Master thesis, 30 hp M.Sc. in Energy Engineering, 300 hp
Department of Applied Physics and Electronics, Spring term 2021
Separate Calcination in
Cement Clinker Production
A laboratory scale study on how an electrified
separate calcination step affects the phase
composition of cement clinker
Separate Calcination in
Cement Clinker Production
A laboratory scale study on how an electrified
separate calcination step affects the phase
composition of cement clinker
Amanda Vikström Amanda.vikstrom@gmail.com
Master Thesis in Energy Engineering Department of Applied Physics and Electronics
Umeå University
Examiner: Robert Eklund
Supervisors: Matias Eriksson, José Aguirre Castillo and Bodil Wilhelmsson
Performed in collaboration with Cementa AB and the Centre for Sustainable Cement and Quicklime Production at Umeå University
Abstract
Cement production is responsible for around 7% of the global anthropogenic carbon dioxide emissions. More than half of these emissions are due to the unavoidable release of carbon dioxide upon thermal decomposition of the main raw material limestone. Many different options for carbon capture are currently being investigated to lower emissions, and one potential route to facilitate carbon capture could be the implementation of an electrified separate calcination step. However, potential effects on the phase composition of cement clinker need to be investigated, which is the aim of the present study. Phases of special interest are alite, belite, aluminate, ferrite, calcite, and lime.
Acknowledgements
Many people have helped and supported me throughout the work of this master thesis; I hope you all know how grateful I am.
This master thesis was done in collaboration with Cementa AB and the Centre for Sustainable Cement and Quicklime Production at Umeå University. I would like to thank my industrial supervisors José Aguirre Castillo and Bodil Wilhelmsson for providing me with an interesting project, as well as your thoughts and advice throughout the process. A special thanks to José who helped me perform the experiments at Cementa’s plant in Slite, as well as the time-consuming analysis of the XRD data. José performed the whole analysis of one of the experiments, and supervised my analysis of the other three experiments. I am very grateful for his commitment to this project and the warm welcome I received during my stay in Slite.
A big thanks to my university supervisor Matias Eriksson, who has always been available to answer my questions and give meaningful advice. Matias has managed to find the perfect balance between challenging me in my academic work, while still providing guidance when I needed it. His support with my writing has been especially important, without him this thesis would have ended up three times as long without ever getting to the point.
I would also like to thank my steering group consisting of Markus Broström, Markus Carlborg, and Tina Hjellström, who have been giving their feedback and thoughts on my work throughout the semester. Especially Markus Carlborg who has spent endless hours trying to teach me XRD analysis prior to the start of my master thesis. His patience and never-failing support has been invaluable.
I would also like to acknowledge the Thermochemical Energy Conversion Laboratory, TEC-lab, at Umeå University for sharing research infrastructure, enabling me to carry out a large part of my work from Umeå. I would also like to express my gratitude to everyone working at TEC-lab for welcoming me into your group and making my office days both educational and tremendous fun. A special thanks to Nils Skoglund and Charlie Ma who introduced me to the exciting field of thermal energy conversion 1.5 years ago, and inspired me to continue my engineering studies when I was hopelessly unmotivated. Without them, I would probably never even have made it to starting my master thesis.
Contents
1 Abbreviations ... 1
1.1 Cement chemical nomenclature ... 1
1.2 Experiment specific abbreviations ... 1
1.3 Other ... 1
2 Introduction ... 2
3 Background ... 3
3.1 The cement production process... 3
3.1.1 Raw materials and pre-treatment ... 4
3.1.2 Preheating and calcination ... 4
3.1.3 Cement kiln... 4
3.1.4 Cooling and post-treatment ... 5
3.2 The main phases in cement clinker ... 5
3.2.1 Alite ... 5
3.2.2 Belite ... 6
3.2.3 Aluminate ... 6
3.2.4 Ferrite ... 6
3.3 Calcination in cement production ... 6
3.4 Clinker phase chemistry ... 7
3.5 Quantitative analysis of the phases in cement clinker ... 8
4 Experimental ... 8
4.1 Sample preparation ... 8
4.2 In situ HT-XRD experiments... 9
4.3 Qualitative and quantitative phase analysis ... 11
5 Results and discussion ... 12
5.1 Temperature determination ... 12
5.2 Clinker formation ... 12
5.3 Influence of separate calcination ... 19
5.3.1 Temperatures up to calcination ... 19
5.3.1 Temperatures from calcination to clinkering ... 19
5.3.2 Cooled clinker ... 22
5.3.3 Effects on the process... 23
5.4 Sources of error ... 23
6 Conclusions ... 24
7 Future work ... 24
1
1 Abbreviations
1.1 Cement chemical nomenclature
C = CaO S = SiO2
A = Al2O3
F = Fe2O3
C3S = (CaO)3∙SiO2 = clinker mineral alite
C2S = (CaO)2∙SiO2 = clinker mineral belite
C3A = (CaO)3∙Al2O3 = clinker mineral aluminate
C4AF = (CaO)4∙Al2O3∙Fe2O3 = clinker mineral aluminoferrite
1.2 Experiment specific abbreviations
IRM = industrial raw meal sampled at Cementa
RM1 = synthetic raw meal 1, limestone calcined prior to raw meal preparation RM2 = synthetic raw meal 2, calcined post raw meal preparation
REF = reference experiments
SEP = experiments with a separate calcination step
1.3 Other
LSF = lime saturation factor SM = silica modulus
AM = alumina modulus XRD = X-ray diffraction
HT-XRD = high-temperature X-ray diffraction XRF = X-ray fluorescence
2
2 Introduction
Concrete is one of the most used manufactured products in the world and is extensively used in the construction industry. Concrete is both affordable, strong, and durable, and there is currently no other material available in the necessary quantities to meet these demands. Cement is the binder in concrete, and therefore rudimental for the concrete manufacture and supply, and modern life as we know it [1].
The cement industry is the world’s third-largest industrial energy consumer and is responsible for around 7% of the global anthropogenic carbon dioxide emissions [1-3]. In Sweden, the cement industry is responsible for 5% of the national emissions [4], and in 2019 Cementa, Sweden’s sole cement manufacturer, emitted 2.0 Mt of carbon dioxide. Of these emissions, 40% were from the combustion of fuels and 60% were process emissions [5], a partitioning similar to production at other plants in Europe [6]. The high process emissions are due to the thermal decomposition of the main component in the raw material calcium carbonate (𝐶𝑎𝐶𝑂3⇌ 𝐶𝑎𝑂 + 𝐶𝑂2) upon heating [7].
There is continuous work within the cement industry to lower carbon dioxide emissions, but due to the high process emissions switching to renewable fuels or improving energy efficiency alone will not decrease emissions sufficiently. There are several initiatives in place to investigate opportunities for carbon capture and storage (CCS), as well as carbon capture and utilization (CCU) [8-14]. Electrification in combination with carbon capture is one option described in CemZero [14], a research programme by Cementa, Vattenfall and Umeå University. By electrifying cement production, either as a whole or in parts, both the demand for fuels and emission of carbon dioxide can be reduced. In addition, the concentration of carbon dioxide in the flue gas would increase [15], facilitating the capture of carbon dioxide. The potential to electrify the process as a sustainability measure is facilitated by Sweden’s nearly fossil-free electricity supply, which is expected to be completely fossil-free in the future [14].
The thermal decomposition of the raw material, called calcination, is a central chemical reaction in the production of cement clinker, the main constituent of cement. Calcination is responsible for about half of the heat input into the kiln system and 60% of the carbon dioxide emissions. Studies have shown that the use of alternative raw material sources are likely to be limited and hence limestone is thought to continue being the most common raw material in the future [14]. Efficient CCS or CCU has the potential to lower carbon dioxide emissions while still using limestone as the primary raw material. Today cement clinker is produced in one continuous process, illustrated by Figure 1 a), but a process where the calcination is separated from the clinkering would enable many new options, illustrated by Figure 1 b) and c).
