Effect of different amounts of slag on the crystallization behavior of glass-ceramics produced by natural cooling yellow phosphorus slag

Abstract

CaO–Al₂O₃–SiO₂ (CAS) glass-ceramics were prepared by a melting method using different amounts of natural cooling yellow phosphorus slag as the main material, and afterwards the effects of the slag on the crystallization and properties of glass-ceramics were studied by differential thermal analysis (DTA), X-ray diffraction (XRD) and scanning electron microscopy (SEM). The crystallization activation energy E of the glass-ceramic samples was calculated by modified Johnson-Mehl-Avrami (JMA) and Ozawa methods, and Avrami indices n of the samples were analyzed with Augis–Bennett equations. The results showed that the crystallization behavior of glass ceramics was surely affected with the addition of different amounts of slag. The crystallization activation energy calculated with the modified Johnson-Mehl-Avrami (JMA) equation ranged from 200.8678 kJ mol⁻¹ (Slag-1) to 380.0184 kJ mol⁻¹ (Slag-3) and from 220.7858 kJ mol⁻¹ (Slag-1) to 400.6569 kJ mol⁻¹ (Slag-3) by the Ozawa method. The crystallization indices increased from 3.20 (Slag-1) to 4.07 (Slag-3). The main crystalline phases were gehlenite (Ca₂Al₂SiO₇) and akermanite (Ca₂MgSi₂O₇) in Slag-1 and Slag-2, while anorthite (CaAl₂Si₂O₈) and akermanite (Ca₂MgSi₂O₇) were the main crystalline phases in Slag-3. The structure of the grains in the samples changed from spherical to columnar. The Slag-3 sample showed stronger acid resistance, alkali resistance and lower water absorption than Slag-1 and Slag-2.

---

Yellow phosphorus slag is formed in the generation process of yellow phosphorus with phosphate rock, coke and silica.¹ It is estimated that 8–10 tons of yellow phosphorus slag will be generated with the production of 1 ton of phosphorus.^2,3 Yellow phosphorus slag not only occupies a great deal of space but also seriously pollutes the environment in many countries. Consequently, how to convert yellow phosphorus slag into more valuable and environment-friendly materials is important for the protection of the environment and the sustainable development of our society.

Glass-ceramic is a fine-grained polycrystalline material synthesized by controlling the composition⁴ and the heat-treatment process.⁵ This kind of material is characterized by low price, low water absorption, high acid resistance and high mechanical properties. The major constituents of yellow phosphorus slag are CaO and SiO₂; the slag also contains oxides such as Al₂O₃, TiO₂, Fe₂O₃, P₂O₅, MgO, K₂O, Na₂O, etc. The yellow phosphorus slag can be suitable for glass-ceramic preparation as a new material with proper composition. Many kinds of metallurgical slag, such as blast furnace slag,^6–8 cooper slag,^4,9 ferrous tailing slag,¹⁰ sludge,¹¹ have been studied for the preparation of glass-ceramic. Cheng¹² introduced the CaO–Al₂O₃–SiO₂ glass-ceramics prepared by sintering with the addition of 42.32 wt% yellow phosphorus slag. The author pointed out that phosphorus and fluorine can decrease the crystallizing temperature and promote the nucleation and crystallization of glass-ceramics. Liu¹³ found that Al³⁺ cation in Al₂O₃ plays an important role in the formation of [AlO₄] tetrahedral or [AlO₆] octahedral structure. The higher Al₂O₃ content is, the better the performance of glass-ceramics will be, for the increase in binary basicity. Toya T. et al.¹¹ pointed out that the phase of gehlenite (Ca₂Al₂SiO₇) and wollastonite (CaSiO₃) can be generated above 900–950 °C in the preparation process of glass-ceramics with water purification process sludge. The characterization analysis of the glass-ceramics shows an excellent durability in alkali solution and Vickers microhardness. Yang et al.¹⁴ investigated the effect of Fe²⁺ and Fe³⁺ on the crystallization properties of CaO–Al₂O₃–SiO₂–MgO glass-ceramics. The results indicated that Fe²⁺ can decrease the crystallization temperature, and Fe³⁺ will improve the crystallization ability as a nucleating agent. It can be concluded that most work only focused on the sintering process of glass-ceramics, but the melting process of glass-ceramics from natural cooling yellow phosphorus slag has not been introduced and the crystallization process of different amounts of slag has not been clarified. To date, the CaO–Al₂O₃–SiO₂ glass ceramic system has been widely applied in the glass ceramic industry, where only a small amount of slag was used with plenty of adjuvant materials added. The present research aims at finding an easy and cost-effective way to make the best use of the yellow phosphorus slag and achieve the JC/T872-2000 standard at the same time. In consideration of the further application in industry, it is a cost-effective way to decrease the extra addition of the SiO₂ and Al₂O₃ during the production process of glass using the yellow phosphorus slag.

