CRYSTALLIZATION KINETICS OF CaO-SiO2-Al2O3-MgO SLAGS
Shaghayegh Esfahani1, Mansoor Barati2
1,2Materials Science and Engineering, University of Toronto 184 College Street
Toronto, Canada M5S 3E4
Keywords: Crystallization kinetics, Nucleation, Growth, Activation energy, Avrami, Single Hot Thermocouple Technique (SHTT), Phase transformation
Abstract
Crystallization behavior of blast furnace slag is of great interest for generating value-added products from slag, such as cement feedstock, where the slag structure determines the material quality. Aided with a Hot Thermocouple Device, the kinetics of crystallization of CaO-SiO2- Al2O3-MgO slags was determined. The rate of nucleation and growth of crystals were measured for a range of slag basicities, temperatures, and hold times.
Introduction
Blast furnace slags constitute over 50% of the total slag produced worldwide, amounting to 320 million tonnes per year. [1]. Blast furnace slags are most desired as feedstock for Portland cement production due to the large contents of silica and lime in these slags. For this application blast furnace slag has to be cooled rapidly to form an amorphous structure, as otherwise it will cause swelling of the concrete made from slag cement. Therefore, understanding of the crystallization behavior of BF slag is crucial with the purpose of controlling the cooling conditions or manipulating the slag composition so that the desired amorphous structure is achieved. Crystallization kinetics of mold fluxes has been previously studied by other researchers [1-7] due to their important role in steel continuous casting mold. However, similar investigations for metallurgical slags are limited [8-11]. In a recent study, [11] the transformation kinetics of a slag with composition close to the BF slags has been studied but to this date, a comprehensive study on the effect of composition on the rate and extent of crystallization is lacking.
Experimental Apparatus and Method
The kinetic study was carried out using the Single Hot Thermocouple Technique (SHTT). In this technique, a small piece of sample, is placed in tip of a type-B thermocouple (D=5mm) which is controlled by a thermocouple driver. The thermocouple acts as the heating element at the same time that it measures the temperature of the sample. This technique allows rapid heating/ cooling of the materials as well as in-situ observation of the crystallization kinetics. Also, higher accuracy ofF temperature measurement is possible due to the direct contact of the thermocouple with the sample.
Advances in Molten Slags, Fluxes, and Salts: Proceedings of The 10th International Conference on Molten Slags, Fluxes and Salts (MOLTEN16) Edited by: Ramana G. Reddy, Pinakin Chaubal, P. Chris Pistorius, and Uday Pal TMS (The Minerals, Metals & Materials Society), 2016
Four CSAM slag samples with varying basicity (C/S=CaO/SiO2) were prepared (Table 1). For each slag composition (wt %), 10 g of oxide powders were blended and melted in a graphite crucible (D=4cm, h=4cm) by heating the material to approximately 50°C above the slag liquidus temperature for 4 hours in an argon atmosphere. To avoid segregation during solidification, the molten slag was quenched in water, producing a glassy material from which small quantities were used for the crystallization study.
Table 1- Composition of the slag samples Sample
number
1 2 3 4
CaO 47 45 44 42
SiO2 33 35 36 38
Al2O3 15 15 15 15
MgO 5 5 5 5
TL (°C)* 1578 1444 1430 1396
C/S 1.4 1.3 1.2 1.1
* Liquidus temperature determined from FactSage™ calculations
The isothermal experiments were carried out by heating 3-5 mg of sample to 1600°C and holding it for 1 minute then cooling rapidly 70-75°C/s to the target temperature. The transformation of glassy to crystalline slag was recorded until the sample was fully crystalline.
Isothermal Transformation Kinetics
Theoretical Treatment of Transformation
For quantifying the transformed volume fraction during isothermal transformation, the theoretical treatment was carried out through JMAK equation (1) [12-16].
𝑋 = 1 − exp[−(𝑘𝑡)𝑛] (1)
This expression describes the relationship between the transformed volume fraction and time as a sigmoidal curve. The coefficient K is a kinetic parameter which includes both nucleation and growth rates and Avrami exponent, 𝑛, can provide useful information on the mechanism of transformation.