3
Figure 1: Simplified schematic of three different process setups; a) conventional setup where the combustion of fuels introduces N2 into the gas stream rendering the flue gas unsuitable for direct storage or utilisation; b) and c) a two-step process setup where
the calcination step is electrified, producing a highly concentrated CO2 gas suitable for direct storage and utilization. In b)
limestone is calcined with sand and in c) limestone is calcined without sand
3 Background
3.1 The cement production process
Portland cement is the most common cement type produced today [16]. Cement production consists of several process steps, an overview of the process is shown in Figure 2. The exact process can vary between different plants, the schematic shown being representative of Cementa’s plant in Slite [17].
4
3.1.1 Raw materials and pre-treatment
The raw materials used for cement production consist of 80 wt.-% calcium carbonate, the rest being mainly oxides of silicon, aluminium, and iron. In addition, minor compounds will be included as well, depending on the raw material. Raw material choice is largely driven by what is closely available to the plant, as long transportation of the large volumes required would cost too much [18, 19]. The calcium carbonate usually originates from limestone, an abundant rock that is largely available all over the world [20]. Other common raw materials are chalk, marl, clay, and shale, among others [18].
The first step of the process is to extract the raw materials from the quarry. The raw material is then transported to the cement plant where it is crushed, homogenized, ground in a raw meal mill, and stored in silos. The recipe of the raw meal is calculated using the standard modules of the cement industry; lime saturation factor (LSF), silica modulus (SM), and alumina modulus (AM). LSF is the ratio between limestone and other recipe components, usually expressed as wt.-%.
𝐿𝑆𝐹 = 100 ∙ 𝐶𝑎𝑂/(2.8 ∙ 𝑆𝑖𝑂2+ 1.18 ∙ 𝐴𝑙2𝑂3+ 0.65 ∙ 𝐹𝑒2𝑂3)
If the cement clinker has an LSF value of 100, the proportion of the principal strength-giving calcium silicate alite is maximised. Industrial clinker usually has an LSF value ranging between 94-98 wt.-%. SM and AM influence both the cement clinker quality and energy consumption of the kiln [7].
𝑆𝑀 = 𝑆𝑖𝑂2/(𝐴𝑙2𝑂3+ 𝐹𝑒2𝑂3)
𝐴𝑀 = 𝐴𝑙2𝑂3/𝐹𝑒2𝑂3
3.1.2 Preheating and calcination
The powdered raw meal, with a size of approximately 10-15% 90 µm and 2% 200 µm sieve residue, is fed into the preheating tower. The preheater consists of a series of cyclones through which the raw meal is heated with hot flue gases from the kiln in counter-current flow. The raw meal starts to calcine as it moves down the preheating tower. Calcination of the main raw material calcium carbonate and can be described as
𝐶𝑎𝐶𝑂3(𝑠) ⇌ 𝐶𝑎𝑂 (𝑠) + 𝐶𝑂2 (𝑔) 𝛥𝐻 = +177.8 𝑘𝐽/𝑚𝑜𝑙
At the bottom of the preheating tower, a combustion chamber called the calciner is located, after which about 95% of the raw meal is calcined. The temperature of the calciner is around 900°C, a higher temperature would create problems in the combustion chamber such as sticking, and therefore the rest of the calcination takes place in the rotary kiln. During calcination calcium carbonate loses 44% of its mass, resulting in a large specific area available for reaction. Some calcium oxide will react with silicate minerals in the calciner, especially when the raw meal particles contain both lime and silicates that are in direct contact, resulting in the formation of belite as well as some aluminate and ferrite phases. Unreacted calcium oxide is denoted free lime [7, 17].
3.1.3 Cement kiln
5
Figure 3: Schematic of a precalciner kiln system showing the locations of the principal chemical reactions occurring. DOC: degree of calcination. Adapted from [7].
3.1.4 Cooling and post-treatment
The main clinker mineral alite is only thermodynamically stable over 1250°C, and hence the hot cement clinker leaving the kiln is rapidly cooled to avoid the breakdown of alite into belite and free lime. The cooling is also characterized by the crystallization of aluminates and ferrites. The cooling is done with large quantities of air, and the heated air can then be re-used in the process [7, 17]. After the clinker is cooled it is ground in a cement mill to a fine powder, and during grinding additives such as gypsum are added [18]. The composition of the final cement varies depending on raw material, production process, and fuel type, but Portland cement typically consists of 50-70% alite, 15-30% belite, 5-10% aluminate, and 5-15% ferrite. However, minor phases are usually also present in low amounts [21].
3.2 The main phases in cement clinker
The four main phases in cement clinker are alite, belite, aluminate, and ferrite. Other minor phases such as alkali sulphates, periclase, and free lime are normally present as well. Alite is the main strength-giving mineral and as such needs to be present in large enough amounts. The strength of the clinker phases is related to how they react with water, as it is the hydrated phases that provide strength in concrete [18].
3.2.1 Alite
Alite is the nonstoichiometric form of C3S. Alite present in cement clinker normally contains 3-4%
substituent ions, typically Mg2+, Al3+, and Fe3+ [21]. Alite is formed at high temperatures as the
high-temperature polymorph R, and can undergo reversible polymorphic transformations upon cooling and heating, described as [22] 𝑅1060°𝐶↔ 𝑀3 1050°𝐶 ↔ 𝑀2 990°𝐶 ↔ 𝑀1 980°𝐶 ↔ 𝑇3 920°𝐶 ↔ 𝑇2 600°𝐶 ↔ 𝑇1
The letters express the formal crystallographic symmetry of the polymorph were T=triclinic, M=monoclinic, and R=rhombohedral, and the numbers variations within the same symmetry [23]. Which polymorphic phases that are present is highly dependent on both impurities and the heating and cooling history [24], in industrial cement clinker it is usually M1 and M3 that are present at ambient
6
3.2.2 Belite
Belite is the nonstoichiometric form of C2S. Belite normally contains 4-6% substituent ions, mainly Al3+
and Fe3+ [21]. Belite has several polymorphs [27], the transformation sequence being
The low-temperature polymorph γ-C2S can cause a problematic phenomenon called dusting due to it
being less dense than the other polymorphs. Luckily, the most common polymorph in Portland cement, β-C2S, is normally stabilized at lower temperatures, either by quenching or the formation of a solid
solution containing a large amount of impurities [23, 28]. Belite reacts much slower with water than alite, however, it substantially contributes to the later strength development in concrete [21].
3.2.3 Aluminate
Aluminate, or tricalcium aluminate, is the nonstoichiometric form of C3A and the most reactive of the
four clinker phases. Stoichiometric C3A does not have any polymorphs, whereas through substitutions
many different polymorphs have been discovered. Ca2+ is typically substituted by Mg2+, 2 K+, 2 Na+, but
also the aluminium ion is commonly substituted. Cement clinkers most often contain cubic or orthorhombic forms of aluminate, with up to 20% substituent ions [21, 26, 29].
3.2.4 Ferrite
Ferrite, or calcium aluminoferrite, is the nonstoichiometric form of C4AF. Ferrite present in cement
clinker differs notably from the pure form and contains up to 10% of substituent ions, mostly Mg2+. The
typical composition of ferrite present in cement clinker can be approximated as Ca2AlFe0.6Mg0.2Si0.15Ti0.05O5 [21].
