...‘.{.§
‘1: s...
: x

91:
V a: 7
:11)!
z: 1.12.... ‘1.
. 11%....557
t‘ 1.: It

‘ .2 i...
. .3...‘ a:
'1. a

v -. .\ ‘
.. tax»-.. .. Y...
. v .\1-‘ v .0
‘ _ i .1
vii... A 5.
39V 2

r

 

. .4... .ufl.!~§~hmr«uvrfliiz [v «-
‘ . (xv r J EH;
.3 as. .mivu
an,

_.. 5...?
3*:
. 2. fisfi
32.5.}... .o i.»
a .7. ii}?! y
.L3; {3.4
as...

.9» ‘ ‘
. 3%.; ‘ .. . .
behave zinmmuy . .

la.— .2 . 3.
. V J
. nub—ma“. a r#5v
this.
1:15.)
its"... .

$96 L36:

:5. .1 .
«354v

.. ...:
van?!
‘ $.-

max...

I345 . a . 3,...
:2...me

,r.

V.
iv

Qua.“
..

 

 

LIBRARY
, Michigan State
10 0 5 . University

This is to certify that the
dissertation entitled

ENGINEERED SOY OILS FOR NEW VALUE ADDED
APPLICATIONS

presented by

PHUONG T. TRAN

has been accepted towards fulfillment
of the requirements for the

PhD. degree in CHEMICAL ENGINEERING

 

CW W [\I (UK/(1% 0“”

\ Major Professor’s Signature I W
May 12‘“, 2005

Date

MSU is an Affinnative Action/Equal Opportunity Institution

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

 

DATE DUE

DATE DUE

DATE DUE

 

JUN 2 .1 2007

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 p:/CIRC/DateDue.indd-p.1

ENGINEERED SOY OILS FOR NEW VALUE ADDED APPLICATIONS

By

Phuong T. Tran

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTORATE IN PHILOSPOHY
Department of Chemical Engineering and Material Science

2005

ABSTRACT
ENGINEERED SOY OILS FOR NEW VALUE ADDED APPLICATIONS

By
Phuong T. Tran

Soybean oil is an abundant annually renewable resource. It is composed of
triglycerides with long chain saturated and unsaturated fatty acids. The presence of
unsaturated fatty acids allows for chemical modification to introduce new functionalities
to soybean oil. A portfolio of chemically modified soy oil with suitable functional
groups has been designed and engineered to serve as the starting material in applications
such as polyamides, polyesters, polyurethanes, composites, and lubricants. Anhydride,
hydroxyl, and silicone functionalities were introduced to soy oil. Anhydride functionality
was introduced using a single-step free radical initiated process, and the chemically
modified soy oils were evaluated for potential applications as a composite and lubricant.
Hydroxyl functionalities were introduced in a single-step catalytic ozonolysis process
recently developed in our labs, which proceeds rapidly and efficiently at room
temperature without solvent. The transformed soy oil was used to successfully prepare
bio-lubricants with good thermal/oxidative stability and bio-plastics such as polyamides,
polyesters, and polyurethanes. A new class of organic-inorganic hybrid materials was
prepared by curing vinyltrimethoxysilane functionalized soy oil. This hybrid material
could have potential as biobased sealant through a moisture initiated room temperature
cure. These new classes of soy-based materials are competitive both in cost and

performance to petroleum based materials, but offer the advantage of being biobased.

ACKNOWLEDGEMENTS

I would like to first thank my advisor, Dr. Ramani Narayan, for the opportunity to work
on such an interesting project both in the lab and in the business world. For giving me
guidance with numerous aspects of the project, I would like to thank Dr. Daniel Graiver
and Ken Farminer. For help in the lab, I would like to thank Dr. Yogaraj Nabar, Dr. Jean-
Marie Raquez, Dr. Sunder Balakarishnan and Madhu Srinivasan. For their valuable
comments and suggestions during the course of this work, I would like to thank my
guidance committee members. I would also like to thank all of my friends in the
department for helping me in the many ways that you did, especially Laura, Christina,
Madhu, Sharad, Guoren, Steve, and Ty. And finally I would like to thank my family for

all of the support they have provided.

iii

TABLE OF CONTENTS

 

 

 

 

 

 

List of Tables vii
List of Figures - viii
List of Abbreviations xiii
Chapter 1: Introduction to Soybean Oil Based Feedstocks 1
1.1 BACKGROUND AND RATIONALE .................................................................. 1
1.1.1 Agricultural Based Feedstock ............................................................................. 1
1.1.2 Carbon Cycle ...................................................................................................... 1
1.1.3 Soy Oil as a Feedstock ........................................................................................ 2
1.1.4 Soy Oil Statistics ................................................................................................. 2
1.2 PROBLEM DESCRIPTION .................................................................................. 5
1.2.1 Variability in Soy Oil .......................................................................................... 5
1.2.2 Fatty Acid Distribution in Soy Oil ...................................................................... 6
1.3 SPECIFIC GOAL .................................................................................................. 6
1.4 RESEARCH OBJECTIVE .................................................................................... 8
1.5 ORGANIZATION OF THESIS .................................... .. ...................................... 9
Chapter 2: Maleic Anhydride Modified Soy Oil 10
2.1 LITERATURE REVIEW .................................................................................... 10
2.1.1 Research on Maleic Anhydride Modified Soy Oil ........................................... 10
2.2 EXPERIMENTAL PROCEDURES .................................................................... 15
2.2.1 Materials and Methods ...................................................................................... 15
2.2.2 Analysis Equipment .......................................................................................... 15
2.3 RESULTS AND DISCUSSION .......................................................................... 16
2.3.1 Investigation of Processing Conditions ............................................................ 16
2.3.2 Diels-Alder Addition to Maleic Anhydride to Soy Oil ..................................... 18
2.3.3 Characterization of Maleated Soy Oil ............................................................... 19
2.4 CONCLUSIONS .................................................................................................. 24
2.5 REFERENCES .................................................................................................... 25
Chapter 3: Biocomposites from Modified Soy-Oil and Biofibers 26
3.1 LITERATURE REVIEW .................................................................................... 26
3.1.1 Research in Soy-based Polymers ...................................................................... 26
3.2 EXPERIMENTAL PROCEDURES .................................................................... 29
3.2.1 Reagents and Equipment ................................................................................... 29
3.2.2 Synthesis Methods ............................................................................................ 29
3.3 RESULTS AND DISCUSSION .......................................................................... 33
3.3.1 Characterization of Soy-based Thermosets ...................................................... 33
3.3.2 Mechanical Properties of Soy-based Composites ............................................. 35
3.3.3 Silane Surface Treatment of Biofibers .............................................................. 37
3.4 CONCLUSIONS .................................................................................................. 39
3.5 REFERENCES .................................................................................................... 40

iv

Chapter 4: Ozone Mediated Polyol Synthesis _- 42

 

 

4.1 INTRODUCTION ............................................................................................... 42
4.1.1 Properties of Ozone ........................................................................................... 42
4.1.2 Ozone Generation ............................................................................................. 43
4.1.3 Ozone Chemistry .............................................................................................. 44
4.2 LITERATURE REVIEW .................................................................................... 47
4.2.1 Methods of Polyol Preparation from Vegetable Oils ........................................ 47
4.3 EXPERIMENTAL PROCEDURES .................................................................... 49
4.3.1 Materials and Methods ................................................ 49
4.3.2 Catalytic Ozonolysis ........................................... . ............................................. 49
4.3.3 Characterization Procedures ............................................................................. 50
4.4 RESULTS AND DISCUSSION .......................................................................... 51
4.4.1 Reactions with Ozone ....................................................................................... 51
4.4.2 Composition of Ozone Treated Soy Oil ............................................................ 53
4.4.3 Reactions with Heterogeneous Catalysis .......................................................... 59
4.5 CONCLUSIONS .................................................................................................. 63
4.6 REFERENCES .................................................................................................... 64
Chapter 5: Continuous Process for the Catalytic Ozonolysis of Soy Oil ............ 65
5.1 INTRODUCTION ............................................................................................... 65
5.1.1 Soy Polyol Production ...................................................................................... 65
5.2 REACTOR DESIGN THEORY .......................................................................... 67
5.2.1 Batch and Continuous Reactors ........................................................................ 67
5.3 REACTOR DESIGN METHODOLOGY ........................................................... 69
5.3.1 Comparison of Reactors .................................................................................... 69
5.4 RESULTS AND DISCUSSIONS ........................................................................ 71
5.4.1 Processing Parameter Considerations ............................................................... 71
5.5 CONTINUOUS REACTOR DESIGN ................................................................ 77
5.5.1 Reactor Selection and Sizing ............................................................................ 77
5.5.2 Process Scale—up ............................................................................................... 79
5.5.3 Process Economics ............................................................................................ 82
Chapter 6: Thermal-Oxidatively Stable Biobased Lubricants 83
6.1 LITERATURE REVIEW .................................................................................... 83
6.1.1 Soy Oil Based Lubricants ................................................................................. 83
6.2 EXPERIMENTAL PROCEDURES .................................................................... 85
6.2.1 Materials and Methods ...................................................................................... 85
6.2.2 Synthesis Techniques ........................................................................................ 85
6.2.3 Characterization Procedure ............................................................................... 87
6.3 RESULTS AND DISCUSSION .......................................................................... 90
6.3.1 Transformation of Double Bonds via Maleation .............................................. 90
6.3.2 Transformation of Double Bonds via Catalytic Ozonolysis ............................. 92
6.3.3 Evaluation as a Biobased Lubricant .................................................................. 96
6.4 CONCLUSIONS ................................................................................................ 101
6.5 REFERENCES .................................................................................................. 102

Chapter 7: Silane Functionalized Soybean Oil - 103

 

 

7.1 LITERATURE REVIEW .................................................................................. 103
7.1.1 Silane Addition to Soy Oil .............................................................................. 103
7.2 EXPERIMENTAL PROCEDURES .................................................................. 104
7.2.1 Materials and Methods

7.2.2 Synthesis Techniques ...................................................................................... 104
7.2.3 Characterization Methods ............................................................................... 105
7.3 RESULTS AND DISCUSSION ........................................................................ 106
7.3.1 Evaluation of Processing Parameters ................................................... _ ........... 1 06
7.3.2 Characterization of Silane Functionalized Soy Oil ......................................... 110
7.3.3 Room Temperature Moisture Activated Cure Studies .................................... 115
7.3.4 Swelling of Soybased Thermosets .................................................................. 116
7.4 CONCLUSIONS ................................................................................................ 121
7.5 REFERENCES .................................................................................................. 122
Chapter 8: Conclusions and Recommendations 123
8.1 MALEIC ANHYDRIDE FUNCTIONALIZED SOY OIL .............................. 123
8.2 BIOFIBER REINFORCED SOYBEAN COMPOSITES ................................ 123
8.3 SYNTHESIS OF SOY POLYOLS VIA OZONE CHEMISTRY .................... 124
8.4 EVOLUTION SOY POLYOL PRODUCTION ............................................... 125
8.5 SOY OIL BASE OIL LUBRICANT PROPERTIES EVALUATED .............. 127
8.6 SILANE FUNCTIONALIZED SOY OIL ........................................................ 127
8.7 RECOMMENDATIONS .................................................................................. 128
Appendix A: Verification with a Model Compound .............................................. 129
Appendix B: Ozonolysis in an Alkaline Median .................................................... 136
Appendix C: Scale-up Process Economics for Polyol Production ......................... 139
Appendix D: NMR of Silane Treated Soy Oil ........................................................ 141

vi

LIST OF TABLES

Table 1.1: Fatty Acid Profile of Natural Oils ............................................................... 3
Table 3.1: Batch Curing Experiments in the Presence of a Base Catalyst ................... 30
Table 3.2: Batch Curing Experiments in the Presence of a Diamine ............................ 30
Table 3.3: Soy-based Biofiber Reinforced Natural Composite .................................... 31
Table 4.1: Statistical Distribution of Soy Polyol .......................................................... 52
Table 5.1: Hydroxyl and Acid Value of Soy Polyols ................................................... 74
Table 5.2: Apparent Residence Time ............................................................................ 79
Table 6.1: Calculated Molecular Weight Distribution .................................................. 93
Table 6.2: Molecular Weight Distribution for Reactions with Methanol ..................... 93
Table 7.1: Solubility Parameter for Various Solvents .................................................. 117

vii

LIST OF FIGURES

Figure 1.1: Natural Oilseed Production Worldwide in 2002 ........................................ 3
Figure 1.2: Annual Production of Soybean Oilseeds .................................................... 4
Figure 1.3: Annual Oil Consumption of Natural Oils ................................................... 5
Figure 1.4: Major Fatty Acid Components of Soybean Oil .......................................... 6
Figure 1.5: Theoretical Fatty Acid Distribution of Soy Oil ........................ ................. 7
Figure 2.1: 1-4 Diels-Alder Adduct .............................................................................. 10
Figure 2.2: Peroxide Initiated Conjugation of Double Bonds ...................................... 11
Figure 2.3: 1,2 Diels-Alder Addition of MA to Oleic Acid .......................................... 12
Figure 2.4: Radical Initiated Addition of MA .............................................................. 12
Figure 2.5: Addition of a Second Mole of MA by a 1,4 Diels-Alder Adduct ............... 13
Figure 2.6: Addition of a second mole of MA by a 1,2 Diels-Alder Adduct ................ 13
Figure 2.7: Crosslinking of SO with MA ....................................................................... 14
Figure 2.8: Effect of Reaction Time on Acid Value at 150°C using different MA/SO

Ratio A02; I 0.6; and O 1.1 ..................................................................... 16
Figure 2.9: Effect of Peroxide Concentration on Acid Value after 60 minutes

at 150°C ANo Initiator; I 0.005 wt%; 9 0.01 wt% ................................. 17
Figure 2.10: Change in Viscosity with Increases in MA Concentration ...................... 19

Figure 2.11: Effect of Reaction Temperature on Maleation of Soy Oil

A 100°C I 120°C 0 150°C ..................................................................... 20
Figure 2.12: Temperature Profile during Maleation

(A) 0 wt%, (B) 0.005 wt% and (C) 0.01 wt% L101 ................................. 22
Figure 2.13: FT IR of Soybean Oil and Maleated Soybean Oil ..................................... 23
Figure 3.1: Preparation of Soy-based Composites by Resin Transfer Molding ........... 32
Figure 3.2: FTIR Spectra of Epoxidized and Maleated Functional Soy Oils ............... 34

viii

Figure 3.3:

Figure 3.4:

Figure 4.1:
Figure 4.2:
Figure 4.3:
Figure 4.4:
Figure 4.5:
Figure 4.6:
Figure 4.7:
Figure 4.8:
Figure 4.9:
Figure 4.10:
Figure 4.11:
Figure 4.12:

Figure 4.13:

Figure 4.14:

Figure 4.15:
Figure 4.16:
Figure 5.1:
Figure 5.2:
Figure 5.3:

Figure 5.4:

Tensile Strength of Composites Prepared by Compression Molding ........ 36
Tensile Strength of Soy Composites after a Silane Surface Treatment of
Kenaf Fibers ............................................................................................... 38
Generation of Ozone .................................................................................. 42
Resonance Structure of Ozone ................................................................... 42
Ozone Generation by Corona Discharge .................................................... 43
Formation of Primary Ozonide ................................................................... 45
Ozonolysis Reaction Mechanism ............................................................... 45
Ozonolysis Products of Methyl Oleate in Methanolic NaOH .................... 46
Ozonolysis Reaction with a Model Compound .......................................... 51
Catalytic Ozonolysis of Alkene with Ethylene Glycol .............................. 53
Catalytic Ozonolysis Products of Soy Oil with Ethylene Glycol ............... 54
Statistical Distribution of Soy Polyols ..................................................... 55
Average Functionality .............................................................................. 55
FTIR of Soy Oil (A) and Soy Polyol (B) ................................................. 57

Hydroxyl Number (---) and Double Bonds (—) as a Function of
Ozonolysis Time ....................................................................................... 58

Hydroxylation of Soy Oil using CaCO3 at Different Ozonolysis Times:
(A) t=0; (B) t=0.25; (C) t=0.5; (D) t=1.0; and (E) t=2.0 min/g of soy oil 60
130mm of Soy 011 ................................................................................. 61
13C-NMR of Soy Polyol ........................................................................... 62
Competing Reactions of Ozone .................................................................. 66
General Mole Balance ................................................................................ 67
Batch Reactor Mole Balance ...................................................................... 67
Batch Reactor Design Equation ................................................................. 67

ix

Figure 5.5: Continuous Stirred Tank Reactor Design Equation ................................... 67

Figure 5.6: Plug Flow Reactor Design Equation .......................................................... 68
Figure 5.7: Ozonolysis Mole Balance ........................................................................... 70
Figure 5.8: Rate of Reactant Disappearance ................................................................. 70
Figure 5.9: Rate of Product Formation ......................................................................... 70
Figure 5.10: FT IR of Soy Polyol using Different Frit Locations ................................. 71
Figure 5.11: FTIR of Soy Polyol for Catalyst Loadings of 1 wt% & 10 wt% .............. 73
Figure 5.12: FTIR of Soy Polyols using Different Particle Sizes of CaCO3 ................ 74
Figure 5.13: Variation of Catalyst Particle Size using a Model Compound ................. 75
Figure 5.14: Sensitivity Analysis of Soy Polyol Cost .................................................. 76
Figure 5.15: Reactor Design Selection ......................................................................... 77
Figure 5.16: Continuous Production of Soy Polyol ...................................................... 78
Figure 5.17: Process Flow Diagram for Continuous Polyol Production ...................... 79
Figure 5.18: Polyol Production Downstream from Soy Oil Extraction ........................ 80
Figure 5.19: Ozone Injected to a Mixture of Soy Oil and Ethylene Glycol ................. 81
Figure 5.20: Ozone Injected into Ethylene Glycol ....................................................... 81
Figure 5.21: Ozone Injected by Gas Sparging .............................................................. 82
Figure 6.1: Soy Oil Maleation and Esterification Apparatus ........................................ 86
Figure 6.2: FTIR of Soybean Oil (A); Maleated Soybean Oil (B); and Esterified
Maleated Soybean Oil (C) .......................................................................... 90
Figure 6.3: Esterification of Maleated Soybean Oil with an Alcohol ........................... 91
Figure 6.4: Catalytic Ozonolysis of Soy Oil in the Presence of an Alcohol ................. 92

Figure 6.5: Rate of Decomposition of Ozone Treated MEG350 and Soy Oil
0 min reaction (A); 10 min reaction (B); and 30 min reaction (C) ............. 94

Figure 6.6:

Weight Loss of Ozone Treated MEG350 and Soy Oil

0 min reaction (A); 10 min reaction (B); and 30 min reaction (C) ............ 95
Figure 6.7: FT IR of Soy Oil (A) and Ozonized Soy Oil with Methoxy-PEG350 (B).. 96
Figure 6.8: Deposits Formation of Lubricant Sample .................................................. 97
Figure 6.9: Cold Flow Properties of Lubricants ........................................................... 98
Figure 6.10: Failure Load for Soy-based Lubricants .................................................... 99
Figure 6.11: Wear Plot of Esterified-Maleated Soy Oi] ............................................... 99
Figure 6.12: Wear Plot of Ozone Modified Soy Oil ..................................................... 100
Figure 7.1: Diels-Alder and Ene Addition Product ...................................................... 107
Figure 7.2: Addition of VTMS to S0 at 150C with Time

(A) no catalyst; (B) L101; (C) BF3-methanol ............................................ 108
Figure 7.3: Addition of VTMS to SO with No Catalyst for 30 minutes

with Increasing Temperature ...................................................................... 109
Figure 7.4: Double Bonds Remaining Following the Reactions of VTMS with S0

at 150°C for 120 minutes (and other model compounds) .......................... 110
Figure 7.5: FT IR of SO and VTMS Functionalized SO ............................................... 111
Figure 7.6: FTIR of LA (A); LA/VTMS (B); and LAIVTMS/LlOl (C) ...................... 112
Figure 7.7: Room Temperature Cure of VTMS Functionalized Soy Oil ..................... 116
Figure 7.8: Degree of Swelling of the Soy Therrnoset with Different Solvents ........... 118
Figure 7.9: Thermal-Oxidative of Soy Oil and RT Cured Silated Soy Oil ................... 120
Figure A1: Ozonolysis of Methyl Soyate in Triethlamine and CH3OH ....................... 130
Figure A2: GC-MS of Methyl Soyate ........................................................................... 131
Figure A3: GC-MS of Ozonolysis Products ................................................................. 132
Figure A4: Mass Spectra of C9 Ozonolysis Products ................................................... 133
Figure A5: Mass Spectra of C9 Base Fragment ............................................................ 134

xi

Figure A5: Mass Spectra of C6 Fragment ..................................................................... 135

Figure B1: Formation of Primary Ozonide ................................................................... 136
Figure B2: Ozonolysis Reaction Mechanism ............................................................... 136
Figure B3: Possible Reactions of Carbonyl Oxides ...................................................... 137
Figure B4: Ozonolysis of Methyl Oleate ...................................................................... 138
Figure B5: Ozonolysis of Methyl Oleate in the Presence of Methanolic N aOH .......... 138
Figure D.1: lH — NMR Spectra of Soybean Oil ............................................................. 141
Figure D2: ‘H — NMR of Silylated Soybean 011 .......................................................... 142
Figure D3: 13(2 — NMR Spectra of Silylated Soybean 011 ............................................ 143
Figure D4: HMBC Correlation Spectrum of Silated Soy Oil ....................................... 144

xii

4-DMAP

MSO

OA
PFR
RT
RTD
RTM
SO
SSO
TAG

VTMS

LIST OF ABBREVIATIONS

4—Dimethylamino Pyridine

Boron Trifluon'de

Corona Discharge

Deteurarted Chloroform
Continuous Stirred Tank Reactor
Epoxidized Soybean Oil

Fourier Transform Infrared Spectroscopy
Hexamethylene Diamine
2,5-BiS(tert-butylperoxy)-2,5-dimethylhexane
Linoleic Acid

Maleic Anhydride

Maleated Methyl Soyate

Number Average Molecular Weight
Maleated Soybean Oil

Weight Average Molecular Weight
Nuclear Magnetic Resonance

Oleic Acid

Plug Flow Reactor

Room Temperature

Residence Time Distribution

Resin Transfer Molding

Soybean Oil

Silated Soybean Oil

Triacyl glycerol

Ultraviolet

Vinyltrimethoxysilane

xiii

Chapter 1: Introduction to Soybean Oil Based Feedstock

1.1 BACKGROUND AND RATIONALE
1.1.1 Agricultural Based Feedstock

A growing awareness of the environment has led to the emergence of
biodegradability and/or the usage of renewable resources as being an important design
criterion. Issues such as sustainability, industrial ecology, biodegradability, and
recyclability are becoming major considerations in a company’s product. Agricultural
based chemicals are one of the most attractive alternatives because they are a sustainable
source of development.
1.1.2 Carbon Cycle

As the demand for petroleum exceeds production, alternative feedstock will have
to be found. In contrast to petroleum based materials, agricultural based materials are
not depleted and are readily regenerated by photosynthesis. Materials based on annually
renewable agricultural feedstock can form the basis for a portfolio of sustainable, eco-
efficient products that is an environmentally preferable, sustainable alternative to current
materials based exclusively on petroleum feedstock. The use of annually renewable
biomass, as the feedstock for the production of polymers, chemicals, and fuels/lubricants
can be rationalized from a global carbon cycle perspective. Carbon is present in the
atmosphere as C02. Plants fix this carbon by photosynthesis using sunlight as the energy
source. Over geological time frames (>106 years) these plant materials are fossilized to
provide petroleum/natural gas. These fossil resources are used to make polymers,
chemicals and fuels/lubricants which release carbon back into the atmosphere as C02 in a

Short time frame of 1-10 years. The C02 emissions problem is a kinetic rate issue. The

rate at which biomass is converted to fossil resources is not in balance with the rate at
which they are consumed and liberated (>106 years vs. 1-10 years). Alternatively, the use
of annually renewable crops as feedstock maintains/balances the global carbon cycle as
the rate at which C02 is fixed is equal to or greater than (if more biomass is planted than
harvested) the rate at which it is consumed and liberated. Thus, the use of annually
renewable crop feedstock to produce the polymers, chemicals, and fuels/lubricants as an
adjunct to fossil resources would begin to move the rate of C02 release more in balance

with the rate at which C02 is fixed.

