- Email: [email protected]
Removal of free and chelated Cu(II) ions from water by a nondispersive solvent extraction process
Water Research 36 (2002) 3611–3619
Removal of free and chelated Cu(II) ions from water by a nondispersive solvent extraction process Su-Hsia Lin, Ruey-Shin Juang* Department of Chemical Engineering, Yuan Ze University, Chung-Li 320, Taiwan Received 31 May 2001; received in revised form 18 January 2002; accepted 10 February 2002
Abstract The possibility of extraction of free and the ethylenediaminetetraacetic acid (EDTA)-chelated Cu(II) ions from water through a microporous hollow fiber to an organic phase containing extractants LIX64N and Aliquat 336 (a quaternary amine) was studied, in which Cu(II) was simultaneously back-extracted across another hollow fiber to a stripping solutions containing HCl. Experiments were carried out at different flow rates of feed (1.4–4.3 cm3/s), organic (1.2– 4.1 cm3/s), and stripping phases (1.4–4.3 cm3/s), and temperatures (298–318 K). It was shown that the extraction rate increased with decreasing feed phase flow rate, but the effect of flow rates of organic and stripping phases was not pronounced under the ranges studied. Mechanism of the mass transfer at each step was discussed and the effect of mutual interaction of the two extractants in the organic phase was also examined. The application potentials of this novel extraction process for this subject appeared to be technically promising. r 2002 Elsevier Science Ltd. All rights reserved. Keywords: Nondispersive extraction; Microporous hollow fibers; Cu2+; EDTA-chelated anions; LIX64N; Aliquat 336
1. Introduction Solvent (liquid–liquid) extraction has been widely used to remove efficiently toxic heavy metals from industrial effluents to environmentally acceptable levels or to recycle those metals back to the original process [1]. In the past two decades, nondispersive solvent extraction based on the microporous membranes has been developed to avoid many of the problems associated with conventional solvent extraction including solvent loss, phase separation, emulsion formation, and the shortcomings due to flooding and loading in conventional column contactors (e.g., packed towers) and mixer-settlers [2–7]. Although such nondispersive process may result in a slower rate of mass transfer compared to traditional operation if membrane resistance is significant, it can compromise possible advantages such as mass transfer rate offered by higher surface *Corresponding author. Fax: +886-3-4559373. E-mail address: [email protected] (R.-S. Juang).
area/volume associated with hollow fiber modules [2,3,8]. Besides this, the hollow fiber membrane contactors present several advantages such as a larger contact area without direct mixing of the aqueous and organic phases, the lower solvent holdup, and reduction in equipment space and volume. Strong chelating agents such as ethylenediaminetetraacetic acid (EDTA) and nitrilotriacetic acid (NTA) are often present in process and waste streams [9]. Effluents discharged from metal finishing industries and the manufacturing process of printed-circuit boards for electroless copper plating are the examples [9,10]. This is also encountered in washing solutions for remediation of metal-contaminated soils [11,12] and chemical cleaning solutions for power plant boilers [13,14]. The streams containing chelating agents may make chemical precipitation methods such as sulfide and hydroxide for heavy metals inefficient [15]. Furthermore, the high buffer capacity provided by these chelating agents requires excessive amounts of chemicals to neutralize alkalinity.
0043-1354/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 0 4 3 - 1 3 5 4 ( 0 2 ) 0 0 0 7 4 - X
S.-H. Lin, R.-S. Juang / Water Research 36 (2002) 3611–3619
3612
Nomenclature CuL2 DCu H4L HX
EDTA-chelated anions extraction distribution ratio of Cu(II) defined in Eq. (1) EDTA (ethylenediaminetetraacetic acid) LIX65N
In this work, the aqueous solutions containing Cu2+ and fewer amounts of EDTA were focused; i.e., they contained both Cu2+ cations and the EDTA-chelated anions. Such types of solutions make the extraction of both species difficult by a single extractant. In general, the metal cations can be recovered by acidic extractants such as LIX-series reagents and organophosphorus acids, whereas the metallic anions such as CrO2 and 4 EDTA-chelated anions by tertiary and quaternary amines [16,1]. An attempt was made here to evaluate the possibility of using the mixture of extractants LIX64N and Aliquat 336 for this purpose in two hollow fiber extractors. It was expected that this process could simultaneously extract free and chelated Cu(II) ions from aqueous feed solution to an organic solution, and then back-extract to a stripping solution. Process variables including the flow rates of feed (1.4–4.3 cm3/ s), organic (1.2–4.1 cm3/s), and stripping phases (1.4– 4.3 cm3/s), and temperature (298–318 K) were examined. Kinetics of individual extraction system and the effect of mutual interaction of the two extractants in the organic phase were discussed.
NR4Cl Q T
Aliquat 336 flow rate (cm3/s) temperature (K)
Subscripts f, s feed and stripping phases, respectively. 0 initial
Fig. 1. Schematic representation of experimental setup for nondispersive extraction in hollow fiber contactors. Table 1 Characteristics of the hollow fiber module used for extraction experiments 1. Description
2. Materials and methods 2.1. Apparatus, hollow fibers, and solutions Fig. 1 shows the experimental setup for extraction and back-extraction processes. The microporous hollow fiber module subject to analysis was the Hoechst Celanese Liqui-Celt Extra-Flow 2.5 8 Membrane Contactor (Model G-261). These modules are small laboratory-scale versions (with three 0–1 dm3/min pumps and flow meters) that are specifically designed for experimental purposes. Their small size makes it possible to evaluate the performance of the technology without having to prepare large amounts of the aqueous and organic phases. Additional information about this module is listed in Table 1. LIX64N was supplied from Henkel Co (now, Cognis Co.), which is a mixture of 99 vol% LIX65N (2hydroxy-5-nonyl-benzophenone oxime) and 1 vol% LIX63 (5,8-diethyl-7-hydroxy-6-dodecanone oxime). The former is very effective for Cu2+ extraction and the latter act as a kinetic accelerator [1]. Aliquat 336
2. Shell characteristics Material Length Inner diameter Outer diameter 3. Fiber characteristics Material Number of fibers Effective length Inner diameter Outer diameter Effective surface area Effective area/volume Average pore size Membrane porosity Membrane tortuosity
Hoechst Celanese LiquiCelt Extra-Flow 2.5 8 Membrane Contactor Model G-261 Polypropylene 20.3 cm 6.3 cm 7.7 cm Celgard X-30 240 polypropylene hollow fibers B10,200 19.8 cm 240 mm 300 mm 1.4 m2 29.3 cm2/cm3 0.03 mm 0.4 2.6