Experimental Determination of Phase Change
Using image analysis techniques, the crystallized fraction of each slag was determined for 9-10 different isothermal temperatures ranging from 1100 ̶ 1300ºC and repeated 3-4 times at each temperature. As an example, the crystallized fraction during isothermal transformation for slag sample with C/S=1.2 at 1200ºC is shown in Figure 1. In these figures, the light regions are opaque crystals which form in the transparent amorphous slag matrix. The bright ring appearing in the middle of some of these figures is due to the additional light source used to increase the image resolution.
𝑡 − 𝜏 =1s 𝑡 − 𝜏 = 3s 𝑡 − 𝜏 =8s 𝑡 − 𝜏 =14s Figure 1- Crystallization of slag with C/S=1.2 at 1200°C isotherm (𝜏 =incubation time).
Determination of the Avrami parameters
The crystallized fraction from the isothermal treatments were found to follow a sigmoidal form Figure 2-a), described by the JMAK equation (1). In this figure it is observed that with the increase of temperature the crystallization is slower. This is due to the fact that nucleation is the dominant mechanism of crystallization and at higher temperatures there is less nucleation driving force thus lower kinetics. The mechanism of crystallization for all the slags has been discussed further in the following sections.
Constants 𝑛and 𝐿𝑛𝑘could be obtained by plotting 𝐿𝑛[−𝐿𝑛(1 − 𝑋)] verses 𝐿𝑛(𝑡 − 𝜏) (2) and Figure 2-b with values presented in Table 2.
𝐿𝑛[−𝐿𝑛(1 − 𝑋)] = 𝑛𝐿𝑛(𝑘) + 𝑛𝐿𝑛(𝑡 − 𝑡0)] (2)
0 5 10 15
0.0 0.2 0.4 0.6 0.8 1.0
1150C 1200C 1250C
Crystallized fraction (X)
t-t0
0.5 1.0 1.5 2.0 2.5 3.0 3.5 -3.2
-2.4 -1.6 -0.8 0.0 0.8 1.6
Ln[-Ln(1-X)]
1150C 1200C 1250C
ln(t-t0
(a) (b)
Figure 2- (a) Crystallized fraction as a function of time (b) JMAK plots for the slag with C/S=1.4.
Table 2- Avrami exponent (n) and Lnk
C/S=1.4 C/S=1.3 C/S=1.2 C/S=1.1
Temp (°C) n Lnk n Lnk n Lnk n Lnk
1100 ‒ ‒ 1.9 -1.8 2.0 -1.8 ‒ ‒
1120 3.1 -0.8 1.4 -2.3 1.5 -1.9 ‒ ‒
1130 ‒ ‒ 1.3 -2.6 1.6 -2.0 1.5 -5.4
1140 2.9 -1.1 1.3 -2.7 1.7 -2.1 1.5 -5.4
1150 2.7 -1.4 2.0 -2.8 1.9 -2.2 1.4 -4.9
1160 3.9 -1.3 1.4 -2.1 1.4 -2.3 2.0 -5.0
1170 3.1 -1.2 1.7 -2.0 1.8 -2.0 ‒ ‒
1180 2.8 -1.3 1.6 -2.0 1.9 -2.2 1.9 -4.7
1200 4.0 -1.9 1.7 -2.4 1.9 -2.2 2.2 -4.7
1250 3.8 -2.2 2.1 -3.0 1.7 -2.2 2.0 -5.1
1270 ‒ ‒ ‒ ‒ ‒ ‒ 2.0 -5.3
1300 4.0 -2.6 1.6 -2.9 1.8 -2.9 ‒ ‒
Nucleation Rate
Depending on the nucleation mechanism, the nucleation rate may increase, decrease or remain constant with time. The dependence of nucleation rate on time was investigated by image analysis. The resulting nucleation rates for all slag compositions at T= 1200°C are presented in Figure 3. As seen, the slag with C/S=1.4 experiences a significant increase in the nucleation rate with time. On the other hand, for the slag with C/S=1.2 and 1.3, the nucleation rate decreases with time. As the basicity is further decreased (C/S=1.1), the nucleation rate becomes close to zero.
Figure 4 shows 𝐿𝑛𝑘 verses 1/𝑅𝑇 for various slags. As seen, at a constant temperature the rate constant increase with the increase of C/S. This is due to the oxide structure becoming depolymerized and the bonds weakened as the C/S increases resulting in lower viscosities. With the decrease of viscosity the ionic mobility increases and it becomes easier for the structure to rearrange and accommodate the changes induced by crystallization.
