Abstract
The electrochemical behaviour of Si(IV) ions on Fe electrode in the NaCl-NaF-KCl-SiO2 molten salt system (0.177 mol·L−1 SiO2) at 1,103 K was studied by using cyclic voltammetry, square wave voltammetry, chronoamperometry, and chronopotentiometry．Experiments conducted using the first two methods indicated that the reduction of Si(IV) occurred in two steps: Si(IV) → Si(II) → Si(0). Furthermore, the electrochemical reaction is a quasi-reversible process, which is controlled by both the ion diffusion and electron transport rates. The electrochemical crystallization of Si was found to be a transient, the three-dimensional nucleation process. The cyclic voltammetry curves indicate that the diffusion coefficient of Si(IV) is 1.16×10–5 cm2·s−1. The phases formed on the surface of the deposit were analysed by scanning electron microscopy and X-ray diffractometer. The results show that Fe and Si have formed intermetallic compounds Fe3Si, FeSi, and Fe5Si3.
Introduction
Silicon steel is an important soft magnetic alloy used in the electric power and electronic industries and military applications, mostly for various kinds of motors, generators, and transformers. Research has shown that increasing the content of silicon could improve the properties of silicon steel. Especially, excellent performance could be achieved at 6.5 wt% Si: the magnetostrictive rate is almost zero, and the maximum permeability and resistivity both reach the highest values [1]. Typically, this material is prepared by silicon deposition methods, such as electron beam physical vapour deposition [2], laser cladding [3], hot dip [4], and molten salt electrodeposition [5]. Since the 1990s, electrodeposition in molten salt has attracted considerable attention [6, 7, 8]. Many researchers reported successful electrodeposition of Si(IV) from fluoride melts such as NaF-KF-Na2SiF6 [9] and LiF-KF-K2SiF6 [10], and the chloride–fluoride melts of NaCl-KCl-NaF-K2SiF6 [11]. In addition, the electrodeposition of Si from molten BaF2-CaF2-SiO2 at 1,200°C has also been demonstrated [12]. But the mechanism of Si(IV) in the molten salt system of NaCl-NaF-KCl-SiO2 has not been reported.
The electrochemical reduction and nucleation mechanism of Si(IV) in molten salt NaCl-NaF-KCl-SiO2 were studied by electrochemical method. The electrochemical behaviour of Si(IV) on a Fe electrode in the molten salt system of NaCl-NaF-KCl-SiO2 (cSiO2=0.177 mol·L−1) with various electrochemical methods [13, 14, 15] were investigated. The phases formed on the surface of the deposit were analysed by scanning electron microscopy and X-ray diffractometry. It provides a theoretical basis for the preparation of functional materials for silicon steel.
Experimental process
KCl, NaCl, NaF, and SiO2 (analytical grade) with the mass ratio of NaCl : KCl : NaF=1 : 1: 2 were mixed thoroughly, and the content of SiO2 was 2 wt%. The mixed chemicals were dried in a DZF-6020 (Bo Xun Industrial, Shanghai) vacuum drying furnace for 8 h at 473 K, and then were cooled to room temperature.
The experiments with molten salts were performed in an argon atmosphere with a 3KL10·BYL tubular resistance furnace (Yunjie Electric Furnace Factory, Baotou) and a P4000 electrochemical workstation (Princeton, USA). All electrode surfaces were polished mechanically to a mirror finish before measurements. In the three-electrode system used for the electrochemical tests, an iron wire (1 mm diameter) and two platinum wires (0.5 mm in diameter) were used as the working, reference, and auxiliary electrodes, respectively. Figure 1 shows the equipment of electrochemical process. A zirconia crucible filled with the mixed salt was placed in the furnace, heated to 1,103 K and maintained at this temperature for about 3.5 h. Then, the three electrodes were inserted into the molten salt for electrochemical measurements. The electrodeposition was carried out at a constant potential of −1.7 V for 1 h . After deposition, the samples were analysed by an S-4800 scanning electron microscope (Hitachi, Japan) and a Noran 7 X-ray diffractometer (Thermo Fisher, Waltham, MA, USA).
