Selfhealing of Engineered Cementitious [603846]
MSc‐ thesis
Self‐healing of Engineered Cementitious
Composites (ECC) in Concrete Repair System
Xia Hua
July 2010
Graduation committee
Prof. dr. ir. K. van Breugel (CMD/Microlab)
Dr. Guang Ye (CMD/Microlab)
Dr. ir. P. C. J. Hoogenboom (Structural Mechanics)
Ir. L. J. M. Houben (Coordinator)
To my parents and Xi
TABLE OF CONTENTS
ACKNOWLEDGEMENTS i
ABSTRACT ii
1 INTRODUCTION 1
1.1 Problem definition 1
1.2 Objective of the research 2
1.3 Outline of the thesis 2
2 LITERATURE STUDY 3
2.1 Review of the Engineered Cementitious Composites (ECC) material 3
2.1.1 Characteristics of ECC 3
2.1.2 Additives in ECC 5
2.2 Self ‐healing in concrete materials 6
2.2.1 Introduction 6
2.2.2 Mechanisms of self‐healing 6
2.2.3 Conditions for self‐healing 7
2.2.4 Self‐healing approaches 8
2.3 Short summary 10
3 METHODOLOGY 11
3.1 Starting point 11
3.2 Approach 11
3.3 Challenge 12
4 EXPERIMENTAL STUDY 13
4.1 Sealing material and manufacture 13
4.1.1 Introduction 13
4.1.2 Investigation of sealing materials 13
4.2 Functional performance of ECC 17
4.2.1 Introduction 17
4.2.2 Materials and mix proportion 17
4.2.3 Specimen preparation 18
4.2.4 Three‐ point bending test 19
4.2.5 Nano‐computer tomography (nano‐CT) 20
4.2.6 Environmental scanning electron microscopy (ESEM) 22
4.2.7 Light microscope 23
4.3 Short summary 23
5 RESULTS AND DISCUSSION 24
5.1 Recovered mechanical properties in ECC 24
5.1.1 Load‐displacement relation 24
5.1.2 General results of stress‐deflection curves 25
5.1.3 Deflection hardening behavior influenced by the capsules 26
5.1.4 Deflection capacity and recovery 27
5.1.5 Flexural strength and recovery 28
5.1.6 Flexural stiffness and recovery 30
5.2 Nano ‐CT observation 32
5.2.1 Identification of microcracks in nano‐CT image 32
5.2.2 Opening of capsules 34
5.3 Light microscope and ESEM observations 35
5.3.1 Multiple cracking behavior 35
5.3.2 Observation of interface zone 36
5.3.3 Observation of crack inside 36
5.3.4 Observation of crack surface 37
5.4 Discussion of potential of self‐healing in ECC 39
5.4.1 Influencing factors 39
5.4.2 Healing efficiency 40
5.4.3 Short summary 41
6 CONCLUSIONS AND RECOMMANDATIONS 42
6.1 General conclusions 42
6.2 Recommendations 43
REFERENCES 44
APPENDIX 46
ACKNOWLEDGEMENTS
This research has been carried out at the Microlab of the Faculty of Civil Engineering and
Geosciences, Delft University of Technology, and continued for 8 months from December 2009
until July 2010. I wish to take this opportunity to express my gratitude to all those who helped me
in this research and successful completion of this thesis, especially the following people:
First of all, I would like to express my sincere thanks to Dr. G. Ye for introducing me to this project,
for his daily supervisions, discussions and constant encouragements throughout the research. I
am also very grateful for his effort to make this project go on smoothly.
Second, I really appreciate Prof. K. van Breugel and Dr. P. C. J. Hoogenboom for providing me the
general supervisions, for their valuable comments on the research.
Third, I would like to thank PhD student H. Huang for his stimulation and collaboration on sealing
material investigation and manufacture. During the study, PhD student J. Zhou also provided me
some interesting ideas and guided me to perform some experiments.
In addition, I also thank Mr. A. Thijssen from Microlab for his assistance on the nano‐CT scan
investigation and the coordinator of Structural Engineering Ir. L. J. M. Houben for his
administration of the graduation aspect.
Last but not least, I appreciate my parents and friends for their support and love in every step I
have taken during the two years of studying abroad.
Xia Hua
Delft, July 2010
i
ABSTRACT
Since the concept of high performance concrete was raised in the late 1980’s, it is well known
that concrete properties have been greatly improved. However, a larger number of existing
concrete structures are suffering deterioration resulting from external or internal causes even
during the early stage of service life. Durable repair of concrete has thus drawn more attention.
To this end, the main objective of this research is to develop and test a novel method for
promoting the self‐healing behavior in concrete repair system.
A new type high performance fiber reinforced cementitious composites called Engineered
Cementitious Composites (ECC) has been developed in recent years, characterized by high
ductility and improved durability due to the multiple micro‐cracking behavior. In this study, it was
proposed that the original ECC with local waste materials was embedded with capsules to
investigate the self‐healing potential of this modified ECC material. To realize this self‐healing
concept, Super Absorbent Polymers (SAP) can be used as the water reservoir enclosed in the
capsules, and then provide available water for self‐healing process when the capsules are
ruptured by cracking. Based on this idea, the preliminary experiments, concerning sealing
materials and encapsulation procedure, were first carried out. Three ECC mixtures focusing on
the influence of capsule content and capsule size were involved. In order to induce artificial
cracks, three‐point bending tests have been used to preloaded ECC specimens to different
deflection levels. After healing for 28 days, the specimens have been tested again in the
three‐point set‐up.
The experimental results reveal that the recovered deflection capacity of damaged modified ECC
specimens can arrive at 65%‐95% of control specimens, which is higher than that of specimens
without capsule. While the recovery of flexural strength and stiffness rarely show improvement.
Compared to the case of coarse capsules, the specimens with fine capsules show more
reasonable performance on mechanical properties including the deflection capacity, the flexural
strength and stiffness as well as their recoveries. The nano‐CT investigation confirmed that
moisture transportation took place in more than half of capsules. Under EDX observation, the
relatively high concentration of calcium on the crack surface could be considered that the healing
product was probably presented in terms of hydration product such as calcium hydroxide or
calcium silicate hydrate. However no apparent healed crack was found under ESEM investigation,
thus it could be inferred that the healing efficiency was not remarkable in this study. For further
research, it is suggested to investigate the contribution of the optimal capsule size and the
sufficient water supplement to more effective self‐healing behavior.
ii
1 INTRODUCTION
1.1 Problem definition
Concrete is a strong, versatile and economical material that has been widely utilized for
constructions over the world. Since the concept of high performance concrete was raised in the
late 1980’s, concrete properties have been greatly improved. However, a large number of existing
concrete structures, such as bridge decks and pavements are suffering deterioration resulting
from external and internal causes even during the early stage of service life. In the civil
engineering sector of the Netherlands, premature failure of structures leads to a situation where
about half the budget is spent annually on maintenance and repair. For concrete structures in
particular, 90% of the repair works focus on repair of cracks caused by reinforcement corrosion
[1]. Due to the brittleness of concrete, cracking is unavoidable in concrete structures. Cracking
can introduce chlorides, sulphates, oxygen, alkali or moisture into the concrete and accelerate
further deterioration of the whole system. To address this problem, the durability of concrete
repairs has drawn more attention. Even though the quality of concrete repairs has increased a lot
in recent years, realizing durable repairs is still difficult.
As stated above, people have tried to make better and stronger materials, which are capable of
repairing cracks and restoring their functionality. These materials can be defined as self‐healing
materials. The starting point for this study is a newly developed class of high performance fiber
reinforced cementitious composites with called Engineered Cementitious Composites (ECC),
characterized by high ductility, improved durability due to multiple micro‐cracking behavior and
lower fiber content [2]. Fine fibers guarantee a tight crack width and significant increase in strain
capacity. This ability of ECC to achieve tight crack width can contribute to engage self‐healing in a
variety of environmental conditions. Therefore, ECC as a means of healing invisible microcracks
can prolong the service life of structures meanwhile reduce the maintenance cost. Since ECC has
several special properties, the use of ECC for concrete repairs was proposed in the last decade. In
a currently running project at Microlab in TU Delft, a number of experiments based on the basic
ECC with local available materials have been studied. The preliminary results indicated that a
wide range of raw materials could be used as the basis for the ECC‐like repair materials [3].
It is well known that fine cracks under favorable moisture conditions has a potential to heal itself,
since self‐healing phenomenon was first directly observed in cracked water pipes in 1937 [4]. A
common agreement is that continued hydration of cement particles within the cracks is one of
the main reasons for the self‐healing. In fact, a large amount of unhydrated cement is available in
most concrete and especially in those concrete with low water/cement ratio. If cracks occur in
the matrix and water flow through cracked concrete, then unhydrated cement reacts with it,
resulting in new hydration products. This formation and growth of new hydration products will fill
in the cracks. Thus, additional water supply at the locations where cracks are formed is highly
significant for successful completion of self‐healing. The self‐healing can enhance the long‐term
performance of concrete repairs. For these reasons, this project studies the self‐healing potential
and mechanical properties of ECC material by a novel method.
1
1.2 Objective of the research
Since the ECC material has self‐healing potential, the main objective of this research is to develop
and test a novel method for promoting self‐healing behavior in ECC materials. The encapsulation
approach is considered as novel method of this study. More specifically, it is investigated that
when cracks rupture embedded capsules inside the ECC mixture, whether this action can release
healing agent (water) for further hydration of cement, without replying on external supply of
water. This study will be conducted as a preliminary study to get more insight into a
cement‐based self‐healing coating of old concrete in concrete repair system.
1.3 Outline of the thesis
This thesis consists of six chapters. Chapter 1 introduces the motivation, objective and overview
of this research. Chapter 2 gives a review of the literature study about the development of
cementitious materials and self‐healing behavior. Chapter 3 explains the methodology, including
the starting point, challenge and approach of the experimental research. Chapter 4 illustrates the
experimental set‐ups, procedures of the research program. Chapter 5 presents the results from
the mechanical tests and microscopic observations followed by the discussions concerning
influencing factors and healing efficiency of the self‐healing behavior. Chapter 6 summarizes the
general conclusions and recommendations for further investigation.
2
2 LITERATURE STUDY
This chapter intends to review the previous works on the ECC material including its characteristics
and additives in ECC. Moreover, the literature study of self‐healing phenomenon in concrete
materials including the mechanisms, the conditions and the approaches of self‐healing is
introduced.
2.1 Review of the ECC material
2.1.1 Characteristics of ECC
In the last decades, concrete with increasingly high compressive strength have been applied to
civil engineering since modern building constructions rapidly grow towards high‐rise and diversity.
The addition of steel fiber and powders improves a number of concrete properties. However,
most of these materials still remain brittle. In some cases, the brittleness increases as the
compressive strength goes up, which poses potential dangers or fracture failures of the concrete.
