Co-removal of Hexavalent Chromium through Copper Precipitation in

The mechanisms of hexavalent chromium [Cr(VI)] co- removal with copper [Cu(II)] during homogeneous precipitation were studied with batch tests using a...
18 downloads 0 Views 252KB Size
Download PDF
Environ. Sci. Technol. 2003, 37, 4281-4287

Co-removal of Hexavalent Chromium through Copper Precipitation in Synthetic Wastewater JING-MEI SUN,† CHII SHANG, AND JU-CHANG HUANG* Department of Civil Engineering, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong

The mechanisms of hexavalent chromium [Cr(VI)] coremoval with copper [Cu(II)] during homogeneous precipitation were studied with batch tests using a synthetic solution containing Cr(VI) and Cu(II). Metal precipitation was induced by adding Na2CO3 stepwise to different pH, and the respective removals of Cu(II) and Cr(VI) were measured. At the same time, the relative quantities of Cu(II) and Cr(VI) in the precipitates were also analyzed to establish their stoichiometric relationship. The results indicated that, in a solution containing 150 mg/L Cu(II) and 60 mg/L Cr(VI), the initial co-removal of Cr(VI) with Cu(II) began at pH 5.0 and completed at pH 6.2. At pH 5.0-5.2, coprecipitation took place through the formation of copperchromium-bearing solids [such as CuCrO4 and/or CuCrO4‚ 2Cu(OH)2]. Thereafter, the remaining soluble copper started to react with carbonate in a heterogeneous environment to form the negatively charged basic copper carbonate precipitates [CuCO3‚Cu(OH)2], which subsequently adsorbed additional Cr(VI) (or HCrO4-) at pH 5.2-6.2. The maximum Cr(VI) co-removal took place at pH 6.2. Between the two mechanisms, coprecipitation accounted for about 29% of the total chromium’s coremoval while the remaining 71% was attributed to surface adsorption, mainly through electrostatic attraction and ligand exchange. When the solution pH was increased to beyond 7.5, a surface charge reversal took place on the basic copper carbonate solids, and this led to some Cr(VI) desorption. Thus, the extent of Cr(VI) adsorption is highly pH dependent.

Introduction Chemical precipitation (typically involving hydroxide, carbonate, or sulfide precipitation) is commonly employed for heavy metal removal from aqueous solution. It can be achieved by a homogeneous combination of aqueous metal ions with inorganic ligands to form insoluble precipitates under preferred conditions followed by a solid-liquid separation step (1-3). However, application of this technology for heavy metal removal from some industrial waste streams may be limited because it is not effective to remove metals that exist in soluble complex matrixes or in anion forms (3-5). Furthermore, the fine metal precipitates formed during the process settle very slowly, thus requiring a large settling basin, a filtration facility, or the addition of a * Corresponding author e-mail: [email protected]; phone: +(852)2358-7165; fax: +(852)2358-1534. † Present address: Tianjin University, China. 10.1021/es030316h CCC: $25.00 Published on Web 08/12/2003

 2003 American Chemical Society

polyelectrolyte to induce solids aggregation, all of which will increase the treatment cost and/or space. In Hong Kong, space is at a premium, and any treatment facility requiring a large area will not be acceptable. Chemical precipitation of heavy metals can also be achieved by heterogeneous (surface) precipitation (6-10). That is, heavy metals are removed by deposition on the surface of some solid media; as such, the solid-liquid separation step can be easily accomplished by capturing these media. The solids present in heterogeneous precipitation can also lower the activation energy to enable the precipitation to take place at a lower solubility product of metal and ligand. For example, in a heterogeneous environment, precipitation can occur when the product is slightly higher than the theoretical Ksp; while in a homogeneous environment, precipitation can only occur after the product reaches approximately 40-50 times higher than the Ksp (11). Through a series of studies, Huang and co-workers (12-14) developed a single-step and space-saving technology to remove copper, zinc, and nickel from plating wastewater using a fluidized sand column to serve as a heavy metal stripper. In operation, both metal-bearing wastewater and carbonate solution were injected simultaneously to a fluidized sand column to allow soluble metals to achieve heterogeneous precipitation and thus become “coated” (or “plated”) on the sand surface. This is called “nucleated precipitation” or sometimes referred to as “pellet crystallization”. At pH 9.2, more than 90% of soluble copper, zinc, and nickel were removed from wastewater when the initial concentration of each metal was not in excess of 225 mg/L. Quite unexpectedly, a large percentage (50-80%) of aqueous Cr(VI) was also coremoved during a pilot trial of this new technology in treating actual plating wastewater without a prior reduction of Cr(VI) to Cr(III). It is well-known that Cr(VI) does not react with either carbonate or hydroxide ions to form any precipitate unless hexavalent chromium is first reduced to the trivalent form. However, Cr(VI) has been known to react with other metallic ions in aqueous solutions to form insoluble PbCrO4, AgCrO4, and BaCrO4 precipitates (15). The exact mechanisms contributing to the co-removal of Cr(VI) with Cu(II), Zn(II), and Ni(II) have not been reported in the literature. In this study, a series of properly designed experimental tests were conducted to investigate the specific mechanisms that are responsible for the Cr(VI) co-removal with copper at different operating pH. Also, the predominant form of the Cu-Cr precipitates was investigated from the experimental data as well as the SEM and EDAX microscopic characterizations. The co-removal of Cr(VI) with zinc and nickel is not included in this paper.

