- Email: [email protected]
Structural and chemical changes in kaolinite caused by flash calcination: Formation of spherical particles
Applied Clay Science 114 (2015) 247–255
Contents lists available at ScienceDirect
Applied Clay Science journal homepage: www.elsevier.com/locate/clay
Research paper
Structural and chemical changes in kaolinite caused by flash calcination: Formation of spherical particles M. Claverie a,⁎, F. Martin a, J.P. Tardy b, M. Cyr c, P. De Parseval d, O. Grauby e, C. Le Roux a a
ERT Géo-matériaux 1074, GET UMR 5563 CNRS-UPS, OMP, 14 Avenue Edouard Belin, 31400 Toulouse, France Argeco Developpement, Rue Fournier Gorre, 47500 Fumel, France LMDC INSA-UPS Génie Civil 135, Avenue de Rangueil, 31077 Toulouse Cedex 4, France d GET, 14 Avenue Edouard Belin, 31400 Toulouse, France e CINaM, Campus de Luminy, Case 913, 13288 Marseille Cedex 9, France b c
a r t i c l e
i n f o
Article history: Received 6 March 2015 Received in revised form 26 May 2015 Accepted 27 May 2015 Available online 23 June 2015 Keywords: Metakaolin Flash calcination Kaolinite Spherical particles
a b s t r a c t To understand the morphological changes of three commercial kaolins during flash calcination and to compare them with those obtained during traditional heat treatments in the laboratory (an electric furnace at 700 °C for 5 h), this paper presents the physical and chemical characteristics of metakaolins obtained from an industrial flash calciner. In the metakaolin products, kaolinite was not completed dehydroxylated during calcination, and the proportion of untransformed kaolinite was greater in flashed metakaolins than in traditional rotarycalcined metakaolins. Several particle morphologies were discernible in the metakaolins, including spherical particles that were formed in flash calcinations. These spherical particles were cut with a focused ion beam (FIB) and were revealed to contain a vitrified aluminum silicate phase with traces of mullite and various gases. These spherical particles were produced from the direct calcination of several submicron kaolinites near the flame of the calciner. © 2015 Elsevier B.V. All rights reserved.
1. Introduction The calcination of clay minerals has attracted much interest due to their widespread use in a diverse range of applications (paints, rubber, render, plastics etc.) (Harvey et al., 2013). The calcination of kaolins (K) (including kaolinite, quartz and subsidiary materials) produces materials having pozzolanic activity, i.e., metakaolins (MK). Thermal decomposition of kaolinite (Kaol) has been the subject of many studies over the years (Insley and Ewell, 1935; Brindley and Nakahira, 1958). The calcination of Kaol at temperatures between 700 °C and 850 °C generates metakaolinite (MKaol) (Brindley et al., 1967; White et al., 2010; Heller-Kallai, 2013). This heating causes the dehydroxylation of Kaol and leads to the production of MKaol, a material that appears amorphous under X-ray diffraction. Typically, two independent processes occur during calcination: (1) the continuous loss of physisorbed water (dehydration) according the reaction H2 O (l, free) → H2O (g) at 100 °C (Teklay et al., 2015); and (2) the discontinuous loss of structural hydroxyls (dehydroxylation) according the reaction Al2Si2O5(OH)4 (Kaol) → Al2Si2O7 (MKaol) + 2H2O between 450 and 750 °C (Salvador, 1995; Sperinck, 2010; Teklay et al., 2015). MKaol has good properties as a mineral additive in applications such as in cement-based materials (Garcia-Diaz, 1995; He et al., 1995; Barbosa ⁎ Corresponding author. E-mail address: [email protected] (M. Claverie).
http://dx.doi.org/10.1016/j.clay.2015.05.031 0169-1317/© 2015 Elsevier B.V. All rights reserved.
et al., 2000; Cyr, 2000). MK (i.e., a trade name referring to a mixture of MKaol, quartz and other minerals) possess a huge reactive potential in basic environments San Nicolas et al., 2013) and have a SiO2/Al2O3 ratio between 1.5 and 2.5, depending on the purity of the raw K material used in their production. On an industrial scale, MK is obtained from calcination in a rotary kiln for several hours between 650 °C and 700 °C. Another industrial process was introduced in MK production: flash calcination (Davies, 1984; Salvador, 1992; Meinhold et al., 1994). Flash calcination enables the dehydroxylation of powdered Kaol within several tenths of a second, while traditional soak-calcination requires minutes at least. In this process, a solid, usually in a finely divided form, is rapidly heated, held at an elevated temperature (N700 °C) for a short time, and subsequently, swiftly cooled (San Nicolas et al., 2013). The aim of this paper is to highlight the physical and chemical transformations between three commercial K after traditional calcination with a laboratory-scale electric furnace at 700 °C for 5 h and after flash calcination. 2. Background of flash calcination In this study, a flash calcination process used by one of the two French producers of MK was employed. After grinding and drying the clay, the raw material is preheated at 500 °C, allowing its dehydration, in a system of cyclonic separators using gases released from the
248
M. Claverie et al. / Applied Clay Science 114 (2015) 247–255
calcination chamber. Clay is then taken to a 200 μm sieve, then brought to a temperature of approximately 700 °C in a calcination chamber, i.e., a cyclonic oven. In this calciner (Fig. 1), the product swirls around the flame, which has a temperature of 1100 °C. Clay spends only a few tenths of a second in the calciner, just enough time for the transformation into MK (San Nicolas et al., 2013). Once this transformation has taken place, the MK is then separated from the hot gases in a cyclonic separator and then cooled in a system of successive cyclonic separators. Finally, the finished product is collected and packed in silos.
Table 1 Chemical composition (dry basis, wt.%) of the K samples. L.O.I: Loss on ignition (measured by thermogravimetry) Chemical analysis (%)
SiO2
K1 K2 K3
51.62 32.01 1.31 66.24 20.81 2.29 67.67 21.79 0.77
Al2O3 Fe2O3 TiO2 K2O
MgO CaO
1.48 0.06 0.13 1.04 0.21 0.15 0.35 1.26 0.14
H2O L.O.I. total⁎
Total
0.45 12.62 12.71 99.76 0.62 8.53 8.81 100.17 0.06 7.31 7.48 99.52
⁎ Analytical method: Titration Karl Fisher.
3. Materials and methods 3.1. Samples Three commercial K (with different origins, compositions and levels of purity) were studied in this paper. The crystallinity of Kaol in each clay was evaluated by different indices obtained by X-ray diffraction (XRD) (Aparicio and Galan, 1999). - K1 is a K from sedimentary deposits located at Fumel (South of France). The Kaol in this sample has a disordered structure. - K2 is a K from sedimentary deposits located at Fumel (South of France). The Kaol in this sample has a disordered structure. B2 is even less phase pure than K1. - K3 is a K collected at a quarry located in primary deposits developed in the North of France. This sample contains ordered Kaol.
