Boron-Doped Diamond Electrodes in Molten Chloride Systems

10 Boron-Doped Diamond Electrodes in Molten Chloride Systems Takuya Goto*, Yuya Kado{, Rika Hagiwara{ *Department of Environmental Systems Science, Doshisha University, Kyoto, Japan, {Department of Fundamental Energy Science, Kyoto University, Kyoto, Japan

10.1

Introduction

10.1.1 Molten Salts Containing Oxide Ions Molten alkali and alkaline earth chlorides are attractive electrolytes having many features, such as thermal and/or chemical stabilities, wide electrochemical windows, high ionic conductivities, and high solubilities of other substances. The mixtures of different alkali and alkaline earth chlorides have lower melting points than single salts, leading extension of the usable temperature range and wide choice of the materials for the apparatus. This is why electrochemical processes with molten alkali and alkaline earth chlorides are used conventionally for the production of alkali, alkaline earth, and rare earth metals, which are impossible or difficult to obtain by electrochemical processes in aqueous solutions. In recent years, electrochemical reduction of metal oxides [1–4], nitriding of various metals [5–9], formation of rare earth metal alloys [10], synthesis of ammonia at normal pressure [11,12], and thermal activated batteries [13] have been proposed for novel applications of these chlorides. In electrochemical reduction processes of metal oxides, pure metals are formed electrochemically on cathodes from metal oxides (MxO) as shown by Reaction (10.1), and oxide ion generated in the electrolyte is simultaneously oxidized according to Reaction (10.2) or (10.3) in the case of a carbon anode: MOx þ 2xe ! M þ xO2

(10.1)

C þ O2 ! CO þ 2e

(10.2)

1 1 C þ O2 ! CO2 þ 2e 2 2

(10.3)

Generated oxide ion and oxygen gas affect electrode reactions in the processes of metal oxides and have an important role in acid–base reactions expressed as Reaction (10.4) because the oxide ion is a strong Lewis base in molten chlorides [14–16]: Base ¼ Acid þ O2 Molten Salts Chemistry © 2013 Elsevier Inc. All rights reserved.

(10.4)

188

Molten Salts Chemistry

Therefore, it is essential to understand the chemical and electrochemical behavior of oxygen gas and oxide ion. For this purpose, it is necessary to measure the thermodynamic quantities of oxide ions such as chemical potentials in molten chlorides and also to monitor and control the concentration of oxide ions for an effective electrochemical process. However, there are few reports so far on the oxygen evolution reaction in molten chlorides because a stable oxygen evolution electrode has not been developed. In addition, oxide ion is generated according to Reaction (10.5) when water exists in molten chlorides [17]: H2 O þ 2Cl ! O2 þ 2 HCl

(10.5)

In the processes with molten chlorides containing oxide ion, Reaction (10.4) can occur, resulting in deposition of metal oxides, corrosion of electrode and cell materials, and interruption of the objective reactions [18]. From this standpoint, it is important to observe and control the concentration of oxide ion and to remove it smoothly in the electrolytes for the effective processes. On the basis of these backgrounds, some researchers have reported on the oxygen electrode reaction: 1 O2 þ 2e ¼ O2 2

(10.6)

Wrench and Inman [14] made use of platinum or gold as anode materials in a lithium chloride (LiCl)–potassium chloride (KCl) eutectic melt containing lithium oxide (Li2O) used as an oxide ion source. However, the two-electron reaction was not observed and the obtained results were not simply applied for oxygen electrode Reaction (10.6). Consequently, it was suggested that the oxide ion (O2) was unstable in the oxygen atmosphere and the oxygen electrode was influenced by the presence of the peroxide (O2 2 ) or the superoxide (O2) ion [14]. Cherginets and colleagues [17] measured the electromotive force (EMF) using a stabilized zirconia as an oxide ion conductor and a platinum electrode as an anode. They suggested that O2 was stable in a LiCl-KCl eutectic melt containing a low concentration of KOH or Li2O used as oxide ion sources; however, O2 2 was more stable than O2 when high concentrations of KOH or Li2O were added. The standard formal potential of O2/O2 was investigated in a LiCl-KCl eutectic melt at 723 K. The value calculated from the solubilities of refractory metal oxides by Masuko and co-workers [19] was 2.97 V (vs Liþ/Li), while that estimated from the EMF by Kanzaki and Takahashi [15] using a platinum electrode as an anode was 2.543 V (vs Liþ/Li).These results do not coincide with each other and therefore further investigations are required. One of the conceivable reasons is inaccurate identification of the electrode reaction because an inert electrode is not available under an oxygen atmosphere at high temperatures. Platinum anodes dissolve in the positive potential regions and also react with oxide ion to form lithium platinate in molten LiCl and platinum oxide (Pt3O4) in molten CaCl2 [2] according to Reactions (10.7), (10.8), and (10.9). Pt ! Pt2þ þ 2e

(10.7)

2Liþ þ Pt þ 3O2 ! Li2 PtO3 þ 4e

(10.8)

3Pt þ 4O2 ! Pt3 O4 þ 8e

(10.9)

It has also been reported that platinum reacts directly with the atmospheric oxygen gas to form platinum oxide (PtO2) [20]. Therefore, it is highly probable that a mixed potential

Boron-Doped Diamond Electrodes in Molten Chloride Systems

189

measured for a platinum anode affected the accuracy of data in the previous studies. Thus, further investigations, preferably using an inert anode, have been required in order to improve the quality of data; therefore, development of an inert electrode under an oxygen atmosphere at high temperatures is extremely important.

