1 Introduction

Cement industry activities are estimated to contribute 5–7% of global man-made CO2 emissions [1,2,3,4]. Supplementary cementitious materials (SCMs) have been used by the cement industry to reduce CO2 emissions for many years [5,6,7,8,9]. However, since Portland cementitious systems cannot incorporate high amounts of SCMs without compromising mechanical properties, and because SCM properties vary by region, the search for alternative binders has intensified. One such binder is limestone calcined clay cement (LC3). LC3 boasts lower CO2 emissions and cost compared to Portland cement, due to its significantly lowered clinker content. Beyond ~ 7 d, its mechanical properties mirror those of Portland cement. Furthermore, it can be produced using equipment and materials already present in existing cement factories [10,11,12]. Typical LC3 comprises 50% Portland cement clinker, 30% calcined clay, 15% limestone, and 5% gypsum [13, 14]. Its mechanical performance, as well as its mineralogical and calorimetric properties, are related to the type and grade of calcined clay used in the binder, and the fineness of the components that make up LC3 [15, 16]. Common types of clay found in the earth’s crust include kaolinite, illite, and montmorillonite [17]. Calcination or thermal activation is the predominantly adopted method for increasing the reactivity of clays [18,19,20,21]. When thermal activation is applied, the pozzolanic activity of these minerals ranks, from highest to lowest, as kaolinite, montmorillonite, and illite [15]. Mechanical activation has gained attention in recent years as an effective method for activating illitic or montmorillonite clays. In addition to its lower energy requirement compared to calcination, mechanical activation can amorphize the structure of illitic and montmorillonite clays, a feat that most thermal activation methods cannot achieve [22,23,24]. Since kaolinite is richer in aluminosilicates than the other two clay minerals, most studies on LC3 have concentrated on utilizing kaolinite of various grades [10, 16, 25,26,27]. Notably, after 7 d, LC3-50 mortars, which incorporated calcined clays with at least 41.9% kaolinite, demonstrated mechanical properties comparable to those of Portland cement mortars [20]. This performance might be attributed to the formation of calcium silicate hydrate, resulting from the reaction of metakaolin with portlandite, and to the generation of carboaluminate phases, namely hemicarboaluminate and monocarboaluminate [28]. These phases develop through the interaction of alumina in C3A with both portlandite and limestone. In LC3, a similar reaction takes place, which involves an interaction among the reactive aluminosilicate phases derived from metakaolin, calcium carbonate from limestone, and portlandite from Portland cement. The fineness of the constituents directly affects the hydration degree of cementitious systems [29, 30]. A recent study [16] investigated the impact of two different fineness levels of each component of a typical LC3 mix, which included calcined clay (with less than 50% kaolinite), limestone, and cement. The study found that while clinker and calcined clay fineness can impact strength at all ages, limestone fineness significantly impacts only the early-age strength.

Current literature reveals a gap in studies that compare different clay types with significantly varied levels of fineness, aiming to highlight the critical role of component fineness in LC3. Moreover, these studies lack substantiation of their findings through comprehensive statistical analysis. To address this gap, the design of the study focuses on evaluating the performance of LC3, paying special attention to two distinct fineness levels of calcined clay and limestone. Consequently, three distinct types of calcined clays (kaolinite, illite, and montmorillonite) were carefully selected. The impact of the fineness of these clays and limestone on compressive strength was evaluated using univariate statistical analysis. Additionally, the study investigates the influence of component fineness in LC3 on the mineralogical, morphological, and hydration characteristics.

2 Materials and methods

2.1 Materials

The LC3 blends were prepared in a laboratory by mixing Portland cement (PC) clinker, limestone (LS), gypsum, and various calcined clays. These clays included kaolinite clay (Kao, from Zafer Mining, Balıkesir), illite clay (Ill, from Baştaş Cement Plant, Ankara), and montmorillonite clay (Mont, from Votorantim Cement Plant, Ankara). The chemical composition of the raw materials was determined using x-ray fluorescence spectroscopy (XRF, Rigaku ZSX Primus II), and their various physical properties are presented in Table 1.