(tri-n-octylmethylammonium chloride) was offered from Aldrich Co. It was washed with dilute HCl solution to remove water-soluble impurities before use. The solvent
S.-H. Lin, R.-S. Juang / Water Research 36 (2002) 3611–3619
kerosene, supplied from Union Chem. Works Ltd., Taiwan, was washed twice with 20 vol% H2SO4 to remove aromatics and then with deionized water (Millipore Milli-Q) three times. Isodecanol, EDTA, CuCl2, and other inorganic chemicals were supplied by Merck Co. as analytical reagent grade. The organic phase was prepared by diluting LIX64N, Aliquat 336, and isodecanol (with equivolume of Aliquat 336) in kerosene, and the initial concentration of extractants ranged from 50 to 300 mol/m3. Isodecanol was used as modifier to avoid the formation of a third phase. The feed phase was prepared by dissolving CuCl2 and fewer amounts of EDTA in deionized water, in which the pH was adjusted by adding a small amount of 0.1 mol/dm3 HCl or NaOH. The total concentration of Cu(II) was fixed at 5 mol/m3. The stripping phase contained 4 mol/dm3 HCl. 2.2. Experimental procedures 2.2.1. Batch extraction In some cases, the extraction distribution ratios of Cu(II) were measured in batch mode. Equal volumes (30 cm3) of the organic and aqueous phases were mixed in glass flasks by a magnetic stirrer for at least 24 h. Owing to the formation of emulsion the two phases were separated by a pressurized membrane filtration (Amicon YM3 regenerated cellulose filter, 0.5 atm) after it had been allowed to settle for 1 h. The pH was measured using a Horiba pH meter (Model F-23), and the concentration of Cu(II) in the aqueous phase was analyzed by an atomic absorption spectrophotometer (Varian Model 220FS). The concentration of Cu(II) in the organic phase was determined from a mass balance. The distribution ratio of Cu(II), DCu ; was calculated by DCu ¼ ½CuðIIÞtot =½CuðIIÞtot ;
ð1Þ
where the overbar refers to the organic phase and the subscript ‘‘tot’’ represents the total molarity of Cu(II) in all forms. The symbol ‘‘Cu(II)’’ denotes both free Cu2+ and its chelated anions in this work. 2.2.2. Nondispersive extraction In the extraction and back-extraction modules, the aqueous and organic phases were contacted in a countercurrent and wholly recycling mode. The aqueous phase flowed through the tube side and the organic phase flowed across the shell side of the modules. Otherwise specified, the flow rates Q of feed, organic, and stripping phases were 2.9, 2.7, and 2.9 cm3/s, respectively. Pressure gauges and valves were present to control the flow rates and to ensure that a positive pressure of about 2–5 psig was maintained on aqueous sides of the fibers [17]. It is accepted that the aqueousmembrane interface can be stabilized by holding a higher pressure on the nonwetting liquid (i.e., the
3613
aqueous phase) than the wetting liquid (i.e., the organic phase); but the pressure difference must be lower than that necessary to displace the wetting liquid from the pores by the nonwetting liquid [6]. Thus, the amount of solvent loss was negligible during our experiments. Prior to experiments, the organic phase and deionized water were fed into each side of the modules for 30 min. The volume of each solution (feed, organic, stripping) was 1 dm3. The flow rates of the aqueous and organic phases were fixed at required values and were stabilized by controlling pressure gauges. The stripping phase was then replaced by the HCl solution, and the feed phase was replaced by the solution containing metallic species. At this moment, the extraction and back extraction processes were started. Aqueous feed and stripping samples (2 cm3) were taken at preset time intervals, and the concentrations of Cu(II) were similarly analyzed. The concentrations of Cu(II) in the organic phase were directly obtained by a mass balance. The successive two-run was conducted following the same procedures except that the feed phase was wholly replaced by a ‘‘fresh’’ feed after 140-min operation, whereas the organic and stripping phases were maintained. Each experiment was at least duplicated under identical conditions. Reproducibility of the measurements was within 6% (mostly, 3%).
3. Results and discussion 3.1. Chemistry of solvent extraction EDTA (H4L) readily forms stable complexes with most heavy metals in a 1:1 molar ratio at pH >2.5 [18]. For a dilute, equimolar solution of Cu2+ and EDTA (10 mol/m3), it was reported that the divalent anions CuL2 absolutely dominate at pH 2.5B12 and the univalvent anions CuHL at pHo2.5 [19]. In this work, a molar concentration ratio of 4:1 of Cu2+ and EDTA was selected as the feed phase. In other words, the feed phase contains free Cu2+ cations and the anions CuL2 in a molar ratio of 3:1 at pH 2.5B12. A mixture of extractants is proposed to simultaneously extract free and the chelated metal ions from aqueous solutions. The main component in LIX64N, LIX65N (HX) has been shown to be very effective in the extraction recovery of Cu2+ [1,5,20,21] Cu2þ þ 2HX"CuX2 þ 2Hþ :
ð2Þ
On the other hand, the metallic anions such as CuL2 and CrO2 4 can be extracted by basic amine extractant Aliquat 336 (NR4Cl) under suitable pH range according to [16,22] CuL2 þ 2NR4 Cl"ðNR4 Þ2 CuL þ 2Cl :
ð3Þ
3614
S.-H. Lin, R.-S. Juang / Water Research 36 (2002) 3611–3619
Fig. 2. The extent of extraction of Cu2+ and CuL2 at different extractant compositions and initial solution pH.
When the loaded organic phase contacts an aqueous phase containing high HCl concentration, it is evident that the above two reactions will be reversed. Fig. 2 shows the effect of extractant compositions and initial pH on the extraction efficiency. The extraction efficiency at an equal volume of the aqueous organic phases is defined by
Extraction efficiency ð%Þ ¼ 100
DCu : 1 þ DCu
ð4Þ
It is seen that the extraction efficiency increases with increasing mole fraction of LIX64N, and reaches a maximum at a fraction of about 0.8. This optimal organic-phase composition is likely related to the fact that the concentration of Cu2+ is 3 times higher than that of CuL2. The same trends are observed at pH0 3B5, although the efficiency becomes higher at higher initial pH. This composition of extractant mixtures is thus selected for further studies in this work. At lower pH0 (o2), the extraction efficiency decreases first and sharply increases with increasing the mole fraction of LIX64N. This is because the EDTA-chelated anions exist dominantly in the form of CuHL, rather than CuL2, leading to inefficient extraction using Aliquat 336 [16]. In this case, Aliquat 336 base will react to a large extent with the minor LIX65N acid, thereby reducing the extraction efficiency.
Fig. 3. Effect of feed phase flow rate on the extent of Cu(II) extraction in hollow fiber contactors.
