Solubility of Na₂O·xAl₂O₃ in NaF-KF-AlF₃-based low-temperature electrolyte

Abstract

The solubility of Na₂O·xAl₂O₃ in a NaF-KF-AlF₃-based low-temperature electrolyte was investigated, and the influence of electrolyte temperature at 750 ℃, CR = 1.3–1.4, CaF₂, and LiF on the alumina solubility was studied. The dissolution rate of alumina in the electrolyte was determined by measuring the alumina concentration in the electrolyte at different times with an oxygen analyzer. The results showed that in the NaF-KF-AlF₃-based low-temperature electrolyte, the dissolution rate of alumina in Na₂O·xAl₂O₃ was faster than that of Al₂O₃. As the Na₂O content in alumina increased, the dissolution rate increased. Similarly, a higher CR of NaF-KF-AlF₃ and a higher temperature of the molten salt (750–850 °C) increased the dissolution rate of alumina in the electrolyte. The dissolution rate of alumina in NaF-KF-AlF₃ decreased after adding CaF₂ and LiF. According to the kinetics analysis, the dissolution rate of alumina followed the zero-order reaction rate law from 0–5 min, which showed that it depended on the amount of alumina absorbed in the electrolyte in the first five minutes.

1 Introduction

Hall-Héroult cryolite-alumina molten salt electrolysis is the primary method for producing aluminum, but it uses carbon anodes that are directly consumed, which generates about 1.5 tons of CO₂ per ton of aluminum. Furthermore, this process emits about 0.75 (t CO₂e/t Al) of perfluorocarbons (PFCs), and the primary aluminum industry is faced with increasingly strict global restrictions on greenhouse gas emissions [1]. The Hall-Héroult method has become very mature after more than 130 years of continuous improvements, and there is very limited room for technological progress.

Carbon-free aluminum electrolysis based on inert anodes uses low-consumption inert electrodes instead of carbon anodes. This process emits oxygen during electrolysis and produces no greenhouse gases or hazardous gases such as CO₂, SO₂, or PFCs. Low-temperature aluminum electrolysis can extend the service life of inert electrodes; thus, inert electrode aluminum electrolysis is more likely to be successful only when combined with low-temperature aluminum electrolysis to achieve zero carbon emissions during aluminum electrolysis [2, 3].

Since Sleepy et al [4] first proposed the concept of low-temperature aluminum electrolysis in 1979, researchers have carried out many subsequent studies on it [5,6,7,8]. However, the lower the temperature, the slower the dissolution rate of alumina, especially below 750 ℃. The biggest problem of low-temperature aluminum electrolysis is that the dissolution rate of alumina in the electrolyte is too slow to keep pace with the electrolytic consumption rate of alumina. This can lead to catastrophic corrosion of the anode and contamination of primary aluminum using non-carbon inert anodes [9].

Beck [10] proposed the concept of suspended alumina electrolysis to increase the alumina dissolution rate during low-temperature aluminum electrolysis. The method involved placing the anode at the bottom of the electrolytic cell. During electrolysis, the gases produced by the anode suspend alumina particles, but this increases the costs and reduces the current efficiency.

Although the KF-AlF₃ system has a lower liquidus temperature and better alumina solubility, it inevitably transforms into a NaF-KF-AlF₃ system as electrolysis proceeds, and sodium in the alumina feedstock accumulates in the electrolyte [11,12,13]. A preliminary study [8] showed that the introduction of a small amount of NaF (0–5 mol%) into the KF-AlF₃ electrolyte with a cryolite ratio (CR = M_KF/M_AlF3) of 1.2–1.7 sharply increased the initial crystallization temperature. This easily caused the electrolyte to crust at the cathode during electrolysis, making the KF-AlF₃ electrolyte unsuitable for industrial applications. The NaF-KF-AlF₃ system is a low-temperature electrolyte system that has attracted attention only in recent years. The NaF-KF-AlF₃ electrolyte with a low cryolite ratio (CR = (M_NaF + M_KF)/M_AlF3) has a lower initial crystallization temperature, which solves the problems caused by the accumulation of NaF during electrolysis of the KF-AlF₃ electrolyte system. This system is the most promising low-temperature aluminum electrolyte system for industrial applications [14,15,16]. However, at temperatures below 800 ℃ and high contents of CaF₂ and LiF in the electrolyte, the solubility of alumina in NaF-KF-AlF₃ at mole ratios in the range of 1.3–1.4 and NaF content of 30 mol% is low, and its dissolution rate is slow [16, 17].

