- Email: [email protected]
Tungsten recovery from alkaline leach solutions as synthetic scheelite
Hydrometallurgy 85 (2007) 110 – 115 www.elsevier.com/locate/hydromet
Tungsten recovery from alkaline leach solutions as synthetic scheelite J.I. Martins a,⁎, José L.F.C. Lima b , A. Moreira a , S.C. Costa a a
Faculdade de Engenharia, Departamento de Engenharia Química, Universidade do Porto, Rua Roberto Frias, 4200-465 Porto, Portugal b REQUIMITE, Departamento de Química Física, Faculdade de Farmácia, Universidade do Porto, Rua Aníbal Cunha 164, 4099-030 Porto, Portugal Received 9 May 2006; received in revised form 12 August 2006; accepted 17 August 2006 Available online 25 September 2006
Abstract The recovery of tungsten from alkaline leach solutions has been studied examining the effect of temperature, pH, Ca/WO3 molar ratio and nature of the precipitated calcium tungstate. The precipitation kinetics of calcium tungstate, upon the addition of aqueous sodium tungstate to calcium solutions, was followed by potentiometric measurements using a calcium ion-selective electrode. Two models, a crystal growth model and a second-order reaction opposed by zero-order reaction, have been used to test the experimental data. Both models show that the apparent activation energy of CaWO4 precipitation falls in the range 58 to 67 kJ mol− 1. The kinetic data shows that the maximum recovery of precipitated calcium tungstate occurs at pH ≥ 8.5 with a 10% excess of CaCl2 at 50 °C over a period of 20 min using sodium tungstate solutions of 100 g L− 1 and 150 g L− 1 WO3. © 2006 Elsevier B.V. All rights reserved. Keywords: Calcium tungstate; Synthetic scheelite; Kinetics; Crystal growth
1. Introduction Tungsten is one of the most important refractory metals used in several industrial applications taking into account its hardy and heat-resistant performance in the form of tungsten carbide. The main sources of tungsten are the high-grade concentrates of wolframite and scheelite ores. Currently, synthetic scheelite is a strong contender as a third source of tungsten. Under normal circumstances the medium-grade concentrates of tungsten have reduced commercial value and are difficult to place in the market (Martins et al., 2003; Martins, 2003). ⁎ Corresponding author. Fax: +351 22 5081449. E-mail address: [email protected] (J.I. Martins). 0304-386X/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.hydromet.2006.08.007
One of the methods to refine these products is through their conversion to ortho-tungstate in alkaline medium, followed by its precipitation to calcium tungstate by addition of calcium salts or solutions. The different methodologies (Delmas and Casquinha, 1976; Burwell, 1955; Fruchter and Moscovici, 1982) used by industry to obtain synthetic scheelite have varying costs and mainly affect the grain size of tungstic acid (Martins et al., 2002). Hence there is interest in the kinetics and crystal growth of its precipitation. The precipitation reaction kinetics depends on the predominance of a slow step in the mechanism of crystal growth. If the equilibrium at the crystal surface is rapid, the transport of the lattice ions from solution to the primary nuclei will be the rate-limiting step. The rate of
J.I. Martins et al. / Hydrometallurgy 85 (2007) 110–115
crystal growth is then proportional to the supersaturation, through Fick's law of diffusion. Conversely, if the interface controls the reaction, the rate of growth is proportional to the square of the supersaturation. Many precipitation reactions follow this latter mechanism, such as calcium carbonate (Nancollas and Reddy, 1971; Reddy, 1977) and calcium sulphate (Liu and Nancollas, 1971). The kinetics of crystal growth of calcium tungstate from sodium tungstate melts have been performed by Roy and Appalasami (1978). They found that the reaction is rate-controlled by the development rate of excess solute concentration. In the case of aqueous solutions, no work appears to have been performed on the kinetics of calcium tungstate, except for the synthesis and characterization of calcium tungstate nanocrystals for application as luminescent material (Chen et al., 2003; Sun et al., 2006). This paper reports the kinetics of calcium tungstate precipitation from alkaline sodium solutions and the influence of temperature, supersaturation, pH and nature of calcium tungstate on tungsten recovery using a calcium ion-selective electrode. This potentiometric technique has the advantage of being non-destructive and measures high to low concentrations directly.
111
technique (Moody et al., 1970; Davies et al., 1972). The sensor composition was 0.17 g of PVC powder and 0.40 g of ion exchanger Orion (92-20-01). The sensor solution immobilized in PVC was a mixture of calcium bis-di(n-decyl)phosphate as ion exchanger and di-n-octylphenyl-phosphonate as solvent mediator and plasticizer. The ion-selective electrode was calibrated by increasing the calcium ion concentration stepwise from zero to 14.7 mM and evaluating the respective potentials over three consecutive days. The prepared electrodes showed a slope between 28 and 33 mV and a linear response from 0.05 mM Ca2+ using an Orion millivoltmeter model 801A. 2.3. Equipment
2. Experimental
To obtain the kinetic data the reaction was performed in a glassware with the solution stirred magnetically and monitored for pH and calcium ion concentration. Calcium tungstate was then precipitated from sodium tungstate solutions of 100 g L− 1 and 150 g L− 1 WO3 using 150 g L− 1 and 250 g L− 1 calcium chloride solutions in a 1-L batch reactor thermostatically controlled at 50 ± 1 °C and stirred with a pitched blade turbine. Scanning electron microscopy micrographs were obtained on a JEOL JSM-6301F instrument.
