The Feasibility of Using

Activated Paper Recycled Minerals

(Top-Crete) as Partial Cement
Replacement

ii

Delft University of Technology

Faculty of Civil Engineering and Geosciences
Microlab

Feasibility of Using Activated Paper Recycled minerals

(Top-Crete) as Partial Cement Replacement
Master thesis by Madiha Ahammout

Delft, March 2011 January 2011

Graduation Committee:

Prof. Dr. Ir. K. van Breugel

Dr. Msc. Guang Ye
Ir. S.A.A.M Fennis Huijben
Ir. L.J.M. Houben
B.M. Piscaer
iii

Indeed, ease comes with every hardship

Verily, with this hardship there is ease
[The Holy Quran, Surah ash-Sharh 94:5 and 94:6]

iv

Acknowledgements
All praise and thanks to my lord Allah, the source of all knowledge and wisdom, who guided me to
start this journey and gave me the strength and fortitude to complete this research work. He also
bestowed me with the potential and the ability to contribute a drop of information to the existing
ocean of scientific knowledge.
There are a number of people who helped me to achieve my goal through their support and
encouragement.
I would like to thank my dear mother for her constant support, motivation, patience and
understanding throughout the entire research. She was a guiding light in the most difficult and
stressful times.
I would like to extend my sincere gratitude to my daily supervisor Dr. MSc. Guang Ye not only for his
continuous supervision and suggestions, but also for the freedom he gave me to perform this project
as I desired. I would also like to thank my other committee members, Prof. Dr. Ir. K. van Breugel and
Ir. S.A.A.M. Fennis for their availability and comments during my project.
I am also grateful to Piscaer Boudewijn and Luis Miguel Ordoez for their suggestions throughout my
research.
I appreciate the support provided by the technicians Ger Nagtegaal, Arjen Thijssen and John van de
Berg during the measurements. Arjen Thijssen also taught me to use the S.M.A.R.T principle (Specific,
Measurable, Achievable, Realistic and Timely) in my project. Well, I hope I worked as SMART as
possible. I also want to thank the other employees at the university, especially Nynke verhulst from
the secretary.
Finally, I greatly appreciate my colleagues (PHD students) from the section and lab members for their
assistance, encouragement and friendship, which have been instrumental in my fulfillment of this
research work. I also want to give special thanks to Natascha Poeran, who has reviewed my work.

Whoever Trusts in Allah, will find Him sufficient.

Verily, Allah will accomplish his purpose.
[The Holy Quran, Surah Al Talaq 65:3]

vi

Abstract
Global warming, which is mainly caused by mankind with the emissions of greenhouse gases, is
nowadays one of the major topics in the world. The reduction of these greenhouse gas emissions has
become a primary focus of the environmental organizations and the government in many countries.
Since Portland cement production is accompanied by a significant CO2 emission, which is a principle
greenhouse gas, it is useful to look for ecological alternatives of cement. Portland cement is normally
replaced with fly ash or blast furnace slag cement, which are both byproducts from other industrial
waste. The sources of these byproducts are limited. Therefore, seeking for other alternatives of
Portland cement is significant from both economical and environmental point of view. In this
research the use of thermally activated paper sludge minerals as partial Portland cement
replacement is investigated. The replacement of cement with paper sludge minerals is proposed to
reduce the environmental burdens that are associated with the concrete productions and with the
disposal of paper sludge waste. By calcination and dehydration of paper sludge, which is a by-product
of recycled paper waste, a non-toxic mineral with a highly pozzolanic activity is formed. Besides the
pozzolanic activity this mineral also shows hydraulic potential. Therefore, the production and usage
of thermally activated paper sludge minerals has attracted increasing interest due to its
environmental advantages and excellent cementitious potential.
The use of thermally activated paper sludge minerals (known as Top-Crete) as a partial cement
replacement is a relatively new research area. Therefore, the main goal of this study is to examine
the feasibility of using Top-Crete (TC) as a partial cement replacement and especially as a leverage to
use more supplementary cementing materials such as fly-ash. The cement paste and mortar mixtures
have been composed with TC ranging from 0% to 15%, with water to binder ratio of 0.5.
Furthermore, fly ash was used with replacement levels of 35% and 70%
TC influences the characteristics of paste and mortar due to both physical and chemical effects.
Several experiments were performed to characterize the physical properties of TC, such as the
specific surface area and the particle size distribution. Furthermore, the microstructure of TC
particles was studied by scanning electron microscopy (SEM). The results and images showed that TC
particles are composed of tiny particles which are agglomerated into rigid structures of different
shapes and sizes. The particle size distribution of TC is slightly finer than that of PC and FA. TC has a
specific surface area of 7.2262 m2/g, which is about 27 times higher than PC. The ignition loss of TC is
21%, which mainly arises from the release of carbon dioxide by the thermal decomposition of the
large CaCO3 content. Knowledge on these physical properties will increase the ability to explain the
influence of TC on various properties of blended cement pastes.
Experiments have been performed on the cement paste and mortar mixtures to gain more insight in
the effect of TC on the properties of blended cement pastes and mortars. The properties include
workability, setting time, bending strength and compressive strength. These properties are tested
according to the procedure outlined in the European standards. Further, the hydration heat and
degree of hydration were studied by isothermal calorimeter. Mercury intrusion porosimetry was
vii

used to measure the pore size and pore size distribution. Finally, the chemically bound water and CH
content were determined with the thermo gravimetric analysis. The results showed that the addition
of TC led to a decrease in workability, lower setting time, lower bending strength and lower
compressive strength. These effects are stronger for higher levels of replacement. The chemically
bound water of the various specimens coincides within the same range of 12%, even though each
mixture has a different amount of TC. Therefore, the great water demand of TC is mainly attributed
to its large specific surface area and high pozzolanic activity. The hydration heat slightly increases
with increase of the TC content. Finally, the results from MIP tests show that the blended pastes
incorporating TC create a denser microstructure with a lower total porosity and finer pore size
distribution
The second objective was to define the amount of free lime in cement pastes blended with TC before
and after hydration to ensure the safeness of structures. Knowledge of the amount of free lime in
concrete is essential because of its influence on the properties of the material. The European
standards have set the requirement that the amount of free lime should be less than 3%. The
glycerol extraction test offers a very rapid and simple method to determine the free lime in
cementitious materials. However, the data obtained from this test demonstrates that the glycerol
extraction method is not an accurate and absolute measure of the free lime content. It is concluded
that methods other than glycerol extraction should be employed to measure the free lime content in
various blends.
The last objective of this study was to define the contribution of TC to the autogenous and chemical
shrinkage of the cement paste. The chemical shrinkage was measured according to procedure
outlined by Geiker and the sealed corrugated tubes were used to measure the autogenous shrinkage
of the specimen. The results indicate that partial replacement of cement with TC decreases the
magnitude of the chemical shrinkage with 25%, particularly at higher replacement levels. The
decrease in chemical shrinkage is a direct consequence of the combined effect of the diluted cement
hydration and the increased pozzolanic reaction of TC. The TC hydrates compensate for the chemical
shrinkage. It was also observed that the replacement of PC with 10% TC substantially increases the
autogenous shrinkage of the blended cement paste compared to that of the control paste. This might
be attributed to the acceleration of the PC hydration, the highly pozzolanic reaction of TC and the
refinement of porosity by TC. The first two contribute to the removal of water from the specimen
which results in self-desiccation of the specimen. While the refinement of the porosity results in the
increase of the capillary tension.
The increase in FA content from 35% to 70% does not result in the improvement of the properties of
TC blended cement pastes. Therefore, the higher FA content combined with the use of TC does not
compensate for the large PC deficiency. Hence, TC cannot be used as a leverage to use more
supplementary cementing materials.
Overall, from this research, it has been found that a certain percentage of TC can act as a
replacement of cement without significant loss in strength and with a considerable shorter setting
time. Replacing cement with 5 % TC in blended pastes is advantageous because of (i) lower setting
time, (ii) the increase in chemical resistance due to finer pore structure, (iii) the lower chemical
shrinkage avoiding micro cracks in the early age of hydration and (iv) the increase in environmental
benefits.
viii

Table of contents
Acknowledgements ...................................................................................................................... v
Abstract ...................................................................................................................................... vii
List of Abbreviations................................................................................................................... xiv
Chapter 1 Introduction
1.1 Background .................................................................................................................................... 1
1.2 Problem description ...................................................................................................................... 1
1.3 Goal, objectives and scope ............................................................................................................ 2
1.4 Outline ........................................................................................................................................... 3
Chapter 2. Literature Study
2.1 Portland cement ............................................................................................................................ 5
2.1.1 Origin .................................................................................................................................. 5
2.1.2 Clinkers ............................................................................................................................... 5
2.1.3 Hydration Process............................................................................................................... 6
2.1.3.1 Chemical reactions of Portland clinkers ...................................................................... 6
2.1.3.2 Hydration products ..................................................................................................... 7
2.1.3.3 Microstructure development of Portland cement ...................................................... 9
2.1.4 Particle morphology of PC ................................................................................................ 10
2.2 Fly ash .......................................................................................................................................... 11
2.2.1 Origin ................................................................................................................................ 11
2.2.2 Chemical composition and classification ......................................................................... 11
2.2.3 Hydration process of FA ................................................................................................... 12
2.2.4 Particle morphology ......................................................................................................... 12
2.3 Activated paper sludge minerals (Top-Crete) ............................................................................. 14
2.3.1 Origin ................................................................................................................................ 14
2.3.2 Paper sludge ..................................................................................................................... 14
2.3.2.1 Introduction............................................................................................................... 14
2.3.2.2 The composition of Paper sludge .............................................................................. 15
2.3.3 Thermal activation process of paper sludge .................................................................... 16
2.3.4 Composition of Top-Crete ................................................................................................ 16
2.3.5 The role of the different components and hydration process of TC ................................ 17
2.3.6 Metakaolin........................................................................................................................ 17
2.3.6.1 Introduction............................................................................................................... 17
ix

2.3.6.2 Structure of Metakaolin ............................................................................................ 18

2.3.6.3 Pozzolanic reactions of metakaolin ........................................................................... 19
2.3.7 The replacement of Portland cement by MK ................................................................... 20
2.3.8 Ecological benefits of the production and usage of TC .................................................... 21
2.4 Early age properties of Portland cement .................................................................................... 24
2.4.1 Setting of Portland cement .............................................................................................. 24
2.4.2 The development of the pore structure ........................................................................... 24
2.4.2.1 Capillary and gel pores .............................................................................................. 24
2.4.2.2 Different types of water in the cement paste ........................................................... 26
2.4.3 Hydration heat.................................................................................................................. 26
2.4.4 Degree of hydration ......................................................................................................... 28
2.4.5 Shrinkage .......................................................................................................................... 28
2.4.5.1 Chemical shrinkage.................................................................................................... 29
2.4.5.2 Autogenous shrinkage ............................................................................................... 30
Chapter 3. Materials and experimental program
3.1 Materials...................................................................................................................................... 33
3.1.1 Cement ............................................................................................................................. 33
3.1.2 Mineral admixtures .......................................................................................................... 34
3.1.2.1 Fly ash ........................................................................................................................ 35
3.1.2.2 Top-Crete ................................................................................................................... 36
3.1.3 Water ................................................................................................................................ 37
3.1.4 Aggregates ........................................................................................................................ 37
3.2 Mixture proportions .................................................................................................................... 37
3.3 Experimental techniques ............................................................................................................. 39
3.3.1 Particle size analysis ......................................................................................................... 40
3.3.2 Specific surface area (SSA) ............................................................................................... 41
3.3.3 Free lime content ............................................................................................................. 43
3.3.3.1 Introduction............................................................................................................... 43
3.3.3.2 Glycol extraction method .......................................................................................... 43
3.3.4 Setting............................................................................................................................... 45
3.3.5 Pore structure................................................................................................................... 45
3.3.6 Microstructure development ........................................................................................... 46
3.3.7 Hydration heat and degree of hydration ......................................................................... 46

3.3.8 Chemical bound water and CH content ........................................................................... 47

3.3.9 Chemical shrinkage........................................................................................................... 48
3.3.10 Autogenous shrinkage .................................................................................................... 49
3.3.11 Workability ..................................................................................................................... 50
3.3.12 Strength .......................................................................................................................... 51
3.4 Physical properties of Top-Crete ................................................................................................. 52
3.4.1 Particle size distribution and specific surface area of TC ................................................. 52
3.4.1.1 Results ....................................................................................................................... 52
3.4.2 Microstructure of Top-Crete ............................................................................................ 53
3.4.2.1 Results ....................................................................................................................... 53
3.4.1.3 Conclusion ..................................................................................................................... 54
Chapter 4. Results of the early age properties and microstructure development of blended cement
pastes with addition of TC
4.1 Introduction................................................................................................................................. 57
4.2 Workability .................................................................................................................................. 59
4.2.1 Introduction...................................................................................................................... 59
4.2.3 Results and discussion ...................................................................................................... 59
4.2.4 Conclusion ........................................................................................................................ 60
4.3 Setting.......................................................................................................................................... 61
4.3.1 Introduction...................................................................................................................... 61
4.3.2 Results and discussion ...................................................................................................... 61
4.2.3.1 Results and discussion of the specimen containing 35 % fly ash. ............................. 61
4.2.3.2 Results and discussion of the specimen containing 70 % fly ash. ............................. 62
4.3.3 Conclusion ........................................................................................................................ 63
4.4 Strength ....................................................................................................................................... 65
4.4.3 Results and discussion ...................................................................................................... 65
4.4.3.1 Results of the bending strength ................................................................................ 65
4.4.3.2 Results of the compressive strength ......................................................................... 67
4.4.3.3 Discussion of the compressive strength results ........................................................ 69
4.4.4 Conclusion ........................................................................................................................ 70
4.5 Hydration heat and degree of hydration .................................................................................... 71
4.5.1 Introduction...................................................................................................................... 71
4.5.2 Results and discussion ...................................................................................................... 72
4.5.2.1 The heat evolution of Top-Crete ............................................................................... 72
xi

4.5.2.2 The heat evolution of binary TC-PC blends ............................................................... 73

4.5.2.3 The heat evolution of ternary TC-PC-FA blends ........................................................ 76
4.5.3 Conclusion ........................................................................................................................ 79
4.6 Porosity........................................................................................................................................ 81
4.6.1 General ............................................................................................................................. 81
4.6.2 Results and discussion ...................................................................................................... 81
4.6.2.1 Total porosity of the mixtures with 35% and 70%FA ................................................ 81
4.6.2.2 Effective porosity....................................................................................................... 89
4.6.3 Conclusions....................................................................................................................... 90
4.7 Thermal gravimetric analysis (TGA)............................................................................................. 91
4.7.1 Results and discussion ...................................................................................................... 91
4.7.1.1 Ignition loss of PC,TC and FA ..................................................................................... 91
4.7.1.2 Chemically bound water............................................................................................ 93
4.7.1.3 Calcium hydroxide content ....................................................................................... 94
4.7.2 Conclusion ........................................................................................................................ 96
Chapter 5. Experimental results of the amount of free lime
5.1 Introduction................................................................................................................................. 97
5.2 Results and discussion ................................................................................................................. 97
5.3.1 The percentage of free lime in de cementitious materials .............................................. 97
5.3 Conclusion ................................................................................................................................... 98
Chapter 6. Chemical and autogenous shrinkage of fly ash cement paste blended with Top-Crete
6.1 Introduction................................................................................................................................. 99
6.2 Hardening Shrinkage ................................................................................................................. 100
6.2.1 Results and discussion .................................................................................................... 100
6.2.1.1 Chemical shrinkage of TC ........................................................................................ 100
6.2.1.2 Chemical shrinkage of the blended pastes ............................................................. 100
6.3.3.2 Autogenous shrinkage ............................................................................................. 102
6.3.4 Conclusion ...................................................................................................................... 104
7. Evaluation
8. Conclusions and recommendations
8.1 Conclusions................................................................................................................................ 109
8.2 Recommendations .................................................................................................................... 111
xii

References ................................................................................................................................ 113

Appendix A: Sample Preparation and Experimental Procedure ................................................... 119
Appendix B: Pore size distribution of the blends with 70% FA ..................................................... 141

xiii

xiii

List of Abbreviations
BSE
CH
CSH
EDTA
ESEM
FA
ITZ
LOI
MIP
MK
PSD
RH
SEM
SSA
TC
TGA

Backscattered electron
Calcium hydroxide
Calcium Silicate Hydrates
Ethylenediaminetetraacetic
Environmental Scanning Electron Microscope
Fly ash
Interfacial Transition Zone
Loss on ignition
Mercury Intrusion Porosimetry
Metakaolin
Particle size distribution
Relative humidity
Scanning Electron Microscope
Specific surface area
Top-Crete
Thermogravimetric analysis

w/c water-to-cement
w/b water-to-binder

xiv

xv

Chapter 1 Introduction
1.1 Background
Concrete is the most commonly used building material in the world. Its huge popularity is a
consequence of several advantages, such as general availability, wide applicability and low cost.
These advantages are also accompanied by a great environmental burden. The billions of tons of raw
materials mined and processed each year leave a mark on the environment. Furthermore, during the
production of Portland cement large quantities of CO2 are released into the atmosphere [1] and
enormous amounts of energy are required. Portland cement is one of the most important ingredient
of concrete. The environmental load of concrete can be reduced by the partial replacement of
Portland cement with other cement alternatives or additives. These cement replacing materials could
be fillers or waste products. Among of them, activated paper sludge waste has been proposed to be
a promising cement replacement. Large amounts of paper sludge waste are generated in highly
urbanized cities due to the recycling of tons of paper. The traditional method of the disposal of the
paper sludge waste is by landfilling or by dumping in the sea. Scarcity of land makes it necessary to
find other possibilities to use this waste. Recycling of this paper sludge waste is the best solution and
will be beneficial for the environment and interesting for the government, since the environmental
impact of new materials and the costs for disposing those waste products will be reduced [2].
MinPlus CDEM is a company in the Netherlands that recycles 181,000 tons of paper sludge waste
into a non-toxic mineral Top-Crete (TC) with a high pozzolanic activity. Top-Crete is formed by
calcination and dehydration of the minerals in paper sludge, including kaolin clay and calcium
carbonate, at a specific temperature and time range. When calcining paper sludge at 700 to 750
degrees after 2 or 5 hours a very reactive metakaolin is created [3]. Metakaolin reacts with CH and
creates a dense structure with low porosity, improving the properties of paste, mortar and concrete.
This study focuses on activated paper sludge minerals as partial replacement of cement. The
activated paper sludge by CDEM Minplus is used for study purposes at different replacement levels.

1.2 Problem description

To make an effort in reducing the environmental load of concrete and the disposal of large amounts
of papers sludge, thermally activated paper sludge is used as a partial replacement of cement. The
effects of this replacement on the properties of concrete have not been fully investigated or are
unknown:
I.

The effects of a partial cement replacement with TC on the properties of the cement paste,
i.e. workability, hydration heat, strength and so on, are not reported.

II.

Most studies have been concentrated on the chemical reactions of metakaolin blended with
cement. The hydration process of TC with Portland cement is much more complex and not
fully understood. Especially, the reaction of free lime in TC is unclear. Free lime can
chemically react with water to form Ca(OH)2 or stay as free lime in the concrete. Lime that
does not react in the fresh cement paste will cause problems in the hardened concrete.
When free lime reacts during the later stage of hydration, the volume of the hydration
products expands by a factor of 1.98.The unexpected expansion potentially causes cracks in
1

concrete. Another negative effect of free lime is that the reaction of free lime could lead to
pop-out of the aggregates on the surface.
III.

The contribution of the highly reactive TC to the chemical shrinkage and autogenous
shrinkage is unclear and no detailed work has been reported.

To prevent disadvantages occurring when using TC in blended cement pastes, a part of PC was
replaced by fly ash (FA). The reaction of the large amount of alkaline in TC is expected to generate
more hydration heat than Portland cement. Furthermore, TC might require a large amount of water
due to its high reactivity and large specific surface area. Also the amount of free lime in TC might
produce a large amount of CH in the blended paste. On the other hand, FA has a much lower heat
dissipation and requires CH for the pozzolanic reaction. Moreover, more water should become
available for the physical and chemical binding of TC because FA does not react with water. The
combined use of pozzolans and TC enables TC to act as a leverage to use higher amounts of
supplementary cementing materials such as FA. It is expected that the use of fly-ash and Top-Crete
as cement replacement will make the cement more sustainable, enhance the quality of the concrete,
minimize landfill waste and minimize the environmental pollution.

1.3 Goal, objectives and scope

Goal
The main goal of this study is to examine the feasibility of using Top-Crete (TC) as a partial cement
replacement and especially as a leverage to use higher amounts of supplementary cementing
materials such as fly-ash.
General Objective
The general aim of this research is to gain more insight in the effect of a partial cement replacement
with Top-Crete on the early age properties and mechanical properties of pastes and mortars blended
with 35% and 70% FA. The results of this research may help to improve the usage of activated paper
sludge minerals (Top-Crete) as partial cement replacement in concrete products.
The objectives are:
-

To investigate the early age properties and mechanical properties of the cement paste and
mortar made with addition of Top-Crete.
To define the amount of free lime in the cement paste before and after hydration and
explain what that will mean for the concrete durability.
To investigate the chemical and autogenous shrinkage of the blended cement pastes with TC.

Scope
There are many factors that influence properties of the blended cement paste. The most important
parameters that may affect the composition of the samples are the water to powder ratio, the
amount of Top-Crete and fly ash, admixtures, curing temperature and curing age.
-

Water to powder ratio of 0.5

Curing temperature of 20 C

All parameters will stay constant, except the amount of cement, Top-Crete and fly ash, otherwise it
would be too complicated and too difficult to compare the experimental results and conclude which
parameter is responsible for which effect.

1.4 Outline
This thesis consists of 8 chapters and is organized as follows:
Chapter One provides a general overview of the subject discussed and outlines the problems in the
area of this study, objectives and scope.
Chapter Two provides an overview of the literature study. This study deals with various issues, which
are relevant for this research, starting with general information on the materials used. It also
presents theoretical background information on the setting, pore structure, hydration heat, degree
of hydration and chemical and autogenous shrinkage.
After the literature study an experimentally study was executed to investigate the influence of the
amount of Top-Crete and fly ash on the properties of cement pastes and mortars. The experimental
program is documented and discussed in Chapter 3. It includes the description of the laboratory
testing setup, mixture proportions and the physical properties of Top-Crete.
All the results obtained from the experiments are presented in Chapter 4, 5 and 6. Results regarding
the development of the early age properties and microstructure of the blended pastes, results from
free lime content before and after 28 days of hydration, and results from the chemical and
autogenous shrinkage are included.
A general evaluation of the results on the properties of the blended cement pastes presented in
chapter 4, 5 and 6 is provided in Chapter 7. A correlation between the various properties of the
specimen is described.
Conclusions and recommendations are offered in Chapter 8.

Chapter 2. Literature Study

The primary purpose of the literature study was to provide general information about the materials
used and about the parameters that influence the properties of the cement paste. This section also
presents a summary of the research that has been done so far on cement replacements by thermally
activated paper sludge waste.

2.1 Portland cement

2.1.1 Origin
Portland cement is a rapid- curing binder which was first fabricated in Great Britain in the early 19th
century. The name Portland is derived from the Portland formation, a layer of rocks with the same
properties, from which Portland stone was mined. Portland stone is a white sandy limestone.
Portland cement clinker is obtained from the calcination of calcium carbonate (limestone or chalk)
and aluminosilicate (clay or shale) in a kiln at a temperature of approximately 1450 C. Partial fusion
occurs and nodules of clinker are produced. The clinker, containing CaO, SiO2, Al2O3, Fe2O3 and small
quantities of other oxides, is mixed and grinded with gypsum to make the cement. Gypsum will slow
down the hydration process and thus improve the workability. The mineral compounds of Portland
cement are always the same, but the proportions can be different (Table 2.1).
Table 2.1: Composition of Portland cement
Oxide
CaO
SiO2
Al2O3
Fe2O3
MgO
K2O
Na2O
SO3
Free lime*
* also included in total CaO

%
63 67
19 23
3 7
1.5 4.5
0.5 2.5
0.1 1.2
0.07 0.4
2.5 3.5
0.5 1.5

2.1.2 Clinkers
The components of the clinker consist of dicalcium silicate (C2S), tricalcium silicate (C3S), tricalcium
aluminate (C3A) and tetracalcium aluminate ferrite (C4AF). The relative composition of the clinker
depends on the temperature and the relative proportions of the raw materials. This composition can
easily be determined with the phase diagram, see figure 2.1, and with the equations given by Bogue.
The chemical formulae represents the mass percentage of each oxide:

Aluminate (3CaO.Al2O3) is the first component of the clinker to react with water and can cause
undesirable setting. However, the strength of concrete does not depend on the hydration of
aluminate and ferrite (4CaO.Al2O3.Fe2O3), but on the hydration of alite (3CaO.SiO2) and belite
(2CaO.SiO2). Ferrite and aluminate are only necessary to control the hydration process of the cement
paste [35]. Alite is the most important constituent of Portland cement, which contributes to the early
age strength and heat development. Belite contributes to the later age strength of concrete due to
the slower reaction with water.

Figure 2.1: CaO-SiO2 phase diagram

2.1.3 Hydration Process
Portland cement is a hydraulic binder which, when mixed with water, forms a paste. The cement
paste sets and hardens by means of hydration reactions and processes. During the hydration process
the cement paste gains its strength and stability due to generation and binding of stable hydrates.
2.1.3.1 Chemical reactions of Portland clinkers
The hydration reactions of Portland cement have been reported by many researchers. It has been
found that many different complex chemical reactions take place when water is added to the
cement. The hydration process commences when cement is dispersed in water. C3A dissolves and
starts to react with the dissolved gypsum.

The first hydration products to be formed by the reaction of C3A with water are needle-shaped
crystals of ettringite (tricalcium sulphate hydrate). The reaction of C3A is followed by the hydration
of C3S and C2S. The hydration products formed by these reactions are calcium hydroxide and calcium
silicate hydrates.

Finally, C3A and C4AF react simultaneously after C3S.

2.1.3.2 Hydration products

There are three main products formed in the reaction of cement with water: ettringite, calcium
hydroxide (CH) and silicate hydrate (C-S-H). The formation of these products in time is shown in
figure 2.2.

Figure 2.2: The formation of hydration products in the different stages (after Locher et al. [1976])
Ettringite is present in the early age of the hydration process and is important for the binding of the
cement paste and the beginning of hardening. The formation of ettringite after hardening of
concrete will lead to cracks in the concrete, because the volume increases by a factor 8. The needleshaped crystals of ettringite are illustrated in figure 2.3.

Figure 2.3: SEM Micrograph of ettringite, CH and CSH crystals in PC pastes (Mr. Jim Margeson, NRCIRC)
Calcium silicate hydrate (CSH) is the main reaction product that contributes to the strength of the
concrete. CSH gels are nonstoichiometric colloidal gels that are essentially amorphous (see figure
2.4). CSH is widely used for calcium silicates hydrates and for poorly crystalline calcium silicate
hydrates with a wide range of possible compositions. Both consist of calcium (Ca) and silica (Si) only
the ratio Si/Ca is different. A generally accepted value is between 1.6 and 1.9 [31], lower values will
lead to weaker CSH.

Figure 2.4: SEM Micrograph of CSH crystals in PC pastes (Mr. Jim Margeson, NRC-IRC)
Calcium hydroxide Ca(OH)2 (CH), which has a hexagonal shape, also contributes to the strength of
concrete by filling the pore space. The CH content of ordinary Portland cement paste depends on the
water to cement ratio (w/c) and environmental conditions. It has been found that in practice about
8

20% of ordinary Portland cement is converted to CH [77]. This large amount might lead to
disadvantages such as solubility in water and reactivity with acidic substances, causing a reduction in
the durability of the hardened cement paste. Hence, it is important to reduce the amount of CH by
using pozzolanic materials. The hydration products formed by the pozzolanic reaction are nonsoluble and less chemically reactive.
2.1.3.3 Microstructure development of Portland cement
As the reactions of PC with water proceed more hydration products are formed and bound together
in slow stages [35]. The conversion of the microstructure of cement is schematically given in figure
2.5.

Figure 2.5: The alteration of the microstructure (after Locher et al. [1976])
The hydration process can be divided into three stages [35]:
Dormant stage: The hydration reaction starts when the particles are in contact with water
with a very short and strong reaction period (pre-induction period). Hereafter follows a period of low
reactivity. Due to the formation of hydration products at the surface of the cement particles further
reaction is prevented, this is called the dormant stage. The cement paste stays plastic in this stage
and is therefore flowable. The duration of this stage depends on the amount of water used for
mixing, the chemical composition, mineral admixtures, geometrical aspects and the curing
temperature, and ends when the initial setting time is reached. Ettringite is the hydration product
mainly formed in this stage.
Setting stage: At this stage, the cement particles become accessible to water and start to
actively form reaction products. The hydration products calcium silicate hydrate (C-S-H) and
crystalline calcium hydroxide (CH) are formed. The first binding starts between the hydration
products. The final setting time is the end of this stage.
Hardening stage: More CSH is formed to strengthen the concrete. However, the hydration
slows down because a dense layer is formed around the cement particles preventing further
reaction.