Thus, the use of renewable crop/biomass feedstock allows for:

0 Sustainable development of carbon based polymer materials

0 Control and even reduce CO2 emissions

1.1.3 Soy Oil as a Feedstock

One potential agriculture feedstock is soybean oil (SO). Soybean oil is an abundant
renewable resource composed of a mixture of triglycerides consisting of glycerin and
saturated and unsaturated long chain fatty acids. Some typical natural oil fatty acid

compositions are presented in Table 1.1.

 

 

 

 

 

 

 

 

 

 

 

 

 

Fatty Acid Canola Corn Linseed Sunflower Soybean
Myristic - C14:0 0.04 - 0.03 0.06 0.1
Palmitic - Cl6:0 4.1 10.8 5.5 5.6 11
Palmitioleic - C16:1 0.21 0.11 0.06 0.1 0.1
Margaric - C17:0 0.1 0.09 0.06 0.06 -
Stearic - Cl8:0 1.9 1.9 1.9 4.5 4
Oleic - C1821 60.9 28.2 20.2 23.2 23.4
Linoleic - Cl8:2 21.4 57.5 16.3 64.9 51.4
Linolenic - C1822 8.5 0.7 54.1 0.5 7.8
Arachidic - C20:0 0.6 0.3 0.12 0.3 0.3
Eicosatrienoic - C20:3 0.1 - - - -
Behenic - C22:0 0.2 0.03 0.41 0.4 0.1

 

 

 

 

 

 

 

Table 1.1: Fatty Acid Profile of Natural Oils

1.1.3 Soy Oil Statistics

The annual production of soybean worldwide reached over 400 billion pounds in

2002 (Figure 1.1).

450

 

Billion Pounds

 

 

 

Soybean Com Canola Peanut Sunflower Palm Kernel Capra
(Rapeseed)

Figure 1.1: Natural Oilseed Production Worldwide in 2002

3

0f the annual production of soybean oilseeds, the United States is the world’s largest
producer as seen in Figure 1.2. Nearly 1/2 of the world’s soybean oilseed production is
by the United States.

The annual production of soybean oilseed in the United States is over 170 billion
pounds. Soybean oilseeds contain about 30-35% oil by mass, amounting to about 120
billion pounds of soybean oil annually produced globally. 0f the nearly 120 billions

pounds, only 63 billion pounds are currently being consumed as presented in Figure 1.3.

 

200

150 b

100*

Billion Pounds

SO~

 

 

           

United States Brazil Argentina China India Paraguay

Figure 1.2: Annual Production of Soybean Oilseeds
The net result is a surplus of over 57 billion pounds of soybean oil. The most
common usage of soybean oil is for edible applications such as salad/cooking oils,
baking/frying fats, and margarine. These applications make up nearly 98%. Currently,

only 2% of the total soybean oil usage is for application such as resins/plastics,

paints/varnishes, and fuels/lubricants. The development of new value-added applications

of soy oil is beneficial both environmentally and economically.

70

 

Billion Pounds

 

 

 

Soybean Palm Rapeseed Sunflower Peanut Cotton Coemut Palm Ker-rial Olive

Figure 1.3: Annual Oil Consumption of Natural Oils

1.2 PROBLEM DESCRIPTION
1.2.1 Variability among Soy Oil

Soybean oil is a complex mixture of primarily fatty acids, which can vary from
region-to-region and from season-to-season. Typically, soybean oil contains 4.3—4.6
“cis” double bonds per mole depending on the origin of the seed (Figure 1.4). To
develop new value added applications for soy oil, chemical modification is used to
incorporate reactive groups in the soy oil molecules. The presence of unsaturation allows
for modifications through specific reactions with these double bonds and the introduction

of reactive functional groups.

 

 

 

 

 

 

 

 

 

/\/\/\/\/\/\/\
0
ll Palmitic acid (16:0) -- 11%
— o R /\/\/\/\/\/\/\/V
o Stearic acid (18:0) -- 4%
/W\/\=/\/W\
-0 LR R: Oleic acid (18:1, 9C) -- 23%
0 WW
II Linoleic acid (18:2; 9C, 12C) -- 51%
" 0 R /\=/\_—_/\=/\/\/\/\
Linolenic acid (18:3 9C,12C,15C) -- 7%
Carbons Common Name - IUPAC Name Double Bonds
16 Palmitic Acid hexadecanoic acid 0
18 Stearic Acid octadecanoic acid 0
18 Oleic cis -9-octadecenoic acid 1
18 Linoleic Acid cis,cis -9,12-octadecadienoic acid 2
18 Linoleneic Acid cis,cis,cis -9,12,15-octadecatrienoic acid 3

 

 

 

 

 

Figure 1.4: Major Fatty Acid Components of Soybean Oil
1.2.2 F atly Acid Distribution of Soy Oil

The Statistical distribution of fatty acids within soy oil suggests about 71% of all
triacylglycerol (TAG) Species contain at least one linoleic acid chain (Figure 1.5).
Furthermore, 25% contain only 1 linoleic acid, while 32% and 14% contain 2 and 3
linoleic acid chains, respectively. Additionally, it can be seen that 29% of the TAG
species contain only 1 polyunsaturated fatty acid (linoleic or linoleneic).

The successful application of soy oil as an alternative to petroleum based products
depends on the cost competitiveness and end-use performance. Additionally, any
proposed chemical modification must be robust enough to account for the inherent
variability among soy oil.

1.3 SPECIFIC GOAL

The goal of this research is to design and engineer a portfolio of chemically modified

soy oil with reactive groups that can serve as the starting material for applications such as

polyesters, polyamides, polyurethanes, composites, and lubricants.

6

 

SSS
121.212
881.2
SSP

””3 S = Stearic Acid (C13:0)
PPP . o o o
850 ' P = Palnntrc Acid (C13:0)
rrs % O = Oleic Acid (C13:1)
121-2" L1 = Linoleic Acid (C13:2)
PS” ‘ L2 = Linolenic Acid (c183)

PPLZ
SSLl
L2L20
$1.20

81.10
LlLlS
LILZO
LlL1L2
PLlO
LlLlP
00L]
LlLlLl
LlLlO

l

0.00 0.05 0.10 0.15
Frequency

Figure 1.5: Theoretical Fatty Acid Distribution of Soy Oil

 

0.20

1.4 RESEARCH OBJECTIVE:

1. Preparation of maleic anhydride functionalized soy oil
A. Optimizing of processing parameters (time, temperature, catalyst type, and
catalyst concentration)
1. Structural determination of maleated soy oil
B. Prepare a soy-based resin matrix by reacting maleated soy oil with a
commercially available epoxidized soy oil
C. Create a structurally strong thermoset by increasing the cross-link
densities by using a spacer molecule
D. Curing in the presence of biofibers to create an all natural composite
1. Optimization of process parameters
2. Enhancement of mechanical properties by a Silane surface
treatment of the biofibers
E. Develop soy-based lubricants
1. Esterification with a series of alcohols to enhance stability
2. Determine base oil lubricant properties

II. Develop a process to synthesize primary soy-based polyols using ozone chemistry
A. Develop a novel single step catalytic ozonolysis to prepare polyols
1. Effects of catalysts type and concentration
2. Effects of processing parameters (time and temperature)
3. Structural determination using model compound
B. Prepare rigid polyurethane foams using soy polyols
C. Develop soy polyols for flexible foam applications
1. Variation of molecular architecture of soy polyol
a. Synthesize polyols with increased molecular weight
b. Prepare polyols with varying hydroxyl content
D. Prepare thermal-oxidatively stable lubricants
1. Variation of molecular architecture by using different alcohols
2. Determine base oil lubricant properties
E. Establish proof-of-concept for the continuous production of soy polyols

HI. Synthesize a Silane functionalized soy oil
A. Develop a fast efficient process to functionally modify soy oil with
vinyltrimethoxysilane
1. Optimizing processing parameters (time, temperature, and
catalyst type, and catalyst concentration)
2. Structural determination of Silated soy oil
B. Prepare a soy-based thermoset by a moisture activated curing of the Silane
functionalized soy oil at room temperature
1. Investigate effects of cure time and catalyst type on the sol-gel
properties of the thermoset

1.5 ORGANIZATION OF THESIS

The thesis is divided into three main parts. The first part, Chapter 2, discusses the
development of a new process to synthesize maleic anhydride functionalized soy oil, and
subsequent chapters discuss applications of the maleated soy oil to prepare all-natural
biofiber reinforced soybean oil based composites (Chapter 3) and as a base oil stock for a
lubricant (Chapter 6).

The second part, Chapters 4 through 6, details the development of a new ozone
mediated synthesis of soy polyols, which were further engineered to have varying
hydroxyl functionalities and molecular weight for potential use in polyesters, polyamides,
and polyurethanes. Chapter 5 discusses the development of a continuous production of
the soy polyols, and Chapter 6 evaluates the lubricating properties of the ozone treated
soy oil.

The third part, Chapter 7, discusses the synthesis of Silane functionalized soy oil.
Soy based therrnosets were prepared using a room temperature moisture activated cure of
this organic-inorganic hybrid.

In Chapter 8, the overall conclusions and recommendations are presented.

Chapter 2: Maleic Anhydride Modified Soy Oil

2.1 LITERATURE REVIEW:
2.1.1 Research on Maleic Anhydride Modified Soy Oil

There is a great deal of literature on functionalization of vegetable oils using
maleic anhydride (MA) to enhance their reactivity. Morrel and Samuels (1) studied the
reaction between maleic anhydride and conjugated oil systems using oils from china
wood and oiticica containing fatty acids with conjugated double bonds. The reaction was

found to follow a typical Diels-Alder addition as seen in Figure 2.1.

F 7
/ A /
II——-—- QM -——--

Figure 2.1: 1-4 Diels-Alder Adduct

 

 

Root (2) investigated the functionalization of soybean oil with maleic anhydride
and described the use of benzoyl peroxide in the maleation reaction. He concluded that
the use of benzoyl peroxide as a catalyst was advantageous as it allowed the reaction to
proceed at a lower reaction temperature of 110°C as compared 160-l90°C when benzoyl
peroxide was not used. Root hypothesized that the presence of a peroxide leads to
electron delocalization, resulting in conjugation of the double bonds within the
unsaturated fatty acids in the soybean oil to form a conjugated system as shown in Figure

2.2. However, very little data was provided to support this claim.

10

H
R0. ROH
0
CH'
_O/“\/\/\/\/\/ W013
ROH R0.

Figure 2.2: Peroxide Initiated Conjugation of Double Bonds

11

However, the mechanism of MA addition to soy oil and the products of this
reaction are still ill defined and are the subject of a great deal of debate in the literature.

Clocker (5) postulated a cyclobutane Structure for oleic acid adducts (Figure 2.3).

O O O

 

 

 

 

CH3

Figure 2.3: 1,2 Diels-Alder Addition of MA to Oleic Acid
Bickford and coworkers (6) studied the reaction between maleic anhydride with

methyl oleate and suggested that MA was added at the eighth or eleventh carbon atom

(Figure 2.4).
O
H30--0 \ CH3
0 O 0
MA Addition to the Eighth Carbon
0
H3C—0 \ CH3

MA Addition to the Eleventh Carbon

Figure 2.4: Radical Initiated Addition of MA

12

The results of Teeter (3) and Kappelmeier (4) also support succinic type Structures
on the basis of model compounds studies using methyl linoleate and methyl linolenate.
Accordingly, the reactions with di-unsaturated, nonconjugated double bonds proceeded
by the ‘ene’ reaction mechanism and the succinic type addition occurs at the eighth,
eleventh, and fourteenth carbon atoms, with a major addition product at the eleventh
carbon. Based on these results, Kappelmeier and Van Den Nuet (5) postulated that
conjugation might occur simultaneously with an attack at the methylene carbon by which

a second mole of maleic anhydride can be added by a Diels-Alder addition (Figure 2.5).

o 000

ego—0W
.41.

0

CH3

 

Figure 2.5: Addition of a Second Mole of MA by a 1,4 Diels-Alder Adduct

Somewhat similarly, Plimmer (6) suggested a second mole of maleic anhydride

could be added to the hexene ring (Figure 2.6).

o 000

H3C_O
CH3

 

 

 

 

Figure 2.6: Addition of a second mole of MA by a 1,2 Diels-Alder Adduct

l3

Nagakura and Yoshitomi (7), however, suggested the presence of free radicals
leads to copolymerization of maleic anhydride with the triglyceride and based their

conclusion on the increase in the viscosity of the reaction product (Figure 2.7).

0
O
O
0

Figure 2.7 : Crosslinking of SO with MA

It is apparent that although maleation of soybean oil has been reported previously, most
of the work was focused on the mechanism of this reaction and almost no attention was
given to process optimization of this free radical initiated reaction. It is therefore our goal
in this work, is to asses the process and determines the usefulness of this reaction.
Uniquely, the study will focus on maleation within a closed system under elevated
temperatures and pressures. The closed system synthesis prevents premature
sublirnination of the MA, which was a recurring problem in the past and caused lower

yields.

14

2.2 EXPERIMENTAL PROCEDURES:
2.2.] Materials and Methods

Unhydrogenated soybean oil was purchased from Spectrum Chemicals (Gardena,
CA). Maleic anhydride, octanol, isopropanol, 2,5-biS(tert-butylperoxy)-2,5-
dimethylhexane (L101) and di-tert-butyl peroxide were purchased from Sigma-Aldrich
(Milwaukee, WI).

Reactions were carried out by charging a 2L Parr Reactor equipped with a
motorized stirrer, thermocouple, and heating element with MA, SO, and the initiator.
The initial reaction time was taken at the time the reactor temperature reached a preset
temperature. At the end of each reaction, the reactor was connected to a vacuum pump
for 30 minutes and volatile products were collected in a solvent trap immersed in an
isopropanol dry ice bath. Acid numbers were determined according to ASTM D1980 (8)
and iodine number was measured by ASTM D1959 (9).

2.2.2 Analysis Equipment
FTIR analyses were run on a Perkin-Elmer model 1000 with a NaCl window. Viscosity

measurements were done using a Brookfield Viscometer model EX100 at 25°C.

15

2.3 RESULTS AND DISCUSSION:
2.3.1 Investigation of Processing Conditions

The effects of reaction time, initiator concentration, and MA concentration were
studied in details using two different initiators, L101 and di-tert-butyl peroxide. The
degree of maleation was determined using acid value, and the products were further
characterized using FTIR spectroscopy and iodine number.

The maleation of soybean oil in a closed system under high temperatures and
pressures in the presence of peroxides proceeded relatively fast. In all cases the reaction
proceeded to completion within ~30 minutes independent of the MA concentrations in

the reaction mixture (MA/SO molar ratio of 0.2, 0.6, and 1.1) as shown in Figure 2.8.

 

 

 

 

 

 

 

 

 

1.5 —--
~ Di-tert -Butyl
1.0 1 § § 1
0.5 P I 5 fl
- z a 5
o _
3 0.0
a: I
> i
'5 F
.8 +-
< » L101
E § 1
I 5 i
I _ r
- i I

 

 

 

30 40 50 60 70 80 90

 

Reaction Time (min)

Figure 2.8: Effect of Reaction Time on Acid Value at 150°C using different MA/SO
Ratio A0.2; I 0.6; and o 1.1

16

It is further interesting to note that the type of peroxide had little effect on the progress of
the reaction and essentially the same reaction rates were observed for both L101 and di-
tert—butyl initiators. However, the presence of the initiator in the reaction mixture greatly
impacted the progress of the reaction (Figure 2.9) and led to acid value of 1.0 compared

to acid value of 0.25 when no peroxide initiator was used.

 

 

 

 

 

 

 

 

 

1.2 :
10 Di-tert-Butyl
' g f % i
0.8 ;
0.6 1 i
0.4
o .
3 0.2 ; g L g I I
a _
> >—
12 0.0 :
in” 1.0 5- UO‘ 5 F i fl
0.8 E
0.6 E
0.4 E
0.2 3 5‘: l i E 4
0.0 i i 1 i 4 i 1 1 M 1 l 1 I 1 1 4 1‘1 1 1
0.0 1.0 2.0 3.0 4.0

Molar Ratio (MA/SO)

Figure 2.9: Effect of Peroxide Concentration on Acid Value after 60 minutes
at 150°C ANo Initiator; I 0.005 wt%; 0 0.01 wt%
Yet, doubling of the initiator concentration from 0.005 wt% to 0.01 wt% had no effect on
the acid values, which were statistically similar. Furthermore, when the reaction was run
with no peroxide, the resulting iodine number indicates the presence of about 4.5 moles

double bonds per triglyceride, meaning no net change in the number of double bonds

l7

after the addition of MA. This same number of double bonds suggests that reactions at
lower temperatures favor the “ene” addition over Diels Alder addition.
2.3.2 Diels-Alder Addition of Maleic Anhydride to Say Oil

It is apparent from Figures 2.2 and 2.3 that even in the presence of a peroxide
and large excess MA (up to molar ration of MAISO=4), the maximum acid value never
exceeded ~l. Experimentally, it was observed that when excess MA was used (e.g.
reactions containing MA/SO molar ratio of 2:1, 3: 1, and 4:1), unreacted MA remained in
the reaction mixture and then was accumulated and collected in the solvent trap during
the work-up procedure. Attempts to run the reaction for longer times (up to 240 minutes)
did not lead to higher acid values and no further increase in the maleation was observed.
The iodine number of the products with acid value around 1.0 was always correlated to
soy oil having about 3.5 double bonds. This net reduction of 1 double bond from the
unmodified soy oil suggests that 1 mole of MA was added to SO by a Diels-Alder
addition (Figure 2.1). Although, further maleation is theoretically possible, it was found
experimentally that only 1 mole of MA had reacted as determined by the acid number.
Although the reason for this upper limit addition of MA is not clear at this time, one
possible explanation could be steric hindrance effects that impede further reaction of MA.
2.3.3 Characterization of Maleated Soy Oil
Maleation of soy oil is accompanied by a noticeable increase in the room temperature

viscosity from 50 cPs of neat SO to 3500 cPs for the maleated SO (Figure 2.10).

18

 

 

   
  
  
   

    

 

 

 

4000 .
E
3500 — i
C OT=100C
3000 : IT=120C
AT=150C
E 2500 C
3 .
Q 2000 E
U)
c _
a :
1000 E
500 E
i t
O 1 1 1 1 l 1 1 1 1 1 g 1
0.0 0.2 0.4 0.6 0.8 1.0 1.2

Molar Ratio (MA/SO)

Figure 2.10: Change in Viscosity with Increases in MA Concentration
The increased viscosity is most likely due to an increase in hydrogen bonding as MA is
grafted into the triglyceride structure. It is not due to opening of the anhydride ring and
copolymerization through reaction of the carboxylic groups as has been suggested
previously since esterification of the maleated S0 with methanol yielded a product with
an acid value of 0.1 and a viscosity of only 100 cPS.

There were no Si gnificant differences observed when the reaction temperature was
adjusted to 100°C, 120°C, or 150°C. Under these conditions, MA was completely
consumed within 30 minutes as Shown in Figure 2.11. The degree of maleation was not
affected by the reaction temperature and similar acid values were obtained when the

initial temperature was set between 100°C to 150°C.

19

Acid Value

 

1.2
Di-tert -Buty1 L101

1

0.8 ~

i} H

0.4 -

 

 

 

00 lJllIIILllllLl 1111111¥11_LJLII
.

 

50 100 150 50 100 150
Temperature (C)

Figure 2.11: Effect of Reaction Temperature on Maleation of Soy Oil
A 100°C I 120°C 0 150°C

20

However, the temperature profile of the reaction was affected by the peroxide
concentration as Shown in Figure 2.12. Three distinct zones were clearly observed in the
course of these reactions; Zone 1 represents the time it took to heat the reactor to the
preset reaction temperature (e.g. 150°C in Figure 2.12), Zone 2 is the time interval
(usually, 30 minutes) the reaction was allowed to proceed after it reached the preset
temperature and Zone 3 is the time period heating was Stopped and vacuum was applied
to remove any unreacted MA. The typical temperature profile of reactions with no
initiator is shown in Figure 2.12A where the reaction proceeded to a set point of 150°C,
which was sustained for 30 minutes until heating was terminated and vacuum was
applied, resulting in Slow cooling. Zone 1 is the same for all the different reactions and is
determined by the applied heating. However, the temperature profile in Zone 2 is
different when peroxide was included in the reaction mixture (Figure 2.12B and 2.12C).
Here, the temperature continued to rise at the end of Zone 1 and although it was set to
150°C, it reached 200°C. This overheating beyond the preset temperature was
particularly noticeable in the reactions with 0.01 wt% peroxide and was accompanied by
an increase in the reactor pressure (25 psi). Although this exotherm may be partially
attributed to the decomposition of the peroxide, we believe it is most likely due to the

maleation reaction itself.

21

Temperature (C)

 

Zone 1 Zone 2 Zone 3

 

 

 

200

100

 

 

b-----------

1

 

 

0 20 40 60
Time (min)

Figure 2.12: Temperature Profile during Maleation
(A) 0 wt%, (B) 0.005 wt% and (C) 0.01 wt% L101

22

80

The maleated soybean oil was further characterized by FT IR spectroscopy
(Figure 2.13) and exhibits peaks at ~1775 cm'1 and 1850 cm'1 not present in 80. These
absorption peaks are related to the symmetric and asymmetric stretching of C=O in the
MA (succinic anhydride derivative), respectively (1 l, 12). Furthermore, no absorption
peaks related to —OH stretching due to the opening of the anhydride is observed, clearly

indicating that the anhydride ring remained intact.