Table 1 Oxide composition and physical properties of materials used
Full size table

Figure 1 shows the x-ray diffraction (XRD) patterns for the starting materials. Their mineral phase compositions were determined through Rietveld refinement using MAUD [31], as shown in Fig. 2.

Fig. 1
figure 1

XRD patterns for: a PC clinker, b Kao, c Ill, d Mont, e LS, f Gypsum (Legend: K-Kaolinite; Q-Quartz; I-Illite; Ga-Garronite; Mt-Montmorillonite; C-Calcite; Al-Albite; G-Gypsum; He-Hemihydrate; A-Anhydrite)

Full size image
Fig. 2
figure 2

Phase compositions of the materials used

Full size image

Figure 3 showcases the Blaine fineness of the materials and the grinding duration required to achieve the desired levels of fineness. The order of hardness of the LC3 mineral admixtures used in this study appears to be Ill > Mont > LS > Kao, consistent with findings from prior research [32].

Fig. 3
figure 3

Blaine fineness of the materials used and corresponding grinding durations

Full size image

Prior to milling, 5 kg of material was dried at 110 °C for each run. The ball mill, measuring 465 mm in length and 370 mm in diameter, had a volume of ~ 50 dm3. The milling media consisted of 15 hard steel balls of 50 mm diameter, 62 balls of 40 mm, 174 balls of 29 mm, 160 balls of 22 mm, and 210 cylpebs of 16 mm length, totaling 621 balls with a combined mass of ~ 57 kg, as described in [33]. The milling process, conducted using this laboratory ball mill, was performed at a rotation speed of 70 rpm. After the clays were ground, they underwent a calcination process. The optimum calcination temperatures for each type of clay were determined based on the strength activity index (SAI) test, following the specifications outlined in ASTM C311 [34]. SAI is defined as the ratio of the compressive strength of mortar incorporating 20% by mass of the calcined clay in the binder to the strength of the control mortar without any clay. A higher SAI indicates higher activity of the calcined clay. In determining the optimal calcination temperature, SAI was determined for clays calcined at every 50 °C increment between 600 and 900 °C. All the clays were heated at 5 °C/min, held at the target temperature for 3 h, and then cooled naturally in the furnace. The results indicated that the optimum calcination temperatures were 650 °C for Kao, 850 °C for Ill, and 600 °C for Mont. Although “optimum” temperatures reported in the literature for clays seem to vary, similar thermal activation temperatures have been reported for Kao [35, 36], Ill [37, 38], and Mont [39, 40]. However, it is important to note that the optimum calcination temperature is influenced by various factors, including the type of calcination (flash or static), the fineness of the clay being calcined, the amount of each mineral (kaolinite, illite, or montmorillonite), and the impurities present in the clay [41, 42].

2.2 Mix design

The gypsum content of the PC control mixture, needed to regulate the setting of the PC clinker, was determined using strength tests as outlined in ASTM C 563 [43], resulting in a PC gypsum:clinker ratio of 5:95. For LC3 containing 30% calcined clay and 15% limestone, the gypsum:clinker ratio was increased slightly and set as 3:52. The ratio used here is lower than in the literature that attempt to achieve a “sulfate balance” or to increase strength. The “proper” amount of added gypsum in LC3 depends on many factors, including SCM fineness and chemical composition. The use of calorimetry curves have been proposed to identify this amount [44, 45]. The decision to use a low amount of added gypsum allows the investigation of the effect of clay type and fineness and of limestone fineness on the sulfate need of the system. The samples were labeled according to the following convention: “LC3-Calcined Clay TypeCalcined Clay Fineness-LimestoneLimestone Fineness”. Table 2 gives the Blaine finenesses chosen for the powder components of the different mixtures.

