Influence of F⁻ on the electrochemical properties of titanium ions and Al–Ti alloy electrodeposition in molten AlCl₃–NaCl

Abstract

The present study reports the electrochemical properties of titanium ions in molten AlCl₃–NaCl with various concentrations of fluoride anions. Transient electrochemical techniques, such as chronopotentiometry, cyclic voltammetry and square wave voltammetry, were performed to reveal the diffusion coefficients of the electroactive species and reduction steps. Results suggest that there are three steps for titanium ion reduction in molten AlCl₃–NaCl: Ti(IV) → Ti(III), Ti(III) → Ti(II) and Ti(II) → Ti. The reduction processes transferred to Ti(IV) → Ti(III) and Ti(III) → Ti when NaF was added in the melt in increasing amounts. The influence of fluoride anions on the Al–Ti alloy deposition was also investigated. A constant potential electrolysis was carried out for depositing Al–Ti alloys and results illustrated that the Al–Ti alloys with different molar ratios were obtained under different amounts of NaF in the melt.

---

Introduction

Interest in light metals such as aluminum and titanium has increased steadily during the recent decades as their physical and chemical properties provide solution to many engineering problems especially in the aircraft, aerospace and automobile sectors. Among the many procedures that can be carried out for the deposition of a particular metal or alloy on a substrate, electrodeposition is found to be a good tool because it offers the possibility to control the thickness of the deposit, the composition and microstructure.^1–3

The chloroaluminate electrolytes possess a number of interesting features, such as relatively high ionic conductivity, low melting point and a wide electrochemical potential window.^4,5 Electrodeposition from chloroaluminate electrolytes has been used to produce aluminum alloys, and it offers significant benefits for improving energy efficiency and reducing the environmental impact in the production of Al alloys. Stafford et al. reported the electroplating of Al–Ti alloys from mixture of 2AlCl₃–NaCl in which titanium was anodically dissolved.^6,7 The composition, morphology, and crystallographic microstructure of Al–Ti alloys electrodeposited were investigated. It was also revealed that the apparent limit on alloy composition is apparently due to a mechanism by which Al₃Ti forms through the reductive decomposition of [Ti(AlCl₄)₃]⁻. The other work done on the same system by Stafford et al. demonstrated that alloys containing up to 28% atomic fraction Ti could be electrodeposited with controlled additions of Ti(II).⁸ Al–Ti alloy electrodeposition in other molten salts like AlCl₃–NaCl–KCl at 200 °C has also been studied.^9,10 It was seen that titanium content in the deposited film increased with the addition of TiCl₃, and smooth deposit of Al-40.2% Ti could be obtained. In contrast, it was revealed that the divalent species may be used to electrodeposit Al–Ti alloys and the trivalent species is sparingly soluble in 2AlCl₃–NaCl.⁸

The electrochemical behavior of titanium ions in electrolytes is very complex, partly due the possible presence of titanium in the 2, 3, and 4 oxidation states.^11–13 The stability of titanium ions depends on the properties of the electrolyte.^14–16 Although Ti(II), Ti(III) and Ti(IV) are existing in both chloride and fluoride melt, Ti(IV) and Ti(III) are much more stable than Ti(II) and the reduction of trivalent titanium directly leads to titanium metal deposition in the presence of fluoride anions in molten salt.^15–18 Thus, the use of fluoride media results in stabilization of the high oxidation states because the metal ions in these oxidation states form very stable complexes with fluoride anions.

It has been shown by electrochemical and spectroscopic methods that molten salt systems containing AlCl₃, have a tendency to stabilize lower valence states of metal ions.^19,20 However, when fluoride anions are present in the molten melts, the Al ions in molten AlCl₃–NaCl, results in the direct reduction of Al(III) to Al in a single-step process exchanging three electrons.²¹ As per the authors' knowledge, no work is available which illustrates the effect of fluoride anions on titanium ions reduction in molten AlCl₃–NaCl. The purpose of this paper is to study the influence of fluoride anions on the electrochemical properties of titanium ions in AlCl₃–NaCl electrolyte at a low temperature. In addition, the electrodeposition of aluminum–titanium alloys from this electrolyte is reported.

