The Hydrogen Evolution Reaction on Rhenium ...

Electrocatalysis DOI 10.1007/s12678-014-0240-z

The Hydrogen Evolution Reaction on Rhenium Metallic Electrodes: A Selected Review and New Experimental Evidence R. Garcia-Garcia & G. Ortega-Zarzosa & M. E. Rincón & G. Orozco

# Springer Science+Business Media New York 2014

Abstract The volcano plots reported in the field of electrocatalysis utilize data with a difference of three orders of magnitude between the worst and the best rhenium electrocatalytic activity toward the hydrogen evolution reaction (HER). However, the commonly used mean value of the exchange density current (j0) of the HER on rhenium is log j0 =−2.9 A cm−2, which is higher than the value used for platinum (log j0 =−3.3 A cm−2). This fact seems to contradict Sabatier’s principle and points to the possibility that this value corresponds more to rhenized surfaces than to metallic rhenium. Rhenized surfaces are primarily composed of a mixture of oxides; therefore, the electrocatalytic behavior is attributed to these thin films rather than to metallic rhenium. In addition, a selected review of rhenized electrodes is included herein because these issues have not been considered in the electrocatalysis literature at the present time. We initially believed that the kinetic parameters might have been overestimated due to the formation of rhenide ion or rhenium hydride species; however, no evidence of the formation of these species was found. Our experimental mean value of the R. Garcia-Garcia Universidad Tecnológica de San Juan del Río, Av. La Palma No. 125, Col. Vista Hermosa, San Juan del Río, Qro, Mexico R. Garcia-Garcia : G. Ortega-Zarzosa Facultad de Ciencias, Universidad Autónoma de San Luis Potosí, Av. Dr. Manuel Nava, Zona Universitaria, San Luis Potosí, Mexico M. E. Rincón Instituto de Energías Renovables-Universidad Nacional Autónoma de México, Privada Xochicalco S/N, Col. Centro, 62580 Temixco, Morelos, Mexico G. Orozco (*) Centro de Investigación y Desarrollo Tecnológico en Electroquímica, Parque Tecnológico Querétaro, Sanfandila, Pedro Escobedo C. P. 76703, Querétaro, Mexico e-mail: [email protected]

exchange current density of the HER on metallic rhenium is 7×10−5 A cm−2 in acidic solution. Therefore, our results are in accordance with Sabatier’s principle, which states that a weak adsorption energy of hydrogen on rhenium (energy, 6.9 kJ mol−1) results in a slow rate of reaction (log j0 = −4.2 A cm−2), whereas an intermediate adsorption energy of hydrogen on platinum (12 kJ mol−1) produces a fast reaction (log j0 =−3.3 A cm−2). Keywords Hydrogen evolution reaction . Sabatier’s principle . Rhenized surface . Metallic rhenium . Volcano plot

Introduction The hydrogen evolution reaction (HER) has been comprehensively investigated on several metals. However, there are few studies that have examined the electrocatalytic activity of rhenium electrodes towards the HER [1–5]. Mendez et al. [6] reviewed the electrocatalytic activity of rhenized surfaces and noted that the kinetics of the HER on these surfaces is not fully understood. Furthermore, Trasatti [7] has noted that the value of 30 mV per decade for the Tafel slope of rhenium is a notable exception. The major drawbacks of rhenium as a catalyst include its limited availability in the Earth’s crust (1 ppb) and the need for intricate processing methods, which result in high fabrication costs. Consequently, researchers have studied rhenium thin films (rhenized surfaces) more extensively than metallic rhenium. Note that the rhenized electrodes are typically obtained by electrodeposition of soluble oxides of rhenium. This study provides a new experimental data for the HER on pure metallic electrodes. In addition, a selected review of rhenized electrodes and the rhenide ion is also included in this study because these issues have not been considered in the present electrocatalysis literature [7–12]. In a forthcoming second

Electrocatalysis

study, the analyses will be expanded to include kinetics data at different acids and bases concentrations. These discrepancies caught our attention and provided the motivation for this study. Moreover, the aim of this work is to contribute to the understanding of hydrogen reactions on metallic rhenium. To achieve this objective, the following steps were performed: (1) the possible formation of rhenide ion was studied by coulometric measurements; (2) new exchange currents densities were obtained by polarization curves and impedance spectroscopy; and (3) the electrocatalytic activity of metallic rhenium for the HER was analyzed in the context of Sabatier’s principle.

Electrocatalysis in the Cathodic Hydrogen Evolution Reaction The overall reaction of the HER in acidic media is described by reaction (1): 2Hþ þ 2e− ↔H2 :

ð1Þ

Fig. 1 Hydrogen chemisorption energy per atom calculated by Nørskov [11] (one monolayer of hydrogen coverage) and the exchange current densities reported in [10]. The dash line is drawn as a guide to the eye

inclusion of impurities, which are unstable in aqueous solutions and gradually transform into rhenium oxides. −

− 4ReO4 ðadsÞ þ 20Hþ ðadsÞ þ 16e →Re þ 3ðReO2 •2H2 OÞ

The electrocatalytic activity for the HER (following Sabatier’s principle) shows a periodic variation of the exchange current densities with increasing atomic number of the metals across the periodic table. The electrocatalytic activity is typically described using a volcano plot, which shows the standard exchange current densities versus the adsorption energies of a hydrogen atom on several metals. The volcano plots commonly displayed in textbooks show similar electrocatalytic activities for rhenium and platinum [7–12], as seen in Fig. 1. Although modern versions of the volcano plot have been recently calculated using Density Functional Theory [8, 11, 12], the trend of the electrocatalytic activities of rhenium and platinum remains unchanged.

