QXRD Studies of Phases Formed

Abstract

As shown earlier 980 °C reaction products of kaolinite constitute poorly crystallized spinel phase, weakly crystallized mullite, and amorphous SiO₂. The SiO₂ phase was estimated by alkali-leaching technique as described in Chap. 17. The estimation of the other two weakly crystalline phases was attempted by different researchers using QXRD technique. Both crystallinity and quantity of mullite develop largely near 1250 °C exotherm. β-cristobalite usually developed at 1200 °C. The growth rates of both mullite and cristobalite were subsequently determined by the same QXRD technique.

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Introduction

As shown earlier 980 °C reaction products of kaolinite constitute poorly crystallized spinel phase, weakly crystallized mullite, and amorphous SiO₂. The SiO₂ phase was estimated by alkali-leaching technique as described in Chap. 17. The estimation of the other two weakly crystalline phases was attempted by different researchers using QXRD technique. Both crystallinity and quantity of mullite develop largely near 1250 °C exotherm. β-cristobalite usually developed at 1200 °C. The growth rates of both mullite and cristobalite were subsequently determined by the same QXRD technique.

2 Estimations of 980 °C Spinel Phase and Mullite Phase

Tsuzuki et al. (1969) used QXRD technique to estimate spinel and mullite phases only. They heated one halloysite and two other kaolinites isothermally and noted the diffraction peaks of the high-temperature phases namely spinel and mullite repeatedly by oscillatory scanning with a Geigerflex X-ray diffractometer provided with a high-temperature attachment. The amounts of high-temperature phases were estimated by comparing with standards. The change in the amount of high-temperature phases with time is shown in Fig. 14.1.

Fig. 14.1

The result of isothermal X-ray diffraction experiments at the temperatures indicated. The changes in the amounts of high-temperature phases are shown as a function of time (After Tsuzukiand Nagasawa 1969). Reproduced by permission of The Clay Science Society of Japan

The amounts of high-temperature phases are of semi-quantitative accuracy because of weak intensity and broad shape of the diffraction peaks. It was noted that spinel crystallized from halloysite and disordered kaolinites, whereas both spinel and mullite crystallized simultaneously from kaolinite. In the high-temperature runs, the rate of crystallization was the largest at the beginning and then it slowed down with time. Whereas, in the low temperature runs, the crystallization began after an induction period and then the rate became larger following a sigmoid shape of the curve. Quantitative assessment of 970 °C heated product out of Mese Alta kaolinite is as follows. Amount of spinel = 35 % approximately. Amount of mullite = 5 % approximately. Total amount of crystalline phase = 40 % approximately.

Okada et al. (1986) estimated spinel, mullite, and cristobalite by heating to various temperature from several kaolin groups of minerals by X-ray quantitative analysis. X-ray measurements were made using a graphite monochromatized (Cuk_α radiation).

For X-ray quantitative analysis CaF₂ was used as an internal standard. A step scanning technique of 0.05° in 2θ intervals and fixed time of 40 s was applied for measurement of areas of spinel phase because of the faint and diffused reflections. Reflections used for quantitative analysis were (440) of the spinel phase, the (121) of mullite, the (101) of cristobalite, and the (220) of CaF₂. Substances used for calibration of each phase were synthetic ones: spinel phase (fired Al₂O₃ gel at 900 °C for 24 h) mullite (fired 3Al₂O₃ 2SiO₂ xerogel at 1600 °C for 24 h) and cristobalite (fired SiO₂ gel at 1500 °C for 24 h). Figure. 14.2 shows the results of X-ray quantitative analysis of the specimens.

Fig. 14.2

Amount of spinel phase, mullite, and cristobalite in each specimen as a function of firing temperature. Calculated amounts for spinel phase based on the compositions of (1) 3SiO₂.2Al₂O₃, (2) 2SiO₂.3Al₂O₃, and (3) Al₂O₃ are represented by arrows with their numbers, and those for mullite and cristobalite are represented by Mu_c and Cr_c, respectively, Sp: spinel phase; Mu: Mullite; Cr: cristobalite (After Okada et al. 1986). Reprinted by permission of the American Ceramic Society

Chakraborty and Ghosh (1991a, b) measured semi-quantitative values of spinel and mullite phases by QXRD technique. The spinel phase was estimated by comparing the 0.137 nm. Bragg diffraction peak height of 980 °C heat-treated kaolinite with that of a standard γ-Al₂O₃ and by using pure.