The aim of this work is to discuss the influence of slag amount on the crystallization and material properties like crystallization kinetics, phase composition, density, water absorption, acid resistance and alkali resistance, thus providing theoretical knowledge for increasing the amount of slag in natural cooling yellow phosphorus slag glass-ceramic.

1 Experimental design and methodology

The natural cooling yellow phosphorus slag (from Yunnan Province in China) was ground and sieved to 180 mesh, dried at 105 °C for 24 h, and then mixed with SiO₂, Al₂O₃. The powder mixture was heated at 1350 °C for 2 h in electric furnace. In order to eliminate internal stress, the as-cast and treated samples of the parent glass were firstly annealed at 600 °C for 2 h and then naturally cooled to the room temperature. Table 1 shows the chemical composition of the natural cooling yellow phosphorus slag analyzed by X-ray fluorescence (XRF).

Table 1 The composition of natural cooling slag from a chemical phosphorus enterprise in Yunnan Province

Component	CaO	SiO2	Al2O3	MgO	P2O5	F	Fe2O3	Na2O	K2O	Others
Content (wt%)	47.98	30.45	2.89	5.35	3.79	2.93	0.059	0.21	0.37	5.971

---

As shown in Table 1, the main components of the yellow phosphorus slag are CaO and SiO₂. In order to investigate the influence of slag amount, different amounts of slag was mixed with certain percentage of SiO₂ and Al₂O₃. Table 2 shows the composition of CAS glass-ceramics containing different amounts of slag.

Table 2 The composition of CAS glass ceramics with different amounts of slag

Sample no.	Slag	SiO2	Al2O3	Amount
Slag-1	82.35376	8.429412	9.216832	100
Slag-2	79.4455	12.32302	8.231481	100
Slag-3	61.05248	15.81371	23.1338	100

---

The thermal behaviors of the CAS glass samples during phase transformation were analyzed by differential thermal analyzer (TG/DTA, HCT-3) at different heating rates ranging from 10 °C min⁻¹ to 25 °C min⁻¹. The alumina powder was used as a reference material and 20 mg of glass sample after crushing and annealing to the size of 180 mesh (<0.1 mm) was put in a nitrogen atmosphere. The crystalline phase was investigated using a X-ray diffraction meter (D/max-2200) between 5° and 90° at a step of 3° min⁻¹ operating at 36 kV and 30 mA using Cu K_α radiation. The scanning electron microscopy (SEM; VEGA3-SBH) was used to observe the microstructure of the glass-ceramic samples, which had been etched for 30 s in 1.5% HF and then dried at 105 °C. The bulk density was determined by Archimedes method. According to the building material industry standard “JC/T872-2000 standard”, the water absorption as well as acid and alkaline resistance of the samples were analyzed.

2 Results and discussion

2.1 Phase diagram analysis

The phase diagram is a relevant source and valuable for the synthesis of glass-ceramic, and the CaO–Al₂O₃–SiO₂ is a common ternary system for glass-ceramic. In this study, the natural cooling yellow phosphorus slag is a component for the (CaO)_0.4459(SiO₂)_0.2641(Al₂O₃)_0.01475–SiO₂–Al₂O₃ ternary system. In order to get the phase of ASlag-liquid via (CaO)_0.4459(SiO₂)_0.2641(Al₂O₃)_0.01475–SiO₂–Al₂O₃ equilibrium diagram, the pure chemical reagent SiO₂ and Al₂O₃ must be added in natural cooling yellow phosphorus slag in proper proportion to make the composition falls in the red region. The samples of Slag-1, Slag-2 and Slag-3 were selected as shown in Fig. 1, and the compositions of CAS glass-ceramics containing different amount of slag were showed in Table 2.

Fig. 1 Phase diagram of (CaO)_0.4459(SiO₂)_0.2641(Al₂O₃)_0.01475–SiO₂–Al₂O₃ system.

