Effective Fuel Cell Electrocatalyst with Ultralow Pd Loading on Ni-N-Doped Graphene from Upcycled Water Bottle Waste

Abstract

Environmental pollution due to the excessive consumption of fossil fuels for energy production is a critical global issue. Fuel cells convert chemical energy directly into electricity in a clean and silent electrochemical process, but face challenges related to hydrogen storage, handling, and transportation. The direct borohydride fuel cell (DBFC), utilizing sodium borohydride as a liquid fuel, is a promising alternative to overcome such issues but requires the design of cost-effective nanostructured electrocatalysts. In this study, we synthesized nitrogen-doped graphene anchoring Ni nanoparticles (Ni@NG) by thermal degradation of polyethylene terephthalate bottle waste with urea and metallic Ni, and evaluated it as a sustainable carbon support. Electrocatalysts were prepared by incorporating ultralow amounts (0.09 to 0.27 wt.%) of Pd into the Ni@NG support. The resulting PdNi@NG electrocatalysts were characterized using ICP-OES, XPS, TEM, N₂-sorption analysis, XRD, and Raman and FTIR spectroscopy. Voltammetry assessed the materials’ electrocatalytic activity for oxygen reduction and borohydride oxidation reactions in alkaline media, corresponding to the anodic and cathodic reactions in DBFCs. The electrocatalyst with 0.27 wt.% Pd loading (PdNi_15@NG) exhibited the best performance for both reactions. Consequently, it was employed as an anodic and cathodic material in a lab-scale DBFC, achieving a specific power of 3.46 kW g_Pd⁻¹.

1. Introduction

Direct borohydride fuel cells (DBFCs) are a type of polymer electrolyte membrane (PEM) fuel cell that use sodium borohydride (NaBH₄) as a fuel in alkaline media and hydrogen peroxide or oxygen as the oxidant. It was studied for the first time in 1962 by Indig and Snyder [1]. NaBH₄ has gained a lot of attention from researchers due to its high specific capacity (5.7 Ahg⁻¹) [2,3], low toxicity, safer transportation and storage, chemical stability, high hydrogen content (10.6 wt.%) [2,3], and high theoretical open circuit voltage (OCV) [3]. However, the commercialization of DBFCs is limited by using expensive noble-metal-based electrode materials as electrocatalysts [4,5,6,7,8]. One effective solution to the cost problem is developing highly active and stable anode and cathode electrocatalyst materials. Recent research has proven that borohydride oxidation reaction (BOR) on a Pd-based electrocatalyst has faster kinetics than on a Pt-based electrocatalyst. In addition to this, it has a lower onset potential than Au [9]. Pd-based electrocatalysts’ stability and electrocatalytic activity are improved by alloying them with lower-cost metals (e.g., Ni, Co, Fe, Sb, As, Bi, Ti) [10,11,12,13,14,15,16,17]. Among the nonprecious transition metals, nickel has unique qualities [18]. For example, the PdNi electrocatalyst’s active sites are renewed when OH-adsorbed groups are present on the nickel oxide surface. PdNi alloys’ enhanced activity and stability result from the synergistic interaction between Pd and Ni [9,14,15,19,20].

Hybrid materials with meso- or micropores have emerged as interrelated porous materials. Being connected to porous materials enhances structural porosity, specific surface area, mass transfer capacity, and shape compatibility [18]. In terms of different structural features [21], linked porous materials have unique homogenous morphologies. These substances could resemble sheets, as is the case with graphene, frequently used to support Pd-based electrocatalysts. Choosing a suitable electrocatalyst support material is critical to a bimetallic electrocatalyst’s stability, activity, and cost. Hosseini et al. prepared an anode composed of Pd/Co and Pd/Ni supported on nitrogen-doped reduced graphene oxide in a DBFC and realized power densities at 60 °C equal to 275.35 and 353.84 mW cm⁻² [22].

