You are on page 1of 76

RELATIONSHIP BETWEEN THE PHYSICAL PROPERTIES AND CURING SYSTEM OF AN EPOXY MATRIX MATERIAL by

Robert Rainer Pittroff Submitted in partial fulfilment of the requirements for the degree

MAGISTER TECHNOLOGIAE: ENGINEERING: MECHANICAL in the Department of Mechanical Engineering FACULTY OF ENGINEERING AND THE BUILT ENVIRONMENT

TSHWANE UNIVERSITY OF TECHNOLOGY

Supervisor: Prof O.C. Vorster Co-Supervisor: Mr. D Louwrens December 2007

ii DECLARATION BY CANDIDATE I hereby declare that the dissertation submitted for the degree M Tech: Engineering: Mechanical, at Tshwane University of Technology, is my own original work and has not previously been submitted to any other institution of higher education. I further declare that all sources cited or quoted are indicated and acknowledged by means of a comprehensive list of references.

Robert R Pittroff

Copyright Tshwane University of Technology 2007

iii Acknowledgement I would like to express my sincere gratitude and appreciation to the following: My supervisor, Prof. O.C. Vorster for his patience and guidance. My co-supervisor, Mr. D. Louwrens for his positive attitude and guidance. Tshwane University of Technology for financial assistance. National Research Foundation for financial assistance. Plastomax Supplies for supplying the material at no cost for this study. Members of the Tshwane University of Technology, specifically the Department of Polymer Technology and Department of Mechanical Engineering for their assistance. My family for their continued support

iv Abstract The effect of curing conditions such as time and temperature on the physical properties of an epichlorohydrin-derived, aliphatic amine cured epoxy system has been investigated. The cure kinetics and thermal behaviour of a setting system was monitored experimentally. The effect of the reaction temperature on gelation (tgel) and vitrification times (tvit) were determined using parallel plate rheology and Differential Scanning Calorometry (DSC) respectively. Rheology was used to determine tgel, by monitoring the complex viscosity, complex modulus and tan against time, whereas the effect of the curing temperature (Tcure) on the glass transition temperature (Tg) of the system was determined from DSC. The cure kinetics and thermal behaviour as determined were used to calculate a time-temperature-transformation (TTT) isothermal cure diagram. The TTT isothermal cure diagram was used as a framework to monitor and characterize changes during the curing process of a thermosetting system. The gelation contour corresponds to the point where viscosity tends to infinity, while the vitrification contour represents Tg rising to Tcure. Rheology measurements under isothermal conditions indicate that Tgel decreases with an increase in Tcure. The effect of the curing temperature (Tcure) and curing time (tc) on Tg, dynamic mechanical properties, and indentation hardness of the system was determined on cured samples characterized with the TTT isothermal cure diagram. Values for Tg, and information on the dynamic mechanical properties of the epoxy system was determined by means of dynamic mechanical thermal analysis (DMTA). Indentation hardness of the cured epoxy system was determined by means of measurements on samples using a Hand-Held Portable Hardness Tester (Barcol Impressor, Model No. 934-1). It is shown that there is a relationship between the cure conditions and the mechanical properties of an epoxy matrix material.

v Samevatting Die effek van die verharding kondisies soos tyd en temperatuur op die fisiese einskappe van n epichlorohydrin-derived, aliphatic amine verhardingsreaksie epoksie systeem word in hierdie studie pepaal. Die verhardingreaksiekinetika en termiese verloop verharding systeem word deur eksperimente bepaal. Die effek op jelpunt (tjel) en vitrifikasie (tvit) punt word deur parallelplaatspanningsreometrie en differensiele skanderingskalorimetrie eksperimente bepaal. Die komplekse viskositeit asook komplekse modulus wat deur spanningsreometrie pebaal is, word gebruik om tjel te definer, terwyl differensile temperature. Die verhardingsreaksiekinetika en termiese verloop wat bepaal is was gebruik om n tyd-temperatuur-transformasie (TTT) isotermiese verharding diagram te op te trek. Die TTT diagram is gebruik as n raamwerk om die verharding reaksie van n reaktiewe systeem te beskryf. Die effek op die verharding temperatuur en verhading tyd op die glastransformasie temperatuur (Tg) sowel as dinamiese meganiese einskappe is deur dinamiese meganiese termiese analise (DMTA) bepaal. Die hardheid van verharde stelsels is bepaal deur hardheid analise met n Hand-Held Portable Hardness Tester (Barcol Impressor, Model No. 934-1). Daar is bewys dat daar wel n verhouding tussen die verharding kondisies en meganiese einskappe van n epoksie system bestaan. skandeerkalorimetrie gebruik word om glas transformasie temeratuur uitg te werk, deur bepaling van verhardingstye en verhardings

vi

CONTENTS
PAGE ACKNOWLEDGEMENTS. ABSTRACT. EKSERP.. LIST OF TABLES LIST OF FIGURES. GLOSSARY. iii iv v vi vii viii

CHAPTER 1 1.1 1.2 1.3 INTRODUCTION ................................................................. BACKGROUND INFORMATION.................................................... OBJECTIVES.. 1 2 3

CHAPTER 2 2.0 2.1 2.2 2.2.1 2.2.1.1 2.2.1.2 2.2.2 2.3 2.3.1 2.3.2 2.3.3 2.3.3.1 2.4 2.4.1 2.4.2 2.4.3 2.4.3.1 LITERATURE SURVEY................................................................... EPOXY RESIN................................................................................. KINETICS OF CURE......................................................... FACTORS INFLUENCING CURE................................................... FUNCTIONALITY............................................................................ MOLECULAR MASS....................................................................... COMPETITION BETWEEN CURE AND DEGRADATION.. GEL POINT..................................................................................... MICROSCOPIC GELATION........................................................... MACROSCOPIC GELATION.......................................................... GELATION STUDY.................................................................. CURING ANALYSIS: RHEOLOGICAL ANALYSIS......................... VITRIFICATION.............................................................................. CURE TEMPERATURE.................................................................. GLASS TRANSITION TEMPERATURE......................................... VITRIFICATION STUDY................................................................. CURING ANALYSIS: DIFFERENTIAL SCANNING CALORIMETRY............................................................................... 17 4 4 5 8 9 9 10 10 11 12 12 13 15 16 16 17

vii 2.5 2.6 2.6.1 2.6.1.1 2.6.1.2 2.6.2 2.6.2.1 CHARACTERIZATION OF THE CURE PROCESS......................... MECHANICAL PROPERTIES.......................................................... MECHANICAL PROPERTIES: MOLECULAR INFLUENCES.. MOLECULAR WEIGHT.................................................................... CROSS-LINKING DENSITY............................................................. MECHANICAL PROPERTY ANALYSIS TECHNIQUES.................. DYNAMIC MECHANICAL ANALYSIS.............................................. 18 19 20 20 20 21 21 22 23 24

2.6.2.1.1 INSTRUMENTATION....................................................................... 2.6.2.1.2 DMA AND DMTA ANALYSIS........................................................... 2.6.2.2 INDENTATION HARDNESS............................................................

CHAPTER 3 3.0 3.1 3.2 3.3 3.3.1 3.4 3.4.1 3.5 3.4.2 3.4.3 RESEARCH METHODOLOGY....................................................... MATERIALS USED AND THEIR PREPARATION......................... GELATION SYUDY........................................................................ VITRIFICATION STUDY................................................................ SAMPLE PREPARATION.............................................................. DYNAMIC MECHANICAL THERMAL ANALYSIS......................... SAMPLE PREPARATION............................................................. INDENTATION HARDNESS......................................................... SAMPLE PREPARATION............................................................. EXPERIMENTAL PROCEDURE 25 25 25 27 27 28 28 31 31 31

CHAPTER 4 4.0 4.1 4.2 4.3 4.3.1 4.3.1.1 4.3.1.2 4.3.1.3 RESULTS AND DISCUSSION........................................................ INTRODUCTION............................................................................. GELATION STUDY......................................................................... VITRIFICATION STUDY................................................................. DETERMINING Tg.......................................................................... Tg VERSUS TCure............................................................................ Tg VERSUS tCure............................................................................. ISO- Tg CONTOURS ...................................................................... 32 32 32 34 35 36 37 39

viii 4.3.1.4 4.4 4.5 4.5.1 4.5.2 4.5.2.1 4.5.2.2 4.5.2.3 4.5.2.4 PREDICTION OF tvit........................................................................ TIME TEMPERATURE TRANSFORMATION DIAGRAM MECHANICAL PROPERTY VS. CURE RELATIONSHIP.............. AREAS FOR ANALYSIS................................................................. DYNAMIC MECHANICAL THERMAL ANALYSIS (DMTA)............. STORAGE MODULUS..................................................................... GLASS TRANSITION TEMPERATURE (Tg)................................... STORAGE MODULUS (E') RELATED TO STIFFNESS LOSS MODUDLUS RELATED TO IMPACT RESISTANCE. 40 45 46 46 48 49 51 52 53 56

4.5.3

HARDNESS ANALYSIS...............................................................

CHAPTER 5 5.0 5.1 5.2 CONCLUSIONS AND RECOMMENDATIONS............................... CONCLUSIONS.. RECOMMENDATIONS.. 58 58 60 61

BIBLIOGRAPHY. APPENDIX A: Material data sheet for Epon Resin 826.

ix LIST OF FIGURES PAGE 2.1 A generalized time-temperature-transformation (TTT) cure diagram. Times to gelation and vitrification during isothermal cure against temperature demarks the boundaries of the four distinct regions states of matter: liquid, gelled rubber, gelled glass, and ungelled glass. (Taken from Prime (1997:1384)).......................... 3.1 Rheometric Dynamic Stress Rheometer (RS-500)............................... 3.2 Typical plot of the loss and storage moduli, tan and complex viscosity (*) against time for Rheological measurements................... 3.3 Typical plot of a DSC scan of endothermic heat flow (mW) against temperature(C)................................................................................... 3.4 Perkin-Elmer Differential Scanning Calorimeter (DSC-7).................... 3.5 Typical plot of the storage modulus (E') and tan against temperature. (Taken from Laza et al. (1998:43)) ................................ 3.6 Rheometric Scientific Solid Analyzer RSA II (DMTA) in dual cantilever mode................................................................................... 3.7 The Impressor: A Hand-Held Portable Hardness Tester (Barcol Impressor, Model No. 934-1)................................................... 4.1 Time-temperature-transformation (TTT) diagram: Gelation contour.... 4.2 Tg (C) vs. Tcure (C) for different isothermal heating times (1 hour through to 72 hours). The reference line of Tg = Tcure is also included............................................................................................... 4.3 Tg (C) vs. Log Time (min) for different isothermal temperatures (Tcure).................................................................................................... 4.4 Limiting Tg (C) vs. Tcure (C),............................................................... 4.5 Time-temperature-transformation (TTT) diagram: Iso-Tg contours...... 4.6 Ln time (min) to fixed Tg values (from Figure 4.5) vs. 1/Tcure (K)......... 4.7 Ln time (min) to macroscopic gelation vs. 1/Tcure (K)........................... 4.8 Summary time-temperature-transformation (TTT) diagram................. 4.9 Detailed TTT-Isothermal cure diagram showing areas for DTMA analysis................................................................................................ 47 37 38 39 41 44 46 36 31 34 30 30 28 28 26 7 26

x 4.10 DMTA scans. Tan and storage modulus (E') vs. temperature (C) for samples cured for different cure times (tcure, min) at different cure temperatures (Tc). 4.10-a: 4.10-b: 4.10-c: DMTA scans for Tcure = 70C.......................................... DMTA scans for Tcure = 90C.......................................... DMTA scans for Tcure = 110C.......................................... 49 49 50 51 53 54 55 56

