Although zero-valent iron represents a promising approach for reduction of nitrate (NO3−) in water, its application in concentrated nitrate is limited by surface passivation. In this study, an alternative approach using in situ synthesized zero-valent copper (Cu0) produced by borohydride (NaBH4) was investigated. Complete reduction was observed within 55 min by reacting 677 mg-N/L of NO3− with CuO (0.312 g/L) and NaBH4 (4.16 g/L) at 60 °C. The pseudo-first-order rate constant was 0.059 min−1, and it increased threefold when the CuO dose was increased to 1.24 g/L. Increasing the NaBH4 dose produced less nitrite (NO2−) throughout the experiments, indicating that it is the primary agent for reducing NO2−. The initial pH exerted a significant effect on the reaction rate, and NO3− was rapidly reduced when the initial pH was less than 4. Based on the research findings, possible reaction pathways for NO3− reduction by Cu0 are proposed in this work.
High concentration of nitrate is found in wastewater from several industries, such as synthetic fiber, mineral processing, fertilizer, metal finishing, ammunitions and explosives, and nuclear industries (De Filippis et al. 2013). For example, nuclear industries produce wastewater that contains 50,000 mg-N/L nitrate from processing molybdenum-99, which is used in various medical procedures. In the uranium industry, the principal source of nitrate wastewater is uranium nitrate raffinate (UNR) in a liquid–liquid extraction process. This UNR contains up to 77,000 mg/L (Biradar et al. 2008). Thus, treatment and disposal of concentrated has become a major concern for these industries.
Zero-valent metals (ZVMs) exhibit suitable electron-donating tendencies, and they can reduce various contaminants, including . Most studies involving ZVMs have focused on Fe0 (Rodríguez-Maroto et al. 2009; Suzuki et al. 2012). Although Fe0 possesses certain advantages, it has some limitations in reducing , such as relatively high dose requirement, long reaction times, and rapid surface passivation (Kumar & Chakraborty 2006). Recent studies have examined alternative ZVMs. Luk & Au-Yeung (2002) reported that Al0 efficiently reduced , but it was unfeasible for treating industrial wastewater with high levels of (>30 mg-N/L). Kumar & Chakraborty (2006) used Mg0 to reduce (30–150 mg/L). Several studies have employed zero-valent copper (Cu0) to reduce various contaminants. For example, Huang et al. (2012, 2013) used Cu0 produced by sodium borohydride (NaBH4) and titanium citrate for dichloromethane and 1,2-dichloroethane reduction. Fanning et al. (2000) studied the reduction of by using Cu(OH)2 in the presence of NaBH4. Results of these studies have illustrated that combining Cu with NaBH4 can effectively reduce various contaminants. Although removing concentrated is a major problem, particularly for industrial wastewater, little is known about the mechanism of reduction with in situ zero-valent copper. No systematic study has been conducted on the application of Cu0 for reducing .
This study investigated the reaction rate and mechanism of reduction of concentrated solution (677 mg-N/L or 3,000 mg/L) by in situ synthesized Cu0 from CuO using NaBH4 under different copper dosage, borohydride dosage, and initial pH conditions. To the best of our knowledge, this study is the first to investigate the kinetics and mechanism of reducing concentrated by using a combination of Cu0 and NaBH4.
Previous studies show that is mainly reduced to , and N2. Even though N2 is a more desirable end product than the other two considering its benign character, has been found to constitute the majority of end products in reduction of using ZVMs (Kumar & Chakraborty 2006; Liou et al. 2012; Suzuki et al. 2012).
MATERIALS AND METHODS
All chemicals were analytical grade and were used as-received without further purification. Sodium nitrate (NaNO3; 99 + %, Acros), NaBH4 (98 + %, Acros), CuO (Showa Chemical Ltd), hydrochloric acid (HCl; 99.5%, Fisher Scientific), and sulfuric acid (H2SO4; >95%, Fisher Scientific) were used in this study. Aqueous solutions of were prepared by dissolving appropriate weights of NaNO3 with ultrapure water in 1 L volumetric flasks. The diameter of CuO was >10 μm as measured by a Zeta PALS analyzer (Brookhaven).
