Fe⁰ and Fe⁰ fully covered with Cu⁰ (Fe⁰ + Fe/Cu) in a fixed bed reactor for nitrate removal

Abstract

To develop a cost-effective, feasible and robust technology for nitrate removal by chemical degradation, a Fe⁰ and Fe⁰ fully covered with Cu⁰ (i.e., Fe⁰ + Fe/Cu) fixed reactor was set up in this study. The performance and mechanism of the Fe⁰ + Fe/Cu system for nitrate reduction were investigated thoroughly. First, the initial pH of the solution and the mass ratio of Fe⁰ and Fe/Cu (i.e., M(Fe⁰)/M(Fe/Cu) ratio) in the medium materials were optimized, and then two control experiments (Fe⁰ + quartz sand system and Fe⁰ + Cu⁰ system) were set up to confirm the superiority of the Fe⁰ + Fe/Cu system. The results suggest that high NO₃⁻–N removal (>99.0%), K_obs (0.403 min⁻¹) and low NH₄⁺–N generation rate (61.1%) could be obtained by the developed system (Fe⁰ + Fe/Cu) with a short hydraulic retention time (HRT = 16 min). In addition, the operational life of the Fe⁰ + Fe/Cu system and Fe⁰ + quartz sand system were investigated comparatively, which shows that the operational life of the former was much longer than that of the latter. Finally, characteristics of the medium materials in the Fe⁰ + Fe/Cu system and Fe⁰ + quartz sand system were also observed by scanning electron microscopy (SEM) and energy dispersive spectrometry (EDS) before and after long-term operation. Therefore, the developed system in this study is a promising technology for nitrate contaminated water treatment.

---

1. Introduction

Nowadays, with deepening research into the hazards of inorganic salts, nitrate, as a main contaminant in natural waters and wastewater, has become a focus of concern in the environmental field. NO₃⁻–N could be discharged by agricultural runoff, landfill leachate, animal feeding operations and industrial waste.^1–3 Higher NO₃⁻–N concentration than Maximum Concentration Level (MCL) in drinking water, ground water and direct contact water could cause significant risk to public health, such as blue baby syndrome, liver damage, methemoglobinemia and the development of cancer when it is reduced in the form of nitrite.^4–7 Reverse osmosis,⁸ electrodialysis,⁹ biological denitrification¹⁰ are usually used to treat nitrate contaminated waters. However, all of these processes suffer from the limitation of high costs or low effect.

Zero valent iron (ZVI or Fe⁰), a kind of common material for contaminated water treatment, has been used to treat the water contaminated with chlorihated organic compounds (COCs), nitroaromatic compounds (NACs), arsenic, heavy metals, nitrate, dyes, phenol etc.^5,6,11 Also, ZVI system was developed as the medium materials in fixed bed due to the high effect and low costs. In particular, nitrate was a kind of typical pollutant which could be effectively treated by fixed bed by chemical degradation with ZVI based system.^12,13 Nitrate can mainly be reduced to NH₃, N₂ and NH₄⁺ by ZVI, and the ZVI may be oxidized to Fe²⁺, Fe³⁺, Fe₂O₃, or Fe₃O₄.^14,15 In literature, it has been reported that due to the high potential difference (0.78 V) between Cu⁰ and Fe⁰, mixture of Fe⁰ (ZVI) and Cu⁰ can enhance the rates of pollutants reduction remarkably.^16,17 These have been studied by previous authors, and the results show that catalytical metal can assist the electron transfer from Fe⁰ to NO₃⁻ by forming galvanic cell, which can enhances the effect.⁴

Furthermore, the treatment efficiency of ZVI system is relatively low, and it is also limited extremely by the pH value.¹⁸ In addition, Fe⁰ particles could be passivated after long-term operation because of the production of iron oxides.¹⁹ Even though the inert materials, such as sand were added in the system, this problem still exists. In order to mitigate these defects of Fe⁰ system, Fe⁰ + transition metal (e.g., Cu⁰, Ni⁰ and Pd⁰) system were investigated,^20–22 which could not only enhance treatment effect and broaden pH range for application, but also restrict the phenomenon of harden and caking. For example, in Fe⁰ + Cu⁰ system, due to the high potential difference (0.78 V) between Cu and Fe and the galvanic corrosion reacted on the surface of particles,^22–25 the treatment efficiency could be enhance extremely, and the pH range for application could also be broaden. In addition, because the Cu⁰ particles do not react with pollutants (such as nitrate), the phenomenon of harden and caking could also be restricted.²¹ However, the high cost of these transition metals is a significant limiting factor of this system for practical application.

To resolve the high cost problem of these transition metals, the Fe/Cu bimetallic particles (i.e., Fe⁰ fully covered with Cu⁰) was prepared in this study. Since only a thin copper film was deposited on Fe⁰ surface, the cost of Fe/Cu bimetallic particles is much lower than that of Cu⁰ particles. Furthermore, mixture of Fe⁰ and Fe/Cu bimetallic particles was used as medium materials to develop a Fe⁰ + Fe/Cu system in this study with high potential difference (between Cu⁰ on Fe/Cu bimetallic particles and Fe⁰ on Fe particles)^26,27 and inert materials. In particular, Fe⁰ + Fe/Cu system for NO₃⁻–N removal and NH₄⁺–N generation were investigated thoroughly. Effects of initial pH and mass ratio of Fe⁰ and Fe/Cu (i.e., M(Fe⁰)/M(Fe/Cu) ratio) on the treatment efficiency were studied, respectively. Meanwhile, to confirm the superiority of Fe⁰ + Fe/Cu system, two control experiments (Fe⁰ + Cu⁰ system and Fe⁰ + quartz sand system) were setup, and the operational life of Fe⁰ + Fe/Cu system and Fe⁰ + quartz sand system were investigated. Finally, the particles in Fe⁰ + Fe/Cu system and Fe⁰ + quartz sand system were observed by scanning electron microscope (SEM) and energy dispersive spectrometer (EDS) after operational life experiments.

