Experimental Protocol for HOR and ORR in Alkaline Electrochemical Measurements

The anion-exchange membrane as electrolyte for alkaline fuel cells (AFCs) provides opportunities to develop non-precious metal (NPM) catalytic centers. This is certainly an advantage that AFC can offer to activate NPM toward the cathodic oxygen reduction reaction (ORR), with a performance approaching that of platinum (Pt). 1 in alkaline medium the hydrogen oxidation re- action is challenging, platinum that shows slug-gish kinetics active and stable nanostructured electrocatalyst and for AFCs the the The glass-corrosion products obtained by adding a known amount of borosilicate glass (Pyrex) powder into the fresh-electrolyte. Then, the polluted solution was stored at 60 ◦ C for 2 h to accelerate the corrosionofglass-powder.Oncethesolutionwascooled-downat25 ◦ C, measurements were performed. An aliquot of electrolyte was extracted for ICP-OES analysis.

The anion-exchange membrane as electrolyte for alkaline fuel cells (AFCs) provides opportunities to develop non-precious metal (NPM) catalytic centers. This is certainly an advantage that AFC can offer to activate NPM toward the cathodic oxygen reduction reaction (ORR), with a performance approaching that of platinum (Pt). 1 Nonetheless, in such an alkaline medium the hydrogen oxidation reaction (HOR) is challenging, even for platinum that shows a sluggish kinetics and regardless of the fact the mechanism is still not well-understood. 2 Numerous efforts to tailor high active and stable nanostructured electrocatalyst (including precious and non-precious centers) for AFCs have been performed. Moreover, important experimental issues have not yet being normalized. This is essential, in order to compare data generated in various laboratories. A series of experimental questions, such as, ink formulation, thin-film coated quality, catalyst loading, electrochemical cell, reference electrode, temperature, electrolyte cation and so on, can indeed affect considerably the electrochemical measurements. One can, moreover, find in the literature excellent protocols to normalize fuel cell electrocatalytic reactions in acid measurements, [3][4][5][6][7] using commercial Pt/C as benchmark material. Protocols for the alkaline medium are seldom, however, a recent work analyses the impact of impurities in alkaline electrolytes. 8 Moreover, to the best of our knowledge, the reported data to obtain reproducible and comparable results in alkaline media are still lacking, since different experimental protocols used for the evaluation of, e.g., benchmark Pt/C in alkaline medium, make it difficult to compare the HOR, and ORR kinetics. Reproducibility and replicability in the measurements is of paramount importance in order to assess good performance of advanced electrocatalytic materials. This work provides a protocol to perform, as a baseline, the HOR and ORR processes, generated on carbon supported platinum NPs in alkaline electrolyte.

An Overview on Alkaline HOR and ORR Data in the Last Decade
The research on fuel cell reactions (HOR and ORR) in alkaline medium reflects the principal development of the proton-exchange membrane fuel cells (PEMFC). The former approach is more focused on novel NPM catalytic centers. Commercial-based platinum electrocatalytic material in massif, single crystal surfaces or carbon supported nanoparticles have been used as a benchmark material to judge the performance of novel electrocatalysts. In this context, if some key * Electrochemical Society Member. z E-mail: Nicolas.Alonso.Vante@univ-poitiers.fr experimental parameters are not utterly considered, the reported kinetics, on the so-called best material (benchmark), is affected. For the above-mentioned fuel cell reactions on Pt-based catalysts, one can take the diffusional limiting current density, j dif , as a diagnostic parameter, if this parameter is independent of the applied electrode potential at a defined electrode rotating rate. Otherwise, one should consider it at an applied electrode potential in the diffusional region to evaluate the electrocatalysts. 9 Rotating-disk electrode (RDE) technique is commonly used in order to uniformly control the mass transfer by regulating the electrode rotation speed during experiments. The difference in the limiting current between measurements reveals the existence of several unexpected factors that affect the electrochemical kinetics. The measured limiting diffusion current for HOR and ORR, reported in the literature throughout the last decade, are summarized in Tables I and II, respectively. These data are further visualized in Figure 4. In the last decade, the data in the tables show some interesting facts: (i) the technological interest in AFC's has grown; (ii) the research on HOR catalysis has become very active since 2014, (iii) few HOR contributions, even on noble metals are scarce, and (iv) in spite of considering platinum as HOR, ORR benchmark material, the variation of the results proves the lack of experimental standardization in alkaline medium. This latter issue is further considered herein.

