Industrial & Engineering Chemistry Research - ACS Publications

---

Ind. Eng. Chem. Res. 1997, 36, 399-406

399

SEPARATIONS Zinc, Cadmium, and Lead Separation from Aqueous Streams Using Solid-Phase Extractants Nandkumar V. Deorkar and Lawrence L. Tavlarides* Department of Chemical Engineering and Materials Science, Syracuse University, Syracuse, New York 13244

The separation of cadmium, zinc, and lead from aqueous solutions using inorganic solid-phase extractants has been examined. An inorganic solid-phase extractant (ISPE-302) was prepared by attachment of bis (2,4,4-trimethyl)monothiophosphonic acid on a functionalized silica surface. Results of preparation and characterization studies are presented here. FTIR spectral studies show van der Waals type interactions between the alkyl chain of the extractant and the functional groups bonded to the silica surfaces. ISPE-302 exhibits the selectivity series of cadmium > lead > zinc. The initial capacity of 0.175 × 10-3 mol of cadmium/g of ISPE-302 achieved is 89% of the theoretical capacity. The capacity of ISPE-302 decreases to 0.5 × 10-4 mol/g, which is likely due to loss of extractant. Cadmium, lead, and zinc were separated by selective stripping from the bed saturated with these metal ions. These results indicate ISPE-302 has potential for separation of these metal ions from dilute aqueous solutions. Introduction Disposal of hazardous ions in aqueous waste streams is a significant industrial waste problem. The basic metals aluminum, cadmium, chromium, cobalt, copper, iron, lead, mercury, nickel, and zinc have been classified as the 10 metals of primary importance for recovery from industrial waste streams (Patterson, 1987). Solvent extraction is a very useful method for selectively separating and concentrating metal ions from complex aqueous solutions because a great number of selective ligands are available which can be used as metal extractants. It requires mixing of the phases to provide sufficient interfacial area for a satisfactory rate of extraction. Separation by solvent extraction processes is generally considered to be economical in the range of aqueous metal concentration from 0.01 to 1.0 mol/dm3 (Akita and Taekuchi, 1990). The recovery of metal ions from dilute solutions using selective adsorbents has been proposed as a technological alternative to solvent extraction techniques for dilute solutions. Recently, various materials like iron oxide (Theis et al., 1992; Wang et al., 1994), activated alumina (Sood et al., 1989), and activated carbon (Ferro-Garcia et al., 1990) were used for removal of metal ions. However, efficiency of adsorption is limited by the phase equilibrium properties of various solutes and the ease with which the sorbent can be regenerated (Tedder et al., 1992). Therefore, there is a need to develop solid-phase extractants which can have high selectivity identical to that of the liquid organic phase with the ability to treat dilute aqueous solutions. It was recognized that chelating ion-exchange resins could fulfill these requirements. Ion-exchange processes between aqueous solutions and organic resins have been extensively studied and are well documented. Over the last decade, great efforts were made to develop polymeric chelating resins and solvent impregnated resins. The polymeric support in impregnated resins serves as a reservoir which can be filled to a certain capacity with an extractant. In order to have high mass-transfer S0888-5885(96)00415-0 CCC: $14.00

rates, impregnated resins should present a high mobility of the extractant in the resin phase, a high mobility of the metal ion between the resin and aqueous phase, and a high hydrophilic balance of the resin (Warshawsky, 1981). The extraction efficiency of tributyl phosphate (extractant) in solvent-impregnated resins is reduced at high extractant concentrations due to the restricted mobility of the reagent in the resin phase (Warshawsky et al., 1979). Other processing problems such as irreversible adsorption of organics on the hydrophobic polymeric network, swelling of resin, slow kinetics, low mechanical strength, and radiolytic stability may need to be overcome for a universal use of ion-exchange and impregnated macroporous resins for treatment of aqueous waste streams. It is desirable to overcome these limitations and maintain the orders of magnitude of selectivity and the convenience of columnar separations. Immobilization of the desired extractant on inorganic polymers could accomplish this goal. Inorganic supports have been used successfully as the stationary phases for extraction chromatographic separations (Braun and Ghersini, 1975). This technique has mainly been applied for preconcentration and separation of trace elements and radiochemical separations. Despite the promising performance, this technique remains largely an analytical tool. Over the last decade a great effort has been made toward immobilization of various chelating agents on silica gels by covalent bonding (Jezorek and Freiser, 1979; Marshall and Mottola, 1983; Plueddeman, 1984; Bradshaw et al., 1992; Lindoy and Eaglen, 1993). The studies were concentrated on the immobilizations of specific extractants such as crown ethers, hydroxyquinoline, etc. We also have developed various chemically active inorganic adsorbents by chemical bridging/covalent attachment of desired chelating agents to ceramic supports (Deorkar and Tavlarides, 1994). This paper describes the preparation of an inorganic solid-phase extractant (ISPE) by attaching bis(2,4,4trimethylpentyl)monothiophosphinic acid (Cyanex-302; © 1997 American Chemical Society