Samples were named MKx for the metakaolins with x being the number of the K. 3.2. Characterization Bulk elemental composition (Table 1) was determined with inductively coupled plasma-optical emission spectrometry (fusion glass prepared with Pt crucibles along with LiBO 2 at 980 °C in an automatic tunnel oven, dissolved in a HNO3–H2O2 mixture and analyzed with ICP-OES, Thermo Fischer Icap 6500, SARM Laboratory of Nancy). XRD analysis was performed on dried K. The XRD patterns were recorded on G3000 INEL equipment (GET laboratory, University of Toulouse) using Cu-Kα radiation. Through these methods, including a Rietveld analysis (Rietveld, 1967, 1969) of the XRD data using the MAUD program (Lutterotti et al., 1999), the amounts of Kaol and other minerals in the K and MK were estimated (Bish and von Dreele,
1989). However, this quantification by Rietveld refinement is a semiquantitative estimate involving some uncertainties. The amount of mineral phases by Rietveld analysis was compared with the corresponding amount estimated by stoichiometry based on the results of the chemical analysis (Bich, 2005). Fourier transformed infrared spectra in the near-infrared region (NIRFTIR) spectra were recorded at a resolution of 4 cm−1 between 4000 and 11,000 cm−1 using a Nicolet 6700 FTIR spectrometer (HydrASA, IC2MP, University of Poitiers) with a smart NIR integrating sphere (CaF2 beam splitter and InGaAs detector). Particle size distributions were measured in ethanol dispersions using a laser analyzer CILAS 1090 (LMDC, University of Toulouse) under ultrasound polydispersion. The sizes determined from these measurements were likely too large due to the incorrect assumption of spherical particles (rather than plates) in the data analysis. An environmental scanning electron microscope equipped with an energy-dispersive X-ray spectrometer (SEM-EDX FEG JEOL 6700F, Centre de Microcaractérisation, Toulouse) was used for the morphological analysis of the surface structure and to determine the Al, Si, Fe, Ti, Ca and O elemental compositions of the particles. An accelerating voltage of 20 keV was used for the EDX microanalysis. Focused ion beam (FIB) thin section was extracted from spherical particle, using a dual beam SEM/FIB FEI Helios NanoLab 600 (septum; CP2M, Marseille). The FIB sectioning consisted of the following steps: (i) a protective platinum sheet of about 1.5 μm thickness was deposited upon the surface in the region of interest; (ii) trenches were then milled at 30 kV on both sides of this sheet to create a 300–500 nm thick foil (reducing current from 6 nA to 80 pA, while thinning); (iii) using a micromanipulator, in situ lift-out was performed by sputtering a platinum sheet to connect the micromanipulator tip with the TEM foil; (iv) the TEM foil was attached to a Cu TEM grid and a final FIB thinning and cleaning at 5 kV–40 pA were carried out in order to obtain a foil of less than 100 nm thick.
FIB section was then observed on a transmission electron microscope (TEM) JEOL 3010 at 300 kV (CINaM, Marseille) for bright-field and dark-field imaging and recording selected area electron diffraction (SAED) patterns. Energy-dispersive X-ray spectroscopy (EDX) analysis was achieved on both FIB section with a JEOL 2011 at 200 kV (CINaM, Marseille). 4. Results and discussion 4.1. Mineralogical characteristics
Fig. 1. Schematic representation of a flash calciner used in MK production (San Nicolas, 2011) and photography of the flash process tower.
Through chemical analyses (Table 1) and XRD (Fig. 2), including the Rietveld analysis of the XRD data (Table 2), the amounts of Kaol, quartz and muscovite in the K samples were estimated. The K1 has a high Kaol content at 89%, and the K2 and K3 have moderate Kaol contents of 56%
M. Claverie et al. / Applied Clay Science 114 (2015) 247–255
249
Fig. 2. X-ray diffraction profiles for (a) the three K and for (b) the six MK. “Labo” means “traditional calcined MK in laboratory”.
Table 2 Estimated phase composition of parent K (calculated from multi-phase Rietveld refinement of powder diffraction profiles) compared to the values deduced using assumed phase compositions and EDX elemental analysis. Kaolinite
Quartz
Muscovite
wt.%
Rietveld
Stoichiometry
Rietveld
Stoichiometry
Rietveld
Stoichiometry
K1 K2 K3
89 ± 1.12 56 ± 0.91 53 ± 0.99
88 55 49
11 ± 0.93 41 ± 1.08 35 ± 0.91
9 40 37
0 ± 0.02 1 ± 0.41 12 ± 0.63
0 2 12
and 53%, respectively. These results correlated with the mineralogical composition determined by stoichiometry using chemical analysis (Table 2). Quartz and muscovite were the main “impurities” (industry term) in these raw K. The X-ray diffractograms of the MK (Fig. 2b) show that both methods of calcination were efficient, as evidenced by the presence of X-ray amorphous halo focused on about 3.8 Å (indicative of MKaol). According to the mineralogical composition determined by the Rietveld analysis of MK (Table 3), unlike Kaol, the relative amounts of quartz and muscovite were unchanged by calcination treatment relative to the fed material. These results also showed that there was no significant difference in the mineralogical compositions between flash calcined MK and traditionally calcined MK. However, in the flash calcination process of MK1, mullite was formed. This mullite formation suggests a calcination temperature exceeding 950 °C (Brindley and Nakahira, 1958). This crystalline phase is generally considered to be unreactive in cement-based materials.
Mullite can be formed from particles that passed directly through the flame rather than swirling around the periphery.
4.2. Trace amounts of Kaol after flash calcination The XRD analysis revealed the presence of remnant Kaol in the three MK after flash calcination, unlike the traditionally calcined MK (Fig. 2b). This observation was confirmed by FTIR (Fig. 3). The infrared spectra of the flashed MK exhibited one small band at approximately 7065 cm−1, which is attributed to inner hydroxyl groups and other hydroxyl groups located on the exterior surface of the octahedral sheet of Kaol (Petit et al., 1999); these bands were proof that Kaol was present after flash calcinations. However, results from these analyses did not clarify whether this trace of Kaol corresponds to a Kaol particle unaffected in its core or to a complete larger particle of Kaol which was unaffected by the dehydroxylation step.
Table 3 Estimated phase composition of six MK (calculated from multi-phase Rietveld refinement of powder diffraction profiles). wt.% MK1 MK2 MK3
Flash Electric furnace in laboratory Flash Electric furnace in laboratory Flash Electric furnace in laboratory
Amorphous phase
Quartz
Muscovite
Kaolinite
Mullite
Anatase
84 ± 2.15 88 ± 1.89 55 ± 1.11 57 ± 0.82 49 ± 0.91 51 ± 1.12
10 ± 0.95 10 ± 0.88 41 ± 1.09 41 ± 1.08 37 ± 0.97 36 ± 0.97
0 ± 0.01 0 ± 0.01 1 ± 0.04 1 ± 0.04 11 ± 0.56 12 ± 0.85
1 ± 0.41 0 ± 0.03 2 ± 0.52 0 ± 0.01 2 ± 0.32 0 ± 0.01
3 ± 0.21 0 ± 0.01 0 ± 0.01 0 ± 0.01 0 ± 0.01 0 ± 0.01
2 ± 0.4 2 ± 0.14 1 ± 0.02 1 ± 0.01 1 ± 0.01 1 ± 0.01
250
M. Claverie et al. / Applied Clay Science 114 (2015) 247–255
Fig. 3. FTIR spectra of the K1, K2, K3 and their MK.
Additionally, trace amounts of Kaol can also be present in MK produced in an industrial rotary kiln (San Nicolas et al., 2013).
