International Journal of Mining, Materials, and Metallurgical Engineering (IJMMME)

ISSN: 2562-4571

The cement industry is responsible for high environmental impacts. The exploitation of natural raw materials can lead to resource depletion, deforestation, soil destabilisation, and other harmful effects [1]. In this scenario, the co-processing of industrial by-products in cement manufacture reduces the extraction of natural sources, promotes proper waste disposal, and often improves sustainability in production [2]. However, it is noteworthy that these by-products contain elements besides those essential for the Portland cement composition (Ca, Si, Fe, Al, O), which can alter the phase development during production.

Several studies have explored the co-processing of alternative raw materials in Portland clinker production [3]–[6]. This approach can improve manufacturing sustainability and product properties by optimizing clinker composition and burnability [7]. On the other hand, the presence of potentially contaminating elements can promote limitations in the production process on an industrial scale, change the stability of clinker phases, emission of polluting gases during clinkering, and contaminate the environment by the leaching of hazardous compounds [3], [8]. In this sense, the co-processing of alternative raw materials should always be associated with studying the effects of impurities in the clinker phase assemblage and the environmental risk assessment.

The solidification of metals in clinker was reported as altered by the added content, industrial technology, processing time, and other impurities [9], [10]. The presence of non-metals can reduce the stabilisation temperature of Ca₂SiO₄ polymorphs, altering the hydraulicity of the cement and the mechanical properties of the pastes [11]. Several studies have investigated the effect of alkaline and transition metals on the stability of aluminates and their polymorphs [12], [13]. Although extensive research has been carried out on the effect of metals and non-metals on clinker, the presence of lanthanides needs further investigation.

The limited number of research on the influence of lanthanides in clinker production was justified by the restricted contents of these elements in ordinary alternative materials (< 0.1 wt.%) [14]. However, the increasing volume of waste from the oil industry being co-processed in cement kilns in recent years demands new investigations addressing the effect of these impurities on clinkering and gas emissions. As an example, lanthanum and cerium oxides comprised up to 5.1 and 1.6 wt.% of the spent fluid catalytic cracking catalyst (SFCC), respectively [15], [16]. SFCC was co-processed in previous investigations to replace bauxite in clinker production [4]–[6]. Nevertheless, the findings examine neither the effects of lanthanides on the clinker mineralogical composition nor their emission potential during clinkering.

Understanding the action mechanisms of lanthanides can be technologically challenging. It demands the execution of an extensive and complex experimental program, consuming financial resources, and time. In this sense, thermodynamic modelling is a useful solution. It applies the Gibbs energy minimisation principle to estimate the clinker composition throughout manufacturing. Previous investigations have established it as an accurate tool for phase determination [17], [18]. Modelling allows for optimizing experimental approaches by analysing the effect of dopants on the composition of solid, liquid, and gaseous phases along clinkering [19]–[21]. It was applied to evaluate the effect of FeO, K₂O, MgO, Na₂O, SiO₂, SO₃, and TiO₂ [22]–[24]. However, further investigations could broaden the understanding of the effect of lanthanides incorporated in clinker raw meals due to the co-processing of alternative raw materials.

In this sense, this study applies thermodynamic modelling to investigate the influence of lanthanides on Portland clinker production. The effects of impurities on the stability of silicates and aluminates were investigated, in addition to the formation of new compounds and the polluting gases emission.

2. Methodology

2. 1. Thermodynamic modelling

Thermodynamic modelling was adopted to estimate the mineral composition of the clinkers produced in the theorical study. The software FactSage version 8.2 were applied to simulations. It contains optimised model parameters for the Gibbs free energy minimisation of solution phases to clinker equilibrium calculations [19]. Thermodynamic databases for gaseous components (FactPS) and oxides in solid, liquid, and solution phases (FToxid) were used for calculating the phase assemblage during production [19], [20]. The system pressure was fixed at 1 atm and the firing temperature was 1400 °C. As input data, the composition of clinker raw meal for an ordinary Portland cement (OPC) (69.5% CaO, 22.0% SiO₂, 5.0% Al₂O₃, and 3.5% Fe₂O₃) was used. The lanthanides were added individually as oxides (La₂O₃, Ce₂O₃, Eu₂O₃, Gd₂O₃, Pr₂O₃, and Nd₂O₃) in up to 10 wt.% of the raw meal in 0.5 wt.% increments. The influence of dopants on the phase assemblage was evaluated at 1200, 1300, and 1400 °C during clinkering. The processing was applied using the software Equilibrium module and adopting the entire set of products available in the system. Calcium silicates are quantified as Ca₃SiO₅ (C₃S) and Ca₂SiO₄ (C₂S). Thermodynamic modelling using FactSage considers solid solutions of calcium aluminium ferrites, such as Ca₂(Al,Fe)₂O₅ (C₂(A,F)) and Ca₃(Al,Fe)₂O₆ (C₃(A,F)) [17], [19], [20]. Where (Al,Fe) means Al and Fe are variable in the structure. The modelled solid solution C₂(A,F) is mainly associated with the OPC ferrite content, Ca₂(Al_xFe_1-x)₂O₅, often named as C₄AF [25]. The tricalcium aluminate or C₃A (Ca₃Al₂O₆) content is included in the C₃(A,F) solid solution, and the complementary content corresponds to intermediate phases in the development of C₄AF [20]. Thermodynamic modelling allowed quantifying other solids, including free calcium oxide (CaO) and new compounds formed from the chemical combination of lanthanides (LaAlO₃, NdAlO₃, PrAlO₃, EuAlO₃, GdAlO₃, Gd₄Al₂O₉, CeO₂, Ce₆O₁₁, and Ce₁₈O₃₁) [19]. The calculated liquid phase corresponded to the melt phase content formed during clinkering. The melt fraction is mainly developed through the melting of aluminates (Ca(Al,Fe)₂O₄, Ca₂(Al,Fe)₂O₅ and Ca₃(Al,Fe)₂O₆) from 1250 °C, reaching their maximum at the final clinkering temperature and then resolidifying during clinker cooling [14], [25].