3.2. Effect of process flow rates on nondispersive extraction The effects of flow rates of the three phases on the extent of extraction are shown in Figs. 3–5, respectively. It is found that the extent of extraction increases with decreasing flow rate of feed phase Qfeed (Fig. 3). However, Seibert et al. [8] examined the effect of flow rate in extraction system of toluene/acetone/water at low superficial velocity ranges. The mass transfer efficiency is independent of the flow rate of solution flowing on tube side (aqueous phase), but dependent of one flowing shell side (organic phase). Breembroek et al. [23] showed that the permeability of Cu2+ by LIX84 increases with increasing flow rate of feed phase in either tube or shell side. In addition, Teramoto and Tanimoto [24] found that the flow rate of feed phase has little effect on the permeation rate of Cu2+ by LIX reagents. On the other hand, Campderros et al. [25] studied the extraction of Cu2+ with LIX864 in hollow fibers, and indicated that the amount of metal transport decreases with increasing feed phase velocity. The present results can be explained by mass transfer mechanism. As we will see later (Fig. 7), Cu2+ ions, rather than CuL2 anions, are preferentially extracted at the early stage of operation. At such specified condition, the process is controlled by combined interfacial chemical reaction and feed layer diffusion; and about 75% of the overall mass transfer resistance is ascribable
S.-H. Lin, R.-S. Juang / Water Research 36 (2002) 3611–3619
Fig. 4. Effect of organic phase flow rate on the extent of Cu(II) extraction in hollow fiber contactors.
3615
Fig. 6. Time profiles of the extent of Cu(II) extraction in feed and stripping phases at different operation temperatures in hollow fiber contactors.
residence time. If the dominant resistance of mass transfer is purely the aqueous layer diffusion, increasing the flow rate of feed phase will increase the extraction efficiency. On the other hand, the effects of flow rates of the organic and stripping phases are negligible under the ranges tested (Figs. 4 and 5). These agree with those obtained earlier [8,23]. The overall mass transfer coefficient in the hollow fiber extractors remained nearly constant at higher flow rates. This is because the resistance of organic stagnant layer diffusion is insignificant in the extraction process and membrane diffusion dominates in the back-extraction process [21]. As indicated above, the flow rates must be adjusted to ensure a positive pressure of 2–5 psig on aqueous sides of the fibers, thus an easily controllable combination of Qfeed ¼ 2:9 cm3/s, Qorganic ¼ 2:7 cm3/s, and Qstripping ¼ 2:9 cm3/s is selected in this work. 3.3. Effect of temperature on nondispersive extraction Fig. 5. Effect of stripping phase flow rate on the extent of Cu(II) extraction in hollow fiber contactors.
to chemical reaction [21]. Decreasing Qfeed leads to not only a decrease in feed-layer mass transfer efficiency but also an increase in extent of reaction because of a longer
As clearly shown in Fig. 6, the extent of extraction in the extraction process increases with increasing temperature but is not so pronounced in the back-extraction process. In the case of extraction module, this effect is mainly caused by an increased rate of interfacial chemical reaction. In the back extraction module, however, membrane diffusion plays an important role
3616
S.-H. Lin, R.-S. Juang / Water Research 36 (2002) 3611–3619
[21]. Although the rate of membrane diffusion increases with increasing temperature due to a decrease in viscosity of organic solution, such effect is generally less apparent than chemical reactions. 3.4. Mass transfer nature in this nondispersive extraction Fig. 7 compares the time profiles of extent of extraction in the feed and stripping phases using different operation types. One is the case of interest, i.e., the organic phase containing mixture of extractants (open symbols). It is seen that the extraction of Cu(II) is comparatively fast at to10 min and then gradually slows down at t > 10 min. According to the color change in the feed phase during experiments, we judge that Cu2+ cations may be preferentially extracted due to different mass transfer rates of two extraction systems of Cu2+-LIX64N-HCl and CuL2-Aliquat 336-HCl. Another experiment was hence made to check this point; i.e., the organic phase consists of LIX64N alone in the first 30 min and changes to Aliquat 336 alone in the following 60 min (filled symbols). This argument is justified although some deviation occurs between about 15 and 30 min. Such deviation is believed to be a result of mutual interaction of the two extractants in the organic phase, leading to a drop in effective extractant concentration compared to individual systems.
Fig. 7. Time profiles of the extent of Cu(II) extraction in feed and stripping phases at different operation types in hollow fiber contactors.
Fig. 8 shows the extraction efficiency of single Cu2+ and CuL2 in batch mode with different composition of extractant mixtures. It is seen that the efficiency decreases when the second extractant is present. Moreover, the extraction efficiency of single CuL2 with Aliquat 336 is lower so such effect is weaker compared to Cu2+/LIX64N. This proves the formation of ‘‘inert’’ species for Cu(II) extraction between LIX64N and Aliquat 336. Another evidences of FTIR spectra of the extractants are shown in Fig. 9. It is found that LIX64N shows a broad band at 3262 cm1 assigned to the H-bonded OH group. In the mixture of LIX64N and Aliquat 336, this band disappears; and several bands between 1300 and 1000 cm1 are observed assigned to the C–O stretching. This likely indicates that the interaction occurs between –OH group in LIX65N and –CH3Cl in Aliquat 336. It is reminded that the extent of back extraction reveals different trends in these two operation types (Fig. 7). The extent continuously increases using the extractant mixture. However, the extent of back extraction first increases and decreases after 20-min operation in another type (filled symbols). Such decrement at the latter stage of operation is explained by the extraction of Cu2+ as CuCl2 from the stripping phase to the 4 organic phase with Aliquat 336, in which Cu2+ has been
Fig. 8. Extraction efficiencies of single Cu2+ and CuL2 at different compositions of LIX64N/Aliquat 336 mixtures.
S.-H. Lin, R.-S. Juang / Water Research 36 (2002) 3611–3619
3617
Fig. 9. FTIR spectra of LIX64N, Aliquat 336, and the mixture of LIX64N/Aliquat 336 in kerosene.
transported at the initial 30 min with LIX64N to the stripping phase. 3.5. Feasibility of nondispersive extraction removal To evaluate the feasibility of the proposed process, the results of successive two-run operation for 260 min is shown in Fig. 10. It appears to be technically promising. Under the conditions investigated, both free Cu2+ ions and their EDTA-chelated anions CuL2 can be effectively extracted and recovered to the stripping phase. In addition, the concentration of Cu(II) in the organic phase remains a reasonable value, implying that the mutual interaction of LIX64N and Aliquat 336 in the organic phase is not so serious. It is expected that the concentration factor of Cu(II) between the feed and stripping phases will be enhanced if the volume of stripping phase is reduced. Although the concentration of Cu(II) in the organic phase maintains a nearly constant value in the recycling operations, the Cu(II) is somewhat accumulated. This is probably a result of extremely slow mass transfer rate in the back-extraction module. If the hydrophilic fibers replace the present hydrophobic ones, this shortcoming can be overcome because the resistance of membrane diffusion in the back extraction module will be reduced [6]. Similar processes have been actually documented for recovery and separation of one metal cation and
Fig. 10. Time profiles of the extent of Cu(II) extraction in feed, organic, and stripping phases for successive two-runs in hollow fiber contactors. Conditions: ½Cu2þ f ;0 ¼ 3:75 mol/m3, ½CuL2 f ;0 ¼ 1:25 mol/m3, [LIX64N]0=160 mol/m3, [Aliquat 336]0=40 mol/m3, and [HCl]s,0=4 mol/dm3.