Plastic waste is considered a dangerous environmental threat because it degrades slowly. Upcycling waste plastic is a cost-effective and simple technique to prepare electrocatalysts based on carbon [23,24,25,26]. Both the energy issue and insufficient management of plastic waste can be solved simultaneously by commercializing alternatives for clean energy supply devices and eliminating the accumulation of plastic waste. Graphene is commonly used as an electrode support in energy supply devices because of the high surface area of its layer structure, good conductivity, and sufficient thermal and chemical durability [27,28,29,30,31,32]. Generally, nitrogen-doped graphene metal-based electrocatalysts include a carbon backbone with defects containing a nitrogen/metal framework inserted into the graphitic/graphene-like structure [27]. When N atoms are incorporated into the graphene skeleton, the electron density in the nearby C atoms is rearranged, which changes the graphene geometry and improves its stability to donate electrons [33,34,35].

Here, nitrogen-doped graphene anchoring Ni nanoparticles (Ni@NG) is synthesized by a waste-to-value, scalable, one-pot pyrolysis of polyethylene terephthalate (PET) waste bottles with nickel metal and urea. This highly porous and stable substrate was used to support an ultralow amount of Pd nanoparticles, thus combining two approaches toward more efficient, cost-effective, and sustainable electrocatalysts. The electrocatalysts demonstrated improved performance as anodic and cathodic electrocatalysts in DBFCs, and therefore, fundamental studies were complemented with fuel cell testing, delivering high specific power densities. A literature search shows no previous work has been performed on creating high-performance NG from plastic wastes to support bimetallic PdNi nanocatalysts for DBFCs.

2. Materials and Methods

2.1. Synthesis of the Electrocatalysts

The samples were prepared by pyrolyzing waste PET drinking water bottles with urea and metallic nickel at 800 °C for 1 h. The produced black precipitate was collected and then ground to produce Ni@NG powder. PdNi@NG electrocatalysts with three different Pd metal nominal loading ratios, namely 0.5, 1, and 1.5 wt.%, were produced by mixing Ni@NG powder and tetrachloropalladinic acid solution (H₂PdCl₄.8wt.%H₂O) purchased from Sigma-Aldrich (St. Louis, MO, USA). This was followed by adding 1 mL of a mixture of 1 M NaBH₄ in 1 M NaOH solution and then stirring for 24 h. The obtained electrocatalyst was filtered, rinsed with deionized water until no chloride was detected, and then left to dry for 12 h at 60 °C. Figure 1 schematically shows the steps followed for preparing the PdNi@NG nanocomposites.

2.2. Characterization of the Electrocatalysts

Several methods were employed to characterize the prepared electrocatalysts, including XRD analysis (Schimadzu-7000, Columbia, MA, USA), Raman spectroscopy (SENTERRA, Bruker, Bremen, Germany), FTIR spectroscopy (Shimadzu FTIR-8400S, Kyoto, Japan), Brunauer–Emmett–Teller (BET) analysis, XPS analysis (K-Alpha, Thermo Fisher Scientific, Waltham, MA, USA), and transmission electron microscopy (TEM, JEOL JEM-1230, Tokyo, Japan). The metal loadings were determined using ICP-OES (Optima 7000DV, Perkin Elmer, Waltham, MA, USA).

2.3. Electrochemical Measurements

A PGSTAT100 potentiostat (Metrohm Autolab B.V., Utrecht, The Netherlands) was used to evaluate the electrocatalysts’ activity for BOR and ORR. Electrochemical measurements were performed in an electrochemical cell containing a glassy carbon electrode (GC) coated with the prepared electrocatalysts as the working electrode, a Pt sheet counter electrode, and a saturated calomel electrode (SCE) as the reference electrode. All recorded experiments were run in triplicate. The potentials given in the work refer to the SCE reference unless otherwise noted.

The electrocatalyst inks contained 5 mg of each synthesized electrocatalyst dispersed ultrasonically for 2 h in 100 μL of 2% solution of polyvinylidene fluoride (PVDF) in N–methyl–2–pyrrolidone (NMP). Then, 3 μL of the obtained ink was dropped on a glassy carbon electrode surface with a geometric area of 0.2 cm² and dried at 80 °C for 4 h.

Cyclic voltammograms (CVs) were run in a 0.03 M NaBH4 + 2 M NaOH solution to assess the electrocatalysts’ activity for BOR. CVs were also recorded in deaerated (using an argon flow) and O₂-saturated 0.1 M KOH to access their activity for ORR. Linear scan voltammograms (LSVs) were also recorded in O₂-saturated 0.1 M KOH solution at a scan rate of 10 mV s⁻¹ using the rotating disk electrode (RDE) method, with rotation rates ranging from 0 to 2400 rpm.

