Investigation on the Micro-Grinding Induced Crystallographic Variations of Nine Different Clinkers

Although clinker has been used for many years, complicated mineralogical properties of clinker pose challenges for the precise quantification. In this study, the mineralogical and crystallographic properties of nine different clink‑ ers according to grinding procedures were investigated. With the dry‑grinding for 2 h, particle size reduction to 3 μm of median particle size with a substantial phase transition to an amorphous phase observed, to which alite (C 3 S) mainly contributed to the transition. Meanwhile, the crystallographic properties of the clinker phases were barely changed during the wet‑grinding. In the wet‑grinding program, the amount of ferrite solid solution (C 4 AF) with a high linear absorption coefficient was not underestimated. Furthermore, well‑corrected preferred orientation effect on C 3 S was positively contributed to the analysis result of clinkers with the wet‑grinding. Hence, it was suggested that the crystallographic effects observed in the wet‑grinding program could produce more reliable results in phase analysis for the clinkers.


Introduction
Worldwide, approximately 4 billion tons of Portland Cement (PC) is being consumed every year because of its suitable material performance and price competitiveness (Alyaseen et al., 2023;Bamigboye et al., 2020;Li et al., 2003;Shah et al., 2023;Tran et al., 2023).PC consists of various minerals, such as alite (C 3 S), belite (C 2 S), aluminate (C 3 A) and ferrite solid solution (C 4 AF).The type of cement varies and is classified based on the content of clinker phases.Depending on the relative contents of these minerals, the material properties of PC develop differently (Miller, 2018;Xie & Visintin, 2018).Hence, the minerals present in cement should be thoroughly calculate to design the performance of cement-based materials.For instance, to understand the mineralogical characteristics and hydration of cement, thermogravimetric analysis, nuclear magnetic resonance, and X-ray powder diffraction (XRPD) techniques have been actively utilised (Alarcon-Ruiz et al., 2005;Hesse et al., 2011;Naik et al., 2006;Parry-Jones et al., 1988).Amongst them, XRPD is one of the most adopted methods for analysing the mineralogical properties of PC (Goergens et al., 2020;Scarlett et al., 2001;Scrivener et al., 2004).
For the optimal investigation of XRPD, various factors must be considered, such as measurement condition, analysis method, and sample preparation.The adequate measurement conditions including X-ray scanning range, spinning speed, type of slit, maximum sample thickness, and step size should be determined empirically or theoretically (Wang, 1994;Whitfield & Mitchell, 2009).One analysis method, the Rietveld refinement method, is used to quantify the crystalline and amorphous phases combined with the internal standard or external standard method.If the amorphous phase is contained in the powder, then the crystalline and amorphous content can be indirectly quantified with the known information of a standard material (Fiebich & Mutz, 1999;Phillips, 1997;Scarlett et al., 2002;Shafi et al., 1997;Young, 1993).Accordingly, various discussion and proposals on the optimal measurement conditions and subsequent analysis method for quantitative XRPD (QXRPD) have been reported (Brindley, 1945;De la Torre & Aranda, 2003;Taylor et al., 2002;Whitfield & Mitchell, 2009).
Likewise, the effect of the particle size distribution in XRPD analysis has been also investigated (Gordon & Harris, 1955;Mitchell et al., 2007;Taylor et al., 2002;Whitfield & Mitchell, 2009).A large particle size could introduce errors into the XRPD pattern, which is closely related to the large crystallite size and the poorly refined preferred orientation (Dongxu et al., 2000;Ermolovich & Ermolovich, 2016;Kumar et al., 2004;Mejdoub et al., 2017).If the linear absorption coefficient of the specific crystalline phase is higher than those of the other phases, then the phase content of the phase can be underestimated (Guirado et al., 2000;Peterson et al., 2006).In addition, it was reported that larger particle size of clinker may yield the different shapes of patterns for identical measurement.Therefore, the patterns would not be able to be well-unified (Whitfield & Mitchell, 2009).Therefore, the accuracy of the XRPD analysis is highly affected by the particle size distribution of the powder.Recently, research has investigated whether the amorphous phase is contained in ordinary Portland Cement (PC), focusing on the factors mentioned above (León-Reina et al., 2009;Suherman et al., 2002;Walenta & Füllmann, 2004).Torre et al. reported that a remarkable amount of amorphous content was found in PC.Using the QXRPD technique combined with the internal standard method, the amount of the amorphous phase was calculated.And, to identify the degree of error, the amount of contained internal standard powder and the scan range of the XRPD pattern were assigned to experimental parameters.It was concluded that PC and clinker contained significant amounts of amorphous content (De la Torre et al., 2001).However, Jensen et al. conducted investigations into the effect of the physical properties of C 3 S in QXRPD analysis.In addition, the reliability of QXRPD combined with the internal standard and external standard methods was discussed.It was concluded that amorphous content was not found in PC (Jansen et al., 2011).Furthermore, the effect of the particle size distribution of powder was investigated by Snellings et al. ().To verify the reliability of the QXRPD results, Snellings and colleagues adopted the external and internal standard methods, and two types of software (i.e.Highscore and TOPAS software) for QXRPD were employed to cross-check the results: but the significant differences were not confirmed.It has been also reported that the amorphous content was not included in the investigated PC in the study.
As summarised above, there is a still active debate on the nature of amorphous phase in PC.It is due to the analytical complexity of QXRPD and material uncertainty or variability of PC.To elucidate the issue, nine as-received clinkers (not PC) were investigated with two different programs of dry-and wet-grindings.It includes particle size analysis and the amount of the amorphous phase which was investigated by quantitative Rietveld analysis with the internal standard method.The increase of amorphous phase by the dry-grinding process was correlated with the decrease of certain mineral phase in the clinkers.The crystallographic difference in different grinding programs was interpreted with the degrees of preferred orientation of monoclinic C 3 S (M3) and micro-absorption of C 4 AF in the clinkers with different particle size.