2. Experimental

2.1 Reagents

NaNO₃ (analytical reagent), CuSO₄·5H₂O (analytical reagent), ethylene diamine tetraacetic acid disodium salt (Na₂EDTA·2H₂O) (analytical reagent), copper powders (Cu⁰) and zero valent iron (ZVI or Fe⁰) powders purchased from Chengdu Kelong chemical reagent factory were used in the experiments. ZVI powders have a mean particle size of approximately 120 μm, and their Fe content is above 99%. Cu powders have a mean particle size of approximately 50 μm, and their Cu content reaches above 99%. Two kinds of quartz sand have different particle size of 200 μm and 1 mm, respectively, and their SiO₂ content reaches above 95%. Other chemicals used in the experiment were of analytical grade. Deionized water was used in all experiments.

2.2 Preparation of Fe⁰ fully covered with Cu⁰ (i.e. Fe/Cu) particles

The micron-scale ZVI powder was used as the substrate material for the preparation of Fe/Cu bimetallic particles. The Fe/Cu bimetallic particles were prepared by displacement plating, via adding 400 mL solution with CuSO₄·5H₂O (29.30 g L⁻¹) and Na₂EDTA·2H₂O (8.72 g L⁻¹) to Fe particles (30 g L⁻¹) and keeping certain plating temperature (30 ± 1 °C) and stirring speed (250 rpm).²⁸ Then the prepared Fe/Cu bimetallic particles were separated from the supernatant immediately after plating process. Finally, the separated Fe/Cu bimetallic particles were rinsed three times with deionized water, three times with ethanol, and then they were dried under vacuum protection at 50 ± 1 °C for 2 hours.

2.3 Experimental procedures

The experiments of nitrate reduction were conducted in a column reactor with a diameter of 30 mm and total length of 150 mm. Coarse quartz sand with particle size of 1 mm was used as fillers on the bottom height of 50 mm, and 60 g medium materials (i.e., Fe⁰ and Fe/Cu, Fe⁰ and Cu⁰, or Fe and quartz sand) were fully mixed, and then they were added above the coarse quartz sand in the column, respectively. Thus three experimental systems with different medium materials were set up, respectively. 200 mL NaNO₃ solution ([NO₃⁻–N] = 50 mg L⁻¹) was used as model contaminated water, and the initial pH was adjusted by adding NaOH or H₂SO₄ solution (0.1 mol L⁻¹). The feed water was circulated through the packed column by a peristaltic pump, and the empty bed hydraulic retention time was calculated by the medium materials volume and treatment time. The column reactor was performed at 25 ± 1 °C by water batch heating, and the flow rate of pump was 50 mL min⁻¹. The configuration of reactor is shown in Fig. S1.† To investigate the Fe⁰ + Fe/Cu system thoroughly, the NaNO₃ solution with different initial pH and the medium materials with different mass ratio of Fe⁰ and Fe/Cu (i.e., M(Fe⁰)/M(Fe/Cu) ratio) were studied. Furthermore, the effect was evaluated by the results of hydraulic retention time versus NO₃⁻–N removal and hydraulic retention time versus NH₄⁺–N generation. In addition, to confirm the superiority of Fe⁰ + Fe/Cu system, two control experiments, (a) Fe⁰ + quartz sand system and (b) Fe⁰ + Cu⁰ system were setup. The mass ratio of Fe⁰ of medium materials and initial pH of NaNO₃ solution in the control experiments were same to the optimal conditions of Fe⁰ + Fe/Cu system. Samples were taken from the system at intervals by withdrawing 1 mL of sample solution, and NO₃⁻–N concentration and NH₄⁺–N concentration were detected. Furthermore, the operational life of Fe⁰ + quartz sand system and Fe⁰ + Fe/Cu systems was also investigated comparatively. A certain amount of NaNO₃ solution was used as feed water once, and it was treated according to the best hydraulic retention time, which could reach 99.0% NO₃⁻–N removal (obtained in above experiments). The operational conditions were same as them in the control experiments. Feed water was changed, and NO₃⁻–N concentration and released iron ions concentration in solution were detected after each circulation. Finally, the particles were taken of SEM and EDS after operational life experiments.

2.4 Analytical method

Nitrate measurement was achieved by the ultraviolet spectrophotometric method at 220 nm and 275 nm. Ammonium was measured using its reaction with Nessler to form complex compound at 420 nm.¹⁴ The released Fe²⁺ was determined as phenanthroline method at 510 nm. The released iron ions (Fe²⁺ and Fe³⁺) was analyzed as Fe²⁺ after its reduction using hydroxylamine hydrochloride.¹⁴

The particles were observed by JSM-7500F scanning electron microscopy (SEM, JEOL, Japan). In addition, the surface elementary composition of the particles was analyzed by energy dispersive spectrometry (EDS). EDS analysis was carried out by a permanent thin film window link (Oxford Instruments) detector and WinEDS software in a JEOL JSM-7500F scanning electron microscope (SEM). At the end, this instrument was operated at 25 kV and emission current of 60–70 μA. The pH was measured by a pHS-3C meter (Rex, China).

