance is an indicator of its cloud point, for example, methyl elaidate (trans double bond ester) has a much higher melting point of 10?C (Knothe and Dunn, 2009) than that of cis double bond ester methyl oleate of -20?C (Knothe, 2008). This trans/cis isomerization ratio increases as reaction temperature increases, and decreases with increasing reaction pressure. The methyl oleate (MO) runs in Table 2.3 show a substantial increase in cloud point from -8.1?C (Run 2.4) to 16.2?C (Run 2.5) which was due to the sharp increase in trans bond formation of 28% to 40%, respectively, and negligible increase in branch chain formation.  A comparison of the cloud point results for branch chain versus trans isomer isomerization show that trans isomers have a much greater negative effect on the cloud point than the potential positive effect of branching the hydrocarbon. This is illustrated with Runs 2.8 and 2.13, where there is a much greater branched chain product formed versus trans isomer, approximately 25% for branched chain and 5% for trans isomers; however, the cloud point was still increased from the control.  2.3.3 Oleic Acid Study  Isomerization reactions were carried out on oleic acid under the reaction conditions of 200-300?C and 0.1-3 MPa with 0-2 wt% water co-catalyst. After the isomerization reaction was completed, the mixture was centrifuged and esterified to create an ester as previously stated in Section 2.2.2. The oleic acid (OA) cloud point results, shown in Table 2.3 under the heading ?OA study?, increased with increasing reaction temperature, ranging from -15.1?C for Run 2.1, compared to 2.9?C for Run 2.9.  The increase in pressure from 0.1 to 1.5 MPa increased the cloud point in every case; whereas, an increase from 1.5 to 3.0 MPa only caused a rise the cloud point in two of the cases (Run 2.6 to Run 2.11 and Run 2.7 to 2.12). The rise in cloud point can be explained again by the unwanted side reaction of cis double bond configuration conversion to trans configuration.      39  Table 2.3 FTIR results of methyl oleate and oleic acid isomerization  Experimental Conditions Experimental Results MO Study OA Study (mass %) CP (mass %) CP Run # Temperature (?C) Pressure (MPa) H2O or Methanol (wt%) BC-isomer  trans-isomer  (?C)  BC-isomer  trans-isomer  (?C)  C1 23 0.0 0 0 0 -15.6 0 0 -15.6 C2 250 1.5 0 0 0 -15.7 0 0 -15.6 2.1 200 0.1 0 1 1 -15.2 4 12 -15.1 2.2 225 0.1 2 14 3 -12.9 20 39 -13.4 2.3 250 0.1 1 21 8 -11.6 24 52 -8.5 2.4 275 0.1 0 22 28 -8.1 36 51 -6.7 2.5 300 0.1 2 23 40 16.2 42 54 -1.3 2.6 200 1.5 1 3 0 -14.9 5 57 -8.1 2.7 225 1.5 0 19 3 -14.0 21 57 -8.3 2.8 250 1.5 2 25 8 -13.3 22 57 -0.1 2.9 275 1.5 2 31 16 -12.2 32 57 2.9 2.10 300 1.5 0 32 30 -11.7 50 54 0.3 2.11 200 3.0 2 6 0 -15.2 3 35 -7.4 2.12 225 3.0 1 12 0 -14.7 21 39 -4.9 2.13 250 3.0 0 23 5 -14.9 30 42 -6.2 2.14 275 3.0 1 27 13 -13.3 37 45 -0.7 2.15 300 3.0 1 32 23 -11.8 45 64 0.7 C1 ? Control #1; C2 ? Control #2; BC ? Branch Chain; MO ? Methyl Oleate; OA ? Oleic acid; CP ? Cloud Point   The OA branch chain yield results (Table 2.3) are very erratic, with no clear pattern that emerges. The values range from 1% for Run 2.1 to 50% yield for Run 2.10, which correspond to a study by Zhang and Zhang (2007) showing a branch chain yield range of approximately 8-40% using a variety of different catalysts.  Most of the results from this work show conversion to branch chain products to be in the 30-40% range. This is consistent with the results of Ha et al. (2009) and Tolvanen et al. (2007), who have reported yields of 32% and 30% branch chain isomers, respectively, using a zeolite based catalyst and oleic acid and linoleic acid as a starting material.   An emerging pattern is the correlation between the trans isomer conversion, branch chain conversion and cloud point, as summarized in Table 2.3. There are two cases where cloud point decreases with increasing temperature for the OA study. First case, by increasing the temperature from 275 to 300?C at 1.5 MPa (Runs 2.9 and 2.10), the cloud point of 2.9?C  40  decreases to 0.3?C, while the branch chain yield increases from 32 to 50% with a small change in trans isomer conversion. Therefore, the drastic increase in branch chain isomer could have a positive effect on the cloud point. A second case is the decrease in cloud point from -4.9?C (Run 2.12) to -6.2?C (Run 2.13), while the trans isomer yield is negligible and the branch chain yield increases from 21 to 40%, as shown in Table 2.3. The second case shows less increase in branch chain isomers to affect cloud point because the trans isomer yield is lower: 57% compared to 42% in cases one and two, respectively. These two results indicate that branching has the potential to decrease the cloud point by at least 2.5?C given the optimal conditions.   The co-catalyst variable is combined with the pressure variable in the results, this was done for simplification purposes. Additionally, literature studies have been vague about the effects that pressure and co-catalysts have on zeolites in isomerization reactions. Therefore, the variables pressure and co-catalyst are combined, then if there is a significant increase in BC isomer yield, the two variables can be separated and further studied.  It was shown in the results (Table 2.3) that the maximum difference between the pressure and co-catalyst variables with respect to the BC isomer yield is 9% and 8% for the MO and OA study, respectively. Whereas the maximum difference in the temperature values is 29% and 45% for the MO and OA study, respectively. This shows that temperature is the major variable in the reaction, and pressure and co-catalyst are not as important to the reaction. In addition to the low variability in the results of the co-catalyst, it was found in the optimization study (appendix A) that the optimal co-catalyst amount was 0%. For this reason the co-catalyst variable will not be further pursued.   2.3.4 GC/MS Analysis A GC/MS analysis was performed on the products of the reaction conditions 2.1-2.15 for both starting materials in order to identify the compounds and verify the FTIR results. The MS analysis of the results showed various compounds, some detected by the FTIR analysis and others that were not. Figure 2.2 shows a representative GC spectra obtained at reaction conditions 250?C and 0.1 MPa. GC/MS analysis results are summarized in Tables 2.4 and 2.5  41  for the methyl oleate and oleic acid starting material, respectively. The MS analysis has identified four different trans methyl octadecenoates. These products are formed from carbocation formation, migration and reforming as a trans double bond at the various carbons without branching (Cirey and Sundburg, 2007). The concentration of the various trans bond esters decreases as the double bond position is shifted farther from the original 9 position (trans methyl-9-ocadecenoate), with the concentration of position 9 > position 11 > position 8 > position 6. The trend is held constant, with the overall amount increasing with reaction temperature. The increase in temperature increases the degree of double bond migration with the high temperatures having the greatest amount of methyl 6-octadecenoate due to the higher reaction rates at higher temperatures.   1. cis methyl octadecenoate; 2. trans methyl  octadecenoate; 3. methyl isooctadecenoates; 4. methyl hydroxyoctadecenoate; 5. ?-stearaolactone; 6. Shorter chain methyl hexa, penta and tetradecenoates (cracked products) Figure 2.2 GC analysis of reaction products for the isomerized ester at reaction conditions 250?C, 0.1 MPa N2 and 6 h reaction time.  The overall conversion of methyl oleate increases with increasing temperature and is the highest with the median pressure value of 1.5 MPa. As conversion increases so does the cloud point of the sample in all but one of the cases. The oleic acid starting material reaction  42  at 275?C and 1.5 MPa (Run 2.9) has a conversion of 75% with a cloud point of 2.9?C; while, 300?C and 1.5 MPa (Run 2.10) reaction condition has a cloud point of 0.3?C with a conversion of 80%. This decrease in cloud point is most likely due to the significant rise in methyl isooctadecenoates from 25 to 50%, respectively. The methyl isooctadecanoates are grouped together in the results due to the inability of the MS to differentiate between the limited branched fatty acids in the library.  The slight discrepancy between the GC/MS and FTIR results can be attributed to two factors. First the FTIR detects the trans configuration of the carbon carbon double bond in both the branched and unbranched compounds; whereas, the GC/MS separates the compounds and measures them individually (i.e., headings (3) and (2.1-2.4) in Tables 2.4 and 2.5). Second reason is the effect of the CH3 group on the IR absorbance of the C=C bond. The peak used to quantify the trans double bond is the group R1HC=CHR2. The group frequency for the unbranched molecule (R1HC=CHR2) is in the range of 980-960 cm-1. When a methyl branch is added, this changes the group to R1R3C=CHR2, thus affecting the intensity and widening the peak (Alpert et al., 1970). However, overall results from the GC/MS analysis support the FTIR findings of bond rearrangement with respect to the reaction conditions. Three groups of products identified by the GC/MS analysis and not through FTIR analysis were methyl hydroxyoctadecenoates, stearolactones and other cracked products. These products were not picked up by the FTIR because they do not differ significantly in bond structure and exist in small concentrations of <3% for the hydroxyl esters and the stearolactones, and <5% for the cracked products. These compounds also only exist in samples run at high temperatures, i.e., >250?C. However, their effect on cloud point would be minimal, since C14-C16 esters and hydroxyl esters have a lower cloud point than that of the methyl octadecenoates (Knothe, 2008; Knothe and Dunn, 2009), in addition to their low concentrations.   43  Table 2.4 GC/MS analysis of isomerized methyl oleate products  Run # Temperature (?C) Pressure (MPa) Water (wt%) Conversion* (%) Area (wt%)^  Cloud Point (?C)  (3) (2.1) (2.2) (2.3) (2.4) (5) (6) (4) C1 23 0.0 0 0 0 0 0 0 0 0 0 0 -15.6 C2 250 1.5 0 3 0 0 0 0 0 0 2 0 -15.7 2.1 200 0.1 0 14 3 2 0 0 0 0 0 0 -14.6 2.2 225 0.1 2 15 8 3 0 0 0 0 0 0 -12.9 2.3 250 0.1 1 30 19 4 1 0 0 0 1 0 -11.6 2.4 275 0.1 0 35 11 8 2 2 1 2 1 1 -8.1 2.5 300 0.1 2 37 10 10 3 3 2 2 3 2 16.2 2.6 200 1.5 1 8 3 1 0 0 0 0 0 0 -14.9 2.7 225 1.5 0 20 9 4 2 0 0 0 0 0 -14.0 2.8 250 1.5 2 35 25 3 1 0 0 1 0 0 -13.3 2.9 275 1.5 2 42 23 4 3 0 0 2 2 0 -12.2 2.10 300 1.5 0 47 22 5 3 2 0 2 3 1 -11.7 2.11 200 3.0 2 9 6 1 0 0 0 0 0 0 -15.2 2.12 225 3.0 1 15 10 1 0 0 0 1 1 0 -14.7 2.13 250 3.0 0 24 18 3 2 0 0 1 1 1 -14.9 2.14 275 3.0 1 29 17 4 1 0 0 2 1 0 -13.3 2.15 300 3.0 1 33 22 4 2 1 0 1 2 1 -11.8 * Conversion is defined as (100-unreacted methyl oleate)*100%  ^ GC separated area curves using dodecane as internal standard  3 - methyl Isooctadecenoates 2.1 ? trans methyl 9-octadecenoate      2.2 ? trans methyl 11-octadecenoate   2.3 ? trans methyl 8-octadecenoate    2.4 ? trans methyl 6-octadecenoate  5 ? stearo-lactone 6 ? shorter chain methyl esters (cracked products)   4 ? methyl hydroxyoctadecenoate C1 ? control #1 no reaction pure product C2 - control #2 exposed to median reaction conditions for 5 h with no catalyst 44  Table 2.5 GC/MS analysis of isomerized esterified oleic acid Run # Temperature (?C) Pressure (MPa) Water (wt%) Conversion* (%) Area (wt%)^  Cloud Point (?C)  (3) (2.1) (2.2) (2.3) (2.4) (5) (6) (4) C1 23 0.0 0 0 0 0 0 0 0 0 0 0 -15.6 C2 250 1.5 0 3 0 0 0 0 0 0 2 0 -15.7 2.1 200 0.1 0 21 12 2 1 2 1 0 2 0 -15.1 2.2 225 0.1 2 32 19 5 4 3 2 0 3 0 -13.4 2.3 250 0.1 1 53 24 8 3 3 2 1 3 1 -8.5 2.4 275 0.1 0 57 34 7 4 3 2 0 2 1 -6.7 2.5 300 0.1 2 69 42 9 5 3 2 0 3 2 -1.3 2.6 200 1.5 1 32 14 5 1 1 0 1 2 0 -8.1 2.7 225 1.5 0 50 29 5 3 2 1 3 4 0 -8.3 2.8 250 1.5 2 67 32 10 5 2 3 3 3 2 -0.1 2.9 275 1.5 2 75 35 15 6 5 3 3 4 2 2.9 2.10 300 1.5 0 80 50 8 5 4 3 0 4 2 0.3 2.11 200 3.0 2 32 17 3 1 1 1 1 3 0 -7.4 2.12 225 3.0 1 63 25 9 4 2 2 1 3 1 -4.9 2.13 250 3.0 0 65 41 8 6 3 1 3 2 1 -6.2 2.14 275 3.0 1 70 41 10 6 3 3 2 2 2 -0.7 2.15 300 3.0 1 71 35 16 8 6 5 2 2 2 0.7 * Conversion is defined as (100-unreacted methyl oleate)*100%  ^ GC separated area curves using dodecane as internal standard  3 - methyl Isooctadecenoates 2.1 ? trans methyl 9-octadecenoate      2.2 ? trans methyl 11-octadecenoate   2.3 ? trans methyl 8-octadecenoate    2.4 ? trans methyl 6-octadecenoate  5 ? stearo-lactone 6 ? shorter chain methyl esters (cracked products)   4 ? methyl hydroxyoctadecenoate C1 ? control #1 no reaction pure product C2 - control #2 exposed to median reaction conditions for 5 h with no catalyst 45  2.3.5 Melting Point and Mechanism Melting points of a substance are accurate indicators of cloud points; therefore, melting point data from the literature is used where cloud point data is lacking. The melting point of methyl stearate is 38?C and the branched chain isomer methyl 16-methylheptadecanoate is 27?C (Knothe and Dunn, 2009); therefore, branching the chain provides advantages in lowering the melting point of a substance, and subsequently the cloud point.  However, the undesired side reaction of cis to trans double bond isomerization suppresses any reduction in cloud point caused by the branching of the molecule. For example, methyl oleate (cis configuration) has a melting point of -20?C and methyl elaidate (trans configuration) has a melting point of 10?C (Knothe and Dunn, 2009). Thus the potential 11?C reduction from branching is negated by the possible 30?C increase from the cis to trans double bond isomerization.   The results of the study all showed an increase in cloud point from the control values (-15.6 and -15.7?C), which is not the intended goal of the study. Though the goal was not met, an examination of the reaction products and research into the isomerization mechanism provides an understanding as to why the cloud points increased. The mechanism discussed earlier for the branching of unsaturated fatty acids as stated by Zhang et al. (2004), Biermann and Metzger (2008) and Martens and Jacobs (2008), converts the double bond to a carbocation and back to the double bond. This process of double bond reforming changes the configuration of the double bond from the cis to trans configuration. This is due to the fact that it is thermodynamically more stable to reform as a trans bond, due to the negative steric interactions of the cis configuration (Veldsink et al., 1997; Garcia-Pino et al., 2006).  The experimental results show that there are trans isomers formed without branches; however, there is no evidence of branching and retaining the cis configuration. Therefore, when the double bond is reformed it may or may not reform with a branch chain, depending on the reaction conditions. A trans bond ester without a methyl side chain is a negative by-product because of its high melting point and can be attributed to the rise in cloud point of the experimental runs. Therefore, future efforts will have to focus on finding the optimal conditions for increasing the branched isomers, while reducing the unbranched trans isomer by-products.   46  2.3.6 Additional Isomerization/Esterification Reactions  The results of the cloud point analysis of the isomerization of OA/MO showed no reduction in cloud point under any of the conditions studied. A factorial design (Appendix A) was undertaken to find any trend that could lead to a positive cloud point reduction. The design inferred that a possible reduction could be achieved at lower reaction temperatures. Therefore, additional reaction conditions were studied to test this hypothesis. The additional parameters studied were lower reaction temperatures, longer reaction times and increased catalyst loadings. Run in tandem with the additional reaction conditions, a study on the combined isomerization/esterification was tested. Literature studies by Kiss et al. (2007) and Diaman et al. (2006) show the effectiveness of solid acids on the esterification of fatty acids. This led to test whether a production using zeolites would mutually esterify and improve the cloud point of a fatty acid sample.  Isomerization reactions were applied to other reaction conditions and catalyst loadings. Results were similar to the modeled study, with lower temperatures yielding lower conversions and less cloud point change compared to higher temperatures. Additional zeolite catalyst loadings yielded higher cloud points with a negligible increase in yield of branched chain products. The major problem is the cis/trans isomerization, increasing the melting point of the molecule, which cannot recover with an increase in branched chain products. Results of the combined isomerization/esterification study show >98% esterification to FAME in all temperature conditions. However, the isomerisation results (Table 2.6) show only minor yields with the reactions conducted at 200 and 150?C, while, no yields were observed for the 100?C sample at 6 h. Reaction times shorter than 6 h show no isomerisation conversion. This is due to the high temperatures required to activate zeolite catalysts; however, increasing the temperature increases the methanol pressure exponentially, preventing higher temperatures from being reached, due to limitations in the reaction vessel.   The transesterification reactions show a very small yield of esters in the product stream, of up to 15%. This is not sufficient to warrant further study. The reason for the low yield is due to the high temperatures required for zeolites and the large size of the triglyceride. Zeolites  47  require temperatures greater than 250?C and have small pores that prevent oil triglycerides from entering and contacting the acid sites.  Table 2.6 Additional isomerization/esterification reactions performed using beta zeolite catalysts Reaction CP (?C) Conversion (mass%) Trans (mass%) BC (mass%) Isomerization of OA, 150?C, 0.1 MPa, 6hrs, 4wt% Catalyst (beta Zeolite) -15.5 0 0 0 Isomerization of OA, 150?C, 3.0 MPa, 6hrs, 4wt% Catalyst (beta Zeolite) -14.8 0 0 0 Isomerization of OA, 150?C, 1.0 MPa, 6hrs, 25wt% Catalyst (beta Zeolite) -15.2 0 0 0 Isomerization of OA, 200?C, 1.0 MPa, 6hrs, 4wt% Catalyst (beta Zeolite) -4.4 38 38 12 Isomerization of OA, 200?C, 1.0 MPa, 6hrs, 25wt% Catalyst (beta Zeolite) 7.0 24 24 8 Isomerization of OA, 150?C, 1.0 MPa, 24hrs, 4wt% Catalyst (beta Zeolite) -12.8 8 8 0 Isomerization of OA, 200?C, 1.0 MPa, 24hrs, 4wt% Catalyst (beta Zeolite) -0.1 52 52 24 Isomerization of OA, 250?C, 1.0 MPa, 24hrs, 25wt% Catalyst (beta Zeolite) 14.8 99 99 32 Isomerization of MO, 150?C, 0.1 MPa, 6hrs, 4wt% Catalyst (beta Zeolite) -15.8 0 0 0 Isomerization of MO, 150?C, 3.0 MPa, 6hrs, 4wt% Catalyst (beta Zeolite) -14.1 0 0 0 Isomerization of MO, 150?C, 1.0 MPa, 6hrs, 25wt% Catalyst (beta Zeolite) -14.9 4 4 0 Isomerization of MO, 150?C, 1.0 MPa, 24hrs, 4wt% Catalyst (beta Zeolite) -12.8 8 8 0 Isomerization of MO, 200?C, 1.0 MPa, 6hrs, 4wt% Catalyst (beta Zeolite) -14.2 5 5 0 Isomerization of MO, 200?C, 1.0 MPa, 6hrs, 25wt% Catalyst (beta Zeolite) -13.2 3 3 0 Isomerization of MO, 200?C, 1.0 MPa, 24hrs, 4wt% Catalyst (beta Zeolite) -8.6 15 15 7 Isomerization of MO, 250?C, 1.0 MPa, 24hrs, 25wt% Catalyst (beta Zeolite) 4.0 81 81 19 Transesterification of palm oil, 100?C, 1.0 MPa, 5hrs, 5 wt% catalyst (Zeolite), 7:1 methanol to oil ratio N/A 0 N/A N/A   48  Reaction CP (?C) Conversion (mass%) Trans (mass%) BC (mass%) Transesterification of palm oil, 150?C, 1.5 MPa, 5hrs, 5 wt% catalyst (Zeolite), 7:1 methanol to oil ratio N/A 10 N/A N/A Transesterification of palm oil, 200?C, 2.5 MPa, 5hrs, 5 wt% catalyst (Zeolite), 7:1 methanol to oil ratio N/A 15 N/A N/A Transesterification of palm oil, 150?C, 4.5 MPa, 24hrs, 5 wt% catalyst (Zeolite), 7:1 methanol to oil ratio N/A 15 N/A N/A Esterification, 100?C, 1.0 MPa, 2 hrs, 10:1 methanol: fatty acid (OA), 5 wt% catalyst (Zeolite) -18.2 83 0 0 Esterification, 150?C, 2.5 MPa, 2 hrs, 10:1 methanol: fatty acid (OA), 5 wt% catalyst (Zeolite) -17.0 100 0 0 Esterification, 200?C, 4.5 MPa, 2 hrs, 10:1 methanol: fatty acid (OA), 5 wt% catalyst (Zeolite) -12.8 100 6.2 6.6 Esterification, 150?C, 2.5 MPa, 2 hrs, 10:1 methanol: fatty acid (OA) ? Isomerization, 260?C, 1.0 MPa, 3hrs, 5 wt% catalyst (Zeolite) 8.4 100 11 17 CP = Cloud point Trans ? Yield of trans fatty acids/esters BC ? Yield of branched chain fatty acids/esters  2.3.7 Kinetic Study A kinetic study was performed on the isomerisation of OA with beta zeolite catalyst.  For the reaction rate equation, conditions of temperature of 260?C and pressure of 1.5 MPa H2 was chosen. Hydrogen was chosen over nitrogen as it has shown better conversion and fewer side reactions (i.e., cracking, oligomerization, etc.) These conditions were chosen as no optimal cloud point was reached within the range of variables studied; while, it appears that the optimal branched chain yield along with the smallest cloud point increase is between 250 and 275?C at 1.5 MPa pressure. For activation barrier calculations temperature conditions were varied between 200-300?C. Figure 2.3 show a low initial rate for branch chain production, reaching an apparent equilibrium of 29% yield. The theoretical equilibrium of the reaction cannot be accurately calculated, due to the exact compounds in the product stream not being identified. Additionally, there are side reactions occurring, including the cis/trans isomerization and carbocation migration, which makes identifying the exact spot of the final  49  carbocation and subsequent branch position difficult. This gives too many unknown variables to accurately calculate an equilibrium concentration (Wilde, 1990). The branch chain concentration leveled off at 6-10 h slowly decreasing up to 24 h. This could be the result of an increase in the cracking compounds or other side reactions destroying the branch chain compounds along with straight chain OA (Ngo et al., 2007).  