- Email: [email protected]
Electrochemical purification and GFAAS analysis of heavy metal fluoride glass
] O I I R N A L OF
ELSEVIER
Journal of Non-CrystallineSolids 184 (1995) 194-199
Electrochemical purification and GFAAS analysis of heavy metal fluoride glass Shanci Bao, P.J. Newman, A. Voelkel, Zhiping Zhou, D.R. MacFarlane * Department of Chemistry, Monash University, Wellington Road, Clayton, V1C 3168, Australia
Abstract
An electrochemical method for the in situ purification of heavy metal fluoride melts has been developed. Several factors which influence the efficiency of purification such as electrode materials, rotation rate, temperature, deposition time and purification times have been studied. The purified glasses were analyzed using graphite furnace atomic absorption spectroscopy (GFAAS). The results show that Fe, Cu and Ni impurities in fluoride glass melts can be reduced to the level of 0.3-0.5 ppm.
1. Introduction
One of the main technological difficulties confounding the application of heavy metal fluoride glasses is obtaining ultrahigh-purity glass. Substantial effort has been devoted to the purification of fluoride raw materials, but little to the purification of the glass melt. Electrochemical methods investigated in our laboratories [1] provide a new way to purify the glass melt directly. The major applications for infrared glass fibres are optical amplifiers and optical sensors for chemical and temperature sensing, where the loss requirements are of the order of 10 d B / k m . To satisfy this requirement, the 3d transition metal elements such as Fe, Ni, Cu and Co in the glass must be < 1 ppm and in some cases < 0.1 ppm. Although some commer-
* Corresponding author. Tel: + 61-3 905 4540. Telefax: + 61-3 565 4597. E-mail: [email protected].
cial fluoride materials purified by sublimation and distillation methods can obtain this standard, often the glasses made from these materials still contain impurities > 1 ppm, due to possible contamination introduced in the transport, storage and glass-making process. Electrochemical methods of purification have been investigated in our laboratories over the past few years [1]. They can be used to purify the fluoride melt during the glass-making process, hence reducing the possible contamination prior to fibre drawing. This method also opens the possibility of using raw materials with high transition metal impurities (i.e., > 1 ppm) to prepare glass containing levels < 1 ppm after purification. This method of electrochemical purification may be justified when supplies of ultrahigh-purity fluorides diminish due to the reduction in market demand [2]. Another advantage of this method is that purification of the glass may be more convenient and efficient than purification of the individual raw materials. This paper describes and discusses the effect of several electro-
0022-3093/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDI 0022-3093(95)00005-4
195
S. Bao et al. /Journal of Non-Crystalline Solids 184 (1995) 194-199
c h e m i c a l parameters on the process o f purification o f the m e l t by direct e l e c t r o c h e m i c a l reduction o f transition metal impurities.
-]
2. E x p e r i m e n t a l
Working
Electrode ' Counter Electrode
- - ] [
Dry Box
--~ Motor
~
PolarographieI
.m-
T h e glass c o m p o s i t i o n used was 5 3 Z r F a - 2 0 B a F 2 4 L a F 3 - 3 A 1 F 3 - 2 0 N a F m o l % ( Z B L A N 2 0 ) . The starting materials w e r e high-purity c o m m e r c i a l fluorides supplied by Morita, e x c e p t for N a F w h i c h was supplied by B D H . The g l a s s - m a k i n g process was the s a m e as that o f Z h o u et al [1], the only difference b e i n g that the g l a s s - m a k i n g process and the electroc h e m i c a l purification process w e r e carried out in the s a m e glassy carbon crucible, to reduce the likelihood o f c o n t a m i n a t i o n during transfer o f glass f r o m one crucible to the other. A t w o - e l e c t r o d e cell was e m p l o y e d for electroc h e m i c a l purification. T h e w o r k i n g electrodes w e r e either a 10 × 10 m m 2 platinum plate, a 3 m m diameter s p e c t r o s c o p i c carbon graphite rod or a 4 m m d i a m e t e r glassy carbon rod. The glassy carbon crucible was used as the counter electrode. The crucible was placed in a resistance furnace. This apparatus was set up in a dry box under a nitrogen atmosphere. The instrument used for controlling the deposition process was a p o l a r o g r a p h i c analyser ( A M E L 433-A). Rotation of the w o r k i n g electrode at a constant and reproducible rate was a c h i e v e d by using a digital
[
l xx I
Analyser
I
m
I;-
Speed
Controller
Computer :-
I
-~Thermometer
-1 I
I
-~
Variac
Fig. 1. Schematic diagram of the apparatus used for electrochemical melt purification.