The Hydrogen Evolution Reaction on Rhenized Electrodes The electrodeposition of rhenium has been extensively reviewed by Gileadi and Eliaz [13], who have concluded that the electroplating of pure rhenium suffers from a low Faradaic efficiency and poor coating quality. Pure metallic surfaces are difficult to produce by electrolysis from aqueous solutions, mainly due to the poor reaction kinetics, leading to the formation of rhenium and rhenium oxides. Reaction (2) is an example of the electrodeposition reaction of rhenium from its soluble oxides [14]; these reactions are influenced by the adsorption of protons during rhenium electrodeposition [6, 13, 14]. Therefore, the adsorption of hydrated rhenium oxides or rhenium hydrides on the electrode surface leads to the

þ 4 H2 O

ð2Þ

Table 1 shows the kinetic parameters of the HER on rhenized surfaces reported in the literature [1–5, 15]. In 1950, Pecherskaya et al. [1] studied the HER on 15 different cathodes, including rhenium electrodes. It is worth mentioning that the results obtained by these authors must be revised because the metallic character of the electrodes is unclear. These authors measured an exchange current density of approximately 10−3 A cm−2 and an overpotential of 0.09 V at a current density of 5×10−3 A cm−2. In 1981, Krasikov [3] obtained a Tafel slope of 30 mV per decade on various rhenized electrodes. In 1998, Szabo [4] claimed that a rhenized-platinum electrode exhibits a Tafel slope of 31 mV per decade. Although the slopes observed by Krasikov [3] and Szabo [4] suggest the same mechanism for various rhenized electrodes, the slope observed by Pecherskaya et al. [1] suggests another mechanism (see Table 1). In 2001, Schrebler [16] reported that the HER on a rhenized-Au electrode occurred preferentially on the electrodeposited rhenium rather than on the Au surface. A comparison of the electrocatalytic activity of the rhenized surfaces toward the HER, as shown in Table 1, is challenging because of several factors: (1) the surfaces are a mixture of metallic rhenium and various rhenium oxides (ReO3, ReO2, and Re2O3) [17]; (2) the evolution of hydrogen gas during the electrodeposition process affects the surface morphology (i.e., surface area and crystallinity) and

Electrocatalysis Table 1 Reported kinetic parameters for the HER on various electrodes of rhenium in acidic solutions at 25 °C

Electrode

Media

Exchange current density (A cm−2)

Tafel slope (mV per decade)

Reference

Rhenized Metallic Metallic Rhenized Rhenized Metallic

1 M H2SO4 0.037 M HCl 0.145 M HCl 0.25 M H2SO4 0.5 M H2SO4 0.5 M H2SO4

1.34×10−3 1.14×10−5 7.3×10−6 No data No data No data

140 43 29 30 31 66

[1] [2] [2] [3] [4] [5]

consequently the electrocatalytic activity; and (3) the presence of metallic rhenium with occluded hydrogen gas is very commonly reported [4, 6, 14, 17, 18]. Consequently, the absorption of hydrogen must be taken into account when evaluating the electrocatalytic activity.

The Hydrogen Evolution Reaction on Rhenium Metallic Electrodes In 1965, Joncich et al. [2] studied the HER in dilute HCl solutions using polarization curves recorded at different temperatures. The exchange current densities at 25 °C were in the range of 10−6 to 10−5 A cm−2. In 1976, Miles and Thomason [19] used cyclic voltammetry to study the HER in 0.1 M H2SO4 at 80 °C on 31 metals. These authors observed lower electrocatalytic activity for rhenium than platinum and classified the elements according to the following order of decreasing activity: Pd>Pt>Rh>Ir>Re>Ni. In 1989, Gomez et al. [20] determined for the first time the electrochemical behavior of metallic rhenium in aqueous solutions. These authors reported that the electrocatalytic activity at potentials of less than 0.1 V was evident at pH −0.7 and postulated that this was a result of metallic rhenium because the surface was not covered by oxide species. However, no information regarding the kinetics of the HER was reported by in that study. More recently, in 2005, Chun et al. [5] studied the overpotential deposited hydrogen for the cathodic HER on metallic rhenium in acidic media, although no exchange currents of the HER were reported [5]. In parallel research in alkaline media, Miles [21] and Tilak [22] studied the kinetic parameters for the HER on several metals. In 1975, Miles [21] observed that palladium and platinum were the best electrocatalysts and the performance of rhenium was similar to that of Ni, Fe, Rh, and Ir. In 1986, Tilak [22] determined that the exchange current density of rhenium (0.002 mA cm−2) was an order of magnitude lower than that of platinum (0.022 mA cm−2) at low polarization. However, at high polarization, the exchange current density of rhenium (1.0 mA cm−2) was slightly higher than that of platinum (0.8 mA cm−2).