4.11 Trends in Tg values for samples cured at different temperatures (Tc) and for different times (tcure)................................................................. 4.12 Trends in E' values for samples cured at different temperatures (Tc) and for different times (tcure)................................................................. 4.13 Determination of the area under the storage modulus curve............... 4.14 Trends in impact strength values for samples cured at different temperatures (Tc) for different cure times (tcure)................................... 4.15 Trends in Barcol Hardness values for samples cured at different temperatures (Tc) and for different times (tcure)....................................

xi LIST OF TABLES PAGE 2.1 Outline and classification of epoxy reactive groups, constituents, and ammonia 2.2 and amine reaction groups. Taken from Ellis 5 8 (1993:1,3&12) Glossary of curing terms and characteristic curing parameters. Taken from Prime (1997:1386)............................................................ 4.1 4.2 4.3 4.4 4.5 Average gel times (Log Time) for different frequencies at the respective cure temperatures.............................................................. Average glass transition temperatures for samples cured for different times at different cure temperatures...................................... Numerical relationship between Ln time to a fixed Tg and the reciprocal of Tcure (1/K)......................................................................... Calculated times to vitrification for Tcure = Tg corresponding to isothermal cure temperatures from Figure 4.2..................................... Results from DMTA............................................................................. 43 50 40 35 33

CHAPTER 1
1.1 Introduction Thermoset systems possess good dimensional stability, thermal stability, chemical resistance, and electrical properties. Prime (1997:1381) believes that the increasing use of thermoset material as adhesives, matrix material in reinforced composites, and protective coatings in aeronautical, automotive, construction, electrical and medical application is consequence to thermosets possessing these properties. Using composite materials technology in design and manufacture in engineering fields, as explained by Jones (1993:256), exploits the high specific moduli and strength of reinforcing fibres, supported and held together by epoxy matrices to produce low density high performing structures. According to Morgan (1997:2097), epoxy resin systems are generally used for composite matrices, due to the suitability in composite fabrication. Optimal component design requires an in-depth knowledge of the properties and mechanical capabilities and limitations of constituent materials of composites, and their correct application. The reinforcing fibers provide strength and stiffness, while the epoxy accounts for the desired shape and load distribution between the reinforcing fibers. Morgan (1997:2198), Yasmin and Daniel (2004:8211), and Nicholson et al. (1) are in general agreement that the failure or degradation of composite component performance can be linked to matrix dominated properties. The weak link in the composite is the structural integrity, mechanical, and physical response of this interfacial region. Favorable response of composite structures to applied loads requires competent structural design based on knowledge of the mechanical properties of the constituent parts of the composite.

2 1.2 Background information

Epoxy is a title given to the class of material that has an epoxide group or oxirane ring in its chemical structure. Epoxies exist either as liquids with low viscosity or as solids. Curing of an epoxy system as explained by Sun (2002:11), is through the ring opening reaction of epoxides where addition of molecules produces a higher molecular mass and finally resulting in a three-dimensional structure. The active epoxide groups in uncured epoxy resins react with various curing agents or hardeners that contain, amongst other, hydroxyl, carboxyl, amine, and amino groups. Epoxy resins can be divided into two major groups: epichlorohydrin-derived resins and cycloaliphatic resins. This particular study will focus on the former group, where there is a reaction with a common polyhydroxy compound, such as bisphenol A, which is cured by means of an aliphatic amine curing agent. Processing of thermosetting based composite structures involve curing cycles of the epoxy systems at different isothermal curing temperatures. The degree of cure of thermosetting systems is determined by these curing cycles and has an important effect on the mechanical properties of the final product. To determine, or to a certain extent predict, the mechanical properties through optimized curing cycles requires thorough understanding of cure kinetics and characteristics of epoxy systems. The kinetics of the curing process was studied by means of parallel plate rheology (Halsz and Vorster (2000, 2004), Lange et al. (1996), Sun (2002)) and differential scanning calorimetry (DSC) (Gan et al. (1989), Gillham (1993), Gumen et al. (2000) and Nez et al. (2000)). Information gained from the rheology and DSC analysis was then used to generate additional information on the physical state of the epoxy resin. The information was used to develop an isothermal time-temperature-transformation (TTT) cure diagram (Gillham (1979, 1982, 1990 and 1993), Gan et al.(1989), Jiawu et al. (2000), Nez et al. (2000), Sun (2002), and Reghunadhan Nair

3 (2004:38)). The TTT cure diagram facilitates characterizing the thermosetting system. By making use of the isothermal TTT cure diagram epoxy samples were characterized and cured according to specific cure temperatures (Tcure) and curing times (tcure). These samples were then subjected to dynamic mechanical thermal analysis (DMTA) scans as well as Barcol hardness test to determine mechanical properties of the cured samples under different curing conditions. 1.3 Objectives

This study focuses on gaining insight into the properties and mechanical capabilities by determining the relationship between the curing conditions and physical properties of an epoxy matrix. It aims to: Establish the kinetics of cure of the epoxy thermosetting system through the use of rheology and differential scanning calorimetry. Use the kinetics of cure to develop and construct an isothermal time-temperature-transformation (TTT) cure diagram. To ascertain the mechanical properties of the cured samples through dynamic mechanical thermal analysis (DMTA) and Barcol hardness indentation measurements. Using the TTT diagram and mechanical properties from DMTA scans and Barcol indentation tests to establish a relationship between the curing conditions and physical properties of the epoxy thermosetting system.

4 CHAPTER 2 2.0 2.1 Literature Survey Epoxy Resin

Commercial epoxy resins were first marketed in the 1940s as result of independent work by Pierre Castan in Switzerland, and Sylvan Greenlee in the United States. Even though patents for similar resins had been in existence since the 1930s, Ellis (1993:1) explains that products that would be called epoxy resins in modern times, were synthesized as early as 1891. It is interesting to note that early epoxy resins were reaction products of bisphenol A and epichlorohydrin, and although various other types of resins are available, this is still the major manufacturing route for most resins available today. Epoxy resin is generally found in the form of liquids with low viscosities or as solids. Prime (1997:1716) and Sun (2002:11) report that epoxy resins form part of a class of material classified and characterized according to the reactive epoxide group, or oxirane ring that forms part of the chemical structure. Ellis (1993:1) points out that the term epoxy resin is generally applied to both the unreacted prepolymers and cured resins, even thought all the reactive groups may have reacted in the cured resins. Epoxy resins can be divided into two major groups: epichlorohydrin-derived resins and cycloaliphatic resins. This particular study will focus on the former group, where there is a reaction with a common polyhydroxy compound, such as bisphenol A, which is cured by means of an aliphatic amine curing agent. The reactive epoxy group, bisphenol A and epichlorohydrin reagents, as well as the aliphatic amine group reaction is outlined and classified in Table 2.1.

5
Table 2.1 Outline and classification of epoxy reactive groups, constituents, and ammonia and amine reaction groups. Taken from Ellis (1993:1,3&12)

Ellis (1993:72) asserts that the cure of an epoxy resin involves reaction between epoxy and hardener reactive groups, and that during cure, a liquid or fluid resinhardener mixture is converted to a solid. 2.2. Kinetics of cure

Thermosetting resin is an appropriate title in that it describes the final infusible and insoluble state of a set, or cured system. Liquid resin is mixed with a hardener or catalyst according to a prescribed ratio in order to initiate the curing process. This process is known as polymerization, curing or cross-linking, and is a permanent change from a viscous liquid to a viscoelastic fluid, to an elastic gel

6 (or rubber), and finally to a glassy state. Heat or another form of energy is often required to initiate cure. Proper formulation and complete processing will result in a polymer with a highly cross-linked, infinite three-dimensional network structure, a consequence of the chemical reactions that accompany cure. It is the polymerization or curing process that separates thermosets from other polymer materials such as thermoplastic materials. According to Ellis (1993:72-74) and Prime (1997:1380-1388) the complexity of the curing process of a thermoset is due to steps that are necessary during the handling and storage of the material, prior to cure, and situational factors that influence the process, and state or phase changes that the material undergoes during cure. The major changes to the properties of thermosetting resins from being a viscous fluid to a glassy solid are a result of the cure process. These changes as cure proceeds can be represented in a time-temperature-transformation (TTT) diagram as represented in Figure 2.1 below.

Figure 2.1. A generalized time-temperature-transformation (TTT) cure diagram. Times to gelation and vitrification during isothermal cure against temperature demarks the boundaries of the four distinct regions states of matter: liquid, gelled rubber, gelled glass, and ungelled glass. (Taken from Prime (1997:1384))

The first phase transformation from a viscoelastic fluid to an elastic gel is known as the gel point, and occurs at the gel time (or gelation times), tgel. A distinction must be made between molecular gelation and macroscopic gelation. Molecular gelation occurs at a specific stage during the chemical reaction, and marks the start of the formation of a cross-linked network, whereas the macroscopic consequence of gelation shows a rapid increase in viscosity, and development of elastic properties. The next phase transition from a gel to a glassy solid is vitrification, with glass transition measurements of Tg and times to vitrification, tvit. Prime (1997:1383) points out that vitrification occurs as a result of the chemical reaction that accompanies cure, and is independent of gelation. Vitrification of the system can occur without gelation.

Table 2.2 Glossary of curing terms and characteristic curing parameters. Taken from Prime (1997:1386) gel ult Chemical conversion (e.g., of epoxide groups), fraction reacted, degree of cure at gel point Maximum achievable extent of conversion 1 Time to gelation (gel time) Time to vitrification Cure temperature, a process parameter Temperature below which no significant reaction of uncured resin mixture occurs in a reasonable time period (cf. storage temperature for uncured resin mixture) Lowest temperature at which gelation occurs in the liquid state Minimum temperature at which ultimate conversion occurs in a reasonable time Glass transition temperature, a material property Tg for thermoset with degree of conversion = 0 Tg for thermoset with degree of conversion gel Tg for thermoset with degree of conversion = 1

tgel tvit Tcure Tcure,0


gelTcure

Tcure, Tg Tg0
gelTg

Tg,

Prime (1997:1383) and Gillham (1989:804, 1993) repeatedly point out that the onset of vitrification marks a decrease in the rate of reaction. This decrease is a result of the change in reaction mechanism from being kinetically controlled to diffusion control. 2.2.1. Factors influencing cure Manufacturing and processing of composite parts destined for commercial applications calls for the least amount of cycle/process time, yielding improved properties. It is therefore imperative that the finished product demand set by the market be met; Production and cost effective manufacture therefore compels the use of additives such as catalysts, accelerators, initiators, hardeners, or fillers. According to Prime (1997:1662), catalysts, accelerators, and initiators catalyze the cure process, and in doing so reduce the processing times. Hardeners are incorporated in the network structure of the polymer chains; thereby having an

9 effect on the reaction rate, as well as the ultimate properties of the material, such as the network structure, glass transition temperature, and dynamic mechanical properties. Fillers are generally incorporated to provide or enhance physical properties such as modulus, thermal expansion and dimensional stability. 2.2.1.1 Functionality

According to Odian (1991:40, 106-108), polymerization of bi-functional monomers (f=2) will form linear polymers, whereas polymerization of monomers with more than two functional groups per molecule (f>2) will result in the polymer being cross-linked. With certain monomers, cross-linking will take place, resulting in a network structure in which one or more branches from one molecule become attached to surrounding molecules. 2.2.1.2 Molecular Mass

