Investigation on the electrochemical behavior of neodymium chloride at W, Al and Cd electrodes in molten LiCl-KCl eutectic

Accepted Manuscript Title: Investigation on the electrochemical behavior of neodymium chloride at W, Al and Cd electrodes in molten LiCl-KCl eutectic Author: S. Vandarkuzhali Manish Chandra Suddhasattwa Ghosh Nibedita Samanta S. Nedumaran B. Prabhakara Reddy K. Nagarajan PII: DOI: Reference:

S0013-4686(14)01703-4 http://dx.doi.org/doi:10.1016/j.electacta.2014.08.069 EA 23267

To appear in:

Electrochimica Acta

Received date: Revised date: Accepted date:

10-6-2014 8-8-2014 21-8-2014

Please cite this article as: S. Vandarkuzhali, M. Chandra, S. Ghosh, N. Samanta, S. Nedumaran, B. Prabhakara Reddy, K. Nagarajan, Investigation on the electrochemical behavior of neodymium chloride at W, Al and Cd electrodes in molten LiCl-KCl eutectic, Electrochimica Acta (2014), http://dx.doi.org/10.1016/j.electacta.2014.08.069 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Investigation on the electrochemical behavior of neodymium chloride at W, Al and Cd electrodes in molten LiCl-KCl eutectic S. Vandarkuzhali, Manish Chandra, Suddhasattwa Ghosh, Nibedita Samanta, S.

ip t

Nedumaran, B. Prabhakara Reddy and K. Nagarajan*

Tamilnadu, India-603102

us

*Corresponding author

an

K.Nagarajan, Email: [email protected] Phone: +91-44-27480500 (Ext: 24289)

M

Fax: +91-44-27480065

cr

Chemistry Group, Indira Gandhi Centre for Atomic Research, Kalpakkam,

d

Abstract

te

Electrochemical behavior of neodymium (III) ion was studied in LiCl-KCl eutectic melt in

Ac ce p

the temperature range 723-798 K on inert tungsten electrode using various transient electrochemical techniques such as cyclic voltammetry, convolution voltammetry, chronopotentiometry and square wave voltammetry. The reduction of Nd(III) ion to Nd metal on tungsten electrode takes place in two steps- Nd(III)/ Nd(II) and Nd(II)/ Nd(0). The diffusion coefficient of Nd(III) and Nd (II) ions were determined. Reduction of Nd(III) to Nd(II) showed reversible electrode behavior and that for Nd(II) to Nd metal followed the quasi-reversible behavior. Heterogeneous rate constant for the reduction, Nd(II)/Nd(0) was estimated from the convoluted voltammograms. The apparent standard electrode * * * E Nd potentials, E Nd and E Nd were estimated from the cyclic ( III ) / Nd ( II ) , ( II ) / Nd ( 0 ) ( III ) / Nd ( 0 )

voltammograms.

1 Page 1 of 60

The electrode behavior of Nd(III) ion on solid aluminium electrode and liquid cadmium electrode was studied by cyclic voltammetry. Under-potential reduction of Nd(III) ion takes place on Al and Cd cathodes in a single step with three electron transfer. The

ip t

* * apparent standard electrode potentials, E Nd ( III ) / Nd ( Al ) and E Nd ( III ) / Nd ( Cd ) were estimated for

different temperatures in the range 698-773 K. The formation of intermetallics, Al11Nd3 and

cr

Cd11Nd were studied from open circuit potential measurement on Al and Cd film electrode

us

respectively. The Gibbs energy formation for these intermetallics was evaluated. The activity of neodymium in Al/Cd, the excess Gibbs energy and the activity coefficient of neodymium

an

in Al/Cd were estimated from the open circuit potential measurement. SEM-EDX analysis and the XRD pattern of the electro-deposit revealed the formation of Al11Nd3 and Cd11Nd on

M

Al and Cd cathodes respectively.

Ac ce p

te

circuit chronopotentiometry

d

Keywords: Molten LiCl-KCl; Neodymium; Cyclic voltammetry; Alloy formation; Open

2 Page 2 of 60

Investigation on the electrochemical behavior of neodymium chloride at W, Al and Cd electrodes in molten LiCl-KCl eutectic S. Vandarkuzhali, Manish Chandra, Suddhasattwa Ghosh, Nibedita Samanta, S.

ip t

Nedumaran, B. Prabhakara Reddy and K. Nagarajan*

Tamilnadu, India-603102

us

*Corresponding author

an

K.Nagarajan, Email: [email protected] Phone: +91-44-27480500 (Ext: 24289)

M

Fax: +91-44-27480065

cr

Chemistry Group, Indira Gandhi Centre for Atomic Research, Kalpakkam,

d

1. Introduction

te

Pyrochemical process is now considered as one of the most promising options in

Ac ce p

nuclear fuel cycle for advanced fuel [1-2]. The focus of research institutes worldwide is to establish a method based on pyrochemical process for a more compact fuel cycle, with minimization of nuclear waste [3-4]. In the electrorefining method for spent metallic fuel, the lanthanides and actinides are dissolved into molten salt by anodic dissolution. The actinides are selectively deposited at the cathode (steel/ cadmium) due to the difference in the redox potentials of the elements. The fission products remain in the anode and in the electrolyte [57]. Lanthanides are the major fission products in the spent fuel and are the most difficult element to separate from actinides due to similar chemical properties [8]. In this study we have investigated the electrochemical behavior of neodymium chloride in molten LiCl-KCl since the yield of Nd is significant in the spent fuel [9].

3 Page 3 of 60

Electrochemical transient techniques are powerful method to determine the thermodynamic properties in molten salt. In recent times, electrochemical transient techniques are widely used to estimate the formal standard potential, Gibbs energy formation

ip t

of chlorides, their activity coefficient etc [10-14]. Cyclic voltammetry and open circuit potentiometry are quick and reliable method to study the formation of alloys [15-17]. The

cr

Gibbs energy formation of the intermetallics and the activity of metals in solvents estimated from transient electrochemical technique are in very good agreement with those estimated

were

used

to

evaluate

the

apparent

standard

electrode

potentials

an

techniques

us

from equilibrium measurements and calorimetric method. The transient electrochemical

* * * ( E Nd ( III ) / Nd ( II ) , E Nd ( II ) / Nd ( 0 ) and E Nd ( III ) / Nd ( 0 ) ), Gibbs energy formation of NdCl3 and the

M

thermodynamic properties of Nd-Al and Nd-Cd systems.

Quite a good number of studies have been carried out on the redox behavior of NdCl3

d

in various molten salt media. Masset et al. had studied the electrochemical reduction of

te

lanthanum and neodymium chlorides at a tungsten working electrode in LiCl-KCl melt using

Ac ce p

cyclic voltammetry and chronopotentiometry [12]. They had derived the thermochemical data from their measurements. Castrillojo et al. had studied the stable oxidation states of La, Ce, Nd and Y and derived the standard potentials of the redox couples in LiCl-KCl and CaCl2NaCl molten salts [11]. Fukusawa et al. and Yamana et al. had studied the electrochemical and spectrophotometric behavior of Nd in chloride media [18, 19]. Yamamura et al. had studied the electrochemical behavior of the chlorides of La, Ce, Nd, Sm and Dy in LiCl-KCl and NaCl-KCl media [20]. Novoselova et al. had studied the electrochemical behavior of NdCl3 in LiCl-KCl-CsCl eutectic in the temperature range 573-943 K by cyclic voltammetry and linear sweep voltammetry. They had estimated the apparent standard redox potential of the Nd(III)/Nd(II) couple by direct potentiometric method [21]. Taxil et al. have reported the reduction behavior of lanthanide ions (Ce, Gd, Sm and Nd) on molybdenum electrode in LiF4 Page 4 of 60

CaF2 melt [22-23]. All the authors have reported the reduction of Nd(III) ion to Nd metal to take place in two steps with the formation of Nd(II) ion in the intermediate step in chloride media. However, there is large difference in the values of the standard potentials reported.

ip t

Also the electrode kinetics for the process is not much studied. In recent days, there is interest in exploring the suitability of aluminium cathode for

cr

efficient recovery of actinides over lanthanides in the electrorefining process [24-28].

us

Investigations are being carried out on elements pertaining to the electrorefining process, to understand the electrode reaction, kinetics, derive the standard potentials and activities of

an

these elements on Cd and Al electrode in order to evaluate the efficiency of these electrodes [29-33]. In this paper we have reported a detailed elucidation on the redox behavior of NdCl3

M

on W inert electrode and its electrode kinetics using electrochemical transient techniques. We have also carried out a detailed study on the electrode reaction at Al and Cd electrodes using

d

LiCl-KCl-NdCl3 melt and estimated the apparent standard potentials of NdCl3 in these

te

electrodes. The thermodynamic properties of the Nd-Al and Nd-Cd system were estimated

Ac ce p

from the open circuit potential measurement. 2. Experimental 2.1. Chemicals

Anhydrous lithium chloride (AR grade, M/s. Chempure Private Ltd, India), anhydrous

potassium chloride (AR grade, M/s. Ranbaxy Fine Chemicals, India), NdCl3 (Alfa Aesar 99.99%), high purity Al wire/ sheet (Alfa Aesar 99.99%) and anhydrous cadmium chloride (M/s. Merck, Germany 99%) were used for the studies. 2.2. Preparation of electrolyte

5 Page 5 of 60

LiCl–KCl eutectic mixture was pre-dried under vacuum for 120 h at 420 K. The salt mixture was purified by passing chlorine gas at 673 K for 20 minutes and the salt was melted under chlorine atmosphere in a quartz set up discussed elsewhere [34]. Addition of NdCl3 to

ip t

the purified salt was done at 723 K in a high purity argon atmosphere glove box. The concentration of neodymium in the LiCl-KCl eutectic was measured by dissolving a sample

cr

in 0.1 M HNO3 and analyzing by ICP-AES.

us

2.3. Electrochemical apparatus and electrodes

A three electrode assembly was used for the electrochemical measurements and the

an

cell is described elsewhere [35]. Different working electrodes were used for the study: (i) 1.5 mm diameter tungsten wire; (ii) 1.5 mm diameter Al wire; (iii) Cd pool electrode and (iv) Cd

M

film electrode. The W and Al working electrodes were sheathed with an alumina sleeve exposing 40 mm of the wire and the area of working electrode was calculated from the depth

d

of immersion. The cadmium pool electrode was prepared by placing cadmium granules in an

te

alumina crucible which was in turn placed in an SS 430 holder assembly. The crucible was

Ac ce p

immersed in molten LiCl-KCl to melt the cadmium. A 1mm diameter tungsten wire covered with alumina sheath was immersed in the cadmium melt through the SS assembly and this served as the lead wire. The Cd film electrode were prepared in-situ by electrodepositing Cd on a W wire of 1.5 mm in diameter as working electrode from melt containing CdCl2. The electrodeposition was carried out at about -1 V vs. Ag/Ag+ reference electrode for a period of 300 s.

