Graphdiyne-based metal atomic catalysts for synthesizing ammonia

Abstract Development of novel catalysts for nitrogen reduction at ambient pressures and temperatures with ultrahigh ammonia (NH3) yield and selectivity is challenging. In this work, an atomic catalyst with separated Pd atoms on graphdiyne (Pd-GDY) was synthesized, which shows fascinating electrocatalytic properties for nitrogen reduction. The catalyst has the highest average NH3 yield of 4.45 ± 0.30 mgNH3 mgPd−1 h−1, almost tens of orders larger than for previously reported catalysts, and 100% reaction selectivity in neutral media. Pd-GDY exhibits almost no decreases in NH3 yield and Faradaic efficiency. Density functional theory calculations show that the reaction pathway prefers to perform at the (Pd, C1, C2) active area because of the strongly coupled (Pd, C1, C2), which elevates the selectivity via enhanced electron transfer. By adjusting the p–d coupling accurately, reduction of self-activated nitrogen is promoted by anchoring atom selection, and side effects are minimized.


INTRODUCTION
Ammonia (NH 3 ), an indispensable raw material in industrial production, has always occupied an important position in basic chemical industry, and is also an important source of chemical fertilizer in agricultural production [1][2][3]. Unfortunately, traditional production of NH 3 must operate at high pressures and temperatures, which is very energyconsuming. Electrochemical catalytic nitrogen reduction reaction (ECNRR) in aqueous electrolytes at ambient conditions provides an ideal method for clean and efficient conversion of N 2 to energy-rich NH 3 [4][5][6][7]. However, there are some key scientific and technical issues with this approach, and the NH 3 yield (Y NH3 ) and Faradaic efficiency (FE) are still very low. To solve these major issues and promote efficient conversion of N 2 to NH 3 , efficient catalysts that can drive ECNRR at low overpotential with high selectivity, Y NH3 , FE and stability are in great demand.
Electrocatalysts comprising singly-dispersed metal atoms and supporting materials have shown high catalytic activity and selectivity in various reactions, because of their atomically distributed active sites [8][9][10][11][12][13][14][15][16][17][18][19][20][21]. Although the rapid development of single-atom catalysts has introduced many new opportunities in terms of new catalytic science, knowledge and concepts, there remain some important scientific issues to be solved, such as the precise structure of the support materials, the valence of the supported metal atoms and the structure, the high dispersion distribution of the metal single atoms and so on.
Very recently, Li and coworkers reported the first zero-valent atomic catalysts (AC) [20,21]. An important finding that has drawn great attention is that zero-valent metal atoms can be anchored on graphdiyne surfaces [22][23][24][25][26][27][28][29][30][31]. The AC exhibits high stability and catalytic activity and represents the emergence of a new generation of catalysts, and there is a drive to understand the mechanism, reaction process and properties of atomic catalysis. In this study, we demonstrate that highly selective and active NH 3 production can be achieved using a zero-valence atom catalyst based on graphdiyne, Pd-GDY, to activate the N 2 and react with water at room temperature and pressure. Experimental and theoretical results solidly confirm the unique chemical and electronic structures and zero-valence state of this electrocatalyst. The strong orbital interactions between Pd atoms and neighboring C sites lead to a strong electronegative reduction character for ECNRR. The downshifted s-band from the electronic structure arises from the elimination of non-bonding lone-pair N2-2s orbitals through N-hydrogenation to suppress HER electronically. The designed Pd-GDY AC presents significantly improved ECNRR performance compared with previously reported catalysts. This study could provide a promising strategy for designing and synthesizing highly efficient electrocatalysts for producing NH 3 at room temperature and ambient pressures.

RESULTS AND DISCUSSION
Pd-GDY was synthesized through a self-reduction strategy. Figure 1 shows the synthesis and reusability schematic of Pd-GDY for the electrochemical nitrogen reduction reaction. As can be seen from the scanning electron microscopy (SEM) images (Supplementary Fig. S1), graphdiyne fibers with porous surface interweave in a three-dimensional manner forming three-dimensional flexible electrodes.
