- Email: [email protected]
Dissolution reactivity of metal oxide mixtures containing ZnO
Hydrometallurgy 60 Ž2001. 17–24 www.elsevier.nlrlocaterhydromet
Dissolution reactivity of metal oxide mixtures containing ZnO T. Grygar a,) , Z. Klımova ´ ´ a,b, J. Jandova´ b b
a ˘ ˘ Czech Republic Institute of Inorganic Chemistry, Academy of Sciences of the Czech Republic, 250 68 Rez, Department of Metals and Corrosion Engineering, Prague Institute of Chemical Technology, Technicka´ 6, 166 28 Prague, Czech Republic
Received 15 December 1999; received in revised form 21 August 2000; accepted 9 September 2000
Abstract The kinetics of zinc dissolution from oxide mixtures in 0.5 M H 2 SO4 was studied. The oxide mixtures were prepared by acid leaching of galvanic sludge, precipitation of the extracts by NaOH followed by calcination of the produced Zn hydroxides. A series of experiments with synthetic oxide mixtures of zinc and other common metals were also performed. The phase composition of the oxide mixtures significantly affected the kinetics of the zinc leaching. The most reactive phases were ZnO and Zn 2 SiO4 , less reactive were substituted CuO and NiO, and the least reactive were spinel-type oxides. Due to the different rates of dissolution of individual oxide phases, the described process can be used for purification of Zn from the galvanic sludge. q 2001 Elsevier Science B.V. All rights reserved. Keywords: ZnO; Dissolution; Industrial waste
1. Introduction The problem of recycling metal-containing industrial waste is not satisfactorily solved in the Czech Republic. Even the neutralisation sludges from Zn electroplating and mechanoplating plants are not recycled, although they contain as much as 5–10% of Zn mainly in the form of hydroxides. Nowadays, these sludges are only deposited in landfills as dangerous waste. Partial dissolution of the oxide mixtures obtained by heating impure Zn hydroxides or basic salts was recently proposed by Jandova´ and Maixner w1x for the recovery of metals from Zn-containing poly-
)
Corresponding author. E-mail address: [email protected] ŽT. Grygar..
metallic galvanic sludge w2,3x. The basis of the separation of Zn is calcination of the mixed hydroxides to transform the metal impurities, especially Al, Cr and Fe, to the oxides, which dissolve much slower in acidic solution than ZnO and Zn 2 SiO4 . The list of oxides, which can be expected andror found in those mixtures w3x, together with their dissolution reactivities, w4–9x are given in Table 1. A similar procedure to extract Ni w4x and Cu w5x preferentially from oxide mixtures was already proposed. In contrast to NiO and the oxides of trivalent metals Al, Cr and Fe, the difference in reactivities of ZnO and Zn 2 SiO4 and the oxides of trivalent metals is much larger and, hence, Zn can be more selectively leached from the oxide mixture. The economical evaluation of this proposed treatment can only be done if more details were known about its applicability for particular impurities present in Zn sludges.
0304-386Xr01r$ - see front matter q 2001 Elsevier Science B.V. All rights reserved. PII: S 0 3 0 4 - 3 8 6 X Ž 0 0 . 0 0 1 5 3 - 5
18
T. Grygar et al.r Hydrometallurgy 60 (2001) 17–24
Table 1 The most common oxide phases formed by heating impure metal hydroxides and basic salts together with evaluation of their dissolution reactivity Žin solutions of HCl or H 2 SO4 . published in Refs. w3–9x Phase ZnO Žzincite. Zn 2 SiO4 Žwillemite. MgO Žpericlase. and ŽZn,Mg.O CuO Žtenorite. ZnFe 2 O4 Žfranklinite. NiO Žbunsenite. NiCr2 O4 a-ŽAl,Cr,Fe. 2 O 3
Dissolution behaviour
Reference
fast diffusion-controlled or chemical dissolution fast diffusion-controlled or chemical dissolution fast diffusion-controlled or chemical dissolution relatively fast chemical dissolution chemical dissolution chemical dissolution sensitive to actual chemical composition very slow chemical dissolution very slow chemical dissolution
w6x w7x w6,8x w5x w9x w4,6x w3x w3x
The aim of this work was to improve our knowledge about the phase composition of the oxide mixture obtained by calcining the contaminated Zn hydroxides and about the dissolution reactivity of the resulting oxides. The problem of the kinetics of the dissolution of the mixed oxides is a rather complex one w6x, and so the reactivity of the mixed oxides cannot be directly predicted from the reactivity of the pure components. The experimental data are rarely available for non-stoichiometric oxides. More experimental studies are, hence, necessary of the dissolution involving complex mixtures of nonstoichiometric oxides that are formed from polymetallic hydroxides.
2. Experimental procedures Synthetic oxide mixtures for dissolution studies were prepared by precipitation of mixtures of metal salts with NaOH solution followed by filtration, drying and calcining Ž10008C, 2 h. of the precipitate. Additionally, two samples of galvanic sludge were obtained from the landfill in Dolnı´ Rozınka in the ˘´ Czech Republic, and from a galvanic Zn plating plant in Letky in the Czech Republic. Each sludge
sample was homogenised and its composition was characterised by semi-quantitative emission spectral analysis and by atomic absorption spectroscopy. The approximate elemental composition is summarised in Table 2. The most important difference between these two samples is in the SirZn ratio, which is significantly increased in sludge A from the waste deposit due to the subsequent contamination of the sludge in the landfill. In sample A, the weight ratio of Si to Zn is 1.1:1. Solid samples were characterised by X-ray powder diffraction ŽSiemens D5005. and specific surface area measurement ŽCoulter SA3100, Beckman Coulter.. For the chemical analysis, the sludges and hydroxides were dissolved in diluted HCl. The insoluble residues after zinc leaching were decomposed by fusing them with NaOH or by evaporating with HF–H 2 SO4 to expel SiO 2 followed by acidic fusion of the residue ŽK 2 S 2 O 7 .. The chemical composition was then obtained by atomic absorption spectroscopy, gravimetry and chelatometric titration. The flowsheet of the sludge process used to obtain the oxide mixtures and refined zinc hydroxide is given in Fig. 1. The solution denoted the first extract in Fig. 1 was prepared by leaching the sludge in aqueous H 2 SO4 at pH s 3 or 3.9 in order to extract
Table 2 Approximate chemical composition of the original galvanic sludge Sample
10 0 –10 1 %
10y1 –10 0 %
10y2 –10y1 %
A Žwaste dump in Dolnı´ Rozınka ˘´ . B Žgalvanic plant in Letky.
