1

CALIBRATION OF BOND DURABILITY FACTOR FOR EXTERNALLY BONDED CFRP SYSTEMS IN CONCRETE STRUCTURES

By

JOVAN TATAR

A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT

OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE

UNIVERSITY OF FLORIDA

2013

2

© 2013 Jovan Tatar

3

To my parents

4

ACKNOWLEDGMENTS

I would like to express sincere gratitude to my committee chair Dr Trey Hamilton,

for his continuous support during this research; his patience, guidance, motivation, and

overall extraordinary mentoring skills. Aside from influencing the development of my

technical skills, Dr. Hamilton took a great part in my personal development.

I would like to thank Dr. Gary Consolazio for serving on my committee and

providing valuable comments regarding the thesis. In addition, I would like to

acknowledge Dr. Chris Ferraro for his insights and help in the experimental phase of the

project, Dr. Kurt Gurley for his suggestions in statistical analysis of data; and Dr. Elliot

Douglas for his support in carrying out the project research activities.

I would also like to express thanks to undergraduate and graduate students that

helped in specimen fabrication and testing phases of the project: Alex Randell, James

McCall, Paige Blackburn, Zack Workman, Juan Ponce, Peter Whitfield, Anthony

Kiessling, Charbel Raad, Philip Strauss, Garrett Littlejohn, Brad Krar.

Financial support for this research project was provided by Florida Department of

Transportation (FDOT). I am grateful for the technical support provided by staff at FDOT

State Materials Office in Gainesville, FL: Patrick Carlton, Dale DeFord, Richard

DeLorenzo, and Bill Baumann, in particular.

I would like to thank my parents, Slavica and Stevo Tatar for their unconditional

love and support during the course of my studies.

Last, but definitely not the least, I would like to express sincere gratitude to Gino

Blanco, for being of immense emotional support; a great guide and true help during my

adjustment to American culture. Thanks to him, my years at University of Florida, so far,

have been filled with joy.

5

TABLE OF CONTENTS page

ACKNOWLEDGMENTS .................................................................................................. 4

LIST OF TABLES ............................................................................................................ 7

LIST OF FIGURES .......................................................................................................... 9

ABSTRACT ................................................................................................................... 13

CHAPTER

1 INTRODUCTION .................................................................................................... 15

2 RESEARCH SIGNIFICANCE AND MOTIVATION .................................................. 18

3 LITERATURE REVIEW .......................................................................................... 19

Bond of FRP Composites to Concrete .................................................................... 19 Bond Test Methods ................................................................................................. 20 Durability of Bond ................................................................................................... 21 Effects of High Temperature, Moisture, UV, Alkaline Environment and Cycling ..... 24

4 DURABILITY TESTING APPROACH ..................................................................... 59

5 DURABILITY TEST SPECIMEN DESIGN AND FABRICATION ............................. 61

Specimen Design .................................................................................................... 61 Three-point Bending Test Bond Strength Index ...................................................... 63 CFRP Composites .................................................................................................. 64 Epoxy Adhesives .................................................................................................... 66 Specimen Fabrication ............................................................................................. 68 Specimen Preparation ............................................................................................ 75 Durability Test Procedures ...................................................................................... 83

6 RESULTS AND DISCUSSION ............................................................................... 86

7 DATA ANALYSIS .................................................................................................... 91

NCHRP Database ................................................................................................... 91 Analysis of Data for CFRP Wet-layup Without Putty ............................................... 97

Data Subset 1 for Wet-layup Without Putty Analysis ...................................... 100 Data subset 2 for Wet-layup Without Putty Analysis ...................................... 106

CFRP Wet-layup With Putty (Composite C) .......................................................... 113 Data Subset 1 for Wet-layup With Putty Analysis ........................................... 115 Data subset 2 for Wet-layup With Putty .......................................................... 117

6

CFRP Laminates .................................................................................................. 122 Literature Data ...................................................................................................... 123

8 INFLUENCE OF FAILURE MODE ON BOND STRENGTH INDEX...................... 126

9 BOND DURABILITY FACTOR .............................................................................. 132

Environmental Reduction Factor and Bond Durability Factor ............................... 132 Characteristic Bond Durability Factor ................................................................... 133 Bond Durability Factor for Wet-layup Without Putty .............................................. 139 Bond Durability Factor for Wet-layup With Putty ................................................... 147

10 SUMMARY AND CONCLUSIONS ........................................................................ 154

11 FUTURE WORK ................................................................................................... 157

APPENDIX

TOLERANCE FACTORS FOR CHARACTERISTIC BDF ..................................... 158

LIST OF REFERENCES ............................................................................................. 159

BIOGRAPHICAL SKETCH .......................................................................................... 166

7

LIST OF TABLES

Table page 3-1 Distribution of failure modes ............................................................................... 54

3-2 Change in ultimate bond strength after 12 months of exposure ......................... 57

4-1 Summary of conditioning protocols ..................................................................... 59

5-1 Properties of composite systems (information provided by manufacturers) ....... 66

5-2 Adhesive A mechanical properties 72 hours post cure at 140˚ F (60˚ C) – as reported by the manufacturer ............................................................................. 66

5-3 Adhesive B mechanical properties (14 day cure at 73˚F (23˚C) and 50% R.H.) – as reported by the manufacturer ............................................................ 67

5-4 Adhesive C mechanical properties (based on cured samples at 72˚F (20˚C) and 40% R.H.) – as reported by the manufacturer ............................................. 67

5-5 Adhesive E mechanical properties (at 73˚F (23˚C) and 50% R.H.) – as reported by the manufacturer ............................................................................. 68

5-6 Exemplary 10,000 psi mix proportions ................................................................ 69

5-7 Exemplary 4,000 psi mix proportions .................................................................. 69

5-8 Concrete mechanical properties ......................................................................... 74

7-1 NCHRP exposure conditions .............................................................................. 93

7-2 Specifics of Group 1 and Group 2 test specimens.............................................. 94

7-3 Statistical models parameters........................................................................... 103

7-4 Statistical test results ........................................................................................ 103

7-5 Normal distribution parameters for 30˚C and 60˚C of Data subset 1 ................ 106

7-6 Statistical tests results for Group 1 data ........................................................... 108

7-7 Levene’s and F-test results .............................................................................. 110

7-8 ANOVA results ................................................................................................. 113

7-9 Statistical tests results ...................................................................................... 116

7-10 Statistical tests results ...................................................................................... 119

8

7-11 Theoretical normal distribution parameters for different groups of data from literature ........................................................................................................... 125

8-1 Comparison of failure loads and failure modes for control specimens .............. 130

9-1 Environmental reduction factors as per ACI 440.2R-08 .................................... 132

9-2 Results of ANOVA analysis .............................................................................. 138

9-3 Conditioning protocol subgroups for wet-layup without putty ............................ 141

9-4 Normal distribution parameters for conservative and overly conservative characteristic BDFk ........................................................................................... 145

9-5 Conditioning protocol subgroups for wet-layup with putty ................................. 149

9-6 Normal distribution parameters for conservative and overly conservative characteristic values ......................................................................................... 152

A-1 Tolerance factors, K (Hahn and Meeker 1991) ................................................. 158

9

LIST OF FIGURES

Figure page 1-1 Number of publications in Web of Science ......................................................... 17

3-1 Possible failure modes in FRP-concrete bond system ....................................... 20

3-2 a) direct pull-off (bottom) and direct torsion (top) test; and b) direct shear (α=0˚), peel (α=90̊) and mixed-mode (0̊

10

3-22 Beam flexure specimen ...................................................................................... 53

3-23 Direct shear test setup ........................................................................................ 55

3-24 Direct sehar test specimen ................................................................................. 57

3-25 Failure surfaces for normal strength concrete .................................................... 57

3-26 Failure surfaces for high strength concrete ........................................................ 58

5-1 Three-point bending test setup ........................................................................... 62

5-2 Loading modes in interfacial region .................................................................... 64

5-3 Fine aggregate, coarse aggregate and cement (left to right) .............................. 70

5-4 Darex AEA, Adva Cast 600 and WRDA 60 (left to right) .................................... 70

5-5 Beam production procedures.............................................................................. 71

5-6 Specimen sandblasting ...................................................................................... 75

5-7 Procedure for repair with Composite systems A, B and D .................................. 76

5-8 Procedure for repair with Composite C system .................................................. 79

5-9 Procedure for repair with Composite E system................................................... 81

5-10 Exposure of test specimens ............................................................................... 83

5-11 Concrete beam in test fixture .............................................................................. 84

5-12 a) Debonding failure; b) shear failure ................................................................. 85

6-1 Representative averaged load-displacement plots for control and exposed samples .............................................................................................................. 86

