INFRASTRUCTURE AND
CONCRETE TECHNOLOGY
Activated clays as supplementary
cementitious material
Gilles Plusquellec, Arezou Babaahmadi,
Emilie L’Hopital, Urs Mueller
Activated clays as supplementary
cementitious material
Gilles Plusquellec, Arezou Babaahmadi,
Emilie L’Hopital, Urs Mueller
Abstract
Activated clays as supplementary
cementitious material
Concrete is the most used material in the world (buildings, infrastructure, transport) and its production is continuously increasing over the years because of the growth of the population, the urbanisation, and the infrastructure development. Unfortunately, the production of the main component of concrete, cement, causes inevitable CO2 emissions, accounting for 6% of the total anthropogenic CO2 emissions. The most efficient way to reduce this environmental footprint is to reduce the clinker factor in cement or to reduce the cement content in concrete, which is done by replacing a part of the cement by Supplementary Cementitious Materials (SCM). However, the most commonly used SCM (fly ash and ground granulated blast furnace slag) are only available in a low amount in Sweden. New SCM must be find.
The objective of this project was to evaluate the potential of using Swedish clays as SCM. An inventory of available clays was performed in a first step. Then, as clays need to be activated before use with cement, different activation procedures were tested. A selection of clays was mixed with cement either in binary mixes (cement + activated clay) or in ternary mixes (cement + activated clay + limestone). The hydration properties and the microstructure of binder pastes were investigated, as well as the strength development of mortars. Finally, a life cycle analysis (LCA) was performed to evaluate the positive impact on the CO2 emissions when clays are used as SCM.
The results of the project highlighted the good potential of using Swedish clays in concrete to decrease the environmental footprint due to the cement and concrete industries. In particular, the clays can be activated through mechanical and thermal treatment, depending on the type of clay. Thermal treatment in temperature ranges between 600-800 degrees is preferred for sedimentary clays, while a mechanical treatment by ball milling gives better results with marine clays. A satisfactory strength is achieved in mortar samples cast with calcined clays. This was achieved by replacing the cement with 30% of calcined clay and 15% of limestone. Finally, the LCA calculation shows that the use of clay in a ternary binder lead to a reduction of approx. 34% of the CO2 emissions.
Key words: Cement; Clay; Activation; Supplementary cementitious material; sustainability
RISE Research Institutes of Sweden AB RISE Report 2021:25
ISBN: 978-91-89385-10-8 2021
Content
Abstract ... 1
Content ... 2
Preface and acknowledgement ... 4
1 Introduction... 5
2 Generalities on clays ... 7
3 Inventory of clays available in Sweden ... 9
3.1 Clays in sedimentary bedrock ... 9
Ordovician and Silurian bentonites ... 11
Silurian mixed clays (marls) ... 11
Mesozoic kaolinite deposits ... 11
Rhaetian and Jurassic clays ... 11
3.2 Marine deposited clays ... 13
4 Materials and methods ... 17
4.1 Materials ... 17 Clay sampling ... 17 Cement ... 18 4.2 Activation methods ... 18 Calcination ... 18 Mechanical treatment ... 19 4.3 Pozzolanicity tests ... 19
Calcium hydroxide consumption test ... 19
R3 test ... 19
4.4 Paste and mortar mixes ... 20
paste ... 20
Mortar ... 21
4.5 Characterization of the clays ... 21
Clay content ... 21
Mineralogy ... 22
4.6 Calorimetry ... 23
4.7 Thermogravimetric analysis (TGA) ... 23
4.8 Mercury Intrusion porosimetry (MIP) ... 24
4.9 Inductively coupled plasma (ICP) ... 24
4.10 Mechanical tests ... 24
4.11 Life cycle analysis (LCA) ... 24
5 Clay: characterization and activation ... 26
Chemical composition ... 26
Clay content ... 27
Mineralogical properties ... 28
5.2 Activation of the clays ... 30
Effect of thermal and mechanical treatment on the mineralogy ... 30
Pozzolanicity after activation ... 34
Conclusion ... 37
6 Hydration properties of cement paste containing calcinated clays ... 38
6.1 Binary binders ... 38
Calorimetry ... 38
XRD ... 40
6.2 Development of ternary binders containing cement, activated clays and limestone ... 42
Calorimetry ... 42
XRD results ... 45
TGA results ... 48
7 Pore structure and strength development ... 50
7.1 Pore structure ... 50
7.2 Strength development... 51
8 Life cycle analysis (LCA)... 53
9 Conclusions ... 54
10 References ... 55
11 Appendices ... 58
11.1 Activation of the clays ... 58
Preface and acknowledgements
This report combines the results obtained during two projects performed at RISE, Department Infrastructure and Concrete Technology:
- Clay-Cem “Application of calcined clays as supplementary cementitious materials”, initially planned to run 2 years (2017-2018) with a budget of 950 000 SEK. The project was then extended to 2019 for a budget of 1 900 000 SEK under the name “Ternary blends of calcined clays and limestone: Hydration and durability”. Clay-Cem was co-funded by the Consortium for the Financing of Fundamental Concrete Research (consisting of Cementa, Thomas Betong, ABetong, Strängbetong, Betongindustri and Swerock). The authors warmly acknowledge the Consortium for the economic support but also for the helpful discussions during the project reference group meetings, always source of advices and inputs.
- Clay-Bind “Activated clays in future binders for effective and sustainable concrete infrastructures”, started in May 2018 and ended in April 2020 with a total budget of 2 000 000 SEK. Half of that budget was financed by the strategic innovation program InfraSweden2030, a joint venture of Vinnova, Formas and Energimyndigheten. InfraSweden is acknowledged for the financial support and also for the several opportunities they gave us to present our project to an audience.
Both projects were initiated by Arezou Babaahmadi, who was the project leader until September 2019, replaced by Gilles Plusquellec.
The authors would also like to acknowledge the Geological Survey of Sweden (SGU) for collaborating on Clay-Cem, especially the efforts by Mats Enghdal and Mikael Erlström on the preparation of the state of art report on clays in Sweden.
All the colleagues at RISE are also welcome here for their intellectual and practical help: Monica Lundgren, Patrick Fontana, Otto During, Nelson Dias Ferreira Da Silva, Leif Fjällberg, Camilla Lindström, Lovise Sjöqvist, Sandra Davidsson, Alexander Oliva Rivera and Lina Giménez.
1
Introduction
Cement is one of the most used and cheapest construction materials available: approximately 50% of built structures around the world are made in concrete. Limited availability of some construction materials like timber or the questionable durability and costs attributed to steel has caused concrete to be the most demanded construction material around the world. However, the worldwide cement production causes approx. 6 % of the global carbon dioxide emissions (Andrew, 2018), which corresponds to 650-900 Kg CO2/t clinker . This has led to a lot of attention to the sustainability issues of the concrete industry in order to reduce its high CO2 footprint. Approx. 40% of the emissions are attributed to the production process and a lot effort have been made in the recent decades to adjust the cement kilns or to find alternative fuels. The other 60% portion of emissions is only due to a chemical reaction, known as calcination process, happening during the production of cement clinker. During calcination, the limestone (CaCO3) is decomposed at 1400°C to CaO and CO2. The most viable solution to reduce these part of CO2 emissions is to reduce the amount of Portland cement in clinker and replace it with supplementary cementitious materials (SCM). SCM are reactive industrial by-products such as fly ash and blast furnace slag or, if available, natural reactive materials such as volcanic ash (Lothenbach et al., 2011). This leads to blended cements, also called blended binder systems, with a lower environmental footprint due to a reduced amount of calcined limestone within the material flow of cement production (Gartner, 2004). Although this solution has improved the sustainability issues and helped to reduce the amount of clinker to an acceptable level, the decrease of the SCM availability seems to interfere with all this. Most of the well-known SCM like fly ash or slag is going to be rarely available in near future, while the demand for construction is increasing. One of the recent introduced solutions to this problem is to look for other possible alternatives SCM such as clays, which are available naturally in many countries including Sweden. It is shown that clays calcined at a temperature between 500 and 800°C can develop pozzolanic activity, due to the formation of reactive amorphous phases. It can be noted that the needed temperature for producing calcined clays is much lower than the one of cement production, i.e. 1400°C, meaning that less fuel sources are needed. More importantly, the calcination of clay will not release any CO2 (no limestone).
