- Email: [email protected]
Application of the hybrid complexation–ultrafiltration process for metal ion removal from aqueous solutions
Journal of Hazardous Materials 161 (2009) 1491–1498
Contents lists available at ScienceDirect
Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat
Application of the hybrid complexation–ultrafiltration process for metal ion removal from aqueous solutions Jianxian Zeng a,b,∗ , Hongqi Ye b , Zhongyu Hu a a b
College of Chemistry and Chemical Engineering, Hunan University of Science and Technology, Xiangtan 411201, PR China College of Chemistry and Chemical Engineering, Central South University, Changsha 410083, PR China
a r t i c l e
i n f o
Article history: Received 14 November 2007 Received in revised form 23 March 2008 Accepted 30 April 2008 Available online 7 May 2008 Keywords: Complexation Ultrafiltration Poly(acrylic acid) sodium salt Mercury Cadmium
a b s t r a c t Complexation–ultrafiltration process was investigated for mercury and cadmium removal from aqueous solutions by using poly(acrylic acid) sodium salt (PAASS) as a complexing agent. The kinetics of complexation reactions of PAASS with the metal ions were studied under a large excess PAASS and pH 5.5. It takes 25 and 50 min for mercury and cadmium to get the complexation equilibrium, respectively, and the reaction kinetics can be described by a pseudo-first-order equation. Effects of various operating parameters such as loading ratios, pH values, etc. on metal rejection coefficients (R) were investigated. In the process of concentration, membrane fluxes decline slowly and R values are about 1. The concentrated retentates were used further for the decomplexation. The decomplexation ratio of mercury-PAASS complex is about 30%, whereas that of cadmium-PAASS complex reaches 93.5%. After the decomplexation, diafiltration experiments were carried out at pH 2.5. Cadmium can be diafiltrated satisfactorily from the retentate, but for mercury it is the contrary. Selective separation of the both metal ions was studied from a binary solution at pH 5. When mercury, cadmium and PAASS concentrations are 30, 30 and 40 mg L−1 , respectively, mercury is retained by ultrafiltration while almost all cadmium passes through the membrane. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Many wastewaters containing heavy metal ions occur most often with flux of large volume and relatively low concentration of the metal ion, which frequently makes the operation of traditional separation techniques difficult. In order to treat better these streams, one of the most promising methods is the complexation–ultrafiltration process by the addition of watersoluble polymers in the wastewaters [1–3]. This process is based on the principles that the polymers with a larger molecular weight than molecular weight cut-off (MWCO) of ultrafiltration membranes, such as polyethyleneimine [4,5], poly(acrylic acid) [1] and poly(vinyl alcohol) [6] etc. can bind heavy metal ions to form macromolecular complexes. Heavy metal ions, which are bound to the polymers, can be retained and concentrated by ultrafiltration membranes, whereas unbound metal ions pass through the membranes [7,8].
∗ Corresponding author at: College of Chemistry and Chemical Engineering, Hunan University of Science and Technology, Xiangtan 411201, PR China. Tel.: +86 732 8290045; fax: +86 732 8290509. E-mail address: [email protected] (J. Zeng). 0304-3894/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.jhazmat.2008.04.123
The complexation–ultrafiltration process has been applied to treat various wastewaters [2,3,9,10], ground and underground waters [11,12], as well as the removal of radionuclides [13]. Trivunac and Stevanovic [14] tried to remove cadmium and zinc ions by using diethylaminoethyl cellulose as a complexing agent. It has also been reported by Mimoune and Amrani [15] that copper ion can be fixed by employing poly(vinyl alcohol). Korus et al. [16] also demonstrated that it is possible to concentrate zinc and nickel solutions by using this process. In majority of these cited studies, the effects of operational parameters and the selection of polymers were mainly considered, but the kinetics of complexation reactions for the polymer with heavy metal ions were scarcely investigated, and the integration of three steps (concentration, decomplexation and diafiltration) was seldom reported. In this work, mercury and cadmium ions were removed by using poly(acrylic acid) sodium salt (PAASS) as a water-soluble polymer. The first part of this study was performed to investigate the kinetics of complexation reaction of PAASS with the metal ions. Further works were done to study the effects of operational parameters on metal rejection coefficients, and the developments of experimental conditions integrating concentration, decomplexation and diafiltration processes. In the final part of this paper, selective sep-
1492
J. Zeng et al. / Journal of Hazardous Materials 161 (2009) 1491–1498
Table 1 Characteristics of the used hollow fiber ultrafiltration membranes Type
Material
MWCO (kDa)
Area (m2 )
Inside and outside diameter (mm)
Length (mm)
UEIP-503 UEOS-503
Polyethersulfone Polysulfone
20 6
0.3 1.5
0.8/1.2 0.25/0.4
260 240
aration of mercury and cadmium was investigated from a binary aqueous solution.
2. Experimental 2.1. Chemicals, membranes and apparatus The PAASS with average molecular weight 250,000 Da (WAKO, Japan) was used as received. Triethanolamine, tartrate sodium, metal nitrates (Hg(NO3 )2 , Cd(NO3 )2 ) and other inorganic chemicals (HNO3 , NaOH, NaCl) were all supplied by Shanghai Guanghua Sci. &Tech. Co., Ltd. (China), as analytical reagent grade. Various solutions were prepared with deionized water of conductivity less than 1 S cm−1 . If necessary, the solution pH was adjusted by adding a small amount of 1 mol L−1 HNO3 or NaOH. Two hollow fiber ultrafiltration membranes were used, all of which were offered from Tianjin Motian Membrane Eng. & Tech. Co., Ltd. (China). The characteristics of these membranes were listed in Table 1. The membrane surfaces have been modified from hydrophobic to hydrophilic and carried a great deal of negative charge, which may result in a low PAASS adsorption on the membrane surfaces. In addition, the membranes have good thermal and chemical stabilities, and are important in industrial application. The polyethersulfone membrane was used to prefiltrate the PAASS, whereas the polysulfone was selected for further use. A laboratory-scale ultrafiltration system shown in Fig. 1 was employed. The system consists of a 10 L volume capacity reservoir with a jacket, a feed gear pump which provides the flow, a flowmeter, a ultrafiltration module with axial outlets, a back pressure valve which enables optimal running conditions, pressure and temperature sensors. To keep the temperature constant during the experiment, a thermostatic water bath connected to the jacket of the reservoir was used, and a pH meter was employed to control pH value.
2.2. Experimental procedures The PAASS must be pre-filter by employing UEIP-503 membrane before use, in order to eliminate the shorter macromolecules which can go through the membrane. The aim of the choice of macromolecular weight of the PAASS was for reducing the loss of polymer during this pre-filtration. In fact, after the pre-treatment, the loss of polymer was only 2%, and much lower than that reported by Baticle et al. [2] (maximum to 25%). The pre-treated polymer was used for all complexation–ultrafiltration experiments with UEOS-503 membrane. 2.2.1. Kinetics experiments Metal ion feed solutions containing no macrocomplexant were introduced in the reservoir, and experiments were performed with recycling of the permeate and of the retentate into the feed. When metal ions were contacted with the membrane, some were adsorbed on the membrane surface, which made metal ion concentrations in the feed solutions reduce slightly. In order to eliminate the effect of this adsorption, the feed solutions were run for 1 h in the ultrafiltration system. This ensured that metal ion adsorption on the membrane surface reached the saturation, and did not change again with time. Then, metal ion concentrations in the retentate or the permeate were measured to obtain the initial concentrations. Further, a large excess of PAASS, which could be retained completely by the membrane, was mixed fully in the reservoir when the pump was operating vigorously. Once the complexation reactions began, free metal ion concentrations in the retentate reduced gradually and were obtained by means of measuring metal ion concentrations in the permeate. This was based on the assumption that the ultrafiltration membrane does not retain free metal ions. The atomic absorption spectrophotometer used for the measurement of metal concentration was a Shimadzu AA-670 instrument. The determination of mercury was done by cold vapor atomic absorption spectrometry. Each experiment was repeated three times, the average value being adopted.
Fig. 1. The laboratory-scale ultrafiltration system.