2.1. Reagents and solutions
2.4. Experimental procedure
All chemicals were analytical grade from Merck and used without further purification. Orion (92-20-01) was the ion exchanger used to prepare the calcium ion-selective membrane electrodes. All solutions were prepared by diluting a stock concentrate. Since the ionic strength influences the kinetics and morphology of crystal growth (Nancollas and Liu, 1975; Nielsen, 1964) and electrode potential, potassium nitrate was added to maintain a constant ionic strength. Since the ion selective electrode shows a linear response between 0.001 mM to 100 mM Ca2+, the initial concentration of calcium ion was chosen as 10 mM, and the ionic strength was fixed as 0.1 M. The titration of the calcium and tungstate concentrated solutions was performed by potentiometry using a solution of EDTA (0.1 M) and lead perchlorate (0.1 M) (Orion 94-82-05), together with calcium and lead ion selective electrodes, respectively.
Since sodium ions can act as interfering species in calcium ion selective electrodes it was necessary to evaluate the extent of sodium ion interference in the potentiometric measurements performed with the constructed electrodes. The calcium ion-selective electrode potential, E, in the presence of sodium ion is given by the modified Nernst equation,
2.2. Preparation of selective electrode The construction of the calcium ion-selective electrode was performed according to Moody and Thomas'
E ¼ const þ
RT pot ln½aCa þ kCa;Na ðaNa Þ2 2F
ð1Þ
pot where is kCa,Na the potentiometric selectivity coefficient. The interference of sodium ion was analysed according to the Srinivasan and Rechnitz (1969) method E/RT by plotting eΔ against (aNa)2 / (aCa). The calculated pot kCa,Na value at 0.1 M ionic strength was 0.011 which fully agrees with literature data (Moody et al., 1978; Ammann et al., 1975). We conclude that the constructed electrode picks up interference from Na+ when the [Na+]/[Ca2+] ratio is higher or equal than 135, where the yield of the precipitation reaction is already about 99%. Nevertheless, to avoid any memory effect of the membrane
112
J.I. Martins et al. / Hydrometallurgy 85 (2007) 110–115
where k1 is the constant rate and C and Ce, respectively, the reactant molar concentrations at time t and at equilibrium. Let us consider the reactant concentration expressed in terms of the conversion x in Eq. (4). Integration gives the following relation ln y ¼ 2k1 Co ð1−xe Þt
ð5Þ
where y¼ Fig. 1. Calcium concentration vs time for precipitation at 10 °C, 15 °C, 25 °C and 40 °C and pH = 8 (stirring speed 200 min− 1, at CaCl2 14.7 mM; Na2WO4 14.7 mM).
to sodium ion, a new membrane was used for each experiment. However, the membranes may be recuperated by immersion in 0.1 M Ca2+ solution over 3 days. In the trials performed on a larger scale, either both reactants were added simultaneously, or calcium chloride solution was added to the bulk sodium tungstate solution. Reaction time was 20 min in both procedures. The filtered solid residue was washed with hot water to eliminate chloride and the filtrate analyzed for WO3 by colorimetric and atomic absorption spectrometry methods. 3. Description of theoretical models: Establishment of kinetics law of precipitation The precipitation of calcium tungstate from sodium tungstate solutions by calcium chloride solutions is given by the following reaction: k1
CaCl2ðaqÞ þ Na2 WO4ðaqÞ p Y k2 CaWO4 ðsÞ þ 2NaClðaqÞ
ð2Þ
xe ð2−x−xe Þ ðxe −xÞð2−xe Þ
ð6Þ
Looking at the reaction (3) according to the models proposed for precipitation and crystal growth based on the theory of surface controlled reactions, the equation derived by Reich and Kahlweit (1968) under conditions of low supersaturation is: −
dC ¼ kr ðC−Ce Þ2 dt
ð7Þ
where kr is the rate constant for a surface-controlled reaction, which integration gives the following relation: ðC−Ce Þ−1 ¼ ðCo −Ce Þ−1 þ kr t
ð8Þ
Thus at low supersaturation the growth rate dependence on supersaturation is parabolic. 4. Results and discussion 4.1. Kinetics reaction In Fig. 1, a typical plot of calcium concentration against time is shown for the precipitation of calcium tungstate at temperatures of 10, 15, 25 and 40 °C and pH = 8.
or by its net ionic equation: k1
2− p CaWO4 ðsÞ Ca2þ k2 ðaqÞ þ WO4ðaqÞ Y
ð3Þ
Tungstate precipitation can be seen as a second-order reaction opposed by a zero-order reaction since the product is a solid. Considering both reactants as having the same 2+ and B as initial concentration and designating A as Ca(aq) 2WO4(aq) , the rate of reaction can be written as follows: −
dC ¼ k1 ðC 2 −Ce2 Þ dt
ð4Þ
Fig. 2. Plots of Eq. (8) for the results of Fig. 1 at temperatures of 10 °C, 15 °C, 25 °C and 40 °C.
J.I. Martins et al. / Hydrometallurgy 85 (2007) 110–115
Fig. 3. Plots of Eq. (5) for the results of Fig. 1 at temperatures of 10 °C, 15 °C, 25 °C and 40 °C.