Ca, Fe, Si, Zn Ca, Zn
Al, Ba, Mg, Mn, Cr, Cu, P, Pb, Ni Al, Fe, Mg, Mn, Si
B, Cd B, Sn
T. Grygar et al.r Hydrometallurgy 60 (2001) 17–24
19
Fig. 1. The flowsheet of the processing of the Zn-sludge.
the maximum amount of Zn on one hand, and to leave the majority of the trivalent metal impurities in the solid residue on the other hand. Four different kinds of solid phases were investigated: a raw Zn hydroxide obtained by neutralisation of the first extract to pH s 7, a calcined oxide mixture, an insoluble residue after the second leaching, and finally, a refined Zn hydroxide, obtained by the second precipitation to pH s 7. Dissolution kinetics was performed in a glass reactor at 408C in 0.5 M H 2 SO4 . Solid to liquid ratio of 1:40 was chosen. The suspension was stirred with a propeller to avoid settling. To monitor the dissolution rate, the suspension was sampled at appropriate times, filtered and its content of dissolved metals was determined by atomic absorption spectroscopy.
3. Results Leaching the galvanic sludge at pH as high as 3–3.9 led to a significant decrease in the amount of
dissolved Fe, and so among three-valent metals in the raw hydroxide, mainly Al and Cr dominated. The Mn content was significantly decreased if the sludge was extracted under strongly oxidising conditions by the addition of H 2 O 2 Ž1% of the final concentration.. The extract obtained was precipitated with NaOH solution. In the presence of a large amount of Si as in sludge A, the raw Zn hydroxide was practically XRD amorphous or several unidentified broad diffraction peaks were found in the XRD pattern. When the amount of Si was smaller as in sludge B, the raw hydroxide contained bechererite, ŽZn,Cu. 6Zn 2 ŽOH.13wŽSi,S.ŽO,OH.4 x 2 ŽICDD card 46-1438.. In the absence of Si the in synthetic solution, basic zinc sulphate Zn 4 SO4ŽOH. 6 P 4H 2 O ŽICDD card 44673. also precipitated. All these hydroxide species were dehydrated for 3 h at 8008C or 10008C to the oxide mixtures denoted as calcine in Fig. 1. At temperature below 8008C, poorly crystalline and, hence, highly reactive oxides were formed w3x. Well-defined mixed oxides including oxides with spinel structures, which bound the majority of Al, Cr
20
T. Grygar et al.r Hydrometallurgy 60 (2001) 17–24
and other metal impurities, were formed at 10008C w3,4,10x. The exact elemental composition of the phases, identified by XRD, could not be determined on the basis of the available data, because the oxides were not stoichiometric, and each element can occur in more than one phase. The overall specific surface area of the samples was in the range from 0.3 to 6 m2 gy1 . The calcines obtained at a given temperature have generally higher specific surface areas if they contain higher levels of Si and metal impurities. Despite the greater surface area, the reactivity of more impure samples was generally lower. An example of the kinetics of leaching of Zn, Al, Cu, and Fe from the calcined raw hydroxides is shown in Fig. 2. In the dissolution experiments, leaching time was usually in the order of minutes for a significant fraction of Cd, Cu, Mg, Si, Zn and a few tens of minutes for Ni. Practically all the trivalent metals and Mn remained in the insoluble residue. In the samples obtained by galvanic–sludge processing, only prevailing elements, occurring in the amounts bigger than about 1%, are discussed below. The elemental composition and specific surface areas of the individual components of the oxide mixtures were not available and, hence, the dissolution kinetics cannot be fully and exactly described. The evalu-
ation of the dissolution kinetics given below is, hence, valid only in the frame of the given experimental conditions. To interpret the kinetics of leaching of the individual metal components from the calcines, a series of experiments with both Zn oxide mixtures obtained by processing galvanic sludge and synthetic mixtures of metal oxides with varying elemental composition was performed. The elemental and phase composition of the synthetic mixtures and their dissolution behaviour are summarised in Table 3. Dissolution kinetics were evaluated with respect to the overall phase and elemental composition of the oxide mixtures. The following phases were identified in the synthetic oxide mixtures: very quickly dissolving zincite ŽZnO. and tenorite ŽCuO., very inert corundum and Al,Cr-containing spinel type oxides, and mainly Cu,Mg,Zn-substituted bunsenite ŽNiO., of which reactivity differs from sample to sample, most likely due to the substitution effect of metal ions. Very reactive willemite ŽZn 2 SiO4 . was also formed at the expense of zincite in Si-rich samples obtained by processing galvanic sludge. As it was already shown w4x, bunsenite doped with trivalent metals Al, Cr and Fe is less reactive than pure NiO and NiO doped with divalent impurities. In the presence of
Fig. 2. An example of dissolution kinetic curves of Zn, Cu, Ni, and Al. For oxide composition see Table 3, run 12. Percentage of dissolution, y, is related to the original amount of the corresponding element in the sample. The sample was dissolved in 0.5 M H 2 SO4 at 408C for 30 min, solid to liquid ratio was 1:40.
