2. 2830 ds+gl final
17 Canadian Metallurgical Quarterly, Vol 42, No 1 pp 17-28, 2003 © Canadian Institute of Mining, Metallurgy and Petroleum Published by Canadian Institute of Mining, Metallurgy and Petroleum Printed in Canada. All rights reserved
DISSOLVED COPPER AND COBALT RECOVERY FROM DILUTE ACIDIC SOLUTIONS BY ADSORPTION ON SILICA SOL E. EL-AMMOURI, S. RODGERS and P.A. DISTIN McGill University, Department of Mining, Metals and Materials Engineering, 3610 University Street, Montreal, QC H3A 2B2
(Received February, 2002; in revised form September, 2002)
Abstract — The ability of silica sol to neutralize and adsorb copper and cobalt from dilute, acidic solutions has been studied on a laboratory scale. At the pH for maximum adsorption (pH 7.0 for Cu and 8.3 for Co), more than 1 ppm metal remains dissolved which rises to about 30 ppm if pH is further increased. This effect is attributed to desorbed colloidal species at high pH. Final pH control with lime gave less than 1 ppm Cu or Co. The replacement of lime with sol reduces the formation of waste hydroxide/gypsum sludge. The adsorption of metal ions coagulates the sol and produces a sharply defined underflow and clear overflow. Adsorbed metal was recovered into a concentrated solution (relative to the feed) by reversing the adsorption reaction through acidification of centrifuged underflow. With feeds containing from 58 ppm to 2.83 g/L metal, product solutions from 339 ppm to 17.5 g/L metal were generated, although at 17.5 g/L a viscous suspension of agglomerates was produced. Use of recycled silica was demonstrated. Résumé — On a étudié en laboratoire la capacité d’un sol de silice à neutraliser et à adsorber le cuivre et le cobalt à partir de solutions acides diluées. Au pH d’adsorption maximum (pH 7.0 pour le Cu, 8.3 pour le Co), plus d’une ppm de métal restait dissoute, mais la quantité s’élevait à environ 30 ppm si on augmentait davantage le pH. Cet effet est attribué aux espèces colloïdales désorbées à pH élevé. Le contrôle du pH final avec de la chaux a donné moins d’une ppm de Cu ou de Co. Le remplacement de la chaux avec le sol réduit la formation de déchets d’hydroxyde/de sédiments de gypse. L’adsorption d’ions métalliques coagule le sol et produit un flot inférieur nettement défini et un flot supérieur clair. On a récupéré le métal adsorbé dans une solution concentrée (par rapport à la source) en renversant la réaction d’adsorption par l’intermédiaire de l’acidification du flot inférieur centrifugé. Avec des sources contenant entre 58 ppm et 2.83 g/L de métal, on générait des solutions de produit ayant de 339 ppm à 17.5 g/L de métal, bien qu’à 17.5 g/L, on produisait une suspension visqueuse d’agglomérés. On a démontré l’utilisation de la silice recyclée.
INTRODUCTION Dilute, weak acidic waste streams, such as acid mine drainage and mine effluents, are treated by adding lime and flocculants to a pH between 9 and 11. A voluminous metal hydroxide/gypsum tailings is formed from which metals may slowly redissolve due to gradually decreasing pH in tailings ponds. Previous laboratory scale studies [1-3] have shown how, in principle, solids waste generation can be significantly reduced, with recovery of associated metal values, by largely replacing lime with a stabilized silica sol. This
latter material acts mainly as an adsorption medium and removes dissolved metal in a manner that avoids extensive hydroxide precipitation. If adsorption/desorption on activated silica can be operated as a reversible pH-controlled process, an opportunity exists for transferring dissolved metal from a feed stream to a product stream as shown in Figure 1. Such a process has merit only if the activated silica is recyclable and if metal in the product solution is more concentrated than in the feed. When activated silica is added to a heavy metal solution to reach a controlled pH (‘Metal recovery’ step in Figure 1), coagulation of colloidal silica creates a metal bearing silica CANADIAN METALLURGICAL QUARTERLY
18
E. EL-AMMOURI, S. RODGERS and P.A. DISTIN
OH OH OH OH | | | | HO-Si-OH + HO-Si-OH Æ HO-Si-O-Si-OH + H2O (1) | | | | OH OH OH OH or in abbreviated form as | | | 2 (-Si-OH ) Æ - Si-O-Si- + H2O | | |
(2)
As described more fully in the following results and discussion section, polymerization may be initiated by adding sulphuric acid to an alkaline silicate solution. While continued polymer growth eventually gives a gel, polymerization can be effectively halted by dilution which creates a stabilized sol as is commonly done when preparing sols for water clarification [4-8]. Fig.1. Conceptual flowsheet for metals recovery using activated silica sol.