A tantalum sheet of 1mm thickness was used as the counter electrode. The reference electrode consisting of a silver wire (2 mm dia) dipped in 0.8 wt % AgCl-LiCl-KCl mixture contained in a pyrex glass tube was used for all the measurements. All the potentials mentioned in this paper were measured with respect to this reference electrode. Preparation 6 Page 6 of 60

and assembling of the cell was carried out in a high pure inert atmosphere glovebox. The cell was taken out of the glove box and kept in a kanthal wire wound resistance furnace which was used for heating the cell. Argon gas purified using calcium getter at 823 K was passed

ip t

through the cell. The temperature of the cell was controlled by a PID controller with an accuracy ±1 K. The temperature of the melt was monitored by a K-type thermocouple kept in

cr

thermo well and immersed in the electrolyte. Cyclic voltammograms, square wave voltammograms, chronopotentiograms and open circuit potential measurements were

us

obtained using AUTOLAB PGSTAT30 from M/s Eco Chemie, Netherlands equipped with IF

an

030 interface. The voltammograms were processed using GPES 4.9 software. 3. Results and Discussion

M

3.1. Reduction behavior of NdCl3 in LiCl-KCl on inert W electrode

d

3.1.1. Analysis of the cyclic voltammograms

te

Cyclic voltammograms were recorded using tungsten as working electrode in molten

Ac ce p

LiCl-KCl mixture containing NdCl3, (6.66 x 10-5 mol cm-3) in the temperature range 723 – 798 K. Cyclic voltammograms at 723 K for different scan rates are shown in Fig.1. It may be seen that there are two peaks in the cathodic cycle and two corresponding peaks in the anodic cycle. Castrilljo et al., Masset et al. and Fukusawa et al. have also made similar observation in their studies involving the redox behavior of NdCl3 in chloride media at inert electrode [11, 12, and 18]. They had deduced that the reduction of Nd(III) ion to Nd metal takes place in two steps with the formation of Nd(II) ion in the intermediate step. We have carried out a detailed analysis of the voltammograms to elucidate the redox behavior of NdCl3 in LiClKCl. INSERT: FIGURE 1

7 Page 7 of 60

Analysis of peak I Cyclic voltammograms were obtained by switching the potential before peak IIc appeared as shown in Fig.2. It may be observed that peak Ia is the corresponding anodic peak

ip t

for peak Ic. The shape of the voltammogram for the redox couple Ic and Ia resemble closely to reversible soluble-soluble system. Graph (a) of Fig.3 shows the plot of the cathodic peak

cr

potentials, EpIc, as a function of logarithm of the scan rates at 723 and 798 K. The peak

us

potentials do not vary appreciably in the scan rate range 10 - 200 mVs-1 indicating that the electrode process is facile. For a reversible electrochemical process involving soluble-soluble

an

species, the anodic peak potential, cathodic peak potential and the number of electron transferred are related as [36, 37] RT nF

M

E PA  E PC  2.22

(1)

d

Value of n estimated using the above equation is close to 1 suggesting that peak Ic and

te

Ia correspond to the redox couple Nd(III)/Nd(II). Novoselova et al. had observed that the

Ac ce p

reduction of NdCl3 in LiCl-KCl-CsCl melt too followed the two step mechanism [21]. However, they did not observe a clear peak for the first redox couple. From the analysis of linear sweep voltammograms at very low scan rates, they had obtained the value of n as unity for the reduction of Nd(III)/Nd(II) using the Heyrovsky-Ilkovic equation applicable for a reversible electrode process.

INSERT: FIGURE 2 INSERT: FIGURE 3

Analysis of peak II

8 Page 8 of 60

The shape of peak II in Fig.1 during the cathodic and anodic cyclic is suggestive of soluble-insoluble couple, Nd(II)/Nd(0). Eq. (2) gives the relationship among the peak potential, the half peak potential and the number of electrons transferred for a reversible

ip t

electrochemical reaction involving soluble-insoluble species. The value of n estimated from the cathodic peak IIc using Eq. (2) was not about 2 for all scan rates which is expected for the

RT nF

us

E p  E p 2   0.77

cr

reduction, Nd(II)/Nd(0) [38, 39]. The value of n was about 1.7 for scan range 10 – 30 mVs-1.

(2)

an

It was found that the peak width, Ep – Ep/2 increased with increase in scan rate. Also the ratio of anodic peak current (ipIIa) to cathodic peak current (ipIIc) is greater than 1 for all

M

scan rates. Graph (b) of Fig.3 shows the plot of the cathodic peak potentials, EpIIc, as a function of logarithm of the scan rates at 723 and 798 K. It was observed that the peak

d

potential varies with increasing scan rate. These observations suggest that the reduction of

te

Nd(II) ion to Nd metal on W electrode does not involve a perfect reversible exchange with

Ac ce p

rate of reduction controlled by mass transfer for all scan rates. The reduction of Nd(II)/Nd(0) at the electrode was influenced by the charge transfer process also. Eq. (3) which is applicable for an irreversible electrochemical process was used for the peak analysis [40].

E p  E p 2   1.857

RT n F

(3)

Where ‘α’ is the charge transfer coefficient. It was observed that the values of ‘αnα’

obtained from Eq. (3) at various scan rates and at different temperatures ranged between 2.4 to 3.6 and did not have any significant meaning for an electrode reaction with two electron transfer. This fact suggests that the reduction of Nd(II)/Nd(0) is not controlled by the charge transfer process alone. Hence we had studied the reduction behavior of Nd(II)/Nd(0) using convolution technique. 9 Page 9 of 60

Semi-integral of the cyclic voltammetry data gives the convoluted curves, as given in Eq. (4) [41, 42]. t

i ( )

 t   d

(4)

0

ip t

 d 1 2 I (t )  1 m(t )    1 2   dt 

where i(τ) is the current measured at the time τ of the I(t) transients. For an electrode

cr

process, where the rate transfer is limited by diffusion, a potential plateau, m*, occurs in the

us

region of cathodic peak potential. The forward and backward waves superimpose for a reversible electrode process. Curve c of Fig.4 shows the convoluted curve for the

an

voltammogram obtained by switching the potential before appearance of peak IIc. It may be observed in Curve c of Fig. 4 that the forward and backward waves more or less superimpose

M

indicating that the electrode process for Nd(III)/Nd(II) may be considered as reversible [43]. Curve b of Fig.4 shows the convoluted curve for the cyclic voltammogram involving both the

d

redox couples. A potential plateau appears about the cathodic peak potential for Nd(II)/Nd(0).

te

However, there is some hysteresis between the cathodic and the anodic cycles. The hysteresis

Ac ce p

could be attributed to the non-reversibility of the electrode process and indicates that the rate of the reduction is not limited by diffusion of Nd(II) ions alone. INSERT: FIGURE 4

3.1.2. Kinetic study of the electrode reaction Nd(II)/Nd(0) – reversibility study The behavior of Nd(II)/Nd(0) on W cathode is similar to our observation for

La(III)/La(0) in our previous study [44]. Though there was no indication of large overpotential from the cyclic voltammograms and convoluted curves such as cross over in the cathodic branch, the cathodic peak potential was found to shift cathodically with increasing scan rates and the peak potential difference Ep − Ep/2 increased with increasing scan rates for the reduction of Nd(II)/ Nd(0). We had used the convoluted curve to study the electrode 10 Page 10 of 60

kinetics and evaluate the kinetic parameter, the heterogeneous rate constant, ks for this reaction. The following relationship is valid in the case of a quasi-reversible exchange with



(5)



( m *  m ) D 1 2  nFA exp nF RT E  E 0  I

us

B

RT RT log k s  2.3 log B n F n F

cr

E  E 0  2.3

ip t

formation of an insoluble product [45, 46]

(6)

an

where ks and α are the charge-transfer rate constant and the transfer coefficient, respectively, E0 is the standard potential of the Nd(II)/Nd(0) system, A is the surface area of

M

the electrode, m is the convoluted current and m* is its limiting value. Analysis of the convoluted data in the region corresponding to reduction of Nd(II)/Nd(0) showed that the

d

reduction followed the quasi-reversible hypothesis in the scan range 75-250 mVs-1 as per Eq.

te

(5) and (6). Plot of E vs. log B were obtained for polarization rates 75, 100, 200 and 250 mV

Ac ce p

s−1 using values of m from 10% to 90% of m* and they were found to be linear. The half wave potential, E1/2, is considered close to the standard potential, E0. The half wave potential seen in Fig. 4 was obtained graphically from the convoluted curves as described by Goto et al. [47, 44] and the value was used for E0 in Eq. (6). The transfer coefficient, α, was obtained from the slope and the charge-transfer rate constant, ks, was obtained from the intercept. The average value of ‘αnα’ was found to be 1.78 ± 0.1 and this value was used for calculation of the charge-transfer rate constant, ks. The results are shown in Table 1. However, the value of ‘αnα’ obtained from the slope of the plots of E vs. log B for scan rates 25 and 50 mVs-1 were above 2 and did not have any significance for a two electron transfer process. So we consider that the redox couple did not show quasi-reversible behavior for scan rates below 50 mVs-1.