T. Grygar et al.r Hydrometallurgy 60 (2001) 17–24
21
Table 3 Elemental and phase composition and dissolution behaviour of synthetic oxide mixtures without Si Ž1–10. and mixtures obtained by sludge A processing Ž11 and 12. in 0.5 M H 2 SO4 at 408C within 30 min Run 1 2 3 4 5 6 7 8 9 10 11 12
Elemental and phase composition Zn Ž80%. Cr Ž52%., Fe Ž25%. Cr Ž43%., Ni Ž32%. Cr Ž43%., Mn Ž32%. Cr Ž65%., Mg Ž8%. Al Ž33%., Mn Ž25%. Al Ž3%., Cr Ž16%., Cu Ž15%., Mg Ž1%., Mn Ž3%., Ni Ž18%., Zn Ž15%. Al Ž13%., Cr Ž27%., Cu Ž3%., Mg Ž1%., Mn Ž8%., Ni Ž3%., Zn Ž3%. Cu Ž14%., Mg Ž2%., Mn Ž13%., Ni Ž13%., Zn Ž39%. Al Ž5%., Cu Ž1.5%., Cr Ž1%., Mn Ž2%., Ni Ž2%., Zn Ž65%. Al Ž2%., Mg Ž3%., Mn Ž4%., Ni Ž3%., Zn Ž50%. Al Ž4%., Cr Ž2%., Mg Ž1%., Mn Ž1%., Ni Ž2%., Zn Ž45%.
Z C S S S, C C, S S, B, T S S, Z, B, T Z, S Z, S, B Z, W, S
Dissolution extent
Insoluble residue
100% - 1% - 1% - 1% - 1% Cr, 3% Mg - 1% Al, 5% Mn 60% Cu, 4% Ni, 3% Mg, 2% Mn and Zn,- 1% Al and Cr 2% Al,- 1% others
none C S S S, C C, S S, B
72% Cu, 64% Zn, 37% Mg and Ni, 1% Mn 97% Zn
S
98% Zn, 87% Mg, 82% Cu, 80% Ni,- 5% others 97% Zn, 87% Mg, 80% Cu, 54% Ni,- 3% others
S
S S S
Oxides calcination at 10008C. Elemental composition is expressed in weight percent, components with content below 1% are omitted. Abbreviations of phase compositions: B bunsenite Ždoped NiO., C corundum, S spinel, T tenorite ŽCuO., W willemite ŽZn 2 SiO4 ., Z zincite ŽZnO.. Dissolution extent is related to the total amount of metals in the calcine.
sufficient amounts of the trivalent metals Al, Cr, Fe and Mn, the spinel is formed with the appropriate divalent metals — in our case, preferentially with Ni and Zn rather than Cu. The phases with the structures of spinel and bunsenite were the only two phases in the insoluble residue after the second leaching in processing the galvanic sludge. The general tendencies of oxide formation and their dissolution reactivities are summarised in Fig. 3. Semiquantitative evaluation of the dissolution kinetics is shown in Table 3. It is obvious, that the dissolution of Zn from zincite and willemite, Cu from tenorite, and Ni from bunsenite can be significant within the time scale of several tens of minutes. The dissolution kinetics of these elements can be directly used to evaluate dissolution reactivity of these mentioned phases. Reactivity was expressed as rate coefficients defined by the exponential kinetic law: y s y TOT 1 y exp Ž yk diss t . ,
Ž 1.
where y is the percentage dissolved of an element in a given dissolving phase, y TOT is the total percent-
age of the element in the dissolving phase, k diss is the rate coefficient in miny1 , and t is time in minutes. The exponential kinetic law represents a simplified case of the kinetics of a general reaction order with the reaction order equal to 1, which is applicable for the linear dissolution of a population of spherical or cubic particles with the logarithmic– normal size distribution w11x. In this case, the time necessary for dissolution of a half of the available reactant is equal to 0.69rk diss . If the dissolution extent was too small with respect to the time scale of the experiment, the rate coefficient was estimated from the initial linear portion of the dissolution curve: y f k diss t.
Ž 2.
This latter approach was used in the case of spinel oxides. The definition of the rate coefficient is the same as in the former formula. The results of the quantitative dissolution data processing are given in Table 4. Because, for example, MgO forms solid solutions with NiO, the dissolution of Mg and Ni can
22 T. Grygar et al.r Hydrometallurgy 60 (2001) 17–24
Fig. 3. The oxide phases formed by calcination of the impure Zn hydroxide and the dissolution reactivity of these phases under the test conditions Žcf. Section 2 for details..
T. Grygar et al.r Hydrometallurgy 60 (2001) 17–24 Table 4 Case relevant rate coefficients of dissolution of the oxides formed by the calcination Ž10008C. of synthetic hydroxides and raw hydroxides from sludge processing Ž k diss , miny1 . Run
Z
T
B
S
1 Žsynthetic, pure ZnO. 10 Žsynthetic. 11 Žsludge. 12 Žsludge. 7 Žsynthetic. 9 Žsynthetic.
0.8 1.0 0.35 0.6 n.p. )1.5
n.p. n.p. n.p. n.p. 0.12 0.12
n.p. n.p. 0.14 0.02 1.5=10y3 0.02
n.p. - 4=10y4 -1=10y3 6=10y5 4=10y5
Abbreviations: phase denotation as in Table 3, n.p. — not present.
be interrelated, as follows from similar shapes of the dissolution curves. As shown by the data in Table 4, the actual reactivity of all the phases varies from experiment to experiment. There is no direct relationship between the overall specific surface areas of the samples and the corresponding rate coefficients, a result that is most likely due to the significant influence of the actual elemental composition of the oxides. Especially the dissolution reactivity of bunsenite varies significantly, most likely due to the varying percentage of Al and Cr in the oxide mixture: in accordance with Ref. w4x, the larger the amount of the trivalent metal, the lower the rate coefficient of bunsenite. For example, in experiment 7, when the total amount of Cr was as high as 16% ŽTable 3., the corresponding bunsenite is least reactive ŽTable 4.. The half-times of the dissolution of the bunsenite components of the oxide mixtures are comparable to those observed for pure and doped bunsenite samples in Ref. w4x. Also, the reactivity of tenorite components corresponds to that of synthetic samples as described elsewhere w5x.
