All chemicals purchased were of the purest grade and used as received from Sigma-Aldrich unless otherwise stated.

Coconuts were purchased from a local market (Walmart, USA) and their shells were grinded. 25 g of powdered coconut shell was chemically soaked overnight with 50 wt% HNO₃, followed by drying at 70˚C - 80˚C for 6 h. The carbonization was carried out under nitrogen flow at 550˚C for 2 h. The obtained carbon was washed with hot water and then immersed in 3 M HNO₃ overnight and finally filtered and dried at 105˚C. To enhance the interaction of urea with the activated carbon surface, the sample was treated with 2 M HNO₃/1 M H₂SO₄ (1:1 v/v) at 120˚C for 2 h (functionalization treatment). For nitrogen doping of the activated carbon, a post-treatment was performed by dispersion in 1 M urea solution overnight. The resultant dispersion was filtered and dried, followed by pyrolysis at 900˚C for 2 h at 5˚C∙min⁻¹. Figure 1 shows synthesis method of N-doped activated carbon and its usage for the electrochemical applications.

The morphologies of the obtained samples were firstly characterized by using scanning electron microscopy (SEM) (JSM 6610LV, JEOL, Japan) operating at 15 KV. Kratos Axis 165 X-ray photoelectron spectroscope (XPS) was used to measure XPS spectra of the samples using Al Kα 1486.6 eV X-rays. High resolution spectra of C-1s, N-1s, O-1s were collected on the samples. During data acquisition runs, passing energy of 160 eV, current at 10 mA, and a time of 20 ms per step were used. Thermal Gravimetric Analysis (TGA) was performed using a thermogravimetry/differential Seiko analyzer TGA/DTA 6200 in a temperature range of 25˚C - 1000˚C at a heating rate of 10˚C∙min⁻¹ under air condition. Raman spectroscopy was carried out using a Renishaw in via Reflex Raman Microscope. Raman shifts within 150 cm⁻¹ of the excitation frequency can be measured at excitation wavelengths of either 532 nm, 633 nm, or 785 nm, with a spectral resolution of 1 cm⁻¹. Fourier-transform infrared spectroscopy (FTIR) (Thermoscientific Nicolete iS-10) was carried out to characterize the functional chemical groups of the char. Also, X-ray diffraction (XRD) analysis was performed using a Mini Flex 600 benchtop machine (Regaku, Japan) with 2Ɵ scanning range from 5˚ to 90˚.

All measurements were carried out at room temperature using a potentiostat (A-METEC) with a three-electrode electrochemical cell. We used an interchangeable glassy carbon (GC) disk electrode with a 5 mm diameter (area of 0.196 cm²) as the working electrode, a platinum (Pt) wire as the counter electrode, and Hg/HgO (in saturated KOH solution) as the reference electrode. The GC disk was mechanically polished to achieve a mirror-like finish on the top surface of the electrode. Finally, the GC electrode was sonicated in ultrapure water, ethanol, and then again in ultrapure water and dried at room temperature before applying the catalyst ink. Linear sweep voltammetry (LSV) measurements were performed in O₂-saturated 0.1 M KOH solutions at a scan rate of 10 mV∙s⁻¹. RDE measurements were performed at rotation rates varying from 500 to 2500 rpm depending on the measurement, with the scan rate of 10 mV∙s⁻¹.

To prepare the catalyst ink for electrochemical measurements, 5 milligrams of the synthesized sample was mixed with the following solution: 1.25 mL of 3:2 Isopropanol/water solution, and 50 μl Nafion (5 wt%); this solution was sonicated for 1 h to achieve homogenous ink. Then a 20 μl drop of the catalyst ink was loaded on the polished surface of the glassy carbon electrode (GCE) and dried slowly at room temperature to achieve a uniform surface. Once the sample was completely dried, we estimated the amount of the sample delivered to the surface was 400 µg/cm². The sample was used in electrochemical measurements immediately after drying. Finally, Pt/C catalyst was used for comparisons and prepared exactly as the synthesized sample to achieve a similar loading of 400 μg/cm² of Pt/C (20 wt%).

The surface morphology of the activated carbon was studied using scanning electron microscope (SEM), as shown in Figure 2. The raw coconut shell shows bulky morphology with a smooth surface, and no pores can be observed from Figure 2(A), however, smaller three-dimensional structures including pores were created upon doping with Nitrogen (Figure 2(B)). Lack of porosity of raw coconut shells is due to strong cross-linking of cellulose, hemicellulose, and

lignin [26]. The presence of numerous voids allows the activation agent (herein urea) to freely access and disperse uniformly in coconut shells, resulting in the formation of porous materials [27]. The special structure of coconut biomass shortens the mass transport path and minimizes the diffusion resistance of target substances, resulting in significant improvement in properties.