No Pd particles or clusters can be observed in the SEM ( Fig. 2a and Supplementary Fig. S2), high-resolution transmission electron microscopy (HRTEM) ( Fig. 2b and Supplementary Fig. S3),   Supplementary Fig. S4) images. Energy-dispersive X-ray spectroscopy mapping results (Fig. 2c) reveal the uniform distribution of Pd and C atoms in the sample. In the HAADF images, the bright dots gave an average size of 0.36 ± 0.01 nm ( Supplementary Fig. S5), confirming that each dot is an individual Pd atom. The possible anchoring sites of Pd atoms on GDY are shown in Supplementary Fig. S6, as revealed by our detailed computational calculations in which all possible placements of the Pd on GDY were fully examined. The optimal anchoring site is determined based on comparison of formation energies between different anchoring sites, in that the lowest formation energy of the anchoring site of −0.99 eV as the presented position will be the optimal placement for Pd on GDY. Inductively coupled plasma optical emission spectroscopy (ICP-OES) measurements showed an average mass loading of 1.02 ± 0.04 wt.%. Figure 2h and i shows the X-ray absorption nearedge structure (XANES) profiles for Pd-GDY along with corresponding reference samples (Pd foil and PdO). For the Pd K-edge spectrum, the absorption threshold position of Pd-GDY is located at the same position as Pd metal but a more negative position than PdO (Fig. 2h). In addition, the first derivative XANES for Pd-GDY is similar to that of metallic Pd (Fig. 2i). These results strongly indicate that the valence states for the Pd atoms in Pd-GDY metals are zero. The extended X-ray absorption fine structure (EXAFS) spectrometry results ( Supplementary  Fig. S7) showed that there was only one peak at around 1.5 to 2Å arising from the Pd∼C contribution, and that no peak contributing to the Pd-Pd (around 2∼3Å) could be observed. These demonstrate that the Pd atoms are individually anchored on GDY and exhibit zero-valence state ( Supplementary  Fig. S8).
We used X-ray photoelectron spectroscopy (XPS) and Raman spectra to characterize the chemical composition and quality of as-synthesized samples. The XPS survey spectra show the only existence of carbon ( Supplementary Fig. S9), except for oxygen signal from the adsorption of air. The C 1s XPS spectrum of Pd-GDY can be divided into sp 2 -C (284.5 eV), sp-C (285.1 eV), C=O (288.3 eV), COO (286.7 eV) and π -π transition (291.2 eV) peaks ( Supplementary Fig. S10a). Compared with pure GDY (Supplementary Fig. S10b), a newly formed peak at 291.2 eV was observed, which can be attributed to the electronic transitions induced by the anchoring of Pd atoms. The ratio of intensity of sp 2 /sp-carbon is 0.5, revealing the integrity of the GDY structure in Pd-GDY. As shown in Supplementary Fig. S11, the Pd-GDY shows a lower valence band (VB) with a smaller energy at 1.73 eV than that of pure GDY (2.26 eV), indicating the enhanced conductivity of Pd-GDY. Raman spectra provide information on the structural changes of carbon materials. The intensity of the D and G bands of GDY and Pd-GDY changed ( Supplementary  Fig. S12). The D band to G band intensity ratio increased from 0.73 (GDY) to 0.82 (Pd-GDY), which suggests the presence of more defective sites in Pd-GDY, indicating the formation of more active sites, which have been demonstrated to be helpful for improving catalytic activity [32,33]. Based on the above inspiring findings, the EC-NRR experiments were carried out in 0.1 M Na 2 SO 4 (pH = 7) aqueous solution at room temperature and atmospheric pressure using an H-type electrolytic cell separated by the Nafion 117 membrane ( Supplementary Fig. S13). The catalyst loading of Pd-GDY is 2.7 μg metal cm −2 , and the geometric area is 2.0 cm 2 . The NH 3 was determined using an indophenol blue method [34] and calibration curves with reliable sensitivity and good linear relationship were obtained ( Supplementary Fig. S14). The ECNRR of Pd-GDY starts at around 0.2 V versus RHE in N 2 -saturated 0.1 M Na 2 SO 4 solution ( Supplementary Fig. S15). Only NH 3 (no byproduct of N 2 H 4 ) can be detected in this work (Supplementary Figs S16 and S17), revealing the ultra-high selectivity of Pd-GDY in the ECNRR process. The NH 3 yielding rate (Y NH3 ) and the Faradaic efficiency (FE) rise with increasing cathodic potential until −0.16 V versus RHE ( Fig. 3a and b), at which the highest Y NH3 of 4.45 ± 0.30 mg NH3 mg Pd −1 h −1 (1.97 ± 0.13 × 10 −10 mol cm -2 s -1 , normalized by geometric surface area) and FE up to 31.62 ± 1.06% were achieved (Supplementary Table S1). These values are larger than for all previously reported ECNRR catalysts working under ambient conditions including Au/TiO 2 (Y NH3 = 21.4 μg NH3 mg cat.
