I Dipped Natural Rubber Latex Thin Films

Home , Polysulfide, Sesquiterpene

Dipped Natural Rubber Latex Thin Films: Hypoallergenic Accelerator Formulations for Crosslinking, and Composites with Waste-Derived Fillers

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

Jessica Lauren Slutzky

Graduate Program in Food, Agricultural and Biological Engineering

The Ohio State University

2019

Dissertation Committee:

Advisor: Dr. Katrina Cornish

Dr. John Lannutti

Dr. Frederick Michel Jr.

Dr. Alfred Soboyejo

Copyrighted by

Jessica Lauren Slutzky

2019

ii Abstract

Bio-based polymeric materials are of great commercial interest to attain environmental sustainability. Natural rubber (cis-1,4 polyisoprene) (NR) is a commodity that is extensively used in industrial, consumer, and medical industries. Over 5,000 plants produce natural rubber, but over 90% of the world’s supply of natural rubber is extracted from one plant species: the Brazilian rubber tree, Hevea brasiliensis. Hevea natural rubber (NR) has a high molecular weight, and can be produced in yields sufficient to meet market demands. NR contains a high amount of allergic proteins that can cause severe allergic reactions. One alternative sources of NR can be derived from the shrub

Parthenium argentatum, commonly known as guayule. Guayule natural rubber (GNR) has a high molecular weight and does not contain proteins associated with allergic reactions, rendering it circumallergenic. However, previous work has shown that GNR and NR have differences in various properties, preventing GNR from being a direct substitute for NR in many applications and giving it advantages in others. Therefore, the proposed work focuses on creating bio-based elastomeric materials by optimizing vulcanization chemistries, as well as creating composites using fillers derived from waste streams.

Vulcanization chemistries of circumallergenic GNR and hypoallergenic NR

(made by removal of soluble proteins which reduce allergic potential) for thin film applications was optimized using the chemical accelerators diisopropyl xanthogen polysulphide (DIXP) and zinc diisononyl dithiocarbamate (ZDNC). DIXP and ZDNC do not induce Type IV allergic reactions such as contact dermatitis, a unique benefit in

ii comparison to other rubber chemical accelerators. The thin films were manufactured from natural rubber latex using traditional coagulation dipping methods onto stainless steel formers. The effects of chemical accelerator concentration on mechanical, rheological, morphological, and thermal properties of circumallergenic GNR and hypoallergenic NR thin films were investigated. Many of the GNR and NR thin films possessed mechanical properties superior to ASTM standards for surgical gloves and condoms. Multivariate models of mechanical properties as a function of film thickness and chemical accelerator concentration were generated to quantify differences between

GNR and NR thin films. Data analytic methods such as canonical correlation analysis further quantified differences in mechanical properties between GNR and NR thin films, with GNR having greater elongation at break than NR, but NR having a higher Young’s

Modulus and strength at break than GNR vulcanized films. Thicker films for both NR and GNR showed an increase in Young’s Modulus and strength at break. Scanning electron microscopic (SEM) analysis of GNR and NR thin films showed that smooth thin films without void spaces or defects were created. The glass transition temperatures and thermal degradation curves of GNR and NR thin films were determined to quantify differences in vulcanization, with GNR having a lower glass transition temperature than

NR vulcanized thin films. By comparing vulcanization chemistries of GNR and NR thin films, differences in thin film properties attributed to species origin can be quantified.

Commercial thin film products made from NR often contain fillers from non- renewable resources to improve mechanical properties and thermal stability. Fillers from agricultural and industrial waste streams were compounded into thin films, using traditional coagulation dipping methods. Fillers included guayule bark bagasse, carbon

iii fly ash, and calcium carbonate derived from eggshells, utilized at various particle sizes and loadings. The chemical accelerators used in these composites include zinc diethyldithiocarbamate (ZDEC), diphenyl guanidine (DPG), and dipentamethylene thiuram polysulfide (DPTT), which are traditional chemical accelerators associated with increased contact dermatitis risk, but create vulcanized thin films with superior stability compared to DIXP and ZDNC accelerators. The vulcanization chemistries in these films were not optimized in order to determine the sole effect of fillers on GNR and NR thin film properties. In addition, NR latex with and without soluble protein was utilized to determine how protein content impacts film properties. Due to the increased allergic potential of these films in comparison to those manufactured in the first method, these films have applications as industrial coatings and should not be implemented in medical or consumer applications.

Reinforcement of NR and GNR compounds were achieved using fillers that were nano sized, especially at loadings below 2 parts per hundred rubber (phr) of carbon fly ash. Adding fillers to GNR typically caused increased elongation at break, whereas NR had a decreased elongation at break. The differences in bulk mechanical properties of

NR and GNR compounds with fillers can be attributed to variances in the polymer-filler interaction; non-rubber components such as proteins and phospholipids vary between

GNR and NR and can affect surface activity of a filler. Variances in bulk mechanical properties of GNR due to different fillers are attributed to properties of the filler, including particle structure, size, bulk density, alkalinity, and surface activity. Particles of larger sizes, such as 300 microns, can provide texture to NR and GNR thin films, which could be utilized for the commercialization of industrial non-slip surfaces. These

iv results can assist in successful commercialization of GNR, and create more sustainable

NR and GNR composites.

Dedicated to my family.

vi Acknowledgements

I would like to thank my committee members Drs. Katrina Cornish, John

Lannutti, Frederick Michel, and Alfred Soboyejo. Their guidance and support made this dissertation possible. I would like to express my utmost gratitude for Dr. Katrina Cornish and the department of Food, Agricultural and Biological Engineering at Ohio State.

Thank you to my high school math teacher, Ferd Schneider, for encouraging me to pursue engineering. In addition, I would like to thank Dr. Louis Chicoine of Nationwide

Children’s Hospital for my first research job. I would also like to thank John Shepherd, who funded my undergraduate scholarship. Above all, I want to thank my family.

I would like to thank Ohio Agricultural Research and Development Center

(OARDC) and the Institute for Materials Research (IMR) for the funding of this project.

I would also like to thank the University of Akron Research Foundation for their mentorship in entrepreneurship.

vii Vita

June 2006……………………………………Graduated from Walnut Hills High School

Cincinnati, Ohio

June 2011……………………………………B.S., Food, Agricultural & Biological

Engineering, The Ohio State University

June 2011……………………………………B.S., Psychology, The Ohio State University

September 2011-August 2014………………Ohio Agricultural Research and

Development Center (OARDC), Graduate

Research Associate, Doctoral Student, The

Ohio State University

August 2014-May 2015…………………….OARDC, Graduate Research Associate,

Charles Thorne Memorial Associateship,

Doctoral Candidate, The Ohio State

University

May 2015- March 2018……………………..Research Scientist, Battelle Memorial

Institute, Columbus, Ohio

July 2018- present…………………………..Research Scientist, Checkerspot, Berkeley,

California

viii Publications

Cornish, K., Bates, G.M., Slutzky J.L., Meleshcuk A., Xie W., Sellers K., Mathias R., Boyd M., Castaneda

R., Wright M., Borel L., 2019. Extractable protein levels in latex products, and their associated

risks, emphasizing American dentistry. Biology and Medicine. 11:2 (7 pages). DOI:

10.4172/0974-8369.1000456.

Slutzky J.L., Baral N., Shah A., Ezeji T., Cornish K., Christy A., 2016. Acetone-Butanol-Ethanol

Fermentation of Corn Stover: Current Production Methods, Economic Viability, and Commercial

Use, FEMS Microbiology Letters. 363:6.

Chen B., Xue J., Meng X., Slutzky J.L., Calvert A.E., Chicoine L.G., 2014. Resveratrol prevents hypoxia-

induced arginase II expression and proliferation of human pulmonary artery smooth muscle cells

via Akt-dependent signaling, American Journal of Physiology - Lung Cellular and Molecular

Physiology. 307:L317-L325.

Fields of Study

Major Field: Food, Agricultural & Biological Engineering

ix Table of Contents

Abstract………………………………………………………………………………...... ii

Acknowledgments…………………………………………………………………....….vii

Vita……………………………………………………………………………………...viii

List of Tables……………………………………………………………………..…….xiii

List of Figures……………………………………………………………………………xv

Chapter 1: Introduction…………………………………………………………...……….1

Chapter 2: Statement of Problem………………………………………………………….3

Chapter 3: Literature Review…………………………………………………………...…6

3.1 Introduction to Polymers…………………………………………………..6

3.2 Introduction to Elastomers……………………………………………....15

3.3 Specific Elastomer Structure and Properties……………………………31

3.4 Elastomer Compounding………………………………………….……..61

3.5 Manufacturing Methods for Latex……………………………………….79

3.6 Manufacturing Methods for Rubber……………………………………..82

Chapter 4: Mechanical properties of Type I circumallergenic & Type IV hypoallergenic guayule natural rubber latex thin films…………………………………………………..86

4.1 Introduction………………………………………………………………87

4.2 Experimental…………………………………………………………….91

4.3 Results and Discussion…………………………………………….…….94

4.4 Conclusion…………………..………………………………………….103

Chapter 5: Mechanical and thermal properties of type I & type IV hypoallergenic Hevea natural rubber latex thin films.………………………………………………………..107

x 5.1 Introduction……………………………………………………………..108

5.2 Experimental……………………………………………………….…..112

5.3 Results and Discussion…………………………………………………116

5.4 Conclusions……………………………………………………………..126

Chapter 6: Canonical correlation analysis of type I and type IV circumallergenic guayule natural rubber thin films and type I and type IV hypoallergenic ultra-low protein Hevea natural rubber thin films.………………………………………………………………..128

6.1 Introduction……………………………………………………………..129

6.2 Experimental………………………………………………………..….133

6.3 Results and Discussion………………………………………….…..….139

6.4 Conclusions……………………………………………………………..143

Chapter 7: Mechanical properties of guayule natural rubber latex thin film composites with biobased fillers.………………………………………………………………...….145

7.1 Introduction………………………………………………………….….146

7.2 Experimental…………………………………………………………...150

7.3 Results and Discussion……………………………………………...….154

7.4 Conclusions…………………………………………………………….161

Chapter 8: Mechanical properties of Hevea natural rubber latex thin film composites with biobased fillers.………………………………………………………………………...162

8.1 Introduction………………………………………………………….…163

8.2 Experimental………………………………………………………..…165

8.3 Results and Discussion…………………………………………………171

8.4 Conclusions……………………………………………………………180

xi Chapter 9: Conclusion.…………………………………………………………………181

Chapter 10: Future Studies…………………………………………………………..…183

References………………………………………………………………………………184

xii List of Tables

Table 3. 1. Common elastomers……………..……….……………………………...….26

Table 4. 1. Latex Compound Recipe……………………………………………………91

Table 4. 2. Dwell time and average thin film thickness, SEM < 0.01 mm……………....94

Table 4. 3. Tensile data for GNRL thin films, which will be used for subsequent thermal analysis via DMA, DSC, and TGA……………………………………………………101

Table 4. 4. Glass transition temperatures obtained from DMA for GNRL thin films....100

Table 4. 5. Glass transition temperatures obtained from DSC for GNRL thin films.....101

Table 4. 6. Thermal degradation temperatures for GNRL thin films……………..…...102

Table 5. 1. Latex Compound Recipe…………………………………………………..112

Table 5. 2. Dwell time and average thin film thickness……………………………..…116

Table 5. 3. Tensile data for ultra-low protein Hevea NRL thin films, which will be used for subsequent thermal analysis via DMA, DSC, and TGA……………………..……..120

Table 5. 4. Glass transition temperatures obtained from DMA for ultra-low protein

Hevea NRL thin films………………………………………………………………..…122

Table 5. 5. Glass transition temperatures obtained from DSC for ultra-low protein Hevea

NRL thin films…………………………………………………………………………123

Table 5. 6. Thermogravimetric data for ultra-low protein Hevea NRL thin films……125

Table 6. 1. Latex Compound Recipe for Guayule Natural Rubber Latex………….….133

Table 6. 2. Latex Compounding Recipe for Ultra-low Protein Hevea NRL…….…….133

Table 6. 3. Dwell time and average thin film thickness……………………………..…135

Table 6. 4. Notation of different experimental data used for statistical analysis……....136

xiii Table 6. 5. Mixed linear and non-linear stochastic multivariate regression models using film thickness (mm) as the significant predictor…………………………….………….139

Table 6. 6. Mixed linear and non-linear stochastic multivariate regression models using

DIXP (phr) as the significant predictor…………………………………………….…...140

Table 6. 7. Mixed linear and non-linear stochastic multivariate regression models using

ZDNC (PHR) as the significant predictor………………………………………………140

Table 6. 8. Canonical correlation values for multivariate models of tensile properties.142

Table 7. 1. Latex Compound Recipe…………..………………………………………151

Table 8. 1. Filler bulk densities………………………………..………………………167

Table 8. 2. Latex Compound Recipe……………………………...…………………..168

xiv List of Figures

Fig. 3. 1: Macromolecular structure of amorphous (A), crystalline (B), and semi- crystalline polymers (C)………………………………………………………………..….7

Fig. 3. 2: Free volume of a polymer as a function of temperature………….………..…...8

Fig. 3. 3: Stress-strain curve of polymeric materials……..………………………..…….12

Fig. 3. 4: Polymer crazing…..…………………………………………………………...12

Fig. 3. 5: SEM of fracture surfaces for a brittle fracture of natural rubber (A), and ductile fracture in a natural rubber-thermoplastic blend (B) …………………………...……….14

Fig. 3. 6: Stress-strain curve of elastomer thermosets with increasing crosslink density.16

Fig. 3. 7: Ideally elastic polymers represented by Hooke’s Law……………...... 18

Fig. 3. 8: Kelvin-Voight model of viscoelastic materials……….……………………....21

Fig. 3. 9: Creep deformation of polymers………………………………………...….….21

Fig. 3. 10: Loading curve and associated hysteresis……………………..……………...23

Fig. 3. 11: Typical stress relaxation curve of viscoelastic materials………...………..…24

Fig. 3. 12: Polymer chemical structure, consisting of carbon backbone and pendant groups, such as methyl and aromatic rings………………………………………………25

Fig. 3. 13: Polymerization of polybutadiene rubber……………………………..……....33

Fig. 3. 14: Polymerization of polyisobutylene rubber………………………………..….31

Fig. 3. 15: Structure of isobutylene-isoprene rubber, or butyl rubber…………………...35

Fig. 3. 16: Structure of ethylene propylene diene terpolymer (EPDM)………….….…..36

Fig. 3. 17: Structure of styrene butadiene rubber (SBR)……………………………..….37

Fig. 3. 18: Polybutadiene linked through the 1-and 4-carbon atoms, and through the 1- and 2-carbon atoms…………………………………………………………………..…..39

xv Fig. 3. 19: Structure of polychloroprene…………………………….………………..…40

Fig. 3. 20: Structure of acrylonitrile butadiene rubber (NBR)…………………………..42

Fig. 3. 21: Structure of polysulfide Thiokol A…………………………………………..44

Fig. 3. 22: Structure of poly(vinylidene fluoride-co-hexafluoropropylene)……….. …..46

Fig. 3. 23: Structure of polydimethylsiloxane (PDMS)…………………………………47

Fig. 3. 24: Structure of styrenic block copolymer elastomers………………………...…49

Fig. 3. 25: Structure of polyamide thermoplastic elastomers……………………………51

Fig. 3. 26: Chemical reaction for polyurethane synthesis…………………………...…..54

Fig. 3. 27: The mevalonate pathway, which produces isopentenyl pyrophosphate (IPP), the monomer for cis-1,4-polyisoprene……………………………………………….…..57

Fig. 4. 1. 3D graphs of mechanical properties for guayule thin films…………………..97

Fig. 4. 2. Glass transition temperatures obtained from DMA for GNRL thin films…...100

Fig. 4.3. Differential scanning calorimetry of GNRL formulations……………………101

Fig. 4. 4. Thermogravimetric analysis of GNRL formulations………………………...102

Fig. 5. 1. 3D graphs of mechanical properties for ultra-low protein Hevea NRL thin films…………………………………………………………………………………….119

Fig. 5. 2. Glass transition temperatures obtained from DMA for ultra-low protein Hevea

NRL thin films…………………………………………………………………………122

Fig. 5. 3. Differential scanning calorimetry of ultra-low protein Hevea NRL formulations…………………………………………………………………………….124

Fig. 5. 4. Thermogravimetric analysis of ultra-low protein Hevea formulations………125

Fig. 7.1. Mechanical Properties of GNRL with eggshells…………………………...…155

Fig. 7.2. Mechanical Properties of GNRL with guayule bagasse………………………157

xvi Fig. 7.3. Mechanical Properties of GNRL with carbon fly ash………………………...158

Fig. 8. 1. Macro (solid line) and micro sized (dashed line) fillers’ particle size distribution……………………………………………………………………………...166

Fig. 8. 2. Particle size distribution of nano sized fillers………………………………..171

Fig. 8. 3. Transmission electron micrographs of nano sized fillers……..……………...172

Fig. 8. 4. Tensile Properties of Eggshell-NRL composites……………………...……..174

Fig. 8. 5. Tensile Properties of Carbon fly ash-NRL composites……………………...176

Fig. 8. 6. Tensile Properties of guayule bagasse-NRL composites…………………….178

xvii Chapter 1: Introduction

Thin film elastomers are extensively used in a variety of applications including industrial coatings, consumer coatings, and medical products such as surgical balloons, gloves, and condoms (Cornish et al., 2007). Natural rubber (NR) is predominately used in thin film applications due to its superior stretch and softness compared to synthetic elastomers such as nitrile. However the proteins found in NR derived from the Brazilian rubber tree, Hevea brasiliensis, are severely allergenic and associated with Type I IgE- mediated allergies (Cornish, 2012; Hamilton and Cornish, 2010; Siler et al., 1996). As a result, synthetic elastomers became widely used in thin film elastomer applications despite their inferior mechanical properties (Cornish, 2012; Cornish et al., 2007). NR from other plant sources, such as the shrub Parthenium argentatum, commonly known as guayule, is of commercial interest partly due to its lack of allergic proteins (Hamilton and

Cornish, 2010; Siler et al., 1996). However, a key step to increasing the commercial potential of guayule natural rubber (GNR) includes rigorous material characterization to determine its differences to NR and define niche applications where GNR is the most ideal material to use.

Key differences between NR and GNR include polymer macromolecular structure and biochemical composition (Monadjemi et al., 2016). Macromolecular structural differences between NR and GNR are attributed to differences in plant species metabolism and biosynthesis of isoprene units, which are still undefined (Puskas et al.,

2014). Biochemical components entrapped during the extraction of GNR and NR include proteins, polyphenols, alkaloids, fatty acids, and phospholipids, which vary among plant species (McMahan et al., 2015). As a result, despite natural elastomers being comprised

1 of cis-1,4-polyisoprene, there are vast differences in structure and chemical composition which have a subsequent impact upon polymer compounding, manufacture, and final product properties (Monadjemi et al., 2016).

Therefore, the engineering and characterization of GNR must be compared to NR in order to define material differences attributed to species origin. GNR is known to have a more linear polymer structure compared to NR (Hager et al., 1979). The lack of polymer branching in GNR decreases molecular entanglements compared to NR. This allows for greater chain slippage and higher ultimate elongations. Since GNR has fewer chain entanglements than NR, crosslinking optimization is different between GNR and

NR to avoid brittle fracture. In addition, biochemical components entrapped in aqueous natural rubber can interfere, as well as improve crosslink formation (Cornish et al., 2007).

For example, proteins act as a natural surfactant and can improve the incorporation of hydrophilic compounding chemicals into the hydrophobic, hydrocarbon structure of cis-

1,4-polyisoprene (McMahan et al, 2015). Other biochemical components such as phospholipids can initiate stress-induced crystallization behavior of GNR and NR, which subsequently impacts tensile strength (Cornish, 2001; Steinbuchel, 2003). Therefore, it is imperative to optimize GNR and NR independently for a given manufacturing process due to their vast differences, despite both being comprised of high molecular weight cis-

1,4-polyisoprene.

2 Chapter 2: Statement of the Problem

Current global demand for elastomers exceeds the natural rubber supply, which is supplemented by synthetic elastomers derived from petroleum (Cornish, 2014). Global demand for elastomers will continue to grow, especially in emerging markets such as

India and China as consumers in those markets gain purchasing power (Cornish, 2014).

Currently, over 90% of the global natural rubber supply is derived from Hevea brasiliensis, the Brazilian rubber tree, which is predominately grown in tropical regions of Southeast Asia, thus requiring extensive exports for rubber products in North America and Europe (van Beilen and Poirier, 2007). In addition, Hevea natural rubber (NR) and natural rubber latex (NRL) is capable of eliciting severe type I allergic responses, attributed to allergic proteins entrapped during the extraction process (Cornish et al.,

1999). Type I allergies to NR and NRL proteins have caused great concern in the medical industry, prompting bans of NR and NRL products in many hospitals. Current commercial alternatives to NR and NRL include nitrile and other synthetic elastomers, which often have inferior mechanical properties compared to NR and NRL (van Jole,

2007). As a result, alternative, renewable materials to alleviate type I sensitization and subsequent allergic reaction need to be developed to promote consumer safety and environmental sustainability.

The hypothesis of this work focuses on finding applications of guayule natural rubber latex (GNRL) and hypoallergenic NRL for consumer and medical industries by optimizing chemical compounding and manufacturing techniques, while providing a comparative analysis to NRL products. Understanding the differences between GNRL, hypoallergenic NRL, and traditional NRL will provide insight into differences in

3 intermolecular interactions between rubber molecules and cross-linking chemicals, as well as solid fillers. Once these latex chemical formulations are optimized, they can be processed via coagulated dipping to create thin film elastomer finished products.

Examples of finished products include surgical balloons, medical gloves, condoms, catheters, dental dams, and biocompatible coatings.

Fillers are often used in compounding of natural rubber latex, to make natural rubber latex composites. Fillers can provide a reinforcing effect, improving mechanical properties. This work will also focus on evaluating waste-derived agro-industrial residues for natural rubber latex composites, for both guayule and Hevea.

Characterization of mechanical properties of NRL films attributed to type of filler, particle size, and loading will be assessed in different natural rubber lattices.

The overall goal of this work is to promote natural rubber latex products, through improved mechanical properties attributed to cross-link optimization, and the use of new fillers for natural rubber latex composites. This goal will be accomplished via the following objectives:

• Objective 1: Understand the physico-chemical behavior changes of vulcanized

GNRL, hypoallergenic NRL, and stabilized NRL containing soluble protein,

using the chemical accelerators diisopropyl xanthogen polysulfide (DIXP) and

diisonyl dithiocarbamate (ZDNC), therefore eliminating a type IV, contact

dermatitis allergy.

• Objective 2: Optimize cross-linking and natural rubber latex mechanical

properties using mathematical modeling.

4 • Objective 3: Engineer new natural rubber latex composites using low-cost,

waste-derived material.

5 Chapter 3: Literature review

3.1. Introduction to Polymers

A polymer is molecule containing chemically bonded repetitive units, termed monomers. Polymeric materials have molar masses in excess of 103 g/mol, and are larger than 1 nm in size (Young and Lovell, 1991). The macromolecular morphology of polymers varies due to chemical composition and secondary bonding, with the resulting three-dimensional aggregate macromolecular structure dictating physical properties

(Young and Lovell, 1991). Polymer molecules generally pack together in a non-uniform fashion, creating a combination of ordered, crystalline regions mixed with disordered, amorphous regions (Figure 3.1). Crystallinity occurs when linear polymer chains are structurally oriented in a uniform three-dimensional matrix, with polymer chains extending out from crystalline domains into amorphous regions where they are coiled and tangled (Young and Lovell, 1991). The extent, or degree, of crystallinity in a polymer depends on polymer chain length, branching, and secondary bonding (such as hydrogen bonding, and dipole-dipole interactions) (Young and Lovell, 1991). Some polymers lack crystalline regions, and are completely amorphous.

Fig. 3. 1: Macromolecular structure of amorphous (A), crystalline (B), and semi- crystalline polymers (C)

3.1.1. Thermal Transitions of Polymers

Material specific thermal transitions such as the glass transition temperature (Tg) and the melting temperature (Tm) provide insight into polymer morphology. Tm is the temperature at which crystalline domains lose their structure and melt (Young and

Lovell, 1991). Tm is indicative of the degree of crystallinity, as well as the type of crystal structure in a polymer (Young and Lovell, 1991). Tg is the temperature above which amorphous domains of a polymer are capable of segmental chain motion; below the Tg a polymer is capable of translational motion only and behaves like a brittle, rigid glass

(Young and Lovell, 1991). At temperatures above the Tg, the polymer is capable of segmental and translational motion, resulting in viscoelastic rubbery behavior (Young and Lovell, 1991). Tg varies among polymers due to differences in the small amount of unfilled volume associated with the end of a polymer chain, termed free volume (Young

7 and Lovell, 1991). As temperature of a polymer system is increased, the amount of free volume increases drastically above Tg (Figure 3.2), which is attributed to segmental motion of the polymer chains (Young and Lovell, 1991). Polymers with large steric bulk from pendant groups will have a higher Tg due to increased thermal energy required for segmental motion (Young and Lovell, 1991).

Fig. 3. 2: Free volume of a polymer as a function of temperature.

3.1.2. Introduction to thermoplastics, thermosets, and elastomers

Polymeric materials are classified into thermoplastics, thermosets, and elastomers based upon their material properties, which are dictated by macromolecular structure and ultimately chemical composition (Dick, 2009).

Unprocessed thermoplastics are pellets, powders, or granules and become viscous during heating which allows them to be molded. Thermoplastics are viscous and

8 moldable above a specified Tm, and solidify when cooled below the Tm. Thermoplastics have a Tm and a Tg that are attributed to the properties of their crystalline and amorphous regions, respectively (Young and Lovell, 1991). For example, thermoplastics with a higher degree of crystallinity have a higher Tm, requiring more thermal energy to melt their crystalline regions (Young and Lovell, 1991). Thermoplastics with longer chain lengths, more extensive chain branching, and stronger secondary bonding have a higher

Tg due to segmental chain motion being hindered, requiring more energy (Young and

Lovell, 1991).

Thermosets are polymers that are chemically crosslinked to create intermolecular and intramolecular bonds, and therefore lack the ability to be reshaped at any temperature

(Dick, 2009). The crosslinks in thermosets prevent chain slippage, which improve bulk mechanical properties (Dick, 2009). The type of crosslinker used in a thermoset depends on the type of active sites found in a polymer, such as unsaturated carbons or epoxy groups (Dick, 2009). Differences in chemical composition among thermosets can provide differences in morphology as well: some thermosets have crystalline regions, or are completely amorphous in structure. Thermosets that have crystalline regions have distinct Tm and Tg, whereas amorphous thermosets only have a Tg (Young and Lovell,

1991). Thermosets with crystalline regions, termed thermoset resins, are typically liquid chemicals with a low molecular weight and low viscosity that are polymerized into long chains and high molecular weight molecules with a high viscosity which are subsequently crosslinked to stiffen the thermoset (Young and Lovell, 1991). Thermosets that are completely amorphous in structure are typically large molecular weight molecules previously polymerized, and then subsequently crosslinked into a thermoset (Young and

9 Lovell, 1991). In both thermoset resins and amorphous thermosets, the finished thermoset product has a three-dimensional network of chemical crosslinks with better mechanical properties than thermoplastics (Dick, 2009).

Elastomers are characterized by their ability to be stretched and return to its original shape without permanent deformation. Elastomers can be thermoplastics, or lightly crosslinked thermosets. Elastomers are typically able to undergo reversible strain at temperatures above Tg due to external forces producing intramolecular conformational changes, or long-range segmental motion, in amorphous regions (Dick, 2009).

3.1.3. Mechanical Properties of Polymers

Stress-strain curves of macromolecular materials further distinguish thermoplastics, thermosets, and elastomers (Fig. 3.3). When subjected to a uniaxial load, thermoplastics deform reversibly (elastically) until a maximum stress, the yield point, is reached. Thermoplastics stressed beyond the yield point deform irreversibly, until the material ultimately fractures (Young and Lovell, 1991). Thermoplastics are capable of undergoing both ductile and brittle fracture, depending on the conditions of the mechanical stress (Young and Lovell, 1991). Thermoplastics stressed in reduced temperatures, with increased strain rates, sharp notches, and increased thickness are more susceptible to brittle fracture (Young and Lovell, 1991).

Polymers can undergo yielding mechanisms, such as shear yielding and crazing while undergoing mechanical stresses (Young and Lovell, 1991). Crazing is characterized by polymer chains rearranging in highly stressed regions under uniaxial loads, creating localized plastic deformations with microvoids connected by polymer fibrilliar bridges (Fig. 3.4) (Young and Lovell, 1991). These polymer fibrilliar bridges

10 eventually coalesce to form a crack, which can ultimately lead to mechanical failure.

Shear yielding is a yielding mechanism that occurs parallel to the direction of the force, whereas crazing occurs normal to the direction of the force.

Crazing in thermoplastics is attributed to changes in physical and chemical bonding (Calhoun and Peacock, 2006). Mechanical stress can cause the weak physical

Van der Waals forces between polymer chains to separate, creating a microscopic void space (Calhoun and Peacock, 2006). The covalent bonds in the polymer backbone chain compensate for the mechanical stress, reducing the polymer chain length and increase the void space initiated by disruption in the Van der Waals forces (Young and Lovell, 1991).

These void spaces are bridged by fibrils a few nanometers in diameter, which are molecules of the stretched polymer chain. Crazing occurs internally in a polymeric material, absorbing fracture energy and therefore improving the fracture toughness of the material (Young and Lovell, 1991). Crazes typically form in regions associated with flaws, and most frequently occur in amorphous, brittle polymers (Young and Lovell,

1991). Crazing can ultimately lead to crack propagation, and these cracks may grow until bulk mechanical failure occurs (Grellman and Seidler, 2001).

Fig. 3. 3: Stress-strain curve of polymeric materials.

Fig. 3. 4: Polymer crazing.

12 Polymers can undergo brittle or ductile fracture, depending on the failure conditions. Polymers favor brittle fracture with reduced temperatures, increased strain rates, geometric notches, and increased thickness (Grellman and Seidler, 2001). Brittle fracture in polymers occurs during high strain rate impacts, and is characterized by a smooth fracture surface (Grellman and Seidler, 2001). Polymers that undergo ductile failure display necking and permanent deformation, which is irreversible (Grellman and

Seidler, 2001). Necking in ductile fracture is attributed to polymeric molecular chains unfolding and re-aligning in the direction of the applied stress (Grellman and Seidler,

2001). Amorphous polymers undergo molecular uncoiling, followed by polymer chains alignment, causing the neck to propagate until polymer chain scission(Grellman and

Seidler, 2001). Crystalline polymers undergo chain unfolding in the amorphous regions between the lamellae of the crystals, followed by crystal fracture and re-alignment of crystalline regions(Grellman and Seidler, 2001). Amorphous polymers that are crosslinked into thermosets ultimately undergo brittle fracture, attributed to the covalent bonds of crosslinks being severed (Grellman and Seidler, 2001). The fracture mechanisms vary between elastomers depending on the respective microarchitecture of the material, which is a result of the aggregate structures based on chemical composition.