J. Zeng et al. / Journal of Hazardous Materials 161 (2009) 1491–1498
2.2.2. Total recirculation experiments Prior to the ultrafiltration experiments, the desired amounts of metal salts and PAASS were dissolved separately in water at a certain pH, and then metal ion solutions were mixed with the PAASS solution. This feed mixture was stirred fully for 2 h to make sure that the complexation equilibrium between the metal ion and PAASS was reached. The pH of the mixture was adjusted to a desired value during the mixing period. The reservoir was filled with feed solution, and the thermostatic water bath was adjusted at a certain temperature. The system was run under the desired experimental conditions. The permeate stream was returned back to the reservoir to keep the concentration constant in the feed. Permeate flux was measured by weighting the permeate volume produced in a determined quantity of time. Samples of permeate and retentate streams were taken to determine metal concentrations. 2.2.3. Concentration experiments This procedure is similar to the previous one (Section 2.2.2), except that permeate stream was sent to a separate tank. According to the total recirculation mode, polymer/metal ratio, pH value, temperature, transmembrane pressure and feed flow rate were selected. 3. Results and discussions 3.1. Kinetics of complexation reactions for PAASS with metal ions In general, since the rate of complexation reaction depends on the concentrations of metal ion, ligand and hydrogen ion, the formation rate of metal-PAASS complex is assumed to be −
d[M]t c = k[M]at [L]bt [H+ ] dt
d[M]t = k [M]at dt
Table 2 Kinetic parameters of the complexation reactions of PAASS with mercury and cadmium, respectively, (experimental conditions are the same as those in Fig. 2) Metal ions
Correlation coefficients
k1 (min−1 )
−ln [M]0
Mercury Cadmium
0.993 0.997
0.302 0.104
−2.43 −2.34
as follows −
d[M]t = k1 [M]t dt
(3)
where k1 is the pseudo-first-order rate constant. By integrating Eq. (3), taking t = 0 to t and [M]t = [M] 0 to [M]t into consideration, Eq. (4) can be derived, which is rewritten as Eq. (5). −ln[M]t = −ln[M]0 + k1 t
(4)
[M]t = [M]0 exp(−k1 t)
(5)
where [M]0 is the initial concentration of metal ion. Fig. 2 shows a kinetic curve of the complex formation of PAASS with mercury and cadmium ions respectively. It is cleanly seen that free metal ion concentrations in the retentate [M]t decrease with increasing the reaction time, reach a minimum value and do not change again. It takes 25 and 50 min for mercury and cadmium to get the complexation equilibrium, respectively. A −ln [M]t versus t plot was prepared in which the datum was fitted according to a line (not shown). Kinetic parameters were obtained and given in Table 2, which indicated that the rates of complexation reactions are first order with respect to metal ion concentrations, i.e. a = 1 in Eq. (2) under the experimental conditions. 3.2. Removal of metal ions by complexation–ultrafiltration process
(1)
where M is the free metal ion, t the reaction time, k the rate constant, L the PAASS, and a, b and c the reaction orders. Under a large excess of the PAASS and at a constant pH value, the reaction rate can be expressed as −
1493
(2)
where k is the observed rate constant. Assuming a = 1, the rate equation for the formation metal-PAASS complex can be written
Fig. 2. Kinetic curves for the formation of PAASS-mercury and PAASS-cadmium complexes, respectively, (MWCO, 6 kDa; pH, 5.5; initial metal ion concentrations, 10 mg L−1 ; initial PAASS concentration, 5000 mg L−1 ; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
3.2.1. Effect of loading ratios on the removal of metal ions In order to determine the binding capacity of PAASS (maximum metal ion amount, e.g. grams, which can be complexed by a fixed amount, e.g. 1 g, of polymer), experiments were performed with a polymer concentration of 500 mg L−1 and a change of metal ion concentration at a fixed pH value. In Fig. 3, the effect of loading ratio (LR) on mercury rejection coefficient (R) is presented at pH 5.02. A wide flat plateau exists at very high R, followed by a linear decrease of the retention with the increasing LR. Binding capacity of the polymer can be defined as the critical LR at which the decrease of R starts. As can be seen in Fig. 3, the binding capacity of PAASS
Fig. 3. Effect of loading ratio on mercury rejection coefficient (MWCO, 6 kDa; PAASS concentration, 500 mg L−1 ; pH, 5.02; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
1494
J. Zeng et al. / Journal of Hazardous Materials 161 (2009) 1491–1498
Fig. 4. Effect of loading ratio on cadmium rejection coefficient (MWCO, 6 kDa; PAASS concentration, 500 mg L−1 ; pH, 6.03; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
to mercury is 1.0 g mercury/g PAASS. Fig. 4 shows that the behavior of cadmium retention as a function of LR is very similar to that of mercury at pH 6.03. It should be noted that the binding capacity of PAASS to cadmium is 0.033 g cadmium /g PAASS, and much lower than that of mercury. 3.2.2. Effect of pH values on the removal of metal ions The pH value is one of the most important factors in the interaction of a metal ion with a binding polymer [8,17,18]. In case of mercury, Fig. 5 shows the effect of pH on R at different LR. Compared to the results obtained at high pH, R decreases significantly at a low pH. This is due to the competition of hydrogen ion with mercury trapped in the PAASS structure. It is also observed that at low LR (such as 0.4), R decreases rapidly if pH is reduced below 3.5; at high LR (such as 2), when pH varies from7.5 to 2.5, R decreases gradually with decreasing pH, and changes slowly compared to the behavior at low LR. The phenomenon may be explained as follows: At low LR, there are many available complexing sites due to high PAASS concentrations. Decreasing pH values leads to strong binding of hydrogen ion with PAASS, and some of mercury may be replaced, which makes that R reduces significantly. At high LR, since the available complexing sites are not enough to bind all mercury, a large amount of free metal ions exist. When pH is lowered, the competition between great free mercury ions and hydrogen ion results in a small change of R. In case of cadmium, the effect
Fig. 5. Effect of pH on mercury rejection coefficient at different loading ratios (MWCO, 6 kDa; PAASS concentration, 500 mg L−1 ; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
Fig. 6. Effect of pH on cadmium rejection coefficient at different loading ratios (MWCO, 6 kDa; PAASS concentration, 500 mg L−1 ; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
of pH on R is presented in Fig. 6. Since general trends and curve shapes are similar, the complexation behavior of cadmium can also be interpreted with the previous explanation for mercury. However, compared to the mercury case, cadmium complexation is more sensitive with changing pH, and this sensitivity occurs mainly in pH range 5–6.
3.2.3. Effect of the added salt on the removal of metal ions The increase of salt concentration leads to the compression of double electrical layer, and thus to the reduction of electrostatic attraction between metal ions and charged polymers [19,20]. As a result, the added salts may affect the retention of metal ions. Figs. 7 and 8 show the effects of added salt concentrations on R values in the presence of sodium chloride. It can be seen that, the influence of Cl− is great at low pH, whereas the influence is weak at high pH, i.e. this effect decreases gradually with increasing R. When R reaches the maximum value, almost no effect on R is observed, regardless of using what metal ions. This may be mainly that the polymer binds strongly metal ions at high pH, and has a good resistance to the added salts. It is also observed that, compared to the mercury case, R values of cadmium are more sensitive with increasing the salt concentration.
Fig. 7. Effect of Cl− concentration on mercury rejection coefficient at different pH values (MWCO, 6 kDa; PAASS concentration, 500 mg L−1 ; mercury concentration, 500 mg L−1 ; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
J. Zeng et al. / Journal of Hazardous Materials 161 (2009) 1491–1498
Fig. 8. Effect of Cl− concentration on cadmium rejection coefficient at different pH values (MWCO, 6 kDa; PAASS concentration, 500 mg L−1 ; cadmium concentration, 16.5 mg L−1 ; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
3.2.4. Effect of low-molecular competitive complexing agents on the removal of metal ions Low-molecular competitive complexing agents, such as ammonia, tartaric acid, citric acid or triethanolamine etc., can form strong complexes with metal ions, and pass easily through the ultrafiltration membrane. If these agents are present and compete with the polymer for metal ions, the removal of metal ions is also complicated. In Figs. 9 and 10, the influences of tartrate sodium and triethanolamine as representative ligands on the retention of mercury and cadmium are presented at different pH values. One can see that the retention of mercury is not influenced by any of the investigated competing complexing agents. R values observed in the different solutions are almost equal. In contrast to mercury, R values of cadmium are significantly reduced in the presence of triethanolamine at pH range 5–7 or tartrate sodium at pH range 5.3–6.5. Nevertheless, in solutions with pH 7, the PAASS can successfully compete with the both ligands for cadmium and a high R value is obtained. This means that the influence of competitive complexing agents can be eliminated by increasing pH to a certain value.