23
4. Discussion and conclusions The reactivity of individual components of the oxide mixtures toward their dissolution in 0.5 M H 2 SO4 decreases in the order: zincite,willemite ) tenoriteG bunsenite4 spinel. The differences in the reaction times for the dissolution of these species are the orders of magnitude. The rate of the leaching of Zn, which is present in these oxides, decreases in the same order: ZnO,Zn 2 SiO4 ) Ž Zn,Cu . O G Ž Zn,Ni . O 4 Ž Zn,Ni . Ž Al,Cr . 2 O4 . In the dissolution experiments lasting for several tens of minutes, e.g. in times much shorter than 1rk diss , Zn contained in spinels is not dissolved significantly. The insoluble fraction of Zn increases with increasing amount of trivalent metals Al, Cr, and Fe in the calcine. The phase composition of the calcine depends on the original chemical composition. The silicates are transformed to Zn 2 SiO4 , which leads to the dissolution of silica together with Zn. Unfortunately, the separation of Si and Zn is one of the common problems of Zn hydrometallurgy. In the presence of a large excess of Cu in the oxide mixture, very reactive CuO nucleates during calcination and its presence then worsens the separation of Zn and Cu. If a sufficient amount of trivalent metal ions ŽAl, Cr, Fe, Mn. is available in the oxide mixture, mixed aluminates, chromites and ferrites with the structure of spinel are formed. These spinels are very slowly
Table 5 The weight ratios of the content of selected impurities to the zinc content in solids and solutions processed according to the procedure shown in Fig. 1
Galvanic sludge A 1st extract at pH s 3.9 Calcine 2nd extract Žafter 10 min. Insoluble residue after 120 min Refined hydroxide ŽpH s 7.
Zn
AlrZn
CurZn
FerZn
NirZn
7.67% 8.18 grl 45.1% 11.3 grl 26.9% 50.1%
0.20 0.074 0.086 0.0018 0.55 0.0012
0.055 0.039 0.033 0.024 0.039 0.015
1.32 0.0008 0.0027 - 0.0001 0.017 - 0.00002
0.046 0.075 0.047 0.016 0.16 0.0027
24
T. Grygar et al.r Hydrometallurgy 60 (2001) 17–24
dissolved, however, this is also valid for Zn-containing spinels, and so the leaching efficiency of Zn decreases in the excess of these impurities. The presence of Ni promotes the formation of substituted bunsenite. Its reactivity depends on its actual composition, as already reported w4,10x, but is generally at least one order of magnitude smaller with respect to ZnO. The last type of oxides formed from the metal impurities is sesquioxide of Al, Cr, Fe andror Mn, which also exhibits very low dissolution reactivity. However, because this oxide is thermodynamically unstable in the presence of the excess divalent metal ions, trivalent metal impurities in ZnO are preferentially bound in more stable spinel oxides. Due to the different reactivity and composition of the components of the oxide mixtures, partial dissolution of the oxide mixture can be used for purification of zinc hydroxide. An example of the content of Zn and the relative amount of four selected impurities in the species shown in Fig. 1 is given in Table 5. From the table, it is clear that the different dissolution kinetics permits one to decrease the amount of trivalent impurities in Zn intermediates to about 1r100 of their original relative amount.
Acknowledgements This work was supported by the Grant Agency of the Czech Republic Žgrant No. 104r97r0705. and by the Research Intention CEZ: 19r28 22300002.
References w1x J. Jandova, ´ J. Maixner, Utilisation of the formation of spinels for removal of impurities from metal-bearing sludges, in: T. Havlık ´ ŽEd.., Proc. Int. Conf. Metallurgy, Refractories and Environment, Herlany-Kosice, Institute of Metallurgy and ˇ Materials, Technical University Kosice, Slovak Republic, ˇ 1997, pp. 177–182. w2x R. Malecky, ´ Dissolution of zinc from contaminated zinc oxide Žin Czech., Thesis, Institute of Chemical Technology, Prague, Czech Republic, 1997. w3x Z. Klımova, ´ ´ Utilization of hardly soluble spinels for processing zinc sludge Žin Czech., Thesis, Institute of Chemical Technology, Prague, Czech Republic, 1999. w4x T. Grygar, J. Jandova, ´ Z. Klımova, ´ ´ Dissolution reactivity of NiO obtained by calcination of pure and contaminated Ni-hydroxides, Hydrometallurgy 52 Ž1999. 137. ˇ w5x J. Jandova, ´ T. Stefanova, ´ R. Niemcyzkova, ´ Recovery of Cu-concentrates from waste galvanic copper sludges, Hydrometallurgy 57 Ž1. Ž2000. 77–84. w6x M. Blesa, A. Morando, J. Regazzoni, Chemical Dissolution of Metal Oxides, CRC Press, Boca Raton, FL, 1994, Chapter 17: Dissolution of mixed oxides. w7x B. Terry, Specific chemical rate constants for the acid dissolution of oxides and silicates, Hydrometallurgy 11 Ž1983. 315. w8x J. Guspiel, W. Reisenkampf, Kinetics of dissolution of ZnO, ´ MgO and their solid solutions in aqueous sulphuric acid solutions, Hydrometallurgy 34 Ž1993. 203. w9x F. Elgersma, G. Kamst, F. Witkamp, J. Rosmalen, Acidic dissolution of zinc ferrite, Hydrometallurgy 29 Ž1992. 173. w10x L. Doubrava, Utilization of hardly soluble spinels for processing nickel sludge Žin Czech., Thesis, Institute of Chemical Technology, Prague, Czech Republic, 1998. w11x T. Grygar, Electrochemical dissolution of ironŽIII. hydroxyoxides. More Information about the Particles, Coll. Czech. Chem. Commun. 61 Ž1996. 93.