A specially designed cementitious material termed as Engineered Cementitious Composites (ECC)
has been developed by Li and continuously evolved over the last twenty years. ECC is
characterized by a high ductility in range of 3‐7%, a tight crack width of around 60μand
relatively low fiber content of 2% or less by volume [3]. In terms of main material constituents,
ECC has characteristics similar to regular Fiber Reinforced Concrete (FRC), including water,
cement, sand, fiber and some additives. Coarse aggregates are not used because they tend to
have negative effects on ductile behavior of the composite. So far, various fiber types and
different cementitious matrixes have been used in ECC, but the detail composition of ECC must
obey certain principles imposed by micromechanics considerations. The most fundamental
mechanical property of ECC is of the ability to carry higher levels of loading after first cracking
while undergoing large deformation. The fibers used in ECC are tailored to work with the matrix
for the purpose of constraining localized brittle fracture and guaranteeing more uniform
distribution of microcracks. Due to the slip‐hardening behavior of fibers, ECC can take increasing
load that generates new cracks at other sites. It can be observed from Figure 2.1 that first
cracking in ECC is followed by increasing stress accompanied by a rise in strain. This
strain‐hardening behavior of ECC is similar to ductile metals. m
Figure 2.1 Types of failure modes in cementitious materials [6]
3
The crack width is another important indicator, reflecting the durability of a concrete structure.
ECC exhibits a well crack width self‐controlled in terms of a flat steady state microcracks
propagation, see Figure 2.2. After the tensile deformation up to around 1% strain, the early
microcracks stop widening and remain more or less constant with crack width of around 60μ.
ECC material can be tailored to form numerous closely spaced microcracks. The crack width in
ECC is much smaller than the typical crack width observed in the reinforced concrete. Moreover,
the self‐control of crack width can be seen as intrinsic properties of ECC material, rather than
depending on steel reinforcement ratio and structural dimensions [7]. Figure 2.2 also shows the
tensile strain capacity of 5% that is about 300‐ 500 times great than normal concrete [8]. m
Figure 2.2 Typical tensile stress‐strain curve and crack width development of ECC [8]
Figure 2.3 The conceptual trapping mechanism with load‐displacement relation [9]
From the aspect of ECC‐concrete repair system, the advantage of using ECC as repair material is
that the trapping mechanism of ECC can serve as a means for enhancing the durability of repair
system. It was reported that microcracks emanated from the tips of defects on the ECC‐concrete
interface, kinked into and subsequently were trapped in the ECC material (Figure 2.3). Due to the
rapidly rising toughness of the ECC, additional load can drive further crack extension into the
interface after kinked crack arrest, followed by subsequent kink and arrest [9].
4
2.1.2 Additives in ECC
2.1.2.1 Blast furnace slag (BFS) and limestone powder (LP)
In order to develop a new version of ECC with locally available materials, a number of mixtures
with blast furnace slag (BFS) and limestone powder (LP) instead of fly ash and silica sand
respectively have been investigated at Delft University of Technology. Portland cement, BFS and
LP are used to produce ECC as matrix materials, which can enhance the mechanical properties
and durability of ECC [3]. There are only a small amount of LP reacting with cement clinker or
hydration products, thus the limestone powder usually behaves as an inert filler material. The
incorporation of limestone powder and Portland cement is conducive to early compressive
strength, workability and durability of concrete. When BFS is mixed with Portland cement, it
reacts with the calcium which is called the pozzolanic reaction. It was reported that the addition
of BFS leads to a lower strength at early age, however it does not have any side effect on the final
compressive strength. Besides LP , BFS is able to improve the durability of ECC and results in a well
homogenous fiber distribution. The experiments with different BFS, LP contents and different
water‐powder ratios were discussed [3]. The optimal results of the ECC mix proportion with
Portland cement, BFS and LP were used as a reference in this study.
2.1.2.2 Super Absorbent Polymer (SAP)
Super Absorbent Polymer (SAP) is a low cross‐linked polyelectrolyte which starts to swell when it
comes into contact with water, leading to the formation of a hydrogel (see Figure 2.4). At the end
of the last century, SAP added into concrete was mainly used to counteract autogenous shrinkage
or self‐desiccation of cement paste and internal curing agent as well. The application of SAP can
be attributed to its ability to absorb amounts of water a few hundred times its own weight,
release free water when the relative humidity (RH) of the concrete pore system decreases caused
by the cement hydration [10]. In fact, several studies verified that the addition of SAP in concrete
not only results in a reduction of autogenous shrinkage but also modifies other microstructure
properties of concrete such as porosity, connectivity of interface transition zone between cement
paste and SAP, workability and durability. The main reasons inducing the change of
microstructure especially the pore structures, are summarized [11],
i. When SAP is fully filled with water, it acts as soft aggregate. When it is empty, SAP acts as air
void in the concrete.
ii. Non‐uniformity of the dispersion of SAP during mixing.
iii. Water uptake of SPA changes the effective water‐cement ratio in the early hydration stage
and water release from SAP leads cement to further hydrate.
Figure 2.4 Dry, collapsed and a swollen suspension polymerized SAP particle [10]
5
2.2 Self‐healing in concrete materials
2.2.1 Introduction
Self‐healing is generally defined as the ability to repair or heal damage of material itself [12]. In
natural materials, skin tissue and bone structures are perfect examples of self‐healing behavior.
Although the mechanisms of healing in natural materials cannot be copied exactly, some forms of
healing in concrete materials have been observed based on the similarity theory.
2.2.2 Mechanisms of self‐healing
Self‐healing behavior in cementitious materials has been demonstrated by numerous
experimental investigations and practical experiences [7,12‐14]. The autogenous healing
phenomenon is that the material has the ability to seal itself without external monitoring or
human intervention. Self‐healing of cracks in concrete is a combination of the complicated
chemical and physical processes. Up to now, several possible causes can be illustrated as follows
(also schematised in Figure 2.5):
i. Formation of calcium carbonate or calcium hydroxide.
ii. Blocking cracks by impurities in the water and loose concrete particles resulting from crack
spalling.
iii. Further hydration of the unreacted cement or cementitious materials.
iv. Expansion of the hydrated cementitious matrix in the crack flanks (swelling of C‐S‐H).
i) ii) iii) iv)
Figure 2.5 Possible mechanisms for self‐healing [14]
Among these, most studies have indicated that the primary mechanism is attributed to the
crystallization of calcium carbonate [7,13]. This view is sustained by the fact that precipitated
calcium carbonate can often be observed at the outside surfaces of the crack as some white
residue. As one of the cement hydration products dissolved in water, the calcium hydroxide is
liberated between the crack surfaces. Then free calcium ions from cement hydration react with
carbon dioxide presented at water surface, so that self‐healed crystal is formed, grows at both
surfaces of crack and finally fills the gap. This process can be described as following formulas,
6
↔+22 + ‐Ca(OH) Ca 2OH
−+→2+ 2
33 Ca CO CaCO
According to Neville [15], self‐healing was mainly owing to continued hydration in his opinion at
first. But later he stated that this is only applied to very young concrete [16] and believed that the
formation of calcium carbonate is the most likely cause of self‐healing. Besides, loose particles
blocking the crack path was also mentioned in some studies as a reason for healing cracks. Since
this was considered to cause the first fast decrease of water flow through the cracks [17].
2.2.3 Conditions for self‐healing
From the literature study, it is pointed out that five general criteria should be satisfied to ensure
self‐healing. These necessary conditions to experience healing of cracks are:
i. Presence of water
All the studies so far state that the presence of water is essential to facilitate healing of the cracks.
Without water, it is impossible for the calcium hydroxide to be leached out of the bulk material
into crack [13].
ii. Presence of chemical species
Adequate concentrations of certain critical chemical species for instance carbonate ions or
bicarbonate ions and free calcium ions dissolved in a flow of water, play a direct role to exhibit
healing mechanisms. This is readily available due to the chemical makeup of cementitious
materials and incomplete hydration [12,15].
iii. Crack width
Another important condition is the controlled crack width, which is associated with the efficiency
of self‐healing in cementitious materials. The crack width to engage noticeable self‐healing
behavior falls below 150μ and preferably lower than 50μ[7]. The smaller microcrack width
requires less self‐healing products to fill the crack and easier to grow from both surfaces of the
crack to get connected. m m
iv. Water pressure
If the water flows go fast through the crack, self‐healing will not take place. Therefore the water
pressure should be not too large and this condition is influenced by the ratio between water
height and the thickness of the structure for a certain crack width [14].
v. Stable crack
To guarantee that a crack is not damaged again, the crack should be under stable condition and
the crack width has to be constant instead of variation with time.
7
2.2.4 Self‐healing approaches
With the developments of smart materials, several innovative approaches of self‐healing have
been promoted in recent years. The core of these approaches is capable of continuously offering
materials or energy. For another, an ideal healing agent is supposed to continuously sense and
respond to damage, and recover the material performance without adverse affecting the matrix
material properties [18]. Several approaches based on this principle can be discussed below.
2.2.4.1 Encapsulation
The microcapsules can be defined as “particles, spherical or irregular, in the size range of about
50 nm to 2000μor larger, and composed of an excipient polymer matrix (shell or wall) and
incipient active polymer (core substance) ” [19]. The microencapsulated approach of
incorporation of healing agent was demonstrated by White [12], and Figure 2.6 illustrates this
autonomic healing concept. When the crack ruptures embedded microcapsules, the healing
agent is released into the crack plane through capillary action. Then the healing agent contacts
the embedded catalyst, triggering polymerization that bonds the crack faces closed. However, a
successful completion of the healing process is not easily realized since it combines a complex set
of requirements on storage, rupture, release, transport and healing. Furthermore, some studies
indicated that specific problems in terms of the size of microcapsules and surface morphology
significantly influenced the healing efficiency [12]. m
i)
ii)
iii)
Figure 2.6 Basic method of the microcapsule approach: i) cracks form in the matrix; ii) the crack
ruptures the microcapsules, releasing the healing agent into the crack plane through capillary
action; iii) the healing agent contacts the catalyst, triggering polymerization that bonds the crack
faces closed [12].
8
2.2.4.2 Hollow glass fibers
The use of hollow glass fibers (Figure 2.7 (A)) follows the similar concept as the microcapsules.
Glass is a typical brittle material, once the glass fibers break, the healing agents flow into the
matrix cracks and heal them so that the mechanical properties of concrete can be regained to a
certain extent. The key advantage of hollow fibers approach is that the fibers can be placed at any
location depending on the operational requirement to deal with specific failure threats (Figure
2.7 (B)). In order to quickly and easily see the internal damage in composite materials, a damage
visual enhancement method was designed by Pang and Bond [12]. In their work, the fibers filled
with healing agent were mixed with fluorescent dye to monitor the healing process (Figure 2.7
(C)).