Fig. 9. Effect of competitive complexing agents on mercury rejection coefficient at different pH values (MWCO, 6 kDa; PAASS concentration, 500 mg L−1 ; mercury concentration, 500 mg L−1 ; competitive complexing agents concentration, 5000 mg L−1 ; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
1495
Fig. 10. Effect of competitive complexing agents on cadmium rejection coefficient at different pH values (MWCO, 6 kDa; PAASS concentration, 500 mg L−1 ; cadmium concentration, 16.5 mg L− 1; competitive complexing agents concentration, 5000 mg L−1 ; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
3.3. Concentration test Figs. 11 and 12 show the effects of volume concentration factors on permeate fluxes (J) and metal rejection coefficients. In case of mercury, J value reduces from 6.2 to 5.4 L m−2 h−1 during the concentration process, i.e. the membrane permeability decreases from 27.9 to 24.3 L m−2 h−1 bar−1 , which is close to the deionized water permeability of the new membrane 30.6 L m−2 h−1 bar−1 . Therefore, for the concentration range studied, the effect of the PAASS on flux is not dominant. A similar result has been obtained by Uludag et al. [21] in the study of polymer-enhanced ultrafiltration using polyethyleneimine as a complexing agent. In case of cadmium, the behavior of J value as a function of volume concentration factor is very similar to that of mercury. While analyzing the retention curves, it should be emphasized that R values of both metal ions are very high (approximated to 1) in the process. Mercury concentrations in the retentate and the permeate are 1499.6 and 0.03 mg L−1 , respectively, which corresponds to a volume concentration factor of 15. As for cadmium, metal concentrations in the retentate and the permeate are 99.6 and 0.05 mg L−1 respectively when volume concentration factor reaches 10. Therefore, both metal concentrations in the permeates are lower than the maximum value fixed by the integrated wastewater discharge standard of china.
Fig. 11. Effect of volume concentration factor on membrane permeate flux and mercury rejection coefficient (MWCO, 6 kDa; pH, 5; initial PAASS concentration, 100 mg L−1 ; initial mercury concentration, 100 mg L−1 ; feed volume, 30 L; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
1496
J. Zeng et al. / Journal of Hazardous Materials 161 (2009) 1491–1498
Fig. 12. Effect of volume concentration factor on membrane permeate flux and cadmium rejection coefficient (MWCO, 6 kDa; pH, 6; initial PAASS concentration, 300 mg L−1 ; initial cadmium concentration, 10 mg L−1 ; feed volume, 30 L; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
3.4. Decomplexation test At this step, the decomplexation of metal-polymer complex was carried out by acidification of the retentate to a low pH value. For this reason, the final concentrated solutions (in Section 3.3) were acidified until pH value reached 2.5, and then subjected to an ultrafiltration operation in total recirculation mode. Metal concentrations in the retentate and the permeate were compared, and the change of permeate flux with time and rejection coefficient of polymer (RPAASS ) were measured. The results are presented in Figs. 13 and 14. In case of mercury, the decomplexation of mercury-PAASS complex occurs actually at pH 2.5. Mercury concentration in the permeate is 440 mg L−1 and equal to free mercury concentration in the retentate, whereas the total mercury concentration in the retentate is about 1500 mg L−1 . This means that about 30% of ¨ mercury-PAASS complexes are dissociated. Muslehiddino˘ glu et al. [17] studied the effect of pH values on the retention of mercury at different LRs. They indicated that the decrease of mercury retention starts if pH is reduced below 3. This is because that some of mercury-polymer complexes are destroyed at low pH. In case of cadmium, cadmium concentrations in the retentate and the permeate are 99.6 and 93.1 mg L−1 , respectively, which makes that 93.5% of cadmium-PAASS complexes are dissociated. In addition, it is clearly observed that mercury or cadmium concentrations in the retentates are always greater than those in the permeates. This
Fig. 13. Effect of time on membrane permeate flux, PAASS rejection coefficient, mercury concentrations in the retentate and the permeate, respectively, (MWCO, 6 kDa; pH, 2.5; PAASS concentration, 1500 mg L−1 ; initial mercury concentration, 1499.6 mg L−1 ; feed volume, 2 L; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
Fig. 14. Effect of time on membrane permeate flux, PAASS rejection coefficient, cadmium concentrations in the retentate and the permeate, respectively, (MWCO, 6 kDa; pH, 2.5; PAASS concentration, 3000 mg L−1 ; initial cadmium concentration, 99.6 mg L−1 ; feed volume, 2 L; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
indicates that some of mercury or cadmium ions are still trapped by the polymer. Permeate fluxes decline with operating time, and it does not change again after 30 min. It has been also observed that permeate flux is rather lower than that measured at pH 5 or 6 (in Section 3.3). This may be explained by the formation of a polymer cake layer on the membrane surface, following a change of the conformation of the polymer in strong acidic media [22]. The membrane with MWCO 6 kDa can retain the PAASS. Thus, the polymer does not have a leakage during the process of decomplexation, and may be recycled completely. 3.5. Diafiltration test In order to extract the maximum amount of metal ions and make PAASS pure as possible as, diafiltration experiments were carried out after the decomplexation. Two liters of the retentate (in Section 3.4) was treated by eight successive additions of 1 L of pH 2.5 nitric acid solutions. The changes of metal concentrations of the retentates and permeates are shown in Figs. 15 and 16. To describe the diafiltration quantitatively, the metal extraction ratio (MER) is defined as follow:
MER =
1−
[M]r,d [M]f,d
100%
(6)
Fig. 15. Effect of diafiltration volume on mercury concentrations in the retentate and the permeate, respectively, (MWCO, 6 kDa; pH, 2.5; PAASS concentration, 1500 mg L−1 ; initial mercury concentration, 1499.6 mg L−1 ; feed volume, 2 L; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
J. Zeng et al. / Journal of Hazardous Materials 161 (2009) 1491–1498
Fig. 16. Effect of diafiltration volume on cadmium concentrations in the retentate and the permeate, respectively, (MWCO, 6 kDa; pH, 2.5; PAASS concentration, 3000 mg L−1 ; initial cadmium concentration, 99.6 mg L−1 ; feed volume, 2 L; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
where [M]r,d is metal concentration of the retentate at a certain diafiltration volume, [M]f,d is initial metal concentration in the retentate. In case of mercury, it is observed that, if a diafiltration volume is equal to four times than that of the retentate, i.e. 8 L, mercury concentration in the retentate decreases from the beginning of 1499.6 mg L−1 to the end of 810.4 mg L−1 , which corresponds to mercury MER of 46%. In case of cadmium, after 8 L of the diafiltration volume, cadmium concentration in the retentate reduces from 99.6 to 3.87 mg L−1 , which makes that cadmium MER reaches as high as 96.1%. Figs. 15 and 16 also show that, when the diafiltration volume reaches 8 L, mercury concentration in the retentate is much higher than that in the permeate, whereas cadmium concentration of the retentate is almost equal to that of the permeate. Therefore, cadmium can be extracted effectively and the purification of polymer is acceptably satisfactory, but for mercury case it is the contrary. 3.6. Selective separation of binary metal ions To investigate the selective separation of mercury and cadmium ions by the complexation–ultrafiltration process, experiments were carried out at pH 5 using a binary solution in which each metal ion
1497
Fig. 18. Effect of volume concentration factor on rejection coefficient during the selective separation of mercury and cadmium (experimental conditions are the same as those in Fig. 17).