11 Page 11 of 60

According to Matsuda and Ayabe the following criteria is applicable for an electrochemical process [40].

ip t

For Reversible (Nernstian) process:

cr

ks  15 DnF RT 1 2

ks  15 DnF RT 1 2

an

10 2 (1 ) 

us

For Quasi-reversible process:

Totally irreversible process:

M

ks  10 2 (1 ) 12 DnF RT 

d

Using the Matsuda-Ayabe criteria, it was found that ks obeyed the quasi reversible

te

hypothesis for the scan rates in the range 75 mVs-1 to 250 mVs-1 in this study. The following

Ac ce p

observations were made for 6.66 x 10-5 mol cm-3 NdCl3 in LiCl-KCl melt at 773 K at various scan rates using the value of 1.78 for ‘αnα’ and 5.01 x 10-6 cm s-1 for ks obtained from the convolution method:

1.87 x 10-6 < ks < 0.169 for scan rate 250 mVs-1 1.68 x 10-6 < ks < 0.152 for scan rate 200 mVs-1 1.19 x 10-6 < ks < 0.107 for scan rate 100 mVs-1 1.03 x 10-6 < ks < 0.092 for scan rate 75 mVs-1 Hence it may be concluded that the reduction of Nd(II) ion to Nd metal on W

electrode shows quasi-reversible behavior for polarization rates in the range 75 - 250mVs-1. 12 Page 12 of 60

INSERT: TABLE 1 3.1.3 Estimation of diffusion coefficient The cathodic peak current for the reduction of Nd(III)/Nd(II) and that for

ip t

Nd(II)/Nd(0) were plotted against square root of the scan rate. Linear dependence of peak

cr

current, ipc, on the square root of scan rate, 1/2 was observed for the scan rate in the scan range 25 ≤  ≤ 150 mV s−1 for Nd(III)/Nd(II) redox couple and in the scan range 10 ≤  ≤ 50

us

mV s−1 for Nd(II)/Nd(0) couple as seen in Fig. 5 and Fig. 6 respectively. Based on our contention that the reduction of Nd(III)/Nd(II) is a reversible electrochemical process, the

an

diffusion coefficient, of Nd(III) ion was evaluated using the Randles-Shevchik equation for a

i pc  0.446nF 

RT 1 2 AC M0 ( x ) DM1 2( x ) 1 2

(7)

d

32

M

diffusion controlled process involving soluble-soluble couple [48]

te

where ipc is the cathodic peak current corresponding to peak Ic (A), A is the electrode area (cm2), C M0 ( x ) is the concentration of Nd(III) ion in the bulk (molcm-3) and DM(x) is the

Ac ce p

diffusion coefficient of Nd(III) ion (cm2s-1),  is the scan rate (Vs-1). Fig.6 shows the plot of current densities as a function of square root of scan rate for the redox couple Nd(II)/Nd(0). It was observed that a linear dependence was not found for the entire scan range studied. Inset of Fig.6 shows the plot of current densities as a function of square root of scan rate in the range 10-50 mVs-1. Since the plot of peak current against square root of scan rate was found to be linear for the electrode reaction Nd(II)/Nd(0) at low polarization rates and also from our discussion in the previous section, it is presumed that the electrode reaction shows reversible behavior in the scan range 10 ≤  ≤ 50 mVs−1. The diffusion coefficient, of Nd(II) ion was evaluated using the Berzins-Delahay equation, Eq. (8) applicable for a diffusion controlled process involving soluble-insoluble species [49]. 13 Page 13 of 60

i pc  0.61nF 

32

RT 1 2 AC M0 ( x ) DM1 2( x ) 1 2

(8)

where ipc is the cathodic peak current corresponding to peak IIc, DM(x) is the diffusion

ip t

coefficient of Nd(II) ion and the other terms have their usual nomenclature. Since the Nd(II) ion is formed from Nd(III) ion, it is assumed that CNd(II)= CNd(III) for calculation of diffusion

[51].

cr

coefficient of Nd(II) [50]. Similar assumption was made by Glatz et al. for estimating DNd(II) Estimation of the diffusion coefficient using the melt LiCl-KCl-NdCl2 should yield

us

more accurate values for the diffusion coefficient of Nd(II) ion. Since we had not carried out any studies using LiCl-KCl-NdCl2 melt, we have adopted the procedure used by Glatz et al. ions calculated at different

an

The values of diffusion coefficient of Nd(III) and Nd(II)

temperatures are shown in Table 1. The diffusion coefficient value at 733 K reported by Glatz

M

et al. is close with that of ours [51]. Kuznetsov et al. had estimated the diffusion coefficient

d

of Eu(III) ion using LiCl-KCl-EuCl3 melt and that for Eu(II) ion using LiCl-KCl-EuCl2 melt

te

[37]. It was found that the order of the values of DEu(III) and DEu(II) obtained by Kuznetsov et al. are similar to the values of DNd(III) and DNd(II) in this study indicating that the assumption

Ac ce p

may not result in large error in the value of DNd(II) . INSERT: FIGURE 5 INSERT: FIGURE 6

From the above discussion it may be summarized that the peaks Ic and Ia correspond

to Nd(III)/Nd(II) redox couple. The cathodic peak IIc and the associated anodic peak IIa correspond to the reduction of Nd(II) ion to Nd metal in the cathodic cycle and subsequent oxidation of Nd metal to Nd(II) ion in the anodic cycle. The electrode reaction for reduction of Nd(III)/Nd(II) is reversible and that for Nd(II)/Nd(0) is quasi reversible. Yamana et al. have studied the stability of Nd(II) ion by spectroscopic method. Nd(II) ions were formed in-

14 Page 14 of 60

situ by applying a cathodic potential sufficient to reduce Nd(III) ion without forming metallic Nd [19]. They had observed the peaks for Nd(II) ion in their spectroscopic studies as long as the potential was applied and the peak intensity for Nd(II) ion gradually fell down with time

ip t

once the potential was switched off. This showed that NdCl2 is unstable and disproportionate

cr

to Nd(III) and Nd metal as given by Eq. (9).

(9)

us

3Nd(II)  2Nd(III) + Nd(0)

However, we have observed peak for formation of Nd(II) ion at about -1.85 V and its

an

subsequent oxidation to Nd(III) during the anodic cycle in the concentration range studied. This could be due to the short time scale of our transient measurement. Novoselova et al.[21]

potentiometric method

M

had estimated the apparent standard redox potential of Nd(III)/Nd(II) couple by direct in LiCl-KCl-CsCl melt containing NdCl3 and NdCl2. They had

d

reported that NdCl2 is unstable above 798 K and disproportionate to Nd(III) and Nd metal.

te

Hence they observed that the potential of the electrode stabilized at the rest potential of Mo

Ac ce p

electrode in their studies above 798 K. But they were able to obtain the equilibrium potential of the Nd(III)/Nd(II) couple in the temperature range 573-723 K. This suggests that the disproportionation reaction is not significant in LiCl-KCl-CsCl melt for temperatures below 798 K. Fukusawa et al. had studied the thermodynamic stability of the Nd(III) complex in different molten alkali chloride melts at 923 K [18]. The Nd(Cl6)3- complex was more stable in an alkali chloride mixture with larger averaged cationic radius which implies that the stability is controlled by the polarizing power of solvent cations. The stability of Nd(III) and Nd(II) complex in the study of Novoselova et al. may be due to the presence of Cs(I) ions. Apart from the disproportionation reaction, corrosive reaction due to the interaction of Nd(II) ions with the materials used in the cell such as alumina and pyrex glass has been reported leading to the formation of oxychlorides [14]. However from our studies we found that it 15 Page 15 of 60

became significant only when the same electrolyte was used for repeated measurement and the duration of the measurement were too long. We did not intend to study this reaction in the present work. Hence we had made measurements with fresh electrolyte and electrode set-up

ip t

for each set of measurement. Nd metal in contact with Nd(III) ions in the melt forms Nd(II) ions as shown in Eq.

cr

(9). This prevents the attainment of stable equilibrium potential for the Nd(III)/Nd couple by emf method as reported by Fusselman et al. [52]. But the cyclic voltammograms and open

us

circuit potential measurements allowed us to estimate the equilibrium potentials of

an

Nd(III)/Nd(II), Nd(II)/Nd(0) and Nd(III)/Nd(0) redox couples which is discussed in the following section. Since our measurement is by transient technique with short time period of

M

perturbation, significant amount of Nd(II) ions will not form in the system. The Nd(II) ions

d

are generated in-situ at the electrode surface alone.

Ac ce p

For soluble-soluble couple

te

3.1.4. Estimation of apparent standard potentials from cyclic voltammogram

For soluble-soluble species, the half wave potential, E1/2, is related to the standard

potential and apparent standard potential by the following relation [37, 48] E Cp  E1 2  1.11RT F 

(10)

E pA  E1 2  1.11RT F 

(11)

E1 2  E Cp  E pA  2

(12)

Where 16 Page 16 of 60

D Nd ( II ) RT   Nd ( III )  RT   ln ln  F F   Nd ( II )  D Nd ( III )

0 E1 2  E Nd ( III ) / Nd ( II ) 

(13)

RT   Nd ( III )  ln F   Nd ( II ) 

(14)

cr

* 0 E Nd ( III ) / Nd ( II )  E Nd ( III ) / Nd ( II ) 

ip t

And the apparent standard potential is described as

D Nd ( III ) RT ln F D Nd ( II )

(15)

an

* E Nd ( III ) / Nd ( II )  E1 2 

us

Hence we get,

0 * Where E Nd ( III ) / Nd ( II ) and E Nd ( III ) / Nd ( II ) are the standard potential and apparent standard

M

potential of the Nd(III)/Nd(II) redox couple. The apparent standard potential for reduction of Nd(III)/Nd(II) was obtained from the cyclic voltammograms using Eq. (10) and Eq. (15). The

Ac ce p

te

d

value of diffusion coefficients obtained from this study was used for the calculations.

For soluble - insoluble couple

Taking into account only the voltammograms recorded at low scan rate, 25 mVs-1, for

which the reduction of Nd(II)/Nd(0) is being considered reversible, the apparent standard potential was evaluated using Eq. (16) [11, 12].