Experimental

A. Chemical and electrochemical apparatus

All experiments were performed in a 2 : 1 mole ratio of molten AlCl₃–NaCl maintained at 180–230 °C. The AlCl₃ (Sinopharm Chemical Reagent Co., Ltd., anhydrous, 99.9%) was sublimed at 250 °C in a helium atmosphere prior to use. The NaCl (Sinopharm Chemical Reagent Co., Ltd., analytical reagent, 99.9%) was dried at 300 °C in vacuum (0.1 Pa) for 24 hours. Subsequently, it was heated up to melt at 850 °C in a high-purity argon atmosphere (Ar, Sinopharm Chemical Reagent Co., Ltd., 99.999%), and the flowrate was 30 mL min⁻¹. Additionally, high-purity hydrogen chloride gas (HCl, Sinopharm Chemical Reagent Co., Ltd. analytical grade 99.999 pct) was bubbled into the salt to remove the O²⁻. Details have been published elsewhere.^11,12 The Ti(II) was added into the melt through the powder NaCl–TiCl₂, and the method for the preparation of a salt containing titanium sub-chloride was described in our previous work.^{11,12,14–16} NaF was used as the source of fluoride anions, and it was processed as NaCl. The powders of AlCl₃, NaCl and NaCl–TiCl₂ were thoroughly mixed in a crucible and sealed in the cell.

The schematic diagram of electrochemical cell shows in Fig. 1. The entire cell, including the cell top, was maintained at elevated temperature to prevent the selective removal of AlCl₃ by condensation on the cooler surfaces above the electrolyte.

Fig. 1 Schematic diagram of electrochemical cell.

B. Electrodes and electrochemical techniques

The reference electrode was a 2.0 mm diameter aluminum wire (99.998%) immersed in AlCl₃–NaCl, while the counter electrode was a 6.0 mm diameter graphite rod. The working compartment contained a 1.0 mm diameter molybdenum wire for analytical measurements. Transient electrochemical techniques, such as chronopotentiometry (CE), cyclic voltammetry (CV) and square wave voltammetry (SWV) were performed using an electrochemical work station (Solartron 1287/Solartron 1255B, AMETEK Advanced Measurement Technology, UK).

Al–Ti alloy depositing experiment was carried out in the AlCl₃–NaCl melt and the mass of salt was 200 g containing 3.0 wt% of Ti(II). A titanium plate and a stainless steel plate (20 mm × 10 mm × 2 mm) were used as anode and cathode, respectively. A constant potential electrolysis was carried out for depositing Al–Ti alloys. Then, the product on the cathode was washed in an ultrasonic tank containing a solution of water and ethanol for dissolving the salt, and the deposit was detached from the cathodic substrate. Finally, the cathode products were characterized by X-ray diffraction (XRD, Rigaku, D/max-RB) and field emission scanning electron microscopy (FESEM, JEOL, JEM-6701F).

Results and discussion

A. Electrochemical properties of titanium ions in AlCl₃–NaCl

The chronopotentiograms of the cathodic behavior of Ti(II) in AlCl₃–NaCl melt under various current densities are shown in Fig. 2. As can be seen from Fig. 2 that the transition time is decreasing with the increase of the current density.

Fig. 2 Chronopotentiograms on a Mo electrode (S = 0.35 cm²) in AlCl₃–NaCl (C_Ti(II) = 0.5 mol dm⁻³) at temperatures of (a) 180 °C; (b) 200 °C; (c) 230 °C, and (d) linear relationship between jτ^1/2 and current density corresponding to the chronopotentiograms at different temperatures.

The eqn (1) demonstrates the relationship between the transition time and the current density.²²

where i is the current, A; τ the transition time, s; the concentration of Ti(II), mol dm⁻³ (the density of the salt can be expressed as: ρ_{2AlCl3–NaCl} = 2011 − 0.92 × T g dm⁻³ (ref. 23)); n the electron charged number; A electrode area, cm²; D₀ diffusion coefficient, cm² s⁻¹.

The transition time was determined by measuring the duration of the first part of the chronopotentiogram. From eqn (1), iτ^1/2 is constant, when the concentration of Ti(II) is constant. Fig. 2(d) exhibits the relation of jτ^1/2 and current density under different temperatures. It shows that the data can be fitted to a line.