2.1.4 Particle morphology of PC

Most of the PC particles have an irregular, angular shape with sharp edges and the diameters of the
particles range over a large interval (figure 2.6). The size and shape of various cements differ because
of possible different grinding methods, which are used to break larger particles into smaller ones.

Figure 2.6: SEM Images of the particles of Cem I 42.5 at a magnification of 100x and 1000x [W. Vries]

10

2.2 Fly ash

2.2.1 Origin
Fly ash (FA) is the residue that originates from the combustion of pulverized coal at high
temperatures and pressures in power stations to generate electric power. The combusted ash is
captured from the flue gases by highly efficient electrostatic precipitators before the finely divided
ash reaches the atmosphere. Rapid cooling of the ash from the molten state causes the ash to be
highly non-crystalline or glassy with small amounts of crystalline products [16]. Therefore, FA is
pozzolanic in nature and can be used as partial cement replacement in concrete. The calcium
hydroxide which is produced by Portland cement serves as activator for the pozzolanic material.
Pozzolans are amorphous siliceous or siliceous and aluminous materials, which in itself possesses little
or no cementitious property but which will, in finely divided form and in the presence of moisture,
chemically react with calcium hydroxide at ordinary temperature to form compounds possessing
cementitious properties [17]. The term pozzolan is derived from the Italian word Pozzuoli, a place
near Naples, and was first used to describe the volcanic ash mined at Pozzuoli. The ash was produced
by the eruption of the Vesuvius in 79 A.D. The romans added the pozzolan to lime composites to
create hydraulic mixtures for the construction of structures throughout the Roman empire. The
general use of fly ash derived from the combustion of pulverized coal in power plants came after
World War II [16].
In recent years, pollution regulations require that fly ash should be captured prior to release. While
in the past, this ash was released into the atmosphere. The relatively large volume of fly ash is
stored at coal power plants, disposed in landfills or used as partially cement replacement. Fly ash
may contain high levels of contaminants compared to bottom ash. Bottom ash is the ash that does
not rise during the combustion of pulverized coal. The amount of toxic constituents depends upon
the specific coal bed makeup and on the composition of the pulverized coal. However, fly ash mixed
with the bottom ash brings the level of contaminants within the range to qualify as nonhazardous
waste.
2.2.2 Chemical composition and classification
The chemical and mineralogical composition of fly ash depends on the nature of coal, the way it is
pulverized, the combustion process, the process of precipitation of ash from the combustion of gases
and the extent of classification of the particles in the exhaust system [16]. Although, the composition
of fly ash varies significantly, all fly ash include a significant amount of silica (SiO2) which is present in
two forms: amorphous, which is rounded and smooth, and crystalline, which is sharp, pointed and
hazardous, alumina (Al2O3), iron oxide (Fe2O3) and calcium oxide (CaO). The main constituents of fly
ash are given in table 2.2.

11

Table 2.2: The main constituents of fly ash

Fly ash oxide (%)
SiO2
Al2O3
Fe2O3
CaO
MgO
SO3
C

40-60
20-30
4-10
5-30
-

Fly ash is classified into two categories according to the ETA 09/0285, NEN-EN 450-1 + A1 and CUR
94, namely Class C and Class F.

Class C
The Class C fly ash has been produced from lignite or subbituminous coal and has a total
content of SiO2, Al2O3, and Fe2O3 greater than 50%. This type of fly ash has a relatively high
calcium content (sometimes more than 10% of CaO). Therefore, this fly ash shows pozzolanic
and hydraulic properties.
Class F
The Class F fly ash has been produced from anthracite or bituminous coal and has a total
content of SiO2, Al2O3, and Fe2O3 greater than 70%,.This type of fly ash has a relatively low
calcium content (always lower than 10% of CaO). Therefore, this fly ash only shows
pozzolanic properties.

2.2.3 Hydration process of FA

The hydration process of pozzolans combined with Portland cement starts with the reaction of the
clinker with water. This reaction forms the hydration products CSH and CH. The pozzolanic material,
in this case FA, reacts with the CH from the cement hydration to additional CSH phases. The fly ash
needs activation to form CSH hydration products.
The hydration is usually described by the following two reactions:

The hydration process depends on the pH value and the amount of alkali hydroxide. A high pH value
is necessary to achieve a good pozzolanic performance of the FA. The alkali hydroxide content is
necessary to disintegrate the glassy phases to produce silicate and aluminate units, which react with
CH to CSH.
2.2.4 Particle morphology
Fly ash particles are generally spherical in shape and range in size from 0.5 m to 100 m, which is
shown in figure 2.7. The ball bearing shape significantly improves the workability of the concrete.
Hence, less water is required in blended PC-FA pastes to achieve similar consistency.

12

Figure 2.7: Microscope view of fly ash particles at magnification of 2000x [ G.K. Hofmann ]
It is suggested that the hollow spherical particle is formed by the expansion of CO2 and H2O gas,
generated by the minerals within the coal. The main driving forces contribute to spherical shape are
the pressure, the surface tension and the gravity. The particles retain their shape, because the
particles are cooled down rapidly after the combustion. Fly ash has a lower bulk density than PC due
to the hollow particles. Therefore, fly ash has a lower solid volume than PC.

13

2.3 Activated paper sludge minerals (Top-Crete)

2.3.1 Origin
Thermally activated paper sludge (Top-Crete) is a stable mineral that originates from the combustion
of paper sludge under controlled conditions in a fluidized bed combustion system in power plants,
see figure 2.8. This combustion process is based on an innovative energy technology that realizes the
disposal of paper waste by the activation of the latent pozzolanic properties of its mineral content
without the negative environmental effects. The technology is developed by four paper producers in
the Netherlands, called CDEM, after the research work of Pera et al [3]. The research work of Pera et
al. confirmed that controlled combustion of paper sludge brings forth a highly reactive pozzolanic
material, which might be used as cement replacement in concrete. Besides the pozzolanic properties,
this mineral also shows hydraulic potential. The production and usage of thermally activated paper
sludge minerals has attracted increasing interest due to its environmental advantages and both
pozzolanic and hydraulic properties.

Figure 2.8: The combustion of paper sludge under controlled conditions in a fluidized bed combustion
system developed by CDEM.
2.3.2 Paper sludge
2.3.2.1 Introduction
The application of the thermally activated paper sludge in concrete requires a basic understanding of
the nature of the paper waste and paper sludge. Each year, 11 million tons of paper waste are
generated by the pulp and paper industry from the recycling process of paper. The paper waste
14

mainly includes different types of paper sludge, combustion ash and organic and inorganic rejects [6].
Paper sludge is defined as the residue that is left over from the deinking or re-pulping of paper during
the recycling process. About 300 kg of paper sludge is produced for each ton of recycled paper [7].
Paper can practically be recycled only a limited number of times into high quality paper, before the
fibers get disintegrated. The short and weak fibers are separated from the total waste to become a
non-hazardous sludge. Thereafter, the sludge is subjected to deinking in a froth flotation process.
This sludge, known as de-inked sludge, is generally composed of water, fiber (cellulose, hemicellulose
and/or lignin), ink and a mineral loads, such as calcium carbonate, kaolin, talc and titanium oxide.
The relatively large volume of paper sludge is burned, disposed in landfills or used as agricultural
fertilizer and remediation. The paper sludge has also high potential to be used as an alternative for
Portland cement in concrete. Although there are advantages of transforming the paper sludge in a
highly active mineral, still a lot of work has to be performed on the utilization of paper sludge in
concrete production.
2.3.2.2 The composition of Paper sludge
The chemical and mineralogical composition of the paper sludge depends on the function of the
industrial process, the composition of the recycled paper used as raw material and the inorganic
components present [8]. An example of the chemical and mineralogical composition of paper sludge
provided by the Spanish manufacturer Holmen Paper Madrid is listed in table 2.3 [6]. The company
uses 100% recycled paper as raw material for the recycling process.
Table 2.3: The chemical and mineralogical composition of paper waste
Chemical composition (%)
CaO
25.43
SiO2
10.79
Al2O3
6.82
MgO
0.86
Fe2O3
0.46
SO3
0.33
TiO2
0.28
Na2O
0.13
K2O
0.24
P2O3
0.13
LOI
54.24

Organic Material
32.34

Mineralogical composition (%)

Calcite
Kaolinite
45.27
13.67

Silicates and quartz

8.72

The chemical composition of the paper sludge was studied by X- ray fluorescence and the
mineralogical composition by X-ray diffraction. The chemical analysis shows that the sludge is formed
mainly by high amount of lime, silica and alumina contents. From the mineralogical composition it
follows that the sludge mainly consists of organic matter (cellulose fibers ), calcite and kaolinite.

15

Other minerals are present in minor proportions, such as silicates and quartz. The Paper sludge has
also a high ignition loss due to presence of calcite and organic matter [9].
2.3.3 Thermal activation process of paper sludge
The paper sludge has to be subjected to a process of thermal activation to provide it with pozzolanic
and hydraulic properties. The thermal conditions and activation process have therefore a direct
influence on the properties of the activated minerals. One of the most important studies was
performed by Pera et al. [3] on establishing optimal calcination and dehydration conditions that
guarantee total elimination of the organic matter, an appropriate transformation of the kaolinite into
highly reactive MK, as well as a minimum content of free lime [3]. The results showed that a highly
reactive pozzolanic material was created when paper sludge was calcined at 750 C for 2 hours. A
decrease of the activation temperature results in lower pozzolanic activity, which is attributed to the
morphological change in the metakaolinite structure and to the start of decarbonation of the calcite.
The decarbonation liberates free lime in dissolution that overlaps with the pozzolanic reaction [6].
Whereas an increase in activation temperature leads to high lime contents and to recrystallization
into quartz and mullite. The Dutch approach of the thermal activation of paper sludge is based on
this research.
CDEM calcines and dehydrates paper sludge in a fluidized bed combustion system under controlled
conditions. During the calcination and dehydration process, kaolinite is changed into metakaolinite
due to the loss of the structural water. Removal of the OH groups leads to a more disordered
structure, which is stated as the metastable phase. Therefore, the structure of metakaolinite is
disordered and highly amorphous, while kaolinite is crystalline [10]. This metakaolin proved to be a
very valuable product that offers good pozzolanic properties when used as cement replacement [8],
[12], [14] and [15]. The thermally activated ash, called Top-Crete (TC), is collected from the flue
gases of the combustor with electrostatic filters. This process used by CDEM is compared to the
conventional combustion process clean, locks up highly toxic chlorine, prevents the formation of
dioxins and captures heavy metal emissions.
2.3.4 Composition of Top-Crete
After the calcination and dehydration of paper sludge, the resulting ash (top-Crete) is composed
mainly of metakaolin (silica and alumina), free lime, calcium carbonate and inert filler. The
characterization is carried out by CDEM and shown in table 2.4.
Table 2.4: The mineralogical composition of Top-Crete
Mineralogical composition (%)
Calcium carbonate
Metakaolin
Free lime
Inert fraction

41
29
23
7

One of the most important properties for a waste product to be accepted and used as a mineral in
composite cement is its pozzolanic nature. The metakaolin in TC contributes to the pozzolanic
properties of TC, while the free lime in TC contributes to the hydraulic properties of TC.

16

2.3.5 The role of the different components and hydration process of TC

The role of each component in TC depends on the mixture proportions of the composite blend, the
water content and the environmental conditions. However, the main role is described below:

Calcium carbonate
The calcium carbonate will behave as inert filler when added to the composite paste,
assuming that the calcium carbonate appears in its stable form calcite. The inert filler can
affect the concrete on physical level, when the particles fill the intergranular voids between
cement particles, and on the surface chemical level, when the particles enhance the
hydration process by acting as nucleation sites [11].

Metakaolin
Metakaolin is a known pozzolanic material that is used in pure form as additional admixture
in concrete. Metakaolin reacts fast with lime and forms hydrated compounds in the presence
of water. This product proved to be more reactive than conventional metakaolins available in
the market [9] and [10].

Free lime
Free lime might react with water and convert to portlandite, Ca(OH)2, or remain as free lime
in the hardened concrete. This free lime has the tendency to react at a later age with the
surrounding material, like free water or CO2 in the air. The reaction with water and CO2 leads
to the formation of CH and carbonate. The volume of the these hydration products is bigger
than the volume of the original CaO.

Inert fraction
The inert fraction has the same effect as the calcium carbonate. The presence of dolomite
might lead to alkali carbonate reaction (ACR). This reaction is unlikely to occur due to the
fineness of TC.

The hydration process of TC mainly depends on the following reactions:

The generated CH, CSH and crystalline products contribute to the hardened properties of the paste,
such as strength, porosity and shrinkage.
2.3.6 Metakaolin
2.3.6.1 Introduction
Kaolin, which is mainly composed of kaolinite, quartz and mica, is a sedimentary rock formed by the
alteration of different crystalline and amorphous rocks (such as feldspar and volcanic ash). Kaolinite
is a white or greyish white hydrous aluminium silicate formed by rock weathering of kaolin [12].
Metakaolin (MK) is obtained from calcination and dehydroxilation of the raw clay mineral kaolinite
under controlled conditions. MK can be generated in pure or impure form, which depends on the
mineralogical composition of the kaolinite. Currently, researches have been focused on impure MK
due to the stringent regulations on environmental protection. Impure MK is acquired from red mud,
17

which is a waste material from the refining process of bauxite, or from the combustion of paper
sludge. The application of pure MK in concrete has expanded rapidly over the years, because MK
prevents alkali silica reactions, improves the strength and water absorption properties and improves
the durability in aggressive environments [12].
2.3.6.2 Structure of Metakaolin
The chemical structure of kaolinite is Al2Si2O5(OH)4 and the morphology structure of the crystal is
plate-like. The lamellar structure and shape are presented in figures 2.9 and 2.10. One crystal
contains up to approximately 200 layers with layers of silica in tetrahedral coordination and alumina
in octahedral coordination. The layers are very close to each other, which prevents the passage of
water molecules between the layers [13].

Figure 2.9: The lamellar structure of kaolinite crystals and an ESEM image of the shape of the crystals
[G. Varga, 2007]

Figure 2.10: Lamellar structure of the crystals [G. Varga, 2007]

18

The kaolinite loses its water when it is calcined and dehydroxylated to form metakaolinite. The
alumina and silica layers get disordered (amorphous), through the breaking of the unstable bonds,
and become chemically reactive. The metakaolinite does not collapse but remains the layered
structure [12].
2.3.6.3 Pozzolanic reactions of metakaolin
From several studies the pozzolanic reaction of MK proved to be a highly complex process. Three
types of pozzolanic reactions of MK are distinguished: (a) with alkali metal hydroxides and soluble
alkali metal silicates, (b) with calcium hydroxide and (c) with PC, which generates alkalis (CH) among
its hydration products [12].
(a)
Composite cements made from alkali metal hydroxides or soluble alkali metal silicates with
MK are known as geo-polymers, ceramics or inorganic polymers. The initial hydration product of the
pozzolanic reaction is aluminosilicate zeolite. The aluminosilicate zeolites have a three-dimensional
amorphous zeolite structure of alkali silicates, which increases the durability (strength and
permeability) of these cements compared to Portland cement. These cements are most applied in
heat resistant concrete and for encapsulating hazardous waste.
(b)
When MK cements are activated by an alkaline solution, such as calcium hydroxide, the
hydration starts with a rapid reaction followed by a slow reaction. The pozzolanic reaction in
MK-lime-water systems was found to be [12]:

In which C5AS2H5 is an average composition of the hydration products CSH, C2ASH8 (gehlenite
hydrate), C4AH13 and C3AH6 (hydrated calcium aluminates). The composition and structure of CSH is
variable. C2ASH8 has the ability to bind large amounts of Na+ and K+ to the structure, which causes
the displacement of Ca2+ in the lattice, and this might affect the chemistry and structure of the
cement. These blends are used to make hydraulic lime mortar, which is used in new constructions
and also in the restoration of old buildings.
(c)
The presence of MK in Portland cements has a significant effect on the chemistry of the
cement paste due to rapid reaction of MK with CH generated by PC. The formed hydration products
are affected by the composition of the blended cement paste. At low MK to PC ratios the pozzolanic
reaction of MK with a large supply of CH leads to the hydration products C4AH13 and CSH gel,
whereas at high MK to PC ratios the MK reaction leads to production of C2ASH8 and CSH gel. The
formation of C4AH13 contributes to a higher water demand than the formation of C2ASH8 [28].
It also has been found that the rate of reaction of MK with CH is similar in both MK/lime and MK/PC
pastes. However, MK accelerates the hydration of PC clinkers in MK/PC pastes [12]. Besides the
acceleration effect, MK also influences the chemical structure and the nanostructure of CSH. The CSH
from PC pastes typically shows a C/S ratio from 1.2 to 2.3, while the C/S ratio of MK/PC pastes ranges
from 0.7 to 2.4. This is attributed to the incorporation of AlO45- into the CSH structure [12]. It also has
been suggested that the CSH phase has a foil-like structure, rather than the needle-like structure
of normal CSH. The foil structure is assumed to be more efficient at filling space without leaving large
interconnected capillary pores [12]. The changes in the CSH structure might improve the durability
and permeability properties of MK/PC blended pastes.
19

2.3.7 The replacement of Portland cement by MK

The benefits of pure MK and MK in thermally activated paper sludge, have been widely studied for its
use in cementitious materials. Although TC has a different influence on the properties of the
composite binder than pure MK, these studies are of great interest for this research. A brief summary
of the results in [8], [12], [14] and [15] is given here.
Effects on the consistency and setting
It is generally acknowledged that MK significantly reduces the normal consistency of the paste and
slightly reduces the initial and final setting time of the cement paste for replacement levels over 10%
of MK. The decrease in consistency is attributed to a greater specific surface area of MK in
combination with a distribution of finer sized particles. Both of these properties of MK require an
increase in water demand. The acceleration in setting times is assigned to the joint presence of
metakaolinite and calcium carbonate.
Effects on the strength
Numerous studies about the influence of MK on the strength have presented an enhancement in the
compressive strength. The optimum replacement level to enhance the compressive strength is
achieved at 10% or 20%, which depends on the MK composition. Two factors that contribute to the
strength enhancement have been identified. These are the filler effect, which accelerates the PC
hydration, and the pozzolanic reaction.
Effects on the pore structure
It was found that the addition of MK in the cement paste led to a refinement of the pore structure
and to a reduction of the threshold pore diameter. However, an increase in MK beyond 30% leads to
an increase of the total porosity. Furthermore, the total intruded pore volumes by MIP tests
increased between the ages of 14 and 28 days for metakaolin paste. The slight increase in pore
volume between 14 and 28 days is explained by the formation of an inhibiting layer of reduction
products on MK pastes.
Effects on the hydration heat
The cumulative heat generated by MK is approximately the same as the total heat evolved by PC.
However, the hydration heat evolves more rapidly in MK blended cement pastes than in the PC
pastes. Therefore, the temperature rise in concrete structures, where the heat dissipation is of great
influence, might be slightly higher for MK concrete.
Effects on shrinkage
Results indicate that the autogenous shrinkage increases with increase in MK content up to a
maximum of 10% MK. This phenomenon was assigned to a combined effect of cement hydration and
pozzolanic reaction of MK. Meanwhile, the chemical shrinkage was also found to increase with an
increase in MK content up to a maximum of 10% MK. This was also attributed to the cement
hydration in combination with the pozzolanic reaction.
Furthermore, an increase in drying shrinkage is obtained when the cement is replaced by 20% MK or
more. The greater shrinkage is explained on the basis of three phenomenas:
1. Nucleation of hydration products on the smaller sized minerals, which accelerates the
cement hydration. Hence, the drying shrinkage of the binder is increased.
20

2. The pozzolanic reaction of MK requires more water and therefore increases the drying
shrinkage.
3. Increase in capillary pressure due to refinement of the pore size.
Effects on durability properties
Chloride ion diffusion
Results show that the chloride diffusion rate is reduced by a factor 3 when PC was partially replaced
by MK. Thus, the increase in MK increases the chloride penetration resistance of concrete. The
results obtained where attributed to the relative changes in intrinsic diffusivity and chloride binding
capacity presented by the different binder compositions. The refinement of the pore structure was
also found to reduce the chloride penetration.
Alkali silica reaction (ASR)
Different studies showed that the expansion due to the alkali silica reaction may sufficiently be
inhibited when PC is replaced with approximately 15% MK. The mechanism that suppresses the
expansion seemed to be due to the entrapment of the available CH and the reduction of the CH/SiO2
ratio by the MK hydrates.
Sulfate resistance
The addition of MK in the cement paste has proved to reduce or eliminate the harmful effects of
sulfate attack. The sulfate resistance increases with increase of MK replacement levels. The main
factor of higher chemical resistance is considered to be the possible reduction of CH caused by the
addition of MK. It was also suggested that the alumino silicate phases of hydrated cement blends are
more stable and therefore more resistant to sulfate attack.
Short summary of the results
The influence of MK on the binder properties in blended cement pastes and mortars has been widely
studied. The research studies clearly show that MK is an effective pozzolan that contributes to a
decrease in setting time, an enhancement of the strength, a refinement of the pore structure, a
reduction of the threshold pore diameter, elimination of the expansion due to ASR and an
improvement in the chemical resistance. Unfortunately, the use of MK as partial cement
replacement also has some negative side effects, such as an increase in chemical, autogenous and
drying shrinkage and a reduction of workability.
2.3.8 Ecological benefits of the production and usage of TC
The production and usage of TC have many advantages from an environmental point of view
compared to the production and use of Portland cement and other pozzolanic materials. Some
positive effects on the environment will be discussed in this paragraph:

CO2 emission

During the production process of Portland cement greenhouse gasses are emitted. A significant part
of these emissions includes the emission of CO2. The production of one ton cement causes
approximately 1.0 ton of CO2 to be emitted into the atmosphere. In the determination of this
number all CO2 sources in the entire production line of cement have been taken into account, such as
21

the use of electricity and blast furnaces [35]. With 1.6 billion tons of cement produced each year, this
production is responsible for 7% - 10% of the total CO2 emission worldwide. This puts the production
of cement in the top three of the largest sources of CO2 emission.
One of the most important advantages of the usage of Top-Crete as cement replacement is that it
allows the realization of a substantial CO2 reduction. In order to reduce the global warming and the
greenhouse effect, it is worldwide of great importance to reduce the amount of CO2 emission. CDEM
Minplus has proven with its innovative technology that the CO2 emission is considerably lower in the
production of TC than in the production of Portland cement. The CDEM plant in Duiven has the
capacity to produce more than 43.000 ton of TC. Worldwide, it is estimated that the current supply
of paper sludge could be transformed into 7 million ton of cementitious additives. Therefore, the
partial replacement of PC in concrete by TC will prevent approximately 5.71 million ton of CO2
emission each year.

Landfill and incineration by paper sludge waste

Nowadays, industrial waste is in some cases combusted and dumped to landfill or disposed of by
spread. Dumping of this waste leads to overcrowded landfills and the need for more landfills. The
methods of disposal are expensive and cause concern by their impact on the environment. TopCrete might be an environmental friendly and cost-effective solution for dumping or incinerating
paper sludge waste. Therefore, the utilization of TC in civil construction may lead to a reduction of
the environmental pollution caused by the paper and lime production waste. This is important for
both the environment and the government, because of cost savings on landfill and incineration of
waste paper and a reduction of the negative impact on the environment.

The use of natural resources

The strong increase in the use of concrete in civil construction entails a strong increase in the use of
PC. The production of PC consumes significant amount of natural resources. The excessive
consumption will lead to depletion of these resources and to a major load on the environment due to
the pollution and loss of nature by mining of these natural resources. CDEM Minplus uses paper
waste products, which saves valuable resources.
In addition, a very valuable product high reactive metakaolin is obtained from the production of TopCrete. Pure metakaolin is a calcined product that is obtained from the dehydroxilation of natural raw
materials under controlled conditions. Today, the use of raw materials has become more stringent
due to the regulations on environmental protection. The impure metakaolin obtained from paper
sludge proved to be a high valuable product that offers good pozzolanic properties when used as
cement replacement.

Green energy

Still a lot of gray energy is used worldwide. Gray energy is defined as the energy generated by the
combustion of fossil fuels, such as coal or gas. These energies are harmful to the environment due to
the emission of harmful and dangerous substances during the production process. This energy is also
not sustainable, because natural gas fields are consumed. The new production process used to
produce Top-Crete generates a surplus of energy without the emission of harmful dioxins or other
pollutants. Each year, over 28,000 MWh of green (clean and zero carbon) energy is generated by
22

CDEM and sold to other companies to meet commercial heating and power needs. This energy is
attractive to society because it is 100% green and reliable. Reliable energy is considered to be that
energy that does not depend on unpredictable weather conditions, such as wind - and solar energy.

23

2.4 Early age properties of Portland cement

2.4.1 Setting of Portland cement
Every cementitious mixture sets when mixed with water by means of a set of chemical reactions. The
different components will hydrate, but each consituent will contribute differently to the hydration
process. Therefore, the setting time will depend on a number of factors, namely the composition of
the material, water to binder ratio, the fineness of the powders and the temperature. The major
components that will affect the setting time are tricalcium silicate (C3S), tricalcium aluminate (C3A),
gypsum, free lime and pozzolans

C3S and C3A cause the cement to harden immediately, reducing the setting time. This is due
to the forming of ettringite, CH and CSH [24].
Gypsum prevents cement from flash setting, because cement would react with water into
tetra calcium aluminate hydrate (4 CaO.Al2O3. 19 H2o) which has the form of large plate
crystals. These large crystals would connect the cement particles with each other, preventing
further movement of the particles. Thus increasing the setting time.
Free lime reacts with water in CH, reducing the setting time.
Highly reactive pozzolans also cause the cement to harden with a high rate. This is due to the
formation of CSH and crystalline products.

A finer material increases the hydration, since the surface area per unit mass increases and the
hydration takes place at the surface of a particle. Therefore, an increase in fineness will speed up the
setting. The mixtures should also be prepared and cured in rooms with constant temperatures and
humidities. An increment of the temperature and a decrease in humidity is followed by a decrement
of the setting time, due the acceleration of the hydration process [25].
The setting of cement is a continuous process, but to distinguish setting from hardening two
parameters are used for research purposes. The initial setting time indicates the time that is needed
to change the mix from liquid to a stiff mix with a certain degree of stiffness. The final setting is the
time at which the cement changes from plastic into a solid state. After the final setting the cement
has acquired a sufficient stiffness to resist a certain pressure without alteration of shape. The initial
and final setting varies for each mixture and can take anywhere from 20 minutes to 24 hours.
2.4.2 The development of the pore structure
2.4.2.1 Capillary and gel pores
Pores in hardened cement paste (hcp) are categorized in gel pores and capillary pores [31]. A
detailed classification of the pores with the role of water in them is given by Young et al. in table 2.5.

24

Table 2.5: Classification of pores in hardened cement paste (Young et al, 1995)
Designation

Diameter

Description

Role of water

Properties influenced

Capillary pores

10-0.05 m
50 - 10 nm

Large capillaries
Medium
capillaries

Gel pores

10 - 2.5 nm

Small capillaries

Strength, permeability
Strength, permeability,
shrinkage at high
humidities
Shrinkage to 50% rh

2.5 - 0.5 nm

Micro pores

< 0.5 nm

Micro pores

Bulk water
Menisci generate
moderate surface
tension forces
Menisci generate
moderate surface
tension forces
Strongly adsorbed
water, no menisci
Structural water
involved in bonding

Shrinkage, creep
Shrinkage, creep

Figure 2.11 shows a schematic overview of the porosity in the cement paste. Capillary pores (4) have
a greater width than gel pores (2 and 3), see table 2.5, and are defined by the space between the
cement particles taken by air or water not used for hydration. The formation of new CSH gel particles
in capillary pores are allowed at any temperature if there is condensed water available [32]. The
development of these pores depends on the water to binder ratio and on the rate of hydration. Gel
pores are spaces within the CSH particles. Diamond [32] distinguished two sorts of gel pores, namely
the open gel pores (3), which exist in the outer CSH gel product, and the dense gel pores (2) which
exist in the inner CSH gel product. The formation of hydration products in the gel pores may take
place with increase in temperature [32].

Figure 2.11: Schematic image of the porosity in the cement paste [32]
The total volume of the gel pores will increase as long as the hydration goes on, while the total
volume of capillary pores decreases with the progress of hydration [32].

25

The pore structure is defined by the pore size distribution, porosity and the pore connection. Porosity
is the ratio of the total volume of the pores in the concrete to the total volume of the concrete:

In which:

= Total volume pores; m3

= Total volume concrete; m3

2.4.2.2 Different types of water in the cement paste

Water in cement paste can be distinguished in three types of water:
1. Chemically bound water
The water that is required to form a certain part of hydration products is called chemically
bound water. About 25% of the water used for mixing is required for the chemical reactions
in the cement paste [35]. This is the amount of water that can only be driven off when the
blended cement paste is exposed to a temperature higher than 105C, which means that
chemically bound water does not depend on the relative humidity. Assumed is that all
chemically bounded water will be lost when heating the sample to a temperature of 1000C.
2.