3. Results and discussion

3.1 Effect of initial pH of solution

In literature, it is reported that the performance of ZVI system can be affected seriously by the initial pH.^29,30 Effect of the initial pH (3.0–9.0) on the NO₃⁻–N removal and NH₄⁺–N generation by Fe⁰ + Fe/Cu system were evaluated thoroughly. Meanwhile, M(Fe⁰)/M(Fe/Cu) ratio was 1.00 in Fe⁰ + Fe/Cu system. Fig. 1(a) shows the logarithmic plots of residual concentration of NO₃⁻–N in solution versus the hydraulic retention time with different initial pH, and a good linear fitting was observed in the treatment system. The results indicate that the rates of NO₃⁻–N removal with different initial pH of the solution were described by the observed pseudo-first-order as shown as follows:where [NO₃⁻–N] is the NO₃⁻–N concentration at instant t (mg L⁻¹), [NO₃⁻–N]₀ is the initial NO₃⁻–N concentration, K_obs is the observed pseudo-first-order degradation rate constant (min⁻¹), and t is the hydraulic retention time (min). The observed pseudo-first-order degradation rate constant (K_obs) increased from 0.137 min⁻¹ to 0.0351 min⁻¹ when the initial pH decreased from 9.0 to 3.0. Fig. 1(b) shows that the values of K_obs for NO₃⁻–N reduction decreased with the increase of initial pH. The K_obs and initial pH have a good linear relationship and the correlation of determination (R²) reaches 0.90, and the kinetic expression can be presented as follows:

K_obs = −0.037pH + 0.447 (2)

Fig. 1 Reduction kinetics (a) and K_obs values (b) for NO₃⁻–N removal, and [NH₄⁺–N]/[NO₃⁻–N removal] ratio (c) obtained by Fe⁰ + Fe/Cu system with different initial pH (experiment conditions: initial NO₃⁻–N concentration of 50 mg L⁻¹, M(Fe⁰)/M(Fe/Cu) ratio of 1.00, and total mass of Fe⁰ and Fe/Cu of 60 g).

Eqn (2) can be used to estimate the potential effect of initial pH variability on K_obs of PNP reduction. However, the slope of this equation was lower compared with that in Fe⁰ system for pollutant degradation in previous studies.³¹ Fig. 1(c) shows that there is no obvious difference about NH₄⁺–N generation among the systems with different initial pH of solution, and [NH₄⁺–N]/[NO₃⁻–N removal] ratio of all of these systems reached 63.0% to 65.0% at the end of the treatment processes.

The low initial pH could support enough [H⁺] which could accelerate the corrosion rate of Fe particles in the Fe⁰ + Fe/Cu system and keep their fresh surface. Suzuki et al.³² and Xiaomeng et al.³³ reported that NO₃⁻–N directly received electrons from ZVI through an iron corrosion product layer and ions in ZVI system. In addition, the existence of Fe²⁺ could enhance the reduction effect of nitrate remarkably,^15,34 and high H⁺ concentration can accelerate the generation of Fe²⁺. According to the conclusions from the studies of previous authors,^33,35 the main reaction equations of nitrate reduction was shown in eqn (3)–(8) as follows:

10Fe⁰ + 6NO₃⁻ + 3H₂O + 6H⁺ → 5Fe₂O₃ + 3H₂O + 3N₂ (3)

5Fe⁰ + 2NO₃⁻ + 12H⁺ → 5Fe²⁺ + N₂ + 6H₂O (4)

4Fe⁰ + NO₃⁻ + 10H⁺ → 4Fe²⁺ + NH₄⁺ + 3H₂O (5)

2.82Fe⁰ + NO₃⁻ + 0.75Fe²⁺ + 1.75H₂O + 0.50H⁺ → 1.19Fe₃O₄ + NH₄⁺ (6)

4Fe⁰ + NO₃⁻ + 10H⁺ → 4Fe²⁺ + NH₄⁺ + 3H₂O (7)

8Fe²⁺ + NO₃⁻ + 10H⁺ → 8Fe³⁺ + NH₄⁺ + 3H₂O (8)

The equations show that the reduction processes consume H⁺, which could also explain the better effect under lower initial pH condition. However, initial pH influences the NO₃⁻–N removal effect remarkably with Fe particles treatment^32,36 while relatively slightly in Fe⁰ + Fe/Cu system. This result could be explained by the theory that the existence of Fe/Cu bimetallic particles formed lots of Fe–Cu cells, and the catalytic performance of Cu could make the galvanic corrosion happen on the surface, which could obtain relatively high reactivity even though the initial pH is high in the solution.³⁷ The similar results were also concluded in our previous studies.³⁸

Also, researchers reported that the selectivity of nitrate reduction (ammonia or nitrogen) could be influenced by the initial pH of solution in iron based system, because nitrate is first reduced to nitrite as an intermediate product, and then nitrite was converted to nitrogen or ammonia.^39,40 Therefore, the hydrogenation of NO₂⁻ is a key process in determining the selectivity, and researchers have found that lower pH contribute to less ammonia generation. On the contrary, in Fe⁰ + Fe/Cu system, initial pH of the solution was not influence the selectivity remarkably, which could be explained by that the main influence factor of hydrogenation process was Fe⁰ and Fe/Cu particles conditions,⁴⁰ and thus initial pH in this system influences the selectivity relatively slightly. Lower pH in solution accelerates the corrosion rate of Fe particles, which can enhance the consumption rate of the Fe particles, and thus the costs of this system will increase. In addition, the initial pH of natural water and wastewater water is usually neutral, and adjusting to lower pH can also increase the costs and cause secondary pollution. The H⁺ in waters also need to be neutralized after the system treatment. Consequently, considering these factors, initial pH of 7.0 was selected as the optimal condition in the subsequent experiments to investigate the effect of M(Fe⁰)/M(Fe/Cu) ratio on NO₃⁻–N removal and NH₄⁺–N generation.