Results and discussion
Cyclic voltammetry
Figure 2 shows the cyclic voltammograms in the molten salt at 1,103 K with and without SiO2. In the absence of SiO2 (Curve a), there are no reduction or oxidation peaks in the voltage range between −2.5 and −1 V. Hence, no redox reactions occur in the base molten salt of NaCl-NaF-KCl. In the presence of 0.177 mol·L−1 SiO2 under the same conditions, there are two redox peaks A’/A and B’/B; Curve b contains two obvious oxidation peaks A´ and B´ at −0.2 and −1.72 V vs. Pt, reduction peaks A and B at −0.8 and −1.7 V vs. Pt, respectively, which indicate that Si(IV) is reduced on the Fe electrode in two steps. This is consistent with the findings of G et al. [16].
The equipment of electrochemical process.
Cyclic voltammograms of NaCl-NaF-KCl melts (a) before and (b) after the addition of 0.177 mol·L−1 SiO2 at 1,103 K on Fe electrode. Scan rate: 0.9 V·s−1.
Square wave voltammetry
In the square wave voltammetry curve of NaCl-NaF-KCl-SiO2 on the Fe electrode (Figure 3), there are also two reduction peaks A and B at −0.8 and −1.7 V vs. Pt, respectively; in accordance with the two steps of Si(IV) reduction on the electrode. According to the literature [17], the half-width of a peak (W1/2) at low frequencies is linked to the number of exchanged electrons (n):
Square wave voltammogram on a Fe electrode in NaCl-NaF-KCl-SiO2 (cSiO2=0.177 mol·L−1) melt at 1,103 K. Pulse height: 180 mV, potential step: 72 mV, and frequency: 20 Hz.
where T is the experimental temperature (1,103 K), R is the gas constant, and F is Faraday’s constant. The W1/2 value found for peak A is about 0.18 V. Using eq. (1), the calculated n is 1.9 for peak A. Since this value is close to 2, the corresponding reaction of Si(IV) on the Fe electrode can be expressed as follows:
The value of W1/2 for peak B is about 0.21 V, leading to n=1.6 (which is also close to 2). Therefore, the corresponding electrochemical reaction on the Fe electrode is:
In summary, the above analysis shows that the reduction of Si(IV) on the Fe electrode occurs in two steps (i. e. Si(IV) → Si(II) → Si(0)), which is consistent with the conclusion from cyclic voltammetry.
Chronopotentiometry
The electrochemical reduction of Si(IV) on Fe electrode in the molten salt was further examined with chronopotentiometry. Figure 4 displays the constant-current transient curves. Again, there are two obvious steps between −0.3 and −0.9 A. When the current exceeds −0.5 A, there appears a step A which is consistent with the reduction peak A at −0.8 V in Figures 2 and 3. In this step, Si(IV) is reduced to Si(II) on the Fe electrode. When the current increases to −0.9 A, another step B appears that is consistent with the reduction peak B at −1.7 V in Figures 2 and 3. This step reduces Si(II) to Si(0) on the Fe electrode.
Chronopotentiometry results at different current intensities on a Fe electrode in NaCl-NaF-KCl-SiO2 (cSiO2=0.177 mol·L−1) melt at 1103 K.
The diffusion coefficient of Si(IV) ions in the molten salt can be calculated by the Sand equation using the current curve of −0.6 A [18]:
where I is the current (A), τ is the transition time (s), S is the electrode surface area (cm2), C0 is the solute concentration (mol·m−3), and D is the diffusion coefficient (cm2·s−1). The estimated value of DSi(IV) is 7.9×10−3 cm2·s−1.
Chronoamperometry
Figure 5 displays the potentiostatic transient curves obtained using chronoamperometry at 1,103 K for the NaCl-NaF-KCl-SiO2 (cSiO2=0.177 mol·L−1) molten salt system. In Figure 5(a), there is a distinct step A between the potentials of −0.7 and −0.8 V. Because silicon was reduced continuously in the molten salt, there was imbalanced Si(IV) concentration near the cathode after applying the voltage, causing the current step. This is in agreement with the reduction peak A at −0.8 V observed in Figures 2 and 3, i. e. Si(IV) + 2 e → Si(II). In Figure 5(b), there is another obvious step B between −1.6 and −1.7 V, indicating the presence of another electron transfer. This is consistent with the reduction peak B at −1.7 V in Figures 2 and 3. Namely, Si(II) accepts two electrons and is reduced to Si(0).
Chronoamperometry results at different voltage on a Fe electrode in NaCl-NaF-KCl-SiO2 (cSiO2=0.177 mol·L−1) melt at 1,103 K. (a) Peak A, shown in the voltage range of −0.5 to −0.8 V, and (b) peak B in the voltage range of −1.4 to −1.7 V.