XRD analysis was performed to identify the phase formed. XRD measurement was performed from an angle of 5˚ to 90˚ at a step rate of 0.2. Figure 3(a) shows XRD pattern obtained for raw coconut particles before carbonization and after N-doped. The predominant diffraction maxima observed for raw coconut are around 16.6˚, 22˚ and 34.6˚, which represents Cordierite (Mg₂Al₄Si₅O₁₈), Quartz (SiO₂), and moissanite (SiC), respectively [28], with quartz having the highest intensity. The diffraction maxima occurring at 2 theta of 27˚ and 44˚ represents silicate minerals and sodalities respectively [29]. The wide X-ray diffraction (XRD) pattern of N-doped Activated Carbon (Figure 3(b)) further confirms its amorphous nature. Two weak and broad diffraction maxima at around 23˚ and 43˚ were observed for this sample, which correspond to the (002) and (100) diffraction patterns of amorphous graphitic carbon [30]. Some diffraction maxima disappeared after N doping which is due to overlapping with (002) broad peak.

XPS spectra of N-doped activated carbon sample are shown in Figure 4. C1s binding energy (BE) can be seen in Figure 4(A), indicating sp² C (284.7 eV), sp³ C (284.9 eV), C-O/C-N (286.2 eV), and π excitation (290.7 eV). We can see O1s BE from the XPS spectra in Figure 4(B); O-C (531.5 eV), O=C (532.4 eV), and O-C=O (534.1 eV) [31]. Figure 4(C) also shows the XPS spectra of functionalized graphene in N1s BE range. The XPS spectra have been fitted to get detailed chemical bonding information of the elements N and O with carbon. Nitrogen is observed in the sample confirming its incorporation into carbon structure. Generally, there are several nitrogen functional groups in nitrogen-functionalized carbon structure. These include pyridinic-N (BE = 396.1 eV),

pyrrolic-N (BE = 400.2 eV), quaternary nitrogen (BE = 401.9 eV), and N-oxides of pyridinic-N (BE = 403.2 eV) [32] [33] [34].

The nitrogen functional groups are usually in the following molecular structures (chemical states) [32] [33] [34]: pyridinic-N refers to nitrogen atoms at the edge of graphene planes, each of them is bonded to two carbon atoms and donates one π-electron to the aromatic π-system; pyrrolic-N refers to nitrogen atoms that are bonded to two carbon atoms and contribute to the p system with two π-electrons; quaternary nitrogen is also called “graphitic nitrogen” or “substituted nitrogen”, in which nitrogen atoms are incorporated into the graphene layer and replace carbon atoms within a graphene plane; N-oxides of pyridinic-N (pyridinic-(N⁺ - O)) are bonded to two carbon atoms and one oxygen atom. Figure 4(D) shows the survey of the presence elements

Raman peak detected in Figure 5 between 1339 and 1348 cm⁻¹ can be assigned to D band characteristic of the disorder while the peak observed between 1586 and 1593 cm⁻¹ can be attributed to the graphitic G band according to literature [35]. The D-band indicates the presence of defects in the graphitic materials due to bond-angle disorder, bond-length disorder, vacancies edge defects, etc. On the other hand, G-band is related to the stretching motion of the pairs of carbon sp² atoms while the broad peak in 2D-band indicates the number of layers formed [36].

The ID/IG ratio provides information about the crystallite dimension, plane defects, edge defects, and the nature of disorder of the carbon derivative [37]. It is well-known that the addition of heteroatoms in the graphitic carbon lattice structure results in more defects and a larger D-band intensity [38].

To gain insight into the purity of activated carbon material, thermogravimetry coupled to mass spectrometry studies (TGA-MS) was performed, and the corresponding thermogram is presented in Figure 6. The curve is composed of two main stages with mass losses. According to mass spectrometry data, the first stage with the weaker mass loss below 100˚C can be attributed to the loss of water molecules adsorbed on the surface of the activated carbon, while a significant loss of CO₂ after 600˚C - 650˚C and carbon monoxide (CO) after 700˚C is also observed.