Fig. 3. 5: SEM of fracture surfaces for a brittle fracture of natural rubber (A), and ductile fracture in a natural rubber-thermoplastic blend (B) (Grellman and Seidler, 2001).

14 3.2. Introduction to Elastomers

Elastomers are materials that can exhibit a rapid and large reversible strain in response to mechanical stress, and therefore have high resiliency (Dick, 2009). Elastic strain can be attributed to chemical bond stretching, bond angle deformation, or crystal structure deformation (Dick, 2009). Elastomers are classified as thermoplastics with semi-crystalline regions, or amorphous polymers that utilize crosslinking to become thermosets (Dick, 2009). Amorphous regions of elastomers contain predominately amorphous structures oriented into random coils. However, there are crystalline regions in thermoplastic semi-crystalline polymers, and amorphous regions of elastomers can phase transform into strain-induced crystals (i.e., strain-induced crystallization) (Dick,

2009). Within amorphous regions, unstrained elastomers exist in a random, coiled structure. As strain is applied, mechanical energy is dissipated by the re-orientation of molecular chains into an uncoiled, aligned, crystalline structures within amorphous regions (Grellman and Seidler, 2001). Mechanical failure of elastomers occurs when chains are completely uncoiled and chemical bonds begin to break typically in very high strains of 750%-2000%, dependent upon the chemical composition and three-dimensional structure of the elastomer (Dick, 2009). Crosslinking of elastomers can hinder polymer chain uncoiling under strain by restricting macromolecular translational movement, thus improving mechanical properties and elastomer utility (Figure 3.6) (Dick, 2009).

Fig. 3. 6: Stress-strain curve of elastomer thermosets with increasing crosslink density.

3.2.1. Elastomer Thermodynamics

Thermodynamic behavior of elastomers is characterized by immediate and reversible strain, and is not time dependent (Mark et al., 2013). Elastomer thermodynamics are summarized by the Helmholtz free energy equation (derived from

Gibbs free energy), where free energy is calculated in terms of the change of length for an elastomer (Mark et al., 2013):

푑퐴 = 푑푈 − 푇푑푆

Where A = Helmholtz free energy

U = internal energy

T = Temperature

S = entropy

Entropy of an elastomeric system is determined by macromolecular conformations. Entropy is maximized when the elastomer is contracted, due to the infinite macromolecular conformations due to dynamic flexibility (changes in spatial orientation while in equilibrium) (Mark et al., 2013). In contrast, when an elastomer is elongated, the entropy of the system is minimized due to the macromolecular structure becoming more linear, and thus losing conformational freedom (Mark et al., 2013).

When an elastomer mechanically fails, the elastomer has minimized the entropy of the system (Mark et al., 2013).

Enthalpy, or the internal energy, is dictated by the thermal energy introduced into a system. If a stretched elastomer is heated, its length will decrease as the system dissipates thermal energy by increasing coiled macromolecular conformations (Mark et al., 2013). Similarly, if a stretched elastomer is allowed to contract, its temperature will decrease (Mark et al., 2013). In contrast, if an elastomer is stretched, its temperature will increase (Mark et al., 2013).

Ideal, purely elastic polymers obey Hooke’s law (Figure 3.7), equilibrating their macromolecular structure proportionally to the amount of strain applied (Mark et al.,

2013). Elastic deformations are attributed to polymer chain segmental motions, dependent upon free volume and are a thermodynamic phenomenon (Mark et al., 2013).

At low strains, elastomers more closely follow Hooke’s law due to impeding uncoiling of polymer chains attributed to macromolecular entanglement (Mark et al., 2013).

However, elastomers deviate from Hooke’s law at high strain due to the elongation of polymer chains, crosslinks’ bonds becoming strained, and the phase transformation of

17 amorphous regions into crystalline structures (Mark et al., 2013). Elastomers also deviate from Hooke’s law under severe compression, attributed to limited free volume (Mark et al., 2013).

Fig. 3. 7: Ideally elastic polymers represented by Hooke’s Law

The non-ideal behavior of elastomers is attributed to their viscous component which, in turn, is attributed to permanent or non-reversible deformation caused by translational motion of the polymer structure while under strain (Mark et al., 2013). This translational, segmental motion of polymers is dependent upon kinetics and causes deviations from ideal thermodynamics associated with a polymeric system (Mark et al.,

2013).

3.2.2. Elastomer Kinetics

Viscoelastic materials such as elastomers are non-Newtonian (non-ideal) and exhibit a temperature dependent, non-linear response to a strain rate (Mark et al., 2013).

Segmental motion of macromolecules is the cause of non-ideal behavior associated with

18 the viscous component of an elastomer (Mark et al., 2013). The Tg of a polymer determines the thermal energy required for segmental motions in macromolecular systems, defining the boundary between elastic and visco-elastic behavior (Mark et al.,

2013). When a polymer is below its Tg, segmental motions are not possible and the polymer is glassy and behaves elastically like an ideal (Bingham) plastic (Mark et al.,

2013). When the temperature of a polymer is above its Tg, long-range segmental motion occurs and the polymer behaves as an elastomer (Mark et al., 2013). Phenomena such as creep, hysteresis, and stress relaxation are attributed to the viscous component in polymers.

3.2.2.1. Creep

Creep is a time dependent increase in deformation under constant stress (Mark et al., 2013). The amount of creep is dependent upon material properties, exposure time and temperature, as well as the applied structural load (Mark et al., 2013). In general, if the stress is removed within a certain time (specified by TA in Fig. 3.9), the strain recovers partially very rapidly (specified by B in Fig. 3.9) and is termed elastic recovery (Mark et al., 2013). When a material is exposed to a stress below the material’s elastic limit beyond a certain time (specified by C in Fig. 3.9), permanent deformation occurs which is due to the slow viscous component of the polymer (Mark et al., 2013). Creep of polymers is an important mechanical characteristic to consider when characterizing dimensional stability.

Creep of polymers can be modeled using the Kelvin-Voight model, represented by a Hookean spring and a Newtonian dashpot in parallel (Figure 3.8). The creep strain is given by the following convolution integral (Mark et al., 2013):

19 ∞ 푡 (− ) 휀(푡) = 휎퐶표 + 휎퐶 ∫ 푓(휏) (1 − 푒 휏 ) 푑휏 0

Where:

σ = applied stress

Co = instantaneous creep compliance

C = creep compliance coefficient

τ = retardation time f(τ) = distribution of retardation times

Viscoelastic materials experience a time-dependent increase in strain when subjected to a step constant stress. The time-dependent increase in strain varies with different amounts of stress. The transition from a linear to non-linear viscoelasticity is characterized by a material specific amount of stress, termed the critical stress (Mark et al., 2013). Below the critical stress, the viscoelastic material has linear viscoelasticity whereas above the critical stress the polymer’s creep rate increases non-linearly (Mark et al., 2013). As constant stress is applied to the system, the material undergoes strain until the material fails. If the stress is maintained on the material for a short amount of time, the material behaves elastically. If the stress is held on the material above its critical value of applied stress, the creep modulus is dependent upon the stress applied (Mark et al., 2013). A viscoelastic creep modulus-time curve can represent multiple strain versus time responses for various stress loads that are under the material’s critical stress value.

Fig. 3. 8: Kelvin-Voight Model of viscoelastic materials.

Fig. 3. 9: Creep Deformation of Polymers

Polymers vary in their creep behavior, attributed to differences in secondary bonding and segmental motion between polymer chains due to variations in molecular

21 weight and chemical composition. In general, more thermally stable polymers, such as those with a higher molecular weight or containing aromatic rings, are more creep resistant (Mark et al., 2013).

3.2.2.2. Elastic hysteresis

Elastic hysteresis is characterized by the deformation of a material due to current stress and past stresses (Mark et al., 2013). In context of viscoelastic mechanics, force loading and subsequent unloading has a time-dependent behavior deviant from the ideal elastic behavior described by Hooke’s law. As forces are loaded onto an elastomer, energy is dissipated by extension of the elastomer. However, when that same force is unloaded, less energy is required by the elastomer to retain its original shape and the excess energy is dissipated as heat (Mark et al., 2013). The thermal energy dissipated by the system is due to internal macromolecular friction (Mark et al., 2013). As a result, elastomers are commonly used as stress dampeners due to an elastomer’s ability to absorb mechanical compression energy and dissipate this energy as heat (Mark et al.,

2013). The large area contained within the hysteresis loop (Figure 3.10) shows that an elastomer dissipates energy by internal macromolecular friction, compared to stiff materials that would have a smaller hysteresis loop (Mark et al., 2013).

Fig. 3. 10: Loading Curve and associated hysteresis.

3.2.2.3. Stress Relaxation

Stress relaxation describes how polymers dissipate stress under constant strain, whereas creep describes polymer behavior under a constant state of stress with an increasing amount of strain (Mark et al., 2013). Stress relaxation is attributed to viscous components in the polymer, and is characterized by a decrease in stress under constant strain (Mark et al., 2013). High stress relaxation rates are characteristic of poor elastic properties. A typical uniaxial stress relaxation curve (i.e. load-time decay curve or stress- time) for viscoelastic materials is shown in Figure 3.11. Let σ and ε be the stress and strain, respectively at position, S, where the loading process was stopped and the stress relaxation begins. The simplest theoretical description of relaxation as a function of time, t, is an exponential law (exp-t/τ) (Mark et al., 2013).

Fig. 3. 11: Typical stress relaxation curve of viscoelastic materials

3.2.3. Elastomer Structure

Polymer structure consists of repeating units, monomers, covalently bonded together. Organic polymer backbones consist of carbon to carbon single or double bonds, with pendant groups such as aromatic rings, or methyl groups attached (Figure 3.12).

Polymers assemble into a three-dimensional aggregate structure based upon their chemical composition and manufacturing conditions (Mark et al., 2013). Elastomers can be lightly crosslinked thermosets, or thermoplastics. Thermoset elastomers typically are amorphous polymer structures, with crosslinks to improve macromolecular translational hinderance and reversible strain (Mark et al., 2013). Many of the amorphous thermoset elastomers (Table 3.1) do not have commercial utility prior to crosslinking (Mark et al.,

2013). Thermoplastic elastomers are typically copolymers, made from two chemically distinct monomers, respectively containing continuous soft (amorphous) segments and

24 hard (crystalline) segments (Table 3.2). Thermoplastic elastomers typically have ultimate strains less than thermoset elastomers, due to their crystalline regions not being able to uncoil to dissipate stress like amorphous regions (Mark et al., 2013). However, unlike some thermoset elastomers, thermoplastic elastomers can be used in injection and other molding processes. Thermoplastic elastomers can also be engineered to optimize their ratio of soft and hard copolymer segments, enabling precise tailoring of properties for given applications (Mark et al., 2013).

Fig. 3. 12: Polymer chemical structure, consisting of carbon backbone and pendant groups, such as methyl and aromatic rings.

25 Table 3. 1: Common elastomers (Mark et al., 2013)

26 Elastomers are predominately hydrocarbon molecules that lack polar groups and do not possess strong intermolecular forces such as hydrogen bonding and dipole-dipole interactions (Mark et al., 2013). The chemical structure of the amorphous regions in elastomers is important due to its ability to provide reversible strain through polymer chain segmental motion under mechanical stress, typical of elastomers (Mark et al.,

2013). The minimal intermolecular forces among polymer chains allow for chain uncoiling (Mark et al., 2013). Rapid strain of elastomers requires minimal steric hindrance within the macromolecular structure (Mark et al., 2013). As a result, elastomer structures often lack large pendant groups in order to allow for rapid uncoiling of amorphous polymer structures when under strain (Mark et al., 2013). In addition, crosslinking is required for reversible strain in amorphous elastomers. Crosslinking inhibits translational motion, contributing to strain reversibility. Ideally, crosslinking to the extent of a percolation network (a random pathway of crosslinks throughout the bulk polymer material) provides reversible strain, whereas crosslinking beyond a percolation network (multiple, dense pathways of crosslinks) can decrease elastomeric performance and result in brittle fracture (Mark et al., 2013).

3.2.4. Elastomer Polymerization Processes

Natural rubber (NR) is enzymatically synthesized in vivo by rubber producing plants such as Hevea brasiliensis Muell. Arg. and Parthenium argentatum Gray (Cornish,

2014). All other elastomers are synthesized from petrochemicals that require the polymerization of monomers, a combination of monomers, or further modification of an existing polymer (Mark et al., 2013). Synthetic elastomers are homopolymers when derived from one type of monomer, such as polyisoprene or polybutadiene (Young and

27 Lovell, 1991). Copolymers are comprised of two or more different monomers, and include elastomers such as styrene-butadiene rubber (SBR) (Mark et al., 2013). There are four methods for elastomer polymerization: bulk, solution, suspension, and emulsion

(Mark et al., 2013).

3.2.4.1. Bulk Polymerization

Bulk polymerization utilizes a single monomer and a suitable catalyst, and is heated or pressurized to initiate polymerization. Bulk polymerization produces solid polymer through polymerization mechanisms such as polycondensation, free radical, or coordination (Mark et al., 2013). Solid polyurethane elastomers utilize a polycondensation mechanism, whereas ethylene acrylic elastomers use the free radical polymerization mechanism (Mark et al., 2013). Ethylene-propylene-diene rubber

(EPDM) uses a coordination polymerization method to create an ethylene-propylene

(EPM) copolymer that is subsequently terpolymerized to create EPDM (Mark et al.,

2013). In general, temperature and molecular weight control is difficult in bulk polymerization due to the solid state of the polymer and uncontrollable auto acceleration of the polymerization reaction (Mark et al., 2013).

3.2.4.2. Solution Polymerization

Solution polymerization uses an inert solvent as the medium for the monomers and catalysts. Catalysts in solution polymerization are soluble, or are finely suspended in solution. Solution polymerization allows for precise control of the polymerization reaction, with improved temperature and viscosity control compared to bulk polymerization (Mark et al., 2013). Catalyst type and concentration can dictate the molecular weight and structure of the final polymer, and allow some control of the auto

28 acceleration of the polymerization reaction (Mark et al., 2013). Solution polymerization uses polymerization methods such as polycondensation, free radical, cationic, anionic, and coordination (Mark et al., 2013). In addition to bulk polymerization methods, polyurethanes can be synthesized by solution polymerization using a condensation mechanism (Mark et al., 2013). Ethylene-vinyl acetate rubber (EVM) uses a free radical solution polymerization method (Mark et al., 2013). Epicholorohydrin rubber (ECO), and butyl rubber (IIR) use a cationic polymerization method, whereas SBR, polybutadiene rubber (BR), and polyisoprene rubber (IR) use an anionic polymerization method (Mark et al., 2013). Coordination mechanisms are used to create BR, IR, EPM, and EPDM in solution (Mark et al., 2013). Disadvantages of solution polymerization include removal of excess solvent from the finished polymer (Mark et al., 2013).

3.2.4.3. Suspension Polymerization

Suspension polymerization uses monomers and a monomer soluble catalyst suspended as monomer/catalyst droplets in water. Suspending agents are used in the aqueous medium, to prevent coalescence during polymerization. Polymerization occurs within the monomer droplets, creating larger polymer beads. The final polymer beads are insoluble in water, and can be retrieved by filtration (Vivaldo-Lima, et al., 1997).

Polymer mechanisms used in suspension polymerization include free radical, cationic, and coordination (Vivaldo-Lima, et al., 1997). EVM is made by a free radical mechanism, in solution and suspension (Vivaldo-Lima, et al., 1997). EPM is made by a coordination reaction in solution and suspension polymerization (Vivaldo-Lima, et al.,

1997). Disadvantages of suspension polymerization include auto acceleration, and

29 difficulties associated with filtering and isolating the polymer product beads (Vivaldo-

Lima, et al., 1997).

3.2.4.4. Emulsion Polymerization

Emulsion polymerization occurs in an aqueous medium, in micelles of monomers.

Surfactants are used to create an emulsion of monomers, catalysts, and modifiers.

Catalysts initiate the polymerization reaction, whereas modifiers can control polymer structure and molecular weight (Lovell and El-Asser, 1997). Stabilizers such as antioxidants are added to the emulsion, and unreacted monomers are recovered from the emulsion. This creates a final emulsion polymer product that is stable, and can be directly used in manufacturing of thin films and coatings or coagulated into solid state

(Lovell and El-Asser, 1997). Free radical polymerization is the method primarily used in emulsion polymerization, and is used to create rubbers such as SBR, chloroprene rubber

(CR), acrylonitrile-butadiene rubber (NBR), EVM, and fluorocarbon rubber (FPM)

(Lovell and El-Asser, 1997). Disadvantages associated with emulsion polymerization include difficulty in isolation of the polymer product, and removal of surfactants to generate pure polymer products (Lovell and El-Asser, 1997).

30 3.3. Specific Elastomer Structure and Properties

Natural rubber (NR), cis-1,4-polyisoprene, is the only naturally produced elastomer with high molecular weights sufficient for industrial use and has a completely amorphous structure, when not under strain (Tanaka, 2001). All other elastomers are synthetic, and can have an amorphous or semi-crystalline structure (Mark et al., 2013).

Synthetic elastomers are polymerized from petroleum byproducts, or alternatively are modified polymerized synthetic materials. Elastomers are further classified according to their utility: general purpose, solvent resistant, or temperature resistant (Mark et al.,

2013). General-purpose elastomers are predominately aliphatic and aromatic hydrocarbons with amorphous structures and include natural and synthetic elastomers

(Mark et al., 2013). General-purpose elastomers have an unsaturated polymer backbone and can be crosslinked with sulfur, with the exception of ethylene propylene rubber that is saturated and therefore is cured using peroxides (Mark et al., 2013). Petroleum and solvent resistant elastomers generally incorporate nitrile, amide, or chloride groups into their structures, and include rubbers such as nitrile, polychloroprene, epichlorohydrin, polyurethane, and chlorinated polyethylene (Mark et al., 2013). Temperature resistant elastomers incorporate molecules such as fluorine, sulfur, or silicon into their polymer structure to produce elastomers resistant to degradation at high temperatures (Mark et al.,

2013). Temperature resistant elastomers include polyacrylate, fluorocarbon, chlorosulfonated polyethylene, and silicone (Mark et al., 2013). Since solvent and temperature resistant elastomers lack unsaturated sites, these polymers must utilize non- sulfur cures, which will be detailed in subsequent sections.

31 The following sections go into further detail about various types of elastomers, their respective structures and unique properties.

3.3.1. Synthetic Elastomer Structures and Properties

Synthetic elastomers account for more than half of the world’s annual 27.5 million metric tons consumption of rubbers (Mark et al., 2013) . Synthetic elastomers are polymerized from petroleum-derived monomers. Synthetic rubbers have temperature and solvent resistance compared to natural rubber (NR) and are used predominately in applications such as tire treads, hoses, belts, flooring, dampeners, and medical devices

(Mark et al., 2013). Synthetic elastomers are thermosets or thermoplastics.

3.3.1.1. Aliphatic and Aromatic Hydrocarbon Elastomers

3.3.1.1.1. Aliphatic Hydrocarbon Elastomers

Synthetic hydrocarbon elastomers are general-purpose elastomers, as they have poorer temperature and solvent resistance than other elastomers that contain halogens, nitrogen, esters, or ethers. Hydrocarbons elastomers are divided into two classes: aliphatic (non-aromatic), and aromatic elastomers (Mark et al., 2013).

Synthetic non-aromatic hydrocarbon elastomers include polybutadiene rubber

(PBR), polyisobutylene rubber (PIB), isobutylene isoprene rubber (IIR), and ethylene propylene diene monomer rubber (EPDM) (Mark et al., 2013). Bulk differences between synthetic non-aromatic hydrocarbon elastomers are attributed to variations in chemical structure, and are specified in detail below.

PBR is a synthetic rubber comprised from the monomer 1,3-butadiene (Figure

3.13) (Mark et al., 2013). PBR can be polymerized in three forms: cis, trans, and vinyl.

Butadiene monomers polymerized from end-to-end, create the cis and trans forms (Mark

32 et al., 2013). The ratio of trans vs. cis isomers in PBR is controlled during synthesis, and impacts both microarchitecture and bulk properties (Mark et al., 2013). Trans double bonds formed during polymerization cause rigidity in the polymer chain, allowing for formation of microcrystalline regions (Mark et al., 2013). The cis double bonds formed during polymerization generate a bend in the polymer chain, preventing efficient polymer chain packing and therefore increasing amorphous regions in the polymer. PBR with a high percentage of cis double bond configurations (over 92%) is an elastomeric material, and is manufactured by using a Ziegler-Natta catalyst during synthesis (Mark et al.,

2013).

PBR accounted for about 25% of all synthetic rubbers consumed in 2012 (Mark et al., 2013). PBR has a high resistance to wear, and is commonly used in tires or as an additive to improve the mechanical strength of plastics (Mark et al., 2013). PBR also has high electrical resistivity, and is used in electrical component coatings (Mark et al.,

2013). In addition, PBR has high resilience and forms the elastic cores of golf balls.

Other applications for PBR include inner tubing for hoses, railway pads, and bridge blocks (Mark et al., 2013).

Fig. 3. 13: Polymerization of polybutadiene rubber.

33 Polyisobutylene (PIB) is a homopolymer of isobutylene, consisting of a saturated hydrocarbon backbone with two pendant methyl groups attached to alternating carbons on the polymer backbone (Fig. 3.14) (Mark et al., 2013). PIB is polymerized using cationic addition (Mark et al., 2013). PIB has good flex properties and is colorless to a light-yellow color (Mark et al., 2013). PIB is gas impermeable, and is primarily utilized as an inner liner in tires and other inflatable items (Mark et al., 2013).

Fig. 3. 14: Polymerization of polyisobutylene rubber.

Butyl rubber (IIR) is a copolymer, comprised of 98% isobutylene and 2% isoprene (Mark et al., 2013). The small amounts of isoprene allows IIR to be crosslinked with sulfur, improving bulk mechanical properties (Mark et al., 2013). Butyl rubber is used for shock absorption applications, and has low gas and moisture permeability (Mark et al., 2013). In addition, butyl rubber has outstanding resistance to heat, aging, weather, ozone, chemicals, flexing, abrasion, and tearing (Mark et al., 2013). Applications that use butyl rubber materials include shock mounts, sealant for rubber roof repair, tubeless tire liners, inner tubes, stoppers, sealants, adhesives, and liners (Mark et al., 2013).

Fig. 3. 15: Structure of isobutylene-isoprene rubber, or butyl rubber.

Ethylene propylene diene terpolymer (EPDM) is a synthetic elastomer, with an

M-class designation (ASTM D-1418) due to its saturated polyethylene chains (Fig. 3.16)

(Mark et al., 2013). Typical dienes used in EPDM include dicyclopentadiene, ethylidene norbornene, and vinyl norbornene (Mark et al., 2013). EPDM has outstanding resistance to heat, ozone, steam, and weather (Mark et al., 2013). It is an electrical insulator.

EPDM is used commonly in the automotive industry for seals: door seals, window seals, trunk seals, and hood seals (Mark et al., 2013). Additional applications for EPDM include appliance hoses and seals, vibrators, electrical insulation, belts, O-rings, and solar panel heat collectors (Mark et al., 2013).

Fig. 3. 16: Structure of EPDM

3.3.1.1.2. Aromatic Hydrocarbon Elastomers

Aromatic compounds contain an aromatic-ring configuration of atoms, whereas aliphatic compounds do not. Synthetic elastomers copolymerized with aromatic monomers, such as styrene, have improved hardness (Mark et al., 2013). In 2012, over

5.4 million tons of styrene butadiene rubber (SBR) were processed worldwide, and comprise about 50% of the rubber used in car tires (Mark et al., 2013). SBR contains butadiene monomers copolymerized with styrene (Fig. 3.17), and is polymerized from monomers in an emulsion or a solution (Mark et al., 2013).

Fig. 3. 17: Structure of styrene butadiene rubber (SBR)

Emulsion polymerized SBR (E-SBR) is produced using a free radical mechanism, not an anionic polymerization reaction like in S-SBR (Mark et al., 2013). In E-SBR, a free radical initiator such as potassium persulfate or hydroperoxide is used with ferrous salts to generate a free radical species (Mark et al., 2013). The free radical initiator then generates a free radical on a monomer, propagating a chain-growth mechanism until the chain transfer agent in the reaction, such as an alkyl mercaptan, caps the growing polymer chain (Mark et al., 2013). The chain transfer agent therefore controls the final molecular weight and viscosity of the E-SBR (Mark et al., 2013).

Solution polymerized SBR (S-SBR) is produced by an anionic polymerization method (Mark et al., 2013). Nucleophilic initiators, such as alkyl lithium, add a negative nucleophile to a monomer in a hydrocarbon solvent (Mark et al., 2013). Chain propagation in the anionic addition polymerization reaction utilizes the carbanion active site (created by the initiator) where subsequent monomers are attached and added (Mark et al., 2013). Since the carbanion active site is not very stable, the reaction is performed

37 at low temperatures close to 0oC (Mark et al., 2013). The reaction has no formal termination mechanism, but the carbanionic active site is often quenched by trace impurities such as oxygen, carbon dioxide, or water (Mark et al., 2013). Spontaneous termination occurs as well due to carbanion decay, resulting in hydride elimination (Mark et al., 2013). The spontaneous termination of the polymerization reaction can create issues with S-SBR processability, specifically with respect to molecular weight distributions and long chain branching (Mark et al., 2013). To improve processability, coupling agents such as SiCl4 and SnCl4 are used during the carbanionic polymerization to broaden the molecular weight distributions and create branched polymer structures

(Mark et al., 2013). S-SBR is typically used in extruded and molded rubber goods including specialty tire applications due to its better wet grip and rolling resistance than

E-SBR (Mark et al., 2013).

Additional variations in SBR polymerization include the ratio of styrene to butadiene monomers and the type of butadiene isomer used (Mark et al., 2013). Higher concentrations of styrene create a rubber that is harder, less elastic, and has better abrasion resistance than BR (Mark et al., 2013). Butadiene monomers can be added in

1,4- (including cis-1,4 and trans-1,4 isomers) or in 1,2- units to the growing polymer backbone (Fig. 3.18) (Mark et al., 2013). The relative concentration of 1,2 vs. 1,4- addition during polymerization is dependent upon the type of polymerization reaction.

Fig. 3. 18: Polybutadiene linked through the 1-and 4-carbon atoms, and through the 1- and 2-carbon atoms.

As a result, SBR is predominately used in tire treads, cables, and footwear, among other applications where abrasion resistance is needed (Mark et al., 2013). Increasing the styrene content in SBR also increases the Tg of the material and therefore can impact physical properties of the material.

SBR has good chemical resistance to weak organic acids, alcohols, moderate chemicals, and ketones (Mark et al., 2013). SBR has poor resistance to ozone, strong acids, fats, oils, and other hydrocarbons (Mark et al., 2013). SBR has poor heat resistance compared to most other elastomers, and is used in temperatures ranging from -

60oF to 250oF (Mark et al., 2013).

3.3.1.2. Halogen and Nitrile Substituted Elastomers

Halogen and nitrile substituted elastomers have superior oil and organic solvent resistance compared to hydrocarbon elastomers. The most popular halogen substituted elastomer is polychloroprene (CR) (Fig. 3.19) (Mark et al., 2013). In addition to superior resistance to organics, CR has low flammability, good toughness, and good ozone and

39 weather resistance (Mark et al., 2013). CR is used in the construction and automotive industries for belts, hoses, and gaskets.

Fig. 3. 19: Structure of polychloroprene.

A variety of blends can be made with CR, to improve specific properties. Natural

Rubber (NR) improves building tack, low-temperature flexibility, elasticity, and reduces

CR material when compounded with CR (Mark et al., 2013). BR can reduce mill sticking of CR, and improves low-temperature brittleness; however, the CR flex-fatigue life-time may be reduced when compounded with BR (Mark et al., 2013). SBR reduces crystallization hardening, and cost when compounded with CR (Mark et al., 2013). NBR improves the oil resistance of CR (Mark et al., 2013). EPDM vulcanizates can be compounded with CR to improve oil resistance, reduce cost, and improve ozone resistance (Mark et al., 2013).

In some cases, organic accelerators are used to promote monosulfidic bridges in

CR curing in addition to accelerators (Mark et al., 2013). Traditionally thioureas, such as ethylene thiourea (ETU) and its chemical derivatives diethylene thiourea (DETU), and diphenyl thiourea (DPTU) were developed as organic accelerators for CR curing (Mark et

40 al., 2013). However, ETUs have been linked to cancer in laboratory animals, and alternatives to thioureas have been developed, such as N-methyl-thiazolidine-2-thione

(Mark et al., 2013). Removing thioureas from curing systems can result in a slower cure, and vulcanizates with higher set properties and lower heat resistance (Mark et al., 2013).

CR can be crosslinked by metal oxides, and do not need accelerators such as other diene rubbers. Zinc oxide (ZnO) and magnesium oxide (MgO) are the most frequently used metal oxides in CR crosslinking; a combination of ZnO and MgO is used to cure CR to optimize mechanical properties (Mark et al., 2013). Typically 5 phr ZnO and 4 phr

MgO are used (Mark et al., 2013).

Nitrile substituted elastomers include acrylonitrile butadiene rubber (NBR), a copolymer of butadiene and acrylonitrile (Fig. 3.20). NBR is used commonly in the automotive non-tire and industrial rubber business due to its oil and heat resistance (Mark et al., 2013). The composition of acrylonitrile in NBR impacts the physical and chemical properties. There are over 188 different grades of NBR manufactured internationally, varying due to polymerization method (batch, continuous, hot, cold), and molecular weight distribution (Mark et al., 2013). Modifications to NBR include carboxylated, precrosslinked, ACN/isoprene/butadiene, liquid, carbon black masterbatches, plasticizer extended, and nitrile/pvc blends (Mark et al., 2013). When compounding NBR, the acrylonitrile (ACN) content and the viscosity of the NBR grade should be taken into consideration first (Mark et al., 2013). NBR grades with a broad molecular weight distribution are used for extrusion and calendaring processes, while narrow molecular weight distributions are used for molding (Mark et al., 2013).

41 Reinforcing fillers are commonly compounded with NBR, to improve tensile and tear strength, abrasion resistance, chemical resistance, resilience and low compression set

(Mark et al., 2013). Carbon black is commonly used in NBR, as well as non-black fillers such as silica, silicate, clays, talc, and calcium carbonate (Mark et al., 2013). Plasticizers are used with NBR to reduce costs, and utilize polar plasticizers such as highly aromatic mineral oils, and esterized oils (Mark et al., 2013).

NBR is crosslinked using sulfur and peroxide cures, and zinc oxide/peroxide for carboxylated nitriles (Mark et al., 2013). Sulfur cures with NBR are best for dynamic applications at moderate temperatures, where heat resistance and low compression set are not major factors. Sulfur-free cure systems are limited in application (Mark et al., 2013).

Peroxide cure systems provide the best NBR heat and compression set resistance, its ability to return to its original geometry after a prolonged compressive stress at elevated temperatures (Mark et al., 2013).

Fig. 3. 20: Structure of acrylonitrile butadiene rubber (NBR).

42 3.3.1.3. Sulfide Elastomers

Polysulfide rubber (PSR) has superior chemical resistance towards hydrocarbons.