10.1.2 Inert Oxygen Evolution Electrode: A Boron-Doped Diamond Electrode For aqueous solution systems, a dimensionally stable electrode (DSE) has been developed and used industrially today [21]. In molten salts, tantalum, tungsten, silicon, platinum, DSE, and a boron-doped diamond (BDD) electrode were investigated as candidates for inert oxygen evolution electrodes in a previous study of the authors’ laboratory. As a result, it was found that the BDD electrode is able to be used as an inert oxygen evolution electrode at 723 K in a LiCl-KCl eutectic melt containing Li2O as an O2 source; however, other candidates were consumed during electrolysis [22]. If the BDD electrode is used as an inert anode, the equilibrium potential of Reaction (10.6), the chemical potential, and the solubility of oxides can be measured. In this report, the thermal and electrochemical stability of the BDD electrode was evaluated, as well as the electrochemical behavior of oxide ions on the BDD electrode in a LiCl-KCl eutectic melt and in a LiCl-NaCl-CaCl2 eutectic melt (52.3:13.5:34.2 mol%). Typical features of a boron-doped diamond contacting with molten chlorides melts were described. Then, durability tests of a BDD were performed in molten chloride melts containing oxide ions (O2) to examine the possibility of an inert anode for oxygen gas evolution. After confirming the inert property of a boron-doped diamond, standard electrode potentials for an oxygen gas electrode reaction were measured with the use of the BDD in chloride melts, and thermodynamic data of oxide ion in molten chlorides such as enthalpy, entropy, chemical potential, and activity coefficient were determined.

10.2

Stability of a Boron-Doped Diamond Electrode in Molten Chloride Systems

It is extremely important to develop an inert oxygen evolution electrode for an effective reduction process of metal oxides as mentioned in Section 10.1. It has been reported that a BDD electrode was found to act as an oxygen evolution electrode in the molten LiCl-KCl-Li2O system at 723 K in the authors’ laboratory [22] and a similar result was obtained in the molten LiCl-NaCl-CaCl2-Li2O system at 773 K [23]. However, the thermal degradation of diamond has been reported and its decomposition temperature has been considered to depend significantly on the crystallinity, plane direction, and amount of doped boron [24–27]. Therefore, it is important to evaluate the thermal and electrochemical stability of the BDD electrode in several molten chlorides. From these backgrounds, in this section, the thermal stability of the BDD electrode was investigated in an oxygen atmosphere by thermogravimetry and its electrochemical stability was examined in eutectic LiCl:KCl (58.5:41.5 mol%), LiCl-KCl (75:25 mol%), eutectic LiCl-CaCl2 (64:36 mol%), and eutectic LiCl-NaCl-CaCl2 (52.3:13.5:34.2 mol%) at 773 K.

190

Molten Salts Chemistry

10.2.1 Thermal Stability of a Boron-Doped Diamond A BDD electrode was crushed and the fragments were immersed in aqueous potassium hydroxide (>50 wt%) at 60  C for a week to eliminate the silicon substrate. Elemental silicon dissolves in a dense alkaline aqueous solution according to the following reaction: 

Si þ 2OH þ H2 O ! SiO3 2 þ 2H2 ↑

(10.10)

The result of thermogravimetric analysis of the recovered BDD is shown in Figure 10.1. The weight loss was observed at 870 to 1100 K, indicating that the BDD electrode is at least thermally stable and has the possibility of being used as an oxygen evolution electrode at lower than 870 K. The sample of about 0.9 mg remaining at higher than 1200 K is attributed to residual silicon and/or silicon dioxide. After the measurement, no traces of BDD were observed in the cell. Diamond is considered to react with O2 gas to form CO and/or CO2 according to Reactions (10.11) and (10.12). 1 C þ O2 ! CO 2

(10.11)

C þ O2 ! CO2

(10.12)

10.2.2 Oxygen Gas Evolution on a Boron-Doped Diamond Electrode in Various Melts Figure 10.2 shows cyclic voltammograms of the BDD electrode after the addition of Li2O into eutectic LiCl-KCl (58.5:41.5), LiCl-KCl (75:25), eutectic LiCl-CaCl2, and eutectic LiCl-NaCl-CaCl2 at 773 K. Anodic currents are observed at around 1.2 V vs Cl2/Cl in eutectic LiCl-KCl (58.5:41.5), LiCl-KCl (75:25) and 0.7 V vs Cl2/Cl in molten LiClCaCl2 and LiCl-NaCl-CaCl2. In order to investigate the reaction, gases generated during potentiostatic electrolysis were analyzed by recoding the variation of the concentration of oxygen gas during the potentiostatic electrolysis at 0.8 V vs Cl2/Cl in eutectic

Figure 10.1 Thermal gravimetric analysis data of BDD in an oxygen atmosphere (heating rate: 1 K min1).

1.6

Weight / mg

1.4

1.2

1.0

0.8 400

600

800

1000

Temperature / K

1200

Boron-Doped Diamond Electrodes in Molten Chloride Systems

Figure 10.2 Cyclic voltammograms on the BDD electrode in molten chlorides containing 1.0 mol% Li2O at 773 K. The scan rate is 0.1 V s1. (a) Eutectic LiCl-KCl, (b) LiCl-KCl (75:25 mol%), (c) eutectic LiCl-CaCl2, and (d) eutectic LiCl-NaCl-CaCl2.

(a) (b)

Current density / mA cm-2

191

(c)

(d)

50 mA cm-2

-2.0

-1.5

-1.0

-0.5

0.5

0.0 /Cl-

Potential / V vs Cl2

Figure 10.3 Variation of the concentration of oxygen gas during potentiostatic electrolysis at 0.8 V vs Cl2/Cl in a LiCl-KCl eutectic melt containing 2.0 mol% Li2O at 773 K.

5000

O2 concentration / ppm

4000

3000

2000

1000

0 0

10

20

30

40

50

60

Time / min

LiCl-KCl. As shown in Figure 10.3, it increased as a passage of time. CO and CO2 were the only gaseous species detected by infrared (IR) spectroscopy, as shown in Figure 10.4, where IR spectra for CO and CO2 were also drawn as references. Supposing all the currents are due to oxygen gas evolution, the current efficiency is calculated to be almost 100% from the deviation of pressure during the electrolysis in an evacuated cell. In addition, no weight loss of the BDD electrode was observed. Thus, these currents are attributed to the oxidation of oxide ion to form oxygen gas, which is expressed as Reaction (10.6). In Figure 10.2, the potentials for oxygen gas evolution are more positive by 0.5 V in LiCl-CaCl2 and LiCl-NaCl-CaCl2 than in LiCl-KCl systems. This result suggests that the surrounding environment of an oxide ion in the melts is different among them. A conceivable reason is the difference of a coulombic interaction between oxide ion and its coordinating cations. The interaction of an O2 ion with a Ca2þ ion is considered to be stronger than that with Liþ and Kþ ions. The thermodynamic stabilities of solid oxides improve in the following order: K2O < Li2O < CaO [28].