Rhenide Ion (Re−) The studies on the HER [1–5, 7–12, 19, 21] summarized in the previous sections have not considered the formation of rhenide ions (Re−), whose standard formation potential lies in the potential region of the HER (Fig. 2a). Only Gomez [20] has noted that the HER at rotating ring-disk electrodes confirms the participation of Re− when the electrode is polarized at potentials close to −0.2 V. Consequently, there is a possibility that the standard exchange current densities have been overestimated and that the electrocatalytic activity of rhenium is actually lower than previously reported. The first claim for the existence of a monatomic metal anion, rhenide ion (Re−), in an aqueous solution was made in 1937 by Lundell [23]. Some years later, Pauling [24], Bravo [25], Cobble [26], and Ginsberg [27] postulated that the aqueous rhenide ion was almost certainly not a simple anion, such as halogen ions, but rather some type of oxygenated complex or a hydride. In 1973, Magee [15] reviewed the early polarographic studies of the perrhenate ion reduction. These studies seemed to support the existence of Re− in solution because the polarograms systematically showed an 8-electron reduction according to the following reaction: −

ReO4 þ 4H2 O þ 8e− →Re− þ 8OH−

ð3Þ

Recent studies have indicated that the solid of a “rhenide” ion is actually a hydridorhenate complex, K2ReH9 (oxidation state VII (d0)) [28]. The solid compound K2ReH9 slowly yields hydrogen gas when dissolved in water. In 2006, Ellis [29] reviewed the studies of highly reduced carbonyl complexes formally containing transition metals in their lowest known oxidation states. Although he concluded that there was no compelling evidence for the rhenide ion in solution, alternative reactions for the electrochemical reduction of the perrhenate ion (reaction (3)) were not provided. Therefore, the nature of the species in solution still remains unknown, although it is unlikely to be a low-valent species such as Re−. Despite the fact that there was no convincing confirmation of the existence of rhenide ion in solution, the Pourbaix diagram [30] and many encyclopedias of electrode potentials

Electrocatalysis Fig. 2 Variation of the potential redox couples of metallic rhenium with pH at 25 °C. a The Nernst potentials of the Re/Re− redox couple. b The Nernst potentials of rhenium hydride species. The potentials were calculated based on the equations listed in Table 3

[31–35] include the Gibbs energy and the standard potentials of the rhenide ion formation reactions. For example, Takeno [35] reported the Re(I+)/Re(I−) redox couple at −0.23 V. The Nernst potentials of the Re/Re− redox couple, reaction (4), are summarized in Table 2. Re þ e− →Re−

ð4Þ

possibility that the currents observed by Tilak [22] at low polarization only included reaction (1) and that the currents measured at high polarization may involve reactions (1) and (6).

Experimental Materials and Chemicals

E ¼ E 0 −0:0591logaRe−

ð5Þ

Figure 2a shows the equilibrium potentials and pH dependence of various reactions of metallic rhenium. However, Bratsch [36] has stated that the standard chemical potentials correspond to rhenium hydride species Re þ e‐ þ Hþ →ReH

ð6Þ

rather than rhenide ions (Eq. (4)), where ReH in Eq. (6) corresponds to various rhenium hydride species. The potentials can then be estimated using the following equation: E ¼ E0 þ

2:303RT aHþ aRe log 1F aReH

ð7Þ

Equation (7) was used to create the diagram shown in Fig. 2b, which shows that the standard potential values for the half-reactions of Re(0) decrease as the pH value increases, neglecting the erroneous prediction of Fig. 2a (Eq. (5)) that metallic rhenium disproportionates above pH 6 [36]. Figure 2b establishes the possibility that the kinetic parameters of the HER on metallic rhenium were overestimated due to the formation of hydride species. For example, there is a

The working electrode was a rhenium wire (Sigma-Aldrich (USA), 99.9 %) with a geometric area of 0.161 cm2. A Hg|Hg2SO4|K2SO4 sat. electrode was used as the reference electrode and a platinum wire (99.99 %) was used as the auxiliary electrode. All potentials were recalculated with respect to the reversible hydrogen electrode (RHE) in the working solution. All reagents were of analytical grade purchased from commercial sources and were used without further purification. For example, sulfuric acid (98.08 %) was purchased from J.T. Baker. Aqueous solutions were prepared with Millipore Milli-Q Plus treated water with a conductance of 11 MΩ cm−1. Hydrogen gas (99.999 %) was supplied by Infra Praxair México. Carbide sandpaper (2000 SIC B-99), purchased from Buehler LTD, was used for polishing the working electrode, which was then rinsed with distilled water and allowed to dry in air. The sandpaper was cleaned with solutions of H2SO4/HNO3 before the polishing process. Immediately after the polishing process, an electrochemical characterization was performed. Alternatively, after polishing the electrode, the surface was also cleaned by immersion in a Table 2 Standard potentials and energies of formation of the rhenide ion at 25 °C

E0 (V)

ΔG (kJ mol−1)

Reference

−0.4 −0.1 −0.1 −0.4 −0.1

38 10 No data No data 10.1

[8] [9] [10] [11] [12]