(A) (B)
(C)
Figure 2.7 (A) Hollow glass fibers; (B) hollow glass fibers embedded in carbon fiber‐reinforced
composites laminate; (C) damage visual enhancement in composite laminate by the bleeding
action of a fluorescent dye from hollow glass fibers [12]
Li [20] utilized hollow glass fibers embedded in ECC material to investigate the feasibility of
passive smart self‐healing via experimental tests. The sensing and actuation mechanisms were
validated by observation of ESEM and the effect of recovery was validated by regaining of flexural
stiffness. He also concluded that the study of the biomimetic self‐healing technique presently was
still in the experimental stage, and plenty of issues remain such as how to effectively fill and place
the hollow glass fibers in large‐scale applications.
9
2.2.4.3 Bacteria
On the other hand, it is found that bacteria incorporated in the concrete matrix as self‐healing
agent probably catalyzes the autonomous repair of cracks [21]. Basically, bacteria of the genus
bacillus were used for the biological production of calcium carbonate ‐based minerals. Such
bacteria added in the cement matrix prior to casting should keep viable for prolonged periods.
Once integrated in the concrete matrix, it should be able to produce amounts of minerals needed
to plug or seal freshly formed cracks. In this sense, integrated bacteria would thus represent an
internal self‐healing agent which autonomously decreases matrix permeability upon cracks
formation. The scenario is schematically shown in the following figure.
Figure 2.8 Scenario of crack‐healing by concrete‐immobilized bacteria. Bacteria on fresh crack
surface become activated due to water ingression, start to multiply and precipitate minerals such
as calcite, which eventually seal the crack and protect the steel reinforcement from further
external chemical attack [12].
2.3 Short summary
From the above literature study, it is known that the ECC material itself has healing potential
primarily for the fine cracks. On the other hand, the self‐healing process can be only realized in
the presence of water, thus it is proposed to use SAP based on its high capacity of absorption as a
water carrier in the mixture to realize the self‐healing of ECC. The detailed methodology will be
introduced in the next chapter.
10
3 METHODOLOGY
3.1 Starting point
The idea of this research comes from an “embedded capsules” approach to repair material itself.
Two starting points were proposed in this study to realize the self‐healing process in concrete
repair system. As mentioned in the introduction, the first starting point is to use ECC material in
studying the healing potential. Because ECC exhibits the high strain capacity and tight crack width
control, those unique properties can promote the occurrence of self‐healing. The second starting
point is related to the saturated SAP, here SAP is considered as a water carrier enclosed in
capsules since it is able to absorb a large amount of water, and the water released from SAP has
the function of promoting the further hydration of the cement.
3.2 Approach
The core question of this thesis can be simply stated as whether water can be released from the
embedded capsules thereby promoting self‐healing in ECC material. The first task is to find a
proper way to seal the saturated SAP. To realize healed cracks in laboratory conditions, the
capsules are ruptured by inducing artificial cracks for releasing water. Finally, mechanical test and
microscopic observation will be carried out to assess the quality of self‐healing. The overview of
this approach can be illustrated in the following Figure 3.1.
Figure 3.1 Basic approach of the self‐healing concept
11
3.3 Challenge
From the approach presented in the previous section, there are three main challenges involving
sealing material and manufacture, cracking pattern during preloading and self‐healing
observation, which will be discussed as below,
iv. Sealing material and manufacture
The suitable sealing material should try to meet a set of requirement on physical, chemical
features and mechanical properties at the same time. For instance, it is expected to has a
stable capacity of water storage, and be sensitive to cracks whilst it is asked for a good bond
strength between the sealing material and the matrix. On the other hand, the manufacture
of capsules could be difficult without rolling machine. Since dry SAP powders used in this
experiment has a small particle size of 300μin diameter, after swelling it becomes softer,
such that it is difficult to be gathered to form a ball. m
v. Cracking pattern during preloading
The second challenge is to determine the crack propagation through the capsules. In this
research, the capsule consists of saturated SAP particles as the core and sealing shell as the
outer surface. When the capsules are incorporated into ECC, the bond strength at interfacial
transition zone (ITZ) between the sealing shell and the cement‐based matrix needs to be
stronger than the strength of capsule itself, to ensure the artificial cracks propagate through
the capsules instead of passing around the capsules. This can be illustrated in Figure 3.2.
Cement‐based matrix
ITZ Sealing shell
Saturated SAP
Figure 3.2 Ideal pattern of the crack passing through SAP
vi. Self‐healing observation
Since the mechanical test until final failure will damage the specimen, how to monitor the
internal crack pattern in the case of ensuring the integrity of specimen, that is another
problem for this study to assess the quality of self‐healing in ECC material.
12
4 EXPERIMENTAL STUDY
4.1 Sealing material and manufacture
4.1.1 Introduction
Since further hydration can only be realized in the presence of water or solution, SAP particles
can be introduced as a water reservoir in cementitious composites. In order to cause the capsules
to release the entrained water at the right time, the outer surface of saturated SAP needs to be
sealed by a protected layer. In this research, the ideal sealing material can be defined as that
which meets the following three requirements. The first is to appear impervious to leakage of
water before inducing the microcracks. Second, this material should be sensitive to cracking,
allowing the broken of capsules occurs at a certain level before arriving at ultimate strength of
ECC material. Last requirement is the proper interfacial bond strength. This strength of interface
between the capsule and the matrix requires being stronger than the strength of the capsule, to
guarantee that cracks can propagate through the capsules rather than around them. Thus high
bond strength at the interface is one of the important factors contributed to cracks passing
through the capsules.
Besides the intrinsic properties of sealing material, the diameter of capsule and the surface
morphology also highly influence capsule cracking behavior. Normally the use of capsules
embedded in composite materials has a negative effect on the mechanical properties such as
strength and ductility, especially when the diameter is relatively larger compared to the specimen
size, this disadvantage should therefore be minimized in this research.
4.1.2 Investigation of sealing materials
4.1.2.1 Selection of sealing materials
In this experiment, two sealing materials were prepared: paraffin wax and epoxy‐cement material.
The reason for taking advantage of wax is that the wax generally has an excellent water resistant
property, stable chemical characteristics. But it is a brittle material that cannot be mixed via a
mixer. Moreover the bond capacity of wax could be relatively weaker due to the smooth surface.
In this research, keep water available for self‐healing behavior is the crucial requirement. Another
alternative is to use the epoxy‐cement material, since the water‐soluble epoxy blended with a
number of cement can maintain better compatibility with the surrounding matrix. The
composition of this epoxy‐cement is presented in Appendix A.
4.1.2.2 Encapsulation procedure
As above mentioned, the size of capsule is not allowed to be large. Here the capsules used in this
experiment were made into two groups with an average diameter of 8mm and 5mm, called
“coarse capsule” and “fine capsule” respectively. The complete procedure of sealing saturated
SAP particles is shown in Figure 4.1. In the first step, every single saturated SAP particle with a
13
small size could be more easily gathered and shaped into a ball when CEM I 52.5N was utilized to
form a surface cover. To finish this process, the saturated SAP particles were sieved by 2.4mm size
of sieve, to separate them from the excess cement. It is important to control the rate of shaking.
If the amount of cement is less, the thin surface cover would not form. However, excess cement
will absorb more water from saturated SAP particles. To maximize the contained water inside
capsules, one method of avoiding water loss was to cure these balls in water at a temperature of
20 for 7 days, in order to achieve the hydration of cement and keep the SAP particles fully
absorbing water. Afterwards, the out surface of ball was sealed by a shell of wax or epoxy‐cement,
respectively. For the first case, paraffin wax was heated up to 105 and then kept a ball into
this hot solution for 2 seconds. Finally, wax microsphere was obtained from rapid cooling of the
suspension of molten wax droplets and it was cured under room condition. For the second case,
5 wt% epoxy and 100 wt% cement was mixed by hand and then rolled a ball in the epoxy‐cement
paste until smooth. After this, the ball was cured in RH 100% at 20 for 7 days. In such a way,
saturated SAP particles were made into two types of capsules (Figure 4.2). C°
C°
C°
Figure 4.1 Flow chart of the encapsulation
14
(A) Saturated SAP particles (B) Shaped by cement cover
(C) Sealed by paraffin wax (D) Sealed by epoxy‐cement paste
Figure 4.2 Manufacture of the capsules
4.1.2.3 Evaluation of sealing effect
To investigate the sealing effect of two different materials, 10 groups of capsules for each sealing
material were cured under room condition (RH 70% at 23 ) and the mass losses of these
capsules due to the evaporation were measured. The weight of capsules was recorded within
three consecutive days after encapsulation procedure. C°
Figure 4.3 demonstrates the results of water loss in the capsules against time for two sealing
materials respectively. As seen in Figure 5.1 (A), the curves of remaining water seem to
approach a horizontal. The lowest point still keeps water inside above 99% by weight after 3 days’
curing in air, which means almost no water loss from the capsules. While it can be observed in
Figure 5.1(B) that, when the capsules sealed with epoxy‐cement paste, the mass of remaining
water decreases rapidly within the first 24 hours and finally the average weight percentage is
approximately below 85%. Compared with sealed with paraffin wax, there is a relatively large
water loss in the capsules when using epoxy‐cement paste. Thus paraffin wax was preferred as
the sealing material in the later experiment.
15
9596979899100101102103104105
0 2 04 06 0Time (h) Remaining water (%)
80
(A) Capsules sealed by paraffin wax, curing at 23 RH 70% C°
7580859095100105
0 2 04 06 08Time (h)Remaining water (%)
0
(B) Capsules sealed by epoxy‐cement paste, curing at 23 RH 70% C°
Figure 4.3 Comparison of mass loss of capsules sealed different sealing materials
16
4.2 Functional performance of ECC
4.2.1 Introduction
This chapter focused on proposing the experimental program and set‐ups used in this research.
To study the functional performance of healed ECC, the overview of the experimental program
can be designed as following steps. After preparation of ECC specimens embedded with the
capsules, the principal task is to introduce cracks to these specimens. For this end, the ECC
specimens are under three‐point bend to form cracks inside. After healing for 28 days, the
specimens will be tested again in three‐point bending. Meanwhile, the reference without
capsules is parallel tested for comparison. The crack pattern and the healing products can be
observed by Nano‐CT, environment scanning electron microscopy (ESEM) and light microscope
techniques to verify whether the self‐healing phenomenon takes place.