The results of the kinetic study are analyzed to generate a rate equation for the isomerization of oleic acid to branched chain C18 fatty acids at the conditions stated using a non-linear regression of the conversion versus time data. The general rate equation (Eq 2.1) was used for the overall rate equation. ??? = ????          (2.1) Where: -ra = rate of reaction of species a (mol L-1) h-1 k = reaction rate constant (mol L-1)-? h-1 Ca = concentration of species a (mol L-1) ? = order of reaction After completion of the non-linear regression using least squares method, the rate constant (k) and reaction order are 0.116 (mol L-1)-0.92 h-1 and 1.92, respectively, giving an overall rate equation for the conversion of oleic acid (Eq 2.2). The confidence level in the results was >95%.  ??? = 0.116??1.92             (2.2)   The simple rate equation (Eq 2.1) was chosen for the analysis to obtain a general rate equation for the isomerization reaction. The reaction has to proceed through many steps in order to go to completion.  ? Adsorption onto the catalyst surface ? Protonation and carbocation formation ? Rearrangement of the carbocation into and alkyl branch ? De-protonation  ? Desorbing into the bulk fluid  50  The mechanism is more thoroughly described in Chapter 1. Each step in the reaction will have a separate rate, however, the rate limiting step is the rearrangement of the carbocation into an alkyl branch (Biermann and Metzger, 2008; Zhang et al., 2004; Ono, 2003). The rate limiting step will determine the overall rate of the reaction which can be modelled in the form of A ? Products. This allows for the use of the general reaction equation (Eq 2.1). Activation energy (E) is the energy barrier that a reaction has to overcome to form products. It can be calculated from the rate constant formula:  ?(?) = ??????           (2.3) k(T) = reaction rate constant  F = frequency factor E = activation energy (J/mol) R = universal gas constant (8.314 J/mol K) T = temperature (K)   Results from the temperatures studies of 200-300?C were used to calculate the different rate constants at their respective temperatures based on the rate equation (Eq 2.2). The rate constant equation (Eq 2.3) is rearranged to plot ln(k) vs (1/T) as shown in Figure 2.4. From this plot the slope of (-E/R) can be calculated to solve for the activation energy (E). The activation energy was calculated to be 48,729 J/mol. This value is relatively low as literature studies have shown values of 52,000-89,000 J/mol for isomerisation of shorter chain alkenes (Boronat et al., 1998; Matsuhashi et al., 1999). The literature studies were performed with different zeolite catalysts, which could explain a reduction in activation energy when using a beta zeolite. In addition, compounds studied were shorter chain alkenes as opposed to the longer chain unsaturated fatty acids used in this study.  The other factor involved in heterogeneous catalyst kinetics is the internal effectiveness factor (?), which is a measure of the overall reaction rate over the rate if the interior of the catalyst were exposed to the surface conditions (Eq 2.4). To estimate the effectiveness factor, the diffusion of the reactant into the pores of the catalyst must first be calculated using the diffusion equation (Eq 2.5) (Satterfield, 1981).    51  ? =  Overall rate of reactionRate of reaction if interior of the catalyst was exposed to surface condtions      (2.4) ???? =  ?2???????           (2.5) Deff = diffusion (cm2/s) ? = void fraction ? = tortuosity Sa = surface area (cm2/g) ? = particle density (g/cm3) T = temperature (K) M = molecular weight (g/mol) Once the diffusion is calculated, the modulus (?) can be calculated from Eq 2.6. With a modulus calculated the effectiveness factor can be estimated from empirical data from standard curves contained in Satterfield (1981). ? =  ?2????(?1??????)1??          (2.6) R = radius of catalyst particle (cm) (?1??????) = observed rate of reaction (mol/s*cm3)  Cs = reactant concentration at catalyst surface (g/cm3)  These values calculate to Deff = 7.63*10-4 cm2/s and a modulus of ? = 72.14 and from interpretation of empirical data that gives a ? of 0.09 (Sample calculation shown in Appendix B). A value of 0.09 is low and means that the observed rate is only 9% of intrinsic rate.  The initial reaction rate and activation energy were calculated with the reaction conversion from 0-6 h reaction time, based on the equation: A ? products. A non-linear regression of the conversion versus time data was used to calculate the rate equation based on the system {-ra=kCa?} for 260?C and 1.5 MPa H2 pressure. The reaction rate is relatively low, explaining a 27% conversion after 6 h of reaction time. The slow reaction rate could most likely be caused by two reasons: low internal mass transfer as indicated by an effectiveness  52  factor of 0.09; and a high activation barrier. Increasing temperature is one way to increase the rate constant and overcome the activation barrier. In addition, increasing temperature increases the diffusion of the reactant into the pore space of the catalyst. However, the reaction rate will increase at a faster rate than the diffusion. The reason is that reaction rate constant is exponentially proportional to temperature and diffusion is square root proportional to temperature. With increased temperature overcoming the activation barrier faster than diffusion improving internal mass transfer the effectiveness factor will likely decrease even further. This is due to increased reaction rates converting the reactant compounds on the catalyst surface before they can enter the pores. This fact leads to the conclusion that the reaction is under kinetic control. The cloud point of the samples increases significantly with time. This effect, however, is caused by the cis/trans isomerisation which overshadows the isomerisation reaction. The increases in cloud points are not consistent with increases in conversion to branch chain products or conversion of OA. This is because the fatty acid profile does not affect the cloud point in a linear manner (Reaume and Ellis, 2012).  Figure 2.3 Kinetic analysis of isomerization of OA studying the yield of branched chain (BC) fatty acids versus reaction time. 01020304050607080901000 5 10 15 20 25Yield of FA (%)Time (h)BC-FA(I) OA trans-FA other 53   Figure 2.4 Arrhenius plot of ln(k) vs inverse temperature for the determination of activation energy   2.3.8 Catalyst Reusability The catalysts used for the isomerization reactions were studied for their reusability. The beta zeolite was tested for surface area, carbon deposition, and reactivity. Results in Table 2.7 show a correlation between carbon deposition and surface area. Calcination was used to remove the carbon/coke deposits. After calcination in air, the used catalysts? surface area increased substantially (e.g 2.17 to 2.18in Table 2.7). This indicates that the carbon deposits are blocking pore spaces leading to a surface area decrease. The carbonaceous material that covers the catalyst is predominantly soluble in fatty acids. The uncalcined catalyst was immersed in a fatty acid solution and heated to 150?C for 15 min. The catalyst was then cooled and washed with acetone and analyzed for surface area, carbon deposition and reactivity. As a result, the fatty acid washed catalyst (2.19) performed much better than the unwashed catalyst, as shown in Table 2.7. All three tests show an improvement in catalyst qualities, however, not as high as the calcination. The fatty acid wash was performed to investigate whether the coke deposited on the catalyst dissolves off in the first minutes of the reaction. This explains why there is still product yield from the used, uncalcined catalyst. The ratio of soluble to insoluble y = -5861.1x + 8.8676R? = 0.9932-4-3.5-3-2.5-2-1.5-1-0.500.0015 0.0017 0.0019 0.0021 0.0023ln (k(T))1/T (K-1) 54  coke (i.e., the ratio of coke content from the reacted catalyst to coke content of the fatty acid washed catalyst) was 17:3 by weight, as shown in Table 2.7. This result is in agreement with a study by Villegas et al. (2006), which shows an insoluble coke content on a hydroisomerization Pt zeolite catalyst of 13-20%.   The catalyst was also tested using an alkene hydrocarbon (hexadecene) to see if the carboxylic acid had a major effect on the reusability of the catalyst. From the tests, the carboxylic acid did not adversely affect the coke formation or surface area coverage. Subsequently, the increased activity of the catalyst using hexadecene created slightly more coke material on the catalyst surface. The catalyst activity causing the coking is further confirmed through the 150?C reaction where there is almost no activity, and accordingly almost no fouling of the catalyst. Further testing was accomplished using an acetic acid wash to test whether the acid group washes away the coke deposited on the surface or if it has more to do with the hydrocarbon tail. As the tests confirm, with almost no improvement after the acetic acid wash, the hydrocarbon tail is the principle reason for the improved catalyst properties.   Table 2.7 Catalyst properties of original and recycled catalyst   Entry # Sample Yield of BC Ester (% w/w) Cloud point (?C) Surface area (m2/g) Carbon Deposition (% w/w catalyst) 2.16 Original 28 2.3 556 N/A 2.17 R1  22 -3.8 165 17 2.18 R1c 24 1.8 374 0 2.19 R1fa 22 -3.2 294 3 2.20 R1aa 18 -8.1 185 17 2.21 R1h 48 N/A 144 18 2.22 R1(at 150?C) 1 -16.2 402 1 2.23 R2 17 -4.2 126 18 2.24 R2c 20 -1.8 294 0 R1 ? catalyst used once before current reaction R2 ? catalyst used twice before current reaction BC ? Branched chain c ? calcined in air at 500?C for 2 h fa ? heated in a fatty acid bath at 150?C for 10 min aa ? heated in acetic acid bath at 80?C for 10 min h ? reaction run with hexadecane properties tested after set iteration, yield values correspond to reaction after iteration   55   Figure 2.5 shows the inverse correlation between the surface area and the amount of carbon deposits on the catalyst surface. It is clearly shown that with increasing carbon deposits, available surface area on the catalyst is severely reduced. (Tables 2.4 and 2.5), the cloud point of the isomerized product increases with BCFA yield. This is caused by the negative side reaction of cis/trans isomerization, which turns the cis oleic acid to trans elaidic acid (Reaume and Ellis, 2011). The melting points for the oleic acid ester is -20?C, and for the elaidic acid ester is 9?C (Knothe and Dunn, 2009). In addition, the cis/trans isomerization reaction rate is higher than the reaction rate for branching (Reaume and Ellis, 2011); therefore, with increased catalyst activity, more elaidic acid is produced raising the cloud point of the product mixture.  Figure 2.5 Correlation of surface area and carbon deposits of beta zeolite catalyst     56  2.4 Conclusion  Branch chain isomerization was achieved for all reaction conditions using a beta zeolite catalyst. Both oleic acid and methyl oleate were isomerized with oleic acid showing higher conversion values. The optimal conditions for skeletal isomerization are oleic acid run at 300?C and 1.5 MPa. However, when the cloud point of the ester is taken, the optimal results are methyl oleate run at 200?C and 3 MPa. The optimal cloud point and branch chain production based on experimental results were -15.2?C (methyl oleate at 200?C and 3 MPa) and 50% (oleic acid at 300?C and 1.5 MPa). This contrast between optimal branching and optimal cloud point is due to the fact that increasing the temperature had the negative effect of increasing the cloud point of the samples. The rise in cloud point was due to the undesired side reaction of trans isomerization, which straightens the hydrocarbon chain thereby increasing the cloud point. Although GC/MS analysis detected small concentrations of hydroxyl side chains and cracked products, they did not possess the high melting point of the trans bond methyl esters or large enough concentrations to significantly affect the cloud point. Though there is a potential benefit to creating a branch chain isomer, efforts are needed to reduce the amount of unbranched trans isomer production. The ideal goals of this study were to branch the FA and FAME, as well as to lower the cloud point. Although the first objective was accomplished, the cloud point of the FAME was not improved. However, this study provided an insight as to why creating a branch chain can negatively affect the cold flow properties of the methyl oleate. Reusability tests on the catalyst show that carbon deposits can adversely affect the catalyst activity, reducing conversion. The deposits can be partially removed in a fatty acid wash and virtually all removed through calcinations. With UFAs showing an increase in cloud point the next step is to experiment with the already high saturated fatty acids (SFAs).        57  Chapter 3 Hydroisomerization of Saturated Fatty Acid Palmitic Acid Using Pt-Doped Solid Acid beta-Zeolite Catalyst for Cloud Point Reduction of Methyl Ester  3.1 Introduction Straight isomerization of oleic acid is unsuccessful at reducing the cloud point of the product, shown in Chapter 2. The major problem was due to the effect of the double bond. Cis to trans configuration changes increased the cloud point substantially more than the effect caused by branching. The next step is to remove the source of the problem being UFAs and shift the focus of study to SFAs. The objective of this work is to use hydroisomerization catalysts in order to create branched chain fatty acids and study the effect of these branched species on the cloud point. Reaction conditions of temperature, pressure and time are studied to lower the cloud point of the fatty acid methyl esters. Additional conditions to be studied are the effect of molecular size and whether it is optimal to start with the fatty acid or methyl ester. To test these conditions the reactants used are C16 palmitic acid, C6 hexanoic acid and the methyl ester methyl palmitate. These compounds are tested for their cloud point to study the effect of the iso-species on the cloud point of a mixture. Once the best conditions are established, a kinetic study is performed in order to model the conversion of palmitic acid to isopalmitic acid, while catalyst reusability was studied to test the reaction effects on the catalyst.   3.2 Experimental Methods 3.2.1 Experimental Design  The experimental study was set up to test four factors of the hydroisomerization reaction: temperatures of 200-285?C; pressures of 1.0-4.0 MPa; reaction time of 1-16 h; and starting material. The choice of starting material (hexanoic acid, palmitic acid or methyl palmitate) was to test two different variables: the effect of molecular size (hexanoic acid (C6) versus palmitic acid (C16)); and fatty acid versus fatty acid methyl ester (palmitic acid versus methyl palmitate). The second variable was to test whether the fatty acid (FA) or fatty acid  58  methyl ester (FAME) was preferred for the hydroisomerization reaction. This was useful for determining whether the hydroisomerization should come before or after biodiesel production in order to ensure the lowest cloud point of the substance.  3.2.2 Hydroisomerization  Hydroisomerization reactions were conducted in an Autoclave Engineers (Erie, PA) mini-reactor. The reactor was charged with 25 g of the reactant (hexanoic acid, methyl palmitate, or palmitic acid) (98% Acros Organic, New Jersey USA) and 1.25 g of 0.5 wt% platinum infused beta zeolite catalyst to give a catalyst loading of 5 wt%. The reactor was purged with N2 for 5 min, and then heated to the desired temperature (200, 250 or 285?C) within 10 min. While the reactor was in the heating stage, H2 gas was switched from N2; and once the desired temperature was attained the pressure was set to the desired point (1.0, 2.5 or 4.0 MPa) with H2 gas. The reaction was then allowed to run for the desired time period (1, 3, 6, 10 or 16 h). Once the reaction time was complete, the slurry was cooled and centrifuged to separate the catalyst using a Mandel Heraeus Megafuge 16 centrifuge.    3.2.3 Esterification  For the palmitic acid to be tested for cloud point and composition, it was esterified with methanol to create a fatty acid methyl ester (FAME). This is due to the high cloud point of saturated fatty acids, which is above the limit of the cloud point analyzer (-40 to 50?C). Furthermore, the boiling point of the fatty acid is well above the maximum temperature of the GC column. The reacted palmitic acid from the autoclave was added to a STEM Omni Reacto Station (Thermo Fisher, Ottawa, ON) 250 mL reaction vessel and combined with methanol at a 10:1 molar ratio of methanol to fatty acid and 95% sulphuric acid catalyst at 2 wt% based on the fatty acid. The mixture was then heated to 65?C under reflux for 3 h. After the reaction was complete the mixture was cooled and water-washed three times to remove methanol, acid and any other impurities. The ester was dried with anhydrous CaCl2 (Fisher, Fairlawn NJ) and filtered. The last step was to take the acid value of the samples in order to ensure full conversion of the acid to ester so that this is not a factor in the analysis. The conversion data for the esterification of the fatty acid reaction products to their methyl esters are shown in Table 3.1.  59  Table 3.1 Acid values of palmitic acid hydroisomerization reaction products after undergoing esterification Temperature (?C) Pressure (MPa) Time (h) Acid Value  (mg KOH/g oil) Mass conversion to methyl palmitate* (%) 200 4.0 16 0.148 99.92 250 4.0 16 0.165 99.91 285 4.0 16 0.113 99.94 285 1.0 16 0.086 99.95 285 2.5 16 0.184 99.90 285 4.0 1 0.105 99.94 285 4.0 3 0.163 99.91 285 4.0 6 0.142 99.92 285 4.0 10 0.075 99.96 * the rest is palmitic acid  3.2.4 Catalyst Preparation  Catalyst used for this study was the beta zeolite CP814E purchased from Zeolyst International (Conshohocken, PA). The beta zeolite has a Si:Al molar ratio of 25 and a surface area of 680 m2/g. The catalyst was calcined in air at 500?C for 3 h to remove organic impurities and change the NH4+ ion to an H+ ion creating an acid catalyst. After calcinations, the catalyst was impregnated with platinum by incipient wetness. A volume of 2.5 mL of a 3.5 g/L solution of tetra-ammonium platinum chloride per gram of zeolite was added to obtain a platinum loading of 0.5 wt% on the zeolite support. The platinum solution was pulled into the pores of the zeolite through incipient wetness. Then the catalyst was dried at 60?C for 24 h to drive off all the water, leaving the platinum deposited on the surface. Next the platinum on the surface was reduced in the presence of H2 gas. The catalyst was loaded into a tube furnace and a stream of H2 at 50 mL/min was passed over the catalyst, and heated to 350?C for 3 h to ensure the platinum was reduced to an oxidation state of zero.       60  3.2.5 Analytical Techniques 3.2.5.1 Cloud Point Analysis *see Section 2.2.4.1 3.2.5.2 Gas Chromatography/Mass Spectrometry (GC/MS)  Analysis of the reaction products was carried out with the use of a Varian CP 3800 GC and Varian 4000-8 MS. The column used was a 60 m * 0.25 mm i.d. CP 50 wax column. For the C16 samples the program was set so that the injector temperature remained constant at 230?C. The column oven temperature initially set at 50?C and was held for 0.5 min; increased to 150?C at 10?C/min and held for 5 min; then increased to 220?C at 2?C/min and held for 10 min. The carrier gas (He) began at a flow rate of 0.5 mL/min for 50 min; then increased to 2.0 mL/min at 0.3 mL/min2 and held for 10 min for a total run time of 65 min. For the C6 samples the temperature was held at 150?C for the extent of the 65 min. The reaction conversion was determined by the equation (98 ? [palmitate])/98 * 100%. The quantification of product was interpolated using the area of the peak from a standard curve generated using methyl palmitate as the standard molecule.  3.2.5.3 Fourier Transform Infrared Spectroscopy (FTIR)  *see Section 2.2.4.2 3.2.5.4 Carbon Deposition, Surface Area and CO Adsorption   *For carbon deposition and surface area analysis description see Section 2.2.4.4.  CO adsorption was taken with the use of a TA instruments SDT 600 TGA unit. Catalyst samples were degassed overnight at 120?C in a N2 flow of 10 mL/min. The degassed samples were loaded into an alumina sample cup and placed into a TGA furnace. The furnace was set to 30?C and held for 10 min at a N2 flowrate of 25 mL/min. Next, the gas was switched to 50 mL/min CO/CO2/N2 mix and held for 15 min. The gas was then switched back to N2 at 25 mL/min and held for 5 min. Finally, the temperature was ramped up to 500?C at 20?C/min for a total time of 53.5 min. Although a gas mixture of N2/CO2/CO was used, it is assumed that the major adsorption species is CO. The reason behind that is N2 is an inert gas and will not chemisorb onto Pt, additionally, Pt has a much higher affinity towards CO over CO2 (Kortsdottir, 2013).   