400 Ind. Eng. Chem. Res., Vol. 36, No. 2, 1997

American Cyanamide Co.) on a ceramic support through van der Waals interactions for separation of cadmium, lead, and zinc. Methods of functionalization of supports and attachment of extractants to optimize functionalization and to increase stability and efficiency of adsorption of metal ions are discussed. The effect of pH on the adsorption of metal ions with ISPE-302, and adsorption capacity, stability, and regeneration efficiency of ISPE-302 were determined. In addition, the separation of cadmium, lead, and zinc by selective stripping with this material is demonstrated. Experimental Section Materials and Apparatus. Bis(2,4,4-trimethylpentyl)monothiophosphinic acid (Cyanex-302; assay 84%) was obtained from American Cyanamide Co. (Wayne, NJ) and used without further purification. Organic solvents such as toluene, xylene, carbon tetrachloride, and chloroform were obtained from Aldrich Chemical Co. in the highest purity, typically g99%. Silica gel (particle size, 70-270 mesh; pore size, 60 Å; surface area, 500 m2/g) and dichlorodimethylsilane were obtained from Aldrich Chemical Co. Analytical-grade metal salts 3CdSO4‚8H2O, Zn(NO3)2‚ 6H2O, and Pb(NO3)2 (Aldrich Chemical Co.) were used to prepare 1.8 × 10-2 mol/dm3 stock solutions. All other chemicals were certified reagent grade unless specified otherwise. Buffer solutions covering the pH range of 3.5-6.5 were prepared with acetic acid and sodium acetate. A chemcade pH meter (Cole Parmer) and an atomic absorption spectrophotometer (Model No. 2380, Perkin Elmer) were used for pH adjustment and metal ion analysis. A Nicolet FTIR impact 400 (4000-400 cm-1) equipped with a Spectratech collector diffuse reflectance unit was used to record FTIR spectra. The spectra were plotted as Kubelka-Munk format. Elemental analysis of functionalized silica was performed by Oneida Research Services, Whitesboro, NY. Preparation of Inorganic Solid-Phase Extractant (ISPE). ISPE-302 was prepared by immobilization of Cyanex-302 on the functional groups bonded to the silica gel surface. The bonded functional group imparts hydrophobic character to the surface. Various silylating reagents can be used to prepare bonded phases. In general, surface hydroxyl groups can be reacted with various silylating agents to prepare functionalized silica (eq 1). Similar polar bonded phases and nonpolar bonded

tSiOH + XnSi(R)4-n f t(SiO)fSiXn-f(R)4-n (1) phases are widely used in reverse-phase liquid chromatography (Unger, 1990) where X ) Cl, OCH3, OC2H5; f ) 1-3; n ) 1-3; and R ) -(CH2)mCH3 where m ) 0-8. The silica gel was reacted with dichlorodimethylsilane (DCMS) to prepare bonded functional groups (eq 2).

2tSiOH + Cl2Si(CH3)2 f (tSiO)2Si(CH3)2

(2)

After reaction, 100 g of silica yielded 111.4 g of methyl group bonded silica. The functionalized silica (20 g) was mixed well with a solution of Cyanex-302 in toluene (5% v/v) with a rotating flask (for about 1 h) so that the Cyanex-302 solution penetrates well inside the pores of the silica gel particles. After complete removal of toluene by evaporation at full vacuum (23-24 mm), the silica (ISPE-302) was found to weigh 24.3 g. Characterization of ISPE-302. The synthesized materials were characterized for coverage density of

bonded functional groups and chelating molecules by performing elemental analysis and potentiometric titrations. The effect of pH on the adsorption of cadmium, zinc, and lead on ISPE-302 was determined by the following batch equilibrium procedure. A total of 25 cm3 of a 1.78 × 10-3 mol/dm3 metal ion solution, adjusted to suitable pH with acetate buffer, was equilibrated with 1.0 g of ISPE-302 for 1 h at room temperature (∼20 °C). The solution was then filtered through Whatmann filter paper, and the concentration of metal ions in the filtrate was determined with an atomic absorption spectrophotometer. These are not equilibrium experiments, but they provide information on relative levels of adsorption with pH to define the range of suitable pH for use of the adsorbent. The corresponding moles of metal ion present in solution is approximately the amount needed to complex with 10-20% of the theoretical sites. These experiments provide the relative degree of extraction of a given metal ion with the solution of a given pH. Metal ions adsorbed on the ISPE-302 bed were desorbed with varying concentrations of different mineral acids to determine suitable stripping agents for regeneration of the bed. The bed comprised of =5.0 g of ISPE-302 was first loaded with a known amount of the desired metal ions. Then, the stripping agent solution was passed through the bed to strip out metal ions, and metal ion concentration in the effluent was measured periodically to evaluate the regeneration efficiency and percent recovery. The capacity of ISPE-302 for a given metal ion was determined by performing breakthrough studies on the packed beds. A glass column (1 cm in diameter, 25 cm in length) was packed with a known amount of ISPE302 (4.87 g), supported between glass wool giving a bed height of 10 cm (bed volume ) 7.85 cm3). An aqueous solution containing cadmium ions (=1.78 × 10-3 mol/ dm3) was passed through the column at a flow rate of 1-1.2 cm3/min to provide a residence time of about 7-8 min. The cadmium ion concentration in the effluent was measured periodically, with an atomic absorption spectrophotometer, to determine the column breakthrough volume, i.e., the volume at which the minimum threshold concentration (8.0 × 10-6 mol/dm3) of cadmium ions in the effluent is exceeded. The capacity of ISPE-302 was calculated from the total amount of cadmium adsorbed on the bed at saturation of the bed. Similar breakthrough curve studies were executed with an aqueous stream containing lead, cadmium, and zinc to determine the effectiveness of ISPE-302 for simultaneous removal of these metal ions. Further, differences in the adsorption affinity of ISPE-302 for zinc, lead, and cadmium were exploited to separate zinc, lead, and cadmium by selective stripping. It should be noted that these breakthrough curves provide the adsorbent equilibrium capacity at the pH and ion concentrations of the feed solutions. This information is of importance for industrial application. FTIR Spectroscopic Studies. FTIR spectroscopic studies were carried out to investigate extractantfunctionalized silica interactions and complexation of ISPE-302. FTIR spectra of the functionalized silica, ISPE-302, ISPE-302-cadmium complex, ISPE-302-lead complex, and ISPE-302-zinc complex were recorded with a Nicolet 5000 Fourier transform infrared spectrometer. DRIFT (diffuse reflectance infrared Fourier transform)