m o t o r speed controller. The furnace temperature was adjusted with a Variac, and was m e a s u r e d with a t h e r m o c o u p l e attached to a digital thermometer. A s c h e m a t i c diagram o f the entire apparatus is s h o w n in Fig. 1. B e f o r e the purification process, the temperatures of the melt and furnace w e r e calibrated. It was found that the temperature o f the melt was constant approximately 45 min after the temperature o f the furnace reached the set value. In the treatment process, the crucible c o n t a i n i n g the glass was placed into the furnace; the temperature was increased to a set value and maintained for 1 h to obtain a uniform melt temperature; then a j a c k was used to raise the fur-
Table 1 Electrochemical purification conditions Melt
Temperature ( + 2) (°C)
1
.
2 3 4 5 6 7 8 9 10 11 12 13 14
600 600 600 600 600 600 600 600 600 600 560 640 600
.
Rotation rate ( _+3) (rpm) .
0 300 600 300 300 300 900 1200 1500 1800 300 300 1200
.
Deposition time ( + 5) (s)
Deposition potential (mV)
Electrode
2700 3600 3600 3600 7200 7200 3600 3600 2400 2400 3600 3600 3600 × 2
- 1100 - 1100 1100 - 1100 - 1100 - 1100 - 1100 - 1100 1100 - 1100 - 1100 - 1100 1100
glassy carbon graphite graphite platinum platinum graphite graphite graphite graphite graphite graphite graphite graphite
.
196
S. Bao et aL /Journal of Non-Crystalline Solids 184 (1995) 194-199
Table 2 The influence of electrode materials on final impurity level achieved Melt Electrode Fe Cu Ni (ppm+ 10%) (ppm+ 10%) (ppm± 10%) 1
-
4.12
3 5 6 7
graphite platinum platinum graphite
1.80 2.03 1.51 2.07
1.01 0.34 0.37 0.30 0.32
1.62 0.78 0.75 0.58 0.60
that the reproducibility and detection limits for Fe, Cu and Ni were 4, 2 and 3 ppb, respectively. Sample 1 was a reference sample which was given no electrochemical treatment but which was otherwise handled identically.
3. Results
nace to the point that the melt contacted the working electrode. The furnace was than raised a further 11 mm. This allowed the surface of the working electrode to be submerged about 1 cm into the melt. At the end of the deposition process, the working electrode was removed from the melt while the deposition potential was maintained, in order to prevent the deposit from redissolving. The electrochemical parameters for purification are shown in Table 1. The purified samples were then quenched and analysed using graphite furnace atomic absorption spectroscopy (GFAAS) [3]. The GFAAS procedure has been described in detail elsewhere [3]: briefly, samples were digested and extracted prior to analysis. Blanks were run through the entire digestion, extraction and analysis procedure which indicated
The samples investigated and the conditions used are summarized in Table 2 along with the analytical results obtained. All of the impurities investigated were reduced after the purification process. Comparison with the reference sample (sample 1) shows that an improvement in Fe, Ni and Cu levels of approximately a factor of three was obtained. From the data in Table 2, it appears that there is not a large influence of the type of electrode material used in the purification (graphite or platinum). The graphite rod electrode was found to be the easiest to clean after each purification process and most cost effective, therefore most of the experiments were conducted with this type of electrode. The rotation rate of the electrode was found to have an effect on purification. Under stationary conditions, the metal ion concentrations remaining in the melt were almost the same as the original glass. By
Table 3 The influenceof rotation rate and time on final impurity level achieved Melt
Rotation rate (rpm)
Time (s)
Fe (ppm + 10%)
Cu (ppm + 10%)
Ni (ppm + 10%)
2 3 4 8 9 10 11 5 6
0 300 600 900 1200 1500 1800 300 300
2700 3600 3600 3600 3600 2400 2400 3600 7200
4.29 1.80 1.73 1.53 1.35 2.07 1.30 2.03 1.51
1.16 0.34 0.25 0.36 0.29 0.32 0.26 0.37 0.30
2.00 0.78 0.69 0.68 0.63 0.81 0.48 0.75 0.58
Table 4 The influenceof temperature on final impurity level achieved Melt Temperature (°C) Fe (ppm + 10%)
Cu (ppm + 10%)
Ni (ppm + 10%)
12 3 13
0.37 0.34 0.23
0.75 0.78 0.62
560 600 640
1.87 1.80 1.91
198
S. Bao et al. /Journal of Non-Crystalline Solids 184 (1995) 194-199
