<%BANNER%>

Study of Hygrothermal Effects and Cure Kinetics on the Structure-Property Relations of Epoxy-Amine Thermosets

Permanent Link: http://ufdc.ufl.edu/UFE0042863/00001

Material Information

Title: Study of Hygrothermal Effects and Cure Kinetics on the Structure-Property Relations of Epoxy-Amine Thermosets Fundamental Analysis and Application
Physical Description: 1 online resource (148 p.)
Language: english
Creator: CHOI,SUNGWON
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: ADHESIVES -- EPOXY -- HYGROTHERMAL -- INTERFACES -- KINETICS -- POLYMERIZATION -- THERMOSETS
Materials Science and Engineering -- Dissertations, Academic -- UF
Genre: Materials Science and Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: For decades, epoxy systems have shown outstanding physical properties for various applications such as coatings, adhesives, and structural composites, providing high performance of strength, stiffness, and resistance to creep. They have numerous advantages over other thermoset resins such as lower cure shrinkage, lower residual stress, chemical resistance, and insulating properties. However, property degradation due to hygrothermal exposure is a critical issue in such applications, in which their nature to absorb moisture up to 7-8 wt% results in a significant loss in glass transition temperature (Tg), modulus, tensile strength, and adhesive strength. These detrimental effects often include some complex behavior in that property increase happens simultaneously with property loss by water plasticization. It has been shown often that the glass transition temperature or modulus changes in a complicated manner during hygrothermal exposure, in that both increase and decrease in Tg and/or modulus is observed during the course of exposure. Various mechanisms have been suggested to describe such complex behavior, but there is no direct experimental evidence supporting them, and a clear explanation of the mechanism for this anomalous behavior remains unanswered. In this study, to specifically describe the mechanism for this complex hygrothermal behavior, direct experimental measurement has been conducted employing thermodynamic analysis and spectroscopic measurements. From the investigation of a model epoxy-amine thermoset and two commercial epoxy products which are used as a seal coat and impregnating resin for structural strengthening applications, Fourier transform near-infrared (FT-NIR) spectroscopic measurements showed that an increase in conversion was responsible for the increase in glass transition, while plasticization occurred simultaneously, rendering the hygrothermal behavior to be complex. To evaluate those factors affecting this complex hygrothermal behavior separately, a relationship between Tg and conversion was constructed for the unexposed system and compared to the corresponding Tg values for the exposed system at the same point of each conversion value. With this method, the conversion was successfully excluded to compare the Tg values directly between the unexposed and exposed systems. This analysis indicates the effects of other factors are negligible compared with the large plasticization effect. The plasticization effect was also evaluated quantitatively by observing Tg differences between the unexposed and the exposed system as a function of absorbed water amount estimated by a characteristic water peak in the NIR spectra. For another part of this dissertation, the effect of water on cure kinetics was investigated with the hypothesis that water significantly accelerates the cure kinetics of epoxy-amine systems. The near FT-IR demonstrated that a small amount of water addition significantly accelerated the cure reaction in terms of epoxide conversion, in which water acts as a strong catalyst. A modified mechanistic model was successfully used to directly compare the effect of hydroxyl from water addition and that from auto-acceleration by the reaction. Comparison of the kinetic parameters obtained from this model shows that the two influences are very close, with the difference in the kinetic parameter values from the Arrhenius relation being less than one order of magnitude. In the final section of this work, in an effort to enhance the durability of bonding properties of an epoxy adhesive for structural strengthening application, silane coupling agents were applied to the surface of concrete before the adhesive was applied. Slant shear testing results show that the durability of the interfacial bonding was significantly enhanced with the use of an epoxy-functional silane coupling agent, which is attributed to the replacement of the weak hydrogen bonds between epoxy and cement by strong covalent bonds. The results indicate that the chemical nature of bonding plays a large role at the interface bonding of epoxy resin for structural strengthening applications.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by SUNGWON CHOI.
Thesis: Thesis (Ph.D.)--University of Florida, 2011.
Local: Adviser: Douglas, Elliot P.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0042863:00001

Permanent Link: http://ufdc.ufl.edu/UFE0042863/00001

Material Information

Title: Study of Hygrothermal Effects and Cure Kinetics on the Structure-Property Relations of Epoxy-Amine Thermosets Fundamental Analysis and Application
Physical Description: 1 online resource (148 p.)
Language: english
Creator: CHOI,SUNGWON
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: ADHESIVES -- EPOXY -- HYGROTHERMAL -- INTERFACES -- KINETICS -- POLYMERIZATION -- THERMOSETS
Materials Science and Engineering -- Dissertations, Academic -- UF
Genre: Materials Science and Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: For decades, epoxy systems have shown outstanding physical properties for various applications such as coatings, adhesives, and structural composites, providing high performance of strength, stiffness, and resistance to creep. They have numerous advantages over other thermoset resins such as lower cure shrinkage, lower residual stress, chemical resistance, and insulating properties. However, property degradation due to hygrothermal exposure is a critical issue in such applications, in which their nature to absorb moisture up to 7-8 wt% results in a significant loss in glass transition temperature (Tg), modulus, tensile strength, and adhesive strength. These detrimental effects often include some complex behavior in that property increase happens simultaneously with property loss by water plasticization. It has been shown often that the glass transition temperature or modulus changes in a complicated manner during hygrothermal exposure, in that both increase and decrease in Tg and/or modulus is observed during the course of exposure. Various mechanisms have been suggested to describe such complex behavior, but there is no direct experimental evidence supporting them, and a clear explanation of the mechanism for this anomalous behavior remains unanswered. In this study, to specifically describe the mechanism for this complex hygrothermal behavior, direct experimental measurement has been conducted employing thermodynamic analysis and spectroscopic measurements. From the investigation of a model epoxy-amine thermoset and two commercial epoxy products which are used as a seal coat and impregnating resin for structural strengthening applications, Fourier transform near-infrared (FT-NIR) spectroscopic measurements showed that an increase in conversion was responsible for the increase in glass transition, while plasticization occurred simultaneously, rendering the hygrothermal behavior to be complex. To evaluate those factors affecting this complex hygrothermal behavior separately, a relationship between Tg and conversion was constructed for the unexposed system and compared to the corresponding Tg values for the exposed system at the same point of each conversion value. With this method, the conversion was successfully excluded to compare the Tg values directly between the unexposed and exposed systems. This analysis indicates the effects of other factors are negligible compared with the large plasticization effect. The plasticization effect was also evaluated quantitatively by observing Tg differences between the unexposed and the exposed system as a function of absorbed water amount estimated by a characteristic water peak in the NIR spectra. For another part of this dissertation, the effect of water on cure kinetics was investigated with the hypothesis that water significantly accelerates the cure kinetics of epoxy-amine systems. The near FT-IR demonstrated that a small amount of water addition significantly accelerated the cure reaction in terms of epoxide conversion, in which water acts as a strong catalyst. A modified mechanistic model was successfully used to directly compare the effect of hydroxyl from water addition and that from auto-acceleration by the reaction. Comparison of the kinetic parameters obtained from this model shows that the two influences are very close, with the difference in the kinetic parameter values from the Arrhenius relation being less than one order of magnitude. In the final section of this work, in an effort to enhance the durability of bonding properties of an epoxy adhesive for structural strengthening application, silane coupling agents were applied to the surface of concrete before the adhesive was applied. Slant shear testing results show that the durability of the interfacial bonding was significantly enhanced with the use of an epoxy-functional silane coupling agent, which is attributed to the replacement of the weak hydrogen bonds between epoxy and cement by strong covalent bonds. The results indicate that the chemical nature of bonding plays a large role at the interface bonding of epoxy resin for structural strengthening applications.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by SUNGWON CHOI.
Thesis: Thesis (Ph.D.)--University of Florida, 2011.
Local: Adviser: Douglas, Elliot P.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0042863:00001


This item has the following downloads:


Full Text

PAGE 1

1 STUDY OF HYGROTHERMAL EFFECTS AND CURE KINETICS ON THE STRUCTURE PROPERTY RELATIONS OF EPOXY AMINE THERMOSETS: FUNDAMENTAL ANALYSIS AND APPLICATION By SUNGWON CHOI A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNI VERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2011

PAGE 2

2 2011 Sungwon Choi

PAGE 3

3 To my wife Kyongah

PAGE 4

4 ACKNOWLEDGMENTS First of all, I deeply appreciate my advisor, Dr. Dou glas for his intellectual guidance and encouragement which enable all this work to be possible. Without his support, none of this work would be successful. I would also like appreciate the valuable advice and encouragement from my committee members for dev eloping on my research: Dr. Hamilton, Dr. Mecholsky, Dr. Sigmund, and Dr. Craciun. I would also like to express my gratitude to all the members of the Dr. Douglas group: Sangjun Lee, Changhua Liu, Andrew Stewart, Laura Diers, Justin Dennison, Andrew Janiss e, Sasha Perkins, and Andrea Phantu. They really gave me a lot of kind support and collaboration necessary to make out all these things. The research work carried out in this dissertation was funded by National Science Foundation and the Alumni Graduate Aw ard from the University of Florida.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ .......... 8 ABSTRACT ................................ ................................ ................................ ................... 11 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .... 13 2 REV IEW OF LITERATURE ................................ ................................ .................... 18 Effect of Water on Physical Properties of Epoxy Amine Thermosets ..................... 18 Plasticization with H ygrothermal E xposu re in E poxy S ystems ......................... 18 Anti Plasticization with H ygrothermal E xposure in E poxy S ystems .................. 21 Effect of C rosslink D ensity on G lass T ransition in E poxy A mine T hermosets ........ 24 Cure Kinetics of E poxy A mine T hermosets ................................ ............................ 25 Mechanism of E poxy C ure R eaction ................................ ................................ 25 Cure K inetic M odels for E poxy A mine C ure R eactions ................................ .... 28 Functional G roup A nalysis using FT IR S pectrometry ................................ ...... 30 Epoxy R esins for S tructural S trengthening A pplication ................................ ........... 32 Durability of Interfacial Bonding between E poxy and C eme n titious M aterials with H ygrothermal E xposure ................................ ................................ ......... 32 Adhesion Mechanism of E poxy B onding with C ementitious M aterials ............. 35 Silane C oupling A gents to E nhance the D urability of A dhes ive S trength during H ygrothermal E xposure ................................ ................................ ...... 36 3 COMPLEX HYGROTHERMAL EFFECTS ON GLASS TRANSITION OF EPOXY AMINE THERMOSETS: A STUDY OF A MODEL EPOXY AMINE ........... 39 Background ................................ ................................ ................................ ............. 39 E xperimental Section ................................ ................................ .............................. 40 Materials ................................ ................................ ................................ ........... 40 Techniques ................................ ................................ ................................ ....... 42 R esults ................................ ................................ ................................ .................... 46 D iscussion ................................ ................................ ................................ .............. 50 Summary ................................ ................................ ................................ ................ 57 4 COMPLEX HYGROTHERMAL EFFECTS ON GLASS TRANSITION OF EPOXY AMINE THERMOSETS: A STUDY OF COMMERCIAL PRODUCTS ....... 59 Background ................................ ................................ ................................ ............. 59

PAGE 6

6 Experimental S ection ................................ ................................ .............................. 60 Materials ................................ ................................ ................................ ........... 60 Techniques ................................ ................................ ................................ ....... 62 Results ................................ ................................ ................................ .................... 65 Discussion ................................ ................................ ................................ .............. 71 Summary ................................ ................................ ................................ ................ 79 5 EFFECT OF WATER ADDITION ON CURE KINETICS OF AN EPOXY AMINE THERMOSET ................................ ................................ ................................ ......... 81 Background ................................ ................................ ................................ ............. 81 Experimental Sec tion ................................ ................................ .............................. 82 Materials ................................ ................................ ................................ ........... 82 Techniques ................................ ................................ ................................ ....... 84 R esults and Discussion ................................ ................................ ........................... 87 Reaction Chemistry Epoxide Conversion ................................ ...................... 87 Reaction Kinetics ................................ ................................ .............................. 93 Sum mary ................................ ................................ ................................ .............. 105 6 ROLE OF CHEMICAL BONDING ON THE DURABILITY OF EPOXY ADHESION FOR STRUCTURAL STRENGTHENING APPLICATION ................. 107 Background ................................ ................................ ................................ ........... 107 Experimental Section ................................ ................................ ............................ 108 Sample Preparation ................................ ................................ ........................ 108 Exposure Env ironments ................................ ................................ ................. 112 Slant Shear Testing ................................ ................................ ........................ 113 Analytical Method: FT IR Spectroscopy ................................ ......................... 115 R esults ................................ ................................ ................................ .................. 115 Discussion ................................ ................................ ................................ ............ 126 Summary ................................ ................................ ................................ .............. 131 7 GENERAL CONCLUSIONS AND FUTURE WORKS ................................ ........... 133 LIST OF REFERENCES ................................ ................................ ............................. 138 BIOGRAPHICAL SKETCH ................................ ................................ .......................... 148

PAGE 7

7 LIST OF TABLE S Table page 2 1 Effects of continuous water exposure on the durability of epoxy/concrete or FRP bond strengths ................................ ................................ ............................ 33 2 2 Effect of continuous exposure to synthetic solutions and sea water on the durability of epoxy/concrete or FRP bond strengths ................................ ........... 34 2 3 Effects of cyclic exposure on the durability of epoxy/concre te or FRP bond strengths ................................ ................................ ................................ ............. 34 3 1 Cure and exposure conditions for hygrothermal exposure. ................................ 41 3 2 Cure conditions for the master pl ot of T g vs. crosslink density for unexposed epoxy. ................................ ................................ ................................ ................. 41 4 1 Comparison of observed T g s and epoxide conversion for unexposed systems between systems I and II ................................ ................................ .................... 61 4 2 Cure and exposure conditions for hygrothermal exposure in monitoring the changes in T g s using DSC and changes in cross link density using FT IR. ........ 62 4 3 Cure conditions for the master plot of T g vs. conversion for unexposed systems I and II ................................ ................................ ................................ 62 4 4 Assignment of observed peak for systems I and II in NIR spectra ...................... 65 5 1 Summary of kinetic rate constants, kinetic constants ratio and estimated values of amount of initially absorbed moisture for each system ...................... 104 5 2 Values of act ivation energies and related pre exponential factor for each system ................................ ................................ ................................ .............. 104 6 1 Cure and exposure conditions for preliminary slant shear testing. ................... 113 6 2 Cure and exposure conditions for main slant shear testing. ............................. 113 6 3 Failure loads from neat cylindrical concrete specimen compressive testing ..... 116 6 4 Possible assignment of DRIFT spectra ................................ ............................ 128

PAGE 8

8 LIST OF FIGURES Figure page 2 1 Suggested cure mechanism for ep oxide amine reaction involving the formation of termolecular intermediates ................................ ............................. 27 2 2 Reaction p rocess of Alkoxy Silane ................................ ................................ ..... 37 3 1 Chemi cal structures of diglycidyl ether of bisphenol A (DGEBA) and poly(oxypropylene)diamine (POPDA). ................................ ................................ 41 3 2 Typical NIR spectra for unexposed systems with different cure conditions. ....... 43 3 3 Experimental confirmation of reliability of quantitative analyses from FTIR ........ 44 3 4 Typical DSC curves after water immersion for 2 days at different temperatures compared with no exposure. ................................ ........................ 46 3 5 Changes in DSC curves as a function of immersion time for each exposure condition ................................ ................................ ................................ ............. 47 3 6 Changes i n T g a fter water immersion at different temperatures .......................... 48 3 7 Typical Near Infrared spectra and changes in conversion for samples with different exposure conditions ................................ ................................ .............. 49 3 8 T g ................................ ... 52 3 9 Direct comparison of T g between unexposed and exposed syst ems using the master plot of T g versus c onversion ................................ ................................ ... 54 3 10 Relative amount of water absorption with exposure time estimated by normalized area of the characteristic water peak from NIR. ............................... 55 3 11 T g differences between unexposed and exposed systems as a function of estimated amount of absorbed water. ................................ ................................ 56 4 1 Comparison of t ypical NIR spectra for each system with different cure conditions ................................ ................................ ................................ ........... 61 4 2 Typical DSC curves after water immersion for 7 days at different temperatures compared with no exposure ................................ ......................... 66 4 3 Changes in T g at different exposure conditions after curing 4 weeks at room temperature in air ................................ ................................ ............................... 67 4 4 Near infrared spectra for samples immersed in water for 7 days at different temperatures compared wit h an unexposed / control sample .............................. 68

PAGE 9

9 4 5 Changes in conversion at different exposure conditions after curing 4 weeks at room tem perature in air ................................ ................................ .................. 70 4 6 T g ............................... 72 4 7 Direct comparison of T g between unexposed and exposed for system I using the master plot of T g versus conversion ................................ .............................. 74 4 8 Direct comparison of T g between unexposed and exposed for system II using the master plot of T g versus conversion ................................ .............................. 75 4 9 Relative amount of water absorption with exposure time estimated by normalized area of the characteristic water peak from NIR ................................ 76 4 10 T g differences between unexposed and exposed systems as a function of estimated amount of absorbed water ................................ ................................ 78 5 1 Chemical structures of diglycidyl ether of bisphenol A (DGEBA, top ) and poly(oxypropylene)diamine (POPDA, bottom). ................................ ................... 83 5 2 Typical FT IR Spectra for DGEBA POPDA system ................................ ............ 84 5 3 Experimental confirm ation of the reliability of quantitative analyses from FTIR .. 86 5 4 Comparison of normalized area for epoxide around 4530cm 1 obtained from IR spectra for different water contents and cure cond itions. ............................... 88 5 5 Epoxy conversion versus cure time with different contents of added water ........ 89 5 6 Concentration profiles for epoxid e and primary amine ................................ ....... 90 5 7 Concentration profiles for secondary and tertiary amine ................................ .... 91 5 8 Reduced rate of epoxy primary amine re action as a function of [OH] chain ........... 97 5 9 Plotting of intercept values (k 1 water ) from Figure 5 8 as a function of the contents of added water ([OH] water ) ................................ ................................ ..... 98 5 10 Comparison of experimental and predicted values for reaction rate of primary amine ................................ ................................ ................................ .................. 99 5 11 Comparison of experimental and predicted values for reaction ra te of epoxide 100 5 12 Arrhenius relations for the rate constants of k 1 C 1 W 2 C 2 W ............... 105 6 1 Pictures of apparatus to prepare ....................... 109 6 2 Chemical structures of DGEBA, POPDA, epoxy functional silane, and amino functional silane. ................................ ................................ ............................... 11 1

PAGE 10

10 6 3 Schematic of slant shear testing method. ................................ ......................... 114 6 4 Results of preliminary slant shear testing ................................ ......................... 116 6 5 Failure sur faces of control specimens for preliminary slant shear testing ......... 118 6 6 Failures of exposed specimens for preliminary slant shear testing ................... 119 6 7 Failures for primary slant shear testing ................................ ............................. 121 6 8 Results of primary slant shear testing for control specimens. ........................... 122 6 9 Results of primary slant shear testing for water immersion at 30C. ................ 123 6 10 Results of primary slant shear testing for water immersion at 40C. ................ 124 6 11 Results of primary slant shear testing for water immersion at 50C. ................ 125 6 12 Results of primary slant shear testing for water immersion at 60C. ................ 126 6 13 FT IR spectra for concrete powders before and after the modification using epoxy functional silane compared with those of the utilized silane ................... 127 6 14 Percent increase of failure load for primary testing with the silane modification of concrete bonding surfaces. ................................ ...................... 130

PAGE 11

11 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy STUDY OF HYGROTHERMAL EFFECTS AND CURE KINETICS ON THE STRUCTURE PROPERTY RELATIONS OF EPOXY AMINE THERMOSETS: FUNDAMENTAL ANALYSIS AND APPLICATION By Sungwo n Choi May 2011 Chair: Elliot P. Douglas Major: Materials Science and Engineering For decades epoxy systems have shown outstanding physical properties for various applications with numerous advantages over other thermoset resins such as lower cure shr inkage lower residual stress, chemical resistance, and insulating properties. However, property degradation due to hygrothermal exposure is a critical issue in applications, in which their nature to absorb moisture results in a significant loss in glass t ransition temperature (T g ), modulus, tensile strength and adhesive strength. These detrimental effects often include some complex behavior i n that both increase and decrease in T g and/or modulus is observed during the course of exposure. In this study, to specifically describe the mechanism for this complex hygrothermal behavior, direct experimental measurement has been conducted employing thermodynamic analysis and spectroscopic measurements. Fourier transform near infrared (FT NIR) spectroscopic measurem ents showed that an increase in conversion was responsible for the increase in T g while plasticization occurred simultaneously, rendering the hygrothermal behavior to be complex. To evaluate those factors affecting this complex hygrothermal behavior sep a r ately, a relationship between T g and

PAGE 12

12 conversion was constructed for the unexposed system and compared to the corresponding T g values for the exposed system at the same point of each conversion value. For another part of this dissertation, the effect of wat er on cure kinetics was investigated with the hypothesis that water sig nificantly accelerates the cure kinetics of epoxy amine systems The near FT IR demonstrated that a small amount of water addition significantly accelerated the cure reaction in terms o f epoxide conversion. A modified mechanistic model was successfully used to directly compare the effect of hydroxyl from water addition and that from auto acceleration by the reaction In the final section, in an effort to enhance the durability of bonding properties of an epoxy adhesive for structural strengthening application silane coupling agents were applied to the surface of concrete before the adhesive was applied. Slant shear testing results show that the durability of the interfacial bonding was s ignificantly enhanced with the use of a n epoxy functional silane coupling agent. The results indicate that the chemical nature of bonding plays a large role at the interface bonding of epoxy resin for structural strengthening application s

PAGE 13

13 CHAPTER 1 INTR ODUCTION Over past decades, epoxy resins have shown outstanding performances in various applications such as surface coatings, structural adhesives, packaging, and engineering composites, providing high performances of strength, stiffness, and resistance t o creep. 1, 2 E poxy resins have numerous advantages over other thermoset resins such as lower cure shrinkage and residual stress, chemical resistance, insulating properties and availability of the resin ranging from low viscous liquid to tack free solids, but the most a convenient use of a wide range of temperatures by judicious selection of curing agents enabling good control over the degree of cross lin 3 In the se use s of epoxy materials, additional variables are often included into a given system such as solvent, concentrations, viscosity, and mixing rates, providing another method to control degree of cross linking or the speed of reaction. 4 9 Also, an additional incorporation of catalysts such as imidazole have extensively been utilized as a convenient way to regulate the condition of manufacturing processes. 10 1 3 This has been preferred in many industrial areas such as outdoor or electronic applications where high temperature cannot be used easily as a means of accelerating the curing reaction in order to take advantages of easy installation Addition of hydroxy l containing coupounds such as water or various types of alcohols and phenols has also known to increase the rate of cure reaction. C atalysis effects of hydroxyl groups in those types of added compounds contribute the ring opening of epoxides although the exact impact of added water has not been suggested quantitatively. 4, 14, 15