10 120 130 140 150
-4 -2 0 2
Ln(N)
Time (s)
C/S=1.4 C/S=1.3 C/S=1.2 C/S=1.1
Figure 3- Change of nucleation rate with time for all slags at T=1200°C.
7.5 8.0 8.5 9.0
-6 -4 -2
0 C/S=1.4
C/S=1.3 C/S=1.2 C/S=1.1
Lnk
1/RT×105
Figure 4-𝐿𝑛𝑘 verses 1/𝑅𝑇 for all slag compositions
Growth Rate
The growth rate can also be determined by measuring the thickness of the crystals at different times, using image analysis, as shown in Figure 5. For each slag, the measurements were made for three temperatures (1150°C, 1200°C, 1250°C) and the growth rate was measured until the crystal under study grows independently and before colliding with any neighboring crystals. The result of this analysis is presented in Figure 6. The slope of these curves which are all constant with time corresponds to the growth rate of individual crystals (i.e. constant growth rate), and are presented in Table 3.
(a) (b)
Figure 5- Growth rate measurement for sample with C/S=1.2 at T=1250°C (a) 𝑡 − 𝜏= 6s (b ) 𝑡 − 𝜏 = 9𝑠.
0 1 2 3 4 5 6
0.0 0.1 0.2 0.3 0.4
1250C 1200C 1150C
Crystal thickness (mm)
Time (s) (a)
0 2 4 6 8 10 12
0.0 0.1 0.2 0.3 0.4
1250C 1200C 1150C
Crystal thickness (mm)
Time (s) (b)
Figure 6- Crystal thickness verses time (a) C/S=1.4 (b) C/S=1.3.
Table 3- Crystal growth rate (mm/s) at different temperatures.
Basicity (C/S)
Growth rate (mm/s)
1250°C 1200°C 1150°C
1.4 68 58 44
1.3 58 53 40
1.2 38 32 29
1.1 10 8 4
The dependence of growth rate on time may provide some indications to the mechanism of growth. In most cases, linear dependence corresponds to interface-controlled growth whereas when growth is proportional to the square root of time, the growth can be considered to be diffusion-controlled. However, in the former case, it has been suggested that a constant growth rate does not always translate to interface-controlled growth [17-19] and for determining the mechanism of growth it is best to examine the morphology of the interface in atomic scale. In general, when faceted interface is observed, the rate-controlling step is the interface reaction.
However, when the interface is rough and cellular, such as dendritic growth, the rate limiting mechanism for growth is diffusion [20]. In the case of the compositions in the current study, the morphology of the interface was not examined in near atomic scale but an observation through a microscope clearly shows that the interface is not faceted which confirms a dendritic growth (Figure 7).
(a) (b)
Figure 7- Fully dendritic interface morphology for sample with C/S=1.1 at a) T=1270°C b) 1250°C
Mechanism of Slag Crystallization
The 𝑛 values obtained from experiments were compared with theoretical values to determine the crystallization mechanism pertaining to the slags of the present study. As discussed earlier, the growth for slags appear to involve a diffusion-controlled process. For slag with C/S=1.4, the nucleation rate is increasing, which according to the 𝑛 values suggested by Christian [21], indicates that 𝑛 must be above 2.5. This is in agreement with the measured value of 4.0 > 𝑛 >
2.8 presented in Table 2. However, for the slag with C/S=1.2 and C/S=1.3, the nucleation rate is
decreasing which would suggest 𝑛 values between 1.5-2.5. This is also consistent with the measured values of 𝑛 which falls within the same range as demonstrated in Table 2. For the sample with C/S=1.1 the nucleation rate is close to zero which again shows a match between the measured 𝑛 values and the those suggested according to Christian [21] for temperatures below 1160°C (n=1.5). The only existing discrepancy is for sample with C/S=1.1 at temperatures above 1160°C which is attributed to the lower contrast between the matrix and the crystals which results in difficulty in defining the starting point of nucleation of this sample.
Conclusion
In this study, crystallization kinetics of synthetic CaO-SiO2-Al2O3-MgO (CSAM) slags with different basicities were studied by Single Hot Thermocouple Technique (SHTT) during isothermal treatment. Kinetic parameters were obtained by analysis of images from in-situ observation of glassy to crystalline transformation. Also, the dependence of nucleation and growth rates of crystalline phases were quantified as a function of time, temperature, and slag basicity. Together with the observations of crystallization front, they facilitated establishing the dominant mechanisms of crystallization.