3.2 Effect of M(Fe⁰)/M(Fe/Cu) ratio of medium materials

The M(Fe⁰)/M(Fe/Cu) ratio of medium materials is also a significant factor influencing the efficiency of NO₃⁻–N removal in Fe⁰ + Fe/Cu system, and thus the effect of M(Fe⁰)/M(Fe/Cu) ratio of medium materials on NO₃⁻–N removal and NH₄⁺–N generation was evaluated keeping the initial pH of 7.0 ± 0.1 in solution. Fig. 2(a) and (b) shows the NO₃⁻–N removal with different M(Fe⁰)/M(Fe/Cu) ratio, which demonstrates that all followed the pseudo first-order kinetics model. It is clear that K_obs increased gradually to the maximum (0.404 min⁻¹) when M(Fe⁰)/M(Fe/Cu) ratio increased from 0.50 to 0.67. And then it began to decrease with the further increase of M(Fe⁰)/M(Fe/Cu) ratio (from 0.67 to 2.00). The hydraulic retention time was 16 min reaching 99.0% NO₃⁻–N removal with M(Fe⁰)/M(Fe/Cu) ratio of 0.67. Also, the system with Fe⁰ alone (i.e. Fe particles system) had low K_obs value, which was only 0.0749. In addition, as shown in Fig. 2(c), at the end of the treatment processes, the ratio of [NH₄⁺–N]/[NO₃⁻–N] decreased (0.50 to 0.85), increased (0.85 to 1.00) and then kept steady (1.00 to 2.00) with the M(Fe⁰)/M(Fe/Cu) ratio increase. The lowest ratio obtained from the experiments is 56.4% with the M(Fe⁰)/M(Fe/Cu) ratio of 0.85. Higher ratio of [NH₄⁺–N]/[NO₃⁻–N removal] was detected with Fe⁰ alone (72.3%).

Fig. 2 Reduction kinetics (a) and K_obs values (b) for NO₃⁻–N removal, and [NH₄⁺–N]/[NO₃⁻–N removal] ratio (c) obtained by Fe⁰ + Fe/Cu system with different M(Fe⁰)/M(Fe/Cu) ratios (experiment conditions: initial NO₃⁻–N concentration of 50 mg L⁻¹, initial pH of 7.0 ± 0.1, and total mass of Fe⁰ and Fe/Cu of 60 g).

The influence of theoretical Cu mass loading on ZVI in Fe/Cu bimetallic particles system for wastewater treatment has been investigated thoroughly in our previous studies.²⁸ In this study, the influence of M(Fe⁰)/M(Fe/Cu) ratio of medium materials has the similar theory with that. Due to the dense, uniform and fully covered Cu on the Fe/Cu bimetallic particles, the galvanic couple formation and electron transfer were between Cu layer on Fe/Cu metallic particles and Fe particles. The results can be explained as follows. When M(Fe⁰)/M(Fe/Cu) ratio is higher (more than 0.67), the decrease of the ratio can enhance the contact area between Cu and Fe⁰ by enhancing the proportion of Cu (on Fe/Cu metallic particles), which can form more Fe–Cu cell and improve the rate of electron transfer, and thus the reduction rate of nitrate can also be increased. When M(Fe⁰)/M(Fe/Cu) ratio is lower (less than 0.67), with the decrease of this ratio, the contact area between Cu and Fe⁰ will be decreased gradually, which goes against the electron transfer. Also, the Fe particles mass was not enough for the reaction of the nitrate reduction process when the M(Fe⁰)/M(Fe/Cu) ratio was little, which also caused the decrease of the efficiency of NO₃⁻–N removal. The efficiency of NO₃⁻–N removal was low with Fe⁰ alone, which concluded that Fe⁰ + Fe/Cu system was much superior to Fe particles system on nitrate reduction.

Furthermore, the kind of reduction product of nitrate could be controlled by optimizing M(Fe⁰)/M(Fe/Cu) ratio of medium materials. The change of this ratio might influence the path of nitrate reduction, and thus influence the generation of ammonia, which has the similar result with the (Pd–Cu)–nFe⁰ system investigated by the previous authors.⁴¹ It has been known that the activated hydrogen adsorbed producing by electron transfer process on the active sites of catalytic metal can abstract oxygen from nitrite, and then two nitrogen atoms are consecutively bonded to form nitrogen gas in bimetallic particles system.³⁹ When the amount of Fe/Cu bimetallic particles were increased, the concentration of atomic hydrogen adsorbed on active sites of Cu layer might be increased, resulting in the enhancement of nitrogen gas selectivity. However, too low ratio of M(Fe⁰)/M(Fe/Cu) can cause the decrease of the amount of reactant, Fe particles, which could restrict the reaction rate and path, and thus increased the ratio of [NH₄⁺–N]/[NO₃⁻–N removal]. According to the results, with the M(Fe⁰)/M(Fe/Cu) ratio of 0.85, the degradation process achieved better path for less NH₄⁺–N generation. Considering the results of NO₃⁻–N removal and NH₄⁺–N generation, M(Fe⁰)/M(Fe/Cu) ratio of 0.67 was chosen as the optimal condition in the following experiments.