electroactive species is transported by migration; diffusion is expected to be the main mechanism by which these impurity species reach the electrode [5]. Thus the deposition rate depends on the diffusion coefficient of the ions, their initial concentration (or their concentration beyond the diffusion layer) and the diffusion layer thickness. Among these factors, diffusion layer thickness is the main parameter which can be adjusted easily in a purification process. Under stirring conditions, the transfer of ions beyond the diffusion layer is controlled by convection and the transfer in the narrow stagnant diffusion layer is controlled by diffusion. From Fick's law, the concentration gradient determines the diffusion rate. The thinner the diffusion layer, the larger the concentration gradient and in turn the larger the diffusion rate will be. The diffusion layer thickness, 6,0, is dependent on the rotation rate of the electrode as expressed by the empirical equation [5] 6,0 = 6o(O ",
(1)
where to is the angular velocity of the rotating electrode, 60 is a constant and n = 0.5. Hence, the deposition rate is predicted to increase with increasing rotation rate of the electrode. The concentration gradient also depends on the concentration beyond the diffusion layer, this decreasing gradually during the purification process. Thus, the deposition rate would be expected to decrease gradually during the process, as is observed in this work. Under steady-state conditions, i.e., where the concentration gradient does not change with time, the quantity deposited is proportional to (oo5 for fixed deposition time, as was demonstrated by Zhou et al. [1]. In the purification process, where the impurity concentration in the melt decreases gradually and thus the concentration gradient in the diffusion layer changes with time, the total deposited mass will depend on to in a more complex way (Table 3).
4.3. The influence of temperature Whether the temperature is expected to influence the purification process depends on how large the corresponding diffusivity and ionic conductivity changes are. The ZBLAN20 melt in the range 560-
640°C has a viscosity and conductivity of the order of 10 1 p and 10 -1 (1~ cm) -1, respectively [7]. Over this limited temperature range, both parameters are not strong functions of temperature. The results given in Table 4 indicate slight improvement in impurity levels obtained at the highest temperature (except for Fe, where it is suspected that the crucible may be leaching Fe during the process).
4.4. Purification repetition and solubility The deposition process has two distinct phases. The first stage is the formation of nuclei of the metal being deposited on the electrode. This stage is important, for it determines the primary structure of the deposit layer and thereby influences the structure of the final electroplate [8]. The formation of nuclei occurs on the electrode surface sparsely, in very dilute solutions, because it requires a high local supersaturation, and the nuclei whose radii are smaller than the critical size, if formed, will redissolve at once. Thus, very dilute solutions tend to give incoherent, powdery deposits [9]. In the second stage, the growth of the crystals on these nuclei forms a porous deposit layer, whose surface area does not equal the surface area of the electrode and increases with the growth of the crystals. Since the dissolution rate is proportional to the surface area of the deposit, the dissolution rate is thus expected to increase with extent of deposition time. The dissolution of the deposited metal is an adverse process from the point of view of purification. Initially, the dissolution rate is zero, and the deposition rate is at a maximum. Deposition results in a decrease in the concentration of metal ions in the melt and so the deposition rate is also reduced, while the dissolution rate of the deposited metal increases. At some point in time, the two processes can be expected to reach equilibrium. Further prolonging the deposition time would not reduce the remaining metal concentration in the melt. In this case, replacing the old electrode with a new one (Table 5) revitalises the deposition process, and establishes a new equilibrium at a lower impurity concentration. The thermodynamic solubility of the deposit metal in a melt is a further important factor in the electrochemical process. When the solubility of the metal is
s. Bao et al. /Journal of Non-Crystalline Solids 184 (1995) 194-199
higher, the balance between deposition and dissolution will leave higher concentrations of transition metals in the melt. In the present work, the Cu could not be reduced further than 0.25-0.30 ppm; the reason may be that the concentration of Cu cannot reach supersaturation under the present electrochemical conditions.