Free water
Water present in capillary pores is called free water or capillary water. Free water is available
for the hydration process to form hydrates. Exposing free water to a temperature of 20C to
105C will cause the water to evaporate. Capillary water depends to a large extent on the
outside relative humidity (RH). Outside RH values lower than 30% will induce the sample to
evaporate water to the environment, whereas at higher humidity levels the sample is
induced to absorb water causing the concrete to shrink and to extend.

3. Physically bound water

The water present in gel pores is called physically bound water or gel water, this water is not
available for hydration. 15% of the gel water is used to fill the gel pores. Another part of the
gel water can be physically adsorbed on the surface of hydration products and can be partly
located in spaces between hydrated products, which are defined as interlayer hydrate water
[69]. The "' inter-crystallite "' bonds do not separate when exposed to physically adsorbed
water [70]. The volume of gel water also depends on the relative humidity. When the
relative humidity is low, one layer of water is physically bound on the hydration products. On
the other hand when the relative humidity is relatively high, approximately 80%, it will
physically bind more water molecules and completely fill the gel pores. Physically bound
water can condensate to free water when hydration products are formed close to each
other.
2.4.3 Hydration heat
The reactions that occur when water is added to the cement paste, are mostly exothermic, that
means that the reactions generate heat. An illustrative example of the heat evolution of Portland
cement is shown in figure 2.12.The hydration process can be divided in three stages as discussed
before. In the first stage the lime in PC is dissolved and ettringite is formed from the reaction of C3A
26

with gypsum and water. This reaction corresponds to the first high peak in the curve and shows a
large heat liberation. This stage only takes place for a few minutes. Hereafter follows the dormant
stage, in which further reaction is prevented due to formation of a layer around the particles. This
corresponds to the constant line in the heat curve [35].

Figure 2.12: Schematic curve of the heat evolution of PC.

The dormant stage is followed by the reaction of C3A and C3S, due to breaking of the layer around the
particles. The particles continue to hydrate and form CSH and CH which corresponds with the second
wide peak in the curve. The speed at which this happens depends on the accessibility of the particles
for the water. A third peak in the heat curve can also occur when ettringite is formed by the reaction
of aluminates, which is converted into monosulphate at a later stage. As long as the reaction of water
with cement takes place at the surface of the particle, the reaction will go quickly. If the water has to
penetrate through the formed gel into the grains, the reaction will proceed more slowly and less heat
will be produced. In the last stage a dense structure is formed around the particles, slowing down
further reaction thus slowing down the hydration heat of the paste.
The heat evolution gives information on the initial hydration reactions and on the degree of
hydration and depends on the composition of the cement, the temperature, pressure, water to
cement ratio, additives, pozzolanic activity of the pozzolan and the fineness of the particles

C3S: Results have shown that compositions containing more C3S have a higher heat
production, increasing the second peak. The heat evolution of the different clinkers in PC is
listed in table 2.6.
Table 2.6: Heat release of the clinkers in PC
Clinker
2 C3S + 6H
2 C2S + 4H
C3A + 12H + CH
C4AF + 9H + 4CH
C+H

hydrates
C3S2H3 + 3CH +
C3S2H3 + CH +
C4AH13 +
C4(AF)H13 +
CH +

+ heat
230 kJ
86 kJ
362 kJ
240 kJ
66 kJ

J/g
500
250
1340
400
1170
27

Fineness: Finely ground cement reacts faster than coarser ground cement. Smaller grain sizes
are inherent to a larger surface. Since the initial reaction occurs at the surface the peak value
will increase with decreasing grain size.
W/c ratio: The water to cement ratio also appears to have effect on the hydration heat
production. The more water available, the easier the reaction will proceed. This is explained
as follows, the cement grains must come in physical contact with water for cement to
hydrate. However, the cement matrix can contain spaces in which water is present without
having contact with the cement particles. To ensure physical contact between the water and
cement particles, there must be more water available then the minimum required water
content for complete hydration.
Pozzolanic activity: The increase in pozzolanic activity will accelerate the pozzolanic reaction,
increasing the hydration heat due to the exothermic effect of the pozzolanic reaction.
Materials with low pozzolanic activity, such as fly ash, decrease the hydration heat.

2.4.4 Degree of hydration

Full hydration is in practice barely achieved. The amount of reacted cement relative to the original
amount of cement is called the degree of hydration, max [35]. The degree of hydration is also
defined as the ratio of the cement content that has hydrated to the total amount of cement, which is
the same as the first definition.
Many parameters can be used to indicate the degree of hydration, such as by the amount of CH, the
quantity of the chemically bound water, fraction of un-hydrated cement and by the heat of
hydration. From these methods the amount of CH is most commenly used to estimate the degree of
hydration of cement [81]. Each method has proven well in determining the degree of hydration of
100% PC, but when mineral admixtures are incorporated none is capable to estimate the exact max
of binary or ternary blends [81].
Theoretically, a w/c ratio of about 0.4 is sufficient to obtain complete hydration of 100% cement (i.e.
max=1.0). The hydration slows down with lower w/c ratios due to absence of water.
2.4.5 Shrinkage
Shrinkage of concrete is described by Gilbert [86] as the time dependent strain which is measured in
an unloaded and unrestrained specimen at constant temperature. Many researchers distinguish
various types of shrinkages in cementitious blends due to loss of moisture at different stages, such as
hardening shrinkage, drying shrinkage and plastic shrinkage. Hardening shrinkage is the reduction in
the volume of blended pastes due to occurrence of chemistry or internal structural changes, where
moisture transmission with the environment is not allowed. Whereas the other types of shrinkage
take place due to changes in environmental conditions, like water evaporation to the environment or
absorption of CO2. Two shrinkage types are directly responsible for hardening shrinkage, namely
chemical shrinkage and autogenous shrinkage.
The hardening shrinkage of blended cement pastes is influenced by numerous factors including
composition and fineness of cement and minerals, degree of hydration, water to binder ratio and the
presence of mineral admixtures [35].

28

2.4.5.1 Chemical shrinkage

The chemical reaction between water, minerals and cement is a complex phenomenon, which is
related to the progressive hardening and to the volume change of the cement matrix. Chemical
shrinkage is defined as the reduction in absolute volume of the reaction products in the blended
hydrating paste during hydration [87]. Chemical shrinkage occurs as long as the hydration process of
the blended cement paste continues.
Chemical shrinkage depends on the chemical composition and degree of hydration of the cement
paste. The influence of the individual clinker components on the chemical shrinkage is modeled by
Paulini [1996] based on Boyles equations, the plot is given in figure 2.13. It has been found that C2S
produces hydration products which are larger in size than the original C2S causing swelling during the
first days, while the hydration products of the other clinkers shrink in size. This is attributed to the
surface reaction of C2S, its polymorphic behaviour and the presence of retarding agents [89]. C3A also
has higher shrinkage during the first days compared to C3S and C4AF. Therefore, Portland cement
with a high C3A content causes greater shrinkage than a lower C3A content. Combinations of larger
C2S fractions and smaller C3A fractions result in a lower shrinkage due to the compensation of the
shrinkage of one clinker by the swelling of the other clinker.
The water to binder ratio and the fineness of the cement particles will not have any influence on the
amount of chemical shrinkage [89]. These factors only effect the rate of the chemical shrinkage.
Therefore, the maximum volume change is only influenced by the chemical composition and the
degree of hydration of the cement.

Figure 2.13: The influence of the composition of Portland cement on the chemical shrinkage (Paulini
1996)
Chemical shrinkage can be determined with the equation based on the molecular weights of the
initial and final products, since the chemical shrinkage is the absolute volume change of the

29

hydration products [89]. Additional experiments are required to determine the exact volume of
different components in the equation:

In which CS stands for chemical shrinkage, Vci is the volume of cement before mixing, Vc is the
volume of the hydrated cement, Vwi is the volume of water before mixing, Vw is the volume of
reacted water and Vhy is the volume of hydrated products.
2.4.5.2 Autogenous shrinkage
Autogenous shrinkage has generally been defined by different researchers as the macroscopic
volume change of concrete that does not include the moisture exchange between the material and
exterior surrounding environment, i.e. external water is not available for the hydration process and
pore water is utilized for the hydration process. This phenomenon is known for a long time, but the
contribution of autogenous shrinkage to the total shrinkage in concrete structures has recently been
acknowledged [94] and accurately investigated and documented. Autogenous shrinkage is part of the
chemical shrinkage [88] and occurs over three different stages. Figure 2.14 shows the volume change
of a sealed cement blend due to the hydration reactions.

Figure 2.14: Volume changes due to chemical shrinkage and autogenous shrinkage [Ambrosia et al,
[87]]
Many researchers have found a correlation between autogenous shrinkage and the degree of
hydration of Portland cement [89], [95] and [96]. Therefore, autogenous shrinkage is a result of
chemical shrinkage related to the hydration of cement particles. The relative shrinkage, ratio
between autogenous shrinkage and chemical shrinkage, as function of the degree of hydration is
illustrated in figure 2.15.

30

Figure 2.15: The relative shrinkage as function of the hydration degree.

Three stages of autogenous shrinkage can be distinguished during the hydration process. These three
stages are shown in figures 2.14 and 2.15. The first section AB in figure 2.15 shows a linear behavior
between the relative shrinkage and the degree of hydration. In this stage the chemical shrinkage is
identical to the autogenous shrinkage (pre set) due to the liquid cement paste. The fluid paste cannot
support internal voids created by chemical shrinkage and causes an external volume change. Section
BC starts at initial setting time with the development of more stiffness without resisting deformation.
The solid skeleton can resist some chemical shrinkage stresses and chemical shrinkage becomes
more restrained. After final setting time the paste becomes rigid in section CD. Empty pores are
formed and autogenous shrinkage includes less chemical shrinkage. The volume reduction is now
only due to self-desiccation of the paste. Self-desiccation of paste is defined by Holt [89] as the
localized drying resulting from a decline in internal relative humidity, which in turn is caused by the
need of extra water for the continuous hydration of the unhydrated cementitious particles. This gives
rise to the formation of water- air menisci. These menisci will continue to be sucked into the capillary
pores, causing tensile stresses in the pore fluid. Therefore, stresses are generated on the capillary
pore walls and the cement paste wants to reduce in volume.
The rigidity of the blended cement paste as function of time is given in figure 2.16. After initial
setting the paste has developed a rigid structure, causing the formation of voids in the
microstructure when the hydration continues. This internal volume reduction is called chemical
shrinkage, while the external volume change is caused by autogenous shrinkage.

31

Figure 2.16: Rigidity of the cement paste as function of time

Lura [97] has described in his report three main driving forces that may contribute to the selfdesiccation of autogenous shrinkage. These three mechanisms have been identified by other
researchers, which are the surface tension, disjoining pressure and capillary water. For more detailed
information on these mechanisms, see the dissertation of P. Lura [97].

Figure 2.17: Stresses pulling the water meniscus lower between two cement particles due to water
suction and capillary pressure [89]
Other obvious factors directly influencing the magnitude of autogenous shrinkage are the w/b ratio,
the inclusion of mineral admixtures, the pozzolanic activity of the mineral admixtures, additives (such
as superplasticizers) and most important the composition and fineness of the cementitious materials.
Lower w/b ratios will cause the unhydrated cementitious particles to consume water from the larger
pores and the more refined pores. The finer porosity due to lower w/b ratio causes the water- air
meniscus to have a greater radius of curvature (figure 2.17), thus increasing the autogenous
shrinkage. Also the use of chemical additives, such superplasticizers, may create the possibility to
reduce the w/b ratio and to produce workable mixtures. However, the water content is reduced and
the concrete is more susceptible to self-desiccation.

32

Chapter 3. Materials and experimental program

3.1 Materials
Throughout the study, three different types of cementitious materials were used in producing a
composite binder, i.e. Portland cement, fly ash and Top-Crete. FA and TC were prepared to replace
the cement content in the mixtures by different percentages. This section provides a summary of the
characteristics of all materials used to produce cement pastes or mortars.
The following materials were used:
-

Portland cement
Top-Crete
Fly ash
Water
Fine aggregates

No chemical admixtures were added to the mixtures to improve the workability

3.1.1 Cement
One type of standard Portland cement (PC) CEM I 42.5 N is used during the whole experimental study
from a single batch. This cement is manufactured by ENCI Heidelberg cement. The mineralogical
composition of CEM I 42,5 N is shown in table 3.1 [5].
Table 3.1: Chemical composition of Portland cement
Chemical
CaO
SiO2
Al2O3
Fe2O3
SO3
MgO
Na2O
K2O
TiO2
P3O5
Mn3O4
LOI
Free lime

Weight (%)
64.4
20.36
4.96
3.17
2.57
2.09
0.14
0.64
0.35
0.18
0.14
0.88
0.6

The phase compositions of PC are calculated with Bogues equations using the data of the oxide
compositions in table 3.1. The results of the phase compositions are presented in table 3.2.

33

Table 3.2: Clinker compound

Clinker compound
C3A
C2S
C3S
C4AF
Max heat (J/g)

Weight (%)
8.31
14.76
65.83
11.11
487.44

The specific surface area (SSA) and the median size of PC are listed in table 3.3 and its particle size
distribution is illustrated in figure 3.1. The measurement of the particles size distribution is
performed by BAS research & technology using the laser diffraction method.
Table 3.3: Specific surface area and median size
Surface area
[m2/g]
0.27

PC

D10 [m]

D50 [m]

D90 [m]

1.89

16.56

70.06

120

Undersize weight % [kg/kg]

100

80

60

Cem I 42,5 N

40

20

0
1

10

Diameter [um]

100

1000

Figure 3.1: Particle size distribution of Portland cement by BAS

3.1.2 Mineral admixtures
The mineral admixtures fly ash and Top-Crete are used in this research as possible cement
replacement. The combination of these two minerals is expected to have a positive effect on both
materials and on the properties of the composite binder. Most of the properties of fly ash are
investigated by many researches, but Top-Crete is a new kind of mineral admixture with unknown
properties.

34

3.1.2.1 Fly ash

A class F fly ash with a low free lime content obtained from the company EON - Benelux has been
used in the mixtures. Table 3.4 shows the results of the chemical analysis of FA measured by X- ray
Diffraction according to EN 196.
Table 3.4: Chemical composition of Fly ash
Chemical
SiO2
Al2O3
CaO
Fe2O3
P2O5
K2O
MgO
TiO2
SO3
Na2O

Weight (%)
48.36
31.36
7.14
4.44
1.90
1.64
1.35
1.24
1.18
0.718

The specific surface area (SSA) and the median size of FA are listed in table 3.5 and its particle size
distribution is illustrated in figure 3.2. The particle size distribution is also performed by BAS research
& technology.
Table 3.5: Specific surface area and median size
Surface area
[m2/g]
0.25

FA

D10 [m]

D50 [m]

D90 [m]

3.53

21.46

93.72

120

Undersize weight % [kg/kg]

100

80

60

Fly ash

40

20

0
1

10

Diameter [um]

100

1000

Figure 3.2: Particle size distribution of fly ash by BAS

35

3.1.2.2 Top-Crete
Top-Crete, produced by MinPlus CDEM, is principally formed of metakaolin and calcium compounds.
The remaining oxides are present in an amount of less than 1%. The chemical composition of TC as
provided by MinPlus CDEM is presented in Table 3.6. The mineral metakaolin is generally considered
as a mineral with a relatively high pozzolanic activity. The pozzolanic activity of TC depends on the
chemistry of the paper residues and on the thermal condition applied [79]. This has been discussed in
the literature study.
Table 3.6: Chemical composition of Top-Crete
Composition
CaCO3
Al2O3. 2 SiO2
CaO
MgO
TiO2
Fe2O3
SO3
K2O
P2O5
CuO
Na2O
BaO
Cl
MnO
ZrO2
ZnO
NiO
PbO
Cr2O3
As2O3
HgO
CdO

Weight (%)
41%
29%
23%
2%
1%
0.7%
0.6%
0.4%
0.5%
0.1%
< 0.1%
< 0.1%
< 0.1%
< 0.1%
< 0.1%
0.03%
0.002%
0.005%
0.002%
< 0.001%
< 0.0001%
0.00005%

The particle size distribution and specific surface area of Top-Crete are provided by Minplus CDEM .
The results are presented in table 3.7 and illustrated in figure 3.3.
Table 3.7: Specific gravity and median size

TC

Surface area
[m2/ g]
13.6

D10 [m]

D50 [m]

D90 [m]

8.12

10.05

50.09

36

120

Undersize weight % [kg/kg

100

80

60
Top-Crete
40

20

0
1

10

100

1000

Diameter [um]

Figure 3.3: Particle size distribution of Top-Crete by CDEM Minplus

3.1.3 Water
Fresh, clean tap water was used in all cases. The water does not contain materials that are harmful
for the blended cement paste.
3.1.4 Aggregates
Some experiments, like the strength and workability, were performed using mortar instead of
cement pastes. The sand used in the mortar specimen meets the requirements for standard norm
sand in mortar according to EN- 13139

3.2 Mixture proportions

A total of 7 binders were prepared by mixing the materials in different ratios. All mixtures were
composed of three cementitious materials except the control paste, which contained two materials.
In order to analyse the effects of the constituents on the binder properties of the mixture, a control
mixture was made for comparison purposes. Each set of blended cement pastes or mortars was
subjected to different curing conditions depending on the test method and/or the Dutch standard.
The proportions of the 7 mixtures are as follows*:

{1}

0% Top-Crete; 65% Portland cement; 35% Fly ash

{2}

5% Top-Crete; 60% Portland cement; 35% Fly ash

{3}

10% Top-Crete; 55% Portland cement; 35% Fly ash

{4}

15% Top-Crete; 50% Portland cement; 35% Fly ash

= Control paste/Reference

37

High fly ash mixtures

{5}

5% Top-Crete; 25% Portland cement; 70% Fly ash

{6}

10% Top-Crete; 20% Portland cement; 70% Fly ash

{7}

15% Top-Crete; 15% Portland cement; 70% Fly ash

*The mass of each component is expressed in percentages of the total blended cement.
Portland cement in the pastes and mortars was replaced by 5%, 10% and 15% TC and by 35% and
70% FA, respectively. The use of 35% FA in the ternary blends was expected to increase the
availability of the water content for the TC hydration. It was also expected that the heat dissipated
and CH produced by Top-Crete could serve as activator for high fly ash contents. The influence of the
increase of FA from 35% to 70% on the properties determines whether TC can be used as leverage
for supplementary cementing materials.
The amount of TC in the mixtures was optimized according to the results of previous experimental
work on activated paper sludge performed by different researchers [8], [12], [14], [15] and [79]. The
best results were achieved at the replacement level of 10% activated paper sludge. The mixture
compositions and designations which were used in this study are shown in table 3.8.
Table 3.8: Mixture proportions and designations
Mix
{1}
{2}
{3}
{4}
{5}
{6}
{7}

0% Top-Crete;
5% Top-Crete;
10% Top-Crete;
15% Top-Crete;
5% Top-Crete;
10% Top-Crete;
15% Top-Crete;

65% Portland cement;

60% Portland cement;
55% Portland cement;
50% Portland cement;
25% Portland cement;
20% Portland cement;
15% Portland cement;

35% Fly ash

35% Fly ash
35% Fly ash
35% Fly ash
70% Fly ash
70% Fly ash
70% Fly ash

PC(65)FA(35)
T(5)PC(60)FA(35)
T(10)PC(55)FA(35)
T(15)PC(50)FA(35)
T(5)PC(25)FA(70)
T(10)PC(20)FA(70)
T(15)PC(15)FA(70)

The designations T, PC and FA represent the materials Top-Crete, Portland cement and Fly ash. The
numbers following the designations indicate the mass percentages of the total blended cement paste
or mortar.
According to the European standard (EN 196-1): methods for testing cement, the proportions of the
components in the blends are listed in kg/m3 and outlined in tables 3.9 and 3.10. All specimen have a
constant w/b ratio of 0.5, despite of the different water absorption capacities of the mineral
admixtures.

38

Table 3.9: Mixture proportions of the composite pastes or mortars

CEM I 42,5 N (kg/m )

Water (kg/m3)
FA (kg/m3)
Top-Crete (kg/m3)
Normsand (kg/m3)
W/b
Curing temp (C)

{1}
292.5
225
157.5
0
1350
0.5
20

{2}
270
225
157.5
22.5
1350
0.5
20

{3}
247.5
225
157.5
45
1350
0.5
20

{4}
225
225
157.5
67.5
1350
0.5
20

Table 3.10: Mixture proportions of the high fly ash mixtures

CEM I 42.5 N (kg/m )

Water (kg/m3)
FA (kg/m3)
Top-Crete (kg/m3)
CEN Normsand (kg/m3)
W/B
Curing temp (C)

{5}
112.5
225
315
22.5
1350
0.5
20

{6}
90
225
315
45
1350
0.5
20

{7}
67.5
225
315
67.5
1350
0.5
20

3.3 Experimental techniques

In order to accomplish the objectives elucidated in the first chapter, an experimental program was
undertaken. The experimental program was designed to quantify the parameters established in the
objectives and to investigate the influence of TC on the early age properties of CEM IIB-V and high FA
blends. The tests performed on Top-Crete, the cement paste and the mortar are provided in table
3.11, 3.12 and 3.13.
Table 3.11: Tests performed on the mineral admixture TC
Test performed on TC

Method or standard

Particle size analysis

Laser Obscuration Time (LOT)

Specific surface area

BET nitrogen absorption method

Free lime

Glycol extraction method

39

Table 3.12: Tests performed on the cement paste

Test performed on the cement paste

Method or standard

Setting

NEN-EN-196-3

Porosity

Mercury intrusion porosimetry

Microstructure

Environmental scanning electron microscopy

Hydration heat and degree of hydration

Isothermal calorimeter

Chemical bound water and CH content

Thermo gravimetric analysis

Chemical shrinkage

Dilatometry method [geiker]

Autogenous shrinkage

ASTM C 1698 09

Table 3.13: Tests performed on the mortar

Tests performed on the mortar

Method or standard

Workability

Shake table test according to NEN 3534

Strength

NEN-EN 196

The following section describes the methods used for each test. A detailed description of the sample
preparation and the experimental procedure of the experiments is given in appendix A, except for
the unknown test methods.
3.3.1 Particle size analysis
The particle size distribution of TC was measured with the Eye Tech machine from Ankersmid. The
Eye Tech uses two methods to measure the particle size of materials. Particle size measurement by
means of Laser Obscuration Time (LOT) and by acquiring video images of moving particles. In this
study the LOT is used for the particle size distribution of TC and the settings of the apparatus are
listed in table 3.14. The Laser obscuration time method is based on the rotation of the laser beam
with a constant speed. The laser beam scans individual particles and is obscured as the particles are
encountered. The duration of the obscuration yields in the size of the particle.

40

Table 3.14: Setting of the Eye Tech machine

Measurement channel:
Liquid flow cell:

Laser
ACM-104A

Lens:
Flow control unit:

A100
AC-10

Min. particle size:

Max. particle size:
Stirrer:
Flow:

0.1 m
200 m
20
Backwards

Amount of cycles:
Time per cycle:
Speed:
Sonicator:

3
300 sec
30
On

Figure 3.4: Eye Tech machine

The time per cycle is of significant influence, because a short cycle will not give the whole size range
of the particles and a long cycle will give a distorted representation of the particle size distribution
due to difference in the speed of the particles. The amount of cycles was chosen to be three in the
settings to obtain accurate and stable results.
Before the first measurement of the PSD of TC, the apparatus was rinsed a couple times with ethanol
to ensure the cleanness of the system. Also a reference test was performed with micronized sand to
conform the functioning and accuracy of the machine. Finally, approximately 1.0 gram of the sample
is added to the ethanol.
3.3.2 Specific surface area (SSA)
The nitrogen adsorption method was applied to determine the specific surface area of a powder. The
N2 - adsorption was carried out using a Micrometrics Gemini VII 2390 V1.03 (V1.03 p). The
experimental apparatus is shown in figure 3.5.

41

Figure 3.5: Micrometrics Gemini VII 2390 V1.03 [Micrometrics]

This method is based on the physical adsorption of a gas on the surface of the solid material. Physical
adsorption depends on the forces between the adsorbate gas molecules and the adsorbent surface
of the powder. The calculation of the amount of adsorbed gas of a monomolecular layer on the
surface determines the SSA. The determination is performed at the temperature of liquid nitrogen
(-195,75 C).
The specific surface area was determined according to the BET- equation:

Where:

V
Vm
P
P0
C

= Volume adsorbed;
= Volume of monolayer;
= Sample pressure;
= Saturation pressure;
= Constant related to the enthalpy of adsorption;

The specific surface area (SBET) is than calculated with the following equation:

Where:

na
am
m
Vl

= Avogrado constant (6.022 1023 mol-1);

= Cross sectional area of nitrogen molecule (0.162 nm2);
= Weight of the sample
= molar volume of nitrogen gas (22414 cm3);

The measurements were conducted on two parallel samples to achieve accurate results.
42

3.3.3 Free lime content

3.3.3.1 Introduction
Various methods have been introduced for the determination of free lime (CaO) in cement and
cement based compositions. These methods show that the amount of free lime can be measured
chemically with chemical solutions or analytically with new computer techniques, such as X- ray
diffraction and X- ray fluorescence. To reduce the time for the determination of free lime, rapid
chemical techniques have been proposed by different researchers [40][41][42]. It has been found
that the methods using water or dilute acids have to be precluded, because cement compounds are
dissolved differently in water or acids. The results from these measurements are untrustworthy.
Dissolutions in acids or Ethylenediaminetetraacetic (EDTA) are methods that have been used
generally for the determination of free lime. These methods are accepted as a quantitative
indication, but solvent and kinetic problems of the dissolutions still occur. In the presence of calcium
hydroxide the methods also dissolve calcium hydroxide making the obtained results inaccurate.
Some of the methods also cant distinguish hydrated lime and hydrated calcium silicate [43].
The glycerol technique of Lerch and Bogue has been used for many years [44] for the determination
of uncombined lime. The rate of dissolution of lime in glycerol was at that time unknown. Later it was
found that glycerol has the ability to solve free lime, but also to solve calcium from the cement
compounds and calcium hydroxide. From other researches, Schlpfer and Bukowski, it is known that
previous mentioned problems can be reduced by using glycol instead of glycerol [45]. Therefore,
methods using glycol give satisfactory results. The free lime amount is therefore measured with the
Glycol extraction method in accordance with the procedure described in ENCI BV, Maastricht; August
1978.
3.3.3.2 Glycol extraction method
The glycol method used by ENCI is based on extraction of CaO in clinker and cement with ethylene
glycol. This method can only be used when the hydrated pastes are grinded into an impalpable
powder. The glycol extraction method does not yield in an absolute value.
3.3.3.2.1 Experimental procedure
The powders TC, PC and FA or the cement paste after 28 days have to be grinded into an impalpable
powder. 1,000 gram of each mixture is accurately weighed on an analytical balance with an accuracy
of 1 mg. The different powders are weighed separately and transferred to a dry Erlenmeyer flask of
100 ml. 50 ml of ethylene glycol and a magnetic stirring are added in the Erlenmeyer. After shaking
the composition to disperse the calcium oxide, the Erlenmeyer flask is closed with a cork stop and
placed in a water bath of 65 C to 70 C. The solution is boiled and shaken automatically and
continuously by the magnetic stirring for 30 minutes. Subsequently the solution is swiftly filtered
through a Whatman No. 42 filter paper by vacuum filtration. The filtrate, particles smaller than 2,5
m, is collected in a Buchner funnel of 250 ml. The Erlenmeyer is rinsed with 10 ml ethylene glycol
and filtered in the Buchner funnel. This activity is repeated twice. After the filtration has finished, 5
droplets of Broomthymolblauw are added to the filtrate. Then the filtrate is titrated by 0,10 mol(M)
hydrochloric acid until the color changes from blue into yellow. The experiment has been completed
if no color change appears.

43

3.3.3.2.2 Determination of free lime

This method determines the free lime by first converting the free lime to Ca-glycolate. Then the
dissociation of Ca-glycolate takes place and the glycolate ion is titrated with HCl. The equations are
presented below.
(1)

C2H4(OH)2 + CaO

C2H4O2Ca + H2O

(Reaction boiling for 30 minutes)

(2)

C2H4O2Ca

(C2H4O2)2- + Ca2+

(Dissociation Ca - glycolate)

(3)
(C2H4O2)2- + 2H3O+

C2H4(OH)2 + 2H2O
------------------------------------------------------------------CaO + 2H3O+

Ca2+ + 3H2O

(Titration with 0,1 M HCl)

(Total reaction)

The total reaction shows that for each M CaO two M H3O+ are added. The Molmass of H3O+ is 56,08
and the equivalent mass of CaO is 28,04. For every gram sample there is x ml 0,100 M HCl titrated to
change the color from blue to yellow.

