Article pubs.acs.org/est
New Strategy To Enhance Phosphate Removal from Water by Hydrous Manganese Oxide Bingcai Pan,* Feichao Han, Guangze Nie, Bing Wu, Kai He, and Lv Lu State Key Laboratory of Pollution Control and Resource Reuse, School of the Environment, Nanjing University, Nanjing, Jiangsu 210023, PR China S Supporting Information *
ABSTRACT: Hydrous manganese oxide (HMO) is generally negatively charged at circumneutral pH and cannot effectively remove anionic pollutants such as phosphate. Here we proposed a new strategy to enhance HMO-mediated phosphate removal by immobilizing nano-HMO within a polystyrene anion exchanger (NS). The resultant nanocomposite HMO@NS exhibited substantially enhanced phosphate removal in the presence of sulfate, chloride, and nitrate at greater levels. This is mainly attributed to the pHpzc shift from 6.2 for the bulky HMO to 10.5 for the capsulated HMO nanoparticles, where HMO nanoparticles are positively charged at neutral pH. The ammonium groups of NS also favor phosphate adsorption through the Donnan effect. Cyclic column adsorption experiment indicated that the fresh HMO@NS could treat 460 bed volumes (BV) of a synthetic influent (from the initial concentration of 2 mg P[PO43−]/L to 0.5 mg P[PO43−]/L), while only 80 BV for NS. After the first time of regeneration by NaOH-NaCl solution, the capacity of HMO@NS was lowered to ∼300 BV and then kept constant for the subsequent 5 runs, implying the presence of both the reversible and irreversible adsorption sites of nano-HMO. Additional column adsorption feeding with a real bioeffluent further validated great potential of HMO@NS in advanced wastewater treatment. This study may provide an alternative approach to expand the usability of other metal oxides in water treatment.
■
be directly employed in flow-through systems because of the excessive pressure drop and poor mechanical strength. The size of nanoparticles has been implicated as a major influence on the reactivity and adsorptivity of metal (hydro)oxides.21 Abbas et al.22 predicted a shift in the pHpzc of metal oxides occurring with the decrease of size particularly when it is less than 5 nm. Madden et al.23 observed a significant shift in pH edge for the 7 nm hematite as compared to the 25 and 88 nm samples. Vayssieres24 found that pHpzc of maghemite increased with the decrease of size when comparing the pHtitration diagrams for three particle sizes, 12, 7.5, and 3.5 nm. All these studies indicate that surface charge of metal oxides is size-dependent at nanoscale. Nevertheless, the size-dependent properties of hydrated manganese oxide (HMO) have rarely been reported. The objective of the current study is to fabricate a nanosized HMO oriented composite material to overcome the inapplicability of bulk HMO in practical application and to explore the possibility of the resultant composite for preferable removal of phosphate with simultaneously considering the potentially sizedependent property of nanosized oxides. To achieve the goal, we encapsulated HMO nanoparticles inside a strongly basic anion exchanger (NS) and then obtained a new composite adsorbent HMO@NS. The change of pHpzc value of HMO
INTRODUCTION Phosphate is an essential nutrient in aquatic ecosystems, but excessive phosphate in water instead leads to eutrophication and other environmental problems. Currently, various techniques have been developed to remove phosphate from water, such as chemical precipitation, biological process, and adsorption.1−4 Chemical precipitation and biological process are widely used in practical application, but they are difficult to effectively remove phosphate at trace levels.1,2 Moreover, a mass of sludge generated from both processes needs further disposal. Comparatively, adsorption has attracted increasing interest in phosphate removal from wastewater.3,4 Also, adsorption has been considered as an effective approach to recycle phosphate resources from the effluents.5 Over the past decade, various adsorbents have been developed for phosphate removal, including slag,6 lanthanum-modified materials,7,8 and metal (hydro)oxides.9−11 Manganese oxides, a transition-metal oxides with high surface area and strong surface reactivity,12 have been widely applied to remove heavy metals from water.13−15 However, manganese oxides cannot be directly used to adsorb anions (such as chloride and phosphate) because of their specific surface properties. Point of zero charge (pHpzc) of most manganese oxides is 2−3,16,17 and some hydrated manganese oxides have their pHpzc values of 4−6.18,19 Consequently, most of the manganese oxides are negatively charged near neutral pH and thus cannot effectively capture anionic species due to the electrostatic repulsion.20 Furthermore, manganese oxides generally are present as fine or ultrafine particles and cannot © 2014 American Chemical Society
Received: Revised: Accepted: Published: 5101
September 27, 2013 April 11, 2014 April 15, 2014 April 15, 2014 dx.doi.org/10.1021/es5004044 | Environ. Sci. Technol. 2014, 48, 5101−5107
Environmental Science & Technology
Article
Figure 1. Characterization of the as-obtained nanocomposite HMO@NS: (a) morphology; (b) TEM image; (c) radical distribution of elemental manganese of the cross section of HMO@NS; (d) XRD spectra.