5. Conclusions (i) The impurity levels of Fe, Cu and Ni in zirconium-based fluoride glasses can be reduced to 0 . 3 - 0 . 5 ppm by electrochemical purification of the melt. (ii) The rotation rate of the electrode, deposition time and number of repetitions are the main factors which determine the extent of purification. (iii) The behaviour of the process can be rationalized in terms of a mechanism involving rate-limiting diffusion of the impurity across the Nernst layer and nucleation and growth of the metal on the electrode.
199
References [1] (a) Zhou Zhiping, P.J. Newman and D.R. MacFarlane, J. Non-Cryst. Solids 161 (1993) 36; (b) Z. Zhou, P.J. Newman, D.K.Y. Wong and D.R. MacFarlane, J. Non-Cryst. Solids 140 (1992) 297; (c) Z. Zhou and D.R. MacFarlane, J. Non-Cryst. Solids 140 (1992) 215; (d) Z. Zhou, P.J. Newmanand D.R. MacFarlane,J. Non-Cryst. Solids 161 (1993) 27. [2] P.J. Melling and O.H. El-Bayoumi, Proc. Int. Soc. Optical Eng. 1761 (1992) 298. [3] P.J. Newman, A.T. Voelkel and D.R. MacFarlane, these Proceedings, p. 324. [4] E. Gileadi, Electrode Kinetics (VCH, New York, 1993) pp. 4-55. [5] L. Antropov, Theoretical Electrochemistry (Mir, Moscow, 1972) pp. 327-329. [6] R.C. Weast, ed., CRC Handbook of Chemistry and Physics, 65th Ed. (CRC, Cleveland,OH, 1984) pp. D155-162. [7] W.C. Hasz and C.T. Moynihan, J. Non-Cryst. Solids 140 (1992) 285. [8] D. Pletcher, Industrial Electrochemistry (Chapman, London, 1982) pp. 176-180. [9] A.J. Allmand, The Principles of Applied Electrochemistry (Arnold, London, 1924) pp. 120-124.
ELSEVIER
Journal of Non-CrystallineSolids 184 (1995) 194-199
Electrochemical purification and GFAAS analysis of heavy metal fluoride glass Shanci Bao, P.J. Newman, A. Voelkel, Zhiping Zhou, D.R. MacFarlane * Department of Chemistry, Monash University, Wellington Road, Clayton, V1C 3168, Australia
Abstract
An electrochemical method for the in situ purification of heavy metal fluoride melts has been developed. Several factors which influence the efficiency of purification such as electrode materials, rotation rate, temperature, deposition time and purification times have been studied. The purified glasses were analyzed using graphite furnace atomic absorption spectroscopy (GFAAS). The results show that Fe, Cu and Ni impurities in fluoride glass melts can be reduced to the level of 0.3-0.5 ppm.
1. Introduction
One of the main technological difficulties confounding the application of heavy metal fluoride glasses is obtaining ultrahigh-purity glass. Substantial effort has been devoted to the purification of fluoride raw materials, but little to the purification of the glass melt. Electrochemical methods investigated in our laboratories [1] provide a new way to purify the glass melt directly. The major applications for infrared glass fibres are optical amplifiers and optical sensors for chemical and temperature sensing, where the loss requirements are of the order of 10 d B / k m . To satisfy this requirement, the 3d transition metal elements such as Fe, Ni, Cu and Co in the glass must be < 1 ppm and in some cases < 0.1 ppm. Although some commer-
* Corresponding author. Tel: + 61-3 905 4540. Telefax: + 61-3 565 4597. E-mail: [email protected].
cial fluoride materials purified by sublimation and distillation methods can obtain this standard, often the glasses made from these materials still contain impurities > 1 ppm, due to possible contamination introduced in the transport, storage and glass-making process. Electrochemical methods of purification have been investigated in our laboratories over the past few years [1]. They can be used to purify the fluoride melt during the glass-making process, hence reducing the possible contamination prior to fibre drawing. This method also opens the possibility of using raw materials with high transition metal impurities (i.e., > 1 ppm) to prepare glass containing levels < 1 ppm after purification. This method of electrochemical purification may be justified when supplies of ultrahigh-purity fluorides diminish due to the reduction in market demand [2]. Another advantage of this method is that purification of the glass may be more convenient and efficient than purification of the individual raw materials. This paper describes and discusses the effect of several electro-
0022-3093/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDI 0022-3093(95)00005-4
195
S. Bao et al. /Journal of Non-Crystalline Solids 184 (1995) 194-199
c h e m i c a l parameters on the process o f purification o f the m e l t by direct e l e c t r o c h e m i c a l reduction o f transition metal impurities.