Heating up clays leads to a process called de-hydroxylation, which causes the formation of reactive amorphous phase (He et al., 1995). As reported in the literature, during the hydration of cement blended with activated clays, those latter can react with calcium hydroxide producing calcium silicate hydrates (C-S-H), calcium aluminium silicate hydrates (C-A-S-H), or calcium aluminium hydrates (C-A-H), which are phases known for contributing to the strength and durability of the hydrated binder system (Fernandez et al., 2011, Sabir et al., 2001).
Considering all the benefits, clays have not been used in their full potential as SCM, maybe due to limited availability in some parts of the world. However, considering the shortage of other types of SCM, it is the right time to initiate an investigation about this potential SCM in Sweden.
This project aimed initially to account for an inventory of available and possible to extract pozzolanic materials, specifically clays, in Sweden.
The main goals of the project are:
• to generate an inventory of available Swedish clays with potentials to be used as SCM,
• to initiate a pilot-scale testing of the activation and hydration properties of calcined clays as binary blends with cement or ternary blends combined with limestone, in terms of changes in microstructure and strength development, • to evaluate the positive impact of the use of clays on the CO2 emissions.
The project was divided in following steps:
Step I: Inventory of available materials:
In this phase an inventory of possible available Swedish clays was done. The work was performed in cooperation with Geological Survey of Sweden (SGU). The uptake and characterisation of the clays were also made during this step.
Step II: Activation of the Swedish clays:
In this phase a laboratory scale testing was performed in order to activate clays by calcination. Mechanical activation was also tested on a limited selection of sample. After activation, the pozzolanicity of the clays was evaluated.
Step III: Hydration properties:
This step includes an investigation of the hydration properties of binders blended with calcined clays. Ternary binders containing limestone were also investigated. The analysis includes microstructural and strength properties.
2
Generalities on clays
The use of activated clays as SCM in ancient and recent structures like water tanks, walls and bridges is reported in literature (Mielenz, 1983). These constructions were built applying activated clays and limestone mortars. Such blends were already used by the Romans who also discovered that blended lime binders with SCM like ash, diatomaceous earth or crushed ceramics improves hardened strength properties (Justnes and Østnor, 2015).
Although calcined clays have been used as SCM and are abundant in many places of the world, not so much research has been devoted to the topic until recent years, probably because of the recent need of alternative types of SCM for sustainability reason. Clays became thus a possible option due to this high availability on earth. However, their reactivity must be carefully investigated, the mineralogy influencing the activation procedure.
Clays minerals are categorized in various groups, in which we can differentiate 2 main “families”: the two sheets (1:1) and the three sheets (2:1) clay minerals. Those minerals have different structures but are all made from layers of silicon tetrahedra and layers of aluminium or magnesium octahedra. The 1:1 clays have one tetrahedra and one octahedra sheet, while the 2:1 clays have one octahedra sheet sandwiched between two tetrahedra sheets (see Figure 1). In Sweden, four main clay minerals compose most of the clays: kaolinite (1:1), smectite (2:1), chlorite (2:1) and illite (2:1).
Figure 1. Schematic representation of the structure of 1:1 and 2:1 clay mineral. From (Marchuk, 2016).
Prior to be used with cement, clays need to be activated. This is usually done by calcination. At a specific temperature depending on the clay mineral, the octahedral layer will be dehydroxylated. During dehydroxylation, hydroxyl groups (OH) are released by forming water (Wesley, 2014). This lead to a reduction in the bonding coordination number of the atoms in the octahedral sheet (Khalifa et al., 2020). Other parameters than the calcination temperature can play a role, such as heating rate, calcination duration, atmosphere, and cooling rate. Mechanical treatments (e.g. grinding) are also
known to cause the dehydroxylation of the octahedral sheet due to a local heat induced by the mechanical treatment (Khalifa et al., 2020).
In recent years, many researchers around the world have emphasized on the importance of activated clays when aiming for alternative SCM. The magnitude of research in different countries is largely dependent on the availability of clays within those countries. As a result, China, India and several South American countries having vast amounts of available clays have already pilot scaled related projects to further develop the topic (Karen Scrivener and Favier, 2015). Several European countries have also already started to either support the research in those countries or start looking in to the possibilities in Europe (Karen Scrivener and Favier, 2015). One of the leading European supporters of the research developments on the topic is Switzerland. The Swiss Agency for Development and Cooperation allocated 4 000 000 CHF in 2014 for the research and development of Limestone Calcined Clay Cement (LC3), developed by the EPFL (Ecole Polytechnique de Lausanne) and partners in India and Cuba (Draft 2018, in preparation). In Nordic countries, Norway has devoted resources into research on the topic. For example, SINTEF has reported results within COIN (Concrete Innovation Centre program) on application of calcined marls as an alternative SCM (Justnes and Østnor, 2015, Østnor and Justnes, 2014, Tone Østnor and Justnes, 2013). In Denmark, Alborg Cement together with Aarhus University reported results on differences in calcination processes for the activation of clays (Rasmussen et al., 2015). In Sweden, beside the present projects, Luleå University of Technology (LUT) is running a project studying the mechanical activation of Swedish clays (Tole et al., 2018)
3
Inventory of clays available in
Sweden
This work was performed in cooperation with SGU (Geological Survey of Sweden).
Two main categories of clays exist in Sweden: clays in the sedimentary bedrock and Marine deposited clays.
3.1 Clays in sedimentary bedrock
Clays occur in a wide range of sedimentary environments and display consequently a complex composition, which reflects the physical and chemical conditions during their formation. In addition, those clays have commonly undergone a post depositional alteration. The clay deposits in the sedimentary bedrock have, except for the authigenic kaolin deposits in Skåne, very seldomly a mono-mineral composition and contain beside clay minerals also accessory minerals such as quartz, feldspar and carbonate as well as organic components. This gives quite different characteristics, including thermal properties and usefulness as SCM.
In Sweden, sedimentary bedrocks are found (on land) in Skåne, Västergötland, Östergötland, Närke, Siljan area in Dalarna, in the Caledonides, and on the islands of Öland and Gotland (Figure 2). Outside Skåne, these strata include merely lower Palaeozoic rocks (Cambrian–Silurian), dominated by shale, limestone, and sandstone. Pure clay deposits are subordinate in comparison, and when they occur, they are often mixed with other detrital components. Below follows a short description of clays found in the Swedish sedimentary bedrock.
Ordovician and Silurian bentonites
Scattered thin beds of bentonite clay occur in the Ordovician and Silurian succession. These occurrences are, however, commonly not larger than a few decimetres in thickness and primarily found at great depth. One exception is a two-meter-thick bentonite bed at Kinnekulle, occurring below 30 m dolerite and 90 m Silurian shale. The Ordovician and Silurian bentonites are dominated by a mixed clay mineral association of illite, smectite as well as some kaolinite. The mineral composition reflects sub-depositional diagenetic alteration of the original smectite dominated ash deposit. More detailed description of the mineralogical composition can be found in the literature (Brusewitz, 1986, Snäll, 1977).
