The Effect of Different Particle Sizes of SiO₂ in Sintering on the Formation of Ternesite

Abstract

Ternesite is synthesized through sintering a mixture of CaCO₃, SiO₂, and CaSO₄ in a molar ratio of 4:2:1. Ternesite has a hydration rate between ye’elimite and belite in an aluminum-containing environment, and is considered to be a new material that can be used to enhance the performance of calcium sulphoaluminate cements. This experiment investigated the influence of different particle sizes of SiO₂ on ternesite formation. Controlled partial pressure sintering was employed within the temperature range from 1100 °C to 1200 °C, with a 72 h incubation period. The highest purity of ternesite in the samples reached 99.47% (500 nm SiO₂ sample). The analysis results from scanning electron microscopy and an energy dispersive spectrometer indicated that the particle size of SiO₂ exerted a significant influence on the formation of ternesite. In the preparation of ternesite from 10 μm particle size SiO₂, traces of calcium silicate were found in the product. The results of a thermal analysis further demonstrated significant distinctions in the thermal stability of ternesite prepared with SiO₂ of different particle sizes. Additionally, the crystallinity of ternesite was influenced by the particle size of SiO₂, consequently impacting the hydration performance of ternesite–calcium sulphoaluminate cement.

1. Introduction

Ternesite was initially discovered within the low-temperature zone of cement rotary kilns [1,2,3], and it was later confirmed to have a structure similar to silicocarnotite(Ca₅(PO₄)₂SiO₄) [4]. Ternesite possesses the chemical formula Ca₅(SiO₄)₂SO₄, denoted as C₅S₂S⁻, and belongs to the orthorhombic crystal system with the space group Pnma [5]. The macroscopic morphology of C₅S₂S⁻ presents itself as a light green solid, while its microscopic morphology exhibits a notable diversity, encompassing various forms such as irregular plate-like crystals, columnar crystals, and polyhedral crystals. These morphological variations are influenced by the temperature and duration of the sintering process of C₅S₂S⁻ [2,5,6,7]. Initially, C₅S₂S⁻ was regarded as an inert mineral with a negligible technological significance when encountered as an intermediate phase during the calcination of calcium sulfoaluminate (CSA) cement clinker [8,9,10,11,12,13]. However, recent research has unveiled that C₅S₂S⁻ exhibits hydration activity in aluminum-containing environments, positioning it as an intermediary between ye’elimite (C₄A₃S⁻) and belite (C₂S) [6,14,15,16,17,18,19].

Henceforth, the incorporation of C₅S₂S⁻ into CSA clinker, primarily composed of C₂S and C₄A₃S⁻, serves to bridge the substantial gap in the hydration activities between C₂S and C₄A₃S⁻. This strategic integration enables components with distinct hydration propensities to collaborate synergistically, thereby enhancing the overall performance of the cement [7,17,20,21,22]. Furthermore, due to the considerably lower sintering temperatures required for C₅S₂S⁻ in CSA clinker as compared to Portland cement (OPC), ternesite–calcium sulphoaluminate cement (T-CSA) founded on the C₅S₂S⁻ within CSA cement emerges as a highly promising low-carbon construction material. Upon establishing the hydration activity of C₅S₂S⁻ in aluminum-rich environments, numerous scholars have redirected their focus toward the comprehensive exploration of this mineral, instigating a plethora of investigations encompassing its formation and hydration kinetics.