-]
2. E x p e r i m e n t a l
Working
Electrode ' Counter Electrode
- - ] [
Dry Box
--~ Motor
~
PolarographieI
.m-
T h e glass c o m p o s i t i o n used was 5 3 Z r F a - 2 0 B a F 2 4 L a F 3 - 3 A 1 F 3 - 2 0 N a F m o l % ( Z B L A N 2 0 ) . The starting materials w e r e high-purity c o m m e r c i a l fluorides supplied by Morita, e x c e p t for N a F w h i c h was supplied by B D H . The g l a s s - m a k i n g process was the s a m e as that o f Z h o u et al [1], the only difference b e i n g that the g l a s s - m a k i n g process and the electroc h e m i c a l purification process w e r e carried out in the s a m e glassy carbon crucible, to reduce the likelihood o f c o n t a m i n a t i o n during transfer o f glass f r o m one crucible to the other. A t w o - e l e c t r o d e cell was e m p l o y e d for electroc h e m i c a l purification. T h e w o r k i n g electrodes w e r e either a 10 × 10 m m 2 platinum plate, a 3 m m diameter s p e c t r o s c o p i c carbon graphite rod or a 4 m m d i a m e t e r glassy carbon rod. The glassy carbon crucible was used as the counter electrode. The crucible was placed in a resistance furnace. This apparatus was set up in a dry box under a nitrogen atmosphere. The instrument used for controlling the deposition process was a p o l a r o g r a p h i c analyser ( A M E L 433-A). Rotation of the w o r k i n g electrode at a constant and reproducible rate was a c h i e v e d by using a digital
[
l xx I
Analyser
I
m
I;-
Speed
Controller
Computer :-
I
-~Thermometer
-1 I
I
-~
Variac
Fig. 1. Schematic diagram of the apparatus used for electrochemical melt purification.
m o t o r speed controller. The furnace temperature was adjusted with a Variac, and was m e a s u r e d with a t h e r m o c o u p l e attached to a digital thermometer. A s c h e m a t i c diagram o f the entire apparatus is s h o w n in Fig. 1. B e f o r e the purification process, the temperatures of the melt and furnace w e r e calibrated. It was found that the temperature o f the melt was constant approximately 45 min after the temperature o f the furnace reached the set value. In the treatment process, the crucible c o n t a i n i n g the glass was placed into the furnace; the temperature was increased to a set value and maintained for 1 h to obtain a uniform melt temperature; then a j a c k was used to raise the fur-
Table 1 Electrochemical purification conditions Melt
Temperature ( + 2) (°C)
1
.
2 3 4 5 6 7 8 9 10 11 12 13 14
600 600 600 600 600 600 600 600 600 600 560 640 600
.
Rotation rate ( _+3) (rpm) .
0 300 600 300 300 300 900 1200 1500 1800 300 300 1200
.
Deposition time ( + 5) (s)
Deposition potential (mV)
Electrode
2700 3600 3600 3600 7200 7200 3600 3600 2400 2400 3600 3600 3600 × 2
- 1100 - 1100 1100 - 1100 - 1100 - 1100 - 1100 - 1100 1100 - 1100 - 1100 - 1100 1100
glassy carbon graphite graphite platinum platinum graphite graphite graphite graphite graphite graphite graphite graphite
.
196
S. Bao et aL /Journal of Non-Crystalline Solids 184 (1995) 194-199
Table 2 The influence of electrode materials on final impurity level achieved Melt Electrode Fe Cu Ni (ppm+ 10%) (ppm+ 10%) (ppm± 10%) 1
-
4.12
3 5 6 7
graphite platinum platinum graphite
1.80 2.03 1.51 2.07
1.01 0.34 0.37 0.30 0.32
1.62 0.78 0.75 0.58 0.60
that the reproducibility and detection limits for Fe, Cu and Ni were 4, 2 and 3 ppb, respectively. Sample 1 was a reference sample which was given no electrochemical treatment but which was otherwise handled identically.