concentration was the same. The feed volume was 30 L, the concentrations of mercury, cadmium and PAASS were 30, 30 and 40 mg L−1 , respectively, and there were not the added salts and low-molecular competitive complexing agents in the solution. After the metal concentration in the retentate is measured, metal concentration ratio (MCR) is calculated by using the following expression: MCR =
[M] [M]
r, s
(7)
f, s
where [M]r,s is metal concentration of the retentate at a certain volume concentration factor, [M]f,s is initial metal concentration in the feed. Fig. 17 shows the effect of volume concentration factor on MCR value. There is a linear growth of MCR for mercury in the process of concentration. When volume concentration factor reaches 15, mercury MCR is 14.83 and mercury concentration in the retentate is 444.9 mg L−1 . In contrast to mercury, cadmium concentration in the retentate grows very slowly during this process. Under the experimental conditions, final cadmium MCR is 1.18 and cadmium concentration of the retentate is only 35.4 mg L−1 . This means that mercury is retained efficiently by the ultrafiltration membrane while almost all cadmium passes through it. Fig. 18 shows the effect of volume concentration factor on the retention of both metal ions. It is cleanly seen that R value of mercury is very high (>0.99), while that of cadmium is fairly low (about 0.1). Practically, at pH 5, PAASS forms very stable complexes with mercury but less stable with cadmium. R values of both metal ions correspond to the accumulation of their macromolecular complexes in the retentate. 4. Conclusions
Fig. 17. Effect of volume concentration factor on the metal concentration ratio during the selective separation of mercury and cadmium (MWCO, 6 kDa; initial PAASS concentration, 40 mg L− 1; initial mercury concentrations, 30 mg L− 1; initial cadmium concentrations, 30 mg L− 1; pH, 5; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
The complexation–ultrafiltration process is viable and suitable to remove mercury and cadmium ions from aqueous solutions with the help of water-soluble polymeric ligand PAASS. At pH 5.5, the kinetics of complexation reactions of PAASS with the metal ions were investigated in the present of a large excess PAASS. It takes 25 and 50 min for mercury and cadmium to get the complexation equilibrium, respectively, and the reaction kinetics can be described by a pseudo-first-order equation. Effects of various operating parameters on metal rejection coefficients were investigated in detail. The binding capacities for PAASS to both metal ions are 1.0 g mercury/g PAASS and 0.033 g cadmium/g PAASS. Compared to the mercury case, the cadmium complexation is more sensitive with changing pH. The influence of the added salt is great at low pH, whereas the influence is
1498
J. Zeng et al. / Journal of Hazardous Materials 161 (2009) 1491–1498
weak at high pH. Low molecular competitive complexing agents do not affect the complexation behavior of PAASS with mercury, but reduce the retention of cadmium. In the process of the concentration, permeate fluxes decline insignificantly and rejection coefficients of both metals are very high. At the decomplexation stage, the decomplexation ratio of mercury-PAASS complex is only about 30%, whereas that of cadmium-PAASS complex reaches as high as 93.5%. In the diafiltration process, cadmium in the retentate can be extracted effectively and a purified PAASS is obtained, but for mercury case it is the contrary. At a special pH value, the selective separation of mercury and cadmium ions can be achieved by controlling the concentrations of metal ions and PAASS. Almost all cadmium is in the permeate, while mercury is retained in the retentate. Acknowledgements This research is supported by a Project Supported by Hunan Provincial Natural Science Foundation of China (no. 05JJ40020). Also, the authors acknowledge the financial supports from a Project Supported by Scientific Research Fund of Hunan Provincial Education Department (no. 05C170). References [1] Y.F. Zhang, Zh.L. Xu, Study on the treatment of industrial wastewater containing Pb2+ ion using a coupling process of polymer complexation–ultrafiltration, Sep. Sci. Technol. 38 (2003) 1585–1596. [2] P. Baticle, C. Kiefer, N. Lakhchaf, O. Leclerc, M. Persin, J. Sarrazin, Treatment of nickel containing industrial effluents with a hybrid process comprising of polymer complexation–ultrafiltration – electrolysis, Sep. Purif. Technol. 18 (2000) 195–207. [3] C.R. Tavares, M. Vieira, J.C.C. Petrus, E.C. Bortoletto, F. Ceravollo, Ultrafiltration/complexation process for metal removal from pulp and paper industry wastewater, Desalination 144 (2002) 261–265. [4] A. Kryvoruchko, L. Yurlova, B. Kornilovich, Purification of water containing heavy metals by chelating-enhanced ultrafiltration, Desalination 144 (2002) 243–248.
[5] J. Barron-Zambrano, S. Laborie, P. Viers, M. Rakib, G. Durand, Mercury removal from aqueous solutions by complexation–ultrafiltration, Desalination 144 (2002) 201–206. [6] C. Dilek, H.O. Ozbelge, N. Bicak, L. Yilmaz, Removal of boron from aqueous solutions by continuous polymer-enhanced ultrafiltration with polyvinyl alcohol, Sep. Sci. Technol. 37 (2002) 1257–1271. [7] R. Molinari, S. Gallo, P. Argurio, Metal ions removal from wastewater or washing water from contaminated soil by ultrafiltration–complexation, Water Res. 38 (2004) 593–600. [8] R.S. Juang, C.H. Chiou, Ultrafiltration rejection of dissolved ions using various weakly basic water-soluble polymers, J. Membr. Sci. 177 (2000) 207–214. [9] S. Petrov, V. Nenov, Removal and recovery of cooper from wastewater by a complexation–ultrafiltration process, Desalination 162 (2004) 201–209. [10] M. Bodzek, I. Korus, K. Loska, Application of the hybrid complexation–ultrafiltration process for removal of metal ions from galvanic wastewater, Desalination 121 (1999) 117–121. [11] S. Verbych, M. Bryk, A. Alpatova, G. Chornokur, Ground water treatment by enhanced ultrafiltration, Desalination 179 (2005) 237–244. [12] J. Bohdziewicz, Removal of chromium ions (VI) from underground water in the hybrid complexation–ultrafiltration process, Desalination 129 (2000) 227–235. [13] A. Kryvoruchko, L. Yurlova, I. Atamanenko, B. Kornilovich, Ultrafiltration removal of U (VI) from contaminated water, Desalination 162 (2004) 229–236. [14] K. Trivunac, S. Stevanovic, Effects of operating parameters on efficiency of cadmium and zinc removal by the complexation–filtration process, Desalination 198 (2006) 282–287. [15] S. Mimoune, F. Amrani, Experimental study of metal ions removal from aqueous solutions by complexation–ultrafiltration, J. Membr. Sci. 298 (2007) 92–98. [16] I. Korus, M. Bodzek, K. Loska, Removal of zinc and nickel ions from aqueous solutions by mean of the hybrid complexation–ultrafiltration process, Sep. Purif. Technol. 17 (1999) 111–116. ¨ Ozbelge, ¨ ¨ [17] J. Muslehiddino˘ glu, Y. Uluda˘g, H.O. L. Yilmaz, Effect of operating parameters on selective separation of heavy metals from binary mixture via polymer enhanced ultrafiltration, J. Membr. Sci. 140 (1998) 251–266. [18] K. Trivunac, S. Stevanovic, Removal of heavy metal ions from water by complexation-assisted ultrafiltration, Chemosphere 64 (2006) 486–491. [19] A. Tabatabai, J.F. Scamehorn, S.D. Christian, Water softening using polyelectrolyte-enhanced ultrafiltration, Sep. Sci. Technol. 30 (1995) 211– 224. [20] R.S. Juang, M.N. Chen, Retention of copper-EDTA chelates from dilute aqueous solution by a polyelectrolyte-enhanced ultrafiltration process, J. Membr. Sci. 119 (1996) 25–37. ¨ Ozbelge, ¨ [21] Y. Uludag, H.O. L. Yilmaz, Removal of mercury from aqueous solutions via polymer enhanced ultrafiltration, J. Membr. Sci. 129 (1997) 93–99. ´ ´ [22] P. Canizares, A.D. Lucas, A. Perez, R. Camarillo, Effect of polymer nature and hydrodynamic conditions on a process of polymer enhanced ultrafiltration, J. Membr. Sci. 253 (2005) 149–163.