The formation and microscopic morphology of C₅S₂S⁻ have been subject to extensive scholarly investigations. C₅S₂S⁻ can be synthesized through the reaction between C₂S and CaSO₄, with the formation temperature range spanning from 900 °C to 1200 °C. Notably, decomposition commences when the temperature surpasses 1200 °C, and upon exceeding 1250 °C, C₅S₂S⁻ undergoes complete decomposition into C₂S and CaSO₄ [2,20,23,24,25]. Jing et al. explored the impact of different crystalline forms of C₂S on the synthesis of C₅S₂S⁻, revealing that both β-C₂S and γ-C₂S were capable of yielding C₅S₂S⁻. Notably, β-C₂S exhibited a higher reactivity in this process compared to γ-C₂S, while α’L-C₂S demonstrated no discernible influence on C₅S₂S⁻ generation [26]. Skalamprinos delved into the effects of elemental doping on C₅S₂S⁻ formation, elucidating that the introduction of MgO, Na₂O, K₂O, SrO, MnO₂, TiO₂, and ZnO could induce a transition in the microscopic morphology of C₅S₂S⁻ from polyhedral crystals to nodular crystals [7]. Liu et al. discerned that both sintering temperature and sintering duration exerted a pronounced impact on the microscopic morphology of C₅S₂S⁻. Elevating the sintering temperature and extending the sintering duration facilitated the transformation of C₅S₂S⁻ from polyhedral crystals into columnar crystals, and subsequently into irregular plate-like crystals [2]. Zhang et al. reported that C₅S₂S⁻ grains produced through primary sintering exhibited an irregular morphology with generally larger sizes (~10 μm). In contrast, secondary sintering could reduce the grain size of C₅S₂S⁻ (~5 μm) and result in a more regular morphology. Rapid cooling with blast air at the conclusion of sintering was observed to reduce the crystallinity of C₅S₂S⁻ [27]. Furthermore, Zhang et al. investigated the influence of alkali metal doping on the crystal structure of C₅S₂S⁻. Their findings revealed that the solid solution of alkali metals within the crystal structure of C₅S₂S⁻ led to unit cell contraction or expansion, contingent upon the radius of the dopant ions, ultimately resulting in a reduction in the overall crystallinity of C₅S₂S⁻ [28]. Meanwhile, plenty of investigations have been conducted on the purity of C₅S₂S⁻. Hanein et al. determined that the partial pressures of both SO₂ and O₂ during sintering, in addition to the sintering temperature, played pivotal roles in controlling the synthesis of C₅S₂S⁻. Through the careful control of partial pressures (1175 °C, 72 h), they successfully synthesized C₅S₂S⁻ samples with a purity as high as 98% in a single step [5]. Likewise, Liu et al. employed a controlled partial pressure method, utilizing reagent-grade CaCO₃, CaSO₄·H₂O, and SiO₂, resulting in C₅S₂S⁻ samples with up to a 96.3% purity after 12 h of sintering at 1200 °C [2]. Wang et al. opted for a one-step calcination method within their hydrothermal synthesis approach, subjecting raw materials to a hydrothermal reaction at 120 °C for 3 h in a hydrothermal reactor, followed by compression and a subsequent 2 h sintering phase. They determined that the optimal sintering temperature for this method was 1150 °C, achieving a purity level of 94.9% for C₅S₂S⁻ [29].

In their research [2], Liu et al. observed that the formation of C₅S₂S⁻ occurs through a reaction in which SiO2 serves as the core component. At 1100 °C, SiO₂ forms the outer layer, which is subsequently enveloped by calcium silicate(CS) and C₂S. Notably, the formation of CS as a precursor to C₂S precedes the formation of C₂S. As the temperature is raised to 1200 °C, and with the extension of sintering time, all the previously generated C₂S and the outer layer of CS gradually transform into C₅S₂S⁻, starting from the outermost layer and progressing inward. However, it is important to note that Liu et al.’s study did not account for the influence of the raw material particle size on the reaction, which can significantly affect the reaction activation energy and specific surface area. Both of these factors collectively impact the overall reaction efficiency. Simultaneously, in an aluminum-rich environment, the hydration rate of C₅S₂S⁻ falls between that of C₂S and C₄A₃S⁻. This property can be harnessed to bridge the substantial discrepancy in the hydration reaction rates between C₂S and C₄A₃S⁻, ultimately improving the compressive strength at mid ages in T-CSA cement. Consequently, in this experiment, four different particle sizes of SiO₂ were deliberately selected to systematically investigate the impact of varying SiO₂ particle sizes, acting as the reaction core, on the formation of C₅S₂S⁻. So, in this experiment, four different particle sizes of SiO₂ were selected to investigate how SiO₂, as the reaction nucleus, affects the purity and crystallinity of C₅S₂S⁻. Additionally, the influence of ternesite prepared from SiO₂ with different particle sizes on the hydration performance of T-CSA cement was also studied.