The activity of PdNi_15@NG as a DBFC anode and cathode electrocatalyst was examined at temperatures between 25 and 55 °C. The assembled DBFC used 1 M NaBH₄ + 2 M NaOH solution as the fuel, O₂-saturated 1 M NaOH as the oxidant solution, and a Fumasep FAA-3-PK-130 (FUMATECH BWT GmbH, Bietigheim-Bissingen, Germany) anion-exchange membrane (AEM) separating the two compartments (Figure 2). Each fuel cell compartment contained 100 mL of the corresponding aqueous electrolyte. The AEM membrane’s active area was ca. 30 cm². All electrochemical measurements were performed with a Zennium electrochemical workstation (ZAHNER-Elektrik GmbH & Co. KG, Kronach, Germany). The performance of the fuel cell was evaluated by recording the cell polarization and obtaining the corresponding power density curves.

3. Results and Discussion

The present research work aimed to create sustainable and cost-effective nanostructured electrocatalysts for BOR and ORR in DBFCs. The masses of the elements (metal loadings) on the NG-based electrocatalyst were determined by ICP-OES analysis (

Table 1. The metal loading in the electrocatalysts as determined by ICP-OES analysis.

Catalyst	Metal Loading in the Electrocatalysts/wt.%
	Pd	Ni
Ni@NG	-	92.28 ± 0.14
PdNi_5@NG	0.09 ± 0.005	56.94 ± 0.23
PdNi_10@NG	0.12 ± 0.01	53.32 ± 0.21
PdNi_15@NG	0.27 ± 0.05	58.06 ± 0.26

). From ICP data, the fabricated bimetallic PdNi@NG electrocatalysts were determined to possess 0.09–0.27 wt.% Pd, which was lower than originally targeted. Still, these electrocatalysts with such low Pd amounts exhibited promising performance. In fact, high activity at low platinum-group metal loading is a general goal in the development of novel electrocatalysts, and thus, a further increase of Pd loading was not attempted.

Figure 3a displays the TEM image of PdNi_15@NG. Metallic particles are shown as dark spots scattered on irregularly, non-uniformly distributed graphitic sheets. Differences in the shape of metallic nanoparticles can be attributed to various factors, including heteroatomic intermix during crystal growth and differences in surface free energy. Pd and Ni nanoparticle sizes play an important role in the support interactions and the anchoring onto the N-doped graphene-based substrate. Although particle size estimation from the TEM image in Figure 3a indicates an average dimension at the nanoscale, it is not possible to quantify Pd and Ni particle sizes exactly.

As shown in Figure 3b, it is evident that the synthetic samples’ nitrogen adsorption-desorption isotherms have a distinctive IUPAC-type IV character. The Ni@NG adsorption isotherms exhibit meso- and microporosity, which is explained by both sheet defects and interlayer gaps between the graphene sheets. For PdNi_15@NG, the N₂-sorption isotherm climbs above 0.9 relative pressure and does not reach saturation. It was also noticed that there is no meeting at the endpoint, referred to as “open hysteresis”. That phenomenon is well known for microstructure changes in carbonaceous materials, mainly when the N₂ sorption occurs at low temperatures, entering the pores and leading to deformation. PdNi_5@NG and PdNi_10@NG also demonstrate the presence of meso- and macropores based on their significant specific volumes adsorbed at low relative pressures.

Table 2. Pore structure parameters of samples.

Catalyst	SBET/m2 g−1	Total Pore Volume/cm3 g−1	Average Pore Diameter/nm
Ni@NG	97	0.06	2.6
PdNi_5@NG	71	0.08	4.3
PdNi_10@NG	40	0.09	9.0
PdNi_15@NG	6	0.3	21.5

displays the samples’ specific surface area, along with the total pore volume and the average diameter of the pores. The total pore volume is automatically reported, setting the final data point at the highest experimental relative pressure within the instrument’s range of accuracy. The average pore diameter in tenths of a nanometer was obtained using multilayer adsorption and capillary condensation of N₂ gas within the narrow-diameter pores below a few tens of nanometers. The average pore diameter of PdNi_15@NG was found to be significantly higher than that of PdNi_10@NG and PdNi_5@NG.