The experimental data was tested considering a surface-controlled reaction for the models represented by Eqs. (4) and (7). The model 1, based on the principles of crystal growth, represents satisfactorily the tungstate precipitation, which has been found to hold for a number of sparingly soluble 2-2 electrolytes (Nancollas and Reddy, 1971; Reddy, 1977; Liu and Nancollas, 1971). The integrated form of Eq. (7) is shown in Fig. 2 where the initial calcium ion concentration is 14.7 mM. This may be compared to the solubility values at 10, 15, 25 and 40 °C of 1.13 × 10− 4, 1.07 × 10− 4, 9.33 × 10− 5, and 8.35 × 10− 5 mol L− 1, respectively. It is notable that this equation explains the precipitation at 25 °C and 40 °C with a correlation factor of 0.99, but at temperatures of 10 °C and 15 °C an induction time of about 150 s had to be allowed, to adjust the experimental data to obtain a correlation factor of 0.98. During induction, the primary nucleation takes place resulting in small nuclei. For temperatures of 25 °C and 40 °C the initial
113
Fig. 5. SEM micrographs of the calcium tungstate particles obtained after 700 s of precipitation. Magnification ×5500 (CaCl2 14.7 mM; Na2WO4 14.7 mM; at pH = 8 and 25 °C).
rapid growth surges may be associated with secondary nucleation at the beginning of the reaction, as has been observed in other 2-2 electrolytes growth experiments. The fitting of the experimental data according to the model 2, based on the second-order reaction opposed by a zero-order reaction, is presented in Fig. 3. The results show that the model is adjustable during the time at temperatures of 25 °C and 40 °C with a correlation factor of 0.99, but at temperatures of 10 °C and 15 °C this fits only after about 200 s and 100 s induction time, respectively. The calculation of apparent activation energy with the Arrhenius equation provides a further criterion for determining the controlling step of the reaction. The Arrhenius plot constructed for models 1 and 2, shown in Fig. 4, gives, respectively, apparent activation energies of 67 kJ mol− 1 and 58 kJ mol− 1. These values are considerably larger than the value expected for a pure bulk diffusion-controlled reaction of about 20 kJ mol− 1. In Fig. 4, k = kr (mol− 1 L s− 1) for model 1, and k = 2k1 Co (1 − xe) (s− 1) for model 2. Further evidence for an interfacial mechanism as the rate-step controlling the reaction is provided by the insensitivity of the yield induced by stirring speed rates higher than 200 min− 1 in the precipitation cell. Scanning electron micrographs of the crystals, Fig. 5, clearly show the presence of secondary nucleation and the tendency for particle agglomeration at the end of the growth experiments due to low supersaturation. 4.2. Reaction parameters
Fig. 4. Arrhenius plot for the precipitation of calcium tungstate using Eq. (5), model 2, and Eq. (8), model 1: k = kr (mol− 1 L s− 1) for model 1, and k = 2 k1 Co (1 − xe) (s− 1) for model 2.
Many factors can affect the precipitation yield of tungstate by calcium ions from leaching solutions including particle size, the calcium salt, excess of reagent, temperature, pH, reaction time, and stirring. This study is only concerned with the reaction yield, since the effect of reaction parameters on the particle size has already
114
J.I. Martins et al. / Hydrometallurgy 85 (2007) 110–115 Table 2 Influence of excess calcium chloride and pH on tungsten recovery from sodium solutions at 50 °C in a 1-L batch reactor Molar ratio Na2WO4 CaCl2 Stirring speed pH % WO3 CaCl2/WO3 (g.L− 1 WO3) (g.L− 1) (min− 1) recovery 1.1 a 1.1 a 1.1 a 1.1 a 1.1 a 1.1 a 1.15 b 1.15 b 1.2 b 1.2 b
Fig. 6. Precipitation yield of calcium tungstate vs time at 25 °C, using excess calcium chloride into 30 mM sodium tungstate solutions.
been described (Martins et al., 2002). In particular, whilst the recovery is enhanced by increasing temperature (Fig. 1), which is in agreement with the endothermic character of the precipitation reaction, Martins et al. (2002) showed that the grain size of synthetic scheelite increases with temperature. Fig. 6 shows the importance of excess reactant on the yield of the precipitation reaction at 25 °C with a stirring speed of 200 min− 1. In this case, a 20% excess of CaCl2 is required to precipitate practically all the tungstate after about 15 min of reaction. Table 1 shows the significant importance of pH on the tungsten recovery. Upon acidification of a solution containing the WO42− ion, polyanions develop very quickly and the velocity increases remarkably with concentration. One way of rationalising the formation of isopolytungstates is to regard them as being produced by the addition of acid to the ortho-tungstate ion. The general equation can be written as follows: 2− n− aHþ ðaqÞ þ bWO4ðaqÞ X cHz Wx OyðaqÞ þ dH2 OðlÞ
ð9Þ
Apart from the common tungstates, the para-tung5− states (a/b = 7/6, HW6O21 , pH ≈ 6) and meta-tungstates Table 1 Influence of pH on tungsten recovery from sodium solutions at 25 °C and 50 °C by potentiometric technique Temperature (°C)
pH
WO3 recovery (%)
25 25 25 50 50 50
5.0 6.0 7.0 5.0 6.0 7.0
3.20 8.20 25.1 5.02 15.8 35.2
Molar ratio CaCl2/WO3 = 1.0; time = 20 min; stirring rate 200 min− 1.
a b
100 100 100 150 150 150 100 150 150 150
150 150 150 250 250 250 250 250 250 150
200 200 200 200 300 500 200 200 200 200
8.5 7 6 8.5 8.5 8.5 8.5 8.5 8.5 8.5
99.9 39.2 17.3 99.9 99.8 99.8 99.8 99.8 99.8 99.9
CaCl2 solution is added to Na2WO4 solution. Both CaCl2 and Na2WO4 solutions are added to the reactor.