192

Molten Salts Chemistry

Figure 10.4 Infrared spectra of gas sampled after the potentiostatic electrolysis at 0.8 V vs Cl2/Cl in a LiCl-KCl eutectic melt containing 2.0 mol% Li2O at 773 K. Transmittance / A.U.

sampled gas

CO

CO2

4000

3000

2000

1000

Wavenumber / cm-1

In addition, the standard potential of O2/O2 against Cl2/Cl, E0O =O2 ðvs Cl2 =Cl Þ, in 2 LiCl-Li2O and CaCl2-CaO systems is calculated based on thermodynamic data. The standard potential is given by the following equation: E0O2 =O2 ðvs Cl2 =Cl Þ ¼ E0Mnþ =M ðvs Cl2 =Cl Þ  E0Mnþ =M vs O2 =O2




(10.13)


where E0Mnþ =M vs O2 =O2 is the standard potential of Mnþ/M (M ¼ Li, Ca) against O2/O2  0 and EMnþ =M ðvs Cl2 =Cl Þ is that against Cl2/Cl. These standard potentials are also expressed as follow:
DG0f ðLi2 OÞ E0Liþ =Li vs O2 =O2 ¼ 2F

(10.14)


DG0f ðCaOÞ E0Ca2þ =Ca vs O2 =O2 ¼ 2F

(10.15)

DG0f ðLiClÞ F

(10.16)

DG0f ðCaCl2 Þ 2F

(10.17)

E0Liþ =Li ðvs Cl2 =Cl Þ ¼ E0Ca2þ =Ca ðvs Cl2 =Cl Þ ¼

where DGf 0 (Li2O), DGf 0 (CaO), DGf 0 (LiCl), and DGf 0 (CaCl2) are the standard free energy of formation of pure chemicals at 773 K [28]. 1 2Li þ O2 ! Li2 O 2

(10.18)

Boron-Doped Diamond Electrodes in Molten Chloride Systems

193

Table 10.1 Calculated Values of Standard Formal Potentials at 773 K(M ¼ Li in LiCl, Ca in CaCl2) Melt

E0Mnþ =M =V vs O2/O2

E0Mnþ =M =V vs Cl2/Cl

E0O

LiCl

2.5742

3.5741

0.9999

CaCl2

2.8705

3.5127

0.6422

2 2 =O

=V vs Cl2/Cl

1 Ca þ O2 ! CaO 2

(10.19)

1 Li þ Cl2 ! LiCl 2

(10.20)

Ca þ Cl2 ! CaCl2

(10.21)

The calculated values are summarized in Table 10.1. The difference of E0O =O2 ðvs Cl2 =Cl Þ 2 obtained theoretically from Equations (10.13)(10.17) is 0.358 V ascribed to the difference of the melts between LiCl-KCl mixtures and the melts containing CaCl2, and this value is smaller than that obtained from the observed potentials. It suggests that the overpotential for oxygen gas evolution in LiCl-CaCl2 and LiCl-NaCl-CaCl2 is higher than that in LiCl-KCl mixtures. A large overpotential might be due to a strong adsorption of generated oxygen gas, which is caused by the low wettability of melts containing CaCl2 [2].

10.2.3 Electrochemical Stability in a LiCl-KCl Eutectic Melt Galvanostatic electrolysis at 12 mA cm2 for 20 h was performed in order to investigate the electrochemical stability of the BDD electrode in LiCl-KCl eutectic melts containing 1.0 and 2.0 mol% of Li2O at 773 K. The quantity of electricity was 800 C cm2. The potential of the BDD electrode was kept roughly around 1.0 V (vs Cl2/Cl) during the galvanostatic electrolysis accompanied by oxygen gas evolution. Oxygen gas evolution was considered to occur continuously during the electrolysis, as CO and CO2 were not observed by IR spectroscopy in the gaseous sample from the electrolysis cell. Figure 10.5 shows scanning electron microscopy (SEM) images of the BDD electrode before and after galvanostatic electrolysis at 773 K, where the morphologies are almost identical. X-ray diffraction (XRD) patterns and micro-Raman spectra of the electrode before and after electrolysis

a

5 mm

b

5 mm

c

5 mm

Figure 10.5 SEM images of the BDD electrode before and after galvanostatic electrolysis in a LiClKCl eutectic melt containing Li2O [(a) as received, (b) 1.0 mol% Li2O, and (c) 2.0 mol% Li2O].

194

Molten Salts Chemistry

Figure 10.6 XRD patterns of the BDD electrode before and after galvanostatic electrolysis in a LiCl-KCl eutectic melt containing Li2O at 773 K [(a) as received, (b) 1.0 mol% Li2O, and (c) 2.0 mol% Li2O].

Si (400) Diamond (111)

Diamond (220)

Intensity / A.U.

(c)

(b)

(a) 40

50

60

70

80

1500

1600

2q / deg.

Figure 10.7 Raman spectra of the BDD electrode before and after electrolysis in a LiCl-KCl eutectic melt containing Li2O at 773 K [(a) as received, (b) 1.0 mol% Li2O, and (c) 2.0 mol% Li2O].

diamond

Intensity / A.U.

(c)

(b)

(a)

1100

1200

1300

1400

Raman shift / cm-1

also reveal that the diamond structure is preserved after electrolysis, as shown in Figures 10.6 and 10.7. In addition, these results suggest that the stability of the BDD electrode does not depend on the O2 concentration in the melt. However, apparent consumption of the electrode was observed after electrolysis at temperatures higher than 823 K. Thus the BDD electrode is concluded to be electrochemically stable and act as an oxygen evolution electrode at temperatures lower than 773 K in molten LiCl-KCl-Li2O systems. It is considered that oxygen gas generated by electrolysis on the BDD electrode is adsorbed on the electrode surface and that its high activity leads to chemical consumption of the BDD electrode at lower temperatures.