2.2 Kinetic analysis

A certain amount of energy is required for the glass to overcome the phase and structure unit rearrangement barrier to transform into the crystalline in the crystallization process, i.e., the activation energy. Therefore, the glass research for the preparation of glass-ceramics can provide an important theoretical basis for the estimation of crystallization activation energy. Two methods, namely isothermal and non-isothermal methods, were used in studying the crystallization kinetics of glass. According to the isothermal method, the samples were heated above the glass transition temperature and kept at the optimal temperature for several hours. The glass samples were heated at a fixed heating rate and crystallized during the heating-up process with in the non-isothermal method. The isothermal method was slower and more difficult than the non-isothermal method. The modified Johnson-Mehl-Avrami method and Owaza method were widely used for the kinetic analyses of non-isothermal DTA data. Activation energy, one of the most important kinetic parameters, was obtained by experimental DTA data through modified Johnson-Mehl-Avrami method and Owaza method in this study. Fig. 2 shows the DTA curves at different heating rates of 10, 15, 20 and 25 °C min⁻¹ for different samples. It can be seen that the peak temperature of crystallization (T_p) shifted to a higher value as the heating rate increased. During the heating-up process, glass state materials will be transformed from short-range order atomic amorphous state to long-range order crystalline state. When the heating rate is slow, there is enough time for the proceeding of the transformation, thus the T_p value will be low. On the contrary, when the rate is fast, the T_p value will be high for the hysteresis of the transformation, but the instantaneous transformation rate is rapid as well as the crystallization rate.¹⁵

Fig. 2 DTA curves of Slag glass-ceramics. (a) Slag-1 sample, (b) Slag-2 sample, (c) Slag-3 sample.

The activation energy for the crystallization of the parent glass samples can be estimated from the DTA curves. The modified Johnson-Mehl-Avrami (JMA) equation^16,17 and Ozawa method¹⁸ were adopted to describe the glass crystallization ability via DTA data:

x = 1 − exp[−(kt)ⁿ].

x is the volume fraction of the transformed phase, n is the crystal growth index, k is the crystallization rate constant.

The crystallization rate constant k can be calculated with the Arrhenius equation:

According to Kissinger,^19,20 the activation energy of crystallization (E) can be obtained as follows:

where T_p is the crystallization peak temperature of the DTA curve, β is the heating rate, R is the gas constant, E is the crystallization activation energy and v is the frequency factor.

The equation proposed by Ozawa can be expressed as:

ln β = −E/RT_p + constant,

where T_p is the crystallization peak temperature of the DTA curve, β is the heating rate, R is the gas constant, E is the crystallization activation energy.

Hu Lili²¹ and Yang Qiuhong²² proposed the crystallization kinetic factor k as a more reasonable criterion than the activation energy of crystallization E for crystallization of glass. The higher the crystallization transition rate coefficient k value is, the worse the glass stability will be, and the greater the tendency to crystallization ability will be. As shown in Table 3, the crystallization activation energy of the CAS system glass ceramics increased with more slag. The crystallization transition rate coefficient of Slag-3 glass-ceramic is larger than other samples, so the Slag-3 glass-ceramic is easier to crystallize.

Table 3 The kinetic parameters of CAS glass ceramics with different amount of slag

Sample no.	β/(°C min^−1)	Tp (°C)	Modified Johnson-Mehl-Avrami (JMA)		Owaza
			E/(kJ mol^−1)	k	E/(kJ mol^−1)
Slag-1	10	904.9	200.8678	1.3211 × 10^8	220.7858
	15	937.1
	20	946.4
	25	949.2
Slag-2	10	920.1	205.2163	1.5702 × 10^8	225.3898
	15	954.1
	20	959.3
	25	965.5
Slag-3	10	955.6	380.0184	4.1645 × 10^15	400.6569
	15	968.7
	20	978.2
	25	984.8

---

Crystallization index²³ can be obtained through Augis–Bennett equation:

ΔT is the half-peak width.

The crystallization mechanism will be surface crystallization, one-dimensional crystallization, two-dimensional crystallization or three-dimensional crystallization when n = 1, 2, 3 or 4, respectively.²⁴

The vs. diagram plotted according to Fig. 2 is shown in Fig. 3 and the diagram of ln β vs. from Fig. 2 is shown in Fig. 4 illustrating the relationship between crystallization index change and the amount of slag.

Fig. 3 The diagram of vs. for slag glass-ceramics.

Fig. 4 The diagram of ln β vs. for slag glass-ceramics.

The crystallization indexes of CAS glass-ceramics with different amounts of slag are shown in Table 4. Slag-1 and Slag-2 are two-dimensional crystallization, and Slag-3 is three-dimensional crystallization.