A three-neck flask (reactor) equipped with a magnetic stirrer was used for batch experiments. The flask was connected to an ammonia (NH3) capturing flask filled with 0.05 M H2SO4 (100 mL) to collect the vaporized NH3 gas during the reduction process. CuO was placed into the three-neck flask containing 250 mL of solution and continuously stirred (100 rpm) for 12 h at room temperature for pre-hydrolysis. The suspensions were then heated using a water bath and purged with argon gas throughout the experiment. Subsequently, NaBH4 was introduced when the temperature reached 60 °C. Periodically, 5–7 mL samples were withdrawn, filtered, and analyzed for , and dissolved Cu concentrations. For studying the effect of pH, the initial pH was adjusted to the range of 2.1–4.7 by adding 1 M of HCl. No acid was added to control the pH during the reaction. All results are shown as the mean of duplicate experiments.
and concentrations were measured using an ion chromatograph (Dionex ICS-1000). The eluent was a mixture of 3.2 mM sodium carbonate (Na2CO3) and 1.0 mM sodium bicarbonate (NaHCO3). Prior to measurement, the samples were filtered using a 0.45 μm polyvinylidene difluoride membrane (Acrodisc). was analyzed using an ultraviolet-visible spectrophotometer (Model V-500, Jasco) by the indophenol method (APHA 1995). The final pH was mostly higher than 11, which is a favorable condition for NH3 stripping. Therefore, to avoid loss of NH3 for total nitrogen calculation, the amount of NH3 reported here is the summation of its concentration in the reactor and capture column. The initial and final pH levels of the solutions were measured using a pH meter (Thermo Electron Corporation Orion 420A). Inductively coupled plasma–atomic emission spectrometry (ICP-AES, Horiba Jobin Yvon, JY 2000) was employed to detect the dissolved Cu concentration. The surface morphology and elemental analysis of the catalyst were determined using a scanning electron microscope with an energy-dispersive X-ray spectrometer (EDS; JSM 6500F, JEOL) at 15 kV. Crystal structures of CuO/Cu0 were investigated using an X-ray diffractometer (XRD; Bruker) with the Cu Kα radiation set at 30 kV and 100 mA. The data were collected from 20 to 80° in 2θ scale at a scan rate and sampling interval of 0.5° min−1 and 0.05° min−1, respectively. Crystalline phases were identified according to the Joint Committee on Powder Diffraction Standards. All of the samples for characterization were filtered and repeatedly washed with distilled water, dried, and then stored in a desiccator with silica gel for at least 24 h.
RESULTS AND DISCUSSION
Reduction of nitrate
Figure 1 shows the reduction of (677 mg-N/L) using 0.312 g/L and 4.16 g/L of CuO and NaBH4, respectively, at 60 °C. As depicted, concentrations decreased exponentially with a concomitant increase in and concentrations. The high amount of at the beginning of the reaction can be attributed to its low affinity to Cu (Liou et al. 2005). concentrations began decreasing after approximately 45 min, whereas concentrations increased approximately linearly throughout the entire experiment. The results show that the total mass of measured nitrogenous species () was less than 677 mg-N/L, indicating the presence of other nitrogen species as products, such as N2. Control experiments exhibited insignificant reduction of when Cu0 was used in the absence of NaBH4, and NaBH4 in the absence of Cu0.
Previous studies (Liou et al. 2005, 2012) have shown that the reaction rate of reduction follows a pseudo-first-order model with respect to concentration. The observed rate constant was 0.059 min−1, which is higher than zero-valent iron reported for nitrate reduction (Liou et al. 2005; Rodríguez-Maroto et al. 2009).