PSR is used in sealant applications due to its superior dimensional stability, flexibility, low moisture vapor transmission, low gas transmission and weatherability (Mark et al.,

2013). Commercial brands of PSR include Thiokol FA, Thiokol ST, and Thiokol LP, which are synthesized via nucleophilic substitution with sodium polysulfide and dichlorides (Mark et al., 2013). Thiokol FA is made from di-2-chloroethyl formal and ethylene dichloride, and is used in specialty roller applications requiring resistance to ketones, aromatic solvents, and some chlorinated solvents (Fig. 3.21) (Mark et al., 2013).

Thiokol ST is a branched polysufide formed from di-2-chloroethyl formal with about 2%

1,2,3-trichloropropane as trifunctional branching units, used for mechanical goods.

Thiokol LP is formed from the cleavage of Thiokol FA and Thiokol ST (mercaptan- terminated polymers) (Mark et al., 2013). Thiokol LP is liquid and is primarily used as a sealant, coating, and binder (Mark et al., 2013).

Solid polysulfides (Thiokol ST and Thiokol FA) are typically vulcanized to provide excellent low temperature properties (Mark et al., 2013). Using an ester type plasticizer can further improve low-temperature properties (Mark et al., 2013).

Polysulfides provide good impermeability to solvents and gases (Mark et al., 2013).

Thiokol FA has excellent weathering and ozone resistance, making it ideal for weather strips and sealants (Mark et al., 2013).

Compounding of polysulfide includes vulcanization agents, fillers, and plasticizers (Mark et al., 2013). Zinc oxide is the predominant vulcanization agent used in Thiokol FA (Mark et al., 2013). Thiokol ST is optimally cured using zinc peroxide.

43 Carbon black, and other elastomers are good reinforcing agents for polysulfides (Mark et al., 2013). Non-black fillers are not as effective in reinforcement of polysulfide elastomers (Mark et al., 2013). Low pH fillers, such as clay, should be avoided and will slow the rate of vulcanization (Mark et al., 2013). Polysulfides are blended with other elastomers such as nitrile rubber, NBR, or neoprene to improve their physical properties and processability (Mark et al., 2013). Thiokol FA is typically blended with neoprene to improve strength and processing, but reduces solvent resistance (Mark et al., 2013).

Fig. 3. 21: Structure of polysulfide Thiokol A.

3.3.1.4. Fluorocarbon Elastomers

Fluorocarbon elastomers (FKM) are typically used in harsh applications where chemical and heat resistance are needed, in the automotive and aerospace industries

44 (Mark et al., 2013). A common type of FKM is the dipolymer of poly(vinylidene fluoride-co-hexafluoropropylene), synthesized from monomers of hexafluoropropylene

(HFP) and vinylidene fluoride (VDF or VF2), shown in Figure 3.22. An additional monomer tetrafluoroethylene (TFE) can be used instead of VF2, to increase the fluorine level in the FKM (Mark et al., 2013). FKM are typically prepared by high-pressure, free- radical emulsion polymerization (Mark et al., 2013). Organic or inorganic peroxy- compounds, such as ammonium persulfate are used as initiators (Mark et al., 2013).

Compounding FKM is different from most other elastomer compounding; plasticizers are not tolerated, and most chemicals used in rubber compounding are not recommended due to FKM’s extreme resistance to chemicals and heat. FKM is typically cured using Bisphenol AF and accelerator salts (Mark et al., 2013). Amines can also be used to cure FKM, but confer an inferior scorch safety and compression set resistance

(Mark et al., 2013). Peroxide curing systems can only be used with FKM that contain a peroxide cure site. All cure systems for FKM need metal oxides, which act as acid acceptors and capture HF formed during vulcanization (Mark et al., 2013). Magnesium oxide (3 phr) and calcium hydroxide (6 phr) are typically used for bisphenol cures, whereas lower levels of zinc oxides (1 to 3 phr) are used for peroxide cures (Mark et al.,

2013).

Fig. 3. 22: Structure of poly(vinylidene fluoride-co-hexafluoropropylene).

3.3.1.5. Polysiloxane Elastomers

Polysiloxanes, or silicone rubber (SR), are elastomers with inorganic backbones containing silicon and oxygen. Common polysiloxanes include polydimethylsiloxane

(PDMS), and fluorosilicone. Linear siloxane polymers, such as PDMS, are generally synthesized using a ring opening polymerization of a trimer or tetramer, containing alkyl or aryl groups (Figure 3.23) (Mark et al., 2013). The spatial orientation of the alkyl and aryl groups post-synthesis can generate isotactic and syndiotactic stereoregular forms.

The ability of PDMS to change spatial arrangements by rotations on its skeletal bonds is termed dynamic flexibility. Polymers with dynamic flexibility have flexible chains, and thus low glass transition temperatures. The Tg of PDMS is approximately -125°C, the

46 lowest of all polymers (Wypych, 2012). In addition to dynamic flexibility, PDMS has chain flexibility along its backbone; the oxygen skeletal atoms (Si-O-Si) are the smallest atoms with multi-valency required for a polymer chain structure and allows for rotational movement along the polymer backbone (Wypych, 2012). The bond angle of Si-O-Si is readily deformable, due to its large bond angle of 143°, compared to a stable tetrahedral bond angle of 109.5° (Wypych, 2012). The ability for chain rotations and chain deformity in PDMS results in an ability of the polymer chain to be compact while in a random coil; therefore PDMS has a very low melting temperature (40°C) due to the resulting equilibrium flexibility (Mark et al., 2013). Therefore, strain induced crystallization rarely occurs in PDMS (Wypych, 2012). To improve stiffness in polysiloxanes, large side groups are added to the backbone, or rigid units are polymerized into the backbone (Wypych, 2012).

In addition to a low glass temperature, other properties of siloxane polymers include high permeability to gases. As a result, siloxanes are frequently used in applications such as gas separation membranes and soft contact lenses (Wypych, 2012).

Fig. 3.23: Structure of polydimethylsiloxane.

47 3.3.1.6. Thermoplastic Elastomers

Thermoplastic elastomers are typically segmented block copolymers that are phase separated into amorphous and semi-crystalline regions. The hard segments are semi-crystalline regions acting as physical crosslinks, reducing chain slippage (Mark et al., 2013). The soft segments are the amorphous regions that create a matrix, which contributes to the flexibility and resiliency of the material. The properties of thermoplastic elastomers vary due to the proportion of hard and soft segments, molecular weight distribution, method of preparation, and thermal history that affects the degree of phase separation and domain formation (Mark et al., 2013). Thermoplastic elastomers lack chemical crosslinks, and therefore utilize physical crosslinks to make sufficient molecular entanglements (Mark et al., 2013).

48 3.3.1.6.1. Styrenic Block Thermoplastic Elastomers

The styrenic block copolymers (SBCs) are the most widely used thermoplastic elastomer. Styrenic block copolymers are comprised of at least three blocks in a A-B-A type, with two hard polystyrene end blocks and a soft, elastomeric midblock. The midblock is typically a polydiene such as polybutadiene or polyisoprene. Common SBCs include styrene-butadiene-styrene (SBS), styrene-isoprene-styrene (SIS), styrene- ethylene-butylene-styrene (SEBS), styrene-ethylene-propylene-styrene (SEPS), styrene- isobutylene-styrene (SIBS), and styrene-ethylene-ethylene-propylene-styrene (SEEPS), and are summarized in Fig. 3.24 (Mark et al., 2013).

Fig. 3. 24: Structure of styrenic block copolymer elastomers.

49 SBCs are used in footwear, and adhesives. SBCs are also used as a hardener in asphalt to improve rut resistance. SBCs are also compounded to create grips and cosmetic finishes for consumer goods, automotive parts, and packaging (Mark et al.,

2013).

3.3.1.7.2. Polyamide Thermoplastic Elastomers

The polyamide elastomers are segmented block copolymers that include poly(esteramides) (PEA), poly(etheresteramides) (PEEA), poly(carbonateesteramides)

(PCEA), and polyether-block-amides (PE-b-A) (Mark et al., 2013). Typical of other segmented coblock polymers, polyamides are phase separated into hard and soft segments. The hard segments of polyamides are aliphatic polyamides, and the soft segments are aliphatic polyethers and/or polyesters. Hard and soft segments are linked by ester or amide groups.

Polyamides have superior temperature resistance compared to other thermoplastic elastomers. Polyamides also have superior abrasion resistance, with a hardness in the range of Shore 80A to Shore 70D (Mark et al., 2013). Polyamides are also used for insulation materials. Additional applications include industrial applications, consumer goods, automotive applications, electronics, hot melt adhesives, powder coatings for metals, and impact modifiers in thermoplastics (Mark et al., 2013). Since polyamides are biodegradable, they are also being developed for biomedical applications including drug delivery systems, hydrogels, and tissue engineering (Mark et al., 2013).

Fig. 3. 25: Structure of polyamide thermoplastic elastomers.

3.3.1.6.3. Polyether Ester Thermoplastic Elastomers

Polyether ester elastomers are coblock polymers with hard polyester segments embedded in a soft polyether matrix. The polyester crystalline regions act as physical crosslinks, which provide polyether ester elastomers with superior heat resistance with a wide useful temperature range. Polyether ester elastomers have good resistance to hydrocarbon based greases and oils, but has limited resistance to halogenated solvents and aqueous acids and bases (Mark et al., 2013). Polyether esters have excellent wear properties and high tensile strength, but have a higher and more limited hardness range

(30 to 82 Shore D) compared to other thermoplastic elastomers (Mark et al., 2013).

3.3.1.6.4. Polyolefin Thermoplastic Elastomers

Polyolefins are hydrocarbon thermoplastics formed from the polymerization of olefins including propylene, ethylene, isoprenes, and butenes (Mark et al., 2013).

Polyolefin thermoplastic elastomers (TPEs) depend on the crystallization of polymer chains to produce elastomeric characteristics. Properties of polyolefins depend on the type of monomers and route of polymerization, resulting in distinct types of polyolefin

TPEs such as: blends or thermoplastic polyolefins (TPOs), dynamically vulcanized blends (TPVs) of ethylene-propylene random copolymer (EPM) or ethylene-propylene diene monomer (EPDM) with an olefin, random block copolymers (e.g., ethylene α-

51 olefin copolymers), block copolymers (e.g., hydrogenate polybutadiene-isoprene- butadiene block copolymer), stereo polymers (e.g., stereoblock polypropylene) and graft copolymers (e.g., polyisobutylene-g-polystyrene) (Mark et al., 2013).

Polyolefin TPEs have different types of crystallization due to structural differences. Random block copolymers polyolefin TPEs contain long ethylene segments that crystallize and form physical crosslinks for amorphous chain segments (Mark et al.,

2013). Stereoblock copolymers utilize tacticity to create crystalline regions (Mark et al.,

2013). Polyolefin graft copolymers contain crystalline polyolefin chains grafted to an amorphous polyolefin backbone. Most graft and block copolymers are elastic due to reversible, physical crosslinks and rubber chain entanglements, utilizing an entropic retractive force (Mark et al., 2013). TPOs are mechanical blends, and therefore are co- continuous with both an elastomeric and crystalline polyolefin phase (Mark et al., 2013).

TPVs are crosslinked, and therefore have a continuous polyolefin phase that surrounds the discontinuous elastomer regions.

Polyolefin blend TPEs (TPOs) are predominately made from elastomeric ethylene-propylene random copolymer (EPM), and crystalline isotactic polypropylene

(iPP). EPM-iPP blends are typically shear mixed in the range of 100-1000 s-1 (Mark et al.,

2013). The viscosity ratios of EPM and iPP can be adjusted to provide continuous phases over a wide range of volume fraction of the blend (e.g., 80/20 to 20/80) (Mark et al.,

2013). TPOs are compounded with additional additives such as fillers, reinforcing agents, plasticizers, lubricants, processing aids, flow modifiers, antioxidants, heat stabilizers, etc. (Mark et al., 2013). Most additives are contained in the rubber phase of

TPOs, softening and extending the rubber (Mark et al., 2013).

52 3.3.1.6.5. Polyurethane Elastomers

Polyurethanes have superior abrasion resistance and tear strength, better oxygen resistance, but are susceptible to rapid breakdown from heat and water. Polyurethanes are produced by the polyaddition reaction of a diisocyanate or a polymeric isocyanate with a polyol, with appropriate catalysts and additives (Mark et al., 2013). Urethane groups are formed along the polymer backbone, typically in a reaction between isocyanate and hydroxyl groups.

Polyurethane elastomers (PUs) are formed by reacting a diisocyanate (aromatic or aliphatic), a long-chain diol, and a small molecule chain-extender diol or diamine (Fig.

3.26). The resulting polyurethane consists of hard diisocyanate-chain extender segments in the soft macrodiol matrix (Mark et al., 2013). The microphase separation between the hard and soft segments vary between types of PUs due to differences in chemical composition and processing.

PUs utilize physical crosslinks from its hard domains, but can be chemically crosslinked as well using a tri-or multifunctional constituents (Mark et al., 2013). Once

PUs are crosslinked, they cannot be remolded and become a semi-crystalline thermoset.

As a result, thermoplastic PUs with only physical crosslinks are of interest due its ability to be remolded and thus recycled (Mark et al., 2013).

Fig. 3. 26: Chemical reaction for polyurethane synthesis.

3.3.2. Natural Rubber Latex and Natural Rubber

Over 2,500 species of plants produce cis-1,4- polyisoprene (Mooibroek and

Cornish, 2000). Natural rubber latex (NRL) is an aqueous emulsion comprised of rubber particles comprised of high molecular weight cis-1,4-polyisoprene macromolecules, and other biochemical metabolites such as proteins, fatty acids, and antioxidants that are all synthesized in vivo (Puskas et al., 2014). NRL can be coagulated to form solid natural rubber (NR), which entraps non-rubber particle latex biochemical metabolites within the rubber matrix (Cornish et al., 2008). NRL is used to make thin film products such as gloves, balloons, and coatings whereas NR is used to make bulk rubber items such as tires and gaskets (Cornish et al., 2008).

The only commercial sources of NRL and NR are available from Hevea brasiliensis, the Brazilian or Para rubber tree. Other plant species which make high molecular weight rubber include Parthenium argentatum (commonly called guayule), and Taraxacum koksaghyz (rubber dandelion) (Cornish et al., 2008). It is currently

54 unknown why plants produce rubber, and some of the underlying mechanisms of NR biosynthesis remain unclear (Puskas et al., 2014).

3.3.1.1. Extraction of Natural Rubber Latex

NRL is an aqueous emulsion of rubber particles present in laticiferous vessels

(ducts) or parenchymal (single) cells of rubber producing plants (Cornish et al., 2005).

The NRL from Hevea (HNRL) is extracted by “tapping” the rubber tree, characterized by making incisions along the trunk and collecting the latex as it bleeds from the laticifers.

The latex vessels are comprised of an anastomosed cell system, located in the secondary phloem in the trunk and arranged as a paracirculatory system (Cornish et al., 2005).

Sieve tubes create a series of circular rings in the plant by fusing to one another, and the incisions of the “tapping” process typically penetrate the contact points between sieve tubes. Rubber particles, as well as cytoplasmic components, are released from the cut laticifers, which retain most larger cellular components such as organelles and nuclei.

In the guayule shrub, rubber particles accumulate in the parenchymal cells of the shoot and root bark, and so tapping is not possible (Cornish et al., 1999). Thus, guayule

NRL (GNRL) is produced by homogenizing fresh plants in aqueous medium to release rubber particles from individual cells, and the particles are then purified while maintaining them in aqueous suspension (Cornish et al., 1999). Solvent extraction also can be used to extract rubber from dried guayule shrub. However, solvent extraction of rubber from guayule has many technical issues, including difficult separation of the viscous extractant from finely dispersed solids, accumulation of terpenes in recycled solvents, and the separation of a low molecular weight rubber and resin fractions from high quality GNR (Cornish et al., 1999).

55 GRNL contains terpene resins, but otherwise does not contain a low molecular- weight rubber fraction (unlike GNR produced by solvent extraction) (Cornish et al.,

2008). However, highly purified GNRL must be stabilized with emulsion additives such as ammonia, potassium hydroxide, and/or amine and phenolic antioxidants, and surfactants (Cornish et al., 2008). Stabilized GNRL in sealed containers has an impressive shelf life of at least 10 years (Cornish et al, 2008).

3.5.2. Rubber Bio-synthesis

Rubber particles are comprised of a monolayer proteo-phospholipid membrane encasing hydrophobic cis-1,4-polyisoprene chains (Puskas et al., 2014). The monolayer proteo-phospholipid membrane stabilizes the rubber particles, preventing aggregation in the aqueous environment of the plant cell or laticifer (Puskas et al., 2014). The biosynthesis of NR is catalyzed by the rubber transferase enzyme complex (cis-prenyl transferase, RT-ase), which is bound to the rubber particle’s membrane. The polymerization of rubber occurs at the active sites of the amphiphilic RT-ase which contains glycosylated hydrophilic regions that regulate the placement of subunits in the rubber particle membrane (Puskas et al., 2014).

The monomer of cis-1,4-polyisoprene is isopentenyl pyrophosphate (IPP) which is synthesized from acetyl-CoA via the cytoplasmic mevalonate pathway (Fig. 3.27.). IPP is isomerized to 1,1-dimethylallyl pyrophosphate (DMAPP) by the enzyme pyrophosphate isomerase (Kumar et al., 2012). DMAPP is then catalyzed by specific trans-prenyl transferases, adding 1-3 IPP monomers to form oligomeric allylic pyrophosphates (APPs), creating the initiator for the polymerization of IPP monomers

56 (Kumar et al., 2012). The RT-ase requires divalent cation co-factors such as Mg2+, and

Mn2+ (Kumar et al., 2012).

Rubber transferase, an amphiphilic enzyme, is located at the interface between the rubber particle core and the aqueous phase of the cytoplasm (Puskas et al., 2014).

Hypothesized polymerization mechanisms for natural rubber include a combination of chain growth and polycondensation mechanisms (Puskas et al., 2014).

The molecular weight and molecular weight distributions of cis-1,4-polyisoprene are species dependent and are impacted by plant age, genotype, environment, and extraction process (Cornish, 2001). NR also contains a gel phase, which is comprised of insoluble rubber that has been naturally crosslinked (Cornish, 2001). There are two different components to gel: a hard and a soft gel. The hard gel is formed by radical reactions between sulfur containing proteins and the dimethyl allyl double bond in the head group of NR (McMahan et al, 2015). The soft gel is produced by hydrogen bonding between phosphates and phospholipids at the end group of NR (McMahan et al, 2015).

The NR gel phase varies between plant species and environmental conditions. NR and

TNR (NR from T. kok-saghyz) are very similar in gel content and composition, whereas

GNR is different, having little protein or gel (McMahan et al, 2015).

Fig. 3. 27: The mevalonate pathway, which produces isopentenyl pyrophosphate (IPP), the monomer for cis-1,4-polyisoprene.

57 3.3.1.2 Hevea brasiliensis Natural Rubber Latex

Hevea rubber is obtained as latex (HRNL) which is about 35% by weight rubber particles. It also includes about 0.5% proteins, 0.6% phospholipids, and 0.09% tocotrienols as non-rubber components. Solid rubber from coagulated latex contains about 2.8% acetone-soluble fraction (tocotrienols, fatty acids, sterols, etc.), 2.5% protein fraction, 0.2% ash fraction, and 95-98% of hydrocarbon rubber (Amnuaypronsri et al.,

2008; McMahan and Lhamo, 2015; Tanaka, 2001).

Lutoids are the most abundant non-rubber particle in Hevea latex (Tanaka, 2001).

Lutoids are spherical membrane-bound vacuoles ranging in diameter from 0.5-5 um and have been completely characterized (Wititsuwannakul and Wititsuwannakul, 2005).

Within the lutoid membrane, there are dissolved metabolites such as acids, minerals, proteins, and sugars. Lutoids have been found to contain acid phosphatase, lysozyme, and acid hydrolases, which are characteristic of lysosomes. Lutoids regulate the homeostasis of the laticiferous system, and contribute to latex coagulation. Lutoid membranes contain high levels of phophatidic acid, and therefore have a strong negative charge that promotes their stabilization in aqueous media. Contents of lutoids include anionic and cationic protein Hevein is an anionic protein that is 70% of the total proteins found in lutoids (Gidrol et al., 1994). Hevein causes latex to coagulate by agglutination with the 23 kDA protein in rubber particles (rubber transferase). Lutoids also contain numerous acid hydrolases and peroxidases, which are utilized during cell homeostasis

(Gidrol et al., 1994).

58 3.3.1.3. Guayule Natural Rubber Latex

Guayule is a non-laticiferous plant, the rubber particles are biosynthesized and accumulate in the parenchyma cells of the stems and roots (Cornish, 2014). Guayule, however, does have resin canals within its parenchyma tissue, which produce resins that contain sesquiterpene ethers, triterpenoids, and fatty acid triglycerides but not rubber particles (Cornish, 2014). Histochemical staining and fluorescent microscopy has shown that rubber particles are localized in epithelial cells in parenchyma tissue and in pith, as well as around resin canals (Cornish and Backhaus, 1990).

3.3.1.4. Other sources of Natural Rubber

Taraxacum kok-saghyz (rubber dandelion)

Natural rubber is also produced by Taraxacum kok-saghyz, which is a diploid (n=16) sexually reproducing dandelion species native to Kazakhstan. The rubber particles produced by 4-month and 12- month old T. kok-saghyz has been characterized as having a unimodal distribution, with particle sizes ranging from 0.2 to 0.7 μm with an average particle size of 320 nm, and an average molecular weight of 4,750 kDa (Schmidt et al., 2010). NMR analysis has confirmed that the purity of poly(cis-1,4-isoprene) in T. kok-saghyz is >95% (Schmidt et al., 2010). The yield of natural rubber in T. kok-saghyz during plant growth increases rapidly during the first 8 months of growth; after 8 months the yield plateaus at 130-150 mg dry rubber/mL latex. Rubber particles from T. kok- saghyz contain cis-1,4-polyprenylcistransferase (rubber transferase), which is the same enzyme identified in the biosynthesis of H. brasiliensis rubber (Schmidt et al., 2010), and it has been hypothesized that biosynthesis of natural rubber in T. kok-saghyz is analogous to H. brasiliensis biosynthesis.

59 Other proteins found in natural rubber latex from T. kok-saghyz have been characterized, and show significant cross-reactivity between T. kok-saghyz proteins and anti-H. brasiliensis latex antibodies (Cornish et al., 2015). For instance, the H. brasiliensis rubber particle membrane bound protein and severe allergen Hev b1, has specific antibodies that cross-react with at least seven proteins found in T. kok-saghyz latex. Therefore, it can be inferred that the natural rubber from T. kok-saghyz has potential to trigger allergic reactions, and perhaps even sensitize people to Type I latex allergy. Therefore, commercial applications utilizing T. kok-saghyz should follow similar precautions to those currently in place for H. brasiliensis with regards to allergenic potential and safety.

60 3.4. Elastomer Compounding

Compounding is the mixing of a base polymer with additives or other polymers to make the base polymer perform better, cost less, improve processability, and/or improve its physical appearance (Donnet and Custodero, 2013; Rattanasom et al., 2007).

Thermoplastics and thermoset elastomers are compounded with additives such as fillers, antidegradants, blowing agents, desiccants, flame retardants, odorants and deodorants, peptisers, pigments, processing aids and plasticizers, retarders, and tackifiers (Dick,

2009). Elastomers are compounded to optimize polymer properties for specific applications, and variations in elastomer compounding are due to polymer structure and specific additives. Thermosets are compounded to generate crosslinks, which, like the elastomers, require chemical activators and accelerators in addition to the crosslinking agent.

3.4.1. Crosslinking

Thermoset elastomers require curing to create a three-dimensional network of chemically bonded intramolecular and intermolecular crosslinks, forming a continuous polymer network (Donnet and Custodero, 2013). The type of cure used to crosslink a thermoset depends on the chemical structure and intended application of the polymer.

Types of cures include peroxide, sulfur, metal oxides, phenolic resins, quinones, etc. and have different reaction chemistries, resulting in varying structure and quality of crosslinks

(Kruzelak et al., 2016). Sulfur and peroxide cures are the most commonly used for crosslinking rubber materials.

3.4.1.1. Peroxide cures

61 Peroxide cures are often used in fully saturated elastomers, such as silicones

(Kruzelak et al., 2016). Organic peroxides are thermally decomposed to create free radicals, which can create an active site on the polymer carbon backbone. The subsequent reaction between two active sites creates a carbon-carbon crosslink.

Polymers cured with peroxides have good heat-aging stability and low compression set, which is attributed to the carbon-carbon crosslinks (Kruzelak et al., 2016). Organic peroxides can be used in a wide variety of elastomers including saturated and unsaturated hydrocarbons (EPDM), fluoroelastomers, nitrile rubbers, and silicones (Kruzelak et al.,

2016).

Coagents (activators) are compounded with the polymer to optimize the peroxide cure. Coagents increase the crosslinking efficiency of the peroxide, typically not exceeding 0.3 phr (Dick, 2009). Coagents generally increase the hardness and tensile strength, while decreasing the elongation at break. Coagents are classified based on their contribution to the cure. Type I coagents increase both the rate and the state of the cure.

Type I coagents are typically polar, multifunctional low molecular weight compounds that form very reactive radicals, which are subsequently homopolymerized or grafted to polymer chains (Dick, 2009). Type II coagents form less reactive radicals and only contribute to the state of the cure (Dick, 2009). Type I coagents homopolymerize and form crosslinks through radical addition reactions, whereas Type II coagents typically contain an extractable allylic hydrogen and form crosslinks through intramolecular cyclization and intermolecular propagation reactions (Dick, 2009). Crosslinking at higher temperatures in the range of 100°C-160°C can be done with peroxides, with formulations typically containing a small amount of vinyl groups (Dick, 2009).

62 3.4.1.2. Sulfur cures

Sulfur is the most common crosslinking agent used with unsaturated rubbers to produced vulcanized rubber. The rhombic form of sulfur is typically used for vulcanization; it exists as an eight member- sulfur ring structure (Dick, 2009). The amorphous form of sulfur is a metastable high polymer, with a molecular weight of

100,000 to 300,000 (Dick, 2009). Amorphous sulfur is insoluble in most solvents and rubber, and therefore is typically only used prevent surface blooming on uncured rubber surfaces where tack is desired (Dick, 2009).

Typically, 1 to 3 phr of sulfur is used for most rubber products (Dick, 2009). For dry rubber crosslinking, the dispersion of sulfur for is improved using sulfur treated with carbon black, magnesium carbonate, or oils (Dick, 2009). In addition, masterbatches of sulfur with rubbers or other grafted polymers are used to improved dispersion. For latex crosslinking, emulsions of sulfur are used, therefore incorporating oils for improved dispersion.

Vulcanization with sulfur alone is an inefficient process, and requires the use of chemical activators and accelerators. The chemical reaction between sulfur and the double bonds of various polymers requires 40 to 55 sulfur atoms, without a chemical accelerator, and can take up to 6 hours to complete the reaction at 140°C (Dick, 2009).

The crosslinks of vulcanizates made without chemical accelerators are extremely prone to oxidative degradation and have inadequate mechanical properties for rubber applications

(Dick, 2009).

Chemical accelerators are used in rubber compounding to increase the speed of vulcanization, and permit vulcanization to proceed at lower temperature and with greater

63 efficiency. There are two categories of accelerators: primary and/or secondary accelerators. Primary accelerators are used at a concentration of 0.5 to 1.5 phr in most rubber compounds, and most are from six chemical classes: thiazoles, sulfenamides, thiurams, guanidines, dithiocarbamates, xanthates, and thioureas, with sulfonamides and thiazoles being the main classes of accelerators used (Dick, 2009). Primary accelerators usually have a long scorch time and cure quickly during vulcanization. Secondary accelerators are typically used at 10-40% loading of the primary accelerator, and boost the cure and increase the crosslink density (Dick, 2009). The most common secondary accelerators include guanidines, thiurams, and dithiocarbamates (Dick, 2009).

Thiazoles are a commonly used class of primary accelerators. Thiazoles have improved scorch safety and allow for high temperature cures with a short cure time and broad vulcanization plateau (Dick, 2009). The most common commercial thiazoles are bis(2-benzothiazole) disulfide (MBTS), 2-mercaptobenzothiazole (MBT), and the zinc salt of mercaptobenzothiazole (ZMBT) (Dick, 2009). The thiazoles typically need activation from zinc oxide or stearic acid, and require the use of a second accelerator to expedite the cure speed (Dick, 2009). Secondary accelerators are typically used at 10-

40% loading of the primary accelerator, and include diphenyl guanidine (DPG) or diothrotolyl guanidine (DOTG) (Dick, 2009).

Sulfenamides are another commonly used class of primary accelerators.

Sulfenamides are relatively safe, and are known for their scorch delay (Dick, 2009).

Sulfenamides allow for easy processing and molding of rubber compounds, providing a broad vulcanization plateau and good aging resistance (Dick, 2009). However, sulfenamides lack good storage ability, and will decompose when exposed to high

64 humidity and heat, as well as acids (Dick, 2009). Sulfenamides are typically used with secondary accelerators such as DPG, DOTG, or tetramethylthuiram mono or disulfide

(TMTM, TMTD) (Dick, 2009). These co-accelerators increase the cure rate but also reduce the scorch safety (Dick, 2009).

Thuirams are effective sulfur cure accelerators that contain two or more sulfur atoms. As a result, these compounds not only function as accelerators but also act as sulfur donors, and can allow for a “sulfur-less” cure therefore not needing elemental sulfur in the rubber compound (Dick, 2009). Rubbers with little to no sulfur content, or low unsaturation such as IIR and EPDM, typically use thuirams (Dick, 2009). The most popular thiurams are tetramethyl thiuram monosulfide (TMTM), tetramethyl thiuram disulfide (TMTD), and dipentamethylene thiuram tetrasulfide (DPTT) (Dick, 2009).

Rubbers cured with thuirams typically have excellent heat and water vapor resistance

(Dick, 2009). However, thiurams are expensive compared to other rubber accelerators, and also have a tendency to bloom to the surface (Dick, 2009).

Guanidines are condensation products of aromatic amines (aniline) and carbondisulfide with subsequent substitution of the thione functionality (>C=S) for a primary ketimine group (>C=NH) (Dick, 2009). Guanidines have a slow cure rate, and require the use of zinc oxide for activation (Dick, 2009). Guanidines are commonly used in thick walled rubber products, but are most commonly used as secondary accelerators with thiazoles (Dick, 2009). The most common guanidines used are diphenyl guanidine

(DPG), and N,N’-diorthotolyl guanidine (DOTG) (Dick, 2009). Guanidines provide a high crosslink density, and good mechanical properties including a high modulus and good compression set (Dick, 2009). Guanidines can cause a brown discoloration of

65 rubber goods, and therefore are not recommended for lightly colored objects (Dick,

2009).