layer which rapidly settles to give a sharp interface with a clear overflow. The silica additions are much higher (several g/L) than those used for water treatment where activated silica at 1 ppm to 2 ppm SiO2 is added to flocculate suspended solids [4-8]. Acidification of the metal bearing silica layer, after discarding the overflow, gives metals desorption and regeneration of activated silica for recycle to the metal recovery step. The production of a concentrated metal solution, relative to the feed, is promoted by using concentrated sulphuric acid and water rejection steps (centrifugation in the present work) as shown in Figure 1. Direct soluble silicate addition, instead of activated silica, may give precipitates that yield solid silica upon acidification. Although solid silica also adsorbs dissolved metal ions, the mixing of silica as a sol with an aqueous solution of metal ions is expected to give more efficient silica/metal ion contact than obtainable with a suspension of solid silica particles. Copper was chosen because adsorption could be monitored in situ using a specific ion electrode for the cupric ion. When the cobaltous ion is used, adsorption can be readily distinguished from precipitation due to major differences in the colour of the products. DISSOLVED METAL ADSORPTION BY SILICA SOL, SILICA GEL OR SILICA Silica sols are created by the polymerization of silicic acid monomers, Si(OH)4, generated upon dissolution of a soluble silicate. Using dimer formation as the simplest example, polymerization can be represented as: CANADIAN METALLURGICAL QUARTERLY
Iler [9] studied the coagulation of silica sols resulting from adsorption of calcium ions from nitrate solutions. The objective was to determine the critical concentration of dissolved calcium ions needed for the coagulation of colloidal silica as a function of particle size. The present work uses this same coagulation effect for the adsorption of dissolved heavy metal ions to form a metal bearing underflow. When silica sol is added to an acidic metal-salt solution, the pH increases due to the Na2O content of the sol and adsorption sites are created for dissolved metal species. With a continuously increasing pH, the reaction path followed by dissolved metal ions passes through several intermediate hydroxy complexes before reaching the point where precipitation would be expected. Iler [10] and Falcone [11] observed that silica sols and silicate solutions began to adsorb dissolved multivalent metal ions at pH values up to 2 units below the pH at which metal hydroxide precipitated. Falcone [11] compared sodium hydroxide and silicates with various SiO2/Na2O ratios when added to copper perchlorate solutions at pH 4 to reach various pH levels up to 7. While a precipitate formed upon neutralization with sodium hydroxide, there was no visible precipitate when using silicate, although some precipitate appeared after ageing the solution for a month. Using a cupric ion selective electrode, cupric ion activities at a given pH were found to be lower after silicate than after sodium hydroxide addition. Colloidal silica adsorbs dissolved multivalent metal ions in a manner analogous to the interaction between a metal ion and a silica gel surface [10,11]. Dissolved metal adsorption on to silica gel (or silica) [12-17] may be represented as an ion exchange process. | | Mn+ + m (-SiOH ) ´ M (O Si-)m(n-m)+ + m H+ | |
(3)
DISSOLVED COPPER AND COBALT RECOVERY FROM DILUTE ACIDIC SOLUTIONS BY ADSORPTION ON SILICA SOL 19
Regardless of whether the starting material is silica sol, silica gel or solid silica, the reacting species is an acidic surface silanol group. Thus Reaction 3 can be taken as the sum of the following two reactions: | | m (-SiOH ) ´ M (SiO- ) + m H+ | |
(4)
| | Mn+ + m (- SiO-) ´ M (O Si-)m(n-m)+ | |
(5)
Reaction 3 has also been presented [18] as | | MLn+ + m (- SiOH) ´ ML (O Si-)m(n-m)+ + m H+ | |
(6)
where L is a species such as OH-. Dugger et al. [16] have studied the stoichiometry of Reaction 3 when applied to silica gel powder adsorbing a wide range of metals dissolved in buffered nitrate or perchlorate solutions. Adsorption was carried out at low pH where n could be taken as the charge on the simple unhydrolyzed metal ion. The quantity of hydrogen ions released per mole metal adsorbed was always such that n = m in Reaction 3 at equilibrium. Vydra and Galba [18] studied Reaction 6 at pH levels where hydrolyzed metal ions would be expected to form. Metals were adsorbed on to silica gel powder from buffered chloride or nitrate solutions. By assuming n = m in Reaction 6, the charge on the adsorbed complex could be deduced by measuring CM/CH where CM and CH are moles metal adsorbed and hydrogen ions released, respectively. A fractional value of CM/CH other than 1.0, 0.50 and 0.33 was taken to indicate adsorption of at least two cations of different valencies. In an experimental study of the ligand properties of surface silanol groups on silica, Schindler et al. [17] determined values for the stability constants of the surface complexes formed with several heavy metal ions, M(OSi-)m(n-m)+. These latter stability constants were compared with the corresponding values for simple hydroxy complexes and were found to be consistent with the occurrence of adsorption and hydrolysis at similar pH levels. With increasing pH, the surface silanol groups effectively captured dissolved metal hydroxy complexes just prior to nucleation of hydroxide precipitate. James and Healy [19,20] concluded that dissolved metals are adsorbed on silica as cationic metal hydroxy complexes which are separated from the surface by at least one layer of water molecules thus preventing direct chemical bond-
