Oxygen Contents in Molten Ni in Equilibrium with Slag of CaO–SiO₂–MgO–NiO System

Abstract

A study has been carried out to measure the oxygen contents in molten Ni in equilibrium with NiO in the top slag, contained in a magnesia crucible, in the temperature range higher than 1500°C. The slag was composed of CaO–SiO₂–MgO satd.-NiO system with (mass% CaO)/(mass% SiO₂) = 1 and 8 mass% NiO. Firstly, it was found that oxygen content rapidly increased after adding the slag of CaO–SiO₂–MgO–NiO system into molten Ni and that it took 30 minutes to attain equilibrium. The measured oxygen contents were lower than those in equilibrium with pure solid NiO and increased with an increase in temperature. Thermodynamic analysis was made on the basis of the above results. It was found that activity of NiO in the present slag of CaO–SiO₂–MgO satd.-NiO system was positively deviated from ideal solution. In addition, Henry constant of NiO in this oxide system was obtained as the following relation under the assumption that Henry’s law is available:   

under the slag of CaO–SiO₂–MgO satd.-NiO with (mass% CaO)/(mass% SiO₂) = 1.

O content could be estimated as follows by applying obtained in this study:   

under the slag of CaO–SiO₂–MgO satd.-NiO with (mass% CaO)/(mass% SiO₂) = 1.

1. Introduction

Ni shows excellent corrosion resistance in alkali environments because Ni is a metal nobler than Fe. One of the applications of Ni alloy such as UNS N02201 is an electrode for a caustic soda plant to produce soaps. During the refining process of those Ni alloys, oxygen contents in equilibrium with the top slag at the oxidizing stage of molten Ni should be understood to accurately control the compositions at the subsequent deoxidation stage.

So far, a number of measurements have been reported as for the solubility of oxygen in molten Ni. In other words, oxygen content in molten Ni in equilibrium with pure solid NiO was clarified. Some representative results of the solubility at 1600°C are described below.

A phase diagram of Ni–O binary system tells us that the solubility is approximately 8300 mass ppm.¹⁾ Compiled data measured by various researchers²⁾ shows that the solubility ranges from 6300 to 9500 mass ppm. 5500 mass ppm has been deduced by combining the thermodynamic data of oxidation of molten Ni and dissolution of oxygen gas into molten Ni.³⁾ It can be pointed out that, although the data exhibit a relatively large extent of scattering, the solubility of oxygen in molten Ni in equilibrium with pure solid NiO is much higher than that of molten Fe in equilibrium with FeO which is 2300 mass ppm.⁴⁾

Accounting for the above results, it is presumed that oxygen contents in molten Ni in equilibrium with NiO in the top slag are higher than those after decarburization stages during the steelmaking processes of Fe based alloys. However, very few researches have been so far carried out for the purpose of the refining process of Ni alloys. A study was previously performed by taking various systems into account by Woo et al.⁵⁾ Through an extensive literature review, they carried out the thermodynamic optimization of phase equilibrium and thermodynamic properties of all available oxide phases CaO–NiO–SiO₂, CaO–MgO–NiO–SiO₂, CaO–MgO–NiO, NiO–SiO₂ and MgO–NiO systems.

In the investigations in the field of geology, some studies were performed focusing on NiO-bearing rock and silicate melt systems as summarised in Table 1.^5,6,7,8,9) In the studies^6,7,8,9) of geology, their work was on revealing standard free energy changes for formation of compounds, such as CaNiSi₂O₆ with clinopyroxene structure, and activity coefficients of NiO in the silicate melts. It is realised that the experiments were conducted under relatively lower temperatures than those for the typical refining process. This was because researchers of geology paid attention to magmas. The temperature ranges of interest as for magmas are usually below 1400°C. Therefore, these studies were not conducted in the temperature range of the refining process of molten Ni. Refining of Ni has to be proceeded in the temperature ranges higher than 1500°C because the melting point of Ni is 1452°C. The above literature survey shows that slag chemistry for the Ni refining process is still lacking.

Table 1. Previous investigations.

Investigators	Oxide systems	Temperature range (°C)	Ref. No.
Woo, Lee and Jung	CaO–NiO–SiO2, CaO–MgO–NiO–SiO2, CaO–MgO–NiO, NiO–SiO2, MgO–NiO	1340–1550	5
Biggar	CaO–NiO–SiO2	1340–1550	6
Pretorius and Muan	CaO–NiO–SiO2, CaO–MgO–NiO–SiO2	1400–1435	7
Mukhopadhyay and Jacob	CaO–NiO–SiO2	1100	8
Holzheid, Palme and Chakraborty	CaO–MgO–SiO2–NiO–FeO–Al2O3–CoO (magma composition)	1400–1600	9

The refining processes of molten Ni alloys consist not only of deoxidation stages¹⁰⁾ but also of oxidising stages, which come at the beginning of melting processes. Based on this back ground, a study has been carried out to measure the oxygen contents in molten Ni in equilibrium with NiO contained in the top slag composed of CaO–SiO₂–MgO–NiO system in the temperature range higher than 1500°C. Finally, Henry constant of NiO in this system was obtained to estimate the oxygen contents under various conditions of NiO contents and temperatures.