were prepared according to the methods reported by Subramanian et al.25 Potentiometric Titrations. The acid−base surface chemistry of HMO bulk phase and the HMO-oriented composites was investigated through potentiometric titration, using 0.01 M NaNO3 solution as the background electrolyte. An automatic titration system (T50, Mettler Toledo) with a combined glass electrode (DGi115-SC) was employed. All titrations were performed in 100 mL vessels at 298 ± 0.2 K. Prior to titration, the desired amount of solid samples was suspended in 40 mL of NaNO3 solution and purged with high purity N2 gas for 2−4 h. The pH values were initially controlled at 3.0 with 0.20 M HNO3 solution. After 1 h of equilibrium, the suspensions were slowly back-titrated at a constant increment of pH with 0.03 M NaOH until pH = 11.5, where the addition of NaOH (0.005− 0.1 mL) was automatically adjusted to obtain desired pH. Each step was allowed to stabilize (the pH drift was less than 0.005 pH unit per minute). To avoid the disturbance of carbonate species, all solutions were prepared using preboiled deionized water and then kept under a nitrogen atmosphere, and high purity N2 gas was bubbled into solutions throughout the titration. Batch Adsorption. Batch adsorption experiments were carried out in 250 mL glass bottles. To start the experiment, 50 mg of adsorbent was introduced to 100 mL of solution with known solute concentration. KCl, K2SO4, or KNO3 were added to examine their effect on phosphate uptake at different concentrations. The reactors were then transferred to a G-25 model incubator shaker with thermostat (New Brunswick Scientific Co. Inc.) under 160 rpm for 24 h to ensure the adsorption equilibrium at the desired temperature. The solution pH values were adjusted with 0.1 M HCl and 0.1 M NaOH solution throughout the experiments. One milliliter of supernatant was sampled from the reactors at various time intervals to determine the adsorption kinetics. The amount of solute loaded on adsorbent particles was calculated by conducting a mass balance on the solute before and after test. All the batch runs were performed in duplicate for data analysis.
upon immobilization was investigated, and phosphate removal by HMO@NS was systematically examined as a function of solution pH and the concentration of the coexisting anions. Column adsorption experiments were also carried out to elucidate the applicability of HMO@NS in phosphate removal from synthetic and industrial biotreated wastewater effluent.
■
MATERIALS AND METHODS Materials. Chemicals used in this study were of analytical reagent grade and purchased from Shanghai Reagent Station (Shanghai, China). The phosphate stock solution (1.0 g P[PO43−]/L) was prepared by dissolving KH2PO4 into deionized water. Three polystyrene hosts of spherical beads, NS, a strongly basic anion exchanger with 3.75 mmol/g quaternary ammonium groups, SS, a strongly acidic cation exchanger with 3.8 mmol/g sulfonate groups, and CS, the neutral precursor to NS with chloromethyl groups, were provided by Zhengguang Electrical Resin Co. Ltd. (Hangzhou, China). The relationship between the three hosts is depicted in Scheme S1 (in the Supporting Information). All the polymeric hosts were subjected to extraction with alcohol in a Soxhlet extraction apparatus to remove possible residue impurities. Then they were rinsed with deionized water several times and air-dried at T ≤ 323 K. Spherical beads of the polymeric hosts ranging from 0.7 to 1.0 mm in diameter were sieved for use. Preparation of HMO-Based Materials. HMO was immobilized into polymeric hosts through the following steps: A mixture of 20 g of NS (or CS, SS) and 360 mL of 0.2 M KMnO4 solution were shaken at an appropriate rate for 12 h at 298 ± 1 K to ensure the sufficient uptake of the anionic MnO4− into the polymeric beads. Then, the resultant beads were filtered and added into 200 mL of 50% (V/V) ethanol/ water and further shaken for 12 h, where the preloaded MnO4− inside the polymer phase was reduced to HMO.24 Finally, the as-obtained nanocomposites (namely HMO@NS, HMO@CS, and HMO@SS respectively) were filtered and rinsed with deionized water until the filtrate reached a pH of ∼7, and then vacuum desiccated at 323 K for 12 h. The bulky HMO particles 5102
dx.doi.org/10.1021/es5004044 | Environ. Sci. Technol. 2014, 48, 5101−5107
Environmental Science & Technology
Article
Fixed-Bed Column Adsorption. Fixed-bed column experiments were conducted in a glass column (30 mm in diameter and 130 mm in length) equipped with a water bath to maintain the temperature constant. Forty milliliters of wet HMO@NS or NS beads was packed within the column. A speed-adjustable peristaltic pump (BT01−40, China) was used to control flow rate. The in situ regeneration of exhausted HMO@NS beads was performed by the binary NaOH-NaCl solution (both 5% in mass) as regenerant. The corresponding superficial liquid velocity (SLV) and the empty bed contact time (EBCT) were described in the corresponding figure captions. Analysis. The concentrations of phosphate were determined by molybdenum blue spectrophotometric method with a UV/vis spectrometer (T6, PGENERAL China). The manganese amount of the composite materials or HMO powder was determined by an atomic absorption spectrometer (TAS990, PGENERAL China) after digesting the sample into a binary solution of perchloric acid and concentrated nitric acid (V1:V2 = 1:1). HMO distribution within the cross section of HMO@NS was performed on a scanning electron microscope (S-3400NII, Hitachi Japan) equipped with an energy-dispersive X-ray analyzer (EX-250, HoribaJapan). HMO dispersion within HMO@NS was observed with transmission electron microscopy (JEM-200CX, Japan). X-ray diffraction analysis was carried out on an X-ray diffractometer (XRD6000, Japan). The zeta potentials of HMO@NS and HMO powder were measured by Malvern Zetasizer3000 HSa. Prior to the experiment, 50 mL of the solid sample was ground and added into 500 mL of 0.010 M NaNO3 solution at different pH. XPS analysis was performed on a PHI 5000 Versa probe system using monochromatic Al Kα radiation (1486.6 eV), and all binding energies (BE) were referenced to the C 1s peak at 284.5 eV.
