Thermodynamics of Autogenous Self-healing in Cementitious Materials

(1)

(2)

Thermodynamics of Autogenous Self-healing in

Cementitious Materials

Proefschrift

ter verkrijging van de graad van doctor aan de Technische Universiteit Delft,

op gezag van de Rector Magnificus prof. ir. K.C.A.M. Luyben, voorzitter van het College voor Promoties,

in het openbaar te verdedigen op maandag 17 november 2014 om 15.00 uur

door

Haoliang HUANG

Master of Science, Wuhan University of Technology geboren te Guangdong, China

(3)

Dit proefschrift is goedgekeurd door de promotor: Prof. dr. ir. K. van Breugel

Copromotor: Dr. G. Ye

Samenstelling promotiecommissie:

Rector Magnificus, voorzitter

Prof. dr. ir. K. van Breugel, Technische Universiteit Delft, promotor Dr. G. Ye, Technische Universiteit Delft, copromotor Prof. ir. R.P.J. van Hees, Technische Universiteit Delft

Prof. dr. ir. D. Damidot, Ecole des Mines de Douai, Frankrijk Prof. dr. ir. N. De Belie, Universiteit Gent, België

Prof. dr. ir. C. Qian, Southeast University, China

Dr. L. Pel, Technische Universiteit Eindhoven Prof. dr. ir. H.E.J.G. Schlangen Technische Universiteit Delft, reservelid

ISBN: 978-94-6186-397-3

Keywords: Thermodynamics, autogenous self-healing, coupled transport-reaction model, nuclear magnetic resonance (NMR), cementitious materials

Printed by Haveka B.V.

Cover design: Haoliang Huang, Yinglin Cao and Yun Zhang

All rights reserved. No part of the material protected by this copyright notice may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying, recording or by any information storage and retrieval system, without written permission from the author.

(4)

(5)

(6)

The research work reported in this thesis was sponsored by Chinese Scholarship Council (CSC) and Delft University of Technology (TU Delft). I would like to express my appreciation to these institutes.

Many people have been supporting me in this research and they are recognized here. I would like to express my deep gratitude to my promotor Prof. Klaas van Breugel for giving me the opportunity to work in Microlab at TU Delft. Without his guidance, advice and support, I could not accomplish this work.

My greatest appreciation goes to my copromotor Dr. Guang Ye for his guidance, support, encouragement and patience throughout the whole course of my research. I am also very grateful to him for giving me many opportunities to be involved in different research activities and get to know experts in our field. I am thankful that the friendships with him and his family make my life abroad more colorful and warm.

I would like to express acknowledgement to Prof. Zhonghe Shui, Prof. Ji Wang and Prof. Wei Chen at Wuhan University of Technology, who recommended me as a PhD candidate at TU Delft and have been encouraging me to destine for a career as a scientific researcher.

All committee members of the defense for this thesis are greatly acknowledged for their constructive comments. Particular gratitude must go to Prof. Denis Damidot from France who taught me patiently how to use the geochemistry code CHESS. I am also thankful that he gave me a lot valuable suggestions on this thesis and relevant journal articles. I would like to express my great acknowledgement to Dr. Leo Pel from Eindhoven University of Technology for his help with the NMR experiments in this work. Discussions with him on the NMR experimental results are very helpful.

I am very thankful to Prof. Geert De Schutter for his valuable suggestions on my research work in the meetings at Ghent University. My thanks also go to Prof. Erik Schlangen for his suggestions and the discussion on my work.

Gerrit Nagtegaal, Arjan Thijssen and John van den Berg are highly appreciated for their help with my experimental work at TU Delft. Our secretaries Nynke Verhulst, Melanie Holtzapffel

(7)

They are greatly recognized as well.

I would like to gratitude Dr. Henk Jonkers, Dr. Oguzhan Copuroglu, Dr. Sanjay Pareek, Dr. Kris Sisomphon, Dr. Virginie Wiktor, Dr. Jian Zhou, Dr. Jie Hu, Dr. Liangyue Ji, Dr. Lupita Sierra Beltran, Dr. Qingliang Yu, Dr. Zhiwei Qian, Dr. Qi Zhang, Dr. Mingzhong Zhang, Dr. Ying Wang, Dr. Yuwei Ma, Zhengxian Yang, Branko Savija, Jure Zlopasa, Zhuqing Yu, Hua Dong, Dr. Yun Gao and Zhijun Tan for their suggestions and the discussions with them on my research work. I am very thankful to Mladena Lukovic who helps me to correct the English writing of this thesis, Rene Veerman for the translations of summary, Renee Mors for the translation of proposition, Yun Zhang and Yinglin Cao for their help with the design of the cover of this thesis, Natalie Carr for her suggestions on the propositions.

Appreciations also go to Dr. Gang Liu, Dr. Quantao Liu, Dr. Yue Xiao, Dr. Fangliang Xiao, Jingang Wang, Yong Zhang, Chunping Gu, Bei Wu, Hao Huang, Tianshi Lu, Jiayi Chen, Xiaowei Ouyang, Xu Ma, Zhipei Chen, Jitang Fan, Qingzhi Liu, Ning An, Xiaoyu Zhang, Haiqiang Wang, Huajie Shi and Lu Zhang for their supports and helps in my daily life during these years in the Netherlands.

I want to express my greatest appreciation to my parents for bringing me up and endlessly supporting me. I am also very thankful to my younger brother who has been taking care of our parents in these years.

Last but not least, I would like to express my deepest gratitude to my wife, Qiwen Deng, for her unconditional love, support and endless encouragements. Without her, I would never have had this colorful and enjoyable life in last 9 years.

Haoliang HUANG October, 2014

(8)

List of Symbols .………...………..………...v

List of Abbreviations ………....viii

Chapter 1 General Introduction ...1

1.1 Research background ...1

1.2 Aim and objectives of this research ...4

1.3 Scope of this research ...4

1.4 Strategy of this research ...5

1.5 Outline of this thesis ...6

Chapter 2 Literature Review ...9

2.1 Introduction ...9

2.2 Mechanisms of self-healing ...9

2.2.1 Self-healing based on adhesive agents ...10

2.2.2 Self-healing based on bacteria ...11

2.2.3 Self-healing based on mineral admixtures ...13

2.2.4 Autogenous self-healing ...14

2.2.5 Challenges about autogenous self-healing ...16

2.3 Methods to supply healing agents to cracks ...17

2.3.1 Encapsulation technique ...17

2.3.2 Technique of vascular systems ...18

2.3.3 Challenges concerning the supply of water or solution to the cracks in concrete structures exposed to the atmosphere ...19

2.4 Conclusions ...20

Chapter 3 Characterization of the Reaction Products of Autogenous Self-healing ...21

3.2 Materials and experiments ...22

3.2.1 Materials and sample preparations ...22

3.2.2 Experimental techniques for characterization of reaction products formed in the self-healing process ...24

3.3 Results and discussion ...25

3.3.1 Self-healing of Portland cement paste cured in water ...25

3.3.2 Self-healing of Portland cement paste cured in saturated Ca(OH)2 solution ...30

(9)

