Viscosity Measurements of CaO–SiO₂–CrO Slag

Abstract

The viscosities of CrO-containing slags obtained using high temperature equilibrium technique were measured by the rotating cylinder method at 1823 to 1948 K (1550 to 1675°C). The slag viscosity decreased with increasing CrO contents. CrO also exhibits a strong basic character in slags, and can act as a network modifier to depolymerize the slag network structure in CaO–SiO₂-based slags. The dependence of the slag degree of polymerization (DOP) on the CrO content further validates these observations.

1. Introduction

Cr is an important alloying element for many different metallic materials. Good fluidity of a Cr-containing slag is critical to the smelting and refining of such materials. Previous studies^{1,2,3,4,5,6,7,8,9)} have been performed regarding the viscosity of Cr₂O₃-containing slags, demonstrating that an increase in Cr₂O₃ content leads to an increase in slag viscosity. However, Cr tends to exist primarily as Cr²⁺ in molten slag, particularly when equilibrating with metal melts under higher temperatures.^{10,11,12,13,14)} As such, the fluidity of CrO-containing (CrO is the oxide of Cr²⁺) slags is much more important for the smelting and refining operations of Cr-containing metal materials. However, experimental data on the viscosity of CrO-containing systems is greatly limited. Forsbacka et al.^7,8,9) measured the viscosity of CrOx-containing slag systems featuring both Cr²⁺ and Cr³⁺, and found that the addition of chromium oxide (CrO_x) into the slag decreases the viscosity; this effect becomes weaker at higher CrO_x contents. To get a better understanding of the property and structural information of CrO-containing slags, the viscosity of CaO–SiO₂–CrO (R, (%CaO)/(%SiO₂)=0.5) was measured by the rotating cylinder method at 1823–1948 K (1550–1675°C). Measurements were carried out in a totally closed system to ensure that the Cr in the molten slag existed solely as CrO. The degree of polymerization (DOP) of this slag was also analyzed.

2. Experimental Methods

The experimental arrangement is presented in Fig. 1. The measurement system contains a MoSi₂ furnace with a maximum temperature of 1973 K (1700°C), a rotary digital viscometer (Brookfield DV3T LV, Brookfield company, USA), a gas protection system (High purity Ar, ≥99.999%), and data collection software (Viscometer Test Software, Version Number HA1.10). The error of viscometer is within ±1% of the full-scale torque. The furnace and sample temperatures are measured by two B-type thermocouples (Omega, USA) with measurement accuracies of ±0.5%. The viscometer is protected by a transparent glass tube installed on an aluminum platform, which can be connected with the furnace tube to form a complete cavity. A vacuum gauge and vacuum pump are used to quickly evacuate the air inside the cavity. This equipment setup sufficiently ensures that the slag can be measured in a closed and controllable atmosphere at high temperatures.

Fig. 1.

Diagram of high temperature viscometer and its auxiliary device.

For the Brookfield’s viscometer, the constant is called the SMC (spindle multiplier constant).^7,9) It is calibrated using standard silicone oils (offered by Brookfield Company) with the viscosities of 100 cP, 500 cP, and 5000 cP. A constant temperature water bath is used to maintain the oils at 298 K during calibration, and for each type of silicone oil, its corresponding torque is measured using different speeds to obtain its optimum SMC. The final SMC is the average of the instrument constants calibrated for different standard silicone oils. A similar calibration operation has been described in detail elsewhere.⁹⁾

Reagent-grade CaO (CaCO₃ dissociation), SiO₂, Cr₂O₃ and Cr powder (≥ 99.95%, 400 mesh) with a total mass of 45 g were accurately weighed, mixed, and pressed into a cylindrical block according to the target slag composition. CrO in the slags was obtained in a stoichiometric amount as shown by reaction (1) using Cr₂O₃ and Cr powder. A high purity Cr sample (≥99.95%, 30 mm diameter, 2 mm thickness) was placed in the bottom of a molybdenum (≥99.7%) crucible, which ensured that the slag was in equilibrium with metallic Cr.   