−1 h −1 , FE = 29.6%) [37] and Ru NC (Y NH3 = 3.6 mg NH3 mg cat. −1 h −1 , FE = 21%) [38], and Pd-based EC-NRR catalysts such as Pd/C (Y NH3 = 4.9 μg NH3 mg cat. [41], and even compared favorably with those working at higher temperatures/pressures ( Fig. 3c; Supplementary Tables S2 and S3) [42]. With further increase of the negative potentials, the Y NH3 and FE decreased sharply, which can be attributed to competition between EC-NRR and HER [21,43]. Control experiments were performed to determine the origin of the detected NH 3 . The pristine GDY yielded limited NH 3 (Y NH3 : 5.44 × 10 -12 mol cm -2 s -1 ; FE: 1.65%; Supplementary Fig. S17a,b) and CC substrate yielded almost no NH 3 ( Supplementary Fig. S17a). No NH 3 can be detected in Ar-saturated electrolyte (Supplementary Fig. S17c) and at the open-circuit potential ( Supplementary Fig. S17d). 15 N-labelling experiments were performed using 15 N 2 as the feeding gas to confirm the ammonium formed from the reduction of N 2 . Only 15 NH 4 + (doublet peak) was observed from the 1 H nuclear magnetic resonance ( 1 H-NMR) results (Fig. 3d). These results revealed that all detected NH 3 comes from reduction of N 2 molecules by the Pd-GDY catalyst. The fact that the Y NH3 and FE of Pd-GDY are almost 10 and 3 times larger than those of palladium nanoparticle-modified GDY (Pd NP/GDY, Supplementary Fig. S18) demonstrates the superiority of isolated zero-valence Pd atoms over bulk Pd nanoparticles toward efficiently catalyzing ECNRR (Fig. 3e). In addition to excellent catalytic activity and selectivity, long-term stability is another essential criterion for an electrocatalyst in practical applications. It was observed that the NH 3 yield rate and FE of Pd-GDY exhibited no obvious change after six successive catalytic cycles (Fig. 3f, Supplementary Fig. S19), and the total NH 3 yield increased linearly with reaction time (Supplementary Fig. S20). Comprehensive characterizations ( Supplementary  Fig. S21) on the sample obtained after stability test showed good preservation of the morphology and structure, revealing its robust nature.
Electrolytes are considered to have significant effects on the ECNRR selectivity and activity [39,[43][44][45]. Neutral and basic electrolytes with limited proton transfer rate can effectively improve the ECNRR selectivity. However, in acidic electrolytes, the ECNRR selectivity could be kinetically limited because of the high availability of protons; the rate of the HER increases significantly with the increase of applied potentials, in which most protons or H 2 O are converted to H 2 rather than NH 3 , resulting in great decreases in reaction selectivity. Efforts have recently been devoted to improving ECNRR performance in acidic electrolytes [37,44,[46][47][48]. However, the Y NH3 and Faradaic efficiencies of reported ECNRR electrocatalysts are still very low (Supplementary Tables S2 and S3). Because of these limitations, it is of significant importance to develop an electrocatalyst with high selectivity and activity in acidic electrolytes for efficient ECNRR.
