Documente Academic
Documente Profesional
Documente Cultură
ii
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.
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
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.
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.
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.
-
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.
%
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.
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.
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.
Figure 2.6: SEM Images of the particles of Cem I 42.5 at a magnification of 100x and 1000x [W. Vries]
10
11
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.
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
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
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
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 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]
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. 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.
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 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
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:
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.
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].
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.
28
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
31
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
Portland cement
Top-Crete
Fly ash
Water
Fine aggregates
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
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
100
80
60
Cem I 42,5 N
40
20
0
1
10
Diameter [um]
100
1000
34
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
100
80
60
Fly ash
40
20
0
1
10
Diameter [um]
100
1000
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
100
80
60
Top-Crete
40
20
0
1
10
100
1000
Diameter [um]
{1}
{2}
{3}
{4}
= Control paste/Reference
37
{6}
{7}
*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;
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
{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
{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
Method or standard
Free lime
39
Method or standard
Setting
NEN-EN-196-3
Porosity
Microstructure
Isothermal calorimeter
Chemical shrinkage
Autogenous shrinkage
ASTM C 1698 09
Method or standard
Workability
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
Laser
ACM-104A
Lens:
Flow control unit:
A100
AC-10
0.1 m
200 m
20
Backwards
Amount of cycles:
Time per cycle:
Speed:
Sonicator:
3
300 sec
30
On
41
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
The measurements were conducted on two parallel samples to achieve accurate results.
42
43
C2H4(OH)2 + CaO
C2H4O2Ca + H2O
(2)
C2H4O2Ca
(C2H4O2)2- + Ca2+
(Dissociation Ca - glycolate)
(3)
(C2H4O2)2- + 2H3O+
C2H4(OH)2 + 2H2O
------------------------------------------------------------------CaO + 2H3O+
Ca2+ + 3H2O
(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:
44
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.
46
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
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.
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.
51
TC (CDEM)
TC (Aidicio)
D10 [m]
D50 [m]
D90 [m]
8.12
3.53
10.05
21.46
50.09
93.72
120
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
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
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)
10,6
10,4
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
18
Diameter (cm)
17
16
15
14
13
12
11
10
0
10
12
14
16
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
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
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
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
65
7,0
6,0
5,0
4,0
3,0
2,0
0,0
0
10
12
14
16
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
3,5
3,0
2,5
2,0
1,5
1,0
0,5
0,0
0
10
12
14
16
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
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
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
67
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
60
50
30
20
10
0
0
10
12
14
16
Figure 4.11: Compressive strength against the replacement level of TC for the mixtures with 35% F.
68
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
100% Top-Crete
TC(100)
0% Top-Crete;
5% Top-Crete;
10% Top-Crete;
15% Top-Crete;
PC(100)
T(5)PC(95)
T(10)PC(90)
T(15)PC(85)
71
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
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
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
0,2
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
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
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
0,4
0,3
dV/dlogD [ml/g]
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
0
0,001
0,1
0,2
10
1000
0,3
0,2
0,1
0,1
0
0,001
0,4
0,1
10
1000
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
0
0,001
0,3
0,2
0,2
0,1
0,1
0,1
10
1000
0
0,001
10
1000
0,1
0,4
0,3
0
0,001
0,1
10
1000
0,1
10
1000
0,4
0,3
dV/dlogD [ml/g]
T(5)PC(65)FA(35)
0,5
0,4
0,3
0,2
0,2
0,1
0,1
0
0,001
0,1
10
0,4
1000
0
0,001
0,1
10
0,2
1000
0,3
0,2
0,1
0,1
0
0,001
0,4
0,1
10
1000
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
0
0,001
0,3
0,2
0,2
0,1
0,1
0,1
10
1000
0
0,001
10
1000
0,1
0,4
0,3
0
0,001
0,1
10
1000
0,1
10
1000
0,4
0,3
dV/dlogD [ml/g]
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
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
0
0,001
0,3
0,3
0,2
0,2
0,1
0,1
0
0,001
0,4
0,1
10
1000
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
10
1000
0,1
10
1000
0,1
10
1000
0,4
0,3
dV/dlogD [ml/g]
T(15)PC(65)FA(35)
0,5
0,3
0,2
0,1
0,1
0
0,001
0,4
0,1
10
1000
0
0,001
0,1
0,2
10
1000
0,3
0,2
0,1
0,1
0
0,001
0,4
0,1
10
1000
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
0
0,001
0,3
0,2
0,2
0,1
0,1
0,1
10
1000
0
0,001
10
1000
0,1
0,4
0,3
0
0,001
0,1
10
1000
0,1
10
1000
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
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
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.
90
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
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
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
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
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
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
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
99
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)
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
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
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.
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.
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
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]
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
119
In which:
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.
120
Where:
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
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.
125
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
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.
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
The distribution of the pores can be determined with the following formule:
(
130
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
dV/dlogD [ml/g]
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
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
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
138
139
140
141
142
0,4
0,3
0,7
dV/dlogD [ml/g]
T(5)PC(65)FA(70)
0,5
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
10
1000
0,3
0,2
0,2
0,1
0,1
0
0,001
0,4
0,1
10
0
0,001
0,4
1000
0,3
0,2
0,2
0,1
0,1
0
0,001
0,1
0,4
10
0
0,001
1000
0,3
0,2
0,2
0,1
0,1
0,1
10
1000
0
0,001
10
1000
0,1
0,4
0,3
0
0,001
0,1
10
1000
0,1
10
1000
0,4
0,3
0,6
dV/dlogD [ml/g]
T(10)PC(65)FA(70)
0,5
0,5
0,4
0,3
0,2
0,2
0,1
0
0,001
0,4
0,1
0,1
10
1000
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
0
0,001
0,3
0,2
0,2
0,1
0,1
0,1
0,4
10
1000
0
0,001
10
1000
0,1
0,4
0,3
0
0,001
0,1
10
1000
0,1
0,3
10
1000
0,3
0,2
0,2
0,1
0,1
0
0,001
0,1
10
1000
0
0,001
0,1
10
1000
0,4
0,3
dV/dlogD [ml/g]
T(15)PC(65)FA(70)
0,5
0,7
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
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
0,3
0,1
10
1000
0,1
10
1000
0,3
0,2
0,2
0,1
0,1
0
0,001
0,1
0,4
10
0
0,001
1000
0,3
0,3
0,2
0,2
0,1
0,1
0
0,001
0,1
10
0,1
0,4
1000
0
0,001
10
1000
0,1
10
1000
146