Ind. Eng. Chem. Res., Vol. 36, No. 2, 1997 401

sampling technique was used to avoid interference due to the spectrum of the bulk silica. The complexes of cadmium, lead, and zinc were prepared by equilibrating 1.0 g of ISPE-302 with 100 cm3 of a 2.23 × 10-1 mol/dm3 solution of metal ions for 10 h to saturate all the adsorption sites. After filtration, the solids (ISPE-302-metal complex) were washed with water (25 cm3) and were air dried to remove excess water from the pores. Results and Discussion Coverage Density of Functional Groups on Silica. The ability of silica gel supports to bind a chelating agent (Cyanex-302) satisfactorily depends on several factors: the properties of the support surface (surface area, pore size), the nature of bonded groups which cover the surface and its coverage density, and the properties of the chelating agent (solubility in aqueous solution, viscosity, hydrocarbon chain, and its interaction with groups which cover the surface of the support). The hydroxyl groups of the silica were converted to methyl groups to obtain functionalized silica. Elemental analysis is used to quantitatively evaluate the degree of the surface treatment. Further, the surface concentration, R (coverage density), of surface-bonded species was calculated from carbon analysis by the following equation (Unger et al., 1976):

Rexp ) W/MSBET

(3)

W ) weight of the functional group (per g of silica), M ) molecular weight of the bonded functional group (g/ mol), and SBET ) specific surface area of the starting silica (m2/g). The carbon load (3.99%) and coverage density (4.21 × 10-6 mol/m2) for silica was obtained when it was treated with DCMS. Assuming the stoichiometry of the surface reaction with DCMS is 2, the coverage density of 4.21 × 10-6 mol/m2 indicates that most of the hydroxyl groups are converted to methyl because the surface hydroxyl concentration on the silica surface is about 8.0 × 10-6 mol/m2 or 4.8 hydroxyl groups/nm2 (Unger et al., 1976). In the event of low surface coverage, the large concentration of surface hydroxyl groups will decrease the stability of the ISPE due to nonuniform coverage of methyl groups. Also, these silanol groups will adsorb metal ions which lead to a lack of selectivity. In order to achieve maximum conversion, we have employed a molar excess of DCMS. As the reaction rate is controlled by diffusion of the reactant to the active surface sites within the porous silica particles, the reaction temperature and time were maintained as high as possible. However, the temperature should not exceed dehydroxylation temperatures at which unreactive siloxane bonds are formed by dehydroxylation at hydroxyl groups. Theoretically, the functionalization of porous silica leads to a decrease in the mean pore diameter and surface area. These effects are more pronounced with an increase in the chain length of the functional group and/or multilayered functionalization. Thus, proper functionalization is necessary to override complete blocking of the pores which will decrease the rate of metal ion diffusion into the pores although it might increase the stability and capacity of the ISPE. At present no attempts were made to quantitatively determine pore structure due to an increase in the chain

Figure 1. Effect of pH of aqueous solutions on the adsorption of metal ions on ISPE-302; equilibration time ) 1 h; wt of ISPE-302 ) 1 g; Vaq ) 25 mL; [Cd2+]initial ) 1.78 × 10-3 mol/dm3.

length of the functionalized silica and its effect on kinetics, stability, and capacity. However, care has been taken to avoid multilayer or polymer layer formation of DCMS on the silica surface. Physisorbed water was quantitatively removed from the surface; otherwise, the DCMS may be hydrolyzed and condensed, yielding physisorbed organosilicon polymers. Theoretical Capacity of ISPE-302. The amount of Cyanex-302 immobilized on the functionalized silica was confirmed by potentiometic titration of ISPE-302 in ethanol with 0.1 N NaOH. The concentration of Cyanex-302 in ISPE-302 is (0.4 ( 0.02) × 10-3 mol/g. Assuming 1:2 stoichiometry of cadmium ions with Cyanex-302, the capacity of ISPE for cadmium should be approximately (0.20 ( 0.01) × 10-3 mol/g. This capacity does not consider equilibrium limitations of eq 4 and possible unaccessability of ligands due to diffusional limitations in small pores. In order to achieve high rates of adsorption, we intend to have a monolayer of Cyanex-302 on the functionalized silica to minimize pore restrictions on diffusion through pores. The functionalized silica might hold higher than 0.4 × 10-3 mol/g of Cyanex-302, which may lead to multilayers of Cyanex-302. By comparing the coverage density of DCMS on silica (4.21 × 10-6 mol/m2 ≡ 2.1 × 10-3 mol/g), the concentration of Cyanex-302 on the functionalized silica (0.4 × 10-3 mol/g), the molar volume of Cyanex-302 (0.4 × 10-3 mol ≡ 0.13 cm3), and the pore volume of silica (0.75 cm3/g), it can be concluded that Cyanex-302 is likely to be immobilized as a monolayer and that the pores are not completely filled with Cyanex-302. No further attempts are made at this time to determine the porosity of the ISPE-302 silica particles and the status of coverage. Metal Ion Uptake as a Function of pH. Since the metal cations react with the acidic chelating agent to form complexes by releasing protons, the mechanism of metal ion adsorption on ISPE-302 can be described as z Maq + nRH h RnM(z-n)+ + nH+ +