3.3 Calcination in cement production
The decomposition of carbonates is a highly endothermic reaction [30]. The calcination of a limestone particle involves several steps, and the rate determining step depends on the calcination conditions. These steps can be summarized as i) heat transfer from the bulk gas to the external surface of the particle and from the external surface to the reaction interface; ii) chemical reaction at the reaction interface; and iii) mass transfer of carbon dioxide from the reaction interface to the bulk gas [31]. If the particles are small and the bulk gas temperature is high, the rate of heat and mass transfer will however generally be high. For the conditions in cement production, the rate-determining step is the chemical reaction [30, 32].
7
Figure 4: Calcination temperature of calcium carbonate as a function of the partial pressure of carbon dioxide at 1 atm. Calculated with FactSage 8.0 software [39].
3.4 Clinker phase chemistry
High-temperature chemistry in cement clinker production mainly involves oxides of calcium, silicon, aluminium, and iron. A simplified schematic of the principal clinker formation reactions in the kiln system at different temperatures can be seen in Figure 5. Clinker formation in rotary kilns is a very complex process due to the many parameters affecting the outcome, such as raw meal composition, kiln dimensions, temperature profile, and gas composition. Vastly simplified cement clinker formation can be described in three stages; i) belite, aluminate, and ferrite formation, ii) melt formation, and iii) alite formation [40].
Figure 5: Approximate proportions of the principal phases present in the kiln system at different temperatures. Cr: cristobalite. Adapted from [7].
8
some extent, belite, leading to solid-liquid and liquid-liquid reactions starting to occur [26]. In industrial conditions, a salt melt will in addition form at lower temperatures [41]. Free lime dissolves partly into the oxide melt and hence the rate of the reaction between lime and silica is strongly improved by the mobility provided by the liquid phase [7]. During this stage, the hot raw meal starts to agglomerate into nodules. In addition, free lime and belite start to react forming alite. The process of alite formation is very complex as it can proceed via solid-solid, solid-liquid, and liquid-liquid reactions. Since the formation of alite is believed to be rate controlled by the diffusion of calcium ions the solid-solid reaction can however be neglected due to slow diffusion through solids [26].
3.5 Quantitative analysis of the phases in cement clinker
X-ray diffraction (XRD) in combination with Rietveld refinement has emerged as a reliable and precise method of quantitative analysis of the crystalline phases with long range order present in cement clinker [42-48]. The Rietveld method uses least squares refinement to fit models of the crystalline structures to the observed diffraction pattern, varying the parameters of the structures and minimizing the difference between observed and calculated patterns [49-51]. The Rietveld method can only be used for quantitative analysis, and prior qualitative analysis is required. Rietveld analysis normalizes the total of the determined crystalline phases to 100% [52], and hence if unidentified, amorphous, or phases with short range order are present the estimated amounts of the crystalline phases will differ from the actual amounts. Amorphous material does not give diffraction peaks as crystalline material does and is only visible in the diffractogram as a broad hump or an elevated background, but can be quantified by methods such as spiking [53, 54]. XRD is often complemented with elemental analysis by X-ray fluorescence (XRF) [43].
If a high-temperature chamber is used in combination with XRD, so-called high-temperature X-ray diffraction (HT-XRD), one can follow the phase transformations and clinker reactions in situ [55-58]. By combining HT-XRD and Rietveld refinement a phase analysis can be obtained as a function of temperature.
4 Experimental
4.1 Sample preparation
Three different raw meals were used for the experiments, one industrial raw meal (IRM) and two synthetic raw meals prepared from limestone and laboratory grade chemicals. The IRM was sampled at Cementa’s plant in Slite at an air chute before the elevator to the top cyclones. The two other raw meals, denoted RM1 and RM2, were mixed using limestone from Tromsdalen (55 wt.-% CaO - Miljøkalk) and laboratory-grade SiO2 (99.7 wt.-% - Merck), Al2O3 (99.4 wt.-% - Honeywell), Fe2O3 (97 wt.-% - VWR),
as well as industrial gypsum (95.8 wt.-% CaSO4), the recipes and goal clinker modules for RM1 and RM2
can be seen in Table 1. For raw meal RM2, a 1-2 mm fraction of the limestone was calcined by heating to 1000°C at 10°C/min in a CO2 atmosphere (1 l CO2/min) using a muffle furnace. The temperature was
kept at 1000°C for 20 min before cooling with 10 l air/min to ambient temperature.
Table 1: Recipes and goal clinker modules used for preparing raw meals RM1 and RM2. Clinker modules calculated according to section 3.1.1.
Raw meal components (wt.-%) Goal clinker modules
Limestone Calcined
limestone SiO2 Al2O3 Fe2O3 Gypsum LSF SM AM
RM1 80.51 - 13.98 3.32 2.13 0.06 100 2.6 1.6
9
The components used for raw meals RM1 and RM2, as well as raw meal IRM, were dried at 105°C prior to weighing, except the calcined limestone used for RM2 to avoid carbonation. All components were then milled individually at 1000 rpm for 1.5 min using a Retsch RS 200 vibratory disc mill before mixing the raw meals. All raw meals were then milled and homogenized at 700 rpm for 1 min and at 1000 rpm for 1.5 min. The size distribution of the raw meals can be seen in Figure 6, measured by laser diffraction with a Panalytical Mastersizer 2000.
Figure 6: Size distribution of the raw meals obtained by laser diffraction.
The elemental composition of the raw meals expressed as oxides, as well as the loss on ignition (LOI), can be seen in Table 2. The elemental composition was determined by XRF using a Panalytical Axios spectrometer. The LOI was determined by heating the sample to 950°C and calculating the weight loss. Clinker modules calculated from the XRF results can also be seen in Table 2.
Table 2: Chemical composition, LOI, and clinker modules of the raw meals. Composition obtained by XRF analysis, LOI by thermogravimetric method, and clinker modules calculated according to section 3.1.1. XRF results normalised to 100% Ca, Si, Al, Fe, Mg, S, K, Na, and O.
Raw meal composition (wt.-%) Clinker modules
LOI CaO SiO2 Al2O3 Fe2O3 MgO SO3 K2O Na2O LSF SM AM
IRM 33.69 42.82 13.51 3.42 2.01 2.40 1.06 0.90 0.19 99.17 2.48 1.70
RM1 35.39 44.40 13.99 3.66 2.17 0.28 <0.01 0.07 0.03 98.91 2.40 1.69
RM2 1.40 67.21 21.64 5.52 3.29 0.43 <0.01 0.11 0.04 97.07 2.46 1.67
If the chemical composition of the raw meal is known an estimated phase composition of the cement clinker can be calculated according to Bogue [59]. In addition, an estimation of the amount of melt formed at different temperatures can be calculated according to the work of Lea et al. [60]. The estimated clinker phase composition and melt formation for the raw meals can be seen in Table 3.
Table 3: Estimated phase composition of the cement clinker calculated according to Bogue [59] and estimated melt formation at 1338°C, 1400°C and 1450°C according to Lea et al. [60] for the three raw meals IRM, RM1, and RM2.
Main clinker minerals (wt.-%) Melt formation (wt.-%)
C3S C2S C3A C4AF 1338°C 1400°C 1450°C
IRM 68.34 6.34 8.46 9.15 21.96 25.32 25.72
RM1 71.73 7.44 9.25 10.14 23.56 24.49 24.94
RM2 68.19 11.41 9.17 10.15 20.92 24.78 25.23
4.2 In situ HT-XRD experiments
In situ HT-XRD measurements were performed using a Panalytical X´Pert diffractometer, in θ-θ
(HT-10
chamber). The diffractometer was equipped with Cu-Kα radiation and a PIXcel1D detector. Data was
collected in an angular range of 8-65° (2θ), with a 0.013° step size and a counting time of 0.13 s/step.