* E p ,c  E Nd ( II ) / Nd ( 0 ) 

RT RT ln X Nd ( II )   0.854 2F 2F

(16)

where XNd(II) is the mole fraction of the electroactive species Nd(II) ion and it is considered that the mole fraction of NdCl2 is the same as the mole fraction of NdCl3 added [11, 50]. 17 Page 17 of 60

The apparent standard potential thus obtained versus the Ag/AgCl (0.31 mole %) was converted to the scale of Cl2/Cl- reference electrode. Assuming unit activity of the pure metal and unit activity coefficient for AgCl at low concentration of AgCl as reported by Flengas et

RT ln X AgCl  nF

(17)

cr

0  E AgCl  E AgCl

ip t

al. [53], the potential of the Ag/AgCl reference electrode is defined as

an

0 E AgCl   1.0910  0.0002924  T ( K )

us

0 relative to the Cl2/Cl- is given by Yang et al. [54] The value of E AgCl

(18)

M

0 vs. Cl2/Cl- and E(0.31 mole % AgCl) vs. Cl2/Cl- at various temperatures The potentials E AgCl

were evaluated using Eq. (18) and Eq. (17) respectively. The apparent standard potentials of

d

the redox couples Nd(III)/Nd(II), Nd(II)/Nd(0) vs. Cl2/Cl- reference electrode at different

te

temperatures were thus evaluated and are shown in Table 2. The apparent standard potential for the redox couple Nd(III)/Nd(0) was obtained by combining the reactions for the two

Ac ce p

redox couples, Nd(III)/Nd(II) and Nd(II)/Nd(0), using the Luter equation given below [55]

E

* Nd ( III ) / Nd ( 0 )



* * E Nd ( III ) / Nd ( II )  2 E Nd ( II ) / Nd ( 0 )

(19)

3 INSERT: TABLE 2

3.1.5. Chronopotentiometry

Fig. 7 shows the chronopotentiograms of NdCl3 in LiCl-KCl eutectic at different current densities at 773 K. The chronopotentiograms showed two plateau which correspond to reduction of Nd(III) to Nd(II) and Nd(II) to Nd(0). The Sand’s equation for a reversible

18 Page 18 of 60

electrode process gives the relationship between the current density and the transition time as given by Eq. (20) [48] i c1 2  0.5nFDM1 2( x ) 1 2C M0 ( x )

ip t

(20)

where i is the current density, Acm-2 and c is the transition time, s for the cathodic

cr

polarization. The product ic1/2 should remain constant for a reversible electrode process. Fig.

us

8 shows the plot of iτ1/2 as a function of the current density for the redox couples Nd(III)/Nd(II) and Nd(II)/Nd(0) at 773 K. It may be observed that the value of iτ1/2 was

an

constant for the current density range studied for the first redox couple whereas the same was not constant for the redox couple Nd(II)/Nd(0). Hence Sand’s equation cannot be used for

M

evaluating the diffusion coefficient of Nd(II) ion. We have not used chronopotentiometry for evaluation of apparent standard potentials and diffusion coefficient of the redox couples in

d

this study. The data were used for qualitative purpose only.

Ac ce p

te

INSERT: FIGURE 7 INSERT: FIGURE 8

3.1.6. Square Wave Voltammetry

Fig. 9 shows the square wave voltammograms of NdCl3 in LiCl-KCl eutectic at 723 K

at various frequencies. The voltammograms indicate the presence of two step process for the reduction of Nd(III) ion to Nd metal. For a reversible electrochemical reaction, the net current-potential curve is bell shaped and the width W1/2, of the peak at half of its height is given by Eq. (21) [56, 57]

19 Page 19 of 60

W1 2  3.52

RT nF

(21)

The peak should be purely Gaussian in nature for a reversible system with the plot of

ip t

the peak current density against square root of frequency being linear [58]. Fig. 10 shows the plot of peak current density as a function of square root of frequency for the redox couple

cr

Nd(III)/Nd(II) and that for Nd(II)/Nd(0). It may be observed that a linear dependence is seen in case of the first redox couple only. This indicates that the reduction of Nd(III) to Nd(II)

us

shows reversible behavior and that for Nd(II)/Nd(0) does not obey a perfect reversible

voltammograms and chronopotentiograms.

an

electrode process. The observations are in agreement with the results obtained from cyclic

M

INSERT: FIGURE 9

d

INSERT: FIGURE 10

te

3.2. Voltammertic study of NdCl3 in LiCl-KCl melt on reactive electrodes

Ac ce p

3.2.1. Electrode reaction of NdCl3 at the Al and Cd interface Fig. 11 compares the cyclic voltammograms for molten LiCl-KCl on tungsten inert

electrode with those on solid aluminium and liquid cadmium pool electrodes. It may be observed that the electrochemical window of the LiCl-KCl melt is limited on aluminium and cadmium surface. Formation of Li-Al alloy due to underpotential deposition of lithium on aluminium restricts the cathodic limit and dissolution of aluminium restricts the anodic limit when aluminium was used as working electrode. Similarly, on a cadmium electrode, the potential window is limited by formation of Cd-Li alloy on the cathodic side and dissolution of cadmium on the anodic side. Behavior of NdCl3 in LiCl-KCl melt was studied on aluminium and cadmium electrodes at various temperatures. Fig. 12 and Fig. 13 show the

20 Page 20 of 60

cyclic voltammograms of NdCl3 in LiCl-KCl melt at 748 K on solid aluminium cathode and liquid cadmium pool cathode respectively. Reduction of NdCl3 on these electrodes takes place at much positive potential than on the W electrode. The activity of Nd is reduced on

ip t

aluminium and cadmium cathodes due to formation of their respective intermetallic compound and this result in underpotential deposition of neodymium on these electrodes. It

cr

may be recalled that reduction of Nd(III) ion to Nd metal takes place in two steps on W electrode with the formation of Nd(II) ion as the intermediate. Whereas the reduction of

us

Nd(III) ion takes place in a single step with three electron transfer on aluminium and

an

cadmium electrodes forming their respective intermetallic compound during the cathodic cycle and dissolution of the alloy during the anodic cycle.

M

INSERT: FIGURE 11

d

INSERT: FIGURE 12

te

INSERT: FIGURE 13

Ac ce p

Fig. 14 and Fig. 15 show the phase diagram of the Nd-Al system and Nd-Cd system respectively [59, 60]. Neodymium forms various intermetallic compounds with aluminium and cadmium. However we have observed only one redox peak predominantly for both the electrodes. It may be envisaged from the phase diagram and also form the extent of potential shift towards the anodic direction that the substrate rich intermetallic is formed in both the cases. The electrode reaction for the reduction of Nd(III) ion on Al and Cd electrode is given by, Nd(III) + 11/3Al +3e  Al11/3Nd

(22)

Nd(III) + 11Cd +3e  Cd11Nd

(23)

21 Page 21 of 60

Formation of Al11RE3 and Cd11RE has been reported for other lanthanides in the literature [25, 61, 62]. Further, the formation of Al11Nd3 and Cd11Nd has been supported by

INSERT: FIGURE 14

cr

INSERT: FIGURE 15

ip t

the XRD pattern of the electrodeposits, details of which has been discussed in section 3.4.

us

3.2.2. Cylic voltammograms on Al electrode and estimation of apparent standard potential Shape of the voltammogram in Fig. 12 resemble closely to that for a soluble-soluble

an

reversible system. The reduction peak current is due to reduction of Nd(III) ions on the aluminium surface forming the intermetallic Al11Nd3 and the oxidation peak current is due to

M

the dissolution of the alloy to Nd(III) ions into the molten salt. Inset of Fig. 12 shows the cyclic voltammograms recorded for a solution of NdCl3 in LiCl-KCl at 698 K at various scan

d

rates on an aluminium cathode. It may be observed that the cathodic and anodic peak for

te

formation and dissolution of alloy does not vary significantly between sweep rates 10-100

Ac ce p

mVs-1. In the cathodic cycle, the magnitude of Ep,c-Ep,c/2 agree with the criteria for solublesoluble reversible system (Eq. 1) yielding the value of n close to 3. The reduction may be presumed close to reversibility at low scan rates and show Nernstian behavior. However, it may not be appropriate to estimate the half wave potentials from the cathodic and anodic peak potentials as is done for a reversible soluble-soluble species in this case. The rate of diffusion of Nd(III) ions in salt phase is not comparable to the rate of diffusion of Nd in solid aluminium phase. Also, when the concentration of Nd exceeds its solubility, precipitation of the solid phase, Al11Nd3 would result and the electrode surface losses its homogeneity. Hence, using Eq. (12) for estimating the half wave potential may not be appropriate in this case.

22 Page 22 of 60

We consider it is more appropriate to derive the standard potential of Nd on aluminium electrode as adopted in case of reduction of La(III) ion on Al in our previous paper [44]. The equilibrium potential of the redox system Nd(III)/Nd(Al) is given by the

ip t

following expression : RT a Nd ( III ) ln 3F a Nd ( Al )

(24)

cr

eq 0 E Nd ( III ) / Nd ( Al )  E Nd ( III ) / Nd ( 0 ) 

us

0 where E Nd ( III ) / Nd ( 0 ) is the standard potential of Nd(III)/Nd(0) in LiCl-KCl melt. aNd(III)

and aNd(Al) are the activity of NdCl3 in salt and that of Nd in aluminium respectively.

RT RT ln X Nd ( III )  ln a Nd ( Al ) 3F 3F

M

eq * E Nd ( III ) / Nd ( Al )  E NdIII ) / Nd ( 0 ) 

an

* Introducing the apparent standard potential, E Nd ( III ) / Nd ( 0 ) , Eq. (24) is rearranged as

(25)

d

The apparent standard potential for reduction of NdCl3 on Al electrode is given as RT ln a Nd ( Al ) 3F

te

Ac ce p

* * E Nd ( III ) / Nd ( Al )  E Nd ( III ) / Nd ( 0 ) 

(26)

Hence, equation 25 is rewritten as

eq * E Nd ( III ) / Nd ( Al )  E Nd ( III ) / Nd ( Al ) 

RT ln X Nd ( III ) 3F

(27)

The equilibrium potential was obtained graphically as shown in Fig. 12. Using Eq.