4.3. Physical characteristics The morphological characteristics of the natural clays obtained by SEM are shown in Fig. 4. K3 clays show the typical microstructure of Kaol, consisting of pseudo-hexagonal plates and clusters of plates (Tironi et al., 2014). These plates are larger (N 4 μm) and more structured compared with the other clay samples. In the K2 and K1 samples, the microstructures are present in fine particle sizes (b 1 μm), irregular Kaol forms and very small flakes. In Fig. 5a, the particle size distributions of the clays are shown. A particle size analysis of the three K and their MK (Fig. 5a) showed that calcination induced a decrease in the relative amount of finer particle sizes (approximately 4 μm for K1 and K2) in favor of coarser particles (approximately 50 μm). This phenomenon seems to be accentuated with flash calcination. Traditional and flash calcination of K3 does not lead to a significant increase in particle size unlike K1 and K2 (K3 shows a slight increase from 15 μm to 30 μm after flash calcination while particle size is unchanged after traditional calcination). This can be explained by the presence of coarser particles (approximately 15 μm) in K3 unlike K1 and K2 (richer in finer particle sizes). Thus,
K1
flash calcination seemed to favor the agglomeration of submicron particles. Several particle morphologies were discernible in the three MK samples (Fig. 5). Quartz (Fig. 5e, h) was accompanied by aggregates of aluminum silicate particles with diameters ≥ 20 μm (Fig. 5d, f, i) covered by nanometer-sized particles of aluminum silicates. A few of these nanoparticles were crystallized and preserved the layered hexagonal structures of Kaol (Fig. 5c, g), whereas other nanoparticles were observed to have amorphized layers (dulled contours) (Fig. 5b). However, observed spherical 10 μm particles (Fig. 6a) in the flash-calcined products were not present in the traditional MK samples. These spherical particles in flashed MK accounted for approximately 19% in MK1, 15% in MK2 and 13% in MK3 (by visual estimation on about 50 SEM investigations and in agreement with San Nicolas, 2011). A previous study had been conducted to determine the internal composition of these spherical particles using hydrochloric acid etching (San Nicolas et al., 2013).These spherical particles seemed to improve mortar blends by compacting the granular skeleton (Jillavenkatsea, 2001).The smooth surface of the sphere shown in the SEM image in Fig. 6a appears to be composed of interlinked thin particles (Fig. 6b). According to EDX analyses (Fig. 6d), the surfaces of these thin particles were composed of aluminum silicates, similar to MKaol or mullite. Moreover, on the surface of these spheres, obscured stains were observable in the SEM image (Fig. 6c). Because the same spectra were obtained from the EDX analyses (Fig. 6d) between the obscured
K2
1 µm
K3
1 µm Fig. 4. SEM images of Kaol (× 10,000).
M. Claverie et al. / Applied Clay Science 114 (2015) 247–255
251
Fig. 5. (a) Particle size distribution for K1, K2, K3 and their MK. Two groups of particle size (grey circles) were highlighted: Group I = submicronical particles in MK with (b), (c) TEM images and (g) electron diffraction pattern of the submicronical Kaol observable in (c). Group II = coarser particles in MK with (d), (e) and (f) SEM images. (h) and (i) are EDX analyses of different particles observed in MK1, MK2 and MK3.
zones and the surfaces of the sphere, it was assumed that these stains were indicative of cavities under the surface. 4.4. Structure of spherical particle In order to analyze the composition of a spherical particle, a focused ion beam was used to cut a spherical particle in half (Fig. 7a) to prepare samples for TEM. The nm-scale resolution of the FIB allowed us to choose the exact region for thinning. A transverse section of a sphere with a platinum coat is observable in Fig. 7b. The lower part of this section was concave, highlighting the presence of a substantial cavity
within the spherical particle. The presence of small bubbles was also noticed (Fig. 7b—Cavity). The top of this section (Fig. 7c) shows a combination of platinum and nanometer-sized aluminum silicate particles as shown in the EDX analysis (Fig. 7—EDX-MKaol). These particles were previously observed (Fig. 6b—EDX-A) and correspond to MKaol enriched with iron and titanium. Several analyses were carried out on the transverse section of the spherical particle (Fig. 7b). The halo diffraction pattern (Fig. 7—SAD-A, B and C) from within the spherical particle indicated its amorphous character. The EDX analysis of this amorphous material (Fig. 7—EDX-A,
252
M. Claverie et al. / Applied Clay Science 114 (2015) 247–255
Fig. 6. (a) SEM images (in flashed MK2) of spherical particle (× 3700) and (b) details of the surface (× 22,000); (c) SEM image (in flashed MK3) of spherical particle (x 9,000) and (d) EDX analyses: EDX-A analysis of thin particles; EDX-B analysis of the smooth surface of sphere; EDX-C analysis of dark stains and EDX-D analysis of clear zone.
B and C) shows that it was primarily aluminum silicate with significant amounts of titanium and iron oxides. However, this spherical particle was not exclusively composed of an amorphous phase. A variegated texture was also observed (Fig. 7d). The diffraction pattern (Fig. 7—Sad-D and E) and the EDX analysis (Fig. 7—EDX-D and E) showed that this area was composed of crystallized aluminum silicate, and unlike the amorphous area, this crystalline phase contained fewer iron, calcium and titanium elements. This observation was confirmed by the chemical mapping analyses (Fig. 8). The crystalline phase was dark for the iron, calcium and titanium mappings, which indicated low contents for these elements. Spherical particles were the seat of a dynamic diffusion of iron, calcium and titanium elements in favor of the amorphous area. In addition to these crystallized aluminum silicates, spherical areas were also observed in the spherical particles (Fig. 7e—EDX-F). The electronic diffraction patterns (Fig. 7—SAD-F) indicated the presence of mullite in these areas. In summary, the spherical particles were composed primarily of amorphous aluminum silicates with traces of crystallized and pure aluminum silicates. These particles had cavities and mullite zones. The surfaces of these polyphasic particles were covered with submicronic MKaol particles. In a second spherical particle approximately 5 μm in diameter cut by FIB (Fig. 9), a large cavity was observed and was surrounded by other smaller cavities. Based on these various analyses, it is predicted that these spherical particles originated from the agglomeration of submicron Kaol during
flash calcination. At a certain distance from the flame (Fig. 10a), the temperatures were still sufficiently high enough to dehydroxylate Kaol to form MKaol. Nearer the flame, two phenomena likely occurred: - a vitrification of submicron aluminum silicates produced droplets due to a minimization of the surface energy; and - MKaol recrystallized into mullite.
Within the flash calciner, the cyclonic air movement allowed the mixing of these different phases (Fig. 10b). During the cooling step, a few of the submicron MKaol particles adsorbed onto the surfaces of the spherical particles. The cavities in the spherical particles showed evidence of gases, which were caused either by gas emissions during combustion (i.e., gas trapped in silicate melt) or by gas produced inside the droplet during the calcination reactions. Studies have been conducted on mortars blended with flash MK compared to rotary-kiln MK. The spherical morphology seemed to improve mortar blends by compacting the granular skeleton (Jillavenkatsea, 2011), and enhancing their workability (San Nicolas et al., 2013). Furthermore, flash calcination technology “consumes 2.2 MJ/t of MK, which is 80% less that the energy consumed during cement production” (San Nicolas, 2011). Recently, Teklay et al. (2015) developed a model for clay particle calcination and confirmed that a “residence time of 0.5 s and temperatures of about 950 °C are likely to be the most appropriate condition to produce high-quality calcines”.