PAGE 14

14 Although the epoxy resins have shown the excellent performances in various engineering fields, it is a major limitation of cross linked ep oxy resins that their tendency to absorb considerable amounts of moisture up to 7 ~8 wt% because the absorbed moisture induces detrimental effects on their physical properties. A number of investigations have reported that hygrothermal exposure of epoxy sys tems results in significant loss of properties such as glass transition temperatures (T g s) 16 26 modulus 27 31 tensile 28, 29, 31 and adhesive strength 32 Plasticization by absorbed water is considered to be a primary reason to induce property loss, whereas other mechanisms such as hydrolysis and hygrothermal stres ses causing swelling, formation of microcracks and crazes, and polymer chain scission may also be factors depending on the material and exposure condition 25, 26, 28, 29, 31 Regarding the interface bonding degradat ion of epoxy adhesives with hygrothermal exposure, the property loss can be explained by general features of bond degradation, in which water penetrates through a permeable adherend or adhesives and causes deterioration of the bond by altering mechanical p roperties of adhesives. 33 On the other hand, there can be another viewpoint describ ing that disruption of hydrogen bo nds at the interface between the adhesive and given adherent play s a n important role. In this description, displacement of epoxy with water under water/moisture exposure induces broken hydrogen bonds between epoxy and given adherent, resulting in reduced i nterface bond strength. 34, 35 Although some studies on epoxy metal adhesion have supported the th eory that broken hydrogen bonds at the interface results in significant loss of adhesive bond strength 34, 35 the chemical nature of epoxy adhesive for structural strengthening

PAGE 15

15 applications at the interface with ceramic materials still remains unanswered and almost nothing known about how that bond is affected by moisture. In addition hygrothermal ex posure often introduces complex behaviors of epoxy resin itself which lead the overall understanding of the system under environmental exposure to be complicated. In such complex behaviors under the exposure, not only the property decrease by plasticizati on is shown, but an increase in physical properties is also observed, i.e., an increase in T g s and/or modulus with hygrothermal exposure. 29, 36 40 Particularly in order to obtain the advantages of easy installation in electronic or structural applications, many adhesives or coatings have been often cured at low temperature where complete curing reactions of thermosets have not been reached. In such cases, with changes in environments from initial condition such as e levated temperature or with passage of time, additional cure can occur which has resulted in an increase of the glass transition with property decrease by plasticization at the same time, hence rendering the hygrothermal behavior to be complex Various mec hanisms have been suggested to describe such complex hygrothermal behavior s such as reactivation of the post curing reaction 29, 36 40 different types of hydrogen bonded water 41 43 and micro structural effect 41 but none of these explanations is there any direct experimental evidence supporting them, and a clear explanation of the mechanism for this anomalous behavior remains unanswered In this study, hygrothermal behaviors of both epoxy resin itself and the interface between epoxy and cementitious material have been investigated. The epoxy amine systems being examined in this study are the generic system including DGEBA and

PAGE 16

16 polyoxyprop ylene diamine with different commercial epoxy products which used as a seal coat, impregnating resins and adhesives for structural applications. For the first part of this work, hygrothermal effects on physical properties of epoxy amine thermosets were inv estigated to evaluate each possible factor which affects the hygrothermal behaviors of the systems. In this investigation, direct experimental evaluations with a thermodynamic analysis and spectroscopic measurements have been proposed to specifically descr ibe the mechanism for the complex hygrothermal behavior. The study has also focused on trying to quantify different factors separately (Chapter s 3 and 4 ). Afterwards, in order to evaluate effect of hydroxyl from water on cure kinetics with the hypothesis t hat water can significantly accelerates the cure kinetics of epoxy amine systems, small amount of water was added to initial mixture of an epoxy amine system. This study has focused on the direct evaluation of the hydroxyl effect from water addition quanti tatively using a modified mechanistic modeling with a s pectroscopic analysis (Chapter 5 ). At the final section of this work, in an effort to enhance the durability of bond ing properties of an epoxy adhesive for structural strengthening application silane coupling agents were utilized to concrete surface s before the adhesive applied. In this work, the hypothesis was examined that the coupling agents form covalent bonding at the interface and improve durability by preventing the water from displacing the e po xy at the interface (Chapter 6 ). The specific aims of this work are summarized and shown as follows. (1) Examine and modify the mechanisms for the description of the complex hygrothermal behavior,

PAGE 17

17 i.e. anti plasticization in terms of increase in glass tran sition temperatures with hygrothermal exposure (Chapter s 3 and 4 ); (2) How each factor affecting complex hygrothermal behavior can be separately evaluated (Chapter s 3 and 4 ); (3) Investigate the effect of moisture/water addi tion on cure kinetics (Chapter 5 ); (4) Test the hypothesis that hydrogen bonding plays a large role for the adhesive bonding (Chapter 6 ); (5) Utilize silane coupling agents between the epoxy resins and concrete to improve the durability of the bond between epoxy and concrete under water/ moisture exposure (Chapter 6 ) We hope that this study lead us to advance our fundamental understanding of structure property relationships in epoxy systems under hygrothermal exposure, allowing us to improve the performance of epoxy systems in various appl ications under humid environments.

PAGE 18

18 CHAPTER 2 REVIEW OF LITERATURE E ffect of Water on Physical Properties of Epoxy Amine Thermosets Plasticization with H ygrothermal E xposure in E poxy S ystems In epoxy resins, property degradation due to hygrothermal exposu re is a critical issue, in which their nature to absorb moisture up to 7 8 wt% results in a significant loss in T g 16 26 modulus 27 31 tensile 28, 29, 31 and adhesive strength 32 For th ese detrimental effect s with hygrothermal exposure in epoxy systems the property loss is mostly attributed to the plasticization e ffect, whereas other mechanisms such as hydrolysis, polymer chain scission, formation of microcracks may also be factors depending on the materials and exposure condition. 25, 26, 28, 29, 31 Plasticization is general ly referred to the phenomenological behavior associated with the depression of glass transition temperature or the decrease of mechanical properties, which were caused by moisture sorption, in which water molecules act as a plasticizer or crazing agents, s ignificantly affecting the properties of the system under the changes in temperature and humidity. 27 In this detrimental effect, combined with the external factors of exposure conditions such as degree of humidity, exposed tempera tures and time, the chemical structure of the polymer matrix as well as processing also plays larger roles in that different degree of cure, network formation and hydrogen bonding sites induce different amount of water absorption. For the studies on the ef fect of water sorption on physical properties of epoxies, t he role of absorbed water associated with plasticization has been often described from two different viewpoints : (1) absorbed water as molecules forming an ordinary polymer

PAGE 19

19 diluent solution and ( 2) absorbed water bonded to specific polar groups in polymer network with interaction 20 For t he first viewpoint the structure of the epoxy is considered to be the pr imary factor to degrade the physical properties with water sorpt ion, generally describing that the absorbed The absorbed water molecules is described to simply reside in the free volume of the network without any interaction between the polymer and water, which enhances the chain mobility, resulting in a structural relaxation transiti on at lower temperatures 44 46 With this point of view, in order to predict the decrease in T g which is often s uggested as a parameter indicating the degree of plasticization quantitatively, the composition dependence of T g in polymer diluent systems have been employed via the free volume approach 47 The compositional model was derived by Kelley and Bueche, and then firstly developed by McKague et al. to describe the decrease in T g of epoxy systems 21, 47, 48 In this description, the degree of pla sticization of the polymer network structure was expected by the diluent volume fraction and the thermal expansion coefficients at the glass transition through the following expression (2 1) where subscripts p and d indicates diluent or water and polymer, and V is thermal expansion at glass transition and free volume, respectively. Ellis and Karas z reported the T g depression for a series of epoxies, in w hich o verall, they observe T g depressions of 8 15 C/wt% water for stoichiometric compositions, but only 4 5 C/wt% water for non stoichiometric compositions. 18 They

PAGE 20

20 also suggested the depression in stoichiometric compositions could be modeled with a compositional model for the T g depression Browning and Wright also used the compositional model with the free volume theory to relate their experiment al observation of the T g depression in epoxy resins. 49, 50 From another point of view, the absorbed water in polymer network bonds to specific polar groups, such as hydroxyl groups, disrupts interchain hydrogen bonds, also resu lting in a loss in properties 19, 51, 52 In this description it assumes a second mechanism of plasticization, and polar nature 53 This approach often employs the configurational entropy model rather than free volume as the temperature dependent function to propose an important role of the absorbed water, which localized at strongly polar molecular groups, sugges ting the formation of hydrogen bonding. 54 Moy et al. successfully applied this type of model to predict their T g depressions in a high T g epoxy system where other supplemental experiment also support ed their assumption that strong interaction between the dispersed water and some specific segments or groups in the polymer 22 Feng also found si gnificant decreases in T g due to absorbed moisture in which he found that the compositional approach could not predict the T g s of the plasticized materials. 55 A n entropy model was used, and this was explained as being caused by the decrease in entropy caused when water molecules become hydrogen bonded to the epoxy. Adamson also identifi ed thre e distinct stages of water sorption, and postulated that those stages correspond to water first entering the free volume, then water becoming bound to hydrogen bonding sites, and finally water entering the more densely

PAGE 21

21 c rosslinked phase. 56 In a series of papers, Soles et al. carefully examined the role of hydrogen bonding of water to the epoxy. 57 59 In some cases, both effects of ordinary polymer diluent solution and disruption of interchain hydrogen bonds are considered to describe property loss under hygrothermal exposure in the same system. 22, 29, 56, 60 Thus, the same amount of absorbed water could result in different degree of plasticization due to the multiplicity of the polymer water interactions in whic h absorbed moisture could be present in the network in different forms. However, i n any case, the absorbed water seems to undoubtedly act as a critical stimulus to degrade the properties of epoxy resins. In addition to the presence of absorbed water, the elevated temperature also contributes to the property loss in that the increased ra te of water diffusion results in a larger amount of water absorption, whereas chain mobility is also enhanced at higher temperatures. Anti P lasticization with H ygrothermal E xposure in E poxy S ystems Although the numerous reports have suggested that water ab sorption induces property loss in physical properties of epoxy systems as previously stated, some studies have revealed anomalous behavior, i.e., an increase in physical properties of epoxy systems after hygrothermal exposure, such as an increase in T g or tensile modulus. There have been many studies which report about some abnormal anti plasticization for various polymer systems such as polycarbonate dibutyl phthalate, polystyrene, poly (vinyl chloride), tricesyl phosphate starch and nylon water A nti pl asticization of a polymer has been conventionally designated as a polymer mixed with a low molar mass additive (plasticizers), which found to be harder and more brittle than

PAGE 22

22 the pure polymer. 61 Generally i n this observation of the behavior for various polymers, the T g of the system and its free volume decreases, but the mechanical properties altered, causing the polymer become stiff and brittle. 62 To describe such unusual behavior of anti plasticization which is opposite to that expected of plasticized materials various mechanisms and explanations have been suggested: suppression of the secondary relaxation transitions, polyme r diluent interaction s decreasing segmental mobility of the polymer chain, a chain end effect by a decrease of fractional free volume at the chain ends, re orientation of polymer chain which decrease free volume under high stress reduced mobility of the d iluents possible solid diluents with higher glass transition temperatures 62 68 In epoxy systems, there have been also some reports associated with anti plasticization behavior during hygrothermal exposure. Nogueir a et al. reported that in an epoxy system, the tensile modulus was slightly increased at certain amounts of water sorption, although it immediately decreased as water content increased. 29 They attribute these results to the reactivation of the post curing reaction. W u and coworkers also observed anomalous hygrothermal behavior for a nonstoichiometric epoxy system cured at room temperature, in which after 500 days of water immersion at 45C, the T g values and moduli of the specimens with a lower ratio of hardener incre ased 40 For an epoxy based primer, Devasahayam also found increases in T g up to 10C for exposure to 95% RH at 55 and 68C, whereas almost no change in T g was found for samples exposed at lower temperatures 36 Other authors have found similar result s, in which an increase of cross link density by the reactivation of the post curing reaction is generally attributed as the mechanism for this anomalous behavior. 37 39

PAGE 23

23 There have also been other descriptive hypothe ses for the anomalous increase in physical properties with hygrothermal exposure. Zhou and Lucas found initial stages of sharp T g depression followed by significant increases of T g in three different epoxy resins d uring hygrothermal exposure 43 Combined with their previous work investigating two types of bound water in epoxy resins through absorption/desorption profiles 42 a theory of different types of hydrogen bonded water was suggested to explain the anomalous behavior. In this theory, single hydrogen bonded water acts as a plasticizer, increasing chain mobility and contributing to property loss (Type 1), whereas water molecules interconnected with multiple sites of the resin network through hydrog en bonds create secondary cross linking, resulting in an increase in physical properties (Type 2). Papanicolaou et al. also employed this model to describe anomalous hygrothermal behavior in an epoxy resin. From dynamic mechanical thermal analysis (DMTA), they found increases in T g values with decreases in tan values during hygrothermal exposure, whereas storage modulus values showed almost no change 41 They explained this phenomena with the theory of different type s of hydrogen bonded water in that Type 1 bound water contributed to the decrease of tan by disrupting initial hydrogen bonds, whereas Type 2 water was responsible for the increase in T g It has also been suggested that the biphasic structure in epoxy systems contributes to this anomalous behavior 41 A ccording to this hypothesis, distribution of compression and expansion forces due to differences in water diffusion between hard and soft phases causes an anomalous increase in mechanical properties. The three suggested mechanisms above for the description of the anomalous hygrothermal exposure are summarized as follows: (1)

PAGE 24

24 post curing reactions caused by the elevated temperature of the water used for exposure; (2) different states of hydrogen bonded water molecules, one o f which creates secondary cross linking; and (3) effect of biphasic structure. It is important to note, however, that for none of these explanations is there any direct experimental evidence supporting them, and a clear explanation of the mechanism for thi s anomalous behavior remains unanswered. Effect of C rosslink D ensity on G lass T ransition in E poxy A mine T hermosets Cross link density is known to be one of the important factors to determine the T g in network polymers. An increase in the cross link density reduces the chain mobility resulting in an increase in T g whereas an increase in molecular weight as well as a decrease in low molecular weight components during the curing reaction also increases the T g 69 The re lationship between T g of a network polymer and conversion is well described else where. 70 72 For derivation of the relationship, a thermodynamic theory is used i n which the reduction in configurational entropy descri bes the relationship as cross linking increased as shown in following equation (2 2) where T g (X) is the T g of a polymer with a degree of cross l inking X, T g (0) is the T g of un cross linked linear polymer, and K is a dimensionless constant that is independent of the material. For another approach, Couchman and Karas z suggested the use of a thermodynamic theory in which changes in heat capacity of each component and its T g

PAGE 25

25 are correlated to describe the glass t ransition in copolymer systems as shown in following equation 70 (2 3) where C pi is the change in heat capacity of the component i at the glass transition temperature T gi and x i is the amount of the component i in solution. The features of the E quation 2 3 include the description of a partially cured polymer being a sol ution of the polymer in the monomer and the composition being related to the degree of conversion. The theory was then developed into the following semi empirical equation to describe T g in cross linked systems with the variation in conversion, assuming C p is inversely proportional to temperature where is conversion, T g0 and T g are the T g values of the monomer and the fully cured network, and C p0 and C p are the heat capacity changes at T g0 and T g respectively. (2 4) Using a similar approach assuming C p is constant, another semi empirical equation can be derived as shown below (2 5) Cure K inetics of E poxy A mine T hermosets Mechanism of E poxy C ure R eaction For the epoxy amine reaction, it is well know that the main features include (1) reaction of primary amine with an epoxide to form a secondary amine and (2) further reaction of the secondar y amine with another epoxide to form a tertiary amine. 73, 74

PAGE 26

26 Although other possib ilities can be also employed such as homo polymerization by epoxide epoxide and etherification by epoxide hydroxyl reactions, those o nly occur under certain conditions, such as in the presence of acid/base catalysts or at higher cure temperatures. 75 80 When homo polymerization and etherification reactions are neglected, the main cure mechanism c an be generally described by two reaction paths: (1) reactions catalyzed by compounds containing nitrogen (2) those by oxygen hydrogen bonds ( designated as catalytic ). 74, 81 85 For the mechanism in this cure reactions, it has been generally proposed that the reactions occur through a concerted termolecular intermediate consisting of epoxides, amines, and/or hydroxyls, although some reports sug gested other possibilities of more complexes can affect the reactions. 86, 87 In this mechanism, the ring opening of epoxide groups is catalyzed by hydrogen bond between the oxygen of the epoxide and the hydrogen of the donor. This protonation is followed by nucleophilic attack of the amine, in which epoxide ring is now easily open e d via a termolecular transition step. Subsequently, the reaction completed with a rapid displacement of the proton. 14 The reactions with the assu mption of the formation of termolecular intermediates can be displayed as shown in Figure 2 1 Since the epoxy amine reaction can be greatly affected by the catalyzed effect of the compounds which generate hydrogen bonds as the mechanism suggested, many s tudies also found that the addition of such compounds could significantly accelerate

PAGE 27

27 the reaction between epoxide and amines or other nucleophilic reagents, such as water, alcohols, phenols and imidazole. 10 13, 86, 8 8 Figure 2 1. Suggested cure mechanism for epoxide amine reaction involving the formation of termolecular intermediate s : (a) non catalytic reaction, (b) catalytic reaction 74 One of the unique features in epoxy amine reaction is the in situ formed hydroxyl groups during the reaction, resulting in typical autocatalytic behavior. Since the ring opening of epoxide groups was protonated by hydroxyl containing compound as suggested, the generated hydroxyl group by the reaction accelerates the entire reaction. 14

PAGE 28

28 Another critical feature of epoxy amine reaction is the changes of the reaction from chemically controlled in early stages to diffusion contro lled in later stage of curing as the cure progresses and a rigid three dimensional network is formed Thus, for the kinetic studies of epoxy amine reaction, only the early stage of reactions should be consider ed to obtain pure kinetic parameters while many other studies also suggested modified kinetic models to evaluate the diffusional effect of the later stage of curing reactions. Cure K inetic M odels for E poxy A mine C ure R eactions The epoxy amine systems are o ne of the oldest cross linked glassy polymers, but the kinetic study of epoxy amine reaction still remains an attractive field of investigation due to the increasing range of applications of epoxy based products 89, 90 To represent a possible kinetic scheme of the cure reaction in epoxy resins, the various kinetic models have been developed, in which the studies have generally followed two approaches: (1) phenomenological ; and (2) mechanistic modeling. 91, 92 In phenomenological or empirical modeling, it employs empirical or semi empirical kinetic relationships which tend to ignore the chemical detail of the reactions. In this method, the simple n th order kinetic expression has been utilized combined with another expression describing auto catalytic behavior for the epoxy amine reaction. The well known Kamal s model combined two kinetic expressions as shown below, and many studies have showed successful applications, in which t hey have used either original 93 96 or modified expressions 97, 98 (2 6 ) Although the studies employing empirical relationships for the reaction also provide valuable information for process design, the application should be confined to those

PAGE 29

29 conditions in that the approach has not been based upon the fundamental reaction mechanisms. 74 Also, the both expressions of n th order and autocatalytic behavior assume the existence of a single rate constant throughout the cure, which limited accuracy of the models. 91 On the other hand, the approach from mechanistic modelin g provides a more accurate method, which is analyzed from the balance of chemical species involved in the entire chemical reaction to form mathematical relations connecting the reaction rate path to cure time and temperature. 91 The mechanistic modeling provides advantageous method to understand the given kinetic reaction in terms that they can also be applied into the same systems, but employing different f ormulations or different temperature histories. For the construction of a kinetic model using this approach, the reaction scheme which represent s the cure mechanism should be established in accordance with theoretical and experimental evidences, which need to apply multiple parameters estimation for a non analytical expression of kinetics. 91 A number of kinetic models with this mechanistic modeling method h ave been established to describe the cure reaction of either the pure epoxy amine systems or the systems in the presence of other compounds such as hydroxyl containing accelerators. Xu et al. derived a kinetic model with three parameters to describe the cu re mechanisms of pure epoxide amine reactions assuming the non catalyzed reaction of secondary amines are much slow, thus ignored. 74 Mijovic et al. suggested a modified kinetic model with four parameters based upon their investigation that different types of hydrogen bonded transition complexes could influence the entire reaction in various epoxy amine systems. 87, 99, 100 Paz Aubin used a four parameter model, which consider

PAGE 30

30 the effect of free volume of polymer network on cure kinetics. 101 Ram an and Palmese suggested a two parameter model to describe the effect of added solvent on epoxy amine cure kinetics. 6 Although various kinetic models have been established through the mechanistic modeling method to show the kinetic relations in their own system, the generalized featu re can be expressed as descir bing the two principal paths of non catalytic reaction and catalytic reaction based on the existence of different transition complexes Thus, when the general feature of the two reaction is considered into the b asic epoxy amine system s the kinetic equations can be derived as follows, which take account of the catalyzed effect of compounds containing nitrogen hydrogen bonds (non catalytic) and those by oxygen hydrogen bonds (catalytic) (2 7 ) (2 8 ) where k 1 1 indicates that the non catalytic and catalytic kinetic parameters for epoxy primary amine addition while k 2 2 are the those for epoxy secondary amine addition, respectively. In this work, the general kinetic expression shown above has been modifie d to directly compare the effect of hydroxyl from water addition and that from auto acc eleration by the reaction. The detailed description is discussed in Chapter 4. Functional G roup A nalysis using FT IR S pectrometry In order to monitor the cure kinetics of the epoxy amine systems, various techniques have been utilized such as epoxide titrat ion, size exclusion chromatography,

PAGE 31

31 differential scanning calorimetry, nuclear magnetic resonance (NMR), and Fourier Transform infrared (FT IR) spectroscopy FT IR spectroscopy has been mostly used for this purpose, providing a con venient way to observe a quantitative amount of functional groups involved in the reaction with sufficient quality to allow differential analysis of the data 6, 79 In order to investigate of cure kinetics of epoxy amine thermoset, concentrations o f various functional groups involved in th is reaction should be revealed. To monitor the changes in the concentrations of primary amines and epoxides, near IR spectra directly provide quantitative profiles of concentrations. From typical NIR spectra for DGEBA Jeffamine system, the area of the peak around 4530 cm 1 provides the information about the concentration of epoxide while the area of the peak around 4930cm 1 presents the concentration about the primary amine. In addi tion, the phenyl peak around 4622 cm 1 gives the convenient method to normalize the each area of epoxide and primary amine, which used as an internal standard. 6, 74, 102, 103 On the other hand, for the concentration profiles of secondary and tertiary amines, mass balance equation shown below has been utilized 6, 79 (2 9 ) (2 10 ) where [E], [A 1 ], [A 2 ], [A 3 ], and [OH] are the concentrations of epoxide, primary amine, secondary amine, tertiary amine and hydroxyl groups present in the system, respectively. [OH] auto refers to the hydroxyl groups generated during the chemical reaction.