We then studied the ECNRR activity of Pd-GDY in 0.1 M HCl solution (Fig. 4a, Supplementary  Fig. S22). NH 3 and N 2 H 4 were determined using a spectrophotometry method, and the calibration curves are shown in Supplementary Fig. S23. Pd-GDY also shows high selectivity toward formation of NH 3 (without N 2 H 4 formation) in acidic media ( Supplementary Figs S22b and S24). As shown in Fig. 4b Table S1). The 15 N isotope labelling experiments confirmed that the NH 3 was formed from reduction of N 2 (Fig. 4c), which indicates that the Pd-GDY is highly selective and active toward ECNRR in acidic media. The FE value keeps decreasing as the applied potentials become more negative, which can be attributed to competition from the HER in acidic media. The rapid increase of current densities in polarization curves recorded in both N 2 -and Ar-saturated 0.1 M HCl (Supplementary Fig. S25) also gives supportive evidence for this major interference. Despite all this, the ECNRR activity of Pd-GDY in acidic conditions still compared favorably with those of most previously reported catalysts (Supplementary Tables S2 and S3) such as MoS 2 (Y NH3 = 8.48 -11 mol cm -2 s -1 , FE = 0.096%). In addition, both Y NH3 and FE of Pd-GDY show negligible degradation after six consecutive tests, confirming reliable stability of Pd-GDY in acidic media (Fig. 4d). Electrochemical impedance spectroscopy (EIS) was conducted to offer more insight into the catalytic behaviors. The impedance data were recorded and analyzed using a R(QR)(QR) equivalent circuit model (Supplementary Table S4). From Fig. 4e, Pd-GDY shows the lowest solution resistance (R s , 10.41 ) and charge transfer resistance (R ct , 3501 ) over GDY (R s , 15.70 ; Rct, 4658 ) and CC (R s , 26.5 ; Rct, 8933 ), indicating a more kinetically favorable ECNRR for Pd-GDY than others. Electrochemically active surface area (ECSA) was further estimated by determining the electrochemical double-layer capacitance (C dl ) through a cyclic voltammetry method ( Supplementary Fig. S26). The C dl value of Pd-GDY is 2.1 mF cm -2 (Fig. 4f), which is larger than those of pure GDY (1.7 mF cm -2 ) and CC (1.3 mF cm -2 ), implying an increase of the active site number in the Pd-GDY catalyst.
The origin of the outstanding catalytic activities of the samples was further studied with computational calculations (see Methods for details). The bonding and anti-bonding orbitals near the Fermi level (E F ) are demonstrated for the cases of Pd-GDY, HER and ECNRR. We found that the (Pd, C1, C2) sites are electron-rich regions, exhibiting the localized p-d coupled electronic orbital. The variation of the real-spatial distribution of charge density distribution indicates site-to-site charge migration and redistribution between the Pd-GDY and N/H-species (Fig. 5a). The interplay of Pd-4d and N 2 -2p orbitals was illustrated with the projected partial density of states (PDOSs). We found that the preference of N 2 -adsorption on the Pd-GDY interacting with Pd-4d orbitals indeed follows the farthest p-d separation rather than p-d orbital overlapping. This is evidently different from the adsorption preference of O-species. Such an anomalous trend implies that the on-site effective screening Coulomb repulsion potential is a determining factor existing between Pd-4d 10 and N 2 -(2s 2 , 2p 3 ). This arises because the non-bonding lone-pair N 2 -2sorbitals are highly active, participating within the ECNRR process for hydrogenation bonding. The fulfilled Pd-4d 10 is an electron-rich center for evident site-to-site charge migrations, exhibiting substantially strong electronegative activity (ENA). Such strong ENA induces high chemical potential contrast for favorite directional electron transfer, which favors electron transfer from Pd onto N-species for N-hydrogenation rather than HER. We further interpret the underlying electronic mechanism to redirect the HER-suppression trend. The p-d coupled effective negative correlation energy matters to the subtle interplay between on-site Coulomb repulsion and ENA, which overcomes on-site Coulomb repulsion and is energetically favorable to accumulate ENA (Fig. 5a). The Pd-4d band center, Pd-4d band dominant peak, and splitting gap between bonding and anti-bonding (BA) of N 2 -2pband all exhibit the same trend to monotonically decrease with N 2 -adsorption energy increases. The deep localized Pd-4d 10 -t 2g component strongly couples with the N 2 -2p band, resulting in a narrowing of the N 2 -2p band BA splitting gap (Fig. 5b). We further demonstrate the s-and p-band variations among the N-hydrogenation process. The downshifted s-band arises from the elimination of non-bonding lone-pair N 2 -2s orbitals through N-hydrogenation. The N-hydrogenation performs 0.8 eV more than the H-adsorption at the C site. Therefore, the original HER performance has been electronically suppressed (Fig. 5c). Considering the contribution of individual C sites within Pd-GDY, both C1 and C2 sites clearly show a strong 2p electronic state at E V -5.0 eV as relay-center-like level promoting electron transfer between Pd and (C1, C2). The evident bonding and anti-bonding splitting feature at the C1 site confirms that the strong Pd-C1 interaction is indeed enhanced and further stabilized via charge transfer (Fig. 5d). The PDOSs of Pd-4d and (C1, C2)-2p orbitals presented a large overlap near the E F , and the two dominant BA orbitals of (C1, C2) coupled the Pd-4d bands. The metallic fcc-Pd 0 , Pd 0−δ , and Pd 0+δ states (0 < δ < 1) were compared within the PDOSs. The Pd 0−δ shows 1 eV lower than the fcc-Pd 0 , while Pd 0+δ stays 1.5 eV higher. This reveals that Pd-GDY preserves the Pd 0 state and possesses an even higher ENA than the fcc-Pd 0 , which dominates the strong electronegative reduction character for ECNRR (Fig. 5e). We further determined the Pd-4d orbital information under the electrode potential of the planar and round constant potential distributions. Strong orbital interaction between Pd and neighboring C sites induces a close shell effect by way of crossover. The crossover orbital energy turns to be the equivalent point for repulsion and ENA energetic competition, which stays at 6.55 eV beyond the original energy found from fcc-Pd (2.51 eV).