Electrocatalysis

concentrated H2SO4/HNO3 solution or concentrated NaOH solution and rinsed with triply distilled water. No significant differences were observed in the data described in the following sections using the different procedures. A three-electrode thermostat jacketed glass cell from Provitex (Mexico) with a volume of 25 ml was utilized. Quartz cuvettes were used for UV-visible spectroscopy. All experiments were performed in triplicate at 25 °C.

measurements were obtained in the frequency range of 0.1 to 105 Hz. Seven frequencies per decade were scanned using a 10 mV peak-to-peak sinusoidal signal. A complex nonlinear least squares fit of the impedance data were obtained using the ZView 3.0 software package. The experimental data points were fitted using a different model circuit and verified by the Kramers–Kronig relations. The exchange current (j0) was calculated from the equation

Coulometry

j0 ¼

Coulometric experiments were performed at a constant potential of −0.3 V for a duration of 3 h, and the total amount of charge (Q) was obtained by integration of the current (I) versus time (t) plot. According to Faraday’s law, the amount of rhenide ion formed during the electrolysis is directly proportional to some quantity of electricity that passes through the cell. The Faradaic yield was arbitrarily considered to be 20 %. A standard three-electrode configuration was employed, and the experiments were performed using a BioLogic model EPP-4000 potentiostat-galvanostat. The solutions of 0.5 M H2SO4 were saturated with purified hydrogen for 40 min before the experiments, and a hydrogen gas atmosphere was maintained in the cell during the experiments. After the electrolysis, the solutions were analyzed using a Thermo Scientific (USA) model 222–262000 UV-visible spectrophotometer. The ultraviolet spectrum (200–800 nm) of the electrolysis solutions was compared to those of standard solutions of 0.5 M sulfuric acid (Sigma-Aldrich (USA)).

RT nFRCT

ð8Þ

where j0 is the exchange current density, F is the Faraday constant, T is the temperature, R is the gas constant, RCT reflects the charge transfer kinetics, and n=2 is the number of electrons involved. Equation (8) is valid when the overpotential is very small [37, 38]. E-pH Diagrams E-pH diagrams were calculated from the standard potentials reported in [30–33]. The Nernst equations used for the E-pH diagrams are listed in Table 3. The diagrams were calculated assuming a temperature of 298.15 K and a concentration of soluble species of 10−6 M.

Results Electrochemical Characterization The polarization curves were measured using the following method. The electrode was prepared as previously described and the solutions were bubbled for 40 min with hydrogen gas before each experiment. The electrode was then immersed in the solution and a potential of −0.1 V was immediately applied for 2 h. Hydrogen gas was flowed over the solution during the equilibration time (2 h) and during the evaluation of the polarization curve. Next, potential sweeps were performed from 0 or 0.1 to −0.22 V using scan rates of 0.116, 1, or 10 mV s−1. Impedance spectroscopy measurements were conducted using a fresh electrode, prepared as described in the previous section. The experiments were performed using an Autolab PGSTAT 302 potentiostat-galvanostat (Ecochemie, Netherlands) with a frequency analyzer module. The solutions were bubbled for 40 min with hydrogen gas before each experiment, after which a potential of −0.1 V was applied for 2 h under a hydrogen atmosphere. Subsequently, alternating current impedance measurements were performed at a −0.1 V cathodic overpotential. A hydrogen atmosphere over the solution was maintained during the experiments. The

Coulometric Experiments and Ultraviolet–Visible Absorption Spectroscopy Coulometric experiments at constant potentials were performed to obtain ions in solutions according to Eq. (4), an example of which is shown in Fig. 3. We speculated that if the Faradaic yield was 20 %, then the concentration of soluble rhenium was approximately 10−3 M, which may be detectable by UV–vis spectroscopy. Indeed, in aqueous solutions, charge transfer bands of soluble Re ions have been reported in the − UV-visible region: blue ReO4 [15, 23, 27], red soluble rhenous ion (Re3+) [15], violet mixed Re(V)-Re(VI) species [39–41], and blue [Re2Cl8]2− [15]. In our experiment, no UV– vis signals were observed in the range of 200–800 nm (the spectrum is not shown for the sake of brevity). The solutions were colorless before and after the experiments, and considerable bubble formation was observed at potentials lower than −0.25 V. These results eliminate the possibility of the formation of acidic rhenide solutions or hydride complexes and are in contrast with the studies of rhenized surfaces with large percentages of oxides, in which these formations have been

Electrocatalysis Table 3

Nernst equations and half-reactions involving rhenium and hydrogen species

Reaction

Number of reaction

Nernst equation

Number of equation

2H+ +2e− →H2 Re+e− →Re‐ Re+e− +H+ →ReH Re3+ +3e− →Re

(1) (4) (6) (10) (12) (14) (16) (18)

E=−0.059pH E ¼ −0:400−0:0591logaRe− E=−0.1−0.0591pH E ¼ 0:300 þ 0:0197log aRe3 E ¼ 0:369−0:0675pH þ 0:0084logaReO‐4 E=0.227−0.0591pH E=0.276−0.0591pH E=0.399−0.0591pH

(9) (5) (7) (11) (13) (15) (17) (19)