Dithiocarbamates are ultra-fast accelerators that have a minimal induction time, and therefore require a retarder to avoid scorch when dithiocarbamates are used as a primary accelerator (Dick, 2009). Dithiocarbamates also require activators such as zinc oxide or fatty acids (Dick, 2009). Common dithiocarbamates include zinc dimethyl dithiocarbamate (ZDMC), zinc diethyl dithiocarbamate (ZDEC), and zinc diburyl dithiocarbamate (ZDBC) (Dick, 2009). As the alkyl group of the dithiocarbamates are lengthened, the scorch safety of the compound increases, with ZDBC having the highest scorch safety (Dick, 2009). However, the zinc salt of the dithiocarbamates decreases solubility in the non-polar rubber matrix, and therefore have a tendency to bloom to the surface at high concentrations (Dick, 2009). Dithiocarbamates are typically used in low sulfur cures, low temperature cures, and for white, transparent, or brightly colored rubber goods (Dick, 2009). Dithiocarbamates are also used as secondary accelerators to speed cure (Dick, 2009).

Xanthates are ultra-fast primary accelerators, predominately used for the vulcanization of rubber latex and rubber in solution at low temperatures (Dick, 2009).

Xanthates are polar, and some are even soluble in water (Dick, 2009). The most common xanthates are zinc xanthate (ZIX), and sodium isopropyl xanthate (NaIX) (Dick,

2009).

Thioureas are ultra-fast primary or secondary accelerators. Thioureas are predominately used for the vulcanization of polychloroprene rubbers. Commonly used

66 thioureas include ethylene thiourea (ETU), dipentamethylene thiourea (DPTU), and dibutyl thiourea (DBTU) (Dick, 2009).

3.4.1.3. Polyurethane crosslinking

Flexible and semi-flexible foams, cast elastomers, and coating systems are crosslinked segmented polyurethanes, consisting of soft and hard segments.

Polyurethanes are prepared using polyurethane prepolymer, which is combined with a chain extender such as a short chain glycol and/or a crosslinking agent having a functionality of three or more (Mark et al., 2013). Isocyanurates of different diisocyanates, polyamides, and polyols are used as crosslinking agents (Mark et al, 2013).

Polyurethane systems can also be crosslinked without the addition of crosslinking agents, attributed to allophanate and biuret bonds made during polyurethane synthesis (Mark et al., 2013).

67 3.4.1.4. Metallic oxides

Carboxylated nitrile, butadiene, and styrene-butadiene rubbers can be crosslinked via the reaction of zinc oxide with the carboxylated groups on the polymer chains, forming a zinc salt with the carboxylate groups (Dick, 2009). Polychloroprene

(neoprenes) and chlorosulfonated polyethylene are also vulcanized using metal oxides, zinc oxide being most common (Dick, 2009).

3.4.1.5. Addition cure

Silicone rubber can use addition or condensation cures for room temperature vulcanization for molding and casting applications (Dick, 2009). Addition cures are catalyzed using platinum, and are favored for their toughness and use in high temperature applications (Dick, 2009). Product types that use addition curing include: solid silicone rubber, liquid silicones, silicone gels, 2-part silicone rubber, and UV-curable silicone rubbers (Dick, 2009). The platinum catalyst creates a three-dimensional polymer network by reacting the crosslinker’s Si-H groups with the vinyl groups of the silicone polymer (Dick, 2009). Platinum catalysts used in silicone addition cures can be deactivated, or poisoned with chemicals such as with nitrogen, sulfur, phosphorus, sulfur vulcanized rubbers, polyurethanes, and tin associated with condensation cured silicone rubbers (Dick, 2009).

3.4.1.6. Condensation-based cures

Silicone rubber is cured by condensation reactions for general mold making and prototype applications. In condensation curing, terminal hydroxyl groups of silicone polymers react with a siloxane curing agent in presence of a tin or organotitanium catalyst and a small amount of water (Dick, 2009). This reaction releases volatile

68 compounds such as alcohols, acetic acid, and amines (Dick, 2009). Thus condensation silicones undergo slight shrinkage during cure from release of volatiles, whereas shrinkage is negligible for addition cures where there are no volatiles (Dick, 2009).

Product types that utilize condensation curing include 2-part silicone rubber, and 1- component silicone rubbers (Dick, 2009). Tin-catalyzed or organotitanium condensation cures are typically cheaper than platinum-catalyzed addition cures, and therefore condensation cures are used more widely in economic manufacturing (Dick, 2009).

Condensation cured silicones are more tear resistant than addition cured silicones.

3.4.2. Fillers

Addition of inert materials dispersed in the bulk of an elastomer prior to curing, can improve strength and hardness but typically at the expense of elasticity and resilience

(Leblanc, 2002). Size and shape of the fillers, as well as degree of dispersion are important factors. Some fillers will improve properties of an elastomer; providing a reinforcing effect that is characterized by improved stiffness, high resistance to tearing and abrasion, and enhanced tensile strength (Donnet and Custodero, 2013). Other fillers are diluents, and typically are used to reduce cost of the bulk product, typically without any improvements on polymer performance (Rattanasom et al., 2007).

Particle size is the most fundamental property of a filler which affects reinforcement of the elastomer the most. Particle sizes ranging from 1000-5000nm provide a small reinforcement; particles less than 1000nm provide a medium reinforcement; particles smaller than 100nm provide the strongest reinforcement (Hamed,

2000).

69 Fillers are typically hydrophilic, whereas most polymer matrices are more hydrophobic due to the large quantity of hydrocarbons. To strengthen the reinforcing effect, the interfacial bonding between a filler and its polymer matrix needs to be improved (Kohls and Beaucage, 2002). To improve interactions between filler and matrix, fillers are often surface modified. Modification of filler surface to improve hydrophobicity is most commonly achieved through silane coupling agents, via the hydrolysis of hydroxyl groups found on a filler’s surface.

3.4.2.1. Carbon Black/ Carbon Fly Ash

Carbon black is useful in increasing strength and hardness for elastomers, and is the most common filler used in the rubber industry. Carbon black is a colloidal form of elemental carbon, typically produced by the combustion of oil or natural gases. The average particle sizes of carbon blacks for industrial use typically range from 10nm to

500nm (Kausar, 2017). Aggregate size also varies with particle size, and can consist of four shape categories: spheroidal, ellipsoidal, linear, and branched (Kauser, 2017).

Particle size and shape can affect the filler-rubber interface and hence the reinforcement of carbon black in elastomer systems; Van Der Waals forces between the carbon black surface and the rubber matrix, mechanical interlocking of the rubber onto the filler surface, and rubber chains grafted to carbon black surfaces via free radical reactions between carbon atoms all provide a strong reinforcing effect (Dick, 2009). Smaller sized carbon blacks have a greater adhesion to the rubber matrix and therefore provide a stronger reinforcing effect, compared to larger sized carbon blacks with weaker adhesion to the rubber matrix (Dick, 2009).

70 3.4.2.2. Clays/Silicas

Silica is the second most common filler used in the rubber industry, and is commonly called a “white” filler. However, since silica is not quite as reactive with rubber as carbon black, silane coupling agents are used to modify the surface of silica particles. Clay is typically used as a cheap filler to reduce costs, and has a poor reinforcing ability because of its large particle size and low surface activity (Tohsan and

Ikeda, 2014; Dick, 2009). To improve the reinforcing effect of clays in polymers, the clays need to be intercalated or exfoliated. Successful intercalation or exfoliation of clays improve mechanical, barrier, and thermal properties (Dick, 2009). This can be done via solution blending, latex compounding, direct intercalation of molten polymer, and in-situ polymerization (Dick, 2009).

3.4.2.3. Calcium Carbonate and Whitening Agents

Calcium carbonate is used in its most common form, calcite, as a filler in elastomers. Calcite is produced from three different mineral sources: chalk, limestone, and marble (Dick, 2009). Calcite produced from chalk is typically the purest form of calcite, and consists of loosely bonded, uniform crystals that are about 3 microns in diameter. Milling breaks bulk chalk deposits into its basic, 3 microns crystal size (Dick,

2009).

To achieve calcium carbonate particle sizes smaller than 3 microns, calcium carbonate is precipitated by carbonation of a calcium hydroxide slurry (Dick, 2009). The crystal phase, particle size, and shape can all be controlled by the reaction conditions and additives (Dick, 2009). Most commonly, rhombic calcite crystals with a size of 50 – 100

71 nm, coated with fatty acids for improved compatibility with polymers are used for nano fillers (Dick, 2009).

3.4.2.4. Natural Fibers

Natural or lignocellulosic fibers consist of helically wound cellulose micro fibrils, embedded in amorphous lignin matrix. Cellulose (α -cellulose), lignin, pectins, hemicellulose, and waxes are the major components of natural fibers. Lignocellulosic fibers have been used extensively in composites, because of their low cost, low density per unit volume, and sometimes acceptable specific strength (Abraham et al., 2013b).

Natural fibers have poor fiber/matrix interactions, water resistance and lower durability

(Abrahamn et al., 2013a). This is attributed to the weak interfacial bonds between hydrophilic natural fibers and non-polar organophilic polymer matrices (Abraham et al.,

2011; Angellier et al., 2005). Commonly, natural fibers are surface treated to improve interfacial adhesion with polymer matrices using physical, mechanical and/or chemical approaches (Liu et al., 2008). Thermal stability of fibers is important when considering their application as fillers. The manufacturing and processing of composites at high temperatures can lead to degradation of the natural fiber, which can result in unfavorable bulk properties.

3.4.3. Anti-degradants

Polymers are subject to degradation on exposure to environments such as: storage, oxygen, heat, UV light and weathering, ozone, catalytic degradation from heavy metal ions (Cu, Mn, Fe, etc.), and dynamic flex-fatigue (Dick, 2009). Aging due to heat results in loss of elasticity and tensile strength, whereas oxygen-mediated failures (such as caused by ozone) can result in extensive cracking as well as loss in elasticity and tensile

72 strength (Dick, 2009). Most anti-degradants include antioxidants and antiozonants, typically classified into amine type anti-degradants and phenolic type anti-degradants

(Dick, 2009).

3.4.4. Blowing Agents

Blowing agents are used to create a cellular structure via a foaming process in materials that undergo hardening or a phase transition, including polymers. A blowing agent is typically added to a polymer in a liquid stage, undergoing a reaction that foams and subsequently hardens to create a porous structure. The cellular structure in a matrix reduces density while increasing thermal and acoustic insulation, with an overall increase in relative stiffness compared to the original polymer (Dick, 2009).

There are typically two classes of blowing agents: physical, and chemical.

Physical blowing agents are endothermic, requiring heat to volatilize the liquid blowing agent. Physical blowing agents in industrial applications include hydrocarbons such as pentane, isopentane, and cyclopentane, and liquid carbon dioxide (Dick, 2009).

Chemical blowing agents create a cellular structure via the release of gaseous products and typically undergo an exothermic reaction. Common chemical blowing agents include isocyanate and water for polyurethanes, azodicarbonamide for vinyls, hydrazine and other nitrogen-based materials for thermoplastic and elastomer foams, and sodium bicarbonate for thermoplastic foams (Dick, 2009). Mixtures of physical and chemical blowing agents are used in the production of flexible polyurethane foams with very low densities. Using both physical and chemical blowing agents minimizes temperature increases during synthesis, mitigating thermal degradation or damage (Dick, 2009). For

73 example, polyurethane foams utilize chemical blowing agents, isocyanate and water, in combination with the physical blowing agent of liquid carbon dioxide.

3.4.5. Bonding promoters

Adhesion between two surfaces is the result of interatomic and intermolecular interactions at its respective surfaces. Silane coupling agents are most commonly used to improve adhesion between rubber and inorganic fillers or substrates (Liu et al., 2008).

Primers or other additives are used typically in coating formulations, and are most commonly silicone-based (Dick, 2009). Other types of bonding promoters are used in complex composites such as rubber reinforced with steel wires, or nylon fibers. Bonding promoters used to improve adhesion of steel and nylon consist of resorcinol and a formaldehyde donor (Dick, 2009). The type of formaldehyde donor impacts the adhesion strength, and the kinetics of vulcanization. Therefore, it is important to optimize the type of bonding promoter and loading in a formulation for an individual application.

3.4.6. Desiccants

Desiccants are used in rubber compounding to remove water that may be from fillers, the bulk material, or a result of the vulcanization reaction. Having water present in a rubber compounds can lead to porosity in structure, and therefore desiccants are added. Calcium oxide is the most commonly used desiccant in rubber compounds, and can be used across a wide variety of rubber chemistries (Dick, 2009). Calcium oxide does influence the kinetics of vulcanization, and formulations need to be adjusted to maintain desired physical properties (Dick, 2009).

74 3.4.7. Flame retardants

Transition metal materials, including oxides, salts, organometallics or metal chelate complexes, are used as flame retardants for polymers. Metal halides such as antimony oxide, zinc borate, aluminum hydroxide, and chlorinated paraffins are commonly used as flame retardants (Dick, 2009).

3.4.8. Odorants and deodorants

Odorants and deodorants are often used to mask the distinct aromas associated with natural and synthetic elastomers, often associated with sulfur based crosslinkers or stabilizers (Dick, 2009). Odorants and deodorants are used to make the final product acceptable to the user. Since fragrances are volatile at higher temperatures, most are used in low temperature thermoplastics such as olefins, for extrusion and molding applications.

3.4.9. Peptisers

Peptisers are added in the earliest stage of natural rubber manufacturing, helping to break down the rubber rapidly during mastication to create a homogenous masticate

(Dick, 2009). Peptisers reduce the viscosity of natural rubber and latex by catalyzing rubber mastication. Reducing the viscosity of natural rubber to a workable level can reduce mixing time and thus provides time, energy, cost, and environmental benefits

(Dick, 2009). Traditionally, pentachloro triophenol based peptisers were used, but are replaced by dibenzoamido diphenyl disulfide chemistries to reduce environmental health risks (Dick, 2009).

75 3.4.10. Pigments

Pigments are used to improve heat and light resistance, as well as to add color to elastomeric products. The selection of a pigment depends on the application requirements for light stability, heat stability, resistance to bleed and migration, and shade (Dick, 2009). Pigments are classified as organic, inorganic, or hybrid pigments.

Pigments are available in a variety of forms, including dry powders, color concentrates, and liquids.

Organic pigments commonly used in plastics include quinacridones (red, violet, orange), dioxazines (violet), isoindolines (yellow, orange, red), perylenes, flavanthrones, and anthraquinones (Dick, 2009). However, organic dyes are sensitive to heat and chemicals, and their color can fade with long-term sun exposure.

The most common inorganic pigments include oxides, sulfides, hydroxides, chromates, and other metal complexes such as cadmium, zinc, titanium, lead, and molybdenum (Dick, 2009). Inorganic pigments are more thermally stable than organic pigments, and are opaquer and more resistant to migration, chemicals, and fading. The most widely used white pigment is titanium oxide, which is used alone or with other colorants to produce pastel shades. Inorganic pigment powders are insoluble in the rubber matrix and therefore cannot phase separate and do not bloom to the surface of the elastomer with use.

Hybrid pigments are formed from a stable dye sphere that surrounds an inorganic substrate such as silica, titanium dioxide, or aluminosilicates (Dick, 2009). Silica is commonly used as a filler in polymers, but has poor dispersion due to its innate hydrophilic properties and preference to agglomerate (Liu et al, 2008). Dyes have been

76 used to modify silica to improve dispersion while providing visually appealing end product colors.

3.4.11. Plasticizers/processing aids

Plasticizers and processing aids have the function of reducing the energy required for processing while assisting in the dispersion of rubber additives (Bergman and

Trimbach, 2014). In general, plasticizers reduce hardness, stiffness, and mechanical properties, but increase elongation at break and improve flexibility at low temperatures

(Petrovic et al., 2013). Plasticizers are often used to reduce the viscosity of the polymer compound, allowing for the addition of fillers (Dick, 2009). Process aids which do not modify the viscosity of the compound are commonly a combination of waxes, and/or fatty acid salts at a concentration of 1-3 phr (Bergman and Trimbach, 2014).

The most common plasticizers are mineral oils, consisting of aromatic, napthenic, and paraffinic oils (Petrovic et al., 2013). When selecting a plasticizer, compatibility between the type of oil and the polymer matrix needs to be considered (Petrovic et al.,

2013).

3.4.12. Retarders

Retarders, also known as pre-vulcanization inhibitors, readily react with accelerators and slowly release them for vulcanization. Ideal retarders increase the scorch time at processing temperatures, without delaying the time to reach 90% cure and without significantly changing the maximum rheometer torque (Dick, 2009). The most common use of retarders is to increase the scorch time, therefore gaining flow time in a mold or

77 calendaring process. This is useful in applications where the polymer is underfilling the mold, and increasing the preform weight results in thick flashes and flow lines.

3.4.13. Tackifiers

Tackifiers are chemical compounds used to increase tack, the stickiness of the surface of an adhesive. Tackifiers are amorphous, glassy, low molecular weight hydrocarbon polymers. Tackifiers can influence the color, odor, and stability of an adhesive. Tackifers are used in applications where adhesion between two surfaces is required. Resins, polymerized materials with a molecular weight less than 10,000, are used with rubber to improve flow properties, in addition to increasing tack (Dick, 2009).

However, not all resins are tackifiers. Resins that are tackifiers as well include aromatic resins (phenolic), petroleum-based resins, and plant based resins (wood rosin and terpenes) (Dick, 2009).

3.4.14. Type IV allergy associated with elastomer compounding

Some elastomer compounding chemicals can cause a type IV allergic reaction, clinically presented as non-immunological irritant contact dermatitis. Rubber additives associated with a clinical type IV allergic response include over 30 chemicals, associated with the chemical classifications of thiurams, benzothiazoles, mercaptos, ureas, thiocarbamates, diamines, and others (Nettis et al., 2002). A type IV allergy is clinically identified using a skin patch test, with readings taken 2 to 4 days post-exposure. Contact dermatitis, characterized by erthematosquamous eczema, is characteristic of a type IV allergic reaction.

78 3.5. Manufacturing Methods for Latex

Latex is an aqueous emulsion that contains a dispersion of polymer particles.

Elastomer latex systems include natural polyisoprene, and synthetic polyisoprene, nitrile, polybutadiene-styrene, and acrylics (Dick, 2009). Latex systems need to be compounded prior to manufacture, and vary in formulations due to polymer chemistry and structure, as well as the presence of other chemical additives. In the case of natural rubber latex, biochemical metabolites (such as proteins) entrapped in the latex during removal from feedstock can affect compounding and therefore manufacturing.

Latex systems will need an emulsifying agent to maintain the emulsion, and prevent phase separation of the polymer and water phases. Proteins and other biochemical metabolites are emulsifying agents that stabilize natural latex, whereas anionic and non-ionic soaps or detergents are typically used as emulsifying agents for synthetic latex systems. Natural latex is typically buffered at a basic pH, whereas synthetic latex requires a more neutral pH; this is attributed to the isoelectronic point and colloidal stability of the emulsions. As a result, variations in polymer latex chemistry can vary greatly and should be taken into consideration when developing a manufacturing technique.

3.5.1. Dipping

Latex dipping is the most commonly used process for a variety of thin film products. The dipping process consists of the following steps: insertion of a former into compounded latex, removing the former, drying to remove liquids, followed by vulcanization and finally stripping the product from the former. Formers are typically made from aluminum, stainless steel, ceramics, or glass. Automated dippers are

79 commonly used for products such as gloves, balloons, and condoms, generating several thousand per hour.

There are a few variations in the dipping process, depending on how thick the final product needs to be. Straight dipping is used for the thinnest products, typically for items 0.6 mm and thinner. Condoms are typically made using straight dipping, where a glass former is typically dipped in latex, dried, a second layer is dipped, dried, followed by the bead rolling, vulcanization, washing and stripping of the former using water spray.

Products should then be leached in water to remove excess compounding ingredients, and then surface treated via chlorination to reduce stickness and tack. Condoms are then tested according to ASTM and ISO standards, and then packaged for consumer use.

Coagulant dipping is used for products that have a thickness up to 1.5 mm, such as gloves. Coagulant dipping varies from straight dipping in that the former is dipped into an alcoholic solution of calcium salt, dried, and then dipped into latex. Coagulants such as positively charged calcium ions, neutralize the negative electrical charge on natural rubber particles and synthetic latex with anionic emulsifiers, which destabilize the forces keeping colloids apart. Coagulant dipping is less effective on synthetic latex systems that use non-ionic emulsifiers.

Heat sensitive dipping is used for products with a thickness up to 5 mm, including baby nipples and some industrial products. To achieve such a thick dipped item, latex is compounded with a heat sensitivity compound, such as polyvinylmethyl ether.

80 3.5.2. Open Cell Foams (Dunlop, Talalay)

Open cell elastomer foams, foams with interconnected pores, are typically formulated from NRL or synthetic polyisoprene latex. There are two methods that are used, the Dunlop and the Talalay methods which vary in how the foam is gelled prior to vulcanization. The Dunlop process utilizes a chemical gelling agent, such as zinc oxide whereas the Talalay process utilizes carbon dioxide gas and the formation of carbonic acid as the gelling agent. The morphology of the open cell architecture also varies between the methods, with the Talalay method having a more uniform cell structure.

The Dunlop method compounded latex with a gelling agent such as zinc oxide, and is whipped to incorporate air into the compounded latex. After the latex is foamed, poured into a mold and closed, and subsequently vulcanized. The foam is washed post- vulcanization, and then dried.

In the Talalay method, latex is compounded and foamed without a gelling agent, and then partially fills a large mold. The mold is closed, and the latex is expanded by vacuum to fill the mold, followed by cooling of the mold to freeze the foam structure, and introduction of carbon dioxide gas, with subsequent heating for vulcanization. The carbon dioxide causes gelation in the Talalay method, via formation of carbonic acid and lowering of the pH.

81 3.6. Manufacturing Methods for Rubber

Common dry rubber manufacturing techniques include closed cell foams, extrusion, and molding. The chemical composition of rubber needs to be taken into account during manufacturing, as the process of manufacture varies among the types of elastomer used. Thermoplastic elastomers are more readily molded due to their thermal transitions around their respective melting points, whereas thermosets do not have a melting point. Thermoset elastomers typically are cured or vulcanized while being manufactured, a major distinction from thermoplastics which do not need a cure cycle.

Natural rubber has multiple biochemical metabolites, such as proteins and phospholipids, entrapped in its dry rubber matrix, which can affect manufacturing and need to be taken into account during compounding.

3.6.1. Closed cell foams (semi-open cell foams)

Closed cell foams, foams with pores that are not connected, are dense foam products. Closed cell foams are used in the automotive industry, construction for insulation and other thermal management applications, and for consumer products such as wetsuits, gloves, and orthopedic braces. EPDM, silicones, natural rubber, and neoprene are all commonly used rubbers for closed cell foams.

Closed cells foams are made using a one or two step process with a gas, typically either nitrogen or supercritical carbon dioxide. In one-step foaming, carbon dioxide dissolution, expansion and cooling of samples takes place inside a high pressure mold.

The two-step foaming process, expansion of the sample into a foam can be done inside

82 the mold used to initially dissolve the carbon dioxide, or outside of the initial mold by heating carbon dioxide-saturated samples using an oil-bath or hot press.

3.6.2. Extrusion

Extrusion is a process in which a polymer is forced through a die of the required cross section under pressure. Typically, rubber products are extruded and then crosslinked or vulcanized. There are two types of extrusion processes: ram extrusion and screw extrusion. Both extrusion processes begin with unvulcanized rubber compound being fed into an extruder. In screw extrusion, the revolving screw within the extruder will begin to bring rubber into the die, increasing the pressure and temperature of the material as it approaches the die. In ram extrusion, a batch approach is used where a piston pushes a given volume of material through the orifice of the extrusion die. When the material reaches the die in both processes, the pressure forces the material through the die orifice, causing swelling of the materials. Since this swelling occurs, extruded parts typically have tolerances on their cross sections. After extrusion, the material is vulcanized, and may shrink. Extrusion dies are precise and specific tools, often custom made for a given product. Commonly extruded rubber materials include EPDM, neoprene, SBR, nitrile and silicone.

3.6.3. Molding

3.6.3.1. Compression Molding

Compression molding is the most common type of mold used in the rubber industry. Typically, compression molding consists of placing a precut or shaped polymer or composite into a two-piece mold that is closed. The pressure applied by the press

83 forces the material to fit the shape of the mold. Any excess material flows out of the mold and is termed flash. Tire treads are typically cured in a compression mold.

3.6.3.3. Transfer Molding

Transfer molding is used to mold complicated shapes, not possible using conventional compression molding. Transfer molding has a more complicated mold than those used in compression molding; transfer molds have an area of the mold which holds uncured polymer and then distributes this uncured polymer into the mold cavity where curing occurs.

3.6.3.3. Injection Molding

Injection molding traditionally were used for thermoplastics, but has been developed so that rubber compounds can be molded and vulcanized by this method.

Injection molding follows a cycle of feed, injection, and demolding, with low rejection rates and lower finishing costs.

3.6.4. Blown Film Extrusion

Blown film extrusion is the most common method to make plastic films, especially for the packaging industry. Elastomers are typically not processed using blown film extrusion, due to the lack of crystallization from thermal cooling by most elastomers which is required for successful blown film extrusion. Elastomers such as natural rubber latex can be used in a blend with highly crystalline thermoplastics, such as low-density polyethylene (Mahapram and Poompradub, 2011). In general, blown film extrusion involves the extrusion of a tube of molten polymer through a die, and inflating it to several times its initial diameter to form a thin film bubble. This bubble is then collapsed and used a lay-flat film or can be made into bags.

84 Blown film extrusion consists of four main steps:

1. Thermoplastic material starts in a pellet form, where it is compacted and melted to

form a continuous, viscous liquid. The molten thermoplastic is then extruded

through an annular die.

2. Air is then injected through a hole in the center of the die, and the resulting

constant and even pressure gradient causes the extruded melt to expand into a

bubble. Air leaving the bubble is replaced as a constant rate, to ensure uniform

thickness of the film.

3. The bubble is cooled via convective air, and begins to solidify. The line at which

the polymer transitions from molten to solid is termed the frost line, due to the

change in polymer opaqueness attributed to the diffraction of light by crystallites

in the polymer that are formed during cooling.

4. After solidification, the film moves into a set of nip rollers that collapse the

bubble and flatten it into two flat film layers.

85 Chapter 4: Mechanical properties of type I circumallergenic & type IV

hypoallergenic guayule natural rubber latex thin films.

J. Lauren Slutzky a, and Katrina Cornish a,b aOhio State University, Department of Food, Agricultural and Biological Engineering,

1680 Madison Avenue, Wooster, Ohio 44691 USA bOhio State University, Department of Horticulture and Crop Science, 1680 Madison

Avenue, Wooster, Ohio 44691 USA

*Corresponding author: Katrina Cornish, Williams Hall, 1680 Madison Avenue,

Wooster, Ohio 44691, USA. Tel: 330-263-3982 Fax: 330-263-3887 Email:

[email protected]

Abstract

Type I latex allergy sensitization and subsequent allergic reactions to Hevea natural rubber latex proteins have created an industry demand for thin film barriers that circumvent the allergic response (i.e. are circumallergenic). Other allergens associated with natural rubber and synthetic polymer thin films are attributed to residual thiazole, thiuram, and carbamate accelerators that can cause type IV allergies, characterized by contact dermatitis or delayed contact hypersensitivity. Type I circumallergenic and type

IV hypoallergenic thin films were created utilizing natural rubber latex from the plant species Parthenium argentatum Gray (guayule), cured with the accelerators diisopropyl xanthogen polysulphide (DIXP) and zinc diisononyl dithiocarbamate (ZDNC). Guayule latex is circumallergenic with respect to type I latex allergy, because its proteins do not cross-react with Hevea associated allergic proteins. Type IV allergies are diminished because DIXP is consumed during the vulcanization process, and skin tests have shown

86 that ZDNC does not cause dermal reactions or delayed contact hypersensitivity, because it does not bloom to the surface. These type I circumallergenic and type IV hypoallergenic thin films have mechanical properties superior to those described in

American Standard for Testing Materials D 3577, the standard for rubber surgical gloves, and could be used to make commercial medical thin film barriers, such as medical gloves, condoms, balloons and dental dams.

Keywords: Guayule, natural rubber latex, type I latex allergy, type IV latex allergy, elastomer thin films

4.1. Introduction

Elastomeric medical products, such as examination and surgical gloves, catheters, masks, dental dams, orthodontic rubber bands, and condoms often are made from natural rubber latex (NRL). NRL is derived from the plant species, Hevea brasiliensis, Muell

Arg. commonly known as the Brazilian or Para Rubber Tree. However, NRL contains numerous allergens capable of eliciting a spectrum of type I allergic responses, mediated by IgE, ranging in severity from contact dermatitis, contact urticaria, and delayed hypersensitivity through systemic reactions, such as hives and edema, to life-threatening anaphylaxis and death. Estimates of healthcare workers with immunogenic NRL immediate hypersensitivity have ranged from 3% to 17.9%, depending on populations studied and criteria on what defines an allergic response (Hamann et al., 2001; Haman et al., 1998; Nabavizadeh et al., 2009; Liss and Sussman, 1999). Direct skin contact by vulcanized elastomer products also can induce type IV delayed hypersensitivity reactions mediated by antigen specific sensitized T lymphocytes, and are attributed to residual accelerators used to crosslink natural, as well as synthetic, diene elastomers (Pak et al.,

87 2012). Type IV contact dermatitis to rubber products is one of the most common causes of occupational contact dermatitis (Meyer et al., 2000). Also, eczema associated with a type IV allergic response can increase susceptibility to a type I sensitization and immunogenic response for individuals with high exposure to medical NRL products, such as gloves (Miri et al., 2007).

Natural rubber latex derived from guayule is a circumallergenic (circumvents the allergic response) alternative to Hevea derived NRL. Purified guayule NRL contains less than 1% of the proteins in Hevea NRL, and 90% of the trace proteins are a single protein, cytochrome P450 oxidase, an allene oxide synthase (Cornish et al., 2008; Cornish et al.,

2006; Pan et al., 1995). This protein is known to be poorly immunogenic, and the P450- protein family has not been associated with allergic reactions in humans (personal communication, Dr. H.P.Rihs, BGFA, Ruhr-University-Bochum, Germany). There are no detailed studies of the remaining 10% of guayule NRL proteins because the amount of these proteins is too small to either evoke an immunogenic response or to concentrate the individual proteins sufficiently to permit an effective study. As a group, concentrated total guayule latex proteins were found to be less than normally immunogenic in rabbits

(Siler et al., 1996). Anti-guayule protein murine polyclonal IgG did not cross-react with

Hevea latex proteins (Siler et al., 1996). Furthermore, in vitro studies have shown that

Hevea latex-specific human IgE antibodies do not bind to guayule latex proteins, and in vivo studies confirmed guayule latex proteins to not cause reactions in Hevea-sensitized healthcare workers (Siler et al., 1996; Carey et al., 1995).