A total of four experiments were performed. Two of the experiments, REF-IRM and REF-RM1, were performed to obtain reference cases (REF) and cement clinker was produced by the conventional method of calcining and heating to clinker temperature in one continuous process. In the other two experiments, SEP-RM1 and SEP-RM2, the samples were calcined in a process separated from the clinkerization (SEP). For SEP-RM1 the calcination was performed in the HT-chamber, whereas the calcination for SEP-RM2 was performed in a muffle furnace as described in section 4.1.
For each experiment, the raw meal was mixed with ethanol and Zapon lacquer to form a paste that was spread thinly on the platinum plate of the HT-chamber. All experiments were initiated with a scan of the sample at 25°C. REF-IRM was heated and cooled according to Figure 7a) to a maximum temperature of 1450°C, REF-RM1 and SEP-RM2 according to Figure 7b) to a maximum temperature of 1500°C, and SEP-RM1 according to Figure 7c) to a maximum temperature of 1000°C for the calcination and 1500°C for the clinkerization. RM1 and RM2 contained lower amounts of minor components compared to IRM (see Table 2), causing them to require a higher clinkering temperature [30]. The scanning temperatures for all experiments are marked in Figure 7 with triangles. The valve of the HT-chamber was left open for all experiments except the calcination of SEP-RM1 (Figure 7c). When the valve was open the gas atmosphere of the HT-chamber was 100% air, whereas when the valve was closed the gas atmosphere was 100% air at the start of the experiment, but as the calcination started and CO2 was emitted the gas
atmosphere consisted of a high percentage of CO2, and also got slightly pressurized. After the calcination
and cooling of SEP-RM1 (Figure 7c) the valve was opened immediately after the sample had cooled to avoid carbonation, and left open for the clinkerization stage.
Figure 7: The heating and cooling of the samples over time in the high-temperature chamber of the XRD. a) was used for REF-IRM, b) was used for REF-RM1 and SEP-RM2, and c) was used for SEP-RM1. Scanning temperatures marked with
11
4.3 Qualitative and quantitative phase analysis
Collected HT-XRD scans from experiments REF-RM1, SEP-RM1, and SEP-RM2 were analysed by the author, whereas experiment REF-IRM was analysed at Cementa Research [61]. Qualitative phase analysis was performed using EVA 5.1 software [62], and quantitative phase analysis was performed by the Rietveld method using Topas-Academic 4.2 software [63]. For the Rietveld refinement, the background was fitted with a Chebyshev function with five terms. The analysis was computed for the 10-65° 2θ range, and for the samples where the platinum plate of the HT-chamber gave clear diffraction peaks the platinum was normalized out of the final phase composition. The use of preferred orientation was kept to a minimum in the analysis due to the temperature difference between the start and finish of the scans collected during heating. All the structural models of the phases used for the Rietveld refinement are shown in Table 4. When required, thermal expansion data [64-69] was used to verify the changes in the cell parameters at high temperatures for the phases lacking high temperature structure files.
Table 4: Structural models of phases used for the Rietveld refinement of the in situ HT-XRD experiments. Phases marked with * were supplied by Bruker [70], and are not available on ICSD [71].
Phase Temperature (°C) Reference ICSD #
C3S R 1200 Nishi et al. [72] 30889 M3 25 De la Torre et al. [73] 94742 M1 25 Noirfontaine et al. [74] * C2S α 1545 Mumme et al. [75] 82999 αH 1250 Mumme et al. [75] 82997 αL 1060 Mumme et al. [75] 82996 β 25 Mumme et al. [76] 81096
C3A Cubic 25 Mondal et al. [77] 1841
Orthorhombic 25 Takeuchi et al. [78] 100220
C4AF 25 Colville et al. [79] 9197
Calcite 25 Markgraf et al. [80] 40107
200 Markgraf et al. [80] 40112
400 Markgraf et al. [80] 40113
600 Markgraf et al. [80] 40114
750 Markgraf et al. [80] 40115
800 Markgraf et al. [80] 40116
Corundum 27 Ishizawa et al. [81] 10425
Hematite 25 Antipin et al. [82] 22505
Quartz α 25 Kihara, K. [83] 89276 125 Kihara, K. [83] 89277 225 Kihara, K. [83] 89278 324 Kihara, K. [83] 89279 424 Kihara, K. [83] 89280 540 Kihara, K. [83] 89282 565 Kihara, K. [83] 89283 β 581 Kihara, K. [83] 89285
Lime 25 Smith et al. [84] 52783
Portlandite 25 Chaix-Pluchery et al. [85] 202220
110 Chaix-Pluchery et al. [85] 202223
Cristobalite 310 Peacor, D. R. [86] 34926
Periclase 25 Schmahl et al. [87] 61325
Platinum 25 Swanson et al. [88] 76153
12
5 Results and discussion
5.1 Temperature determination
The temperature of the system can be verified by inspecting how the unit-cell volume of CaO changes over temperature. CaO is especially suitable for this since it does not accept large amounts of dopants that could otherwise affect the unit-cell. The unit-cell volume of CaO as a function of temperature is shown in Figure 8 for all four experiments. It is important to keep in mind that this temperature determination is only approximate since the scans were performed simultaneously as the temperature was increased, and hence the temperature varied between the start and finish of the scan. It is however useful to give an indication of the obtained temperature and to detect if any experiment exhibits deviant unit-cell volumes, indicating a problem with the experiment or analysis.
Figure 8: The unit-cell volume of CaO obtained through Rietveld refinement of the collected data plotted against temperature for all four experiments.
A linear relationship between the unit-cell volume and temperature can be given as V(T) = n + mT. From Figure 8 (with the temperature changed to K) the unit cell volume of CaO is given as V(T) = 109.66 + 0.0047 T. The volumetric lattice thermal expansion coefficient is given as αv = m/n = 42.9 × 10-6 K-1. If
the material is assumed to be isotropic the linear expansion coefficient is given as αL = αV/3 = 14.3 × 10 -6 K-1, which agrees well with the reported bulk linear expansion for CaO, 13.6 × 10-6 K-1 [90].
The CaO unit-cell volume of SEP-RM2 and REF-IRM deviates from that of the other experiments at some temperatures, indicating that there has been a problem with either the heating, heat transfer or Rietveld refinement of the collected data. This will be discussed further for REF-IRM in section 5.2 and SEP-RM2 in section 5.3.
5.2 Clinker formation
13
14 Rietveld quantitative phase analysis was performed on all collected scans, and Figure 10 shows a representative example of the fit of a performed Rietveld analysis, with the main peaks labelled. HT-XRD causes the Rietveld refinement to be somewhat complicated by the thermal expansion of the unit-cell, and in addition the clinker minerals accept a large amount of substituent ions that also affect the unit-cell. The Rietveld refinement was further complicated by the experimental method used in the present work, as the temperature was ramped simultaneously as the scans were performed for all temperatures except 25°C, 1450°C and/or 1500°C, and 25°C* (cooled clinker), as well as 1000°C for SEP-RM1. This problem was especially evident at temperatures where rapid reactions took place, such as during calcination, as phases present when the scan started might have disappeared at the end of the scan, or new phases formed. Nonetheless, an overall satisfying fit of the Rietveld refinements were obtained. Figure 11 shows the amount of the major phases calcite, lime, C2S, C3A, C4AF,
and C3S, as a function of temperature for all experiments.
Table 5 shows the phase composition at 25°C, clinkering temperature (1450°C/1500°C) and 25°C*. A table with the complete analysis can be found in Appendix 1.
At low temperatures CaCO3 constituted a large part of the
sample. Calcite is the only thermodynamically stable CaCO3 polymorph at ambient conditions [91], and was the
only one detected in the current study. REF-IRM contained a high amount of clay minerals, as would be expected from an industrial raw meal [30]. For REF-RM1 only SiO2, Al2O3, and Fe2O3 were observed in addition to
calcite, as expected from the raw meal preparation.