FactSage is the software with the broadest database for thermodynamic modelling of reactions at high temperatures [26]. The accuracy of thermodynamic calculations has been reported in previous investigations that simulated the clinkering process. The results were experimentally validated in the following applications: Influence of bauxite residue co-processed in Fe-rich calciumsulfoaluminate-ferrite clinker (CSAF) [27]; Effects of co-processing impurities, including MgO, ZnO, Cr₂O₃, and NiO in different zones of an industrial rotary kiln [28]; Influence of the kiln atmosphere on the quality of clinker produced on an industrial scale [21]; Effects of impurities containing phosphorus on the stability of silicates in clinker [20]; Valorisation of blast furnace slag in clinker production [29]; Effect of mineralising agents (CaF₂, AlF₃, MgSiF₆, Na₂SiF₆, CaCl₂, ZnO, and CaSO₄) on clinker production [24]; Influence of the gradual substitution of Al₂O₃ with Fe₂O₃ in the raw material of an ordinary Portland clinker [30]. In this regard, all these applications have demonstrated that thermodynamic modelling results in better agreement with the phase composition of real clinkers compared to estimations from Bogue equations. These findings demonstrate the value of thermochemical calculations as an important tool for expanding the understanding of clinker production.

This study focused on analysing the effect of lanthanides during heating on clinker production (1200, 1300, and 1400 °C). The emission of compounds containing lanthanides was identified by quantifying gaseous phases during clinkering at 1400 °C through thermodynamic calculations on FactSage.

3. Results and discussions

3. 1. Composition of an Ordinary Portland Clinker (OPC) during heating

Table 1 presents the composition of an OPC at different temperatures to illustrate the phase evolution in an undoped system. Tricalcium silicate (C₃S) is the main phase of OPC, responsible for the early-age compressive strength of hydrated cement [2]. It is formed during clinkering after 1250 °C and obtained through the reaction between dicalcium silicate (C₂S) and calcium oxide (CaO) [14]. C₂S is generally the first silicate formed in clinkering. This compound increases the later-age compressive strength of cementitious materials [2]. Modelling using FactSage identifies its gamma (T < 735 °C), alpha prime (735~1435 °C), and alpha (T > 1435 °C) polymorphs [25]. If the system has free CaO, it combines with C₂S at about 1280 °C to form C₃S [14]. This reaction is verified in Table 1, where CaO is depleted at 1300 and 1400 °C as C₃S increases.

During heating, the proportion of C₃(A,F) tends to be higher than C₂(A,F). However, previous investigations have shown that this ratio tends to reverse during cooling [17]. The reason is the combination of CaO dissolved in the melt phase with C₂S forming additional C₃S [14]. Thereby, Ca availability is reduced for aluminates, promoting the formation of C₂(A,F) instead of C₃(A,F). In this sense, although the modelled OPC has a high content of C₃(A,F) during heating at 1200 and 1400 °C, the system tends predominantly to form C₂(A,F) during cooling [17]. At 1400 °C, C₂(A,F) and C₃(A,F) reached the melting point and then were converted to the melt phase.

Table 1. Predicted ordinary Portland clinker composition (wt.%) during clinkering at 1200, 1300 and 1400 °C.