Silurian mixed clays (marls)
The clays are variably mixed with carbonate and characterized as calcareous clay or marl. The clay mineral content varies greatly but can locally exceed 70%. The clay mineralogy is dominated by illite and chlorite. Besides carbonate, pyrite and quartz are common accessory minerals. Marls are widely distributed on Gotland, where the Hemse Marl and Mulde Brick Clay Member constitute the most clay mineral-rich deposits. The Mulde Brick Clay has a thickness of ca 25 m and has been quarried for brick and ceramic purposes at Mulde. Any detailed mineralogical data of the clays have not been found more than general description made by Snäll (1978).
Mesozoic kaolinite deposits
Kaolinite is formed in nature by acid reactions in anaerobic medium. The genesis of the kaolinite group of clay minerals is commonly related to intensively weathered soil profiles in subtropical and sub humid climates. These conditions prevailed in Skåne during Jurassic and Cretaceous times and have resulted in authigenic kaolinite deposits occurring beneath variably thick younger sedimentary deposits. One of the most famous localities is the kaolin deposit at Ivö Klack and at Åsen in northeast Skåne (Figure 3). These deposits have been used as raw material for the ceramic industry. Similar kaolin deposits are identified at several places in central Skåne, of which the one at Billinge (Figure 3) is planned to be quarried with a maximum production rate of 200 000 tons kaolin/year if the environmental issues are solved. The judged reserve is estimated to 20-million-ton clay.
Rhaetian and Jurassic clays
Shallow occurrences of mixed clay deposits of Late Triassic (Rhaetian) and Jurassic age are widely distributed in northwest Skåne (Figure 3). These clays have previously been evaluated regarding their suitability as buffer material in the Swedish Nuclear Waste Programme (Erlström & Pusch, 1987). Their study was based on the following delimiting factors:
• Bed thickness >1 m,
• Easy access, limited overburden in relation to clay thickness,
• Clay content > 20% and at least 30% of the clay fraction being either smectite, illite or kaolinite.
Out of their study there emerged four major clay deposits of interest, i.e. the Vallåkra, Höganäs, Kågeröd, and Fyledalen-Vitabäck clays. An overview of the mineralogy in the Triassic and Jurassic clays in Skåne is also presented in Ahlberg et al. (2003).
• The Vallåkra clay
The Vallåkra clay is found beneath the Quaternary deposits in a narrow zone bordering the Rhaetian-Liassic depositional basin in NW Skåne (Figure 3) at depths ranging between 20 m and 400 m. The thickness of the clay unit is commonly between 10 and 20 m. The clays have been quarried at Vallåkra for ceramic purposes, e.g. clay pipes, ceramic tiles and pottery (1956). Here they quarried strata consists of dark grey and grey silty clay with some organic material. The clay content is 50–60% and dominated by kaolinite. The other minerals occurring are primarily quartz. The Vallåkra clay has in addition been quarried at several other places in northwest Skåne, often in combination with open pit mining of coal. The clay in the quarries around Billesholm have primarily been used for production of refractory tiles and bricks (1968).
• The Höganäs clay
The Höganäs clay has a similar age as the Vallåkra clay. The name derives from the open pit at Margreteberg, east of Höganäs (Figure 3) where Höganäs AB has quarried the deposit. The clay is different in comparison to the Vallåkra clay as it contains a significant amount of smectite (beidellite) besides kaolinite. The clay also contains iron-rich oxides which gives the clay a reddish tint. The clay is in the Margreteberg area up to 15 m thick. It was already during the 19th century used as raw material, together with ash, for production of mortar. The mineralogy and distribution are presented in (Norin, 1949, Erlström and Pusch, 1987).
• The Kågeröd clay
The Triassic Kågeröd clay constitutes an up to 50 m thick unit in the upper part of the Kågeröd Formation. The clay underlines the Vallåkra clay. The clay is mottled red–green and contains a highly variable amount of feldspars and quartz in the sand fraction. It can, thus, be classified as an arenaceous clay. It is widely distributed in west and northwest Skåne and is present at high depth. It is also found as subcrop to the Quaternary deposits in a large area between Kågeröd, Kävlinge and Lund. The clay mineralogy is dominated by smectite and minor amount of kaolinite. The clay has not yet been quarried for any industrial purposes. A description of the distribution and properties is available (Erlström and Pusch, 1987).
Figure 3. Distribution of Mesozoic clays in Skåne of potential industrial interest.
• The Fyledal-Vitabäck clay
The Upper Jurassic Fyledal clay and the Upper Jurassic–Lower Cretaceous Vitabäck clay are thoroughly characterized in various studies (Erlström et al., 1991, Erlström and Pusch, 1987). The clays are primarily found at great depth in southwest and west Skåne. However, the clays occur at shallow depth beneath the Quaternary deposits in a 1 to 2 kilometres-wide zone along the Fyledalen Fault Zone (Figure 3). The clays have been quarried for ceramic purposes at Eriksdal during the 20th century. In the Eriksdal area the total quantity of the clay deposit is estimated to exceed 700 million tonnes (Erlström and Pusch, 1987). The colour of the clays varies between green, grey, dark grey and black. This is coupled to variations in the amount of Fe2+ and Fe3+ as well as organic material. The clay mineralogy is a mixture of 10 Å minerals (hydro mica and illite) and a minor amount of mixed layer illite/smectite, especially in the slightly younger Vitabäck clay unit. The Fyledal clay has been used for ceramic purposes, as it possesses unique properties for stoneware products.
3.2 Marine deposited clays
Fine-grained sediments (clay and silt) have different properties that are partly linked to their formation environment and therefore it is convenient to divide them into glacial or post-glacial clay. The fine-grained glacial sediments were deposited during ice melting and sometime later in an Arctic sea or a cold inner sea. The postglacial clays are formed from re-shaped clay particles and organic matter in a temperature equivalent to today or slightly warmer. The only certain method of categorizing the clays in the west coast is
through fossil findings, such as shell of mussels and foraminifers. Foraminifers are single-cell organisms that live in the ocean and have a lime shell. Different species thrive in different environments, which are controlled by, for example, salinity and water depth. The clay and silt in western Sweden have been deposited in three different environments (Stevens, 1987, Stevens et al., 1991): proximal glaciomarin (glacial clay), distal glaciomarin (glacial clay) and shallow marine (mainly postglacial clay) (Fig.3). The sea has been mostly salty when clay particles settled to the bottom throughout this period. The clays on the West Coast lack a clear multi-layering property (vervighet). The layering is caused by seasonal fluctuations of meltwater inflow to the sea. In winter, the flow decreases, and clay particles settle. In summer, the flow is large and fine sand and silt layers can be formed.
In the Baltic Sea Basin, conditions have shifted between periods of brackish water and fresh water, from the inland ice melting up to today (Schoning, 2016) (Figure 4). During the Baltic Sea lake stage, 16 000 - 11 700 years ago, dominating melt water and the seawater were sweet. The multi-layered clays were deposited during this period. During the next stage, of the Baltic Sea development, the Yoldia Sea (Yoldiahavet) (about 200 years), the conditions were brackish water and the multi-layering of the clay declined. The boundary between glacial and postglacial clays is not synchronous, i.e. The glacial clays in the Södra Östersjön were deposited in the Baltic Sea, while the corresponding clays in the Bothnian Sea were deposited in the Ancy Sea (Ancylussjön).