Table 2 Finenesses chosen for the components in the LC3 blends
Full size table

2.3 Test methods

Mortars were prepared using silica sand in accordance with ASTM C305 [46], with a water-to-binder ratio of 0.48 and a flow of 110 ± 5%, achieved by adjusting the superplasticizer content as specified by ASTM C1437 [47]. The LC3 blends incorporated a superplasticizer (Sikament FFN) with a density of 1.15–1.19 kg/l at 20 °C. The amount of superplasticizer varied depending on the fineness of the calcined clays and limestone, with 1.5% by weight of cement used when both materials were fine, 1.3% when calcined clays were fine and limestone was coarse, and 1.0% when calcined clays were coarse, and limestone was fine. No superplasticizer was used when both materials were coarse. Strength development of these mortars was evaluated in accordance with ASTM C109 [48], using 50 mm cubes. After being removed from their molds, the mortar cubes underwent a curing process in water at room temperature (23 ± 2 °C) until reaching specified test age (1, 3, 7, and 28 d). The microstructures of the hardened LC3 pastes were investigated using scanning electron microscopy (SEM) coupled with energy dispersive x-ray (EDX) composition analysis (Zeiss Evo LS10). For the SEM analysis, paste samples were prepared following the mixing procedure outlined in ASTM C 305 and cured in water at room temperature until the day of testing. The 28-day bulk paste samples were examined at magnifications of 250× and 3000×. Additionally, 28-day bulk paste samples, prepared according to the same mixing procedure as the SEM analysis, were used for the mercury intrusion porosimetry (MIP, Quantachrome Poremaster 60) analysis. Pore size distribution was determined using a maximum pressure of 375 MPa. XRD analysis (Olympus BTX-II) was conducted in the range of 5–55°2θ with a resolution of 0.25°2θ at room temperature (23 ± 2 °C). For the XRD analysis of raw materials, no curing was applied, whereas for the hardened paste samples, the tests were prepared in accordance with ASTM C 305 and cured in water at room temperature (23 ± 2 °C) until the day of testing. On the testing day, the hardened paste samples were ground, and their microstructure was examined under the specified initial settings. The heat evolution of pastes was measured using an isothermal calorimeter (TAM Air, TA Instruments). Consequently, 10 g paste samples were prepared at room temperature (23 ± 2 °C) and subsequently loaded into the device. The hydration kinetics of the paste samples were monitored for up to 48 h.

3 Results and discussion

3.1 Strength development

Figure 4a, c, and e depict the strength development of LC3 produced using combinations of fine or coarse ingredients, Kao/LS, Ill/LS, and Mont/LS, up to 28 d. Meanwhile, Fig. 4b, d, and f show both the main effect plot and statistical outputs (i.e. F-value and p-value) for calcined clay and LS, considering two different fineness levels for the aforementioned test ages. The main effect plot represents the mean responses corresponding to each level of an independent variable, thereby illustrating the impact of these variables on the dependent variable within the context of an experiment.

Fig. 4
figure 4

Strength development (a, c, e), and main-effect plots and statistical results (b, d, f) for LC3 mixtures produced with fine and/or coarse Kao/LS, Ill/LS, and Mont/LS

Full size image

The 1-d and 3-d strength of LC3 mortars prepared using various clays are similar irrespective of the clay type and its fineness. However, strengths at 7 d and 28 d reveal that the clay type directly influences the performance of LC3. The difference in performance can be attributed to the distinct structural characteristics of the clays. For instance, kaolinite has a 1:1 structural layer, while illite and montmorillonite feature a 2:1 structural layer. Kaolinite has more OH groups in unique positions within its structure compared to illite and montmorillonite, leading to a unique decomposition that reduces crystallinity significantly, making it more reactive at later ages. Conversely, in illite and montmorillonite, the O2− groups from one layer align next to those of another layer, leading to fewer OH ions being released into the solution. Additionally, the structure of these minerals features an octahedral alumina sheet encased between two silicon tetrahedron sheets. Consequently, illite and montmorillonite are less reactive than kaolinite [15].