3.4 Discussion on the mechanisms of self-healing ...41

3.4.3 Self-healing of slag cement paste cured in saturated Ca(OH)2 solution ...44

Chapter 4 Quantification of the Kinetics of Autogenous Self-healing ...47

4.2 Experimental techniques to evaluate the efficiency of self-healing ...48

4.2.1 X-ray computed tomography ...48

4.2.2 Permeability tests ...49

4.2.3 Ultrasonic pulse transmission tests ...50

4.2.4 Strength tests ...50

4.2.5 Methods used in this research ...51

4.3 Materials and experiments ...51

4.3.1 Materials and sample preparation ...51

4.3.2 Environmental conditions for self-healing ...53

4.3.3 Experiments ...54

4.4 Results and discussion ...55

4.4.3 Self-healing of slag cement paste cured in saturated Ca(OH)2 solution ...66

Chapter 5 Simulation of Autogenous Self-healing by a Coupled Transport-Reaction Model ...71

5.2 State-and-art of simulation on autogenous self-healing in cementitious materials ...72

5.3 Coupled transport-reaction model on self-healing ...72

5.3.1 Computational algorithm ...73

5.3.2 Dissolution of the reactive material present at the crack surfaces ...76

5.3.3 Ion diffusion ...80

5.3.4 Thermodynamic modeling of chemical reactions in the crack ...84

5.4 Model parameters ...86

5.4.1 Sizes of paste sample and pixels and the length of time step for modeling self-healing ...86

5.4.2 Dissolution rate of the reactive material ...87

5.4.3 Ion diffusion coefficients ...90

5.4.4 Chemical reactions and equilibrium constants ...91

5.5 Validation of coupled transport-reaction model ...93

5.5.1 Validation of coupled transport-reaction model for simulating self-healing in Portland cement paste ...93

5.5.2 Validation of coupled transport-reaction model for simulating self-healing in slag cement paste ...102

5.5.3 Summary of the validation of the coupled transport-reaction model ...106

5.6 Prediction of self-healing of real cracks in slag cement paste ...107

(10)

5.6.2 Effect of initial crack width on self-healing ...110

5.7 Effect of carbonation on self-healing ...111

5.7.1 Effect of carbonation on self-healing in Portland cement paste ...111

5.7.2 Effect of carbonation on self-healing in slag cement paste ...112

Chapter 6 Effect of Absorption of Water by the Bulk Paste on Autogenous Self- healing ...115

6.2 Materials and experimental methods ...116

6.2.1 Materials and specimen preparation ...116

6.2.2 Nuclear magnetic resonance (NMR) ...116

6.2.3 Test procedures ...120

6.3 Experimental results and discussion ...121

6.3.1 NMR tests on water migration from cracks into bulk paste ...121

6.3.2 Further hydration of unhydrated cement particles promoted by additional water ...123

6.3.3 Quantification of further hydration of unhydrated cement particles promoted by additional water ...126

6.4 Simulation of additional hydration in the bulk paste promoted by additional water .128 6.4.1 Additional volume of hydration products caused by extra water ...128

6.4.2 Evolution of microstructure caused by supply of additional water ...134

6.5 Influence of absorption of water by the bulk paste on self-healing ...136

6.5.1 Analyzed case of cement paste made of CEM І 42.5N with w/c ratio of 0.3 ....136

6.5.2 Discussion for general situation ...137

Chapter 7 Retrospection, Conclusions and Prospects ...139

7.1 Retrospection ...139

7.3 Contributions to science and engineering ...142

7.4 Prospects ...143 References………...145 Summary………...157 Samenvatting .……….161 List of Publications ……….….165 Curriculum Vitae ……….169

(11)

(12)

List of Symbols

Roman lower case letters

f Filling fraction of a crack [-]

kp Apparent coefficient of permeability [m2]

e The self-healing efficiency [%]

fn A normalized value of filling fraction corresponding to the

uniform width of 10 µm

[-]

wc Crack width [µm]

ci The concentration of the ith ion [mol/m3]

ci,crack the concentration of the ith ion in the solution in the crack [mol/m3]

l The number of independent chemical reactions [-] be,i The number of atoms of element e in the chemical formula of

the ith species before reactions

[-]

be,j The number of atoms of element e in j th species after

reactions

[-]

vki The coefficient for the ith species in the kth chemical reaction [-] mc,t The total mass of CO2 needed for complete carbonation of the

solids formed in the crack

[g]

mc The mass of CO2 consumed by the carbonation of the solids at

a certain stage

[g]

fr Frequency of the alternating field [MHz]

Roman capital letters

K Equilibrium constants of chemical reactions [-]

Ahp The area of reaction products formed in a crack on BSE image [µm2] Ac,empty The area of the crack that is still not yet filled with reaction [µm2]

(13)

products on BSE image

L Thickness of the sample for air permeability tests [m] A cross-sectional area of the sample for air permeability tests [m2] Qi the measured gas flow in an air permeability test [m3/s]

Pi inlet pressure in an air permeability test [Pa]

Patm atmospheric pressure [Pa]

Acr,rm Area of reactive materials at the crack surfaces [m2]

Acr,se Area of “seed” (non-reactive material) at the crack surfaces [m2]

Acr Area of crack surfaces [m2]

R(t) The dissolution rate of the reactive material as a function of

time

[m/s]

R0 The basic dissolution rate of cement clinker or slag [m/s]

V(t) The volume of the reactive material that has dissolved [m3]

Di The diffusion coefficient in the solution in the crack for the ith

ion.

[m2/s]

DCSH The ion diffusion coefficient of C-S-H [m2/s]

Dpore The ion diffusion coefficient in pore solution [m2/s]

N The number of chemical species [-]

Zi The valence (including sign) of the ith species at the initial

stage

[-]

B0 The magnitude of the externally applied static magnetic field [MHz/T]

Greek letters

µ The coefficient of viscosity [Pa·s]

P

δ The penetration depth of the reactive material [m]

out

δ The thickness of outer products [m]

δ The total thickness of the layer of reaction products covering the reactive materials

[m]

1

ε The parameter used to describe the “induction period” of the reaction as affected by the large amount of water or solution in the crack

[-]

2

ε The parameter used to described the acceleration of the reaction due to the seeding effect of non-reactive material at the crack surfaces [-] tr δ Transition thickness [m]

( )

t i

θ The mole number of the ith ions [mol]

( )

t

inp i,

(14)

( )

t

crp i,

θ The mole number of the reaction products formed in the crack [mol] σ The coefficient of the ion exchange between the solution in the

crack and pore solution in the bulk paste

[m2/s]

,0

P

δ The penetration depth of the reactive material before self-healing

[m]

,

P a

δ The additional penetration depth reached during self-healing [m] ,

P tr

δ The transition thickness describing the change from a phase-boundary reaction to a diffusion-controlled reaction at flat surfaces

[m]

i

β The activity coefficient of the ith species [-]

ϕ Capillary porosity [-]

car

α The degree of carbonation [%]

γ The gyromagnetic ratio of the nuclei [MHz/T]

(15)