 

Soybean Oil

 

Absorbance

Maleated
Soybean Oil

 

 

 

 

 

; I 1 1 1 1 4 1 1 1 1 l 1 1 4 L

 

3500 3000 2500 2000 1500 1000
Wavenumber (cm-1)

Figure 2.13: FTIR of Soybean Oil and Maleated Soybean Oil

23

2.4 CONCLUSIONS

Maleation of SO in a closed system at elevated temperatures and pressures is
advantageous over open system processes. The maleated soy oil product we obtained is a
pale yellow in color in contrast other studies where a red-orange liquid was obtained (13).
Apparently, the use of high pressure in a close system is more advantageous and prevents
premature decomposition and degradation of the soy oil. Furthermore, this closed system
is also advantageous as it allows for a fast and simple one-step process. Unlike previous
studies that required a reaction time of up to 7 hours for the maleation, with the MA
being added to soy oil in aliquots (14), we have found that the reaction is completed
within 30 minutes. Optimum process conditions for maleated SO in a closed system
synthesis appear to be 30 minutes reaction time at 150°C in the presence of 0.005 wt %
peroxide. The use of a pressurized closed system prevents loss of MA due to sublimation
during the reaction. It was further found that the concentration of MA in the reaction
mixture had the most pronounced effect on the extent of the reaction while the acid
values were independent of the concentration and type of peroxide used. However, even
when excess MA was present, the acid value did not increase above ~1.0 mole. The data
appear to indicate that maleation reaction in the absence of a peroxide proceeds primarily
through “ene chemistry” whereas a Diels Alder addition is favored when a peroxide is
present. Thus, conjugation of the double bonds in the linoleic and linolenic fatty acid

residues is expected to occur only in the presence of a catalyst and at high temperatures.

24

2.5 REFERENCES

1.

10.

11.

12.

13.

14.

Morrel RS, and Samuels H., “Doubly conjugated system in alpha and beta-
eleostearic acids”, J. Chem. Soc. Abstracts 2251-2254 (1932)

Root F.B., Improvement of drying oils, US. Patent 2374381 (1945).

Clocker E.T., Condensation products for use in enamels, etc., US. Patent #2188882
(1940)

Bickford W.G., Fisher G.S., Kyame L., and Swift C.E., Autoxidation of fats. II.
Preparation and oxidation of methyl oleate-maleic anhydride adduct, J. Amer. Oil

Chem. Soc. 25, 254-257 (1948).

Teeter H.M., Geerts M.J., and Cowan J.C., Polymerization of drying oils. 111. Some

observations on reaction of maleic anhydride with methyl oleate and methyl linoleate,
J. Amer. Oil Chem. Soc. 25, 158-162 (1948)

Kappelmeier C.P.A., Van Der Nuet J.H., Van Goor W.R., Maleic-treated oils, Paint
Oil Chem. Rev. 113, 11-18 (1950)

Plimmer H. and Robinson E.B., Drying or semidrying oils and their derivatives,
British Patent 565432 (1944)

Nagakura M. and Yoshitomi K., Reaction of maleic anhydride with unsaturated fatty
acids in the presence of organic peroxides, Yukagaku 21, 83-91 (1972)

ASTM D1980, Standard Test Method for Acid Value of Fatty Acids and Polymerized
Fatty Acids, Annual Book of ASTM Standards 06.03: 418-419 (2000)

ASTM D1959, Standard Test Method for Iodine Value of Drying Oils and Fatty
Acids, Anmgil Book of ASTM Standards 06.03: 399—401 (2000)

Bellamy L.J., The InerRed Spectrg of Complex Molecules, Wiley: New York,
(1964).

Pouchert 0]., The Aldrich Library of InfraL-Red Spwtrg, Aldrich Chemical Company;
Milwaukee, WI, (1970).

Tarik E., Kuesefoglu S.H., Wool R., Polymerization of maleic anhydride-modified
plant oils with polyols, J. Appl. Polymer Science 90, 197-202 (2003)

Warth H., Mulhaupt R., Hoffmann B., Lawrence S., Polyester network based upon
epoxidized and maleinated natural oil, Angewandte 249, 79-92 (1997)

25

Chapter 3: Biocomposites from Modified Soy Oil and Biofibers

3.1 LITERATURE REVIEW
3.1.1 Research in Soy-based Polymers

Both thermoplastic and thermoset matrix biocomposites have been described in the
literature; the resin matrix is generally a biopolymer or a synthetic biodegradable polymer
and the reinforcing component is usually some biofibers (e.g. jute, flax, hemp, ramie,
sisal, coir, cotton, etc). It Should also be mentioned here that the use of natural fibers
instead of traditional reinforcing materials such as glass fibers, talc, and mica provides
further advantages in terms of lower cost, lower density, reduced tool wear, acceptable
strength properties, energy recovery, and recyclability.

Currently, the most common chemical modification of soybean oil is the
introduction of epoxy groups that replace some of the double bonds in the fatty esters.
This epoxidation process was commercialized recently and the modified oil is being used
primarily as a process aid in polyvinyl chloride to improve its stability, flexibility and
processability. It is also being used in the preparation of adhesives for carpet backing, as
a source of polyols for the preparation of rigid polyurethanes and as a reactive modifier
or diluent of epoxy resin system (1). The reaction of maleic anhydride with soybean oil
has also been reported in the literature and these chemically modified oils are used
primarily in cosmetic applications as a component in shower gels, bar soaps, sunscreens,
and skin treatment products (2).

Several researchers have investigated the use of chemically modified soybean oil
in composites. However, it was found that the mechanical properties, primarily the tensile

strength and stiffness, were too low, and thus limited their use in the market place. For

26

example, an early work with epoxidized soybean oil (ESO) produced flexible, semi-
flexible, and rigid crosslinked polyesters (3, 4). Williams and Wool (5) prepared soy
composites by free radical polymerization of acrylated and epoxidized soybean oil, which
was then cured with styrene and divinylbenzene by resin transfer molding in the presence
of hemp and flax. These composites were reported to have mechanical properties
comparable to commercially available synthetic resins. However, the use of petroleum-
based polymers was essential to enhance the mechanical properties. Lu et al. (6) reported
the preparation of clay nanocomposites based on functionalized triglycerides such as
acrylated epoxidized soybean oil, maleated acrylated epoxidized soybean oil, and
soybean oil pentaerythritol maleates combined with styrene. Another type of composite
was prepared by Taveeme-Valduizen et al. (7) from soybean oil by thermoplastic
processing in the presence of a binding agent such as urea-formaldehyde copolymer or a
protein that was added to the soybean oil. Can et al. (8, 9) focused on the preparation of
copolymers from soy oil monoglycerides with Styrene and soybean oil monoglyceride
maleates with styrene. Zhu et al. (10) reported that epoxidized methyl soyate and
epoxidized allyl soyate yielded resins that were highly reactive and provided a high
degree of intermolecular crosslinking due to the addition of allyl soyate. These resins had
higher modulus than Similar resins obtained from unmodified E80 and thus, it was
suggested that they could be suitable as composites for structural applications. Warth et
al. (11) reported on the preparation of natural fibers reinforced soybean oil in polyester
resin. Their work was based on networks of epoxidized and maleated soybean oil with
hemp and flax fibers (10 wt %), where they were able to achieve tensile strengths of 6.0

and 6.2 MPa, respectively.

27

The goal of this work was to synthesize tough composites from natural biofibers
and chemically modified soybean oil with no additional styrene, polyester or other
petroleum-based polymers to supplement the soy matrix in order to enhance the
mechanical properties. The soy-based composites were prepared by combining
epoxidized soybean oil (ESO) with maleated soybean oil (MSO), or maleated methyl
soyate (MMS), in the presence of selected biofibers. In contrast to previous studies, this
work was focused on increasing the efficiency of the crosslink density, minimizing the
sol fraction of the network in order to enhance the stiffness of the composites and
eliminate possible “bleeding out” of the sol fraction from the gel network due to
incomplete cure in order to avoid any adverse effects on the surface properties of the
composites. It was thought that one common reason for a low gel fraction and incomplete
cure in previous studies of epoxidized soy-based composites was related to incomplete
cure between the epoxidized soy modified oil and the anhydride groups primarily due to
the poor reactivity of the epoxy groups and steric hindrances. Thus, in this study, we
explored the use of flexible spacer crosslinking amine catalysts (e.g. hexamethylene
diamine (HMD) and 4-dimethylamino pyridine (4-DMAP)) that were expected to react

with both epoxy and anhydride functional groups.

28

3.2 EXPERIMENTAL PROCEDURE
3.2.] Reagents and Equipment

The soybean oil (Spectrum), 1/3 partially epoxidized soybean oil (Elf Atochem),
and methyl soyate (Michigan Soybean Board) were used without further purification.
Methylene chloride, maleic anhydride, isopropanol, hexamethylene diamine (HMD), 4-
dimethylamino pyridine, and 2,5-biS(tert-butylperoxy)-2,S-dimethylhexane (L101) were
purchased from Sigma-Aldrich. Biofibers were obtained from American Fillers &
Abrasives, Inc. and (2-aminoethyl)-3-aminopropyl-trimethoxysilane was purchased from
Gelest.

FTIR spectra were run on a Perkin-Elmer Model 1000 by evaporating a solution
of a sample from the surface of a NaCl crystal. The composites were cured in Wabash
Metal Products Model CM02871. Mechanical pr0perties were run after aging the samples
for 3 days at 23°C and 50% relative humidity on a tensile instrument made by United
Testing Systems Model SFM-20.

3.2.2 Synthesis Methods
Maleation of Soy Oils

As reported in Chapter 2 (Page 15)
Cure Reactions of M80 (or MMS) with ESO

Approximately equimolar M80 (or MMS) and ESO were combined with a
catalyst in a beaker and placed in a silicone oil bath heated at 5°C/min while stirring
using a magnetic stirrer. The gels were then removed and extracted in a soxhlet

apparatus for 3 days using methylene chloride as a solvent and the gel fraction was

29

determined. The results of the batch cure experiments are presented in Table 3.1 and

3.2.

M80 (mol) MMS (mol) ESO (mol) 4—DMAP (wt%) T (°C ) Gel Time (8) Gel Wt (%)

0 200 - -
0.05 172 68 0.99
0. l 170 40 0.98
0. 15 162 37 0.97
0.2 148 0.98
0.2 room -
0 200 -
0.05
0. l 164
0. 15
0.2
0.2

Gelation could have occurred over a of time or

0
0
0
0
0
0
l
l
l
l
l

cccco——-—-—-——-
~———---~_

O

 

Table 3.1: Batch Curing Experiments in the Presence of a Base Catalyst

MSO (mol) MMS (mol) ESO (mol) HMD (wt%) T (°C ) Gel Time (s) Gel Wt (%)

0 200 - -
0.05 162 52 0.98
0. 1 158 35 0.99
0.15 152 27 0.99
0.2 155 25 0.99
0.2 room - -
0 200 - -
0.05 164 61
0. 1 162 58 0.99
0.15 159 46
0.2 148 33 0.99
0.2 room - -

 

u—tu—on—n—n—nu—Icooooo

~fl-—_fl—_—~_

l
l
l
l
l
l
0
0
0
0
0
0

Table 3.2: Batch Curing Experiments in the Presence of a Diamine
From these results, further cure studies were done in an oven at 150°C. Each mixture
was visually checked every few minutes. Initially, the color of the reaction mixture was
light yellow and the catalyst was insoluble in the oil. However, after about 15 minutes at
this temperature, a significant drop in the viscosity was noticed and all the catalyst
appeared to have dissolved in the matrix. Thereafter, the viscosity of the reaction mixture
continuously increased and the color of the mixture turned slightly darker. The reaction

was allowed to continue for 3 hours and the gel was then removed from the oven,

30

extracted in a soxhlet apparatus for 3 days using methylene chloride as a solvent and the
gel fraction was determined. Based on these experimental conditions, biocomposites

containing various fibers were prepared as shown in Table 3.3.

 

 

 

Reactants Resin Matrix Characteristics Biofiber
ESO + MSO Viscous Liquid -
ESO + MSO + 4-DMAP Soft and Sticky Solid Kenaf, Kayocell, Protein Grits, & Solka-floc
ESO + HMD Transluscent Highly Viscous Liquid -
MSO + HMD Tranparent Highly Viscous Liquid -
ESO «1» M80 + HMD Semi-Rigid Solid Kenaf, Kayocell, Protein Grits, & Solka-floc
ESO + MMS Liquid -
ESO + MMS + 4-DMAP Soft Solid Kenaf, Kayocell, Protein Grits, & Solka-floc
MMS + HMD Tranparent Viscous Liquid -
ESO + MMS + HMD Semi-Rigid Solid Kenaf, Kayocell, Protein Grits, & Solka-floc

 

 

Table 3.3: Soybased Biofiber Reinforced Natural Composite
Resin Transfer Molding
A measured weight of dried kenaf fiber mats sufficient to make 10 wt%
composites was placed inside a rectangular mold. A flow media was placed over the top
of the fiber mat, and a vacuum bag was placed over the entire system as shown in Figure
3.1. The reactants and catalyst were mixed at 60°C for 30 minutes, then, added to the
fiber mat and vacuum was applied to ensure the fibers were completely saturated. The

cure reaction was allowed to proceed for 3 hours in a vacuum oven at 150°C.

31

VI bag

 
 
  

 

Flow

  

mat
Tape

   

Peel ply

  
 
 
  
 

Wire wrap

Vacuum pump
Top View

Resin

Figure 3.1: Preparation of Soy-based Composites by Resin Transfer Molding

Compression Molded Cures

The catalyst was combined with F30 and impregnated into the biofiber (10 wt %)
and the mixture was heated with stirring until the catalyst was completely dissolved.
Then, MSO was added, thoroughly mixed and the reaction mixture was poured into a
rectangular mold and held at 150°C under 11 tons of pressure for 3 hours. Samples were
removed from the press after it was cooled to RT by circulating cooling water through the
platens.
Surface Treatment of Kenaf Fibers

Kenaf fibers were treated by dipping them briefly in a 2 wt% freshly prepared
aqueous solution of (2-aminoethyl)-3-aminopropyl-trimethoxysilane. The fibers were

then placed in a vacuum oven at 80°C and allowed to dry for 4 hours.

32

 

3.3 RESULTS AND DISCUSSION
3.3.1 Characterization of Soy-based Thermosets

The crosslinking reaction between epoxy functional polymers and anhydrides to
yield a polyester resin is well known (12) and depends primarily on the reactivity of the
anhydride, its molar ratio with respect to epoxy groups, the concentration and efficiency
of the cure catalyst, and the reaction conditions. In our study, the resin network was
obtained by reacting epoxy functional soy oil (E80) and succinic anhydride functional
soy oil (MSO or MMS) catalyzed by HMD or 4-DMAP. The cure was easily followed by
the disappearance of the characteristic FTIR anhydride absorptions at 1780 and 1850 cm'1
(Figure 3.2). It was observed that the onset of gelation was directly proportional to the
catalyst concentration and always appeared at lower temperatures when higher
concentrations of the catalyst were used (Table 3.1). However, the nature of the gels was
greatly affected by the choice of catalyst and the type of soy derivative that was used as
Shown in Tables 3.1 and 3.2. Thus, HMD catalyzed reaction between MMS and ESO
produced a gel with very high gel fraction (low sol content), but the same reaction
catalyzed with 4-DMAP produced a poor gel having unacceptable high sol content.
Reactions between MSO and ESO for both catalyst systems yielded high gel fraction
networks with essentially no weight loss after soxhlet extraction. In contrast, 4-DMAP

catalyzed reaction between MMS and ESO yielded a poor network with high sol fraction.

33

 

 

 

Absorbance

A .1 M
M

 

 

 

 

3700 3200 2700 2200 1700 1200 700
Wavenumber [cm'l]

Figure 3.2: FTIR Spectra of M80 (A); MSO + ESO (B); and
M80 + ESO + HMD (C)

These observations are in agreement with the curing mechanism of epoxy resins.
It is generally agreed (13) that a non-catalyzed cure starts by the reaction of anhydride
with hydroxyl (which is usually found in ESO) to yield a monoester and a carboxylic
group. The carboxylic acid group then reacts with an epoxide to form a diester with a
hydroxyl group, which subsequently reacts with another anhydride. However, in the
presence of amine catalysts the cure reaction is initiated by complex formation of the
amine with a proton donor and the active centers are the carboxylate and hydroxylate
anions. The reaction mechanism in these cases depends to a great extent on the reactivity
of the anhydride and, more important, on the type of amine catalyst. Apparently, in our

case the cure catalysis with 4-DMAP was not as efficient compared with HMD catalysis.

34

The poor network with high sol fraction that was obtained in the case of
ESO/MMS was most likely due to the presence of the non reactive saturated palmitic and
Stearic components in the methyl soyate oil. Since maleation occurs across the double
bonds, these fatty acids, which are present in the soy oil at about 15 wt%, do not react
with maleic anhydride and thus, are not incorporated into the network. Hence, they
merely act as a plasticizer and are easily removed out of the resin by the soxhlet
extraction. The fact that reactions between MMS and E80 in the presence of HMD also
resulted in high gel fractions with nearly no weight loss after extraction is most likely due
to rearrangement and formation of amide linkages with the carbonyl esters of the methyl
soyate. Such ester saponification and amide formation allows for the saturated
components that carry no anhydride groups to be incorporated into the gel matrix through
amide linkages.

The cured soy-based resins, especially those cured with HMS, were very rigid and
brittle unlike “flexible and rubber-like” resins prepared previously from ESO with
succinic anhydride”. Apparently, the use of the succinic anhydride functional soy oil in
combination with a flexible amine catalyst led to a more efficient network formation
compared with Similar ESO matrices that were cured with low molecular weight
cycloaliphatic anhydrides and catalyzed by benzyldimethylanrine.

3.3.2 Mechanical Properties of Soy-based Composites

The mechanical properties were greatly enhanced when these soy-based resins
were reinforced with biofibers. Generally, the tensile strengths of such composites cured
with HMD as a catalyst had the best tensile strengths. The type of biofibers or the soy

matrix (ESO with M80 or with MMS) had only secondary impact (Figure 3.3). In

35

 

contrast, composites cured using 4-DMAP as a catalyst had relatively poor mechanical
properties as a result of lower resin crosslink density. In fact, it was observed that
composites made from MMS-ESO and cured with 4-DMAP had a significant sol fraction

that diffused out of the sample upon aging.

 

E Kenaf

C] Kayocell

I Soy Protein
Solka-floc

.—

0 .

Tensile Strenght (MPa)

 

 

 

ESO+MSO+4-DMAP ESO+MSO+HMD EO+MMS+4-DMAP ESO+MMS+HMD

Figure 3.3: Tensile Strength of Composites Prepared by Compression Molding

Preparation of biocomposites by incorporation of kenaf fibers provided the best
reinforcement due to good interactions between the soy matrix and these lignocellulosic
fibers (Figure 3.3). Although the tensile strength of composites cured with 4-DMAP
were generally lower than those cured with HMD, good tensile strength was observed in
the samples that were reinforced with soy proteins. It is possible that in this particular
case the presence of free amino groups in the protein grits contributed to the fiber/matrix

adhesion and consequently led to higher then expected tensile strength

36

3.3.3 Silane Surface Treatment of Biofibers

Further enhancement in the mechanical properties was realized by surface
treatment of the biofibers with (2-aminoethyl)-3-aminopropyl-trimethoxysilane coupling
agent prior to incorporating them into the matrix (Figure 3.4). This coupling agent has
been used extensively for surface treatment of various fibers (14). It is effective at low
concentrations whereby the amine functional group enhances the solubility of the Silane
in water and at the same time provides multiple hydrogen bonding through interactions
with hydroxyl groups on the surface of the fiber. After deposition, the amine groups
further promotes hydrolysis and condensation of the alkoxy groups on the Silane to yield
insoluble oligomeric polysiloxanes and a hydrophobic network on the surface of the fiber
(15). This surface treatment improved the adhesion between the resin and kenaf fibers
and led to about 20% improvement in the tensile strength compared with Similar
composite where kenaf fibers were dried but was not treated. The effect of surface
treatment of the fibers was somewhat less pronounced in composites that were prepared
by resin transfer molding simply due to the difference in the wetting of the fibers in this

process.

37

 

Tensile Strenght (MPa)

16

14*

12*-

10

 

 

I Kenaf
Silated Kenaf

     

 

 

 

 

 

 

   

// %

 

 

 

ESO+MSO+HMD ESO+MMS+HMD ESO+MSO+HMD ESO+MMS+HMD
Compression Molding Resin Transfer Molding

Figure 3.4: Tensile Strength of Soy Composites after a Silane Surface
Treatment of Kenaf Fibers

38

 

3.4 CONCLUSIONS

Soybean oils and methyl soyate chemically modified with suitable functional
groups (e.g. epoxy and anhydride) react in the presence of a flexible amine catalyst to
create a three-dimensional crosslinked polyester matrix. The formation of these networks
was followed by FTIR under different temperatures and reaction times. Biocomposites
were prepared by curing these soy-based resins in the presence of various biofibers (e.g.
flax, kayocell, protein grit, and solka-floc) using compression molding and resin transfer
molding techniques. Rigid network resins with high gel fractions were obtained when
ESO was reacted with MSO using HMD as a catalyst. Composites with high mechanical
strength were achieved when kenaf fibers were first treated with (2-aminoethyl)-3-

aminopropyl-trimethoxysilane in the reinforcement of these soy-based matrix resins.

39

 

3.5 REFERENCES

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

. Kamani, R.; Krishnan, M.; Narayan, R. Polym Eng and Sci 1997, 37, 476

Mustata F. J Polym Eng 1997, 17, 491

Gripp, A.; Steinberg, D. In-Cosmetics Exhibition and Conference Proceedings,
Barcelona, 1994, 293

Crivello, J .V.; Narayan, R. Chem Mat 1992, 4, 692

Crivello, J .V.; Narayan, R.; Stemstein, S.S. Chem Mat 1997, 64, 2073
Williams, G.; Wool, R.P. Appl Compos Ma_t 2000, 7, 421

Lu, J .; Hong, C.K.; Wool, R.P. J Polym Sci. Part B: Polym Phys 2004, 42, 1441
Taveme-Velduizen, W.; Simka, H.; Feil, H. Eur. Pat. Appl. 976790 (2000)
Can, E.; Kusefoglu, S.; Wool, R.P. J Appl Poly Sci 2001, 81, 69

Can, E.; Kusefoglu, 8.; W001, R.P. J Appl PolySci 2002, 83, 972

Zhu J .; Chandrashekhara, K.; Flanigan, V.; Kapila, S. J Appl Polym Sci 2004, 91,
3513

Warth, H.; Mulhaupt, R.; Hoffman, B.; Lawson, S. Ang Makomol Chem 1997, 249,
79

Root F.B. US. Patent 2374381 (1945)
Nagakura M.; Yoshitomi K. Yukagaku 1972, 21, 83

ASTM D1980, Standard Test Method for Acid Value of Fatty Acids and Polymerize
Fatty Acids, Annual Book of ASTM Standards 2000, 6, 418

ASTM D1959, Standard Test Method for Iodine Value of Drying Oils and Fatty
Acids, Annual Book of ASTM Standards 2000, 6, 399

Takahashi K.; Hitoshi F. Adv Polym Sci 1986, 80, 173
Trappe V.; Buchrd W.; Steinmann B. Macromolecules 1991, 24, 4738
Steinmann B. J Appl Polym Sci 1990, 39, 2005

Steinmann B. J Appl Polym Sci 1989, 37, 1753

40

 

21. Roesch J .; Muelhaupt R. Polym Bull 1993, 31, 679

22. Plueddemann E.P. In Silanes Surfaces and Interfaces, Leyden D.E., Ed.; Gordon and
Breach Sci. Pub.: New York, 1986, Chap. 2.