(4)

where an overbar represents surface sorbed species. Figure 1 illustrates that near-quantitative removal of cadmium is possible at all pH values studied, whereas the percent extraction of lead and zinc is very low at pH 1.0. The removal efficiency of lead and zinc increased with pH, and near-quantitative removal is

402 Ind. Eng. Chem. Res., Vol. 36, No. 2, 1997

Figure 2. Elution curve of cadmium from the ISPE-302 bed: flowrate ) 1.0-1.2 cm3/min; % recovery ) 98.7 and 49.0 with 2.0 mol/dm3 HCl and 1.0 mol/dm3 HCl, respectively.

possible at pH 2.5 and 3.5, respectively. ISPE-302 shows selective adsorption of cadmium, zinc, and lead over calcium and magnesium as the percent extraction of these metals is very low over the pH range 1-6.5. Thus, this pH dependency can be exploited for selective removal and separation of these metal ions. Regeneration of ISPE-302. The regeneration of ISPE-302 saturated with metal ions was carried out by desorbing the metal ions from the ISPE bed by hydrogen exchange with various acidic stripping agents (reverse step of eq 4). Mineral acids such as hydrochloric acid, nitric acid, and sulfuric acid were used to strip out the metal ions from the beds. As shown in Figure 2, cadmium can be quantitatively recovered by stripping with 3.0 bed volumes of 2.0 mol/ dm3 hydrochloric acid, whereas 9.0 bed volumes of 1.0 mol/dm3 hydrochloric acid could recover only 49% of the adsorbed cadmium. It is not possible to desorb cadmium ions with 0.5 mol/dm3 or lower concentrations of nitric acid and hydrochloric acid. Similar studies with lead show that lead can be recovered quantitatively by stripping with 8 bed volumes of 0.5 mol/dm3 nitric acid. In order to increase the concentration factor, bed volumes of stripping agents for quantitative recovery can be decreased by increasing the concentration of stripping agents. For example, lead was quantitatively recovered by stripping with 3 bed volumes of 1.0 mol/ dm3 nitric acid. However, the immobilized extractant must be hydrolytically stable at that acid concentration. Stability studies with Cyanex-302 showed no significant degradation after contacting with 3.0 mol/dm3 sulfuric acid for 640 h at 50 °C (Rickelton and Boyle, 1990). In general, thiophosphates may hydrolyze after prolonged contact with highly concentrated nitric acid (>6 mol/ dm3). Stripping studies with zinc show that zinc can be quantitatively recovered by stripping with 8.2, 4.0, and 2.0 bed volumes of 0.05, 0.1, and 0.5 mol/dm3 nitric acid, respectively. Capacity and Stability of ISPE-302. Breakthrough curve studies on the prepared ISPE-302 were conducted to determine the behaviors of its chelation (adsorption) sites and its maximum capacity. As shown in Figure 3, the breakthrough curve of cadmium at cycle 1, cadmium ions are removed to less than 8.0 × 10-6 from a 1.78 × 10-3 mol/dm3 aqueous stream for up to 50 bed volumes. An initial capacity of 0.175 × 10-3 mol/g (19.7 mg/g) is achieved. This value is 89% of the theoretical capacity calculated based on the concentration of Cyanex-302 on the surface. The sharp break-

Figure 3. Breakthrough curves of cadmium on the ISPE-302 bed: flowrate ) 1.0-1.2 cm3/min; 1st cycle, [Cd2+]feed ) 2.0 × 10-3 mol/dm3, pH ) 6.5 (acetate buffer), uptake capacity ) 0.175 × 10-3 mol/g; 20th cycle, [Cd2+]feed ) 1.5 × 10-3 mol/dm3, pH ) 6.5 (acetate buffer), uptake capacity ) 0.5 × 10-4 mol/g. Plateau data after saturation are not shown.

through curve profile at a residence time of 7 min suggests that favorable equilibrium and rates of adsorption are achieved. It can also be concluded that the hydrophobic carbon chain of Cyanex-302 might be strictly adhered to the surface due to van der Waals interactions with bonded methyl groups, leaving the pores open to facilitate fast diffusion of cadmium ions to the thiophosphic acid site. Detailed kinetic rate date experiments and analysis are needed to support this conclusion. Similar breakthrough studies of zinc from sulfate solutions with D2EHPA-impregnated macroporous resin and zinc from chloride solutions with octylamine-impregnated macroporous resin showed no clear breakthrough point due to slow sorption rates and a wide sorption zone (Juang and Su, 1992; Akita and Taekuchi, 1990). In order to determine the stability of ISPE-302, multiple cylces of adsorption and stripping were conducted. Figure 3 illustrates that, at the 20th cycle, cadmium ions are removed to less than 8.0 × 10-6 mol/ dm3 from a 1.5 × 10-3 mol/dm3 aqueous solution for up to 16 bed volumes. Thus, the capacity after 20 cycles is decreased to 0.5 × 10-4 mol/dm3. The decrease in capacity can be attributed to the loss of extractant due to solubilization into aqueous solution and/or destruction of the complexing moiety. Studies on the distribution of Cyanex-302 between the functionalized silica and aqueous phase will determine the extent of loss of Cyanex-302 during the breakthrough curve due to solubilization. The distribution ratio can be increased by immobilizing extractants having long straight hydrocarbon chains on large-pore silica functionalized with ethyl, propyl, butyl, or hexyl groups. Alternatively, analogs of Cyanex-302 can be covalently bonded to the ceramic support to minimize the loss of extractant and increase the stability (Tavlarides and Deorkar, 1994). Removal of Cadmium, Zinc, and Lead from Dilute Solutions. The adsorption efficiency of ISPE302 for simultaneous removal of cadmium, zinc, and lead was determined by executing breakthrough curves on the ISPE-302 bed. As shown in Figure 4, 140 bed volumes of an aqueous solution at pH 6.5 containing 3.5 × 10-4 mol/dm3 cadmium, 6.11 × 10-4 mol/dm3 zinc, and 9.6 × 10-5 mol/dm3 lead were passed through the bed. The pH value was selected based on Figure 1 as suitable to adsorb all the metals simultaneously. It is