Dissolution reactivity of metal oxide mixtures containing ZnO T. Grygar a,) , Z. Klımova ´ ´ a,b, J. Jandova´ b b
a ˘ ˘ Czech Republic Institute of Inorganic Chemistry, Academy of Sciences of the Czech Republic, 250 68 Rez, Department of Metals and Corrosion Engineering, Prague Institute of Chemical Technology, Technicka´ 6, 166 28 Prague, Czech Republic
Received 15 December 1999; received in revised form 21 August 2000; accepted 9 September 2000
Abstract The kinetics of zinc dissolution from oxide mixtures in 0.5 M H 2 SO4 was studied. The oxide mixtures were prepared by acid leaching of galvanic sludge, precipitation of the extracts by NaOH followed by calcination of the produced Zn hydroxides. A series of experiments with synthetic oxide mixtures of zinc and other common metals were also performed. The phase composition of the oxide mixtures significantly affected the kinetics of the zinc leaching. The most reactive phases were ZnO and Zn 2 SiO4 , less reactive were substituted CuO and NiO, and the least reactive were spinel-type oxides. Due to the different rates of dissolution of individual oxide phases, the described process can be used for purification of Zn from the galvanic sludge. q 2001 Elsevier Science B.V. All rights reserved. Keywords: ZnO; Dissolution; Industrial waste
1. Introduction The problem of recycling metal-containing industrial waste is not satisfactorily solved in the Czech Republic. Even the neutralisation sludges from Zn electroplating and mechanoplating plants are not recycled, although they contain as much as 5–10% of Zn mainly in the form of hydroxides. Nowadays, these sludges are only deposited in landfills as dangerous waste. Partial dissolution of the oxide mixtures obtained by heating impure Zn hydroxides or basic salts was recently proposed by Jandova´ and Maixner w1x for the recovery of metals from Zn-containing poly-
)
Corresponding author. E-mail address: [email protected] ŽT. Grygar..
metallic galvanic sludge w2,3x. The basis of the separation of Zn is calcination of the mixed hydroxides to transform the metal impurities, especially Al, Cr and Fe, to the oxides, which dissolve much slower in acidic solution than ZnO and Zn 2 SiO4 . The list of oxides, which can be expected andror found in those mixtures w3x, together with their dissolution reactivities, w4–9x are given in Table 1. A similar procedure to extract Ni w4x and Cu w5x preferentially from oxide mixtures was already proposed. In contrast to NiO and the oxides of trivalent metals Al, Cr and Fe, the difference in reactivities of ZnO and Zn 2 SiO4 and the oxides of trivalent metals is much larger and, hence, Zn can be more selectively leached from the oxide mixture. The economical evaluation of this proposed treatment can only be done if more details were known about its applicability for particular impurities present in Zn sludges.
0304-386Xr01r$ - see front matter q 2001 Elsevier Science B.V. All rights reserved. PII: S 0 3 0 4 - 3 8 6 X Ž 0 0 . 0 0 1 5 3 - 5
18
T. Grygar et al.r Hydrometallurgy 60 (2001) 17–24
Table 1 The most common oxide phases formed by heating impure metal hydroxides and basic salts together with evaluation of their dissolution reactivity Žin solutions of HCl or H 2 SO4 . published in Refs. w3–9x Phase ZnO Žzincite. Zn 2 SiO4 Žwillemite. MgO Žpericlase. and ŽZn,Mg.O CuO Žtenorite. ZnFe 2 O4 Žfranklinite. NiO Žbunsenite. NiCr2 O4 a-ŽAl,Cr,Fe. 2 O 3
Dissolution behaviour
Reference
fast diffusion-controlled or chemical dissolution fast diffusion-controlled or chemical dissolution fast diffusion-controlled or chemical dissolution relatively fast chemical dissolution chemical dissolution chemical dissolution sensitive to actual chemical composition very slow chemical dissolution very slow chemical dissolution
w6x w7x w6,8x w5x w9x w4,6x w3x w3x
The aim of this work was to improve our knowledge about the phase composition of the oxide mixture obtained by calcining the contaminated Zn hydroxides and about the dissolution reactivity of the resulting oxides. The problem of the kinetics of the dissolution of the mixed oxides is a rather complex one w6x, and so the reactivity of the mixed oxides cannot be directly predicted from the reactivity of the pure components. The experimental data are rarely available for non-stoichiometric oxides. More experimental studies are, hence, necessary of the dissolution involving complex mixtures of nonstoichiometric oxides that are formed from polymetallic hydroxides.
2. Experimental procedures Synthetic oxide mixtures for dissolution studies were prepared by precipitation of mixtures of metal salts with NaOH solution followed by filtration, drying and calcining Ž10008C, 2 h. of the precipitate. Additionally, two samples of galvanic sludge were obtained from the landfill in Dolnı´ Rozınka in the ˘´ Czech Republic, and from a galvanic Zn plating plant in Letky in the Czech Republic. Each sludge
sample was homogenised and its composition was characterised by semi-quantitative emission spectral analysis and by atomic absorption spectroscopy. The approximate elemental composition is summarised in Table 2. The most important difference between these two samples is in the SirZn ratio, which is significantly increased in sludge A from the waste deposit due to the subsequent contamination of the sludge in the landfill. In sample A, the weight ratio of Si to Zn is 1.1:1. Solid samples were characterised by X-ray powder diffraction ŽSiemens D5005. and specific surface area measurement ŽCoulter SA3100, Beckman Coulter.. For the chemical analysis, the sludges and hydroxides were dissolved in diluted HCl. The insoluble residues after zinc leaching were decomposed by fusing them with NaOH or by evaporating with HF–H 2 SO4 to expel SiO 2 followed by acidic fusion of the residue ŽK 2 S 2 O 7 .. The chemical composition was then obtained by atomic absorption spectroscopy, gravimetry and chelatometric titration. The flowsheet of the sludge process used to obtain the oxide mixtures and refined zinc hydroxide is given in Fig. 1. The solution denoted the first extract in Fig. 1 was prepared by leaching the sludge in aqueous H 2 SO4 at pH s 3 or 3.9 in order to extract
Table 2 Approximate chemical composition of the original galvanic sludge Sample
10 0 –10 1 %
10y1 –10 0 %
10y2 –10y1 %
A Žwaste dump in Dolnı´ Rozınka ˘´ . B Žgalvanic plant in Letky.