3.3 Control experiments

As a traditional system for nitrate contaminated water treatment in fixed bed, Fe⁰ + quartz sand system has been used in industry for several years. In recent years, the Fe⁰ + transition metal systems was also developed, and this new system (i.e. Fe⁰ + Fe/Cu system) was based on the Fe⁰ + transition metal systems (Fe⁰ + Cu⁰ system). Therefore, to investigate the superiority of Fe⁰ + Fe/Cu system, two control experiments, (a) Fe⁰ + Cu⁰ system and (b) Fe⁰ + quartz sand system were setup. Both of these two control experiments have same initial pH (7.0 ± 0.1) in solution and M(Fe⁰)/M(Fe/Cu) ratio (0.67) of medium materials. Fig. 3 shows that best efficiency of nitrate reduction was obtained by Fe⁰ + Fe/Cu system treatment. The K_obs value of NO₃⁻–N removal of Fe⁰ + Fe/Cu system (0.404) was much higher than that obtained in Fe⁰ + Cu⁰ system (0.109) and Fe⁰ + quartz sand system (0.0282). The NO₃⁻–N removal reached 99.0% with the hydraulic retention time of 16 min by Fe⁰ + Fe/Cu system, however, which of that reached 99.0% with the hydraulic retention time of 32 min and 128 min by Fe⁰ + Cu⁰ particle system and Fe⁰ + quartz sand system, respectively. Furthermore, the NH₄⁺–N generation in solution was also detected in these three systems. As shown in Fig. 3(c), the ratio of NH₄⁺–N generation versus NO₃⁻–N removal was lowest in the solution treated by Fe⁰ + Fe/Cu system, [NH₄⁺–N]/[NO₃⁻–N removal] ratio of which were only 61.07% at the end of treatment. And that of the solution by Fe⁰ + Cu⁰ system and Fe⁰ + quartz sand system were 70.2% and 72.3%, respectively.

Fig. 3 Reduction kinetics (a) and K_obs values (b) for NO₃⁻–N removal, and [NH₄⁺–N]/[NO₃⁻–N removal] ratio (c) obtained by different treatment systems (experiment conditions: initial NO₃⁻–N concentration of 50 mg L⁻¹, initial pH of 7.0 ± 0.1, M(Fe⁰)/M(other materials) of 0.67, and total mass of Fe⁰ and other materials of 60 g).

The results show stronger efficiency of NO₃⁻–N removal and less NH₄⁺–N generation in the solution by Fe⁰ + Fe/Cu system. Compared with Fe⁰ + quartz system, the superiority of that was obvious. The reaction rate increased with the existence of Cu layer by forming Fe–Cu galvanic cell, which could enhance the rate of electron transfer.^26,42 In addition, the contribution of more nitrogen and less ammonia generation by catalytic metal with the formation of galvanic cell in bimetallic particles system has been reported.^39,43,44 The reaction path could change with the Cu as catalyst, which cause reduction of NH₄⁺–N generation. And compared with the Fe⁰ + Cu⁰ system, the Fe⁰ + Fe/Cu system also shows better effect of NO₃⁻–N removal and NH₄⁺–N generation. These results could be explained by the following theories. (i) Because particle size of the Cu particles is relatively small, the Cu particles and Fe particles have different mechanics characteristics, which could cause the uneven mixture between Fe⁰ and Fe/Cu bimetallic particles, and thus could cause the decrease of contact between Fe⁰ and Cu⁰ and goes against the electron transfer. (ii) The Cu layer on Fe/Cu particles was fresh, and little of it was oxidized by oxygen or other oxidant. However, Cu particles might have copper oxide and lower reactivity on their surface, which reduce the amount of Cu particles with catalytic activity, and thus worse effect was obtained by Fe⁰ + Cu⁰ system.

3.4 The operational life of Fe⁰ + quartz sand system and Fe⁰ + Fe/Cu system

In order to confirm the superiority of operational life of Fe⁰ + Fe/Cu system, the operational life of Fe⁰ + Fe/Cu system and Fe⁰ + quartz sand system were investigated. A certain amount of NaNO₃ solution was used as feed water once, and it was treated with the obtained best hydraulic retention time in Fe⁰ + Fe/Cu system (64 min per circulation) and Fe⁰ + quartz sand system (512 min per circulation), respectively. The operational conditions were same as them in the control experiments (initial pH of 7.0 ± 0.1 and M(Fe⁰)/M(Fe/Cu) ratio of 0.67). Feed water was changed, and NO₃⁻–N concentration and released iron ions concentration in solution were detected after each circulation. 200 mL solution treatment is perceived as one time.

The results are shown in Fig. 4. The Fe⁰ + Fe/Cu system was operated 180 times, and the Fe⁰ + quartz sand system was operated 85 times. The NO₃⁻–N removal and released iron ions concentration in effluent were 95.3% and 9.72 mg L⁻¹ after first 4 times treatment, decreased to 49.2% (less than 50%) and 7.09 mg L⁻¹ at 56 times, and decreased to 8.2% (less than 10%) and 3.38 mg L⁻¹ at 80 times in solution by Fe⁰ + sand quartz system treatment, respectively. And they were 95.8% and 10.98 mg L⁻¹ after first 4 times treatment, decreased to 44.7% (less than 50%) and 5.92 mg L⁻¹ at 96 times, and decreased to 9.5% (less than 10%) and 2.27 mg L⁻¹ at 160 times in solution by Fe⁰ + Fe/Cu system treatment, respectively. The total nitrate (NO₃⁻–N) reduction was 615 mg by Fe⁰ + quartz sand system and 1127 mg by Fe⁰ + Fe/Cu system, respectively. The Cu²⁺ or Cu¹⁺ was not detected in whole experiments.

Fig. 4 NO₃⁻–N removal and released iron ions concentration in the effluent obtained by two treatment systems (i.e., Fe⁰ + Fe/Cu and Fe⁰ + quartz sand) with different frequency wastewater circulation (experiment conditions: initial NO₃⁻–N concentration of 50 mg L⁻¹, initial pH of 7.0 ± 0.1, M(Fe⁰)/M(other materials) of 0.67, and total mass of Fe⁰ and other materials of 60 g).