Contents lists available at ScienceDirect
Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat
Application of the hybrid complexation–ultrafiltration process for metal ion removal from aqueous solutions Jianxian Zeng a,b,∗ , Hongqi Ye b , Zhongyu Hu a a b
College of Chemistry and Chemical Engineering, Hunan University of Science and Technology, Xiangtan 411201, PR China College of Chemistry and Chemical Engineering, Central South University, Changsha 410083, PR China
a r t i c l e
i n f o
Article history: Received 14 November 2007 Received in revised form 23 March 2008 Accepted 30 April 2008 Available online 7 May 2008 Keywords: Complexation Ultrafiltration Poly(acrylic acid) sodium salt Mercury Cadmium
a b s t r a c t Complexation–ultrafiltration process was investigated for mercury and cadmium removal from aqueous solutions by using poly(acrylic acid) sodium salt (PAASS) as a complexing agent. The kinetics of complexation reactions of PAASS with the metal ions were studied under a large excess PAASS and pH 5.5. It takes 25 and 50 min for mercury and cadmium to get the complexation equilibrium, respectively, and the reaction kinetics can be described by a pseudo-first-order equation. Effects of various operating parameters such as loading ratios, pH values, etc. on metal rejection coefficients (R) were investigated. In the process of concentration, membrane fluxes decline slowly and R values are about 1. The concentrated retentates were used further for the decomplexation. The decomplexation ratio of mercury-PAASS complex is about 30%, whereas that of cadmium-PAASS complex reaches 93.5%. After the decomplexation, diafiltration experiments were carried out at pH 2.5. Cadmium can be diafiltrated satisfactorily from the retentate, but for mercury it is the contrary. Selective separation of the both metal ions was studied from a binary solution at pH 5. When mercury, cadmium and PAASS concentrations are 30, 30 and 40 mg L−1 , respectively, mercury is retained by ultrafiltration while almost all cadmium passes through the membrane. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Many wastewaters containing heavy metal ions occur most often with flux of large volume and relatively low concentration of the metal ion, which frequently makes the operation of traditional separation techniques difficult. In order to treat better these streams, one of the most promising methods is the complexation–ultrafiltration process by the addition of watersoluble polymers in the wastewaters [1–3]. This process is based on the principles that the polymers with a larger molecular weight than molecular weight cut-off (MWCO) of ultrafiltration membranes, such as polyethyleneimine [4,5], poly(acrylic acid) [1] and poly(vinyl alcohol) [6] etc. can bind heavy metal ions to form macromolecular complexes. Heavy metal ions, which are bound to the polymers, can be retained and concentrated by ultrafiltration membranes, whereas unbound metal ions pass through the membranes [7,8].
∗ Corresponding author at: College of Chemistry and Chemical Engineering, Hunan University of Science and Technology, Xiangtan 411201, PR China. Tel.: +86 732 8290045; fax: +86 732 8290509. E-mail address: [email protected] (J. Zeng). 0304-3894/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.jhazmat.2008.04.123
The complexation–ultrafiltration process has been applied to treat various wastewaters [2,3,9,10], ground and underground waters [11,12], as well as the removal of radionuclides [13]. Trivunac and Stevanovic [14] tried to remove cadmium and zinc ions by using diethylaminoethyl cellulose as a complexing agent. It has also been reported by Mimoune and Amrani [15] that copper ion can be fixed by employing poly(vinyl alcohol). Korus et al. [16] also demonstrated that it is possible to concentrate zinc and nickel solutions by using this process. In majority of these cited studies, the effects of operational parameters and the selection of polymers were mainly considered, but the kinetics of complexation reactions for the polymer with heavy metal ions were scarcely investigated, and the integration of three steps (concentration, decomplexation and diafiltration) was seldom reported. In this work, mercury and cadmium ions were removed by using poly(acrylic acid) sodium salt (PAASS) as a water-soluble polymer. The first part of this study was performed to investigate the kinetics of complexation reaction of PAASS with the metal ions. Further works were done to study the effects of operational parameters on metal rejection coefficients, and the developments of experimental conditions integrating concentration, decomplexation and diafiltration processes. In the final part of this paper, selective sep-
1492
J. Zeng et al. / Journal of Hazardous Materials 161 (2009) 1491–1498
Table 1 Characteristics of the used hollow fiber ultrafiltration membranes Type
Material
MWCO (kDa)
Area (m2 )
Inside and outside diameter (mm)
Length (mm)
UEIP-503 UEOS-503
Polyethersulfone Polysulfone
20 6
0.3 1.5
0.8/1.2 0.25/0.4
260 240
aration of mercury and cadmium was investigated from a binary aqueous solution.
2. Experimental 2.1. Chemicals, membranes and apparatus The PAASS with average molecular weight 250,000 Da (WAKO, Japan) was used as received. Triethanolamine, tartrate sodium, metal nitrates (Hg(NO3 )2 , Cd(NO3 )2 ) and other inorganic chemicals (HNO3 , NaOH, NaCl) were all supplied by Shanghai Guanghua Sci. &Tech. Co., Ltd. (China), as analytical reagent grade. Various solutions were prepared with deionized water of conductivity less than 1 S cm−1 . If necessary, the solution pH was adjusted by adding a small amount of 1 mol L−1 HNO3 or NaOH. Two hollow fiber ultrafiltration membranes were used, all of which were offered from Tianjin Motian Membrane Eng. & Tech. Co., Ltd. (China). The characteristics of these membranes were listed in Table 1. The membrane surfaces have been modified from hydrophobic to hydrophilic and carried a great deal of negative charge, which may result in a low PAASS adsorption on the membrane surfaces. In addition, the membranes have good thermal and chemical stabilities, and are important in industrial application. The polyethersulfone membrane was used to prefiltrate the PAASS, whereas the polysulfone was selected for further use. A laboratory-scale ultrafiltration system shown in Fig. 1 was employed. The system consists of a 10 L volume capacity reservoir with a jacket, a feed gear pump which provides the flow, a flowmeter, a ultrafiltration module with axial outlets, a back pressure valve which enables optimal running conditions, pressure and temperature sensors. To keep the temperature constant during the experiment, a thermostatic water bath connected to the jacket of the reservoir was used, and a pH meter was employed to control pH value.
2.2. Experimental procedures The PAASS must be pre-filter by employing UEIP-503 membrane before use, in order to eliminate the shorter macromolecules which can go through the membrane. The aim of the choice of macromolecular weight of the PAASS was for reducing the loss of polymer during this pre-filtration. In fact, after the pre-treatment, the loss of polymer was only 2%, and much lower than that reported by Baticle et al. [2] (maximum to 25%). The pre-treated polymer was used for all complexation–ultrafiltration experiments with UEOS-503 membrane. 2.2.1. Kinetics experiments Metal ion feed solutions containing no macrocomplexant were introduced in the reservoir, and experiments were performed with recycling of the permeate and of the retentate into the feed. When metal ions were contacted with the membrane, some were adsorbed on the membrane surface, which made metal ion concentrations in the feed solutions reduce slightly. In order to eliminate the effect of this adsorption, the feed solutions were run for 1 h in the ultrafiltration system. This ensured that metal ion adsorption on the membrane surface reached the saturation, and did not change again with time. Then, metal ion concentrations in the retentate or the permeate were measured to obtain the initial concentrations. Further, a large excess of PAASS, which could be retained completely by the membrane, was mixed fully in the reservoir when the pump was operating vigorously. Once the complexation reactions began, free metal ion concentrations in the retentate reduced gradually and were obtained by means of measuring metal ion concentrations in the permeate. This was based on the assumption that the ultrafiltration membrane does not retain free metal ions. The atomic absorption spectrophotometer used for the measurement of metal concentration was a Shimadzu AA-670 instrument. The determination of mercury was done by cold vapor atomic absorption spectrometry. Each experiment was repeated three times, the average value being adopted.
Fig. 1. The laboratory-scale ultrafiltration system.