The formula for the determination of free lime follows from the next equation:

In which:

x = amount of titrated 0,100 M (ml)

a = weight sample; (g)

3.3.3.2.3 Comments on the experiment

This test method does not always provide accurate results because the glycerol method was
developed primarily for the extraction of free lime in Portland cement. Therefore, this method has
some limitations [48]:
Fineness and Ca(OH)2
The fineness of the powders has influence on the amount of free lime. More free lime is determined
when the particles of the blends are finer. From a previous report [48] it is observed that hydrated
cement grinded into an impalpable powder shows 0.4 0.7 % more free lime than the hydrated
cement from the original specimen. It is suggested that larger particles could be surrounded partially
or completely by other constituents, making the first reaction (1) going slow. According to this
information it is impossible to compare the free lime of the blended powders before hydration and
after hydration because the fineness of the powders is larger than the fineness of the grinded
cement paste.
CH may dissolve in ethylene glycol, therefore this method is not an absolute method for materials
containing and producing CH.

44

Influence of the temperature

The mentioned reactions in paragraph 4.2.2.3 occur at a temperature between 65 and 70C. At lower
temperatures the reactions will slow down, so that the results of free lime obtained are reduced.
Also when the solution is filtrated the temperature cannot cool too much, this will result in lower
values.
Influence of the titration
When titration with HCl is not performed immediately after removing the Bchner funnel this could
lead to disadvantages. First the temperature of the filtrate can cool down or when the filtrate is still
boiled the filtrate can evaporate. Furthermore carbon dioxide can penetrate in the Erlenmeyer and
react with the free lime.
Influence of gypsum
From previous reports [45] it is suggested that Gypsum is soluble in ethylene glycol and leads to
higher free lime results.
3.3.4 Setting
The initial setting time and the final setting time of the mixtures were performed and determined
according to the procedure outlined in NEN-EN-196-3 Beproevingsmethoden voor cement Deel 3:
Bepaling van begin en einde van de binding en bepaling van de vormhoudendheid (see appendix A
for a detailed description). For each mixture the initial and final setting time was determined from
two tests. The average is taken as setting time in the results.
3.3.5 Pore structure
Mercury Intrusion Porosimetry (MIP) ( figure 3.6) was used to examine the pore size distribution, the
porosity and the effective porosity of the blends at curing periods of 3, 14, 28 and 56 days. The MIP
provides information on the geometry of the pores and on the connectivity of the pores. This
method is based on a liquid that enters a porous sample under pressure. The pressure that is needed
to enter the pores is determined with the Washburn equation:

In which is the surface area of the liquid (480 N/m), is the contact angle (139) and d is the pore
diameter (m). The Washburn equation assumes that the pores are cylindrical.

45

Figure 3.6: the Micrometrics PoreSizer 9320 and supplies for sample preparation
3.3.6 Microstructure development
The microstructure development of the hardened cement blends has been studied with the scanning
electron microscope Philips-XL30-ESEM (figure 3.7) at the curing ages 3, 14, 28 and 56 days. The SEM
is one of the most versatile instruments available for the analysis of the microstructural
characteristics of solid objects. The images obtained from the backscattered mode (BSE) provide
information on the distribution of various phases presented in the hydrated sample.

Figure 3.7: The Philips-XL30-ESEM at the TU Delft

3.3.7 Hydration heat and degree of hydration
The measurements of the heat evolution of the blended cement pastes have been carried out with a
thermometric isothermal conduction calorimeter (TAM- Air-314) as shown in figure in 3.8. The
calorimeter monitors continuously the heat evolved from the specimen with time when the
temperature of both the specimen and the surrounding environment are maintained at isothermal
conditions. The hydration heat and the degree of hydration of the blended cement pastes were
measured for a period of 7 days.

46

Figure 3.8: A thermometric isothermal conduction calorimeter (TAM- Air-314)

After the measurement of 7 days, the rates of heat evolution per gram binder were calculated from
every mix design. The average value from four samples was used for the results.
The development of the degree of hydration of the blends at a specific time t can be estimated by
the ratio between the generated heat at time t and the total heat of the binder [34]:

Where (t) is the degree of hydration at age t, Q(t) [J/g] is the accumulated generated heat at time t
during the hydration process, t [h] is the time that is converted to an age that expresses the maturity
of the material during the hydration process, and Qpot is the potential heat produced by the different
materials separately. Which can be defined with the material components and chemical
compositions. Consequently, the potential heat is determined using the following formula [34]:

In which:

Pcem and Ppoz1,2 are the mass percentages of the materials; (%)
Qcem is the potential heat of cement (453 J/g);
Qpoz1 is the potential heat of fly ash (209 J/g [18]);
Qpoz2 is the potential heat of Top-Crete ( 460 J/g);

The total heat release by the reactions of TC is unknown and therefore assumed to be slightly higher
than the total heat released by PC.
3.3.8 Chemical bound water and CH content
The chemically bound water and the CH content of the specimen were measured with the thermo
gravimetric analysis (TGA) at the curing periods of 3, 14, 28 and 56 days. TGA was carried out using a
47

NETZSCH Simultaneous Thermal Analyzer TG-449-F3-Jupiter. The experimental apparatus is shown in

figure 3.9. This method is based on the combustion of specimen in the furnace of the thermo balance
of the TGA under a controlled atmosphere to obtain thermo gravimetric curves. The TGA measures
the amount and rate of weight loss in a material as function of time or temperature changes up to a
temperature of 1090 C. The results are used to define the composition of materials. Only materials
that exhibit weight change due to decomposition or dehydration can be characterized.

Figure 3.9: NETZSCH Simultaneous Thermal Analyzer TG-449-F3-Jupiter

The chemically bound water is calculated according to the formulas presented in appendix A. The
relation between TG curve and the CH content is determined by a graphical technique mentioned by
Marsh et al. [36]. More information on this method is also provided in appendix A.
3.3.9 Chemical shrinkage
The accurate measurement of chemical shrinkage in blended cement pastes is very difficult. The
measurement has to take place immediately in sealed condition without the disturbance of the
structure of the blended cement paste. Over the years, new methods have been developed to
determine the chemical shrinkage of cement pastes. Henri Le Chatelier was the first to evolve a
method for the determination of chemical shrinkage of 100% cement with a w/c ratio of 0.4. Le
Chatelier measured the volume contraction of the paste in a glass beaker in connection with a waterfilled tube. This contraction is explained as the difference in density between the initial products and
the formed hydration products. The absolute volume change is almost 10% at full hydration, which is
in practice never achieved due to resistance of the rigid cement paste. However, Chateliers method
gives a significant and accurate part of the chemical shrinkage before setting [90]. Powers [91] also
provided another measuring method to quantify the chemical shrinkage in cement. The method is
based on the measurement of shrinkage of the individual clinker compounds in cement. The results
were in agreement with the method of Le Chatelier. A direct method which is used nowadays for the
measurement of chemical shrinkage is the dilatometry test. This method follows Le Chateliers
contraction principle and has been applied by many researchers, such as Geiker [92]. In this study the
chemical shrinkage of cement blends was tested according to the method described by Geiker, see
figure 3.10. The tests were performed for a period of seven days and the average of three samples
was taken as result for the chemical shrinkage of the mixture.

48

Figure 3.10: Chemical shrinkage test method according to Geiker with a small glass jar enclosed with
a rubber stop encasing a pipette [Ambrosia et al, [87]]
3.3.10 Autogenous shrinkage
Different test methods have been developed to determine the autogenous strain of the cement
pastes using either volumetric or linear measurements. Autogenous strain is a synonym of
autogenous shrinkage, but the term autogenous strain is more general since it enables expansions to
also be considered [93]. The volumetric strain is defined by the weight change of a paste casted in a
rubber balloon immersed in a fluid, while the linear strain is determined by the measurement of the
length change of the paste placed in a rigid mold with displacement transducers at each end. Both
methods have advantages and disadvantages [93].
Volumetric measurement
Advantage: The measurements can start immediately after mixing the materials.
Disadvantage: Lack of contact between the cement paste and the material of the balloon
due to water or air. This water or air can be sucked back into the cement paste and reduce
the volume change. An additional problem is the osmosis of water through the balloon.
Linear measurement
Advantage: The length change is clearly measured by firm anchorages at each measuring
point.

Disadvantage: The measurement can only start when the paste has set. There is also the risk
of restraining the cement paste to the mold. Bleeding of the paste has also effect on the
results due to reabsorption of the water.

The autogenous strain of the blended cement paste is measured in this study according to ASTM C
1698 09 [94]. The measurement of each mixture takes place from final setting time until 28 days of
hardening. This method combines the advantages of both the linear and volumetric measurements,
49

while avoiding their disadvantages. The autogenous strain is measured with a dilatometer bench
using sealed, flexible corrugated plastic tubes with two end plugs (see figure 3.11). The material of
the tube prevents osmosis of water and the triangular-shaped corrugations minimize the restraint of
the paste to the mold. The bulk strain of the specimen is measured in sealed conditions at constant
environmental conditions.

Figure 3.11: A dilatometer bench and corrugated plastic tubes with two end plugs [94]
3.3.11 Workability
The consistency for each specimen is measured with the flow table test ( see figure 3.12)
immediately after mixing the materials with water. The tests were performed according to NEN EN196-1 methods of testing cement Part 1: Determination of strength.

Figure 3.12 :Flow table for the measurement of the workability.

50

3.3.12 Strength
The bending strength and the compressive strength of the mortars were determined according to
the procedure outlined in NEN-EN-196-1 Beproevingsmethoden voor cement Deel 1: Bepaling van
de sterkte. Three samples for each mixture were tested at 7, 14, 28 and 56 days. The attributes
which were used to determine the strength are illustrated in figure 3.13.

Figure 3.13: Attributes for the strength

51

3.4 Physical properties of Top-Crete

The use of Top-Crete as partial cement replacement influences the characteristics of paste and
mortar owing to both physical and chemical effects. The physical effects act upon the packing
characteristics of the specimen, which depends on size, shape and specific surface area of the TC
particles. Several experiments were performed to determine the physical properties of TC, such as
the specific surface area, the particle size distribution and microstructure. It is generally accepted
that an increase of the specific area and a decrease in particle size will expose a greater surface of
the particle to chemical reaction enhancing reactivity [19]. Knowledge of the microstructure of TC is
also important to distinguish TC particles form PC and FA particles.
3.4.1 Particle size distribution and specific surface area of TC
3.4.1.1 Results
The particle size distribution of Top-Crete was measured in the laboratory of Aidicio in Spain. The
summary of the median size results obtained by Aidico and CDEM are presented in table 3.15 and the
results are illustrated in figure 3.14.
Table 3.15: Specific surface area and median size

TC (CDEM)
TC (Aidicio)

D10 [m]

D50 [m]

D90 [m]

8.12
3.53

10.05
21.46

50.09
93.72

120

Undersize weight % [kg/kg]

100

80

60

Top-Crete (Aidicio)
Top-Crete (CDEM)

40

20

0
0,1

10

100
Diameter [um]

1000

10000

Figure 3.14: Particle size distribution of Top-Crete by CDEM Minplus and Aidicio
The particle size distribution by both measurements shows almost the same trend, but with different
values. The results from Aidicio show a much coarser distribution. The difference in the distribution
can be assigned to the various methods and apparatus used for the measurement and to the time
52

between the measurement and the production of Top-Crete. It is most likely that for long periods the
smaller particles are agglomerated together.
The specific surface area was measured for two samples using the nitrogen adsorption method in the
Microlab. The results of the specific surface area of both measurements are included in table 3.16.
Table 3.16: Specific surface area and median size
Surface area [m2/g]
TCsample1 (microlab)
TCsample2 (microlab)
TC (CDEM)

7.1553 0.038
7.2971 0.0404
13.6

The choice was made to use the PSD results provided by Aidicio and the SSA results measured in the
microlab. These results have recently been defined contrary to the results of CDEM, which were
determined a few years ago.
3.4.2 Microstructure of Top-Crete
3.4.2.1 Results
The experimental technique ESEM was used to observe the microstructure of TC. Figure 4.44 shows
an overview of the BSE images from the TC sample. It has been perceived that TC particles contain
irregular shapes, varies sizes and different grey colors. The shape differs from spherical to oblong as
shown in image (a). TC particles consist of tiny particles which are agglomerated into rigid structures
of different shapes and sizes. Such particles were subsequently observed in other parts of the
sample. This discovery distinguishes TC particles from the PC and FA.

53

(a)

(b)

(c)

(d)

(e)
Figure 3.15:(a) ESEM image of the cross section of Top-Crete particles at a magnification of 100x. (b)(e) Close ups of image (a) at a magnification of 800x
3.4.1.3 Conclusion
From the results it can be concluded that TC has a slightly finer particle size distribution and a much
larger specific surface area than PC and FA. The PSD curves are compared and illustrated in figure
3.16. Moreover, TC particles appear to be an agglomeration of very tiny particles which are mainly
spherical in the whole size range.

54

120

Undersize weight % [kg/kg]

100

80
Cem I 42,5 N
60

Fly ash
Top-Crete (aidicio)

40

Top-Crete (CDEM)

20

0
0,1

10

100
Diameter [um]

1000

10000

Figure 3.16: Comparison of the particle size distribution of TC (Aidicio and CDEM), PC and FA.
The mineral TC has particles with a small average diameter, has a relatively high specific surface area
and contains large properties of silica and aluminates (paragraph 3.1.2.2). Therefore, it can be
concluded that TC exhibits high pozzolanic activity.

55

56

Chapter 4. Results of the early age properties and microstructure

development of blended cement pastes with addition of TC
4.1 Introduction
The first objective of this research was to investigate the early age properties and the microstructure
development of the blended cement pastes made with the addition of Top-Crete. The partial
replacement of cement with Top-Crete will influence the early age properties of the cement paste,
i.e. workability, setting, bending strength and compressive strength, hydration heat and degree of
hydration. The microstructure development including the pore structure, CH phase content ant the
chemically bound water were also measured. These tests will improve the understanding of the
partial replacement of cement by Top-Crete. In this chapter the experimental results are presented
and discussed.

57

58

4.2 Workability
4.2.1 Introduction
The workability of concrete plays an important role in the construction process of structures. It has a
direct influence on the properties of hardened concrete, such as strength and porosity. Currently, the
workability can easily be increased with superplasticizers, but the usage of superplasticizers could
entail disadvantages like bleeding and segregation [20].
The workability of mortars is defined by different properties, such as the consistency, plasticity and
cohesion [21]. In this study the consistency by the flow table test is applied as a measure for the
workability, as described in section 3.3.11.
4.2.3 Results and discussion
The initial diameter expressed against the replacement level of cement with Top-Crete for the
mixtures containing 35% and 70% fly-ash is illustrated in figure 4.1. The initial diameter is the spread
of the paste after a small cone is filled with paste and lifted. The use of Top-Crete results in an
increase in initial diameter at 5% replacement level compared to the reference mixture, while at
higher replacement levels the initial diameter decreases compared to the reference mixture. The
increase in initial diameter might be assigned to the change made in the experimental procedure.
The reference mixture (PC(65)FA(35)) was compacted with 30 strokes, while the other mixtures were
compacted with 20 strokes. In all specimens, including the reference sample, the water to binder
ratio was 0.5. Due to the large amount of water available for the cement hydration, bleeding
immediately occurs in the reference sample. With every compaction stroke, the excess water caused
by bleeding will flow away from under the cone. When the number of strokes is increased, the total
amount of water that will flow away will also increase. Hence, the reference mixture shows less
consistency compared to the mixture containing 5% TC. The expected consistency is displayed as the
dash dotted line in figure 4.1.
10,8
The initial spread (35%
FA)
Initial spread (70% FA)

Initial diameter (cm)

10,6
10,4

The expected initial

spread (35% FA)

10,2
10
9,8
9,6
0

6
8
10
Replacement level with TC (%)

12

14

16

Figure 4.1: Initial spread for the mixes containing 35% and 70% Fly Ash

59

The increase in the amount of FA to 70% shows an increase in the initial diameter. As for the
mixtures containing 35% FA, the initial diameter decreases systematically with increasing
replacement levels of TC.
20
Spread with 35% FA

19

Spread with 70% FA

18

Diameter (cm)

17
16
15
14
13
12
11
10
0

10

12

14

16

Replacement level with TC (%)

Figure 4.2: Spread after 12 times shaking the table for mixes with 35% and 70% Fly Ash
Figure 4.2 shows the spread after 12 times drops of 10 mm of falling height expressed against the
replacement level of cement with Top-Crete. The influence of Top-Crete and fly ash is clearly
illustrated in this figure. The replacement of PC with TC has a substantial effect on the consistency.
An increase in the TC content shows a significant reduction in the consistency of the mixtures. Thus,
an increase in w/b ratio is required to maintain the same consistency as the reference sample. This
suggest that TC has a greater water demand than PC. On the other hand, the increase of the FA
content has a negligible effect on the flow value measured after 12 drops.
4.2.4 Conclusion
From the results it follows that at constant ambient temperature the increase in Top-Crete from 5%
to 15% leads to a substantial decrease in consistency. This is assigned to the high pozzolanic activity
of Top-Crete, which allows quick binding and hardening, and to the specific surface area of TC
particles, which is much higher than the specific surface area of PC and FA particles. Both physical
properties result in greater consumption of water.
The outputs also show that an increase in FA content does not result in a significant improvement in
workability. Therefore, the loss in workability due to the presence of Top-Crete cannot be
compensated by increasing the amount of fly ash from 35% to 70%.. And the workability depends
significantly upon the amount of Top-Crete and not on the fly ash content.
The use of superplasticizers will have a significant influence on these results, especially because TC
has a finer particle size distribution.

60

4.3 Setting
4.3.1 Introduction
The setting (change from plastic to solid state) is also an important property of concrete in the
construction field. Knowledge of the setting characteristics will improve scheduling the various stages
in concrete construction operations such as transporting, placing, compacting and finishing of
concrete [23].
4.3.2 Results and discussion
4.2.3.1 Results and discussion of the specimen containing 35 % fly ash.
The results from the initial and final setting time as function of the replacement level of TC are
presented in figure 4.4. For evaluation of the measurement results, it is assumed that the
contribution of FA to the hydration process in the first few hours is significantly lower than the
contribution of PC and TC due to the low pozzolanic activity of FA. Therefore, in this section the
effect of fly ash is not included and only the effect of Top-Crete will be discussed. When 5% of
cement is replaced with TC both the initial setting time and the final setting time show a small
increase. As the replacement level is increased to 10% TC a substantial decrease of almost 50%
appears in the setting time. However, the further increase of TC to 15% shows a small increase in the
setting time. It can be deduced that Top-Crete contributes to the setting in a non-systematic manner.
The influence of the different constituents in each mixture on the uptake of water by physical
absorption is partly determined by the workability test as mentioned in paragraph 4.2.3. It was
revealed that the mixtures with a higher TC content need more water to achieve the same
consistency as the reference sample. The large water requirement is a result of the large specific area
of TC (7.2262 m2/g), which is much bigger in comparison to that of FA ( 0.25 m2/g) and PC (
0.27m2/g), and the high chemical activity of TC. The increase in the amount of TC will cause the water
demand to increase, which leads to a decrease in the setting time.
14
Initial setting time
(35% FA)

12

Setting (hour)

10
8
6
4
2
0
0

10

12

14

16

Replacement level with TC (%)

Figure 4.4: Initial and final setting times for mixtures with 35% fly ash.

61

The results in figure 4.4 might be explained as follows. At low TC to PC ratios, the reaction of MK with
the abundant supply of CH will lead to the hydration products C4AH13 and CSH gel. The formation of
C4AH13 contributes to a high water demand. However, at the level of 5% TC the supply of water is
limited due to the rapid consumption of water by the hydration of PC [28]. Hence, the hydration
process is and the setting time is increased. As the level of TC is increased to 10%, the setting time is
significantly decreased. This decrease is attributed to the increase in available water due to the
reduction of the PC content for the pozzolanic reaction of MK. The reaction of TC will go much faster
as the water supply becomes larger and more cementitious reaction products are formed between
the cement particles. This leads to a decrease in setting time. The setting time suddenly increases as
the TC level is further increased to 15%. Although, the water available for the pozzolanic reaction is
further increased the supply of CH has become limited due to the reduction of the PC content. The
supply of CH will become insufficient to sustain the pozzolanic reaction of MK. This delays the
formation of hydration products and increases the setting time. To confirm these suggestions more
experiments are needed to determine the chemical reaction process.
4.2.3.2 Results and discussion of the specimen containing 70 % fly ash.
The final setting time expressed against the replacement level of cement with FA is shown in figure
4.5. Figure 4.5 displays a substantial increase in final setting time with increase in FA content for the
mixtures containing 10% and 15% TC. The specimen with 5% TC have a negligible effect on the
setting time the with the increase in FA content. The inconsistent increase in setting time at the
various replacement levels suggests that FA and TC do not behave independently.
25

FInal setting time (hours)

20

15
5% TC
10% TC

10

15% TC

0
0

10

20

30
40
50
Replacement level with FA

60

70

80

Figure 4.5: Final setting times as function of the replacement level with FA at 5%, 10% and 15% TC.
Most of the mineral admixtures with a low pozzolanic activity and a low specific surface area, like FA,
have a delayed effect on the setting time. There are numerous studies in which the retardation of the
hydration process of FA blends is explained. Hydration reactions of the constituents, which form an
interconnected network, are controlled by the amount of available water and the space between the
reacting particles [29]. To achieve a reduction in setting time the formation of an interconnected
62

network is of great importance. In the early stage of hydration, FA is considered to be an inert filler
due to the low pozzolanic activity of FA. The increase in FA content increases the distance between
the reacting particles, wherefore the formation of an interconnected network by PC is delayed and so
is the setting time.
4.3.3 Conclusion
Based on the experimental results of the setting time, the following concluding remarks can be
made:
The results clearly show that both minerals have different influences on the setting time.
I.

II.

The increase in the TC content contributes to the setting time in a non-systematic manner.
This behavior is most likely caused by the pozzolanic reaction of TC with CH, the PC
hydration and the availability of water and CH that sustains these reactions.
The increase in the amount of FA leads to a substantial increase in setting time at all TC
levels. In the early age of hydration, FA is considered to act as inert filler and less hydration
products are formed due to the reduction in PC content. Hence, increasing the time period at
which hydration products are connected.

63

64

4.4 Strength
The strength of concrete is defined as the ability of concrete to withstand forces, stresses and
pressures without failure. Among all properties of concrete, strength is generally considered one of
the most valuable. The strength gives an overview of the quality of the concrete because it is directly
related to the hydrated cement matrix.
The strength of concrete is separated into flexure, compression, tension and splitting tension. The
strength does not only depend on the strength properties of the materials, but also on the bond
strength between the aggregates and binder. The most commonly performed experiments are the
compression and flexural tests.
4.4.3 Results and discussion
4.4.3.1 Results of the bending strength
The results of the strength test are an average of three prism measurements. The prisms have a size
of 40x40x160 mm. Figure 4.6 presents the bending strength data obtained from the flexure tests for
the blended mortars containing 35% FA as a function of the curing time. The profile of the bending
strength becomes shallower with an increase in curing time. Hence, the mixtures are approaching a
constant value.
7,0

Bending strength (MPa)

6,0
5,0
4,0
PC(65)FA(35)

3,0

T(5)PC(60)FA(35)
2,0

T(10)PC(55)FA(35)
T(15)PC(50)FA(35)

1,0
0,0
0

10

20

30
Time (days)

40

50

60

Figure 4.6: Bending strength of mortars with 35% FA

The effect of the amount of TC on the bending strength at 7, 14, 28 and 56 days is illustrated in
figures 4.7 and 4.8 for the mortars with 35%. All curves show a loss in bending strength with increase
in Top-Crete content. There is a slight reduction in bending strength between the replacement levels
5% and 10%. However, the rate of reduction is lower between the replacement levels 0% - 5% and
10% - 15% TC at all curing ages.

65

7,0

Bending strength (MPa)

6,0
5,0
4,0
3,0

7 days with 35% FA

14 days with 35% FA

2,0

28 days with 35% FA

1,0

56 days with 35% FA

0,0
0

10

12

14

16

Replacement level with TC (%)

Figure 4.7: Bending strength against the replacement level with TC for the mortars with 35% FA
4,0
7 days with
70% FA
14 days with
70% FA
28 days with
70% FA

Bending strength (MPa)

3,5
3,0
2,5
2,0
1,5
1,0
0,5
0,0
0

10

12

14

16

Replacement level with TC (%)

Figure 4.8: Bending strength against the replacement level with TC for the mortars with 70% FA
Figure 4.9 displays the bending strength versus the curing time for the mortars containing 70% FA.
The mortars with 35% FA show a 70% higher bending strength compared to the mortars with 70% FA
at all times. Although, the mortars with 70% FA have a lower bending strength, the rate of strength
increases is higher than the mortars with 35% FA. Therefore, the difference in strength between the
two mixtures becomes smaller at every curing age. Especially from 28 to 56 days the mortars with
70% FA still show an increase in strength, while the mortars containing 35% FA are approaching a
constant value.

66

4,5
4,0

Bending strength (MPa)

3,5
3,0
2,5
2,0
T(5)PC(25)FA(70)

1,5

T(10)PC(20)FA(70)

1,0

T(15)PC(15)FA(70)
0,5
0,0
0

10

20

30
Time (days)

40

50

60

Figure 4.9: Bending strength of mortars with 70% FA

4.4.3.2 Results of the compressive strength
The results of the compressive strength expressed against the curing time for the mortars containing
35% and 70% FA are shown graphically in figure 4.9 and figure 4.10. Both figures show an increase in
compressive strength with increasing curing time at all three TC levels. Figure 4.9 shows that the
difference between the reference sample and the mixture T(5)PC(60)FA(35) is quite small. After 56
days this difference increases abruptly. A possible explanation for this phenomenon is that FA might
start to contribute to the strength after 28 days, but the availability of CH in the specimen
T(5)PC(60)FA(35) is reduced by the pozzolanic reaction of TC.
60,0

Compressive strength (MPa)

50,0
40,0
30,0
PC(65)FA(35)
20,0

T(5)PC(60)FA(35)
T(10)PC(55)FA(35)

10,0

T(15)PC(50)FA(35)

0,0
0

10

20

30
Time (days)

40

50

60

Figure 4.9: Compressive strength of mortars with 35% FA

67

The increase in FA content to 70% leads to a decrease in compressive strength of approximately

61%. In contrast to the bending strength, the strength of the samples with 70% FA increases with a
higher rate compared to the rate of the specimen with 35% FA. However, the mortars with 35% FA
show at 56 days an increase of the compressive strength at a higher rate.

Compressive strength (MPa)

25,0

20,0

15,0

T(5)PC(25)FA(70)
T(10)PC(20)FA(70)

10,0

T(15)PC(15)FA(70)
5,0

0,0
0

10

20

30
Time (days)

40

50

60

Figure 4.10: Compressive strength of mortars with 70% FA

Figure 4.11 illustrates the effect of TC replacement on the compressive strength at 7, 14, 28 and 56
days for the mortars with 35% FA. At higher replacement levels the strength is substantially lower
than the strength of the reference mortar, which is consistent at all curing ages. Therefore, the use of
TC in FA blended mortars has a negative effect on the strength of the blended mortars. However,
after 10% TC the strength did not change.

Compressive strenght (MPa)

60

7 days with 35% FA

14 days with 35% FA

50

28 days with 35% FA

40

56 days with 35% FA

30
20
10
0
0

10

12

14

16

Replacement level with TC (%)

Figure 4.11: Compressive strength against the replacement level of TC for the mixtures with 35% F.

68

4.4.3.3 Discussion of the compressive strength results

In this paragraph only the results of the compressive strength will be discussed, because the results
of both flexural strength and compressive strength show the same trend. The relative loss of the
compressive strength after 56 days versus the replacement level with TC is shown in figure 4.12. The
relative strength loss of the blended mortar compared to the strength of the reference mortar is
called the relative strength loss. This plot gives insight in the change of the strength with increasing
amounts of TC.
70

Relative strength loss (%)

60
Compressive strength
56d (35% FA)

50
40

Compressive strength
56d (70% FA)

30
20
10
0
0

6
8
10
Replacement level with TC (%)

12

14

16

Figure 4.12: The relative compressive strength loss as function of the replacement level at 56 days for
35% FA and 70% FA.
It is generally observed that when TC or FA substitutes PC, lower strengths than the reference
mixture are obtained at all curing ages up to 56 days. The degree of the early strength retardation is
in direct relation to the amount of TC or FA. Although the compressive strength of the pastes was
reduced by both materials, it could be assumed that the strength loss is provoked by different
mechanisms. The assumption made is supported by the different properties of the materials. It is
also noted that the different TC replacement levels in high fly ash mixtures have similar effect on the
strength development.
There are a couple of factors influencing the contribution of TC to the strength when it partially
replaces PC in the cement paste. The strength might be influenced by the dilution effect of PC, the
pozzolanic reaction of TC with CH and the combination of the PC hydration and the pozzolanic
reaction of TC. It is also suggested that the strength loss might be influenced by the formation of new
weaker compounds, the coarsening of the pore structure in the hardened cement paste and the
deterioration of the microstructure in the transition zone between the aggregate and the cement
paste [49]. In order to understand the mechanism of the compressive strength development in TCPC-MK ternary systems, further analysis has to be conducted on paste specimens to characterize the
presence of phases.