In this paper, a small amount of Na₂O was added to Al₂O₃ to produce β-Al₂O₃ by calcination to promote the dissolution rate of alumina in the electrolyte. Na₂O·xAl₂O₃ (x is the MR of Al₂O₃ to Na₂O) samples with different molar ratio (MR) values were formulated, pressed, and calcined. Then, the solubility of the calcined pressed disc samples in NaF-KF-AlF₃ low-temperature electrolyte was measured. The Na₂O·xAl₂O₃ raw material with the lowest possible Na₂O addition amount and fast Al₂O₃ dissolution rate was obtained.

2 Experimental

2.1 Chemical and experimental setup

The experimental materials included potassium fluoride (KF), sodium fluoride (NaF), aluminum hydroxide (Al(OH)₃), polyvinyl alcohol ([-CH₂CHOH-]_n), sodium hydroxide (NaOH), and aluminum fluoride (AlF₃). All reagents were placed into a vacuum chamber and dried at 120 ℃ for 24 h and stored in an airtight container until use. A LECO oxygen analyzer (RO500, LECO Corporation, St. Joseph, MI, USA) was used to determine the oxygen content in Na₂O·xAl₂O₃ in the electrolyte. The alumina content of the sample was determined by the oxygen content.

1—stainless steel sampling rod; 2—corundum tube (argon inlet); 3—copper pipe (cooling water inlet); 4—O-ring; 5—corundum furnace tube; 6—heat insulation sheet group; 7—graphite crucible; 8—molten salt; 9-Na₂O·xAl₂O₃ tablets; 10—stainless steel positioning rod; 11—brass ring; 12—fastening bolts; 13—brass cover.

As shown in Fig. 1, the experimental setup and procedure for determining the alumina concentration in the electrolyte were similar to those used by Yan et al. [15, 16]. Dried NaF, KF, and AlF₃ were weighed according to the cryolite ratio (CR). The mixed materials were quickly transferred into a high-purity graphite crucible. The calcined Na₂O·xAl₂O₃ pressed disc was fixed on a tungsten wire.

Fig. 1

Schematic diagram of the device for measuring aluminum solubility

To reduce the volatilization of molten salt during the experiment, a graphite lid was attached to the graphite crucible. A Na₂O·xAl₂O₃ disc was hung above the melt before starting the furnace. The furnace was started, and then the air-cooled industrial chiller was turned on, and high-purity argon gas was introduced. After heating to the required temperature, the furnace was held for 1 h to ensure that all of the electrolyte was melted. Then, the Na₂O·xAl₂O₃ disc was placed into the melt, and a stainless steel rod was used to take samples at intervals through the hole in the graphite lid.

2.2 Experimental process

The experimental protocol was performed by using Na₂O and Al₂O₃ in MR of 1: 50, 1: 70, and 1: 90. The blank Al₂O₃ samples were formed from Al(OH)₃. The sodium dioxide content of the investigated ratios is shown in Table 1.

Table 1 Effect of Na₂O·xAl₂O₃ as raw material on electrolytic aluminum

Metallurgical-grade alumina was produced by the Bayer process. Due to the use of sodium hydroxide solution as the solvent, the Na₂O content in the alumina products is generally 0.2–0.4 wt%. During alumina production by the Bayer process, alumina with a higher Na₂O content can be obtained by controlling the washing degree of aluminum hydroxide. [18]

The proportioned reagent was mixed with 5 wt% of total reagent mass deionized water and 5 wt% polyvinyl alcohol. This mixture was heated until boiling and then placed into a vacuum-drying oven at 100 ℃ for 12 h. After that, it was pressed into flakes, which were loaded into a corundum crucible for calcination at 1000 ℃.