Materials and Experimental Methods
Nine types of clinker were collected from different cement plants in the Republic of Korea.Clinker powders were pre-treated as follows.The acquired clinker nodules underwent an actual cement grinding process.In specific, the nodules were initially crushed in a roll crusher, a preliminary grinding process, without the addition of gypsum.Subsequently, the obtained coarse clinker powder was pulverised into a very fine powder using a cement ball mill, which is the main grinding facility.Last, pulverised homogeneous powder forms of clinkers were prepared with a Henschel mixer.Particle size distributions of the as-received clinkers will be compared with those of micro-ground clinker in Sect.3.1.The X-ray fluorescence results (S8 Tiger, Bruker Co. Ltd., Land Baden-Württemberg, Germany) are presented in Table 1, with the loss of ignition (LOI) obtained from a weight loss percentage by heating at 1050 °C for 2 h.The variations were observed in Al 2 O 3 , Fe 2 O 3 , and MgO content according to the type of clinker.Al 2 O 3 variations could be linked to both cooling rate and the addition of a significant Fe source.Discrepancies in Fe 2 O 3 are likely due to variations of content of raw material, a crucial iron source in clinker production.
Dry-or wet-grinding programs were performed using lab-scale micro-ball mill equipment (XRD-Mill McCrone, Retsch Co. Ltd., Retsch-Allee, Germany) with a 125-mL jar and 48 cylindrical grinding balls were adopted for grinding clinker powders.Based on the preliminary experiments, 5 g of clinker was pulverised for 2 h for dry-grinding and methanol solution (6 mL) was added for wet-grinding program.In more detail, the grinding time was determined based on numerous preliminary experiments, set at the minimum duration where the preferred orientation effect of C 3 S remains constant.After wet grinding, the powders were separated from methanol using a negative pressure machine and then dried for 40 min at 70 °C.The grinding durations were separated for 1 min after 10 min of grinding to prevent potential temperature rise.To measure the particle size distribution of the clinkers, the particle size analyser (Malvern Instrument Ltd., Malvern, UK) was used.Two measurements were taken per sample, and the measurements were made after achieving sufficient dispersion in isopropanol.The as-received clinker, dry-ground clinker, and wet-ground clinker are notated as ARC, DGC, and WGC, respectively.
XRPD patterns were measured with an X-ray diffractometer (D2 Phaser, Bruker Co. Ltd., Land Baden-Württemberg, Germany), and the obtained patterns are presented in Additional file 1: Figs.S1-S3.The device was equipped with Cu-Kα radiation (λ = 1.5418Å), and the measurement range was set from 5˚ to 60˚ (2θ) with an optimal step size of 0.02˚ per 2θ.The generator voltage and the tube current were assigned to 30 kV and 10 mA, respectively (Jeong et al., 2020;Maheswaran et al., 2016).The goniometer radius used was 141 mm.To mitigate the effects of X-ray scattering at a low angle range, a slit with a size of 0.1 mm was inserted on the generator side.The clinker samples were rotated to improve particle statistics (Shi et al., 2017;Sun & Vollpracht, 2018).
TOPAS software version 7.0 (Bruker Co. Ltd., Land Baden-Württemberg, Germany) was used to conduct the QXRPD analysis of the clinkers.The clinkers were measured three times for one sample in the form of lightly pressed powder for assuring the reliability of measurement.The background of the XRPD pattern was determined with the Chebyshev polynomial function and 1/X term in the TOPAS software.The scale factors, lattice parameters, and full-width at half-maximum (FWHM) factors were carefully refined to stabilise the Rietveld refinement.In particular, except for the C 3 S, C 2 S, and C 4 AF phases, the FWHM profile parameters were constrained with the value established in the previous literature (Chaix-Pluchery et al., 1987;De la Torre & Aranda, 2003;Hazen, 1976b;Mondal & Jeffery, 1975;Smyth, 1975).March-Dollase preferred orientation corrections were refined to the phases of C 3 S (M3) and olivine (Dollase, 1986).The simulated diffraction patterns for ARC7 and WGC 7 are presented in Fig. 1a and b, respectively.
To quantify the exact amount of the crystalline phase and the amorphous phase contained in the clinkers, the internal standard method was adopted (Sakai et al., 2005;Shanahan & Zayed, 2007).The internal standard and the clinker samples were mixed with a specific weight ratio (internal standard reference: clinker powder = 1:9) by hand mixing for 30 min, with a mortar.The calibration process for the weight fractions was calculated according to Eqs. (1-2) (Guirado et al., 2000): where Corr(W a ), STD known , STD measured , and W amorphous represent the corrected weight percentage, weighed concentration of the standard, analysed concentration of the standard, and weight percentage of amorphous material, respectively.Zincite (NIST SRM 674b) was used (1) as the internal standard reference (Chrysochoou et al., 2010;Gualtieri et al., 2012;Snellings et al., 2010).The QXRPD results of ARC, DGC, and WGC are presented in Tables 2, 3, and 4, respectively.Furthermore, the effect of inclusion of C 3 S polymorphs (M3, triclinic [T1], and gamma [γ]) on the QXRPD analytical result of all samples of ARC, DGC, and WGC was investigated.Since the powder diffraction peaks corresponding to the C 3 S polymorphs are severely overlapped in the 32.1˚ and 32.6˚ for 2θ, it is challenging to accurately identify the existence of minor C 3 S polymorphs (T1 and γ) in the clinker (Bigare et al., 1967;Dunstetter et al., 2006;Jansen et al., 2011).Therefore, the minor C 3 S polymorphs were generally excluded in qualitative XRPD analysis (Jansen et al., 2011;Snellings et al., 2014;Whitfield & Mitchell, 2009).However, in this study, QXRPD analysis has been conducted twice (i.e. with the minor C 3 S polymorphs and without the polymorphs).The QXRPD analysis result without the minor C 3 S polymorphs is shown in Tables 2, 3 and 4. The analytical results with the inclusion of C 3 S polymorphs of T1 and γ have been presented in Additional file 1: Tables S1-S3.