6-2 Failure modes: a) 0% adhesive; b) 50% adhesive; c) 100% adhesive ............... 87

6-3 Bond strength index: a) immersion in water at 30˚C; b) immersion in water at 60˚C; and c) RH=100% at 60˚C .......................................................................... 89

6-4 CFRP laminate failure modes: a) adhesive failure between CFRP laminate and epoxy; b) CFRP laminate decohesion ......................................................... 90

7-1 a) Wet-layup without putty; b) wet-layup with putty; c) CFRP laminates ............. 96

7-2 Histogram and density distribution of test data ................................................... 99

11

7-3 Q-Q plot for all test data ................................................................................... 100

7-4 Histogram and density distribution of test data ................................................. 100

7-5 Q-Q plot for test data ........................................................................................ 101

7-6 Statistical fit models for Data subset 1: probability density function (left) and cumulative distribution function (right) .............................................................. 104

7-8 Histogram and probability density estimate for samples exposed to 60˚C ....... 105

7-9 Comparison of probability density estimates for 30˚C and 60˚C of Data subset 1 ............................................................................................................ 106

7-10 Histogram and probability density estimate for Data subset 2 .......................... 107

7-11 Q-Q plot for Data subset 2 ................................................................................ 107

7-12 Normal distribution fit for Data subset 2 ............................................................ 108

7-13 Comparison of normal distribution models for two data subsets (wet-layup without putty) .................................................................................................... 109

7-14 Histogram and probability density estimate for wet-layup with putty ................. 113

7-15 Q-Q plot for wet-layup with putty ...................................................................... 114

7-16 Histogram and probability density estimate for Data subset 1 .......................... 115

7-17 Q-Q plot for Data subset 1 ................................................................................ 115

7-18 Normal distribution fit for Data subset 1 ............................................................ 116

7-19 Histogram and probability density estimate for Data subset 2 .......................... 117

7-20 Q-Q plot for Data subset 2 ................................................................................ 118

7-21 Normal distribution fit for Data subset 2 ............................................................ 119

7-22 Comparison of normal distribution models for two data subsets (wet-layup with putty) ......................................................................................................... 121

7-23 Histogram and probability density function for CFRP laminates ....................... 122

7-24 Small-scale data distribution (literature) ........................................................... 124

7-25 Large-scale data distribution (literature) ........................................................... 124

7-26 Comparison of distributions for different groups of data from literature ............ 125

12

8-1 Failure mode of concrete specimens ................................................................ 127

8-2 Cohesive failure mode (0% adhesive) .............................................................. 128

8-3 Adhesive failure mode (100% adhesive) .......................................................... 128

8-4 Partially adhesive failure mode (70% adhesive) ............................................... 129

8-5 Comparison of control to conditioned (60 days immersion in water at 60˚C) specimen load capacity for Composite B .......................................................... 131

9-1 Tolerance factor, K vs. number of tests, n ........................................................ 135

9-2 Illustration of bond strength index degradation with respect to time from NCHRP study ................................................................................................... 136

9-3 Illustration of bond strength index degradation with respect to time from FDOT study ...................................................................................................... 137

9-4 Characteristic values for wet-layup without putty excluding Group 1 exposure to water immersion to 30˚C through 60˚C ......................................................... 140

9-5 Whisker plot ...................................................................................................... 142

9-6 Whisker plots for each subgroup for conservative characteristic values ........... 144

9-7 Whisker plots for each subgroup for overly conservative characteristic values 144

9-8 Fitted cumulative distribution function for wet-layup without putty excluding Group 1 exposure to water immersion to 30˚C through 60˚C ........................... 146

9-9 Cumulative distribution functions for small-scale (left) and large-scale (right) data for wet-layup from literature ...................................................................... 147

9-10 Characteristic values for wet-layup with putty ................................................... 148

9-11 Whisker plots for each subgroup for conservative characteristic values ........... 150

9-12 Whisker plots for each subgroup for overly conservative characteristic values 151

13

Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science

CALIBRATION OF BOND DURABILITY FACTOR FOR EXTERNALLY BONDED CFRP

SYSTEMS IN CONCRETE STRUCTURES

By

Jovan Tatar

December 2013

Chair: Homer R. Hamilton III Major: Civil Engineering

Use of carbon fiber-reinforced polymer (CFRP) composites in civil infrastructure

has shown an increase in the past two decades. Externally bonded CFRP materials

(wet-layup and laminate) have found their application in repair of damaged concrete

structures, particularly bridges.

Lack of understanding and confidence in the long-term performance of externally

bonded FRP systems in concrete structures still inhibits the use of FRP composites to

repair bridges. The most significant issue with externally bonded FRP composites is

their susceptibility to degradation when exposed to moisture. One method to gage this

sensitivity to moisture is to accelerate the effect of the moisture by exposing the material

to high heat and moisture.

This research utilized small-beam specimens to study FRP-concrete bond

performance when subjected to accelerated conditioning environments (immersion in

water and exposure to high humidity at elevated temperatures). Small beams with

bonded CFRP reinforcement were conditioned and then tested to failure under three-

point bending. The bond strength index was determined by dividing the average

conditioned strength by the average control strength.

14

Test results from the present research were combined with other test data to

form a database of over 900 test results. By utilizing an apparent analogy of FRP-

concrete bonded systems to adhesive anchors, a bond durability factor that quantifies

loss in bond capacity due to accelerated conditioning is determined equivalently as

characteristic test value for adhesive anchors.

For the purpose of the analyses, and based on available data, it was determined

that all FRP systems may be split into three categories. Bond durability factors

corresponding to wet-layup without putty and wet-layup with putty were determined to

be 0.6 and 0.4, respectively. Bond durability factor for CFRP laminate was not

established due to a lack of prolonged exposure data corresponding bond failure; CFRP

laminate specimens failed prematurely due to low material quality. Analysis of failure

modes revealed a dependency between ultimate load and failure mode. Cohesive

failure modes allowed for higher ultimate loads than adhesive failure modes. Moreover,

it was noted that failure mode and ultimate strength is not directly a function of concrete

strength, but rather its porosity and surface preparation.

15

CHAPTER 1INTRODUCTION

Aging infrastructure in the US, bridges in particular, is in need of reliable,

economical, and fast repair method. Eleven percent of bridges in 2012 were classified

as structurally deficient (ASCE 2013 Report Card definition: “Bridges that require

significant maintenance, rehabilitation, or replacement.”), and 24.9% were functionally

obsolete (ASCE’s 2013 Report Card definition: “Bridges that no longer meet the current

standards that are used today.”). In Florida, situation is somewhat better, with 262 or

2.2% of bridges being structurally deficient and 1,764 or 14.7% functionally obsolete.

Application of externally bonded fiber reinforced polymer (FRP) composite

reinforcement is becoming one of the most popular repair techniques in aging bridges

due to ease of installation, and its cost-effectiveness. FRP composites are typically

composed of glass, Kevlar, aramid, or carbon fibers oriented in specific directions and

embedded in an epoxy matrix. Most-commonly used fibers in bridge repair and

strengthening is carbon.

CFRP composites are light-weight, resistant to corrosion and have a high

strength-to-weight ratio relative to traditional construction materials, such as steel.

Currently, there are two types of CFRP based on application methods: wet-layup CFRP

and CFRP laminate. Wet-layup CFRP application is performed by saturating a dry fiber

fabric with epoxy. The wet fabric is then placed on the concrete surface. Primers are

sometimes used to facilitate adhesion to the concrete substrate.

CFRP laminates are formed by pultrusion. In this process the fibers undergo

impregnation by epoxy resin and are pulled through a heated stationary die which

allows for epoxy polymerization. CFRP laminate is usually adhered to concrete surface

16

with paste epoxy with higher viscosity than epoxies normally used in wet-layup

applications. CFRP compositesare mostly used as flexural and shear reinforcement, or

as a confinement in columns.

ACI 440R-08 characterizes the typical failure modes for members strengthened

with bonded CFRP composites as follows:

1. Crushing of concrete in compression before yielding of reinforcing steel 2. Steel yielding followed by rupture of CFRP 3. Yielding of the steel in tension followed by concrete crushing 4. Concrete cover delamination 5. Debonding of the FRP from concrete substrate

From design standpoint failure modes 2 and 3 are favorable because they are

less brittle than the other failure modes. Failure mode 1 is of a brittle nature and is not

desirable. Finally, modes 4 and 5 are debonding modes and are the most common due

to high stress concentrations at the ends of external CFRP reinforcement and due to

development of stresses at the bond line as the flexural cracks open. CFRP debonding

is considered a premature failure mode and is not desirable.