Ca, Fe, Si, Zn Ca, Zn
Al, Ba, Mg, Mn, Cr, Cu, P, Pb, Ni Al, Fe, Mg, Mn, Si
B, Cd B, Sn
T. Grygar et al.r Hydrometallurgy 60 (2001) 17–24
19
Fig. 1. The flowsheet of the processing of the Zn-sludge.
the maximum amount of Zn on one hand, and to leave the majority of the trivalent metal impurities in the solid residue on the other hand. Four different kinds of solid phases were investigated: a raw Zn hydroxide obtained by neutralisation of the first extract to pH s 7, a calcined oxide mixture, an insoluble residue after the second leaching, and finally, a refined Zn hydroxide, obtained by the second precipitation to pH s 7. Dissolution kinetics was performed in a glass reactor at 408C in 0.5 M H 2 SO4 . Solid to liquid ratio of 1:40 was chosen. The suspension was stirred with a propeller to avoid settling. To monitor the dissolution rate, the suspension was sampled at appropriate times, filtered and its content of dissolved metals was determined by atomic absorption spectroscopy.
3. Results Leaching the galvanic sludge at pH as high as 3–3.9 led to a significant decrease in the amount of
dissolved Fe, and so among three-valent metals in the raw hydroxide, mainly Al and Cr dominated. The Mn content was significantly decreased if the sludge was extracted under strongly oxidising conditions by the addition of H 2 O 2 Ž1% of the final concentration.. The extract obtained was precipitated with NaOH solution. In the presence of a large amount of Si as in sludge A, the raw Zn hydroxide was practically XRD amorphous or several unidentified broad diffraction peaks were found in the XRD pattern. When the amount of Si was smaller as in sludge B, the raw hydroxide contained bechererite, ŽZn,Cu. 6Zn 2 ŽOH.13wŽSi,S.ŽO,OH.4 x 2 ŽICDD card 46-1438.. In the absence of Si the in synthetic solution, basic zinc sulphate Zn 4 SO4ŽOH. 6 P 4H 2 O ŽICDD card 44673. also precipitated. All these hydroxide species were dehydrated for 3 h at 8008C or 10008C to the oxide mixtures denoted as calcine in Fig. 1. At temperature below 8008C, poorly crystalline and, hence, highly reactive oxides were formed w3x. Well-defined mixed oxides including oxides with spinel structures, which bound the majority of Al, Cr
20
T. Grygar et al.r Hydrometallurgy 60 (2001) 17–24
and other metal impurities, were formed at 10008C w3,4,10x. The exact elemental composition of the phases, identified by XRD, could not be determined on the basis of the available data, because the oxides were not stoichiometric, and each element can occur in more than one phase. The overall specific surface area of the samples was in the range from 0.3 to 6 m2 gy1 . The calcines obtained at a given temperature have generally higher specific surface areas if they contain higher levels of Si and metal impurities. Despite the greater surface area, the reactivity of more impure samples was generally lower. An example of the kinetics of leaching of Zn, Al, Cu, and Fe from the calcined raw hydroxides is shown in Fig. 2. In the dissolution experiments, leaching time was usually in the order of minutes for a significant fraction of Cd, Cu, Mg, Si, Zn and a few tens of minutes for Ni. Practically all the trivalent metals and Mn remained in the insoluble residue. In the samples obtained by galvanic–sludge processing, only prevailing elements, occurring in the amounts bigger than about 1%, are discussed below. The elemental composition and specific surface areas of the individual components of the oxide mixtures were not available and, hence, the dissolution kinetics cannot be fully and exactly described. The evalu-
ation of the dissolution kinetics given below is, hence, valid only in the frame of the given experimental conditions. To interpret the kinetics of leaching of the individual metal components from the calcines, a series of experiments with both Zn oxide mixtures obtained by processing galvanic sludge and synthetic mixtures of metal oxides with varying elemental composition was performed. The elemental and phase composition of the synthetic mixtures and their dissolution behaviour are summarised in Table 3. Dissolution kinetics were evaluated with respect to the overall phase and elemental composition of the oxide mixtures. The following phases were identified in the synthetic oxide mixtures: very quickly dissolving zincite ŽZnO. and tenorite ŽCuO., very inert corundum and Al,Cr-containing spinel type oxides, and mainly Cu,Mg,Zn-substituted bunsenite ŽNiO., of which reactivity differs from sample to sample, most likely due to the substitution effect of metal ions. Very reactive willemite ŽZn 2 SiO4 . was also formed at the expense of zincite in Si-rich samples obtained by processing galvanic sludge. As it was already shown w4x, bunsenite doped with trivalent metals Al, Cr and Fe is less reactive than pure NiO and NiO doped with divalent impurities. In the presence of
Fig. 2. An example of dissolution kinetic curves of Zn, Cu, Ni, and Al. For oxide composition see Table 3, run 12. Percentage of dissolution, y, is related to the original amount of the corresponding element in the sample. The sample was dissolved in 0.5 M H 2 SO4 at 408C for 30 min, solid to liquid ratio was 1:40.