M. Claverie et al. / Applied Clay Science 114 (2015) 247–255
253
Fig. 7. (a) SEM image (in flashed MK1) of spherical particle cut in half (× 5000); (b) TEM image of the transverse section; (c) TEM image of the top of the section; (d) TEM image of the crystalline phase; (e) TEM image of mullite spheres; elemental quantifications (using the Cliff-Lorimer ratio technique) from EDX analysis and selected area electron diffraction patterns of different phases observed inside the spherical particle.
254
M. Claverie et al. / Applied Clay Science 114 (2015) 247–255
Fig. 8. Schematic representation (same sphere section than Fig. 7b) and elemental mappings of the section of the sphere (Fig. 7b) (Si, Al, Ti, Ca and Fe).
5. Conclusions
treated by both calcination processes and the resultant products were analyzed. The following conclusions can be drawn:
The purpose of this study was to compare the chemical, physical and morphological characteristics of flash calcined and traditional rotarycalcined metakaolins and to determine the contents of the spherical particles formed after flash calcination. Three kaolin sources were
PLATINUM
CAVITY
Fig. 9. SEM image of second sphere cut in half by FIB.
1. The two calcination techniques produced metakaolins without significant mineralogical or chemical differences. However, the agglomeration of submicron particles after calcination was accentuated in flash calcination. 2. Unlike the traditionally calcined metakaolins in the laboratory, kaolinite was present in metakaolins after flash calcination. However, trace amounts of kaolinite have been reported in traditionally calcined metakaolins in industry 3. Unlike traditional metakaolins, flash metakaolins contained spherical particles (between 10% and 20% in this study). These spherical particles account for the enhanced workability of mortars blended with flash metakaolins compared with mortars blended with rotary-kiln metakaolins. 4. These spherical particles were produced from the agglomeration of submicron kaolinites during flash calcination. These spherical particles were composed of gases and a mixture somewhat pure aluminum silicates with varied crystallinities. This heterogenic composition was primarily due to the presence of a temperature gradient and cyclonic air movement in the flash calciner. Future works should study ability for these spherical particles to dissolve and their effects on the properties of cement matrices. These spherical particles could also have potential as reservoirs capable of reacting with calcium hydroxide.
M. Claverie et al. / Applied Clay Science 114 (2015) 247–255
255
Fig. 10. (a) Schematic representation of phenomena taking place in a flash calciner and (b) schematic representation of the cyclonic air movement allowing the mix of different phases.
Acknowledgments The authors are grateful for the support provided by Argeco Développement and sincerely thank those who volunteered their time to participate in this study. We thank David Bish and the anonymous reviewers for their constructive comments. References Aparicio, P., Galan, E., 1999. Mineralogical interference on kaolinite crystallinity index measurements. Clay Clay Miner. 47, 12–27. Barbosa, V.F.F., MacKenzie, K.J.D., Thaumaturgo, C., 2000. Synthesis and characterisation of materials based on inorganic polymers of alumina and silica: sodium polysialate polymers. Int. J. Inorg. Mater. 2, 309–317. Bich, C., 2005. Contribution à l'étude de l'activation thermique du kaolin: evolution de la structure cristallographique et activité pouzzolanique PhD thesis. Institut National des Sciences Appliquées, Lyon, France. Bish, D.L., von Dreele, R.B., 1989. Rietveld refinement of non-hydrogen atomic positions in kaolinite. Clay Clay Miner. 37, 289–296. Brindley, G.W., Nakahira, M., 1958. A new concept of the transformation sequence of kaolinite to mullite. Nature 4619, 1333–1334. Brindley, G.W., Sharp, J.H., Nakahira, M., Patterson, J.H., 1967. Kinetics and mechanism of dehydroxylation processes. Am. Mineral. 52, 1697–1705. Cyr, M., 2000. Contribution à la caractérisation des fines minérales et à la compréhension de leur rôle joué dans le comportement rhéologique des matrices cimentaires PhD thesis. University of Sherbrooke, Canada. Davies, T.W., 1984. Equipment for the study of flash heating of particle suspension. High Temp. Technol. 2 (3), 6. Garcia-Diaz, A., 1995. Réactivité pouzzolanique des métakaolinites: corrélation avec les caractéristiques minéralogitologiques des kaolinites PhD thesis. Génie des Procédés. Alès, l'école des Mines, France. Harvey, C.C., Lagaly, G., 2013. Industrial application Chapter 4.2. Handbook of Clay Science 5B pp. 451–490. He, C., Makovicky, E., Osbaeck, B., 1995. Pozzolanic reactions of six principal clay minerals: activation, reactivity assessments and technological effects. Cem. Concr. Res. 25 (8), 1691–1702.
Heller-Kallai, L., 2013. Thermally modified clay minerals Chapter 10.2. Handbook of Clay Science 5A pp. 411–427. Insley, H., Ewell, R., 1935. Thermal behavior of the kaolins minerals. J. Res. Natl. Bur. Stand. 14, 331–335. Jillavenkatsea, A., Dapkunas, S.J., Lum, L.H., 2001. Particle size characterisation. NIST Recommend. Prac. Guide 906 164 pp. Lutterotti, L., et al., 1999. MAUD (Material Analysis Using Diffraction): a user friendly Java Program for Rietveld Texture Analysis and more. Proceeding of the Twelfth International Conference on Textures of Materials. 1, p. 1559. Meinhold, R.H., Salvador, S., Davies, T.W., Slade, R.T.C., 1994. A comparison of kinetics of flash calcination of kaolinite in different calciners. Chem. Eng. Res. Des. 72 (8). Petit, S., Madejova, J., Decarreau, A., Martin, F., 1999. Characterization of octahedral substitutions in kaolinites using near infrared spectroscopy. Clay Clay Miner. 47, 103–108. Rietveld, H.M., 1967. Line profiles of neutron powder-diffraction peaks for structure refinement. Acta Cristallographica 22, 151–152. Rietveld, H.M., 1969. A profile refinement method for nuclear structures. J. Appl. Cristallogr. 2, 65-57. Salvador, S., 1992. Production de pouzzolanes de synthèse par calcination flash de sols argileux; étude des produits et conception d'une installation PhD thesis,. Institut National des Sciences Appliquées, Mécanique, Toulouse, p. 230. Salvador, S., 1995. Pozzolanic properties of flash-calcined kaolinite: a comparative study with soak-calcined products. Cem. Concr. Res. 25, 102–112. San Nicolas, R., 2011. Approche performantielle des bétons avec métakaolins obtenus par calcination flash PhD thesis. University of Toulouse, France. San, Nicolas R., Cyr, M., Escadeillas, G., 2013. Characteristics and applications of flash metakaolins. Appl. Clay Sci. 83–84, 253–262. Sperinck, S., Raiteri, P., Marks, N., Wright, K., 2010. Dehydroxylation of kaolinite to metakaolin—a molecular dynamics study. J. Mater. Chem. 21, 2118–2125. Teklay, A., Yin, C., Rosendahl, L., Lindequist, Køhler L., 2015. Experimental and modeling study of flash calcination of kaolinite rich clay particles in a gas suspension calciner. Appl. Clay Sci. 103, 10–19. Tironi, A., Castellano, C.C., Bonavetti, V., Trezza, M.A., Scian, A.N., Irassar, E.F., 2014. Kaolinic calcined clays–cement system: hydration and properties. Constr. Build. Mater. 64, 615–621. White, C.E., Provis, J.L., Proffen, T., Riley, D.P., van Deventer, J.S.J., 2010. Density functional modeling of the local structure of kaolinite subjected to thermal dehydroxylation. J. Phys. Chem. A 114, 4988–4996.