6− , pH ≈ 4) have been known for a (a/b = 3/2, H2W12O40 long time. Our results present a poor yield on tungsten recovery at pH between 5 and 7 that is explained by the increasing para-tungstate formation at lower pH. Duncan and Kepert (1962) show that in the case of 25 mM tungstate solution, the ratio between the tungsten atoms as normal tungstate and the sum of normal tungstate plus paratungstate is 0.02 and 0.15 at pH 5 and pH 6, respectively. The use of 15% excess of Ca(OH)2 to precipitate tungstate at 50 °C over 20 min has shown the presence of 1, 2.4 and 3% free CaO as impurity in the tungstate for pH 8, 10 and 11, respectively. Table 2 shows the results obtained on a larger scale using a 1-L batch reactor at 50 °C. The data matches very well with the “micro” scale study based on the potentiometric technique to follow the calcium concentration. It is important to note the insensitivity of the reaction to stirring speed higher than 200 min− 1, the poor recovery for solutions with pH lower than 8, and the total recovery of tungstate after 20 min of reaction using an excess of 10% of calcium chloride. The calcium tungstate obtained at pH = 8.5 was easily decanted and washed, and was practically free of foreign cations that normally adhere to the tungsten salt precipitate.
5. Conclusions The yield of calcium tungstate precipitation is dependent on pH, temperature and excess of calcium reactant. The main conclusions to emerge from this work are as follows: 1) the proposed kinetics equations are identical for C ≫ Ce, so the kinetics findings in the paper present similar results; 2) the experimental results fit both models at temperatures of 25 and 40 °C; 3) the
J.I. Martins et al. / Hydrometallurgy 85 (2007) 110–115
proposed models do not satisfactorily explain the kinetic relationship at temperatures of 10 and 15 °C due to the presence of an induction time linked to the nucleation; 4) the apparent activation energy of reaction is of the order of 58–67 kJ mol− 1; 5) the rate is independent stirring above 200 min− 1; 6) by decreasing pH from 8 to 5 the tungsten recovery is reduced because of the formation of para-tungstate ions; 7) the precipitation can be carried out in industry with a very good recovery over 20 min under the following conditions: pH ≥ 8.5; 10% excess of CaCl2; temperature 50 °C; stirring speed N 200 min− 1. Acknowledgements We acknowledge the Portuguese Science and Technology Foundation (FCT) for the Project POCTI/ECM/ 39646. References Ammann, D., Bissig, R., Güggi, M., Pretsch, E., Simon, W., Borowitz, I.J., Weiss, L., 1975. Preparation of neutral ionophores for alkali and alkaline-earth metal cations and their application in ionselective membrane electrodes. Hel. Chim. Acta 58, 1535–1548. Burwell, B.T., 1955. Synthetic scheelite. Min. World 8, 44–49. Chen, Shu-Jian, Li, Jing, Chen, Xue-Tai, Hong, Jian-Ming, Xue, Ziling, You, Xiao-Zeng, 2003. Solvothermal synthesis and characterization of crystalline CaWO4 nanoparticles. J. Cryst. Growth 253, 361–365. Davies, J.E.W., Thomas, J.D.R., Moody, G.J., 1972. Nitrate ion selective electrodes based on poly(vinyl-chloride) matrix membranes. Analyst 97, 87–104. Delmas, F., Casquinha, H., 1976. Chemical treatment of wolframites. Int. Rep. Portuguese Atomic Energy Comission/LFEN. Duncan, J.F., Kepert, D.L., 1962. Polyanion equilibrium in aqueous solution. 2. Thermodynamic study of paratungstate anion. J. Chem. Soc. 205–214.
115
Fruchter, I.M., Moscovici, A., 1982. Process for the recovery of tungsten in a pure form from tungsten-containing materials. Israel Patent 65958. Liu, S.T., Nancollas, G.H., 1971. Kinetics dissolution of calcium sulfate dihydrate. J. Inorg. Nucl. Chem. 33, 2295–2311. Martins, J.I., 2003. Leaching of synthetic scheelite by nitric acid without the formation of tungstic acid. Ind. Eng. Chem. Res. 42, 5031–5036. Martins, J.I., Delmas, F., Casquinha, H., Costa, S., 2002. Influence of particle size of raw materials on the final WO3 grain size. Hydrometallurgy 67, 117–123. Martins, J.I., Moreira, A., Costa, S.C., 2003. Leaching of synthetic scheelite by hydrochloric acid without the formation of tungstic acid. Hydrometallurgy 70, 131–141. Moody, G.J., Oke, R.B., Thomas, J.D.R., 1970. Calcium-sensitive electrode based on a liquid ion exchanger in a poly vinyl chloride matrix. Analyst 95, 910–918. Moody, G.J., Nassory, N.S., Thomas, J.D.R., 1978. Calcium ionselective electrodes based on bis[di(P-1,1,3,3-tetramethylbutylphenyl)phosphate] sensor and trialkyl-phosphate mediators. Analyst 103, 68–71. Nancollas, G.H., Liu, S.T., 1975. Crystal growth and dissolution of barium sulfate. Soc. Pet. Eng. J. 15, 509–516. Nancollas, G.H., Reddy, M.M., 1971. The crystallization of calcium carbonate. 2. Calcite growth mechanism. J. Colloid Interface Sci. 37, 824–829. Nielsen, A.E., 1964. The Kinetics of Precipitation. Pergamon, New York. Reddy, M.M., 1977. Crystallization of calcium carbonate in the presence of trace concentrations of phosphorus-containing anions. J. Cryst. Growth 41, 287–295. Reich, R., Kahlweit, M., 1968. Kinetics of crystal growth in aqueous solutions. 1. Ber. Bunsenges. Phys. Chem. 72, 66–69. Roy, B.N., Appalasami, S., 1978. Kinetics of crystal growth of calcium tungstate from solutions in sodium tungstate melts by continuous cooling. J. Am. Ceram. Soc. 61, 38–41. Srinivasan, K., Rechnitz, G.A., 1969. Selective studies on liquid membrane, ion-selective electrodes. Anal. Chem. 41, 1203–1208. Sun, Lingna, Cao, Minhua, Wang, Yonghui, Sun, Genban, Hu, Changwen, 2006. The synthesis and photoluminescent properties of calcium tungstate nanocrystals. J. Cryst. Growth 289, 231–235.