Chemicals and Materials
Electrolyte solutions were prepared with KOH and NaOH (pellets, ≥ 90% and ≥ 97% respectively, Sigma-Aldrich) and Milli-Q water (18.2 M cm). The electrolyte solutions were prepared in a polypropylene volumetric flask, cf. Electrochemical cell section. Isopropanol (≥ 99%) and 5% wt Nafion (Sigma-Aldrich) were used for the ink preparation. 20wt% Pt/C from Johnson-Matthey was used as a benchmark material.
Electrochemical measurements were carried out in a homemade three-electrode glass, and Teflon (PTFE) cells. Each measurement was carried out with fresh electrolyte, using a H 2 or O 2 bubbling-pressure of 100 kPa as previous discussed by Rheinländer et al. 12 A Pt mesh and a Hg/HgO (E = −0.92 V vs. SHE) were used as counter and reference electrodes, respectively, using a potentiostat/galvanostat (Autolab PGSTAT 30). For comparison, a reversible hydrogen electrode (RHE) was used as reference electrode. ORR polarization curves were background corrected following previous protocols. 3,10

Ink Preparation
High-quality electrochemical measurements require a thin and uniform film on the entire glassy carbon disk surface. 10 The film's quality relies on a delicate balance between the chemical nature and quantity of alcohol, catalyst-powder dispersion, drying conditions and the glassy carbon surface. 3 Glassy carbon disk electrodes must be polished with 0.05 μm alumina suspension in an "eight shape" pattern until obtaining a mirror-finished surface. This latter is rinsed with deionized water, sonicated for 10 min in deionized water and dried in air at room temperature at least 20 min before use. In order to prepare the catalytic suspension and the coating of glassy carbon electrode, the procedure proposed by Garsany et al. 3 was followed, with few modifications. In short, 5 mg of catalyst powder were suspended in 1 mL of stock solution, 3 and sonicated at least 30 min, Table III. An aliquot of 3 μL of the ink was carefully dropped onto the mirror-finished glassy carbon (GC) disk, geometric surface of 0.07 cm 2 . The ink was, thereafter, dried during 20 min, using a rota- tion rate of 700 rpm at room temperature. Figure 1 compares our final coated thin film catalyst, dried using rotation and without rotation. One can appreciate a significant difference. Definitively, the rotationdrying thin-film leads to a uniform layer covering the GC surface in comparison to a lack of rotation. Homogeneous thin-films of the order of 0.2 μm are ideal, since thicker ones produce high mass-transport resistance through the catalyst layer, 3 which limits the whole utilization of the electrode's electrochemical active surface. The evaluation of coated catalyst thin-films was already discussed by Garsany et al. 3  The metal loading, L M [μg cm −2 ], is another important factor used to normalize the kinetic parameters. This parameter is calculated using: where %m is the mass percentage of metal (and/or electrocatalysts) content in the powder, w is the mass (mg) of catalyst powder, V' is the volume (μL) dropped onto the disk, V is the suspension volume, (mL) and A geo is the geometric surface (cm 2 ) of the glassy carbon. Considering the nominal loading of Pt for the benchmark Pt/C, L M is ca. 42 μg cm −2 geo . The analytical techniques, e.g., inductively coupled plasma-optical emission spectroscopy (ICP-OES) and thermogravimetric analysis (TGA), are widely used to determine the real metal content in powder form electrocatalysts.

Reference Electrode
Per definition, an electrode used as a reference electrode (RE), referred to the standard hydrogen electrode (SHE) must have a constant potential, independent of the electrolyte nature, and of the analyte concentration. In alkaline measurements, a reference electrode must not contaminate and not be contaminated by the species in the experiment. Common references are electrodes, reported in the literature for alkaline measurements, include the reversible hydrogen electrodes (RHE), saturated calomel electrodes (SCE) and Ag/AgCl electrodes. However, these electrodes as reference do not completely satisfy the conditions of a real reference electrode in alkaline environment. For example, RHE, which is a popular reference in acid electrolytes, 3 is not stable at high pH values, due to the fact that Pt/H 2 equilibrium is perturbed by OH − species. This is certainly, a reason to use chloride-containing reference electrodes, such as Ag/AgCl, and saturated calomel electrode (SCE) for alkaline measurements. We must consider that there is a probable contamination by the chloride ions, a concern for correct measurements on benchmarking materials or novel electrocatalytic material. In addition, if Ag/AgCl electrodes are exposed to a strong alkaline solution, AgCl could be converted to Ag x O, leading to a gradual shift of the reference electrode potential in the positive direction due to the mixed-potential of Ag/AgCl and Ag/Ag x O interfaces. 9 Hg/HgO reference electrode is, in this sense, a highly recommended RE to be used in alkaline environment. To illustrate the advantage of Hg/HgO and disadvantage of RHE, Figure  2 depicts the cyclic voltammograms on three independent measurements made on 20wt% Pt/C (JM). The measurements, performed in the same alkaline electrolyte, were done in a Teflon cell provided with Hg/HgO (RE), Figure 2a, and in a glass (Pyrex) cell with RHE (RE), Figure 2b.
Although similar finger prints for Pt in alkaline medium were obtained, we can, however, appreciate important differences, namely, the geometric current density of the cyclic voltammetry profile measured in the PTFC cell is higher than that obtained in the glass cell. This fact suggests that the Pt surface can be blocked by products of glass (corrosion) etching. 11 On the other hand, the surface reaction waves are not at the same electrode potential when using the RHE. This is not observed when using Hg/HgO. It is clear that a Hg/HgO reference electrode is the best option in alkaline medium.

Electrochemical Cell
Most electrochemical measurements performed in acid or alkaline electrolytes are carried out in standard three-electrode electrochemical cells made of borosilicate glass. It is well established that any trace of contamination in acid electrolytes impacts negatively the experimental measurements conducted, e.g., via the rotating disk electrode (RDE). 4 The pH of the electrolyte is a critical parameter to be considered in the use of glass-based electrochemical cells. It is established that silicateglasses are less stable at pH > 9, and at pH 11.5, the hydrolysis of Si-O-X (X: Si, Al, B) bonds can dramatically increase, 12 accelerating the process of destruction of the glass network. 13 Even though it is well-known that glass corrodes in alkaline solution, the corrosion products of glass are underestimated. Mayrhofer et al. 11 discussed the effect of the borosilicate glass corrosion in alkaline electrochemical medium. These authors concluded that silicates (SiO 2 ), as well as other glass components, used in the glass production process by, e.g., Pyrex, Schott Duran, Kimax, or Kavalier Simax, are dissolved in the electrolyte. The presence of such impurities in the electrolyte dramatically affects the active electrode surface of electrocatalysts, especially platinum, and consequently its kinetics. 14 The recorded RDE limiting diffusional current density, j dif , for HOR and ORR, is a function of the electrode rotation rate, ω (rad s −1 ). The theoretical j dif is given by Levich equation: where n is the number for transferred-electrons, F is the Faraday's constant (96487 C mol −1 ), D 0 is the oxygen diffusion coefficient, ν 0 is the kinematic viscosity, C 0 is the solubility of the gas (H 2 or O 2 ) in the electrolyte. In 0.1M NaOH and 0.1M KOH solutions, 15 D 0 = 1.9 × 10 −5 cm 2 s −1 , ν 0 = 8.7 × 10 −3 cm 2 s −1 , and C 0 = 1.22 × 10 −6 mol cm −3 . Pt surfaces are rather sensitive to trace impurities in the electrolyte, affecting dramatically the reaction kinetics and the magnitude of j dif . 3,4,10,11 This phenomenon usually occurs when SiO 2 , 11 Cl − , and Br − , 16 are species competing for the free adsorption sites on the catalyst's surface with the reaction substrate: molecular oxygen, or hydrogen. The HOR and ORR polarization curves were recorded in a PTFE cell. The glass-corrosion product, stored in alkaline solution, was added to a PTFE electrochemical cell to obtain the concentration reported in Table IV. Borosilicate species (Al, B, Si) were obtained by inductively coupled plasma/optical emission spectrometry (ICP-OES). The alkaline electrolyte solutions were prepared in a polymeric volumetric flask. Figure 3 shows the CVs, RDE measurements for hydrogen reactions (HOR/HER), and ORR. Figure 3 clearly shows the negative impact of the concentration of glass corrosion products. The well-defined surface reactions on Pt The iR-corrected HOR and ORR polarization curves, in absence of glass corrosion products, are within 10% experimental error 3 with respect to theoretical value for j dif , i.e., 6.16 mA cm −2 and 2.52 mA cm −2 , respectively, for ORR and HOR. Figure 3d displays, for the ORR, an initial limiting current j dif,0 decay of ca. 30%, whereas for the HOR this decay is ca. 20% at the highest concentration of Si species in the alkaline electrolyte. Under this condition, the twoelectron reaction pathways for the ORR is favored. The data of Tables I, and II are contrasted in Figure 4. A high dispersion of data for ORR is clearly observed, Figure 4b. Besides, most reported data lie below the theoretical j dif . On the other hand, the HOR is, apparently, less sensitive to the electrolyte contamination by glass corrosion  products. One has to recall that electrocatalytic measurements (e.g., ORR) is readily affected by, e.g., 1h residence time of fresh electrolyte in a glass cell.

Ohmic Drop
In an electrochemical measurement, the total measured potential, E T , is defined as: Where, E k , is the kinetic-potential, and the potential-drop (ohmic drop) is due to the contributions, e.g., electrolyte (iR elec ); electrical connections (iR cable ); and working to reference distance (iR distance ), so that Eq. 3 can be simply rewritten as Eq. 4: where i is the measured current and R the system's resistance. This iR term must be subtracted from E T to obtain E k for every electrochemical kinetic measurement. The most common technique to determine R is the electrochemical impedance spectroscopy (EIS) by linearly extrapolating at the realaxis intercept in the Nyquist diagram or the independent variation of the |Z| at high frequency (Bode plot), Figure 5. For Pt electrodes in acid electrolytes, R was estimated within 2 to 10 . 3 The ohmic drop in alkaline medium is more important than in acid medium. For Pt, in 0.1M KOH, we obtained a resistance within 40 to 45 .
The EIS in both media at 0 V/RHE shows that the adsorption behavior at the Pt/C is dominated by the nature of the interacting species, namely, H 3 O + , OH − and H 2 O. Figure 6 shows the HER/HOR and ORR polarization curves Pt/C (JM), measurements carried out in the PTFE cell, using Hg/HgO RE, before and after ohmic-drop compensation in alkaline medium. For practical purposes, the applied electrode potential is quoted to the reversible hydrogen electrode (RHE) scale, using (E RHE = E Hg/HgO + 0.059 pH + E • Hg/HgO ).

Cation Electrolyte
Sodium hydroxide (NaOH) and potassium hydroxide (KOH) are common electrolytes employed in half-cell alkaline medium. Due to the different atomic nature between sodium and potassium cations, some physicochemical properties of the solution are completely different. Therefore, the surface electrochemistry on Pt in presence of NaOH or KOH is expected to be different. This effect is also underestimated in the literature. Jin et al. 17 concluded that the ORR kinetics is superior in KOH than in NaOH solutions. In both cases the ORR is not favored with the alkaline concentration increase, since the protonation of the superoxide is suppressed. The results, on Pt/C in 0.1M NaOH   17 results. This effect is not noticeable for the hydrogen oxidation reaction, Figure 7b.
The characteristic surface reactions of platinum, in KOH and NaOH electrolytes, are well defined. Moreover, the intensity and position of the peaks are slightly different. Taking as reference the profile recorded in KOH, one can appreciate that the main surface reaction peaks, in the H upd region, Figure 7a, is shifted by +10 mV (anodic peak) and −20 mV (cathodic peak). Based on DFT calculations, Mc-Crum et al. 18 anticipated that hydrogen is not the only species adsorbed in the H upd region. The OH − species and cation also participate in the adsorption process. The wave assigned to the formation of Pt-OH is shifted +15 mV in NaOH. This observation suggests a stronger binding energy between Pt surface atoms and OH species. The oxide reduction wave in both electrolytes is centered at the same potential.
As mentioned above, one assesses that HOR and HER processes are not significantly affected by the nature of the alkali cations, see inset in Figure 7b. On the other hand, for ORR one can appreciate an important negative effect on the half-wave potential in the NaOH electrolyte. Certainly, the hydrated cations interact with adsorbed oxygen species affecting the ORR electrocatalytic performance. 19 The interaction between oxygen-species covalently bound to Pt surface atoms, in the inner Helmholtz plane (IHP), and the non-specifically adsorbed hydrated cations in the outer Helmholtz plane (OHP), Figure 7d, at the origin of the observed phenomenon, is in agreement with previous studies on Pt(111) surfaces. 19 The interaction of ions and water molecules are strong for the system: O 2 -M( = Na or K)-OH-H 2 O due to the charge that the ionic species carry and the polarity of water molecules. 20 Molecular oxygen is hydrophobic in nature, thus poorly dissolved. Further, K + possesses a bigger atomic radius and a smaller charge density than Na + . Therefore, the interaction K + /H 2 O is lower than that of Na + /H 2 O, 20 Figure 7d. In other words, under the same conditions, the ORR performance is inhibited by the lower solubility of O 2 in a NaOH solution as compared with a KOH solution.

Conclusions
Uncertainty of experimental results reported so far in alkaline medium desires the need to standardize and validate the measured performance of the Pt benchmark catalyst, in order to guarantee the experimental results of the novel electrocatalysts. Herein, we provide experimental evidences regarding the effect of some factors, e.g., cell, reference electrode, catalyst mass loading, etc. that impinges the electrocatalytic materials' performance in alkaline medium. Summing up, the electrochemical cell of choice for measurements in alkaline medium is Teflon and Hg/HgO as reference electrode. The electrochemical measurements in acid should be iR-corrected, as well as in alkaline electrolytes. The use of KOH as an electrolyte is, apparently, the electrolyte of choice for measurements in alkaline electrolyte.