Electrodeposited NiCu bimetal on carbon paper as stable non-noble anode for efficient electrooxidation of ammonia

Electrochemical remediation of ammonia-containing wastewater at low cell voltage is an energyeffective technology which can simultaneously recover energy via hydrogen evolution reaction. One of the main challenges is to identify a robust, highly active and inexpensive anode for ammonia electrooxidation. Here we present an alternative anode, prepared by electrochemical co-deposition of Ni and Cu onto carbon paper. This NiCu bimetallic catalyst is characterised by scanning electron microscope, scanning transmission electron microscope, X-ray diffraction, x-ray photoelelectron spectroscopy, cyclic voltammetry, linear sweep voltammetry and chronoamperometry techniques. The stability and activity of NiCu bimetallic catalyst are largely improved in comparison with Ni or Cu catalyst. Moreover this noblemetal-free NiCu catalyst even performs better than Pt/C catalyst, as NiCu is not poisoned by ammonia. An astewater i-Cu bimetal ammonia electrolysis cell is fabricated with NiCu/carbon paper as anode for ammonia electrolysis. The influences of pH value, applied cell voltages and initial ammonia concentration on cell current density, ammonia removal and energy efficiency are tested. An ammonia removal efficiency of ∼80% and coulombic efficiency up to ∼92% have been achieved. Ni-Cu bimetal on carbon paper is a stable non-noble anode for efficient electrooxidation of ammonia. © 2016 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license


Introduction
The discharge of ammonia-containing wastewater into natural environment can damage ecological balance, which is known as eutrophication [1,2]. Normally biological methods are widely used for nitrogen removal in wastewater with the complex nitrification-denitrification processes [3,4]. The biochemical chain of nitrogen removal includes two steps, aerobic oxidation NH3→NO -2 → NO -3 and anaerobic reduction NO -3 →NO -2 →N2 [5]. Although biological treatment is a cost-saving way, it usually takes a long time. Besides, additional carbon source (e.g. methanol) is required when the carbon/nitrogen ratio of wastewater is low [6]. Many wastewater sources such as industry and landfill leachate are produced with a high value of ammonia concentration but a low value of C/N ratio [7,8], which limit the application of biological treatment.
Electrochemical treatment of ammonia has attracted more and more attention because it is easyoperation, environmental-friendly and applicable to tough conditions [9][10][11][12]. Ammonia can be directly converted to nitrogen gas at anode through a three-electron oxidation reaction [13][14][15].
Compared to biological method, electrooxidation of ammonia needs less step and simpler setup.
Furthermore, hydrogen gas can be generated at cathode according to Reaction (2). In this way, the chemical energy in ammonia is retrieved in the form of hydrogen energy during ammonia electrolysis.
The generated hydrogen can be utilized to produce power by combustion, fuel cells and hydrogenforced engines, compensating for parts of the energy consumption in ammonia electrolysis [16][17][18].
Theoretically, only a small value of external energy (0.06 V) is required to crack ammonia to nitrogen and hydrogen [15]. It is much lower than the energy (1.23 V) required by water electrolysis: Water electrolysis: However, ammonia electrooxidation at anode is reported to be sluggish and have a large overpotential, thus it is important to develop high-performance electrocatalysts in order to solve this fundamental challenge [19]. Platinum and Pt-based catalysts are most active with small overpotential, but they are less affordable and easily poisoned by the adsorbed Nads, and thus limiting the electrolysis current [20][21][22]. Transition metal Ni and Ni(OH)2 based anode catalyst is a promising and inexpensive material for ammonia electrooxidation in alkaline condition [13,23]. Nevertheless they can be corroded and deactivate in ammonia solution resulting in a secondary pollution of metal ions [23,24]. Besides, copper also has high activity toward ammonia electrooxidation theoretically, which is comparable to Pt [22].
However it binds N atom too weakly so that it cannot catalytically oxidize ammonia experimentally.
One strategy to enhance activity and stability of metal catalyst is to incorporate another element into pure metal. Bimetallic NiCu catalyst is reported to have improved activity for methanol electrooxidation and water electrolysis [25][26][27][28], as well as enhanced stability [29]. In this work we present a method of electrodeposition of NiCu bimetal on carbon paper for ammonia electrooxidation. Carbon paper is used as supporting substrate for NiCu catalyst, because it has large surface area, high electric conductivity and chemical stability in alkaline ammonia electrolytes. NiCu coated carbon paper electrode is further used as noble-metal-free anode to fabricate ammonia electrolysis cell (AEC). This AEC can work under low cell voltage to achieve an energy-effective degradation, which is especially capable of ammoniarich wastewater.