This implies that the ENA has been promoted by the negative correlation to overcome the orbital repulsion (Fig. 5f).
We move on to the energetic preference of ECNRR. Benchmarked from different adsorption configurations, the pathway prefers to perform at the (Pd, C1, C2) active area. This is because the strongly coupled (Pd, C1, C2) elevates the selectivity via enhanced electron transfer, as discussed above. The Pd site dominated preferable pathway was then considered. The Pd-GDY possesses potential of U = −0.37 V for ECNRR and shows a mainly downhill energetic trend. Between these two parts, the Pd site dominates the optimal N-fixation and the C sites along the C-chain distribute the Hadsorption. For U = −0.37 V, the overall energy gain is −3.01 eV, where the NH 3 -desorption is almost barrier-free of merely 0.03 eV for each NH 3 . For U = 0 V, the path confronts the barrier at the asymmetrical N-hydrogenation step until the formation of * NH-NH 2 + 3(H + +e − ), which starts a downhill trend. This hydrogenation step determines the overall reaction barrier of 0.66 eV. The formation of * N = NH + 5(H + +e − ) also shows an uphill step, but a lower barrier of 0.45 eV. Therefore, the asymmetric H-desorption from the (C1, C2) sites controls the barrier acting as a potential determining step (PDS) for the ECNRR (Fig. 6a). We compared site-dependent H-adsorption energies. The C1 and C2 sites are confirmed as energetically favorable sites for H-adsorption. Further H-chemisorption energies summarize that both C1 and C2 contribute two optimal active sites for Hadsorption for efficient proton-electron charge exchange ( Fig. 6b and c). From pathway analysis, we found that the ECNRR prefers parallel hydrogenation (late N-N cleavage) over serial hydrogenation (early N-N cleavage). Further local structural configurations of the ECNRR process demonstrate that late N-N bond dissociation starts at the step of NH 2 −−NH 3 +(H + +e − ), which is advantageous for better HER suppression. This shows a consistent trend that the intermediate nitrogen-nitrogen bonding (i.e. N=N and N-N) variation ensures energetic compensation for H-desorption from the local active adsorption sites (Fig. 6d).

CONCLUSION
In summary, our experimental and theoretical results strongly support a zero-valence state of the palladium atom that can be anchored on graphdiyne surfaces. The anchoring site of a zero-valence metal atom was observed directly by XANES, and the position of the atomic site as judged by sight is in agreement with the theoretical calculation. The derivative X-ray absorption near-edge structure (XANES) of Pd-GDY is similar to that of metallic Pd, which confirms that the stable valence state of Pd atoms in the sample is zero. The zero-valence Pd atoms separated from graphdiyne show high ECNRR activity and selectivity with the highest N 2 reduction activity and average yield of NH 3 at 4.45 ± 0.30 mg NH3 mg Pd −1 h −1 under environmental conditions. Our work represents a new concept of catalysis, which is of great significance for a deeper understanding of the catalytic process and mechanism of zero-valence atomic catalysts, especially the discovery of highly selective atomic catalysts for conversion of nitrogen to ammonia at high efficiency.