Along the North Coast, there is a type of postglacial clay that has been formed in an oxygen-poor environment, which is called sulphide clay. On the east coast, one can often distinguish between the two clay types, as the glacial clay has a clear multi-layering. Clay is found below the highest coastline (HK) (Fig.3). HK is in the Helsingborg area about 40 meters above sea level, in Gothenburg stretch about 100 meters above sea level, in the Stockholm area about 150 meters above sea level and in the Luleå area about 230 m. On Gotland and Öland there is clay with limited propagation and strength. At levels above the highest coastline, fine-grained sediments can be found as glacial sediments. The total spread of glacial and postglacial clay in Sweden is about 5% of the total area, this is while this value for the glacial is about 70%. The thickness of the clay is usually less than 20 m on the East Coast while on the West Coast it can be around 100 m. The thickness of the postglacial clays inland, which is in fact located on the multi-layered glacial clay on the East Coast, is 1-4 m. In Gothenburg, the thickness of the postglacial clay can be about 10 m. The glacial clay on the west coast has no pronounced multi-layering property because the clay particles were sedimented in a saltwater.
The occurrence and propagation of clays can be seen in the Earth Map, which can be accessed via SGU's Map Viewer (Draft 2018, in preparation).
The clay content of glacial clays usually varies between 20 and 70% and the colour is grey brown. A postglacial clay usually has a clay content between 30 and 60%. In the area around Lake Mälaren, the differences in clay content were significantly smaller between the different clays (Sohlenius and Eriksson, 2009). The colour of the postglacial clay is usually dark grey or almost black due to more organic materials in a postglacial clay.
Other fractions found in both types of clay are mainly silt but sometimes small amounts of fine sand are found. In glacial clays there are also layers of fine sand and silt.
Studies of clay content in the topsoil in Sweden show that clay content is highest in Östergötland, Sörmland. Västmanland and Uppland (Eriksson et al., 1999). There is a clear connection between an area's soil and the underlying origin of the soil in which SGU's soil samples are collected from (Sohlenius and Eriksson, 2009). The archipelago's clay content will soon be presented as Map Viewer at SGU.
The calcium content (calcium carbonate) is usually 1-4% in Western Swedish clays if the sample has been taken deeper than about 2 m below ground level (Engdahl, 1997). In the surface, the calcium carbonate content has been drained away. The lime comes from shell and limestone particles in the clay. Calcium clays (almost 10% calcium carbonate) are also found in Upplands and Östgötaslätterna.
The organic content of postglacial clays is 0-2%.
Illite is the common secondary clay mineral in Western Swedish clays (Stevens and Bayard, 1994). Mixture, vermiculite and kaolinite are also present. Illite is also the most common clay mineral in the clays on the east coast followed by chlorite and kaolinite (Sohlenius, 2006).
The salt content of pore water from western Swedish clays is about 3% from samples taken about 3 m below ground (Söderblom, 1969). The initial content may have been lower in depth where the melting water part of the water has been greater during the sedimentation of clay particles. Almost half of the original salt content can disappear through leaching which happens due to precipitation and groundwater passing through the clay (Löfroth et al., 2011).
The salt content of clays formed during the Baltic Sea lake and the Yoldia Sea was about 1% (Schoning, 2001). In the clays formed during the Baltic Sea and Ancy lake, the salinity should be lower.
Analysis of grain size, calcium content, organic content and pH have been made on samples taken in conjunction with SGU's soil mapping operations in southern and central Sweden.
4
Materials and methods
4.1 Materials
Clay sampling
Sample uptake was performed in 15 selected locations as illustrated in Figure 5. The chosen sites of sampling are:
1. Västervik 2. Nyköping 3. Östhammar 4. Uppsala 5. Örebro 6. Linköping 7. Lilla Edet 8. Göteborg 9. Vitting I Uppland 10. Mariestad 11. Laholm 12. Sandviken 13. Skara 14. Eriksdal 15. Vallåkra
Figure 5. Sampling locations.
The locations are chosen to enable a screening of marine clay sources both in east and west of Sweden as well as sedimentary clays in the south. Figure 6 shows the sampling locations in Eriksdal (top) and Vallåkra (bottom).
Västervik Nyköping Östhammar Uppsala Örebro Linköping Lilla Edet Laholm Eriksdal Valåkra Sandviken Skara Mariestad Vittinge West Coast
Marine glacial and postglacial clays
Eastern central Sweden
Brackish marine glacial clays, freshwater deposited clays occur
Littorina Sea
Brackish marine postglacial clays, including sulfide clays
Areas below the highest shoreline, where the salinityis expected to have been at its lowest during clay deposition
Areas above the highest shoreline
Figure 6. Sampling in Eriksdal (top) and Vallåkra (Bottom).
Cement
The Portland snabbcement CEM I 52.5 R from Cementa (Skövde) was used in all this study. Its composition is given in Table 1.
Table 1. composition of the cement used (provided by the producer). Blaine
(m2/kg) Al2O3 CaO Fe2O3 K2O MgO Na2O SO3 SiO2
531 5.2 62.9 3.0 1.3 1.3 0.17 4 19.1
4.2 Activation methods
Calcination
Prior to calcination, the clays were dried at 70°C and sieved in order to only keep the particle below 5 µm. The calcination was performed during one hour in a furnace (from Carbolite) with a possibility of controlled temperature up to 1000°C. Five different temperatures were used: 600, 650, 700, 750, 800 and 1000°C.
Mechanical treatment
Ball milling was performed on a different set of samples to check the potential influence of mechanical action.
Mechanical grinding can induce changes in the material, e.g. delamination, leading to a more disordered/amorphous structure and thus a lowering of the calcination temperature needed for dihydroxylation (Wesley, 2014).
Two different treatment were applied, both by using a planetary ball mill PM 100 from Retsch:
1. Ball milling during 5 min at 450 rpm with. This was made prior to calcination. 2. Ball milling during 20 min at 500 rpm, using 20 mm diameter balls and a
ball-to-powder ratio of 25 by mass. Those parameters were taken from a study in which the authors successfully activate illitic phase by grinding (Tole et al., 2018).
Note: it has to be pointed that this last procedure has been tested towards the end of the project and on a single clay. Therefore, the clay from Linköping tested with this method comes from a different batch (and a different supplier) than the rest of the study. It is thus not recommended to not compare the results obtained on those two batches directly. The conclusions drawn in this report on this type of activation are nonetheless still valid.
4.3 Pozzolanicity tests
Calcium hydroxide consumption test
To investigate the pozzolanic activity of the calcined clays, suspensions of calcined clay and calcium hydroxide solution was prepared with a liquid-to-solid ratio of 50. The changes in the concentration of calcium ions in the suspensions over time was measured through inductive coupled plasma spectrometry (ICP) analysis. The measurements were performed at 14 and 28 days. The consumption of calcium is proportional to the pozzolanicity of the tested material.
R3 test
A second method was used, the R3 (“Rapid, Relevant and Reliable”) Rilem test (Avet et al., 2016). It consists of an isothermal calorimetry study carried out on model mixes at 40°C. Those mixes are composed of clay, portlandite and gypsum, with the following proportions: portlandite-to-clay ratio of 3:1 and SO3/Al2O3 molar ratio of 1. Gypsum is added to reach that ratio. Finally, the powder is mixed with a 0.5 mol/l KOH solution to reach a water-to-solid ratio of 1. The heat flow was recorded at 40°C up to 7 days of hydration. This test was performed on the clays after a calcination at 800°C. The R3 test
allows to rank the pozzolanicity of SCM in time frame much shorter than the previous method. The repeatability is also better according to (Avet et al., 2016).
4.4 Paste and mortar mixes
paste
Binder paste were casted as binary binder (cement/activated clay) and ternary binder (cement/activated clay/limestone). For the latter, a sulphate adjustment is required, which is made by adding gypsum (more information in section 6.2). The water-to-binder ratio is set at 0.45. The cement is provided by Cementa, the limestone by Nordkalk and the gypsum by Merck (CaSO4.2H2O, minimal purity of 99%).
The binary binders were investigated by calorimetry at 20°C during 7 days of hydration. Different cement replacement rates were used: 10, 20 and 30% of activated clay. The pastes were analysed by XRD after stopping the hydration right after the calorimetry ended.