3. Results
nace to the point that the melt contacted the working electrode. The furnace was than raised a further 11 mm. This allowed the surface of the working electrode to be submerged about 1 cm into the melt. At the end of the deposition process, the working electrode was removed from the melt while the deposition potential was maintained, in order to prevent the deposit from redissolving. The electrochemical parameters for purification are shown in Table 1. The purified samples were then quenched and analysed using graphite furnace atomic absorption spectroscopy (GFAAS) [3]. The GFAAS procedure has been described in detail elsewhere [3]: briefly, samples were digested and extracted prior to analysis. Blanks were run through the entire digestion, extraction and analysis procedure which indicated
The samples investigated and the conditions used are summarized in Table 2 along with the analytical results obtained. All of the impurities investigated were reduced after the purification process. Comparison with the reference sample (sample 1) shows that an improvement in Fe, Ni and Cu levels of approximately a factor of three was obtained. From the data in Table 2, it appears that there is not a large influence of the type of electrode material used in the purification (graphite or platinum). The graphite rod electrode was found to be the easiest to clean after each purification process and most cost effective, therefore most of the experiments were conducted with this type of electrode. The rotation rate of the electrode was found to have an effect on purification. Under stationary conditions, the metal ion concentrations remaining in the melt were almost the same as the original glass. By
Table 3 The influenceof rotation rate and time on final impurity level achieved Melt
Rotation rate (rpm)
Time (s)
Fe (ppm + 10%)
Cu (ppm + 10%)
Ni (ppm + 10%)
2 3 4 8 9 10 11 5 6
0 300 600 900 1200 1500 1800 300 300
2700 3600 3600 3600 3600 2400 2400 3600 7200
4.29 1.80 1.73 1.53 1.35 2.07 1.30 2.03 1.51
1.16 0.34 0.25 0.36 0.29 0.32 0.26 0.37 0.30
2.00 0.78 0.69 0.68 0.63 0.81 0.48 0.75 0.58
Table 4 The influenceof temperature on final impurity level achieved Melt Temperature (°C) Fe (ppm + 10%)
Cu (ppm + 10%)
Ni (ppm + 10%)
12 3 13
0.37 0.34 0.23
0.75 0.78 0.62
560 600 640
1.87 1.80 1.91
198
S. Bao et al. /Journal of Non-Crystalline Solids 184 (1995) 194-199
electroactive species is transported by migration; diffusion is expected to be the main mechanism by which these impurity species reach the electrode [5]. Thus the deposition rate depends on the diffusion coefficient of the ions, their initial concentration (or their concentration beyond the diffusion layer) and the diffusion layer thickness. Among these factors, diffusion layer thickness is the main parameter which can be adjusted easily in a purification process. Under stirring conditions, the transfer of ions beyond the diffusion layer is controlled by convection and the transfer in the narrow stagnant diffusion layer is controlled by diffusion. From Fick's law, the concentration gradient determines the diffusion rate. The thinner the diffusion layer, the larger the concentration gradient and in turn the larger the diffusion rate will be. The diffusion layer thickness, 6,0, is dependent on the rotation rate of the electrode as expressed by the empirical equation [5] 6,0 = 6o(O ",
(1)
where to is the angular velocity of the rotating electrode, 60 is a constant and n = 0.5. Hence, the deposition rate is predicted to increase with increasing rotation rate of the electrode. The concentration gradient also depends on the concentration beyond the diffusion layer, this decreasing gradually during the purification process. Thus, the deposition rate would be expected to decrease gradually during the process, as is observed in this work. Under steady-state conditions, i.e., where the concentration gradient does not change with time, the quantity deposited is proportional to (oo5 for fixed deposition time, as was demonstrated by Zhou et al. [1]. In the purification process, where the impurity concentration in the melt decreases gradually and thus the concentration gradient in the diffusion layer changes with time, the total deposited mass will depend on to in a more complex way (Table 3).