ing that would give silicate precipitation. As noted by several authors [19-25], these adsorbed metal ions do not lose their ‘primary hydration sheaths’ and are in the same form as in the aqueous solution. In addition to the fundamental studies described above, more recent laboratory scale work [26] has demonstrated the use of silica gel in an application analogous to that of the present work with silica sol. Chromium was adsorbed from synthetic effluents followed by desorption using a mineral acid to regenerate the gel and produce a more concentrated solution than the feed. Dodwell and Smith [27] have shown that certain titanium or tin silicate gels can be used to adsorb lead, cadmium, chromium and mercury both rapidly and selectively over calcium and magnesium coexisting in contaminated water. Various coatings of ion exchange materials adsorbed on to silica gel which acts as an inert substrate have been tested for removal of metal ions from aqueous solution. Such coatings, which include pyridylazo naphthol [28], calcium phosphate [29] and dipicolylamine [30], are more suitable for analytical purposes than for large scale treatment of industrial effluents. There are no studies more recent than the fundamental work of Iler [9,10] and Falcone [11] known to the authors that deal with dissolved metal adsorption on to silica sols. The mixing of what is effectively two aqueous solutions (colloidal sol and effluent) provides more intimate contact between adsorption sites and metal ions than obtainable with a gel. The handling problems associated with the fragility of silica gel are also avoided. EXPERIMENTAL Reagents and Materials Activated silica sols were prepared from sodium silicate solutions provided by National Silicates Ltd. These reagents had a SiO2/Na2O weight ratio (m) of 3.22 or 2.00. Most of the work was carried out with a silicate solution of m = 3.22 due primarily to its lower cost which would be a significant factor in a commercial application involving effluent/waste treatment. As-supplied silicate with m = 3.22 contained 28.7% SiO2, 8.90% Na2O at pH 11.3 and was designated ‘N’ silicate by National Silicates Ltd. Silicate solution with m = 2.00 (‘D’ silicate) contained 29.4% SiO2, 14.7% Na2O at pH 12.4. All chemicals, other than sodium silicate solutions, were reagent grade materials. Aqueous solutions of copper or cobalt were prepared from cupric or cobaltous sulphate salts. Lime was used in the form of an aqueous suspension containing either 0.40 g/L CaO or 28 g/L CaO depending on the experiment. CANADIAN METALLURGICAL QUARTERLY
20
E. EL-AMMOURI, S. RODGERS and P.A. DISTIN
Preparation of Silica Sols Silica sols were prepared by, firstly, diluting ‘N’ silicate by 50 v/v % water to lower viscosity and promote ease of handling. The pH of the diluted ‘N’ silicate was 11.15. While stirring with a magnetic stirrer, gelation was initiated by adding 10 w/w % sulphuric acid solution such that a 2.04 w/w % silica solution at pH 8.2 was produced. Agitation was stopped 45 seconds after acid addition. This gave a gel time of 18 minutes ± 30 seconds. Gelation was considered complete upon loss of uniform fluid flow with the simultaneous appearance of internal breakage planes when the mixture was tilted. Also, the gel adhered to the glass wall and depression of the surface was noted when a glass rod was placed on the gel. In general, gelation was arrested by dilution with 50 v/v % water at various fractions of the gel time to provide a 1.0 w/w % silica reagent in the form of a stabilized, activated sol (hereafter referred to as ‘sol’). Most tests were carried out with sol stabilized at half the gel time (9 minutes after initiating gelation). Some tests were also performed using 2.0 w/w % silica sols based on ‘N’ silicate. A few experiments with sols prepared from ‘D’ silicate were carried out. Adsorption/Desorption Cycles The experimental set up for the adsorption tests is shown in Figure 2. In most experiments, 100 mL of 0.001 M cupric or cobaltous solution (63 ppm Cu and 58 ppm Co) at pH 5.5 for copper and pH 6.8 for cobalt was stirred at room temperature with a magnetic stirrer. A few tests were carried out with feed solutions containing up to 0.315 g/L Cu and 2.83 g/L Co. Sol was added incrementally using an autotitrator and the pH was continuously monitored. Apparent equilibrium was approached about 20 minutes after each addition. In some tests with 63 ppm Cu, cupric potentials were followed with an Orion specific ion electrode. In studies with the cobaltous ion, the cupric and reference electrodes seen in Figure 2 were removed. At the end of adsorption, agitation was stopped and adsorbed copper (or cobalt)
appeared as a faintly blue (or pink) slightly opaque layer that rapidly settled to give a sharp interface with a clear overflow. The underflow was centrifuged at 2000 rpm for 10-15 minutes (IEC International Centrifuge, Model CL) to give concentrated metal loaded sol. Up to 90 v/v % of the feed was rejected as overflow after the settling and centrifuge steps following metal recovery (Figure 1). Thus, with 100 mL sulphate feed solution, the amount of concentrated sol available for desorption was too small to permit use of the arrangement shown in Figure 2. The addition of a few drops of concentrated acid to reach about pH 4 was sufficient for metal desorption. By using concentrated acid, the dilution effect inherent in using an acid solution is avoided. One goal was to produce as concentrated a solution relative to the feed as possible. A second centrifuge step gave concentrated metal solution and regenerated sol. In some cases, sol was recycled (after pH adjustment with lime). Analytical Methods Dissolved Metal Concentrations: Dissolved metal levels in solution samples after filtration were determined using a Perkin Elmer 3110 atomic absorption spectrometer (AAS). In some tests, a cupric ion selective electrode provided instantaneous in situ measurements of the approach to equilibrium when sol was added to a cupric solution. The electrode distinguishes ‘free’ dissolved cupric ion from other forms of copper such as Cu(OH)+ or copper adsorbed by the sol. In general, the electrode was used for solutions containing 63.5 ppm Cu (0.001 M) or less. At this level, the activity coefficient for the cupric ion was taken to be unity. Viscosity: The gelation process initiated by acidification of a sodium silicate solution is due to polymerization of silicate species and gives significant increases in viscosity. A 50-V705 Canon-Fenske viscometer was used to determine viscosities of sols stabilized at various fractions of the gel time. RESULTS AND DISCUSSION Sol Characteristics