The main effect plots in Fig. 4b, d, and are utilized to analyze the differences in strength from the mean strength at different levels of the fine/coarse factor, represented by a line. A horizontal line indicates the absence of a main effect. Slight deviations from the horizontal line may significantly impact the response. A steeper slope signifies a greater magnitude of the main effect. The p-values helps interpret the significance of the result. For a confidence level of 95%, a p-value greater than 0.05 suggests that the result is not significant. A p-value less than 0.05 indicates significance, a value less than 0.01 denotes high significance, and so on [49]. Based on this information, whenever the fineness of clay or limestone at various ages is statistically significant, the corresponding F-value and p-value for these materials are highlighted with a yellow dashed rectangle. Figure 4b shows that in LC3-Kao-LS, the impact of Kao fineness on 28-d strength, and that of LS at 3 and 28 d, is statistically significant. The statistical significance of Kao fineness in the 28-d strength could be linked to the aluminate content and the structural layer of kaolinite. Meanwhile, the 1:1 structural layer of kaolinite fosters a more alkaline environment by releasing additional OH− ions into the cementitious matrix, as reported in [15, 50]. On the other hand, the statistical significance of LS fineness in LC3-Kao-LS on 3- and 28-d strengths is attributed to the filler effect of LS at 3 d and its role in forming carboaluminate phases at 28 d. Figure 4d shows that the fineness of Ill significantly affects strength at 1 d, 3 d, and 28 d, while the fineness of LS influences strength at 1 d and 3 d in LC3-Ill-LS. Interestingly, coarser Ill resulted in higher 1-d strength compared to its finer counterpart. This phenomenon was also observed in LC3 mixtures made with Kao, though the effect of Kao fineness on 1-d strength in LC3-Kao-LS did not reach statistical significance. In LC3 mixtures containing Kao and Ill, the result is linked to sulfate content. As reported by [51, 52], the sulfate requirement in cementitious mixtures is determined by their aluminate content. Therefore, the addition kaolinite, which is rich in aluminate, calls for adjusting the sulfate levels. Incorporating additional gypsum into the LC3 may slow down the hydration of aluminate components but improves that of silicate components, highlighting the pivotal role of gypsum in early-age strength. Consequently, in LC3 mixes with coarse Kao and Ill, lower sulfate demand for finer cases results in higher 1-d strength compared to mixes with finer Kao and Ill. Similar observations have been reported in other studies [53]. The statistical significance of Ill fineness in the 3- and 28-d strengths of LC3-Ill-LS is attributed to its inert nature and resulting physical effects, such as cement dilution, dispersion, and nucleation. This inert characteristic of Ill is further indicated by the lack of statistical significance in LS fineness, suggesting a limited interaction among Ill, LS, and Portland cement at 28 d. Figure 4f shows that the effect of the fineness of Mont and LS on strength of LC3-Mont-LS is statistically significant at different intervals: Mont at 3 d and 7 d, and LS at 3 d and 28 d. The significance of Mont fineness for strength at early ages might be due to the nucleation and filling effect of Mont. Although montmorillonite shows greater pozzolanic activity than illite, the statistical insignificance of Mont fineness for strength at later ages could be related to with the low amount of montmorillonite in this clay (Fig. 2).

3.2 Mineralogical changes

The 28-d XRD curves for LC3 pastes prepared with clays and limestone of varying fineness are shown in Fig. 5.