Odian (1991:19-20) emphasises the importance of the molecular mass of a polymer in its synthesis and application. Also that, the useful mechanical properties uniquely associated with polymeric materials are a consequence of their high molecular weight. It can generally be said that the use of polymer material in practical applications require high molecular weights to obtain high strengths. It should be noted that polymer chains with strong intermolecular forces develop sufficient strength to be useful at lower molecular weights than polymers having weaker intermolecular forces (with higher molecular weights). There is a significant dependency of the properties (other than strength) of a polymer on its molecular weight, albeit different quantities for different properties. Optimum or maximum values of properties will be reached at different molecular weights and can decrease with a further increase. It is therefore necessary for the polymerization process to result in a compromise molecular weight, where sufficient strength is obtained without sacrificing other

10 properties. The control of the molecular weight of a cross-linking process is essential for the practical application. 2.2.2 Competition between Cure and Degradation. Prime (1997:1505) explains that the cure process can often become complicated as a result of competition between degradation and chemical cross-linking reactions. This can often occur as a result of the high cure temperatures required to achieve high glass transition temperatures of high performance systems. Yuan and Mazeika (1991) as reported by Prime (1997:1431), found that isothermal degradation occurs in three stages; during the first stage, there is a decrease in modulus and glass transition temperature (Tg), a probable result due to scission in the epoxy network. The second stage marks the formation of char, and during the third stage char stabilizes, and shows an increase in modulus and Tg. The increase may be caused by an increase in the cross-linking density of the degraded network. The char region can be seen on Figure 2.1. 2.3 Gel point Gelation is the first noticeable change that is encountered during the cure process. Prime (1997:1496) describes gelation as an irreversible change from a liquid to a rubber, and that this change marks the start of the formation of the cross linked network. Malkin and Kulichikhin (1996:277) used gelation to denote the process of fluidity loss, which is a result of the formation of a network of chemical bonds. Prime further defines gelation on a molecular basis, but states that the phenomenon is measured as a macroscopic phenomenon. A distinction between molecular gelation and macroscopic gelation may be drawn; macroscopic gelation is the consequence of the microscopic gelation. Gelation will be discussed on a microscopic and macroscopic level.

11

2.3.1 Microscopic gelation Microscopic gelation marks the start of the formation of the three dimensional network structures that is typical of thermosetting systems, and will have a direct influence on the final structure of the cured resin. It is vital to understand the influence of the occurring changes, and their effect on the final properties of the cured resin. Microscopic analysis of gelation focuses on molecular and chemical changes that take place during polymerization. Ellis (1993:4&72-74), Malkin and Kulichikhin (1996:277), and Prime (1997:13801383, 1496) are in agreement that the cure of thermoset material on a microscopic level is the formation and linear growth of molecular chains. These linear chains form branches, cross-links with other branched chains, and result in the formation of a rigid three dimensional molecular network, all due to the chemical reaction between the epoxy and hardener reactive groups. Malkin and Kulichikhin (1996:277) add that the microcrystals formed act as stable points of the three-dimensional network. These strong physical bonds will cause fluidity loss. Flory (as quoted by Gillham (1989:804)) and Winter (as quoted by Prime (1997:1381)) defines the microscopic gel point of a cross-linked system, as the point where the weight average molecular weight of the network of chains approaches infinity. Molecular gelation is therefore a point that may be measured by the initial identification of an insoluble gel. Flory and Miller et al. (taken from Prime (1997:1382)) are in agreement that molecular gelation generally occur during a well-defined and calculable stage in the course of the chemical reaction. The only requirements are that the reaction mechanism is not dependent on temperature and that there are no non-crosslinking side reactions. Molecular gelation is dependent on the functionality,

12 reactivity, and stoichiometry of the reactants, therefore the theoretical gel point can be calculated if the chemistry of the reactants is known. 2.3.2 Macroscopic gelation Ellis (1993:5&72) points out that the first critical feature of a curing system is gelation and can be macroscopically described as the point where the resins viscosity rapidly approaches infinity. This happens as a consequence of microscopic gelation, which is also defined as such by Prime. Gan et al (1989:804) explain that the transformation is from a viscous fluid to a state of infinite viscosity, or a change from a fluid liquid to a rubbery state. Gillham (1990:1) adds that there is a transformation of liquids with low molecular weight to an amorphous solid with a high molecular weight by means of chemical reactions. This irreversible transformation from a fluid to a rubber means that beyond the gel point a thermosetting resin loses all ability to flow. This marks a very important stage, especially for manufacturing and processing. 2.3.3 Gelation Study The measurements of isothermal times to gelation are important to cure studies and will be used to form the basis of a cure diagram and is discussed in later chapters of the study. The progressive macroscopic change from the liquid state to the rubbery state and/or glassy state which occurs during the setting of reactive systems has been described on both a molecular and macroscopic level. Prime (1997:1496) explains that the detection of an insoluble material, i.e. a cross-linked material (gel) that forms from a reacting soluble material (sol), may be taken as the starting point of measure of gelation.

13 According to Winter, Holly et al., Feve, Boiteux et al. (as quoted by Prime (1997:1496)), the point where the mechanical loss tangent becomes independent of frequency can be used as an alternative measurement starting point. Molecular gelation is taken as the point where the loss modulus (G') and storage modulus (G'') cross each other, i.e., tan =1. The approximation of gelation macroscopically is based on rheological and mechanical changes that occur during the transition from liquid to rubber. The rheological aspect accounts for the changing viscosity, while the mechanical aspect accounts for the evidence of an equilibrium modulus. Initial work by Gan et al. (1989) reported on the use of the torsional braid analysis (TBA) technique to monitor and study gelation and vitrification of a setting system. TBA is a resonant instrument, which according to Gallagher (1997:137), was widely used in the past. The detailed mechanics of the system will not be discussed here, but it is important to understand that the frequency of oscillation is related to the modulus, and that the logarithmic decrement (ratio of the amplitude of successive cycles) is related to tan . Nez et al. (2000) proposed three methods to experimentally determine the time at which gelation occurs (tgel), namely; solubility test, gel-timer and dynamic mechanical analysis (DMA). The solubility test, as described by Hagnaver (1983), is used to determine the time taken for the setting material to reach a fibriform structure in tetrahydrofurane. The DMA measures the dynamic mechanical properties of the epoxy. 2.3.3.1 Curing analysis: Rheological analysis used

Malkin and Kulichikhin (1996:142), Lange et al. (1996), Kulichikin (1997), Halsz et al. (2001), Vorster (1995), Vorster et al. (2004), and Sun (2002) rheology, the study of the change of viscosity, as a mean to determine the point of gel formation. Ellis (193:87-89) explains that even though the initial rate of reaction may be high, the molecular growth is slow. There is a relationship

14 between the degree of cure and viscosity; this allows for cure monitoring by measuring the change in shear viscosity of the setting system. The shear viscosity of a resin increases during cure as a result of gelation and/or the onset of vitrification. According to Halsz et al. (2001:898), rheological analysis of the viscosity of curing systems may be carried out using a small strain (5%) at different frequencies and temperatures. The components of complex modulus, complex viscosity and tan should be measured and recorded against frequency and therefore as a function of time, as depicted in work performed by Halsz et al. (2001:898). It is interesting to note that prior to gelation the behaviour of the liquid resin can be characterized by the zero shear viscosity. Vorster et al. (2004) quoted Flory (1941 & 1959) and Stockmayer (1943 & 1944) who predicted theoretical gel point values using the following equation;
2 1 g = R( f 1)( g 1)
1

[2.1]

where R is the molar ratio between the epoxy and hardener, f is the functionality of epoxy and g the functionality of the hardener. From the data obtained from experimental rheological measurements the gelation times (tg) can be determined using three methods: Extrapolating the inverse of zero viscosity Equilibrium (intersection) of G and G curves of the dynamic modulus Intersection of the tan curves. Halsz et al. (2001:899) obtained and compared gel times according to the three methods discussed above. The time differences were found to be insignificant, thereby permitting single test validity.

15 Times to gelation (Log time in minutes) as described by Gan et al. (1989:807) can be plotted against isothermal cure temperatures to obtain time temperature transformation (TTT) diagrams with gelation contours. Gelation contours on a TTT diagram shows the relationship between tgel and Tcure. The relationship that exists between tgel (Ln time in minutes) to macroscopic gelation versus the reciprocal temperature (1/K) is linear. The slope of the linear relationship represents the apparent activation energy (Ea) for the system. 2.4 Vitrification Wunderlich (1997:460) defines vitrification of a setting system as the transition from a viscous liquid or rubber to an amorphous solid or glass. Gan et al. (1989:803) describes vitrification as the point where the system solidifies. This phase transition is a consequence of the cure reactions, and as Prime (1997:1383 & 1569) explains, vitrification can occur at any stage during polymerization. Glass formation (vitrification) is as a result of the glass transition temperature (Tg) increasing with conversion. It is accepted that vitrification occurs at times to vitrification, tvit, when Tg rises to the cure temperature (Tcure). The measurements of the times to vitrification, similar to times to gelation, are important and will be used in cure studies. It forms the basis of cure diagrams. With reference to Figure 2.1, Malkin and Kulichikhin (1996:120), and Prime (1997:1384) explain that in the glassy state beyond the vitrification line, the cure reaction rate significantly drops. Wisanrakkit and Gillham (as quoted by Nez et
al. (2001:3581), Adabbo and Williams, Dillman and Seferis, and Wisanrakkit and

Gillham (taken from Prime (1997:1623)) explain that isothermal curing of thermosets occur in two distinct stages; the first stage reaction mechanism is kinetically controlled, but changes to diffusion controlled soon after vitrification. There is an increase in the diffusion time scale, whereas segmental mobility, and reacting groups decrease.

16 According to Prime (1997:179), gel conversion is dependent on functionality and stoichiometry, whereas the glass transition is dependent on the frequency of measurements. Therefore indicating the dependence on the dynamic mechanical technique used for monitoring it. 2.4.1 Cure temperature According to Prime (1997:1381), most thermosetting formulations require heat or irradiation to cure. Energy in the form of heat can be supplied to the reaction and will be elevating the cure temperature of the system. The cure temperature (Tcure) is a processing parameter and is the temperature at which the polymerisation or cure process takes place. 2.4.2 Glass transition temperature Scherer, Rekhson and Elliott (from Ellis, 1993:81) point out that glass-forming systems exhibit a glass transition phenomenon. Prime (1997:1401) further defines the point as a transition from a supercooled liquid or rubbery state to a metastable glassy state. According to Wunderlich (1997:226 & 380), there is a definite link between the liquid state and glassy state of a thermosetting resin. The glass transition temperature (Tg) is the main characteristic temperature of the two states, and is the process where the molecular orientation of the liquid state is frozen in position as a glass. Because there is no change in molecular order at Tg, the macroconformation of the molecules in the glass state correspond to that of the liquid. Odian (1991:24 & 29) defines the glass transition temperature, Tg, of a polymeric system, as the temperature where the amorphous or unordered regions of a polymer take on the characteristic properties of the glassy state, typically brittleness, stiffness, and rigidity that is associated with the solid phase.