Ind. Eng. Chem. Res., Vol. 36, No. 2, 1997 403

Figure 4. Simultaneous removal of zinc, lead, and cadmium with a ISPE-302 bed: flowrate ) 1.0-1.2 cm3/min; pH ) 6.3; [M2+]feed: [Zn2+] ) 6.11 × 10-4 mol/dm3, [Cd2+] ) 3.55 × 10-4 mol/dm3, [Pb2+] ) 0.96 × 10-4 mol/dm3.

noted that various leaching solutions such as dilute sulfuric acid (pH ) 1.5-3.0), sulfuric acid/lime (pH ) 6.0), sodium acetate (pH ) 5.0), and ammonium carbonate are utilized for leaching of metals from soil and sludge (Radha Krishnan et al., 1993). All three metal ions are removed to less than 4.8 × 10-6 mol/dm3 for up to 61 bed volumes. The breakthrough of zinc occurred at 61 bed volumes, although both lead and cadmium continue to be removed to less than 8.0 × 10-6 mol/dm3. After 70 bed volumes the concentration of zinc in the effluent from the bed is greater than the feed concentration, while lead and cadmium are being adsorbed. The increase in the zinc concentration in the effluent may be due to displacement of zinc from the bed by lead and cadmium, since lead and cadmium have higher affinities for the active sites than zinc. Similar behavior of displacement of adsorbed lead after 104 bed volumes shows that cadmium has the highest affinity for ISPE-302. Separation of Cadmium, Zinc, and Lead. The effect of pH on adsorption and stripping studies on ISPE-302 shows the selectivity series of Cd2+ > Pb2+ > Zn2+. This selectivity is further exploited to separate zinc, lead, and cadmium by selective stripping. As shown in Figure 5A, zinc, lead, and cadmium were separated by selective stripping with 0.1 mol/dm3 nitric acid, 1.0 mol/dm3 nitric acid, and 2 mol/dm3 hydrochloric acid, respectively, from the ISPE-302 bed loaded with 1.6 × 10-5, 4.4 × 10-5, and 2.4 × 10-5 mol/g of zinc, cadmium, and lead, respectively. When zinc was stripped with 0.1 mol/dm3 nitric acid, lead was costripped, whereas no cadmium was removed. The concentrations of zinc and lead in the third effluent fraction were 8.44 × 10-3 and 3.4 × 10-5 mol/dm3, respectively. The total volume of each effluent fraction of stripping agent was 5.0 mL. In all other effluent fractions the lead concentration did not exceed 4.5 × 10-5 mol/dm3. Similarly, when lead was stripped with 1 mol/dm3 nitric acid, trace/minor amounts of cadmium were costripped with lead. However, costripping occurred after the fourth fraction, when most of the lead was stripped. The concentration of cadmium in all the effluent fractions did not exceed 6 ppm. Finally, cadmium was recovered by stripping with 2 mol/dm3 hydrochloric acid, with no traces of zinc or lead in any of the effluent fractions. In order to minimize the costripping of lead with zinc and cadmium with lead, different concentrations of stripping solutions were used (Figure 5B). When zinc was stripped with 9 bed volumes of 0.05 mol/dm3 nitric acid, no lead and cadmium was costripped. The percent

Figure 5. Separation of zinc, lead, and cadmium by selective stripping from a ISPE-302 bed: flowrate ) 1.0-1.2 cm3/min. (A) Stripping agents: zinc ) 0.1 mol/dm3 nitric acid, lead ) 1.0 mol/ dm3 nitric acid, cadmium ) 2.0 mol/dm3 hydrochloric acid. (B) Stripping agents: zinc ) 0.05 mol/dm3 nitric acid, lead ) 0.5 mol/ dm3 nitric acid, cadmium ) 2.0 mol/dm3 hydrochloric acid.

recovery of zinc was 99%. The percent recovery of 89% for lead indicates that lead could not be stripped completely with 9 bed volumes of 0.5 mol/dm3 nitric acid. The remaining lead ions were stripped when 2 mol/dm3 hydrochloric acid was passed through the column to strip cadmium. Most of the lead ions were recovered in the first fraction of this strip. Thus, separation of zinc, lead, and cadmium is feasible by selective stripping. Evaluation of the Interaction between Cyanex302 and the Functionalized Silica. The IR spectra of neat Cyanex-302, the functionalized silica, and ISPE302 are shown in Figure 6. The spectrum for Cyanex302 exhibits triplet bands at 2975-2800 cm-1 due to stretching modes of aliphatic CH groups (CH3), a band at 1490 cm-1 due to CH2 scissoring, and a band at 1375 cm-1 due to CH3 deformation. Also a weak and diffuse band at 2360 cm-1 and sharp bands at 909 and 810 cm-1 are observed due to POH stretching, PO stretching, and PdS stretching, respectively. The IR spectrum of the functionalized silica exhibits three bands at 2975 (strong, sharp), 2915, and 2800 cm-1 attributed to the stretching modes (asymmetric and symmetric) of aliphatic CH groups. Also, a sharp and strong band at 1350 cm corresponding to CH deformation of CH3 and at 840 cm-1 corresponding to SiC rocking and SiC stretching are observed. These bands correspond to the infrared frequencies for organic silicon compounds (Socrates, 1980). The IR spectrum of ISPE-302 (Cyanex-302 immobilized on the functionalized silica) shows that the retention of Cyanex-302 by the functionalized silica involves significant shift in the normal modes of bands associated with the methyl group (-CH3) of silane. For example, a sharp band at 1350 cm-1 associated with deformation of CH of CH3 has shifted to 1300 cm-1. Also, a band at 2975 cm-1 associated with CH stretching has been broadened. This broadening might be due to the overlap between absorption due to chains of Cyanex-302 mol-