(27) the apparent standard potentials on Al electrode were determined and the results are listed in Table 3. INSERT: TABLE 3 3.2.3. Cylic voltammograms on Cd electrode and estimation of apparent standard potential

23 Page 23 of 60

Fig. 13 shows the cyclic voltammograms recorded for a solution of NdCl3 in LiClKCl at 698 K on liquid cadmium pool electrode and inset of Fig. 16 shows that on cadmium film electrode at different scan rates. The reduction peak and the oxidation peak currents

ip t

appeared at much positive potentials than those on tungsten inert electrode. On the cadmium pool electrode, the cathodic and anodic peaks are similar to those for soluble-soluble

cr

reversible system. The peaks are much broader than those obtained on Al electrode. The cathodic peak potentials do not shift appreciably at higher polarization rates also and the

us

magnitude of Ep,c-Ep,c/2 meets with the condition for soluble-soluble couple (Eq. 1) for a three

an

electron transfer. Thus, the reduction is presumed to show Nernstian behavior. INSERT: FIGURE 16

M

Assuming that the concentration of Nd(III) ions are very less than that required for exceeding saturation, we have used the voltammetric curves to derive the half wave potential

d

using Eq. (12) for a soluble-soluble reversible system. Since cadmium is liquid under the

te

temperature of study, it is assumed that the diffusion coefficient of Nd in salt phase and

Ac ce p

cadmium phase are identical unlike in the case of aluminium electrode. Similar assumption has been made by other authors earlier [33, 63]. The values of apparent standard potential estimated using Eq. (12) – Eq. (15) are shown in Table 4. INSERT: TABLE 4

The difference between the cadmium pool electrode and film electrode is that the

amount of Cd in the film electrode is limited. So the deposited neodymium exceeds its solubility and the electrode loses its homogeneity. Appearance of separate solid phase makes the voltammograms on cadmium film electrode less accurate for estimating the half wave potential than those obtained from cadmium pool electrode. It may be observed in Fig. 16

24 Page 24 of 60

that the onset potential for the reduction on film electrode is more cathodic than on pool electrode. The half wave potentials on Cd film electrode are shifted cathodically by ~ 30 mV.

ip t

3.3 Open-circuit chronopotentiometry

cr

Open-circuit chronopotentiometry is a suitable technique to investigate the formation of intermetallic compounds in molten salt systems and to calculate their Gibbs energies of

us

formation [61-62, 64-65]. In this method, a thin film of the intermetallic compound is prepared on the reactive cathode by depositing the electroactive species at sufficient cathodic

an

potential for a short period (30-120 sec). The polarization is stopped and the open-circuit potential is monitored as a function of time when no current flows. The deposited metal

M

reacts with the substrate and diffuses into the bulk of the electrode. It may be observed that the electrode potential gradually shifts to more positive values. A series of potential plateau

d

appears when two phases are in equilibrium at the electrode surface and the potential shifts

Ac ce p

electrode.

te

further towards anodic direction and stabilizes at the equilibrium potential of the reactive

3.3.1. OCP measurement on Al electrode Fig.17 shows the open-circuit chronopotentiograms obtained with a solution of NdCl3

in LiCl-KCl at Al electrode for a cathodic polarization of -1.8 V for various durations. A significantly distinct plateau was observed at  -1.4 V vs. Ag/Ag+ reference electrode ( 0.6 V vs. Nd(III)/Nd) in all the transients, the potential about which a peak was observed in the cyclic voltammograms. Though Nd forms various intermetallic compounds with Al, only one intermetallic is formed under our experimental condition. The plateau corresponds to the coexistence of the phases Al11Nd3 and Al. The potentials that were measured with respect to

25 Page 25 of 60

the reference electrode 0.8 wt % AgCl in LiCl-KCl were calibrated to the equilibrium potential of the Nd(III)/Nd(0) couple which was obtained as described below. INSERT: FIGURE 17

ip t

Open-circuit chronopotentiograms were earlier obtained with a solution of NdCl3 in

cr

LiCl-KCl melt at W electrode by applying a cathodic polarization (-2.3 V) for a short period (60-300 s) to form metallic Nd on the W surface. The open circuit potential was monitored as

us

a function of time. The Nd metal coated on W undergoes the reaction shown in Eq. (9).

an

Fig. 18 shows the open circuit potential after depositing Nd metal on W electrode at 2.3 V for 120 s at 723 K. Considering the above mentioned reaction, it is deduced that the

M

potential plateau at  -2.02 V correspond to the redox couple Nd(II)/Nd(0) and is the equilibrium potential for the redox couple. This potential is close to the potential of peak IIc

d

in the cyclic voltammogram. Similar behavior was observed by Cordoba et al. for their

te

studies on NdCl3 in CaCl2-NaCl on liquid Al electrode and they have used the potential values of the plateau to estimate the apparent standard potential of the Nd(II)/Nd(0) couple

Ac ce p

from OCP measurement [17]. Osipenko et al. have estimated the apparent standard potential of the Cm(III)/Cm(0) couple from OCP measurement [66]. The potential plateau at  -1.84 V is attributed to the redox couple Nd(III)/Nd(II) from the peak potentials in the cyclic voltammograms. The equilibrium potential for the redox couple Nd(III)/Nd(0) was estimated using ENd(II)/Nd(0) and ENd(III)/Nd(II) using the Luter equation given in Eq. (19). The ENd(III)/Nd(0) estimated was  -1.957 V. Hence the plateau  -1.96 V seen in Fig. 18 could be attributed to equilibrium potential of Nd(III)/Nd(0). The equilibrium potentials for Nd(III)/Nd(II) and Nd(II)/Nd(0) estimated from open circuit potential measurements at various temperatures are listed in Table 5. The equilibrium potentials for Nd(III)/Nd(0) calculated using Eq. (19) are also shown in Table 5. As stated earlier, estimation of equilibrium potential for the redox 26 Page 26 of 60

couple Nd(III)/Nd(0) using equilibrium technique such as emf method is prevented by the formation of Nd(II) ion by Eq. (9). From our study, we find that open circuit potential measurement could be a suitable method for estimation of the equilibrium potentials of redox

INSERT: FIGURE 18

us

INSERT: TABLE 5

cr

be obtained from emf measurement due to corrosive reaction [17, 52].

ip t

species such as Nd(III)/Nd(0), Am(III)/Am(0) etc whose stable equilibrium potentials cannot

an

The equilibrium potential of the redox couple Nd(III)/Nd(0) was measured and the * apparent standard potential, E Nd ( III ) / Nd ( 0 ) , was determined using the Nernst equation :

RT a Nd ( III ) ln a Nd ( 0 ) 3F

M

0 eq E Nd ( III ) / Nd ( 0 )  E Nd ( III ) / Nd ( 0 ) 

(28)

d

Defining the apparent standard potential as RT ln  Nd ( III ) 3F

Ac ce p

te

* 0 E Nd ( III ) / Nd ( 0 )  E Nd ( III ) / Nd ( 0 ) 

(29)

We get ,

eq * E Nd ( III ) / Nd ( 0 )  E Nd ( III ) / Nd ( 0 ) 

RT ln X Nd ( III ) 3F

(30)

The apparent standard potential for the redox couple Nd(III)/Nd(0) obtained from

cyclic voltammetry and OCP method are listed in Table 6. The values are compared with those reported in literature. Our values are comparable to those reported by Castrillejo et al. INSERT: TABLE 6 The equilibrium potential ENd(III)/Nd(0) thus obtained from the OCP measurement was used to rescale the equilibrium potential ENd(III)/Nd(Al) obtained on Al electrode. The potential of the plateau referred to the neodymium electrode is the emf of the cell represented as: 27 Page 27 of 60

Nd ( s ) / NdCl3 in LiCl  KCleut (l ) / Al11 / 3 Nd  Al  Gibbs energies of formation of the intermetallic Al11Nd3 at different temperatures

agreement with those reported in literature [51, 67].

cr

INSERT: TABLE 7

ip t

were estimated from the emf values and are shown in Table 7. Our values are in good

us

The activity of neodymium on aluminium was calculated from Eq. (31) and the values are given in Table 7.

RT ln a Nd ( Al ) 3F

an

emf  

(31)

ex

M

The excess Gibbs energy of Nd in aluminium,  G Nd was evaluated using Eq. (32)

d

[68, 69]

(32)

te

G exNd  RT ln  Nd ( Al )  3FE  RT ln X Nd ( Al )

Ac ce p

Since the potential plateau correspond to coexistence of Al11Nd3 and (Al) phases,

assuming saturation solubility, the mole fraction of Nd in aluminium, XNd(Al) correspond to that of Nd in Al11Nd3. Solubility of Nd in solid Al is negligible as reported in literature [70]. Hence we assumed that the neodymium formed electrochemically interacts with aluminium and forms Al11Nd3. Similar approach was made for our measurement on La-Al system in our previous work [44]. The activity coefficients of neodymium in solid aluminium calculated using Eq. (32) is shown in Table 7. The low activity coefficient values show the strong interaction of neodymium with aluminium. 3.3.2. OCP measurement on Cd electrode

28 Page 28 of 60

Similarly, open-circuit potential measurements were recorded after depositing Nd on cadmium film electrode by polarizing the electrode at -1.7 V. Fig.19 shows the open circuit potential transient curve on the cadmium film cathode monitored at different temperatures.

ip t

Only one potential plateau is seen as in case of Nd-Al system. From the phase diagram and our observations from the cyclic voltammograms, it may be concluded that the plateau

cr

correspond to the coexistence of the phases, Cd11Nd and (Cd). The potentials of the plateau were calibrated with respect to Nd(III)/Nd(0) potential obtained earlier. The potential of the

us

plateau referred to the neodymium electrode is the emf of the cell represented by

an

Nd ( s ) / NdCl3 in LiCl  KCleut (l ) / Cd11 Nd  Cd 

M

The emf for formation of Cd11Nd at different temperatures was evaluated as done in the case of aluminium electrode. The data was used for deriving the Gibbs energy formation

d

of the intermetallic, Cd11Nd, activity of neodymium in liquid cadmium. The excess Gibbs

te

energy and activity coefficient of neodymium in cadmium were calculated using Eq. (32). The mole fraction of Nd in cadmium, xNd(Cd) was taken from the solubility data reported in the