Debonding of CFRP, and CFRP-concrete bond behavior has been an

increasingly popular research topic in the past decade. Figure 1-1 presents number of

published papers relating to search terms “FRP” and “FRP” AND “Bond” in Web of

Science achieve, showing increasing trends in the past years. This demonstrates high

level of interest in FRP composite research.

17

Figure 1-1. Number of publications in Web of Science

Year

No.

of p

ublic

atio

ns

1980 1985 1990 1995 2000 2005 2010 20150

100

200

300

400

500Search: "FRP"Search: "FRP" AND "Bond"

18

CHAPTER 2RESEARCH SIGNIFICANCE AND MOTIVATION

Short-term performance and behavior of FRP-repaired concrete systems has

been extensively studied. However, research on issues that may arise due to exposure

of FRP-concrete systems to highly aggressive environments.

Since use of CFRP composite reinforcement in concrete structures is a relatively

novel repair method information on long-term performance of these systems is scarce.

ACI 440R-07 indicates that research and development are needed in the

following areas:

1. Identification of appropriate environments for durability testing 2. Durability studies of externally bonded FRP repair or retrofit measures 3. Service life prediction of structures using FRP 4. Development of standardized test methods with an accent on durability testing 5. Design and construction guidelines and specifications

The above-mentioned facts motivated development of a research plan that would

address durability issues in FRP materials’ performance in concrete structures when

exposed to harsh environmental conditions. This thesis presents research that is

primarily concerned with the development of a bond durability factor. This factor would

quantify the loss in bond properties between CFRP (carbon fiber reinforced polymer)

and concrete over time, and eventually affect the design factors in ACI 440.2R-08.

Furthermore, the intent of presented work is to promote development of a standardized

test method for durability of FRP-concrete bond.

19

CHAPTER 3LITERATURE REVIEW

Bond of FRP Composites to Concrete

Bond between FRP and concrete is established through the epoxy adhesive.

Bond is formed by means of mechanical interlocking and chemical bonding.

Mechanical interlock is established by flow of epoxy into the holes, crevices and

pores of concrete substrate. After it cures, epoxy locks in mechanically to the surface.

However, due to plasticization effects in epoxy, caused by its exposure to water,

mechanical bond may be weakened. Plasticization is a change in the thermal and

mechanical properties of a given polymer which involves: (a) lowering of rigidity at room

temperature; (b) lowering of temperature, at which substantial deformations can be

effected with not too large forces; (c) increase of the elongation to break at room

temperature; (d) increase of the toughness (impact strength) down to the lowest

temperature of serviceability (Immergut and Mark 1965).

Chemical bond is established through hydrogen bonding. Hydrogen bond forms

as a result of interaction between positively charged hydrogen atom and highly

negatively charged atom like oxygen (O) or nitrogen (N). In general, hydrogen bond

represents a combination of electrostatic, covalent and Van der Vaals interaction, and

should not be mistaken with covalent bond. In case of concrete, hydrogen bonds are

established through oxygen atoms on concrete surface and hydrogen atoms of epoxy

hydroxyl groups. Once the bond is exposed to moisture, hydrogen bonds get replaced

by water molecules (Lefebvre 2000) causing chemical bond to degrade.

Furthermore epoxy structure is affected by temperatures higher than Tg which

causes epoxy to lose stiffness due to increased chain mobility. Tg (glass transition

20

temperature) is approximate midpoint of the temperature range over which reversible

change in an amorphous polymer, or in amorphous regions of a partially crystalline

polymer, to a rubbery or viscous condition from a glassy or hard condition

(ISO 22768:2006)

As a result of aforementioned behavior, distinguishing failure modes occur in

FRP-concrete joints. Failure mode that is the most common and inherent to dry ambient

condition is cohesive failure mode in concrete substrate (Figure 3-1). Due to exposure

to moisture failure mode shifts to the interface between concrete substrate and

adhesive (Figure 3-1). This failure mode is termed adhesive failure mode. Failure

modes corresponding to adhesive decohesion or adhesive failure between FRP and

adhesive are typically not experienced (Figure 3-1). FRP decohesion (Figure 3-1) is

usually not an issue and may occur in underreinforced members, due to development of

high interlaminar stresses, or due to exposure to aggressive environments.

Figure 3-1. Possible failure modes in FRP-concrete bond system

Bond Test Methods

Bond of FRP-concrete joints is usually a critical component of the repaired

system. Debonding of FRP from concrete substrate is most commonly adopted limit

state. A number of test methods was developed in an attempt to study the phenomenon

of debonding of FRP from concrete substrate. These tests can generally be separated

into two groups.

Cohesive failureAdhesive decohesion

FRP decohesion

Adhesive failureAdhesive failure betweenFRP and adesive

Concrete

Adhesive

FRP

21

Direct bond tests (Figure 3-2), where FRP is subjected to controlled loading

mode, consist of:

• Direct pull-off • Direct torsion • Direct shear pull-off • Peel test • Mixed-mode test

Indirect bond tests, that utilize flexural test setup to test the bond performance,

are usually performed as:

• Three-point bending beam tests • Four-point bending beam tests

The following literature review discusses different test setups in more detail.

a) b)

Figure 3-2. a) direct pull-off (bottom) and direct torsion (top) test; and b) direct shear (α=0˚), peel (α=90̊) and mixed-mode (0̊

22

specific material’s properties. When referred to durability in terms of concrete, one

usually assumes ability of the material to withstand deterioration caused by weather and

chemical exposure, or surface abrasion. For the purpose of this study, by bond

durability of FRP-to-concrete, its long-term resistance to aggressive environmental

conditions is implied (e.g. high temperature, high moisture levels, UV, alkali, etc.). The

following chapter presents findings from the literature related to the durability of FRP-to-

concrete joints.

Since a wide range of tests methods and test specimens at different size scales

are utilized to study the topic, the term “loss in capacity” needs to be defined. By “loss in

capacity” is considered the quantitative decay in exposed specimen’s ultimate failure

load, ultimate stress, or fracture energy, when compared to control specimen with the

same characteristics that was not conditioned in an aggressive environment. It should

be noted that loss in capacity, as defined here, may not be necessarily equivalent to the

loss in bond capacity. This is mainly dependent the specimen design. For example, in

the ultimate load capacity of reinforced concrete beams that are repaired with FRP, R/C

beam’s residual capacity provided by the steel reinforcement at the tension face will

have a portion in both exposed and unexposed FRP-reinforced specimen. In order to

quantify the loss of bond capacity, residual load capacity of R/C concrete beam ought to

be eliminated when calculating the loss in bond capacity. In other words, only the

portions of capacity above the dotted line in Figure 3-3 shall be compared, as they

represent the contribution of FRP. This approach assumes that residual capacity is

determined from unexposed specimen. It shall be regarded that this is a more

conservative approach, as opposed to using the exposed specimen residual capacity in

23

calculations. In case of concrete-only flexural specimens, assumption is that its capacity

to resist tensile stresses comes from FRP only (concrete contribution is neglected).

Therefore, specimen’s load capacity is assumed to be representative of FRP-concrete

bond strength.

Figure 3-3. Specimen capacity

Cap

acity

Decrease due to exposure

Effective strengthIncrease due to FRP

24

Effects of High Temperature, Moisture, UV, Alkaline Environment and Cycling

One of the first times that the issue of long-term performance of the composite-

to-concrete bond was mentioned in the literature was in paper by Saadatmanesh and

Ehsani (1990). However, amongst the first attempts to conduct tests that included

environmental considerations was by Xie et al. (1995). They subjected multiple

reinforced concrete specimens, with an addition of CFRP laminates at the beam soffit to

both accelerated and long-term environmental tests (Figure 3-4). The exposure

conditions were as follows: (1) 2 weeks water immersion at room temperature followed

by 10 days of drying at room conditions; and (2) heating in oven at 40˚C for a week

followed by refrigeration at -23˚C for a week, for a total of two months. Authors found

that specimens exposed to water experienced a slight increase in strength (about 2%),

which they credited to increased fracture toughness of epoxy due to plasticization. On

the other hand, specimens that were exposed to hot-cold cycles experienced a

decrease in strength of about 10%.

Figure 3-4. Three-point bending test setup

Chajes et al. (1995) studied the influence of aggressive environmental

conditioning on durability of bond between FRP and concrete. The concrete test

25

specimens measuring 1.5 x 1.1 x 13 in. (38.1 x 28.6 x 330 mm) were reinforced with

one threaded steel bar placed 0.75 in. (19.1 mm) from the compression face of the

beam. Graphite FRP strip was adhered to the beam with the same width as the test

specimen – 1.5 in. (38.1 mm). Beams were first covered with a calcium chloride solution

and then conditioned at: (a) 50 and 100 freeze/thaw cycles as per ASTM C672-84

(freezing at -17˚C for 16 h, followed by thawing at room temperature for 8 h); and (b) 50

and 100 wet/dry cycles (immersion in calcium chloride solution for 16 h, followed by

drying at room temperature for 8 h). After conditioning, beams were loaded until the

total failure in four point bending test setup. Authors recorded a loss in additional

capacity provided by FRP from control (not exposed samples) to the samples subjected

to 100 wet/dry cycles peaking at around 13%.