C-S-H Calcium silicate hydrates

UHC Unhydrated cement

ESEM Environmental scanning electron microscope

BSE Back-scattered electron

EDS Energy dispersive spectroscopy

FTIR Fourier transform infrared spectroscopy

TGA Thermogravimetric analysis

XRD X-ray diffraction analysis

NMR Nuclear magnetic resonance

MMA Methylmethacrylate

PMMA Poly methylmethacrylate

LWA Lightweight aggregates

BFS Blast furnace slag

w/c Water to cement ratio

DTG Differential thermogravimetric

CLP Crystal-like product

GLP Gel-like product

CH Portlandite

Hc Hemicarboaluminate Ca4Al2(CO3)0.5(OH)13·5.5H2O Mc Monocarboaluminate Ca4Al2(CO3)(OH)12·5H2O C3AH6 Tricalcium aluminate hydrate Ca3Al2(OH)12

CT-scanning X-ray computed tomography

ECC Engineered cementitious composites

SAP Super-absorption polymers

LVDT Linear variable differential transducers

CW Crack width

NBO/T The ratio of the Non-Bridging Oxygen Atoms to Tetrahedrally Coordinated Atoms

(16)

C3S Tricalcium silicat 3CaO·SiO2

C2S Dicalcium silicate 2CaO·SiO2

C3A Tricalcium aluminate 3CaO·Al2O3

C4AF Calcium ferroaluminate 4CaO·Al2O3·Fe2O3

(17)

(18)

1

General Introduction

1.1 Research background

It is well known that concrete is a brittle composite cementitious material that easily fractures under tensile loading. For this reason, reinforcement is installed in order to carry the tensile cross sectional forces after cracking. From this point of view, reinforced concrete is always designed to allow the occurrence of cracks.

It must be emphasized that cracks as such are not regarded as failure of reinforced concrete [1]. However, the existence of cracks can decrease the durability of reinforced concrete structures. In reinforced concrete structures exposed to the atmosphere, cracks provide preferential access for aggressive agents, such as chloride, sulphate and moisture (see Figure 1.1). In general, when the relative humidity in the atmosphere is higher than 75%, the sections of reinforcing bars, which are exposed to the atmosphere because of the cracks (as shown in Figure 1.1), may start to corrode [2]. Meanwhile, because of the cracks, the carbonation of concrete, sulphate attack and alkali-silicate reaction can take place deep inside the matrix. Both the corrosion of reinforcing bars and the degradation of concrete can shorten the service life of reinforced concrete structures.

For this service life problem caused by cracks, man-made repair is a common solution [1]. Although man-made repair can prolong the service life of reinforced concrete structures, it has several drawbacks. For instance, even though the quality of repairs has substantially increased in recent years, it is also known that realizing durable repairs is difficult. Most of these repairs can only last for ten to fifteen years [1]. Furthermore, it is difficult to repair cracks which are not accessible, such as the cracks in underground concrete structures [3]. Apart from these technical aspects, the costs of man-made repairs are usually very high [4]. Moreover, if structures, like bridges and tunnels, have to be taken out of service for repair, the indirect costs are generally several times higher than the direct costs [1]. This is a high financial burden for the society.

(19)

In comparison, self-healing of cracks in concrete structures could be beneficial [3]. As a novel idea, self-healing of cracks has attracted much attention in recent years. Reinforced concrete structures should be made smart enough to detect their own damage and repair themselves [5]. These self-healing concrete structures have significant potential to extend their service life and reduce their economic, social and environmental costs [5]. From this point of view, the whole society would benefit from self-healing concrete structures. In particular, owners would be interested in concrete structures with self-healing capacity and long service life [1]. Meanwhile, construction companies might also be interested in such products, which can dramatically reduce the potential maintenance cost and thus enable them to take a competitive position in the market [1].

Autogenous self-healing is one of the main mechanisms of self-healing in cementitious materials. The phenomenon of autogenous self-healing was noticed many years ago. According to Hearn [6], autogenous self-healing had already been observed in water retaining structures, culverts and pipes by Hyde [7] by the end of nineteenth century. In 1920s, a more systematical analysis of autogenous self-healing was executed by Glanville [8]. After that, autogenous self-healing of cracks in bridges was also investigated [9, 10]. Although the autogenous self-healing phenomena were noticed many years ago and have stimulated many investigations, the physico-chemical process of autogenous self-healing is not completely understood yet and the potential of autogenous self-healing in cementitious materials is not clear either. Atmosphere Reinforcement steel bar Crack Concrete matrix

Moisture combining with Cl-_{and SO}

4

2-The section of steel bar exposed to atmosphere

Crack mouth

(20)

By now, the results of studies about the composition of reaction products formed in autogenous self-healing processes are not consistent. Jacobsen and Sellevold [11] found some newly formed portlandite and ettringite in cracks in high strength concrete under freeze/thaw cycles. Schlangen and Ter Heide [12] detected newly formed C-S-H in cracks after cracked samples were cured in water. As shown in Figure 1.2, when the water/cement ratio of Portland cement paste is 0.3, about 20% cement clinker remains unhydrated at the age of 3 years and hydration almost stops. According to their findings, it could be concluded that autogenous self-healing was caused by further hydration of unhydrated cement clinker. Edvardsen [13] found calcium carbonate (CaCO3) in cracks after autogenous self-healing. Obviously the

precipitation of CaCO3 can also be a mechanism of autogenous self-healing. The investigations

by Yang et al. [14] and Qian et al. [15] also support this opinion. The different findings about the reaction products formed in cracks indicate that the physico-chemical process of autogenous self-healing is not completely understood. Further investigations on the physico-chemical process of autogenous self-healing are still necessary.

The potential of autogenous self-healing in cementitious materials is not completely clear either. It has been reported that the higher fraction of reactive materials, i.e., unhydrated cement clinker, slag and fly ash, the higher the potential of autogenous self-healing [16]. However, there is still a lack of quantitative information about the relationship between the potential of autogenous self-healing and the amount of reactive materials. Furthermore, the influence of various factors on autogenous self-healing, such as the initial ion concentrations of the solution in cracks and the crack width, are insufficiently studied. The study of the effect of water migration (from cracks into the bulk paste) on autogenous self-healing is completely absent. All these factors affect the potential of autogenous self-healing and should be investigated further.

Figure 1.2 Percentage of unhydrated cement (UHC) left in Portland cement paste with the w/c of 0.3, calculated by HYMOSTRUC [17-19].

(21)

1.2 Aim and objectives of this research

The aim of this research is to better understand the physico-chemical process and to quantify the potential of autogenous self-healing in cementitious materials. In order to achieve this aim, the main objectives can be outlined as follows:

1. The mineralogy of the reaction products formed in cracks is still not clear. The first objective of this research is to determine the mineralogy of the reaction products formed in cracks and the corresponding fraction of each mineral.

2. By now, there is still a lack of information about the kinetics of autogenous self-healing, i.e. the amount of reaction products formed in cracks as a function of time. The second objective of this research is, therefore, to quantify the amount of reaction products formed in cracks as a function of time.

3. So far there is no model that can be used to simulate autogenous self-healing explicitly considering the physico-chemical process. The third objective of this research is to develop a model for simulating autogenous self-healing and to predict the kinetics of autogenous self-healing influenced by several factors, such as the crack width, the amount of unreacted cement or blast furnace slag in the bulk matrix.

4. By now, there is no information about the effect of absorption of water (present in cracks) by the bulk matrix on autogenous self-healing. The fourth objective of this research is to investigate the effect of absorption of water by the bulk matrix on autogenous self-healing.