−

ReO4 þ 8Hþ þ 7e− →Re þ 4H2 O Re2O3 +6H+ +6e− →2Re+3H2O Re+2H2O→ReO2 +4H+ +4e− ReO3 +2H+ +2e− →ReO2 +H2O

Redox couples containing rhenium or hydrogen and their respective Nernst equations [15], in which the activities of solid species are unity and the activities of soluble species are approximated as equal to the concentration (10−6 mol kg−1 ). The Nernst equations were used for the calculations of the lines on the pH-E diagram in Fig. 2

reported [14–18]. In addition, the measured voltammograms (Fig. 4a) do not show any peaks in the zone of stability of rhenium in aqueous media. A very small crossover of the current plots appeared at −0.16 mV, which demonstrates that the electrodeposition of Re− ion onto its own metal surface did not occur. Electrochemical Characterization Figure 2 shows the potential of the redox couples of metallic rhenium as a function of pH. The vertical dashed lines show that metallic rhenium exhibits a very small zone of stability in aqueous media. Furthermore, it is well known that all metals, with the exception of gold, are covered by a native oxide layer in air. When immersed in an electrolyte, in favorable cases (e.g., the Pt group metals), this oxide can be reduced. In all of the experiments, once the electrode was introduced into the solution, a potential corresponding to the zone of stability of Re(0) was applied as quickly as possible. This potential was applied over several hours with bubbling of an inert gas to

Fig. 3 Coulometry at a constant potential of −0.3 V for a duration of 3 h. The total amount of charge (Q) is obtained by integration of the I versus t plot

reduce the native oxide layer. Therefore, it is assumed that all of the results described in the following discussion correspond exclusively to a metallic rhenium surface. Additionally, we considered the corrosion of rhenium to be negligible at open circuit potentials because this metal is resistant to sulfuric acid, hydrochloric acid, seawater, and aqua regia at room temperature [42]. Gomez et al. [20] observed that the open-circuit potential decays over several hours in solutions for a wide pH range. Therefore, the potential applied once the electrode was immersed in the solution was more cathodic than the corrosion potential determined by these authors. The voltammogram in Fig. 4a shows that the HER began at −0.11 V in acidic media. It was observed a crossing over of the cyclic voltammogram at −0.14 mV. This cyclic voltammetry was obtained under slow scan rate (10 mV s−1.), while the voltammogram in Fig. 4b was recorded at a potential scan rate of 500 mV s−1. When the upper potential limit was increased from 0.2 to 0.88, new peaks appeared in the voltammogram. The currents’ intensities of both peaks increase as the upper potential limit moves towards more positive values (Fig. 4c). Also, high scan rates (100–500 mV s−1) create peaks with large currents; for example the current at 1 V is produced due to the formation of massive oxide ReO4. Gomez [20] postulated that a Re(III) soluble specie was formed on the electrode at 0.5 V. According to our results, this soluble specie was only obtained if the upper limit was superior to 0.6 V. The deposition of this soluble oxide establishes the crossovers in the cyclic voltammograms. Therefore, the potential–log current plot in Fig. 4d was obtained with an upper potential limit of 0.1 V to avoid the formation of soluble oxides. The experimental potential–log current plots in hydrochloric acid and sulfuric acid are comparable, and their values of exchange current densities are listed in Table 4. These potential–log current plots in sulfuric and hydrochloric acids are similar to the plots obtained by Joncich [2] in acid chloride media for metallic rhenium. Figure 4c shows that the current at −0.36 V was −40 mA cm−2, whereas the current was approximately 0.3 mA cm−2 when the potential range was between 0 and

Electrocatalysis

Fig. 4 a Second cycle of the voltammetry of metallic rhenium in 0.5 M H2SO4 at a scan rate of 10 mV s−1; b voltammetry of metallic rhenium in 0.5 M H2SO4 at a scan rate of 500 mV s−1. After one cycle, the upper limit was increased; c voltammetry of metallic rhenium in 0.5 M H2SO4 at a scan rate of 500 mV s−1. The red line is a zoom of the cathodic sweep when the lower limit was increased up to −0.36 V. This voltammetry is

Table 4

presented with its respective current y-scale. d Potential–log current plot of metallic rhenium in 0.5 M H2SO4 in the region from 0.1 to −0.0 V at a scan rate of 0.116 mV s−1. The dotted line is the Tafel slope lying in the vicinity of −0.1 Vand the short dash-dot line is a linear region at the more cathodic potentials

Experimental exchange current densities for the HER on metallic rhenium at 25 °C in different acidic solutions (saturated with hydrogen gas)

Solution

The scan rate of the polarization curve or impedance potential (mV/s)

Exchange current density a A cm−2 curvature region

Exchange current density b A cm−2 linear region

0.037 M HCl 0.037 M HCl 0.145 M HCl 0.145 M HCl 0.5 M H2SO4 0.5 M H2SO4

1 0.116 1 0.116 10 1

5.26×10−5 ±2.41×10−5 7.11×10−5 ±3.70×10−5 2.27×10−5 ±1.27×10−5 8.72×10−5 ±1.93×10−5 1.16×10−5 ±3.2×10−6 2.51×10−5 ±1.8×10−5