T. Grygar et al.r Hydrometallurgy 60 (2001) 17–24
21
Table 3 Elemental and phase composition and dissolution behaviour of synthetic oxide mixtures without Si Ž1–10. and mixtures obtained by sludge A processing Ž11 and 12. in 0.5 M H 2 SO4 at 408C within 30 min Run 1 2 3 4 5 6 7 8 9 10 11 12
Elemental and phase composition Zn Ž80%. Cr Ž52%., Fe Ž25%. Cr Ž43%., Ni Ž32%. Cr Ž43%., Mn Ž32%. Cr Ž65%., Mg Ž8%. Al Ž33%., Mn Ž25%. Al Ž3%., Cr Ž16%., Cu Ž15%., Mg Ž1%., Mn Ž3%., Ni Ž18%., Zn Ž15%. Al Ž13%., Cr Ž27%., Cu Ž3%., Mg Ž1%., Mn Ž8%., Ni Ž3%., Zn Ž3%. Cu Ž14%., Mg Ž2%., Mn Ž13%., Ni Ž13%., Zn Ž39%. Al Ž5%., Cu Ž1.5%., Cr Ž1%., Mn Ž2%., Ni Ž2%., Zn Ž65%. Al Ž2%., Mg Ž3%., Mn Ž4%., Ni Ž3%., Zn Ž50%. Al Ž4%., Cr Ž2%., Mg Ž1%., Mn Ž1%., Ni Ž2%., Zn Ž45%.
Z C S S S, C C, S S, B, T S S, Z, B, T Z, S Z, S, B Z, W, S
Dissolution extent
Insoluble residue
100% - 1% - 1% - 1% - 1% Cr, 3% Mg - 1% Al, 5% Mn 60% Cu, 4% Ni, 3% Mg, 2% Mn and Zn,- 1% Al and Cr 2% Al,- 1% others
none C S S S, C C, S S, B
72% Cu, 64% Zn, 37% Mg and Ni, 1% Mn 97% Zn
S
98% Zn, 87% Mg, 82% Cu, 80% Ni,- 5% others 97% Zn, 87% Mg, 80% Cu, 54% Ni,- 3% others
S
S S S
Oxides calcination at 10008C. Elemental composition is expressed in weight percent, components with content below 1% are omitted. Abbreviations of phase compositions: B bunsenite Ždoped NiO., C corundum, S spinel, T tenorite ŽCuO., W willemite ŽZn 2 SiO4 ., Z zincite ŽZnO.. Dissolution extent is related to the total amount of metals in the calcine.
sufficient amounts of the trivalent metals Al, Cr, Fe and Mn, the spinel is formed with the appropriate divalent metals — in our case, preferentially with Ni and Zn rather than Cu. The phases with the structures of spinel and bunsenite were the only two phases in the insoluble residue after the second leaching in processing the galvanic sludge. The general tendencies of oxide formation and their dissolution reactivities are summarised in Fig. 3. Semiquantitative evaluation of the dissolution kinetics is shown in Table 3. It is obvious, that the dissolution of Zn from zincite and willemite, Cu from tenorite, and Ni from bunsenite can be significant within the time scale of several tens of minutes. The dissolution kinetics of these elements can be directly used to evaluate dissolution reactivity of these mentioned phases. Reactivity was expressed as rate coefficients defined by the exponential kinetic law: y s y TOT 1 y exp Ž yk diss t . ,
Ž 1.
where y is the percentage dissolved of an element in a given dissolving phase, y TOT is the total percent-
age of the element in the dissolving phase, k diss is the rate coefficient in miny1 , and t is time in minutes. The exponential kinetic law represents a simplified case of the kinetics of a general reaction order with the reaction order equal to 1, which is applicable for the linear dissolution of a population of spherical or cubic particles with the logarithmic– normal size distribution w11x. In this case, the time necessary for dissolution of a half of the available reactant is equal to 0.69rk diss . If the dissolution extent was too small with respect to the time scale of the experiment, the rate coefficient was estimated from the initial linear portion of the dissolution curve: y f k diss t.
Ž 2.
This latter approach was used in the case of spinel oxides. The definition of the rate coefficient is the same as in the former formula. The results of the quantitative dissolution data processing are given in Table 4. Because, for example, MgO forms solid solutions with NiO, the dissolution of Mg and Ni can
22 T. Grygar et al.r Hydrometallurgy 60 (2001) 17–24
Fig. 3. The oxide phases formed by calcination of the impure Zn hydroxide and the dissolution reactivity of these phases under the test conditions Žcf. Section 2 for details..
T. Grygar et al.r Hydrometallurgy 60 (2001) 17–24 Table 4 Case relevant rate coefficients of dissolution of the oxides formed by the calcination Ž10008C. of synthetic hydroxides and raw hydroxides from sludge processing Ž k diss , miny1 . Run
Z
T
B
S
1 Žsynthetic, pure ZnO. 10 Žsynthetic. 11 Žsludge. 12 Žsludge. 7 Žsynthetic. 9 Žsynthetic.
0.8 1.0 0.35 0.6 n.p. )1.5
n.p. n.p. n.p. n.p. 0.12 0.12
n.p. n.p. 0.14 0.02 1.5=10y3 0.02
n.p. - 4=10y4 -1=10y3 6=10y5 4=10y5
Abbreviations: phase denotation as in Table 3, n.p. — not present.
be interrelated, as follows from similar shapes of the dissolution curves. As shown by the data in Table 4, the actual reactivity of all the phases varies from experiment to experiment. There is no direct relationship between the overall specific surface areas of the samples and the corresponding rate coefficients, a result that is most likely due to the significant influence of the actual elemental composition of the oxides. Especially the dissolution reactivity of bunsenite varies significantly, most likely due to the varying percentage of Al and Cr in the oxide mixture: in accordance with Ref. w4x, the larger the amount of the trivalent metal, the lower the rate coefficient of bunsenite. For example, in experiment 7, when the total amount of Cr was as high as 16% ŽTable 3., the corresponding bunsenite is least reactive ŽTable 4.. The half-times of the dissolution of the bunsenite components of the oxide mixtures are comparable to those observed for pure and doped bunsenite samples in Ref. w4x. Also, the reactivity of tenorite components corresponds to that of synthetic samples as described elsewhere w5x.