23. Ishida H.; Naviroj S.; Tripathy S.K.; Fizgerald J .J .; Koenig J .L. J Polym Sci. Part B:
Polym Phys 1982, 20, 701

41

 

Chapter 4: Ozone Mediated Polyol Synthesis

4.1 INTRODUCTION
4.1.1 Properties of Ozone

Ozone is a natural component in the earth’s atmosphere, where it is formed
photochemically. It can be recognized by the clean, fresh odor of air after a

thunderstorm, where lighting produces low levels of ozone (Figure 4.1).
3 02 + energy ——'—" 2 03

Figure 4.1: Generation of Ozone
The first odor of ozone was reported by Van Mauren in 1785, in the presence of as
electrical discharge. The identity of the structure of the compound was confirmed in

1867 as triatomic oxygen. Ozone is an allotrope of oxygen (Figure 4.2).

0

0+ 0+
0/ \O- <———-> O-/ \
Figure 4.2: Resonance Structure of Ozone

Ozone (03) is formed by recombination of atomic (O) and molecular oxygen (02).
It is the strongest oxidizing agent, able to be produced. A comparison of oxidation
strengths of standard oxidants, ozone is 2.07 V as compared to chlorine (1.36 V) and
chlorine dioxide (1.50 V). However, the molecules are unstable and cannot be stored for
future use; thus, it must be generated near point of application and then immediately
used. Ozone is relatively stable in air, with a half-life of several hours at low
concentrations. It has a limited solubility; therefore the concentration of ozone in the gas

stream becomes very important. Consequently, the solubility of ozone increases with an

42

increase in ozone concentration in the gas Stream. Also, the ozone solubility decreases
with an increase in temperature.
4.1.2 Ozone Generation

Ozone has been commercially used since 1893 when the first full scale drinking
water treatment application was implemented. Today, it is used in thousands of water
treatment applications, cooling towers, ultra pure water, marine aquaria, beverage
industries, swimming pools, bottled water plants, agriculture, food processing, and
chemical synthesis.

Ozone is generated by passing pure oxygen of dry air through an electric field,
where a portion of the oxygen is converted into ozone. There are several other methods
to produce ozone, such as corona discharge (CD), UV irradiation, and electrolysis. The
CD generation is capable of producing high concentrations of ozone. An energy potential
is used to split oxygen-oxygen double bonds, producing atomic oxygen. Collisions

between atomic and diatomic oxygen forms ozone (Figure 4.3)

Water Cooled Low Voltage Electrode

 

High Voltage Electrode

Figure 4.3: Ozone Generation by Corona Discharge

43

The quantity of ozone produced is dependent on several factors, such as voltage
and frequency of the alternating current applied to the CD cell. When enough high
energy electrons bombard gas molecules, ionization occur emitting gaseous plasma,
referred to as a corona. A typical CD ozone system consists of 4 fundamental
components, an air preparation or oxygen production unit, a CD generator, an ozone
diffuser/contactor, and an ozone off-gas destruction unit. Ozone generation by CD is an
exothermic physico-chemical reaction, where most of the energy used for ozone
generation is lost as heat, therefore cooling efficiency is an important factor. Recent
advances in CD have achieved greater ozone production with less electrode surface area,
reduced power consumption per kg of ozone produced, and the generator efficiency may
be manipulated by varying frequency, wave form, voltage, etc.

Air preparation is very important to efficient and reliable operation of CD
ozonators. Moisture and particulate matter have a detrimental effect on the generation
cell electrodes and dielectric material. Moist air in the ozone generator will form nitric
acid decreasing ozone production, and corroding components. Also, the presence of wet
air will cause arcing and sparks within the generator. Although, oxygen feed results in
higher ozone concentrations and reduced power consumption, for economic reasons most
applications use air as the feed gas.

4.1.3 Ozone Chemistry

Ozone was discovered more than 150 years ago, but the first ozonolysis reaction
to cleave carbon-carbon double bonds was only carried out in 1903 by Harris. Ozone is a
very powerful oxidant, which leaves only oxygen as a Side product after oxidation. It can

be generated on-site without “chemistry”. No other oxidizing agent is capable of

cleaving double bonds in such a fast, clean, and selective way to yield carbonyl
compounds. While studying the resonance structure it became clear to Criegee that
ozone can act as a dipole adding to double bonds such that the first step in the reaction is

a 1,3-dipolar cycloaddition, forming a primary ozonides (Figure 4.4).

 

/C—C\

Primary Ozonide
Figure 4.4: Formation of Primary Ozonide
However, the primary ozonide is a very unstable 5 member ring, 1,2,3-trioxolane.
Practical ozonolyses are run at -78 — 25°C, and under these conditions the primary
ozonide decomposes immediately by selective cleavage of the carbon-carbon bond and
one oxygen-oxygen bond leading to a carbonyl compound and the Criegee “Zwitterion”,

a peroxidic carbonyl oxide, as a second fragment (Figure 4.5).

 

0\ /o
05 P _. <3 . ..
/(|:_(|:\ /?\ /C\
Carbonyl Oxide
+ /O + 0
ll < ; O_(l:_
/ \ |

Figure 4.5: Ozonolysis Reaction Mechanism

45

Due to the large oxygen content of these molecules, the carbonyl oxide is very unstable
and tends to decompose. Polymeric peroxides are of particular concern because of their
low solubility and tendency to precipitate. Therefore for safety consideration maintaining
low concentrations of these species is important. This can be achieved by using a
continuous process or working in dilute solutions.

It was proposed by Marshall and Garafalo that under basic conditions, the
ozonolysis of alkenes in methanolic NaOH could redirect the decomposition to form
methyl esters (Figure 4.6). Their work included an ozonloysis of methyl oleate in
methanolic NaOH, which yielded dimethyl azelate and methyl nonanoate at 77 and 78

wt%, respectively.

0

H3C—O/U\/\/\/\/\/\/\/\/\CH

Methyl Oleate (C l 8: l )

3

03 , NaOH/CH30H/CH2C12

H30 W CH3
\0 O/

Dimethyl Azelate (77 wt%)

+
0

H30 WCHa
\0

Methyl Nonanoate (78 wt%)

Figure 4.6: Ozonolysis Products of Methyl Oleate in Methanolic NaOH

46

Nominally, the ozonolysis of methyl oleate would lead to a mixture of aldehydes and
carboxylic acids. A new synthesis was developed to prepare soy polyols by reacting soy
oil with a diol in an alkaline medium using ozone.
4.2 LITERATURE REVIEW
4.2.1 Methods of Polyol Preparation of Vegetable Oils

One of the early methods to prepare polyols from various vegetable oils is based
on transesterification of the fatty acids in the triglycerides with a polyol such as glycerol,
glycerin, pentaerythritol, a-methylglucoside or sucrose (1, 2). Unfortunately, premature
degradation occurs by this process due to high temperatures and a relatively long period
of time in this reaction. Another method is based on reacting vegetable oils with peroxy
acid to yield epoxidized fatty acids. The epoxide rings are then opened (3, 4) or
hydroxylated (5) with polyfunctional alcohols to yield secondary alcohols. Epoxidized
soy oil is available commercially but the reactivity of this oil is low due to the nature of
the secondary alcohols. Furthermore, several hydroxyl groups per fatty acid residue are
obtained (at least these fatty acids that contain more than one double bond).
Consequently, multiple numbers of hydroxyl groups having varying reactivity are
present, which tend to complicate subsequent reactions that could lead to premature
gelation. Hydroforrnylation offers another method to prepare polyols whereby an
aldehyde functional vegetable oil is first obtained, which is then hydrogenated to alcohols
(6, 7). Polyurethanes prepared from these polyols had different mechanical properties
depending on the hydroformylation catalyst that was used. Thus, rigid materials at room
temperature were obtained with a rhodium catalyst while cobalt catalyzed

hydroformylation led to rubbery materials (8). An alternative method to prepare primary

47

 

polyols is based on oxidizing an olefin having a carbonyl group with molecular oxygen
followed by hydrolysis and reduction of the acetal (or ketal) to an alcohol (9). Although
this method appears somewhat complicated and must be run at high pressure, good yields
were reported. A somewhat less complicated method, also based on an oxidation process
to yield polyols, is based on catalytic oxidation using an organic hydroperoxide in the
presence of OSO4 and a NaBr cocatalyst (10). An alternative process is to use ozone to
cleave and oxidize the double bonds in the vegetable oil and then reduce the
decomposing ozonides to alcohols using NaBHa or Similar reducing agents.

It is apparent that there is a need to improve the chemistry as well as the process
and devise a more efficient method to prepare vegetable oils containing primary alcohols.
Reported is a new method to synthesize such soy-based primary polyols in a single-step
ozonolysis procedure. Our method takes advantage of the ease and efficiency of the
oxidation methods mentioned earlier but it is relatively simple, fast and gives high yields

in a single step.

48

4.3 EXPERIMENTAL PROCEDURES
4.3.1 Materials and Methods

Degummed soybean oil was purchased from Spectrum Chemicals in a 20L
container and was used without further purification. Linoleic acid (purity>99%) was
purchased from Eastman Chemical. Ethylene glycol, sodium hydroxide, pyridine, 4-
dimethylamino—pyridine (4-DMAP) and CaCO3 were purchased from Sigma-Aldrich.
Ozone was produced by passing dry oxygen (0.25 ft3/min) through a Praxair Trailigaz
generator (Cincinnati, OH), model number OZC-1001. The exit port of the ozone
generator was connected to the bottom of the reaction bottle and allowed the ozone
(about 6 wt% in oxygen) to bubble through a fritted disc as fine gas bubbles in the
reaction mixture. The very small gas bubbles allowed good dispersion of ozone within
the reaction medium. Excess gas and unreacted ozone was vented through the top of the
reaction bottle through aqueous KI solution to destroy any unreacted ozone. Unless noted
all the ozonolysis reactions were run at 0°C and maintained at this temperature with
ice/water bath.
4.3.2 Catalytic Ozonolysis
Reactions with a Model Compound

Linoleic acid was used as a model compound for evaluation of the ozone attack
on the double bonds. Linoleic acid and ethylene glycol were placed in the 500 ml
reaction bottle and cooled with ice/water to 0°C. Ozone was allowed to bubble through
while samples were removed periodically and analyzed to determine the extent of double

bond disappearance by iodine test.

49

Ozonolysis of Soybean Oil

Soybean oil, ethylene glycol and various amounts of catalysts (NaOH, pyridine,
4-DMAP or C3CO3) were placed in a 500 ml gas wash bottle and cooled to 0°C in an ice
water bath and then ozone was bubbled through the reaction mixture. The product was
then washed with distilled water 5 times (twice with 0.2M HCl to neutralize the base) in
order to remove any excess ethylene glycol and catalyst (when CaC03 was used as a
catalyst, it was filtered out through a fine filter paper). The product was then dried over
molecular sieves for 48 hours prior to testing.

4.3.3 Characterization Procedures

The number of double bonds was determined by measuring the iodine value
according to ASTM test method D1959. The iodine value is defined as centigrams of 12
per gram of sample. The hydroxyl value was determined by ASTM test method D1957.
Hydroxyl value is defined according to this test as the number of milligrams KOH
equivalent to hydroxyl content of 1 g of sample.

Functional groups were identified using Perkin-Elmer FT IR model 1000 using at
least 64 scans. The spectra were obtained after polyol samples were dissolved in
chloroform and a drop of the solution was placed on a sodium chloride plate. The solvent
was allowed to evaporate leaving a thin film on the surface of the crystal. The sample cell
was constantly purged with dry nitrogen while the signal was acquired.

NMR spectra were obtained on a 500 MHz model INOVA 500 instrument.
Samples were dissolved in deuterated chloroform, and the 13C NMR spectra were

obtained at room temperature.

50

4.4 RESULTS AND DISCUSSION
4.4.1 Reactions with Ozone

Ozone is a very powerful oxidation agent and is well known to attack and cleave
double bonds in alkenes. Aside from its high oxidation potential, ozone oxidation is
convenient Since unreacted ozone simply decomposes back to oxygen and no Special
neutralization or separation are required at the end of the reaction. Furthermore, since
ozone is being produced “on site” it does not require complex logistics in transport,
Storage and disposal as many other oxidation agents. The oxidation of linoleic acid under
our reaction conditions was used as a model compound for the ozonolysis reaction by

following the disappearance of carbon-carbon double bonds (Figure 4.7).

 

2.0

1.5 -

1.0 ~ 1

Double Bonds

.5 _ f

 

 

 

00 1 1 1 ‘
0.0 0.5 1.0 1.5 2.0

Ozonolysis Time [min/g of sample]

Figure 4.7 : Ozonolysis Reaction with a Model Compound

51

It is observed that under these conditions all the double bonds were cleaved within 2

rrrinutes (g of soy oil basis). It is well known that in the absence of participating solvents,

unstable 1, 2, 4-trioxolane and peroxide oligomers are formed by the Criegee mechanism

and these species eventually decomposed to a mixture of aldehydes and carboxylic acids.

Some products containing hydroxyl groups were also noted when ethylene glycol was

present, even after excessive washing and separation to remove any traces of this alcohol

from the oil. This is not surprising since alcohols (and protic solvents in general) are

known to impact the rate of the reaction and react with the ozonide intermediate (11).

However, the course of the reaction is changed in the presence of an alkaline catalyst,

which leads to an ester linkage between the decomposing ozonide intermediate and the

 

 

alcohol (12).

Fraction Composition Symbol Frequency -OH
Nonanoate, Nonanoate, Nonanoate NNN 0.23520 3
Nonanoate, Nonanoate, Palmitate NNP 0.09240 2
N onanoate, Nonanoate, Stearate NNS 0.04200 2

I Nonanoate, Palmitate, Stearate NPS 0.01100 1
Palmitate, Palmitate, Nonanoate PPN 0.01210 1
Palmitate, Palmitate, Palmitate, PPP 0.00053 0
Palmitate, Palmitate, Stearate PPS 0.00072 0
Stearate, Stearate, Nonanoate SSN 0.00250 1
Stearate, Stearate, Palmitate SSP 0.00033 0
Stearate, Stearate, Stearate SSS 0.00005 0
Bis(2-hydroxylethyl) malonate EE 0.26984 2
(2-hydroxylethyl) nonanoate E1 0.09524 1

II (2-hydroxylethyl) hexanoate E2 0.20635 1
(2-hydroxylethyl) propionate E3 0.03175 1

 

 

 

Table 4.1: Statistical Distribution of Soy Polyol

52

Nominally, the ozonolysis reaction of alkene with excess ethylene glycol and an

alkaline catalyst (e.g. NaOH) is expected to yield a mixture of polyols as shown in

Figure 2.8:
03 O
R/VR + Hung—0112-014 W R/\n/O—CH2-CH2-OH + HO—CH5CH50—“\/R'

0

Figure 4.8: Catalytic Ozonolysis of Alkene with Ethylene Glycol

4.4.2 Composition of Ozone Treated Soy Oil

Similarly, when soy oil was exposed to ozone in the presence of ethylene glycol and
NaOH the cleavage of the carbon-carbon double bonds and the reaction of hydroxyl
groups with the ozonide intermediate yielded a mixture of polyols as Shown in Figure
4.9. This product mixture was clear, and had a lower viscosity than the Starting soy oil. A
closer examination into the composition of this polyol mixture indicates two distinct
fractions of about equal weights (Table 4.1); a relatively high molecular weight
triglyceride fraction and a low molecular weight fraction. The high molecular weight
fraction consists of the triglycerides that were cleaved at the 9th position of the
unsaturated fatty acid residues and reacted with ethylene glycol. Thus, NNN denotes a
triglyceride composed of only 2-hydroxyethyl nonanoate esters. This triglyceride was
obtained from any triglycerides that contained only unsaturated fatty esters (e.g. oleic,
linoleic, or linolenic). Since the first double bond is always found at the 9th position,
cleavage occurs there and this active center reacts with ethylene glycol as indicated in
scheme 1. Similarly, NNS denotes a triglyceride composed of two 2-hydroxyethyl

nonanoate esters and one stearic ester. Since stearic acid does not contain double bonds, it

53

 

r fans—:1;

is not affected by the ozonolosis reaction and remains unchanged. However, the other
two fatty acids in the original triglyceride contained double bonds and, thus, produced 2-
hydroxyethyl nonanoate esters. The low molecular weight fraction is composed of the

fragments that were cleaved beyond the double bonds of the unsaturated fatty acid.

 

 

0 0 0 0 0
—0-J-LR —0—”—[—01~12}—JJ—0,/~0H Ho’VO‘JWLm/‘OH
7
_” 03 cat. 0 0 + o
—0 -n > —0-JL[-0H2}—“-0./~OH CH3-[—CH,-]—”—O./~OH
0 HO’VOH 7 1,4 or7
-oJLR _R.

 

 

Palmitic acid (16:0) - 11%
/\/\/W\/\/\

 

Stearic acid (18:0) - 4% Palmitic 3019 (1630)
R'- /\/\/\/\=/\/\/\/\ R: /\/\/\/\/\/\/W
Oleic acid (18:1, 90) - 23% Stearic acid (1810)

W
Linoleic acid (18:2; 90, 12C) - 51%

Linolenic acid (18:3 90,120,15C) - 7%

Figure 4.9: Catalytic Ozonolysis Products of Soy Oil with Ethylene Glycol

It Should be noted that about 45% of these fractions are composed of the diol bis(2-
hydroxyethyl) malonate and the rest is a mixture of mono functional alcohols (Figure
4.10). Fraction 1 contains about 24 wt% triols (NNN), 13 wt% diols (NNP and NNS),
less than 3 wt% mono functional alcohols (NPS, PPN and SSN) and a very small fraction

(less than 0.2 wt %) of unreactive triglycerides having no hydroxyl groups.

54

 

 

 

 

 

 

 

 

 

 

 

 

30
: EE
25 ’_ Soy Polyol m
. N N N
P = Palmitic acid
52 S = Stearic acid

20 : N = Nanorc acrd
>‘ EB: bis(2-hydroxyethyl) malonate
3 E1: (2-hydroxyethyl) nonanoate
g 15 : E2: (2-hydroxyethy1)hexanoate
E; E3: (2-hydroxyethyl) propionate
L“ [m

10 ~ 5‘ L N P

£11
5 i 53
N P S
N P P
N S S
0 i -1 1 i i I i 121 . .
100 200 300 400 500 600 700 800 900

Molecular Weight [g]

Figure 4.10: Statistical Distribution of Soy Polyols
The average functionality of the soy polyol can be calculated based on these statistical

data using:

Jar
f“‘" 2...

Figure 4.11: Average Functionality

 

Where, 122, is the number of moles of species containing functionality f,-. Accordingly, the
maximum hydroxyl functionality is 1.9 per mole of soybean oil (Figure 4.11).

In practice, the hydroxyl functionality may be higher if the ozonolysis reaction is
not terminated after all the double bonds are consumed and hydrocarbon chains are
cleaved. Similarly, a hydroxyl functionality lower than 1.9 is expected if the ozonolysis

reaction is terminated before all the double bonds are consumed. Additionally, our

55

calculation did not take into account complications such as the reaction of both hydroxyl
groups of ethylene glycol. In this case a high molecular weight product is expected if
ethylene glycol reacts with two triglycerides. On the other hand, intramolecular reactions,
where ethylene glycol reacts with two fatty acids in the same triglyceride, will lead to
cyclization and loss of hydroxyl functionality.Similar complications could arise when
any of the hydroxyl terminated fragments in fraction H reacts with another triglyceride
instead of ethylene glycol. Most of this complication could be avoided, or at least greatly
minimized, by using a large excess ethylene glycol in the reaction mixture.

The process conditions can further be adjusted to control the molecular weight
and the hydroxyl number of the product by using triols or other common low molecular
weight polyols instead of ethylene glycol. Alternatively, the catalyzed ozonolysis reaction
with polyethylene glycols of different molecular weights, instead of ethylene glycol,
yields products containing different polyether chain lengths. Subsequent reaction of these
polyols with isocyanates gives flexible polyurethane foams with various degrees of
flexibility, depending on the molecular weight of the polyethers used in the ozonolysis
reaction.

AS expected, the soy polyol mixture is characterized by a broad hydroxyl
Stretching peak around 3500 cm", the complete disappearance of the =C---H band at
3005 cm'l and the C=C stretch at 1650 cm'1 (Figure 4.12). This set of FT IR spectra
further indicates that at the end of the ozonolysis reaction the carbonyl stretch at 1743
cm‘1 became broader suggesting the formation of new carbonyl compounds. A Significant

increase in the 1105 cm"1 band was also observed and that was assigned to C---O

56

stretching. It is important to note here that no absorptions were noted around 2900—2700

cm’1 indicating the absence of aldehydic C---H stretch.

 

i

 

 

 

 

 

Absorbance

 

 

 

 

 

L A J 1

3700 3200 2700 2200 1700 1200 700
Wavenumber [cm"]

Figure 4.12: FTIR of Soy Oil (A) and Soy Polyol (B)

It was also observed that while keeping the reaction time constant, the hydroxyl number
was directly proportional to the concentration of the NaOH and increased to 1.96 when
0.4 M NaOH was used. However, further attempts to use higher NaOH concentrations
caused saponification of the triglyceride ester linkages and the formation of free glycerol.

Possible saponification and the difficulty associated with removing traces of
NaOH from the reaction mixture even after neutralization and repeated washing led us to
examine alternative base catalysts. One class of potential catalysts we examined was
various amines (e.g. 4-dimethylamino pyridine (4-DMAP), pyridine, triethylamine,

tributylamine, dimethylamine and hexamethylene diamine). It was observed that similar

57

soy polyol products were obtained when any of these amines were used (Appendix A).
As was the case with NaOH, the number of double bonds continuously decreased as a

function of ozonation time and the hydroxyl number simultaneously increased (Figure

 

 

4.13).
1.0 t
r 5.0

0.8 r .
5 ~ 4.0
E L.
>. .8
e ' S
u: 0.6 ~ « z
'O
I: r 3.0 '3‘
B ><
0 O
:5 6
3 0.4 - ‘ E
8 . 2.0

0'2 h j 1.0

> 4
I
00 se-e- - e L 00

 

 

 

0.0 0.5 1.0 1.5 2.0
Ozonation Time [min/g of sample]

Figure 4.13: Hydroxyl Number (---) and Double Bonds (—)
as a Function of Ozonolysis Time
It is important to note that unlike previous methods in the literature describing
preparation of soy polyols, our one step ozonolysis process appears to yield high
hydroxyl numbers and primary alcohols in a simple, fast reaction. The rate of cleavage of
the double bonds was independent of the amine catalyst and was directly proportional to
the amount of ozone and the dispersion of the ozone bubbles in the reaction medium.
Hydroxyl numbers were directly related to the type of amine used and its concentration.