Tungsten recovery from alkaline leach solutions as synthetic scheelite J.I. Martins a,⁎, José L.F.C. Lima b , A. Moreira a , S.C. Costa a a
Faculdade de Engenharia, Departamento de Engenharia Química, Universidade do Porto, Rua Roberto Frias, 4200-465 Porto, Portugal b REQUIMITE, Departamento de Química Física, Faculdade de Farmácia, Universidade do Porto, Rua Aníbal Cunha 164, 4099-030 Porto, Portugal Received 9 May 2006; received in revised form 12 August 2006; accepted 17 August 2006 Available online 25 September 2006
Abstract The recovery of tungsten from alkaline leach solutions has been studied examining the effect of temperature, pH, Ca/WO3 molar ratio and nature of the precipitated calcium tungstate. The precipitation kinetics of calcium tungstate, upon the addition of aqueous sodium tungstate to calcium solutions, was followed by potentiometric measurements using a calcium ion-selective electrode. Two models, a crystal growth model and a second-order reaction opposed by zero-order reaction, have been used to test the experimental data. Both models show that the apparent activation energy of CaWO4 precipitation falls in the range 58 to 67 kJ mol− 1. The kinetic data shows that the maximum recovery of precipitated calcium tungstate occurs at pH ≥ 8.5 with a 10% excess of CaCl2 at 50 °C over a period of 20 min using sodium tungstate solutions of 100 g L− 1 and 150 g L− 1 WO3. © 2006 Elsevier B.V. All rights reserved. Keywords: Calcium tungstate; Synthetic scheelite; Kinetics; Crystal growth
1. Introduction Tungsten is one of the most important refractory metals used in several industrial applications taking into account its hardy and heat-resistant performance in the form of tungsten carbide. The main sources of tungsten are the high-grade concentrates of wolframite and scheelite ores. Currently, synthetic scheelite is a strong contender as a third source of tungsten. Under normal circumstances the medium-grade concentrates of tungsten have reduced commercial value and are difficult to place in the market (Martins et al., 2003; Martins, 2003). ⁎ Corresponding author. Fax: +351 22 5081449. E-mail address: [email protected] (J.I. Martins). 0304-386X/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.hydromet.2006.08.007
One of the methods to refine these products is through their conversion to ortho-tungstate in alkaline medium, followed by its precipitation to calcium tungstate by addition of calcium salts or solutions. The different methodologies (Delmas and Casquinha, 1976; Burwell, 1955; Fruchter and Moscovici, 1982) used by industry to obtain synthetic scheelite have varying costs and mainly affect the grain size of tungstic acid (Martins et al., 2002). Hence there is interest in the kinetics and crystal growth of its precipitation. The precipitation reaction kinetics depends on the predominance of a slow step in the mechanism of crystal growth. If the equilibrium at the crystal surface is rapid, the transport of the lattice ions from solution to the primary nuclei will be the rate-limiting step. The rate of
J.I. Martins et al. / Hydrometallurgy 85 (2007) 110–115
crystal growth is then proportional to the supersaturation, through Fick's law of diffusion. Conversely, if the interface controls the reaction, the rate of growth is proportional to the square of the supersaturation. Many precipitation reactions follow this latter mechanism, such as calcium carbonate (Nancollas and Reddy, 1971; Reddy, 1977) and calcium sulphate (Liu and Nancollas, 1971). The kinetics of crystal growth of calcium tungstate from sodium tungstate melts have been performed by Roy and Appalasami (1978). They found that the reaction is rate-controlled by the development rate of excess solute concentration. In the case of aqueous solutions, no work appears to have been performed on the kinetics of calcium tungstate, except for the synthesis and characterization of calcium tungstate nanocrystals for application as luminescent material (Chen et al., 2003; Sun et al., 2006). This paper reports the kinetics of calcium tungstate precipitation from alkaline sodium solutions and the influence of temperature, supersaturation, pH and nature of calcium tungstate on tungsten recovery using a calcium ion-selective electrode. This potentiometric technique has the advantage of being non-destructive and measures high to low concentrations directly.
111
technique (Moody et al., 1970; Davies et al., 1972). The sensor composition was 0.17 g of PVC powder and 0.40 g of ion exchanger Orion (92-20-01). The sensor solution immobilized in PVC was a mixture of calcium bis-di(n-decyl)phosphate as ion exchanger and di-n-octylphenyl-phosphonate as solvent mediator and plasticizer. The ion-selective electrode was calibrated by increasing the calcium ion concentration stepwise from zero to 14.7 mM and evaluating the respective potentials over three consecutive days. The prepared electrodes showed a slope between 28 and 33 mV and a linear response from 0.05 mM Ca2+ using an Orion millivoltmeter model 801A. 2.3. Equipment
2. Experimental
To obtain the kinetic data the reaction was performed in a glassware with the solution stirred magnetically and monitored for pH and calcium ion concentration. Calcium tungstate was then precipitated from sodium tungstate solutions of 100 g L− 1 and 150 g L− 1 WO3 using 150 g L− 1 and 250 g L− 1 calcium chloride solutions in a 1-L batch reactor thermostatically controlled at 50 ± 1 °C and stirred with a pitched blade turbine. Scanning electron microscopy micrographs were obtained on a JEOL JSM-6301F instrument.