17 According to Gan et al. (1989:803), Gillham (1990, 1993), and Wang (1992), the sensitivity and one-to-one relationship between Tg and the conversion for thermosetting resins, implies that Tg can be used as an index to measure conversion during cure. The glass transition temperature is a useful, sensitive, easily measured macroscopic parameter for monitoring the cure process of thermosetting systems. Prime (1997:1506&1414) points out that softening of a thermoset is affiliated with Tg, defines the upper use temperature of the system. This information is critical when designing and manufacturing composite structures to be used at elevated temperatures. 2.4.3 Vitrification study As stated before, onset of vitrification of a thermosetting system is when there is a change from a liquid to a glassy solid. This transition is further defined as the point where Tg rises to Tcure. It is therefore important to be able to determine Tg. Ellis (1993:72) points out that when Tcure is too low, vitrification may occur before gelation and inhibit further reaction. 2.4.3.1 Curing analysis: Differential scanning calorimetry

Differential scanning calorimetry (DSC) has been used in work by various groups such as Gillham (1990 & 1993), Nez et al. (2001) and Mafi et al. (2005) to determine the glass transition temperature (Tg) of setting systems. According to Turi (1981) and Wunderlich (1990) as quoted by Prime (1997:1388-1390), DSC forms part of a family of measuring techniques that record the response of a material to temperature scans. This is achieved by measuring and recording the heat flow into a material (endothermic) or out of a material (exothermic). DSC can therefore be used to determine Tg for thermosets; by assuming that the measured heat flow (dH/dt) is proportional to the reaction rate (d/dt), or after integration, the total heat detected is identical to the heat evolved by the curing

18 reaction. Other characteristics that can be determined through the use of DSC are cure, and physical ageing of systems. This technique monitors the setting system during the change from a liquid to rubbery state and finally glass state. According to Gan et al (1989:804) Tg can be measured in three types of experiments:1 2 3 Times to vitrification are obtained when Tg rises to Tcure The progress of setting before vitrification at Tcure can be obtained from measurements of Tg obtained during intermitted cooling Temperature scans first to a temperature below and then to temperatures above Tcure, give Tg values which reflect the state of cure obtained at Tcure, and the stability of the material to temperatures higher than Tcure. Tg values for samples cured at Tcure for tcure determined from DSC scans can be used to monitor cure conversion simply by plotting Tg against Tcure for isothermal heating times. The relationship is linear at lower isothermal cure temperatures. All Tcure versus Tg curves features a maximum. Gan et al. (1989:807) explain that this inflection is due to the competition between cure (which raises Tg) and thermal degradation (which decreases Tg). 2.5 Characterization of the cure process Gillham (taken from Prime (1997:1568)) provides a useful frameworks for the understanding and conceptualization of the changes that occur during the curing process of a thermosetting resin. This is through the development of the isothermal time-temperature-transformation (TTT) cure diagram as depicted in Figure 2.1. The diagram depicts the various physical states of a thermoset (liquid, rubber, and glass) in relation to the two critical phenomena: gelation and vitrification. Sun (2002:91) proposes that the characteristics of the cure process can be understood by combining and describing the rheological analysis and

19 DSC thermal analysis by a time-temperature-transformation (TTT) isothermal cure diagram. Construction of a TTT isothermal cure diagram requires that the relationship between degree of cure and cure time at isothermal cure temperatures be established. The relationship between gel times (tgel), defined as the elapsed time to achieve equal storage and loss moduli, and isothermal cure temperature is represented by the gelation contour. Earlier discussions verify that Tg is a function of degree of cure and dependent on cure temperature. The relationship between Tg and the isothermal cure time can be determined. This allows that the time taken for isothermal cure, or vitrification (tvit), when Tg rises to the cure temperature (Tcure), can be determined. The relationship between tvit and Tcure is represented by the vitrification curve. By plotting the gelation and vitrification curves on the same graph, the isothermal TTT cure diagram for epoxy resin system is constructed. 2.6 Mechanical properties Proper application of epoxy resins requires that the material be a rigid solid at the use temperature Tuse. The use temperature in all these applications is below the glass transition temperature of the cured resin, TuseTg. Work by various authors (Nielsen (1969:69-103), Ellis (1993:159) Paul (1989:187), Gillham (1993), Lange et al. (1996), Morgan (1997:2103), Fabrice and Redford (2002:340), Ahamed et al. (2004), and Mafi et al. (2005)) confirm the influence of molecular weight and cross-linking density on the final mechanical properties of thermosetting systems. Ellis explains that epoxies with higher cross-linking densities have higher Tgs, which often succumbs to brittle failures, whereas epoxies with low cross-linking densities exhibit structures with more mobility that allows for greater plastic deformation. Higher levels of deformation often result in higher fracture toughness. Misev (1991:196-197) substantiates that the mechanical properties are directly proportional to the

20 molecular weight. But only to a point, as the molecular weight is dependent on the resin and cross-linker stoichiometry. Bell as well as Misra et al. (as reported by Ellis (1993:159)) found significant increases in impact strength with an increase in molecular mass between crosslinks. A higher molecular masss between cross-links denotes a lower cross-link density. 2.6.1 Mechanical properties: Molecular influences 2.6.1.1 Molecular weight According to Misev (1991:176), the molecular weight of a thermoset influences the following mechanical properties; tensile strength, impact resistance and melt viscosity. This was confirmed through studies performed by Nielsen, Matsuoka, Walsh and Termonia (taken from Nicholson et al. (date:1)), the molecular weight distribution effects fracture toughness, and impact strength of the polymer. The critically molecular weight dependent property is the glass transition temperature (Tg), which represents the temperature region where there is a phase change, either form a glass to a rubber, or from a rubber to a glass. Tg generally increases with the extend of cure, which denotes molecular growth, and confirms the relationship. Good mechanical properties after curing requires that the molecular weight be high. There is however a constraint; thermosetting systems with higher molecular weights consequently have high viscosities, which causes production problems.

2.6.1.2

Cross-linking density

The cross-linking density is representative of the number of bonds or cross-links that has formed as result of the chemical reaction between the constituents reagents during the cure. Ellis (1993:159), Nielsen (1969:69-103), Paul (1989:187) and Morgan (1997:2103) are in agreement that the fundamental

21 physical and mechanical properties of epoxy resins are determined by the nature and density of its cross-linked network. There is influence on Tg of the polymer, the ability to undergo plastic deformation, therefore its toughness characteristics at a given temperature. Gillham (1993) explains that the maximum cross-linking density for a fully-cured material will yield the minimum internal stress, and maximum modulus. There is of course a limit to the cross-linking density, explained by Reghunadhan Nair (2004:406) and Mimura and Ito (2002:7559), as epoxies with high crosslinking densities tend to have high Tgs, but often exhibit brittle structures. Almeida and Monteiro (1996:330) add that higher cross-linking densities generally result in stiffer structures with higher ultimate strengths. Excessive cross-linking reactions result in a rigid, brittle network structure with a diminished load bearing capacity. 2.6.2 Mechanical property analysis techniques 2.6.2.1 Dynamic mechanical analysis

Methods such as torsional braid analysis (TBA), dynamic mechanical analysis (DMA), and dynamic thermal mechanical analysis (DMTA) are methods of measuring changes in complex modulus, compliance, and viscosity of samples during oscillatory deformation of material samples. The different modes of deformation that a sample material can be subjected to are as follows; flexure, tensile, shear, bending, and compression deformations. Principal applications of DMA for thermosetting material is the measurement of Tg, analysis of crosslinking, mechanical properties developed during the curing process and characterizing cured or partially cured materials. Prime (1997:1464-1465) states that Tg for thermosetting materials can be defined to occur at the maximum value of the loss modulus (E'), the loss compliance, or

22 tan . Mechanical properties such as storage and loss moduli or compliances for samples during or after cure can be measured. The dynamic mechanical modulus, E*, is defined as the ratio of the stress, *, to the strain, *, in the material being tested. The strain is a response to the stress applied, or the stress is a response to the applied strain. E* is a complex quantity and is defined as: E* =

* = E '+iE ' ' *


E' ' E'

[2.2]

tan =

[2.3]

where E' is the storage modulus, a measure of the stress stored in the sample as mechanical energy; E'' is the loss modulus, a measure of stress dissipated as heat; tan is the phase lag between the stress and strain. 2.6.2.1.1 Instrumentation o Torsional braid analysis (TBA) TBA is a dynamic mechanical technique where the specimen for a freely oscillating torsion pendulum experiment consists of an inert substrate or braid coated with the reactive system. The changes from the liquid or rubbery state to the glassy state can be monitored during cure. Gan et al. (1989) and Jiawu et al.(2000) used the TBA to study the cure behaviour and measure macroscopic gelation times tgel and glass transition measurements such as Tg and times to vitrification (tvit) to characterize reactive coatings. Dynamic mechanical analysis (DMA)

DMA is a thermal analysis system module that is designed for four modes of oscillation: fixed frequency oscillation, resonant frequency oscillation, stress relaxation, and creep. Direct measurements of frequency, temperature, time, stress, stain, and phase angles are taken during scans. Measurements can be

23
made in the following modes; three-point bending, dual cantilever, fiber extension, film extension, and parallel plate. Yasmin and Daniel (2004:8213) determined thermomechanical properties such as storage modulus (E), loss modulus (E), damping factor tan, and Tg of samples through DMA scans. Park et al. (2004) used DMA purely to determine the temperature dependence of the loss factor tan for samples modified through synthesis of epoxidized soybean oil. Ahamed et al. (2004) used DMTA to study the effect of stoichiometry on the dynamic mechanical properties of resins systems. Cain et al. (2004:3) performed DMA in dual-cantilever mode on samples. o Dynamic mechanical thermal analysis (DMTA) The DMTA is a fixed frequency thermal analysis module that measures time, temperature, frequency, sample, and suspension stiffness (in-phase and out-ofphase) directly. Three measurement modes are possible: bending mode, singlecantilever (and double cantilever), and shear sandwich geometry. Dynamic mechanical thermal analysis (DMTA) can be used to measure the complex modulus, compliance, and viscosity of materials as functions of temperature in several different modes of oscillatory deformation. Mafi et al. (2005) and Remiro et al. (2001) used DMTA to gauge the effect of curing conditions such as time and temperature on dynamic mechanical behaviour (adhesion, viscoelasticity, and stiffness) of a thermoset system. Laza et al. (1998) used DMTA to study the effect of amine concentrations on the dynamic mechanical properties of resins systems. 2.6.2.1.2 DMA and DMTA analysis DMA according to Prime (1997:1581-1597), is commonly used in characterization of polymeric materials. This is due to the viscoelastic behaviour of the material which changes with time. A reacting system therefore requires a description of the kinetics of the system and the effect of the changing system on the viscoelastic model. The non-isothermal dynamic mechanical characterization of cured or partially cured systems will include effects of additional cure. The

24
viscoelastic model is described by the dynamic mechanical behaviour shown when the measured storage (E), and loss (E) moduli are plotted against temperature. 2.6.2.2 Indentation hardness

The indentation hardness of cured epoxy samples can be measured by means of a Barcol Impressor. Pharr (1998) reviewed and discussed commonly ultra-low used for load indentation (or

nanoindentation)

techniques

measuring

mechanical

properties of thin films and small volumes of material. The different indentation techniques vary and can be classified according to the type of indenter used in the measurement. Although the above mentioned work deals with small quantities of material on a sub-micron scale, it still bears justice to this study as mechanical properties can be determined by analysis of the indentation loaddisplacement data. According to Reghunadhan Nair (2004:89), enhanced modulus, compressive strength and hardness are desirable improvements in epoxy resins. Good mechanical performance offer immediate solutions for challenging problems. Zhang et al. (2004) investigated the enhancement of wear resistance of epoxies and found that the friction and wear behaviour can be improved through enhancement of hardness, stiffness and compressive strength.