Table 4 The crystallization indexes of glass ceramics with different amounts of slag

Sample	Slag-1	Slag-2	Slag-3
Crystallization index n	3.20	3.38	4.07

---

2.3 Phase analysis

From the comprehensive XRD profiles it can be seen that the amount of slag has an effect on the internal structure of CAS glass-ceramics too. Before the characterization analysis, the glass specimen were heated at 790 °C for 2 h, and then at 1080 °C for 1.5 h to obtain microcrystalline glass. The glass-ceramics powder was ground in an agate mortar. XRD analysis was then performed to identify the crystalline phase. The results are shown in Fig. 5:

Fig. 5 XRD patterns of glass-ceramics. (a) Slag-1 sample, (b) Slag-2 sample, (c) Slag-3 sample.

The main crystal phases of each sample were confirmed in Jade 5.0. Fig. 5 shows that the two main components in the samples are: gehlenite (Ca₂Al₂SiO₇ PDF: 35-0755) and akermanite (Ca₂MgSi₂O₇ PDF: 35-0592) in Slag-1 and Slag-2; anorthite (CaAl₂Si₂O₈ PDF: 41-1486) and akermanite (Ca₂MgSi₂O₇ PDF: 35-0592) in Slag-3. The higher the peak height is, the better the crystallization effect inside will be, thus the crystallization degree in Slag-2 is larger than Slag-1 as the peaks of Slag-2 are higher. This result is in agreement with the crystal growth index n calculated through the modified JMA equation. Slag-3 has a different crystalline phase, for the main crystalline phase in the sample is anorthite.

2.4 The macro morphology and micrograph analysis

The section photos of different samples are shown in Fig. 6. The color of Slag-1 sample shows a little yellow. With the decrease of the amount of slag, the colors of the samples gradually turn white. All samples are compact free of pores or interlayer glass.

Fig. 6 The photos of (a) Slag-1 sample, (b) Slag-2 sample, (c) Slag-3 sample.

The micrographs of glass-ceramics with different amounts of slag after corrosion in 1.5% HF solution 30 s are shown in Fig. 7 from which it can be seen that the crystal grain size changes with the addition of slag in glass-ceramics. For Slag-1 the grain of internal glass is smaller than Slag-2, and the results are the same as those shown in the XRD profiles. Slag-1 sample is two-dimensional crystallization, the grain is spherical, fine and uniform according to Fig. 7(a); the Slag-2 is also two-dimensional crystallization, but the grain size is larger and the structure is dispersible in Fig. 7(b); Slag-3 is three-dimensional crystallization, and the grain is big and shows a columnar distribution while the structure is compact which are the same as the results shown in the XRD profiles. With the aid of the macro morphology images and micrographs, summary can be made that decreasing the amount of slag in the preparation of glass-ceramics can promote the crystallization in CAS systems.

Fig. 7 SEM images of (a) Slag-1 sample, (b) Slag-2 sample, (c) Slag-3 sample.

2.5 Material properties

According to the JC/T872-2000 standard the material property tests of the Slag-1, Slag-2 and Slag-3 samples were undergone. The results are shown in Table 5.

From Table 5, the water absorption, acid resistance and alkali resistance of Slag-3 were less than Slag-1 and Slag-2 samples for their differences in crystalline phase composition and grain structure. Concretely, SiO₂ is a glass-forming oxide. Most silicon-oxygen tetrahedron [SiO₄] can form a continuous network of irregular structural units constituting the glass skeleton. SiO₂ in Slag-2 is more than that in Slag-1 and so as the silicon oxygen tetrahedron. The decrease of Ca/Si ratio will reduce the accumulation function of calcium oxide thus loosening the network structure of the residual glass phase. In CaO–Al₂O₃–SiO₂ glass-ceramic melts, Al³⁺ can fix the silicon-oxygen network by seizing nonbridging oxygen, forming aluminum–oxygen tetrahedral structure and finally reconnecting the broken network. The SiO₂ and Al₂O₃ in Slag-3 are more than those in Slag-1 and Slag-2, which will make a small amount of alkali metal ions or alkaline earth metal ions fill the gaps in aluminum-oxygen and silicon–oxygen tetrahedras and promote the formation of a dense structure of the melts. The stronger movability and binding of alkali metal ions in the glass phase can enhance the reactivity of the phase and simultaneously lower the resistance to chemical corrosion comparing to those in the crystalline phase. Sodium hydroxide can dissolve the glass phase,¹⁹ as the proportion of crystallization increases, the residual glass phase in the glass-ceramics decreases, and consequently the amount of glass-ceramics samples dissolved by sodium hydroxide decreases correspondingly.