Inorganic solid-phase extractants. e Capacity after 20 adsorption/stripping cycles. d

Immobilized macroporous resins. b 1,4-Dithia-19-crown-6. c 1,4,7,10-Tetrathia-18-crown-6.

5.0 × 10-5 mole of cadmium/g functionalized silica ISPEd

silica

silica

inorganic adsorbent inorganic adsorbent

chelating resins

DVB-acrylonitrile copolymer DVB copolymer chelating resins

chelating resins

Cyanex-302 (surface deposition)

for T219C6b 0.4 × 10-3 mol/g of Ag(I), Hg(II), and Pb(II); for T418C6c 0.3 × 10-3 mol/g 0.175 × 10-3 mol of Cd(II)/g

P capacity 1.64 × 10-3 mol/g, 1.70 × 10-3 calcium/g P capacity 0.68 × 10-3/g

=2.2 × 10-5 mol of Cd(II)/g, =0.1 × 10-3 mol of Cu(II)/g 4.4 × 10-6 mol of Cu(II)/g, 4.8 × 10-6 mol of Pb(II)/g 0.189 × 10-3 mol of Cd(II)/g gel-type styrene-2% DVB Amberlite XAD-4 chelating resins

azobenzylphosphoric acid (covalent bonding) poly(aminophosphonic acid) (covalent bonding) poly(acrylamidoxime) (covalent bonding) diphosphonic acid (covalent bonding) gem-diphosphonic acid groups (covalent bonding) thiamacrocycles (covalent bonding)

=0.55 × 10-4 mol of Zn/g =9.8 × 10-5 mol of Cu(II)/g tri-n-octylamine hydroxyoxime XAD-2 XAD-2 IMRa IMRa

a

this study

Bruening et al., 1991

due to high binding constant, complex stripping agents such as thiourea and EDTA are required separation of lead, cadmium, and zinc by selective stripping

Chiarizia et al., 1995

Colella et al., 1980 for preconcentration of metal ions

Chiarizia et al., 1995

Yebra-Biurrun et al., 1992 for preconcentration of metal ions

Ueda et al., 1988

ref

Akita and Taekuchi, 1990 Warshawski, 1981

no clear breakthrough point at 2.56 and 28.6 min residence time no clear breakthrough reagent complexations efficiency was 28%

remarks capacity

=0.174 × 10-3 mol of Zn/g D2EHPA (1.03 × 10-3 mol/g) XAD-2 IMRa

extractant/active group (immobilization method) support adsorbent

Table 1. Comparisons of Adsorbents Used for Metal Ion Removal

ecules and methyl groups bonded to the silica surface. These observations suggest that there are interactions between Cyanex-302 and the alkyl chain (methyl) of the functionalized silica which are responsible for Cyanex302 retention on the support. Similar IR spectroscopic studies with impregnated microporous resins reported weak interactions of extractant with polymeric supports such as Amberlite XAD-7 and XAD-2 (Cortina et al., 1994; Bokobza and Cote, 1985). Recently, it has been shown that the bonded alkyl chain is of primary importance in the retention of organic compounds on reverse-phase packing materials as retention time increases with an increase in the alkyl chain length (Barrett et al., 1996). Therefore, stable ISPE-302 can be produced by using large-pore functionalized silica with longer alkyl chain length for immobilization of Cyanex-302. The functionalization can be achieved by silinization reaction shown in eq 1. In addition, the IR spectrum for ISPE-302 shows a broad and diffuse band at 2360-2300 cm-1 corresponding to POH stretching and a band at 1490 cm-1 due to CH2 scissors vibration. No bands are observed in PO and PdS stretching because the sampling technique used (DRIFT) could not detect weak bands below 900 cm-1. Also, in this region, sharp bands due to SiC rocking vibration were present which interfere with these frequencies. Information about the interaction between the metal ions and the chelating groups (e.g., POH and PdS) can be obtained from the IR spectra of ISPE-302 and its metal complexes. IR spectra of the ISPE-302 complex with cadmium, lead, and zinc are shown in Figure 7. It appears that the 2300-2360 cm-1 regions of ISPE-302 and of the cadmium, lead, and zinc are different. The absorption in this region is due to the stretching of POH. Replacement of the hydrogen atom with a metal forming a covalent bond will diminish this band. Cadmium and lead complexes show no band in this region, indicating complete replacement of hydrogen atom and strong complexation. However, the zinc complex shows a reduced band in this region, indicating weak interaction with the POH group. This observation

stability

Figure 6. FTIR spectra of (A) neat bis(2,4,4-trimethylpentyl)monothiophosphinic acid (Cyanex-302); (B) functionalized silica; (C) ISPE-302.