The XPS elemental survey spectrum of the PdNi_15@NG surface is shown in Figure 4a. The graphitic C 1s peak was shown at 287.7 eV owing to the carbon bonds in the graphene architecture. Moreover, an N 1s peak at 400 eV demonstrates that the graphene was successfully doped with nitrogen. O, Ni, and Pd were also observed in the composite. As shown in Figure 4b, three peaks at 284.1, 285.3, and 287.4 eV may be identified in the high-resolution spectrum of C 1s. The last peak most likely originates from the C=O bond. That was confirmed with the high-resolution O 1s XPS spectrum (Figure 4c). The other two peaks correspond to sp³ hybridized single carbon bonds, C-C, and either C-O or C-N bonds, or a combination of both signals. As shown in Figure 4d, the high-resolution N 1s XPS spectrum provides further evidence of the presence of C-N bonds. The peak at 399.6 eV in the deconvoluted N 1s spectrum represents the pyridinic nitrogen groups [31,36]. Nitrogen-doped materials have been shown to have increased electrocatalytic activity toward BOR and ORR. Graphene has a larger density of π states near the Fermi level due to the pyridinic N, which can significantly increase ORR activity. The high-resolution Ni 2p XPS spectra shown in Figure 4e exhibit Ni with various chemical valences, detected at 855.6 eV, 861.5 eV, 873.1 eV, and 878.6 eV, corresponding to Ni²⁺ 2p_3/2, Ni³⁺ 2p_3/2, Ni²⁺ 2p_1/2, and Ni³⁺ 2p_1/2, respectively. These results confirm that the peak positions for the oxidation states of Ni²⁺ and Ni³⁺ are in close agreement with prior reports [37]. However, the strength of the Ni²⁺ 2p_3/2 or 2p_1/2 peaks is larger than that of the Ni³⁺ peaks. This indicates that there are more Ni²⁺ ions present than Ni³⁺ ions, which is beneficial for the electronic conductivity [3]. It is evident from the high-resolution spectra of Pd 3d shown in Figure 4f that Pd is mainly found as Pd⁰. As compared to the standard spectra of metal Pd⁰, the binding energies of Pd increased from 335.1 eV to 336.9 eV and from 340.4 eV to 341.1 eV corresponding to Pd 3d_5/2 and Pd 3d_3/2, respectively [38].

Raman spectroscopy confirmed the samples’ graphitic nature. Figure 5a shows that all the prepared electrocatalysts had two broad peaks corresponding to the G and D bands, at approximately 1590 cm⁻¹ and 1370 cm⁻¹, respectively. The sp³ hybridized carbon atoms, corresponding to the D band, are the reason for the structural defects. Conversely, organized sp²-linked carbon atom vibrations are the source of the G band, also known as the growth band [39]. The D band of PdNi_15@NG shows a shift to a higher wavelength contrasted with other PdNi@NG and Ni@NG electrocatalysts, and it could refer to the sustained damage caused to the C-C bond conjugation during the high-concentration Pd precipitation reaction [40]. However, several peaks were observed at low-frequency Raman bands, ascribed to laser-induced oxidation of the metal nanoparticles with a resultant formation of nickel oxide and palladium oxide. The degree of irregularity in the graphitic carbon structure, i.e., the concentration of the defect in the graphitic structure, is represented by the I_D/I_G intensity ratio. As can be seen from Figure 5a, the PdNi_15@NG sample shows the most disordered nanographitic structure with the I_D/I_G ratio of 0.93, i.e., higher than that of PdNi_10@NG (0.88), PdNi_5@NG (0.73), and Ni@NG (0.66), indicating the interaction between Ni@NG and Pd nanoparticles. This interaction is responsible for the increased degree of defects with an increase in the Pd concentration, and that causes this difference in peak ratio [40,41].