Effects on the Particle Size Distribution of Clinker Powder According to Dry-or Wet-Grinding Program
The particle size distributions of all investigated clinkers are presented in Fig. 2(a-i).In the case of ARC, the particle size distribution is wider than the other samples.
It was observed that the particle size range was between 3 μm and 395 μm, and it exhibited 2 peaks.Furthermore, the median particle size of ARC ranged from 20 μm to 60 μm.For the DGC samples, a narrow particle size distribution ranging from 1 μm to 8 μm was observed, with a median particle size of about 3 μm.Meanwhile, in the case of the WGC sample, a broad particle size distribution was observed, and two peaks were identified.However, a significant particle size reduction effect was noted when compared to that of the ARC.The median particle size for these samples was around 10 μm.Therefore, a substantial reduction in particle size was clearly confirmed in all clinkers, regardless of the type of grinding program.Interestingly, differences were observed in the particle size depending on the type of grinding program (Fig. 3).The particle size distributions of the DGC were smaller than those of the WGC.This trend suggests that highvolume energy supplied by the ball mill can directly transfer to the clinker powder, and thus the particles could be finely pulverised with dry-grinding process.In addition, it should be noted here that the particle size distribution of the WGC was wider than that of the DGC (Fig. 2).In particular, in the case of the WGC, it was confirmed that there were significant amounts of particles between 0.2 μm and 1.0 μm, which indicated as a 2nd mode size (Figs. 2 and 3).This phenomenon might occur due to the difference in brittleness according to the type of clinker phase.In other words, it was shown that the particles containing a number of clinker phases with relatively low micro-structural stability were partially and more finely pulverised in a state of saturation with methanol (Le Saoût et al., 2011).