3618
S.-H. Lin, R.-S. Juang / Water Research 36 (2002) 3611–3619
one metallic anion [26,27]. Yang et al. [26] used a module containing two sets of hollow fibers to separate and extract Cu2+ and CrO2 from aqueous 4 solutions. The aqueous phase flows on the shell side. One organic phase containing acidic extractant LIX84 flows through the bore of the fiber set 1 and the other organic phase containing basic extractant tri-noctylamine in fiber set 2. They further employed a module containing three sets of hollow fibers to extract and recover Cu2+ and CrO2 from aqueous 4 solution to separate basic and acidic stripping solutions [27]. In these two studies, the two extractants are not mixed during experiments. Certainly, the present process offers a simplified, promising alternative for this purpose.
4. Conclusions Nondispersive extraction removal of free Cu2+ and the EDTA-chelated anions CuL2 from chloride solution was investigated with a mixture of LIX64N and Aliquat 336 in two hydrophobic hollow fibers. The following results were obtained: 1. For a given free and chelated Cu(II) ion solution at initial pH 3–5, the composition of extractant mixture could be determined from batch experiments. At lower pH (o2), the extraction was less efficient because the EDTA-chelated anions mainly existed in the form of CuHL, rather than CuL2, leading to inefficient use of Aliquat 336. 2. The extent of extraction increased with decreasing flow rate of the feed phase, which could be explained by the mass transfer mechanism. The effects of flow rates of the organic and stripping phases were negligible under the ranges examined. Owing to the requirement of operation stability, a controllable flow rate condition of 2.9 (feed), 2.7 (organic), and 2.9 cm3/s (stripping) was selected. 3. The extent of Cu(II) extraction was first faster during the experiments and then slowed down. This was due to the preferential extraction of Cu2+ by LIX64N to that of CuL2 by Aliquat 336 at the early stage of operation (o10 min). Batch extraction data and FTIR spectra showed that the weak interaction took place between –OH group in LIX65N and –CH3Cl in Aliquat 336, leading to a decrease in effective extractant concentration in contrast to the individual systems. 4. The successive two-run operation for 260 min revealed that the proposed novel process seemed to be technically promising under the conditions studied. The concentration of Cu(II) in the organic phase remained reasonably low, indicating that the inter-
action of LIX64N and Aliquat 336 in the organic phase was not serious.
References [1] Ritcey GM, Ashbrook AW. Solvent extraction. Principles and applications to process metallurgy, Part I. Amsterdam: Elsevier, 1984. p. 87–171. [2] Daiminger UA, Geist AG, Nitsch W, Plucinski PK. Efficiency of hollow fiber modules for nondispersive chemical extraction. Ind Eng Chem Res 1996;35:184–91. [3] Gabelman A, Hwang ST. Hollow fiber membrane contactors. J Membr Sci 1999;159:61–106. [4] Guha AK, Yun CH, Basu R, Sirkar KK. Heavy metal removal and recovery from wastewater by a hollow fiber contained liquid membrane permeator. AIChE J. 1994;40:1223–37. [5] Kim BM. Membrane-based solvent extraction for selective removal and recovery of metals. J Membr Sci 1984;21: 5–19. [6] Prasad R, Sirkar KK. Membrane-based solvent extraction. In: Ho WSW, Sirkar KK, editors. Membrane handbook. New York: Van Nostrand Reinhold, 1992. p. 727–63. [7] Yun CH, Prasad R, Guha AK, Sirkar KK. Hollow fiber solvent extraction removal of toxic heavy metals from aqueous waste streams. Ind Eng Chem Res 1993;32: 1186–95. [8] Seibert AF, Py X, Mshewa M, Fair JR. Hydraulics and mass transfer efficiency of a commercial-scale membrane extractor. Sep Sci Technol 1993;28:343–59. [9] Conway M, Holoman S, Jones L, Leenhouts R, Williamson G. Selecting and using chelating agents. Chem Eng (NY) 1999;106(3):86–9. [10] Spearot RM, Peck JV. Recovery process for complexed copper-bearing rinse waters. Environ Prog 1984;3:124–8. [11] Abumaizar R, Khan LI. Laboratory investigation of heavy metal removal by soil washing. J Air Waste Manage Assoc 1996;46:765–8. [12] Sun B, Zhao FJ, Lombi E, McGrath SP. Leaching of heavy metals from contaminated soils using EDTA. Environ Pollut 2001;113:111–20. [13] Holcombe LJ, Behrens GP, Micheletti WC. Methods for removal of copper and iron from boiler chemical cleaning wastes. Environ Prog 1987;6:74–81. [14] Huang CP, Hsu MC, Miller P. Recovery of EDTA from power plant boiler chemical cleaning wastewater. J Environ Eng ASCE (American Society of Civil Engineers) 2000;126:919–32. [15] Tunay O, Kabdasli I, Tasli R. Pretreatment of complexed metal wastewaters. Water Sci Technol 1994;29(9): 265–77. [16] Juang RS, Chen YJ, Huang IP. Amine-based extraction recovery of Cu(II) from aqueous solution in the presence of EDTA. Equilibrium studies. Sep Sci Technol 1999;34:3099–112. [17] Juang RS, Huang IP. Hollow-fiber membrane extraction of copper(II) from aqueous ethylenediaminetetraacetic acid solution with Aliquat 336. Ind Eng Chem Res 2000;39:1409–15.
S.-H. Lin, R.-S. Juang / Water Research 36 (2002) 3611–3619 [18] Morel FMM, Hering JG. Principles and applications of aquatic chemistry. New York: Wiley, 1993. p. 332–43. [19] Juang RS, Shaiu LD. Ion exchange equilibria of metal chelates of ethylenediaminetetraacetic acid with Amberlite IRA-68. Ind Eng Chem Res 1998;37:555–71. [20] Harada M, Miyake Y, Kayahara Y. Kinetics and mechanism of metal extraction with hydroxyoximes. J Chem Eng Jpn 1989;22:168–77. [21] Lin SH, Juang RS. Mass-transfer in hollow-fiber modules for extraction and back-extraction of copper(II) with LIX64N carriers. J Membr Sci 2001;188: 251–62. [22] Alonso AI, Galan B, Irabien A, Ortiz MI. Separation of Cr(VI) with Aliquat 336: chemical equilibrium modeling. Sep Sci Technol 1997;32:1543–55. [23] Breembroek GRM, van Straalen A, Witkamp GJ, van Rosmalen GM. Extraction of cadmium and copper using
[24]
[25]
[26]
[27]
3619
hollow fiber supported liquid membranes. J Membr Sci 1998;146:185–95. Teramoto M, Tanimoto H. Mechanism of copper permeation through hollow fiber liquid membranes. Sep Sci Technol 1983;18:871–92. Campderros ME, Acosta A, Marchese J. Selective separation of copper with LIX 864 in a hollow fiber module. Talanta 1998;47:19–24. Yang ZF, Guha AK, Sirkar KK. Novel membrane-based synergistic metal extraction and recovery processes. Ind Eng Chem Res 1996;35:1383–94. Yang ZF, Guha AK, Sirkar KK. Simultaneous and synergistic extraction of cationic and anionic heavy metallic species by a mixed solvent extraction system and a novel contained liquid membrane device. Ind Eng Chem Res 1996;35:4214–20.