Catalyst characterization
The surface morphology of the as-obtained NiCu/CP electrode was studied by scanning electron microscope (SEM) (Zeiss SUPRA 55-VP) equipped with an energy-dispersive X-ray (EDX) spectrometer that allows elemental composition analysis. Compositional analysis was carried out using a FEI Titan Themis scanning transmission electron microscope (STEM) equipped with a Super-X windowless EDX detector system. Samples for STEM characterisation were prepared by dispersing specimen powder on a carbon film supported on a gold grid. The X-ray diffraction (XRD) patterns of NiCu/CP were recorded with Panalytical X-Pert Pro MPD with Cu Kα1 radiation. The x-ray photoelelectron spectroscopy (XPS) data were collected using an Omicron Multiprobe with the sample being illuminated using an XM1000 monchromatic Al kα x-ray source (Omicron Nanotechnology).
Electrochemical characterization of the NiCu/CP electrode was determined by cyclic voltammetry (CV), linear sweep voltammetry (LSV) and chronoamperometry (CA) techniques conducted by Solartron 1287A Electrochemical Interface in a three-electrode cell. A platinum foil and an Ag/AgCl electrode (saturated KCl) were used as the counter and reference electrode, respectively. The CV measurements, from 0.05 V to 0.7 V with a scan rate of 25 mV s -1 , were recorded after at least 5 cycles till stable results. The sweep rate of LSV is 1 mV s -1 . Electrochemical impedance spectroscopy (EIS) measurements were carried out by a Solartron 1287A/1250 in the frequency range of 60 kHz to 0.01Hz at a bias of 0.55 V.

Fabrication of ammonia electrolysis cell and measurements
Carbon supported Pt (Pt/C) particles were dispersed in a mixture of Nafion ® solution, isopropanol and deionized water under ultrasonic to get catalysts ink. Then it was brushed onto carbon paper (2 cm 2 ) to prepare cathode of AEC for hydrogen evolution reaction. The Pt/C was also used as anode for comparison. The Pt loading on carbon paper was 0.3 mg cm -2 . NiCu/CP without any pre-activation was directly used as anode of AEC in a sealed glass cell containing 11 mL electrolyte. The initial ammonia

Characterization of NiCu/CP electrode
The XRD patterns of pure carbon paper and NiCu/CP were shown in Figure 1 in order to identify the corresponding crystalline structures of the catalyst. Pure carbon paper had strong peak at 2θ angle of 54º. Typical Ni(111) and Cu2O(111) peaks were observed in the Ni/CP and Cu/CP patterns when sole Ni or Cu element was electrodeposited. However Cu(111) peak was too weak in the Cu/CP patterns to be seen. It indicated that the main deposition product was Cu2O rather than Cu in Cu/CP. When Ni and Cu were co-electrodeposited on carbon paper (NiCu/CP), the Cu2O(111) peak became very weak and Cu(111) and Cu(200) peaks were newly observed at 2θ angles of 43º and 50.3º, respectively. Besides, Ni(111) and Ni(200) peaks were also found at 2θ angles of 44.5º and 52.1º in NiCu/CP patterns. The Ni(111) peak of NiCu/CP was in accordance with that of Ni/CP, and no peak shift was observed, indicating that the NiCu catalyst synthesised on CP by electrochemical co-deposition was bimetal that was composed of two separate metals rather than alloy.