25

CHAPTER 3
3.0 3.1 Research Methodology Materials used and their preparation

The base resin used in this study was prepared from the reaction between epichlorohydrin and bisphenol-A (Epon Resin 826 (ARH)) manufactured by Resolution Performance Products, and supplied by Plastomax. The curing agent used in this study was an aliphatic amine (Ancamine) also from Plastomax. Data sheets representing the materials are presented in Annexure A. The components of the system, the epoxy resin and hardener (curing agent), were carefully weighed on an electronic scale and homogeneously mixed to a ratio of 100:25 by mass (as per instruction from the supplier).

3.2

Gelation Study

A Rheometric Dynamic Stress Rheometer (RS-500) (Figure 3.1) equipped with Orchestrator software was used under oscillatory dynamic mode to obtain the experimental data reported in this study. A parallel 25 mm diameter plate measuring head with a 0,2 mm gap was employed. The measurements were carried out by using a small strain (5%) at four frequency values (50, 15, 10 and 1 rad/s) and nine temperatures for each frequency (30, 40, 50, 60, 70, 80, 90, 100 and 110C). The components of complex modulus, complex viscosity and tan were measured and plotted against frequency and therefore as a function of time. The gel point was taken as the crossover value of the two components of the complex modulus curves, namely the storage and loss modulus. (See Figure 3.2). Gelation occurs at time tgel=196 s, indicated by the crossover of the components of complex modulus.

26

Figure 3.1. Rheometric Dynamic Stress Rheometer (RS-500)

G'/G ' Crossover Point: 196

Figure 3.2. Typical plot of the loss and storage moduli, tan and complex viscosity * against time for Rheological measurements

27

The average time for the system to reaction to reach gelation, time Tgel, was calculated from the data and plotted in Figure 3.2. This shows the relationship between the time to gelation (tgel) and curing temperature (Tcure) of the system.

3.3

Vitrification Study

3.3.1 Sample preparation Small amounts of the initial reactive mixture as per section 3.1 (approximately 10 mg) were transferred into small DSC aluminium pans. These samples were grouped in batches of 10 and sealed in plastic zip lock bags. The plastic bags were stored in a freezer, at approximately -15C to prevent any cure taking place. Sample pans were removed from the sealed plastic bags, and were allowed at least 20 minutes to reach room temperature before placing in an oven for isothermal curing at 60, 80, 100,120 and 140C for periods up to 4320 minutes. A Perkin Elmer differential scanning calorimeter, Model DSC 7 (see Figure 3.4), equipped with Pyris software was employed to determine the glass transition temperatures. The heating profile consisted of a heating run from 25C to 150C at 10C/min. The test was carried out in a nitrogen atmosphere with a flow rate of 20 5 ml/min. The parameters measured were the glass transition temperature (Tg) and residual exotherm (Hr) of the reaction after the material had been subjected to isothermal curing at the different curing temperatures. Tg appears as an endothermic shift over a temperature interval in the DSC scan as shown in Figure 3.3. In this study, Tg was taken as the mid-point of the steptransition, while Hr of the remaining reaction appears as an exothermic peak in the temperature range of the rubbery state of the material.

28

Figure 3.3. Typical plot of a DSC scan of endothermic heat flow (mW) against temperature(C).

Figure 3.4 Perkin-Elmer Differential Scanning Calorimeter (DSC-7)

3.4

Dynamic Mechanical Thermal Analysis

3.4.1 Sample preparation The reactive mixture (prepared as described in section 3.1) was poured into specially created flat moulds made from impression material (Zetalabor Platinum Laboratory High Precision Addition Silicone. Manufactured by Zhermack). These samples were stored in a freezer, at approximately -15C to prevent any cure taking place.

29 Samples were removed from the freezer, and were allowed at least 20 minutes to reach room temperature before moving to an oven for isothermal curing. The cured samples were cut to approximately 5mm x 25mm x 1.5 mm, carefully measured and mounted in the DMTA. A Rheometric Scientific Solid Analyzer RSA II (DMTA), equipped with Orchestrator software was used to obtain the experimental data (see Figure 3.6). A Dynamic Temperature Ramp using the dual cantilever mode was employed and measurements of the storage modulus (E'), loss modulus (E) and loss tangent (tan ) of fully cured samples were carried out using a small strain (0.01 %) at a frequency of =6.28 rad/s. The cured samples were subjected to temperature scans from 40 to 160C at a rate of 10C/min. DMTA analyses were performed on samples cured at three different temperatures (70, 90 and 110C) for three different periods (30, 60 and 120 minutes). A typical DMTA scan is shown in Figure 3.5

30

Figure 3.5. Typical plot of the storage modulus (E') and tan against temperature. (Taken from Laza et al. (1998:43))

Figure 3.6. Rheometric Scientific Solid Analyzer RSA II (DMTA) in dual cantilever mode

31 3.5

Indentation hardness

3.5.1 Sample preparation The samples were prepared as described in Section 3.4 under Dynamic Mechanical Thermal Analysis 3.5.2 Experimental procedure A Hand-Held Portable Hardness Tester (Barcol Impressor, Model No. 934-1) was used to obtain the experimental data (see Figure 3.7). The indentor consisted of a hardened steel truncated cone with an angle of 26 with a flat tip of 0.157 mm in diameter. The test method used to determine the indentation hardness was performed in accordance with Standard Test Method for Indentation Hardness of Rigid Plastics by Means of a Barcol Impressor, ASTM Designation: D 2583 87. Barcol indentation analyses were performed on samples cured at three different temperatures (70, 90 and 110C) for three different times (30, 60 and 120 minutes). It must be noted that the impression hardness tests were carried out on samples at room temperature.

Figure 3.7. The Impressor: A Hand-Held Portable Hardness Tester (Barcol Impressor, Model No. 934-1)

32

CHAPTER 4
4.0 4.1 Results and discussion Introduction

The results obtained from the experimental work will be dealt with individually in this chapter. Results from gelation and glass transition temperature studies are used to monitor the extent of cure of the thermoset system, and forms the basis of characterizing the material. The results will be summarized in the form of an isothermal time-temperature-transformation (TTT) diagram. The results obtained from dynamic mechanical thermal analysis (DMTA) and hardness tests will be used to illustrate the relationship between mechanical properties and the molecular structure of the cured material as the molecular structure of the system is dependent on curing conditions.

4.2

Gelation Study

The aim of the gelation study is to determine at what point during the curing process of the thermosetting system gelation (Tgel) occurs. The experimental results obtained from parallel plate rheology tests, as per Section 3.1., were used to obtain crossover values of the storage and loss modulus curves as a function of time. A derived summary of the average gel times (Log Time) for tests performed at the four frequencies (50, 15, 10 and1 rad/s) and seven cure temperatures (30, 40, 50, 60, 70, 80, 90, 100 and 110C) are shown in Table 4.1. Times to gelation for samples tested at low curing temperatures (30, 40 and 50C) are extrapolated from storage and loss modulus curves, as the system solidified prior to reaching gelation. We can classify these as having vitrified before the curing system could form a gel. The greyed out cells represent measurements that could not be recorded due to premature solidification of the resin system. The values of pre-gelation and premature solidification will not form part of the calculations.

33

Table 4.1. Average gel times (Log Time) for different frequencies at the respective cure temperatures Gelation Times (Log t) 30 40 50 60 70 80 90 100 110 50.00 8326.95 4628.70 3007.05 1504.70 1020.21 663.49 443.39 344.07 224.10 Frequency (rad/s) 15.00 8273.35 4830.05 3107.15 1677.15 1126.38 688.20 435.13 297.42 243.00 10.00 8973.95 4597.20 3273.30 1727.45 893.51 650.42 416.90 312.99 289.00 1.00 10009.80 5722.45 3615.80 1661.25 1008.60 680.04 409.87 Average (s) 8896.01 4944.60 3250.83 1642.64 1012.18 670.54 358.65 318.16 252.03 Average (min) 148.27 82.41 54.18 27.38 16.87 11.18 5.98 5.30 4.20 Log Time (min) 2.17 1.92 1.73 1.44 1.23 1.05 0.78 0.72 0.62

Further analysis of data from Table 4.1 presented in Figure 4.1 shows the decrease in gelation times (Tg) for samples cured at higher cure temperatures (Tcure). This indicates that there is a definite relationship between the time taken for the polymer system to react and form a three dimensional network structure and the temperature at which the system is cured. It is obvious that the time taken for the system to form a gel decreases as the cure temperature increases. The oscillation frequency at which the tests were conducted does not have a real effect on gel times.

34

120 110 100 90

Tcure (C)

80 70 60 50 40 30 20 0.40

0.60

0.80

1.00

1.20

1.40

1.60

Log Time (min)


Figure 4.1. Time-temperature-transformation (TTT) diagram: Gelation contour.

4.3

Vitrification Study

The aim of the vitrification study is to determine at which point the thermosetting system vitrifies during the curing process. Vitrification is a transition from a viscous liquid or rubber to a glass and according to Gan et al. (1989:803), vitrification is the point where the system solidifies. According to Prime (1997:1383 &1569), vitrification can occur at any stage during polymerization, either as a consequence of solvent evaporation or chemical reaction. It is therefore important to be able to track and monitor the cure reaction and define the phase changes that occur during cure and in doing so, demonstrate the conditions where vitrification takes place. Work by Gan et al. (1989:803), Gillham (1990, 1993), and Wang (1992), illustrated that there is a sensitive one-to-one relationship between the glass transition temperature, Tg, and the conversion in thermosetting resins. This implies that Tg can be used as an index for measuring conversion during cure. The glass transition temperature is a useful, sensitive, easily measured macroscopic parameter for monitoring the cure process of thermosetting systems. Measurements taken on samples cured at different cure temperatures, Tcure, for different periods, tcure, were measured by differential

35 scanning calorimetry (DSC). The DSC scans were performed as described in Section 3.2.

4.3.1 Determining Tg A typical DSC scan is represented by Figure 3.3. Tg appears as an endothermic shift in the DSC scan. In this study Tg was taken as the mid-point of the endothermic step-transition. The resulting Tg values for samples cured for different times (tcure) at different cure temperatures (Tcure) are summarized in Table 4.2 below.
Table 4.2. Average glass transition temperatures for samples cured for different times at different cure temperatures.
Tg (C) 60 80 100 120 140 Tcure (C) Curing Time in minutes (tcure) 45 82.78 93.23 98.35 101.74 99.86 60 84.84 94.1 98.6 102.47 100.53 90 91.43 96.55 100.28 104.47 102.73 180 92.6 97.56 101.5 106.21 105.37 360 93.05 98.78 102.45 106.57 106.49 600 93.42 102.52 103.84 107.3 107.65 1440 92.5 102.67 104.75 106.51 105.25 2880 94.87 103.4 105.14 104.66 102.34 4320 95.63 104.9 106.35 103.78 101.53

The increase in Tg for systems cured at an isothermal Tcure results in an increase in tcure as indicated in Table 4.2. This confirms the relationship that exists between the glass transition temperature, the cure temperature, and the cure time.

36

4.3.1.1 Tg versus Tcure Values of Tg determined by DSC and the corresponding Tcure values for cure times tcure (1 to 72 h) are plotted in Figure 4.2. The contours show the relationship between Tg, and Tcure.
Tg=107,6C

Tg=Tcure

Figure 4.2. Tg (C) vs. Tcure (C) for different isothermal heating times (1 hour through to 72 hours). The reference line of Tg = Tcure is also included.