The electrolyte was mixed and put into the graphite crucible, two calcined Na₂O·xAl₂O₃ flakes were perforated, and the tungsten wire and fixed with clips above the interior of the graphite crucible. The high-temperature well resistance furnace was controlled to heat up from room temperature to different temperatures at a heating rate of 3 ℃·min⁻¹. When the temperature reached the value, the furnace was held for 1 h. After the temperature was constant, immersed the sample in the electrolyte to start the test, and samples were at 3 min, 5 min, 10 min, 60 min, 120 min, and 240 min. When collecting the solidified electrolyte, the crusted material on the stainless steel rod was scraped off. The cooled solid electrolyte sample was put into a sample bag for drying and storage. After drying, the sample was measured by a LECO oxygen meter to determine the oxygen content in the electrolyte, which was then converted to alumina concentration. The formulas for calculating the Al₂O₃ content of Na₂O·xAl₂O₃ in the melt are:

$$W_{{Al_{2} \,O_{3} }} = 2.125 \cdot {\text{O}}_{{Al_{2\,} \,O_{3} }}$$

(1)

$$o_{{Al_{2} \,O_{3} }} = {\text{O}}_{LECO} - {\text{O}}_{{Na_{2} o}}$$

(2)

$$o_{LECO} = a + x \cdot a \cdot b \cdot \frac{16}{{62}}$$

(3)

$$o_{{Na_{2} o}} = X \cdot b \cdot a \cdot \frac{16}{{62}}\quad \left( {b = \frac{16}{{X \cdot 164}}} \right)$$

(4)

where O_LECO is the result detected by the oxygen analyzer, x is the MR of Al₂O₃ to Na₂O, b refers to the percentage of Na₂O in Na₂O·70Al₂O₃, and a refers to the percentage of O in Al₂O₃.

The dissolution rate over a specific time period was calculated as follows. The dissolution rate of Na₂O·xAl₂O₃ in the electrolyte under each condition was calculated in segments by calculating the dissolution rate of alumina at each time period. For example, the dissolution rate of alumina in the time period T₂–T₁ is given by (5):

$$\overline{{\text{V}}} = \frac{{S_{2} - S_{1} }}{{T_{2} - T_{1} }}$$

(5)

where S₁ and S₂ are the alumina content in the electrolyte at T₁ and T₂, respectively.

3 Results and discussion

3.1 Study of the structure of Na₂O·xAl₂O₃ at different MR

The samples of Na₂O·xAl₂O₃ (x = 0, 50, 70, 90) were calcined at 1000 ℃ and then characterized by XRD. Powder diffraction data were obtained using a PANalytical X’pert PRO diffractometer with CuKα radiation. The detector was equipped with a graphite monochromator. All the XRD patterns were baseline adjusted and smoothed.

As shown in Fig. 2, phases in Na₂O·x(50, 70, 90) Al₂O₃ calcined at 1000 ℃ included β-Al₂O₃ and α-Al₂O₃, and Na₂O·x(50, 70) Al₂O₃ contained a small amount of γ-Al₂O₃. In Fig. 2, the intensity of the β-Al₂O₃ diffraction peak in Na₂O·50Al₂O₃ was the highest, while Na₂O·90Al₂O₃ was the smallest. As shown in supplementary Fig. S-1, because of the formation of β-Al₂O₃ when Na₂O in Na₂O·xAl₂O₃ solidified, the β-Al₂O₃ system formed a Na/O layered structure that the other phases of alumina did not have. The different structure led to the faster dissolution rate of β-Al₂O₃ in the electrolyte [19].

Fig. 2

Effect of MR on the structure of Na₂O·xAl₂O₃ after calcination at 1000 ℃

3.2 Effect of different MR on the solubility of Na₂O·xAl₂O₃ in NaF-KF-AlF₃

Na₂O·xAl₂O₃ was put into the electrolyte, and the concentration of alumina in NaF-KF-AlF₃ with a CR = 1.6 and a NaF content of 30 mol% was determined at 850 °C. The results are shown in Fig. 3.

Fig. 3

Effect of different MR on the dissolution properties of Na₂O·xAl₂O₃ in NaF-KF-AlF₃

In NaF-KF-AlF₃, the alumina concentration of Na₂O·xAl₂O₃ increased gradually upon increasing the dissolution time. From 0 to 10 min, the concentration of alumina changed the most, indicating that the dissolution rate of Na₂O·xAl₂O₃ in the electrolyte was the fastest during this period. From 0 to 60 min, it was found that the concentration of the blank Al₂O₃ sample in the electrolyte was lower than that of Na₂O·xAl₂O₃. The difference between the two concentrations was about 1.43 wt%, which was about 2.6 times. When dissolved for 240 min, the concentration of Na₂O·(50, 70, 90) Al₂O₃ and blank Al₂O₃ sample decreases gradually. The lower the MR, the higher the Na₂O content, and the faster the dissolution rate.