Acknowledgements
The funding for this research was provided by Natural Science and Engineering Council of Canada (NSERC) and Ontario Government through an Early Researcher Award and an Ontario Government Scholarship for Science and Technology (OGSST).
References
1. M. Barati and S. Esfahani, “Slag crystallization behavior pertaining to heat recovery”
(Paper presented at the 6th International Congress on the Science and Technology of Steelmaking, Beijing, China, 2015).
2. Y. Kashiwaya, C. E. Cicutti, and A. W. Cramb, "Investigation of the crystallization of a continuous casting mold slag using the single hot thermocouple technique," ISIJ International, 38 (1998), 357-365.
3. Y. Kashiwaya, C. E. Cicutti, A. W. Cramb, and K. Ishii, "Development of double and single hot thermocouple technique for in situ observation and measurement of mold slag crystallization," ISIJ International, 38 (1998), 348-356.
4. J. L. Klug, R. Hagemann, N. C. Heck, A. C. F. Vilela, H. P. Heller, and P. R. Scheller,
"Fluorine-free mould powders for slab casting: Crystallization control in the CaO-SiO2- TiO2-Na2O-Al2O3 system," Steel Research International, 83 (2012), 1186-1193.
5. G.-H. Wen, H. Liu, and P. Tang, "CCT and TTT diagrams to characterize crystallization behavior of mold fluxes," Journal of Iron and Steel Research, International, 15 (2008) 32-37.
6. Z. T. Zhang et al. , "Observations of crystallization in mold slags with varying Al 2O3/SiO2 ratio," Steel Research International, 81 (2010), 516-528.
7. Y. G. Maldonado et al., "Kinetic study of the devitrification of mold powder slags," Iron and Steel Technology, 10 (2013), 65-75.
8. P. F. James, "Kinetics of crystal nucleation in silicate glasses," Journal of Non- Crystalline Solids, 73 (1985), 517-540.
9. P. Rocabois et al., "Crystallization kinetics of Al2O3-CaO-SiO2 based oxide inclusions,"
(Paper presneted at the 6th International Conference on Molten Slags, Fluxes and Salts, 12-17 June 2000, Netherlands, 2001) 98-109.
10. A. A. Francis, "Non-isothermal crystallization kinetics of a blast furnace slag glass,"
Journal of the American Ceramic Society, 88 (2005), 1859-1863.
11. Y. L. Qin, X. W. Lv, J. Zhang, J. L. Hao, and C. G. Bai, "Determination of optimum blast furnace slag cooling rate for slag recycling in cement manufacture," Iron and Steelmaking, 42- 5 (2015), 395-400.
12. M. Avrami, "Kinetics of phase change. I: General theory," The Journal of Chemical Physics, 7 (1939), 1103-1112.
13. M. Avrami, "Kinetics of phase change. II Transformation-time relations for random distribution of nuclei," The Journal of Chemical Physics, vol. 8 (1940), 212-224.
14. M. Avrami, "Granulation, phase change, and microstructure kinetics of phase change.
III," The Journal of Chemical Physics, 9 (1941), 177-184.
15. W. A. Johnson and R. F. Mehl, "Reaction kinetics in processes of nucleation and growth," American Institute of Mining and Metallurgical Engineers - Transactions, 135 (1939) 416-442.
16. A. N. Kolmogrorov, " A statistical theory for the recrystallization of metals," Bull. Acad.
Sci. URSS, Cl. Sci. Math. Nat., (1937), 355.
17. D. R. Uhlmann, Glass: Current Issues (Boston: Martinus Nijhoff Publishers), 1985.
18. H. R. Swift, "Some experiments on crystal growth and solution in glasses," American Ceramic Society -- Journal, 30 (1947), 165-169.
19. D. R. Uhlmann and G. W. Scherer, "Crystallization kinetics of Na2O·3SiO2," Journal of Crystal Growth, 29 (1975), 12-18.
20. R. J. Kirkpatrick, "Crystal growth from the melt: A review," American Mineralogist vol.
60 (1975), 798-814.
21. J. W. Christian, The Theory of Transformations in Metals and Alloys (Oxford: Pergamon, 1975).