2. Experimental Section

2.1. Materials and Samples Preparation

Given the chemical formula of C₅S₂S as Ca₅(SiO₄)₂(SO₄), it is evident that the ratio of Ca, Si, and S is 5:2:1. This allows us to determine the ratio of raw materials as CaCO₃:SiO₂:CaSO₄ = 4:2:1. In this experiment, high-purity reagent-grade CaCO₃ and CaSO₄ were chosen as raw materials, both of which had purity levels of 99+%. Additionally, four different types of high-purity SiO₂ with varying particle sizes were selected as variables. These reagents were sourced from Beijing InnoChem Science & Technology Co. Ltd., Beijing, China. Figure 1 illustrates the particle size distribution curves for the SiO₂ with varying particle sizes.

In accordance with the ratio of CaCO₃:SiO₂:CaSO₄ = 4:2:1, the constituent raw materials were introduced into a ball mill and alcohol was used as a grinding medium for the comminution process. This milling process lasted for a period of 3 h, resulting in a homogeneous blend of the raw materials. Subsequently, the treated raw materials were removed from the ball mill and subjected to a drying process in an oven set at 50 °C for a duration of 24 h. During this experiment, the partial pressure of SO₂ was carefully controlled under one-step sintering conditions. The thoroughly mixed raw material was then placed into a mold and compressed into a tablet with a diameter of 2 cm and a thickness of 3 mm at a pressure of 15 Mpa [30]. Following the compression of the tablet, the SO₂ partial pressure was controlled as depicted in Figure 2. The tablet was then placed into a muffle furnace and maintained at the target temperatures, ranging from 1100 °C to 1200 °C (1100 °C, 1110 °C, 1120 °C... 1190 °C, and 1200 °C). The heating rate employed for this process was 3.3 °C/min, maintained at the target temperature for 72 h with a cooling rate of 6 °C/min.

2.2. Characterization

2.2.1. X-ray Diffraction (XRD) Characterization and Analysis

To explore the impact of the SiO₂ with varying particle sizes on the formation of C₅S₂S⁻, the samples underwent a thorough comminution process followed by an X-ray Diffraction (XRD) analysis. The XRD analysis was conducted using a Bruker D8 Advance XRD diffractometer(Bruker Corporation, Karlsruhe, Germany), and the tests were carried out at room temperature. The XRD tests for C₅S₂S⁻ involved scanning in the 10°–80° (2θ) range with a step size of 0.02° and a scanning rate of 2°/min. This analytical approach allowed for the assessment of the crystallographic structure and phase composition of the samples, providing valuable insights into the formation of C₅S₂S⁻ and any potential variations associated with different SiO₂ particle sizes.

The prepared samples were subjected to a quantitative analysis using the Rietveld method with the assistance of the FullProf 64b software. In this analytical process, the data obtained from the XRD experiments were fitted using the Pearson VII function. Additionally, the crystal structure files from

Table 1. ICSD codes used for Rietveld refinements.

Mineral Name	Phase	ICSD Codes
Ternesite	C5S2S−	85123
Belite	C2S	79552
Anhydrite	CaSO4	40043
Lime	CaO	52783

were refined based on information from the ICSD (Inorganic Crystal Structure Database) database. This methodology allowed for a detailed examination and refinement of the crystallographic properties and structures of the samples, providing a comprehensive understanding of their composition and characteristics.

2.2.2. Scanning Electron Microscope (SEM) and Energy Dispersive Spectrometer (EDS) Analysis

A HITACHI Regulus8100 scanning electron microscope(HITACHI, Tokyo, Janpan) was used to characterize the microscopic morphology of the prepared samples with an accelerating voltage of 5 kV and a working distance of 8.3 mm.

2.2.3. Thermal Analysis

Differential thermal analysis-thermogravimetry (DTA-TG) was tested using a NETZSCH STA2500 Differential Thermal Analyzer(NETZSCH, Free State of Bavaria, Germany) with a ramp rate of 10 °C/min under a nitrogen atmosphere.