Type IV allergens in medical natural and synthetic elastomer products are attributed to residual accelerators used to increase the rate and efficacy of sulfur cross-

88 links formation and, in some cases, to anti-oxidants. The commonly used rubber accelerators, thiazoles, thiurams, and carbamates, are recognized by the U.S. Food and

Drug Administration as sensitizing agents capable of eliciting type IV allergic reactions

(van Jole, 2008). Other health hazards associated with thiuram and carbamate rubber accelerators include formation of fugitive N-nitrosamines, known carcinogens and teratogens. Current alternatives for diene elastomer crosslinking include gamma and UV irradiation, organic peroxide cures, zinc oxide activators without accelerators (utilizing the carboxyl-zinc ionic bond), and placement of functional groups onto the polymer backbone that can form crosslinks post product fabrication (van Jole, 2008). However, since gamma and UV irradiation, and organic peroxides create carbon and not sulfur crosslinks, they generally possess inferior mechanical properties compared to vulcanized rubbers (van Jole, 2008).

The rubber accelerators, zinc diisononyl dithiocarbamate (ZDNC) and diisopropyl xanthogen (DIXP), utilize sulfur in crosslinks, and have diminished residual chemicals associated with type IV allergens (Chakraborty and Couchman, 2006). ZDNC has a lower allergenic potential than conventional dithiocarbamates due to its high molecular weight (Chakraborty and Couchman, 2006). This limits its ability to bloom to the surface of latex films, and less ZDNC can be extracted from finished rubber articles compared to common industry accelerators such as ZBEC (Chakraborty and Couchman,

2006). DIXP is a fugitive xanthate accelerator which, unlike other fugitive accelerators, is consumed completely into sulfur crosslinks, and the byproducts of this reaction are volatile isopropanol and carbon disulfide (Chakraborty and Couchman, 2006). DIXP contains no nitrogen, and so cannot form carcinogenic and teratogenic N-nitrosamines.

89 The objective of this study was to optimally vulcanize guayule NRL (GNRL) with the accelerators DIXP and ZDNC to create type I circumallergenic and type IV hypoallergenic thin film elastomers with mechanical properties suitable for surgical gloves, and other thin film elastomer products, such as catheters and condoms. The materials made in this study are the first type I circumallergenic and type IV hypoallergenic NRL thin films reported with mechanical properties superior those specified in American Standard for Testing Materials (ASTM) D 3577, the standard for rubber surgical gloves.

90 4.2. Experimental

4.2.1. Materials and sample preparation

4.2.1.1. Emulsion chemistry/compounding

GNRL was made using a base compounding recipe (Table 4.1) while varying the amount of the accelerators DIXP and ZDNC (Robinson Brothers Ltd, UK). The amount of accelerators and other materials were added to GNRL at specified concentrations based on parts per hundred dry rubber (phr). The sulfur emulsion, antioxidant dispersion, and zinc oxide dispersion were supplied by Akron Dispersions (Akron, OH, USA) and the ammonium hydroxide was supplied by W.W. Grainger, Inc (Salt Lake City, UT,

USA). Deionized water was added to the compounded emulsion until 48% solids by volume was achieved, and was prevulcanized for 2.5 hours while stirring with a 30 rpm hand mixer.

Table 4.1. Latex Compound Recipe. Quantity (phr)

Guayule NRL 100

Sulfur 2

Ammonium Hydroxide 1

ZnO 0.3

Antioxidant 2

ZDNC 0.2, 0.4, 0.6, 0.8, 1.0, 1.2 or 1.4

DIXP 1, 1.2, 1.4, 1.6, 1.8, 2.0, 2.1 or 2.2

4.2.1.2. Thin film manufacture by dipping

Thin film elastomer products are manufactured by dipping formers into emulsions with subsequent heating to remove liquids and vulcanize the NRL. Initially, a stainless

91 steel former was heated to 70 ºC and dipped (10 second dwell) into a coagulant solution

(25% aqueous calcium nitrate in 70% isopropyl alcohol). After the solvent evaporated, the coagulant-coated former was reheated to 70 ºC then dipped into a compounded latex emulsion and held there for different dwell times during which a thin film of coagulated latex was deposited (dwell times varied from 5, 15, 30, 45, and 60 sec). Once the latex gelled via heating (15 min at 100 ºC), it was leached in water (30 min at 55 ºC), followed by stripping of the former and subsequent vulcanization of the rubber article (20 min at

105 ºC). The rubber article was then placed in a tumble dryer post-vulcanization (60 min at 60 ºC). All guayule natural rubber latex (GNRL) thin films in this study were made using a Diplomat automated dipper (DipTech Systems Inc., Kent, OH, USA).

4.2.2. Tensile Properties

Four dumbbell samples of each compound at each dwell time were cut using Die

C according to ASTM D 412 (ASTM International, 2013a). Evaluation of the tensile mechanical properties followed ASTM D 412 and was determined using an Instron 3366 with Bluehill v. 2.17 software package (Instron, Norwood, MA, USA) [15]. Samples were tested using a crosshead speed of 500 mm/min at room temperature (26°C). The reported mean ± the standard error from the mean (SE) values are averages of at least 4 samples. In addition, statistical analysis was performed using Minitab, version 16.0

(State College, PA). Significant differences (P-values < 0.05) in tensile data amongst

GNRL compounds for each film thickness were determined using one-way analysis of variance (ANOVA) and the Fisher method.

4.2.3. Dynamic Mechanical Analyzer (DMA)

92 The glass transition temperatures of the GNRL thin films were studied using a TA

Instrument Q800 (New Castle, DE), as previously described (Monadjemi et al., 2016).

The samples were equilibrated at -110°C and subsequently heated to 150°C with a heating rate of 5°C/min. The sample dimensions were approximately 25 x 3.0 x 0.7 mm3 with an amplitude of 15 μm, preload of 1N, and frequency of 1 Hz using the tension film clamps. The values presented are averages of four samples.

4.2.4. Differential Scanning Calorimetry (DSC)

TA Instruments Q100 (New Castle, DE) differential scanning calorimetry (DSC) was used to determine the state of the crosslinked elastomers as a function of temperature, using a previously described procedure (Musto et al., 2016). The average glass transition temperatures were determined from thermograms obtained by heating 5 to 10 mg samples from -20 to 200 ºC at a rate of 10 ºC/min. The samples were allowed to anneal at 200 ºC for 3 to 5 min to remove the thermal history of the polymers, and subsequently cooled to -85 ºC, held for 3 to 5 min, and reheated from -85 to 200 ºC with a heating rate of 10 ºC/min. The reported values are averages of four samples.

4.2.5. Thermogravimetric Analysis (TGA)

Thermogravimetric analysis (TGA) was used to study thermal decomposition properties using a TA Instruments TGA 5000 (New Castle, DE). The samples were heated under nitrogen from room temperature to 500 ºC with a heating rate of 20 ºC/min.

The reported values are averages of four samples.

93 4.3. Results and Discussion

4.3.1. Dwell time and film thickness

When the film thickness for each respective dwell time was measured for four replicate samples, each set had a standard error less than 0.01 mm (Table 4.2). The relationship between dwell time and average thin film thickness showed a linear relationship (R2=0.98, df = 3).

Table 4. 2. Dwell time and average thin film thickness, SEM < 0.01 mm

Dwell Time (s) Average Film

Thickness (mm)

5 0.15

15 0.19

30 0.21

45 0.23

60 0.26

4.3.2. Tensile properties

3D surface graphs were created using mesh points based on an average of four samples, to show how modulus at 500% elongation (MPa), tensile strength (MPa), and ultimate elongation (%) varied with different levels of DIXP and ZDNC for each film thickness (Fig. 4.1).

The modulus at 500% elongation typically was not significantly different across film thicknesses for each formulation. Exceptions are films made with 1.0 phr DIXP and

0.4 phr ZDNC, which showed significant differences in modulus at 500% elongation for the 150 and 190 μm thinner films and the thicker films of 210, 230, and 260 μm. The

94 modulus was statistically higher for the thinnest films of 150 μm compared to the thickest films of 260 μm for the following formulations: 1.4 phr DIXP for all ZDNC loadings above 0.6 phr, and 2 phr DIXP for ZDNC loadings of 1 phr, 1.2 phr. There was no statistical difference in modulus for film thicknesses for the formulations of 2 phr DIXP with 0.6 phr ZDNC, and 2 phr DIXP with 0.8 phr ZDNC, nor at comparative film thicknesses.

The modulus at 500% elongation in the thinnest films was particularly sensitive to the ratio and concentration of the two accelerators (Figure 4.1(a) and (d)). The thinnest film (Figure 4.1(a)) has a modulus made with 1.4 phr DIXP was statistically significantly higher than for films made with 2.0 phr DIXP, with ZDNC concentrations of 0.4, 0.6, and

0.8 phr. However, this difference was not observed in treatments containing higher concentrations of ZDNC, such as 1.4 phr DIXP with 1.0 phr and 1.2 phr ZDNC compared to 2.0 phr DIXP with 1.0 phr and 1.2 phr ZDNC. In the thicker films (Figure

4.1 (g), (j) and (m)) modulus at 500% elongation became quite uniform over the different accelerator concentrations and ratios, with the exception of the 1 phr DIXP, with 0.4 phr

ZDNC film was being statistically lower than films from all other formulations at the same film thicknesses.

Tensile strength varied considerably among all film thicknesses in the tested accelerator treatments. Tensile strength (Figure 4.1 (b), (e), (h), (k) and (n)) was strongest in the films made with 2.0 phr DIXP/0.6 phr ZDNC and 2.0 phr DIXP/0.8 phr

ZDNC for all film thicknesses, and these had similar strength to each other. The films made with 2.0 phr DIXP/0.6 phr ZDNC, and 2.0 DIXP/0.8 phr ZDNC, were significantly stronger than films from the other formulations and across the different thicknesses.

95 Relative minima in tensile strength were observed in 1.4 phr DIXP/1.2 phr ZDNC films

(Figure 4.1 (b) (e) and (k)), which were significant lower than films made with 2.0 phr

DIXP/0.8 phr ZDNC, as well as 2.0 phr DIXP/0.6 phr ZDNC.

The largest ultimate elongations were obtained in 2.0 phr DIXP/0.8 phr ZDNC

(Figure 4.1 (c), (f) and (i)) across most thicknesses. The thinnest 2.0 phr DIXP/0.8 phr

ZDNC films had an elongation at break of 1888 ± 78.68 % (Figure 4.1(c)), significantly higher than all other films except those made with 2.0 phr DIXP/0.6 phr ZDNC. The lowest ultimate elongations for most films occurred in 1.4 phr DIXP/1.2 phr ZDNC, and

1.6 phr DIXP/1.2 phr ZDNC (Figure 4.1 (c), (f), (i) and (l)), and these were statistically lower than 2.0 phr DIXP/0.8 phr ZDNC and the 2.0 phr DIXP/0.6 phr ZDNC formulations. In contrast, the thickest films (Figure 4.1(o)) had maximum ultimate elongations in 1.4 phr DIXP regardless of the level of ZDNC, and were statistically significantly higher than the minimum ultimate elongations in 2.0 phr DIXP/0.6 phr

ZDNC.

Film Thickness Modulus at Tensile Strength (MPa) Ultimate Elongation (%) (µm) 500% Elongation (MPa)

150

(a) (b) (c)

190

(d) (e) (f)

210

(g) (h) (i)

230

(j) (k) (l)

260

(m) (n) (o)

Fig. 4. 1. 3D graphs of mechanical properties for guayule thin films: modulus at 500% elongation (MPa), tensile strength (MPa), and ultimate elongation (%), while varying the amount of rubber accelerators DIXP (phr) and ZDNC (phr) for different film thicknesses.

DIXP: diisopropyl xanthogen polysulphide; phr: parts per hundred dry rubber; ZDNC: zinc diisononyl dithiocarbamate.

97 Three formulations in Fig. 4.1 were selected for thermal analysis (Table 4.3). The formulation with the best tensile properties (DIXP 2.0 phr/ ZDNC 0.8 phr) was selected, as well as two formulations that had statistically inferior mechanical properties (Table

4.3) and a total accelerator concentration higher (DIXP 2.2 phr/ZDNC 0.4 phr) or lower

(DIXP 1.0 phr/ ZDNC 0.4 phr) than the optimal formulation (Table 4.3). This allows for thermal analysis across formulations with different crosslinking densities, including optimally cured films, an under-crosslinked material that undergoes ductile fracture, and an over-crosslinked material that undergoes brittle fracture.

Table 4.3. Tensile data for GNRL thin films, which will be used for subsequent thermal analysis via DMA, DSC, and TGA.

Sample Tensile Properties: Modulus at Tensile Ultimate 500% Elongation Strength (MPa) Elongation (%) (MPa) Guayule NRL: DIXP 1.0; ZDNC 0.4 0.58 ± 0.04 0.76 ± 0.06 884 ± 41 DIXP 2.0; ZDNC 0.8 1.58 ± 0.02 28.96 ± 2.57 1888 ± 78 DIXP 2.2; ZDNC 0.4 1.68 ± 0.07 18.70 ± 5.64 961± 334

4.3.3. Dynamic Mechanical Analysis (DMA)

The glass transition temperature (Tg) is the temperature at which polymers transition from glassy behavior, where segmental motion is restricted, to rubbery behavior that is characterized by molecular relaxations involving cooperative segmental motion. The segmental motion of polymers allows for reversible strain in a material, the defining behavior of an elastomer. As a result, characterization of the Tg can provide insight into the physio-chemical interactions of materials, and how molecular motion can

98 become hindered due to vulcanization. The Tg typically increases in polymers that undergo more extensive crosslinking, as molecular segmental motion becomes hindered by inter- and intra-molecular crosslinks.

A low frequency of 1 Hz was used to test film samples, since higher frequencies can increase the Tg due to reduced molecular relaxation of the polymeric system. Storage moduli transitions were observed between -54 ºC and -49 ºC for all GNRL films (Table

4.4 and Fig. 4.2) but the initial temperature transition occurred at a lower temperature in the optimal film than in the others. Storage modulus decreased with increasing temperature above Tg (Fig. 4.2). The two suboptimal films, which had different tensile properties (Table 4.3, p <0.05) did not have statistically significant different storage moduli at respective temperatures (Fig. 4.2). However, the film with the optimal performance (DIXP 2.0 phr/ ZDNC 0.8 phr) maintained much higher storage modulus over temperatures above its Tg than the other two films.

Tan delta (δ) is the ratio of loss modulus (E”) to storage modulus (E’), and reflects the energy dissipated as heat during a heating cycle (Fig. 4.2). The damping peak occurred between -39 °C and -37 °C for all three films, higher than the glass transition values derived from the storage modulus (Table 4.4), which is very common in the analysis of elastomers. The loss modulus peak is another way to determine the glass transition: an increase in the loss modulus represents an increase in viscous behavior due to increased cooperative segmental motion. The loss modulus peak occurred between -49

°C and -47 °C (Table 4.4). The storage modulus (Fig. 4.2) of the optimized film (DIXP

2.0 phr/ZDNC 0.8 phr) is significantly stronger above Tg compared to the suboptimial films.

99 Table 4.4. Glass transition temperatures obtained from DMA for GNRL thin films. Sample Glass Transition from DMA o o o Tonset ( C) Tmax ( C) Tmax ( C) from from from Storage Loss Tan Delta Modulus Modulus Guayule NRL: DIXP 1.0; ZDNC 0.4 -51.73 ± 1.23 -49.26 ± 0.91 -38.67 ± 0.70 DIXP 2.0; ZDNC 0.8 -49.55 ± 0.75 -47.40 ± 0.17 -37.82 ± 0.76 DIXP 2.2; ZDNC 0.4 -54.16 ± 0.18 -49.16 ± 0.55 -39.38 ± 0.89

Fig. 4. 2. Glass transition temperatures obtained from DMA for GNRL thin films.

4.3.3. Differential Scanning Calorimetry (DSC)

Glass transition temperatures from DSC were calculated from thermal transitions upon cooling and re-heating optimal and suboptimal cured films, and uncrosslinked

GNRL outlined in Table 4.3. The uncrosslinked GNRL had the lowest glass transition temperature at -60 °C (Table 4.5, Fig. 4.3). As the amount of chemical accelerators increased, the glass transition temperature also increased for GNRL compounds. This is attributed to increased crosslink densities for compounds with increased accelerator concentrations. Glass transition temperatures from DSC (Table 4.5) are significantly lower than the glass transition temperatures calculated from DMA (Table 4.4). Since there are no mechanical forces applied in DSC, it is no surprise that it would require less

100 energy to undergo cooperative segmental motion, resulting in a lower glass transition temperature when compared to DMA data.

Table 4.5. Glass transition temperatures obtained from DSC for GNRL thin films.

Average of four samples.

Sample Glass Transition (Second Melting) o o Teig ( C) Tmg ( C) Guayule NRL: Uncrosslinked -60.39 ± 1.02 -60.25 ± 1.95 DIXP 1.0; ZDNC 0.4 -56.75 ± 2.58 -54.42 ± 2.14 DIXP 2.0; ZDNC 0.8 -54.01 ± 2.12 -49.96 ± 0.74 DIXP 2.2; ZDNC 0.4 -52.45 ± 1.18 -48.25 ± 0.86 Teig = extrapolated onset temperature, Tmg = midpoint temperature

Fig. 4. 3. Differential scanning calorimetry of GNRL formulations. Average of four samples.

4.3.3. Thermogravimetric Analysis (TGA)

Thermal degradation was observed from TGA curves depicting the weight loss of

GNRL (Fig. 4.4) outlined in Table 4. There were no distinctive differences between the three different films’ peak temperature for major weight loss, which was 371 ºC. (Fig.

4.4). The uncrosslinked GNRL had a significantly lower percent residue remaining,

101 compared to formulated GNRL compounds which are attributed to residue from chemical accelerators, initiators, and sulfur used in crosslinking.

Table 4.6. Thermal degradation temperatures obtained from TGA for GNRL thin films.

Average of four samples.

o o Sample Tonset ( C) Tpeak ( C) % Residue Guayule NRL: Uncrosslinked 312.56 ± 2.51 371.60 ± 0.25 0 ± 0.23 DIXP 1.0; ZDNC 0.4 329.41 ±1.23 371.99 ± 1.54 8.33 ± 3.59 DIXP 2.0; ZDNC 0.8 337.43 ± 2.20 371.01 ± 2.59 8.641 ± 2.46 DIXP 2.2; ZDNC 0.4 326.11 ± 1.25 370.25 ± 3.21 4.54 ± 0.89 Tonset = beginning mass loss; Tpeak = maximum mass loss; % Residue= mass remaining (%)

Fig. 4. 4. Thermogravimetric analysis of GNRL formulations. Average of four samples.

102 4.4. Conclusion

Dipped thin film elastomers are employed in medical applications such as surgical balloons, catheters, dental dams, condoms and gloves. One of the most taxing specifications is for surgical gloves, which must protect patients and surgeons during hours of surgery without failing. The ASTM D3577 standard specification for natural rubber surgical gloves demands tensile strength greater than 24 MPa, modulus at 500% elongation of less than 5.5 MPa and ultimate elongation greater than 750% (ASTM

International, 2013b). All three performance parameters are exceeded by many of the thin films formulated and made in this study (Figure 4.1(a) to (o)).

Crosslinking GNRL to confer good physical properties requires a synergistic amount of primary and secondary accelerator and this has to be in balance with the added sulfur crosslinker (Gao et al., 2013, Vennemann et al, 2013). The strongest GNRL films had an accelerator/sulfur ratio of 1.4, which indicates that the accelerator package of 2 phr DIXP/0.8 phr ZDNC is semi-efficient. Semi-efficient accelerator systems are able to promote a higher rate of crosslinking compared to conventional systems, incorporating polysulfidic, di and monosulfidic bonds (Nocil Ltd., 2010).

Vulcanization temperature also significantly impacts crosslink structure and tensile properties. Optimum crosslinks are formed using the lowest temperature possible, regardless of the type of accelerator used (Nocil Ltd., 2010). In our accelerator package, the primary accelerator is DIXP and it has a very short vulcanization plateau which occurs from 80-100 °C (Chakraborty and Couchman, 2006). Therefore, the use of ultra- fast accelerators such as ZDNC increases the rate of cure despite the relatively low cure temperature required. ZDNC, a dithiocarbmate, induces a faster cure rate and higher

103 crosslink density, which can be attributed to higher crosslink density and improved tensile strength of films (Fig. 1B, E, H, K and N) compared to using DIXP as a sole accelerator (Nocil Ltd., 2010).

There are over 50 different types of vulcanization accelerators commercially available, with various accelerator classes utilizing different crosslinking mechanisms

(Nocil Ltd., 2010). Within accelerator classes, reaction kinetics can vary due to differences in molecular weight and solubility in the polymer emulsion (Nocil Ltd.,

2010). Therefore, this study of DIXP and ZDNC is unique to GNRL with its emulsion chemistry containing high amounts of potassium hydroxide. A different stabilization package for GNRL may change accelerator reaction kinetics, crosslink density, and, ultimately, tensile properties.

The GNRL thin films in this study have mechanical properties that exceed other glove films of natural or synthetic origin (Krutzer et al., 2014). Studies of commercially available polychloroprene, anionically polymerized polyisoprene, and zinc polymerized polyisoprene thin films reported lower tensile strengths and higher ultimate elongations than NRL thin films (Krutzer et al., 2014). The crosslinking mechanisms for the commercial films tested were unspecified, and so it is not known whether these films contain accelerators capable of causing a type IV allergic response (Krutzer et al., 2014), although it seems likely.

Type I circumallergenic elastomeric thin films have been created using synthetic polymers such as polychloroprene, and polyisoprene, yet all have poorer tensile properties than the GNRL films in this study (Chen et al., 2011; Tao, 2002).

Polychloroprene films have been reported utilizing DIXP and ZDNC, in addition to the

104 accelerator DPG (a known type IV allergen) but these films contain residual DPG, increasing their type IV allergic potential, and did not even meet the considerably lower performance requirements set by the ASTM standards for synthetic surgical gloves (van

Jole, 2006). Polychloroprene was also used to make type I circumallergenic and type IV hypoallergenic thin films using only DIXP as accelerator, but did not meet ASTM standards for synthetic surgical gloves (Krutzer et al., 2014). However, the GNRL films created in our study have superior ultimate elongations, and are softer than the polychloroprene films (Chen et al., 2011). Softer materials require less force for material deformation, minimizing hand fatigue for surgeons during hours of surgery.

Other type I circumallergenic and type IV hypoallergenic thin film elastomers have been made from nitrile butadiene rubber. Carboxyl functional side groups were added to nitrile monomer units to induce crosslinking without sulfur or accelerators (Tao,

2002). However, these nitrile films have inferior ultimate elongations, and are 2.5x stiffer than the GNRL thin films made in our study (Tao, 2002).

Overall, GNRL thin films have superior properties for glove applications than other type I circumallergenic and type IV hypoallergenic elastomers (Cornish et al., 2007;

Krutzer et al., 2014; Chen et al., 2011; Tao, 2002), creating a unique combination of strength, high elongation and softness (as shown by its low modulus at 500% elongation).

Thus, gloves made from GNRK filkms combine outstanding disease protection with consumer comfort unmatched by other elastomeric materials.

Future research to optimize DIXP and ZDNC in other polymer lattices will determine the extent to which emulsion chemistry and polymer origin impact vulcanization reaction kinetics and ultimately tensile properties. Future quantification of

105 crosslink density, and the type of crosslink (polysulifidic, monosulfidic, or disulfidic) in

GNRL thin films can distinguish the underlying mechanisms of how accelerator concentrations impact tensile properties.

106 Chapter 5: Mechanical and thermal properties of type I & type IV hypoallergenic

Hevea natural rubber latex thin films.

J. Lauren Slutzky a*, and Katrina Cornish a,b aOhio State University, Department of Food, Agricultural and Biological Engineering,

1680 Madison Avenue, Wooster, Ohio 44691 USA bOhio State University, Department of Horticulture and Crop Science, 1680 Madison

Avenue, Wooster, Ohio 44691 USA

*Corresponding author: Katrina Cornish, Williams Hall, 1680 Madison Avenue,

Wooster, Ohio 44691, USA. Tel: 330-263-3982 Fax: 330-263-3887 Email:

[email protected]

Abstract:

Type I latex allergy sensitization and subsequent allergic reactions to Hevea natural rubber latex proteins have created an industry demand for thin film barriers that are hypoallergenic. Other allergens associated with natural rubber and synthetic polymer thin films are attributed to residual thiazole, thiuram, and carbamate accelerators that are prone to cause type IV allergies, characterized by contact dermatitis or delayed contact hypersensitivity. This work focuses on the manufacturing of type I and type IV hypoallergenic thin films utilizing natural rubber latex from the plant species Hevea brasiliensis, cured with the accelerators diisopropyl xanthogen polysulphide (DIXP) and zinc diisononyl dithiocarbamate (ZDNC). Ultra- low protein Hevea latex is hypoallergenic with respect to type I latex allergy, because its extractable protein and rubber bound proteins have been precipitated and discarded prior to compounding and subsequent manufacture. Type IV allergies are diminished because DIXP is consumed

107 during the vulcanization process, and skin tests have shown that ZDNC does not cause dermal reactions or delayed contact hypersensitivity. Fabrication of the type I and type

IV hypoallergenic thin films was carried out using a coagulant dipping method, and the effects of the accelerator loading on the mechanical and thermal properties of cured ultra- low protein Hevea thin films were characterized. Many of the formulated films have mechanical properties superior to those described in American Standard for Testing

Materials D 3577, the standard for rubber surgical gloves, and could be used to make commercial medical thin film barriers such as medical gloves, condoms, balloons and dental dams. Thermal analysis of films also confirmed that accelerator loading affected crosslinking.

Keywords: ultra-low protein Hevea, natural rubber latex, type I latex allergy, type IV latex allergy, elastomer thin films

5.1. Introduction

Over 40,000 consumer products are made from natural rubber, including tires, gloves, catheters, condoms, and clothing elastic (Kelly, 1995). NRL is derived from the plant species, Hevea brasiliensis, commonly known as the Brazilian or Para Rubber Tree.

However, NRL contains over 200 different proteins, over 60 of which are human allergens, capable of eliciting a spectrum of type I allergic responses, ranging in severity from contact dermatitis, contact urticarial, and delayed hypersensitivity through systemic reactions, such as hives and edema, to life-threatening anaphylaxis and death (Spaner et al., 1989).

Repeated exposure to Hevea NRL allergenic proteins can sensitize individuals, inducing IgE antibodies. With increased exposure to Hevea NRL products, an individual

108 is more likely to create IgE antibodies associated with Hevea NRL allergenic proteins: At the height of the crisis 68% of children with spina bifida were sensitized due to frequent spinal catheterization with Hevea NRL products, 5-15% of healthcare workers were sensitized from Hevea NRL gloves, 10% of industrial rubber workers were sensitized, whereas 0.2%-8.2% of the general population was sensitized (Grzybowski et al., 1996;

Alenius et al., 2002).

Direct skin contact by vulcanized elastomer products also can induce type IV delayed hypersensitivity reactions mediated by antigen specific sensitized T lymphocytes, and are attributed to residual accelerators used to crosslink natural as well as synthetic diene elastomers (Pak et al., 2012). Type IV contact dermatitis to rubber products is one of the most common causes of occupational contact dermatitis (Sasseville, 2008). In addition, 50% of individuals with type IV allergy will develop a type I allergy to Hevea

NRL proteins, since contact dermatitis has compromised their dermal barrier and therefore is more susceptible to allergenic protein absorption (Hayes et al., 2000).

Ultra-low protein Hevea NRL is type I hypoallergenic, due to an aluminum hydroxide treatment which removes proteins from the NRL. A slurry of aluminum hydroxide is introduced at the processing stage, binding to the non-rubber particles and soluble protein in NRL (Swason, 2008). These aluminum bound protein particles are subsequently removed by centrifugation (Swason, 2008). In addition to removing antigenic proteins, this process also removes other non-rubber biochemical metabolites and impurities, creating a cleaner and stable ultra-low protein Hevea NRL (Swason,

2008).

109 Type IV allergens in medical natural and synthetic elastomer products are attributed to residual accelerators used to increase the rate and efficacy of sulfur crosslinks formation and in some cases, to anti-oxidants. The commonly used rubber accelerators, thiazoles, thiurams, and carbamates, are recognized by the U.S. Food and

Drug Administration as sensitizing agents capable of eliciting type IV allergic reactions

(Taylor and Leow, 2000). Additional concerns with sulfur vulcanization accelerators include the potential formation of the carcinogenic and teratogenic N-nitrosamines.

Nitrosamines form when nitrites from the environment react with nitrosatable substances, such as secondary amines found in many vulcanization chemistries (Mutsuga, 2013).

Current alternatives for diene elastomer crosslinking include gamma and UV irradiation, organic peroxide cures, zinc oxide activators without accelerators (utilizing the carboxyl-zinc ionic bond), and placement of functional groups onto the polymer backbone that can form crosslinks post product fabrication. However, since gamma and

UV irradiation, and organic peroxides create carbon and not sulfur crosslinks, they generally possess inferior mechanical properties compared to vulcanized rubbers (van

Jole, 2008).

Novel sulfur vulcanization chemistries continue to be developed by chemists, with aims of reducing allergenic potential, mitigating the formation of nitrosamines, and improving vulcanization efficiency. The new chemical accelerators diisopropyl xanthogen polysulfide (DIXP) and zinc diisononyl dithiocarbamate (ZDNC) (developed by Robinson Bros. Ltd., U.K.) have reduced allergenic potential, lack secondary amines, and can be used in tandem for the efficient vulcanization of NRL (Chakraborty, 2005).

DIXP is a fugitive xanthate accelerator which is completely consumed into sulfur

110 crosslinks, and its byproducts are volatile isopropanol and carbon disulfide (Chakraborty,

2005). ZDNC has a lower allergic potential than conventional dithiocarbamates because its high molecular weight renders it soluble in the rubber matrix, and therefore less

ZDNC can be extracted from finished rubber articles compared to common industry dithiocarbamates (Chakraborty, 2005).

The objective of this study was to vulcanize ultra-low protein Hevea NRL with the accelerators DIXP and ZDNC to create type I and type IV hypoallergenic thin film elastomers with mechanical properties suitable for surgical gloves, and other thin film elastomer products, such as catheters and condoms. The materials made in this study are type I and type IV hypoallergenic NRL thin films with mechanical properties superior those specified in American Standard for Testing Materials (ASTM) D 3577, the standard for rubber surgical gloves.

111 5.2. Experimental

5.2.1. Materials and sample preparation

5.2.1.1. Emulsion chemistry/compounding

Ultra-low Hevea NRL (Vystar, U.S.A) was formulated using a base compounding recipe (Table 5.1) while varying the amount of the accelerators DIXP and ZDNC. The amount of accelerators and other materials were added to ultra-low Hevea NRL at specified concentrations based on parts per hundred dry rubber (phr). The sulfur emulsion, antioxidant dispersion, and zinc oxide dispersion were supplied by Akron

Dispersions (Akron, OH, USA) and the ammonium hydroxide was supplied by W.W.

Grainger, Inc (Salt Lake City, UT, USA). Deionized water was added to the compounded emulsion until 48% solids by volume was achieved, and was prevulcanized for 2.5 hours while stirring with a 30 rpm hand mixer.

Table 5.1. Latex Compound Recipe.