Experimental Section Synthetic Metal Solutions. All metal solutions used in this study were prepared from reagent-grade chemicals: CuCl2‚ 2H2O from Riedel-Dehae¨n and K2Cr2O7 from Nacalai Tesque. Stock solutions of Cu(II) and Cr(VI), each 4000 mg/L, were prepared by dissolving the metal reagents with doubledistilled, deionized water. In this preparation, HNO3 was added to lower the solution pH to less than 2.0 to prevent metal precipitation during storage in polypropylene bottles. Then the stock solutions were diluted to target aqueousphase concentrations with double-distilled, deionized water before each test, with its pH adjusted to a proper value. The standard metal solutions for calibration purpose were also prepared by diluting commercial standard solutions (Fluka Chemie AG CH-9470 Bchs, 1.000 g/L ( 0.3%, 20 °C). VOL. 37, NO. 18, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4281

FIGURE 1. Soluble metal [Cr(VI) and Cu(II)] concentrations at equilibrium as a function of pH measured in (a) Cu-Cr co-removal test (solid symbol) and (b) pure copper precipitation test (open symbol). Analytical Methods. pH was measured using an Orion pH meter (model 420-A). Daily calibration with proper buffer solutions (pH 4.01, 7.00, and 10.01) was performed to ensure its accuracy. Soluble metal concentrations were determined by the Standard Method 3500 (16) with the injection of sample filtrates to an atomic absorption spectrophotometer (HITACHI Z-8200 Polarized Zeeman). The sulfate ion (SO42-) concentration was determined by ion chromatography (DIONEX DX-500) based on the Standard Method 4500 (16). The morphology and compositions of the metal precipitates were examined spectroscopically using both scanning electron microscope (SEM) (PHILIPS, JSM-6300F) and energydispersive analysis of X-ray (EDAX) (PHILIPS, XL30). Experimental Procedures. Unless otherwise specified, all batch experiments were carried out in 500-mL well-mixed reactors. The co-removal of chromium with copper was mainly evaluated with a synthetic metal solution containing 150 mg/L Cu(II) and 60 mg/L Cr(VI) with its initial pH adjusted to approximately 2.0. These two metal levels were employed because they were typical as observed in our previous field study (13, 14). Although no data are reported in this paper, a few tests were conducted using Cu(II) as high as 15 000 mg/L and Cr(VI) as high as 6000 mg/L. During each test, the solution was continuously stirred while sodium carbonate (0.5 N) was added dropwise to the solution using a titration buret to progressively increase pH to induce metal precipitation. In each pH increment, the solution was mixed for 30 min to allow the reaction to reach equilibrium. Then a 10-mL aliquot was withdrawn and passed through a 0.45µm membrane filter. The filtrate was acidified with concentrated HNO3 and stored in a polypropylene bottle for subsequent metal analyses. The test solution pH was then increased stepwise to 10.0. At each increased pH, the filter paper together with the collected precipitates was immersed in a 10-mL, 0.3% HNO3 solution (pH 1.51) to redissolve metals. Again, the solution was then stored in a polypropylene bottle for later metal analyses. The filter paper was found to contain no copper or chromium in a blanket test. Some of the precipitate samples on the filter paper were air-dried and collected for subsequent spectroscopy characterizations with both SEM and EDAX. For chromium adsorption tests, the adsorbent (i.e., basic copper carbonate precipitates) was obtained by dosing Na2CO3 (0.5 N) to a 150 mg/L pure copper [Cu(II)] solution to induce complete copper precipitation at pH 7.0. The precipitates were then settled and washed 3 times with double-distilled, deionized water with its pH adjusted to 7.0. The adsorption testes were conducted with 60 mg/L Cr(VI) 4282