2.2.4. Hydration Heat

An eight-channel isothermal calorimeter of TMA AIR was used to test the heat of hydration of the CSA cement blended with 10% of the sample with a water–cement ratio of 0.4.

3. Results and Discussion

3.1. Formation of Ternesite

After 72 h of sintering, a light green colored sample was prepared. To confirm the successful preparation of C₅S₂S⁻ and assess the purity of C₅S₂S⁻, the samples were subjected to an XRD analysis. A quantitative analysis was conducted using the Rietveld method with the assistance of the full prof software, as illustrated in Figure 3 and Figure 4. The obtained Rwp values fell within the acceptable range [31]. From Figure 3, it can be observed that the purity of the C₅S₂S⁻ samples synthesized using the 500 nm SiO₂ particles (referred to as the 500 nm_1170 °C_sample) reached an impressive 99.47% at a sintering temperature of 1170 °C. Furthermore, the highest purities of C₅S₂S⁻ among the samples prepared using SiO₂ particles of different sizes were as follows: 87.22% for the 20 nm_1160 °C_sample, 94.02% for the 25 nm_1140 °C_sample, and 67.01% for the 10 μm_1170 °C_sample, respectively. Additionally, the presence of CS [32] was solely detected in the sample with the 10 μm particle size SiO₂ as the raw material. The content of CS amounted to a mere 0.15%, and this occurred only after an increase in the sintering temperature to 1120 °C. These findings align with Liu’s research [2], indicating that CS acts as a precursor to C₂S and is fully converted into C₂S upon further elevation of the sintering temperature. Conversely, no traces of CS were detected in the C₅S₂S⁻ samples derived from the other three particle sizes of SiO₂. This discrepancy may be attributed to the significantly larger particle size and reduced specific surface area of the 10 μm SiO₂, resulting in an elevated reaction activation energy and, consequently, a reduced reaction efficiency. In general, the purity of C₅S₂S⁻ in the 10 μm sample was consistently low across the sintering temperatures ranging from 1100 °C to 1200 °C. This supports the notion of a low reaction efficiency associated with SiO₂ particles of a 10 μm size. On the other hand, when comparing the samples prepared from the SiO₂ particles of different sizes, the 500 nm particle size exhibited the highest purity at all sintering temperatures. Notably, the 500 nm_1170 °C_sample achieved an impressive purity of 99.47%.This intriguing phenomenon leads to speculation: the 500 nm particle size SiO₂ possessed a lower reaction activation energy when compared to the 10 μm particle size SiO₂. Additionally, compared to the SiO₂ particles of 20 nm and 25 nm sizes, the 500 nm particle size SiO₂ offered a larger reaction contact area. Consequently, under the synergistic effect of these two factors, the 500 nm SiO₂ outperformed the other three particle sizes in terms of reaction efficiency, making it the most suitable choice for the preparation of C₅S₂S⁻. In addition, except for the 500 nm SiO₂ sample, CaO was observed in all other samples. This is a product of the decomposition of CaCO₃ at high temperatures. Due to an inadequate reaction activation energy and reaction contact area for the SiO₂ with particle sizes of 20 nm, 25 nm, and 10 μm, the raw materials did not react fully, resulting in the presence of residual CaO. Meanwhile, the 10 μm SiO₂ exhibited the smallest specific surface area and highest reaction activation energy, leading to the lowest purity of C₅S₂S⁻ in the samples it produced.