Fig. 2. Metals recovery/redissolution experimental set up. CANADIAN METALLURGICAL QUARTERLY
A simplified version of the dissolution behaviour of amorphous silica under equilibrium conditions at 25 °C [10] is shown in Figure 3 which also includes information relating to reagents used in the present work. Below 0.002 M SiO2, Si(OH)4 monomer predominates in neutral and weakly acidic or weakly alkaline solution, while Si(OH)5– becomes important at pH > 11. Above both 0.002 M SiO2 and pH 8, a multimeric domain exists which contains dissolved monosilicic acid and low molecular weight polymers such as tetramer and decamer. Above the upper limit of the multimeric domain, larger polymers are formed giving colloidal particles. The as-received reagent used in most of the
DISSOLVED COPPER AND COBALT RECOVERY FROM DILUTE ACIDIC SOLUTIONS BY ADSORPTION ON SILICA SOL 21
present work (‘N’ silicate) is located as indicated on Figure 3. This latter location implies that as-supplied ‘N’ silicate is colloidal. However, ‘N’ silicate is an impure commercial product and given the proximity of location 1 to the boundary with the multimeric domain, the true location may be within this latter region thus indicating total silica solubility. The equivalent position for as-received ‘D’ silicate is clearly within the area for dissolved silica comprising low molecular weight polymers. The production of stabilized sols involved dilution of as-received reagent (to lower the viscosity), then acidification (to promote polymerization) followed by a further dilution. This latter dilution essentially arrested further polymerization and prevented gelation that would otherwise have occurred, thus creating a stabilized colloidal sol. The stabilized sol most commonly used in the present work is at position 3 on Figure 3, and comprises 1.0 w/w % silica based on ‘N’ silicate at pH 9.0.
Table I – Gel times at 25 °C for various concentrations of sodium silicate (% SiO2 / Na2O, m = 3.22 by weight) when mixed with sulphuric acid to give initial pH values shown. Weight % SiO2 1.0 1.0 1.0 1.0 1.0 1.0 2.0 2.0 2.0 2.0 2.0 2.0 4.0 4.0 4.0 4.0
Initial pH 0.5 4.9 5.6 6.7 7.6 9.0 4.5 5.7 7.1 8.2 8.7 9.9 4.8 5.8 8.9 10.6
Gel time 21 days 1005 min 85 min 33 min 187 min No gelation after 42 days 178 min 20 min 5 min 18 min 112 min 29 days 10 min 2 min 4 min 5 days
ter sols (Figure 5). The expected increase in viscosity as gel times are approached is attributable to an increasing degree of polymerization [10]. Neutralization with Sol or Lime Dissolved copper values obtained by AAS and the copper electrode were compared following sol additions that gave increased pH levels as shown in Table II. Following each incremental sol addition, stable electromotive force (emf) Fig. 3. Solubility diagram for amorphous silica at 25 °C [10]. 1. Asreceived ‘N’ silicate. 2. As-received ‘D’ silicate. 3. 1.0 w/w % SiO2 stabilized ‘activated’ silica sol.
Stabilization of acidified sol by dilution may greatly delay gelation such that a metastable structure is retained for a useful time period. This allows practical application of the sols’ properties. Examples of gel times measured in the present work are seen in Table I and compared in Figure 4 with previous results [26] for silicate with the same molar SiO2/ Na2O ratio of 3.22 as in ‘N’ silicate. Gel times are highly sensitive to silica concentration and pH. The fastest polymerization or gelling rate occurs near the pH of least ionization where repulsion forces between charged colloidal particles are at a minimum [26]. Viscosity measurements were carried out on 1.0 w/w % SiO2 sols produced by 50 v/v % water dilution of acidified 2.0 w/w % SiO2 sols. The dilutions were made to arrest gelation at various proportions of the gel times for the lat-
Fig. 4. Gel times at 25 °C for various concentrations of sodium silicate (SiO2/Na2O, m = 3.22 by weight) when mixed with sulphuric acid to give initial pH values as shown. Comparison of present data (present) with previous study (previous) [20].
CANADIAN METALLURGICAL QUARTERLY
22
E. EL-AMMOURI, S. RODGERS and P.A. DISTIN
undetectable. The most probable explanation is that some copper adsorbed previously in the experiment below pH 7 forms a colloidal product such as Cu(OH)2 which desorbs from the sol. Above pH 7 there was no visible precipitate, although colloidal copper compounds not trapped during filtration may have been reported by AAS as dissolved copper. The copper electrode gave emf readings of zero at above pH 7.5 regardless of whether sol or lime was added thus confirming the expected absence of dissolved cupric ion. If the behaviour of the two silicates is compared, ‘N’ silicate has a higher SiO2/Na2O ratio than ‘D’ silicate so that the amount of the former needed to reach a specified pH is higher than for the latter. This is seen in Figure 7 Fig. 5. Viscosity of ‘N’ silicate sols (1.0 w/w % SiO2) stabilized at different gel time fractions.