Contents lists available at ScienceDirect
Applied Clay Science journal homepage: www.elsevier.com/locate/clay
Research paper
Structural and chemical changes in kaolinite caused by flash calcination: Formation of spherical particles M. Claverie a,⁎, F. Martin a, J.P. Tardy b, M. Cyr c, P. De Parseval d, O. Grauby e, C. Le Roux a a
ERT Géo-matériaux 1074, GET UMR 5563 CNRS-UPS, OMP, 14 Avenue Edouard Belin, 31400 Toulouse, France Argeco Developpement, Rue Fournier Gorre, 47500 Fumel, France LMDC INSA-UPS Génie Civil 135, Avenue de Rangueil, 31077 Toulouse Cedex 4, France d GET, 14 Avenue Edouard Belin, 31400 Toulouse, France e CINaM, Campus de Luminy, Case 913, 13288 Marseille Cedex 9, France b c
a r t i c l e
i n f o
Article history: Received 6 March 2015 Received in revised form 26 May 2015 Accepted 27 May 2015 Available online 23 June 2015 Keywords: Metakaolin Flash calcination Kaolinite Spherical particles
a b s t r a c t To understand the morphological changes of three commercial kaolins during flash calcination and to compare them with those obtained during traditional heat treatments in the laboratory (an electric furnace at 700 °C for 5 h), this paper presents the physical and chemical characteristics of metakaolins obtained from an industrial flash calciner. In the metakaolin products, kaolinite was not completed dehydroxylated during calcination, and the proportion of untransformed kaolinite was greater in flashed metakaolins than in traditional rotarycalcined metakaolins. Several particle morphologies were discernible in the metakaolins, including spherical particles that were formed in flash calcinations. These spherical particles were cut with a focused ion beam (FIB) and were revealed to contain a vitrified aluminum silicate phase with traces of mullite and various gases. These spherical particles were produced from the direct calcination of several submicron kaolinites near the flame of the calciner. © 2015 Elsevier B.V. All rights reserved.
1. Introduction The calcination of clay minerals has attracted much interest due to their widespread use in a diverse range of applications (paints, rubber, render, plastics etc.) (Harvey et al., 2013). The calcination of kaolins (K) (including kaolinite, quartz and subsidiary materials) produces materials having pozzolanic activity, i.e., metakaolins (MK). Thermal decomposition of kaolinite (Kaol) has been the subject of many studies over the years (Insley and Ewell, 1935; Brindley and Nakahira, 1958). The calcination of Kaol at temperatures between 700 °C and 850 °C generates metakaolinite (MKaol) (Brindley et al., 1967; White et al., 2010; Heller-Kallai, 2013). This heating causes the dehydroxylation of Kaol and leads to the production of MKaol, a material that appears amorphous under X-ray diffraction. Typically, two independent processes occur during calcination: (1) the continuous loss of physisorbed water (dehydration) according the reaction H2 O (l, free) → H2O (g) at 100 °C (Teklay et al., 2015); and (2) the discontinuous loss of structural hydroxyls (dehydroxylation) according the reaction Al2Si2O5(OH)4 (Kaol) → Al2Si2O7 (MKaol) + 2H2O between 450 and 750 °C (Salvador, 1995; Sperinck, 2010; Teklay et al., 2015). MKaol has good properties as a mineral additive in applications such as in cement-based materials (Garcia-Diaz, 1995; He et al., 1995; Barbosa ⁎ Corresponding author. E-mail address: [email protected] (M. Claverie).
http://dx.doi.org/10.1016/j.clay.2015.05.031 0169-1317/© 2015 Elsevier B.V. All rights reserved.
et al., 2000; Cyr, 2000). MK (i.e., a trade name referring to a mixture of MKaol, quartz and other minerals) possess a huge reactive potential in basic environments San Nicolas et al., 2013) and have a SiO2/Al2O3 ratio between 1.5 and 2.5, depending on the purity of the raw K material used in their production. On an industrial scale, MK is obtained from calcination in a rotary kiln for several hours between 650 °C and 700 °C. Another industrial process was introduced in MK production: flash calcination (Davies, 1984; Salvador, 1992; Meinhold et al., 1994). Flash calcination enables the dehydroxylation of powdered Kaol within several tenths of a second, while traditional soak-calcination requires minutes at least. In this process, a solid, usually in a finely divided form, is rapidly heated, held at an elevated temperature (N700 °C) for a short time, and subsequently, swiftly cooled (San Nicolas et al., 2013). The aim of this paper is to highlight the physical and chemical transformations between three commercial K after traditional calcination with a laboratory-scale electric furnace at 700 °C for 5 h and after flash calcination. 2. Background of flash calcination In this study, a flash calcination process used by one of the two French producers of MK was employed. After grinding and drying the clay, the raw material is preheated at 500 °C, allowing its dehydration, in a system of cyclonic separators using gases released from the
248
M. Claverie et al. / Applied Clay Science 114 (2015) 247–255
calcination chamber. Clay is then taken to a 200 μm sieve, then brought to a temperature of approximately 700 °C in a calcination chamber, i.e., a cyclonic oven. In this calciner (Fig. 1), the product swirls around the flame, which has a temperature of 1100 °C. Clay spends only a few tenths of a second in the calciner, just enough time for the transformation into MK (San Nicolas et al., 2013). Once this transformation has taken place, the MK is then separated from the hot gases in a cyclonic separator and then cooled in a system of successive cyclonic separators. Finally, the finished product is collected and packed in silos.
Table 1 Chemical composition (dry basis, wt.%) of the K samples. L.O.I: Loss on ignition (measured by thermogravimetry) Chemical analysis (%)
SiO2
K1 K2 K3
51.62 32.01 1.31 66.24 20.81 2.29 67.67 21.79 0.77
Al2O3 Fe2O3 TiO2 K2O
MgO CaO
1.48 0.06 0.13 1.04 0.21 0.15 0.35 1.26 0.14
H2O L.O.I. total⁎
Total
0.45 12.62 12.71 99.76 0.62 8.53 8.81 100.17 0.06 7.31 7.48 99.52
⁎ Analytical method: Titration Karl Fisher.
3. Materials and methods 3.1. Samples Three commercial K (with different origins, compositions and levels of purity) were studied in this paper. The crystallinity of Kaol in each clay was evaluated by different indices obtained by X-ray diffraction (XRD) (Aparicio and Galan, 1999). - K1 is a K from sedimentary deposits located at Fumel (South of France). The Kaol in this sample has a disordered structure. - K2 is a K from sedimentary deposits located at Fumel (South of France). The Kaol in this sample has a disordered structure. B2 is even less phase pure than K1. - K3 is a K collected at a quarry located in primary deposits developed in the North of France. This sample contains ordered Kaol.