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 18, 2003

solution in two different manners. In test I, all of the copper precipitates (obtained from precipitating 1 L of 150 mg/L copper) were added to 1 L of Cr(VI) solution (with its initial pH adjusted to 6.0) to allow the initial adsorption to take place. Upon reaching equilibrium, a 10-mL test mixture was withdrawn to measure its remaining soluble meals. Then, the pH of the remaining test mixture was increased stepwise to higher levels by dosing additional 0.5 N Na2CO3. The procedures were repeated until the solution pH reached 10.0. The samples collected at different pH were subjected to soluble Cr(VI) determinations. In test II, it was conducted by adding the basic copper carbonate precipitates obtained from precipitating 100 mL of 150 mg/L Cu solution to a series of 100-mL chromate solutions [containing 60 mg/L Cr(VI)], each with a different initial pH (from 6.0 to 10.0). Thus, the concentration of copper carbonate precipitates added to each chromate solution was the same as that used in test I. Again, after equilibrium adsorption, the soluble Cr(VI) concentration remaining in each test solution was determined. Besides, a representative precipitate sample was obtained from each test pH for spectroscopic characterizations.

Results and Discussion For the pH normally encountered in dilute plating wastewater, Cr(VI) mainly exists in soluble anions as HCrO4- or CrO42- (depending on solution pH), neither of which can react with CO32- or OH- to form precipitates unless Cr(VI) is first reduced to Cr3+. Thus, at the beginning of this study, it was first postulated that the co-removal of Cr(VI) during copper precipitation could be attributed to either adsorption or coprecipitation or a combination of both. To avoid any confusion, it is necessary to first define some terms being used here. The term “adsorption” is used to describe the circumstance wherein Cr(VI) is adsorbed onto the surface of some copper-bearing precipitates, while “coprecipitation” refers to the condition that involves a direct Cu-Cr interaction, leading to a simultaneous binding of the two metals through the formation of copper-chromium-bearing precipitates. The term “co-removal” refers to the total Cr(VI) co-removal by both coprecipitation and adsorption. Co-removal of Chromium with Copper. Figure 1 shows the quantities and percentage of Cu(II) and Cr(VI) removed at different pH both in a mixed Cu-Cr solution (solid symbol) and in a pure Cu solution (open symbol). Before the actual mechanism for the Cr(VI) removal is further discussed, the term of co-removal is used here first. For the mixed Cu-Cr solution, at pH below 5.0, essentially no Cr(VI) or Cu(II) removal was observed. However, with progressive pH incre-