Although the reaction proceeded as anticipated it was noted that the polyol product was

58

dark-yellow to brown depending on the type of amine that was used due to its oxidation
in the course of the reaction (13). It was further noted that strong bases like 4-DMAP led
to saponification of the fatty esters in the soy oil as observed with NaOH.

4.4.3 Reactions with Heterogeneous Catalysis

Another class of potential catalysts that appeared most useful was inorganic alkaline
oxides and carbonates (such as MgO and CaC03). These heterogeneous catalysts are
effective and can simply be filtered out and removed from the reaction mixture with no
need for neutralization. These heterogeneous catalysts can further be recycled and reused
without any special treatment. To the best of our knowledge, only zeolites (14) were
claimed to be heterogeneous catalysts in ozonolysis reactions although these catalysts
were used without additional hydroxyl compounds to in the production of dicarboxylic
acid from alkenes by ozonolysis. The progress of the reaction whereby double bonds
were cleaved and hydroxyl groups were incorporated is clearly seen by FTIR (Figure
4.14). In addition to convenient removing and recycling this heterogeneous catalyst, the
soy polyol product is light-yellow, low viscosity oil suitable for further use in the

production of polyurethanes and polyesters without further purification.

59

 

 

 

Absorbance

 

 

 

 

 

C
B
A
A... l _i_._ 1 r 1 i i i 1 1 l 1 1
3750 3500 3250 3000 2750 2500

Wavenumber [cm'l]

Figure 4.14: Hydroxylation of Soy Oil using CaCO3 at Different Ozonolysis Times:
(A) t=0; (B) t=0.25; (C) t=0.5; (D) t=1.0; and (E) t=2.0 min/g of soy oil

60

Further confirmation of the soy polyol structure was obtained from 13C NMR
spectra (Figure 4.15 and 4.16). The characteristic double bonds peaks at 130 ppm,
related to the unsaturated fatty acid in the soy oil, are missing in the soy polyol Spectrum
indicating complete cleavage of the double bonds. The carbonyl ester peaks (177 ppm)
and the various methylene peaks (between 25 and 36 ppm) remained unchanged as were
the glycerol carbons (64 and 69 ppm). Additionally, new resonance peaks appear at 66
ppm related to ethylene oxide carbons as well as peaks at 60 ppm related to the new C-
OH functional groups. It iS apparent from the NMR data that the hydroxylation of

soybean oil progressed as expected to yield the desired polyols.

-CH2_

-C= - -c—0-

 

 

 

 

 

 

 

 

 

 

 

 

i .
i
l
1
W V

[Iiifllirgr'ii'.i_rrlililzrilr.rl:r]:;7filtvIT‘TIfi£{IIIET—rIglei‘TlIIIfrI:T‘TlliIIIIiT'Iil'rr7

 

180 160 140 120 100 80 60 40 20 ppm

Figure 4.15: ”GM of Soy Oil

61

'CHZ'

-00-

 

 

 

 

 

 

 

 

 

Figure 4.16: l3C-NMR of Soy Polyol

62

4.5 CONCLUSIONS

A new class of soy-based polyols was prepared using a simple single step
synthesis. This novel synthesis involves an ozone mediated reaction between soy oil with
a multi-functional hydroxy compound in an alkaline medium. The reaction proceeded
rapidly at room temperature, without an additional solvent. The alkaline medium
facilitated the redirection of the decomposition of the ozonides to form ester linkages
between the soy oil molecules and a hydroxy compound. The reaction proceeded with
both strong bases (NaOH) and mild bases (pyridine). Sodium hydroxide was not a
desirable catalyst because of its limited solubility and side reactions such as
saponification. Pyridine was used because it had good solubility, however, its negative
health effects make it an unlikely candidate for potential large scale commercial
applications. Thus, triethylamine was used as an alternative. The resulting polyols were
a dark yellow to brown color due to oxidation of the amine. Tributylamine, was
evaluated because it has better oxidative stability, however, these polyols also changed
color over time. Using an alkaline heterogeneous catalyst, such as CaCO3, solved
previously mentioned difficulties. CaC03 did not cause saponification of the triglyceride;
however, it was sufficient to promote the formation of the ester linkages. Furthermore,
the CaCO3 could easily be removed and recycled by filtering. Using CaC03 a series of
soy polyols with varying hydroxyl functionalities was prepared varying the hydroxyl
compound (eg — ethylene glycol, 1,4-butanediol, glycerol, PEG, PVOH, starch).
Additionally, soy-based amino compounds were prepared by using ethanolamine. Using
ozone chemistry as a platform, a portfolio of biobased building blocks were prepared as

an alternative to traditional petroleum feedstocks.

63

4.6 REFERENCES
l. Stanton, J .M., Isocyanate-Modified Drying Oils, J. Am. Oil Chem. Soc. 36: 503-507
2. Wells, ER, and J .C. Hixenbaugh, Am. Paint J. 46: 88 (1962).

3. Sherringham, J .A., A]. Clark, and B.R.T. Keene, New Chemical Feedstocks from
Unsaturated Oils, Lipid Technology 12: 129-132 (2000).

4. Gruber, B., R. Hoefer, H. Kluth, and A. Meffert, Polyols on the Basis of Oleochemical
Raw Materials, Fett Wissenschaft Technologie 89: 147-151 (1987).

5. Heidbreder, A., R. Gruetzmacher, U. Nagomy, and A. Westfechtel, Use of Fatty
Esfter-bgsed Polyols for Polyurethgne CgtinLResins and Cgting Materials

6. Frankel, E.N., and EH. Pryde, Catalytic Hydroformylation and Hydrocarboxylation of
Unsaturated Fatty Compounds, J. Am. Oil Chem. Soc. 54: 873-881 (1977).

7. Khoe, T.H., F.H. Otey, and EN. Frankel, Rigid Urethane Foams from
Hydroxymethylated Linseed Oil and Polyol Esters, J. Am. Oil Chem. Soc. 49: 615-618
(1972).

8. Guo, A., D. Demydov, W. Zhang, and 2.8. Petrovic, Polyols and Polyurethanes from
Hydroformylation of Soybean Oil, Polym. grid the Environ. 10: 49-52 (2002).

9. Takahara, J ., and T. Setoyama, Process for Preparation of Polyhydric Alcohols by
Oxidation of Olefins having a Carbonyl Group and Catalytic Hydrogenation, W0
Application Patent 2002049999 (2002).

10. Michaelson, R.G., and RC. Austin, Hydroxylation of olefins, US Patent 4314088
(1982).

11. Ledea, 0., J. Molerio, M. Diaz, D. Jardines, A. Rosado, and T. Correa, Revista
CENIC, Ciencias Ouimicas 29: 75-78 (1998).

12. Marshall, J .A., and A.W. Garofalo, Oxidative Cleavage of Mono-, Di-, and
Trisubstituted Olefines to Methyl Esters through Ozonolysis of NaOH, J. Org. Chem. 58:
3675-3680 (1993).

13. Bailey, P.S., J .E. Keller, D.A. Mitchard, and HM. White, Ozonation of Amines,
AMnces in Chemistry Series 77: 58-64 (1968).

14. Rebrovic, L., Catalyzed Process for Oxidation of Ozonides of Unsaturates to
Carboxylic Acids, US Patent 5,399,749 (1995).

Chapter 5: Continuous Process for the Catalytic Ozonolysis of Soy Oil

5.1 INTRODUCTION
5. 1 . 1 Soy Polyol Production
Present day commercial applications of continuous ozonolysis are primarily for water
treatment applications and odor removal on farms. Recently, Cognis has commercialized
the preparation of nonanoic and azelaic acid by the ozonolysis of oleic acid in water.
Common to all the processes is the low organic concentration within the aqueous
mixture. Among the novelties of the proposed catalytic ozonolysis iS the absence of a
solvent. In the proposed process a high fatty acid mass composition is possible because
the alkaline medium facilitates the decomposition of the ozonides, to prevent the
formation of polymeric peroxides (Chapter 4). The ozone mediated synthesis of soy
polyols involves a series of competitive reactions as seen in Figure 5.1. Competitive
reactions of ozone are:

1) Decomposition back to O2

2) Reaction with H2O to form H202

3) Reaction with double bonds to form a molozonide

65

0

03 > H202 AOH
\\ / 0
/°\ 1
> H

 

 

Figure 5.1: Competing Reactions of Ozone

Of these competing reactions the selectivity of ozone favors the reaction with double
bonds. The molozonide being an unstable 5 member ring, further undergoes
decomposition resulting in the formation of aldehydes and carboxylic acids. However, in
the presence of an alkaline medium it is further observed that the decomposition of the
ozonides can be redirected to form ester linkages with hydroxyl groups. Furthermore if
reactions are carried out in the presence of multifunctional compounds, such as ethylene
glycol, soy-based polyols can be prepared. The soy polyols can further undergo
oligomerization reactions with decomposing ozonides. To increase the yield of the soy
polyols and/or oligomers of the soy polyol various process parameters were studied and
optimized, and a proof-of-concept was established for the continuous production of soy

polyols using ozone chemistry.

66

 

5.2 REACTOR DESIGN THEORY
5.2.1 Batch and Continuous Reactors

In —- Out + Generation = Accumulation
V dN
Fm — F0“, +1 adv = d—' “Figure 5.2: General Mole Balance”
1

Batch Reactor:
0 There is no flow in or out of the system.
0 Perfect mixing so there is no Spatial variation in the reaction rate.
0 In this case the mole balance simplifies.

dN,
7 = riV “Figure 5.3: Batch Reactor Mole Balance”

There are 2 primary types of batch reactors: Constant Volume and Constant Pressure

| dc,
dt

dC. C.d ln V
co ‘ + ' = r,- “Constant Pressure”
dt d:

Figure 5.4: Batch Reactor Design Equation

= r,- “Constant Volume”

 

 

 

 

 

 

Batch reactors are not favorable for large-scale ozone application. Due to the large
oxygen content the reaction intermediates are unstable. Thus, the ozonide concentration
should be minimized. Gaseous ozone becomes explosive when concentrations reach 240
g/m3. Most ozonation systems are in the range of 50-200 g/m3. So, for safety concerns
large-scale batch applications would require the use of solvents. Alternatively, the
ozonide concentration can be minimized using a continuous flow reactor. There are 2

general types of flow reactors.

67

Continuous Stirred Tank Reactor (CSTR):
0 Good mixing resulting in no spatial variations in concentration, temperature, or
reaction rate throughout the reactor
Easy maintenance and good temperature control
However, CSTR has the lowest conversion per reactor volume for flow reactors
For a CSTR the mole balance simplifies to:

 

—-.1..—.

UI V = in out
" ’1

Figure 5.5: Continuous Stirred Tank Reactor Design Equation

 

 

 

Plug Flow Reactor (PFR)
0 The concentration does not vary in the radial direction
0 Reactants are consumed as the flow though the length of the reactor (Reaction
rate varies as the concentration changes axially)
The extent of reaction depends on the reactor volume, and not shape
The mole balance for a PFR simplifies to:

 

 

 

Figure 5.6: Plug Flow Reactor Design Equation

0 For heterogeneous reactions, the reaction rate is based on the mass of the solid
catalyst

68

5.3 REACTOR DESIGN METHODOLOGY
5.3.1 Comparison of Reactors

A CSTR offers the advantage of no spatial variations in temperature,
concentration, or reaction rate. We will also consider non-ideal mixing which can result
from the increased viscosity of the polyol. If mixing is determined to be highly non-ideal,
alternative techniques such as residence time distribution (RTD) will be used. The major
drawback of CSTR is they have the lowest conversion per reactor volume. A CSTR
operates under the condition of lowest reactant concentration (eg — exit concentration),
and thus the lowest value of the reaction rate.

A PBR has increased conversion per reactor volume. The main parameters
when using a PBR are the temperature control, pressure drop, and reduced reactant
concentration. Although, there is expected to be a pressure drop across the PBR,
however, Since the reaction is liquid phase; we are going to ignore the pressure drop
because the concentration of reactants is insignificantly effected by changes in pressure
(incompressible flow). 80, the pressure drop is ignored for the liquid-phase kinetics.
Alternatively, the decrease in concentration of ethylene glycol is expected to hinder the
reaction because of oligomerization effects, or ozonides decomposition to a carboxylic
acid (presence of H20) or the aldehyde.

Stoichiometry:
For our proposed process, if we consider each reaction between ozone and the double

bonds present in soybean oil. The general reaction stoichiometry is:

69

R R

1 O 0
M + 203 + 2CH3OH ————> i CH3 + i /CH3 + 02+ 2H20
H H R 0/ R1 0

Figure 5.7: Ozonolysis Mole Balance
From the stoichiometry the relative reaction rates are: -1:-2:-2: 1:1:122, respectively.

Nominally, the rate of reactant disappearance becomes:

_ d[DB] = _2 d[03] = _2 d[EG]
dt dt dt

 

 

“Figure 5.8: Rate of Reactant Disappearance”

Where DB = Double Bonds
EG = Ethylene Glycol

While the rate of product formation equals:

2d[Polyol] = 2d[HzO] = d[02]

d: dt d t “Figure 5.9: Rate of Product Formation”

For every double bond 1 mole of ozone will react to form a primary ozonide which
readily decomposes to form an aldehyde and carbonyl oxide (Figure 4.3). It is proposed
that in an alkaline median the decomposition can be redirected to for ester linkages with
hydroxyl groups (Appendix A). Furthermore, if a diol such as ethylene glycol is used,
ester linkages with terminal hydroxyl groups will form. The reaction products are 2
moles of polyol, 2 moles of water and 1 mole of oxygen. Absent any Side reactions 1
mole of soy oil (~4.6 double bonds per mole) could theoretically yield ~9 moles of

polyol.

70

5.4 RESULTS AND DISCUSSIONS

5.4.1 Processing Parameters Consideration

Preliminary studies were done to further investigate the formation of the soy polyols
using a heterogeneous catalyst. Variables such as reaction time, temperature, catalyst
concentration, and catalyst particle Size were studied. Based on the batch synthesis inside .
the gas wash bottle, it was observed that after 2 minutes of reaction time per gram of soy
oil, 30% of the double bonds remained. Considering the reaction stoichiometry, a mass
balance of the system shows only 30% of the ozone charged into the reacted with the soy
oil. Although, some of the ozone could have reacted with moisture in the system, it is
believed that most of the ozone simply exits the system into the destruction unit. A
change of the frit location to face downward (initially ~130 degree angle), nearly doubled

the reaction rate of the soy oil as seen in Figure 5.10

 

— Soybean Oil
— SO/EG/CCl (new frit)
-- SO/EG lCCl (old frit)

 

Absorbance

 

 

 

 

 

M
”op.

/‘\ flit! \\
1 i R ._ . g.

4000 3800 3600 3400 3200 3000 2800 2600 2400 2200 2000
Wavenumber (cm'l)

 

 

Figure 5.10: FTIR of Soy Polyol using Different Frit Locations

71

From these data it appeared that changing the frit location increased the hydroxyl content
by increasing the mass transfer of ozone within the reaction mixture and thus improving
the reaction yield. Positioning the frit at the bottom of the reactor increased the ozone
dispersion within the reaction mixture. Under this condition the double bonds consumed
within 1 minute were comparable to the 2 minute reaction time as noted previously.
Alternatively, for the reaction temperatures considered (0-25°C) the reaction rate
was generally unaffected. In general terms the reaction rate tends double for 10°C
increases in the reaction temperature. However, in the case of ozone reaction with double
bonds, the activation energy is very small (ozone has a very high oxidation potential), so
the reaction temperature does not Si gnificantly impact the reaction rate constant (which is
related through the Arrhenius equation). Alternatively, increasing the temperature
adversely effects the reaction rate since the activation energy is low, the reactions occur
instantaneously, whereas the solubility of ozone is inversely related to the temperature.
Consequently, these reactions appear to be mass transfer limited and not reaction limited
Furthermore, the catalyst concentration was found to be directly proportional to
the hydroxyl content of the product. It was found that increasing the catalyst loading
from 1 wt% to 10 wt% increased the rate of hydroxyl formation as seen by the intensity

of the stretching at 3500 cm'1 as seen in Figure 5.11.

72

 

 

 

 

 

 

 

 

—- Soybean Oil

— SO/EG/CC 1 (new frit)

-- SO/EG/CClO (new frit)
8
:1
ii
8
<

1 4 x 1 - A 1 4::
4000 3800 3600 3400 3200 3000 2800 2600 2400 2200 2000

Wavenumber (cm'l)

Figure 5.11: FTIR of Soy Polyol for Catalyst Loadings of 1 wt% & 10 wt%

Further increase in the catalyst loading may further increase the hydroxyl formation but
operation above 10 wt% becomes difficult due to the resulting slurry of soy oil, ethylene

glycol, and CaCO3).

FTIR data indicate that the conversion was better when 10 mm CaC03 particles
were used compared with smaller particle size CaCO3 (micron) were used (Figure 5.12).
Since, for equal catalyst loading (~10 wt %), the larger CaCO3 particle will have less
surface area than the smaller particle CaCO3. These experiments suggest that the reaction

is diffusion limited and not surface limited.

73

 

 

 

 

 

 

 

 

— Soybean Oil

— SO/EG/CC 10 (new frit)

-— SO/EG/nCC 10 (new frit)
8
1:
a
i
<

J 1 A
4000 3800 3600 3400 3200 3000 2800 2600 2200 2000

Wavenumber (cm!)

Figure 5.12: FTIR of Soy Polyols using Different Particle Sizes of CaCO3

To verify this observation hydroxyl value and acid value were determined by titration.

 

 

 

 

 

 

 

CaC03 Hydroxyl Acid
10 mm 50 82
50 pm 128 10
10 pm 172 8

 

 

Table 5.1: Hydroxyl and Acid Value of Soy Polyols

The results of the titration suggest the larger absorption around 3500 cm’1 in the FTIR

using large particle size CaC03 is derived from the combination of —0H and —COOH

functional groups. Alternatively, when smaller particles of CaCO3 were used, the

-COOH formation was considerably less. It should be noted that when no catalyst was

used the hydroxyl number was very low (< 8 mgs of KOH/g of sample) indicating very

74

little polyol formation. Also, the moisture content was ~8 wt% (determined by TGA).
The titrations suggest the hydroxyl is affected by the catalyst surface area. To rule out
the hydroxyl number was not due to ethylene glycol remaining in the polyol, a repeat of
the earlier experiment was done using methanol instead of ethylene glycol (residual

methanol was Stripped off) as seen in Figure 5.13.

 

 

 

 

 

 

 

 

SO
5
ILmCC za-
8
D
J <
mmCC
3700 3200 2700 A 2200 l700 1200 700

Wavenumber [em'l]

Figure 5.13: Variation of Catalyst Particle Size using a Model Compound
FT IR from the reaction using methanol indicates using larger particle size CaC03 had a
substantially absorbance at 3500 cm'1 whereas there is only a Slight absorbance when
smaller particle CaCO3 is used. It should be further noted, that even finer CaC03 particle
size lead to an increase in the hydroxyl content and decrease in the carboxyl content,

however, removing the CaC03 became very difficult. Commercially, the extent of

increased hydroxyl formation has to be weighed against the difficulty in removing the
catalyst.

Heterogeneous catalyst offers the advantages of recyclability and ease of removal.
Preliminary studies suggest the optimal reaction conditions to be 10 wt% loading of 50
um particle size CaCO3. Due to the reactivity of ozone towards doubles bonds (low
activation energy), the optimal temperature is room temperature (with cooling water
available for safety). The greatest impact to the reaction rate is the mass transfer of
ozone, and Specifically the residence time of ozone within the system as seen by the
changing of the frit location. Additionally, an economic evaluation of the current polyol

production cost is most sensitive to production cost of ozone (Figure 5.14).

 

 

 

 

$0.80
Power Cost

$0.75 —
’5
g Calcium Carbonate Cost /
é \ Soybean Oll Cost
0. . . - . . _ — ”dam-Wu - - V- “.4... ~-~:~1-r~'°"“f""“5W '"'”“"‘:""""“’ '
2'
O
a.

$0.65

Ethylene Glycol Cost
$0.60 1 1 L
-33% -17% 0% 17% 33%

Deviation from Base Case

Figure 5.14: Sensitivity Analysis of Soy Polyol Cost

76

5.5 CONTINEOUS REACTOR DESIGN
5.5.1 Reactor Selection and Sizing

There are several difficulties of associated to continuously producing polyol such
as maintaining uniform contact between the gas and liquid, different residence times for

the gas and liquid, and obtaining the residence times as long as those in a batch reactor.

 

 

Batch: No further investigation

 

 

 

 

 

CSTR:

 

 

 

PFR (PBR)

 

Figure 5.15 : Reactor Design Selection
A PBR would further offer the advantage of removing an additional unit operation for
catalyst recovery. However, due to the constraint of ozone production (equipment
specification) the maximum volumetric flow rate is not sufficient to generate turbulent
flow for enhanced mass transfer of the gas and liquid. Thus the reactor decided upon for
the continuous production of soy polyol was CSTR. The 3 impellers are used to enhance
mass transfer by creating turbulent flow. The ozone enters into the reactor through a
Sparging ring located at the bottom of the reactor. Directly above the gas inlet is flat

blade turbine used to further shear the gas bubbles, followed by 2 pitch blade turbines

77

used to direct the flow downward to further increase residence times. Additionally, 4

baffles are used to create turbulent flow.

 

 

 

 

 

 

 

 

 

 

 

 

 

Gas Ir

 

 

Figure 5.16: Continuous Production of Soy Polyol

78

The reactor was sized according to the apparent residence time:
Based on the available ozone flow rate (~0.013 moles/min) a maximum volumetric flow
rate of 4.03 mein can be used.

reactor volume

 

apparent residence time = .
volumetnc flow rate

Residence Time

 

Height Volume (3" diameter)

 

 

 

 

 

 

 

 

inches cm ems L min hr
6 15.24 695 0.69 173 2.9
8 20.32 927 0.93 231 3.8
10 25.4 1158 1.16 288 4.8
12 30.48 1390 1.39 346 5.8

 

 

 

 

Table 5.2: Apparent Residence Time
The residence times were selected to allow for sufficient reaction time.
5.5.2 Process Scale-up

The current continuous soy polyol process uses a CSTR as seen in Figure 5.17.

 

 

 

 

 

 

 

 

 

 

     

Soy Polyol <
Cooling Destruction Unit
Water

Soy Oil

+

Ethylene Glycol

. +

Cooling 1 C 21C 0,
Water 0:
Ozone Generator .
Stir Tank
Oxygen Tank

Figure 5.17 : Process Flow Diagram for Continuous Polyol Production

79

For reduced production cost, the polyol synthesis should be down stream from soy oil

extraction (Figure 5.18).