Table II – Comparison of analysis methods. Electrode measurements vs atomic absorption. Sol added as 1.0 w/w % SiO2 with m = 3.22 (‘N’ silicate) stabilized at 0.5 gel time (total gel time 18 minutes). Initial Cu solution contained 63.5 ppm Cu at pH 5.5. Sol addition (mL)
pH
ppm Cu from E (mV)
ppm Cu from AA
3 6 9 12 15 18
6.14 6.17 6.23 6.45 6.76 7.48
44.13 29.28 18.45 7.84 1.49 0.17
48.32 35.83 21.71 8.86
DISSOLVED COPPER AND COBALT RECOVERY FROM DILUTE ACIDIC SOLUTIONS BY ADSORPTION ON SILICA SOL E. EL-AMMOURI, S. RODGERS and P.A. DISTIN McGill University, Department of Mining, Metals and Materials Engineering, 3610 University Street, Montreal, QC H3A 2B2
(Received February, 2002; in revised form September, 2002)
Abstract — The ability of silica sol to neutralize and adsorb copper and cobalt from dilute, acidic solutions has been studied on a laboratory scale. At the pH for maximum adsorption (pH 7.0 for Cu and 8.3 for Co), more than 1 ppm metal remains dissolved which rises to about 30 ppm if pH is further increased. This effect is attributed to desorbed colloidal species at high pH. Final pH control with lime gave less than 1 ppm Cu or Co. The replacement of lime with sol reduces the formation of waste hydroxide/gypsum sludge. The adsorption of metal ions coagulates the sol and produces a sharply defined underflow and clear overflow. Adsorbed metal was recovered into a concentrated solution (relative to the feed) by reversing the adsorption reaction through acidification of centrifuged underflow. With feeds containing from 58 ppm to 2.83 g/L metal, product solutions from 339 ppm to 17.5 g/L metal were generated, although at 17.5 g/L a viscous suspension of agglomerates was produced. Use of recycled silica was demonstrated. Résumé — On a étudié en laboratoire la capacité d’un sol de silice à neutraliser et à adsorber le cuivre et le cobalt à partir de solutions acides diluées. Au pH d’adsorption maximum (pH 7.0 pour le Cu, 8.3 pour le Co), plus d’une ppm de métal restait dissoute, mais la quantité s’élevait à environ 30 ppm si on augmentait davantage le pH. Cet effet est attribué aux espèces colloïdales désorbées à pH élevé. Le contrôle du pH final avec de la chaux a donné moins d’une ppm de Cu ou de Co. Le remplacement de la chaux avec le sol réduit la formation de déchets d’hydroxyde/de sédiments de gypse. L’adsorption d’ions métalliques coagule le sol et produit un flot inférieur nettement défini et un flot supérieur clair. On a récupéré le métal adsorbé dans une solution concentrée (par rapport à la source) en renversant la réaction d’adsorption par l’intermédiaire de l’acidification du flot inférieur centrifugé. Avec des sources contenant entre 58 ppm et 2.83 g/L de métal, on générait des solutions de produit ayant de 339 ppm à 17.5 g/L de métal, bien qu’à 17.5 g/L, on produisait une suspension visqueuse d’agglomérés. On a démontré l’utilisation de la silice recyclée.
INTRODUCTION Dilute, weak acidic waste streams, such as acid mine drainage and mine effluents, are treated by adding lime and flocculants to a pH between 9 and 11. A voluminous metal hydroxide/gypsum tailings is formed from which metals may slowly redissolve due to gradually decreasing pH in tailings ponds. Previous laboratory scale studies [1-3] have shown how, in principle, solids waste generation can be significantly reduced, with recovery of associated metal values, by largely replacing lime with a stabilized silica sol. This
latter material acts mainly as an adsorption medium and removes dissolved metal in a manner that avoids extensive hydroxide precipitation. If adsorption/desorption on activated silica can be operated as a reversible pH-controlled process, an opportunity exists for transferring dissolved metal from a feed stream to a product stream as shown in Figure 1. Such a process has merit only if the activated silica is recyclable and if metal in the product solution is more concentrated than in the feed. When activated silica is added to a heavy metal solution to reach a controlled pH (‘Metal recovery’ step in Figure 1), coagulation of colloidal silica creates a metal bearing silica CANADIAN METALLURGICAL QUARTERLY
18
E. EL-AMMOURI, S. RODGERS and P.A. DISTIN
OH OH OH OH | | | | HO-Si-OH + HO-Si-OH Æ HO-Si-O-Si-OH + H2O (1) | | | | OH OH OH OH or in abbreviated form as | | | 2 (-Si-OH ) Æ - Si-O-Si- + H2O | | |
(2)
As described more fully in the following results and discussion section, polymerization may be initiated by adding sulphuric acid to an alkaline silicate solution. While continued polymer growth eventually gives a gel, polymerization can be effectively halted by dilution which creates a stabilized sol as is commonly done when preparing sols for water clarification [4-8]. Fig.1. Conceptual flowsheet for metals recovery using activated silica sol.