2. Experimental

300 grams of pure Ni contained in an MgO crucible (ϕ50 mm×ϕ70 mm×90 mmH) was placed in a vertical furnace as illustrated in Fig. 1. The purity of Ni was confirmed to be higher than 99.99 mass%. The specimen was heated up to the aimed temperatures of 1500, 1550, 1600 and 1650°C under an Ar gas atmosphere. Ar gas was flown at a rate of 1 L/minute. After attaining the aimed temperature, it was kept for 30 minutes to homogenise the temperature of the molten Ni. This was followed by the first sampling whose timing was treated as time zero. Sampling method was that the silicon rubber cap of the top lid was taken off and a silica tubing, inside which was evacuated, was inserted to the melts. Therewith, the molten Ni was sucked into the tubing.

Fig. 1.

Schematic illustration of vertical furnace.

Thereafter, 40 grams of the mixture of CaO–SiO₂–MgO–NiO oxide powder was added onto the molten Ni from the top of the furnace. The pure reagents of each oxide were uniformly mixed before the experiments. Usually pre-melting is done to create mother slags. But, it was difficult to pre-melt for the present system because NiO is easily reduced or decomposed.

For the slag compositions, it should be noted that the mass ratio of (mass% CaO)/(mass% SiO₂) was fixed as unity and MgO was fixed as 10 mass% simulating the slag system for oxidizing stage of molten Ni. The NiO concentrations were varied depending on the temperatures; the target content was set to be 8 mass% in the slags after the decomposition to supply oxygen into the molten Ni. The blend conditions of each oxide are listed in Table 2. For example, it was 39 mass% CaO – 39 mass% SiO₂ – 10 mass% MgO – 12 mass% NiO in the case of 1600°C. These concentrations were more precisely adjusted based on the results obtained by the first experiment of 1550°C.

Table 2. Slag compositions (mass%).

Temperature (°C)		CaO	SiO2	MgO	NiO	XNiO
1500	Blend	39.5	39.5	10.0	11.0	0.091
	Analysis	39.8	40.7	11.8	7.7	0.061
1550	Blend	40.0	40.0	10.0	10.0	0.082
	Analysis	40.3	40.5	12.4	6.8	0.053
1600	Blend	39.0	39.0	10.0	12.0	0.101
	Analysis	38.6	38.8	14.9	7.7	0.060
1650	Blend	38.7	38.7	10.0	12.6	0.106
	Analysis	38.1	38.5	15.6	7.8	0.061

After the addition of the top slag, the molten Ni was sampled at 5 (partly), 10, 30, 60, 120 and 180 minutes by a silica tube, as described above, to analyse oxygen contents. After understanding how long it would take to attain equilibrium state, the samples were taken till 30 minutes. Besides, at 30 minutes, a small amount of slag was sampled using a quartz rod of 2 mm in diameter.

Analysis of oxygen was conducted by an inert gas fusion method (Horiba; EMGA 530). The slags were mounted, polished and finally coated by Au to measure NiO contents in the slags by SEM (Scanning Electron Microscope; HITACHI S-3500N) equipped with EDS (Energy Dispersive X-ray Spectrometry; HORIBA EMAX-7000). Analyses were conducted with a low magnification of 100 times to measure the relatively wide areas to obtain as average values as possible.

3. Results

3.1. Variation of Oxygen Content in the Molten Ni

The variation of O content as time at 1550°C is shown in Fig. 2. It is obvious that O content rapidly increases to 1100 ppm after 10 minutes of slag addition. Further, O content takes almost constant as 1200 ppm after 30 minutes. Thereby we judged that 30 minutes was long enough to have equilibrium state. Figure 3 shows O content of every temperature indicating that O content increases with an increase in temperature. Hereinafter the values of 30 minutes were applied for the analyses.

Fig. 2.

Variation of O content as time at 1550°C.

Fig. 3.

Variation of O content as time at every temperature.

It was very important to check the mass balance of oxygen. O content in the molten Ni was calculated to verify if NiO in the slag had an only role to supply O into the molten Ni. Figure 4 and Table 2 show the NiO contents of the sampled slags. In Fig. 4, the horizontal line at 8 mass% was the target value implying that NiO contents were mostly controlled at 8 mass%. To further confirm it, oxygen content in the gas atmosphere was measured in the experiment of 1550°C by an oxygen sensor. The sensor was connected with the rubber tubing from the outlet of Ar gas, showing that it did not detect oxygen gas. These results successfully proved that the surplus NiO over 8 mass% was consumed only to supply oxygen atoms into the molten Ni.