1 Oil Seed

Crusher H exan e
. Storage
Solrds

 

 

 

 

 

 

Pump

 

 

Hexane

V

 

 

 

 

O

8

5
on

 

 

 

 

Figure 5.18: Polyol Production Downstream from Soy Oil Extraction
For “large” scale productions where flow rates are sufficient for turbulent flow a PBR
could offer further cost reduction. Unit operations can be removed for making a slurry of
the catalyst and reactants before entering the reactor and another for removal/recycling of
the catalyst. With the larger flow rates, alternative methods for ozone delivery can
further enhance ozone mass transfer. Three methods for ozone delivery are proposed.
Figure 5.19 uses a venturi injector to deliver ozone to a mixture of soy oil and ethylene

glycol.

80

Gas Outlet
Polyol

__________ .’ -_--_--_——
CoolingWater

   

Venturi

__.M_. :f" :36
SOandEG I g

Ozone

PBR

Figure 5.19: Ozone Injected to a Mixture of Soy Oil and Ethylene Glycol
Experimentally it was found that in the absence of a catalyst the reaction would

proceed to form aldehydes and carboxylic acids. A possible solution is to inject the
ozone into the ethylene glycol stream, and charge the reactor with soy oil (Figure 5.20).

Gas Outlet

Polyol

Soy Oil

 

Figure 5.20: Ozone Injected into Ethylene Glycol

81

The third option is deliver ozone by gas Sparging through a diffuser as seen in

 

Figure 5.21.
Gas Outlet
——o

Polyol

---------- o - - - - - - - - - - ->

Cooling Water Cooling Water

___p ‘—
EG Soy Oil

 

Ozone

Figure 5.21: Ozone Injected by Gas Sparging
5.5.3 Process Economics
Based on an annual polyol production rate of 500 lbs/hr the production cost is
$0.51 per pound. A capital investment of $1.2 million dollars was assumed with a 5 year
debt write-off. The ozone reaction efficiency was taken to be 80%, and the plant would

operate 70% of the year. The calculations are detailed in Appendix C.

82

Chapter 6: Thermal-Oxidatively Stable Biobased Lubricants

6.1 LITERATURE REVIEW
6.1.1 Soy oil based lubricants

Soybean oil lubricants are preferred to petroleum oils because they are cheap,
renewable and have performance advantages (1). Soybean oil has low volatility due to
the high molecular weight. In addition the presence of polar ester groups results in good
boundary lubrication due to adhesion with metal surfaces. Alternatively, one of the major
challenges to using SO as a lubricant is poor oxidative stability (2). The poor oxidative
stability can be attributed to the reactivity of the double bonds present within SO. On
average SO contains 4.3-4.6 double bonds per mole depending on the origin of the seed.
The presence of unsaturation can undergo reactions such as free radical oxidation,
fragmentation, rearrangement, disproportionation and polymerization, which reduces
oxidative stability. Various methods such as chemical modification and genetic
modification have been used to improve the oxidative stability. Although, additives
technologies can be used to improve the overall oxidative stability of the lubricant, it has
only has an indirect effect on the oxidative stability of SO.

Several researches have investigated the use of chemical modification as a means
of enhancing oxidative stability (3). Adhvaryu and coworkers (4, 5) reported on the use
of epoxidized soybean oil and alkoxylated triacylglycerols as lubricants. Other chemical
modifications were done by sulfurization and phosphate modification (6, 7). However,
these schemes did not account for the complexity of SO. Soybean oil is a complex
mixture of primarily five fatty acids (palmitic, stearic, oleic, linoleic, and linolenic acids)

with very different melting points, oxidative stability, and chemical functionalities. Thus,

83

there is expected to be variability in performance resulting from the varying selectivity of
the chemical modification. In other Studies the chemically modified SO were just
evaluated in less severe applications, such as emulsions in water for a cutting fluid (8).
Other studies have used genetic modification to improve the oxidative stability of SO.
Researchers at Dupont increased the oleic acid content of S0 to ~80% (9). Conventional
SO contains oleic acid levels of ~25% and polyunsaturated fatty acid levels of over 60%.
The presence of bis allylic protons is highly susceptible to radical attack and
subsequently undergoes oxidative degradation. More recently, Buhr et al. (10) achieved
SO with greater than 85% oleic acid content and relatively low polyunsaturated fatty acid
content (3-5%). While improving the oxidative stability of S0 is possible through
genetic modification; these approaches are generally longer routes which are not
necessarily welcomed due to concerns regarding genetically modified crops.

High oxidative stability is a critical property for lubricants. The purpose of this
study is to use novel chemical modifications to improve the oxidative stability of SO and
to evaluate their lubricant base oil properties. Two different chemical modifications were
investigated. The first scheme used a free radical initiated addition of maleic anhydride
to S0, followed by an esterification with an alcohol. The second scheme involved
passing ozone through a solution of soybean oil and a mono-functional alcohol in the

presence of an alkaline catalyst.

84

6.2 EXPERIMENTAL PROCEDURES
6. 2.1 Materials and Methods

Degummed soybean oil was purchased from Spectrum Chemicals in a 20L
container and was used without further purification. Maleic anhydride (MA), 2,5-
BiS(tert-butylperoxy)-2,5-dimethylhexane (L101), Methanol, butanol, methoxy-
polyethylene glycol (Mw = 350), sodium hydroxide, pyridine, potassium iodide, 4-
dimethylamino-pyridine (4-DMAP), CaCO3 and all other reagents were purchased from
Sigma-Aldrich (Milwaukee, WI). Ozone was produced by passing dry oxygen (0.25
ft3/min) through Praxair Trailigaz generator (Cincinnati, OH), model number OZC-1001.
The exit port of the ozone generator was connected to the bottom of the reaction bottle
and allowed the ozone (about 6 wt% in oxygen) to bubble through a fritted disc as fine
gas bubbles in the reaction mixture. The very small gas bubbles allowed good dispersion
of ozone within the reaction medium. Excess gas and unreacted ozone was vented
through the top of the reaction bottle through aqueous KI solution to destroy any
unreacted ozone. Unless noted all the ozonolysis reactions were run at 0°C and
maintained at this temperature with ice/water bath.
6.2.2 Synthesis Techniques
Maleation of Soybean Oil

Reactions were carried out by charging a 2L Parr Reactor equipped with a
motorized stirrer, thermocouple, and heating element with MA, SO, and the initiator.
The reactor was purged with N2 for 5 minutes, sealed and heated. The initial reaction
time was taken at the time the reactor temperature reached 150°C, where it was

maintained for 30 minutes. At the end of each reaction, the reactor was connected to a

85

vacuum pump for 30 rrrinute and any volatile products were collected in a solvent trap

immersed in isopropanol dry ice bath (Figure 6.1).

 

\
l
I

\~ ---
- ,——
~—-—.._—- \

s

s_‘-
~.

.-....-_

 

Controller Parr Reactor Nitrogen Vacuum Pump

Figure 6.1: Soy Oil Maleation and Esterification Apparatus

Esterification of Maleated Soybean Oil

Maleated SO and a large excess of alcohol (10:1 molar ratio alcohol to SO) was
added to a 2L Parr Reactor equipped with a motorized stirrer, thermocouple, and heating
element. The reactor was purge with N2 for 5 minutes, sealed and heated. The initial
reaction time was taken at the time the reactor temperature reached 200°C, where it was
maintained for 30 minutes. At the end of each reaction, the reactor was connected to a
vacuum pump for 30 minutes and any volatile products were collected in a solvent trap

immersed in isopropanol dry ice bath.

86

Ozonolysis of Soybean Oil

Soybean oil, an alcohol (methanol, butanol, octanol, dodecanol, or methoxy-
PEG350) and various amounts of catalysts (NaOH, pyridine, 4-DMAP or CaC03) were
placed in a 500 ml gas wash bottle and cooled to 0°C in an ice water bath and then ozone
was bubbled through the reaction mixture. Unless otherwise noted, all reactions were run
for 2 hours and the product was then washed with distilled water 5 times in order to
remove any excess alcohol and catalyst (when CaCO3 was used as a catalyst, it was
filtered out through a fine filter paper). The product was then dried over molecular sieves

for 48 hours prior to testing.

6.2.3 Characterization Procedures

The number of double bonds was determined using iodine value testing,
according to ASTM test method D1959. The iodine number in this test is defined as
centigrams of 12 per gram of sample.

The acid value was determined according to ASTM test method D1980. The acid
number in this test is defined as milligrams of KOH per gram of sample.

Functional groups were identified using Perkin-Elmer FTIR model 1000 using at
least 64 scans. Prior to recording a spectrum, the sample holder was flushed with nitrogen
to remove moisture and CO2 and the background was recorded. The spectra were
obtained after polyol samples were dissolved in chloroform and a drop of the solution
was placed on sodium chloride crystals. The solvent was allowed to evaporate leaving a
thin film on the surface of the crystal. The sample cell was constantly purged with dry

nitrogen while the signal was acquired.

87

Thermo-gravimetric analysis (TGA) was carried out with a TA Instruments (New
Castle, DE) Model TGA 2950 thermobalance. The instrument was operated in the
dynamic mode with a heating rate of 5°C/min, where 40°C and 350°C were the initial
and final temperatures, respectively. On average, 4 mg was the initial mass of each
sample analyzed. Air was used instead of inert atmosphere, to study the effect the
thermal oxidation.

NMR spectra were obtained on a 500 MHz model INOVA 500 instrument.
Samples were dissolved in deuterated chloroform, and the 1H NMR and 13C NMR spectra

were obtained at room temperature.

Thermo-oxidation Engine Oil Simulation Test (TEOS) was used to simulate
oxidation in a moderately high temperature deposit conditions in the piston ring zone of
modern smaller, more highly, stressed engines. Due to the known poor oxidation stability
of soy oil in this test, a sample was run for only 10 hours at 285°C rather than the normal
24 hours. Accordingly, 8.5 g of oil was recirculated continuously over the special steel
rod heated to 285°C. Air is circulated continuously over the rod to increase exposure to
oxygen. In addition, any volatile material is caught by the walls of a surrounding mantle
and collected separately, thus increasing the stress on the remaining oil. The weight of the
rod before and after the test was then determined as the main criterion of the extent of
oxidation.

Scanning Brookfield Technique (SBT) was used to determine the low temperature
profile of the lubricant by continuously measuring the viscosity as the temperature is
slowly (1°C/hr) decreased from 0°C to the maximum viscosity of the viscometer head.

Any structures that were built-up in the sample caused an increase in viscosity above the

88

exponential relationship expected from a Newtonian fluid, which by definition, is free of
gel-forming tendencies. The presence of such structures was found by taking the first
derivative of a linearized treatment of the viscosity-temperature curve. In the case of
these studies, the temperature range chosen was from 0°C to the lowest temperature
possible of the viscometer head.

Friction and Wear were measured by a modified Falex pin and Vblock. In this
test, Two V-blocks press against a rotating pin from opposite sides, ‘pinching’ the pin
between them with a force that is progressively increased in steps by the test operator.
The contact between the V-blocks and the pin are four straight lines and permit
evaluation of the lubricant tested in the so-called quasihydrodynamic region of
lubrication. This region can produce wear and ultimate seizure of the contiguous
contacting surfaces. The test was conducted with increasing 50 lb steps of force with five
minutes residence time at each step. Wear, friction and pin temperature were measured at

each step.

89

6.3 RESULTS AND DISCUSSION

The presence of double bonds in vegetable oils is largely responsible to the poor
oxidative and thermal stability of these oils, two chemical routes were explored to
remove the double bonds and introduce ester functionality. The first route involves
insertion of maleic anhydride across the double bond followed by esterification with an
alcohol. The second route consisted of catalytic ozonation of the double bonds in the
presence of an alcohol to cleave the double bonds and attach the alcohol to the new
terminal carbon via ester linkages.

6. 3.1 Transformation of Double Bonds via Maleation

JL M

1700 1200 7

3700 3200 2700 2200 r
Wavenumber [cm‘ ]

 

Absorption

 

 

 

 

 

00

Figure 6.2: FTIR of Soybean Oil (A); Maleated Soybean Oil (B); and Esterified
Maleated Soybean Oil (C)

90

The first modification was done by introducing an anhydride functionality to S0,
followed by an esterification using an alcohol. Soybean oil was functionalized using
single-step free radical initiated chemistry. The anhydride functionality was confirmed
by FTIR spectroscopy (Figure 6.2). The spectra shows the maleated soybean oil exhibits
peaks at ~1775 cm'1 and 1850 cm”1 not present in S0. These absorption peaks are related
to the symmetric and asymmetric stretching of C=O in the MA, respectively.
Furthermore, no absorption peaks related to -OH stretching due to the opening of the
anhydride is observed, clearly indicating that the anhydride ring remained intact.
Additionally, the alcohol esterified maleated SO also shows no absorption peaks ~3500

1

cm' , suggesting complete esterification and half ester formation. Nominally, the

esterication proceeds as presented in Figure 6.3.

O 0

° l I
Fl Fl

 

l

 

Figure 6.3: Esterification of Maleated Soybean Oil with an Alcohol

To favor the formation of the ester over the acid, the esterification was done at elevated
temperature (200°C) and pressure with excess alcohol. A series of aliphatic alcohols with
increasing molecular weight were considered (methanol, ethanol, butanol, octanol and
methoxy-PEG). Experimentally, it was oberved that the oxidative stability increased
with the molecular weight of the alcohol. However, when alcohols with molecular

weight greater than butanol were used, the oxidative stability was found to decrease.

91

Additionally, the use of linear aliphatic alcohols was preferred over branched alcohols
such as t-butyl alcohol as it offered greater thermo oxidative stability.

The structural modification with the best thermo oxidative stability is carried out
in 2 stages:

1) Free radical maleation of soybean oil in a closed system

2) High temperature and pressure esterification of maleated soybean oil using

excess butanol.
6.3.2 Transformation of Double Bonds via Catalytic Ozonolysis

Somewhat similarly the same series of aliphatic alcohols with increasing
molecular weight were (methanol, hexanol, octanol and methoxy-PEG350) used in a
catalytic ozonolysis to transform the bonds to more thermal-oxidatively stable ester

linkages (Figure 6.4).

 

 

W 03 O O 0 GR
-0 + HO-R —- ~0 _
se 0 R + O
° ° 0 AAJLO-R
Toll—MW TOLMA/MAA/
0
CH—u—o-n

Figure 6.4: Catalytic Ozonolysis of Soy Oil in the Presence of an Alcohol
The lower molecular fragment derived from cleavage of the double bonds must be
removed because of their volatility. These low molecular weight fragments are expected
to have low boiling points, reducing the thermal stability. However, these fragments
compose almost ~40 wt% of the mixture. It can also be seen the fragments have a

significant effect on the molecular weight distribution (Table 6.1).

92

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Whole Distribution Fragments Removed
Alcohol
Mn Mw Mn Mw
Methanol 350 553.3 676.3 679.6
Hexanol 480. 1 669.6 850.2 850.3
Octanol 532.2 721.2 919.8 920.6
MEG350 736.9 921 - -
Soy 011 871.7 871.8 - -
n=—Z——n M
Zn
WzZn M 2
wZn M
Where: n; = mole fraction of species i

M; = molecular weight of species i

Table 6.1: Calculated Molecular Weight Distribution
Where Mn and Mw are the number average molecular weight and the weight average
molecular weight, respectively. The molecular weights were calculated as:
The mole fractions were calculated based on the fatty acid distribution as seen in Figure

1.4. A sample calculation for reactions with methanol is summarized in Table 6.2.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Triglyceride Frequency MW Fragments Frequency MW
NNN 0.2352 644 R0 0.0952 172
NNP 0.1008 698 RL 0.2698 132
NNS 0.0336 727 RLl 0.2063 130
PPP 0.0007 807 RL2 0.0317 88
PPN 0.0007 753
PPS 0.0144 835 _2___C0m lete Mixture
SSS 0 391 Mn = 350 Mw = 553
SSN 0.0016 809 W
SSP 0.0002 863 Mn = 676 MW = 680
NPS 0.0096 781

 

Table 6.2: Molecular Weight Distribution for Reactions with Methanol

93

 

To reduce the volatility of the lower molecular weight fragments, higher molecular
weight alcohols were used as apposed to removal of low boilers. Experimentally, it was
observed that methoxy-PEG350 had the best thermal stability because of the volatility of

the lower molecular weight species resulting from the cleavage of the double bonds

(Figure 6.5).

 

 

 

Rate of Change [wt%/min]

 

 

 

_O.6 1111111111111iiiiiiriiiirnmiiriiiiriiur
0 5 10 15 20 25 30 35 40

Tlme [min]

Figure 6.5: Rate of Decomposition of Ozone Treated MEG350 and Soy Oil
0 min reaction (A); 10 min reaction (B); and 30 nrin reaction (C)
As the reaction time increases, the rate of decomposition decreases. Both ozone treated
samples, rate of decomposition levels off after an initial drop in weight. This initial
weight loss is probably due to residual moisture, and lower weight fragments. However,
in the case of plain soy oil, the decompositions progresses further. This is most likely
due to further oxidation resulting in a breakdown of soy oil. The corresponding weight

loss in an oxidizing atmosphere is presented in Figure 6.6.

94

100 _.
99 °\\“‘-\\

98

 

97}

Weight (%)
8

 

 

 

95 f C
94 f A
93 3 e
92 I 1 1 1 1 1 1 1 L4 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
0 5 10 15 20 25 30 35 40

Time (min)
Figure 6.6: Weight Loss of Ozone Treated MEG350 and Soy Oil
0 min reaction (A); 10 min reaction (B); and 30 min reaction (C)
The FTIR of the ozonized soy oil Shows near complete removal of the double bonds in
soy oil (Figure 6.5). The stretching at 3005 cm‘1 from the C=CH is significantly less in
the ozone treated sample than from soy oil. It can also be seen that there is negligible

acid formation (~3500 cm’l).

95

 

 

 

 

L
Absorption

 

 

 

A 1 k 1 J l

3700 3200 2700 2200 r 1700 1200 700
Wavenumber [cm' ]

 

Figure 6.7: FTIR of Soy Oil (A) and Ozonized Soy Oil with Methoxy-PEG350 (B)

6.3.3 Evaluation as a Biobased Lubricant

Based on improved oxidative stability, the butanol esterified maleated soy oil and
methoxy-PEG350 ozonized samples lubricating properties were evaluated. Both modified
soy oils Showed reduced tendency to form deposits as compared to soy oil (Figure 6.8).
This is most likely due to the reduced number of double bonds present. This is further
supported by the ozone modified soy oil having the best performance with the least
number of double bonds among the samples. In the case of the maleated sample, there is
a net reduction of 1 double bond, whereas near complete removal of double bonds in the

ozone modified soy oil.

96

 

Maleated

587.5

 

Ozonized 6

 

Soy Oil 1590

 

 

 

O

500 1000 1500 2000
Deposits [mg]

Figure 6.8: Deposits Formation of Lubricant Sample

Similarly, the ozone modified soy oil also had the best the cold flow property
performance as seen in Figure 6.9. The cleavage from the ozone treatment results in a
series of fragments improving the cold flow properties of standard soy oil. Alternatively,
the esterified-maleated soy oil structure still resembles that of soy oil, thus their cold flow

properties remain relatively unchanged.

97

100000

 

 

 

 

 

 

 

 

\
\
\ _\ ‘,
’ Ozonized / 1
80000 P - - - Maleation \ I ‘,
h — -Soybean 011 ‘ -‘
l '.
E 60000 ~ ' '
£0 ‘ 0
b ‘
g ‘ ‘1
.2 l ‘
> 5 0
40000 - l 1
i l ‘,
t | 1
0
. ‘ ‘ .
20000 5 ‘ “
l- e
. l ‘.
0 :3
-35 -30 -25 -20 -15 -10 -5 0

Temperature (°C]

Figure 6.9: Cold Flow Properties of Lubricants

Both modified soy oil could greater loads before failure than untreated soy oil
(Figure 6.10). Additionally, it can also be seen that durability the ozone modified soy oil
is better than the esterifed-maleated soy oil. The wear plots for the two modified soy oils

can be seen in Figures 6.11 and 6.12, the ozone modified oil Shows less wear with

increasing load than the esterified-maleated soy oil.

98

Maleated

Ozonized

Soy 011

0.02

 

 

 

 

1222

1712

942

 

500 1000 1500 2000
Failure Load [lbs]

Figure 6.10: Failure Load for Soy-based Lubricants

 

0.015 -

Wear [in]

0.005 _

 

0.01 r

+ Wear Step
+ Cumlative Wear

 

 

 

500 1000 l 500 2000
Actual Load [lbs]

Figure 6.11: Wear Plot of Esterified-Maleated Soy Oil

99

Wear [in]

0.02

0.015

0.01

0.005

 

 

 

 

 

+ Wear Step
+ Cumlative Wear
0 500 1000 1500 2000
Actual Load [lbs]

Figure 6.12: Wear Plot of Ozone Modified Soy Oil

100

6.4 CONCLUSIONS
The oxidative degradation of soybean oil was greatly reduced by transforming the
double bonds into more oxidatively stable esters were done using to two chemical

modifications.

0 Maleation followed by an esterification using an butanol
0 Catalytic ozonolysis in the presence of an methoxy-PEG350

The present study shows that both processes were found to Significantly reduce soy oil
tendency to form deposits (insoluble matter) by removing oxidation-susceptible sites
from the soy molecules. However, from the viewpoint of effects on low-temperature
properties — dependent on the physical configuration of the soy molecules and their
tendency to associate into larger, more flow-resisting structures at low-temperature — the
processes have little effect. Additionally, both modifications were found to significantly

improve lubrication performance in comparison to the untreated soy oil.