69

The influence of 70%FA on the strength is defined by the available amount of CH, which is in these
mixtures assumed to be too small to enable the TC and FA pozzolanic reactions to compensate for
the large PC deficiency.
4.4.4 Conclusion
From the results obtained from the compressive and bending strength tests, the following
conclusions can be drawn:

The increase in curing age leads to an increase in strength, whereas the increase in Top-Crete
leads to a decrease in bending and compressive strength.
The mixtures with larger amounts of fly ash even show a larger reduction in strength of
almost 60%. This is attributed to the low available CH content for the pozzolanic reaction of
TC and FA, due to the reduction of the PC content.
The replacement of PC with different levels of TC in high fly ash mixtures does not affect the
strength.

In the performed test, the presence of Top-Crete does not compensate for the strength that is
reduced by the replacement of Portland cement by higher amounts of fly-ash.

70

4.5 Hydration heat and degree of hydration

4.5.1 Introduction
The heat release during cement hydration is an important factor affecting the setting and other
properties of the blended paste. When cement is partially replaced with the highly reactive TC, the
reaction of TC could lead to an excessive heat development. The total heat generated by the blended
cement paste could cause a substantial rise in the temperature of massive concrete structures during
curing. The temperature rise occurs in the inner core of the element, because the surface of the
element is able to release its heat to the environment. The non-uniform cooling might generate
thermal stresses, which can exceed the tensile strength of the concrete. These tensile stresses lead
to thermal cracks, loss of strength and reduced durability of the structure. Therefore, knowing the
effect of TC on the heat release is important in the construction phase especially in massive
structures.
The behavior of TC when it was mixed only with water is interesting. When 5 gram of TC was mixed
with 2.5 gram of water the mixture did not change into a paste, because all of the water was
physically adsorbed by the surface of the TC particles at the surface of the mixture. After the raise of
the water to binder ratio to 1.0, the mixture became sticky but did not have a workable consistency.
Furthermore, some parts of TC did not come into contact with the added water because the water
content was too low to fully saturate all TC particle. Hence, the water to binder ratio was increased
to 1.5. With this w/b ratio TC finally became a paste with a workable consistency. Therefore this
water to binder ratio was used to determine the heat evolution and degree of hydration of the TC
samples in the isothermal calorimeter. The greater water demand of TC has been discussed
previously in paragraph 4.2 and 4.3.
The heat evolution of TC is measured for unary, binary and ternary mixtures. The unary cement
pastes contain 100% TC with a w/b ratio of 1.5. The binary blends were composed of Portland
cement and Top-Crete in amounts of 5%, 10% and 15% by mass with a w/b ratio of 0.5. The mixture
proportions of the binary blended cement pastes used in this experimental study are outlined in
table 4.2. The ternary blends consisting of PC, TC and FA are similar to the mixtures described in
chapter 3 with a water to binder ratio of 0.5 (see tables 3.8 and 3.9)
All unary, binary and ternary blended cement pastes have a similar preparation procedure.
Table 4.1: Mixture proportion and designation of the TC paste
Mixture
[1]

100% Top-Crete

TC(100)

Table 4.2: Mixture proportion and designation of binary blends

Mixture
{1}
{2}
{3}
{4}

0% Top-Crete;
5% Top-Crete;
10% Top-Crete;
15% Top-Crete;

100% Portland cement;

95% Portland cement;
90% Portland cement;
85% Portland cement;

PC(100)
T(5)PC(95)
T(10)PC(90)
T(15)PC(85)

71

4.5.2 Results and discussion

4.5.2.1 The heat evolution of Top-Crete
The rate of heat evolution of TC and PC as function of curing time is plotted in figure 4.13. The results
of the hydration heat are an average of four samples. The cumulative heat and the degree of
hydration of TC compared to PC expressed against the curing time are presented in figures 4.14.

Figure 4.13: Rate of heat evolution in the first seven days of Top-Crete and Portland cement
From figure 4.13, it was observed that the first peak in the hydration heat curve of 100% PC appears
to be from the reaction of gypsum and C3A with water. However, the first peak in the hydration heat
curve of TC cannot be attributed to the reaction of gypsum or C3A with water because TC is only
composed of free lime and MK. It is assumed that the reaction of CaO makes minor contribution to
the heat evolution, although the lime is consumed at a very rapid rate in the initial period of reaction
[82].Therefore, the first large peak in the hydration heat curve of TC is mainly attributed to the
pozzolanic reaction of MK with CH.
The course of the heat curve after the first peak in TC sample is explained by the initial rapid reaction
of MK with CH which is delayed. This is possibly due to the formation of hydration products at the
surface of TC particles thus reducing the heat evolution. The breakdown of this formed layer at a
later stage allows a renewed hydration of MK with CH, producing a second peak. This phenomenon
also occurs in the hydration process of PC. However, the second peak in the hydration process of TC
occurs earlier and faster, due to higher pozzolanic activity of TC.

72

0,7

250

0,6

Heat (J/g)

200
PC(100)

150

Top-Crete
100
50

Degree of hydration (%)

300

0,5
0,4
Top-Crete
0,3

PC(100)

0,2
0,1

0
0

100
200
Time (hrs)

300

100
200
Time (hrs)

300

Fig
Figure 4.14: Isothermal heat evolution and the degree of hydration of TC and PC as function of time
TC will generate more heat in seven days than PC when a w/c ratio is used of 1.5. This however does
not necessarily mean that the ultimate heat of hydration and the ultimate degree of hydration of TC
will be higher as well. Figure 4.14 shows that the degree of hydration and heat development initially
develops faster but the ultimate values might be lower. This phenomenon is mainly attributed to the
lower available CH content in the TC samples that is needed for the pozzolanic reaction of MK.
4.5.2.2 The heat evolution of binary TC-PC blends
Figures 4.15 represents the evolution of the hydration heat of the binary blended cement pastes
made with addition of 0%, 5%, 10% and 15% TC. The first heat evolution peak is not distinguished
clearly due to the scale implemented.
2,5
PC(100)
T(5)PC(95)

T(10)PC(90)

Heat (mW/g)

T(15)PC(85)
1,5

0,5

0
0

20

40

60

80
Time (hrs)

100

120

140

Figure 4.15: Rate of heat evolution of the binary blended cement pastes as function of time

73

Figure 4.15 clearly shows the effect of TC on the heat of hydration development in binary TC-PC
blends. The increase in the amount of TC replacement leads to a shorter and an earlier appearance of
the second peak. With the increase of TC content in binary blends the heat of hydration curve of PC
is shifting more towards the curve of TC as shown in figure 4.16. Moreover, the replacement of
Portland cement with 10% TC or more introduces a third shoulder in the heat evolution curve before
the end of the deceleration of the curve. This peak becomes more prominent for larger replacement
levels.
The earlier and shorter appearance of the second peak signifies the acceleration of the hydration
process of the binary blended cement paste due to the presence of the highly reactive TC. This
acceleration is explained by the fact that the sulfate concentration in the pore solution would drop
more rapidly as the MK levels increase [84]. Sulfate in the pore solution is necessary to maintain
stability of the ettringite phase. The presence of TC compromises the ettringite layer around the
reacting particles causing the layer to break down much faster. The particles now continue to
hydrate much earlier leading to an earlier appearance of the second peak.
As mentioned in the literature study a similar, but less distinct, third peak appears in the hydration
heat curve of some Portland cement pastes when ettringite is converted to mono sulphate. However,
this large peak did not appear in the paste with 100% PC. Therefore it cannot be appointed to the
conversion of ettringite to mono sulphate. The occurrence of the third peak might be attributed to
the C3S reaction and to the renewed pozzolanic reaction of MK with the relatively high concentration
of CH produced by both PC and TC. The hydration of C3S produces a large amount of CH that
stimulates the pozzolanic reaction of MK in TC and can therefore be considered as the reason for the
occurrence of the third peak in the combined hydration heat curve.
Figure 4.16 demonstrates the cumulative heat of hydration versus the curing time for the binary
pastes up to ten days. The heat evolution rate of the binary blends varies with increase in curing age.
It is important to indicate that between 0 and 20 hours the cumulative heat evolution increases with
the increase in the amount of TC. This is opposite to the cumulative heat evolution after 20 hours,
where the total generated heat decreases with increase in TC content. The cumulative heat of the
blends is influenced by the reaction of TC with the available supply of water at the different
compositions and curing ages. The increase in TC levels requires a larger water demand, reducing the
water supply for the PC hydration. The reduction of the PC hydration diminishes the formation of
hydration products and thus the total heat of hydration.

74

300

250

Heat (J/g)

200

150

PC(100)
T(5)PC(95)

100

T(10)PC(90)
T(15)PC(85)

50

0
0

50

100

Time (hrs)

150

200

250

Figure 4.16: Isothermal heat evolution in the first seven days of the cement pastes made with PC and
TC
Figure 4.17 shows the total generated heat at 10 days as function of the replacement level with TC.
The increase in the amount of TC in the blended cement pastes reduces the total heat generated by
the constituents of the blends. It is also observed that the cumulative heat of the pastes containing
5% of TC is similar to the total heat of pastes with 10% TC. However, the total heat of each binary
blend at 10 days is lower than the heat generated by the reference sample. This was expected since
the reaction of TC is limited by the availability of water, which is lower with increase in the TC
content.
275
10 days

Heat (J/g)

270

265

260

255

250
0

6
8
10
Replacement level with TC (%)

12

14

16

Figure 4.17: Isothermal heat evolution as function of the replacement level at 10 days

75

4.5.2.3 The heat evolution of ternary TC-PC-FA blends

The results of the rate of heat evolution of all ternary blended cement pastes are illustrated in figures
4.18 and 4.19. Comparison of figures 4.15 and 4.18 shows a remarkable similarity in the shape of the
curves. Therefore, the effect of the TC content in the ternary blends on the heat of hydration heat
curve is similar to that of the binary blends. The mixtures with 70% FA show a drop in the maximum
rate of heat evolution relative to the reference mixture and the specimen containing 35% FA, as
shown in figure 4.19. This indicates that FA has a retarding effect on the hydration process and thus
on the development of the hydration heat.
1,6
PC(65)FA(35)
1,4

T(5)PC(60)FA(35)

Heat (mW/g)

1,2

T(10)PC(55)FA(35)

T(15)PC(55)FA(35)

0,8
0,6
0,4
0,2
0
0

20

40

60

80

100

120

140

Time (hrs)

Figure 4.18: Rate of heat evolution of the blended cement pastes with 35% FA as function of time
0,6
T(5)PC(25)FA(70)
0,5

T(10)PC(20)FA(70)
T(15)PC(15)FA(70)

Heat (mW/g)

0,4
0,3
0,2
0,1
0
0

20

40

60

Time (hrs)

80

100

120

140

Figure 4.19: Rate of heat evolution of the blended cement pastes with 70% FA as function of time

76

The discrepancy in the generated heat due the increase in FA content is simply explained with basic
reaction kinetics. The pozzolanic reaction of FA with CH proceeds very slow as result of the relatively
low specific surface area of FA and the low solubility of the alumino- silicate glass in the alkaline
environment created by the hydration of PC [85]. Hence, FA is considered to act as an inert
supplement in the first 160 hours during the test and more water is made available for the hydration
of TC and PC. However, the reduced amount of PC diminishes the amount of produced hydration
products and the total heat of hydration.

200

100

180

90

160

80

140

70

120

60

Heat (J/g)

Heat (J/g)

Figure 4.20 shows the cumulative heat of the ternary blended cement pastes with 35% and 70% FA
over a period of seven days. The degree of hydration as function of time for all ternary specimens is
plotted in figure 4.21.

100

50
40

80

PC(65)FA(35)

60

T(5)PC(60)FA(35)

30

40

T(10)PC(55)FA(35)

20

T(15)PC(55)FA(35)

20

T(5)PC(25)FA(70)
T(10)PC(20)FA(70)
T(15)PC(15)FA(70)

10
0

0
0

50

100
Time (hrs)

150

50

100
Time (hrs)

150

Figure 4.20: Cumulative heat evolution in the first seven days of the blended cement pastes with 35%
and 70% FA.

0,35

0,18

0,3

0,25
0,2

PC(65)FA(35)

0,15

T(5)PC(60)FA(35)

0,1

T(10)PC(55)FA(35)

0,05

T(15)PC(55)FA(35)

0
0

50

100
Time (hrs)

150

Degree of hydration (%)

0,2

Degree of hydration (%)

0,4

0,16
0,14
0,12
0,1
0,08

T(5)PC(25)FA(70)

0,06

T(10)PC(20)FA(70)

0,04

T(15)PC(15)FA(70)

0,02
0
0

50

100
Time (hrs)

150

Figure 4.21: The degree of hydration of the specimens with 35% and 70% FA at 7 days
77

The specimens containing 35% FA show an increase in the cumulative heat and the degree of
hydration with increase in TC level at short hydration times (< 20 hrs) and long hydration times (>100
hrs). This is in contrast with the development of the total heat and degree of hydration between 20
and 80 hours, where the cumulative heat and the degree of hydration decrease with increasing TC
level. It can be deduced that the cumulative heat and the degree of hydration behave in a nonsystematic manner. This non-systematic behavior is suggested to be a function of the pozzolanic
reaction of TC with CH and water at the different compositions and curing times. The supply of CH is
governed by the rate of PC hydration, by the amount of PC and by the free lime hydration in TC,
whereas the rate of PC hydration and free lime hydration is influenced by the amount of available
water. With increase in TC level, there is an increase in water demand due to the large specific
surface area of TC and there is also less PC available to hydrate. Hence, the heat evolution reduces.
The complexity of the TC reactions in ternary blends requires a detailed study of the hydration
products to explain the heat evolution.
Meanwhile, the specimens containing 70% FA show a decrease in the cumulative heat and the
degree of hydration with increase in TC level for the measurement period after 20 hrs. The
replacement of PC with 70% FA and 5%,10% and 15% TC significantly diminishes the supply of CH
produced by PC. Thus, the pozzolanic reaction of TC and FA with CH is reduced and this significantly
reduces the total generated heat. Although, the pastes with 70% FA have lower initial heat, these
pastes could generate more heat than the control paste later. This follows from the slope of the
extrapolation. The heat measurement up to 7 days might therefore be insufficient to accurately
explain the heat evolution in the mixtures with 70% FA.
Figure 4.22 illustrates the cumulative heat at 7 days as function of the replacement level with TC for
specimens containing 35% and 70% FA. The increase in TC level in ternary blends slightly increases
the cumulative heat of the blended cement pastes containing 35%. This is in contrast to the heat
development of binary blends, which reduces with increase in replacement level.
200
180
160
7 days (35% FA)

140

7 days (70% FA)

Heat (J/g)

120
100
80
60
40
20
0
0

6
8
10
Replacement level with TC (%)

12

14

16

Figure 4.22: Isothermal heat evolution as function of the replacement level with TC at 7 days

78

The behavior of TC in the high fly ash mixtures is similar to that of binary blends. However, it is
suggested that the reaction of TC in compositions with 70% FA is governed by the CH content and not
by the available moisture content. The increase in replacement level of PC with TC reduces the
supply of CH produced by PC and hence reduces the pozzolanic reaction of TC and FA with CH.
Therefore, the cumulative generated heat diminishes with increase in TC level for the specimen with
70% FA.
4.5.3 Conclusion
The main conclusions that can be drawn from this study are summarized as follows:

Pastes containing 100% TC will generate more heat in seven days compared to the heat
evolution of 100% PC when these pastes are fully saturated with water. This however does
not necessarily mean that the ultimate heat of hydration and the ultimate degree of
hydration of TC will be higher as well.

Binary blends

The increase in the amount of TC in the binary blended cement pastes reduces the total heat
generated by the constituents of the blends. It is suggested that TC diminishes the PC
hydration due to the high water demand of TC. The reduction of the PC hydration contracts
the formation of hydration products and thus the total heat of hydration.

Ternary blends
The two pozzolans TC and FA in ternary blends influence the heat of hydration in different ways:

The increase of the TC level in the specimen containing 35% FA slightly increases the
cumulative heat of these ternary blends. This is in contrast to the heat development of
binary blends, which reduces with increase in replacement level.
As the FA replacement level increases to 70%, the cumulative heat and the degree of
hydration of the ternary blended pastes decrease by almost 50%. This is suggested to be due
the low pozzolanic activity of FA at the early age of hydration and the reduced CH content for
the pozzolanic reaction of TC and FA.

79

80

4.6 Porosity
4.6.1 General
Many physical and mechanical properties of concrete are influenced by the pore structure of the
cement paste. Moreover, the ability of concrete to resist diverse weather conditions, chemical
attacks and deterioration depends also on the pore structure. The permeability and strength are the
most important properties of concrete that are influenced by the pore structure. The purpose of this
section was to investigate the influence of Top-Crete and the influence of the increase in FA content
on the pore structure of the blended cement pastes.
4.6.2 Results and discussion
4.6.2.1 Total porosity of the mixtures with 35% and 70%FA
Experimental techniques differ in the accuracy of the measurements. In every measurement there is
obviously some measurement error involved. However, according to Ye [34] the variation of the
results obtained from the MIP experimental technique for multiple samples from one specimen is
less than 2%. Other researchers have found that this percentage is less than 4% [75] or 6% [76].
Therefore, the results obtained from one MIP measurement are sufficiently accurate.
The cumulative pore size distribution and the differential pore size distribution are plotted in figures
4.23 - 4.26 for the specimens with 35% FA. The plots of the blends containing 70% FA can be found in
appendix B. Table 4.6 summarizes the total porosity, i.e. the capillary porosity and gel porosity
excluding the dead end pores, of all mixtures at curing ages of 3, 14, 28 and 56 days. In the next
paragraph, the porosity and threshold pore diameter of the specimen with 35% FA will be analyzed
firstly, assuming that FA does not affect the hydration process up to 28 days and that only PC and TC
have an influence on the hydration process. After that, the porosity of the blends with 70% will be
examined. When the amount of FA is increased to 70%, the influence of FA will be discussed as well.
Finally, the effective porosity will also be described.
Table 4.6: Results of the total porosity at the age of 3, 14, 28 and 56 days
Age,
days
35% FA

70% FA

3
14
28
56
3
14
28
56

Total porosity (%)

Reference sample
39,9341
35,4956
35,3308
34,4664
-

5%TC
37,079
30,2586
27,128
23,581
39,4791
30,0304
23,1781
20,3496

10% TC
38,1487
33,3525
31,6865
28,6786
41,1686
35,5754
34,3725
30,4095

15% TC
37,6371
31,8985
29,5988
26,9769
40,922
33,2303
27,5728
22,235

81

3 days 1st intrusion

3 days 2nd intrusion

0,4

0,3

dV/dlogD [ml/g]

cum intr vol [vol/vol]

PC(65)FA(35)
0,5

0,5

0,4

0,3

0,2

0,2

0,1

0,1

0
0,001
0,4

0,1

10

1000

14 days 1st intrusion

14 days 2nd intrusion

3 days 1st intrusion

3 days 2nd intrusion

0
0,001

0,1

0,2

10

1000

14 days 1st intrusion

14 days 2nd intrusion

0,3

0,2

0,1

0,1

0
0,001
0,4

0,1

10

1000

28 days 1st intrusion

28 days 2nd intrusion

0
0,001
0,4

0,3

0,3

0,2

0,2

0,1

0,1

0
0,001

0,1

0,4

10

1000

56 days 1st intrusion

56 days 2nd intrusion

0
0,001

0,3

0,2

0,2

0,1

0,1

0,1

10

pore diameter [m]

1000

0
0,001

10

1000

28 days 1st intrusion

28 days 2nd intrusion

0,1

0,4

0,3

0
0,001

0,1

10

1000

56 days 1st intrusion

56 days 2nd intrusion

0,1

10

pore diameter [m]

1000

Figure 4.23: Pore size distribution of mixtures PC(65)FA(35)

82

cum intr vol [vol/vol]

3 days 1st intrusion

3 days 2nd intrusion

0,4

0,3

dV/dlogD [ml/g]

T(5)PC(65)FA(35)
0,5

0,4

3 days 1st intrusion

3 days 2nd intrusion

0,3

0,2

0,2
0,1
0,1

0
0,001

0,1

10

0,4

1000

14 days 1st intrusion

14 days 2nd intrusion

0
0,001

0,1

10

0,2

1000

14 days 1st intrusion

14 days 2nd intrusion

0,3

0,2

0,1

0,1

0
0,001
0,4

0,1

10

1000

28 days 1st intrusion

28 days 2nd intrusion

0
0,001
0,4

0,3

0,3

0,2

0,2

0,1

0,1

0
0,001
0,4

0,1

10

1000

56 days 1st intrusion

56 days 2nd intrusion

0
0,001

0,3

0,2

0,2

0,1

0,1

0,1

10

pore diameter [m]

1000

0
0,001

10

1000

28 days 1st intrusion

28 days 2nd intrusion

0,1

0,4

0,3

0
0,001

0,1

10

1000

56 days 1st intrusion

56 days 2nd intrusion

0,1

10

1000

pore diameter [m]

Figure 4.24: Pore size distribution of mixtures T(5)PC(65)FA(35)

83

3 days 1st intrusion

3 days 2nd intrusion

0,4

0,3

dV/dlogD [ml/g]

cum intr vol [vol/vol]

T(10)PC(65)FA(35)
0,5

0,3

0,2

0,2
0,1
0,1

0
0,001
0,4

0,1

10

1000

14 days 1st intrusion

14 days 2nd intrusion

0
0,001

0,1

0,2

10

1000

14 days 1 st intrusion
14 days 2nd intrusion

0,3

0,2

0,1

0,1

0
0,001

0,1

0,4

10

1000

28 days 1st intrusion

28 days 2nd intrusion

0
0,001

0,3

0,3

0,2

0,2

0,1

0,1

0
0,001
0,4

0,1

10

1000

56 days 1st intrusion

56 days 2nd intrusion

0
0,001
0,4

0,3

0,3

0,2

0,2

0,1

0,1

0
0,001

0,1

10

1000

0,1

0,4

0
0,001

Pore diameter (m)

10

1000

28 days 1st intrusion

28 days 2nd intrusion

0,1

10

1000

56 days 1st intrusion

56 days 2nd intrusion

0,1

10

1000

Pore diameter (m)

Figure 4.25: Pore size distribution of mixtures T(10)PC(65)FA(35)

84

3 days 1st intrusion

3 days 2nd intrusion

0,4

0,3

dV/dlogD [ml/g]

cum intr vol [vol/vol]

T(15)PC(65)FA(35)
0,5

0,3

3 days 1st intrusion

3 days 2nd intrusion
0,2

0,2
0,1
0,1

0
0,001
0,4

0,1

10

1000

14 days 1st intrusion

14 days 2nd intrusion

0
0,001

0,1

0,2

10

1000

14 days 1st intrusion

14 days 2nd intrusion

0,3

0,2

0,1

0,1

0
0,001
0,4

0,1

10

1000

28 days 1st intrusion

28 days 2nd intrusion

0
0,001
0,4

0,3

0,3

0,2

0,2

0,1

0,1

0
0,001

0,1

0,4

10

1000

56 days 2nd intrusion

56 days 1st intrusion

0
0,001

0,3

0,2

0,2

0,1

0,1

0,1

10

pore diameter [m]

1000

0
0,001

10

1000

28 days 1st intrusion

28 days 2nd intrusion

0,1

0,4

0,3

0
0,001

0,1

10

1000

56 days 2nd intrusion

56 days 1st intrusion

0,1

10

pore diameter [m]

1000

Figure 4.26: Pore size distribution of mixtures T(15)PC(65)FA(35)

85

The evolution of the total porosity with hydration time is illustrated in figure 4.27. These values
follow from the cumulative intrusion pore volume curves in figures 4.23 - 4.26. There is a continuous
reduction in total porosity over time for the reference mixture as well as for the mixtures containing
TC. However, the rate of reduction is higher for the specimen containing TC. The decrease in total
porosity with time can be explained by the progression of the hydration process. With the increase in
hydration time, more water has reacted into hydration products. These hydration products fill up the
pore space inducing a decrease in total porosity. Hence, the total porosity reduces when the
hydration process continues. The lower porosity in the specimen containing TC suggests that TC
produces cementitious compounds that create a denser paste or that the hydration process of PC
and FA is accelerated by TC. The acceleration of PC is also shown in the heat of hydration curve in
section 4.5.2.3.
45
40
35

Porosity (%)

30
25
20
PC(65)FA(35)

15

T(5)PC(60)FA(35)

10

T(10)PC(55)FA(35)
5

T(15)PC(50)FA(35)

0
0

10

20

30
Time (days)

40

50

60

Figure 4.27: Total porosity, including inkbottle porosity, as function of curing age.
The influence of the TC content on the total porosity at the curing periods 3, 14, 28 and 56 days is
displayed in figure 4.28. At 3 days of curing , all blended specimen have an almost similar total
porosity as the control paste. However, beyond 3 days the differences in total porosity between the
reference mixture and the TC blended cement pastes becomes larger. A substantial decrease of the
porosity is observed at the level of 5% TC replacement followed by an increase at 10% TC
replacement. The further increase of TC to 15% shows anew a slight decrease in the total porosity. A
similar trend is noticed at all curing ages. However, the blended cement pastes in figure 4.29 always
have a lower pore volume than the control sample at all tested ages. The lowest value of porosity
corresponds to the blended paste containing 5% TC. The lower pore volume of cement paste blended
with may be explained on the basis of phenomena such as:

The nucleation of hydration products on the particles of the mineral admixtures (FA and TC),
which accelerates the hydration of cement, and therefore, increases the formation of
hydration products and decreases the total porosity.
The pozzolanic reaction between metakaolin and calcium hydroxide, either from the reaction
of TC or from hydration of the cement clinker. This reaction converts the lime into
cementitious compounds that create a denser paste.
86

45
40
35
Porosity (%)

30
25
3 days

20

14 days

15

28 days

10

56 days

5
0
0

6
8
10
Replacement level with TC (%)

12

14

16

Figure 4.28: Total porosity, including inkbottle porosity, as a function of the replacement level with
TC.
A possible explanation for the increase in pore volume at 10% of TC compared to the pore volume at
5% of TC is the larger supply of water available for the reaction with TC. When there is sufficient
water available to induce pozzolanic reaction of TC, MK will form the less dense C4AH13 and C2ASH8
hydration products [36]. These hydration products can transform to a dense hydrogamet, which is
associated with a decrease in the solid volume and hence an increase in pore volume. However,
there is no evidence of formation of the hydrogamet at 20C and at short curing time [36]. To
confirm this suggestions more experiments are needed to determine the chemical reaction process
of TC blended pastes.
The differential curves in figures 4.23 - 4.24 show an initial large peak at 3 days of hydration. The
pore diameter corresponding to this large peak is called 1st critical pore diameter. The pore diameter
coinciding with the smaller peak in these curves is called the 2nd critical pore diameter. The first
critical pore diameter is found at about 1-2 m and the second critical pore diameter is found at
about 0.01 - 0.03 m. For all specimens, the 1st peak becomes smaller and shifts to the left when the
hydration continues. In contrast to the first peak the second peak slowly grows for the specimens
when the hydration time increases. This means that the pore structure is becoming much more fine
with an increase in curing age. The amount of TC also strongly influences the critical pore diameters.
The 1st and 2nd critical pore diameters of TC blended pastes have a lower peak and occur at lower
pore diameters compared to the reference specimen. Therefore, the incorporation of TC in the
cement paste led to a refinement of the pore structure.
Figure 4.29 provides information on the total porosity as a function of time for the blended cement
pastes containing 70% FA. While the total porosity versus the replacement level with TC is plotted in
figure 4.30. The influence of the amount of TC on the total porosity is similar to that of the specimen
containing 35% FA.

87

45
40
35
Porosity (%)

30
25
20
15

T(10)PC(20)FA(70)

10

T(15)PC(15)FA(70)

T(5)PC(25)FA(70)

0
0

10

20

30

40

50

60

Time (days)

Figure 4.29: The total porosity versus curing time for the mixtures containing 70% FA.
45
40
35
Porosity (%)

30
3 days

25

14 days

20

28 days

15

56 days

10
5
0
0

10

12

14

16

Replacement level with TC (%)

Figure 4.30: Total porosity, including inkbottle porosity, as a function of the replacement level with TC
for the samples with 70% FA.
The total porosity at 56 days expressed against the replacement level of PC with FA is displayed in
figure 4.31. The inconsistent increase and decrease in total porosity at the various replacement levels
suggests that FA and TC do not behave independently. It is interesting to note that, except for 10%
TC, the increase in FA content reduces the total porosity of the blend. While it was expected that a
higher FA content would increase the porosity, because the assumed CH content is much lower in
the mixtures with 70% FA. The lower porosity of the mixtures might be explained by the microaggregate effect of unreacted FA and the acceleration of the interaction between FA an TC.