Samples were named MKx for the metakaolins with x being the number of the K. 3.2. Characterization Bulk elemental composition (Table 1) was determined with inductively coupled plasma-optical emission spectrometry (fusion glass prepared with Pt crucibles along with LiBO 2 at 980 °C in an automatic tunnel oven, dissolved in a HNO3–H2O2 mixture and analyzed with ICP-OES, Thermo Fischer Icap 6500, SARM Laboratory of Nancy). XRD analysis was performed on dried K. The XRD patterns were recorded on G3000 INEL equipment (GET laboratory, University of Toulouse) using Cu-Kα radiation. Through these methods, including a Rietveld analysis (Rietveld, 1967, 1969) of the XRD data using the MAUD program (Lutterotti et al., 1999), the amounts of Kaol and other minerals in the K and MK were estimated (Bish and von Dreele,
1989). However, this quantification by Rietveld refinement is a semiquantitative estimate involving some uncertainties. The amount of mineral phases by Rietveld analysis was compared with the corresponding amount estimated by stoichiometry based on the results of the chemical analysis (Bich, 2005). Fourier transformed infrared spectra in the near-infrared region (NIRFTIR) spectra were recorded at a resolution of 4 cm−1 between 4000 and 11,000 cm−1 using a Nicolet 6700 FTIR spectrometer (HydrASA, IC2MP, University of Poitiers) with a smart NIR integrating sphere (CaF2 beam splitter and InGaAs detector). Particle size distributions were measured in ethanol dispersions using a laser analyzer CILAS 1090 (LMDC, University of Toulouse) under ultrasound polydispersion. The sizes determined from these measurements were likely too large due to the incorrect assumption of spherical particles (rather than plates) in the data analysis. An environmental scanning electron microscope equipped with an energy-dispersive X-ray spectrometer (SEM-EDX FEG JEOL 6700F, Centre de Microcaractérisation, Toulouse) was used for the morphological analysis of the surface structure and to determine the Al, Si, Fe, Ti, Ca and O elemental compositions of the particles. An accelerating voltage of 20 keV was used for the EDX microanalysis. Focused ion beam (FIB) thin section was extracted from spherical particle, using a dual beam SEM/FIB FEI Helios NanoLab 600 (septum; CP2M, Marseille). The FIB sectioning consisted of the following steps: (i) a protective platinum sheet of about 1.5 μm thickness was deposited upon the surface in the region of interest; (ii) trenches were then milled at 30 kV on both sides of this sheet to create a 300–500 nm thick foil (reducing current from 6 nA to 80 pA, while thinning); (iii) using a micromanipulator, in situ lift-out was performed by sputtering a platinum sheet to connect the micromanipulator tip with the TEM foil; (iv) the TEM foil was attached to a Cu TEM grid and a final FIB thinning and cleaning at 5 kV–40 pA were carried out in order to obtain a foil of less than 100 nm thick.
FIB section was then observed on a transmission electron microscope (TEM) JEOL 3010 at 300 kV (CINaM, Marseille) for bright-field and dark-field imaging and recording selected area electron diffraction (SAED) patterns. Energy-dispersive X-ray spectroscopy (EDX) analysis was achieved on both FIB section with a JEOL 2011 at 200 kV (CINaM, Marseille). 4. Results and discussion 4.1. Mineralogical characteristics
Fig. 1. Schematic representation of a flash calciner used in MK production (San Nicolas, 2011) and photography of the flash process tower.
Through chemical analyses (Table 1) and XRD (Fig. 2), including the Rietveld analysis of the XRD data (Table 2), the amounts of Kaol, quartz and muscovite in the K samples were estimated. The K1 has a high Kaol content at 89%, and the K2 and K3 have moderate Kaol contents of 56%
M. Claverie et al. / Applied Clay Science 114 (2015) 247–255
249
Fig. 2. X-ray diffraction profiles for (a) the three K and for (b) the six MK. “Labo” means “traditional calcined MK in laboratory”.
Table 2 Estimated phase composition of parent K (calculated from multi-phase Rietveld refinement of powder diffraction profiles) compared to the values deduced using assumed phase compositions and EDX elemental analysis. Kaolinite
Quartz
Muscovite
wt.%
Rietveld
Stoichiometry
Rietveld
Stoichiometry
Rietveld
Stoichiometry
K1 K2 K3
89 ± 1.12 56 ± 0.91 53 ± 0.99
88 55 49
11 ± 0.93 41 ± 1.08 35 ± 0.91
9 40 37
0 ± 0.02 1 ± 0.41 12 ± 0.63
0 2 12
and 53%, respectively. These results correlated with the mineralogical composition determined by stoichiometry using chemical analysis (Table 2). Quartz and muscovite were the main “impurities” (industry term) in these raw K. The X-ray diffractograms of the MK (Fig. 2b) show that both methods of calcination were efficient, as evidenced by the presence of X-ray amorphous halo focused on about 3.8 Å (indicative of MKaol). According to the mineralogical composition determined by the Rietveld analysis of MK (Table 3), unlike Kaol, the relative amounts of quartz and muscovite were unchanged by calcination treatment relative to the fed material. These results also showed that there was no significant difference in the mineralogical compositions between flash calcined MK and traditionally calcined MK. However, in the flash calcination process of MK1, mullite was formed. This mullite formation suggests a calcination temperature exceeding 950 °C (Brindley and Nakahira, 1958). This crystalline phase is generally considered to be unreactive in cement-based materials.
Mullite can be formed from particles that passed directly through the flame rather than swirling around the periphery.
4.2. Trace amounts of Kaol after flash calcination The XRD analysis revealed the presence of remnant Kaol in the three MK after flash calcination, unlike the traditionally calcined MK (Fig. 2b). This observation was confirmed by FTIR (Fig. 3). The infrared spectra of the flashed MK exhibited one small band at approximately 7065 cm−1, which is attributed to inner hydroxyl groups and other hydroxyl groups located on the exterior surface of the octahedral sheet of Kaol (Petit et al., 1999); these bands were proof that Kaol was present after flash calcinations. However, results from these analyses did not clarify whether this trace of Kaol corresponds to a Kaol particle unaffected in its core or to a complete larger particle of Kaol which was unaffected by the dehydroxylation step.
Table 3 Estimated phase composition of six MK (calculated from multi-phase Rietveld refinement of powder diffraction profiles). wt.% MK1 MK2 MK3
Flash Electric furnace in laboratory Flash Electric furnace in laboratory Flash Electric furnace in laboratory
Amorphous phase
Quartz
Muscovite
Kaolinite
Mullite
Anatase
84 ± 2.15 88 ± 1.89 55 ± 1.11 57 ± 0.82 49 ± 0.91 51 ± 1.12
10 ± 0.95 10 ± 0.88 41 ± 1.09 41 ± 1.08 37 ± 0.97 36 ± 0.97
0 ± 0.01 0 ± 0.01 1 ± 0.04 1 ± 0.04 11 ± 0.56 12 ± 0.85
1 ± 0.41 0 ± 0.03 2 ± 0.52 0 ± 0.01 2 ± 0.32 0 ± 0.01
3 ± 0.21 0 ± 0.01 0 ± 0.01 0 ± 0.01 0 ± 0.01 0 ± 0.01
2 ± 0.4 2 ± 0.14 1 ± 0.02 1 ± 0.01 1 ± 0.01 1 ± 0.01
250
M. Claverie et al. / Applied Clay Science 114 (2015) 247–255
Fig. 3. FTIR spectra of the K1, K2, K3 and their MK.
Additionally, trace amounts of Kaol can also be present in MK produced in an industrial rotary kiln (San Nicolas et al., 2013).