PAGE 32

32 Epoxy R esins for S tructural S trengthening A pplication Durability of Interfacial Bonding between E poxy and C eme n titious M aterials with H ygrothermal E xposure Epoxy resin as a seal coat and impregnating resin has been widely used for structural strengthening application such as a rehabilitation of concrete structures using e xternally bonded carbon fiber reinforced polymer (CFRP) composites providing high performances of strength, stiffness, and resistance to creep. 15, 104 Although epoxy resins provide the high qualities of tensile and adhesive properties for the structural strengthening the strengthened systems with epoxies has shown critical degradation of its performance with environmental exposure conditions such as moisture, temperature, and various solutions like chloride, alkali, and salt water 105 112 Many studies have revealed that degrada tion of interfacial bonding between epoxy resin and cemetitious materials with environmental exposure is the main reason for the loss of strength in th is structural application 106, 108, 112 These autho rs showed that environmental exposure induced adhesive failure which occurred between epoxy resin and concrete while the unexposed system displayed cohesive failure i.e. the failure occurred from a crack, which is initiated from the concrete specimen w hich supports the concept that interfacial bond degradation is most responsible for the loss of performance in this system. The studies of environmental effects on the durability of bond strength between cementitious materials and epoxy resins can be gener ally categorized into three groups, based upon the exposure environment into which a given system has been introduced: (1) Continuous exposure to pure water (immersion or high humidity); (2) Continuous exposure to various synthetic chemical solutions other than pure water such as alkali,

PAGE 33

33 chloride, salt water, and sea water; and (3) Cyclic exposure (wet/dry or freeze/thaw). Tables 2 1, 2 2, and 2 3 summarize results from the literature regarding the durability of bond strength of either CFRP bonded concrete/ mortar using epoxy adhesives or epoxy bonded concrete/mortar without the presence of carbon fibers. As shown in the results, it i s observed that either significant losses in bond strength or change in failure mode from cohesive to adhesive occurs with expo sure. Alt hough there are many studies reporting a decrease in bond strength between epoxy and cemetitious materials under various environmental exposure, few studies investigate methods to effectively maintain or enhance the durability of the bond strength under the exposure. Table 2 1. Effects of continuous water exposure on the durability of epoxy/concrete or FRP bond strengths Materials Exposure conditions Testing Results Pre cured CFRP laminate bonded with epoxy 100% RH, C, / 0 56 days Shear / peel tests Dry specimens failed cohesively. Wet specimens failed adhesively. 50 60% loss in fracture toughness 106 Concrete bonded with epoxy Immersion / 135 days Bond strength ( ASTM C882 ) 20 50% loss in bond strength 113 Concrete bonded with epoxy Immersion / 48 hrs + 48 hrs drying Bond strength ( ASTM C882 ) 17 35% loss in bond strength 105 CFRP wet layup on mortar Immersion at ambient temp / 60 days 4 point flexure 24 33% loss in strength 109 CFRP wet layup on mortar Immersion at ambient temp. / 60 days Peel test 10 20% decrease in G IIc Fracture mode changes from primarily cohesive to adhesive. Almost no change in G Ic 110 CFRP wet layup on concrete Immersion / 3 8 weeks Modified double cantilever beam Dry specimens failed cohesively Wet specimens failed adhesively. 35 75% loss in interfacial fracture toughness 112

PAGE 34

34 Table 2 2. Effect of continuous exposure to synthetic solutions and sea water on the durability of epoxy/concrete or FRP bond strengths Materials / Fluid Exposure conditions Testing Results Concrete bonded with epoxy / Seawater RT / 6, 12, 18 months Split tensile slant shear 25% decrease in bond strength a fter 18 months. Failure mode changes from cohesive to adhesive with aging time 108 Concrete bonded with e poxy / Seawater C / Up to 18 months Split tensile, slant shear, flexure, compression No apparent difference in strengths 114 CFRP wet layup on m ortar / Synthetic seawater Ambient temp / 60 days 4 point flexure 13 26% loss in strength 109 CFRP wet layup on mortar / Synthetic seawater Ambient temp / 60 days Peel test 25 30% decrease in G IIc Almost no change s in G Ic 110 Table 2 3. Effects of cyclic exposure on the durability of epoxy/concrete or FRP bond strength s Materials /Fluid Exposure conditions Testing Results CFRP or GFRP wet layup / Simulated seawater 4hrs immersion / 2hrs at 35 C, 90%RH (75 days) 4 point flexure 7 33% loss in flexural strength 111 CFRP wet layup, one steel bar in concrete / Calcium chloride solution 16 hrs at C / 8 hrs at RT ; or 16 hrs immersion / 8 hrs dry ( 50 / 100 days ) 4 point flexure 18 21% loss in strength for fre eze thaw, 10 19% loss in strength for immersion 107 Concrete bonded with epoxy / Seawater Immersed during high tide only ( 6, 12, 18 months ) Split tensile slant shear 25% decrease in bond strength after 18 months. Specimens become protected by build up of shells. Failure mode changes from cohesive to adhesive with aging time 108 CFRP wet layup / Salt water outdoor exposure 17 months salt water at 10~60 C / 6 months outdoor ; 23 months outdoor Torsion t ension ( metal disk / CFRP ) 0 55% loss in bond strength for salt water exposure; 0 45% loss in bond strength for outdoor exposure 115 CFRP wet layup on mortar / Freeze thaw 15.5C for 24 hrs / 20C for 24 hrs (60 days) 4 point Flexure 0 15% loss in strength 109 CFRP or GFRP wet layup on mortar / Freeze thaw 15.5C for 24 hrs / 20C for 24 hrs (60 days) Peel test 0 30% increase in both G Ic and G IIc 110

PAGE 35

35 Adhesion Mechanism of E poxy B onding with C ementitious M aterials For the general description of the adhesion between various types of materials, several categories of mechanisms have been propos ed : m echanical i nterlocking, c hemical bonding, e lectrostatic adhesion, i nter diffusion, and a dsorption. 116 120 In the case of the structural strengthening application which employs an adhesi ve bonding between organic adhesives and cemetitious materials, the theories of mechanical interlocking and chemical bonding can be generally accepted as operative mechanisms. 121 The mechanical interlocking theory suggest s that good adhe sion between two materials occurs when an adhesive penetrates into the pores, holes and crevices and other irregularities of the adhered surface of a substrate, and locks mechanically to the substrate. 110, 116 118, 1 20 Lin et al have shown that, in their strengthening application of epoxy resins, more rough and deep surface topography of concrete surface before bonding, attained by water jet or blast sanding, enhanced the bond strength of the system. 122 M omayez e t al. also suggested that rough surface preparation of concrete lead to higher bond strength through their pull off and slant shear testing method. 123 Although the mechanical interlocking seems to promote the bonding strength between two materials, when considering that good adhesion is also accompanied with smooth adherened as well, there should be another type of mechanism which strongly affects the bonding. For another type of the adhesion mechanism to describe the bonding between epox y adhesives and cementitious materials, the mechanism of chemical bonding is taken into account Based upon the chemical nature of both materials in which the concrete surface contains silanol groups and epoxy adhesives contains hydroxyl

PAGE 36

36 groups, it is high ly probable that hydrogen bonding plays a larger role for the bonding. A few studies that examine the chemical nature of the interface also suggest that hydrogen bonds are formed between epoxy and concrete, and that water tends to cluster at the interface and form hydrogen bonds 124 126 W hen considering similar mechanism proposed for the description used for the degradation in epoxy metal adhesion under humid environments in which a decrease in bond strength have be en ascribed to displacement of epoxy with water, which preferentially forms hydrogen bonds with the adherend, the hydrogen bonding at the epoxy concrete interface could also be considered as a factor to affect the durability in the bonding of epoxy cememti tous materials 34, 35 Silane C oupling A gents to E nhance the D urability of A dhesive S trength durin g H ygrothermal E xposure To enhance the durability of bond strength between various organic and inorganic materials, silane coupling agents has been extensively attempted, in which the coupling agents connect two materials through covalent bonds which are m ore durable against water attack than hydrogen bonds. Adhesion enhancement is also attributed to the blocking of lateral diffusion of water by barriers due to the presence of the hydrophobic silane. 127 In addition, capture of the excess moisture coming from the surrounding environment occurs by the hydrolysis reactions of active alkoxy groups in modified resins, by which the free state water absorbed can be consumed. 128 Silane coupling agents are silicon based chemicals that contain two types of reactants inorganic and organic in the same molecule. A typical general structure is (RO) 3 Si(CH) n X, where RO is a hydroly z able group such as methoxy, ethoxy, and X is

PAGE 37

37 an organofunctional group such as amino, epoxy. A silane coupling agent acts at an interface between an inorganic substrate such as glasses, metals or minerals and organic materials to bond two incomp atible materials. 129 Figure 2 2 Reaction p rocess of Alkoxy Silane, R is organofunctional group in t his figure. 129 There are four steps for the silane coupling reactions (Figure 2 2 ). First, the alkoxy (RO) groups are hydrolyzed. Second, after the hydrolysis of the first and second alkoxy groups, condensation to oligomers occurs. Third, the hydroxyl group is hydrogen bonded with hydroxyl sites on the substrate after orientation of third methoxy group on hydrolysis. Finally, a covalent bond is formed with the substrate during drying or curing with water liberated. At the interface, there is usually only one bond from each silicon of the organosilane to the substrate surface. A moderate cure cycle wil l leave silanol groups remaining in free form and, along with any silane organofunctionality, may bond with the subsequent topcoat, forming an interpenetrating polymer network and providing improved adhesion. 129

PAGE 38

38 Silane coupling agents as a primer are widely used for improving adhesion between organic materials and inorganic materials. Many papers have repo rted enhanced durability of bond strength between metals/ceramics and epoxy resins under humid environments. 130 132 Lin Ye et al reported the use of a silane coupling agent to improve the adhesion bonding between CF RP and concrete using epoxy adhesive. 122 3 G lycido xypropyltrimethoxysilane o ne of the epoxy functional silane coupling agents was employed to the system, and its use was not very helpful for the mechanical properties of the system. The fractu re energy of crack growth was slightly increased by the silane coupling agent. However, the system was not under environmental exposure such as water and temperature before the fracture testing. If the system had been under water/moisture exposure, it is expected that the system would have demonstrated improved durability of mechanical properties by the effects of the silane coupling agent which prevents the water from displacing the epoxy at the interface.

PAGE 39

39 CHAPTER 3 COMPLEX HYGROTHERMAL EFFECTS ON GLASS TRANSITION OF EPOXY AMINE THERMOSETS: A STUDY OF A MODEL EPOXY AM INE Background The glass transition temperature ( T g ) of a polymer is the temperature at which dramatic changes in properties occur, particularly reduction in strength, stiffness, and resistance to creep. Thus, there is a desire among structural engineers a nd designers to use the T g as a materials specification in codes and standards. However, for thermosetting materials used as adhesives and as a component of FRP, such as epoxies and vinyl esters, the glass transition is not a single value. In these materia ls, the glass transition can vary with factors such as degree of cure and extent of water absorption. These factors are exacerbated by the fact that in civil engineering applications, the material is cured under ambient conditions, and thus is typically no t fully cured. Thus, an understanding of how these factors affect the glass transition is of particular importance for defining upper temperature use limits in these applications. Water absorption into various types of polymer systems has been extensively studied including amorphous 133 136 semi crystalline 137 139 cross linked polymers 18, 21 and copolymers 140, 141 because of the general tendency of water absorption to induce a loss in physical properties such as a decrease in T g Plasticization by absorbed water has been generally regarded as the major factor resulting in this property loss. As previously discussed in Chapter 2, although property degradation due to hygrothermal exposure is also a critical issue in epoxy resins, many studies have reported anomalous behavior, i.e., an increase in physical properties of epoxy systems after hygro thermal exposure, such as an increase in T g or tensile modulus. There has been mechanisms suggested to describe th is abnormal behavior, such as post curing

PAGE 40

40 reactions caused by the elevated temperature of the water used for exposure; different states of hyd rogen bonded water molecules, one of which creates secondary cross linking; and effect of biphasic structure. It is important to note, however, that for none of these explanations is there any direct experimental evidence supporting them, and a clear expla nation of the mechanism for this anomalous behavior remains unanswered. In this study, we have also observed similar anomalous increases in T g with hygrothermal exposure. We focus on direct experimental measurements that allow us to specifically describe t he mechanism for this behavior E xperimental Section Materials The material used in this work is a two part epoxy, diglycidyl ether of bisphenol A (DGEBA) and poly(oxypropylene) diamine (POPDA, Jeffamine D 230), purchased from Huntsman. Figure 3 1 illustrat es the chemical structures of DGEBA and POPDA. The mass ratio of DGEBA to POPDA was 100 to 32.9, corresponding to stoichiometric equivalence between functional groups. The two liquid components were mixed vigorously for 10 min to ensure even mixing. The mi xed material was then degassed for 30 mins under vacuum to remove air bubbles. Specimens were cured at room temperature (RT, 22 23C) for 4 weeks before the exposure was started, ensuring enough time for the curing reaction at RT. Exposure environments con sisted of immersion into clean tap water at temperatures of 30, 40, 50, and 60C. Details of the cure and exposure conditions for hygrothermal exposure are shown in Table 3 1. To construct a master plot of T g versus cross link density for the unexposed epo xy, we employed various cure conditions to obtain specimens with a wide range of

PAGE 41

41 conversion values. The specific information for cure conditions for these specimens is shown in Table 3 2. Figure 3 1. Chemical structures of diglycidyl ether of bisphenol A (DGEBA) and poly(oxypropylene)diamine (POPDA). Table 3 1. Cure and exposure conditions for hygrothermal exposure. Type of Experiment Cure c ondition Exposure Condition Monitoring of changes in T g using DSC RT/28days Water immersion at 30, 40 50, and 60C / 1, 2, 4, 7, 14 and 28 days Monitoring of changes in crosslink density using FT IR Table 3 2. Cure conditions for the master plot of T g vs. crosslink density for unexposed epoxy. Cure condition T g ( C) Epoxide conversion RT/ 0~28days 56.2~46.8 0.00~0.8 6 80 C/110min 54.4 0.9 0 90 C/80min 61.1 0.9 3 90 C/110min 69.0 0.97 50 C/60min + 80 C/120min + 125C/180min 76.7 1.00

PAGE 42

42 Techniques Differential scanning calorimetry (DSC; DSC 220, Seiko instruments) was utilized to identify c hanges in T g with hygrothermal exposure. Samples were prepared by depositing a small amount of the epoxy formulation directly into the bottom of a DSC aluminum pan before the system was cured in order to maximize the thermal contact between the sample and the pan and avoid the artifacts often observed during first DSC scans when there is poor thermal contact. After the sample was cured in the bottom of the pan either the pan with the sample in it was immersed in water for hygrothermal exposure, or the pan w as immediately used for T g measurement. Following hygrothermal exposure, the surface of the sample was wiped carefully with a paper tissue to remove any excess surface water. The pan was immediately sealed and placed directly into the DSC chamber for the m easurement. To reduce the potential for any cure occurring during the measurement, we determined the T g from the first run of each specimen. After placing the sample in the DSC, the heat flow signal was allowed to fully equilibrate before beginning the tem perature scan. The temperature range for the measurement was 30 to 130C at a heating rate of 10C/min, and three replicas were fabricated for each exposure condition to obtain the average value. The T g was chosen from the midpoint of the tangent between the extrapolated baselines before and after the transition. To measure cross link density and the amount of absorbed water, we used Fourier transform infrared spectrometry (FTIR; Nicolet Magna 760, Thermo Electron Cooperation) with a CaF beamsplitter and a n MCT detector. Near infrared spectra were recorded over the range of 3800 7400 cm 1 using 32 scans at a resolution of 4 cm 1 Thin films with 0.5 mm thickness were cast between glass plates using Teflon spacers.

PAGE 43

43 After the selected cure condition as shown in Tables 3 1 and 3 2, the films were either immersed in water for hygrothermal exposure or directly used for the FTIR measurement. Figure 3 2. Typical NIR spectra for unexposed systems with different cure conditions: Cured at RT for 28 days (top) and c ured at 50C for 1 h followed by 80C for 2 h and 125C for 3 h (bottom). Typical NIR spectra for stoichiometric systems without exposure are shown in Figure 3 2. The reduction in the epoxide absorbance at 4530 cm 1 for the higher temperature cure is evid ent in this figure. For the quantitative analysis of cross link density changes, the area of the epoxide peak around 4530 cm 1 normalized to the phenyl peak around 4622 cm 1 was calculated for each exposure condition. The epoxy conversion was then calculat ed by = 1 A ( t )/ A (0), where A ( t )/ A (0) is the ratio of normalized peak area with respect to the uncured and unexposed system. For quantitative analysis in accordance with the Beer Lambert law, the absorbance was less than 1.0. The baselines for area calculatio n of those peaks were chosen as suggested

PAGE 44

44 by Dannernberg with out any baseline correction 1 03 In Figure 3 2, there is a small unknown peak at around 4560 cm 1 that overlaps the epoxide peak. The area of this peak as measured on a stoichiometric, fully cured sample was subtracted from the total area of the peaks in that region in order to det ermine the absorbance due only to epoxide groups. Figure 3 3. Experimental confirmation of reliability of quantitative analyses from FTIR : (a) Plot of normalized area assigned to epoxide at 4530 cm 1 with variation in epoxide/amine ratio. The corre lation coefficient value, R 2 is 0.99. (b) Normalized area for characteristic water peak around 5230 cm 1 for various unexposed and exposed systems.

PAGE 45

45 Although the reliability of this quantitative analysis of the epoxy cure reaction was confir med by previous workers 102, 103 we reconfirmed it for our system. To demonstrate the reliability, normalized peak areas in the spectra were measured with respect to the epoxide concentration. In this experiment, specimens with different epoxide concentrations were cured at high temperature for enough time to ensure complete reaction between epoxide and amine functional groups. The cure condition was 50C for 1 h our followed by 80C for 2 h ours and 125C for 3 h ours Formul ations were smeared onto KBr IR disposable cards for the samples with pure DGEBA and those with a lower amine ratio because they did not solidify after the cure. The other samples were prepared by casting thin films with a 0.5 mm thickness using glass plat es and Teflon spacers. As shown in Figure 3 3 ( a ) the normalized absorbances decrease linearly as epoxide concentration decreases, which indicates that our analysis is reliable and obeys the Beer Lambert law. To estimate the amount of absorbed water for each exposure condition, the normalized area of the characteristic water peak around 5230 cm 1 was obtained, which has been assigned to a combination of asymmetric stretching ( as) and in plane deformation ( ) of water 142 144 The normalized area of the characteristic water peak is known to be a way to measure the amount of absorbed water in various polymer water systems, where the normal ized area is proportional to the real amount of absorbed water 142 145 In this work, the region from 4972 to 5342 cm 1 was taken to measure the area and normalized by the phenyl absorbance at 4622 cm 1 Because thes e are normalized values we can not use them to determine absolute concentrations of water present; however, because the normalized areas are proportional to the amount of

PAGE 46

46 water they can be used for relative comparisons. It is also important to note that th is peak includes hydroxyl groups generated in the epoxy network due to the cure reaction. To check the contribution of these hydroxyl groups to the total absorbance, we made measurements for both exposed and unexposed samples over a range of conversions. A s shown in Figure 3 3 ( b ) for the range of epoxide conversion between 0.86 and 1.00, the absorbance due to the hydroxyl groups is approximately constant, indicating that the change in absorbance of this peak is due to the change in amount of absorbed wate r. R esults Figure 3 4. Typical DSC curves after water immersion for 2 days at different temperatures compared with no exposure. Our results, shown in Figure 3 4, clearly indicate that defined as an increase in glass transition afte r exposure to water, occurs. As shown in Figure 3 4, exposure to 40, 50, and 60C water for 2 days results in significant increases in T g as compared to the control specimen that was measured before the exposure was started. On the contrary, when the speci mens were exposed to 30C water for 2 days, they showed a lower T g than the control.

PAGE 47

47 Figure 3 5. Changes in DSC curves as a function of immersion time for each exposure condition. From top of each set of DSC scans: Exposed for 1, 2, 4, 7, 14, and 28 days. Exposed in (a) 30, (b) 40, (c) 50, and (d) 60C water. From Figure 3 5 showing the entire DSC scans for each exposed system, it is shown that endothermic aging peaks are superimposed around the glass transition region. Amorphous polymers, which are cooled down to below T g or isothermally cured at a temperature below T g typically show enthalpy relaxation upon reheating, which is observed as an endothermic peak in the DSC scan. This behavior is well known as a structural enthalpy relaxation or physic al aging, and is a result of the recovery of enthalpy trapped in the glassy state of t he material during aging 146 151 Figure 3 5 shows that the intensity of the endothermic peaks rapidly decreased upon exposure at 60C,

PAGE 48

48 which is attributed to the fact that the exposure temperature is above T g of the system, leading the system closer to thermodynamic equilibrium as exposure continues 147, 149 Figure 3 6. Changes in T g after water immersion at different temperatures. Samples were cured for 4 weeks at room temperature in air before exposure was started. In this picture, the size of the error bars displaying the standard deviation is too small to be shown at some points. To fo llow the changes in T g for each exposure condition, we measured T g as a function of time for each exposure condition. As shown in Figure 3 6, samples exposed at 40 and 50C show continuous increases in T g with exposure until the rates of increase slow down and approach saturation points. Exposures at 30 and 60C show some interesting results that are contrary to each other. For samples exposed at 30C, an initial decrease in T g was followed by an increase, whereas samples exposed at 60C showed an increase in T g at the beginning stage, but a slight decrease in T g at the later stage of exposure.