Ac ce p

literature [71, 72]. The thermodynamic properties for Nd-Cd system estimated from OCP measurements are shown in Table 8. The values of Gibbs energy formation of Cd11Nd and the activity coefficient of neodymium in cadmium are in very good agreement with those reported by Kurata et al. and Sakamura et al. [72, 73]. INSERT: FIGURE 19 INSERT: TABLE 8 3.4. Potentiostatic electrolysis To examine the alloy formed on aluminium surface, potentiostatic electrolysis was carried out at -1.5 V vs. Ag/Ag reference electrode on aluminium sheet, for 5 hrs at 773 K. 29 Page 29 of 60

the aluminium sheet containing the deposit was washed with ethylene glycol to remove the adhering salt. Fig. 20 shows SEM micrograph and the corresponding EDX analysis of the deposit. The XRD analysis of the sample confirms the formation of the intermetallic Al11Nd3

ip t

as seen in Fig. 21. Similarly, to examine the alloy formed on cadmium electrode, we had used the LiCl-KCl-NdCl3-CdCl2 melt. A Ta sheet was polarized at -1.55 V vs. Ag/AgCl reference

cr

electrode for 5 hours. This potential is slightly cathodic to the potential plateau observed in the OCP measurement for the Nd-Cd system. Co-deposition of Cd and Nd occurs to form the

us

stable intermetallic on the W surface. The cathode deposit adhered with salt was powdered,

an

placed in a glass slide and sealed thoroughly to ensure it was air tight for characterization by XRD analysis. Fig.22 shows the XRD pattern of the sample and confirms the formation of

M

the intermetallic Cd11Nd.

INSERT: FIGURE 20

te

d

INSERT: FIGURE 21

Ac ce p

INSERT: FIGURE 22

4. Conclusions

Cyclic voltammograms, chronoptentiograms and square wave voltammograms show

that the reduction of Nd (III) ion to Nd metal on tungsten electrode takes place in two steps via formation of Nd(II) in the intermediate step. The diffusion coefficient of Nd(III) and Nd (II) ions were determined from the cyclic voltammograms in the temperature range 723-798 K. The electrode kinetics for the reduction of Nd(II) to Nd metal was studied from the convoluted curves. The reduction of Nd(III)/Nd(II) showed reversible electrode kinetics whereas the reduction of Nd(II) to Nd metal followed the quasi-reversible hypothesis. The

30 Page 30 of 60

heterogeneous rate constant for the reduction of Nd(II)/Nd(0) were estimated for the temperature range 723-798 K. The apparent standard electrode potential for the redox couples Nd(III)/Nd(II),

ip t

Nd(II)/Nd(0) and Nd(III)/Nd(0) were estimated from the cyclic voltammograms and open

cr

circuit potential transients.

The electrode behavior of Nd(III) ion on solid aluminium electrode and liquid

us

cadmium electrode was studied by cyclic voltammetry. The reduction occurred at more positive potential than that on an inert electrode. The activity of Nd is reduced on the Al and

an

Cd electrode due to the formation of intermetallic compound. The apparent standard electrode potentials of Nd(III)/Nd(Al) and Nd(III/Nd(Cd) were estimated for the temperature

M

range 698-773 K. The Gibbs energy formation of the alloy, activity of neodymium in Al/Cd, the excess Gibbs energy and the activity coefficient of neodymium in Al/Cd were estimated

d

from the open circuit potential measurement. SEM-EDX analysis and the XRD pattern of the

te

electro-deposit showed the formation of intermetallic compound Al11Nd3 and Cd11Nd on Al

Ac ce p

and Cd cathodes respectively.

5. References

[1] J.P. Ackerman, Ind. Eng. Chem. Res. 30 (1991) 141. [2] J. E. Battles, W. E. Miller, E. C. Gay, ANL/CP—70796. [3] J.L. Willit, W.E. Miller, J.E. Battles, J. Nucl. Mater. 195 (1992) 229. [4] Z. Tomczuk, J. Ackerman, W. Miller, J. Electrochem. Soc. 139 (1992) 3523. [5] L. Burris, R. K. Steunenberg, W. E. Miller, CONF-861146—14. 31 Page 31 of 60

[6] T. Toda, T. Maruyama, K. Moritani, H. Moriyama, H. Hayashi, J. Nucl. Sci. Technol. 46 (1) (2009) 18.

ip t

[7] T. Koyama, M. Iizuka, H. Tanaka, M. Tokiwai, Y. Shoji, R. Fujita, T. Kobayashi, J. Nucl. Sci. Technol. 34 (1997) 384.

cr

[8] A. F. Laplace, J. Lacquement, J. L. Willit, R. A. Finch, G. A. Fletcher, M. A.

us

Williamson, Nucl. Tech. 163 (2008) 366.

an

[9] C. E. Johnson, I. Johnson, P. E. Blackburn, C. E. Crouthamel, Reactor Tech. 15 (4) (1972) 303.

M

[10] S.A. Kuznetsov, M. Gaune-Escard, J. Electroanal. Chem. 595 (2006) 11.

d

[11] Y. Castrillejo, M.R. Bermejo, E. Barrado, A.M. Martıneza, P. Dıaz Arocas, J.

te

Electroanal. Chem. 545 (2003) 141.

Ac ce p

[12] P. Masset, R.J.M. Konings, R. Malmbeck, J. Serp, J. P. Glatz, J. Nucl. Mater. 344 (2005) 173.

[13] G. Bourges, D. Lambertin, S. Rochefort, S. Delpech, G. Picard, J. Alloys Compd. 444 (2007) 404.

[14] S. A. Kuznetsov, M. Gaune-Escard, J. Nucl. Mater. 389 (2009) 108. [15] S. Kobayashi, K. Kobayashi, T. Nohira, R. Hagiwara, T. Oishi, H. Konishi, J. Electrochem. Soc. 158 (12) (2011) 142. [16] Y. Castrillejo, M.R. Bermejo, P. Dıaz Arocas, A.M. Martınez, E. Barrado, J. Electroanal. 32 Page 32 of 60

Chem. 579 (2005) 343. [17] G. De Cordoba, A. Laplace, O. Conocar, J. Lacquement, C. Caravaca, Electrochim. Acta

ip t

54 (2008) 280. [18] K. Fukasawa, A. Uehara, T. Nagai, T. Fujii, H. Yamana, J. Alloys Compd. 509 (2011)

cr

5112.

us

[19] H. Yamana, B. G. Park, O. Shirai, T. Fujii, A. Uehara, H. Moriyama, J. Alloys Compd.

an

66 (2006) 408.

[20] T. Yamamura, M. Mehmood, H. Maekawa, Y. Sato, Chem. Sustainable Development

M

12 (2004) 105.

te

d

[21] A. Novoselova, V. Smolenski, Electrochim. Acta 87 (2013) 657.

Ac ce p

[22] P. Taxil, L. Massot, C. Nourry, M. Gibilaro, P. Chamelot, L. Cassayre, J. Fluorine Chem. 130 (2009) 94.

[23] M. Gibilaro, L. Massot, P. Chamelot, P. Taxil, Electrochim. Acta 55 (2009) 281. [24] P. Soucek, R. Malmbeck, E. Mendes, C. Nourry, D. Sedmidubsky, J. P. Glatz, J. Nucl. Mater. 394 (2009) 26.

[25] P. Soucek, R. Malmbeck, C. Nourry, J. P. Glatz, Energy Procedia 7 (2011) 396. [26] L. Cassayre, R. Malmbeck, P. Masset, J. Rebizant, J. Serp, P. Soucek, J. P. Glatz J. Nucl. Mater. 360(2007) 49.

33 Page 33 of 60

[27] P. Soucek, L. Cassayre, R. Malmbeck, E. Mendes, R. Jardin, J.P. Glatz, Radiochim. Acta 96 (2008) 315.

ip t

[28] J. Serp, M. Allibert, A. Le Terrier, R. Malmbeck, M. Ougier, J. Rebizant, J. P. Glatz , J. Electrochem. Soc. 152 (2005) 167.

cr

[29] P. Soucek, R. Malmbeck, E. Mendes, C. Nourry, J.P. Glatz, J. Radioanal. Nucl. Chem.

us

286 (2010) 823.

an

[30] L. Cassayre, C. Caravaca, R. Jardin, R. Malmbeck, P. Masset, E. Mendes, J. Serp, P. Soucek, J.P. Glatz, J. Nucl. Mater. 378 (2008) 79.

M

[31] O. Conocar, N. Douyere, J.P. Glatz, J. Lacquement, R. Malmbeck, J. Serp, Nucl. Sci.

d

Eng. 153 (2006) 253.

Ac ce p

360 (2007) 32.

te

[32] Y. Castrillejo, R. Bermejo, A.M. Martınez, E. Barrado, P. Dıaz Arocas, J. Nucl. Mater.

[33] Y. Castrillejo, M.R. Bermejo, P. Dıaz Arocas, A.M. Martınez, E. Barrado, J. Electroanal. Chem. 575 (2005) 61.

[34] T. Subramanian, B. Prabhakara Reddy, P. Venkatesh, R. Kandan, S. Vandarkuzhali, K. Nagarajan, P.R. Vasudeva Rao, Studies on the head-end steps for pyrochemical reprocessing of oxide fuels, in: Proceedings of the Workshop on Pyrochemical Separations, Avignon, France, March 14–16, 2000, p. 187. [35] B. Prabhakara Reddy, S. Vandarkuzhali, T. Subramanian, P. Venkatesh, Electrochim. 34 Page 34 of 60

Acta 49 (2004) 2471. [36] Z. Galus, Fundamentals of Electrochemical Analysis, Ellis Horwood, London, 1994.

ip t

[37] S.A. Kuznetsov, M. Gaune-Escard, Electrochim. Acta 46 (2001) 1101. [38] G. Mamantov, D.L. Manning, J.M. Dale, J. Electroanal. Chem. 9 (1965) 253.

cr

[39] P. Masset, D. Bottomley, R. Konings, R. Malmbeck, A. Rodrigues, J. Serp, J. P. Glatz, J.

us

Electrochem. Soc.152 (6) (2005) A1109.

an

[40] H. Matsuda, Y. Ayabe, Z. Elektrochem. 59 (1955) 494.