4.3. Physical characteristics The morphological characteristics of the natural clays obtained by SEM are shown in Fig. 4. K3 clays show the typical microstructure of Kaol, consisting of pseudo-hexagonal plates and clusters of plates (Tironi et al., 2014). These plates are larger (N 4 μm) and more structured compared with the other clay samples. In the K2 and K1 samples, the microstructures are present in fine particle sizes (b 1 μm), irregular Kaol forms and very small flakes. In Fig. 5a, the particle size distributions of the clays are shown. A particle size analysis of the three K and their MK (Fig. 5a) showed that calcination induced a decrease in the relative amount of finer particle sizes (approximately 4 μm for K1 and K2) in favor of coarser particles (approximately 50 μm). This phenomenon seems to be accentuated with flash calcination. Traditional and flash calcination of K3 does not lead to a significant increase in particle size unlike K1 and K2 (K3 shows a slight increase from 15 μm to 30 μm after flash calcination while particle size is unchanged after traditional calcination). This can be explained by the presence of coarser particles (approximately 15 μm) in K3 unlike K1 and K2 (richer in finer particle sizes). Thus,
K1
flash calcination seemed to favor the agglomeration of submicron particles. Several particle morphologies were discernible in the three MK samples (Fig. 5). Quartz (Fig. 5e, h) was accompanied by aggregates of aluminum silicate particles with diameters ≥ 20 μm (Fig. 5d, f, i) covered by nanometer-sized particles of aluminum silicates. A few of these nanoparticles were crystallized and preserved the layered hexagonal structures of Kaol (Fig. 5c, g), whereas other nanoparticles were observed to have amorphized layers (dulled contours) (Fig. 5b). However, observed spherical 10 μm particles (Fig. 6a) in the flash-calcined products were not present in the traditional MK samples. These spherical particles in flashed MK accounted for approximately 19% in MK1, 15% in MK2 and 13% in MK3 (by visual estimation on about 50 SEM investigations and in agreement with San Nicolas, 2011). A previous study had been conducted to determine the internal composition of these spherical particles using hydrochloric acid etching (San Nicolas et al., 2013).These spherical particles seemed to improve mortar blends by compacting the granular skeleton (Jillavenkatsea, 2001).The smooth surface of the sphere shown in the SEM image in Fig. 6a appears to be composed of interlinked thin particles (Fig. 6b). According to EDX analyses (Fig. 6d), the surfaces of these thin particles were composed of aluminum silicates, similar to MKaol or mullite. Moreover, on the surface of these spheres, obscured stains were observable in the SEM image (Fig. 6c). Because the same spectra were obtained from the EDX analyses (Fig. 6d) between the obscured
K2
1 µm
K3
1 µm Fig. 4. SEM images of Kaol (× 10,000).
M. Claverie et al. / Applied Clay Science 114 (2015) 247–255
251
Fig. 5. (a) Particle size distribution for K1, K2, K3 and their MK. Two groups of particle size (grey circles) were highlighted: Group I = submicronical particles in MK with (b), (c) TEM images and (g) electron diffraction pattern of the submicronical Kaol observable in (c). Group II = coarser particles in MK with (d), (e) and (f) SEM images. (h) and (i) are EDX analyses of different particles observed in MK1, MK2 and MK3.
zones and the surfaces of the sphere, it was assumed that these stains were indicative of cavities under the surface. 4.4. Structure of spherical particle In order to analyze the composition of a spherical particle, a focused ion beam was used to cut a spherical particle in half (Fig. 7a) to prepare samples for TEM. The nm-scale resolution of the FIB allowed us to choose the exact region for thinning. A transverse section of a sphere with a platinum coat is observable in Fig. 7b. The lower part of this section was concave, highlighting the presence of a substantial cavity
within the spherical particle. The presence of small bubbles was also noticed (Fig. 7b—Cavity). The top of this section (Fig. 7c) shows a combination of platinum and nanometer-sized aluminum silicate particles as shown in the EDX analysis (Fig. 7—EDX-MKaol). These particles were previously observed (Fig. 6b—EDX-A) and correspond to MKaol enriched with iron and titanium. Several analyses were carried out on the transverse section of the spherical particle (Fig. 7b). The halo diffraction pattern (Fig. 7—SAD-A, B and C) from within the spherical particle indicated its amorphous character. The EDX analysis of this amorphous material (Fig. 7—EDX-A,
252
M. Claverie et al. / Applied Clay Science 114 (2015) 247–255
Fig. 6. (a) SEM images (in flashed MK2) of spherical particle (× 3700) and (b) details of the surface (× 22,000); (c) SEM image (in flashed MK3) of spherical particle (x 9,000) and (d) EDX analyses: EDX-A analysis of thin particles; EDX-B analysis of the smooth surface of sphere; EDX-C analysis of dark stains and EDX-D analysis of clear zone.
B and C) shows that it was primarily aluminum silicate with significant amounts of titanium and iron oxides. However, this spherical particle was not exclusively composed of an amorphous phase. A variegated texture was also observed (Fig. 7d). The diffraction pattern (Fig. 7—Sad-D and E) and the EDX analysis (Fig. 7—EDX-D and E) showed that this area was composed of crystallized aluminum silicate, and unlike the amorphous area, this crystalline phase contained fewer iron, calcium and titanium elements. This observation was confirmed by the chemical mapping analyses (Fig. 8). The crystalline phase was dark for the iron, calcium and titanium mappings, which indicated low contents for these elements. Spherical particles were the seat of a dynamic diffusion of iron, calcium and titanium elements in favor of the amorphous area. In addition to these crystallized aluminum silicates, spherical areas were also observed in the spherical particles (Fig. 7e—EDX-F). The electronic diffraction patterns (Fig. 7—SAD-F) indicated the presence of mullite in these areas. In summary, the spherical particles were composed primarily of amorphous aluminum silicates with traces of crystallized and pure aluminum silicates. These particles had cavities and mullite zones. The surfaces of these polyphasic particles were covered with submicronic MKaol particles. In a second spherical particle approximately 5 μm in diameter cut by FIB (Fig. 9), a large cavity was observed and was surrounded by other smaller cavities. Based on these various analyses, it is predicted that these spherical particles originated from the agglomeration of submicron Kaol during
flash calcination. At a certain distance from the flame (Fig. 10a), the temperatures were still sufficiently high enough to dehydroxylate Kaol to form MKaol. Nearer the flame, two phenomena likely occurred: - a vitrification of submicron aluminum silicates produced droplets due to a minimization of the surface energy; and - MKaol recrystallized into mullite.
Within the flash calciner, the cyclonic air movement allowed the mixing of these different phases (Fig. 10b). During the cooling step, a few of the submicron MKaol particles adsorbed onto the surfaces of the spherical particles. The cavities in the spherical particles showed evidence of gases, which were caused either by gas emissions during combustion (i.e., gas trapped in silicate melt) or by gas produced inside the droplet during the calcination reactions. Studies have been conducted on mortars blended with flash MK compared to rotary-kiln MK. The spherical morphology seemed to improve mortar blends by compacting the granular skeleton (Jillavenkatsea, 2011), and enhancing their workability (San Nicolas et al., 2013). Furthermore, flash calcination technology “consumes 2.2 MJ/t of MK, which is 80% less that the energy consumed during cement production” (San Nicolas, 2011). Recently, Teklay et al. (2015) developed a model for clay particle calcination and confirmed that a “residence time of 0.5 s and temperatures of about 950 °C are likely to be the most appropriate condition to produce high-quality calcines”.
M. Claverie et al. / Applied Clay Science 114 (2015) 247–255
253
Fig. 7. (a) SEM image (in flashed MK1) of spherical particle cut in half (× 5000); (b) TEM image of the transverse section; (c) TEM image of the top of the section; (d) TEM image of the crystalline phase; (e) TEM image of mullite spheres; elemental quantifications (using the Cliff-Lorimer ratio technique) from EDX analysis and selected area electron diffraction patterns of different phases observed inside the spherical particle.
