You are on page 1of 447

MULTI-SCALE DESIGN FOR DURABLE REPAIR

OF CONCRETE STRUCTURES
by

Mo Li

A dissertation submitted in partial fulfillment


of the requirements for the degree of
Doctor of Philosophy
(Civil Engineering)
in The University of Michigan
2009

Doctoral Committee:
Professor Victor C. Li, Chair
Professor Richard E. Robertson
Associate Professor Jerome P. Lynch
Assistant Professor Adda Athanasopoulos-Zekkos

UMI Number: 3392967

All rights reserved


INFORMATION TO ALL USERS
The quality of this reproduction is dependent upon the quality of the copy submitted.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.

JUMT
Dissertation Publishing

UMI 3392967
Copyright 2010 by ProQuest LLC.
All rights reserved. This edition of the work is protected against
unauthorized copying under Title 17, United States Code.

uest
ProQuest LLC
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, Ml 48106-1346

Mo Li
All rights reserved
2009

To my parents,
Whom I love more than I can write.

Acknowledgements
This dissertation has been completed under the guidance of my advisor, Professor
Victor C. Li. Without his support, inspiration, dedication of time and energy throughout
the past five years I could have never completed this work. I owe forever my sincerest
gratitude to him, for opening my eyes to the innovative material technology world, and
for challenging me with novel research ideas capable of solving real-world problems.
I would like to thank my dissertation committee members, Professor Richard E.
Robertson, Professor Jerome P. Lynch, and Professor Adda Athanasopoulos-Zekkos for
their intellectual guidance, valuable comments, great patience and encouragement. The
research presented in this dissertation is only possible with their help.
I would also like thank Mr. Roger Till at the Michigan Department of
Transportation, for his tremendous help and valuable comments that made the
development and field implementation of HES-ECC possible, Professor Shuichi
Takayama and his graduate students in the Department of Biomedical Engineering, for
their selfless sharing of knowledge in fluoresccence technology and assistance with the
fluoresccence microscopic equipments, Professor Erik Schlangen at Delft University of
Technology in the Netherlands, for generously providing the softeware FEMMASSE and
guidance of use, Professor Petr Kabele at Czech Technical University in Prague for
helpful discussions.
I must acknowledge the National Science Foundation for funding of Materials
Use: Science, Engineering, and Society (MUSES) Biocomplexity Grants CMS-0223971
and CMS-0329416, the National Science Foundation for funding of Cyberinfrastructure
Training, Education, Advancement, and Mentoring for Our 21 st Century Workforce (CI-

iii

TEAM) Grant OCI-0636300, the Michigan Department of Transportation, and


Metamorphix Global, Inc., all of which supported much of my research.
I would like to thank my colleagues and my friends at the Advanced Civil
Engineering Materials Research Laboratory, Dr. Shuxin Wang, Dr. Yun Y. Kim, Dr.
Martin Weimann, Dr. Michael D. Lepech, Dr. Shunzhi Qian, Dr. En-Hua Yang, Dr.
Andrea Spagnoli, Dr. Yingzi Yang, Dr. Estela Garces, Mr. Jian Zhou, Ms. Lili Kan, Mr.
Matthew Fadden, Mr. Ravi Ranade, and Mr. Michael Stults for making my life at ACEMRL and CEE concrete mixing labs full of inspiration and joy. I would also like to thank
Dr. Tsung-Chin Hou for his friendship and great help during my studies.
I am grateful to Ms. Debbie Hemmeter, Ms. Janet Lineer, Ms. Reta Teachout, Ms.
Linda Fink, Ms. Kimberly Bonner, Ms. Kimberly Smith, Ms. Jill Miller, Ms. Kimberly
Gauss, and Ms. Patricia Mackmiller for unparalleled assistance in administrative matters,
and to Mr. Merrick Burch, Mr. Jan Pantolin, Mr. Robert Fischer, and Mr. Robert Spence
for their greatly helpful technical support.
To my dearest and best friends of life, Dr. Hao Chen, Zhiyu Luo, Dr. Minhua
Shao, Dr. Jiexin Lian, Xiaoqun Huang, Chonglong Fu, and Yan Li, only with your
encourgements and love was I able to start such an wondrful adventure in the U.S.
Finally, I could never have attempted any of this without the love and support of
the greatest parents in the world, Ping Gao and Gongda Li. I owe a great deal to
Alphonse Anderson. Without his tremendous help and unconditional support I would
have never been able to achieve this.

IV

I cannot thank you enough Fonz.

Table of Contents
Dedication

ii

Acknowledgements

iii

List of Tables

List of Figures

xiii

Chapter
1. INTRODUCTION
1.1 Deterioration in Concrete Structures - Major Durability Concerns
1.1.1 Bridges

1
4
5

1.1.2 Highways and Roadways

13

1.1.3 Parking Structures

16

1.2 Limitation of Current Approaches to Improve Durability of Concrete Structures 18


1.2.1 Effectiveness of current material "durability" solutions

19

1.2.2 Compatibility of new repair with existing concrete substrate

22

1.2.3 Some considerations of current repair performance characterization

25

1.3 Proposed Approach

27

1.4 Engineered Cementitious Composites

30

1.5 Research Motivation and Objectives

33

1.6 Research Scope and Dissertation Organization

35

2. Micromechanics Based Design of ECC Repair Materials

63

2.1 High Performance Fiber Reinforced Cementitious Composites

65

2.2 ECC Micromechanics Based Design Theory

67

2.2.1 ECC Design Framework - Scale Linking

67

2.2.2 Conditions for Tensile Strain Hardening

68

2.2.3 Condition for Saturated Multiple Microcracking

71

2.3 High Early Strength Engineered Cementitious Composites

75

2.3.1 Background

75

2.3.2 High Early Strength Requirements

77

2.3.3 Methods of Achieving High Early-Age Compressive Strength

79

2.3.4 Initial Material Composition Selection Based on Strength Requirements 84


2.3.4.1 Experimental Program

84

2.3.4.2 Compression Test Results and Discussion

87

2.3.5 Microstructure Tailoring for Tensile Ductility

90

2.3.5.1 Initial Composite Testing in Uniaxial Tension

90

2.3.5.2 Microstructure Tailoring of Mix 6 for Tensile Ductility

91

2.3.6 HES-ECC Material Characterization


2.4 Conclusions

98
104

3. ECC Processing and Quality Control for Optimized Material Properties and
Robustness

145

3.1 Introduction

146

3.2 Effect of Processing on ECC Micromechanical Parameters

149

3.3 Experimental Program

152

3.3.1 Materials

153

3.3.2 Mixing, Testing of Fresh Properties, and Casting

154

3.3.3 Testing of Hardened Tensile Properties

156

3.3.4 Measurement of Fiber Dispersion Coefficient

157

3.4 Experimental Results and Discussion

158

3.5 Rheology Control Concept and Case Study

163

3.6 Conclusions

165

4. Effect of Material Ductility on Restrained Volume Change Induced Cracking and


Interfacial Delamination

182

4.1 Introduction

183

4.2 Proposed Approach

188

4.3 Free Shrinkage of HES-ECC

191

VI

4.3.1 Background

191

4.3.2 Materials and Experimental Procedure

194

4.3.3 Experimental Results and Discussion

196

4.4 Restrained Shrinkage Ring Test

197

4.4.1 Materials and Experimental Procedure

197

4.4.2 Experimental Results and Discussion

198

4.5 Restrained Shrinkage Test on a Simulated Repair System


v

200

4.5.1 Materials

200

4.5.2 Repair Specimen Configuration and Surface Preparation

201

4.5.3 Experimental Results and Discussion

203

4.6 Effects of Surface Preparation on the Repaired System Performance

207

4.6.1 Experimental Procedure

208

4.6.2 Experimental Results and Discussion

209

4.7 Numerical Study on the Simulated Repair System

212

4.7.1 Problem Formulation and FEM Model

212

4.7.2 Numerical Simulation Results

214

4.8 Simplified Analytical Model for the Repair Systems under Restrained Volume
Change

216

4.9 Conclusions

220

5. Influence of Material Ductility on Reflective Cracking in Overlay Repairs

252

5.1 Introduction

253

5.2 Ductile ECC Overlay

257

5.3 Experimental Program

259

5.3.1 Specimen Configuration and Loading Conditions

259

5.3.2 Materials and Specimen Preparation

260

5.4 Experimental Results

262

5.4.1 Overlay System Monotonic Flexural Performance

262

5.4.2 Overlay System Fatigue Performance

266

5.5 Conclusions

268

vn

6. Transport Properties of ECC under Chloride Exposure

287

6.1 Steel Corrosion in Concrete Structures

288

6.2 Proposed Approach - ECC Microcracking Phenomenon

290

6.3 Transport Properties of Cementitious Materials

292

6.4 Experimental Program

297

6.4.1 Materials and Mixture Proportions

297

6.4.2 Specimen Preparation and Testing

298

6.5 Experimental Results and Discussion

302

6.6 Conclusions

307

7. Effect of Cracking and Healing on the Durability of ECC under Combined


Aggressive Chloride Environment and Mechanical Loading

328

7.1 Introduction

329

7.2 Evidence of Self-Healing of ECC Microcracks and Recovery of Chloride Transport


Properties

331

7.3 Experimental Program

334

7.3.1 Testing of Composite Properties

334

7.3.2 Measurement of Micromechanical Properties

335

7.4 Experimental Results and Conclusions

337

7.5 Conclusions

341

8. Large-Scale Processing and Field Demonstration of High Early Strength


Cementitious Composites in Bridge Repair

353

8.1 Background

354

8.2 Optimization of HES-ECC Ingredients and Mixing Procedure for Larger Scale
Applications

356

8.2.1 Fiber Length Change

356

8.2.2 Hydration Stabilizing Admixtures

358

8.2.3 Batching Sequence

359

8.3 HES-ECC Larger Scale Trial Batching


8.3.1 One Cubic Feet Batching

361
361

vm

8.3.2 Three Cubic Feet Batching and Six Cubic Feet Batching

362

8.4 HES-ECC Material Cost Effectiveness

364

8.5 HES-ECC Patch Repair Demonstration

371

8.6 Long-Term Monitoring of HES-ECC Patch Repair Performance

372

8.7 Conclusions

374

9. Concluding Remarks

402

9.1 Research Overview and Findings

402

9.1.1 Material Engineering for Durable Repair Materials

403

9.1.2 Durability Assessment of ECC/Concrete Repair System

407

9.1.3 Field Application of ECC Repair Materials

415

9.2 Impact of Research

417

9.3 Recommendations for Future Research

420

IX

List of Tables
Table 2.1

Three phases of ECC microstructure and corresponding micro-parameters.


107

Table 2.2

List of materials used.

108

Table 2.3

Mixing proportions for initial material composition selection based on


strength requirements.

109

Table 2.4

Properties of REC 15 PVA fiber from Kuraray, Co, Japan.

109

Table 2.5

Compressive strength of Mix 1 and 2 at different ages.

110

Table 2.6

Compressive strength of Mix 3, 4 and 5 at different ages.

110

Table 2.7

Compressive strength of Mix 6 (w/ and w/o insulation) at different ages.


Ill

Table 2.8

Mixing proportion of HES-ECC before / after adding artificial flaws. 111

Table 2.9

Mix proportion of HES-ECC.

Table 2.10

HES-ECC tensile properties at different ages (mean standard deviation).

112

112
Table 2.11

HES-ECC compressive properties at different ages (mean standard


deviation).

Table 2.12

113

HES-ECC flexural properties at different ages (mean standard deviation).


113

Table 3.1

Mix proportion of materials.

168

Table 3.2

Material fresh and hardened properties.

169

Table 3.3

Mix proportion of WHITE ECC.

169

Table 4.1

General requirements of patch repair materials for structural compatibility.


224

Table 4.2

Mix proportion of the investigated materials.

224

Table 4.3

Mechanical properties of the investigated materials.

225

Table 4.4

HES-ECC tensile properties at different ages.

225

Table 4.5

Theoretical parametric values of the investigated materials.

226

Table 4.6

Restrained drying shrinkage crack width development with age.

226

Table 4.7

Estimated material cracking potentials, RH=45%.

226

Table 4.8

Interface delamination and surface cracking of HES-Concrete, HES-SFRC,


and HES-ECC layered repair systems at the age of 60 days.

Table 4.9

Material composition, 28-day mechanical properties

227
and cracking

potential.
Table 4.10

228

Interface delamination and surface cracking of concrete, SFRC and ECC


layered repair systems.

228

Table 4.11

Interface property values in FEM model.

229

Table 5.1

Modulus of rupture of HES-ECC and HES-Concrete overlays under


monotonic loading.

Table 5.2

270

Center deflection (upon failure) of HES-ECC and HES-Concrete overlays


under monotonic loading.

Table 5.3

270

Interfacial delamination (upon failure, including initial delamination) of


HES-ECC and HES-Concrete overlays under monotonic loading.

Table 6.1

271

Permissible crack widths for reinforced concrete under service loads


according to ACI Committee 224.

310

Table 6.2

Mixture proportions and properties of ECC and mortar.

310

Table 6.3

Mechanical properties of PVA fiber.

311

Table 6.4

Crack widths, numbers and depths of pre-loaded ECC and mortar prisms.
311

Table 6.5

Chloride ponding test - Effective diffusion coefficient.

311

Table 7.1

Tensile properties of ECC under different combinations of pre-loading and


chloride exposure conditions.

343

Table 8.1

ECC material mixing proportions.

377

Table 8.2

Tensile properties of HES-ECC with 0.33 in. (8.4 mm) PVA

Table 8.3

Larger scale HES-ECC mixing proportions and batching weights.

378

Table 8.4

Larger scale HES-ECC batching sequence and time.

378

Table 8.5

fiber.

377

Flowability test results from one & six cubic feet (0.03 & 0.17 cubic meter)
batching.

379

XI

Table 8.6

Compressive Strength development of HES-ECC from one & six cubic


feet (0.03 & 0.17 cubic meter) batching.

Table 8.7

Table 8.8

379

Tensile properties of HES-ECC from one & six cubic feet (0.03 & 0.17
cubic meter) batching.

380

Minimum requirements of HES-ECC material.

380

xn

List of Figures
Figure 1.1

Concrete Structure Deterioration, Damage, and Defect.

40

Figure 1.2

Deterioration of Kurtsubo Bridge, Japan: (a) Rust signs before the first
repair; (b) Tendon corrosion and breakage before the second repair.

41

Figure 1.3

Life cycle cost of Kurtsubo Bridge, Japan.

41

Figure 1.4

Bridge deck transverse cracking.

42

Figure 1.5

Steel corrosion and spalling in concrete bridges in Michigan.

42

Figure 1.6

Pavement overlay reflective cracking.

43

Figure 1.7

Deterioration in parking structures.

43

Figure 1.8

Model of concrete slab restrained against shrinkage and (b) the crack
width w controlled by the tension-softening curve.

44

Figure 1.9

Schematic plot of crack width w as a function of shrinkage strain es.

44

Figure 1.10

Integrated Multi-Scale Material & Structural Engineering Framework for


durability of repaired concrete structures.

Figure 1.11

45

Disconnected Material Engineering and Structural Engineering Design


Philosophies for durability of repaired concrete structures.

46

Figure 1.12

Typical tensile stress-strain curve of ECC.

47

Figure 1.13

Driving surface static friction tester.

48

Figure 1.14

Relative dynamic modulus vs. number of freeze thaw cycles.

48

Figure 1.15

Mihara Bridge with ECC/steel composite deck opened to traffic in April


2005,Japan:

Figure 1.16

49

ECC link slab on the Grove Street Bridge over 1-94 opened to traffic in
October 2005, Michigan.

Figure 2.1

Figure 2.2

49

Uniaxial tensile stress-deformation relation of concrete, FRC, and


HPFRCC.

114

Scale linking in ECC Material Development.

115

xin

Figure 2.3

Typical fiber bridging stress vs. crack opening a(5) curve for a tensile
strain-hardening composite.

115

Figure 2.4

Steady-state cracking with a constant crack opening .


c

116

Figure 2.5

Theoretically predicted effect of initial flaw size on the cracking strength


ofPVA-ECC.

Figure 2.6

State

and federal

116
compressive

strength requirements,

and

target

compressive strength at different ages for HES-ECC.

117

Figure 2.7

Uniaxial compressive test setup.

117

Figure 2.8

Effect of cement type on compressive strength development.

118

Figure 2.9

Effect of water/cement ratio on compressive strength development.

118

Figure 2.10

Effect of accelerator on compressive strength development.

119

Figure 2.11

Effect of curing condition on compressive strength development.

119

Figure 2.12

Uniaxial tensile test setup.

120

Figure 2.13

Dimensions of tensile test specimen.

120

Figure 2.14

Tensile stress-strain curves of Mix 6 at ages of 4h, 24h, 3d, 7d, and 28d.
122

Figure 2.15

Age dependency of tensile strain capacity of Mix 6.

123

Figure 2.16

Age dependency of tensile strength of Mix 6.

123

Figure 2.17

Matrix fracture toughness test set-up.

124

Figure 2.18

Single fiber pull-out test set-up.

124

Figure 2.19

General profile of a single fiber pullout curve.

125

Figure 2.20

Age dependency of matrix fracture toughness.

125

Figure 2.21

Age dependency of matrix/fiber interface chemical bond.

126

Figure 2.22

Age dependency of matrix/fiber interface frictional stress.

126

Figure 2.23

Age dependency of slip hardening coefficient.

127

Figure 2.24

Evolution of tensile strain capacity and Jb'/Jtip ratio with age (Mix 6). 128

Figure 2.25

Optimization scheme for pre-existing flaw size distribution in matrix. 129

Figure 2.26

Aggregates used as artificial flaws: polystyrene (PS) beads.

129

Figure 2.27

Age dependent tensile stress vs. strain curve of Mix 7.

131

Figure 2.28

Age dependency of tensile strain capacity.

132

Figure 2.29

Effect of PS beads on compressive strength development.

133

xiv

Figure 2.30

HES-ECC tensile strength development with age.

134

Figure 2.31

HES-ECC Young's modulus (E) development with age.

134

Figure 2.32

HES-ECC tensile strain capacity development with age.

135

Figure 2.33

HES-ECC multiple micro-cracking pattern at different ages.

135

Figure 2.34

HES-ECC compressive strength development with age.

136

Figure 2.35

Compressive

strength

(average

value)

development

of

HES-ECC

compared with requirements by DOTs and FHWA.

136

Figure 2.36

Third-point bending test setup and specimen dimensions.

137

Figure 2.37

HES-ECC flexural stress - displacement curves at 4 hours, 24 hours, and


28 days.

Figure 3.1

137

Variation in tensile stress vs. strain curves of ECC with same mixing
proportion.

170

Figure 3.2

Effect of fiber dispersion on a(5) curve and cmc.

170

Figure 3.3

Experimental program.

171

Figure 3.4

Rheology properties tests: (a) Viskomat NT rotational viscomter (b) marsh


cone flow rate test (c) mini-slump test.

Figure 3.5

171

Specimen preparation for measuring fiber dispersion at the final failure


section.

172

Figure 3.6

Fluorescence imaging technology.

173

Figure 3.7

Fluorescence imaging test setup.

173

Figure 3.8

Marsh cone flow time versus plastic viscosity.

174

Figure 3.9

Mini-slump flow diameter of the seven ECC mixes before/after adding


fibers.

Figure 3.10

174

Tensile stress versus strain curves of ECC with different VMA content.
175

Figure 3.11

Tensile strain capacity vs. VMA/cement ratio (error bars describe standard
error of the data).

Figure 3.12

Figure 3.13

176

Tensile strength vs. VMA/cement ratio (error bars describe standard error
of the data).

176

Fluorescence image of a sample specimen.

177

xv

Figure 3.14

Relation between marsh cone flow time before adding fiber, fiber
dispersion coefficient, and tensile strain capacity.

178

Figure 3.15

Effect of rheology control on tensile properties of WHITE ECC.

179

Figure 4.1

Typical deterioration process of concrete repair.

230

Figure 4.2

Schematics of typical failure modes in layered repair system.

231

Figure 4.3

Illustration of the Kelvin Equation.

231

Figure 4.4

Free drying shrinkage test.

232

Figure 4.5

Experimental data and theoretical curves of hygral deformation vs. relative


humidity.

232

Figure 4.6

Restrained shrinkage steel ring test.

233

Figure 4.7

Restrained shrinkage crack width development as a function of drying


time (at RH = 45 5%).

234

Figure 4.8

Layered repair system test set-up under restrained shrinkage.

235

Figure 4.9

Free shrinkage strain of repair materials at different ages.

235

Figure 4.10

Repair surface cracking at the age of 60 days.

236

Figure 4.11

Interface delamination profiles of repaired systems.

237

Figure 4.12

Specimen end delamination height at different ages.

238

Figure 4.13

FEM model of layered repair system.

239

Figure 4.14

Assumed tensile behavior of (a) HES-Concrete, (b) HES-SFRC, (c) HESECC materials, and (d) combined for all three materials.

Figure 4.15

Interface model: combined uniaxial stress-strain relation in the normal


direction with influence of shear slip.

Figure 4.16

241

Predicted stress distribution (axx) in the (a) HES-Concrete, (b) HES-SFRC


and (c) HES-ECC repaired system at the age of 60 days.

Figure 4.18

240

Predicted crack width of (a) HES-Concrete, (b) HES-SFRC and (c) HESECC repair layer at the age of 60 days.

Figure 4.17

239

242

Predicted interface shear stress (oxy) of (a) HES-Concrete, (b) HES-SFRC


and (c) HES-ECC repaired system.

243

Figure 4.19

Concrete repair subjected to differential shrinkage.

244

Figure 4.20

Simplified tensile stress-strain curve of ECC.

244

xvi

Figure 4.21

Analytical model predicted tensile stress in ECC repair axx, repair/old


interfacial tensile stress a yy and shear stress o xy .

245

Figure 5.1

Crack-and-seated pavement surface.

272

Figure 5.2

Rubblized pavement surface.

272

Figure 5.3

Paving fabric installation.

273

Figure 5.4

HMA interlayer installation.

273

Figure 5.5

Overlay saw and seal.

274

Figure 5.6

Specimen confuguration and loading conditions.

274

Figure 5.7

Specimen preparation.

275

Figure 5.8

Concrete substrate surface preparation.

276

Figure 5.9

"Kinking and trapping" mechanism.

276

Figure 5.10

Flexural behavior of HES-ECC and HES-Concrete layered repair system


under monotonic loading: (a) smooth interface; (b) rough interface.

Figure 5.11

Modulus of rupture (MOR) of HES-ECC and HES-Concrete overlay


systems under monotonic loading,

(a) Overlay age of 6 hours,

Overlay age of 28 days.


Figure 5.12

277

(b)
278

Deflection at failure of HES-ECC and HES-Concrete overlay systems


under monotonic loading, (a) Overlay age of 6 hours, (b) Overlay age of
28 days.

Figure 5.13

279

Interfacial delamination upon failure of HES-ECC and HES-Concrete


overlay systems under monotonic loading, (a) Overlay age of 6 hours.
(b) Overlay age of 28 days.

Figure 5.14

280

Crack pattern of HES-ECC and HES-Concrete overlay systems at overlay


age of 28 days. Epoxy glue was spread over the cracking sections of
HES-ECC overlays (a) and (b), to better visually reveal multiple
microcracking.

Figure 5.15

281

Fatige stress-fatigue life relation (log) for HES-ECC and HES-Concrete


overlay systems.

282

Figure 6.1

Corrosion in Concrete Structures.

312

Figure 6.2

Corrosion process of reinforcing steel according to Tuutti.

313

xvn

Figure 6.3

Coefficient of permeability versus crack width for ECC and reinforced


mortar series deformed to 1.5% in uniaxial tension.

313

Figure 6.4

Typical tensile stress-strain response of ECC mixture.

314

Figure 6.5

Immersion test, (a) Cylinders were stored under continuous exposure to


3% NaCl solution, (b) One cylinder was taken out, split open, and sprayed
A

with 0.1-N silver nitrate (AgN03) solution, to measure the chloride


penetration depth.

315

Figure 6.6

Four point bending test.

315

Figure 6.7

Salt ponding test on uncracked and cracked beam specimens.

316

Figure 6.8

Typical crack pattern on positive moment surface of ECC beams at 2 mm


(0.079 in.) deformation.

317

Figure 6.9

Chloride penetration depth variation measured by immersion test.

317

Figure 6.10

Chloride profiles of uncracked mortar and uncracked ECC prisms after 30


and 90 days in 3 % NaCl solution.

Figure 6.11

Chloride profiles of mortar and ECC prisms in cracked zone at 30 days


exposure.

Figure 6.12

318

319

Diffusion coefficient versus crack width for mortar deformed under


bending load.

320

Figure 6.13

Diffusion coefficient versus number of cracks for ECC.

320

Figure 6.14

Diffusion coefficient versus pre-loading deformation level for ECC and


mortar.

Figure 6.15

321

Self-healing products in ECC microcracks before and after salt ponding


test at 30 days exposure.

322

Figure 6.16

ESEM micrograph of rehydration products in a self-healed crack.

322

Figure 7.1

Comparison of allowable crack widths under marine exposure.

344

Figure 7.2

Typical pre-cracking tensile stress-strain curves of ECC.

344

Figure 7.3

Tensile stress-tensile strain curves of ECC specimens before and after


exposure to 3% NaCl solution.

Figure 7.4

346

Influence of NaCl exposure time and applied strain level on ECC ultimate
tensile strength.

347

xvm

Figure 7.5

Influence of NaCl exposure time and applied strain level on ECC tensile
strain capacity.

Figure 7.6

347

Influence of NaCl solution exposure time on ECC matrix fracture


toughness.

Figure 7.7

348

Influence of NaCl solution exposure time on ECC matrix/fiber interfacial


chemical bond.

Figure 7.8

348

Influence of NaCl solution exposure time on ECC fiber/matrix interfacial


frictional bond.

349

Figure 7.9

Influence of NaCl solution exposure.time on ECC PSH Index.

349

Figure 8.1

PVA fiber, 0.5 in. (12.7 mm) and 0.33 in. (8.4 mm).

381

Figure 8.2

Tensile stress-strain curves of HES-ECC with 0.33 in. (8.4 mm) PVA fiber.
381

Figure 8.3

Loss of deformability of HES-ECC with and without hydration stabilizer.


382

Figure 8.4

Gravity mixer with capacity of two cubic feet (0.057 cubic meters).

382

Figure 8.5

HES-ECC one cubic feet (0.03 cubic meter) mixing.

384

Figure 8.6

Loss of deformability of HES-ECC processed in one cubic feet (0.03 cubic

meter) batching.
Figure 8.7

384

Compressive strength development of HES-ECC from one cubic feet


batching.

Figure 8.8

385

Tensile stress-strain curves of HES-ECC from one cubic feet batching.


385

Figure 8.9

Gravity mixer with capacity of 9 cubic feet (0.25 cubic meters).

386

Figure 8.10

HES-ECC six cubic feet (0.17 cubic meters) mixing.

388

Figure 8.11

Loss of deformability of HES-ECC processed in six cubic feet (0.17 cubic


meters) batching.

Figure 8.12

389

Compressive strength development of HES-ECC from 6 cubic feet (0.17


cubic meter) batching.

Figure 8.13

Figure 8.14

389

Tensile strain capacity development of HES-ECC from 6 cubic feet (0.17


cubic meters) batching.

390

Uniaxial tensile test plate dimensions.

390

xix

Figure 8.15

Ellsworth Road Bridge over US-23 (S07 of 81074).

391

Figure 8.16

Bridge deck cracking.

691

Figure 8.17

Partial depth removal.

391

Figure 8.18

Sandblasting of patch area.

392

Figure 8.19

Replacement of damaged reinforcement.

3 92

Figure 8.20

Completed patch repair preparation.

392

Figure 8.21

Separated patch areas for HES-ECC repair and Thoroc 10-60 repair.

393

Figure 8.22

12 cubic foot (0.34 cubic meter) capacity gas mixer.

393

Figure 8.23

Pouring HES-ECC into wheelbarrow for transportation.

393

Figure 8.24

HES-ECC was poured into patch.

394

Figure 8.25

HES-ECC shows self-compacting property.

394

Figure 8.26

Placement of Thoroc 10-60.

394

Figure 8.27

Comparison of HES-ECC patch repair and Thoroc 10-60 patch repair.


395

Figure 8.28

Completed patch repair work.

Figure 8.29

Maximum surface crack width of HES-ECC and Thoroc 10-60 patch


repairs.

Figure 8.30

395

396

Maximum interfacial delamination width of HES-ECC and Thoroc 10-60


patch repairs.

396

Figure 8.31

Surface cracking in HES-ECC patch repair, on September 18, 2009.

397

Figure 8.32

Surface cracking in Thoroc 10-60 patch repair, on September 18, 2009.


398

Figure 8.33

Interfacial delamination between HES-ECC patch repair and existing


concrete, on September 18, 2009.

Figure 8.34

Interfacial delamination between Thoroc 10-60 patch repair and existing


concrete, on September 18, 2009.

Figure 8.35

399

399

Exposure of aggregates in Thoroc 10-60 patch repair close to the Thoroc


10-60/HES-ECC interface area, on September 18, 2009.

xx

400

CHAPTER 1

INTRODUCTION

Concrete infrastructure supports the operation of society, and provides the


services and facilities necessary for an economy to function1.

Concrete is the most

heavily used man-made material in the world2. It has been the predominant construction
material for centuries because of its widely available ingredients (e.g. cement, sand,
gravel, water), capacity for taking on different forms, and low cost. Annually, more than
one cubic meter of concrete per person on Earth is made3, and more than 380 million
cubic meters of concrete (almost 1.5 cubic meters per person) are placed every year to
support the U.S. infrastructure4. The volume of in-place concrete in the U. S. is estimated
at 7 billion cubic meters (25 cubic meters per person) and most of the concrete is more
than 20 years old4.
Concrete infrastructure, both new and repaired, is suffering deterioration, damage,
and defects 5 ' 6 ' 7 ' 8 (Figure 1.1).

Concrete deteriorates over time under service loading

conditions when subjected to combined environmental and mechanical conditions. For

1
c

example, saturated concrete exposed to freeze-thaw cycles may experience disintegration;


Restrained shrinkage often causes early-age cracking in concrete; Chloride penetration
from deicing salts or seawater through cracks, carbonation of concrete, or inadequate
concrete cover depth can result in reinforcement corrosion; Expansion of corrosion
products can cause spalling and delamination of the concrete cover; Mechanical loading
such as fatigue combining with environmental effects will accelerate and aggravate the
deterioration process. Concrete damage is normally caused by short-term severe loading
conditions such as fire, earthquakes, chemical spills, overloading, or impact. This may
result in cracking, immediate spalling, disintegration, or complete failure of the concrete
element or structure.

Concrete defects are caused by improper detailing or design,

construction practices, human error, or materials with improper quality control that result
in premature deterioration, aesthetic issues, or inadequate structural capacity. The lack of
durability in concrete infrastructure due to deterioration, damage, and defects drives a
repair industry that supports researchers, engineers, architects, material manufacturers,
equipment suppliers, testing companies, contractors, lawyers, and educators. For owners,
the annual cost of concrete repair (including protection and strengthening) is estimated at
$18 billion to $21 billion in the U.S. alone9.
Lack of durability in concrete infrastructure has become a looming threat that
could jeopardize the world's prosperity and our quality of life. Although infrastructure
deterioration might not be as dramatic as damage from fires or earthquakes, the
magnitude of this problem in terms of structural life cycle economic, social, and
environmental impacts under service loading conditions dwarfs those associated with
failures under severe loading conditions. The economic impacts are associated with the

high cost of maintaining, repairing or replacing deteriorating structures, which is


exceedingly burdensome on building owners, departments of transportation, the
insurance industry and society. According to the American Society of Civil Engineers
(ASCE) 2009 Report Card10 for U.S. Infrastructure, an average grade of D (poor) was
assigned over 12 infrastructure categories. ASCE estimated that USD 2.2 trillion is
needed over the next five years for repair and retrofit.

Correspondingly, repair and

retrofit cost has been estimated to be USD 2 trillion for Asia's infrastructure. In Europe,
Japan, Korea, and Thailand, the annual cost for repair has exceeded that for new
construction11.

The social impacts include construction-related congestion, operations

interruption, wasted time and fuel cost, vehicle damage due to bad road conditions, and
safety concerns. Environmental sustainability, which has become a compelling issue
with the current depletion of raw materials and climate change, has been greatly
compromised by infrastructure deterioration, repair and reconstruction cycles, which
involve significant amounts of raw materials and energy consumption as well as
greenhouse gas emissions. To battle these impacts, an effective repair methodology that
addresses the underlying concrete durability problems and extends the service life of
deteriorating concrete structures is urgently needed.
While countless materials and techniques are used in practice to meet the demand
for rapid, inexpensive and durable concrete repairs, few of them target the inherent
material shortfall of concrete as a brittle material that cracks and fractures in many
applications, leading to further deterioration. It has been estimated that almost half of all
concrete repairs fail prematurely12. Many undesirable repair behaviours can be observed
in the field in the form of early age surface cracking, spalling, or interface delamination

between the repair and the concrete substrate. Delamination and cracking provide routes
for chlorides, oxygen, moisture, alkali, and sulphates to penetrate into the repaired system
and accelerate further deterioration. Additionally, the loss of structural integrity impairs
load transfer between the repair and the concrete substrate. As a result, the repaired
concrete structure with unsatisfactory performance and a short service life must be further
maintained or repaired, leading to repeated repairs and significantly increased structural
life cycle cost.
The Kurtsubo Bridge is a case in point. This bridge, located in a coastal area in
Japan, experienced severe corrosion of the post-tensioned tendons13 due to ingress of
chloride through the concrete cover (Figure 1.2). After the first repair the corrosion was
not inhibited and the deterioration continued, until severe corrosion and breakage of the
tendons necessitated the second repair. After two repairs and one rehabilitation (the latter
is more targeted on increasing structural capacity besides restoring serviceability14), the
bridge was finally replaced. Consequently, after 40 years of service life, the total cost of
the prematurely deteriorated bridge was 3.8 times greater than the initial cost of its design
and construction (Figure 1.3).

The life cycle cost analysis of the Kurtsubo Bridge

illustrates that the lack of durability of repairs can result in high maintenance costs that
greatly exceed initial construction costs.

1.1

Deterioration in Concrete Structures - Major Durability Concerns


Concrete structure deterioration results from the combined effects of various

environmental and mechanical loading conditions throughout the structure's service life.
In this section, common deterioration modes in typical concrete structures, i.e. bridges,

roadways and highways, and parking structures, are reviewed. According to Emmons
and Sordyl 20064, the total annual cost of repairs, rehabilitation, strengthening, and
protection of concrete structures in the U.S. is $18-21 billion, with $8 billion, $4 billion
and $1 billion needed for bridges, highways and roadways, and parking structures,
respectively.

1.1.1 Bridges
Repair need and cost for bridges in the U.S. has been rising as the bridge fleet age
increases. In the U.S., bridges were typically built with an expected service life of 50
years, and their average age is now 43 1 5 .

According to the U.S. Department of

Transportation, of the 600,905 bridges in the U.S. as of December 2008, 12.1% were
categorized as structurally deficient, and 14.8% were categorized as functionally
obsolete16. A structurally deficient bridge may either be closed or restrict traffic due to
limited load carrying capacity. Although not unsafe, they pose limits to vehicle speed and
weight. A functionally obsolete bridge has outdated design geometries and features.
Although not unsafe, they cannot accommodate current traffic volumes, vehicle sizes and
weights. These structurally deficient and functionally obsolete bridges require frequent
maintenance, repair, and reconstruction, leading to construction-related traffic congestion
and lengthy detours, and eventually resulting in significant economic, social, and
environmental costs to society17. To address bridge needs, AASHTO18 estimated in 2008
that the repair cost for bridges in the U.S. would amount to $140 billion - about $48
billion to repair structurally deficient bridges and $91 billion to improve functionally
obsolete bridges.

Bridge deck transverse cracking


Transverse cracking (Figure 1.4) in bridge decks is one of the dominant
deterioration problems in bridges in many states in the U.S. ' ' ' ' ' . It is estimated
that more than 100,000 bridges in the United States have early-age transverse cracks ,
and these cracks have been observed in most geographical locations on many
superstructure types22. Some of these cracks form right after construction, and some after
the bridge has been opened to traffic for a period of time. Transverse cracks usually form
when concrete is set ' ' , and then widen with time. These cracks are typically full
depth, are often observed over transverse reinforcement ' ' ' ' , and are spaced 1-3 m
apart along the length of the bridge ' ' . These cracks reduce the service life of the
bridge deck, and increase maintenance and repair costs during the bridge's life cycle.
They accelerate chloride penetration into the concrete and initiate corrosion of
reinforcements, especially in marine environments or regions where deicing salts are
applied in winter26'29.

Corrosion damage has even been observed on epoxy coated

reinforcing bars in bridge decks with such cracks ' . Furthermore, freeze-thaw cycles of
water in cracks and leakage of water to supporting structures may also reduce the service
life of a bridge.
The cause of bridge deck transverse cracking is tensile stress induced by
restrained volume change of concrete, such as drying shrinkage, autogeneous shrinkage,
thermal effects, and creep. The magnitude of volume change in concrete depends on its
ingredients, composition, and environmental conditions, such as ambient temperature and
humidity. The degree of restraint, which is most commonly due to the deck and girder
composite behavior, depends on bridge design characteristics, such as boundary

conditions and relative stiffness of the superstructure. Although the relative importance
of these factors is not completely quantified yet, drying shrinkage is considered by many
99

studies to be the major cause of concrete deck cracking

9^

"^ 1 ^ 9

"\~X "XA "} ^ "\f\

' ' > _

Although transverse cracking in bridge decks has been one of the main concerns
of designers and researchers for decades, effective mitigation procedures have not been
developed19'20. Most research studies on the causes of transverse cracking over the past
several decades have focused on three main categories of factors: (i) material and mix
design, and their effects on cracking tendency, (ii) construction practices, and (iii)
structural design factors.
In terms of material and mix design, it is recommended to: (a) reduce cement
content to 385-390 kg/m3 22' H

25 32 35 31 37

' '

' ' , (b) use AASHTO specification Type II

9 9 9 4 ^ 9 ^"X "\R

cement for bridge deck construction ' ' ' ' (c) limit the water/cement ratio to 0.4-0.45
and use water reducers to decrease water content ' ' ' ' ' ' , (d) use fly ash, (e) avoid
using high early strength concrete if "early-opening" is not an issue , (f) use the largest
possible aggregate size as specified in ACI-318, use crushed stone as a coarse aggregate,
9 ^ 9A ^ 1 9 9

"XQ "X^i 1 8

A(\

and maximize aggregate content ' ' ' ' ' ' , and (g) increase concrete strength .
In terms of construction practices, it is recommended
42,43,44,45 tQ . ^
temperatures,

usg

evap0rati0n

22 23 24 27 28,31 35 36 37 38 39 41

' ' ' '

' ' ' ' ' ' '

rate chart proposed by ACI and cast the deck in mild

(b) start curing immediately after finishing, then cure for at least 7

consecutive calendar days or 14 days if the construction schedule allows, (c) pour the
complete deck at one time whenever feasible, within the limitation of the maximum
placement length, based on drying shrinkage considerations; if multiple placements must
be made and the bridge is continuous span, then place concrete in the center of positive
7

moment regions first and observe a 72 hour delay between placements; when deck
construction joints are created, prime existing interface surfaces with a bonding agent
prior to placement of new concrete.
In terms of design factors, the following recommendations19'40'46'47'48'49'50'51 are
made: (a) boundary restraints should be consistent with the design - construction
practices should not introduce undue boundary restraints on the girders, (b) minimize the
ratio of girder/deck stiffness by adjusting deck thickness, girder spacing, and girder
moment of inertia, (c) employ flexible superstructures, and (d) limit the skew angle of
bridges to prevent corner cracking. The increase of reinforcement volume above code
requirements is discouraged because research19 shows that it does not help mitigate
cracking.
Despite a significant effort among the bridge engineering community to eliminate
or at least reduce transverse bridge deck cracking, such cracking still exists and poses a
great challenge to engineers and researchers48'52. Most of the past efforts have focused
on modifying concrete material properties and construction practices to reduce volume
change in concrete, especially drying shrinkage. Such efforts, however, have limitations
because they don't fundamentally address the inherent brittleness of concrete materials.
Therefore, the potential of reducing concrete volume change to a level less than the
material's deformation capacity is marginal and unreliable 12 ' 53 ' 54 , resulting in (often
"unexpected") cracks, especially when the combined effects of drying shrinkage,
temperature change, and mechanical loading are present.

The same brittle material

limitation also leads to early-age cracking in partial-depth patching and full-depth


concrete repairs of existing bridge decks with transverse cracks 55 ' 56 , which has a major

impact on the long-term durability of a repair system. Early-age cracking often occurs as
a result of restrained volume change of the new repair material, due to its dimensional
incompatibility with existing concrete 57 ' 58 ' 59 ' ' 61 . These shortcomings point to an urgent
need for developing a new material technology that effectively reduces the cracking
potential through removal of the material brittleness, and eventually eliminates restrained
volume change induced cracking in new and repaired bridge structures.

Reinforcing and prestressing steel corrosion


Corrosion of reinforcing or prestressing steel (Figure 1.5) is another pressing
deterioration problem in bridge structures and other reinforced concrete structures,
demanding significant amounts of repair and rehabilitation62'6 . Corrosion starts when
chloride ions transport through the sound or cracked concrete cover, reach the steel, and
accumulate beyond a critical concentration level (pH falls below 11). At this point the
natural protective concrete layer surrounding the steel is depassivated. Corrosion then
initiates if oxygen and moisture are present at the steel-concrete interface. The oxidation
products generate expansive forces within the concrete, eventually causing the concrete
cover to spall64.

Every year the U.S. spends about $8.3 billion to directly address

corrosion in American highway bridges65. Indirect costs related to traffic delays and
productivity losses are estimated to exceed 10 times the direct cost of maintenance, repair,
and rehabilitation.
According to the FHWA Office of Research, Development & Technology66,
reinforced concrete bridges functioned reasonably well until the late 1960s, when
premature concrete delamination and spalling, which used to be encountered only in
coastal areas, became common in many of the reinforced concrete decks in the "snow

belt."

Concrete bridge decks were beginning to require maintenance after being in

service for as little as 5 years. The emergence of this type of concrete deterioration,
which was first observed in marine structures and chemical manufacturing plants,
coincided with the increased application of deicing salts to bridges and roads during
winter months. By the mid 1970s, it was recognized that this problem is caused by
corrosion of reinforcing steel in concrete, which is induced by the intrusion of even small
amounts of chloride from the deicing salts into the concrete. Now it is widely recognized
that corrosion of the reinforcing steel has led to premature deterioration of many concrete
bridges in the United States before their design life is exceeded. Although corrosion is
not the sole cause of all structural deficiencies, it is a significant contributor and has
become a major concern to transportation agencies nationwide67'68.
The magnitude of this corrosion problem in transportation infrastructure systems
has increased significantly in the last three decades and will continue to increase69. Even
though the cost of maintaining bridge decks is becoming prohibitively expensive, the
benefits to public driving safety provided by deicing salts are so important that their use
is not likely to decrease in the future66. In fact, the use of road deicing salts, which are
extremely corrosive due to the disruptive effects of its chloride ions on protective films
on metals, has increased in the first half of the 1990s after leveling off during the 1980s70.
Although an effective alternative deicing agent, calcium magnesium acetate (CMA), is
available and less corrosive, its high price greatly limits its wide-spread application71'72.
Therefore, it can be expected that the road environment will likely remain corrosive. In
response to the tremendous economic burden that corrosion of reinforcing steel in
concrete bridges placed on the national economy, the Structure Division of FHWA has

10

placed emphasis on finding effective and economical solutions that can be easily
implemented by state and local transportation agencies.
Compared to reinforcing steel, corrosion of prestressing steel and its impact on
structural performance can be more severe. Prestressed concrete technology was only
applied to bridges relatively recently.

Therefore, the existing prestressed concrete

members in bridges are still relatively young, and the corrosion and concrete
deterioration problems associated with prestressed concrete members didn't become
evident until the early 1980s73. Although prestressed concrete members were generally
manufactured with higher strength concrete, time has shown that they are subject to the
same adverse effects of reinforcement corrosion as reinforced concrete members74.
Because the cross section of each prestressing wire or strand is relatively small and is
already under significant stress, a very small cross sectional loss from corrosion will
cause the wire or strand to debond from the concrete and eventually break75.

Since

prestressed concrete members rely on the tensile strength of the strands to resist loads,
loss of even a few strands per member can be catastrophic76'77. Additionally, because the
strands are subjected to high stress, corrosion effects can be accelerated78. Even small
corrosion pits may cause a strand to fracture, as compared to non-prestressed reinforcing
steel that will literally rust away before breaking. Documented cases of prestressing
strands breaking as a result of corrosion make this a pressing problem.
The rapid premature deterioration of many concrete bridge decks in the late 1960s
raised concern among state highway agencies. As a result, the use of non-corroding
fusion bonded epoxy-coated reinforcement has become a standard practice since the late
1970s79. For additional protection, low-permeability concrete, such as low water-cement

11

ratio Portland cement concrete, latex-modified concrete, and other specialty concrete
began to be used, along with increased concrete cover thickness over coated reinforcing
steel80.

Additionally, the use of waterproof membranes over concrete decks81, in

conjunction with an asphalt overlay, has produced mixed results and been adopted
sparingly. A few states have used epoxy-coated rebars in conjunction with a waterproof
membrane and asphalt overlay on bridge decks.
FHWA research and field data have indicated that the use of overlays, waterproof
membranes, and sealers only marginally extend the life of a structure, to a variable
extent66. On the other hand, cathodic protection has proven to be successful at retarding
and

controlling

chloride-induced

components ' ' ' ' .

corrosion

in

reinforced

concrete

bridge

In addition, electrochemical chloride extraction is another

alternative rehabilitation technique that is being explored by a number of state highway


agencies89. More research is under way to bring this technology up to par with cathodic
protection.
Current approaches to preventing steel reinforcement corrosion in concrete
bridges have mixed results66. These approaches include the use of good construction
design and procedures, adequate concrete cover depth, corrosion-inhibiting admixtures,
and the adoption of low-permeability concrete. The reason for the limited success is due
to the fact that concrete is brittle and tends to crack in a variety of combined mechanical
and environmental loads.

In fact, it has been observed recently that the new low

permeability concrete (high-performance concrete), which is made from partial


substitution of Portland cement with silica fume or fly ash, has a more pronounced
tendency than conventional concrete to crack ' 9 ' 91 . As a result, a faster, gravity-assisted

12

flow of salt-laden water channels through the cracks, rather than a slow diffusion of
chloride through the uncracked concrete. Corrosion-inhibiting admixtures for concrete
will not be useful once concrete cracks , and the reinforcing steel solely relies on itself
as the last line of defense against corrosion. In this situation, the use of a barrier system
on the reinforcing steel, such as an epoxy coating or other organic coating, becomes more
critical in abating corrosion. However, it is likely that there may never be an organic
coating that can resist the extreme combination of constant wetting, high temperature,
and high humidity that reinforcing steel is exposed to in marine environments, especially
in the splash zone in coastal regions66. Epoxy-coated steel bars will have to be used in
conjunction with sound, crack-free concrete in bridge decks89 where the concrete is not
constantly wet and the other exposure conditions are not as severe. The limitations of
current approaches to reducing steel corrosion in concrete calls for a fresh approach in
tackling this problem - a new methodology that addresses the intrinsic brittleness of
concrete material for the purpose of enhancing structural durability.

1.1.2 Highways and Roadways


Highway and roadway infrastructure involve enormous material consumption,
energy usage, and capital investment. Currently, the U.S. consumes more than 35 million
metric tons of asphalt and 48 million metric tons of cement annually, at a cost of nearly
$65 billion in the transportation infrastructure system alone. Repair and reconstruction of
deteriorated and aging highway and roadway infrastructure in the United States leads to
significant economic cost, environmental burdens, and social costs, which motivate
sustainable engineering initiatives. Highway and roadway systems should provide a safe,

13

reliable, efficient, and comfortable driving environment.

While the transportation of

people and goods has expanded significantly in recent decades due to the existence of an
extensive highway and roadway system in the U.S., deteriorating and aging highway and
roadway pavement systems have become a burden to society. Poor road conditions lead
to excessive wear and tear on motor vehicles and increased numbers of crashes and
delays, and cost U.S. motorists $67 billion ($333 per motorist) annually in vehicle repair
and operating costs. Currently, 33% of America's major roads are in poor or mediocre
condition and 36% of the nation's major urban highways are congested93. The current
spending level of $70.3 billion for highway capital improvements9 is well below the
estimated $186 billion needed per year to substantially improve the nation's highways .
Typical deterioration mechanisms in highways and roadways include transverse
cracking, longitudinal cracking, D-cracking, and alkali silica reaction9'9 ' 9 7 Transverse
cracking results from restrained volume change, such as drying shrinkage and
temperature gradient, similar to the causes of transverse cracking in bridge decks98' 9.
These environmental effects, combined with repeated vehicle loading, aggravate the
deteriorated area and further increase cracking and spalling. Longitudinal cracks are
caused by a combination of repeated heavy loading, loss of foundation support, and
curling and warping stresses, or by improper construction of longitudinal joints.
Durability, or "D", cracking manifests as a "closely spaced crescent-shaped hairline
cracking pattern", and occurs near joints, cracks, and free edges100. D-cracking is caused
by freeze-thaw expansion of some types of coarse aggregate, and occurs frequently with
limestone in Midwestern states in the United States. The alkali silica reaction occurs
between alkali hydroxyl ions in Portland cement and certain siliceous minerals found in

14

the aggregate, such as opaline chert, and strained quartz. This reaction produces a gel
that surrounds the aggregate in the concrete matrix. With the introduction of moisture,
the gel expands and causes cracking and distress of the pavement101.
In developed countries such as the U.S., much of the highway and roadway
construction activities involve maintenance, repair, and reconstruction of existing
pavements.

Three main types of repair methods have been used: (i) partial-depth

patching (partial-depth repair), (ii) full-depth repair, and (iii) unbonded concrete overlay
or hot mixed asphalt (HMA) overlay.
Partial-depth patching is a pavement restoration technique that repairs surface
distress and localized failures that do not extend to the bottom of the slab, re-establishes
well-defined joint reservoirs, or reduces impact loading at transverse joints, with the goal
of extending pavement service life1 . More specifically, it is used for mid-slab surface
cracking and spalling, joint spalling, localized scaling, and localized distress from
corroding steel reinforcement. Partial-depth patching improves the rideability of concrete
pavement, and reduces intrusion of moisture, chloride, or incompressibles into cracks or
joints.
Full-depth repair is for joint deterioration larger than 1/3 of slab depth, severe
transverse cracking, longitudinal cracking, and broken slabs or corner breaks103. Partialdepth and full-depth repairs are subjected to restrained volume change induced cracking
and debonding from existing concrete, which are most frequently caused by inappropriate
selection of repair materials with higher cracking tendency, material variation, poor
construction practice and quality control, late curing, and lack of bond between the repair
and the existing concrete.

15

Concrete or HMA overlay is one of the major rehabilitation methods used for
severely deteriorated pavements104. Reflective cracking is often the ultimate failure mode
that limits the overlay service life105'106'107. Reflective cracking is characterized as new
cracks reflected through the overlay from pre-existing cracks or joints in the substrate
(Figure 1.6) under repeated traffic loads. Current methods to abate reflective cracking
include the use of a bond-breaking layer, i.e. HMA, between the concrete overlay and the
substrate, to release stress concentration at the source of reflective cracking, rubblization
(breaking the substrate pavement into a well graded high modulus aggregate base
material), or extensive repair of existing pavement deterioration before overlaying.
Currently, none of these approaches has been able to effectively eliminate reflective
cracking in rigid (concrete) pavement overlays, mainly due to the brittle nature of the
overlay material.

1.1.3 Parking Structures


Apart from bridge decks and highway pavements, parking structures represent
another category of structures that suffers early deterioration (Figure 1.7). A combination
of parking lot, building, and bridge structures, the modern parking structure108 is designed
to meet an ever-expanding array of criteria to support vehicular traffic loads while
enduring challenging environmental conditions. A parking structure frame can be made
of cast-in-place concrete, pre-cast concrete, or steel.

A parking structure deck is

commonly cast-in-place or pre-cast concrete. The concrete members are either reinforced
or pre-stressed1 .

Cast-in-place decks are usually a single, continuous reinforced

concrete deck spanning between steel or concrete framing. This type of deck is generally

16

more prone to deterioration from restrained shrinkage cracking, chloride attack, and
corrosion induced spalling110'111'112.

Therefore, performance of this type of deck is

directly related to provisions that mitigate chloride attack1 3. While there is no substitute
for good design and materials, combating chloride attack has become a challenge over the
years 114 ' 115 ' 115 .

Current protection methods include reinforcement coating, adding

corrosion indicators to concrete, applying sealers and densifiers to the concrete surface,
and using traffic bearing membranes 117 ' 118 ' 119 ' 120 . For decks that are left unprotected,
maintenance dollars are spent chasing defects

Pre-cast deck elements are cast in a factory and use dense and high quality
concrete produced with strict quality control methods. Therefore, they are expected to
have higher resistance to chloride penetration 08. Because these members are typically
pre-stressed in compression, shrinkage cracking should not be of concern.

However,

maintenance requirements for pre-cast parking garages can be far more demanding than
for cast-in-place parking garages. This is because pre-cast elements are often thinner and
have a low tolerance against construction caused defects.

They tend to crack when

1 99 19^

restrained from thermal expansion and contraction

'

. Cracks, once formed, pose a

serious threat to the pre-cast elements as they allow a direct path for chloride to penetrate
through and reach the reinforcement, despite the dense microstructure of pre-cast
members.

As the pre-cast members typically use wire that is thinner than the

reinforcement bar used for cast-in-place decks, a relatively small amount of corrosion can
cause a significant loss of load carrying capacity.

Even worse, repair of these thin

elements is often difficult and costly123. Moreover, pre-cast members typically require
sealant joints as connections. These joints, extending around each pre-cast member, can
17

very quickly add up to miles of sealant that must be maintained and replaced periodically.
Water migration through failed joints can cause problems for cars below and accelerate
deterioration of the pre-cast members124.
In summary, water leakage through cracks and joints, chloride ingress, and freezethaw cycles are common mechanisms that cause reinforcement corrosion, spalling, and
other distress in parking structures. The risk associated with deterioration in parking
structures is severe, and the threat of costly and time-consuming lawsuits associated with
driving and tripping hazards is substantial.

Current repair methods include use of

waterproofing membranes, joint sealants to control water intrusion, strengthening


systems to correct design and construction errors, surface repair systems for spalling
damage, and slab replacement.

1.2

Limitation of Current Approaches to Improve Durability of Concrete

Structures
Summarizing from the previous discussion, early-age cracking due to restrained
volume change, reflective cracking, chloride penetration, corrosion, spalling, and alkali
silica reaction are major deterioration phenomena in new and repair concrete structures.
At the root of these durability concerns is the inherent brittleness of concrete materials.
Concrete tends to crack and spall under applied structural loads, restrained volume
change (shrinkage, thermal deformation, etc.), and stress concentration induced by
existing cracks in the substrate concrete. Once cracked, wider crack width tends to
exacerbate water, gas, and chloride ion transport and accelerate the deterioration
mechanisms mentioned above. The factors driving wider crack width include the level of

18

deformation (determined by ambient environmental conditions and structural loading


conditions), constraints (determined by boundary conditions), member geometry, and
material characteristic length. Crack width has a direct influence on transport properties.
Large cracks easily allow water and corrosives into the material and accelerate
deterioration, such as leeching, steel corrosion, spalling, sulfate attack and freezing-andthawing damage.
Current methods to increase durability of concrete structures include (a)
increasing density of concrete material microstructure through well-graded particle size
distribution to reduce permeability and transport of corrosives to the steel125'126, (b)
increasing concrete compressive strength, (c) modifying the concrete matrix by adding
entrained air, sulfate resistance cements, and corrosion-inhibiting admixture, and (d)
adding cracking-control reinforcement.

These methods, except (d), do not target the

inherent brittleness and cracking behavior of concrete, and instead rely upon the concrete
to remain uncracked within a structure throughout its service life so that the denser
concrete matrix is able to resist penetration of aggressive agents. As discussed above,
concrete cracking behavior and crack width play the most important role in a majority of
the concrete durability problems in the field. Therefore, a fundamental solution, which
reduces the brittle nature of concrete and controls crack width while maintaining cost
effectiveness and ease of application, is necessary.

1.2.1

Effectiveness of current material "durability" solutions


It is often perceived that the solution to the repair durability problem is improving

the compressive strength of the repair material or accelerating its strength gain at material

19

early ages.

This misconception encourages the seeking of various expensive, high

strength repair materials that do not behave as expected, or sometimes even worse in the
field12. High strength concrete (with 50-140 MPa compressive strength), for example, is
often believed to be a highly durable material because of its very dense microstructure,
which lowers permeability and reduces transport of corrosives to the steel 25 '

. This is

achieved with low water/cement ratios 7, silica fume and fly ash 28, and a well-graded
particle size distribution129. This concept, however, is based on the assumption that high
strength concrete remains uncracked throughout the structure's expected service life, so
that it can resist the ingress of water, chloride ions, oxygen, and other aggressive agents
through its dense microstructure.
In reality, however, high strength concrete is particularly susceptible to cracking
i

due to its high brittleness and small creep deformation

i n

. Containing more finely ground

particles than regular concrete, high strength concrete typically has increased early age
shrinkage and thermal volume change. Restrained volume change induces large tensile
stress in the high strength concrete because of its high elastic modulus and small creep
deformation, consequently resulting in cracking. Cracks provide paths for the aggressive
agents to penetrate into the high strength concrete, and ultimately lead to steel corrosion,
freezing and thaw disintegration, alkali or sulfate attack12.
The higher cracking tendency of high strength concrete can be explained using
fracture mechanics theory. Stang and Li131 investigated the cracking stability of concrete
material under restrained shrinkage conditions using a simple model of a concrete slab of
length L, restrained on both ends as shown in Figure 1.8. When the shrinkage strain
Eshrink stays below concrete's elastic tensile strain capacity et) no cracking occurs,
20

although tensile stress develops in the restrained direction.

Once et is exceeded, the

magnitude and stability of the crack opening depend on the brittleness of the material.
For an ideally brittle material, the crack width will instantly jump from zero to a crack
width equal to the shrinkage strain times the length of the slab, as shown in Figure 1.9. If
the concrete material has tension softening behavior, then the predicted crack width is
dependent on the material characteristic length lCh through Equation 1.1.
L
w=

i shrink

~,)

,lV)

(\-LI2lch)
where w is the crack width, L is the length of the restrained member, Shrink is the
shrinkage strain, et is the elastic tensile strain capacity, and lCh is the characteristic length
as defined by Hillerborg et al.132 in Equation 1.2.
EG

l*d-

(1-2)

Jt

where lCh is the characteristic length, E is the modulus of elasticity, Gf is the fracture
energy, and ft is the tensile strength.
Figure 1.9 suggests that the crack width will grow from zero linearly proportional
to the excess of shrinkage strain over the tensile strain capacity (es - et). This slope has
two limits. As characteristic length drops to L/2, the slope becomes infinity and the
material behaves as an ideally brittle material by conforming immediately to the Les line.
As characteristic length increases to infinity, the slope approaches L. This demonstrates
that for the same shrinkage potential (es - et), a material with small characteristic length
(i.e. very brittle with low toughness) will show larger crack widths. The development of
cracks also tends to be unstable, as indicated by the infinite slope of the line. A material
with higher toughness and characteristic lengths will have smaller crack widths and more
21

stable crack development. Cracking of a quasi-brittle material will fall between these two
limits.

Characteristic length values for ordinary Portland cement concrete and high

strength concretes are calculated to be about 190 and 120 mm (7.5 and 4.7 in.),
respectively131. Because high strength concrete has lower material characteristic lengths,
it should exhibit larger cracks after early age shrinkage.

This is supported by field

evidence of very early age cracking of bridge decks constructed using high strength
concrete in Colorado and California133'134'135, highlighting the susceptibility of high
strength concretes to early age cracking due to restrained volume change.
Based on the above discussion, the effectiveness of current solutions for concrete
structure and repair durability is questionable. Instead, material durability should be
highly dependent on its cracking resistance and cracking stability. A material with high
tensile strain capacity (ductility) will be able to accommodate restrained volume change
due to shrinkage, temperature effects, structural loading, etc., and relieve the tensile stress
buildups in the repair without cracking failure.

1.2.2

Compatibility of new repair with existing concrete substrate


Unlike new structures, the durability of a repaired system highly depends on the

compatibility of the repair material with existing concrete136. Compatibility is defined as:
"a balance of physical, chemical and electrochemical properties, and dimensions
between a repair material and the existing substrate that will ensure that the repair
can withstand all the stresses induced by volume changes and chemical and
electrochemical effects without distress and deterioration over a designated period
of time."4
Dimensional compatibility, which is the most important among these considerations, is
the ability of the repair to withstand restrained volume change without cracking and loss

22

of bond. It also refers to the ability of the repair to diffuse stress concentration from
existing cracks in the substrate concrete without reflective cracking.

These types of

cracking, either due to restrained volume change or reflective cracking, are the main
causes of reduction in service lives of bridges, highways and roadways, parking
structures, and other concrete structures, as discussed earlier in Section 1.1.
Dimensional compatibility includes drying shrinkage, thermal expansion or
contraction, creep, modulus of elasticity, and repair geometry. An ideal repair material
should undergo zero shrinkage, and the same amount of thermal expansion or contraction,
creep, and modulus of elasticity as the existing surrounding concrete.

In reality,

shrinkage always takes place in a new repair as soon as it is installed, and is restrained by
existing reinforcements and the surrounding old concrete, which already underwent
shrinkage years ago. Restrained shrinkage results in tensile stress in the repair, and
tensile and shear stresses at the repair/old interface. These stresses are the cause of repair
cracking and interfacial debonding. In such a situation, a relatively larger creep strain
and lower elastic modulus are actually preferred for repair materials, as they aid in
relieving the stress buildup in the repair and at the repair/old interface.
It should also be noted that the interaction between the new repair and the existing
concrete is a time-dependent process. Repair material shrinkage, creep, elastic modulus,
tensile strength, and the repair/old interfacial bond are all age-dependent properties, and
develop at different rates once the new repair is installed.

Cracking and interfacial

debonding time and extent are determined by competition among these age-dependent
parameters. In the past years, there have been many efforts made on improving repair
material

properties

by

reducing

repair

23

material

shrinkage,

increasing

its

compressive/tensile strength, or enhancing its bond strength with concrete. These efforts,
although meaningful, have resulted in only marginal improvements.

For example,

increasing repair material strength may delay repair cracking, as long as the restrained
shrinkage induced tensile stress remains below the material tensile strength over time.
However, it poses an elevated risk of interfacial debonding because higher levels of shear
and tensile stress build-up at the new/repair interface, which could have otherwise been
relieved through cracking in the repair layer. On the other hand, accelerating interfacial
bond development may delay the debonding process, but makes the repair more prone to
cracking because debonding reduces the level of constraints and partially relieves the
tensile stress in the repair. Reducing repair material shrinkage seems to be the most
effective method.

However, considering the low deformation capacity (elastic strain

capacity = 0.01%) for most of the currently available concrete repair materials, most
efforts can only reduce the shrinkage to a small amount below 0.01%.

In reality,

shrinkage is not the only factor that contributes to the deformation of concrete repairs.
When thermal effects, structural loading, and other factors are combined, it is highly
possible that the total imposed deformation of the repair will surpass its deformation
capacity, and cracking and/or debonding will inevitably follow.
To ensure dimensional compatibility within the repaired system, the repair
material should be able to deform to an extent that fully accommodates the imposed
deformation due to a combination of many factors, without debonding or cracking locally.
These criteria should be satisfied at all ages throughout the repair's life. In this sense, a
repair material can be considered truly dimensionally compatible with existing concrete if
the following characteristics are satisfied at all ages: (a) high tensile ductility

24

(deformation capacity), (b) adequate bonding with concrete, (c) similar thermal
coefficient and creep as concrete, and (d) same or lower elastic modulus as concrete.

1.2.3

Some considerations of current repair performance characterization


The technical guideline published by the International Concrete Repair Institute

(ICRI)137 provides a standardized protocol for testing and reporting data for inorganic
concrete repair material.

In this guideline, test methods for characterizing repair

materials contain three main categories: (a) mechanical properties (i.e., compressive,
flexural, splitting tensile and direct tensile strength), (b) dimensional compatibility (i.e.,
modulus of elasticity, bond strength, length change, compressive creep, and coefficient of
thermal expansion), and (c) durability (i.e., freezing and thawing resistance, scaling
resistance, rapid chloride permeability, chloride ponding, sulfate resistance, chemical
resistance, and cracking resistance).

While these test methods provide standardized

characterization of repair material properties, and insights on material selection for the
objective of improving durability of concrete repairs, appropriate adoption of these test
methods and adequate interpretation of the test results require special consideration of a
number of issues detailed below.
The performance and durability of concrete repairs and new concrete structures
should not be evaluated in the same way. New structures are typically well-defined
structural systems that have been designed and put into service. Therefore, it is not
necessary to consider pre-existing cracks. However, in repair and rehabilitation, preexisting conditions such as cracking, delamination, and spalling are already present from
the beginning of the repair system's service life. As an example, concrete substrate

25

cracks can act as stress concentrators in the repair layer, resulting in reflective cracking
and shortening of service life. These pre-existing conditions should be accounted for
when laboratory testing is conducted to measure repair performance.
In addition, the repair may experience cracking shortly after placement due to
restrained shrinkage or other factors, which would affect repair durability. One example
of the inadequate measurement of repair performance is the permeability test, the results
from which are widely used as durability indicators of repair materials. This method,
conducted using uncracked laboratory specimens, fails to take the early-age cracking that
usually occurs in repairs due to restrained volume change into consideration. Moreover,
permeability is a measurement of the transport of water under a hydrostatic pressure head,
and is not the only parameter that describes the transport properties of repair materials.
For surface repairs of bridge decks or road pavements, for example, the diffusion of
chloride ions under an applied concentration gradient is the governing transport process
that impairs the durability of the repair, and should be measured using the saltwater
ponding test on uncracked as well as cracked specimens.
Another

example of tests that may result in inadequate

performance

characterization of repaired systems is the free shrinkage test. This test is generally used
to measure shrinkage strain of repair materials. Repair materials with lower shrinkage
strain are considered to have higher resistance to shrinkage cracking, and therefore are
believed to be more durable. However, the shrinkage cracking potential of a repair is
determined by a combination of different factors, including the repair material's
shrinkage strain, tensile strain capacity, creep, elastic modulus, and tensile strength at all
ages. A repair material with relatively low shrinkage strain, but low creep strain, low

26

tensile strain capacity, and high elastic modulus can be more susceptible to shrinkage
cracking. Therefore, a restrained shrinkage test such as ASTM CI581 "Standard Test
Method for Determining Age at Cracking and Induced Tensile Stress Characteristics of
Mortar and Concrete under Restrained Shrinkage"138 will be more meaningful than a free
shrinkage test for evaluating a repair material's resistance to shrinkage cracking.

1.3

Proposed Approach
The challenges posed by deteriorating concrete structures and widely-observed

repair failures require an effective and practical approach that breaks down the concrete
deterioration process, which typically begins with restrained volume change induced
cracking or reflective cracking, followed by penetration of aggressive agents, and
eventually results in concrete spalling, disintegration, and loss of structural capacity. This
dissertation research focuses on the development and implementation of an innovative
material technology based on an integrated multi-scale material/structural engineering
framework, as illustrated in Figure 1.10.

This framework begins with material

microstructure tailoring at the micrometer scale, links to repaired system durability


assessment through composite material properties at the centimeter scale, and ultimately
relates construction processes and performance evaluation at the meter scale of the
repaired infrastructure.

To achieve the holistic durability of such engineered systems,

durability concepts are interwoven across all material engineering, system durability
assessment, and structural application scales.
At the material engineering scale, innovative cementitious composite materials
are developed to have desired material properties through microstructure tailoring and

27

processing techniques.

The critical target material properties are large tensile strain

capacity and tight self-controlled crack width, which are essential to address the major
causes of concrete repair deterioration, e.g. lack of dimensional compatibility, cracking
due to restrained volume change, reflective cracking, and penetration of aggressive
agents. In this way, the typical deterioration process in concrete repairs can be disrupted.
Additionally, the materials proposed herein are designed with special functionalities
favorable for specific repair applications, without sacrificing durability.

These

functionalities include workability, high early-age strength, lightweightness, greenness,


and cost effectiveness.
At the repaired system durability assessment scale, repair material properties are
linked to repaired system durability through the interaction between the repair and the
existing concrete. Experimental measurements are designed to take into consideration
the interaction between new repair and old concrete as well as the presence of cracks.
The durability performance of the repaired systems is investigated under various
environmental and mechanical loading conditions.

This work bridges the gap of

translation from the properties of repair material properties to the durability of a repaired
system, which has not been fully understood through previous approaches.
The structural application stage focuses on bringing repair materials designed and
tested in the laboratory to full-scale field applications. Before the newly developed repair
materials can be considered a viable construction material for use in concrete repairs,
they must be capable of processing and placement with controlled quality at a larger scale
using commercial batching equipment common within the repair industry.

This is

achieved through further tailoring of the composite ingredients and processing procedure

28

for larger scale mixing, scaled up trial batches, and workability measurements and control.
The ability of full-scale processing and placement of a newly developed repair material is
demonstrated in a bridge patch repair project in Michigan. Monitoring the long-term
durability of the HES-ECC repair further validates the applicability of the repair
technology.
At the core of this integrated multi-scale material/structural

engineering

framework is Engineered Cementitious Composites (ECC), which is a unique class of


high-performance fiber-reinforced cementitious composites (HPFRCC).

The tailor-

ability of ECC at the microstructure level for prescribed material properties and structural
performance lends itself towards easy adoption with the integrated multi-scale
material/structural engineering philosophy.
In contrast to the presently proposed approach to durable repair development,
current practice can be illustrated by Figure 1.11, which shows the disconnect between
material engineering and structural engineering in current concrete repair practices.
Within this philosophy, material engineering is often focused on developing new repair
materials or modifying currently available repair materials to achieve desired properties,
such as higher compressive strength, a denser matrix, low permeability, and low
shrinkage. These "durable" materials, once available, are adopted in concrete structural
repair applications for "durable performance".

However, without bridging the gap

between "repair material properties" and "repaired structure performance" through


appropriate durability assessment that addresses the interaction between repair and
existing concrete, the transferability between the two concepts is questionable.
Furthermore, the disconnected philosophy hinders the ability of performance-based

29

material development through microstructure tailoring for target concrete repair


performance (e.g. cracking resistance, corrosion prevention, early re-opening, etc.). This
disconnect may be seen as a major reason why current concrete repair technology has not
been satisfactory.

This thesis attempts to address this technological gap from an

integrated material/structural system viewpoint.

1.4

Engineered Cementitious Composites


Engineered Cementitious Composites (ECC) is a unique class of high-

performance fiber-reinforced cementitious composites (HPFRCC) that features high


intrinsic tensile ductility and moderate fiber content139'140. ECC exhibits tensile strainhardening behavior through multiple microcracking with self-controlled crack width,
leading to fracture toughness similar to aluminum alloys \ Tensile strain capacity in the
3-5% range, which is around 300-500 times that of concrete and fiber reinforced concrete
(FRC), has been demonstrated in ECC materials using polyethylene fibers and polyvinyl
alcohol (PVA) fibers with fiber volume fraction no greater than 2%142'143. A typical
uniaxial tensile stress-strain curve for ECC is shown in Figure 1.12, along with crack
width development at different loading stages. The tensile strain-hardening behavior, i.e.
increasing load capacity with increasing straining without localizing into a fracture plane,
distinguishes ECC from FRC, which shows tension-softening after first cracking. ECC
materials also feature high ductility when subjected to large shear stress144 and impact
loading145'146, and exhibit high damage resistance to fracture failure under severe stress
concentrations induced by steel/concrete interactions147'148.