Removal of free and chelated Cu(II) ions from water by a nondispersive solvent extraction process Su-Hsia Lin, Ruey-Shin Juang* Department of Chemical Engineering, Yuan Ze University, Chung-Li 320, Taiwan Received 31 May 2001; received in revised form 18 January 2002; accepted 10 February 2002
Abstract The possibility of extraction of free and the ethylenediaminetetraacetic acid (EDTA)-chelated Cu(II) ions from water through a microporous hollow fiber to an organic phase containing extractants LIX64N and Aliquat 336 (a quaternary amine) was studied, in which Cu(II) was simultaneously back-extracted across another hollow fiber to a stripping solutions containing HCl. Experiments were carried out at different flow rates of feed (1.4–4.3 cm3/s), organic (1.2– 4.1 cm3/s), and stripping phases (1.4–4.3 cm3/s), and temperatures (298–318 K). It was shown that the extraction rate increased with decreasing feed phase flow rate, but the effect of flow rates of organic and stripping phases was not pronounced under the ranges studied. Mechanism of the mass transfer at each step was discussed and the effect of mutual interaction of the two extractants in the organic phase was also examined. The application potentials of this novel extraction process for this subject appeared to be technically promising. r 2002 Elsevier Science Ltd. All rights reserved. Keywords: Nondispersive extraction; Microporous hollow fibers; Cu2+; EDTA-chelated anions; LIX64N; Aliquat 336
1. Introduction Solvent (liquid–liquid) extraction has been widely used to remove efficiently toxic heavy metals from industrial effluents to environmentally acceptable levels or to recycle those metals back to the original process [1]. In the past two decades, nondispersive solvent extraction based on the microporous membranes has been developed to avoid many of the problems associated with conventional solvent extraction including solvent loss, phase separation, emulsion formation, and the shortcomings due to flooding and loading in conventional column contactors (e.g., packed towers) and mixer-settlers [2–7]. Although such nondispersive process may result in a slower rate of mass transfer compared to traditional operation if membrane resistance is significant, it can compromise possible advantages such as mass transfer rate offered by higher surface *Corresponding author. Fax: +886-3-4559373. E-mail address: [email protected] (R.-S. Juang).
area/volume associated with hollow fiber modules [2,3,8]. Besides this, the hollow fiber membrane contactors present several advantages such as a larger contact area without direct mixing of the aqueous and organic phases, the lower solvent holdup, and reduction in equipment space and volume. Strong chelating agents such as ethylenediaminetetraacetic acid (EDTA) and nitrilotriacetic acid (NTA) are often present in process and waste streams [9]. Effluents discharged from metal finishing industries and the manufacturing process of printed-circuit boards for electroless copper plating are the examples [9,10]. This is also encountered in washing solutions for remediation of metal-contaminated soils [11,12] and chemical cleaning solutions for power plant boilers [13,14]. The streams containing chelating agents may make chemical precipitation methods such as sulfide and hydroxide for heavy metals inefficient [15]. Furthermore, the high buffer capacity provided by these chelating agents requires excessive amounts of chemicals to neutralize alkalinity.
0043-1354/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 0 4 3 - 1 3 5 4 ( 0 2 ) 0 0 0 7 4 - X
S.-H. Lin, R.-S. Juang / Water Research 36 (2002) 3611–3619
3612
Nomenclature CuL2 DCu H4L HX
EDTA-chelated anions extraction distribution ratio of Cu(II) defined in Eq. (1) EDTA (ethylenediaminetetraacetic acid) LIX65N
In this work, the aqueous solutions containing Cu2+ and fewer amounts of EDTA were focused; i.e., they contained both Cu2+ cations and the EDTA-chelated anions. Such types of solutions make the extraction of both species difficult by a single extractant. In general, the metal cations can be recovered by acidic extractants such as LIX-series reagents and organophosphorus acids, whereas the metallic anions such as CrO2 and 4 EDTA-chelated anions by tertiary and quaternary amines [16,1]. An attempt was made here to evaluate the possibility of using the mixture of extractants LIX64N and Aliquat 336 for this purpose in two hollow fiber extractors. It was expected that this process could simultaneously extract free and chelated Cu(II) ions from aqueous feed solution to an organic solution, and then back-extract to a stripping solution. Process variables including the flow rates of feed (1.4–4.3 cm3/ s), organic (1.2–4.1 cm3/s), and stripping phases (1.4– 4.3 cm3/s), and temperature (298–318 K) were examined. Kinetics of individual extraction system and the effect of mutual interaction of the two extractants in the organic phase were discussed.
NR4Cl Q T
Aliquat 336 flow rate (cm3/s) temperature (K)
Subscripts f, s feed and stripping phases, respectively. 0 initial
Fig. 1. Schematic representation of experimental setup for nondispersive extraction in hollow fiber contactors. Table 1 Characteristics of the hollow fiber module used for extraction experiments 1. Description
2. Materials and methods 2.1. Apparatus, hollow fibers, and solutions Fig. 1 shows the experimental setup for extraction and back-extraction processes. The microporous hollow fiber module subject to analysis was the Hoechst Celanese Liqui-Celt Extra-Flow 2.5 8 Membrane Contactor (Model G-261). These modules are small laboratory-scale versions (with three 0–1 dm3/min pumps and flow meters) that are specifically designed for experimental purposes. Their small size makes it possible to evaluate the performance of the technology without having to prepare large amounts of the aqueous and organic phases. Additional information about this module is listed in Table 1. LIX64N was supplied from Henkel Co (now, Cognis Co.), which is a mixture of 99 vol% LIX65N (2hydroxy-5-nonyl-benzophenone oxime) and 1 vol% LIX63 (5,8-diethyl-7-hydroxy-6-dodecanone oxime). The former is very effective for Cu2+ extraction and the latter act as a kinetic accelerator [1]. Aliquat 336
2. Shell characteristics Material Length Inner diameter Outer diameter 3. Fiber characteristics Material Number of fibers Effective length Inner diameter Outer diameter Effective surface area Effective area/volume Average pore size Membrane porosity Membrane tortuosity
Hoechst Celanese LiquiCelt Extra-Flow 2.5 8 Membrane Contactor Model G-261 Polypropylene 20.3 cm 6.3 cm 7.7 cm Celgard X-30 240 polypropylene hollow fibers B10,200 19.8 cm 240 mm 300 mm 1.4 m2 29.3 cm2/cm3 0.03 mm 0.4 2.6
(tri-n-octylmethylammonium chloride) was offered from Aldrich Co. It was washed with dilute HCl solution to remove water-soluble impurities before use. The solvent
S.-H. Lin, R.-S. Juang / Water Research 36 (2002) 3611–3619