FIGURE 2. [Cr]/[Cu] molar ratios in the Cu-Cr precipitates as a function of pH. ments from pH 5.0 to pH 6.2, both Cu(II) and Cr(VI) were phenomenally transformed into some insoluble forms and became removed from solution. Also, the co-removal of Cr(VI) appeared to synchronize with Cu(II) precipitation. The maximum Cr(VI) co-removal was 36 mg/L (reduced from 60 to 24 mg/L), which occurred at pH 6.2-6.8. This was accompanied by more than 96% of Cu(II) removal. However, at pH higher than 6.8, the removal of Cu(II) continued to increase slightly until it reached almost 100% at pH 7.8. On the other hand, Cr(VI) co-removal was found to decrease when pH was raised to beyond 6.8. A total of 6 mg/L Cr(VI) was released back to solution when pH was increased from 6.8 to 10.0. This means that 16% of the total co-removal (6 out of 36 mg/L total Cr removal) could be desorbed at a higher pH. These data suggest that the co-removal of Cr(VI) with copper involves multiple mechanisms, one being responsible for inducing the “simultaneous removal” of both chromium and copper at pH 5.0-5.2, and the other responsible for the additional co-removal of Cr(VI) at pH from 5.2 to 6.2. A portion of the latter could become desorbed when the pH was elevated to beyond 6.8. It was first postulated that the simultaneous co-removal of Cr(VI) with Cu(II) at pH 5.0-5.2 involved a direct Cu-Cr interaction leading to the formation of CuCrO4 precipitates (i.e., a phenomenon of coprecipitation). To verify this postulation, the MINEQL+ software (17) was used for chemical equilibrium calculations. In the program, a Ksp of 3.63 × 10-6 is indicated for CuCrO4. On the basis of this Ksp, however, this is no CuCrO4 precipitation shown in the output under the test condition. It is suspected that the actual Ksp for CuCrO4 may be somewhat smaller than the one indicated on the MINEQL+ program. Unfortunately, in the literature there is no other Ksp value reported for CuCrO4. As will be explained later in this paper, upon characterization of the initial precipitates collected at pH 5.0-5.2, the molar stoichiometric ratio of Cu to Cr was 0.74 under the test condition (Figure 2). In an additional test, both Cu(II) and Cr(VI) were increased by 100 times, and it was observed that the initial precipitation started approximately at pH 4.0. The collected precipitates were found to have a [Cr]/[Cu] molar ratio of close to 0.9. Thus, it is reasonable to assert that the initial simultaneous removal of Cu and Cr at pH 5.0-5.2 was indeed due to a direct reaction of these two metals leading to the formation of CuCrO4 precipitates. When the pH was increased to 5.2-6.2, copper precipitation continued to increase at a fast pace while chromium’s co-removal tapered off considerably, indicating that there was no direct coprecipitation. It was postulated that, as pH was further increased to beyond 5.2, additional precipitation of Cu(II) could take place by reacting with carbonate and hydroxide ions under

a heterogeneous environment because of the presence of the initially produced CuCrO4 solids. The newly formed copper-bearing precipitates had a tendency to adsorb some of the remaining soluble Cr(VI) through electrostatic attraction and ligand exchange. To verify the above postulation, a second experiment was conducted using the same procedures but with only a pure copper solution, also containing 150 mg/L copper without chromium. The soluble copper concentrations remaining at different pH are also presented in Figure 1 (open symbol). In this case, basic copper carbonate precipitates, mainly consisting of CuCO3‚Cu(OH)2 [Ksp ) 10-33.31 (17)] as previously reported by us (14), were formed as soon as the solution pH approached 6.0, and this was one pH unit higher than the starting precipitation pH observed for the mixed Cu-Cr solution. It is then clear that, in the absence of Cr(VI), copper precipitation cannot take place until the solution pH reaches 6.0. From MINEQL+, the ion products at pH 6.0 under the test condition are 10 times higher than the Ksp of the basic copper carbonate but only slightly higher than that of copper hydroxide [Ksp ) 10-19.33 (17)]. Thus, the further precipitation of copper at pH beyond 5.2 was mainly attributed to the formation of the basic copper carbonate. Pure copper carbonate precipitates [Ksp ) 10-11.5 (17)] cannot be formed under this condition. The above finding suggests that the initial copper precipitation which took place at pH 5.0-5.2 in the mixed Cu-Cr solution was not attributed to the formation of basic copper carbonates. Instead, it was a result of direct coprecipitation with the formation of insoluble CuCrO4 precipitate. Adsorption of Cr(VI) at this initial pH was unlikely since very little basic copper carbonate precipitates were produced in this pH region. For this reason, it is reasonable to conclude that, at pH 5.0-5.2, coprecipitation is the main mechanism responsible for the co-removal of Cr(VI). It was also distinctly observed that once CuCrO4 precipitates were formed at pH 5.0-5.2, the test solution turned into a heterogeneous environment, and the produced CuCrO4 crystallites were able to catalyze an early formation of the basic copper carbonate at a pH much lower than 6.0. This was because the heterogeneous environment was able to lower the solubility product needed for initiating the first basic copper carbonate precipitation. This argument is also supported in the literature (11) that the required product of metal and ligand concentrations can be significantly reduced in initiating the first metal precipitation in a heterogeneous environment. The suggested heterogeneous precipitation of basic copper carbonate starting at pH slightly higher than 5.2 was also confirmed with MINEQL+. That is, under the experimental conditions, the solid phase of the basic copper carbonate could exist at pH 5.5 at equilibrium. When this occurred, adsorption of some remaining Cr(VI) by the newly formed basic copper carbonate precipitates could then take place since the produced solids were positively charged at this pH region (will be discussed later) while Cr(VI) existed in anionic HCrO4- and CrO42- forms. To evaluate the chemical forms of the produced precipitates and also to confirm the above postulated mechanisms for the Cr(VI) co-removal, the precipitate solids collected at different pH were washed, redissolved, and then analyzed for the Cu and Cr concentrations. Figure 2 shows the [Cr]/[Cu] molar ratios of the precipitates obtained at different pH. If the Cr(VI) co-removal was indeed only caused by the formation of CuCrO4, the [Cr]/[Cu] molar ratio would be equal to 1.0. However, the [Cr]/[Cu] molar ratios shown in Figure 2 were found to vary considerably with pH. At pH 5.0, wherein the initial precipitation took place, a molar ratio of 0.74 was observed. As pH was increased to 5.2, the ratio rapidly decreased to 0.35. This finding seems to confirm our previous postulation that, immediately after the first production of VOL. 37, NO. 18, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4283