30

Due to its unique high tensile ductility as well as compatibility with ordinary
Portland cement concrete in other respects, such as compressive strength and thermal
properties, ECC possesses characteristics sought by engineers for improving performance
of civil infrastructure, such as buildings, bridges and pipes149. The high tensile ductility
and toughness of ECC material greatly elevates the mechanical performance of
reinforced ECC (R/ECC) structures by preventing brittle failure and loss of structural
integrity, which are common in traditional reinforced concrete (R/C) structures under
excessive loading.

It has been demonstrated experimentally that R/ECC structural

members, such as beams150, columns151, walls152, and connections153 surpass normal R/C
structural members in structural load carrying capacity, deformability, and energy
absorption capacity under monotonic and reverse cyclic loading.
An extensive body of literature exists on the durability of ECC materials. A brief
review of durability characteristics of ECC under abrasion and wear, freezing and
thawing, and hot and humid environments, is given below. ECC resistance to chloride
penetration will be a focus of the present thesis (Chapter 6). A review of other durability
related studies on ECC, including salt scaling resistance, permeability, and absorption,
can be found in Chapter 6 as well.
To evaluate ECC's abrasion and wear resistance for roadway applications, surface
friction and wear track testing according to Michigan Test Method 111 was conducted in
conjunction with MDOT 154 using a static friction tester (Figure 1.13). These tests were
conducted on a wet pavement surface with vehicle tires operated at 65 kph (40 mph).
Initial friction forces between the tire and the ECC surfaces were determined. The ECC
specimens were then subjected to 4 million tire passes to simulate long-term wear. After

31

wearing, friction forces were again determined to evaluate deterioration or surface


polishing during wearing. These final measured frictional forces are called the Aggregate
Wear Index (AWI). AWI values for the textured ECC samples range from 1.6kN to
2.3kN, which are higher than the required minimum AWI of 1.2kN for Michigan
trunkline road surfaces. The test results show that ECC has good abrasion and wear
resistance to heavy traffic.
Freeze and thaw testing was conducted on ECC and normal concrete specimens
based on ASTM C666A155 (Figure 1.14). In addition, a series of ECC tensile specimens
were also subjected to freeze-thaw exposure to evaluate the effect on ECC's tensile strain
capacity.

Testing of non-air-entrained ECC and concrete prisms was conducted

concurrently over 14 weeks. After 5 weeks (110 cycles), the concrete specimens had
severely deteriorated, requiring removal from the test.

However, ECC specimens

survived 300 cycles with no degradation of dynamic modulus. The computed durability
factor of ECC was 100, far larger than the value of 10 for the non-air-entrained concrete.
The uniaxial tension tests performed on freeze and thaw exposed ECC specimens after
200 cycles showed no significant drop in strain capacity (3%).
Hot water immersion tests, performed on single fibers embedded in ECC matrix,
and on composite ECC material specimens, were conducted to simulate hot and humid
environments156. After immersion in hot water at 60C for 26 weeks, little change was
seen in fiber properties. Interfacial properties, however, experienced significant changes,
particularly between 13 and 26 weeks of immersion, resulting in a drop of ultimate
composite strain capacity from 4.5% at early age to 2.8% after 26 weeks of hot water
immersion.

The residual strain capacity of ECC after exposure to this accelerated

32

weather environment (equivalent to more than 70 years of natural weathering) was still
over 250 times that of normal concrete.
ECC is currently emerging in full-scale structural applications, including the
composite deck of a cable-stayed bridge in Japan157 (Figure 1.15). This bridge employed
an estimated 800 cubic meters of ECC for the 38 mm (1.5 in.) thick ECC/Steel composite
deck. In the United States, the first full-scale ECC link slab has been constructed in
Michigan158'159, as shown in Figure 1.16.
The combination of high performance and moderate fiber content is achieved by
micromechanics-based

composites

optimization1 '

60

(also

see

Chapter

2).

Micromechanics provides guidance in selection and tailoring of the type, size, amount,
and proprieties of ingredients at micrometer and nanometer scales. The fibers, matrix,
and interface interact synergistically, thus resulting in a composite material with
controlled multiple microcracking when overloaded in tension, as well as other target
functionalities.

The ECC micromechanics-based design framework elevates the

cementitious materials design from conventional trial-and-error empirical combination of


individual constituents to systematic material "engineering".

Its ability to suppress

localized fracture modes through controlled microcracking, and its micromechanicsbased design philosophy and tailor-ability, differentiates ECC from other fiber reinforced
cementitious composites.

1.5

Research Motivation and Objectives


The huge flow of material driven by concrete global production and consumption

and the lack of durability of concrete infrastructure has created large economic, social

33

and environmental impacts, and poses great challenges for the civil engineering
profession. Presently it is very difficult to meet this challenge of durable development
with traditional concrete materials, which have inherently low resistance to cracking.
Also, the limitations of currently available repair materials and technologies are widely
evident in the common occurrence of early deterioration of repairs, which fail to address
the underlying concrete deterioration problems and prolong the service life of concrete
structures.

Moreover, the current disconnect between material engineering, system

durability assessment, and structural application often causes breakdowns within the
overall design.
This research looks to develop and demonstrate an integrated multi-scale
material/structural engineering approach for improving durability of concrete structures.
This will be accomplished through the establishment of links between the development,
assessment, and implementation of an innovative material technology that fundamentally
tackles major

concrete deterioration problems.

Specifically,

the

deterioration

mechanisms of concrete repairs under various environmental and mechanical loading


conditions should be understood first. Through identifying causes and effects of each
deterioration stage, the critical repair material properties that influence each stage need to
be accurately determined. These properties should be designed into special versions of
ECC materials, with other functionalities that are preferred for specific repair applications,
through constituent ingredients screening, micromechanical tailoring of matrix and
fiber/matrix interfacial properties, and laboratory testing. Repair systems containing the
developed repair material should be properly evaluated under various environmental
exposure and mechanical loading conditions. The influence of material properties on the

34

interaction between the repair and the existing concrete as well as the durability of the
repair system should be fully addressed. Finally, the performance and practicality of this
material technology should be justified not only in laboratory studies, but also in fullscale field applications.

1.6

Research Scope and Dissertation Organization


Within this dissertation, the development, evaluation, and implementation of an

innovative ductile repair technology on all three levels of the integrated multi-scale
material/structural

engineering

framework

will be

addressed,

namely

material

engineering, repair system durability assessment, and field application (Figure 1.10).
Within each of these levels, the tailoring of ECC materials is highly focused on meeting
specific structural performance requirements while adhering to additional system
constraints (i.e. cost, processing, material availability, existing cracks, etc.). A variety of
micromechanics based tools, along with carefully designed experiments and finite
element methods, are employed to understand and optimize ECC materials development
for specific repair applications. In addition to laboratory development and experimental
studies, the newly-developed

ductile repair technology is transferred

to

field

implementation by optimizing workability and processing techniques, accommodating


traditional large-scale mixing equipment, and demonstrating in a bridge repair project in
Southern Michigan.
In Chapter 2 the ECC micromechanical design theory is introduced, and the
development of special versions of ECC materials for concrete repair applications will be
presented. The development of high-early-strength ECC (HES-ECC) is first discussed.

35

This material is designed for fast and durable repair of transportation infrastructure (e.g.
bridges, roadways and highways), or other types of concrete structures, which prefer a
minimal interruption of operations and are often exposed to challenging environmental
and mechanical loading conditions. The material design begins with matrix screening for
strength and workability requirements, and the initially selected mixing proportion is
further tailored to achieve the target tensile ductility.
Chapter 3 focuses on maximizing ECC material properties (e.g. tensile strain
capacity, tensile strength) and minimizing material variation through processing
technique. ECC laboratory testing and application in construction projects will rely on
the capacities to consistently produce ECC materials with controlled quality. To address
this need, this work develops a simple and practical quality control method for ECC
processing. The influence of fresh material processing on the microstructure of ECC
composites is first discussed.

Through measurement of ECC mortar rheological

parameters, fiber dispersion, and ECC tensile properties, the correlation between the three
is established.

The optimal range of marsh cone flow rate, an indicator of plastic

viscosity of ECC mortar, was then indentified and used to guide ECC quality control. By
controlling the plastic viscosity of ECC mortar within this optimal range, the ECC
material can achieve optimized material properties as originally designed through the
micromechanical methodology described in Chapter 2. Collectively, Chapters 2 and 3
comprise the "material engineering" portion of the integrated material/structural
engineering framework shown in the bottom triangle in Figure 1.10.
Chapters 4, 5, 6 and 7 shift the focus from "material engineering" to "repair
system durability assessment" (the middle triangle in Figure 1.10). Chapter 4 addresses

36

one of the major durability concerns in concrete repairs, i.e. repair cracking and repair/old
interfacial delamination due to restrained volume change and lack of dimensional
compatibility. Experimental and numerical studies are conducted on a simulated repair
system, which consists of a layer of old concrete substrate and a layer of new repair
material, under restrained shrinkage conditions. The influence of repair material ductility
on the durability of the repaired system is identified. The effect of substrate surface
preparation on the repair performance is also investigated. Experimental and numerical
results verify that when adequate interfacial bonding is present, the high tensile ductility
of HES-ECC can relieve shrinkage induced stresses in the repair layer and at the
repair/old interface, thereby simultaneously suppressing large repair surface cracks and
interface delamination.

This research departs from the traditional emphasis on high

compressive strength repair materials, and moves toward balancing strength, ductility and
the repair material's compatibility with the old surrounding concrete. This concept of
translating the ductility of the repair material to the durability of the repair system, when
subjected to restrained volume change, can be widely applied to concrete structure repair
applications for minimizing maintenance requirements and reducing repair costs.
Chapter 5 addresses the reflective cracking issues prevalent in concrete repairs,
especially overlay repairs. As discussed in Section 1.1, reflective cracking and restrained
volume change induced cracking are two major cracking mechanisms affecting concrete
repair service life.

Experimental studies are conducted on an overlay repair system,

which consists of a layer of overlay repair cast on a concrete substrate with an existing
crack and small amount of interfacial debonding. This overlay repair system is subjected
to monotonic flexural loading to measure its load carrying capacity, and fatigue flexural

37

loading to measure its S-N curve. The influence of repair material ductility on the load
carrying capacity and service life of the repaired system is investigated. Findings verify
that the tensile ductility of ECC effectively diffuses the stress concentration in the zone
above the existing crack, therefore preventing reflective cracking.

Through this

mechanism, the ECC repaired system exhibits 100% increased load carrying capacity and
significantly prolonged fatigue life, compared to the control concrete repaired system.
This methodology of suppressing reflective cracking through material ductility can be
generalized as translating the ductility of the repair material to the durability of the repair
system, when it is subjected to stress concentration from existing cracks.
Chapter 6 addresses the chloride penetration stage in a typical repair deterioration
process, before and after cracking occurs at different straining levels. Chloride diffusion
driven by a chloride concentration gradient, which is the predominant mechanism of
chloride transport for most of the concrete structures exposed to deicing salts and
airborne chloride ions, is considered in this study. Experiments are conducted under
combined mechanical (flexural) and environmental (3% chloride solution ponding)
conditions, to simulate actual conditions experienced by a structural member when it is
subjected to both structural loading and chloride exposure. Influences of material tensile
strain capacity, crack width, and applied deformation level on the effective chloride
diffusion coefficient are investigated.

The results conclude that ECC is effective at

slowing down the diffusion process of chloride ions under combined mechanical and
environmental loading, by virtue of its ability to achieve a self-controlled tight crack
width, even under large applied straining levels. This study verifies the concept of

38

improving repaired system durability with the built-in crack self-controlling mechanism
of ECC repair materials.
Chapter 7 investigates the influence of cracking and self-healing on the durability
of ECC under aggressive chloride exposure combined with mechanical loading
conditions. The results show that after various levels of pre-applied deformation and
time-lengths of chloride exposure, the reloaded ECC still remains its large tensile
ductility accompanied by multiple microcracking behavior.

Recovery of ECC

mechanical properties in term of stiffness, tensile strain capacity, and tensile strength was
observed, indicating strong evidences of self-healing of the microcracked ECC material
under chloride exposure.

This study confirms that ECC, whether uncracked or

microcracked, remains durable under severe marine environment.


Chapter 8 shifts the focus from "repair system durability assessment" to the
"structural application" portion of the integrated engineering methodology (upper triangle
in Figure 1.10). Efforts are made to transfer innovative ECC repair technology from the
laboratory to field implementation through a bridge patch repair demonstration project.
Large-scale processing and construction of an HES-ECC repair using commercial
facilities are realized through optimization of the HES-ECC ingredients and mixing
procedure, trial batches, and quality control methods. The long-term durability of the
HES-ECC patch repair under field conditions is monitored.

Through this work, the

linkages between material engineering, repair system durability assessment, and


structural application are further forged.
Overall conclusions from this study are summarized in Chapter 9, and some
future works worthy of further investigation are outlined.

39

-. n

:v.> *

t--{*

t- ::- , ft. -n
.-j4

' *

Deterioration due to steel corrosion,


concrete cracking and spalling

Damage caused by earthquake

Defect associated with improper detailing


and construction practice
Figure 1 . 1 - Concrete Structure Deterioration, Damage, and Defect.

40

(b)
(a)
Fig. 1.2 - Deterioration of Kurtsubo Bridge, Japan: (a) Rust signs before the first repair;
(b) Tendon corrosion and breakage before the second repair.

Removing old one

w
a
o
~m

New construction

o
o
a

Rust signs

cr

Monitoring
^2nd Repair + rehab.

1 st Repair

Load regulation

0
0

20
40
Service period (years)

Figure 1.3 - Life cycle cost of Kurtsubo Bridge, Japan.

41

80

(b)

(a)

Figure 1.4 - Bridge deck transverse cracking, (a) Transverse cracking of a normal
strength concrete deck in Ontario. Cracks 0.5 mm wide spaced at 200 mm. (b) Early-age
transverse cracking of a high strength concrete deck in Colorado. Cracks 0.4 mm wide
while construction is being completed (concrete < 2 months old).

Figure 1.5 - Steel corrosion and spalling in concrete bridges in Michigan (Courtesy of
MDOT).

42


Overlay

Existing Crack
or Joint In Qlcter

, Pavement
i _^. ,-,^~,

Existing Craetad
Pavement

.,_ Jfc!. _ _,

Figure 1.6 - Pavement overlay reflective cracking.

(b)
Figure 1.7 - Deterioration in parking structures, (a) Cracking in concrete slab, (b)
Corrosion and spalling of the pans and ribs on a parking deck soffit.

43

(a)

i_

crack width: w

(b)

Figure 1.8 - (a) Model of concrete slab restrained against shrinkage and (b) the crack
width w controlled by the tension-softening curve.

w=

LE

/ L=21

Figure 1.9- Schematic plot of crack width w as a function of shrinkage strain es. The
crack width development is shown for three materials with varying toughness.

44

Structural
Performance

Environmental Loads
& Structural Loads

Repair System
Durability

Construction
Quality

Environmental Loads
& Structural Loads

Material
Properties

Repair/Old
Interaction

Environmental; Loads
& Structural Loads:

Material iV;
Microstructure

Processing

Figure 1.10 - Integrated Multi-Scale Material & Structural Engineering Framework for
durability of repaired concrete structures.

45

Structural
Performance

Material
Properties

"

"~"

""

Construction
Quality

Material
Properties

Material
Microstructure

'

- - . - . -

Processing

Figure 1.11- Disconnected Material Engineering and Structural Engineering Design


Philosophies for durability of repaired concrete structures.

46

100

5 T

ECC
80

+ 60

40-^

20

Concrete Strain 10 times expanded

Strain (%

Figure 1.12 - Typical tensile stress-strain curve of ECC. Above 3 MPa, ECC shows a
distinct strain-hardening response up to about 5% strain. In contrast, normal concrete
fails at 0.01% strain.

47

fe
W<M
JP"1

Figure 1.13 - Driving surface static friction tester.

Testing in accordance with ASTM C-666A


1.20.0 j
f

i oo.o k
\

80.0
60.0
40,0
20.0

HES-ECC

: Regular Con crew

o.o '
0

SO

100

150

200

250

300

Number of Freeze Thaw Cycles

Figure 1.14 - Relative dynamic modulus vs. number of freeze thaw cycles.
48

II

I
Figure 1.15 - Mihara Bridge with ECC/steel composite deck opened to traffic in April
2005,Japan.

*%i

*-*"fc

Figure 1.16 - ECC link slab on the Grove Street Bridge over 1-94 opened to traffic in
October 2005, Michigan.

49

References:
1

Infrastructure, Online Compact Oxford English Dictionary,


http://www.askoxford.com/concise_oed/infrastrucrure, accessed on June 24, 2009.
2

Lomborg, B., The Skeptical Environmentalist: Measuring the Real State of the World,
2001, pp. 138.
3

"Minerals Commodity Summary - Cement - 2007",


http://minerals.usgs.gov/minerals/pubs/commodity/cement/index.html, accessed on June
24, 2009.
4

Emmons, P. H., and Sordyl, D. J.,"The State of the Concrete Repair Industry, and a
Vision for its Future," Concrete Repair Bulletin, July/August 2006.
5

Frangopol, D. M., and Furuta, H., Proceedings of the US-Japan Workshop on LifeCycle Cost Analysis and Design of Civil Infrastructure Systems, Honolulu, Hawaii, 2000.

Narayan, W., "Infrastructure, Climate Change, Sustainability: The Challenge of Design


- Strength or Durability," Concrete, Vol. 41, No. 9, October 2007, pp. 31-33.
7

Mays, G., Durability of Concrete Structures: Investigation, Repair, Protection, 1992.

"Identifying and Evaluating Concrete Defects",


http://www.cement.org/tech/cct_con_design_defects.asp, accessed July 10, 2009.
9

Vision 2020: A Vision for the Concrete Repair Protection and Strengthening Industry,
the Strategic Development Council (SDC). 2006.
http://www.concretesdc.org/_pdfs/Vision2020-Versionl.0_%20May2006.pdf
10

Report Card for America's Infrastructure. 2009.


http://www.infrastructurereportcard.org/
11

Li, V. C , "Engineered Cementitious Composites", Proceedings of ConMat'05,


Vancouver, Canada, August 22-24, 2005, CD-documents/l-05/SS-GF-01_FP.pdf.
12

Vaysburd, A. M., Brown, C. D., Bissonnette, B, and Emmons, P. H., ""Realcrete"


versus "Labcrete"", Concrete International, Vol. 26 No.2, pp 90-94, 2004.
13

Tanaka, Y., Kawano, H., Watanabe, H and Nakajo, T., "Study on required cover depth
of concrete highway bridges in coastal environment", 17th U.S. - Japan Bridge
Engineering Workshop, Tsukuba, 2001.
14

Federal Highway Administration, "Pavement Preservation


Definitions,"
http://www.fhwa.dot.gov/pavement/preservation/091205.cfm, accessed on August 13,
2009.

50

The American Association of State Highway and Transportation Officials (AASHTO),


"America's Top Five Transportation Headaches - and Their Remedies," January 2009,
www.tripnet.org/Transportation_Headaches_Report_Jan_2009.pdf.
16

Report Card for America's Infrastructure, http://www.infrastructurereportcard.org/factsheet/bridges


17

Keoleian, G.A., Kendall, A., Dettling, J.E., Smith, V.M., Chandler, R.F., Lepech, M.D.
and Li, V.C., "Life Cycle Modeling of Concrete Bridge Design: Comparison of ECC
Link Slabs and Conventional Steel Expansion Joints," Journal of Infrastructure Systems,
ASCE, March 2005, pp.51-60.
18

American Association of State Highway and Transportation Officials (AASHTO).


Bridging the Gap. July 2008. http://www.transportationl.org/BridgeReport/frontpage.html
19

Saadeghvaziri, M. A., and Hadidi, R., "Transverse Cracking of Concrete Bridge Decks:
Effects of Design Factors," ASCE Journal of Bridge Engineering, Vol. 10, No. 5,
September/October 2005, pp. 511-519.
20

Hadidi, R., and Saadeghvaziri, M. A., "Transverse Cracking of Concrete Bridge Decks:
State-of-the-Art," ASCE Journal of Bridge Engineering, Vol. 10, No. 5,
September/October 2005, pp. 503-510.

21

Cady, P. D, and Carrier, R. E., "Final Report on Durability of Bridge Deck Concrete:
Part 1: Effect of Construction Practice on Durability," PennDOT Contract No. 31057-H,
Dept. of Civil Engineering, Pennsylvania State Univ., University Park, PA, 1971.
22

Krauss, P. D., and Rogalla, E. A., "Transverse Cracking in Newly Constructed Bridge
Decks," NCHRP Rep. No. 380, Transportation Research Board, National Research
Council, Washington, D. C , 1996.
23

Portland Cement Association (PCA), "Durability of Concrete Bridge Decks - A


Cooperative Study "Final Rep., Ill, 1970.

24

Kosel, H. C , and Michols, K. A., "Evaluation of Concrete Deck Cracking for Selected
Bridge Deck Structures of Ohio Turnpike," Rep., Ohio Turnpike Commission,
Construction Technology Laboratory, Ohio Department of Transportation, Columbus,
Ohio, 1985.
25

Iowa Department of Transportation (Iowa DOT), "A Study of Transverse Cracks in the
Keokuk Bridge Deck," Final Rep., Ames, Iowa, 1986.
26

McKeel, W. T.,"Evaluation of Deck Durability on Continuous Beam Highway Bridges."


Rep. No. VHTRC 85-R32, Virginia Highway and Transportation Research Council,
Charlottesville, VA, 1985.

51

27

Ramey, G. E., Wolff, A. R., and Wright, R. L., "Structural Design Actions to Mitigate
Bridge Deck Cracking," Practice Periodical on Structural Design and Construction, Vol.
2, No. 3, pp. 118-124, 1997.
28

Cheng, T. T., and Johnson, D. W., "Incidence Assessment of Transverse Cracking in


Bridge Decks: Construction and Material Consideration," Rep. No. FHWA/NC/85-002,
Federal Highway Administration, Washington, D. C , Vol. 1, 1985.
29

Perfetti, G. R., Johnson, D. W., and Bingham, W. L., "Incidence Assessment of


Transverse Cracking in Concrete Bridge Decks: Structural Considerations," Rep. No.
FHWA/NC/85-002, Vol. 2, Federal Highway Administration, Washington, D. C , 1985.
30

Perragaux, G. R., and Brewester, D. R., "In-Service Performance of Epoxy Coated


Steel Reinforcement in Bridge Decks," Final Technical Rep. No. 92-3, New York State
Department of Transportation, Albany, N. Y., 1992.
31

Babaei, K., and Hawkins, N. M.,"Evaluation of Bridge Deck Protective Strategies,"


NCHRP Rep. No. 297, Transportation Research Board, National Research Council,
Washington, D. C , 1987.
32

La Fraugh, R. W., and Perenchio, W. F., "Phase I Report of Bridge Deck Cracking
Study West Seattle Bridge," Rep. No. 890716, Wiss, Janney, Elstner Associates,
Northbrook, 111, 1989.
33

Babaei, K. and Purvis, R., "Prevention of Cracks in Concrete Bridge Decks: Report on
Laboratory Investigation of Concrete Shrinkage," Research Project No. 89-01,
Pennsylvania Department of Transportation, Harrisburg, PA, 1994.
34

Ducrete, J., Lebet, J. , and Monney, C , "Hydration Effect and Deck Cracking During
the Construction of Steel Concrete Composite Bridges," Proc. ICOM-Construction
Mettalique, Article 359, 1997.
35

French, C , Eppers, L., Le, Q., and Hajjar, J. F., "Transverse Cracking in Concrete
Bridge Decks," Transportation Research Record 1688, Transportation Research Board,
Washington, D. C , 1999, pp. 21-29.
36

Frosch, R. J., Radabaugh, R. D., and Blackman, D. T., "Investigation of Transverse


Deck Cracking," Proc. Structures Congress, ASCE, Reston, VA, 2002.
37

Schmitt, T. R., and Darwin, D., "Cracking in Concrete Bridge Decks," Rep. No. KTRAN.KU-94-I, Kansas Department of Transportation, Topeka, KAN, 1995.

38

Horn, M. W., Stewart, C. F., and Boulware, R. L., "Factors Affecting the Durability of
Concrete Bridge Decks: Construction Practice," Interim Rep. No. 3, CA-DOT-ST-4101-475-3, Bridge Department, California Division of Highways, Sacramento, California, 1975.

52

Kochanski, T., Parry, J., Pruess, D., Schuchardt, L., and Ziehr, J., "Premature Cracking
of Concrete Bridge Decks Study," Final Rep., Wisconsin Department of Transportation,
Madison, Wisconsin, 1990.
40

Ramey, G. E., Wolff, A. R., and Wright, R. L., "Structural Design Actions to Mitigate
Bridge Deck Cracking," Practice Periodical on Structural Design and Construction, Vol.
2, No. 3, 1997, pp. 118-124.
41

Meyers, C , "Survey of Cracking on Underside of Classes B-l and B-2 Concrete


Bridge Decks in District 4," Investigation No. 82-2, Division of Material and Research,
Missouri Highway and Transportation Department, Jefferson City, Mo, 1982.
42

Stewart, C. F., and Gunderson, B. J., "Factors Affecting the Durability of Concrete
Bridge Decks," Interim Rep. No. 2, Research and Development Section of Bridge
Department, California Department of Transportation, Sacramento, California, 1969.
43

American Concrete Institute (ACI), "Hot Weather Concreting," ACI 305 R, Committee
305, Detroit, 1999.

44

New York State Department of Transportation (NYDOT), "The State of the Art Bridge
Deck," Final Rep, The Bridge Deck Task Force, NYDOT, Albany, N. Y., 1995.
45

Issa, M., "Investigation of Cracking in Concrete Bridge Decks at Early Ages," Journal
of Bridge Engineering, Vol. 2, No. 2, 1999, pp. 116-124.

46

Hadidi, R., Saadeghvaziri, M. A., and Hsu, C. T., "Practica Tool to Accurately
Estimate Tensile Stresses in Concrete Bridge Decks to Control Transverse Cracking,"
Practice Periodical on Structural Design and Construction, Vol. 8, No. 2, 2003, pp. 7482.
47

Houde, J., "Study of Force-Displacement Relationship for the Finite Element Analysis
of Reinforced Concrete," Rep. No. 72-3, Department of Civil Engineering and Applied
Mechanics, McGill University, Montreal, 1973.
48

Saadeghvaziri, M. A., and Hadidi, R., "Cause and Control of Transverse Cracking in
Concrete Bridge Decks," Final Rep., FHWA-NJ-2002-19, 2002.

49

Yam, L. C. P., and Chapman, J. C , "The Inelastic Behavior of Simply Supported


Composite Beam of Steel and Concrete," Proc.-Inst. Civ. Eng., Vol. 41, 1969, pp. 651683.

50

Gilani, A. and Jansson, P., "Link Slabs for Simply Supported Bridges." MDOTReport
Number MDOT SPR-54181, Structural Research Unit, Construction and Technology
Support Area, Michigan Department of Transportation. Lansing, Michigan, 2004.
Fu, G., Feng, J., Dimaria, J., and Zhuang, Y., "Bridge Deck Corner Cracking on
Skewed Structures," MDOT Report RC-1490, Sept. 2007.
53

CTC & Associates LLC, WisDOT Research & Library Unit, "Early Concrete Cracking
on Bridge Decks and Overlays," prepared for WHRP Structures Technical Oversight
Committee, January 30, 2008.
http://on.dot.wi.gov/wisdotresearch/database/tls/tlsdeckcracking.pdf.
53

Li, ML, and Li, V. C , "Behavior of ECC/Concrete Layered Repair System under
Drying Shrinkage Conditions", Journal of Restoration of Buildings and Monuments, Vol.
12, No. 2,2006,ppl43-160.
54

Li, V.C., "High Performance Fiber Reinforced Cementitious Composites as Durable


Material for Concrete Structure Repair," Proc. of ICFRC Int'l Conference on Fiber
Composites, High Performance Concretes, and Smart Materials, Ed. By V.S.
Parameswaran, Pub. Allied Publishers Private Limited, New Delhi, India, 2004, pp. 5774.
55

Beushausen, H. D., and Alexander, M. G., "Performance of Concrete Patch Repair


Systems," Advances in Construction Materials, 2007, No. 255-262.
56

Simpson, J. M., and Burdetted, E. G., "Predicting Thermal Cracking in Bridge Deck
Repairs," Publication #C02C067,Copyright 2002 Hanley-Wood,LLC,
ftp://imgs.ebuild.com/woc/C02C067.pdf.
57

Czarnecki, L., Garbacz, A., Lukowski, P., and Clifton, J. R., "Polymer Composites for
Repairing Portland Cement Concrete: Compatibility Project," Technical Report NISTIR
6394, National Institute of Standards and Technology, 1999.
58

Emmons, P. H., Vaysburd, A. M., Poston, R. W., and McDonald, J. E., "Performance
Criteria for Concrete Repair Materials, Phase II, Field Studies," Technical Report REMRCS-60, U. S. Army Waterways Experiment Station, Vicksburg, MS, September 1998.
59

Pigeon, M. and Bissonnette, B., "Bonded Concrete Repairs: Tensile Creep and
Cracking Potential," Concrete International, Vol. 21, No. 11, November 1999, pp. 31-35.
60

Vayburd, A. M., "Research Needs for Establishing Material Properties to Minimize


Cracking in Concrete Repairs, Summary of a Workshop," ICRI Publication No. Y320001,
1996.
61

Vaysburd, A. M., Emmons, P. H., McDonald, J. E., Poston, R. W., and Kesner, K. E.,
"Performance Criteria for Concrete Repair Materials, Phase II Summary Report,"
Technical Report REMR-CS-62, U. S. Army Engineers Waterways Experiment Station,
Vicksburg, MS, March 1999.
62

Batis, G., Pantazopoulou, P., and Routoulas, A., "Corrosion Protection Investigation of
Reinforcement by Inorganic Coating in the Presence of Alkanolamine-Based Inhibitor,"
Cement and Concrete Composites, Vol. 25, 2003, pp. 371-377.

54

Elsener, B., "Macrocell Corrosion of Steel in Concrete - Implications for Corrosion


Monitoring," Cement & Concrete Composites, Vol. 24, 2002, pp. 65-72.
64

Scheissl, P., "Corrosion of Steel in Concrete," RILEM Report.


London. 1988.

Chapman and Hall.

65

Bavarian, B., and Reiner, L., "Migrating Corrosion Inhibitors for Steel Rebar in
Concrete," Materials Performance, February 2003, pp. 3-5.
66

Virmani, Y. P., and Clemena, G. G., "Corrosion Protection - Concrete Bridges,"


FHWA-RD-98-088, US Department of Transporation Federal Highway Administration,
Research and Development Turner-Fairbank Highway Research Center, VA, September
1998.
67

The State of the Nation's Highway Bridges: Highway Bridge Replacement and
Rehabilitation Program and National Bridge Inventory," Thirteenth Report to the United
States Congress, Federal Highway Admistration, Washington, D. C , May 1997.
68

Hime, W. G., "The Corrosion of Steel - Random Thoughts and Wishful Thinking,"
Concrete International, Vol. 15, No. 10, 1993, pp. 54-57.

69

Lewis, D. A., "Some Aspects of the Corrosion of Steel in Concrete," Proceeding of the
First International Congress on Metallic Corrosion, London, 1962, pp. 547-555.
70

Berman, H. A., "The Effects of Sodium Chloride on the Corrosion of Concrete


Reinforcing Steel and on the PH of Calcium Hydroxide Solution," Report No. FHWARD-74-1, Federal Highway Administration, Washington, D. C , 1974.
71

Santagata, M. C , and Collepardi, M., "The Effect of CMA Deicers on Concrete


Properties," Cement and Concrete Research, Vol. 30, No. 9, September 3000, pp. 13891394.
72

Dunn, S. A., and Schenk, R., "Alternate Highway Deicing Chemicals," Publication No.
FHWA-RD-79-108, Federal Highway Administration, Washington, D. C , October 1979.
73

Novokshchenov, V., "Salt Penetration and Corrosion in Prestressed Concrete


Members," Report No. FHWA-RD-88-269, Federal Highway Administration,
Washington, D. C , July 1989.
74

Whiting, D., Stejskal, B., and Nagi, M., "Condition of Prestressed Concrete Bridge
Components: Technology Review and Field Surveys," Report No. FHWA-RD-93-037,
Federal Highway Admiistration, Washington, D. C , 1993.
75

Lankard, D. R., Thompson, N., Sprinkel, M., and Virmani, Y. P., "Grouts for PostTensioned Concrete Construction: Protecting Prestressed Steel in Concrete," ACI
Materials Journal, September/October, 1993, pp. 406-413.

55

Ghorbanpoor, A., "Evaluation of Post-Tensioned Concrete Bridge Structures by the


Impact-Echo Technique," Report No. FHWA-RD-90-096,
Federal Highway
Administration, Washington, D. C , 1993.
77

Wagner, J., Young, W., Scheirer, S., and Fairer, P., "Cathodic Protection Development
for Prestressed Concrete Components - Interim Report," Report No. FHWA-RD-92-056,
Federal Highway Administration, Washington, D. C , March, 1993.
78

Funahashi, M., Wagner, X., and Young, W. T., "Cathodic Protection Developments for
Prestressed Concrete Components," Report No FHWA-RD-94-001, Federal Highway
Administration, Washington, D. C , July, 1994.
79

Virmani, Y. P., Clear, K. C , and Pasko, T. J., "Time-to-Corrosion of Reinforcing Steel


in Concretes, Vol. 5: Calcium Nitrite Admixture or Epoxy-Coated Reinforcing Bars as
Corrosion Protection Systems," Report No. FHWA/RD-83/012, Federal Highway
Administration, Washington, D. C , 1983.
80

Hay, R. E., and Virmani, Y. P., "North American Experience in Concrete Bridge
Deterioration and Mechanism," Proceeding of the Concrete Society, United Kingdom,
September, 1985.
81

Manning, D. G., National Research Council (U. S.), Transportation Research Board,
National Cooperative Highway Research Program, American Association of State
Highway and Transportation Officials (U. S.), Federal Highway Administration,
"Waterproofing Membranes for Concrete Bridge Decks," No. 220, 1995.
82

Stratfull, R. F., "Experimental Cathodic Protection of a Bridge Deck," Transportation


Research Record No. 500, 1974, pp. 1-15.
83

Fromm, H. J., "Electrically Conductive Asphalt Mixes for the Cathodic Protection of
Concrete Bridge Decks," Presented at the 1976 Meeting of the Association of Asphalt
Paving Technologiests, New Orleans, February, 1976.
84

Fontana, J. J., Reams, W., and Elling, D., "Conductive Overlay in Conjunction with an
Active Cathodic Protection System," Report No. FHWA-RD-88-145, Federal Highway
Administration, Washington, D. C , 1989.
85

Clemena, G. G., and Jackson, D. R., "Performance of a Conductive-Paint Anode in


Cathodic Protection Systems for Inland Concrete Bridge Piers in Virgina," Report No.
FHWA/VTRC 98-R7, Virginia Transportation Research Council, Charlottesville, Virgina,
1997.
86

Whiting, D., Nagi, M., and Broomfield, J. P., "Evaluation of Sacrificial Anode for
Cathodic Protection of Reinforced Concrete Bridge Decks," Report No. FHWA-RD-95041, Federal Highway Administration, Washington, D. C , May, 1995.

56

Hartt, W., Joubert, E., and Kliszowski, S., "Long-Term Effects of Cathodic Protection
on Prestressed Concrete Bridge Components," Report No. FHWA-RD-96-029, Federal
Highway Administration, Washington, D. C , November, 1996.
88

Bennett, J. E., and Shue, T. J., "Field Evaluation of Cathodic Protection on Prestressed
Concrete Bridge Members: Final Report," Report No. FHWA-RD-97-153, Federal
Highway Administration, Washington, D. C , 1997.
89

Clemena, G. G., and Jackson, D. R., "Pilot Applications of Electrochemical Chloride


Extraction on Concrete Piers in Virginia - Interim Reports," Report No. VTRC-96-IR4,
Virginia Transportation Research Council, Charlottesville, Virginia, 1996.
90

Mehta, P.K., "Durability-Critical Issues for the Future", Concrete International, Vol.
19, No. 6, 1997, pp. 27-33.

91

Xi, Y., B. Shing, N. Abu-Hejleh, A. Asiz, A. Suwito, Z. Xie, A, Ababneh, "Assessment


of the Cracking Problem in Newly Constructed Bridge Decks in Colorado," Colorado
Department of Transportation Research Branch. Denver, Colorado. CDOT-DTD-R-20033. March, 2003.
92

Kondratova, I., and Bremner, T. W., "Field and Laboratory Performance of EpoxyCoated Reinforcement in Cracked and Uncracked Concrete," Presented at the 77the
Annual Meeting of the Transportation Research Board, Washington, D. C , 1998.
93

The Road Information Project (TRIP), Key Facts About America's Road and Bridge
Conditions and Federal Funding, updated August 2008
94

U.S. Department of Transportation, Status of the Nation's Highways, Bridges and


Transit: Conditions and Performance, 2006.
95

Report of the National Surface Transportation Policy and Revenue Study Commission
- Transportation for Tomorrow, December 2007. Volume II.

96

Al-Qadi, I. L., Scarpas, T., and Loizos, A., Pavement Cracking, 2008, p. 930.

97

Delatte, N. J., Concrete Pavement Design, Construction, and Performance, 2007, p.


372.

98

Huang, Y. H., Pavement Analysis and Design (2" Edition), 2003, p. 792.

99

Miller, J. S., and Bellinger, W. Y., Distress Identification Manual for the Long-Term
Pavement Performance Program (Fourth Revised Edition), FHWA-RD-03-031, June 2003.
100

Schwartz, D. R., and Brown, B. C , "D-Cracking of Concrete Pavements," National


Transportation Research Board, 1987, p. 34.
101

Carpenter, A. J., and Cramer, S. M., "Mitigation of Alkali-Silica Reaction in


Pavement Patch Concrete that Incorporates Highly Reactive Fine Aggregate,"
57

Transportation Research Record, Annual Meeting of the Transportation Research Board,


No. 78, 1999, pp. 60-67.
102

U.S. Department of Transportation, Federal Highway Administration, Partial-Depth


Repairs, http://www.fhwa.dot.gov/pavement/concrete/index.cfm.
103

U.S. Department of Transportation, Federal Highway Administration, Full-Depth


Repairs, http://www.fhwa.dot.gov/pavement/concrete/full.cfm.
104

MDOT, Pavement Design and Selection Manual, Michigan Department of


Transportation, Lansing, Michigan, 2005, p. 65.
105

Tayabji, S. D. and Okamoto, P. A., "Thickness Design of Concrete Resurfacing,"


Proc, 3rd Int'l Conf. on Concrete Pavement Design and Rehabilitation, 1985, pp. 367379.
106

ERES Consultants, Inc., NCHRP Report 415, Evaluation of Unbonded Portland


Cement Concrete Overlays, Transportation Research Board, Washington, D. C , 1999.
107

Huang, Y. H., Pavement Analysis and Design, 2nd edition, Pearson Education, Upper
Saddle River, NJ 07458.
108

Keenan, L. E., and Soden, B. R., "Parking Structure Maintenance: Early Detection,
Early Cure," Journal of Architectural Technology published by Hoffmann Architects,
Vol. 19, No. 2, Feb. 2001.
109

Peterson, C. A., "Designing and Building Durable Parking Structures," Aberdeen's


Concrete Construction, Vol. 40, No. 3, March 1999.
110

Bhuyan, S. (Walker Parking Consultants, Inc.), Sabnis, G. M., Shiu, K. N., "A
Systematic Approach to Extending Service Life of Parking Structures," Indian Concrete
Journal, Vol. 75, No. 1, January 2001, pp. 58-64.
111

Bhuyan, S. (Walker Parking Consultants & Restoration Engineers, Kalamazoo, MI,


USA), "Repairing Concrete Parking Structure," Concrete Construction - World of
Concrete, Vol. 33, No. 2, February 1998.
112

O'Connor, J. P., and Olson, C. A., "Deterioration in Precast Prestressed Concrete


Parking Garage," Concrete International, Vol. 12, No. 11, Nov 1990, pp. 52-54.
113

Bickley, J. A., "Deterioration of Parking Structures," Annals of Biomedical


Engineering, 1980.
114

Ma, R., Xiao, Y., "Full-Scale Testing of a Parking Structure Column Retroffited with
Carbon Fiber Reinforced Composites," Construction and Building Materials, Vol. 14, No.
2, pp 63-71, March 30, 2000.

58

115

Aalami, B. O., Swanson, D. T., "Innovative Rehabilitation of a Parking Structure,"


Concrete International, Vol. 10, No. 2, February 1988, pp. 30-35.
116

Brainerd, M. L., "Evaluation and Reabilitation of Deteriorated Parking Structures,"


Construction Specifier, Vol. 42
117

Litvan, G. G., "Waterproofing of Parking Garage Structures with Sealers and


Membranes: The Canadian Experience," Construction and Building Materials, Vol. 10,
No. 1 Spec. Iss., February 1996, pp. 95-100.
118

Ojha, S. K., "Rehabilitation of a Parking Garage," Concrete International, Vol. 8, No.


4, 1986, pp. 24-28.
119

Meyers, M., "Structures at Risk. Preparing and Repairing Parking Garages," Public
Works, Vol. 121, No. 9, 1990, pp. 46-47.
120

Steele, M., "Protecting Concrete Parking Structures: From Asphaltics to


Polyurethanes," Journal of Protective Coatings and Linings, Vol. 23, No. 11, 2006, pp.
52-55.
121

Tighe, M. R., Hemba, D., "Industry Faces the Parking Problem," Civil Engineering
New York, N. Y., Vol. 60, No. 11, November 1990, pp. 65-66.
122

Delaney, T. J., "Repairing Post-Tensioned Parking Structures," Aberdeen's Concrete


Repair Digest, Vol. 7, No. 4, Aug-Sep 1996, pp. 186-189.
123

Nehil, T. E., "Rehabilitating Parking Structures with Corrosion-Damaged ButtonHeaded Post-Tensioning Tendons," Concrete International, Vol. 13, No. 10, Oct 1991,
pp. 66-73.
124

Anon, "Avoid Joint Deterioration in Concrete Parking Structures," Aberdeen's


Concrete Construction, Vol. 36, No. 8, Aug 1991, pp. 616-618.
125

Beeldens, A., and Vandewalle, L., "Durability of high strength concrete for highway
pavement restoration," In CONSEC '01: Third International Conference on Concrete
under Severe Conditions, Vancouver, BC, Canada, 2001, pp. 1230-1238.
126

Oh, B.H., Cha, S.W., Jang, B.S. & Jang, S.Y., "Development of high-performance
concrete having high resistance to chloride penetration," Elsevier Science SA, Nuclear
Engineering and Design (Switzerland), Vol. 212, No. 1-3, 2002, pp. 221-231.
Mehta, P. K., Concrete: Structure, Properties, and Materials,
Englewood Cliffs, New Jersey, 1986, pp. 353-367.
128

Prentice-Hall,

Chang, P. K., Peng, Y. N., and Hwang, C. L., "A Design Consideration for Durability
of High-Performance Concrete," Cement & Concrete Composites, Vol. 23, No. 4-5, Aug.
-Oct. 2001,pp. 375-380.

59

Hwang, C. L., Liu, J. J., Lee, L. S., and Lin, F. Y., "Densified Mixture Design
Algorithm and Early Properties of High Performance Concrete, " Journal of the Chinese
Institute of Civil and Hydraulic Engineering, Vol. 8, No. 2, 1996, pp. 217-229.
130

Shah, S. P., Wang, K., and Weiss, W. J., "Is High Strength Concrete Durable?"
Concrete Technology for a Sustainable Development in the 21st Century Eds. O. E. Gjorv
and K. Sakai, 2000, pp. 102-114.
131

Li, V. C , and Stang, H., "Elevating FRC Material Ductility to Infrastructure


Durability," Proceedings of 6th R1LEM symposium on FRC, Varenna, Italy, 2004, pp.
171-186.
132

Hillerborg, A., Modeer, M., and Petersson, P.E. "Analysis of Crack Formation and
Crack Growth in Concrete By Means of Fracture Mechanics and Finite Elements."
Cement and Concrete Research, Vol. 6, No. 6, 1976, pp. 773-782.
133

Mehta, P.K., "Durability-Critical Issues for the Future", Concrete International, Vol.
19, No. 6, 1997, pp. 27-33.
134

Mehta, P.K. and R.W. Burrows, "Building durable infrastructures in the 21st century."
Concrete International, Vol. 23, No. 3, 2001, pp. 57-63.
135

Xi, Y., B. Shing, N. Abu-Hejleh, A. Asiz, A. Suwito, Z. Xie, A, Ababneh


"Assessment of the Cracking Problem in Newly Constructed Bridge Decks in Colorado"
Colorado Department of Transportation Research Branch. Denver, Colorado. CDOTDTD-R-2003-3. March 2003.
136

Morgan, D. R., "Compatibility of Concrete Repair Materials and Systems,"


Construction and Building Materials, Vol. 10, No. 1, 1996, pp. 57-67.
137

International Concrete Repair Institute, "Guideline for Inorganic Repair Material Data
Sheet Protocol," Guideline No. 03740, 2003.
138

ASTM C1218/C1218M "Standard Test Method for Water-Soluble Chloride in Mortar


and Concrete" Vol. 4.02.
139

Li, V. C , "Reflections on the Research and Development of Engineered Cementitious


Composites (ECC)," Proceedings of the JCI International Workshop on Ductile Fiber
Reinforced Cementitious Composites (DFRCC) - Application and Evaluation (DRFCC2002), Takayama, Japan, Oct. 2002, pp. 1-21.
140

Li, V.C., "Integrated Structures and Materials Design," RILEM J. of Materials and
Structures, Vol. 40, No. 4, 2007, pp. 387-396.
141

Maalej, M., Hashida, T., and Li, V.C., "Effect of Fiber Volume Fraction on the OffCrack Plane Energy in Strain-Hardening Engineered Cementitious Composites," Journal
of American Ceramics Society, Vol. 78, No. 12, 1995, pp. 3369-3375.

60

Li, V. C , "Engineered Cementitious Composites - Tailored Composites Through


Micromechanical Modeling," Fiber Reinforced Concrete: Present and the Future, N.
Banthia, A. A. Bentur, and A. Mufti, eds., Canadian Society for Civil Engineering,
Montreal, Quebec, Canada, 1998, pp. 64-97.
143

Li, V. C , Wu, C , Wang, S., Ogawa, A, and Saito, T., "Interface Tailoring for StrainHardening Polyvinyl Alcohol-Engineered Cementitious Composites (PVA-ECC)," ACI
Materials Journal, Vol. 99, No. 5, Sept.-Oct. 2002, pp. 463-472.
144

Fischer, G., and Li, V. C , "Deformation Behavior of Fiber-Reinforced Polymer


Reinforced Engineered Cementitious Composites (ECC) Flexural Members under
Reversed Cyclic Loading Conditions," ACI Structural Journal, Vol. 100, No. 1, Jan-Feb.
2003, pp. 25-35.
145

Zhang, J., Maalej, M. and Quck, S. T., "Hybrid Fiber Engineered Cementitious
Composites (ECC) for Impact and Blast-Resistant Structures," Proceedings of the First
International Conference on Innovative Materials and Technologies for Construction and
Restoration-IMTCR04, Lecce, Italy, Vol. 1, June 2004, pp. 136-149.
Maaley, M., Quck, S. T., and Zhang, J., "Behavior of Hybrid-Fiber Engineered
Cementitious Composites Subjected to Dynamic Tensile Loading and Projectile Impact,"
ASCE Journal of Materials in Civil Engineering, Vol. 17, No. 2, 2005, pp. 143-152.
147

Qian, S., Lepech, M. D., Kim, Y. Y., and Li, V. C , "Introduction of Transition Zone
Design for Bridge Deck Link Slabs Using Ductile Concrete," ACI Structural Journal,
Vol. 106, No. 1, Jan. 2009, pp. 96-105.
148

Qian, S. and Li, V. C , "Influence of Concrete Material Ductility on Headed Anchor


Pullout Performance," ACI Materials Journal, Vol. 106, No. 1, Jan. 2009, pp. 72-81.
Li, V. C , "Engineered Cementitious Composites," Proceedings of ConMat'05,
Vancouver, Canada, August 22-24, 2005, CD-documents/l-05/SS-GF-01_FP.pdf.
150

Li, V.C., and Wang, S., "Flexural Behavior of GFRP Reinforced Engineered
Cementitious Composites Beams," ACI Materials Journal, Vol. 99, No.l, 2002, pp.11-21.
151

Fischer, G., and Li, V.C., "Effect Of Matrix Ductility On Deformation Behavior of
Steel Reinforced ECC Flexural Members Under Reversed Cyclic Loading Conditions,"
ACI Structural Journal, Vol. 99, No. 6, 2002, pp.781-790.
152

Kesner, K., and Billington, S. L., "Experimental Response of Precast Infill Panels
Made with DFRCC," DFRCC-2002 International Workshop, Takayama, Japan, 2002, pp.
289-298.
153

Parra-Montesinos, G., and Wight, J.K., "Seismic Response of Exterior RC Column-toSteel Beam Connections," Journal of Structural Engineering, Vol. 126, No. 10, 2000, pp.
1113-1121.

61

Li, V.C. & Lepech, M. 2004. "Crack resistant concrete material for transportation
construction," In TRB 83rd Annual Meeting, Washington, D.C., CD ROM, Paper 044680.
155

Li, V.C., Fischer, G., Kim, Y.Y., Lepech, M., Qian, S., Weimann, M. & Wang,
S., "Durable Link Slabs for Jointless Bridge Decks Based on Strain-Hardening
Cementitious Composites," Reportfor MDOTRC-1438. 2003.
156

Li, V.C, Horikoshi, T., Ogawa, A., Torigoe, S. & Saito, T., "Micromechanics-based
Durability Study of Polyvinyl Alcohol-Engineered Cementitious Composite (PVA-ECC),"
ACI Materials Journal, Vol. 101, No. 3, 2004, pp. 242-248.
157

Kunieda, M., and Rokugo, K., "Recent Progress on HPFRCC in Japan," Journal of
Advanced Concrete Technology, Vol. 4, No. 1, 2006, pp.19-33.
158

Li, V.C, Lepech, M., and Li, M., "Field Demonstration of Durable Link Slabs for
Jointless Bridge Decks Based on Strain-Hardening Cementitious Composites, " Michigan
DOT Report RC-1471, Dec. 2005.
159

Li, V. C , Li, M., Lepech, M., "High Performance Material for Rapid Durable Repair
of Bridges and Structures," Michigan DOT Report RC-1484, Dec. 2006.
160

Lin, Z., and Li, V. C , "Crack Bridging in Fiber Reinforced Cementitious Composites
with Slip-Hardening Interfaces," Journal of Mechanics and Physics of Solids, Vol. 45,
No. 5, 1997, pp. 763-787.

62

CHAPTER 2

Micromechanics Based Design of ECC Repair Materials

Within the material engineering phase of the Integrated Material and Structural
Engineering framework, shown in Figure 1.10, microstructure tailoring and material
processing

converge

to

ultimately

produce

composite

materials

with

desired

combinations of properties. This chapter focuses on the microstructure tailoring portion


of this framework through the development of a new version of ECC material called
HES-ECC, specifically for fast and durable repair applications. Besides tensile ductility,
HES-ECC also possesses high early-age strength properties, which are not achieved by
ordinary type I Portland cement based ECC (e.g. ECC M45). High early-age strength is
necessary for repaired structures to be able to return to service with minimum operations
interruption; tensile ductility is crucial for the repaired structures to be durable with
minimum repair and maintenance frequency during their service lives.
Wang and Li, 2006' first introduced the idea of simultaneously designing tensile
ductility and high early strength into cementitious materials. Their work mainly focused

63

on the micromechanical design of HES-ECC based on proprietary rapid-hardening


cement binders. Although other binder systems (i.e. Type I Portland cement and Type III
Portland cement) were also included in the study, the emphasis was on composite scale,
and a detailed study of the effects of these binder systems on micromechanical
parameters was not reported.

Type III Portland cement, however, is more widely

available, less expensive, more compatible with existing commercially available


admixtures, and more familiar to the construction industry compared to proprietary rapidhardening cement.

Driven by these reasons, the research described in this chapter

focuses on the detailed micromechanical design and material characterization of HESECC based on Type III Portland cement binder systems.
Within this chapter, the ECC design philosophy and scale linking are described
first, followed by a review of strain-hardening criteria and conditions for saturated
multiple cracking. After the initial mix design to meet the target early-age strength and
workability requirements, effects of the Type III Portland cement binder system on the
age dependency of the micromechanical parameters (i.e. matrix, fiber/matrix interfacial
properties) and the pseudo strain-hardening index (PSH) were investigated. Based on
these factors, the initial mix was then further tailored through introducing artificial flaws
to maximize multiple-cracking behavior and tensile strain capacity while maintaining
self-controlled crack width below 100 \im. Finally, this new Type III Portland cement
based HES-ECC was characterized based on composite traits such as tensile, compressive,
and flexural properties to provide a database for future structural applications. Durability
of this material, such as shrinkage properties, was also investigated and will be reported
in later chapters.

64

2.1

High Performance Fiber Reinforced Cementitious Composites


In the past decade great strides have been made in developing high performance

fiber reinforced cementitious composites (HPFRCC). HPFRCC were first classified by


Naaman and Reinhardt2 as materials that achieved different degrees of tensile ductility,
often accompanied by a macroscopic pseudo-strain hardening response after first
cracking.

Under this definition, HPFRCC is distinguished from an ordinary Fiber

Reinforced Concrete (FRC) that has a tension-softening response. Figure 2.1 compares
the uniaxial tensile stress-deformation relation of concrete, FRC, and HPFRCC. For
concrete, crack formation results in a sudden drop in load carrying capacity. For FRC,
after cracking occurs, tensile load capacity drops at a relatively slow rate as the single
crack enlarges. This is called "tension-softening". For HPFRCC, after the first crack
forms, tensile load capacity continues to rise with increasing strain ("strain-hardening")
through formation of multiple cracks. The deformation during the elastic and strainhardening stages in HPFRCC can be treated as "volumetric straining" rather than
"localized crack opening"3.
Efforts to design tensile ductility into cementitious materials started in the 1970s,
and have mainly focused on using continuous aligned fibers or large-volume
discontinuous fibers.

Aveston et al., 19714and Krenchel and Stang, 19895 achieved

tensile ductility hundreds of times that of normal concrete using continuous aligned fibers
in concrete materials. Textile reinforced concrete materials, representing the modern
version of continuous fiber reinforced concrete, were developed by Curbach and Jesse,
19996, and Reinhardt et al, 20037.

Moreover, pultruded continuous fiber reinforced

concrete was recently developed by Mobasher et al, 20068. Additionally, cementitious

65

composite materials using discontinuous fibers at high dosage (4-20%) in cement


laminates9 or in SIFCON (Slurry Infiltrated Fiber Concrete)10'11 attain higher tensile
strength and strain capacity than normal concrete, but much lower tensile ductility than
continuous fiber and textile reinforced cementitious materials.
The HPFRCC materials described above have so far found limited field
applications. This is due to their considerable cost and difficulty in processing, especially
in on-site construction.

Considering the huge consumption of materials in the

construction industry, cost effectiveness and construction efficiency are of foremost


importance. These requirements can hardly be met using aligned fiber or large-volume
discontinuous fiber materials. Various methods have been proposed for overcoming
processing difficulties.

For example, Shah et al. employed an extrusion process to

produce HPFRCC with tensile strain capacity below 1%, at greater than 4% fiber volume
content. Most of these research efforts have focused on approaches to embed a large
volume of fibers in order to attain high composite performance. The market, however,
demands low cost and ease of processing.
In recent years, a new class of HPFRCC has emerged.

ECC (Engineered

Cementitious Composites), originally developed at the University of Michigan12, features


high ductility with a moderate amount (no more than 2 vol%) of short discontinuous
fibers.

Despite its apparent similarity in composition (e.g. cementitious binder, fine

aggregates, water, and fiber) compared with other HPFRCC materials, ECC's design
principle is quite distinct. Instead of relying on a high fiber content to achieve strainhardening behavior, the development approach for ECC is to design synergistic
interactions between the fiber, matrix, and fiber/matrix interface to maximize tensile

66

ductility of the composite through formation of closely spaced multiple microcracks


while minimizing fiber content. Through this micromechanics and fracture mechanics
based material design approach, ECC can achieve tensile strain capacities of 3-5%,
approximately 300-500 times that of concrete or FRC, with 2 vol% or less discontinuous
polymer fibers. The theoretical framework of ECC material design is reviewed in the next
section.

2.2

ECC Micromechanics Based Design Theory

2.2.1

ECC Design Framework - Scale Linking


Micromechanics based design theory of ECC was first established in the early

90's 1 3 ' 1 4 . This theory links the measurable constituent parameters to the cracking
propagation mode, and then to conditions for composite tensile strain-hardening. Scale
linking (Figure 2.2) is a fundamental characteristic of the ECC design approach, in which
understanding and tailoring of microscale constituent parameters are the keys to
achieving target macroscale composite behavior.
The scale linking between ECC microstructure and composite tensile behavior is
illustrated in Figure 2.2. As a composite material, ECC contains three main phases:
libers, matrix (including pre-existing flaws), and fiber/matrix interface. Each phase can
be defined by a set of micro-parameters as shown in Table 2.1. Under uniaxial tension,
ECC composites exhibit tensile strain hardening behavior at the macroscale (mm - cm)
through a multiple cracking process.

Steady-state crack propagation is a necessary

condition to ensure multiple cracking, which is governed by the fiber bridging properties
across cracks at the mesoscale (\im - mm). The fiber bridging spring law across a crack,

67

quantified by the fiber bridging stress vs. crack opening relationship 0(8), is the
integration of the bridging force contributed by every single fiber bridging the crack. For
an individual fiber, its bridging force for a given crack opening is determined by its
pullout behavior from the surrounding matrix, and governed by fiber and interface
properties at the microscale (nm - \im), as well as by fiber embedment length and
inclination angle between the fiber axis and the crack face normal. In summary, the
micromechanics model links microscale constituent parameters to fiber bridging
constitutive behavior on the mesoscale; steady-state crack analysis links fiber bridging
properties to tensile strain-hardening behavior on the composite macroscale.

Once

established, the model-based linking provides a systematic framework for optimizing


composite tensile properties with the minimum amount of fibers by strategically tailoring
the microstructure at the smallest scales.

2.2.2

Conditions for Tensile Strain Hardening


The tensile strain-hardening behavior of ECC is realized by tailoring the

synergistic interaction between the fibers, matrix, and fiber/matrix interface using
micromechanics theory. As a fiber reinforced brittle mortar matrix composite, ECC's
pseudo strain-hardening behavior is achieved through sequential formation of matrix
multiple cracking. The fundamental requirement for matrix multiple cracking is that
steady-state flat crack propagation prevails under tension, which was first characterized
by Marshall and Cox15 for continuous aligned fiber-reinforced ceramics, and extended to
discontinuous fiber-reinforced cementitious composites by Li and Leung13. To ensure
steady-state cracking, the crack tip toughness JtjP must be less than the complementary

68

energy Jt,' calculated from the fiber bridging stress o versus crack opening 6 curve, as
illustrated in Figure 2.3 and shown in Equations 2.1 and 2.2.
Jtip*o060-f<?(S)dd
o

= J'b

(2.1)

J*-f=-

(2-2)

where ao is the maximum bridging stress corresponding to the opening 5o, Km is the
matrix fracture toughness, and Era is the matrix Young's modulus. Equation 2.1 employs
the concept of energy balance during flat crack extension between external work (ao So),
crack flank energy absorption through fiber/matrix interface debonding and sliding (
f o(d)dd),

and crack tip energy absorption through matrix breakdown (JtiP).

This

energy-based criterion determines whether the crack propagation mode is steady-state flat
cracking or Griffith cracking16, as illustrated in Figure 2.4.
The fiber bridging stress versus crack opening relationship o"(8), which can be
viewed as the constitutive law of fiber bridging behavior, is analytically derived from
fracture mechanics, micromechanics, and probabilistics tools.

In particular, the

energetics of tunnel crack propagation along the fiber/matrix interface is used to model
the debonding process of a single PVA fiber from the surrounding cementitious matrix.
After debonding is complete the fiber pullout stage begins, and is modeled as sliphardening behavior with the assumption that non-linear frictional stress increases with
slip distance. By these means, the full debonding-pullout process of a single fiber, with
given embedment length, is quantified as the fiber bridging force vs. fiber displacement
relationship13.

Probabilistics is then introduced to describe the randomness of fiber

69

location and orientation with respect to a crack plane, with the assumption of uniform
random fiber distribution17. The random orientation of the fibers also necessitates the
accounting of the mechanics of interaction between an inclined fiber and the matrix crack.
In addition, the snubbing coefficient f and strength reduction factor f are introduced to
account for the interaction between fiber and matrix, as well as the reduction of fiber
strength when pulled at an inclined angle. As a result, the a(5) curve is expressible as a
function of micromechanics parameters, including: fiber volume content Vf, fiber
diameter df, fiber length Lf, fiber Young's modulus Ef, matrix Young's modulus Em,
interface chemical bond Gd, interface frictional bond x0, and slip-hardening coefficient p,
as well as f and f.
Apart from the energy criterion (Equation 2.1), another condition for pseudo strainhardening is that the matrix tensile cracking strength a c must not exceed the maximum
fiber bridging strength 00.
ac<a0

(2.3)

where crc is determined by the matrix fracture toughness Km and pre-existing internal flaw
size ao. While the energy criterion (Equation 2.1) governs the crack propagation mode,
the strength-based criterion (Equation 2.3) controls the initiation of cracks. Satisfaction
of both Equation 2.1 and Equation 2.3 is necessary to achieve ECC strain-hardening
behavior; otherwise, the composite behaves as a normal fiber reinforced concrete (FRC)
and tension-softening behavior results.

70

2.2.3

Condition for Saturated Multiple Microcracking


For ECC materials with pseudo strain-hardening behavior, high tensile strain

capacity results from saturated formation of multiple microcracks. Material tensile strain
capacity increases as the number of microcracks increases.

While the steady-state

cracking criteria ensures the occurrence of multiple cracking, it is not directly related to
the intensity of multiple cracking.

Aveston et al.18 and later Wu and Li 19 derived

minimum crack spacing Xd for aligned continuous fiber and short discontinuous fiber
reinforced composites respectively, assuming that the matrix cracking strength is uniform
at each section. Under this assumption, crack spacing between Xd and 2xd is predicted
after crack satuation.

The minimum crack spacing is determined by the distance

necessary for transferring load from the bridging fibers at one crack back into the matrix
through the fiber/matrix interface shear, so that the next crack can be formed. However,
in ECC uniaxial tension specimens, a wide distribution far exceeding two times the
minimum crack spacing is often observed, due to the variation in ECC matrix properties
and non-uniform fiber dispersion.

Large crack spacing means that the potential of

reinforcing fibers is not effectively utilized, and maximum tensile strain capacity is not
achieved.
Once the steady-state cracking criteria is satisfied, the number of microcracks that
can be developed is determined by (i) the maximum fiber bridging stress Oo, and (ii) the
matrix properties, in particular the pre-existing flaw size distribution, and the matrix
fracture toughness. Considering that ECC is a non-homogenous brittle-matrix composite,
first-cracking strength is determined by the largest flaw size in the section where the first
microcrack is initiated. Its ultimate tensile strength is determined by the "weakest"

71

section, where the fiber bridging capacity (maximum bridging stress a 0 ) is the lowest
among all sections subjected to the same level of stress. Therefore, the maximum fiber
bridging stress Oo at the "weakest" section imposes a lower bound of critical flaw size cmc,
so that only those flaws larger than cmc can be activated and contribute to multiple
cracking. There also exists a minimum crack spacing controlled by interface properties,
which imposes an upper bound for the density of multiple cracking.

Matrix randomness - flaws


Matrix imperfections, e.g. random distribution of pre-existing flaws, is one cause
of the variation in crack spacing and tensile strain capacity. In ECC composites with a
quasi-brittle matrix, cracks initiate from pre-exisitng flaws in the matrix. Most of the
flaws have sizes below 4 mm, and their existance reduces the cracking strength of the
cementitous matrix. Li and Wang, 200520 computed the effect of initial flaw size on the
theoretical cracking strength of an infinite two-dimensional ECC plate under uniaxial
tension. The ECC contains poly-vinyl-acohol (PVA) fiber at a volume content of 2%.
As shown in Figure 2.5, the theoretical tensile strength of the composite without
macrodefects is assumed to be 6.5 MPa. If the composite contains an initial crack 1 mm
wide, the matrix tensile cracking strength is reduced to 5.4 MPa. A larger initial crack
with 4 mm width reduces the cracking stress to 4.8 MPa. This reduction in matrix tensile
cracking strength is important to make Equation 2.3 hold, because the matrix tensile
cracking strength a c must be lower than the maximum fiber bridging strength Oo, which is
5.5 MPa for this particular PVA-ECC theoretical specimen, to satisfy the strength-based
strain-hardening criterion. Therefore, the critical flaw size c mc can be determined as the

72

flaw size that corresponds to the cracking stress 5.5 MPa (o c = o 0 = 5.5MPa). The critical
flaw size is what separates inert and active flaws - only flaws larger than cmc can be
activated and contribute to multiple cracking. The pre-existing flaws in ECC can be
entrapped air pores, weak boundaries between phases, and cracks induced by material
differential shrinkage, which all possess a random nature and strongly depend on
processing details and environmental effects. The number of cracks that can form before
reaching the maximum fiber bridging stress may therefore be limited, and can vary
significantly from batch to batch. Therefore, a large number of flaws slightly larger than
cmc (flaws much larger than cmc will lead to a reduction in the net cross section and fiber
bridging stress at the crack section) are preferred for saturated multiple cracking and high
tensile strain capacity.

Fiber dispersion non-uniformity


Fiber dispersion non-uniformity is another contributor to the variation in tensile
strain capacity and unsaturated multiple cracking. With a fixed fiber volume percentage,
the maximum fiber bridging stress Oo at the weakest section is determined by the degree
of fiber dispersion uniformity in the composite. Fiber dispersion uniformity is directly
influenced by the rheology characteristics of the fresh ECC during processing.

The

uniformity of fiber dispersion determines the maximum bridging stress o 0 , the shape of
the o(5) curve at the weakest section, and the critical flaw size cmc. Non-uniform fiber
dispersion leads to a reduction of the value of ao at the weakest section, which increases
the critical flaw size cmc. Therefore, less pre-existing flaws with sizes larger than cmc can
be triggered and contribute to multiple cracking, resulting in a relatively lower tensile

73

strain capacity. Non-uniform fiber dispersion also shifts the 0(6) curve downwards, and
may reduce complementary energy J b ' to less than Jtip. In this case, the steady-state
criteria are violated and tension-softening behavior results. ECC then loses its ductile
behavior and becomes a regular FRC material.

Ideally, processing of ECC should

optimize fiber dispersion to achieve a uniform random distribution state, therefore


minimizing the probability of creating "weak" sections with lower fiber content. By
these means, the largest possible tensile strain capacity can be achieved by maximizing
multiple cracking behavior.
The random nature of pre-existing flaw size and fiber distribution in ECC leads to
variation in J b and JtiP. A large margin between J b and Jtip, and between OQ and a c , is
preferred to increase the tendency toward saturated multiple cracking. The pseudo strainhardening (PSH) performance index is used to quantitatively evaluate the margin, and is
defined as follows (Equation 2.4).

PSH energy =

~
(2.4)

P W
_ ^0
n strength ~

The micromechanics-based strain-hardening criteria, Equation 2.1 and 2.3, are


used to guide ECC design in the following sections. The evaluation of PSH is helpful in
quantifying the saturation of multiple cracking and the robustness of tensile ductility for
the newly developed ECCs.

74

2.3

High Early Strength Engineered Cementitious Composites

2.3.1

Background
There is an increasing demand for fast and durable repair of concrete structures

(e.g. bridge decks, highway pavements, parking structures, and airport runways), where
the least disruption of operations is needed. This requires that repair materials be able to
gain strength rapidly during early ages, and also remain durable throughout the repaired
structure's service life. For example, rapid concrete pavement repairs, including fulldepth repairs and patch repairs, have become common on many busy highways
throughout North America21. Highway transportation authorities often require the repair
job to be completed in 6-8 hours at night so that the lane can be reopened to traffic the
next morning.

Overnight closures are complicated by the fact that these pavements

typically carry heavy traffic, and heavy delays and user costs are induced by construction.
For freeway and toll roads it is very often that only overnight closures are permitted22.
Overnight construction is also common for airport pavements23'24, especially for bridging
areas where taxiways cross, high traffic volume appron areas, and runaway intersections
25

. Very fast setting concrete is used in such circumstances. Beside short closure time,

low maintenance and repair frequency are similarly important to avoid operations
interruption. This requires the repair to be durable under combined environmental and
structural loading conditions. Therefore, high early-age strength as well as durability (e.g.
resistance to cracking, chloride penetration, disintegration due to ASR, etc.) are critical
characteristics desired for concrete repair materials. Additionally, the repair materials
should be cost effective and easy to be processed and placed for various construction
situations.

75

Over the past two decades, high early strength concrete materials have been
successfully developed by academic and industry groups with various strength
(compressive and flexural) gain rates, depending on the types of cement binders and
accelerating admixtures used 26 ' 27 ' 28,29 ' 30 ' 31 ' 32 ' 33 . While possessing the desired high early
strength properties, these materials have been perceived to be more prone to early-age
cracking due to their higher thermal and autogeneous shrinkage, caused by faster earlyage hydration and heat release 34 ' 35 ' 36 . Cracking leads to early deterioration and great
reduction in the service life of these repairs.

Additionally, reduced freeze-thaw

resistance28 has also been found in some very high early strength concrete mixes, limiting
their applications in cold regions. Achieving high early-age strength and preventing
early-age cracking are conflicting goals that must be carefully balanced through new
material technology to achieve acceptable performance of concrete repairs.
This work targets on designing an innovative repair material, called HES-ECC
(High Early Strength Engineered Cementitious Composites), based on the ECC
micromechanical design theory discussed in Section 2.2. Conventional ECC mixes use
Type I ordinary portland cement (OPC), and have relatively slow strength development.
The new HES-ECC material should have the high early-age strength desired for rapid
repair applications, while possessing large tensile ductility and self-controlled tight crack
width below 100 um during the strain-hardening stage. Wang et al.37 reported that as
crack width increases from 100 um to 500 um, the permeability coefficient increases
nearly seven orders of magnitude, from 1.0 x 10"n m/sec to 1.0 x 10"4 m/sec. However,
for crack widths under 100 um, the permeability coefficient remains nearly identical to
that of uncracked concrete, suggesting that for crack widths below this threshold there is

76

no significant increase in permeability after cracking. The tensile ductlity and selfcontrolled crack width of repair materials are essential for achieving repair durability by
preventing cracking and penetration of aggressive agents, which has been validated in
this dissertation work and will be discussed in Chapters 4, 5, and 6. Besides high earlyage strength and tensile ductility, the HES-ECC material should also be cost-effective,
and possess workability that make it easy to be applied in different construction
applications.

2.3.2

High Early Strength Requirements


When a repair material is applied to a structure, such as a highway overpass, the

strength gain rate during early age determines when the repaired structure can be
reopened to traffic.

Different repair applications have different minimum strength

requirements, which must be reached before the structure can be returned to service. For
example, large deck patches are typically given a 24 hour cure; small deck patches
generally use fast setting mortar and open to traffic after 4 to 6 hours; and for prestressed
concrete beam end repairs, both the early age strength and the 28 day strength are
specified; for airport runway spall repairs, overnight closure is often preferred. Selection
of a repair material depends on the ability of the material to meet early age strength
performance targets.
Currently, there is no general standard for minimum high early strength. The
California Department of Transportation (Caltrans) specifies a minimum flexural strength
of 400 psi (2.8 MPa) prior to opening to highway traffic for full depth highway pavement
repairs38.