10.2.4 Dependence of Electrochemical Stability on Melt Compositions The stability of the BDD electrode as an oxygen evolution electrode in molten LiCl-KCl (75:25), eutectic LiCl-CaCl2, and eutectic LiCl-NaCl-CaCl2 was examined in the same

Boron-Doped Diamond Electrodes in Molten Chloride Systems

a

195

b

c

d

e

5 mm

Figure 10.8 SEM images of BDD electrodes before and after galvanostatic electrolysis in molten LiCl-KCl systems containing 1.0 mol% Li2O [(a) as received, (b) eutectic LiCl-KCl, (c) LiCl-KCl (75:25), (d) eutectic LiCl-CaCl2, and (e) eutectic LiCl-NaCl-CaCl2].

manner as described for a LiCl-KCl eutectic melt. The potentials during the galvanostatic electrolysis were around 1.0 V (vs Cl2/Cl) in LiCl-KCl (75:25) and 0.05 V (vs Cl2/Cl) in eutectic LiCl-CaCl2 and eutectic LiCl-NaCl-CaCl2. Figure 10.8 shows SEM images before (a) and after galvanostatic electrolysis at 12 mA cm2 for 20 h in eutectic LiCl:KCl (58.5:41.5) (b), LiCl-KCl (75:25) (c), eutectic LiCl-CaCl2 (d), and eutectic LiCl-NaCl-CaCl2 (e) containing 1.0 mol% Li2O. The BDD electrode is as stable in LiCl-KCl (75:25) as in eutectic LiCl-KCl (58.5:41.5). This result suggests that the stability of the BDD electrode does not depend on the melt composition in molten LiCl-KCl systems and that the BDD electrode is considered to act as an oxygen evolution electrode in any compositions of LiCl-KCl at 773 K. Although no notable change was observed by XRD and Raman spectroscopy in molten LiCl-CaCl2 and LiCl-NaCl-CaCl2, SEM images in Figure 10.8 apparently show that the BDD electrode is less stable as an oxygen evolution electrode than in LiCl-KCl systems. However, the anodic currents attributed to CO and/or CO2 evolution are not observed in Figure 10.2. A conceivable explanation is chemical consumption of the BDD electrode by oxygen gas or atoms generated electrochemically on the electrode. This result might be attributed to the low wettability of the melts containing CaCl2, leading to a stronger adsorption of oxygen gas on the electrode surface and a higher overpotential for oxygen gas evolution. Therefore, the high activity of oxygen gas is considered to facilitate the consumption of the BDD electrode chemically. Another conceivable reason is the catalytic effect of calcium in the electrolyte. It is possible that calcium deposited at the cathode diffuses into the electrolyte and interacts chemically with the carbon atom of the diamond surface.

196

Molten Salts Chemistry

10.3

Thermodynamics of Oxygen Electrode Reaction on a Boron-Doped Diamond Electrode

An inert electrode under an oxygen atmosphere at high temperatures is essential for accurate identification of the oxygen electrode reaction and measurement of the standard formal potential of O2/O2. However, as described in our previous report [22], a boron-doped diamond electrode was found to act as an oxygen gas evolution electrode in molten chloride systems. Accordingly, the BDD electrode used as an inert electrode is prospective for the investigation of chemical and electrochemical behavior of oxide ions in molten chlorides. Molten LiCl-KCl mixtures of the following compositions were used as electrolytes: LiCl: KCl ¼ 58.5:41.5 (eutectic), 65:35, 70:30, and 75:25 mol%. In this chapter, the oxygen electrode reaction on the BDD electrode was investigated and the standard formal potential of O2/O2 was determined by measurement of the EMF for evaluation of the BDD electrode.

10.3.1 Standard Formal Potential of O2/O2 Figure 10.9 shows cyclic voltammograms of the BDD electrode before and after the addition of 2.0 mol% Li2O into a LiCl-KCl eutectic melt (LiCl:KCl ¼ 58.5:41.5 mol%) at 773 K. Anodic currents were observed at around 3.6 V vs Liþ/Li before the addition of Li2O, corresponding to the evolution of chlorine gas at the anode limit of the melt. After the addition of Li2O, anodic currents were observed at around 2.5 V and these are attributed to the oxidation of oxide ion to form only oxygen gas, which was confirmed by gas analyses after the potentiostatic electrolysis. This result indicates that it is possible to use the BDD electrode as an inert anode for electrochemical investigation of the oxygen electrode reaction. From these results, it is considered that the reaction on the BDD electrode is the oxygen reduction reaction as described by Reaction (10.6). Its Nernstian equation is expressed as Equation (10.22) using the gas constant, R; absolute temperature, T; Faraday constant, F; pressure of O2, PO2 ; and the anion fraction of O2, XO2 : 1 =2 RT PO2 þ ln 2F XO2

Figure 10.9 Cyclic voltammograms of the BDD electrode before and after the addition of 2.0 mol% Li2O into a LiCl-KCl eutectic melt (LiCl: KCl ¼ 58.5:41.5 mol%) at 773 K. The scan rate is 0.1 V s1.

(10.22)

120 100 Current density / mA cm-2

EO2 =O2 ¼

0 E0O2 =O2

(b)

80 60 40 20

(a)

0 -20 2.0

2.5

3.0

3.5

Potential / V vs Li+ /Li

4.0

Boron-Doped Diamond Electrodes in Molten Chloride Systems

197

0

where E0O =O2 is the standard formal potential of O2/O2, which is given by the following 2 equation using the standard potential of O2/O2, E0O =O2 ; the fugacity coefficient of O2, gO2 ; 2 and the activity coefficient of the O2 anion, gO2 : ¢ E0O2 =O2

¼

E0O2 =O2

1 =2 RT gO2 þ ln 2F gO2

(10.23)

0

According to Equation (22), E0O =O2 is obtained experimentally from the relationship 2   1 =2 between EO2 =O2 and ln PO2 =XO2 . Therefore, the dependence of the EMF of the BDD