Quantity (phr)

Ultra-low protein Hevea NRL 100

Sulfur 2

Ammonium Hydroxide 1

ZnO 0.3

Antioxidant 2

ZDNC 0.2, 0.4, or 0.6

DIXP 0.2, 0.4, 0.6, 0.8, or 1.0

112 5.2.1.2. Thin film manufacture by dipping

Thin film elastomer products are manufactured by dipping formers into emulsions with subsequent heating to remove liquids and vulcanize the NRL. Initially, a stainless steel former was heated to 70 ºC and dipped (10 second dwell) into a coagulant solution

105 ºC). The rubber article was then placed in a tumble dryer post-vulcanization (60 min at 60 ºC). All ultra-low protein Hevea NRL thin films in this study were made using a

Diplomat automated dipper (DipTech Systems Inc., Kent, OH, USA).

5.2.2. Tensile Properties

Four dumbbell samples of each compound at each dwell time were cut using Die

C according to ASTM D 412 (ASTM International, 2013a). Mechanical properties were tested following ASTM D 412 and were determined using an Instron 3366 with Bluehill v. 2.17 software package (Instron, Norwood, MA, USA) (ASTM International, 2013a).

Samples were tested using a crosshead speed of 500 mm/min at room temperature

(26°C). The reported mean ± the standard error from the mean (SE) values are averages of at least 4 samples. In addition, statistical analysis was performed using Minitab, version 16.0 (State College, PA). Significant differences (p-values < 0.05) in tensile data

113 amongst ultra-low protein Hevea NRL compounds for each film thickness were determined using one-way analysis of variance (ANOVA).

5.2.3. Dynamic Mechanical Analyzer (DMA)

The mechanically induced glass transition temperatures of the cured films were studied using a TA Instrument Q800 (New Castle, DE), as previously described

(Monadjemi et al., 2016). The samples were equilibrated at -110°C and subsequently heated to 150 °C with a heating rate of 5 °C/min. The sample dimensions were approximately 25 x 3.0 x 0.7 mm3 with an amplitude of 15 μm, preload of 1N, and frequency of 1 Hz using the tension film clamps. The values presented are averages of four samples.

114 5.2.4. Differential Scanning Calorimetry (DSC)

TA Instruments Q100 (New Castle, DE) differential scanning calorimetry (DSC) was used to determine the state of the crosslinked elastomers as a function of temperature, using a previously described procedure (Musto et al., 2016). The average glass transition temperatures were determined from thermograms obtained by heating 5 to 10 mg samples from -20 to 200 °C at a rate of 10 °C/min. The samples were allowed to anneal at 200 °C for 3 to 5 min to remove the thermal history of the polymers, and subsequently cooled to -85 °C, held for 3 to 5 min, and reheated from -85 to 200 °C with a heating rate of 10 °C/min. The reported values are averages of four samples.

5.2.5. Thermogravimetric Analysis (TGA)

Thermogravimetric analysis (TGA) was used to study thermal decomposition using a TA Instruments TGA 5000 (New Castle, DE). The samples were heated under nitrogen from room temperature to 500 °C with a heating rate of 20° C/min. The reported values are averages of four samples.

115 5.3. Results and Discussion

5.3.1. Dwell time and film thickness

When the film thickness for each respective dwell time was measured for four samples, each set had a standard error less than 0.01 mm (Table 5.2). The relationship between dwell time and average thin film thickness showed a linear relationship

(R2=0.98, df=3).

Table 5.2. Dwell time and average thin film thickness, standard error from mean <10 μm Dwell Time (s) Average Film

Thickness (μm)

5 150

15 190

30 210

45 230

60 260

5.3.2: Tensile properties

The modulus at 500% elongation in the thinnest films was particularly sensitive to the ratio and concentration of the two accelerators (Figure 5.1(a) and (d)). The thinnest films (Figure 5.1(a)) had statistically significant lower modulus for formulations containing 0.2 phr DIXP with ZDNC concentrations of 0.0 and 0.2 phr, in comparison to films containing 0.4 phr DIXP with ZDNC concentrations of 0.2, 0.4, and 0.6 phr. There

116 are no statistical differences in modulus at 500% elongation for the thinnest films (Figure

5.1(a)) between formulations containing higher concentrations of DIXP, such as 0.6 phr

DIXP with 0.4 or 0.6 phr ZDNC, compared to 0.8 phr DIXP with 0.2, 0.4, and 0.6 phr

ZDNC. In the thickest films (Figure 5.1 (e)), modulus at 500% elongation became uniform for formulations containing 1.0 phr DIXP with 0.2, 0.4, and 0.6 phr ZDNC

(Figure 5.1 (e)), whereas in thinner films (Figure 5.1 (b), (c), (d)) the 1.0 phr DIXP/0.2 phr ZDNC formulations are statistically significantly softer for modulus at 500% elongation compared to the formulations containing 1.0 phr DIXP/0.6 phr ZDNC.

Tensile strength varied considerably among all film thicknesses in the tested accelerator treatments. Tensile strength (Figure 5.1 (b), (e), and (h)) was strongest, and similar, in films made with 0.4 phr DIXP/0.4 phr ZDNC, and 0.6 phr DIXP/0.4 phr

ZDNC for the film thicknesses of 150, 190, and 210 μm. The thinnest films of the 0.4 phr

DIXP/0.4 phr ZDNC polymer compound were 150 μm thick and had a tensile strength of

31.38 ± 1.29 MPa (Figure 5.1(b)). Films made with 0.4 phr DIXP/0.4 phr ZDNC, and 0.6 phr DIXP/0.4 phr ZDNC were weaker than films from other formulations at film thicknesses of 150, 190, and 210 μm. Films made with higher concentrations of DIXP in the thicker 230 and 260 μm films had tensile properties similar to films from 0.4 phr

DIXP/0.4 phr ZDNC and 0.6 phr DIXP/0.4 phr ZDNC. Relative minima in tensile strength were observed in 0.2 phr DIXP/0.0 phr ZDNC (Figure 5.1 (b) (e) and (k)), which were statistically significantlky weaker than films made with 0.4 phr DIXP/0.4 phr

ZDNC and 0.6 phr DIXP/0.4 phr ZDNC.

There was little statistical significance among formulations for ultimate elongation. The largest ultimate elongations, in the thinnest films of 150 μm, were

117 obtained in 0.8 phr DIXP/0.4 phr ZDNC (Figure 5.1 (c)), and had an elongation at break of 1686 ± 64%. This film was only significantly different from the group of films with the lowest ultimate elongations, i.e., 0.2 phr DIXP/0.0 phr ZDNC, 0.2 phr DIXP/0.2 phr

ZDNC, and 0.4 phr DIXP/0.6 phr ZDNC. For the films 190 μm thick, the largest ultimate elongations were in films made with 0.6 phr DIXP/0.2 phr ZDNC formulation

(Figure 5.1 (f)), and had an elongation at break of 2107 ± 42%, and was significantly higher than the other films of this thickness. The highest ultimate elongation in thickest films (260 μm) were also obtained from this fomulation, (Figure 5.1 (o)), and had an elongation at break of 2204 ± 299%. However, only the films with the ultimate elongation minimum, (made with 0.2 phr DIXP/0.0 phr ZDNC) had significantly lower elgonation at break.

118 Film Thickness Modulus at 500% Elongation Tensile Strength (MPa) Ultimate Elongation (%) (µm) (MPa)

150

(a) (b) (c)

190

(d) (e) (f)

210

(g) (h) (i)

230

(j) (k) (l)

260

(m) (o) (n) Fig. 5. 1. 3D graphs of mechanical properties for ultra-low protein Hevea NRL thin films: modulus at 500% elongation (MPa), tensile strength (MPa), and ultimate elongation (%), while varying the amount of rubber accelerators DIXP (phr) and ZDNC

(phr) for different film thicknesses. DIXP: diisopropyl xanthogen polysulphide; phr: parts per hundred dry rubber; ZDNC: zinc diisononyl dithiocarbamate.

119 Three formulations in Figure 5.1 were selected for thermal analysis (Table 5.3).

The formulation with the best tensile properties (DIXP 0.4 phr/ZDNC 0.4 phr) was selected, as well as two formulations that had statistically inferior mechanical properties and a total accelerator concentration higher or lower than the optimal formulation (Table

5.3). This allows for thermal analysis across formulations with different crosslinking densities, including optimally-cured films (DIXP 0.4 phr /ZDNC 0.4 phr), an under- crosslinked material (DIXP 0.2 phr /ZDNC 0.2 phr) that undergoes ductile fracture, and an over-crosslinked material (DIXP 1.0 phr/ZDNC 0.2 phr) that undergoes brittle fracture.

Table 5.3. Tensile data for ultra-low protein Hevea NRL thin films, which will be used for subsequent thermal analysis via DMA, DSC, and TGA.

Sample Tensile Properties: Modulus at Tensile Ultimate 500% Elongation Strength (MPa) Elongation (%) (MPa) Ultra-low Protein Hevea NRL: DIXP 0.2; ZDNC 0.2 1.19 ± 0.08 6.30 ± 0.44 1238 ± 36 DIXP 0.4; ZDNC 0.4 2.14 ± 0.04 31.38 ± 1.29 1590 ± 31 DIXP 1.0; ZDNC 0.6 2.51 ± 0.08 23.35 ± 2.52 1496 ± 45

120 5.3.3. Dynamic Mechanical Analysis (DMA)

A low frequency of 1 Hz was used for the testing of samples, since higher frequencies can increase the Tg due to reduced molecular relaxation of the polymeric system. The storage moduli transitions were observed between -56ºC and -53ºC for all ultra-low protein Hevea NRL (Table 5.4 and Figure 5.2). Ultra-low protein Hevea NRL showed a significant decrease in storage modulus with increasing temperature (Figure

5.2). However, ultra-low protein Hevea NRL formulations with statistically significant tensile properties (Table 5.3, p <0.05) did not have statistically significant different storage moduli at respective temperatures (Table 5.4).

Tan delta (δ) is the ratio of loss modulus (E”) to storage modulus (E’), and reflects the energy dissipated as heat during a heating cycle (Figure 5.2). The dampening peak for ultra-low protein Hevea NRL formulations occurred between -44 °C and -37 °C, and tan delta peak values are higher than the glass transition values derived from the storage modulus (Table 5.4), which is very common in the analysis of elastomers. The

121 loss modulus peak is another way to determine the glass transition: an increase in the loss modulus represents an increase in viscous behavior due to increased cooperative segmental motion. The loss modulus peak for ultra-low protein Hevea NRL formulations occurred between -54 °C and -50 °C (Table 5.4).

Table 5.4. Glass transition temperatures obtained from DMA for ultra-low protein Hevea

NRL thin films.

Sample Glass Transition from DMA o o o Tonset ( C) Tmax ( C) Tmax ( C) from from from Storage Loss Tan Delta Modulus Modulus Ultra-low Protein Hevea DIXP 0.2;NRL: ZDNC 0.2 -56.65 ± 0.23 -54.37 ± 1.88 -41.74 ± 3.16 DIXP 0.4; ZDNC 0.4 -53.20 ± 0.77 -50.30 ±0.34 -36.95 ± 0.59 DIXP 1.0; ZDNC 0.6 -56.60 ± 0.37 -54.00 ± 0.51 -44.05 ± 1.53

Fig. 5. 2. Glass transition temperatures obtained from DMA for ultra-low protein Hevea

NRL thin films.

122 5.3.3. Differential Scanning Calorimetry (DSC)

Glass transition temperatures from DSC were calculated from thermal transitions upon cooling and re-heating of ultra-low protein Hevea samples outlined in Table 5.3.

Uncrosslinked ultra-low protein Hevea NRL was included for DSC analysis; the uncrosslinked material is ultra-low protein Hevea NRL latex dried at ambient temperatures for 1 week, and does not contain any formulation additives. The uncrosslinked ultra-low protein Hevea NRL had the lowest glass transition temperature at

-67.5°C (Figure 5.3, Table 5.5). As the amount of chemical accelerators increased, the glass transition temperature also increased for ultra-low protein Hevea NRL compounds.

This is attributed to increased crosslink densities for compounds with increased accelerator concentrations. Glass transition temperatures from DSC (Table 5.5) are significantly lower than the glass transition temperatures calculated from DMA (Table

5.4). Since there are no mechanical forces applied in DSC, it is no surprise that it would require less energy to undergo cooperative segmental motion, resulting in a lower glass transition temperature than estimated by DMA data.

Table 5.5. Glass transition temperatures obtained from DSC for ultra-low protein Hevea

NRL thin films. Average of four samples.

Sample Glass Transition (Second Melting) o o Teig ( C) Tmg ( C) Ultra-low Protein Hevea NRL: Uncrosslinked -67.50 ± 1.25 -64.81 ± 0.69 DIXP 0.2; ZDNC 0.2 -59.86 ± 0.23 -58.30 ± 1.68 DIXP 0.4; ZDNC 0.4 -59.26 ± 0.84 -57.31 ± 1.48 DIXP 1.0; ZDNC 0.6 -57.35 ± 1.54 -56.51 ± 0.87 Teig = extrapolated onset temperature, Tmg = midpoint temperature

123

Fig. 5. 3. Differential scanning calorimetry of ultra-low protein Hevea NRL formulations. Average of four samples.

5.3.3. Thermogravimetric Analysis (TGA)

Thermal degradation was determined from TGA curves depicting the weight loss of the ultra-low protein Hevea NRL films outlined in Table 5.5, and uncrosslinked

(ambient dried without additives) ultra-low protein Hevea NRL (Figure 5.4). The onset temperature and peak temperature of degradation were significantly different from uncrosslinked ultra-low protein Hevea NRL compared to all formulated samples with crosslinking. The film with the best mechanical properties (0.4 phr DIXP/0.4 phr ZDNC) had significantly lower onset temperature and peak temperature of degradation than all other films (Table 5.5). The uncrosslinked ultra-low protein Hevea NRL had a significantly lower percent residue remaining, compared to formulated ultra-low protein

Hevea NRL compounds which are attributed to residue from chemical accelerators, initiators, and sulfur used in crosslinking.

124

Table 5.6. Thermogravimetric data for ultra-low protein Hevea NRL thin films. Average of four samples.

o o Sample Tonset ( C) Tpeak ( C) % Residue Ultra-low Protein Hevea NRL: Uncrosslinked 307.37 ± 3.43 365.38 ± 3.45 1.018 ± 0.12 DIXP 0.2; ZDNC 0.2 323.58 ± 2.32 373.44 ± 2.57 4.697 ± 1.57 DIXP 0.4; ZDNC 0.4 285.93 ± 1.34 359.09 ± 1.97 5.537 ± 1.29 DIXP 1.0; ZDNC 0.6 318.28 ± 2.75 372.48 ± 1.56 7.727 ± 2.25

Fig. 5. 4. Thermogravimetric analysis of ultra-low protein Hevea formulations. Average of four samples.

125 5.4. Conclusions

Type I and type IV hypoallergenic thin film elastomers can be used in applications such as surgical gloves. The specifications for surgical gloves is detailed in

ASTM D3577: natural rubber surgical gloves demands tensile strength greater than 24

MPa, modulus at 500% elongation of less than 5.5 MPa and ultimate elongation greater than 750% [16]. All three performance parameters are exceeded by many of the thin films formulated in this study (Figure 5.1).

The films had significantly different tensile properties based on formulation.

Thermal differences between films were not statistically significant; however all films had a higher glass transition temperature compared to uncrosslinked ultra-low protein

Hevea NRL. Therefore accelerator ratio and concentration affects tensile properties; however, the accelerator ratio and concentration does not affect thermal properties of ultra-low protein Hevea NRL films, such as glass transition temperature. Thermal measurements such as glass transition temperature cannot provide the resolution of crosslink density and quality, but do provide significant changes compared to uncrosslinked, ambient dried ultra-low protein Hevea NRL.

To quantify significant property differences between ultra-low protein Hevea

NRL films, mechanical properties from tensile testing should be prioritized as a result to determine the effect of crosslinking on film properties. The storage modulus of ultra-low protein Hevea NRL films is not significantly different between films, and therefore only tensile properties should used to assess films to determine optimal formulation and properties.

126 The strongest ultra-low protein Hevea NRL films (DIXP 0.4 phr/ ZDNC 0.4 phr) has an accelerator ratio of 1:1; whereas the strongest guayule NRL films in Chapter 4

(DIXP 2.0 phr/ ZDNC 0.8 phr) has an accelerator ratio of 2.5:1. This vast difference in chemical accelerator loadings and ratios between optimized formulations can be attributed to differences in the NRL chemistry. All NRL emulsions are buffered to a alkaline pH to extend shelf-life; Ultra-low protein Hevea NRL is stabilized with ammonium hydroxide, whereas GNRL is stabilized with potassium hydroxide. The differences in the stabilization packages of NRL needs to be considered when determining an ideal accelerator concentration and ratio, specifically due to the role of ammonium hydroxide and zinc in creating efficient carbon-sulfur bonds. NRL stabilized with ammonium hydroxide such as ultra-low protein Hevea NRL have excess ammonium hydroxide; ammonium hydroxide is an activator for zinc-based accelerators (such as

ZDNC) and can promote carbon-sulfur bond formation. Future work could focus on understanding how ammonium hydroxide concentration in NRL affects accelerator concentration and ratio.

127 Chapter 6: Canonical correlation analysis of type I and type IV circumallergenic guayule natural rubber thin films and type I and type IV hypoallergenic ultra-low protein Hevea natural rubber thin films.

J. Lauren Slutzky a, Alfred Soboyejo a , and Katrina Cornish a,b* aOhio State University, Department of Food, Agricultural and Biological Engineering,

1680 Madison Avenue, Wooster, Ohio 44691 USA bOhio State University, Department of Horticulture and Crop Science, 1680 Madison

Avenue, Wooster, Ohio 44691 USA

*Corresponding author: Katrina Cornish, Williams Hall, 1680 Madison Avenue,

Wooster, Ohio 44691, USA. Tel: 330-263-3982 Fax: 330-263-3887 Email:

[email protected]

Abstract:

Parthenium argentatum Gray, commonly called guayule, is an industrial crop that produces low protein laex containing proteins that do not cross-react with antibodies raised against Hevea brasiliensis Muell. Arg. associated allergic proteins. In Chapter 4, guayule natural rubber latex (GNRL) (Chapter 4) and hypoallergenic ultra-low protein

Hevea natural rubber latex (NRL) (Chapter 5) thin films were cured with the accelerators diisopropyl xanthogen polysulphide (DIXP) and alkyldithiocarbamate (ZDNC), and the tensile properties were assessed according to the American Standard for Testing

Materials (ASTM) 3577, which defines surgical glove specifications. The aim of this chapter was to develop multivariate stochastic regression models using ASTM 3577 specifications such as modulus at 500% elongation (MPa), tensile strength (MPa), and ultimate elongation (%) as a function of the concentration of chemical accelerators DIXP

128 and ZDNC, as well as film thickness, for GNRL and ultra-low protein Hevea NRL. In addition canonical correlation analysis (CCA) was used to demonstrate the individual effect of DIXP, ZDNC, and film thickness on the tensile properties. CCA was used to determine Pearson correlation (R2) relationships within and between the input variables of DIXP, ZDNC, and film thickness and the output variables of the tensile properties.

CCA was also used in this study to maximize the correlation between input variables and output variables by calculating a singular value decomposition correlation (R2), which was used to compare the linear, multiplicative, and predictor transformed multivariate regression methods. The abstraction and integration of correlations from multivariate stochastic regression analysis and CCA can reduce the time needed to improve product development by reducing the range of input variable which must be used to test performance.

Keywords: Guayule, natural rubber latex, multivariate stochastic regression, canonical correlation analysis

6.1. Introduction

Over 2,500 plants produce rubber, but only a few plants produce rubber with high molecular weight capable of macromolecular entanglement for commercial applications

(Perumal et al., 2013). Currently, Hevea brasiliensis Muell. Arg. accounts for over 90% of the global supply of natural rubber (Perumal et al., 2013). However, latex from Hevea contains proteins that are capable of sensitizing and eventually eliciting a type I allergic reaction, characterized by skin inflammation, nausea, and perhaps anaphylaxis and ultimately death (Cornish and Siler, 1996). There are over a dozen serious human allergens found in Hevea latex (Carey et al., 1995; Miri et al., 2007).

129 Natural rubber latex derived from the guayule shrub circumvents the Hevea type I latex allergy; guayule does not contain any of the Hevea associated allergens (Cornish et al., 2006; Cornish, 2012). Also, GNRL contains very little protein. Therefore, guayule natural rubber latex (GNRL) does not pose a threat to consumers with regards to developing a type I latex allergy, and is safe to use by consumers with type I latex allergies (Cornish et al., 2007).

Natural rubber is typically crosslinked in order to have mechanical properties suitable for commercial applications. Natural rubber latex thin film elastomers utilize the diene bonds in the cis-1,4-polyisoprene structure for crosslinking with sulfur, which hinders macromolecular translational movement and improves bulk mechanical properties. Activators and accelerators are used to improve the rate and efficiency of sulfur crosslink formation. Commonly used accelerators such as thiurams, thiazoles, and carbamates are acknowledged by the U.S. Food and Drug Administration as capable of eliciting type IV allergies, characterized by contact dermatitis and compromise of the dermal barrier (Miri et al., 2007).

As a result, alternative accelerators have been developed by Robinson Brothers

Ltd (West Bromwich, U.K.), namely zinc diisononyl dithiocarbamate (ZDNC) and diisopropyl xanthogen (DIXP). ZDNC has a higher molecular weight compared to other thiocarbamates, which hinders ZDNC from blooming to the finished rubber article’s surface (Chakraborty et al., 2006). DIXP leaves no residual chemicals, generating the volatile byproducts of isopropanol and carbon disulfide (Chakraborty et al., 2006; van

Jole, 2007). In Chapter 4, DIXP in conjunction with ZDNC was used to create type I circumallergenic and type IV hypoallergenic guayule natural rubber latex (GNRL)

130 elastomeric thin films; In Chapter 5, type I and type IV hypoallergenic ultra-low protein

Hevea natural rubber latex (NRL) elastomeric thin films were created.

This chapter uses the data from Chapter 4 and 5, in multivariate stochastic regression analysis, to establish quantitative structure-property relationships (QSRPs), between the predictor variables of accelerator concentration and film thicknesses, and the corresponding response variables of the tensile properties outlined in ASTM 3577.

Multivariate models using parametric regression, via the linear least squares method are used. Both linear and nonlinear regression models were used for the simulation, as well as models developed from a predictor variable transformation. To compare models, coefficient analysis to quantify the effects of each independent variable on individual tensile properties, and the Pearson correlation coefficient (R2), were compared among models to determine the best fit.

This study used multivariate ordination methods to reduce dimensionality as well as to recognize patterns in these multivariate data sets, specifically by using canonical correlation analysis (CCA). Ordinations were divided into two groups: unconstrained and constrained methods. Unconstrained methods included principle component analysis, and reduced dimensions on the basis of minimizing residual variance (Soboyejo,

2011). Unconstrained ordination methods were used to recognize relative dispersions among groups of data (Soboyejo, 2011). In contrast, constrained ordination methods relate a matrix of response variables with a matrix of quantitative predictor variables, and are used in this study (Soboyejo, 2011). Specifically, CCA utilizes ordination axes that maximize their correlation with linear combinations of some quantitative predictor variables, producing a singular value decomposition correlation (R2) (Rickman et al.,

131 2017). Constrained ordination methods have been termed a “direct” gradient response

(Rickman et al., 2017).

For each type of multivariate stochastic regression modeling technique used, the transformed data was organized into matrices of predictor variables and response variables for CCA. CCA is widely used in parametric multivariate data analysis to analyze correlations within and between multiple independent (predictor) variables and multiple dependent (response) variables. In this study, the predictor variables are the accelerator concentrations of DIXP and ZDNC, and film thickness; the response variables are the tensile properties of modulus at 500% (MPa), tensile strength (MPa), and ultimate elongation (%). CCA was used in the regression analysis to comprehensively evaluate the reliability of the proposed models, to determine if one type of regression modeling could be used for all of the response variables by comparative analysis of singular value decomposition correlation (R2) for each type of data transformation.

132 6.2. Experimental

6.2.1. Materials and sample preparation

6.2.1.1. Emulsion chemistry/compounding

GNRL and ultra-low protein Hevea NRL were made using a base compounding recipe (Table 6.1 and Table 6.2) while varying the amount of the accelerators DIXP and

ZDNC (Robinson Brothers Ltd, UK). The amount of accelerators and other materials were added to GNRL at specified concentrations based on parts per hundred dry rubber

(phr). The sulfur emulsion, antioxidant dispersion, and zinc oxide dispersion were supplied by Akron Dispersions (Akron, OH, USA) and the ammonium hydroxide was supplied by W.W. Grainger, Inc (Salt Lake City, UT, USA). Deionized water was added to the compounded emulsion until 48% solids by volume was achieved, and was prevulcanized for 2.5 hours while stirring with a 30 rpm hand mixer.

Table 6.1. Latex Compound Recipe for Guayule Natural Rubber Latex

Dry Weight Guayule NRL 100 Sulfur 2 Ammonium Hydroxide 1 ZnO 0.3 Antioxidant 2 ZDNC 0.2, 0.4, 0.6, 0.8, 1.0, 1.2 or 1.4 DIXP 1, 1.2, 1.4, 1.6, 1.8, 2.0, 2.1 or 2.2

Table 6.2. Latex Compounding Recipe for Ultra-low Protein Hevea NRL

Dry Weight Ultra-low Protein Hevea NRL 100 Sulfur 2 Ammonium Hydroxide 1 ZnO 0.3 Antioxidant 2 ZDNC 0.2, 0.4, or 0.6 DIXP 0.2, 0.4, 0.6, 0.8, or 1.0

133 6.2.1.2. Thin film manufacture by dipping

105 ºC). The rubber article was then placed in a tumble dryer post-vulcanization (60 min at 60 ºC). All thin films in this study were made using a Diplomat automated dipper

(DipTech Systems Inc., Kent, OH, USA).

6.2.2. Tensile Properties

Four dumbbell samples of each compound at each dwell time were cut using Die

C according to ASTM D 412 (ASTM International, 2013a). Tensile mechanical properties were determined followed ASTM D 412 using a tensiometer (Model 3366,

Instron, Norwood, MA, USA). As outlined in ASTM 3577, the surgical glove standard, tensile properties of modulus at 500% (MPa), tensile strength (MPa), and ultimate elongation (%) were calculated for each sample (ASTM International, 2013b). The reported mean ± the standard error from the mean (SE) values are averages of at least 4 samples. In addition, statistical analysis was performed using Minitab, version 16.0

(State College, PA). Significant differences (P-values < 0.05) in tensile data amongst

134 GNRL compounds for each film thickness were determined in Chapter 4 using one-way analysis of variance (ANOVA), and ultra-low protein Hevea NRL compounds were assessed for significant differences (P-values < 0.05) in tensile data in Chapter 5.

6.2.3. Statistical Analysis

6.2.3.1. Data

6.2.3.1.1 Dwell Time and Film Thickness

When the film thickness for each respective dwell time was measured for four samples, each has a standard error less than 0.01mm (Table 6.3). The relationship between dwell time and average thin film thickness showed a linear relationship

(R2=0.98, df=3). This relationship was consistent between GNRL and ultra-low protein

Hevea NRL thin films.

Table 6.3. Dwell time and average thin film thickness, SEM < 0.01mm.

Dwell Time (s) Average Film Thickness (mm) 5 0.15 15 0.19 30 0.21 45 0.23 60 0.26

6.2.3.1.2 Tensile Data

Mechanical properties for GNRL and ultra-low protein Hevea NRL thin films used in this study were determined in Chapters 4 and 5. The data used includes mechanical properties of vulcanized thin films of varying thicknesses, using the type IV hypoallergenic accelerators, DIXP and ZDNC.

135 The variables were categorized into response variables (Yi) which represent the different mechanical properties of the vulcanized thin films, and explanatory variables

(Xi), which represent the film thicknesses, and loading of chemical accelerators DIXP and ZDNC (Table 6.4).

Table 6.4. Notation of different experimental data used for statistical analysis.

Variable Notation Experimental Data for Model X1 Film Thickness (mm) X2 DIXP (PHR) X3 ZDNC (PHR) Y1 Stress at 500% Elongation (MPa) Y2 Tensile Strength (MPa) Y3 Ultimate Elongation (%) 6.2.3.2. Multivariate stochastic regression models

Multivariate stochastic regression models were made using JMP 11 Statistical

Analysis Software (SAS Institute Inc., Cary, NC). The approach used film thickness and concentration of DIXP and ZDNC accelerators as explanatory variables (Table 6.4) to predict the thin films’ mechanical properties as designated by ASTM 3577.

The first multivariate stochastic regression models were developed using a linear model, summarized by the notation:

푘

푌 = 푎 + ∑ 푏푖푥푖 푖=1 where the predictors X1, X2, X3 are the same for each independent model of Y1, Y2, and

Y3 (Soboyejo, 2011).

Next, multiplicative non-linear models were created. The notation for a multiplicative non-linear model is summarized by the following:

136 푘 푏푖 푌 = 푎 ∏ 푥푖 푖=1 where the predictors X1, X2, X3 are the same for each independent model of Y1, Y2, and

Y3 (Soboyejo, 2011).

The mixed models of non-linear and linear regression models are summarized by the following notation: 푋푖 푌 = 푘 푎 + ∑푖=1 푏푖푥푖

푙푛푋푖 ln 푌 = 푘 푎 + ∑푖=1 푏푖푙푛푥푖

푘 푋 푍 = 푖 = 푎 + ∑ 푏 푥 푖 푌 푖 푖 푖=1

푘 푙푛푥푖 = 푎 + ∑ 푏푖푥푖 푙푛푦푖 푖=1

For i=1, 2…k

Using this notation, mixed model regressions were made with each correlated predictor value (termed the significant predictor variable) (Soboyejo, 2011).

Modeling techniques for the multivariate regression analysis were assessed using the square of multiple correlation (R2) for a multivariate regression model for each tensile property assessed. The R2 value for each mechanical property was calculated for a series of regression models including linear, nonlinear, and mixed significant predictor modeling.

137 6.2.3.3. Canonical correlation analysis (CCA)

The relationships between the sets of processing variables (DIXP concentration,

ZDNC concentration, and film thickness) and tensile property variables (modulus at 500% elongation, tensile strength, and ultimate elongation) were investigated to create linear composites of the respective variable sets. In CCA, the correlation of interest was between the linear variates created for thin film manufacturing and tensile sets of data

(Rickman et al., 2017). CCA initially calculates a canonical loading for each individual data set (i.e., the manufacturing data, and tensile data), and then calculates a canonical correlation from the canonical variates for each respective data set (Rickman et al., 2017).

CCA can determine correlations between manufacturing data and tensile data, as well as show correlations within tensile data (Rickman et al., 2017).

138 6.3 Results and Discussion

6.3.1. Multivariate stochastic regression models

The R2 values for the linear regression for all three models in both types of latex were below statistical significance of 0.95 and therefore, the data was modeled using a multiplicative model. The R2 values for all multiplicative non-linear models were below statistical significance of 0.95, and so the mixing of linear and non-linear models was used next to model the tensile data. Mixed model (linear and non-linear) regressions were made with each correlated predictor value (termed the significant predictor variable), as summarized in Table 6.5, Table 6.6 and Table 6.7.