The different ternary binders are summarized inError! Reference source not
found. Table 2. For each binder, 8 samples (2 duplicates for 4 different ages) were casted
in 50 ml plastic container and stored at room temperature. After the different hydration times, i.e. 2, 7, 28 and 56 days, samples were crushed in centimetric pieces and dried by solvent exchange. The pieces were submerged in isopropanol (2-propanol ≥ 98 % technical, VWR chemicals) for 7 days. During that period, the isopropanol was renewed two times the first day, and one time the 2nd, 3rd and 5th day. The pieces were then dried in an oven at 35 °C for 24 h. The samples were stored in a desiccator with silica gel and soda lime as CO2 trap. Approx. 8g of centimetric pieces of each sample was used for MIP while the rest were dry grinded during 3 min with a mortar mill RM 200 form Retsch equipped with sintered aluminium oxide mortar and pestle. The obtained powders were measured by TGA and XRD.
Table 2. Paste mixes. The indicated % relates only to the dry binder.
Vallåkra Eriksdal Göteborg
% m (g) % m (g) % m (g) Cement 53 69.0 54 70.3 54.5 70.9 Clay 30 39.0 30 39.0 30 39.0 Limestone 15 19.5 15 19.5 15 19.5 Gypsum 2 2.60 1 1.30 0.5 0.651 Total 100 130 100 130 100 130 Water - 58.6 - 58.6 - 58.6
Mortar
Mortar samples mixes were designed according to the Svensk standard SS-EN 196-1:2016 with an adaptation of the water to binder ratio from 0.5 to 0.45. Cement, limestone, calcinated clay and gypsum were used (same material as described in section 4.4.1). The sand was a CEM standard sand EN 196-1 from Normsand.
The mixes are summarised in Table 3. Each sample was casted in order to have 3 prisms of 40*40*160 mm at each following age: 7, 28 and 56 days. The samples were stored at room temperature under water. After the different hydration times, flexural and compressive strength were measured.
For those ternary binders (cement, clay, limestone), only mixes with 30% of clay have been used. The amount of limestone, 15%, has been chosen based on the literature (Antoni et al., 2012, Scrivener et al., 2018).
Table 3. Mortar mixes. The indicated % relates only to the dry binder.
Vallåkra Eriksdal Göteborg Reference % m (g) % m (g) % m (g) % m (g) Cement 53 239 54 243 54.5 245 100 450 Clay 30 135 30 135 30 135 0 0 Limestone 15 67.5 15 67.5 15 67.5 0 0 Gypsum 2 9.0 1 4.50 0.5 2.25 0 0 Sand - 1350 - 1350 - 1350 - 1350 Water - 202.5 - 202.5 - 202.5 - 202.5
4.5 Characterization of the clays
Clay content
The clay content of each sample was provided by the Swedish geological survey (SGU). It was determined by sedimentation/hydrometer method as follow.
The grain size analysis of the samples has been performed by 1) using sieves (fraction 20 – 2 mm) and 2) sedimentation (fractions < 2 mm).
After sieving, the amount of sample in each sieve is weighed and the result is used to calculate the grain size fraction.
The fraction smaller than 2 mm is run through a sedimentation analysis. The sample is first dispersed in water. The material then slowly settles in the container. Because fine clay particles settle more slowly than coarser silt particles, it is possible to determine the fraction of clay material in the material. This is done by measuring the density of the suspension at given times. If the clay fraction is high, the density will decrease slowly with time. The clay fraction is defined as particle size below 0.002 mm. When the density
corresponding to this particle size is obtained, the amount of particles that has not sedimented is measured. It represents the clay fraction of the sample.
Mineralogy
The characterization of crystalline phases was performed by X-Ray Diffraction (XRD) measurements with a Rigaku Miniflex 600 with a fast 1d solid-state detector. The sample were front loaded. The diffractogram was performed between 2 and 42° 2Ɵ with an increment of 0.02 and a scanning speed of 1°/min. To allow intensity and peak shift comparison, 10 wt.% of corundum was added as internal standard.
However, a special sample preparation prior to XRD is needed in order to prevent preferred orientation:
Step 1: Preparation of a suspension by mixing approx. 0.1 g of clay with 10 ml of deionised water, and dispersion of the suspension by ultrasound during approx. 30 s.
Step 2: Calculation of the time needed for a specific particle size (R = 1 µm here) to sediment down until the desired depth (here 3 mm), using Stokes low, see Eq. (1).
v = 2/9 ∙ ((ρp - ρf)/μ) ∙ g ∙ R2 (1)
Where v is the speed of the clay particles (m/s), ρp and ρf are the density of the clay and water (kg/m3), µ the dynamic viscosity of water (Pa.s), g the gravitational acceleration (m/s2) and R the radius of the considered particle (m).. This give us a waiting time of 30 min.
Step 3: Pipetting 1 ml of the of the suspension at 3 mm of depth from the surface, 30 min after preparing the suspension, and dropping it on a glass film. The glass film is left to dry out in a plastic box containing silica gel to fasten the drying process.
Step 4: Using the dried glass film for XRD. The analysis was performed between 2 and 42° 2Ɵ with an increment of 0.02 and a scanning speed of 1°2Ɵ/min.
Figure 7. Sample preparation for XRD analysis.
Finally, an ethylene glycol treatment was performed in order to distinguish between clays containing smectite or chlorite. For this treatment, the dried glass film prepared for XRD were placed in a desiccator above ethylene glycol, and a weak vacuum were made in the desiccator. The samples stayed in the ethylene glycol atmosphere for several days prior to XRD analysis.
Smectite can absorb ethylene glycol between its sheet, causing a shift in the XRD pattern. Chlorite is not affected by the treatment.
4.6 Calorimetry
The early heat development was measured by isothermal calorimetry in form of time vs. cumulative heat and time vs. heat flow curves over 7 days. The analyses were performed with a TAM Air isothermal calorimeter. The dry components were precisely weight (precision up to 0.0001 g) and mixed together. The desired amount of solution (Milli-Q water or KOH solution in the case of R3 test) was taken with a propipette, added to the solid mix and its weight were also measured. Right after, the mix was stirred for 2 min with a spatula before the introduction of approx. 5 g of the paste in a calorimetry vial with a funnel. The mass of the sample in the vial was precisely measured before sealing the vial and introduce it into the calorimeter. The heat release was measured during the hydration of the paste.
4.7 Thermogravimetric analysis (TGA)
For the TGA measurements a Mettler Toledo TGA-DSC3+ was used. Approximately 30 to 50 mg of the powdered paste samples were weighed into 70 µL alumina crucibles. The
Sedimentation/Decantation 30 min Sampling Stocks Law: V=2/9*((ρp-ρf)/μ)*g*R2 3 mm 2 µm particles Drying XRD
samples were heated from 20 to 1000 °C at a rate of 10 °C/min and purging flow of 50 mL/min N2.
4.8 Mercury Intrusion porosimetry (MIP)
The analysis of the pore structure was performed by MIP analysis. The method itself is described in (Rübner and Hoffmann, 2006). The samples were analysed at the age of 2, 7, 28 and 56 days.
4.9 Inductively coupled plasma (ICP)
The solutions were analysed by ICP by ALS Scandinavia. All samples were acidified (0.1 mol/l) before analysis. The dilution, acidification and calibration procedures were not provided by ALS.
4.10 Mechanical tests
Mechanical strength was tested on standard mortar bars cast according to: EN 196-1 in form of compressive strength tests. Tests were performed on a standard mortar press for cement mortars (load cell 300 kN, Tony Technik).