The titanium ions in molten AlCl₃–NaCl detected by chemical analysis are primary Ti(II) and the average valence of titanium ions is around 2.08. More details on the chemical analysis could be found everywhere.^11,12,15,16 By using the average value jτ^1/2, the diffusion coefficients of Ti(II) were calculated using eqn (1) and the results are shown in Table 1. The diffusion coefficient of Ti(II) obtained in various melts and temperature were also listed in this table. The results suggest that the diffusion coefficient of titanium ions fluctuated from 2.5 × 10⁻⁶ cm² s⁻¹ to 35.8 × 10⁻⁶ cm² s⁻¹ in various molten salts and from 0.16 × 10⁻⁶ cm² s⁻¹ to 0.43 × 10⁻⁶ cm² s⁻¹ in ionic liquid at temperature range of 80–598 °C. While considering our experiments, it can be concluded that the diffusion coefficient of Ti(II) increased with the increasing of temperature range of 180–230 °C.

Table 1 The diffusion coefficient of Ti(II) ion obtained by different electrochemical methods at various molten salts/ionic liquid systems

Electrolyte	T/°C	Electrochemical method	D0 (×10^−6 cm^2 s^−1)
AlCl3–NaCl (65 : 35 mol%)	185	Cyclic voltammetry	2.6 (ref. 24)
		Chronopotentiometry	2.5 (ref. 24)
		Liner sweep voltammetry	3.1 (ref. 24)
	260	Liner sweep voltammetry	6.0 (ref. 24)
2AICI3–NaCI	150	Cyclic voltammetry	4.0 (ref. 7)
	180		6.51 [this work]
2AICI3–NaCI	200	Chronoamperometry	8.11 [this work]
	230		9.88 [this work]
AlCl3–EtMeImCl	80	Voltammetry	0.22 (ref. 25)
AlCl3–DMSO2	130	Cyclic voltammetry	0.16 (ref. 26)
		Chronoamperometry	0.43 (ref. 26)
LiCl–KCl	470	Chronoamperometry	19.5 (ref. 27)
LiCl–KCl	456	Chronoamperometry	10.3 (ref. 28)
	598	Chronoamperometry	35.8 (ref. 28)
LiCl–KCl	500	RDE	4.6 (ref. 29)
NaCl–MgCl2–KCl	475	RDE	1.7 (ref. 29)
		Cyclic voltammetry	6.7 (ref. 30)
CaCl2–NaCl	550	Chronopotentiometry	18.0 (ref. 30)
		Convolution	17.0 (ref. 30)

---

B. Influence of fluoride on the electrochemical properties of Ti(II)

The F/Ti ratio was used to reveal the influence of fluoride anions on the electrochemical properties of titanium ions, and it was defined as the molar ratio of F⁻ and Ti²⁺ in molten AlCl₃–NaCl. The definition was used to illustrate the influence of fluoride anions in molten NaCl–KCl.¹⁸ Fig. 3 shows the cyclic voltammograms of titanium ions in molten AlCl₃–NaCl in various F/Ti ratios at 200 °C.

Fig. 3 Cyclic voltammograms of titanium ions in molten AlCl₃–NaCl at 200 °C on a Mo working electrode obtained at various F/Ti ratios, reference: Al.

As shown in Fig. 3, the reduction peaks around the left dash line (0.29 V) transferred to a negative direction from 0.29 V to 0.17 V when F/Ti ratio increased from 0 to 4.0. The results also suggest that the redox peaks become apparent around 1.15 V when F/Ti ratios are higher than 2.0.

In order to confirm the reduction reactions, the cathodic behavior of titanium ions in molten AlCl₃–NaCl was studied by the square wave voltammetry. The results observed by the square wave voltammetry under various F/Ti ratios are shown in Fig. 4. They had a same potential range of 0–1.9 V vs. Al as that in the cyclic voltammograms.

Fig. 4 Square wave voltammograms for reduction titanium ions under various F/Ti ratios. Pulse height: 25 mV, potential step: 3 mV, frequency: 20 Hz.

The reduction potentials from square wave voltammograms in Fig. 4 are corresponding well with the results illustrated in cyclic voltammograms. Moreover, it can be seen that the current density decreased with the increasing of fluoride concentration in molten salt. The reduction peaks become apparent around 1.18 V (the second dash line) when the F/Ti ratios are higher than 2.0. Then, the current density increased with the increasing of fluoride anions in molten salt.

The result observed by the mathematical analysis of square wave voltammogram (F/Ti equals to 4.0) is shown in Fig. 5. Using the half peak (W_1/2) of Gaussian wave, the number of exchanged electrons can be calculated. The relationship between the half peak of Gaussian wave and the number of exchanged electrons is shown in eqn (2).

Fig. 5 Experimental results and square wave voltammogram after Gauss fitting for the reduction peaks. Pulse height: 25 mV, potential step: 3 mV, frequency: 20 Hz.