layer which rapidly settles to give a sharp interface with a clear overflow. The silica additions are much higher (several g/L) than those used for water treatment where activated silica at 1 ppm to 2 ppm SiO2 is added to flocculate suspended solids [4-8]. Acidification of the metal bearing silica layer, after discarding the overflow, gives metals desorption and regeneration of activated silica for recycle to the metal recovery step. The production of a concentrated metal solution, relative to the feed, is promoted by using concentrated sulphuric acid and water rejection steps (centrifugation in the present work) as shown in Figure 1. Direct soluble silicate addition, instead of activated silica, may give precipitates that yield solid silica upon acidification. Although solid silica also adsorbs dissolved metal ions, the mixing of silica as a sol with an aqueous solution of metal ions is expected to give more efficient silica/metal ion contact than obtainable with a suspension of solid silica particles. Copper was chosen because adsorption could be monitored in situ using a specific ion electrode for the cupric ion. When the cobaltous ion is used, adsorption can be readily distinguished from precipitation due to major differences in the colour of the products. DISSOLVED METAL ADSORPTION BY SILICA SOL, SILICA GEL OR SILICA Silica sols are created by the polymerization of silicic acid monomers, Si(OH)4, generated upon dissolution of a soluble silicate. Using dimer formation as the simplest example, polymerization can be represented as: CANADIAN METALLURGICAL QUARTERLY
Iler [9] studied the coagulation of silica sols resulting from adsorption of calcium ions from nitrate solutions. The objective was to determine the critical concentration of dissolved calcium ions needed for the coagulation of colloidal silica as a function of particle size. The present work uses this same coagulation effect for the adsorption of dissolved heavy metal ions to form a metal bearing underflow. When silica sol is added to an acidic metal-salt solution, the pH increases due to the Na2O content of the sol and adsorption sites are created for dissolved metal species. With a continuously increasing pH, the reaction path followed by dissolved metal ions passes through several intermediate hydroxy complexes before reaching the point where precipitation would be expected. Iler [10] and Falcone [11] observed that silica sols and silicate solutions began to adsorb dissolved multivalent metal ions at pH values up to 2 units below the pH at which metal hydroxide precipitated. Falcone [11] compared sodium hydroxide and silicates with various SiO2/Na2O ratios when added to copper perchlorate solutions at pH 4 to reach various pH levels up to 7. While a precipitate formed upon neutralization with sodium hydroxide, there was no visible precipitate when using silicate, although some precipitate appeared after ageing the solution for a month. Using a cupric ion selective electrode, cupric ion activities at a given pH were found to be lower after silicate than after sodium hydroxide addition. Colloidal silica adsorbs dissolved multivalent metal ions in a manner analogous to the interaction between a metal ion and a silica gel surface [10,11]. Dissolved metal adsorption on to silica gel (or silica) [12-17] may be represented as an ion exchange process. | | Mn+ + m (-SiOH ) ´ M (O Si-)m(n-m)+ + m H+ | |
(3)
DISSOLVED COPPER AND COBALT RECOVERY FROM DILUTE ACIDIC SOLUTIONS BY ADSORPTION ON SILICA SOL 19
Regardless of whether the starting material is silica sol, silica gel or solid silica, the reacting species is an acidic surface silanol group. Thus Reaction 3 can be taken as the sum of the following two reactions: | | m (-SiOH ) ´ M (SiO- ) + m H+ | |
(4)
| | Mn+ + m (- SiO-) ´ M (O Si-)m(n-m)+ | |
(5)
Reaction 3 has also been presented [18] as | | MLn+ + m (- SiOH) ´ ML (O Si-)m(n-m)+ + m H+ | |
(6)
where L is a species such as OH-. Dugger et al. [16] have studied the stoichiometry of Reaction 3 when applied to silica gel powder adsorbing a wide range of metals dissolved in buffered nitrate or perchlorate solutions. Adsorption was carried out at low pH where n could be taken as the charge on the simple unhydrolyzed metal ion. The quantity of hydrogen ions released per mole metal adsorbed was always such that n = m in Reaction 3 at equilibrium. Vydra and Galba [18] studied Reaction 6 at pH levels where hydrolyzed metal ions would be expected to form. Metals were adsorbed on to silica gel powder from buffered chloride or nitrate solutions. By assuming n = m in Reaction 6, the charge on the adsorbed complex could be deduced by measuring CM/CH where CM and CH are moles metal adsorbed and hydrogen ions released, respectively. A fractional value of CM/CH other than 1.0, 0.50 and 0.33 was taken to indicate adsorption of at least two cations of different valencies. In an experimental study of the ligand properties of surface silanol groups on silica, Schindler et al. [17] determined values for the stability constants of the surface complexes formed with several heavy metal ions, M(OSi-)m(n-m)+. These latter stability constants were compared with the corresponding values for simple hydroxy complexes and were found to be consistent with the occurrence of adsorption and hydrolysis at similar pH levels. With increasing pH, the surface silanol groups effectively captured dissolved metal hydroxy complexes just prior to nucleation of hydroxide precipitate. James and Healy [19,20] concluded that dissolved metals are adsorbed on silica as cationic metal hydroxy complexes which are separated from the surface by at least one layer of water molecules thus preventing direct chemical bond-
ing that would give silicate precipitation. As noted by several authors [19-25], these adsorbed metal ions do not lose their ‘primary hydration sheaths’ and are in the same form as in the aqueous solution. In addition to the fundamental studies described above, more recent laboratory scale work [26] has demonstrated the use of silica gel in an application analogous to that of the present work with silica sol. Chromium was adsorbed from synthetic effluents followed by desorption using a mineral acid to regenerate the gel and produce a more concentrated solution than the feed. Dodwell and Smith [27] have shown that certain titanium or tin silicate gels can be used to adsorb lead, cadmium, chromium and mercury both rapidly and selectively over calcium and magnesium coexisting in contaminated water. Various coatings of ion exchange materials adsorbed on to silica gel which acts as an inert substrate have been tested for removal of metal ions from aqueous solution. Such coatings, which include pyridylazo naphthol [28], calcium phosphate [29] and dipicolylamine [30], are more suitable for analytical purposes than for large scale treatment of industrial effluents. There are no studies more recent than the fundamental work of Iler [9,10] and Falcone [11] known to the authors that deal with dissolved metal adsorption on to silica sols. The mixing of what is effectively two aqueous solutions (colloidal