Electrochemical properties of the NiCu/CP electrode
Electrochemical measurements were conducted to further characterize the activity and stability of NiCu/CP electrode. Figure 3a demonstrated the CVs of NiCu/CP electrode in 0.5 M NaOH electrolyte with the presence and absence of 55 mM NH3 respectively. A pair of redox peaks were generated in the CV plots of NiCu/CP in 0.5 M NaOH, which were attributed to the transformation between Ni(II) and Ni(III) species [30,31]. When ammonia was added, a dramatic increase of anodic current density with onset potential about 0.47 V vs. Ag/AgCl could be observed. The anodic current density reached 52 mA cm -2 at 0.7 V vs. Ag/AgCl. It demonstrated that NiCu catalyst had obvious activity toward electrooxidation of ammonia.
For comparison, CVs of Ni/CP and Cu/CP were also carried out which are presented in Figure S3.
The black curve in Figure S3a was the CV curve of Ni/CP in 0.5 M NaOH, which was similar to that of NiCu/CP with a pair of redox peaks. However, there was no obvious change in the red curve of Figure S3a when ammonia was added, which meant that metal Ni alone did not show catalytic activity in this potential range. This was possibly due to the large overpotential of ammonia electrooxidation on Ni catalyst. In fact Ni catalysts were based on the β-Ni(OH)2↔NiOOH mechanism in electrocatalytic oxidation reaction [32][33][34]: NiOOH would be formed from β-Ni(OH)2 via electrochemical reaction (6) to catalyse the NH3 oxidation and get reduced back to β-Ni(OH)2. According to reaction (7), NiOOH was the real activated substance.
Thus Ni catalyst had to be pre-activated with a transformation routine of Ni → α-Ni(OH)2 → β-Ni(OH)2 ↔ (β, γ)-NiOOH in alkaline solution [9,10,23]. Moreover, Cu/CP without Ni element was found to have a negligible activity, which demonstrated a low anodic current density (5 mA cm -2 at 0.7 V vs. Ag/AgCl) as shown in Figure S3b. This was because Cu bound the N atom in ammonia too weakly as mentioned in introduction, which could be improved by Ni doping. to be used as anode at this potential, in accordance with the CV results. Higher voltage had to be applied to achieve a desired current density in AEC when using Ni/CP and Cu/CP electrodes, but in that case oxygen evolution reaction might occur and energy efficiency would be very low. This outcome indicated that the NiCu/CP electrode had much higher electrocatalytic activity than that of either Ni/CP or Cu/CP electrode for ammonia oxidation, demonstrating its deep potential to be used as electrocatalyst with enhanced performance. The multi-CVs in 0.5 M NaOH+55 mM NH4Cl shown in Figure S4 further testified the stability of NiCu/CP electrode. From the 1 st cycle to the 50 th cycle, there was an obvious increase of both oxidation and reduction current density. This was attributed to the further activation of NiCu catalyst by possibly formed NixCu1-xOOH double oxyhydroxides. Although the copper oxidation/reduction reaction might also happen, it was to unable to observe the separated peaks corresponding to these reactions, respectively, implying that Ni, Cu peaks and ammonia oxidation current were partially overlapped. From the 50 th cycle to 100 th cycle, there was only a slight difference in current density, indicating the strong stability of NiCu/CP in ammonia electrooxidation.
From the XRD patterns of NiCu/CP after electrochemical tests in Figure S5, new β-Ni(OH)2 phase was observed, indicating that NiCu bimetallic catalyst might gradually transform to the isomorphous compound as β-Ni(OH)2 under alkaline condition. In order to investigate the surface composition change of NiCu catalysts after electrochemical experiments, XPS was conducted as shown in Figure 4.
Analysis of the Ni 2p and Cu 2p regions was conducted using the work of Biesinger and co-workers as a guide [35,36]. The surface metals of the as-prepared NiCu catalysts before electrochemical tests were oxidized to metal oxides and hydroxide in air according to XPS data in Figure 4a and Figure 4b. The oxide and hydroxide phases were not observed in XRD because the signal was too weak to be picked up by X-ray or, they are in amorphous state. Oxygen was also picked up by high angle annular dark field (HAADF) imagines generated by STEM ( Figure S2) indicating the existing of oxide and/or hydroxide, likely on the surface as the major phase are still poorly crystallised Ni and Cu as shown in Figure 1. After electrochemical tests, Ni oxyhydroxides (NiOOH) was newly formed via Reaction (7) as shown in Figure 4c, and more Ni(OH)2 was observed. Similarly, in Figure 4d, most of Cu2O was transformed to Cu(OH)2 in alkaline solution after tests. The transformation of Cu(I) to Cu(II) was more clearly proved in Cu LMM Auger spectra in Figure 4e. The "before" sample had a relatively sharp peak at a kinetic energy of 916.6 eV, whereas the "after" sample seemed more broad and has a peak centred at 917.1 eV. This was due to a shift from a mix of Cu2O and Cu(OH)2 to being predominantly Cu(OH)2.
Note that there was also a general shift in the Ni and Cu binding energies (between Figure 4a  It was reported that Fe incorporating into Ni-based catalysts during the crystallization process would greatly improve the catalysts activity and lower the anode overpotential toward water oxidation reaction [37,38]. This effect might be due to the greatly improvement of conductivities of NiOOH and increase of active sites after Fe dopping [37,38]. In this experiment, Cu incorporating was possible to have a similar effect as Fe to form NiCu double oxyhydroxides and enhance the activity by increasing active sites and conductivity [39]. where ip was peak current (A cm -2 mg -1 ); n was the number of electrons involved in the reaction (1); A was the ECSA; C was the reactant concentration (10 -5 mol cm -3 ), D was diffusion coefficient (6.3×10 -6 cm 2 s -1 ), v was scan rate (0.05 V s -1 ). The ECSA of NiCu/CP reached 18.4 cm 2 mg -1 , which was much larger than that of Ni/CP (8.6 cm 2 mg -1 ). Higher ECSA might lead to an enhancement of catalytic activity. In addition, the catalyst conductivities could also affect the catalytic activity. In order to test the conductivity, EISs of NiCu/CP and Ni/CP were conducted as shown in Figure S7. The catalytic activities of NiCu/CP prepared at different potentials were tested by CVs in Figure S8. It was demonstrated that the current density increased when lower potential was applied, but the using Pt/C had been decreasing from 8.5 mA cm -2 to 5 mA cm -2 (in Figure S10), indicating poor stability when compared to NiCu catalyst (in Figure 3b). This was mainly because that Pt had a potential to be poisoned by the Nads due to the very strong bond between Pt and N atoms, and thus limiting the current density [19,22,40]. In terms of stability and activity for electrooxideation of ammonia, the NiCu bimetal is better than Pt/C.