soluble bichromate and chromate ions were readily adsorbed by the positively charged copper carbonates through electrostatic attraction. Due to adsorption of chromium, the [Cr]/[Cu] molar ratio in the precipitates did not show a substantial further reduction when pH was raised to 7.0, which was the final pH for completing copper precipitation. If the pH was elevated to beyond 7.5, a charge reversal took place on copper carbonate precipitates, and this led to some Cr(VI) desorption, which resulted in a corresponding decrease in the [Cr]/[Cu] ratio (Figure 2). These arguments are also consistent with the data of Figure 1, which shows some Cr(VI) desorption at an elevated pH. Adsorption of Cr(VI) onto Basic Copper Carbonate Precipitates. To evaluate the significance of adsorption in contributing to the total Cr(VI) co-removal and also to verify the effect of pH on such phenomena, adsorption experiments were conducted by using basic copper carbonate precipitates as the adsorbent and aqueous Cr(VI) as the adsorbate at pH ranging from 6.0 to 10.0. As described in the Experimental Procedures, the adsorption tests were conducted in two different manners. In the first test, the initial adsorption was allowed to take place at a fixed pH of 6.1; thereafter, the pH of the test solution was progressively increased to observe Cr(VI) desorption. In the second test, the initial adsorption was allowed to take place at different pH, which yielded different speciation, or [HCrO4-]/[CrO42-] ratios, in the test solution. In both tests, the extents of soluble chromium removal at different pH were monitored and recorded in Figure 3. It is worth pointing out that a blank test using only double-distilled, deionized water without Cr(VI) was also conducted to ensure that there was no copper dissolution at the tested pH range of 6.0-10.0. Since no copper dissolution was observed in the blank test, it could be concluded that any change in Cr(VI) concentrations in the test solution was mainly attributed to the adsorption and desorption reactions. From the data of test I (Figure 3), about 25 mg/L Cr(VI) was initially adsorbed at pH 6.1. The subsequent progressive increase of pH to 10.0 caused only a small extent of chromium desorption. At pH 10.0, the amount of chromium removal was reduced to 18 mg/L. This corresponded to a 28% [i.e., (25-18)/25 ) 0.28] desorption. It is interesting to note that the maximum Cr(VI) co-removal reported in Figure 1 was 35 mg/L, which occurred at pH 6.2. Out of this, 25 mg/L was attributed to surface adsorption (as shown in Figure 3). Thus, the mechanism of adsorption actually accounted for 71% (i.e., 25/35 ) 0.71) of the total chromium removal. The remaining 29% of the co-removal was attributed to coprecipitation. However, these calculations may be arguable since it is based on the assumptions that the adsorption process is not involved in the coprecipitation stage and also that the total Cr(VI) co-removal by both coprecipitation and adsorption is additive. The adsorption mechanism could be further elucidated by the experimental data from test II (Figure 3) wherein the initial chromium adsorption was allowed to take place at