The processed samples’ FTIR spectra are displayed in Figure 5b, where a clear broad absorption at 3450 cm⁻¹ is indicative of either the humidity absorbed during the samples’ preparation or the stretching vibration mode of –OH. The nitrogen-doped graphitic structure of the samples is evident by the presence of an in-plane C=C/C=N vibration at 1680 cm⁻¹; this is a feature inherent in sp² graphitic materials [42]. The C-O bond is attributed to the peak at roughly 664 cm⁻¹. Moreover, two more peaks were seen between 610 and 714 cm⁻¹; these peaks may be related to the stretching vibration of metal oxide/C–O–C [43].

Figure 5c shows the XRD patterns of the prepared electrocatalysts. The Ni face-centered cubic (fcc) structure was confirmed to be thermally incorporated into the graphitic structure by observations made at 2θ values of 44.2°, 51.5°, and 75.8°, corresponding to the (111), (200), and (220) planes, respectively (JCPDS No. 70-1849). PdNi_10@NG and PdNi_15@NG spectra show a distinctive Pd peak at 37.6° corresponding to the (111) plane (PDF 03-065-6174), and no discernible Pd diffraction peaks are seen in PdNi_5@NG; this could be caused by a low amount of Pd present that is not detectable by XRD [44].

Figure 6 shows the cyclic voltammograms of the prepared electrocatalysts recorded in the supporting 2 M NaOH electrolyte and in 0.03 M NaBH₄ + 2 M NaOH solution, revealing their stable response with typical BH₄⁻ oxidation peaks. The peak current densities of the BH₄⁻ oxidation increase with the increase in Pd loading in the PdNi@NG electrocatalysts. The highest current density for the oxidation of BH₄⁻ ions was achieved using the PdNi_15@NG electrocatalyst (3.9 mAcm⁻² at −0.24 V vs. SCE), being ~2.3 times higher than at PdNi_5@NG electrocatalyst.

With an increase in the potential scan rate from 5 to 1000 mVs⁻¹, current density values increased from ~2 to 10 times for the prepared PdNi@NG electrocatalysts in the NaBH₄ solution (Figure 7). The plots of the peak current (j_p) versus the square root of scan rate ν^1/2 were constructed and are shown in Figure 7d. As is evident, the peak current j_p is linearly related to the square root of scan rate, ν^1/2, with a good correlation coefficient, R², of 0.907, 0.996, and 0.987 for PdNi_5@NG, PdNi_10@NG, and PdNi_15@NG, respectively. The shift of peak potential, E_p, to more positive values with the increasing scan rate is typical for irreversible processes. An increase of j_p with ν^1/2 and a slight shift of E_p suggest that the oxidation process of BH₄⁻ ions on the prepared electrocatalysts proceeds as an irreversible oxidation of a bulk species and is thus controlled by diffusion [45].

CVs were recorded in deaerated and O₂-saturated 0.1 M KOH solutions (Figure 8), the latter showing the typical cathodic peak for ORR. Current densities obtained with PdNi_15@NG are ca. 1.5–2 times higher than for PdNi_10@NG, PdNi_5@NG, and Ni@NG electrocatalysts.

The RDE LSVs recorded in O₂-saturated 0.1 M KOH (Figure 9) illustrate that the current densities increase with the rotation rate increase from 0 to 2000 rpm. With the rotation rate increase from 0 to 2000 rpm, the current density values increased ca. 2.7 to 4 times for the prepared PdNi@NG electrocatalysts in 0.1 M KOH solution. The obtained current densities are up to 2.1 times higher on PdNi_15@NG compared to the Ni@NG support and the other PdNi@NG electrocatalysts.

The number of exchanged electrons per O₂ molecule, n, was calculated at different potentials in the 0.40–0.60 V range using the Koutecký–Levich (K–L) equation. The obtained K–L plots are provided in Figure 9a′–d′ and Figure 10b. A comparison of ORR LSVs recorded for the Ni@NG and different PdNi@NG electrocatalysts in O₂-saturated 0.1 M KOH solution at 1600 rpm and K–L plots of the studied electrocatalysts determined at −0.5 V are presented in Figure 10a and Figure 10b, respectively. The studied samples are active ORR electrocatalysts in alkaline media, showing onset potential (E_onset, denoted as the potential at an ORR current density of −0.1 mA cm⁻² [46]) values ranging from −0.213 to −0.320 V and corresponding half-wave potentials, E_1/2, from −0.366 to −0.410 V (Figure 10a and

Table 3. Comparison of the ORR performance of different electrocatalysts in 0.1 M KOH.