At temperatures above 700°C rapid calcination started for both reference cases, with formation of a high amount of free lime. During calcination in industrial conditions the raw meal is intimately mixed, causing much CO2 to be lost
before any free lime can be detected due to the formation of belite and other early phases [30]. This was not observed in the present experiments, most likely due to the lack of mixing. Another possible explanation is the poor time resolution, as scans were performed every 100°C upon heating. REF-IRM contained dolomite which should calcine at lower temperatures than calcite [30], but due to the poor time resolution and low amounts of dolomite, this was not observed in the present study.
Simultaneously as free lime was observed, the formation of C2S, C3A and C4AF could be observed. For the synthetic
raw meals, RM1 and RM2, only cubic C3A could be
observed, whereas for the industrial raw meal IRM a mixture of cubic and orthorhombic C3A was observed. The
15
formation of orthorhombic aluminate can only occur if sufficient alkali is available [21], and since the alkali content of RM1 and RM2 was low, the absence of orthorhombic C3A was expected.
The amounts of both C3A and C4AF determined by Rietveld refinement was lower than would be
expected according to the Bogue calculations (Table 3) for all experiments. Phase composition according to Bogue can only be seen as approximate, and it is not unusual for the actual phase composition to differ from the theoretical Bogue composition [30]. However, others have obtained higher amounts of C3A and C4AF from HT-XRD studies [55, 58] more in accordance with Bogue, and
hence the low amounts in the present study are unexpected. No clear explanation has been suggested, but is most likely due to some issue with the experimental setup, as it is consistent for all four experiments. It should be noted that the phase composition obtained by Rietveld refinement appears not to account for all the Al2O3 and Fe2O3 present in the raw meal according to the XRF analysis, and
hence some aluminate and ferrite phases might be lacking from the analysis. In theory, many aluminate and ferrite phases other than C3A and C4AF could be formed [92], but none have been
16
Figure 11: Amount of a) calcite, b) lime, c) C2S, d) C3A, e) C4AF, and f) C3S obtained by Rietveld refinement of collected
HT-XRD data as a function of temperature for all four experiments. Data normalised to exclude Pt, as well as the pure Al and Fe in REF-IRM.
Upon heating belite polymorphs αL, αH and α were observed. For the synthetic raw meals RM1 and RM2,
αL was the first polymorph to be observed, and at higher temperatures αH, and later α, were formed,
which is consistent with observed belite formation in other HT-XRD studies [58]. For REF-IRM the belite polymorphic transitions overlapped more, and αH was the first polymorph to be observed. The αL
and αH belite polymorphs are similar in structure, but can be distinguished by αL having weak additional
reflections at d-spacing values of 3.578, 3.092, 2.711, and 2.177 Å (measured at 1329K) [69]. This was however difficult in the present study due to the relatively high background, and as the amount of αH
observed was so low it is difficult to say with certainty that αH was formed prior to αL in REF-IRM. As
17
Table 5: Phase composition (wt.-%) obtained by Rietveld refinement of isothermal scans at temperatures 25°C, 1450/1500°C, and 25°C* (cooled clinker). In addition, the composition at 25°C** (cooled calcined raw meal) is shown for SEP-RM1.
REF-IRM REF-RM1 SEP-RM1 SEP-RM2
Temperature (°C) 25 1450 25* 25 1500 25* 25 25** 1500 25* 25 1500 25* C3S R 72.7 82.1 61.6 M3 27.8 46.2 57.9 39.6 60.5 M1 7.7 27.3 16.2 39.5 5.2 T3 45.4 C2S α 0.3 6.5 6.5 11.0 αH 3.8 16.9 9.2 20.5 αL 4.5 β 14.4 16.2 7.3 19.1 C3A Cubic 1.1 4.3 1.6 4.6 4.8 Orthorhombic 1.2 C4AF 6.1 2.3 5.1 3.6 3.7 Calcite 75.1 82.8 82.7 6.2 1.0 Corundum 1.2 4.1 4.5 5.4 6.2 Hematite 0.1 3.9 3.8 2.1 5.5 Quartz α 9.5 9.2 8.9 10.1 14.3 β 0.6 Lime 0.9 0.9 3.2 3.9 53.0 1.8 4.7 71.6 5.3 5.2 Portlandite 1.4 Cristobalite 0.2 Periclase 2.5 4.3 0.4 0.2 0.8 Platinum 0.7 0.7 0.2 0.2 0.5 0.4 0.1 0.9 0.6 Dolomite 3.8 Epidote 1.2 Magnetite 0.3 Microcline 3.4 Chlorite 0.4 Åkermanite 2.4 Spinel 0.7 Monticellite 1.3 Mullite 0.7 Aluminium 7.0 11.7 Iron 0.1 0.9
C3A and C4AF could not be detected at temperatures above 1300°C, which was expected as the aluminate
and ferrite phases, and to some extent belite, melt around 1200-1300°C forming an oxide melt [30]. The temperature at which melting begins is affected by the amount of minor components present [30], and thus melting will start at a lower temperature for IRM than RM1 and RM2. This was however difficult to observe with the present time resolution. Shortly before C3A and C4AF could no longer be observed,
C3S could be detected, with the exception of REF-IRM where it was detected just above 800°C. Before
melt formation, C3S is formed through solid state reactions between C2S and lime. Alite formation and
growth is, however, much faster in the presence of the liquid phase [30]. Alite formation is faster still in the presence of a melt of low viscosity, which occurs if the raw meal contains a high amount minor components [30], such as IRM.
At high temperatures, only the high-temperature alite polymorph R could be observed for REF-IRM, SEP-RM1, and SEP-RM2, whereas in REF-IRM M1, M3, and T3 could be observed. As expected [21], alite
was found in the polymorphic forms M1 and M3 in the cooled clinker for all four experiments. The
18
windows in the XRD patterns that can be used to set the two apart [93], the main focus of the present work will be the total alite content and as such no discussion will be made on the partitioning between M1 and M3 in the different experiments.
SiO2 was observed in the polymorphic forms α-quartz, β-quartz and cristobalite. The phase
transformation from α to β-quartz is usually rapid at temperatures above 573°C [30], but in all experiments α-quartz could be detected at much higher temperatures than would be expected. A transformation to cristobalite could be seen around 1300°C for REF-RM1 and SEP-RM1, and around 1450°C for SEP-RM2, whereas for REF-IRM it was not observed at all. Cristobalite normally forms above 1470°C, but admixed with many other materials it can be formed already above 1000°C [30].
Upon cooling, C3A and C4AF crystallised from the melt. For some experiments, the alite content was
higher in the cooled clinker than in the last scan before cooling, this is however most likely due to the poor time resolution, as alite content should not increase upon cooling [30]. Rapid quenching of cement clinker can lead to the formation of glass [21], this was however not observed during any of the experiments despite the high cooling rate. The phase composition of the final clinkers varies from that of the Bogue calculations in Table 3. As mentioned earlier, it is not uncommon for the actual phase composition to differ markedly from that of Bogue, especially in alite being underestimated and belite overestimated. Two explanations for this is cooling under non-equilibrium conditions and assuming stoichiometric clinker minerals and ignoring substitutions [30].