2.64

2.64

2.62

2.62 EMF / V vs Li+/Li

EMF / V vs Li+/Li

electrode on O2 pressure and O2 concentration was investigated. At first, the change of the EMF on the BDD electrode in the melt containing 0.5 mol% Li2O was observed when oxygen gas was introduced into the cell. Figure 10.10 shows changes of the EMF on the BDD electrode when oxygen pressure was changed from 0.5 to 0.7 atm and from 0.7 to 0.5 atm. The EMF responded immediately to the change of the oxygen pressure. This result suggested that the reaction on the BDD electrode is attributed to the O2/O2 redox couple. Figure 10.11 shows the logarithmic plot of the EMF against oxygen pressures from 0.05 to 0.7 atm. It increases with an increase of the oxygen pressure and changes linearly with logPO2 . The slope is calculated by the least-square method, and the stoichiometric number of electrons determined is 1.91, which is regarded as 2 within the experimental error to be consistent with the theoretical value in Equation (10.22). In addition, Figure 10.12 shows the logarithmic plot of the EMF against the amount of Li2O added into the melt when the oxygen pressure was kept constant at 0.5 atm. The potential shifts to the negative direction almost linearly with logXO2 as the additive amount of Li2O increases and finally shows a constant value corresponding to the saturation of Li2O in the melt. As a result, the solubility of Li2O in a LiCl-KCl eutectic melt at 773 K is determined to be 1.15 mol%. Moreover, the stoichiometric number of electrons in the electrode reaction is determined to be 2.03 by a least-square method, which is regarded as 2 within the experimental error and coincides with the theoretical value in Equation (10.22). It is concluded from these results that the reaction on the BDD electrode is the oxygen electrode reaction as expressed by Reaction (10.6).

O2 0.7 atm

2.60

O2 0.5 atm

2.58

a

O2 0.5 atm

2.58

2.56

2.56

2.54

O2 0.7 atm

2.60

0

10

20

Time / min

2.54

30

b

0

10

20

30

Time / min

Figure 10.10 Electromotive force of the BDD electrode under oxygen pressures of 0.5 and 0.7 atm in a LiCl-KCl eutectic melt containing 0.5 mol% Li2O at 773 K.

198

Molten Salts Chemistry

Figure 10.11 Logarithmic plot of the EMF against oxygen pressures from 0.05 to 0.7 atm.

2.62

EMF / V vs Li+/Li

2.60 2.58 2.56 2.303RT

2.54

2F

2.52 0.1

1.0 Po21/2/ atm1/2

Figure 10.12 Logarithmic plot of the EMF against the amount of Li2O added into the melt when the oxygen pressure was kept constant at 0.5 atm. EMF / V vs Li+/Li

2.65

2.60

2.303RT

2.55 0.1

2F

1.0 10.0 Amount of Li2O added / mol%

The standard formal potential of O2/O2 is determined to be 2.424  0.003 V vs Liþ/Li in a LiCl-KCl eutectic melt at 773 K. The error value denotes 95% confidence. In the same manner, the standard formal potential of O2/O2 was measured from 673 to 803 K. The result is shown in Figure 10.13. It decreases almost linearly with temperature elevation, and the temperature dependence is given by the following equation using a least-square method: ¼ 2:995ð0:001Þ  7:38ð0:14Þ  104 T=V vs Liþ =Li E0¢ O2 =O2

(10.24)

where the value at 723 K is 2.463  0.001 V vs Liþ/Li. This result indicates that the previous study by Kanzaki and Takahashi [15], in which the standard electrode potential was determined by the measurement of EMF using a platinum electrode, was more precise than the result obtained from the solubility of refractory metal oxides in the melt reported by Masuko and colleagues [19], although the influence of a mixed potential was not avoided completely on a platinum electrode as described in Section 10.1. It is also suggested that the result estimated from the Li2O solubility [29] is almost precise. In the same manner, temperature dependence of the standard formal potential of O2/O2 in LiCl-KCl mixtures is shown in Figure 10.14: LiCl:KCl ¼ 65:35, 70:30, and 75:25 mol% containing 0.5 mol% Li2O. The result in the eutectic melt is also included as a reference. In each melt, the standard formal

Boron-Doped Diamond Electrodes in Molten Chloride Systems

Figure 10.13 Temperature dependence of the standard formal potential of O2/O2 in a LiCl-KCl eutectic melt.

2.55

EO02' /O2- / V vs Li+/Li

199

2.50

2.45

2.40

2.35 650

700

800

750

850

Temperature / K

Figure 10.14 Temperature dependence of the standard formal potential of O2/O2 vs the standard formal potential of Liþ/Li in molten LiCl-KCl mixtures; LiCl:KCl ¼ (a) 58.5:41.5 (eutectic), (b) 65:35, (c) 70:30, and (d) 75:25 mol% containing 0.5 mol% Li2O. (Error values denote 95% confidence intervals.)

2.48

2.44

(d)

2.42 2.40

(c)

2.38

(b)

'

0 0+ EO 2/O2- / V vs E Li /Li

2.46

2.36

(a)

2.34 2.32 650

700

750

800

850

Temperature / K

Table 10.2 Standard Formal Potential of O2/O2 Against Standard Potential of Liþ/Li in Molten LiCl-KCl Mixtures; LiCl:KCl ¼ 58.5:41.5 (Eutectic), 65:35, 70:30, and 75:25 mol% Containing 0.5 mol% Li2O(Error Values Denote 95% Confidence Intervals) 0

E0O2 =O2 ¼ a  bT vs E0Liþ =Li LiCl:KCl /mol%

Temperature/K

a

b

58.5:41.5

673–803

2.972  0.001

7.84  0.13

65:35

723–803

2.976  0.001

7.52  0.27

70:30

723–803

2.987  0.001

7.38  0.29

75:25

773–803

2.995  0.001

7.10  0.47

potential of O2/O2 decreases almost linearly with the elevation of temperature. The temperature dependences of the standard formal potentials are summarized in Table 10.2. In addition, the potential is more positive in the melt that has a larger content of LiCl. Here, these potentials are referenced to the standard potential of the Liþ/Li redox couple regarding the activity of Liþ as follows. The potential of Liþ/Li is expressed as