The New Jersey State Department of Transportation (NJDOT) specifies a

77

minimum compressive strength of 3000 psi (20.7 MPa) in 6 hours, and a minimum
flexural strength of 350 psi (2.4 MPa) in 6 hours for the "Fast-tract mix" developed in the
mid-90's39. The Michigan Department of Transportation (MDOT) specifies minimum
compressive strengths of 2000 psi (13.8 MPa) in 2 hours, 2500 psi (17.2 MPa) in 4 hours,
and 4500 psi .(31.0 MPa) in 28 days (higher for prestressed concrete applications) for
prepackaged hydraulic fast-set materials used in structural concrete repairs40. Moreover,
Parker et al.41 suggested a minimum compressive strength of 2000 psi (17.2 MPa) for
road patching repair to prevent damage when initially reopened to traffic. An FHWA
national research program report42 on high performance concretes designates three
categories based on strength: (a) very early strength, (b) high early strength, and (c) very
high strength. The very early strength concretes have strength of at least 3000 psi (20.7
MPa) within 4 hours after placement.

The high early strength concretes have a

compressive strength of at least 5000 psi (34.5 MPa) within 24 hours. In addition, the
FHWA43 recommends a minimum compressive strength of 1000 psi (6.9 MPa) at 3 hours
and 3000 psi (20.7 MPa) at 24 hours for rapid-setting cementitious concretes. To meet
these various requirements, the target compressive strength for HES-ECC is prescribed as
below:
> 2500 psi (17.2 MPa) at 4 hr
> 3000 psi (20.7 MPa) at 6 hr
> 5000 psi (34.5 MPa) at 24 hr
> 7000 psi (48.3 MPa) at 28 d
The minimum compressive strength of 2500 psi (17.2 MPa) at 4 hr and 3000 psi
(20.7 MPa) at 6 hr enables the repaired bridge or structure to be opened to traffic within

78

4-6 hours after placement. Consequently a fast repair can be done at night, and then
reopened to traffic the next morning. The early age strength gain rate within 24 hours is
considered sufficient for repair jobs in heavy traffic areas. The minimum compressive
strength of 7000 psi (48.3 MPa) at 28 days should be adequate for prestressed concrete
applications. As shown in Figure 2.6, the target compressive strength at different ages
for HES-ECC serves as an envelope of the requirements by the DOTs and FHWA.
Besides compressive strength requirements, a minimum 15 minutes working time is
required for repair handling and placement.

2.3.3

Methods of Achieving High Early-Age Compressive Strength


High early strength of concrete materials is mostly attained by using traditional

concrete ingredients and concreting practices, although sometimes special materials or


techniques are needed. Depending on the age at which the specified strength must be
achieved and on application conditions, high early strength can be achieved by using one
or a combination of the following:
a) Type III Portland Cement
b) Proprietary rapid hardening cements
c) High cement content
d) Low water/cementing materials ratio
e) Chemical admixtures (accelerator, superplasticizer, retarder, etc.)
f) Silica fume (or other supplementary cementing materials)
g) Insulation to retain heat of hydration
h) Higher freshly mixed concrete temperature

79

i) Higher curing temperature


j)

Steam or autoclave curing


Type III Portland Cement is widely used in construction where a more rapid rate

of strength gain is desirable. Type III Portland Cement and Type I Ordinary Portland
Cement (OPC) are both normal Portland cement, and the same manufacturer will produce
each with the same composition. The strength gain of normal Portland cement is mainly
contributed by the hydration of Tricalcium Silicate (C3S) and Dicalcium Silicate (C2S),
which produces a calcium silicate hydrate (C-S-H). By grinding the cement more finely
compared with Type I cement (540 m2/kg vs. 370 m2/kg), the resultant cement surface
area in contact with water will increase, leading to faster hydration and more rapid
development of strength. By this method, Type III cement can develop strength much
faster than Type I cement - within the first 3-8 hours. The amount of strength gain of
Type III cement over the first 24 hours almost doubles that of Type I cement 4.
Proprietary rapid hardening cements such as magnesium phosphate-based cement,
gypsum-based cement, and calcium aluminate cement have also been used to obtain very
high early strength in some concrete mixes26'27'28'29'30. However, many proprietary rapid
hardening cements often perform unpredictably under various conditions. For example,
magnesium phosphate based cement is extremely sensitive to water content, and its
strength can be severely reduced by a very small amount of extra water45; concrete
containing gypsum-based cement may lose durability in freezing weather or when
exposed to moisture , and the presence of sulfates in its mixture may promote embedded
steel corrosion45; calcium aluminate cement may lose strength at high curing temperature
due to chemical conversion . Furthermore, most commercially available proprietary

80

rapid-hardening cements are expensive compared with ordinary type III cement. Due to
the unpredictable performance and higher cost, proprietary rapid-hardening cements will
not be considered in this research, unless HES-ECC based on Type III cement fails to
achieve the target strengths.
Chemical admixtures play a key role in producing high early strength concrete
mixtures that meet strength and workability criteria; they include water-reducing
admixtures, accelerating admixtures, and hydration controlling admixtures.

Water-

reducing admixtures increase early strength by lowering the quantity of water required
while increasing the workability for appropriate concrete placement and finishing
techniques, and increase strength at later ages by dispersing the cement to allow more
efficient hydration. Accelerating admixtures aid early strength development and reduce
initial setting times by increasing the rate of hydration. Hydration controlling admixtures
allow time for transportation of the concrete from the ready-mix plant to the jobsite, and
provide time for adequate placement and finishing.

Chemical admixtures and their

potential effects on material performance are reviewed below.

Water-reducing admixtures
As high cement content and a low water/cement ratio are favorable for achieving
high early strength, a water-reducing admixture should be necessary to effectively lower
the water/cement ratio in HES-ECC by reducing the water required to attain a given
slump. Consequently, it improves the strength and impermeability of the cement matrix.
The strength increase can be observed in as early as one day if excessive retardation does
not occur 47 .

Furthermore, since the cement is dispersed and a more uniform

81

microstructure is developed, the compressive strength can be as much as 25% greater


than that achieved by the decrease in w/c alone48.
Water-reducing admixtures are often observed to increase the rate of shrinkage
and creep of concrete, depending on the cement type and the particular admixture.
However, after 90 days of drying, there is little difference in shrinkage compared to a
control concrete49'50. The shrinkage properties of the developed HES-ECC including the
use of water-reducing admixture and related durability issues will be investigated in
Chapter 4.

Accelerating admixtures:
Accelerating admixtures accelerate the normal strength development and
processes of setting. Regular accelerators are used to speed construction by permitting
earlier attainment of sufficient strength to allow removal of formwork and to carry
construction loads. Quick-setting admixtures provide setting times of only a few minutes.
They are generally used in shotcreting applications and for emergency repair in general,
where very rapid development of rigidity is required. Accelerators are also beneficial
during winter concreting because they can partially overcome the slower rate of hydration
due to low temperatures, and shorten the period for which protection against freezing is
required51'52.
Accelerating admixtures can be divided into three groups: (1) soluble inorganic
salts, (2) soluble organic compounds, and (3) miscellaneous solid materials53. Many
soluble inorganic salts, including chlorides, bromides, nitrites, nitrates, fluorides,
carbonates, thiosulfates, silicates, aluminates, and alkali hydroxides, can accelerate the

82

setting and hardening of concrete to some degree, calcium salts generally being the most
effective. Soluble carbonates, aluminates, fluorides, and ferric salts have quick-setting
properties and they are commonly used in shotcreting applications. Calcium chloride is
the most widely used because it is the most cost effective, in terms of giving more
strength acceleration at a particular rate of addition when compared to other
accelerators54. However, the use of calcium chloride will increase the rate of corrosion of
metals embedded in concrete. The ACI Building Code places limits on the chloride
content of concrete that preclude its use for both prestressed and reinforced concrete.
Chloride-free accelerators should be used in such cases.
Organic compounds for most commercial uses include triethanolamine, calcium
formate55. Calcium acetate, calcium propionate, oxalic acid, lactic acid and various ring
compounds are also in this group.

These compounds accelerate the hydration of

tricalcium aluminate and produce more ettringite at early age. The reaction of
triethanolamine with Portland cement is complex. However, excessive amounts of these
compounds can result in retardation effects56.
Solid materials such as calcium fluoroaluminate or calcium sulfoaluminate can be
used to obtain rapid-hardening characteristics. In certain instances, finely ground silicate
minerals and hydrated Portland cement itself have been found to act as accelerators due
to the effects of seeding. Also, introduction of small amounts of calcium-aluminate
cement can significantly shorten the setting time of Portland cement concrete

55 57 58 59

' ' ' .

Solid materials are not often used in concrete materials for accelerating hydration.
Although accelerating admixtures can be expected to increase early strength, the
increase diminishes with time, and later strengths (at 28 days or more) are likely to be

83

lower than the strength of concretes without an accelerating admixture . This reduction
in later strength is more pronounced when the initial accelerating effects are large.
Accelerating admixtures may also increase the rate of drying shrinkage and creep,
although the ultimate values are not affected61. The early shrinkage leads to high tensile
stresses in restrained repair material, often resulting in cracking or interface delamination.
The undesireable effects of accelerating admixture use on concrete material performance
negate many of its benefits.

2.3.4

Initial Material Composition Selection Based on Strength Requirements


The material development of HES-ECC contains three phases. The first phase is

matrix design for high early strength (Section 2.3.4). In this phase, the main factors that
affect early strength gain rate, including cement type, water/cement ratio, accelerator, and
curing conditions, are experimentally investigated.

The second phase is composite

tailoring for saturated multiple cracking and high tensile ductility (Section 2.3.5). The
third phase is detailed characterization of mechanical properties, including tensile,
compressive, and flexural properties, of the developed HES-ECC composite material
(Section 2.3.6).

2.3.4.1 Experimental Program


The ingredient materials used in this study are listed in Table 2.2 with information
on product name, manufacturer, chemical composition and particle size. The trial mixes
for initial material composition selection are listed in Table 2.3. Two types of cements
were investigated.

Type III Portland cement has the same composition as Type I

84

Ordinary Portland Cement (OPC), except that it is finely ground. ASTM class F fly ash
is a pozzolantic material that is an integrated part of conventional ECC. It is eliminated
from some of the mixes in this study to improve early age compressive strength.
Fine silica sand is the main aggregate in ECC, with a maximum particle size of
270 um and average size of 110 urn. Polystyrene (PS) bead is another type of aggregate
that forms a very weak bond with the cementitious matrix so that it behaves as an
artificial flaw under tensile loading. As the uniaxial tensile test decribed in Section 2.3.5
shows, the initially selected mix was not able to meet the target tensile strain capacity
requirements, so polystyrene beads were deliberately introduced into the mix to control
initial flaw size and distribution at the material tailoring stage. By these means, the
composite material can develop saturated multiple microcracks during the strainhardening stage, leading to a large tensile strain capacity.
The chemical admixtures used in this study include accelerating admixture and
superplasticizer. The accelerating admixture is Pozzolith

NC 534 from Master Builders,

Inc. The accelerating species in this brand of admixture are calcium nitrate and sodium
thiocyanate. The recommended maximum effective dosage of Pozzolith

NC 534 is 4

wt% to cement. Superplasticizer is necessary to achieve sound workability when the


mixing composition has a low water/cement ratio. The superplasticizer used in this
research project was Glenium 3200 HES from Master Builders Inc. Glenium 3200
HES is a late generation superplasticizer with powerful dispersion capacity at low
water/cement ratios. It can better facilitate hydration of cement particles, resulting in a
stronger accelerating effect. Due to combined electrostatic and steric repulsion forces,

85

Glenium 3200 HES requires a lower dosage than other types of superplasticizer, such as
melamine formaldehyde based ML330, to achieve the same workability.
The fiber in this study is REC15 Polyvinyl-alcohol (PVA) fiber from Kuraray Co.,
Japan, used at a moderate volume fraction of 2% in ECC materials. PVA fiber possesses
great potential as a reinforcing fiber because of its high strength, high modulus of
elasticity, and relatively low cost. However, directly applying commercially available
PVA fiber has resulted in composites with low tensile strain capacities, on the order of
0.5-1.0%, even when as much as 4% fiber by volume is used62. This poor composite
performance can be attributed to the hydrophilic behavior of PVA fiber, that leads to
excessive chemical bonding (i.e. interface debonding fracture energy Gd) with
cementitious hydrates and strong slip-hardening response during pullout, which
consequently results in premature rupture of the fibers ' . In order to maximize Jt,', the
micromechanical models suggested that the fiber/matrix interfacial bond should be
significantly reduced. This leads to the adjusted fiber manufacturing process that applies
a 1.2% oil agent content coating to the fiber surface65. This interfacial modifying process
reduces both chemical bond Gd and frictional bond To, and increases Jb . The dimensions
of the PVA fiber are 12 mm in length and 39 urn in diameter, on average. The nominal
tensile strength of the fiber is 1620 MPa and the density of the fiber is 1300 kg/m3. The
detailed properties of REC 15 PVA fiber are listed in Table 2.4.
A Hobart mixer with 13L capacity was used in preparing all ECC mixtures. Solid
ingredients, including cement, fly ash, and sand, were first mixed for approximately 1
minute. Water and superplasticizer were slowly added into the dry mixture and mixed
for another 3 minutes. Once a consistent and uniform state of mortar was reached, fibers

86

were slowly added into the mortar matrix and mixed until they were evenly dispersed.
The mixture was then cast into molds. Accelerating admixture, if used, was added after
the fiber and just before casting. The molds were covered with plastic sheets and cured
in air at room temperature (203C). The specimens were demolded at the earliest
possible testing age, ranging from 4 hours to 24 hours.
Compressive testing was carried out according to ASTM C39 "Standard Test
Method for Compressive Strength of Cylindrical Concrete Specimens" on standard
cylinders measuring 75 mm (3 in.) in diameter and 150 mm (6 in.) in length. The ends of
the specimens were capped with a sulfur compound to ensure a flat and parallel surface,
and better contact with the loading device. Tests were conducted on a FORNEY F50
compression test system (Figure 2.7). The tests began 4 hours after casting or once
adequate strength had been developed. The age of the specimen was recorded as the time
elapsed from the end of casting to the start of testing.

2.3.4.2 Compression Test Results and Discussion


The compressive strength development of various mixes are summarized in
Tables 2.5, 2.6, and 2.7 and plotted in Figures 2.8, 2.9, 2.10, and 2.11. Three specimens
were tested for each test series and the average values were plotted.
The replacement of Type I by Type III cement greatly improved the early age
compressive strength gain rate (Figure 2.8). For example, the 12h and 24h compressive
strengths both increased by more than 60%, even though the 14d and 28d compressive
strengths exhibited little difference. However, Mix 2 containing Type III cement was
unable to achieve the target high early strength. Its 4h compressive strength was 3.4 - 3.6
MPa, which is far below the target 17.2 MPa at 4h. The 6h compressive strength (17.3 87

19.1 MPa) also failed to meet the requirement of 20.7 MPa. The high content of fly ash
was considered to be the factor that delayed the early-age compressive strength
development. Therefore, fly ash was eliminated from the following mixes.
Lowering the water/cement (w/c) ratio was shown to be an effective way to
improve early-age and late-age compressive strength. This is more effective for ECC
mixes than regular concrete, as ECC mixes often contain higher content of cement. As
shown in Figure 2.9, reduction of w/c from 0.53 to 0.36 leads to a greater than 100%
increase in compresive strength at 4h and 28d. The curing time required to reach the 4h
compressive strength target of 17.2 MPa was 24h, 8.5h, and 5.5h for w/c 0.53, 0.45, and
0.36 respectively. The curing time required to reach the 6h compressive strength target of
20.7 MPa was 36h, llh, and 6h for w/c 0.53, 0.45, and 0.36 respectively. Mix 5, with
w/c = 0.36, met the 6h, 24h, and 28d compressive strength requirements, but not the 4h
compressive strength requirement.
The use of polycarboxylate based superplasticizer, e.g. GL3200, was crucial to
achieve very high early strength when a reduced w/c ratio was used in a type III cementbased system. It was also noticed that a higher dosage has a strong retarding effect. The
effectiveness of superplasticizer is not only dependent on its composition, but also the
chemistry of the cementitious phase and the presence of other admixtures, e.g.
accelerators, retarders, etc. The compatibility of commercially available superplasticizers
and other admixtures with type III portland cement is one compelling driver of the
preference toward portland cement based-systems over proprietary rapid hardening
cement binder systems.

88

To further improve strength development within the first 6 hours, additional


accelerator was incorporated (Mix 6). As shown in Figure 2.10, the use of calcium
nitrate based accelerator NC534 at dosage of 4.0% reduced the age to reach 17.2 MPa
compressive strength from 5.5h to less than 4h. The 6h, 24h, and 28d compressive
strengths also satisfied the target strength requirements.
The curing condition was found to have a significant influence on the strength
gain rate at early ages. As shown in Figure 2.11, Mix 6 developed strength much faster
in the first 12 hours under insulated conditions, where the specimens were stored in
plastic containers with insulation foam. The target 17.2 MPa at 4h compressive strength
could be achieved as early as 3 hours.

This can be explained by the temperature

controlled hydration rate of C3S and formation of C-S-H. It should also be noted that
compressive strength was measured using relatively small cylinders in this experimental
program. As hydration heat can be better preserved in larger volume, the compressive
strength development of Mix 6 when processed in a larger volume should be between the
non-insulated and insulated lab curing conditions.
Mix 6 remained workable for 20 minutes, which was considered sufficient for
repair jobs. For larger-scale mixing that requires a longer working time, a hydration
control agent can be used to retard the hydration process and retain workability. Delvo
stabilizer from Master Builders has been tested to work well for this purpose; the details
of processing optimization for larger-scale mixing are described in Chapter 8.
As a conclusion, Mix 6 was initially selected as the mix proportion that satisfied
the early-age and late-age compressive strength and workability requirements. Further

89

evolution of this mix is necessitated by tensile property demands, as described in the


following sections.

2.3.5

Microstructure Tailoring for Tensile Ductility

2.3.5.1 Initial Composite Testing in Uniaxial Tension


With Mix 6 satisfying target compressive strength requirements, initial tensile
testing was conducted to evaluate its tensile properties at different material ages.

The

direct uniaxial tensile test is considered the most convincing method for evaluating
material strain-hardening behavior66 because some quasi-brittle fiber reinforced concrete
can also show apparent hardening behavior under flexural loading - a phenomenon
known as "deflection hardening". Figure 2.12 shows the uniaxial tensile test setup and
Figure 2.13 illustrates the specimen dimensions. The plate specimens are 228.6 mm (9
in.) long, 76.2 mm (3 in.) wide and 12.7 mm (0.5 in.) thick.

Before testing, four

aluminum plates were glued to the four edges of the plate specimen to facilitate gripping.
Tests were conducted on an MTS machine with 25 KN (5.62 kip) capacity, under a
displacement control with rate of 0.005 mm/s (1.97><10~4 in./s). Two external LVDTs
(Linear Variable Displacement Transducers) were attached to the specimen surface, with
a gage length of 101.6 mm (4 in.), to measure the displacement. The mixing procedure
was the same as described in Section 2.3.4. Specimens were demolded at 4 hours, cured
in air at room temperature (203C), and tested at different ages from 4 hours to 28 days.
The tensile stress-strain curves of Mix 6 at ages of 4h, 24h, 3d, 7d, and 28d are
shown in Figure 2.14. The age dependencies of the tensile strain capacity and tensile
strength of Mix 6 are shown in Figure 2.15 and 2.16. Three specimens were tested for

90

each age. Mix 6 exhibited a rapid strain capacity decrease with age, accompanied by an
increase in first cracking strength. The tensile strain capacity was around 4% at 4h,
decreased to 3% at 24h, and further decreased to less than 1% after 3d. This reduction in
tensile strain capacity at later ages is unfavorable for durable performance of this repair
material when undergoing large deformation due to shrinkage, thermal effects, structural
loading, and expansion of corroded reinforcement, under restrained conditions.
Therefore, Mix 6 needs further tailoring to maintain tensile ductility at all ages.

2.3.5.2 Microstructure Tailoring of Mix 6 for Tensile Ductility


To overcome the gradual loss of tensile ductility in Mix 6, it is necessary to
investigate the fundamental micromechanisms responsible for this undesirable composite
behavior, and to provide a pathway to minimize this tendency. Toward these goals, the
age dependency of micromechanical parameters, including matrix and fiber/matrix
interface properties, were measured at different ages and used to guide the tailoring
process. Based on the conditions for strain-hardening and saturated multiple cracking, it
is evident that high tensile strain capacity requires a high Jb'/Jtip ratio and sufficient
number of pre-exsiting flaws larger than cmc. The matrix toughness JtjP and flaw size
distribution are matrix properties, while the complementary energy Jt,' is mainly
controlled by fiber and interface properties.
To measure the age dependency of JtjP, matrix toughness tests were conducted on
the Mix 6 matrix (without fibers) at different ages. This test was similar to ASTM E399
"Standard Test Method for Plane-Strain Fracture Toughness of Metallic Materials". The
ASTM E399 allows one to use different geometry specimens, such as bending specimens

91

and compact tension specimens, to measure the Km value. The Mix 6 matrix was
prepared and cured as described in Section 2.3.4, except that fibers were not added. The
fresh mix was cast into notched beam specimens measuring 305 mm (12 in.) in length, 76
mm (3 in.) in height, and 38 mm (1.5 in.) in thickness. The matrix fracture toughness Km
at different ages was measured by the three point bending test, as shown in Figure 2.17.
The span of support is 254 mm (10 in.) and the notch depth to height ratio is 0.4. Three
specimens were tested for each test series. JtiPwas calculated from the measured K m by
use of Equations 2.1 and 2.2.
To calculate the age dependency of Jb' of Mix 6, single fiber pullout tests were
conducted to measure three important interfacial parameters at different ages: chemical
bond strength Gd, frictional bond strength To, and slip hardening coefficient |3. As shown
in Figure 2.18, single fiber pull-out tests were conducted on small-scale prismatic
specimens with dimensions of 10 mm x 5 mm x 0.5 mm (0.4 in. x0.2 in. x0.02 in.). A
single fiber was aligned and embedded into the center of a Mix 6 mortar prism with an
embedment length of 0.5 mm (0.02 in.). Three specimens were tested for each test series.
The load P versus displacement 5 curve was obtained through quasi-static testing and
used to determine the interfacial parameters. These interfacial parameters, along with
fiber volume fraction, length and diameter, were then used to calculate the fiber bridging
law a(5). (More details on o(5) are given below). The snubbing coefficient is closely
related to fiber surface frictional coefficient, which decreases significantly with increase
of oiling coating content65. The snubbing coefficient f = 0.2 was assumed in this study
for PVA fiber with 1.2% oiling coating, and decreases significantly with increase of
oiling coating content. The fiber strength reduction coefficient f = 0.33 was measured

92

by single fiber in-situ strength test67. The resulting complimentary energy J b calculated
from the a(5) curve combined with the matrix fracture toughness JtiP obtained from the
Km measurement were used as inputs to evaluate composite material behavior (i.e. strainhardening or tension-softening) and to calculate the PSH index as defined in Equation 2.4.
The general profile of a single fiber pullout curve (P vs. 5) can be decomposed
into three major regimes as shown in Figure 2.19 . Initially, a stable fiber debonding
process occurs along the fiber/matrix interface (Figure 2.17 a). While the load carried by
the fiber increases up to P a , the fiber embedded end, 1 = le, does not become displaced.
The debond length, Id, increases towards Id = le. During this "debonding" stage, the
displacement of the fiber end is the sum of the elastic stretching of the debonded fiber
segment and of the fiber free length (the portion of the fiber outside the matrix). This
debonding process results in a tunnel-like crack that propagates stably from the free end
towards the embedded end of the fiber. This tunneling process is stable until the tunnel
crack tip approaches the embedded end of the fiber at which stage it loses stability and
the load suddenly drops from P a to Pb. At this moment the fiber is held in the matrix only
by friction. The chemical debonding energy value, Gd, is calculated from the Pa to Pb
difference using Equation 2.5:

n Efdf
where Ef is the fiber Young's modulus, and df is the fiber diameter.
At the point Pb, the embedded fiber end is just debonded and the frictional bond
strength xoat the onset of fiber slippage can be caculated using Equation 2.6:
Ph

*o = - 7 y

(2-6)

93

During the fiber slippage stage, the fiber can undergo sliding with slip hardening,
constant friction, or slip-softening, characterized by the positive, zero or negative
coefficient (3, respectively69. Slip-hardening occurs often with polymer fibers including
PVA fibers used in this study. Because polymer fibers are less hard than the surrounding
matrix, they tend to be damaged during pullout process and therefore a jamming effect
can take place inside the matrix, leading to an increasing load. Conversely, constant
friction or slip-softening are often observed when the fiber hardness is higher than that of
the surrounding matrix. The |3 value is calculated from the initial (5" approaching 0)
slope of the P versus S' curve using:
P = (df /lf)[l/T0JTdf)(AP /AS')|S,_>0 +1]

(2.7)

Lin et al.69 proposed the formulation of the crack bridging stress versus opening
relationship a(5) based on single fiber pullout P(5) relation
4V,
0B(5) = f f ' (

P(5)ef*p(4>)p(z)dzd<l>

(2.8)

where Vf is the fiber volume fraction, df is the fiber diameter, is the orientation angle of
the fiber relative to a crack plane, Lf is the fiber length, z is the centroidal distance of a
fiber from the crack plane, f is a snubbing coefficient, and p((j)) and p(z) are probability
density functions of the fiber orientation angle and centroidal distance from the crack
plane, respectively. P(6) is the pullout load versus displacement relation of a single fiber
aligned normal to the crack plane, as experimentally measured in this study from the
single fiber pullout test. It should be noted that during fiber debonding and pullout stages
(Figure 2.17), the fiber may rupture if the load P exceeds the fiber strength, Of, which is
often the case in the PVA-ECC system when the interfacial chemical bond and frictional
bond are strong.

94

Figure 2.20 presents the age dependency of matrix fracture toughness Km


determined from the notched beam bending test. The development of Km at early age is
fastest during the first 24 hours, and the Km-age curve becomes relatively flat after 3 days.
Figures 2.21, 2.22, and 2.23 present the age dependency of Gd, To and (3 measured
from the single fiber pullout test. Despite the scattering of data, the trend clearly shows
that the development of interface chemical bond Gd and frictional bond to remained
relatively slow during the first 24 hours, accelerated considerably after 24 hours, and
finally reached a plateau after 7 days.
Compared with Km, the development of interface bonding strength was much
lower at the early age until about 3 days. The different development rates between
fracture toughness and interfacial properties can be due to the accelerated hydration at
early ages in the cementitious matrix, and development of the fiber/matrix interfacial
bond is less affected by the accelerated hydration process. These findings were different
from the age-dependent development of interfacial properties of polyethylene (PE) fibers,
which are hydrophobic, in a cementitious matrix70. It was found that the interfacial bond
strength of PE fibers matured much faster, in less than 7 days, in comparison with the
development of matrix bulk properties that typically took 14-28 days. The difference
between the present study and Chan and Li70 can be attributed to the adoption of a rapidstrength-gain binder system and different type of fibers (PVA fibers instead of PE fibers)
in Mix 6.
It was the difference between the development rates of matrix toughness and
matrix/fiber interface properties that resulted in the quick change of the pseudo strainhardening (PSH) performance index, defined by the Jb'/Jtip ratio. PVA fibers form strong

95

chemical and frictional bonds with the cementitious matrix because of their hydrophilic
nature.

Moreover, PVA fiber has strong slip-hardening behavior during the pullout

process.

Therefore, the fiber often ruptures instead of being pulled out from the

surrounding matrix, leading to a reduction in complementary energy V . For Mix 6, the


rapid increase in Gd, To and (3 after 24 hours led to more occurances of fiber rupture than
pullout, resulting in a significant decrease in V (Figure 2.24). This was accompanied by
an increasing JtiP (Figure 2.24) corresponding to increased matrix toughness Km. As
shown in Figure 2.24, JbVJtiP dropped from 39 at 4 hours to 5 at 24 hours, then to around
1 at 3 days, and flattened after 3 days, leading to the deteriorated composite tensile strain
capacity. Tensile strain capacity of Mix 6 at different ages is also plotted in Figure 2.24.
Strong correlation was found between the age dependent development of tensile strain
capacity and the JbVJtiP ratio (PSH index).
The low tensile strain capacity of Mix 6 after 24 hours is due to a low value of the
Jb'/Jtip ratio, resulting in unsaturated multiple cracking. When the margin between V and
Jtip is small, it becomes critical to control pre-existing flaw distribution in the matrix to
retain high strain capacity.

As discussed in Section 2.2.3, a large number of flaws

slightly larger than cmc are preferred for saturated multiple cracking and high tensile
strain capacity. The critcal flaw size cmc is set such that only those flaws larger than cmc
can be activated before reaching Oo and contribute to multiple cracking. The approach
used in this study is to superimpose a prescribed flaw system to the pre-existing flaws in
the Mix 6 matrix, as illustrated in Figure 2.25. The artifical flaws can be any inclusion
with a weak bond to the cementitious matrix or with relatively low tensile strength
compared to the tensile strength of the surrounding matrix, such as lightweight

96

aggregates, air bubbles, polymer foams and rubber powder. In this study, Polystyrene
(PS) beads (Figure 2.26) with a size of 4 mm were added as artificial flaws at a dosage of
5% by volume. The size of 4 mm was chosen because it was slightly larger than the
calculated critical flaw size cmc71, so that the PS beads are large enough to initiate
cracking without significantly lowering the first cracking strength and interrupting fiber
dispersion. Possessing a cylindrical shape with sharp edges and extremely weak bond
with the surrounding cement binder, the PS beads serve as crack initiators. The modified
mixing proportion - Mix 7 - is the same as Mix 6 except for the presence of PS beads
(Table 2.8).
Figure 2.27 shows the tensile stress-strain curves of Mix 7 at the ages of 4h, 24h,
3d, 7d, 28d and 50d. Significant tensile strain-hardening behavior with strain capacity
above 3% was observed at all ages up to 50d. Figure 2.28 compares the tensile strain
capacity of Mix 6 and Mix 7 at different ages. Greatly improved tensile strain capacity
was found in Mix 7 at all ages, especially at ages after 24h, compared to Mix 6.
Moreover, introduction of 5 vol% artificial flaws does not seem to affect the first
cracking strength and ultimate tensile strength. This is because of the small dosage, and
the similarity of the artificial flaw size and pre-existing flaw size in Mix 6.

The

introduction of artificial flaws with prescribed size was shown to be an effective


approach to improve tensile strain capacity through increased saturatation of multiple
cracking.
The effect of adding PS beads on ECC compressive strength was investigated and
shown in Figure 2.27. Comparing the compressive strength development curves for Mix
6 and 7, this figure shows the minor effect of PS bead presence on the early compressive

97

strength gain rate, although inclusion of the beads greatly improves the multiple cracking
behavior when the specimen is under tension.

Specifically, small differences in the

strength gain rate were observed in the first 6h, while compressive strength after 12h was
reduced by approximately 10%. Even with this reduction, Mix 7 still satisfied the
compressive strength requirements at early and late ages as well as the workability
requirement. Up to now, HES-ECC (Mix 7) has been successfully developed with the
target early-age compresseive strength, workability, and high tensile ductility at all ages.
Its mechanical properties will be characterized in detail in the next section.

2.3.6

HES-ECC Material Characterization


In this section, the mechanical properties of the newly developed HES-ECC

material, including tensile, compressive, and flexural properties, were tested with a larger
number of specimens and documented in detail.
Materials
The mix proportion of the newly developed HES-ECC is shown in Table 2.9. It
contains type III portland cement, fine silica sand Fl 10, water, PVA fibers, PS beads,
superplasticizer, accelerating admixture, and hydration control admixture (when a longer
working time is needed). Mixes were prepared with a Hobart-type mixer with 13 liter
capacity according to following mixing procedure:
1) Mix type III Portland cement and silica sand for approximately 1 minute;
2) Add water slowly and continue mixing for 1-2 minutes;
3) Add superplasticizer; continue mixing for 1-2 minutes until a consistent mixture is
reached;

98

4) Add PVA fibers slowly and mix for 2 minutes until fibers are well distributed;
5) Add polystyrene beads and mix for 1-2 minutes;
6) Add hydration control admixture and mix for 1 minute unless casting will be
completed within 15 minutes of completion of mixing;
7) Add accelerating admixture and mix for 1 minute before casting into molds.
The whole mixing procedure for each batch took 10-15 minutes. After mixing,
the mixture was cast into tensile, compressive and flexural specimen molds with
moderate vibration applied. The molds were then covered with plastic sheets, demolded
in 4 hours, cured in air at room temperature (203C) and relative humidity (RH 405%),
and tested at different ages.

Uniaxial Tensile Test and Results


Age-dependent uniaxial tension tests were conducted to measure the tensile
strength, Young's modulus, tensile strain capacity, crack width and crack pattern of HESECC. Testing procedure and specimen preparation were the same as described in Section
2.3.5. Specimens were tested at ages of 4, 6, 12 and 24 hours, and 3, 7, 14, 28 and 60
days, measured as the time between the end of casting and start of testing. Ten to
fourteen specimens each were tested at each age.
Under uniaxial tensile loading, HES-ECC specimens exhibit significant tensile
strain-hardening behavior. HES-ECC first undergoes "elastic straining" until the first
microcrack appears.

Then, the material undergoes "strain-hardening" similar to the

"plastic yielding" in metals. The strain-hardening stage is accompanied by the formation


of many closely-spaced microcracks with controlled crack width.

These cracks are

unlike the localized fracture in concrete or other brittle materials as they continue to carry

99

increasing stress across the crack faces after formation.

The multiple microcracks

eventually "saturate" (reach the minimum spacing) along the specimen and a localized
fracture occurs due to exhaustion of the local fiber bridging capacity. At this moment,
the load starts dropping and the HES-ECC enters the "tension softening" stage until final
failure.

The peak stress prior to the tension-softening stage is defined as the tensile

strength of the HES-ECC.


Table 2.10 and Figures 2.30, 2.31, 2.32 and 2.33 summarize the uniaxial tensile
test results of HES-ECC. As shown in Figure 2.30, The tensile strength of HES-ECC
increases rapidly during the first 24 hours, from 3.5 0.1 MPa (501 12 psi) at 4 hours
to 4.7 0.1 MPa (680 1 2 psi) at 24 hours. It continues increasing at a lesser rate during
later ages, and reaches 5.8 0.2 MPa (840 28 psi) at 60 days.
The Young's modulus (E) of HES-ECC is determined from its tensile stress-strain
curve, defined as the slope of the line drawn between the starting point and the point
corresponding to a strain of 0.015%. The 0.015% tensile strain is an approximate elastic
strain capacity of ECC material32.

The age dependency of the HES-ECC Young's

modulus is plotted in Figure 2.31. Young's modulus of HES-ECC increases with age,
from 13.1 0.8 GPa (1900 111 psi) at 4 hour to 23.8 0.7 GPa (3452 103 psi) at 60
days.

The Young's modulus of HES-ECC is generally lower than that of concrete

materials, due to the absence of coarse aggregates in its mix.

This lower Young's

modulus is in fact desirable for a concrete repair material because it lowers the induced
tensile stress build-up caused by shrinkage restrained by the existing surrounding
concrete ' ' , thereby reducing the tendency toward cracking and repair/substrate
material interface delamination.

100

The tensile strain capacity of each specimen was defined as the tensile strain
value corresponding to tensile strength (prior to tension softening). Figure 2.32 shows
the age-dependent development of HES-ECC tensile strain capacity that exceeds 5% at 4
hours, decreases to 4% at 3 days, and stablizes to an average of 3.5% beyond 7 days (test
data extend up to 60 days). This long-term high tensile strain capacity is several hundred
times larger than the 0.01% assumed for normal concrete materials.
The crack width of HES-ECC microcracks at the strain-hardening stage was
measured because it determines the HES-ECC's resistance to the penetration of
aggressive agents. Research has shown that the narrow crack width of ECC reduces
water permeability75 and chloride diffusion penetration76; these issues will be discussed in
detail in Chapter 6. Figure 2.33 shows the multiple microcracking behavior and crack
width of HES-ECC at the ages of 4 hours, 24 hours, 3 days and 28 days. To enhance
resolution of these microcrack images, epoxy glue was applied to the surfaces of the
specimens before the photos were taken. Near-saturated multiple microcracks with crack
width below 60 ^m were observed at all ages. It should be noted that the crack width at
the very early age of 4h is as low as 10 ^m while the tensile strain capacity is as high as
6%. This high tensile ductility and very tight crack width at early age of the repair
material are crucial for the repair's resistance to early-age cracking due to restrained
shrinkage and chloride penetration, which start as soon as the repair is re-opened to
service.

Uniaxial Compressive Test and Results


Age dependent tests in uniaxial compression were conducted to establish a
database of HES-ECC compressive strength development as a function of material age.

101

Compressive testing was conducted according to the procedure described in Section 2.3.4.
Specimens were tested at ages of 4, 6, 12 and 24 hours, and 3, 7, 14, 28 and 60 days. Ten
to fourteen specimens each were tested at each age.
Table 2.11 and Figure 2.34 summarize the compressive strength of HES-ECC at
different ages. HES-ECC reaches compressive strengths of 23.6 1.4 MPa (3422 203
psi) at 4 hours, 34.2 1.4 MPa (4963 182 psi) at 6 hours, and 42.3 1.4 MPa (6130
202 psi) at 24 hours. The increasing trend slows down at later ages, leveling off at 55.6
2.2 MPa (8063 315 psi) at 28 days, and 56.8 1.7 MPa (8233 250 psi) at 60 days.
The compressive strength development of HES-ECC is shown in Figure 2.35 along with
the target requirements, and requirements given by NJSDOT 199732, MDOT QA/QC
200340, Zia et al. 199142 and FHWA 199943.

Flexural Test and Results


A minimum flexural strength of the repair material, apart from its compressive
strength, is usually specified as a requirement for reopening a roadway to traffic. NJDOT
specified a target flexural strength of 350 psi (2.4 MPa) in 6 hours for the "Fast-track mix"
developed in mid-90's 32 . Caltran requires a minimum flexural strength of 400 psi (2.8
MPa) prior to opening to traffic for full depth pavement repairs33. FHWA 43 recommends
a minimum flexural strength of 450 psi (3.1 MPa) for Rapid-setting Cementitious
Concretes.
For high early strength concrete, the flexural strength requirements are more
stringent than the compressive strength requirements because of the material's brittle
nature. In contrast, ECC materials feature very high flexural strength, or modulus of

102

rupture (MOR), because of its high ductility and strain-hardening behavior .

The

flexural strength of ECC normally approaches three times the flexural strength of its
matrix. It was expected, and verified in this study, that once the compressive strength
and tensile ductility requirements are met, the flexural strength targets of HES-ECC can
be easily attained.
A flextual test was conducted on HES-ECC prism specimens to measure the
stress-deformation curves and material flexural strength (MOR) at different ages.
Flexural testing procedures followed the ASTM C78 78 standard. Figure 2.36 shows the
three-point bending test set-up and specimen dimensions. Each beam specimen measured
406.4 mm (16 in.) long, 76.2 mm (3 in.) wide and 101.6 mm (4 in.) thick, with a test span
length of 304.8 mm (12 in.). Specimens for flexural testing were demolded at 4 hours,
cured in air, and tested at different ages from 4 hours to 60 days.

The air-curing

condition was different from that specified in ASTM C78 because it was intended to
simulate the curing condition in field applications, where the repair was expected to be
exposed to air and open to traffic in 4 hours. The three-point bending test was conducted
with an MTS 810 machine under displacement control at a loading rate of 0.02 mm/s
(7.87x10"4 in./s). Five specimens were tested at each age.
Table 2.12 summarizes the flexural strength (MOR) and deflection at failure of
HES-ECC. Figure 2.37 shows the typical flexural stress versus deflection curves of
HES-ECC at ages of 4 hours, 24 hours, and 28 days. Significant deflection hardening is
observed at these ages. The flexural strength (MOR) of HES-ECC increases rapidly
during the first 24 hours from 9.8 0.2 MPa (1422 34 psi) to 11.4 0.2 MPa (1656
30 psi). This increasing trend slows down at later ages and reaches 15.1 0.4 MPa (2188

103

50 psi) at 28 days. At the same time, specimen flexural ductility (defined as the
maximum deflection when local failure starts) decreases after 6 hours, but approaches a
constant after 3 days. It should be noted that HES-ECC attains a flexural strength of 9.8
0.3 MPa (1422 34 psi) at 4 hours, which is significantly higher than the requirements
of NJDOT, Caltranand FHWA. The 28-day flexural strength of HES-ECC is 15.1 0.3
MPa (2188 50 psi), more than double the flexural strength of normal concrete with
similar compressive strength.

2.4

Conclusions
Material engineering for durable repair materials is possible through the

application of micromechanical tools developed for ECC materials.

This has been

demonstrated through the development of the high early strength and ductile repair
material HES-ECC. Micromechanics can be used as a tool for designing HES-ECC
composites that meet the two strain-hardening criteria as well as conditions for saturated
multiple cracking to achieve large tensile ductility above 3% and self-controlled crack
widths below 100 u.m at both early and late ages. These properties are crucial for repair
durability, as they prevent restrained volume change or stress concentration induced
cracking and penetration of aggressive agents, which are typical deterioration causes in
concrete repairs.
Through strategic combination of type III Portland cement, low water/cement
ratio, polycarboxylate based superplasticizer and calcium nitrate based accelerator, the
newly developed HES-ECC material is capable of attaining compressive strength of 23.6
1.4 MPa (3422 203 psi) in 4 hours and 55.6 2.2 MPa (8062 315 psi) in 28 days,

104

which exceeds HES-ECC design targets as well as federal and state transportation agency
requirments. HES-ECC attains flexural strength of 9.8 0.2 MPa (1422 34 psi) in 4
hours and 15 0.3 MPa (2187 50 psi) in 28 days, which exceeds twice the flexural
strength of concrete with similar compressive strength. Its tensile strength is 3.5 0 . 1
MPa (501 12 psi) at the age of 4 hours and 5.7 0.2 MPa (823 20 psi) at age of 28
days.
The age dependencies of micro-parameters and their effects on the age
dependency of composite tensile strain capacity were investigated using experimental
methods and micromechanics-based analytical tools.

The significant increase of

interfacial properties To, Gd and (3 after 24 hours led to more fiber ruptures during the
pullout process, resulting in a diminishing complementary energy Jt,'. Meanwhile, the
matrix fracture toughness increased at a faster pace within the first 24 hours and achieved
a relatively high value after 24 hours. The micromechanics model revealed that the quick
deterioration in composite tensile strain capacity after 24 hours was attributed to the rapid
drop in complementary energy V and continuous rise of J tiP> which leads to a significant
reduction in the PSH index measured by JbVJtip.
When there is a small margin of Jb'/JtiP, it is important to have large amounts of
flaws in the matrix with sizes above cmc, which can be activated during material strainhardening stage before the maximum fiber bridging stress is reached. Introduction of
artificial flaws (PS beads) with prescribed size distribution proved to be an effective
approach to achieve saturated multiple cracking, therefore retaining large tensile strain
capacity at late ages. The presence of small volume fraction (e.g. 5%) of PS beads did

105

not impair the early age compressive strength gain rate before 24 hours, although it did
decrease the compressive strength at late ages by approximately 10%.
Detailed characterization of mechanical properties of HES-ECC further validated
the material performance, and provided a database of the tensile, compressive, and
flexural properties of the newly developed repair material for future engineering
applications.

106

Table 2.1 - Three phases of ECC microstructure and corresponding micro-parameters.


Phase

Micro-Parameters

Fiber

Length Lf, Diameter df, Volume Fraction Vf


Tensile Strength Of, Elastic Modulus Ef

Matrix

Fracture Toughness Km, Elastic Modulus Em, Initial


Flaw Size Distribution ao

Fiber/Matrix Interface

Chemical Bond Gd , Frictional Bond to, Slip Hardening


Property (3, Snubbing Coefficient f, Fiber Strength
Reduction factor f

107

Table 2.2 - List of materials used.


Product
Name

Manufacturer

Portland
Cement
Type I

LaFarge Co.,
USA

Portland
Cement
Type III

Holcim Co.,
USA

Finely ground
OPC

Blaine Surface area:


5000 cm2/g

Normal
Fly Ash

Boral Material
Tech. Inc.,
USA

ASTMC618
Class F

Mean size:
10-20 um

Pozzolith
NC534

Master
Builders,
Inc., USA

Calcium nitrate
<43%,
Sodium
thiocyanate
<5%

Silica
sand Fl 10

U.S. Silica Co.

Silica

PS Beads

Dow Chemical
Co., USA

Polystyrene

Superplasticizer

Glenium
3200 HES

Master
Builders,
Inc. USA

Polycarboxylate
Based

Hydration
Control

Delvo
Stabilizer

Master
Builders,
Inc. USA

Fiber

REC15

Kuraray, Co,
Japan

Category

Cementitious
Material

Accelerating
Admixture

Aggregate

108

Chemical
Composition
Ordinary
Portland
Cement (OPC)

Polyvinyl
Alcohol (PVA)
Fiber

Particle Size
Blaine Surface area:
3300 cm2/g

Specific weight:
1.39g/ml

Mean size: 110 urn,


Max. size: 250 um
Size: 4mm, Density:
1.4g/cm3
Specific weight:
1.05 g/ml

Table 2.3 - Mixing proportions for initial material composition selection based on
strength requirements.
Cement

Mix

Silica

Fly

Sand

Ash

Water

Admixtures

Fiber,
V f (%)

Type I 1.0

0.8

1.2

0.53

GL3200 0.75%

2.0

Type III 1.0

0.8

1.2

0.53

GL3200 0.75%

2.0

Type III 1.0

1.0

0.53

2.0

Type III 1.0

1.0

0.45

GL3200 0.35%

2.0

Type III 1.0

1.0

0.33

GL3200 0.75%

2.0

Type III 1.0

1.0

0.33

GL3200 0.75%,
NC534 4.0%

2.0

Table 2.4 - Properties of REC 15 PVA fiber from Kuraray, Co, Japan.
Fiber Type

REC15

Nominal
Strength
(MPa)

Apparent
Strength
(MPa)

Fiber
Diameter

1620

1092

(^im)

Fiber
Length
(mm)

Young's
Modulus
(GPa)

39

12

42.8

109

Elongation
(%)

6.0

Table 2.5 - Compressive strength of Mix 1 and 2 at different ages.


Age
(Hour)
4
5
6
10
12
24
48
72
144
672

17.8
22.0
31.1
40.9
55.9
62.7

Compressive Strength (MPa)


Mixl
Mix 2
3.4
3.4
7.1
8.2
17.3
18.8
24.3
25.5
17.9
18.3
28.8
30.4
35.4
22.6
24.4
38.5
32.3
32.6
41.6
42.6
44.3
44.5
46.9
46.8
57.9
60.2
55.1
56.6
65.5
59.7
66.7
62.0

3.6
9.4
19.1
28.3
32.3
38.6
44.8
48.4
57.7
65.7

Table 2.6 - Compressive strength of Mix 3, 4 and 5 at different ages.


Age
(Hour)
4
5
6
7
8
9
10
12
24
48
72
96
120
144

1.4
3.2
4.0
9.7
16.4
24.8
24.8
25.7
27.6
27.1

168
336
480
672

27.7
28.3
29.2
29.0

Mix 3

1.9
3.4
4.7
10.9
17.7
25.5
26.7
27.0
28.6
29.1
29.4
29.7
30.3
30.3

Compressive Strength (MPa)


Mix 4
1.8
2.0
4.7
1.9
5.2
5.4
5.2
8.9
5.9
5.9
6.3
23.1
9.7
12.8
30.3
11.2
2.5
14.0
15.2
17.1
32.8
4.1
18.6
20.0
20.0
33.1
5.5
18.8
20.0
21.1
37.3
21.4
12.4
22.3
22.8
36.7
29.3
30.8
19.2
30.3
41.9
33.1
34.1
31.8
26.9
47.9
28.1
33.1
34.2
34.7
52.6
36.4
37.7
28.0
35.6
53.7
37.2
28.7
38.5
39.9
56.3
43.0
43.5
30.6
40.1
55.8
29.8
42.7
43.7
45.1
56.9
47.2
30.7
46.1
48.0
59.9
48.0
49.3
29.6
49.6
62.3
47.6
49.0
31.8
50.6
62.9

no

Mix 5
5.0
10.2
26.1
31.1
34.0
36.0
37.7
38.8
43.5
50.2
55.3
55.2
57.4
57.9
58.7
62.3
64.3
65.4

5.3
11.7
26.4
31.6
35.7
38.3
39.0
41.0
44.5
51.9
56.2
56.4
58.2
60.0
61.4
63.8
64.5
66.7

Table 2.7 - Compressive strength of Mix 6 (w/ and w/o insulation) at different ages.
Age
(Hour)
2.7
3
4
5
6
7
8
9
10
12
24
48
72
96
120
144
168
336
480
672

Compressive Strength (MPa)


Mix 6 (w/ insu ation) Mix6( w/o insulation)
12.4
12.5
12.6
21.3
7.6
7.7
8.7
22.0
22.7
17.4
17.9
18.0
26.6
28.1
29.3
28.6
29.7
31.7
34.3
34.6
34.6
33.0
34.3
34.7
38.3
40.1
41.0
37.4
39.5
34.8
36.4
40.8
41.7
40.4
37.1
39.8
41.5
42.5
38.6
39.3
40.5
41.8
44.5
40.2
43.3
39.5
41.1
42.5
45.6
40.0
43.9
44.8
39.9
42.3
44.8
45.1
41.1
45.7
44.5
45.7
47.5
48.0
50.5
47.6
50.0
51.0
53.2
49.3
51.8
53.2
51.8
54.2
56.4
53.0
57.8
52.9
54.1
54.7
55.9
56.6
58.3
54.4
56.2
55.2
56.0
57.2
57.4
54.3
56.5
55.9
56.9
58.6
59.0
57.2
57.0
58.1
58.9
58.5
58.9
59.8
60.7
61.0
59.5
60.0
60.6
60.7
62.2
63.4
62.7
63.2
64.0
62.4
61.6
63.0
64.4
63.7
65.3

Table 2.8 - Mixing proportion of HES-ECC before / after adding artificial flaws.
Mix

Cement

Silica
Sand

Water

Other
Aggregates

Type III, 1.0

1.0

0.33

Type III, 1.0

1.0

0.33

PS Beads
0.064

111

Admixtures
(wt% of cement)
GL3200 0.75%,
NC534 4.0%
GL3200 0.75%,
NC534 4.0%

Fiber Vf
(%)
2.0
2.0

Table 2.9 - Mix proportion of HES-ECC.


HES-ECC Mix Design Parameter

Value (kg/m3)

Portland Cement, Type III

918

Silica sand, Fl 10

918

Water

301

Poly-vinyl-alcohol Fiber

26.1

Polystyrene Beads

58.8

Superplasticizer, Glenium 3200HES

6.89

Accelerating Admixture,
36.7

Pozzolith NC 534
Hydrating Control Admixture,

Optional

Delvo Stabilizer

Table 2.10 - HES-ECC tensile properties at different ages (mean standard deviation).
Age
4h
6h
12 h
24 h
3d
7d
14 d
28 d
60 d

Young's Modulus,
xlO3 MPa (ksi)
13.10 0.77
(1900.12 111.44)
14.98 0.79
(2173.37 114.11)
16.05 1.02
(2328.26 147.29)
18.30 0.59
(2654.19 84.87)
19.100.81
(2770.22 116.91)
20.59 0.70
(2986.32 100.90)
22.18 1.02
(3216.27 147.47)
23.20 0.96
(3364.88 139.77)
23.80 0.71
(3451.90 103.45)

Tensile Strength,
MPa (psi)
3.46 0.08
(501.23 11.64)
4.21 0.13
(610.41 19.42)
4.57 0.17
(662.27 25.18)
4.69 0.08
(680.35 11.50)
5.09 0.16
(738.32 23.62)
5.56 rb0.ll
(806.82 16.42)
5.880.16
(852.42 23.01)
5.98 0.14
(867.33 20.39)
5.99 0.19
(868.78 28.11)
112

Strain Capacity,
%

5.97 0.22
4.97 0.38
4.41 0.33
3.99 0.27
3.61 0.28
3.52 0.29
3.64 0.37
3.47 0.62
3.52 0.57

Table 2 . 1 1 - HES-ECC compressive properties at different ages


(mean standard deviation).
Age (h)

Compressive Strength,
MPa (psi)

3h
4h
6h
12 h
24 h
3d
7d
14 d
28 d
60 d

9.82 0.56
(1424.82 81.54)
23.59 1.40
(3422.16 203.33)
34.22 1.40
(4963.22 181.51)
37.01 1.84
(5367.54 266.43)
42.26 1.39
(6129.62 201.55)
44.70 2.37
(6482.65 343.99)
47.47 1.89
(6884.56 274.83)
50.79 2.38
(7366.43 344.71)
55.59 2.17
(8062.90 315.03)
56.76 1.72
(8233.13 250.08)

Table 2.12 - HES-ECC flexural properties at different ages (mean standard deviation).
Age(h)
4h
6h
24 h
3d
7d
28 d

Flexural Strength / MOR,


MPa (psi)

Deflection at Failure,
mm (in)

9.81 0.24
(1422.19 34.22)
11.02 0.36
(1598.67 51.94)
11.41 0.21
(1655.54 29.78)
12.96 0.35
(1879.63 50.79)
13.55 0.43
(1965.28 62.61)
15.08 0.34
(2187.64 49.87)

14.73 0.50
(0.58 0.02)
15.49 0.50
(0.61 0.02)
12.45 0.50
(0.49 0.02)
10.67 0.50
(0.42 0.02)
9.91 0.30
(0.39 0.01)
9.91 0.50
(0.39 0.02)
113

High Performance Fiber

Reinforced Ccmeritious Composites


(HPFRCC)

wore
Figure 2.1 - Uniaxial tensile stress-deformation relation of concrete, FRC, and HPFRCC.
For concrete, formation of the crack corresponds to the sudden drop of tensile load
carrying capacity. For FRC, after a crack is formed, deformation is associated with crack
opening 5. Tensile load capacity drops at a slower rate (compared to concrete) as a single
crack enlarges. This is a typical "tension softening" behavior.
For HPFRCC,
deformation during the elastic and strain-hardening stages is described as straining. After
the first crack forms, tensile load capacity continues to rise with increasing strain through
multiple micro-cracking behavior. The strain capacity of HPFRCC is defined as the
strain value at which peak tensile load is reached.

114

mm - c m

Composite Tensile
Strain-Hardening
Steadv-Stat
urn - mm

Fiber Bridging Properties


Across uracKS
Vlicromechanics

Fiber
Matt ix

inttiifsce

Figure 2.2 - Scale linking in ECC material development.

#0

6SS

Sc

Figure 2.3 - Typical fiber bridging stress vs. crack opening o(8) curve for a tensile strainhardening composite. Hatched area represents complimentary energy Jb- Shaded area
represents crack tip toughness Jtip.

115

ttttmtmtr tttttttttttttl
fco Qs.v

<-<-<:

0-o ( I )

yyvfl.,

wwuuuuu uuuuwu*
(b)

(a)

Figure 2.4 - (a) Steady-state cracking with a constant crack opening Sss accompanied by a
constant ambient load oss; (b) Griffith cracking with a widening crack opening 5m
accompanied by a descending ambient load.

fiber bridging strength

cracking strength

Initial Flaw Size (mm)


Figure 2.5 - Theoretically predicted effect of initial flaw size on the cracking strength of
PVA-ECC containing 2% by volume PVA fiber. The maximum fiber bridging strength is
also shown in the plot. The critical flaw size (cmc) seperates inert and active flaws.

116

9000
8000
7000
6000
5000
4000
3000
2000
1000
0

B1

D NJSDOT, 1997
O MDOT QA/QC, 2003
A Ziaetal., 1991
* FHWA.1999
-Target

o
_!

I....I-I.J..U,,,,

'

'

10

'

'

100

1000

Age (hour)

2.6 - State and federal compressive strenght requirements, and target compressive
strength at different ages for HES-ECC.

#^

<

~(>.2mm
(*in.l

i.
fa

Figure 2.7 - Uniaxial compressive test setup.

117

* * *

70
TO
Q.

-Mix 1: Type I Cement

60 --

-Mix 2: Type III Cement

50 -C

40
>
v>
V
Q.

E
o
u

30
20
10 +
0
1

10

100

1000

Age (hour)

Figure 2.8 - Effect of cement type on compressive strength development.

10

100

1000

Age (hour)
Figure 2.9 - Effect of water/cement ratio on compressive strength development.

118

70
Q.

- Mix 5: w/o Accelerator

60 -

-Mix 6: w/Accelerator

%
en 50
2

40

SJ 30
"w
8 20

Io 10
0

-J

10

1000

100
Age (hour)

Figure 2.10 - Effect of accelerator on compressive strength development.

70
Q-

O
c
a>
a_
+-
</)
<u
>
'en
t/>
a)
i_

a.
E
o

-Mix 6: w/o Insulation

60 +

-Mix 6: w/ Insulation

50
40
30 -f
20
10
-J

1I Il,

10

4100

-J

I J LL.

1000

Age (hour)

Figure 2 . 1 1 - Effect of curing condition on compressive strength development.

119

rt

A*

i*
,. .

;)
i

*-v T

'

-y\
'-

\
,4*.

-lit
" Vr'W',

iTK
w ? " ;*'.'-<

Figure 2.12 - Uniaxial tensile test setup.

LOAD
LVDT

LVDT
Holder

^ T

II

JgJ

Aluminum

12.7 mm (0.5

76.2 mm (3 in)

in)

Figure 2.13 - Dimensions of tensile test specimen.


120

4h

IB

6-

a.
2

S-

V)

s
1_

<w

Tens

0B
-

4V

^m****~~~
^***^*^ ^^^^^v""**

4.lfmq~r*!tf*mr~~******i*&*Z.

1
*

2%

0%

T\
\i~"^\

2n

~\

4%

3%
Strain

5%

7
c?

a.
2

65-

W)

4-

0)

<5>

2-

1-

1-

24 h

me^,

v&Sf

Qrt&P2

%*;

0)

n
3%

1%

4%

5%

6%

Strain

3d

I
!

s
to

c
\a/

2%

3%
Strain

121

4%

6%:

7d
cd
Q.

w
a>
CO

"m
c

h*

!U*

[_

2%

4%

8%

Strain

7
*^
CS

28 d

Q.

w
w
a>

* w *

^O
->
C

jSU

0!

1
0%

1%

2%

3%

4%

5%

6%

Strain

Figure 2.14 - Tensile stress - strain curves of Mix 6 at ages of 4h, 24h, 3d, 7d, and 28d.

122

8%
7% -

I
1

6%
5%
4%
!

3%
2%

1%
0%

.J

_ _ ! 1 MX.

1- 1 1 1 1 I I

10

1000

100

10000

Age (hour)

Figure 2.15 - Age dependency of tensile strain capacity of Mix 6.

^^ 6 (0
a.
5 5 -

"*w-*

x:
*
o> 4
c
0)

fc-

(t) 3 a>

'in ? c
a>
t- 1

10

100
Age (hour)

Figure 2 . 1 6 - Age dependency of tensile strength of Mix 6.

123

1000

76 mm
38 mm

Figure 2.17 - Matrix fracture toughness test set-up.

0.5 mm

10 mm

Figure 2.18 - Single fiber pull-out test set-up.

124

5 mm

$=(>. I..-I -i
.^L,,^

single fiber
pulimtt load
(P)

, *?

<*? &

slip hardening. P>0

. ' .

dcborsdcd length

(6) displacement
*

defending j

whole fiber slippage

I, embedment length

Figure 2.19 - General profile of a single fiber pullout curve.

10

100

1000

Age {hour)
Figure 2.20 - Age dependency of matrix fracture toughness.

125

(mini

,j

5 -+
2.5

2 -

. ^

1 55 --O

/ :

1 -

/ *

6 -

nH

t - T TT T f l T | " T

10

100

1000

Age fhour)

Figure 2.21 - Age dependency of matrix/fiber interface chemical bond.

44
A

CO

a.

24

i
A

1
A

i_

10

100

1000

Age (hour)

Figure 2.22 - Age dependency of matrix/fiber interface frictional stress.

126

10

100

1000

Age (hour)

Figure 2.23: Age dependency of slip hardening coefficient.

127

120

0 -t

10

100

1000

Age (hour)
(a)

Tensile Strain Capacity


n.Jb'/Jtip

I
1

-\

--

'

'

'

i ' ' ' i

'

i i

ii r

(b)
ure 2.24: Evolution of (a) V and JtiP, (b) Jb'/Jtip ratio and tensile strain capacity with
age (Mix 6).

128

P(c)

P(c)

artificial flaws
natural flaws

flaw size c

flaw size c

Figure 2.25 - Optimization scheme for pre-existing flaw size distribution in matrix: (a)
natural flaw size distribution with random nature inherent from processing; (b) artificial
flaws larger than critical size cmc imposed to ensure saturation of multiple-cracking
(Wang and Li, 2004).

t .A

Figure 2.26: Aggregates used as artificial flaws: polystyrene (PS) beads.

129

OJ

t
61

(A

CD

-fe.
f9

00
^3

^*jN

Tensile Stress (MPa)

fit

(A

Ctt

OJ

rr

O - i M d J k O l f f l ^

Tensile Stress (MPa)

Tensile Stress (MPa)

!
k>

a>

Tl
OQ

^1

CO

B'

CO

C/3

CD

en

S3

-+
CO

o-

aCO
^a

-Age

-*

CO

en

03

*5*

Or1-

Tensile Stress (MPa)

5*

Tensile Stress (MPa)

Tensile Stress (MPa)

8%
> 7%
10

a.
(0

Mix 6 (w/o artifical flaws)

6%

trai

O 5%
c

4%

CO

Ten

J) 3%
"S5
2%

1%

0%

100

10

1000

10000

Age (hour)

(a)
C7o

Mix 7 (w/ artifical flaws)

7%
6%

<f 5%
c
" 4%

10

3%
tfl

2% 1%
AO/

I,I

,i

I.J.IJ

10

100

i,

1000

10000

Age (hour)

(b)
ure 2.28 - Age dependency of tensile strain capacity: (a) before adding artificial flaws
(Mix 6); (b) after adding artificial flaws (Mix 7).

132

10000
"as 9000
0L

8000

-Mix 6: w/o PS beads


Mix 7: w/PS beads

* "

7000

o>
1_

6000

5000

res siv

+J

4000

a>

3000

a.
E 2000
o

1000

0
10

100

1000

Age (hour)

Figure 2.29 - Effect of PS beads on compressive strength development.

133

(0
Q.

c
03

100

1000

10000

Age (hour)

Figure 2.30 - HES-ECC tensile strength development with ages.

4000
25

3500
3000 |-

20

<*

_ 2500
i2 2000

15 "fa

w 1500

10

1000
-I 5

500
0

i_

10

100

1000

0
10000

Age (hour)

Fig. 2.31 - HES-ECC Young's modulus (E) development with ages.

134

7
6

Vu

"o 4
D.
10

^
o 3

c
2 2
<s>

1
I I I I I I

10

' '

1LI

I IU

1000

100

1 !_)!

10000

Age (hour)
Figure 2.32 - HES-ECC tensile strain capacity development with ages.

24h: crack w idlh 30 - 60 inn


(I.IK " 1 0 J -2.3(^10"' in)

4li: uruoU width around 10 um


l0.3 l >-10 :i in.)
1 !';

3d: cuck widll' 20


IO'^MO'

^0 um

l.'JT-M)

in)

Final failure

28d: cuick widlh 30 60 um


fl 18 *l()
2.36 10 'in)

Figure 2.33 - HES-ECC multiple microcracking pattern at different ages.

135

a.

01

(/)
<D

.>
</>
a.

o
U

10

100

1000

10000

Age (hour)

Figure 2.34 - HES-ECC compressive strength development with age.

a NJSDOT, 1 i97
O MDOTQAf"QC, 2003
A 2ia el a!., 1991
* FHWA.199i9
"""'Target
^ HES"liCC

Age (hour)

ure 2.35- Compressive strength (average value) development of HES-ECC compared


with requirements by DOTs and FHWA.

136

Heml of Testing Machine;

Load-applying apci
Support Block

Span Length = 304.Kmm (12 in)


Bed of Testing Machine

Figure 2.36 - Third-point bending test setup and specimen dimensions.

28 days
24 hours

a.

^rK h o u r s

m
m
a
w
A3
X

:t

4H"-4--4--l~HMI-

_#_|~_+.

10
15
Displacement (mm)

20

Fig 2.37 - HES-ECC flexural stress - displacement curves at 4 hours, 24 hours, and 28
days.

137

References:
1

Wang, S., and Li, V. C , "High Early Strength Engineered Cementitious Composites,"
ACI Materials Journal, Vol. 103, No. 2, 2006, pp. 97-105.
2

Naaman, A. E., and Reinhardt, H. W., "Characterization of High Performance Fiber


Reinforced Cement Composites HPFRCC," High Performance Fiber Reinforced
Cementitious Composites, RILEM Proceeding 31, Eds. A. E. Naaman and H. W.
Reinhardt, 1996, pp. 1-23.
3

Li, V.C., and Wang, S., "On High Performance Fiber Reinforced Cementitious
Composites," Proc. of the JCI Symposium on Ductile Fiber Reinforced Cementitious
Composites (DFRCC), Tokyo, Japan, pp. 13-23, 2003.

Aveston, J., Cooper, G., and Kelly, A., "Single and Multiple Fracture," Properties of
Fiber Composites, Guidford, UK, 1971, pp. 15-24.
5

Krenchel, H. and Stang, H., "Stable microcracking in cementitious materials,"


Brittle Matrix Composites 2. A.M. Brandt and J.H. Marshall, eds., 1989, pp. 20-33.

In

Curbach, M. and Jesse, F.,"High-Performance Textile-Reinforced Concrete," Structural


Engineering International, Vol. 9, No. 4, 1999, pp. 289-291.
7

Reinhardt, H.W., Kriiger, M. and GroBe, C.U., "Concrete prestressed with textile
fabric," Journal of Advanced Concrete Technology, Vol. 1, No. 3, 2003, pp. 231-239.
8

Mobasher, B., Peled, A. and Pahilajani, J., "Distributed cracking and stiffness
degradation in fabric-cement composites," Materials and Structures, Vol. 39, No. 3,
2006, pp. 317-331.
9

Allen, H.G., "Stiffness and strength of two glass-fiber reinforced cement laminates,"
Journal of Composite Materials, Vol. 5, No. 2, 1971, pp. 194-207.
10

Lankard, D.R., "Preparation, properties and applications of cement based composites


containing 5-20 percent steel fiber reinforcement," Steel Fiber Concrete, S.P. Shah, S.P.
and A. Skarendahl, eds. Elsevier Publishers, 1986.
11

Naaman, A.E., "SIFCON:


Tailored properties for structural performance,"
Proceedings of the International Workshop on High Performance Fiber Reinforced
Cement Composites, H.W. Reinhardt and A.E. Naaman, eds., Published by E & FN Spon,
London, 1992, pp. 18-38.
Li, V.C., "From Micromechanics to Structural Engineering - the design of
cementitious composites for Civil Engineering applications," JSCE J. of Struc.
Mechanics and Earthquake Engineering, Vol. 10, No. 2, 1993, pp. 37-48.

138

13

Li, V. C , and Leung, C. K. Y., "Theory of Steady State and Multiple Cracking of
Random Discontineous Fiber Reinforced Brittle Matrix Composites," Journal of
Engineering Mechanics, ASCE, Vol. 118, No. 11, 1992, pp. 2246-2264.
14

Li, V. C , and Wu, H. C., "Conditions for Pseudo Strain-Hardening in Fiber Reinforced
Brittle Matrix Composites," Journal of Applied Mechanics Review, Vol. 45, No. 8,
August, 1992, pp. 390-398.
15

Marshall, D.B., and Cox, B. N., "A J-Integral Method for Calculating Steady-State
Matrix Cracking Stresses in Composites," Mechanics of Materials, Vol. 7, No. 2, 1988,
pp. 127-133.
16

Griffith, A. A., "The phenomena of rupture and flow in solids," Philosophical


Transactions of the Royal Society of London, A 221, 1921, pp. 163-198,
http://www.cmse.ed.ac.uk/AdvMat45/Griffith20.pdf.
17

Wang, Y., Backer, S., and Li, V.C., "A Statistical Tensile Model of Fiber Reinforced
Cementitious Composites," Journal of Composites, Vol. 20, No. 3, 1990, pp.265-274.
18

Aveston, J., Cooper, G. A., and Kelly, A., "Single and Multiple Fracture," Proceedings
of Properties of Fiber Composites, National Physical Laboratory, IPC Science and
Technology Press, Buildford, UK, 1971, pp. 15-24.
19

Wu, H. C , and Li, V. C , "Stochastic Process of Multiple Cracking in Discontinuous


Random Fiber Reinforced Brittle Matrix Composites," International Journal of Damage
Mechanics, Vol. 4, No. 1, 1995, pp. 83-102.
20

Li, V. C , and Wang, S., "Microstructure Variability and Macroscopic Composite


Properties of High Performance Fiber Reinforced Cementitious Composites,"
Probabilistic Engineering Mechanics, Vol. 21, No. 3, 2006, pp. 201-206.
21

American Concrete Pavement Association, "How do I obtain "high early strength"


concrete?",
http://www.pavement.com/Concrete_Pavement/Technical/FATQ/Design/High_Early_Str
engthC oncrete. asp.
Delatte, N. J., Concrete Pavement Design, Construction, and Performance, 2007,
372pp.
Portland Cement Association, http://www.cement.org/pavements/pv_cp_airports.asp,
accessed on August 15, 2009.
24

Elhindy, E., Chevalier, F. and Khayat, K. H., "Airfield Pavement Repair Using High
Early Strength Concrete: Experience of the Aeroports De Montreal," Concrete for
Infrastructure and Utilities by Ravindra, L, D. and Henderson, N. A., 1996, pp. 601-603.

139

25

Kohn, S. D., Tayabji, S., and Okamoto, P., Best Practices for Airport Portland Cement
Concrete Pavement Construction Rigid Airport Pavement, American Concrete Pavement
Association, Washington, D. C , 2003, 146 pp.
26

Seehra, S. S., Gupta, S., and Kumar, S., "Rapid Setting Magnesium Phosphate Cement
for Quick Repair of Concrete Pavements - Characterization and Durability Aspects,"
Cement and Concrete Research, Vol. 23, No. 2, 1993, pp. 254-266.

27

Knofel, D., and Wang, J. F., "Properties of Three Newly Developed Quick Cements,"
Cement and Concrete Research, Vol. 24, No. 5, 1994, pp. 801-812.

28

Whiting, D., and Nagi, M., "Strength and Durability of Rapid Highway Repair
Concretes," Concrete International, Vol. 16, No. 9, 1994, pp. 36-41.
29

Sprinkel, M. M., "Very-Early-Strength Latex-Modified Concrete Overlay,'''ReportNo.


VTRC99-TAR3, Virginia Department of Transportation, Richmond, Virginia, 1998, 11pp.

30

Baraguru, P. D., and Bhatt, D., "Rapid Hardening Concrete," Report No. FHWA NJ
2001-3, New Jersey Department of Transportation, Trenton, New Jersey, 2000, 22pp.
31

Parker, F., and Shoemaker, M. L., "PCC Pavement Patching Materials and
Procedures," ASCE Journal of Materials in Civil Engineering, Vol. 3, No. 1, 1991, pp.
29-47.
32

Kurtz, S., Balaguru, P., Consolazio, G., and Maher, A., "Fast Track Concrete for
Construction Repair," Report No. FHWA 2001-015, New Jersey Department of
Transportation, Trenton, New Jersey, 1997, p. 67.
33

Anderson, J., Daczko, J., and Luciano, J., "Producing and Evaluating Portland CementBased Rapid Srength Concrete," Concrete International, Vol. 25, No. 8, 2003, pp. 77-82.
34

Bentz, D. P., and Peltz, M. A., "Reducing Thermal an Autogeneous Shrinkage


Contributions to Early-Age Cracking," ACI Materials Journal, Vol. 105, No. 4, JulyAugust 2008, pp. 414-420.
35

Mehta, P. K., and Burrows, R. W., "Building Durable Structures in the 21 st Century,"
Concrete International, Vol. 23, 2001, pp. 57-63.

Mihashi, H., Nishiyama, N., Kobayashi, T., and Hanada, M., "Development of a Smart
Material to Mitigate Thermal Stress in Early Age Concrete," Control of Cracking in
Early Age Concrete, Eds. H. Mihashi & F. H. Wittmann, Balkema, Rotterdam, 2002, pp.
385-392.
37

Wang, K., Jansen, D, Shah, S., and Karr, A., "Permeability Study of Cracked
Concrete", Cement and Concrete Research, Vol. 27, No. 3, 1997, pp. 381-393.

