Chemical reduction was firstly employed to treat synthetic wastewaters of various compositions prepared to simulate the retentate stream of polyelectrolyte enhanced ultrafiltration (PEUF). With fixed Cu:polyethylenimine (PEI) monomer:dithionite molar ratio, increasing copper concentration increases copper removal efficiency. Under fixed Cu:dithionite molar ratio and fixed Cu concentration, increasing PEI monomer:copper molar ratio decreases copper removal efficiency. The formation of nano-sized copper particles, which readily pass through 0.45 μm filter used for sample pretreatment before residual copper analysis, might be the reason behind the decreasing copper removal efficiency observed. Particle size analysis shows that the size of copper particles, which are formed through reduction reaction, increases with decreasing pH value and increasing reaction time. As ultrafiltration is capable of removing these nano-sized particles, integration of chemical reduction and PEUF is proposed to simultaneously achieve regeneration of polyelectrolyte and recovery of copper in one process. Results show that the proposed process could achieve almost complete copper removal without being affected by reaction pH.
Metal removal by low-pressure membrane filtration with the addition of soluble macromolecular chelating agents, denoted as polyelectrolyte enhanced ultrafiltration (PEUF), is based on complexation of metal ions with natural or synthetic polyelectrolytes, followed by rejection of the metal–polyelectrolyte complexes by ultrafiltration (Camarillo et al. 2010; Jiao et al. 2013). PEUF becomes attractive due to its high membrane permeability along with high efficacy in metal removal. However, the retentate of PEUF containing both metal and polyelectrolyte needs further treatment.
Several regeneration methods, such as acidification followed by membrane separation and electrochemical regeneration, were proposed to recover metal and to regenerate polyelectrolytes from PEUF retentate. With acidification/membrane separation (Juang & Chiou 2000; Camarillo et al. 2010), pH values as low as 2 are needed to breakdown metal–polyelectrolyte complexes, and an additional membrane is required to separate metal ions from polyelectrolyte. Acidification/membrane separation is affected by severe membrane fouling (Jellouli Ennigrou et al. 2009). With electrochemical regeneration, other than an acidic pH value (Jiao et al. 2013), it is necessary to support electrolytes to obtain efficient electrochemical regeneration efficiency. However, the process is still plagued by low current efficiency (Llanos et al. 2009), and vigorous mixing is needed to facilitate mass transfer of metal ions between bulk solution and electrode surface (Llanos et al. 2009; Jiao et al. 2013).
Chemical reduction has been frequently employed for production of metallic nanoparticles (Hashemipour et al. 2011; Chang et al. 2013). Polymers such as PVA (polyvinyl alcohol), PVP (polyvinylpyrrolidone), and PEI (polyethylenimine) are frequently used as coordination agents to control the size of particles produced (Singh et al. 2010; Wang et al. 2014). It has been shown that the higher the concentration of polymer added, the smaller the size of particles produced (Wang et al. 2014). In view of the similarity of the compositions in PEUF retentate and solutions prepared for making metallic nano-sized particles, chemical reduction is used in this study to treat PEUF retentate for simultaneous removal and recovery of metal (copper), and regeneration of polyelectrolyte. PEI-bound copper ions could be reduced into metallic copper particles, resulting in freeing the binding sites in PEI, i.e., regeneration of PEI.
Depending on the PEUF operation scheme, concentrations of Cu and PEI, and their molar ratios in the retentate, might vary. For example, with a fixed molar ratio of Cu and PEI in the feed water and an assumption of copper and PEI being effectively retained by the UF membrane, concentrations of copper and PEI will increase simultaneously in the retentate while the Cu:PEI molar ratio is the same as that in the feed solution. Therefore, the objective of this study is to investigate the effects of Cu and PEI concentrations and their ratio on copper reduction efficiency. It is well known that dithionite decomposes rapidly in the acidic solution via disproportional reaction (Geoffroy & Demopoulos 2009), and the binding of metal ions with PEI is affected profoundly by pH (Li et al. 2008). Thus, the effect of pH on copper removal efficiency by chemical reduction is explored. Finally, as ultrafiltration is capable of removing nano-sized particles, chemical reduction is integrated into PEUF to achieve simultaneous regeneration of polyelectrolyte and recovery of copper. In the proposed process, the reduction process could act as an insurance policy by reducing free copper ions to metallic copper particles and preventing these ions leaking through the membrane. The proposed process could potentially achieve excellent treatment efficiency, and the efficient utilization of reductant and polyelectrolyte. Thus, copper removal efficiency by chemical reduction and PEUF integrated process is also investigated.
Chemicals and materials
All chemicals were of reagent grade. Stock copper solution of 100 mmol/L was prepared using copper sulfate pentahydrate (Yakuri Pure Chemicals Co. Ltd, Singapore, Malaysia). Polyelectrolyte stock solution of 15 mmol/L was prepared by diluting PEI (50% w/v in water, number average molecular mass of 60 kDa, Sigma-Aldrich, St Louis, MO, USA) in DI (deionized water). PEI is a branched polymer. Based on the molecular weight (MW) and the structure of the alternating unit, there are around 412 monomer units in each PEI molecule (Li et al. 2008). Sodium dithionite (Na2S2O4, Alfa Aesar) was used as the reducing agent. Due to instability of dithionite in solution (reaction with molecular oxygen), chemical reduction reaction was started by adding sodium dithionite in powder form or by adding sodium dithionite solution fresh prepared before each experiment. Sodium hydroxide (1 mol/L) and nitric acid (1 mol/L) were used for pH adjustment.