200

Molten Salts Chemistry

ELiþ =Li ¼ E0Liþ =Li þ

RT lnaLiþ F

(10.25)

where E0Liþ =Li is the standard potential of Liþ/Li and aLiþ is the activity of Liþ. Since ELiþ =Li was defined to be zero in each melt, E0Liþ =Li is expressed as E0Liþ =Li ¼ 

RT lnðgLiþ XLiþ Þ F

(10.26)

where gLiþ is the activity coefficient and XLiþ is the cation fraction of Liþ in the melt. The standard potential of Liþ/Li is calculated using the following relation between the activity coefficient of Liþ and the cation fraction in molten LiCl-KCl [30]: RT lngLiþ ¼ 13390ð1  XLiþ Þ2

(10.27) 0

Thus, the standard formal potential of O2/O2, E0O =O2 , against the potential of Liþ/Li in 2 each melt obtained in the present study is calculated to be the normalized value against þ 0 the standard potential of Li /Li, ELiþ =Li , from Equation (10.26):   ¢ ¢ E0O2 =O2 vs E0Liþ =Li ¼ E0O2 =O2  E0Liþ =Li

(10.28)

10.3.2 Change of Gibbs Free Energy, Entropy, Enthalpy The standard Gibbs free energy change for Reaction (10.29) is given by Equation (10.30): 1 2Li ðlÞ þ O2 ðgÞ ¼ 2Liþ þ O2 2

(10.29)

 0  0 DG0 ðLi2 O in LiCl  KClÞ ¼ 2F E0O2 =O2 E0Liþ =Li

(10.30)

The Gibbs free energy change increases almost linearly with the elevation of temperature as shown in Figure 10.15. In addition, it decreases with an increase of the LiCl content in the

-450

(a) -455 DG0' / kJ mol-1

Figure 10.15 Temperature dependence of the standard Gibbs free energy change for the reaction 2Li þ 1/2O2 ¼ 2Liþ þ O2 in molten LiCl-KCl mixtures; LiCl:KCl ¼ (a) 58.5:41.5 (eutectic), (b) 65:35, (c) 70:30, and (d) 75:25 mol% containing 0.5 mol% Li2O. (Error values denote 95% confidence intervals.)

(b) -460

(c) -465

(d) -470 -475 650

700

750

800

Temperature / K

850

Boron-Doped Diamond Electrodes in Molten Chloride Systems

201

melt: 456.4  0.5, 462.1  0.8, 466.1  0.6, and 472.1  0.5 kJ mol1, respectively, at 773 K in each melt, indicating that Reaction (10.29) is more favorable to proceed to the right direction in the melt with a larger content of LiCl. The standard formal entropy change, DS0¢ (29) and the standard formal enthalpy change, DH0¢ (10.29), are also calculated according to Equations (31) and (32): 

0

DS0 ðLi2 O in LiCl  KClÞ ¼  0

 0 @DG0 ðLi2 O in LiCl  KClÞ @T p

(10.31)

0

DH 0 ðLi2 O in LiCl  KClÞ ¼ DG0 ðLi2 O in LiCl  KClÞ 0 þ TDS0 ðLi2 O in LiCl  KClÞ

(10.32)

The obtained standard formal entropy and enthalpy are summarized in Table 10.3. DS0¢ (10.29) and DH0¢ (10.29) are practically constant in these temperature ranges. These values are plotted against the cation fraction of Liþ in Figures 10.16 and 10.17. DS0¢ (10.29) increases and DH 0¢ (10.29) decreases with an increase of the cation fraction of Liþ. DS0¢ (10.29) and DH 0¢ (10.29) in a LiCl single salt (XLiþ ¼ 1) are estimated to be 115  6 J K1 mol1 and 584.8  0.1 kJ mol1, respectively, by the extrapolation

Table 10.3 Standard Formal Entropy and Enthalpy of Formation of Li2O in Molten LiCl-KCl Mixtures; LiCl:KCl ¼ 58.5:41.5 (Eutectic), 65:35, 70:30, and 75:25 mol% Containing 0.5 mol% Li2O(Error Values Denote 95% Confidence Intervals) 0

0

LiCl:KCl

T/K

DSf 0 (Li2O in LiCl-KCl) /J K1 mol1

58.5:41.5

673–803

151  3

573.5  0.1

65:35

723–803

145  5

574.2  0.1

70:30

723–803

142  6

576.4  0.2

75:25

773–803

137  8

578.0  0.1

Figure 10.16 Relationship between the cation fraction of Liþ and the standard formal entropy change for the reaction 2Li þ 1/2O2 ¼ 2Liþ þ O2 in molten LiCl-KCl mixtures; LiCl:KCl ¼ 58.5:41.5 (eutectic), 65:35, 70:30, and 75:25 mol% containing 0.5 mol% Li2O. (Error values denote 95% confidence intervals.)

-80

DS0' / J mol-1 K-1

-100 -120 -140 -160 -180 -200 0.5

DHf 0 (Li2O in LiCl-KCl) /kJ mol1

0.6

0.8

0.7

XLi

+

0.9

1.0

202

Molten Salts Chemistry -570

DH0' / kJ mol-1

Figure 10.17 Relationship between the cation fraction of Liþ and the standard formal enthalpy change for the reaction 2Li þ 1/2O2 ¼ 2Liþ þ O2 in molten LiCl-KCl mixtures; LiCl:KCl ¼ 58.5:41.5 (eutectic), 65:35, 70:30, and 75:25 mol% containing 0.5 mol% Li2O. (Error values denote 95% confidence intervals.)