PAGE 49

49 Figure 3 7. Typical Near Infrared spectra and changes in conversion for samples with different exposure conditions: (a) Near infrared spectra for samples immer sed in water for 1 day at different temperatures compared with an unexposed control sample. Cure condition was 4 weeks at room temperature in air before the exposure was started. (b) Changes in conversion at different exposure conditions after curing 4 wee ks at room temperature in air. conversion were investigated using NIR spectroscopy. Typical NIR spectra are displayed in Figure 3 7 ( a ) in which samples exposed to water immersion fo r 1 day at different temperatures are compared with control samples which had been exposed to

PAGE 50

50 air for 1 day. The cure condition was the same as used for measuring changes in T g namely, 4 weeks at room temperature in air before the exposure was started. As shown in Figure 3 7 ( a ) the area of the epoxide peak around 4530 cm 1 decreased after hygrothermal exposure compared with the control, indicating a post curing reaction occurred during the exposure that resulted in an increase in conversion. To follow th e changes in conversion for each exposure condition, we measured conversion as a function of time for each exposure condition. Figure 3 7 ( b ) shows that the epoxide conversion increased for all the exposure conditions, resulting in significant increases in cross link density compared with the unexposed sample. This result shows that the elevated temperature of the water caused additional cure, even at a low temperature such as 30C. When this figure is compared to the changes in T g with hygrothermal exposur e in Figure 3 6, it is obvious that the increase in cross increase in T g with hygrothermal exposure. Careful comparison of Figures 3 6 and 3 7 ( b ) also shows that the hygrothermal behavior is complex. For example, the samples exposed at 30C showed an initial decrease in T g and samples exposed at 60C showed a decrease in T g at the latter stages of exposure, even though the conversion is continuously increasing with exposure. Thus, we can conclude that plasticization a lso plays a significant role in determining the properties of this system. D iscussion For the epoxy system being examined in this study, hygrothermal exposure induces complex behavior in which an increase in glass transition due to increasing cure simultan eously occurs with a decrease in glass transition due to plasticization. From the comparison of the two figures showing changes in T g and conversion with

PAGE 51

51 hygrothermal exposure (Figures 3 6 and 3 7 ( b ) ), we can see that the increase in T g due to additional cure and the decrease in T g due to plasticization are in competition with each other during exposure. When observing the initial stage of the sample exposure between 0 and 1 day at 30C, plasticization has the predominant effect on T g compared with additi onal cure, causing a decrease in T g despite an increase in cross link density. However, as exposure continues, the additional cure becomes the dominating factor, resulting in an increase in T g On the contrary, for exposure at 60C, a continuous increase i n T g at the initial stage of the exposure indicates that additional cure dominates the change in T g The effect of plasticization becomes more important at the later stage of exposure, resulting in a slight decrease in T g With regards to exposure at 40 an d 50C, it is observed that the increase in cross link density is dominant throughout the exposure, resulting in a monotonic increase in T g When comparing samples at 50 and 60C, the T g for 50C samples is larger than T g for 60C samples after 8 days expo sure even though both had already reached 100% conversion, which indicates that at 60C there is more absorbed water, resulting in greater plasticization. As described previously, two other factors that have been hypothesized to cause vior are creation of a hydrogen bonded network of water and the influence of the epoxy microstructure. To determine the relative importance of these effects, T g for the unexposed system was plotted as a function of conversion. This e used to exclude the influence of crosslink density by plotting the results from Figures 3 6 and 3 7 ( b ) onto

PAGE 52

52 to compare the T g values directly between exposed and unexposed systems at the same conversion value. Figure 3 8. T g experimentally obtained and solid line s are fits to theoretical : (a) Equation 3 1 and (b) Equation 3 2. Cross link density is known to be one of the important factors to determine the T g in network polymers. An increase in the cross link density reduces the chain mobility resulting in an increase in T g whereas an increase in molecular weight as well as a decrease in low molecular weight components during the curing react ion also increases

PAGE 53

53 the T g 69 The relationship between T g of a network polymer and conversion is well described elsewhere. 70 72 For derivation of the relationship, a thermodyn amic theory is used in which changes in heat capacity of each component and its T g are correlated to describe the glass transition in copolymer systems. The theory was then developed into the following semi empirical equation to describe T g in cross linked systems with the variation in conversion, assuming C p is inversely proportional to temperature (3 1) where is conversion, T g0 and T are the T g values of the monomer and the fully cured network, and C p0 and C are the heat capacity changes at T g0 and T respectively. Using a similar approach assuming C p is constant, another semi empirical equation can be derived as shown below (3 2) To construct the master plot of T g as a function of conversion, T g and conversion values under identical cure conditions were experimentally obtained for the unexposed system. Equations 3 1 and 3 2 were u sed for fitting the experimentally obtained data as shown in Figure 3 8. Both equations have good agreement to the experimental data, with coefficients of determination of 0.99 for both cases. The values of C / C p0 as a fitting parameter were 0.69 and 0.88 for E q uation s 3 1 and 3 2, respectively. Finally, the values of T g and conversion for exposed systems, shown in Figures 3 6 and 3 7 ( b ) were applied to the master plots using E q uation 3 2 for the fitting As shown in Figure 3 9, all the T g points for the exposed systems are below the line, showing the effect of plasticization independent of crosslink density. This result also

PAGE 54

54 suggests that the contribution of as different types of hydrogen bonded water and the influence of microstructure are small compared to plasticization. Figure 3 9. Direct comparison of T g between unexposed and exposed systems using the master plot of T g versus conversion. The section marked with dotted lines in (a) is magnified in (b).

PAGE 55

55 To further evaluate the plasticization effect quantitatively, we determined T g the difference in T g between unexposed and exposed samples at the same conversion value. Glass transition temper atures for the exposed system were determined directly from DSC measurements, whereas T g values at the corresponding conversion for the unexposed system were obtained from the values on the fitted curve as indicated in Figure 3 9 ( b ) Finally, to correlate T g with the absorbed water amount at each exposure condition, the relative amount of water uptake was determined from the characteristic water peak around 5230 cm 1 which was well resolved in the NIR spectra of this system. Figure 3 10. Relative amoun t of water absorption with exposure time estimated by normalized area of the characteristic water peak from NIR. When the relative amount of absorbed water is evaluated by this method, it shows the typical features of water uptake with exposure time in wh ich water amount increases as exposure continues, ultimately approaching some saturation point, as shown in

PAGE 56

56 Figure 3 10. The rate of increase in absorbed water at the initial stage of the exposure is enhanced as exposure temperature increases, which is als o a typical feature in that water absorption of the epoxy system is accelerated with an increase in diffusion rate as temperature increases. Figure 3 11. T g differences between unexposed and exposed systems as a function of estimated amount of absorbed water. Finally, T g was plotted as a function of the relative absorbed water amount for each exposure condition (Figure 3 11). For samples exposed at 30 and 40C, T g increases with water amount, showing the expected behavior of increase in plasticization as the amount of absorbed water increases. However, for the 50 and 60 C samples, some anomalous behavior is observed in which T g changes independently of water content. This anomalous behavior possibly indicates that there are other factors affecting h as secondary types of hydrogen bonded water and microstructure, although these effects are very small compared with plasticization.

PAGE 57

57 Summary For the anomalous hygrothermal behavior of an epoxyamine thermoset, in which T g increased with exposure to water i mmersion at different temperatures, FT NIR spectroscopic studies demonstrated that increases in cross link density due to additional cure during the elevated temperature exposure led to an increase in properties. However, from the feature that the T g decre ased at certain stages of the exposure, it is evident that plasticization by water occurred simultaneously, rendering the hygrothermal behavior to be more complex than originally hypothesized. In this in crease in cross link density and occurred simultaneously. Constructing the plot of T g versus conversion for the unexposed system provided an excellent method to consider contributions of other factors and exclude the facto r of crosslink density. By applying the results for the exposed it is possible to directly compare the T g values between exposed and unexposed samples while ruling out the factor of cross link density. The result indicates t hat the effect of other factors, such as different types of hydrogen bonded water and influence of the microstructure, are very small compared to the plasticization effect. To evaluate the plasticization effect quantitatively, we measured T g the difference of T g between unexposed and exposed samples at the same conversion value, as a function of absorbed water amount. It is observed that T g increases with absorbed water amount at lower exposure temperatures, whereas T g changes independe ntly of water amount at higher exposure temperatures. Thus, factors other than additional cure and plasticization may affect T g at these higher exposure temperatures. Examination of these other factors such as changes in the relative

PAGE 58

58 amounts of the hard an d soft phases and different states of hydrogen bonded water molecules are currently ongoing in order to provide an in depth understanding of the anomalous behavior in this system.

PAGE 59

59 CHAPTER 4 COMPLEX HYGROTHERMAL EFFECTS ON GLASS TRANSITION OF EPOXY AMINE T HERMOSETS: A STUDY OF COMMERCIAL PRODUCTS Background In order to take advantage of easy installation in outdoor or electronic applications, many epoxy amine adhesives and coatings are cured under ambient conditions, and thus are typically not fully cured. For these systems, changes in the environments such as an increase in temperatures or exposure to water can cause additional cure resulting in an increase in physical property while properties can decrease simultaneously due to plasticization, rendering th is hygrothermal behavior to be complex. For such systems with complex hygrothermal behaviors, it is complicated to evaluate their own effect of degradation by environments in that the materials properties are simultaneously influenced by two combinational effects. From the previous study (Chapter 3) a tool was successfully proposed to evaluate those two factors separately with a model epoxy amine system. 152 When the glass transition temperatures (T g s) are measured to describe the current states of materials specifications for the given system, constructing the plot of T g versus conversion for the unexposed system provided an excellent method to excl ude the factor of cross link possible to directly compare the T g values between exposed and unexposed samples while ruling out the factor of cross link density. In this study, we have also found complex hygrothermal behaviors of two commercial epoxy products, both used as a seal coat and impregnating resin for structural strengthening applications, where T g has been changed in a complicated manner due to competing effects of both additional cure and plasticization. The study

PAGE 60

60 has focused on how to quantify those two effects separately with the suggested method, using a thermodynamic analysis and spectroscopic measurements. Experimental S ection Materials Both of the epoxy ad hesive systems used in this work were for the structural strengthening application as a seal coat and impregnating resin, purchased from Sika Corporation and BASF, respectively. Each system consists of two liquid components, where part A and part B refers to epoxy resin component and hardener component, respectively. In system I part A consists of diglycidyl ether of bisphenol A (DGEBA) and Part B consists of blend of amines. In system II part A consists of epoxy resin, alkyl glycidyl ether, aliphatic dig lycidyl ether, and ethyl benzene, while part B consists of polyoxypropylene diamine, Isophorone diamine, eposy resin, benzyl alcohol, suggested by the manufacturers, the mix ratio between part A and part B was 1 to 0.345 and 1 to 0.3 by weight for the systems I and II respectively. For the sample preparation, the two liquid components were mixed vigorously for 5 mins to ensure even mixing. The mixed material was then degassed for 30~60 mins under vacuum to remove air bubbles. Specimens were cured at room temperature (RT, 22 23C) for 4 weeks prior to exposure, ensuring enough time for the curing reaction at RT. Table 4 1 shows the basic thermal properties of systems I and II witho ut any exposure to water immersion. The corresponding values of T g s and conversion were measured by differential scanning calorimetry and Fourier transform infrared spectrometry (techniques of measurements are described in the following section). As

PAGE 61

61 shown in Table 4 1, T g values between two systems are very close for various cure condition, which indicates the similarity in thermal properties between both of the unexposed systems. Exposure environments consisted of immersion into de ionized water at tempera tures of 30, 40, 50, and 60C. Details of the cure and exposure conditions for hygrothermal exposure are shown in Table 4 2. Fig ure 4 1. Comparison of t ypical NIR spectra for each system with different cure conditions: System I c ured at 50C for 1 h foll owed by 80C for 2 h and 125C for 3 h System I c ured at RT for 28 days System II c ured at 50C for 1 h followed by 80C for 2 h and 125C for 3 h System II c ured at RT for 28 days (from bottom to top). Table 4 1. Comparison of observed T g s and epoxide conversion for unexposed systems between systems I and II Cure Condition System I System II T g (C) Conversion T g (C) Conversion Right after mixing 45.9 0.00 48.6 a 0.00 28days at RT 52.5 0.80 53.0 0.88 Fully Cured 82.1 1.00 74.2 1.00 a Estimated from a T g conversion construction b Cured at 50 C for 60min s followed by for 120min s followed by for 180min s

PAGE 62

62 Table 4 2 Cure and exposure con ditions for hygrothermal exposure in monitoring the changes in T g s using DSC and changes in cross link density using FT IR Type of Experimental Work Condition Temp. ( C) Time (days) Cure Air RT 28 Hygrothermal Exposure Water Immersion 30/40/50/60 1/2/4/ 7/14/28 To construct a master plot of T g versus cross link density for each system of the unexposed epoxies, we employed various cure conditions to obtain specimens with a wide range of conversion values. The specific information for cure conditions for these specimens is shown in Table 4 3. Table 4 3 Cure conditions for the master plot of T g vs. conversion for unexposed systems I and II Cure Condition System I System II T g ( C) Epoxide C onversion T g ( C) Epoxide C onversion RT/0~28days 45.9~52.5 0. 00~0.80 30.7~53.0 0.23~0.88 80 C/110min 59.7 0.89 56.6 0.90 90 C/80min 65.5 0.91 66.3 0.95 90 C/110min 76.9 0.96 50 C/60min + 82.1 1.00 74.2 1.00 Techniques Differential scanning calorimetry (DSC; DSC 220, Seiko instruments) was utilized to identify changes in T g with hygrothermal exposure. Samples were prepared by depositing a small amount of the epoxy formulation directly into the bottom of a DSC aluminum pan before the system was cured in order to maximize the thermal contact

PAGE 63

63 between the sample and the pan and avoid the artifacts often observed during first DSC scans when there is poor thermal contact. After the sample was cured in the bottom of the pan either the pan with the sample in it was immersed in water for hygrothermal exposure, or the pan was immediately used for T g measurement. Following hygrothermal exposure, the surface of the sample was wiped carefully with a paper tissue to remove any excess surface water. The pan was immediately sealed and placed directly into the DSC chamber for measurement. To reduce the potential for any cure occurring during the measurement, the T g was determined from the first run of each specimen. After placing the sample in the DSC, the heat flow signal was allowed to fully equilibrate before beginning the temperature scan. The temperature range for the measurement was 20 to 120~130C at a heat ing rate of 10C/min, and three replicas were fabricated for each exposure condition to obtain the average value. The T g was chosen from the midpoint of the tangent between the extrapolated baselines before and after the transition. To measure cross link d ensity and the amount of absorbed water, we used Fourier transform infrared spectrometry (FTIR; Nicolet Magna 760, Thermo Electron Cooperation) with a CaF2 beamsplitter and an MCT detector. Near infrared spectra were recorded over the range of 3800 7400 cm 1 using 32 scans at a resolution of 4 cm 1 Thin films with 0.5 mm thickness were cast between glass plates using Teflon spacers. After the selected cure condition as shown in Tables 4 2 and 4 3, the films were either immersed in water for hygrothermal e xposure or directly used for the FTIR measurement.

PAGE 64

64 Typical NIR spectra for each system without exposure are shown in Figure 4 1. The reduction in the epoxide absorbance at 4530 cm 1 for the higher temperature cure is evident in this figure. For the quantit ative analysis of cross link density changes, the area of the epoxide peak around 4530 cm 1 normalized to the phenyl peak around 4622 cm 1 was calculated for each system I and II The epoxy conversion was then calculated A(t)/A(0), where A(t)/A( 0) is the ratio of normalized peak area with respect to the uncured and unexposed system. For quantitative analysis in accordance with the Beer Lambert law, the absorbance was less than 1.0. The baselines for area calculation of those peaks were chosen as suggested by Dannernberg without any baseline correction 102, 103 To estimate the amount of absorbed water for each exposure condition, the normalized area of the characteristic water peak around 5230 cm 1 w as obtained, which plane 142 144 The normalized area of the characteristic water peak provide another method to measu re the amount of absorbed water in various polymer water systems, where the normalized area is proportional to th e real amount of absorbed water 142 145 In this work, the area of this region was also normalized by t he phenyl absorbance at 4622 cm 1 Because these are normalized values we can not use them to determine absolute concentrations of water present; however, because the normalized areas are proportional to the amount of water they can be used for relative co mparisons. Although this peak includes hydroxyl groups generated in the epoxy network due to the cure reaction, the change in absorbance of the peak by the cure reaction is negligible compared with the amount of absorbed water. 152

PAGE 65

65 A summary of characteristic NIR absorption band for both systems I and II is shown in Table 4 4. Table 4 4. Assignment of observed peak for systems I and II in NIR spect ra Peak position (wavenumber, cm 1 ) NIR Assignment 4530 Combination of C H stretching and bending of epoxide ring 4622, 4680 Combination of aromatic conjugated C=C stretching (1626cm 1 ) with aromatic CH fundamental stretching (3050cm 1 ) 52 30 Characteris tic water peak by combination of asymmetric stretching ( as ) and in plane deformation ( ) of water 5986 Overtone of Phenyl CH stretching 606 5 First overtone of the terminal CH fundamental stretching Results As shown in Fig ure 4 2, the DSC curves for each system after 7 days of hygrothermal exposure show th at the changes of T g s are difficult to understand due to the fact that both increase and decrease in T g s were observed during the course of exposure. Samples exposed at 40, 50, and 60C in system I show significant increases in T g s compared with unexposed, while T g decreased for exposed at 30C. On the other hand, decreases in T g s for all the exposure conditions were observed for system II after 7 days of exposure, compared to unexposed samples. From the DSC curves in Fig ure 4 2, it is also revealed that en dothermic aging peaks are superimposed around the glass transition region. This well known behavior is structural enthalpy relaxation or physical aging, a result of the recovery of enthalpy trapped in the glassy state of the material during aging. 146 151 This phenomena is easily observed for the amorphous polymers being cooled down to below T g or isothermally cured at a temperature below T g typically showing enthalpy relaxation upon reheating.

PAGE 66

66 Fig ure 4 2. Typic al DSC curves after water immersion for 7 days at different temperatures compared with no exposure: (a) System I (b) System II The arrows indicate T g for each condition. In order to monitor the changes in T g s during the entire course of the exposur e, the T g s of both systems are plotted as a function of exposure time as shown in Fig ure 4 3. The results clearly indicate complex hygrothermal behaviors in terms that both increase and decrease in T g values were observed in a complicated manner during the course of exposure although each system exhibits a different behavior in terms of the manner of changes in T g s for each exposure condition. For system I the samples exposed at 50 and 60C show continuous increases in T g with exposure until the rates

PAGE 67

67 of i ncrease slow down and approach saturation points while, for samples exposed at 30 and 40C, an initial decrease in T g was followed by an increase. Fig ure 4 3. Changes in T g at different exposure conditions after curing 4 weeks at room temperature in air : (a) System I (b) System II In this picture, the size of the error bars displaying the standard deviation is too small to be shown at some points. On the other hand, for system II the T g values for all the exposure condition starts from decreases at the b eginning stage of exposure, but followed by increases after 1~2 days of exposure. At the later stage of exposure, the samples exposed at 30 and 40C

PAGE 68

68 keep continue to increase T g until the rates of increase slow down and approach saturation points. On the c ontrary, samples exposed at 50 and 60C exhibit additional decrease of T g at the later stage and approach saturation points. Fig ure 4 4 Near infrared spectra for samples immersed in water for 7 day s at different temperat ures compared wit h an unexposed / control sample : (From bottom to top ) control (exposed in air) exposed at 30C, expos ed at 4 0C expos ed at 5 0C and expos ed at 6 0C ) : (a) System I (b) System II Cure condition was 4 weeks at room temperature in air before the exposure was started.

PAGE 69

69 In order to figure out how post curing reactions affect hygrothermal behaviors, changes in conversion in terms of decreases in epox ide group were also monitored during the same course of the exposure. The cure condition, in this experiment, was also the same to that used for measuring changes in T g namely, 4 weeks at room temperature in air before the exposure was started. As shown i n Fig ure 4 4 where the samples exposed to water immersion for 7 days at different temperatures are compared with control samples which had been exposed to air for 7 days, both systems clearly show that decreases in the area of epoxide group around 4530cm 1 with increase of exposure temperature, indicating post curing reaction obviously occurred. In monitoring the changes in conv ersion during the entire course of exposure as shown in Fig ure 4 5, it shows that the epoxide conversion in both systems increased due to post curing reaction for all the exposure conditions, resulting in significant increases in cross link density compare d with the unexposed sample. This result shows that the elevated temperature of the water caused additional cure, even at a low temperature such as 30C. From the comparison of the features of changes in T g s and changes in conversion with hygrothermal exposure for each system (Fig ures 4 3 and 4 5), it clearly indicates that the increase in cross link density by post curing reaction leads to the increase in T g s with exposure. At the same time, it shows that property loss by plasticization simultaneously occurs with additional cure, also playing a significant role in determining the properties. It is evident tha t two phenomena are in competition with each other during the exposure.

PAGE 70

70 Fig ure 4 5. Changes in conversion at different exposure conditions after curing 4 weeks at room t emperature in air : (a) System I (b) System II For example, in system I the samples exposed at 30 and 40C shows an initial decrease in T g despite continuous increase in cross link density, which indicates that plasticization effe ct by water has the predominant effect on T g at initial stage, compared with additional cure. However, as exposure continues, the additional cure becomes the dominating factor, resulting in an increase in T g With regards to exposure at 50C, it is

PAGE 71

71 observe d that the increase in cross link density is dominant throughout the exposure, resulting in a monotonic increase in T g When comparing samples at 50 and 60C, the T g for 50C samples is larger than T g for 60C samples after 1 day exposure despite the fact that samples exposed at 60 C always had a higher conversion values. This indicates that at 60C there is more absorbed water, and therefore greater plasticization. In a similar manner, system II also shows the combinational effect between two competing fac tors during the exposure. When observing the samples exposed at 30 and 40C, they show the initial decreases in T g s indicating the larger plasticization effect dominant over the effect of additional cure, which was followed by increases in T g s indicating t he changes in dominating factor from plasticization to additional cure. For the samples exposed at 50 and 60C, it is shown that the dominating factor was changed from plasticization at the initial stage (causing decreases in T g s at 1~2 days of exposure) t o additional cure at the middle stage (causing increases in T g s at 2~7 days of exposure) and returned back to plasticization later (causing decreases in T g s at 7~14 days of exposure). In this system, the larger plasticization effect can also be observed at higher temperatures of exposure, in which the samples exposed at 50 and 60C shows lower T g values than those exposed at 30 and 40C at the later stage of exposure despite higher conversion values. Discussion From the comparison of the two figures showing changes in T g and conversion with hygrothermal exposure (Fig ures 4 3 and 4 5), we can see that the increase in T g due to additional cure and the decrease in T g due to plasticization are in competition with each other during exposure. In our previous study we have successfully proposed

PAGE 72

72 the tool for the evaluation of the relative importance of two effects separately, in which T g can then be used to exclude the influence of cross link density by plotting the results from Fig ures 4 3 and 4 T g values directly between exposed and unexposed systems at the same conversion value. Fig ure 4 6. T g systems: (a) System I (b) System II Points were experimentally obtained and solid line s are fits to theoretical

PAGE 73

7 3 For the construction of T g conversion master plot, the well known semi empirical expression of the relationship between T g and conversion for cross linked systems are utilized as shown below 70 72 (4.1) where is conversion, T g0 and T g are the T g values of the monomer and the fully cured network and C p0 and C p are the heat capacity changes at T g0 and T g respectively. To construct the master plot of T g as a function of conversion, T g and conversion values under identical cure conditions were experimentally obtained for unexposed samples of each system. The semi empirical equation shown above was used for fitting the exp erimentally obtained data as shown in Fig ure 4 6. Both systems show good agreements to the experimental data, with coefficients of determination of 0.99 for both c ases. The values of C p / C p0 as a fitting parameter were 0.88 and 0.66 for systems I and II resp ectively. II the samples with lower values of conversion exhibited multiple T g values in DSC curves, which m ight be due to phase separation of mixed specimens between epoxy resin and hardener. Since system II consisted of different blends of amine, it is estimated that the system under low value of conversion still has phase separation after hand mixing which re sulted in having multiple Tgs. This behavior, however, disappeared from the point with the conversion value of 0.23 and showed a single value of T g s thus, for system II, samples with conversion values greater than 0.23 were used to construct the master pl ot On the contrary, system I does not have the phase separation from the initial mixture between

PAGE 74

74 epoxy resin and hardener, which can be estimated that blends of amines used in this product was well mixed even at initial condition. T g0 for system II was estimated by fitting using the semi Fig ure 4 7. Direct comparison of T g between unexposed and exposed for system I using the master plot of T g versus conversion. The section mark ed with dotted lines in (a) is magnified in (b).