kerosene, supplied from Union Chem. Works Ltd., Taiwan, was washed twice with 20 vol% H2SO4 to remove aromatics and then with deionized water (Millipore Milli-Q) three times. Isodecanol, EDTA, CuCl2, and other inorganic chemicals were supplied by Merck Co. as analytical reagent grade. The organic phase was prepared by diluting LIX64N, Aliquat 336, and isodecanol (with equivolume of Aliquat 336) in kerosene, and the initial concentration of extractants ranged from 50 to 300 mol/m3. Isodecanol was used as modifier to avoid the formation of a third phase. The feed phase was prepared by dissolving CuCl2 and fewer amounts of EDTA in deionized water, in which the pH was adjusted by adding a small amount of 0.1 mol/dm3 HCl or NaOH. The total concentration of Cu(II) was fixed at 5 mol/m3. The stripping phase contained 4 mol/dm3 HCl. 2.2. Experimental procedures 2.2.1. Batch extraction In some cases, the extraction distribution ratios of Cu(II) were measured in batch mode. Equal volumes (30 cm3) of the organic and aqueous phases were mixed in glass flasks by a magnetic stirrer for at least 24 h. Owing to the formation of emulsion the two phases were separated by a pressurized membrane filtration (Amicon YM3 regenerated cellulose filter, 0.5 atm) after it had been allowed to settle for 1 h. The pH was measured using a Horiba pH meter (Model F-23), and the concentration of Cu(II) in the aqueous phase was analyzed by an atomic absorption spectrophotometer (Varian Model 220FS). The concentration of Cu(II) in the organic phase was determined from a mass balance. The distribution ratio of Cu(II), DCu ; was calculated by DCu ¼ ½CuðIIÞtot =½CuðIIÞtot ;
ð1Þ
where the overbar refers to the organic phase and the subscript ‘‘tot’’ represents the total molarity of Cu(II) in all forms. The symbol ‘‘Cu(II)’’ denotes both free Cu2+ and its chelated anions in this work. 2.2.2. Nondispersive extraction In the extraction and back-extraction modules, the aqueous and organic phases were contacted in a countercurrent and wholly recycling mode. The aqueous phase flowed through the tube side and the organic phase flowed across the shell side of the modules. Otherwise specified, the flow rates Q of feed, organic, and stripping phases were 2.9, 2.7, and 2.9 cm3/s, respectively. Pressure gauges and valves were present to control the flow rates and to ensure that a positive pressure of about 2–5 psig was maintained on aqueous sides of the fibers [17]. It is accepted that the aqueousmembrane interface can be stabilized by holding a higher pressure on the nonwetting liquid (i.e., the
3613
aqueous phase) than the wetting liquid (i.e., the organic phase); but the pressure difference must be lower than that necessary to displace the wetting liquid from the pores by the nonwetting liquid [6]. Thus, the amount of solvent loss was negligible during our experiments. Prior to experiments, the organic phase and deionized water were fed into each side of the modules for 30 min. The volume of each solution (feed, organic, stripping) was 1 dm3. The flow rates of the aqueous and organic phases were fixed at required values and were stabilized by controlling pressure gauges. The stripping phase was then replaced by the HCl solution, and the feed phase was replaced by the solution containing metallic species. At this moment, the extraction and back extraction processes were started. Aqueous feed and stripping samples (2 cm3) were taken at preset time intervals, and the concentrations of Cu(II) were similarly analyzed. The concentrations of Cu(II) in the organic phase were directly obtained by a mass balance. The successive two-run was conducted following the same procedures except that the feed phase was wholly replaced by a ‘‘fresh’’ feed after 140-min operation, whereas the organic and stripping phases were maintained. Each experiment was at least duplicated under identical conditions. Reproducibility of the measurements was within 6% (mostly, 3%).
3. Results and discussion 3.1. Chemistry of solvent extraction EDTA (H4L) readily forms stable complexes with most heavy metals in a 1:1 molar ratio at pH >2.5 [18]. For a dilute, equimolar solution of Cu2+ and EDTA (10 mol/m3), it was reported that the divalent anions CuL2 absolutely dominate at pH 2.5B12 and the univalvent anions CuHL at pHo2.5 [19]. In this work, a molar concentration ratio of 4:1 of Cu2+ and EDTA was selected as the feed phase. In other words, the feed phase contains free Cu2+ cations and the anions CuL2 in a molar ratio of 3:1 at pH 2.5B12. A mixture of extractants is proposed to simultaneously extract free and the chelated metal ions from aqueous solutions. The main component in LIX64N, LIX65N (HX) has been shown to be very effective in the extraction recovery of Cu2+ [1,5,20,21] Cu2þ þ 2HX"CuX2 þ 2Hþ :
ð2Þ
On the other hand, the metallic anions such as CuL2 and CrO2 4 can be extracted by basic amine extractant Aliquat 336 (NR4Cl) under suitable pH range according to [16,22] CuL2 þ 2NR4 Cl"ðNR4 Þ2 CuL þ 2Cl :
ð3Þ
3614
S.-H. Lin, R.-S. Juang / Water Research 36 (2002) 3611–3619
Fig. 2. The extent of extraction of Cu2+ and CuL2 at different extractant compositions and initial solution pH.
When the loaded organic phase contacts an aqueous phase containing high HCl concentration, it is evident that the above two reactions will be reversed. Fig. 2 shows the effect of extractant compositions and initial pH on the extraction efficiency. The extraction efficiency at an equal volume of the aqueous organic phases is defined by
Extraction efficiency ð%Þ ¼ 100
DCu : 1 þ DCu
ð4Þ
It is seen that the extraction efficiency increases with increasing mole fraction of LIX64N, and reaches a maximum at a fraction of about 0.8. This optimal organic-phase composition is likely related to the fact that the concentration of Cu2+ is 3 times higher than that of CuL2. The same trends are observed at pH0 3B5, although the efficiency becomes higher at higher initial pH. This composition of extractant mixtures is thus selected for further studies in this work. At lower pH0 (o2), the extraction efficiency decreases first and sharply increases with increasing the mole fraction of LIX64N. This is because the EDTA-chelated anions exist dominantly in the form of CuHL, rather than CuL2, leading to inefficient extraction using Aliquat 336 [16]. In this case, Aliquat 336 base will react to a large extent with the minor LIX65N acid, thereby reducing the extraction efficiency.
Fig. 3. Effect of feed phase flow rate on the extent of Cu(II) extraction in hollow fiber contactors.
3.2. Effect of process flow rates on nondispersive extraction The effects of flow rates of the three phases on the extent of extraction are shown in Figs. 3–5, respectively. It is found that the extent of extraction increases with decreasing flow rate of feed phase Qfeed (Fig. 3). However, Seibert et al. [8] examined the effect of flow rate in extraction system of toluene/acetone/water at low superficial velocity ranges. The mass transfer efficiency is independent of the flow rate of solution flowing on tube side (aqueous phase), but dependent of one flowing shell side (organic phase). Breembroek et al. [23] showed that the permeability of Cu2+ by LIX84 increases with increasing flow rate of feed phase in either tube or shell side. In addition, Teramoto and Tanimoto [24] found that the flow rate of feed phase has little effect on the permeation rate of Cu2+ by LIX reagents. On the other hand, Campderros et al. [25] studied the extraction of Cu2+ with LIX864 in hollow fibers, and indicated that the amount of metal transport decreases with increasing feed phase velocity. The present results can be explained by mass transfer mechanism. As we will see later (Fig. 7), Cu2+ ions, rather than CuL2 anions, are preferentially extracted at the early stage of operation. At such specified condition, the process is controlled by combined interfacial chemical reaction and feed layer diffusion; and about 75% of the overall mass transfer resistance is ascribable
S.-H. Lin, R.-S. Juang / Water Research 36 (2002) 3611–3619
Fig. 4. Effect of organic phase flow rate on the extent of Cu(II) extraction in hollow fiber contactors.