Figure 3. Normalized adsorption capacity of HMO toward phosphate before and after being loaded into three polystyrene hosts at 298 K and pH 7.0. Each adsorbent dose was 0.070 g Mn/L and the background SO42− was 600 mg/L.
Figure 4. Schematic illustration of phosphate adsorption by HMO before and after being loaded into NS (pH = 7.0).
■
RESULTS AND DISCUSSION Characterization of HMO@NS. The as-obtained nanocomposite HMO@NS remained the original spherical geom-
Figure 5. Effect of pH on phosphate adsorption by three adsorbents in the background of 600 mg/L SO42−at 298 K. Initial phosphate: 5 P (PO43−)/L; solid dosage: 1.0 g/L of HMO@NS; and 0.24 g/L of HMO + 0.76 g/L of NS.
Figure 2. Potentiometric titration curves of HMO@NS, HMO@CS, and the bulky HMO particles.
Similar compositions of the Mn-based adsorbents (Figure S2) as well as their similar preparation processes imply that they have similar chemical structure. The presence of low-valence Mn (II, III) is supposed to result in the lattice defects such as oxygen vacancies26 and produce abundant hydroxyl oxygen,27 which would increase effective sites of HMO@NS. Zeta potentials of the bulky HMO and HMO@NS were illustrated in Figure S3. In general, the zeta potential of HMO decreased as pH increased and the isoelectric point (IEP) is 4.5, which is consistent with the value reported elsewhere.28,29 For
etry of NS, but its color was changed from white to brown. HMO nanoparticles were successfully loaded and uniformly dispersed inside the NS beads with the average particle size of 5.0−7.0 nm (Figure 1 and Figure S1 in the Supporting Information). The X-ray diffraction spectra indicated that HMO loaded into NS was amorphous in nature (Figure 1d). Mn content of HMO@NS and HMO was 14.08% and 57.25%, respectively. Three valence states of Mn, i.e., Mn(IV), Mn(III), and Mn(II), were detected in both HMO@NS and HMO. 5103
dx.doi.org/10.1021/es5004044 | Environ. Sci. Technol. 2014, 48, 5101−5107
Environmental Science & Technology
Article
Figure 6. Effects of ubiquitous anions on the phosphate removal by NS and HMO@NS at 298 K, pH 7.0. Each adsorbent dose was 0.5 g/L with 5 mg P (PO43−)/L of phosphate: (a) NO3−; (b) SO42−; (c) Cl−.
HMO@NS, a slight drop of the zeta potential was observed when increasing pH from 3.0 to 7.0, and a further increase to 12 did not affect its zeta potential significantly. The zeta potential of HMO@NS was higher than 15 mV in the tested pH ranges possibly because of the positively charged ammonium groups that are covalently bound to the host of HMO@NS. To further probe the change of the surface chemistry of HMO upon encapsulating inside polymeric hosts, we carried out potentiometric titration on the bulky HMO, HMO@NS, and HMO@CS and thus obtained the pH-dependent apparent net surface charge (σ0) of the solid samples (Figure 2). Note that σ0 means the net amount of H+ or OH− consumed by the solid samples through protonation/deprotonation or ion exchange during titration. As shown in Figure S4, the neutral CS showed no surface charge since its titration curve overlaps with the pure water. As for NS with positively charged functional groups, it has negligible surface charge at pH 10 due to the ion exchange process of NS with OH−. Thus, it suggests that σ0 of HMO@CS and σ0 of HMO@NS at pH