Effect of CuO loading and borohydride dose
Figure 2(a) presents a summary of the results of the reduction experiment with various amounts of CuO and a constant dose of NaBH4 (4.16 g/L). The reaction rate increased in conjunction with the CuO dose (Figure 2(b)), indicating that the reaction occurred on the metal surface. Table 1 shows that increasing the CuO concentration from 0.312 to 1.25 g/L resulted in a threefold increase in the pseudo-first-order rate constant (kobs). A previous study reported similar results when Cu0 was used for dichloromethane degradation (Huang et al. 2012). The final concentrations of and in each experiment were used to evaluate the reactions.
|[CuO] (g/L)||kobs (min−1)||R2|
|[CuO] (g/L)||kobs (min−1)||R2|
Increasing the amount of CuO improved the removal of and decreased the final concentration of and . At 120 min, a fraction of the total amount of measured nitrogenous species () decreased in the order of 85, 83, 79, and 77% when using 0.312, 0.624, 0.936, and 1.25 of CuO, respectively. This shows that increasing the CuO dose produced more N2, which could be attributed to the high surface availability (as reported in a previous study involving noble metal catalysts) which favors two adjacent NOx molecules that interact to form N–N bonds (Chen et al. 2003). The final percentage of after a 120 min reaction was 42, 30, and 22.3 when 0.312, 0.624, and 0.936 g/L of CuO was used. No further significant decrease was observed when CuO dose was increased again, indicating that the quantity of was insufficient for further reduction.
Figure 3 shows the effect of changing the NaBH4 dose while maintaining a constant CuO dose. The kobs values were 0.069 and 0.076 when the NaBH4 concentration was increased from 8.32 to 12.48 g/L, which is insignificant in comparison with the results from using various CuO loadings. At 120 min, the total amount of measured nitrogenous species () was 89 and 90%, and the decreased from 25.3 to 23.0% when the NaBH4 dose was increased from 8.32 to 12.48 g/L. Throughout the experiment, the concentration was lower in this case, implying that NaBH4 had a more significant effect than CuO did on reduction. To investigate this effect further, additional experiments were conducted using various doses of NaBH4 and CuO. At 120 min, 3.2% concentration was observed when 0.936 g/L and 12.48 g/L of CuO and NaBH4 were used, respectively, and the was completely removed at 180 min. At higher concentrations of NaBH4 (16.64 g/L), was completely removed within 120 min. The results indicate that the reaction products can be free of by increasing the reaction time, CuO dose, and NaBH4 dose.
Effect of initial pH
Experiments were conducted to determine the effects of the initial pH on removal with an initial concentration of 677 mg-N/L, CuO dose of 0.624 g/L, and dose of 8.32 g/L. The initial pH of the solution was adjusted to an acidic range, and another control experiment was conducted under identical conditions, but without adjusting the pH (initial pH = 6.8; Figure 4). The results show that reduction was significantly affected by the initial pH of the solution. It was found that lowering the initial pH causes an increase in the rate of reduction. Because OH− forms when the concentration of is reduced, the metal surface becomes inactive with metal oxide deposition, and the surface reactivity decreases (Choi & Kim 2009). Using Cu0 to reduce consumes a stoichiometric ratio of protons (H+). Thus, we anticipated that the reduction reaction could be improved by lowering the pH of the aqueous solution. Furthermore, acidic conditions increase the activity of Cu by removing the copper oxide layer that passivates the surface. Moreover, acid accelerates the hydrolysis reaction of NaBH4 (Liu & Li 2009), which could enhance the reduction of . However, at initial pH lower than 2.1, an increase in the accumulation of was observed. Hence, an initial pH between 2.1 and 4.7 was applied in this study. A total of 80% was removed when the initial pH was 6.8 within 25 min; however, all of it was removed within 25 min when the initial pH ranged from 2.1 to 4.7. This observation is in agreement with previous results using Fe0 to reduce (Huang & Zhang 2004). In all of our experiments, the pH increased over time, and the final pH ranged from 11.0 to 11.6. In the control run in the absence of CuO and , no removal of was found over the time period of a typical experiment. Table 2 lists the rate constants based on the pseudo-first-order model. Decreasing the initial pH from 6.8 to 2.1 increased the reaction rate by approximately 20 times, indicating that lowering the pH of the aqueous solution significantly enhances the reduction of by Cu0.