MgO as an internal standard according to the formulae used by Verduch and Estrade (1962). The calibration curve for spinel was first obtained by measuring mixtures of 10–40 wt%. γ-Al₂O₃ with SiO₂ (A) as diluents. Thereafter, R values of different heat-treated samples were determined separately from their individual XRD runs and then R values versus % spinel was plotted. Where,

$$ {\mathbf{R}} \, = \frac{{{\mathbf{H}}_{\text{sp}} \times {\mathbf{W}}_{\text{sp}} }}{{{\mathbf{H}}_{\text{MgO}} \times {\mathbf{W}}_{\text{MgO}} }} $$

where H _sp = Height of spinel peak, W _sp = Width at half height; and H _MgO and W _MgO are the corresponding values for standard MgO.

From the calibration curve, approximately, values of spinel content formed on heating different clays have been estimated. Table 14.1 shows that the content of spinel in different kaolinite varies. The average value lies within 30–40 wt%.

Table 14.1 Semi-quantitative estimation of Ai–Si spinel phase DERIVED from kaolinites heated at 980 °C

Mackenzie and Brown (1987) calculated the spinel content formed out of kaolinite marked as BDH “light’’ by quantitative X-ray diffraction measurements of the 440 reflection, using γ -Al₂O₃ as celebrant phase and elemental Si as internal standard. Their estimated value is ~19 wt%, which is considerably less than that found by Okada et al. (1986).

3 Calculated Value of Aluminosilicate (A) Phase

Weakly crystalline mullite formed during decomposition of kaolinite at 980 °C is different from that of well-crystalline mullite which is usually taken as a standard. Even in poorly developed stage the intensity of the 0.537 nm Bragg reflection peak varies and it is dependent on the origin of the clay. Keler and Leonov (1955) observed more spinel and little mullite, where as Mikheev and Stulov (1955) noted comparatively more mullite and less spinel. Under these circumstances, the percentage of poorly crystalline mullite which forms in different kaolinites at 980 °C may be taken as 4–6 wt%, approximately. The amount of SiO₂ (A) liberated during kaolinite was 35–37 % as estimated by Chakraborty and Ghosh (1978a, b). Sum of the two crystalline phases and SiO₂(A) is still short of 100, which predicts the existence of another amorphous phase in 980 °C heated products of kaolinite. It was identified by taking powder diffraction photograph using Lindamen glass tube as sample holder in an 11.46 cm Philips circular camera. An amorphous band was still observed in XRD patterns of the alkali leached residue during the extraction process of kaolinite heated to 980 °C. The approximate percentage of this aluminosilicate phase (A) is = [100−(amount of spinel) + (amount of mullite) + (amount of SiO₂ (A))]. By substituting the estimated values in the equation, the amount of alumino silicate (A) is calculated as 30–40 %. The varying percent of this phase is due to the variations of spinel and mullite formed from different varieties of kaolinite found in nature. This aluminosilicate (A) is designated as mullite (A).

4 Estimation of Mullite and Cristobalite Phases

Konopicky (1959, 1962) identified mineral phases formed during heating wide range of German and other continental commercial fire bricks and then estimated the amounts of mullite and cristobalite by X-ray methods and the glass content by difference. Grimshaw (1971) reviewed the works of the above authors. The amount of mullite content is found to be related to the alumina present in the clay and the relative ratio of cristobalite to glass may be dependent upon the content of fluxing oxides.

Johnson and Andrews (1962) studied the mineralogical compositions of a range of British alumino-silicate fired products made from high- and low-grade fire clays. Later on they compared the composition of fired clay products with fired products made out of calcined Kyanite or Sillimanite and calcined Bauxite as raw materials.