61  3.3 Results and Discussion 3.3.1 Catalyst Characterization  Transmission Electron Microscopy (TEM) was conducted on the beta zeolite catalyst (Figure 3.1) before and after impregnation with platinum in order to determine the size and dispersion of the platinum particles on the zeolite surface. The TEM images clearly display Pt particles observable as small dark spots on the beta zeolite material at an average size of approximately 5-7 nm in diameter. The particle spacing between two highly observable Pt particles is approximately 60 nm. In samples not impregnated with Pt no such dark spots are visible on the micrograph.  3.3.2 MP and PA Results  Results from the cloud point analysis show a reduction in all conditions studied. The mass conversion on all figures was calculated from the GC/MS data using the formula (98 ? [palmitate])/98 * 100%. The value of 98 was used as the initial purity of the reactants was 98%. The conditions of high temperature and pressure yielded the highest conversion to branch chain products as reported by Modhera et al. (2009) and Campelo et al. (1995). Higher conversion was observed for increasing the reaction temperature from 250 to 285?C (Figure 3.2); while, a relatively small increase in conversion (+4%) and cloud point decrease of -0.3?C occurred when the reaction temperature increased from 200-250?C. The dramatic increase in conversion from 250-285?C means that either the temperature required to reach the activation energy is most likely above 250?C and/or increases the diffusion on the reactant onto the catalyst. As for pressure, there was a significantly greater increase in reaction conversion of 15% and decrease in cloud point of 2.7?C (Figure 3.3) when pressure was changed from 1.0 to 2.5 MPa, than there was from 2.5-4.0 MPa where conversion only increased by 6% and cloud point decreased by 1.1?C. The conversion trend for pressure is due to the fact hydrogen is required to initiate the dehydrogenation reaction; however, as hydrogen partial pressure increases this shifts the equilibrium towards hydrogenation, removing reactant species from the isomerization steps. This is why the conversion increases at a greater rate at low pressure and increases slower at high hydrogen pressure, in the study by Ono (2003).  62     Figure 3.1 TEM images of zeolites: (A) with platinum; and (B) without platinum   63    The starting material had a strong effect on the hydroisomerization reaction conversion and cloud point of the products. For the cloud point analysis on hexanoic acid (the C6 species), the starting cloud point of -11?C was low enough that the acid did not have to be esterified and could be tested directly. As for the C16 species, palmitic acid, its melting point is 60?C; therefore, it had to be esterified after the reaction in order to fall into the temperature range of the cloud point analyzer. After hydroisomerization of the hexanoic acid, there was a 7.5?C reduction in the cloud point; a 9.8?C reduction for palmitic acid; and a 5.3?C reduction using methyl palmitate as a starting material. The maximum conversions were 59%, 46% and 16% for the hexanoic acid, palmitic acid and methyl palmitate starting materials, respectively. The methyl palmitate and palmitic acid both had the lowest cloud point and highest conversion at conditions of maximum time, temperature and pressure. Hexanoic acid used as a starting material achieved the lowest cloud point at high pressure (4.0 MPa), medium temperature (250?C) and 6 h, whereas the highest conversion was achieved at high pressure (4.0 MPa), high temperature (285?C) and 6 h.   In the reaction time study (Figure 3.4), the C6 species hexanoic acid run at 16, 10 and 6 h yielded conversions of 51%, 53% and 50% respectively, with cloud points of -15.5, -17.0 and -16.0?C, respectively. The conversion trend might indicate that equilibrium is reached with the smaller hexanoic acid. This trend did not appear for the larger C16 species, palmitic acid, within the time period studied. The higher conversions observed for the C6 species may be due to the smaller size of the molecule, therefore reducing the resistance into the pores of the catalyst and increasing mass transfer of the material as compared to the C16 species. Studies by Claude et al. (2001) and Jim?nez et al. (2003) show similar results using different sized hydrocarbons. Investigating the effect of pore size of catalysts, Zhang and Zhang (2007) showed a 6% increase in reaction conversion for the isomerization of oleic acid using mesoporous catalyst versus beta zeolites. In this study, a 40% increase in conversion was obtained by increasing the temperature. Thus, this study focused more on the reaction conditions, rather than different catalysts. However, future studies could benefit from an examination of different catalysts, including mesoporosity.   64    Figure 3.2 (A) Effect of temperature on cloud point; and (B) effect of temperature on mass conversion for C16 species methyl palmitate (MP) and palmitic acid (PA) and C6 species hexanoic acid (HA) at conditions of 4.0 MPa H2 pressure, 16 h reaction time, 5 wt% catalyst.   65    Figure 3.3 (A) Effect of pressure on cloud point; and (B) effect of pressure on mass conversion for C16 species methyl palmitate (MP) and palmitic acid (PA) and C6 species hexanoic acid (HA). Reaction conditions: 285?C, 16 h reaction time and 5 wt% catalyst.   66     Figure 3.4 (A) Effect of time on cloud point; and (B) effect of time on mass conversion for the hydroisomerization of palmitic acid (PA) and hexanoic acid (HA) at conditions of 285?C, 4.0 MPa H2 pressure and 5 wt% catalyst.    67  The results of the C16 species showed that palmitic acid had a much higher conversion and lower cloud point of the reaction products than the methyl palmitate reactions were able to achieve. This indicates that the fatty acid is the preferred compound over the ester when performing the hydroisomerization reaction. The reason behind this could be the increased polarity of the acid site versus the ester site. Zeolites contain ionic sites on the surface which would attract the polar carboxylic acid site over the less polar ester, yielding greater affinity to the fatty acid. The conversion curve (Figure 3.4B) indicates that the reaction is moving towards equilibrium as the trend in conversion versus time appears to be levelling off. This is supported by the C6 study where the reaction conversion does not increase from 6 to 16 h, indicating that the conversion reaches a steady-state. The equilibrium constant was calculated from the Gibbs Free Energy of Formation (?Gf) for HA and isoHA. The ?Gf was estimated from the Van Krevelen-Chermin equation. The equilibrium constant (Keq) was calculated to be 1.034, giving an equilibrium concentration of mass fraction 50.8 to 49.2% for the isoHA to HA, respectively. These results are in agreement with the experimental results showing an equilibrium concentration of approximately mass fraction 50-53 % isoHA.    The GC/MS analysis yielded four reaction products during the hydroisomerization reaction: (1) unreacted methyl palmitate; (2) methyl isopalmitate; (3) shorter chain methyl esters; and (4) unsaturated C16 methyl esters. The major contributors are compounds (1) and (2), with compounds (3) and (4) at low concentration of < 3% and < 2%, respectively. These compounds, due to their low concentrations and their lower cloud points than methyl isopalmitate or methyl palmitate, did not negatively affect the cloud point results.    In order to determine the precision of the results, triplicate tests were run for 5 samples: methyl palmitate at 285?C, 4.0 MPa for 16 h; hexanoic acid at 250?C, 2.5 MPa for 6 h; palmitic acid at 200?C, 4.0 MPa for 16 h; palmitic acid at 285?C, 1.0 MPa for 16 h; and palmitic acid at 285?C, 4.0 MPa for 16 h. There is less than a 5% difference between the results in all of the triplicate samples run. Error bars on Figures 3.2-3.4 indicate the standard deviation where the triplicate samples were tested.     68  3.3.3 Kinetic Study  A kinetic study was performed on the reaction conditions which yielded the highest conversion (285?C, 4.0 MPa) and starting material (palmitic acid), as shown in Figure 3.5. The analysis of the cloud point of the reaction products at different times shows that there is a substantial decrease from 10-16 h of 4.6?C with only a 10% increase in conversion. This comes after a 1?C decrease in cloud point from 3-10 h with a 25% increase in conversion. This could be caused by the reaction products containing multiple C16 fatty acid compounds with different branching positions. As shown in a study by Knothe and Dunn (2009), the isopalmitate (C14 branch position) and anteisopalmitate (C13 branch position) esters have different melting points due to the different branch positions. The ratio of these compounds in the product mixture would affect the cloud point in different ways. Therefore, the cloud point decrease varies with reaction conversion.  The results of the kinetic study are analyzed to generate the rate equation for the hydroisomerization of palmitic acid to branched chain C16 fatty acids at the conditions stated using a non-linear regression of the conversion versus time data. The general rate equation (Eq 3.1) was used for the overall rate equation.  ??? = ????          (3.1) -ra = rate of reaction of species a (mol L-1) h-1 k = reaction rate constant (mol L-1)-? h-1 Ca = concentration of species a (mol L-1) ? = order of reaction  After completion of the non-linear regression using least squares method, the rate constant (k) and reaction order came out to be 0.027 (mol L-1)-0.43 h-1 and 1.43, respectively, giving an overall rate equation for the conversion of palmitic acid (Eq 3.2). The results had a confidence level of >95%.  ???   =  0.027??1.43         (3.2)   69   The reaction mechanism is more thoroughly described in Chapter 1. Each step in the reaction will have a separate rate, however, the rate limiting step is the rearrangement of the carbocation into an alkyl branch (Biermann and Metzger, 2008; Zhang et al., 2004; Ono, 2003). The rate limiting step will determine the overall rate of the reaction which can be modelled in the form of A ? Products. This allows for the use of the general reaction equation (Eq 3.1). Activation energy (E) is the energy barrier that a reaction has to overcome to form products. It can be calculated from the rate constant formula:  ?(?) = ??????           (3.3) k(T) = reaction rate constant  F = frequency factor E = activation energy (J/mol) R = universal gas constant (8.314 J/mol K) T = temperature (K)   Results from various temperature runs (200, 250, 285?C) were used to calculate the different rate constants at their respective temperatures based on the rate equation (Eq 3.2). The rate constant equation is rearranged to form ln (k) = ln(A) ? (1/T)(E/R) in order to plot ln(k) vs (1/T) as shown in Figure 3.6. From this plot a slope of (-E/R) can be calculated to solve for the activation energy (E). The activation energy was calculated to be 63,037 J/mol. The activation energy is within the range that is reported in the literature of 33,500 ? 144,400 J/mol (Matsuhashi et al., 1999; Calemma et al., 2000). The literature studies were performed with different zeolite catalysts, and different chain length alkane molecules. The variation in reactants could account for the wide range in values reported. The mechanism though, is the same in fatty acids and alkanes, which would account for the experimental results falling within the range of values reported in the literature.   The other factor involved in heterogeneous catalysis kinetics is the internal effectiveness factor (?). The internal effectiveness factor was calculated to be 0.09 (*See Section 2.3.7 for calculation as the only difference in the catalyst was the Pt).   70  The rate equation described above is for the platinum impregnated beta zeolite catalyzed hydroisomerization of palmitic acid to branched chain C16 fatty acids at reaction conditions 285?C and 4.0 MPa. The rate indicates a relatively slow reaction which is shown by the long time (16 h) needed for a significant reaction conversion to take place. The slow reaction rate could most likely be due to two reasons: low internal mass transfer as indicated by an effectiveness factor of 0.09; and a higher activation barrier than in the isomerization reaction. The higher activation barrier is likely caused by the bi-functional requirement of the reaction. The reactants have to contact two different catalytic sites, Pt and Bronsted acid or the reaction will not occur.     Figure 3.5 Kinetic analysis of isomerization of OA studying the yield of branched chain (BC) fatty acids versus reaction time.   01020304050607080901000 5 10 15 20 25Yield of FA (%)Time (h)BC-FA(HI) PA other 71   Figure 3.6 Arrhenius plot of ln(k) vs. inverse temperature for the determination of activation energy  3.3.4 FTIR Results FTIR analysis was used to measure the change in bond structure. Figure 3.7 shows the heights of the peaks of interest methyl (CH3 at 2955 cm-1), methylene (CH2 at 2864 cm-1) (Doumenq et al., 1990; Van de Voort et al., 1994; Ripoche and Guillard, 2001) as well as other major peaks. The heights of the peaks were utilized for all absorbencies due to poor peak resolution, especially useful between 3100 cm-1 to 2750 cm-1 where there was no full peak separation. Taking the heights simplifies the analysis and removes errors of area integration due to poor peak resolution (Doumenq et al., 1990).  The second safeguard against error was taking the ratio of two peaks: i.e., 2955:1745 cm-1 for branch chain bonds; 2864:1744 cm-1 for straight chain bonds; and 1463:1745 cm-1 for the carbon oxygen single and double bonds. This last ratio shows consistency within the results and should not change through the experiments. This method is a quantitative way of testing whether the isomerization has occurred. From the data presented, it was shown that the heights at 2955 cm-1 increase while the peaks at 2864 cm-1 decrease, showing that methyl branching has occurred. Figure 3.6 shows the change in peak height between three samples, the increase in the 2955 cm-1 peak and decrease in the 2864 cm-1 peak with reaction conditions. Furthermore, Table 3.2 shows that no change in the 1463:1745 cm-1 peaks was observed. This indicates that the major reaction was the isomerization reaction; and any other reaction was too minor to be observed by the FTIR analysis. The FTIR results y = -7583.2x + 9.6784R? = 0.9551-7-6-5-4-3-2-100.0017 0.0018 0.0019 0.002 0.0021 0.0022ln (k(T))1/T  (K-1) 72  confirm the GC/MS results of conversion increasing with increasing temperature, pressure and time as can be seen by the CH3 peaks increasing while the CH2 peaks decrease with these trends. The absence of any other recordable change in the other peaks supports the GC/MS results where other reaction by-products are minimal and that the main components are methyl palmitate and branched C16 methyl esters.    Table 3.2 FTIR results for the hydroisomerization of the C16 species MP and PA Conditions Ratio of Absorbencies (cm-1/cm-1) Starting Material Temperature (?C) Pressure (MPa) Time (h) CH3:CH2 (2955:2864) CH2:C=O (1171:1745) C-OH:C=O (1463:1745) Control 0.1149 0.4605 0.2987 Methyl  200 4.0 16 0.1162 0.4621 0.2965 Palmitate 250 1.0 16 0.1208 0.4363 0.2968  250 4.0 16 0.1294 0.4028 0.2988  285 4.0 16 0.1396 0.3282 0.2951 Palmitic  200 4.0 16 0.1185 0.4597 0.2987 Acid 250 4.0 16 0.1308 0.3456 0.2964  285 4.0 16 0.1463 0.2058 0.2935  285 1.0 16 0.1334 0.2557 0.2982  285 2.5 16 0.1397 0.2351 0.2987  285 4.0 1 0.1135 0.4583 0.2915  285 4.0 3 0.1184 0.4542 0.2946  285 4.0 6 0.1276 0.3826 0.2971  285 4.0 10 0.1331 0.2769 0.2968   73    Figure 3.7 FTIR plot for the peak range of 3050 to 2735 cm-1 for 3 samples, (purple) unreacted PA, (red) PA reacted at conditions of 200?C and 1.0 MPa H2 pressure, and (blue) PA reacted at conditions of 285?C and 4.0 MPa H2 pressure.   3.3.5 Reactions Performed on Vegetable Oils  With a successful test through hydroisomerization to reduce the cloud point of palmitic acid methyl esters, the next step is to test the reaction on a set of vegetable oils to apply the technique. Vegetable oils consisting of canola and palm were chosen as canola oil is widely produced in Canada, and palm oil is one of the highest yielding oils, based on land use. The oils were directly hydroisomerized and isomerized to test the effects the reactions had on the oils. The oils were also pre-transesterified to study the effects on the esters. The results of the oil analysis are shown in Table 3.3. This section focuses on hydroisomerization; however, isomerization was performed on the oils to test if there was an unknown variable in the vegetable oils that could generate a positive or negative result.      74  Table 3.3 Vegetable oil hydroisomerization reactions performed using Pt zeolite catalysts Reaction CP (?C) Conversion (mass%)  Trans (mass%) BC (mass%) Transesterification 65?C, atm, 2 hrs, 7:1 methanol to oil (canola oil), KOH catalyst  ?  Hydroisomerization, 300?C, 4.0 MPa, 16 hrs, 4 wt% catalyst (Pt-zeolite) 32 100 0 12 Isomerization, 250?C, 2.5 MPa, 6 hrs, Canola oil, 4wt% catalyst (beta zeolite)  ?  Transesterification 65?C, atm, 2 hrs, 7:1 methanol to oil, KOH catalyst 2.8 12 6 0 Hydroisomerization, 300?C, 4.0 MPa, 16 hrs, Canola oil 4 wt% catalyst (Pt-zeolite)  ?  Transesterification 150?C, 2.5 MPa, 2 hrs, 7:1 methanol to oil 21 52 0 0 Transesterification 65?C, atm, 2 hrs, 7:1 methanol to oil (Palm oil), KOH catalyst  ?  Hydroisomerization, 300?C, 4.0 MPa, 16 hrs, 4 wt% catalyst (Pt-zeolite) 22 100 0 18 Isomerization, 250?C, 2.5 MPa, 6 hrs, palm oil, 4wt% catalyst (beta zeolite)  ?  Transesterification 65?C, atm, 2 hrs, 7:1 methanol to oil, KOH catalyst 18 16 12 0 Hydroisomerization, 300?C, 4.0 MPa, 16 hrs, palm oil 4 wt% catalyst (Pt-zeolite)  ?  Transesterification 150?C, 2.5 MPa, 2 hrs, 7:1 methanol to oil 20 34 0 0 CP = Cloud point Trans ? Yield of trans fatty acids/esters BC ? Yield of branched chain fatty acids/esters   The results show that hydroisomerization is not effective at reducing the cloud point of biodiesel made from palm and canola oils. The cloud points in all six cases increased from the original cloud point of -1.0 and 15?C for canola and palm oil biodiesels, respectively. This again can be attributed to the cis/trans isomerization for the isomerization reactions (Cirey and Sundburg, 2007). In the case of the hydroisomerization reactions, there is significant increase  75  in cloud point which is most likely due to hydrogenation of the UFAs (Veldsink et al., 1997). An additional factor affecting the results could be the size of the vegetable oil triglyceride. The reaction conversions and subsequent yield of the samples that were reacted under I/HI reactions first are significantly lower than the samples that were pre-transesterified. With the size of the triglyceride preventing access to the pore space of the catalyst, this would undoubtedly cause mass transfer problems and affect reaction conversions (Claude and Martens, 2000).   3.3.6 Catalyst Recycle The catalysts used for the hydroisomerization reactions were studied for their reusability. The catalyst was washed with acetone after the reaction to remove any unbound material and tested for surface area, carbon deposition, and CO adsorption. Results in Table 3.4 show a correlation between carbon deposition and surface area. Catalyst was calcined to remove coke deposits as confirmed in Table 3.4 (sample 3.2 to 3.3). After calcination in air, the used catalysts surface area increased substantially. This indicates that the carbon deposits are blocking pore spaces leading to surface area decrease after use. The effect of reaction on Pt attached to the particle surface is tested by CO adsorption. To test whether the beta zeolite would adsorb CO, the gas mixture was passed over the catalyst without Pt. With no discernible increase in mass, it was concluded that Pt must be the adsorbing species. Although a gas mixture of N2/CO2/CO was used, it is assumed that the major adsorption species is the CO. The reason behind this is that N2 is an inert gas and will not chemisorb onto Pt, additionally, Pt has a much higher affinity towards CO over CO2 (Kortsdottir, 2013). Adsorption of CO shows that the carbon deposits also block the accessibility of the Pt on the catalyst surface. The carbonaceous material that covers the catalyst was predominantly soluble in fatty acids. The uncalcined catalyst was immersed in a fatty acid solution and heated to 150?C for 15 minutes. The catalyst was then cooled and washed with acetone and analyzed for surface area, carbon deposition and CO adsorption. The ester washed catalyst (3.4) performs much better than the acetone-only washed catalyst, as shown in Table 3.4. All three tests show an improvement in carbon deposition, surface area and CO adsorption, however, not as well as calcination. The fatty acid wash was performed to investigate whether the coke deposited on the catalyst dissolves off in the first minutes of the reaction. This explains why there was still  76  catalytic activity from the used, uncalcined catalyst. The ratio of soluble to insoluble coke (i.e., the ratio of coke content from the reacted catalyst to coke content of the fatty acid washed catalyst) was 17:3 by weight, as shown in Table 3.4. This result is in agreement with a study by Villegas et al. (2006), which shows an insoluble coke content on a hydroisomerization Pt zeolite catalyst of 13-20%. On the other hand, the reduction in product yield for the recycled catalyst was 44% to 27%, giving a 40% reduction. This is comparable with a study by Lee et al. (1995), which found a 47% reduction using a recycled hydroisomerization Pt zeolite catalyst.  