Catalyst	Eonset/V vs. RHE	E1/2/V vs. RHE	jd/ mA cm−2	n	Tafel Slope/V dec−1	Source
Ni@NG	0.69	0.60	−1.16	2.3–2.5	0.100	This work
PdNi_5@NG	0.71	0.63	−0.71	1.9–2	0.135	This work
PdNi_10@NG	0.73	0.61	−1.41	2.1	0.108	This work
PdNi_15@NG	0.80	0.67	−1.53	3.2–3.6	0.157	This work
PdNi-450	0.95	0.87	−0.42	4.2–4.3	–	[47]
PdNi-NS/C	0.99	0.93	−0.21	3.7	0.080	[48]
PdNiCu/NG	0.87	0.76	1.79	4	–	[49]
PdNi EN/CB	0.99	0.92	0.19	4	–	[50]

).

The ORR RDE voltammograms at 10 mV s⁻¹ and 1600 rpm were further used to obtain the Tafel plots (Figure 10c), as the Tafel slope is a key parameter in electrochemical kinetics that reflects the rate-determining step of a reaction. For the ORR, the Tafel slope provides insight into the reaction mechanism and the kinetics of the involved steps. Typically, lower Tafel slopes are associated with faster reaction kinetics and, thus, higher activity. However, in some cases, the highest activity may correlate with the highest Tafel slope for several potential reasons. Namely, the Tafel slope for Ni@NG was 0.100 V dec⁻¹, lower than that for PdNi_5@NG (0.135 V dec⁻¹), PdNi_10@NG (0.108 V dec⁻¹), and PdNi_15@NG (0.157 V dec⁻¹) (Table 3). While a high Tafel slope generally indicates slower reaction kinetics, the overall activity of a catalyst for the ORR can still be high due to factors such as surface properties, morphology, electronic effects, and mass transport considerations. Additionally, the electrocatalysts with higher Pd loading show more positive onset potentials, confirming their enhanced activity for ORR.

Namely, PdNi_15@NG electrocatalyst exhibited the highest ORR electrocatalytic activity, with E_onset of −0.161 V and E_1/2 of −0.366 V; these values are 0.132 V and 0.044 V, respectively, shifted to more positive potentials than when using the Ni@NG electrocatalyst. The number of electrons transferred per O₂ molecule (Figure 9d′) ranged between 3.2 and 3.6 in the 0.40–0.60 V potential range. This further suggests that ORR proceeds by mixed two- and four-electron mechanisms in the studied electrocatalysts.

The best performance of the PdNi_15@NG electrocatalyst for both BOR and ORR results from the highest Pd loading and the highest number of defects in its structure, accelerating the charge transfer process. Moreover, this electrocatalyst’s significantly higher pore volume and pore size diameter enable the unobstructed flow of reactants (BH₄⁻, O₂) and products (BO₂⁻, OH⁻) to and from the Pd and Ni active centers, enhancing the mass transfer process.

The practical applicability of PdNi_15@NG as a fuel cell electrocatalyst was examined in an alkaline single-cell DBFC using NaBH₄ as the fuel and O₂ as the oxidant. This DBFC was constructed using PdNi_15@NG both as an anode and cathode electrocatalyst. Figure 11 presents the fuel cell polarization and corresponding power density curves at 25, 35, 45, and 55 °C. All polarization curves were run in triplicate without any visible loss of activity; curves overlapped for each repetition, showing that the PdNi_15@NG electrocatalyst exhibited good stability.

The main determined fuel cell parameters are summarized in

Table 4. Operational parameters of the DBFC employing the PdNi_15@NG electrocatalysts as the anode and cathode.