According to the results, the Fe⁰ + Fe/Cu system also shows much longer operational life than the Fe⁰ + quartz sand system. The released iron in solution could indicated the corrosion degree of Fe particles, and the previous authors have reported that Fe²⁺ and Fe³⁺ in solution could increase the effect of nitrate reduction remarkably,^45,46 and thus, the theories could explain the better performance of Fe⁰ + Fe/Cu system and the positive correlation change law of NO₃⁻–N removal and released iron ions concentration in effluent. During the late treatment, the caking and weaker reactivity could cause the reduction of released iron ions concentration in the solution.

In addition, the results about NO₃⁻–N removal effect could be explained by following theories. (i) The Fe⁰ + Fe/Cu system could change the rate and path of electron transfer,⁴ which could influence the path of nitrate reduction. The process in Fe⁰ + quartz sand system might cause the production of iron oxides cover the Fe particles, and the existence of Cu might reduce the amount of iron oxides on the surface of Fe particles. (ii) Due to the difference mechanics characteristics of Fe particles and quartz sand, the treatment process can cause the uneven mixture between Fe⁰ and quartz sand, and thus the system became harden and caking and formed solidified zone after several times treatment.⁴⁷ However, the Fe particles and Fe/Cu bimetallic particles in Fe⁰ + Fe/Cu system have similar mechanics characteristics, and dense, uniform and fully covered Cu layer is also inert material in this system. Therefore, it could relief that phenomenon and increases the operational life of this system. (iii) Fig. 4 shows that released iron ions concentration in solution treated by Fe⁰ + Fe/Cu system were higher than that treated by Fe⁰ + quartz sand system at the same number of treatment, and thus better performance was shown in Fe⁰ + Fe/Cu system.^15,45 (iv) During the late treatment process, the Cu layer could abscise, and the Fe/Cu bimetallic particles also could react with the NO₃⁻. Consequently, the Fe⁰ + Fe/Cu system has more activated Fe⁰, and could support longer operational life. (v) The Fe⁰ + Fe/Cu system has much higher rate for nitrate reduction which cut the treatment time of this system, and thus the extra corrosion consumption of Fe particles could be reduced. The total amount of nitrate reduction could hence be enhanced.

3.5 SEM and EDS analyses

In order to identify the morphologies and composition of particles in Fe⁰ + Fe/Cu system and Fe⁰ + quartz sand system, SEM and EDS analysis were used in this study. The SEM and EDS of fresh Fe particles and quartz sand and Fe particles in Fe⁰ + quartz sand system after 50 times reacted and 85 times reacted are shown in Fig. 5 and them of fresh Fe/Cu bimetallic particles and Fe particles and Fe/Cu bimetallic particles in Fe⁰ + Fe/Cu system after 50 times reacted and 180 times reacted are shown in Fig. 6.

Fig. 5 SEM-EDS spectra of (a) fresh Fe particles before reacted, (b) quartz sand particles and (b*) Fe particles from Fe⁰ + quartz sand system after 50 times reacted, and (c) quartz sand particles and (c*) Fe particles from Fe⁰ + quartz sand system after 85 times reacted.

Fig. 6 SEM-EDS spectra of (a) fresh Fe/Cu bimetallic particles before reacted, (b) Fe particles and (b*) Fe/Cu bimetallic particles from Fe⁰ + Fe/Cu system after 50 times reacted, and (c) Fe particles and (c*) Fe/Cu bimetallic particles from Fe⁰ + Fe/Cu system after 180 times reacted.

As shown in Fig. 5(a), no O element was detected on Fe particles before reacted. In Fig. 5(b*) the composition of the Fe particle in Fe⁰ + quartz sand system after 50 times reacted was shown and the punctiform iron oxides could be observed in SEM imaging. The particle-state Fe particles in Fe⁰ + quartz sand system after 85 times reacted were difficult to find, and thus the Fig. 5(c) is imaging and composition of quartz sand with iron oxides on its surface, and Fig. 5(c*) is them of Fe particles from caking.

Fig. 6(a) shows the fresh Fe/Cu bimetallic particle in Fe⁰ + Fe/Cu system. No O element was detected on the particle, and the particle was spongy, the shape of which was same as that of Fe particles.²⁸ This could also confirm that the Fe/Cu bimetallic particles have fresh, dense, uniform and fully covered Cu layer on it. Fig. 6(b) is about the Fe particle and Fig. 5(b*) is about the Fe/Cu metallic particles in Fe⁰ + Fe/Cu system after 50 times reacted. 6.56% Cu and 39.55% O was detected on the Fe particle, respectively. The Fe/Cu metallic particle still keep spongy and O was also detected on it. Fig. 5(c) and (c*) are Fe particle and Fe/Cu bimetallic particle in Fe⁰ + Fe/Cu system after 180 times reacted, respectively. More than 40% of O was detected and oxides can be observed on both of these two particles. In addition, Fe particles shows more Cu on it while Fe/Cu shows less compared with them in Fe⁰ + Fe/Cu system after 50 times reacted.

In Fe⁰ + Fe/Cu system, the Fe/Cu bimetallic particles after 50 times reacted are still spongy and little Cu abscised. Little O element was detected on the surface of Fe/Cu bimetallic particles, which could be explained by that the O element might be from the formed iron oxides adsorbed on the surface of Fe/Cu bimetallic particles. The Cu layer could relief the formation of solidified zone, which has been mentioned above. The main heterogeneous reaction by Fe⁰ for NO₃⁻–N at this phase are shown in Fig. 7(a). Therefore, most Fe and Fe/Cu still kept particle-state after 50 times reacted, while the Fe particles were harden and caking with same times treatment. The O element weight (%) was higher on the Fe particles in Fe⁰ + Fe/Cu system than them in Fe⁰ + quartz sand system. The total NO₃⁻–N removal was much higher in Fe⁰ + Fe/Cu system, which cause more iron oxides formed. However, the Fe⁰ + Fe/Cu system still had better performance compared with the Fe⁰ + quartz sand system, which could be explained by that Cu layer in this system could assist the transfer of electron, and thus indicated better effect.