3615
Fig. 6. Time profiles of the extent of Cu(II) extraction in feed and stripping phases at different operation temperatures in hollow fiber contactors.
residence time. If the dominant resistance of mass transfer is purely the aqueous layer diffusion, increasing the flow rate of feed phase will increase the extraction efficiency. On the other hand, the effects of flow rates of the organic and stripping phases are negligible under the ranges tested (Figs. 4 and 5). These agree with those obtained earlier [8,23]. The overall mass transfer coefficient in the hollow fiber extractors remained nearly constant at higher flow rates. This is because the resistance of organic stagnant layer diffusion is insignificant in the extraction process and membrane diffusion dominates in the back-extraction process [21]. As indicated above, the flow rates must be adjusted to ensure a positive pressure of 2–5 psig on aqueous sides of the fibers, thus an easily controllable combination of Qfeed ¼ 2:9 cm3/s, Qorganic ¼ 2:7 cm3/s, and Qstripping ¼ 2:9 cm3/s is selected in this work. 3.3. Effect of temperature on nondispersive extraction Fig. 5. Effect of stripping phase flow rate on the extent of Cu(II) extraction in hollow fiber contactors.
to chemical reaction [21]. Decreasing Qfeed leads to not only a decrease in feed-layer mass transfer efficiency but also an increase in extent of reaction because of a longer
As clearly shown in Fig. 6, the extent of extraction in the extraction process increases with increasing temperature but is not so pronounced in the back-extraction process. In the case of extraction module, this effect is mainly caused by an increased rate of interfacial chemical reaction. In the back extraction module, however, membrane diffusion plays an important role
3616
S.-H. Lin, R.-S. Juang / Water Research 36 (2002) 3611–3619
[21]. Although the rate of membrane diffusion increases with increasing temperature due to a decrease in viscosity of organic solution, such effect is generally less apparent than chemical reactions. 3.4. Mass transfer nature in this nondispersive extraction Fig. 7 compares the time profiles of extent of extraction in the feed and stripping phases using different operation types. One is the case of interest, i.e., the organic phase containing mixture of extractants (open symbols). It is seen that the extraction of Cu(II) is comparatively fast at to10 min and then gradually slows down at t > 10 min. According to the color change in the feed phase during experiments, we judge that Cu2+ cations may be preferentially extracted due to different mass transfer rates of two extraction systems of Cu2+-LIX64N-HCl and CuL2-Aliquat 336-HCl. Another experiment was hence made to check this point; i.e., the organic phase consists of LIX64N alone in the first 30 min and changes to Aliquat 336 alone in the following 60 min (filled symbols). This argument is justified although some deviation occurs between about 15 and 30 min. Such deviation is believed to be a result of mutual interaction of the two extractants in the organic phase, leading to a drop in effective extractant concentration compared to individual systems.
Fig. 7. Time profiles of the extent of Cu(II) extraction in feed and stripping phases at different operation types in hollow fiber contactors.
Fig. 8 shows the extraction efficiency of single Cu2+ and CuL2 in batch mode with different composition of extractant mixtures. It is seen that the efficiency decreases when the second extractant is present. Moreover, the extraction efficiency of single CuL2 with Aliquat 336 is lower so such effect is weaker compared to Cu2+/LIX64N. This proves the formation of ‘‘inert’’ species for Cu(II) extraction between LIX64N and Aliquat 336. Another evidences of FTIR spectra of the extractants are shown in Fig. 9. It is found that LIX64N shows a broad band at 3262 cm1 assigned to the H-bonded OH group. In the mixture of LIX64N and Aliquat 336, this band disappears; and several bands between 1300 and 1000 cm1 are observed assigned to the C–O stretching. This likely indicates that the interaction occurs between –OH group in LIX65N and –CH3Cl in Aliquat 336. It is reminded that the extent of back extraction reveals different trends in these two operation types (Fig. 7). The extent continuously increases using the extractant mixture. However, the extent of back extraction first increases and decreases after 20-min operation in another type (filled symbols). Such decrement at the latter stage of operation is explained by the extraction of Cu2+ as CuCl2 from the stripping phase to the 4 organic phase with Aliquat 336, in which Cu2+ has been
Fig. 8. Extraction efficiencies of single Cu2+ and CuL2 at different compositions of LIX64N/Aliquat 336 mixtures.
S.-H. Lin, R.-S. Juang / Water Research 36 (2002) 3611–3619
3617
Fig. 9. FTIR spectra of LIX64N, Aliquat 336, and the mixture of LIX64N/Aliquat 336 in kerosene.
transported at the initial 30 min with LIX64N to the stripping phase. 3.5. Feasibility of nondispersive extraction removal To evaluate the feasibility of the proposed process, the results of successive two-run operation for 260 min is shown in Fig. 10. It appears to be technically promising. Under the conditions investigated, both free Cu2+ ions and their EDTA-chelated anions CuL2 can be effectively extracted and recovered to the stripping phase. In addition, the concentration of Cu(II) in the organic phase remains a reasonable value, implying that the mutual interaction of LIX64N and Aliquat 336 in the organic phase is not so serious. It is expected that the concentration factor of Cu(II) between the feed and stripping phases will be enhanced if the volume of stripping phase is reduced. Although the concentration of Cu(II) in the organic phase maintains a nearly constant value in the recycling operations, the Cu(II) is somewhat accumulated. This is probably a result of extremely slow mass transfer rate in the back-extraction module. If the hydrophilic fibers replace the present hydrophobic ones, this shortcoming can be overcome because the resistance of membrane diffusion in the back extraction module will be reduced [6]. Similar processes have been actually documented for recovery and separation of one metal cation and
Fig. 10. Time profiles of the extent of Cu(II) extraction in feed, organic, and stripping phases for successive two-runs in hollow fiber contactors. Conditions: ½Cu2þ f ;0 ¼ 3:75 mol/m3, ½CuL2 f ;0 ¼ 1:25 mol/m3, [LIX64N]0=160 mol/m3, [Aliquat 336]0=40 mol/m3, and [HCl]s,0=4 mol/dm3.