$$\mathrm{C}{\mathrm{r}}_2{\mathrm{O}}_{3(\mathrm{s})}+\mathrm{C}{\mathrm{r}}_{(\mathrm{s})}=3\mathrm{C}\mathrm{r}{\mathrm{O}}_{(\mathrm{l})}$$(1)

When the furnace temperature was raised to 1948 K, the slag viscosity was measured each hour for five hours to determine if the system had reached an equilibrium state. For the three CrO-containing slags, the slag viscosity at 5 h was ≤ 1.5% different than the value at 4 h, and the deviation of the two viscosities was close to the system error of the viscometer (±1%). Thus, five hours was set as the equilibrium time. The furnace was then programmed to cool at a rate of 3 K/min to certain lower temperatures, and the temperature was equilibrated for 20–30 min before each measurement. The average of the 60 stable viscosity values recorded at three rotating speeds was the viscosity of the slag at the corresponding temperature (1948 K, 1923 K, 1898 K, 1873 K, 1848 K, 1823 K). Finally, the crucible was removed and broken, the slag was removed and crushed to 200 mesh for composition measurements (XRF, Zetium). The measured composition and viscosities of the CaO–SiO₂–CrO system in equilibrium with Cr are shown in Table 1.

Table 1. Composition and viscosities measured of CaO–SiO₂–CrO slags in equilibrium with Cr.

No.	Composition wt.%			R	Viscosity (Pa‧S)
	CaO	SiO2	CrO		1948 K (1675°C)	1923 K (1650°C)	1898 K (1625°C)	1873 K (1600°C)	1848 K (1575°C)	1823 K (1550°C)
0	33.8	66.2	0	0.51	0.990	1.146	1.343	1.573	1.864	2.221
1	32.7	63.5	3.8	0.52	0.683	0.777	0.894	1.037	1.212	1.432
2	31.9	61.8	6.3	0.52	0.561	0.639	0.734	0.847	0.989	1.170
3	30.5	60.0	9.5	0.51	0.440	0.494	0.564	0.646	0.750	0.879

3. Results and Discussion

3.1. Divalent Chromium Ratio Prediction

A phase diagram of the CaO–SiO₂–CrO system in equilibrium with metallic Cr at the lowest measurement temperature (1823 K) was calculated using the thermochemical software FactSage 7.0¹⁵⁾ as shown in Fig. 2(a), and was used to further determine the valence state of Cr in slag under our experimental conditions. The compositions selected for viscosity measurements were projected into this phase diagram, all of which were in the complete liquid phase zone. The basicity (R) of slags was very close to 0.5, which promotes the existence of Cr²⁺ based on the X_CrO/X_CrO1.5 ratio prediction equation obtained by Wang et al.¹⁶⁾ Furthermore, the ratio of divalent to trivalent chromium ((%Cr²⁺)/(%Cr³⁺)) in the slag with a basicity of 0.5 was predicted by the software under different CrO contents and temperatures. As shown in Figs. 2(b) and 2(c), this value favored Cr²⁺ by several orders of magnitude, which reflects that the slags studied here were CaO–SiO₂–CrO ternary systems.

Fig. 2.

(a) Slag compositions shown in the phase diagram of the CaO–SiO₂–CrO system in equilibrium with metallic Cr; (b) (c) effects of CrO content and temperature on (%Cr²⁺)/(%Cr³⁺) at a slag basicity of 0.5.