Mcgee (1966) heat-treated three clays, a high duty plastic fire clay, a flint fire clay, and a kaolin at 1350 °C for 5 h after which these were cooled in variety of ways. The mineralogical composition of the fired products namely mullites and cristobalite were quantitatively calculated by X-ray diffraction method using CaF₂ and BaF₂ as internal standards. It is shown those mullites content in respective clays are not affected by difference in cooling schedule. Further it also noted that the rate of crystallization from glassy silica phase to β-cristobalite is very much influenced by the cooling rate. Generally, rapid cooling resulted in a high-glass content in contrast to the more crystalline phase developing under slow conditions. Four classes of fire clay refractories, viz., (1) high heat duty, (2) intermediate heat duty, (3) moderate heat duty, and (4) low heat duty is tentatively defined by some physical tests under the A.S.T.M. Performance of these bricks might be predicted out of the knowledge of the crystal development of mullite out of fire clay. Harvey and Birch (1936) fired super heat duty bricks at 1280, 1320, and 1430 °C and examined for petrographic microscope for microstructure developments. Percentages of mullite formed were ascertained by eye estimating the relative intensities of X-ray lines of the known and unknown specimens. Qualitative X-ray tests conclusively show mullite formation at all three successive heat-treated specimens. The percentage of mullite was, approximately, 50 % which is independent of firing temperatures. Petrographic examination showed no indication of mullite development at 1280 °C but indication of incipient crystallization in the form of minute skeletal outlines was noted at 1320 °C. On high temperature calcination at ~1430 °C, they observed an interlocking mass of very fine mullite crystals of needle like habit in the matrix. The sizes of them are ~0.01 mm in length and 0.0007 mm in width. Thus, they came to the conclusion that the size of the mullite crystals increases with increase of calcination temperature but probably not the total amount of mullite content.

Chakraborty (1993) estimated the semi-quantitative amount of cristobalite formed in Rajmohol kaolinite heated to different temperatures with 2 h soaking by XRD technique. The standard cristobalite used was prepared by heating pure quartz at 1650 °C for 8 h followed by leaching with 5 % HF in an ice both for 5 min and lastly by X-ray analysis of the washed residue. This process is repeated till the height of the 0.404 nm XRD peak of β-cristobalite became nearly constant. Figure. 14.3 shows the growth curve of β–cristobalite.

Fig. 14.3

Amount of β-cristobalite formed versus temperature of heating of Rajmohol kaolinite under static heating conditions

5 Assessment of Vitreous Phase

Kaolinite generally contains some associated impurity oxides namely Fe₂O₃, FeO, TiO₂, CaO, MgO, K₂O, and Na₂O to the extent about 2.5 % by weight. These oxides may undergo solid state reaction during heating with free SiO₂ liberated out of kaolinite and forms a silica-rich vitreous phase. The vitreous phase may have played a role in mullite. Major fluxing oxides reduce the refractoriness of kaolinite body during heating process. These fluxing oxides control the viscosity of the resultant vitreous phase. Finally, the impurity oxides also influence the quantity of the glassy phase developing. Thus, the nature and quantity of impurities exert influence on the formation process of mullite and its crystal growth. Generally higher the kaolinite content and lower the flux concentration, higher will be the content of mullite. However, the amount and nature of vitreous phase developed and the way it reinforced the mullite needless is the determining factor in obtaining a definite microstructure. The vitreous phase improves homogeneity of the matrix containing various sizes of mullite. Thus, mullite/glass ratio is an important index in determining the strength characteristics, since the matrix contains a well-growth nest like interwoven network of mullite crystals in glassy mass (Fig. 14.4).

Fig. 14.4

Mineralogical composition versus firing temperature of pioneer kaolin (After McGee1966). Reprinted by permission of the American Ceramic Society