All experiments were repeated in triplicate with the average and one standard deviation from the mean reported. For investigation of the effect of initial copper concentration on copper reduction efficiency, solutions with copper concentration ranging from 0.1–5.0 mmol/L and fixed Cu:PEI monomer molar ratio of 1:1.5 were prepared by diluting copper and PEI stock solutions. After adjustment of solution pH to 3.5 under magnetic stirrer (100 rpm) and N2 gas purging, reaction was started with the addition of dithionite (Cu:dithionite molar ratio of 1:3). After 30 min, the reaction was terminated with a drop of hydrogen peroxide (H2O2, Sigma-Aldrich, St Louis, MO, USA) to oxidize remaining dithionite. The sample for residual copper analysis was filtered with 0.45 μm mixed cellulose esters (MCE) filter (Advantec Co. Ltd, Ehime, Japan), and acidified with a drop of concentrated HNO3.
To determine the effect of PEI concentration on copper removal efficiency, PEI monomer:Cu molar ratios of 1.5:1, 3:1, and 6:1 were prepared with initial copper concentration of 5 mmol/L. Reaction was initiated by dosing dithionite with Cu:dithionite molar ratio of 1:3 at pH 6.0 under magnetic stirrer (100 rpm) and N2 gas purging. After 60 min, reaction was terminated and samples for copper analysis were filtered (0.45 μm filter) and acidified.
Effects of pH (3.5–8.5) and reaction time (10–60 min) on copper removal efficiency were studied with Cu:PEI monomer:dithionite molar ratio of 1:1.5:3, and initial copper concentration of 5 mmol/L. Samples for residual copper analysis were taken, quenched (a drop of H2O2), filtered (0.45 μm), and acidified at various reaction times. Particle size analysis was also conducted for samples taken from experiments with pH values of 3.5 and 8.5. The values of pH chosen are based on the fact that copper/PEI complexes will not form at pH of 3.5 (Juang & Chiou 2000; Li et al. 2008), allowing the effect of forming copper complexes on particle size to be investigated.
Finally, the effect of pH (3.5 to 8.5) on copper removal efficiency of the process combining chemical reduction and PEUF was studied. Experiments were conducted using a stirred ultrafiltration cell (Model 8200, Merck Millipore, Billerica, MA, USA). Polyethersulfone UF membrane with MW cutoff of 10 kDa (courtesy of Hydranautics, Oceanside, CA, USA) was cut into a 62-mm diameter circle and assembled into the stirred cell. A 100-mL solution with initial copper concentration of 5 mmol/L and Cu:PEI monomer molar ratio of 1:1.5 was prepared and placed in the cell. After solution pH was adjusted, reaction was started by adding dithionite with Cu:dithionite molar ratio of 1:3. With the stirred cell pressurized at fixed pressure of 2 bars (N2 gas cylinder), permeate samples (5 mL) were collected at reaction time of 10 and 60 min. After a drop of H2O2 was added, samples were acidified with a drop of concentrated nitric acid for copper analysis.
Transmission electron microscope (TEM) images were obtained by a field emission electron microscope (JEM-2100F, JEOL, Tokyo, Japan) operated at 200 kV. TEM samples were prepared by placing a drop of sample on a silicon wafer covered by carbon-coated copper grid, followed by drying under vacuum at ambient temperature. A nanoparticle analyzer (SZ-100, Horiba, Kyoto, Japan) was used for particle size analysis.
RESULTS AND DISCUSSION
Effect of copper concentration
Solutions with fixed Cu:PEI monomer molar ratio of 1:1.5 were prepared to simulate retentate generated from PEUF process operated with fixed molar ratio of Cu and PEI in the feed water. With the assumption of copper and PEI being effectively retained by the membrane, concentrations of copper and PEI will increase simultaneously in the retentate while the Cu:PEI molar ratio remains the same as that in the feed solution. With initial copper concentration varying from 0.078 to 5.0 mmol/L, Figure 1 shows that the higher the initial copper concentration, the higher the removal efficiency of copper by reduction reaction, revealing the benefit of employing the PEUF process to concentrate the copper concentration followed by chemical reduction. The low reduction efficiency at low initial copper concentration, i.e., low dithionite concentration, is due to domination of side-reactions such as oxidation of dithionite by oxygen and rapid disproportional reaction of dithionite under acidic conditions (Rinker et al. 1960) over copper reduction reaction.