4.3. The influence of temperature Whether the temperature is expected to influence the purification process depends on how large the corresponding diffusivity and ionic conductivity changes are. The ZBLAN20 melt in the range 560-
640°C has a viscosity and conductivity of the order of 10 1 p and 10 -1 (1~ cm) -1, respectively [7]. Over this limited temperature range, both parameters are not strong functions of temperature. The results given in Table 4 indicate slight improvement in impurity levels obtained at the highest temperature (except for Fe, where it is suspected that the crucible may be leaching Fe during the process).
4.4. Purification repetition and solubility The deposition process has two distinct phases. The first stage is the formation of nuclei of the metal being deposited on the electrode. This stage is important, for it determines the primary structure of the deposit layer and thereby influences the structure of the final electroplate [8]. The formation of nuclei occurs on the electrode surface sparsely, in very dilute solutions, because it requires a high local supersaturation, and the nuclei whose radii are smaller than the critical size, if formed, will redissolve at once. Thus, very dilute solutions tend to give incoherent, powdery deposits [9]. In the second stage, the growth of the crystals on these nuclei forms a porous deposit layer, whose surface area does not equal the surface area of the electrode and increases with the growth of the crystals. Since the dissolution rate is proportional to the surface area of the deposit, the dissolution rate is thus expected to increase with extent of deposition time. The dissolution of the deposited metal is an adverse process from the point of view of purification. Initially, the dissolution rate is zero, and the deposition rate is at a maximum. Deposition results in a decrease in the concentration of metal ions in the melt and so the deposition rate is also reduced, while the dissolution rate of the deposited metal increases. At some point in time, the two processes can be expected to reach equilibrium. Further prolonging the deposition time would not reduce the remaining metal concentration in the melt. In this case, replacing the old electrode with a new one (Table 5) revitalises the deposition process, and establishes a new equilibrium at a lower impurity concentration. The thermodynamic solubility of the deposit metal in a melt is a further important factor in the electrochemical process. When the solubility of the metal is
s. Bao et al. /Journal of Non-Crystalline Solids 184 (1995) 194-199
higher, the balance between deposition and dissolution will leave higher concentrations of transition metals in the melt. In the present work, the Cu could not be reduced further than 0.25-0.30 ppm; the reason may be that the concentration of Cu cannot reach supersaturation under the present electrochemical conditions.
5. Conclusions (i) The impurity levels of Fe, Cu and Ni in zirconium-based fluoride glasses can be reduced to 0 . 3 - 0 . 5 ppm by electrochemical purification of the melt. (ii) The rotation rate of the electrode, deposition time and number of repetitions are the main factors which determine the extent of purification. (iii) The behaviour of the process can be rationalized in terms of a mechanism involving rate-limiting diffusion of the impurity across the Nernst layer and nucleation and growth of the metal on the electrode.
199
References [1] (a) Zhou Zhiping, P.J. Newman and D.R. MacFarlane, J. Non-Cryst. Solids 161 (1993) 36; (b) Z. Zhou, P.J. Newman, D.K.Y. Wong and D.R. MacFarlane, J. Non-Cryst. Solids 140 (1992) 297; (c) Z. Zhou and D.R. MacFarlane, J. Non-Cryst. Solids 140 (1992) 215; (d) Z. Zhou, P.J. Newmanand D.R. MacFarlane,J. Non-Cryst. Solids 161 (1993) 27. [2] P.J. Melling and O.H. El-Bayoumi, Proc. Int. Soc. Optical Eng. 1761 (1992) 298. [3] P.J. Newman, A.T. Voelkel and D.R. MacFarlane, these Proceedings, p. 324. [4] E. Gileadi, Electrode Kinetics (VCH, New York, 1993) pp. 4-55. [5] L. Antropov, Theoretical Electrochemistry (Mir, Moscow, 1972) pp. 327-329. [6] R.C. Weast, ed., CRC Handbook of Chemistry and Physics, 65th Ed. (CRC, Cleveland,OH, 1984) pp. D155-162. [7] W.C. Hasz and C.T. Moynihan, J. Non-Cryst. Solids 140 (1992) 285. [8] D. Pletcher, Industrial Electrochemistry (Chapman, London, 1982) pp. 176-180. [9] A.J. Allmand, The Principles of Applied Electrochemistry (Arnold, London, 1924) pp. 120-124.