PAGE 75

75 Fig ure 4 8. Direct comparison of T g between unexposed and exposed for system II using the master plot of T g versus conversion. The section marked with dotted lines in (a) is magnified in (b). Into these constructed master plots for unexposed systems, the values of T g and conversion for each exposed system I and II which shown in Fig ures 4 3 and 4 5, were finally applied as displayed in Fig ures 4 7 and 4 g the difference in T g between unexposed and exposed samples at the same conversion value can successfully propose a parameter indicating plasticization effect quantitativel y

PAGE 76

76 while excluding the influence of changes in cross link density with exposure. In this calculation of T g differences, T g s for the exposed system were determined directly from DSC measurements, whereas T g values at the corresponding conversion for the unex posed system were obtained from the values on the fitted curve, as indicated in Fig ures 4 7 (b) and 4 8 (b). Fig ure 4 9. Relative amount of water absorption with exposure time estimated by normalized area of the characteristic water peak from N IR : (a) System I (b) System II

PAGE 77

77 In order g with the absorbed water amount at each exposure condition, the relative amount of water uptake was determined from the characteristic water peak around 5230 cm 1 which was well resolved in the NIR spectra of each system. When the relativ e amount of absorbed water is evaluated by this method, both systems exhibit the same typical trends that the amounts of water absorption increases as exposure continues and that the rate of increase in absorbed water at the initial stage of the exposure i s enhanced as exposure temperature increases by an increase in diffusion rate as temperature increases (Figure 4 9) However, for the features of the amount of water uptakes varied with different temperatures at the later stage of exposure, each system show a different behavior, contrary to each other. After water absorption saturated at some points, system II shows larger water uptakes for exposed at higher temperatures, which is not unusual, while system I shows the opposite behavior, that is larger water uptakes for exposed at lower temperatures. Regarding this abnormal behavior of wat er absorption in system I, it is probably related to an imperfect cure of the system. 153 Since system I still has larger difference in conversion value at the later stage of exposure between the samples exposed at different temperatures, it is estimated that there is also larger difference i n density of the network structure of polymer chains, where the samples exposed at lower temperatures which has lower conversion value have loose network with more free volume, resulted in easier ingress of water On the other hand, for system II, althoug h there is also difference in conversion values between the samples exposed at different temperatures at the later stage of exposure, the amount of differences in percent conversion is not so much as that for

PAGE 78

78 system I. In addition, system II had already re ached a higher conversion value compared with system I before the exposure started It is estimated that there was not huge differences in polymer networks between the samples exposed at different conditions of post curing reactions, resulted in the normal hygrothermal behavior having a larger water uptakes with higher temperature of water exposure Fig ure 4 10. T g differences between unexposed and exposed systems as a function of estimated amount of absorbed water : (a) System I (b) System II

PAGE 79

79 Finally, T g was plotted as a function of the relative absorbed water amount for each exposure system As shown in Fig ure 4 10 T g increases with water amount, showing the expected behavior of increase in plasticization as the amount of absorbed water increases. F rom the comparison of two systems in this figure, system II shows larger deviation of T g values with different exposure temperatures, indicat ing that small chages in water uptakes resulted in larger plasticization effect in this system. Summary In this work, complex hygrothermal behaviors of two commercial products, both used as a seal coat and impregnati ng resin for structural strengthening applications have been investigated, in which property loss by plasticization simultaneously occurs with additional cure during the exposure. From the comparison of two features describing changes in T g s and cross link density with water immersion at different temperatures, it clearly shows that the plasticization effect due to water absorption and the effect of additional cure due to post curing reaction are in competition with each other during the exposure. Construct ion of the plot of T g versus conversion for the unexposed system successfully provided an excellent method to exclude the factor of cross link density. By directly comp are the T g values between exposed and unexposed samples while ruling out the factor of cross link density. Combined with the estimated amounts of water absorption, it provided a tool to evaluate plasticization effect quantitatively, which was due to their own environmental exposure.

PAGE 80

80 The suggested method will contribute to the understanding of the complex behavior having property increase and decrease at the same time for various cross linked polymer areas employing low temperature curing under environmental effects

PAGE 81

81 CHAPTER 5 EFFECT OF WATER ADDITION ON CURE KINETICS OF AN EPOXY AMINE THERMOSET Background Although epoxy systems have numerous advantages over other thermoset resins such as lower cure shrinkage lower residual stress, chemical resistance, and insulating properties t a convenient use of a wide range of temperatures by judicious selection of curing agents enabling good control over the degree of cross 3 In the se use s of epoxy materials, additional variables are often included into a given system such as solvent, concentrations, viscosity, and mixing rates providing another method to control degree of cross linking or the speed of reaction. 4 9 Also, an additional incorporation of catalysts such as imidazole have extensively been utilized as a convenient way to regulate the condition of manufacturing processes. 10 13 This has been pref erred in many industrial areas such as outdoor or electronic applications where high temperature cannot be used easily as a means of accelerating the curing reaction in order to take advantage s of easy i n stallation A ddition of hydroxyl containing coupound s such as water or various types of alcohols and phenols has also known to increase the rate of cure reaction 4, 14, 15 C atalysis effects of hydroxyl groups in those types of added compounds contribute the ring open ing of epoxides although the exact impact of added water has not been suggested quantitatively I n the previous stud ies in C hapters 3 and 4 it has also been revealed that water may largely affect s the k inetics of the post curing reaction in which conver sion of an epoxy amine thermoset has been significantly accelerated with the help of water immersion even at low temperatures such as 30C. 152

PAGE 82

82 Und erstanding of the moisture effect on cure kinetics is very important, especially in the case of the systems employing slow curing at ambient condition for easy installation such as the application of structural adhesives, where absorption of moisture at in itial condition of storage or during the course of curing could result in huge different states of cure degree at the final products. Although some previous studies have reported that added water can significantly contribute to accelerate curing reaction i n epoxy amine thermosets, the effect of added water itself with a view to the kinetic aspect has not been quantitatively evaluated. 4, 14 In this study, we have focused on how the added water inf luences the polymerization kinetics and reaction chemistry. This suggests the direct comparison of the kinetic values between catalyzed hydroxyl effect by added water and that by auto catalysis, using a modified mechanistic modeling. Such an effort will co ntribute to the understanding of the relationship between the materials properties and the processing environments, which will also be very useful in designing materials with tuned properties. Experimental Section Materials The material used in this work is a two part epoxy, diglycidyl ether of bisphenol A (DGEBA EPON 826, epoxy equivalent weight=178~186g/equiv, n=0.08, information provided by manufacturer, ) and poly(oxypropylene) diamine (POPDA, Jeffamine D 230 n=2~3, information provided by manufacture r ), purchased from Hexion and Huntsman, respectively Figure 5 1 illustrates the chemical structures of DGEBA and POPDA. The mass ratio of DGEBA to POPDA was 100 to 32.9, to reach stoichiometric equivalence between functional groups based upon manufacturer s information.

PAGE 83

83 The two liquid components with selected amount of deionized water (0, 1, 2 and 3wt%) were mixed vigorously for 3 min to ensure even mixing. The limit of the amount of added water was determined by the solubility of water in the epoxy resin. When the excessive water of 4wt% is added into the epoxy resin, the system no longer shows intimate mixing, in whi ch phase separated forms of microdroplets of water are easily observed by the naked eye after being cured. Since the intent of this study is to investigate the water or moisture effect by spontaneous absorption of the system itself from the environment which could occur during the service of the products the limit of the water content was set to 3wt%. T he mixed material was put into 15~18 rep licates of preheated teflon spacers at different cure temperatures of 5 0, 55 60 65 and 7 0C with a convection oven. After curing for the selected time, the specimens were periodically removed from the oven followed by immediate quenching using liquid n itrogen to minim ize the effect of residual heat and were then the spectra were ready to be taken F igure 5 1. Chemical structures of diglycidyl ether of bisphenol A (DGEBA top ) and poly(oxypropylene)diamine (POPDA bottom ).

PAGE 84

84 Techniques To monitor the c ure kinetics of the epoxy amine system, we used Fourier transform infrared spectrometry (FTIR; Nicolet Magna 76 0, Thermo Electron Cooperation) The prominent advantages of FTIR to monitor the cure kinetics of epoxy systems have been well established, in wh ich quantitative amount of functional groups involved in the reaction can be easily observed. 73, 74, 154 In this study, n ear infrared spectra were recorded with a CaF 2 beamsplitter and an MCT detector over the range of 3800 7 5 00 cm 1 using 32 scans at a resolution of 4 cm 1 For the recording of IR spectra for the initial cure stage of the samples which did not solidify after the selected cure time, KBr IR disposable cards were used. The other samples which solidifie d after the selected cure time were prepared by casting thin films with a pproximate 0.5 ~1 mm thickness using Teflon spacers. F igure 5 2 Typical FT IR Spectra for DGEBA POPDA system (from top to bottom): Cured at 60C for 40 min, Cured at 60C for 120 m in, Cured at 60C for 450 min, and Cured at 50C/1hr followed by 80C/2hrs and 125C/3hrs S pectra were moved parallel to the y axis for the display.

PAGE 85

85 Typical NIR spectra for the systems are shown in Figure 5 2. To measure the concentrations of primary amin es and epoxides, IR spectra have been directly observed. For the quantitative analysis of the epoxide group the area of the epoxide peak around 4530 cm 1 was normalized to the phenyl peak around 4622 cm 1 The epoxy conversion was then calculated by = 1 A ( t )/ A (0), where A ( t )/ A (0) is the ratio of the normalized peak area with respect to the uncured system. The baselines for area calculation of those peaks were chosen as suggested by Dannernberg without any baseline correction 102, 103 For the measurement of concentrations of primary amines, the area of the primary amine peak around 4930cm 1 was also normalized to the phenyl peak around 4622 cm 1 Although the reliability of this quantitative analysis of the epoxy cure reaction was confirmed by previous workers, 102, 103, 152, 155 we reconfirmed it for our system. In this experiment, to demonstrate the validity of using Beer peak areas for both t he epoxide and primary amine in the spectra were measured with respect to differ e nt epoxide concentration s To monitor the normalized area for the epoxide group specimens with different epoxide concentrations were cured at high temperature s for enough tim e to ensure complete reaction between epoxide and amine functional groups. The cure condition was 50C for 1 h r followed by 80C for 2 h rs and 125C for 3 h rs On the other hand, for the primary amine, IR spectra of specimens with different epoxide concent rations were recorded right after mixing the 2 components. As shown in Figure 5 3, the normalized absorbance either decrease s or increase s linearly as epoxide concentration decreases, which indicate that our analysis is reliable and

PAGE 86

86 obeys the Beer Lambert law. For quantitative analysis in accordance with the Beer Lambert law, th e absorbance of all the specimens was less than 1.0. Figure 5 3. Experimental confirmation of the reliability of quantitative analyses from FTIR : (a) Plot of normalized ar ea assigned to epoxide at 4530 cm 1 with variation in epoxide/amine ratio. The correlation coefficient value, R 2 is 0.99. (b) Plot of n ormalized area assigned to primary amine at 4930 cm 1 with variation in epoxide/amine ratio. The correlation coefficient value, R 2 is 0.99. In addition, when Figure 5 3 (a) is carefully observed, it reveals that the determined mixing ratio between part A and part B, which was based upon the manufacture s information, was slightly off stoichiometric ratio. From the intercept ion of the fitted linear regressions between excess epoxide condition (square) and excess

PAGE 87

87 amine condition (circle), it shows that the stoichiometric mass ratio is 100 to 33.4 between DGEBA and POPDA, which slightly deviated from the selected ratio of 100 t o 32.9 in this study. For the calculation of concentrations in this study, we follow the information obtained from the IR spectra, which corresponding to the concentration ratio of 4.14 to 2.03 mol/kg between epoxide and primary amine. R esults and Discussi on Reaction Chemistry Epoxide Conversion For the epoxy amine reaction, it is well known that the main features include (1) reaction of primary amine with an epoxide to form a secondary amine and (2) further reaction of the secondary amine with another ep oxide to form a tertiary amine 73, 74 Although other possible mechanisms can be also included such as homo polymerization by epoxide epoxide and etherification by epoxide hydroxyl reactions, those only occur under c ertain conditions, such as in the presence of acid/base catalysts or at higher cure temperatures. 75 80 Given the system in our study employs only neutral water and relatively low cure temperatures they can be safel y ignored. In the meantime, water in the system could influence the reaction chemistry in terms that weaker nucleophiles of water can even react with epoxides in the presence of acid catalysts Although no acid or base catalyst was included in this study, in order to estimate the possibilities of the reaction between pure epoxide and water, only part A containing pure epoxides were mixed with deionized water and cured at various conditions. As shown in Figure 5 4, the normalized areas for the epoxide around 4530cm 1 obtained from IR spectra did not vary with all the cure conditions, which indicates there is no reaction between epoxide and water in this study.

PAGE 88

88 F igure 5 4 Comparison of normalized area for epoxide around 4530cm 1 obtained from IR spectra fo r different water contents and cure conditions. Finally, to evaluate the effect of added water on cure kinetics, conversion in terms of a decrease of epoxide functional groups has been plotted as a function of cure time with different water contents and d ifferent cure temperatures. As shown in Figure 5 5, FT IR demonstrates that the rate of cure reactions significantly increased with increases in water contents. Furthermore, it reveals that the added water also affects the final conversion values at the la ter stage of cure, which also slightly increased with added water. Combined with the prior results in Figure 5 4, it can be concluded that water only contributes to the ring opening reaction of epoxides as a catalyst to increase curing rate. In order to un derstand the entire chemical reaction and kinetic relations of the system, the changes in concentration s of the epoxide and primary amine were directly calculated from the monitoring of IR absorption bands. In this calculation, the initial concentration of each specie before the curing reaction proceeds allows us convert the changes in the area of IR absorption bands to the concentration profiles, based on our validation of Beer as shown in Figure 5 3. 74

PAGE 89

89 Figure 5 5. Epoxy conversion versus cure time with different contents of added water : (a) Cured at 50 C, (b) Cured at 55 C, (c) Cured at 6 0 C (d) Cured at 65 C, (e ) Cured at 70 C

PAGE 90

90 Figure 5 6. Concentration profiles for epoxide and primary amine. The profiles were directly obtained from FT IR measurement : (a) Cured at 50 C, (b) Cured at 55 C, (c) Cured at 60 C, (d) C ured at 65 C, (e) Cured at 70 C

PAGE 91

91 Figure 5 7 Concentration profiles for secondary and tertiary amine. The profiles were obtained using mass balance equations: (a) Cured at 50 C, (b) Cured at 55 C, (c) Cured at 60 C, (d) Cured at 65 C, (e) Cured at 70 C

PAGE 92

92 On the other hand, the mass balance equations also allow us to monitor the changes in other functional groups of interest which are involved in the reaction such as secondary amines and tertiary amines. In this mass balance equation shown below, it is assumed that the total number of nitrogen atoms is constant during the entire reaction (Equation 5 1), while the number of the generated hydroxyl functional groups counted by the epoxide consumption corresponds to that counted by the generation of seconda ry and tertiary amines (Equation 5 2). I n this hydroxyl balance, etherification reactions are ignored. 6, 74 ( 5 1) ( 5 2) where [E], [A 1 ], [A 2 ], [A 3 ], and [OH] are the concentrations of epoxide, primary amine, secondary amine, tertiary amine and hydroxyl groups present in the system, respectively. [OH] auto refers to the hydroxyl groups generated by autocatalysis during the chemical reaction. In Figure s 5 6 and 5 7 the changes in concentration for the functional groups of interest can be finally monitored during the entire course of the curing reaction. As shown in F igure 5 6 primary amines are consumed faster with the increase of the added water for all the reaction temperatures, indicating faster reaction s of primary amines with epoxides to form secondary amines. In the meantime, tertiary amines exhibit the faster production with the increase of the added water (Figure 5 7) It shows that further reaction of secondary amines with epoxides to form tertiary amines also speed up with the added water, which also confirms that the rate of the cure reaction is accelerated by the addition of water. In accordance with the faster consumption and

PAGE 93

93 production of primary amines and tertiary amines, respectively, secondary amines also exhibit the faster rate of cure reaction with the in crease of the added water. In this concentration profile of secondary amines, they were produced faster with the increase of water contents, which was followed by faster consumption to form tertiary amines. Reaction Kinetics As has been suggested in a numb er of previous studies, the main cure mechanism for epoxide amine reactions has been generally described by two reaction paths: (1) reactions catalyzed by compounds containing nitrogen hydrogen bonds (designated as non catalytic by previous workers) and (2) those by oxygen hydrogen bonds ( designated as catalytic ). 74, 81 85, 156 In this kinetic expression, it is often described that the reaction of an epoxide group with an amine occurs through the formation of ter molecular intermediates consisting of an amine, epoxide, and hydroxyl due to the existence of hydrogen bonding in which the formation of the termolecular intermediate is the rate determining step. 7, 14, 74, 79, 156 158 The involved reactions with the assumption of the formation of termolecular intermediates can be displayed as shown in Figure 2 1 When the reactions shown in Figure 2 1 are written in a compact form which containing the reaction rate constants, they can be displayed by following equations ( 5 3) ( 5 4) ( 5 5) ( 5 6)

PAGE 94

94 where k 1 and k 1 indicates that the non catalytic and catalytic kinetic parameters for epoxy primary amine a ddition while k 2 and k 2 are those for epoxy secondary amine addition, respectively. Thus, t he well known kinetic expression for the above epoxy amine reactions can be described as follows. ( 5 7) ( 5 8) In this expression, the first term of each equation represents the non catalytic reaction while the second term shows the catalytic reac tion by hydroxyl functional groups. For the second term, the effect of hydroxyl generated from added water or absorbed moisture from air has n ot been separated from that generated from polymer chain such as hydroxyl effect by autocatalysis or the effect of hydroxyls already present in the reactant molecules. However, since the intent of this study is to evaluate the effect of hydroxyl from added water itself, the given equations should be further developed which separates those two effects In the meantime, the non catalytic reaction term can be neglected when considering that the catalyzed effect of amine group is much smaller than that of hydroxyl. 74 Thus, for the purpose of evaluating how much the added water affects the cure kinetics in this work, E quations 5 7 and 5 8 can be further developed into following E quations 5 9 and 5 10 with the assu mption of negligible effect of the non catalytic reaction. In these equations, hydroxyl groups are divided into two groups: (1) [OH] chain which yielded from the polymer chain by initially present ([OH] initial chain ) or auto generated during the reaction ([ OH] auto ) and (2) [OH] water which given by initially

PAGE 95

95 absorbed moisture ([OH] initial moisture ) or intentionally added water ([OH] add ). Thus, k 1 C and k 2 C represents the catalytic kinetic parameters with regard to the hydroxyl groups generated from the polym er chain while k 1 W and k 2 W are with regard to the hydroxyl groups by added in the form of moisture or added water. ( 5 9) ( 5 10) where total amounts of [OH] chain and [OH] water can be derived as follows ( 5 11) ( 5 12) Also, [OH] initial chain which is present due to the polymer chain of DGEBA (n=0.08) can be calculated using the following equation. ( 5 1 3 ) E quations 5 9 and 5 10 can be also represented as Equations 5 1 4 or 5 1 5 in the form where all the relative kinetic parameters of k 1 C k 1 W k 2 C and k 2 W are combined in one equation. ( 5 14) ( 5 15)

PAGE 96

96 In order to evaluate the kinetic parameters regarding epoxy primary amine addition, E quation 5 9 was developed into the following reduced form. ( 5 16) When the reduced reaction rate (the left term of (16)) is plotted as a function of [OH] chain then k 1 C and k 1 W [OH] water can be derived from the slope and intercept of t he linear regression of the early stage of cure for each reaction condition, respectively, as shown in Figure 5 8 In order to obtain pure intrinsic kinetic parameters, the later stage of the reaction was removed from the linear regression of the fitting, in which the main reaction is under the control of diffusion effect. The limit of the linear regression was set up to approximately 50~60% conversion in terms of the values of epoxide conversion. 159 161 Using this m ethod, the determined kinetic parameter of k 1 C is shown in Table 5 1 and Figure 5 1 2 Also, each intercept value from Figure 5 8 representing k 1 W [OH] water for all the conditions of different water contents and cure temperatures can be plotted as a functi on of the amount of added water ([OH] add ) as shown in Figure 5 9 From this graph, the slop e obtained by the linear regression with regard to each reaction temperature indicates the kinetic parameter k 1 W while each intercept represents the value of k 1 W [O H] moisture, initial In this estimation, we assumed that 1 mole of the added water produces 1 mole of hydroxyl group which have the catalyzed effect on the curing reaction. The determined kinetic parameter of k 1W is shown in Figure 5 12 and Table 5 1. The initial amount of absorbed water by moisture ([OH] moisture, initial ) for each condition was also estimated and shown in Table 5 1.

PAGE 97

97 Figure 5 8 Reduced rate of epoxy primary amine reaction as a function of [OH] chain : (a) Cured at 50 C, (b) Cured at 55 C, (c) Cured at 6 0 C (d) Cured at 65 C, (e ) Cured at 70 C

PAGE 98

98 Figure 5 9 Plotting of intercept values (k 1 [OH] water ) from Figure 5 8 as a function of the contents of added water ([OH] water ) With the calculated values of k 1 C k 1 W and [OH] water by the method suggested above (Figures 5 8 and 5 9), Equation 5 9 presents the predicted values of the reaction rate for the primary amine. These predicted values are compared with the experimental values obtained by differentiation of the concentration ch anges of primary amines which were directly observed by FT IR measurement, as shown in Figure 5 6. I t is observed that the rate equations derived according to the proposed cure mechanism describe the experimental phenomena during the course of the reaction For the evaluation of the kinetic parameters with regard to the epoxy secondary amine reaction (k 2 C k 2 W ), the concentration profiles for tertiary amines should be obtained by the calculation of the mass balance equation with the values of the concentr ations of epoxide and primary amines directly measured by FT IR.