Fig. 4.

Analysed NiO contents compared with the target value.

3.2. Temperature Dependence

The data at 30 minutes were plotted against temperature as shown in Fig. 5. It can be seen that O content in equilibrium with the present slag is much lower than that in equilibrium with NiO. The temperature dependence is the same; O content increases with increasing temperature. The lower O content under the condition with slag could be owing to the activity of NiO in the slag system that is below unity. Therefore, thermodynamic considerations were made in the following sections.

Fig. 5.

Temperature dependence of O content in molten Ni.

3.3. Slag Composition

As provided in Table 2, it was confirmed that the ratios of (mass% CaO)/(mass% SiO₂) were kept as almost unity within measurement error. However, it was found that MgO content increased with increasing temperature. According to the CaO–SiO₂–MgO–NiO quaternary phase diagram,⁵⁾ the measured MgO contents were close to the liquidus lines of each temperature. Thus the present slag systems should be expressed to be CaO–SiO₂–MgO satd.-NiO with (mass% CaO)/(mass% SiO₂) = 1.

4. Discussion

4.1. Activity of NiO

It has been confirmed that NiO in the slag phase decomposes according to the following reaction:   

$$\text{(NiO)}=\text{Ni(l) }+\underset{\_}{\text{O}}$$(1)

Equilibrium constant of the above reaction is expressed as the following equation:³⁾   

$$\text{log }K=log\left(\frac{a_{\text{O}}}{a_{\text{NiO}}}\right)=\frac{-8\mathrm{\hspace{0.17em} }870}{T}+4.48$$(2)

where a_O is the activity of oxygen in molten Ni taking standard state as infinite dilute solution using mass% as unit, and a_NiO is the activity of NiO taking pure solid as standard state. This equation is written as the following expression considering activity coefficient of oxygen in the molten Ni (f_O):   

$$\text{log}\left(f_{\text{O}}\left[\text{mass\%O}\right]\right)-log{a}_{\text{NiO}}=\frac{-8\mathrm{\hspace{0.17em} }870}{T}+4.48$$(3)

The definition of activity coefficient gives the following relation using interaction parameter between oxygen atoms in molten Ni.   

$$\text{log }{f}_{\text{O}}=e_{\text{O}}^{\text{O}}[\text{mass\%O}]$$(4)

Here, because $e_{\text{O}}^{\text{O}}=0$¹¹⁾ is reported, finally one can obtain the following simple equation:   

$$log{a}_{\text{NiO}}=\text{log}\left[\text{mass\%O}\right]+\frac{8\mathrm{\hspace{0.17em} }870}{T}-4.48$$(5)

Activities of NiO can be calculated by Eq. (5) because oxygen contents were measured as shown in Fig. 3. The O contents were 1047, 1179, 1470 and 1821 mass ppm at 1500, 1550, 1600 and 1650°C, respectively. Substituting these values, NiO activities were obtained to be 0.35, 0.29, 0.26 and 0.25 at 1500, 1550, 1600 and 1650°C, respectively. Figure 6 shows the temperature dependence of activity of NiO indicating that NiO activity decreases with an increase of temperature. The NiO content of 8 mass% corresponds to 0.06 as molar fraction in the present slag. The dotted line shows the state of ideal solution. This can prove that activity of NiO is deviated positively. It is well known that the more ideal any solution becomes, the higher the temperature becomes. It can be said that our result is consistent with this rule.

Fig. 6.

Temperature dependence of NiO activity.

4.2. Activity Coefficient of NiO

According to the work by Pretorius,⁷⁾ it is quite appropriate to assume that Henry’s law is available within the range of NiO contents which are 0.06 as molar fraction. Henry constant ${\gamma }_{\text{NiO}}^{°}$ can be obtained by the following equation with substituting the X_NiO values in Table 2.   

$$\text{log }{\gamma }_{\text{NiO}}^{°}=log{a}_{\text{NiO}}-log{X}_{\text{NiO}}$$(6)

Henry constant ${\gamma }_{\text{NiO}}^{°}$ was plotted in Fig. 7 as a function of reciprocal of absolute temperature. Clearly, it is expressed as a linear function and ${\gamma }_{\text{NiO}}^{°}$ decreases with increasing temperature. Least square regression of these values gives us the following relation:   

$$\text{log }{\gamma }_{\text{NiO}}^{°}=\frac{3\mathrm{\hspace{0.17em} }710}{T}-1.33$$(7)

under the slag of CaO–SiO₂–MgO satd.-NiO with (mass% CaO)/(mass% SiO₂) = 1.

Fig. 7.