101

6.5 REFERENCES

1. Erhan 8.2. and Asadauskas S., Lubricant basestocks from vegetable oils, DEL-
Crogs Prod. 11, 277-282 (2000)

2. Asadauskas S., Perez J .M., and Duda J .L., Oxidative stability and anti-wear
properties of high oleic vegetable oils, Lubr. Eng. 52, 877-882 (1996)

3. D’Sousa V., DeMan L., and DeMan J .M., Polymeric behavior of high melting
glycerides from hydrogenated canola oil, J. Am. Oil Chem. Soc. 68, 907-911
(1991)

4. Adhvaryu A. and Erhan S.Z., Epoxidized soybean oil as a potential source of
high-temperature lubricants, Ind. Crops Prod. 15, 247-254 (2002)

5. Adhvaryu A., Liu Z., and Erhan S.Z., Synthesis of novel alkoxylated
triacylglycerols and their lubricant base oil properties, Ind. Crops Prod. 21, 113-
119 (2005)

6. Schwab A.W. and Gast L.E., Free radical unsaturated oils addition of hydrogen

sulfide to conjugated and nonconjugated methyl esters of vegetable oils, J. Am.
Oil Chem. Soc. 47, 371-373 (1970)

7. Schwab A.W., Gast LE, and Rohwedder W.K., Nucleophilic addition of
hydrogen sulfide to methyl oleate, methyl linoleate, and soybean oil, J. Am. Oil
Chem. Soc. 52, 236-239 (1975)

8. Jacob J., Bhattacharaya M., and Raynor P., Emulsions containing vegetable oils
for cutting fluids application, Colloids and Surf. 237: 141-150 (2004)

9. Adhvaryu A. and Erhan S.Z., Epoxidized soybean oil as a potential source of
high-temperature lubricants, Ind. Crops Prod. 15, 247-254 (2002)

10. Warth H., Mulhaupt R., Hoffmann B., Lawrence S., Polyester network based
upon epoxidized and maleinated natural oil, Angewagidte 249, 79-92 (1997)

11. Bellamy L.J., The InMed Spectrg of Complex Molecules, Wiley: New York,
(1964).

12. Pouchert 0]., The Aldrich Library of Infra-Red Spectra, Aldrich Chemical
Company; Milwaukee, WI, (1970).

102

Chapter 7: Silane Functionalized Soybean oil

7.1 LITERATURE REVIEW
7.1.1 Silane Addition to Soy Oil

Several strategies have been used in the past to enhance the reactivity of vegetable
oils by manipulation of the double bonds such that they can be crosslinked at room
temperature which were then used as coating, adhesives, or sealants. Preparation of these
Silane-modified vegetable oils involved methods that initially yielded polyols derivatives
of the vegetable oil, either by transesterification, epoxidation, or oxidation. Typically,
these polyols were then reacted with excess diisocyanate and end-capped with different
amino functional Silane coupling agents (1). It was noted that coatings based on this
epoxidized 10-undecenoic acid triglyceride had excellent UV stability. Alternatively, a
solution of soybean oil and Cl3SiH could be irradiated with Hg lamps to hydrosilylate the
oil and incorporate the trichlorosilane onto a double bond (2). In this case too, exposure
to moisture will hydrolyze the chlorosilane to silanols, which will then condense to form
the siloxane crosslinking bridges. The added advantage of silicone species covalently
bonded to the oil is their reactivity with inorganic silicates surfaces. Thus, it was claimed
that Silane modified soy oil has good adhesion to concrete and could be used as a water
repellent surface coatings for masonry structures.

Here reports the results of our studies where vinyltrimethoxy Silane is reacted
with soy oils under anhydrous conditions and allowed then, to cure upon exposing to

ambient atmosphere.

103

7.2 EXPERIMENTAL PROCEDURES
7.2.1 Materials and Methods

Unhydrogenated soybean oil (SO) was purchased from Spectrum Chemicals
(Gardena, CA). Linoleic Acid (LA), Oleic Acid (0A), Dibutylin dilaurate, Tyzor 131,
2,5-Bis(tert-butylperoxy)-2,5-dimethylhexane (L101), BF3-methanol complex and all
other reagents were purchased from Sigma-Aldrich (Milwaukee, WI).
Vinyltrimethoxysilane (VTMS) was obtained from GE Silcones (Friendly, WV).
7.2.2 Synthesis Techniques
Silane F unctionalization of Soy Oil (or Fatty Acids Derivatives)

Silylation reactions were carried out by charging a 2L Parr Reactor equipped with
a motorized stirrer, thermocouple, and heating element with the VTMS, SO, and the
initiator (Figure 6.1). The initial reaction time was taken at the time the reactor
temperature reached a preset temperature. The effects of reaction time, temperature,
catalyst type and concentration were investigated. At the end of each reaction, the reactor
was connected to a vacuum pump for 30 minute and any volatile products were collected
in a solvent trap immersed in isopropanol dry ice bath.
Room Temperature Cure Studies

Room temperature curing reactions were done by adding the 50 grams of Silane
functionalized soy oil (or fatty acid derivative) to a glass petri dish and left in controlled
environment at 25°C and 50% relative humidity. The effects of cure time and catalyst
concentration were investigated. The cure was allowed to continue and the gel was then
removed and extracted in a soxhlet apparatus for 3 days using methylene chloride as a

solvent and the gel fraction was determined.

104

 

7.2.3 Characterization Methods

Infra red spectra were recorded using a Perkin-Elmer model 1000 with a NaCl
window and 64 scans. Prior to recording a spectrum, the sample holder was flushed with
nitrogen to remove moisture and CO2 and the background was recorded. The sample cell

was constantly purged with dry nitrogen while the Signal was acquired.

NMR spectra were obtained on a Varian 500 MHz model INOVA instrument.
Samples were dissolved in deuterated chloroform, and lH-NMR and l3C-NMR spectra

were obtained at room temperature.

TGA was performed using TA Instruments Q-2950 high resolution TGA
equipment. The samples were raised to a temperature of 600°C at 20°Can in an
atmosphere of air and then held isothermal for 20 minutes.

The degree of swelling was determined by immersing the cured sample for 24
hours in a series of solvents with varying CED (cohesive energy density). The cured
samples were removed and the swollen weight was determined. The samples were then

placed in a vacuum oven and dried to constant weight (polymer dry weight).

105

7.3 RESULTS AND DISCUSSIONS

Soybean oil is an abundant biorenewable resource composed of a mixture of
triglycerides consisting of glycerin and saturated and unsaturated long chain fatty acid. Some
typically natural oil fatty acid compositions are presented in Table 1.1. Depending on the
origin of the oil seed, the number of double bonds can vary from 4.3-4.6 per mole. Although
the fatty acid composition may vary, common among all samples of 80 are the internal

position of the double bonds and the cis configuration (Table 1.2).
7.3.1 Evaluation of Processing Parameters

The effects of reaction time, initiator concentration, and MA concentration were
studied in details using two different initiators, L101 and BF3-methanol. The degree of
VTMS addition was determined gravimetrically, and the products were further
characterized using FTIR spectroscopy, lH-NMR, l3C-NMR and iodine value. The
VTMS functionality is calculated as the moles of VTMS per mole of SO. Similarly, the
number of double bonds is calculated from the iodine value (cgs of I2/g of sample).

In previous work we found that maleic anhydride (MA) could react with the
double bonds within SO by an “ene” addition or Diels-Alder adduct depending on the
reaction conditions. It was found that reactions between MA and SO in the presence of a
peroxide facilitated conjugation of the double bonds resulting in a Diels-Alder type
adduct, whereas when no peroxide was used the “ene” addition product was formed.
Nominally, the addition of VTMS to 80 is expected to yield products as shown in Figure

7.1.

106

 

 

 

4» _,1 < —-e
\\ R-0/'\O-R _\\\ _ 8(0—R
R-O/ \O-R

Scheme 1: Diels-Alder Addition

' H ‘ ' H
i O—R .13 ‘><:rll l
+ ./ ——> —>
\ /31\ \fl O-R
R-O O_R _ _ Si/
R-O/ ‘o-R

Scheme 2: Ene Addition

I

 

 

Figure 7.1: Diels-Alder and Ene Addition Product
The addition of VTMS to S0 in a closed system under high temperatures and pressures
proceeded relatively fast with and without an initiator. Three systems were considered;
reactions with no catalyst, a peroxide initiated reaction, and a Lewis acid catalysis. Since
soy oil contains internal non-conjugated double bonds the “ene” addition type is
expected, which could be further promoted using Lewis acid (e.g. — BF3-methanol).
Similarly, a peroxide (e.g. L101) was used to promote conjugation of the double bonds
within the polyunsaturated fatty acids. In all cases the reaction proceeded to completion
within ~120 minutes as shown in Figure 7.2. Gravimetric determination of the reaction
product between SO and VTMS after pulling a V2 hour vacuum at 150°C correlated to the
addition 3 moles of VTMS per mole of SO. It is further interesting to note that the type
of catalyst had little effect on the progress of the reaction and essentially the same

reaction rates were observed for both L101 and BF3-methanol.

107

 

 

 

 

 

 

 

 

 

4
A i
2
g 0
1: 4
8
o B T ’
= 1
1?.
m 2
2
I"
>
0
4
r J
C t
2
0 I fi' T Y
0 50 100 150 200 250

Reaction Time [min]

Figure 7.2: Addition of VTMS to SO at 150C with Time
(A) no catalyst; (B) L101; (C) BF3-methanol

Similarly, no significant differences were observed when the reaction temperature
was adjusted from room temperature to 100°C, 150°C, 200°C, or 250°C for 30 minutes.
Under these conditions, for reaction at 150°C and above, 1 mole of VTMS was added to
80 as shown in Figure 8.3. In these cases the VTMS addition is kinetically controlled,

as opposed to thermodynamic control.

108

 

1.2

 

1.0 - ~ - ,3 , 7- ~-

P
on

VTMS Functionality
p
0‘

P
J;
l
l

0.2 - , 1

 

 

 

 

0.0

   

25

150
Temperature (°C)

Figure 7.3: Addition of VTMS to SO with No Catalyst for 30 minutes
with Increasing Temperature

Experimentally, it was observed the addition of VTMS to S0 (or fatty acid
residue) by an “ene” addition. In the case of the Diels-Alder adduct, there is expected to
be a net reduction of 1 double bond from the unmodified soy oil (e. g. ~3.6 double bonds).
This same number of double bonds suggests that reactions favor the “ene” addition over
the Diels Alder addition. Further verification was done by using the fatty acid residues of
SO as model compounds. In these reactions OA and LA were reacted with VTMS with
and without the peroxide. Similar, to the reaction of VTMS with S0 at 150 °C for 120
minutes yielded no reduction in the number of double bonds; the reactions with the model
compounds also retained the original number of double bonds (Figure 7.4). Although,

further addition of VTMS is theoretically possible, it was found experimentally that only

109

 

3 mole of VTMS had reacted as determined by gravimetric analysis. Although the reason
for this upper limit addition of VTMS is not clear at this time, one possible explanation

could be steric hindrance effects that impede further reaction of VTMS.

 

    
    
 

  

 

  

 

 

 

5
Dnocatalyst
I0.01wt% L101
4 1.
8 3 ~
E
13
2
1D
’8"
2 " ’ ,
, //

 

 

 

 

 

 

Oleic Acid Linoleic Acid Soy Oil

Figure 7.4: Double Bonds Remaining Following the Reactions of VTMS with SO
at 150 °C for 120 minutes (and other model compounds)

7.3.2 Characterization of Silane F unctionalized Soy Oil

The VTMS functionalized SO was further characterized by FT IR spectroscopy
(Figure 7.5) and exhibits a splitting of the carbonyl peak at ~1750 cm'l present in neat
soy oil to another peak at ~1710 cm‘1 not present in S0. Additionally, there are
significant differences within the finger print region. Especially, noticeable is the peak at
1085 cm’1 and its intensity. Further verification was done using a model compound. As

seen in Figure 7.6, the reaction between VTMS and LA also shows a new peak at ~1710

110

 

cm" and an intense peak at 1085 cm’1 not seen in FTIR of pure LA. The narrowing of the
broad peaks of hydrogen bonded acid groups in LA to that in the silylated LA suggests

there is some transesterification between the methoxy group from the Silane and the

carboxylic acid group of LA.

 

 

— Soybean 011
— SON'I'MS/LlOl .

 

 

 

Absorbance

 

 

 

 

 

 

 

 

 

 

Wavenumber [cm'l]

Figure 7.5: FTIR of SO and VTMS Functionalized SO

111

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

L

3700 3200 2700 2200 1700 1200 700

Wavenumber [cm'l]

Figure 7.6: FTIR of LA (A); LA/VTMS (B); and LAIVTMS/LlOl (C)

112

Absorbance

Further characterization was done using NMR using CDC13 as the solvent.
Carbon, proton and HMBC spectra were recorded, in order to confirm the grafting of the
vinyl group onto soy ester and elucidate the structure of the resulting compound. Since
native soybean oil is a mixed ester of different fatty acids, the NMR spectra are complex
and are not easily resolvable. The methylene protons of glyceryl backbone appear at 4.1
ppm and 4.3 ppm. The methine proton appears downfield at 5.2 ppm and the positions
vary depending on the fatty acid attached to the backbone. The methyl groups on each of
the fatty acids also vary in position and appear in the region 0.88 — 1.00 ppm. The double
bonds in the soy fatty acids are all non-conjugated and the protons appear in the region
5.2 — 5.6 ppm. The methylene protons in the fatty acids appear over a wide range
depending on their position in the acid. The methylenes adjacent to the carbonyl appear
around 2.1 — 2.3 ppm, methylenes sandwiched between double bonds appear downfield at
2.6 — 2.8 ppm. Methylenes B to the carbonyl appear at 1.6 — 1.7 ppm. Methylenes on
either side of double bonds (not sandwiched) appear around 2.0 ppm and the rest of the
methylenes appear in the region 1.2 — 1.5 ppm.

The proton spectrum of the Silane modified soy oil shows clear differences from
that of native soy oil. The intense peak at 3.5 ppm is characteristic of the methoxy groups
of the Silane. The additional peaks at around 6.1 ppm correspond to vinyl protons. Since
no such peaks are found in native soy oil, these have to result either from the residual
unreacted Silane in the mixture or from some bond migration due to grafting. The NMR
Spectrum of the pure Silane does not Show any peaks in this region and hence it can be
indirectly concluded that these peaks are a result of grafting reactions. It is also known

that conjugated double bonds appear in the 6.1 ppm region. From these results it can be

113

concluded that the grafting does occur and it occurs by an ene addition mentioned
elsewhere in this work. Also integration of the spectrum shows the relative proton
number for the 6.1 ppm peak to be around 2 (indicating at least one ene addition) and that
of the methoxy peaks of the siloxane moiety to be around 18. Since each siloxane unit
carries three methoxies, this corresponds to 2 moles of Silane grafted onto soy. Thus it
can be inferred that at least 2 moles of Silane have grafted onto the backbone. The ene
addition produces conjugation only in the polyunsaturated acid groups and does not affect
oleic acid or the saturates. The methoxy groups are by far the best means to ascertain the
extent of grafting through NMR. The absence of any broad peaks at 4.4 ppm indicates the
absence of silanols which can be formed by hydrolysis of the siloxane groups. The
possibility of self condensation of methoxy groups activated by heat is not ruled out and
this might account for the discrepancy in grafting efficiency as ascertained by thermo
gravimetric and gravimetric methods previously mentioned. (Appendix D)

The carbon spectra of native soy oil and silylated soy oil are also complex and
were used only to confirm the methoxy carbons of the siloxane moiety. These carbons
show up around 52.0 ppm as a broad peak. The vinyl carbons appear in the region 130.0
— 140.0 ppm and the carbonyls at 170.0 ppm. (Appendix D)

HMBC spectra were recorded in order to confirm the structure of the silylated soy
molecule. Since soy oil is a mixed ester, the spectrum was complex and no Significant
information could be inferred from the spectrum than from the 1-D experiments. Also
from the postulated mechanism (ene addition) and the resultant structure, it can be seen
that the protons on the carbon attached to silicon have a 4 bond coupling to the protons

on the soy carbons. Since a 4 bond coupling is very weak, it is extremely difficult to spot

114

on the HMBC. However the HMBC Spectra do confirm the results of the 1-D
experiments. The one bond coupling between the protons and carbon of the OCH3 group
is visible as doublets on either Sides of the actual peak (leak through one bond coupling).
The carbonyl group at 170 ppm correlates with the protons of the glyceryl backbone,
indicating that the ester backbone is preserved. It is not very easy to distinguish the
carbons of the vinyl group on siloxane (converted to methylenic after grafting) attached
to silicon in the Spectra as soy oil also has peaks in the region around 10-20 ppm
corresponding to methyl groups. The new peaks at 6.1 ppm which are present in the
silylated soy oil, also correlate to the vinyl carbons in the structure, which is supportive
of the ene addition mechanism. From the NMR data, it can be inferred that Silane does
graft onto the soy backbone and there is enough evidence in favor of an ene addition
mechanism operating and also the ene addition leads to conjugation in the resultant
structure. (Appendix D)
7.3.3 Room Temperature Moisture Activated Cure Studies

Room temperature cure studies of the VTMS functionalized soy oil were done in
a heat and moisture controlled room set at 25C and 50% relative humidity. The moisture
present in the was used to promote the hydrolysis of the silicon alkoxide bonds, to form
silanols which further react to form highly stable siloxane linkages. It was further
observed that this process could be enhanced with a catalyst. This is particularly the case
when an amine catalyst is used. As seen in Figure 7.7, to obtain the same sol fraction
as the amine catalyzed cure, takes ~3 times as long in the absence of any catalyst under
the same cure conditions. The use of a tin or titanate catalyst were effective in reducing

the cure time but not as efficient as the amine catalyst.

115

 

 

 

 

 

100
1 none
\ 0 tin
80 W W , W 3 31,3113” Camrne
; I titanium
“E 60-
.2
1'5
a:
I-
‘1'
8 w- -
A
I
20 ~
I A ‘
A
0
0 25 50 75 100 125 150 175

Cure Time [hr]

Figure 7.7 : Room Temperature Cure of VTMS Functionalized Soy Oil

7.3.4 Swelling of Soybased Thermosets

The degree of swelling of the soy thermoset was determined for a series of
solvents (Table 7.1) and plotted against the solvents’ solubility parameter. Dissolution of
a polymer begins with permeation of solvent molecules into the polymer’s amorphous
regions. The solvent breaks down intra- and inter-molecular bonds between polymer
chains, increasing their mobility and free volume. This results in swelling and softening
of the polymer and extraction of low molecular weight additives and polymer chains.
The thermodynamically driven swelling force is counterbalanced by the retractive force
of the crosslinked structure. Two forces become equal at some point and equilibrium is

reached. As seen in Figure 7.8 the greatest degree of swelling is observed for

116

trichloroethylene and chloroform. Interestingly, the graph shows the degree of swelling
of the soy thermoset in methylene chloride is more than acetone, even though acetone has
a closer solubility parameter. Theoretically, liquids with Similar solubility parameter
should have similar solubility characteristics, and yet the observed behavior in this
instance is different. This is most likely due to the differences in kinds of polar

contributions that give rise to the total cohesive energy densities in each case.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Solvent Density (g/ml) 8[MPam]
n-Hexane 0.659 14.9
Trichloroethylene 1.463 18.7
Toluene 0.865 18.3
Methylene Chloride 1.325 20.2
Chloroform 1.492 18.7
Acetone 0.791 19.7
n-Butanol 0.81 28.7
Ethanol 0.789 26.2
Methanol 0.791 29.7
Water 1 48

67301111 : ggispersion + algolar + alzlydrogen

Table 7.1: Solubility Parameter for Various Solvents

Since, there are 3 types of polar interactions contributing to solubility: dispersion
forces, polar forces, and hydrogen bonding forces. The inconsistencies observed in are
mostly likely contributed to the difference in hydrogen bonding between the chlorinated
solvents and the ketones. The intermolecular forces in the soy thermoset are primarily
due to dispersion forces, with practically no hydrogen bonding involved. These polar
configurations are perfectly matched by the intermolecular forces between chloroform
molecules, thus encouraging interpenetration and swelling of the soy thermoset. Acetone,

however, are more polar molecules, with moderate hydrogen bonding capabilities. Even

117

 

 

though the solubility parameter is more similar chloroform than methylene chloride, the
differences in component forces, primarily hydrogen bonding, lead to the observed

differences. Thus, acetone is attracted to each other rather than to the soy thermoset.

 

 

 

 

O
0 o
3 ..
0 0
60
fl
= O
a:
B
U)
"5
a 2 "
'53
O
5 . °
1 O
0 4 1%
O 5 10 15 20 25 30 35 40 45 50
Solubility Parameter [MPam]

Figure 7.8: Degree of Swelling of the Soy Thermoset with Different Solvents

Thermogravimetry was carried out on the native and cured silylated soybean oils
to determine their relative thermal stability in air and also evaluate the grafting efficiency
of the silane (Figure 7.9). The results of the thermal analysis show that grafting slightly
enhances the temperature stability of the soybean oil as indicated by the absence of any
Sharp drops in weight with temperature and also the ultimate weight after the analysis.
The native soy oil degraded completely in an air atmosphere while the silylated samples
still retained 15% or more of their weights. This gives an indication of the extent of

grafting of silane. When the experiments involving Silylation were run, a simple mass

118

balance around the reaction setup indicated that around 3 moles of silane were used up in
the reaction. This would correspond to an ultimate weight of around 14% after the TGA
run (assuming all the carbon was oxidized in the sample to leave silica). The TGA results
indicate that more than 15% of the sample remains after the experiment which confirms
the grafting efficiencyof the silane. The onset of weight loss however is at a lower
temperature for the silylated materials than for the soy oil. This is due to the fact that
curing reactions with siloxane invariably lead to loss of small compounds like methanol
or ethanol, which account for the early onset of weight loss. From the analysis it can also
be seen that the amine catalyzed cure and plain cure have better thermal stability than the
Tin catalyzed and TYZOR catalyzed cure. This could be explained on the basis that both
tin and TYZOR are capable of catalyzing the hydrolysis of ester backbone in soy leading
to lower degradation temperatures. This is also true for the amine catalyst, but its activity

is much lesser than that of the former.

 

 

    

120
100 « .-\\ - = - _
3 \ Soy Oil
80 ‘ Tin Cured —-——>
, Amine Cured

 

Tyzor Cured

Neat Cured

 

Weight [%1
8

 

 

 

f

0 100 200 300 400 500 600 700
Temperature [°C]

Figure 7.9: Thermal-Oxidative Stability of Soy Oil and RT Cured Silated Soy Oil

119

7.4 CONCLUSIONS

Silane functionalities were introduced to soybean oil for use as a chemical feedstock.
The investigation and optimization of this work resulted in a single step process to react
vinyltrimethoxy silane to soybean oil. Three systems were evaluated, thermal addition, a
peroxide initiated addition, and a Lewis acid catalyst. The effects of reaction time,
catalyst type, and reaction temperature were investigated. The silated soybean oil was
characterized using iodine value, FT IR spectroscopy, lH-NMR, l3C-NMR and TGA.
Characterization suggested that 3 moles of silane were added to each TAG by an “ene”
addition. Soybased thermosets were prepared using a room temperature moisture
activated cure of these silane functionalized soy oils, and degree of swelling in a series of

solvents was investigated.

120

7.5 REFERENCES
1. J. Baghdachi, J. and Li D., Soybean oil-based non-isocyanate containing moisture cure
resins and coatings. Proceedings of the International fiaterbome, High-Solids,aid

Powder Coatings Symposium (2001), 28th 231-233.

2. Reaction product of a silane and an unsaturated fatty substance and its use in coating
compositions for siliceous materials. GB patent No. 1129219 (1968).