3.2. Viscosity of CaO–SiO₂–CrO Slags

The viscosity of the CaO–SiO₂–CrO slag as a function of CrO content is shown in Fig. 3(a). The viscosity of the 0# sample (CaO–SiO₂–0%CrO Slag) was calculated by the viscosity module of FactSage 7.0 based on the quasi-chemical model,^17,18) which reasonably agreed with the measured data. For the four slags measured at a basicity of 0.5, the slag viscosity decreases with increasing CrO contents. Partial viscosity values under similar conditions measured by Forsbacka et al.⁸⁾ are also plotted here, and analysis after each measurement showed that the Cr is mainly present as Cr²⁺. The viscosity data obtained by Forsbacka et al.⁸⁾ is lower than that found in this study, which may due to the higher basicity of their CaO–SiO₂–CrO_x system (R=0.64 and 1); Forsbacka et al.^7,8) eventually concluded that Cr²⁺ has strong basicity. Wang et al.¹⁹⁾ studied the effect of MgO and Al₂O₃ on the valence of Cr in CaO–SiO₂–CrO_x systems by X-ray photoelectron spectroscopy (XPS) and found that Cr²⁺ could be used as a network modifier. In addition, Mills²⁰⁾ suggested that the optical basicity of CrO tends to be 1.0. These findings further illustrate that CrO has strong basic characteristics and can dissociate to Cr²⁺ and O²⁻ in silicate slags. While O⁰ is a bridging oxygen bonded to two Si atoms, O²⁻ is free oxygen not bonded to any Si atoms, and O⁻ is a non-bridging oxygen bonded to one Si atom; free oxygen O²⁻ can depolymerize the complex Si–O network structure according to reaction (2), resulting in decreased viscosity.   

$${\mathrm{O}}^0+{\mathrm{O}}^{2-}=2{\mathrm{O}}^{-}$$(2)

Fig. 3.

(a) Effect of CrO content on the slag viscosity; (b) relationship between slag viscosity and temperature.

The relationship between the logarithm of the slag viscosity and the reciprocal of temperature is shown in Fig. 3(b), denoting a good linear relationship (R²≥0.999) between lnη and 10000/T. The activation energies of the four slags were 191.0, 175.1, 173.2, and 163.7 kJ‧mol^–1 corresponding to 0#, 1#, 2#, and 3# samples, decreasing with increasing CrO contents. This further illustrates that CrO tends to decrease the flow resistance of molten slags and then decrease its viscosity.

3.3. DOP Variation of Slags

The viscosity of silicate slag is closely related to its degree of polymerization (DOP).²¹⁾ The optical basicity (Λ) and NBO/T can characterize the slag DOP to further illustrate variations in slag viscosity. NBO/T represents the number of non-bridging oxygen atoms connected to network forming atoms (such as Si, Al, P, etc.), using a formula established previously.²¹⁾ The optical basicity reflects the activity of free oxygen (O²⁻) in the slag.²²⁾ Figure 4 shows that NBO/T and the optical basicity of slag increases with increasing CrO content, reflecting that the DOP of molten slag gradually decreases with increasing CrO content, which is consistent with the dependence of slag viscosity on CrO contents.

Fig. 4.

Effect of CrO content on non-bridging oxygen and optical basicity of slags.

Cr-containing slags are generally characterized by high viscosity.^7,23) This is primarily due to the fact that Cr₂O₃ has a low solubility in silicate slags, and is easily combined with divalent metal oxides to form high melting point spinels such as FeO‧Cr₂O₃ or MgO‧Cr₂O₃ suspended in the slag^1,2,4) and can also form Cr–O–Cr network structures embedded in Si–O–Si structures.^4,5,6) A large number of Cr-containing phases often leads to a sharp increase in Cr-containing slag viscosity or even slag hardening, which is disadvantageous for chromium recovery. The formation of divalent chromium therefore benefits for the reduction of slag viscosity.

4. Conclusions

The viscosity of CaO–SiO₂–CrO (R=0.5) slag was measured from 1823 to 1948 K. Increasing the CrO content reduces the viscosity of the CaO–SiO₂–based slag. CrO has basic characteristics, and can act as a network modifier to depolymerize slag network structures, based on calculated DOP variations of the slags tested here. The slag exhibited good Newtonian fluid behavior at high temperatures, and the activation energy of viscous flow decreased with increasing CrO content. It is suggested that a reducing atmosphere can change Cr₂O₃ to CrO, which will facilitates the improvement of Cr-containing slag fluidity.