Figure 3-5. Four-point bending test setup

Further work in addressing an issue of long-term performance of FRP-concrete

bond was undertaken by Karbhari and Engineer (1996). They tested small concrete

beams (measuring 2 x 1 x 13 in. or 50.8 x 25.4 x 330.2 mm) reinforced with different

26

FRP materials (GFRP and CFRP) in four point bending tests. All beams were exposed

to the following environments for a period of 60 days before testing: (1) immersion in

water at 20˚C; (2) immersion in synthetic seawater at 20˚C; (3) freezing at -15.5˚C; and

(4) Freeze-thaw cycling (-15.5˚C for 24 hours followed by 50˚C for 24 hours). Control

samples were kept in ambient conditions at 20˚C. Test results showed the highest loss

in capacity in beams immersed in fresh and seawater, peaking at around 35%. The

minimum overall change was observed for specimens exposed to freezing at -15.5˚C.

According to the authors, the loss in bond capacity was due to epoxy’s susceptibility to

plasticization and increase in compliance caused by the water absorption. Moreover, a

significant drop in glass transition temperature (Tg) of resins was noticed after exposure

to continuous water ingress. According to the authors, the reduction in Tg due to water

absorption signifies a degradation in epoxy matrix properties which may lead to

decreased load carrying capacity.

Furthermore, in a separate study, Kabhari et al. (1997) subjected specimens

exposed to the same environmental conditions to a controlled mixed mode test (Figure

3-6). They measured interfacial fracture energy for each exposure condition as a

function of phase angle ψ (mode mixity parameter - I

II

KK1tan −=ψ ) and the angle at

which peel force was applied. It should be noted that the phase angle ψ=0˚ corresponds

to a pure Mode I conditions, and ψ=90˚ relates to pure Mode II loading. They observed

a clear difference in interfacial fracture energy magnitudes between the systems

exposed to water immersion and freeze/freeze-thaw cycles, with the former set of

specimens having higher values of interfacial fracture energy.

27

It was also noted that this difference was not as apparent in specimens

reinforced with carbon fiber as opposed to specimens with glass fibers. Additionally,

specimens reinforced with carbon fibers showed almost no change in interfacial fracture

energy in respect to phase angle. On the other hand, an increase in fracture energy of

up to 300% (from Mode I to Mode II) was observed for specimens with glass fibers.

Analysis of dependency of peel force, Mode I fracture energy (GI), Mode II fracture

energy (GII) and total fracture energy (G=GI+GII) to the peel angle showed that

exposure to water not only resulted in a decrease of peel force and interfacial energy,

but it also causes a shift in overall trends – variation of GII in respect to peel angle

changes from exponential (for ambient conditions) to linear dependency (after exposure

to water). According to the authors, this indicates a change in mechanisms during peel

due to exposure to water. In addition to this, authors observed a change in failure mode

from cohesive to adhesive. Specimens exposed to freezing and freeze-thaw conditions

showed an increase in peel force and interfacial fracture energy when compared to

control specimens.

28

Figure 3-6. Mixed mode peel test setup

Another study pertaining to long-term durability of concrete bonded with external

FRP in marine environments was performed by Toutanji and Gomez (1997). The test

specimen consisted of small concrete beam, measuring 2 x 2 in (51 x 51 mm) with a

total length of 14.4 in (365 mm), reinforced with either a CFRP or a GFRP laminate over

the full length of the beam. Three different adhesives were used to bond FRP sheets to

concrete. Test method consisted of loading the specimens in four-point bending at a

constant crosshead displacement rate until the failure. Specimens were exposed to 300

wet and dry cycles (hot air at 35˚C average and 90% relative humidity) in a salt

environment (35 g of salt per 1 liter of water). Control specimens were kept in standard

room conditions. Test results showed that there was a significant reduction in specimen

capacity ranging from 3 to 33%, depending on the type of fibers and epoxy used. The

loss in capacity was attributed to degradation of epoxy.

29

Figure 3-7. Four-point bending test setup

Beaudoin et al. (1998) performed a durability study on reinforced concrete beams

with external FRP. Beams measured 3.9 x 5.9 x 47.2 in. (100 x 150 x 1200 mm) and

were reinforced with two 0.26 in. (6.5 mm) diameter steel bars (0.1 in2 or 65 mm2) in

addition to stirrups that were placed to avoid shear failure. Beams were reinforced with

Mitsubishi Replak 20 and Sika CarboDur CFRP laminates too. Control specimens were

kept in dry laboratory conditions, while the rest of the samples were exposed to 13 wet-

dry cycles. Cycles consisted of immersion in water at 21˚C for five days, followed by

drying for two days at 27˚C. After conditioning beams were subjected to four-point

bending tests. Test results showed that beams reinforced with Replak 20 had a loss in

ultimate capacity of around 10%, while beams strengthened with CarboDur experienced

an increase in ultimate load capacity of approximately 10%. The loss in FRP bond

capacity was around 20% for Replak 20. CarboDur samples had an increase of 15% in

FRP bond capacity on average.

30

Figure 3-8. Four-point bending test setup

31

Sen et al. (1999) conducted another study on durability of FRP-concrete bond in

marine environments. Test specimens were prepared by bonding two types of carbon

fibers (either bidirectional woven fabric or unidirectional carbon fiber procured sheet) to

a concrete slab using five different epoxy systems. Concrete slabs were each 17.9 x

17.9 in. (455 x 455 mm) with the thickness varying between 2.95 and 3.74 in. (75 and

95 mm). Four different exposure conditions were investigated: (1) combined wet/dry

cycles and hot/cold cycles in 5% salt water for 17 months; (2) wet/dry cycles in 15% salt

water for 17 months; (3) outdoor conditions for 23 months; and (4) air conditioned

laboratory conditions (control) for 23 months. To measure the bond strength destructive

direct pull-off or direct torsion tests were performed. Test results showed that bond

degradation was least under outdoor exposure and greatest under wet/dry cycles.

Consequently, authors concluded that moisture absorption is potentially more damaging

to bond durability, than other environmental factors. Test results indicated that direct

pill-off test generally produced a bond failure at lower stresses than direct torsion test.

As stated by the authors, direct torsion test is, therefore, more appropriate for identifying

bond degradation in flexural applications.

32

a) b)

Figure 3-9. a) torsion test setup; b) pull-off test setup

33

Leung et al. (2001) evaluated environmental impacts on the flexural behavior of

reinforced concrete beam strengthened with CFRP. Beams were made of concrete and

had the following dimensions: 2.95 x 2.95 x 11.8 in. (75 x 75 x 300 mm). Test beams

were subjected to three-point bending until the failure. The following four exposure

conditions were introduced: (1) water immersion at 27˚C; (2) Wetting/drying cycle (water

immersion at 27˚C for half a week, followed by storage in a control room for half a week

– 25±2˚C and RH 65±2%); (3) constant moisture condition (25±2˚C and RH 65±2%);

and (4) heating/cooling cycle (oven at 60˚C for half a week, followed by storage in a

control room for half a week). Authors found that exposure to the aforementioned

environments caused changes in the concrete and the adhesive. Generally, plain

concrete specimens had higher failure loads with decreasing in moisture contents. Also,

longer exposure to the moist environments resulted in an increase in strength of plain

concrete beams. Finally, authors concluded that long-term exposure of CFRP reinforced

beams to highly moist environments affects the adhesive and leads to decrease in load-

carrying capacity and midspan deflection. Observed failure mode for beams reinforced

with CFRP was “shear failure with plate peel-off”.

Myers and Ekenel (2005) conducted a study that investigated the effects of

moisture and temperature of the concrete surface at time of installation on FRP-

concrete bond strength. Direct pull-off and direct torsion tests were performed in order

to identify the critical surface moisture content and R.H. of concrete. Additionally,

flexural tests were performed on precracked reinforced concrete beams to determine

the effects of temperature at installation on the performance of bond between FRP and

concrete. Authors found that specimens that were constructed with at a high surface

34

moisture contents exhibited poor bond performance. Furthermore, it was found that

specimens that were strengthened at a relative humidity higher than 82% may have

lower bond quality. When it comes to the effects of temperature, it was concluded that

the extremely low temperatures affected the bond adversely. However, installation of

FRP in high temperatures did not prove to affect the bond behavior.

a) b)

c)

Figure 3-10. a) flexural specimens; b) direct pull-off test setup; c) direct torsion test setup

Grace and Singh (2005) explored the effects of various environmental conditions

on the performance of reinforced concrete beams externally strengthened with CFRP

35

laminates and fabrics. Reinforced concrete beams specimens used in this study

measured 6 x 10 in. (152.4 x 254 mm) in cross-sectional dimensions and were 108 in.