Around 4.5 wt.-% free lime was observed in the cooled clinker for REF-RM1, SEP-RM1, and SEP-RM2, whereas REF-IRM had a free lime content under 1 wt.-%. A high concentration of free lime causes concrete unsoundness and is a waste of both raw material and energy, and is hence undesirable [26]. For raw meals with an LSF less than 100 high free lime is likely due to inadequate mixing or insufficient residence time at clinkering temperature [92], and since REF-IRM had a low free lime it is likely due to insufficient residence time for the raw meals RM1 and RM2 lacking minor components. It is expected that the synthetic raw meals RM1 and RM2 would require a longer residence time as silica present as quartz is less reactive than that in clay minerals, and a pure limestone consisting of mostly calcite is less reactive than a limestone high in silica or silicate minerals [30], such as the one present in IRM. Telschow et al. [40] concluded that longer residence times were required to obtain a clinker phase composition similar to that of industrial clinker at lab-scale, and that a temperature of 1450°C might be too low to simulate the conditions in the rotary kiln. A higher temperature and longer residence time at clinkering temperature were used for REF-RM1, SEP-RM1, and SEP-RM2, but the high amount of free lime indicates that even longer residence times would have been needed for those three experiments to compensate for the low amounts of minor components in raw meals RM1 and RM2.
In REF-IRM, pure iron and aluminium were identified both in the heated sample and in the cooled clinker, as can be seen in Table 5. Formation of pure iron and aluminium is not expected in cement clinker production, and should not form under present conditions. In addition, the amount of aluminium quantified is much higher than the amount of Al2O3 in the raw meal according to the XRF
analysis in Table 2. The IRM was, as mentioned, much more heterogeneous than RM1 and RM2 and as such the analysis was much complicated, and hence it is not unlikely that the identified pure aluminium and iron were in fact some other phases. In addition, REF-IRM showed some deviating unit-cell volumes in Figure 8, not all of them at the same temperatures as pure aluminium and iron were identified, but the deviating unit cell volume at 25°C* is noteworthy. The highest amount of aluminium and iron were observed at 25°C*, and it is well documented that CaO should have a unit-cell volume above 111 Å3 at
25°C [68, 84, 94], whereas REF-IRM had a unit-cell volume below 110 Å3 at the same temperature. This
19
5.3 Influence of separate calcination
The influence of the separate calcination step in SEP-RM1 and SEP-RM2 will be discussed in this section, focusing especially on the phase composition of the cooled clinker. High carbon dioxide atmosphere during calcination is known to not affect the phase chemistry to any large extent [10, 37, 38], and as such all differences between the reference experiments and the separate calcination experiments are assumed to be due to the separate calcination step.
As Rietveld refinement normalizes the content of the identified phases to 100%, it can be difficult to get a clear view of the phase changes from scatter plots, such as in Figure 11, when the process includes weight change of the sample. In Figure 12 area plots of the phase composition as a function of temperature, normalized for the weight change from calcination, carbonation, and melt formation, can be seen for REF-RM1, SEP-RM1, and SEP-RM2. The amount oxide melt is approximated from the amount C3A and C4AF before melting. It is evident from Figure 12 that the approximated amount oxide
melt is much lower than the estimated amount in Table 3, but as the amount C3A and C4AF formed was
much lower than expected and the oxide melt was approximated from these this is not surprising.
5.3.1 Temperatures up to calcination
The calcination in SEP-RM1 starts at the same temperature as in the reference experiments, but takes somewhat longer to reach completion, as can be seen in Figure 12b. The sample is already completely calcined at 1000°C, but upon cooling some of the lime is carbonated due to the high CO2 atmosphere of
the HT-chamber, causing the cooled raw meal to contain 6 wt.-% calcite. Upon reheating of the calcined raw meal (Figure 12c), the calcite content increases further to just above 8 wt.-% calcite. The phase composition is noticeably similar at 1000°C in Figure 12b and c, with 23.3 wt.-% belite at calcination and 24.5 wt.-% upon reheating.
SEP-RM2 contained around 1 wt.-% calcite at 25°C, probably due to carbonation of the calcined limestone during raw meal preparation. An XRD analysis was performed on the calcined limestone directly after calcination to verify complete calcination, but due to the reactivity of CaO it is nevertheless likely that the calcined limestone contained a low amount of CaCO3 at the time of the raw meal
preparation. This is consistent with the XRF results, since the LSF of RM2 is slightly below that of IRM and RM1, as well as the LOI being higher than would be expected, and carbonation of the calcined limestone prior to raw meal would explain both of these factors. SEP-RM2 also contained some portlandite, which is readily formed as lime is in contact with moisture, such as the humidity in atmospheric air. SEP-RM2 was readily carbonated at temperatures below the calcination temperature, in contrast to SEP-RM1. This could, however, be due to the raw meal preparation, as the muffle furnace used to calcine the limestone for RM2 could not heat at the desired rate of 10°C/min at 800-1000°C. This prolonged the heating process, possibly producing a more reactive lime [95].
Calcination is known to be favoured by the intimate mixing with quartz and clay minerals that take place during calcination in cement production [30]. A difference in calcination rate between SEP-RM1 and SEP-RM2 could, however, not be observed in the present study as the limestone in RM2 was calcined in a muffle furnace ex situ. It is, however, given that early products such as belite, as well as some aluminate and ferrite phases that might commonly form already in the calciner [58, 96, 97] cannot be formed during the calcination step of SEP-RM2. This could potentially affect the clinker reactions taking place at higher temperatures.
5.3.1 Temperatures from calcination to clinkering
C2S, C3A and C4AF could be observed at similar temperatures in SEP-RM1 and SEP-RM2 as in the
20
Figure 12: Results from Rietveld refinement normalized for weight change as a function of temperature. ‘Other phases’ consist of all crystalline phases other than C3S, C2S, C3A, C4AF, lime, and calcite, such as SiO2, Al2O3, Fe2O3, MgO, and clay minerals. Oxide
21
Figure 13: HT-XRD scans at 1393°C for a) REF-RM1, b) SEP-RM1, and SEP-RM2. Main peaks labelled.
Belite and alite have been readily formed in REF-RM1 and SEP-RM1 at 1393°C, whereas in SEP-RM2 a much lower amount of belite has been formed and no alite could be detected. The reason for this is unclear, but as discussed in section 5.1 the unit cell of CaO differed markedly for SEP-RM2 compared to the other experiments. This could imply that there is some problem with the mounting of the sample, causing the heat transfer to be impaired. During trial experiments, it was observed that improper mounting could cause the sample to crack and detach from the platinum plate of the HT-chamber. This resulted in an air gap between the platinum plate and the sample, leading to poor heat transfer and no clinker formation. It is therefore interesting that alite is eventually formed in SEP-RM2, but at a higher temperature than in the other experiments.
SEP-22
RM2. Due to the many factors that might have influenced the clinker formation in SEP-RM2, it is too early to dismiss that method of separate calcination in favour of SEP-RM1.
5.3.2 Cooled clinker
In Figure 14 HT-XRD scan from 25°C* is shown for a) REF-RM1, b) SEP-RM1, and c) SEP-RM2, with the main peaks labelled. The amount of alite in the cooled clinker from SEP-RM1 is similar to that of the reference experiments (Figure 11f), with it being somewhat higher than REF-RM1 and somewhat lower than REF-IRM, if REF-IRM it normalised to not contain pure aluminium and iron. As discussed, there are many sources of uncertainty in the data, but the high amount of alite in SEP-RM1 indicates that it might be possible to produce clinker of similar quality with a separate calcination step in lab-scale. It is however important to note that no replicates were performed in the present study, and hence further experiments are needed to verify this.
23
The amount alite in the cooled clinker for SEP-RM2 is instead somewhat lower than both reference experiments. The level of belite is instead higher, which would to some extent be expected from the Bogue calculations in Table 3. In addition RM2 has a lower LSF, which could also explain the lower alite and higher belite level [18]. As the free lime amount is the highest in RM2 it could be that SEP-RM2 needs the longest residence time at clinkering temperatures, which from an economical perspective would be a clear disadvantage compared to SEP-RM1.