4.11 Life cycle analysis (LCA)
The LCA has been performed by Otto During. Three different recipes have been taken into account:
- One pure cement, with a CEM I 52,5 R from Cementa
- One mix containing a sedimentary clay, Vallåkra. Composition: 53% CEM I, 30% calcined clay, 15% limestone and 2% gypsum
- One mix containing a marine clay, Lilla Edet. Composition: 53% CEM I, 30% calcined clay, 15% limestone and 0.5% gypsum
For the calculation, all the clays were considered calcined at 800°C.
The different emissions during transport (Table 4) and the different processes involved in the production (Table 5) of the three binders (reference CEM I and ternary binders made from a sedimentary clay and a glacial/postglacial clay) were taken into account.
Table 4. CO2 emissions due to transport of the materials
Material From To Distance
(km) g CO2/kg
1 Sedimentär Lera Vallåkra Skövde 327 18,0
2 Glacial Lera Lilla Edet Skövde 125 6,9
3 Bränd lera Skövde Borås 111 6,1
4 Kalksten Vingåker Köping 73 4,0
5 Kalkfiller Limus 25 Köping Borås 319 17,6
6 Cement Skövde Borås 111 6,1
7 Gips Vingåker Köping 73 4,0
8 Köping Borås Köping Borås 319 17,6
Table 5. CO2 emissions due to the various industrials processes
Process Reference Environment impact
Mining Clay Ecoinvent 3.3 Clay (CH)Clay pit operation 7,73gCO2/kg
Calcination of Clay Cassagnebére 2011 304 gCO2/kg
Mining Limestone Ecoinvent 2010 0,665 g CO2/kg
Grinding of Limestone Nordkalk 2011 1,04 g CO2/kg
Production of Cement
Cementa AB. (2019). EPD Portland Cement CEM I 52.5 R, Skövde. EPD-HCG-20190140-CAA1-EN.
824 g CO2-ekv/kg
Production of Gypsums
Ecoinvent 3.3 Gypsum Mineral (CH), no infra
structure, Alloc 2,21 gCO2-ekv/kg
Production of
Electricity Ecoinvent 3.3 Svensk elmix 41,7 g CO2-ekv/kWh
Diesel Ecoinvent 3.3 Diesel in working machine 74,4 gCO2/MJ Coal Värmeforsk 2011 Miljöfaktabok för bränslen 115 g CO2/MJ
5
Clay: characterization and
activation
5.1 Characterization of the clays
Chemical composition
The chemical composition of the pure clay samples obtained from the different locations (Error! Reference source not found.), are presented in Figure 8. The results show that the SiO2 content of the clays lies in the range between 55-70 wt.% except for the clay taken in Östhammar which has a lower silica content (approx. 40 wt.%). The calcium oxide content of the clays is around 2 wt.%, except for the clay from Östhammar which contains 20 wt.% CaO. The average aluminium oxide content is around 16 wt.% and the amount of other alkalis is rather low (less than about 6 wt.%) in all the samples.
Interestingly, the chloride content is low in all the samples, i.e. 0.01 wt.%, and even lower (below detection limit) in the sedimentary clays. This very low chloride content is in accordance with the standards for SCM (SS-EN_197-1_2011), the indicated upper limit for the chloride content being 0.1 wt.%. A higher Cl content was expected considering the marine origin of the clays.
Figure 8. Chemical composition of clay samples.
Clay content
The clay content of the different samples is shown in Table 6. SGU did not provide data for the samples coming from Vallåkra and Eriksdal. Overall, the minimal clay content is 30% for Laholm. Most of the samples we obtained present a clay content above 40%, which is the minimal amount recommended by other authors to obtain a cement with good properties (Alujas et al., 2015).
0 0.01 0.02 0.03 Ö rebr o Lin köpin g Lilla Ed et Vit tin ge Mar iestad Lahol m Sand viken Skar a Ö sth am m ar Up p sala Vä sterv ik N yköpin g Eri ksd al V allåkra Wt . % Cl Glacial Sedimentary 0 2 4 6 8 10 12 14 16 18 Ö rebr o Lin köpin g Lilla Ed et Vit tin ge Mar iestad Lahol m Sand viken Skar a Östh amm ar Up p sala Vä sterv ik N yköpin g Eri ksd al V allåkra Wt . % Fe2O3, MgO, K2O & Na2O Fe2O3 K2O MgO Na2O 0 5 10 15 20 25 30 Ö reb ro Lin köpin g Lilla Ede t Vit tin ge Mar iestad Lahol m Sand viken Skar a Ö sth am m ar Up p sala Vä sterv ik N yköpin g Eri ksd al V allåkra Wt . % Al2O3& CaO Glacial Sedimentary Al2O3 CaO 40 45 50 55 60 65 70 Ö reb ro Lin köpin g Lilla Ede t Vit tin ge Mar iestad Lahol m Sand viken Skar a Ö sth am m ar Up p sala Vä sterv ik N yköpin g Eri ksd al V allåkra Wt . % SiO2 Glacial Sedimentary
Table 6. Clay content depending on the location. Location Clay content (wt.%)
Örebro 53 Linköping 82 Lilla Edet 70 Vittinge 73 Mariestad 70 Laholm 31 Skövde 40 Västervik 37 Nyköping 74 Östhammar 41 Uppsala 35 Skara 40 Sandviken 63
Mineralogy
The XRD results are presented in Figure 9. The glacial clays are mainly a mixture of smectite, illite and kaolinite. No major difference in mineral properties is observed between the east and west of Sweden at the exception of the sedimentary clays that are smectite free.
Figure 10 shows the diffractograms of the clays before and after the ethylene glycol treatment. This experiment allows to distinguish smectite and chlorite which appear at the same diffraction angle at ≈ 6° 2θ. Smectites are expandable, and in presence of ethylene glycol the peak shift to smaller angle, while ethylene glycol will not affect chlorites.
For the Östhammar clay, the peak position is not affected by the ethylene glycol indicating the presence of chlorite. For the 10 other clays, the peak is shifted to lower angle indicating the presence of smectite, probably as mix layered smectite/illite, the peak at ≈ 9° 2θ (corresponding to illite) being hardly influenced by the ethylene glycol treatment.
Figure 9. XRD analysis results on Swedish clays from different area. Dark blue diffractograms are glacial and post-glacial clay from east of Sweden while light blue diffractograms are from west of Sweden. Dotted dark diffractograms are from sedimentary clays. *Chlorite is only present in Östhammar, while all the other clays have smectite.
5 10 15 20 25 30 Angle (2θ) CuKα Eriksdal Nyköping Uppsala Lilla edet Mariestad Laholm Skara Vallåkra Göteborg Östhammar Örebro Linköping Västervik Vittinge Il lit e K aol ini te Il lit e K aol in it e Sm ect it es or chl or it e* Sm ect it es or chl or it e* Qua rt z Sm ect it es or chl or it e* Fe ld sp ar s Q ua rt z + Il lit e
Figure 10. XRD analysis of raw and ethylene glycol treated clays.
5.2 Activation of the clays
Effect of thermal and mechanical treatment on the
mineralogy
The effect of a short mechanical treatment (5 min) followed by calcination with temperatures ranging from 600 to 1000°C on the clay phases (i.e. kaolinite, illite, smectite/chlorite) was follow by XRD. Since all the different studied clays present the same phases (see Figure 9), only the examples of Lilla Edet and Nyköping are presented here in Figure 11 (but all diffractograms are presented in the appendices). The straight lines correspond to a calcination only and the dotted lines correspond to a ball milling followed by calcination.
All the kaolinite peaks disappear after calcination, even at the lowest tested temperature. This indicates that crystalline kaolinite is transformed to an amorphous product called metakaolin. Smectite is also not detected after calcination at 700°C and above.