J. Zeng et al. / Journal of Hazardous Materials 161 (2009) 1491–1498
2.2.2. Total recirculation experiments Prior to the ultrafiltration experiments, the desired amounts of metal salts and PAASS were dissolved separately in water at a certain pH, and then metal ion solutions were mixed with the PAASS solution. This feed mixture was stirred fully for 2 h to make sure that the complexation equilibrium between the metal ion and PAASS was reached. The pH of the mixture was adjusted to a desired value during the mixing period. The reservoir was filled with feed solution, and the thermostatic water bath was adjusted at a certain temperature. The system was run under the desired experimental conditions. The permeate stream was returned back to the reservoir to keep the concentration constant in the feed. Permeate flux was measured by weighting the permeate volume produced in a determined quantity of time. Samples of permeate and retentate streams were taken to determine metal concentrations. 2.2.3. Concentration experiments This procedure is similar to the previous one (Section 2.2.2), except that permeate stream was sent to a separate tank. According to the total recirculation mode, polymer/metal ratio, pH value, temperature, transmembrane pressure and feed flow rate were selected. 3. Results and discussions 3.1. Kinetics of complexation reactions for PAASS with metal ions In general, since the rate of complexation reaction depends on the concentrations of metal ion, ligand and hydrogen ion, the formation rate of metal-PAASS complex is assumed to be −
d[M]t c = k[M]at [L]bt [H+ ] dt
d[M]t = k [M]at dt
Table 2 Kinetic parameters of the complexation reactions of PAASS with mercury and cadmium, respectively, (experimental conditions are the same as those in Fig. 2) Metal ions
Correlation coefficients
k1 (min−1 )
−ln [M]0
Mercury Cadmium
0.993 0.997
0.302 0.104
−2.43 −2.34
as follows −
d[M]t = k1 [M]t dt
(3)
where k1 is the pseudo-first-order rate constant. By integrating Eq. (3), taking t = 0 to t and [M]t = [M] 0 to [M]t into consideration, Eq. (4) can be derived, which is rewritten as Eq. (5). −ln[M]t = −ln[M]0 + k1 t
(4)
[M]t = [M]0 exp(−k1 t)
(5)
where [M]0 is the initial concentration of metal ion. Fig. 2 shows a kinetic curve of the complex formation of PAASS with mercury and cadmium ions respectively. It is cleanly seen that free metal ion concentrations in the retentate [M]t decrease with increasing the reaction time, reach a minimum value and do not change again. It takes 25 and 50 min for mercury and cadmium to get the complexation equilibrium, respectively. A −ln [M]t versus t plot was prepared in which the datum was fitted according to a line (not shown). Kinetic parameters were obtained and given in Table 2, which indicated that the rates of complexation reactions are first order with respect to metal ion concentrations, i.e. a = 1 in Eq. (2) under the experimental conditions. 3.2. Removal of metal ions by complexation–ultrafiltration process
(1)
where M is the free metal ion, t the reaction time, k the rate constant, L the PAASS, and a, b and c the reaction orders. Under a large excess of the PAASS and at a constant pH value, the reaction rate can be expressed as −
1493
(2)
where k is the observed rate constant. Assuming a = 1, the rate equation for the formation metal-PAASS complex can be written
Fig. 2. Kinetic curves for the formation of PAASS-mercury and PAASS-cadmium complexes, respectively, (MWCO, 6 kDa; pH, 5.5; initial metal ion concentrations, 10 mg L−1 ; initial PAASS concentration, 5000 mg L−1 ; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
3.2.1. Effect of loading ratios on the removal of metal ions In order to determine the binding capacity of PAASS (maximum metal ion amount, e.g. grams, which can be complexed by a fixed amount, e.g. 1 g, of polymer), experiments were performed with a polymer concentration of 500 mg L−1 and a change of metal ion concentration at a fixed pH value. In Fig. 3, the effect of loading ratio (LR) on mercury rejection coefficient (R) is presented at pH 5.02. A wide flat plateau exists at very high R, followed by a linear decrease of the retention with the increasing LR. Binding capacity of the polymer can be defined as the critical LR at which the decrease of R starts. As can be seen in Fig. 3, the binding capacity of PAASS
Fig. 3. Effect of loading ratio on mercury rejection coefficient (MWCO, 6 kDa; PAASS concentration, 500 mg L−1 ; pH, 5.02; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
1494
J. Zeng et al. / Journal of Hazardous Materials 161 (2009) 1491–1498
Fig. 4. Effect of loading ratio on cadmium rejection coefficient (MWCO, 6 kDa; PAASS concentration, 500 mg L−1 ; pH, 6.03; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
to mercury is 1.0 g mercury/g PAASS. Fig. 4 shows that the behavior of cadmium retention as a function of LR is very similar to that of mercury at pH 6.03. It should be noted that the binding capacity of PAASS to cadmium is 0.033 g cadmium /g PAASS, and much lower than that of mercury. 3.2.2. Effect of pH values on the removal of metal ions The pH value is one of the most important factors in the interaction of a metal ion with a binding polymer [8,17,18]. In case of mercury, Fig. 5 shows the effect of pH on R at different LR. Compared to the results obtained at high pH, R decreases significantly at a low pH. This is due to the competition of hydrogen ion with mercury trapped in the PAASS structure. It is also observed that at low LR (such as 0.4), R decreases rapidly if pH is reduced below 3.5; at high LR (such as 2), when pH varies from7.5 to 2.5, R decreases gradually with decreasing pH, and changes slowly compared to the behavior at low LR. The phenomenon may be explained as follows: At low LR, there are many available complexing sites due to high PAASS concentrations. Decreasing pH values leads to strong binding of hydrogen ion with PAASS, and some of mercury may be replaced, which makes that R reduces significantly. At high LR, since the available complexing sites are not enough to bind all mercury, a large amount of free metal ions exist. When pH is lowered, the competition between great free mercury ions and hydrogen ion results in a small change of R. In case of cadmium, the effect
Fig. 5. Effect of pH on mercury rejection coefficient at different loading ratios (MWCO, 6 kDa; PAASS concentration, 500 mg L−1 ; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
Fig. 6. Effect of pH on cadmium rejection coefficient at different loading ratios (MWCO, 6 kDa; PAASS concentration, 500 mg L−1 ; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
of pH on R is presented in Fig. 6. Since general trends and curve shapes are similar, the complexation behavior of cadmium can also be interpreted with the previous explanation for mercury. However, compared to the mercury case, cadmium complexation is more sensitive with changing pH, and this sensitivity occurs mainly in pH range 5–6.
3.2.3. Effect of the added salt on the removal of metal ions The increase of salt concentration leads to the compression of double electrical layer, and thus to the reduction of electrostatic attraction between metal ions and charged polymers [19,20]. As a result, the added salts may affect the retention of metal ions. Figs. 7 and 8 show the effects of added salt concentrations on R values in the presence of sodium chloride. It can be seen that, the influence of Cl− is great at low pH, whereas the influence is weak at high pH, i.e. this effect decreases gradually with increasing R. When R reaches the maximum value, almost no effect on R is observed, regardless of using what metal ions. This may be mainly that the polymer binds strongly metal ions at high pH, and has a good resistance to the added salts. It is also observed that, compared to the mercury case, R values of cadmium are more sensitive with increasing the salt concentration.
Fig. 7. Effect of Cl− concentration on mercury rejection coefficient at different pH values (MWCO, 6 kDa; PAASS concentration, 500 mg L−1 ; mercury concentration, 500 mg L−1 ; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
J. Zeng et al. / Journal of Hazardous Materials 161 (2009) 1491–1498
Fig. 8. Effect of Cl− concentration on cadmium rejection coefficient at different pH values (MWCO, 6 kDa; PAASS concentration, 500 mg L−1 ; cadmium concentration, 16.5 mg L−1 ; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
3.2.4. Effect of low-molecular competitive complexing agents on the removal of metal ions Low-molecular competitive complexing agents, such as ammonia, tartaric acid, citric acid or triethanolamine etc., can form strong complexes with metal ions, and pass easily through the ultrafiltration membrane. If these agents are present and compete with the polymer for metal ions, the removal of metal ions is also complicated. In Figs. 9 and 10, the influences of tartrate sodium and triethanolamine as representative ligands on the retention of mercury and cadmium are presented at different pH values. One can see that the retention of mercury is not influenced by any of the investigated competing complexing agents. R values observed in the different solutions are almost equal. In contrast to mercury, R values of cadmium are significantly reduced in the presence of triethanolamine at pH range 5–7 or tartrate sodium at pH range 5.3–6.5. Nevertheless, in solutions with pH 7, the PAASS can successfully compete with the both ligands for cadmium and a high R value is obtained. This means that the influence of competitive complexing agents can be eliminated by increasing pH to a certain value.