140

Anderson, J., "Paving Repair Finds a Four-Hour Champion," Concrete Construction,


Vol. 46, No. 12, 2001, pp. 69-70.
39

Kurtz, S., Balaguru, P., Consolazio, G., and Maher, A., "Fast Track Concrete for
Construction Repair," Report No. FHWA 2001-015, New Jersey Department of
Transportation, Trenton, N.J., 1997, p. 67.
40

Michigan Department of Transportation (MDOT), "Qualification Procedure For


Prepackaged Hydraulic Fast-Set Materials for Patching Structural Concrete," MDOT
Quality Assurance and Quality Control (QA/QC), 2003.
41

Parker, F., and Shoemaker, M. L., "PCC Pavement Patching Materials and
Procedures," ASCE Journal or Materials in Civil Engineering, Vol. 3, No.l, 1991, pp.
29-47.
42

Zia, P., Leming, M. L., and Ahmad, S. H., "High Performance Concretes, a State-ofthe-Art Report," Report No. SHRP-C/FR-91-103, Strategic Highway Research Program,
National Research Program, Washington, DC, 1991.
43

Federal Highway Administration (FHWA), "Manual of Practice: Materials and


Procedures for Rapid Repair of Partial-Depth Spalls in Concrete Pavements," Federal
Highway Administration (FHWA), 1999, p. 135.
44

Neville, A. M., and Brooks, J. J., Concrete Technology, Longman Scientific and
Technical, 1987.
45

Smith, K. L., Peshkin, D.G., Rmeili, E. H., Dam, T. V., Smith, K. D., Darter, M. I.,
"Innovative materials and Equipment for Pavement Surface Repairs", Strategic Highway
Research Program (SHRP), National Highway research Council, Washington D.C., 1991,
205 pp.

46

Stingley, W. M., "NCHRP Synthesis of Highway Practice No. 45: Rapid Setting
Materials for Patching of Concrete," National Cooperative Highway research Program
(NCHRP), Transportation Research Board, Washington D.C., 1977, 13 pp.
47

Collepardi, M., "Water Reducers/Retarders", Concrete Admixtures Handbook, ed. V.


S. Ramachandran, Noyes Publications, Park Ridge, NJ, 1984, pp. 116-210.

48

Malhotra, V. M., Berry, E. E., and Wheat, T. A., "Superplasticizers in Concrete",


Proceedings of International Symposium, Ottawa, 1988 (2 vols.), CANMET, Department
of Energy, Mines and Resources, Ottawa, Canada, also published in part as ACI SP-62,
American Concrete Institute, Detroit, MI, 1989.
49

Nmai, C. K., Schlagbaum, T., and Violetta, B., "A History of Mid-Range Water
Reducing Admixtures", Concrete International, Vol. 20, No.4, 1998, pp. 45-50.

141

Ramachandran, V. S., Malhotra, V. M., Jolicoeur, C. and Spriato, N., "Superplasticizer:


Properties and Applications in Concrete," Publ. No. MTL-14 (TR), CANMET National
Resources Canada, Ottawa, Canada, 1998.
51

Edmeades, R. M. and Hewlett, P. C , "Cement Admixtures", Lea's Chemistry of


Cement and Concrete, (4th ed.), ed. P. C. Hewlett, Arnold, London, 1998, pp. 837-902.
52
Mindess, S., Young, J. F., and Darwin, D., Concrete (2ndEd.), 2003.
53

ACI Committee 212, "Admixtures for Concrete (ACI 212.3R-91)", ACI Manual of
Concrete Practice, Part 1, American Concrete Institute, Detroit, MI, 2001.
5

Ramachandran, V. S., Calcium Chloride in Concrete Science and Technology, Applied


Science Publishers, London, 1986.
55

Rear, K., and Chin, D., "Non-Chloride, Accelerating Admixtures for Early
Compressive Stength", Concrete International, Vol. 12, No. 10, 1990, pp. 55-58.
5

Hewlett, P. C , and Young, J. F., "Physico-chemical Interactions between Chemical


Admixtures and Portland Cement", Journal of Materials Education, Vol. 9, No. 4, 1989,
pp. 389-433.
en

Ramachandran, V. S., Concrete Admixtures Handbook, Noyes Publications, Park


Ridge, NJ, 1984.
5

Rixom, M. R., and Mailvaganam, N. P., Chemical Admixtures for Concrete (2nd ed.),
E & FN Spon, London, 1986.
59

Rear, K., and Chin, D., "Non-Chloride, Accelerating Admixtures for Early
Compressive Stength", Concrete International, Vol. 12, No. 10, 1990, pp. 55-58.
60
Ramachandran, V. S., Concrete Admixtures Handbook, Noyes Publications, Park
Ridge, NJ, 1984.
1

Rixom, M. R., and Mailvaganam, N. P., Chemical Admixtures for Concrete (2nd ed.),
E & FN Spon, London, 1986.
62

Shao, Y. and Shah, S. P., "Mechanical Properties of PVA Fiber Reinforced Cement
Composites Fabricated by Extrusion Processing," ACI Material Journal, Vol. 94, No. 6,
1997, pp. 555-564.
63

Kanda, T., and Li, V. C , "Interface Property and Apprarent Strength of a High
Strength Hydrophilic Fiber in Cement Matrix," ASCE Journal of Materials in Civil
Engineering, Vol. 10, No. 1, 1998, pp. 5-13.

142

64

Redon C , Li, V. C , Wu, C , Hoshiro, H., Saito, T., and Ogawa, A., "Measuring and
Modifying Interface Properties of PVA Fibers in ECC Matrix," ASCE Journal of
Materials in Civil Engineering, Vol. 13, No. 6, 2001, pp. 399-406.
65

Li, V. C , Wu, C , Wang, S., Ogawa, A., and Saito, T., "Interface Tailoring for Strainhardening PVA-ECC," ACIMaterials Journal, Vol. 99, No. 5, Sept.-Oct, 2002, pp. 463472.
66

Stang, PL, "Scale Effects in FRC and HPRFCC Structural Elements," High
Performance Fiber Reinforced Cementitious Composites, RILEM Proceedings Pro 30,
Eds. Naaman, A.E. and Reinhardt, H.W., 2003, pp. 245-258.
67

Kanda, T., Design of Engineered Cementitious Composites for Ductile Seismic


Resistant Elements, Ph.D. Dissertation, University of Michigan, 1998, 329 pp.
68

Redon, C , Li, V.C., Wu, C , Hoshiro, H., Saito, T., and Ogawa, A., "Measuring and
Modifying Interface Properties of PVA Fibers in ECC Matrix," Journal of Materials in
Civil Engineering, Vol. 13, No. 6, 2001, pp. 399-406.
69

Lin, Z., Kanda, T., and Li, V. C , "On Interface Property Characterization and
Performance of Fiber Reinforced Cementitious Composites," Journal of Concrete
Science and Engineerings RILEM, Vol. 1, 1999, pp. 173-184.
70

Chan, Y. W., and Li, V. C , "Age Effect on the Characteristics of Fibre/Cement


Interfaical Properties," Journal of Materials Science, Vol. 32, 1997, pp. 5287-5292.
71
Wang, S., "Micromechanics Based Matrix Design for Engineered Cementitious
Composites," Ph.D. Dissertation, University of Michigan, 2005.
72

Vaysburd, A. M., Brown, C. D., Bissonnette, B., and Emmons, P. H., ""Realcrete"
Versus "Labcrete"", Concrete International, Vol. 26, No. 2, 2004, pp. 90-94.
73

Li, M., and Li, V. C , "Durability of HES-ECC Repair Under Mechanical and
Environmental Loading Conditions," Proc, High Performance Fiber Reinforced Cement
Composites, Eds. H.W. Reinhardt & A.E. Naaman, Mainz, Germany, 2007, pp. 399-408.
74

Li, M., and Li, V. C , "Influence of Material Ductility on the Performance of Concrete
Repair," submitted to ACI Structural Journal, June, 2008.
75

Lepech, M., and Li, V. C , "Water Permeability of Cracked Cementitious Composites,"


Paper 4539 of Compendium of Papers CD ROM, ICF 11, Turin, Italy, March 2005.
76

Sahmaran, M., Li, M., and Li, V. C , "Transport Properties of Engineered Cementitious
Composites Under Chloride Exposure," ACI Materials Journal, Vol.104, No. 6, 2007,
pp.604-611.

143

Qian, S., and Li, V. C , "Simplified Inverse Method for Determining the Tensile Strain
Capacity of Strain Hardening Cementitious Composites," J. Advanced Concrete
Technology, Vol. 5, No. 2, 2007, pp. 235-246.
78

ASTM C78 - 08 Standard Test Method for Flexural Strength of Concrete (Using
Simple Beam with Third-Point Loading).

144

CHAPTER 3

ECC Processing and Quality Control for Optimized Material Properties and
Robustness

The capability of consistently processing and producing robust ECC materials


plays a crucial role in its ascendancy as a new construction material in various structural
applications.

ECC strain-hardening behavior and tensile strain capacity are closely

related to fiber dispersion uniformity. Non-uniform fiber dispersion within ECC can
result in huge variation in its tensile properties, and even switch ECC from a strainhardening material to a tension-softening normal fiber reinforced concrete (FRC) material.
This can lead to material properties less than originally designed for based on the
micromechanical theory described in Chapter 2; this theory assumes uniform fiber
dispersion.
This study focuses on the development of a simple and practical quality control
method for ECC processing. Through measuring ECC mortar rheological parameters,
fiber dispersion, and ECC tensile properties, the correlation between the three was
established. The optimal range of Marsh cone flow rate, which is an indicator of plastic

145

viscosity of ECC mortar, was then indentified and used to guide ECC quality control.
Based on these experimental findings, it is proposed that ECC material can achieve
maximized fiber dispersion and hardened tensile properties, as well as improved
robustness, through plastic viscosity control of ECC mortar within this optimal range.
With this method, more widespread adoption of ECC material in structural applications
as well as laboratory testing can be realized with confidence.

A material with more

robust properties will also lead to enhanced reliability in structural design, and more
complete utilization of the material performance (e.g. by use of a smaller factor of safety).
Within the material engineering phase of the Integrated Material and Structural
Engineering framework, the material processing technique developed in this chapter,
together with the microstructure tailoring emphasized in Chapter 2, converge to
ultimately produce composite materials with desired combinations of properties and
robustness.

3.1

Introduction
The laboratory and on-site processing and mixing capacities of ECC play a crucial

role in its ascendancy as a construction material for structural applications. The ability to
produce robust ECC materials with controlled quality in laboratories and on construction
sites using existing equipment is essential for standardized testing of this new material
and its successful introduction to the construction industry. Variation in ECC tensile
behavior has been observed in laboratory small-scale and on-site large-scale production
depending on the type of mixers, mixing procedure, source of local ingredients, and
mixing personnel's experience. This variation is reflected in the standard deviation of

146

measured material tensile strain capacity, ultimate tensile strength, and width and density
of microcracks. Figure 3.1 shows an extreme example of such variation in tensile stressstrain curves of one ECC mix produced from different batches using a gravity mixer with
capacity of 0.6 m3 (2 ft.3). Although the mixing proportion is identical, the specimens
showed tensile strain capacity ranging from 0.6% to 3.4%.

Some specimens formed

closely spaced "saturated" microcracks during the strain-hardening stage, and therefore
had larger tensile strain capacity before the tension-softening stage started.

Other

specimens reached the tension-softening stage soon after a limited number of


microcracks formed, leading to a lower tensile strain capacity. Although this example
reveals a somewhat exaggerated variation, it serves to illustrate the need for a simple and
effective quality control methodology in ECC processing for consistent and optimized
properties.
The variation in ECC tensile ductility is mainly due to the influence of processing
details on composite microstructure that governs ECC material properties1.

Notably,

ECC utilizes short, randomly distributed polymer fibers at a moderate volume fraction
(2% or less in general). During ECC mixing, fibers are added after the other ingredients,
i.e. cementitious and pozzolanic ingredients, water, and admixtures are all mixed and
have achieved a consistent mortar state. Based on past experience, it is known that
processing details such as mixer type, mixing speed, time and sequence, and mixing
personnel's experience level can influence the rheological properties of ECC mortar
(before adding fibers) while in fresh state, which in turn can strongly affect fiber
dispersion uniformity, size distribution of entrapped air pores in the cementitious matrix,
and bonding properties at the fiber/matrix interface in the hardened state. Using ECC

147

ingredients from different sources can also lead to a change in ECC rheological
properties and composite microstructure.
Research has been conducted on the influence of rheological characteristics on the
fiber dispersion in cementitious materials.

Ozyurt et al, 2006 2 found that fresh

rheological properties affect segregation of steel fibers and mechanical properties of steel
fiber reinforced concrete. Standard deviation of the steel fiber contents throughout the
specimen decreases when viscosity increases.

Chung 2005 3 assessed the degree of

dispersion of short microfibers, e.g. carbon and steel fibers, in cement mortar or paste
through measuring the volume electrical resistivity, and found that the degree of
dispersion is improved by the use of silica fume, acrylic particle dispersion,
methylcellulose solution, silane, and use of fiber surface treatment.

Yang et al 4

investigated the effects of fly ash type, water-binder ratio, and chemical admixtures on
the fresh and hardened properties of ECC by means of design of experiments. Test
results indicated that ECC tensile properties correlated well with the viscosity of fresh
ECC mortar.
Within this chapter, a simple and effective quality control approach is presented
for the production of robust ECC materials with optimized tensile properties and
minimized material variation. To achieve this goal, we build on the findings of Yang et
al .

Specifically, a systematic experimental framework is designed in this study to

establish the correlation between ECC mortar rheological properties in fresh state, fiber
dispersion in hardened state, and composite tensile properties in hardened state. Based
on this correlation, the desired range of ECC mortar rheological properties is indentified.
Through controlling ECC mortar rheology within this range, it is confirmed that

148

optimized composite properties can be achieved through more uniform fiber dispersion.
Intended applications for this quality control approach are standardized laboratory
production of ECC for research purposes, as well as ECC factory precast, truck mixing,
and on-site processing for various construction projects.

3.2

Effect of Processing on ECC Micromechanical Parameters


As discussed in Chapter 2, tensile strain-hardening behavior of ECC is realized by

tailoring the synergistic interaction between fiber, matrix, and fiber/matrix interface
based on micromechanics theory. As a fiber-reinforced brittle mortar matrix composite,
ECC's pseudo strain-hardening behavior is achieved through sequential formation of
matrix multiple cracking. The fundamental requirement for matrix multiple cracking is
that steady-state flat crack propagation prevails under tension, which was first
characterized by Marshall and Cox5 for continuous aligned fiber-reinforced ceramics, and
extended to discontinuous fiber-reinforced cementitious composites by Li and Leung6.
To ensure steady-state cracking, the crack tip toughness JtiP must be less than the
complementary energy V calculated from the fiber bridging stress a versus crack
opening 8 curve, as illustrated in Figure 3.2. This energy-based criterion determines
whether the crack propagation mode is steady-state flat crack or Griffith crack7, as
illustrated in Figure 2.4.
The fiber bridging stress versus crack opening relationship o(S), which can be
viewed as the constitutive law of fiber bridging behavior, is analytically derived based on
fracture mechanics, micromechanics, and probabilistics tools.

In particular, the

energetics of tunnel crack propagation along fiber/matrix is used to model the debonding

149

process of a single PVA fiber from the surrounding cementitious matrix.

After

debonding is completed, the fiber pullout stage begins, and is modeled as slip-hardening
behavior assuming a non-linear frictional stress increase with slip distance. By these
means, the full debonding-pullout process of a single fiber with given embedment length
is quantified as the fiber bridging force vs. fiber displacement relation6. Probabilistics is
then introduced to describe the randomness of fiber location and orientation with respect
to a crack plane, with the assumption of uniform random fiber distribution8. The random
orientation of the fibers also necessitates the accounting of the mechanics of interaction
between an inclined fiber and the matrix crack. In addition, snubbing coefficient f and
strength reduction factor f are introduced to account for the interaction between fiber and
matrix as well as the reduction of fiber strength when pulled at an inclined angle. As a
result, the 0(6) curve is expressible as a function of micromechanics parameters,
including fiber volume content Vf, fiber diameter df, fiber length Lf, fiber Young's
modulus Ef, matrix Young's modulus Em, interface chemical bond Gd, interface frictional
bond To, and slip-hardening coefficient p, as well as f and f.
Apart from the energy criterion (1), another condition for pseudo strain-hardening
is that the matrix tensile cracking strength ac must not exceed the maximum fiber
bridging strength o 0 .
For ECC materials with pseudo strain-hardening behavior, high tensile strain
capacity results from saturated formation of multiple microcracks.
strength increases as the number of microcracks increases.

Material tensile

While the steady-state

cracking criteria ensure the occurrence of multiple-cracking, it is not directly related to


the intensity of multiple cracking. Once the steady-state cracking criteria is satisfied, the

150

number of microcracks that can be developed before reaching the maximum bridging
stress oo is determined by (i) the maximum fiber bridging stress ao, and (ii) the matrix
properties, in particular the pre-existing flaw size distribution, and the matrix fracture
toughness. Considering that ECC is a non-homogenous brittle-matrix composite, firstcracking strength is determined by the largest flaw size in the section where the first
microcrack is initiated.

Its ultimate tensile strength is determined by the "weakest"

section, where the fiber bridging capacity (maximum bridging stress OQ) is the lowest
among all sections subjected to the same level of stress. Therefore, the maximum fiber
bridging stress Oo at the "weakest" section imposes a lower bound of critical flaw size cmc
so that only those flaws larger than c mc can be activated and contribute to multiple
cracking. There also exists a minimum crack spacing controlled by interface properties,
which imposes an upper bound for the density of multiple cracking.
Rheologistic characteristics of fresh ECC during processing affect fiber dispersion,
which influences the possibility and density of multiple-cracking. During ECC material
design through microstructure tailoring, determination of the fiber bridging stress versus
crack opening relationship o(5) from single fiber pullout test results is based on the
assumption of uniform random fiber distribution, which can hardly be ensured in actual
processing.

The uniformity of fiber dispersion in reality determines the maximum

bridging stress a 0 , the shape of the a(5) curve at the "weakest" section, and the critical
flaw size cmc, as illustrated in Figure 3.2.

Non-uniform fiber dispersion leads to a

reduction of the value of Oo at the "weakest" section which increases the critical flaw size
Cmc. Therefore, less pre-existing flaws with sizes larger than cmc can be triggered and
contribute to multiple cracking, resulting in a relatively lower tensile strain capacity.

151

Non-uniform fiber dispersion also shifts the a(5) curve downwards, and may reduce
complementary energy Jb' to less than Jtip. In this case, the steady-state criteria are
violated and tension-softening behavior results.

Therefore, ECC loses its ductile

behavior and becomes a regular FRC material.

Ideally, processing of ECC should

optimize fiber dispersion to achieve a uniform random distribution state, therefore


minimizing the probability of creating "weak" sections with lower fiber content. By
these means, the largest possible tensile strain capacity can be achieved through
maximizing multiple cracking behavior.
In the following experimental study, the correlation between fiber dispersion
uniformity and tensile strain capacity, as described above, is validated. The optimal
range of ECC fresh properties that correspond to the largest fiber dispersion coefficient
and tensile strain capacity are determined.

3.3

Experimental Program
The experimental program contains four sets of investigations.

First, fresh

viscosity and flowabilty of seven ECC mortar mixes were measured using the Marsh
cone flow test and mini slump test. Second, ECC tensile specimens made from the seven
mixes were tested under uniaxial tensile load to measure their tensile stress-strain curves
and tensile strain capacity.. Third, fiber dispersion at the "weakest" section was measured
using fluorescence imaging technology.

Based on data from the three tests, the

relationships between ECC fresh properties, tensile strain capacity, and fiber dispersion
were established, and the optimal range of fresh properties was indentified. Finally, to
prove the effectiveness of this method, several versions of ECC were processed by

152

optimally controlling rheology to achieve improved and robust hardened material


properties. The experimental program is shown in Figure 3.3.

3.3.1

Materials
Seven mixtures with the same mix proportion but increased viscosity were

produced. The mix proportion was determined according to ECC micromechanics-based


design theory described above for strain-hardening behavior. Increased viscosity of the
seven mixtures was achieved by adding a Viscosity Modifying Agent (VMA), V-MAR,
with increased dosage of 0, 0.01%, 0.015%, 0.02%, 0.025%, 0.03% and 0.04% of cement
weight. V-MAR 3 is a high efficiency, milky white liquid admixture designed by W. R.
Grace, which enables production of Self Consolidating Concrete by modifying the
rheology of concrete9. It works by increasing the viscosity of the concrete while still
allowing the concrete to flow without segregation. V-MAR 3 is recommended for use
in conjunction with ADVA Cast 530, which is the common superplasticizer
incorporated into different versions of ECC materials, to achieve self-consolidating
properties.
The material mix proportions of the seven mixes are shown in Table 3.1. The
cement is Ordinary Portland Cement (OPC) type I from Holcim Inc., USA, with Blaine
Surface Area of 3300 cm2/g (2.320xl0 5 in2/lb). The fly ash is ASTM standard type F fly
ash with mean grain size of 10-20 [im (3.937xl0~4-7.874xl0"4 in.) from Boral Material
Tech. Inc., USA. F-110 fine silica sand from US silica, with 250 urn (9.843xl0~3 in.)
maximum grain size and 110 iim (4.33lxlO"3 in.) mean grain size, was used as fine
aggregate.

The superplasticizer is a polycarboxylate-based high range water reducer

153

(ADVA Cast 530) from W.R. Grace & Co. Polyvinyl Alcohol (PVA) fiber REC-15
from Kuraray Co., Ltd., Japan, is used at a volume fraction of 2% in this study. This
particular volume fraction and other fiber properties were determined by the ECC
micromechanical model to satisfy strain-hardening criteria. The PVA fibers are 8 mm
(3.150x10"' in.) long and 39 um (3.150xl0~4 in.) in diameter, with nominal tensile
strength of 1600 MPa (232 ksi) and density of 1300 kg/m3 (81.156 lb/ft3). The fiber is
surface-coated with oil (1.2% by weight) to reduce the interfacial bond with the
surrounding cementitioius matrix.

3.3.2

Mixing, Testing of Fresh Properties, and Casting


In this study, all mixtures were prepared using the same 12 L (3.170 gallon)

capacity Hobart mixer, and followed the same mixing sequence, speed, and time under
controlled room temperature 201 C (66-70 F) and relative humidity conditions 505
% RH. By these means, the isolated effect of plastic viscosity could be investigated.
Solid ingredients, including cement, fly ash, and silica sand, were first mixed at 100 rpm
for one minute. Water and chemical admixtures, including superplasticizer and VMA,
were then added into the dry mixture and mixed at 150 rpm for three minutes to produce
a consistent and uniform ECC mortar (w/o PVA fiber).

The fresh mortar was then

measured for its workability and rheological parameters.


Mortar is generally assumed to be a non-Newtonian fluid, and the Bingham model
is often used to describe its rheology10. According to this model, fresh mortar must
overcome a yield stress before it can flow. Once the mortar starts to flow, shear stress
increases linearly with an increase in strain rate, the slope of which defines the plastic

154

viscosity. The "plastic viscosity" measures how easily the material can flow, once the
yield stress is overcome. The "deformability" measures the ultimate extent to which the
material can flow. In this study, plastic viscosity of the seven mortars was measured
using a rotational viscometer as a direct method, and Marsh cone flow time as an indirect
method. While the direct method is more complicated and requires computer software to
analyze data, the indirect method is simpler and the device is portable to mixing sites.
The objective was to validate that the Marsh cone flow time correlates well with the
plastic viscosity of ECC mortar measured using the rotational viscometer, so that the
Marsh cone flow rate test can be adopted as a practical and reliable method for quality
control during ECC processing on-site. The deformability of ECC mortar was measured
using mini-slump flow test.
Fresh rheology test equipment is shown in Figure 3.4. A Viskomat-NT rotational
viscometer11, (Figure 3.4 (a)) was used to measure the plastic viscosity and yield stress of
seven ECC mortars at a controlled temperature of 201 C (66-70 F). The material was
stirred at a rotation rate N, generating shear resistance of torque T. A series of data
points of T and N were recorded by a computer. Through linear regression, the relative
yield stress g (N-mm) and relative plastic viscosity h (N-mm/rpm) can be determined1213
by:
T = g + Nh

(3.1)

Relative plastic viscosity (N-mm/rpm) can be converted to plastic viscosity (Pa.s) using
a calibration liquid14.
A Marsh cone (Figure 3.4 (b)), also called a "Marsh funnel" or "Marsh funnel
viscometer", is a simple device for measuring viscosity according to the time it takes a
known volume of liquid to flow from the bottom of the cone though a short tube15,16. It is

155

often used in the concrete and oil industry to check the quality of cement grout mixtures
or drilling mud. In this test, a funnel was filled completely with ECC mortar and then the
bottom outlet was opened, allowing the mortar to start flowing. The Marsh cone flow
time of mortar was determined as the elapsed time (t) in seconds between the opening of
the bottom outlet until light first became visible at the bottom, when observed from the
top.
The mini-slump test measures the consistency of mortar, and its results are
correlated with the yield stress17. As shown in Figure. 3.4 (c), the mini-slump cone has a
top diameter of 70 mm (2.76 in.), a bottom diameter of 100 mm (3.94 in.), and a height of
60 mm (2.36 in.). The cone was placed in the center of a square piece of glass and filled
with ECC mortar. The cone was then lifted up to allow the mortar to flow. Once the
flowing stopped, the spread of the mortar was recorded as the average of the diameters
measured along two diagonals.
After these tests of the fresh properties of ECC mortar, another batch of the same
mix was made following the same procedure, with the exception that fibers were added to
the ECC mortar and mixed at 150 rpm for three more minutes. The mixtures were then
cast into tensile plate molds, covered with plastic sheets, and demolded after 24 hours.
The specimens were then moisture-cured in plastic bags at 95+5 % RH and 201 C (6670 F) for 7 days, and air cured at 505 % RH and 201 C (66-70 F) for 21 days until
the age of 28 days for testing. Five specimens were prepared for each mixture.

3.3.3

Testing of Hardened Tensile Properties


At the age of 28 days, five specimens from each mixture were subjected to

uniaxial tensile test. The direct uniaxial tensile test is considered the most convincing
156

method to evaluate material strain-hardening behavior18. Fig. 7 illustrates the test setup
and specimen dimensions. The plate specimens were 228.6 mm (9 in.) in length, 76.2
mm (3 in.) in width and 12.7 mm (0.5 in.) in thickness. Before testing, four aluminum
plates were glued to both ends of the plate specimen to facilitate gripping. Tests were
conducted using an MTS machine with a 25 kN (5.620xl0 3 lbf) capacity under a
displacement control of rate of 0.0025 mm/s (9.843xl0"5 in./s) to simulate a quasi-static
loading condition. Two external LVDTs (Linear Variable Displacement Transducers)
were attached to the specimen surface with a gage length of 101.6 mm (4 in.) to measure
the displacement. The tensile stress-strain curve of each specimen was recorded.

3.3.4

Measurement of Fiber Dispersion Coefficient


After the uniaxial tensile test, a 5 mm-thick small sample was cut from each

tensile specimen at the final failure section, as shown in Figure 3.5. This failure section
is considered as the "weakest section" where Go is the lowest (among all the multiple
microcracks present at failure) for the specimen. Fluorescence imaging technique was
used to quantify the fiber distribution at this section. It is known that many materials
show fluorescence when irradiated with light19'20. Compared to the optical microscopic
method and the X-ray method, fluorescence technology has a very high level of
sensitivity to detect organic fibers from the surrounding cementitious matrix due to a low
background signal21'22.
Figure 3.6 illustrates the fluorescence imaging technology, and Figure 3.7 shows
the testing devices and setup. Without polishing, each of the forty-two samples from
seven mixtures were observed through a fluorescence microscope (TE300 Nikon). The

157

light source of the fluorescence microscope was the mercury lamp, which generated light
with a broad range of wavelengths. The sample was illuminated with light at a specific
wavelength, which was absorbed by the luorophores, causing them to emit longer
wavelengths of light as the result of well-established physical phenomena described as
fluorescence or phosphorescence. The emission of light through the fluorescence process
is nearly simultaneous with the absorption of the excitation light due to less than a
microsecond delay between photon absorption and emission23. The illumination light
was separated from the much weaker emitted fluorescence through the use of a UV filter.
Through this process, under the fluorescence microscope the fibers appeared brightly
colored while the surrounding cementitious matrix appeared dark grey. The fluorescence
image was then captured by a CCD camera (Hamamatsu 1394 ORCA-ER).
The whole cross sectional 76.2 mm x 12.7 mm (3 in. x 0.5 in.) image of each
sample, with 23005 x 4186 pixels, was constructed by connecting 105 (21 x 5) images
with image processing software.

3.4

Experimental Results and Discussion


The Marsh cone flow time of the seven ECC mortar mixtures is shown in Figure

3.8. Through adding an increased dosage of VMA, ECC mortar mixtures with the same
material composition but a broad range of plastic viscosity were achieved. From ECCO
(0% VMA content) to ECC_0.04% (0.04% VMA content), the Marsh cone flow time
increased from 9 seconds to 40 seconds due to increased plastic viscosity. The same
trend was observed for the plastic viscosity measured using the rotational viscometer. A
strong correlation (R = 0.95) was found in this study between the Marsh cone flow time

158

and plastic viscosity, showing that the Marsh cone flow test should be a reliable indirect
method for onsite or laboratory viscosity measurement.
The mini-slump flow diameters of these seven mixes before and after adding
fibers are shown in Figure 3.9. For the same mix, adding 2% fibers reduced the minislump flow diameter. Increasing VMA content seems to have a very slightly or no effect
on the mini-slump diameter, except for ECC_0 that showed little change before and after
adding fibers.

These test results indicate that including VMA can adjust the plastic

viscosity of ECC mortar without sacrificing ECC flowability, as indicated by the nonchanging mini-slump flow diameter.
Figure 3.10 shows the tensile stress-strain curves of the seven ECC mixtures from
the uniaxial tensile test. For ECC_0, only one specimen shows slight strain-hardening
behavior with 0.44% tensile strain capacity, while the other specimens show tensionsoftening behavior. Large variation exists not only in material tensile strain capacity, but
also in the ultimate tensile strength. For ECC_0.01%, although all of the specimens show
tensile strain-hardening behavior, their tensile strain capacity is relatively low (0.32% 1.19%). Large variation in tensile strain capacity and ultimate tensile strength was also
found in ECC_0.015%. While some specimens had improved tensile strain capacity up
to 1.97%, others had tensile strain capacity lower than 1%. ECC_0.02% exhibited more
consistent tensile behavior as well as greatly improved tensile strain capacity, with an
average value of 3.1%. Specimens from ECC_0.025%, ECCJ).03%, and ECC_0.04%
also all have tensile strain capacity around 3%, and greatly improved consistency in
tensile test data.
Figure 3.11 plots the mean value and standard error (as described by the error bar)

159

of tensile strain capacity versus VMA/cement ratio. The trend contains two parts: (i)
From VMA/cement ratio of 0 to 0.02%, tensile strain capacity increases while the
standard error decreases. Note that the tensile strain capacity standard error of ECC_0 is
lower than ECC 0.01% because of its lower mean value (standard error is an absolute
value); (ii) From VMA/cement ratio of 0.02% to 0.04%, there is no significant increase
in tensile strain capacity and a decrease in the standard error.
Figure 3.12 shows the mean value and standard error (as described by the error
bar) of tensile strength versus VMA/cement ratio. Again the trend contains two parts: (i)
From VMA/cement ratio of 0 to 0.02%, tensile strength increases while the standard error
decreases; (ii) From a VMA/cement ratio of 0.025% to 0.04%, tensile strength decreases
while the standard error remains low. Microscopy of the failure section revealed that
bigger air pores were entrapped during processing in ECC_0.04% compared with other
mixes due to a larger plastic viscosity. This effect, combined with less area of net cross
section available for fiber bridging, resulted in the lowered first-cracking strength and
ultimate tensile strength of ECC_0.04%. Future research is needed for detailed study of
the effect of plastic viscosity on ECC pore structure and tensile behavior.
Figures 3.11 and 3.12 together suggest an optimal VMA/Cement Ratio of 0.02%
for maximizing the composite tensile strain capacity without reducing the tensile strength.
Figure 3.13 shows the assembled fluorescence image of the whole cross section of
a sample specimen. Image processing software was used to increase the contrast of the
image to better reveal the fibers. This whole image was then divided into 6 x 26 unit
areas.

The number of fibers in each unit area was counted.

The fiber dispersion

uniformity of the whole cross section was quantified by a fiber dispersion coefficient,

160

which expresses the deviation of the number of fibers in a unit area from the average
number of fibers. The fiber dispersion coefficient a is calculated using Equations 3.2 and
3.3 24 :

S(s,--*) 2 / \
n
/

(3.2)

a = exp[-i/>(x)]

(3.3)

where \|/(x) is the coefficient of variation, x; is the number of fibers in the unit area
i, x is the average number of fibers in each unit area, and n is the number of unit area, a
equals to 1 when the fiber dispersion is uniform, and approaches 0 when the coefficient
of variation of fiber dispersion becomes larger.
Table 3.2 summarizes the Marsh cone flow rate of fresh ECC mortar before
adding fibers, fiber dispersion, and hardened tensile strain capacity of ECC after adding
fibers. The relationships between the three parameters is illustrated in Figure 3.14. A
strong positive correlation is found between the fiber dispersion coefficient and
composite tensile strain capacity (Figure 3.14 (a)). This observation validates the concept
discussed earlier that when the fibers are more-uniformly distributed in an ECC specimen,
the fiber bridging capacity a 0 at the "weakest" section is larger, leading to a smaller
critical flaw size ccm. Therefore, more pre-existing flaws with size larger than ccm can be
activated to form microcracks before the fiber bridging capacity at the "weakest" section
is exhausted, resulting in a larger tensile strain capacity.
Figure 3.14 (b) also shows that the fiber dispersion coefficient increases with
Marsh cone flow time in ECC mortar. This increasing trend reaches a plateau after the
Marsh cone flow time reaches 24 s. At the Marsh cone flow time of 39 s, a reduction in
161

the first cracking strength and ultimate tensile strength was found despite a high value for
the fiber dispersion coefficient. This means that there exists an "optimal" range of Marsh
cone flow time between 24 s and 33 s for desired good fiber distribution, but without the
attendant larger air voids as discussed earlier.

This Marsh cone flow time vs. fiber

dispersion coefficient relation can be used as a guideline for quality control during ECC
processing to achieve close-to-uniform fiber dispersion. It should be noted that mortar
from different versions of ECC can have different plastic viscosities at fresh state due to
variation in ingredients and composition. In this study, variation in VMA dosage was
used to achieve various plastic viscosities for the same ECC mix (Figure 3.8). For this
particular mix, an optimal dosage of 0.02-0.03% is suggested by this study. However,
this optimal dosage value may not apply to other ECC mixes, which have different
original viscosities (before viscosity control) due to different ingredients and composition.
The optimal range of viscosity is the most important factor in achieving uniform fiber
dispersion, which should not depend on the difference in mixes. In this sense, addition of
various dosages of VMA, together with other potential viscosity control methods (e.g.
modifying ingredient particle size distribution, adjusting water/cement ratio, using
superplasticizer), are only tools to achieve the objective of controlling viscosity within
the identified optimal range.
Figure 3.14 (c) shows the effect of ECC mortar Marsh cone flow rate on tensile
strain capacity. Due to the strong positive correlation between tensile strain capacity and
the fiber dispersion coefficient, the effect of Marsh cone flow time on tensile strain
capacity is similar to its effect on the fiber dispersion coefficient. Tensile strain capacity
increases with Marsh cone flow time, and reaches a plateau at 24 s. This trend indicates

162

that by controlling ECC mortar plastic viscosity so that its Marsh cone flow time falls
into the range between 24 s and 33 s before adding fibers, the hardened ECC material can
achieve

maximum

tensile

strain

capacity

as

originally

designed,

based

on

micromechanics tools which assume fiber dispersion uniformity.

3.5

Rheology Control Concept and Case Study


Once the correlations between ECC mortar Marsh cone flow rate, fiber dispersion,

and ECC tensile strain capacity were established, Figure 3.14 (b) became available as a
guideline to control the plastic viscosity of ECC mortar before adding fibers so that
maximized fiber dispersion and tensile strain capacity can be achieved. Plastic viscosity
is indirectly quantified as the Marsh cone flow time (in seconds) in this figure, and the
fiber dispersion coefficient is strongly correlated with tensile strain capacity. While the
rheology control methodology was demonstrated using an ECC of the composition given
in Table 3.1, the methodology itself is general and can be applied to other ECC
compositions.

This concept is illustrated with a case study of a newly developed

pigmentable ECC for achitectural applications, as described below.


WHITE ECC has the mix proportions shown in Table 3.3. This material was
developed based on ECC strain-hardening theory, but was different from the ECC used to
establish the rheology control methodology in terms of mix proportion and cement type.
WHITE ECC contains type I white cement from Lehigh Cement Company instead of
type I ordinary Portland cement. Also, fly ash was deliberately removed in its mix design.
No VMA was used in this initial mix design. Using white cement and eliminating fly ash
both reduce plastic viscosity of the mortar of WHITE ECC, making it challenging to

163

achieve robust composite properties due to the observed tendency of fiber clumping
during mixing. As shown in Figure 3.15(a), tests on three specimens of WHITE ECC
without rheology control showed large variation in tensile properties, with tensile strain
capacity ranging from 1.1% to 2.7%.
To conduct rheology control on WHITE ECC, the Marsh cone flow rate test was
performed on WHITE ECC mortar. Without any VMA, the Marsh cone flow rate for this
mix was experimentally determined to be 12 s, corresponding to a fiber dispersion
coefficient around 0.53 based on Figure 3.14 (b). This indicates that the fiber dispersion
in this material was relatively poor. By adding VMA at a dosage of 0.05% cement
content, the Marsh cone flow rate of WHITE ECC mortar was increased to 28 s,
corresponding to a fiber dispersion coefficient of around 0.8 (Figure 3.14 (b)). As shown
in Figure 3.15(b), this method greatly improved the average tensile strain capacity of
WHITE-ECC from 2% to 3%, and reduced the variation in tensile properties. Adding
more VMA, at a dosage of 0.1% of cement content, brought the Marsh cone flow rate up
to 38 s, which was close to the upper bound of the optimal range of Marsh cone flow time.
Tensile test results (Figure 3.15(c)) again showed greatly improved average tensile strain
capacity as well as reduced variation. However, the first cracking strength and ultimate
tensile strength were slightly reduced due to the higher value of plastic viscosity and
likely entrappment of air. For this WHITE ECC mix, therefore, the optimal amount of
VMA is 0.05% of cement content.

This is more than twice the optimal amount

recommended for the ECC mix discussed earlier (Table 3.1 E C C 0 ) , due to the
substantially lower plastic viscosity of the WHITE-ECC mortar mix.
This case study verified the effectiveness of the rheology control methodology.

164

This methodology can be used for quality control of any fresh ECC mortar before adding
fibers in order to achieve maximum fiber dispersion uniformity and composite tensile
strain capacity. The findings above reiterate the fact that the optimal amount of VMA
varies for different ECC mix compositions. However, the optimal plastic viscosity (or
Marsh cone flow rate) is a universal parameter.

3.6

Conclusions
Within this work, a simple and practical quality control method for ECC

processing was developed. Seven ECC mixes with the same mix proportions but with
viscosities varied by viscosity agent content were experimentally investigated to
determine rheological properties, hardened tensile properties, and fiber dispersion at the
final failure section. The following conclusions can be drawn:
A strong correlation, with R = 0.95, was found between the Marsh cone flow time
and plastic viscosity measured using a Viskomat-NT rotational viscometer, showing that
the simple Marsh cone flow test should serve as a reliable indirect method for onsite or
laboratory viscosity measurement.
Through measurements of the ECC mortar plastic viscosity and Marsh cone flow
rate, and ECC mini-slump flow diameter, it was found that incorporation of VMA can be
an effective method to control plastic viscosity of ECC mortar without sacrificing ECC
flowability after adding fibers.
Fluorescence imaging technique is a useful tool for quantifying the distribution of
short discontinuous PVA fibers within a cementitious matrix.

Fiber dispersion

coefficient for the seven mixes was found to range from 0.29 to 0.89. This quantifies the

165

effectiveness of adding VMA as a rheology control method for achieving more uniform
fiber dispersion in hardened ECC. This is the first time that fiber dispersion has been
directly and quantitatively correlated with deliberate viscosity control for optimal
composite tensile capacity performance.
Fiber uniformity as measured by the fiber dispersion coefficient was found to
have a strong effect on ECC tensile strain capacity. A lower fiber dispersion coefficient
not only reduces ECC tensile strain capacity and ultimate tensile strength, but also
increases the variation in both properties. A very low value for the fiber dispersion
coefficient can switch ECC from a strain-hardening material to a tension-softening
material.
ECC mortar plastic viscosity is established experimentally as a fundamental
rheological parameter that affects dispersion of PVA fibers in ECC mixes. An ECC
mortar with low plastic viscosity and Marsh cone flow time tends to have poorly
distributed fibers. Increasing plastic viscosity of ECC mortar improves fiber dispersion,
and the improvement reaches a plateau once the Marsh cone flow time reaches 24 s.
Further increasing viscosity can potentially lead to a reduction in ECC first cracking
strength and ultimate tensile strength due to more larger-size entrapped air pores.
Therefore, an optimal range of plastic viscosity (and corresponding optimal Marsh cone
flow time) was revealed, which can be used to guide ECC rheology control during
processing before fibers are added.
For ECC to achieve robust tensile strain-hardening behavior with designed tensile
ductility, micromechanical material design should be combined with controlled material
processing - neither of the two should be ignored. This is because material processing

166

strongly affects the composite microstructure, which in turn determines whether the
strain-hardening criteria are satisfied in the produced material as originally designed.
This concept was verified by the experimentally established relationship between ECC
mortar Marsh cone flow rate and ECC composite tensile properties, accompanied by
rheology control as demonstrated in this study.
Through the establishment of correlation between Marsh cone flow rate of ECC
mortar, fiber dispersion at the "weakest" section, and ECC tensile strain capacity, the
optimal range of Marsh cone flow time was identified. The present study results in a
practical methodology to control the quality of ECC during laboratory or onsite largescale processing for maximized material tensile properties and robustness, by simple use
of a portable Marsh cone and adjustment of VMA content to result in a flow time
between 24 s and 33 s. The effectiveness of this method was further confirmed using a
special version of WHITE-ECC.

167

Table 3.1 - Mix proportion of materials.


Mix#
ECC_x

Cement
1.0

Silica

Fly

Sand

Ash

0.8

1.2

Water

Fiber,

Superplasticizer

(Vf)
0.66

2.0%

VMA
(%)

0.013%

* Content of PVA fibers is expressed as volume proportion of the mix, while all the other ingredients are
expressed as weight proportion of cement content. Seven mixes were investigated, with x = 0, 0.01, 0.015,
0.02, 0.025, 0.03 and 0.04.

168

Table 3.2 - Material fresh and hardened properties.


Marsh Cone Flow Time (s)
ECC_0

. 9

Fiber Dispersion Coefficient

0.5

0.42

0.38 0.39

0.29

Tensile Strain Capacity (%)

0.44

0.06

0.13

0.09

0.03

0.57

0.53

0.54

1.19

0.69

0.52

12

Marsh Cone Flow Time (s)


ECC_0.01%

Fiber Dispersion Coefficient 0.49 0.52


Tensile Strain Capacity (%)

0.34

Marsh Cone Flow Time (s)


ECC_0.015% Fiber Dispersion Coefficient 0.66
Tensile Strain Capacity (%) 1.97

0.32

17
0.55

0.6

0.58

0.65

0.76

0.85

0.75

1.38

Fiber Dispersion Coefficient 0.84 0.77

0.74 0.73

0.83

Tensile Strain Capacity (%)

2.87

3.2

3.0

24

Marsh Cone Flow Time (s)


ECC_0.02%

3.5

Marsh Cone Flow Time (s)


ECC_0.025% Fiber Dispersion Coefficient 0.89
Tensile Strain Capacity (%)

3.35

3.1

30
0.81

0.75

0.83

0.72

3.27

2.94

3.01

2.68
0.67

Marsh Cone Flow Time (s)


ECCO.03%

33

Fiber Dispersion Coefficient 0.84

0.75

0.7

0.8

Tensile Strain Capacity (%)

3.18

2.82

3.12 2.71

3.25

Marsh Cone Flow Time (s)


ECC_0.04%

39

Fiber Dispersion Coefficient 0.88

0.79

0.82

0.80

0.77

Tensile Strain Capacity (%)

3.27

3.30

2.83

3.23

3.33

Table 3.3 - Mix proportion of WHITE ECC.


Material

Cement

Sand

Water

Superplasticizer

PVA Fiber,
(Vf)

WHITE
ECC

0.5

0.31

169

0.003

2.0%

1000

800
w

- 600
litniu,inlumu,u,l L

US

400
C
1=

200

Tensile Strain (%)

Figure 3.1 - Variation in tensile stress vs. strain curves of ECC with same mixing
proportion.

j'^a^-jOniSydS

Itt

P(c)

Worse fiber dispersion


lowers a(b) curve

Natural flaw
distribution
Activated flaws
<%

mc

flaw size c

Reduced u 0 increases c

Figure 3.2 - Effect of fiber dispersion on o(5) curve and cmc.

170

Mix fresh materials


Seven mixtures with increased VMA content
Same mix proportion, mixing equipment, speed, sequence and time

Measure mortar (w/o fiber) Fresh Properties


- Plastic Viscosity (Pa.s)
- Marsh Cone Flow Rate (s)
- Mini Slump Diameter (cm)

Measure composite (w/ fiber) Fresh Properties


- Mini Slump Diameter (cm)

Cast five tensile plate specimens for each of the seven mixtures

I
Demold specimens after 24 hours, moisture-cure for 7 days, aircure for 21 days

Measure specimen tensile stress-strain curve at age of 28 days

Cut small specimens from the final failure section, measure fiber
dispersion coefficient using fluorescence imaging technique

Figure 3.3 - Experimental program.

70 mm
(2.8 in.)

*\

60 mm
(2.4 in.)
300 mm
(11.8 in.)

20 mm
(0.8 in.)

(a)

100 mm
(4 in.)

25 mm
(1 in.)

(b)

(c)

Figure 3.4 - Rheology properties tests: (a) Viskomat NT rotational viscomter (b) marsh
cone flow rate test (c) mini-slump test.

171

Figure 3.5 - Specimen preparation for measuring fiber dispersion at the final failure
section.

172

Sample

x2 objective

Mercury Lamp

Excitation
Filter

Beam
Splitter

flllllMIMMiflie
Emission
Filter

CCD
Camera

Image
Output

Figure 3.6 - Fluorescence imaging technology.

7 . * " * < * i

Mercury
Lamp

4 |
j -f

1
WKmr-

Lamp Ignition
"Device
Figure 3.7 - Fluorescence imaging test setup.

173

45

40

~Z 35
E
H 30
3
25
g 20 I
o

ECC 0.025%

ECC 0.03%

O
ECC 0.02%

y = 2.3737x + 5.9715
R2 = 0.9546

ECC 0.015%

15"
S 10
5

ECC 0.01%
ECC 0

0
6

10

14

12

16

Plastic Viscosity (Pa.s)

Figure 3.8 - Marsh cone flow time versus plastic viscosity.

b 45 FlowD iameter

40
35
30
25
20
15
10
=
vi
5
E
0
;=
5

'
"
~
"

* * " *

^"~~""'"

'f-'"-"*<

o\o

<?>

<

.o\o

Mix it

'

<?

o\o

"c*
16 :=.
14
12
10
8
6 u.
4
.=
2 J.
0
c

1
1

Ow/o fiber

//fiber

Figure 3.9 - Mini-slump flow diameter of the seven ECC mixes before/after adding fibers.

174

ECC_0

$<L.

V^

*N

\ .

- 800
700
600
500
400
300
200
100
0

6
-2T

&
2

I1

tA*^n^V

11

l 11
0

800
700
600
500
400 5!
- 300
200
- 100
<-^^.^
0

ECC 0.015%

1!

tflflf
0

W,WL-.YM.

.^^f-r.<.

^,^.^^,..^w

2
3
Tensile Strain (%)

Tensile Strain (%)

Tensile Strain (%)

- 800
700
600
500
400
" 300
200
100
0

ECC 0.01%

5 -

II

ECC_ 0.02%

5 4

800

600
500
400
300
200
100
)! 0

1
0

i 700

2
3
Tensile Strain (%)

ECC_0.025%

800
700
600 3
500
400
300 !
200 $
100

1 1.

1I

'

2
3
Tensile Strain (%)

ECC_0.04%
^

ijrfrfS&ffiiffiffi^^fy

3
2
1
0

0
0

ECC_0.03%

5 -

Tensile Strain (%)

800
700
600
500
400
300
200
100

-o
Tensile Strain {%)

gure 3.10 - Tensile stress versus strain curves of ECC with different VMA content.

175

800
700
600
500
400
300
200
100
0

0.01

0.02

0.03

0.04

0.05

VMA/Cement Ratio (%

Figure 3.11 - Tensile strain capacity vs. VMA/cement ratio (error bars describe standard
error of the data).

Tensile Strength (MPa

5.5

5.0 4.5 4.0 3.5 3.0


()

0.01

0.02

0.03

0.04

0.05

VMA/Cement Ratio (%)

Figure 3.12 - Tensile strength vs. VMA/cement ratio (error bars describe standard error
of the data).

176

mti&GfiBW

pin
MP

HEBl'
BlHBI^ilHiNB 1

HBBHjl
12.7 mm
(0.5 in.)

Figure 3.13 - Fluorescence image of a sample specimen.


177

4-
<*9tia* "

f 3.0
re

Q.
TO

2.0

're
A

=
c

1.0

,.A A"

0.0
0.2

0.4

0.6

0.8

(a)

Fiber Dispersion Coefficient

1.0
0.9
0.8
S 0.7
u 0.6
.2 0.5

*>

ffici

OJ

0.4
^^
. 0.3
Q
^
0.2
0)
5LL. 0.1
0.0

o
$
o

i+

10

20

30

40

50

(b)

Marsh Cone Flow Time (s)

4.0
^9

>'u
ra

I
X

3.0

B
D

Q.

re
u
're

2.0 -

i
+

i/>

1.0
c

0.0
10

20

30

Marsh Cone Flow Time (s)

40

50

(c)

Figure 3.14 - Relation between marsh cone flow time before adding fiber, fiber
dispersion coefficient, and tensile strain capacity.

178

'-cMV&A^-u

^ s } ^ ^ ^ - ^
Control fresh
rheology

OM^VMA;
1

(b)

Strain (%)

Control fresh
rheology

0.1%VMA
1

;(c)

Strain (%)

Figure 3.15 - Effect of rheology control on tensile properties of WHITE ECC: (a) A
version of WHITE ECC - "MMX-ECC v4" with 0 VMA content and Marsh cone flow
rate of 12 s in fresh state; (b) "MMX-ECC v4" with 0.05% VMA content and Marsh cone
flow rate of 28 s in fresh state; (c) "MMX-ECC v4" with 0.1% VMA content and Marsh
cone flow rate of 38 s in fresh state.

179

References:
1

Li, V. C , "Integrated Structures and Materials Design," RILEM Journal of Materials


and Structures, Jan. 2006, pp. 1-10.
2

Ozyurt, N., Mason, T. 01, and Shah, S. P., "Correlation of fiber dispersion, rheology
and mechanical performance of FRCs," Cement and Concrete Composites, Vol. 29, No.
2, Feb. 2007, pp. 70-79.
3

Chung, D. D. L., "Dispersion of Short Fibers in Cement," ASCE Journal of Materials in


Civil Engineering, Vol. 17, No. 4, Jul./Aug. 2005, pp. 379-383.

Yang, E.H., M. Sahmaran, Y. Yang and V.C. Li, "Rheological Control in the Production
of Engineered Cementitious Composites," ACI Materials Journal, Vol. 106, No. 04, JuneJuly 2009, pp. 357-366.
5

Marshall, D.B., and Cox, B. N., "A J-Integral Method for Calculating Steady-State
Matrix Cracking Stresses in Composites," Mechanics of Materials, No. 8, 1988, pp. 127133.
6

Li, V. C , and Leung, C. K. Y., "Theory of Steady State and Multiple Cracking of
Random Discontinuous Fiber Reinforced Brittle Matrix Composites," Journal of
Engineering Mechanics, ASCE, Vol. 118, No. 11, 1992, pp. 2246-2264.
7

Griffith, A. A., "The phenomena of rupture and flow in solids", Philosophical


Transactions of the Royal Society of London, Vol. A 221, 1921, pp. 163-198,
http://www.cmse.ed.ac.uk/AdvMat45/Griffith20.pdf.
8

Wang, Y., Backer, S., and Li, V.C, "A Statistical Tensile Model of Fiber Reinforced
Cementitious Composites," Journal of Composites, Vol. 20, No. 3, 1990, pp. 265-274.
9

Grace Chemical Company, www.grace.com, accessed on August 14, 2009

10

Bingham, E. C. Fluidity and Plasticity. New York: McGraw-Hill, 1922.

11

Schleibinger Testing Systems, http://www.schleibinger.com. Accessed on August 13,


2009.
12

Banfill, P. F. G., "The Rheology of Fresh Mortar," Magazine of Concrete Research,


Vol. 43, No. 154, 1991, pp. 13-21.
13

Ferrais, C. F., and de Larrard, F., "Testing and Modeling of Fresh Concrete Rheology,"
Interagency Report 6094, National Institute of Standards and Technology, 1998.
14

Flatt, R., Larosa, D., and Roussel, N., "Linking Yield Stress Measurements: Spread
Test Versus Viskomat," Cement and Concrete Research, Vol. 36, No. 1, 2006, pp. 99109.

180

15

ASTM D6910-04 Standard Test Method for Marsh Funnel Viscosity of Clay
Construction Slurries
16

Roussel, N., and Roy, R. L., "The Marsh Cone: a test or a rheological apparatus?"
Cement and Concrete Research, Vol. 35, No. 5, May 2005, pp. 823-830.
17

Clayton, S., Grice, T. G., and Boger, D. V., "Analysis of The Slump Test for On-Site
Yield Stress Measurement of mineral Suspensions," International Journal of Mineral
Processing, Vol. 70, No. 1-4, 2003, pp. 3-21.
18

Stang, H., "Scale Effects in FRC and HPFRCC Structural Elements," High
Performance Fiber Reinforced Cementitious Composites, RILEM Proceedings Pro 30, A.
E. Naaman and H. W. Reinhardt, eds., 2003, pp. 245-258.
19

Bradbury, S., and Evennett, P., "Fluorescence microscopy," Contrast Techniques in


Light Microscopy, BIOS Scientific Publishers, Ltd., Oxford, United Kingdom, 1996.
20

Rost, F. and Oldfield, R., "Fluorescence microscopy," Photography with a Microscope,


Cambridge University Press, Cambridge, United Kingdom, 2000.
21

Torigoe, S., Horikoshi, T., Ogawa, A., and Saito, T., "Study on Evaluation Method for
PVA Fiber Distribution in Engineered Cementitious Composites," Journal of Advanced
Concrete Technology, Vol. 1, No. 3, 2003, pp. 265-268.
22

Lee, B. Y., Kim, J. K , Kim, J. S., and Kim, Y. Y., "Quantitive Evaluation Technique
of Polyvinyl Alcohol (PVA) Fiber Dispersion in Engineered Cementitious Composites,"
Cement and Concrete Composites, Vol. 31, No. 6, July 2009, pp. 408-417.

23

Spring, R. S., and Davidson, M.W., "Introduction to Fluorescence Microscopy," Nikon


MicroscopyU - The Resource for Microscopy Education,
http://www.microscopyu.com/articles/fluorescence/fluorescenceintro.html., Retrieved on
June 6, 2009.
24

Torigoe, S., Horikoshi, T., Ogawa, A., Saito, T., and Hamada, T., "Study on Evaluation
Method for PVA Fiber Distribution in Engineered Cementitious Composite," Journal of
Advanced Concrete Technology, Vol. 1, No. 3, 2003, pp. 265-268.

181

CHAPTER 4

Effect of Material Ductility on Restrained Volume Change Induced Cracking and


Interfacial Delamination

Starting with this Chapter, the focus will shift from "material engineering" to
"repair system durability assessment", as shown in the middle triangle in Figure 1.10.
This chapter addresses one of the major durability concerns in concrete repairs, i.e. repair
cracking and repair/old interfacial delamination due to restrained volume change and lack
of dimensional compatibility.

To begin, the causes and effects of typical concrete

deterioration processes will be discussed. Cracking, especially, is the biggest single


factor affecting the durability of concrete repair because it often initiates a repair
deterioration process and causes premature repair deterioration. Experimental studies
were carried out to characterize free drying shrinkage properties of HES-ECC at various
RHs, and to determine its restrained shrinkage cracking properties using the steel ring test.
Furthermore, in order to fully understand the interaction between the repair and concrete
substrate, experimental, numerical and analytical studies were conducted on a simulated
repair system subjected to drying shrinkage. The influence of repair material ductility, as

182

well as substrate surface preparation, on the performance of the repaired system was
identified. This research departs from the traditional emphasis on isolated repair material
properties testing, and moves toward the age-dependent interaction between the new
repair and old concrete. This research proposes and demonstrates, on a laboratory scale,
a new repair technology that relies on the repair material's ductility and dimensional
compatibility with existing concrete to minimize cracking and interfacial delamination.

4.1

Introduction
Concrete repair is a complex process with unique challenges that differ from those

associated with new concrete construction, as discussed in Chapter 1. The durability of a


concrete repair depends, to a large degree, on the correct choice and application of repair
materials1. The concrete repair must successfully integrate new materials with existing
concrete to form a composite system capable of resisting service loads and various
environmental factors.
Concrete repair failure results from a combination of physical, chemical and
mechanical processes, as illustrated in Figure 4.1. Generally the repair deterioration
process starts with cracking and repair/old interfacial delamination due to restrained
volume change such as drying shrinkage or thermal deformation, when there exists a
dimensional incompatibility between the repair and existing concrete. Cracking and
delamination are insidious causes of many repair durability problems2'3'4'5. The presence
of cracks dramatically alters the transport properties of the repair material6'7'8'9'10'11,
facilitates the ingress of aggressive agents into the repaired system, and eventually causes
premature deterioration and repair failure such as corrosion and disintegration. Large

183

cracks accelerate chloride penetration into the concrete and initiate corrosion of
reinforcements, especially in regions where deicing salts are applied in winter or in
marine environments12'13. Corrosion damage has even been observed on epoxy coated
reinforcing bars in bridge decks with such cracks14'15. In overlay repair applications,
delamination of concrete bridge overlays from the substrate deck is one of two primary
causes of ultimate overlay failure16. The loss of structural integrity impairs load transfer
between the repair and old concrete structure. Both cracking and interfacial delamination
caused by restrained volume change can reduce the durability and load-carrying capacity
of the repaired structure.
The ability of the repair material to withstand restrained volume change without
cracking and interfacial delamination depends on the dimensional compatibility between
17 18

1 Q ")C\ 9 1

the repair material and existing concrete ' ' ' ' .

Table 4.1 lists the properties

generally required of patch repair material to achieve structural compatibility ' . Among
these properties, dimensional compatibility includes strain capacity, drying shrinkage,
coefficient of thermal expansion, creep, modulus of elasticity, and Poisson's ratio. An
ideal repair material should undergo zero shrinkage, and have the same amount of
thermal deformation, creep, modulus of elasticity and Poisson's ratio as the existing
surrounding concrete. However in reality, shrinkage always takes place in a new repair
as soon as it is installed, and is considered as the most adverse factor ' > > > >
among those described above that promotes dimensional incompatibility of cement-based
repair materials.
The shrinkage of a newly installed repair is immediately restrained by the
substrate old concrete, which already underwent shrinkage years ago. Consequently,

184

tensile stress is developed in the repair layer, and a combination of tensile and shear
stresses are built up along the interface between the repair and the concrete substrate. As
schematically illustrated in Figure 4.2, tensile stress at the repair/substrate interface is the
y-direction stress that opens the interface (delaminates the interface in the y-direction).
Shear stress at the repair/substrate interface is the x-direction stress that causes the repair
layer to slide along the surface of the substrate (delaminates the interface in the xdirection).

These stresses can cause repair surface cracking, and/or interface

delamination (Figure 4.2).


A detailed discussion of stress distribution and failure mechanism in a concrete
repair system undergoing drying shrinkage can be found in Wittmann and Martinola,
200329' The immediate moisture loss of the cementitious repair material through its
exposed surface results in moisture gradients along the repair cross section normal to the
exposed repair surface. The highest tensile stresses occur in the exposed repair surface
where the moisture gradients are at a maximum, and the tensile stresses increase as the
drying process proceeds. Depending on the repair material properties and the severity of
the moisture gradients, the maximum tensile stresses can easily surpass the tensile
strength (or the maximum tensile strains surpass the tensile strain capacity) of the repair
material. As a result, cracks will first occur at the repair surface, and propagate further
down through the repair cross section as the drying process proceeds. Simultaneously,
the surface crack enlarges. The moisture gradients will also result in the tendency of
warping of the repair layer at both ends, causing a combination of tensile and shear
stresses at the repair/old interface and potential repair delamination.

185

It should also be noted that the interaction between the new repair and the existing
concrete is a time-dependent process30'31.

Repair material shrinkage, creep, elastic

modulus, tensile strength, and the repair/old interfacial bond are all age-dependent
properties, and develop at different rates once the new repair is installed. Cracking and
interfacial debonding time and extent are determined by competition among these agedependent parameters.
In the past decades, there have been significant efforts on reducing repair
shrinkage cracking through improving repair material properties such as: (a) reducing
material free shrinkage, (b) increasing compressive/tensile strength, and (c) enhancing its
bond strength with concrete. Reduced cement-based repair material shrinkage is often
achieved by using shrinkage control admixture, expansive cement, short random fibers,
lower water/cement ratio, proper and thorough compaction to eliminate air bubbles, large
coarse aggregates, or fly ash. In terms of the curing and construction methods, an earlier
and longer curing time, the use of sprayed liquid aliphatic alcohol as an evaporation
retardant onto the repair surface upon finishing, the use of effective curing procedures
including the application of a single pack, membrane (thick) liquid ready-to-use curing
compound, and early (sometimes within 8 hours) saw cutting of joints, are important
procedures in reducing shrinkage cracking. Increased compressive/tensile strength of
cement-based repair materials, e.g. high strength concrete, is typically achieved through a
very dense microstructure by employing low water/cement ratios32, silica fume and fly
ash33, and a well-graded particle size distribution34.
These efforts, although meaningful, have resulted in only marginal improvements.
Cracking due to restrained volume change is still a prevalent phenomenon, and remains

186

as a big challenge to the repair industry.

For example, high strength concrete is

particularly more susceptible to cracking due to its high brittleness, high elastic modulus,
and small creep deformation35. Secondly, containing more finely ground particles than
regular concrete, high strength concrete typically has increased autogeneous shrinkage
and thermal volume changes. Finally, increasing repair material strength might delay
repair cracking, as long as the restrained shrinkage induced tensile stress remains below
the material tensile strength over time. However, it poses an elevated risk of interfacial
debonding because higher levels of shear and tensile stress build-up at the new/repair
interface, which otherwise could have been relieved through cracking in the repair layer.
On the other hand, accelerating interfacial bond development may delay the debonding
process, but makes the repair more prone to cracking because debonding reduces the
level of constraints and partially relieves the tensile stress in the repair. Reducing repair
material shrinkage appears to be the most effective method. However, most efforts can
only reduce the shrinkage to slightly below 0.01%. This puts a significant demand on the
tensile capacity (about 0.01%) for most of the currently available concrete repair
materials. In reality, shrinkage is not the only factor that contributes to the deformation
of the concrete repair. When thermal effects, structural loading, and other factors are
combined, it is highly possible that the total imposed deformation of the repair will
surpass its deformation capacity, and cracking and/or debonding will inevitably follow.
The limitations in current approaches to reducing concrete repair failure are that
they do not fundamentally address the inherent brittleness of concrete materials. With
such brittleness, the potential of preventing cracking by reducing concrete volume change
to a level less than the material's deformation capacity is low and unreliable5'36'37. These

187

shortcomings point to an urgent need for developing a new material technology that
effectively reduces the cracking potential through removal of material brittleness, and
eventually eliminates restrained volume change induced cracking and interfacial
delamination in concrete repairs.

4.2

Proposed Approach
To ensure dimensional compatibility within the repaired system, the repair

material should possess deformation capacity that fully accommodates the imposed
deformation due to a combination of factors, without debonding or fracturing locally.
These criteria should be satisfied at all ages throughout the repair's life. In this sense, a
repair material can be considered truly dimensionally compatible with existing concrete if
the following characteristics are satisfied at all ages: (a) high tensile train capacity, (b)
adequate bonding with concrete, (c) similar thermal coefficient and creep as concrete, and
(d) same or lower elastic modulus as concrete.
Li and Stang8 illustrated the effect of the inelastic tensile strain capacity of
cementitious material on the cracking behavior of a 2-D slab geometry of length L,
restrained at both ends. For brittle or quasi-brittle repair material with tension softening
behavior, the cracking potential under restrained shrinkage is defined as:
p = (ssh-(se+scp))

(4.1)

where sSh is the material's shrinkage strain, ee is its elastic tensile strain capacity, and scp
is its tensile creep strain. A high positive value of p signifies a strong potential for
cracking due to restrained shrinkage. In this case, a single crack forms in the repair
material with crack width w proportional to the cracking potential p

188

w = L((esh-(ee + Ecp))/(l-L/2lch)
w

= L^sh 'cP)f

for(ee+ecp)zesh<;wc/L

, *C/L

(4.2)
(4.3)

where w is crack width, lch is Hillerborg's material characteristic length9:

The variables E, Gf, ot, wc stand for the material's Young's modulus, fracture energy,
tensile strength, and critical crack opening. A linear tension-softening law is assumed
where strength retention decreases from at to 0 as the crack width opens from zero to wc.
Equation 4.2 indicates that crack width w depends on the cracking potential p, the degree
of brittleness L/21ch, and the repair dimension L10. For instance, highly brittle material
such as high strength concrete, which has a relatively small characteristic lengths (lch), is
expected to exhibit more severe cracking with larger crack width. Like other brittle
concrete or quasi-brittle fiber-reinforced concrete materials, the crack width of high
strength concrete will increase with increasing structural dimensions.
For a repair layer, the boundary conditions are different than those described
above. Restraints are applied at the base of the slab rather than at the ends. This
difference results in a number of distributed cracks along the repair layer instead of one
single crack. For most of the repair materials that are brittle, traction-free cracks will
open with a crack width proportional to the cracking potential p. Through the formation
and free-opening of these surface cracks, stresses formerly built-up in the repair layer and
especially at the interface, can be relaxed.

By these means, interface delamination

tendency is greatly reduced.


Fibers have been used in concrete repairs to control drying shrinkage and service
load-related cracking. For tension-softening FRC materials, shrinkage induced stresses

189

can result in surface cracking similar to concrete. However, since the cracks are not
traction-free because they are bridged by fibers, some amount of tensile and shear
stresses remain in the repair layer. As a result, interface delamination can be more
prominent in FRC repairs than in concrete or mortar repairs. This phenomenon was first
observed by Kabele38 through a numerical study of a similar repair system.
To suppress both repair surface cracking and repair/old interface delamination,
the repair material needs to exhibit "inelastic straining" to accommodate its shrinkage
deformation, thus relieving the stresses built-up under restrained drying shrinkage
conditions. By these means, surface crack width and interface delamination can both be
minimized. Inelastic straining in the form of microcrack damage has been demonstrated
in newly developed HES-ECC in Chapter 2. This material has been optimized for high
inelastic tensile strain capacity s; and tight microcrack width at minimum fiber content.
For HES-ECC, the cracking potential10 is modified as
(4-5)

/> = ( * , - ( * , + , + < )

The formation of many fine microcracks in HES-ECC can be regarded as inelastic


straining on the macroscopic repaired-system-size scale. The large value of s; (>3%) for
HES-ECC both in early and late ages results in a highly negative cracking potential p in
Equation 4.5. A negative cracking potential p indicates that localized fracture due to
restrained shrinkage will be inhibited. Instead, the repair material will undergo inelastic
straining by developing multiple microcracks with controlled crack width. Therefore,
HES-ECC is suggested as a promising material for durable repair jobs by minimizing
repair surface cracking and interface delamination.

190

4.3

Free Shrinkage of HES-ECC

4.3.1

Background
Drying shrinkage refers to the strain caused by the loss of water from the

hardened cementitious material39.

It is a cement paste property; aggregates have a

restraining influence on the volume change that takes place within the cement paste.
Proposed mechanisms for causes of drying shrinkage of cement paste include: (a)
capillary stress, (b) disjoining pressure, and c) surface tension. The importance of these
mechanisms depends on the relative humidity (RH).
Capillary Stress
Within a small capillary pore in the hardened concrete cementitious matrix, water
is partially under the influence of surface interactions exerted by the pore walls39.
Therefore, the water cannot evaporate unless RH is lowered by an amount depending on
the pore radius (r), given by the Kelvin equation :
\n(RH) = K(^)
r

(4.6)

where y is the surface free energy (surface tension) of the water, and AT is a constant with
units of (m3/J).
The Kelvin equation40 describes the maximum radius of pores that are saturated
with water as a function of RH and time t. Pores in a cementitious composite are water
saturated even when the water vapor partial pressure is lower than the saturated water
vapor partial pressure, due to the effect of capillary condensation. As illustrated in Figure
4.3, at and below 40% RH all pores are empty according to the Kelvin equation due to
the fact that no pores with a radius smaller than 1 nm exist in a cementitious material .
For RH below 40%, the removal of hydrostatic stress should be accompanied by a

191

relaxation in the solid and an increase in length. Although such behavior is observed in
materials with a rigid skeleton (e.g., porous glass), it is obscured in cement paste by the
slower concomitant loss of water from micropores (subsequently discussed).
Disjoining Pressure
Water is adsorbed on the surfaces of C-S-H, and the thickness of the adsorbed
water layer increases with increasing humidity. The assembly of colloidal particles in CS-H attracts adjacent particles through Van der Waals' forces and brings the adjacent
surfaces closer.

Adsorption of water between C-S-H surfaces creates a disjoining

pressure, which is the combination of steric force and electrostatic force.

Disjoining

pressure increases with increasing relative humidity and thickness of the adsorbed water
between particles.

Once surpassing the Van der Waals' attractions, it will force the

particles apart and result in a dilation of the composite material. During hydration, C-SH is formed in the dilated state and the micropores are filled with water. Upon first
drying, the decreasing disjoining pressure, due to a lowered relative humidity, will cause
the particles to be drawn together by the Van der Waals' forces, resulting in a volume
decrease. Disjoining pressure is only a significant factor when RH > 40%.'
Surface Free Energy
When RH < 40%, capillary stress and disjoining pressure are no longer operating;
drying shrinkage is explained by significant changes in surface free energy as the most
strongly adsorbed water is removed. The change in solid volume with changing surface
energy y can be described by the Bangham equation:
y = *AX

(4.7)

192

where A/// is the observed relative length change, Ay is the change in surface free energy
during adsorption, and & is a constant. Ay is dependent on RH, and can be expressed
42,43,44

as

m)

R-T *u

^ -ir-AM-AJP
hip

(4g)

Where R is the universal gas constant, T is absolute temperature, M is mol volume of


water, A is specific surface area, P is water vapor partial pressure that depends on RH,
and u is adsorbed water volume.
Hygral deformation at a given RH can be experimentally measured as the
difference between the steady state drying shrinkage value of the dry state, and the steady
state drying shrinkage value at the given RH.