PAGE 99

99 Figure 5 1 0 Comparison of experimental and predicted values for reaction rate of primary amine (symbols: experimental, lines: predicted) : (a) Cured at 50 C, (b) Cured at 55 C, (c) Cu red at 6 0 C (d) Cured at 65 C, (e ) Cured at 70 C

PAGE 100

100 Figure 5 1 1 Comparison of experimental and predicted values for reaction rate of epoxide (symbols: experimental, lines: predicted) : (a) Cured at 50 C, (b) Cured at 55 C, (c) Cured at 6 0 C (d) Cu red at 65 C, (e ) Cured at 70 C

PAGE 101

101 However, it is actually difficult to obtain precise values for the changes in concentration of tertiary amine in that the n umerical values of concentrations are much smaller compared with that of the epoxide and primary amin e. Given the ex situ condition to obtain periodical changes in normalized area for epoxide and primary amine while handling the FT IR spectra, it is inevitable that very small variation of the normalized area of epoxide and primary amine resulted in somewh at large deviation in the calculation of tertiary amine concentrations, which consequently produces much larger statistical errors for the linear regression with the first derivative value as a function of time using E quation 5 10. For another approach to estimate the kinetic parameters regarding the epoxy secondary amine reaction, it is suggested that the parameters can be obtained using E quation 5 15, assuming the ratios of k 2 /k 1 and k 2 C 1 C (as well as k 2 W /k 1 W in this study) are independent of the re action path 83, 84, 101, 162, 163 In this assumption, the ratio of the reaction constant between epoxide secondary amine and epoxyde primary amine addition are assumed to be the same whether the reaction occurs thro ugh non catalytic reaction, catalytic reaction via hydroxyl from polymer chain, or catalytic reaction via hydroxyl from water/moisture ( k 2 /k 1 = k 2 C 1 C = k 2 W /k 1 W ). Thus, from E quation 5 15, when the concentration of the secondary amine is at the maximu m value, it follows that the left hand side is equal to 0. The n with the assumption of the ratios being independent of the path, the following equation can be derived 101 ( 5 17)

PAGE 102

102 When the values of [A 1 ] and [A 2 ] at the maximum point of [A 2 ] are directly taken from the curves shown in Figure s 5 6 and 5 7 for each cure condition, kinetic parameters k 2 C and k 2 W can be obtained as shown in Figure 5 1 2 and Table 5 1 Combined with the obtained kinetic parameters of k 1 C k 1 W k 2 C and k 2 W the validity of the utilized model for the entire reaction can be compared using E quation 5 14 as shown in Figure 5 1 1 Again, it is shown that the rate equations derived according to the proposed cure mechanism well describes the experimental changes of concentration in this system. With the deriv ed values of each kinetic parameter, Arrhenius temperature dependencies can be analyzed for the catalytic reaction rate constants of k 1 C k 1 W k 2 C and k 2 W as shown in Figure 5 1 2 Activation energies and corresponded values of pre exponetial factors were also calculated from the resulting slopes and intercepts, as shown in Table 5 2. From Figure 5 1 2 it clearly shows that the rate constants of k 1C and k 2C have slightly higher values than k 1W and k 2W respectively, indicating that the catalyzed effect by hydroxyl s from the polymer chain is larger than that by hydroxy l s from moisture or water in that the reaction is faster for all the reaction temperatures, where auto catalyzed nature still has a larger influence to the cure acceleration than the effect of water addition The comparison of the calculated activation ene rgy which have comparable values of 55.0 and 61.3 kJ/mol for the epoxide primary amine addition and 46.3 and 52.5 kJ/mol for the epoxide secondary amine addition also supports the above comparison of rate constants. Considering the acidity of water is very comparable to that of alcohols (polymer chain ) from the fact that pKa values of those are in the range of 14~16, 164 the

PAGE 103

103 comparable values of rate constants between the reaction cataly zed by [OH] chain and [OH] water are reasonable in which difference in the logarithm value of the reaction constant is less than one order of the magnitude. Regarding the relative kinetic constant ratio between k 2C 1C k 2W 1W t he estimated values ranges from 0.19 to 0.23 as shown in Table 5 1. Th is value of re lative kinetic constant ratio represening the reactive ratio of the primary to secondary amine hydrogens which is also known as substitution effect suggests an important f actor in the formation of network morphology, in which i t determines whether extensive branching already occurs in the early stages of curing, or linear polymer chains are predomina ntly formed. 82, 100, 165 The value of the kinetic constant ratio can be termed negative when the secondary amine is less reactive than the primary amine by sterical hindrance and reduced mobility resulting from the addition of epoxide while the positive substitution effect is termed when the secondary amine is more reactive. 82, 157, 165, 166 The low value of the kinetic constants ratio in this study indicat es that a lower reactivity of the hydrogen in a secondary amine group with respect to a hydrog en in a primary amine group result in initial formation of linear polymer chains predominantly 100, 166, 167 The obtained results of strong negative substitution effect corresponds with the previous report about th e epoxy systems containing polyoxypropylene diamine, in which the value of the kinetic constants ratio as low as 0.2. 37 The strong negative substitution effect of epoxy amine systems is known to be mainly due to the sterical crowding on the nitrogen atoms after the reaction between epoxide and primary amine. 100, 166, 167

PAGE 104

104 Table 5 1. Summary of kinetic rate constants, kinetic constants ratio and estimated values of amount of initially absorbed moisture for each system Cure Temp. (C) Added Water Rate Constants (kg 2 /(mol 2 min)) Relative Kinetic Constants Ratio Estimated Amount of [OH] moisture, initial wt% (mol/kg) k 1 C k 2 C k 1 W k 2 W 50 0 (0.00) 0.00290 0.00072 0.00206 0.00048 0.25 0.10 1 (0.56) 0.00335 0.00078 0.23 2 (1.11) 0.00414 0.00093 0.22 3 (1.67) 0.00522 0.00122 0.23 55 0 (0.00) 0.00400 0.00095 0.00390 0.00082 0.24 0.0 9 1 (0.56) 0.00498 0.00118 0.24 2 (1.11) 0.00692 0.00109 0.16 3 (1.67) 0.00615 0.00131 0.21 60 0 ( 0.00) 0.00552 0.00104 0.00494 0.00096 0.19 0.01 1 (0.56) 0.00644 0.00141 0.22 2 (1.11) 0.00894 0.00184 0.21 3 (1.67) 0.01152 0.00189 0.16 65 0 (0.00) 0.00705 0.00157 0.00617 0.00121 0.22 0.10 1 (0.56) 0.00896 0.00176 0.20 2 (1.11) 0.01111 0.00197 0.18 3 (1.67) 0.01142 0.00213 0.19 70 0 (0.00) 0.01059 0.00201 0.00860 0.00166 0.19 0.16 1 (0.56) 0.01063 0.00211 0.20 2 (1.11) 0.01323 0.00294 0.22 3 (1.67) 0.01810 0.00289 0.16 Table 5 2. Value s of activation energies and related pre exponential factor for each system k 0 (kg 2 /(mol 2 min)) E a (kJ/mol) k 1 C 3.03 10 6 55.0 k 2C 2.65 10 4 46.3 k 1W 1.92 10 7 61.3 k 2 W 1.62 10 5 52.5

PAGE 105

105 Figure 5 12. Arrhenius relations for the rate constants of k 1 C k 1 W k 2 C and k 2 W Summary In this study, we examined the effect of water addition on cure kinetics of an epoxy amine thermoset. The near FT IR measurement suggested an excellent method to monitor the cure kinetics of the system. Experimental data clearly demonstrated that the cure reaction was significantly accelerated with the addition of a small amount of water. With the modified mechanistic modeling, it successfully compared the kinetic values catalyzed by hydroxyl group generated from polymer chain with those from added water or moisture absorption. From the comparison, it can be concluded that the catalyzed eff ect by hydroxyl from water is as large as that by hydroxyl from polymer chain. Regarding the huge effect of water which catalyzed cure reaction observed in this system, it can be estimated that the well known property of hydrophilic ity for polyoxypropylene diamine induce intimate mixing between the polymer chains and water which maximize the influence of hydroxyl from water or moisture

PAGE 106

106 The obtained results in this study propose that even small amount of water which is absorbed during the course of curing or at the storage condition could largely influence the rate of cure reactions resulting in different states of cure degree at the final products. The water influence could cause more impact especially on the slow curing system. In the field of real applic ation such as outdoor or electronic applications, i n order to take advantage of easy installation, m any epoxy amine adhesives and coatings are cured under ambient conditions thus undergoes slow cure reactions. In such cases, the small changes in humidity during the given cure time could significantly influenced the different rate of cure reaction.

PAGE 107

107 CHAPTER 6 ROLE OF CHEMICAL BONDING ON THE DURABILITY OF EPOXY ADHESION FOR STRUCTURAL STRENGTHENING APPLICATION Background Epoxy resin as a seal coat, adhesiv e and impregnating resin has been widely used for structural strengthening applications such as a rehabilitation of concrete structures using e xternally bonded carbon fiber reinforced polymer (CFRP) composites providing high performances of strength, stif fness, and resistance to creep. 15, 104 One of the most significant advantages of this method is easy installation and considerable cost savings using a wet lay up or pre cured laminate. Compared to the conven tional repair method of concrete structures by using steel, workers can easily handle the light and flexible FRP materials which involve short labor time and rapid bonding of FRP materials using heat curing epoxy adhesives. 168, 169 Although epoxy resins provide the high qualities of adhe sion for the structural strengthening the strengthened systems ha ve shown critical degradation of its performance with environmental exposure such as moisture, temperature and variou s solutions like chloride, alkali, and salt water. 105 112 A number of studies have revealed that degradation of interfacial bonding between c eme n titious materials and the utilized adhesives with environmental exposure is the main reason for the loss of strength in these system s 106, 108, 112 Although many studies have reported a loss in performance few studies have attempted a method to effectively maintai n or enhance the durability of the bond strength under the exposure. As previously stated in Chapter 2, in an att empt to enhance the durability of interfacial bonding properties between an epoxy and cementitious materials with hygrothermal exposure, the me chanism of chemical bonding is taken into account

PAGE 108

108 B ased on the hypothesis that hydrogen bonding plays a large role for the adhesion at the interface between cementitious materials and epoxies, and is responsible for the loss of adhesion under humid enviro nments by displacement of epoxy with water it is expected that chemical modification of the surface of adherend with coupling agents contributes to the enhancement of bonding properties. In this work, in order to investigate the effect of the silane treat ment of concrete surface on the durability of bonding, slant shear testing method was employed, in which the changes in failure strength for silane modified samples was compared to those for samples without modification. This work also focused on spectrosc opic analysis to identify the chemical modification of concrete. S uch an effort will contribute to the understanding of the adhesive bonding properties for structural strengthening applications which will also be very useful in designing materials with tu ned properties. Experimental Section Sample Preparation To prepare the samples in accordance with the slant shear testing method based upon ASTM C882, two equal sections of a cylindrical concrete specimen were f abricated with dimension s of 7.6cm 15.2cm, each section having a diagonally cast bonding area at a 30 angle from vertical For the fabrication of each specimen a dummy section which fits to a cylind rical mold was prepared (Figure 6 1 (a) ) After inserting the dummy section into the mold, the concrete was added into the mold and then cured. The ratios of water/cement and cement/sand were selected as 0.37:1 and 1:1.91, respectively, based upon ASTM C109 and C305. Cement from the same bag was used to produce all of the specimens to ensure consistency in the chemical and mechanical properties of the specimens.

PAGE 109

109 Figure 6 1 Pictures of apparatus to prepare composite cylinder samples : (a) dummy section (left) and mold (right), (b) supporting tool to make the bonding surface of composite cylinder samples horizontal after epoxy application (photo courtesy of Sungwon Choi) I n this work, different cur ing schedule s of each concrete mi xture were employed, which fit the different purp oses of each slant shear test : (1) preliminary slant shear testing and (2) primary slant shear testing. The main objective of the preliminary slant shear testing was to find the silane coupling agent that demonstrated the best performance under hygrothermal exposure. In this preliminar y tes t to simply compare the durability between the samples with silane modification and those without modification, only one exposure condition was chosen: water immersion at 50 C. For the sa mple preparation of this test in order to shorten the fabricat ion time, the concrete mixture was cured for a total of 3 weeks in air (1~2 days in a cylinder mold at RT followed by 19~20 days in air after de molding). For the primary slant shear testing, samples were prepared to monitor the changes in durability with the selected s ilane coupling agent which show ed the best performance In this primary test the changes in failure load with different exposure

PAGE 110

110 conditions were monitored, after a full cure of concrete specimens The cure conditions of the concrete specimen s for this test include d cur ing in a cylinder mold for 1 week, followed by immers ion into a lime solution for 3 weeks af ter de molding, followed by curing in air for 1 week. Two different types of silane coupling agent s were utilized in the preliminary sla nt shear testing to compare their contribution to the enhance ment of the durability with hygrothermal exposure: (1) epoxy functional silane coupling agent (3 glycidoxypropyltrimethoxysilane) and (2) amino functional silane coupling agent ( 3 amino propyltri methoxysilane ) To prepare each silane solution, 1wt% of the selected coupling agent was added to a 90:10 mixture by weight of ethanol:deionized water. After the pH was adjusted to approximately 5 by the addition of a few drops of acetic acid, the solution s were stirred for 60 minutes to allow complete hydrolysis of the coupling agent and were then ready to be applied. To prepare the composite cylinders with treatment of the silane coupling agent, the surfaces were wiped with a piece of clean fabric befo re the treatment without employing any method of surface roughening such as sand blasting. Subsequently, both surfaces of two equal sections of concrete were painted with the prepared silane solution using a brush for 10 mins. After painting, the samples w ere put into an oven at 6 0 C for hydrolysis. Then the surface was ready for bonding, preferably within 1 hour. A two part epoxy, diglycidyl ether of bisphenol A (DGEBA, EPON 826 ; Hexion ) and poly(oxypropylene) diamine (POPDA, Jeffamine D 230 ; Huntsman ) was used as an adhesive to bond two sections of concrete mortar in both the preliminary and primary slant shear testing The mass ratio of DGEBA to POPDA was 100 to 32.9, to reach

PAGE 111

111 stoichiometric equivalence between functional groups based upon the manufacture information. The two liquid components were mixed vigorously for 5 min s to ensure even mixing. The mixed material was then degassed for 30 mins under vacuum to remove air bubbles and ready to be applied. For the application of the epoxy adhesive, the cylinder halves were jacketed at the bonding surface using duct tape. To ensure a uniform separation of 1~2mm between the two cylinder halves, a small piece of plastic was used as a spacer. Finally, the composite cylinder samples for slant shear testing were prepared by injecting the epoxy adhesive using a syringe through an opening in the jacketing and allow ing it to wick into the space between the two bo n ding surfaces. Subsequently, remaine d horizontal and the epoxy was allowed to cure (Figure 6 1 (b) ) In this work, the epoxy adhesive was cured at RT for 3 weeks before exposure was begun for both the preliminary and primary slant shear testing Figure 6 2 Chemical structures of DGEBA, P OPDA, epoxy functional silane ( 3 g lycidoxypropyltrimethoxysilane ), and amino functional silane ( 3 amino propyltrimethoxysilane ).

PAGE 112

112 Chemical structures of DGEBA, POPDA, epoxy functional silane ( 3 g lycidoxy propyltrimethoxysilane ), and amino functional silane ( 3 amino propyltrimethoxysilane ) are shown in Figure 6 2 With regard to the fabrication timeline of the composite cylinders for the primary testing, the mix number shown in Table 6 2 represents the time sequence as the concrete mixture was prepared. Thu s, 6 specimens were fabricated at one time, and the time interval between each mix number was approximately one week. Since this work was conducted over a long period of time changes in temperature or humidity due to changes in weather varied over the cou rse of the study The surface condition of the concrete is highly dependent on the humidity and temperature of its environment, which could thus be varied since the concrete cylinders are cured in air for one week before the epoxy adhesive and/or the silan e coupling agent are applied. On the contrary, compared to the long term fabrication time for sample preparation for primary testing, the preliminary testing consumed a very short term period for fabrication, thus, the changes in weather were not considere d to have influenced the results The detailed impact of this factor on the primary mechanical testing results is discussed in the Result s section. Exposure Environments Exposure environments used for the preliminary slant shear testing consisted of immers ion into tap water at 50 C for 3 weeks. For the primary testing, they consisted of water immersion at temperatures of 30, 40, 50, and 60C for time periods of 4, 8, and 12 weeks Details of hygrothermal exposure conditions used for the preliminary and prim ary slant shear testing are shown in Table s 6 1 and 6 2

PAGE 113

113 Table 6 1. Cure and exposure conditions for preliminary slant shear testing. Cure Condition Exposure Conditions # of specimens Concrete Epoxy Type s Time (weeks) Temp. (C) W/ O Silane treatment W/ A mino Silane W/ Epoxy Silane 1~2 days in mold + 3 weeks in air 3 weeks in air Air ( c ontrol) 3 RT 3 3 3 Water i mmersion 3 5 0 3 3 3 Table 6 2. Cure and exposure conditions for main slant shear testing. Mix # Cure Condition Exposure Conditions # of sp ecimens Concrete Epoxy Types Time (weeks) Temp. (C) W/O Silane treatment W/ Epoxy Silane 1 1 week in mold + 3 weeks in lime solution + 1 week in air 3 weeks in air Air (control) 0 RT 3 3 2 12 RT 3 3 3 Water immersion 12 30 3 3 4 12 40 3 3 5 12 50 3 3 6 12 60 3 3 7 Air (control) 8 RT 3 3 8 Water immersion 8 30 3 3 9 8 40 3 3 10 8 50 3 3 11 8 60 3 3 12 Air (control) 4 RT 3 3 13 Water immersion 4 30 3 3 14 4 40 3 3 15 4 50 3 3 16 4 60 3 3 Slant She ar Testing To investigate t he effects of silane coupling agents on the durability of interfacial bonding properties between epoxy and concrete with hygrothermal exposure a slant shear mechanical testing method was employed based upon ASTM C882. T he streng th was determined by using the epoxy adhesive to bond the two equal sections of a 7.6cm15.2cm cyl inder, each section of which had a diagonally cast bonding area at a

PAGE 114

114 30 angle from vertical. A schematic of the slant shear testing method is shown in Figure 6 3 T he surfaces of the two concrete sections were pretreated by the silane coupling agent a nd bonded by the epoxy adhesive as described above, and were then compared to those without any pretreatment by the silane. The loading rate was 0.25 0.05 MPa/sec per ASTM C39 and the test was performed by determining the average compressive failure load of 3 rep To determine the compressive strength of the concrete itself, cylindrical concrete specimens were fabricated with the same dimension s of 7.6 cm 15.2cm The cure condition s w ere the same as those of the slant shear testing method. The same loading rate of 0.25 0.05 MPa/sec w as used for the determination of the compressive strength of the epoxy bonded cylindrical composite, based upon ASTM C39 Compared with the fabrication time line of the composite cylinders used for the primary slant shear testing the neat cylindrical concrete specimens were fabricated with the 6 th and 16 th mixes of the composite cylinders (Table 6 2) F igure 6 3 Schematic of slant shear testing method

PAGE 115

115 Analytical Method: FT IR Spectroscopy In order to identify the presence of silane coupling agent deposited on concrete Fourier transform infrared spectro scopy was used (FTIR; Nicolet Magna 760, Thermo E lectron Cooperation) in the d iffuse reflectance infrared Fourier transform (DRIFT) mode. The DRIFT spectra were recorded over the range of 400 4 0 00 cm 1 with a KBr beamsplitter and a DTGS detector using 64 scans at a resolution of 4 cm 1 For the sample pr eparation, powders of cyinders were obtained from parts of the sample unexposed to the epoxy or silane coupling agent, after the slant shear testing. The DRIFT spectra of the neat powder was compare d with that of the powder treated b y the silan e coupling agent with the same application procedure as used for composite cylinders No baseline correction or subtraction method was used on the spectra In this experiment, t he DRIFT spect r um for each powder sample mixed with KBr powder in a proportio n of 1:100 were obtained to facilitate qualitative measurement. To compare those spectra with the spectroscopic features of the silane coupling agent itself, the attenuated total reflection ( ATR ) mode with a KBR beamsplitter and a DTGS detector was utiliz ed. In this experiment, ZnSe was utilized as the internal reflection element with an aperture angle of 63 using 64 scans at a resolution of 4 cm 1 R esults Since this stud y focused on interfacial bond degradation between epoxy and concrete with environment al exposure, the strength of the concrete should be high enough to ensure that the concrete itself does not fail, thus forcing adhesive failure during loading This enables monitor ing of the changes in interfacial bond degradation. Table 6 3 shows the comp ressive failure loads of neat cylindrical concrete each of which had the same cure conditions as the concrete mixture used for the preliminary

PAGE 116

116 and primary slant shear testing respectively From the results they show higher failure loads compared to the adhesive failure loads of the composite cylinders for each test which will be described in the next section. Table 6 3. Failure loads from neat cylindrical concrete specimen compressive testing Specimens Failure Load (kN) Concrete used for p r elimin ary t esting Concrete used for m ain t esting Mix 6 Mix 16 Values Average Values Average Values Average #1 155.4 135. 5 362.1 352. 2 33 3 0 335. 7 #2 107. 1 342.7 331. 1 #3 137.8 351.7 343.0 #4 141.5 F igure 6 4 Results of preliminary slant s hear testing In order to identify how each coupling agent affect s the bond durability the changes in failure load from the preliminary slant shear testing was ob serv ed with a single exposure condition of water immersion at 5 0C for 3 weeks, compared to c ontrol

PAGE 117

117 specimens ex posed in air for the same time period As shown in Figure 6 4 it is obvious that interfac ial bond degradation resulted in significant decrease in the durability of the bond strength with hygrothermal exposure. The control specimens show no difference in failure loads with type of silane w hen the failure load for the control system of unexposed specimens without any silane application is compared to that of the exposed specimens, over a 40% residual strength decrease was seen after hygrot hermal exposur e This difference appears to be significant, as the failure loads for the control (no silane) and exposed (no silane) specimens are different by more than one standard deviation (as shown from the error bars). Figure 6 4 also clearly shows t hat the application of epoxy functional silane coupling agent leads to a significant enhancement of durability with hygrothermal exposure. When the failure loads of the exposed specimens for three groups, which varied with different surface conditions, are compared with each other, it is shown that the average value of the failure load was higher for the exposed specimens modified by the epoxy functional silane. This difference between the failure loads for the exposed epoxy functionalized specimens and the exposed unfunctionalized specimens appears to be significant, as the difference is greater than one standard deviation. In the comparison of the average value of failure load, c ompared to a control group of the exposed specimens without any silane couplin g agent, which revealed 61.6% of residual strength with the exposure, the specimens with the epoxy functional silane showed a significantly higher residual strength (92.8%).