The relationship between the square wave voltammogram and the half peak of the Gaussian wave are shown in Fig. 5. For reduction of A, the number of electrons transferred is close to 1.16, and it is 2.78 for peak B. The results suggest that the cathodic reaction of Ti(IV) in molten AlCl₃–NaCl precedes two steps:

Ti(IV) + 1e⁻ = Ti(III) (reaction A) (3)

Ti(III) + 3e⁻ = Ti (reaction B) (4)

There are two disproportionations existing in the melt while the presence of metallic titanium which can be described as below:

3Ti(II) = 2Ti(III) + Ti (5)

4Ti(III) = 3Ti(IV) + Ti (6)

The Ti(III) ion has a competition with Ti(II) in reaction (5) and it also has competition with Ti(IV) in reaction (6) in the molten salt. The equilibrium constant for these reactions can be expressed as follow:where K_ci is equilibrium constant and a_i is the activity of a species i.

The equilibrium will transfer to the right direction for above two reactions when fluoride was added in the melt. That is due to relatively higher stability of the complex ions of higher valance cations with the fluoride ion (F⁻), compared to lower valance cations (Ti²⁺, Ti³⁺). It was investigated by our previous work that the titanium ions in equilibrium with metallic titanium are Ti²⁺ and Ti³⁺ in the chloride melt with low concentration of fluoride. In the melts with high fluoride concentration, Ti²⁺ was undetectable and Ti⁴⁺ was detected.¹² It is because that almost all Ti³⁺ and Ti⁴⁺ forms the complex with 6 coordination number, TiF₆³⁻ and TiF₆²⁻ in the high fluoride concentration melts. Thus, stable complexes of Ti(III) and Ti(IV) with F⁻ were formed when NaF was added in the melt.

In order to interpret the mechanisms about the influence of fluoride anions on the electrochemical behavior of titanium ions, a schematic was employed shown in Fig. 6. The reaction (Ti³⁺ + xF⁻ = TiF_x^3−x) occurred when the presence of fluoride ion in the melt. It could lead to the disproportionation (5) transfer to the right direction.

Fig. 6 Schematic of the influence of fluoride anions on the electrochemical behavior of titanium ions in molten AlCl₃–NaCl.

Electrochemical analysis data in this work revealed that there are two steps for Ti(III) ions reducing in AlCl₃–NaCl melt. A single step was observed when F/Ti ratio equals to 4.0. It means Ti(III) is much stable under a higher fluoride concentration in the melt, and it also suggests that the equilibrium for the disproportionation transferred to the right direction.

Therefore, the electroactive species for reaction B shown in Fig. 5 is probably Ti(III) complexes. The same ideas can be used to explain the results when F/Ti equals to 2.0. The diffusion coefficient of the species for each situation was clarified by chronopotentiometry, and the results are shown in Fig. 7. The current densities ranging from 0.45 mA cm⁻² to 1.14 mA cm⁻² were used. The current density and jτ^1/2 were plotted in Fig. 8. It can be found that the data can be fitted a line for each results under various F/Ti ratios.

Fig. 7 Chronopotentiograms for electroactive species reduction on a Mo electrode (S = 0.35 cm²) in AlCl₃–NaCl (C_Ti(II) = 0.5 mol dm⁻³) under F/Ti molar ratio of (a) 0.5; (b) 1.0; (c) 2.0; (d) 4.0.

Fig. 8 Linear relationship between jτ^1/2 and current density corresponding to the chronopotentiograms at various F/Ti ratios.

By using the average value jτ^1/2, the diffusion coefficients for the electroactive species were calculated by using eqn (1), and the results are shown in Table 2. The electroactive species as mentioned in Table 2 are expected Ti(II) and Ti(III), respectively. The judgment is based on the analysis of square wave voltammograms.

Table 2 The diffusion coefficient of the expected electroactive species under various F/Ti ratios

F/Ti	Expected electroactive species	D0 (cm^2 s^−1 × 10^6)
0.5	(Ti(II))	8.98
1.0	(Ti(II))	6.70
2.0	(Ti(III))	2.19
4.0	(Ti(III))	5.03

---

The trend for the diffusion coefficient changes in Table 2 can offer another perspective on the change in state of titanium ions. The diffusion coefficient of Ti(II) decreased when the F/Ti ratios increased from 0.5 to 1.0 which means the diffusion for the species, such as [Ti(AlF₄)₃]⁻, is more difficult than [Ti(AlCl₄)₃]⁻. The Ti(III) will become more stable when F/Ti ratio is higher than 2.0 and continues to increase. The competition between Ti(II) and Ti(III) was reduced, and the diffusion coefficient for Ti(III) was enhanced in this situation. They are according with the changes of current peaks shown in Fig. 3 and 4.