The total hygral deformation of a

cementitious material can be described by the Munich model, which was originally
developed for cement paste45'46, and later found to also work for concrete47 and an
ordinary ECC (M45) , whose cementitious matrix contains Portland type I cement,
ASTM class F fly ash, and fine silica sand. This model assumes that the capillary stress
is approximately independent of RH49,50. Therefore, the total hygral deformation can be
described as the deformation due to a change in surface energy of C-S-H gel particles,
plus the deformation due to the disjoining pressure between C-S-H gel surfaces when RH
> 40%.

hyg,totaiyKn) =

hyg,surf ace-energy\Ktl)

+ hyg,disjoining-pressure\^^

>

4")

(4-")

According to Equation 4.7, the first term of Equation 4.9 can be expressed as:
%g,Surfa^nergy(RH) = A' AY(RH)

(4.10)

where X is a function of the stiffness of the porous material51.

193

The second term of Equation 4.9 can expressed as:

hyg,disjoininzrpressure

( ^ x

) =

' (^#x

~ RH^f

' (RH

x ~ ^#40

for RH 2 RH40 = 40%

(4.11)

where c and d are empirical parameters.


Combining Equations 4.8, 4.9, 4.10, and 4.11, the total hygral deformation can be
rewritten as:
hyg,,olai = A {A \n(RH)

B) + c-(RHx-RH40)2+d-(RHx-

RH40)

when RH>RH4o = 40%


ehyg,tolal = *.-(A-ln(RH) + B)
when RH<RH40 = 40%

(4.12)

In this study, the total hygral deformation of HES-ECC at different RHs will be
experimentally determined to characterize its free drying shrinkage properties.
Additionally, by fitting Equation 4.12 to the experimental data, the parametric values of A,
A, B, c, and d can be determined for the physical model that describes the hygral
deformation of HES-ECC.

4.3.2

Materials and Experimental Procedure


HES-ECC and HES-Concrete (control mix) are investigated in this study. Their

mixing proportions and mechanical properties are summarized in Table 4.2, 4.3 and 4.4.
The HES-Concrete mixture consisted of coarse aggregate (CA) with 10 mm (0.4 in)
nominal grain size, Portland type III cement (C), sand (S) and water (W).
Superplasticizer (SP) was used to achieve sound workability. Accelerating admixtures
(AC) were also included to accelerate the material's strength development and setting
194

processes. HES-Concrete specimens were tested to have average compressive strength


(fc') of 49.9 MPa (7234 psi) at the age of 7 days, and 54.2 MPa (7860 psi) at the age of
28 days. Under uniaxial tensile loading, HES-Concrete is a brittle material with sudden
fracture failure on a single crack plane.
The HES-ECC mixture, developed in Chapter 2, is comprised of Type III Portland
cement (C), water (W), silica sand (S) with 110 ^im (3.94* 10"3 in) nominal grain size,
polystyrene beads with a size of 4 mm (0.157 in) as aggregates, superplasticizer,
accelerating admixtures, and 2% (Vf) polyvinyl-alcohol (PVA) fibers. These PVA fibers
had length of .12 mm (0.472 in) and diameter of 40 \xm (1.54xl0"3 in). The HES-ECC is
self-consolidating in the fresh state.

The polystyrene beads serve as deliberately

introduced initial flaws to assist in triggering the multiple microcracks6 during tensioning
of the composite.

As measured in Chapter 2, the HES-ECC mixture had average

compressive strength of 47.5 MPa (6885 psi) at the age of 7 days, and 55.6 MPa (8063
psi) at the age of 28 days. Its average Young's modulus was 20.6 GPa (2986 ksi) at the
age of 7 days, and 23.2 GPa (3365 ksi) at the age of 28 days, which was lower than that
of HES-Concrete (26.2 GPa (3803 ksi) at 7 days; 27.8 GPa (4025 ksi) at 28 days) due to
the absence of coarse aggregates (CA). A lower modulus repair material is desirable in
that it limits the tensile stress induced by restrained drying shrinkage. The HES-ECC
mixture also has a tensile strain capacity between 3 - 6% at all ages.
Free drying shrinkage tests were conducted on HES-ECC and HES-Concrete
(control) prism specimens with dimensions of 200 mm (7.9 in.) x 40 mm (1.6 in.) x 40
mm (1.6 in.), in accordance with ASTM C157/C157-99 and C596-0152 standards. The
test setup is shown in Figure 4.4. Twenty-one specimens for each mix were cast and

195

demolded after 24 hours. After storage at 100% relative humidity (RH) for two days, the
specimens reached hydral equilibrium. Three specimens of each mix were then stored in
RH = 0%, 12%, 33%, 66%, 75%, 85%, and 93% by using saturated salt solutions in
different desiccators53'54. The drying shrinkage deformation of the HES-ECC specimens
was then measured as a function of drying time, until hygral equilibrium was reached.
After all specimens had reached hygral equilibrium, the specimens stored at 0% RH were
dried in an oven at 105 C (221 F) until a constant mass was achieved. The specimens
were then removed from the oven and allowed to cool down to room temperature 201
C (66-70 F). After the specimens reached thermal equilibrium with room temperature,
the length changes were determined as the shrinkage values at RH = 0%.

4.3.3

Experimental Results and Discussion


The hygral deformation of HES-ECC and HES-Concrete is calculated as the

difference between the steady state drying shrinkage value in the dry state (0% RH) and
the steady state drying shrinkage value at specific RHs. These data are plotted as a
function of RH in Figure 4.5, together with the measured hygral deformation values of an
ordinary ECC (ECC 45) and a concrete mix (both are Portland type I cement based) from
Weimann and Li, 200348. The parametric values of HES-ECC, HES-Concrete, ordinary
ECC and concrete

are summarized in Table 4.5. The good fit between theoretical

(Equation 4.12) and experimental data demonstrated that the drying shrinkage of HESECC can be explained by the contraction due to the increase in surface free energy of CS-H particles, as well as the decrease in disjoining pressure between the surfaces of C-SH, the latter mechanism only operating above 40% RH. These results clarified that the

196

drying shrinkage behavior of HES-ECC is governed by the same mechanisms as concrete


materials and Type I cement based ECC M45.
The experimental results reveal that the drying shrinkage of HES-ECC is
approximately twice that of the HES-Concrete control mix, and 15% higher than the
ordinary ECC. This is due to the very high cement content, finer cement particles, and
absence of large volume of coarse aggregates and fly ash in the HES-ECC mix.
Nonetheless, Figure 4.5 shows that the shrinkage strain of HES-ECC is below 0.3%,
which is one order of magnitude lower than its tensile strain capacity 3-6%. This implies
that when the drying shrinkage of HES-ECC is restrained, the material's ductility can
accommodate shrinkage deformation by forming multiple microcracks at its strain
hardening stage without localized fracture failure.

Once microcracked, the effective

modulus of the HES-ECC will be substantially reduced, as seen in the reduced slope of
the tensile stress-strain curve in the strain-hardening stage (Figure 2.27), thus further
suppressing tensile stress build-up in the repair material due to continued restrained
shrinkage. This predicted behavior is verified in further experimental studies where
HES-ECC drying shrinkage deformation is restrained. These investigations are described
in the next sections.

4.4

Restrained Shrinkage Ring Test

4.4.1

Materials and Experimental Procedure


Due to its simplicity and versatility, the ring test has become often used over

the last two decades to measure restrained shrinkage cracking of concrete materials and
fiber reinforced cementitious composites 55 ' 56 ' 57 .

197

It was adopted in this study to

investigate the number and width of shrinkage-induced cracks in HES-ECC compared to


a control HES-Concrete with the mixing proportion shown in Table 4.2.
The restrained shrinkage ring test followed AASHTO58 PP-34.

Figure 4.6

shows the test set-up and specimen dimensions. For each specimen, a 25.4 mm (1 in.)
thick layer of HES-ECC or HES-Concrete material was cast around a rigid steel ring. A
plastic covered paper cylinder was used as an outer mold during casting. The outer mold
was removed 6 hours after casting and the specimen was then exposed to 45 5% RH
and 201 C (66-70 F). It should be noted that the curing conditions for the ring tests
are different from the free drying shrinkage test, which was described earlier in Section
4.3. The purpose of the early demolding and short (6 hours) curing time in the ring tests
is to simulate the early re-opening conditions in field application, while the free shrinkage
tests employ curing conditions based on the ASTM C157/C157-99 and C596-01
standards to characterize HES-ECC shrinkage strain at various RHs.
Once exposed to the ambient relative humidity and temperature conditions, the
drying shrinkage deformation of the HES-ECC and HES-Concrete, when restrained by
the steel ring, results in internal radial pressure. Consequently, the HES-ECC and HESConcrete layers are subjected to a circumferential tensile stress state that can cause
cracking. The cracking pattern, crack number and crack width were measured as a
function of age with a portable microscope. Measurements were taken at three different
locations along each crack and the average value was plotted. Three HES-ECC and three
HES-Concrete specimens were tested.

198

4.4.2

Experimental Results and Discussion


Test results are summarized in Table 4.6 and Figure 4.7. For the three HES-

Concrete control specimens, one localized crack was observed in each specimen. The
average crack width of the three specimens was 310um (0.012214 in.) at 2 days, 610 um
(0.024034 in.) at 4 days, 1210 um (0.047674 in.) at 28 days and 1250 um (0.049250 in.)
at 60 days. For each of the three HES-ECC specimens, 25, 23, and 19 cracks were
formed at 60 days, respectively. The average crack width was 40 um (0.001576 in.) at 2
days, and remained at 50 um (0.00197 in.) for the remaining 2 day to 60 day test period.
The crack width of these HES-ECC microcracks was significantly lower than the crack
width of HES-Concrete localized cracks, and the former retains its load carrying capacity
after crack formation. Wang et al.59 reported that as crack width increases from 100 urn
to 500 um, the permeability coefficient increases nearly seven orders of magnitudes from
1.0 x 10"u m/sec to 1.0 x 10"4 m/sec. However, for crack widths under 100 um, the
permeability coefficient remains nearly identical to that of uncracked concrete,
suggesting that for crack widths below this threshold there is no significant increase in
permeability after cracking. The test results show that HES-ECC has significantly greater
resistance to restrained shrinkage cracking than HES-Concrete, despite its higher free
(drying) shrinkage value. This is due to the negative shrinkage cracking potential of HESECC, in contrast to the positive shrinkage cracking potential of HES-Concrete. This point
will be discussed in detail in the next section.
The shrinkage deformation process in HES-ECC occurs during the strainhardening stage, and the shrinkage crack width is an intrinsic material property
independent of structural dimensions or the reinforcement ratio.

199

Additionally, the

shrinkage crack width of HES-ECC becomes independent of age (drying time) once the
material enters the strain-hardening stage. Further shrinkage deformation of HES-ECC
will be accommodated by forming more microcracks with controlled crack width under
50 urn (0.00197 in.), instead of increasing the width of existing cracks (Figure 4.7). In
contrast, continued shrinkage deformation of HES-Concrete with age is accommodated
by forming of localized cracks with increasing crack width. The crack width of HESConcrete is not only related to age (drying time), but also highly depends on structural
dimensions - larger repair members tend to have larger shrinkage deformation that needs
to be accommodated by crack opening, as experimentally demonstrated by Li and Stang
(2004). Thus, while the restrained shrinkage tests in this study demonstrate a 25 fold
reduction in the 60-day shrinkage crack width in HES-ECC compared with that in HESConcrete, the advantage of ECC can be even more significant in field repair applications.

4.5

Restrained Shrinkage Test on a Simulated Repair System


The lack of widely agreed upon testing methods leaves repair materials subject to

limited evaluations driven by manufacturers rather than by users5. Frequently, only the
isolated properties of repair materials are emphasized, whereas the more important
properties of the composite system are neglected. In this study, the restrained shrinkage
test was conducted on a simulated repair system, which contains one layer of old concrete
substrate and one layer of new repair material to simulate real repair conditions. The
performance of the composite repair system as well as the interaction between the repair
and concrete substrate were investigated.

200

4.5.1

Materials
Three different repair materials HES-Concrete, HES-Steel Fiber Reinforced

Concrete (HES-SFRC) with tension softening stress-strain curve, and HES-ECC - were
investigated in this study.

The HES-Concrete and HES-SFRC were employed as

controls. All three materials use Portland type III cement to achieve high early strength.
Their mixing proportions are summarized in Table 4.2. Six specimens of each mix were
tested to achieve the material mechanical properties reported in Table 4.3.
The HES-Concrete and HES-ECC mixes were the same as those investigated in
the free shrinkage test and restrained shrinkage ring test. The HES-SFRC mix had the
same composition as HES-Concrete, except that it contained 1% (Vf, volume fraction)
steel fibers. The steel fiber, with length of 30 mm (1.18 in) and diameter of 0.5 mm (0.02
in), had a smooth surface and hooked ends. The averaged 7-day and 28-day compressive
strength of the HES-SFRC specimens were 51.5 MPa (7462 psi) and 56.9 MPa (8254
psi).

Under tensile loading, HES-SFRC is a quasi-brittle material with a tension

softening stress-strain response. The ultimate tensile strain capacity (su) of both HESConcrete and HES-SFRC is around 0.01%. The substrate concrete has the same material
composition, except that it used ordinary concrete containing Portland type I cement
while the HES-Concrete repair used Portland type III cement.

4.5.2

Repair Specimen Configuration and Surface Preparation


In this study, layered repair systems were experimentally investigated with the

three types of repair materials described above; HES-Concrete, HES-SFRC and HESECC. Concrete substrates were initially cast with dimensions of 1600 mm x 100 mm x

201

100 mm (63"x 4"x 4"), as shown in Figure 4.8. They were moisture-cured until the age
of 28 days, and then left to dry in ambient conditions for an additional 150 days before
the repair layers were placed.

The additional 150 days allowed for any potential

shrinkage in the substrates to occur before bonding the repairs. This procedure was
followed to simulate the actual repair conditions of actual repair of old concrete
structures.
The contact surfaces of the concrete substrates were roughened while the material
was still in a fresh state by using a chisel to remove slurry cement from the external
surfaces of coarse aggregates. The estimated roughness amplitude was 3.8 mm to 5.1
mm (0.15 to 0.20 in). Before placing the repair layer, the substrate surface was recleaned with a brush and high-pressure air to ensure a clean bonding surface, and was
then dampened with water fog. The moisture level of the contact surface was critical to
achieve bonding; excessive moisture in a contact surface may clog the pores and prevent
absorption of the repair material. On the other hand, an overly dry substrate contact
surface may absorb water from the repair material, resulting in an undesirable magnitude
of shrinkage60. After dampening the surface, a 50 mm (2 in.)-thick repair layer made of
each of the three repair materials was cast on top of the concrete substrate. The repair
layers were moisture cured for 6 hours and then demolded. After demolding, the layered
specimens were moved into a room with ambient conditions of 45 5% RH and 201 C
(66-70 F).
For each specimen, two dial gauges (Figure 4.8) were used to record the interface
vertical separation distance at the ends of the specimens as a function of drying time after
delamination began.

Additionally, a portable microscope was used to measure the

202

delamination at 30 different locations along the interface, from which the delamination
crack profile was derived. The microscope was also employed to observe crack pattern,
crack number and crack width of the top surface of the repair layer, as a function of age.
Both delamination and surface cracking were measured daily.
Free shrinkage testing was also carried out in order to measure the free shrinkage
strain of the HES-Concrete, HES-SFRC and HES-ECC mixtures for the calculation of
cracking potential p. The free shrinkage test specimens were made from the same batch
as the repair layer mix for each of the three repair materials. Differing from the free
shrinkage test in Section 4.3, these specimens were domolded as early as after 6 hours,
and shrinkage began. The testing environment of the specimens were also modified to be
the same as the layered repair specimens, with ambient condition of 15-21C (60-70F)
and 45 5% RH. The same curing, storing and testing environment were used for
relating the free shrinkage test results to the observed behavior of the layered specimens.

4.5.3

Experimental Results and Discussion

Cracking Potential of Repair Materials


Three specimens were tested for HES-ECC, HES-Concrete and HES-SFRC and
the average free shrinkage strain (sSh) values are summarized in Figure 4.9. The data
show the shrinkage strain of each material at the age of Id, 3d, 7d, 28d, and 60d. HESECC has the highest shrinkage strain value because of higher cement content and the
absence of large coarse aggregates, while the HES-SFRC has the lowest shrinkage strain
value because of the constraining effect of steel fibers.

203

The cracking potential p for HES-Concrete, HES-SFRC and HES-ECC can be


estimated based on measured values of sShand s; according to Equation 4.1 and 4.5, and is
summarized in Table 4.7. The other parametric values (se and scp) were not measured in
this study, but were adopted by estimation as the se and scp of normal concrete and SFRC
from Li61. Although HES-ECC had the highest free shrinkage, its negative p-value
verifies that it had entered the strain hardening stage under restrained drying shrinkage,
and will experience microcracking damage without fracture localization. In contrast,
HES-Concrete and HES-SFRC are subject to tensile fracturing due to their positive pvalues.
It should be noted that the cracking potential p was calculated assuming a
boundary condition that is generally different from what repairs experience within
structures. In repair conditions, restraints are applied at the base of the repair and the
restraint level might be lower, depending on the stiffness of the substrate, presence of
reinforcements, boundary conditions, etc. However, the calculated p values in Table 4.7
provide revealing insight on the significant influence of repair material ductility (inelastic
strain capacity) on the repair cracking potential. Most of the currently available concrete
repair materials are either quasi-brittle (e.g. concrete) or tension-softening (e.g. fiber
reinforced concrete), therefore having an e; equal to 0. This means that reducing the free
shrinkage esh of these materials is not an effective approach to preventing cracking due to
restrained volume change, because it can only lower the cracking potential by a minimal
amount below zero. When shrinkage is combined with other causes of volume change
such as thermal deformation and deformation due to structural loads, very often the total
strain imposed on the repair, etotai, leads to a positive cracking potential p.

204

( 4 - 13 )

P = O w - ( e e + ec))

This is not the case for a ductile repair material such as HES-ECC, with inelastic strain
capacity one order of magnitude higher than its shrinkage strain. The highly negative
cracking potential of HES-ECC provides a large safety margin for resisting localized
cracking due to restrained volume change.
(414)

/ > - 0 w - ( * . + e, + O )

Surface Cracking and Interface Delamination of Repaired Systems


Table 4.8 summarizes the surface crack pattern, i.e. crack number and crack width,
of the three repaired systems, respectively at the age of 60 days. Three specimens of each
repair material were tested. When HES-Concrete was used as the repair material, 3-4
surface cracks localized at the age of 60 days. The maximum crack width observed for
the three specimens was 490 urn (19.3 x 10"3 in). When HES-SFRC was used as the
repair material, 1-4 localized surface cracks formed, and the maximum crack width of the
three specimens was 280 urn (1 l.Ox 10"3in). The smaller crack width in the HES-SFRC
repair can be attributed to the steel fiber's bridging effect.

Note that the restrained

shrinkage induced crack width for HES-Concrete or HES-SFRC repair is a structural


property, which is dependent on structural dimensions.
In contrast, when HES-ECC was used as the repair material, 83-113 microcracks
were formed on the surface of the repair layer, with a maximum crack width of 60 |am
(2.36x 10" in) - much narrower than the cracks found in the HES-Concrete and HESSFRC repairs. The average crack width of the HES-ECC repair was around 30 um
(1.18x10" in). No localized fracture was observed. The shrinkage strain of HES-ECC

205

(less than 0.3 %) was far below its tensile strain capacity of 3-6 %. Therefore, the
restrained shrinkage cracking of HES-ECC occurred in its strain-hardening stage, during
which the material formed multiple microcracks with steady state crack width. In this
deformation stage, the restrained shrinkage crack width of the HES-ECC repair was selfcontrolled, and is a material property independent of structural dimensions. Even for
larger scale repair applications with different restraint conditions, HES-ECC repair is still
expected to exhibit tight crack width, below 60 urn (2.36x 10~3 in). This tight crack width
should greatly reduce transport of aggressive agents by permeation10 and chloride
penetration11. Figure 4.10 shows the surface-cracking pattern of each type of repair layer.
HES-Concrete and HES-SFRC repairs exhibited localized fractures, in contrast to the
multiple microcracks of HES-ECC repair with significantly smaller crack width.
The measured interface delamination heights and lengths for the nine layered
repair specimens are summarized in Table 4.8.

As illustrated in Figure 4.11, the

delamination height is the distance between the crack faces of the crack running along the
repair-substrate boundary, at the two ends of the repaired system; the delamination length
is the length of the delaminated section along the long axis of the specimen. At the age
of 60 days, both the HES-ECC and the HES-Concrete repaired systems exhibited
relatively low delamination heights at the specimen ends, which were a maximum of 80
um (3.15x 10"3 in) for the former and 90 um (3.54* 10'3 in) for the latter. The maximum
delamination length was 80 mm (3.23 in) for the HES-ECC repair and 170 mm (6.69 in)
for the HES-Concrete repair.

The HES-SFRC repaired system had much larger

delamination height than the HES-ECC or HES-Concrete repaired system at the age of 60
days, which was 310 um (12.2x 10"3 in).

206

Its delamination length was also larger,

measured to be 340 mm (13.46 in). Figure 4.12 shows the interface delamination profiles
of

the

three

repair

systems

at

different

ages,

which

are

the

vertical

displacement/delamination heights of the repair layers at different locations along the


repair/substrate interface. These profiles are approximately symmetric about the midpoint of the specimen, as would be expected.
The interface delamination height development of the three repaired systems as a
function of time is shown in Figure 4.12. It can be seen that HES-ECC and HESConcrete repaired systems completed most of their interface delamination at very early
ages - within 7 days, when surface cracking had already completed and helped to release
tensile and shear stresses at the specimen interface. However, for the SFRC repaired
system, delamination behavior continued to evolve for up to 60 days (Figure 4.12), at
which time the SFRC repair material had undergone most of its shrinkage (Figure 4.9).
This further confirms that the fiber-bridged cracks of HES-SFRC repair could only
release a fraction of the stresses at the interface, so that delamination continued along
with shrinkage. As can be seen in the figure, the HES-SFRC repair exhibited interface
delamination heights several times greater than those of the HES-ECC and HES-Concrete
repairs.

4.6

Effects of Surface Preparation on the Repaired System Performance


The same simulated repair system described in Section 4.5 was adopted in this

section to investigate the effects of surface preparation (interfacial bond) on repair


cracking and interfacial delamination due to restrained drying shrinkage. Three types of
repair material were used: concrete (brittle), FRC (tension-softening) and ECC M45

207

(ductile). These materials do not have the high early-age strength properties, as did the
materials investigated in the last section, but still well represent their material categories
with respect to specific tensile behavior. This study clarified the influence of surface
preparation on the performance of repairs with different tensile behaviors. It also further
validated the effects of repair material tensile ductility on the performance of the repaired
system. It also further validated the effects of repair material tensile ductility on the
performance of repaired system - To prevent restrained shrinkage induced repair
cracking and interfacial delamination, it is the tensile properties of the repair material, not
any specifically designed mixes, that governs.

4.6.1

Experimental Procedure
Three different type I Portland cement based repair materials concrete, steel

fiber reinforced concrete (SFRC) with tension softening stress-strain curve, and ECC
(M45) were investigated in this study. Concrete and SFRC were employed as controls
since they have been used in repair applications and included in the ACI Concrete Repair
Guide ACI 546R-0462.

The mix proportion, mechanical properties and shrinkage

cracking potentials are summarized in Table 4.9. Experimental details in determining


these properties can be found in Li and Li, 200631.
The concrete substrate was cast and cured as described in Section 4.5. In this
experiment, the surfaces of the concrete substrates were prepared in four different ways:
(a) normally cast (smooth); (b) roughened to 4~5mm; (c) roughened to 7~8mm; and (d)
roughened to 7~8mm + cement bonding slurry. For case (b), (c) and (d), the substrate
surfaces were roughened while in the fresh state by using a chisel to remove slurry

208

cement from the external surfaces of coarse aggregates. Before placing the repair layers,
the substrate surfaces were re-cleaned with a brush and high-pressure air to ensure a clean
bonding surface, and then they were damped to an adequate moisture level. After that,
repair layers made of each of the three repair materials were cast on top of the concrete
substrates. If cement-bonding slurry was used, a thin coating of "creamy" grout was
vigorously and thoroughly brushed into the prepared surface immediately before placing
the repair material.
The repair layers were moisture cured for 24 hours and then demolded. After
demolding, the layered specimens were moved into a room with ambient conditions of 45
5% RH and 201 C (66-70 F). Both delamination and surface cracking were
measured on a daily basis using dial gauges and a portable microscope (Figure 4.8).

4.6.2

Experimental Results and Discussion


Surface cracking and interface delamination of the three types of repair systems

are summarized in Table 4.10. The experimental results show that with a normally cast
(smooth) substrate surface, concrete, SFRC and ECC repairs all exhibited relatively large
delamination heights and lengths. The delamination values (height and length) of the
ECC repair were significantly higher than for the other two repair materials, probably due
to the large fly ash content in ECC, which leads to lower bonding strength with the
concrete substrate at early ages when shrinkage is most significant.
For the concrete repair, enhancing interfacial bonding by roughening the substrate
surface and/or adding bonding agent (cement-based slurry) reduced the delamination
values. However, the crack pattern did not change and several localized cracks formed

209

with crack width larger than lOOum. Furthermore, larger crack width was observed as
the interfacial bond became stronger. This is explained by the positive cracking potential
of concrete, resulting in shrinkage deformation that cannot be accommodated by concrete
itself through "yielding". Therefore, the shrinkage deformation must be accommodated
by interfacial delamination, repair cracking, or a combination of both to different extents.
As a stronger interfacial bond reduces delamination, bigger cracks naturally form in the
concrete repair layer to accommodate shrinkage deformation.
For the SFRC repair, no significant changes in interfacial delamination values
were observed among the four types of surface preparation.

Even with a deeply

roughened substrate surface and application of bonding agent, the delamination height
and length were still larger than 310^m. This happened because after shrinkage cracks
formed in SFRC repair, the cracks were bridged by steel fibers. Unlike the traction-free
cracks in the concrete repair, these cracks were not able to open freely to relax the stress
built up in the repair layer, and consequently forced the interface to delaminate. This
phenomenon dominated the interaction between the SFRC repair and substrate, even
when a strong interfacial bond was present.
For the ECC repair with enhanced interfacial bonding, the delamination height
and length were significantly reduced, and the multiple cracking phenomenon became
more predominant. For the "roughened to 7-8mm" and "roughened to 7-8mm + cement
bonding slurry" cases, the ECC repaired system exhibited the most desirable performance,
with both crack width and interfacial delamination below 60 um. No localized fracture
was observed in the ECC repair layer.

210

This study further verified the outstanding cracking and delamination resistance
of a ductile material repaired system under restrained volume change conditions. When
an adequate bond was provided, the ECC repair developed multiple microcracks rather
than several localized cracks, which consequently suppressed interfacial delamination
under restrained drying shrinkage. Unlike other brittle or quasi-brittle materials, the tight
crack width of ECC is a material property, which is independent of structural dimensions.
This implies that with increasing structural scale, the advantage of using ECC as a repair
material will become even more important.
The bond strength of repair materials with concrete is a requirement of ACI
546R-0462. In ACI 546R-04 Section 2.7, in-place tensile pull-off tests are recommended
to evaluate the adequacy of the surface preparation and bonding of repair materials.
Failure in substrate is preferred, which means that materials having high bonding strength
with the old concrete are more likely able to make repairs with sound performance and
durability. Experimental results from this study show that sufficient bonding strength is
necessary for ECC repair material to fully utilize its tensile ductility to suppress repair
cracking and interfacial delamination.

However, for brittle or quasi brittle repair

materials with large cracking potential values, enhancing interface bonding strength will
have a very limited effect on achieving durability in the repaired structure. This can be
explained by the time-dependent interaction between the repair and substrate. Once
sufficient stresses have been built up in the repaired system due to restrained repair
volume change, a stronger bond can reduce the trend of interface delamination, but also
promote the tendency of surface cracking, potentially increasing cracking number and/or
width. In this sense, special attention needs to be paid when interpreting results from the

211

pull-off test, because it over-simplifies the interaction between the repair and concrete
substrate, and neglects the subtle time-dependent competition between surface cracking
and interfacial delamination. Simply seeking a strong bond while ignoring the critical
influence of the repair material's tensile ductility on reducing cracking potential cannot
fundamentally improve repair durability.

4.7

Numerical Study on the Simulated Repair System

4.7.1

Problem Formulation and FEM Model


A numerical study was carried out to simulate the layered repair specimen by

using the finite element program MLS (Multi-Layer Systems) of FEMMASSE (Finite
Element Modules for Materials Science and Structural Engineering)63. The module MLS
can compute the physical and structural behavior of composite structures, taking varying
environmental conditions into account.
This numerical study considered a concrete substrate that was repaired by a
freshly cast layer of material. The dimensions of the concrete substrate and the repair
layer adopted were the same as those of the experimental specimen. The geometry and
mechanical boundary conditions of this model are shown in Figure 4.13. The objectives
of this numerical study were to simulate the restrained shrinkage induced repair cracking
and interfacial delamination process using three types of repair material models (brittle,
tension-softening, and strain hardening). The simulation results were used to explain the
experimental findings, by providing insights on the time-dependent stress distribution in
the repair layer and at the repair/substrate interface.

212

Three material models were applied, respectively to simulate the three types of
repair material: brittle material model - HES-Concrete, tension softening model - HESSFRC, and strain hardening model - ECC (Figure 4.14). In these models, a bi-linear
stress-strain relation was used to describe the elastic straining and strain-hardening
response, while a stress-crack opening relation was used to describe the tension-softening
response. Note that the parameters of the material models are age-dependent, and were
fitted as curves according to experimentally measured results from material of ages 4 h to
28 d. For example, the initial slope of the three curves (the material Young's modulus E)
changes with material age according to experimental results, as shown in Table 4.3 and
Table 4.4. The tensile strength of HES-ECC from material age of 4 h to 28 d was
adopted from the testing results in Table 4.4. The first cracking strength of HES-ECC
was assumed to be 80 percent of the ultimate tensile strength. The tensile strength of
HES-Concrete and HES-SFRC was assumed to be the same as HES-ECC at different
ages. Statistical standard deviation of material tensile strength of the three materials was
assumed to be 0.1 in the FEMMASSE FEM model. The tensile strain capacity of HESECC was conservatively assumed to be 3% at all ages. Compressive strength of the three
materials was adopted from the experimental data in Table 4.3. Poisson's ratios of the
three materials were all assumed to be 0.2.
The concrete substrate was assumed to have the same mechanical and fracture
properties as the HES-Concrete repair layer. No cracking or shrinkage was considered in
the old concrete substrate.
The repair/old interface used the interface type A model in FEMMASE, with the
properties listed in Table 4.11 and illustrated in Figure 4.15. No direct measurements

213

were made of the interface properties, but they were assumed to be the same in all three
cases. ft was assumed to be 4 MPa38. FEMMASE type A model default values were used
for the other parameters.
In this numerical model, loading derives from the repair layer material drying
shrinkage deformation that changed with material age as well. The measured timedependent shrinkage strain of HES-Concrete, HES-SFRC and HES-ECC shown in Figure
4.9 was used as the input loading for this FEM model.

4.7.2

Numerical Simulation Results


The FEM predicted cracking behavior of the HES-Concrete, HES-SFRC and

HES-ECC repairs at the 60-day age is shown in Figure 4.16. It can be seen that the HESConcrete repair developed several localized cracks with a maximum crack width of 229
um. The HES-SFRC repair developed more localized cracks, but a smaller maximum
crack width of 111 um. Despite having shrinkage strain nearly twice as high as that of
HES-Concrete and HES-SFRC, HES-ECC exhibited saturated cracking behavior with
crack width under 46 um. These results from FEM simulation are quite similar to the
experimental results. Although the cracking numbers predicted by the FEM models are
not exactly the same as the experimental results due to the unavoidable material
variability (e.g. initial flaw sizes), the cracking trends (localized fractures with larger
crack width in the HES-Conerete and HES-SFRC repairs and microcracks in the HESECC repair) are accurately predicted by the FEM model.
Figure 4.17 shows the tensile stress (oxx) distribution in the HES-Concrete, HESSFRC, and HES-ECC repair layers at the age of 60 days.

214

Figure 4.18 shows the

repair/old interface shear stress (axy) distribution of the three repaired systems at the age
of 60 days. In the HES-Concrete repair, a xx is zero at the cracking locations because the
cracks are traction free. Opening these cracks relieves part of the interface normal and
shear stresses, and reduces interface delamination at the ends of repair. However, it can
be seen that c xy remains at a significant level. The tensile component Cyy (not shown)
shows a similar trend. This high stress is due to the tendency of the repair elements
between the widely separated cracks to shrink while being restrained by the underlying
substrate. Therefore, although the HES-Concrete repair exhibited a similar amount of
interface delamination as the HES-ECC repair in the experiments, the FEM simulation
indicates that the tendency of interface delamination of the HES-Concrete repair is much
higher than the HES-ECC repair due to the higher level of interface stresses. In the HESSFRC repair, axx is non-zero at the cracking locations due to tension softening associated
with steel fiber bridging. The resulting high interface stresses (Figure 4.18) induce high
interface delamination in the HES-SFRC repair.
A significantly different behavior can be observed in the HES-ECC repair. The
saturated microcracks of HES-ECC are tight (below 46 \im) and uniformly distributed
through the repair layer. After cracking, these microcracks can still carry significant
amounts of stress, as seen in Figure 4.17. By developing a large number of microcracks,
the HES-ECC repair deformed uniformly in a ductile manner, and relieved the stresses
everywhere along the interface. Thus, the shrinkage deformation was accommodated by
the "plastic" straining of the ECC repair layer. Figure 4.18 shows that small interface
shear stresses (cxy) (and similarly the tensile stress <%, not shown) are almost uniformly
distributed along the HES-ECC repair. A small amount of interface delamination exists

215

only at the repair ends.

The repair cracking and interface delamination behavior

experimentally measured and numerically predicted in this study are consistent with
those numerically modeled by Kabele38.

4.8

Simplified Analytical Model for the Repair Systems under Restrained

Volume Change
An analytical model was developed by Zhou et al64 to calculate the stresses in the
concrete and ECC repairs as well as at their interfaces with concrete substrate due to
restrained volume change. For the concrete repair, the model development procedure
contains four steps, assuming the restrained volume change is due to differential
shrinkage65 as illustrated in Figure 4.19. Firstly, the repair material is assumed to be
unbonded from the substrate concrete. Instead, the shrinkage of the repair material is
restrained at the both ends, resulting in a tensile stress
ffSh(y) = Er x esh{y)

(4.15)

where esh(y) and Er are the shrinkage and elastic modulus of the repair material.
Secondly, the repair material was bonded to the substrate concrete. For the equilibrium
of the repair system, a compressive stress ~ashiy) is then applied at the ends of the repair
material. This stress is simplified to be constant through the depth of the repair material,
which can be written as
oeq=--fhEresh(y)dy

(4.16)

where hr is the thickness of repair material. Thirdly, the stresses and strains in the repair
system subjected to the external stress aeq are calculated based on the plate theory and the
assumption of a linear relation between shear stress and slip at the interface. Finally, the

216

restrained shrinkage stress oshiy) is superimposed to the stress in the repair material
calculated in the third step. According to this model, the highest tensile stress oxx in the
repair material is located at the middle bottom (in x direction) of the repair layer, and the
highest interfacial tensile stress (ryy and interfacial shear stress <% are located at the two
ends (in x direction). The maximum values of the stresses can be calculated using the
following equations:

1+

iT

e +e

' '

s s

FE
e^ - 1 2
~T*
(1--!:)
Xhr E,

(4.18)

WM"-1

(4 . 19)

EX
where X = J - ^ - + - ^ \Ehr r
Eh
s s

(4.20)

Es is the elastic modulus of substrate concrete, hs is the thickness of substrate concrete, K


is the shear stiffness of interface, andL is the length of repair system.
If restrained volume change is due to combined effects of shrinkage, temperature
deformation and structural loads, the eSh can be replaced by etotai in this model.
For the ECC repair, the tensile stress-strain curve of ECC is assumed to be
bilinear, as illustrated in Figure 4.20. When the stress is lower than the first cracking
strength o/c, ECC exhibits elastic straining like concrete, as defined by Ei; When the
stress is between the first cracking strength qyc and the ultimate strength (7CU, ECC shows
strain-hardening behavior as defined by E2.
The calculation of stresses in the ECC repair system was divided into two stages.

217

In the first stage, i.e. when the tensile stress is smaller than the first cracking strength of
ECC, the stresses in the ECC repair system increase as the differential volume change
increases. The stresses can be calculated using Equations 4.17, 4.18, 4.19, 4.20, by
replacing Es with the elastic modulus of ECC Ej. The critical value of the differential
volume change s/c corresponding to the stresses leading to first cracking of ECC can be
calculated with the following equation:

^-T^-^^r)-^
where k 1 =

[AV EA

+ E

sK

(421)

(4.22)

With the calculated value of differential volume change e/c, the maximums of stresses at
the interface can be calculated by:
E

i /c

yy

{ i

e K L

J__^>
\K Es

(423)

e^+\\L}

ElfMre^-l

Once the tensile stress is higher than the first cracking strength of ECC, ECC
starts the multiple cracking process and enters the strain hardening stage. During this
stage, the stresses induced by the increase in the differential volume change from e/c to sSh
can be calculated by:
.2

< -

2(sH-fc)n
FJ

/ (1

2"'r

E2hr
Eshs

2_

1
^ T
x,k

r)
.Kh

* Eii1sh-f)nE2
e^Le^+lhL
-\ 2
yy = ~ i E z F - d
w , ,r)

CT

KK

(4-25)

^2~

218

(4.26)

2 _E2^sh-f)KKehL-\
*y
E2hr
ew + i
1+
Eh
where ^

jJL

(421)

+ JL

(4.28)

By superimposing the stresses calculated in these two stages, the stresses in the
ECC repair system induced by the differential volume change can be determined by the
following equations:

= <t + o2xx

(4.29)

43

-< <

( - )

* , - < +<

(4-31)

The experimentally measured ECC material properties31 were used as inputs to


calculate stresses in the ECC repair system, which were experimentally and numerically
investigated in previous sections, using Equations 4.29, 4.30 and 4.31.

The shear

stiffness of interface if was assumed to be 10 N/(mm2><mm)66. The ECC shrinkage strain


at different ages was assumed to be constant through the depth of the repair layer
(independent of y). Figure 4.21 shows the calculated results. As the shrinkage of ECC
increases to 392 |j.strain (0.0392%), the tensile stress in ECC reaches the first cracking
stress of 5.0 MPa. At this point, ECC enters the strain-hardening stage and starts inelastic
straining similar to metals. As the shrinkage strain continued increasing up to 1800
ustrain (0.18%), which corresponds to the measured free shrinkage of ECC M45 (28day, 45 5% RH, 201 C (66-70 F)), the stresses in the ECC repair only increases a
small amount. This is because of the low stress-strain gradient E2 in the strain-hardening

219

stage. As the maximum imposed shrinkage strain 0.18% is greatly lower than ECC
tensile strain capacity 3-6%, the material is still experiencing the early part of its strainhardening stage through multiple microcracking, and there is no danger of tensile failure.
Additionally, Figure 4.21 shows the interfacial stresses increase as shrinkage increases to
392 |istrain (0.0392%). Beyond this strain level, the increase rates interfacial stresses
become extremely slow.

This shows that the inelastic straining of ECC material

effectively reduces the stresses built up at the repair/old interface, therefore suppressing
interfacial delamination.
The modeling results correspond well with the experimental findings and FEM
results described in previous sections, in term of the stress development in the ECC repair
layer and at the interface as well as their effects on repair cracking and interfacial
delamination.

Although some assumptions (e.g. interfacial shear stiffness as 10

N/(mm2xmm)) need to be further validated, this analytical model provides a useful tool to
study the influence of various parameters (e.g. interfacial shear stiffness, repair length
and thickness, repair material tensile strain capacity, tensile strength and elastic modulus)
on the stresses and strains in the repair and at the repair/old interface, and to predict the
performance of different repair systems.

The details will be included in future

publications.

4.9

Conclusions
Simultaneously suppressing repair cracking and interfacial delamination, which

are induced by restrained volume change, can be achieved using ductile ECC repair
materials as proposed in this study. The free drying shrinkage properties of HES-ECC

220

were characterized at different RHs. Results showed that the shrinkage strain of HESECC was below 0.3%, which was one order of magnitude lower than its tensile strain
capacity of 3-6%, therefore leading to a highly negative cracking potential value. This
implied that when the drying shrinkage of HES-ECC was restrained, the material's
ductility can accommodate shrinkage deformation by forming multiple microcracks
during its strain hardening stage without localized cracking failure. This contention was
proven later by the results of the restrained shrinkage ring test and the restrained
shrinkage test on a simulated repair system.
An experimental study was conducted on a simulated repair system subject to
drying shrinkage, in order to investigate the effects of repair material ductility and surface
preparation on the repair's performance. When an adequate bond was present, 3 - 4 large
surface cracks with width up to 520 \xm formed in the HES-Concrete repair layer,
accompanied by a small amount of interfacial delamination up to 90 \xm. For the HESSFRC repair system, 1 ~ 4 large surface cracks formed, and the presence of steel fibers
reduced the maximum crack width to 280 \xm. However, the interfacial delamination of
HES-SFRC repair was the largest among the three repair materials. This was explained
by the significant amount of restrained shrinkage induced tensile and shear stresses at the
interface, which had no avenue of release through material inelastic straining as does
HES-ECC, or crack free-opening as does HES-Concrete. For the HES-ECC repair, more
than 100 fine microcracks with width less than 50 um formed in the repair layer,
accompanied by interface delamination less than 80 ^m at the ends. Similar constrasts in
behavior were also found in concrete, SFRC, and ECC repair systems, suggesting that the

221

ductility of a repair material is the most critical factor contributing to the resistance to
repair cracking and delamination.
Experimental results also showed that sufficient bonding strength was necessary
for ECC repair materials to fully utilize their tensile ductility to achieve desired
performance. However, for brittle or quasi-brittle repair materials with large cracking
potential values, increasing bonding strength will provide only limited aid in improving
the durability of the repaired system. This is due to the interaction between the repair and
concrete substrate, and the time-dependent competition between surface cracking and
interfacial delamination. Therefore, simply seeking a strong bond while ignoring the
critical influence of the repair material's tensile ductility on reducing cracking potential
cannot fundamentally improve repair durability.
The results from an FEM study further confirmed the influence of repair material
ductility on the repair system's resistance to cracking and interfacial delamination. The
analysis provided additional insights on the stress distribution in the repair systems using
brittle, quasi-brittle, and ductile repair materials, respectively. It revealed that the
shrinkage deformation of ECC was accommodated by the "plastic" strain of the ECC
repair layer by forming saturated microcracks, which still carried significant amounts of
stress after formation. The uniform deformation of the ECC repair in a ductile manner
relieved the stresses everywhere along the interface, so that a small amount of interface
delamination existed only at the repair ends. In the concrete repair, cracking happened
when the restrained shrinkage caused the tensile stress in the repair layer to reach the
material tensile strength. Opening these tract-free cracks relieves part of the interface
normal and shear stresses, and reduces interface delamination at the ends of repair.

222

However, the model showed that axy and (% remained at a significant level. In the SFRC
repair, axx is non-zero at the cracking locations due to its tension-softening behavior
associated with steel fiber bridging. This resulted in high interfacial stresses between the
SFRC repair and old concrete.

The FEM study provided understanding on the

mechanical mechanism behind the repair/substrate composite performance that was


observed in the experimental study.
Additionally, a simplified analytical model was derived to calculate the tensile
stress in the repair layer, and the tensile and shear stresses at the interface. The modelpredicted results corresponded well with the experimental findings and the FEM results.
More importantly, this analytical model can be used to study the influence of different
parameters (e.g. interfacial shear stiffness, repair length and thickness, repair material
tensile strain capacity, tensile strength and elastic modulus) on the stresses and strains in
the repair and at the repair/old interface.
This research departs from the traditional emphasis on isolated material properties
such as high compressive strength and low shrinkage, and intends to shift the repair
industry's attention toward the age-dependent interaction between the new repair and old
concrete. To ensure dimensional compatibility between the repair and existing concrete,
a repair material with large inelastic tensile strain capacity (ductility) is necessary. When
the material ductility requirement is satisfied, cracking control reinforcements and repair
material free drying shrinkage limits will become less important, and surface preparation
methods to enhance the interface bond will be more meaningful toward achieving
durability of repaired concrete structures.

223

Table 4.1 - General requirements of patch repair materials for structural compatibility.
Relationship of repair mortar (R)
to concrete substrate (C)

Property
Strength in compression, tension,
R>C
and flexure
Modulus in compression,
R~C
tension, and flexure
Poisson's ratio

Dependent on modulus and type of repair

Coefficient of thermal expansion

R~C

Adhesion in tension and shear

R>C

Curing and long-term shrinkage

R<C

Strain Capacity

R>C

Creep

Dependent on whether creep causes desirable or


undesirable effects

Fatigue performance

R>C
Should not promote alkali-aggregate reaction,
sulfate attack, or corrosion of reinforcement in the
substrate
Dependent on permeability of patch material and
chloride ion content of substrate

Chemical reactivity
Electrochemical stability

Table 4.2 - Mix proportion of the investigated materials.


Material
HES-Concrete
HES-SFRC
HES-ECC

C(a)

1.0
1.0
1.0

0.4
0.4
0.33

S
1.3
1.3
1.0

CA
1.3
1.3
0.064 (b)

SP
0.005
0.005
0.0075

Portland type III cement


Polystyrene beads as "coarse aggregates" for HES-ECC
Fiber volume content
Steel hooked fibers used for HES-SFRC
PVA fiber used for HES-ECC

224

AC
0.04
0.04
0.04

Vf(c)

0.01 (d)
0.02 (e)

Table 4.3 - Mechanical properties of the investigated materials.


f J (a)

Age

Material

7d
HES-Concrete
28 d
7d
HES-SFRC
28d
7d
HES-ECC
28 d

MPa(psi)
49.9 1.6
(7234 228)
54.2 2.4
(7860 345)
51.5 2.4
(7462 348)
57.0 2.0
(8254 290)
47.5 1.9
(6885 275)
55.6 2.2
(8063315)

Ew,
GPa(ksi)
26.21.4
(27.8 196)
27.81.5
(4025 221)
25.72.0
(3725 287)
29.21.3
(4231 189)
20.60.7
(2986 101)
23.21.0
(3365 140)

eu %

Tensile
Behavior

0.01

brittle

0.01

quasi-brittle

3-6

ductile

Table 4.4 - HES-ECC tensile properties at different ages.


Age
4h
6h
12 h
24 h
7d
28 d

Ew,
GPa(ksi)
13.10 0.77
(1900 111)
14.98 0.79
(2173 114)
16.05 1.02
(2328 147)
18.30 0.59
(2654 85)
20.59 0.70
(2986 101)
23.20 0.96
(3365 140)

Tensile
Strength(a)MPa(psi)
3.46 0.08
(501 12)
4.21 0.13
(610 19)
4.57 0.17
(662 25)
4.69 0.08
(680 12)
5.560.11
(807 16)
5.68 0.14
(823 20)

Strain Capacity (%)


5.970.22
4.970.38
4.410.33
3.990.27
3.520.29
3.470.62

Mean standard deviation (measured from 9 specimens for each of the properties)

225

Table 4.5 - Theoretical parametric values of the investigated materials.

HES-ECC
HES-Concrete
ECC (M45)*
Concrete

A [J/gl
8.4
2.9
8.2
2.3

B [J/g]
13.1
8.5
17.9
5.1

X [g/J] xlO 3 c [g/J] xlO s


4.3
3.1
3.8
1.6
2.3
2.0
4.0
1.9

d [g/J] xlO 4
-1.0
1.2
+7.1
-3.0

Values adopted from Weimann and Li, 2003

Table 4.6 - Restrained drying shrinkage crack width development with age.
Age (day)
0
1
2
3
4
14
28
40
60

Average cracking width, urn (in)


HES-Concrete
HES-ECC
0(0)
0(0)
30(0.001182)
0(0)
310(0.012214)
40(0.001576)
590 (0.023246)
50 (0.00197)
610 (0.024034)
50 (0.00197)
1070 (0.042158)
50 (0.00197)
1210 (0.047674)
50 (0.00197)
1250 (0.049250)
50 (0.00197)
1250 (0.049250)
50 (0.00197)

Table 4.7 - Estimated material cracking potentials, RH=45%.


Properties
sh(%)
Cp ( % )
Ee(%)
Ei ( % )
HES-Concrete 0.105
0.02 - 0.06
0.01
0
0.02 - 0.06
HES-SFRC
0.077
0.01
0
0.07
HES-ECC
0.242
0.015
3-6
*Properties were measured at specimen's age of 28 days.

226

P = sh" (e+i+ cp) ( % )

0.035 to 0.075
0.007 to 0.047
(-5.343) to (-2.843)

Table 4.8 - Interface delamination and surface cracking of HES-Concrete, HES-SFRC,


and HES-ECC layered repair systems at the age of 60 days.

Repair
Material

Specimen
Number
(1)

HESConcrete

(2)
(3)
(1)

HES-SFRC

(2)
(3)
(1)

HES-ECC

(2)
(3)

Delamination
Height, Length,
[im
mm
(inxio- 3 )
(in)
90
170
(3.54)
(6.69)
30
27
(1.06)
(1.18)
65
73
(2.56)
(2.87)
323
310
(12.72)
(12.2)
301
260
(10.2)
(11.85)
342
300
(13.46)
(11-8)
80
82
(3.15)
(3.23)
50
47
(1.97)
(1.85)
75
79
(2.95)
(3.11)

227

Surface Cracking
Width,
Number
(am
(inxlO"3)
160 (6.30), 520 (20.5),
3
370 (14.6)
190 (7.48), 340 (13.4),
4
360 (14.2), 490 (19.3)
700 (2.76), 380 (15.0),
4
420 (16.5), 450 (17.7)
2

110(4.33), 120(4.72)

50 (1.97), 90 (3.54),
120(4.72), 130(5.12)

280(11.0)

83
109
113

10-50
(0.39-1.97)
10-60
(0.39 - 2.36)
10-50
(0.39-1.97)

Table 4.9: Material composition, 28-day mechanical properties and cracking


potential.

%
0.01

(MPa)
601

E
(GPa)
261

0.01 (c)

0.01

0.01

632

261

0.02 (d)

0.02

3-5

622

201

FA

CA

SP

V/b)

vf

1.0

0.4

1.3

1.3

0.01

SFRC

1.0

0.4

1.3

1.3

0.01

ECC

1.0

0.53

0.8

1.2

0.03

Material

C(a)

Concrete

P
0 to 0.04
(-0.017)
to 0.023
(-4.908)
to (-2.408)

Portland type I cement


Fiber volume content
(c)
Steel hooked fibers used for SFRC
td)
PVA fiber used for ECC

Table 4.10 - Interface delamination and surface cracking of concrete, SFRC and ECC
layered repair systems.

Repair
Material

Concrete

SFRC

ECC

Surface
Preparation
Smooth
4-5mm roughened
7-8mm roughened
7-8 mm roughened
+ bonding slurry
Smooth
4-5mm roughened
7-8mm roughened
7-8 mm roughened
+ bonding slurry
Smooth
4-5mm roughened
7-8mm roughened
7-8mm roughened
+ bonding slurry

Delamination
Height (um) Length (mm)

Cracking
Number Width (urn)

410
190
65

320
172
50

2
3
4

120-130
130-190
220-270

32

23

210-360

550
370
275

397
375
350

2
5
3

90-120
70-110
120-140

310

354

90-150

1225
425
53

722
402
50

0
21
76

~
10-60
10-60

40

31

103

10-60

228

Table 4 . 1 1 - Interface property values in FEM model.


ft (MPa)
Factor for SI
wl (mm)
w2 (mm)
Friction angle (Degr.)
Maximum slip (mm)
Normal stiffness (N/ram3)
Shear stiffness (N/mm )

229

4.0
0.25
0.05
0.2
35.0
3.0
100
10

Shrinkage

Dimensional
Incompatibility

Thermal
Effect

Existing Cracks
in Substrate

Structural
Fatigue Loading

% rf
Stress V
Concentration

Restrained
Volume Change

Impaired Transport
:; Resistance;

ij

Chloride
Ingress

Oxygen
. .
Diffusion

Moisture _ .
..
.
Carbonation
Ingress

~H
Water

~0"~

e_t

t l ^_
Saturation

< Freeze & T h a w )

Alkali
Ingress
Alkali
Attack

"s,

Disintegration "j

/*z.
End of Repair Life

Figure 4.1 - Typical deterioration process of concrete repair.

230

Sulfate
Ingress

Sulfate
Attack

^\
/

Surface Cracks

/ )

Delamination ^-L^

Height

J^ j^__

Dll'lfill
Delamination Profile

Delamination
Length

y
- x

Figure 4.2 - Schematics of typical failure modes in layered repair system.

1000

20

40
60
Relative Humidity RH (%)

80

Figure 4.3 - Illustration of the Kelvin Equation.

231

100

Figure 4.4 - Free drying shrinkage test.

0.4

ra

0.3 4

2
c"
_o
re

"re

0.2 -f

0.1 +

hyg,Dis

>>
hyg,Sur

0.0
0

20

40

60

80

100

Relative Humidity RH (%)

Figure 4.5 - Experimental data and theoretical curves of hygral deformation vs. relative
humidity for the investigated materials (data of HES-ECC and HES-Concrete are
measured in this study; data of ECC and Concrete are adopted from Weimann and Li,
2003).

232

Steel Ring
Specimen
Crack width
measuremer

279.4 mm (11 in.)

5\
304.8 mm (12 in.)

K-

355.5 mm (14 in.)

Figure 4.6 - Restrained shrinkage steel ring test.

233

Q.S6
1400
1200

.3.
JC

1000
800

20

30

40
SO
Age (day)

OHES-Conciete

606

400

AS

2.00

10

i
HES-ECC

as

>
<

60

Figure 4.7 - Restrained shrinkage crack width development as a function of drying time
(at RH = 45 5%).

234

I"

I I' 1

m
Dial Gage

Microscope

Figure 4.8 - Layered repair system test set-up under restrained shrinkage.

0.30%
0.25%
c
0.20%

Ar

(0

HES-Concrete
o HES-SFRC
* HES-ECC

J 0.15%

c
0.10%
0

0.05%

0.00%
()

10

20

30

40

50

60

70

Age (days)

Figure 4.9- Free shrinkage strain of repair materials at different ages.

235

HES-Concrcte

* -

I'JII | i l l l

\
3

>

U'

'.

in

llli

i
HES-SFRC

/ .
/
I
I

' 1 IM

11 111
l'l im

I U - ou \xm
(0.39-2.36* I ( P i n )

IIES-ECC
A

-A

Figure 4.10 - Repair surface cracking at the age of 60 days.

236

Distance From Specimen's End (mm)

200

400

600

800

1000

1200

1400

at

IUJ

- --Id
A - -find

- - / d

TO
V

10

20

30

40

50

Distance From Specimen's End (in.)


Distance From Specimen's End (mm)

-I

400

600

800

1000

1200

1400

1
HES-SFRC

IS

0)

"Id

~~ 7 d

-*14d
28d
-o_60d

c
<a

ionfoj

200

5
+
20

+40

30

50

Distance From Specimen's End (in.)


Distance From Specimen's End (mm)
400

600

800

10

20

30

1000

40

1200

1400

50

Distance From Specimen's End (in.)

60

00

Figure 4.11 - Interface delamination profiles of repaired systems.

237

- HES-SFRC
-AHES-ECC
-HES-Concrete

Figure 4.12 - Specimen end delamination height at different ages.

238

Repair

Figure 4.13 - FEM model of layered repair system.

(a)
u(A)
0.0625 mm

Of

(b)
ufA)

-HK
0.03 mm

15 mm

a Ai
Of

1 \
! \

- - -

0.8 Of

|
i
1
!
!

i
i
i
i

\
\
\
\
s

6 mm

(c)

u[A)

Figure 4.14 - Assumed tensile behavior of (a) HES-Concrete, (b) HES-SFRC, (c) HESECC materials, and (d) combined for all three materials.

239

Normal

Normal
Displacement

Shear Slip

Figure 4.15 - Interface model: combined uniaxial stress-strain relation in the normal
direction with influence of shear slip.

240

0.000

0.229mm
^ g (9x10" 3 in)
0.229

(a) HES-Concrete

^ ^ , 1 ^ * ^ . 3 = ^

0.122 mm
4.8x10"3in)

0.000
0.122

o.oo

(b) HES-SFRC

M K t a

MWHHUMM

0.046 mm
0*04(5 (1.8x10'3in)

(c) HES-ECC
Figure 4.16 - Predicted crack width of (a) HES-Concrete, (b) HES-SFRC and (c) HESECC repair layer at the age of 60 days.

241

(a) HES-Concrete

(a) HES-SFRC

(b) HES-ECC

Figure 4.17 - Predicted stress distribution (o n ) in the (a) HES-Concrete, (b) HES-SFRC
and (c) HES-ECC repaired system at the age of 60 days

242

Maximum 3.243 MPa (470 psi)

Minimum -3.192 MPa (-463 psi)

(c) HES-Concrete

Maximum 2.272 MPa (330 psi)

Minimium -2.196 MPa (-320 psi)

(b) HES-SFRC

^Maximum 3.753 MPa (544 psi)

dllliiuiSVww^^p^'w^/^^v^

Minimum -3.708 MPa (-538 psi)

(c) HES-ECC
Figure 4.18 - Predicted interface shear stress (axy) of (a) HES-Concrete, (b) HES-SFRC
and (c) HES-ECC repaired system.

243

Repair material

Substrate concrete

L/2

Ssh

'

r^

hs

L/2

Figure 4.19 - Concrete repair subjected differential shrinkage.

Sfc

Figure 4.20 - Simplified tensile stress-strain curve of ECC.

244

500

1000
1500
Shrinkage (^strain)

2000

Figure 4.21 - Analytical model predicted tensile stress in ECC repair axx, repair/old
interfacial tensile stress (% and shear stress cr>xy.
x

245

References:
1

Vaysburd, A. M., Emmons, P. H., McDonald, J. E., Poston, R. W., and Kesner, K. E.,
"Selecting Durable Repair Materials: Performance Criteria - Field Studies," Concrete
International, Vol. 22, No. 12, December 2000, pp. 39-45.
2

Vaysburd, A. M., Emmons, P. H., Mailvaganam, N. P., Mcdonald, J. E., and


Bissonnette, B., "Concrete Repair Technology - A Revised Approach is Needed,"
Concrete International, Vol. 26, No. 1, January 2004, pp. 59-65.
3

Banthia, N. and Gupta, R., "Repairing with Fiber-Reinforced Concrete - Cellulose


Fiber Enhance Overlay Performance," Concrete International, Vol. 28, No. 11,
November 2006, pp. 36-39.
4

Cusson, D., Qian, S., and Hoogeveen, T., " Field Performance of Concrete Repair
Systems on Highway Bridges," ACI Materials Journal, Vol. 103, No. 5, 2006, pp. 366373.

Vaysburd, A. M., Brown, C. D., Bissonnette, B., and Emmons, P. H., " "Realcrete"
versus "Labcrete" - Searching for Tests that Give Reliable Results," Concrete
International, Vol. 26, No. 2, February 2004, pp. 90-94.
6

Lawler, J.S., Zampini,D., and Shah, S. P., "Permeability of Cracked Hybrid FiberReinforced Mortar under Load," ACI Materials Journal, Vol. 99, No. 4, pp. 379-385.
7

Hearn, N. "Effect of Shrinkage and Load-Induced Cracking on Water Permeability of


Concrete," ACI Materials Journal, Vol. 96, No. 2, pp. 234-241.
8

Gerard, B., Reinhardt, H. W., and Breysse, D., "Measured transport in cracked concrete"
Penetration and Permeability of Concrete: RILEMReport 16. Ed. H.W. Reinhardt. 1997,
pp. 265-331.
9

Gerard, B., Jacobsen, S., and Marchand, J., "Concrete Cracks II: Observation and
Permeability - A Review," Concrete Under Severe Conditions 2, Environment and
Loading, Vol. 1, Gjorv, E & FN Spon, eds. Sakai, K and Banthia, N., 1998, pp. 183-197.
10

Lepech, M., and Li, V. C , "Water Permeability of Cracked Cementitious Composites,"


Paper 4539 of Compendium of Papers CD ROM, ICF 11, Turin, Italy, March 2005.
11

Sahmaran, M., Li, M., and Li, V. C , "Transport Properties of Engineered Cementitious
Composites Under Chloride Exposure," ACI Materials Journal, Vol.104, No. 6, 2007,
pp. 604-611.
Iowa Department of Transportation (Iowa DOT), "A Study of Transverse Cracks in the
Keokuk Bridge Deck," Final Rep., Ames, Iowa, 1986.

246

1^

Perfetti, G. R., Johnson, D. W., and Bingham, W. L., "Incidence Assessment of


Transverse Cracking in Concrete Bridge Decks: Structural Considerations," Rep. No.
FHWA/NC/85-002, Vol. 2, Federal Highway Administration, Washington, D. C , 1985.
14

Cheng, T. T., and Johnson, D. W., "Incidence Assessment of Transverse Cracking in


Bridge Decks: Construction and Material Consideration," Rep. No. FHWA/NC/85-002,
Federal Highway Administration, Washington, D. C , Vol. 1, 1985.
15

Perragaux, G. R., and Brewester, D. R., "In-Service Performance of Epoxy Coated


Steel Reinforcement in Bridge Decks," Final Technical Rep. No. 92-3, New York State
Department of Transportation, Albany, N. Y., 1992.
16

Kriviak, G. W., Skeet, J. A., and Carter, P. D., "Service Life Prediction of Protective
Systems for Concrete Bridge Decks in Alberta," International Association of Bridges and
Structural Engineers Symposium, San Francisco, Calif., 1995, pp. 469-474,.
17

Czarnecki, L., Garbacz, A., Lukowski, P., and Clifton, J. R., "Polymer Composites for
Repairing Portland Cement Concrete: Compatibility Project," Technical Report NISTIR
6394, National Institute of Standards and Technology, 1999.
18

Emmons, P. H., Vaysburd, A. M., Poston, R. W., and McDonald, J. E., "Performance
Criteria for Concrete Repair Materials, Phase II, Field Studies," Technical Report REMRCS-60, U. S. Army Waterways Experiment Station, Vicksburg, MS, September 1998.
19

Pigeon, M. and Bissonnette, B., "Bonded Concrete Repairs: Tensile Creep and
Cracking Potential," Concrete International, Vol. 21, No. 11, November 1999, pp. 31-35.
20

Vayburd, A. M., "Research Needs for Establishing Material Properties to Minimize


Cracking in Concrete Repairs, Summary of a Workshop," ICRI Publication No. Y32000I,
1996.
21

Vaysburd, A. M., Emmons, P. H., McDonald, J. E., Poston, R. W., and Kesner, K. E.,
"Performance Criteria for Concrete Repair Materials, Phase II Summary Report,"
Technical Report REMR-CS-62, U. S. Army Engineers Waterways Experiment Station,
Vicksburg, MS, March 1999.
22

Emberson, N. K., and Mays, G. C , "Significance of Property Mismatch in the Patch


Repair of Structural Concete. Part I: Properties of Repair Systems," Magazine of
Concrete Research, No. 152, September 1990, pp. 147-160.
23

Babaei, K., and Hawkins, N. M.,"Evaluation of Bridge Deck Protective Strategies,"


NCHRP Rep. No. 297, Transportation Research Board, National Research Council,
Washington, D. C , 1987.

247

La Fraugh, R. W., and Perenchio, W. F., "Phase I Report of Bridge Deck Cracking
Study West Seattle Bridge," Rep. No. 890716, Wiss, Janney, Elstner Associates,
Northbrook, 111, 1989.
25

Babaei, K. and Purvis, R., "Prevention of Cracks in Concrete Bridge Decks: Report on
Laboratory Investigation of Concrete Shrinkage," Research Project No. 89-01,
Pennsylvania Department of Transportation, Harrisburg, PA, 1994.
26

Ducrete, J., Lebet, J. , and Monney, C , "Hydration Effect and Deck Cracking During
the Construction of Steel Concrete Composite Bridges," Proc. ICOM-Construction
Mettalique, Article 359, 1997.
27

French, C , Eppers, L., Le, Q., and Hajjar, J. F., "Transverse Cracking in Concrete
Bridge Decks," Transportation Research Record 1688, Transportation Research Board,
Washington, D. C , 1999, pp. 21-29.
28

Frosch, R. J., Radabaugh, R. D., and Blackman, D. T., "Investigation of Transverse


Deck Cracking," Proc. Structures Congress, ASCE, Reston, VA, 2002.
29

Wittmann, F. H., and Martinola, G., "Decisive Properties of Durable Cement-Based


Coatings for Reinforced Concrete Structures," International Journal for Restoration of
Buildings and Monuments, Vol. 9, No 3, pp 235-264, 2003.
30

Li, M., and Li, V. C , "Durability of HES-ECC Repair Under Mechanical and
Environmental Loading Conditions," Proc, High Performance Fiber Reinforced Cement
Composites (HPFRCC5), Eds. H.W. Reinhardt and A.E. Naaman, Mainz, Germany,
2007, pp. 399-408.
31

Li, M., and Li, V. C , "Behavior of ECC/Concrete Layered Repair System under
Drying Shrinkage Conditions", International Journal of Restoration of Buildings and
Monuments, Vol. 12, No. 2, 2006, pp. 143-160.
32

Mehta, P. K., Concrete: Structure, Properties, and Materials, Prentice-Hall,


Englewood Cliffs, New Jersey, 1986, pp. 353-367.
33

Chang, P. K., Peng, Y. N., and Hwang, C. L., "A Design Consideration for Durability
of High-Performance Concrete," Cement & Concrete Composites, Vol. 23, No. 4-5, Aug.
-Oct. 2001, pp. 375-380.
34

Hwang, C. L., Liu, J. J., Lee, L. S., and Lin, F. Y., "Densified Mixture Design
Algorithm and Early Properties of High Performance Concrete, " Journal of the Chinese
Institute of Civil and Hydraulic Engineering, Vol. 8, No. 2, 1996, pp. 217-229.

248

35

Shah, S. P., Wang, K., and Weiss, W. J., "Is High Strength Concrete Durable?"
Concrete Technology for a Sustainable Development in the 21s' Century Eds. O. E. Gjorv
and K. Sakai, 2000, pp. 102-114.
Li, M., and Li, V. C , "Behavior of ECC/Concrete Layered Repair System under
Drying Shrinkage Conditions", Journal of Restoration of Buildings and Monuments, Vol.
12, No. 2, 2006, ppl43-160.
37

Li, V.C., "High Performance Fiber Reinforced Cementitious Composites as Durable


Material for Concrete Structure Repair," Proc. of ICFRC Int 7 Conference on Fiber
Composites, High Performance Concretes, and Smart Materials, Ed. By V.S.
Parameswaran, Pub. Allied Publishers Private Limited, New Delhi, India, 2004, pp. 5774.
38

Kabele, P., "Assessment of Structural Performance of Engineered Cementitious


Composites by Computer Simulation", CTU Report 4, Czech Tech. University, Prague,
2001.
39

Mindess, S., Young, J. F., and Darwin, D., Concrete, 2nd Edition, 2003.

40

Adamson, A. W., Physical Chemistry of Surfaces, 6th Edition, John Wiley and Sons,
Inc., 1997.
41

Wittman, F. H., "Interaction of Hardened Cement Paste and Water," Journal of the
American Ceramic Society, Vol. 56, No. 8, 1973, pp. 409-415.
42

Yates, D. J. C , "Molecular Specificity in Physical Adsorption," Advances in Catalysis


and Related Subjects, Academic Press Inc., Vol. 12, 1960, pp. 265-312.

43

Flood, E .A., and Heyding, R. D., "Stresses and Strains in Adsorbent-Adsorbate


Systems," Canadian Journal of Chemistry, Vol. 32, 1954, pp. 660-682.

44

Wittmann, F. H., "Surface Tension, Shrinkage and Strength of Hardened Cement


Paste," Materials and Structures, Vol.1, No. 6, 1968, pp. 547-552.
45

Wittmann, F. H., "The Structure of Hardened Cement Paste - A Basis for a Better
Understanding of the Materials Properties," Hydraulic Cement Paste: Their Structure
and Properties, Cement and Concrete Association, Slough, United Kingdom, 1976, pp.
96-117.
46

Wittmann, F. H., "Creep and Shrinkage Mechanisms," Creep and Shrinkage in


Concrete Structures, Eds. Z. Bazant and F. H. Wittmann, J. Wiley and Sons Inc., 1982,
pp.129-161.

249

Weimann, M. B., "Vergleichende Studie der hygrischen Eigenschaften ausgewahlter


Werkstoffe des Bauwesens," Building Materials Reports No. 14, Aedificatio Verlag
GmbH, Freiburg, Germany, 2001.
48

Weimann, M. B., and Li, V. C., "Hygral Behavior of Engineered Cementitious


Composites (ECC)," International Journal for Restoration of Buildings and Monuments,
Vol. 9, No 5, 2003, pp513-534.
49

Beltzung, F., Wittmann, F. H. and Holzer L., "Influence of Composition of Pore


Solution on Drying Shrinkage," Creep, Shrinkage and Durability Mechanics of Concrete
and other Quasi-Brittle Materials, Eds. Ulm, F. J., Bazant, Z .P. and Wittmann, F. H.,
Elsevier Science Ltd., 2001, pp. 39-48.
50

Schubert, H., Kapillaritat in porosen Feststoffsystemen, Springer-Verlag, Berlin,


Germany, 1982.

51

Hiller, K. H., "Strength Reduction and Length Changes in Porous Glass Caused by
Water Vapour Adsorption," Journal of Applied Physics, Vol. 35, No. 5, 1964, pp. 16221628.
52

American Society for Testing and Materials (U.S. Code Organization).

53

Wittmann, F. H., and Martinola, G., "Decisive Properties of Durable Cement-Based


Coatings for Reinforced Concrete Structures," International Journal for Restoration of
Buildings and Monuments, Vol. 9, No 3, 2003, pp. 235-264.
54

Greenspan, L., "Humidity Fixed Points of Binary Saturated Aqueous Solutions,"


Journal of Research of the National Bureau of Standard - American Physics Chemistry,
81A, 1, January-February 1977.
55

Weiss, W. J., and Ferguson, S., "Restrained Shrinkage Testing: the Impact of Specimen
Geometry on Quality Control Testing for Material Performance Assessment," Concreep
6: Creep, Shrinkage, and Curability Mechanic of Concrete and Other Quasi-Brittle
Materials, ed., Ulm, F. J., Bazant, Z. P., and Wittman, F. H., Elsevier, August 22-24
Cambridge MA 2001, pp. 645-651.
56

Shah, S.P., Karaguler, M.E., and Sarigaphuti, M., "Effects of shrinkage-reducing


admixtures on restrained shrinkage cracking of concrete," ACI Materials Journal, Vol.
89, No. 3, 1992, pp 291-295.
en

Bentur, A., and Mindess, S., Fibre Reinforced Cementitious Composites, 1990, 449 p.
CO

American Association of State Highway and Transportation Officias, 1999.

250

Wang, K., Jansen, D, Shah, S., Karr, A., "Permeability Study of Cracked Concrete,"
Cement and Concrete Research, Vol. 27, No. 3, pp. 381-393, 1997
60

Emmons, P. H., "Concrete Repair and Maintenance Illustrated", American Society for
Testing and Materials (U.S. Code Organization), 1994, pp. 155-164.
61

Li, V. C , "High Performance Fiber Reinforced Cementitious Composites as Durable


Material for Concrete Structure Repair," International Journal for Restoration of
Buildings and Monuments, Vol. 10, No 2, pp 163-180, 2004.
62

American Concrete Institute, Concrete Repair Guide (ACI546R-04), American


Concrete Institute. Farmington Hills, MI. 2004

63

FEMMASSE MLS: Computer Program for the Analysis of the Thermal and
Mechanical Behavior of Hardening Concrete - User Manual 2006.
64

Zhou, J., Li, M., Ye, G., Schlangen, E., Breugel, K. V., and Li, V. C , "Modeling the
Performance of ECC Repair Systems under Differential Volume Changes," Proc. 2nd
Int'l Conf on Concrete Repair (ICCRRR), Rehabilitation and Retrofitting, Cape Town, S.
Africa, November, 2008, ppl005-1009. Leiden: CRC Press/Balkema.
65

Zhou, J., Ye, G., Schlangen, E., and van Breugel, K., "Modelling of Stresses and
Strains in Bonded Concrete Overlays Subjected to Differential Volume Changes,"
Theoretical Applied Fracture Mechanics, No. 49, Vol. 2, pp. 199-205.