(2743 mm) long. Specimens were exposed to the following environmental conditions:

(a) 100% humidity; (b) dry heat; (c) saltwater solution; (d) alkaline solution; (e) freeze-

thaw cycles; and (f) thermal expansion. Beams were then tested in four-point bending

by loading and unloading in two stages until the complete failure. All beams failed either

due to debonding of FRP or “onset of delamination (shear-tension failure)”. Again, the

highest loss in capacity (of up to 30%) was observed in beams that were exposed to

highly moist environments. Beams experienced either a smaller loss or an increase in

capacity due to exposure to other environmental conditions. Authors noted that CFRP

laminates are more susceptible to aggressive environmental conditions than the beams

reinforced with CFRP fabric. Furthermore, they observed that duration of exposure for

the beams exposed to humidity and saltwater solution had no significant influence on

beams that were reinforced with CFRP fabric, while beams strengthened with CFRP

laminates experienced further deterioration in strength due to the prolonged exposure.

36

Figure 3-11. Drawings of test specimen

37

Attempt in determining the influence of temperature only on the bond

performance of concrete externally strengthened with FRP was undertaken by Klamer

et al. (2005). They conducted both shear-lap tests and three point bending tests at

different environmental temperatures. Concrete specimens were strengthened with

CFRP laminate and paste epoxy (Adhesive B in this report) with a glass transition

temperature Tg=62˚C. Change in the failure mode, from cohesive to adhesive, was

observed for temperatures higher than 50˚C. In shear-lap specimens, an increase in

temperature (below the Tg) produced an increase in the bond capacity. Authors

concluded that the bond capacity is affected by the increase in temperature due to the

decrease in adhesive stiffness and reduction in adhesive strength (especially for

temperatures above the Tg). It should be regarded that the measured force in CFRP

was lower and the displacement was higher at higher temperatures in the three-point

bending test, which ultimately resulted in a higher specimen capacity. Authors explained

this by effects introduced by the test setup.

38

Figure 3-12. Double lap shear test setup (left); and three-point bending test setup (right)

39

Au and Buyukozturk (2006) performed a study on influence of moisture on the

bond behavior in peel (Mode I) and shear (Mode II) test configurations. Pre-cracked

peel and direct shear fracture specimens were conditioned at: (a) RH=100% at 23˚C;

and (b) RH=100% at 50˚C. Specimens were conditioned for 2, 4 and 8 weeks. Control

samples were kept in dry conditions. Results from these tests were presented in terms

of specialized fracture energy release rate, based on the tri-layer fracture model

developed by Au and Buyukozturk (2006). Under the peel conditions exposed

specimens experienced a sudden drop in bond capacity of around 60%. Shear fracture

specimens, however, experienced a more gradual drop in capacity peaking at around

50% loss. Higher temperature did not significantly affect the peel properties of test

specimens, whereas shear fracture specimens achieved lower capacities at higher

conditioning temperature. Additionally, Tuakta and Buyukozturk (2010) extended this

study to include specimens conditioned by immersion in water at same temperatures

and wet-dry cycling. They also explored the influence of drying before testing the

specimens (referred to as moisture reversal tests). They noted that decrease in fracture

energy of FRP-concrete bond may be up to 70%. Furthermore, they concluded that

bond properties cannot be fully regained after drying or successive wet-dry cycles.

Authors observed a shift in failure mode from cohesive to adhesive in all conditioned

samples.

40

Figure 3-13. Direct shear/peel test specimen

41

Study on assessment of quality of bond between FRP and concrete with the

presence of water at the time of installation was undertaken by Wan et at. (2006). They

used the modified double cantilever beam (MDCB) test to obtain the energy release rate

of FRP debonding when subjected to mixed mode loading conditions. Details on the test

setup and analytical model may be found in the referenced journal article. To simulate

the presence of water at time of installation of wet-layup CFRP, four different surface

moisture conditions were introduced: (a) dry; (b) saturated surface dry 1 (SSD1); (c)

saturated surface dry 2 (SSD2); and (d) wet. For dry surface condition specimens were

left in ambient conditions to cure before and after priming. For SSD1 condition

specimens were submerged in water for 3 days, followed by drying the surface with a

paper towel and then applying the primer. Additionally, specimens were submerged in

water after priming with the water level below the concrete surface. For SSD2 condition,

the same procedure was followed, however, specimens were left in ambient conditions

after priming. For the wet condition, primer was applied to the wet concrete surface

directly. Specimens were then submerged in water again. Based on the test data

authors concluded that the bond capacity decreases with the amount of water present

at the surface. Namely, loss in capacity for SSD2, SSD1, and wet specimens was 58,

38, and 8% of the specimens kept in dry conditions, respectively. The prevailing failure

mode for all conditioned specimens was mixed or adhesive. In addition to the previously

described test program, authors performed a series of tests to investigate the influence

of water on bond after FRP cures. The specimens were conditioned in water for 3, 6,

and 8 weeks and then subjected to MDCB test. Results showed that FRP-concrete

42

bond degrades in respect time, with a loss of up to about 75% in ultimate energy

release rate when compared to results for specimens kept in dry conditions.

Figure 3-14. Mixed mode peel test setup

Alfar (2006) studied the durability of CFRP strengthened reinforced concrete

members subjected to real-time exposure in environments typical for Amman city (at

Building Research Center of the Royal Scientific Society of Jordan), Dead Sea region

(salt extracting plant), and Aqaba region (splash and tidal zone) in Jordan. According to

the authors, the latter two locations provided exposure to some of the most severe

marine environments. Salinity of Dead Sea is close to 34% (340 parts per thousand)

compared to only around 3.5% (35 parts per thousand) that is the salinity of North

Atlantic ocean. Specimens were also subjected to high temperatures ranging from 37 to

49.5˚C in the summer and RH=80%. Additionally, some samples were exposed to

43

artificially created laboratory conditions that included: (a) wetting specimens with

chloride solution (9.6% NaCl) and storing them at RH=65% at 20˚C; and (b) RH=65% at

20˚C (control). Specimens consisted of reinforced concrete slabs measuring 63 x 19.7 x

4.7 in. (1600 x 500 x 120 mm) and concrete prisms measuring 5.9 x 5.9 x 17.7 in. (150

x 150 x 450 mm) in nominal dimensions.

Slab specimens were notched on both sides, at 200 and 300 mm from the

midspan. Moreover, after curing, each pair of slabs was tied together, and applied a

sustained load in excess of theoretical cracking load in order to produce cracking. Then,

pairs of slab specimens were conditioned in previously described environments. After

four months of exposure specimens were repaired with CFRP laminates and three

types of epoxies. Then, sustained load was increased by 20%, and specimens were

conditioned for additional 12 months. Eventually, slab specimens were tested in three

point bending test setup. Authors noted the highest loss in slab capacity in specimens

conditioned in Dead Sea environment of about 12%. Most of the specimens failed by

debonding of CFRP laminate caused by an intermediate flexural crack. Moreover, a

shift from cohesive to adhesive failure mode was observed in all specimens after

conditioning in severe environments.

Concrete prism specimens were used to perform direct shear test, and were

subjected to the same conditioning and repair protocol as slab specimens. Results from

direct shear tests showed that exposure to severe environments did not have

detrimental effect on FRP-concrete bond.

Frigione et al. (2006) studied the efficiency of bond in concrete joints adhered by

epoxy when affected by moisture. The slant shear tests were performed based on

44

ASMT C882-91. Two different concrete mixes (25 MPa and 50 MPa) and three different

epoxy adhesives used mainly applied to concrete members for different bonding

purposes. Test results for only one of the epoxies will be included, as that is the only

one that can be used to bond CFRP to concrete. Adhesive thickness in concrete-

concrete joint was varied at 0.5, 2, and 5 mm. Shear slant specimens were conditioned

in distilled water at 23±1˚C for 2, 7, 14, and 28 days before testing. Test results

indicated that the bond strength decay plateaus after 14 days of exposure, peaking at

around 35%. Relatively slight decrease in bond strength was observed as the epoxy

thickness was increased.