FIGURE 3. Cr(VI) removal through adsorption onto basic copper carbonate precipitates as a function of pH with initial [Cr(VI)] ) 60 mg/L and [basic copper carbonate] ) 150 mg/L as Cu. Test I: Adsorption at pH 6.1 followed by desorption with pH increase (solid symbol). Test II: Adsorption at different pH (open symbol). CuCrO4 solids at pH 5.0, the test solution became a heterogeneous system which then allowed some Cu(II) to react with carbonate and hydroxide ions to form the basic copper carbonate precipitates. The formation of the latter would of course lower the [Cr]/[Cu] molar ratio. Since in the experiment a molar ratio of unity was never observed, it is believed that once the initial CuCrO4 crystallites are formed, even in only a small quantity, these finely divided precipitates are able to catalyze an immediate formation of a mixture of some other complex species, such as CuCO3‚Cu(OH)2 and Cu(OH)2, and this mixture contains no chromium. The production of the latter precipitates would of course greatly reduce the [Cr]/[Cu] molar ratio, as reflected in Figure 2. In a separate test, we intentionally raised the Cu(II) level to 15 000 mg/L and raised the Cr(VI) level to 6000 mg/L in order to induce the first solids production at a lower pH of approximately 4.0. The collected solids from this test at pH 4.0 showed a [Cr]/[Cu] molar ratio of 0.9. This further verifies that the first production of precipitated solids is indeed CuCrO4. Because once such type of precipitates was produced, the test solution became heterogeneous, which then triggered subsequent formation of other types of precipitates. Therefore, it is difficult to establish a precise stoichiometry for the Cu-Cr-precipitated solids because the test system involved the production of multiple precipitated species within a very narrow pH range. Through ζ-potential analysis (data not shown), the basic copper carbonate precipitates were found to carry positive surface charges at pH below 7.5. As a comparison, the Cu-Cr precipitates were also found to carry some positive surface charges but in a much smaller quantity at pH below 6.0. However, when pH was further increased, a charge reversal took place. Thus, it is reasonable to conclude that, upon the immediate production of the basic copper carbonates at pH 5.2-6.0, a considerable fraction of the remaining

TABLE 1. Equilibrium Concentrations of Sulfate and Chromium Adsorbed onto Basic Copper Carbonate Precipitatesa sulfate

aq adsorbate species (%) total removal (mequiv/g) test A test B test C average a

chromium

pH 6.5

pH 10.0

2-

2-

SO4

1.68 1.58 1.68 1.65

(100%)

SO4

1.02 0.71 1.11 0.95

(100%)

pH 6.5 2-(50%),

CrO4 4.70 4.65 4.80 4.72

Conditions: [adsorbent] ) 0.3 g/L and [adsorbate] ) 2.0 mequiv/L at pH 6.5 and pH 10.0.

4284

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 18, 2003

pH 10.0

HCrO4

-(50%)

CrO42- (100%) 0.51 0.10 0.13 0.25

different pH. The maximum adsorption of Cr(VI) was maintained at pH 6.0-7.3 with a similar quantity as that observed in test I. However, the extent of the initial adsorption decreased sharply to less than 5 mg/L when the initial adsorption pH was increased to 8.7 or higher. The results suggest that adsorption of Cr(VI) is a complex phenomenon and it is highly pH-dependent. Most likely it is a combination of electrostatic attraction and ligand exchange depending on the surface complexes formed at different pH. The prevailing mechanism of adsorption is pH-dependent. From the perspective of surface chemistry in aqueous phase, the surfaces of metal precipitates (both in the forms of hydroxide and carbonate) are generally covered with hydroxyl groups but vary in forms at different pH (18). Although no data are shown here, it was found in this study that the ζ-potential of the basic copper carbonate precipitates varied strongly with pH, ranging from +80 mV at pH 6.0 to -20 mV at pH 10.0. Thus, the adsorption of Cr(VI) through electrostatic attraction will be strongly pH-dependent, being more significant at a lower pH when the precipitates are positively charged. The data obtained from test II have already confirmed this trend. At the pH range of 6.0-7.5, the precipitates are positively charged, thereby resulting in a high adsorption of HCrO4- and CrO42-. When pH is raised to beyond 7.5, a charge reversal takes place. Thus, the electrostatic adsorption becomes very limited or even impossible to occur because both the adsorbent and the adsorbate carry the same negative charges. Therefore, at pH higher than 7.5, Cr(VI) adsorption is most likely caused by ligand exchange. In this case, the hydroxylated oxide particles [CuCO3‚Cu(OH)2 precipitates] can be considered as a polymeric oxo acid (or base) and the replacement of its coordinative ligands (OH2+, OH, or O-) by other ions (HCrO4- or CrO42-) can be interpreted as ligand exchange at the oxidewater interface (18). As such, although the adsorbent is negatively charged, a small quantities of Cr(VI) adsorption can still occur at the alkaline pH range. The speciation of adsorbate (HCrO4- or CrO42-) is also considered important in affecting the adsorption capacity. No matter whether adsorption is attributable to electrostatic attraction or ligand exchange, HCrO4- is expected to be the preferential species to be adsorbed, provided that its adsorption affinity onto the basic copper carbonate precipitates is the same as CrO42-. To verify this hypothesis, a similar test using a sulfate solution (Na2SO4 from RiedelDehae¨n), in lieu of the bichromate plus chromate solution, was conducted. All experimental conditions were identical to the second Cr(VI) adsorption test as described earlier except that only two pH levels (6.5 and 10.0) were used. These two pH values were selected based on the characteristic Cr(VI) species present in the test solution. At pH 6.5, HCrO4- and CrO42- are present at equal molar quantities, while at pH 10.0, CrO42- is almost the exclusive species. On the other hand, sulfate is present in the form of SO42- at both pH levels. Although the adsorption capacity of sulfate onto the basic copper carbonate precipitates is different from that of chromate due to their different adsorption affinities, the qualitative trends of adsorbing divalent anions onto the basic copper carbonate precipitates at pH 6.5 and pH 10.0 can be well illustrated in this simplified case. Again, the aqueous sulfate concentrations in solution prior to and after adsorption were measured, and the difference was interpreted as the corresponding concentrations of sulfate adsorbed onto the basic copper carbonate precipitates. The experiments were repeated three times, and the results are shown in Table 1, accompanied with the corresponding results of chromium adsorption (also in triplicate) at the same pH conditions. At pH 6.5, the copper precipitates were indeed capable of adsorbing substantial quantities of divalent SO42- ions, which could not be coprecipitated by forming copper sulfate