4.2.2 Materials and mix proportion
As mentioned above, various modified ECC incorporating local waste materials have been
developed to optimize ECC mix proportion and investigate the self‐healing properties. BFS as the
main cement replacement material in the Netherlands shows a potential of pozzolanic reaction
but these reactions need to be activated by the hydration products of Portland cement. In mix
design, the Portland cement and BFS were used as cementitious materials, and the limestone
powder was considered as inert filler material. In this thesis, one mixture of ECC was chosen as a
reference and other two mixtures of ECC mixed with blending additional capsules compared with
the reference. As shown in Table 4.1, three mixtures were prepared in all for comparison purpose,
including M1‐M3. M1 without capsules was a trial mixture as the reference, M2 and M3 contains
the same capsule content of 2% of cement weight, but with capsule size of 8mm in diameter
(coarse capsule) and 5mm in diameter (fine capsule) respectively. The water to binder ratio,
water to powder ratio and superplasticizer content were therefore constants to reach similar
workability for each mixture. The details for the material properties and the composition of each
mixture are presented in Appendix B. In all mixtures, the polyvinyl alcohol (PVA) fiber with a
length of 8mm and a diameter of 40μwas used with the contents of 2% by total volume. In
order to control the amount of water inside capsules and prepare the specimen better, the
weight of saturated SAP (shown in Appendix B) was converted into the number of capsules, see
Appendix C. m
17
Table 4.1 Mix proportion of ECC (by weight)
Mix CEM I Limestone BFS Saturated Water Water/powdera Super‐ PVA fiber
Number 42.5N Powder SAP ratio plasticizer (by volume)
M1b 1 2 1.2 0 1.092 0.26 0.030 2%
M2c 1 2 1.2 0.02
(0.46% by volume) 1.092 0.26 0.030 2%
M3d 1 2 1.2 0.02
(0.23% by volume) 1.092 0.26 0.030 2%
Remark: a Powder includes cement, BFS and limestone powder
b Mixture without capsules (reference)
c Mixture with capsule size of 8mm (coarse capsule)
d Mixture with capsule size of 5mm (fine capsule)
4.2.3 Specimen preparation
The ECC specimen preparation followed the procedure described in [3]. In the first place, the
solid materials, CEM I 42.5, BFS and limestone powder were mixed with a HOBART mixer for 2
minutes at low speed. Then water and superplasticizer were added at low speed mixing for 1
minute, followed by high speed for 2 minutes. At last, the PVA fibers were added at low speed
and the sample was mixed at high speed for another 2 minutes. In the case of ECC specimen with
capsules inside, the capsules were mainly arranged in the center region of the specimen within
the stacking sequence to increase the opening probability of capsules. The distribution of the
capsules embedded in ECC specimen can be illustrated in Figure 4.4. Because the capsules were
not strong enough to resist the mixing process, they were added into the specimen layer by layer
by hand in the final step. After 1 day curing in moulds covered with plastic sheet, the specimens
were cured under sealed condition at a temperature of 20 for another 27 or 55 days before
testing. The specimens with the dimension of 160mmC°
×40mm×40mm were carefully ground
with P120 sand papers before the mechanical test.
Front Profile Capsule
Top
Specimen
Figure 4.4 Distribution of capsules embedded in the specimen
18
4.2.4 Three‐point bending test
In this research, the three‐point bending test was the main method to induce artificial cracks and
also characterize the mechanical properties of the modified ECC material.
4.2.4.1 Experimental set‐up
As seen in Figure 4.5, the support span of three‐point bending test set‐up is 110mm and the load
is located in the middle of the specimen. The configuration of three‐point bending test set‐up is
explained more in Appendix D. Two linear variable differential transducers (LVDTs) are fixed on
both sides of the set‐up to measure the vertical deformation at mid cross‐section of the specimen.
The test was conducted under deformation control at a constant speed of 0.01 mm/s. At least
two measures were done for each mixture, and the flexural strength and deflection were
calculated based on the average results of these measures.
Figure 4.5 Three‐point bending test set‐up
4.2.4.2 Experimental program
As shown in Figure 4.6, the overall program for three‐point bending test includes three different
schemes for each mixture, resulting in nine combinations in total. In order to roughly estimate
how much deflection is preloaded to, the samples from scheme A was bended until final failure
at first to derive the flexural stress‐deflection relation. Based on the results from scheme A
(reference), the preloaded levels of schemes B and C were then determined. The deflection of
1.0mm was selected since it is approximately equal to half the ultimate strength. Meanwhile the
deflection of 1.3 mm is well below ultimate deflection of the mixture. For schemes B and C, the
samples were cured in sealed condition at 20 for 28 days and preloaded to the specified
deflection levels, afterwards, the preloaded samples were further cured under sealed condition
for another 28 days before testing. C°
19
Remark: X stands for M1, M2, and M3 respectively. Therefore, 3 mixtures and 3 schemes for each
mixture, resulting in 9 combinations in total.
Figure 4.6 Bending test program of ECC material
4.2.5 Nano‐computer tomography (nano‐CT)
4.2.5.1 Experimental set‐up
The nano‐CT instrument is based on the same general principle as other micro‐CT systems, while
the improvement of nano‐CT is that it provides the focal spot size in the submicron range and
capable of resolving image features as small as 200nm, thereby achieving stable and ultra
high‐resolution images. The principle of nano‐CT system is depicted in Figure 4.7. It can be seen
that the sample is placed on the object stage between X‐ray source and array detector. When the
sample is rotated within the x‐ray cone beam, the CT scanner firstly acquires the complete
geometry of the component by generating a series of x‐ray images. The resulting X‐ray absorption
image including information about sample features for instance the position and the density is
projected onto the digital detector, which will be available for the numerical 3D reconstruction of
the volume data. This output is a 2D X‐ray image, where differences in gray‐scales reflect changes
in density of the sample. Darker regions correspond to a material of low‐density such as pores
whilst lighter regions correspond a material of high‐density such as cement. A full 3D data set is
created by collecting 2D images though a 360 degrees rotation, then these volumetric data are
reconstructed to visualize a virtual 3D volume.
The NANOTOM is a very compact laboratory system, which is particularly suit for the 3D
inspection of small and complex samples such as mineral samples and composite materials, etc.
The sample size is allowed to 120mm in diameter, 150mm in height and 1kg in weight. Based on
3D analysis with submicrometer resolution, any internal difference in material, density and
porosity can be measured. Hence this system is perfect tool for examination of the samples for
internal cracks, inclusions and porosity and more. On the other hand, the NANOTOM includes the
software package for system control, data acquisition, image analysis and realistic 3D
visualization, to ensure optimum image quality and high volume reconstruction speeds. Software
20
Image J (free license) was applied to generate 3D image and further analyze the cracking pattern.
Figure 4.7 Schematic representation of nano‐CT system
4.2.5.2 Experimental program
As mentioned in Chapter 3.3, the important advantage of nano‐CT is that the non‐destructive
technique can show the internal crack pattern meanwhile keep the integrity of the specimens.
Therefore, nano‐CT was used to view that how the cracks propagated within ECC material in this
research. The experimental program is shown in Figure 4.8, the nano‐CT only aims at a situation
that the specimen contains capsules. The scanning was carried out within the specified location
(Figure 4.4) where all the capsules were distributed. In order to study the internal differences
before and after healing, the images from the same position and the same scan direction are
guaranteed for each specimen. On the other hand, the images from front, top and profile
direction (Figure 4.4) are conducive to observe the healed cracks from multiple directions.
Figure 4.8 Nano‐ CT program of the preloaded ECC
21
4.2.6 Environmental scanning electron microscopy (ESEM)
4.2.6.1 Experimental set‐up
Environmental scanning electron microscopy (ESEM) was preferred in this study to analyze the
quality of self‐healing products formed inside the crack. This technique can provide insight into
the chemical composition of healing products and therefore identify the self‐healing behavior in
ECC material. The ESEM retains all performance advantages of a conventional SEM (Figure 4.9),
moreover eliminates the high vacuum constraint on the sample environment. The electron gun at
the top of the column creates a electron beam, and then the electrons are accelerated and
focused by a series of magnetic lenses and apertures. A set of scanning coils deflects the electron
beam in a scanning pattern over the sample surface and the objective lens offers the final
focusing. The interactions between the beam electrons and the sample atoms will generate a
variety of signals in forms of secondary electrons (SE), backscattered electrons (BSE) and
characteristic X‐rays, and emerging signals can be detected and reconstructed into a virtual image
displayed on the monitor screens.
Figure 4.9 Schematic of a scanning electron microscopy
4.2.6.2 Sample preparation
To observe the crack inside, samples from M2 and M3 series (with capsules) after 28 days’ curing
were examined under ESEM. Since it is desired that components inside the capsule can be clearly
seen, the sample was not impregnated with a low‐viscosity epoxy. Sample preparation for BSE
imaging followed the steps below. At first, the specimen was cut into a small piece with the
dimension of 20mm×20mm×30mm, and then was ground with sand paper from p300‐ p4000 for
about half minute each. During the grinding process, ethyl glycol was used instead of water in
order to keep its probable products integral. For the case of EDX (Energy dispersive X‐ray
spectroscopy), samples were dealt with by a small chisel to maintain the fractal surface.
22
4.2.7 Light microscope
As shown in Figure 4.10, a transmitted light microscope Leica MZ6 with cold light sources (CLS150)
was used in this study to observe the typical crack pattern. This modular stereomicroscope can
create brilliant three‐dimensional images of spatial object and Leica cold light sources provide
strong light intensity even within small space. In addition, the CLS 150 has been specially adapted
for automated control of the new transmitted light base via the powerful software of LAS (Leica
Application Suite). Through the serial interface, brightness and the electronic shutter can be
controlled using the computer.
Figure 4.10 Light microscope
4.3 Short summary
At the beginning of this chapter, the preliminary experiment for investigating two sealing
materials was introduced. Based on using the paraffin wax as the sealing material, the mechanical
testing program and several microscopic observations were purposed and introduced to explore
the functional performance of healed ECC. The next chapter will present the experimental results
and discussions.
23
5 RESULTS AND DISCUSSION
5.1 Recovered mechanical properties in ECC
Several techniques have been used in examining self‐healing behavior. In this section, the
self‐healing in ECC is evaluated from the point of view of mechanical properties.
5.1.1 Load‐displacement relation
The displacement controlled three‐point bending test records the load‐displacement relationship.
One example of load‐displacement curve at three different stages of loading is given in Figure 5.1
(A). As indicated, there is an initial linear‐elastic part up to the first crack strength. The following
is of the propagation of cracks, more microcracks are formed and developed in the specimen but
the loading continues to increase during this stage which is called hardening and the material is
still capable of resisting higher levels of loading up to a maximum. After the peak load is reached,
the applied load becomes to go down, a single macrocrack has appeared and the material has
started to soften.
Compared with the control samples bended until final failure (scheme A), the preloaded samples
(schemes B and C) have different stages, as described in Figure 5.1 (B). When the desired
preloading level is arrived the specimen then is unloaded. After healing period, the specimen is
reloaded under three‐point bending test again. Because of reopening of the cracks resulted from
the preloading stage, the load‐displacement curve of reloading presents differences in terms of
deflection, stress, and stiffness, which are discussed further below respectively. The calculation of
conversion from load‐displacement relation stress‐deflection curve is explained in Appendix D.