88

35
30

Porosity (%)

25
20
15
5% TC (56 days)
10

10% TC (56 days)

15% TC (56 days)

5
0
0

10

20

30
40
50
Replacement level with FA (%)

60

70

80

Figure 4.31: Total porosity at 56 days as function of the replacement level with FA.
The unreacted FA contributes to the properties of the hardened cement blend with its microaggregate effect. This property is reflected in three aspects [53]:
(1) The self-strength of the glass microbeads in FA particles is very high, over 700 Mpa.
(2) FA particles have excellent interfacial property. The weakest place in hardened cement paste
is the interfacial zone between micro aggregate and cement stone. The results [53] have
shown that the hardness of the interfacial zone between a FA particle and cement stone is
higher than that of cement gel.
(3) FA has good dispersibility. It improves the uniformity of the cement paste and fills and fines
the capillary pores as well. Besides the physical effect of FA, the difference in porosity is also
considered to be due to the acceleration of the FA hydration, the available water for the
reaction with TC and the acceleration of the combined effect of TC and FA. With increase in
FA, more water becomes available for the physical adsorption and chemical reaction of TC.
More TC will be able to react which leads to an increasing amount of hydration products and
might accelerate the hydration of FA. As mentioned before the compounds formed by the
pozzolanic reaction of TC have a more dense structure.
4.6.2.2 Effective porosity
The effective porosity is the interconnected pore volume that contributes to the fluid or gas flow, i.e.
permeability. The permeability or in this case the effective porosity is the most important transport
property that is associated with the durability behavior. The results of the effective porosity are
obtained from the second intrusion curves, as illustrated in figures 4.23 - 4.26. The effect of TC on the
effective porosity is similar to the effect of TC on the total porosity. Therefore, these results will not
be discussed.

89

4.6.3 Conclusions
The effect of Top-Crete (TC) on the pore structure of the blended cement pastes including fly ash is
studied with mercury intrusion porosity (MIP). The following conclusions can be drawn from the
results and discussion:
I.

II.

Compared to the reference paste, blended pastes including TC resulted in a denser

microstructure with a lower total porosity and finer pore size distribution. The lowest value
of porosity corresponds to the specimen with 5% TC and 35% and 70% FA.
The increase in the FA content to 70%, except for the mixtures with 10% TC, resulted in a
lower total porosity. This might be attributed to the micro-aggregate behavior of FA, the
increase in available water for the pozzolanic reaction of TC and the acceleration of the
interaction between FA an TC.

90

4.7 Thermal gravimetric analysis (TGA)

From the results of the workability test and setting time test it follows that larger amounts of water
are needed for the chemical and physical binding of TC for increasing TC contents. The chemically
bound water is necessary for the reaction with/of free lime as well as for the pozzolanic reaction of
metakaolin. The large specific surface area of TC particles requires a large amount of physical water.
So far no work has been reported on the amount of chemically bound water required for the
chemical reaction of blended pastes made with addition of TC. Hence, it is unknown whether the
large amount of water is only needed for the physical binding or also for the chemical binding of TC.
The replacement of cement with large amounts of TC and FA has a great influence on the CH content
in the blended paste. The free lime in TC has the potential to produce CH, while the pozzolan MK in
TC has the potential to react with CH to form hydration products. Moreover, the pozzolan FA also
needs CH to chemically react and form hydration products. Therefore it is of great significance to
investigate the chemically bound water and the CH content at different curing periods from the
thermal decomposition. The thermal decomposition of the specimens was investigated using the
thermo gravimetric analysis (TGA).
4.7.1 Results and discussion
4.7.1.1 Ignition loss of PC,TC and FA
The loss on Ignition (LOI) is the amount of moisture or impurities lost when the sample is ignited
under specified conditions. The moisture or impurities may be caused by improper and prolonged
storage or during transport. The sample is ignited until a constant weight loss is obtained. The loss on
ignition of the various minerals (FA, PC and TC) is presented in table 4.7 and is determined using the
curves from TGA by taking the loss in weight between 105C and 1000 C. The percentage of
cumulative weight loss curves as determined by TGA for the raw materials PC, TC and FA are
illustrated in figure 4.32.
Table 4.7: Determination of ignition loss of Portland cement, Top-Crete and Fly ash.

Weight loss between 105C and 1000C (%)

PC
1,99188

TC
21,09348

FA
1,81718

Fly ash and Portland cement show a small drop in weight loss of about 2%. The ignition of TC is
accompanied by a weight loss of almost 21%. The main chemical compound of TC by 41 % is calcium
carbonate. Calcium carbonate undergoes thermal decomposition when it is heated intensively to
approximately 800C. By this decomposition process the CaCO3 is converted into lime and carbon
dioxide is released into the atmosphere. Therefore, this initial weight loss can arise from both the
desorption of water from the surface of TC particles and from the release of carbon dioxide by the
thermal decomposition of CaCO3.

91

(A)

(B)

(C)
Figure 4.32: TGA results for the determination of ignition loss of the raw materials Portland cement
(A) , Fly Ash (B) and Top-Crete (C).
92

4.7.1.2 Chemically bound water

The variation of the results obtained from the TGA technique for parallel samples of one blend is very
small. Therefore, the implementation of one measurement is sufficiently accurate. The chemically
bound water as a function of time for the mixtures with 35% is presented in figure 4.33. The
chemically bound water content is expressed as a percentage of the total amount of water used in
the mixtures. Figure 4.34 shows the chemically bound water as a function of the cement
replacement with TC for the specimen with 35% FA.
14

Chemically bound water (%)

12
10
8

PC(65)FA(35)
T(5)PC(60)FA(35)

T(10)PC(55)FA(35)
4

T(15)PC(50)FA(35)

2
0
0

10

20

30
Time (days)

40

50

60

Figure 4.33: Change in chemically bound water as a function of time for samples with 35% FA
14

Chemically bound (%)

12
10
8
14 days
6

28 days
56 days

4
2
0
0

6
8
10
Cement replacement with TC (%)

12

14

16

Figure 4.34: Change in chemically bound water as afunction of cement replacement for samples with
35% FA
With increase in curing time the increase in chemically bound water diminishes for the blends with
35% FA. The decline in rate of the chemical bound water with time can be explained by the
progression of the hydration process, which slows down as the concrete is sufficiently hydrated. In
addition, the chemical binding of water by the different constituents at different curing hydration
93

periods for the reference specimen is similar to that of the blended cement paste containing 5% TC.
The distinction in chemical bound water between the reference paste and the other blends
containing 10% and 15% TC is also very small. These findings suggest that TC needs the same amount
of water as PC to chemically react. Therefore, the replacement of PC by different levels of TC does
not significantly affect the measured amount of water used for the chemical reactions of the
components in the blends. This suggestion is confirmed by figure 4.35, where the replacement of TC
in the blends containing 70% FA does not have a significant effect on the chemical binding of water.
9
Chemically bound water (%)

8
7
6
5
4
T(5)PC(25)FA(70)

T(10)PC(20)FA(70)

T(15)PC(15)FA(70)

1
0
0

10

20

30
Time (days)

40

50

60

Figure 4.35: The chemically bound water as function of time for samples with 70% FA
However, the increase in FA content has a significant influence on the chemically bound water in the
blends. The samples with 70% FA show a decrease of 33 % in the amount of the chemically bound
water compared to the samples containing 35% FA. This is mainly attributed to the pozzolanic
property of FA that allows the formation of cementitious products without needing water for the
chemical reaction . The replacement of PC with FA also slows down the hydration rate of the paste
and reduces the CH content produced by PC. CH production requires a large amount of water. The
reduction of the PC amount will therefore automatically reduce the chemically bound water.
Furthermore, due to the reduction of PC more water will become available for the reaction of TC.
However, the CH content is assumed to be too small in the blends with 70% of FA to enable the
pozzolanic reaction of all TC.
4.7.1.3 Calcium hydroxide content
Figure 4.36 displays the calcium hydroxide content plotted against time for all blended cement
pastes with 35% and 70% FA. In order to compare the results of the different specimens, the CH
content is expressed as a percentage of the weight of Portland cement. Stating the results per unit
weight of cement instead of unit weight ignited cement gives a better estimate of the start of
calcium hydroxide depletion caused by the pozzolanic reaction [71]. The specimens containing
different levels of TC and FA reaches similar percentages in the beginning of hydration. However the
samples do not produce a similar CH content per gram sample, but the same CH amount per gram
cement.

94

0,16

0,2
0,18

Ca(OH)2 (g CH/g cement)

Ca(OH)2 (g CH/g cement)

0,14
0,12
0,1
0,08

PC(65)FA(35)

0,06

T(5)PC(60)FA(35)

0,04

T(10)PC(55)FA(35)

0,02

T(15)PC(50)FA(35)

T(5)PC(25)FA(70)

0,16

T(10)PC(20)FA(70)

0,14

T(15)PC(15)FA(70)

0,12
0,1
0,08
0,06
0,04
0,02

0
0

20

Time (days)

40

60

20

Time (days)

40

60

Figure 4.36: Change in CH content as a function of time for samples with 35% FA and 70% FA
It is observed that the CH content of the reference sample increases in time, while the CH content of
the TC blended cement pastes with both 35% and 70% FA shows a decrease in time. It is suggested
that PC in the reference mixture keeps on producing CH as the hydration continues without the
consumption of this CH by the pozzolanic reaction of FA in the beginning. The similar assumption
about FA is made for the TC blended pastes. However, the pozzolanic reaction of TC with increase in
time consumes a large amount of CH in the blends containing various levels of TC. Furthermore, the
total CH content is much lower when the cement is replaced with 70% FA instead of 35% FA. This is
attributed to the dilution effect of PC. The CH liberated by the cement hydration appears to be
consumed more rapidly in these blends than in the blends with 35% FA. However, the CH content is
plotted per gram cement and therefore seems to be consumed more rapidly than in the blends with
35% FA. The consumption of the CH content is almost similar in the blends with 35% and 70% FA.
0,16

Ca(OH)2 (g CH/g cement)

0,14
0,12
0,1
0,08
3 days
0,06

14 days

0,04

28 days
56 days

0,02
0
0

10

12

14

16

Cement replacement (%)

Figure 4.37: Change in CH content as function of cement replacement by TC for samples with 35% FA

95

The CH amount expressed against the cement replacement at the curing ages 3, 14, 28 and 56 days is
plotted in figure 4.37. The various levels of TC cause the CH content to increase with an average of
50% per gram cement at the curing period of three days. This percentage reduces as the hydration
process continues and is even smaller than the reference specimen after 28 days. It is noticed that
the use of different levels of TC has little influence on the amount of CH in the blends. This might be
attributed to the lower effective w/c ratio. The increase in TC content slightly influences the rate at
which CH is consumed in the blends.
4.7.2 Conclusion
Based on the experimental results and discussion, the following conclusions can be drawn:
i.

ii.
iii.
iv.

v.

Top-Crete shows an ignition loss of approximately 21%, which can arise from both the
desorption of water from the surface of TC particles and from the release of carbon dioxide
by the thermal decomposition of CaCO3.
The replacement of cement with different levels of TC does not significantly affect the
measured amount of chemically bound water.
The increase in the amount of FA to 70% decreases the amount of chemically bound water to
with 33%. This is mainly attributed to the low pozzolanic activity of FA.
The addition of TC in the cement paste increases the amount of CH with 50% per gram of
cement during three days. The various levels of TC have a similar effect on the CH content
per gram cement, which is probably due to the lower effective w/c ratio.
The total CH content is much lower when the cement is replaced with 70% FA instead of
35% FA due to dilution effect of PC.

It can be concluded that TC needs a large amount for the physical binding, because the different
levels of TC do not affect the measured amount of chemically bound water.

96

Chapter 5. Experimental results of the amount of free lime

5.1 Introduction
One important characteristic of Top-Crete is its large amount of free lime. Knowledge of this amount
in concrete is essential because of its influence on the properties of the material. Free lime has the
tendency to react with surrounding material, like water or CO2. The reaction of lime with CO2 or
water leads to the formation of carbonate and CH. Both hydrates have a larger volume than the
initial hydration products and therefore lead to an excessive volume change. This swelling is followed
by cracking that results in unusable concrete. NEN, the Dutch normalization institute, has set the
requirement that the amount of free lime in cement should be less than 3%. Therefore the second
objective of this study was to investigate the amount of free lime before and after hydration to
ensure the safeness of the structure.

5.2 Results and discussion

The glycol extraction experiment is performed three times for each mixture. The final values
presented in this chapter are an average result determined from the three tests.
5.3.1 The percentage of free lime in de cementitious materials
The results of the free lime content obtained from the glycol extraction method and from references
[3], [4] and [5] for the different materials used in the mixtures are summarized in table 4.4. The
results obtained from references were carried out with analytical technique methods (XRF and XRD)
by other technicians. This comparison gives insight in the accuracy of the method.
Table 4.4: Determination of free lime in the powders
Sample no.

Free lime (%)

from
extraction
method

Free lime (%)

from
references

Top-Crete

14

23

Fly-Ash

1,35

0,45

Portland cement

0,4

0,6

The free lime contents attained by the two methods are not in agreement with each other. Top-Crete
has a percentage of 14% free lime which is much lower than the 23% acquired from the references.
Meanwhile, fly ash contains a higher and Portland cement a lower free lime content compared to
the results from the references. This test method also indicates that FA has a larger free lime content
than PC, which is in contrast with the reference results. The data demonstrates that the glycerol
extraction method does not compare well with the analytical methods. Although the results are not
in agreement with each other, both methods show that TC has a much larger free lime content
compared to PC and FA.

97

5.3 Conclusion
The glycerol extraction test offers a very rapid and simple method to determine the free lime in
cementitious materials. However, the data obtained from this test demonstrates that the glycerol
extraction method is not an accurate and absolute measure of the free lime content. The use of this
method results in significant lower quantities of free lime due to the limitations of this test as
described in paragraph 3.3.3. This method is therefore not suitable for materials which are stored
for a long time, leading to the accumulation of smaller particles into a larger one, or which will
produce a lot of CH that may dissolve in glycerol [40]. Methods other than glycerol extraction should
be employed to measure the free lime content in various blends.

98

Chapter 6. Chemical and autogenous shrinkage of fly ash cement paste

blended with Top-Crete
6.1 Introduction
The last objective of this study was to investigate the influence of Top-Crete and the increase in FA
content on the hardening shrinkage of the blended cement pastes. Early volume changes, within 24
hours, have a negative effect on the tensile stress of the hardening blended cement paste and on the
crack formation in the hardened paste in particular when the cement paste is restrained. At early age
two phenomena are directly responsible for the volume change of the material. The first is the
volume change caused by chemical reactions of the individual constituents in the blended paste in
the first hours and that is called chemical shrinkage. The second volume change is known as
autogenous shrinkage which is a result of the change in the capillary tension in the blend. These
shrinkage types are of great concern when aiming for retaining the durability of structures.
To improve the understanding of the chemical and autogenous shrinkage of the blended cement
paste, it is necessary to study the parameters that determine these types of shrinkage. The main
objective of this chapter is therefore to investigate the development of the chemical shrinkage by
means of the Geiker testing method [92]. Also the development of the autogenous shrinkage is
investigated according to the ASTM standard [94] by using corrugated tubes.
One of the unary pastes investigated in this study was already described in section 4.5.1. The paste
contained 100% TC with a w/b ratio of 1.5.

99

6.2 Hardening Shrinkage

6.2.1 Results and discussion
6.2.1.1 Chemical shrinkage of TC
The chemical and autogenous shrinkage test were performed on three parallel samples. The values
and curves presented in this paragraph are an averaged result from the three tests. Figure 6.1
represents the development of chemical shrinkage of TC as function of curing time up to 7 days. The
curves are initialized 20 minutes after the first watercement contact.
0,06

Chemical shrinkage (ml/g)

0,05
0,04
0,03
Top-Crete
0,02
0,01
0
0,00

50,00

100,00

150,00

200,00

250,00

Time (hrs)

Figure 6.1: Chemical shrinkage of the samples containing only TC with a w/b ratio of 1.5
The Top-Crete specimen experiences an expansion with a high rate in the first 24 hours, as the
hydration continues to 100 hours the TC paste still expands but with a lower rate. After 100 hours
the paste starts to shrink. It is suggested that the reaction of MK with CaO and water produces
hydrates that occupy a greater volume than the volume of the initial products. The large expansion in
the beginning of hydration was especially attributed to the reaction of CaO. This reaction provides CH
for the pozzolanic reaction of MK in TC. Previous studies on MK have also confirmed that an increase
in volume of the paste is attributed to the hydration products formed as a result of the pozzolanic
reaction of MK [28], [88]. The pozzolanic reaction of MK leads to an increased formation of the
C2ASH8 content, which would result in larger expansion of the paste.
6.2.1.2 Chemical shrinkage of the blended pastes
The chemical shrinkage as function of the curing time for the blended pastes with 35% FA is
illustrated in figure 6.2. This figure clearly demonstrates that as the replacement level of PC with TC
rises the total chemical shrinkage reduces. It is also observed that for the specimen with 10% TC,
chemical shrinkage is minimal in the first 24 hours. Moreover, the specimen with 15% TC even shows
an expansion of the blend during the first day. After 24 hours all specimens shrink with increase in
curing time. The decrease in chemical shrinkage is a direct consequence of the combined effect of
the diluted cement hydration and the increased pozzolanic reaction of TC. The dilution effect of PC
100

leads to a reduction in the total hydration products formed by PC and thus to lower chemical
shrinkage. It is assumed that the hydrates formed by the pozzolanic reaction of TC have different
volumes than the compounds formed by the hydration of PC. Suggested is that the hydrates of PC
and FA occupy a smaller volume than the initial products. While the hydrates formed by the reaction
of TC occupy a larger volume than the initial components. Therefore, the expansion of the hydration
products formed by TC compensate for the chemical shrinkage due to the hydration of cement. High
TC percentages can even lead to an expansion of the blend in the first hours, because the total
expansion of the hydrates is larger than the total shrinkage. It is suggested that FA does not
influences the chemical shrinkage of the blended cement paste due to its slow reaction.
0,015

T(5)PC(60)FA(35)

Chemical shrinkage (ml/g of binder)

0,01

PC(65)FA(35)
T(10)PC(55)FA(35)

0,005

T(15)PC(50)FA(35)

0
0,00

50,00

100,00

150,00

200,00

250,00

-0,005
-0,01
-0,015
-0,02
-0,025
-0,03

Time (hrs)

Figure 6.2: Chemical shrinkage of cement pastes blended with TC and 35% FA
The chemical shrinkage of the blends with 70% FA is plotted in figure 6.3. This figure shows that the
chemical shrinkage decreases with increase in FA amount. However, the expansion of the specimens
containing 10% and 15% TC has become bigger in the first 75 hours compared to the expansion of
the mixtures containing 35% FA. This is attributed to the slow reaction of FA and the increase in
available water for the reaction of TC. The slow reaction of FA causes a lower formation of hydration
products with a smaller volume than the initial component. Moreover, the increase in available
water allows more TC to react producing more hydrates with a larger volume than TC.

101

0,02
Chemical shrinkage (ml/g of binder)

T(5)PC(25)FA(70)
0,015

T(10)PC(20)FA(70)
T(15)PC(15)FA(70)

0,01
0,005
0
0

50

100

150

200

250

-0,005
-0,01
-0,015

Time (hrs)

Figure 6.3: Chemical shrinkage of cement pastes blended with TC and 70% FA
6.3.3.2 Autogenous shrinkage
Figure 6.4 illustrates the effect of the water to binder ratio on the autogenous shrinkage of the
blended paste as function of time, initialized after final setting time. As mentioned in the
introduction, the autogenous shrinkage is influenced by the w/b ratio due to the consumption of
water from the refined pores by the unhydrated particles. This would result at lower w/b ratio in the
increase of the capillary tension of the pore water, hence increasing the autogenous shrinkage of the
paste. It is also suggested by [90] that the large ionic concentration difference between the pore
water and the adsorbed water at the hydrating front of the particles is the main cause of autogenous
shrinkage. As a consequence of this large difference in concentration, water will be displaced by the
suction effect. The movement of water initiates an increase in capillary tension, which therefore
results in a large autogenous shrinkage.
1,00

Measurement from the final setting

Autogenous shrinkage (mm/m)

0,00
-1,00
-2,00
-3,00

100

200

300

400

500

600

700

Time (hrs)
PC(65)FA(35)
(w/b=0,4)
PC(65)FA(35)
(w/b=0,5)

-4,00
-5,00
-6,00

Figure 6.4: Autogenous shrinkage of cement pastes blended with 35% FA and with w/b ratios of 0.4
and 0.5 as function of time.
102

Figure 6.5 illustrates the effect of 10% TC on the autogenous shrinkage as a function of time, starting
from the final setting time. From this figure is observed that in the first hours the specimens expand
and then rapidly increase in autogenous shrinkage. After 200 hours, the increase in autogenous
shrinkage is much slower. The rate of increase in autogenous shrinkage depends on the composition
of the blends. The replacement of PC with 10% TC and 35% FA shows a larger increase of the
autogenous shrinkage at 600 hours compared to the control paste. However, contradictory results
are found by some authors concerning autogenous shrinkage in MK blended cement pastes. These
results showed that the autogenous shrinkage decreased with increase in MK replacement level. The
specimen with 70% FA and 10% TC shows a slight decrease in autogenous shrinkage compared to the
specimen T(10)PC(55)FA(35).
The main phenomena for an increase in autogenous shrinkage in the blend with 35% FA compared to
the reference paste are the acceleration of the PC hydration, the highly pozzolanic reaction of TC and
the refinement of porosity by TC [97]. The first two contribute to the removal of water from the
specimen which results in self-desiccation of the specimen. While the refinement of the porosity
results in the increase of the capillary tension.
1,00

Measurement from the final setting

Autogenous shrinkage (mm/m)

0,00
0
-1,00

100

200

300

400

500

600

700

Time (hrs)
T(10)PC(55)FA(35)

-2,00
Control paste (w/b=0,5)

-3,00

T(10)PC(20)FA(70)

-4,00
-5,00
-6,00

Figure 6.5: Chemical shrinkage of cement pastes blended with TC and 70% FA
The decrease in autogenous shrinkage of the specimen with 70% FA in comparison with that of the
specimen with 35% FA is attributed to the slow pozzolanic reaction of FA and the reduced pozzolanic
reaction of TC with CH, which is lower in the specimens with 70% FA (paragraph 4.7.3). The slow
pozzolanic activity of FA requires a lower water demand. The increase of the available water in the
blended cement pastes with 70% FA leads to a decrease in autogenous shrinkage. On the other hand,
there is a larger amount of water available for the reaction with TC when PC is replaced by FA. The
physically or chemically binding of TC with water could increase the autogenous shrinkage. However,
the available CH content and the pozzolanic reaction with CH become more dominant as the
replacement of PC increases. Therefore, the autogenous shrinkage decreases when PC is replaced

103

with a larger amount of FA. More research is needed to fully understand the effect of TC on the
autogenous shrinkage of blended cement pastes.
6.3.4 Conclusion
The following conclusions can be drawn from the effect of Top-Crete on the chemical and
autogenous shrinkage of the blended cement paste:

The results indicate that partial replacement of cement with TC decreases the magnitude of
the chemical shrinkage, particularly at higher replacement level. The decrease in chemical
shrinkage is a direct consequence of the combined effect of the diluted cement hydration
and the increased pozzolanic reaction of TC. The TC hydrates compensate for the chemical
shrinkage.
The replacement of PC with 70% of FA results in a lower chemical shrinkage compared to the
specimen with 35% FA, because FA has almost no influence on the chemical shrinkage due
to its low pozzolanic activity. However, the expansion of these blends is significantly
increased in the first hours due to the pozzolanic reaction of TC.
The replacement of PC with 10% TC and 35% FA shows a substantial increase in autogenous
shrinkage compared to the control paste. This might attributed to the acceleration of the PC
hydration, the highly pozzolanic reaction of TC and the refinement of porosity by TC [97]. The
first two contribute to the removal of water from the specimen which results in selfdesiccation of the specimen. While the refinement of the porosity results in the increase of
the capillary tension.
The specimen with 70% FA and 10% TC shows a slight decrease in autogenous shrinkage
compared to the specimen T(10)PC(55)FA(35). Due to the slow pozzolanic reaction of FA and
the reduced pozzolanic reaction of TC with CH.

104

7. Evaluation
In chapter 4 and 6 the influence of TC and the influence of the amount of FA on the different
properties of the blended cement pastes was discussed separately . The aim of this chapter is
therefore to relate the properties of TC blended cement pastes presented in previous chapters. Since
these properties are related to each other from one way to another.

Physically bound water and Chemically bound water

The influence of TC, PC and FA on the uptake of water by physical absorption was partly determined
by the workability test and partly by the setting time test. Meanwhile, the chemically bound water
was measured by TGA. From these tests it was found that TC consumes a large amount of water. This
high water consumption is mainly attributed to the physical adsorption of the water due to its large
specific surface area and high pozzolanic activity of TC. The different levels of TC did not affect the
chemically bound water in the blended cement pastes compared with the reference mixture.
Although, the chemically bound water was similar in all mixtures, the setting time decreased with the
increase in the amount of TC. This means that TC leads to higher rate of the chemically binding of
water.

Porosity and Strength

According to literature, the compressive strength often may be associated with other properties of
concrete. Especially, the relationship between uniaxial strength and porosity in cement pastes is
sometimes defined linearly. Because the strength and pore structure development are both related
to the amount of hydration products and the spatial distribution of these products. Therefore, it is
useful to show the relationship between porosity and compressive strength of blended cement
pastes made with addition of TC. The correlation between the porosity and compressive strength for
the reference specimen and the T(5)PC(60)FA(35) specimen at 56 days is illustrated in figure 7.1.

Figure 7.1: The correlation between porosity and compressive strength for the reference specimen
and the T(5)PC(60)FA(35) specimen at 56 days
There is a clearly reversed proportional relationship between the compressive strength and the
porosity. With the increase in porosity, the compressive strength decreases for the blended cement
105

mortars. This is contrast to the literature where a porous microstructure is logically associated with a
lower compressive strength.
In the discussion of the strength results, the strength loss was also assigned to the coarsening of the
pore structure in the hardened cement paste. This factor does not hold for TC blended cement
pastes, because the porosity is increasing.

Autogenous shrinkage and Physical adsorbed water

The autogenous shrinkage of the cement paste is strongly influenced by the moisture absorbed by
the aggregates and mineral admixtures. It was found that the autogenous shrinkage decreases with
increasing moisture content of the aggregates and mineral admixtures. It is suggested that the water
lost by self-desiccation of cement paste is replaced by the physical adsorbed water from the
aggregates or mineral admixtures [98]. Therefore, the autogenous shrinkage is directly related to
physical adsorption capacity of aggregates and mineral admixtures. From the nitrogen adsorption
test followed that TC has a large specific surface area. Hence, TC particles will physically adsorb a
large amount of water. The literature suggested that the addition of mineral admixtures, which
adsorb a large amount of physically water, would mitigate the autogenous shrinkage of the mixtures.
Obviously TC does not release its large physical adsorbed water for the hydration of the particles,
because the autogenous shrinkage increased of the blended cement pastes with TC. Therefore, it
does not act as internal curing agent in the blended cement past and the autogenous shrinkage
increases.

Porosity and Autogenous shrinkage

The autogenous shrinkage is also influenced by the pore structure of the cement paste. It has been
found that the total autogenous shrinkage slowly increase due to the refinement of the capillary
pores, which forces the formation of more curved menisci, generating an overall compression of the
skeleton [100]. The correlation between the porosity and autogenous shrinkage for the reference
mixture and the blended cement paste with 10% TC and 35% FA at 7 days is illustrated in figure 7.2.

Figure 7.2: The correlation between porosity and autogenous shrinkage at 7 days for the reference
specimen and the blended cement paste with 10% TC and 35% FA at 7 days
106

The figure shows that the autogenous shrinkage increases with the increase in the refinement of the
pores. This behavior is similar to that of the literature. It is well know from the results that the
pozzolanic reaction of TC contributes to the refinement of the capillary pore structure. Thus, the use
of TC leads to an increase of the autogenous shrinkage.