-575

-580

-585 0.5

0.6

0.7

0.8

0.9

1.0

XLi+

of the obtained data. The standard formal Gibbs free energy change and the standard formal potential of O2/O2 are given by Equations (10.33) and (10.34): 0

DG0 ðLi2 O in LiCl  KClÞ ¼ 116ð6Þ  103 T  584:8ð0:1Þ=kJ mol1 ¢

E0O2 =O2 ¼ 3:031ð0:001Þ  5:96ð0:31Þ  104 T=V vs Liþ =Li

(10.33) (10.34)

At 923 K, which is the typical temperature for the reduction of uranium oxides in a pure LiCl melt, they are, respectively, 477.7  5.4 kJ mol1 and 2.474  0.028 V vs Liþ/Li. From this result, the EMF for the electrochemical reduction of UO2 expressed by Reaction (10.35) is calculated to be 0.13  0.03 V vs Liþ/Li in a LiCl melt containing 0.52 mol% Li2O: UO2 ðsÞ þ 4e ¼ UðsÞ þ 2O2

(10.35)

This value is estimated from the standard formal potential of O2/O2 for Reaction (10.36) obtained in the present study and the theoretical standard potential of U4þ/U calculated from the standard Gibbs free energy of formation of solid UO2 for Reaction (10.37)[28], corresponding to the potential of the redox couple of O2/O2 estimated from the polarization curve in the previous report by Sakamura and colleagues [2]. 1 O2 þ 2e ¼ O2 2

(10.36)

U ðsÞ þ O2 ðgÞ ¼ UO2 ðsÞ

(10.37)

10.3.3 Activity Coefficient of Oxide Ion The obtained formal thermodynamic quantities are compared with literature data for the formation of liquid Li2O from liquid Li and gaseous O2 [28]: 1 2Li ðlÞ þ O2 ðgÞ ¼ Li2 O ðlÞ 2

(10.38)

Boron-Doped Diamond Electrodes in Molten Chloride Systems

Figure 10.18 Activity coefficients of O2 ion in molten LiCl-KCl mixtures; LiCl:KCl ¼ 58.5:41.5 (eutectic), 65:35, 70:30, and 75:25 mol% containing 0.5 mol% Li2O. (Error values denote 95% confidence intervals.)

3.5 3.0

(a)

lngO2–

2.5 2.0

(b)

1.5

(c)

203

1.0 0.5

(d)

0.0 1.20

1.25 1.30

1.35 1.40

1.45

1.50

1000T -1 / K -1

1 2Li ðlÞ þ O2 ðgÞ ¼ 2Liþ þ O2 (10.39) 2 When the standard state is defined to be pure Li2O liquid, the standard Gibbs free energy of formation of Li2O (l), DGf 0 (liquid Li2O), corresponds to the standard Gibbs free energy change of Reaction (10.38), DG0 (10.38). Therefore, DGf 0 (liquid Li2O) is expressed by Equation (10.40) using the relation given by Equation (10.38): DG0f ðliquid Li2 OÞ ¼ DG0 ðLi2 O in LiCl  KClÞ

1 =2 P O2 0 ¼ DG ðLi2 O in LiCl  KClÞ  RT ln XO2 0

1 =2 gO 2 DG ðLi2 O in LiCl  KClÞ ¼ DG ðLi2 O in LiCl  KClÞ  RT ln gO2

00

0

(10.40)

(10.41)

where the activity coefficient of oxygen gas can be regarded as unity because only oxygen gas is used as a gas phase and does not depend on oxygen pressures. Therefore, gO2 in the dilute O2 region is estimated as shown in Figure 10.18. The activity coefficient decreases with the increase of LiCl content in the melt and depends more strongly on the melt composition than on the temperature. It is considered that the attractive force between Liþ and O2 ion is stronger than that between Kþ and O2 affecting the activity of oxide ions in the melt. This tendency agrees with the thermodynamic stability of the corresponding oxides, that is, Li2O > K2O. In a LiCl melt, which is a typical system for the reduction of uranium oxides, the activity coefficient of O2 is estimated to be 0.07 0.05. The result is explained by the repulsive force between O2 and Cl, which is weaker than that between the two oxide ions.

10.4

Conclusions

Quantitative analysis of the evolved gas formed on a boron-doped diamond anode in a LiClKCl melt containing Li2O shows that the oxidation of oxide ions occurs at the potential range of 2.5 to 3.5 V (vs Liþ/Li) to produce pure diatomic oxygen quantitatively according to the following reaction: 1 O2 ! O2 þ 2e 2

(10.42)

204

Molten Salts Chemistry

The loss of the electrode by the electrolysis is negligibly small. Thus the diamond electrode has been found unexpectedly to be used as insoluble anode material for the electrochemical processes in molten salt systems. The stability of the BDD electrode as an oxygen evolution electrode was evaluated in molten LiCl:KCl (58.5:41.5 mol%), LiCl-KCl (75:25 mol%), LiCl-CaCl2 (64:36 mol%), and LiCl-NaCl-CaCl2 (52.3:13.5:34.2 mol%) at 773 K. In molten LiCl-KCl systems, the BDD electrode is stable and its stability does not depend on the concentration of oxide ion and the melt composition. Thus, the BDD electrode has the potential to be employed for the inert anode in molten LiCl-KCl systems with any compositions at 773 K. In molten LiCl-CaCl2 and LiCl-NaCl-CaCl2, however, the BDD electrode is less stable than in molten LiCl-KCl systems, which is conceivably due to the low wettability of the melts containing CaCl2. As a result in this chapter, it is suggested that molten LiCl-KCl (75:25) is one of the best candidate electrolytes for the reduction processes of metal oxides when combined with the BDD counterelectrode, taking into account factors such as high stability of the electrode and the large solubility of oxides at relatively low temperatures. Standard formal potentials of O2/O2 were measured using the BDD electrode as an inert anode, and activity coefficients of oxide ion were determined in molten LiCl-KCl mixtures: LiCl:KCl ¼ 58.5:41.5 (eutectic), 65:35, 70:30, and 75:25 mol% containing 0.5 mol% Li2O at 673 to 803 K. The standard formal potential of O2/O2 is more positive in the melt that has a larger content of LiCl, and the activity coefficient of oxide ions decreases with the increase of the LiCl content in the melt. From these results, in a LiCl melt at 923 K, which is a typical system for the reduction of uranium oxides, the standard formal potential and the activity coefficient are estimated by extrapolation to be E0¢ ¼ 2:474  0:028 V vs Li þ =Li O2 =O2 and 0.07  0.05, respectively. Thus, this study has given available data for thermodynamic consideration on chemical and electrochemical processes involving oxygen gas and/or oxide ion such as reduction processes of metal oxides in molten LiCl-KCl systems.