The dissolution rate of alumina in the electrolyte was calculated based on the change between each time sampled, as shown in Fig. 4.

Fig. 4

The dissolution rate of different MR of alumina in electrolyte

Upon increasing the reaction time, the dissolution rate of alumina in the electrolyte gradually decreased. At 0–5 min, the dissolution rate of Na₂O·50Al₂O₃ in the electrolyte was faster than the dissolution rate of alumina in samples with other MR values. In the first 5 min, the dissolution rate of the blank Al₂O₃ sample was the slowest. The dissolution rate of Na₂O·(50, 70, 90)Al₂O₃ in the electrolyte was 3.2–2.7 times higher than that of the blank Al₂O₃ sample.

3.3 Effect of CR on the solubility of Na₂O·70Al₂O₃ in electrolytes

In the NaF-KF-AlF₃ system with a CR = 1.6, the structure and the dissolution rate of Na₂O·70Al₂O₃ remained at intermediate stable values. For Na₂O·xAl₂O₃ and the first blank Al₂O₃ sample β-Al₂O₃, the peak intensities were 1934, 1452, 1271, and 0 cps respectively. The dissolution rate of aluminum oxide in the electrolyte was 0.72, 0.64, 0.61, and 0.22 g Al₂O₃/100 g melt from 0–5 min. However, due to the lack of the variables of temperature and CaF₂ and LiF additives, the solubility of Na₂O·xAl₂O₃ in this system cannot be further explored. The solubility of this specimen under these conditions was verified.

Experiments were performed to determine the solubility of the Na₂O·70Al₂O₃ sample’s temperature at 850 °C in NaF-KF-AlF₃ electrolytes with CR = 1.2, 1.3, 1.4, 1.5, and 1.6. The variables were the CR and sodium fluoride amounts of 30 and 40 mol%. The experimental results are shown in Fig. 5.

Fig. 5

Effect of CR on the dissolution properties of aluminum oxide. The numbers in the figure indicate that sample a was Na₂O·70Al₂O₃, and sample b was blank

The alumina solubility of Na₂O·70Al₂O₃ and Al₂O₃ in the electrolyte was greatly improved as the reaction time increased. Compared with the blank Al₂O₃ sample, the dissolution performance of Na₂O·70Al₂O₃ in the electrolyte was better than that of the blank Al₂O₃ sample from 0–10 min. The concentration difference between Na₂O·70Al₂O₃ and the blank Al₂O₃ sample gradually decreased and eventually stabilized from 60–240 min. As the CR of the electrolyte increased, so did the KF content in the electrolyte, while the AlF₃ content decreased. Increasing the KF content increased the saturation solubility point from 5.65 to 7.46 wt%, which reduced the mass-transfer limitations of alumina in the electrolyte and accelerated the dissolution rate of Na₂O·70Al₂O₃ in the molten salt. This has also been proven by other researchers [20,21,22].

For each time period, the experimental conditions were set as CR = 1.2–1.6 and NaF = 30 and 40 mol% (Fig. 6).

Fig. 6

The dissolution rate of alumina in each time period. The numbers in the figure indicate that sample a was Na₂O·70Al₂O₃, and sample b was blank

As the dissolution time progressed, the dissolution rate of alumina decreased gradually. Form 0–3 min (CR = 1.6, and NaF = 30 and 40 mol%), the dissolution rate of alumina was higher than that of other CR. Figure 6 shows that changing the CR also affected the dissolution rate of alumina in the electrolyte between Na₂O·70Al₂O₃ and the blank Al₂O₃ sample. The effect on the blank Al₂O₃ sample was stronger. At the same NaF content, increasing the CR increased the dissolution rate of Na₂O·70Al₂O₃ and the blank Al₂O₃ sample in the electrolyte by about 1.5 times.

The dissolution rate of alumina in the electrolyte changed obviously at 0–3 min for NaF = 30 mol% and 40 mol%. The dissolution rates of Na₂O·70Al₂O₃ and the blank Al₂O₃ sample increased by 0.25 and 0.16 (g Al₂O₃/100 g melt), respectively, when CR = 1.2–1.6. According to the above analysis, when the NaF content was constant and the CR increased, the solubility of the blank Al₂O₃ sample and the dissolution rate of alumina were lower than those of Na₂O·70Al₂O₃.