Qualitative X-ray Diffraction Analysis Results
According to the Inclusion of C 3 S Polymorphs C 3 S has known to have seven different polymorphs (i.e.M3, M2, M1, T3, T2, T1, and γ) (Bigare et al., 1967;Dunstetter et al., 2006).However, all the polymorphs are not usually considered in the QXRPD analysis with a medium resolution X-ray source (De Noirfontaine et al., 2006; Ren Hydrogrossular (Basso, 1983;Ferro et al., 2003) 1.1 (0.1) 1.6 (0.1) 1.6 (0.7) 0.9 (0.8) Olivine (Hazen, 1976b;Smyth, 1975) 3.0 (0.1) ) also exist in that range.Therefore, it is challenging to quantify the existing amounts of the polymorphs of C 3 S and C 2 S. In this regard, the effect of the inclusion of C 3 S polymorphs on the QXRPD result should be tested for optimal interpretation of subsequent measurements.From the conducted QXRPD of the all nine clinkers, the averaged values of weighted profile R-factor (R WP ) were compared in Fig. 4. When the calculated R WP indices are about 10.0, it can be considered that the result of QXRPD is reliable (Toby, 2006).The calculated R WP values of QXRPD considering three types of C 3 S (M3, T1, and γ) were lower than that of QXRPD considering only one polymorph of C 3 S (M3) for all three grinding conditions (ARC, DGC, and WGC).It shows that less accurately refined C 3 S area can underestimate the total amount of C 3 S compared to the case where only one polymorph was considered during the QXRPD analysis.
When the C 3 S polymorphs (T1 and γ) are included in the QXRPD analysis, the amount of C 3 S (γ) was quite high in the WGC (Table 4 and Additional file 1: Table S3).However, it was reported that the negligible amount of C 3 S (γ) generally exists in a clinker owing to low thermodynamic stability at very high temperature (De Noirfontaine et al., 2006;Dunstetter et al., 2006).This abnormal phenomenon also led to underestimate the amount of C 2 S. Simply, the overestimation of C 3 S (γ) makes the actual amount of C 2 S underestimated.Although the R WP indices of QXRPD results without C 3 S polymorphs were higher than those analysed with C 3 S polymorphs, more reasonable results were derived when the C 3 S polymorphs were excluded in the analysis.The higher R WP value of the QXRPD results of WGC without C 3 S Hydrogrossular (Basso, 1983;Ferro et al., 2003) 2.1 (0.2) 0.5 (0.5) 0.5 (0.4) 0.9 (0.4) Olivine (Hazen, 1976b;Smyth, 1975) 1.0 (1.1) polymorphs can be caused by the existence of incomplete C 3 S (M3) superstructure (Courtial et al., 2003;De La Torre et al., 2002;Nishi et al., 1985;Snellings et al., 2014).