PAGE 118

118 Figure 6 5 Failure surfaces of control specimens for preliminary slant shear testing: (a) without silane treatment, (b) and (c) with amino functional silane treatment, (d) and (e) with epoxy functional silane treatment (photo courtesy of Sungwon Choi) The large deviation within the set of control specimens which employed the amino and epoxy functional silane treatment as sho wn by the error bars in Figure 6 4 was ascribed to the partial breaking of the concrete itself combined with mostly adhesive

PAGE 119

119 failure due to the lower strength of concrete (Figure 6 5 ) However, all the expos ed samples showed completely adhesive failure, and exhibited relatively low standard deviations which indicates the interfacial bond degradation with exposure (Figu re 6 6 ) resulting in lower standard deviations due to the more consistent adhesive failure mode Figure 6 6. Failures of exposed specimens for preliminary slant shear testing: (a) without silane treatment, (b) with amino functional silane treatment, (c) with epoxy functional silane treatment (photo courtesy of Sungwon Choi) On the other hand the effect of the amino fu n ctional silane on durability was negative, where the average residual strength was 56.3%, which is al mo st the same as the control. A possible explanation for this behavior can be found in the literature in

PAGE 120

120 which Harun et al co mpared the effect of an amino functional silane coupling agent used as an adhesion promoter for an epoxy coating on steel substrate on a pull off adhesion test with water immersion. 170 They found that the adhesion strength deteriorated rapidly with water immersion when the amino functional silane coupling agent was utilized, even compared to no silane treatment. In their analysis from angle resolved X r ay photoelectron spectroscopy, it was estimated that the coupling agent adsorbs onto the steel substrate with amine groups predominantly oriented towards the substrate while the silanol groups oriented upwards to the organic coating layer. Thus, the amine group was hydrogen bonded to the steel rather than the coupling agent forming the desired Si O Si linkage This hydrogen bond is considerably less resistant to hydrolysis, result ing in the lower failure strength. Another possible explanation to describe t he negative effect of the amino functional silane is the possibility of the formation of several silane layers which physically adsorbed on the concrete surfaces during the modification process. Since the multiple silane layers are usually cohesively weak, in such case the failure could easily occur through the silane layers with hydrolysis during water exposure, result ing in even lower failure strength. 35 In the primary slant shear testing, with the selected epoxy functional silane coupling agent, changes in failure load fr om the slant shear testing was monit ored with water immersion at different temperatures of 30, 40, 50, and 60C compared with control sp ecimens Exposure in these conditions was conducted over time periods of 4 8 and 12 weeks, and the changes in failure load with the specimens treated by the silane modification were compared to those without any modification In this slant shear

PAGE 121

121 testing, all the specimens showed adhesive failure, in which failure occurr ed between the epoxy and concrete surface at a maxim um failure load although some specimens were accompanied with small broken pieces of concrete (Figure 6 7 ) Figure 6 7 Failures for primary slant shear testing: (a) control specimens prior to exposure, (b) exposed at 30 C for 4 weeks, (c) exposed at 40 C for 12 weeks, (d) exposed at 50 C for 8 weeks, (e) exposed at 60 C for 8 weeks (photo courtesy of Sungwon Choi) Figure 6 8 shows the changes in failure load for the control specimens which were exposed to air after 3 weeks of epoxy adhesive curing at RT. For samples modified by the silane coupling agent, higher average values of failure load s than those without the modification throughout the entire exposure timeline are evident for the 0

PAGE 122

122 4 and 12 week exposure although the difference at 12 weeks i s less than one standard deviation and thus may not be significant Figure 6 8 Results of primary slant shear testing for control specimens. With regard s to the lower failure load at 4 and 8 week exposure, compared to the initial and the higher failur e load at 12 week exposure, some possible explanations can be suggested. Firstly, the concrete specimens for the 8 week exposure showed features of salt leaching out found at the outer surface of the specimen after exposure despite the same experimental co nditions with other specimens. We assume that different cure condition s caused by changes in the weather influenc ed the chemical nature of the concrete surfaces, wh ich induce d some abnormal phenomena of salt leaching out of concrete specimens, result ing in the lower strength. A nother possible description with respect to the much higher failure load at the 12 week exposure could be related to the properties of the epoxy resins which cured under ambient conditions. Since the tests were also conducted over rel atively short exposure time periods, post curing reaction of

PAGE 123

123 the utilized epoxy adhesive might produce some unusual behavior in which additional cure induces a property increase after the exposure. 152, 171 On top of that, the larger deviation of the failure load between each exposure time despite t he same experimental condition suggests that the changes in the chemical nature of the concrete surface caused by changes in fabrication environments could result in larger d ifference in interfacial bond properties. As described in the experimental section, each set of the specimens were fabricated at a different time, and include d curing in air for 1 week. Since this work was been conducted over a long time period, changes in temperature or humidity varied significantly over the fabrication time Thus, it is estimated that the fabrication environments strongly influenced the chemical nature of the concrete specimens, which resulted in a large deviation of the failure load A dditional testing would be needed to confirm this result. Figure 6 9 Results of primary slant shear testing for water immersion at 30C.

PAGE 124

124 Figure 6 10 Results of primary slant shear testing for water immersion at 40 C. As shown in Figure 6 9 the sp ecimens with silane modification and water immersion at 30C exhibit higher failure load compared to those without any modification with the exception of the 12 week exposure. For the 4 and 8 week exposure the differences between the modified and unmodif ied specimens is greater than one standard deviation. Figure 6 10 also shows that the application of the epoxy functional silane coupling agent leads to a significant enhancement of durability with hygrothermal exposure at 40 C The average values of the failure loads were higher for all exposure times As with the data for 40 C exposure, it appears that only the 4 and 8 week exposure times show significant differences (greater than one standard deviation). This might indicate a trend that the silane effec t diminishes as exposure time increases. A detailed discussion of this is given in the Discussion section.

PAGE 125

125 For the specimens exposed at 50 and 60 C, the average values of the failure loads clearly shows that the silane modification improves the interfacial bonding properties with the exception of the 12 week exposure at 60 C (Figures 6 1 1 and 6 1 2 ). Although the comparison of the average values for failure loads suggests the epoxy functional silane clearly contributes to the enhancement of the durability, t hese differences do not appear to be significant as they are all less than one standard deviation, other than the 4 week exposure at 50 C. Combined with the similar trend that the differences are less than one standard deviation for the samples with longer exposure times of 12 weeks at lower temperatures of 30 and 40 C, it appears that the effectiveness of the silane is reduced under relatively severe exposure conditions. The increase of the exposure temperature from 30 to 40 to 50 C did not result in large r degradation of failure loads in these short term exposure time periods, but, f or the samples at 60C a large degradation due to the higher temperature is evident (Figure 6 1 2 ) Figure 6 11 Results of primary slant shear testing for water immersion a t 50 C.

PAGE 126

126 Figure 6 12 Results of primary slant shear testing for water immersion at 60 C. D iscussion In this study, we observed that the surface modification of concrete surfaces using an epoxy functional silane coupling agent resulted in a significant increase in the durability of interfacial bonding properties between the epoxy adhesive and concrete. In order to evaluate the possible chemical changes of bonding before and after the silane application, DRIFT spectra were recorded for both the concrete p owder sample with the modification using the epoxy functional silane and with out the modification, and were compared with ATR spectra of the silane coupling ag ent itself. As shown in Figure 6 1 3 the DRIFT spectra for the concrete powders before the silane treatment shows the typical features of the material composed of cement and sand, in which the primary bands that appear are due to calcium hydroxide OH around 3643 cm 1 silanol groups around 1620 and 3450 cm 1 CaCO 3 around 1350~1550 cm 1 and SiO 4 arou nd 1055 cm 1 The detailed assignment of the infrared spectra is summarized in Table 6 4.

PAGE 127

127 Figure 6 13 Comparison of FT IR spectra: (a) F rom top to bottom: FT IR spectr a of concrete powder after the epoxy functional silane modification (record ed in DRIFT mode) and those before the silane modification (recorded in DRIFT mode), compared to those of the utilized silane coupling agent (recorded in ATR mode), (b) FT IR spectra for concrete powder from top picture is magnified in the range of 750~105 0cm 1 (top: after the modification, bottom: before the modification).

PAGE 128

128 Table 6 4. Possible assignment of DRIFT spectra Peak Position (Wabenumber, cm 1 ) DRIFT IR Assignment 695 Stretching vibration of Si O Si and Si O from quartz or coupling agent 172, 173 798 Bending mode of Si OH from quartz or coupling agent 173 175 858 Deformation mode of Al O or Al OH 176, 177 877 Asymmetric str etching vibration of CO 3 177, 178 915 Combination of C H stretching/bending of epoxide ring and Si OH stretching from GPS species 179, 180 1055 Asymmetric stretching vibration of Si O 181, 182 1101 Stretching vibration of S O 177, 182, 183 1350~1550 Asymmetric stretching vibration of CO from CaCO 3 or CO 2 176, 182 1620 Bending mode of OH from H 2 O 176, 177 1795 CaCO 3 177 2840~2940 Stretching vibration of CH, CH 2 184, 185 3450 Symmetric and asymmetric stretching vibration of OH 172, 186, 187 3643 Ca(OH) 2 which formed as silicate phases in the cement dissolve 172, 186, 187 Compared to the spectr um of the concrete powder without any silane treatment, the sample with silane treatment clearly shows the surface of the concrete powder was chemically modified with the utilized silane coupling agent. When comp ar ed to the ATR spectr um of the silane coupling agent, the new absorption peak s around 2840~2940 cm 1 are reasonably attributed to the deposition of the silane coupling agent on the concrete surfaces, due to the stretching vibration of C H in the alkoxy gr oup from the coupling agent. In addition, although the IR bands due to silanol, silicate, and SO 4 contained in concrete mixtures obscures the identification of Si O bond formation by the

PAGE 129

129 coupling agent around 455, 970, and 1055 cm 1 the increased absorpti on bands around 915 cm 1 which are due to c ombination of C H stretching / bending of epoxide ring and Si OH stretching from the species of the epoxy functional silane clearly indicates that the silane molecules have been successfully grafted onto the surfa ce of concrete powders. From the previous slant shear testing results, it is evident that the influences of chemical modification can be strongly affected by the fabrication environment due to the changes in humidity or temperature at the curing s tage of concrete, which obscure the actual contribution to the enhancement of the durability. In order t o exclude the effect of the fabrication environment and estimate the pure contribution of the silane coupling agent to the failure strength, the percent increas e of the failure load by virtue of the silane modification was evaluated. In this calculation the percent increase of the failure load due to the presence of the coupling agent at a given exposure time was calculated with the average values of failure loa d per specimen group for each mix number, in which the failure load of the composite cylinders with the silane modification was divided by that without silane modification In this calculation, the failure load was used for specimens with and without the coupling agent that had each undergone the same exposure (e.g. immersion in water at 30C ). Since each mix number represents the same fabrication time line, this method excludes the e ffect of the fabrication environment and evaluate s the silane contributi on of its own effect. As shown in Figure 6 1 4 the silane modification increased the failure load up to 180~190%, of the failure load with no silane treatment. In this figure, until the exposure time reached 8 weeks the percent value with silane modificat ion were over 100% for all

PAGE 130

130 the exposure conditions, which clearly indicates that the silane coupling agent significantly enhances the durability against water immersion. Furthermore, the increase of failure load was also observed for control samples which were exposed to air. Interestingly, Figure 6 1 4 also shows a trend that the silane effect diminishes as exposure time increases. Although the silane layer has a high permeability to water, providing a protective barrier against water absorption, continuous and severe water exposure induces the silane bonds of the substrate to be reactive with water, causing hydrolysis. 188 Thus, it can be estimated that the formation of silane bonds have been broken by hydrolysis with further exposure t ime which diminishes the ability of the silane to enhance durability. Figure 6 14 Percent increase of failure load for primary testing with the silane modification of concrete bonding surfaces. Evidence from this study suggests that enhanced bonding a t the interface results in improved durability. The improved adhesive properties might also be attributed to the

PAGE 131

131 barrier effect of blocking the diffusion of water due to the presence of the hydrophobic silane Given that hydrogen bonding is the sole candid ate for the description of chemical bonding for the interface between concrete and epoxies when there is no silane treatement, hydrogen bonding is most likely responsible for the degradation under water attack. Thus, it can be concluded that the observed r esults of the increase in durability with silane treatment confirms the hypothesis that the displacement of epoxy hydroxyl groups from the surface of the concrete by water disrupts the interfac ial bonds, resulting in a reduced interfacial bond strength. S ummary In this study, the effect of a silane coupling agent on the durability of bonding properties between epoxy and concrete was investigated. T his investigation intended to examine the hypothesis that hydrogen bonding plays a large role in the degradati on of interfacial bonding between epoxy and concrete, in which displacement of epoxy with water induces a disruption of the interfac ial hydrogen bond which results in reduced interfacial bond properties T hus the durability can be enhanced with the use of a coupling agent by changing the interfacial hydrogen bonds to stronger covalent bonding. From the slant shear testing, it is clear that the use of the epoxy functional silane coupling agent leads to significant improvement in durability for all the expos ure condition s which is attributed to the replacement of the weak hydrogen bonds between epoxy and cement by strong covalent bonds. Given that hydrogen bonding is the sole c andidate for the description of chemical bonding for the interface between concret e and epoxies when there is no silane treatement, disruption of this hydr ogen bond is responsible for the degradation under water attack.

PAGE 132

132 The spectroscopic features from FT IR confirmed the changes in chemical nature of concrete after the silane modificat ion. With the calculation of the percent contribution to the failure strength for each set of specimens, it is confirmed that the silane application contributes to improve the durability of bonding properties with exposure for all condition s but the effec t diminished as exposure time increase d

PAGE 133

133 CHAPTER 7 GENERAL CONCLUSIONS AND FUTURE WORKS For the first part of this work, the complex hygrotheraml behavior in epoxy amine thermosets was investigated, in which glass transition temperatures changed in a com plicated manner during hygrothermal exposure in terms that both increase and decrease in T g s were observed during the course of exposure. FT NIR spectroscopic studies with a model epoxy amine system demonstrated that increases in cross link density due to additional cure during the elevated temperature exposure led to an increase in properties. How ever, from the feature that the T g decreased at certain stages of the exposure, it is evident that plasticization by water occurred simultaneously, rendering the hygrothermal behavior to be more complex. Constructing the plot of T g versus conversion for the unexposed system provided an excellent method to consider contributions of other factors and exclude the factor of crosslink density. By applying the results fo it is possible to directly compare the T g values between exposed and unexposed samples while ruling out the factor of cross link density. The result indicates that the effect of other factors, such as different types of hydrogen bonded water and influence of the microstructure, are very small compared to the plasticization effect. The study about commercial epoxy products which used as a seal coat and impregnating resin for structural strengthening applications confirmed that the same complex hygrothermal behavior occurs, in which the increase in T g due to additional cure and the decrease in T g due to plasticization are in competition with each other during exposure despite sufficient cure time at RT before expos ure started.

PAGE 134

134 Actually, in order to take advantage of easy installation in outdoor or electronic applications, m any adhesives and coatings including epoxy amine thermosets are cured under ambient conditions, and thus are typically not fully cured. For these systems, changes in the environments such as an increase in temperatures or exposure to water can cause additional cure resulting in an increase in physical property while properties can decrease simultaneously due to plasticization, rendering this hygrot hermal behavior to be complex. Thus, the suggested mechanism of the two competitive factors and the method for the quantitative evaluation of those in separation will contribute to the understanding of the complex behavior having property increase and decr ease at the same time for various cross linked polymer areas employing low temperature curing under environmental effects, which will also be very useful in designing a material by providing a new method to suggest a materials specification in standards. A t the second part of this work, the effect of hydroxyl from water on cure kinetics was investigated with the hypothesis that water can significantly accelerate the cure kinetics of epoxy amine systems. From the study of the complex hygrothermal behaviors, it shows that the post curing reaction of the epoxy amine thermosets was accelerated by water. Thus, this acceleration renders the effect of additional cure is no longer small, finally introducing complex behavior despite sufficient cure time at RT before exposure started. Thus, the intent of this study was to evaluate the individual water effect on cure kinetics. Experimental data employing near FT IR spectroscopic analysis in this study clearly demonstrated that the cure reaction accelerated with the addi tion of a small

PAGE 135

135 amount of water, in which water acts as a strong catalyst To evaluate the hydroxyl effect from water separately in this study, a modified mechanistic modeling was utilized. In this modeling, the non catalytic kinetic term was neglected due to much slower reaction, and the catalytic kinetic term was separated into two different terms: (1) catalytic kinetic term by hydroxyl from auto catalysis (2) catalytic kinetic term by hydroxyl from added water. With the suggested modeling, it successfull y compared the kinetic values generated from catalyzed hydroxyl group by polymer chain with those by added water. The predicted kinetic scheme obtained from the calculated kinetic parameters using the suggested modeling well matched to the experimental val ues, which confirms that the proposed mechanism well describes the cure reaction of this system. From the comparison, it can be concluded that the catalyzed effect by hydroxyl from polymer chain is larger than that by hydroxyl from moisture or water. Also, this kinetic study confirmed the strong negative substitution effect in DGEBA J effamine system, which is mainly due to the steric crowding on the nitrogen atoms of Jeffamine. For the last part of this study, role of chemical bonds in the durability of int erfacial bonding between epoxy and concrete during hygrothermal exposure was investigated with the application of silane coupling agents. In this attempt of the application of silane coupling agent, it was intended to examine the hypothesis that hydrogen b onding plays a larger role for the degradation of interfacial bonding at the interface between epoxy and concrete, in which displacement of epoxy with water induces the disruption of the interfacial hydrogen bonding, resulting in reduced interfacial bond p roperties, thus durability can be enhanced with the use of the coupling agent by changing the main type of bonds to stronger covalent bonding.

PAGE 136

136 The slant shear testing results show that the durability of the interfacial bonding was significantly enhanced w ith the use of a epoxy functional silane coupling agent, which is attributed to the replacement of the weak hydrogen bonds between epoxy and cement by strong covalent bonds. The result s proposed that the bonding systems with epoxy concrete interface also h ave the similar mechanism proposed for the description used for the degradation in epoxy metal adhesion under humid environments, in which a decrease in bond strength have been ascribed to displacement of epoxy with water, which preferentially forms hydrog en bonds with the adherend, the hydrogen bonding at the epoxy concrete interface is a factor to affect the durability in the bonding of epoxy ceme n titous materials Thus, it can be conclude that the chemical nature of bonding plays a n important role at the interface bonding of epoxy resin for the structural strengthening application. The spectroscopic features using FT IR also confirmed the changes in chemical nature of concrete after the silane modification. With the calculation of the percent contribution to the failure strength for each set of specimens, it is confirmed that the silane application contributes to improve the durability of bonding properties with exposure for all exposure condition, but over time the improvement diminished As in any investigation through this work there are several issues which remain unanswered. First of all, from the studies of complex hygorthermal behavior of epoxy amine thermosets (Chapters 3 and 4), although it was revealed that the influences of other factors such as different states of hydrogen bonded water or microstructural effect are very small the exact impact of these factors remains still unanswered. In order to evaluate the effects of those factors quantitatively it would be useful if the decrease in

PAGE 137

137 T g during hygrothermal exposure is correlated with those factors N e ar infrared spectroscop y over a wide range of wavenumbers 6000 ~7400cm 1 which employ s a subtraction method between dried and wet samples would be suggest ed as a method to identify different states of hydrogen bonded water during the exposure. Also, the quantitative analysis of microstructural changes with AFM tapping mode such as changes in relative fractions of hard and soft ph ases during exposure can suggest an indicator which correlate the hygrothermal behavior with the effect of microstructures in epoxy amine thermosets. For a practical issue which need to be addressed from the kinetic stud y in Chapter 5 in situ investigatio n of cure kinetics using IR spectroscopy would be helpful to monitor the cure reactions precisely. When the changes in rate reaction can be obtained in situ IR spectroscopy, sufficient quantities of data with the limited effect of residual heat enable to m inimize experimental error s which can largely affect the calculation of concentration of tertiary amines using mass balance equations. For another issue relating to the work about the enhancement of the durability of bonding properties between epoxies and cementitious materials (Chapter 6) it is suggested that if surface roughness is also varied with different sample conditions, and different types of silane coupling agents are utilized in which different chemical end s are create d with different amount of hydroxyl s at the surface the fractional contribution of each mechanism of the chemical bonding and the mechanical interlocking could be assessed separately.

PAGE 138

138 LIST OF REFERENCES (1) Chen, J. S.; Ober, C. K.; Poliks, M. D.; Zhang, Y.; Wiesner, U.; Cohen, C. Polymer 2004, 45, 1939 1950. (2) Ding, Y.; Liu, M.; Li, S.; Zhang, S.; Zhou, W. F.; Wang, B. Macromol. Chem. Phys. 2001, 202, 2681 2685. (3) Ratna, D., Handbook of Thermoset Resins Smithers Rapra Technology: 2009. (4) Chen, J.; Naka mura, T.; Aoki, K.; Aoki, Y.; Utsunomiya, T. J. Appl. Polym. Sci. 2001, 79, 214 220. (5) Chiao, L.; Lyon, R. E. J. Compos. Mater. 1990, 24, 739 752. (6) Raman, V. I.; Palmese, G. R. Macromolecules 2005, 38, 6923 6930. (7) Shechter, L.; Wynstra, J.; Kurkjy, R. P. Ind. Eng. Chem. 1956, 48, 94 97. (8) Tung, C. Y. M.; Dynes, P. J. J. Appl. Polym. Sci. 1982, 27, 569 574. (9) Varley, R. J. Polym. Int. 2004, 53, 78 84. (10) Hale, A.; Macosko, C. W.; Bair, H. E. J. Appl. Polym. Sci. 1989, 38, 1253 1269. (11) Liu, X D.; Sudo, A.; Endo, T. J. Polym. Sci., Part A: Polym. Chem. 2011, 49, 250 256. (12) Omrani, A.; Simon, L. C.; Rostami, A. A.; Ghaemy, M. Int. J. Chem. Kinet. 2008, 40, 663 669. (13) Wartewig, S.; Helmis, G.; Strehmel, V.; Zimmermann, E.; Husler, K.; Fed tke, M. Progr. Colloid Polym. Sci. 1992, 90, 206 208. (14) May, C. A., Epoxy resins: chemistry and technology CRC Press: 1988. (15) Wu, L.; Hoa, S. V.; Ton That, M. T. J. Appl. Polym. Sci. 2004, 92, 2261 2268. (16) Carter, H. G.; Kibler, K. G. J. Compos. Mater. 1977, 11, 265 275. (17) De'Neve, B.; Shanahan, M. E. R. Polymer 1993, 34, 5099 5105. (18) Ellis, T. S.; Karasz, F. E. Polymer 1984, 25, 664 669. (19) Jelinski, L. W.; Dumais, J. J.; Cholli, A. L.; Ellis, T. S.; Karasz, F. E. Macromolecules 1985, 18, 1091 1095. (20) LaPlante, G.; Lee Sullivan, P. J. Appl. Polym. Sci. 2005, 95, 1285 1294.

