1. Introduction

2. Effect of Zr and B ions on K3AlF6-Na3AlF6-AlF3 molten salt system

2.1. Computational Methods

This calculation uses the Material Studio software to set a given number of various ions in a periodic box. The Foeite module in MS (UFF force field with a time step of 1 fs) was used to simulate the dynamic processes of the molten salt. The Coulomb effect of the simulated ions was processed using the Ewald summation algorithm for the system, with the buffer width and energy calculation accuracy set to 0.5 Å and 0.5 kcal/mol. The cut-off radius for short-range interactions was set to half the length of the simulated box sides. The simulated molten salt tank was heated to 4000 K at a pressure of 1.01 MPa under an isothermal isobaric system and a structural relaxation of 100 ps was performed. The particle number, pressure and temperature of the molten salt system were set to remain constant during the simulation. The superheated liquid was then cooled to the target molten salt melting point of 1173 K at 1 K/ps in the canonical system and the molten salt model continued to undergo 100 ps of structural relaxation.

Table 1. he number of atoms in the simulation system and the size of the system.

Table 1. he number of atoms in the simulation system and the size of the system.

Additive	Number of particles						Box side length/ Å
	Al	Na	K	F	Zr	B
No addition	610	780	270	2890	0	0	43.6
0.1wt%Zr4+	610	780	270	2890	1	0	43.6
0.2wt%Zr4+	610	780	270	2890	2	0	43.6
0.3wt%Zr4+	610	780	270	2890	3	0	43.6
0.4wt%Zr4+	610	780	270	2890	4	0	43.6
0.5wt%Zr4+	610	780	270	2890	5	0	43.6
0.1wt%B2-	610	780	270	2890	0	2	43.6
0.5wt%B2-	610	780	270	2890	0	4	43.6
0.3wt%B2-	610	780	270	2890	0	6	43.6
0.4wt%B2-	610	780	270	2890	0	8	43.6
0.5wt%B2-	610	780	270	2890	0	10	43.6
0.1wt%ZrB2	610	780	270	2890	1	2	43.6
0.2wt%ZrB2	610	780	270	2890	2	4	43.6
0.3wt%ZrB2	610	780	270	2890	6	6	43.6
0.4wt%ZrB2	610	780	270	2890	4	8	43.6
0.5wt%ZrB2	610	780	270	2890	5	10	43.6

3. Effect of Zr, B, C, W, Ti, Hf, Nb, and Ta ions on K₃AlF₆-Na₃AlF₆-AlF₃ molten salt system

3.1. Computational Methods

Since the electrolysis system did not change, the parameters of the box used were set as follows: Density 2 g/cm³. Then, selected the Ensemble of Dynamics as NPT to keep the box size constant for parameter setting, Temperature as 4000K, Pressure as 0.00101GPa, and Total simulation time as 100ps. And then selected the Ensemble of Dynamics as NVT to keep the box size constant for parameter setting, Temperature as 1173K.

Table 2. The number of atoms in the simulation system and the size of the system.

Table 2. The number of atoms in the simulation system and the size of the system.

AdditiveZrB2	Number of particles												Box side length/ Å
	Al	Na	K	F	Zr	B	C	W	Ti	Hf	Nb	Ta
0.1wt%	610	780	270	2890	1	2	2		2	1	2	1	43.7
0.2wt%	610	780	270	2890	2	4	2		2	1	2	1	43.7
0.3wt%	610	780	270	2890	3	6	2		2	1	2	1	43.7
0.4wt%	610	780	270	2890	4	8	2		2	1	2	1	43.8
0.5wt%	610	780	270	2890	5	10	2		2	1	2	1	43.8
0.1wt%	610	780	270	2890	1	2	2						43.6
0.2wt%	610	780	270	2890	2	4	2						43.6
0.3wt%	610	780	270	2890	3	6	2						43.7
0.4wt%	610	780	270	2890	4	8	2						43.7
0.5wt%	610	780	270	2890	5	10	2						43.7
0.1wt%	610	780	270	2890	1	2		1					43.6
0.2wt%	610	780	270	2890	2	4		1					43.7
0.3wt%	610	780	270	2890	3	6		1					43.7
0.4wt%	610	780	270	2890	4	8		1					43.7
0.5wt%	610	780	270	2890	5	10		1					43.7

3.2. Radial distribution function

The radial distribution is mainly used to observe the bonding between two molecules or atoms. First, select two kinds of atoms or molecules, set them into two kinds of Sets, and then analyze the radial distribution of the two Sets. For this analysis, this paper select two kinds of atoms (Al and F). The formula of the radial distribution function is as follows [16].

$$\mathrm{g}\left(\mathrm{r}\right)=\frac{1}{\mathsf{\rho }\mathrm{N}}\left({\displaystyle \sum }_{\mathrm{ij}}\mathsf{\delta }\left|\mathrm{r}-{\mathrm{r}}_{\mathrm{ij}}\right|\right)$$

(4)

Table 3. Average bond lengths of Al-F ion pairs in molten salt systems using different additives.