Effects of Dry-or Wet-Grinding Program on QXRPD Analysis
As can be seen in Fig. 5a, an unusually high C 3 S peak was observed at 32.1˚ for 2θ in the XRPD pattern of the ARC, unlike in the previous studies (De La Torre et al., 2002;Mumme, 1995;Stutzman et al., 2016).& Aranda, 2003;Snellings et al., 2014).As a result, the amorphous content calculated in the ARC using QXRPD ranges from 0.2 wt.% to 23.4 wt.% (Table 2).There was no correlation between particle size and the calculated amount of amorphous content in the ARC.To investigate micro-grinding effect on the mineralogical properties of clinkers, grinding time should be chosen.Based on several preliminary tests, it was confirmed that 2 h of grinding yield consistent amount of amorphous content in three identical measurements either by dry-or wetgrinding program.
The calculated amount of amorphous phase after 2 h of micro-grinding is presented in Fig. 6.A noticeable trend was observed depending on the selected grinding program.In the case of the DGC, the amount of amorphous phase substantially increased compared to the ARC case.On the other hand, it decreased to a negligible quantity in the WGC.These results can be explained as follows.When the dry-grinding program was additionally applied to the ARC for 2 h, it had a significant effect on the XRPD Pyrite (Bayliss, 1989;Buerger, 1937 measurement results.Overall, the intensity of almost all clinker phases got lower than those of the ARC (Fig. 5a).
In particular, the intensity of one of the main peaks (32.1° of C 3 S) was significantly reduced whilst the relative intensity was measured similarly as reported in the previous studies (Jansen et al., 2011;Stutzman et al., 2016).The reduction of the intensities of these XRPD peaks definitely affected the quantification analysis.The amount of amorphous phase surprisingly increased, which ranges from 18.7 wt.% to 30.9 wt.% (Fig. 6).This result can be explained by not only applying the high-volume energy delivered by the ball mill to particle size reduction but also by partially modifying the crystallographic information of the clinker phases (Huang et al., 2017;Mejdoub, Hammi, Khitouni, Suñol, & M'nif, 2017).That is, crystallographic defects such as the dislocation and vacancies of atoms could arise during the drygrinding program.These abnormal variations are sources that could change the crystalline phase to the amorphous phase (Dongxu et al., 2000;Juhász & Opoczky, 1990;Mejdoub, Hammi, Khitouni, Suñol, & M'nif, 2017;Snellings et al., 2014).In particular, it was confirmed that the increased amount of amorphous phase was remarkably correlated with the decreased amount of C 3 S (Fig. 7).As the prolonged dry-grinding program was introduced, the preferred orientation of C 3 S (M3) to the (0 0 3) plane was substantially reduced which could be confirmed by the increased March-Dollase coefficient value (i.e.fitted to 0.91 for DGC).Therefore, it is safe to conclude that C 3 S, which has relatively lower spatial stability (Le Saoût et al., 2011), substantially contributed to the observed amorphisation of clinkers.
When wet-grinding program was applied to the ARC, well-corrected preferred orientation effect of C 3 S was observed, and it was also confirmed by identically measured XRPD patterns for the same sample.The obtained March-Dollase coefficients for the ARC, DGC, and WGC were about 0.84, 0.91, and 0.96, respectively.In the case of WGC, the biased crystal growth with specific direction was hard to be observed (i.e. the fitted coefficient was almost close to 1.0 (Whitfield & Mitchell, 2009)).Furthermore, the broadness of the XRPD peaks of the WGC was wider than those of the ARC with lowered intensities.This result can be well-explained by the reductions of the particle size as well as the crystallite size of the clinker phases (Pourghahramani & Akhgar, 2015;Uvarov & Popov, 2007).
However, in the case of C 4 AF (i.e. an interstitial mineral), the opposite result was observed by the wet grinding.When the crystallite size decreased, the peak broadened and as such the intensity should have been lowered.However, the intensity was measured to be rather high for the C 4 AF.The obtained results originated from the high linear absorption coefficient of C 4 AF ().Guirado et al., 2000;Peterson et al., 2006In the case of C 4 AF, the linear absorption coefficient is high at about  7 Relation between C 3 S and the amorphous phase according to the dry-grinding program 500 cm −1 , but for other clinker phases, it is mostly low at 300 cm −1 .Accordingly, in the case of C 4 AF, which has a large crystallite size, the amount could be underestimated because it absorbs a relatively large amounts of irradiated X-rays.That is, the effect of reducing the degree of absorption of X-rays is more dominant than the effect of decreasing the peak intensity owing to the decrease in the crystallite size (De la Torre & Aranda, 2003).Therefore, the phenomenon observed in WGC (i.e.particle size reduction without causing the amorphisation effect), was possibly owing to the reduced preferred orientation of the C 3 S (M3) phase, and the mitigated effect of linear absorption coefficient of C 4 AF in finer size.Therefore, unlike the QXRPD analysis of the ARC and DGC, the amorphous content was substantially reduced in the results of the WGC (Fig. 8).