1.3 Scope of this research

This research only focuses on autogenous self-healing of cracks in cement-based materials. Therefore, the term “self-healing” in this thesis particularly refers to “autogenous self-healing”. Moreover, only autogenous self-healing in Portland cement paste and blast furnace slag cement paste is investigated in this research. “Slag” in this thesis refers to “blast furnace slag”. It is known that high performance concrete usually has a w/c ratio lower than 0.3. As mentioned before, when the w/c ratio of Portland cement paste is 0.3, about 20% cement clinker remains unhydrated at the age of 3 years and hydration almost stops (see Figure 1.2). In this research, the cement pastes with w/c ratio of 0.3 are investigated. The possible influence of w/c ratio on autogenous self-healing is briefly addressed. Only microcracks (about 10 µm) are investigated experimentally. Larger cracks (about 30 µm) are studied by modeling.

In this research, although the methods to supply healing agents to cracks for self-healing have not been studied, saturated Ca(OH)2 solution as a potential healing agent has been

investigated. In addition to water, a saturated Ca(OH)2 solution is used to heal Portland cement

paste. It is well known that slag can be activated by solutions with a pH higher than 12 [20]. To determine the potential of self-healing in slag cement paste, saturated Ca(OH)2 solution can

(22)

be used to activate slag maximally. In this research, slag cement paste is only cured in saturated Ca(OH)2 solutions for self-healing.

Because the diffusion coefficient of CO2 in water is 4 orders of magnitude smaller than that

in air [21], in this thesis carbonation is ignored when cracks are filled with water or a healing solution. However, when the crack is not filled with water or healing solution anymore (because of drying or the consumption of water by chemical reactions), the reaction products formed in cracks are exposed to the atmosphere. Carbonation of these reaction products could evolve much faster. In this thesis the carbonation of reaction products in cracks after the exposure to the atmosphere will be investigated only by modeling (see Chapter 5).

1.4 Strategy of this research

The strategy to achieve the aim of this research is described below:

- Firstly, the mineralogy of the reaction products formed in cracks is characterized and the fraction of each mineral in the reaction products is quantified experimentally. The physico-chemical process of autogenous self-healing can be analyzed based on the mineralogy of the reaction products formed in cracks.

- Secondly, the filling fraction of cracks as a function of time is quantified by means of back-scattered electron (BSE) image analysis. Meanwhile, the reduction of the air permeability through cracks due to autogenous self-healing is determined. As a result, the kinetics of autogenous self-healing is experimentally evaluated.

- After that, a coupled transport-reaction model is developed for simulating the autogenous self-healing process. The model is verified with experimental results. Consequently, the kinetics of autogenous self-healing influenced by different factors, such as the crack width, the amount of unhydrated cement or slag and the initial ion concentrations of the solution in cracks, can be predicted by the model.

- The migration of water from the crack into the bulk paste is tested by means of nuclear magnetic resonance (NMR) technique. The effect of water migration (from the crack into the bulk paste) on self-healing is investigated.

The coupled transport-reaction model developed in this research can be used to predict the kinetics of self-healing in cement paste made of different types of cement, i.e. types of cement with different fineness. This model can also be used to calculate the capacity of self-healing in cementitious materials with blended slag or fly ash. Moreover, for self-healing based on blended expansive agents, the coupled transport-reaction model can easily be extended for the calculation of healing capacity.

The study of the influence of water transport (from the crack to the bulk paste) on self-healing provides useful information for service life prediction of reinforced concrete structures, especially of underground concrete structures or concrete structures submersed in water.

(23)

1.5 Outline of this thesis

As shown in Figure 1.3, this dissertation includes 7 chapters. Chapter 2 presents a literature survey on self-healing in cementitious materials and specifies the knowledge gaps that have to be bridged by this research.

In Chapter 3 the reaction products formed in cracks due to autogenous self-healing are characterized by using environmental scanning electron microscope (ESEM) equipped with energy dispersive spectroscopy (EDS), Fourier transform infrared spectroscopy (FTIR), thermogravimetric analysis (TGA) and X-ray diffraction analysis (XRD). Through the characterization, not only the mineralogy of the reaction products is determined, but also the fraction of each mineral in the reaction products is quantified. With the mineralogy of the reaction products formed in cracks, the physico-chemical processes of autogenous self-healing are analyzed and hence understood better.

In Chapter 4, the kinetics of autogenous self-healing in Portland and slag cementitious materials is investigated. The amount of reaction products formed in cracks as a function of time is quantified by means of ESEM image analysis. Meanwhile, the reduction of the air permeability through cracks after autogenous self-healing is evaluated by performing air permeability tests.

A coupled transport-reaction model for simulating autogenous self-healing is developed in Chapter 5. This model is verified with the experimental results presented in Chapter 3 and 4. Moreover, the kinetics of autogenous self-healing influenced by several factors, such as the crack width, the amount of unhydrated cement or slag and the initial ion concentrations of the healing agents, is predicted by this model.

Chapter 6 focuses on the effect of the water migration (from cracks to the bulk paste) on self-healing. In order to do that, the migration of water from cracks into the bulk paste is monitored by using nuclear magnetic resonance (NMR). The kinetics of further hydration of unhydrated cement in the bulk paste nearby the crack is quantified. In addition, the amount of reaction products formed in cracks versus the amount of additional water penetrating into cracks is determined.

Chapter 7 summarizes the results in this research and presents the conclusions. Recommendations on enhancing the potential of autogenous self-healing in Portland and slag cementitious materials are given as well.

(24)

Chapter 1 Introduction

Chapter 2 Literature review

Chapter 3 Characterization of reaction products of autogenous

self-healing

Chapter 4 Quantification of the kinetics of autogenous

self-healing

Chapter 5 Simulation of autogenous self-healing by a

coupled transport-reaction model

Chapter 6 Effect of the migration of water on autogenous

self-healing

Chapter 7 Conclusions

(25)

(26)

2

Literature Review

2.1 Introduction

Self-healing of cracks has a significant potential to extend the service life of concrete structures and reduce economic, social and environmental costs [5]. Motivated by these advantages, self-healing in concrete has attracted much attention in recent years. As shown in Figure 2.1, journal publications related to self-healing in concrete increased dramatically in the last decade and this trend is likely to continue. Various aspects related to self-healing in concrete, such as mechanisms, various healing agents and evaluation methods etc., are being investigated worldwide.

This chapter gives an overview of the various mechanisms of self-healing and healing agents (in Section 2.2), followed by the methods to supply healing agents to cracks for self-healing in cementitious materials (in Section 2.3). By doing this, the knowledge gap that has to be bridged in this research is specified.

2.2 Mechanisms of self-healing

Based on the literature survey, self-healing of cracks in cementitious materials can be grouped into four categories:

1) self-healing based on adhesive agents, 2) self-healing based on bacteria,

3) self-healing based on mineral admixtures, 4) autogenous self-healing.

The physico-chemical healing process, influencing factors, advantages and disadvantages of these categories of self-healing are described in this section.

(27)

2.2.1 Self-healing based on adhesive agents

The hardening of adhesive agents in cracks has the potential to seal the cracks and connect both crack surfaces. These adhesive agents can be either an one-component or two-component agent, or even a multi-component agent. The hardening processes vary with the properties of the agent. Various adhesive agents applied for self-healing in recent studies are discussed below.