C. Electrodeposition of Al–Ti alloy in molten AlCl₃–NaCl

According to the Al–Ti phase diagram as shown in Fig. 9, titanium can form four solid intermetallic compounds with aluminum at our experimental temperature namely TiAl₃, TiAl₂, TiAl, and Ti₃Al, respectively.

Fig. 9 Titanium–aluminum phase diagram.³¹

The electrodeposition of aluminum–titanium alloys in this paper was carried out in molten AlCl₃–NaCl under various concentrations of fluoride anions. Considering the co-reduction potentials, the potential −0.5 V vs. Al was applied in the series of experiments. Fig. 10 shows the SEM photographs of the electrodeposit. The EDX analysis of individual particles in Fig. 10 is presented in Table 3.

Fig. 10 SEM photographs of the surface of electrodeposited Al–Ti alloy films under the F/Ti molar of (a) 0.5; (b) 1.0; (c) 2.0; (d) 4.0.

Table 3 The EDX analysis results of particles shown in Fig. 10

Point	Composition of particle (EDX analysis)
	Elements	at pct
#1	Al	71.81
	Ti	28.19
#2	Al	79.06
	Ti	20.94
#3	Al	81.33
	Ti	18.67
#4	Al	87.68
	Ti	12.32

---

It can be clearly seen from Fig. 10 and Table 3 that, with the increase of F/Ti from 0.5 to 4.0, the titanium content in the Ti–Al alloy decreases from 28.19 to 12.32 at pct. Thus, the analysis confirms that Al and Ti–Al alloy were co-deposited from electrolytes. The titanium content in the alloy was higher when F/Ti ratio was at a lower value, and Ti and Al may be present as Al₃Ti intermetallic from the EDX atomic percent analysis. It can be explained by the nobler deposition voltage of titanium with compared to aluminum.³² It is also noticeable that the average particle size decreases with the increasing of fluoride anions in the melt. The co-reduction potentials under F/Ti ratios of 0.5, 1.0, 2.0 and 4.0 ​estimated from a linear sweep voltammetry are around −0.23 V, −0.17 V, −0.14 V and −0.11 V, respectively. The over potentials for metallic titanium depositing were applied in these experiments can be calculated as 0.27 V, 0.33 V, 0.36 V and 0.39 V, respectively. A higher over potential leads to a finer particle size, and this can be explained by considering the nucleation phenomenon.

Fig. 11 concludes the average atomic percent of Ti in Ti–Al alloy under various F/Ti ratios. It suggests that the titanium contents decreased with the increasing of the molar ratio of fluoride versus aluminum.

Fig. 11 Variation of average atomic percents of Ti and Al under different F/Ti ratios.

According to the above analysis, an alloy with a higher content of titanium could be prepared under a lower concentration of fluoride anions (the F/Ti is smaller than 0.5). The XRD pattern of the alloy shows in Fig. 12. The products are primary TiAl₃, and Al was also observed in the XRD pattern.

Fig. 12 X-ray diffraction patterns of electrodeposited Al–Ti alloy films F/Ti = 0.5.

The polarization power between Ti(III) and fluoride anions is more powerful than that between Ti(II) and fluoride anions. It means that it is easier for the reductive decomposition of [Ti(AlF₄)₃]⁻ than [Ti(AlF₄)₄]⁻. Thus, the content of titanium in the alloy decreased with the increasing of fluoride anions. For a lower content of fluoride anions in the melt, the reductive decomposition of [Ti(AlCl₄)₃]⁻ contributes to a higher titanium content in the alloy. The phenomenon was also confirmed by Stafford et al.⁸

Conclusions

The influence of fluoride anions on the electrochemical behavior of titanium ions in molten AlCl₃–NaCl was revealed in the study. The fluoride ions, essentially F/Ti ratio was altered by adding NaF in the molten salt. The reduction steps for titanium ions were investigated by transient electrochemical techniques such as cyclic voltammetry and square wave voltammetry. Results showed that there are three steps for cathodic reduction of titanium ions when the F/Ti ratio is lower 2.0. Two steps process was observed when the F/Ti ratio is higher 2.0. The diffusion coefficients for the expected electroactive species were calculated and their significance was discussed at various conditions. The production of Al–Ti alloys was successfully carried out on a stainless steel plate cathode from molten AlCl₃–NaCl at a molar ratio of F/Ti ranging from 0.5 to 4.0. Alloys with a relatively high content of titanium was produced when the F/Ti was lower than 2.0.