Figure 7 shows the FTIR pattern with functional groups of the samples. The FTIR characterization used transmittance spectrum data and was recorded between 750 - 4000 cm⁻¹. The characterization results show that a functional group on the sample surface was close to each other. The peak at approximately 3400 cm⁻¹ is due to O-H vibrations, indicating the phenolic groups [39] [40] [41]. The peak at 2350 cm⁻¹ and 1550 cm⁻¹ corresponds to carboxylic acid with a stretching vibration of C-O [42] [43]. And the peak at 1120 cm⁻¹ can be attributed to C-O [44] [45]. The functional groups of activated carbon, comprising O-H, C-O, and C-O stretching vibration, have the essential role as an absorbent of pollutant [46]. Furthermore, the functional groups containing oxygen on the sample surface are a depositional location to absorb the metal particle for catalyst application [44]. The functional groups observed in our coconut shell-based activated carbon are similar to previous research and commercial activated carbon [46].

We compared the performance of our sample for ORR and OER catalyses in alkaline medium (0.1 M KOH) saturated in O₂ using LSV as shown in Figure 8. It is also common to perform the two reactions in the same vessel with the aim of seeking a good catalyst for regenerative fuel cell [47] [48] [49]. Comparisons of LSV data for raw coconut sample and N-doped activated carbon sample revealed that the presence of nitrogen can enhance the electrocatalytic performance, i.e., higher current density and more positive onset potential (Figure 8(A)). The results indicate that the intercalation of nitrogen in the carbon structure can improve electrocatalytic activities for the two reactions.

To obtain insight into the number of electrons involved in the reaction, which is an important factor for fuel cell applications, Koutecky-Levich plots were used. The corresponding Koutecky-Levich plots (j⁻¹ vs. (angular rotation speed, ω)^−1/2) at various electrode potentials show good linearity for N-doped activated carbon (Figure 8(B)); linearity and parallelism of both plots are considered as typical first-order reaction kinetics with respect to the concentration of dissolved O₂. The kinetic parameters can be analysed based on the Koutecky-Levich equations [47]:

1 j = 1 j L + 1 j K = 1 B ω 1 / 2 + 1 j K B = 0.201 n F C 0 ( D 0 ) 2 / 3 ν − 1 / 6 J K = α n F C 0

Here, j is the measured current density; j_K and j_L are the kinetic and diffusion-limiting current densities, respectively; ω is the angular rotation speed of the disk (rpm); n is the overall number of electrons transferred in oxygen reduction; F is the Faraday constant (F = 96,485 C∙mol⁻¹); C₀ is the bulk concentration of O₂; ν is the kinematic viscosity of the electrolyte; and α is the transfer coefficient. The curve j⁻¹ vs. ω^−1/2 is a straight line characterized by slope (nB)⁻¹ and intercept ( j K − 1 ). The slope of the straight line enables the extraction of the total number of exchanged electrons n and the intercept at the origin gives the inverse of the kinetic current j_K by using parameters C₀ = 1.2 × 10⁻³ mol∙L⁻¹, D₀ =1.9 × 10⁻⁵ cm∙s⁻¹, and ν = 0.1 m²∙s⁻¹ in 0.1 M KOH. ORR can occur either via a direct 4-electron reduction pathway where O₂ is reduced to H₂O, a 2-electron reduction pathway where it is reduced to hydrogen peroxide (H₂O₂), or an indirect 4-electron reduction pathway where the generated H₂O₂ is further reduced to H₂O. The number of electrons transferred for N-doped activated carbon is approximately 4 while that for raw coconut is around 2 (Figure 8(G)). Thus, our N-doped sample shows better activity towards ORR compared to the raw sample, forming about 4 electrons during ORR, which agrees with previous papers [50] [51] [52] [53]. In this regard, Yang et al. [54] showed that the presence of quaternary N (graphitic N, N-bonded to three carbon atoms) atoms which have relatively low energetic barrier for donating electrons from the surface of the catalyst can also be responsible for ORR. The quaternary N atoms in graphene-typed structure could give electrons to the π-conjugated system so that it can increase nucleophile strength of the adjacent carbon rings [C(δ⁻)] and improve O₂ adsorption. Since O₂ has high densities of O lone pair electrons [O(δ⁺)], the presence of [C(δ⁻)] can enhance ORR [55] [56]. The role of the real “electrocatalytically active sites” still remains in debate since their contribution to the catalytic activity is not well defined [57]. In some studies, the enhanced electrocatalytic activity is attributed to pyridinic-N and/or pyrrolic-N [32] [33] [58]. Miyata suggested that graphitic nitrogen is more important for the electrocatalytic activity of nitrogen-doped carbon [59]. In parallel to this data, our XPS results indicate that N-graphene also exists in the N-doped activated carbon sample. It is believed that carbon atoms adjacent to nitrogen dopants possess a substantially higher positive charge density to counterbalance the strong electronic affinity of the nitrogen atom [60], which results in an enhanced adsorption of O₂ and reactive intermediates (i.e., superoxide, hydroperoxide) that proceeds to accelerate the ORR [61] [62]. The nitrogen-induced charge delocalization could also change the chemisorption mode of O₂—from monoatomic end-on adsorption on undoped carbon to a diatomic side-on adsorption at nitrogen functionalized carbon structure which effectively weakens the O-O bond to facilitate ORR [60]. This is also true for H₂O₂ reduction because breaking the O-O bond is also a key step for the electrocatalytic reduction of H₂O₂. In addition, the presence of nitrogen in the structure enhances the ability of graphene sheets to donate electrons [33], which is advantageous for reduction reactions.