3618
S.-H. Lin, R.-S. Juang / Water Research 36 (2002) 3611–3619
one metallic anion [26,27]. Yang et al. [26] used a module containing two sets of hollow fibers to separate and extract Cu2+ and CrO2 from aqueous 4 solutions. The aqueous phase flows on the shell side. One organic phase containing acidic extractant LIX84 flows through the bore of the fiber set 1 and the other organic phase containing basic extractant tri-noctylamine in fiber set 2. They further employed a module containing three sets of hollow fibers to extract and recover Cu2+ and CrO2 from aqueous 4 solution to separate basic and acidic stripping solutions [27]. In these two studies, the two extractants are not mixed during experiments. Certainly, the present process offers a simplified, promising alternative for this purpose.
4. Conclusions Nondispersive extraction removal of free Cu2+ and the EDTA-chelated anions CuL2 from chloride solution was investigated with a mixture of LIX64N and Aliquat 336 in two hydrophobic hollow fibers. The following results were obtained: 1. For a given free and chelated Cu(II) ion solution at initial pH 3–5, the composition of extractant mixture could be determined from batch experiments. At lower pH (o2), the extraction was less efficient because the EDTA-chelated anions mainly existed in the form of CuHL, rather than CuL2, leading to inefficient use of Aliquat 336. 2. The extent of extraction increased with decreasing flow rate of the feed phase, which could be explained by the mass transfer mechanism. The effects of flow rates of the organic and stripping phases were negligible under the ranges examined. Owing to the requirement of operation stability, a controllable flow rate condition of 2.9 (feed), 2.7 (organic), and 2.9 cm3/s (stripping) was selected. 3. The extent of Cu(II) extraction was first faster during the experiments and then slowed down. This was due to the preferential extraction of Cu2+ by LIX64N to that of CuL2 by Aliquat 336 at the early stage of operation (o10 min). Batch extraction data and FTIR spectra showed that the weak interaction took place between –OH group in LIX65N and –CH3Cl in Aliquat 336, leading to a decrease in effective extractant concentration in contrast to the individual systems. 4. The successive two-run operation for 260 min revealed that the proposed novel process seemed to be technically promising under the conditions studied. The concentration of Cu(II) in the organic phase remained reasonably low, indicating that the inter-
action of LIX64N and Aliquat 336 in the organic phase was not serious.
References [1] Ritcey GM, Ashbrook AW. Solvent extraction. Principles and applications to process metallurgy, Part I. Amsterdam: Elsevier, 1984. p. 87–171. [2] Daiminger UA, Geist AG, Nitsch W, Plucinski PK. Efficiency of hollow fiber modules for nondispersive chemical extraction. Ind Eng Chem Res 1996;35:184–91. [3] Gabelman A, Hwang ST. Hollow fiber membrane contactors. J Membr Sci 1999;159:61–106. [4] Guha AK, Yun CH, Basu R, Sirkar KK. Heavy metal removal and recovery from wastewater by a hollow fiber contained liquid membrane permeator. AIChE J. 1994;40:1223–37. [5] Kim BM. Membrane-based solvent extraction for selective removal and recovery of metals. J Membr Sci 1984;21: 5–19. [6] Prasad R, Sirkar KK. Membrane-based solvent extraction. In: Ho WSW, Sirkar KK, editors. Membrane handbook. New York: Van Nostrand Reinhold, 1992. p. 727–63. [7] Yun CH, Prasad R, Guha AK, Sirkar KK. Hollow fiber solvent extraction removal of toxic heavy metals from aqueous waste streams. Ind Eng Chem Res 1993;32: 1186–95. [8] Seibert AF, Py X, Mshewa M, Fair JR. Hydraulics and mass transfer efficiency of a commercial-scale membrane extractor. Sep Sci Technol 1993;28:343–59. [9] Conway M, Holoman S, Jones L, Leenhouts R, Williamson G. Selecting and using chelating agents. Chem Eng (NY) 1999;106(3):86–9. [10] Spearot RM, Peck JV. Recovery process for complexed copper-bearing rinse waters. Environ Prog 1984;3:124–8. [11] Abumaizar R, Khan LI. Laboratory investigation of heavy metal removal by soil washing. J Air Waste Manage Assoc 1996;46:765–8. [12] Sun B, Zhao FJ, Lombi E, McGrath SP. Leaching of heavy metals from contaminated soils using EDTA. Environ Pollut 2001;113:111–20. [13] Holcombe LJ, Behrens GP, Micheletti WC. Methods for removal of copper and iron from boiler chemical cleaning wastes. Environ Prog 1987;6:74–81. [14] Huang CP, Hsu MC, Miller P. Recovery of EDTA from power plant boiler chemical cleaning wastewater. J Environ Eng ASCE (American Society of Civil Engineers) 2000;126:919–32. [15] Tunay O, Kabdasli I, Tasli R. Pretreatment of complexed metal wastewaters. Water Sci Technol 1994;29(9): 265–77. [16] Juang RS, Chen YJ, Huang IP. Amine-based extraction recovery of Cu(II) from aqueous solution in the presence of EDTA. Equilibrium studies. Sep Sci Technol 1999;34:3099–112. [17] Juang RS, Huang IP. Hollow-fiber membrane extraction of copper(II) from aqueous ethylenediaminetetraacetic acid solution with Aliquat 336. Ind Eng Chem Res 2000;39:1409–15.
S.-H. Lin, R.-S. Juang / Water Research 36 (2002) 3611–3619 [18] Morel FMM, Hering JG. Principles and applications of aquatic chemistry. New York: Wiley, 1993. p. 332–43. [19] Juang RS, Shaiu LD. Ion exchange equilibria of metal chelates of ethylenediaminetetraacetic acid with Amberlite IRA-68. Ind Eng Chem Res 1998;37:555–71. [20] Harada M, Miyake Y, Kayahara Y. Kinetics and mechanism of metal extraction with hydroxyoximes. J Chem Eng Jpn 1989;22:168–77. [21] Lin SH, Juang RS. Mass-transfer in hollow-fiber modules for extraction and back-extraction of copper(II) with LIX64N carriers. J Membr Sci 2001;188: 251–62. [22] Alonso AI, Galan B, Irabien A, Ortiz MI. Separation of Cr(VI) with Aliquat 336: chemical equilibrium modeling. Sep Sci Technol 1997;32:1543–55. [23] Breembroek GRM, van Straalen A, Witkamp GJ, van Rosmalen GM. Extraction of cadmium and copper using
[24]
[25]
[26]
[27]
3619
hollow fiber supported liquid membranes. J Membr Sci 1998;146:185–95. Teramoto M, Tanimoto H. Mechanism of copper permeation through hollow fiber liquid membranes. Sep Sci Technol 1983;18:871–92. Campderros ME, Acosta A, Marchese J. Selective separation of copper with LIX 864 in a hollow fiber module. Talanta 1998;47:19–24. Yang ZF, Guha AK, Sirkar KK. Novel membrane-based synergistic metal extraction and recovery processes. Ind Eng Chem Res 1996;35:1383–94. Yang ZF, Guha AK, Sirkar KK. Simultaneous and synergistic extraction of cationic and anionic heavy metallic species by a mixed solvent extraction system and a novel contained liquid membrane device. Ind Eng Chem Res 1996;35:4214–20.