Acknowledgements

The authors would like to express appreciation to the National Natural Science Foundation (No. 51674022) for its financial support of this research.

References

• 1)   E.  Minami,  M.  Amatatsu and  N.  Sano: Tetsu-to-Hagané, 73 (1987), S871 (in Japanese).
• 2)   G.  Qiu,  L.  Chen,  J.  Zhu,  X.  Lv and  C.  Bai: ISIJ Int., 55 (2015), 1367.
• 3)   C.  Xu,  W.  Wang,  L.  Zhou,  S.  Xie and  C.  Zhang: Metall. Mater. Trans. B, 46B (2015), 882.
• 4)   W.  Huang,  Y.  Zhao,  S.  Yu,  L.  Zhang,  Z.  Ye,  N.  Wang and  M.  Chen: ISIJ Int., 56 (2016), 594.
• 5)   R. Z.  Xu,  J. L.  Zhang,  Z. Y.  Wang and  K. X.  Jiao: Steel Res. Int., 87 (2016), 1.
• 6)   Q.  Li,  J.  Gao,  Y.  Zhang, Z. An and Z. Guo: Metall. Mater. Trans. B, 48B (2017), 346.
• 7)   L.  Forsbacka: Dr. thesis, Helsinki University of Technology, (2007), http://urn.fi/urn:nbn:fi:tkk-010975, (accessed 2007-12-07).
• 8)   L.  Forsbacka and  L.  Holappa: VII Int. Conf. on Molten Slags, Fluxes and Salts, South African Institute of Mining and Metallurgy, Johannesburg, (2004), 129.
• 9)   L.  Forsbacka,  L.  Holappa,  A.  Kondratiev and  E.  Jak: Steel Res. Int., 78 (2007), 676.
• 10)   F.  Körber and  W.  Oelsen: Mitt. K.-Wilh.-Inst. Eisenforsch. Düsseld., 17 (1935), 231.
• 11)   P.  Herasymenko: Trans. Faraday Soc., 34 (1938), 1254.
• 12)   W. J.  Rankin and  A. K.  Biswas: Arch. Eisenhuttenwes., 50 (1979), 7.
• 13)   S. T.  Tsai: Dr. thesis, The Pennsylvania State University, (1981), https://search.proquest.com/docview/303167026?accountid=28857, (accessed 2016-05-23).
• 14)   R. H.  Pan and  R. H.  Eric: J. S. Afr. Inst. Min. Metall., 104 (2004), 583.
• 15)  FactSage 7.0: http://www.factsage.com/, (accessed 2015-10-08).
• 16)   L.  Wang and  S.  Seetharaman: Metall. Mater. Trans. B, 41B (2010), 946.
• 17)   S. A.  Decterov,  A. N.  Grundy and  A. D.  Pelton: Proc. VIII Int. Conf. on Molten Slags, Fluxes and Salts, Gecamin, Santiago, (2009), 423.
• 18)   A. N.  Grundy,  H. C.  Liu,  I. H.  Jung,  S. A.  Decterov and  A. D.  Pelton: Int. J. Mater. Res., 99 (2008), 1185.
• 19)   L.  Wang,  J.  Yu,  K.  Chou and  S.  Seetharaman: Metall. Mater. Trans. B, 46B (2015), 1802.
• 20)   K.  Mills: Short course presented as part of Southern African Pyrometallurgy 2011, South African Institute of Mining and Metallurgy, Johannesburg, (2011), 1.
• 21)   K. C.  Mills: ISIJ Int., 33 (1993), 148.
• 22)   X. H.  Huang: Principles of Steel Metallurgy, 3rd ed., Metallurgical Industry Press, Beijing, (2013), 211.
• 23)   M.  Kekkonen,  H.  Oghbasilasie and  S.  Louhenkilpi: Research Report, Aalto University, Helsinki, (2012), 28.