Illite is remaining after calcination up to 900°C. This shows a poor answer of this phase towards thermal treatment. This might be due to a too low calcination temperature: Hollanders et al. (2016) showed that pure illite can be activated at 900°C, but only a
4 6 8 10 12 14 16 18 20
Angle (2θ) CuKα
Raw clay
Ethylene glycol treated clay
Eriskdal Nyköping Uppsala Lilla edet Mariestad Laholm Skara Vallåkra Göteborg Östhamma r Örebro Linköping Västervik Vittinge Angles affected with ethylene glycol
partial and limited amorphization was observed. Illite is not visible after a calcination at 1000°C according to the XRD results presented in Figure 11. This might indicate the successful activation of this phase. However, considering the high temperature, it can also be a sign for its vitrification. This latter possibility must be avoided because it will lead to a decrease in reactivity. This was checked by isothermal calorimetry using cement paste containing 30% of calcined clay. Figure 12 show the result of the test. A cement paste with a substitution rate of 30% of clay from Lila Edet was measured up to 7 days after mixing. The evolution of the total heat is presented here. One can see that an increase of the activation temperature from 800°C to 1000°C leads to a decrease of the total heat after 7 days of hydration. This indicates a lower reactivity of the clay.
Therefore, a calcination at 1000°C should not be used to activate the Swedish illitic clays.
In addition, the short mechanical treatment helped to decrease the amount of crystalline illite to some extent, as presented in Figure 11 (dotted lines). When comparing the effect of calcination with or without pre-grinding, we can see a difference only at a temperature of 800°C for Lila Edet. In this case, the diffractogram shows large “bumps” indicating that the sample is much more amorphous if the sample is mechanically ground. Therefore, this sample should be more reactive if used in a binder. Few other samples show this amorphization after ball milling (see appendices).
Finally, the long mechanical treatment was applied on a clay coming from Linkoping. The results are shown in Figure 13. We can clearly see that all the intensity of the peaks corresponding to the different clay phase (i.e., smectite, illite and kaolinite) are significantly decreasing after a ball milling. Only a small amount of non-amorphized illite remains.
Figure 11. XRD analysis of Lilla Edet (A) and Nyköping (B) clay at different calcination temperatures with (dotted line) and without (solid line) ball milling.
5 10 15 20 25 30
2θ
A. Lilla Edet clay
Reference 1000°C 900°C 850°C 800°C 750°C 700°C 650°C 600°C Ca l ci nation Ba l l milling + ca l cination Sm e cti te Sm e cti te Sm e cti te Il lite Il lite K a o li n ite K a o li n ite F e ld sp a rs 5 10 15 20 25 30 2θ B. Nyköping Clay Reference 1000°C 900°C 850°C 800°C 750°C 700°C 650°C 600°C Ca l ci nation Ba l l milling + ca l cination Sm e cti te Sm e cti te Sm e cti te Il lite Illite K a o li n ite K a o li n ite Fe ld sp ar s
Figure 12. Calorimetry investigation of the effect of the activation of illitic clays at 800°C and 1000°C on the pozzolanicity. Example of a cement paste with 30% substitution of Lila Edet clay. Values normalized by the amount of powder (cement + clay) in the sample.
Figure 13. Activation of a clay from Linköping by a long ball milling treatment. Before (black) and after (blue). The intensities are normalized to the intensity of the Quartz peak at approx. 21 °2Θ.
0 50 100 150 200 250 300 350 0 48 96 144 H ea t (mJ /g pow der ) Time (h)
Heat
Lila Edet 30 -800C Lila Edet 30 -1000C5
10
15
20
25
30
Angle (2θ) CuKα
Linkoping Linkoping BMIl
li
te
K
ao
li
n
it
e
Il
li
te
Sm
e
ct
it
e
Sm
e
ct
it
e
Qu
ar
tz
Qu
ar
tz
+
Il
li
te
K
ao
li
n
it
e
Sm
e
ct
it
e
Pozzolanicity after activation
In this section, the two different pozzolanicity test are presented.
5.2.2.1 Calcium hydroxide consumption test
The consumption of calcium hydroxide by the different clays calcinated at 650 and 800°C with or without mechanical treatment (ball milling) is shown after 14 and 28 days in Figure 14. The initial solution of calcium hydroxide is shown as a reference to visualize the Ca consumption of the samples. The difference between this curve (“No clay”) and the curves at 14 and 28 days (after the clay addition) corresponds to the consumption of calcium hydroxide.
The kaolinitic clays, i.e. Vallåkra and Eriksdal, consume more calcium hydroxide than the other clays. This is linked to higher kaolinite content. The calcium hydroxide consumption does not increase significantly after 14 days, indicating a quick clay reaction. The different activation temperatures do not have a significant impact for those clays. This implies that a calcination at 650°C is enough for both Vallåkra and Eriksdal (sedimentary) clays.
The glacial and post glacial clays show lower calcium hydroxide consumption, meaning that their pozzolanicity is lesser than sedimentary clays. This is particularly visible for Göteborg, which presents a very limited pozzolanicity.
For this type of clay, the time of immersion does not have a clear impact for Göteborg, Lilla Edet and Linköping: the results are similar after 14 and 28 days. On the contrary, for Niköping, Mariestad and Vittinge, more calcium hydroxide is consumed after 28 days than 14 days. This indicates a slower reaction for those clays.
The calcination temperature does not have the same impact depending on the location. While an increased temperature lead to a clear increase of calcium hydroxide consumption for Nyköping and Vittinge, this is not observed for the other clays.
For every clay and for the temperatures tested here, we did not measure any significant change of pozzolanicity after the mechanical treatment. This is also true for Lilla Edet, even though this clay presented a higher amorphous content after a short ball milling (Figure 11).
Figure 14. Consumption of Ca by the activated clays after 14 and 28 days of immersion in Ca(OH)2
saturated solution. The activation method is indicated in the x axis: calcination at 650°C or 800°C with or without short ball milling prior to calcination (BM).
0 5 10 15 20 25 [C a] m m ol /l Nyköping No clay 14 days 28 days 0 5 10 15 20 25 [C a] m m ol /l Lilla Edet No clay 14 days 28 days 0 5 10 15 20 25 [C a] m m ol /l Linköping No clays 14 days 28 days 0 5 10 15 20 25 [C a] m m ol /l Mariestad No clays 14 days 28 days 05 10 15 20 25 [C a] m m ol /l Vittinge No clays 14 days 28 days 0 5 10 15 20 25 [C a] m m ol /l Vallåkra No clays 14 days 28 days 0 5 10 15 20 25 [C a] m m ol /l Göteborg No clays 14 days 28 days 0 5 10 15 20 25 [C a] m m ol /l Eriksdal No clay 14 days 28 days
5.2.2.2 “R3” test
Figure 15 shows the results of the R3 test performed by isothermal calorimetry. The data are presented as the evolution of the cumulative normalized heat (W/g total powder) up to 168 hours, i.e. 7 days. The ranking of the pozzolanicity of the clays is directly visible: the higher is the cumulative heat at 168 hours, the higher is the pozzolanicity.
The sedimentary clays coming from Vallåkra and Eriksdal have clearly the highest pozzolanicity, as also seen in the calcium consumption test (Figure 14). The glacial and post glacial clays form a second group with a lower pozzolanicity. However, Mariestad and Linköping clays shown a slightly higher reactivity than Vittinge, Nyköping and Lilla Edet.
Figure 15. R3 pozzolanicity test: evolution of the normalized heat flow (W/g powder) as a function of the time. Clay calcined at 650°C.
The clay from Linköping that underwent the long activation by ball milling was also tested with this test. The result is compared to the same batch of clay but after calcination in Figure 16. One can see that the long ball milling activation results in a better pozzolanicity, confirming the activation of the different phases, including illite.
For this type of clay (glacial/ post glacial), it is thus recommended to perform the long ball milling activation procedure.