The Crystallographic Variations of C 3 S (M3) and C 4 AF Phases Induced by the Grinding Program
The averaged crystallite size can be calculated based on the value corresponding to the FWHM or the integral breadth of the XRPD peak.In general, as the peak intensity decreases and broadens, the crystallite size decreases (Pourghahramani & Akhgar, 2015;Uvarov & Popov, 2007).These shape changes of C 3 S (M3) and C 4 AF could be identified in Fig. 9a and b.More quantitatively, the variation of calculated FWHM by grinding program was analysed using peaks at 29.4˚ and 12˚ for C 3 S (M3) and C 4 AF, respectively (Fig. 9c).Under the dry-grinding program, no noticeable change of the FWHM was confirmed.However, the value of C 3 S (M3) was significantly increased with the wet-grinding program.The increase in FWHM of C 3 S (M3) indicated that the existing preferred orientation was appropriately corrected in the WGC, which had a bias concerning crystallinity in the [0 0 1] direction in ARC.On the other hand, the FWHM of C 4 AF slightly increased in the DGC but decreased in WGC.The result of WGC found herein was not an effect normally caused by the grinding process (Pourghahramani & Forssberg, 2006;Radoi et al., 2004;Sheng et al., 2011).It can be explained by the micro-absorption coefficient of C 4 AF.When the wet-grinding program was applied, the characteristics of X-ray absorption were significantly alleviated, and as a result, the decrease in particle size had a remarkable effect on the XRPD patterns (De la Torre & Aranda, 2003;Snellings et al., 2014).Therefore, unlike the C 3 S (M3) phase, the C 4 AF phase would not have been able to confirm the variations in the fine FWHM by the effect of pattern change due to the correction of micro-absorption by wet-grinding.

Conclusions
This study systematically investigated the nature of amorphous content in clinkers under different grinding programs, an issue still under debate.The relationship between the amorphous phase contained in the clinkers and the two types of grinding methodologies was covered through particle size measurement and QXRPD analysis.The conclusion of this study is as follows: When the C 3 S (M3, T1, and γ) polymorphs were included in the QXRPD analysis, the improved R WP indices were obtained.In particular, it is mostly enhanced in the WGC.However, it was due to the misfitted C 3 S (γ) when all C 3 S polymorphs were included.It led to the overestimated amount of C 3 S(γ) and unrealistic amount of C 2 S. Therefore, it was proposed that QXRPD without considering the C 3 S polymorphs is more reliable method and it was adopted for the subsequent analyses.
In the case of the DGC, the median particle size reduction to 3 μm resulted in a significant change in the QXRPD analysis results.It was confirmed that the intensity of the XRPD peaks was lowered, suggesting the amorphisation effect of clinker phases.In particular, a remarkable amorphisation effect of C 3 S was confirmed.
Meanwhile, the wet-grinding program produced the median particle size of 10 μm.The correction of preferred orientation effect was supported by the crystallite size reduction of the C 3 S (M3) phase in the WGC.In addition, the amount of C 4 AF could be accurately calculated in the WGC owing to the particle size reduction.That is, the reduced degree of absorption of X-rays as the particle size decreases favourably contributed to the optimal estimation of C 4 AF in WGC.There was negligible amount of amorphous phase estimated in the WGC.On the other hand, substantial amorphisation including C 4 AF was observed in DGC and the amount of C 4 AF was underestimated in ARC with large particle sizes.
The implementation of the method demonstrated in this study suggests that the wet-grinding program can be an adequate sample treatment methodology for obtaining more accurate mineralogical composition of clinker materials.In specific, it can lead to the negligible amount of amorphous phase and the more accurate estimation of C 4 AF amount.Along with the application of this methodology to clinkers, it can be extended to quantitatively analyze the complex hydration reaction of PC especially focusing on the degree of reaction of C 3 S and C 4 AF.

Fig. 1
Fig. 1 Rietveld refinement result of XRPD pattern of ARC7 (a) and WGC 7 (b) It might be caused by the preferred orientation effect [i.e.reflected by a March-Dollase coefficient of 0.84 for all ARC of C 3 S (M3)] originating from the large crystallite size of C 3 S, in which the crystals were preferably developed in the [0 0 1] direction.It can yield a partial error in quantification of phases at the QXRPD analysis stage (De la Torre

Fig. 5 Fig. 6
Fig. 5 Comparison of XRPD patterns of the ARC7 and DGC7 (a) and the ARC7 and WGC7 (b).Symbols A, B, F, and T represent C 3 S, C 2 S, C 4 AF, and C 3 A, respectively

Fig. 8 Fig. 9
Fig. 8 Relation amongst C 3 S, C 4 AF and amorphous phases according to the wet-grinding program

Table 1
X-ray fluorescence results for clinkers

Table 2
QXRPD analysis results of ARCThe averaged weight percentage and standard deviation of three analysed results are presented

Table 3
QXRPD analysis results of DGCThe averaged weight percentage and standard deviation of three analysed results are presented

Table 4
QXRPD analysis results of WGCThe averaged weight percentage and standard deviation of three analysed results are presented