Table 6.5. Mixed linear and non-linear stochastic multivariate regression models using film thickness (mm) as the significant predictor. Regression Equations R2 Guayule NRL Formulations: [ln(X1)/ln(Y1 )]= -7.27 – 2.03(ln(X1)) + 1.16(ln(X2)) – 0.21(ln(X3)) 0.014 [ln(X 1)/ln(Y2)] = -.60 -0.438(ln(X1)) – 1.71(ln(X2)) – 0.459(ln(X3)) 0.138 [ln(X 1)/ln(Y3)] = 0.019 + 0.152(ln(X1)) + 0.0175(ln(X2)) + 0.886 0.00925(ln(X3)) Ultra-low Protein Hevea NRL Formulations: [ln(X1)/ln(Y1 )]= 1.47 + 3.26(ln(X1)) + 4.04(ln(X2)) – 3.21(ln(X3)) 0.957 [ln(X 1)/ln(Y2)] = 0.358 + 0.507(ln(X1)) + 0.0887(ln(X2)) + 0.726 [ln(X0.00862(ln(X1)/ln(Y3)]3)) = 0.0214 + 0.148(ln(X1)) + 0.00093(ln(X2)) + 0.977 0.000149(ln(X 3))

139 Table 6.6. Mixed linear and non-linear stochastic multivariate regression models using DIXP (phr) as the significant predictor.

Regression Equations R2 Guayule NRL Formulations:

[ln(X2)/ln(Y1 )]= 1.14 + 0.621(ln(X1)) + 1.96(ln(X2)) +0.175(ln(X3)) 0.396

[ln(X 2)/ln(Y2)] = -0.119 – 0.0792(ln(X1)) +0.337(ln(X2)) -0.0158(ln(X3)) 0.680

[ln(X 2)/ln(Y3)] = -0.00611-0.00344ln(X1)) + 0.137(ln(X2)) – 0.00161(ln(X3)) 0.996

Ultra-low Protein Hevea NRL Formulations:

[ln(X2)/ln(Y1 )]= 0.41 + 1.29(ln(X1)) + 4.69(ln(X2)) - 2.72(ln(X3)) 0.944

[ln(X 2)/ln(Y2)] = 0.149 + 0.0740(ln(X1)) +0.382(ln(X2)) + 0.00831(ln(X3)) 0.947

[ln(X 2)/ln(Y3)] = 0.00837+ 0.00459ln(X1)) + 0.137(ln(X2)) – 0.00161(ln(X3)) 0.999

Table 6.7. Mixed linear and non-linear stochastic multivariate regression models using ZDNC (PHR) as the significant predictor.

Regression Equations R2 Guayule NRL Formulations: [ln(X3)/ln(Y1 )]= -2.74 -1.45(ln(X1)) + 0.97(ln(X2)) + 2.15(ln(X3)) 0.146 [ln(X 3)/ln(Y2)] = -0.346 – 0.451(ln(X1)) -0.893(ln(X2)) +0.150(ln(X3)) 0.176 [ln(X 3)/ln(Y3)] = 0.00267+0.00376ln(X1)) + 0.00899(ln(X2)) + 0.141(ln(X3)) 0.992

Ultra-low Protein Hevea NRL Formulations: [ln(X3)/ln(Y1 )]= -50.6 -2.2(ln(X1)) + 14.6(ln(X2)) - 51(ln(X3)) 0.986 [ln(X 3)/ln(Y2)] = 0.442 + 0.120(ln(X1)) +0.0235(ln(X2)) +0.546(ln(X3)) 0.997 [ln(X 3)/ln(Y3)] = 0.0193 + 0.00849ln(X1)) + 0.00053(ln(X2)) + 0.140(ln(X3)) 0.999

The multivariate models created showed that DIXP and ZDNC were both

significant predictor variables for response variables of stress at 500% elongation (MPa),

tensile strength (MPa), and ultimate elongation (%) in GNRL and ultra-low protein

Hevea NRL thin films. Comparison of GNRL and ultra-low protein Hevea NRL

compounding data and multivariate models give insight into differences in their emulsion

chemistry. Multivariate models of GNRL were only significant for response variables of

ultimate elongation when using a mixed-model with DIXP and ZDNC as the significant

predictors. In contrast, similar multivariate models for ultra-low protein Hevea NRL were

140 statistically significant for all three response variables of stress at 500% elongation, tensile strength, and ultimate elongation. This indicates that latex origin must be considered during compounding studies.

6.3.2. Canonical Correlation Analysis

A CCA was performed on the data used to make the aforementioned models

(Tables 6.5-6.7). The matrices of predictor variables and response variables for CCA were the same as used in multivariate analysis (Table 6.4).

141 Table 6.8. Canonical correlation values for multivariate models of tensile properties.

Ultra-low

Type of Model Guayule Protein Hevea

NRL Canonical NRL Canonical

Correlation (R2) Correlation

(R2)

Linear Models 0.6226 0.8780

Multiplicative Models 0.6349 0.9085

Mixed Models with Thickness as Significant Predictor 0.9527 0.9918

Mixed Models with DIXP as Significant Predictor 0.9986 0.9998

Mixed Models with ZDNC as Significant Predictor 0.9972 0.9999

The canonical correlations in Table 6.8 show an overall improvement in correlation as the data were transformed by the different models. Canonical correlations of untransformed data (linear models) have an R2 of 0.6226, and 0.8780 for GNRL and ultra-low protein Hevea NRL, respectively. Natural log transformations of the data

(multiplicative models) only slightly improved canonical correlation correlations. Further data transformations, such as the significant predictor transformations, improved the canonical correlation values. Film thickness, DIXP, and ZDNC as significant predictors further improved the canonical correlation values, with DIXP and ZDNC achieving over

0.99 (Table 6.8) for both GNRL and ultra-low protein Hevea NRL. Therefore, the significant predictor transformations of the data were effective, and improved correlations among the data sets of predictor variables correlated to response variables.

142 6.4. Conclusions

The statistical modeling approach proved to be an effective way to estimate how film thickness, and accelerator amount and type individually contribute to the mechanical properties of both GNRL and ultra-low protein Hevea NRL thin film elastomers. This is notable due to the complexity of GNRL and ultra-low protein Hevea NRL systems and the large number of chemical accelerators that can be optimized in these lattices. In addition, this modeling approach can be used to predict properties in various lattices, compounding formulations, and film geometry (such as film thickness). Loading of DIXP and ZDNC were significant for elongation at break in GNRL thin films. However, we were not able to model the effect of DIXP and ZDNC on mechanical stress, including stress at 300% elongation, and stress at break. This is most likely due to complexity in the GNRL system, such as the role of the latex buffering system on crosslink formation, and biochemical metabolites in the lattice such as phospholipids. Therefore, it is clear that modeling of

GNRL systems is complex and needs to account for additional predictor variables outside the scope of this study to increase statistical significance.

In contrast, this modeling approach using the ultra-low protein Hevea NRL data set was effective in estimating the individual contributions of chemical accelerator loading and film thickness on thin film mechanical properties. The ultra-low protein Hevea NRL system found that all tensile standards as outlined in ASTM 3577 are effectively modeled, in contrast to only ultimate elongation in the GNRL data set. This shows that ultra-low protein Hevea NRL is more predictive lattice in formulation and process conditions modeling; this may be attributed to a less complex lattice compared to GNRL. Ultra-low protein Hevea NRL is buffered in ammonia hydroxide, which is ideal for crosslinking of

143 thin film elastomers (and may lead to more predicative performance) compared to potassium hydroxide stabilized GNRL. Including predictor variables that quantify buffering systems in lattices may improve the effectiveness of the mixed modeling of these latex systems and provide insight into what predictor variables dictate mechanical properties.

144 Chapter 7: Mechanical properties of guayule natural rubber latex thin film

composites with biobased fillers.

J. Lauren Slutzky a*, and Katrina Cornish a,b aOhio State University, Department of Food, Agricultural and Biological Engineering,

1680 Madison Avenue, Wooster, Ohio 44691 USA bOhio State University, Department of Horticulture and Crop Science, 1680 Madison

Avenue, Wooster, Ohio 44691 USA

*Corresponding author: Katrina Cornish, Williams Hall, 1680 Madison Avenue,

Wooster, Ohio 44691, USA. Tel: 330-263-3982 Fax: 330-263-3887 Email:

[email protected]

Abstract:

Many natural rubber latex consumer products contain fillers used to diminish costs, or create new desirable physical properties that aid in processability or utility. Yet these commonly used fillers are not sustainable industries, and dedicate energy and natural resources such as petroleum for their production. Thus, sustainable alternatives for natural rubber latex composites need to be developed. This work focuses on the manufacturing of natural rubber latex composites from the plant species Parthenium argentatum, commonly referred to as guayule, compounded with fillers derived from waste streams: calcium carbonate from eggshells (mostly calcium carbonate), carbon fly ash, and guayule bark bagasse. Tensile properties of composites were assessed according to ASTM D 412 to assess how type of sustainable filler, as well as filler loading and particle size effect mechanical properties.

145 In general, all latex composites had stronger tensile properties with smaller particle sizes at lower loadings. The tensile strength of latex composites at high loadings were often inferior compared to latex films without fillers, but a reinforcing effect was found in low loadings of nano-sized fillers. In general, the elongation at break was increased in latex composites compared to latex films without fillers. Many of the composites still exceeded the tensile requirements described in ASTM D 3577, the surgical glove standard. Latex composites with sustainable fillers can create polymer films with qualities desirable in many applications. The use of such fillers can create novel materials and decrease cost of manufacture.

Keywords: Guayule, natural rubber latex composites, biofillers, calcium carbonate, carbon fly ash, guayule bagasse, elastomer thin films

7.1. Introduction

Guayule natural rubber latex (GNRL) is a colloidal suspension of both rubber and non-rubber constituents, derived from the rubber producing perennial shrub, Parthenium argentatum. Guayule stores natural rubber latex in its parenchyma cells, and therefore the shrub must be homogenized in an aqueous buffer from which the latex is then purified and stabilized to create a colloidal suspension of both rubber particles and non- rubber constituents (Cornish et al., 2008). Non-rubber constituents in GNRL include rubber particle associated proteins and resins (low-molecular weight, acetone-extractable material) which include saponifiable lipids (Cornish et al., 2008).

Fillers are inert materials that are dispersed into the bulk of a material prior to manufacture and curing. Fillers can improve properties of a material, providing a

146 reinforcing effect that is characterized by improved stiffness, higher resistance to tearing and abrasion, and enhanced tensile strength (Donnet, 2003; Kohls and Beaucage, 2002).

Other fillers are diluents, and typically are used to reduce cost of the bulk product, typically without any improvements in polymer performance (Rothon, 2000). Particle size is the most fundamental property of a filler, which affects reinforcement of the elastomer the most (Rothon, 2000; Dick, 2009). Particle sizes ranging from 1000-5000 nm provide a small reinforcement; particles less than 1000 nm provide a medium reinforcement; particles smaller than 100 nm provide the strongest reinforcement

(Rothon, 2000; Dick, 2009). Other properties can affect the interactions between a polymer and filler, including surface chemistries, and particle geometry (Rothon, 2000;

Dick, 2009).

The most common reinforcement filler for natural rubber is carbon black

(Frohlich et al., 2005). Carbon black is produced by the thermal decomposition or partial combustion of petroleum or natural gas, and typically contains more than 95% pure carbon with minimal impurities of oxygen, hydrogen and nitrogen. Carbon black is manufactured to create controlled particle sizes that range from 10 nm to 500 nm, which fuse into chain-like aggregates (Rothon, 2000; Dick, 2009). Carbon fly ash is a coal combustion product that is composed of fine particulates of burned fuel that are driven out of coal-fired boilers with flue gases. In the past, carbon fly ash was released in to the atmosphere, but air pollution standards now require its captured, resulting in a large global supply. Other researchers have investigated the possibility of replacing carbon black with the cheaper carbon fly ash in elastomers such as styrene-butadiene rubber, and dry natural rubber (Barrera and Cornish, 2015).

147 The most common diluent filler for natural rubber is the calcite form of calcium carbonate. In addition to being a diluent, calcium carbonate is a detacktifying agent, can increase stiffness, and provides abrasion resistance (Dick, 2009). Calcium carbonate is typically dry milled or ground to crystallites as small as 1 micron, whereas nano- precipitated calcium carbonate creates crystallites with sizes ranging from 1- 100 nm.

(Dick, 2009). Ground calcium carbonate is derived from chalk, sediments of crushed marine shells (Dick, 2009). Alternative natural sources of calcium carbonate, such as eggshells, have shown promise as reinforcing or diluent for natural rubber when used to partially or completely replace carbon black (Barrera and Cornish, 2016) or silica (Ren et al., 2019).

The production of GNRL results in a significant amount of waste fiber, or guayule bark bagasse (Cornish et al., 2008). Guayule bark bagasse has successfully used in applications such as termite resistant wood products, chemical derivatives, paper products, and energy sources (Cornish et al., 2008). Guayule bagasse has been shown to have higher chemical extracts, especially the amount of resins, compared to other sources of wood fibers such as maple and milkweed (Cornish et al., 2008). Dry natural rubber composites have been made with guayule bagasse previously (Barrera and Cornish,

2015).

The utilization of alternative fillers such as carbon fly ash, calcium carbonate-rich eggshells, and guayule bagasse could reduce the final cost of a manufactured rubber product. However, mechanical properties may decline as a result of insufficient filler- rubber interactions, compared to crosslinked GNRL films without any fillers, especially at high loadings. The aim of this study was to analyze GNRL composite films made with

148 the alternative fillers carbon fly ash, calcium carbonate from eggshells, or guayule bagasse. The effect of particle size and loading level were evaluated.

149 7.2. Experimental

7.2.1. Materials and sample preparation

GNRL was extracted as described (Cornish et al., 2008), from fresh guayule shrub grown in Arizona. Compounding chemicals, specifically the accelerators zinc diethyldithiocarbamate (ZDEC), diphenyl guanidine (DPG), and dipentamethylenethiuram tetrasulfide (DPTT), activator zinc oxide, sulfur, and antioxidant were purchased from Akron Dispersions (Akron, OH). The ammonium hydroxide was purchased from W.W. Grainger, Inc (Salt Lake City, UT, USA).

The waste filler raw materials were generously donated as follows: eggshells (ES) by Troyer’s Home Pantry (Apple Creek, OH), carbon fly ash (CFA) by Cargill Salt

(Akron, OH), and guayule bagasse (GB) was generated as a co-product of latex extraction at our facility from guayule shrubs donated by PanAridus (Casa Grande, AZ).

7.2.1.1. Emulsion chemistry/compounding

Fillers were ground using an IKA A11 basic mill (Wilmington, NC). Macro sized particles were separated using a size 50 and 400 mesh sieve from Fisher Scientific

(Pittsburgh, PA), with resulting particles ranging from 300 µm to 38 µm. Micro sized particles were separated using a size 400 mesh, isolating particles 38 µm and smaller.

Nano sized fillers were made by dispersing the micro sized particles in distilled water, and then wet milling to submicron size using a Planetary Ball Mill 100 manufactured by

Glen Mills (Clifton, NJ). The carbon fly ash and guayule bark bagasse were dry milled as received, whereas the eggshells were washed and membranes were peeled from the eggshells prior to milling.

150 GNRL was made using a compounding recipe (Table 7.1) while varying the amount, type, and particle size of filler. The amount of fillers and other materials were added to GNRL at specified concentrations based on parts per hundred dry rubber (phr).

Fillers were dispersed in deionized water (amount of water needed to achieve 48% solids by volume was used to disperse fillers), and stirred for 1 hour with a 30 rpm hand mixer.

The latex compound recipe without fillers was prevulcanized for 2.5 hours while stirring with a 30 rpm hand mixer. The filler-water dispersion was then added to the prevulcanized latex compound, and stirred for an additional 1 hour with a 30 rpm hand mixer.

Table 7.1. Latex Compound Recipe.

Dry Weight

Guayule NRL 100

Sulfur 2

Ammonium Hydroxide 1

ZnO 1

Antioxidant 2

Zinc Diethyldiothiocarbamate (ZDEC) 0.5

Diphenyl guandineAccelerator (DPG) Accelerator 0.4

Dipentamethylenethiuram (DPTT) 0.6

AcceleratorFiller 1, 2, 5 or 10

151 7.2.1.2. Thin film manufacture by dipping

105 ºC). The rubber article was then placed in a tumble dryer post-vulcanization (60 min at 60 ºC). All GNRL thin films in this study were made using a Diplomat automated dipper (DipTech Systems Inc., Kent, OH, USA).

7.2.2. Tensile Properties

Four dumbbell samples of each compound at each dwell time were cut using Die

C according to ASTM D 412 (ASTM International, 2013b). Film thickness varied from

0.23 mm to 0.24 mm for all composites. Evaluation of the tensile mechanical properties followed ASTM D 412 and was determined using an Instron 3366 with Bluehill v. 2.17 software package (Instron, Norwood, MA, USA) (ASTM International, 2013b). Samples were tested using a crosshead speed of 500 mm/min at room temperature (26°C). The reported mean ± the standard error from the mean (SE) values are averages of at least 4 samples. In addition, statistical analysis was performed using Minitab, version 16.0

152 (State College, PA). Significant differences (P-values < 0.05) in tensile data amongst

GNRL compounds for each film thickness were determined using one-way analysis of variance (ANOVA) and the Fisher method.

7.2.3. Scanning Electron Microscopy (SEM)

The morphology of the various fillers was investigated using a Hitachi S-3500N scanning electron microscope (Tokyo, Japan). Samples were sputter-coated with a thin layer of platinum (0.2 KÅ) by an Anatech Hummer 6.2 Sputtering system prior to analysis.

153 7.3. Results and Discussion

7.3.1. Tensile properties

Eggshell-GNRL composites showed a range of varied tensile mechanical properties, compared to unfilled GNRL thin films. Eggshell-GNRL composites showed a significant stiffening in some composites, namely macrosized eggshell at 1 phr and microsized eggshell at 2 phr, compared to unfilled GNRL thin films. However other eggshell-GNRL composites had a significant softening effect compared to unfilled

GNRL films; this included all nanosized eggshell composites, as well as microsized and macrosized eggshell composites at high loadings, 5 and 10 phr (Fig. 7.1a). Elongation at break was significantly increased in all nanosized eggshell composites, and in higher loadings (5 and 10 phr) of microsized eggshell-GNRL composites (Fig. 7.1b).

Macrosized eggshell-GNRL composites, in all loadings, did not significantly change elongation at break, compared to unfilled GNRL thin films (Fig. 7.1b). There is no reinforcement effect in any of the eggshell-GNRL composites; tensile strength is not improved significantly compared unfilled GNRL thin films (Fig. 7.1c). Eggshells at 2 phr loading, in any size, provide a diluent effect; there is no significant change tensile strength compared to unfilled GNRL thin films (Fig. 7.1c).

154

Fig. 7.1. Mechanical Properties of GNRL with eggshells. (a) Modulus at 500%

Elongation (MPa). (b) Elongation at Break (%). (c) Tensile Strength (MPa). Control has no filler. n=6. * statistically significant (P>0.95) from control

155 Guayule bagasse (GB)-GNRL composite films all have decreased tensile strength compared to unfilled GNRL films (Fig. 7.2c). GB-NRL films are softer in composites containing high loadings of macro fillers and nanofillers; whereas GB-NRL films are stiffer in composites with low loadings of macro fillers (Fig. 7.2a). GB-NRL films that are softer (Fig. 7.2a), also have a higher elongation at break (Fig. 7.2b); this can be attributed to the fillers interfering in crosslinking and resulting in a softer, more ductile composite film. The guayule bagasse fillers were not able to provide a reinforcement effect, but GB-NRL composites such as 1 phr micro are able to produce a diluent effect for modulus at 500% and elongation at break, however are not able to maintain tensile strength properties.

156

Fig. 7.2. Mechanical Properties of GNRL with guayule bagasse. (a) Modulus at 500%

Elongation (MPa). (b) Elongation at Break (%). (c) Tensile Strength (MPa). Control has no filler. n=6. * statistically significant (P>0.95) from control

Low loadings of carbon fly ash (CFA) provides a diluent effect for GNRL films at low loadings, such as 1 and 2 phr (Fig. 7.3c). However at high micro and macro filler loadings, CFA-GNRL composite films, are softer (Fig. 7.3a), have higher elongations at break (Fig. 7.3b), and diminished tensile strength (Fig. 7.3c). CFA-NRL composites with nano fillers are stiffer (Fig. 7.3a), with higher elongations at break (Fig. 7.3b), without compromising tensile strength (Fig. 7.3c). CFA-NRL composites with 1 phr micro filler

157 provides no change in tensile properties compared to unfilled films (Fig. 7.3 a, b, c), providing a diluent effect.

Fig. 7.3. Mechanical Properties of GNRL with carbon fly ash. (a) Modulus at 500%

Elongation (MPa). (b) Elongation at Break (%). (c) Tensile Strength (MPa). Control has no filler. n=6. * statistically significant (P>0.95) from control

Composites made with carbon fly ash, guayule bagasse, and eggshells did not provide a reinforcing effect. However, a diluent effect did occur in composites with low loadings (1 and 2 phr) of fillers, specifically for 1 phr micro GB-NRL (Fig. 7.1), 1 phr micro CFA-NRL (Fig. 7.2), and 2 phr macro eggshell-NRL composite films (Fig. 7.3).

For smaller sized particles, the increased surface area and high degree of irregularity

158 restricts the polymer chain motion under applied strain. This behavior is not observed in macro sized fillers due to the reduced surface area and more regular, laminar shape. In 5 and 10 phr filler loadings properties were decreased; the NRL acted as a binder material for the fillers. Nano sized composites with high loadings (5 and 10 phr) could not be successfully made; the emulsions were unstable and coagulated during mixing due to insufficient wetting of the fillers.

Composites made with micro waste-derived fillers performed better than composites made with macro particles (Figs. 7.1-7.3). This is attributed to the larger particle size of the macro fillers; macro fillers have less surface area per unit weight than smaller particles, having less interaction with the polymer, and also generating flaws within the composite. Micro sized particles have a lower bulk density than macro sized particles, most likely attributed to the broader particle size distribution of the macro sized particles. The broader particle distribution allows smaller particles to fit between the void spaces of bigger particles.

Composites that provided a diluent effect contained micro sized fillers.

Interaction between these waste-derived fillers and interactions between the particles and rubber, help maintain physical properties. Hydrophobicity of the waste-derived fillers contribute to the filler’s affinity for the rubber.

Filler-filler interactions existing between particles with polar surfaces are attributed to hydrogen bonding; whereas non-polar fillers have filler-filler interactions due to van der Waals forces. Filler-filler interactions can promote the formation of a percolation network, that can contribute to the overall reinforcement of the material by

159 restricting chain mobility. The linear chain structure of GNRL allows it to flow more easily into a filler network than the more branched HNRL.

Surface chemistry of fillers can effect the efficiency of crosslinking. CFA is composed of alumino-silicate (>50%); silanol groups can react with rubber compounding ingredients and lead to slower cures. Eggshells are over 95% calcium carbonate, which can expedite cure rates due to its alkalinity. Since processing conditions were the same for all films, slower cure rates may contribute to a lower crosslink density, which results in lower mechanical properties.

160 7.4. Conclusions

Agro-industrial residues such as eggshells, carbon fly ash, and guayule bagasse are abundant solid wastes with heterogeneous compositions. These characteristics make them promising materials for NRL composites, comparable in properties to unfilled NRL films. In addition, the renewable character of agro-industrial residues can improve the sustainability of NRL products while adding value to these waste-derived fillers.

Therefore, the results encourage the further study of utilizing renewable waste-derived materials to reduce the amount of non-renewable fillers in NRL composites. Future work includes tailoring mechanical properties of NRL composites by modification of fillers such as surface functionalization with silane coupling agents, and blending particles of various origins, and sizes to achieve specific requirements for various end applications.

161 Chapter 8: Mechanical properties of Hevea natural rubber latex thin film

composites with biobased fillers.

J. Lauren Slutzky a*, and Katrina Cornish a,b aOhio State University, Department of Food, Agricultural and Biological Engineering,

1680 Madison Avenue, Wooster, Ohio 44691 USA bOhio State University, Department of Horticulture and Crop Science, 1680 Madison

Avenue, Wooster, Ohio 44691 USA

*Corresponding author: Katrina Cornish, Williams Hall, 1680 Madison Avenue,

Wooster, Ohio 44691, USA. Tel: 330-263-3982 Fax: 330-263-3887 Email:

[email protected]

Abstract:

Natural rubber latex (NRL) products are often compounded with non-sustainable reinforcing fillers such as carbon black. This work focuses on the manufacturing of NRL composites using natural rubber latex from the plant species Hevea brasiliensis with and without soluble proteins, compounded with sustainable fillers from waste streams: calcium carbonate-rich eggshells, carbon fly ash, and guayule bark bagasse. The soluble protein in NRL contains severe allergens, but serves as a naturally occurring surfactant that may improve filler dispersion in NRL. To make the NRL composite thin films, fillers were compounded into NRL, followed by coagulant dipping, and vulcanization.

The effect of filler type, size, and loading in ultra-low protein and standard protein NRL composites on mechanical properties were characterized, according to ASTM D 412.

Ultra-low protein NRL composites were more susceptible to softening compared than standard protein NRL composites. Nano sized fillers increase elongation at break of

162 ultra-low protein NRL composites across all filler types; whereas nano fillers in standard

NRL composites have no effect on elongation at break compared to unfilled standard

NRL films. Overall, none of the fillers in any loading had a reinforcement effect. Many composites did not have significantly different properties compared to unfilled NRL thin films, indicating a diluent effect.

Keywords: Hevea, Natural rubber latex composites, biofillers, calcium carbonate, carbon fly ash, guayule bagasse, elastomer thin films

8.1. Introduction

Natural rubber latex (NRL) is a colloidal suspension of rubber and non-rubber constituents, derived from the plant species Hevea brasiliensis. Standard NRL typically contains 60% dry rubber, with an additional 1.6% dry non-rubber constituents that includes proteins and phospholipids (Sansatsadeekul et al., 2011). Ultra-low protein

Hevea NRL undergoes an aluminum hydroxide treatment that removes soluble protein and other impurities, resulting in 0.5% dry non-rubber constituents (Swason, 2008). The effect of non-rubber constituents in NRL thin films have been shown to have profound effects on mechanical properties, attributed to protein-polymer interactions (Monadjemi et al., 2016). However, the effect of soluble protein in NRL composites with biofillers has not been previously investigated.

Fillers are expected to adhere to rubber particles during thin film dipping and vulcanization to provide a reinforcing effect. However, many fillers are not able to adhere to rubber particles without a coupling agent, which can add cost to the final product. Fillers that are not effectively adhered to rubber particles provide a diluent effect; diluent fillers are inert and added to polymer materials at levels that don’t cause

163 premature material failure, to reduce the cost of the final product. Diluent fillers such as calcium carbonate, sodium fluorohectorite, and kaolinite clay are common bulk cheapeners used in NRL products (Dick, 2009) but frequently reduce product performance. However, calcium carbonate and kaolinite clay requires extensive mining, which is dangerous and can ultimately cause detriment to ecological systems (Dick,

2009). Sodium fluorohectorite is synthesized at extremely high temperatures, using glass precursors, which is energy intensive and a dangerous synthesis (Stoter et al., 2013).

Therefore diluent fillers derived from waste streams are of interest; fillers derived from waste streams require little energy for preparation, and are essentially negligible in cost.

Waste streams that are applicable as diluent fillers for NRL composites include calcium carbonate-rich eggshells, carbon fly ash, and guayule bark bagasse. Eggshells, carbon fly ash, and guayule bark bagasse have successfully been incorporated into natural rubber composites as diluent fillers, and in NRL and GNRL composites as partial replacements of carbon black (Barrera and Cornish, 2015; Barrera and Cornish, 2016) or silica (Ren et al., 2019) when they can also act as reinforcing fillers and also improve the dispersion of the conventional filler. This study analyzed the effect of NRL soluble protein and filler type and loading, on mechanical properties of NRL composites made with the alternative fillers carbon fly ash, calcium carbonate-rich eggshells, or guayule bagasse. The fillers varied in particle sizes, loading, and type of latex (with or without soluble protein).

164 8.2. Experimental

8.2.1. Materials

Ultra-low protein NRL (trade name Vytex) and standard protein NRL (trade name

Centex) were purchased from Centrotrade (Chesapeake, VA). Compounding chemicals, specifically the accelerators zinc diethyldithiocarbamate (ZDEC), diphenyl guanidine

(DPG), and dipentamethylenethiuram tetrasulfide (DPTT), activator zinc oxide, sulfur, and antioxidant were purchased from Akron Dispersions (Akron, OH). The ammonium hydroxide was purchased from W.W. Grainger, Inc (Salt Lake City, UT, USA).

The waste filler raw materials were generously donated as follows: eggshells (ES) by Troyer’s Home Pantry (Apple Creek, OH), carbon fly ash (CFA) by Cargill Salt

(Akron, OH), and guayule bagasse (GB) was generated as a co-product of latex extraction at our facility from guayule shrubs donated by PanAridus (Casa Grande, AZ).

8.2.2. Preparation of Fillers

Raw materials were dried at 55 oC and then ground using an IKA A11 basic mill

(Wilmington, NC). Macro sized particles were separated using a size 50 and 400 mesh sieve from Fisher Scientific (Pittsburgh, PA), with resulting particles ranging from 300

µm to 38 µm. Micro sized particles were separated using a size 400 mesh, isolating particles 38 µm and smaller. Nano sized fillers were made by dispersing the micro sized particles in distilled water, and then wet milling to submicron size using a Planetary Ball

Mill 100 manufactured by Glen Mills (Clifton, NJ). The carbon fly ash and guayule bark bagasse were dry milled as received, whereas the eggshells were washed and membranes were peeled from the eggshells prior to milling. Particle size distributions were verified using a particle size analyzer LA-950V2, Horiba Scientific (Irvine, CA) (Fig. 8.1). Filler

165 bulk densities were determined by weighing a known volume of each filler and calculating the bulk density as mass/volume occupied (Table 8.1).

Fig. 8. 1. Macro (solid line) and micro sized (dashed line) fillers’ particle size distribution. a) Carbon fly ash; macro (median 76.20 μm, mean 89.32 μm, SD 61.94 μm), micro (median 11.25 μm, mean 12.12 μm, SD 4.93 μm), b) eggshells; macro )median 230.33 μm, mean 241.46 μm, SD 111.01 μm), micro (median 16.07 μm, mean 23.19 μm, SD 20.79 μm), c) guayule bagasse; macro (median 243.10 μm, mean 279.33 μm, SD 188.56 μm), micro (median 49.09 μm, 54.75 μm, SD 30.63 μm).

166 Table 8.1. Filler bulk densities Filler Size Density (g/cm3) Carbon fly ash Macro 0.74 ± 0.02 Carbon fly ash Micro 0.57 ± 0.01 Guayule bagasse Macro 0.53 ± 0.02 Guayule bagasse Micro 0.36 ± 0.07 Eggshells Macro 1.32 ± 0.02 Eggshells Micro 0.79 ± 0.05

8.2.3. Emulsion chemistry/compounding

NRL composites were made using a compounding recipe (Table 8.1) while varying the amount, type, and particle size of filler. The amount of fillers and other materials were added to NRL at specified concentrations based on parts per hundred dry rubber (phr). Fillers were dispersed in deionized water (amount of water needed to achieve 48% solids by volume was used to disperse fillers), and stirred for 1 hour with a

30-rpm hand mixer. The latex compound recipe without fillers was prevulcanized for 2.5 hours while stirring with a 30-rpm hand mixer. The filler-water dispersion was then added to the prevulcanized latex compound, and stirred for an additional 1 hour with a

30-rpm hand mixer.