FIGURE 4. Representative EDAX spectra of (a) basic copper carbonate precipitates (pH 7.0), (b) precipitates from the co-removal test (pH 10.0), and (c) precipitates from the adsorption test I (pH 10.0). precipitates at the experimental conditions. The sulfate adsorbed by the copper precipitates at pH 10.0 was relatively limited, which was also smaller as compared to that at pH 6.5. This again demonstrates that the surface charge of the basic copper carbonate precipitates significantly affected sulfate adsorption. Besides, at pH 6.5 the adsorption capacity of divalent SO42- ions was still much less than that of the chromate adsorption in a test solution containing equal amounts of HCrO4- and CrO42-. On the contrary, at pH 10.0, the SO42- adsorption was considerably larger than that of CrO42-. This finding confirms the previous argument that chromium speciation is important in affecting the Cr(VI) VOL. 37, NO. 18, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4285

Combining the results from the first and the second Cr(VI) adsorption tests (Figure 3), it is known that the adsorption and desorption equilibria in tests I and II did not yield the same results in the basic pH range. In general, Cr(VI) adsorption was much less in test II than that in test I at any given basic pH level. This suggests that, once Cr(VI) is adsorbed onto the basic copper carbonate precipitates, it becomes difficult to desorb even when the solution pH is subsequently increased to a point at which a reversal of the surface charge occurs. The data of test I show that only 28% of the adsorbed Cr(VI) was desorbed. On the contrary, when the initial adsorption took place at an alkaline pH, the capacity of Cr(VI) adsorption was drastically reduced, mainly because chromate existed as a divalent anion. Thus, the initial chromium speciation has a profound impact on the capacity of Cr(VI) adsorption. To further shed some light on the incorporation of Cr(VI) during Cu(II) precipitation, both EDAX and SEM analyses were made on the representative precipitates obtained from the co-removal (Figure 2) and the first adsorption (Figure 3) tests, as well as from the precipitation of pure copper solution which yielded the basic copper carbonate precipitates. The results are shown in Figures 4 and 5. From the EDAX analysis, it is clear that, for a given amount of copper content, the precipitates from the co-removal test display a higher level of Cr(VI) as compared to that from the adsorption test. There is no Cr(VI) peak observed for the solids of basic copper carbonate precipitates. This finding confirms that, other than adsorption, coprecipitation also accounts for some coremoval of Cr(VI). Furthermore, the photomicrographs of SEM analysis show that the lattice structures of the precipitates from the coremoval test (including both coprecipitation and adsorption) and the first adsorption test (including only adsorption) appear to be similar but distinctly different from that of the basic copper carbonate precipitates. This finding suggests that adsorption of chromium can significantly alter the original copper carbonate lattice structure and that adsorption seems to be the main mechanism accounting for the Cr(VI) co-removal. All of the above discussions seem to suggest that the main mechanism for Cr(VI) co-removal during copper precipitation is through adsorption by the basic copper carbonate precipitates, both during and after their formation. Nevertheless, coprecipitation with copper in producing CuCrO4 is also crucial in the initial co-removal, which occurs at pH 5.0-5.2. With respect to adsorption, both electrostatic attraction and ligand exchange are responsible, with the former playing a greater role at pH below 7.5 because the adsorbent carries positive surface charges.