[41] K.B. Oldham, G.D. Zoski, Anal. Chem. 52 (13) (1980) 2116.

M

[42] L. Nadio, J.M. Saveant, D. Tessier, J. Electroanal. Chem. 52 (1974) 403.

te

(1998) 797.

d

[43] M. Mohamedi, S. Martimet, J. Bouteillon, J.C. Poignet, Electrochim. Acta 44

Ac ce p

[44] S. Vandarkuzhali, Nibedita Gogoi, Suddhasattwa Ghosh, B. Prabhakara Reddy, K. Nagarajan, Electrochim. Acta 59 (2012) 245.

[45] Y. Castrillejo, M.R. Bermejo, A.M. Martinez, P. Diaz Arcas, J. Min. Metall. B 39 (1-2) (2003) 109.

[46] Y. Castrillejo, S. Palmero, M.A. Garcia, L. Deban, P. S. Batanero, Electrochim. Acta 41 (15) (1996) 2461. [47] M Goto, K. B. Oldham, Anal. Chem. 45 (12) (1973) 2043. [48] A.J. Bard, L.R. Faulkner, Electrochemical Methods. Fundamentals and Applications, 35 Page 35 of 60

Wiley New York, 1980. [49] T. Store, G.M. Haarberg, R. Tunold, J. Appl. Electrochem. 30 (2000) 1351.

ip t

[50] P. Delahay (Ed.), New Instrumental Methods in Electrochemistry. Theory, Instrumentation and Applications to Analytical and Physical Chemistry, Interscience,

cr

New York, 1954.

us

[51] J. P. Glatz, R. Malmbeck, C. Pernel, C. Scheppler, J. Serp, PYROREP Contract No.

an

FIKW- CT-2000-00049, Project No. FIS5-1999-00199.

[52] S.P. Fusselman, J.J. Roy, D.L. Grimmett, I.F. Grantham, C.L. Krueger, C.R. Nabelek, T.

M

S. Storvick, T. Inoue, T. Hijikata, K. Kinoshita, Y. Sakamura, K. Uozumi, T. Kawai, N.

d

Takahashi, J. Electrochem. Soc. 146 (7) (1999) 2573.

te

[53] S.N. Flengas, T.R. Ingraham, J. Electrochem. Soc. 106 (8) (1959) 714.

Ac ce p

[54] L. Yang, R.G. Hudson, J. Electrochem. Soc. 106 (11) (1959) 986. [55] S.A. Kuznetsov, L. Rycerz, M. Gaune-Escard, J.Nucl. Mater. 344 (1-3) (2005) 152. [56] J. Osteryoung, R.A. Osteryoung, Anal. Chem. 57 (1985) 101. [57] L. Ramaley, M.S. Krasue, Anal. Chem. 41 (1969) 1362. [58] A. Saila, M. Gibilaro, L. Massot, P. Chamelot, P. Taxil, A. M. Affoune, J. Electroanal. Chem. 642 (2010) 150. [59] H. Okamoto: Al-Nd Phase Diagram, ASM Alloy Phase Diagrams Center, P. Villars, Editor-in-chief; http://www1.asminternational.org/AsmEnterprise/APD, ASM 36 Page 36 of 60

International, Materials Park, OH, USA, 2006-2014. [60] K.A. Gschneidner: Cd-Nd Phase Diagram, ASM Alloy Phase Diagrams Center, P.

ip t

Villars, Editor-in-chief; http://www1.asminternational.org/AsmEnterprise/APD, ASM International, Materials Park, OH, USA, 2006-2014.

cr

[61] O. Shirai, A. Uehara, T. Fujii, H. Yamana, J. Nucl. Mater. 344 (2005) 142–145.

us

[62] Y. Castrillejo, M.R. Bermejo, E. Barrado, A.M. Martınez, Electrochim. Acta 51 (2006)

an

1941.

[63] O. Shirai, K. Uozumi. T. Awai, Y. Arai, Analytical Sciences 17 (2001) i959.

M

[64] P. Soucek, R. Malmbeck, E. Mendes, C. Nourry, D. Sedmidubsky, J. P. Glatz, J. Nucl.

d

Mater. 394 (2009) 26.

te

[65] T. Iida, T. Nohira, Y. Ito, Electrochim. Acta 46 (2001) 2537.

Ac ce p

[66] A. Osipenko, A. Maershin, V. Smolenski, A. Novoselova, M. Kormilitsyn, A. Bychkov J. Electroanal. Chem. 651 (2011) 67.

[67] M. C. Gao, N. Unlu, G.J. Shiflet, M. Mihalkovic, M Widom Metallurgical and Materials Trans. A 36 (A) (2005) 3269.

[68] H. Konishi, T. Nishikiori, T. Nohira, Y. Ito, Electrochim. Acta 48 (2003) 1403. [69] P. Taxil, J. Mahenc, J. Appl. Electrochem. 17 (1987) 261. [70] Landolt-Börnstein, Elements and Binary Systems from Ag-Al to Au-Tl - Group IV Physical Chemistry Volume 19B1, 2002, p. 178. 37 Page 37 of 60

[71] I. Johnson, K.E. Anderson, and R.A. Blomquist: Trans ASM 59 (1966), 352. [72] Masaki Kurata and Yoshiharu Sakamura, J. Phase Equilibria 22 (3) (2001) 232.

ip t

[73] Y. Sakamura, T. Inoue, T.S. Storvick, and L.F. Grantham: Proc. Int. Conf. on Evaluation of Emerging Nuclear Fuel Cycle Systems, Global ‘95, Versailles, France, Sept. 11–14,

Ac ce p

te

d

M

an

us

cr

1995, vol. 2, p. 1185.

38 Page 38 of 60

Figure captions Fig.1. Cyclic voltammograms of LiCl-KCl-NdCl3 melt on W electrode for different scan rates. Temperature: 723 K; Concentration of NdCl3: 6.66 x 10-5 mol cm-3; Electrode area:

ip t

0.49 cm2.

cr

Fig.2. Cyclic voltammograms of LiCl-KCl-NdCl3 melt on W electrode obtained at different

us

switching potentials.

an

Fig.3. Plot of the cathodic peak potentials as a function of logarithm of the scan rates at 723

M

and 798 K.

Fig.4. Cyclic voltammograms of LiCl-KCl-NdCl3 melt on W electrode at 723 K and

te

d

corresponding convoluted curve.

Fig.5. Plot of current density for peak Ic vs. square root of scan rate for LiCl-KCl-NdCl3

Ac ce p

melt.Temperature: 723 K; Concentration of NdCl3: 6.66 x 10-5 mol cm-3.

Fig.6. Plot of current density for peak IIc vs. square root of scan rate for LiCl-KCl-NdCl3 melt.Temperature: 723 K; Concentration of NdCl3: 6.66 x 10-5 mol cm-3.

Fig.7. Chronopotentiograms on a tungsten electrode for LiCl-KCl-NdCl3 melt. Temperature: 723 K; Concentration of NdCl3: 6.66 x 10-5 mol cm-3; Electrode area: 0.49 cm2.

Fig.8. Variation of iτ1/2 with different current densities for the redox couples Nd(III)/Nd(II) and Nd(II)/Nd(0) for the chronopotentiograms shown in Fig. 7. 39 Page 39 of 60

Fig.9. Square wave voltammograms of NdCl3 in LiCl-KCl eutectic at 723 K at various frequencies. Amplitude of pulse: 10 mV. Concentration of NdCl3: 6.66 x 10-5 mol cm-3;

ip t

Electrode area: 0.49 cm2.

cr

Fig.10. Plot of peak current density as a function of square root of frequency for LiCl-KCl-

us

NdCl3 melt. Temperature: 723 K; Concentration of NdCl3: 6.66 x 10-5 mol cm-3.

M

Concentration of NdCl3: 6.66 x 10-5 mol cm-3.

an

Fig.11. Cyclic voltammograms of LiCl-KCl-NdCl3 melt at different electrodes;

Fig.12. Cyclic voltammograms for NdCl3 in LiCl-KCl melt at aluminium electrode.

d

polarization rate: 10 mVs-1; Concentration of NdCl3: 6.66 x 10-5 mol cm-3. Inset: Cyclic

te

voltammograms at different scan rates.

Ac ce p

Fig.13. Cyclic voltammograms for NdCl3 in LiCl-KCl melt at cadmium pool electrode. Polarization rate: 10 mVs-1; Concentration of NdCl3: 6.66 x 10-5 mol cm-3.

Fig.14. Al-Nd phase diagram [59].

Fig.15. Cd-Nd phase diagram [60].

Fig.16. Comparison of cyclic voltammograms for NdCl3 in LiCl-KCl melt at cadmium pool and film electrode. Polarization rate: 10 mVs-1; Concentration of NdCl3: 6.66 x 10-5 mol cm-3.

40 Page 40 of 60

Inset: Cyclic voltammograms at cadmium film electrode at different scan rates. Temperature: 723 K.

ip t

Fig.17. Open circuit potential transient curve for NdCl3 in LiCl-KCl melt on Al electrode. Cathodic polarization: -1.8 V vs. (Ag/Ag+) reference electrode for different durations. Inset:

us

cr

OCP transient curve at 723 K. Cathodic polarization: -1.8 V for 600 s.

Fig.18. Open circuit potential transient curve for NdCl3 in LiCl-KCl melt on W electrode.

an

Cathodic polarization: -2.3 V vs. (Ag/Ag+) reference electrode. Inset: OCP transient curve

M

showing attainment of open circuit potential for W electrode after long duration.

Fig.19. Open circuit potential transient curve for NdCl3 in LiCl-KCl melt on Cd film

te

d

electrode. Cathodic polarization: -1.7 V vs. (Ag/Ag+) reference electrode.

Ac ce p

Fig.20. (a) Cross-sectional SEM image of Nd–Al film formed by potentiostatic electrolysis at -1.5 V vs. (Ag/Ag+) reference electrode at 773 K and (b) corresponding EDX analysis confirming the formation of Al11Nd3.