Notes and references

1. M. Ueda, D. Susukida, S. Konda and T. Ohtsuka, Surf. Coat. Technol., 2004, 176, 202.
2. L. Legrand, A. Chaussé and R. Messna, J. Electrochem. Soc., 1998, 145, 1.
3. L. G. Boxall, H. L. Jones and R. A. Osteryoung, J. Electrochem. Soc., 1973, 120, 2.
4. Z. Wang, H. Kan, Z. Shi, B. Gao, Y. Ban and X. Hu, J. Mater. Sci. Technol., 2008, 24, 915.
5. G. R. Stafford and C. M. Haarberg, Plasmas Ions, 1999, 1, 35.
6. G. R. Stafford, T. Tsuda and C. L. Hussey, J. Min. Metall., Sect. B, 2003, 39, 23.
7. G. R. Stafford and T. P. Moffat, J. Electrochem. Soc., 1995, 142, 10.
8. G. R. Stafford, J. Electrochem. Soc., 1994, 141, 945.
9. J. Uchida, H. Seto and A. Shibuya, J. Surf. Finish. Soc. Jpn., 1995, 46, 1167.
10. M. Jafarian, F. Gobal, I. Danaee and M. G. Mahjani, Electrochim. Acta, 2007, 52, 5437.
11. J. Song, Q. Wang, M. Kang, S. Jiao and H. Zhu, Electrochemistry, 2014, 82, 1047.
12. J. Song, Q. Wang, J. Wu, S. Jiao and H. Zhu, Faraday Discuss., 2016, 190, 421.
13. D. M. Gruen and R. L. Mcbeth, Pure Appl. Chem., 1963, 6, 23.
14. M. Kang, J. Song, H. Zhu and S. Jiao, Metall. Mater. Trans. B, 2014, 46, 162.
15. J. Song, Q. Wang, G. Hu, X. Zhu, S. Jiao and H. Zhu, Int. J. Miner., Metall. Mater., 2014, 21, 660.
16. J. Song, Q. Wang, X. Zhu, J. Hou, S. Jiao and H. Zhu, Mater. Trans., 2014, 55, 1299.
17. N. Ene and S. Zuca, J. Appl. Electrochem., 1995, 25, 671.
18. F. Lantelme and A. Salmi, J. Electrochem. Soc., 1995, 142, 3451.
19. T. C. F. Munday and J. D. Corbett, Inorg. Chem., 1996, 5, 1263.
20. U. Anders and J. A. Plambe, Can. J. Chem., 1969, 47, 3055.
21. M. Gibilaro, L. Massot, P. Chamelot and P. Taxil, J. Alloys Compd., 2009, 471, 412.
22. H. J. S. Sand, Philos. Mag., 1901, 1, 45.
23. R. W. Berg, H. A. Hjuler and N. J. Bjerrum, J. Chem. Eng. Data, 1983, 28, 251.
24. K. W. Fung and G. Mamantov, J. Electroanal. Chem., 1972, 35, 27.
25. T. Tsuda, C. L. Hussey, G. R. Stafford and J. E. Bonevich, J. Electrochem. Soc., 2003, 150, C234.
26. L. Legrand, A. Chausse and R. Messna, J. Electrochem. Soc., 1998, 145, 110.
27. D. M. Ferry, G. S. Picard and B. L. Tremillon, J. Electrochem. Soc., 1988, 135, 1443.
28. G. M. Haarberg, W. Rolland and J. Thonstad, J. Appl. Electrochem., 1993, 23, 217.
29. K. Matsmoto, H. Numata, S. Haruyama and I. Ohno, Mater. Trans. JIM, 1999, 40, 1429.
30. A. M. Martinez, Y. Castrillejo, E. Barrado, G. M. Haarberg and G. Picard, J. Electroanal. Chem., 1998, 449, 67.
31. A. Kostov, B. Friedrich and D. Živković, J. Min. Metall., Sect. B, 2008, 44, 49.
32. D. Pradhan, R. G. Reddy and A. Lahiri, Metall. Mater. Trans. B, 2009, 40, 114.

---