Ammonia electrolysis at different pH values
The bulk electrolysis of 500 ppm ammonia at pH=8, 10, 12 in AEC was conducted during 16 h at constant cell voltage of 1.1 V. In a 16-h experiment, current density was recorded and shown in Figure   6a. All the current densities of the three tests decreased along with electrolysis time, owing to the consumption of ammonia in bulk electrolytes. However, the current responses varied greatly with the pH condition. Higher initial pH value of electrolyte was observed to have a higher current density, leading to faster reaction rate. This highly pH-dependent phenomenon might be caused by the difference of the NiCu catalyst activity. As the activity of NiCu catalyst was based on the formation of NixCu1-xOOH in basic condition, a high pH value would facilitate the transformation of NixCu1-x(OH)2 to NixCu1-xOOH to improve catalytic activity toward ammonia electrooxidation. What's more, the redox potential of ammonia oxidation in Reaction (1), was related to pH value according to Nernst equation: E= -0.77 + 0.0592(14-pH). High pH value would lead to a negative potential (E) and make ammonia electrooxidation reaction easier to proceed in terms of thermodynamics. This was in accordance with the reported work about electrooxidation of ammonia on Ni-based catalyst which shown poor current responses when pH was below 8 [23]. Thus higher pH value could diminish the anode polarization to enhance the current output. The influence of pH on Pt/C anode was similar to NiCu/CP anode as shown in Figure S11a. On the contrary, higher pH value would lead to a larger polarization on Pt/C cathode as shown in Figure S11b, but this disadvantage was negligible as cathodic reaction was much easier to reach enough current density with low overpotential when compared to ammonia electrooxidation reaction (anodic reaction)..
Ammonia was successfully removed from water during the electrolysis and the concentration change with time was shown in Figure 6b. Normally there was a positive correlation between AEC current density and ammonia removal rate. The electrolysis at pH=8 demonstrated a much low current density and ammonia removal efficiency than those at pH=10 and 12. During a 16-h test, the ammonia concentration with pH=8 decreased from 500 ppm to ~450 ppm, achieving only 10% efficiency. When the pH rose to 10, the final ammonia concentration reached 210 ppm with a removal efficiency of 58%.
For pH=12, the ammonia concentration reduced quickly in the initial 5 hours, and then decreased gradually to 108 ppm at 16 hours. However, more NO -3 tended to generate in higher pH condition.
Nitrate ions were detected after ammonia electrolysis, and the results were shown in Table S1. In the test of pH=12, 14.6% of the degraded NH3 was recovered as NO -3 . In comparison, the conversion percent of NH3 to NO -3 was only 6.1% and 7.9% when the pH was 8 and 10, respectively.