sol and effluent) provides more intimate contact between adsorption sites and metal ions than obtainable with a gel. The handling problems associated with the fragility of silica gel are also avoided. EXPERIMENTAL Reagents and Materials Activated silica sols were prepared from sodium silicate solutions provided by National Silicates Ltd. These reagents had a SiO2/Na2O weight ratio (m) of 3.22 or 2.00. Most of the work was carried out with a silicate solution of m = 3.22 due primarily to its lower cost which would be a significant factor in a commercial application involving effluent/waste treatment. As-supplied silicate with m = 3.22 contained 28.7% SiO2, 8.90% Na2O at pH 11.3 and was designated ‘N’ silicate by National Silicates Ltd. Silicate solution with m = 2.00 (‘D’ silicate) contained 29.4% SiO2, 14.7% Na2O at pH 12.4. All chemicals, other than sodium silicate solutions, were reagent grade materials. Aqueous solutions of copper or cobalt were prepared from cupric or cobaltous sulphate salts. Lime was used in the form of an aqueous suspension containing either 0.40 g/L CaO or 28 g/L CaO depending on the experiment. CANADIAN METALLURGICAL QUARTERLY
20
E. EL-AMMOURI, S. RODGERS and P.A. DISTIN
Preparation of Silica Sols Silica sols were prepared by, firstly, diluting ‘N’ silicate by 50 v/v % water to lower viscosity and promote ease of handling. The pH of the diluted ‘N’ silicate was 11.15. While stirring with a magnetic stirrer, gelation was initiated by adding 10 w/w % sulphuric acid solution such that a 2.04 w/w % silica solution at pH 8.2 was produced. Agitation was stopped 45 seconds after acid addition. This gave a gel time of 18 minutes ± 30 seconds. Gelation was considered complete upon loss of uniform fluid flow with the simultaneous appearance of internal breakage planes when the mixture was tilted. Also, the gel adhered to the glass wall and depression of the surface was noted when a glass rod was placed on the gel. In general, gelation was arrested by dilution with 50 v/v % water at various fractions of the gel time to provide a 1.0 w/w % silica reagent in the form of a stabilized, activated sol (hereafter referred to as ‘sol’). Most tests were carried out with sol stabilized at half the gel time (9 minutes after initiating gelation). Some tests were also performed using 2.0 w/w % silica sols based on ‘N’ silicate. A few experiments with sols prepared from ‘D’ silicate were carried out. Adsorption/Desorption Cycles The experimental set up for the adsorption tests is shown in Figure 2. In most experiments, 100 mL of 0.001 M cupric or cobaltous solution (63 ppm Cu and 58 ppm Co) at pH 5.5 for copper and pH 6.8 for cobalt was stirred at room temperature with a magnetic stirrer. A few tests were carried out with feed solutions containing up to 0.315 g/L Cu and 2.83 g/L Co. Sol was added incrementally using an autotitrator and the pH was continuously monitored. Apparent equilibrium was approached about 20 minutes after each addition. In some tests with 63 ppm Cu, cupric potentials were followed with an Orion specific ion electrode. In studies with the cobaltous ion, the cupric and reference electrodes seen in Figure 2 were removed. At the end of adsorption, agitation was stopped and adsorbed copper (or cobalt)
appeared as a faintly blue (or pink) slightly opaque layer that rapidly settled to give a sharp interface with a clear overflow. The underflow was centrifuged at 2000 rpm for 10-15 minutes (IEC International Centrifuge, Model CL) to give concentrated metal loaded sol. Up to 90 v/v % of the feed was rejected as overflow after the settling and centrifuge steps following metal recovery (Figure 1). Thus, with 100 mL sulphate feed solution, the amount of concentrated sol available for desorption was too small to permit use of the arrangement shown in Figure 2. The addition of a few drops of concentrated acid to reach about pH 4 was sufficient for metal desorption. By using concentrated acid, the dilution effect inherent in using an acid solution is avoided. One goal was to produce as concentrated a solution relative to the feed as possible. A second centrifuge step gave concentrated metal solution and regenerated sol. In some cases, sol was recycled (after pH adjustment with lime). Analytical Methods Dissolved Metal Concentrations: Dissolved metal levels in solution samples after filtration were determined using a Perkin Elmer 3110 atomic absorption spectrometer (AAS). In some tests, a cupric ion selective electrode provided instantaneous in situ measurements of the approach to equilibrium when sol was added to a cupric solution. The electrode distinguishes ‘free’ dissolved cupric ion from other forms of copper such as Cu(OH)+ or copper adsorbed by the sol. In general, the electrode was used for solutions containing 63.5 ppm Cu (0.001 M) or less. At this level, the activity coefficient for the cupric ion was taken to be unity. Viscosity: The gelation process initiated by acidification of a sodium silicate solution is due to polymerization of silicate species and gives significant increases in viscosity. A 50-V705 Canon-Fenske viscometer was used to determine viscosities of sols stabilized at various fractions of the gel time. RESULTS AND DISCUSSION Sol Characteristics
Fig. 2. Metals recovery/redissolution experimental set up. CANADIAN METALLURGICAL QUARTERLY
A simplified version of the dissolution behaviour of amorphous silica under equilibrium conditions at 25 °C [10] is shown in Figure 3 which also includes information relating to reagents used in the present work. Below 0.002 M SiO2, Si(OH)4 monomer predominates in neutral and weakly acidic or weakly alkaline solution, while Si(OH)5– becomes important at pH > 11. Above both 0.002 M SiO2 and pH 8, a multimeric domain exists which contains dissolved monosilicic acid and low molecular weight polymers such as tetramer and decamer. Above the upper limit of the multimeric domain, larger polymers are formed giving colloidal particles. The as-received reagent used in most of the
DISSOLVED COPPER AND COBALT RECOVERY FROM DILUTE ACIDIC SOLUTIONS BY ADSORPTION ON SILICA SOL 21
present work (‘N’ silicate) is located as indicated on Figure 3. This latter location implies that as-supplied ‘N’ silicate is colloidal. However, ‘N’ silicate is an impure commercial product and given the proximity of location 1 to the boundary with the multimeric domain, the true location may be within this latter region thus indicating total silica solubility. The equivalent position for as-received ‘D’ silicate is clearly within the area for dissolved silica comprising low molecular weight polymers. The production of stabilized sols involved dilution of as-received reagent (to lower the viscosity), then acidification (to promote polymerization) followed by a further dilution. This latter dilution essentially arrested further polymerization and prevented gelation that would otherwise have occurred, thus creating a stabilized colloidal sol. The stabilized sol most commonly used in the present work is at position 3 on Figure 3, and comprises 1.0 w/w % silica based on ‘N’ silicate at pH 9.0.