4.4×10−7 ±1.48×10−7 1.22×10−6 ±1.10×10−6 3.29×10−7 ±7.41×10−7 2.02×10−6 ±9.13×10−7 1.21×10−6 ±9.58×10−7 6.51×10−7 ±4.28×10−7

0.5 M H2SO4 0.5 M H2SO4

0.116 −0.1 (impedance potential)

1.15×10−5 ±3.64×10−6 2.75×10−4 ±1.12×10−4

8.10×10−7 ±3.55×10−7 No data

Atmospheric pressure was 0.8 Atm. The electrical resistances of the solutions were determined by impedance spectroscopy and the average value was 141 s cm−1 Interval of potential where the current was determined was 0.0 to −0.12 V (doted line in Fig. 4b). The average current was j0 =7×10−5 or log j0 = −4.2 A cm−2

a

Interval of potential where the current was determined was −0.12 to −0.2 V (short dash-dot line). The average current was j0 =1×10−6 or log j0 = −6.0 A cm−2

b

Electrocatalysis

0.2 V. Consequently, Fig. 4a–c shows that metallic rhenium exhibits no electrocatalytic activity toward the hydrogen oxidation reaction. Thus, the real surface area of the polycrystalline rhenium electrode cannot be determined from the hydrogen adsorption or desorption charges. In our work and in the study by Joncich [2], the geometric area was used as a reference for the density current. It is not uncommon that when the real surface area is accounted for in the density current, the apparent enhancement of the electrocatalytic activity disappears. Consequently, the real current density can be smaller than the average density current of 7×10−5 A cm−2 reported in Table 4. The potential–log current plot in Fig. 4d shows a wide range of values compared to the data reported by Joncich [2]. Nevertheless, the exchange currents (Table 4) were determined according to the criteria of Reference [2], i.e., exchange currents were determined in the potential periphery around −0.1 V (dotted line). It is worth mentioning that a linear region (short dash-dot line) is clearly observed at more cathodic potentials, that is, the exchange currents vary from one behavior (dotted line) to another (short dash-dot line). However, both regions take into account slow kinetics and are far from the great performances observed in Fig. 1 for the metals belonging to the platinum group. The linear region in Fig. 4d (short dash-dot line) will be discussed later in this section. The experimental data points of the impedance measurements were fitted using a Randles model circuit because only a single time constant (one semicircle) was observed on the Nyquist plots. Such plots are not shown here for the sake of brevity. Chi-squared tests were used to measure the quality of the fit, resulting in values of approximately 10−4. The validity of the impedance measurements was demonstrated by performing a Kramers–Kronig transform analysis (pseudo chi-squared values were less than 10−5). The impedance spectra provide the electrolyte resistance and exchange currents shown in Table 4, which were calculated considering this resistance. The resistance of the electron-transfer process was used in Eq. (8) for the calculation of the exchange current density. Equation (8) was used because the electrode reaction is completely controlled by the electron-transfer kinetics at −0.1 V. Therefore, in Table 4, the average current density for the HER on metallic rhenium includes data obtained from the impedance spectra. Table 1 shows the large discrepancy between the reported kinetic parameters of the HER on rhenium in acidic media. This disagreement can be explained because the rhenized surfaces have a dissimilar nature compared to metallic rhenium. The volcano plot reported by Trasatti [10] used the standard exchange current density, as reported by Pecherskaya [1] in sulfuric acid media. However, the volcano plot by Kita [43] used data obtained from Joncich [2], who studied the HER in acid chloride media. Therefore, these plots display

data that differ by three orders of magnitude between the best and the worst rhenium electrocatalytic activity (see Table 1). Our study shows that the exchange current densities with chloride and sulfate (Table 4) ions are similar, indicating that the affinities of these ions for adsorption sites are comparable. Therefore, the differences between the volcano plots of Trasatti [10] and Kita [43] mainly originate from the surface oxides of rhenium, i.e., the data of rhenized electrodes were incorrectly attributed to metallic rhenium. To explain the behavior of metallic rhenium toward the HER, it is useful to consider the data compiled by Kopylets [44], who calculated the energy needed for the activation of hydrogen adsorption based on the experimental heats of adsorption of hydrogen on several metals, including rhenium. Figure 5 compares plots of the exchange current densities of metallic rhenium obtained in this work, the energies calculated by Kopylets [44], and the data reported by Appleby [45]. It is observed that in each row of d-metals, the energies increase as the d-shell is filled, except for nickel. The experimental current densities and the energies of hydrogen adsorption increased from left to right across each d-metal row in the periodic table, especially in the third-row metals where metallic rhenium is located. Therefore, Fig. 5 is consistent with Sabatier’s principle, which states that a weak adsorption (energies between 6 and 11 kJ mol−1) or strong adsorption (energies between 15 and 16 kJ mol−1) will cause the reaction to proceed too slowly (current densities between 10−8 and 10−4 A cm−2), whereas an intermediate energy of adsorption (energies between 12 and 14 kJ mol−1) is required for a rapid reaction (current densities between 10−3 and 10−2 A cm−2). The ability of a given metal to electrocatalyze the HER is routinely measured by the exchange current density, which is the rate of reaction (1) per surface area at the equilibrium potential. For over 50 years, the exchange current density of the HER has been plotted against the metal hydrogen bond energy, from which a volcano-shaped curve is obtained. Recently, a systematic approach that uses adsorption free energies calculated using DFT has been introduced. However, in this approach, the calculated adsorption free energies need to be paired with the experimental exchange current densities. Therefore, this study is focused on providing reliable exchange current densities for the corresponding theoretical calculations. Consequently, a discussion regarding the mechanism of the HER will be reported in a future publication. Given that the reported volcano plot in Fig. 1 was derived from data obtained from rhenized electrodes, Table 1, and the fact that rhenium, cobalt, and nickel have very similar hydrogen chemisorption energies [8, 11], it can be falsely concluded that the electrocatalytic activity of rhenium does not rigorously follow Sabatier’s principle. Our data confirms that metallic rhenium possesses the same electrocatalytic activity toward the HER as that of cobalt and nickel, i.e., metallic rhenium does indeed follow Sabatier’s principle. Therefore, it is