Figure 3-15. Slant shear test specimen

Fava et al. (2007) used the direct shear test in Mode II loading to assess the

performance of FRP-concrete bond after conditioning in multiple environments. Test

setup essentially follows the one described in Taljsten (1996) and Au and Buyukozturk

(2006). Test specimens were exposed to: (a) standard conditions – at 20˚C and

RH=60%; and (b) one month of salt spray fog at 50˚C. Authors noted an increase in

strength of around 30% in specimens conditioned in salt fog environment when

compared to the specimens kept in standard conditions. They attributed this rise in bond

45

capacity to “beneficial effects on epoxy resin of high humidity level” as found by Wu et

al. (2004).

Figure 3-16. Direct shear test setup

Soudki et al. (2007) tested 11 reinforced concrete beams repaired with CFRP

laminate or wet-layup. Eight beams were precracked before repairing, while the

remaining three beams were kept intact as a control. Each beam was 5.9 in. (150 mm)

wide, 9.8 in. (250 mm) deep and 94.5 in. (2,400 mm) long (Figure 3-17). Beams were

lightly reinforced, with a reinforcing ratio of 0.6%. Control beams were kept at a room

temperature, while the rest of the beams were conditioned in 100, 200 and 300 wet/dry

cycles with a 3% solution of NaCl. One wet/dry cycle lasted two days – one day of

wetting followed by one day of drying. This condition protocol was established to

achieve active corrosion of reinforcing steel in a reasonable time. After exposure

specimens were tested in four-point bending. In addition, corrosion rates of reinforcing

steel, and chloride contents at different depths were measured. Authors noted that all

specimens failed by debonding of FRP followed by a maximum loss in capacity of 11

and 28% for wet-layup sheets and laminates, respectively. Furthermore, CFRP and the

resin system seemed to decrease the corrosion rate of the reinforcing steel.

46

Figure 3-17. Four-point bending test beam design

Silva and Biscaia (2008) developed an experimental program to evaluate the

degradation of bond between FRP and concrete. They tested hinged concrete

specimens in four-point bending. Specimens were conditioned as follows: (a) salt fog

cycles at 35˚C – 16 h dry followed by 8 h of fog; and (b) moisture cycles – RH=20% for

12 h, followed by RH=90% for 12 h. Specimens from the first group were conditioned for

3000 and 6000 h, while the specimens in the second group were exposed for 1000,

5000 and 10,000 hours. Authors noted that failure mode was affected by the exposure

environment. They observed cohesive failure mode in specimens exposed to moisture

cycles, while the specimens conditioned in salt fog cycles experienced adhesive failure

along the interface. However, they noted that both groups of specimen had almost the

same reduction in load capacity of about 20%.

47

Figure 3-18. Hinged beam specimen

48

Garmage et al. (2009) performed a durability study of FRP-concrete bond using

75 x 75 x 250 mm3 reinforced with wet layup CFRP sheets. All specimens were

conditioned at temperature cycles ranging from 20 to 50˚C within 4.5 hours period, with

1.25 hours soaking time at minimum and maximum temperatures. Relative humidity

was kept constant at 90%. Some of the conditioned specimens were subjected to

different levels of sustained loading. As this study is not concerned with effect of

sustained loading on bond performance, findings related to those samples will be

omitted in the literature review. Control samples were not conditioned. Authors tested

specimens in single lap shear test setup after 175, 325, 1250, and 2400 hours of

conditioning and reported a reduction in ultimate load capacity peaking at close to 30%.

Banthia et al. (2010) performed a series of infrared thermography measurements

and pull-off tests on four concrete bridges in Canada that were repaired with FRP

materials. This literature review will include bridges repaired with CFRP only. The

following bridges were included in the study: (a) St. Etienne de Bolton Bridge near

Sherbrooke, QC “exposed to temperatures ranging from -18 to 24˚C, notable amounts

of deicing salts, and physical impact during snow cleaning”; (b) Leslie Street Bridge in

Toronto, ON exposed to varying temperatures (from -10˚C to 27˚C) with multiple

freezing and thawing cycles; (c) Maryland Bridge in Winnipeg, MB exposed to “a

relatively dry climate with temperatures between -23 and 6˚C”. Authors noted that

debonded areas determined based off of thermographs corresponded to areas with

reduced bond strength. Furthermore, they observed that cohesive failures were related

to pull-off strengths higher than 330 psi. Relatively low bond strengths were observed

on Safe Bridge “probably because the geometry of the girders allowed for the infiltration

49

of water behind the FRP layer”. Authors also reported that aside from water infiltration it

was not clear why many of the bond strengths were low. They suspected that the low

values were related to general deterioration; however, this could not be proven as the

initial bond strength data was not obtained.

To determine the sensitivity of FRP-concrete bond to chloride content Pan et al.

(2010) performed direct shear tests on specimens conditioned in water solutions with

the following concentrations of NaCl: (a) 3%, (b) 6%; (c) 10%; and (d) 15%. Specimens

were conditioned for 15, 30, 60, 90, and 120 days. Authors concluded that concrete

compressive strength significantly increases with the immersion time in NaCl solution.

However, the chloride concentration did not have significant effect on the compressive

strength of concrete. Furthermore, authors noted a slight decrease in initial and ultimate

debonding loads after 15 and 30 days of conditioning. Specimens that were conditioned

longer, however, experienced a slight increase in bond strength when compared to

unconditioned samples. Moreover, there was no apparent correlation between the bond

strength and the chloride concentration of the water solution.

Figure 3-19. Direct shear test specimen

50

Dai et al. (2010) performed a study on influence of moisture of concrete surface

during the application of FRP, and in service moisture on performance of FRP-concrete

bond. Authors utilized direct pull-off and bending tests to assess the bond performance

after conditioning for 8, 14 and 24 months in wet-dry cycles consisting of four day

immersion in 60˚C followed by drying for three days in standard laboratory conditions.

Additional variables were: (1) curing conditions after repair (RH=48% vs. RH=90%); (2)

wet vs. dry substrate at the time of repair; (3) normal vs. hydrophilic primer; and (4)

normal vs. ductile adhesive. It should be noted that for the purpose of this research

special form of CFRP, called carbon stranded sheet (CSS), was used. The CSS

consists of 1 to 2 mm in diameter circular carbon microbars, formed by pultruding dry

carbon fibers with epoxy. In pull-off test specimens a loss in capacity of up to 50% was

noted, while the loss in capacity in bending specimens peaked at around 40%. Based

on the results of the experimental program authors concluded the following:

1. Different curing conditions were not of critical importance on the bond capacity.

2. Wet concrete substrate at time of installation detrimentally affected the bond performance, however, only when the normal primer was used.

3. Wet-dry cycling caused shift of failure mode from cohesive to adhesive (between primer and concrete). This may be due to microcracks observed by microscope that formed at the primer-to-concrete interface that formed after wet-dry cycling.

4. No general trend was observed in bond capacity in respect to duration of wet-dry cycling. Bending specimens experienced both increase and decrease in capacity over time of conditioning.

5. Pull-off tests are not indicative of the overall bond condition, because they rather capture the local weaknesses at the interface. Pull-off test is, however, deemed sufficient to provide a conservative estimate of durability of the FRP-concrete bond capacity.

51

Figure 3-20. Three-point bending test setup

Cromwell et al. (2011) investigated influence of multiple aggressive environments

on performance of FRP reinforcement in concrete structures. Authors used two types of

CFRP in their test program: a laminate and a wet-layup system. Test program included

three different types of specimens: (1) tension coupon specimens – prepared as per

ASTM D3039; (2) bond specimens – “two 2 in. (51 mm) concrete cubes spaced 1 in. (25

mm) apart and bonded together using 0.75 in. (19 mm) wide by 5 in. (127 mm) long

FRP strips on opposing faces”; (3) beam flexure (three-point bending) specimens –

concrete beams reinforced with 2 #3 bars at top and bottom and U shaped stirrups

W2.9 spaced at 5.98 in. (152 mm) on center, and the following dimensions

D:W:L=6.1:8:96.1 in (154:203:2440 mm); reinforced with CFRP at the soffit. Specimens

were conditioned in the following environments: (a) water – RH=100% at 38˚C for 1000,

3000, and 10,000 as per ASTM D2247; (b) salt water solution - prepared as per ASTM

D1141, for 1000. 3000, and 10,000 h at 22˚C; (c) alkaline (CaCO3 solution) – at 22˚C for

1000, 3000, and 10,000 h; (d) dry heat - 60˚C in a forced-draft circulation-air furnace as

per ASTM D3045, for 1000 and 3000 h; (e) diesel fuel – immersion in diesel fuel at 22˚C

52

for 4 h as per ASTM C581; (f) weathering – UV340 light at 63˚C for 2 h, followed by

RH=100% at 22˚C for 2 h, for the total of 2000 h (1000 cycles); (g) freeze-heat – cycling

between -18˚C for 15 h, and RH=100% at 38˚C for 15 h (20 cycles in total), following the

exposure to RH=100% at 38˚C for 500 h and drying for 48 h; (h) freeze-thaw cycling –

360 cycles as follows: “(1) 70 min at -18˚C at 30% RH; (2) 20 min ramp up to 4.5˚C

(resulting in 90% RH); (3) 70 min at 4.5˚C at 50% RH with UV lights on; and (4) UV

lights off and 80 min ramp down to -18˚C (resulting in 40% RH)”. Control specimens

were conditioned in standard laboratory relative humidity and temperature. It should be

noted that only beam flexure specimens were exposed to freeze-thaw cycling. Tension

and bond specimens were conditioned in all other aforementioned environments.