Table 5 The material properties of the glass-ceramics with different amounts of slag

Sample	Density (g cm^−3)	Water absorption (wt%)	Acid resistance (wt%)	Alkali resistance (wt%)
Slag-1	2.3279	0.046	0.1334	0.0046
Slag-2	2.2469	0.015	0.1549	0.0067
Slag-3	2.5206	0.003	0.0954	0.0022

---

3 Conclusions

(1) Different amount of slag can influence the crystallization kinetic behavior of glass-ceramics. As the addition of natural cooling yellow phosphorus slag decreases, the crystallization activation energy of CAS system, the crystallization rate constant k and the crystallization indexes n will increase. This study provides theoretical instruction for the comprehensive utilization of natural cooling yellow phosphorus slag.

(2) The crystalline phase is a mixture of gehlenite (Ca₂Al₂SiO₇) and akermanite (Ca₂MgSi₂O₇) in Slag-1 and Slag-2. The phases of Slag-3 sample are anorthite (CaAl₂Si₂O₈) and akermanite (Ca₂MgSi₂O₇) prepared with the natural cooling yellow phosphorus slag.

(3) The macro morphology images and micrographs show that there are different forms of grain in different samples generated with different amounts of slag, the colors of slag glass-ceramics gradually turn white and the structure of the grains changes from spherical to columnar. Slag-3 sample shows stronger acid resistance and alkali resistance than Slag-1 and Slag-2, and a lower water absorption. This kind of material will prove a promising application in building decoration.

Acknowledgements

The authors thank the support of Fund for Analyzing and Testing of Kunming University of Science and Technology (20150471).

References

1. Z. Wang, M. Jiang, P. Ning and G. Xie, Ind. Eng. Chem. Res., 2011, 50, 12194–12202.
2. J. Chen, B. Zhao, X. Wang, Q. Zhang and L. Wang, Int. J. Miner., Metall. Mater., 2010, 17, 121–126.
3. Y. Su, G. Li and J. Xia, Can. J. Chem. Eng., 2009, 87, 610–613.
4. Z. Yang, Q. Lin, S. Lu, Y. He, G. Liao and Y. Ke, Ceram. Int., 2014, 40, 7297–7305.
5. J. Cao and Z. Wang, J. Alloys Compd., 2013, 557, 190–195.
6. Y. Zhao, D. Chen, Y. Bi and M. Long, Ceram. Int., 2012, 38, 2495–2500.
7. L. Gan, C. Zhang, J. Zhou and F. Shangguan, J. Non-Cryst. Solids, 2012, 358, 20–24.
8. H. Liu, H. Lu, D. Chen, H. Wang, H. Xu and R. Zhang, Ceram. Int., 2009, 35, 3181–3184.
9. Z. Yang, Q. Lin, J. Xia, Y. He, G. Liao and Y. Ke, J. Alloys Compd., 2013, 574, 354–360.
10. C. Liu, P. Shi, D. Zhang and M. Jiang, J. Iron Steel Res. Int., 2007, 14, 73–78.
11. T. Toya, A. Nakamura and Y. Kameshima, Ceram. Int., 2007, 33, 573–577.
12. J. Cheng, H. Qiu, H. Li and J. Xie, Key Eng. Mater., 2008, 368, 1412–1414.
13. Z. Liu, Y. Zong, H. Ma, W. Dai and D. Cang, Adv. Appl. Ceram., 2014, 113, 394–403.
14. Z. Yang, Y. Li, D. Cang, M. Diao and W. Guo, J. Mater. Sci. Technol., 2012, 20, 45–51.
15. Y. Wang, Crystallization behavior of glass-ceramics prepared from Baotou steel blast furnace slag[M], Metallurgical Industry Press, 2014.
16. M. Avrami, J. Chem. Phys., 1939, 7, 1103–1112.
17. W. Johnson and R. Mehl, Trans. Metall. Soc. AIME, 1939, 135, 396–415.
18. T. Ozawa, Polymer, 1971, 12, 150–158.
19. H. Kissinger, J. Res. Natl. Bur. Stand., 1956, 2l7–221.
20. H. Kissinger, Anal. Chem., 1957, 29, 1702–1706.
21. L. Hu and Z. Jiang, J. Chin. Ceram. Soc., 1990, 18, 315–321.
22. Q. Yang and Z. Jiang, J. Chin. Ceram. Soc., 1994, 22, 419–426.
23. J. A. Augis and J. E. Bennett, J. Therm. Anal. Calorim., 1978, 13, 283–292.
24. K. Matusita and S. Sakka, Phys. Chem. Glasses, 1979, 20, 81–84.

---