Table 3. Average bond lengths of Al-F ion pairs in molten salt systems using different additives.

Other elements	ZrB2 Additive
	0.1wt%	0.2wt%	0.3wt%	0.4wt%	0.5wt%
No addition	3.15	3.13	3.13	3.11	3.11
W	3.03	3.03	3.03	3.01	3.01
C	3.15	3.13	3.13	3.11	3.11
Ti, Hf, Nb, Ta, C	3.19	3.19	3.17	3.17	3.15

Figure 4 shows the radial distribution functions among the Al-F ions when adding ZrB₂ at the proportions of 0 wt%, 0.1 wt%, 0.2 wt%, 0.3 wt%, 0.4 wt%, 0.5 wt%, and different trace elements to the K₃AlF₆-Na₃AlF₆-AlF₃ molten salt electrolyte system. In (a), the position of the first wave of Al-F varies with the ZrB₂ content from 0~0.5% with 3.17 A, 3.15 A, 3.13 A, 3.13 A, 3.11 A, and 3.11 A. In (b), the average bond length of Al-F decreases with increasing ZrB₂ content. In the Al-F comparison plot (a), (b), (c), and (d). it can be found that plot (a) and (c) are not significantly different from each other, indicating that the addition of C ions has no significant effect on the F and Al ions of this molten salt system. With the addition of different additives in the (b) and (d) plots, it can be obviously found that the sharpness of the first front of Al-F has changed, which indicates that the W ions and Ti, Hf, Nb, and Ta ions in the molten salt species make the distribution of F ions around Al ions more uniform. Taking the (d) as an example, it can be found that the sharpness of the first front of Al-F becomes significantly slower with the increase of ZrB₂ content, indicating that the distribution of F ions around Al ions is more uniform under the joint effect of Zr ions and B ions [17].

The radius of the first peak of the radial distribution function indicates the average bond length of the ion pair [18]. Table 3 shows the average bond length of Al-F ion pairs in the molten salt system with different additives, which also corresponds to the four figures of abcd, respectively. It can be noticed that the average bond length of the molten salt system with the addition of W decreases, indicating that the addition of W results in a weaker interaction between Al-F, while the average bond length of the molten salt system with the addition of Ti, Hf, Nb, Ta, and C increases, indicating that the addition of Ti, Hf, Nb, Ta, and C leads to an enhanced Al-F interaction. The interaction between Al and F represents the coordination ability between Al and F. It is therefore possible to predict the trend of the effect of the addition of W, Ti, Hf, Nb, Ta, and C elements on the coordination number of Al and F in the electrolyte system by virtue of these variation patterns.

3.3. Coordination number

In Table 4, the coordination number between Al-F decreases after adding ZrB₂ to the system, indicating that the interaction between Al-F is gradually decreasing with the addition of ZrB₂ and the number of fluorine bridges formed in the system is also gradually decreasing. In the addition of 0.1% ZrB₂ into the melt system, the coordination number F between Al was significantly reduced, which also indicated a shortening of the interaction F between Al [19].

Table 4. Coordination number between Al-F in molten salt systems using different additives.

Table 4. Coordination number between Al-F in molten salt systems using different additives.

Other elements	ZrB2 Additive
	0.1wt%	0.2wt%	0.3wt%	0.4wt%	0.5wt%
No addition	5.52	5.20	5.12	4.84	4.83
W	5.45	5.07	5.06	4.93	4.77
C	5.49	5.20	5.16	4.80	4.78
Ti, Hf, Nb, Ta, C	5.98	5.86	5.75	5.73	5.50

In Figure 5, it can be concluded that the addition of C and W ions decreases the coordination number of F- around Al³⁺ in the K₃AlF₆-Na₃AlF₆-AlF₃ molten salt electrolyte system, and the effect of W ions in decreasing the coordination number of F- around Al3+ is more obvious, presumably because of C and W ions. The hypothetical reason is that the C and W ions intensify the transition process of [AlF₆]^3-→[AlF₄]^-, where the effect of W is more effective. With the addition of Ti, Hf, Nb, and Ta ions, the coordination number of F- around Al³⁺ in this molten salt system increases significantly, probably because Ti, Hf, Nb, and Ta ions inhibit the transition process of [AlF₆]^3-→[AlF₄]^-.