Compared with Na₂O·70Al₂O₃, the different dissolution performances and rates of the blank Al₂O₃ sample increased with the NaF content. When the CR was constant, the dissolution rate of alumina and its concentration in the electrolyte with 30 mol% NaF were higher than that of alumina in the electrolyte with 40 mol% NaF.

According to the experimental results, increasing the NaF content slowed the dissolution rate of alumina in the electrolyte. According to the dissolution rate of alumina, upon gradually increasing the NaF content and CR, the influence of CR had a greater effect than the NaF content.

3.4 Study on dissolution kinetics of Al₂O₃

Study on the influence of dissolution rate of alumina with time. T = 850 ℃, CR = 1.2 and 1.6, and NaF = 30 mol%. The results are shown in Fig. 7.

Fig. 7

Relationship between solubility of Na₂O·70Al₂O₃ and dissolution time in molten salt

Figure 7 shows that the concentration of aluminum oxide in the molten salt increased with time. In the first 5 min, upon increasing the experimental time, the solubility of aluminum oxide in the electrolyte showed a linear increase. However, after 5 min, this increase gradually slowed.

It can be seen in ref. [19] that the dissolution rate of aluminum oxide in the electrolyte increased linearly with the reaction time. The part in which the dissolution rate of aluminum oxide in the electrolyte increased in a curve with the reaction time and followed the zero-order law. The part of the curve in which the dissolution rate of aluminum oxide in the electrolyte increased followed the first-order law. The reaction rate constants of these two laws reflect the dissolution rate of alumina in the electrolyte.

A zero-order reaction refers to a chemical reaction whose reaction rate is proportional to the zeroth power of reactant concentration. The reaction rate equation is expressed as:

$$C_{a} - C_{0} = - kt_{a}$$

(6)

where t_a is the time, C_a is the dissolution rate of alumina at the time t_a, C₀ is the initial dissolution rate, and k is the reaction rate constant.

It can be seen from the formula of the zero-order reaction rate that the concentration of Na₂O·70Al₂O₃ in the electrolyte increased with the reaction in the first 5 min of the experiment. When the reaction continued, the concentration of reactants in the molten salt decreased gradually. Based on these conditions, the alumina dissolution rate at 10 min can be used as the initial dissolution rate. The sampling time of the linear portion was 0, 1, 3, and 5 min respectively, as shown in Supplementary Table S-I.

The alumina dissolution rate at 5 min was used as the initial dissolution rate. To facilitate the formula calculation, the time was arranged in reverse order. That is, the fifth minute of the original experiment is time zero, the third minute of the original experiment is the first minute, the first minute of the original experiment is the third minute, and the zero time of the original experiment is the fifth minute. The results of data conversion are shown in Supplementary Table S-II.

To linearly fit the relationship between alumina dissolution rate and time with CR = 1.2 and 1.6, Supplementary Table S-II and Table S-II were combined, and the results are shown in Supplementary Fig. S-2.

After fitting, the zero-order reaction rate expression of CR = 1.2 and 1.6 can be obtained:

$$C_{a} - C_{0} = - 0.13t_{a} - 0.0121\quad R^{{2}} = \, 0.{9865}$$

(7)

$$C_{a} - C_{0} = - 0.47t_{a} - 0.0561 \quad R^{{2}} = \, 0.{9781}$$

(8)

The zero-order reaction rate equation had a good fit when CR = 1.2 and 1.6. Through calculation, when the CR = 1.2 and 1.6, the k values in the zero-order reaction rate equation were 0.13 and 0.47 ((g Al₂O₃/100 g melt)·min⁻¹·L⁻¹), respectively.

The first-order reaction rate equation is:

$$1n\left( {C_{a} } \right) - 1n\left( {C_{0} } \right) = - kt_{a}$$

(9)

It can be seen from the first-order reaction rate formula that the concentration of Al₂O₃ in the electrolyte decreased gradually as the reaction proceeded after 5 min of the experiment. Based on this, the alumina dissolution rate after 120 min was used as the initial dissolution rate. The sampling time of the linear portion was 30, 60, 90, and 120 min respectively, as shown in Supplementary Table S-III.

Similarly, at 120 min, the dissolution rate of alumina was taken as the initial dissolution rate. The logarithm of the data in supplementary Table S-III was taken to obtain supplementary Table S-IV.

To linearly fit the relationship between alumina dissolution rate and time with CR = 1.2 and 1.6, Supplementary Table S-III and Table S-IV were combined, and the results are shown in Supplementary Fig. S-3.