0
20
40
60
80
100
120
140
0 24 48 72 96 120 144 168
N
or
m
al
iz
ed
he
at
J/
g
pow
de
r
Time (h)
Vallåkra Eriksdal Mariestad Linköping Vittinge Nyköping Lilla EdetNote: Unfortunately, the activation by ball milling was successfully tested towards the end of the project, and it was not possible to run more investigation using a clay activated this way (e.g. hydration, mechanical properties etc).
Figure 16. R3 pozzolanicity test: comparison of different activation treatment on the clay from Linköping – Calcination or long ball milling.
Conclusion
After activation of the clays, the two clay groups (sedimentary and glacial) show very different behaviour.
- After calcination, the sedimentary clays present a much higher pozzolanicity compared to the marine clays. For all clays, the activation temperature ranges between 650 and 800°C.
- The marine clays can be activated more efficiently with the mechanical treatment described in section 4.2.2.
The activated clays can now be blended with cement to investigate the hydration properties.
0
10
20
30
40
50
60
70
0 24 48 72 96 120 144 168
N
or
m
al
iz
ed
he
at
f
low
W
/g
pow
de
r
Time (h)
Linkoping - BM Linkoping - Calcined6
Hydration properties of cement
paste containing calcinated clays
In this section we investigate the hydration properties of binary and ternary binders. Binary binders contain cement and calcined clays, while ternary binders contain limestone in addition.
6.1 Binary binders
Calorimetry
Figure 17 presents the calorimetry curves obtained on cement pastes containing clays (0, 10, 20 and 30% clay replacement), with the heat flow on the left and the total heat on the right. All the values were normalised by the total amount of initial dry powder in the sample. The black curves always show the reference (100% cement), while the coloured curves show the blended mixes. Figure 18 shows the heat flow of the same samples, but this time normalized to the amount of cement in the sample.
Only a selection of clays has been investigated, with 2 sedimentary clays and 2 glacial clays: Vallåkra, Eriksdal, Lilla Edet and Göteborg.
All the heat flow curves show the 2 main hydration peaks, i.e. the silicate (C3S, C2S) peak followed by the aluminate (C3A) peak. The replacement of a part of the cement by clay lead to an acceleration of the hydration reaction (curves shifted to the left). The effect is however less pronounced with the clay coming from Göteborg. In presence of calcinated clay, a reduction of the intensity of the peaks can also be noticed, which is mainly due to the replacement of cement by the clay. However, this decrease in intensity is less pronounced for the Vallåkra and Eriksdal binders, indicating a better reactivity of those clays which is in accordance with the results of the pozzolanicity tests.
The evolution of the cumulative heat development with time (Figure 17, right side) gives an important information: the higher the total heat is, the higher strength will be achieved. With that in mind, the clay from Vallåkra exhibit the more promising results: it has the lesser decrease in total heat after 7 days. The heat reduction is about 11% with 30% of cement replacement for Vallåkra, when comparing with the 100% cement sample. The reduction increases to approx. 13% for Eriksdal, 19% for Lila Edet, and 16% for Göteborg clays. These results agree with the results of the R3 test (Figure 15).
The Figure 18 presents the heat flow of the same samples but relative the amount of cement in the samples, to assess the changes in the cement hydration. The curves of the samples containing clay are now above the reference (100% cement), indicating more hydration. This can be explained by different effects : 1) the fine clay particles can act as nucleation sites for C-S-H (leading to more heat) ; 2) the “dilution effect”, i.e. the fact that more water is available for hydration of the cement if a portion a cement is replaced by a pozzolanic material, and 3) the normalisation by the amount of cement.
Figure 17. Investigation of the hydration of cement pastes blended with clays, with a cement replacement of 10, 20 or 30%. Left: heat flow during the first 24h of hydration. Right: cumulative heat development over 7 days after mixing. All values are normalized by the total amount of powder (cement + clay). All clays were activated at 800°C.
0 1 2 3 4 0 24 He at fl ow (m W /g so lid ) Time (h) Heat flow Vallåkra 10 Vallåkra 20 Vallåkra 30 CEM I 0 50 100 150 200 250 300 350 400 0 24 48 72 96 120 144 168 He at (m J/ g so lid ) Time (h) Heat A 0 1 2 3 4 0 24 He at fl ow (m W /g so lid ) Time (h) Heat flow Eriksdal 10 Eriksdal 20 Eriksdal 30 CEM I 0 50 100 150 200 250 300 350 400 0 24 48 72 96 120 144 168 He at (m J/ g so lid ) Time (h) Heat B 0 1 2 3 4 0 24 H ea t flow (mW/ g ceme n t) Time (h) Heat flow Lila Edet 30 CEM I C 0 50 100 150 200 250 300 350 400 0 24 48 72 96 120 144 168 He at (m J/ g cem en t) Time (h) Heat 0 1 2 3 4 0 24 He at fl ow (m W /g so lid ) Time (h) Heat flow Göteborg 10 Göteborg 20 Göteborg 30 CEM I 0 50 100 150 200 250 300 350 400 0 24 48 72 96 120 144 168 He at (m J/ g so lid ) Time (h) Heat D
Figure 18. Heat flow during the first 24h of hydration of cement pastes blended with clays, with a cement replacement of 10, 20 or 30%. The values are normalized by the amount of cement in the sample. All clays were activated at 800°C.
XRD
After the calorimetry measurement, the samples were recovered, and the hydration was stopped with the protocol explained in section 4.4 prior to XRD analysis with an internal standard. The results are compiled in Figure 19. This technique being expensive and time consuming, only the samples containing Vallåkra and Eriksdal were analysed this way. XRD results without internal standard are available but only presented in the appendices.
As shown, a decrease of the portlandite content with the substitution rate is observed in Figure 19 for both clays. this can be related to 1) the decrease of the amount of cement and therefore a decrease of the amount of formed hydrates and/or 2) the pozzolanic reaction induced by the presence of clay. It is however difficult to conclude since we do not have samples containing filler (e.g. quartz) to mimic the decrease of cement without adding other reactive material.
0 1 2 3 4 0 24 H ea t fl o w (m W /g c emen t) Time (h) Vallåkra Vallåkra 10 Vallåkra 20 Vallåkra 30 CEM I 0 1 2 3 4 0 24 H ea t fl o w (mW /g c eme n t) Time (h) Eriksdal Eriksdal 10 Eriksdal 20 Eriksdal 30 CEM I 0 1 2 3 4 0 24 H ea t fl o w (mW /g c eme n t) Time (h) Göteborg Göteborg 10 Göteborg 20 Göteborg 30 CEM I 0 1 2 3 4 0 24 H ea t fl o w (mW /g c eme n t) Time (h) Lilla Edet Lilla Edet 30 CEM I
It is important to notice that Figure 19 also shows an increase of aluminate phases (hemicarboaluminate and monocarboaluminate) with the substitution rate, meaning that the Al present in clay are taking part in hydration reactions. No conclusions can be done about ettringite content, considering the small variation in its intensity, coupled with the influence of a nearby peak when clay is present (overlapping effect).
Figure 19. XRD analysis of paste samples containing clay (0, 10, 20 and 30% cement substitution), after 7 days of hydration., with zoom on specific areas. The results concerning Eriksdal are on the left, and on the right for Vallåkra. An internal standard was used to allow the comparison of peak intensities. The main phases are indicated above the peaks: Ett: ettringite ; Qz: quartz ; CH: portlandite ; CC: calcium carbonate ; Hc: hemicarboaluminate ; Mc: monocarboaluminate ; Fe: ferrite ; *: internal standard.
5 10 15 20 25 30 35 40 45 Angle (2θ) CuKα CEM I Eriskdal 10 % Eriskdal 20 % Eriskdal 30 % CH CH Qz CC Qz Ett Ett CH * * *