Effect of PEI concentration
Investigation of PEUF for cadmium removal using PEI, Li et al. (2008) reported a stoichiometric complexation ratio between PEI and cadmium around 1.55 mole of monomer in PEI for one mole of cadmium removed. Kuncoro et al. (2005) investigated binding capacity of PEI for Hg(II), reporting 2.08 monomer units of PEI per Hg(II) ion removed. To achieve efficient metal removal, PEI with a concentration higher than the stoichiometric values shown above might be added. Thus, retentate with various PEI monomer to copper ratios is expected to be encountered. In this study, the effect of PEI monomer to copper molar ratios of 1.5:1, 3:1, and 6:1 on copper reduction efficiency was investigated with an initial copper concentration of 5 mmol/L. As indicated in Figure 2, increasing PEI to copper molar ratio decreases copper removal efficiency with an average of 83% for a molar ratio of 1.5:1 and merely around 10% for ratios of 3:1 and 6:1.
Several authors (Song et al. 2009; Chang et al. 2013; Tan & Cheong 2013; Sierra-Ávila et al. 2014) have stated the importance of capping agent or stabilizers in control of the size of particles produced during production of metallic nanoparticles by chemical reduction. In this study, chemical reduction is used to treat the retentate of PEUF, and PEI in the retentate might play the role of stabilizer, which might cause the formation of nano-sized metallic copper particles. Since samples for residual copper analysis were filtered with 0.45 μm filter and acidified with concentrated HNO3 in these experiments, the low copper removal efficiency observed is probably the result of these nano-sized particles passing through 0.45 μm filter and being counted on the residual copper analysis. As indicated in Figure 3, TEM images reveal that the size of particles produced decreases with increasing PEI monomer:Cu molar ratios. For particles generated under molar ratio of 1.5:1, reduced particles are aggregated and particles size is around 4 μm. With molar ratio of 6:1, the particles are much smaller with particle size of around 60 nm. Although removal efficiency of copper is not effective due to formation of nanoparticles, it is expected that the removal efficiency will be improved if UF membrane is used, i.e., integration of chemical reduction and PEUF process in one process.
Effect of pH and reaction time
It is well known that dithionite decomposes rapidly in acidic conditions in solution via disproportional reaction (Geoffroy & Demopoulos 2009), therefore, it is expected that pH will affect copper reduction efficiency. As indicated in Figure 4, removal efficiency by chemical reduction decreases with increasing pH values and increases with reaction time. The highest copper reduction efficiency observed at acidic pH value of 3.5 indicates that reduction reaction is the dominant reaction, and disproportional reaction of dithionite has a very small effect on copper reduction efficiency.
Differences in removal efficiency observed might be caused by the extent of PEI and copper complexation in different pH values. It has been shown that the binding of metal ions with PEI occurs via ion exchange reactions (Li et al. 2008). In acidic pH, hydrogen ions compete strongly with metal ions for an available binding site of PEI, resulting in more free copper ions than PEI-bound copper in the solution. It might be that reduction reaction kinetics between PEI-bound copper ions with dithionite is slower and/or nano-sized particles are formed due to the formation of PEI/copper complexes, resulting in decreasing copper removal efficiency with increasing pH values.
Particle size analysis was then employed to elucidate the change of particle size as a function of reaction time for pH values of 3.5 and 8.5. As shown in Figure 5(a) and (b), particles grow with reaction time and the size of particles is bigger for those produced at pH of 3.5 than those at pH of 8.5. If the low removal efficiency observed at higher pH values is due to nano-sized particles produced, incorporation of chemical reduction with PEUF process will solve the problem.
Copper removal efficiency by chemical reduction and PEUF integrated process
As mentioned above, if the formation of nano-sized particles is the only culprit responsible for the low removal efficiency observed, copper removal efficiency will increase dramatically with the incorporation of UF membrane. Copper removal by chemical reduction and PEUF integrated process was conducted to test the hypothesis. Figure 6 shows that copper removal efficiency still slightly decreases with increasing pH values but all are higher than 90% at reaction time of 10 min and almost 100% at reaction time of 60 min. In view of the removal efficiency, which ranges from 0 to 30% for pH values of higher than 5.5 without UF membrane being incorporated at reaction time of 10 min (see Figure 4), one could conclude that formation of nano-sized particles due to formation of PEI/copper complexes is the main reason behind the low copper removal efficiency observed above, and the effect due to slow reduction reaction kinetics between PEI-bound copper ions with dithionite might not be so important.
Synthetic wastewater prepared to simulate retentate stream from PEUF was treated with chemical reduction to recover copper in metallic form. Increasing copper concentration increases chemical reduction efficiency, revealing the benefit of employing PEUF process to concentrate copper first and then subsequently treating by chemical reduction. Increasing PEI monomer:copper ratio decreases copper removal efficiency due to formation of nano-sized copper particles, which is related to the extent of PEI/copper complexes formed. Increasing pH values increases the extent of PEI and copper complexation, resulting in decreasing removal efficiency of copper by chemical reduction. The pitfall of forming nano-sized particles resulting in low copper removal efficiency is solved by incorporating chemical reduction with PEUF in one process in which almost 100% copper removal efficiency could be achieved.
The study was supported by the Ministry of Science and Technology of Taiwan under grant number 103-2221-E-032-001.