Displacement (mm)Load (kN)
Softening
Propagation of cracks
Linear‐elastic
(A) Control samples (scheme A)
24
Displacement (mm)Load (kN)
Unloading
Reloading Preloading
Curing for 28 days
(B) Preloaded samples (schemes B and C)
Figure 5.1 Comparison of load‐displacement curves for different schemes
5.1.2 General results of stress‐deflection curves
To represent the typical feature of different mixtures and schemes, the general results of
stress‐deflection curves from three‐point bending test are presented in Figures 5.2‐5.4. For
comparison of mechanical properties in Figures 5.6, 5.7 and 5.9, the value shown in column
charts were calculated based on the average results of at least two experimental results, after
removal of the maximum and minimum. All results in Figures 5.2‐5.4, 5.6, 5.7 and 5.9 were
derived from the stress‐deflection curves which are given in Appendix E.
M1 series (no capsule)
0246810121416
01234Deflection (mm)Bending stress (MPa)Scheme A
Scheme B
Scheme C
Figure 5.2 Bending stress‐deflection curves of M1 series
25
M2 series (coarse capsule)
0246810121416
0123Deflection (mm)Bending stress (MPa)
4Scheme A
Scheme B
Scheme C
Figure 5.3 Bending stress‐deflection curves of M2 series
M3 series (fine capsule)
0246810121416
01234Deflection (mm)Bending stress (MPa)Scheme A
Scheme B
Scheme C
Figure 5.4 Bending stress‐deflection curves of M3 series
5.1.3 Deflection hardening behavior influenced by the capsules
Figures 5.5 shows the bending stress‐deflection curves of series M1‐M3 under scheme A (control
scheme). As expected, M1 (no capsule) present a typical deflection hardening behavior, which
characterized by a straight, linear part up to first cracking, then followed by a bent over and
consequent plateau curve until final failure. When comparing the linear‐elastic curves between
reference (M1) and other two mixtures (M2 and M3), it can be observed that once linear part of
M2 and M3 (with capsules) ends, a shorter bend over part is followed by a decrease of flexural
stress at a rapid rate, which implies final failure takes place quickly. Thus one conclusion can be
drawn that the capsules have a negative effect on deflection hardening behavior of ECC.
26
Scheme A (bend to final failure)
0246810121416
0123Deflection (mm)Bending stress (MPa)
4M1
M2
M3
Figure 5.5 Comparison of deflection hardening behavior from different mixtures
5.1.4 Deflection capacity and recovery
In this study, deflection capacity is a concern to evaluate the self‐healing behavior in ECC material.
Deflection capacity is defined as the deflection which corresponds to the maximum bending
stress (flexural strength). And the recovery of the deflection capacity can be computed according
to the following formula:
−=deflection capacityscheme controlNormalized value ( )controlY (1)
where scheme Y stands for scheme A, scheme B and scheme C, control = scheme A
Figure 5.6 (A) provides the results of the deflection capacity for three mixtures under three
schemes. It is found that M3 (fine capsule) has a higher deflection capacity compared with M2
(coarse capsule). In most cases, the deflection capacity of M3 under each scheme exceeds that of
the corresponding parallel scheme of M2. The possible reason is that the properties of matrix
such as toughness in ECC are changed to more extent by the capsules with large size.
As shown in Figure 5.5 (B), M2 and M3 (with capsules) both clearly reveal the desirable recovery
of deflection capacity in schemes B and C. The preloaded samples from M2 and M3 (with
capsules) can reach about 70%‐90%, 80%‐90% of its control deflection capacity, respectively.
While the reference from M1 (no capsule) only arrives at 60%‐80% that of the control specimen.
Moreover, the recovery level of M3 (fine capsule) is slightly higher than M2 (coarse capsule), and
this is another support for that the large capsule size exerts more reduction of deflection
capacity.
It also can be observed from Figure 5.6, scheme B (bend to 1.0mm deflection) shows higher
values of deflection capacity as well as its recovery, compared with scheme C (bend to 1.3mm
deflection). From the self‐healing point of view, the results reflect that the self‐healing tend to
occur in small cracks.
27
01234
M1 M2 M3
Mixture designationDeflection capacity (mm)Scheme A Scheme B Scheme C
(A) Deflection capacity
0%20%40%60%80%100%120%140%
M1 M2 M3
Mixture designationNormailized deflection capacityScheme A Scheme B Scheme C
(B) Normalized deflection capacity
Figure 5.6 Comparison of deflection capacity and its recovery from different mixtures
5.1.5 Flexural strength and recovery
The recovery of flexural strength can be calculated as followed:
−=flexural strengthscheme controlNormalized value ( )controlY
(2)
where scheme Y stands for scheme A, scheme B and scheme C, control = scheme A
The recovery of flexural strength is shown in Figure 5.7. It is noticed that the flexural strength
from M1 (no capsule) under all schemes exceeds the performance of M2 and M3 (with capsules).
This indicates that the capsules inside the specimen will reduce the flexural strength of ECC
composite due to the disturbance of fiber, matrix and fiber‐matrix interface properties. When
compared the difference between the mixtures with capsules inside, the flexural strength of M3
28
(fine capsule) under three schemes is higher than that of M2 (coarse capsule), especially for
scheme C shown in Figure 5.7 (A). From the view of normalized flexural strength (Figure 5.7 (B)),
the values from M3 (fine capsule) nearly remain at a level of 100% of control value, which almost
arrives at the same recovery level of M1 (no capsule). These seem to indicate that small capsule
size attains beneficial influence on flexural strength and its recovery compared with large capsule
size.
On the other hand, it should be note that the flexural strength of M2 (with capsules) under
scheme C (bend to 1.3mm deflection) exhibits a sharp reduction, and is recovered only about
60% of the control specimen. This may be explained that the flexural strength was already
reached during the preloading stage for the case of 1.3mm deflection level. A single macrocrack
was likely to be generated and this can be further confirmed by the microscopic observation in
the later section.
02468101214161820
M1 M2 M3Mixture designationFlexural strength (MPa)Scheme A Scheme B Scheme C
(A) Flexural strength
0%20%40%60%80%100%120%140%
M1 M2 M3Mixture designationNormailized flexural strength Scheme A Scheme B Scheme C
(B) Normalized flexural strength
Figure 5.7 Comparison of flexural strength and its recovery from different mixtures
29
5.1.6 Flexural stiffness and recovery
Stiffness measurement was used to monitor the extent of self‐healing within preloaded ECC
specimens. In this research, the flexural stiffness is the equivalent slope of initial linear‐elastic
stage of flexural stress‐deflection curve as shown in Figure 5.8, stiffness and its recovery can be
calculated by the following formulas respectively:
σθ==ΔStiffness tan [MPa / mm]L (3)
−=flexural stiffnessscheme controlNormalized value ( )controlY
(4)
where scheme Y stands for scheme A, scheme B and scheme C, control = scheme A
0246810121416
012345Deflection (mm)Bending stress (MPa)
Flexural strength
Linear‐elastic stage
σ
θ 0.4*Flexural strength
ΔL
Figure 5.8 Stress‐deflection curve for stiffness calculation
Figure 5.9 presents the comparison of flexural strength and its recovery from different mixtures.
From Figure 5.9 (A), it can be seen that the flexural stiffness from M1 (no capsule) under control
scheme A is relatively lower compared with that of M2 and M3 (with capsules) under the same
scheme. This is as expected, the capsules behave as the normal aggregates, and their presence
increases the flexural stiffness of ECC. However, this trend is reversed in preloaded schemes B and
C, M1 (no capsule) has slight higher stiffness than M3 and M2 (with capsules) in most cases. This
is probably because once the cracks formed during the preloading stage, there is little resistance
to load and tend to cause a reduction of stiffness.
By comparing the recovery of flexural stiffness from Figure 5.9 (B), the level of stiffness retained
of M3 (fine capsule) is not remarkable than that of the reference M1 (no capsule), but it has an
enhancement compared with M2 (coarse capsule). It can be considered that fine capsule leads to
more reasonable recovery compared with coarse capsule since fine capsules has relatively less
influence on the matrix toughness and fiber‐matrix interface properties. This finding is also
supported by the deflection and flexural strength recovery as described earlier.
30
051015202530
M1 M2 M3Mixture designationFlexural stiffness (Mpa/mm)Scheme A Scheme B Scheme C
(A) Flexural stiffness
0%20%40%60%80%100%120%140%
M1 M2 M3Mixture designationNormailized flexural stiffness Scheme A Scheme B Scheme C
(B) Normalized flexural stiffness
Figure 5.9 Comparison of flexural stiffness and its recovery from different mixtures
31
5.2 Nano‐CT observation
5.2.1 Identification of microcracks in nano‐ CT image
As mentioned before, in order to check how the cracks develop in the interface zone between the
capsule and the cement‐based matrix, the crack pattern before and after curing was investigated
with nano‐CT scan. First of all, Figure 5.10 illustrates the components of the ECC material
represent in the nano‐CT 2D images according to the grey‐level histogram. The pore and void
inside the capsule are presented by the darkest, while the lightest corresponds to cement‐based
matrix and the medium gray is an indication of the capsule. Due to the similarity of density
between the paraffin wax and the SAP, the wax shell and the saturated SAP particles are very
difficult to be distinguished in this gray‐level image.
Void inside Capsule Matrix
Crack
Saturated SAP
Pore
Figure 5.10 Explanation of 2D nano‐CT images
On the other hand, how the cracking actually propagates (passing through or around capsules)
seems difficult to be observed from these 2D images. However, the difference inside the capsule
before and after curing can provide an insight into the moisture transportation aspect. Here it
could be assumed, that capsules after 28 days’ curing are generally divided into two types:
opened and closed. Two examples in Figure 5.11 compare the difference between two images
before and after curing at the same location of the specimen. It is clearly seen that the opened
capsule can be considered as in terms of an obvious enlargement of dark region compared with
the initial image. On the contrary there is little change in the closed one. The conclusion can be
made from the opened capsule is that the water is already released from the capsules and the
moisture transport takes place due to relative humidity (RH) gradient.
32
8 mm
“Opened”
“Closed”
(A) Before curing_28 days old After curing_56 days old
“Opened”8 mm
“Closed”
(B) Before curing_28 days old After curing_56 days old
Figure 5.11 Two examples for comparison of nano‐CT images before and after 28 days’ curing
In order to further verify whether the opening of capsules is induced by the cracking, it is
interesting to check the crack pattern by 3D reconstruction view. Figure 5.12 corresponding to
above two examples shows that, some apparent cracking patterns are revealed on the wax shell
of the opened capsule, and it can be seen from the closed capsule, there is also cracking pattern
on the shell. Therefore it cannot be established that once the cracks pass through the wax shell
the water will be released from the capsules, this may occur due to the dense structure of
paraffin wax and the smaller preloaded level. Based on above reasons, the useful information can
be inferred from this 3D view is that the capsules with the wax shell is possible to be opened by
the cracking.