Mcgee (1966) calculated the percentage of liquid-phase by subtracting total crystalline phases from 100. The result shows a constant value of about 60 % of liquid-phase in three types of clay Fig. 14.5. The chemical composition of the liquid-phase was calculated based on the assumptions that Al₂O₃ to SiO₂ ratio of mullite is 3:2, Fe₂O₃ and TiO₂ replaced alumina only and is shown in Table 14.2. A few problems are noted in estimation process. There are variations in both mullite and cristobalite contents of fired kaolinites. Mcgee (1966) showed % mullite is in the range 27–33 which is lower than the mullite developed in a commercially fired aluminosilicate products reported by Konopicky (1962) and Johnson and Andrew (1962). The main source of error in estimation of mullite is the variation of degree of crystallinity of mullite formed from kaolinite during different stages of heating as pointed out by McGee (1966). In contrast, standard mullite used in estimation work generally of sintered/or fused variety. Konopicky and Kohler (1958) suggested that for mullite determination in heated clays, the 100 % mullite should be obtained from alumino silicate in question by separation process. McGee (1966) synthesized cristobalite to be used as a standard by calcining reagent grade silicic acid at 1600 °C for 4 h. Chakraborty and Ghosh heat-treated pure quartz at 1650 °C/1 h. to develop major cristobalite.

Table 14.2 Composition of liquid phase of fireclays and kaolin (After McGee, 1966) Reprinted with kind permission of American Ceramic Society

Fig. 14.5

Mullite development versus temperature from different sources : (C2D) mixture of Al₂O₃(A) and SiO₂ (A), (B) Bhandak kaolinite (After Chakraborty and Ghosh 1991). Reprinted by permission of the American Ceramic Society

Chakraborty and Ghosh (1991a, b) compared the mullite formation behavior of kaolinite with that of synthetic mixture. An artificial mixture (C_2D) of Al₂O₃ and SiO₂ (ratio 3:2) was first prepared by mixing amorphous alumina and amorphous SiO₂ in an agate mortar for 2 h. This mixture and Bhandak kaolinite were heat treated at successively higher temperatures, cooled, ground to −325 mesh, and analyzed by XRD for quantitative identification of the phases present as well as far quantitative estimation of mullite. In the synthetic mixture C_2D, γ-Al₂O₃ is the first crystalline phase and it exists between 700 and 900 °C. On further heating, γ-Al₂O₃ gradually starts to disappear and transforms to corundum (major) at 1100 °C, leaving behind a minor amount of unconverted α-Al₂O₃ phase. More and more corundum crystallizes, as is evident from the increase in intensity of the 0.208 nm Bragg reflection of the corundum pattern with an increase of temperature up to 1200 °C and then it decreases. The development of corundum takes place independently whether amorphous Al₂O₃ is present in a mixture with silica (A) or alone. Cristibalite appears at about 1075 °C and goes on increasing in between 1200 and 1300 °C after that it decreases. Mullite can be identified at 1200 °C from its faint X-ray pattern, but with a progressive increase of temperature both the intensity and the crystallinity of mullite increase (Fig. 14.5).

Thus, the results show that component oxides first crystallites into their respective high-temperature crystalline forms, e.g., alumina (A) to γ-Al₂O₃ and then to α-Al₂O₃. Side by side silica (A) to β–cristobalite. Afterward, as the temperature rises, the intensities of the two newly formed crystals, namely, corundum and cristobalite decrease gradually with the formation of an aluminosilicate phase (A). Thereafter, mullite nucleation takes place and then steady crystallization as recorded in the sequential phase developments versus temperature curves (Fig. 14.6). In comparison, quantitative mullite formation from Bhandak clay shows different growth rate.

Fig. 14.6

Sequential phase developments during heating of artificial mixture of amorphous Al2O3 and amorphous SiO₂: (1) cristobalite, (2) corundum, (3) mullite (After Chakraborty and Ghosh 1991). Reprinted by permission of the American Ceramic Society

6 Summary

Amount of spinel phase formed in K–M reaction series was measured by QXRD technique. Tszuki et al. (1969) noted difference in formation behavior of spinel phase and its quantity in ordered/disordered kaolinite and halloysite. Instead of dynamic heat-treated kaolinite, Chakraborty and Ghosh (1991a, b) made quantitative study of spinel phase on soaked kaolinite and noted a variation in the spinel content from 30–40 %.They indicated the existence of 35–40 % mullite (A) phase in the 980 °C decomposition product of kaolinite. Okada et al. (1986) also estimated spinel phase formed in different kaolin minerals by QXRD technique. They compared the estimated value of spinel formed with theoretical amount expected for three types of spinel formation models, and finally showed that observed value of spinel content best fitted with calculated ones based on the assumption of γ-Al₂O₃ spinel formation as intermediary.