107

108

8. Conclusions and recommendations

8.1 Conclusions
Within the framework of this master thesis the use of thermally activated paper sludge minerals
(Top-Crete) as a partial cement replacement was examined. The effects of a replacement on the
properties of ternary cement pastes, such as the development of the strength and hydration heat,
were determined and analyzed. Another objective of this research was to define the free lime
content in the blended pastes before and after hydration. Furthermore, the contribution of TopCrete to the chemical and autogenous shrinkage was studied. The following conclusions are derived:
1. Top-Crete (TC) particles are composed of tiny particles that are agglomerated into rigid
structures of different shapes and sizes. The particle size distribution of TC is slightly finer
than that of Portland cement (PC) and fly ash (FA). TC has a specific surface area of 7.2262
m2/g, which is about 27 times higher than PC. The ignition loss of TC is 21%, which mainly
arises from the release of carbon dioxide by the thermal decomposition of the large CaCO3
content.
2. The use of TC results in a large water requirement for its physical and chemical binding. The
chemical bound water content of all mixtures are within the same range of 12%, even though
each mixture has a different amount of TC. It can be concluded that TC chemically binds
almost the similar amount of water as PC. Therefore, the great water demand is mainly
attributed to the physical properties of TC, in particular to its large specific surface area and
high pozzolanic activity compared to that of FA and PC.
3. The addition of TC shows a decrease in workability and setting time, which can possibly be
assigned to its large water requirement. The effects are stronger when more TC is used.
4. Compressive and bending strength tests show a loss in strength of 8% - 20% for the mortars
blended with various amount of TC. The strength loss increases with increase in TC
replacement level. When PC is replaced by TC, the strength of the mortars is only maintained
for the mixtures with 5%TC at 28 days.
5. The results from MIP tests show that the blended pastes incorporating TC create a denser
microstructure with a lower total porosity and finer pore size distribution. The lowest values
of porosity correspond to the specimen with 5% TC. The porosity of the specimen with 5% TC
shows a reduction of almost 35% at 56 days compared to the reference specimen.
6. The addition of 5% or more TC to the cement paste has increased the amount of CH with
50% per gram cement during three days. Increasing the amount of TC from 5% to 15% does
not significantly affect the CH content.
7. The isothermal calorimeter test shows a slight increase in the hydration heat and degree of
hydration of the blended paste containing 35% FA, especially at higher replacement levels.
8. The glycerol extraction test offers a very rapid and simple method to determine the free lime
in cementitious materials. However, the data obtained from this test demonstrates that the
glycerol extraction method is not an accurate and absolute measure of the free lime content.
This method is therefore not suitable for materials which are stored for a long time, leading
to the accumulation of smaller particles into a larger one, or which will produce a lot of CH
that may dissolve in glycerol [40]. Methods other than glycerol extraction should be
employed to measure the free lime content in various blends. With the estimated free lime

109

content of about 23%, the maximum replacement level of Top-Crete when used for
structural concrete is about 6% of the total cement content.
9. The results indicate that partial replacement of cement with TC decreases the magnitude of
the chemical shrinkage with 25% or more, particularly at higher replacement levels. The
decrease in chemical shrinkage is a direct consequence of the combined effect of the diluted
cement hydration and the increased pozzolanic reaction of TC. The TC hydrates compensate
for the chemical shrinkage.
10. The replacement of PC with 10% TC and 35% FA shows a substantial increase in autogenous
shrinkage compared to the control paste. This might attributed to the acceleration of the PC
hydration, the highly pozzolanic reaction of TC and the refinement of porosity by TC [97]. The
first two contribute to the removal of water from the specimen which results in selfdesiccation of the specimen. While the refinement of the porosity results in the increase of
the capillary tension.
Furthermore, the use of thermally activated paper sludge minerals (Top-Crete) as a leverage to use
higher amounts of supplementary cementing materials was also examined. The FA content of the
blended cement pastes was increasing to 70%. It was expected that combining TC with high
percentages of FA would lead to better material properties. The influence of an increase in FA
content on the properties of the blended cement pastes made with addition of TC was determined.
The main conclusions are given below:
11. The loss in workability due to the presence of Top-Crete cannot be compensated by
increasing the amount of fly ash from 35% to 70%. The increase in FA content also does not
result in a significant improvement of the setting time.
12. The increase of the amount of FA leads to lower strength, lower hydration heat, slow degree
of hydration, lesser chemically bound water content and CH content and a slightly lower
autogenous shrinkage. This decrease is suggested to be due to the low pozzolanic activity of
FA and the reduced CH content for the pozzolanic reaction of TC and FA. The chemical
shrinkage is also reduced, but the chemical expansion has significantly increased. From these
findings it can be concluded that TC will not compensate for the large PC deficiency by the
increase of FA from 35% to 70%.
Overall, from this research, it has been found that a certain percentage of TC can act as a
replacement of cement without significant loss in strength and with a considerable shorter setting
time. Replacing cement with 5 % TC in blended pastes is advantageous because of (i) lower setting
time, (ii) the increase in chemical resistance due to finer pore structure, (iii) the lower chemical
shrinkage avoiding micro cracks in the early age of hydration and (iv) the increase in environmental
benefits.

110

8.2 Recommendations
The replacement of cement by activated paper sludge minerals (Top-Crete) is a relatively new
research area. The research work that already has been done on this subject is very rear. This thesis
contributes to the understanding that is needed to explain the effect of TC on the properties of the
blended cement pastes. However, more investigation needs to be done before the replacement of
cement by TC is accepted by the industry. To obtain a more detailed understanding of the effect of
TC on the properties of concrete, the following additional research work is recommended.
In order to study the effect of TC on the properties of blended cement pastes, i.e. strength, hydration
heat and so on, it is necessary to investigate the chemical reactions of Top-Crete with various
amounts of water and calcium hydroxide. The use of different w/b ratios in the blended cement
pastes might cause the production of various amounts of CH by the free lime in TC. Meanwhile, the
CH content has an effect on the hydrates formed by TC and on the rate of pozzolanic reaction. It is
also of great importance to investigate the hydration reactions of binary and ternary blended cement
pastes made with addition of TC. These investigations would result in a better understanding of the
behavior of TC and the hydration products formed by TC in various environments.
The results obtained from the experiments clearly show that the increase in Top-Crete content leads
to a substantial decrease in consistency due to the large specific surface area of TC and the high
pozzolanic activity of TC. Therefore, the workability needs to be improved by the increase in w/b
ratio or by the usage of super plasticizers. Super plasticizers are available in several types with
various properties. The usage of super plasticizers or the increase in w/b ratio in the mixtures will
induce different effects on the properties of the blended cement paste. The increase in water also
has a different influence on the hydration process of TC and PC. Further investigations are needed to
determine the effects of super plasticizers on the binder properties of cement pastes blended with
TC and to determine the influence of the w/b ratio on the hydration process of the cement pastes
blended with TC.
Proper analytical techniques need to be used to accurately determine the free lime amount in
cementitious minerals and in the hardened blended cement paste made with addition of TC. The
existing X-ray diffraction combined with the X-ray fluorescence could be used in future research
work.
Almost all experiments were performed on blended cement pastes, only the strength and workability
tests were carried out on blended mortars. However, the TC blended cement pastes is strongly
influenced by the presence of aggregates. Thus, the effects of aggregates become important. First,
the influence of aggregates on the CH production is not well understood or easily predictable. The
aggregate particles might serve as nucleation sites for CH and cause nuclei to grow into crystals [77].
Secondly, the cement paste-aggregate interface has an effect on the interfacial transition zone (ITZ)
between aggregates and the hardened cement paste [77]. The absorption of water by the
aggregates, the tendency of bleeding due to downwards movement of the aggregates and the lower
packing density of cement grains are likely to cause this effect. Therefore, further investigations are
needed on these effects of aggregates in concrete mixtures.
It is well known that MK improves the density and permeability of the cement matrix. The results
obtained in this research confirmed that the use of TC leads to a decrease in total porosity and a
lower effective porosity (permeability). Therefore, cement pastes with addition of TC might have a
111

better effect on durability properties. Further investigations are needed to study the effect of TC in
blended cement pastes on the chloride penetration depth, carbonation development, alkali silica
reaction, sulfate resistance and other durability properties.
The use of FA in the ternary cement blends may compensate for the heat produced by TC and may
increase the available water for the physical adsorption of TC particles. However, alternative
pozzolanic materials of FA could also be used to investigate the influence of TC on the behavior of
these materials. It would be very interesting to replace FA with blast furnace slag or rice husk ash and
to determine the change in strength, setting, pore structure and hydration heat. The investigation on
the synergetic effect between TC and the different materials is also important.
The application of concrete blended with TC in structural elements.

112

References
[1]
[2]
[3]
[4]
[5]
[6]

[7]
[8]

[9]

[10]

[11]
[12]
[13]
[14]
[15]

[16]

[17]
[18]

[19]
[20]

[21]
[22]

K. Humphreys and M. Mahasenan, Towards a sustainable cement industrysub-study 8:

climate change, accessed on 12 January 2003.
C. Thomark, Conservation of energy and natural resources by recycling building waste
Resources, Conservation & Recycling, 33 (2001), pp. 113130
J. Pera, A. Amrouz, Development of highly reactive metakaolin form paper sludge, Adv. Cem.
Based Mater., 7 (1998), pp. 4956
J. Klootwijk, E.ON Benelux NV, 2011
W. de Vries, The use of micronized sand as cement replacement, Delft University of
technology, 2008
M. Fras, I. Vegas, R. V. de la Villa and R. G. Gimnez, Recycling of Waste Paper Sludge in
Cements: Characterization and Behavior of New Eco-Efficient Matrices, Universidad
Autnoma de Madrid, Spain
S.A. Balwaik and S. P. Raut, Utilization of Waste Paper Pulp by Partial Replacement of Cement
in Concrete, Department of Civil Engineering, YCCE, Nagpur-10, Maharashtra, India
M. Fras, M.I. Sanchez de Rojas, O. Rodriguez, R. G. Gimnez, R. V. de la Villa, Characterisation
of calcined paper sludge as an environmental friendly source of metakaolin for manufacture
of cementitious materials, Eduardo Torroja Institute, Madrid.
I. Vegas, J. Urreta, M. Fras, O. Rodrguez, S. Ferreiro, B. Nebreda, R. Garca and R. Vigil,
Engineering properties of cement mortars containing thermally activated paper sludge,
LABEIN-Tecnalia, Derio, Spain
R. V. de la Villa, M. Fras, M. I. S. de Rojas, I Vegas, R. Garca, Mineralogical and morphological
changes of calcined paper sludge at different temperatures and retention in furnace,
Departamento de Geologa y Geoqumica, Facultad de Ciencias, Universidad Autnoma,
28049 Madrid, Spain, 2006
H. Moosberg-Bustnes, B. Lagerblad and E. Forssberg, The function of fillers in concrete,
Swedish Cement and Concrete Research Institute, Stockholm, Sweden, 2004
T.R. Jones, Metakaolin as a pozzolanic addition to concrete.
G. Varga, The structure of kaolinite and metakaolinite, Constantine the Philosopher
University, 2007
R. Siddique and J. Klaus, Influence of metakaolin on the properties of mortar and concrete: A
review, Thapar university, india, 2008
B.B Sabir, S Wild and J Bai, Metakaolin and calcined clays as pozzolans for concrete: a review,
School of Technology, University of Glamorgan, Pontypridd, Mid.Glamorgan, Wales CF37 1DL,
UK, 2001
K. Yazan, Effects of retempering with superplasticizer on properties of prolonged mixed
mineral admixture containing concrete at hot weather conditions, School of Natural and
Applied Sciences, 2005
V.M. Malhotra and P.K. Mehta, Pozzolanic and cementitious materials, Taylor and Francis.
A. K. Schindler and K. J. Folliard, Influence of supplementary cementing materials o the heat
of hydration of concrete, Advances in Cement and Concrete IX Conference Copper Mountain
Conference Resort in Colorado, August 10-14, 2003
R. Walker and S. Pava, Physical properties and reactivity of pozzolans, and their
influence on the properties of limepozzolan pastes, Materials and Structures (2010)
Ravindra Gettu, Joana Roncero & Migual A. Martin, New Chemical Admixtures for Concrete:
Effectiveness and Practical Implication, Structural Technology Laboratory, University
Politecnica de Catahinga, Barcelona, Spain.
Panarese WC, Kosmatka SH, Randall FA. Concrete masonry handbook for architects,
engineers, builders. Portland Cement Association, 5th ed. USA; 1991.
EN 12350-5:2009

113

[23]

[24]
[25]
[26]
[27]
[28]
[29]

[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]

[38]

[39]

[40]
[41]

[42]

[43]
[44]

[45]

J. J. Brooks, M. A. Megat Johari and M. Mazloom, Effect of admixtures on the setting times of
high-strength concrete, School of Civil Engineering, University of Leeds, Leeds LS2 9JT, UK,
2009
Y. Chen and I. Odler, On the origin of portland cement setting, Technical University Clausthal
Fed. Rep. Germany, 12 February 2003
http://theconstructor.org/concrete/setting-of-cement-microstructure-of-hcp/957/
NEN-EN 196-3+A; Methods of testing cement - Part 3: Determination of setting times and
soundness
D. Snelson, S. Wild & M. O'Farrell, Setting times of Portland cementmetakaolinfly ash
blends, 13 Apr 2011
Wild, S.; Khatib, J. M.; Roose, L. J. 1998. Chemical shrinkage and autogenous shrinkage of
Portland cement-metakaolin, Advances in Cement Research 10(3): 109119.
R. F. Feldman, G. G. Carette & V. M. Malhotra, Studies on Mechanism of Development of
Physical and Mechanical Properties of High-Volume Fly Ash-Cement Pastes, Institute for
Research in Construction/NRC, Montreal Road, Ottawa, Ontario, Canada K 1A 0R6, Cement &
Concrete Composites 12 (1990) 245-251
NEN EN-196-1 methods of testing cement Part 1: Determination of strength
G.C. Bye, Portland Cement, Thomas Telford edition 1999
Koichi Maekawa,Tetsuya Ishida and Toshiharu Kishi, Multi-scale modeling of structural
concrete, Taylor and francis
H.F.W. Taylor, Cement Chemistry, Taylor and Thomas Telford Services Ltd, 1997
Y. Guang, Experimental Study & Numerical Simulation Of The Development Of The
Microstructure and Permeability Of Cementitious Materials, DUP Science, 2003
K. van Breugel, Betonkunde,
J.M. Khatib and S. Wild, Pore size distribution of metakaolin paste, Centre for Research in
Built Environment, University of Glamorgan, Pontypridd, UK, 1996
Moiss Frias and Joseph Cabrera, Pore size distribution and degree of hydration of
metakaolincement pastes, Civil Engineering Materials Unit (CEMU), School of Civil
Engineering, University of Leeds, Leeds LS2 9JT, UK, 2000
C.-S. Poona, L. Lama, S.C. Kou, Y.-L. Wong, Ron Wong, Rate of pozzolanic reaction of
metakaolin in high-performance cement pastes, Department of Civil and Structural
Engineering, The Hong Kong Polytechnic University, Hung Hon, Kowloon, Hong Kong, China,
2001
Jean Ambroise, Sandrine Maximilien, and Jean Pera, Properties of Metakaolin Blended
Cements, laboratoire des Mat&iaux Min&aux, Institut National des Sciences Appliqu~es de
Lyon, Villeurbanne, France, 1994
Gebhardt, R.F. Rapid Methods for Chemical Analysis of Hydraulic Cement; American Society
for Testing and material: Philadelphia, PA, 1988
William Lerch and R.H. Bogue, Revised procedure for the determination of uncombined lime
in Portland Cement, Portland Cement Association Fellowship, U.S. Bureau of Standards,
Washington, D.C.
Oi-Wah Lau, Shiu-Fai Luk, Nina L.N. Cheng and Hoi-Yin Woo, Determination of free lime in
Clinker and Cement by lodometry, Department of chemistry, The Chinese University of Hong
Kong, Shatin, N.T., Hong Kong.
FRANK M. BIFFEN, Determination of Free Lime and Carbonate in Calcium Silicate Hydrates by
Thermobalance, Johns-Manville Research Center, Manville, N. J.
P. Arjunan and A. Kumar, Rapid Techniques for determination of free lime and free magnesia
in cement clinker and portlandite in hydrates, Central Glass and Ceramic Research Institute
Jadavpur, Calcutta 700032, India
J.A. Fifield and R.G. Blezard, Rapid routine determination of uncombined lime by an
automatic absorptiometric method, Development laboratory, tunnel cement Ltd, West
Thurroch, Essex.
114

[46]
[47]
[48]

[49]
[50]
[51]

[52]

[53]

[54]
[55]
[56]

[69]
[70]
[71]

[75]

[76]
[77]

[78]

[79]
[80]

[81]

D.Lonquet, L.Burglen and G.Bellina, The use of glycol in establishing the percentage content
of Free Lime in cement clinker, 1977
NEN 3103. Richtlijnen voor het stellen van de titer van titervloeistoffen
Wild et al., 1996 S. Wild, J.M. Khatib and A. Jones, Relative strength, pozzolanic activity and
cement hydration in superplasticised metakaolin concrete. Cement and Concrete Research,
26 10 (1996), pp. 15371544.
D.D. Vu, Strength properties of metakaolin- blended paste, mortar and concrete, Delft
university press, 2002
P. K. Mehta, INFLUENCE OF FLY ASH CHARACTERISTICS ON THE STRENGTH OF PORTLAND-FLY
ASH MIXTURES, University of California, Berkeley, CA 94720, 1985
R.A. Olsen, C.M. Neubauer and H.M. Jennings, Damage to the pore structure of hardened
Portland cement paste by mercury intrusion, Journal of American Ceramic Society, vol 80 (9)
pp. 2454-2458, 1997
Prinya Chindaprasirt, Chai Jaturapitakkul and Theerawat Sinsiri, Effect of fly ash fineness on
compressive strength and pore size of blended cement paste, Department of Civil
Engineering, Khon Kaen University, Khon Kaen 40002, Thailand, 2004
Aiqin Wang, Chengzhi Zhang and Wei Sun, Fly ash effects: III. The microaggregate effect of fly
ash, Department of Materials Science and Engineering, Southeast University, Nanjing
210096, People's Republic of China, 2004
Talyor H.F.W., Cement Chemistry, Acadamic Press, London, 296-3080990
Ken W. Day, Concrete mix design, quality control and specification, Taylor & Francis, 2006 357 pagina's
Mukesh Pokharel and Mamadou Fall, Coupled thermochemical effects on the strength
development of slag-paste backfill materials, Dept. of Civil engineering. Univ of Ottawa, 2010.
P. Van den Heede, Porositeit en transporteigenschappen van groene betonsoorten,
universiteit gent, 2008
R. F. FELDMAN and P. J. SEREDA, A model for hydrated Portland cement paste as deduced
from sorption-length change and mechanical properties,
Bryan K. Marsh 1 and Robert L. Day, POZZOLANIC AND CEMENTITIOUS REACTIONS OF FLY
ASH IN BLENDED CEMENT PASTES, Dept. of Civil Engg., University of Calgary, Alberta, Canada,
1987
Hearn, N., Hooton, R.D. and Mills, R. H., Pore Structure and Permeability, Chapter 25 in
Significance of Tests and Properties of Concrete and Concrete Making Materials, ASTM
STP169C, 1994 (pp. 240-262).
Cook, R.A., Hover, K.C. (1999) Mercury porosimetry of hardened cement pastes, Cement
and Concrete Research, vol.29 (6) pp. 933-943.
S. Wild and J.M. Khatib, PORTLANDITE CONSUMPTION IN METAKAOLIN CEMENT PASTES AND
MORTARS, Centre for Research in the Built Environment, University of Glamorgan,
Pontypridd, UK , 1996
B.B Sabir , S Wild and J Bai, Metakaolin and calcined clays as pozzolans for concrete: a review,
School of Technology, University of Glamorgan, Pontypridd, Mid. Glamorgan, Wales CF37
1DL, UK, 2001
CDEM Minplus, www.CDEM.nl
Fras, M., Snchez de Rojas, M.I., Cabrera, J., 2000. The effect that the pozzolanic reaction of
metakaolin has on the heat evolution in metakaolin-cement mortars. Cement and Concrete
Research 30, 209216.
Nasir Shafiq C and Muhd Fadhil Nuruddin, Degree of Hydration of OPC and OPC/FA Pastes
Dried in Different Relative Humidity, University Technology Petronas, 31750, Tronoh, Perak,
MALAYSIA, 2010
115

[82]
[83]

[84]

[85]

[86]

[87]
[88]
[89]
[90]

[91]
[92]

[93]
[94]
[95]

[96]
[97]
[98]

[99]

J. Bensted, Seventh International Congress on the Chemistry of Cement, vol. IIParis (1980) p.
II/16.
N.Y. Mostafaa and P.W. Brownb, Heat of hydration of high reactive pozzolans in blended
cements: Isothermal conduction calorimetry, a Chemistry department, Faculty of Science,
Suez Canal University, Ismailia, Egypt, b Materials Science and Engineering, Materials
Research Laboratory, The Pennsylvania State university, University Park, PA 16802, USA,
2005
J. Bai and S. Wild, Investigation of the temperature change and heat evolution of mortar
incorporating PFA and metakaolin, School of the Built Environment, University of Glamorgan,
Pontypridd CF37 1DL, UK, 2002
David G. Snelsona, Stan Wilda, Martin O'Farrellb, Heat of hydration of Portland Cement
MetakaolinFly ash (PCMKPFA) blends, aFaculty of Advanced Technology, University of
Glamorgan, Llantwit Road, Trefforest, Pontypridd, Mid Glamorgan, Wales, CF37 1DL, United
Kingdom, 2008
R.I. Gilbert, Shrinkage, Cracking and Deflection-the Serviceability of Concrete Structures,
Professor and Head, School of Civil and Environmental Engineering, The University of New
South Wales, Sydney, NSW, 2052
Matthew DAmbrosia1 and Nathaniel Mohler2, Early-Age Cracking, 1Associate, CTLGroup,
2
Concrete Engineer, PCA, 2011
J. M. Kinuthia, S. Wild, B. B. Sabir and J. Bai, Self-compensating autogenous shrinkage
inPortland cement-metakaolin-fly ash pastes, University of Glamorgan, 2000
Erika E. Holt, Early age autogenous shrinkage of concrete, Technical Research centre of
finland, Vtt Publications, ESPOO 2001,
M. Bouasker, P. Mounanga, P. Turcry, A. Loukili and A. Khelidj, Chemical shrinkage of cement
pastes and mortars at very early age: Effect of limestone filler and granular inclusions,
Cement & Concrete Composites (2007)
Powers, T. C., and Brownyard T. L., (1948) "Studies on the Physical Properties of Hardened
Portland Cement Paste" PCA Bulletin 22, Chicago, Illinois
Geiker, M. Studies of Portland Cement Hydration: Measurements of Chemical Shrinkage and
a Systematic Evaluation of Hydration Curves by Means of the Dispersion Model, Ph.D.
Thesis, Technical University of Denmark, Copenhagen, Denmark, 1983.
G. Sant, P. Lura and J. Weiss, Measurement of Volume Change in Cementitious Materials at
Early Ages: Review of Testing Protocols and Interpretation of Results, 2005
E. Tazawa and S. Miyazawa, Influence of cement and admixture on autogenous shrinkage of
cement paste. Cem Conc Res, 25 2 (1995), pp. 281287.
A. Princigallo, P. Lura, K. van Breugel and G. Levita, Early development of properties in a
cement paste: a numerical and experimental study. Cem. Concr. Res., 33 7 (2003), pp. 1013
1020.
P. Lura, O.M. Jensen and K. van Breugel, Autogenous shrinkage in high-performance cement
paste: an evaluation of basic mechanisms. Cem. Concr. Res., 33 2 (2003), pp. 223232.
P. Lura, Autogenous Deformation and Internal Curing of Concrete, DUP Science, 2003
K. Kohno, T. Okamoto, Y. Isikawa, T. Sibata and H. Moris,Effects of artificial lightweight
aggregate on autogenous shrinkage of concrete, Central Research Institute Kiyosumi,
Taiheiyo Cement Corporation, 2-23, Kiyosumi 1-chome, Koto-ku, Tokyo, 135-8410, Japan,
1999
R Gagn, I Aouad and J Shen, Development of a new experimental technique for the study of
the autogenous shrinkage of cement paste, Materials and Structures, 1999 - Springer

116

117

118

Appendix A: Sample Preparation and Experimental Procedure

A.1 Workability
A.1.1 Sample preparation
The sample preparation was carried out in accordance with NEN EN-196-1 methods of testing
cement Part 1: Determination of strength [30]. After weighting the different components drinking
water was added to the powder and mixed at low speed for 60 seconds, whereof norm sand was
added steadily during the last 30 seconds. Afterwards the mixer was switched to high speed for an
additional 30 seconds. Then the mixer was stopped for 90 seconds, in which the first 30 seconds
where used to remove the mortar from the wall and bottom part of the bowl and placed in the
middle. Thereafter the mixing was continued at high speed for 60 seconds. The total mixing time,
including stops, was 4 minutes.
A.1.2 Experimental procedure
The consistency of the mortar is measured with the flow table test, see figure 1. A cone is filled with
mortar in two layers immediately after the mixing procedure has finished. To ensure uniformly
distribution of the specimen each layer is compacted with 20 strokes of the rod. After striking the top
surface of the cone and cleaning the outer space of the cone, the cone is lifted up vertically. The
mortar will spread out and the first diameter is recorded. Thereafter, the mortar is spread out by
shaking the table 12 times. The diameter is measured again, but this time in two directions
perpendicular to each other. The mixture shows a circle cross section when the cone is lifted up and
removed from the flow table. After shaken the table twelve times the circle will change into an
ellipse cross section. Therefore, the diameter has to be measured twice. First the biggest diameter is
measured and then the diameter perpendicular to the first diameter.

Figure 1: Flow table equipment for measuring the workability.

The average of the diameters, in the different directions, is taken as the flow value. This flow value is
an indicator for the workability [22].

119

In which:

d1 = maximum spread of the mortar

d2 =spread perpendicular to d1

A.2 Setting
A.2.1 Sample preparation
The sample preparation was carried out in accordance with NEN EN-196-1 methods of testing
cement Part 1: Determination of strength. The sample preparation is almost the same as described
in the section A.1.2 sample preparation. Instead on mortar, the setting test has been performed on
cement paste. After mixing at low speed for 60 seconds, the mixer was stopped for 60 seconds, in
which the first 30 seconds where used to remove the paste from the wall and bottom part.
Thereafter the mixing was continued at high speed for 60 seconds. The total mixing time, including
stops, is 3 minutes.
A.2.2 Experimental procedure
There are different methods to measure the initial and final setting, but in this research the Vicat
apparatus, as shown in figures 2, has been used. The Vicat test is a standard test for determining the
setting time according to NEN-EN-196-3. This method is based on measuring a depth of a steel
needle into the cement paste. After mixing the different reactants together, the mixture is instantly
placed in a small oiled mould, which is resting on a glass plate. Subsequently the upper surface is
smoothed with a gentle saw motion and the mould is placed in the Vicat apparatus under the rod
bearing the needle. Then the test is started by releasing the needle quickly through the cement
paste. The time between starting time and the time at which the needle fails to penetrate completely
to the bottom of the paste to a point 6.0 3.0mm from the bottom is called the initial setting. This is
measured as the second thick line from the bottom in figure 3. The time between starting time and
the time at which the fails to penetrate through the top of the paste, 0.5 mm into the specimen, is
called the end setting [26]. This is de second thick line from the top in figure 3.
To minimize the effect of the environment the mixtures were prepared and kept in a room with a
controlled temperature of 20 2C and a controlled relative humidity of 50 5% throughout the
duration of the tests. For each mixture the initial and final setting time was determined from two
tests. The average is taken as setting time in the results.

Figure 2: Automatic Recording Vicat Apparatus used for experiment

120

Figure 3: Automatic Recording results from the vicat apparatus

A.3 Strength
A.3.1 General
A.3.1.1 Flexural strength
The capability of a material to withstand a load under deformation is called flexural strength. This has
been determined with the three point flexural test on rectangular prisms according to the standard
NEN-EN 196-1 given in figure 4. The length of the prism should be at least 50 mm bigger than three
times the depth.

Figure 4: Configuration of flexural strength testing of prisms of 40x40x160 mm

The calculation of the flexural strength is done with the standard formula based on hookes law ():

Where:

= the flexural strength; (N/(mm2))

= the ultimate bending moment;
= the moment of inertia of the cross section;
= the depth of the beam;

With centre loading the formula will become:

Where:

= maximum load;
L, b and d are the length, width and depth of the prism in mm.
121

It represents the highest stress at the moment when the first crack happens.
A.3.1.2 Compressive strength
The compressive strength is the maximum load that is needed to crush the material, see figure 5.
This strength is calculated by dividing the maximum load by the cross-sectional area of the prism:

In which:
F
A

= compressive strength of mortar; (N/(mm2))

= the maximum load
= the cross-sectional area

Figure 5: Configuration of compressive strength testing of prisms of 40x40x160 mm

A.3.2 Sample preparation, moulding, curing and demoulding
The sample preparation is the same as described in paragraph A.1.2 sample preparation according to
EN 196-1. Also the curing, moulding and demoulding are in accordance with this code. The material
was kept for half an hour in the bowl after mixing, to give the free lime in TC the opportunity to
react with water. During this half hour the moulds were prepared and covered with oil.
After covering the mould with oil, it was filled with mortar in two layers. The mould has room for
three prisms of a size 40x40x160 mm, see figure 6. After filling the first layer the mould was clamped
and vibrated for 15 seconds. The second layer is filled during the next vibration of 15 seconds. To
ensure a flat surface the mortar was smoothened with a steel ruler in both directions. The mortar
that is exposed to the open air is sealed off with plastic foil to prevent the water from evaporating.
The specimens were restored in a room with a temperature of 205 C and a humidity of 65% for 24
hours. After 24 hours of hydration the prisms were hard enough to demould. The specimen were
removed from the mould and placed in a wet room which has a relative humidity of 100% and
controlled temperature of 202C.