Temperature/°C	E at Peak Power Density/V	j at Peak Power Density/mA cm−2	Peak Power Density/ mW cm−2	Specific Peak Power Density /kW gPd−1
25	0.50	14	7.1	3.46
35	0.45	18	7.9	3.85
45	0.45	20	8.8	4.29
55	0.50	19	9.5	4.63

. The lab DBFC exhibited an open circuit voltage close to 1.0 V. At 25 °C, the DBFC attained a peak power density of 7.1 mW cm⁻² at a current density of 14 mA cm⁻² and a cell voltage of 0.50 V. Moreover, the peak power density values increased by ca. 1.3 times when increasing the temperature from 25 to 55 °C. Considering the ICP-OES data (cf. Table 1), the Pd loading on the prepared PdNi_15@NG-based fuel cell electrodes was 2.05 µg_Pd cm⁻². This leads to a specific power of 3.46 kW g_Pd⁻¹, significantly higher than for other systems reported in the literature using Pd-based anode electrocatalysts (

Table 5. The performance of DBFCs using different Pd-based electrocatalysts.

Anode (mgPd cm−2)	Cathode (mg cm−2)	Anolyte	Catholyte	P/kW gPd−1	T/°C	Source
PdNi_15@NG (0.002)	PdNi_15@NG	1 M NaBH4 + 2 M NaOH	O2 + 1 M NaOH	3.46	25	This work
Pd50Cu50/C (0.313)	Pt/C (0.5)	1 M NaBH4 + 6 M NaOH	O2 flowrate: 200 mL min−1	0.313	60	[12]
Pd/C (1.08)	Pt/C (0.3)	1 M NaBH4 + 20 wt.%NaOH	Air flowrate: 30 mL min−1	0.018	25	[51]
Au-Pd/C(1:1) (1.75)	Pt/C (4)	2 M NaBH4 + 2 M NaOH	O2 flowrate: 200 mL min−1	0.027	60	[52]
53 wt.% Pd/C (0.31)	50 wt.% Pt/C (0.31)	0.1 M NaBH4 + 1 M NaOH	0.4 M H2O2 + 1 M H2SO4	0.887	25	[53]
Pd/C (3)	Pt/C (3)	1.5 M NaBH4 + 3 M KOH	15 wt.% H2O2 + 1.5 M H2SO4	0.203	25	[54]
NiF@AEI + Pd/C (0.5)	Pt/C (1)	1.5 M NaBH4 + 3 M NaOH	15 wt.% H2O2 + 1.5 M H2SO4	0.898	80	[55]
Pd/C (3) + Ni	Pt/C (3)	1.5 M NaBH4 + 3 M KOH	15 wt.% H2O2 + 1.5 M H2SO4	0.297	70	[56]
Pd@N/C-8 foam (0.15)	Pd@N/C-8 foam (0.15)	0.3 M NaBH4 + 2 M NaOH	1.6 M H2O2 + 2 M H2SO4	0.913	60	[57]

).

4. Conclusions

The present study used one-pot pyrolysis of PET waste bottles with nickel metal and urea as a fast and simple process of preparing nickel on N-doped graphene, used as a sustainable support material for anchoring ultralow loadings of palladium nanoparticles. This straightforward method has substantial benefits, including simple operating stages and reaction setup, allowing scaling up to an industrial scale.

From the produced PdNi@NG electrocatalysts, PdNi_15@NG exhibited the highest electrocatalytic activity for borohydride oxidation and oxygen reduction reactions. This electrocatalyst achieved a current density of 3.9 mA cm⁻² at −0.24 V vs. SCE for the oxidation of BH₄⁻ in 0.03 M NaBH₄ + 2 M NaOH solution. As for ORR in alkaline media, it exhibited an E_onset value of −0.213 V vs. SCE and an E_1/2 of −0.366 V vs. SCE, with a number of electrons transferred of 3.2–3.6 being determined by K-L analysis.

The direct borohydride fuel cells (DBFCs) assembled using PdNi_15@NG as an anode and cathode electrocatalyst in 1 M NaBH₄ + 2 M NaOH anolyte and O₂-saturated 1 M NaOH catholyte attained a peak power density of 7.1 mW cm⁻² at a current density of 14 mA cm⁻² and a cell voltage of 0.50 V at 25 °C. Based on the Pd loading, this corresponds to an outstanding specific power of 3.46 kW g_Pd⁻¹. Accordingly, these data confirm the prepared PdNi@NG electrocatalysts, especially PdNi_15@NG, exhibit promising performance as potential anode and cathode electrode materials for DBFCs. Moreover, N-doped graphene derived from upcycled PET bottle waste was established as a cost-effective and sustainable electrocatalyst support.