66

Ackermann, G., "Structural behaviour of joints in composite structures of precast


elements and in-situ concrete," Adherence of Young on Old Concrete, Proc. 2n Bolomey
Workshop. Sion/Sitten, Switzerland, 1-2 April 1993, F.H. Wittmann (ed.),
Unterengstringen: Aedificatio, 1994, pp. 115-127.

251

CHAPTER 5

Influence of Material Ductility on Reflective Cracking in Overlay Repairs

Chapter 5 addresses the reflective cracking issues that are prevalent in concrete
repairs, especially pavement overlay repairs.

Reflective cracking is another major

cracking mechanism that greatly affects concrete repair service life. It is caused by the
brittle response of overlay material under high stress concentration induced by preexisting joints or cracks in the substrate concrete. The approach in this study emphasizes
the inherent ductility and strain hardening behavior of ECC repair materials to diffuse
stress concentration, thereby fundamentally eliminating reflective cracking.

To

demonstrate this concept, experimental studies were conducted on a simulated HES-ECC


overlay repair system, to evaluate its early-age and late-age behavior subjected to
monotonic and fatigue flexural loading. Compared to the control HES-Concrete overlay
system, the HES-ECC overlay system exhibits 100% increased load carrying capacity
and significantly prolonged fatigue life, both at ages of 6 hours and 28 days. This

252

methodology of suppressing reflective cracking can be generalized as translating the


ductility of the repair material to the durability of the repair system, when subjected to
stress concentration from existing cracks.

5.1

Introduction
In concrete repairs, especially overlay repairs, premature reflective cracking is

often the ultimate failure mode that limits the repair service life1'2'3. Concrete or hot mix
asphalt (HMA) overlays primary methods for rehabilitating distressed asphalt and
Portland cement concrete pavements subjected to moderate and heavy traffic, including
highways, bridge decks and airfields4. Before the more expensive option of removal and
re-construction of existing pavements are undertaken, overlays can be used to provide
smooth surfaces while utilizing the structure of existing concrete pavements 5 ' 6 .
Reflective cracking is often the first distress that occurs in overlays7. It is characterized
as new cracks reflected through the overlay from pre-existing cracks or joints in the
substrate (Figure 1.6) under repeated traffic loads.

In most cases cracks propagate


o

through the new overlay during the first few years of service . Under severe conditions, a
reflective crack can reach the surface of a 150 mm (6 in.) overlay in less than one year9.
Reflective cracking not only reduces the overlay load carrying capacity and service life,
but also provides a path for aggressive agents to infiltrate the pavement or bridge deck
structure and cause early deterioration10.
Traditionally, efforts undertaken to reduce reflective cracking in overlay repairs
have fallen into three broad categories11: (a) concrete slab fracturing, which includes
rubblization, crack and seat (for plain concrete pavement), and break and seat (for

253

reinforced concrete pavement); (b) stress-relieving interlayer, which is used between the
concrete substrate and overlay, and includes paving fabric, modified HMA interlayer,
geogrid, and other proprietary interlayers; and (c) modified overlay, which includes
increasing overlay strength, increasing overlay thickness, adding fiber reinforcement,
introducing saw and seal joints in HMA overlay, etc.

If used appropriately, these

methods can delay reflective cracking to a certain extent, although they will not
fundamentally eliminate the fracture failure of the overlay.

The advantages and

drawbacks associated with these techniques will be briefly discussed below.


Through the crack and seat technique (Figure 5.1), numerous cracks are created in
the existing concrete pavement slab to allow deformation of the concrete to occur at more
locations and diffuse the stress concentration12. Similarly, the break and seat technique
breaks the pavement into shorter slabs, thereby reducing the movements at the joints and
cracks to prevent reflective cracking. However, these initial fracturing processes have
not been successful13. Even with the reduced movements of the smaller slabs, cracks
eventually reflect through the overlay. Rubblization (Figure 5.2) involves an extreme
reduction of the slab length. It breaks the existing pavement into fragments that range in
size from grains of sand to approximately 100 mm (4 in.) to 200 mm (8 in.) in width. As
it completely destroys the concrete's slab action by breaking it down into crushed-stone
sized fragments, it is an attempt to turn the existing pavement into base course.
Rubblization is generally considered to be a more favorable approach to reducing
reflective cracking among these three repair options14'15. However, there is one notable
shortcoming to this analogy: a rubblized pavement has no gradation or density control,
which are required by specifications for base course16. Also, rubblization destroys the

254

main structural component of the pavement system and reduces its load carrying capacity,
such that a thicker overlay is required to compensate17'18.

Although perceived as a

rehabilitation technique to delay reflective cracking and quick deterioration of the overlay,
rubblization is actually a destructive procedure; it destroys the structure of the existing
1 O 1 Q ^A

concrete slab and weakens the entire pavement system ' .


As the concrete slab fracturing technique sacrifices the pavement integrity to
reduce stress in the overlay induced by substrate movements, the stress-relieving
interlayer method relies on an additional low-modulus layer of material to diffuse stress
concentration.

Stress-relieving interlayer performance can vary greatly; the type of

interlayer chosen depends on overlay thickness, interlayer placement, base layer stability,
and surface preparation21. Interlayers, such as paving fabrics (Figure 5.3), can form a
waterproof layer of protection above the base layer, even after severe reflective
cracking22, and some studies have found they can retard development of overlay
cracks21'23'24.

However, previous studies, including a large-scale assessment by the

Illinois DOT25, have found that fabric interlayers and HMA interlayers (Figure 5.4) only
retard reflective cracking by at most 3 years and are not cost effective.
As opposed to the repair methods described above, the modified overlay
technique relies on improving the performance of the overlay itself. Increasing overlay
thickness or increasing overlay material strength (e.g. using high strength concrete) both
emphasizes increasing the margin between the traffic-induced stress in the overlay and
the overlay material's tensile strength. A thicker overlay may delay the reflective crack
from reaching the overlay surface if the crack is propagating by fatigue mode. These two
methods have shown marginal improvement on overlay resistance to reflective cracking26,

255

mainly because they do not provide abatement for tensile stress concentration in the
overlay.

Moreover, increasing overlay thickness may compromise cost-effectiveness

because it greatly increases material usage, considering the long miles of pavement.
Increasing the compressive strength of concrete materials also increases the material
Young's modulus, correspondingly increasing the stress level at the same level of
deformation. It also increases brittleness of the concrete material, leading to an increased
susceptibility to fracture failure27. As for the saw and seal joint method (Figure 5.5), the
state DOTs have documented their experiences with marginal to good results28'29.
Secondary reflective cracking can occur, unless the saw cut is made within 25 mm (1 in.)
from the existing joint/crack in substrate. The timing of when the cut is made has been
found to be critical for preventing reflective cracking28'29. Furthermore, the presence of
joints in overlay may necessitate additional maintenance11.
The effectiveness of current approaches in eliminating reflective cracking in
pavement overlays is limited30. For example, according to the Illinois DOT31, on average
HMA overlay with rubblized concrete pavement can extend pavement service life for 1-2
years.

This amount of extension in service life is beneficial, yet not significant

considering the continuously aging and deteriorating US highway infrastructure, and the
declining funding for repair and rehabilitation.

A new approach, which effectively

addresses this repair failure mode and greatly extends the service life of the overlay
system, is necessary for significantly improving the durability of concrete infrastructure.

256
\

5.2

Ductile ECC Overlay


The proposed approach in this study utilizes the inherent tensile ductility and

strain hardening behavior of ECC materials to fundamentally eliminate reflective


cracking initiated from underlying cracks in the substrate concrete. Without relying on
an increase in overlay thickness, fracturing substrate concrete, and application of stress
relieving interlays, a ductile overlay would be able to diffuse stress concentrations and
prevent propagation of cracks to the overlay surface. This concept, which is investigated
in an overlay case within this study, can be generalized to many surface repair scenarios
where underlying cracks or joints exist in substrate concrete, e.g. surface repairs of dams,
building floors, retaining walls and viaducts, tunnels, and pools.
The concept of ductile ECC overlays was first investigated by Lim and Li
(1997)32 and Kamada and Li (2000)33 using a special version of ECC, containing a Type I
Portland cement based matrix and short polyethylene fibers (PE-ECC). It was found that
-under monotonic flexural loading, PE-ECC effectively diffused stress concentrations
through its microcracking behavior at the base of the overlay.

The multiple

microcracking process of the PE-ECC overlay absorbed over ten times the energy that
was absorbed in the brittle fracture (reflective cracking) process of the control concrete
overlay. Later, Zhang and Li (2002)34 conducted an experimental study and analytical
analysis on the overlay performance of another version of ECC, containing Type I
Portland cement based matrix and short polyvinyl alcohol fibers (PVA-ECC), under
monotonic and fatigue flexural loading.

The experimental results revealed great

improvements in both load carrying capacity and fatigue life of the PVA-ECC overlay
system, compared to the control concrete overlay system. Recently, Zhang et al. (2007)

257

and Qian (2007) u looked at the overall sustainability of unbonded concrete overlays,
HMA overlays, and PVA-ECC overlays using economic, environmental, and social
metrics. It was found that ECC overlays were able to improve the sustainability of rigid
pavement overlays by reducing surface roughness (improving vehicle fuel economy) and
extending overlay service life. A 39.2% or 55.7% reduction in total costs was able to be
achieved when using ECC overlay compared with concrete or HMA overlay.

ECC

overlay was also found to reduce total primary energy consumption by 75% compared
with HMA overlay and greenhouse effect was reduced by 32 to 37% compared with
concrete and HMA overlay.
These research findings revealed the great potential of ECC as a durable overlay
material that extends the service life of pavement infrastructure. The motivation behind
this research is to understand the performance of the new HES-ECC material, which was
developed in this thesis specifically for concrete repair applications, in term of its
resistance to reflective cracking. Differing from previously investigated ECC materials,
HES-ECC contains a Type III Portland Cement based matrix, and simultaneously
possesses unique tensile properties as well as rapid early-age strength gain rate. Of
special interest is the early-age performance of HES-ECC overlays. This study looked at
the 6-hour and 28-day cracking behavior, load carrying capacity, deformation carrying
capacity, and fatigue life of the HES-ECC overlay system compared to a control HESConcrete overlay system. Influence of concrete substrate surface preparation on the
overlay system's overall performance was also evaluated and reported here.

258

5.3

Experimental Program

5.3.1

Specimen Configuration and Loading Conditions


Traffic loading induces flexural stresses in the overlay and substrate slab. These

stresses in the overlay are at a maximum at and/or close to the existing cracked section,
because there is no load transfer through the existing cracks36'37. Additionally, due to the
dimensional incompatibility between the cracked pavement substrate and overlay, a
certain amount of interfacial delamination, starting from the existing cracks, will form .
As a result, the maximum flexural stress zone is extended to the region from the existing
cracks to the endpoints of the delamination zone on each side.
To simulate the scenario described above, an overlay repair system subjected to
four-point bending ' '

was employed in this study. As shown in Figure 5.6, this

system consisted of a 51 mm (2 in.) thick layer of repair material cast on a 51 mm (2 in.)


thick old concrete substrate. This concrete substrate, which was 356 mm (14 in.) long
and 76 mm (3 in.) wide, contained a vertical crack that simulated an existing crack or
joint in pavement, and a 51 mm (2 in.) long horizontal interfacial crack that simulates the
initial delamination between the overlay and substrate. The specimen was subjected to a
four-point bending load, with the load-deflection curve monitored during testing. The
deflection of the specimen at the center was measured by two linear variable differential
transducers (LVDTs), which were mounted on both sides at the center of the overlay.
Both the monotonic and cyclic loading tests were conducted on a 56 kip load
capacity MTS-810 testing machine equipped for closed-loop testing. The monotonic
flexural test was carried out with deformation controlled by the displacement of the
actuator.

The displacement was increased at a constant rate of 0.1 mm/min (0.004

259

in./min) according to ASTM CI018 . The fatigue flexural test was conducted with load
controlled using a sinusoidal waveform with a frequency of 2 Hz. Loading followed a
ramp function to the maximum load Pmax at a rate of 0.1 kN/s (22.5 lb/s), followed by
sine waveform fatigue loading cycles. The maximum load (Pmax) was 90%, 80% and
70% of the average load-carrying capacity of the overlay system, which was obtained
from the monotonic flexural test results.

The ratio R between the minimum and

maximum load levels was constant at R=Pmin/Pmax=0.1. Three specimens were tested for
each scenario.

5.3.2

Materials and Specimen Preparation


HES-ECC (developed in chapter 2) and HES-concrete (control mix) were

investigated in this study. They were the same as those high early strength materials
investigated in the repair layer system subjected to restrained shrinkage in Chapter 4.
Their mixing proportions and mechanical properties are summarized in Table 4.2, 4.3 and
4.4. The HES-Concrete mixture consisted of coarse aggregate (CA) with 10 mm (0.4 in)
nominal grain size, Portland type III cement (C), sand (S) and water (W).
Superplasticizer (SP) was used to achieve sound workability. Accelerating admixtures
(AC) were also included to accelerate the material's strength development and setting
processes. HES-Concrete specimens were tested to have average compressive strength
(fc') of 49.9 MPa (7234 psi) at the age of 7 days, and 54.2 MPa (7860 psi) at the age of
28 days. Under uniaxial tensile loading, HES-Concrete is a brittle material with sudden
fracture failure on a single crack plane.
The HES-ECC mixture is comprised of Type III Portland cement (C), water (W),

260

silica sand (S) with 110 \xm (3.94xl0"3in.) nominal grain size, polystyrene beads with a
size of 4 mm (0.157 in.) as aggregates, superplasticizer, accelerating admixtures, and 2%
(Vf) polyvinyl-alcohol (PVA) fibers. These PVA fibers had a length of 12 mm (0.472 in)
and diameter of 40 um (1.54xl0"3 in). The HES-ECC was self-consolidating in the fresh
state. The polystyrene beads served as deliberately introduced initial flaws to assist in
triggering the multiple microcracks during tensioning of the composite. As measured in
Chapter 2, the HES-ECC mixture had average compressive strength of 47.5 MPa (6885
psi) at the age of 7 days, and 55.6 MPa (8063 psi) at the age of 28 days. Its average
Young's modulus was 20.6 GPa (2986 ksi) at the age of 7 days, and 23.2 GPa (3365 ksi)
at the age of 28 days, which was lower than that of HES-Concrete (26.2 GPa (3803 ksi)
at 7 days; 27.8 GPa (4025 ksi) at 28 days) due to the absence of coarse aggregates (CA).
The HES-ECC mixture also had a tensile strain capacity between 3 - 6% at all ages. Its
flexural strength (MOR) was 13.6 MPa (1965 psi) at the age of 7 days, and 15.1 MPa
(2188 psi) at the age of 28 days.
The substrate concrete had the same material composition as HES-Concrete,
except that it used ordinary concrete containing Portland type I cement while the HESConcrete repair used Type III Portland cement.
To prepare the specimens, concrete beams with dimensions of 356 mm x 76 mm
x 102 mm (14 in x 3 in x 4 in) were first cast, and demolded after 24 hours. After
demolding, the beams were cured in water at 20+1 C (66-70 F) for 28 days. Then each
beam was cut using a diamond saw into four concrete blocks with size of 178 mm * 76
mm x 51

m m

(7

x3 i n x 2 in) (Figure 5.7 (a)). The blocks were then stored in

laboratory conditions of 45 5% RH and 201 C (66-70 F) for an additional week, at

261

which time smooth plastic tape was used to form the vertical crack and horizontal
interfacial crack before the repair layer was cast (Figure 5.7 (b)).
The contact surfaces of the concrete substrates were prepared in two different
ways: (a) diamond saw cut smooth, and (b) roughened to 7~8mm (Figure 5.8). For the
latter, the substrate surfaces were roughened while in the fresh state by using a chisel to
remove slurry cement from the external surfaces of coarse aggregates. Before placing the
repair layers, the substrate surfaces were re-cleaned with a brush and high-pressure air to
ensure a clean bonding surface, and then they were dampened to an adequate moisture
level. HES-ECC or HES-concrete repair layer was cast on top of two concrete blocks
that formed the concrete substrate. The repair layers were moisture cured for 6 hours and
then demolded. After demolding, the layered specimens were moved into a room with
ambient conditions of 45 5% RH and 201 C (66-70 F). The test was conducted in
the same room at specimen ages of 6 hours and 28 days. Three specimens were tested for
each case.

5.4

Experimental Results

5.4.1

Overlay System Monotonic Flexural Performance


Figure 5.10 plots the flexural stress vs. midpoint deflection curve of the HES-

ECC repair system compared with the HES-Concrete repair system under monotonic
loading. The flexural stress Ojiexurai is the tensile stress at the center base of the overlay
above the vertical crack, and was determined using linear elastic theory:

-J-

a
flexural

(5-1)

/ g

where M is the bending moment at the center of the overlay, b is overlay width, and h is
the overlay thickness. The dimensions of the overlay, instead of the overlay system, were
262

used for calculation of Ojiexurd because there was no load transfer through the vertical
crack. The flexural strength, or modulus of rupture (MOR), which was defined as the
peak flexural stress from the flexural stress vs. deflection curve, is summarized in Table
5.1 and compared in Figure 5.11.

The corresponding deflection and interfacial

delamination length (including the 2 in. initial horizontal crack) upon overlay failure are
summarized in Table 5.2 and 5.3, and Figure 5.12 and 5.13.
The kinking and trapping mechanism32 was demonstrated in the HES-ECC
overlay system at both early age (6 hours) and late age (28 days).

According to

Hutchinson and Suo (1992)39, interfacial cracks can kink into the repair material when the
crack tip satisfied the "interfacial crack kink condition", which contains two terms relative driving force and relative toughness. The relative driving force term can be
analytically or numerically calculated, and the relative toughness term can be
experimentally measured based on the interface toughness and the toughness of the repair
material. Lim (1996)40 found that low initial fracture toughness with a rapid rise R-curve
behavior of repair material, or low first crack strength with a large margin to ultimate
tensile strength, are the essential requirements of the trapping mechanism in a repair
system. These requirements are met by ECC materials (including HES-ECC), but are not
satisfied by traditional concrete (including HES-Concrete) or FRC materials. Figure 5.9
shows the conceptual kinking and trapping mechanism of the ECC/concrete system with
the load vs. displacement relationship. Large amounts of energy are dissipated during
repeated sequences of kinking, ECC multiple microcracking, trapping, and interfacial
crack propagation, so that fracture failure is prevented. In this study, the kinking and
trapping mechanism was clearly observed in the HES-ECC overlay system at ages of 6

263

hours and 28 days. With this mechanism, the HES-ECC overlay system effectively
diffused the stress concentration from the initial vertical crack and horizontal interfacial
crack by trapping the cracks within the HES-ECC overlay, without reflecting them to the
overlay surface (Figure 5.14). Thus, the HES-ECC overlay exhibited a ductile ultimate
failure mode in flexure. In contrast, the control HES-Concrete overlay system exhibited
a brittle fracture failure mode, with the existing crack quickly reflected to the overlay
surface (Figure 5.10). These results indicate that an HES-ECC overlay can be effective
in resisting reflective cracking as early as 6 hours after placement.
Although HES-ECC and HES-Concrete have similar 28-day compressive strength,
the HES-ECC overlay system achieved significantly higher flexural load carrying
capacity than the HES-Concrete overlay system at the same overlay age (Table 5.1 and
Figure 5.11). This is due to the tensile strain-hardening behavior of HES-ECC and the
resulting different cracking mechanisms discussed above. The flexural strength (MOR)
of the HES-ECC overlay was approximately 100% larger than that of the HES-Concrete
overlay, at both ages of 6 hours and 28 days, and in both smooth surface and rough
surface situations.

Furthermore, the deformation capacity of the HES-ECC overlay

system, represented by the overlay midpoint deflection corresponding to the peak flexural
stress, was approximately 700% (smooth surface) and 500% (rough surface) larger than
those of the HES-Concrete overlay system, respectively. The significantly higher load
carrying capacity as well as deformation capacity of HES-ECC overlay suggests it has a
higher resistance to reflective cracking, even in heavy traffic areas.
It should be noted that the flexural strength (MOR) of HES-ECC overlay at age of
6 hours (9.65 MPa [smooth]; 9.51 MPa [rough]) is slightly higher than that of HES-

264

Concrete overlay at age of 28 days (9.09 MPa [smooth]; 8.67 MPa [rough]). This earlyage load carrying capacity of HES-ECC overlay confirms its capability of early opening
to traffic 6 hours after placement. This advantage of HES-ECC overlay, in terms of its
early-age load carrying capacity and resistance to reflective cracking, has not been
demonstrated in other versions of ECC materials or traditional concrete or FRC materials.
Comparing the flexural curves of the HES-ECC repaired system with two
differently prepared substrate surfaces, higher deformation capacity was noted when the
substrate surface was smooth (Table 5.2, Figure 5.10 and 5.12). This phenomena was
consistent with findings from previous studies33'34, and was additionally revealed at the
early age of the HES-ECC overlay. Due to the weak bond of the smooth interface, under
flexural load a larger delamination area was formed between the HES-ECC repair layer
and the concrete substrate.

Thus, a larger section of the HES-ECC overlay (the

delaminated section) was subjected to maximum bending stress and underwent the
kinking-trapping and multiple microcracking processes, leading to a higher deformation
capacity.

The interfacial delamination was measured, including the initial 2-in.

delamination length, and is shown in Figure 5.13. For the HES-Concrete overlay system,
the influence of interface properties on deformation capacity was negligible. In both
smooth and rough interface cases, the initial interface crack kinked out from the interface
into the HES-Concrete overlay and propagated unstably to the overlay surface. This
phenomenon corresponds to a sudden load drop without any interface delamination
(Figure 5.10 and 5.13).
The typical cracking pattern of the HES-ECC overlay system compared with the
HES-Concrete overlay system is shown in Figure 5.14.

265

Corresponding to the peak

applied flexural stress, the HES-Concrete/Concrete overlay system had a sudden failure
with one single crack propagating through the HES-Concrete overlay. The fractured
halves of the overlay specimens separated completely in both the smooth and rough
interface cases. In contrast, the HES-ECC/Concrete overlay system exhibited multiple
microcracking behavior without reflective cracking failure.

Before the HES-ECC

overlay eventually failed in ductile flexural mode, it experienced deflection hardening


with self-controlled crack width below 50 um (2*10~3 in.), in both smooth and rough
interface cases. The results show that the unique kink-trap mechanism in HES-ECC
overlays effectively suppressed the common reflective failure mode in concrete repairs.
Similar behavior was also found in regular ECC (M45) overlays34.

5.4.2

Overlay System Fatigue Performance


The fatigue flexural performances of HES-ECC and HES-Concrete overlay

systems are shown in Figure 5.15, in terms of flexural stress versus fatigue life (S-N).
There was no obvious difference in the crack patterns at failure between the applications
of fatigue and monotonic flexural loading conditions. At a certain number of cycles of
loading, the HES-Concrete overlay system failed with a single crack unstably
propagating to the overlay surface. In contrast, the HES-ECC overlay system underwent
multiple microcracking at the base of the ECC overlay, preventing the initial crack from
unstably "reflecting" through to the surface. It was also observed that as the fatigue
stress level decreased, the number of microcracks also decreased. In this way, it became
more difficult for the HES-ECC overlay to reach saturated multiple microcracking
because the material was under lower stress and strain levels. Such phenomenon was

266

also observed by Qian (2008)11 and Matsumoto et al. (2002) in regular ECC beam
specimens41.

During a fatigue test with constant stress level, fatigue load cycles

repeatedly re-opened the microcracks in HES-ECC overlay, until the fiber bridging
capacity at one of the microcracks was exhausted under fatigue, leading to the final
overlay failure. Because of significant difference in the failure modes of HES-Concrete
and HES-ECC overlay systems under fatigue, it was clearly observed from Figure 5.15
that at the same overlay age of 28 days, and at the same level of applied maximum
flexural stress, the fatigue life of the HES-ECC overlay system was several orders of
magnitude higher than that of the HES-Concrete repair system. Additionally, comparing
the HES-ECC overlay system at the age of 6 hours with the HES-Concrete overlay
system at the age of 28 days, the former had a fatigue life approximately one order of
magnitude higher at the same stress level. These results suggest that when designed with
equal thickness, HES-ECC overlays can have a drastically longer service life than HESConcrete overlays under the same traffic loading conditions. In other words, under equal
service life and traffic loading conditions, the required overlay thickness of an HES-ECC
overlay can be significantly lower than that of an HES-Concrete overlay, leading to large
savings of overlay material, equipment use, and construction time.
The difference in interface roughness did not significantly affect the fatigue life of
the HES-ECC and HES-Concrete layered repair systems. The hypothesis was that the
flexural fatigue life of a beam was dominated by the single crack propagation behavior
under fatigue loading42. For the HES-ECC overlay, its fatigue life until failure was likely
governed by the fiber bridging properties under cyclic loading at the dominant single
crack, and did not significantly depend of the number of microcracks formed. However,

267

the number of microcracks was positively correlated with the deformation capacity of the
HES-ECC overlay upon fatigue failure, as was similarly observed in the monotonic
flexural testing results.

5.5

Conclusions
The common durability concern in concrete repairs, reflective cracking, can be

fundamentally prevented through the high tensile ductility and strain-hardening behavior
of ECC materials. This was demonstrated in this study of an HES-ECC overlay system
with an existing vertical crack in the substrate concrete and a small amount of interfacial
delamination. Under monotonic and fatigue four-point bending, the HES-ECC overlay
exhibited a "kinking and trapping" mechanism, and multiple microcracking behavior
with controlled crack width below 50 um (2><10"3 in.). Through this process a large
amount of energy was dissipated, thus preventing the brittle reflective cracking failure
mode that was observed in the control HES-Concrete overlay system.
Due to the ductile tensile behavior and unique "kinking and trapping" mechanism
of the HES-ECC overlay, it exhibited 100% higher load carrying capacity and several
orders of magnitude longer fatigue life compared to an HES-Concrete overlay with
similar compressive strength and the same age.

This suggests that significant

improvements are possible in overlay design when HES-ECC is used: assuming the
same loading conditions, (a) if designed with same overlay thickness, the HES-ECC
overlay is expected to have a significantly longer service life than traditional concrete
overlays; (b) if designed with same service life, the HES-ECC overlay thickness can be
greatly reduced compared to traditional concrete overlays, leading to a significant amount

268

of saving of materials, construction equipment, and construction time; or (c) a


compromise of either parameter, i.e. reduced overlay thickness as well as prolonged
overlay service life, can be achieved simultaneously.
It should be noted that the high resistance to reflective cracking of the HES-ECC
overlay through multiple microcracking and the "kinking and trapping" mechanism was
also observed at the early age of 6 hours. The load carrying capacity of the HES-ECC
overlay at the age of 6 hours and the control HES-Concrete overlay at the age of 28 days
was the same. Furthermore, at the same stress level, the fatigue life of the HES-ECC
overlay at the age of 6 hours was approximately one order magnitude larger than that of
the control HES-Concrete overlay.

All of these results highlight the significant

advantages of the newly-developed HES-ECC material in fast and durable repair


applications, with shortened operations shutdown time as well as greatly prolonged
service life. These potentials in HES-ECC have not been demonstrated in other versions
of ECC materials, or traditional concrete or FRC materials.

269

Table 5.1 - Modulus of rupture of HES-ECC and HES-Concrete overlays under


monotonic loading.

Overlay Age
HES-ECC [Smooth]
Average
HES-ECC [Rough]
MOR,
psi (MPa)

Average
HES-Concrete [Smooth]
Average
HES-Concrete [Rough]
Average

6 hours
1513.32 (10.43)
1315.06 (9.07)
1371.62 (9.46)
1400.00 (9.65)
1527.83 (10.53)
1337.68 (9.22)
1270.97 (8.76)
1378.83 (9.51)
657.02 (4.53)
604.81 (4.17)
570.00 (3.93)
610.61 (4.21)
548.24 (3.78)
661.37 (4.56)
510.53 (3.52)
573.38 (3.95)

28 days
2725.38 (18.79)
3095.92 (21.35)
2877.19 (19.84)
2899.49 (19.99)
2620.75 (18.07)
2861.51 (19.73)
2773.58 (19.12)
2751.95 (18.97)
1370.99 (9.45)
1201.88 (8.29)
1381.16 (9.52)
1318.01 (9.09)
1391.18 (9.59)
1318.67 (9.09)
1064.69 (7.34)
1258.18 (8.67)

Table 5.2 - Center deflection (upon failure) of HES-ECC and HES-Concrete overlays
under monotonic loading.
Overlay Age
HES-ECC [Smooth]
Average
HES-ECC [Rough]
in. (mm)

Average
HES-Concrete [Smooth]
Average
HES-Concrete [Rough]
Average
270

6 hours
0.198 (5.02)
0.180 (4.56)
0.206 (5.22)
0.194 (4.93)
0.118 (3.00)
0.100 (2.54)
0.131 (3.34)
0.117 (2.96)
0.011 (0.28)
0.015 (0.37)
0.018 (0.45)
0.014 (0.37)
0.016 (0.41)
0.015 (0.37)
0.015 (0.39)
0.015 (0.39)

28 days
0.138 (3.51)
0.132 (3.35)
0.171 (4.34)
0.747 (3.73)
0.075 (1.91)
0.112 (2.85)
0.095 (2.41)
0.094 (2.39)
0.017 (0.43)
0.016 (0.41)
0.017 (0.43)
0.017 (0.42)
0.015
0.017
0.017
0.016

(0.38)
(0.43)
(0.43)
(0.42)

Table 5.3 - Interfacial delamination (upon failure, including initial delamination) of


HES-ECC and HES-Concrete overlays under monotonic loading.
Overlay Age
HES-ECC [Smooth]
Average
HES-ECC [Rough]
Delamination
in. (mm)

Average
HES-Concrete [Smooth]
Average

6 hours
7.34 (186.44)
7.62 (193.55)
7.93 (201.42)
7.63 (193.80)
4.03 (102.36)
4.51 (114.55)
3.89 (98.81)
4.14 (105.24)
2.00 (50.80)
2.00 (50.80)
2.12 153.85}
2.04 (51.82)

28 days
7.07 (179.58)
6.45 (163.83)
6.12 (155.45}
6.55 (166.29)
4.20 (106.68)
4.05 (102.87)
3.51 (89.15)
3.92 (99.57)
2.08 (52.83)
2.12 (53.85)
2.05 (52.07)
2.08 (52.92)

2.00
2.01
2.07
2.03

2.05
2.10
2.06
2.07

HES-Concrete [Rough]
Average

271

(50.80)
(51.05)
(52.58)
(51.48)

(52.07)
(53.34)
(52.32)
(52.58)

/.-

?*?

.Jl.

,, ~f
Figure 5.1 - Crack-and-seated pavement surface.

'-J

>w

:;J*S***i,,-:

'

^" \ $

,r

' **' 1* %-**'

Figure 5.2 - Rubblized pavement surface.

272

-'

.'.uwy IS

. . t t t t e t ;"
. . "S"* t * * . * " *

#L

3-$$$k

U
,

'%Ni- ...

Figure 5.3 - Paving fabric installation.

'-V ^

Figure 5.4 - HMA inter layer installation.

273

"^MkJ&^^HSiife

Figure 5.5 - Overlay saw and seal

153 mm

XL,
Overlay

Initial Unbonded

Base Concrete
kf

Existing Crack

XT

O
4r->|

76 mm

51 mm
305 mm
356 mm

Figure 5.6 - Specimen configuration and loading conditions.

274

Cut with
diamond saw

76 mm
356 m m

(a)

Repair Layer

Smoothly cut or
roughened

I
Concrete
Substrate

Plastic tapes for pre-introducing crack


in concrete substrate and interfacial
(b)
Figure 5.7 - Specimen preparation, (a) Concrete beams were cut into four blocks, two of
which formed the concrete substrate with an initial vertical crack, (b) Overlay repair layer
was cast onto the concrete substrate with an initial vertical crack and small amount of
horizontal interfacial crack.

275

Figure 5.8 - Concrete substrate surface preparation.

lilltll

Load

Displacement
Figure 5.9 - "Kinking and trapping" mechanism27.

276

Deflection (mm)
-6

Smooth Interface
+ 20
re
Q.

-r 15
D)
C

a>k .
*-

V)

3
X
V

+5

HES-ECC 6 hours

0.05

0.1

0.15

0.2

0.25

Deflection (in)

(a)
Deflection (mm)

3500

Rough Interface
3000 H

20

HES-ECC_28 days
15 ~

HES-ECC 6 hours

, 0 |

-r 5

HES-Concrete 28 days
0.05

0.1

0.15

Deflection (in)

0.2

0.25

(b)

Figure 5.10 - Flexural behavior of HES-ECC and HES-Concrete layered repair system
under monotonic loading: (a) smooth interface; (b) rough interface.
277

JOUU

6 hours

3000

20

2500
\

2000

1500

15 I?
a.

10 i*
o

1000

500 -

o -

1
HES-ECC
(Smooth)

1
HES-ECC
(Rough)

1
HES-Concrete
(Smooth)

1
HES-Concrete
(Rough)

(a)

oouu

28 days

ooo

3000

20

2500
!

2000

1500

15

EL

10 a:

1000
500
0

1
HES-ECC
(Smooth)

1I
HES-ECC
(Rough)

HES-Concrete
(Smooth)

1
HES-Concrete
(Rough)

(b)

Figure 5.11 - Modulus of rupture (MOR) of HES-ECC and HES-Concrete overlay


systems under monotonic loading, (a) Overlay age of 6 hours, (b) Overlay age of 28
days.

278

0.25

6 hours
0.20

- 4

0.15
X
X
X

0.10

- 3
- 2

0.05

0.00 HES-ECC
(Smooth)

HES-ECC
(Rough)

HES-Concrete
(Smooth)

- 1

Deflection (mm)

- 5

o
Deflection (in.)

O
h 0
HES-Concrete
(Rough)

(a)

u.zo

28 days
0.20

o
c
c
o

0.15

X
X
X

0.10

o
o

**a

0.05
0.00 -

[ 6

1
1

HES-ECC
(Smooth)

HES-ECC
(Rough)

HES-Concrete
(Smooth)

0
1

E.
c
o
o

'^
<u

HES-Concrete
(Rough)

(b)

Figure 5.12 - Deflection at failure of HES-ECC and HES-Concrete overlay systems


under monotonic loading, (a) Overlay age of 6 hours, (b) Overlay age of 28 days.

279

y -I
6 hours

(in)

7
6

- 150

linatio

re
0)

- 200

5
4
3

Initial
. Delamination

X
*

- 100

E
E

lation

8 -

"F
A

2 -

- 50

1 0

U i
HES-ECC
(Smooth)

HES-ECC
(Rough)

HES-Con crete
(Smooth)

HES-Concrete
(Rough)

(a)

28 days

200

Initial
Delamination

150 ,,

c
o
100 ra
c
E
50 "55

*
x

HES-ECC
(Smooth)

HES-ECC
(Rough)

HES-Concrete
(Smooth)

HES-Concrete
(Rough)

(b)

Figure 5.13 - Interfacial delamination upon failure of HES-ECC and HES-Concrete


overlay systems under monotonic loading, (a) Overlay age of 6 hours, (b) Overlay age
of 28 days.

280

(a) HES-ECC overlay system (smooth interface)

(b) HES-ECC overlay system (rough interface)

JlllMMlHfMPw'TfiffTT^^

(c) HES-Concrete overlay system (smooth interface)

(d) HES-Concrete overlay system (rough interface)


Figure 5.14 - Crack pattern of HES-ECC and HES-Concrete overlay systems at overlay
age of 28 days. Epoxy glue was spread over the cracking sections of HES-ECC overlays
(a) and (b), to better visually reveal multiple microcracking.
281

3500

re
0.
5,
u>

to
"re
3
X

_o
LL.

E
3
E
"S
re

Cycle (log N)
HES-ECC, 28d, rough

xHES-Concrete, 28d, Smooth

AHES-Concrete, 28d, Rough

OHES-ECC, 28d, smooth

HES-ECC, 6h, Smooth

a"HES-ECC, 6h, Rough"

Figure 5.15 - Fatige stress-fatigue life relation (log) for HES-ECC and HES-Concrete
overlay systems.

282

References:
1

Tayabji, S. D. and Okamoto, P. A., "Thickness Design of Concrete Resurfacing," Proc,


3r Int'l Conf. on Concrete Pavement Design and Rehabilitation, 1985, pp. 367-379.
2

ERES Consultants, Inc., "Evaluation of Unbonded Portland Cement Concrete Overlays,"


NCHRP Report 415, Transportation Research Board, Washington, D. C , 1999.
3

Huang, Y. H., Pavement Analysis and Design, 2nd edition, Pearson Education, Upper
Saddle River, NJ 07458.

MDOT, Pavement Design and Selection Manual, Michigan Department


Transportation, Lansing, Michigan, 2005, 65p.

of

Jones, H. W., "Asphalt Overlays; Crack & Seat vs. SAMI vs. Rubblizing vs. Removal
How to Choose," 2009 Michigan Concrete Repair Seminar, Michigan's Local Technical
Assistant Program, Michigan Concrete Paving Association, Michigan Transportation
Asset Management Council, Howell, Michigan, 2009.
6

National Concrete Pavement Technology Center, Guild to Concrete Overlays:


Sustainable Solutions for Resurfacing and Rehabilitating Existing Pavements, Second
Edition, September 2008.
7

Transport Research Board (TRB) of the National Academies, "Crumb Rubber Modified
(CRM) Stress-Absorbing Membrane Interlayers for New Pavement Construction,
Maintenance and Repair of Existing Pavements and Bridge Decks," Research Needs
Statements, http://rns.trb.org/dproject.asp?n= 13076, accessed on August 31, 2009.
8

Newman, K., and Shoenberger, J. E., "Polymer Concrete Micro-Overlay for Fuel and
Abrasion Resistant Surfacing: Laboratory Results and Field Demonstrations," US Army
Engineer Research and Development Center, Vicksburg, Mississippi, USA, Presented for
the 2002 Federal Aviation Administration Airport Technology Transfer Conference, May
2002.
9

RILEM Technical Committee 157-PRC, "Third RILEM International Conference


Reflective Cracking in Pavements," Materials and Structures, Vol. 30, June 1997, pp.
317-318.
National Concrete Pavemet Technology Center, Guild to Concrete Overlay Solutions,
January 2007.
Qian, S., "Influence of Concrete Material Ductility on the Behavior of High Stress
Concentration Zones," Ph.D. Dissertation, University of Michigan, 2007.

283

Harris, G. K., "Cracking & Seating to Retard Reflective Cracking - Fremont County,"
Final Report, Iowa Highway Research Board Research Project HR-279, 1993.
13

American Concrete Pavement Association, "Rubblizing of Concrete Pavements: A


Discussion of its Use," Concrete Pavement Engineering and Research,\99S.
DelDOT, Design Guidance Memorandum on Concrete Pavement Rubblization,
Delaware Department of Transportation, 2002.
15

Ceylan, H., Mathews, R., Kota, T., Gopalakrishnan, K., and Coree, B. J.,
"Rehabilitation of Concrete Pavements Utilizing Rubblization and Crack and Seat
Methods," Final Report submitted to Iowa Highway Research Board and Iowa
Department of Transportation, 2005, 156p.
16

Guide Specifications for Highway Construction, American Association of State


Highway and Transportation Officials, Washington, DC, 1993.
17

FHWA, "Crack and Seat performance," Review Report, Demonstration Projects


Division and Pavement Division, Federal Highway Administration, 1987.
18

Freeman, T., "Final Report Evaluation of Concrete Slab Fracturing Techniques in


Mitigating Reflective Cracking Through Asphalt Overlays," submitted to Virginia
Transportation Research Council, Charlottesville, Virginia, VTRC 03-R3, 2002, 18p.
19

Buncher, M., and Jones, H. W., "State-of-the-Practice: Rubblization of Airfield


Pavements," Concrete Pavement Engineering and Research, American Concrete
Pavement Association, 1998.
Transportation Research Board, "Rubblization of Portland Cement Concrete
Pavements," Transportation Research Circular E-C087, Washington, DC, January 2006.
21

Zou, W, Wang, Z, and Zhang, H, "Field Trial for Asphalt Pavements Reinforced with
Geosynthetics and Behavior of Glass-Fiber Grids," Journal of Performance of
Constructed Facilities, Vol. 37, Sep/Oct 2007, pp. 361-367.
22

Buttlar, W, Bozkurt, D, Dempsey, B, "Cost-Effectiveness of Paving Fabrics Used to


Control Reflective Cracking," Transportation Research Record, No. 1730, 2000, pp. 139149.

Blankenship, P, Iker, N, Drbohlav, J, "Interlayer and Design Considerations to Retard


Reflective Cracking," Transportation Research Board, No. 1896, 2004, 177-186.
Penman, J, Hook, K.D., "The use of geogrids to retard reflective cracking on airport
runways, taxiways and aprons," Pavement Cracking: Mechanisms, Modeling, Detection,
Testing and Case Histories, 2008.

284

Buttlar, W, Bozkurt, D, Dempsey, B, "Cost-Effectiveness of Paving Fabrics Used to


Control Reflective Cracking," Transportation Research Record, Vol. 1730, pp. 139-149.
26

Jiang, Y., and McDaniel, R. S., "Application of Cracking and Seating and Use of
Fibers to Control Reflective Cracking, " Transportation Research Board, No. 1388, 1993,
pp. 150-159.
27

Li, V. C , and Stang, H., "Elevating FRC Material Ductility to Infrastructure


Durability," Proceedings of 6th RILEM symposium on FRC, Varenna, Italy, 2004, pp.
171-186.
28

Kilareski, W. P., and Bionda, R. A., "Structural Overlay Strategies for Joints Concrete
Pavements," Vol. 1, Report No. FHWA-RD-89-142, FHWA, Washington, D. C , 1990.
29

Hall, K. T., Correa, C. E., and Simpson, A. L., "Performance of Rigid Pavement
Rehabilitation Treatments in the SPS-6 Experiment, " TRB, 1823, 2003, pp. 64-72.
30

Lepech, M. D., Keolain, G. A., Qian, S., and Li, V. C , "Design of Green Engineered
Cementitious Composites for Pavement Overlay Applications," Int'I Association for Life
Cycle Civil Engineering conference (IALCCE 08), Italy, June, 2008, pp. 837-842.
31

Lippert, D. L., Rubblizing in Illinois (Presentation), Illinois DOT, 2003, 65p.

32

Li, V.C. and Lim, Y.M., "Trapping Mechanism of Interface Crack at


Concrete/Engineered Cementitious Composite Bimaterial", in Proc. of First International
Conference on Damage and Failure of Interfaces - DFI-1 Vienna, pp. 405-409, Austria,
22-24 September, 1997.
33

Kamada, T. and V.C. Li, "The Effects of Surface Preparation on the Fracture Behavior
of ECC/Concrete Repair System," J. of Cement and Concrete Composites, Vol. 22, No. 6,
pp.423-431,2000.
34

Zhang J., and Li, V.C, "Monotonic and Fatigue Performance in Bending of Fiber
Reinforced Engineered Cementitious Composite in Overlay System ," Journal of Cement
and Concrete Research, Vol. 32, No. 3, 2002, pp. 415-423.
35

Zhang, H., G. A. Keoleian, S. Qian, and V. C. Li,"Dynamic Life Cycle Modeling of


Pavement Overlay System: capturing the impacts of users, construction, and roadway
deterioration", The 4th Int'I Conference of the International Society for Industrial
Ecology, ISIE, June 17-20, 2007.
36

Cable, J. K., Fanous, F. S., Ceylan, H., Wood, D., Frentress, D., Tabbert, T., Oh, S. Y.,
Gopalakrishnan, K., "Design and Construction Procedure for Concrete Overlay and
Widening of Existing Pavements," FHWA Technical Report DTFH61-01-X-00042
(Project 6), IHRB Project TR-511, September 2005.

285

Kim, J., and Buttlar, W. G., "Analysis of Reflective Crack Control System Involving
Reinforcing Grid Over Base-Isolating Interlayer Mixture," Journal of Transportation
Engineering, Vol. 128, No. 4, 2002, pp. 375-384.
38

ASTM C1018-97 Standard Test Method for Flexural Toughness and First-Crack
Strength of Fiber-Reinforced Concrete (Using Beam With Third-Point Loading).

39

Hutchinson, J. W., and Suo, Z., "Mixed Mode Cracking in Layered Materials,"
Advances in Applied Mechanics, Vol. 29, 1992, pp. 63-191.
40

Lim, Y. M., Interface Fracture Behavior of Rehabilitated Concrete Infrastructures


Using Engineered Cementitious Composites, Ph.D. Thesis, The University of Michigan,
Ann Arbor, 1996.
41

Matsumoto, T. P., Suthiwarapirak, P., and Kanda, T., "Mechanisms of Multiple


Cracking and Fracture of DFRCCs under Fatigue Flexure," Proceedings of the JCI
International Workshop on Ductile Fiber Reinforced Cementitious Composites (DFRCC)
-Application and Evaluation (DFRCC -2002), Takayama, Japan, Oct. 2002, pp. 259-268.
42

Zhang, J., Stang, H., and Li, V. C , "Fatigue Life Prediction of Fiber Reinforced
Concrete Under Flexural Load," International Journal of Fatigue, Vol. 21, No. 10, 1999,
pp. 1033-1049.

286

CHAPTER 6

Transport Properties of ECC under Chloride Exposure

This chapter addresses the chloride penetration stage in a typical repair


deterioration process, before and after cracking occurs at different levels of strain.
Chloride diffusion driven by a chloride concentration gradient, which is the predominant
mechanism of chloride transport for most of the concrete structures exposed to deicing
salts, and airborne or waterborne chloride ions, is the focus of this study. Experiments
are conducted under combined mechanical (flexural) and environmental (3% chloride
solution ponding) conditions to simulate actual conditions experienced by a structural
member when it is subjected to both structural loading and chloride exposure. The
influences of material tensile strain capacity, crack width, and applied deformation level
on the effective chloride diffusion coefficient are investigated. The results show that
ECC is effective at slowing down the diffusion process of chloride ions under combined
mechanical and environmental loading, by virtue of its ability to achieve a self-controlled
tight crack width, even under large applied straining levels. This study verifies the

287

concept of improving repaired system durability with the inherent crack self-controlling
mechanism of ECC repair materials.

6.1

Steel Corrosion in Concrete Structures


The corrosion of steel in concrete is a major durability problem for reinforced and

prestressed concrete structures (Figure 6.1), demanding significant amounts of repair and
rehabilitation1'2'3'4. When concrete is cast around steel reinforcements, a passivation film
will form around the steel bars and protect them from corrosion initiation. However,
chloride ions, whether airborne or from deicing salt or seawater, are often present,
especially in coastal areas and snow regions where deicing salts are used. Corrosion
starts when chloride ions pass through the sound or cracked concrete cover, reach the
steel, and accumulate beyond a threshold concentration level. At this point the natural
protection layer surrounding the steel becomes depassivated, if oxygen and moisture are
present in the steel-concrete interface. The oxidation products generate expansive forces
within the concrete, eventually causing the concrete cover to spall5 (Figure 6.2).
Moreover, the cross section of the reinforcing bar or prestressing strand is now
diminished, leading to a reduction in the load carrying capacity of the concrete member.
It is widely recognized that corrosion of reinforcing steel has led to the premature
deterioration of many concrete structures in the United States, especially bridges and
marine structures, before their design life was exceeded. Although corrosion is not the
sole cause of all structural deficiencies, it is a significant contributor and has become a
major concern3'6.
Current approaches to preventing steel reinforcement corrosion in concrete

288

bridges have had mixed results . These approaches include the use of good construction
design and procedures, adequate concrete cover depth, corrosion-inhibiting admixtures,
and the adoption of low-permeability concrete. The reason for the limited success is due
to the fact that concrete is brittle and tends to crack under a variety of combined
mechanical and environmental loading conditions. In fact, it has been observed recently
that the new low permeability concrete (high-performance concrete), which is made from
partial substitution of Portland cement with silica fume or fly ash, has a more pronounced
tendency than conventional concrete to crack7'8. As a result, a faster, gravity-assisted
flow of salt-laden water channels through the cracks, rather than a slow diffusion of
chloride through uncracked concrete. Corrosion-inhibiting admixtures for concrete are
not useful once concrete cracks9, at which point the reinforcing steel has no further
recourse against corrosion. In this situation the use of a barrier system on the reinforcing
steel, such as an epoxy coating or other organic coating, becomes more critical in abating
corrosion. However, it is likely that there may never be an organic coating that can resist
the extreme combination of constant wetting, high temperature, and high humidity that
reinforcing steel is exposed to in marine environments, especially in the splash zone in
coastal regions . Epoxy-coated steel bars will have to be used in conjunction with sound,
crack-free concrete in bridge decks10, where the concrete is not constantly wet and other
exposure conditions are not as severe. The limitations of current approaches to reducing
steel corrosion in concrete calls for a new methodology that fundamentally controls the
crack width in concrete repairs, therefore preventing penetration of chloride ions through
the repair and substrate concrete, and protecting the reinforcement steel from corrosion.

289

6.2

Proposed Approach - ECC Microcracking Phenomenon


As highlighted previously, ECC is a class of materials that feature high tensile

ductility with moderate fiber volume fraction, typically 2% by volume11. Of special


interest is the capability of ECC to deform to high levels of tensile strain, commonly over
3%, while maintaining self-controlled tight crack widths.

These cracks have

experimentally been shown to be on the order of 60 \xm (0.002 in.). This high tensile
ductility and tight crack width is made possible in ECC through the use of steady state
crack models that provide quantitative links between micromechanical properties, such as
fiber bridging properties and matrix toughness, and composite mechanical behavior, such
as steady state cracking stress and maximum crack width.

To accommodate large

deformations, rather than forming a small number of cracks which widen with increasing
load as seen in concrete or tension-softening FRC, ECC forms numerous microcracks
which allow the material to undergo pseudo-strain hardening.

Once the initial

microcracks widen to the typical 60 ^m (0.002 in.) at roughly 1% composite straining,


additional microcracks form to accommodate further deformation while maintaining this
tight crack width in each crack. This phenomenon was shown previously in Figure 1.12.
The microcracking phenomenon observed in ECC materials (e.g. HES-ECC, ECC
M45) when subjected to restrained volume change and stress concentration from existing
cracks in substrate concrete, has been described in previous chapters.

It should be

emphasized that crack width in ECC material is a result of matrix-fiber interaction at


specific fiber volume fractions and is independent of the steel reinforcement ratio. This
inherent maximum crack width can be seen as a material property, similar to compressive
strength or elastic modulus, rather than a structural parameter, such as structural

290

dimensions or the reinforcing ratio in reinforced concrete. The independence of crack


width in ECC from structural dimensions and the ratio of cracking-control reinforcement
is a unique property different than other cementitious repair materials, which is crucial
for achieving robust resistance to chloride penetration in various application situations
without relying on cracking-control reinforcements.
Reinforced concrete design codes currently do not specify precise limits on
transport properties, such as maximum concrete permeability, or maximum crack width
under loading.

Permissible crack widths at the tensile face of reinforced concrete

structures for service loads under different environmental conditions, according to ACI
Committee 224 specifications12, are given in Table 6.1. The most stringent crack width
limit is placed on water-retaining structures, which are allowed maximum crack widths
no larger than 100 jam (0.004 in.). For exposure conditions of seawater, seawater spray,
wetting, and drying, crack widths smaller than 150 um (0.006 in.) are permitted. For
deicing chemical exposure, a maximum crack width of 180 [im (0.007 in.) is specified.
Within the current ACI code13, a maximum crack width is no longer explicitly specified,
although previous versions14 of the design code suggested 300 \im (0.012 in.) as an upper
limit. The AASHTO15 design code relies on the computation of a "Z" factor, for which
maximum crack width limits are set depending on the type of environmental exposure.
The limit imposed within the AASHTO code corresponds to a minimum reinforcement
spacing that results in a crack width limit of approximately 350 p,m (0.014 in.). As
discussed previously, at these relatively large crack widths the permeability of concrete
material is nearly five orders of magnitude greater than that of uncracked concrete.
While these crack width limits are important in reducing the transport of water and

291

corrosives (e.g. chlorides) into reinforced concrete, they are far larger than that necessary
for blocking corrosives from rapidly deteriorating the reinforcing steel, and thus the
entire reinforced concrete structure. Furthermore, these crack width limits are difficult to
accurately design for or control in real applications with traditional concrete materials.
Through the use of ECC material, which exhibits inherently tight crack widths less than
100 (xm (0.004 in.) even under large tensile deformations, there exists the possibility of
designing durable repairs with consistent resistance to chloride penetration.

6.3

Transport Properties of Cementitious Materials


The transport of chloride through cementitious materials is a complex

phenomenon potentially involving three main mechanisms - permeation, diffusion, and


absorption16. Depending on the conditions, transport of chloride through concrete can be
driven by one or a combination of these three mechanisms. The main driving force
behind permeation is the presence of a hydraulic pressure gradient. Permeation is an
important transport mechanism for underwater concrete structures, such as offshore
structures.

Absorption, driven by capillary pore suction, is the dominant transport

process when dry concrete is exposed to chloride solution.

Diffusion is the most

commonly studied transport process of chloride ions. When a reasonably moist concrete
(sufficient level of pore water exists) is exposed to a chloride solution, a chloride
concentration gradient is created between the concrete element surface and the pore
solution. Under such conditions, the diffusion process prevails. It is rare for a significant
hydraulic head to be maintained on highway structures, parking structures and buildings,
and the effect of absorption is typically limited to a dry and shallow concrete cover

292

region17'18'19. Exposure to deicing salts in snowy regions, or the concomitant presence of


humid climate conditions and high level of airborne and waterborne chlorides in coastal
regions, create an aggressive salt-laden environment for concrete structures. Therefore,
researchers tend to agree that in most cases, diffusion is the dominant chloride transport
mechanism for most concrete structures ' ' ' , and will be the focus of this study.
The development of reliable methods for predicting chloride transport into
concrete is important to determining the service life of reinforced concrete structures. A
number of studies have been carried out to understand the transport mechanism of
chloride ions

'

'

introduced25,26'27'28.

and numerous service life prediction models have been


In these studies, it was common to investigate the transport

properties of concrete in the uncracked state.

However, in reality most reinforced

concrete structures experience cracking in field conditions. This is especially true for
concrete repairs, which are susceptible to cracking due to restrained volume change or
r,

29,30,31,32,33,34,35

reflective cracking ' ' ' ' ' ' .


The formation of cracks increases the transport properties of concrete, so that
water, oxygen and chloride ions easily penetrate and reach the reinforcing steel, and
accelerate the initiation of steel reinforcement corrosion. Crack widths can range from
very small internal microcracks that occur upon application of modest amounts of stress,
to quite large cracks caused by undesirable interactions with the environment36. Wider
crack widths have been found to induce corrosion much faster than relatively smaller
cracks.
Aldea et al.37 conducted rapid chloride permeability tests on normal and high
strength concrete samples pre-cracked under a feedback-controlled splitting tensile test,

293

with crack width up to 400 \xm (0.016 in.). It was concluded that cracks less than 200 |im
(0.008 in.) had no effect on chloride conductivity, while cracks between 200 and 400 \xm
(0.008 and 0.016 in.) resulted in higher chloride conductivity. Tsukamoto 199038, Wang
et al. 199739, and Feng and Peng 200640 found that the permeability of cracked concrete
scales with the third power of crack width, and crack widths below 100 um (0.004 in.)
(50 um (0.002 in.) for gas permeability) tend to behave like sound concrete.
Lepech and Li41 found that cracked ECC exhibits nearly the same permeability as
sound concrete, even when strained in tension to several percent (Figure 6.3). Within this
study, both ECC and reinforced mortar specimens were stretched in tension to 1.5%
deformation, resulting in a variety of crack widths and number of cracks among the
various specimens. The permeability of these cracked materials was then determined
under hydraulic head. As seen in Figure 6.3, there is a dramatic rise in permeability with
increasing crack width. Further, when normalized by the number of cracks within the
specimen, the comparable permeability of cracked ECC with sound concrete becomes
more apparent.

These results suggest that due to the self-controlled crack widths,

replacing concrete with ECC can make significant strides in durability.


Absorption was examined for both cracked and uncracked ECC specimens by
Martinola et al.42. These studies found that absorption in cracked ECC specimens (1.41
kg/m2/h0'5) was much higher than in uncracked ECC or concrete mortar specimens (0.54
kg/m2/h0-5). This was due to the presence of microcracks within ECC specimens, which
promote liquid transport via capillary suction. The addition of an aqueous Silan-Siloxan
hydrophobic agent into the mixing water was proposed to repel capillary suction into the
thin ECC cracks. This agent was found to reduce the absorption coefficients of cracked

294

ECC by over 90%. Sahmaran and Li 2009 also found that microcracks induced by
mechanical loading increases the sorptivity value of ECC when no water repellent
admixture is present. However, capillary suction of ECC is not necessarily a concern
when compared with normal concrete, even when the ECC is loaded into the
microcracked strain-hardening stage. The use of a water-soluble silicone-based water
repellent admixture in the production of ECC further reduced the water sorptivity and
absorption properties of cracked ECC, to a level significantly lower than that of normal
uncracked concrete.
A comparison of diffusion coefficients for cracked and uncracked concrete shows
an increase in the diffusion coefficient for cracked concrete by one to two orders of
magnitude, with wider cracks resulting in higher values44. Chloride diffusion in concrete
pre-cracked under three point bending loading was also studied by Gowripalan et. al45. In
this study, prisms were preloaded, resulting in cracks up to 300 urn (0.012 in.) in width.
The experiment showed that the apparent chloride diffusion coefficient is larger in the
tensile zone than compression zone. Mangat and Gurusamy46 studied the influence of
cracks on chloride diffusion in steel fiber reinforced concrete. Cracks of widths ranging
from 70 to 1080 um (0.003 to 0.043 in.) were produced on prism specimens and exposed
to cycles of splash and tidal zone marine exposure. The authors concluded that crack
widths larger than 500 urn (0.020 in.) have a more pronounced influence on chloride
intrusion while crack widths less than 200 um (0.008 in.) appeared to have nearly no
effect on chloride intrusion. The influence of crack widths, up to 66 um (0.003 in.), on
chloride diffusivity was also studied by Tognazzi et. al47. The experimental results
showed that the chloride diffusion coefficient through cracked concrete is proportional to

295

the crack width.


Diffusion, the transport by an ion concentration gradient, has previously been
examined within HPFRCC materials by Miyazato et al. . Diffusion of both chloride ions
(Cf) and carbon dioxide (CO2) within cracked and uncracked HPFRCC and mortar
specimens was studied. These two chemical species are of particular concern due to their
association with depassivation of reinforcing steel, which is a precursor to active steel
rebar corrosion. These studies found that in uncracked specimens, chloride ions and
carbon dioxide molecules were transported between 1.5 to 3 times faster through an
uncracked HPFRCC matrix than through an uncracked concrete matrix.

In cracked

sections, however, these results were reversed and HPFRCC specimens showed a much
lower rate of diffusion, only 20% to 25% of that exhibited by cracked concrete deformed
to the same strain level.
The objective of the present research was to investigate the potential improvement
of ECC microcracking behavior on the resistance of chloride diffusion, when subjected to
combined mechanical and environmental conditions.

This was achieved through

preloading ECC beam specimens to different deformation levels, and then exposing them
to chloride ponding. The chloride ion diffusion resistance of cracked and uncracked
reinforced (with steel wire mesh) mortar specimens was measured in a control test series.
For ECC or reinforced mortar specimens, the crack number and crack width varied with
the applied deformation. The effects of crack number and crack width on the effective
diffusion coefficient were determined. The performance of cracked and uncracked ECC
and mortar specimens, in terms of chloride penetration profile and depth, and the
effective diffusion coefficient as a function of beam deformation level, was compared.

296

6.4

Experimental Program

6.4.1

Materials and Mixture Proportions


An ordinary version of ECC, ECC M45, is investigated in this study. The mix

proportions for ECC M45 mixtures are summarized in Table 6.2. Similar to typical fiber
reinforced cement composites (FRC), ECC M45 consists of Type I Portland cement, sand,
Class-F fly ash, water, fibers, and a superplasticizer (SP). However, unlike typical FRCs,
the component characteristics and proportions within ECC are carefully determined
through the use of micromechanical design tools49 to achieve the desired strain-hardening
response. To minimize the mortar matrix fracture toughness and crack tip energy JtiP, no
large aggregates were used, and the silica sand had average and maximum grain sizes of
110 um (0.004 in.) and 200 urn (0.008 in.), respectively. The poly-vinyl-alcohol (PVA)
fibers were the same type as used in HES-ECC. They were purposely manufactured with
a tensile strength, elastic modulus, and maximum elongation to match the
micromechanical design criteria for strain-hardening performance.

Additionally, the

surface of the PVA fiber was coated with a small amount (1.2% by weight) of a
proprietary hydrophobic oiling agent to tailor the interfacial properties between fiber and
matrix for strain-hardening performance50.

The mechanical and geometrical properties

of the PVA fibers used in this study are shown in Table 6.3.
In addition to ECC, the resistance of cracked and uncracked reinforced (with steel
wire mesh) mortar specimens to chloride ion penetration was also measured in a control
test series. The mix proportions of the mortar specimens are also shown in Table 6.2.
The components of this material are Type I Portland cement, water, and natural river sand.
The ultimate tensile strain capacity of ECC and the compressive strength test results of

297

ECC and mortar mixtures are also listed in Table 6.2. The compressive strength was
computed as an average of three standard 075x150 mm (03x6 in.) cylinder specimens.
To characterize the direct tensile behavior of the ECC mixtures, 152.4x76.2x12.7 mm
(6.0 in. x 3.0 in. x 0.5 in.) coupon specimens were used. Direct tensile tests were
conducted under displacement control at a loading rate of 0.005 mm/s (0.0002 in./s). The
typical tensile stress-strain curves of ECC mixtures are shown in Figure 6.4.

6.4.2

Specimen Preparation and Testing


From each mixture, 075x150 mm (03x6 in.) cylinders were prepared for the

determination of chloride penetration depth (immersion test), and 355.6 mm x 50.8 mm x


76.2 mm (14 in. x 2 in. x 3 in.) prisms were prepared for the determination of chloride ion
profiles and the diffusion coefficient (salt ponding test). All specimens were demolded at
the age of 24 hours, and moisture cured in plastic bags at 955% RH, 232 C for 7 days.
The specimens were then air cured at 505% RH, 232 C until the age of 28 days for
testing.

Immersion Test
At the end of 28 days, all surfaces of the cylindrical specimens except for their
bottom sides were sealed with a silicon coating so that chloride penetration could only
occur in one direction. Afterwards, they were left in the laboratory for an additional day.
At the age of 29 days, the cylinders were stored under continuous exposure to 3% NaCl
solution at room temperature, as shown in Figure 6.5 (a).

The NaCl solution was

replenished every month to maintain uniform concentration. After 30, 60 and 90 day
immersion periods, one cylinder was taken out and split open before being sprayed with

298

0.1-N silver nitrate (AgN03) solution (Figure 6.5 (b)). In locations where chloride ions
penetrated to at least approximately 0.15% by weight of cement51, silver chloride will
precipitate and cause that portion of the sample to turn white, while the silver nitrate in
the non-chloride penetrated zone turns brown.

Salt Ponding Test


Chloride ion profiles and the diffusion coefficient of ECC were evaluated in
accordance with AASHTO T259-8052. In addition to ECC, the same transport properties
of mortar specimens were also determined as a control test series. Each mortar prism was
reinforced with three levels of steel mesh to achieve cracks of varying widths. At the age
of 28 days, the prism surfaces were abraded using a steel brash. After abrasion, the
prisms were preloaded using a four point bending test (Figure 6.6) to obtain different
crack widths. The ponding test was carried out with the preloaded specimens in the
unloaded state (Figure 6.7). A small amount of crack closure occurred upon unloading.
For consistency throughout the test program, all crack width measurements were
conducted in the unloaded stage. The widths of the cracks were measured on the surface
of the specimens by a portable microscope. The average width of the resulting cracks
was obtained through measurement of crack widths at five points.
Table 6.4 shows the preloaded beam deformation (BD) value, along with the
corresponding average crack widths (CW), depths, and number of cracks for the prism
specimens. Two virgin prisms from each mixture were tested without preloading for
control purposes. Note that preloading beam deformation of the mortar prisms was
limited to 0.83 mm (0.033 in.) due to the large crack width (~ 400 urn [0.016 in.]) and
crack depth 70 mm (2.756 in.) generated in these specimens. In the ECC specimens, the

299

crack width remained at about 50 um (0.002 in.), even after beam deformation of 2 mm
(0.079 in.). Accurate measurement of crack length in the ECC specimens was not
feasible because of the tight crack width. As seen from Table 6.4, when the deformation
applied to the prism specimens increased, the number of cracks in ECC clearly increased
but crack width did not grow beyond 50 um (0.002 in.). Micromechanically designed
ECC changes the cracking behavior from one crack with large width to multiple smaller
cracks (Figure 6.8).
After load application, a Plexiglas embankment was built around the immediate
perimeter of the prism specimens, such that a chloride solution could be held against the
exposed side surfaces of the prism. At 29 days of age, a solution of 3% NaCl by weight
was ponded on the cracked surface of the prisms. In order to retard the evaporation of
solution, aluminum plates were used to cover the upper surface of the specimens. After
30 days of ponding, the salt solution was removed from the prism surface and samples
were taken from each specimen to measure chloride concentration as a function of depth.
In the case of uncracked ECC and mortar specimens, a ponding test was also conducted
after 90 days of NaCl solution exposure in accordance with AASHTO T259-80.
In the mortar beams, chloride was found to concentrate near the pre-crack and
penetrate deeply - to depths of approximately 40-70 mm (1.575-2.756 in.) depending on
the crack width and crack depth of the specimen. Mortar samples were taken from the
cracked zone for chloride analysis at various depths by using a 15 mm (0.591 in.)
diameter rotary drill.

In the ECC beams, chloride penetration occurred at multiple

locations corresponding to where the multiple cracks were formed during preloading.
However, the penetration depth was much shallower - between 0 to 40 mm (0 to 1.575

300

in.) depending on the level of imposed deformation.

Powder samples for chloride

analysis of cracked ECC specimens were taken where the multiple cracks were formed
during preloading. Total chloride (acid-soluble) content by weight of material at each
sampling point was examined according to AASHTO T 260-9753.
Statistical and curve-fitting software was then used to examine the chloride
profiles. Equation 6.1, Crank's solution to Fick's second law54, was fitted to the data.
Regression analysis yielded the values of the effective diffusion coefficient (De) and
surface chloride concentration (Cs) for each specimen.
C

=C

\-erf\

(6.1)

2V*V/

where
C(x,t) = chloride concentration at time t at depth x
Cs = surface chloride concentration
De = effective chloride diffusion coefficient
t = exposure time
erf( ) = error function
This equation would perfectly describe the diffusion process of chloride if no
other transport mechanisms (i.e., absorption and permeation) were present. However,
this is not the case in real field conditions. In this experimental study, the coefficient of
diffusion, found by regression analysis of chloride profiles using Equation 6.1, is referred
to as the effective diffusion coefficient (De), and includes the combined transport
mechanisms. De forms a reasonable basis for comparing the diffusion properties of
cracked and uncracked ECC and mortar specimens.

301

6.5

Experimental Results and Discussion


The results of immersion testing on the cylinders in the 3% NaCl solution for up

to 90 days are shown in Figure 6.9. In this test, cylinder specimens were sealed with a
silicon coating on all but one side so that chloride ions were allowed to ingress only in
one direction.

The main transport mechanism in this test was absorption, initially.

However, after the specimens became saturated, diffusion became the dominant transport
mechanism at later ages. For ECC mixtures with a low water/cementitious material ratio
and high volume of fly ash, chloride ion penetration depths were always lower than that
of the mortar mixture. The use of fly ash generally resulted in a denser matrix, by
reducing the pore sizes and microcracking in the transition zone55.
Figure 6.10 presents the chloride concentration profiles of uncracked mortar and
uncracked ECC from the ponding test. Prism specimens were examined after 30 and 90
days of exposure, and the effective diffusion coefficients (De) and coefficients of
2

determination (R ) were found using regression analysis. The R value represents how
well the estimated regression equation fits the data sample, with a value of one signifying
perfect agreement. In this figure, it can be seen that there is no significant difference
between the chloride concentration profile of the virgin mortar and ECC specimens. The
effective diffusion coefficients of the virgin mortar and ECC specimens exposed for 30
days are higher than for those exposed 90 days, likely due to the continuing hydration of
cement and fly ash, which is beneficial in reducing pore size and increasing matrix
density46.

In addition, the effective chloride diffusion coefficient of virgin ECC is

slightly lower than that of virgin mortar for 30 and 90 days, probably due to the higher
amount of cementitious materials, lower water/cementitious materials ratio, and high
volume fly ash content, which leads to a denser microstructure in ECC.
302

A lower

effective chloride diffusion coefficient value implies increased service life, according to
Figure 6.2, although no service life determination was made in this study to quantify this
effect.
The chloride concentration profiles of cracked mortar and cracked ECC at
different depths, exposed to NaCl solution for 30 days, are shown in Figure 6.11. The
chloride concentration profiles of virgin (BD = 0 mm) ECC and mortar specimens cured
in the same environment and of the same age are also included in this figure. As depth
from the surface increased, the chloride content in the cracked zone decreased. Figure
6.11 shows that the presence of cracks significantly modifies the chloride transport
properties of reinforced mortar. In all cases the chloride concentration increases as the
crack width increases, but this trend is especially pronounced in mortar specimens with
crack widths greater than 150 um (0.006 in.). These conclusions are consistent with
previous research findings46. The multiple-cracking in ECC due to pre-loading alters the
chloride transport properties as a function of beam deformation.

It was found that

chloride ingress increased as the number of cracks in the ECC specimens increased,
although the increase was fairly insignificant when compared to mortar specimens at the
same deformation levels. Thus, it can be concluded that an ECC specimen with multiple
microcracks exhibited greatly increased chloride penetration resistance.
The effective chloride diffusion coefficients of prisms, calculated by using the
chloride ion profiles reported above, together with Fick's second law, are summarized in
Table 6.5. As expected, this table clearly demonstrates the increase in the effective
diffusion coefficient with beam deformation and crack width increase in the mortar
specimens. It was also found that the effective diffusion coefficient increased as the
number of cracks in ECC specimens increased.
303

Cracks within the ECC specimens

opened to a maximum width of approximately 50 urn (0.002 in.). As designed, the ECC
specimens revealed a significantly smaller effective diffusion coefficient compared to
mortar specimens preloaded to the same deformation level, again reflecting the reduction
in chloride ion transport rate as a result of tight crack width control in ECC.
Figure 6.12 shows the relationship between the effective diffusion coefficient of
chloride ions and crack width in mortar specimens exposed to NaCl solution for 30 days.
Salt solution may fill the cracks, which can lead to diffusion from the crack plane45. This
is evident from the high chloride concentrations measured along the depth of crack. The
effective diffusion coefficient generally increased as crack width increased, and became
almost constant when the crack width fell below 150 urn (0.006 in.). It appears that the
effective diffusion coefficient of mortar has the form of a power function with respect to
crack width. The relationship between crack width and chloride diffusivity has also been
examined by other researchers. Despite the numerous tests and experimental methods
used, a consensus on the relationship between crack width and coefficient of diffusion
has not been reached. Some authors have found an increase in diffusivity within cracked
concrete up to ten times higher than in uncracked concrete44,47. Others have found no
effect or even a decrease in the diffusion coefficient56. The previous studies measured
chloride ingress into cracked concrete with roughly constant crack width, but cracks
generated in beams under a bending load will vary in width along the length of the beam.
The variety of experimental methods used to create these cracks, measure chloride
transport properties, and measure the chloride content, as well as the reactivity of the
chemical elements within the solid body, are the main reasons for the diversity of results
obtained from these different studies57.
In the case of mortar beams reinforced with three layers of steel mesh, pitting
304

corrosion was also observed in the first steel mesh layer of specimens, having crack
widths of 400 um (0.016 in.) after 30 days of NaCl solution exposure. Cracks can reduce
the service life of reinforced concrete structures by accelerating the initiation of
corrosion, especially when the crack width is larger than 150 \xm (0.006 in.), which
significantly increases the effective diffusion coefficient.
The crack width in the ECC mix did not grow beyond 50 urn (0.002 in.), even
under high levels of beam deformation and 1% tensile straining. However, the number of
cracks on the tensile surface did increase.