Instant change of current with time corresponds to the electrochemical crystallite nucleation and growth of silicon on the cathode. At the beginning of this instant crystallization process, nuclei are formed at the active sites on the substrate surface. According to the principle of thermodynamics, the nuclei are independent of each other, and can grow to the critical radius in a process controlled by hemispherical diffusion [19]. After forming the first nuclei, the current increases to an extremum maximum value quickly, mainly due to the electrical double layer and the continuous reduction of Si(IV) ions on the cathode to form Si nuclei. When the concentration of Si(IV) ions cannot match the Si nucleation rate, a concentration polarization occurs on the cathode surface and the current is reduced. The crystal nuclei of Si then grow without forming new ones, so the current tends to be stable [20, 21].
According to the three-dimensional nucleation mechanism [20, 22], when the electrical crystallization process is controlled by the diffusion of the electroactive ions and the gain and loss of the electron, and it is continuous nucleation growth, the relationship between I and t can be represented by eq. (5). On the other hand, if the process is instantaneous growth of the formed nuclei, the relation between I and t can be expressed by eq. (6):
where Kn is the nucleation rate constant, N0 is the maximum nuclei density, M is the atomic weight of the electrodeposited species, and ρ is electrodeposition density. According to the chronoamperometry data at the potential −0.8 V, a relationship between I–t3/2 and I–t1/2 was obtained. Figures 6(a) and 6(b) show the corresponding correlation.
Curves of (a) I–t1/2 and (b) I–t3/2 on a Fe electrode in NaCl-NaF-KCl-SiO2 (cSiO2=0.177 mol·L−1) melt at 1,103 K.
Linear regression analysis of the data in Figure 6 leads to eqs. (7) and (8). The R2 values indicate that eq. (7) describes the results better [23].
According to the fitted eq. (8), it can be inferred that Si deposition in the NaCl-NaF-KCl-SiO2 molten salt system is consistent with the model of hemispherical three-dimensional nucleation followed by instantaneous growth. Based on the chronoamperometry data at the potential of −0.8 V, the diffusion coefficient was calculated by the Cornell equation [16].
The calculated value of DSi(IV) is 9.06×10–5 cm2·s−1.
Calculation of diffusion coefficient
Figure 7 shows the cyclic voltammetry curves at various scan rates (0.5–0.9 V·s−1). These curves are consistent with Curve b in Figure 2, showing two reduction peaks A and B at −0.8 and −1.7 V vs. Pt, respectively. The peaks correspond to the deposition and dissolution of Si. The cathode and anodic peak potentials (Epc, Epa) of peaks A and B move toward the negative direction with increasing scan rate v.
Cyclic voltammograms of NaCl-NaF-KCl-SiO2 (cSiO2=0.177 mol·L−1) melt on the Fe electrode at 1,103 K under various scan rates.
According to the cyclic voltammogram at a sweep rate of 0.9 V·s−1, the relation between Ipc and v1/2 (shown in Figure 8) is nonlinear, indicating that the cathodic process corresponding to peaks A and B meets the criterion of the quasi-reversible electrode process. That means that the electrode reactions of Si(IV)→Si(II)→Si(0) are controlled by the diffusion of the electroactive ions, and the gain and loss of electronic. The diffusion coefficient of Si(IV) ions can be expressed by the Berzins–Delahay equation [9, 24]:
Cathodic peak currents at different potential scan rates on a Fe electrode in NaCl-NaF-KCl-SiO2 (cSiO2=0.177 mol·L−1) melt at 1,103 K. (a) Ipc vs. v1/2, corresponding to Si(IV)→Si(II), (b) Ipc vs. v1/2, corresponding to Si(II)→Si(0).
where v is the potential scan rate (V·s−1). At the rate of −0.9 V·s−1, the value of Ipc at peak A is 0.291 A, and the estimated value of DSi(IV) is 1.16×10–5 cm2·s−1.
Surface composition analysis
Based on the results of electrochemical analysis, 1.2 % silicon steel sheet was used as the substrate for the electrodeposition experiments. Figures 9 and 10 are the energy-dispersive X-ray spectroscopy (EDS) mapping results of the silicon steel sheet substrate before and after Si deposition in the NaCl-KCl-NaF-SiO2 molten salt system, respectively. The Si content in the deposition diffusion layer was increased, and Si diffused into the substrate. This indicates that the deposition of Si at 1,103 K is accompanied by diffusion, forming a Fe–Si diffusion layer. The surface scans also showed that Si is uniformly distributed in the Fe–Si alloy before and after the deposition.