Fig. 5
figure 5

28-d XRD patterns of LC3 pastes produced with fine and/or coarse: a Kao/LS, b Ill/LS, c Mont/LS (Legend: E-ettringite; Hc-hemicarboaluminate; Mc-monocarboaluminate; T-Tobermorite; CH-Portlandite; Q-Quartz; G-Gypsum; C-Calcite)

Full size image

Figure 5 demonstrates that hemicarboaluminate which refines the pore structures and contributes to strength development [25, 28, 54], formed in all mixtures, regardless of the type of calcined clay used. However, in mixtures containing finer limestone, its intensity is more prominent. Both monocarboaluminate and hemicarboaluminate were observed in LC3-KaofineLSfine and LC3-KaocoarseLSfine (Fig. 5a). The conversion of hemicarboaluminate to monocarboaluminate in these mixtures occurred at 28 d, driven by the higher aluminate content in Kao compared to other calcined clays, the structural layer of kaolinite in Kao, and the influence of particle size of LS. Figure 5b illustrates that the greater amounts of hemicarboaluminate form in LC3-Ill-LS mixtures with fine LS than in mixtures with coarse LS. The fineness of Ill does not significantly impact the formation of hemicarboaluminate. This observation is also corroborated by the statistically insignificant strength of LS at 28 d in Fig. 4d. Figure 5c indicates that hemicarboaluminate in LC3-MontfineLScoarse is greater than in LC3-MontcoarseLScoarse. This suggests that Mont might be a significant aluminum source for LS. Several studies have highlighted that while montmorillonite possesses a 2:1 structural layer like illite, it is inherently less stable and may provide greater amounts of aluminum. Consequently, montmorillonite emerges as a more reactive component in LC3 [15, 55].

3.3 Pore size distribution and porosity

The pore structures of 28-d-old LC3 pastes determined with MIP are presented in Fig. 6.

Fig. 6
figure 6

Volume intruded and pore size distribution curves for 28-d LC3 pastes: a, b LC3-Kao-LS; c, d LC3-Ill-LS; e, f LC3-Mont-LS

Full size image

The LC3 pastes made with coarse calcined clay, typically exhibit the highest porosity, regardless of the type of clay used (Fig. 6a, c, and e). The pore size distribution confirmed the predominant presence of capillary pores (0.005–5 μm) across all samples. In the coarse samples of LC3-Kao and LC3-Ill, larger pores (> 1 μm) were primarily observed, potentially due to less effective compaction. In contrast, the fine cases of LC3-Kao and LC3-Ill exhibited fewer large pores and a higher concentration of capillary pores in the 0.010–0.1 μm range. Both fine and coarse ingredient cases of LC3-Mont displayed significant amounts of capillary pores, with the coarse case showing a slightly higher presence of larger pores compared to the fine ones. Comparing the threshold pore diameters of the pastes, the largest pore size at which the volume of mercury intruded increases significantly [56], the differences are noted depending on the type of calcined clay used. Specifically, for Kao, the threshold pore diameters range from 0.02 to 0.06 μm; for Ill, they vary from 0.008 to 1.8 μm; and for Mont, they extend from 0.03 to 0.1 μm. Assessment of the threshold diameter and normalized volume reveals differences in strength by highlighting the differences in quantities of large and small pores within the samples. Even though LC3-KaofineLSfine has the highest strength without having the smallest threshold diameter, the varying pore sizes in other mixtures account for their lower strengths.

3.4 Microstructure characterization

Figure 7, 8, and 9 illustrate the effect of calcined clay and limestone fineness on the microstructures of LC3. Figure 7a–e shows the impact of Kaofine/Kaocoarse and LSfine/LScoarse on the microstructures of LC3. LC3-KaofineLSfine matrix, rich in amorphous CSH gel and with minimal portlandite, contrasts with LC3-KaocoarseLScoarse. This difference stems from pozzolanic reaction that fills voids and enhances the microstructure. Despite the identification of needle-like ettringite structures in Fig. 7b and e, the observed densification of the microstructure in Fig. 7b may be predominantly attributed to the concurrent formation of CSH and hemicarboaluminate. This is confirmed by the EDX analysis of the hexagonal particles with smooth surfaces in Fig. 7b and e, which suggests they may be a poorly crystallized kaolinitic phase, in alignment with the clay mineral images in [57, 58]. The labeling of this mineral as hemicarboaluminate rather than calcium hydroxide is related to the amount of aluminum observed in the EDX analysis. Additionally, the presence of hemicarboaluminate in the 28-d XRD results further supports the identification of this mineral as hemicarboaluminate. The abundance of kaolinite in Fig. 7e relative to Fig. 7b might imply a lower hydration degree in LC3-KaocoarseLScoarse compared to LC3-KaofineLSfine.