23
4. Discussion and conclusions The reactivity of individual components of the oxide mixtures toward their dissolution in 0.5 M H 2 SO4 decreases in the order: zincite,willemite ) tenoriteG bunsenite4 spinel. The differences in the reaction times for the dissolution of these species are the orders of magnitude. The rate of the leaching of Zn, which is present in these oxides, decreases in the same order: ZnO,Zn 2 SiO4 ) Ž Zn,Cu . O G Ž Zn,Ni . O 4 Ž Zn,Ni . Ž Al,Cr . 2 O4 . In the dissolution experiments lasting for several tens of minutes, e.g. in times much shorter than 1rk diss , Zn contained in spinels is not dissolved significantly. The insoluble fraction of Zn increases with increasing amount of trivalent metals Al, Cr, and Fe in the calcine. The phase composition of the calcine depends on the original chemical composition. The silicates are transformed to Zn 2 SiO4 , which leads to the dissolution of silica together with Zn. Unfortunately, the separation of Si and Zn is one of the common problems of Zn hydrometallurgy. In the presence of a large excess of Cu in the oxide mixture, very reactive CuO nucleates during calcination and its presence then worsens the separation of Zn and Cu. If a sufficient amount of trivalent metal ions ŽAl, Cr, Fe, Mn. is available in the oxide mixture, mixed aluminates, chromites and ferrites with the structure of spinel are formed. These spinels are very slowly
Table 5 The weight ratios of the content of selected impurities to the zinc content in solids and solutions processed according to the procedure shown in Fig. 1
Galvanic sludge A 1st extract at pH s 3.9 Calcine 2nd extract Žafter 10 min. Insoluble residue after 120 min Refined hydroxide ŽpH s 7.
Zn
AlrZn
CurZn
FerZn
NirZn
7.67% 8.18 grl 45.1% 11.3 grl 26.9% 50.1%
0.20 0.074 0.086 0.0018 0.55 0.0012
0.055 0.039 0.033 0.024 0.039 0.015
1.32 0.0008 0.0027 - 0.0001 0.017 - 0.00002
0.046 0.075 0.047 0.016 0.16 0.0027
24
T. Grygar et al.r Hydrometallurgy 60 (2001) 17–24
dissolved, however, this is also valid for Zn-containing spinels, and so the leaching efficiency of Zn decreases in the excess of these impurities. The presence of Ni promotes the formation of substituted bunsenite. Its reactivity depends on its actual composition, as already reported w4,10x, but is generally at least one order of magnitude smaller with respect to ZnO. The last type of oxides formed from the metal impurities is sesquioxide of Al, Cr, Fe andror Mn, which also exhibits very low dissolution reactivity. However, because this oxide is thermodynamically unstable in the presence of the excess divalent metal ions, trivalent metal impurities in ZnO are preferentially bound in more stable spinel oxides. Due to the different reactivity and composition of the components of the oxide mixtures, partial dissolution of the oxide mixture can be used for purification of zinc hydroxide. An example of the content of Zn and the relative amount of four selected impurities in the species shown in Fig. 1 is given in Table 5. From the table, it is clear that the different dissolution kinetics permits one to decrease the amount of trivalent impurities in Zn intermediates to about 1r100 of their original relative amount.
Acknowledgements This work was supported by the Grant Agency of the Czech Republic Žgrant No. 104r97r0705. and by the Research Intention CEZ: 19r28 22300002.
References w1x J. Jandova, ´ J. Maixner, Utilisation of the formation of spinels for removal of impurities from metal-bearing sludges, in: T. Havlık ´ ŽEd.., Proc. Int. Conf. Metallurgy, Refractories and Environment, Herlany-Kosice, Institute of Metallurgy and ˇ Materials, Technical University Kosice, Slovak Republic, ˇ 1997, pp. 177–182. w2x R. Malecky, ´ Dissolution of zinc from contaminated zinc oxide Žin Czech., Thesis, Institute of Chemical Technology, Prague, Czech Republic, 1997. w3x Z. Klımova, ´ ´ Utilization of hardly soluble spinels for processing zinc sludge Žin Czech., Thesis, Institute of Chemical Technology, Prague, Czech Republic, 1999. w4x T. Grygar, J. Jandova, ´ Z. Klımova, ´ ´ Dissolution reactivity of NiO obtained by calcination of pure and contaminated Ni-hydroxides, Hydrometallurgy 52 Ž1999. 137. ˇ w5x J. Jandova, ´ T. Stefanova, ´ R. Niemcyzkova, ´ Recovery of Cu-concentrates from waste galvanic copper sludges, Hydrometallurgy 57 Ž1. Ž2000. 77–84. w6x M. Blesa, A. Morando, J. Regazzoni, Chemical Dissolution of Metal Oxides, CRC Press, Boca Raton, FL, 1994, Chapter 17: Dissolution of mixed oxides. w7x B. Terry, Specific chemical rate constants for the acid dissolution of oxides and silicates, Hydrometallurgy 11 Ž1983. 315. w8x J. Guspiel, W. Reisenkampf, Kinetics of dissolution of ZnO, ´ MgO and their solid solutions in aqueous sulphuric acid solutions, Hydrometallurgy 34 Ž1993. 203. w9x F. Elgersma, G. Kamst, F. Witkamp, J. Rosmalen, Acidic dissolution of zinc ferrite, Hydrometallurgy 29 Ž1992. 173. w10x L. Doubrava, Utilization of hardly soluble spinels for processing nickel sludge Žin Czech., Thesis, Institute of Chemical Technology, Prague, Czech Republic, 1998. w11x T. Grygar, Electrochemical dissolution of ironŽIII. hydroxyoxides. More Information about the Particles, Coll. Czech. Chem. Commun. 61 Ž1996. 93.