Juang and Su, 1992

404 Ind. Eng. Chem. Res., Vol. 36, No. 2, 1997

Ind. Eng. Chem. Res., Vol. 36, No. 2, 1997 405

available for long-term stability with respect to the metal ion uptake capacity of other adsorbents. Despite the loss in capacity of ISPE-302, this system should be economical. After operation for 20 cycles, the functionalized silica can be used for reattachment of Cyanex302. Concentration factors of about 30 were achieved for aqueous streams containing 1.78 mol/dm3 cadmium ions. It is possible to remove zinc, lead, and cadmium simultaneously and separate these metal ions by selective stripping with 0.05 or 0.1 mol/dm3 nitric acid, 0.5 or 1.0 mol/dm3 nitric acid, and 2 mol/dm3 hydrochloric acid, respectively. IR spectral studies indicate the retention of Cyanex-302 on the functionalized silica is due to interactions between the alkyl chain of the extractant and functional groups on the surface. Optimization of the chain length of the functional group and extractants could yield highly stable ISPE-302 due to increased interactions. Additional studies on the determination of distribution coefficients of metal ions, kinetic rates, and distribution of Cyanex-302 between functionalized silica and an aqueous stream are in progress. Further, studies on the optimization of silica functionalization and hydrocarbon chain length of Cyanex-302 for higher stability will be undertaken. These results will be published elsewhere. Figure 7. FTIR spectra of (A) zinc-ISPE-302 complex, (B) leadISPE-302 complex, (C) cadmium-ISPE-302 complex.

supports the findings of the stripping studies which showed zinc could be stripped out with very weak acids (dilute) due to weak complexation. Since no bands could be recorded for PdO and PdS groups, quantitative assessment cannot be made by considering changes in PdO vibrations with respect to interactions of PO‚‚‚ Cd2+ and PO‚‚‚Pb2+. Conclusions This study reports the feasibility of the use of inorganic solid-phase extractants for selective removal of metal ions from dilute streams. ISPE-302 was prepared by attachment of Cyanex-302 on a functional silica support. The silica was functionalized by silanization with dichlorodimethylsilane, yielding a high monolayer coverage density of 4.21 × 10-6 mol/m2. The Cyanex302 was immobilized on functionalized silica to achieve 0.4 × 10-3 mol/g in the final ISPE-302. The dependence of aqueous solution pH on the removal efficiency of zinc, lead, and cadmium shows that, at low pH, ISPE-302 selectively removes cadmium. The sharp breakthrough curves for a residence time of 7 min suggest favorable equilibrium and kinetic rates are achieved at this flow rate and inlet concentrations. An initial capacity of 0.175 × 10-3 mol of cadmium/g of ISPE-302 achieved is 89% of the theoretical capacity. The capacity of ISPE-302 decreases to 0.5 × 10-4 mol/g after 20 cycles of adsorption/stripping, which is likely due to solubilization of Cyanex-302. Table 1 compares ISPE-302 with other adsorbents used for metal ion removal. These comparisons are not strictly identical, but they provide a relative comparison of these adsorbents. However, ISPE-302 can be best compared with impregnated macroporous resins. Impregnated macroporous resins have shown a slow sorption rate (Akita and Taekuchi; 1990; Juang and Su, 1992). On the contrary, this study suggests that ISPE-302 offers favorable equilibria and kinetic rates. Also, the capacity of ISPE-302 is comparable. No quantitative data are

Acknowledgment The financial support of the U.S. EPA through Grant No. R818630-01-1 is gratefully acknowledged. We thank Mr. James Snyder for his assistance during this project. Literature Cited Akita, S.; Taekuchi, H. Sorption and Separation of Metals from Aqueous Solution by a Macromolecular Resin Containing Trin-octylamine. J. Chem. Eng. Jpn. 1990, 23 (4), 439-443. Barrett, D. A.; Brown, V. A.; Shaw, P. N.; Davies, M. C. Characterization of a Range of Alkyl-Bonded Silica HPLC Stationary Phases: Chromatographic Behavior of Neutral, Acidic and Basic Test Solutes. J. Chromatogr. Sci. 1996, 34, 146-156. Bokobza, L.; Cote, G. Spectroscopic Investigation of Adsorbed 7-(4ethyl-1-methyloctyl)-8-Quinolinal and of its Gallium (III) Complex: Comparison with the Behaviours Observed in Solvent Extraction Systems. Polyhedron 1985, 4 (8), 1499-1505. Bradshaw, J. S.; Izatt, R. M.; Bruening, R. L.; Christensen, J. J. Analysis of Ions Present at Low Concentrations in Solutions Containing Other Ions. U.S. Patent 5,175,100, 1992. Braun, T.; Ghersini, G. Extraction Chromatography; Elsevier: Amsterdam, The Netherlands, 1975. Bruening, R. L.; Tarbet, B. J.; Krakowiak, K. E.; Bruening, M. L.; Izatt, R. M.; Bradshaw, J. S. Quantitation of Cation Binding by Silica Gel Bound Thiamacrocycles and the Design of Highly Selective Concentration and Purification Columns for Palladium(11), Gold(III), Silver(I), and Mercury(II). Anal. Chem. 1991, 63, 1014-1017. Chiarizia, R.; Ferraro, J. R.; D’Arcy, K. A.; Horwitz, E. P. Uptake of Metal Ions by A New Chelating Ion-Exchange Resin. VII. Alkaline Earth Cations. Solvent Extr. Ion Exch. 1995, 13, 1063-1082. Colella, M. B.; Siggia, S.; Barnes, R. M. Synthesis and Characterization of a Poly(acrylamidoxime) Metal Chelating Resin. Anal. Chem. 1980, 52, 967-972. Cortina, J. L.; Miralles, N.; Agwilar, M.; Sastre, A. M. Solvent Impregnated Resins Containing Di-(2-ethylhexyl)phosphoric acid. I. Preparation and Study of the Retention and Distribution of the Extractant on the Resin. Solvent Extr. Ion Exch. 1994, 12 (2), 349-369. Deorkar, N. V.; Tavlarides, L. L. Removal of Copper and Nickel from Dilute Aqueous Streams Using Inorganic Chemically Active Adsorbents. In Metals and Materials Waste Reduction, Recovery and Remediation; Liddel, K. C., et al., Eds.; 1994; pp 3-9.