Fig. 7
figure 7

SEM/EDX micrographs of 28-d pastes: a, b LC3-KaofineLSfine; c EDX result of Spot 1 in (b); d, e LC3-KaocoarseLScoarse

Full size image
Fig. 8
figure 8

SEM micrographs at 28 d: a, b LC3-IllfineLSfine, c, d LC3-IllcoarseLScoarse

Full size image
Fig. 9
figure 9

SEM/EDX micrographs at 28 d: a, b LC3-MontfineLSfine; c, d LC3-MontcoarseLScoarse; e EDX result of Spot 2 in (d)

Full size image

Figure 8 shows a pronounced presence of air voids in LC3 pastes made with Ill. This situation is not fineness-specific, as it persists across different Ill fineness levels (fine and coarse) and is also evident at greater magnification (Fig. 8b and d). The simultaneous presence of portlandite and Ill suggests reduced pozzolanic activity of Ill. In effective pozzolanic reactions, products like C–S–H typically fill voids, resulting in a denser microstructure. A greater amount of voids may imply less efficient reactions leading to reduced strength, similar to [59]. The characterization of the irregular sheet-like shapes with rough surface morphology in Fig. 8b and d as illite, based on [60], further supports the conclusion that LC3 pastes incorporating Ill fail to develop a densified microstructural matrix.

The visibly microcracked microstructure in Fig. 9a contrasts notably with that of Fig. 9c, the disparity is attributed to the expansive properties of montmorillonite (detected by EDX analysis in Fig. 9e) and aligns with findings in previous studies [61, 62]. Figure 9b further reveals the presence of portlandite and calcite structures, indicating incomplete reaction of Mont. Moreover, the microstructural of LC3-MontfineLSfine (Fig. 9b) reveals porous formations resembling holes. This observation suggests a potential delay in the pozzolanic reaction of Mont particles, agreeing well with [63,64,65].

3.5 Heat evolution

The change in normalized heat flow and cumulative heat for the various pastes are systematically presented in Fig. 10.

Fig. 10
figure 10

Normalized heat flow and cumulative heat versus time for: a LC3-Kao-LS; b LC3-Ill-LS; c LC3-Mont-LS