3.2. SEM/EDS

From Figure 5a–c, it can be observed that the morphology of the C₅S₂S⁻ prepared in this experiment bears a resemblance to the samples prepared by Skalamprinos et al. [7], exhibiting an irregular granular appearance. The 10 μm_1170 °C_sample shown in Figure 5d contains numerous spherical and variably sized block-like particles. The EDS analysis in Figure 6d reveals that the central parts of these spherical particles are primarily composed of Ca and Si elements (note that Figure 5 and Figure 6 are not the same location). Considering the sintering conditions and raw materials used, it can be inferred that the main constituent of the central parts of these spherical particles is C₂S. At the edges of the spherical particles, three elements are concurrently present: Ca, Si, and S. This suggests the formation of C₅S₂S⁻ at the periphery of the spherical particles. This is in accordance with the findings of Liu et al. [2], where the formation of C₅S₂S⁻ occurred through a reaction initiated at the core of SiO₂. With an increase in the sintering temperature and a prolonged sintering duration, the C₂S enveloping the exterior of SiO₂ reacts with CaSO₄ to form C₅S₂S⁻. The EDS results for the samples prepared with the other three particle sizes of SiO₂ indicate that the distribution of S elements aligns with the distribution of Ca and Si. This is attributed to the higher specific surface area of the smaller SiO₂ particles, which is more conducive to the reaction leading to the formation of C₅S₂S⁻. Smaller SiO₂ particles facilitate extensive contact and reaction with other raw materials, resulting in an even distribution of S elements in the product. In conclusion, in conjunction with the refinement results in Figure 3, it can be inferred that the formation process of C₅S₂S⁻ begins with SiO₂ as the nucleus, where CS is formed first on the outer layer, followed by the formation of C₂S, and finally, C₂S reacts with CaSO₄ to produce C₅S₂S⁻. Using SiO₂ as the reaction nucleus imposes requirements on the particle size and specific surface area of SiO₂, making 500 nm SiO₂ suitable for the preparation of C₅S₂S⁻.

3.3. Thermal Analysis

Due to the relatively low purity of the C₅S₂S⁻ (67.01%) in the sample prepared from SiO₂ particles with a size of 10 μm, this particular sample was excluded from a direct comparative weight loss analysis with the high-purity samples. Consequently, only three samples with a high C₅S₂S⁻ purity, namely the 20 nm_1160 °C_sample, 25 nm_1140 °C_sample, and 500 nm_1170 °C_sample, were selected for a thermogravimetric analysis. Hanein et al. conducted a similar experimental setup, sintering the raw material in a two-stage process at 1175 °C for 3 days and 1 day, respectively [5]. This resulted in the preparation of a C₅S₂S⁻ sample with a purity of 99% and an upper limit decomposition temperature of approximately 1290 °C. Furthermore, Liu et al. demonstrated that the thermal stability of C₅S₂S⁻ is directly proportional to the holding time during sintering [33]. Longer sintering durations promote an improved crystal development in C₅S₂S⁻, thereby enhancing the thermal stability of the samples. In the current experiment, the holding time for all the samples during sintering was 72 h. Therefore, the particle size of the SiO₂ employed in the preparation of C₅S₂S⁻ emerges as the primary determining factor for the thermal stability of the resulting samples. It appears that, in the thermogravimetry test shown in Figure 7a, there is minimal difference in the weight loss between the 20 nm_sample and 25 nm_sample, while the 500 nm_sample experiences a relatively low weight loss. The refinement results in Figure 3 indicate that the main impurities in the sample are CaSO₄ and C₂S. The decomposition temperature of CaSO₄ falls within the range of 1350–1400 °C, while the decomposition temperature of C₂S is above 1400 °C. Therefore, a temperature of 1340 °C was chosen for the calculation of the weight loss. At this temperature, only C₅S₂S⁻ undergoes decomposition. Based on these findings, the 500 nm_1170 °C_sample demonstrates a high purity and superior thermal stability. The sintering conditions of 1170 °C for 72 h are found to be particularly suitable for the 500 nm particle size SiO₂ to undergo the necessary reaction for the formation of C₅S₂S⁻ and crystal development. This combination results in a higher purity and optimal crystal development, thereby providing the best high-temperature stability. Consequently, the 500 nm_1170 °C_sample is less likely to decompose at high temperatures and exhibit mass loss in the form of SO₂ gas [2] (as indicated by “Equation (1)”). Additionally, the decomposition temperature of the 500 nm_1170 °C_sample, as revealed by the DTA (differential thermal analysis) results shown in Figure 7b, is the highest among the samples, estimated to be around 1280 °C. This further supports the superior thermal stability of the 500 nm_1170 °C_sample. In summary, based on the thermogravimetric and refinement results, along with the DTA analysis, the 500 nm_1170 °C_sample exhibits a high purity, excellent crystal development, and superior thermal stability, making it less prone to decomposition at high temperatures.