Fig. 7 The main heterogeneous reaction mechanism of the Fe⁰ + Fe/Cu system (Fe⁰ for NO₃⁻–N removal).

Furthermore, some Cu layer on Fe/Cu bimetallic particles has abscised during the reacted process, and this caused some Fe⁰ of Fe/Cu metallic particles bare, and some Cu adhering on Fe particles. Consequently, the process of electron transfer for nitrate reduction with the catalysis of Cu could finish on one particle, and thus the rate of NO₃⁻–N removal could further increase. In other word, the galvanic cell could also form between Cu layer of Fe/Cu bimetallic particles and Fe⁰ form Fe/Cu bimetallic particles (as shown in Fig. 7(b)). Also, the electron transfer process still could happen between Cu layer (abscised or on Fe/Cu bimetallic particles) and Fe⁰ from Fe⁰ particles. The main heterogeneous reaction by Fe⁰ for NO₃⁻–N at this phase are shown in Fig. 7(a) (at first) and (b) (at later). And thus, much better performance were still kept in this phase compared with Fe⁰ + quartz sand system. However, with longer times treatment (more than 180 times), most of Cu was abscised and some iron oxide and iron hydroxide might adhered on the surface of Fe particles and Fe/Cu bimetallic particles, which could restrict the reactivity of this system extremely. The particles in Fe⁰ + Fe/Cu system have higher degree of oxidation compared with them in Fe⁰ + quartz sand system. This result could be explained by that the nitrate reduction in Fe⁰ + Fe/Cu system might mainly be achieved by iron ions, which have better performance than Fe⁰ and generate less iron oxides.¹⁴ In Fe⁰ + Fe/Cu system, the released iron ions were higher according to Fig. 4, and the precipitation of iron hydroxides might also be more. And in contrast, the Fe particles in Fe⁰ + quartz sand system might mainly reduce nitrate by producing iron oxides on their surface, and thus caused the formation of solidified zone. The fluidized state Fe particles were difficult to find in Fe⁰ + quartz sand system after 85 times reacted, because residual Fe particles have formed caking, which cannot continue to react for nitrate reduction. Considering above explanations, better performance for once reacted and longer operational life of Fe⁰ + Fe/Cu system could be explained.

3.6 Analysis of preparation cost

The cost of Fe particles, Cu particles and CuSO₄·5H₂O are about 200 USD per t, 4000 USD per t and 1500 USD per t, respectively. EDTA–2Na was not consumed in the Fe/Cu preparation process. The cost of Fe/Cu bimetallic particles are about 1370 USD per t. Therefore, the costs of medium materials of Fe⁰ + Fe/Cu and Fe⁰ + Cu⁰ were about 900 USD per t and 2480 USD per t, respectively. Also the Fe⁰ + Fe/Cu system showed better performance than Fe⁰ + Cu⁰ system and Fe⁰ + quartz sand system. Consequently, Fe + Fe/Cu system was a cost-effective technology for nitrate contaminated water treatment.

4. Conclusions

The model nitrate contaminated water was effectively treated by the Fe⁰ and Fe⁰ fully covered Cu⁰ (Fe⁰ + Fe/Cu system) by chemical degradation. The optimal conditions (initial pH of 7.0 in solution and M(Fe⁰)/M(Fe/Cu) ratio of 0.67 of medium materials) were obtained, and under the optimal conditions, this system could obtain high NO₃⁻–N removal (>99.0%) with short hydraulic retention time (HRT = 16 min), high K_obs value (0.403 min⁻¹) and low NH₄⁺–N generation rate (61.1%). In addition, its treatment ability was much higher than that of two control experiments. Furthermore, Fe⁰ + Fe/Cu system also shows the superiority of operational life. Finally, according to the analysis of SEM and EDS, the results could be confirmed. As conclusions, this system can be considered as a cost-effective, feasible and robust technology for nitrate contaminated water treatment.

Acknowledgements

The authors would like to acknowledge the financial support from National Natural Science Foundation of China (No. 21207094), Fundamental Research Funds for the Central Universities (No. 2015SCU04A09) and Funds for Innovation and Entrepreneurship Training for College Students (Sichuan University).