The reference lines included are of Tcure = Tg, and Tg = 107.6C. The value for Tg is determined in Section 4.3.1.2. By examining Figure 4.2, especially Tg vs. Tcure for the tcure=10 hours contour, we note a definite increase in glass transition temperature with an increase in cure temperature. It is important to note that samples cured at a constant Tcure for increased time increments, show continuous increases in Tg values. Keeping in mind that Tg is used to monitor cure, we can see from the positive contour slopes of Tg vs. Tcure at lower curing temperatures, that the cure is not complete. Contours at higher curing temperatures, for longer curing times, exhibit a flexure point in the slope of the plot and a definite maximum is observed. This is a result of the competition between cure (which increases Tg)

37 and thermal degradation (which decreases Tg) and is indicative of a deteriorating molecular structure. The change is possibly due to de-vitrification, and molecular breakdown as a result of thermal influences on the molecular structures. According to Gan et al. (1989:807), a system that exhibits contrasting values of Tg >Tcure for Tcure >Tg, even for prolonged cure times (tcure = 72 hours) indicates that the system is highly reactive.

4.3.1.2 Tg versus tcure The variation of Tg with log time at different isothermal temperatures is illustrated in Figure 4.3. It shows that samples cured at constant temperatures for increasing cure times exhibit a general increase in Tg which is a result of the curing reaction taking place and, is in accordance with that reported by Gan et al. (1989:808).

Figure 4.3. Tg (C) vs. Log Time (min) for different isothermal temperatures (Tcure).

38 There is a notable decrease in Tg values after 2.7 Log Time for samples cured for longer periods (Tcure=48 and 72 hours) at higher temperatures (Tcure=120 and 140C). This, according to Gan et al. (1989:808), is due to thermal degradation of the system. Prime (1997:1505) suggests that degradation of the system often competes with cross-linking curing reactions, and is an indication that higher Tcure values are required to achieve higher Tg values. In conclusion it can be stated that there is a competition between the extend of molecular cross-linking and degradation, that results in a superior or mediocre Tg value. Figure 4.4 shows the limiting value of Tg versus Tcure. The highest Tg value observed in this set of data is 107.6C obtained at a cure temperature of 140C. Tg values for DSC scans performed on samples cured at 160C were lower than those for cure at 140C. Therefore 107.6C is the operational (maximum Tg) value assigned to Tg.

110 108 106 104


maxTg=107,6C

Tg (C)

102 100 98 96 94 40 60 80 100 120 140 160 180

Tcure (C)

Figure 4.4. Limiting Tg (C) vs. Tcure (C),

39

4.3.1.3 Iso-Tg contours The data derived from Figure 4.3 was used to develop iso-Tg contours in the format of a time-temperature-transformation (TTT) diagram and is shown in Figure 4.5.

Tcure (C) against Log Time (min)

Figure 4.5. Time-temperature-transformation (TTT) diagram: Iso-Tg contours.

As previously stated, the glass transition temperature is a useful tool to monitor the reactions that occur during cure. The iso-Tg contours can therefore be seen as an extention of conversion curves. Vitrification of the system occurs when Tg rises to Tcure. Vitrification will be discussed in the next paragraph.

40

4.3.1.4 Prediction of tvit Data from Figure 4.5 was used to compile Table 4.3, which, in turn, was used to compile Figure 4.6. This illustrates the relationship between Ln time to a fixed Tg and the reciprocal of Tcure (1/K). The apparent activation energies, E, can be obtained from the slopes of the linear relationships in Figure 4.6 (as in Section 4.1.2): E values are 18.19, 16.90, 26.80, and 46.82 kJ/mol, for the iso-Tg values 95, 98, 100, and 102C, respectively.
Table 4.3 Numerical relationship between Ln time to a fixed Tg and the reciprocal of Tcure (1/K).

Ln Time (min)
Tcure (C)
60 80 100 120 140

Iso-Tg
93
5.73 3.73 2.58

Tcure (K)
333 353 373 393 413

1000/Tcure
3.00 2.83 2.68 2.54 2.42

95
8.01 4.14 2.95 2.30

98
5.00 3.82 3.09 2.60

100
6.26 4.49 3.73 3.89

102
7.71 4.79 3.89 4.31

104

105

5.66 4.37 4.81

7.44 5.07 5.04

41

8.00
Tg: 95C

7.00

Tg: 98C Tg: 100C Tg: 102C

6.00

Ln Time (min)

5.00

4.00

3.00

2.00

1.00 2.30

2.40

2.50

2.60

2.70

2.80

2.90

1000/Tcure (K)
Figure 4.6. Ln time (min) to fixed Tg values (from Figure 4.5) vs. 1/Tcure (K).

According to Gan et al. (1989:810) and Prime (1997:1616-1617) (Arrhenius Reference) the slope of the linear part of the Arrhenius plot, Figure 4.6, represents the apparent activation energies, E, of an epoxy (thermoset) system. Prime further explains that the activation energy may be viewed as a time-temperature shift factor where the cure process can be described by a single overall activation energy. The activation energies, E, as determined from the slopes of curves from Figure 4.6 were used to determine vitrification times as illustrated below in equations 4.1 to 4.4.

42

The linear equations for the iso-Tg plots emanating from Figure 4.6 are listed below:
1000 Ln(t ) = 7.8827. 18.186 Tcure

(r = 1) for iso-Tg = 95C (r = 1) for iso-Tg = 98C (r = 1) for iso-Tg = 100C (r = 1) for iso-Tg = 102C

[4.1] [4.2] [4.3] [4.4]

1000 Ln(t ) = 7.7311. 16.904 Tcure 1000 Ln(t ) = 11.672. 26.803 Tcure 1000 Ln(t ) = 19.252. 46.824 Tcure

The times to vitrification, tvit, were calculated by setting Tg = Tcure in the iso-Tg equations, since vitrification is define to occur at Tg = Tcure. For Tcure = 95, 98, 100 and 102C, the logarithm10 of times to vitrification are 1.447, 1.750, 2.077, and 2.26 minutes respectively. The calculated values were used to construct the isothermal TTT cure diagram simply because of the long experimental times for the system to vitrify. The method used by Gan et al. (1989) to calculate vitrification times for a thermosetting system, assumes the absence of diffusion control, solvent evaporation, and thermal degradation. Their method describes the relationship between the rate of reaction (-dx/dt), reactant concentration (x), activation energy (E), temperature (T,K) and time (t) with the Arrhenius equation:

dx = A exp( E a / RT ) f ( x) dt

[4.5]

or by integrating
F ( x ) = A exp( Ea / RT ) t

[4.6]

43 A being the Arrhenius pre-exponential factor, E the activation energy, R the

gas constant (8,314 J/(mol.K)), and F(x), f(x) functions of x. The analysis of thermal data obtained in this study, with support from work performed by other researchers is summarized in an isothermal timetemperature-transformation (TTT) cure diagram. The vitrification curve in a TTT system cure diagram corresponds to a stage in the reaction at temperature Tg, for time tvit. The iso-Tg contour for any corresponding pair of values of Tg and tvit is given by the following equation;

exp( E a / RTcure ) t cure = exp( Ea / RTg ) t vit

[4.7]

Or

1 1 ln t vit = (E a / R ) + ln t cure Tcure Tg

[4.8]

The use of equation [4.8], in conjunction with data for tcure = 3 hours from Figure 4.2, and E/R = 7882.7, obtained from time temperature conversion behaviour for Tg = 95C (assumed to be constant for lower conversions), indicates that cure at 60C for 3 hours gives Tg = 92.5C, with log tvit = 1.88 min for Tcure = 92.5C. Vitrification times were calculated for isothermal cure and presented in Table 4.4.
Table 4.4 Calculated times to vitrification for Tcure = Tg corresponding to isothermal cure temperatures from Figure 4.2.

Tcure (C, From Fig 4.2)


60 70 80 90 100 110

Tg (C)
92.5 95.2 97.7 99.1 101.4 103.2

Tcure (C)
92.5 95.2 97.7 99.1 101.4 103.2

tvit (Log min)


1.88 1.95 2.02 2.05 2.11 2.15

44

Figure 4.7 indicates the relationship between Ln time to gelation and the reciprocal temperature (K);

1000 Ln(t ) = 4,9374. 11,583 Tcure

(r = 0.986)

[4.9]

for which the apparent activation energy is 11.583 kJ/mol. The value of Tg at gelation is expected to vary with Tcure since the measurement of gelation in this study was carried out by means of parallel plate rheology.

3.50 3.00 2.50

Ln Time (min)

2.00 1.50 1.00 0.50 0.00 2.5 2.6 2.7 2.8 2.9 3 3.1

1000/Tcure (K)
Figure 4.7. Ln time (min) to macroscopic gelation vs. 1/Tcure (K).

The isothermal temperature at which the time to vitrification equals the time to macroscopic gelation (extrapolated) is indicated as
gelTg.

This is calculated
gelTg,

using Eq. (8), where tvit corresponds to reaction temperature tcure = 3 hours of Figure 4.2.

equation

[4.9], and equation [4.10] which summarizes isochronal Tg versus Tcure data for

Tcure = 0,25.Tg + 77,5 (Tg <102C)

[4.10]

45

The relevant expression becomes:

1 4937,4 1 11,583 = 7882.7 + 5.193 gelTg ' 0,25. gel Tg '+77,5 gelTg '
Solving the expression, 40.204C.
gelTg

[4.11]

40.204C, log tvit =

1.82 min for Tcure =

4.4

Time-Temperature-Transformation Diagram

The results of the work done in the first two parts of this chapter can be summarized in the form of a time-temperature-transformation (TTT) isothermal cure diagram, as in Figure 4.8 where the progress of transformation of liquid to glass or to rubber is shown by the gelation, vitrification and iso-Tg contours.

46

Figure 4.8. Summary time-temperature-transformation (TTT) diagram.

4.5

Mechanical property Vs. cure relationship

4.5.1 Areas for Analysis The present investigation is aimed at indicating the relationship between cure and mechanical properties. To facilitate this, a TTT isothermal cure diagram (Figure 4.8) was used to determine regions or zones to illustrate the

47
relationship and is presented in Figure 4.9 below. Figure 4.9 is a detailed version of Figure 4.8. Zone 1 through to zone 3 represents samples cured at Tcure = 70C for cure times (tcure) of 30, 60, and 120 minutes, zones 4 to 6 are representative of samples subjected to Tcure = 90C, and zones 7 to 9 to Tcure = 110C, for cure times of 30, 60, and 120 minutes respectively.

Figure 4.9. Detailed TTT-Isothermal cure diagram showing areas for DTMA analysis.

With reference to the generalised time-temperature-transformation (TTT) cure diagram (taken from Prime (1997:1384)), Figure 2.1, zones 1 and 2 fall into a conversion area beyond gelation but ahead of vitrification, zone 3 is after vitrification.

48
4.5.2 Dynamic Mechanical Thermal Analysis (DMTA) As it would have been very time and material consuming, DMTA was chosen as method of determining mechanical properties instead of tensile, impact and bending tests. The results for storage modulus, E', and loss modulus, E as obtained from DMTA scans in this study are dealt with separately for samples cured isothermally at different cure times (tc). The values for Tcure and tc were determined by means of the time-temperature-transformation (TTT) diagram constructed in the previous section of this chapter. The DMTA temperature scans were performed at a rate of 10C/minute, for temperatures between 25 and 160C.

4.5.2.1 Storage Modulus


The scans show that systems cured at different temperatures (Tcure) display different moduli but broadly similar dynamic mechanical behaviour. Distinct changes in log(E') are evident and coincide with the glass transition temperature (Tg) which is represented by the peak tan values.