Fig. 9. Effect of competitive complexing agents on mercury rejection coefficient at different pH values (MWCO, 6 kDa; PAASS concentration, 500 mg L−1 ; mercury concentration, 500 mg L−1 ; competitive complexing agents concentration, 5000 mg L−1 ; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
1495
Fig. 10. Effect of competitive complexing agents on cadmium rejection coefficient at different pH values (MWCO, 6 kDa; PAASS concentration, 500 mg L−1 ; cadmium concentration, 16.5 mg L− 1; competitive complexing agents concentration, 5000 mg L−1 ; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
3.3. Concentration test Figs. 11 and 12 show the effects of volume concentration factors on permeate fluxes (J) and metal rejection coefficients. In case of mercury, J value reduces from 6.2 to 5.4 L m−2 h−1 during the concentration process, i.e. the membrane permeability decreases from 27.9 to 24.3 L m−2 h−1 bar−1 , which is close to the deionized water permeability of the new membrane 30.6 L m−2 h−1 bar−1 . Therefore, for the concentration range studied, the effect of the PAASS on flux is not dominant. A similar result has been obtained by Uludag et al. [21] in the study of polymer-enhanced ultrafiltration using polyethyleneimine as a complexing agent. In case of cadmium, the behavior of J value as a function of volume concentration factor is very similar to that of mercury. While analyzing the retention curves, it should be emphasized that R values of both metal ions are very high (approximated to 1) in the process. Mercury concentrations in the retentate and the permeate are 1499.6 and 0.03 mg L−1 , respectively, which corresponds to a volume concentration factor of 15. As for cadmium, metal concentrations in the retentate and the permeate are 99.6 and 0.05 mg L−1 respectively when volume concentration factor reaches 10. Therefore, both metal concentrations in the permeates are lower than the maximum value fixed by the integrated wastewater discharge standard of china.
Fig. 11. Effect of volume concentration factor on membrane permeate flux and mercury rejection coefficient (MWCO, 6 kDa; pH, 5; initial PAASS concentration, 100 mg L−1 ; initial mercury concentration, 100 mg L−1 ; feed volume, 30 L; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
1496
J. Zeng et al. / Journal of Hazardous Materials 161 (2009) 1491–1498
Fig. 12. Effect of volume concentration factor on membrane permeate flux and cadmium rejection coefficient (MWCO, 6 kDa; pH, 6; initial PAASS concentration, 300 mg L−1 ; initial cadmium concentration, 10 mg L−1 ; feed volume, 30 L; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
3.4. Decomplexation test At this step, the decomplexation of metal-polymer complex was carried out by acidification of the retentate to a low pH value. For this reason, the final concentrated solutions (in Section 3.3) were acidified until pH value reached 2.5, and then subjected to an ultrafiltration operation in total recirculation mode. Metal concentrations in the retentate and the permeate were compared, and the change of permeate flux with time and rejection coefficient of polymer (RPAASS ) were measured. The results are presented in Figs. 13 and 14. In case of mercury, the decomplexation of mercury-PAASS complex occurs actually at pH 2.5. Mercury concentration in the permeate is 440 mg L−1 and equal to free mercury concentration in the retentate, whereas the total mercury concentration in the retentate is about 1500 mg L−1 . This means that about 30% of ¨ mercury-PAASS complexes are dissociated. Muslehiddino˘ glu et al. [17] studied the effect of pH values on the retention of mercury at different LRs. They indicated that the decrease of mercury retention starts if pH is reduced below 3. This is because that some of mercury-polymer complexes are destroyed at low pH. In case of cadmium, cadmium concentrations in the retentate and the permeate are 99.6 and 93.1 mg L−1 , respectively, which makes that 93.5% of cadmium-PAASS complexes are dissociated. In addition, it is clearly observed that mercury or cadmium concentrations in the retentates are always greater than those in the permeates. This
Fig. 13. Effect of time on membrane permeate flux, PAASS rejection coefficient, mercury concentrations in the retentate and the permeate, respectively, (MWCO, 6 kDa; pH, 2.5; PAASS concentration, 1500 mg L−1 ; initial mercury concentration, 1499.6 mg L−1 ; feed volume, 2 L; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
Fig. 14. Effect of time on membrane permeate flux, PAASS rejection coefficient, cadmium concentrations in the retentate and the permeate, respectively, (MWCO, 6 kDa; pH, 2.5; PAASS concentration, 3000 mg L−1 ; initial cadmium concentration, 99.6 mg L−1 ; feed volume, 2 L; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
indicates that some of mercury or cadmium ions are still trapped by the polymer. Permeate fluxes decline with operating time, and it does not change again after 30 min. It has been also observed that permeate flux is rather lower than that measured at pH 5 or 6 (in Section 3.3). This may be explained by the formation of a polymer cake layer on the membrane surface, following a change of the conformation of the polymer in strong acidic media [22]. The membrane with MWCO 6 kDa can retain the PAASS. Thus, the polymer does not have a leakage during the process of decomplexation, and may be recycled completely. 3.5. Diafiltration test In order to extract the maximum amount of metal ions and make PAASS pure as possible as, diafiltration experiments were carried out after the decomplexation. Two liters of the retentate (in Section 3.4) was treated by eight successive additions of 1 L of pH 2.5 nitric acid solutions. The changes of metal concentrations of the retentates and permeates are shown in Figs. 15 and 16. To describe the diafiltration quantitatively, the metal extraction ratio (MER) is defined as follow:
MER =
1−
[M]r,d [M]f,d
100%
(6)
Fig. 15. Effect of diafiltration volume on mercury concentrations in the retentate and the permeate, respectively, (MWCO, 6 kDa; pH, 2.5; PAASS concentration, 1500 mg L−1 ; initial mercury concentration, 1499.6 mg L−1 ; feed volume, 2 L; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
J. Zeng et al. / Journal of Hazardous Materials 161 (2009) 1491–1498
Fig. 16. Effect of diafiltration volume on cadmium concentrations in the retentate and the permeate, respectively, (MWCO, 6 kDa; pH, 2.5; PAASS concentration, 3000 mg L−1 ; initial cadmium concentration, 99.6 mg L−1 ; feed volume, 2 L; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
where [M]r,d is metal concentration of the retentate at a certain diafiltration volume, [M]f,d is initial metal concentration in the retentate. In case of mercury, it is observed that, if a diafiltration volume is equal to four times than that of the retentate, i.e. 8 L, mercury concentration in the retentate decreases from the beginning of 1499.6 mg L−1 to the end of 810.4 mg L−1 , which corresponds to mercury MER of 46%. In case of cadmium, after 8 L of the diafiltration volume, cadmium concentration in the retentate reduces from 99.6 to 3.87 mg L−1 , which makes that cadmium MER reaches as high as 96.1%. Figs. 15 and 16 also show that, when the diafiltration volume reaches 8 L, mercury concentration in the retentate is much higher than that in the permeate, whereas cadmium concentration of the retentate is almost equal to that of the permeate. Therefore, cadmium can be extracted effectively and the purification of polymer is acceptably satisfactory, but for mercury case it is the contrary. 3.6. Selective separation of binary metal ions To investigate the selective separation of mercury and cadmium ions by the complexation–ultrafiltration process, experiments were carried out at pH 5 using a binary solution in which each metal ion
1497
Fig. 18. Effect of volume concentration factor on rejection coefficient during the selective separation of mercury and cadmium (experimental conditions are the same as those in Fig. 17).