PAGE 139

139 (21) McKague, E. L.; Reynolds, J. D.; Halkias, J. E. J. Appl. Polym. Sci. 1978, 22, 1643 1654. (22) Moy, P.; Karasz, F. E. Polym. Eng. Sci. 1980, 20, 315 319. (23) Pe rrin, F. X.; Nguyen, M. H.; Vernet, J. L. Eur. Polym. J. 2009, 45, 1524 1534. (24) Theocaris, P. S.; Kontou, E. A.; Papanicolaou, G. C. Colloid Polym. Sci. 1983, 261, 394 403. (25) Xiao, G. Z.; Delamar, M.; Shanahan, M. E. R. J. Appl. Polym. Sci. 1997, 65, 449 458. (26) Xiao, G. Z.; Shanahan, M. E. R. J. Appl. Polym. Sci. 1998, 69, 363 369. (27) Apicella, A.; Nicolais, L. Adv. Polym. Sci. 1985, 72, 69 77. (28) Lu, M. G.; Shim, M. J.; Kim, S. W. J. Appl. Polym. Sci. 2001, 81, 2253 2259. (29) Nogueira, P.; Ra mirez, C.; Torres, A.; Abad, M. J.; Cano, J.; Lopez, J.; Lopez Bueno, I.; Barral, L. J. Appl. Polym. Sci. 2001, 80, 71 80. (30) Nunez, L.; Villanueva, M.; Fraga, F.; Nunez, M. R. J. Appl. Polym. Sci. 1999, 74, 353 358. (31) Yang, Q.; Xian, G.; Karbhari, V. M. J. Appl. Polym. Sci. 2008, 107, 2607 2617. (32) Lobanov, Y. E.; Shterenzon, A. L. Mech. Compos. Mater. 1971, 7, 654 656. (33) Frigione, M.; Aiello, M. A.; Naddeo, C. Constr. Build. Mater. 2006, 20, 957 970. (34) de Bruyne, N. A. J. Appl. Chem. 1956, 6, 303 310. (35) Schmidt, R.; Bell, J. Adv. Polym. Sci. 1986, 75, 33 71. (36) Devasahayam, S. J. Appl. Polym. Sci. 2006, 99, 3318 3327. (37) Johncock, P. J. Appl. Polym. Sci. 1990, 41, 613 618. (38) Netravali, A. N.; Fornes, R. E.; Gilbert, R. D.; Memory, J. D. J. Appl. Polym. Sci. 1985, 30, 1573 1578. (39) Thomson, K. W.; Wong, T.; Broutman, L. J. Polym. Eng. Sci. 1984, 24, 1270 1276. (40) Wu, L.; Hoa, S. V.; Minh, T.; Ton, T. J. Appl. Polym. Sc i. 2006, 99, 580 588. (41) Papanicolaou, G. C.; Kosmidou, T. V.; Vatalis, A. S.; Delides, C. G. J. Appl. Polym. Sci. 2006, 99, 1328 1339.

PAGE 140

140 (42) Zhou, J.; Lucas, J. P. Polymer 1999, 40, 5505 5512. (43) Zhou, J.; Lucas, J. P. Polymer 1999, 40, 5513 5522. (44) Gupta, V. B.; Drzal, L. T.; Rich, M. J. J. Appl. Polym. Sci. 1985, 30, 4467 4493. (45) Maggana, C.; Pissis, P. J. Polym. Sci., Part B: Polym. Phys. 1999, 37, 1165 1182. (46) Vanlandingham, M. R.; Eduljee, R. F.; Gillespie, J. W. J. Appl. Polym. Sci. 1999, 71, 787 798. (47) Chateauminois, A.; Chabert, B.; Soulier, J. P.; Vincent, L. Polym. Compos. 1995, 16, 288 296. (48) Kelley, F. N.; Bueche, F. J. Polym. Sci. 1961, 50, 549 556. (49) Browning, C. E. Polym. Eng. Sci. 1978, 18, 16 24. (50) Wright, W. W. Comp osites 1981, 12, 201 205. (51) Bellenger, V.; Verdu, J.; Morel, E. J. Mater. Sci. 1989, 24, 63 68. (52) Grave, C.; McEwan, I.; Pethrick, R. A. J. Appl. Polym. Sci. 1998, 69, 2369 2376. (53) Apicella, A.; Nicolais, L.; de Cataldis, C. Adv. Polym. Sci. 1985, 66, 189 207. (54) Carfagna, C.; Apicella, A.; Nicolais, L. J. Appl. Polym. Sci. 1982, 27, 105 112. (55) Feng, J. "Interaction and permeability of water with liquid crystalline thermoset" Ph.D. Thesis, University of Florida, Gainesville, FL, 2001. (56) Ad amson, M. J. J. Mater. Sci. 1980, 15, 1736 1745. (57) Soles, C. L.; Chang, F. T.; Bolan, B. A.; Hristov, H. A.; Gidley, D. W.; Yee, A. F. Journal of Polymer Science Part B: Polymer Physics 1998, 36, 3035 3048. (58) Soles, C. L.; Chang, F. T.; Gidley, D. W. ; Yee, A. F. J. Polym. Sci., Part B: Polym. Phys. 2000, 38, 776 791. (59) Soles, C. L.; Yee, A. F. Journal of Polymer Science Part B: Polymer Physics 2000, 38, 792 802. (60) Mikols, W. J.; Seferis, J. C.; Apicella, A.; Nicolais, L. Polym. Compos. 1982, 3, 118 124. (61) Ngai, K. L.; Rendell, R. W.; Yee, A. F.; Plazek, D. J. Macromolecules 1991, 24, 61 67.

PAGE 141

141 (62) Anderson, S. L.; Grulke, E. A.; DeLassus, P. T.; Smith, P. B.; Kocher, C. W.; Landes, B. G. Macromolecules 1995, 28, 2944 2954. (63) Chang, Y. P.; Che ah, P. B.; Seow, C. C. J. Food Sci. 2000, 65, 445 451. (64) Liu, Y.; Roy, A. K.; Jones, A. A.; Inglefield, P. T.; Ogden, P. Macromolecules 1990, 23, 968 977. (65) Maeda, Y.; Paul, D. R. Journal of Polymer Science Part B: Polymer Physics 1987, 25, 1005 1016 (66) Roy, A. K.; Inglefield, P. T.; Shibata, J. H.; Jones, A. A. Macromolecules 1987, 20, 1434 1437. (67) Tiemblo, P.; Martinez, G.; Gmez Elvira, J. M.; Milln, J. Polym. Bull. (Berlin) 1994, 32, 353 359. (68) Vrentas, J. S.; Duda, J. L.; Ling, H. C. Ma cromolecules 1988, 21, 1470 1475. (69) Venditti, R. A.; Gillham, J. K. Polym. Mater. Sci. Eng. 1993, 69, 434 5. (70) Couchman, P. R.; Karasz, F. E. Macromolecules 1978, 11, 117 119. (71) Pascault, J. P.; Williams, R. J. J. J. Polym. Sci., Part B: Polym. Ph ys. 1990, 28, 85 95. (72) Wise, C. W.; Cook, W. D.; Goodwin, A. A. Polymer 1997, 38, 3251 3261. (73) Mijovic, J.; Andjelic, S. Macromolecules 1995, 28, 2787 2796. (74) Xu, L.; Fu, J. H.; Schlup, J. R. Ind. Eng. Chem. Res. 1996, 35, 963 972. (75) Byrne, C. A.; Hagnauer, G. L.; Schneider, N. S.; Lenz, R. W. Polym. Compos. 1980, 1, 71 76. (76) Cuadrado, T. R.; Macgregor, J. F.; Hamielec, A. E. J. Appl. Polym. Sci. 1990, 40, 867 890. (77) Regnier, N.; Fayos, M.; Moreau, P.; Lafontaine, E.; Mortaigne, B. Polym. Adv. Technol. 1999, 10, 637 646. (78) Riccardi, C. C.; Williams, R. J. J. J. Appl. Polym. Sci. 1986, 32, 3445 3456. (79) Xu, L.; Fu, J. H.; Schlup, J. R. J. Am. Chem. Soc. 1994, 116, 2821 2826. (80) Xu, L.; Schlup, J. R. J. Appl. Polym. Sci. 1998, 67, 895 901. (81) Dusek, K.; Ilavsky, M.; Lunak, S. J. Polym. Sci., Polym. Sypm. 1975, 53, 29 44.

PAGE 142

142 (82) Girard Reydet, E.; Riccardi, C. C.; Sautereau, H.; Pascault, J. P. Macromolecules 1995, 28, 7599 7607. (83) Horie, K.; Hiura, H.; Sawada, M.; Mita, I.; Kambe, H. J. Polym. Sci., Part A: Polym. Chem. 1970, 8, 1357 1372. (84) Riccardi, C. C.; Adabbo, H. E.; Williams, R. J. J. J. Appl. Polym. Sci. 1984, 29, 2481 2492. (85) Sung, C. S. P.; Pyun, E.; Sun, H. L. Macromolecules 1986, 19, 2922 2932. (86) Perrin, F. X.; Ng uyen, T. M. H.; Vernet, J. L. Macromol. Chem. Phys. 2007, 208, 55 67. (87) Mijovic, J.; Fishbain, A.; Wijaya, J. Macromolecules 1992, 25, 979 985. (88) Parker, R. E.; Isaacs, N. S. Chem. Rev. (Washington, DC, U. S.) 1959, 59, 737 99. (89) Zvetkov, V. L. Ma cromol. Chem. Phys. 2002, 203, 467 476. (90) Zvetkov, V. L.; Krastev, R. K.; Samichkov, V. I. Thermochim. Acta 2008, 478, 17 27. (91) Skordos, A. A.; Partridge, I. K. Polym. Eng. Sci. 2001, 41, 793 805. (92) Um, M. K.; Daniel, I. M.; Hwang, B. S. Compos. S ci. Technol. 2002, 62, 29 40. (93) Dusi, M. R.; Lee, W. I.; Ciriscioli, P. R.; Springer, G. S. J. Compos. Mater. 1987, 21, 243 261. (94) Kenny, J. M. J. Appl. Polym. Sci. 1994, 51, 761 764. (95) Michaud, D. J.; Beris, A. N.; Dhurjati, P. S. J. Compos. Mate r. 1998, 32, 1273 1296. (96) Williams, R. J. J.; Benavente, M. A.; Ruseckaite, R. A.; Churio, M. S.; Hack, H. G. Polym. Eng. Sci. 1990, 30, 1140 1145. (97) Karkanas, P. I.; Partridge, I. K. J. Appl. Polym. Sci. 2000, 77, 1419 1431. (98) Karkanas, P. I.; Pa rtridge, I. K.; Attwood, D. Polym. Int. 1996, 41, 183 191. (99) Mijovic, J.; Fishbain, A.; Wijaya, J. Macromolecules 1992, 25, 986 989. (100) Mijovic, J.; Wijaya, J. Macromolecules 1994, 27, 7589 7600. (101) Paz Abuin, S.; Lopez Quintela, A.; Pellin, M. P. ; Varela, M.; Prendes, P. J. Polym. Sci., Part A: Polym. Chem. 1998, 36, 1001 1016.

PAGE 143

143 (102) Chike, K. E.; Myrick, M. L.; Lyon, R. E.; Angel, S. M. Appl. Spectrosc. 1993, 47, 1631 1635. (103) Dannenberg, H. SPE Trans. 1963, 3, 78 88. (104) Qiao, P.; Xu, Y. J. Compos. Constr. 2004, 8, 352 359. (105) Aiello, M. A.; Frigione, M.; Acierno, D. J. Mater. Civ. Eng. 2002, 14, 185 189. (106) Au, C.; Buyukozturk, O. J. Compos. Constr. 2006, 10, 35 47. (107) Chajes, M. J.; Thomson, T. A.; Farschman, C. A. Constr. Build. Mater. 1995, 9, 141 148. (108) El Hawary, M.; Al Khaiat, H.; Fereig, S. Cem. Concr. Res. 2000, 30, 259 266. (109) Karbhari, V. M.; Engineer, M. J. Reinf. Plast. Compos. 1996, 15, 1194 1216. (110) Karbhari, V. M.; Engineer, M.; Eckel Ii, D. A. J. Mater. Sci 1997, 32, 147 156. (111) Toutanji, H. A.; Gomez, W. Cem. Concr. Compos. 1997, 19, 351 358. (112) Wan, B.; Petrou, M. F.; Harries, K. A. J. Reinf. Plast. Compos. 2006, 25, 875 890. (113) Tu, L.; Kruger, D. ACI Struct. J. 1996, 93 26 35. (114) El Hawary, M.; Al Khaiat, H.; Fereig, S. Cem. Concr. Compos. 1998, 20, 41 52. (115) Sen, R.; Shahawy, M.; Mullins, G.; Spain, J. ACI Struct. J. 1999, 96 906 914. (116) Fowkes, F. M. Ind. Eng. Chem. 1964, 56, 40 52. (117) Fowkes, F. M. J. Adhes. Sci. Technol. 1990, 4 669 691. (118) Krump, H.; Hudec, I.; Luyt, A. S. Int. J. Adhes. Adhes. 2005, 25, 269 273. (119) Mei, L.; He, X.; Li, Y.; Peng, Q.; Wang, R.; Xu, J. Surf. Interface Anal. 2010 n/a. doi: 10.1002/sia.3560. (120) Owens, D. K.; Wendt, R. C. J. Appl. Polym. S ci. 1969, 13, 1741 1747. (121) Bakis, C. E.; Uppuluri, V. S.; Nanni, A.; Boothby, T. E. Compos. Sci. Technol. 1998, 58, 1307 1319. (122) Lin, Y.; Friedrich, K.; Weimer, C.; Mai, Y. W. Adv. Compos. Mater. 1998, 7, 47 61.

PAGE 144

144 (123) Momayez, A.; Ehsani, M. R.; Ra mezanianpour, A. A.; Rajaie, H. Cem. Concr. Res. 2005, 35, 748 757. (124) Merlin, F.; Guitouni, H.; Mouhoubi, H.; Mariot, S.; Vall, F.; Van Damme, H. J. Colloid Interface Sci. 2005, 281, 1 10. (125) Perruchot, C.; Chehimi, M. M.; Vaulay, M. J. h.; Benzarti K. Cem. Concr. Res. 2006, 36, 305 319. (126) Viallis Terrisse, H. e.; Nonat, A.; Petit, J. C. J. Colloid Interface Sci. 2001, 244, 58 65. (127) Ramrus, D. A.; Berg, J. C. J. Adhes. Sci. Technol. 2006, 20, 1615 1623. (128) Ji, W. G.; Hu, J. M.; Liu, L.; Z hang, J. Q.; Cao, C. N. Surf. Coat. Technol. 2007, 201, 4789 4795. (129) Witucki, G. L. J. Coating Tech. 1993, 65, 57 60. (130) Kinloch, A.; Korenberg, C.; Tan, K.; Watts, J. J. Mater. Sci. 2007, 42, 6353 6370. (131) Rattana, A.; Abel, M. L.; Watts, J. F. J. Adhes. 2005, 81, 963 988. (132) Ji, W. G.; Hu, J. M.; Zhang, J. Q.; Cao, C. N. Corros. Sci. 2006, 48, 3731 3739. (133) Lassila, L. V. J.; Nohrstr, T.; Vallittu, P. K. Biomaterials 2002, 23, 2221 2229. (134) Oksanen, C. A.; Zografi, G. Pharm. Res. 1993, 10, 791 799. (135) Sung, Y. K.; Gregonis, D. E.; Russell, G. A.; Andrade, J. D. Polymer 1978, 19, 1362 1363. (136) Tan, Y. Y.; Challa, G. Polymer 1976, 17, 739 740. (137) Birkinshaw, C.; Buggy, M.; Daly, S. Polym. Commun. 1987, 28, 286 8. (138) Hodge, R. M .; Bastow, T. J.; Edward, G. H.; Simon, G. P.; Hill, A. J. Macromolecules 1996, 29, 8137 8143. (139) Jin, X.; Ellis, T. S.; Karasz, F. E. J. Polym. Sci., Polym. Phys. Ed. 1984, 22, 1701 1717. (140) Lagaron, J. M.; Powell, A. K.; Bonner, G. Polym. Test. 200 1, 20, 569 577. (141) Pissis, P.; Apekis, L.; Christodoulides, C.; Niaounakis, M.; Kyritsis, A.; Nedbal, J. J. Polym. Sci., Part B: Polym. Phys. 1996, 34, 1529 1539. (142) Musto, P.; Mascia, L.; Ragosta, G.; Scarinzi, G.; Villano, P. Polymer 2000, 41, 565 574.

PAGE 145

145 (143) Musto, P.; Ragosta, G.; Mascia, L. Chem. Mater. 2000, 12, 1331 1341. (144) Musto, P.; Ragosta, G.; Scarinzi, G.; Mascia, L. J. Polym. Sci., Part B: Polym. Phys. 2002, 40, 922 938. (145) Fukuda, M.; Kawai, H.; Yagi, N.; Kimura, O.; Ohta, T. Polym er 1990, 31, 295 302. (146) Bockenheimer, C.; Fata, D.; Possart, W. J. Appl. Polym. Sci. 2004, 91, 369 377. (147) Fraga, F.; Castro Diaz, C.; Rodriguez Nunez, E.; MartInez Ageitos, J. M. Polymer 2003, 44, 5779 5784. (148) Fraga, F.; Payo, P.; Rodriguez Nun ez, E.; Martinez Ageitos, J. M.; Castro Daz, C. J. Appl. Polym. Sci. 2007, 103, 3931 3935. (149) Janssson, J.; Gedde, U.; Ribas, S. Progr. Colloid Polym. Sci. 1992, 87, 78 82. (150) Montserrat, S.; Corts, P.; Calventus, Y.; Hutchinson, J. J. Therm. Anal. Calorim. 1997, 49, 79 85. (151) Shi, X.; Fernando, B.; Croll, S. J. Coat. Technol. Res. 2008, 5, 299 309. (152) Choi, S.; Douglas, E. P. ACS Appl. Mater. Interfaces 2010, 2, 934 941. (153) Abdelkader, A. F.; White, J. R. J. Appl. Polym. Sci. 2005, 98, 254 4 2549. (154) Rigail Cede, A.; Sung, C. S. P. Polymer 2005, 46, 9378 9384. (155) Fu, J. H.; Schlup, J. R. J. Appl. Polym. Sci. 1993, 49, 219 227. (156) Bella, R.; Fenouillot, F. i.; Cassagnau, P.; Falk, L. Polymer 2007, 48, 6902 6912. (157) Rozenberg, B. A Adv. Polym. Sci. 1986, 75, 113 165. (158) Smith, I. T. Polymer 1961, 2, 95 108. (159) Deng, Y.; Martin, G. C. Macromolecules 1994, 27, 5147 5153. (160) Fraga, F.; Soto, V.; Rodriguez Nunez, E.; Martinez Ageitos, J.; Rodriguez, V. J. Therm. Anal. Calorim. 2007, 87, 97 100. (161) Fraga, F.; Vazquez, E. C.; Rodriguez Nunez, E.; Martinez Ageitos, J. M. Polym. Adv. Technol. 2008, 19, 1623 1628. (162) Paz Abuin, S.; Pellin, M. P.; Paz Pazos, M.; Lopez Quintela, A. Polymer 1997, 38, 3795 3804. (163) Barton, J. M Polymer 1980, 21, 603 606.

PAGE 146

146 (164) Ballinger, P.; Long, F. A. J. Am. Chem. Soc. 1960, 82, 795 798. (165) Dusek, K. Adv. Polym. Sci. 1986, 78, 1 59. (166) Matejka, L. Macromolecules 2000, 33, 3611 3619. (167) Strachota, A.; Whelan, P.; Kr, J.; Brus, J.; Urb anov, M.; Slouf, M.; Matejka, L. Polymer 2007, 48, 3041 3058. (168) Hag Elsafi, O.; Alampalli, S.; Kunin, J. Compos. Struct. 2001, 52, 453 466. (169) Khalifa, A.; Gold, W. J.; Nanni, A.; M.I, A. A. J. Compos. Constr. 1998, 2, 195 202. (170) Harun, M. K.; L yon, S. B.; Marsh, J. Prog. Org. Coat. 2003, 46, 21 27. (171) Karbhari, V. M.; Chin, J. W.; Hunston, D.; Benmokrane, B.; Juska, T.; Morgan, R.; Lesko, J. J.; Sorathia, U.; Reynaud, D. J. Compos. Constr. 2003, 7, 238 247. (172) Chollet, M.; Horgnies, M. Sur f. Interface Anal. 2011, 43, 714 725. (173) Sood, R. K.; Krishnamoorthy, S.; Khandelwal, B. L. Int. J. Cem. Compos. Lightweight Concr. 1984, 6, 93 98. (174) Martinez Ramirez, S.; Zamarad, A.; Thompson, G. E.; Moore, B. Build. Environ. 2002, 37, 933 937. (1 75) Zhang, L.; Glasser, F. P. Cem. Concr. Res. 2005, 35, 2252 2260. (176) Trezza, M. A.; Lavat, A. E. Cem. Concr. Res. 2001, 31, 869 872. (177) Ylmen, R.; Jaglid, U.; Steenari, B. M.; Panas, I. Cem. Concr. Res. 2009, 39, 433 439. (178) Kjellsen, K. O.; Jus tnes, H. Cem. Concr. Compos. 2004, 26, 947 956. (179) Chu, L.; Daniels, M. W.; Francis, L. F. Chem. Mater. 1997, 9, 2577 2582. (180) Wood, D. L.; Rabinovich, E. M. Appl. Spectrosc. 1989, 43, 263 267. (181) Lee, T. C.; Wang, W. J.; Shih, P. Y.; Lin, K. L. C em. Concr. Res. 2009, 39, 651 658. (182) Mollah, M. Y. A.; Yu, W.; Schennach, R.; Cocke, D. L. Cem. Concr. Res. 2000, 30, 267 273. (183) Hughes, T. L.; Methven, C. M.; Jones, T. G. J.; Pelham, S. E.; Fletcher, P.; Hall, C. Advn. Cem. Bas. Mat. 1995, 2, 91 104.

PAGE 147

147 (184) Iijima, M.; Sato, N.; Wuled Lenggoro, I.; Kamiya, H. Colloid Surf. A Physicochem. Eng. Asp. 2009, 352, 88 93. (185) Shanmugharaj, A. M.; Sabharwal, S.; Majali, A. B.; Tikku, V. K.; Bhowmick, A. K. J. Mater. Sci. 2002, 37, 2781 2793. (186) Delgad o, A. H.; Paroli, R. M.; Beaudoin, J. J. Appl. Spectrosc. 1996, 50, 970 976. (187) Yilmaz, B. e.; Olgun, A. Cem. Concr. Compos. 2008, 30, 194 201. (188) Wypych, G., Handbook of Materal Weathering 3rd ed.; ChemTee publishing: 2003.

PAGE 148

148 BIOGRAPHICAL SKETCH S ungwon Choi was born in Seoul, Republic of Korea in 1977. He was admitted to Korea University in 1998, and received his Bachelor of Engineering degree in materials science and engineering in 2004. He worked at Korea Institute of Science and Technology in 2 004, where he participated in the research about the organic inorganic hybrid thin films for the application of organic thin film transistors. He was admitted to the D epartment of M aterials S cience and E ngineering at the University of Florida and started P h. D program in 2006. Finally, he received his Ph.D. from the University of Florida in the spring of 2011.