167 Table 8.2. Latex Compound Recipe.

Dry Weight

NRL 100

Sulfur 2

Ammonium Hydroxide 1

ZnO 1

Antioxidant 2

Zinc Diethyldiothiocarbamate (ZDEC) 0.5

Diphenyl guandineAccelerator (DPG) Accelerator 0.4

Dipentamethylenethiuram (DPTT) 0.6

AcceleratorFiller 1, 2, 5 or 10

168 8.2.4. Thin film manufacture by dipping

105 ºC). The rubber article was then placed in a tumble dryer post-vulcanization (60 min at 60 ºC). All NRL thin films in this study were made using a Diplomat automated dipper

(DipTech Systems Inc., Kent, OH, USA).

8.2.5. Tensile Properties

Four dumbbell samples of each compound at each dwell time were cut using Die

C according to ASTM D 412 (ASTM International, 2013a). Film thickness varied from

0.24 mm to 0.23 mm for all composites. Evaluation of the tensile mechanical properties followed ASTM D 412 and was determined using an Instron 3366 with Bluehill v. 2.17 software package (Instron, Norwood, MA, USA) (ASTM International, 2013a). Samples were tested using a crosshead speed of 500 mm/min at room temperature (26 °C). The reported mean ± the standard error from the mean (SE) values are averages of at least 4 samples. In addition, statistical analysis was performed using Minitab, version 16.0

(State College, PA). Significant differences (P-values < 0.05) in tensile data amongst

169 NRL compounds for each film thickness were determined using one-way analysis of variance (ANOVA) and the Fisher method.

8.2.6. Scanning Electron Microscopy (SEM)

170 8.3. Results and Discussion

8.3.1. Filler characterization

Variations in particle size distribution were observed among the different fillers

(Fig. 8.1). Nano sized eggshell fillers had the narrowest particle distribution, whereas the larger micro and macro particle sizes were broader and more likely to agglomerate. All waste-derived fillers have elementary particles with nano sized dimensions (Fig. 8.2).

Particle morphology was also observed using SEM and TEM (Fig. 8.3 and 8.4). CFA and ES had irregular particle contours, whereas GB has a rod-like particle shape.

Fig. 8. 2: Particle size distribution of nano sized fillers. a) Guayule bagasse (median:

1.370 μm, mean: 1.585 μm), b) carbon fly ash (median: 0.218 μm, mean: 1.109 μm), c) eggshells (median: 0.125 μm, mean: 0.757 μm).

171

Fig. 8. 3: Transmission electron micrographs of nano sized fillers. a) guayule bagasse; b) carbon fly ash; c) eggshells.

8.3.2. Tensile properties of NRL composites

Tensile properties vary among eggshell NRL composites due to loading, and type of lattices (with and without soluble protein). Ultimately, no reinforcement effect was found for NRL composites with eggshell fillers at any loading, at any particle size.

Eggshell-standard protein NRL composites was statistically significantly softer than unfilled standard protein NRL in all particle sizes and filler loadings (Fig. 8.4b); but statistically significant softening occured only in higher eggshell filler loadings for ultra-

172 low protein NRL composites (Fig. 8.4a). Eggshell-ultra-low protein NRL composites showed a significant increase in elongation at break with nanosized 1 phr and 2 phr loadings, as well as a microsized 5 phr loading, compared to an unfilled ultra-low protein

NRL film (Fig. 8.4c). However, in eggshell-standard protein NRL composites, elongation at break is improved only with nanosized 2 phr loading (Fig. 8.4d). Eggshell fillers at any size or any loading do not improve the tensile strength of NRL composite films, regardless of presence or absence of soluble latex protein (Fig. 8.4f).

173

Fig. 8. 4. Tensile Properties of Eggshell-NRL composites. Vytex is ultra-low soluble protein NRL; Centex is standard protein NRL.

174 Tensile properties of carbon fly ash NRL composites varied with filler loading and amount of protein in the latex. Compared to unfilled NRL thin films, carbon fly ash- ultra-low protein NRL composites significantly softened, specifically with nanosized fillers at 1 and 2 phr loadings and in macro-sized fillers at 2, 5 and 10 phr loadings (Fig.

8.5a). Microsized carbon fly ash fillers in ultra-low protein NRL also softened, but notably only at higher loadings of 5 and 10 phr. However, carbon fly ash composites with standard protein NRL softened less than the protein films; only carbon fly ash composites with high loadings (5 and 10 phr) of macrosized fillers caused a significant softening effect in thin films (Fig. 8.5b). The amount of protein in the NRL also affected elongation at break of carbon fly ash composites, which significantly increased in 1 and 2 phr nanosizded fillers (Fig. 8.5c). Elongation at break was unaffected by carbon fly ash fillers of any size or any loading in standard NRL-carbon fly ash composites (Fig. 8.5d).

There was no reinforcing effect for carbon fly ash compared to unfilled NRL thin films, regardless of the amount of soluble protein in NRL (Fig. 8.5e, Fig. 8.5f). At higher loadings such as 5 and 10 phr, there is a significant decrease in properties in micro and macro sized fillers in films made from both ultra-low protein and standard protein NRL.

175

Fig. 8. 5. Tensile Properties of Carbon fly ash-NRL composites. Vytex is ultra-low soluble protein NRL; Centex is standard protein NRL.

176 Guayule bagasse was not capable of providing a reinforcing effect, regardless of particle size, loading, or amount of soluble protein in NRL (Fig. 8.6). Guayule bagasse composites softened in nanosized guayule bagasse-ultra-low protein NRL composites, as well as high loadings of macrosized guayule bagasse (Fig. 8.6a). However, guayule bagasse did not soften guayule-bagasse-standard soluble protein NRL composites (Fig.

8.6b). Elongation at break is significantly increased in nanosized guayule bagasse-ultra- low protein NRL composites, compared to unfilled ultra-low protein NRL composites

(Fig. 8.6c). However, elongation at break significantly decreased in 1phr microsized guayule bagasse-ultra-low protein NRL composites, compared to unfilled ultra-low protein NRL composites (Fig. 8.6c). In guayule bagasse at nano and micro sizes did not affect standard-soluble protein NRL composites (Fig. 8.6d), but macro sized guayule bagasse-standard soluble protein NRL composites had significantly lower elongation at break. Guayule bagasse fillers decreased tensile strength of composites especially in those made with standard-protein NRL, most marked with macrosized filled standard- protein NRL thin films (Fig 6.8f). Guayule bagasse did not reinforce ultra-low protein

NRL composites and 1 phr microsized guayule bagasse filled composites were significantly weaker than the ultra-low protein NRL thin films without fillers (Fig 6.8e).

177

Fig. 8. 6. Tensile Properties of guayule bagasse-NRL composites. Vytex is ultra-low soluble protein NRL; Centex is standard protein NRL.

178 Composites made with carbon fly ash, guayule bagasse, and eggshells did not provide a reinforcing effect, regardless of type of latex used. Structure of the fillers can influence reinforcing effect. The eggshells and CFA fillers have a high surface area due to its platy, rough structure and high porosity (Fig. 8.2). The porosity promotes a wetting effect, that can improve interfacial adhesion between the polymer and filler. The fillers in the NRL composite films were not surface treated; surface modification of fillers to improve surface hydrophobicity makes it more compatible with the rubber polymer.

Hydrophobic surface chemistries in fillers can improve polymer-filler interaction and contributed to the observed reinforcing effect. Naturally occurring resins in GB can affect mechanical properties of GB-NRL composites. Resins can potentially provide a plasticizing effect, increasing ductility and properties such as elongation at break and modulus at 500% elongation.

Many of the composites made in this study meet the physical properties required for their application in a range of rubber products. Potential applications include medical surgical gloves, and textured dipped coatings. The filler-NRL composites that provide a diluent effect could be used to manufacture NRL products with a low carbon footprint and are most sustainable than conventional fillers such as mined calcium carbonate, and carbon black.

179 8.4. Conclusions

Waste-derived fillers can provide a diluent effect for NRL composite films, regardless of the level of soluble protein in the NRL. Diluent composite NRL films were achieved by low loadings of micro waste-derived fillers. This is important due to the renewable character of these materials, and possible applications for NRL composites.

180 Chapter 9: Conclusion

The main goals of this research were to develop thin film applications for type I circumallergenic GNRL and type I hypoallergenic ultra-low protein Hevea NRL; this included optimizing a type IV hypoallergenic accelerator package to make medical grade elastomers, and developing latex thin film composites with agro-industrial residues for industrial grade elastomers.

The type IV hypoallergenic accelerator package was successfully used to compound thin films that meet the specifications for surgical gloves (ASTM 3577) for both type I circumallergenic GNRL, and type I hypoallergenic ultra-low protein Hevea

NRL. Instrinsic differences in the composition of GNRL and ultra-low Hevea NRL required markedly different accelerator concentrations and ratios to maximize their mechanical performance and achieve or exceed the tensile properties outlined in ASTM

3577; GNRL required 3x more accelerator than ultra-low protein Hevea NRL formulations, probably due to its even lower protein content as well as differences in stabilizing systems. Statistical modeling of GNRL formulations correlated only one type of accelerator loading and ultimate elongation; whereas ultra-low protein Hevea NRL statistical modeling correlated both accelerator types, and film thickness with all tensile mechanical properties modeled. Thus, GNRL must have other significant properties which would be better predictors of behavior but were not identified or taken in the current models for statistical modeling. Also, the buffering systems of the GNRL and ultra-low protein Hevea NRL are drastically different; GNRL was stabilized with potassium hydroxide to stabilize the emulsion (a strong base) whereas ultra-low protein

Hevea NRL is stabilized by ammonium hydroxide (a weak base). Ammonium hydroxide

181 is known to work as an activator with zinc-based catalyst systems, and therefore buffering systems of the lattices would have an effect on crosslinking and accelerator efficiency and may cause the stronger correlations in formulation-property statistical models seen in this latex.

The composites of GNRL and ultra-low protein Hevea NRL with agro-industrial wastes did not reinforcement the composite latex films, regardless of latex type, filler type, or loading. However, several sizes and loadings could be used without damanging properties until higher loadings were used. Fillers did affect crosslinking in some composites in GNRL and NRL (with and without soluble protein) significantly decreasing tensile strength, and softening the modulus at 500% elongation, compared to unfilled thin films. Soluble protein in latex had an effect on % elongation at break: nano fillers increased elongation at break of GNRL and ultra-low protein Hevea NRL composites, but did not occur in standard soluble protein NRL. Thus, latex composition, including nonrubber components, and polymer-filler interations, are important to the final performance of composite films.

182 Chapter 10: Future Studies

This research revealed the complexity and variation in types of latex and their effect on crosslinking, and ultimately tensile mechanical properties. This work provides a foundation for future research in understanding and predicting how variances in latex chemistry affect chemical compounding and ultimately, mechanical properties.

Future work should include: investigating the crosslinking efficiency of a GNRL stabilized in ammonium hydroxide for chemical accelerator packages, optimizing chemical accelerator packages for the filler composite work, and functionalizing fillers with a hydrophobic or sulfur surface group to improve the reinforcing effect in the latex composites. Additional material characterization could be done of the composites formulated in this work could include degradation studies, oxidation resistance, and analysis of dynamic properties.

183 References:

Abraham, E., Deepa, B., Pothan, L.A., Jacob, M., Thomas, S., Cvelbar, U., Anandjiwala,

R., 2011. Extraction of nanocellulose fibrils from lignocellulosic fibres: A novel

approach. Carbohydr. Polym. 86, 1468–1475. doi:10.1016/j.carbpol.2011.06.034

Abraham, E., Deepa, B., Pothan, L.A., John, M., Narine, S.S., Thomas, S., Anandjiwala,

R., 2013a. Physicomechanical properties of nanocomposites based on cellulose

nanofibre and natural rubber latex. Cellulose 20, 417–427. doi:10.1007/s10570-

012- 9830-1

Abraham, E., Thomas, M.S., John, C., Pothen, L. a., Shoseyov, O., Thomas, S., 2013b.

Green nanocomposites of natural rubber/nanocellulose: Membrane transport,

rheological and thermal degradation characterisations. Ind. Crops Prod. 51, 415–

424. doi:10.1016/j.indcrop.2013.09.022

Alenius, H., Turjanmaa, K., Palosuo, T., 2002. Natural Rubber latex allergy.

Occupational and Environmental Medicine. 59, 419-424.

Amnuaypornsri, S., Sakdapipanich, J., Toki, S., Hsiao, B.S., Ichikawa, N., Tanaka, Y.,

2008. Strain-Induced Crystallization of Natural Rubber: Effect of Proteins and

Phospholipids. Rubber Chem. Technol. 81, 753–766. doi:10.5254/1.3548230

Angellier, H., Molina-Boisseau, S., Dufresne, A., 2005a. Mechanical properties of waxy

maize starch nanocrystal reinforced natural rubber. Macromolecules 38, 9161–

9170. doi:10.1021/ma0512399

184 ASTM International, 2013a. ASTM standard D412-15a. Standard test methods for

vulcanized rubber and thermoplastic elastomers- tension. Doi:10.1520/D0412-

15A

ASTM International, 2013b. ASTM standard D3577. Standard specification for rubber

surgical gloves.

Barrera, C.S., Cornish, K., 2016. High performance waste-derived filler/carbon black

reinforced guayule natural rubber composites. Ind. Crops Prod. 86, 132–142.

doi:10.1016/j.indcrop.2016.03.021

Barrera, C.S., Cornish, K., 2015. Novel Mineral and Organic Materials from Agro-

Industrial Residues as Fillers for Natural Rubber. J. Polym. Environ. 23, 437–448.

doi:10.1007/s10924-015-0737-4

Bergman, C., Trimbach, J., 2014. Influence of Plasticizers on the Properties of Natural

Rubber Compounds. KGK Rubberpoint. 67, 40-49.

Calhoun, A., Peacock, A., 2006. Polymer Chemistry Properties and Applications. Hanser.

Carey, A.B., Cornish, K., Schrank, P.J., 1995. Cross reactivity of alternative plant sources

of latex in subjects with systemic IgE mediated sensitivity to Hevea brasiliensis

latex. Annuals of Allergy, Asthma, Immunology. 74, 317-320.

Chakraborty, K.B., Couchman, R., 2006. Technological and physico-chemical behavior

of safer accelerators in synthetic polyisoprene latex. 17, 1-12.

185 Chen, S.F., Wont, W.C., Low, C.Y., 2011. Vulcanization composition having reduced

allergic potential. Patent Application CA2802956.

Cornish, K., 2001. Similarities and differences in rubber biochemistry among plant

species. Phytochemistry 57, 1123–1134. doi:10.1016/S0031-9422(01)00097-8

Cornish, K., 2014. Biosynthesis of natural rubber (NR) in different rubber-producing

species, in: Chemistry, Manufacture and Applications of Natural Rubber.

Elsevier, pp. 3–29. doi:10.1533/9780857096913.1.3

Cornish, K., 1996. Hypoallergenic natural rubber products from Parthenium argentatum

(Gray) and other non-Hevea brasiliensis species. U.S. Patent, No. 5,580,942.

Cornish, K., Backhaus, R.A., 1990. Rubber transferase activity in rubber particles of

guayule. Phytochemistry. 29, 3809-3813. Doi: 10.1016/0031-9422(90)85337-F.

Cornish, K., Chapman, M.H., Nakayama, F.S., Vinyard, S.H., Whitehand, L.C., 1999.

Latex quantification in guayule shrub and homogenate. 10, 121-136. Doi:

10.1016/S0926-6690(99)00016-3

Cornish, K., Fishman, B., Nguyen, K., Williams, J., 2007. Guayule natural rubber thin

film articles. Patent Application WO2009078883 A1, USA.

Cornish, K., McMahan, C.M., Xie, W., 2006. ASTM D1076 Category 4 Latex

Quantifying Guayule (NRG) and Hevea (NR) Latex Protein. In: International

Latex Conference, 15-16.

186 Cornish, K., Whitehand, L.C., Van Fleet, J.E., Brichta, J.L., Chapman, M.H., Knuckles,

B.E., 2005. Latex yield and quality during storage of guayule (Parthenium

argentatum Gray) homogenates. Ind. Crops Prod. 22, 75–85.

doi:10.1016/j.indcrop.2004.07.004

Cornish, K., Williams, J., Hall, J.L., McCoy, R.G., 2008. Production and properties of

Yulex® - The Natural Solution to Latex Allergy. Rubber Chem. Technol. 81,

709– 722. doi:10.5254/1.3548227

Cornish, K., Williams, J., Nguyen, K., Fishman, B., 2007. Guayule natural rubber latex

thin film articles. U.S. Patent, No 84316676B2.

Dick, J.S., 2009. Rubber Technology: Compounding and Testing for Performance 2E.

Hanser.

Donnet, J.B., 2003. Nano and microcomposites of polymers elastomers and their

reinforcement. Compos. Sci. Technol. 63, 1085–1088. doi:10.1016/S0266-

3538(03)00028-9

Donnet, J.-B., Custodero, E., 2013. Reinforcement of Elastomers by Particulate Fillers,

in: The Science and Technology of Rubber. Elsevier, pp. 383–416.

doi:10.1016/B978-0-12-394584-6.00008-X

Gao, T., Huang, M., Puwang, L., 2013. Effect of accelerator type on dynamic properties

of natural rubber vulcanizates. Advanced Materials Research. 834, 138-141.

187 Gidrol, X., Chrestin, H., Tan, H.L., Kush, A., 1994. Hevein, a lectin-like protein from

Hevea brasiliensis is involved in the coagulation of latex. 25, 9278-83.

Grellmann, W., Seidler, S., 2001. Deformation and Fracture Behaviour of Polymers.

Springer.

Grzybowski, S., Peyser, P.A., Johnson, C.C., 1996. The prevalence of anti-latex IgE

antibodies among registered nurses. Jour. Allergy and Clinical Immuno. 98, 535-

544.

Hager, T., MacArthur, A., McIntyre, D., Seeger, R., 1979. Chemistry and Structure of

Natural Rubbers. Rubber Chemistry and Technology. 52, 693-709.

Doi:10.5254/1.3535234

Hamann, C.P., Rogers, P.A., Sullivan, K., 2001. Latex allergy update. Infection Control

Today. 5, 36-38.

Hamann, C.P., Turjanmaa, K., Rietschel, R., Siew, C., Owensby, D., Gruninger, S.E.,

1998. Natural rubber latex hypersensitivity: incidence and prevalence of type I

allergy in the dental professional. The Journal of the American Dental

Association. 129, 219-220.

Hamed, G.R., 2000. Reinforcement of Rubber. Rubber Chem. Technol. 73, 524-533.

Doi: 10.5254/1.3547603

188 Hamilton, R.G., Cornish, K., 2010. Immunogenicity studies of guayule and guayule latex

in occupationally exposed workers. Ind. Crops Prod. 31, 197–201.

doi:10.1016/j.indcrop.2009.09.012

Hayes, B.B., Afshari, A., Millecchia, L., Willard, P.A., Povoski, S.P., Meade, B.J., 2000.

Evaluation of percutaneous penetration of natural rubber latex proteins.

Toxological Sciences. 56, 262-270. Doi: 10.1093/toxsci/56.2.262

Kauser, A., 2017. Contemporary applications of carbon black-filled polymer composites

overview of essential aspects. Journ. Plastic Film & Sheeting. Doi:

10.1177/8756087917725773

Kelly, K., 1995. Management of the Latex-Allergic Patient. Immunol Allergy Clinical

Am. 15, 139-156.

Kohls, D.J., Beaucage, G., 2002. Rational design of reinforced rubber. Curr. Opin. Solid

State Mater. Sci. 6, 183-194. Doi: 10.1016/S1359-0286(02)00073-6

Krutzer, B., Ros, M., Smit, J., de Jong, W., 2014, A review of synthetic lattices in

surgical glove use. Report, Kraton Innovation Center.

Kruzelak, J., Sykora, R., Hudec, I., 2016. Sulphur and peroxide vulcanization of rubber

compounds- overview. Chemical Papers. 70, 1533-1555. Doi: 10.1515/chempap-

2016-0093.

189 Kumar, S., Hahn, F.M., Baidoo, E., Kahlon, T.S., Wood, D.F., McMahan, C.M., Cornish,

K., Keasling, J.D., Daniell, H., Whalen, M., 2012. Remodeling the isoprenoid

pathway in tobacco by expressing cytoplasmic mevalonate pathway in

chloroplasts. Metabolic Engineering. 14, 19-28. Doi:

10.1016/j/ymben.2011.11.005

Leblanc, J., 2002. Rubber-filler interactions and rheological properties in filled

compounds. Prog. Polym. Sci. 27, 627-687. Doi: 10.5254/1.3538716

Liss, G.M., Sussman, G.L., 1999. Latex sensitization: occupational versus general population prevalence rates. American Journal of Industrial Medicine. 35, 196-200.

Lovell, P.A., El-Aasser, M.S., 1997. Emulsion Polymerization and Emulsion Polymers.

John Wiley & Sons Ltd

Mahapram, S., Poompradub, S., 2011. Preparation of natural rubber (NR) latex/low

density polyethylene (LDPE) blown film and its properties. Polymer Testing. 30:

716-725. Doi: 10.1016/j.polymertesting.2011.06.006

Mark, J.E., Erman, B., Roland, C.M., 2013. The Science and Technology of Rubber 4th

edition. Doi: 10.1016/C2011-0-05820-9

McMahan, C., Kostyal, D., Lhamo, D., Cornish, K., 2015. Protein influences on guayule

and Hevea natural rubber sol and gel. J. Appl. Polym. Sci. 42051, n/a-n/a. 185

doi:10.1002/app.42051

190 McMahan, C., Lhamo, D., 2015. Study of Amino Acid Modifiers in Guayule Natural

Rubber. Rubber Chem. Technol. 88, 310–323. doi:10.5254/rct.15.85931

Miri, S., Pourpak, Z., Zarinara, A., 2007. Prevalence of type I allergy to natural rubber

latex and type IV allergy to latex and rubber additives in operating room staff

with glove-related symptoms. Allergy and Asthma Proceedings. 28, 557-563.

Meyer, J.D., Chen, Y., Holt, D.L., 2000. Occupational contact dermatitis in the UK: a

surveillance report from Epiderm & OPRA. Occupational Medicine. 50, 265-273.

Monadjemi, S.M., McMahan, C.M., Cornish, K., 2016. Effect of Non-Rubber

Constituents on Guayule and Hevea Rubber Intrinsic Properties. Journal of

Research Updates in Polymer Science. 5, 87-96. doi:10.6000/1929-

5995.2016.05.03.01

Mooibroek, H., Cornish, K., 2000. Alternative sources of natural rubber. Appl.

Microbiol. Biotechnol. 53, 355–365. doi:10.1007/s002530051627

Musto, S., Barbera, V., Maggio, M., Mauro, M., Guerra, G., Galimberti, M., 2016.

Crystallinity and crystalline phase orientation of poly(1,4-cis-isoprene) from

Hevea brasiliensis and Taraxacum kok-saghyz. Polymer Advanced Technologies.

7, 1082-1089. Doi: 10.1002/pat.3774

Mutsuga, M., 2013. Analysis of N-Nitrosamine Migration from Rubber Soothers. Journal

of Analytical Chemistry. 4, 277-285.

191 Nabavizadeh, S.H., Anushiravani, A., Amin, R., 2009. Natural rubber latex

hypersensitivity with skin prick test in operating room personnel. Iranian Journal

of Allergy, Asthma and Immunology 8, 219-220.

Nettis, E., Assennato, G., Ferrannini, A., Tursi, A., 2002. Type I allergy to natural rubber

latex and type IV allergy to healthcare workers with glove-related skin symptoms.

Clin. Exp All 32, 441-447.

Nocil Ltd., 2010. Vulcanization and Accelerators. Report, Technical Note.

Pak, V.M., Watkins, M., Green-McKenzie, J., 2012. What is the role of thiurams in

allergy to natural rubber latex products? Journal of Occupational and

Environmental Medicine. 54, 649-650.

Pan, Z., Durst, F., Werck-Reichhart, 1995. The major protein of guayule rubber particles

is a cytochrome P450: enzyme and its reaction products. Journal of Biological

Chemistry. 270, 8487-8494.

Perumal, V., Geetha, N., Palanivel, S., Thulaseedharan, A., 2013. Natural rubber

producing plants: An overview. African Journal of Biotechnology. 12,1297-1310.

Doi: 10.5897/AJBX12.016

Petrovic, Z.S., Ionescu, M., Milic, J., Halladay, J.R., 2013. Soybean Oil Plasticizers as

Repleacement of Petroleum in Rubber. Rubber Chem. and Tech. 86, 233-249.

192 Puskas, J.E., Chiang, K., Barkakaty, B., 2014. Natural rubber (NR) biosynthesis:

perspectives from polymer chemistry, in: Chemistry, Manufacture and

Applications of Natural Rubber. Elsevier, pp. 30–67.

doi:10.1533/9780857096913.1.30

Quirk, R.P., 1988. Overview of curing and crosslinking elastomers. Prog. Rubber Plastic

Technology 4, 31.

Rattanasom, N., Saowapark, T., Deeprasertkul, C., 2007. Reinforcement of natural rubber

with silica/carbon black hybrid filler. Polym. Test. 26, 369–377.

doi:10.1016/j.polymertesting.2006.12.003

Ren, X., Geng, Y., Soboyejo, A.B.O., Cornish, K., 2019. Reinforced mechanical

properties of functionalized silica and eggshell filled guayule natural rubber

composites. Rubber Chemistry and Technology. Doi:10.5254/rct.19.81485

Rickman, J.M., Wang, Y., Rollett, A.D., Harmer, M.P., Compson, C., 2017. Data

analytics suing canonical correlation analysis and Monte Carlo simulation.

Computational Materials. Doi:10.1038/s41524-017-0028-9

Rothon, R.N., 2000. Particulate Fillers for Polymers. Rapra. ISBN: 1-85957-310-X

Sansatsadeekul, J., Sakdapipanich, J., Rojurthai, P., 2011. Characterization of associated

proteins and phospholipids in natural rubber latex. J Biosci Bioeng. 111, 628-634.

Doi: 10.1016/j.biosc.2011.01.013

193 Sasseville, D., 2008. Occupational contact dermatitis. Allergy, Asthma, and Clinical

Immunology: Office Journal of Canadian Society of All. and Clin. Immuno. 54,

649-650. doi:10.1097/JOM.0b013e31824be452.

Schmidt, T., Lenders, M., Hillebrand, A., van Deenen, N., Munt, O., Reichelt, R.,

Eisenreich, W., Fischer, R., Prufer, D., Gronover, C.S., 2010. Characterization of

rubber particles and rubber chain elongation in Taraxacum kok-saghyz. BMC

Biochemistry, 11:11. Doi: 10.1186/1471-2091-11-11.

Seymour, B., 1989. Polymer science before and after 1989: notable development in the

lifetime of Maurtis Dekker. J. Macromols. Sci. Chem. 26, 1023-1032.

Siler, D.J., Cornish, K., 1995. Measurement of Protein in Natural Rubber Latex. Anal.

Biochem. 229, 278–281.

Siler, D.J., Cornish, K., 1994. Hypoallergenicity of guayule rubber particle proteins

compared to Hevea latex proteins. Ind. Crops Prod. 2, 307–313.

doi:10.1016/0926- 6690(94)90122-8

Siler, D.J., Cornish, K., Hamilton, R.G., 1996. Absence of cross-reactivity of IgE

antibodies from subjects allergic to Hevea brasiliensis latex with a new source of

natural rubber latex from guayule (Parthenium argentatum). J. Allergy Clin.

Immunol. 98, 895–902. doi:Doi 10.1016/S0091-6749(96)80005-4

Soboyejo, A.B., 2011. Probabilistic methods in engineering and biosystems engineering.

194 Spaner, D., Dolovich, J., Tarlo, S., Sussman, G., Butto, K., 1989. Hypersensitivity to

natural rubber latex. J. Allergy Clin. Immunol. 83, 1135–1137. doi:

10.1016/0091-6749(89)90457-0.

Steinbüchel, A., 2003. Production of rubber-like polymers by microorganisms. Curr.

Opin. Microbiol. 6, 261–270. doi:10.1016/S1369-5274(03)00061-4

Stoter, M., Kunz, D.A., Schmidt, M., Hirsemann, D., Kalo, H., Putz, B., Senker, J., Breu,

J., 2013. Nanoplatelets of Sodium Hectorite Showing Aspect Ratios of 20000 and

superior purity. Langmuir. 29, 1280-1285.

Swason, M., 2008. Vytex Natural Rubber Latex: An Innovative Source for Natural

Rubber Products. International Latex Conference.

Tanaka, Y., 2001. Structural Characterization of Natural Polyisoprenes: Solve the

Mystery of Natural Rubber Based on Structural Study. Rubber Chem. Technol.

74, 355–375. doi:10.5254/1.3547643

Tao, J., 2002. Soft nitrile gloves having improved glove relaxation properties. Patent

Application EP 1352616 A1.

Taylor, J.S., Leow, Y.H., 2000. Cutaneaous Reactions to Rubber. Rubber Chemistry and

Technology: Rubber Reviews. 73, 427-485.

Tohsan, A., Ikeda, Y., 2014. Generating particulate silica fillers in situ to improve the

mechanical propertie4s of natural rubber (NR), in: Chemistry, Manufacture and

Applications of Natural Rubber. Elsevier, pp. 169-192.

195 van Jole, E., 2007. Latex accelerator composition. European Patent, No 1913071B1. van Beilen, J.B., Poirier, Y., 2007. Establishment of new crops for the production of

natural rubber. Trends Biotechnol. 25, 522–529.

doi:10.1016/j.tibtech.2007.08.009

Vennemann, N., Schwarze, C., Kummerlowe, C., 2013. Determination of crosslink

density and network structure of NR vulcanizates by means of TSSR. Advanced

Materials Research. 844, 482-485.

Vivaldo-Lima, E., Wood, P.E., Hamielec, A.E., 1997. An Updated Review on

Suspension Polymerization. Ind. Eng. Chem. Res. 36, 939-965. Doi:

10.1021/ie960361g

Wititsuwannakul, D., Wititsuwannakul, R., 2005. Natural Rubber and Structure of Latex,

Biochemistry of. Biopolymers Online. Doi: 10.1002/3527600035.bpol2006

Wypych, G., 2012. Handbook of Polymers. ChemTec Publishing.

Yasuyuki, T., 2001. Structural Characterization of Natural Polyisoprenes: Solve the

Mystery of Natural Rubber Study. Rubber Chemistry and Technology. 74, 355-

375.

Young, R.J., Lovell, P.A., 1991. Introduction to Polymers: Second Edition. Springer-

Science. Doi: 10.1007/978-1-4899-3176-4.

196