2.1. Reagents and solutions
2.4. Experimental procedure
All chemicals were analytical grade from Merck and used without further purification. Orion (92-20-01) was the ion exchanger used to prepare the calcium ion-selective membrane electrodes. All solutions were prepared by diluting a stock concentrate. Since the ionic strength influences the kinetics and morphology of crystal growth (Nancollas and Liu, 1975; Nielsen, 1964) and electrode potential, potassium nitrate was added to maintain a constant ionic strength. Since the ion selective electrode shows a linear response between 0.001 mM to 100 mM Ca2+, the initial concentration of calcium ion was chosen as 10 mM, and the ionic strength was fixed as 0.1 M. The titration of the calcium and tungstate concentrated solutions was performed by potentiometry using a solution of EDTA (0.1 M) and lead perchlorate (0.1 M) (Orion 94-82-05), together with calcium and lead ion selective electrodes, respectively.
Since sodium ions can act as interfering species in calcium ion selective electrodes it was necessary to evaluate the extent of sodium ion interference in the potentiometric measurements performed with the constructed electrodes. The calcium ion-selective electrode potential, E, in the presence of sodium ion is given by the modified Nernst equation,
2.2. Preparation of selective electrode The construction of the calcium ion-selective electrode was performed according to Moody and Thomas'
E ¼ const þ
RT pot ln½aCa þ kCa;Na ðaNa Þ2 2F
ð1Þ
pot where is kCa,Na the potentiometric selectivity coefficient. The interference of sodium ion was analysed according to the Srinivasan and Rechnitz (1969) method E/RT by plotting eΔ against (aNa)2 / (aCa). The calculated pot kCa,Na value at 0.1 M ionic strength was 0.011 which fully agrees with literature data (Moody et al., 1978; Ammann et al., 1975). We conclude that the constructed electrode picks up interference from Na+ when the [Na+]/[Ca2+] ratio is higher or equal than 135, where the yield of the precipitation reaction is already about 99%. Nevertheless, to avoid any memory effect of the membrane
112
J.I. Martins et al. / Hydrometallurgy 85 (2007) 110–115
where k1 is the constant rate and C and Ce, respectively, the reactant molar concentrations at time t and at equilibrium. Let us consider the reactant concentration expressed in terms of the conversion x in Eq. (4). Integration gives the following relation ln y ¼ 2k1 Co ð1−xe Þt
ð5Þ
where y¼ Fig. 1. Calcium concentration vs time for precipitation at 10 °C, 15 °C, 25 °C and 40 °C and pH = 8 (stirring speed 200 min− 1, at CaCl2 14.7 mM; Na2WO4 14.7 mM).
to sodium ion, a new membrane was used for each experiment. However, the membranes may be recuperated by immersion in 0.1 M Ca2+ solution over 3 days. In the trials performed on a larger scale, either both reactants were added simultaneously, or calcium chloride solution was added to the bulk sodium tungstate solution. Reaction time was 20 min in both procedures. The filtered solid residue was washed with hot water to eliminate chloride and the filtrate analyzed for WO3 by colorimetric and atomic absorption spectrometry methods. 3. Description of theoretical models: Establishment of kinetics law of precipitation The precipitation of calcium tungstate from sodium tungstate solutions by calcium chloride solutions is given by the following reaction: k1
CaCl2ðaqÞ þ Na2 WO4ðaqÞ p Y k2 CaWO4 ðsÞ þ 2NaClðaqÞ
ð2Þ
xe ð2−x−xe Þ ðxe −xÞð2−xe Þ
ð6Þ
Looking at the reaction (3) according to the models proposed for precipitation and crystal growth based on the theory of surface controlled reactions, the equation derived by Reich and Kahlweit (1968) under conditions of low supersaturation is: −
dC ¼ kr ðC−Ce Þ2 dt
ð7Þ
where kr is the rate constant for a surface-controlled reaction, which integration gives the following relation: ðC−Ce Þ−1 ¼ ðCo −Ce Þ−1 þ kr t
ð8Þ
Thus at low supersaturation the growth rate dependence on supersaturation is parabolic. 4. Results and discussion 4.1. Kinetics reaction In Fig. 1, a typical plot of calcium concentration against time is shown for the precipitation of calcium tungstate at temperatures of 10, 15, 25 and 40 °C and pH = 8.
or by its net ionic equation: k1
2− p CaWO4 ðsÞ Ca2þ k2 ðaqÞ þ WO4ðaqÞ Y
ð3Þ
Tungstate precipitation can be seen as a second-order reaction opposed by a zero-order reaction since the product is a solid. Considering both reactants as having the same 2+ and B as initial concentration and designating A as Ca(aq) 2WO4(aq) , the rate of reaction can be written as follows: −
dC ¼ k1 ðC 2 −Ce2 Þ dt
ð4Þ
Fig. 2. Plots of Eq. (8) for the results of Fig. 1 at temperatures of 10 °C, 15 °C, 25 °C and 40 °C.
J.I. Martins et al. / Hydrometallurgy 85 (2007) 110–115
Fig. 3. Plots of Eq. (5) for the results of Fig. 1 at temperatures of 10 °C, 15 °C, 25 °C and 40 °C.
The experimental data was tested considering a surface-controlled reaction for the models represented by Eqs. (4) and (7). The model 1, based on the principles of crystal growth, represents satisfactorily the tungstate precipitation, which has been found to hold for a number of sparingly soluble 2-2 electrolytes (Nancollas and Reddy, 1971; Reddy, 1977; Liu and Nancollas, 1971). The integrated form of Eq. (7) is shown in Fig. 2 where the initial calcium ion concentration is 14.7 mM. This may be compared to the solubility values at 10, 15, 25 and 40 °C of 1.13 × 10− 4, 1.07 × 10− 4, 9.33 × 10− 5, and 8.35 × 10− 5 mol L− 1, respectively. It is notable that this equation explains the precipitation at 25 °C and 40 °C with a correlation factor of 0.99, but at temperatures of 10 °C and 15 °C an induction time of about 150 s had to be allowed, to adjust the experimental data to obtain a correlation factor of 0.98. During induction, the primary nucleation takes place resulting in small nuclei. For temperatures of 25 °C and 40 °C the initial
113
Fig. 5. SEM micrographs of the calcium tungstate particles obtained after 700 s of precipitation. Magnification ×5500 (CaCl2 14.7 mM; Na2WO4 14.7 mM; at pH = 8 and 25 °C).