49

DMTA Graphs : 70C


1.2
Tan Delta 30 min Tan Delta 60 min 9.5 10.0

Tan Delts 120 min 30 min @ 70C 60 min @ 70C 9.0

0.8 Tan Delta

120 min @ 70C 8.5

0.6

8.0

7.5

0.4
7.0

0.2
6.5

0 20 40 60 80 100 120 140 Temperature (C)

6.0

160

Figure 4.10-a. DMTA scans for Tcure = 70C


DMTA Graphs for 90C
1.4

10.0
Tan Delta 30 min Tan Delta 60 min Tan Delta 120 min

1.2

9.5 9.0 8.5 8.0

60 min @ 90C

Tan Delta

0.8

120 min @ 90C

0.6

7.5
0.4

7.0 6.5 6.0 160

0.2

20

40

60

80

100

120

140

Temperature C

Figure 4.10-b. DMTA scans for Tcure = 90C

Storage Modulus Log E'

30 min @ 90C

Storage Modulus LogE'

50

DMTA Graphs for 110C


1 0.9 0.8 0.7 Tan Delta 0.6 0.5 0.4 0.3 7.0 0.2 0.1 0 20 40 60 80 100 120 140 Temperature (C) 6.5 6.0 160
Tan Delta 30 min Tan Delta 60 min Tan Delta 120 min 60 min @ 110C 120 min 110C 30 min @ 110C

10.0 9.5 9.0 8.5 8.0 7.5

Figure 4.10-c. DMTA scans for Tcure=110C. Tan and storage modulus (E') vs. temperature (C) for samples cured for different cure times (tcure, min) at different cure temperatures (Tc).

The results from Figure 4.10 are tabulated in Table 4.4 below. Subsequent Tables and Figures were derived from this Table.
Table 4.4. Results from DMTA
Tan Cure Temperatures (Tcure, C) 70 90 110 E' (Pa) 30 min Tg (C) 91.30 109.82 110.71 Time (min) Tan 7.95 9.75 9.83 E' (Pa) 60 min Tg (C) 96.06 111.13 114.84 Time (min) Tan 8.45 9.93 10.42 E' (Pa) 120 min Tg (C) 98.14 113.37 112.39 Time (min) 8.65 10.27 10.05

0.74 1.69E+08 0.99 6.01E+07 0.75 1.08E+08

1.10 7.12E+07 0.83 5.65E+07 0.95 1.05E+08

0.83 8.45E+07 1.15 8.60E+07 0.79 5.51E+07

Storage Modulus Log E'

51

4.5.2.2 Glass transition temperature (Tg)


A dynamic mechanical thermal analyser (DMTA) was used to determine the modulus and tan as a function of temperature and/or frequency. The samples were tested under three-point bending mode, known as dual cantilever. According to Laza et al. (1998), and Mafi et al. (2005), the glass transition temperature, Tg, is generally defined as the maximum in the tan curve. The contours of Tg values against tc (log time) as in Figure 4.11 below,

indicates the relationship between Tg, Tcure, and tc as determined by DMTA.

120

110

Tg (C)

100

Tcure : 70C 90 Tcure : 90C Tcure : 110C 80 0 20 40 60 80 100 120 140

Cure Time (min)


Figure 4.11. Trends in Tg values for samples cured at different temperatures (Tc) and for different times (tcure)

The glass transition temperature vs. cure time (tc) contours for Tcure= 70 and 90C show a distinct increase in Tg values with increase in tc. This indicates that the cure reaction is not complete. The Tg values of samples cured at Tcure =70C for tc = 30, 60 and 120 minutes, (zones 1 to 3 on Figure 4.9) are 91.30, 96.06 and 98.14C respectively. The values are higher than expected, when considering that zones 1 and 2 are representative of the area between gelation and vitrification, and zone 3

52
represents the area between iso-Tg contours for Tg = 93, and 95C on the time temperature transformation diagram cure diagram. The Tg values of samples cured at Tcure =90C for tc = 30, 60 and 120 minutes, (zones 4 to 6 on Figure 4.9) are 109.82, 111.13 and 113.37C respectively, and similar to values measured for samples representative of structures zones 1 to 3, higher than expected, considering that zones 4 to 6 fall between iso-Tg contours for Tg = 95, 98 and 100C. Tg values of samples cured at Tcure =110C for tc = 30, 60 and 120 minutes, zones 7 to 9 on Figure 4.9 are 110.71, 114.84 and 112.39C respectively, similar to zones 1 to 3, and zones 4 to 6, are higher than expected when considering that zones 4 to 6 fall between iso-Tg contours for Tg = 98 to 104C. Sircar (1997:1013) is of the opinion that DSC Tg versus DMA Tg values for elastomers differ as a result of the influence of temperature and frequency on the properties of elastomers. DMA scans indicate there an increase in the peak loss factor temperature (Tg) of a material result of an increasing oscillatory frequency during testing: therefore, data obtained from material under static conditions (DSC) cannot foretell behaviour under oscillatory loads. This explains the reason for the difference between DSC Tg and DMA Tg values. The decreasing Tg values observed for the sample cured at 110C for 120 minutes is due to thermal degradation.

4.5.2.3 Storage Modulus (E') (Related to stiffness)


A dynamic mechanical thermal analyser (DMTA) was used to determine the storage modulus and tan as a function of temperature and/or frequency. The samples were tested under a three-point bending mode, known as dual cantilever. Similar to the method of determining the glass transition temperature at the maximum tan curve as described by Laza et al. (1998), and Mafi et al. (2005), values for storage modulus (E') is taken coincidental with the peak values of tan .

53

The storage modulus (E) vs. log time contours in Figure 4.12 depicts the trend in modulus values as determined by DMTA.

200.00 180.00 160.00 140.00

Tcure : 70C Tcure : 90C Tcure : 110C

E' (Mpa)

120.00 100.00 80.00 60.00 40.00 20.00 0.00 20 40 60 80 100 120 140

Time (min)
Figure 4.12. Trends in E' values for samples cured at different temperatures (Tc) and for different times (tcure)

Analysis show that there is a decrease in the storage modulus (E') for samples cured at predetermined temperatures for predetermined times. Samples cured at lower temperatures (Tcure =70 and 90C) show an initial decreasing E for shorter cure times (tc), but an increase for longer tc. Samples cured for longer tc at higher cure temperatures (Tcure =110C) show a decreasing E.

4.5.2.4 Loss Modulus (Related to impact resistance)


Loss modulus (E) and tan values against temperature (C) from DMTA tests for samples cured isothermally for predetermined times were plotted as shown in Figure 4.13 below. The E curve shows in increase with temperature, a maximum value, then an exponential decrease. The maximum E value is measured at the point where the storage modulus E starts to decrease.

54

Figure 4.13. Determination of the area under the storage modulus curve.

Areas beneath the E curves represents the impact strength, energy in joules, of the sample. These values for the samples were plotted and compared as shown in Figure 4.14 below. The curves show the relationship between impact strength, cure time (tc) and cure temperature (Tcure) of the system.

55

7000.00 6500.00
Tc: 70C

6000.00 5500.00 5000.00 4500.00 4000.00 3500.00 3000.00 2500.00 1.40

Tc: 90C Tc: 110C

Impact Strength (KJ)

1.50

1.60

1.70

1.80

1.90

2.00

2.10

2.20

Cure Time (Log Time, min)

Figure 4.14. Trends in impact strength values for samples cured at different temperatures (Tc) for different cure times (tcure)

The curves show that samples cured at higher Tcure and longer tc exhibit greater impact strength. The decrease in impact strength at longer tc show a change in molecular structure, and can be credited to molecular degradation.

56

4.5.3 Hardness Analysis The results obtained from the Barcol hardness tests for this study are shown in Figure 4.15 below. The contours represents the response of the indentation hardness of the polymer samples cured for at the same temperature (Tcure), for different cure times (tc). Tcure and tc were determined according to the timetemperature-transformation (TTT) diagram constructed in the previous section of this chapter.

46 44 42 40 38 36 34 32 30 20 40 60 80
Cure Time (min)

Barcol Hardness (GYZJ 934-1)

70C Cure 90C Cure 110C Cure

100

120

140

Figure 4.15. Trends in Barcol Hardness values for samples cured at different temperatures (Tc) and for different times (tcure)

The Barcol Hardness vs. log time (tc) contours for Tcure =70 and 90C show the relationship that exists between the hardness of the polymer material, Tcure and tc. The contours show that for a constant Tcure and an increasing tc the indentation hardness increases. The hardness vs. log time (tc) for Tcure= 110C shows an initial increase in hardness for samples cured for 30 and 60 minutes. There is however a loss in hardness for the samples cured for 120 minutes.

57
Penetration hardness as determined by the Barcol method is related to the loss modulus. It must be noted that the penetration hardness tests were carried out on samples at room temperature. With this in mind, the results can be correlated with those as determined by the loss modulus, and as this test is far easier and less expensive to carry out that the DTMA investigations will prove to be a useful quality control test. We can conclude that a relationship exists between the hardness of the cured sample, the curing temperature (Tcure) and curing time (tc). The loss in indentation hardness of the samples cured at an elevated temperature (Tcure =110C) and for longer cure times (tc=120 minutes) can be as a result of de-vitrification or molecular degradation due to thermal breakdown.

58

CHAPTER 5
5.0 5.1 Conclusion and recommendations Conclusion

Study of the kinetics of cure and mechanical response of an epoxy system to differing thermal conditions reveals the relationship between the molecular structure of the cured thermoset and cure conditions. This allows for the physical characterization of the system. Samples were cured at different reaction cure temperatures (Tcure) for pre-determined cure times (tcure). The kinetics of cure was determined through the use of rheology and differential scanning calorimetry (DSC). By using parallel plate rheology, times to gelation (tgel) for samples cured at different Tcures were measured and used to determine the relationship between the gelation and cure temperature. The glass transition temperature (Tg) of samples cured at different Tcures was determined using DSC. Tg is used to measure chemical conversion of the setting system. Vitrification or solidification of thermosets transpires when Tg rises to the Tcure. Measured times to vitrification (tvit) for samples cured for different tcures at different Tcures is evidence of the relationship that exists between the variables. The relationship between gel formation at tgel, Tg, and the onset of vitrification at tvit is summarized in the form of a time-temperature-transformation (TTT) isothermal cure diagram. A TTT isothermal cure diagram is a graphical representation of what phases are present at various temperatures and times. Phase transformation of liquid to glass or rubber is shown by the gelation and vitrification contours. The boundaries represented on the TTT isothermal cure diagram were used to set parameters for mechanical tests.

The mechanical property response of the system to different cure conditions concludes the following;

59 o Results from dynamic mechanical thermal analysis (DMTA) indicate that a direct relationship exists between Tg, Tcure, and tcure. Samples cured isothermally for longer tcure showed an increasing Tg. Tg rising is limited to a maximum after which there is a distinct decrease. The decrease is due to thermal molecular degradation. Improved Tg, which results from increased Tcure and longer tcure times allow the system to be used at elevated temperatures (Tuse), but is limited to o Storage modulus (E) values for samples with Tcure= 70 and 90C from DMTA decrease for tcure= 30 and 60 minutes, but show an increase at tcure= 120 minutes. This is not the case for Tcure= 110C, which exhibits a decrease in E values for all cure times. E can be used as an indication of stiffness. o Loss modulus (E) values from DMTA for Tcure= 70C decrease for all cure times. There is an increase in E for Tcure= 90 and 110C between tcure= 30 and 60 minutes, which reaches a maximum and diminishes. It is interesting to note that samples cured at lower temperatures exhibit higher E values, which in turn signifies higher impact strength. o The indentation hardness of samples increased for Tcure= 70, 90 and 110C at between tcure= 30 and 120 minutes, except for samples which represent Tcure= 110C between tcure= 60 and 120 minutes. o The glass transition temperature (Tg) of a thermoset can be used to measure the conversion during cure. When combining and comparing the trend of Tg, E and E against tcure= 60 and 120 minutes for Tcure=70 and 90C from DMTA definite relationships are brought to light;

Increases and decreases in Tg not only from the cure reaction, but as result of variable cure conditions such as Tcure and tcure, implies that the correlation between thermo-mechanical properties and intrinsic structural characteristics of thermosetting systems can be modeled and to a large

60 degree predicted. The molecular weight and cross-linking density influences fracture toughness, impact strength and stiffness respectively.