|Initial pH||kobs (min−1)||R2|
|Initial pH||kobs (min−1)||R2|
ZVMs and advantages of in situ synthesis
Although the ZVM reduction process has been known for years, it has not been applied widely to concentrated , due to the high dosage and surface passivation. The in situ synthesis of ZVM is a better way to avoid surface passivation as shown in the result of this study. From left to right in the transition-metal series, metal ion reduction becomes easier. The most favorable reduction potential is for Cu2+ (Glavee et al. 1994). Hence Cu0 can be produced easily compared to the widely used Fe0 and other ZVM with high reduction potential such as Mg0 and Zn0. Even though a rapid reduction occurred in the system when using copper, generation of a high amount of as an intermediate and production of borate can be main concerns. Previous work showed less when using other metals, such as Fe and Mg. As reported in this study, the higher amount of produced was attributed to a more active reduction and less affinity of to Cu surface, though the amount of decreased over time. Conversely, less amount of was observed in lower initial pH (<4) and high dosage of Cu and . In situ synthesis of Cu0 for the reduction of concentrated can be a promising method. Table 3 compares the advantages and disadvantages of Cu0, Fe0, and Mg0.
|Zero-valent copper (Cu0)|
|Zero-valent iron (Fe0)|
|Zero-valent magnesium (Mg0)|
|Zero-valent copper (Cu0)|
|Zero-valent iron (Fe0)|
|Zero-valent magnesium (Mg0)|
Mechanism and reaction pathway
The results of Figure 1 indicated that is an intermediate in the system, whereas and N2 are final products. This is consistent with other studies that have shown that is an intermediate of reduction (Liou et al. 2005; Kumar & Chakraborty 2006). The experimental results show that Cu0 is the main electron donor in reduction because increasing the CuO dose increased the reduction rate. Moreover, an increase in the NaBH4 dose produced less , which may result from the rapid conversion of to . This indicated that NaBH4 is the main reducing agent for reduction in the system. In addition, H2 (which was produced in the system) may be involved in the reduction of . It has been demonstrated that H2 is a strong reductant in the presence of noble metal catalysts. In the same way, but to a lesser extent, H2 reduced the concentration of in the CuO/ system and contributed, to some extent, to the reduction of . Thus, reduction may proceed in the following two parallel reactions: (1) reduction to , and then to and N2 by Cu0; and (2) reaction between and H2. Cu0 primarily served as an electron donor for reducing and was converted to Cu2+. Once the Cu0 was consumed during the reaction with , Cu2+ was immediately reduced to Cu0 by , based on the proposed reaction mechanism. This cycle continued throughout the experiment, as confirmed by the XRD and EDS results. The hydrogen from the hydrolysis of was used for the reduction. Based on the above discussion and results from analysis and characterization, the mechanism of nitrate reduction is illustrated as a conceptual model in Figure 5.
In situ synthesized Cu0 produced by NaBH4 was used to reduce concentrated . Complete reduction of (677 mg-N/L) with CuO (0.312 g/L) and NaBH4 (4.16 g/L) at 60 °C was observed within 55 min. The good fit of a linear model to the data (R2 > 0.97 in all cases) provides strong evidence that the reaction is pseudo-first-order with respect to concentration. The experimental results and subsequent data analysis show that the reduction of depends on the CuO load and initial pH. Increasing the NaBH4 dose produced less , implying that it is the primary agent for reducing . The results of solid characterization show the sole presence of Cu0 in the reaction of CuO with NaBH4. In all of the experiments, the concentration of copper ions in the solution was less than 0.18 mg/L, which is considerably low. In summary, the combined use of CuO and is effective in reducing concentrated .
The authors are grateful for the financial support from the Institute of Nuclear Energy Research, Atomic Energy Council, Taiwan (contract no. NL-1020163).