Magnetically attracted iron scrap anode based electrocoagulation for phosphate removal

This study shows the effectiveness of a novel electrocoagulation process using magnetically attracted iron scrap anodes for phosphate removal from aqueous solution. The effect of contact time, reaction temperature, dose of iron scrap, initial phosphate concentration, applied voltage, pH, magnetic force, and the species of competing anions on the ef ﬁ ciency of phosphate removal and the reaction products has been investigated. The techniques of XRD, XPS, and VSM were used to characterize the elemental composition and the types of the reaction products in order to clarify the interaction between novel anode and phosphate ions. The removal of phosphate was ﬁ tted by a pseudo ﬁ rst-order reaction kinetic model. The results showed that magnetically attracted iron scrap anodes was electrodissoluted under an applied potential and reacted with phosphate into Fe-hydroxo-phosphate complexes. The work suggested that electrocoagulation using magnetically attracted iron scrap anodes had the potential to become a promising technique for phosphate precipitation. used The effect of different parameters on novel electrocoagulation was investigated. (cid:129) Elemental composition and the type of the reaction product were characterized XRD, XPS, and VSM. (cid:129) Fe-hydroxo-phosphate complexes were formed during iron scrap electrocoagulation. (cid:129) Iron scrap waste was used for high value electrode. of iron scrap, (2) immobilization of iron scraps via magnetic ﬁ eld, (3) effectively avoid passivation compared to the iron plate anode because of higher reaction surface, (4) provid-ing a promising method for reclamation of iron scraps, (5) no additional chemicals for this electrocoagulation, (6) no need to adjust the pH value of the solution. This work provides a promising approach for phosphate removal while this iron scrap anode could also be used for novel electro-Fenton process.


INTRODUCTION
Industrial wastewater and agricultural runoffs with excessive phosphorus directly or indirectly discharge into waterbody in terms of phosphate (PO 4 3À ), hydrogen phosphate (HPO 4 2À ) and dihydrogen phosphate (H 2 PO 4À ) (Fulazzaky et al. 2014). The excess dissolved phosphate is well recognized as a primary source of eutrophication, which leads to algae boom, opacity and oxygen depletion. Furthermore, natural phosphorus has significantly declined both in the quantity and the quality in the past decades (Lei et al. 2017). The potential shortage of phosphorus resources along with the over-discharge of phosphorus-containing wastewater have created an increased awareness of the importance of phosphorus recovery and recycling. Thus, different technologies have been developed to mitigate phosphorus discharge to the environment including biological and chemical treatments (Sun et al. 2021). Struvite formation is regarded as one of the most promising chemical precipitation ways (González et al.). However, it relies on different Mg sources to assist struvite formation, which makes the process less economically attractive because of the low concentration of Mg 2þ in wastewater (Barbosa et al. 2016). Alternatively, a novel process of electrochemical induced CaP precipitation on the cathode has been reported. Unfortunately, the phosphate removal efficiency of the process suffers CaP coverage on the cathode and a saturation of Ca and P (Lei et al. 2017). Electrolysis appears as a viable onsite sanitation solution for phosphate recovery from human waste, however, this strategy requires costly TiO 2 -coated semiconductor anodes and suffers hydroxyapatite coverage on the cathode (Cid et al. 2018;Hou et al. 2021).
Electrocoagulation (EC), as a well-known electrochemical process, has received considerable attention for phosphorus precipitation due to its high feasibility and efficiency (Zeng et al.). Traditionally, zero-valent-iron (ZVI) plates were used as the electrode of electrocoagulation process. The reaction mechanisms of ferrous or ferric ions with phosphate were evaluated using both isotherm models and thermodynamic models (Vasudevan et al. 2008). Recently, a reverse-electric-field/air cathode electrocoagulation reactor was developed for high efficient phosphate removal from domestic wastewater compared to conventional electrocoagulation (Tian et al. 2017). A migration electric-field assisted electrocoagulation system was further developed to increase phosphate removal efficiency from domestic wastewater (Tian et al. 2018).
Unfortunately, electrocoagulation with ZVI plate inevitably suffers the passivation due to the formation of iron (oxyhydr)oxide passive layers on the surface of Fe(0) plates (Veluchamy et al. 2017;Samir Naje et al. 2019), resulting in a decrease in anodic interface potential favoring a larger faradaic current efficiency (Dubrawski et al. 2015). The dissolution and passivation rate of iron were experimentally verified to strongly depend on the electrode potential, ionic strength, ionic type, and the solution pH (Lorenz et al. 2002). Mechanical cleaning (Lakshmanan et al. 2009), the polarity of the electrodes (Timmes et al. 2010), and addition of halide salts (Adamovic et al. 2016) were reportedly the most efficient and reliable methods to clean the scale and fouling on the electrode for electrode maintenance.