The first-order reaction rate expression with CR = 1.2 and 1.6 was obtained after fitting.

$${\text{ln}}\left( {C_{a} /C_{0} } \right) = \, - 0.0108t_{a} + 0.0332\quad R^{2} = 0.{9665}$$

(10)

$${\text{ln}}\left( {C_{a} /C_{0} } \right) = \, - 0.0122t_{a} + 0.0371\quad R^{2} = 0.{9977}$$

(11)

The fitting degree of the first-order reaction rate equations when CR = 1.2 and 1.6 was better. The k values in the first-order reaction rate equation were 0.0108 and 0.0122 ((g Al₂O₃ /100 g melt)·min⁻¹·L⁻¹), respectively.

The zero-order reaction rate constant of CR = 1.2 and 1.6 was larger, which also indicates that the dissolution rate of aluminum oxide in the electrolyte with CR = 1.2 and 1.6 was faster in the first 5 min after the experiment. However, from their first-order reaction rate constants upon increasing the time, the difference between the first-order reaction rate constants for CR = 1.2 and 1.6 gradually decreased, and the dissolution rate of aluminum oxide in the electrolyte gradually became the same.

The dissolution of aluminum oxide in the electrolyte can be explained by oxyfluoride, which was produced by the exchange of oxygen ions and fluoride ions. In the experiment, the KF content of the electrolyte was more than 10 wt%. In the NaF-KF-AlF₃ system, the content of AlF₅²⁻ was high, which promoted the formation of Al₂OF₆²⁻ [14]. At the beginning of the experiment, the content of aluminum oxide in the electrolyte was low. Because the concentration of aluminum oxide in the electrolyte increased rapidly, aluminum ions and oxygen ions complexed and combined into the melt structure. Because the oxygen content in the melt was low, the electrolyte easily absorbed aluminum oxide, and the dissolution of aluminum oxide followed the zero-order law.

Upon gradually increasing the alumina concentration, the AlF₅²⁻ content decreased, and AlF₆³⁻ became the dominant species. The coordination number increased from 5 to 6, that is, the Al-O-F complex ion changed from Al₂OF₆²⁻ to Al₂O₂F₄²⁻, making it difficult to absorb oxygen and reducing the dissolution rate of alumina in the electrolyte [23]. Therefore, the dissolution of alumina in electrolyte followed the first-order reaction law.

To sum up, the dissolution rate of alumina in electrolyte depended on the rate constant of the zero-order reaction, that is, the sum of the dissolved amount of alumina in the first 5 min.

3.5 Effect of temperature on the solubility of Na₂O·70Al₂O₃ in electrolytes

The effect of temperature on the solubility of alumina was also investigated in the NaF-KF-AlF₃ system at CR is 1.2, 1.3, and 1.4, and NaF = 30 mol%. The results are shown in Fig. 8.

Fig. 8

Effect of different temperatures on the dissolution properties of Na₂O·70Al₂O_3. The numbers in the figure indicate that sample a is Na₂O·70Al₂O₃, and sample b was blank

The concentration of alumina in the electrolyte increased significantly when the temperature increased from 750 to 850 °C. After 120 min, the concentration of alumina in the electrolyte increased slowly and gradually became saturated.

Two groups of samples at different temperatures were compared with the blank Al₂O₃ sample. Upon increasing the CR, the concentration difference between Na₂O·70Al₂O₃ and the blank Al₂O₃ sample gradually increased, and the concentration difference between Na₂O·70Al₂O₃ and the blank Al₂O₃ sample was the greatest when the temperature was 850 ℃ and the reaction time was 3 min. When the reaction tended towards saturation (240 min), there was almost no concentration difference. The experimental results show that the lower temperature, the lower the oxygen content of Na₂O·70Al₂O₃ in NaF-KF-AlF₃, and decreasing the oxygen content decreased the alumina concentration.

The dissolution rate of Na₂O·70Al₂O₃ in the NaF-KF-AlF₃ system was investigated under the same CR and sodium fluoride amount with the temperature as the variable. The experimental conditions were: CR = 1.2–1.4, NaF content of 30 mol%, and temperature of 750 °C and 850 °C. The experimental results are shown in Fig. 9.