For this reason, the effective diffusion

coefficient of ECC specimens at different deflection levels, exposed to NaCl solution for
30 days, are plotted against the number of cracks in Figure 6.13. Corresponding values
for virgin ECC specimens, also cured in the same environment for 30 days, are included
in the figure.

The effective diffusion coefficient of ECC increases linearly with the

increase in crack number. This finding agrees with Konin et al., who studied chloride
penetration of reinforced ordinary and high strength concrete in the microcracked state58.
Figure 6.14 shows the relationship between the effective diffusion coefficient of
chloride ions and beam deformation level for mortar and ECC specimens.

Despite

imposition of more than twice the preload deformation, and a higher crack density, the
ECC specimens have an effective diffusion coefficient considerably lower than that of the
reinforced mortar because of tight crack width control. At higher deformation levels, the
effective diffusion coefficient of mortar increased exponentially with beam deformation.
The effective diffusion coefficient of ECC, however, increased linearly with the imposed
deformation value, because the number of microcracks on the tensile surface of ECC is
proportional to the imposed beam deformation. The total chloride concentration profiles,
perpendicular to the crack path, indicate no significant chloride penetration in the ECC
305

specimens, even at large imposed deformation (2 mm [0.079 in.]).


The relatively low diffusion coefficient of cracked ECC specimens can be
explained by the combined effects of the self-controlled tight crack width as well as selfhealing of the microcracks. The self-healing of cracks becomes prominent when the
crack width is small.

Based on experimental results, Evardsen59 and Reinhardt and

Jooss60 proposed that crack widths below 0.1 mm (0.004 in.) can be closed by a selfhealing process. In the case of precracked ECC specimens exposed to salt solution, a
distinct white deposit was visible over the crack surface after one month of exposure
(Figure 6.15).

These deposits were most probably caused by efflorescence due to

leaching of calcium hydroxide (CH) into cracks36 and also to the presence of NaCl ions in
the solution. The white deposits at the crack surfaces effectively blocked the flow path
due to the smaller crack widths present in ECC. An environmental scanning electron
microscope (ESEM) observation of the fractured surface of ECC across a healed crack is
shown in Figure 6.16. This ESEM observation shows that most of the products seen in
the cracks were newly formed C-S-H gels. Calcium hydroxide (CH) and deposited salts
were also observed in the crack path. These results indicate that microcracks of ECC
exposed to NaCl solution for 30 days healed completely by the end of the exposure
period. This can be attributed primarily to the large fly ash content and relatively low
water-to-binder ratio in the ECC mixture. The continued pozzolanic activity of fly ash
caused self-healing of the crack, which in turn reduced the ingress of the chloride ions.
The cracking and healing phenomenon of ECC under combined chloride environment
and mechanical loading will be investigated in the next chapter (Chapter 7).

306

6.6

Conclusions
Significant improvements in resistance to chloride penetration can be achieved

through the built-in microcracking mechanism of ECC materials. When subjected to


increasing deformation levels, which were achieved through the four-point bending test
in this study, ECC exhibited multiple microcracking behavior with increased crack
numbers but self-controlled crack width under 50 um (0.002 in.). This phenomenon is
fundamentally different from traditional concrete repair materials, in which localized
cracks form and the crack width enlarges with increasing strain level. Reinforcements
can limit the width of such localized cracks to some degree, but will not alter the
dependency of crack width on strain level and structural dimensions. This was partially
illustrated in this study through testing on reinforced mortar specimens, which represents
a large category of repair materials. After ECC beams and reinforced mortar beams were
preloaded to various deformation levels, and then exposed to 3% NaCl ponding for 90
days, the effective diffusion coefficient of ECC was measured to be significantly lower
than that of reinforced mortar. Even at a twice-higher preloaded deformation level, the
effective diffusion coefficient of the investigated ECC beam was around % that of the
investigated reinforced mortar beam.
The significantly lower effective diffusion coefficient of ECC compared with
reinforced mortar is attributed to the difference in cracking pattern.

The effective

diffusion coefficient of ECC was found to be linearly proportional to the number of


cracks, whereas the effective diffusion coefficient of reinforced mortar was proportional
to the square of the crack width.

Therefore, crack width in mortar had a more

pronounced effect on chloride transport than did the number of cracks in ECC, leading to

307

a much higher effective diffusion coefficient at the same deformation level.


Additionally, through the formation of small microcracks less than 100 (xm (0.004
in.), a significant amount of self-healing was observed within the microcracked ECC
when subjected to NaCl solution.

This self-healing phenomenon aides in further

reducing the effective diffusion coefficient of microcracked ECC.


In the uncracked beam specimens, the effective chloride diffusion coefficient of
ECC was lower than that of reinforced mortar for 30 and 90 days. This is explained by
the denser microstructure of ECC due to a higher amount of cementitious materials,
lower water/cementitious materials ratio, and high volume fly ash content.

The

immersion test results further proved this by showing that the chloride penetration depth
in uncracked ECC was 2.0 mm (0.79 in.), 3.3 mm (0.13 in.) and 4.1 mm (0.16 in.)
shallower than that of reinforced mortar after 30-day, 60-day and 90-day immersion
periods, respectively.
This study validated the concept that the microcracking mechanism and selfcontrolled crack width in ECC materials are crucial for slowing the transport of chloride
ions under combined mechanical loading and chloride exposure, and providing protection
for steel reinforcements against corrosion. The advantage of ECC is expected to be
greater for large field-scale applications, where repair dimensions and the reinforcing
ratio can both affect crack width in traditional repair materials. Findings from this study
can be used for service life prediction and life cycle cost analysis of ECC repairs or
structures, as well as providing a rational basis for the durable design of ECC
infrastructure systems. The research is particularly relevant given the anticipation that
ECC, especially when applied as a repair material, will be used in the microcracked state

308

during its structural service life.

309

Table 6.1 - Permissible crack widths for reinforced concrete under service loads
according to ACI Committee 224.
Exposure condition

Permissible Crack Width


in.
mm

Dry air or protective


membrane

0.016

0.41

Humidity, moist air, soil

0.012

0.30

Deicing chemicals

0.007

0.18

Seawater and seawater


spray, wetting and drying

0.006

0.15

Water-retaining structures

0.004

0.10

Table 6.2 - Mixture proportions and properties of ECC and mortar.


Properties

Mortar

FA/C
#
W/CM
0.35
3
3
Water (W), kg/m (lb/yd )
215(362)
3
3
Cement (C), kg/m (lb/yd )
614(1035)
3
3
Fly ash (FA), kg/m (lb/yd )
3
3
Sand (S), kg/m (lb/yd )
1535 (2585)
3
3
Fiber (PVA), kg/m (lb/yd )
Superplasticizer (SP), kg/m3 (lb/yd3)
36.6 (5.3)
7-day compressive strength, MPa (ksi)
42.9 (6.2)
28-day compressive strength, MPa (ksi)
7-day tensile strain capacity (%)
28-day tensile strain capacity (%)
CM: Cementitious Materials (Cement + Fly ash)

310

ECC (M45)
1.2
0.27
331 (558)
570 (961)
684(1153)
455(767)
26 (44)
5.1 (8.6)
37.8 (5.5)
53.3 (7.7)
3.90
3.10

Table 6.3 - Mechanical properties of PVA fiber.


Nominal
Strength

Apparent
Strength

Diameter

Length

Young's
Modulus

MPa
(ksi)
1620
(235)

MPa
(ksi)
1092
(158)

um
(in.)
39
(0.002)

mm
(in.)
8
(0.3)

GPa
(ksi)
42.8
(6200)

Elongation

(%)

6.0

Table 6.4 - Crack widths, numbers and depths of preloaded ECC and mortar prisms.

Mix ID

Mortar

ECC
(M45)

Beam
Deformation
mm (in.)

Average
Crack Widths
u.m (in.)

Crack
Depth mm
(in.)

0.50 (0.020)
0.70 (0.028)
0.80(0.031)
0.83 (0.033)
0.5 (0.020)
1.0(0.039)
1.5 (0.059)
2.0 (0.079)

-50 (0.002)
-150(0.006)
-300 (0.012)
-400 (0.016)
-0
-50 (0.002)
-50 (0.002)
-50 (0.002)

20 (0.787)
36(1.417)
55 (2.165)
70 (2.756)
-

Crack
Number
1
1
1
1
0
15
21
35

Table 6.5 - Chloride ponding test: effective diffusion coefficient.

Mix ID

Mortar

ECC
(M45)

Beam
Average Crack
Widths
Deformation
u.m (in.)
mm (in.)
0.00
0.50 (0.020)
0.70 (0.028)
0.80(0.031)
0.83 (0.033)
0.00
0.5 (0.020)
1.0(0.039)
1.5 (0.059)

-50 (0.002)
-150(0.006)
-300 (0.012)
-400 (0.016)
-

-0
-50 (0.002)
-50 (0.002)
311

Diffusion
Coefficient
m 2 /sxl0 12
(in. 2 /sxl0 8 )
10.58(1.64)
33.28(5.16)
35.54(5.51)
126.53 (19.61)
205.76(31.89)
6.75(1.05)
8.10(1.26)
27.99 (4.34)
37.50(5.81)

Bridge

^tW:;1'

Marine Structure

;*trXl'A

17.

>*

Seal Wall

^ _te

*Y
vif *

Figure 6.1 - Corrosion in Concrete Structures.

312

Cracking of concrete

Limit state

>

co2

o
1_
s_

o
O

Concrete
Steel

CJ:

Start of steel

i i 1

wmmmm

cuiiuaiun

^^iBSSB^^W

/
Corrosion
rate

y"' 1
r

Depassivation time

Propagation

Initiation

Time

Service life
Figure 6.2 - Corrosion process of reinforcing steel according to Tuutti (1982)20

1.00E+001.0E-01
1.00E-02-

1.0E-02

M 1.0QE-04o
g l.ODE-OfiO-c

1.Q0E-08-

CI

Observed

<4-<

4)
O

Normalized

1.00E-12-

300
400
Crack Width, w ((Am)

Figure 6.3 - Coefficient of permeability versus crack width for ECC and reinforced
mortar series deformed to 1.5% in uniaxial tension 41 .

313

Tensile Strain (%)

Figure 6.4 - Typical tensile stress-strain response of ECC mixture.

314

3%
NaCl
solution

J?

I
1

rfcf

Figure 6.5 - Immersion test, (a) Cylinders were stored under continuous exposure to 3%
NaCl solution, (b) One cylinder was taken out, split open, and sprayed with 0.1-N silver
nitrate (AgN03) solution, to measure the chloride penetration depth.

Head of Testing Machine

Load-applying and
Support: Block

yyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy/y
vu/rjgt.qm&#6

.191.6mm
(4 In)

191.6jn.ni
(4 in)
(4 in )
Spun Length 304.8mm (1.2 in)

Bed of Testing- Machine

Figure 6.6 - Four point bending test.

315

3 % NaCI Solution
13 mm
Sealed on
Sides

Concrete Sample

> 75 mm

50 % r.h.
atmosphere
AASHTO T259

., \ -ci

iw

Wl

Figure 6.7 - Salt ponding test on uncracked and cracked beam specimens with
dimensions of 355.6 mm x 50.8 mm x 76.2 mm (14 in. x 2 in. x 3 in.).

316

wm

mm

. - , -

'

>

- . * '

' * ' ' " . "

>

, - * *

* > - _ : -

'

"

. . .

. *

'^v'l-'-''.',

. ."
^ ' " ' . '

-..-.

-..'-.J?iV>1'---'."

Figure 6.8 - Typical crack pattern on positive moment surface of ECC beams at 2 mm
(0.079 in.) deformation.

12

0.45
g

4 0.40

.10

0.35

4 0 JO

15
9

0.25

0.20

fit

0.15

5
o 2

0.10
A Mortar

0.05

ECC (M45)
10

20

30

40

50

SO

70

80

90

* 0.00
100

Immersed time |days|


Figure 6.9 - Chloride penetration depth variation measured by immersion test.

317

Distance in in.
0.0

0,2

0,4

0.6

0.8

1.0

1.2

0.40
0.35
0.30

0 .s

1 S 0.25 ^
g * 0.20 o
9 ? 0.15
0.10 OJS 0.00 0

10

15
20
Distance in mm

25

30

Figure 6.10 - Chloride profiles of uncracked mortar and uncracked ECC prisms after 30
and 90 days in 3% NaCl solution.

318

35

Distance in in.
1.5

1.0

0.00

10

15

20

30
35
40
Distance in mm

25

45

50

55

60

65

70

(Mortar)

0.0

0.5

1.0

'

0.25

Distance in in.
1.5

2.0

2.5

'

-*-BD=0.00mm, CW=0|jm
-**-BD=0.50mm, CW=0pm
0.20

"*-BD=1.00mm, CW=50(jm
-*~BD=1.50mm, CW=50|jm

0 .
1 > 0.15
i : 13

-**-BD=2.00mm, CW=50|jm

c p

5 | 0.10

6:g,
0.05

0.00
10

20

30
40
Distance in mm

50

60

(ECC)
Figure 6.11 - Chloride profiles of mortar and ECC prisms in cracked zone at 30 days
exposure.

319

70

0.002

0.004

Crack Width (in.)


0.006 0.008 0.01 0.012

0.014

0.016

250
35
c
h 30 'o

200

JSL

y= 17.36 + 0.0012x"
R2 = 0.99

o
0

^^
W

150

o
o

If"*

T-

20

'w x

1 100

15 5 -,
>
10

>

50
IU

[ I I

50

100

150

200

'

250

300

350

400

450

Crack Width {pm\


Figure 6.12- Diffusion coefficient versus crack width for mortar deformed under
bending load.

60

y = 1.36x + 7.64
R2 = 0.99

50

<v
'o

7
40

c
o

ID
O

(V

O ~~
! > 30

Q S

5 .E

i<
0)

>

c io
O

20 \

>
o

<J

<D

10

10

15

20

25

30

35

40

Number of Cracks

Figure 6.13 - Diffusion coefficient versus number of cracks for ECC.

320

-T-

Deformation (in.)
.00

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

Mortar
Iny = 2.31 + 5.19xs
R2 * 0.99

ECC (M45)
y - 24.87x + 2.04
R2 = 0.95

I Z-.

0.0

'
I,

0.5

jf

1.0
Deformation (mm)

1.5

2.0

Figure 6.14 - Diffusion coefficient versus preloading deformation level for ECC and
mortar.

321

tt*f*r

i 5 l - ' 'All y "

(a) ECC cracks before salt ponding test

(b) ECC cracks after salt ponding test

Figure 6.15 - Self-healing products in ECC microcracks before and after salt
ponding test at 30 days exposure.

*V '

% *
'

'-4

"

-.

' : '
-

*''-:'.!
V

"*

"*.

Figure 6.16 - ESEM micrograph of rehydration products in a self-healed crack.

322

References:
1

Batis, G., Pantazopoulou, P., and Routoulas, A., "Corrosion Protection Investigation of
Reinforcement by Inorganic Coating in the Presence of Alkanolamine-Based Inhibitor,"
Cement and Concrete Composites, Vol. 25, 2003, pp. 371-377.
2

Elsener, B., "Macrocell Corrosion of Steel in Concrete - Implications for Corrosion


Monitoring," Cement & Concrete Composites, Vol. 24, 2002, pp. 65-72.
3

The State of the Nation's Highway Bridges: Highway Bridge Replacement and
Rehabilitation Program and National Bridge Inventory, Thirteenth Report to the United
States Congress, Federal Highway Administration, Washington, D. C , May 1997.
4

Lewis, D. A., "Some Aspects of the Corrosion of Steel in Concrete," Proceeding of the
First International Congress on Metallic Corrosion, London, 1962, pp. 547-555.
5

Scheissl, P. "Corrosion of Steel in Concrete" RILEM Report.


London. 1988.

Chapman and Hall.

Hime, W. G., "The Corrosion of Steel - Random Thoughts and Wishful Thinking,"
Concrete International, Vol. 15, No. 10, 1993, pp. 54-57.

Mehta, P.K., "Durability-Critical Issues for the Future", Concrete International, Vol.
19, No. 6, 1997, pp. 27-33.

Xi, Y., Shing, B., Abu-Hejleh, N., Asiz, A., Suwito, A., Xie, Z., and Ababneh, A.,
"Assessment of the Cracking Problem in Newly Constructed Bridge Decks in Colorado,"
Colorado Department of Transportation Research Branch. Denver, Colorado. CDOTDTD-R-2003-3. March 2003.
9

Kondratova, I., and Bremner, T. W., "Field and Laboratory Performance of EpoxyCoated Reinforcement in Cracked and Uncracked Concrete," Presented at the 77th
Annual Meeting of the Transportation Research Board, Washington, D. C , 1998.
10

Whiting, D., Nagi, M., and Broomfield, J. P., "Evaluation of Sacrificial Anode for
Cathodic Protection of Reinforced Concrete Bridge Decks," Report No. FHWA-RD-95041, Federal Highway Administration, Washington, D. C , May, 1995.
11

Li, V.C., "On Engineered Cementitious Composites (ECC) - A Review of the Material
and its Applications," Journal of Advanced Concrete Technology, Vol. 1, No. 3, 2003, pp.
215-230.
12

ACI Committee 224R-01, "Control of Cracking in Concrete Structures," ACI Manual


of Concrete Practice, Part 3, American Concrete Institute, Farmington Hills, Michigan,
2001.

323

1^

American Concrete Institute "Building Code Requirements for Structural Concrete


(ACI 318-08) and Commentary (ACI 318R-08)," ACI Committee 318, Detroit, Michigan,
2008.
14

American Concrete Institute "Building Code Requirements for Structural Concrete


(ACI 318-95) and Commentary (ACI 318R-95)," ACI Committee 318, Detroit, Michigan,
1995.
15

American Association of State Highway and Transportation Officials "AASHTO


LFRD Bridge Design Specifications: Second Edition," AASHTO, Washington, D.C.,
USA, 1998, pp. 5-41.
16

Uji, K., Matsuoka, Y., and Maruya, T., "Formulation of an Equation for Surface
Chloride Content of Concrete due to Permeation of Chloride," Corrosion of
Reinforcement in Concrete, Society of Chemical Industry, London, 1990, pp. 258-267.
17

Heirman, G., Vandewalle, L., Boel, V., Audenaert, K., de Schutter, G., D'Hemricourt,
J., Desmet, B., and Vantomme, J., "Chloride Penetration and Carbonation in SelfCompacting Concrete," ConcreteLife'06 - International RILEM-JCISeminar on Concrete
Durability and Service Life Planning: Curing, Crack Control, Performance in Harsh
Environments, RILEM Publications SARL, 2006, pp. 13-23.
18

Ghanem, H. A., "Chloride Ion Transport in Bridge Deck Concrete under Different
Curing Durations," M.S.E. Thesis in Civil Engineering, Texas Tech University, 2008, p.
159.
19

Stanish, K. D., Hooton, R. D., and Thomas, M. D. A., "Testing the Chloride
Penetration Resistance of Concrete: A literature Review," FHWA Contract DTFH61-97R-00022 "Prediction of Chloride Penetration in Concrete", 2000.
20

Tuutti, K., "Service Life of Structures with regard to Corrosion of Embedded Steel,"
Performance of Concrete in Marine Environment, ACI SP-65, American Concrete
Institute, Detroit, pp. 223-236.
21

Browne, R. D., and Geoghegan, M. P., "Mechanism of Corrosion of Steel in Concrete


in Relation to Design, Inspection and Repair of Offshore and Coastal Structures,"
Performance of Concrete in Marine Environment, ACI SP-65, American Concrete
Institute, pp. 169-204.
22

Bruenfield, N., "Chloride in Concrete," Concrete Repairs, Paladian, Vol. 2, 1986, pp.
115-130.

23

Hong, K., and Hooton, R. D., "Effects of Cyclic Chloride Exposure on Penetration of
Concrete Cover." Journal of Cement and Concrete Research, Vol. 29, No. 9, Sept. 1999,
pp. 1379-1386.

324

Ababneh, A., Benboudjema, F., and Xi, Y., "Chloride Penetration in Nonsaturated
Concrete," Journal of Materials in Civil Engineering, Vol. 15, No. 2, Mar.-Apr. 2003, pp.
183-191.
25

Hooton, R.D., and McGrath, P.F., "Issues Related to Recent Developments in Service
Life Specifications for Concrete Structures," Chloride Penetration into Concrete,
International RILEM Workshop, St. Remy - Les Chevreuse, 1995, p. 10.
26

Clifton, J., "Predicting the Service Life of Concrete," ACI Materials Journal, Vol. 90,
No. 6, Nov.-Dec. 1993, pp. 611-617.
27

Liang, M. T., Wang, K. L., and Liang, C. H., "Service Life Prediction of Reinforced
Concrete Structures," Journal of Cement and Concrete Research, Vol. 29, No. 9, Sept.
1999, pp. 1411-1418.

28

ACI Committee 365.1R-00, "Service-Life Prediction-State-of-the-Art


American Concrete Institute, Farmington Hills, Michigan, 2000.

Report,"

29

Li, M. and Li, V. C , "Influence of Material Ductility on the Performance of Concrete


Repair", Accepted, ACI Materials Journal, Feb. 2009.
30

Czarnecki, L., Garbacz, A., Lukowski, P., and Clifton, J. R., "Polymer Composites for
Repairing Portland Cement Concrete: Compatibility Project," Technical Report NISTIR
6394, National Institute of Standards and Technology, 1999.
31

Emmons, P. H., Vaysburd, A. M., Poston, R. W., and McDonald, J. E., "Performance
Criteria for Concrete Repair Materials, Phase II, Field Studies," Technical Report REMRCS-60, U. S. Army Waterways Experiment Station, Vicksburg, MS, September 1998.
32

Pigeon, M. and Bissonnette, B., "Bonded Concrete Repairs: Tensile Creep and
Cracking Potential," Concrete International, Vol. 21, No. 11, November 1999, pp. 31-35.

33

Vayburd, A. M., "Research Needs for Establishing Material Properties to Minimize


Cracking in Concrete Repairs, Summary of a Workshop," ICRI Publication No. Y320001,
1996.

34

Vaysburd, A. M., Emmons, P. H., McDonald, J. E., Poston, R. W., and Kesner, K. E.,
"Performance Criteria for Concrete Repair Materials, Phase II Summary Report,"
Technical Report REMR-CS-62, U. S. Army Engineers Waterways Experiment Station,
Vicksburg, MS, March 1999.
35

Li, V. C , Li, M., and Lepech, M., "High Performance Material for Rapid Durable
Repair of Bridges and Structures" MDOT Report RC-1484, Dec. 2006.
36

Mindess, S., and Young, F., Concrete, Prentice Hall Inc., Englewood Cliffs, New
Jersey, 1981.

325

51

Aldea, C. M., Shah, S. P., and Karr, A., "Effect of Cracking on Water and Chloride
Permeability of Concrete," Journal of Materials in Civil Engineering, Vol. 11, No. 3,
Aug. 1999, pp. 181-187.
TO

Tsukamoto, M., "Tightness of Fiber Concrete," Darmstadt Concrete, 5, 1990, pp. 215225.
39

21. Wang, K., Jansen, D, Shah, S., Karr, A., "Permeability Study of Cracked Concrete,"
Cement and Concrete Research, Vol. 27, No. 3, 1997, pp. 381-393.

40

Feng, N.Q., and Peng, G. F., "Effect of Mineral Admixtures on Durability of Concrete
Structure Subjected to Alkaline Saline Corrosions," Key Engineering Materials, Vol.
302-303, 2006, pp. 68-72.

41

Lepech, M. D., and Li, V. C , "Water Permeability of Engineered Cementitious


Composites," Cement and Concrete Composites, Accepted July 23 2009.

42

Martinola, G., Baeuml, M. F., and Wittmann, F. H., "Modified ECC by Means of
Internal Impregnation," Journal of Advanced Concrete Technology, Vol. 2, No. 2, 2004,
pp. 207-212.
43

Sahmaran, M., and Li, V. C , "Influence of Microcracking on water absorption and


sorptivity of ECC," RILEMJ. of Materials and Structures, Vol. 42, 2009, pp. 593-603.

44

Raharinaivo, A., Brevet, P., Grimaldi, G., and Pannier, G., "Relationship between
Concrete Deterioration and Reinforcing-Steel Corrosion," Durability of Building
Materials, Vol. 4, 1986, pp. 97-112.
45

Gowripalan, N., Sirivivatnanon, V., and Lim, C. C , "Chloride Diffusivity of Concrete


Cracked in Flexure," Journal of Cement and Concrete Research, Vol. 30, No. 5, May
2000, pp. 725-730.
46

Mangat, P. S., and Gurusamy, K., "Chloride Diffusion in Steel Fibre Reinforced
Marine Concrete," Journal of Cement and Concrete Research, Vol. 17, No. 3, May 1987,
pp. 385-396.
47

Tognazzi, C , Ollivier, J.-P., Carcasses, M., and Torrenti, J.-M., "Couplage FissurationDe'gradation Chimique des Mate'riaux Cimentaires: Premiers Re'sultats sur les
Proprie'te's de Transfert," Ouvrages, ge'ornate 'riaux et interactions, Ch. Petit, G.
Pijaudier-Cabot, and J. M. Reynouard, eds., Herme's, Paris, 1998, pp. 69-84.
48

Miyazato, S. and Hiraishi, Y. "Transport Properties and Steel Corrosion in Ductile


Fiber Reinforced Cement Composites," Proceedings of the Eleventh International
Conference on Fracture, Turin, Italy, March 20-25, 2005.

326

Li, V. C , Wang, S, and Wu, C , "Tensile Strain-Hardening Behavior of Polyvinyl


Alcohol Engineered Cementitious Composite (PVA-ECC)," ACI Materials Journal, Vol.
98, No. 6, 2001, pp. 483-492.
50

Li, V. C , Wu, C , Wang, S., Ogawa, A., and Saito, T., "Interface Tailoring for Strainhardening PVA-ECC," ACI Materials Journal, Vol. 99, No. 5, Sept.-Oct. 2002, pp. 463472.
51

Otsuki, N., Nagataki, S., Nakashita, K., "Evaluation of AgN03 Solution Spray Method
for Measurement of Chloride Penetration into Hardened Cementitious Matrix Materials,"
Construction and Building Material, Vol. 7, No. 4, 1993, pp. 195-201.
52

Standard Method of Test for Resistance of Concrete to Chloride Ion Penetration,


AASHTO T 259-80, AASHTO, USA, 2001.
53

Standard Method of Test for Sampling and Testing for Chloride Ion in Concrete and
Concrete Raw Materials, AASHTO T 260-97, AASHTO, USA, 2001.
54

Crank, J., The Mathematics of Diffusion (2nd Edition), Oxford University Press,
London, 1975.
55

Mehta, P. K., and Monteiro, P. J. M., Concrete - Microstructure, Properties, and


Materials (Indian Edition), Indian Concrete Institute, Chennai, 1997.
56

Rodriguez, O.G., and Hooton, R.D., "Influence of Cracks on Chloride Ingress into
Concrete," ACI Materials Journal, Vol. 100, No. 2, Mar.-Apr. 2003, pp. 120-126.
57

Gerard, B., Reinhardt, H.W., and Breysse, D., "Measured Transport in Cracked
Concrete," Penetration and Permeability of Concrete, H. W. Reinhardt, ed., RILEM
Report 16, 1997, pp. 123-153.
58

Konin, A., Francois, R., and Arliguie, G., "Penetration of Chlorides in Relation to the
Microcracking State into Reinforced Ordinary and High Strength Concrete," Materials
and Structures, Vol. 31, No. 209, June 1998, pp. 310-316.
59

Evardsen, C. "Water Permeability and Autogenous Healing of Cracks in Concrete,"


ACI Materials Journal, Vol. 96, No. 4, Jul.-Aug. 1999, pp. 448-454.
60

Reinhardt, H. W., and Jooss, M., "Permeability and Self-Healing of Cracked Concrete
as a Function of Temperature and Crack Width," Journal of Cement and Concrete
Research, Vol. 33, No. 7, Mar. 2003, pp. 981-985.

327

CHAPTER 7

Effect of Cracking and Healing on the Durability of ECC under Combined


Aggressive Chloride Environment and Mechanical Loading

ECC's large tensile ductility and self-controlled crack width below 100 [im are
essential for achieving durable repair performance.

This chapter investigated the

durability of ECC in terms of maintaining these tensile properties under combined


mechanical loading conditions and aggressive chloride environment. Specifically, ECC
specimens were preloaded to 0.5, 1.0, and 1.5% tensile strain levels, and then immersed
in 3% chloride solution for 30 days, 60 days, and 90 days, respectively. These specimens,
together with control virgin (without preloading) specimens, were reloaded to evaluate
the potential change in tensile properties including tensile strength, material stiffness, and
tensile strain capacity. This study revealed that the reloaded specimens retained multiple
microcracking behavior and tensile strain capacity of more than 3%, while the average
crack width increased from 40um to lOOum and tensile strength was reduced by 10%.
Furthermore, self-healing of the existing microcracks in ECC under chloride exposure is

328

evident in terms of recovery of initial material stiffness and tensile strain capacity. These
results indicated that even under severe marine environment conditions, ECC remains
durable and can provide reliable tensile ductility and self-controlled crack width to
prevent common cracking failure as well as chloride penetration.

7.1

Introduction
Concrete cracking is a result of the combined effects of mechanical loading

conditions and environmental exposure l .

Cracking can occur at different stages

throughout the life of a concrete structure. The presence of cracks not only causes loss of
load carrying capacity and stiffness of structural members, but also dramatically lowers
the resistance to penetration by aggressive agents that results in further deterioration.
Among all possible forms of environmental exposure for concrete structures, an
aggressive chloride environment (e.g. marine environment) is one of the most severe and
aggressive on earth2.

Serious concrete deterioration has been reported ' ' ' , and the

general cause was noted to be concrete cracking, which resulted in corrosion of


embedded reinforcing steel . Maximum allowable crack widths are thereby required in
various codes and specifications by technical committees for design of reinforced
concrete structures exposed to marine environments.

These requirements are

7 8 Q 10 I I 19

summarized in Figure 7.1 ' ' ' ' ' . The allowable maximum crack width ranges from
150 \xm to 300 urn, with the most stringent requirements specified by JSCE and ACI 224.
Such crack limits, however, are a challenge for structures or repairs using traditional
concrete materials when they are subjected to mechanical loading conditions while also
being attacked in aggressive environments. Despite extensive research, reliable crack

329

width control using steel reinforcements in concrete structures remains difficult to realize
inpractice13'14'15'16.
In previous chapters of this dissertation, it has been proposed and demonstrated
that the inherent large tensile ductility (>3%) of ECC achieved through multiple
microcracking behavior, as well as its self-controlled crack width under 100 |j.m during
strain-hardening stage, are critical properties for achieving durability of concrete repairs.
Durability improvements include preventing cracking and interfacial delamination caused
by restrained volume change (Chapter 4), eliminating reflective cracking under stress
concentration and fatigue (Chapter 5), and enhancing resistance to chloride penetration,
even at large deformation levels (Chapter 6). Through suppression of these three most
common mechanisms during the typical deterioration process in concrete structures, the
ECC repair can effectively protect the underlying concrete from the hostile environment,
and restore/enhance its load carrying capacity. By these means, the structural service life
can be greatly extended with minimized future maintenance and repair frequencies.
While ECCs large tensile ductility and self-controlled tight crack width are
essential for ensuring repair durability, a question that could naturally be raised is
whether ECC can still maintain these properties under combined aggressive
environmental (e.g. chloride) exposure and mechanical loading. This concern is critical,
especially because for most repair applications ECC is expected to be in service during its
strain-hardening stage, as discussed in Chapters 4, 5, and 6. At this stage, the imposed
deformation due to restrained shrinkage, movement of underlying cracks, and structural
load would have activated multiple microcracking behavior with a certain amount of
microcracks. When exposed to an aggressive chloride environment, the presence or

330

potential penetration of chloride ions through these cracks might reach the fibers and the
fiber/matrix interface, leading to potential loss of mechanical properties of ECC. The
objective of the present study is to understand the long-term durability of ECC in both the
cracked and uncracked state, i.e. the influence of different strain levels, combined with
different time-lengths of chloride exposure, on the tensile ductility and cracking behavior
of ECC. Specifically, the tensile ductility and self-controlled steady-stage crack width of
ECC are investigated after a specified preloaded deformation level, and after exposure to
chloride for a specified time period.

7.2

Evidence of Self-Healing of ECC Microcracks and Recovery of Chloride

Transport Properties
Self-healing of cracked concrete, commonly known as autogenous healing, is an
often-studied phenomenon17'18. Self-healing is generally attributed to the hydration of
previously unhydrated cementitious material, calcite formation, expansion of concrete in
the crack flanks, crystallization, closing of cracks by solid matter in water, and closing of
cracks by fragments of concrete from the cracks themselves19. It is believed that, under
certain environmental conditions as well as the presence of adequate concentrations of
certain chemical species, the possibility and consistency of self-healing mechanisms
greatly depend on crack width20. Previous researchers have reported various ranges of
maximum crack widths for self-healing to occur in concrete under various environmental
exposure conditions. These maximum crack width criteria include: 5 to 10 (xm21, 53
00

0"X

0 A.

0^

Of\

OH

|j.m , 100 \xm , 150 |im , 200 |xm , 205 \xm , 300 jam . These current results, despite
their wide range, all imply that a tighter crack width is favorable for self-healing to occur

331

within the crack. However, as discussed earlier, requirements for such tight crack width
are difficult to achieve consistently in concrete. This explains why reliable self-healing
in most concrete structures has not been observed.
The self-healing phenomenon in ECC at the composite scale under various
exposure regimes, including water/air cycle, water/hot air cycle, 90%RH/air cycle, water
submersion, and air, have been reported by Lepech and Li (2005)28, Yang et al. (2005) ,
and Yang (2008)17. The chloride exposure regime, however, has not been previously
studied. Autogenous self-healing in preloaded ECC specimens, in terms of recovery of
chloride transport properties, was observed under chloride exposure in Chapter 6. In
that study, preloaded ECC beams under different bending deformation levels were
subjected to chloride (3% NaCl solution) ponding. As applied deformation increased, the
number of cracks in ECC beams increased with unchanged crack width (approximately
50 urn), in contrast with the increased crack width in the control reinforced mortar beams.
Having the same or higher magnitude of imposed overall deformation, the ECC beams
revealed an effective diffusion coefficient considerably lower than that of the reinforced
mortar beams, especially at higher deformation levels. The total chloride concentration
profiles perpendicular to the crack path indicated no significant chloride penetration up to
90 days of exposure, even at large imposed deformation (2 mm), for the ECC beams.
The significantly lower diffusion coefficient of cracked ECC beams was
attributed to the combined effects of the self-controlled tight crack width as well as the
self-healing of microcracks. Distinct white deposits were visible at the crack surfaces
after one month of exposure to chloride ponding (Figure 6.15), which effectively blocked
the flow path of penetration of chloride ions. It should be noted that a previous water

332

permeability test found no self-healing in cracked ECC specimens that were submersed in
water29. The self-healing phenomenon in cracked ECC under chloride exposure can be
attributed to the presence of NaCl ions, which promoted leaching of calcium hydroxide30
into cracks and sealed the cracks. An environmental scanning electron microscope
(ESEM) observation of the fractured surface of ECC across a healed crack (Figure 6.16)
showed that most of the products seen in the cracks were newly formed C-S-H gels.
Calcium hydroxide (CH) and deposited salts were also observed in the crack path. This
can be attributed primarily to the high fly ash content, unhydrated cement, and relatively
low water-to-binder ratio in the ECC mixture. The continued cement hydration and
pozzolanic activity of fly ash caused self-healing of the microcracks, which in turn
reduced the ingress of the chloride ions. These observations indicate that microcracks of
ECC healed completely after exposure for 30 days to NaCl solution.
The possibility of autogenous self-healing in preloaded ECC specimens exposed
to a chloride environment, in terms of recovery of tensile properties (tensile strain
capacity, Young's modulus and crack width), was not previously understood.

To

investigate this phenomenon in the present study, ECC coupon specimens were preloaded
under uniaxial tension to strain levels of 0.5%, 1.0%, and 1.5% to simulate in-service
loading conditions (e.g., vehicle load, prestressing load, restrained shrinkage, thermal
load, etc). After applying and removing the load, specimens were exposed to a 3%
chloride concentration solution for 1, 2, or 3 months and subsequently reloaded until
failure to measure residual tensile properties. The effect of cracking and autogenous
healing under combined mechanical loading and chloride exposure was assessed by

333

measuring the retained stiffness, ultimate tensile strength, and tensile strain capacity, and
crack width of ECC.

7.3

Experimental Program

7.3.1

Testing of Composite Properties


An ordinary version of ECC, ECC M45, as investigated in Chapter 6, is also

employed in this study. The mix proportions for ECC M45 mixtures are summarized in
Table 6.2. The mechanical and geometrical properties of the PVA fibers used in this
study are shown in Table 6.3.
The ECC mixture was prepared in a standard concrete mixer according to the
following sequence: the cement, fly ash and sand were first mixed for one minute; water
and HRWR were then slowly added and mixed for an additional two minutes until the
cementitious material obtained a homogeneous state; finally, the fibers were added and
mixed for two-three minutes until maximum uniformity of fiber dispersion was achieved.
This mixture was cast into 152.4x76.2x12.7 mm coupon specimens for the direct uniaxial
tensile test. The specimens were covered with plastic sheets and demolded after 24 hours.
Then they were first cured in plastic bags at 955% RH, 201 C (66-70 F) for 6 days,
and then left to cure in laboratory air under uncontrolled humidity 45 5% RH and
temperature 201 C (66-70 F) conditions until the age of 28 days for testing.
The magnitude of initial and recovered mechanical properties was measured
under uniaxial tensile loading. First, deliberate damage was introduced by tensioning the
coupon specimens to predetermined strain levels of 0.5, 1.0 and 1.5% followed by
unloading, before exposure to 3% NaCl solution. A servohydraulic testing system was

334

used in displacement control mode to conduct the tensile test. The loading rate used was
0.0025 mm/s to simulate a quasi-static loading condition. Aluminum plates were glued
both sides at the ends of coupon specimens to facilitate gripping. Two external linear
variable displacement transducers were attached to the specimen to measure the specimen
deformation. Typical stress-strain curves of the preloaded specimens are shown in Figure
7.2.
The precracked ECC specimens were then continuously immersed in a 3% NaCl
solution at room temperature, together with some uncracked specimens (as control) that
were not subjected to preloading, for 30, 60 or 90 days, respectively.

The sodium

chloride solution was replaced with a fresh solution every month. Subsequently, the
specimens were reloaded in direct tension and stress-strain curves were recorded. In the
case of uncracked specimens, the average strain capacity and ultimate strength of ECC
were averaged from four specimens. In the case of precracked specimens, the average
strain capacity and ultimate strength of ECC were calculated from a minimum of four and
maximum of six specimens.

The tensile properties were then compared with those

measured before damage and with those after damage but before potential self-healing
took place.

7.3.2

Measurement of MicroMechanical Parameters


To understand the potential effects of chloride exposure on ECC matrix and

fiber/matrix interfacial properties at the micromechanical scale, matrix fracture toughness


test and single fiber pullout tests were conducted. Based on the conditions for strainhardening and saturated multiple-cracking, high tensile strain capacity requires a high

335

Jb'/Jtip ratio. The matrix toughness Jtip and flaw size distribution are matrix properties,
while the complementary energy Jb' is mainly controlled by fiber and interface properties.
To measure the JtiP, matrix toughness tests were conducted on the ECC matrix
(without fibers) before and after exposure to 3% NaCl solution for 30, 60 and 90 days,
respectively. This test was similar to ASTM E399 "Standard Test Method for PlaneStrain Fracture Toughness of Metallic Materials". The ASTM E399 allows one to use
different geometry specimens, such as bending specimens and compact tension
specimens, to measure the Km value.

The fresh mix was cast into notched beam

specimens measuring 305 mm (12 in.) in length, 76 mm (3 in.) in height, and 38 mm (1.5
in.) in thickness. The ECC matrix specimens were then demolded in 24 hours, and cured
just as the composites tensile specimens, except that fibers were not added. Four different
curing conditions were employed for these beam specimens after being demolded in 24
hours: (a) 28-day curing in air, (b) 28-day curing in air + 30-day curing in 3% NaCl
solution, (c) 28-day curing in air + 60-day curing in 3% NaCl solution, and (d) 28-day
curing in air + 90-day curing in 3% NaCl solution. The matrix fracture toughness Km
was measured by the three point bending test, as shown in Figure 2.15. The span of
support is 254 mm (10 in.) and the notch depth to height ratio is 0.4. Three specimens
were tested for each test series.

Jtip was calculated from the measured Km through

Equations 2.1 and 2.2.


To calculate Jb' of the ECC mix with different curing conditions, single fiber
pullouts tests were conducted to measure three important interfacial parameters: chemical
bond strength Ga, frictional bond strength x0, and slip hardening coefficient (3. As shown
in Figure 2.16, single fiber pull-out tests were conducted on small-scale prismatic

336

specimens with dimensions of 10 mm x 5 mm x 0.5 mm (0.4 in. x0.2 in. x0.02 in.). A
single fiber was aligned and embedded into the center of an ECC mortar prism with an
embedment length of 0.5 mm (0.02 in.). Four different curing conditions were employed
for these prismatic specimens after being demolded in 24 hours: (a) 28-day curing in air,
(b) 28-day curing in air + 30-day curing in 3% NaCl solution, (c) 28-day curing in air +
60-day curing in 3% NaCl solution, and (d) 28-day curing in air + 90-day curing in 3%
NaCl solution.

Three specimens were tested for each test series.

The load versus

displacement curve was obtained through quasi-static testing and used to determine the
interfacial parameters. These interfacial parameters, along with fiber volume fraction,
length and diameter, were then used to calculate the fiber bridging law a(5). The
resulting complimentary energy Jb calculated from the a(5) curve combined with the
matrix fracture toughness JtiP obtained from the Km measurement were used as inputs to
evaluate composite material behavior (i.e. strain-hardening or tension-softening) and to
calculate the PSH (pseudo strain hardening, J t/Jtip) index as defined in Equation 2.4.

7.4

Experimental Results and Discussion


Table 7.1 summarizes the tensile strain capacity, tensile strength, crack width of

ECC with various preloaded strain values, and NaCl exposure conditions. Typical tensile
stress-strain curves obtained for specimens before and after exposure to the NaCl solution
are shown in Figure 7.3. For the uncracked specimens that were not preloaded, exposure
to NaCl appears to reduce the cracking strength and ultimate strength by approximately
10%. However, the tensile strain capacity does not appear to be affected after 30, 60, and
90 days of chloride exposure. Multiple microcracking behavior is retained after 30, 60,

337

and 90 days of chloride exposure, while the average crack width increased from 45 um
(w/o chloride exposure) to 100 urn (w/ chloride exposure).

Exposure to chloride

appeared to similarly affect the precracked specimens; applied tensile strain up to 1.5%
did not exacerbate the deterioration.
Figure 7.3 also shows the tensile properties of ECC specimens that had been
precracked to 0.5, 1.0 and 1.5% strain levels, then unloaded and reloaded 1 day after
precracking.

Thus, these specimens had no time to undergo the crack healing as

compared to those specimens exposed to NaCl solution for 30, 60 and 90 days.
Compared to the latter, the reloaded specimens without possible self-healing exhibited a
remarkable loss in initial stiffness, which is defined by the initial linear part on the stressstrain curve. This can be explained by the initial low load resistance of the existing
microcracks.

After initial microcracks formed in the precracked ECC specimen,

reloading the same specimen will simply re-open the already-formed microcracks until
the crack widths initially reached before unloading are reached again and the fiberbridges re-engage. At first, re-opening these initial microcracks without fiber-bridging
behavior resulted in low initial stiffness of the ECC material. Once fiber bridging was reengaged, however, the load capacity resumed, and further tensile straining of the intact
material could take place.
In contrast, a significant recovery of the initial material stiffness was found in the
precracked ECC specimens after they were exposure to chloride solution for 30, 60, and
90 days and then reloaded. This suggests that between the time of inducing precracking
and the time of reloading, healing of the microcracks took place in the ECC specimens
after exposure to NaCl solution. After crack healing, reloading then stressed the rehealed

338

products in the microcracks. Self-healing can be attributed primarily to the availability of


unhydrated cement due to the high cementitious material (cement and fly ash) content
and relatively low water to binder ratio within the ECC mixture. As a result of the
formation of microcracks due to mechanical loading, unhydrated cementitious particles
were easily exposed to the sodium chloride solution during the immersion period, which
led to further initiation of hydration processes. The newly formed products not only
physically sealed the microcracks, but also recovered the fiber-bridging mechanism by
re-forming the fiber/matrix interfacial bond at the microcracks, and thus restored the
composite tensile properties.
The ultimate tensile strength values are shown in Figure 7.4. Compared to control
specimens cured in laboratory air, the test results indicate that the specimens (precracked
and uncracked) immersed in sodium chloride solution show a 10% reduction in ultimate
tensile strength for all exposure ages; this should be attributed to the effects of calcium
hydroxide leaching on the fiber/matrix interfacial bond and fiber bridging capacity.
Water not saturated with calcium hydroxide (high-calcium hydrated lime) may affect test
results due to leaching of lime from the test specimens31. Presence of chloride ions,
especially, tends to increase the leaching of calcium and opening of the porosity of the
cementitious matrix32'33.
Figure 7.5 shows the average tensile strain capacity of ECC specimens stored in
NaCl solution. The tensile strain capacity reported for these specimens does not include
the residual strain from the precracking load. By neglecting this residual strain, the large
variability in material relaxation during unloading was avoided, and a conservative
estimate of ultimate strain capacity of the material was determined. The tensile strain

339

capacity of uncracked and precracked ECC specimens exposed to NaCl solution averaged
between 2.4% and 3.8%. This value is higher than or similar to that of air cured
specimens (averaging 2.5% to 2.8%).
The influence of chloride exposure on non-preloaded ECC tensile properties can
be explained by the influence of chloride exposure on the microstructure of ECC. Based
on the measurements of the micromechanical parameters, a reduction in matrix fracture
toughness (Figure 7.6) was found after the specimens were exposed to chloride solution
for 30, 60 and 90 days, compared to no exposure to chloride solution. The reduction in
ECC matrix fracture toughness, very probably due to the leaching of calcium hydroxide
in the cementitious matrix to the NaCl solution and the corresponding increase in matrix
porosity, explained the reduction in ECC composite first cracking strength. Additionally,
from Figure 7.7 and 7.8, despite the large error bar typical of this type of test, it was
observed that both the interfacial chemical bond Gd and frictional bond %Q tend to be
reduced after 30, 60 and 60 days of exposure to 3% NaCl solution. The reduction in the
fiber/matrix interfacial bonds can be due to the leaching of calcium hydroxide and
increased porosity (less dense) at the fiber/matrix interfacial transition zone, and potential
presence of chloride ions and moisture at the interface that might harm the interfacial
bonds. The reduction in interfacial bonds explained the increase in the crack width in
ECC composites from less than 45 \im before chloride exposure, to approximately 100
|j,m before 30, 60 or 90 days chloride exposure. Furthermore, the reduction in interfacial
bonds also led to a decrease in fiber bridging capacity, resulting in the reduction
(approximately 10%) in ECC composite ultimate tensile strength.

340

The calculated PSH index, J b/JtiP, is shown in Figure 7.9. Interestingly, it was
found that the PSH index increased from 2.3 (before exposure to 3% NaCl solution), to
2.8 (30-day exposure to 3% NaCl solution), 3.1 (60-day exposure to 3% NaCl solution),
and 2.7 (90-day exposure to 3% NaCl solution). This indicates that after different timelengths of severe chloride exposure, the PSH index increased rather than declined. This
clarified the underlying micromechanical mechanism of the maintenance of tensile strain
capacity of ECC composites after 30, 60, and 90 days of severe chloride (3% NaCl
solution) exposure. Such a mechanism does not exist in traditional concrete materials,
because the leaching of calcium hydroxide only impairs concrete mechanical properties
and promotes potential concrete deterioration32'33.

7.5

Conclusions
This study revealed that ECC maintained its unique tensile ductility and multiple

microcracking behavior under combined mechanical loading conditions (0.5%, 1.0%,


1.5%) tensile straining and reloading to failure) and aggressive chloride exposure (30-day,
60-day, and 90-day immersion in 3% NaCl solution). This indicated that even under
severe marine environment conditions, ECC remains durable and can provide reliable
tensile ductility to prevent common cracking failure due to restrained volume change or
stress concentration. Additionally, its self-controlled microcrack width during the strainhardening stage can offer reliable resistance to chloride penetration, even in a severe
marine environment. Although the average crack width of ECC increases from 45 um to
100 \xm, it is still below the maximum crack width limit for marine environments.

341

The innately tight crack width in ECC is favorable for the self-healing process to
take place. This was demonstrated in this study in terms of recovery of tensile behavior,
especially material initial stiffness. Specimens preloaded with up to 1.5% strain capacity
showed almost complete recovery of stiffness and tensile strain capacity when reloaded
in direct tensile tests, even after periods of 30, 60, and 90 days exposure to NaCl solution.
It was also found that both precracked and uncracked ECC specimens exposed to
NaCl solution lost approximately 10% of their first cracking strength and ultimate tensile
strength. Furthermore, average crack width during reloading increased from 45 urn
(exposed to air) to 100 urn (exposed to chloride solution). These phenomena, observed at
the composite scale, suggested possible changes in the fiber/matrix interfacial bond
properties at the microscale due to chloride exposure. It was suspected that presence of
chloride ions promoted the leaching of calcium hydroxide, and consequently increased
porosity in the cementitious matrix and the fiber/matrix interfacial transition zone.
Through investigation on ECC microparameters, reductions in matrix fracture toughness,
fiber/matrix interfacial chemical bond and frictional bond, were observed after exposure
to 3% NaCl solution for 30, 60, and 90 days, compared to those without the chloride
exposure. However, the PSH index, J t/JtiP, slightly increased. These changes in matrix
and fiber/matrix properties due to exposure to 3% NaCl solution explained the reduction
in first-cracking strength and ultimate tensile strength, increase in crack width during the
strain-hardening stage, and the maintained tensile strain-hardening behavior and tensile
strain capacity.

342

Table 7.1 - Tensile properties of ECC under different combinations of preloading


and chloride exposure conditions.
Environmental
Exposure
Condition

28 days air + 30
days air or NaCl

28 days air + 60
days air or NaCl

28 days air + 90
days air or NaCl

Preloaded
Strain

Tensile Strain
Capacity

(%)

(%)

0.0 (Air curing)


0.0 (3% NaCl)
0.5 (3% NaCl)
1.0 (3% NaCl)
1.5 (3% NaCl)
0.0 (Air curing)
0.0 (3% NaCl)
0.5 (3% NaCl)
1.0 (3% NaCl)
1.5 (3% NaCl)
0.0 (Air curing)
0.0 (3% NaCl)
0.5 (3% NaCl)
1.0 (3% NaCl)
1.5 (3% NaCl)

2.86 0.58
2.79 0.54
3.85 0.61
2.66 0.66
2.48 0.94
2.510.19
2.370.50
3.160.26
3.28 0.42
2.97 0.69
3.02 0.60
3.27 0.76
3.22 0.39
2.61 0.12
2.96 0.87

343

Ultimate Tensile
Strength
(MPa)
4.81 0.64
4.34 0.62
4.59 0.29
3.85 0.09
3.87 0.73
4.75 0.45
4.25 0.47
4.05 0.47
4.18 0.22
4.07 0.47
4.64 0.32
4.41+0.45
4.70 0.35
4.120.17
4.120.17

Average
Crack Width
(um)
-45
-100
-100
-100
-100
-35
-100
-100
-100
-80
-30
-100
-100
-80
-90

U.JO

i"
.. 0.25
e
n

>ack V

s:

"S 0.10
re

s
0.05
n nn

ACI224

ACI222

BS8110

CEB/FIB

CP 110

ENV1991

JSCE

Figure 7.1 - Comparison of allowable crack widths under marine exposure.

5.0
4.5
4.0
3.5
3.0
2.5

2.0
1.5
1.0

ECC_Pre-cracking 0.5%
ECC_Pre-cracking 1.0%
ECC_Pre-cracking 1.5%

0.5
0.0
0.00

0.50

1.00

1.50

Tensile Strain (%)

Figure 7.2 - Typical precracking tensile stress-strain curves of ECC.

344

2.00

Loaded after 30 days


in NaCI Solution

2.0

2.5

3.0

4.0

4.5

5.0

T e n s i l e Strain (%)

(a) Not preloaded (precracked)


Re-loaded after 30 days
In NaCI Solution

2.0

2.5

3.0

Tensile Strain (%)

(b) Precracked to 0.5% tensile strain

345

4.5

5.0

Re-loaded after 60 days


in NaCI Solution

Re-loaded after 30 days


in NaCI Solution

4.5

5.0

(c) Precracked to 1.0% tensile strain


Re-loaded after 30 days
in NaCI Solution

Re-loaded after 90 days


in NaCI Solution

5.0

(d) Precracked to 1.5% tensile strain


Figure 7.3 - Tensile stress-tensile strain curves of ECC specimens before and after
exposure to 3% NaCI solution, (a) Not preloaded (precracked). (b) Precracked to 0.5%
tensile strain, (c) Precracked to 1.0% tensile strain, (d) Precracked to 1.5% tensile strain.

346

6.0

1>5Q

a.
S_

^ ^ ^ ^ ^ ^ - ^ ^ ^
-I^==a..

fc 4

if)

a 3.0

"Si
E

S 2.0
ra
E

--Not pre-cracked
- * - Pre-cracked-0.5%
- * - Pre-cracked-1.0%
->*- Pre-cracked-1.5%

1.0

0.0
10

20

30

40

50

60

70

80

90

100

Exposure Age (days)

Figure 7.4 - Influence of NaCl exposure time and applied strain level on ECC
ultimate tensile strength.

5.0
4.5
4.0
3.5

g..

2? 3.0
"o
ra
O
= 2.0
ra
55

1.5

- - N o t pre-cracked
~" Pre-cracked-0.5%
- * - Pre-cracked-1.0%
- * - Pre-cracked-1.5%

1.0
0.5
0.0

10

20

30

40

50

60

70

80

90

100

Exposure Age (days)

Figure 7.5 - Influence of NaCl exposure time and applied strain level on ECC
tensile strain capacity.

347

E 0.4
re
Q.

e, 0.3 i

30

60

90

120

Days of Immersion in 3% NaCI Solution

Figure 7.6 - Influence of NaCI solution exposure time on ECC matrix fracture toughness.

30

60

90

120

Days of Immersion in 3% NaCI Solution

Figure 7.7 - Influence of NaCI solution exposure time on ECC matrix/fiber interfacial
chemical bond.

348

30

60

90

120

Days of Immersion in 3% NaCI Solution

Figure 7.8 - Influence of NaCI solution exposure time on ECC fiber/matrix interfacial
frictional bond.

v 2
C
Q.

30

60

90

120

Days of Immersion in 3% NaCI Solution

Figure 7.9 - Influence of NaCI solution exposure time on ECC PSH Index.

349

References:
1

ACI Committee 224R-2001, "Control of Cracking in Concrete Structures," ACI Manual


of Concrete Practice, Part 3, American Concrete Institute, Farmington Hills, Michigan,
2001.
2

Tsinker, G. P., Marine Structures Engineering: Specialized Applications, 1995, pp. 548.

Mehta, P. K., and Gerwick, B. C , "Cracking-Corrosion Interaction in Concrete Exposed


to Marine Environment," Concrete International, Vol. 4, No. 10, pp. 45-51.
4

Gilbride, P., Morgan, D. R., Bremner, T. W., "Deterioration and Rehabilitation of Berth
Faces in Tidal Zones at the Port of Saint John," Shortcrete, Fall 2002, pp. 32-38.
5

Mehta, P. K., "Durability of Concrete Exposed to Marine Environment - A Fresh Look,'


ACI Special Publication, Vol. 19, 1988, pp. 1-30.
6

Liu, P. C , "Damage to Concrete Structures in a Marine Environment," Materials and


Structures, Vol. 24, 1991, pp. 302-307.
7

ACI Committee 224R-2001, "Control of Cracking in Concrete Structures," ACI Manual


of Concrete Practice, Part 3, American Concrete Institute, Farmington Hills, Michigan,
2001.

British Standards Institution. BS 8110: Pt.l, BSI, London, 1997.

CEB-FIP Model Code 1990, CEB Information Report No. 213/214, Comite EuroInternational DuBeton, Lausanne, May 1993.
10

Code of Practice for the Structural Use of Concrete - Part 1. Design, Materials and
Workmanship, British Standards Institution Publication CP 110, London, England,
November 1972 (Amended May 1977).
11

Standard Specification for Design and Construction of Concrete Structures - 1986, Part
1 (Design), Japan Society of Civil Engineers, SP-1, Tokyo, Japan, 1986.

12

ACI Committee 222, "Corrosion of Metals in Concrete," ACI 222R-89, American


Concrete Institute, Detroit, Michigan, 1975.
13

Markeset, G. Rostam, S., and Klinghoffer, O., "Guild for the Use of Stainless Steel
Reinforcement in Concrete Structures," Project Report 405, Nordic Innovation Center
Project - 04118: Corrosion Resistant Steel Reinforcement in Concrete Structures
(NonCor), 2006, p. 59.

350

Baysburd, A. M., Emmons, P. H., Mailvaganam, N. P., McDonald, J. E., and


Bissommette, B., "Concrete Repair Technology - A Revised Approach is Needed,"
Concrete International, Vol. 26, No. 1, Jan. 2004, pp. 58-65.
15

Cusson, D., and Mailvaganam, N., "Durability of Repair Materials," Concrete


International, Vol. 18, No. 3, Mar. 1996, pp. 34-38.
16

Grzybowski, M., and Shah, S. P., "Shrinkage Cracking of Fiber Reinforced Concrete,"
ACIMaterials Journal, Vol. 87, No. 2, Mar.-Apr. 1990, pp. 138-148.
17

Yang, E. H., "Designing Added Functions in Engineered Cementitious Composites,"


Ph.D. Dissertation, The University of Michigan, 2008.
18

Yang, Y. Z., Lepech, M. D., and Li, V. C , "Self-Healing of Engineered Cementitious


Composites under Cyclic Wetting and Drying," Proc. Int. Workshop on Durability of
Reinforced Concrete under Combined Mechanical and Climatic Loads (CMCL), Qingdao,
China, Oct. 2005, pp. 231-242.
19

Ramm, W., and Biscoping, M., "Autogenous healing and reinforcement corrosion of
water-penetrated separation cracks in reinforced concrete." Nuclear Engineering and
Design, Vol. 179, 1998, pp. 191-200.
20

Li, V. C , and Yang, E. H., "Self Healing in Concrete Materials," Self Healing
Materials: An Alternative Approach to 20 Centuries of Materials, by Sybrand van der
Zwaag, 2007, pp. 160-194.
21

Jacobsen, S., Marchand, J., and Homain, H., "SEM Observations of The
Microstructure of Frost Deteriorated and Self-healed Concrete," J. of Cement and
Concrete Research, Vol. 25, 1995, pp. 1781-1790.
22

Ismail, M., A.Toumi, R. Francois, and R. Gagne, "Effect of crack opening on local
diffusion of chloride inert materials," Cement and Concrete Research, Vol. 34, 2004, pp.
711-716.
23

Reinhardt, H., Joos, M., "Permeability and Self-healing of Cracked Concrete as a


Function of Temperature and Crack Width," J. of Cement and Concrete Research, Vol.
33,2003,pp.981-985.
24

Sahmaran, M., Li, M., and Li, V.C., "Transport Properties of ECC under Chloride
Exposure," accepted, ACI Materials J., 2007.
25

Edvardsen, C , "Water Permeability and Autogenous Healing of Cracks In Concrete,"


ACI Materials J., Vol. 96, 1999, pp.448-455.

351

26

Aldea, C , Song, W., Popovics, J. S., and Shah, S.P., "Extent of Healing of Cracked
Normal Strength Concrete," J. of Materials In Civil Engineering, Vol. 12, 2000, pp.92-96.
27

Clear, C.A., "The Effects of Autogenous Healing Upon the leakage of Water through
Cracks in Concrete," 1985, p. 28.
28

Lepech, M.D., and Li, V.C., "Water Permeability of Cracked Cementitious


Composites," ICF 11, Turin, Italy, Paper 4539 of Compendium of Papers CD ROM,
2005.
29

Lepech, M.D., "A Paradigm for Integrated Structures and Materials Design for
Sustainable Transportation Infrastructure," Ph.D. Dissertation, Department of Civil and
Environmental, University of Michigan, Ann Arbor, PhD Thesis, 2006.
30

Wang, K., Nelsena, D.E., and Nixon, W., "Damaging effects of deicing chemicals on
concrete materials," Cement and Concrete Composites, Vol. 28, No. 2, 2006, pp.173-188.
31

ASTM C 511, "Standard Specification for Moist Cabinets, Moist Rooms, and Water
Storage Tanks Used in the Testing of Hydraulic Cements and Concretes," American
Society for Testing and Materials, West Conshohocken, PA, Vol. 4, No. 2, 2002.
32

Delagrave, A., Pigeon, M., Marchand, J., and Revertegat, E., "Influence of chloride
ions and pH level on the durability of high performance cement pastes (Part II)," Cement
and Concrete Research, Vol. 26, No. 5, May 1996, pp. 749-760.
33

Wang, K., Nelsen, D. E., and Nixon, W. A., "Damaging Effects of Deicing Chemicals
on Concrete Materials," Cement and Concrete Composites, Vol. 28, No. 2, February
2006, pp. 173-188.

352

CHAPTER 8

Large-Scale Processing and Field Demonstration of High Early Strength


Cementitious Composites in Bridge Repair

This chapter shifts the focus from "repair system durability assessment" to the
"structural application" portion of the integrated engineering methodology (upper triangle
in Figure 1.10). Efforts are made to transfer innovative ECC repair technology from the
laboratory to field implementation through a bridge patch repair demonstration project.
Large-scale processing and construction of an HES-ECC repair using commercial
facilities are realized through optimization of the HES-ECC ingredients and mixing
procedure, trial batches, and quality control methods. A simple lifecycle cost analysis is
also performed to show the relative economic advantage of HES-ECC compared with
concrete and alternative field repairs. The long-term durability of the HES-ECC patch
repair under field conditions is monitored. Through this work, the linkages between
material engineering, repair system durability assessment, and structural application are
further forged.

353

8.1

Background
Before the newly-developed HES-ECC material can be considered as a viable

building material for use in repair or structural designs, it must be capable of onsite
processing and placement with adequate fresh properties and controlled quality at a large
scale, using commercial batching and handling equipments common within the repair
construction industry.
Large-scale mixing of ECC has been investigated in Japan1 for pre-casting and
spraying applications, and Michigan2'3 in a UM (University of Michigan) - MDOT
(Michigan Department of Transportation) project "Field Demonstration of Durable Link
Slabs for Jointless Bridge Decks Based on Strain-Hardening Cementitious Composites".
These investigations, especially those done in Michigan, provided good insights into
large-scale mixing processes for the newly developed HES-ECC material.
Kanda et al.1 investigated the tensile properties of ECC material produced using a
1.3 cubic yard (1 cubic meter) omni-mixer. This mixer is a force-based type that uses
external mixing paddles to deform a rubber drum containing the cementitious material.
This mixing equipment is substantially different from gravity mixers that rely mainly on
gravity to mix a viscous liquid (paddles within the rotating drum lift the material and
agitate it by dropping the material inside the drum). Force based mixers are typically
much more efficient in achieving homogeneity within concrete mixes because of the
greater mixing agitation. However, they are only common at pre-cast concrete plants, but
very uncommon on construction sites where repair work is carried out. Gravity mixers,
instead, are more typical equipment on repair construction sites. The present work
demonstrated and concluded that larger scale on-site production of ECC using a gravity

354

mixer was possible and the performance of the material was similar to that mixed in the
lab. Additionally, other work by Kanda et al.1 used both gravity and omni-mixers to
process ECC material for spraying applications. This limited study concluded that the
mechanical performance of ECC processed using an ordinary concrete gravity mixer was
similar to that using an omni-mixer.
Large scale mixing of a typical version of ECC (M45) using concrete mixing trucks
was carried out for the field demonstration of ECC link slabs for jointless bridge decks in
Michigan2. Before the truck mixing, grain size distribution analysis was conducted for
proportioning the ECC components to produce a free-flowing mixture. The length of
PVA fiber was changed from 0.5 in. (12.7 mm), as used previously in ECC M45, to 0.33
in. (8.4 mm) to promote easier mixing and better fiber dispersion in a gravity mixer. A
local concrete supplier was then contracted to process 1, 2 and 4 cubic yard (0.76, 1.53,
and 3.10 cubic meter) trial batches of ECC M45 material, which provided meaningful
lessons for the following processing of 20 cubic yards (15.3 cubic meters) of ECC M45
for the construction of ECC link slabs. In total, based on mixings ranging from 1 cubic
foot up to 20 cubic yards (0.03 to 15.3 cubic meters), it was concluded that large-scale
mixing of ECC was able to produce fresh material that was homogeneous, Theologically
stable, and had good flowability. Testing of mechanical properties of the hardened ECC
M45 from the large-scale batches showed that the material compressive properties were
similar to those of laboratory mixes, while the material tensile properties were reduced by
an acceptably small amount from those typically seen in laboratory-grade ECC M45
material2'3.

355

The material mix proportions of the ECC mix from Kanda et al.1, ECC M45, and
HES-ECC are compared in Table 8.1.
Compared with the other two ECC mixes that were previously investigated, HESECC contained a much higher content of cement, and no fly ash. In addition, the cement
type was high early strength Portland type III cement rather than normal strength
Portland type I cement, and the former had finer-sized particles (Blaine Surface area 2637
ft2/lb (540 m2/kg) for type III versus 1806 ft2/lb (370 m2/kg) for type I), resulting in a
different fresh rheology and shorter setting time after mixing with water. Furthermore,
an accelerator was used in HES-ECC that further shortened setting time, which put a tight
constraint on the working time. Finally, HES-ECC included polystyrene beads, which
affect the material fresh rheology (i.e. flowability, plastic viscosity, setting time) and
dispersion of fibers. These four distinguishing features of HES-ECC compared with
regular ECC pose extra challenges for its larger scale processing.

8.2

Optimization of HES-ECC Ingredients and Mixing Procedure for Larger

Scale Applications
8.2.1

Fiber Length Change


Prior to conducting larger scale HES-ECC batching, the fresh rheology of HES-

ECC was optimized to adapt to large gravity mixers. To achieve better flowability and
more homogeneous fiber distribution, a change of fiber length was examined.

The

standard poly-vinyl-alcohol (PVA) fiber used in HES-ECC was 0.5 in. (12 mm) long and
1.5 mils (0.039 mm) in diameter. After adding the fibers to the fresh HES-ECC mortar, a
significant increase in viscosity and reduction in flowability of the material was observed.

356

To improve the flowability of fresh HES-ECC, shorter PVA fibers with length of 0.33 in.
(8 mm) and diameter of 1.5 mils (0.039 mm) were used to replace the 0.5 in. (12 mm)long fibers. The two types of fibers (Figure 8.1) were identical with respect to chemical
composition, surface coating, strength, and modulus.
The fresh-state flowability and the hardened mechanical properties of HES-ECC
material using shorter fibers were measured. A flowability test was performed on the
fresh HES-ECC mix immediately after processing the material using a laboratory forcebased mixer with 12 L capacity. In this test, a standard concrete slump cone was filled
with fresh HES-ECC material and emptied onto a level sheet of Plexiglas. The flowable
HES-ECC mix then flattened, only by gravity, into a large pancake-shaped mass. Two
orthogonal diameter measurements of this "pancake" were recorded and averaged,
denoted as Di, which captured the overall deformability, or flowability, of the material in
its fresh state. The flowability test showed that Di of HES-ECC increased from 22 in. to
24 in. (559 mm to 610 mm) by changing from 0.5 in. (12 mm)-long fibers to 0.33 in. (8
mm)-long fibers. This additional flowability of HES-ECC material allowed for easier
mixing in large capacity gravity mixers and easier placement on the construction site.
Based on separate analytical studies5, this fiber length change was shown to have a
minimal effect on the crack bridging stress versus crack width opening curve G(8) and,
therefore, little expected effect on the overall pseudo-tensile strain-hardening behavior of
the ECC composite. To verify this, the tensile properties of HES-ECC with shorter fibers
were evaluated using a uniaxial tensile test, as described in earlier chapters. HES-ECC
specimens containing shorter fibers were demolded at 4h, cured in air, and tested at
different ages from 4h to 28d. Three specimens were tested at each age. Table 8.2
summarizes the measured tensile properties at different ages, and Figure 8.2 shows the
357

typical stress-strain curves at different ages. It was observed that when shorter fibers
were used, HES-ECC maintained tensile strain capacity of more than 3% at all ages.
Therefore, the 0.33 in. (8 mm) fibers were selected for use in the larger scale batching.

8.2.2

Hydration Stabilizing Admixtures


As a gravity mixer instead of a force-based mixer is generally used on repair

construction sites, longer mixing and working times should be necessary in achieving
homogeneity and handling of HES-ECC. Due to a very high amount of type III cement
and incorporation of hydration accelerator, HES-ECC set quickly after mixing was
finished. Additionally, HES-ECC included Glenium 3200 HES admixture as the highrange water-reducing admixture, which was originally developed for extremely high
early strength concrete. The slump retention of HES-ECC containing Glenium 3200
HES admixture was less than that of ECC M45 containing a conventional high-range
water-reducing admixture. For all of the above reasons, HES-ECC began to lose its
flowability and workability beyond the first 15-20 minutes after mixing was complete.
Although such a short setting time would pose no major problems for small-scale
specimens prepared for laboratory testing, this rapid loss of workability is not adequate
for larger scale construction repairs. Therefore, a hydration-retarding admixture was
needed to delay hydration at the very beginning and to extend setting when a longer
working time was desirable, without sacrificing the early-age compressive strength gain
rate.
A number of hydration retarders are commercially available. For this project
Delvo Stabilizer from MasterBuilders was selected due to its compatibility with the

358

accelerator PozzolitrT NC 534 and the superplasticizer HRWR Glenium 3200HES used
in HES-ECC. All three admixtures were produced by the same company. As declared
by MasterBuilders, Delvo Stabilizer can reduce water content as required to allow
workability and reduce segregation.

Additionally, it has been found that concrete

produced with Delvo Stabilizer can develop higher early-age (within 24 hours) and
late-age compressive strengths than plain concrete when used within the recommended
dosage range, and under normal curing conditions.
In the present study, dosage of Delvo Stabilizer in HES-ECC followed the
manufacturer's recommendation as 5 fl oz/cwt (65.2 mL/100 kg = 1.0 oz/cwt; cwt =
hundredweight, or 100 pounds) of cementitious materials.

HES-ECC containing

stabilizer as well as shorter fibers was mixed using a laboratory force-based mixer, and
the change in flowability with time was measured. As shown in Figure 8.3, HES-ECC
without stabilizer exhibited rapid loss of flowability and became hardly workable after
15-20 minutes. However, HES-ECC with stabilizer exhibited a higher flowability and
prolonged setting time, which was desirable for onsite larger-scale processing and
handling. Table 8.3 summarizes the mixing proportions and batching weights of HESECC containing shorter fibers and hydration stabilizer.