rapid growth surges may be associated with secondary nucleation at the beginning of the reaction, as has been observed in other 2-2 electrolytes growth experiments. The fitting of the experimental data according to the model 2, based on the second-order reaction opposed by a zero-order reaction, is presented in Fig. 3. The results show that the model is adjustable during the time at temperatures of 25 °C and 40 °C with a correlation factor of 0.99, but at temperatures of 10 °C and 15 °C this fits only after about 200 s and 100 s induction time, respectively. The calculation of apparent activation energy with the Arrhenius equation provides a further criterion for determining the controlling step of the reaction. The Arrhenius plot constructed for models 1 and 2, shown in Fig. 4, gives, respectively, apparent activation energies of 67 kJ mol− 1 and 58 kJ mol− 1. These values are considerably larger than the value expected for a pure bulk diffusion-controlled reaction of about 20 kJ mol− 1. In Fig. 4, k = kr (mol− 1 L s− 1) for model 1, and k = 2k1 Co (1 − xe) (s− 1) for model 2. Further evidence for an interfacial mechanism as the rate-step controlling the reaction is provided by the insensitivity of the yield induced by stirring speed rates higher than 200 min− 1 in the precipitation cell. Scanning electron micrographs of the crystals, Fig. 5, clearly show the presence of secondary nucleation and the tendency for particle agglomeration at the end of the growth experiments due to low supersaturation. 4.2. Reaction parameters
Fig. 4. Arrhenius plot for the precipitation of calcium tungstate using Eq. (5), model 2, and Eq. (8), model 1: k = kr (mol− 1 L s− 1) for model 1, and k = 2 k1 Co (1 − xe) (s− 1) for model 2.
Many factors can affect the precipitation yield of tungstate by calcium ions from leaching solutions including particle size, the calcium salt, excess of reagent, temperature, pH, reaction time, and stirring. This study is only concerned with the reaction yield, since the effect of reaction parameters on the particle size has already
114
J.I. Martins et al. / Hydrometallurgy 85 (2007) 110–115 Table 2 Influence of excess calcium chloride and pH on tungsten recovery from sodium solutions at 50 °C in a 1-L batch reactor Molar ratio Na2WO4 CaCl2 Stirring speed pH % WO3 CaCl2/WO3 (g.L− 1 WO3) (g.L− 1) (min− 1) recovery 1.1 a 1.1 a 1.1 a 1.1 a 1.1 a 1.1 a 1.15 b 1.15 b 1.2 b 1.2 b
Fig. 6. Precipitation yield of calcium tungstate vs time at 25 °C, using excess calcium chloride into 30 mM sodium tungstate solutions.
been described (Martins et al., 2002). In particular, whilst the recovery is enhanced by increasing temperature (Fig. 1), which is in agreement with the endothermic character of the precipitation reaction, Martins et al. (2002) showed that the grain size of synthetic scheelite increases with temperature. Fig. 6 shows the importance of excess reactant on the yield of the precipitation reaction at 25 °C with a stirring speed of 200 min− 1. In this case, a 20% excess of CaCl2 is required to precipitate practically all the tungstate after about 15 min of reaction. Table 1 shows the significant importance of pH on the tungsten recovery. Upon acidification of a solution containing the WO42− ion, polyanions develop very quickly and the velocity increases remarkably with concentration. One way of rationalising the formation of isopolytungstates is to regard them as being produced by the addition of acid to the ortho-tungstate ion. The general equation can be written as follows: 2− n− aHþ ðaqÞ þ bWO4ðaqÞ X cHz Wx OyðaqÞ þ dH2 OðlÞ
ð9Þ
Apart from the common tungstates, the para-tung5− states (a/b = 7/6, HW6O21 , pH ≈ 6) and meta-tungstates Table 1 Influence of pH on tungsten recovery from sodium solutions at 25 °C and 50 °C by potentiometric technique Temperature (°C)
pH
WO3 recovery (%)
25 25 25 50 50 50
5.0 6.0 7.0 5.0 6.0 7.0
3.20 8.20 25.1 5.02 15.8 35.2
Molar ratio CaCl2/WO3 = 1.0; time = 20 min; stirring rate 200 min− 1.
a b
100 100 100 150 150 150 100 150 150 150
150 150 150 250 250 250 250 250 250 150
200 200 200 200 300 500 200 200 200 200
8.5 7 6 8.5 8.5 8.5 8.5 8.5 8.5 8.5
99.9 39.2 17.3 99.9 99.8 99.8 99.8 99.8 99.8 99.9
CaCl2 solution is added to Na2WO4 solution. Both CaCl2 and Na2WO4 solutions are added to the reactor.
6− , pH ≈ 4) have been known for a (a/b = 3/2, H2W12O40 long time. Our results present a poor yield on tungsten recovery at pH between 5 and 7 that is explained by the increasing para-tungstate formation at lower pH. Duncan and Kepert (1962) show that in the case of 25 mM tungstate solution, the ratio between the tungsten atoms as normal tungstate and the sum of normal tungstate plus paratungstate is 0.02 and 0.15 at pH 5 and pH 6, respectively. The use of 15% excess of Ca(OH)2 to precipitate tungstate at 50 °C over 20 min has shown the presence of 1, 2.4 and 3% free CaO as impurity in the tungstate for pH 8, 10 and 11, respectively. Table 2 shows the results obtained on a larger scale using a 1-L batch reactor at 50 °C. The data matches very well with the “micro” scale study based on the potentiometric technique to follow the calcium concentration. It is important to note the insensitivity of the reaction to stirring speed higher than 200 min− 1, the poor recovery for solutions with pH lower than 8, and the total recovery of tungstate after 20 min of reaction using an excess of 10% of calcium chloride. The calcium tungstate obtained at pH = 8.5 was easily decanted and washed, and was practically free of foreign cations that normally adhere to the tungsten salt precipitate.