Fig. 9

The dissolution rate of alumina at different temperatures

At the same temperature, the dissolution rate of alumina in the two groups increased upon increasing the CR between 0 and 3 min. At 750 ℃, the dissolution rate of Na₂O·70Al₂O₃ in NaF-KF-AlF₃ increased by 0.85 (g Al₂O₃/100 g melt) with a CR of 0.1. The dissolution rate of Na₂O·70Al₂O₃ in the blank Al₂O₃ sample increased by 2.83–1.83 times compared with Na₂O·70Al₂O₃. At 850 ℃, the dissolution rate of Na₂O·70Al₂O₃ in the electrolyte increased by 1.11 (g Al₂O₃/100 g melt). Compared with Na₂O·70Al₂O₃, the dissolution rate of the blank Al₂O₃ sample increased by 2.2–2.5 times. In general, a higher temperature had a greater influence on the dissolution rate of the sample. Upon increasing the CR, the difference between the dissolution rate of alumina increases gradually at both temperatures. From 0–10 min, the dissolution rates of Na₂O·70Al₂O₃ and Al₂O₃ in the electrolyte at 850 ℃ were 1.3–1.4 times higher than those at 750 ℃.

3.6 Effect of CaF₂ and LiF on the solubility of Na₂O·70Al₂O₃ in electrolytes

Since it has been determined in the exploratory experiments that the experimental time had a clear trend at 60 min, the time was set to 60 min in subsequent experiments to study the concentration of Na₂O·70Al₂O₃ in NaF-KF-AlF₃-CaF₂-LiF electrolyte. This was done to reduce the errors caused by gradual saturation over time. The effect of the CR and CaF₂ and LiF contents on the solubility of alumina in the electrolyte melt was investigated in the NaF-KF-AlF₃-CaF₂-LiF electrolyte system under the same experimental conditions (CR is 1.3 and 1.4, NaF addition of 30%, and temperature of 750 °C). The obtained data are shown in Fig. 10.

Fig. 10

Effect of CaF₂ and LiF on the solubility of Na₂O·70Al₂O₃ in NaF-KF-AlF₃ electrolyte. The numbers in the figure indicate that sample a is Na₂O·70Al₂O₃, and sample b was blank

As shown in Fig. 10, the concentration of alumina was negatively correlated with the CaF₂ and LiF content. As the calcium fluoride content increased, the CR decreased, resulting in a decrease in the concentration of alumina in the melt. When the LiF content increased from 0 to 2.5 wt%, for every 1 wt% increase in LiF content, the concentration of alumina decreased by about 0.12 wt%, respectively. It is evident that when the CR was decreased, the effect of LiF and CaF₂ on the alumina dissolution rate also decreased.

In an alumina solubility experiment, Yan et al. [16] reported that for every 1 wt% LiF that was added to a K₃AlF₆ melt with an AlF₃ content of 33 wt% at 700 ℃, the solubility of alumina decreased by 0.37 wt%. In recent years, many scholars have studied the solubility of lithium salts in alumina in traditional aluminum electrolysis systems. [24, 25] The results have shown that in electrolyte melts, Li⁺ has a small ionic radius and is more attractive to anions such as AlF₄⁻ and AlF₅²⁻. Moreover, upon increasing the LiF content, the content of free fluoride ion F⁻ in molten salt increased slightly. Therefore, when Li⁺ is added to the melt, the Al₂O₃ reaction is inhibited, and the solubility of molten Al₂O₃ in the electrolyte is reduced. When CaF₂ is added to the melt, the contents of AlF₄⁻ and AlF₅²⁻ will also be reduced, and the solubility of alumina in the bath will be eventually reduced. In an electrolyte with the same amount of CaF₂ and LiF, upon increasing the CR, the concentration of alumina in the Na₂O·70Al₂O₃ and blank Al₂O₃ sample electrolyte also increased, then the concentration of Na₂O·70Al₂O₃ in the electrolyte underwent a smaller change than that of the blank Al₂O₃ sample. When the amount of additives was the same, the greater the CR, the better the concentration of Na₂O·70Al₂O₃ and the blank Al₂O₃ sample in the electrolyte. Under the same CR, the influence of CaF₂ on the concentration of alumina in the electrolyte is less than that of LiF at the same weight percentage.

The dissolution rate of alumina in NaF-KF-AlF₃-CaF₂ (5 wt%, 4 wt%, 2.5 wt%)-LiF (0 wt%, 1 wt%, 2.5 wt%) electrolyte systems was investigated using calcium fluoride and lithium fluoride as variables. The experimental conditions were: CR = 1.3–1.4, NaF content of 30 mol%, and temperature of 750 °C. The experimental results are shown in Fig. 11.