Epoxy is one of the adhesive agents often used for self-healing. In some investigations, a two-component epoxy was encapsulated and pre-embedded in concrete for self-healing [22-24]. When the capsules became intersected by cracks, both components of the epoxy were released and mixed with each other. These two mixed components then reacted and hardened to heal the cracks. However, the reaction can be negatively influenced as the ratio of these two components leaking into the cracks could not be controlled and the optimum mixing could not be achieved [3]. A one-component epoxy has been reported to be more effective for self-healing. For instance, Thao et al. [25] embedded one-component epoxy in concrete by using sealed glass tubes. After the glass tubes were broken by cracks, the one-component epoxy leaked into the cracks. Once this type of epoxy was exposed to air, it started to harden and finally repaired the cracks effectively. Nishiwaki et al. [26] also reported positive results of self-healing by using one-component epoxy.

Methylmethacrylate (MMA) is another type of adhesive agent with two or multi components that can be utilized for self-healing. Yang et al. [27] used microcapsules to supply methylmethacrylate and triethylborane (TEB) to cracks for self-healing. Dry and McMillan [28] even used three-component MMA for self-healing in concrete. In their study, in order to supply this multi-component MMA to cracks, a multi-channel vascular system was built inside concrete. The liquid components were delivered to cracks through this vascular system from the outside (More details can be seen in Section 2.3.2). Compared with the two-component epoxy, MMA has a lower viscosity and thereby the mixing of components is better. Nevertheless, because the viscosity of MMA is low, it can easily leak out of the cracks. This Figure 2.1 Archived journal publications on self-healing concrete over the last decade [5].

(28)

leaking may decrease the efficiency of self-healing. To overcome this problem, poly methylmethacrylate (PMMA) with higher viscosity was applied to replace MMA as a healing agent [3].

Also, cyanoacrylate has been employed as adhesive agent for self-healing. Cyanoacrylate has a low viscosity and is able to penetrate into the matrix. As a result it can connect the crack surfaces strongly [3, 29]. Furthermore, as a one-component adhesive agent, it can harden within a short time after exposure to air [3]. Therefore, cyanoacrylate is frequently used to develop self-healing concrete [30-36]. Apart from the aforementioned agents, other healing agents such as silicon [37] and tung oil [4] were also used.

In general, the efficiency of self-healing based on adhesive agents mainly depends on the type of adhesive agents. One-component agents are easier to operate with and their efficiency is higher than that of two or multi component agents [3].

In addition, one of the important requirements for this self-healing mechanism is the successful supply of adhesive agents to cracks. These liquid adhesive agents are generally supplied to cracks by using capsules or vascular systems. Details about the methods to supply liquid healing agent to cracks will be presented in Section 2.3.

2.2.2 Self-healing based on bacteria

The idea of bacteria-based self-healing is to utilize bacteria to promote precipitation of calcium carbonate in cracks. In 1990s, Gollapudi et al. [38] suggested to use bacteria to induce the precipitation of calcium carbonate (CaCO3) to repair cracks. The precipitation of calcium

carbonate can be caused by various metabolic pathways, such as the hydrolysis of urea and the oxidation of organic acids [3].

Compared to the other pathways for generating carbonate, the hydrolysis of urea has several advantages. For instance, it can easily be controlled and it has the potential to produce high amounts of carbonate within a short time [39]. Catalyzed by urease, urea is degraded to carbonate and ammonium. The concentration of carbonate increases in the bacterial environment [39]. One mole of urea is hydrolyzed intracellularly to one mole of ammonia and one mole of carbonate (Equation 2.1), which spontaneously hydrolyzes to one mole of ammonia and carbonic acid (Equation 2.2) [39]. These products subsequently reach the equilibrium in water to form bicarbonate and two moles of ammonium and hydroxide ions (Equation 2.3 and 2.4) [39]:

CO(NH2)2 + H2O ⇒ H2COOH + NH3 2.1

NH2COOH + H2O ⇒ NH3 + H2CO3 2.2

2NH3 + 2H2O ⇔ 2NH4+ + 2OH- 2.3

2OH- + H2CO3 ⇔ CO32- + 2H2O 2.4

In the presence of calcium ions, calcium carbonate is deposited, once a certain super-saturation level for the precipitation of calcium carbonate is reached [39]. As shown in Figure 2.2, because of the negative charge of the cell wall, calcium ions are attracted (Equation 2.5). As a

(29)

result, the crystals of calcium carbonate precipitate on the bacterial cell (Equation 2.6) [3, 39]. In addition, precipitation also takes place in the bulk liquid phase [39]:

Ca2+ + Cell ⇒ Cell-Ca2+ 2.5

Cell-Ca2+ + CO32- ⇒ CaCO3 2.6

Another metabolic pathway to produce calcium carbonate is the oxidation of organic acids. As expressed in Equations 2.7, 2.8 and 2.9, calcium carbonate is formed with the conversion of calcium acetate during the bacterial metabolism [40]:

Ca(CH3COO)2 + 4O2 ⇒ Ca2+ + 4CO2 + 2H2O + 2OH- 2.7

CO2 + OH- ⇔ HCO3- 2.8

2HCO3- + Ca2+ ⇒CaCO3 + CO2 + H2O 2.9

Compared to the hydrolysis of urea, which produces excessive ammonium, the oxidation of organic acids has less environmental impact [41]. Moreover, as presented in Equation 2.9, during the precipitation of calcium carbonate through this metabolic way, CO2 is produced as

well. The produced CO2 can also react with portlandite (Ca(OH)2), which is quantitatively an

important hydration product of Portland cement, to form more calcium carbonate [42, 43]. However, in order to use bacteria to heal cracks in concrete, some technical problems have to be solved. The bacteria should be protected not only against the alkaline environment in concrete, but also against the decreasing space in the matrix when hydration of cement proceeds [3, 42]. As found by Jonkers [42], when the bacteria spores, which can be viable for up to 50 years, were directly added into concrete, their lifetime dramatically decreased to only a few months. This is caused by the hydration of cement grains. As the cement grains hydrate, most of pores becomes smaller than bacterium spores with the size of 1 µm, which causes the cell to collapse [42]. It is demonstrated that the immobilization of bacteria in porous clay aggregates before the mixing of concrete can prolong the lifetime of bacteria enormously. Self-healing can be triggered later on when cracks intersect these clay particles [42-44]. Contemporarily, researchers in Ghent University are trying to immobilize bacteria in silica gel or polyurethane [45]. The “food” for the bacteria, urea or organic acids, should also be

Figure 2.2 Simplified representation of carbonate precipitation induced by the hydrolysis of urea with the help of bacteria [39].

(30)

embedded in the matrix in similar ways. It should be mentioned that the essential condition for bacteria-based self-healing is that water is present in cracks. Otherwise the chemical reactions in Equation 2.1 to 2.9 can not take place.

2.2.3 Self-healing based on mineral admixtures

Self-healing can be attributed to reactions of mineral admixtures in cementitious materials. These mineral admixtures are added to the concrete mixture during mixing. After the concrete cracks, some unreacted mineral admixtures are present at crack surfaces. When water penetrates into the cracks, these mineral admixtures start to react with the water in cracks. The cracks are then expected to be filled with reaction products.