121

 

Chapter 8: Overall Conclusions and Recommendations

8.1 MALEIC ANHYDRIDE FUNCTIONALIZED SOY OIL

New value added product was developed from soybean oil for use as a chemical
feedstock. The investigation and optimization of this work resulted in a simple and fast
process to maleate soybean oil. Anhydride functionality was introduced to soybean oil
through a free radical initiated maleation. Two initiators were evaluated, 2,5-bis(tert-
butylperoxy)-2,5-dimethylhexane peroxide and di-tert-butyl peroxide. The effects of
reaction time, initiator concentration, maleic anhydride concentration, and reaction
temperature were investigated. The maleated soybean oil was characterized using acid
value, iodine value and FTIR spectroscopy. The acid value was directly related to the
initial concentration of maleic anhydride, whereas the concentration and type initiator had
little effect on the acid value. The peroxide initiated functionalization of soybean oil with
maleic anhydride in a closed vessel at elevated pressure and temperature was found to
proceed by a Diels-Alder mechanism.
8.2 BIOFIBER REINFORCED SOYBEAN COMPOSITES

Composites with good mechanical properties were prepared from chemically
modified soy oils and biofibers without additional petroleum-based polymers. These
composites were prepared from maleic anhydride and epoxide functionalized soybean
oils that were cured in the presence of various biofibers (e.g. kenaf, kayocell, protein
grits, and solka-floc) by a flexible amine catalyst . Rigid thermosets characterized by high
crosslink density network and a high gel fraction were obtained. FI‘ IR was used to follow
the cure reaction via the disappearenace of the characteristic anhydride adsorptions.

Composites having high tensile strength and low elongation were obtained when kenaf

123

 

fibers were treated with (2-aminoethyl)-3-aminopropyl-trimethoxysilane and then added
to the epoxidized/maleated soy matrix and cured with hexamethylene diamine. These
biobased composites could provide inexpensive epoxy resin alternatives for a wide
veriety of industrial applications.
8.3 SYNTHESIS OF SOY POLYOLS VIA OZONE CHEMISTRY

New soy-based polyols were prepared by passing ozone through a solution of
soybean oil and ethylene glycol in the presence of an alkaline catalyst. Under these
conditions the double bonds in the unsaturated fatty acids are cleaved and the ozonide
intermediates that are formed react with one hydroxyl group of the glycol through an
ester linkage leaving the other hydroxyl group unreacted. The product of this reaction is
composed of triglycerides containing primary alcohols and a mixture of low molecular
weight diols and mono-alcohols. The latter are derived from cleaved fragments of the
unsaturated fatty acids beyond the double bonds. Due to the position of the double bonds
in the unsaturated fatty acid residues of soy oil, the primary triglyceride products contain
a nonanoate ester. The composition of the soy polyols mixture was calculated based on
statistical analysis and characterized by hydroxyl value, iodine value, C-NIVIR, and FT IR
spectroscopy. The effects of experimental conditions with various catalysts were also
investigated. A portfolio of soy polyols were prepared with varying hydroxyl
functionalities (varying reaction time) and molecular weight (varying hydroxy compound
and concentration) were prepared. Rigid soy-based polyurethanes were successfully
prepared using 100% of the soy polyols. Whereas in most current commercial
technologies soy polyol normally contribute 20—30 wt%, with the majority still petroleum

based.

124

8.4 EVOLUTION SOY POLYOL PRODUCTION

There have been 4 major milestones achieved since establishment of the proof-of-
concept. Initial preparation of soy polyols were done in small quantities (~20 grams) in a
chlorinated solvent at low temperatures (-70°C). The solvent was used to minimize the
concentration of the ozonides for safety concerns. However, use of the solvent did allow
for reactions to be run at low temperatures to increase ozone solubility. Commercially, it
would be beneficial to have a process which does not require solvent or low
temperatures.

The first significant milestone was achieved by removal of the solvent and room
temperature synthesis. Concurrently, larger batches of soy polyol were being prepared
(~1000 g). It was recognized that a key variable was the mass transfer of ozone through
the reaction median. A mass balance around the system showed that only 30% of the
ozone fed was reacting, while the majority was exiting the system into the destruction
unit. A preliminary economic evaluation showed the polyol production cost was greatly
influenced by the energy associated with producing the ozone. Running reactions at low
temperatures was advantageous for safety and enhanced mass transfer, since ozone
solubility is inversely related to temperature. However, it was determined to be more
economical running at room temperature because of the high cost of dry ice.

The second significant milestone was the successful application of a heterogeneous
catalyst. Initial studies were done using NaOH as a catalyst. Some difficulties associated
were the breakdown of the triglyceride backbone and difficulties removing residual
NaOH. Next, different amines were considered as catalyst. Amines were selected

because the target application for the soy polyols is polyurethanes, and since the amines

125

are also used to catalyze urethane chemistry; the goal was use the same catalyst for both
applications. This would be beneficial because it completely removes a unit operation for
removal of the catalyst. However, there was severe discoloration of the polyol overtime
which was attributed to the oxidation of the amine. Heterogeneous catalysts were
considered for ease of removal. Alkaline solids such as CaCO3 and MgO were evaluated,
and both were found to be effective at redirecting the decomposition of the ozonides to
form polyols. For cost and availability CaCO3 was determined to be the preferred
catalyst.

The third milestone was the development of soy polyols applicable for use in flexible
polyurethane. These polyols were developed by increasing the amount of C--C linkages,
to allow for more flexibility within the molecule. This was done using 2 different
approaches. The first approach involves preparing the polyols using different molecular
weight PEG. These polyols were prepared using CaCO3, and PEG400, 500, and 600
instead of and/or in combination with ethylene glycol. When used in polyurethane
synthesis, these new polyols were found to produce a more resilient and flexible foam, as
apposed to those prepared using only ethylene glycol. Another scheme to increase the
molecular weight of the polyols is to run the reactions near stoichiometric with ethylene
glycol. Experimentally, it was found that as the reaction progressed and the
concentration of ethylene glycol reduced, oligomerization reactions occur. This was
found to be the preferred method because of the increased soy oil content (increased
biobased content) and economically advantages since soy oil is less expensive. The
newest developed polyol using this approach had a MW = 4100 g/mol and a hydroxyl

value of 48. This newly prepared polyol could be a sustainable alternative to an industry

126

 

standard, DOW’s Voranol 4701. Commonly used for flexible foam applications the
petroleum based polyol has a MW = 4800 g/mol and a hydroxyl value of 34.

The fourth milestone is the development of the continuous production of soy
polyols. A baffled CSTR with 3 impellers was designed to achieved enhance mass
transfer of ozone within the system through creating turbulent flow.

8.5 MODIFIED SOY OIL BASE OIL LUBRICANT PROPERTIES EVALUATED

The oxidative stability of soybean oil was enhanced by transforming the double bonds
into more oxidatively stable esters by two chemical modifications. The first modification
was done by incorporation of anhydride functionality in soy oil in the presence of a
peroxide, followed by an esterification using alcohol. A series of alcohols with
increasing molecular weight were considered, and it was found that butanol was the most
stable of the series. The second modification was done using a catalytic ozonolysis in the
presence of an alcohol. A series of alcohols with increasing molecular weight were
considered, and it was found that the use of methoxy-PEG350 resulted in the most stable
product for the series. Both modifcations were also found to reduce the tendency to form
deposits and improve lubrication performance in comparison to the untreated soy oil.
8.6 SILANE FUNCTIONALIZED SOY OIL

Silane functionalities were introduced to soybean oil in a single step process by
reacting vinyltrimethoxysilane with soybean oil. Three systems were studied and the
effects of reaction time and reaction temperature were investigated. The use of catalyst
did not greatly affect the progress of the reaction as the reaction products from thermal
addition, peroxide initiated addition, and Lewis acid catalyst did not vary. The silated

soybean oil was characterized using iodine value, FTIR spectroscopy, lH-NMR, 13C-

127

 

NMR and TGA. It was determined that 3 moles of silane were added to each triglyceride.
A room temperature moisture activated cure of these silane functionalized soy oils were
used to prepare soybased thermosets, and degree of swelling in a series of solvents was

investigated.

8.7 RECOMMENDATIONS

Ozone Mediated Synthesis
0 Soy polyol with reduced moisture content of soy oil.
0 Prepare amine (ethanolamine) and acid (water) functionalities to soy oils
0 Optimization of continuous polyol production

Application of Polyols
- Evaluation in unsaturated polyesters
0 Formulation of flexible foam polyurethanes
0 Evaluate in concrete preparation

Soy-based lubricants
o Formulation using additives technology.

Silated Soy Oil
0 Synthesis using low sat soy oil to reduce sol content

128

 

Appendix A: Verification with a Model Compound

EXPERIMENTAL PROCEDURE

Ozonolysis of Soybean Oil

Methyl soyate (20 g), methanol (40 ml), methylene chloride (150 ml) and triethylamine
(0.4 M in CH3OH) were placed in a 500 ml gas wash bottle and cooled to -75°C in an
isopropanol dry ice bath and then ozone was bubbled through the reaction mixture.
Samples were taken as the reaction proceeded as noted. The product was then washed
with excess distilled water 5 times in order to remove any excess methanol and catalyst.
The first 2 washes were slightly acidic (0.2M HCl) to neutralize the base. A rotary
evaporator was then used to strip off the solvent and residual methanol. The product was

then dried over molecular sieves for 48 hours prior to testing.

Characterization Procedures

GM-MS were done at the MSU Mass Spectrometry Facility

RESULTS
The reaction products using triethylamine was similar to those formed when

N aOH was used, and these products could be match to the NIST database.

129

 

0 Methyl Oleate (C18: 1 ; 9C)

H 30_0/U\/\/\/\/\/\/\/\/\CH

03
l NaOH/CH30H/CH2C12

3

 

O O O
H C /u\/\/\/\/“\ CH Hsc WOW
3 \o 0/ 3 \o
Dimethyl Azelate Methyl Nonanoate
0 Methyl Linoleate (C18:2 ;9C, 12C)

H39%,JJ\/\/\/\/\/\/\/\/\CH

03
l NaOH/CH3OH/CH2C12

3

O O
H C /u\/\/\/\/“\ CH
3 \O O/ 3
Dimethyl Azelate
O O O
H C M CH H C /u\/\/\
3 \o o/ 3 3 \o CH3
Dimethyl Malonate Methyl Hexanoate

 

Methyl Linloleneate (C 18:2 ; 9C, 12C, 15C)
0

Hfi—OJK/VVWAMACH

03
l NaOH/CH3OH/CH2C12

3

O O
H C /“\/\/\/\/U\ CH
3 \o o/ 3
o 0 Dimethyl Azelate 0
2 H C M CH H C /“\/CH
3 \o o/ 3 3 \o 3
Dimethyl Malonate Methyl Propanoate

Figure Al: Ozonolysis of Methyl Soyate in Triethlamine and CH3OH

130

 

Absorbance . Solvent

 

 

 

 

 

 

 

 

 

 

 

22000000~ g
J !:
200000004 1
4
180000001 i STD ML
16000000~
14000000- 5
4 l
120000004 l;
MO g;
a I i
100000001. 1 l i;
« l i £2
MP l
80000004 1 f 3
. l I ll :
i I. i: :
60000004 a i . l
. i l i
1: i':
4000000~ ‘ I! ‘ MLn
* , E1 M8 l l I
20000004 _ . if; {a l ,3 = ;e
' 'A '. i l l J
‘ i 1% 1 ll 2 J: l i ’1
i '= I i ___. .._.___) k \.,_,-' ‘ i
110 . Y Y A ‘1? TV Y L.__: 5 Y i Y r , r r— t Y I Y W T Y 5 I 1 1—
8am 500 1000 I 1500 2000 I l _
Illtlui‘lltllllllilillillllilll|il’llgl'||lll.' i.l|ll|ll:l1|llllv|tilllllIl IIIIIIIII IIIIDII'I
Min. 5 10 15 20 25 30 35 4o 45

Figure A2: GC-MS of Methyl Soyate

131

Solvent

Abundance

l.|y I .ulylullylilll . ill

24000000-

21000000 §

18000000}

 

l

m |III lIIIIIlIII III :1!
l}... ”I: .IIIHJ Hrllillvll' it

1

 

 

 

15000000-i

 

12000000«
90000001?
6000000-
aooooow

TIC
Scan

10

Min.

GC-MS of Ozonolysis Products

Figure A3

132

Scan: 926 R.T.: 17.47 Base: m/z 74; 36%

100— "'0

% i . 87.0

20.: 55.0 129.0 141
1.48") 172.0

59.0
'4 63") 83.0 98119”J 115.0 11140
c . 111" 111111 rillt I #Ll Y rmxk ‘ g? r "11'
l 7 . I 1 l

T ‘ I I I T '
mlz 60 80 100 120 140 160 180

 

 

 

1r“

NIST MS 1 of 40 (1731-84-6) Base: m/z 74
Nonanoic acid, methyl ester

100—

1

74.0

2 /O\/\\/\/\/\
30i l l
O

 

a:
9

.32
?1Lllxlll

ilLLrL

N
0

141.0

 

 

 

129.0
A 11011.0 A11'SIO 1

 

I . .‘ ' l . ' r r T i i 7
W2 60 80 100 120 140 160 180

Figure A4: Mass Spectra of C9 Ozonolysis Products

133

Scan:1998

1001

% : 55.0 74083.0

 

mlz 50

MS! MS 1 of 40 (1732-10-1)
Nonanedioic acid, dimethyl ester

1

55.0
74.0

83.0

 

 

 

 

111.0

 

111.0

 

 

R.T.: 35.38 Base: m/z 152; 12%.

152.0

143.0

 

 

185.0

 

 

l T 'T' r ' r‘rrr I

150 200 250

Base: mlz 152
152.0
/0Y\/\/\/\n/O\
O O
1 85.0

143.0

 

150

 

l I 171.0 I“
1 ‘ I Y I I V l' f

 

1
200 250

Figure A5: Mass Spectra of C9 Base Fragment

134

 

 

 

 

 

 

Scan: 747 Rm: 15.53 Base: ".1; 74; 37.1%}.‘5
1; 74.0
:1
401 87.0
’ 99.0
2 59.0
55.0 I 710 1
or. ._. , 5.11.11, . .- .1; ix 1. . Page , . . .mgL‘PS-PT, . . ,. . . .133". . . . , . -
mlz 50 60 7o 80 90 100 110 120 130 140
NISTMSIOI‘W (106-70-7) Base: W274
Hexuoic acid, methyl ester
74.0
1
O\/\/\
”1 / ll /
601 0
% j
4 ' 87.0
59.0
2 55.0 99.0
1‘ L 71.0
-5.0
o‘liLv 11%111"? [1:177' 17 11 r1 %rrfr iii 1.! {Tfi 1 r1 1 Yr] 1 v 1 tTfiu
m/z 50 60 70 80 90 100 110 120 130 140

Figure A5: Mass Spectra of C5 Fragment

135

 

Appendix B: Ozonolysis in an alkaline median
PROPOSED CHEMISTRY
Ozone reacts with double bonds such that the first step is a 1,3-dipolar cycloaddition,
forming a primary ozonides (Figure Bl).

o’o‘o
\ /

 

=C\ + 03

Primary Ozonide
Figure B1: Formation of Primary Ozonide
The primary ozonide is a very unstable and decomposes immediately by selective
cleavage of the carbon-carbon bond and one oxygen-oxygen bond leading to a carbonyl

compound and a peroxidic carbonyl oxide, as a second fragment (Figure B2).

 

o/O\o 0/0
\C C/\ ———> I + n
/| I /(E\ / \
Carbonyl Oxide
+ 0/0 + 9
("3 e e O’C—
/ \ |

Figure B2: Ozonolysis Reaction Mechanism

136

 

This carbonyl oxide is a very reactive and can react in various routes, all of them

generating peroxides. Some of the possible carbonyl oxide reactions are shown in

ix

0 Secondary Ozonide o/
/U\O_ I . /l\
Carboxylic Acid \ C)/O / Hydroperoxide
1.
/ \
0—0
x >< ire/Vi
O—O n

Dirneric Peroxide Polymeric Peroxide

Figure B3.

Figure B3: Possible Reactions of Carbonyl Oxides

A reaction with the carbonyl compound forms as a second fragment to yield secondary
ozonides. The resulting 1,2,4-trioxolanes are more stable than the primary ozonides.
Another reaction is with protic molecules (e.g. solvent) to form hydroperoxides. Also,
possible is a rearrangement to carboxylates, as well as dimerization and polymerization.

However, due to the large oxygen content of these molecules, the carbonyl oxide,
secondary ozonides, dimeric and polymeric structures are very unstable and tend to
decompose. Polymeric peroxides are of particular concern because of their low solubility

and tendency to precipitate. Therefore for safety consideration maintaining low

137

 

concentrations of these species is important. This can be achieved by using a continuous
process or working in diluted solutions.

Nominally, the ozonoysis of methyl oleate would lead to a mixture of aldehydes
and carboxycylic acids (Figure B4). It was proposed by Marshall and Garafalo that under
basic conditions, the ozonolysis of alkenes in methanolic NaOH could redirect the

decomposition to form methyl esters (Figure B5).

 

\_\ O3 _.
0.
R1 0 O/ R/< )\
a 1 L —— ° “1
+ +
H R H/ \

R1

 

 

Figure B4: Ozonolysis of Methyl Oleate
The reaction pathway is hypothesized to be by formation of an aldehyde and carbonyl
oxide from the primary ozonide, reacting with methoxide and methanol to form a
hemiacetal and hydroperoxide. The former undergoes a base assisted hydride abstraction

by 03 and the latter dehydrates to form a methyl ester (Figure B5).

 

 

 

 

 

1
—’ R o—cn3
/0 NaOH O/EH)
l C1130H fi "OH
+ R——H
H/ \R 0 K)
\CH3

Figure B5: Ozonolysis of Methyl Oleate in the Presence of Methanolic NaOH

138

Appendix C: Scale-up Process Economics for Polyol Production

 

 

 

Reaction Stoichiometry Polyol
Compund mols MW (g) weight (g) weight (lb) wt%
Soybean Oil 1 870 870 1.9162996 0.5337423
Ethylene Glycol 10 62 620 1.3656388 0.3803681
Calcium Carbonate - - - 0.3281938
Ozone 10 48 480 1.0572687
Water 10 18 180 0.3964758
Oxygen 5 32 160 0.3524229
Total 3.5903084 lbs
Raw Material
Compound weight (lb) wt% Cost ($llb) Cost
Soybean Oil 0.53374233 0.533742331 0.25 $0.1334
Ethylene Glycol 0.3803681 0.380368098 0.46 $0.1750
Calcium Carbonate 0.1 - 0.75 $00750
Oxygen 281 ft3 ~ $4.31 - 30.9509
1 lbs 51.3343
Manufacturing Cost
Ozone Required (g) 122.997791 g 03 required to treat 0.58 lbs of SO
Reaction efficiency (%) 0.8
153.747239 * g of 03 required
Time 4.1329903 hr at 0.62 g O3/min
Ozone consumes <0.07 kW per kg 03
0.00936486 $lhr to run the ozone generator assuming $0.06/kWhr
0.038704876 * energy cost to operate ozone generator per hr

Ozone Generator $200,000
Peripheral equipment: $800,000

5 years write-off. debit per year: $200,000
Production capacity ds: 500 lb/hr
Production hours per year: 6I32 hr

 

debit per lb ds 30.0652

 

 

Labor

three shifts per day, 24 hours

 

 

 

Production hours per year: 6132 hr
Personnel $10 per hr 3 men
Personnel per lb I $0.02]
Per lb product
Raw material $1.334 Raw material $0.383
Energy $0.039 Energy $0.077
Debit $0.065 Debit $0.065
Personnel $0.020 Personnel $0.020
Total $1.46 Total $05460
* using air
* assume doubling power to account for air compressor

 

 

 

139

 

Raw Data for Sensitivity Analysis

Assumptions:

* Energy cost is $0.06 per kWhr

* Ozone is generated from air

* Energy cost is doubled to account for the air compressor and heater

* Synthesis done in a PBR production cost ~$0.47l/lb (CaCO3 cost removed), else $0.546/lb

 

Effects of Extent of Reaction

 

~ Assume 80% ozone reaction efficiency

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

COST Reacted Double Bonds
$ 0.5460 4.6
$ 0.5292 3.6
$ 0.5124 2.6
$ 0.4956 1.6
$ 0.4787 0.6
Effects of Ozone Efficiency
COST efficiency Variation (%)
$ 0.8726 20% 0333333333
$ 0.8106 25% 0166666667
$ 0.7694 30% 0
$ 0.7399 35% 0.166666667
$ 0.7178 40% 0.333333333

 

 

 

 

 

 

Effects of Price Fluctuation on Raw Material Cost

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

COST EC (3) Variation (%)
3 0.6371 5 0.36 -0.2
$ 0.6561 5 0.41 -0.1
3 0.6751 5 0.46 0.0
5 0.6942 3 0.5 1 0.1
3 0.7132 5 0.56 0.2
COST SO (3) Variation (%)
5 0.6218 5 0.15 -0.4
5 0.6485 3 0.20 -0.2
3 0.6751 $ 0.25 0
5 0.7018 3 0.30 0.2
$ 0.7285 5 0.35 0.4
COST CC ($) Variation (%)
3 0.6651 5 0.65 -0. 133333333
$ 0.6701 5 0.70 -0.066666667
$ 0.675 1 $ 0.75 0
3 0.6801 $ 0.80 0.066666667
3 0.6851 $ 0.85 0.133333333

 

 

 

 

140

Appendix D: NMR for Silated Soy Oils

 

 

Figure D.l: lH — NMR Spectra of Soybean Oil

141

 

 

 

 

 

 

 

 

 

, LJ

L_..J

 

 

 

 

 

 

 

 

 

fiIlilfjlllllilrrlllll'

8 7 6 5 4

3

InIWTItrrltrrrjtrrr]

2 1 0

Figure D2: lH - NMR of Silylated Soybean Oil

142

 

       

 

 

 

1 r

.. 1’ i- 1L
1 .‘ i l -

  
     
   

 

 

 

 

 

 

 

 

 

. 51,.

'1 ;‘i'li' " ‘ f‘ijl .
g. . aural. . j i .1111 M.,-.. ‘ uli flunk} -1

1'"

140 120 100 80 60 40 20 ppm

Figure D3: 13C - NMR Spectra of Silylated Soybean Oil

143

 

     
     

 

 

 

J P .JL— 4%

 

 

.'v.‘_ i _,W 2.!“
r"1l' .L‘k

. ' , , ' al.; ; :_ '--‘,‘" ‘_..1;i,ai' upli-
140 120 100 80 6O

    
 

 

 

 

 

 

 

 

:! il ;-
iaikhiuhh" j
4O 20 ppm

Figure D3: 13 C - NMR Spectra of Silylated Soybean Oil

143

 

 

220

 

J.J.;“ L1.L-_L;‘..L .4“

180

L- 1- L.' _.. .'"..J...L.i

 

00 O 3

LL,

..1

I“ L _‘.._

14o

 

..-:,: '3 -‘3 ,1

100

..il.k!‘.l.L :. ‘

 

will:

0.

607

 

 

 

20 .5

..
I.
C
.0
I .‘-.
1
.- O
:C' | :
Q '0 H
.i “2..” .

 

 

 

 

Figure D4: HMBC Correlation Spectrum of Silated Soy Oil

144