References [1] Chen, G.Z., Fray, D.J., Farthing, T.W. (2000). Direct electrochemical reduction of titanium dioxide to titanium in molten calcium chloride. Nature (London), 407, 361–367. [2] Sakamura, Y., Kurata, M., Inoue, T. (2006). Electrochemical reduction of UO2 in molten CaCl2 or LiCl. J. Electrochem. Soc., 153, D31–D39. [3] Nohira, T., Yasuda, K., Ito, Y. (2003). Pinpoint and bulk electrochemical reduction of insulating silicon dioxide to silicon. Nature. Material, 2, 397–401. [4] Yasuda, K., Nohira, T., Amezawa, K., Ogata, Y.H., Ito, Y. (2005). Direct electrolytic reduction of solid silicon dioxide in molten LiCl-KCl-CaCl2 at 773 K. J. Electrochem. Soc., 152, D208–D212. [5] Goto, T., Tada, M., Ito, Y. (1994). Electrochemical surface nitriding of titanium in molten salt system. Electrochim. Acta, 39, 1107–1113. [6] Goto, T., Tada, M., Ito, Y. (1997). Electrochemical behavior of nitride ions in a molten chloride system. J. Electrochem. Soc., 144, 2271–2276. [7] Goto, T., Obata, R., Ito, Y. (2000). Electrochemical formation of iron nitride film in a molten LiClKCl-Li3N system. Electrochim. Acta, 45, 3367–3373. [8] Tsujimura, H., Goto, T., Ito, Y. (2002). Electrochemical formation and control of chromium nitride films in molten LiCl-KCl-Li3N systems. Electrochim. Acta, 47, 2725–2731. [9] Toyoura, K., Tsujimura, H., Goto, T., Hachiya, K., Hagiwara, R., Ito, Y. (2005). Optical properties of zinc nitride formed by molten salt electrochemical process. Thin Solid Films, 492, 88–92. [10] Konishi, H., Nohira, T., Ito, Y. (2003). Kinetics of DyNi2 film growth by electrochemical implantation. Electrochim. Acta, 48, 563–568. [11] Murakami, T., Nishikiori, T., Nohira, T., Ito, Y. (2003). Electrolytic synthesis of ammonia in molten salts under atmospheric pressure. J. Am. Chem. Soc., 125, 334–335.

Boron-Doped Diamond Electrodes in Molten Chloride Systems

205

[12] Murakami, T., Nohira, T., Goto, T., Ogata, Y.H., Ito, Y. (2005). Electrolytic ammonia synthesis from water and nitrogen gas in molten salt under atmospheric pressure. Electrochim. Acta, 50, 5423–5426. [13] Masset, P., Guidotti, R. (2007). A. Thermal activated (thermal) battery technology. II. Molten salt electrolytes. J. Power Sources, 164, 397–414. [14] Wrench, N.S., Inman, D. (1968). The oxygen electrode in molten salts: Potentiometric measurements. J. Electroanal. Chem., 17, 319–325. [15] Kanzaki, Y., Takahashi, M. (1975). The oxygen electrode in fused lithium chloride-potassium chloride eutectic containing oxide ion. J. Electroanal. Chem., 58, 339–348. [16] Littlewood, R. (1962). Diagrammatic representation of the thermodynamics of metal-fused chloride systems. J. Electrochem. Soc., 109, 525–533. [17] Cherginets, V.L., Demirskaya, O.V., Rebrova, T.P. (2004). Potentiometric study of acid-base equilibria in the KCl þ LiCl eutectic melt at temperatures in the range (873 to 1073) K. J. Chem. Thermodynamics, 36, 115–120. [18] Cho, S.H., Zhang, J.S., Shin, Y.J., Park, S.W., Park, H.S. (2004). Corrosion behavior of Fe-Ni-Cr alloys in the molten salt of LiCl-Li2O at high temperature. J. Nuclear Mater., 325, 13–17. [19] Masuko, N., Okada, M., Hisamatu, Y. (1963). E-pO2 diagram in fused chloride solution and its application to UO2 electrolysis (in Japanese). Yoyuen (Fused Salts) 6, 570–582. [20] Alcock, C.B., Hopper, G.W. (1960). Thermodynamics of the gaseous oxides of the platinumgroup metals. Proc. Roy. Soc. A, 254, 551–561. [21] Beer, H.B. (1980). The invention and industrial development of metal anodes. J. Electrochem. Soc., 127, 303C–307C. [22] Goto, T., Araki, Y., Hagiwara, R. (2006). Oxygen gas evolution on the boron-doped diamond electrode in molten chloride system. Electrochem. Solid State Lett., 9, D5–D7. [23] Kado, Y., Goto, T., Hagiwara, R. (2008). Electrochemical behavior of oxide ion in a LiCl-NaClCaCl2 eutectic melt. J. Electrochem. Soc., 155, E85–E89. [24] Sun, Q., Alam, M. (1992). Relative oxidation behavior of chemical vapor deposited and type II a natural diamonds. J. Electrochem. Soc., 139, 933–936. [25] Alam, M., Sun, Q. (1993). Improvement in the thermal oxidation resistance of chemical vapourdeposited diamond films. J. Mater. Sci. Lett., 12, 1389–1391. [26] John, P., Polwart, N., Troupe, C.E., Wilsonb, J.I.B. (2002). The oxidation of (100) textured diamond. Diamond Relat. Mater., 11, 861–866. [27] Li, H.S., Qi, Y.X., Gong, J.H., Wang, M., Li, M.S. (2009). High-pressure synthesis and characterization of thermal-stable boron-doped diamond single crystals. Int. J. Refract. Met. Hard Mater., 27, 564–570. [28] Landolt-Bo¨rnstein, (1999). Thermodynamic properties of inorganic material, scientific group Theremodata Europe (SGTE). Springer-Verlag, Berlin-Heidelberg. [29] Kado, Y., Goto, T., Hagiwara, R. (2008). Dissolution behavior of lithium oxide in molten LiClKCl systems. J. Chem. Eng. Data, 53, 2816–2819. [30] Lumsden, J. (1996). Thermodynamics of Molten Salt Mixtures, p. 47. Academic Press, London.