Full size image

A consistent observation across all samples is the occurrence of a marked exothermic reaction. The initial phase of this reaction, which lasts up to 30 min, is attributed to the wetting and dissolution processes characteristic of the paste composition. The second (alite) peak is directly related to the dissolution of C3S and the subsequent precipitation of C-S–H. In this phase, sulfate is also adsorbed onto the C–S–H. This peak becomes much more prominent when calcined clay and limestone are integrated into the cementitious matrix. This can be observed from Fig. 10 if heat flow is normalized by the amount of clinker instead of total binder. Since LC3 contains nearly half the amount of clinker in PC, the LC3 curves need to be doubled leading to alite peak heights of 5.0–5.5 mW/g clinker compared with ~ 3.9 mW/g clinker peak height for the control. This increase is linked to the filler effect produced by these two SCMs [28]. This physical effect related with the increased surfaces provided by the fine materials promotes the accelerated nucleation of C–S–H and adsorption of sulfate ions from the pore solution [44, 66]. The alite peak also helps assess the sulfate balance in LC3. Given a sufficient sulfate amount in LC3, the aluminate peak (the third peak) invariably succeeds the alite peak [44, 45]. The alite peak emerges earlier in LC3 produced with Kao compared to blends containing other clays, when the same clay/limestone fineness combinations are compared (Fig. 10a vs. Figure 10b or Fig. 10c). This suggests that gypsum is depleted more rapidly in LC3 mixtures containing Kao and that the chemical composition of the calcined clay is a factor influencing the sulfate balance. The enhanced alite reaction results in a higher rate of C–S–H and ettringite precipitation, and consequently, earlier sulfate depletion [67]. Additionally, the time between the aluminate and alite peaks in LC3 mixtures produced with Kao (compare the curves in Fig. 10a) is shorter than in those produced with Ill and Mont. Considering that all pastes contain the same amount of gypsum, this indicates a higher gypsum need (sulfate demand) in Kao-containing mixtures. Kao also causes narrower aluminate peaks. Furthermore, the LC3 paste produced with Kaocoarse has a longer time between these peaks than the paste produced with Kaofine, showing the effect of SCM fineness on the sulfate need. Hence, LC3 with Kaofine requires a greater amount of gypsum than LC3 with Kaocoarse. Similarly, LC3-Kaocoarse-LSfine requires more gypsum than LC3-Kaocoarse-LScoarse. However, the effect of LS fineness appears to be less significant than that of the Kao. This could also be related with the greater amount of calcined clay than limestone (30% and 15%) in LC3. Similar effects of clay fineness and limestone fineness are also observed in Fig. 10b and c. LC3 blends with Kao exhibit higher heat flow during the dormant period compared to those with Ill and Mont which also suggests more extensive sulfate consumption in kaolinite-rich mixtures. Comparing the total heat released by the different pastes, LC3 blends produced with Kao achieve the highest cumulative heat at 48 h, due to combined hydration, pozzolanic reaction, and formation of carboaluminates. LC3 blends with Mont registered the lowest heat. Although montmorillonite is generally more reactive than illite [15], the higher 48-h cumulative heat evolved by pastes containing Ill can be attributed to Ill containing ~ 62% illite but Mont containing only ~ 26% montmorillonite. Past studies [44, 45, 68, 69] offer mixed views on the impact of the calcined clay’s chemical composition on the sulfate need of LC3. The findings in this study suggest that both the fineness and chemical composition of the calcined clay influences the sulfate need of LC3.

4 Conclusions

This study statistically examined the effects of particle size and type of calcined clay, along with limestone particle size, on the mechanical performance of LC3, as well as their effects on morphology, mineralogy, and hydration kinetics. The following conclusions were reached:

  • The fineness of kaolinitic clay significantly influences 28-d strength at a 95% confidence level, but its effect is not statistically significant at earlier ages. Similarly, the fineness of calcined illite clay is significant for late ages, yet does not enhance early strength. The effect of the fineness of these two clays on late strength relates to their structural layers and chemical compositions. The fineness of montmorillonite clay significantly affects strength at 3 and 7 d, while its impact on 1 and 28-d strength is not significant. This could be due to the low montmorillonite content in the clay used, mainly physically affecting hydration.

  • Limestone fineness has a significant impact on 28-d strength of LC3 blends containing calcined kaolinitic and montmorillonite clays but not in blends with illite clay. The interaction between limestone and calcined clay emphasizes the crucial role of the position of aluminates within the structural layers of clays.

  • The hemicarboaluminate phase occurs in all LC3 mixtures, independent of clay type; however, monocarboaluminate formation requires the fine grinding of kaolinitic clay and limestone. Unlike kaolinitic and montmorillonite clay, whose fineness significantly affects the hemicarboaluminate quantity, the fineness of illite clay has little impact.

  • Clay type directly influences threshold pore diameter, with fineness playing a more crucial role in calcined clays of high activity and less so in those of low activity. Calcined clay fineness contributes to a denser microstructure more effectively than limestone fineness.

  • The sulfate requirement in LC3 mixtures is influenced more by the filler effect of calcined clays than by limestone. The aluminate peak intensity depends on the position of aluminum within the structural layer of calcined clay and its physical impact on the hydration of the cementitious system