There has been previous work by Hild et al. [99] investigating the reaction rate of raw meals containing either CaCO3 or CaO, showing that the calcined CaO featured a larger particle size than the original
CaCO3. The larger particle size increased the diffusion distance, affecting the reaction rate negatively.
However, as the original article is in German, and from 1933, the author has not been able to take part of the results other than from secondary sources [26]. It is, in addition, unclear if these results are applicable to the present study as the calcined limestone for RM2 was milled after calcination, but might be relevant for future work if limestone is milled prior to calcination. In addition, Borgwardt [100] found that CO2 gas accelerates the sintering rate of CaO, which might also affect an electrified separate
calcination step with a high CO2 atmosphere.
5.3.3 Effects on the process
It is evident from the experiments that having a separate calcination step creates the need for adequate storage to avoid carbonation and hydration if the raw meal/calcined limestone is to not immediately enter the rotary kiln. CaO is very reactive, and the risk of carbonation is especially high at temperatures just below calcination temperatures. In addition, if the storage is non-heated the raw meal must be reheated when it enters the kiln, which would increase the required retention time in the kiln if it was not preheated prior to entering the kiln. The kiln retention time of an industrial process is normally 20-40 min [101], and if a longer retention time is required it would affect the production capacity negatively. Heat and mass balance calculations of a separate calcination step has not been within the scope of the present work, but is a crucial part of determining whether a separate calcination step is feasible or not. The feasibility study CemZero [14] showed that the production costs were likely to double if Cementa’s plant located in Slite were to be completely electrified compared to the costs in 2018. However, other CO2 mitigation techniques considered in the study were deemed more expensive. The cost of an
electrified separate calcination step would likely be lower, but does on the other hand not affect the emissions from the rotary kiln. While the cost of implementing large scale CO2 mitigation generally is
high, a study by Rootzén et al. [102] showed that the increased cost due to CO2 mitigation decreased
substantially along the value chain of cement, and the production cost of a residential building was calculated to increase by only 1% even though the cement price was assumed to almost double.
While the present work focuses on a process to facilitate carbon capture it is also important to remember that there are many different ways to mitigate CO2 emissions from both cement production and concrete
construction. Examples of such are reducing the clinker content in cement, minimizing the content of cement in concrete, and maximising the design life of buildings and infrastructure [1].
5.4 Sources of error
Several improvements could be made to the experimental method in the present work. Other studies investigating clinker formation as a function of temperature [55, 57, 58] often use a synchrotron X-ray source, resulting in a higher resolution, minimizing peak overlap. In addition, the XRD analysis could be improved by keeping the temperature constant during all scans. The setup could also be improved by instead of closing the valve to the HT-chamber to obtain a high CO2 atmosphere, a steady flow of gas
would be used. This would not only provide a 100% CO2 atmosphere and avoid the pressure buildup,
but also allow SEP-RM1 to be cooled in a low CO2 atmosphere. Furthermore, the atmosphere during
24
The heating and cooling program could also be improved to provide a fairer comparison between the different processes. SEP-RM1 had an equal calcination time as SEP-RM2, and since the size fraction of RM1 was much smaller than that of the limestone in RM2, a shorter residence time at calcination temperature for SEP-RM1 might be preferable. The reference experiments on the other hand had no residence time at all at calcination temperatures. This difference in heating introduces yet another source of uncertainty, making it more difficult to draw conclusions regarding which effects are due to a separate calcination step and which are due to difference in experimental design. The cooling rate could also be improved to better reflect industrial conditions. While no glassy phase was observed in the present work, a slower cooling under 1250°C might be preferable.
It is likely that carbonation of the calcined limestone caused the LSF to be lower for RM2 than for the other two raw meals. It would therefore be preferable to minimize the time between the different experimental steps to improve the experiments further, as identical standard modules of the raw meals would ease the comparison of the experiments.
6 Conclusions
The aim of the present work was to investigate how an electrified separate calcination step affects the phase composition of cement clinker, with emphasis on phases alite, belite, aluminate, ferrite, calcite, and lime. This was investigated by HT-XRD lab-scale experiments, allowing the phase transformations to be observed in situ. Two different methods of separate calcination were investigated, one method where the raw meal was calcined separately, and one method where the limestone was calcined separately. The former yielded an alite amount similar to that of the reference experiments, whereas the latter method yielded a lower amount. It could, unfortunately, not be excluded that the difference was due to poor experimental conditions, and additional experiments are needed to investigate the matter further. The study does, however, provide the following indications
A calcined raw meal might be used to produce a clinker of similar phase composition as the reference concerning major phases belite, aluminate, ferrite, alite, and free lime. A raw meal containing calcined limestone might need longer residence time at
clinkering temperature to obtain similar phase composition as the reference.
For a raw meal containing calcined limestone, alite formation might start at a higher temperature.
A raw meal containing calcined limestone might be carbonated to a greater extent upon reheating than a calcined raw meal.
Adequate storage is needed to avoid carbonation and hydration of a calcined limestone and/or raw meal.
Further experiments are needed to fully understand the effects on clinker composition of an electrified separate calcination step, and several improvements to the experimental method are given in the study. In addition, an electrified separate calcination step needs to be investigated from a heat and mass balance perspective to evaluate its effect on the production process.
7 Future work
Future work could include an improved experimental setup as described in section 5.4, as well as replicate experiments to provide more data. In addition, a more heterogeneous raw meal could be used for the separate calcination experiments, as the minor components are known to affect both calcination and clinkerization.
25
26
8 References
[1] A. F. Pales and Y. Leung, "Technology Roadmap: Low-Carbon Transition in the Cement Industry," International Energy Agency, 2018.
[2] R. M. Andrew, "Global CO2 emissions from cement production, 1928-2017," Earth system
science data, vol. 10, no. 4, pp. 2213-2239, 2018, doi: 10.5194/essd-10-2213-2018.
[3] C. Le Quéré et al., "Global Carbon Budget 2017," Earth system science data, vol. 10, no. 1, pp. 405-448, 2018, doi: 10.5194/essd-10-405-2018.
[4] Naturvårdsverket. "Utsläpp av växthusgaser från industrin efter växthusgas och bransch. År 1990-2018." https://www.naturvardsverket.se/Sa-mar-miljon/Statistik-A-O/Vaxthusgaser-utslapp-fran-industrin/ (accessed 06/11, 2020).
[5] Cementa. "Nollvision för koldioxid." https://www.cementa.se/sv/nollvision2030. (accessed 06/11, 2020).
[6] F. Schorcht, I. Kourti, B. M. Scalet, S. Roudier, and L. Delgado Sancho, "Best available techniques (BAT) reference document for the production of cement, lime and magnesium oxide: Industrial Emissions Directive 2010/75/EU (integrated pollution prevention and control)," vol. 26129, 2013.
[7] P. del Strother, "Portland Cement: Classification and Manufacture," in Lea's Chemistry of
Cement and Concrete, 5 ed. Oxford, England; Cambridge, Massachusetts:
Butterworth-Heinemann, 2019, ch. 2, pp. 31-56.
[8] Leilac, "Public Leilac feed summary report," Low Emission Intensity Lime and Cement project, 2017. Accessed: 07/12, 2020. [Online]. Available: https://www.project-leilac.eu/
[9] HeidelbergCement, "Sustainability Report 2019," 2020. Accessed: 07/12, 2020. [Online]. Available: https://www.heidelbergcement.com/en/sustainability-reports
[10] L.-A. Tokheim, A. Mathisen, L. E. Øi, C. Jayarathna, N. H. Eldrup, and T. Gautestad, Combined
calcination and CO2 capture in cement clinker production by use of electrical energy. 2019.