33
“Opened”“Opened”
“Closed”
“Closed”
(A) (B)
Figure 5.12 3D view of the cracking pattern
5.2.2 Opening of capsules
In this section, the opening possibility of capsules will be discussed. All capsules which might be
opened by the cracks from the preloaded specimens (schemes B and C) of M2, M3 series (with
capsules) were checked based on the principle of “opened” and “closed”. Figure 5.13 gives a
summary of opened ratio of capsules, and point out that there are more than 50% capsules being
opened during the preloading stage in this research. More statistical information is summarized
in Appendix F.
0%10%20%30%40%50%60%70%80%90%100%
M2 M3Mixture designationOpened ratio of capsulesScheme B Scheme C
Figure 5.13 Summary of opened ratio of capsules in preloaded specimens
34
5.3 Light microscope and ESEM observations
5.3.1 Multiple cracking behavior
The multiple cracking behavior of ECC is one of the distinct differences from normal concrete.
Because of the fiber bridging effect, the cracks can progressively open. When the cement‐based
matrix starts to crack, the fibers will slip out the matrix. With the matrix crack extends, the fibers
can be completely pulled out from the matrix. As a result of this process, the ECC is able to take
the increasing load and forms new cracks at other sites. Figure 5.14 obtained by the light
microscope shows the typical crack pattern of multiple cracking behavior. The bridging effect of
fibers observed under ESEM is also shown in Figure 5.15.
200 μm
Figure 5.14 Typical crack pattern in ECC specimen
Figure 5.15 Fibers bridging effect in ECC specimen
35
5.3.2 Observation of interface zone
Besides, the crack path at the interface between the capsule and the cement‐based matrix was
also studied by ESEM technique. One thing is clear from the Figure 5.16: when the crack
propagates across the wax shell, the crack width appears smaller than that of developing in the
matrix. This is most probably associated with the characteristic of paraffin wax with a dense
laminated structure. In some cases, the cracks go around instead of passing through the capsules,
see Figure 5.17. This is caused by the weaker bond strength at the interface zone compared with
that of the capsule itself.
Wax shell
Matrix Crack pathCrack path
on the wax shell
Figure 5.16 Crack pattern on the wax shell
Saturated SAP
Wax shell Saturated SAP
Pass through
Around
Matrix
Figure 5.17 Interface between the capsule and the matrix
5.3.3 Observation of crack inside
Another main purpose of ESEM observation is to investigate whether there is healing product
being formed inside the cracks. Figure 5.18 explains what is “crack inside” and “crack surface”
separately. For M2 and M3 series (with capsules), the failed specimens from schemes B and C
(preloaded) were used to prepare samples for ESEM observation. Unfortunately, no apparent
healing product for instance calcium carbonate crystals with prismatic shape was found within
the cracks. However, as shown in Figure 5.19, calcium hydroxide in the shape of hexagonal
platelet can be observed near the crack surface.
36
Crack
Under ESEM
Cut Crack inside:
Figure 5.18 Illustration of crack inside and crack surface
Figure 5.19 Preloaded sample after 28 days’ curing
5.3.4 Observation of crack surface
Finally, Semi quantitative chemical characterization by EDX of the preloaded ECC specimens using
an ESEM was carried out to reveal the composition inside the microcracks. One sample cut from
cross‐section of the crack and another sample with normal cross‐section (non‐crack) from the
same specimen as a reference for comparison purpose (Figures 5.20 and 5.21). Table 5.1
summarizes the results of weight percentage of element calcium at the specified location. More
detailed information is listed in Appendix G. It can be seen that the calcium content at the crack
surface accounts for more percentage (10% larger) than that of the calcium at the non‐crack
surface. Another evidence is of the elemental mapping (MAPing) by EDX analysis as shown in
Table 5.1 and Figure 5.22. As mentioned in Chapter 2.2.2, the healing products could be
presented in terms of three substances, including calcium carbonate (CaCO 3), calcium hydroxide
(Ca(OH) 2) and calcium silicate hydrate (CSH). In this research, it could be explained that the
specimen was cured in sealed condition, the available carbon dioxide (CO 2) dissolved in water is
little, such that it is difficult to form the calcium carbonate. Hence, the high concentration of
calcium could indicate that the healing product is probably hydration product, such as calcium
hydroxide or calcium silicate hydrate. Under ESEM
Cut
Crack surface:
37
4 3
1
2
Figure 5.20 ESEM images of crack surface
6 5
Figure 5.21 ESEM images of non‐ crack surface (reference)
Table 5.1 Summary of Calcium (Ca) weight percentage
Sample Location Ca Wt% Average Wt%
1 42.05
Crack 2 42.51
surface 3 48.87
4 38.66 43.02
Non‐crack 5 33.06
surface (reference) 6 34.68 33.87
(A) Analysis area (B) Distribution of the element calcium
Figure 5.22 MAPing analysis on the crack surface
38
5.4 Discussion of potential of self‐healing in ECC
5.4.1 Influencing factors
In this research, six mechanical indicators are involved: deflection capacity, flexural strength,
stiffness and their normalized values. Two factors (capsule content and capsule size) were
designed to explore the self‐healing potential of ECC material, and they are discussed as below,
vii. Capsule content (no capsule & with capsules)
As mentioned in Chapter 4.2.2, the ECC specimens without capsule (M1) and with capsules (M2
and M3) were taken into account to investigate how the capsules influence the healing properties.
When comparing the magnitude of deflection capacity between M1 (without capsule) and M2
and M3 (with capsules) at each parallel scheme, the reference M1 (without capsule) always
exceeds the performance of M2 and M3. The same trend also occurs in most cases of the flexural
strength and stiffness. This can be explained that, the capsule itself behaves as the normal
aggregate, and the presence of aggregates disturbs the matrix properties especially the fiber
distribution of the ECC composite and has negative influences on the mechanical properties.
However, from the results of recovered mechanical properties, it is noticed that the deflection
capacity from the specimens with capsules (M2 and M3) can recover about 65%‐95% from
control specimens, which is higher than that of the reference specimens without capsule (M1).
While, there is a slight reduction from aspect of flexural strength recovery expect M2 (coarse
capsule) under scheme 3 (bend to 1.3mm deflection) that is with a significant drop. In this case,
the specimen was already preloaded to a single macrocrack and offered little resistance to reload.
In addition, approximately 60% in average of flexural stiffness is regained for specimens with
capsules (M2 and M3), this recovery level is much lower (about 20%) than that of the reference
without capsule (M1). Hence, there is an enhanced recovery of deflection capacity but a reduced
recovery of flexural strength and stiffness more or less, and it may be explained that the water
released from the capsules is partially contributed to promote the self‐healing of cracks in this
research. Further discussion of healing efficiency will be presented in the next paragraph.
viii. Capsule size (coarse capsule & fine capsule)
From the literature study, it is suggested that the size range of microcapsule is from 50nm to
2000μor larger [19]. But manufacture of microcapsules is difficult in this research due to no
available rolling machine at that time. In such a case, M2 with capsule diameter of 8mm (coarse
capsule) and M3 with capsule diameter of 5mm (fine capsule) were designed in this research, to
compare the influence of capsule size on the performance of ECC. It is clear from Figures 5.5, 5.6
and 5.8 that small capsule size shows higher deflection capacity, flexural strength, stiffness as well
as their recovery for almost all the cases compared with larger capsule size. For the crack pattern,
it can be seen in Figure 5.22 that when the capsules are made in large size, once the preloading is
slightly larger than a certain level, the ECC specimen tends to generate a single macrocrack and
less new cracks are formed after reloading. In contrast, the specimen with small capsule size has m
39
more even distribution of multiple cracks. Thus small capsule size has more benefits on
mechanical properties and mechanical recovery and cracking pattern since it changes matrix and
fiber‐matrix properties to a less extent.
Macrocrack
Coarse_1.0mm Fine_1.0mm Coarse_1.3mm Fine_1.3mm
(A) 28 days old (before reloading)
Coarse_1.0mm Fine_1.0mm Coarse_1.3mm Fine_1.3mm
(B) 56 days old (after reloading)
Figure 5.22 Comparison of crack pattern for M2 (coarse capsules) and M3 (fine capsules) under
scheme B (preloaded to 1.0mm) and scheme C (preloaded to 1.3mm)
5.4.2 Healing efficiency
Although it was indicated that the recovery level of deflection capacity was enhanced while the
recovery of flexural strength and stiffness had a reduction more or less compared with the
reference. Even though the high concentration of calcium on crack surface was found via EDX
analysis, based on this, it could be considered that the healing product was presented in terms of
calcium hydroxide or calcium silicate hydrate. But no apparent healed crack was observed under
ESEM, maybe since the product was too little and the healing efficiency was low. Here the
possible reasons of that are discussed in this section.
The crucial reason of low healing efficiency is that the amount of available water is not sufficient.
Here the healing efficiency can be explained in terms of moisture transportation in the preloaded
ECC specimens. In this study, when the capsule is opened, the moisture is firstly desorbed form
the SAP particles and moves to the crack surface. It can be assumed that the crack is immediately
in high humidity environment (RH 99%), then the moisture movement takes place by diffusion
40
which is driven by the humidity gradient. According to Huang [24], the numerical simulation was
established to show the relationship between the distance of moisture transport into the mortar
and the time, see Figure 5.23. It indicates that the moisture will move far away from the crack
surfaces with the increase of curing time. Similar to that, in this study it was confirmed that
moisture had moved to the crack surface, but when the water supply is limited, the moisture
content left on the crack surfaces is too little to produce more healing products in short curing
period. Thus the sufficient water supplement is essential for desirable healing efficiency. The
amount of water for further hydration of unhydrated cement will be investigated in the further
research.
Since the moisture transportation can be considered as the function of time, the curing time
becomes another factor influencing the healing efficiency. However, when the amount of water is
sufficient, the curing time is not a decisive factor for self‐healing.
0.5h 0.5h1h1h2h 2h
crack8h8h
05101520253035404550
0 2 04 06 08 0 1Z (mm)
00X (mm)
Figure 5.23 Water penetration depth in cracked cementitious mortar during wetting process [24]
5.4.3 Short summary
Back to the core question, whether the self‐healing occurs in the modified ECC material? It can be
viewed from two aspects. On the one hand, both capsule sizes (coarse capsule with 8mm in
diameter and fine capsule with 5mm in diameter) used in this research, are relatively large from
the view of micro scale. The capsules act as the aggregates embedded into ECC, and thus their
influence on the mechanical properties of ECC cannot be ignored. As discussed above, the
capsule itself has a negative effect on the micromechanics. However on the other hand, it is
expected that the capsules can provide water available to heal the cracks. This could enhance the
mechanical properties. Therefore as an access of the water supply, the capsule exerts a positive
effect on mechanical properties. These two aspects seem to be contradictory, that could explain
why the enhancement as well as the reduction were presented in this research. From the view of
healing product, it could be said the self‐healing occur due to a little healing product presented as
the high concentration of calcium, but this behavior is not apparent. It can be inferred that
optimize the capsule size will be contributed to minimize the negative effect and provide
sufficient water will promote more effective self‐healing behavior.