Fig.21. XRD pattern of Nd–Al film formed by potentiostatic electrolysis at -1.5 V vs. (Ag/Ag+) reference electrode for 5 hrs at 773 K.

Fig.22. XRD pattern of Nd-Cd film formed by potentiostatic electrolysis at -1.55V vs. (Ag/Ag+) reference electrode for 5 hrs at 748 K.

41 Page 41 of 60

Table captions Table1. Diffusion coefficient and heterogeneous rate constant at different temperatures. Concentration of NdCl3: 6.66 x 10-5 molcm-3.

ip t

Table 2. Apparent standard potentials (± 0.003 V) and Gibbs free energy formation of NdCl3

cr

estimated from cyclic voltammograms for LiCl-KCl-NdCl3 melt at W electrode.

voltammogram on Al electrode, X NdCl3 = 0.00224.

us

Table 3. Half wave potential (± 0.002 V) for Nd(III)/Nd(Al) estimated from cyclic

M

voltammogram on Cd pool, X NdCl3 = 0.00224.

an

Table 4. Half wave potential (± 0.002 V) for Nd(III)/Nd(Cd) estimated from cyclic

Table 5. Equilibrium potentials (± 0.003 V) on inert W electrode from OCP measurement.

te

d

Table 6. Apparent standard potentials obtained from cyclic voltammetry and OCP method Table 7. Thermodynamic properties of Nd-Al intermetallic compound estimated from OCP

Ac ce p

measurement on Al electrode

Table 8. Thermodynamic properties of Nd-Cd intermetallic compound estimated from OCP measurement on Cd film electrode

42 Page 42 of 60

Tables

Table1. Diffusion coefficient and heterogeneous rate constant at different temperatures. Concentration

748

773

DNd(III) ( 0.05 x 105, cm2s-1)

0.28

0.45

0.54

DNd(II) (0.05 x 105, cm2s-1)

0.98

1.38

1.69

ks, Nd(II)/Nd(0) (x 106, cms-1)

2.95

3.23

798

0.76

cr

723

2.03

us

Temperature, K

ip t

of NdCl3: 6.66 x 10-5 molcm-3.

5.01

7.76

M

an

DNd(III) and DNd(II) (x 105, cm2s-1): 0.95 and 1.25 respectively at 733 K [51].

Table 2. Apparent standard potentials (± 0.003 V) and Gibbs free energy formation of NdCl3

d

estimated from cyclic voltammograms for LiCl-KCl-NdCl3 melt at W electrode. * E Nd ( III ) / Nd ( II )

* E Nd ( II ) / Nd ( 0 )

* E Nd ( III ) / Nd ( 0 )

 G NdCl 3

vs. Cl2/Cl , V

vs. Cl2/Cl , V

vs. Cl2/Cl , V

± 0.9 kJ mol-1

723

-3.058

-3.110

-3.093

-895.4

748

-3.049

-3.094

-3.079

-891.3

773

-3.039

-3.078

-3.065

-887.2

798

-3.029

-3.061

-3.050

-883.1

te

Temperature, K

Ac ce p

-

-

-

43 Page 43 of 60

Table 3. Half wave potential (± 0.002 V) for Nd(III)/Nd(Al) estimated from cyclic voltammogram on Al electrode, X NdCl3 = 0.00224. * E Nd ( III ) / Nd ( Al )

vs. Ag/Ag+, V

* E Nd ( III ) / Nd ( Al )

-

vs. Cl2/Cl , V

-1.406

-2.640

723

-1.384

-2.623

748

-1.362

-2.606

798

-1.318

-2.573

an

us

cr

698

ip t

Temperature, K

M

Table 4. Half wave potential (± 0.002 V) for Nd(III)/Nd(Cd) estimated from cyclic voltammogram on Cd pool, X NdCl3 = 0.00224.

* E Nd ( III ) / Nd ( Cd )

* E Nd ( III ) / Nd ( Cd )

vs. Ag/Ag+, V

vs. Cl2/Cl , V

te

d

Temperature, K

-1.499

-2.733

723

-1.487

-2.726

733

-1.482

-2.723

758

-1.469

-2.716

773

-1.462

-2.711

Ac ce p

698

-

44 Page 44 of 60

Table 5. Equilibrium potentials (± 0.002 V) on inert W electrode from OCP measurement. Temperature, K

E Nd ( III ) / Nd ( II )

E Nd ( II ) / Nd ( 0 )

E Nd ( III ) / Nd ( 0 )

E Nd ( III ) / Nd ( II )

vs. Ag/Ag+, V

vs. Ag/Ag+, V

vs. Ag/Ag+, V

vs. Cl2/Cl , V

698

-1.858

-2.030

-1.973

-3.093

723

-1.842

-2.015

-1.957

-3.082

748

-1.825

-2.001

-1.942

-3.069

773

-1.807

-1.986

-1.926

-3.057

-

vs. Cl2/Cl , V

E Nd ( III ) / Nd ( 0 ) -

vs. Cl2/Cl , V -3.207

-3.254

-3.197

cr

ip t

-3.264

-3.245

-3.186

-3.235

-3.176

M

an

us

-

E Nd ( II ) / Nd ( 0 )

Table 6. Apparent standard potentials obtained from cyclic voltammetry and OCP method.

Temperature, K

d

* E Nd ( III ) / Nd ( 0 )

-

* E Nd ( III ) / Nd ( 0 )

vs. Cl2/Cl , V (OCP)

Literature

-3.093

-3.070

-3.105 [11]

748

-3.079

-3.055

-3.113 at 733 K [12]

773

-3.065

-3.040

-3.079 [11]; -3.25 [19]

798

-3.050

-3.025

Ac ce p

723

te

vs. Cl2/Cl , V (Cyclic voltametry)

-

45 Page 45 of 60

Table 7. Thermodynamic properties of Nd-Al intermetallic compound estimated from OCP measurement on Al electrode log a Nd ( Al )

 G Nd kJ mol-1

log  Nd ( Al )

-12.22

-12.8

-162.1

-12.1

0.583

-12.06

-12.2

-159.5

-11.5

748

0.575

-11.89

-11.6

-156.8

-10.9

773

0.567

-11.73

-11.1

-154.2

-10.4

0.591

723

ip t

698

cr

kJ gatom-1

ex

te

d

M

an

emf(±0.002) V

us

 f G Al11 Nd 3

Temperature, K

Table 8. Thermodynamic properties of Nd-Cd intermetallic compound estimated from OCP

Ac ce p

measurement on Cd film electrode

log a Nd ( Cd )

 G Nd kJ mol-1

log  Nd ( Cd )

-12.80

-11.5

-119.4

-8.9

0.529

-12.76

-11.1

-121.3

-8.8

733

0.527

-12.71

-10.9

-121.6

-8.6

773

0.521

-12.56

-10.2

-123.6

-8.4

 f GCd11Nd

Temperature, K

emf(±0.002) V

kJ gatom-1

698

0.531

723

ex

46 Page 46 of 60

te

d

M

an

us

cr

ip t

Figure 1

Ac ce p

Figure 2

47 Page 47 of 60

ip t cr us an M Ac ce p

te

d

Figure 3

Figure 4 48 Page 48 of 60

ip t cr us an M Ac ce p

te

d

Figure 5

Figure 6 49 Page 49 of 60

ip t cr us an M Ac ce p

te

d

Figure 7

Figure 8

50 Page 50 of 60

ip t cr us an M Ac ce p

te

d

Figure 9

51 Page 51 of 60

d

M

an

us

cr

ip t

Figure 10

Ac ce p

te

Figure 11

52 Page 52 of 60

d

M

an

us

cr

ip t

Figure 12

Ac ce p

te

Figure 13

53 Page 53 of 60

d

M

an

us

cr

ip t

Figure 14

Ac ce p

te

Figure 15

54 Page 54 of 60

te

d

M

an

us

cr

ip t

Figure 16

Ac ce p

Figure 17

55 Page 55 of 60

d

M

an

us

cr

ip t

Figure 18

Ac ce p

te

Figure 19

56 Page 56 of 60

te

d

M

an

us

cr

ip t

Figure 20

Ac ce p

Figure 21

57 Page 57 of 60

ip t cr us an M

Ac ce p

te

d

Figure 22

58 Page 58 of 60

Ac ce p

te

d

M

an

us

cr

ip t

Graphical abstract

60 Page 59 of 60

Research highlights Electrochemical behavior of neodymium (III) ion was studied in LiCl-KCl eutectic melt in the temperature range 723-798 K on inert tungsten electrode using various transient electrochemical

ip t

techniques such as cyclic voltammetry, convolution voltammetry, chronopotentiometry and square wave voltammetry. The reduction of Nd(III) ion to Nd metal on tungsten electrode takes place in two

cr

steps- Nd(III)/ Nd(II) and Nd(II)/ Nd(0). Reduction of Nd(III) to Nd(II) showed reversible electrode

us

behavior and that for Nd(II) to Nd metal followed the quasi-reversible behavior. Heterogeneous rate constant for the reduction, Nd(II)/Nd(0) was estimated from the convoluted voltammograms. The

an

* * * apparent standard electrode potentials, E Nd E Nd and E Nd were ( III ) / Nd ( II ) , ( II ) / Nd ( 0 ) ( III ) / Nd ( 0 )

estimated from the cyclic voltammograms and open circuit potentiograms.

M

The electrode behavior of Nd(III) ion on solid aluminium electrode and liquid cadmium electrode was studied by cyclic voltammetry. Under-potential reduction of Nd(III) ion takes place on

d

Al and Cd cathodes in a single step with three electron transfer. The apparent standard electrode

te

* * potentials, E Nd ( III ) / Nd ( Al ) and E Nd ( III ) / Nd ( Cd ) were estimated for different temperatures in the range

Ac ce p

698-773 K. The formation of intermetallics, Al11Nd3 and Cd11Nd were studied from open circuit potential measurement on Al and Cd film electrode respectively. Thermodynamic properties of NdAl and Nd-Cd systems were evaluated.

61 Page 60 of 60