8.2.3

Batching Sequence
To scale up HES-ECC batching from small laboratory mixers to large gravity

mixers, the mixing sequence must be optimized to promote the best homogeneity of the
material. In typical laboratory mixing, for which a force-based mixer was used, all dry
ingredients of the HES-ECC matrix (cement and sand) were initially added to the mixer

359

and blended. Following a complete mixing of dry ingredients, water and high range
water reducer were slowly added to gradually turn the mixture into a liquid state. Fibers
were then slowly added and dispersed throughout the mixture. The accelerator was then
added. Finally polystyrene beads were slowly added and mixed until they were well
distributed. The overall mixing sequence lasted between 10 and 15 minutes.
For larger-scale mixing, however, the processing sequence described above was
not applicable. Adding all of the dry components, including cement, followed by a small
amount of water (due to the low water/cement ratio) resulted in a large mass of very dry
material, which was difficult for a low-energy gravity mixer to process. Subsequently
adding high range water reducer did not help to convert the dry mix into a liquid state,
mainly due to the low speed and energy of the gravity mixer. Therefore, it was found to
be essential that the mixture remained in a liquid state as long as possible throughout the
mixing process and attained its most viscous state at the end of mixing. To meet this
objective, the processing sequence using gravity mixers (with 2 cubic feet and 9 cubic
feet capacities) was adjusted so that first all of the dry sand was added, along with three
quarters of the water and all of the superplasticizer, as well as the stabilizer. Once these
four components were well mixed, cement was added and mixed until the complete
mortar matrix of ECC achieved a homogenous liquid state. The remaining mix water
was then added with half of the accelerator admixture to wash the cement particles off the
wall of the mixing drum. Fibers were subsequently added gradually into the mixture and
mixed until they were well dispersed. Polystyrene beads were then added slowly and
mixed until they were well distributed. In the end, the remaining half of the accelerator
admixture was added until the complete material was homogenous and ready for

360

placement.

This mixing sequence, which took 25 minutes in total, is shown with

summaries of the time elapsed for each mixing step in Table 8.4.

8.3

HES-ECC Larger Scale Trial Batching

8.3.1

One Cubic Feet Batching


The goal of this one cubic foot (0.03 cubic meter) batching was to verify the

applicability of the new mixing sequence of HES-ECC in a small gravity mixer with
capacity of 2 cubic feet (0.06 cubic meters), as shown in Figure 8.4, and to examine the
material's fresh and hardened properties when including 0.33 in. (8.4 mm) fibers and
hydration stabilizer. Figure 8.5 illustrates the major mixing steps. The whole mixing
process went as anticipated, validating the applicability of the suggested batching
sequence for 1 cubic feet (0.03 cubic meters) batching of HES-ECC.
A flowability test was conducted on the fresh HES-ECC from the one cubic foot
(0.03 cubic meter) batching, and the results are shown in Table 8.5 and Figure 8.6. The
rate of reduction in flowability with time was similar to that of HES-ECC processed
using the laboratory force-based mixer, as shown in Figure 8.3.
HES-ECC compressive cylinder specimens and tensile coupon specimens were
cast from the one cubic foot (0.03 cubic meter) batch and tested for hardened mechanical
properties at different ages. Compressive testing was done according to the ASTM C39
"Standard Test Method for Compressive Strength of Cylindrical Concrete Specimens" on
standard 4 in. x 8 in. (102 mm x 203 mm) cylinders. Three specimens were tested for
each test series. From the compressive test results summarized in Table 8.6 and Figure

361

8.7, it can be seen that HES-ECC processed from one cubic foot (0.03 cubic meter)
batching was able to achieve the target compressive strength at both early and late ages.
Uniaxial tensile tests were conducted on HES-ECC tensile plate specimens cast
from the one cubic foot (0.03 cubic meter) batch, with the same specimen size and test
setup described in previous chapters. Three specimens were tested for each test series
and their tensile properties are summarized in Table 8.7. Figure 8.8 shows representative
tensile stress-strain curves of the HES-ECC mix at different ages. It can be seen that
HES-ECC processed from one cubic foot (0.03 cubic meter) batching was able to achieve
the target >2% tensile strain capacity at both early and late ages.

8.3.2

Three Cubic Feet Batching and Six Cubic Feet Batching


3 and 6 cubic feet (0.08 and 0.17 cubic meter) batches of HES-ECC were

produced using a larger gravity mixer with capacity of 9 cubic feet (0.25 cubic meters),
as shown in Figure 8.9. The processing steps are illustrated in Figure 8.10. As expected,
due to the absence of large aggregates to agitate materials within the mixing drum,
additional mixing time was needed between charging of the cementitious materials and
the fibers. This added 5 to 10 minutes of mixing time provided further agitation and time
for the high-range water-reducer to liquefy the material. Upon completion of mixing, the
HES-ECC material for both batch sizes was evaluated for fiber distribution, flowability
and mixture rheology. Fiber dispersion was evaluated through visual inspection and
random sampling of the material to look for pockets or conglomerates of unmixed matrix
materials or fiber bundles. Both batches produced close to homogenous material without
segregation.

362

For the 3 cubic feet (0.08 cubic meter) batching, after achieving a homogenous
HES-ECC mix, a small additional amount of superplasticizer was added to evaluate the
possibility of further increasing the material's flowability. However, segregation was
observed after adding the extra superplasticizer.

Therefore, the mixed HES-ECC

material was disposed of and was not evaluated for its compressive and tensile properties.
From this experience, it was concluded that flowability higher than Di=30 in. (762 mm)
of HES-ECC without segregation could not feasibly be achieved, if no additional
admixtures (e.g. viscosity modifying agents) are included. Therefore, the following 6
cubic feet (0.17 cubic meter) batching strictly followed the mixing proportions specified
in Table 8.3.
Evaluation of the fresh HES-ECC from the 6 cubic feet (0.17 cubic meter)
batching was done through flowability testing of the material. For this test, a standard
concrete slump cone was filled with fresh HES-ECC and the cone was then removed,
resulting in a "pancake" as shown in Figure 8.10 (f). The average deformation diameter
Di was 24 in. (610 mm). The flowability test results are summarized in Table 8.5. The
loss of deformability of HES-ECC processed in six cubic feet (0.17 cubic meter) batching
is shown in Figure 8.11.
Similar to the 1 cubic feet (0.03 cubic meter) batching, HES-ECC compressive
cylinder specimens and tensile coupon specimens were cast from the 6 cubic feet (0.17
cubic meter) batch and tested for hardened mechanical properties at different ages. Three
specimens were tested for each test series. The compressive test results are summarized
in Table 8.6 and Figure 8.12. The tensile test results are summarized in Table 8.7 and
Figure 8.13. It was concluded that HES-ECC processed from the 6 cubic feet (0.17 cubic

363

meter) batching was able to achieve the target compressive strength and tensile strain
capacity at both early and late ages.
The scaled up batching of HES-ECC demonstrated the success of the modified
HES-ECC mix design and processing procedure. The specifications on quality control
for constructing an HES-ECC repair are listed in Li et al. 20066.

8.4

HES-ECC Material Cost Effectiveness


The material costs (not including installation cost) of HES-ECC compared with

various traditional or new repair materials are compared as below:


Portland cement concrete:

$80/yard3 ($105/m3)

Steel fiber reinforced concrete:

$195/yard3 ($255/m3)

High strength concrete:

$200/yard3 ($262/m3)

Latex modified concrete:

$275-$300/yard3 ($360-$392/m3)

Gypsum based patching material:

$415/yard3 ($543/m3)

Normal PVA-ECC:

$220/yard3 ($288/m3)

Thoroc 10-60:

$595 /yard3 ($778/m3)

HES-ECC:

$615/yard3 ($804/m3)

Polymer concrete/mortar:

$2700/yard3($3531/m3)

Ductal:

$3000/yard3 ($3924/m3)

The relative high cost of HES-ECC mainly comes from the large amount of fine
silica sand, PVA fibers, and accelerators used in the mixture. However, the HES-ECC
repair is still expected to be cost effective compared with other repair materials, from the
viewpoint of service life cost, and in some cases even first cost. First cost includes both

364

material cost and installation cost. It may also include crack sealing cost for concrete
deck repairs, which appears to be an accepted practice. The HES-ECC material cost is
$615/yard3 ($804/m3) currently, which is higher than Portland cement concrete
($80/yard3 ($105/m3)), high strength concrete ($200/yard3 ($262/m3)), Steel Fiber
Reinforced Concrete (SFRC) ($195/yard3 ($255/m3), and Gypsum based patching
material ($415/yard3 ($543/m3)). However, HES-ECC is similar to Thoroc 10-60, which
is a very rapid-setting one-component cement mortar, and is much cheaper than many of
the other currently used "high performance" repair materials, such as Ductal ($3000/yard3
($3924/m3), marketed by LaFarge), and Polymer Concrete/Mortar

($2700/yard3

($3531/m3)). It should be noted that repair material cost only occupies a small portion of
the repair first cost. Research work done by the Virginia Department of Transportation7'8
concluded that the major cost of current repair overlay construction derives from costs
associated with labor, equipment, mobilization, and traffic control, and material cost is
generally less than 10% of the total repair overlay cost. In addition, the construction of
the composite steel/ECC deck of a cable-stayed bridge in Hokkaido, Japan cites the
reduction of installed cost (construction speed and lower labor cost) as one reason behind
the choice of ECC over other cementitious materials.

Other reasons cited included

durability (estimated 100 year service life) and low weight (40% reduction compared
with normal concrete with thicker section). Therefore, despite that the material cost of
HES-ECC is higher than that of concrete, high strength concrete and SFRC, the savings
in labor and equipment costs related to decreased traffic control and congestion during
shortened construction and curing time can minimize the first cost of HES-ECC repairs.
From the service life cost point of view, HES-ECC repairs are expected to last for

365

a significantly longer time, compared with repairs made from traditional mortar or
concrete-type materials, which are brittle. As has been documented in previous chapters,
an HES-ECC repair can effectively suppress repair surface cracking and repair/concrete
interface delamination, reflective cracking, and chloride penetration, thereby preventing
common deterioration in concrete repairs.

The HES-ECC repair has also been

demonstrated to have a significantly longer fatigue life. As a result, the large savings
from reduced future maintenance and repairs should make HES-ECC very competitive
economically.
A preliminary life cycle cost (LCC) analysis was conducted based on the following
assumptions:
a) Deck patch repair incurs the following costs: material, chipping (removal of old
material), traffic control and mobilization.
b) Other related tasks (e.g. pier repair) incur the following costs: material, chipping,
forming, and mobilization.
c) Repair depth = 4 in. (102 mm); each cubic yard (0.03 cubic meters) of material
converts to 9 square yards (7.53 square meters) of patch area.
d) Discount factor 1=3%.
e) All numbers below are based on the following MDOT document:
http://www.michigan.gov/documents/MDOT_14aCSM_Workbook05_126884_7.xls
Materials:
Concrete patch material cost = $115/yard3 ($150/m3) [concrete material] +
$685/yard3 ($896/m3) [repair premium charge] = $800/yard3 ($1046/m3)

366

HES-ECC patch material cost = $615/yard3 ($804/m3) [HES-ECC material] +


$685/yard3 ($896/m3) [repair premium charge] - $1300/yard3 ($1700/m3)
Chipping:
Deck: $125/yard2 = $1125/yard3 (for 4 in. repair depth)
($149/m2 = $1471/m3 (for 102 mm repair depth))
Pier: $65/feet2 - $1755/yard3 (for 4 in. repair depth)
($700/m2 = $2295/m3 (for 102 mm repair depth))
Forming:
$28/feet2 = $2268/yard3 (for 4 in. repair depth)
($301/m2 = $2966/meter3 (for 102 mm repair depth))
Traffic control:
Assuming traffic control related cost is $3000/day; patch repairs typically require
2 days, therefore traffic control related cost = $6000.
Mobilization:
5% of all other costs.

Based on assumptions a) - e), the following calculation can be performed:


Deck patch repair cost per cubic yard:
Normal concrete patch repair: $800/yard3 ($1046/m3) [material] + $1125/yard3
($1471/m3) [chipping] + $6000/yard3 ($7848/m3) [traffic] + $396/yard3 ($518/m3)
[mobilization] = $8320/yard3 ($10882/m3)

367

HES-ECC patch repair:

$1300/yard3 ($1700/m3) [material] + $1125/yard3

($1471/m3) [chipping] + $6000/yard3 ($7848/m3) [traffic] + $421/yard3 ($551/m3)


[mobilization] = $8846/yard3 ($11570/m3)

Other than deck patch repair cost per cubic yard:


Normal concrete patch repair: $800/yard3 ($1046/m3) [material] + $1755/yard3
($2295/m3) [chipping] + $2268/yard3 ($2966/m3) [forming] + $241/yard3 ($315/m3)
[mobilization] = $5064/yard3 ($6600/m3)
HES-ECC patch repair:

$1300/yard3 ($1700/m3) [material] + $1755/yard3

($2295/m3) [chipping] + $2268/yard3 ($2966/m3) [forming] + $266/yard3 ($348/m3)


[mobilization] = $5589/yard3 ($7310/m3)

The question to be discussed here is: Assuming the service life of a normal concrete
patch is 3 or 10 years, how long must an HES-ECC patch last for its life cycle cost to
equal that of a normal concrete patch?
Deck patch repair
(a) Assumption Case 1: Normal concrete patch life = 3 years
Assuming the present value of LCC of a normal concrete deck patch with
service life of 3 years equals to the present value of LCC of a HES-ECC deck
patch with service life of x years:
.
$8320+

$8320

$8320
r- +

$8320
r +

$8320
+

$8320
pr +

TT+ ....

(l+3%) 3 (l+3%) 6 (l+3%) 9 (l+3%) 12 (l+3%) 15


,.,,
$8846
$8846
$8846
$8846
$8846
= $8846 +
+
+
+
+
+ ....
(1+3%)'

(\+3%)2x

(l+3%) 3 *

By solving this equation we get x = 3.2 years


368

(l+3%) 4 *

(1+3%)5X

Therefore, the HES-ECC deck patch must last at least 3.2 years for its LCC to
equal or be less than that of a normal concrete deck patch with service life of 3 years,
(b) Assumption Case 2: Normal concrete patch life =10 years
Assuming the present value of LCC of a normal concrete deck patch with
service life of 10 years equals the present value of LCC of a HES-ECC deck
patch with service life of y years:
*oo

$8320

$8320 +

$8320
T7T +

$8320
zjr +

$8320
rr- +

$8320
-7T +

7^+

(l+3%) 10 (l+3%) 20 (l + 3%)30 a + 3 % ) 4 0 (l+3%) 50


*OOA*
= $8846
+ $
+ $
+ $
+ $
+ $
+ ....
8846

8846

(l+3%) y .

8846

(l+3%) 2y

8846

(l+3%) 3>

8846

(l+3%) 4 *

(l + 3%)5>

By solving this equation we get y = 10.5 years


Therefore, the HES-ECC deck patch must last at least 10.5 years for its LCC to
equal or be less than that of a normal concrete deck patch with service life of 10 years.

Non-deck patch repair


(a) Assumption Case 1: Normal concrete patch life = 3 years
Assuming the present value of LCC of a normal concrete non-deck patch with
service life of 3 years equals the present value of LCC of a HES-ECC nondeck patch with service life of x years:
,Cft
$5064 +

$5064

(l+3%) 3
<.____
$5589

$5064
+

= $5589 +

T+

$5064
+

$5064
TT +

HT +

(l+3%) 6 (l+3%) 9 (l+3%) 12 (l+3%) 15


$5589
$5589
$5589
$5589

+
(1+3%)*

$5064

(l+3%) 2 *

+
(l+3%) 3 *

(1+3%)4X

+ ....

(l+3%) 5 *

By solving this equation we get y = 3.3 years


Therefore, the HES-ECC non-deck patch must last at least 3.3 years for its LCC
to equal or be less than that of a normal concrete non-deck patch with life of 3 years.
369

(b) Assumption Case 2: Normal concrete patch life =10 years


Assuming the present value of LCC of a normal concrete non-deck patch with
life of 10 years equals to the present value of LCC of a HES-ECC non-deck
patch with life of y years:
&cn^A

$5064 +

$5064

$5064
nr +

$5064
+

$5064
rr- +

$5064
2T +

+ ....

(l+3%) 10 (l+3%) 20 (l+3%) 30 (l+3%) 4 0 (l+3%) 50


,m
$5589
$5589
$5589
$5589
$5589
= $5589 +
+
+
r- +
+
^ - + ....
(l+3%) y

(l+3%) 2>

(l+3%) 3y

(l+3%) 4 y

(l+3%)5>

By solving this equation we get y = 10.9 years


Therefore, the HES-ECC non-deck patch must last at least 10.5 years for its LCC
to equal or be less than that of a normal concrete non-deck patch with life of 10 years.
The Thoroc 10-60 Rapid Mortar is a patch repair material used by MDOT. It is a
very rapid-setting cement mortar containing crystalline silica, Portland cement, hydraulic
cement and lithium carbonate. During the HES-ECC patch repair demonstration project,
patch repairs made of Thoroc 10-60 and HES-ECC were poured side by side to compare
their early-age and long-term performance.

The material cost of Thoroc 10-60 is

$595/yard3 ($778/m3), which is similar to the $615/yard3 ($804/m3) material cost for
HES-ECC. Assuming the installation costs for HES-ECC and Thoroc 10-60 are the same,
the preliminary LCC analysis shows that both the Thoroc 10-60 patch and the HES-ECC
patch must last at least 3.2 years (deck patch) or 3.3 years (non-deck patch) for their LCC
to equal or be less than that of a normal concrete patch with service life of 3 years; both
the Thoroc 10-60 patch and the HES-ECC patch must last at least 10.5 years (deck patch)
or 10.9 (non-deck patch) for their LCC to equal or be less than that of a normal concrete
patch with service life of 10 years.
370

The above LCC calculations suggested that regarding economic cost, the use of
HES-ECC is preferable to other repair materials if the repair service life can be extended
by approximately 10%.

The findings from previous chapters of this dissertation

suggested that service life of a HES-ECC patch repair should be significantly longer than
traditional concrete repairs, thus justifying the cost effectiveness of HES-ECC as a patch
repair material.
Note that the above analyses do not account for user or environmental costs. If
these costs were accounted for, the longer service life of HES-ECC over other repair
materials would give additional advantage to HES-ECC.

8.5

HES-ECC Patch Repair Demonstration


Construction of a patch repair on the Ellsworth Road Bridge over US-23 (S07 of

81074), Ann Arbor, Michigan began at 9am on Tuesday, November 28, 2006 with
placement of traffic control devices. Photographs of the construction process are shown
in Figures 8.15 - 8.28. The MDOT repair personnel conducted partial depth concrete
removal, sandblasted the patch area, replaced damaged reinforcement, and completed
patch preparation. Before pouring the patch materials, water fog was sprayed onto the
concrete substrate to enhance repair bonding. The average temperature was 55 F (12.8
C) and relative humidity was 59.5% during this construction.
HES-ECC mixing materials were pre-packed in the laboratory for easy field
processing, and then transported to the repair site. 7 cubic feet (0.20 cubic meter) HESECC was mixed by UM research personnel using a 12 cubic feet (0.34 cubic meter)
capacity concrete gas mixer provided by MDOT (Figure 8.22). The total mixing time

371

was approximately 30 minutes. The fresh HES-ECC exhibited desirable viscosity, fiber
distribution and flowability (Figure 8.25). Flowability testing using a slump cone was
conducted immediately after finishing the mixing and before repair placement. With a
25.6 in. (650.2 mm) slump diameter, the HES-ECC achieved a self-compacting state.
The HES-ECC material was then poured into a wheelbarrow, transported to the patch
area, and poured into the patch. The self-compacting property of HES-ECC allowed it to
easily flow into the corners of the patch area without any vibration. The surface of the
HES-ECC patch repair was finalized by using a steel trowel, by hand.
Mixing of Thoroc 10-60 was performed by MDOT personnel using the same
mixer as for mixing HES-ECC. It can be seen from Figure 8.26 that the fresh Thoroc 1060 was barely flowable, resulting in much more effort and a longer time for construction
workers to place and finish. The two patch repairs using HES-ECC and Thoroc 10-60 are
shown side-by-side in Figure 8.27. Latex-based curing compound was sprayed onto both
patches to reduce evaporation and shrinkage.
The two side-by-side patch repairs on the Ellsworth Road Bridge were completed
at around 1pm on Tuesday, November 28, 2006, as shown in Figure 8.28.

Traffic

reopened at 9am on Wednesday, November 29, 2006.

8.6

Long-Term Monitoring of HES-ECC Patch Repair Performance


After placement, the HES-ECC and Thoroc 10-60 patch repairs were exposed to

identical traffic and environmental loads, thereby allowing for a comparison of their
long-term durability performance. (The HES-ECC patch was on the shoulder lane and
therefore experienced less severe direct traffic loading.) The bridge has a high average

372

daily traffic (ADT) of 22,546, and it is also used frequently by 11-axle trucks that are
heavily loaded with aggregates. Hence, mechanical (traffic) loading on the patch repairs
is significant. Potential environmental loading (S.E. Michigan weather exposure) on the
patch repairs includes shrinkage, temperature change, freezing and thawing, chloride
penetration, salt scaling and potential steel corrosion.
The performance of the HES-ECC and adjacent Thoroc 10-60 patch repairs has
been assessed through a series of twelve site visits. The initial visit was 1 day after
construction (November 29, 2006). The last visit was on September 18, 2009. The
maximum surface crack width and maximum interfacial delamination width of the HESECC and Thoroc 10-60 patch repairs at different ages are plotted in Figure 8.29 and 8.30.
Several microcracks with width of 60 um were observed in the HES-ECC patch 2 days
after construction. After nearly 3 years of service (November 2006 - September 2009),
the width of microcracks in the HES-ECC patch remain less than 90 urn, although more
microcracks have been found over time. The interfacial delamination width between the
HES-ECC patch and existing concrete is less than 70 (am. For the Thoroc 10-60 patch
repair, several cracks with maximum crack width of 300 um were observed 34 days after
construction, and the maximum crack width increased to 1700 urn at the time of the last
visit (September 2009). Microspalling was also observed at some crack locations. The
interfacial delamination width between the Thoroc 10-60 patch and existing concrete was
900 um 34 days after construction, and increased to 2200 um at the time of the last visit.
Therefore, at the same age of approximately 2 years and 10 months, and under the given
mechanical and environmental loading conditions, the maximum surface crack width and
maximum interfacial delamination width of the HES-ECC patch were both much smaller

373

than in the Thoroc 10-60 patch (Figure 8.31, 8.32, 8.33, 8.34, 8.35).

Furthermore,

exposure of aggregates in the Thoroc 10-60 patch repair was observed close to the
Thoroc 10-60/HES-ECC interface area, while no disintegration in the HES-ECC patch
was observed (Figure 8.35). No spalling or other deterioration has been found in the
HES-ECC patch to date.
This field performance data for 2 years and 10 months of patch life reveals the
promise of HES-ECC as a durable repair material in the harsh environments of Michigan
winters (with combined effects of restrained shrinkage, freezing and thawing, exposure to
de-icing salts, temperature change, etc.) in addition to heavy traffic loads.

The

performance of the HES-ECC patch repair and Thoroc 10-60 patch repair on Ellsworth
Road Bridge, Michigan, will continue to be monitored and reported in future publications.

8.7

Conclusions
Within this work, the design and optimization of HES-ECC ingredients and a

processing procedure for large-scale commercial batching were completed.

These

included the reduction of fiber length from 12 mm to 8 mm, incorporation of a hydration


stabilizing admixture, and revision of the mixing sequence. The three scaled-up trial
mixtures of 1, 3, and 6 cubic feet using gravity mixers verified that larger-scale onsite
mixing of HES-ECC is feasible and can result in a material that is both high performance
by retaining overall tensile ductility and micfocracking behavior, and commercially
producible using gravity mixer-based batching operations.
The capability of onsite large-scale processing and placement of HES-ECC was
demonstrated in a bridge patch repair construction project on the Ellsworth Road Bridge

374

over US-23 (S07 of 81074) on November 28, 2006. The construction process went
smoothly, and produced a close-to-homogeneous, self-compacting HES-ECC material
that was easy to pour into the patch area without external vibration.

Testing of

mechanical properties of HES-ECC specimens made from the same batch showed that
the compressive strength and tensile properties (tensile strain capacity, tensile strength,
multiple microcracking behavior) were similar to those of laboratory mixtures. The
repair construction was completed at around 1pm on Tuesday, November 28, 2006.
Traffic reopened early at 9am on Wednesday, November 29, 2006.
A simple lifecycle cost analysis was also performed to show the relative economic
advantage of HES-ECC compared with concrete and alternative field repairs.

This

analysis suggests that just a 10% increase in service life would lead to a lower life cycle
cost if HES-ECC (with higher material cost) were adopted, compared with current patch
repair materials.
The long-term durability of the HES-ECC patch repair under field conditions has
been monitored up to approximately 2 years and 10 months after patch placement. Under
heavy traffic loading, restrained shrinkage, temperature effects, three winters' exposure
to freezing and thawing as well as de-icing salts, and other possible loading conditions in
the field, the HES-ECC patch repair exhibited multiple microcracking behavior with
crack width below 100 jim, repair/old interfacial delamination width no greater than 60
(xm, and no sign of disintegration or fracture failure, which outperformed the adjacent
Thoroc 10-60 patch repair constructed at the same time and which experienced the same
loading conditions. While the shoulder lane placement favors the durability of the HESECC patch compared with the Thoroc 10-60 patch that was in the right hand driving lane,

375

the durable performance of HES-ECC shown in this patch repair was consistent with
another patch repair made with ECC M45 (without the high early age strength) placed on
Curtis Road in Ann Arbor, Michigan in 20029.
Overall, this work proved the capability of onsite large-scale processing and
placement of HES-ECC using traditional equipment, and further validated the long-term
durability of HES-ECC repairs under combined mechanical and environmental loading
conditions under field conditions.

376

Table 8.1 - ECC material mixing proportions.


Material

Kanda et al.

ECC - M45

HES-ECC

Cement

1.0

1.0

1.0

Sand

0.91

0.8

1.0

Fly Ash

0.43

1.2

0.0

Water

0.65

0.53

0.33

PS Beads

0.0

0.0

0.064

High Range Water Reducer

0.0

0.03 Plastol 5000

0.0075 GL3200-HES

Accelerator

0.0

0.0

0.04

Anti-Shrinkage Agent

0.027

0.0

0.00

Fiber (Vol. %)

0.02

0.02

0.02

Table 8.2 - Tensile properties of HES-ECC with 0.33 in. (8.4 mm) PVA fiber.
Age
4h

24h

7d

28d

Young's Modulus
Ksi (GPa)
1728.26 (11.9)
1913.22 (13.2)
1997.68 (13.8)
2553.43 (17.6)
2596.56 (17.9)
2687.32 (18.5)
2901.12 (20.0)
2987.55 (20.6)
3102.28 (21.4)
3245.42 (22.4)
3376.51 (23.3)
3401.63 (23.5)

Ultimate Strength
psi (MPa)
511.54 (3.53)
518.17 (3.57)
525.73 (3.62)
679.11 (4.68)
690.32 (4.76)
702.12 (4.84)
820.33 (5.66)
829.67 (5.72)
851.52 (5.87)
811.67 (5.60)
832.21 (5.74)
908.82 (6.27)

377

Strain Capacity
(%)

4.21
5.52
5.66
3.95
4.29
4.43
3.36
3.51
3.64
3.28
3.55
3.57

Table 8.3 - Larger scale HES-ECC mixing proportions and batching weights.
Amount

Material

Proportion

Cement

1.0

1547.37 (3.32xl0-2)

Sand

1.0

1547.37 (3.32xl0"2)

Water

0.33

507.54 (1.09xl0~2)

PVA Fiber (Vol. %)

0.02

44.01 (9.45xl0"4)

Polystyrene Beads

0.064

99.09 (2.13xl0-3)

High Range Water Reducer

0.0075

11.61 (2.49X10"4)

Accelerator

0.04

61.83 (1.33xl0"3)

Hydration Stablizer

0.0046

7.02 (1.51xl0-4)

lb/yard3 (kg/m3)

Table 8.4 - Larger scale HES-ECC batching sequence and time.


Elapsed Time
including Mixing Time

Activity

(min)
1. Charge all sand

2. Charge 3/4 amount of mixing water, all


HRWR, and all hydration stabilizer
3. Charge all cement slowly

4. Charge remaining 1/4 amount of mixing


water to wash drum
5. Charge 1/2 amount of accelerator, mix
until the whole HES-ECC matrix material is
homogenous
6. Charge PVA fibers slowly, mix until the
fibers are well distributed
9. Charge PS beads slowly, mix until the
beads are well distributed
10. Charge the remaining 1/2 amount of
accelerator, mix until the complete material
is homogenous and ready for placement
Total

378

5
5
4
2
25

Table 8.5 - Flowability test results from one & six cubic feet (0.03 & 0.17 cubic
meter) batching using a gravity mixer.
Average Deformation Diameter, Dj
in. (mm)
Minute
Six Cubic Feet
One Cubic Foot
3
22.83 (580)
24.02 (610)

21.26 (540)
10
18.50 (470)
20.47 (520)
20
17.72 (450)
17.72 (450)
30
15.75 (400)
15.35 (390)
40

Table 8.6 - Compressive Strength development of HES-ECC from one & six cubic feet
(0.03 & 0.17 cubic meter) batching using a gravity mixer.
Comp ressive Strength
psi (MPa)
Age
One Cubic'.Foot
Six Cubic Feet
Target
3012
2500
2623
2718
2932
2511
2732
4h
(17.2)
(20.8)
(17.3)
(18.1) (18.7) (20.2)
(18.8)
4876
5113
5526
4722
3000
5428
5209
6h
(20.7)
(33.6) (35.3) (37.4)
(38.1)
(32.6)
(35.9)
5000
6258
5873
6115
6211
5977
6289
24 h
(34.5)
(42.8)
(41.2)
(43.4)
(43.1V (40.5) (42.2)
8621
8209
8012
8701
8454
7000
8500
28 d
(48.3)
(59.4) (56.6)
(55.2)
(60.0)
(58.3)
(58.6)

379

Table 8.7 - Tensile properties of HES-ECC from one & six cubic feet (0.03 & 0.17 cubic
meter) batching using a gravity mixer.
Young's Modulus Ultimate Strength Strain Capacity
Age
ksi (MPa)
psi (MPa)
(%)
1728 (11916)
512 (3.53)
4.21
1913 (13191)
518 (3.57)
5.52
4h
1998 (13774)
526 (3.62)
5.66
One
2553 (17605)
679 (4.68)
3.95
2597 (17903)
690 (4.76)
4.29
Cubic 24h
2687 (18528)
702 (4.84)
4.43
Foot
3.28
3245 (22376)
812 (5.60)
3.55
3377 (23280)
832 (5.74)
28d
3.57
3402 (23453)
909 (6.27)
4.05
1812 (12493)
497 (3.43)
1901 (13107)
4.53
4h
522 (3.60)
1979 (13645)
4.68
533 (3.67)
601 (4.14)
4.34
2000 (13790)
619 (4.27)
4.52
2103 (14500)
6h
Six
629 (4.34)
5.01
2210 (15237)
Cubic
2499 (17230)
678 (4.67)
3.43
Feet
2587 (17837)
689 (4.75)
3.55
24h
2713 (18705)
695 (4.79)
3.87
3.02
3302 (22766)
813 (5.61)
3.26
3421 (23587)
826 (5.70)
28d
3.62
3499 (24125)
851 (5.87)

Table 8.8 - Minimum requirements of HES-ECC material.


Minimum Requirements
of HES-ECC Material
Compressive Strength
psi (MPa)
Tensile Strength [Uniaxial]
psi (MPa)
Ultimate Tensile
Strain Capacity [Uniaxial]

4 hour

6 hour

24 day

28 day

2500
(17.24)
400
(2.76)

3000
(20.68)
500
(3.45)

5000
(34.47)
600
(4.14)

7000
(48.26)
700
(4.83)

2%

380

Figure 8.1 - PVA fiber, 0.5 in. (12.7 mm) and 0.33 in. (8.4 mm).

Q.

v>
w
<D
.*t

CO

0%

1%

2%

3%

4%

5%

Tensile Strain

Figure 8.2 - Tensile stress-strain curves of HES-ECC with 0.33 in. (8.4 mm) PVA fiber.

381

30
V

25
E

-w/o stablizer

- 700

w/stablizer

- 600

20

- 500

rma
1 (in

(0

15

400

c
o

*>5 ' -'

O
0)

O
0)

ji

300
10
- 200

S)
(0

5
>
<
I

- 100

<

20

40

60

Time,t (min)

Figure 8.3 - Loss of deformability of HES-ECC with and without hydration


stabilizer.

""*-.. pw^i

7**

'4'-*M6_
jk&.

A-

Figure 8.4 - Gravity mixer with capacity of two cubic feet (0.057 cubic meters).

382

a) After adding sand, water, HRWR and hydration stabilizer.


7 ' 71
__S

b) After adding cement (homogenous and creamy HES-ECC mortar matrix).

c) After adding PVA fibers.

383

d) After adding PS beads.


Figure 8.5 - HES-ECC one cubic feet (0.03 cubic meter) mixing.

30
-1 cubic feet batching

25 h

L
E

O Q
<U

20

500

15

400
300

10

- 200

0)
Ol

100
-H
20

t
40

9)
4-1

0)

- 600

>
<

- 700

- 0
60

E
(O

O
c
o ~-

If
Q
4)

O)

a
i

0)

>

Time,t(min)

Figure 8.6 - Loss of deformability of HES-ECC processed in one cubic feet (0.03 cubic
meter) batching.

384

U)

c
0)
4-*

CO
<D

>
'w
(A

0)
Q.

o
O

10000 -9000 -8000 -7000 -6000 5000 4000 3000 2000 1000 0 -L

<0

60
50

+ 40
Target

-ft
4-1

CD

- 30

CO
CD

-- 20

i/)

>
'55
CD

-- 10
10

o
O

100

1000

Age (hour)

ooooo o
ooooo o

400
300
200
100
0

- 6

J i _ ^ ^ - * ^ ^ - ^ * ^ 28 days

l^lmims~*~~~~^^
If

24 hours

- 4

4 hours ^
- 2

1B
Q.

2
</)
a)

Tensile.

2L

<J1 O) ->J 00 CD O

s Stress (psi)

Figure 8.7 - Compressive strength development of HES-ECC from one cubic feet
batching.

0%

1%

2%

3%

4%

- 0
5%

Tensile Strain

Figure 8.8 - Tensile stress-strain curves of HES-ECC from one cubic feet batching.

385

Figure 8.9 - Gravity mixer with capacity of 9 cubic feet (0.25 cubic meters).

386

a) Pre-wetting the mixer.

b) After adding sand, water, HRWR and hydration stabilizer.

c) After adding cement (homogenous and creamy HES-ECC mortar matrix).


387

m #8
ft.,. i

d) After adding PVA fibers.

e) After adding PS beads.

WP

f) Flowability test (5 minutes).


Figure 8.10 - HES-ECC six cubic feet (0.17 cubic meters) mixing.

388

30
0>
4-1

25 r

re

(0

c
o _

20

e. 15 h
0)

10

0>

<l)

Ol

O)

2
>
<

4)

>
<

20

40

60

Time,t(min)

Figure 8.11 - Loss of deformability of HES-ECC processed in six cubic feet (0.17 cubic
meters) batching.

10

100

1000

Age (hour)

Figure 8.12 - Compressive strength development of HES-ECC from 6 cubic feet (0.17
cubic meter) batching.

389

10

100

1000

Age (hour)
Tensile strain capacity development of HES-ECC from 6 cubic feet (0.17
cubic meters) batching.

304 mm
(12 in.)

76 mm
(3 in.)

"'

13 mm
(0.5 in.)

Figure 8.14 - Uniaxial tensile test plate dimensions.

390

_Jr*

>

fr. - 1 ,

*H*

ui""*

Figure 8.15 - Ellsworth Road Bridge over US-23 (S07 of 81074).

Figure 8.16 - Bridge deck crackim

Figure 8.17 - Partial depth removal.

391

,,!,

, g*

#.

>C1

'

V.
Figure 8.18 - Sandblasting of patch area.
"i

'-.

t >4
-*-.";

<

J *-*t.*

Figure 8.19 - Replacement of damaged reinforcement.

Figure 8.20 - Completed patch repair preparation.

392

Figure 8.21 - Separated patch areas for HES-ECC repair and Thoroc 10-60 repair.

*i.\

Figure 8.22 - 12 cubic foot (0.34 cubic meter) capacity gas mixer.

tam

^m

..

S^fcii.

; WM 3K
Figure 8.23 - Pouring HES-ECC into wheelbarrow for transportation.

393

<

Figure 8.24 - HES-ECC was poured into patch.

Figure 8.25 - HES-ECC shows self-compacting property.


(Flowability test slump diameter = 25.6 in. (650 mm))

Figure 8.26 - Placement of Thoroc 10-60.

394

Figure 8.27 - Comparison of HES-ECC patch repair and Thoroc 10-60 patch repair.

Figure 8.28 - Completed patch repair work.

395

-A-HES-ECC
-e-Thoroc 10-60

200

400

600

800

1000

1200

Age (days)

Figure 8.29 - Maximum surface crack width of HES-ECC and Thoroc 10-60 patch
repairs.
2400

I
J

2200

-A-HES-ECC

2000

-B-Thoroc 10-60

1800
1600
1400
1200

t
01

1000
800
600
400
200
0
200

600

400

800

1000

1200

Age (days)

Figure 8.30 - Maximum interfacial delamination width of HES-ECC and Thoroc 10-60
patch repairs.

396

HES-ECC
Patch

in
v

70 nm, with
possible healing
products

HES-ECC
Patch *

/
70 Jim. with possible
healing products

HES-ECC
Patch
90 urn, with nossible healing,
products

Figure 8.31 - Surface cracking in HES-ECC patch repair, on September 18, 2009.

397

mmmsmst
Thoroc 10^-60
'Pateh ., .

is
200-500-p.tn "

"

?*#
**"tpf'

v. .,*$(
'Patch . * , '

'

'"
(.

300-700 urn,

'V-ir""'

.;&

Thbrp^lb : 60
Patch, . f: ,\*

-.

200-1200 p.m,
pii<yaspalling

Figure 8.32 - Surface cracking in Thoroc 10-60 patch repair, on September 18,2009.

398

HES-ECC Patch
interface

Existing Cbncrcte

Figure 8.33 - Interfacial delamination between HES-ECC patch repair and existing
concrete, on September 18, 2009.

Existing Concrete
Interface"

Thoroc 10-60 Patch

Figure 8.34 - Interfacial delamination between Thoroc 10-60 patch repair and existing
concrete, on September 18, 2009.

399

l:\ntMrface

Thoroc 10-60
Patch

:KES-EGC

--*

*.

8fc?";:

Figure 8.35 - Exposure of aggregates in Thoroc 10-60 patch repair close to the Thoroc
10-60/HES-ECC interface area, on September 18, 2009.

400

References:
1

Kanda, T et al, "Tensile and Anti-Spalling Properties of Direct Sprayed ECC," Journal
of Advanced Concrete Technology, Vol. 1, No. 3, 2003, pp. 269-282.
2

Li, V. C , Lepech, M. and Li, M., "Final Report on Field Demonstration of Durable
Link Slab for Jointless Bridge Decks Based on Strain-Hardening Cementitious
Composites," Michigan Department of Transportation, 2005.
3

Lepech, M.D., and Li, V. C , "Large Scale Processing of Engineered Cementitious


Composites," ACI Materials Journal, Vol. 105, No. 4, 2008, pp. 358-366.
4

Kong, J.H., Bike, S. and Li, V.C., "Development of a Self-Consolidating Engineered


Cementitious Composite Employing Electrosteric Dispersion/Stabilization," Journal of
Cement and Concrete Composites, Vol. 25, No. 3, 2003, pp. 301-309.
5

Kanda, T. and V.C. Li, "Effect of Apparent Fiber Strength and Fiber-Matrix Interface
on Crack Bridging in Cement Composites," ASCEJ. of Engineering Mechanics, Vol. 125,
No. 3, 1999, pp. 290-299.
6

Li, V. C , Li, M. and Lepech, M., "High Performance Material for Rapid Durable Repair
of Bridges and Structure," Michigan Department of Transportation, December 2006.
7

Sprinkel, M., "Evaluation of Latex-Modified and Silica Concrete Overlays Placed on


Six Bridges in Virginia," Virginia DOT Final Report, 2000. (a)

Sprinkel, M., "Evaluation of High Performance Concrete Overlays Placed on Route 60


Over Lynnhaven Inlet in Virginia," Virginia DOT Final Report, 2000. (b)
9

Li, V.C. and M. Li, "Durability Performance of Ductile Concrete Structures," Proc. 8th
Int'l Conf. on Creep, Shrinkage and Durability of Concrete and. Concrete Structures, IseShima, Japan, Eds. Tanabe et al, Sept-Oct, 2008, pp. 761-768.

401

CHAPTER 9

Concluding Remarks

9.1

Research Overview and Findings


This dissertation focuses on the development and demonstration of an integrated

multi-scale design framework for durable repair of concrete structures.

This is

accomplished through the establishment of links between the development, assessment,


and implementation of an innovative material technology that fundamentally tackles
major concrete deterioration problems. Within each of the design scales, the tailoring of
ECC repair materials is highly focused on meeting specific structural performance
requirements while adhering to additional system constraints (i.e. cost, processing,
material availability, environmental exposure, existing cracks, etc.).

A variety of

micromechanics based tools, along with carefully designed experiments and finite
element methods, are employed to understand and optimize ECC material development
for specific repair applications. Furthermore, a comprehensive understanding of ECC
material mechanical properties, durability, and its interaction with existing concrete under
various environmental exposure and mechanical loading conditions is achieved in this

402

dissertation. In addition to laboratory development and experimental studies, the newly


developed ductile repair technology is transferred to field implementation by optimizing
workability and processing techniques, accommodating traditional large-scale mixing
equipment, and demonstrating in a bridge repair project in Southern Michigan.
Research activities have been divided into individual chapters depending upon
their placement within the overall integrated multi-scale design framework, namely
material engineering (Chapters 2 and 3), repair system durability assessment (Chapters 4,
5, 6 and 7), and field application (Chapter 8). Significant findings and accomplishments
made in individual research tasks are also summarized in the following.

9.1.1

Material Engineering for Durable Repair Materials

HES-ECC Material Development and Characterization


Material engineering for durable repair materials is possible through the
application of micromechanical tools developed for ECC materials.

This has been

demonstrated through development of the ductile repair material High Early Strength
ECC (HES-ECC) in Chapter 2. HES-ECC is designed for fast and durable repair of
transportation infrastructure (e.g. bridges, roadways and highways), and other types of
concrete structures that prefer a minimal interruption of operations and are often exposed
to challenging environmental and mechanical loading conditions. Micromechanics can
be used as a tool for designing HES-ECC composites that meet the two strain-hardening
criteria as well as conditions for saturated multiple cracking, to achieve large tensile
ductility above 3% and self-controlled crack widths below 100 \xm at both early and late
ages. These properties are crucial for repair durability, as they prevent cracking and

403

penetration of aggressive agents resulting from restrained volume change or stress


concentration induced by pre-existing substrate cracks; cracking and aggressive agent
penetration are typical deterioration causes in concrete repairs.
Through strategic combination of type III Portland cement, low water/cement
ratio, polycarboxylate based superplasticizer and calcium nitrate based accelerator, and
deliberate introduction of a small volume fraction (e.g. 5%) of PS beads, the newly
developed HES-ECC material is capable of attaining compressive strength of 23.6 1.4
MPa (3422 203 psi) in 4 hours and 55.6 2.2 MPa (8062 315 psi) in 28 days, which
generally exceeds federal and state transportation agency requirements.

HES-ECC

attains flexural strength of 9.8 0.2 MPa (1422 34 psi) in 4 hours and 15 0.3 MPa
(2187 50 psi) in 28 days, which exceeds twice the flexural strength of concrete with
similar compressive strength. Its tensile strength is 3.5 0.1 MPa (501 12 psi) at the
age of 4 hours and 5.7 0.2 MPa (823 20 psi) at age of 28 days. The PS beads act as
artificial flaws with controlled size and are included for maintaining tensile ductility of
HES-ECC at all levels of maturity.
Detailed characterization of mechanical properties of HES-ECC further validated
the material performance, and provided a database of the tensile, compressive, and
flexural properties of the newly developed repair material for future engineering
applications.

ECC Processing and Quality Control Optimized Material Properties and Robustness
ECC laboratory testing and application in construction projects will rely on the
ability to consistently produce ECC materials with controlled quality. Chapter 3 focuses

404

on maximizing ECC material properties (e.g. tensile strain capacity, tensile strength) and
minimizing material variation through a standardized processing technique. To achieve
this, a simple and practical quality control method for ECC processing is developed in
this chapter.
Through measurement of ECC mortar rheological parameters, fiber dispersion,
and ECC tensile properties, correlations between these three quantities are established,
which serve as a basis for identifying the optimal range of fresh properties for maximal
hardened properties. During the rheology measurements, a strong correlation, with R =
0.95, is found between the Marsh cone flow time and plastic viscosity measured using a
Viskomat-NT rotational viscometer, showing that the simple Marsh cone flow test should
serve as a reliable indirect method for onsite or laboratory viscosity measurement. It is
also found that incorporation of VMA can be an effective method to control plastic
viscosity of ECC mortar without sacrificing ECC flowability after adding fibers.
The fluorescence imaging technique is a useful tool for quantifying the
distribution of short discontinuous PVA fibers within a cementitious matrix. The fiber
dispersion coefficient for seven mixes with the same mix design but increasing VMA
content increases from 0.29 to 0.89. This quantifies the effectiveness of adding VMA as
a new rheology control method for achieving more uniform fiber dispersion in hardened
ECC, which has not been previously reported by other researchers.
Fiber uniformity, as measured by the fiber dispersion coefficient, has a strong
effect on ECC tensile strain capacity. A lower fiber dispersion coefficient not only
reduces ECC's tensile strain capacity and ultimate tensile strength, but also increases the
variation in both properties. A very low value for the fiber dispersion coefficient can

405

switch ECC from a strain-hardening material to a tension-softening'material.

ECC

mortar plastic viscosity is established experimentally as a fundamental rheological


parameter that affects dispersion of PVA fibers in ECC mixes, and consequently affects
ECC tensile strain capacity. An ECC mortar with low plastic viscosity and Marsh cone
flow time tends to have poorly distributed fibers. Increasing plastic viscosity of ECC
mortar improves fiber dispersion, and the improvement reaches a plateau once the Marsh
cone flow time reaches 24 s. Further increasing viscosity can potentially lead to a
reduction in ECC first cracking strength and ultimate tensile strength due to more largersize entrapped air pores.

Therefore, an optimal range of plastic viscosity (and

corresponding optimal Marsh cone flow time) is revealed, which can be used to guide
ECC rheology control during processing before fibers are added.
The present study results in a practical methodology to control the quality of ECC
during laboratory or onsite large-scale processing for maximized material tensile
properties and robustness, by simple use of a portable Marsh cone and adjustment of
VMA content to result in a flow time between 24 s and 33 s. The effectiveness of this
method was further confirmed using a special version of WHITE-ECC.
In summary of Chapter 2 and 3, for ECC to achieve robust tensile strain
hardening behavior with designed tensile ductility, micromechanics-based material
design should be combined with controlled material processing - neither of the two
should be ignored. This is because material processing strongly affects the composite
microstructure, which in turn determines whether the strain-hardening criteria are
satisfied in the produced material as originally designed.

406

9.1.2

Durability Assessment of ECC/Concrete Repair System


Concrete repair failure results from a combination of physical, chemical and

mechanical processes.

Generally the repair deterioration process starts with repair

cracking and repair/old interfacial delamination due to restrained volume change, or from
reflective cracking due to stress concentration initiated in pre-existing cracks in substrate
concrete. Cracking is an insidious cause of many repair durability problems, because it
dramatically alters the transport properties of the repair material, facilitates the ingress of
chloride into the repaired system, and eventually causes premature deterioration and
repair failure such as corrosion and disintegration. At the repair system durability
assessment level (middle triangle in Figure 1.10), the link between ECC repair material
properties and repaired system durability is explored, in terms of the three major repair
deterioration mechanisms: cracking and interfacial delamination due to restrained volume
change (Chapter 4), reflective cracking (Chapter 5), and chloride penetration (Chapter 6).

Influence of Material Ductility on Restrained Volume Change Induced Cracking and


Interfacial Delamination
Chapter 4 addresses one of the major durability concerns in concrete repairs, i.e.
repair cracking and repair/old interfacial delamination due to restrained volume change
and lack of dimensional compatibility. Through experimental, numerical and analytical
studies of a simulated repair system under restrained shrinkage conditions, this chapter
establishes conclusively that simultaneous suppression of repair cracking and interfacial
delamination can be achieved using ductile ECC repair materials.
Specifically, the free drying shrinkage properties of HES-ECC (large compared

407

with normal concrete but below 0.3%) is found to be one order of magnitude lower than
its tensile strain capacity of 3-6%. This leads to a highly negative cracking potential
value, implying that when the drying shrinkage of HES-ECC is restrained, the material's
ductility can accommodate shrinkage deformation by forming multiple microcracks
during its strain-hardening stage without localized cracking failure. This contention is
proven later by the results of the restrained shrinkage ring test and the restrained
shrinkage test on a simulated repair system.
It is found that repair material ductility has a strong influence on the performance
of the repaired system under restrained volume change. For a brittle repair material such
as concrete or HES-Concrete, several localized surface cracks form with large width up
to 520 [Am, accompanied by a small amount of interfacial delamination up to 90 um. A
tension-softening repair material such as SFRC or HES-SFRC, which includes short steel
fibers, is not able to suppress localized repair cracking with large crack width, although
the presence of steel fibers does help to reduce the crack width to some extent.
Meanwhile, the bridged cracks in an FRC greatly promote the tendency of interfacial
delamination, which is the largest among the three categories of materials. This is
because HES-SFRC has a positive cracking potential value with no inelastic strain
capacity, so that the restrained shrinkage induced tensile and shear stresses at the
interface have no avenue of release through material inelastic straining as does HES-ECC,
or crack free-opening as does HES-Concrete. In contrast, a ductile repair material such
as ECC or HES-ECC exhibits more than 100 fine microcracks with width less than 50
\xm formed in the repair layer, accompanied by interface delamination less than 80 \xm at
the ends. The experimental results correspond well with the numerical simulation and a

408

simple analytical model developed in this study, concluding that the ductility of a repair
material is the most essential factor contributing to the resistance to repair cracking and
delamination.
This research departs from the traditional emphasis on isolated material properties
such as high compressive strength and low shrinkage, and intends to shift the concrete
repair industry's focus onto the age-dependent interaction between the new repair and old
concrete. To ensure dimensional compatibility between the repair and existing concrete,
a repair material with large inelastic tensile strain capacity (ductility) is necessary. When
the material ductility requirement is satisfied, cracking control reinforcements and repair
material free drying shrinkage limits will become less important, and surface preparation
methods to enhance the interface bond will be more meaningful toward achieving
durability of repaired concrete structures.
Although investigated in this study in the restrained shrinkage scenario, this
concept of translating the ductility of the repair material to the durability of the repair
system can be generalized to other restrained volume change (e.g. temperature effects)
situations.

Influence of Material Ductility on Reflective Cracking in Overlay Repairs


Chapter 5 addresses the reflective cracking issues that are prevalent in concrete
repairs - especially pavement overlay repairs. Reflective cracking is another major
cracking mechanism that greatly affects concrete repair service life. It is caused by the
brittle response of overlay material under high stress concentration induced by preexisting joints or cracks in the substrate concrete. This chapter concludes that reflective

409

cracking can be fundamentally prevented through the high tensile ductility and strainhardening behavior of ECC materials.
This was demonstrated in this study of an HES-ECC overlay system ,with an
existing vertical crack in the substrate concrete and a small amount of interfacial
delamination. Under monotonic and fatigue four-point bending, the HES-ECC overlay
exhibited a "kinking and trapping" mechanism, and multiple microcracking behavior
with controlled crack width below 50 urn (2><10"3 in.). Through this process a large
amount of energy was dissipated, thus preventing the brittle reflective cracking failure
mode that was observed in the control HES-Concrete overlay system.
Due to the ductile tensile behavior and unique "kinking and trapping" mechanism
of the HES-ECC overlay, it exhibited 100% higher load carrying capacity and several
orders of magnitude longer fatigue life compared to an HES-Concrete overlay with
similar compressive strength and the same age.

This suggests that significant

improvements are possible in overlay design when HES-ECC is used: assuming the
same loading conditions, (a) if designed with the same overlay thickness, the HES-ECC
overlay is expected to have a significantly longer service life than traditional concrete
overlays; (b) if designed with the same service life, the HES-ECC overlay thickness can
be greatly reduced compared to traditional concrete overlays, leading to a significant
amount of saving of materials, construction equipment, and construction time; or (c) a
compromise of either parameter, i.e. reduced overlay thickness as well as prolonged
overlay service life, can be achieved simultaneously.
It should be noted that the high resistance to reflective cracking of the HES-ECC
overlay through multiple microcracking and the "kinking and trapping" mechanism is

410

also observed at the early age of 6 hours. The load carrying capacity of the HES-ECC
overlay at the age of 6 hours and the control HES-Concrete overlay at the age of 28 days
is almost the same. Furthermore, at the same stress level, the fatigue life of the HESECC overlay at the age of 6 hours is approximately one order of magnitude larger than
that of the control HES-Concrete overlay at the same age. These results highlight the
significant advantages of the newly developed HES-ECC material in fast and durable
repair applications, with shortened operations shutdown time as well as greatly prolonged
service life and increased load carrying capacity. These potentials in HES-ECC have not
been demonstrated in other versions of ECC materials, or traditional concrete or FRC
materials.
This methodology of suppressing reflective cracking can be generalized as
translating the ductility of the repair material to the durability of the repair system, when
subjected to stress concentration from existing cracks.

Transport Properties of ECC under Chloride Exposure


Chapter 6 addresses the chloride penetration stage in a typical repair deterioration
process, before and after cracking occurs at different levels of strain. Chloride diffusion
driven by a chloride concentration gradient, which is the predominant mechanism of
chloride transport for most of the concrete structures exposed to deicing salts, and
airborne or waterborne chloride ions, is the focus of this chapter.
Significant improvements in the resistance to chloride penetration can be achieved
through the built-in microcracking mechanism of ECC materials. When subjected to
increasing deformation levels, ECC exhibited multiple microcracking behavior with

411

increased crack numbers but self-controlled crack width under 50 urn (0.002 in.). This
phenomenon is fundamentally different from traditional concrete repair materials, in
which localized cracks form and the crack width enlarges with increasing strain level.
Reinforcements can limit the width of such localized cracks to some degree, but will not
alter the dependency of crack width on strain level and structural dimensions. This is
partially illustrated in this study through testing on reinforced mortar specimens, which
represent a large category of repair materials. After ECC beams and reinforced mortar
beams are preloaded to various deformation levels, and subsequently exposed to 3%
NaCl ponding for 90 days, the effective diffusion coefficient of ECC is measured to be
significantly lower than that of reinforced mortar. Even at a twice-higher preloaded
deformation level, the effective diffusion coefficient of the investigated ECC beam is
around a quarter of that of the investigated reinforced mortar beam.
The significantly lower effective diffusion coefficient of ECC compared with
reinforced mortar is attributed to the difference in cracking pattern.

The effective

diffusion coefficient of ECC is found to be linearly proportional to the number of cracks,


whereas the effective diffusion coefficient of reinforced mortar is proportional to the
square of the crack width. Therefore, crack width in mortar has a more pronounced
effect on chloride transport than did the number of cracks in ECC, leading to a greatly
higher effective diffusion coefficient at the same deformation level.
Additionally, through the formation of microcracks less than 50 um (0.002 in.), a
significant amount of self-healing was observed within the microcracked ECC when
subjected to NaCl solution. This self-healing phenomenon aides in further reducing the
effective diffusion coefficient of microcracked ECC.

412

In the uncracked beam specimens, the effective chloride diffusion coefficient of


ECC is also lower than that of reinforced mortar for 30 and 90 days. This is explained by
the denser microstructure of ECC due to a higher amount of cementitious materials,
lower water/cementitious materials ratio, and high volume fly ash content.

The

immersion test results further prove this by comparing the chloride penetration depth in
uncracked ECC and reinforced mortar.
The results show that ECC is effective at slowing down the diffusion of chloride
ions under combined mechanical and environmental loading by virtue of its ability to
achieve a self-controlled tight crack width, even under large applied strain levels. This
study verifies the concept of improving repaired system durability with the inherent crack
self-controlling mechanism of ECC repair materials.

Effect of Cracking and Healing on the Durability of ECC under Combined Aggressive
Chloride Environment and Mechanical Loading
As the previous chapters (4, 5, and 6) concluded that ECC's large tensile ductility
and self-controlled crack width below 100 um are essential for achieving durable repair
performance, Chapter 7 investigates the durability of ECC material itself in terms of
maintaining its high ductility and tight crack width characteristics under combined
mechanical loading conditions and aggressive chloride environmental conditions.
This study reveals that ECC maintained its unique tensile ductility and multiple
microcracking behavior under combined mechanical loading conditions (0.5%, 1.0%,
1.5% tensile straining and reloading to failure) and aggressive chloride exposure (23-day,
60-day, and 90-day immersion in 3% NaCl solution). This indicates that even under

413

severe marine environment conditions, ECC remains durable and can provide reliable
tensile ductility to prevent common cracking failure due to restrained volume change or
stress concentration.

Additionally, it maintains self-controlled microcrack width no

larger than 100 [am during the material strain-hardening stage, and can thereby offer
reliable resistance to chloride penetration even in severe marine environment conditions.
Micromechanical scale experimental studies are carried out to understand the
observed composites behavior. After exposure to 3% NaCl solution for 30, 60 and 90
days, changes in ECC matrix fracture toughness, interfacial chemical bond, and
interfacial frictional bond are experimentally determined and compared with those of
ECC without exposure to NaCl. These changes are due to the presence of chloride ions,
which tend to increase the leaching of calcium hydroxide and porosity. The reduction in
fiber/matrix interfacial bonds explains the increase in average crack width of ECC from
45 urn (before chloride exposure) to 100 urn (after chloride exposure for 30, 60 and 90
days); the reduction in matrix fracture toughness explains the reduction in ultimate tensile
strength after chloride exposure for 30, 60 and 90 days. Interestingly, the simultaneous
decreases in fracture toughness and interfacial bonds result in an unchanged (or slightly
increased) PSH pseudo strain-hardening index (JbVJtiP), leading to a non-deteriorated
tensile strain capacity of ECC after 30-day, 60-day, and 90-day chloride exposures.
The innately tight crack width in ECC is favorable for the self-healing process to
take place. This is demonstrated in Chapter 6 in terms of recovery of chloride transport
properties, and then in Chapter 7 in terms of recovery of tensile behavior. Specimens
preloaded to 1.5% strain nearly completely recover their stiffness and tensile strain

414

capacity when reloaded in direct tensile tests, even after periods of 30, 60, and 90 days
exposure to NaCl solution.

9.1.3

Field Application of ECC Repair Materials


Chapter 8 shifts the focus from "repair system durability assessment" to the

"structural application" portion of the integrated engineering methodology (upper triangle


in Figure 1.10). Efforts are made to transfer innovative ECC repair technology from the
laboratory to field implementation through a bridge patch repair demonstration project.
Before the newly-developed HES-ECC material can be considered a viable
building material for use in repair or structural designs, it must be capable of processing
at a large scale using commercial batching equipment common within the repair
construction industry. In this chapter, large-scale processing and construction of an HESECC repair using commercial facilities are realized through optimization of the HESECC ingredients and mixing procedure, trial batches, and quality control methods.
Specifically, workability of HES-ECC is evaluated with scaled-up batching sizes.
Gravity mixers are used to process HES-ECC at batch sizes of 1, 3 and 6 cubic feet (0.03,
0.08, 0.17 cubic meters). Prior to the larger scale batching, preliminary tests in the
laboratory are carried out to improve flowability and workability of the material. The 0.5
in. (12.7 mm) long PVA fiber used previously in HES-ECC is changed to 0.33 in. (8.4
mm) long PVA fiber to promote easy mixing and better fiber distribution in the gravity
mixer, and to improve flowability of the material.

Additionally, the effect of the

hydration stabilizer on the flowability, initial setting time, and hardened mechanical

415

properties of HES-ECC is evaluated. Finally, the batching sequence of this material is


also optimized to facilitate mixing based on larger scale gravity mixers.
Based on the results from 1, 3 and 6 cubic feet (0.03, 0.08, and 0.17 cubic meter)
trial batches, it is concluded that larger scale mixing of HES-ECC material approaching
that of field conditions can be accomplished without difficulty.

Using the optimized

batching sequence, the overall mixing of HES-ECC material proceeds smoothly and
results in a fresh material that is homogeneous, flows well, and is Theologically stable.
Testing of hardened mechanical properties of HES-ECC processed in 1, 3 and 6 cubic
feet (0.03, 0.08, and 0.17 cubic meter) batching, shows that the early age and late age
compressive strength and tensile strain capacity are similar to those of mixes made in a
laboratory size Hobart mixer, and continue to meet all of the performance targets set forth
when developing HES-ECC at the laboratory scale.
A HES-ECC patch repair was constructed on the Ellsworth Road Bridge over US23 (S07 of 81074) on November 28, 2006. A 7 cubic foot (0.20 cubic meter) HES-ECC
batch was mixed by MDOT and UM research personnel using a 12 cubic foot (0.34 cubic
meter) capacity concrete gas mixer provided by MDOT.

The mixing time was

approximately 30 minutes. The mixed HES-ECC exhibited desirable creamy viscosity,


good fiber distribution, and self-compacting properties. The repair construction was
completed at around 1pm on Tuesday, November 28, 2006. Traffic reopened at 9am on
Wednesday, November 29, 2006.
The long-term durability of the HES-ECC patch repair under field conditions was
monitored until September 18, 2009. Under heavy traffic loading, shrinkage, temperature
effects, three winters' exposure to freezing and thawing as well as de-icing salts, and

416

other possible loading conditions in field, the HES-ECC patch repair exhibited multiple
microcracking behavior with crack width below 100 |im, and no sign of disintegration
and fracture failure.

Through this work, the linkages between material engineering,

repair system durability assessment, and structural application are further forged.

9.2

Impact of Research
Lack of durability in concrete infrastructure has become a looming threat that

could jeopardize the world's prosperity and our quality of life.

According to the

American Society of Civil Engineers (ASCE) 2009 Report Card1 for US Infrastructure,
an average grade of D (poor) was assigned over 12 infrastructure categories. ASCE
estimated that USD 2.2 trillion is needed over the next five years for repair and retrofit.
Correspondingly, repair and retrofit cost has been estimated to be USD 2 trillion for
Asia's infrastructure. In Europe, Japan, Korea, and Thailand, the annual cost for repair
has exceeded that for new construction2.
While countless materials and techniques are used in practice to meet the demand
for rapid, inexpensive and durable concrete repairs, few of them target the inherent
material shortfall of concrete as a brittle material that cracks and fractures in many
applications, leading to further deterioration. It has been estimated that almost half of all
concrete repairs fail prematurely3.

The challenges posed by deteriorating concrete

structures and widely-observed repair failures require an effective and practical approach
that breaks down the concrete deterioration process, which typically begins with
restrained volume change induced cracking or reflective cracking, followed by
penetration of aggressive agents, and eventually results in concrete spalling,

417

disintegration, and loss of structural capacity.

The current repair materials and

technologies have had limited success in preventing these deterioration mechanisms.


The main contribution of this dissertation is to develop and implement a novel
integrated multi-scale design framework for durable concrete repairs, which addresses the
aforementioned limitations of current repair technologies. This framework begins with
material microstructure tailoring at the micrometer scale, links to repaired system
durability assessment through composite material properties at the centimeter scale, and
ultimately relates construction processes and performance evaluation at the meter scale of
the repaired infrastructure. To achieve the holistic durability of such engineered systems,
durability concepts are interwoven across all material engineering, system durability
assessment, and structural application scales.
As opposed to traditional concrete and fiber reinforced concrete (FRC) materials
that are susceptible to cracking and fracture failure, a new class of innovative, ductile
ECC repair materials can be designed using micromechanics and processing tools offered
by material engineering technology to exhibit desired multifunctionality (i.e., possessing
more than one of the engineering functionalities) that addresses each typical deterioration
stage in concrete repairs. The main differences between the new and traditional concrete
repair materials are that the former possesses more than 300% larger tensile strain
capacity, inherent self-controlled crack width under 100 \xm without relying on steel
reinforcement, and great tailor-ability with additional functionalities (e.g. high early
strength) based on established micromechanics tools. These properties, as revealed and
validated in this dissertation, are essential for suppressing the three major concrete
deterioration mechanisms, namely (a) cracking and interfacial delamination due to

418

restrained volume change, (b) reflective cracking due to stress concentration from
existing cracks, and (c) chloride penetration that initiates embedded steel corrosion.
Without these properties, traditional repair materials and technologies can only delay, but
not fundamentally suppress, concrete deterioration.
The capability of consistently processing and producing robust ECC materials
plays a crucial role in its ascendancy as a new construction material in various structural
applications. To address this need, this dissertation develops a practical quality control
approach for laboratory or onsite ECC processing in order to maximize material
properties as designed, and minimize material variation. This approach is novel, in that it
is based on the systematically-built relationship between ECC material fresh rheological
properties, dispersion uniformity of short PVA fibers within the cementitious matrix, and
ECC material hardened tensile properties. As opposed to qualitative approaches that
have previously been used often in ECC processing, which involve visual and physical
(by hand) inspection of mix homogeneity and fiber dispersion uniformity, this approach
offers a quantitative and scientifically-based method that relates the marsh cone flow rate
to fiber dispersion coefficient. Furthermore, this dissertation brings to attention that nonuniform fiber dispersion within ECC can result in undesirable variation in its tensile
properties, and even switch ECC from a strain-hardening material to a tension-softening
normal fiber reinforced concrete (FRC) material. This can lead to material properties that
fall short of those originally designed for using the ECC micromechanical theory, which
assumes uniform fiber dispersion.
This dissertation

also provides a comprehensive understanding on the

deterioration mechanisms of concrete repairs under various environmental and

419

mechanical loading conditions.

Through identifying causes and effects of each

deterioration stage, the critical repair material properties that influence each stage are
determined. The influence of material properties on the interaction between the repair
and the existing concrete as well as the durability of the repair system is addressed.
Finally, the performance and practicality of this material technology is justified not only
in laboratory studies, but also in field full-scale application.
This dissertation departs from the traditional emphasis on isolated material
properties such as high compressive strength and low shrinkage, and intends to shift the
repair industry's attention toward the age-dependent interaction between the new repair
and old concrete. This concept of translating the ductility of the repair material to the
durability of the repair system, at different deterioration stages (i.e. restrained shrinkage
cracking, interfacial delamination, reflective cracking, chloride penetration, fatigue, loss
of durability in aggressive chloride environment) can be widely applied to concrete
structure repair applications for minimizing maintenance requirements and reducing
repair costs. The potential impacts of the integrated multi-scale design framework on
society's ability to preserve and extend service life of large volume of existing concrete
infrastructure systems, which represent less harm to future generations in terms of
economic, social and environmental impacts, are indeed tremendous.

9.3

Recommendations for Future Research


Different from new construction, successful concrete repairs require an accurate

understanding of the underlying concrete deterioration mechanisms and causes and


effects, which can be case-sensitive based on specific environmental, loading, and

420

boundary conditions. While the development of HES-ECC was successfully completed


for fast and durable repair applications, which cover a broad range of transportation
infrastructure and other types of structures that prefer minimal operations disruption,
other repair scenarios are still open to consideration. For example, repair of damaged
concrete members due to earthquake, fire, or impact loading may need additional repair
material properties to meet specific requirements, such as resistance to high strain-rate
loading, or non-deterioration in fiber and fiber/matrix interfacial properties under very
high temperature. Additionally, for repair of non-critical structures, while compressive
strength requirements are lower, ECC materials with mix designs incorporating large
volumes of waste materials (e.g. fly ash, slag, rice husk ash) may be preferable to mixes
with the lowest initial cost or highest material greenness. Finally, for regions where only
local ingredients are available, strategic incorporation of these local ingredients into ECC
materials, without sacrificing critical material properties that are desired for repair
durability, will be meaningful.
This dissertation investigated three major deterioration mechanisms in concrete
repairs: repair cracking and interfacial delamination due to restrained volume change,
overlay reflective cracking, and chloride penetration.

However, there are other

deterioration mechanisms that need to be better understood in the future, such as alkali
silica reaction, salt-scaling, concrete spalling and loss of load carrying capacity due to
steel corrosion, and accelerated corrosion in concrete repairs. Furthermore, combined
loading states consisting of mechanical and environmental loads, instead of single
deterioration mechanisms, need to be more comprehensively considered in future studies
of concrete repair durability and service life estimations.

421

Accelerated corrosion in concrete repairs due to electrochemical incompatibility


between "old" and "new" portions of the repaired concrete structures is commonly
present4. The effect of a repair on electrochemical activity in a repaired structure is a
function of the change in voltage potentials, the nature of the repair materials, and the
exterior and interior environments. If the steel in the repair area is partially embedded in
existing, chloride-contaminated concrete, and partially in new repair material, strong
corrosion cells can develop. The part of the steel bar in the existing concrete will become
anodic and corrode at a rapid rate, driven by the other part acting as a cathode. In this
case, repair phase deterioration and failure may develop in less than 1 year. Such an
accelerated corrosion process can be influenced by the cracking pattern of repair
materials. Chloride ions penetrate fastest through cracks, reach the embedded steel, and
turn these parts of steel into anodes. The number of cracks, and the chloride penetration
rate through cracks, will determine whether microcells or macrocells will form and the
rate of corrosion. Therefore, the effects of ECC multiple microcracking behavior, in
contrast with the localized cracks with larger crack width (higher chloride diffusion rate)
in traditional concrete repairs, on modifications of the accelerated corrosion process will
be of great research interest.
Broader impacts (i.e. economic, social, and environmental impacts) resulted from
increased durability of concrete structures containing alternative ECC repair materials,
compared with those containing traditional materials, need to be quantified through lifecycle modeling and analysis. The development and validation of service life models
comprise large opportunities for future research. For example, the measured effective
diffusion coefficient of ECC materials in this dissertation compared with traditional

422

reinforced mortar materials can be used for future service life prediction and life cycle
cost analysis of ECC repairs or structures. While the methods can focus on a single
deterioration mechanism resulting in the most common failure mode, concrete repairs can
fail in a variety of ways at any age. A probabilistic-based model for use in predicting the
service life of concrete repairs and repaired structures is needed. At this time, few of
these values have been incorporated in service life models or established for new
materials such as ECC.
The self-healing phenomenon under chloride exposure was revealed in this
dissertation in terms of recovery of both chloride transport properties and material tensile
properties. The self-healing mechanism under chloride exposure might be different from
other exposure conditions (e.g. immersion in water, wet-dry, high alkali environment),
and would be of research interest. The physical and chemical mechanisms behind the
self-healing phenomena need to be understood at the micro-scale level. For example, the
details of the ECC self-healing process and self-healing product need to be examined
through ESEM (Environmental Scanning Electron Microscope) and X-ray diffraction
technology. The mechanical properties of self-healing products and their influence on
ECC micro-parameters (i.e. matrix fracture toughness, fiber/matrix chemical bond,
frictional bond and slip-hardening coefficient) need to be quantified to guide ingredient
re-selection and material microstructure tailoring to further ensure robust self-healing. At
the macro-scale level, the repeatability of self-healing is still unknown.

Also, the

laboratory self-healing test was performed in the unloaded state. Given that structures in
the field are likely to be in the loaded state, self-healing behavior in the loaded state will
need to be evaluated and established.

In addition, ECC self-healing under field

423

conditions has never been demonstrated and needs to be further confirmed both in the
loaded and unloaded state.
More field demonstration projects and repair applications of ECC materials are
needed in the future to evaluate their performances under various field environmental and
mechanical loading conditions, repair dimension scales, and boundary conditions.
Further exploration of the links between material engineering, system durability
assessment, and structural application within the multi-scale design framework will be
greatly meaningful.

424

References:
1

Report Card for America's Infrastructure. 2009.


http://www.infrastructurereportcard.org/
2

Li, V. C , "Engineered Cementitious Composites", Proceedings of ConMat'05,


Vancouver, Canada, August 22-24, 2005, CD-documents/l-05/SS-GF-Ol_FP.pdf.
3

Vaysburd, A. M., Brown, C. D., Bissonnette, B, and Emmons, P. H., ""Realcrete"


versus "Labcrete"", Concrete International, Vol. 26, No.2, 2004, pp. 90-94.
4

Vaysburd, A. M., and Emmons, P. H., "How to Make Today's Repairs Durable for
Tomorrow - Corrosion Protection in Concrete Repair," Construction and Building
Materials, Vol. 14, No. 20, 2000, pp. 189-197.

425