Composition	Phase	1200 °C	1300 °C	1400 °C
Ca3SiO5	C3S	0.00	68.64	76.26
Ca2SiO4 α’	C2S α’	63.07	11.29	3.39
Ca2(Al,Fe)2O5	C2(A,F)	2.36	1.59	0.00
Ca3(Al,Fe)2O6	C3(A,F)	17.54	18.48	0.00
CaO	CaO	17.03	0.00	0.00
Melt	Melt	0.00	0.00	20.35

3. 2. Composition of lanthanide-doped clinkers at 1200 °C

Previous investigations established that the lanthanides would act similarly on the stability of the clinker phases due to their comparable atomic characteristics [14]. However, thermodynamic modelling indicated that Ce promotes notably different behaviours on the phase evolution during clinkering. On the other hand, La, Pr, Eu, Gd and Nd indeed have similar trends. Figure 1 shows the effect of lanthanide type and content on the phase assemblage during clinkering at 1200 °C. Figure 1a presents the data for La e as a representation of Pr, Eu, Gd, and Nd for simplification purposes, as the effect of these other lanthanides was similar.

La, Pr, Eu, Gd, and Nd stabilised C₂(A,F) rather than C₃(A,F). Clinkers start from 2.4 wt.% C₂(AF) and reach up to 8.2 wt.% when the dopant content is maximised (Figure 2a). For C₃(AF), the content varies from 17.5 wt.% to values below 1.1 wt.%. This behaviour indicates that the presence of lanthanides promotes the formation of clinkers rich in C₄AF. This characteristic is particularly promising for producing high ferrite clinker, in which the C₃A content is limited, to guarantee the requirements of resistance to abrasion and attack by sulphates [31]. The evolution of calcium aluminium-ferrite according to the Ce₂O₃ content stands out among the analysed lanthanides (Figure 2b). Ce destabilised C₂(A,F) by depleting it when 1.8 wt.% Ce₂O₃ was added. C₃(A,F) is partially reduced as the dopant increases. Ce acted as a flux, lowering the melting point, and gradually increasing the melt phase content up to 6.1% Ce₂O₃.

Co-processing impurities in cement kilns promote the formation of compounds beyond those commonly found in OPC. Figure 1 presents the new solid phases of the modelled systems with up to 10 wt.% of dopants. La, Pr, Eu, Gd, and Nd were incorporated as cubic perovskite containing Al and O. The doped systems indicated the formation of LaAlO₃, PrAlO₃, EuAlO₃, GdAlO₃, and NdAlO₃ with increasing values as the dopant content is increased. Previous investigations reported the formation of cubic lanthanum aluminate at temperatures above 540 °C [32], which can be pseudo-cubic or rhombohedral under industrial conditions [33]. Although the modelled structure of PrAlO₃ was cubic, this compound has the highest number of crystalline structure transitions among rare earth perovskites [34]. NdAlO₃ and GdAlO₃ were classified as materials with high optical conductivity, applicable in producing optical amplifiers and solar panels [35], [36]. For Gd, doping above 8 wt.% resulted in the destabilisation of GdAlO₃ and the formation of Gd₄Al₂O₉. This phase was synthesised in previous research by processing Gd₂O₃ and Al₂O₃ at temperatures ranging between 1400 and 1700 °C [37]. Gd₄Al₂O₉ is classified as an enhancer of the mechanical properties of ceramic materials, including the toughness and fracture strength of brittle ceramics [38].

Although the thermodynamic database includes the cerium aluminate perovskite [19], the simulations did not identify it (Figure 1b). Previous studies have shown that the combining potential of Ce in structures containing aluminium depends on several factors, including processing in a reducing atmosphere, pressure of O₂ in the atmosphere, proportion of other impurities in the reagents, and heat treatment conditions [39]–[41]. Thermodynamic calculations showed Ce₂O₃ did not react with Al₂O₃ in the clinker raw meal, remaining in its pure crystalline forms and varying as the dopant content increased (CeO₂, Ce₆O₁₁, and Ce₁₈O₃₁). Additional experimental investigations are necessary to understand the Ce action mechanism during clinkering and the formation of new solid compounds containing this element.

3. 3. Composition of lanthanide-doped clinkers at 1400 °C

The phase evolution at 1400 °C as a function of the dopant increment shows that La, Pr, Eu, Gd, and Nd enhanced C₃S formation (Figure 2a).

Up to 3% dopant, the C₃S content shows a positive development rate, starting from 76.3 wt.% (0% dopant) and reaching up to 79.0 wt.% (3% dopant). This behaviour is related to the C₂S destabilisation, which starts from 3.4 wt.% and is depleted in this range (Figure 2a). In this sense, lanthanides improve the clinker composition, increasing the C₃S content by optimizing the reaction between C₂S and CaO dissolved in the melt phase. A similar effect was observed in experimental samples doped with Praseodymium (Pr) [42].