Table I – Gel times at 25 °C for various concentrations of sodium silicate (% SiO2 / Na2O, m = 3.22 by weight) when mixed with sulphuric acid to give initial pH values shown. Weight % SiO2 1.0 1.0 1.0 1.0 1.0 1.0 2.0 2.0 2.0 2.0 2.0 2.0 4.0 4.0 4.0 4.0
Initial pH 0.5 4.9 5.6 6.7 7.6 9.0 4.5 5.7 7.1 8.2 8.7 9.9 4.8 5.8 8.9 10.6
Gel time 21 days 1005 min 85 min 33 min 187 min No gelation after 42 days 178 min 20 min 5 min 18 min 112 min 29 days 10 min 2 min 4 min 5 days
ter sols (Figure 5). The expected increase in viscosity as gel times are approached is attributable to an increasing degree of polymerization [10]. Neutralization with Sol or Lime Dissolved copper values obtained by AAS and the copper electrode were compared following sol additions that gave increased pH levels as shown in Table II. Following each incremental sol addition, stable electromotive force (emf) Fig. 3. Solubility diagram for amorphous silica at 25 °C [10]. 1. Asreceived ‘N’ silicate. 2. As-received ‘D’ silicate. 3. 1.0 w/w % SiO2 stabilized ‘activated’ silica sol.
Stabilization of acidified sol by dilution may greatly delay gelation such that a metastable structure is retained for a useful time period. This allows practical application of the sols’ properties. Examples of gel times measured in the present work are seen in Table I and compared in Figure 4 with previous results [26] for silicate with the same molar SiO2/ Na2O ratio of 3.22 as in ‘N’ silicate. Gel times are highly sensitive to silica concentration and pH. The fastest polymerization or gelling rate occurs near the pH of least ionization where repulsion forces between charged colloidal particles are at a minimum [26]. Viscosity measurements were carried out on 1.0 w/w % SiO2 sols produced by 50 v/v % water dilution of acidified 2.0 w/w % SiO2 sols. The dilutions were made to arrest gelation at various proportions of the gel times for the lat-
Fig. 4. Gel times at 25 °C for various concentrations of sodium silicate (SiO2/Na2O, m = 3.22 by weight) when mixed with sulphuric acid to give initial pH values as shown. Comparison of present data (present) with previous study (previous) [20].
CANADIAN METALLURGICAL QUARTERLY
22
E. EL-AMMOURI, S. RODGERS and P.A. DISTIN
undetectable. The most probable explanation is that some copper adsorbed previously in the experiment below pH 7 forms a colloidal product such as Cu(OH)2 which desorbs from the sol. Above pH 7 there was no visible precipitate, although colloidal copper compounds not trapped during filtration may have been reported by AAS as dissolved copper. The copper electrode gave emf readings of zero at above pH 7.5 regardless of whether sol or lime was added thus confirming the expected absence of dissolved cupric ion. If the behaviour of the two silicates is compared, ‘N’ silicate has a higher SiO2/Na2O ratio than ‘D’ silicate so that the amount of the former needed to reach a specified pH is higher than for the latter. This is seen in Figure 7 Fig. 5. Viscosity of ‘N’ silicate sols (1.0 w/w % SiO2) stabilized at different gel time fractions.
Table II – Comparison of analysis methods. Electrode measurements vs atomic absorption. Sol added as 1.0 w/w % SiO2 with m = 3.22 (‘N’ silicate) stabilized at 0.5 gel time (total gel time 18 minutes). Initial Cu solution contained 63.5 ppm Cu at pH 5.5. Sol addition (mL)
pH
ppm Cu from E (mV)
ppm Cu from AA
3 6 9 12 15 18
6.14 6.17 6.23 6.45 6.76 7.48
44.13 29.28 18.45 7.84 1.49 0.17
48.32 35.83 21.71 8.86