Results from tension coupon tests showed that properties of both CFRP laminate and

fabric are not significantly affected by exposure to aggressive environments. In any of

the conditioning environments modulus of elasticity and ultimate did not fall below 95%

and 90% for CFRP laminate and fabric, respectively. Bond specimens showed a much

greater variation in results as well as a higher level of sensitivity to aggressive

environmental conditions. Specimens with bonded CFRP laminate reinforcement

showed the greatest reduction in bond capacity (of close to 20%) after exposure in salt

water for 10,000 h, and dry heat for 1000 h. On the other hand, specimens reinforced

with CFRP fabric proved to be mostly affected by dry heat condition where they showed

a reduction of close to 40% of control bond strength. Beam flexure specimens

experienced very low reductions in strength. Therefore, authors concluded that the

“intermediate-crack (IC) debonding” is unaffected by freeze-thaw cycling.

53

Figure 3-21. Bond test setup

Figure 3-22. Beam flexure specimen

Lai et al. (2009) performed a series of direct shear tests to determine effects of

high temperature and water ingress on durability of FRP-concrete bond. Specimens

were immersed for 5, 15 and 30 weeks in water at the following temperatures: (a) 25˚C;

(b) 40˚C; and (c) 60˚C. By digitally processing the visual images of FRP strips, authors

identified three distinguishing failure modes:

A. Failure in concrete

B. Failure at FRP-epoxy interface

C. Failure within adhesive bonding layer

54

Aside from a decrease in ultimate failure load (of up to 30%), authors observed

increase in average delamination (flaws in adhesive layer that form due to exposure to

aggressive environments) sizes in the 40 ˚C and 60˚C specimens, when compared to

control. In addition to that, they observed a shift from failure mode A in control and the

25˚C specimens, to predominantly failure modes B and C in specimens exposed to 40

and 60˚C, as presented in Table 3-1.

Table 3-1. Distribution of failure modes Exposure temperature (˚C)

Average failure mode (%)

Mode A Mode B Mode C Control 75.3 24.7 0 25 63.8 33.3 2.8 40 58.1 17.4 24.5 60 41.3 32.7 26

55

Figure 3-23. Direct shear test setup

In a different study on durability of FRP-concrete bond when conditioned in water

with elevated temperatures, Lai et al. (2013) used infrared thermography in conjunction

with direct shear test. Based on the thermographs, three distinguishing stages in

debonding process can be identified (Smith and Teng 2002, Colombi et al. 2010):

1. “Elastic stage (no interfacial softening or rupture can be found over the entire interface)”;

56

2. “Elastic-softening stage (local softening starts at the loaded end and some parts of the interfacial bonds become softened while the portion neat the fixed end, remains elastic)”;

3. “Elastic-softening-debonding stage and softening-debonding stage (local rupture of the bond layers happens and propagates from the loaded to the fixed ends)”

From the results of durability study authors noted an early occurrence of the

softening-ruptured state for all specimens that were conditioned at 60˚C. Control, the

25˚C and the 40˚C debonding process commenced with elastic-softening stage. The

degradation due to exposure caused a drop in ultimate shear force capacity from control

to specimens exposed at 60˚C of up to around 30%.

Srestha et al. (2013) examined influence of water on FRP-concrete bond in high

strength concrete by utilizing direct shear pull-off test. They tested specimens made of

normal strength concrete for comparison purposes. Specimens (Figure 3-24) were

immersed in water at 20˚C for up to 12 months. Two types of epoxy (Epoxy E:

combination of Bisphenol-A and Bisphenol-F epoxy resins; and Epoxy F: Bisphenol-A

epoxy resin), and one CFRP fabric were used in the study. Linear dependency of

ultimate bond strength in respect to exposure time was observed, with ultimate values

recorded in Table 3-2. Better performance of bond in lower strength concrete was

explained by differences in surface properties between the two. Namely, high strength

concrete is tightly packed and due to lack of pores and voids does not have much

available surface for transfer of frictional forces, whereas this is not the case in normal

strength concrete. As evidence to support this claim, authors compared the failure

surfaces, which revealed that less concrete debris was found on CFRP that debonded

from concrete surface (Figure 3-25 and Figure 3-26).

57

Table 3-2. Change in ultimate bond strength after 12 months of exposure Normal Strength

Concrete High Strength Concrete

Epoxy E ≈+40% ≈-32% Epoxy F ≈0% ≈-30%

Figure 3-24. Direct sehar test specimen

Figure 3-25. Failure surfaces for normal strength concrete

58

Figure 3-26. Failure surfaces for high strength concrete

59

CHAPTER 4DURABILITY TESTING APPROACH

A material characterization study was performed on epoxies (Stewart 2012) to

determine how the material degradation mechanism changes depending on the

conditioning environment. Epoxy samples were exposed to hygrothermal conditions at

30˚C through 90˚C, and UV radiation combined with high humidity. Stewart (2012)

studied differences in behavior that occurred due to the exposure to the aforementioned

conditions and proposed the following set of conditioning protocols for testing of FRP-

concrete bond:

1. Exposure to water immersion at temperatures that are 15˚C above and below the Tg for 1, 2, and 8 weeks

2. Exposure to UV and water since this condition changed epoxy mechanism of degradation from oxidation to hydrolysis

Based on the recommendations, exposure conditions and exposure times in

Table 4-1 were defined for FRP-concrete bond testing. Five composite systems and two

different concrete strengths were selected for the study. Three specimens per exposure

condition were tested. Three-point bending test was utilized to test FRP-concrete bond

durability (Gartner 2007).

Table 4-1. Summary of conditioning protocols

Exposure Condition Temp. (oC) Exposure Times (weeks)

Number of specimens

Immersed in water 30 1, 2, 8 105

Immersed in water 60 1, 2, 8 105

RH=100% 60 1, 2, 8 90 UV & RH=100% 60 1, 2, 8 15

Test matrix for FRP reinforced concrete samples presented in Figure 4-1 was

created for the purpose of durability study. It should be noted, however, that tests on

60

combined effect of UV and moisture at 60˚C were not performed due to unexpected

circumstances. Concrete strength for al samples was 10,000 psi except for Composite

D which was tested for concrete strengths of 4,000 psi and 10,000 psi. Therefore

designations such “D04” and “D10” signify the corresponding concrete compressive

strength.

1 2 8 1 2 8 1 2 8 1 2 81 in wide bonded

CFRP fabric 0.644 Composite A3 3 3 3 3 3 3 3 3 3 3 3 3 39

1 in wide bonded CFRP fabric 0.618 Composite B 3 3 3 3 3 3 3 3 3 3 3 3 3 39

1 in wide bonded CFRP fabric 0.6 Composite C 3 3 3 3 3 3 3 3 3 3 3 3 3 39

1 in wide bonded CFRP fabric 0.644 Composite D04 3 3 3 3 3 3 3 3 3 3 3 3 3 39

1 in wide bonded CFRP fabric 0.644 Composite D10 3 3 3 3 3 3 3 3 3 3 3 3 3 39

1 cm wide laminate N/A Composite E 3 3 3 3 3 3 3 3 3 3 3 3 36

Tota

l 4,0

00 p

si

Tota

l 10,

000

psi

FRP Reinforcement Material

Immersion + 30C

Exposure time (weeks)

Immersion + 60C

Exposure time (weeks)

Immersion + UV + 60CExposure

time (weeks)

100%RHSystem Fiber Weight (kg/m2) C

ontr

ol

Exposure time (weeks)

Figure 4-1. Test matrix

61

CHAPTER 5DURABILITY TEST SPECIMEN DESIGN AND FABRICATION

Specimen Design

All the concrete specimens had a square cross section measuring 4 in. x 4 in.

and were 14 in. long. This specimen size was deemed adequate to capture the bond

behavior in bending at a small scale. The size was also chosen to match the previous

work conducted at the University of Florida by Gartner et al. (2011). The amount of FRP

reinforcement was picked to enable the debonding failure of the FRP instead of a

concrete shear failure. If too much reinforcement is applied to the specimen the bond

strength cannot be fully utilized and the specimen is likely to fail in shear by reaching

the ultimate shear capacity of concrete. Adequate cross sectional area of surface

bonded FRP was determined based on the available experimental data (Gartner et al.