EDS mapping result of the 1.2 % Si silicon steel matrix.
EDS mapping result of the silicon steel sample after deposition.
Phase analysis by x-ray diffraction
The X-ray diffraction (XRD) patterns of sediment coating (Figure 11) show that the intermetallic compounds Fe3Si, FeSi, and Fe5Si3 are present on the surface of the deposited layer. According to our calculation with the HSC thermodynamic software [5] and the Fe–Si alloy phase diagram in Figure 12 [25]. In the range of the temperature of solid fusion, the region of α phase expands to 13wt%Si. With the increase of Si content, the transition temperature of α–γ increases, and the transition temperature of γ–δ decreases. As the silicon content and temperature change, a closed two-phase region formed between α phase and γ phase. The α1 phase is DO3 ordered based on Fe3Si, and the α2 phase is the B2 ordered phase of Fe3Si. Under 540°C, α2 phase eutectoid decomposes into disordered α-Fe(Si) solid solution (bcc phase) and α1 phase. Fe and Si could undergo seven possible reactions at 1,003–1,103 K to form different intermetallic compounds:
XRD patterns on the surface of the deposited layer.
The Fe–Si phase diagram.
The composition of the Fe–Si system is calculated by the phase diagram as shown in Figure 13 at 1103 K. With increasing Si content, the content of Fe in the ferrosilicon system is gradually reduced. Furthermore, the first intermetallic compound to form is Fe3Si. When the amount of silicon reaches 0.1 and 0.75 kmol, Fe5Si3 and FeSi are formed, respectively. The content of Fe3Si first increases and then decreases with the silicon content, reaching a peak value at 0.85 kmol. The same trend occurs for Fe5Si3 with a peak concentration at 2.0 kmol Si. In contrast, the content of FeSi monotonously increases with the Si concentration. Figure 10 shows that Fe3Si, FeSi, and Fe5Si3 are all formed in the deposited coating, and the surface Si content should be at least 0.75 kmol or 11 %, where non-negligible amounts of Fe5Si3 start to form according to Figure 13 [25].
The balance of Fe–Si system, for Si contents up to 3 kmol.
Conclusions
The electrochemical reduction of Si on Fe electrode in the NaCl-NaF-KCl-SiO2 molten salt system was studied by various electrochemical analytic methods. The following conclusions can be drawn:
The electrochemical reduction of Si(IV) occurs in two steps of two-electron transfer: Si(IV)→Si(II)→Si(0).
The electrode reaction of Si(IV) is a quasi-reversible process, which is controlled by both the ion diffusion rate and electron transport rate.
The electrical crystallisation of Si in the NaCl-NaF-KCl-SiO2 system agrees with the three-dimensional nucleation and instantaneous growth mechanisms.
According to the surface scans, silicon is distributed uniformly in the Fe–Si alloy. The XRD spectra confirmed the formation of intermetallic compounds Fe3Si, FeSi, and Fe5Si3.
According to all of the results, the two-step mechanism and the diffusion constants would be the new findings which have not been reported in this or similar systems.