Electrocatalysis

Fig. 5 Activation energies for the adsorbed hydrogen atom calculated by Kopylets [44], the exchange current densities reported in [45], and the exchange current for the HER on metallic rhenium determined in this work. Behavior of the a first-row, b second-row, and c third-row metals

necessary to rewrite the question raised by Schmickler et al. [8] (“Why is Re such a good catalyst, whereas Co and Ni are such bad catalysts?”) when interpreting Fig. 1. In light of the findings of this study, we should instead ask: Why are the oxides of rhenium just as good electrocatalysts as platinum? One possible answer comes from Brewer-Engel theory [46], which indicates that the optimal d-electronic configuration for maximal electrocatalytic activity toward the HER is d8. Metallic rhenium has an electronic configuration of 5d5, whereas rhenium oxides possess configurations of d0, d1, d2, d3, and d4 [28]. This theory for bulk solids cannot directly explain the high electroactivity of rhenium oxides, which is governed mainly by the properties of the surface. Previous reports have determined that rhenized electrodes comprise a mixture of rhenium and mixed-valent rhenium oxides, e.g., Re2O5, ReO3, and ReO2 [6, 13–18, 47–49]. Some of these oxides with crystalline structures appropriate for hydrogen diffusion exhibit very high proton conductivities [50–52]. If these oxides act as proton exchange membranes with mobile protons, then the Tafel slope (30 mV per decade) and exchange currents of the

rhenized electrodes correspond to the platinum substrate. This result explains why the polarization curves in acidic media observed by Szabó [4] for platinum with and without a coating of rhenium oxide are indistinguishable. There is an uncertainty associated with the estimation of the exchange current density (Table 4) because the data selected by Joncich [2] exhibits a high degree of variation in its curvature (dotted line in Fig. 4d). Note that our data show a wider range of values compared to the data reported by Joncich [2]. Therefore, Fig. 4d exhibits a new linear region (short dash-dot line on Fig. 4d) that can be assigned to the overpotentially deposited-hydrogen (OPD-H), which allows a new average exchange current density to be obtained (1× 10−6 A cm−2). This value enables us to claim that the activity of rhenium in Fig. 5c lies in the region that was previously selected, that is, the new datum does not alter the previous discussion. Conway and Jerkiewicz [53, 54] have concluded that the apex of the volcano plot arises at the zero of the standard Gibbs energy of dissociative chemisorption of hydrogen. The values of the energy of chemisorption of

Electrocatalysis

hydrogen on rhenium do not fall near this apex and consequently the exchange current is lower. These authors have found that the volcano relation can be differentiated in terms of four groups of metal surfaces according to their energy of chemisorption, and therefore rhenium can be assigned into one of these groups. Although there is a little evidence that allows for making a rigorous selection of one of these groups, there is one that seems the most likely option: the group of metals on which the HER occurs at metal-hydride phases covered by a barrier-layer oxide film. Additional relevant kinetics data determined were the Tafel slopes, which show notable differences depending on the selected potentials (see Table 5). Moreover, in the literature (Table 1), there are large discrepancies that significantly weaken the discussion presented by Joncich [2]. However, our data contain a wide range of values compared to the data reported in that study, and we observed a new Tafel slope of 67 mV per decade ((short dash-dot line in Fig. 4d), which is plotted on the overpotentially deposited-hydrogen (OPD-H) region. A Tafel slope of 60 mV per decade was observed by Chun [5], who commented that this slope was produced by the barrierless reaction of proton discharge [5]. Consequently, the Tafel slope of 67 mv per decade appears to imply that the proton reaction discharge is the rate-determining step of the HER at the polycrystalline Re in acidic medium. Jerkiewicz [55] has noted that the underpotential deposition of hydrogen (UPD-H) is characteristic of some platinum group metals, whereas other metals do not behave in this way. A pair of redox peaks at 0.24 V observed by Chun et al. [5] was assigned to the UPD-H on rhenium electrodes at pH 0.3. However, this peak pair appears close to the potentials thermodynamically predicted by the Nernst equations of three redox couples at pH 0.3, as seen in Fig. 2 and Table 3. Additionally, Gomez [20] assigned a very small peak at 0.04 V to hydrogen oxidation at pH −0.7. Furthermore, this author commented that redox reactions of rhenium can occur between 0.10 and 0.13 V. On the other hand, Chun et al. [5] cleaned the electrodes using a flame cleaning method and then cooled them in air. Keep in Table 5

mind that the rhenium is oxidized in a moist air atmosphere above 600 °C [13]; therefore, it is highly possible that some rhenium heptoxide was formed in the electrodes studied by Chun [5]. We are suggesting that some regions of the surface were covered by oxide and that these regions were the substrate of the UPD-H observed by Chun [5].