By now, the mineral admixtures for self-healing can be categorized into two groups: expansive additive and crystalline additive [46]. In the case of expansive additive, the volume of the reaction products is larger than that of the admixture itself and the expansion depends on the composition of the admixtures [47]. Regarding the crystalline additive, its components can react with Ca(OH)2 to form crystalline products [46]. Both additives have been investigated.

For instance, Kishi and co-workers [48, 49] used a mixture of expansive agents, i.e. calcium sulfoaluminate (Ca4(AlO2)6SO4), free lime (CaO) and anhydrite (CaSO4). In their studies,

another expansive additive called geo-material, which mainly consists of silicon dioxide, sodium aluminum silicate hydroxide and montmorillonite clay, was also studied [49]. Similar to Kishi’s research in [48-50], various types of minerals as expansive admixtures for self-healing were also investigated by Sisomphon [46]. The effects of the mixing ratio of the mineral compositions on the capacity of self-healing were explored as well [46]. Apart from expansive additives, a synthetic cementitious material called “crystalline additive”, which contains reactive silica and some crystalline catalysts, was also applied for self-healing [46]. It has been reported that these mineral admixtures lead to remarkable improvement of watertightness of cracks due to self-healing [46, 48-50].

There are some advantages by using mineral admixtures for self-healing. Since some minerals are able to react intensively with water, self-healing of crack contributed by these minerals proceeds fast. Moreover, because of their expansive character, the expansive additive can definitely improve the efficiency of self-healing.

However, technical problems of this approach have to be solved. For instance, if the minerals are directly added into the concrete mixtures without any protection, once they come in contact with water during the mixing of concrete, they immediately start to react [47]. As a result, these added minerals are consumed before cracking. Moreover, when an expansive additive is used, expansion always occurs in the interior of the concrete matrix, which could cause damage [51]. Therefore, when applying mineral admixtures to realize self-healing, pre-processing, like encapsulation, is necessary. Sisomphon et al. [52] utilized expanded clay lightweight aggregates (LWA) to store sodium-monofluorophosphate solution for self-healing. Before the mixing of concrete, the LWA particles with healing solution were coated by a layer of cement paste. Ahn and Kishi [53] tried to apply the encapsulation technique to store free lime (CaO) and anhydrite (CaSO4) for self-healing before the mixing of concrete. One more

(31)

available in cracks, the same precondition as required for bacteria-based self-healing [46, 48, 49].

2.2.4 Autogenous self-healing

Concrete is also able to get healed while water is present in cracks without any admixtures such as bacteria or minerals. This is called autogenous self-healing of concrete.

2.2.4.1 Autogenous self-healing in Portland cement concrete

Autogenous self-healing in Portland cement concrete has attracted much attention since it was observed many years ago. According to Hearn [6], the phenomena of autogenous self-healing had already been noticed in water retaining structures, culverts and pipes by Hyde [7] by the end of nineteenth century. In 1920s, a more systematical analysis of autogenous self-healing was executed by Glanville [8]. After that, autogenous self-healing of cracks in bridges was also investigated [9, 10].

The effect of autogenous self-healing on water leaking through cracks was extensively studied by Clear [54], Hearn [55] and Edvardsen [13]. Moreover, Reinhardt et al. [56] correlated this effect with different temperatures and crack widths. The reduction of chloride ingression through cracks due to autogenous self-healing was reported by Fidjestol et al. [57], Ramm et al. [58] and Otsuki et al. [59]. Apart from the durability aspect, the improvement of mechanical properties of concrete due to autogenous self-healing has been explored as well. Lauer and Slate [60] demonstrated that the tensile strength measured perpendicular to the crack plane increased after autogenous self-healing of cracks. In that study, the influence of age and curing conditions were also taken into account. Similarly, the recovery in strength of concrete was also found by Dhir et al. [61] and Granger et al. [62].

Although the effects of autogenous self-healing have been investigated for many years, the results of the study at the components of reaction products formed the autogenous self-healing process are not consistent. Jacobsen and Sellevold [11] found some newly formed portlandite and ettringite in cracks in high performance concrete. Schlangen and Ter Heide [12] detected newly formed C-S-H in cracks after the cracked samples were cured in water. They concluded that autogenous self-healing was caused by further hydration of unhydrated cement clinker. Edvardsen [13] found calcium carbonate (CaCO3) in cracks after autogenous self-healing. The

investigations by Yang et al. [14] and Qian et al. [15] also evidenced the existence of CaCO3 in

cracks. As shown in Figure 2.3, when CO2 in the air dissolves in water, CO32- ions diffuse into

cracks through the crack mouth [13]. CaCO3 precipitates when the concentration of Ca2+ and

CO32- ions reach supersaturation level [13]. Moreover, Sisomphon [46] reported that, since the

concentration of CO32- nearby the crack mouth is higher than that inside the cracks, CaCO3

(32)

2.2.4.2 Autogenous self-healing in slag cement concrete

Blast furnace slag (BFS), as an industrial by-product, can be used to replace Portland cement in concrete mixtures. Energy used to produce cement can be saved and the CO2 emissions

caused by cement manufacturing is reduced [63, 64]. Moreover, the addition of slag can lead to a fine pore system and hence a low permeability of concrete [65]. The low permeability can improve the resistance against both chloride and sulphate, which can prolong the service life of concrete structures [66]. Therefore, blast furnace slag is widely used, at least in the Netherland.

Figure 2.3 Autogenous self-healing due to precipitation of calcium carbonate in the presence

of water and dissolved CO2 [13] (The figure was reprinted from [67]).

CaCO₃ CaCO₃

(33)

However, like Portland cement concrete, slag cement concrete is also a brittle composite material and cracks easily occur under tensile loading. These cracks have similar negative effects as in Portland cement concrete. Zhou et al. [68], Van Tittelboom et al. [3, 69] and Sahmaran et al. [16] explored the effects of slag on autogenous self-healing of cementitious materials.

Blast furnace slag is a latent hydraulic material and needs alkali to activate its pozzolanic reactions. In slag cement concrete, a high fraction of slag remains unhydrated in the matrix even after a long period of hydration if the fraction of slag is high. On the one hand, further reaction of these unreacted slag cores potentially contributes to autogenous self-healing [3, 69]. On the other hand, the consumption of portlandite by the reaction of slag reduces the potential of the precipitation of calcium carbonate and would have a negative effect on autogenous self-healing [3, 69]. Therefore, there are still challenges to improve autogenous self-self-healing of slag cement concrete.

2.2.5 Challenges about autogenous self-healing

As indicated above, there are several mechanisms of self-healing. Among these mechanisms, autogenous self-healing can take place without any “strange” admixtures as long as water or solution is available in cracks. This is different from self-healing based on bacteria or mineral admixtures, for which not only the presence of water in cracks, but also of the admixtures of bacteria or minerals are necessary. Compared to self-healing based on adhesive healing agents described in Section 2.2.1, autogenous self-healing is much cheaper and more environment friendly. Hence, autogenous self-healing is more favorable for concrete structures than other mechanisms of self-healing, although its rate is slow and only small cracks can be completely healed.