41
6 GENERAL CONCLUSIONS AND RECOMMENDATIONS
6.1 General conclusions
In this thesis, the self‐healing potential of the ECC material by means of available water released
from the capsules containing the water saturated SAP has been investigated. Based on the
experimental results of mechanical test and microscopic observation, the following conclusions
can be drawn,
The mass of water is almost not influenced by the curing time when the paraffin wax is used as
sealing material. Due to its better capacity of water storage compared to that of epoxy‐cement
paste, the paraffin wax was preferred as the sealing material in this research.
Recovery of mechanical properties are regarded as the indicators of the self‐healing efficiency.
The recovery of deflection capacity, flexural strength and stiffness were examined in this research.
The recovered deflection capacity was enhanced while the improvement on flexural strength and
stiffness were rarely shown. It could be considered that the self‐healing efficiency was not
remarkable, since the mechanical properties were not significantly improved.
The ECC specimens with the small capsule size of 5mm in diameter have preferable performance
in mechanical properties and their recovery as well as in crack pattern. Similar to normal concrete,
the mechanical properties of a cementitious material are highly influenced by the capsule size.
Based on the results, the smaller capsule size is conducive to the self‐healing behavior.
Under nano‐CT technique and ESEM observation, it could be confirmed that moisture
transportation took place in more than half of capsules. Sufficient water supplement plays a
major role on promoting the self‐healing process. In this research, the low efficiency of healing
was thus mainly attributed to the insufficient water supply.
The observed healing product probably is of the hydration product such as calcium hydroxide or
calcium silicate hydrate. This finding indicates that the self‐healing phenomenon probably
occurred. Unfortunately no apparent healed crack was observed, it could be concluded that the
cracks were likely to undergo the self‐healing process, but it is not a very effective healing in this
research.
42
6.2 Recommendations
Several feasible improvements and further research to realize the self‐healing of cementitious
materials in concrete repair system can be given as following,
i. Alternative sealing material
As discussed before, a suitable sealing material is not only capable of storing water, but also has
high bond strength at the interface between the sealing material and the cementitious matrix.
The capsule is considered as the weakest element in the composite and the interface bond of
sealing material directly determines on the crack pattern, which influences the release of healing
agent. Thus selection of a proper sealing material is crucial as a basis for further study.
ii. Alternative water reservoir
Liapor particle is a promising candidate for carrying water instead of SAP. It has a high water
absorption capacity (30%‐40% by weight) and the particle is roughly spherical with a diameter of
1‐10mm. Therefore, it can be made into a small capsule. As known, the capsule size significantly
influences the mechanical properties of the composite. Moreover, control of capsule size is
essential to the uniform distribution of capsules and the probability of capsule opening. The
application of liapor needs to be studied further.
iii. Encapsulation procedure
From the point of view of producing the capsules, It is important to find an effective method to
manufacture capsules quickly and easily. A rolling machine (Figure 6.1) is suggested in future
study. This rotary barrel tumbler results in economic, quiet and efficient operation for tumbling
small parts [25].
Figure 6.1 Rotary barrel tumbler
iv. Effective healing agent
Currently, sodium silicate used as the microencapsulated healing agent was embedded in a
concrete mixture [26]. The sodium silicate reacts with the calcium hydroxide and forms a gel‐like
material (calcium‐silica‐hydrate) that will heal the crack and block the pores. The advantage is
that the gel hardens in about one week and the recovery of the strength can reach 26 percent of
its original strength. It is believed that a more effective healing agent can contribute to the
self‐healing process.
43
REFERENCES
[1] M. G. Grantham, Diagnosis, inspection, testing and repair of reinforced concrete structures, M.
G. Associates, 1999, p.3‐5
[2] Li V.C., Kanda T., Engineered cementitious composites for structural application, ASCE J.
Materials in Civil Engineering, 1998, Vol.10, No.2, p.66‐69
[3] Zhou,J. Qian, S. Ye, G. Breguel, K van and Li, V.C., Development of Engineered Cementitious
Composites with Limestone Powder and Blast Furnace Slag, submitted to Materials and
Stuctures, 2009
[4] Huan He, Zhangqi Guo, Piet stroeven, Martijn Stroeven, Lambertus Johannes Sluys,
Self‐healing capacity of concrete – computer simulation study of unhydrated cement structure,
Image Anal Stereol 2007, p.137‐143
[5] Tsuji M., Okuyama A., Enoki K. and Suksawang S., Development of new concrete admixture
preventing from leakage of water through cracks, JCA Proc. Of Cement & Concrete 52, 1998,
p.418‐ 423
[6] Li V.C., Engineered Cementitious Composites (ECC) – Tailored Composites though
Micromechanical Modeling, Fiber Reinforced Concrete: Present and the Future, Canadian
Society of Civil Engineers, 1998, p.64‐97
[7] Yingzi Yang, Michael D. Lepech, En‐Hua Yang, Victor C. Li, Autogenous healing of engineered
cementitious composites under wet‐dry cycles, Cement and Concrete Research 39, 2009, p.
382‐ 390
[8] Mustafa Sahmaran, Victor C. Li, Durability properties of micro‐cracked ECC containing high
volumes fly ash, Cement and Concrete Research 39, 2009, p. 1033‐1043
[9] Toshiro Kamada, Victor C. Li, The effects of surface preparation on the fracture behavior of
ECC/concrete repair system, Cement and Concrete Composites 22, 2000, p.423‐431
[10] Romildo D. Toledo Filho, Eugenia F. Silva, Anne N.M. Lopes, Effect of super absorbent polymers
(SAP) on the workability of concrete, RILEM‐TC SAP Chapter 5, 2009
[11] Guang Ye, Klaas van Breugel, Effect of SAP on the Harding Process of Binder Paste and
Microstructure Development in Concrete (Porosity, morphology, connectivity), RILEM‐TC SAP
Chapter 6, 2009
[12] S. van der Zwaag (ed), Self healing materials: an alternative approach to 20 centuries of
material science, 2007
[13] S. Qian, J. Zhou, M.R. de Rooij, E. Schlangen, G. Ye, K. van Breugel, Self‐healing behavior of
strain hardening cementitious composites incorporating local waste materials, Cement and
Concrete Composites 31, 2009, p.613‐621
[14] E. Schlangen, Fracture Mechanics, CT5146 Lecture Notes, Delft University of Technology, 2007
[15] A. Neville, Properties of Concrete, 1995, p. 328
[16] A. Neville, Autogeous Healing‐ A concrete Miracle, Concrete International, 2002, p. 76‐82
[17] Nynke Ter Heide, Crack healing in hydrating concrete, Master thesis, Delft University of
Technology, 2005
[18] Dong Yang Wu, Sam Meure, David Solomon, Self‐healing polymeric materials: A review of
recent developments, Prog. Polym. Sci. 33, 2008, p.479‐522
[19] R. Arshady (ed), Microspheres, microcapsules and liposome, Vol.1: preparation and chemical
44
45
application, Citus Books, 1999
[20] Victor C. Li, Yun Mook Lim, Yin‐Wen Chan, Feasibility study of a passive smart self‐healing
cementitious composite, Compisites part B, 1998, p.819‐827
[21] Henk M. Jonkers, Arjan Thijssen, Gerard Muyzer, Oguzhan Copuroglu, Erik Schlangen,
Application of bacteria as self‐healing agent for the development of sustainable concrete,
Ecological Engineering, 2009
[22] Haoliang Huang, Guang Ye, Klaas van Breugel, Numerical simulation on moisture transport in
cracked cement‐basis materials, 2010
[23] McMASTER ‐CARR, http://www.mcmaster.com/#barrel ‐tumblers/=7yvslm
[24] Susan Wilson, Cost effective self healing concrete developed at URI, 2010
APPENDIX A Mix proportion of epoxy‐cement material
Mix proportion
CEM I Epoxy Hardener Epoxy/hardener Water Water/cement Epoxy/cement
52.5N [g] [g] [g] ratio [g] ratio ratio
20 0.6135 0.7730 1:1.26 5.6135 0.300 0.05
46
APPENDIX B Mix design of ECC
Material property and mix proportion of M1
Density Mix proportion Weight Volume Diameter Component
[g/cm3] (by weight) [g] [cm3] [mm]
Portland cement 3.15 1 550 175
BFS 2.85 1.2 660 232
Limestone
powder 2.70 2 1100 407
Water 1.00 1.092 601 601
Super‐plasticizer 1.17 0.030 16.5 14
PVA fiber 1.30 0.02
(by volume) 37.1 29
Saturated SAP
inside capsule 0
(by CEM weight) 0 0
Number of
capsules 0
(0% by volume)
Capsule size 0
Water/powder
ratioMaterial property and mix proportion of M2
Density Mix proportion Weight Volume Diameter Component
[g/cm3] (by weight) [g] [cm3] [mm]
Portland cement 3.15 1 550 175
BFS 2.85 1.2 660 232
0.26
Total 2964 1457
Limestone
powder 2.70 2 1100 407
Water 1.00 1.092 601 601
Super‐plasticizer 1.17 0.030 16.5 14
PVA fiber 1.30 0.02
(by volume) 37.1 29
Saturated SAP
inside capsule 2%
(by CEM weight) 11.0 6.7
Number of
capsules 25
(0.46% by volume)
Capsule size 8
Water/powder
ratio
0.26
Total 2975 1465
47
Material property and mix proportion of M3
Density Mix proportion Weight Volume Diameter Component
[g/cm3] (by weight) [g] [cm3] [mm]
Portland cement 3.15 1 550 175
BFS 2.85 1.2 660 232
Limestone
powder 2.70 2 1100 407
Water 1.00 1.092 601 601
Super‐plasticizer 1.17 0.030 16.5 14
PVA fiber 1.30 0.02
(by volume) 37.1 29
Saturated SAP
inside capsule 2%
(by CEM weight) 11.0 3.3
Number of
capsules 50
(0.23% by volume)
Capsule size 5
Water/powder
ratio
0.26
Total 2975 1461
48
APPENDIX C Converting the weight of SAP into the number of capsules
Basically, capsules in the same series were controlled to be the similar size according to different
requirements of capsule size (M2: diameter of 8mm, M3: diameter of 5mm). During the making
process of capsules, saturated SAP particles were shaped into a ball by a thin cover of cement, and
less cement was remained especially after curing them in water for 7 days. On the other hand, dry
SAP powder can absorb amounts of free water many times than its own weight. Based on above
two points, the weight of cement cover and SAP particles can be assumed to be neglected, the
weight of ball is therefore considered as the weight of water only. As a result, the calculation of
“saturated SAP” with 2% weight ratio of cement in Appendix B, means the amount of free water
needed for healing, which is represented by Wwater.