ZVI particle has drawn great attention for the remediation of a wide variety of contaminants in groundwater and wastewater because of its high efficiency, low cost and nontoxic characteristics (Li et al. 2018;Jiang et al. 2020). Numerous researches reported on the phosphate removal from water by ZVI focusing on influence of different parameters such as solution composition (Eljamal et al. 2016), initial phosphate concentration ZVI aging (Sleiman et al. 2017), the enhancement of ZVI by different supporters (Khalil et al. 2017), and activation of persulfate by nanosized zero-valent iron (Zhao et al. 2016;Kim et al. 2018). Main focus of ZVI particle related research is how to further enhance the removal efficiency in many applications. This work was motivated by the wide application of ZVI particle in environmental remediation and the disadvantages of ZVI plate electrocoagulation. Therefore, a novel electrocoagulation process using magnetically attracted iron scrap anodes was designed for phosphate removal. This system has several advantages: (1) economies of iron material because of low cost of iron scrap, (2) immobilization of iron scraps via magnetic field, (3) effectively avoid passivation compared to the iron plate anode because of higher reaction surface, (4) providing a promising method for reclamation of iron scraps, (5) no additional chemicals for this electrocoagulation, (6) no need to adjust the pH value of the solution. This work provides a promising approach for phosphate removal while this iron scrap anode could also be used for novel electro-Fenton process.

Analytical techniques
All samples were filtered through 0.22 μm pore diameter syringe filters (polyvinylidenedifluoride, PVDF, 25 mm diameter, Restek Corporation) and analyzed for phosphate using ammonium molybdate spectrophotometric method by continuous flow analyzer (model AA3, SEAL, Germany). Removal rate was calculated based on initial and final concentrations. A pH probe (SevenMulti, Mettler-Toledo International Inc.) and conductometer (Rex Chemical Corp, Shanghai) were used to measure solution pH and conductivities. X-ray diffraction (XRD) patterns were recorded with a Bruker D8 advance diffractometer with monochromatic Cu Kα irradiation in the 2θ angular regions between 10 and 90°. XPS spectra were obtained on a Thermo Fisher Scientific ESCALAB 250Xi XPS spectrometer using Al Kα (1,486.6 eV) as the radiation source. The binding energy was calibrated by means of the peak energy of C 1s at 284.5 eV (Gao et al. 2011), and the elemental compositions were determined from peak area ratios after the correction for the sensitivity factor for each element. A vibrating-sample magnetometer (VSM, USA) is a scientific instrument that measures magnetic properties.
Batch electrocoagulation using magnetically attracted iron scrap Electrocoagulation using magnetically attracted iron scrap anodes were carried out in batch-scale reactor consisted of a 200 mL electrolyte glass tank. The sacrificial anode was 0.2 g iron scraps evenly distributed in a groove (40 mm Â 25 mm Â 2 mm) on the graphite surface (100 mm Â 30 mm Â 5 mm). These iron scraps were magnetically attracted by a magnet (40 mm Â 20 mm Â 5 mm, surface magnetic field strength ∼ 2,004 gauss) outside of the glass tank. The cathode was a 100 mm Â 25 mm Â 1 mm graphite sheet. The interelectrode gap is 10 mm. The applied potential was provided by a PS325DII power supply (0-32 V, 0-5 A, Shenzhen Lodestar precision tools Co., Ltd). The solution pH was varied from acidic to neutral and alkaline (4.0, 5.0, 7.0, 8.0, and 10.0). The initial phosphate concentration ranged from 100 to 2,500 mg·L À1 . All experiments were carried out under constant voltage conditions. A series voltages of 1.2, 5.0, 10.0, 15.0 V were applied for the new type EC. For each sampling event, a volume of suspension was drawn from the reactor at 0, 5, 15, 30, 60, 120, and180 min, and then filtered through a 0.22 μm PVDF membrane for analysis.

RESULTS AND DISCUSSION
Figure 1(a) illustrates the effect of initial phosphate concentration from 100 to 2,500 mg·L À1 (pH values from 5.18 to 4.70) on the phosphate concentration remaining in the solution. It can be obtained that removal rates are 100.00%, 80.05%, 76.13%, 73.41%, and 53.60% for phosphate solutions from 100 mg·L À1 to 2,500 mg·L À1 . Obviously, an increase in concentration at a constant potential of 10 V and 0.2 g iron scrap result in decrease in removal efficiency because of limited iron amount although higher concentration solution has larger current density. The amount of as-formed iron hydroxides was insufficient to coagulate the higher amount of phosphate at higher phosphate concentration.