[11] ECRA, "ECRA CCS Project: Report on Phase IV. Technical report TR-ECRA-128/2016," European Cement Research Academy, Duesseldorf, 2016.
[12] VDZ, "Decarbonising Cement and Concrete: A CO2 Roadmap for the German cement industry," Verein Deutscher Zementwerke, Duesseldorf, 2020.
[13] L.-M. Bjerge and P. Brevik, "CO2 Capture in the Cement Industry, Norcem CO2 Capture Project (Norway)," Energy procedia, vol. 63, pp. 6455-6463, 2014, doi: 10.1016/j.egypro.2014.11.680. [14] B. Wilhelmsson, C. Kollberg, J. Larsson, J. Eriksson, and M. Eriksson, "CemZero – A feasibility
study evaluating ways to reach sustainable cement production via the use of electricity," Project report, Cementa/Vattenfall, 2018.
[15] E. Favre, R. Bounaceur, and D. Roizard, "A hybrid process combining oxygen enriched air combustion and membrane separation for post-combustion carbon dioxide capture,"
Separation and purification technology, vol. 68, no. 1, pp. 30-36, 2009, doi:
10.1016/j.seppur.2009.04.003.
[16] E. A. R. Trout, "The History of Calcareous Cements," in Lea's Chemistry of Cement and
Concrete, 5 ed. Oxford, England Cambridge, Massachusetts: Butterworth-Heinemann, 2019,
ch. 1, pp. 1-29.
[17] Cementa. "Cementproduktion steg-för-steg." https://www.cementa.se/sv/cementproduktion-steg-for-steg (accessed 12/11, 2020).
[18] A. M. Harrisson, "Constitution and Specification of Portland Cement," in Lea's Chemistry of
Cement and Concrete, 5 ed. Oxford, England; Cambridge, Massachusetts:
Butterworth-Heinemann, 2019, ch. 4, pp. 87-155.
[19] A. M. Harrisson, "The importance of raw materials," International Cement Review, 2014;46-8. [20] USGS, "Mineral Commodity Summaries," 2020.
[21] H. F. W. Taylor, "Portland cement and its major constituent phases," in Cement chemistry, 2 ed. London, England: Thomas Telford Ltd, 1997, ch. 1, pp. 1-28.
[22] I. Maki and S. Chromý, "Microscopic study on the polymorphism of Ca 3SiO 5," Cement and
concrete research, vol. 8, no. 4, pp. 407-414, 1978, doi: 10.1016/0008-8846(78)90020-0.
27
[24] H.-M. Ludwig and W. Zhang, "Research review of cement clinker chemistry," Cement and
concrete research, vol. 78, pp. 24-37, 2015, doi: 10.1016/j.cemconres.2015.05.018.
[25] I. Maki and K. Goto, "Factors influencing the phase constitution of alite in portland cement clinker," Cement and concrete research, vol. 12, no. 3, pp. 301-308, 1982, doi: 10.1016/0008-8846(82)90078-3.
[26] S. Telschow, F. Frandsen, K. Theisen, and K. Dam-Johansen, "Cement Formation - A Success Story in a Black Box: High Temperature Phase Formation of Portland Cement Clinker,"
Industrial & engineering chemistry research, vol. 51, no. 34, pp. 10983-11004, 2012, doi:
10.1021/ie300674j.
[27] M. A. Bredig, "Polymorphism of Calcium Orthosilicate," Journal of the American Ceramic
Society, vol. 33, no. 6, pp. 188-192, 1950, doi: 10.1111/j.1151-2916.1950.tb12789.x.
[28] J. G. Fletcher, "Borate fluxes in ordinary portland cement production: a feasibility study," ed: ProQuest Dissertations Publishing, 1991.
[29] L. Gobbo, L. l. Sant'Agostino, and L. Garcez, "C3A polymorphs related to industrial clinker alkalies content," Cement and concrete research, vol. 34, no. 4, pp. 657-664, 2004, doi: 10.1016/j.cemconres.2003.10.020.
[30] H. F. W. Taylor, "The chemistry of Portland cement manufacture," in Cement chemistry, 2 ed. London, England: Thomas Telford Ltd, 1997, ch. 3, pp. 55-87.
[31] C. N. Satterfield and F. Feakes, "Kinetics of the thermal decomposition of calcium carbonate,"
AIChE journal, vol. 5, no. 1, pp. 115-122, 1959, doi: 10.1002/aic.690050124.
[32] R. H. Borgwardt, "Calcination kinetics and surface area of dispersed limestone particles," AIChE
journal, vol. 31, no. 1, pp. 103-111, 1985, doi: 10.1002/aic.690310112.
[33] T. Darroudi and A. W. Searcy, "The effect of CO2 pressure on the rate of decomposition of lime,"
The Journal of Physical Chemistry, vol. 85, no. 26, pp. 3974-3978, 1981.
[34] J. Khinast, G. F. Krammer, C. Brunner, and G. Staudinger, "Decomposition of limestone: The influence of CO2 and particle size on the reaction rate," Chemical engineering science, vol. 51, no. 4, pp. 623-634, 1996, doi: 10.1016/0009-2509(95)00302-9.
[35] J. M. Valverde and S. Medina, "Crystallographic transformation of limestone during calcination under CO2," Physical chemistry chemical physics : PCCP, vol. 17, no. 34, pp. 21912-21926, 2015, doi: 10.1039/c5cp02715b.
[36] J. R. Fernandez, S. Turrado, and J. C. Abanades, "Calcination kinetics of cement raw meals under various CO2 concentrations," Reaction chemistry & engineering, vol. 4, no. 12, pp. 2129-214, 2019, doi: 10.1039/c9re00361d.
[37] ECRA, "ECRA CCS Project: Report on Phase III. Technical report TR-ECRA-119/2012," in "Technical report 119/2012," European Cement Research Academy, Duesseldorf, 2012. [38] L. Zheng, T. P. Hills, and P. Fennell, "Phase evolution, characterisation, and performance of
cement prepared in an oxy-fuel atmosphere," Faraday discussions, vol. 192, pp. 113-124, 2016, doi: 10.1039/c6fd00032k.
[39] C. W. Bale et al., "FactSage thermochemical software and databases," Calphad, vol. 26, no. 2, pp. 189-228, 2002, doi: 10.1016/s0364-5916(02)00035-4.
[40] S. Telschow, "Clinker Burning Kinetics and Mechanism," Ph.D. thesis, Department of Chemical and Biochemical Engineering, Technical University of Denmark, 2012.
[41] B. Hökfors, Phase chemistry in process models for cement clinker and lime production. Umeå: Department of applied physics and electronics, Umeå Universitet, 2014.
[42] G. Walenta and T. Füllmann, "Advances in quantitative XRD analysis for clinker, cements, and cementitious additions," Powder diffraction, vol. 19, no. 1, pp. 40-44, 2012, doi: 10.1154/1.1649328.
[43] J. Anderson, L. De Andrade Gobbo, and H. van Weeren, "X-Ray Diffraction: New Eyes on the Process," IEEE transactions on industry applications, vol. 51, no. 1, pp. 20-27, 2015, doi: 10.1109/TIA.2014.2350557.
[44] H. L. Hong, Z. Y. Fu, and X. M. Min, "Quantitative XRD analysis of cement clinker by the multiphase Rietveld method," (in English), Journal of Wuhan University of
Technology-Materials Science Edition, Article vol. 18, no. 3, pp. 56-59, Sep 2003.
[45] K. L. Scrivener, T. Füllmann, E. Gallucci, G. Walenta, and E. Bermejo, "Quantitative study of Portland cement hydration by X-ray diffraction/Rietveld analysis and independent methods,"
Cement and concrete research, vol. 34, no. 9, pp. 1541-1547, 2004, doi:
