Ionic Liquid-Derived Carbon-Supported Metal Electrocatalysts as Anodes in Direct Borohydride-Peroxide Fuel Cells

: Three different carbon-supported metal (gold, platinum, nickel) nanoparticle (M/c-IL) electrocatalysts are prepared by template-free carbonization of the corresponding ionic liquids, namely [Hmim][AuCl 4 ], [Hmim] 2 [PtCl 4 ], and [C 16 mim] 2 [NiCl 4 ], as conﬁrmed by X-ray diffraction analysis, scanning electron microscopy coupled with energy-dispersive X-ray spectroscopy and Raman spectroscopy. The electrochemical investigation of borohydride oxidation reaction (BOR) at the three electrocatalysts by cyclic voltammetry reveals different behavior for each material. BOR is found to be a ﬁrst-order reaction at the three electrocatalysts, with an apparent activation energy of 10.6 and 13.8 kJ mol − 1 for Pt/c-IL and Au/c-IL electrocatalysts, respectively. A number of exchanged electrons of 5.0, 2.4, and 2.0 is obtained for BOR at Pt/c-IL, Au/c-IL, and Ni/c-IL electrodes, respectively. Direct borohydride-peroxide fuel cell (DBPFC) tests done at temperatures in the 25–65 ◦ C range show ca. four times higher power density when using a Pt/c-IL anode than with an Au/c-IL anode. Peak power densities of 40.6 and 120.5 mW cm − 2 are achieved at 25 and 65 ◦ C, respectively, for DBPFC with a Pt/c-IL anode electrocatalyst. SEM/EDS analysis of the three studied electrocatalysts. and and investigation, and writing—review and and visualization, and


Introduction
Ionic liquids (ILs) are nowadays finding many applications, including as reaction solvents in different synthesis procedures, such as polymerization, esterification, or acidic hydrolysis [1,2]. They have various cation-anion combinations, a carbon-rich nature, low corrosivity, low volatility, tunable solubility, and very good thermal stability [3]. Thus, recent studies are focusing on the combination of excellent properties of carbon nanomaterials and their functionalization with ILs to create new sorbent materials, of high loading capacity, good durability, reliability, selectivity, and sensitivity [4,5]. For instance, the low-temperature synthesis of porous carbon material from pyrolysis of cellulose using 18 types of ILs as a reusable agent was reported by Huang et al. [6]. The synthesis of porous carbon nanomaterials by the direct carbonization of ILs has further advantages for a notably simpler procedure (compared to, for instance, traditionally used carbonization of polymers) [7], leading to carbons of unique features such as high nitrogen content. This fur- Figure 1B,D,F illustrates the Raman spectra of Pt/c-IL, Ni/c-IL, and Au/c-I along with D and G peaks, contain three additional bands associated with the D b D3, and D4 [24]. The ID/IG ratios were calculated to be 2.5, 2.13, and 2.07 for Ni/c-IL, and Pt/c-IL, respectively, indicating carbon structure with a number of defe ID3/IG ratio was the highest in the case of Ni/c-IL (0.8), followed by Au/c-IL (0.4) lowest in the case of Pt/c-IL (0.2). This result implies the low crystallinity of Ni/ Au/c-IL compared to Pt/c-IL [25].
SEM analysis ( Figure S1, in Supplementary Information) revealed the distribution of metals on carbon with islands of agglomerated metal nanoparticl

Borohydride Oxidation Study
CVs of Pt/c-IL and Au/c-IL electrocatalysts in the absence of BH4 − (Figure 2A lines) show low current densities (<0.5 mA cm −2 ), with small peaks that can be as with the formation and reduction of gold and platinum oxides in alkaline media other relevant redox processes were observed on those CVs. CV of Ni/c-IL supporting electrolyte shows an anodic peak at 1.47 V and a cathodic peak a corresponding to the Ni 3+ /Ni 2+ redox couple conversion in alkaline media (Fi [15,27]. The oxidation of metallic Ni in alkaline media is known to involve thr steps in the studied potential range [27]. The first step is the formation of the αaccording to Equation (1) in the potential range 0 ≤ ERHE ≤ 0.5 V.
Ni 0 + 2 OH − ⇌ α-Ni(OH)2 + 2 e − With increasing potential (0.5 ≤ ERHE ≤ 1.4 V), the α-Ni(OH)2 hydroxide could into a less hydrated and more stable phase, β-Ni(OH)2 (Equation (2)). This β-p (C)  Figure 1B,D,F illustrates the Raman spectra of Pt/c-IL, Ni/c-IL, and Au/c-IL, which along with D and G peaks, contain three additional bands associated with the D band: D2, D3, and D4 [24]. The I D /I G ratios were calculated to be 2.5, 2.13, and 2.07 for Ni/c-IL, Au/c-IL, and Pt/c-IL, respectively, indicating carbon structure with a number of defects [24]. I D3 /I G ratio was the highest in the case of Ni/c-IL (0.8), followed by Au/c-IL (0.4) and the lowest in the case of Pt/c-IL (0.2). This result implies the low crystallinity of Ni/c-IL and Au/c-IL compared to Pt/c-IL [25]. SEM analysis ( Figure S1, in Supplementary Information) revealed the uniform distribution of metals on carbon with islands of agglomerated metal nanoparticles.

Borohydride Oxidation Study
CVs of Pt/c-IL and Au/c-IL electrocatalysts in the absence of BH 4 − (Figure 2A,C, solid lines) show low current densities (<0.5 mA cm −2 ), with small peaks that can be associated with the formation and reduction of gold and platinum oxides in alkaline media [26]. No other relevant redox processes were observed on those CVs. CV of Ni/c-IL in the supporting electrolyte shows an anodic peak at 1.47 V and a cathodic peak at 1.37 V, corresponding to the Ni 3+ /Ni 2+ redox couple conversion in alkaline media ( Figure 2B) [15,27]. The oxidation of metallic Ni in alkaline media is known to involve three main steps in the studied potential range [27]. The first step is the formation of the α-Ni(OH) 2 according to Equation (1) in the potential range 0 ≤ E RHE ≤ 0.5 V.
BH3OH -+ 6 OH -⇌ B(OH)4 -+ 3 H2O + 6 e - BH3OH -+ 3 OH -⇌ B(OH)4 -+ 3 /2 H2O + 3 e -(7) Regarding Ni/c-IL, three anodic peaks were observed in the CV performed in the presence of BH4 − ( Figure 2B). Those processes appear in a zone of high overpotential, namely peak a1 at 1.26 V (ja1 = 3.9 mA cm −2 ), peak a2 at 1.49 V (ja2 = 6.2 mA cm −2 ) and peak a3, in the back scan, at 1.14 V (ja3 = 4.1 mA cm −2 ). Gas bubbles were observed on the electrode surface during all measurements, suggesting H2 generation by BH4 − hydrolysis. Moreover, a1 most likely corresponds to BH3OH − oxidation and a2 corresponds to BH4 − oxidation proceeding in parallel with NiOOH generation [27]. The back scan peak (a3) originates from the oxidation of BH3OH − ions (which remained adsorbed on the electrode surface during the anodic scan) after the reduction of nickel oxide, which leads to a reactivation of the catalyst surface. The mechanism of BOR at Ni-based electrodes is not completely known; thus, it was speculated that the BOR occurs only after the formation of NiOOH, as given by Equation (8) [27,34].
Regarding the CVs in the presence of BH 4 − (Figure 2, dashed lines), different behaviors are observed for each electrocatalyst. Pt/c-IL shows two well-defined peaks in the direct scan at 0.48 V (j a1 = 11.1 mA cm −2 ) and 1.09 V (j a3 = 8.0 mA cm −2 ) and a shoulder at 0.79 V (j a2 = 9.1 mA cm −2 ) (Figure 2A). In the back scan, another oxidation peak is observed at 0.59 V (j a4 = 6.4 mA cm −2 ). The presence of those peaks evidences the occurrence of both BOR (Equation (4), peak a 3 ) and its competitive reaction, BH 4 − hydrolysis (Equation (5)), and the subsequent oxidation of the hydrolysis products, H 2 (peak a 1 ) and BH 3 OH − (Equations (6) and (7), peak a 2 and a 4 ) [33].
Regarding Ni/c-IL, three anodic peaks were observed in the CV performed in the presence of BH 4 − ( Figure 2B). Those processes appear in a zone of high overpotential, namely peak a 1 at 1.26 V (j a1 = 3.9 mA cm −2 ), peak a 2 at 1.49 V (j a2 = 6.2 mA cm −2 ) and peak a 3 , in the back scan, at 1.14 V (j a3 = 4.1 mA cm −2 ). Gas bubbles were observed on the electrode surface during all measurements, suggesting H 2 generation by BH 4 − hydrolysis. Moreover, a 1 most likely corresponds to BH 3 OH − oxidation and a 2 corresponds to BH 4 − oxidation proceeding in parallel with NiOOH generation [27]. The back scan peak (a 3 ) originates from the oxidation of BH 3 OH − ions (which remained adsorbed on the electrode surface during the anodic scan) after the reduction of nickel oxide, which leads to a reactivation of the catalyst surface. The mechanism of BOR at Ni-based electrodes is not completely known; thus, it was speculated that the BOR occurs only after the formation of NiOOH, as given by Equation (8) [27,34].
NiOOH + NaBH 4 Ni(OH) 2 + oxidation products of NaBH 4 (8) However, simultaneous measurements of H 2 and the electric current indicated that BOR-and H 2 -involving reactions occur in parallel at Ni-based electrodes, with metallic nickel being active for BH 4 − oxidation as shown by Oshchepkov et al. in a recent work [35]. The theoretical number of electrons exchanged (n) during BOR at the Ni electrode can be different depending on the mechanism pathway. If the stepwise pathway, originally proposed by Lee [36], is considered, BOR occurs according to Equations (9) and (10). Since this pathway involves four steps, a theoretical n of 4 could be reached.
[BH x (OH) 4 A second pathway considers the simultaneous occurrence of BOR and BH 4 − hydrolysis according to Equations (11) and (12) [13,27,37]. The number "y" is highly dependent on the electrocatalyst surface and the potential and can change substantially according to the chemical, morphological and superficial properties of an Ni-based electrocatalyst.
[BH y (OH) 4 The appearance of an oxidation peak in the back scan of Ni/c-IL CV in the presence of BH 4 − evidences a reactivation of the Ni electrode surface ( Figure 2B). Oshchepkov et al. analyzed the effect of the metal oxidation grade on the activity of carbon-supported Nibased electrocatalysts for BOR [35], being observed that a partially oxidized Ni-based electrocatalyst achieved better performance than a highly oxidized one. In the case of Ni/c-IL, it is believed that the catalytic activity of Ni(OH) 2 is associated with a 1 and a 3 , since they appear in a zone where this Ni-oxide species is already formed.
Au/c-IL exhibited the typical behavior of Au-based catalysts ( Figure 2C), presenting three anodic peaks: a 1 at 1.09 V (j a1 = 4.7 mA cm −2 ), an ill-defined shoulder (a 2 ) at 1.32 V (j a2 = 2.8 mA cm −2 ) in the direct scan, and a 3 in the back scan at 1.18 V (j a3 = 4.7 mA cm −2 ) [38]. According to mechanistic studies [39], a 1 is a combination of the oxidation of BH 4 − hydrolysis-derivate species and low-potential BH 4 − direct oxidation, a 2 is associated with an oxidation process involving BH 2 (OH) 2 − or BH(OH) 3 − species. In c 1 occurs the same process of a 1 , which proves that Au oxides are not active for BOR, and after reactivation in the back scan (by a reduction reaction), the electrocatalyst becomes active again.
The irreversibility of the process was probed by observing the effect of the scan rate on the voltammograms of the three electrocatalysts, Figure 3. Pt/c-IL ( Figure 3A) and Au/c-IL ( Figure 3C) showed a shift of the peak to more positive potentials. Ni/c-IL ( Figure 3B) did not demonstrate a considerable deviation of peak a 2 potential value, while a 1 seems to shift slightly to the left ( Figure 3D). Furthermore, it was possible to observe an increase of the current density with an increasing scan rate for all electrocatalysts, especially for Pt/c-IL (+173%) and Ni/c-IL (a 2 , +135%), Figure 3E.
The effect of BH 4 − concentration was also analyzed for the three catalysts ( Figure 4). An expected increase of the current density when increasing BH 4 − concentration for all catalysts is observed. From the slope of ln j vs. ln c regression (inset of Figure 4B), it was possible to calculate the order of the reaction β with respect to BH 4 − for the three catalysts, Table 1. BOR on all c-IL-supported catalysts was found to be of the first order in respect to BH 4 − . Furthermore, the values obtained are close to those reported for similar catalysts, such as Pt/C [40] and Pt/PPy-C (platinum nanoparticles supported on polypyrrole (PPy) + carbon composite) [41], or such as Au/C [42] and bulk Au [43]. During all the experiments described above, Ni/c-IL exhibited unstable behavior. Namely, Ni/c-IL showed a critical decrease in BOR activity with continuous cycling, leading to the ill definition or even disappearance of BOR peak in the voltammograms. Oshchepkov et al. explained [35] that the decrease of BOR catalytic activity in the case of Ni electrodeposited on Vulcan XC-72R (Ni ED /C) catalyst was directly related to a decrease in the electrochemically active surface area (ECSA). When totally reduced, Ni ED /C presented the maximum activity, when partially oxidized, an intermediate activity, and when totally oxidized, the lowest BOR activity. This oxidation resulted in a consequent decrease of the ECSA and a proportional reduction in Ni ED /C catalytic activity for BOR. If the Ni is oxidized at very high potentials (as is the case of Ni/c-IL catalyst, Figure 4B), its reduction is extremely difficult or even impossible when in the presence of BH        The effect of an increasing temperature on BOR was also analyzed for Pt/c-IL and Au/c-IL catalysts ( Figure 5). The enhancement of BOR kinetics at both Pt/c-IL and Au/c-IL by temperature increase was evidenced by the increase of the current density. This behavior allows the application of the Arrhenius equation to calculate the apparent activation energy E a app , Table 1. BOR at Au/c-IL presented a low E a app value (13.8 kJ mol −1 ), comparable to values reported for BOR at gold nanorod-polyaniline (Au-PANI) composites (12.6 and 14.1 kJ mol −1 ) [12] and notably lower than those obtained employing Au disk macroelectrodes (40-72 kJ mol −1 ) [43]. Furthermore, a low E a app value of 10.6 kJ mol −1 was determined for BOR at Pt/c-IL, which is 6% lower than one of the lowest value found in the literature for Pt-based electrocatalysts [41] and ca. 38% lower than the lowest one for Pt/C [14], Table 1. Ni electrodeposited on Vulcan XC-72R (NiED/C) catalyst was directly related to a decrease in the electrochemically active surface area (ECSA). When totally reduced, NiED/C presented the maximum activity, when partially oxidized, an intermediate activity, and when totally oxidized, the lowest BOR activity. This oxidation resulted in a consequent decrease of the ECSA and a proportional reduction in NiED/C catalytic activity for BOR. If the Ni is oxidized at very high potentials (as is the case of Ni/c-IL catalyst, Figure 4B), its reduction is extremely difficult or even impossible when in the presence of BH4 − [35]. Thus, it was decided not to use Ni/c-IL catalyst in further studies, since Ni/c-IL is not stable enough to be used as an anode in a DBFC. The effect of an increasing temperature on BOR was also analyzed for Pt/c-IL and Au/c-IL catalysts ( Figure 5). The enhancement of BOR kinetics at both Pt/c-IL and Au/c-IL by temperature increase was evidenced by the increase of the current density. This behavior allows the application of the Arrhenius equation to calculate the apparent activation energy Ea app , Table 1. BOR at Au/c-IL presented a low Ea app value (13.8 kJ mol −1 ), comparable to values reported for BOR at gold nanorod-polyaniline (Au-PANI) composites (12.6 and 14.1 kJ mol −1 ) [12] and notably lower than those obtained employing Au disk macroelectrodes (40-72 kJ mol −1 ) [43]. Furthermore, a low Ea app value of 10.6 kJ mol −1 was determined for BOR at Pt/c-IL, which is 6% lower than one of the lowest value found in the literature for Pt-based electrocatalysts [41] and ca. 38% lower than the lowest one for Pt/C [14], Table 1. To get further insight into the reaction mechanism, the number of exchanged electrons during BOR was calculated from rotating disk electrode (RDE) measurements (Figure 6) by employing the Koutecky-Levich (KL) equation [14]. The j −1 vs. ω −1/2 regressions shown in the inset of Figure 6B reveal good fitting (linearity) of the experimental points. An n of 5.0 and 2.4 was calculated for BOR at Pt/c-IL and Au/c-IL, respectively. It should be mentioned that the number of electrons exchanged during BOR at commercial Pt/C was found to be similar (5.2) ( Figure 6C).
As it was not possible to run RDE measurements with Ni/c-IL catalysts, the number of exchanged electrons was evaluated from the above-described current density vs. square root of scan rate study [14]. Thus, it was calculated that 2 electrons were exchanged during BOR at Ni/c-IL, with this value being comparable to those reported for similar Ni samples [45].
The calculated n values suggest greater tendency for an indirect BH4 − oxidation (according to Equations (5)-(7)), with an n < 8, or even a mixed direct (Equation (4)) and indirect oxidation of BH4 − at three studied electrocatalysts. The BOR mechanism at Pt-, Ni-, and Au-based electrodes is highly dependent on the electrocatalyst surface properties. The adsorption/desorption of the reduced/oxidized species determines the way in which BOR will occur. To get further insight into the reaction mechanism, the number of exchanged electrons during BOR was calculated from rotating disk electrode (RDE) measurements ( Figure 6) by employing the Koutecky-Levich (KL) equation [14]. The j −1 vs. ω −1/2 regressions shown in the inset of Figure 6B reveal good fitting (linearity) of the experimental points. An n of 5.0 and 2.4 was calculated for BOR at Pt/c-IL and Au/c-IL, respectively. It should be mentioned that the number of electrons exchanged during BOR at commercial Pt/C was found to be similar (5.2) ( Figure 6C).  Figure 7 presents the polarization curves of DBPFC with Au/c-IL ( Figure 7A) and Pt/c-IL anode ( Figure 7B) at different temperatures with the corresponding power density curves. Open circuit voltage (OCV) was found to be ca. 1.6 V for DBPFC, with both Au/c-IL and Pt/c-IL anodes at all temperatures ( Table 2). OCV value is lower than the theoretical value for DBPFCs with acidic catholyte due to the mixed overpotentials created by BH4 − As it was not possible to run RDE measurements with Ni/c-IL catalysts, the number of exchanged electrons was evaluated from the above-described current density vs. square root of scan rate study [14]. Thus, it was calculated that 2 electrons were exchanged during BOR at Ni/c-IL, with this value being comparable to those reported for similar Ni samples [45].

Fuel Cell Tests
The calculated n values suggest greater tendency for an indirect BH 4 − oxidation (according to Equations (5)-(7)), with an n < 8, or even a mixed direct (Equation (4)) and indirect oxidation of BH 4 − at three studied electrocatalysts. The BOR mechanism at Pt-, Ni-, and Au-based electrodes is highly dependent on the electrocatalyst surface properties. The adsorption/desorption of the reduced/oxidized species determines the way in which BOR will occur. Figure 7 presents the polarization curves of DBPFC with Au/c-IL ( Figure 7A) and Pt/c-IL anode ( Figure 7B) at different temperatures with the corresponding power density curves. Open circuit voltage (OCV) was found to be ca. 1.6 V for DBPFC, with both Au/c-IL and Pt/c-IL anodes at all temperatures ( Table 2). OCV value is lower than the theoretical value for DBPFCs with acidic catholyte due to the mixed overpotentials created by BH 4 − hydrolysis and H 2 O 2 decomposition species, i.e., the generation of H 2 on the anode and the generation of O 2 on the cathode side. Similarly, OCV values lower than the expected theoretical value were observed with commercial Pt/C (Table 2), as well as in a previous work involving platinum polypyrrole-carbon (Pt/PPy-C) composites as anodes in DBPFC [41].  Figure 7 presents the polarization curves of DBPFC with Au/c-IL ( Figure 7A) an Pt/c-IL anode ( Figure 7B) at different temperatures with the corresponding power densi curves. Open circuit voltage (OCV) was found to be ca. 1.6 V for DBPFC, with both Au IL and Pt/c-IL anodes at all temperatures ( Table 2). OCV value is lower than the theoretic value for DBPFCs with acidic catholyte due to the mixed overpotentials created by BH hydrolysis and H2O2 decomposition species, i.e., the generation of H2 on the anode an the generation of O2 on the cathode side. Similarly, OCV values lower than the expect theoretical value were observed with commercial Pt/C (Table 2), as well as in a previo work involving platinum polypyrrole-carbon (Pt/PPy-C) composites as anodes in DBPF [41]. The values of the peak power density, Pp, of DBPFC with both Au/c-IL and Pt/canode showed the same increasing trend with increasing temperatures, Table 2. Thu DBPFC with Au/c-IL anode reached Pp of 11.1 mW cm −2 and 43.5 mW cm −2 at 25 and °C, respectively. Pp of 40.6 and 120.5 mW cm −2 was reached with Pt/c-IL anode at 25 an 65 °C, respectively. This behavior could be a consequence of enhanced electron transf kinetics, as well as decreased solution viscosity, increased species diffusion coefficien and decreased ohmic losses. Electrochemical impedance spectroscopy measurements ca ried out in a previous work involving gold-rare earth alloys as DBPFC anodes at 25 an 45 °C showed a 61% decrease of the charge transfer resistance and consequent enhanc ment of BOR kinetics at 45 °C [46]. It should be further noted that Pt/c-IL gave notab (three to four times) higher Pp values than Au/c-IL, at all temperatures. The values of the peak power density, P p , of DBPFC with both Au/c-IL and Pt/c-IL anode showed the same increasing trend with increasing temperatures, Table 2. Thus, DBPFC with Au/c-IL anode reached P p of 11.1 mW cm −2 and 43.5 mW cm −2 at 25 and 65 • C, respectively. P p of 40.6 and 120.5 mW cm −2 was reached with Pt/c-IL anode at 25 and 65 • C, respectively. This behavior could be a consequence of enhanced electron transfer kinetics, as well as decreased solution viscosity, increased species diffusion coefficient, and decreased ohmic losses. Electrochemical impedance spectroscopy measurements carried out in a previous work involving gold-rare earth alloys as DBPFC anodes at 25 and 45 • C showed a 61% decrease of the charge transfer resistance and consequent enhancement of BOR kinetics at 45 • C [46]. It should be further noted that Pt/c-IL gave notably (three to four times) higher P p values than Au/c-IL, at all temperatures.

Fuel Cell Tests
DBPFC with commercial Pt/C as anode electrocatalyst gave a peak power density of 52.4 mW cm −2 at 25 • C ( Figure 7C), i.e., higher than DBFPC with Pt/c-IL. This might be due to a better dispersion of the Pt nanoparticles in the case of the commercial electrocatalyst. As mentioned, some islands of agglomerated Pt nanoparticles have been observed in the case of Pt/c-IL. Moreover, a significant increase in peak power density of DBPFC with Pt/C anode could be observed when increasing the temperature to 35 • C. However, further increasing above 45 • C caused pronounced BH 4 − hydrolysis, as evidenced by vigorous bubbling at the anode side. This resulted in a decrease in peak power density when increasing the temperature from 55 to 65 • C. Thus, the peak power density of a DBPFC with commercial Pt/C on the anode side reached a maximum value of 108.1 mW cm −2 , i.e., lower than the DBPFC with herein prepared Pt/c-IL. * Commercial Pt/C electrocatalyst; ** Pt/G-graphene-supported Pt nanoparticles.
Graphene-supported Pt nanoparticles (Pt/G) and Vulcan XC-72R-supported Pt (Pt/C) nanoparticles were also employed as anodes in DBPFC at 25 • C [47]. P p values of 34.1 and 41.8 mW cm −2 were obtained with Pt/C and Pt/G anodes, respectively, which are lower/comparable with that of the herein tested Pt/c-IL anode [47]. DBPFC with Pt deposited on the nanoporous carbon showed somewhat better results at 25 • C, reaching P p of 54 mW cm −2 [48]. DBPFC with Pt on bulk polypyrrole film (Pt/PPy) reached a lower peak power density of 27 mW cm −2 , with a corresponding current density of 42 mA cm −2 at room temperature [49].
The maximum power density of 28.2 mW cm −2 was obtained for DBPFC, with the carbon-supported Au-based electrocatalysts (Au/C) at 20 • C [50]. Herein, an obtained P p value 11.1 mW cm −2 for DBPFC with Au/c-IL at 25 • C could be caused by the agglomeration of Au nanoparticles or enhanced BH 4 − hydrolysis at high concentration of BH 4 − solution in DBFC/DBPFC and at high temperature [51].

Preparation and Characterization of the (IL-c)-Supported Electrocatalysts
Ionic liquids were prepared without solvent, by mixing and melting appropriate amounts of metal chloride and ionic liquids with chloride anion (1-hexyl-3-methylimidazolium chloride and 1-hexadecyl-3-methylimidazolium chloride) (Scheme 1) at a temperature of 85 • C, according to the reactions described in Equations (13) ), in two steps. In the first step, ionic liquids were heated up to 400 °C and maintained at this temperature for 1 h. In the second step, the materials were heated up to 800 °C for another hour. In both steps, the heating rate was 10 °C min −1 . After the treatment at 800 °C, the materials were left to cool to ambient temperature and denoted as Pt/c-IL, Ni/c-IL, and Au/c-IL.
X-ray diffraction data were collected on a Philips PW 1050 (Philips, The Netherlands) diffractometer with Cu-Kα1,2 radiation (Ni filter) at room temperature. Measurements were done in the 2θ range of 10-90° with a scanning step width of 0.05° and 3 s time per step. Raman spectra were collected on an Advantage 532 Raman spectrometer (DeltaNu Inc., Laramie, WY, USA), with YAG lesser operating at 532 nm. Deconvolution of the spectra was performed using Gaussian fitting. Scanning electron microscope Phenom™ ProX Desktop SEM (ThermoFisher Scientific™, Waltham, MA, USA) with integrated energydispersive X-ray spectroscopy (EDS) detector was used to examine the surface morphology and composition.

NaBH4 Oxidation Studies
The catalytic inks were prepared by ultrasonically dispersing 5 mg of the corresponding electrocatalyst and 0.6 mg of Vulcan carbon in 125 μL of a 2% polyvinylidene difluoride solution in N-methyl-2-pyrrolidone. The electrodes were prepared by depositing 10 μL of the catalytic ink onto a glassy carbon tip and left to dry at 80 °C overnight. This procedure was done immediately after the ultrasonic step, to ensure the good dispersion of the electrocatalyst.
Electrochemical measurements were performed using Ivium V01107 Potentiostat/Galvanostat (Ivium Technologies B.V., Eindhoven, The Netherlands) in a one-compartment cell. A saturated calomel electrode (SCE, Metrohm, Herisau, Switzerland) and a graphite rod served as reference and counter electrode, respectively. All potentials within this paper were converted to the reversible hydrogen electrode (RHE) scale. Electrochemical measurements were done in 2 M NaOH (EKA, 98 wt.%) solution, with or without 0.03 atmosphere (N 2 flow rate of 35 mL min −1 ), in two steps. In the first step, ionic liquids were heated up to 400 • C and maintained at this temperature for 1 h. In the second step, the materials were heated up to 800 • C for another hour. In both steps, the heating rate was 10 • C min −1 . After the treatment at 800 • C, the materials were left to cool to ambient temperature and denoted as Pt/c-IL, Ni/c-IL, and Au/c-IL.
X-ray diffraction data were collected on a Philips PW 1050 (Philips, The Netherlands) diffractometer with Cu-Kα 1,2 radiation (Ni filter) at room temperature. Measurements were done in the 2θ range of 10-90 • with a scanning step width of 0.05 • and 3 s time per step. Raman spectra were collected on an Advantage 532 Raman spectrometer (DeltaNu Inc., Laramie, WY, USA), with YAG lesser operating at 532 nm. Deconvolution of the spectra was performed using Gaussian fitting. Scanning electron microscope Phenom™ ProX Desktop SEM (ThermoFisher Scientific™, Waltham, MA, USA) with integrated energydispersive X-ray spectroscopy (EDS) detector was used to examine the surface morphology and composition.

NaBH 4 Oxidation Studies
The catalytic inks were prepared by ultrasonically dispersing 5 mg of the corresponding electrocatalyst and 0.6 mg of Vulcan carbon in 125 µL of a 2% polyvinylidene difluoride solution in N-methyl-2-pyrrolidone. The electrodes were prepared by depositing 10 µL of the catalytic ink onto a glassy carbon tip and left to dry at 80 • C overnight. This procedure was done immediately after the ultrasonic step, to ensure the good dispersion of the electrocatalyst.
Electrochemical measurements were performed using Ivium V01107 Potentiostat/Galvanostat (Ivium Technologies B.V., Eindhoven, The Netherlands) in a one-compartment cell. A saturated calomel electrode (SCE, Metrohm, Herisau, Switzerland) and a graphite rod served as reference and counter electrode, respectively. All potentials within this paper were converted to the reversible hydrogen electrode (RHE) scale. Electrochemical measurements were done in 2 M NaOH (EKA, 98 wt.%) solution, with or without 0.03 M NaBH 4 (98 wt.%, Scharlau), except for the experiments on the effect of NaBH 4 concentration. Cyclic voltammetry measurements were run in the −0.01-1.6 V potential window, at scan rates ranging from 5 to 250 mV s −1 (25 • C). Additionally, CVs were recorded at 50 mV s −1 in the 25-65 • C temperature range to verify the effect of temperature on the kinetics of the oxidation process. RDE experiments were done by running linear scan voltammetry experiments at 10 mV s −1 and rotation rates ranging from 0 to 1800 rpm. RDE experiments were also done with a commercial Pt/C (46.4 wt.% Pt, Tanaka, Japan) electrocatalyst for comparison purposes. Finally, voltammetric experiments in 2 M NaOH solutions containing different concentrations of NaBH 4 (0.01, 0.03, 0.06, 0.09 and 0.12 M) were run at 50 mV s −1 (25 • C).

Fuel Cell Test
The DBPFC tests were done by employing Pt/c-IL or Au/c-IL, as well as commercial Pt/C, as anode electrocatalyst and a Pt mesh (Johnson Matthey, A = 50 cm 2 , London, England) as the cathode (to ensure that the cathodic reaction is not limiting the DBPFC performance). A Nafion N117 cation-exchange membrane (DuPont, Wilmington, DE, USA) was used as a separator. Moreover, a 1 M NaBH 4 in 4 M NaOH solution was used as the anolyte, while a 5 M H 2 O 2 (30 vol.%, Carlo Erba, Barcelona, Spain) in 1.5 M HCl (37 wt.%, Sigma Aldrich, Missouri, MO, USA) solution was used as the catholyte. The polarization and power density curves were recorded for the DBPFCs, employing different anode electrocatalysts in the 25-65 • C temperature range.

Conclusions
Finally, [C 16 4 ] ionic liquids were successfully carbonized to produce carbon-supported metal nanoparticle electrocatalysts, as confirmed by XRD, Raman spectroscopy, and SEM/EDS analysis. Thus, prepared materials were systematically examined for BOR in highly alkaline media. Pt/c-IL, Ni/c-IL, and Au/c-IL electrocatalysts showed completely different behaviors in the presence of BH 4 − . Pt/c-IL gave the highest BOR current densities with two well-defined oxidation peaks and a shoulder (close to the first peak) on the direct scan. Reaction order was calculated to be close to 1 for all three electrocatalysts, with low apparent activation energy determined for BOR at Pt/c-IL and Au/c-IL. Number of electron transferred during BOR decreased in the order Pt/c-IL (5.0) > Au/c-IL (2.4) > Ni/c-IL (2.0). Moreover, Ni/c-IL showed unstable activity for BOR with time. DBPFCs' test run at 65 • C showed peak power density of 43.5 mW cm −2 for Au/c-IL anode and of 120.5 mW cm −2 for Pt/c-IL anode, indicating Pt/c-IL as the best choice for the anode electrocatalyst leading to the highest DBPFC performance.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.