Acknowledgments This study was supported in part by the Hong Kong Research Grants Council under two contracts (HKUST537/94E and HKUST6034/01E).

FIGURE 5. Representative SEM photo images (×20 000) of (a) basic copper carbonate precipitates (pH 7.0), (b) precipitates from the co-removal test (pH 10.0), and (c) precipitates from the adsorption test I (pH 10.0). adsorption, and HCrO4- is adsorbed more than CrO42- by the basic copper carbonate precipitates. It has been reported in the literature that the adsorption free energy of HCrO4- is from -0.52 to -2.26 kcal/mol while that for CrO42is from -0.34 to -2.13 kcal/mol at a surface loading from 10-5 to 7.7 × 10-5 mol of Cr(VI)/g of TiO2. Thermodynamically, HCrO4- is more favorably adsorbed onto TiO2 than CrO42(19). 4286

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 18, 2003

Literature Cited (1) Kim, B. M. AIChE Symp. Ser. 1980, No. 77, 39. (2) Patterson, J. W.; Allen, H. E.; Scala, J. J. J. Water Pollut. Control Fed. 1977, 49, 2397. (3) Meanally, S.; Benefield, L.; Reed, R. B. Sep. Sci. Technol. 1984, 19, 191. (4) McFadden, F.; Benefield, L.; Reed, R. B. Proceedings of the 39th Industrial Waste Conference; Purdue University: West Lafayette, IN, 1985; p 417. (5) Dean, J. A. Lange’s Handbook of Chemistry, 11th ed.; McGrawHill: New York, 1992. (6) Bailey, R. P.; Bennett, T.; Benjamin, M. M. Water Sci. Technol. 1992, 26, 1239. (7) Benjamin, M. M.; Sletten, R. S.; Bailey, R. P.; Bennett, T. Water Res. 1996, 30, 2609.

(8) Nielsen, P. B.; Christensen, T. C.; Vendrup, M. Water Sci. Technol. 1997, 36, 391. (9) Aktor, H. Water Sci. Technol. 1994, 30, 31. (10) Seckler, M. M.; Bruinsma, O. S. L.; Rosmalen, G. M. Water Res. 1996, 30, 1677. (11) Snoeyink, V. L.; Jenkins, D. Water Chemistry; Wiley: New York, 1980. (12) Huang, J. C.; Li, W. F. Proceedings of the 1997 Asian Industrial Technology Congress, Hong Kong, 1997; p 98. (13) Zhou, P.; Huang, J. C.; Li, W. F.; Wai, K. M. Water Res. 1999, 33, 1918. (14) Sun, J.; Huang, J. C. Water Sci. Technol. 2002, 46, 413. (15) Bjerrum, J.; Schwarzenbach, G.; Sillen, L. G. Stability Constants of Metal-Ion Complexes, with Solubility Products of Inorganic Substances, Part II: Inorganic Ligands with Solubility Products of Inorganic Substances; Royal Chemical Society: London, 1964.

(16) Standard Methods for the Examination of Water and Wastewater, 19th ed.; APHA-AWWA-WEF: Washington, DC, 1995. (17) Schecher, W. D.; McAoy, D. C. MINEQL+, A Chemical Equilibrium Modeling System, Version 4.5; Environmental Research Software: Hallowell, ME, 2003. (18) Stumm, W.; Morgan, J. J. Aquatic Chemistry, 3rd ed.; Wiley: New York, 1996. (19) Weng, C. H.; Wang, J. H.; Huang, C. P. Water Sci. Technol. 1997, 35, 55.

Received for review January 15, 2003. Revised manuscript received June 5, 2003. Accepted June 18, 2003. ES030316H

VOL. 37, NO. 18, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

4287