5. Conclusions The yield of calcium tungstate precipitation is dependent on pH, temperature and excess of calcium reactant. The main conclusions to emerge from this work are as follows: 1) the proposed kinetics equations are identical for C ≫ Ce, so the kinetics findings in the paper present similar results; 2) the experimental results fit both models at temperatures of 25 and 40 °C; 3) the
J.I. Martins et al. / Hydrometallurgy 85 (2007) 110–115
proposed models do not satisfactorily explain the kinetic relationship at temperatures of 10 and 15 °C due to the presence of an induction time linked to the nucleation; 4) the apparent activation energy of reaction is of the order of 58–67 kJ mol− 1; 5) the rate is independent stirring above 200 min− 1; 6) by decreasing pH from 8 to 5 the tungsten recovery is reduced because of the formation of para-tungstate ions; 7) the precipitation can be carried out in industry with a very good recovery over 20 min under the following conditions: pH ≥ 8.5; 10% excess of CaCl2; temperature 50 °C; stirring speed N 200 min− 1. Acknowledgements We acknowledge the Portuguese Science and Technology Foundation (FCT) for the Project POCTI/ECM/ 39646. References Ammann, D., Bissig, R., Güggi, M., Pretsch, E., Simon, W., Borowitz, I.J., Weiss, L., 1975. Preparation of neutral ionophores for alkali and alkaline-earth metal cations and their application in ionselective membrane electrodes. Hel. Chim. Acta 58, 1535–1548. Burwell, B.T., 1955. Synthetic scheelite. Min. World 8, 44–49. Chen, Shu-Jian, Li, Jing, Chen, Xue-Tai, Hong, Jian-Ming, Xue, Ziling, You, Xiao-Zeng, 2003. Solvothermal synthesis and characterization of crystalline CaWO4 nanoparticles. J. Cryst. Growth 253, 361–365. Davies, J.E.W., Thomas, J.D.R., Moody, G.J., 1972. Nitrate ion selective electrodes based on poly(vinyl-chloride) matrix membranes. Analyst 97, 87–104. Delmas, F., Casquinha, H., 1976. Chemical treatment of wolframites. Int. Rep. Portuguese Atomic Energy Comission/LFEN. Duncan, J.F., Kepert, D.L., 1962. Polyanion equilibrium in aqueous solution. 2. Thermodynamic study of paratungstate anion. J. Chem. Soc. 205–214.
115
Fruchter, I.M., Moscovici, A., 1982. Process for the recovery of tungsten in a pure form from tungsten-containing materials. Israel Patent 65958. Liu, S.T., Nancollas, G.H., 1971. Kinetics dissolution of calcium sulfate dihydrate. J. Inorg. Nucl. Chem. 33, 2295–2311. Martins, J.I., 2003. Leaching of synthetic scheelite by nitric acid without the formation of tungstic acid. Ind. Eng. Chem. Res. 42, 5031–5036. Martins, J.I., Delmas, F., Casquinha, H., Costa, S., 2002. Influence of particle size of raw materials on the final WO3 grain size. Hydrometallurgy 67, 117–123. Martins, J.I., Moreira, A., Costa, S.C., 2003. Leaching of synthetic scheelite by hydrochloric acid without the formation of tungstic acid. Hydrometallurgy 70, 131–141. Moody, G.J., Oke, R.B., Thomas, J.D.R., 1970. Calcium-sensitive electrode based on a liquid ion exchanger in a poly vinyl chloride matrix. Analyst 95, 910–918. Moody, G.J., Nassory, N.S., Thomas, J.D.R., 1978. Calcium ionselective electrodes based on bis[di(P-1,1,3,3-tetramethylbutylphenyl)phosphate] sensor and trialkyl-phosphate mediators. Analyst 103, 68–71. Nancollas, G.H., Liu, S.T., 1975. Crystal growth and dissolution of barium sulfate. Soc. Pet. Eng. J. 15, 509–516. Nancollas, G.H., Reddy, M.M., 1971. The crystallization of calcium carbonate. 2. Calcite growth mechanism. J. Colloid Interface Sci. 37, 824–829. Nielsen, A.E., 1964. The Kinetics of Precipitation. Pergamon, New York. Reddy, M.M., 1977. Crystallization of calcium carbonate in the presence of trace concentrations of phosphorus-containing anions. J. Cryst. Growth 41, 287–295. Reich, R., Kahlweit, M., 1968. Kinetics of crystal growth in aqueous solutions. 1. Ber. Bunsenges. Phys. Chem. 72, 66–69. Roy, B.N., Appalasami, S., 1978. Kinetics of crystal growth of calcium tungstate from solutions in sodium tungstate melts by continuous cooling. J. Am. Ceram. Soc. 61, 38–41. Srinivasan, K., Rechnitz, G.A., 1969. Selective studies on liquid membrane, ion-selective electrodes. Anal. Chem. 41, 1203–1208. Sun, Lingna, Cao, Minhua, Wang, Yonghui, Sun, Genban, Hu, Changwen, 2006. The synthesis and photoluminescent properties of calcium tungstate nanocrystals. J. Cryst. Growth 289, 231–235.