concentration was the same. The feed volume was 30 L, the concentrations of mercury, cadmium and PAASS were 30, 30 and 40 mg L−1 , respectively, and there were not the added salts and low-molecular competitive complexing agents in the solution. After the metal concentration in the retentate is measured, metal concentration ratio (MCR) is calculated by using the following expression: MCR =
[M] [M]
r, s
(7)
f, s
where [M]r,s is metal concentration of the retentate at a certain volume concentration factor, [M]f,s is initial metal concentration in the feed. Fig. 17 shows the effect of volume concentration factor on MCR value. There is a linear growth of MCR for mercury in the process of concentration. When volume concentration factor reaches 15, mercury MCR is 14.83 and mercury concentration in the retentate is 444.9 mg L−1 . In contrast to mercury, cadmium concentration in the retentate grows very slowly during this process. Under the experimental conditions, final cadmium MCR is 1.18 and cadmium concentration of the retentate is only 35.4 mg L−1 . This means that mercury is retained efficiently by the ultrafiltration membrane while almost all cadmium passes through it. Fig. 18 shows the effect of volume concentration factor on the retention of both metal ions. It is cleanly seen that R value of mercury is very high (>0.99), while that of cadmium is fairly low (about 0.1). Practically, at pH 5, PAASS forms very stable complexes with mercury but less stable with cadmium. R values of both metal ions correspond to the accumulation of their macromolecular complexes in the retentate. 4. Conclusions
Fig. 17. Effect of volume concentration factor on the metal concentration ratio during the selective separation of mercury and cadmium (MWCO, 6 kDa; initial PAASS concentration, 40 mg L− 1; initial mercury concentrations, 30 mg L− 1; initial cadmium concentrations, 30 mg L− 1; pH, 5; temperature, 25 ◦ C; transmembrane pressure, 22.2 kPa).
The complexation–ultrafiltration process is viable and suitable to remove mercury and cadmium ions from aqueous solutions with the help of water-soluble polymeric ligand PAASS. At pH 5.5, the kinetics of complexation reactions of PAASS with the metal ions were investigated in the present of a large excess PAASS. It takes 25 and 50 min for mercury and cadmium to get the complexation equilibrium, respectively, and the reaction kinetics can be described by a pseudo-first-order equation. Effects of various operating parameters on metal rejection coefficients were investigated in detail. The binding capacities for PAASS to both metal ions are 1.0 g mercury/g PAASS and 0.033 g cadmium/g PAASS. Compared to the mercury case, the cadmium complexation is more sensitive with changing pH. The influence of the added salt is great at low pH, whereas the influence is
1498
J. Zeng et al. / Journal of Hazardous Materials 161 (2009) 1491–1498
weak at high pH. Low molecular competitive complexing agents do not affect the complexation behavior of PAASS with mercury, but reduce the retention of cadmium. In the process of the concentration, permeate fluxes decline insignificantly and rejection coefficients of both metals are very high. At the decomplexation stage, the decomplexation ratio of mercury-PAASS complex is only about 30%, whereas that of cadmium-PAASS complex reaches as high as 93.5%. In the diafiltration process, cadmium in the retentate can be extracted effectively and a purified PAASS is obtained, but for mercury case it is the contrary. At a special pH value, the selective separation of mercury and cadmium ions can be achieved by controlling the concentrations of metal ions and PAASS. Almost all cadmium is in the permeate, while mercury is retained in the retentate. Acknowledgements This research is supported by a Project Supported by Hunan Provincial Natural Science Foundation of China (no. 05JJ40020). Also, the authors acknowledge the financial supports from a Project Supported by Scientific Research Fund of Hunan Provincial Education Department (no. 05C170). References [1] Y.F. Zhang, Zh.L. Xu, Study on the treatment of industrial wastewater containing Pb2+ ion using a coupling process of polymer complexation–ultrafiltration, Sep. Sci. Technol. 38 (2003) 1585–1596. [2] P. Baticle, C. Kiefer, N. Lakhchaf, O. Leclerc, M. Persin, J. Sarrazin, Treatment of nickel containing industrial effluents with a hybrid process comprising of polymer complexation–ultrafiltration – electrolysis, Sep. Purif. Technol. 18 (2000) 195–207. [3] C.R. Tavares, M. Vieira, J.C.C. Petrus, E.C. Bortoletto, F. Ceravollo, Ultrafiltration/complexation process for metal removal from pulp and paper industry wastewater, Desalination 144 (2002) 261–265. [4] A. Kryvoruchko, L. Yurlova, B. Kornilovich, Purification of water containing heavy metals by chelating-enhanced ultrafiltration, Desalination 144 (2002) 243–248.
[5] J. Barron-Zambrano, S. Laborie, P. Viers, M. Rakib, G. Durand, Mercury removal from aqueous solutions by complexation–ultrafiltration, Desalination 144 (2002) 201–206. [6] C. Dilek, H.O. Ozbelge, N. Bicak, L. Yilmaz, Removal of boron from aqueous solutions by continuous polymer-enhanced ultrafiltration with polyvinyl alcohol, Sep. Sci. Technol. 37 (2002) 1257–1271. [7] R. Molinari, S. Gallo, P. Argurio, Metal ions removal from wastewater or washing water from contaminated soil by ultrafiltration–complexation, Water Res. 38 (2004) 593–600. [8] R.S. Juang, C.H. Chiou, Ultrafiltration rejection of dissolved ions using various weakly basic water-soluble polymers, J. Membr. Sci. 177 (2000) 207–214. [9] S. Petrov, V. Nenov, Removal and recovery of cooper from wastewater by a complexation–ultrafiltration process, Desalination 162 (2004) 201–209. [10] M. Bodzek, I. Korus, K. Loska, Application of the hybrid complexation–ultrafiltration process for removal of metal ions from galvanic wastewater, Desalination 121 (1999) 117–121. [11] S. Verbych, M. Bryk, A. Alpatova, G. Chornokur, Ground water treatment by enhanced ultrafiltration, Desalination 179 (2005) 237–244. [12] J. Bohdziewicz, Removal of chromium ions (VI) from underground water in the hybrid complexation–ultrafiltration process, Desalination 129 (2000) 227–235. [13] A. Kryvoruchko, L. Yurlova, I. Atamanenko, B. Kornilovich, Ultrafiltration removal of U (VI) from contaminated water, Desalination 162 (2004) 229–236. [14] K. Trivunac, S. Stevanovic, Effects of operating parameters on efficiency of cadmium and zinc removal by the complexation–filtration process, Desalination 198 (2006) 282–287. [15] S. Mimoune, F. Amrani, Experimental study of metal ions removal from aqueous solutions by complexation–ultrafiltration, J. Membr. Sci. 298 (2007) 92–98. [16] I. Korus, M. Bodzek, K. Loska, Removal of zinc and nickel ions from aqueous solutions by mean of the hybrid complexation–ultrafiltration process, Sep. Purif. Technol. 17 (1999) 111–116. ¨ Ozbelge, ¨ ¨ [17] J. Muslehiddino˘ glu, Y. Uluda˘g, H.O. L. Yilmaz, Effect of operating parameters on selective separation of heavy metals from binary mixture via polymer enhanced ultrafiltration, J. Membr. Sci. 140 (1998) 251–266. [18] K. Trivunac, S. Stevanovic, Removal of heavy metal ions from water by complexation-assisted ultrafiltration, Chemosphere 64 (2006) 486–491. [19] A. Tabatabai, J.F. Scamehorn, S.D. Christian, Water softening using polyelectrolyte-enhanced ultrafiltration, Sep. Sci. Technol. 30 (1995) 211– 224. [20] R.S. Juang, M.N. Chen, Retention of copper-EDTA chelates from dilute aqueous solution by a polyelectrolyte-enhanced ultrafiltration process, J. Membr. Sci. 119 (1996) 25–37. ¨ Ozbelge, ¨ [21] Y. Uludag, H.O. L. Yilmaz, Removal of mercury from aqueous solutions via polymer enhanced ultrafiltration, J. Membr. Sci. 129 (1997) 93–99. ´ ´ [22] P. Canizares, A.D. Lucas, A. Perez, R. Camarillo, Effect of polymer nature and hydrodynamic conditions on a process of polymer enhanced ultrafiltration, J. Membr. Sci. 253 (2005) 149–163.