In Section 2.2.4 many investigations of the efficiency of autogenous self-healing in Portland cement concrete were mentioned. Numerous water permeability tests and chloride transport tests were performed to evaluate the efficiency of autogenous self-healing in Portland cement concrete. The recovery of mechanical properties of Portland cement concrete due to autogenous self-healing has been studied extensively as well. It is known that both the reduction of permeability and the recovery of mechanical properties after self-healing are strongly dependent on the amount of reaction products formed in cracks. However, by now there are still very few investigations carried out to quantify the reaction products formed in cracks as a function of time, i.e. the kinetics of autogenous self-healing.

Regarding the composition of reaction products formed in cracks, several studies have been performed. As indicated in Section 2.2.4, C-S-H, portlandite, ettringite and CaCO3 were found

in cracks after autogenous self-healing. However, the results of these investigations are not consistent. The mineralogy of the reaction products formed in cracks is insufficiently studied and the physico-chemical process of autogenous self-healing is not completely understood yet. The questions concerning autogenous self-healing in Portland cement-based systems also apply to slag cement-based systems. As demonstrated in Chapter 1, all these questions will be explored in this thesis.

(34)

2.3 Methods to supply healing agents to cracks

Except autogenous self-healing, self-healing with other mechanisms usually requires the supply of healing agents to the cracks. By now, the main methods to supply healing agents to cracks artificially are by using capsules or vascular systems. The state-and-art of these two methods is presented below.

2.3.1 Encapsulation technique

The encapsulation technique can be used to store and protect healing agents in the matrix. Once cracks appear and propagate through the matrix, some capsules can be hit and ruptured by the cracks, the healing agents are released into the cracks, provided that the healing agents are in liquid form. This concept has been used for self-healing in polymers for more than 10 years (see Figure 2.5) [70].

Similar to the case in polymer materials, many researchers have used capsules to supply liquid-based adhesive agents to cracks in cementitious materials for self-healing, as mentioned in Section 2.2.1. Also, solid healing agents, such as bacteria and mineral admixtures described in Section 2.2.2 and 2.2.3, respectively, can be protected by using capsules and mixed into concrete.

There are various encapsulation techniques available so far. They can be grouped into physical encapsulation and chemical encapsulation [71]. These techniques are widely used in the food industry and pharmaceutical industry [71]. Experience of encapsulation in the food industry and pharmaceutical industry can be useful for encapsulating healing agents for concrete mixtures.

(35)

In general, the use of capsules for self-healing has the advantage that self-healing of cracks can take place automatically [70]. However, there are some limitations of this method. Since the capsules are randomly dispersed in the matrix, only a small part of them can be hit and ruptured by cracks [72, 73]. In order to guarantee that cracks are able to intersect the capsules, the strength of capsules is usually designed to be lower than the matrix itself. This, however, can lead to a lower overall strength of the material.

2.3.2 Technique of vascular systems

For delivering liquid healing agents to cracks also a vascular system can be applied. As shown in Figure 2.6, if cracks intersect any part of the vascular system inside the material, liquid healing agents can be transported to the cracks through the vascular system [74-77]. As a result, self-healing of cracks takes place.

Dry [78] embedded some pipes in concrete as a simple vascular system to supply healing agents to cracks. It was found that the capacity of self-healing was high when using this method [78]. Stimulated by these positive results, some researchers are now investigating methods to install vascular systems in concrete structures. For instance, Pareek et al. [79] set up vascular systems in concrete structures by creating canal networks. To do that, steel bars with smooth surface are pre-embedded inside the concrete matrix during casting. After the concrete is cured for 1 day, the embedded steel bars are pulled out. As a result, a canal is created inside the concrete matrix. A network is formed when these canals are connected with each other. Sangadji and Schlangen [80] are also investigating the installation of a vascular system in concrete structures. “Porous concrete” (see Figure 2.7) is developed for creating vascular systems. The pores inside the porous concrete are of high connectivity and the pore size is in the order of magnitude of millimeters. Similar to the canal networks proposed by Pareek et al. [79], when any part of the porous concrete network is intersected by cracks, liquid healing agents can be delivered to cracks via this system.

As discussed before, one of the disadvantages of the use of capsules is that only a limited amount of liquid healing agent can be supplied to cracks. In comparison, sufficient liquid agents can be delivered to cracks by using vascular systems, leading to high efficiency of self-healing. However, extra efforts are needed to inject the liquid healing agents into the vascular systems. Moreover, by now it is still difficult to install a vascular system inside the matrix and the vascular system is usually vulnerable.

(36)

Figure 2.6 Material with a vascular system for self-healing [75].

Figure 2.7 Porous concrete for vascular systems [80].

2.3.3 Challenges concerning the supply of water or solution to the cracks in concrete structures exposed to the atmosphere

As indicated in Section 2.2.4, one of the essential conditions for autogenous self-healing is the presence of water or healing solution in cracks. However, for concrete structures exposed to the atmosphere (instead of being immersed in water), water is usually not available in the cracks inside the structures. In order to enhance the capacity of self-healing, it is necessary to supply water or healing agents to the cracks. In Section 2.3.1 and 2.3.2, many investigations were mentioned dealing with the use of capsules and vascular systems for supplying healing agents to cracks. However, there are still some technical problems that have to be solved before these methods can be applied.

(37)

Regarding the use of capsules or vascular system for supplying water or healing agent to cracks, it is necessary to predict the efficiency of self-healing as a function of the amount of water or healing solution. Moreover, when water or healing solution is released into cracks, this liquid can migrate into bulk paste because of capillary forces. In the absence of water or healing solution in cracks, self-healing of cracks can not proceed anymore. Therefore, the efficiency of self-healing could be negatively affected by the migration of this liquid from the crack into the matrix. This effect should be taken into account.

2.4 Conclusions

A general literature survey has been presented in this chapter. Mechanisms of self-healing in concrete structures have been discussed. From the comparisons of the mechanisms of self-healing, it is found that although the rate of autogenous self-healing is slow and only small cracks can be completely healed under this mechanism, autogenous self-healing is more preferential for concrete structures and more environment friendly than other mechanisms of self-healing as no “strange” additives are needed. Remaining scientific challenges concerning the autogenous self-healing in both Portland cement concrete and slag cement concrete are the following:

- The mineralogy of the reaction products formed in cracks has hardly been studied and the physico-chemical processes of autogenous self-healing are not completely understood yet.

- There are very few investigations carried out to determine the amount of reaction products formed in cracks as a function of time, i.e. the kinetics of autogenous self-healing.

- The influence of different factors on the capacity of autogenous self-healing is not exactly known, such as the amount of unreacted cement clinker (slag) and the initial ion concentrations of the solution in cracks.

- The effect of the migration of water or healing solution (from cracks into the bulk paste) on self-healing is not determined yet.

The challenges defined in this chapter will be dealt with in the following chapters. By focusing on these challenges, better insight is gained into the physico-chemical processes of autogenous self-healing. The potential of autogenous self-healing in Portland or slag cementitious materials is determined and recommendations on enhancing the capacity of autogenous self-healing are given.

(38)

3

Characterization of the Reaction Products

of Autogenous Self-healing

3.1 Introduction

Autogenous self-healing takes place when water or a healing solution is present in cracks. The “water to powder (i.e., cement and slag) ratio” in cracks is much higher than that in the bulk paste. It means that there is more water for further reaction of unhydrated cement or slag particles. Moreover, the space for the formation of reaction products in the cracks is larger than that in the bulk matrix, which may influence the nucleation and growth processes of the reaction products. Thus, the mineralogy of the reaction products formed in the cracks may be different from that in bulk paste. From the literature survey in the previous chapter, it was found that the composition of the reaction products formed in cracks is still insufficiently studied and the physico-chemical process of autogenous self-healing in Portland or slag cementitious materials is not completely understood yet.