It is well-known that applied potential dramatically influence the phosphate precipitation efficiency of the iron electrocoagulation. The effect of applied potential on the phosphate removal was shown in Figure 1(b). As expected, the removal efficiency increased significantly with the increase in applied potential within 3 h because of higher coagulant dosage rate, bubble production rate, and size and floc growth rate at larger current density. The highest electrical potential of 15 V produced the highest phosphate precipitation rate of 93.89% and the negligible phosphate removal efficiency of 1.30% occurred in the lowest electrical potential of 1.2 V. When the results were fitted by first-order reaction, the first-order kinetic rate constant decreased from 79.02 min À1 at 15 V to 182.62 min À1 at 5 V. This is ascribed to the fact that at higher voltage the amount of iron oxidized increased, resulting in a greater amount of precipitate for the removal of phosphate.
Many researchers reported that there is an influence of the magnetic field on the rate of various chemical and electrochemical reactions of non-electrochemical dissolution of iron (Bech-Nielsen & Jaskuła 2008). Figure 1(c) is shown that the influence of gauss surface strength of magnet on the phosphate precipitation at applied potential of 10 V. Each magnet has 2004 gauss surface strength. The phosphate rates are 80.28% and 76.13% when five magnets and one magnet are used, respectively. It could be concluded from Figure 1(c) that the strength of magnetic field has slightly impact on the phosphate precipitation. Due to the ferromagnetic property of ZVI, its external the Lorentz force and magnetic gradient force have been proposed to contribute to the enhancement in the reactivity and depassivation of ZVI (Xu et al. 2016). However, the effect of electric field is much greater than that of magnetic field because electrocoagulation is dominated by the applied potential in this case. Figure 1(d)-1(f) depict the effect of various solution temperature from 25 to 45°C and the amount of loaded iron scraps on the electrocoagulation performance. It can be observed that a higher amount of loaded iron scrap results in a greater amount of precipitate for the removal of phosphate due to a larger amount of iron oxidized at each temperature. Moreover, an increase in temperature causes an increase in removal efficiency. This may be attributed to the temperature-induced increase in the mass transferred of Fe 2þ from the anode surface to the solution bulk and the increase of the rate of Fe 2þ hydrolized to Fe(OH) 2 and reacted with PO 4 3À to form ironphosphorus compound. These trends were confirmed by the first-order kinetic rate constants summarized in Table 1. The first-order kinetic rate increases with an increase in solution temperature and the amount of loaded iron scrap.
The electrocoagulation was carried out using magnetically attracted iron scrap under nitrogen aeration. Dissolution of the iron scrap anode leads to iron ions which react with phosphate ions and hydroxo ions. Fe-hydroxophosphate complexes was formed according to the following reactions of xFe 2þ þ PO 4 3À þ (2x À 3)OH À ¼   Figure 2. The peaks at 44.75°and 65.2°represented the characteristic peaks of ZVI correlated to lattice planes of (110) and (200) (Figure 2(a)). The XRD patterns from preciptitate samples showed a typical vivianite phase (Fe 3 (PO 4 ) 2 ·8H 2 O) with a diffraction peak at 13.1°correlated to lattice planes of (020) (Figure 2(b)). However, the XRD reflexes characteristic for Fe(OH) 2 ·nH 2 O could not be identified from bulk preciptitate samples due to its small amount of amorphous phase. The vivianite mineral has been experimentally well studied and can be found in coatings of water pipes, soils, morasses, and sediments. Vivianite has paramagnetic properties as ambient temperature is higher than Neél temperature T N (∼10 K) (Pinto et al. 2014). Vivianite can disintegrate into strongly magnetic magnetite and weakly magnetic hematite upon heating in air through auto-oxidation or by the air when Fe 2þ is oxidized to Fe 3þ . The magnetic properties of the vivianite were investigated using VSM analysis at room temperature, on the magnetic field range from À1,000 to 1,000 mT. Magnetic hysteresis measurements showed that vivianite is of paramagnetic properties (Figure 3(a)). The saturation magnetization (Ms) and remanence (Mr) of vivianite are 0.05 and 0.003 Am 2 ·kg À1 , respectively. The sample has a coercitivity of 4.5 mT and saturation field is 75 mT. Magnetic hysteresis measurements showed that iron scraps contained is of ferrimagnetic material recognized by the opening of the hysteresis loops. The result from magnetic hysteresis measurements confirmed the formation of vivianite during electrocoagulation process by magnetically attracted iron scrap in phosphate solution.