Ca₅(SiO₄)₂SO₄→2Ca₂SiO₄ + CaO + SO₂+ 1/2O₂

(1)

3.4. Hydration Heat

After Montes et al. demonstrated that C₄A₃S⁻ can effectively stimulate the hydration activity of C₅S₂S⁻ [15], Zhang et al. further investigated this phenomenon [27]. Their study revealed that C₅S₂S⁻ with a higher crystallinity exhibits poorer early hydration activity, while C₅S₂S⁻ with a lower crystallinity displays higher early hydration activity. In Figure 8a, it is evident that the sample represented by 500 nm SiO₂ particles has the lowest exothermal peak intensity and the latest formation time compared to the other three particle sizes. This is attributed to the highest purity of the samples being prepared from the 500 nm SiO₂ particles, sintered at 1170 °C for 72 h, representing the most suitable sintering conditions for the formation and development of C₅S₂S⁻ crystals. In other words, the 500 nm_1700 °C_sample exhibits the best crystallinity, indicating lower early hydration activity. The cumulative heat curves in Figure 8b also support this observation, as the 500 nm sample exhibits the lowest total exothermic amount, while the exothermic amounts of the other three groups of samples are relatively similar, further confirming the lower early hydration activity of the 500 nm_1700 °C_sample. Additionally, the three groups of samples prepared with the 20 nm, 25 nm, and 10 μm SiO₂ particles show a shoulder peak before 1 h. This is attributed to the presence of a small amount of CaO impurities in the C₅S₂S⁻ prepared with these SiO₂ particle sizes (20 nm, 25 nm, and 10 μm). These impurities react exothermically with water to form Ca(OH)₂, leading to the appearance of the shoulder peak. The intensity of the shoulder peak is influenced by the CaO content. The 25 nm_1140 °C_sample has the highest CaO content, leading to the highest intensity of its shoulder peak. In summary, due to its superior crystallinity, the 500 nm_1700 °C_sample exhibits the slowest early hydration rate and the lowest hydration reaction intensity, resulting in the lowest cumulative heat. The slower hydration rate and lowest cumulative heat of 500 nmSiO₂-T-CSA can mitigate, to some extent, the disadvantage of poor dimensional stability in alumina cement.

4. Conclusions

In this experiment, we addressed previously overlooked aspects by utilizing SiO₂ of varying particle sizes as raw materials to investigate their impact on the formation of C₅S₂S⁻, changes in morphology, thermal stability, and the influence on the hydration heat of T-CSA cement. This study yields the following conclusions:

• Different particle sizes of SiO₂ have a significant impact on the formation of C₅S₂S⁻. Specifically, C₅S₂S⁻ prepared using 500 nm SiO₂ as a raw material can achieve a purity as high as 99.47%, whereas C₅S₂S⁻ prepared with 10 μm SiO₂ as a raw material exhibits a purity of only 67.01%. Additionally, the presence of CS is detected in the samples prepared with 10 μm SiO₂.
• In the samples prepared using SiO₂ particles with sizes of 20 nm, 25 nm, and 500 nm, S elements are evenly distributed within each particle, facilitating the relatively smooth formation of C₅S₂S⁻. Conversely, in the samples prepared with 10 μm SiO₂ particles, C₅S₂S⁻ containing S elements coats the edges of the spherical particles dominated by SiO₂ and C₂S. Due to the influence of the SiO_2′s specific surface area, the formation of C₅S₂S⁻ is relatively challenging.
• In the TG and DTA tests, the 500 nm_1700 °C_sample exhibits the best thermal stability, primarily due to its superior crystallinity. In the hydration heat test, also influenced by the better crystallinity of the 500 nm_1700 °C_sample, CSA cement doped with this sample displays the slowest hydration rate and lowest cumulative heat release.

In this experiment, the optimal SiO₂ particle size and temperature for ternesite preparation were identified, and the underlying reaction mechanisms were analyzed. Additionally, the hydration performance of ternesite was studied for potential applications, offering valuable insights for future research on low-carbon construction materials and alumina cement.