5.2

Recommendations

This study has demonstrated the need for understanding the influence of the curing conditions on the mechanical properties of thermosetting systems. It is recommended that correct characterization of the thermosetting system form part of material and product development.

Additional work to relate the mechanical reaction to the cure conditions should focus on one mechanical property and not on a spectrum of properties in order to link variables such as cross-linking density, molecular weight, etc. to physical mechanical properties of thermosetting systems.

61 BIBLIOGRAPHY ASTM Designation D2583-87. Standard Test Method for Indentation Hardness of Rigid Plastics by Means of a Barcol Impressor. AHAMED, S., GARG, A., SUNDARARAMAN, S., CHANDRASHEKHARA, K., FLANIGAN, V., KAPILA, S. Dynamic mechanical characterization of soy based epoxy resin system. Proceedings of The Sampe Conference, Long Beach, CA, May 1620. 2004. dALMEIDA, J.R.M., MONTEIRO, S.N. 1996. The effect of the resin/hardener ratio on the compressive behaviour of an epoxy system. In: Polymer Testing. Vol. 15, 329-339. Elsevier. CANTWELL, W.J., KAUSCH, H.H. 1993. Fracture behaviour of epoxy resins. In: ELLIS, B. (ed.). Chemistry and Technology of Epoxy Resins. Glasgow: Blackie Academic & Professional. CAIN, J., POST, N.L., RIFFLE, J.S., LIN, Y.N., LESKO, J.J., CASE, S.W., HESS, P.E. 2004. Post-curing effects on marine VARTM FRP composite material properties for test and implementation. In: Composites 2004 Convention and Trade Show, held in the United States of America on 6-8 October, 2004. Tampa: American Composites Manufacturers Association. COOPER, A.R. (ed.). 1989. Determination of Molecular Weight. New York: John Wiley & Sons, Inc. ELLIS, B. (ed.). 1993. Chemistry and Technology of Epoxy Resins. Glasgow: Blackie Academic & Professional. FABRICE, L., REDFORD, K. 2002. Curing effects on viscosity and mechanical properties of a commercial epoxy resin adhesive. International Journal of Adhesion & Adhesives. Vol. 22. 337-346. Elsevier. FLORY, P.J. 1941. Molecular size distribution in three dimensional polymers. I. Gelation. II. Trifunctional branching units. III. Tetrafunctional branching units. Journal American Chemical Society.36:3083-3100. FLORY, P.J. 1959. Principles of Polymer Chemistry. Cornell University Press, Ithaca.

62

GALLAGHER, P.K. 1997. Thermoanalytical instrumentation, techniques, and methodology. In: Turi, E.A. (ed.). Thermal characterization of polymeric materials, vol.1. 2nd ed. San Diego: Academic Press:1-203. GAN, S., GILLHAM, J.K. & PRIME, B.R. 1989. A methodology for characterizing reactive coatings: Time-TemperatureTransformation(TTT) analysis of the competition between cure, evaporation, and thermal degradation for an epoxy-phenolic system. Journal of Applied Polymer Science, Vol. 37: 803-816. GILLHAM, J. K. 1979. Polymer Engineering Sciences. 19:10 GILLHAM, J.K. 1982. In Developments in Polymer Characterization-3, J. V. Dawkins, Ed., Applied Science Publishers Ltd., London, 1982, Chap. 5, pp. 159-227. GILLHAM, J.K. 1990. The glass transition temperature (Tg) as an index of chemical conversion for a high-Tg amine/epoxy system: chemical and diffusion controlled reaction kinetics. The Journal of Coatings Technology, April. GILLHAM, J.K. 1993. Coating applications of thermosetting cure diagrams: off-stoichiometric high-Tg epoxy/amine systems. The Journal of Coatings Technology, 8/1/1993. GUMEN, V.R., JONES, F.R., ATTWOOD, D. 2000. Prediction of the glass transition temperatures for epoxy resins and blends using group interaction modelling. Polymer, Vol. 42, 5717-5725. HAGNAVER, G. L. 1983. Chermorheology of thermosetting polymers. ACS Symosium Series, vol. 227. American Chemical Society, Washington DC. HALSZ, L., VORSTER, O.C., PIZZI, A., GAUSI, K. 2001. Rheology study of gelling of Phenol-Formaldehyde resins. Journal of Applied Polymer Science, Vol. 80, 898-902. John Wiley & Sons. JAIWU, G., KUI, S., MONG, G.Z. 2000 The cure behaviour of tetraglycidal diaminodiphenyl methane with diaminodiphenyl sulfone. Thermochimicaacta 352-353, 153-158. Elsevier JONES, F.R. Composite materials. In: ELLIS, B. (ed.). 1993. Chemistry and Technology of Epoxy Resins. Glasgow: Blackie Academic & Professional.

63

KULICHIKHIN, S.G. 1997. Rheokinetics of curing of epoxy resins near the glass transition. Polymer Engineering and Science, 8/1/1997. Highbeam Research. LANGE,J., MANSON, J.A.E., HULT, A. 1996. Built-up of structure and viscoelastic properties in epoxy and acrylate resins cured below their ultimate glass transition temperature. Polymer, vol. 37, 26. 58595868. Elsevier. LAZA, J.M., JULIAN, C.A., LARRAURI, E., RODRIGUEZ, M., LEON, L.M. 1998. Thermal scanning Rheometer analysis of curing kinetic of an epoxy resin: 2. An amine as curing agent. Polymer, Vol. 40, 3545. Elsevier. MAFI, R., MIRABEDINI, S.M., ATTAR, M.M., MORADIAN, S. 2005. Cure characterization of epoxy and polyester clear powder coatings using Differential Scanning Calorimetry (DSC) and Dynamic Mechanical Thermal Analysis (DMTA). Progress in Organic Coatings, Vol. 54, 164-169. Elsevier. MALKIN, A.Y. & KULICHIKHIN, S.G. 1996. Rheokinetics. Basel: Hthig & Wepf Verlag. McADAMS, L.V. & GANNON, J.A. 1986. Epoxy Resins. In: Encyclopaedia of Polymer Science and Engineering, vol. 6. Bikales, N. M., Overberger, C. G., Menges G. Wiley-Interscience, New York. MELO, J.D.D., RADFORD, D.W. 2005. Time and temperature dependence of the viscoelastic properties of CFRP by dynamic mechanical analysis. Composite Structures, Vol. 70, 240-253. Elsevier. MIMURA, K., ITO, H. 2002. Characteristics of epoxy resin cured with in situ polymerized curing agent. Polymer, Vol. 43, 7559-7556. Elsevier. MISEV, T.A. 1991. Powder coatings. Chichester: John Wiley & Sons. MORGAN, R.J. 1997. Thermal characterization of composites. In: E.A. Turi, E.A. (ed.). Thermal characterization of polymeric materials, vol. 2. 2nd ed. San Diego: Academic Press: 2091-2261. NICHOLSON, L.M., WHITLEY, K.S., GATES, T.S., HINKLEY, J.A. (Date unknown). How molecular structure affects mechanical properties of an advanced polymer. National Research Council Resident Research

64 Associate & NASA Langley Research Center. Hampton: U.S. Government. NIELSEN, L. E. 1969. J. Macromol. Sci.-Revs Macromol.Chem. C3(1), 69-103. Glasgow: Blackie Academic & Professional NEZ, L., FRAGA, F., CASTRO, A., NEZ, M.R., VILLANUEVA, M. 2000. TTT cure diagram for an epoxy system dilycidyl ether of bisphenol A/1,2 diamine cyclohexane/calcium carbonate filler. Polymer, Vol. 42, 2001: 3581-3587. ODIAN, G. 1991. Principles of Polymerization-3rd ed.. New York: John Wiley & Sons, Inc. PARK, S.J., JIN, F.L., LEE, J.R. 2004. Thermal and mechanical properties of tetrafunctional epoxy resin toughened with epoxidized soybean oil. Materials Science and Engineering, Vol. A 374: 109-114. Elsevier. PAUL, S. 1989. Crosslinking: Chemistry of surface coatings. In: Eastmond, G.C., Ledwith, A., Russo, S., Sigwalt, P. (ed.). Comprehensive polymer science. Oxford: Pergamon Press:149-192. PHARR, G.M. 1998. Measurement of mechanical properties by ultra-low load indentation. Materials Science and Engineering, Vol. A253. 151159. Elsevier. PRIME, B.R. 1997. Thermosets. In: Turi, E.A. (ed.). Thermal characterization of polymeric materials, vol.1. 2nd ed. San Diego: Academic Press:1379-1766. REGHUNADHAN NAIR, C.P., 2004. Advances in addition-cure Phenolic resins. Progress in polymer science, vol. 29. 401-498. Elsevier. REMIRO, P.M., MARIETA, C., RICCARDI, C.CC, MONDRAGON, I. 2001. Influence of curing conditions on the morphologies of PMMAmodified epoxy matrix. Polymer, 42. 9909-9914. Elsevier. SIRCAR, K.A. 1997. Elastomers. In: Turi, E.A. (ed.). Thermal characterization of polymeric materials, vol.1. 2nd ed. San Diego: Academic Press:887-1378. STARING, E., DIAS, A.A., VAN BENTHEM, R.A.T.M. 2002. New challenges for R&D in coating resins. Progress in Organic Coatings, 45:101-117.

65

STOCKMAYER, W.H. 1943. Theory of molecular size distribution and gel formation in branched-chain polymers. Journal of Chemical Physics,11:45-55. STOCKMAYER, W.H. 1943. Theory of molecular size distribution and gel formation in branched-chain polymers. II. General Crosslinking. Journal of Chemical Physics, 12:125-131. SUN, L. 2002. Thermal rheological analysis of cure process of epoxy prepreg. Dissertation for degree of Doctor of Philosophy in Department of Chemical Engineering. Louisiana State University and Agricultural and Mechanical College. TURI, E.A. (ed.). 1997. Thermal characterization of polymeric materials, vol. 1&2. 2nd ed. San Diego: Academic Press. VORSTER, O.C. 1995. The curing of a carboxyl terminated polyestertriglycidyl isocyanurate powder coating. Thesis for degree of Doctorate in Technology: Chemistry in the Faculty of Natural Sciences. Technikon Pretoria. VORSTER, O.C., HALSZ, L. 2004. The use of a rheological technique in the determination of the curing kinetics of a reactive polyester powder coating. The South African Journal of Natural Science and Technology, Vol. 23, 2004: 13-21 WUNDERLICH, B. 1997. The Basis of Thermal Analysis. In: Turi, E.A. (ed.). Thermal characterization of polymeric materials, vol.1. 2nd ed. San Diego: Academic Press:205-482. YASMIN, A., DANIEL, I.M. 2004. Mechanical and thermal properties of graphite platelet/epoxy composites. Polymer, Vol. 45, 8211-8219. Elsevier. ZHANG, Z., BREIDT, C., CHANG, L., HAUPERT, F., FRIEDRICH, K. 2004. Enhancement of the wear resistance of epoxy: short carbon fibre, graphite, PTFE and nano-TiO2. Composites: Part A, Applied Science and Manufacturing. Vol 35. 1385-1392. Elsevier.

You might also like