The elemental compositions of ZVI under various experimental conditions and reaction products were investigated by XPS. The broad survey scans (now shown) reveal that the surface of unreacted ZVI is composed mainly   (Figure 4(a)) shows peaks of oxidized iron [Fe(III)] in the binding energies of Fe 2p 1/2 (ca. 724.2 eV) and Fe 2p 3/2 (ca. 710.6 eV), while the peaks at 718.8 eV represents the binding energies of shakeup satellite Fe 2p 3/2 (Xu et al. 2016). Additionally, the small shoulder of Fe 2p 3/2 (ca. 706.6 eV) suggests the presence of elemental iron (Fe 0 ) (Mondal et al. 2014). Thus, our results are in accordance with those reported in literature where the ZVI was shown to have a core-shell structure with an inner Fe 0 and an outer envelope of oxide layer. The Fe 2p 3/2 spectrum indicated the presence of Fe(II) and Fe(III) on the reacted ZVI and precipitate reaction product (Figure 4(b) and 4(c)). No Fe 0 was detected on the reacted ZVI surfaces, indicating that the reacted ZVI surface was completely covered with a layer of reaction product (mainly iron phosphate). According to the peak-fitting results, the ratio of Fe(II)/Fe(III) on the surface of reacted ZVI (1.87) was higher than that on the surface of reaction product (1.05). The Fe 2p core level (Figure 4(b) and 4(c)) shows binding energy peaks of Fe(III) and Fe(II) are 725.2 eV and 711.3 eV, 722.9 eV and 709.3 eV for Fe 2p 1/2 and Fe 2p 3/2 , respectively. The binding energies of Fe 2p 1/2 and Fe 2p 3/2 in Fe(III) shifted to higher values because of formation of Fe-hydroxo-phosphate complexes. The existence of Fe(II) is consistent with the ZVI coagulation process where Fe(II) ions were firstly generated. XPS result show the presence of this protective layer on the ZVI surface due to the fact that the penetration depth on surface of XPS (∼nm) is much lower than that of XRD (∼mm or μm). The P 2p XPS spectra of ZVI (1.2 V), ZVI (10 V), and reaction product were shown in Figure 4(d). No P was detected for iron scrap anodes at an applied potential of 1.2 V because of unmotivated electrocoagulation. Iron scrap anode (10 V) and reaction product show binding energy peaks of P 2p located at 133.5 eV. This result indicated the existence of phosphate complexes on the surface of the ZVI anode and reaction product. Figure 5 shows the effect of competing anions of nitrate, sulfate, and acetate on phosphate removal by magnetically attracted iron scrap electrocoagulation with an applied potential of 10 V, phosphate concentration of 500 mg·L À1 and electrical conductivity of 850 mS·m À1 . The reason why chloride ion has not been compared herein is that it will be oxidized to Cl 2 at an applied potential of 10 V. Figure 5(a)-5(c) represent that the rate of iron electro-coagulation at acidic and anoxic conditions is faster than that of the iron electro-coagulation by neutral and basic condition which is consistent with the result elsewhere (Huang & Zhang 2004). Nitrate, sulfate, and acetate have similar effect on phosphate removal (98-99%) at low pH of 4 and 5. As pH increases to near neutral (pH of 7 and 8), phosphate removal order follows SO 4 2À (99%, 96%) . NO 3 À (96%, 92%) . CH 3 COO À Figure 4 | XPS spectra of (a) ZVI at an applied potential of 1.2 V for 3 h, (b) ZVI at an applied potential of 10 V for 3 h, (c) reaction products collected at an applied potential of 10 V for 3 h, and (d) phosphate of ZVI (1.2 V), ZVI (10 V), and reaction product.

CONCLUSIONS
In conclusion, a novel electrocoagulation process with magnetically attracted iron scrap anode was employed for phosphate removal. This process follows first-order kinetics and the phosphate removal rate increased significantly with the increase in applied potential (93.89% at 15 V, 1.30% at 1.2 V). The maximum and minimum values of removal rate are 100.00% and 53.60% when phosphate concentration are 100 mg·L À1 and 2,500 mg·L À1 . The effect of co-anions on phosphate removal follows the order of SO 4 2À . NO 3 À . CH 3 COO À at pH , 10 and NO 3 À . SO 4 2À . CH 3 COO À at pH , 10, respectively. The reaction preciptitate was Fe-hydroxo-phosphate complexes, mainly vivianite phase, Fe 3 (PO) 2 ·8H 2 O. This iron scrap anode could easily be extended to novel electro-Fenton process and could also stimulate the reactions using other magnetic nanoparticle electrodes.