4. Conclusion

References

1. Wei Ding. Experimental analysis of the combustion of aluminum foil in air[J]. Mathematical and physical study, 2011(07): 106-107.
2. Yi Wang, Shengxiang Deng. Numerical simulation of magnetic field of 400kA aluminum electrolyzer based on COMSOL[J]. Light Metals, 2022(11): 30-34. [CrossRef]
3. Shengshou Wang. A review of physicochemical properties of aluminum electrolytes[C]// National Aluminum Electrolysis Energy Conservation Technology Exchange. 2006.
4. Yongzhi Wang. Study on the influence of additives on the structure and properties of aluminum electrolytic molten salt system[D]. Kunming University of Science and Technology, 2015.
5. Ning Zhang, Wei Chen. Exploration of energy-saving and emission reduction technology for electrolytic aluminum production[J].China Metal Bulletin, 2017(11): 100-101.
6. Yuping Zhu, Yedong He, DeRen Wang. Oxidation and electrolytic corrosion behavior of Fe-40Cr-2Ce alloy anodes at 900°C[J]. Journal of Materials Heat Treatment, 2011, 32(11): 6. [CrossRef]
7. Xiaojun Lv, Yajing Shuang, Lingyun Hu. Theoretical calculation of the initial crystal temperature and alumina solubility of aluminum electrolytes[J]. Light Metals, 2015, No.443(09): 27-31. [CrossRef]
8. Shuo Han, Honglei Xu, Peng Zhen. Molecular dynamics simulation of Li-rich aluminum electrolyte system[J]. Light Metals, 2021, No.515(09): 29-33. [CrossRef]
9. Guocai Tian, Yongzhi Wang. Molecular dynamics simulation study of the effect of AlF3 on the structure and properties of cryolite molten salt system[J]. Journal of Kunming University of Science and Technology (Natural Science Edition), 2015, 40(03): 1-8. [CrossRef]
10. Yayun Zhang, Mian Chen, Ya Deng. Molecular Dynamics Simulation of Temperature and Pressure Effects on Hydration Characteristics of Montmorillonites[J]. Journal of the Chinese Ceramic Society, 2018(10): 22. [CrossRef]
11. Cikit S, Akdeniz Z, Madden P A. Structure and Raman Spectra in Cryolitic Melts: Simulations with an ab Initio Interaction Potential[J]. The Journal of Physical Chemistry B, 118 (2014), 1064-1070. [CrossRef]
12. Nazmutdinov Renat R, Zinkicheva Tamara T. A spectroscopic and computational study of Al(III) complexes in cryolite melts: Effect of cation nature[J]. Chemical Physics, (412), 22–29. [CrossRef]
13. Nan Ma, Jinglin You, Guangchao Wei. High-temperature Raman spectroscopy and quantum chemical calculations of molten salt microstructures of K₃AlF₆-Al₂O₃ system[C]//. Abstracts of the 19th National Conference on Light Scattering. 2017: 144.
14. R.A. McKee. A generalization of the nernst-Einstein equation for self-diffusion in high defect concentration solids[J]. Solid State Ionics, 5(none), 133–136. [CrossRef]
15. Searles, Debra J, Evans Denis J. The fluctuation theorem and Green–Kubo relations[J]. The Journal of Chemical Physics, 112(22), 9727. [CrossRef]
16. Xiaojun Lv, Zhenming Xu. Theoretical investigation on local structure and transport properties of NaF-AlF3 molten salts under electric field environment[J]. Journal of Molecular Structure, 1117, 105–112. [CrossRef]
17. Xianwei Hu, Zhaowen Wang. Raman spectra of the ionic structure of acidic NaF-AlF₃ molten salts[J]. Chinese Journal of Nonferrous Metals, 2008(10): 1914-1919. [CrossRef]
18. Zongwei Li, Jue Kou, Tichang Sun, Zekun Chen. Molecular dynamics simulation of CaO dephosphorylation mechanism during direct reduction of high phosphorus hematite[J]. Nonferrous Metals(Mineral Processing Part), 2022(06): 35-42.DOI:10.3969/j.issn.1671-9492.2022.06.055.
19. Yanli Jiang, Xin Tang, Fan Gao. Molecular dynamics study of cryolite-alumina molten salts[J]. Journal of Atomic and Molecular Physics, 2016, 33(02): 325-329. [CrossRef]