References

1. K. Hiscock, J. Lloyd and D. Lerner, Water Res., 1991, 25, 1099–1111.
2. V. Joekar-Niasar and B. Ataie-Ashtiani, Environ. Geol., 2009, 57, 1785–1798.
3. S.-J. Liu, Z.-Y. Zhao, J. Li, J. Wang and Y. Qi, Water Res., 2013, 47, 5977–5985.
4. S. M. Hosseini, B. Ataie-Ashtiani and M. Kholghi, Desalination, 2011, 276, 214–221.
5. A. B. Cundy, L. Hopkinson and R. L. Whitby, Sci. Total Environ., 2008, 400, 42–51.
6. F. Fu, D. D. Dionysiou and H. Liu, J. Hazard. Mater., 2014, 267, 194–205.
7. B. T. Nolan, B. C. Ruddy, K. J. Hitt and D. R. Helsel, Environ. Sci. Technol., 1997, 31, 2229–2236.
8. R. Rautenbach, W. Kopp, G. Van Opbergen and R. Hellekes, Desalination, 1987, 65, 241–258.
9. M. Bosko, M. Rodrigues, J. Z. Ferreira, E. Miró and A. Bernardes, J. Membr. Sci., 2014, 451, 276–284.
10. V. Vavilin and S. Rytov, Chemosphere, 2015, 134, 417–426.
11. M. Stefaniuk, P. Oleszczuk and Y. S. Ok, Chem. Eng. J., 2016, 287, 618–632.
12. R. Thiruvenkatachari, S. Vigneswaran and R. Naidu, J. Ind. Eng. Chem., 2008, 14, 145–156.
13. F. Obiri-Nyarko, S. J. Grajales-Mesa and G. Malina, Chemosphere, 2014, 111, 243–259.
14. C. Tang, Z. Zhang and X. Sun, J. Hazard. Mater., 2012, 231, 114–119.
15. J. Xu, Z. Hao, C. Xie, X. Lv, Y. Yang and X. Xu, Desalination, 2012, 284, 9–13.
16. A. Koutsospyros, J. Pavlov, J. Fawcett, D. Strickland, B. Smolinski and W. Braida, J. Hazard. Mater., 2012, 219, 75–81.
17. L. M. Ma and W. X. Zhang, Environ. Sci. Technol., 2008, 42, 5384–5389.
18. S. M. Hosseini and T. Tosco, J. Contam. Hydrol., 2015, 179, 182–195.
19. B. Lai, Y. X. Zhou and P. Yang, Ind. Eng. Chem. Res., 2012, 51, 7777–7785.
20. L. Yin, Y. Dai, J. Niu, Y. Bao and Z. Shen, J. Hazard. Mater., 2012, 209, 414–420.
21. H. Lin, Y. Lin and L. Liu, J. Taiwan Inst. Chem. Eng., 2016, 58, 148–154.
22. L. M. Ma, Z. G. Ding, T. Y. Gao, R. F. Zhou, W. Y. Xu and J. Liu, Chemosphere, 2004, 55, 1207–1212.
23. S. J. Bransfield, D. M. Cwiertny, K. Livi and D. H. Fairbrother, Appl. Catal., B, 2007, 76, 348–356.
24. A. Ghauch and A. Tuqan, Chemosphere, 2008, 73, 751–759.
25. Z. Zheng, S. Yuan, Y. Liu, X. Lu, J. Wan, X. Wu and J. Chen, J. Hazard. Mater., 2009, 170, 895–901.
26. L. Sun, H. Song, Q. Li and A. Li, Chem. Eng. J., 2016, 283, 366–374.
27. A. Koutsospyros, J. Pavlov, J. Fawcett, D. Strickland, B. Smolinski and W. Braida, J. Hazard. Mater., 2012, 219–220, 75–81.
28. B. Lai, Y. Zhang, Z. Chen, P. Yang, Y. Zhou and J. Wang, Appl. Catal., B, 2014, 144, 816–830.
29. Y. Sun, J. Li, T. Huang and X. Guan, Water Res., 2016, 100, 277–295.
30. Y. Nakatsuji, Z. Salehi and Y. Kawase, J. Environ. Manage., 2015, 152, 183–191.
31. B. Lai, Z. Y. Chen, Y. X. Zhou, P. Yang, J. L. Wang and Z. Q. Chen, J. Hazard. Mater., 2013, 250, 220–228.
32. T. Suzuki, M. Moribe, Y. Oyama and M. Niinae, Chem. Eng. J., 2012, 183, 271–277.
33. F. Xiaomeng, G. Xiaohong, M. Jun and A. Hengyu, J. Environ. Sci., 2009, 21, 1028–1035.
34. J. Zhang, Z. Hao, Z. Zhang, Y. Yang and X. Xu, Process Saf. Environ. Prot., 2010, 88, 439–445.
35. C. Della Rocca, V. Belgiorno and S. Meriç, Desalination, 2007, 204, 46–62.
36. J. Rodríguez-Maroto, F. García-Herruzo, A. García-Rubio, C. Gómez-Lahoz and C. Vereda-Alonso, Chemosphere, 2009, 74, 804–809.
37. W. Han, F. Fu, Z. Cheng, B. Tang and S. Wu, J. Hazard. Mater., 2016, 302, 437–446.
38. Y. Ren and B. Lai, RSC Adv., 2016, 6, 58302–58314.
39. J. Shi, C. Long and A. Li, Chem. Eng. J., 2016, 286, 408–415.
40. Y. Lubphoo, J. M. Chyan, N. Grisdanurak and C. H. Liao, J. Taiwan Inst. Chem. Eng., 2016, 59, 285–294.
41. Y. Lubphoo, J.-M. Chyan, N. Grisdanurak and C.-H. Liao, J. Taiwan Inst. Chem. Eng., 2015, 57, 143–153.
42. J. Jung, S. Bae and W. Lee, Appl. Catal., B, 2012, 127, 148–158.
43. U. Prüsse and K.-D. Vorlop, J. Mol. Catal. A: Chem., 2001, 173, 313–328.
44. I. Mikami, Y. Sakamoto, Y. Yoshinaga and T. Okuhara, Appl. Catal., B, 2003, 44, 79–86.
45. D. Shiers, D. Ralph and H. Watling, Hydrometallurgy, 2014, 150, 259–268.
46. J. Shi, S. Yi, H. He, C. Long and A. Li, Chem. Eng. J., 2013, 230, 166–171.
47. L. Carniato, G. Schoups, P. Seuntjens, T. Van Nooten, Q. Simons and L. Bastiaens, J. Contam. Hydrol., 2012, 142, 93–108.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra24014c

---