406 Ind. Eng. Chem. Res., Vol. 36, No. 2, 1997 Ferro-Garcia, M. A.; Rivera-Utrilla, J.; Bautista-Toledo, I.; Mingoranne, M. D. Removal of Lead from Water by Activated Carbons. Carbon 1990, 28 (4), 545-552. Jezorek, J. R.; Freiser, H. Metal Ion Chelation Chromatography on Silica-Immobilized 8-hydroxy-quinoline. Anal. Chem. 1979, 51 (3), 366-373. Juang, R.-S.; Su, J.-Y. Separation of Zinc and Copper from Aqueous Sulfate Solutions Using Bis(2-ethylhexyl)phosphoric AcidImpregnated Macroporous Resin. Ind. Eng. Chem. Res. 1992, 31, 2779-2783. Lindoy, L. F.; Eaglen, P. Ion Complexation by Silica-Immobilized Polyethyleneimenes. U.S. Patent 5,190,660, 1993. Marshall, M. A.; Mottola, H. A. Synthesis of Silica Immobilized 8-quinolinol with (Aminophenyl)trimethoxysilane. Anal. Chem. 1983, 55, 2089-2093. Patterson, J. W. Metal Speciation Separation and Recovery; Lewis Publishers, Inc.: Chelsea, MI, 1987. Plueddeman, E. P. Metal Extraction from Solution and Immobilized Chelating Agents used Therefore. U.S. Patent 4, 448,694, 1984. Radha Krishnan E.; Ultrecht, P. W.; Patkar, A. N.; Davis, J. S.; Pour, S. G.; Foerst, M. E. Recovery of Metals from Sludges and Wastewaters; Noyes Data Corp.: Park Ridge, NJ, 1987. Rickelton, W. A.; Boyle, R. J. The Selective Recovery of Zinc With New Thiophosphiric Acids. Solvent Extr. Ion Exch. 1990, 8 (6), 783-797. Socrates, G. Infrared Characteristic Group Frequencies; John Wiley and Sons, Ltd.: New York, 1980. Sood, A.; Fleming, H. L.; Novak, J. W. Process for Removal, Separation and Recovery of Heavy Metal Ions from Solutions Using Activated Alumina Including Acid Treated Activated Alumina. U.S. Patent 4,824,576, 1989. Tavlarides, L. L.; Deorkar, N. V. Ceramic Compositions with a Phospho Acid Moiety. U.S. Patent Application No. 08/313/019, Sept 1994. Tedder, D. W. Separations in Hazardous Waste Management. Sep Purif. Methods 1992, 21 (1), 23-74.

Theis, T. L.; Iyer, R.; Ellis, S. K. Evaluation of New Granular Iron Oxide for Removing of Lead from Drinking Water. J.sAm. Water Works Assoc. 1992, 101-105. Ueda, K.; Yuko, S.; Yoshimura, O.; Yamamoto, Y. Separation and Concentration of lead, Uranium and Copper Using Polystyrene Resins Functionalized with Azobenzylphosphonic Acid Ligands. Analyst 1988, 113, 773-777. Unger, K. K. Packing and Stationary Phases in Chromatographic Techniques; Marcel Dekker, Inc.: New York, 1990. Unger, K. K.; Becker, N.; Roumeliotis, P. Recent Development in the Evaluation of Chemically Bonded Silica Packings for Liquid Chromatography. J. Liq. Chromatogr. 1976, 125, 115-127. Wang, T. C.; Chandra, K. P.; Anderson, P. R. Evaluation of Fe Oxide-Coated GAC for Removal and Recovery of Copper(II) from Water. In Critical Issues in Water and Waste Water Treatment; National Conference on Environmental Engineering, July 1113, 1994; Ryan, J. N., Edward, M., Eds.; American Society of Civil Engineering: New York, 1994; pp 225-233. Warshawsky, A. Extraction with Solvent-Impregnated Resins. In Ion Exchange and Solvent Extraction; Marinksy, J. A., Mascus, Y., Eds.; Marcel Dekker: New York, 1981; Vol. 8, pp 229-331. Warshawsky, A.; Kalir, R.; Bercovitz, H. Solvent Impregnated Resins Incorporating Organophosphoric Acid Compounds. Trans. Inst. Min. Metal. 1979, 88, 31-35. Yebra-Biurrun, M. C.; Bermejo-Barrera, A.; Bermejo-Barrera, M. P. Synthesis and Characterization of a Poly(aminophosphonic acid) Chelating Resin. Anal. Chim. Acta 1992, 264, 53-58.

Received for review July 18, 1996 Revised manuscript received November 12, 1996 Accepted November 14, 1996X IE960415M

X Abstract published in Advance ACS Abstracts, January 1, 1997.