Funding statement: The National Natural Science Foundation of China: 51401075, 51674120. Natural Science Foundation of Hebei Province: E2016209163. Hebei Provincial Department of Education: BJ2017050
References
[1] D. Ruiz, T. Ros-Yanez, R.E. Vandenberghe, E.D. Grave, M.D. Wulf and Y. Houbaert, Jpn. J. Appl. Phys., Part 1., 93 (2003) 7112–7114.10.1063/1.1558196Search in Google Scholar
[2] X.D. He, X. Li and Y. Sun, J. Magn. Magn. Mater., 320 (2008) 217–221.10.1016/j.jmmm.2007.05.030Search in Google Scholar
[3] D. Dong, C. Liu, S. Chen and B. Zhang, Int. J. Min Met Mater., 16 (2009) 208–214.10.1016/S1674-4799(09)60035-2Search in Google Scholar
[4] J. Barros, T. Ros-Yanez, O. Fischer, J. Schneider and Y. Houbaert, J. Magn. Magn. Mater., 304 (2006) 614–616.10.1016/j.jmmm.2006.02.195Search in Google Scholar
[5] H. Li, J. Li, X.P. Zhao, Y. Li and F. Zhang, Special Steel., 35 (2014) 12–14.Search in Google Scholar
[6] R. Boen and J. Bouteillon, J. Appl. Electrochem., 13 (1983) 277–288.10.1007/BF00941599Search in Google Scholar
[7] T. Oishi, M. Watanabe, K. Koyama, M. Tanaka and K. Saegusa, J. Electrochem. Soc., 158 (2011) E93–E99.10.1149/1.3605720Search in Google Scholar
[8] T. Oishi, K. Koyama and M. Tanaka, J. Electrochem. Soc., 163 (2016) E385–E389.10.1149/2.0391614jesSearch in Google Scholar
[9] A.L. Bieber, L. Massot, M. Gibilaro, L. Cassayre, P. Taxil and P. Chamelot, Electrochim. Acta., 62 (2012) 282–289.10.1016/j.electacta.2011.12.039Search in Google Scholar
[10] A.L. Bieber, L. Massot, M. Gibilaro, L. Cassayre, P. Chamelot and P. Taxil, Electrochim. Acta., 56 (2011) 5022–5027.10.1016/j.electacta.2011.03.099Search in Google Scholar
[11] S.V. Kuznetsova, V.S. Dolmatov and S.A. Kuznetsov, Russ. J. Electrochem., 45 (2009) 742–748.10.1134/S1023193509070052Search in Google Scholar
[12] Y. Hu, X. Wang, J. Xiao, J. Hou, S. Jiao and H. Zhu, J. Electrochem. Soc., 160 (2013) D81–D84.10.1149/2.038303jesSearch in Google Scholar
[13] C. Huang, H. Zhang, B. Zhao and H.T. Guo, Mater Prot., 30 (1997) 13–15.Search in Google Scholar
[14] M. Li, W. Li, W. Han, M. Zhang and Y. Yan, Chem J Chinese U, 35 (2014) 2662–2667.Search in Google Scholar
[15] F. Zhu. Electrochemical behavior of Cerium(III) and Co-deposition of Mg-Li-Ce and Al-Li-Ce in molten salts, M.S. Thesis, Harbin Engineering University, 2012.Search in Google Scholar
[16] G.M. Haarberg, L. Famiyeh, A.M. Martinez, M. Zhang and Y. Yan, Electrochim. Acta., 100 (2013) 226–228.10.1016/j.electacta.2012.11.052Search in Google Scholar
[17] C. Hamel, P. Chamelot, A. Laplace, E. Walle and O. Dugne, Electrochim Acta., 52 (2007) 3995–4003.10.1016/j.electacta.2006.11.018Search in Google Scholar
[18] P. Cao, M.L. Zhang, W. Han, Y. Yan, S. Wei and T. Zheng, J Rare Earths., 29 (2011) 763–767.10.1016/S1002-0721(10)60538-8Search in Google Scholar
[19] Y.G. Li, The Basic Study on Fe-W Functionally Gradient Material Made by Electrodeposition in Molten Salt, Ph.D. Thesis, Northeastern University, 2005.Search in Google Scholar
[20] J. Li, X.Y. Zhang, Y.B. Liu, Y.G. Li and R.P. Liu, Rare Metals, 32 (2013) 512–517.10.1007/s12598-013-0156-4Search in Google Scholar
[21] J. Li, X.Y. Zhang, Y.B. Liu, Y.G. Li and R.P. Liu, Rare Metal Mat Eng, 42 (2013) 2237–2241.10.1016/S1875-5372(14)60028-XSearch in Google Scholar
[22] Y. Castrillejo, M.R. Bermejo, P.D. Arocas, A.M. Martínez and E. Barrado, J. Electroanal. Chem., 575 (2005) 61–74.10.1016/j.jelechem.2004.08.020Search in Google Scholar
[23] M.R. Bermejo, J. Gómez, J. Medina, A.M. Martínez and Y. Castrillejo, J. Electroanal. Chem., 588 (2006) 253–266.10.1016/j.jelechem.2005.12.031Search in Google Scholar
[24] V. Smolenski, A. Novoselova, A. Osipenko, C. Caravaca and G.D. Córdoba, Acta, 54 (2008) 382–387.Search in Google Scholar
[25] Y.G. Li, J.L. Liang, H. Li, G.Z. Tang and W. Tian, Chin J Nonferrous Met, 19 (2009) 714–719.10.1016/S1003-6326(08)60338-6Search in Google Scholar
© 2018 Walter de Gruyter GmbH, Berlin/Boston
This article is distributed under the terms of the Creative Commons Attribution Non-Commercial License, which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