Fig. 11

The dissolution rate of alumina in electrolytes containing CaF₂ and LiF

While the dissolution rate of alumina in the electrolyte gradually decreased. From 0–3 min, the dissolution rate of alumina in the electrolyte with a CR = 1.4 was faster than that of alumina with a CR = 1.3. When CR = 1.4, the dissolution rate of alumina in the electrolyte with Na₂O·70Al₂O₃ was 1.3 times higher than that of the Al₂O₃ sample. When CR = 1.3, the dissolution rate of Na₂O·70Al₂O₃ in the electrolyte was about twice that of the Al₂O₃ sample. At the same CR, there was a significant difference in the dissolution rate of alumina in NaF-KF-AlF₃-CaF₂, and the difference between the two samples was about 0.1 and 0.05 (g Al₂O₃/100 g melt min), respectively. At 3–5 min, the dissolution rate of alumina in the electrolyte followed the above rule. However, the dissolution rate of alumina in different electrolytes in 5–10 min changed irregularly, and there was little difference. At 10–60 min, the dissolution rate of alumina with different electrolytes was small, so no specific analysis was performed.

Changes in the alumina concentration with the CaF₂ and LiF contents in the NaF-KF-AlF₃ electrolyte are studied for Na₂O·70Al₂O₃ when the CR is 1.3 and 1.4, the NaF content is 30 mol%, and the temperature is 750 °C. The experimental results are shown in Fig. 12.

Fig. 12

Effect of NaF-KF-AlF₃ and NaF-KF-AlF₃-CaF₂-LiF on the dissolution properties of Na₂O·70Al₂O₃.

Adding CaF₂ alone or the addition of CaF₂ and LiF mixture to the NaF-KF-AlF₃ electrolyte system decreased the solubility of Na₂O·70Al₂O₃ in alumina. The concentration of alumina in the NaF-KF-AlF₃ electrolyte system with Na₂O·70Al₂O₃ increased gradually upon increasing the CR of the NaF-KF-AlF₃-CaF₂ (2.5 wt%, 4 wt%, 5 wt%)-LiF (2.5 wt%, 1 wt%, 0 wt%) electrolyte. In comparison, the effect of the alumina concentration in the electrolyte with a CR = 1.3 was much greater than that at a CR = 1.4. The concentration and the dissolution rate of alumina in the conventional ternary electrolyte system were higher than those in the electrolytes containing CaF₂ and LiF.

4 Conclusions

Na₂O·xAl₂O₃ samples were formed by adding Na₂O to Al₂O₃. After being calcined at 1000 ℃, the structure changed, and β-Al₂O₃ was produced. However, β-Al₂O₃ promoted the dissolution rate of the sample in the electrolyte, so the dissolution rate of alumina was improved by increasing the Na₂O content in Al₂O₃.

In a NaF-KF-AlF₃-based low-temperature electrolyte system, Na₂O·xAl₂O₃ showed a faster alumina dissolution rate than Al₂O₃. The higher the Na₂O content in alumina, the faster the dissolution rate. When the NaF content in NaF-KF-AlF₃ was in the range of 30–40 mol%, the alumina dissolution rate of Na₂O·xAl₂O₃ in the molten salt was 1–2 times higher than that of Al₂O₃. When the temperature was in the range of 750–850 ℃, the alumina dissolution rate of Na₂O·xAl₂O₃ in the molten salt was 1.5–3 times higher than that of Al₂O₃. The dissolution rate of Na₂O·70Al₂O₃ and the blank Al₂O₃ sample in the electrolyte increased when the temperature and CR of the molten salt increased. However, the dissolution rate of alumina decreased upon increasing the NaF content.

The reaction rate constant of the zero-order reaction was larger than those of the first-order reaction rate during alumina dissolution. The dissolution rate of alumina in each time period and the rate constants of the two rate laws showed that the dissolution rate of alumina depended on the amount of alumina absorbed by the electrolyte in the first five minutes of the experiment.

At a constant CR and temperature, adding CaF₂ and LiF to the ternary electrolyte reduced the difference between the dissolution rate between the blank Al₂O₃ sample and Na₂O·70Al₂O₃ in the molten salt by 1.7–3.2 times. The influence of CaF₂ on the dissolution rate of alumina in the electrolyte was weaker than that of LiF.