In this chapter, in order to better understand the physico-chemical process of autogenous self-healing in Portland or slag cement materials, the mineralogy of the reaction products formed in cracks is investigated experimentally. For self-healing, cracked Portland cement paste is cured in distilled water. In order to investigate the effect of Ca2+ ion concentration of the healing agent on self-healing, cracked Portland cement paste is also cured in saturated Ca(OH)2) solution. Since slag is a pozzolanic material and its reaction needs to be activated by

alkaline solutions with a pH higher than 12 [20], saturated calcium hydroxide (Ca(OH)2)

solution is also utilized to cure the cracked slag cement paste. After the cracks are healed, environmental scanning electron microscope (ESEM) is employed to observe the morphology of the reaction products formed in cracks and energy dispersive spectroscopy (EDS) is applied to detect the chemical elements of these reaction products. In addition, the compounds of each mineral in the reaction products in a healed crack are characterized by Fourier transform infrared spectroscopy (FTIR). Thermogravimetric analysis (TGA) and X-ray diffraction (XRD)

(39)

are used to determine the fraction of each mineral in the reaction products formed in cracks. Based on the characterization of the reaction products in cracks, deeper insight is obtained into the physico-chemical process of self-healing in Portland cement paste and slag cement paste.

3.2 Materials and experiments

3.2.1 Materials and sample preparations

The materials used in this chapter are Portland cement paste and slag cement paste. CEM І 42.5N is used to prepare Portland cement paste, while CEM Ш/B 42.5N is used to prepare slag cement paste. Their chemical compositions are presented in Table 3.1 and 3.2, respectively. In CEM Ш/B 42.5N, clinker accounts for 34.1% by mass. The water to cement ratio (w/c) of the cement paste is 0.3. For self-healing, Portland cement paste is cured in water and saturated Ca(OH)2 solution, while slag cement paste is only cured in saturated Ca(OH)2 solution.

FTIR, TGA and XRD are used to characterize the reaction products formed in cracks. To use these techniques, the reaction products in cracks should be separated from the tiny cracks. However, it is difficult to make sure that the reaction products separated from the cracks do not contain substances from the bulk paste. To be sure that only the newly formed products are measured, artificial cracks (planar gaps, see Figure 3.1) instead of real cracks are made for obtaining them. Cement paste sample at the age of the paste of 7, 14 and 28 days is sliced. In order to assure that the artificial cracks have similar width, the slices of cement paste were carefully ground with P320, P500 and P1200 sand papers. After that the slices are pressed together (The pressure, less than 0.025 MPa, is very low compared to the strength of cement paste. Therefore, its influence on self-healing is ignored.). The width of artificial cracks is about 30 µm, checking with ESEM. The pressed slices at the age of 7 days and 28 days are cured in water or saturated Ca(OH)2 solution. As shown in Figure 3.1 (a), only the bottom

zones of the specimens, with a depth of 5 mm, are submerged in a limited amount of water or solution. Due to capillary forces, the water or solution is absorbed into the artificial cracks. For

Table 3.1: Chemical composition of CEM І 42.5N.

Compound CaO SiO2 Al2O3 Fe2O3 K2O Na2O SO3 MgO Total Weight (%) 64.40 20.36 4.96 3.17 0.64 0.14 2.57 2.09 98.33

Table 3.2: Chemical composition of CEM Ш/B 42.5N and the included slag (slag content: 66%).

Compound CaO SiO2 Al2O3 Fe2O3 K2O Na2O SO3 MgO Total CEM Ш/B 42.5N

(g/100 g cement) 45.52 30.61 10.58 1.42 0.58 0.31 2.66 7.33 99.01 Slag

(40)

self-healing in slag cement paste, the samples are completely submerged in saturated Ca(OH)2

solution (see Figure 3.1 (b)). The curing is under sealed conditions in order to prevent the dissolution of CO2 into the water or the healing solution. The temperature for curing is fixed at

20 ± 1°C.

When the pressed slices have been cured for 200 hours, they are separated and dried by vacuum drying for 2 hours. Some of these untreated slices are directly analyzed by using ESEM. The reaction products on the surfaces of the slices are detected by EDS, while other slices are used to obtain the reaction products for the measurements of FTIR, TGA and XRD. Because the slice surfaces are flat and the newly formed reaction products are weakly bound on the slice surfaces, these reaction products can easily be scratched off with a plastic sheet. The obtained reaction products hardly contain any substances from the original slice surfaces. The obtained reaction products are milled into powders (<50 µm) for FTIR, TGA and XRD tests.

(41)

3.2.2 Experimental techniques for characterization of reaction products formed in the self-healing process

3.2.2.1 ESEM/EDS

ESEM, as a scanning electron microscope (SEM), allows a gaseous environment in the specimen chamber for collecting electron micrographs of specimens that are wet and uncoated [81]. With this equipment, not only the samples after processing, such as drying, epoxy impregnation and polishing, but also the virgin samples can be observed. In addition, EDS equipped together with ESEM enables chemical element analyses of the samples. Its characterization ability is based on the fundamental principle that each element with a unique atomic structure has a unique set of peaks on its X-ray spectrum [82]. Therefore, the mineralogy of the reaction products can be observed by using ESEM, while their chemical elements can be analyzed with EDS.

Images of the morphology of reaction products formed on slice surfaces are taken by the backscattered electron (BSE) detector under moisture vapour mode. In order to acquire high contrast images, a beam with acceleration voltage of 20 kV is used. To analyze chemical elements of the reaction products, random points on the surfaces of the reaction products are detected by EDS. The magnification of the image is 800X. The ratios of the main chemical elements, i.e. Ca/Si, Al/Si and Mg/Si, in the reaction products formed on slice surfaces are then calculated.

3.2.2.2 FTIR

Fourier transform infrared spectroscopy (FTIR) enables the determination of the molecular structure of the compounds of a material by detecting the strength of light that is absorbed by the material. Each compound, except optical isomers, has its own specific IR absorption spectra. The region of IR wavenumber from 600 to 4000 cm-1 is correlative exactly with the chemical composition of a material.

Since long FTIR has been a tool to analyze the compounds in cement paste. The measurements described in this chapter are conducted in transmission mode in which the scan resolution in wavenumber is 4 cm-1. The spectrum of a sample is obtained by taking the average of 20 scans. The result is corrected against the background spectrum of air. Instrument control and spectral collection are performed by using Spectrum software.

3.2.2.3 TGA/DTG

Thermogravimetric analysis (TGA) and differential thermogravimetric (DTG) provide information about the decomposition or dehydration of minerals as a function of temperature. For cement-based materials, TGA/DTG can be adopted to estimate the degree of hydration and to specify quantities of CH in the system [83]. TGA/DTG can also give information about the mineralogy of reaction products formed in the healing process.

The tests are performed in argon atmosphere at 1.5 bars. The heating rate is 10 °C/min and the final temperature is 1100 °C. To charaterize the reaction products from slag cement paste, mass spectroscopy analysis is combined with TGA tests. During the decomposition of the