Table 3.4 Physical properties of isomerization and hydroisomerization catalysts before and after the reactions Entry # Catalyst Sample Yield of BCFA (%) Cloud point of reacted ester (?C) Surface area  (m2/g) Carbon deposition (%w/w catalyst) CO adsorption (mol CO/mol Pt) 3.1 HI(original) 44 20.1 374 0 7.6 3.2 R1 27 23.8 50 17 ?0 3.3 R1c 44 20.4 294 0 6.9 3.4 R1fa 40 21.2 143 3 1.8 3.5 R1aa 32 23.4 62 16 0.4 3.6 R1h 58 N/A 69 16 0 3.7 R1(@150?C) 6 28.7 342 3 5.5 3.8 R2  32 24.0 42 18 ?0 3.9 R2c 41 21.9 219 0 6.6 HI - hydroisomerization R1 ? catalyst used once before current reaction R2 ? catalyst used twice before current reaction BCFA ? branched chain fatty acid c ? calcined in air at 500?C for 2 h fa ? heated in a fatty acid bath at 150?C for 10 min aa ? heated in acetic acid bath at 80?C for 10 min h ? reaction run with hexadecane   The catalyst was also tested using an alkene hydrocarbon (hexadecene) to see if the carboxylic acid had a major effect on the reusability of the catalyst. From the tests shown the carboxylic acid does not adversely affect the coke formation or surface area coverage. Subsequently, the increased activity of the catalyst using hexadecene created slightly more coke material on the catalyst surface. Coke formation through increased catalyst activity is further confirmed through the 150?C reaction where there is almost no activity and accordingly  77  almost no fouling of the catalyst. Further testing was accomplished using an acetic acid wash to test whether the carboxylic acid group washes away the coke deposited on the surface or if it has more to do with the hydrocarbon tail of the ester. As the tests confirm, with almost no improvement after the acetic acid wash, the hydrocarbon tail is the principle reason for the improved catalyst performance.   Figure 3.8 shows the inverse correlation between the surface area and the amount of carbon deposits on the catalyst, indicating that as carbon deposits on the surface increase the available surface area decreases. In contrast to this Figure 3.9 shows the CO adsorption and catalyst surface area share a similar correlation in that as surface area drops so does the CO adsorption. This is indicative of carbon covering the Pt surfaces along with the pore spaces of the bifunctional Pt-zeolite catalyst.   Figure 3.8 Correlation between catalyst carbon deposition and CO adsorption of Pt-zeolite catalyst   78   Figure 3.9 Correlation between catalyst carbon deposition and surface area of Pt-zeolite catalyst    3.3.6 Reaction By-Products The GC/MS results showed minor traces of reaction by-products, i.e., cracking and hydroxyl compounds. Other studies performed on the hydroisomerization of alkanes show a high degree of cracking of reactants (Huybrechts et al., 2005; Soualah et al., 2008). In contast, this study shows that relatively low amounts of cracking products were produced. The reason is mainly the reaction temperature. Usual cracking conditions are temperatures >400?C (Kissin, 1996), where at these temperatures the carbenium ion breaks the molecule in two, creating smaller alkanes or olefins. Since the maximum temperature of this study is run at 285?C, cracking is not a major reaction. Furthermore, cracking is hindered by high hydrogen pressures, which are necessary for the hydroisomerization reactions. Hence, there are very few cracking products detected in the GC/MS analysis.     79  3.4 Conclusions  The objective of the study was to test the effect of hydroisomerization on fatty acids and their methyl esters. Beta zeolite catalyzed hydroisomerization of palmitic acid can be effective at relatively low temperatures of <300?C, to produce high mass yields of 42% branched chain products and a significant cloud point reduction of 10?C. Hydroisomerization showed high conversions of C6 hexanoic acid at all reaction conditions studied. Optimization of the reaction conditions was able to achieve similar results for the longer chain C16 palmitic acid. The optimal reaction conditions for palmitic acid were 285?C, and 4.0 MPa. Reusability studies on the catalyst showed that there is significant carbon deposits formed, negatively affecting surface area, and CO adsorption. This carbon accumulation limits the overall reusability of the catalyst and methods of removal will need to be further examined. This study has shown that creating branched fatty acids has the direct effect of lowering the cloud point of fatty acid methyl esters. Moving forward it has been shown that hydroisomerization is effective with the use of SFAs, however, not UFAs, or vegetable oils feedstocks. Further study is needed to determine the viability of an isomerization/hydroisomerization system, which will be examined in the next chapter.               80  Chapter 4 Synergistic Effects of Skeletal Isomerization on Oleic and Palmitic Acid Mixtures for the Reduction in Cloud Point of their Methyl Esters  4.1 Introduction Branching through isomerization and hydroisomerization reactions can have synergistic effects on the cloud point of a biodiesel sample. Different branching positions (middle or end of chain) can lower the cloud point of the sample more than the effect they have separately. Thus, the objective of this study is to test the effects of the isomerization and hydroisomerization reactions on the cloud point of saturated and unsaturated fatty acid methyl esters. This was accomplished with the use of palmitic (saturated) and oleic (unsaturated) acids. The two acids are reacted separately using hydroisomerization and isomerization reactions individually, as well as together in a mixture. Cloud points of the mixtures are then compared and the effects of the individual reactions are studied. The two different reactions have different effects on the cloud point of the biodiesel. In the isomerization reaction, the UFA is branched by creating a methyl side chain at the point of the double bond (Reaume and Ellis, 2011). However, this branching does not always lower the cloud point of the esters, due to the shift from cis- to trans-configuration in the double bond structure.  The cis/trans isomerization occurs at a much faster rate than the branching due to the thermodynamic stability of the trans double bond (Garcia-Pino et al., 2006; Cirey and Sundburg, 2007). The cloud point increase is due to the cis configuration, which is the configuration in vegetable oils, having a lower melting point, changing to trans configuration which has a higher melting point. For example, the cis configuration molecule, methyl oleate, has a melting point of -20?C and the trans configuration molecule, methyl elaidate, has a melting point of 10?C (Knothe and Dunn, 2009).  The cis/trans configuration is based on the different thermodynamic energy states that play a key role in UFAs. The cis configuration, methyl oleate is by far the most common in vegetable oils (>99%) (Karmakar et al., 2010). However, the trans configuration is the more thermodynamically stable configuration (Garcia-Pino et al., 2006). In nature, enzymes and energy additions from organisms specifically design  81  the fatty acids for a purpose, which is why they exist in cis configuration. In the isomerization reaction of UFAs, thermodynamics is the key driver of the reaction, in terms of the products containing cis or trans bonds. The isomerization reaction has one activation energy, however, there are two products formed with different thermodynamic stabilities. The trans bond is more stable due to the steric interferences of the cis configuration. This effect is shown by the simple reaction of A ?  B or C in Figure 4.1, where, C is the more thermodynamically stable compound and thus preferred in the reaction.  Figure 4.1 Energy levels for the reaction of A ?  B or C   The negative side reaction for hydroisomerization is hydrogenation, which turns UFAs into SFAs. This has detrimental effects on the melting point of the esters, given that the UFA oleic acid methyl ester has a melting point of -20?C; and the subsequent SFA stearic acid methyl ester has a melting point of 37?C (Knothe and Dunn, 2009). The desired and undesired reactions for UFAs are shown in Figure 4.2. The two undesired reactions, hydrogenation and cis/trans isomerization are not present in the reactions with the SFAs as no double bond is present, as shown in Figure 4.3.   82     Figure 4.2 Desired and undesired products from isomerization/hydroisomerization of oleic acid (OA) based on cloud point     Figure 4.3 Hydroisomerization of palmitic acid (PA)     83   4.2 Experimental Methods and Materials 4.2.1 Materials  Reactants used in the isomerization and hydroisomerization were Palmitic acid (98% Acros Organics, New Jersey, USA) (PA) and Oleic acid (NF/FCC, Fisher Scientific, New Jersey, USA) (OA). The catalyst used in the two experiments was Beta Zeolite (CP814E, Zeolyst international, Kansas City, USA), and was impregnated with a solution of tetra ammonium platinum chloride ((NH3)4PtCl2), (Aldrich Chem Co, Milwaukee, USA) using the incipient wetness method. The reactant gas used to pressurize the reaction vessel was hydrogen gas (H2) (Praxair, Mississauga, Canada). In the esterification reactions, methanol (Fisher Scientific, New Jersey, USA) was used with a catalyst of sulphuric acid (BDH, 95-98%). 4.2.2 Catalyst Preparation Two different catalysts were prepared for this study, beta zeolites with and without platinum. The beta zeolite used for both reactions (isomerization and hydroisomerization) was CP814E (Zeolyst International) with the specifications of Si/Al molar ratio of 25, and surface area of 680 m2/g. The catalyst used for the isomerization reaction was calcined at 500?C for 3 h, while that for the hydroisomerization reactions was impregnated with platinum. The compound used for impregnation was tetraammonium platinum(II) chloride, (NH3)4PtCl2. A 10 mL solution of 0.0205 M (NH3)4PtCl2 was added to 8 g of beta zeolite by incipient wetness to give a platinum loading of 0.5 wt%. The sample was then dried in an oven at 110?C for 24 h. In order to achieve platinum in the ground state it was reduced in a tube furnace at 350?C in a H2 environment for 3 h.  4.2.3 Isomerization   The beta zeolite catalyst without Pt is used for isomerization. A Parr 4848 autoclave was charged with 30 g of fatty acid and 1.2 g of catalyst to give 4 wt% catalyst. The reactor was purged with H2 gas for 5 min and heated up to 260?C while pressurized with H2 to 1.5 MPa and stirred at a rate of 600 rpm. The reaction was allowed to proceed for 6 h. When the  84  reaction was complete the reactants were cooled to 80?C. The reactor was then depressurized and contents removed. The products were centrifuged to remove all traces of catalyst from the reacted fatty acid. Lastly, the reaction product was stored in a cool dry dark place till they were analyzed. Due to the small size of the reactor and large volume of material needed for the study, multiple reactions are carried out, with all reaction products being thoroughly mixed before analysis. 4.2.4 Hydroisomerization  The beta zeolite containing 0.5% Pt by weight was used for the hydroisomerization reaction. A Parr 4848 autoclave was charged with 30 g of fatty acid and 1.2 g of catalyst to give 4 wt% catalyst. The reactor was then purged with H2 gas for 5 min, heated up to 300?C while pressurized to 4.0 MPa, and stirred at 600 rpm. After 16 h, the reactor was cooled to 80?C stopping the reaction, and depressurized. The reaction products were removed from the vessel and centrifuged. The supernatant was cooled and stored for further analysis. Just as in the isomerization steps, a large volume of material was collected and thoroughly mixed before analysis. 4.2.5 Centrifugation The products of the isomerization and hydroisomerization were too viscous to be filtered, thus centrifugation was used to remove the catalyst particles. Due to the relatively high melting point of the fatty acids, the mixture was heated to 70?C before centrifugation which was performed at 12,000 rpm for 5 min. This process was completed after every reaction to ensure that there was no further reaction once the sample was removed from the reactor.  4.2.6 Esterification For the fatty acids to be tested for cloud point and composition, they were esterified with methanol to create a fatty acid methyl ester (FAME). This was performed due to the high cloud point of saturated fatty acids, which was above the limit of the cloud point analyzer (-40 to 50?C). Moreover, fatty acids have much higher boiling points leading to very high residence times in the GC column. Esterification was carried out in an Omni Reacto Station (Thermo Scientific) using sulphuric acid as a catalyst. Approximately 20 g of reacted fatty acid were  85  charged into the reactor with 25 g of methanol and 0.4 g of sulphuric acid. This gives a methanol to fatty acid molar ratio of approximately 10:1 and 2 wt% catalyst. The sulphuric acid and methanol were first added and allowed to mix thoroughly at 350 rpm; next, the fatty acid was added to the mixture with the temperature set to 65?C. The reaction was allowed to proceed for 2 h under reflux conditions. Once the reaction was complete the ester was cooled and water washed three times to remove all traces of methanol, acid and other impurities. The washed ester was then dried using anhydrous CaCl2, and was tested for acid number to ensure >99% conversion.  4.2.7 Simulation Calculations  Simulations of the molecules were performed with density functional theory (DFT) calculations using Materials Studio 4.4.0.0 (Accelrys Inc.) software. The software predicts the most stable configuration of a system based on geometric and electronic properties of the molecules. Two systems were chosen to study the effects of adding a methyl branch to an ester: first, a control of two methyl palmitate molecules; and second, one methyl palmitate and one methyl isopalmitate. The simulations were run at 0 K, in order to study the molecules at their lowest energy state. At temperatures higher than 0 K, the increased energy can force the molecules apart, which is not based on electrical or spacial properties. Thus, studying at 0 K gives the orientation of the molecules based purely on these properties. 4.2.8 Experimental Error In order to reduce the risk of experimental error due to the large number of samples and relatively small sample size, a react and mix approach was adopted. This method reacts the two fatty acids, OA and PA, separately in large batches creating four reacted batches of: isomerized oleic acid (IO); hydroisomerized oleic acid (HO); isomerized palmitic acid (IP); and hydroisomerized palmitic acid (HP). These batches were combined in order to create the proper mixtures, and prepared and reacted in triplicates, where the standard deviations are shown in the graphs as error bars.    86  4.2.9 Product Analysis 4.2.9.1 FTIR *see Section 2.2.4.2 4.2.9.2 GC/MS Analysis of the reaction products was carried out with the use of a Varian CP 3800 GC (gas chromatograph) and Varian 4000-8 MS (mass spectrometer). The column used was a 60 m * 0.25 mm i.d. CP 50 wax column. The injector temperature was set constant at 230?C. The column oven temperature started at 100?C and was held for 0.5 min; increased to 150?C at 10?C/min and held for 5 min; then increased to 220?C at 5?C/min and held for 5.5 min. The carrier gas (He) began at a flow rate of 0.5 mL/min for 15 min; then increased to 2.0 mL/min at 0.3 mL/min2 and held for 10 min for a total run time of 30 min. 4.2.9.3 Cloud Point *see Section 2.2.4.1 4.3 Results and Discussion The objective of the study is to find the optimal conditions which give the greatest cloud point reduction. The results of the various isomerization and hydroisomerization reactions on OA and PA are shown in Table 4.1. The optimal conditions for cloud point reduction were obtained when a mixture of the fatty acids were reacted together by both isomerization and hydroisomerization. This condition gave an ester cloud point reduction of 7.5?C, and 50% branched material, which was the highest amount of any sample.       87  Table 4.1 GC/MS analysis of the methyl esters of the isomerization/hydroisomerization reactions on oleic acid (OA) and palmitic acid (PA)  Sample Sample CP (?C) PAME (%) OAME (%) SAME (%) Branched Compounds (%) Other C=C species (%) Di-C=C species (%) Cracking Products (%)       (C16) (C18)    4.1 100PA 29.0 92 0 0 0 1 0 0 0 4.2 100HP 19.1 44 3 0 48 0 0 0 3 4.3 100IP 26.9 88 0 0 3 0 2 0 1 4.4 100PAT 18.6 42 2 0 47 0 0 0 5            4.5 100OA -16.1 5 88 0 0 0 2 0 0 4.6 100IO -3.7 0 41 0 0 34 15 3 2 4.7 100HO 28.2 0 6 52 0 21 0 0 4 4.8 100OAT 5.6 2 5 38 0 44 0 0 4            4.9 55OA/45PA 17.2 43 46 0 0 0 2 2 0 4.10 55IO/45PA 13.5 43 15 0 0 21 7 0 1 4.11 55OA/45HP 12.2 21 49 0 18 0 2 2 3 4.12 55IO/45HP 10.7 22 21 0 20 20 6 2 3 4.13 55HO/45PA 25.2 42 8 25 0 17 0 0 2 4.14 55OA/45IP 16.8 39 48 0 3 0 0 0 1 4.15 55HO/45IP 24.8 40 5 30 4 17 0 0 3 4.16 55HO/45HP 23.6 20 7 28 21 16 2 0 5 4.17 55IO/45IP 13.8 40 25 0 3 16 8 0 0 4.18 55OA/45PAT 9.7 21 22 0 23 27 0 0 5 OA ? Oleic acid    C=C ? unsaturated fatty acid other than oleic acid IO ? Isomerized oleic acid   PAME ? Palmitic acid methyl ester HO ? Hydroisomerized oleic acid  OAME ? Oleic acid methyl ester PA ? Palmitic acid    SAME ? Stearic acid methyl ester IP ? Isomerized palmitic acid   C16? mixture of 16 carbon length molecules containing a methyl branch HP ? Hydroisomerized palmitic acid  C18- mixture of 18 carbon length molecules containing a methyl branch T ? Two reactions run together    55/45 ? 55 to 45 wt% CP ? Cloud point      The isomerization conversions listed in Table 4.1 are in line with previous studies yielding conversions to iso species of 30-40% (Ngo et al., 2007; Ha et al., 2009). The hydroisomerization conversions, on the other hand, are much higher than the patented study of 6% conversion to branch chain products, which only tested hydroisomerization of stearic acid (Kenneally and Connor, 2001). Much higher conversions are obtained as shown in Table 4.1 of up to 48% using palmitic acid as the starting material, and 21% using oleic acid as the starting material. The mixtures achieved a total combined conversion of approximately 50% branch chain products. Additionally, the cloud point of samples 4.10, 4.12 and 4.18 from Table 4.1 were lowered in all cases from the starting material. The reductions of 3.7 to 6.5?C are  88  consistent with a study by Yori et al. (2006) who achieved cloud point reductions of 4.0-6.5?C using solid acid catalysts and a biodiesel mixture. Figure 4.4 shows the major effects of the individual reactions on the mixture of OA and PA based on cloud point. This allows the examination of the effect of either isomerization or hydroisomerization on a single fatty acid while all other conditions remain constant. For example, Figure 4.4(A) shows the effects of isomerization of OA on the cloud point, where a reduction is shown in all but one of the cases. The reason for the increase in cloud point for the 100OA sample is the cis/trans isomerization changing the OA to a much higher cloud point trans configuration molecule (Knothe and Dunn, 2009). Figure 4.4(B), on the other hand, shows the effects of hydroisomerization of OA on the cloud point, which increased for all cases. This is caused by the hydrogenation of the double bond creating a saturated compound with a much higher cloud point than that of OA (Knothe and Dunn, 2009). The hydrogenation reaction has a higher reaction rate than the hydroisomerization reaction; therefore, the hydrogenation is complete well before the hydroisomerization (Zaera and Somorjai, 1984). The melting point data from the literature illustrates a greater increase in the melting point of methyl oleate to methyl stearate than to methyl isostearate; where, the melting points are -20, 38 and 26?C, respectively (Knothe and Dunn, 2009). Therefore this gives a 58?C increase with only a maximum potential decrease of 12?C. Hence, the hydrogenation of unsaturated fatty acids must be avoided if any reduction in cloud point or melting point is to be achieved. Figure 4.4(C) shows the reduction of cloud point through hydroisomerization of PA. The only reaction present was the branching of the PA, where there was no cis/trans isomerization or hydrogenation which could cause an increase in cloud point. Note that isomerization of PA is left out (Appendix C) of the figures as the isomerization has shown negligible effect on the saturated fatty acid.    89     -20-15-10-505101520100X 55X/45P 55X/45HP 55X/45IPCloud Point (?C)(A)UnreactedReacted-20-10010203040100X 55X/45P 55X/45HP 55X/45IPCloud Point (?C)(B)UnreactedReacted 90   (A) - (X=OA for unreacted sample and X=IO for reacted sample) (B) - (X = OA for unreacted sample and X=HO for reacted sample) (C) - (X=PA for unreacted sample and X=HP for reacted sample) (OA= oleic acid, PA= palmitic acid, IO= isomerized oleic acid, HO= hydroisomerized oleic acid, HP= hydroisomerized palmitic acid)  Figure 4.4 (A) Effect of isomerization of OA on the cloud point of the ester. (B) Effect of hydroisomerization of OA on the cloud point of the ester. (C) Effect of hydroisomerization of PA on cloud point of the ester.  