Ammonia electrolysis at different cell voltages
The be oxidation reaction of carbon paper electrode at high potential: C+2H O CO +4H +4e  E=0.207 V vs. SHE (9) As shown in Figure 7b, the coulombic efficiency was only 53.1% when the applied cell voltage was 1.4 V, indicating that a portion of generated current was due to the oxygen evolution and carbon oxidation reaction. The ammonia electrolysis at lower cell voltage would lead to a higher coulombic efficiency and was a more energy-efficient way to remove ammonia.

Electrolysis with high concentration of ammonia
One of the advantages of electrochemical ammonia treatment is that it can work well at high ammonia concentration conditions. As shown in Figure 8a, the current density of AEC increased with the ammonia concentration up to a very high value of 1100 ppm. One reason might be that high concentration could improve the mass transfer, as the only mass transfer force in this batch model AEC is the concentration gradient between bulk electrolyte and electrode surface. In addition, there was a competition of adsorbed species such as OHon the catalyst surface. More ammonia percent in solution could lead to a higher coverage of catalyst surface by ammonia, thus enhancing ammonia electrooxidation. In comparison, increasing ammonia concentration could not lead to much higher current density of anode reaction when using Pt/C anode (in Figure S12a). In the aspect of Pt/C cathode reaction, higher ammonia concentration could also enhance cathodic current density due to the improvement of electrolyte conductivity (in Figure S12b). Figure 8b demonstrated that ammonia was successfully degraded at high concentration condition. For example, after 14-h experiment, the ammonia concentration was reduced to 220 ppm with initial value of 1100 ppm, reaching a removal efficiency of 80%. Besides, the anodic coulombic efficiency also increased a bit with the higher ammonia concentration, as shown in Figure S13.

Conclusion
Ammonia is regarded as a kind of tough contaminant especially for the high-content wastewater. On the contrary, the AEC equipped NiCu/CP electrode is proved to be suitable for ammonia-rich wastewater due to the improved mass transfer and catalyst surface coverage. NiCu bimetal was prepared by an electrochemical deposition method. XRD, XPS and STEM study indicate the bulk is NiCu bimetal while the surface of the catalyst was oxidised to metal oxide and hydroxide. Experiments show high removal efficiency of around 80% is achieved at both low and high initial concentration (from 500 ppm to 1100 ppm). This AEC could work at low cell voltage to avoid the competitive reaction of water electrolysis and save energy. High coulombic efficiency up to 92.8% is realized at applied cell voltage of 1.0 V. The pH value could largely influence the current density of AEC as well as the formation of nitrate. More NO3ions were formed at higher pH value and higher applied voltage. Therefore low applied voltage is a better choice.
By electrochemical co-deposition of Ni and Cu, the current density is improved by more than 10 times compared to Ni and Cu according to the electrochemical characterization. Higher ECSA was observed for NiCu/CP than that for Ni/CP. In addition, this NiCu bimetallic catalyst demonstrates a strong stability with no poison during the electrolysis which is a big advantage over Pt/C electrode for electrooxidation of ammonia. Therefore it is a promising noble-metal-free catalyst for ammonia electrooxidation. Further improvement and reduction of cost of AEC can possibly be achieved by optimising the cell configuration and the application of non-noble cathode.