After 3 wt.% dopant, the C₃S content decreased gradually, reaching 74.9 wt.% when 10 wt.% dopant was added. However, it is noteworthy that this percentage was improved, since the system produced 98% of the reference C₃S amount (instead of 90%) even with 10 wt.% substitution of raw materials by lanthanides. In this range, the depletion of C₂S limits the C₃S growth. After 8 wt.% dopant, part of the melt phase recrystallised as C₂(A,F). The results do not fully explain the recrystallisation occurring at high lanthanide contents. There are no lanthanides in the melt phase. Furthermore, resolidification occurs at different levels for each dopant, starting at 8.7 wt.% for La₂O₃, followed by Pr₂O₃ (8.8 wt.%), Nd₂O₃ (9.0 wt.%), and Eu₂O₃ (9.3 wt.%) (Figure 3). In this sense, high levels of this oxides (8~10%) increased the melting point of the clinker raw meal, promoting a considerably lower content of the melt phase at 1400 °C.

Cerium promoted discrepant effects compared to the analysed lanthanides (Figure 2b). Although the system started with the same C₂S content (3.4 wt.%) and it was depleted in advance (occurring at 2.3 wt.% Ce₂O₃), the C₃S content declined for increasing dopant levels. It reached 64.3 wt.% when 10 wt.% Ce₂O₃ was added, i.e., 84% of the calculated C₃S for OPC (76.3 wt.%). This decrease may be related to the SiO₂ destabilisation. In this sense, C₃S tends to decompose into solid CaO and molten SiO₂. The increased free CaO and melt in the C₃S depletion regions evidenced this behaviour (Figure 2b).

3. 4. Emissions

Table 2 quantifies the emission of potentially polluting gases containing lanthanides accessed by thermodynamic modelling.

Table 2. Emission of compounds containing lanthanides simulated by thermodynamic modelling of the clinkering of raw meals containing 10 wt.% lanthanide oxide (1400 °C).

Lanthanide	Compound	Emission (µg/t of clinker)
La	La	5.7 10-22
	La2	9.8 10-48
	LaO	1.0 10-7
	La2O	8.1 10-31
	La2O2	8.6 10-18
Pr	Pr	2.6 10-20
	PrO	1.7 10-6
Eu	Eu	2.6 10-12
	EuO	8.1 10-7
	Eu2O	1.4 10-22
	Eu2O2	4.6 10-14
Nd	Nd	1.7 10-19
	NdO	1.1 10-7
	NdO2	43.16
Gd	Gd	5.8 10-21

All dopants showed negligible emission levels. Nd had the highest emission potential. However, the volatilised phase (NdO₂) was in the order of micrograms per ton of clinker produced, a value easily mitigated by the protection infrastructure of industrial kilns. On the other hand, the modelling did not detect the emission of gases containing Ce. This behaviour was attributed to the stability of its oxides as solid phases, as discussed in the previous section. For La, Pr, and Eu, the emission tended to higher values when the oxidation state of these elements was 2⁺. According to the modelling, the emission of polluting gases containing lanthanides does not promote environmental risk since the emissions were approximately zero even when co-processing 10 wt.% of lanthanide oxide in the clinker raw meal.

4. Conclusions

According to the results of this study, the following conclusions can be drawn:

La, Pr, Eu, and Nd showed similar effects, enhancing the C₃S content. This effect occurred on a larger scale for up to 3 wt.% of added dopant, when C₂S is depleted in the clinker. Ce promoted notably different behaviours on phase assemblage, decreasing the C₃S development depending on the dopant content.

Increasing lanthanide content generally stabilised C₂(A,F) rather than C₃(A,F), suggesting that these elements may be applicable to produce high-ferrite and sulphate-resistant cements.

Co-processing lanthanides promoted the formation of new compounds, mainly the cubic perovskite containing Al and O (LaAlO₃, NdAlO₃, PrAlO₃, EuAlO₃, and GdAlO₃), with increasing values as the dopant content is increased. However, Ce₂O₃ remained in its pure crystalline forms, varying as the dopant content increased (CeO₂, Ce₆O₁₁, and Ce₁₈O₃₁).

The emission of polluting gases containing lanthanides was low even when co-processing 10 wt.% of lanthanide oxide in the clinker raw meal. However, it needs to be compared with the limits of the environmental regulations of each location. Furthermore, investigations are needed to understand how impurities could be immobilised in the cementitious material.

Acknowledgements

The authors express their sincere appreciation to the Bahia State Research Support Foundation (FAPESB, Brazil - Grant number 0289/2020 to A.R.D.C.) and the Coordination for the Improvement of Higher Education Personnel (CAPES, Brazil - Grant number 88887.608021/2021-00 to L.B.B.B.) for their financial support. J.P.G. and A.P.K. gratefully acknowledge the National Council for Scientific and Technological Development (CNPq, Brazil) and the Foundation for Research Support of the State of Rio Grande do Sul (FAPERGS, Brazil).

References