Table 4.1 shows that when OA is reacted under isomerization conditions, the ester cloud point increases from -16.1 to -3.7?C, Samples 4.5 and 4.6, respectively. However, when mixed with PA, the isomerized OA reduces the ester cloud point of the mixture from 17.2 to 13.5?C, Samples 4.9 and 4.10, respectively. The cloud point is also reduced in Samples 4.11 to 4.12 from 12.2 to 10.7?C, respectively. In both reduction cases the only reaction was isomerization on the OA fraction. This effect is caused by mixing the fatty acids OA and PA, as opposed to using a single unsaturated fatty acid. This leaves a lower initial concentration of unsaturated fatty acid, greatly affecting the cloud point of the ester. The effect of branched and unsaturated compounds on cloud point is shown in Figures 4.5 and 4.6, respectively.  05101520253035100X 55O/45X 55IO/45X 55HO/45XCloud Point (?C)(C)UnreactedReacted 91    Figure 4.5 Effect of increasing iso-fatty acid methyl esters (branched compounds) on the cloud point of a fatty acid methyl ester mixture. (Mixture starts with saturated fatty acid methyl esters)  Figure 4.6 Effect of increasing cis-unsaturated fatty acid methyl esters on the cloud point of fatty acid methyl ester mixture. (Mixture starts with saturated fatty acid methyl esters)  Figure 4.5 shows that increased branching and a reduction in cloud point follow a linear trend. On the other hand, Figure 4.6 shows that increasing the unsaturated portion only largely affects the cloud point at concentrations of 60% unsaturated compounds. Therefore, with a maximum concentration of 55% unsaturated compounds, the effect on the cloud point is lower 051015200 10 20 30 40 50 60Cloud Point (oC)Branched Compounds (wt%)-20-15-10-505101520250 20 40 60 80 100Cloud Point (?C)Unsaturated Compounds (wt%) 92  than increasing from 60% to 100% unsaturated compounds. However, the addition of branched compounds is significant in terms of cloud point reduction. This general trend is shown with hydrocarbons, where the effects of mixing branched alkanes with straight chain alkanes resulted in the reduction of the melting point of the mixture in a linear trend. Similarly, the addition of double bond alkenes does not significantly affect the melting point until concentrations at or above 75% (Gibbs, 1995). Figure 4.7 shows the synergistic effect of combining both isomerization and hydroisomerization to lower the cloud point of a fatty acid mixture. The starting cloud point was the control sample of 55O/45P ester. The cloud point was increased in all the samples where oleic acid was hydroisomerized. This was caused by the hydrogenation side reaction that turns the UFA into an SFA raising the cloud point by a maximum of 57?C. The isomerization of palmitic acid samples showed no effect on the cloud point of the results. This is due to the fact that isomerization requires a double bond and the SFA palmitic acid does not contain a double bond. Therefore, there is no reaction to take place and no improvement. In samples where palmitic acid was hydroisomerized and/or oleic acid was isomerized. The cloud point was decreased as the reactions proceeded from isomerization only (55IO/45P), to hydroisomerization only (55O/45HP), to isomerization and hydroisomerization (55IO/45HP), and lastly to the reactions run together (55O/45PT). The general trend in cloud point reduction was caused by the increase in branching reactions (isomerization/hydroisomerization). The 55IO/45PA and the 55OA/45HP samples have one branching reaction each, i.e., either isomerization or hydroisomerization. The 55IO/45HP sample undergoes two branching reactions. Similarly, in the 55OA/45PAT sample, there are now two branching reactions; however, double the effect is produced due to both isomerization and hydroisomerization occurring on each component, yielding in effect four branching reactions.   93   55O/45P ? 55% oleic acid to 45% palmitic acid 55HO/45P ? 55% hydroisomerized oleic acid to 45% palmitic acid 55HO/45IP ? 55% hydroisomerized oleic acid to 45% isomerized palmitic acid 55HO/45HP ? 55% hydroisomerized oleic acid to 45% hydroisomerized palmitic acid 55O/45IP ? 55% oleic acid to 45% isomerized palmitic acid 55IO/45IP ? 55% oleic acid to 45% palmitic acid 55IO/45P ? 55% isomerized oleic acid to 45% palmitic acid 55O/45HP ? 55% oleic acid to 45% hydroisomerized palmitic acid 55IO/45HP ? 55% isomerized oleic acid to 45% hydroisomerized palmitic acid  55/45 T ? 55% oleic acid to 45% palmitic acid run under both isomerization and hydroisomerization  Figure 4.7 Synergistic effects of methyl branching on improving the cloud point of a mixture of fatty acid methyl esters (arrows indicate reference axis, area% for bar graph and cloud point for line graph)     The lower cloud point from increased branching may be attributed to the reduction in London dispersion forces that occur between the hydrocarbon tails. The increased distance caused by the methyl group reduces the intermolecular dispersion forces causing the molecules not to stack as well as the straight chain molecules, which in effect lowers the cloud point of the mixture. To further confirm the trend, molecular simulations have been conducted to show the increased space between the molecules caused by the branching, as shown in Figure 4.8.  94  The simulations indicate that the average distance between the carbon backbone of the esters increases from 2.85 to 4.35 ?, due specifically to the addition of the methyl branch. This increase in distance causes a reduction in molecular dispersion attractive forces resulting in a more fluid structural arrangement, therefore lowering the cloud point of the mixture. In addition to the increased spacing between the carbon chains, the methyl branch forces the oxygen molecules in the ester bond closer causing the bond to rotate. This rotation forces the methyl group from the ester bond into another spacial plane, which would force molecules in that plane further away from the esters.   Figure 4.8 Simulated stable configurations of: (A) methyl palmitate and methyl palmitate; and (B) methyl palmitate and branched methyl isopalmitate   95   The isomerization and hydroisomerization reactions create several by-products in very low concentrations. These compounds include cracking products, i.e., smaller chain fatty acids, alkanes and olefinic hydrocarbons. The cracking products are grouped together and shown in Table 4.1. One product not present in any sizeable quantity is the dimer compounds. These did not show up in the GC/MS analysis. Additionally, evidence of their structure was shown in the FTIR results (Table 4.2), however in readings of <1%, too small to be significant. The dimer C=O group absorbs at the wavelength 1711 cm-1(Ismail et al., 1993) and is less than 1% of the ester C=O bond group. This may be due to the relatively small pore size of the beta zeolite structure (5.5 to 7.4 ?) not allowing the two fatty acids to come together and form a dimer. Again in this study dimers are not a major concern due to the fact that they will only be formed in isomerization reactions involving the OA, which is less than ? of the samples. Table 4.2 FTIR wavelengths of isomerized and hydroisomerized oleic and palmitic acids at various group frequencies  Group Wavelength  Sample (-CH2-)  (-CH3)  (C=O) Dimer (C=O) Ester  2864 cm-1 2955 cm-1 1711 cm-1 1742 cm-1      100OA 0.4363 0.0260 0 0.6333 100IO 0.4175 0.0344 0.00168 0.6322 100HO 0.4280 0.0301 0 0.6341 100OAT 0.4094 0.0377 0.00082 0.6338 100PA 0.1418 0.0179 0 0.0924 100HP 0.1304 0.0265 0 0.0921 100IP 0.1410 0.0189 0 0.0918 100PAT 0.1293 0.0271 0 0.0929  The goal of the isomerization and hydroisomerization reactions is to create a branched fatty acid without any negative side reactions creating higher cloud point molecules. This could be accomplished by the isomerization reaction first, followed by hydroisomerization. Another option could be to shorten the duration of the hydroisomerization reaction while increasing the isomerization reaction. A third option is to use less Pt in the hydroisomerization reaction to slow down the hydrogenation of the double bonds. Whatever the combination of reactions this study has shown that there are potential benefits to using isomerisation reactions for cloud point improvement of biodiesel.  96  4.4 Cloud Point Model Based on Fatty Acid Content Fatty acid content has major effects on the cloud point of a biodiesel product. However, there are many different combinations of fatty acids in fats and vegetable oils, subsequently yielding a variety of possible cloud points. While there are many different fatty acids, four main categories can be classified. These are ploy-UFAs, mono-UFAs, long chain SFAs, and medium chain SFAs. These fatty acids are present in different ratios in fats and oils and affect the cloud point of biodiesel in different ways. This leaves a gap in the knowledge about exactly how adding/removing one or more of these fatty acids will affect the cloud point of the biodiesel.   This section examines the effect of four fatty acid esters on the cloud point of a biodiesel mixture. The fatty acids chosen were poly-UFA linoleic acid (LA), mono-UFA oleic acid (OA), long chain SFA palmitic acid (PA), and medium chain SFA lauric acid (LAU). These fatty acids were mixed, esterified to make a biodiesel, and tested for cloud point. The ratio of fatty acids and cloud point results are shown in Table 4.3. An equation was generated based on the cloud point results of Table 4.3. The cloud point versus fatty acid data was analyzed in Poly-Math software using a non-linear regression analysis to generate an equation, Eq 4.1 shown below. Two values that were not incorporated into the final equation were Samples 4.33 and 4.34, which were considered outliers and did not fit the general trend of the rest of the data. These two samples did not fit the trend of the data because the LAU samples abruptly changed from decreasing to increasing cloud point at 90%. As this was the only FA to have the change in trend, it could not be fit into the model. The analysis revealed that the cloud point of biodiesel is linearly proportional to the SFAs (PA and LAU) and cubically proportional to the UFAs (LA and OA). This is consistent with the results from Figure 4.4 which shows a non-linear decrease in cloud point with addition of UFA esters.  A plot of the experimental versus predicted cloud point values for the model equation is shown in Figure 4.9. The results are in agreement with the model analysis showing a >95% confidence in the equations predictions with the R2 value of 96%. Vegetable oils biodiesel from literature studies are shown in Figure 4.9 to test the predictability of the model with three oil biodiesels from cottonseed, lard and palm oils (Joshi et al., 2010; Smith et al., 2010).  97  Table 4.3 Experimentally determined cloud points of different fatty acid content biodiesels Sample Fatty Acid (mass %) Cloud Point   Linoleic acid Oleic acid Palmitic acid Lauric acid (?C) 4.19 10 10 70 10 22.2 4.20 25 10 55 10 18.4 4.21 50 10 30 10 8.1 4.22 75 10 5 10 -8.8 4.23 90 5 0 5 -14.2 4.24 100 0 0 0 -27.9 4.25 10 25 55 10 18.9 4.26 10 50 30 10 8.9 4.27 10 75 5 10 -10.3 4.28 5 90 0 5 -13.8 4.29 0 100 0 0 -16.1 4.30 10 10 55 25 16.9 4.31 10 10 30 50 5.4 4.32 10 10 5 75 -8.0 4.33 5 5 0 90 -1.3 4.34 0 0 0 100 1.1 4.35 0 0 100 0 29.0  CP = (-1.5*10-5)(LA3) ? 0.00125LA2 + 0.04LA ? (2.0*10-5)(OA3) ? 0.00115OA2 +  0.06OA + 0.338PA ? 0.13LAU              (4.1)  98   Figure 4.9 Experimental cloud point vs. predicted cloud point (Literature data taken from Joshi et al., 2010; Smith et al., 2010)  4.5 Conclusions Two cloud point improvement reactions: isomerization and hydroisomerization were tested, for their effect on OA and PA, separately and together. Isomerization had a negative effect on the pure oleic acid sample, increasing the ester cloud point; however, when oleic acid was isomerized and mixed with palmitic acid the ester cloud point was reduced.  Hydroisomerization indicated a negative effect on oleic acid, and a positive effect on palmitic acid with respect to ester cloud point reduction. The optimal run combination with respect to ester cloud point reduction was the mixture of oleic acid and palmitic acid run under both reactions together. This optimal condition gave an ester cloud point reduction of 7.5?C and a total branched component of 50%. This study has successfully proven that when combined, the two reactions isomerization and hydroisomerization can lower the cloud point of a mixture of UFAs and SFAs. With a successful test on a mixture of pure compounds of both SFAs and UFAs the next step is to find a suitable vegetable oil that fits this process.  R? = 0.9607-40-30-20-10010203040-40 -30 -20 -10 0 10 20 30 40Predicted CP (?C)Experimental CP (?C)Measured Literature 99  Chapter 5  Use of Isomerization and Hydroisomerization Reactions to Improve the Cold Flow Properties of Vegetable Oil Based Biodiesel  5.1 Introduction Work from previous chapters has been focused on the use of fatty acids as starting material for the isomerization/hydroisomerization (I/HI) reactions. Detailed analysis using model compounds in previous chapters is now applied to vegetable oils which are the starting material used for commercial scale production of biodiesel. The design of the experiments has been developed from previous work on isomerization/hydroisomerization (I/HI) of fatty acid from Chapters 2 and 3. The vegetable oils in their original state are too large for the I/HI reactions, they have to undergo hydrolysis to release the fatty acids. The fatty acids are then put under the I reaction, followed by the HI reaction. This process was found to be the optimal process for cloud point reduction of a SFA/UFA mixture, based on Chapter 4. The products undergo esterification to test for various fuel qualities and composition. The objective of this study is to find an optimal vegetable oil with reference to cloud point improvement based on the I and HI reactions. If a successful improvement process is discovered there will be extra energy and materials associated in utilizing it. This will add to the cost of production. In order to test the validity, a complete energy use comparison along with a total operating cost comparison will be performed of the two processes; standard and improved biodiesel production. Energy and feedstock costs can be calculated based on inputs; however, capital costs of equipment will depend on the reaction conditions and physical requirements. This is where a process simulation such as a HYSYS model is beneficial. The simulation will be used to calculate the capital cost of the system and amortize it over the life of the facility.  In the course of scale up production of biodiesel, simulations can be a useful tool for predicting many of the factors associated with increased production. Sizing of equipment, pricing, feedstock properties and energy requirements can all be estimated using a HYSYS simulation model (Lee et al., 2011). HYSYS models can, in addition, be effective tools to compare different  100  processes and study the operational and economic feasibility of the different processes (West et al., 2008). A subprogram within Aspen/HYSYS is the ICARUS cost estimator, which has been used for over 30 years in commercial design of plants. ICARUS provides estimates of economic data based on provided detailed designs allowing for minimal input and quick analysis (Lee et al., 2011). While there are assumptions made in creating the models, these are minor compared to the assumptions that will have to be made without modelling (Zhang et al., 2003). Two studies have examined the use of HYSYS models to compare different biodiesel production processes. Lee et al. (2011) have examined alkali catalyzed versus supercritical production of biodiesel, using different feedstocks (fresh and waste vegetable oils). The model predicted that the supercritical method would be the most cost effective process based on feedstock cost. An additional study by West et al. (2008) compares the economic feasibility of four different production processes using a similar feedstock. The model predicts that the most cost effective model is the heterogeneous acid catalyzed process, based on lowest capital investment.  These studies show that HYSYS is an effective tool in determining the economic feasibility of different processes based on a variety of factors.  5.2 Methods 5.2.1 Catalyst Preparation *See Section 4.2.2 5.2.2 Hydrolysis  Fatty acids contained in the oils must be hydrolyzed prior to undergoing isomerization and hydroisomerization, due to the large size of the triglyceride molecules. An Autoclave Engineers mini reactor was charged with 20 g of water and a select amount of vegetable oil to give water to oil molar ratio of 20:1. The reactor was heated to 230?C, pressurized to 4.0 MPa with N2 gas and stirred at 500 rpm. The reaction was allowed to proceed for 2 h after which time the mixture was cooled to 80?C, depressurized and moved to a separatory funnel. The mixture separates into two phases: the water and glycerol heavy phase; and the fatty acid light phase. The fatty acids are removed and stored in a cool dry place to await isomerization/hydroisomerization.  101  5.2.3 Isomerization  The beta zeolite catalyst without Pt was used for isomerization. A Parr 4848 autoclave was charged with 30 g of fatty acid and 1.2 g of catalyst to give 4 wt% catalyst. The reactor was purged with H2 gas for 5 min and heated up to 260?C while pressurized with H2 to 1.5 MPa and stirred at a rate of 600 rpm. The reaction was allowed to proceed for 6 h. When the reaction was complete the reactants were cooled to 80?C. The reactor was then depressurized and contents removed. The products were centrifuged to remove all traces of catalyst from the reacted fatty acid. Lastly, the reaction product was stored in a cool, dry, dark place to await analysis.  5.2.4 Hydroisomerization The beta zeolite containing 0.5% Pt by weight was used for the hydroisomerization reaction. A Parr 4848 autoclave was charged with 30 g of fatty acid and 1.2 g of catalyst to give 4 wt% catalyst. The reactor was then purged with H2 gas for 5 min, heated up to 300?C while pressurized to 4.0 MPa with H2, and stirred at 600 rpm. After 16 h, the reactor was cooled to 80?C, and depressurized. The reaction products were removed from the vessel and centrifuged at 12,000 rpm for 5 min. The supernatant was cooled and stored for further analysis.   5.2.5 Esterification Fatty acids were esterified with methanol to create a fatty acid methyl ester (FAME) prior to testing for cloud point and composition. This was due to the high cloud point of saturated fatty acids, which is above the limit of the cloud point analyzer (-40 to 50?C). Additionally, fatty acids have much higher melting points leading to very high residence times in the GC column. Esterification was carried out in an Omni Reacto Station (Thermo Scientific) using sulphuric acid as a catalyst. Approximately 20 g of reacted fatty acids were charged into the reactor with 25 g of methanol and 0.4 g of sulphuric acid. This gave a methanol to fatty acid molar ratio of approximately 10:1 and a 2 wt% catalyst. The sulphuric acid and methanol were added and allowed to mix thoroughly at 350 rpm; next, the fatty acid was added to the mixture and the temperature was set to 65?C. The reaction was allowed to proceed for 2  102  h under reflux conditions. Once the reaction was complete the ester was cooled and water washed three times to remove all traces of methanol, acid and other impurities. The washed ester was then dried using anhydrous CaCl2 and an acid number was taken to ensure >98% conversion.  5.2.6 Analytical Techniques 5.2.6.1 GC/MS Analysis of the reaction products was carried out with the use of a Varian CP 3800 GC (gas chromatograph) and Varian 4000-8 MS (mass spectrometer). The column used was a 60 m * 0.25 mm i.d. CP 50 wax column. The injector temperature was set at 230?C. The column oven temperature started at 100?C and was held for 0.5 min; increased to 150?C at 10?C/min and held for 5 min; then increased to 220?C at 5?C/min and held for 5.5 min. The carrier gas (He) began at a flow rate of 0.5 mL/min for 15 min; then increased to 2.0 mL/min at 0.3 mL/min/min and held for 10 min for a total run time of 30 min. A 10 mg sample was measured out and dissolved into enough heptane to give a total mass of 10 g. The solution was then mixed at 3000 rpm for 30 seconds. A 1 g sample of the solution was then dissolved in 9 g of heptane and mixed at 3000 rpm for 30 seconds, to yield 100 mg/kg concentration of original sample. This process was repeated for all samples.  5.2.6.2 Cloud and Pour Points *see Section 2.2.4.1 5.2.6.3 Viscosity Measurements  Viscosity measurements were taken using a Brookfield DV-E viscometer at 40?C. The rotor was set at 30 rpm giving a shear rate of 36.7/s. Sample measurements were allowed to stabilize for 15 min and a viscosity reading was taken. 5.2.6.4 Acid Number  For the acid number test, 1 g sample was added to 10 mL of propanol with 3 drops of phenylthaline solution and stirred at 180 rpm. A 0.1 M solution of KOH was added to the  103  mixture drop by drop until the end point was detected. The volume of KOH was recorded and used to calculate the acid number. The hydrolyzed sample was split into two: one for original sample; and the second to undergo the isomerization/hydroisomerization reactions. The samples were then esterified and two different acid numbers generated for the original and reacted samples (Table 5.1). The acid number measurement ensures that the esterification reaction conversions are high enough such that only minute amounts of fatty acids are present in the final sample, thus not affecting the cloud point analysis. Table 5.1 Acid number results for hydrolysis and esterification reactions Sample  Hydrolysis Esterification   Acid Number (mg KOH/g oil) Conversion (%) Acid Number (mg KOH/g fatty acid) Conversion (%) Palm oil Original 197.3 95.2 3.71 98.21    Reacted   2.98 98.56 Coconut oil Original 249.6 96.9 4.02 98.64    Reacted   2.94 98.84 Rapeseed oil Original 189.3 94.7 1.72 99.14    Reacted   3.26 98.37 Corn oil Original 192.3 96.2 2.42 98.47    Reacted   3.01 98.50 Soybean oil Original 190.4 95.2 1.98 98.75    Reacted   2.85 98.96 Animal fat Original 187.9 94.9 2.22 98.69    Reacted   2.04 98.84 Lard Original 185.2 92.4 3.51 98.25    Reacted   3.01 98.50 Olive oil Original 189.6 94.9 1.79 99.08    Reacted   3.02 98.47 Palm Kernel  Original 249.6 96.9 4.02 98.64    Reacted   2.94 98.84 Waste Oil Original 185.2 92.4 3.51 98.25    Reacted   3.01 98.50 I/HI ? Isomerized/Hydroisomerized   5.2.6.6 Oligomer Determination  Oligomers (dimers) were analyzed using a TA instruments SDT Q600 Thermogravametric Analyzer (TGA). A 30 mg sample of esterified product was added to an alumina sample cup and inserted into the TGA furnace. The furnace was then flushed with N2  104  at 50 mL/min for 15 min at 30?C. Next the flowrate was set to 25 mL/min N2 and heated to 250?C at 20?C/min and held for 5 min. Then the furnace was heated to 360?C at 20?C/min and held for 5 min, for a total time of 41.5 min. The maximum temperature was chosen to be above the boiling points of fatty acids (approximately 360?C) and below the boiling point of oleic acid dimers (approximately 660?C), in addition to be lower than the temperature required for significant cracking to take place. The leftover residue from 5 runs (required 5 repeats due to small amount of residue leftover) was mixed and analyzed by Fourier Transform Infrared Spectroscopy (FTIR). A scan range was set at 4500 ? 600 cm-1 with 128 scans co-added together. The wavelengths of note were the 1711 cm-1 (C=O group of dimer acids) (Ismail et al., 1993) and the 1742 cm-1 (C=O group of FAMEs) (Doumenq et al., 1990). 5.2.7 HYSYS Set-Up Details  The two economic models were set up using HYSYS plant NetVer3.2 (ASPEN tech Cambridge, MA) to conduct the simulations. The HYSYS library contained information for the following compounds: methanol, triolein, glycerol, oleic acid, methyl oleate, water, KOH, H2SO4 and hydrogen. The vegetable oil was represented by the triolein and the fatty acids by oleic acid. Compounds not available in the HYSYS library were built and properties estimated using the Hypo Manager. The compounds that had to be built were isostearic acid and methyl isostearate. Isostearic acid was used for the branched chain fatty acids and methyl isostearate for the branched chain esters. The models were built to mix, react, separate, clarify and store reactants and products of the reaction based on process and unit designs built within HYSYS. Sizing of the equipment was based on an input of 1000 kg/d and the ICARUS subprogram was used to perform the economic analysis.  5.3 Results 5.3.1 Cloud Point Analysis The results of the GC/MS compositional analysis (Table 5.2) show a substantial reduction in the unsaturated fatty acid (UFA) portion after the I/HI reactions. This was because the HI reaction hydrogenates the carbon carbon double bond creating saturated fatty acid (SFA) compounds. This has the negative effect of raising the cloud point of the high UFA  105  containing vegetable oil biodiesels. However, the low UFA oil (palm and coconut) esters had significant improvements in cloud point, as shown in Figure 5.1. Coconut and palm kernel oil biodiesel show a greater cloud point reduction than palm oil biodiesel. This was due to coconut and palm kernel oil having a higher percentage of medium chain fatty acids (MCFA), which are fatty acids with C10 ? C14 chain lengths.  Results in Table 5.2 show that MCFAs produce greater yields of branch chain species over the long chain fatty acids (LCFA), which are composed of >C14 length fatty acids.  Studies have shown that smaller chain hydrocarbons have higher initial reaction rates over long chain hydrocarbons (Jim?nez et al., 2003; Ono, 2003; Soualah et al., 2008). This is further shown by Claude and Martens (2000), where the initial reaction rate decreases as carbon number increases from C10 to C18 alkane samples. Similar interpretation may be applied to fatty acids, for the reason that it is the hydrocarbon chain that undergoes the hydroisomerization reaction.  The change in cloud point of the samples is in direct correlation with UFA content and the production of branched chain fatty acids (BCFA). The largest increase in cloud point was shown for the high UFA oil biodiesels, rapeseed and olive oil because of high UFA to SFA hydrogenation, in addition to low BCFA yield. The oils with low initial UFA content such as coconut, palm kernel and palm oils benefit from the reactions, yielding biodiesel cloud point reductions of -16.5, -12.9 and -4.7?C, respectively. This is attributed to the high BCFA production with modest UFA to SFA hydrogenation. However, high UFA content oils show a slight benefit from branching. Corn oil, for example, had a BCFA yield of 40%, the largest of the high UFA oils, and only a 4.4?C increase in cloud point, despite 95% of its UFA content converted to SFAs. This is compared to the 10.9-19.5?C increase of other high UFA content oils (e.g., soybean and olive oils). However, high UFA samples do not follow any specific trend with respect to cloud point changes. In other words, cloud point changes cannot be predicted solely from the initial fatty acid profiles. For example, corn and soybean oils (having similar initial fatty acid profiles as shown in Table 5.2) show different cloud point changes of +4.4 and +10.9?C, respectively (Figure 5.1).     106  Table 5.2 GC/MS compositional analysis of reactant and product stream from isomerization and hydroisomerization sequential reactions Sample  Unsaturated FAME (mass %) Saturated Long Chain FAME (>C16 chain length) (mass %) Saturated Medium Chain FAME  (C10-C14 chain length) (mass %)   Poly Mono SC BC SC BC Palm oil Original 10 40 49 0 1 0    Reacted 0 18 33 36 2 2 Coconut oil Original 2 6 12 0 71 0    Reacted 0 0 13 9 34 39 Rapeseed oil Original 32 62 6 0 0 0    Reacted 2 4 57 26 3 0 Corn oil Original 59 28 13 0 0 0    Reacted 0 4 48 40 0 2 Soybean oil Original 61 24 15 0 0 0    Reacted 2 3 43 21 5 14 Beef fat Original 4 43 43 0 3 0    Reacted 0 12 62 17 0 0 Lard Original 10 44 40 0 2 0    Reacted 0 3 59 25 2 2 Olive oil Original 11 71 16 0 0 0    Reacted 0 0 74 21 0 0 Palm Kernel Original 5 7 13 0 73 0    Reacted 0 0 14 8 34 33 Waste Oil Original  60 25 14 0 1 0    Reacted 2 10 49 29 0 0 Original ? Neat fat or oil sample Reacted ? Reacted under Isomerization/Hydroisomerization conditions FAME ? Fatty acid methyl ester Poly - Poly-unsaturated fatty acid methyl esters Mono ? Mono-unsaturated fatty acids SC ? Straight chain fatty acid methyl esters BC ? Branch chain fatty acid methyl esters   107   Figure 5.1 Cloud point changes of the various vegetable oil biodiesels before and after the isomerization and hydroisomerization sequential reactions (error bars represent standard deviation of triplicate tests)   As explained by Knothe (2008), fatty acid profile of vegetable oils strongly affects many properties of the fuel, including cloud point, viscosity, and oxidative stability. Based on Moser (2009), low UFA biodiesel results in a high oxidative stability index (OSI) value. With reacted coconut oil biodiesel showing low UFAs, it is inferred that a biodiesel with high OSI and the low cloud point of -2.3?C is achieved. Table 5.3 shows the effect of the reactions on the flow properties. Note that as the cloud point decreases due to the I/HI reactions, the flow properties (pour point, viscosity and density) improve as well. Conversely, as the cloud point increases due to the I/HI reactions, the flow properties become increasingly inferior.     -5051015202530Cloud Point (?C)Vegetable oilOrignial Reacted 108  Table 5.3 Effect of isomerization/hydroisomerization reactions on fuel quality Sample  Cloud Point (?C) Pour Point (?C) Viscosity(mm2/s) (@40?C) Density(kg/m3) (@15?C) Palm oil Original 17.5 15 4.42 882    Reacted 12.8 9 4.08 864 Coconut oil Original 14.2 9 4.58 850    Reacted -2.3 -3 3.57 824 Rapeseed oil Original -1.0 -15 4.53 874    Reacted 17.3 18 6.95 902 Corn oil Original 11.5 9 5.01 880    Reacted 15.9 12 6.52 894 Soybean oil Original 4.6 0 4.62 882    Reacted 15.5 9 6.68 886 Animal fat Original 16.9 15 5.84 892    Reacted 28.1 24 6.99 905 Lard Original 13.1 0 5.02 873    Reacted 21.1 18 6.04 865 Olive oil Original 7.8 -3 4.21 875    Reacted 27.3 27 7.03 903 Palm Kernel Original 10.9 9 4.58 850    Reacted -2.0 -3 3.57 824 Waste Original 11.5 9 5.01 880    Reacted 15.9 12 6.52 894 I/HI ? reacted under isomerization/hydroisomerization BD - Biodiesel  Branched chain alcohols have been shown to yield esters with lower melting points versus their straight chain counterparts. Several studies (Lee et al., 1995; Marchetti & Errazu, 2008) show melting point reductions using branched alcohols opposed to methyl alcohol. Additionally, studies by Yori et al. (2006) and Ngo et al. (2007) were able to achieve cloud point and melting point reductions using isomerization reactions on biodiesel and oleic acid methyl ester, respectively. This study further cements the idea that branching biodiesel is a sound method of cloud point reduction.  The two classes of fatty acids that cause low cloud points are UFA and BCFA methyl esters; however, these fatty acids lower the cloud point of biodiesel in different ways. As shown in Chapter 3, the addition of BCFA methyl esters lower the cloud point in a linear fashion; whereas, the addition of UFA methyl esters only lowers the cloud point significantly at concentrations greater than 60%. Therefore, unless the oil contains a significant amount of  109  UFAs, the cloud point will remain high, yielding greater importance to increasing BCFA content for biodiesel cloud point reduction.  5.3.2 Reaction By-Products The two main reasons for large cloud point increases are the cis/trans isomerization and hydrogenation side reactions (Reaume & Ellis, 2011; 2012). These two reactions significantly raise the melting point of a UFA methyl ester, by changing it to the trans and hydrogenated compound. For example, cis methyl oleate has a melting point of -20?C; while, the trans methyl elaidate has a melting point of 10?C. The saturated methyl stearate has a melting point of 37?C, and lastly the branched methyl isostearate has a melting point of 26?C (Knothe and Dunn, 2009). The increases in melting point from the side reactions cannot be made up solely by the decrease caused by branching. Therefore, branching must be maximized while minimizing the side reactions.  The I/HI reaction creates other by-products aside from the fatty acids. These are cracking products (smaller chain fatty acids, alkanes and olefinic hydrocarbons), hydroxyls, dimers and coke material left on the catalyst.  Dimers are formed from the oligomerization of two UFA compounds, creating a double fatty acid molecule attached at the carbon carbon double bond site (Tolvanen et al., 2007). These compounds cannot be classified using our GC/MS, due to their high boiling points; therefore, an isolation of dimers was conducted using a TGA unit under N2 atmosphere. During heating of the samples, the FAMEs, cracking products, hydroxyls and any unesterified fatty acids evaporate off (due to their boiling points of < 360?C) leaving a small amount of residue. This residue (boiling point > 360?C) is analyzed with FTIR for dimer C=O bonds. The residue was less than 2% of the overall material and FTIR analysis showed it was >95% dimer fatty acids. The cracking products and hydroxyls were analyzed by the GC/MS analysis. The total weight of the reaction by-products was less than 7 wt%, with small chain fatty acids accounting for the bulk of the material.      110  5.3.3 Optimization of Low Unsaturated Fatty Acid Oils (Coconut/Palm Kernel)   It has been shown that the low unsaturated oils (coconut and palm kernel oils) yield significant cloud point reduction when put through the I/HI process. Due to this fact, these oils are selected to be studied for the improvement process and compared to their standard production process. The standard production process is described as straight transesterification of vegetable oils with methanol, and KOH used as a catalyst. This will be compared to an improvement process utilizing the I/HI reactions. The main issue facing the improvement process is where it will take place, before or after the biodiesel esterification/transesterification reactions. Studies in Chapters 2 and 3 show that there is a significant increase in reaction conversion and branch chain yield in addition to a larger decrease in cloud point when the improvement reactions I/HI are performed on the fatty acids over the esters. Additionally in Section 3.3.5, it is shown that the isomerization reaction has little to no effect when used on vegetable oils. This means that a hydrolysis reaction is necessary to release the fatty acids before the I/HI reactions can take place. The last factor that comes into play in designing a production process for the improvement reactions is that the oils selected contain very low amounts of unsaturated fatty acids. This aspect raises the question of whether the isomerization reaction, which requires UFAs, is necessary. A small scale study was performed to test the difference between the I/HI process and straight hydroisomerization improvement, results are shown in Table 5.4. The conversion results in Table 5.4 were calculated from the equation [C(SFA)initial ? C(SFA)final]/C(SFA )initial * 100%.  Table 5.4 Comparison of I/HI improvement against hydroisomerization improvement   Reaction CP (?C) UFA (mass %) Long Chain SFA (mass %) Medium Chain SFA (mass %) Reaction Conversion (mass%)    SC BC SC BC  Control  (no reaction) 14.2 8 12 0 71 0 N/A I/HI -2.1 0 13 9 34 36 51 HI only -1.9 0 13 7 33 34 48   111  The results in Table 5.4 show that there is a negligible difference between I/HI process and the HI only process when using coconut oil. The miniscule improvement gained from the extra isomerization reaction does not outweigh the extra energy, cost and materials that would be required to perform the isomerization reaction. Therefore the improvement process is described as follows: Hydrolysis ?  Hydroisomerization ? Esterification This process will now be studied against the traditional transesterification process for extra energy and costs required. 5.3.4 Energy Use The I/HI process uses more energy than the standard production process. However, due to the coconut oil yielding the highest BCFA yield and greatest cloud point reduction, there is no need for the isomerization process. Energy calculations were used to examine whether there is still a net energy gain from producing improved biodiesel. Energy use for the hydroisomerization improvement process was calculated against the standard biodiesel production process. Energy calculations were based on the energy it takes to raise the temperature of the reactants to reaction temperatures. This was accomplished using specific heats of the reactants, in addition to calculating the heat losses through insulated reaction vessel walls and pressurization of gases. Heat capacities were calculated using differential scanning calorimetry (DSC), a Safire standard and the equation:                                            ????? = ??????                                                                      (5.1) Cv = specific heat capacity of sample (J/g?C) Cv? = specific heat capacity of Safire standard (J/g?C) m = mass of sample (g) m? = mass of Safire (g) y = distance to baseline of sample (mm) y? = distance to baseline of Safire (mm)  The energy required to heat the materials to reaction temperature was calculated using the formula:  112                                                  ? = ?????                                                                     (5.2) Q = Energy (J) ?T = difference in temperature between reaction temperature and initial temperature (?C) m = mass of sample (g) Cv = specific heat capacity of sample (J/g?C)  The heat loss through the reaction vessel walls was calculated using the formula:                                                     ? =??????                                                                       (5.3) Q = Energy (J) ?T = difference in temperature between reaction temperature and initial temperature (?C) Ri = insulation thermal resistance (m2K/Wm) As = surface area of vessel (m)  The R value was assumed for a high temperature insulation material which had a value of 145.7 m2K/Wm. Lastly, the energy required to compress the hydrogen to 4.0 MPa was calculated using the formula (American Gas Compression Services Inc.). The compression was assumed to be adiabatic (isentropic) with 3 stages of compression.                                ?? = {[144?????33000(??1)] [(?2  ?1 )(??1??) ? 1]}                                            (5.4) HP = horsepower (W) Pf = Pressure drop between inlet and outlet (kPa) V = volume (m3) k = ratio of specific heats  P2 = pressure required (kPa) P1 = initial pressure (kPa) N = number of compression stages  The two processes are defined as:  1. Hydroisomerization improvement: Hydrolysis ?  Hydroisomerization ? Esterification 2. Standard process: Transesterification The process comparison based on energy usage is shown in Table 5.5. Calculations gave total energy usages of 83.19 kJ and 4.04 kJ based on a 25 g sample of vegetable oil for  113  the hydroisomerization and standard process, respectively. The extra 79.15 kJ of energy for the improvement process is substantial compared to the low energy use of the standard process; however, the energy content of biodiesel and glycerol by-product is 1030 kJ and 58 kJ, respectively. This gives the total energy for the additional process at <10% of the energy yielded from the biodiesel for the scale we have examined. Furthermore, other side benefits to the process are: not using the transesterification reactions which eliminates KOH, thus avoiding emulsification problems (Souza et al., 2009); and the hydrolysis method producing a cleaner glycerol product which can be sold as pure glycerol as opposed to crude glycerol (Thompson and He, 2006). Table 5.5 Energy use comparison for the hydroisomerization improvement biodiesel production process versus the standard biodiesel production Production Stage Improvement  (kJ/25g oil) Standard  (kJ/25g oil) Hydrolysis   Heating 14.08 N/A Loss  4.60 N/A Gas compression  0.01 N/A Mixing 0.11 N/A    Hydroisomerization   Heating 14.30 N/A Loss  42.05 N/A Gas compression  0.01 N/A Mixing 0.88 N/A    Esterification/Transesterification   Heating 6.24 3.04 Loss  0.79 0.89 Gas compression  0 0 Mixing 0.11 0.11    Total Energy Use 83.19 4.04  Similar processes are used to upgrade petroleum based fuels. Catalytic cracking and octane upgrading utilize similar solid acid catalysts and H2 gas at high pressures and temperatures >500?C (Ono, 2003). The major difference in the petroleum industry is the lower cost of crude feedstock and much larger volumes available to process, versus the higher cost and lower volume vegetable oils. If a combined biofuel/petroleum process can be integrated  114  into existing petroleum processing facilities, costs of upgrading biofuels can be significantly reduced or absorbed by the higher profit petroleum products. 5.3.5 Economic Analysis  In addition to extra energy requirements, the improvement process will require extra equipment and materials. These extra requirements will increase the cost of production using cloud point improvement over the standard production process. An economic analysis was performed in order to compare the cost of the two different processes, improved and standard. The two production processes described earlier are the standard production process transesterification (Figure 5.2) and the improvement process with hydroisomerization (Figure 5.3). The analysis studies the cost for materials, energy and capital costs of the reactors and other components of the system. The analysis is based on a plant that uses a continuous flow system and processes 1000 kg of coconut oil per day.  A HYSYS model was developed to study the costs associated with both processes using real time flow values. The composition of each of the streams in the models standard and improved processes are shown in Tables 5.6 and 5.7, respectively. The reactors in Figures 5.2 and 5.3 were modelled as conversion reactors and are assumed to be continuous stirred tank reactors. The reactants enter the reactors at their respective flow rates and immediately reach conversions required for the reactor exit. Reaction conversions for hydrolysis, transesterification and esterification are all assumed to be 100%; and hydroisomerization is assumed 45% based on studies conducted in Chapters 2 and 3. In Reactor R2 (Figure 5.3) is set up as a packed bed reactor and proper filter materials ensure no loss of catalyst material during operation. The hydrogen gas used in R2 is re-pressurized after reaction completion and stored for later use. There is assumed only minimal losses of hydrogen gas. The reusability of the catalyst is assumed to be four recycles. Lastly, though there were impurities that were produced in the experimental reactions these were never more than 7% of the total weight. In addition, impurities have shown no effect on reaction conversions or cloud point in the experimental studies. Therefore, it is assumed that they have no effect on the system and carry through to the end of each process.   115   S ? Storage tank P ? Pump MIX ? Mixer R ? Reaction vessel (R1 ? Transesterification) CL ? Clarifier  SEP ? Separator  Figure 5.2 Process flow diagram for standard biodiesel production process  Table 5.6 Composition data for process flow diagram of standard production process Feed Stream Material (mass %) Flow rate (kg/min) Temperature (?C) Pressure (MPa)  Oil CH3OH NaOH Glycerol Ester    1 0 100 0 0 0 0.228 25 0 2 0 0 100 0 0 0.010 25 0 3 10 0 0 0 0 0.694 25 0 4 0 94.6 5.4 0 0 0.238 25 0 5 0 100 0 0 0 0.130 80 0 6 0 0 1.4 10.1 88.5 0.802 60 0 7 0 0 0 0 100 0.710 55 0 8 0 0 12.2 87.7 0 0.092 55 0   The standard transesterification process starts out with KOH and methanol being mixed together in the mixer MIX1. Then they are transferred to the reactor R1 and mixed with the  116  vegetable oil. Inside R1 the transesterification reaction occurs, all values and procedures are described in Section 2.2.3. Once the reaction is complete, methanol is distilled off and recovered. Biodiesel and glycerol mixture is moved to the separator SEP1 where glycerol is settled out and the biodiesel stored for use. The glycerol is then purified in the clarifier CL1 to remove catalyst and produce a pure glycerol product.    S ? Storage tank P ? Pump MIX ? Mixer R ? Reaction vessel (R1-Hydrolysis; R2 ? Hydroisomerization; R3 ? Esterification) SEP ? Separator  Figure 5.3 Process flow diagram for cloud point improvement biodiesel production process       117  Table 5.7 Composition data for process flow diagram of improvement production process Feed Stream (mass %) 1 2 3 4 5 6 7 8 Oil 0 100 82.5 0 0 0 0 0 Water 100 0 17.5 100 0 0 0 0 Glycerol 0 0 0 0 10.7 100 0 0 Fatty acids 0 0 0 0 89.3 0 100 0 Hydrogen 0 0 0 0 0 0 0 100 Iso-fatty acids 0 0 0 0 0 0 0 0 Methanol 0 0 0 0 0 0 0 0 Sulphuric acid 0 0 0 0 0 0 0 0 Esters 0 0 0 0 0 0 0 0 Iso-esters 0 0 0 0 0 0 0 0 Flowrate (kg/min) 0.183 0.694 0.877 0.128 0.749 0.08 0.669 0.050 Temperature (?C) 25 25 25 110 110 80 80 25 Pressure (MPa) 0 0 0 0 500 0 500 600  Table 5.7 cont?d Feed Stream (mass %) 9 10 11 12 13 14 15  Oil 0 0 0 0 0 0 0  Water 0 0 0 6.3 0 75.8 0  Glycerol 0 0 0 0 0 0 0  Fatty acids 0 55 0 0 0 0 0  Hydrogen 100 0 0 0 0 0 0  Iso-fatty acids 0 45 0 0 0 0 0  Methanol 0 0 100 0 100 0 0  Sulphuric acid 0 0 0 2.0 0 24.2 0  Esters 0 0 0 50.6 0 0 55.1  Iso-esters 0 0 0 41.2 0 0 44.9  Flowrate (kg/min) 0.050 0.669 0.585 0.752 0.502 0.046 0.706  Temperature (?C) 200 250 25 90 90 70 70  Pressure (MPa) 600 400 0 0 0 0 0   118  The improvement production process starts out by mixing water and vegetable oil, which is then pumped to reactor R1 for hydrolysis to take place. The values and procedures for hydrolysis are described in Section 5.2.1. Next, the fatty acids, glycerol and excess water are sent to the separator SEP1 for water and glycerol settling; while, the fatty acids are moved to R2. There the fatty acids are mixed with hydrogen in a fixed bed reactor for hydroisomerization, based on the procedure described in Section 5.2.3. The excess hydrogen is recovered and the hydroisomerized fatty acids are sent to the R3 unit. In the R3, unit the fatty acids are mixed with methanol and sulphuric acid to undergo esterification, as described in Section 5.2.5. Next the excess methanol is distilled and recovered; while, the biodiesel, water and acid are sent to SEP2 for water washing. Finally, the biodiesel is separated and stored for use.   The results of the economic analysis are shown in Table 5.8 as a breakdown of the different costs, which are separated into four different sections: Raw Material, Energy, Equipment and Miscellaneous. The raw material and energy costs are based on the daily inputs and calculated to a price per litre of fuel produced. The Equipment costs are generated from the ICARUS subprogram using the HYSYS model. The costs are spread over an estimated 25 year lifespan and then calculated based on the production of fuel to give a cost per litre of fuel produced. Miscellaneous costs are the use of water for reaction/washing, estimated material losses and waste removal.   The improvement process is shown to add significant costs to the production of biodiesel (approximately 17%). This improvement although costly, allows biodiesel made from coconut, palm and palm kernel oils access to northern fuel markets. With average cloud points of ?15?C, colder climates are out of reach for these fuels, but with cloud point reduction the biodiesel could have cloud points of ?-2.0?C allowing for use in colder northern climates. Additionally, these are three of the highest yielding oils (Figure 1.4), which have the added benefit of using less land to produce the same amount of fuel. This can help solve some of the ethical issues of using farm land for fuel. The second overall advantage to using the improvement process is improved yields of the final product. Without the use of KOH or alkaline catalyst, there is no soap formation. Soaps cause emulsification of the biodiesel which is then lost in the water/dry wash or cause the fuel to hold water and not separate out impurities.  119  This means the improvement process will help the biodiesel to meet ASTM standards. Impurities such as glycerol (free and bound), soaps, free fatty acids, water and methanol are lower in the improved production process without using such catalysts. Glycerol is removed in the hydrolysis step; while, soaps are not formed due to lack of alkaline catalyst. Furthermore, free fatty acids are reduced due to the higher reaction conversions of esterification. Water and methanol separate out easier due to the absence of soaps.  Table 5.8 Economic comparison for the hydroisomerization biodiesel production process versus the standard biodiesel production Item Standard Process ($/tonne biodiesel produced) Improvement Process ( $/tonne biodiesel produced) Raw Material   Oil 1350 1350 Methanol 64.5 64.5 Catalyst   Zeolite N/A 236.4 Acid/Base 28.0 14.8 Hydrogen N/A 10.2  Energy (see Section 3.4 for breakdown) (Multiplied to 1000 kg/day production)  3.14 64.94 Equipment (averaged over 25 years) 6.0 92.42 Pumps 1 3.0 Storage tanks 0.5 1.5 Mixers 0.5 1.0 Separators 1.5 4.0 Reactors 2.5 82.92  Miscellaneous  (includes water use/waste, and methanol losses of 5%)  3.9 8.8  Total 1450.14 1847.06   In addition to producing a lower quality product, extra costs are associated with the standard production process that was not included in the model. These costs are associated  120  with the use of fuel additives, required to improve the fuel quality. The additives are pour point depressors and oxidative stability enhancers. Pour point depressors reduce crystal agglomeration and allow the fuel to flow at temperatures below their cloud point. Oxidative stability enhancers are used to prevent the fuel from breaking down into harmful components during storage. These costs along with a 5% loss in product due to soap formation add up to $0.16-0.29/L. This value is based on $0.015/L for the oxidative stability enhancers (Biodiesel Notes, 2013), $0.075-0.205/L for pour point depressors (Reliance Energy, 2013) and $0.07/L for losses due to soap formation (Kiss et al., 2006). This reduces the cost difference from $0.41/L to $0.115-0.25/L. The use of additives does not produce a fuel without any issues. Pour point depressors do not prevent the formation of crystals only their agglomeration, which leaves small wax crystals in the fuel potentially clogging injectors and cause ignition problems in cold weather (Schmidt and Van Gerpen, 1996).   While it is assumed that the reactions of hydrolysis and esterification reach 100% the laboratory studies show that hydrolysis reaches approximately 97% and the esterification reaction reaches approximately 99% (Table 5.1). These conversions have shown no negative observable effects on the product in lab scale. While problems associated with scale up are inevitable any adjustments made will be pure assumptions with no data to back them up. In the case of hydrolysis the conversion was assumed to be 45% based on the approximate lab scale conversion. The presence of both the iso-fatty acid and straight chain fatty acid would not affect the esterification results as shown in Table 5.1, with no observable difference between reacted and unreacted fatty acid esterification in the experimental sections.  5.4 Conclusions A reduction in cloud point of vegetable oil biodiesels has been shown through the use of isomerization/hydroisomerization. Improvements were successful on low UFA containing vegetable oils, yielding reductions of 4.7 to 16.5?C. The reactions were unsuccessful at reducing the cloud points of the high UFA vegetable oils. Of the biodiesels that were improved, other fuel qualities were improved along with the cloud point, including viscosity and pour point. With the implementation of I/HI reactions, there was a sizable energy increase (<10% of the energy contained in biodiesel is needed for the improvement process), due to the high reaction temperatures and additional hydrolysis step. However, cloud point reduction along  121  with reduced biodiesel washing and cleaning help to offset the negatives associated with the energy increase. The economic analysis shows that the hydroisomerization improvement process does add to the cost of production. The extra cost is significant and would most likely be too prohibitive for biodiesel production to be carried out in this manner. However, this is a starting point for any future development in cloud point improvement. The process is successful and future work must be focused on reducing energy and material costs for the cloud point improvement of biodiesel.                     122  Chapter 6 Conclusions  6.1 General Findings Heterogeneous catalytic reactions have shown that straight skeletal isomerization of unsaturated fatty acids (UFAs) was not successful at lowering the cloud point of the product. In contrast, hydroisomerization of saturated fatty acids (SFAs) has shown a clear reduction in the cloud point of the final product. The isomerization of UFAs yielded a product with a cloud point up to 18.5?C higher than the original ester; while, hydroisomerization yielded a product with a cloud point 10?C lower than the original ester at optimal conditions of 285?C, and 4.0 MPa hydrogen pressure. This leads to the conclusion that isomerization is not effective at reducing the cloud point of biodiesel; in fact, the opposite effect of cloud point increase is shown.  This increase is in tandem with increases in reaction conversion. The increase in cloud point is caused by the side reaction cis/trans isomerization, thus, as reaction conversion increases, cis/trans conversion increases and cloud point increases.    Although the isomerization experiments were unsuccessful at reducing the cloud point of UFAs, UFAs are contained in vegetable oils and fats. A more complete analysis was needed to study the effects of branched trans UFAs on the cloud point of a mixture of fatty acids. The mixture study (Chapter 4) showed that while the cloud point of pure UFAs increases after isomerisation, the same cannot be said for a mixture of fatty acids. When isomerized oleic acid (OA) was added to pure SFA palmitic acid (PA) the cloud point was reduced 3.7?C, over a pure OA/PA mixture. This leads to the conclusion that isomerization can reduce the cloud point of a mixture of fatty acids. The study in addition concluded that a combination of isomerization and hydroisomerization reactions performed on a mixture of OA and PA yielded the largest cloud point reduction. These conclusions now give an optimal reaction set to lead into a study on vegetable oils.  The vegetable oil results showed that only low UFA containing oils cloud points were reduced using the isomerization/hydroisomerization reactions. High UFA oils were unsuccessful; however, oils that contained low UFAs yielded biodiesel cloud points that were  123  reduced up to 16?C. Once the improvement was found an optimal process was concluded as Hydrolysis ? Hydroisomerization ? Esterification.   An energy and economic analysis was performed to study extra requirements of the optimal improvement process. It was found that the process adds $0.115-0.41/L to the fuel depending on the amount of post processing that is required for standard biodiesel. This leads into the next section of where future work should focus. 6.2 Strengths and Weaknesses of Research Strengths Weaknesses ? All components of the reaction are available on the open market, there is no need to synthesize any materials. ? Catalysts are benign zeolites, there is no need for dangerous or volatile chemicals. ? Process reduces the cloud point of the already high cloud point low UFA oil biodiesels. ? Products are not limited to the biodiesel industry. Others including lubricants and cosmetics could use BCFAs. ? Requires a hydrolysis step because oil molecule are too large. This adds energy and cost to the production. ? High temperatures and high pressure are required for significant conversions. ? High capital costs for the system.  6.3 Potential Applications 1. Biodiesel production is already widespread, yet two of the highest producing oils yield biodiesels with a cloud point that is too high for northern markets. If a cloud point reduction method can be successfully implemented, that would provide a value added product for these oils.  2. Isomerization and hydroisomerization reactions are already in use in the petroleum industry for octane boosting, wax prevention and hydro-cracking. This work has shown that the biodiesels can withstand harsh conditions of up to 300?C and not break down.  If biodiesel or fatty acids could be added to the petroleum process at a certain point then costs could be shared with the regular diesel pool. This would add a  124  renewable aspect to the fuel pool and give biodiesel the quality fuel properties it needs.  3. Cosmetics and lubricants are two main industries that use oils and fatty acids. BCFAs benefit these industries by providing a better melting point and highly stable product. The applications are smaller than a fuel market; however, this is a higher value market.   6.4 Future Work  The main focus of future work is to reduce the cost associated with the cloud point improvement. This could be accomplished in two possible methods:  The first method is to rework or design a new catalyst to facilitate the hydroisomerization reactions. The current catalyst, beta zeolite requires very high temperatures and pressures to catalyze the reactions. If a strong acid catalyst could be developed that reduces the need for high temperatures and pressures, this would save on reaction vessel material costs. Additionally there is room for improvement in the process design. As it stands the optimal process is Hydrolysis ? Hydroisomerization ? Esterification, if a catalyst could be developed to perform all three reactions it could save material costs.  Lastly, due to the fact that the catalyst cannot process oils and need fatty acids, if a catalyst could be developed to process pure oils that would eliminate the need for hydrolysis and save an entire step from the production.    The second method utilizes that the hydroisomerization reaction has been used in the petroleum industry for more than 50 years. Biodiesel is similar to petroleum fuel in that the only difference is the ester group of the biodiesel. Future work could strive to integrate biodiesel into a section of the petroleum process to try to reduce cost, and blend at source. Studies would have to be performed to test the effects of oxygen in the biodiesel on the petroleum process.    125  References  American gas compression services inc. Retrieved 10/03, 2012, from http://www.agcsi.com/estimate/chart.html  Amonette, J. (2012). Mitigation of climate change with biomass: Technical potentials and factors affecting implementation. 17th International Oil Palm Conference and Exopalma 2012, Cartagena de Indias, Colombia. 17, 94-95.  Beer, T., Grant, T., Williams, D., & Watson, H. (2002). Fuel-cycle greenhouse gas emissions from alternative fuels in australian heavy vehicles. Atmospheric Environment, 36(4), 753-763.  Bernstein, H. J. (1962). Bond energies in hydrocarbons. Trans Faraday Soc. 58, 2258-2306. Berrios, M., & Skelton, R. (2008). Comparison of purification methods for biodiesel. Chemical Engineering Journal, 144(3), 459-465.  Biermann, U., & Metzger, J. O. (2008). Synthesis of alkyl-branched fatty acids. European Journal of Lipid Science and Technology, 110(9), 805-811.  Biloen, P., Dautzenberg, F. M., & Sachtler, W. M. H. (1977). Catalytic dehyrogenation of propane to propene over platinum and platinum-gold alloys. Journal of Catalysis, 50, 77-86. Biodiesel Notes. (2013). Comparison of oxidative stability additives for biodiesel. Retrieved 03/01, 2013, from http://www.biodiesel.org/reports/20110201_gen-426.pdf  Boronat, M., Viruela, P., & Corma, A. (1996). Theoretical study of the mechanism of branching rearrangement of carbenium ions. Applied Catalysis A: General, 146(1), 207-223.  Boronat, M., Viruela, P., & Corma, A. (1998). Theoretical study of the mechanism of zeolite-catalyzed isomerization reactions of linear butenes. J Phys Chem: A, 102, 982-989. Calemma, V., Peratello, S., & Perego, C. (2000). Hydroisomerization and hydrocracking of long chain n-alkanes on Pt/amorphous SiO2-AlO3 catalyst. Applied Catalysis A: General, 190, 207-218. Campelo, J., Lafont, F., & Marinas, J. (1995). Hydroisomerization and hydrocracking of n-hexane on Pt/SAPO-5 and Pt/SAPO-11 catalysts. Zeolites, 15(2), 97-103.  Cerqueira, H. S., Sievers, C., Joly, G., Magnoux, P., & Lercher, J. A. (2005). Multitechnique characterization of coke produced during commercial resid FCC operation. Industrial & Engineering Chemistry Research, 44(7), 2069-2077.   126  Carey, F., & Sundburg, R. (2007). Advanced organic chemistry: Structure and mechanisms 5th ed, Springer Science and Business Media LLC.  Claude, M. C., & Martens, J. A. (2000). Monomethyl-branching of long n-alkanes in the range from decane to tetracosane on Pt/H-ZSM-22 bifunctional catalyst. Journal of Catalysis, 190(1), 39-48.  Claude, M. C., Vanbutsele, G., & Martens, J. A. (2001a). Dimethyl branching of long n-alkanes in the range from decane to tetracosane on Pt/H-ZSM-22 bifunctional catalyst. Journal of Catalysis, 203(1), 213-231.  Claude, M. C., Vanbutsele, G., & Martens, J. A. (2001b). Dimethyl branching of long n-alkanes in the range from decane to tetracosane on Pt/H-ZSM-22 bifunctional catalyst. Journal of Catalysis, 203(1), 213-231.  Collins, K. (2008). The role of biofuels and other factors in increasing farm and food prices. Retrieved March, 11, 2013, from http://www.foodbeforefuel.org/files/roleofbiofuels6-19-08.pdf  Crude oil and commodity prices. (2011). Retrieved August 22, 2011, 2011, from Oil-Price.net, http://www.oil-price.net Dimian, A. C., Bildea, C., Omota, F., & Kiss, A. (2009). Innovative process for fatty acid esters by dual reactive distillation. Computers & Chemical Engineering, 33(3), 743-750.  Doumenq, P., Guiliano, M., Bertrand, J., & Mille, G. (1990). GC/FT-IR analysis of fatty acid methyl esters. Applied Spectroscopy, 44(8), 1355-1359.  Janaun, J., & Ellis, N. (2010). Glycerol etherification by tert-butanol catalyzed by sulfonated carbon catalyst. Journal of Applied Sciences, 10(21), 2633-2637.  Foglia, T. A., Nelson, L. A., Dunn, R. O., & Marmer, W. N. (1997). Low-temperature properties of alkyl esters of tallow and grease. Journal of the American Oil Chemists' Society, 74(8), 951-955.  Forzatti, P., & Lietti, L. (1999). Catalyst deactivation. Catalysis Today, 52(2), 165-181.  Friedlingstein, P., Houghton, R., Marland, G., Hackler, J., Boden, T. A., Conway, T., & Qu?r?, C. (2010). Update on CO2 emissions. Nature Geoscience, 3(12), 811-812.  Garcia-Pino, A., Buts, L., Wyns, L., & Loris, R. (2006). Interplay between metal binding and cis/trans isomerization in legume lectins: Structural and thermodynamic study of P. angolensis lectin. Journal of Molecular Biology, 361(1), 153-167.   127  Gecan, R., Johansson, R., & Fitzgerald, K. (2009). The impact of ethanol use on food prices and greenhouse gas emissions. Retrieved 03/15, 2013, from http://www.cbo.gov/ftpdocs/100xx/doc10007/04-08-ethanol.pdf  Gibbs, A. (1995). Physical properties of insect cuticular hydrocarbons: Model mixtures and lipid interactions. Comparative Biochemistry and Physiology Part B: Biochemistry and Molecular Biology, 112(4), 667-672.  Gonz?lez G?mez, M., Howard-Hildige, R., Leahy, J., & Rice, B. (2002). Winterisation of waste cooking oil methyl ester to improve cold temperature fuel properties. Fuel, 81(1), 33-39.  Guisnet, M. (2002). ?Coke? molecules trapped in the micropores of zeolites as active species in hydrocarbon transformations. Journal of Molecular Catalysis A: Chemical, 182, 367-382.  Guisnet, M., & Ribeiro, F. R. (2011). Deactivation and regeneration of zeolite catalyst. edited by: Hutchings, G.J. Catalytic Science Series 19, Imperial College Press. Ha, L., Mao, J., Zhou, J., Zhang, Z. C., & Zhang, S. (2009). Skeletal isomerization of unsaturated fatty acids on beta zeolites: Effects of calcination temperature and additives. Applied Catalysis A, General, 356(1), 52-56.  Heidbreder, A., H?fer, R., Gr?tzmacher, R., Westfechtel, A., & Blewett, C. W. (1999). Oleochemical products as building blocks for polymers. Lipid/Fett, 101(11), 418-424.  Hossain, A., Salleh, A., Boyce, A. N., Chowdhury, P., & Naqiuddin, M. (2008). Biodiesel fuel production from algae as renewable energy. American Journal of Biochemistry and Biotechnology, 4(3), 250-254.  Huybrechts, W., Vanbutsele, G., Houthoofd, K. J., Bertinchamps, F., Laxmi Narasimhan, C. S., Gaigneaux, E. M., & Baron, G. V. (2005). Skeletal isomerization of octadecane on bifunctional ZSM-23 zeolite catalyst. Catalysis Letters, 100(3), 235-242.  Ismail, A., Van de Voort, F., Emo, G., & Sedman, J. (1993). Rapid quantitative determination of free fatty acids in fats and oils by fourier transform infrared spectroscopy. Journal of the American Oil Chemists' Society, 70(4), 335-341.  Jacobson, K., Gopinath, R., Meher, L. C., & Dalai, A. K. (2008). Solid acid catalyzed biodiesel production from waste cooking oil. Applied Catalysis B: Environmental, 85(1), 86-91.  Jim?nez, C., Romero, F. J., Rold?n, R., Marinas, J. M., & G?mez, J. P. (2003). Hydroisomerization of a hydrocarbon feed containing n-hexane, n-heptane and cyclohexane on zeolite-supported platinum catalysts. Applied Catalysis A: General, 249(1), 175-185.   128  Joshi, H., Moser, B. R., Shah, S. N., Mandalik, A., & Walker, T. (2010). Improvement of fuel properties of cottonseed oil methyl esters with commercial additives.  Eur J. Lipid Sci Technol,  112, 802-809. Kalam, M., & Masjuki, H. (2002). Biodiesel from palmoil--an analysis of its properties and potential. Biomass and Bioenergy, 23(6), 471-479.  Kaltschmitt, M., Reinhardt, G., & Stelzer, T. (1997). Life cycle analysis of biofuels under different environmental aspects. Biomass and Bioenergy, 12(2), 121-134.  Karmakar, A., Karmakar, S., & Mukherjee, S. (2010). Properties of various plants and animals feedstocks for biodiesel production. Bioresource Technology, 101(19), 7201-7210.  Kenneally, C. J., & Connor, D. S. (2001). Process for the branching of saturated and/or unsaturated fatty acids and/or alkyl esters therof US Patent US 20010044550. 2001 Nov 22. Kiss, A. A., Dimian, A. C., & Rothenberg, G. (2006). Solid acid catalysts for biodiesel Production??Towards sustainable energy. Advanced Synthesis & Catalysis, 348(1?2), 75-81.  Kiss, A. A., Dimian, A. C., & Rothenberg, G. (2007). Biodiesel by catalytic reactive distillation powered by metal oxides. Energy & Fuels, 22(1), 598-604.  Kiss, A. A., Omota, F., Dimian, A. C., & Rothenberg, G. (2006). The heterogeneous advantage: Biodiesel by catalytic reactive distillation. Topics in Catalysis, 40(1), 141-150.  Kissin, Y. V. (1996). Chemical mechanism of hydrocarbon cracking over solid acidic catalysts. Journal of Catalysis, 163(1), 50-62.  Knothe, G. (2007). Some aspects of biodiesel oxidative stability. Fuel Processing Technology, 88(7), 669-677.  Knothe, G. (2008). ?Designer? biodiesel: Optimizing fatty ester composition to improve fuel properties. Energy Fuels, 22(2), 1358-1364.  Knothe, G., & Dunn, R. O. (2009). A comprehensive evaluation of the melting points of fatty acids and esters determined by differential scanning calorimetry. Journal of the American Oil Chemists' Society, 86(9), 843-856.  Koivusalmi, E., & Jakkula, J. (2006). Process for the manufacture of hydrocarbons European Patent Application EP 1 795 576 A1 2006.  129  Kortsdottir, K., Fernandez, C. D., & Lindstrom, R. R. (2013). Influence of hydrogen and operation conditions on CO2 adsorption on Pt and PtRu catalysts in a PEMEC. ECS Electrochemistry Letters, 2(5), 41-43. Lammertsma, E. I., de Boer, H. J., Dekker, S. C., Dilcher, D. L., Lotter, A. F., & Wagner-Cremer, F. (2011). Global CO2 rise leads to reduced maximum stomatal conductance in florida vegetation. Proceedings of the National Academy of Sciences, 108(10), 4035-4040.  Lardon, L., He?lias, A., Sialve, B., Steyer, J., & Bernard, O. (2009). Life-cycle assessment of biodiesel production from microalgae. Environmental Science & Technology, 43(17), 6475-6481.  Lee, I., Johnson, L. A., & 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, S., Posarac, D. & Ellis, N. (2011). Process simulation and economic analysis of biodiesel production processes using fresh and waste vegetable oils and supercritical methanol. Engineering Research and Design, 89, 2626-2674. Li, C., Novaro, O., Munoz, E., Bold?, J., Bokhimi, X., Wang, J., & Gomez, R. (2000). Coke deactivation of Pd/H-mordenite catalysts used for C< sub> 5</sub>/C< sub> 6</sub> hydroisomerization. Applied Catalysis A: General, 199(2), 211-220.  Li, Y. F., Zhu, J. Q., Liu, H., He, P., Wang, P., & Tian, H.P. (2000). Bulletin of the Korean Chemical Society, 32(6), 1851-1858. L?nyi, F., & Valyon, J. (2001). A TPD and IR study of the surface species formed from ammonia on zeolite H-ZSM-5, H-mordenite and H-beta. Thermochimica Acta, 373(1), 53-57.  Marchetti, J., & Errazu, A. (2008). Comparison of different heterogeneous catalysts and different alcohols for the esterification reaction of oleic acid. Fuel, 87(15-16), 3477-3480.  Matsuhashi, H., Shibata, H., Nakamura, H., & Arata, K. (1999). Skeletal isomerization mechanism of alkenes over solid superacid of sulphated zirconia. Applied Catalysis A: General, 190 207-218. McHughen, A., & Smyth, S. J. (2012). Regulation of genetically modified crops in USA and canada: American overview. Regulation of Agricultural Biotechnology: The United States and Canada, 1, 35-56.  Mittelbach, M., & Remschmidt, C. (2004). Biodiesel the comprehensive handbook (First ed.). Vienna, Austria: Boersedruck Ges.m.b.H.   130  Modhera B. K., Chakraborty, M., Parikh, P., & Jasra, R. (2009). n-hexane hydroisomerization over nano-crystalline zeolite beta. Petroleum Science and Technology, 27(10-12), 1196-1208.  Moser, B. R. (2009). Comparative oxidative stability of fatty acid alkyl esters by accelerated methods. Journal of the American Oil Chemists' Society, 86(7), 699-706.  Ngo, H. L., Nu?ez, A., Lin, W., & Foglia, T. A. (2007). Zeolite-catalyzed isomerization of oleic acid to branched-chain isomers. European Journal of Lipid Science and Technology, 109(3), 214-224.  O'Brien, D. (2012). World corn supply - demand and trends. Retrieved 03/12, 2013, from http://www.agmanager.info/marketing/outlook/news/letters/archives/GRA2V.OUTLOOK.05-03-11.pdf  Olah, G. (1972). Stable carbocations. CXVIII. General concept and structure of carbocations based on differentiation of trivalent (classical) carbenium ions from three-center bound penta- of tetracoordinated (nonclassical) carbonium ions. Role of carbocations in electrophilic reactions. Journal of the American Chemical Society, 94(3), 808-820. Ono, Y. (2003). A survey of the mechanism in catalytic isomerization of alkanes. Catalysis Today, 81(1), 3-16.  Pachauri, R. K., & Reisinger, A. (2007). Contribution of working groups I, II and III to the fourth assessment report of the intergovernmental panel on climate change. Retrieved 01/15, 2013, from http://www.ipcc.ch/publications_and_data/ar4/syr/en/contents.html  Park, K. C., & Ihm, S. K. (2000). Comparison of Pt/zeolite catalysts for n-hexadecane hydroisomerization. Applied Catalysis A: General, 203(2), 201-209.  Reaume, S. J., & Ellis, N. (2011). Optimizing reaction conditions for the isomerization of fatty acids and fatty acid methyl esters to their branch chain products. Journal of the American Oil Chemists' Society, 88, 661-667.  Reaume, S., & Ellis, N. (2012) Synergistic effects of skeletal isomerization on oleic and palmitic acid mixtures for the reduction in cloud point of their methyl esters. Energy and Fuels, 26(7), 4514-4520. Reliance Energy. (2013). Biodiesel kits online. Retrieved 03/02, 2013, from http://www.biodiesel-kits-online.com/biodiesel-additive.html  Revelle, R., & Suess, H. E. (2010). Carbon dioxide exchange between atmosphere and ocean and the question of an increase of atmospheric CO2 during the past decades. Tellus, 9(1), 18-27.   131  Rigby, A., & Frash, M. (1997). Ab initio calculations on the mechanisms of hydrocarbon conversion in zeolites: Skeletal isomerisation and olefin chemisorption. Journal of Molecular Catalysis A: Chemical, 126(1), 61-72.  Ripoche, A., & Guillard, A. (2001). Determination of fatty acid composition of pork fat by fourier transform infrared spectroscopy. Meat Science, 58(3), 299-304.  Ruston, N. (1952). Commercial uses of fatty acids. Journal of the American Oil Chemists' Society, 29(11), 495-498.  Santana, R. C., Po, P. T., Santikunaporn, M., Alvarez, W. E., Taylor, J. D., Sughrue, E. L., & Resasco, D. E. (2006). Evaluation of different reaction strategies for the improvement of cetane number in diesel fuels. Fuel, 85, 643-656.  Scatterfield, C. N. (1981) Mass Transfer in Heterogeneous Catalysis. (Robert E Krieger Publishing Company Inc. New York. Schmidt, K., & Van Gerpen, J. (1996). The effect of biodiesel fuel composition on diesel combustion and emissions. SAE Paper, 961086.  Schreiner, P. R., Schleyer, P. R., & Schaefer III, H. F. (1993). Mechanisms of electrophilic substitutions of aliphatic hydrocarbons: Methane nitrosonium cation. Journal of the American Chemical Society, 115(21), 9659-9666.  Scott, S. A., Davey, M. P., Dennis, J. S., Horst, I., Howe, C. J., Lea-Smith, D. J., & Smith, A. G. (2010). Biodiesel from algae: Challenges and prospects. Current Opinion in Biotechnology, 21(3), 277-286.  Sinfelt, J.H., Hurwitz, H., & Rohrer, J. C. (1962). Role of dehyrogenation activity in the catalytic isomerization and dehyrocyclization of hydrocarbons. Journal of Catalysis, 12, 481-492. Smith, P. C., Ngothai, Y., Dzuy Nguyen, Q., & O'Neill, B. K. (2010). Improving the low-temperature properties of biodiesel: Methods and consequences. Renewable Energy, 35(6), 1145-1151.  Smith, P. C., Ngothai, Y., Nguyen, Q. D., & O'Neill, B. K. (2009). Alkoxylation of biodiesel and its impact on low-temperature properties. Fuel, 88(4), 605-612.  Soualah, A., Lemberton, J., Pinard, L., Chater, M., Magnoux, P., & Moljord, K. (2008). Hydroisomerization of long-chain n-alkanes on bifunctional Pt/zeolite catalysts: Effect of the zeolite structure on the product selectivity and on the reaction mechanism. Applied Catalysis A: General, 336(1-2), 23-28.   132  Souza, M. S., Aguieiras, E. C. G., da Silva, M. A. P., & Langone, M. A. P. (2009). Biodiesel synthesis via esterification of feedstock with high content of free fatty acids. Applied Biochemistry and Biotechnology, 154(1), 74-88.  StatsCan. (2012). Global availability of agriculture and arable land, 2005. Retrieved 03/15, 2013, from http://www.statscan.gc.ca/pub/15-202-x/2009000/t230-eng.htm  Tamino word press, "death by heat wave". (2012). Retrieved 01/15, 2013, from http://tamino.wordpress.com/2012/11/19/death-by-heat-wave/#more-5984  Tanabe, K. (1989). New solid acids and bases: Their catalytic properties Elsevier Science Limited.  Thompson, J., & He, B. (2006). Characterization of crude glycerol from biodiesel production from multiple feedstocks. Applied Engineering in Agriculture, 22(2), 261.  Tolvanen, P., M?ki-Arvela, P., Kumar, N., Er?nen, K., Sj?holm, R., Hemming, J., & Murzin, D. Y. (2007). Thermal and catalytic oligomerisation of fatty acids. Applied Catalysis A: General, 330, 1-11.  Tyson K. S. (2001) Biodiesel: Handling and Use Guidelines, Golden CO, National Renewable Energy Laboratory. www.biofuels.coop/archive/HandlingAndUse2001.pdf (accessed Dec. 10, 2009). USDA advisory committee on biotechnology & 21st century agriculture (AC21). (2012). Retrieved February, 15, 2013, from http://www.usda.gov/wps/portal/usda/usdahome?contentid=AC21Main.xml&contentidonly=true  Van de Voort, F., Ismail, A., Sedman, J., & Emo, G. (1994). Monitoring the oxidation of edible oils by fourier transform infrared spectroscopy. Journal of the American Oil Chemists' Society, 71(3), 243-253.  Veldsink, J. W., Bouma, M. J., Sch??n, N. H., & Beenackers, A. A. C. M. (1997). Heterogeneous hydrogenation of vegetable oils: A literature review. Catalysis Reviews, 39(3), 253-318.  Villegas, J., Kumar, N., Heikkil?, T., Lehto, V. P., Salmi, T., & Murzin, D. Y. (2006). Isomerization of< i> n</i>-butane to isobutane over pt-modified beta and ZSM-5 zeolite catalysts: Catalyst deactivation and regeneration. Chemical Engineering Journal, 120(1), 83-89.  Wang, P. S., Thompson, J., & Van Gerpen, J. (2011). Minimizing the cost of biodiesel blends for specified cloud points. Journal of the American Oil Chemists' Society, 88(4), 563-572.   133  West, A. H., Posarac, D., & Ellis, N. (2008). Assesment of four biodiesel production processes using HYSYS plant. Bioresourc Technol, 99, 6587-6601. Wilde, J. P. (1990). Enthalpy and entropy of organic compounds from group contributions. SFCC Publications, 58 (Memorandum, LR M-635)  Yori, J. C., D'Amato, M. A., Grau, J. M., Pieck, C. L., & Vera, C. R. (2006). Depression of the cloud point of biodiesel by reaction over solid acids. Energy Fuels, 20(6), 2721-2726.  Zaera, F., & Somorjai, G. (1984). Hydrogenation of ethylene over platinum (111) single-crystal surfaces. Journal of the American Chemical Society, 106(8), 2288-2293.  Zhang, Y., Dube, M. A., McLean, D. D., & Kates, M. (2003). Biodiesel production from waste cooking oils: Process design and technological assesment. Bioresource Technol, 89(1), 1-16. Zhang, Z. C., Dery, M., Zhang, S., & Steichen, D. (2004). New process for the production of branched-chain fatty acids. Journal of Surfactants and Detergents, 7(3), 211-215. Zhang S., & Zhang Z.C. (2007) Skeletal isomerization of unsaturated fatty acids: the role of mesopores in HBeta zeolites. Catalysis Letters, 115:114-121              134  Appendix A: Optimization Study on Isomerization of Oleic Acid   An optimization study was conducted to examine the effects of the co-catalyst on the cloud point and branch chain conversion of the various reaction products. There is evidence in the literature that supports the idea of using water or methanol to improve the conversion of the isomerization reaction by facilitating the transfer of the protons. Since beta zeolite is an acid catalyst it will catalyze the esterification/de-esterification reactions. Therefore, this study uses water for the oleic acid experiments and methanol for the methyl oleate experiments in order to prevent these reactions from taking place. The use of a factorial design allows for the user to get the maximum amount of data from a reasonable amount of experiments. A secondary purpose to the optimizations study is to find an optimal set of reaction conditions. The conditions studied were temperature, pressure and wt% co-catalyst based on Table A.1. The responses that were optimized were the branch chain conversion (BC) and cloud point (CP) of methyl oleate (MO) and esterified oleic acid (OA), as summarized in Table A.2. The optimal predicted conditions were experimentally verified to check the accuracy of the model as shown in Table A.3. The results are organized to show the optimal conditions of each of the responses separately, and in various combinations. The optimal response for the combination of all four factors is 200?C, 1.25 MPa and 0 wt% co-catalyst.   Table A.1 Experimental design for testing different isomerization reaction conditions Run # Temperature (?C) Pressure (MPa) H2O or Methanol (wt%) C1 23 0 0 C2 250 1.5 0 A.1 200 0.1 0 A.2 225 0.1 2 A.3 250 0.1 1 A.4 275 0.1 0 A.5 300 0.1 2 A.6 200 1.5 1 A.7 225 1.5 0 A.8 250 1.5 2 A.9 275 1.5 2 A.10 300 1.5 0 A.11 200 3 2  135  A.12 225 3 1 A.13 250 3 0 A.14 275 3 1 A.15 300 3 1 C1 ? control #1 C2 ? control #2  The two highest predicted values for branch chain conversion and cloud point were 53% and -17.5?C, respectively.  The highest branch chain conversion was for when oleic acid was isomerized; while, the lowest cloud point was when methyl oleate was isomerized. Therefore two contour plots of isomerized methyl oleate and oleic acid branch chain conversion, Figures A.1 and A.2, respectively, were created to show the combined effect of the reaction conditions. From these plots it is clear that the optimal conditions do reflect the conditions predicted by the optimization study.  Figure A.1 Contour plot of two-way interaction of temperature (?C) and pressure (MPa) on cloud point (?C) for isomerization reactions carried out on methyl oleate  136   Figure A.2 Contour plot of two-way interaction of temperature (?C) and pressure (MPa) on branch chain conversion (wt%) using oleic acid as a starting material  Table A.2 Response optimization of reaction conditions based on cloud point (?C) and branch chain conversion (wt%) # Factors Optimized Optimal Reaction Conditions Response  (predicted) Response (experimental) %Difference *   Temperature (?C) Pressure (MPa) wt% co-catalyst    1 CP (MO) 200 0.1 2 -17.5 -15.8 9.7 2 CP (OA) 200 0.1 0 -15.2 -15.5 -2.0 3 BC (MO) 200 3 0 39 36 7.6 4 BC (OA) 300 1.5 0 53 50 5.7 5 CP (MO) & 200 3 0 -16.7 -15.4 7.8  BC (MO)    39 35 10.3 6 CP (OA) & 200 0.1 0 -15.2 -15.5 -2.0  BC (OA)    36 40 -11.1 7 CP (MO), 200 1.25 0 -15.2 -14.9 2.0  137   CP (OA),    -11.6 -12.2 -5.2  BC (MO)&    25 31 -24.0  BC (OA)    32% 40% -25.0 * % difference of predicted versus experimental results CP ? cloud point BC ? branch chain product MO ? methyl oleate starting material OA ? oleic acid starting material esterified post reaction                      138  Appendix B: Sample Calculations  Rate Equation re-arrangement  ??? = ???? ??????= ???? ? ???????????? = ? ???0?? ?1(1 ? ?)??(1??) = (? ? 1)??????? ??(1??) ? ???(1??) = (1 ? ?)?? ??(1??) = (? ? 1)?? + ???(1??) Steps: ? Re-arrange basic rate equation to a form that can be integrated with respect to concentration of species a. ? Once the equation is integrated re-arrange to isolate for concentration of species a. Finding the reaction rate constant for multiple temperatures  Find k at T = 473, 523 and 558 K through ???0.43 ? 3.32?0.43 = 0.43?? ? =  ???0.43 ? 0.59690.43? k473 = 0.00213 k558 = 0.027 k523 = 0.0052  139  ? Using the integrated re-arranged rate equation input the reaction order of 1.43 and solve for k at the various temperatures examined.  Plot ln k versus 1/T  T (K) 1/T (1/K) k ((mol/L)^0.43)/hr) ln k 473 0.00211 0.00213 -6.1516 523 0.00191 0.00520 -5.2590 558 0.00179 0.02700 -3.6119  Slope of plot = (E/R) = -7582 E = 63037 J/mol  (?) Effectiveness Factor ? Start out by finding the diffusion coefficient of the catalyst. Diffusion: ???? = 19400 ?2????????? T = 533 K   tm = 1 M = 282 g/mol  Sa = 680000 cm2/g ? = 0.43   ?? = 0.95 g/cm3 Deff = 7.63 * 10-4 cm2/s Modulus ? The modulus is used to find the effectiveness factor from standard curves generated from experimental data.  ? = ?2????(?1??????)1?? R = 0.01 cm  140  Deff = 7.63 * 10-4 cm2/s Cs = 0.003014 mol/cm2 ? = 72.14 ? The observed rate of reaction is taken from experimental data in the thesis and is re-arranged to fit the proper units. ?1?? ???? -ra = 0.116(3.014)1.92       = 0.9647 mol/Ls      = (dn/dt) ?1??= ?11.252.15 ?1?? ???? = 1.659 mol/g cm3 cm2  The effectiveness factor is then taken from Scatterfield using the calculated modulus ? = 0.09            141  Appendix C: Isomerization of Palmitic Acid Added to Oleic acid    Figure C.1 Effect of the isomerization of palmitic acid on a mixture of fatty acids               142  Appendix D: Calibration Curves for FTIR Quantification of Different Fatty Acids  Figure D.1 Calibration curve for determining branched chain fatty acid composition using FTIR, methyl oleate as straight chain esters and methyl-16-methyloctadecanoate as branched chain esters   Figure D.2 Calibration curve for determining trans bond ester composition using FTIR, methyl oleate cis bond ester and methyl elaidate as trans bond ester  143     Figure D.3 Calibration curve for determining free fatty acid content of oil using FTIR, methyl oleate as fatty acid and canola oil as vegetable oil.               144  Appendix E: Study to Find Optimal Hydrolysis Method  Several methods were studied to find a successful hydrolysis method and one the one that worked was the near supercritical water method. Ultimately the optimal method was 230?C, 4.0 MPa N2 pressure, 2 h with no catalyst.  Table E.1 Hydrolysis Method Testing Temperature (?C) Pressure (MPa) Catalyst Time (h) Conversion (wt%) 100 0.0 H2SO4 1 0 100 0.0 H2SO4 2 2 100 0.0 H2SO4 4 5 100 0.0 H2SO4 6 16 150 1.0 N/A 4 2 170 2.0 N/A 4 21 190 2.5 N/A 4 47 210 3.0 N/A 4 74 230 4.0 N/A 4 98 250 5.0 N/A 4 98 270 6.0 N/A 4 98 230 4.0 N/A 1 87 230 4.0 N/A 2 98 230 4.0 N/A 6 98            

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0103370/manifest

Comment

Related Items