Henry constant of NiO in the slag as a function of temperature along with the previous studies.^5,7)

It is worthy to notice that our results are in good agreement with the previous data^5,7) whose oxide systems are close to ours. In particular, the agreement with Pretorius’ value can assure that the relation of Eq. (7) is able to be extrapolated to lower temperatures.

This agreement can prove that some assumptions described in the previous sections are valid. At first, it is quite appropriate to have employed Eq. (2) for the dissolution reaction of NiO into molten Ni shown by Eq. (1). Secondly, it is quite valid to assume that Henry’s law is available in the present NiO contents.

4.3. O Content Estimated by Using ${\gamma }_{\mathrm{N}\mathrm{i}\mathrm{O}}^{°}$

Combination of Eqs. (5) and (7) provides the following relation to estimate O content under the present slag system.   

$$\text{log}\left[\text{mass\%O}\right]=log{X}_{\text{NiO}}-\frac{5\mathrm{\hspace{0.17em} }160}{T}+3.15$$(8)

under the slag of CaO–SiO₂–MgO satd.-NiO with (mass% CaO)/(mass% SiO₂) = 1.

Estimated O content by the above relation is given in Fig. 8 assuming that the slag melt obeys Henry’s law in the range of X_NiO lower than 0.1. For convenience, mass% NiO is taken at the top axis. One can see that O content after the oxidation stage of molten Ni reaches as high as 2000 to 3500 ppm under the oxidizing stage of molten Ni. It is helpful for the deoxidation stage, that is followed by the oxidizing stage, to determine the amount of deoxidant.

Fig. 8.

Calculated O content as a function of temperature.

5. Conclusions

A study was carried out to measure the oxygen contents in molten Ni in equilibrium with NiO in the top slag composed of CaO–SiO₂–MgO satd.-NiO system with (mass% CaO)/(mass% SiO₂) = 1 and 8 mass% NiO. In addition, Henry constant of NiO in this oxide system was obtained. The following words summarise the present study:

(1) It was found that oxygen content rapidly increased after adding the slag of CaO–SiO₂–MgO–NiO system and that it took 30 minutes to attain equilibrium.

(2) Oxygen contents were lower than those in equilibrium with NiO and increased with increasing temperature.

(3) Activity of NiO in the present slag of CaO–SiO₂–MgO satd.-NiO system was positively deviated from ideal solution.

(4) Under the assumption that Henry’s law is available as for NiO in this system, the obtained activity coefficients of NiO gave us the following relation:   

$$\text{log }{\gamma }_{\text{NiO}}^{°}=\frac{3\mathrm{\hspace{0.17em} }710}{T}-1.33$$

under the slag of CaO–SiO₂–MgO satd.-NiO with (mass% CaO)/(mass% SiO₂) = 1.

(5) O content could be estimated as the following equation by applying ${\gamma }_{\text{NiO}}^{°}$ obtained in this study.   

$$\text{log}\left[\text{mass\%O}\right]=log{X}_{\text{NiO}}-\frac{5\mathrm{\hspace{0.17em} }160}{T}+3.15$$

under the slag of CaO–SiO₂–MgO satd.-NiO with (mass% CaO)/(mass% SiO₂) = 1.

References

• 1)   M.  Hansen and  K.  Anderko: Constitution of binary alloys, McGraw-Hill, New York, (1958), 1050.
• 2)   N.  Kemori,  I.  Katayama and  Z.  Kozuka: J. Chem. Thermodyn., 13 (1981), 313.
• 3)   Y.  Miyazaki,  K.  Mizuno,  A.  Hongo,  T.  Ishii and  H.  Todoroki: Tetsu-to-Hagané, 93 (2007), 518.
• 4)  The 19th Committee on Steelmaking: Steelmaking Data Sourcebook, The Japan Society for the Promotion of Science, Gordon and Breach Science Publications, New York, (1988).
• 5)   D.-H.  Woo,  H.-G.  Lee and  I.-H.  Jung: J. Eur. Ceram. Soc., 31 (2011), 43.
• 6)   G. M.  Biggar: J. Am. Ceram. Soc., 52 (1969), 316.
• 7)   E. B.  Pretorius and  A.  Muan: J. Am. Ceram. Soc., 75 (1992), 1490.
• 8)   S.  Mukhopadhyay and  K. T.  Jacob: Metall. Trans., 26A (1995), 2311.
• 9)   A.  Holzheid,  H.  Palme and  S.  Chakraborty: Chem. Geol., 139 (1997), 21.
• 10)   Y.  Kanbe,  H.  Todoroki,  Y.  Kobayashi and  K.  Shitogiden: ISIJ Int., 49 (2009), 837.
• 11)   F.  Ishii,  S.  Ban-ya and  M.  Hino: ISIJ Int., 36 (1996), 25.