This section interprets that how to convert the amount of free water into the number of capsules.
First, the saturated SAP particles after shaping by cement cover and 7 days’ curing in water is
called a” ball”, and the weight of a ball actually is the weight of water as explained above. In order
to investigate the weight of a ball, 100 balls were divided into 10 groups and each group included
10 balls. Each group of balls was weighed by electric balance, and then the average weight of one
ball (W
ball) can be evaluated. Due to accidental errors, it was proposed that the number of
capsules for different mixtures was multiplied by a magnification factor 1.1, then it follows that:
=water
capsules
ballWN1 . 1W
The following table gives the experimental results and calculation of the number of capsules for
two different mixtures.
Experimental results and calculation of the number of capsules
Capsule size M2: diameter of 8mm M3: diameter of 5mm
Group Weight of saturated SAP [g] Weight of saturated SAP [g]
1 0.8694 0,4101
2 0.8793 0,4310
3 0.7360 0,3874
4 0.8945 0,4363
5 0.8121 0,4102
6 0.7475 0,3976
7 0.6763 0,4227
8 0.7698 0,3830
9 0.7271 0,4039
10 1.0014 0,4701
Average 0.8113 0,4152
Wball 0.0811 0,0416
Wwater 11.0 g 11.0 g
Ncapsules 149.14≈150 291.40≈300
Ncapsules per specimen 150/ 6=25 300/ 6=50
49
50
APPENDIX D Three‐point bending test configuration and bending stress calculation
Bending stress at mid‐span: σ== = = ⋅ =2
22M( F / 2 ) * L ( F / 2 ) * 5 50.002578125F[kN mm ] 2.578125F[MPa]W ( 1/6)*b*h ( 1/6)*40*40
F
ASpecimen
M=F/2*LL=55mm
110mm b=40mm
h=40mm
Test set‐up Section A‐A
Moment diagramA
160mm
APPENDIX E Bending stress‐deflection curves
M1a (no capsule_28days old)
0246810121416
01234
Deflection (mm)Bending stress (MPa)M1a‐1
M1a‐2M1b (preloaded to 1.0mm_28days old)
0246810121416
01234
Deflection (mm)Bending stress (MPa)M1b‐1
M1b‐2M1c (preloaded to 1.3mm_28days old)
0246810121416
01234
Deflection (mm)Bending stress (MPa)M1c‐1
M1c‐2
M1c‐3
M1c‐4
M1b (no capsule_56days old)
0246810121416
01234
Deflection (mm)Bending stress (MPa)M1b‐1
M1b‐2M1c (no capsule_56days old)
0246810121416
01234
Deflection (mm)Bending stress (MPa)M1c‐1
M1c‐2
M1c‐3
M1c‐4
51
M2a (Capsule size of 8mm_28days old)
0246810121416
01234
Deflection (mm)Bending stress (MPa)M2a‐1
M2a‐2
M2a‐3M2b (preloaded to 1.0mm_28days old)
0246810121416
01234
Deflection (mm)Bending stress (MPa)M2b‐1
M2b‐2
M2b‐3
M2b‐4M2c (preloaded to 1.3mm_28days old)
0246810121416
01234
Deflection (mm)Bending stress (MPa)M2c‐1
M2c‐2
M2b (Capsule size of 8mm_56days old)
0246810121416
01234
Deflection (mm)Bending stress (MPa)M2b‐1
M2b‐2
M2b‐3
M2b‐4M2c (Capsule size of 8mm_56days old)
0246810121416
01234
Deflection (mm)Bending stress (MPa)M2c‐1
M2c‐2
52
M3a (Capsule size of 5mm_28days old)
0246810121416
01234
Deflection (mm)Bending stress (MPa)M3a‐1
M3a‐2
M3a‐3
M3a‐4M3b (preloaded to 1.0mm_28days old)
0246810121416
01234
Deflection (mm)Bending stress (MPa)M3b‐1
M3b‐2
M3b‐3M3c (preloaded to 1.3mm_28days old)
0246810121416
01234
Deflection (mm)Bending stress (MPa)M3c‐1
M3c‐2
M3c‐3
M3b (Capsule size of 5mm_56days old)
0246810121416
01234
Deflection (mm)Bending stress (MPa)M3b‐1
M3b‐2
M3b‐3M3c (Capsule size of 5mm_56days old)
0246810121416
01234
Deflection (mm)Bending stress (MPa)M3c‐1
M3c‐2
M3c‐3
53
APPENDIX F Results of opened ratio of capsules under nano‐CT observation
Number of opened capsules Number of closed capsules Sample Front Right Average Front Right Average Total capsules Opened ratio
M2b‐ 1 6 6 6 3 4 4 10 63%
M2b‐ 2 8 7 8 4 4 4 12 65%
M2b‐ 3 4 4 4 6 6 6 10 40%
M2c‐1 5 5 5 3 3 3 8 63%
M2c‐2 4 4 4 4 4 4 8 50%
M2c‐3 5 5 5 4 3 4 9 59%
M3b‐1 5 5 5 6 6 6 11 45%
M3b‐ 2 6 6 6 5 5 5 11 55%
M3b‐ 3 8 7 8 7 6 7 14 54%
M3c‐1 12 11 12 7 6 7 18 64%
M3c‐2 9 7 8 5 7 6 14 57%
M3c‐3 9 9 9 8 9 9 18 51%
54
APPENDIX G Results of EDX of the preloaded ECC specimens
Crack surface‐‐‐Location 1
Element Wt % At % K‐ Ratio Z A F
C 5.14 9.53 0.0156 1.0500 0.2884 1.0010
O 42.74 59.53 0.0625 1.0323 0.1417 1.0001
Mg 1.25 1.15 0.0056 0.9902 0.4497 1.0028
Al 1.64 1.35 0.0093 0.9610 0.5870 1.0050
Si 4.82 3.82 0.0337 0.9890 0.7031 1.0064
S 0.63 0.44 0.0053 0.9767 0.8478 1.0206
K 0.66 0.37 0.0066 0.9394 0.9724 1.0991
Ca 42.05 23.38 0.3994 0.9615 0.9873 1.0005
Fe 1.07 0.43 0.0090 0.8742 0.9643 1.0000
Total 100.00 100.00
Crack surface‐‐‐Location 2
Element Wt % At % K‐ Ratio Z A F
C 6.24 11.58 0.0186 1.0501 0.2831 1.0010
O 40.61 56.61 0.0581 1.0324 0.1385 1.0001
Mg 1.34 1.23 0.0061 0.9903 0.4554 1.0030
Al 1.82 1.50 0.0104 0.9612 0.5920 1.0052
Si 5.35 4.25 0.0376 0.9891 0.7058 1.0065
S 0.67 0.47 0.0057 0.9768 0.8454 1.0206
K 0.70 0.40 0.0070 0.9396 0.9710 1.0988
Ca 42.51 23.65 0.4032 0.9616 0.9860 1.0003
Fe 0.77 0.31 0.0064 0.8744 0.9636 1.0000
Total 100.00 100.00 Crack surface‐‐‐Location 3
Element Wt % At % K‐ Ratio Z A F
C 4.85 9.51 0.0153 1.0546 0.2989 1.0011
O 37.07 54.56 0.0488 1.0368 0.1268 1.0001
Mg 1.13 1.10 0.0050 0.9944 0.4456 1.0029
Al 1.47 1.29 0.0083 0.9651 0.5836 1.0052
Si 4.17 3.50 0.0293 0.9932 0.7015 1.0074
S 0.50 0.37 0.0043 0.9820 0.8513 1.0241
K 0.87 0.53 0.0090 0.9440 0.9743 1.1156
Ca 48.87 28.71 0.4665 0.9660 0.9877 1.0004
Fe 1.06 0.45 0.0089 0.8787 0.9564 1.0000
Total 100.00 100.00
Crack surface‐‐‐Location 4
Element Wt % At % K‐ Ratio Z A F
C 7.12 12.80 0.0209 1.0476 0.2794 1.0009
O 42.87 57.88 0.0649 1.0300 0.1470 1.0001
Mg 1.60 1.42 0.0073 0.9880 0.4600 1.0030
Al 2.01 1.61 0.0115 0.9589 0.5944 1.0050
Si 5.75 4.42 0.0403 0.9868 0.7061 1.0059
S 0.54 0.37 0.0045 0.9739 0.8427 1.0187
K 0.67 0.37 0.0066 0.9371 0.9704 1.0898
Ca 38.66 20.83 0.3656 0.9591 0.9858 1.0003
Fe 0.77 0.30 0.0065 0.8719 0.9681 1.0000
Total 100.00 100.00
55
Non‐crack surface‐‐‐Location 5
Element Wt % At % K‐ Ratio Z A F
C 9.01 15.36 0.0264 1.0431 0.2808 1.0008
O 46.64 59.67 0.0766 1.0256 0.1600 1.0001
Mg 1.60 1.34 0.0073 0.9839 0.4632 1.0028
Al 2.11 1.60 0.0121 0.9549 0.5978 1.0047
Si 5.85 4.26 0.0409 0.9828 0.7084 1.0051
S 0.48 0.31 0.0040 0.9688 0.8435 1.0162
K 0.72 0.38 0.0070 0.9326 0.9717 1.0775
Ca 33.06 16.88 0.3115 0.9546 0.9867 1.0003
Fe 0.54 0.20 0.0045 0.8675 0.9746 1.0000
Total 100.00 100.00 Non‐crack surface‐‐‐Location 6
Element Wt % At % K‐ Ratio Z A F
C 6.65 11.70 0.0193 1.0457 0.2778 1.0009
O 45.88 60.62 0.0807 1.0281 0.1711 1.0002
Mg 1.71 1.48 0.0085 0.9863 0.5037 1.0032
Al 2.24 1.75 0.0137 0.9572 0.6352 1.0053
Si 6.82 5.13 0.0499 0.9851 0.7384 1.0056
S 0.54 0.35 0.0045 0.9717 0.8558 1.0174
K 0.64 0.35 0.0063 0.9352 0.9716 1.0823
Ca 34.68 18.29 0.3273 0.9572 0.9857 1.0004
Fe 0.85 0.32 0.0072 0.8700 0.9763 1.0000
Total 100.00 100.00
56
Copyright Notice
© Licențiada.org respectă drepturile de proprietate intelectuală și așteaptă ca toți utilizatorii să facă același lucru. Dacă consideri că un conținut de pe site încalcă drepturile tale de autor, te rugăm să trimiți o notificare DMCA.
Acest articol: Selfhealing of Engineered Cementitious [603846] (ID: 603846)
Dacă considerați că acest conținut vă încalcă drepturile de autor, vă rugăm să depuneți o cerere pe pagina noastră Copyright Takedown.