Yang et al. showed that the adsorption of OOH* and O* intermediates on carbon atoms next to the pyridinic N is responsible for OER [54]. Pyridinic N (an electron-withdrawing group with the lone pair electrons involved in the resonance to delocalize electrons to make the N atoms electron-deficient) can accept electrons from adjacent C atoms [C(δ⁺)], facilitating the adsorption of water oxidation intermediates (OH⁻, OOH⁻) from catalyst surface [63] [64]. Owing to the fact that Pt is not a good catalyst for OER, we also compared the performance of our samples with Ru catalysts, which are well-known precious metal catalysts for OER. The OER activities of our sample are comparable to those of precious metal catalysts (Figure 8(A)).

The Tafel equation describes the electrochemical kinetics relating the rate of an electrochemical reaction to the overpotential [65]. The Tafel slope provides insight into the reaction mechanism. Larger Tafel slope indicates that a larger resistance (or a large loss of potential) is necessary to accelerate a chemical reaction. Pt/C 20 wt% exhibits Tafel slopes of 60 and 120 mV∙dec⁻¹ at high and low potentials, respectively. Tafel slopes for our sample in ORR and OER regions as shown in Figure 8(C) and Figure 8(D) demonstrate similar results as compared to Pt/C 20 wt%.

The durability/stability of the obtained sample as ORR-OER catalyst was evaluated against Pt/C electrode Figure 8(E) and Figure 8(F); the test was performed using chronoamperometry in 0.1 M KOH solution saturated with O₂. The corresponding current-time chronoamperometric response of N-doped activated carbon exhibits a very slow attenuation with only low decay for ORR and OER regions over 1000 s. In comparison, Pt/C 20 wt% showed higher decay for ORR and OER regions. This demonstrated that the durability/stability of our catalyst is better than that of Pt catalyst. To confirm the reversibility of the catalysts, LSVs were measured using forward and backward currents as illustrated in Figure 8(H). Observation of the same LSV futures of running forward and reverse currents reveals that the sample has a good reversibility for both reactions. This result reveals that performing the electrochemical measurement via starting from ORR region or OER region has no effect on the catalytic performance. Consequently, this catalyst can be used for regenerative fuel cell.

To understand the catalytic performance, we further compared the oxygen electrode activity (OEA) of our sample with the Ru and Pt commercial catalysts. Here OEA is calculated as the difference between the ORR at −3 mA∙cm⁻² and OER at 10 mA∙cm⁻², which is the common value used in literature [49] [66]. A current density of 10 mA∙cm⁻² is the conventionally used in OER literature [3] [67]. Smaller OEA value indicates that the performance is closer to the ideal reversible oxygen electrode. Table 1 shows the accumulated OEA data for all the

materials used in this study. To assess the overall oxygen electrode activity, the difference between the ORR and the OER metrics was considered. The OEA of N-doped activated carbon is favorably comparable to the precious metal catalysts. The presence of electron donating quaternary nitrogen atoms is favorable for ORR region. In addition, high contents of electron-withdrawing functional groups (pyridinic N, C=O) are attractive for OER, [68]. Comparison between N-doped activated carbon and the raw coconut shells indicates that the observed electrocatalytic performance is attributed to the incorporation of nitrogen groups in the sample [60] [69] [70]. The results demonstrate the low-cost and scalable synthesis for producing efficient noble-metal-free, carbon-based oxygen catalysts for a variety of electrochemical applications.