122

Figure 6: Mould (40x40x160 mm) for the strength test and attributes for the strengths (de Vries,
2008)
A.3.3 Experimental procedure
After the specimens had been cured for the proper length of time in the wet room, the prisms were
removed from the wet room and dried with high pressure. Finally the prisms were placed in the
compress attributes (figure 6), which were put in the hydraulic pressure, and tested. The flexural
strength and the compressive strength are both tested at 7, 14, 28 and 56 days. The machine was
operated at a constant rate within the range of 0.125 to 0.50 MPa per second.
A.4 Microstructure
A.4.1 Introduction
The microstructure and hydration process of the cement blends have been studied with a PhilipsXL30-ESEM (figure 8). The images obtained from the backscattered mode (BSE), which is part of the
ESEM, provide information on the distribution of phases such as un-hydrated cement, fly ash and
Top-Crete particles, calcium hydroxide, CSH, porosity, geometrical aspects and the composition of
the hardened cement paste. Other modes will not be satisfactory due to the lack of topography. An
image(2D) from BSE is shown in figure 7 . The different phases are distinguished by their colour.
Black stands for pores, because pores do not reflect the electrons. White displays the unhydrated
core, CSH appears dark grey and CH appears light grey.

123

Figure 7: Image from BSE with different phases. Black = pore; white= unhydrated cement; dark grey =
CSH; light grey = CH.
A.4.2 Sample preparation
The preparation of the samples was done the same way as for the mercury intrusion porosity
experiment. Also the curing and the drying procedure were the same. When a stable mass loss of
0.01% per day was reached, one third of the small specimens were used for the ESEM.

Figure 8: The Philips-XL30-ESEM at the TU Delft and the completed impregnated sample [Guang].
A.4.2.1 Vacuum drying and epoxy impregnation
After the drying period, one third of the small specimens were placed in a plastic bottle and vacuum
dried in a vacuum chamber. The vacuum drying took place at 30mbar for 1,5 hour. The samples for
ESEM were impregnated with a very low viscosity epoxy resin under low vacuum from outside the
vacuum chamber via a plastic tube. Each plastic bottle was filled with epoxy until all specimens were
covered with epoxy. Then the air was step by step released into the chamber to push the epoxy in all
pores. The impregnated sample was cured in an oven at 40 C for 1 day.
A.4.2.2 Cutting, grinding and polishing
After hardening of the epoxy, the impregnated samples were cut into small samples of 1 cm thick
(figure 8). And ground on a lap wheel with p120, p220, p320, p500, p800, p1200 and p4000 sand
papers for 2 minutes. The first minute the sample was put on the wheel and the second minute the
124

sample was put perpendicular to that direction on the wheel. Then the samples were polished on the
lap wheel with 3, 1 and 0.25 m diamond paste each 2 minutes. The polishing was completed with a
low-relief polishing cloth. To prevent damage of the specimens or change in the microstructure,
grinding and polishing should be done very carefully.
A.4.3 Experimental procedure
In order to observe the microstructure of the specimen, the sample was inserted into the microscope
chamber, where the desired pressure and temperature were achieved after approx. 5 min. The
technical operating parameters, i.e. contrast, brightness and magnification were kept the same to
compare the different samples with each other. No polishing or further sample preparation was
performed after imaging with the SEM. The chemical compositions of the different phases were
monitored and recorded in the blended cement paste by using EDXA.
A.5 Hydration heat and degree of hydration
A.5.1 General
Various methods, such as isothermal and adiabatic test setups, are available for measuring the
hydration heat . Also different approaches are used to determine the degree of hydration. In this
study, an isothermal calorimeter as shown in figure 9 was used to determine the hydration heat and
the degree of hydration of the blended cement pastes.The measurements of the heat evolution of
the blended cement pastes have been carried out with a thermometric isothermal conduction
calorimeter (TAM- Air-314). The calorimeter contains eight separate channels for the hydrated
samples and eight channels for the un-hydrated reference sample , each holds a single specimen
during the measurement. The reference has been used to correct for minor temperature
fluctuations. The calorimeter monitors continually the heat evolved from the specimen with time
when the temperature of both the specimen and the surrounding environment are maintained at
isothermal conditions.

Figure 9: A thermometric isothermal conduction calorimeter (TAM- Air-314)

125

A.5.2 Sample preparation and experimental procedure

The isothermal calorimeter has to be calibrated, before each run, at an operating temperature of
20 C. The calibration is performed using a computer program CalForWin. A night before the
measurements takes place or at least 12 h prior to the measurements, the materials are stored in
small quantities in a room with controlled environmental conditions with a temperature of 202 C
to control the placement temperature of the materials. The experimental procedure that has been
followed is according to the report of Sonja Brendle. The main idea of this sample preparation is to
reduce the time between the addition of water to the specimen and inserting the sample in the
calorimeter to a maximum of 30 seconds. In other words, preventing loss of results due to mixing of
the specimen. Approximately 5 gram of each mixtures were weighed on an analytical balance and
put into a glass ampoule. Water was added to the blend in the ampoule and mixed for 20 seconds
with a small vibrating needle maintaining the water to binder ratio of 0.5. These ampoules were
sealed with a cap and inserted in the channels of the calorimeter.

Figure 10: A schematic drawing of an isothermal heat conduction calorimeter.

The sample is placed on top of flow heat sensors, which are in contact with a heat sink (figure 10).
The produced heat was conducted from the sample to the heat sink. The rate of heat evolution is
detected by the thermo sensors as differences in temperature and continually evaluated, processed
and registered by an online computer data acquisition system. The readings were taken every 30
minutes for a period of 7 days.
The reference must approximately have the same thermal properties as the sample. Only it cannot
produce any heat. Common reference materials are dry sand and water. For these measurements
dry blends were used as reference.
A.5.3 Determination of hydration heat and degree of hydration
After the measurement of 7 days, the rates of heat evolution per gram binder were calculated from
every mix design. The average value from four samples of each mix was used for the results. The
determination is presented below:
A.5.3.1 Heat evolution
The heat that was evolved during a reaction was conducted through the thermo sensors. The heat
conduction is converted to a voltage signal (mW/g) proportional to the amount of heat flow as
function of time. Knowing the calibration factor used in the beginning of the measurement, allows
conversion of the voltage signal to the rate of heat evolution (J/g), which is plotted in figure 11.
126

Figure 11: Plot of the recordings from the isothermal calorimeter

Integration of the areas under the rate of heat evolution curves allowes calculation of the total heat
of hydration.

Where Q(t) is the total thermal energy [J/g], P(t) is the thermal power [mW/g] and t is time [h].
A.5.3.2 Degree of hydration
The development of the degree of hydration of the blends at a specific time t can be estimated by
the ratio between the generated heat at time t and the total heat of the binder [34]:

Where (t) is the degree of hydration at age t, Q(t) [J/g] is the accumulated generated heat at time t
during the hydration process, t [h] is the time that is converted to an age that expresses the maturity
of the material during the hydration process, and Qpot is the potential heat produced by the different
materials separately. Which can be defined with the material components and chemical
compositions. Consequently, the potential heat is determined using the following formula [34]:

Where Pcem and Ppoz1,2 (%) are the mass percentages of the materials and Qcem and Qpoz1,2 are the
potential heat of the cement and pozzolans.
A.5.4 Comments on experiment
In order to control the temperature conditions of both the materials and the apparatus during
measurements, the isothermal calorimeter and materials have to be placed in an environment
controlled room. Also disturbances made to the instrument have influences on both the samples and
the reference. Other conditions that may have influence on the technique results is the time during
the sample preparation outside the apparatus.

127

A.6 Pore structure [31], [33] and [34]

A.6.1 Introduction
Mercury Intrusion Porosimetry (MIP) was used to examine the pore size distribution, the porosity
and the effective porosity. MIP is chosen because it provides information on the geometry of the
pores as well as on the connectivity of the pores. MIP measures pore sizes varying from 0.001 m to
1000 m. This method is based on a liquid that enters a pores sample only under pressure. The
pressure that is needed to enter the pores is determined with the Washburn equation.

In which is the surface area of the liquid, is the contact angle and d is the pore diameter. The
Washburn equation assumes that the pores are cylindrical.

A.6.2 Sample preparation and curing

After mixing the samples as described in paragraph A.2.1, the cement paste was poured into 500 ml
plastic bottles. The bottles were sealed with a plastic lid and placed into a rotational machine to
prevent bleeding and segregation in the cement paste. The samples were rotated for 24 hours at a
speed of 5 rotations per minute. The environment had a temperature of 22 2 C. After 24 hours the
samples were removed from the rotational machine and restored in a another room with a
temperature of 20 2 C until the age of testing. The samples were tested after 3, 14, 28 and 56
days. At the end of each curing period, the plastic bottles were broken and the samples were split up
in small pieces of 1 cm3, see figure 12. Only the inner core of the sample was used for testing,
because this part of the sample has been least effected by the environment. This preparation
procedure is the same for other testing methods, like Thermo Gravimetric Analysis (TGA) and
environmental scanning electron microscope (ESEM). Therefore, one sample has been prepared and
used for all three experiments.

Figure 12: Splitting of the sample into small pieces of 1 cm3

A.6.3 Drying procedure
The small pieces were exposed to the air during the breaking of the samples. To minimize the
formation of calcium carbonation, these pieces were immediately frozen by immersion in liquid
nitrogen for 2 minutes (figure 13). Freezing of these small pieces is necessary, not only to prevent the
formation of carbonation, but also to stop further hydration. The frozen water is removed from the
samples by freeze-drying. Ye Guang [34] stated that the freeze-drying method minimizes the damage
128

on the pore structure in comparison with other drying methods, such as oven drying and vacuum
drying. The samples had to be dried in a freeze-dryer (figure 13) with a temperature of -24C and a
vacuum at 0.1 Pa. This drying procedure takes between 10 to 20 days until a stable mass loss of
0.01%/day was reached. Depending on the amount of water used for mixing and the curing age.

Figure 13: Immersion of the pieces in liquid nitrogen to stop hydration and the freeze-dryer.
A.6.4 Experimental procedure
After completing the drying procedure, the small pieces were weighted on an analytical balance with
an accuracy of 1 mg and placed in a 15 ml penetrometer (figure 14) with a stem volume of 1.836 ml.
The weight that is put in the penetrometer depends on the maximum volume of the stem. The
maximum volume of the stem has to be 80% to get the best results. The exact weight of the sample
can be determined with the following formula:

The penetrometer was closed with a steel cap and the weight was measured again including the
steel cap. Finally, the penetrometer was placed in the low pressure chamber of the Micrometrics
PoreSizer 9320 (figure 14) with a pressure increasing from 0.004 to 0.150 MPa.

129

Figure 14: the Micrometrics PoreSizer 9320 and supplies for sample preparation
The pressure in the low pressure chamber is increased to fill the penetrometer with mercury. The
mercury will therefore intrude in the pores, this first procedure is called intrusion. Then the pressure
will decrease, when the highest pressure is reached, this is called extrusion. After the first intrusion a
second intrusion will follow with the same procedure. The penetrometer was removed from the low
pressure chamber, after the second intrusion was completed. The penetrometer is now filled with
mercury and therefore it has to be weighted again on the balance. Hereafter the penetrometer was
placed in the high pressure chamber with a pressure from 0.150 to 200 MPa and again the first and
second intrusion took place. The first intrusion, extrusion and second intrusion are shown in figure
15.

Figure 15: Mercury intrusion and extrusion hysteresis [Ye Guang, 2003]
A.6.5 Determination
The following characteristics can be calculated from the result obtained by MIP:

Total porosity

Pore size distribution

The distribution of the pores can be determined with the following formule:
(

130

Pore size differential

The pore size differential is shown in figure 16, the peaks stand for the pore diameters. The first and
second peak represents the pore diameter of capillary pores and gel pores.

Bulk density

0,4

0,5

14 days 1st intrusion

14 days 2nd intrusion

dV/dlogD [ml/g]

cum intr vol [ml/ml]

The bulk density defines the degree of compactness of a material.

0,3

0,4

0,3

0,2
0,2
0,1

0,1

0
0,001

0,1

10

1000

0
0,001

0,1

10

1000

Figure 16: Pore size distribution (porosity= maximum intruded volume) and pore size distribution
differential curve.

Ink-bottle porosity

There are four types of capillary pores that are distinguished for the interpretation of the results. A
continuous pore (a), a continuous pore with inkbottle (b), a dead-end pore with inkbottle (c) and an
isolated pore (d) (figure 17). The isolated pores cannot be measured by MIP, because these pores
dont have a gap to intrude the mercury. During the extrusion the mercury will be removed from the
pores except from the inkbottles and dead-ends. These pores dont contribute to the permeability of
the cement paste. Therefore ink-bottle is the overestimation of the volume of the fine pores and an
underestimation of the pore volume of the wide ones (Ye Guang, 2003).
The inkbottle is displayed in figure 17. The inkbottle is the volume of the maximum intruded mercury
in the first intrusion minus the maximum extruded mercury in the extrusion divided by the bulk
volume of the sample.

131

Figure 17: a continuous pore (a), a continuous pore with inkbottle (b), a dead-end pore with inkbottle
(c) and an isolated pore (d) [Guang, 2003]
A.6.6 Comments on the experimental method
MIP is a well-known method and is commonly used for the determination of the pore structure of
cement pastes and cement blends. However, this technique has also limitations like most other
techniques used worldwide. The assumption, that pores are cylindrical and equally accessible to
mercury, made in the beginning may be incorrect. Pores dont have to be cylindrical. The MIP
method also measures the pore-entry sizes and not the distribution of the pore sizes. This means
that large pores which have been entered by small pores are registered as small pores. Also before
placing the samples in the low pressure chamber, these small pieces have to be frozen and dried. The
pore structure can change due to the chosen freezing and drying method. Other conditions that may
have influence on the technique are the wrong contact angle and surface tension in the Washburn
equation.
A.7 Chemical bound water and CH content
A.7.1 Introduction
The thermo gravimetric analysis (TGA) has been used to determine the chemically bound water and
the amount of CH of the blends. The TGA measures the amount and rate of weight loss in a material
as function of time or temperature changes under a controlled atmosphere. The results are used to
define the composition of materials up to a temperature of 1000 C. Only materials that exhibit
weight change due to decomposition or dehydration can be characterized.
The relation between TG curve and the amount of chemically bound water and CH content is
determined by a graphical technique mentioned by Marsh et al. [36] based on nods in the curve. The
CH content provides an indication of the progress of the pozzolanic reaction. The chemically bound
water content is used as measure of hydration of cement pastes.
A.7.2 Sample preparation
The sample preparation is similar as that of MIP. After drying the samples in the freeze dryer, one or
two small pieces of 1 cm3 of each mixture were grinded into a fine powder. This powder with a mass
of 65 - 85 mg is used for the TGA experiment.
A.7.3. Experimental procedure
TGA was carried out using a NETZSCH Simultaneous Thermal Analyzer TG-449-F3-Jupiter. The
experimental apparatus is shown in figure 18. An alumina crucible (figure 18), i.e. a small melting pot,
is filled with the fine powder and placed in the machine in a micro furnace. Nitrogen has been used
132

as flow gas to control the atmosphere by passing the crucible. The samples were heated from 40C to
1090C, with a heating rate of 10C/min. The mass of each sample was measured accurately using a
sensor connected underneath the crucible to the crucible by the connection pole.

Figure 18: NETZSCH Simultaneous Thermal Analyzer TG-449-F3-Jupiter and alumina crucible
A.7.4 Determination of chemically bound water and CH content
In figure 19 is a typically weight loss curve shown as function of temperature changes obtained by
TGA. Also the DSC curve, which is the derivative weight (dW/dT) loss, can be seen in the same figure.

Figure 19: Weight loss as a function of temperature changes. DSC and TG curve measured from
thermal gravimetric analysis
Three parameters can be determined from the TG curve. These parameters are the chemical bound
water, the weight loss due to dehydration of CH and the weight loss due to calcium carbonation. The
loss due to CH dehydration is represented in the figure as an abrupt weight loss between 400C and
500C and a peak in the DSC curve. A smaller weight loss between 700C and 800C corresponds to
133

the loss due to calcium carbonate. The weight of CH and calcium carbonate can be defined by the
graphical method of Marsh et al. [36] as illustrated in figure 20. Tangents were drawn along the lines,
which suddenly deflect, causing nodes in the curve. The nodes occur at two temperatures, the
weight loss between the extrapolated lines in the middle of these temperatures is taken as weight
loss due to CH or CaCO3.
The CH amount can be defined by the following chemical reaction, after the graphical method has
been used:
CaO + H2O

Ca(OH)2

(Total reaction)

The total reaction shows that for each mol CaO one mol H2O is added and one mol Ca(OH)2 is
created. The molmass of water is 18,015 g/mol and the molmass of CaOH2 is 74,1 g/mol. The
amount of water that reacted with calcium in the samples is determined with TGA.
The continuous weight loss over the whole range from 105C to 1000C is called the chemically
bound water. The chemically bound water is an interconnected part of the structure of the cement
paste and can only evaporate at temperatures above 105C. This water is assumed to be lost at
heating the sample to 1000C. The difference in mass between 105C and 1000C minus the ignition
loss of cement, fly ash and Top-Crete and minus the loss due to carbonation corresponds to the
chemically bound water.

Where

The ignition loss of the powders is measured by heating the powders at the same temperatures as
the different specimen in the TGA apparatus allowing substances to evaporate, until no mass change
arise. The chemically bound water can be separated into dehydration of CH and dehydration of other
hydrates, like calcium silicate hydrates and calcium aluminate hydrates. CH is considered as one of
the major important phases.

134

Figure 20: Graphical method to calculate the CH content and carbonation content from weight loss
curve (TG-Curve) during thermal analysis [Marsh and Day, 1987]
A.7.5 Comments on procedure [36]
The determination of the CH and CaCO3 content with thermo gravimetric analysis has been accepted
as an accurate method. However also this method shows some limitations and cannot be used as a
precise indicator, because:
a. The actual composition of the hydration products and stoichiometry of the reaction is not
known accurately. This especially holds for the chemical reactions of Top-Crete.
b. The distinction between evaporable and chemically combined waters cannot be made
accurately. For example, drying at 105C is known to result in the loss of some chemically
bound water from the calcium sulpho aluminates and the hexagonal tetracalcium aluminate
hydrates.
c. The interpretation of the obtained results for specimen combined with pozzolanic materials
has been found to be difficult. The CH content at any time does not only depend upon the
amount of reacted Portland cement, but also upon the extent of pozzolanic reaction of the
used pozzolanic material. The CH content tends to increase or decrease due to this
ignorance.
Despite these limitations TGA is still a useful indicator of the CH and chemical water content in
blended cement pastes.
A.8 Chemical shrinkage
A.8.1 General
The accurate measurement of chemical shrinkage in blended cement pastes is very difficult. The
measurement has to immediately take place in a sealed condition without disturbing the blended
cement paste. Over the years, new methods have been developed to determine the chemical
shrinkage. Henri Le Chatelier was first to evolve a method for the determination of chemical
shrinkage in 100% cement with a w/c ratio of 0.4. Chatelier measured the volume contraction of the
135

paste in a glass beaker in connection with a water-filled tube. This contraction is explained as the
difference in density between the initial products and the formed hydration products. The absolute
volume change is almost 10% at full hydration, which is in practice never achieved due to resistance
of the rigid cement paste. However, Chateliers method gives a significant and accurate part of the
chemical shrinkage before setting [90]. Powers [ 91] also provided another measuring method to
quantify the chemical shrinkage in cement. The method is based on the measurement of shrinkage of
the individual clinker compounds in cement. The results were in agreement with the method of Le
Chatelier. A direct method which is used nowadays for the measurement of chemical shrinkage is the
dilatometry test. This method follows Le Chateliers contraction principle and has been applied by
many researchers, such as Geiker [92]. In this study the chemical shrinkage was also measured
according the method described by Geiker.
A.8.2 Procedure
In this method, 50 g of each mixture was blended with water in a cup using a small blender. About 8
g of this freshly blended cement paste was placed at the bottom of a small cylindrical glass jar, with a
diameter of 25mm and a height of 60 mm. Three small glass jars were used for each mixture. After
the blended cement paste was placed at the bottom, the specimen were vibrated for a couple of
seconds to remove air bubbles in the paste and to obtain a uniform thickness spread over the
bottom of the jar. The pastes were then covered with 0.8 ml of water to include a water source in
contact with the paste. The remaining empty part of the jar was filled with paraffin oil and enclosed
with a rubber stop encasing a pipette graduated in 10 ml increments. The change in oil level within
the pipette indicates the volume change due to chemical shrinkage.

Figure 21: Chemical shrinkage test method according to Geiker with a small glass jar enclosed with a
rubber stop encasing a pipette.
A reference filled with oil was used to correct for minor temperature fluctuations. These four small
glass jars were put in an automatically controlled temperature water bath at the required curing
temperature of 20 C. The oil level in the pipette was monitored to the nearest 0.0025 ml over time,
for a period of 7 days. The testing method is given in figure 21.

136

A.8.3 Determination of the chemical shrinkage content

By normalizing the change in volume by mass of blended cement paste in the sample, the chemical
shrinkage per gram of initial specimen (ml/g) was determined. The chemical shrinkage of each
mixtures was taken as average of three samples [92]
A.8.4 Comments on experiment
This simple test procedure is often used, but has several difficulties that can result in inaccurate test
results. Firstly, the method requires manual reading from the pipette every half hour, which takes a
lot of time and accuracy. Secondly, the risk of faulty readings due to the fit of the rubber stop in the
jar. The rubber stop can slide up due to pressure of the oil or cause air bubbles in the oil. The air
bubbles at the surface of the jar can leave the pipette at a later age, which would result in change of
the volume of the oil. Also if the blended cement paste is not vibrated well, the air leaving the
blended cement paste results in change of the submerged weight. The thickness of the paste has also
influence on the results, the blended paste should have enough permeability for the water to
penetrate in all capillary pores. Thicker specimen develop a lower porosity and water cannot
penetrate through the specimen in the capillary pores [89] and [93].
A.9 Autogenous shrinkage
A.9.1 Procedure
The sample was prepared and mixed according to EN 196-1 (see section A.2.1 sample preparation).
After mixing of the specimen, the blended cement paste was immediately cast under vibration into
three corrugated tubes. Each tube was closed with one plug and put with the closed end on the
vibration table in a vertical, rigid support tube. The pourable mixtures were slowly poured in the
corrugated tubes without stopping, while the semi- fluid mixtures were poured in layers. Each layer
was compacted for a couple of seconds by the vibrating table. Then the open ends of the tubes were
closed with another plug and placed on a horizontally supporting corrugated plastic sheets to avoid
damage to the microstructure (see figure 22). All specimens were stored in a room with controlled
environment of 20 2C and 60 5% RH throughout the duration of the tests, to minimize the
influence of temperature fluctuations. The restraining amount of the mixture was cast into a setting
mold and simultaneously measured with the autogenous strain. The setting of the blended cement
paste was determined with the vicat needle test. Only this time the specimen were covered with foil
to prevent moisture loss during the experiment.

Figure 22: Corrugated plastic sheets supporting the specimens during the test [94]
137

The measurement of the length change started about 30 minutes after the addition of water to the
specimen. The measurement was performed every hour and after final setting it was performed
every two hours. The tubes have to be supported over their entire length during the measurements
to avoid damage to the paste, which is shown in figure 23.

Figure 23: Careful measurement of the corrugated tube in the dilatometer bench [94]
A.9.2 Determination of autogenous strain
The length change was measured from the moment the specimen was poured into the corrugated
tubes. However, the calculations of the autogenous strain have been performed at the time of final
setting. The length of the corrugated tube filled with the blended cement paste at time t has been
determined with the following equation:

Where:

Lref
R(t)
Lplug
t

= length of reference bar; (mm)

= reading of length gauge with specimen in the dilatometer; (mm)
= average length of end plugs; (mm)
= time elapsed from the addition of water to the materials; (min)

The time for the determination of the autogenous strain was zeroed at final setting. The autogenous
strain of the blended cement pastes at the time of final setting, expressed as as m/m, is determined
with the following equation:
(
(
where:

tfs

= time of final setting; (min)

138

A.9.4 Comments on experiment

The method used in this ASTM standard has also some limitations and the standard cannot be used
as precise indicator. The two common disadvantages are faulty results due to bleeding and
entrapped air. Bleeding has influence on the autogenous strain due to reabsorption of the water
that is formed on the surface. When this water is sucked back into the specimen it may cause
expansion of the hardened paste. However, bleeding can be prevented by slowly rotating the
blended cement pastes in a horizontal rotational machine. The entrapped air on the other hand has
influence on the deformation of the microstructure of the specimen, hence the autogenous strain.
However, after final setting the entrapped air is negligible, because the specimen has formed a rigid
structure.

139

140

Appendix B: Pore size distribution of the blends with 70% FA

141

142

3 days 1st intrusion

3 days 2nd intrusion

0,4

0,3

0,7

dV/dlogD [ml/g]

cum intr vol [vol/vol]

T(5)PC(65)FA(70)
0,5

3 days 1st intrusion

3 days 2nd intrusion

0,6
0,5
0,4
0,3

0,2

0,2
0,1
0,1
0
0,001
0,4

0,1

10

0
0,001

1000

0,1

0,3

14 days 1st intrusion

14 days 2nd intrusion

10

1000

14 days 1st intrusion

14 days 2nd intrusion

0,3
0,2
0,2
0,1
0,1

0
0,001
0,4

0,1

10

0
0,001
0,4

1000

28 days 1st intrusion

28 days 2nd intrusion
0,3

0,3

0,2

0,2

0,1

0,1

0
0,001

0,1

0,4

10

0
0,001

1000

56 days 1st intrusion

56 days 2nd intrusion

0,3

0,2

0,2

0,1

0,1

0,1

10

pore diameter [m]

1000

0
0,001

10

1000

28 days 1st intrusion

28 days 2nd intrusion

0,1

0,4

0,3

0
0,001

0,1

10

1000

56 days 1st intrusion

56 days 2nd intrusion

0,1

10

pore diameter [m]

1000

Figure 1: Pore size distribution of mixtures T(5)PC(65)FA(70)

143

3 days 1st intrusion

3 days 2nd intrusion

0,4

0,3

0,6

dV/dlogD [ml/g]

cum intr vol [vol/vol]

T(10)PC(65)FA(70)
0,5

3 days 1st intrusion

3 days 2nd intrusion

0,5
0,4
0,3

0,2
0,2
0,1

0
0,001
0,4

0,1

0,1

10

1000

14 days 2nd intrusion

14 days 1st intrusion

0
0,001
0,4

0,3

0,3

0,2

0,2

0,1

0,1

0
0,001

0,1

0,4

10

1000

28 days 1st intrusion

28 days 2nd intrusion

0
0,001

0,3

0,2

0,2

0,1

0,1

0,1

0,4

10

1000

56 days 1st intrusion

56 days 2nd intrusion

0
0,001

10

1000

14 days 1st intrusion

14 days 2nd intrusion

0,1

0,4

0,3

0
0,001

0,1

10

1000

28 days 1st intrusion

28 days 2nd intrusion

0,1

0,3

10

1000

56 days 1st intrusion

56 days 2nd intrusion

0,3
0,2
0,2
0,1
0,1

0
0,001

0,1

10

Pore diameter (m)

1000

0
0,001

0,1

10

Pore diameter (m)

1000

Figure 2: Pore size distribution of mixtures T(10)PC(65)FA(70)

144

3 days 1st intrusion

3 days 2nd intrusion

0,4

0,3

dV/dlogD [ml/g]

cum intr vol [vol/vol]

T(15)PC(65)FA(70)
0,5

0,7

3 days 1st intrusion

3 days 2nd intrusion

0,6
0,5
0,4
0,3

0,2

0,2
0,1
0,1
0
0,001
0,4

0,1

10

1000

14 days 1st intrusion

14 days 2nd intrusion

0
0,001
0,4

0,3

0,3

0,2

0,2

0,1

0,1

0
0,001
0,4

0,1

10

0
0,001

1000

28 days 1st intrusion

28 days 2nd intrusion

0,3

0,1

10

1000

14 days 1st intrusion

14 days 2nd intrusion

0,1

10

1000

28 days 1st intrusion

28 days 2nd intrusion

0,3
0,2
0,2
0,1
0,1

0
0,001

0,1

0,4

10

0
0,001

1000

56 days 1st intrusion

56 days 2nd intrusion

0,3

0,3

0,2

0,2

0,1

0,1

0
0,001

0,1

10

pore diameter [m]

0,1

0,4

1000

0
0,001

10

1000

56 days 1st intrusion

56 days 2nd intrusion

0,1

10

pore diameter [m]

1000

Figure 3: Pore size distribution of mixtures T(15)PC(65)FA(70)

145

146