2011). FRP strips measuring 1 in. in width and 8 in. in length deemed appropriate to

achieve a desired failure mode of the test specimen. In addition, to better simulate the

behavior of FRP in flexural applications, by promoting Mode II loading condition, a 2 in.

deep notch was provided at the specimen midspan to simulate cracked concrete (as

shown in Figure 5-1). Furthermore, notch allows testing beam specimens in a three-

point bending test instead of a four-point bending test because the notch presence

eliminates the need for a constant moment region and reduces the possibility of a shear

failure.

62

Figure 5-1. Three-point bending test setup

4 in. 4 in.

6 in. 6 in.

12 in.

1 in.

4 in.

4 in

.2 in

.2

in.

1 in

.CFRP

notch

63

Three-point Bending Test Bond Strength Index

One of the most important details worth discussion is how the physical value that

is obtained from the three-point bending test relates to the actual bond strength. First, it

should be regarded that bond strength is usually defined in terms of loading condition.

Namely, two most commonly observed loading conditions are Mode I, Mode II, and a

mixture of the two. Mode I or the peel mode starts at the ends of the laminates and

propagates towards the midspan (Figure 5-2). Peeling occurs as a result of the end

offset of FRP reinforcement from the supports, which allows for the ends of FRP

reinforcement to be exposed somewhere along the span. If looked at boundary

conditions in FRP reinforcement at the ends (free end boundary condition) it can be

easily concluded that there is zero curvature at the location. To satisfy the boundary

conditions, and due to the existence of bending stresses in concrete, FRP laminate

needs to bend in the direction opposite to direction of beam bending, which will cause

Mode I loading at the location (Sebastian 2001).

Mode II, on the other hand, is observed in the vicinity of vertical flexural cracks.

According to Sebastian (2001), when concrete’s tensile capacity is reached, flexural

cracks open, which loads the FRP in direct shear (Figure 5-2).

Mixed mode (Figure 5-2) happens as a result of inclined cracks along the span

and is is a function of phase angle ψ (mode mixity parameter - I

II

KK1tan −=ψ ) which is

dependent on crack inclination angle (Sebastian 2001).

With the newly developed anchoring techniques, Mode I loading may be

neglected. Therefore, Mode II crack development is considered the most critical in

flexural FRP applications.

64

Figure 5-2. Loading modes in interfacial region

The purpose of three-point bending test, described previously, is to directly

simulate Mode II loading condition which is deemed representative of the FRP-

reinforced beam behavior when subjected to bending. The quantity recorded from these

tests is the ultimate load that caused failure of the test specimen. Even though the

specimen failure load is not representative of the actual bond stress that caused FRP to

debond, it provides a way to indirectly express the bond strength in flexural applications.

The ultimate exposed specimen failure load normalized to averaged control specimen

failure load will be referred to as bond strength index (BSI). Bond strength index, as

defined here, is representative of both peel mode and Mode II loading conditions, with

Mode II being more dominant. It should be noted that in this study terms such as bond

capacity, bond strength, and bond degradation, when used in relation to three-point

bending test, refer to bond strength index, and not the actual debonding stress.

CFRP Composites

All composite materials used in this study are commercially available by US

manufacturers. In the following subheadings, CFRP composite systems that were used

to reinforce the concrete beam specimens, and corresponding adhesives are described.

Mode I Mixed-modeMode II

65

Composite A is a custom, uni-directional carbon fabric orientated in the 0˚

direction that combined with Adhesive A forms a wet-layup composite system that is

used in infrastructure applications (Table 5-1).

Composite B is a high strength unidirectional carbon fiber fabric that is field

laminated by epoxy to form a carbon fiber reinforced polymer (CFRP). Properties are

specified in Table 5-1.

Composite C consists of Composite C Primer, Composite C Putty, Adhesive C

(Composite C Saturant) and dry fabric constructed of high strength aerospace grade

carbon fibers (Table 5-1):

• Composite C Primer is a low viscosity, 100% solids, polyamine cured epoxy. Being the first component of the system applied to the concrete surface, Primer is used to penetrate the pores of concrete substrate and to provide a high bond base coat for the Composite C system. According to the manufacturer’s product data sheet Primer should be mixed following these ratios: a) 3 parts of component A to 1 part of component B by volume; or b) 100.0 parts of component A to 30.0 parts of component B by weight.

• Composite C Putty is a 100% solid, non-sag paste epoxy material that is applied as a second layer of Composite C system. The purpose of Putty is to level the uneven surfaces before the application of Composite C fibers. Additionally, Putty improves adhesion of subsequent coatings on substrates. Putty shall be mixed according to the following ratios: a) 3 parts of component A to 1 part of component B by volume; or b) 100.0 parts of component A to 30.0 parts of component B by weight.

• Finally, fibers, when saturated with Adhesive C, form a high strength carbon fiber reinforced polymer (CFRP).This material, as stated by the manufacturer, can provide additional strength to concrete, masonry, steel and wood structural elements. Composite D consists of Composite A system carbon fiber fabric and Adhesive D

(Table 5-1).

Composite E a pultruded carbon fiber reinforced polymer (CFRP) laminate.

Composite E is bonded onto the structure as external reinforcement using Adhesive E

(Table 5-1).

66

Table 5-1. Properties of composite systems (information provided by manufacturers) Fabric

weight (g/m2)

Thickness (in)

Ultimate Tensile Strength (psi)

Modulus of Elasticity (106 psi)

Composite A

644 0.04 143,000 13.90

Composite B

618 0.04 123,200 10.24

Composite C

600 0.04-0.06 550,000 33.00

Composite D

644 0.04 n/a n/a

Composite E

n/a 0.047 449,000 23.90

Epoxy Adhesives

Adhesive A is a two component epoxy matrix material for bonding applications. It

a material used in structural applications to provide a wet-layup composite system for

strengthening structural members. Adhesive A provides a long working time, with no

offensive odor. Adhesive A is mixed according to the following ratios: a) 100.0 parts of

component A to 42.0 parts of component B by volume; or b) 100.0 parts of component

A to 34.5 parts of component B by weight. The recommended minimum cure time is 72

hours at 70 F. The preceding description of the material is provided by the

manufacturer.

Table 5-2. Adhesive A mechanical properties 72 hours post cure at 140˚ F (60˚ C) – as reported by the manufacturer

Property ASTM Method Typical Test Value* Tg D4065 180˚ F (82˚ C) Tensile Strength D638 Type 1 10,500 psi Tensile Modulus D638 Type 1 461,000 psi Elongation Percent D638 Type 1 5.0 % Compressive Strength D695 12,500 psi Compressive Modulus D695 465,000,000 psi

67

Adhesive B is a two-component 100% solids, epoxy, that is used as an

impregnating resin with Composite B CFRP fabric; and as a seal coat and impregnating

resin for horizontal and vertical applications. The recommended minimum cure time is

14 days at normal ambient conditions (73˚ F and 50% R. H.).

Table 5-3. Adhesive B mechanical properties (14 day cure at 73˚F (23˚C) and 50% R.H.) – as reported by the manufacturer

Property ASTM Method Typical Test Value* Tg D4065 46 Tensile Strength D638 8,000 psi Tensile Modulus D638 250,000 psi Elongation Percent D638 3.0 % Compressive Strength D695 N/A Compressive Modulus D695 N/A

Adhesive C is a 100% solids, low viscosity epoxy material that is used to saturate

the fibers of Composite C to form a carbon fiber reinforcing laminate (CFRP). Physical

properties of Adhesive C are provided Table 5-4.

Table 5-4. Adhesive C mechanical properties (based on cured samples at 72˚F (20˚C) and 40% R.H.) – as reported by the manufacturer

Tg D4065 163˚F (71˚C) Tensile Strength D638 8,000 psi Tensile Modulus D638 440,000 psi Elongation Percent D638 3.5 % Compressive Strength (28 day)

D695 12,500 psi

Compressive Modulus (7 day) D695 380,000 psi Adhesive D is a two-part epoxy with a known chemical composition, comprised of

diglycidyl ether of bisphenol A (DGEBA) and poly(oxypropylene) diamine (POPDA). The

mix ratio of DGEBA to POPDA was 100 to 32.9.

Adhesive E is a 2-component, 100% solids, structural epoxy paste adhesive.

This adhesive conforms to the current ASTM C881 and AASHTO M-235 specifications.

The epoxy is used as an adhesive for bonding external reinforcement to concrete,

masonry, steel, wood, stone, etc., structural bonding of composite laminates, structural

68

bonding of steel laminate