We note the rapid development of single-atom catalysts in recent years, which has led to the rapid development of catalytic science. Prof. Tao Zhang [8,9,11], Prof. Jun Li [11,49,50] and Prof. Yadong Li [16,17,19] et al. have made many contributions to basic and applied research on single-atom catalysts, leading the progress in this field. Our study cleverly uses the advantages of the electronic and chemical structures of graphdiyne to successfully anchor transition metal and noble metal zero-valent atoms on graphdiyne, so we call this an 'atomic' catalyst. Loading zero-valent metal atoms has been an important challenge in the field of catalysts, especially for anchoring transition metal atoms. The success of loading zero-valent metal atoms on graphdiyne expands the synthesis methodology of single-atom catalysts and enriches the types of functional catalysts. The birth of zero-valent atomic catalysts is of great academic significance for us to clearly understand the anchoring process of single atoms on the support, the new catalytic process and mechanism, the inter-action of metal atoms with the support, the energy, electron transfer and conversion in the system, and the relationship between the catalytic performance and the above scientific content. The emergence of zero-valent catalysts is a real insight into the strategies involved in the development of new single-atom catalysts. The following strategies are the principles of our later design and synthesis of highly efficient catalysts: (i) the design of support is an important basic strategy related to catalyst stability, selectivity and efficiency; (ii) the energy transfer and electron transport between the support and different metal atoms must be considered; (iii) systems theory must be developed for single-atom and atomic catalysts, and how to use machine learning to screen functionalized catalysts; (iv) in particular, it is necessary to develop highly conjugated supports to form donoracceptor (D-A) systems with metal atoms, understand the mode of interaction of metal atoms and supports, and better understand the regular of structure and performance of the system; and (v) it must be considered how to control the amount of charge transfer in a single-atom system to realize the regulation of catalyst activity.

Preparation of GDY electrode
GDY electrodes were prepared according to previous studies [20,21,23]. Typically, 50 mL of a pyridinic solution of hexaethynylbenzene (HEB, 0.4 mg mL −1 ) was added very slowly to a threenecked flask containing several pieces of copper foil and carbon cloth (CC) at 110 • C. After a 3-day reaction (110 • C, protected from air and light), the CC was washed with hot acetone, DMF, KOH (4 M), HCl (6 M), KOH (4 M) and water, sequentially, followed by drying in a 40 • C vacuum oven for 12 hours. The GDY electrodes were obtained.

Preparation of Pd-GDY
A piece of the freshly prepared GDY electrode was immersed into 35 mL H 2 SO 4 aqueous solution containing 12 mg PdCl 2 . During this process, the Pd atoms could spontaneously anchor to the GDY surface. After a 4-hour reaction, the samples were washed with 0.5 M H 2 SO 4 and water in sequence, and then immediately used for ECNRR.

Characterizations
SEM images were obtained from an S-4800 field emission scanning electron microscope. TEM, HRTEM and EDX mapping data were collected using a JEM-2100F electron microscope operating at 200 kV. (HAADF) STEM measurements were conducted on aberration-corrected cubed FET Titan Cubed Themis G2 300 or JEM-ARM200F (JEOL, Tokyo, Japan). XRD patterns were recorded using a D/max-2500 rotation anode X-ray diffractometer (Rigaku, Japan) with Cu Kα radiation (λ = 1.54178Å). Raman measurements were performed on a Renishaw-2000 Raman spectrometer (473 nm excitation laser source). A Thermo Scientific ESCALab 250Xi instrument with monochromatic Al Kα X-ray radiation was used to perform the XPS measurements.

ECNRR measurements
ECNRR measurements were carried out on an electrochemical workstation (CHI 660E) with H-type electrolytic cell, which was separated by the Nafion 117 membrane into two chambers. The freshly prepared Pd-GDY (or other reference sample) was used as a working electrode. The catalyst loading of Pd-GDY catalyst is 2.7 μg metal cm −2 , and the geometric area is 2 cm 2 . The graphite rod and saturated calomel electrode were used as the counter and reference electrodes, respectively. 30 mL N 2 -saturated electrolyte (0.1 M Na 2 SO 4 or 0.1 M HCl) was added into the chambers. N 2 flow was continuously fed into the cathodic side with proper position. The catalytic activities of Pd-GDY normalized by geometric area and catalyst loading were compared with other recently reported catalysts. Chronoamperometry tests were performed under ambient conditions at different potentials.