254
M. Claverie et al. / Applied Clay Science 114 (2015) 247–255
Fig. 8. Schematic representation (same sphere section than Fig. 7b) and elemental mappings of the section of the sphere (Fig. 7b) (Si, Al, Ti, Ca and Fe).
5. Conclusions
treated by both calcination processes and the resultant products were analyzed. The following conclusions can be drawn:
The purpose of this study was to compare the chemical, physical and morphological characteristics of flash calcined and traditional rotarycalcined metakaolins and to determine the contents of the spherical particles formed after flash calcination. Three kaolin sources were
PLATINUM
CAVITY
Fig. 9. SEM image of second sphere cut in half by FIB.
1. The two calcination techniques produced metakaolins without significant mineralogical or chemical differences. However, the agglomeration of submicron particles after calcination was accentuated in flash calcination. 2. Unlike the traditionally calcined metakaolins in the laboratory, kaolinite was present in metakaolins after flash calcination. However, trace amounts of kaolinite have been reported in traditionally calcined metakaolins in industry 3. Unlike traditional metakaolins, flash metakaolins contained spherical particles (between 10% and 20% in this study). These spherical particles account for the enhanced workability of mortars blended with flash metakaolins compared with mortars blended with rotary-kiln metakaolins. 4. These spherical particles were produced from the agglomeration of submicron kaolinites during flash calcination. These spherical particles were composed of gases and a mixture somewhat pure aluminum silicates with varied crystallinities. This heterogenic composition was primarily due to the presence of a temperature gradient and cyclonic air movement in the flash calciner. Future works should study ability for these spherical particles to dissolve and their effects on the properties of cement matrices. These spherical particles could also have potential as reservoirs capable of reacting with calcium hydroxide.
M. Claverie et al. / Applied Clay Science 114 (2015) 247–255
255
Fig. 10. (a) Schematic representation of phenomena taking place in a flash calciner and (b) schematic representation of the cyclonic air movement allowing the mix of different phases.
Acknowledgments The authors are grateful for the support provided by Argeco Développement and sincerely thank those who volunteered their time to participate in this study. We thank David Bish and the anonymous reviewers for their constructive comments. References Aparicio, P., Galan, E., 1999. Mineralogical interference on kaolinite crystallinity index measurements. Clay Clay Miner. 47, 12–27. Barbosa, V.F.F., MacKenzie, K.J.D., Thaumaturgo, C., 2000. Synthesis and characterisation of materials based on inorganic polymers of alumina and silica: sodium polysialate polymers. Int. J. Inorg. Mater. 2, 309–317. Bich, C., 2005. Contribution à l'étude de l'activation thermique du kaolin: evolution de la structure cristallographique et activité pouzzolanique PhD thesis. Institut National des Sciences Appliquées, Lyon, France. Bish, D.L., von Dreele, R.B., 1989. Rietveld refinement of non-hydrogen atomic positions in kaolinite. Clay Clay Miner. 37, 289–296. Brindley, G.W., Nakahira, M., 1958. A new concept of the transformation sequence of kaolinite to mullite. Nature 4619, 1333–1334. Brindley, G.W., Sharp, J.H., Nakahira, M., Patterson, J.H., 1967. Kinetics and mechanism of dehydroxylation processes. Am. Mineral. 52, 1697–1705. Cyr, M., 2000. Contribution à la caractérisation des fines minérales et à la compréhension de leur rôle joué dans le comportement rhéologique des matrices cimentaires PhD thesis. University of Sherbrooke, Canada. Davies, T.W., 1984. Equipment for the study of flash heating of particle suspension. High Temp. Technol. 2 (3), 6. Garcia-Diaz, A., 1995. Réactivité pouzzolanique des métakaolinites: corrélation avec les caractéristiques minéralogitologiques des kaolinites PhD thesis. Génie des Procédés. Alès, l'école des Mines, France. Harvey, C.C., Lagaly, G., 2013. Industrial application Chapter 4.2. Handbook of Clay Science 5B pp. 451–490. He, C., Makovicky, E., Osbaeck, B., 1995. Pozzolanic reactions of six principal clay minerals: activation, reactivity assessments and technological effects. Cem. Concr. Res. 25 (8), 1691–1702.
Heller-Kallai, L., 2013. Thermally modified clay minerals Chapter 10.2. Handbook of Clay Science 5A pp. 411–427. Insley, H., Ewell, R., 1935. Thermal behavior of the kaolins minerals. J. Res. Natl. Bur. Stand. 14, 331–335. Jillavenkatsea, A., Dapkunas, S.J., Lum, L.H., 2001. Particle size characterisation. NIST Recommend. Prac. Guide 906 164 pp. Lutterotti, L., et al., 1999. MAUD (Material Analysis Using Diffraction): a user friendly Java Program for Rietveld Texture Analysis and more. Proceeding of the Twelfth International Conference on Textures of Materials. 1, p. 1559. Meinhold, R.H., Salvador, S., Davies, T.W., Slade, R.T.C., 1994. A comparison of kinetics of flash calcination of kaolinite in different calciners. Chem. Eng. Res. Des. 72 (8). Petit, S., Madejova, J., Decarreau, A., Martin, F., 1999. Characterization of octahedral substitutions in kaolinites using near infrared spectroscopy. Clay Clay Miner. 47, 103–108. Rietveld, H.M., 1967. Line profiles of neutron powder-diffraction peaks for structure refinement. Acta Cristallographica 22, 151–152. Rietveld, H.M., 1969. A profile refinement method for nuclear structures. J. Appl. Cristallogr. 2, 65-57. Salvador, S., 1992. Production de pouzzolanes de synthèse par calcination flash de sols argileux; étude des produits et conception d'une installation PhD thesis,. Institut National des Sciences Appliquées, Mécanique, Toulouse, p. 230. Salvador, S., 1995. Pozzolanic properties of flash-calcined kaolinite: a comparative study with soak-calcined products. Cem. Concr. Res. 25, 102–112. San Nicolas, R., 2011. Approche performantielle des bétons avec métakaolins obtenus par calcination flash PhD thesis. University of Toulouse, France. San, Nicolas R., Cyr, M., Escadeillas, G., 2013. Characteristics and applications of flash metakaolins. Appl. Clay Sci. 83–84, 253–262. Sperinck, S., Raiteri, P., Marks, N., Wright, K., 2010. Dehydroxylation of kaolinite to metakaolin—a molecular dynamics study. J. Mater. Chem. 21, 2118–2125. Teklay, A., Yin, C., Rosendahl, L., Lindequist, Køhler L., 2015. Experimental and modeling study of flash calcination of kaolinite rich clay particles in a gas suspension calciner. Appl. Clay Sci. 103, 10–19. Tironi, A., Castellano, C.C., Bonavetti, V., Trezza, M.A., Scian, A.N., Irassar, E.F., 2014. Kaolinic calcined clays–cement system: hydration and properties. Constr. Build. Mater. 64, 615–621. White, C.E., Provis, J.L., Proffen, T., Riley, D.P., van Deventer, J.S.J., 2010. Density functional modeling of the local structure of kaolinite subjected to thermal dehydroxylation. J. Phys. Chem. A 114, 4988–4996.