Conclusion The present electrocatalysis literature contains unclear information regarding rhenized surfaces obtained by electrodeposition and metallic rhenium. The rhenized surfaces are principally composed of oxides, and therefore their electrocatalytic behavior is produced by these thin films. This has led to a misconception regarding rhenium electroactivity toward the HER. In the particular case of the new volcano plot calculated by Schmickler [8] using Density Functional Theory, this group calculated that there are three metals with very similar exergonic adsorption energies: nickel, cobalt, and rhenium. Of these metals, only the high electrocatalytic activity of rhenium was reported. This behavior seems to contradict Sabatier’s principle, which has been explained to some degree by Schmickler [8]. However, our results show that metallic rhenium does indeed follow Sabatier’s principle. In addition, other important conclusions of this study are: &

&

The values of the exchange current densities obtained by Joncich [2] in 1965 are in agreement with our results and indicate that chloride and sulfate have comparable affinities for adsorption sites on metallic rhenium. In addition, the exchange current density determined in the potentials of the OPD-H confirms that the electroactivity of rhenium towards the HER follows Sabatier’s principle. Rhenium can belong to the group of metals on which the HER occurs at metal-hydride phases covered by a barrierlayer oxide film [52, 53].

Experimental Tafel slopes for the HER on metallic rhenium at 25 °C in different acidic solution

Solution

The scan rate of the (mV/s)

a

b

Tafel slope mV per decade curvature region

Tafel slope mV per decade linear region

0.037 M HCl 0.037 M HCl

1 0.116

102.0±13.4 244.7±6.5

65.0±6.45 89.8±2.8

0.145 M HCl 0.145 M HCl 0.5 M H2SO4 0.5 M H2SO4 0.5 M H2SO4

1 0.116 10 1 0.116

102.0±13.4 189.5±10.2 215.8±28.3 156.8±17.9 184.5±21.4

57.7±2.19 66.4±3.2 67.7±4.45 62.6±4.49 59.9±3.62

The slopes were determinate from Polarization curve. The solutions were saturated with hydrogen gas and atmospheric pressure was 0.8 Atm a

Interval of potential where the Tafel slope was determined was 0.0 to −0.12 V (dotted line in Fig. 4b). The average slope was 112 mV per decade

b

Interval of potential where the Tafel slope was determined was 0.0 to −0.12 V (short dash-dot line in Fig. 4b). The average slope was 67 mV per decade

Electrocatalysis

& & &

&

In our analysis, the possible formation of rhenide ion or soluble hydride was eliminated because no evidence of these species was observed. The Tafel slope of 67 mV per decade suggests that the proton reaction discharge is the rate-determining step of the HER at the polycrystalline Re in acidic medium. No evidence of the UPD-H on metallic rhenium was found using low scan rates to measure the potential–log current plots. It should be noted that a more extensive characterization of the surface will be performed to conclusively determine the UPD-H on rhenium oxide. In this study, a new issue was raised based on our data and analysis: Why are the oxides of rhenium just as good electrocatalysts as platinum? We postulate that some oxides act as proton exchange films with mobile protons; therefore, the kinetic parameters of rhenized electrodes correspond to the metal substrate.

Acknowledgments This work was performed under the support of the CONACYT (project numbers 102018 and 224366).

References 1. G. Pecherskaya, V.V. Stender, Zhur. Fiz. Khim. 24, 856 (1950). Abstracts from SciFinder® ACS (2013) 2. M.J. Joncich, L.S. Stewart, F.A. Posey, J. Electrochem. Soc. 112, 717 (1965) 3. a) V. L. Krasikov, Elektrokhimiya, 17, 1518 (1981); b) V.L. Krasikov, Elektrokhimiya, 19, 209 (1983). 4. S. Szabó, I. Bakos et al., Proceedings of the 7th International Symposium on Scientific Bases for the Preparation of Heterogeneous Catalysts (Elsevier, Amsterdam, 1998), pp. 269–276 5. J.H. Chun, S.K. Jeon, K.H. Ra, J.Y. Chun, Int. J. Hydrogen Energy 30, 485 (2005) 6. E. Méndez, M.F. Cerdá, A.M. Castro-Luna, C.F. Zinola, C. Kremer, M.E. Martins, J. Colloid Interface Sci. 263, 119 (2003) 7. S. Trasatti, in Advances in Electrochemistry and Electrochemical Engineering, ed. by H. Gerisher, C.W. Tobias (Wiley, New York, 1977), p. 19 8. W. Schmickler, E. Santos, Interfacial Electrochemistry, 2nd edn. (Springer, Berlin, 2010), p. 166 9. A.J. Arvia, A.E. Bolzan, M.A. Pasquale, in Catalysis in Electrochemistry: From Fundamentals to Strategies for Fuel Cell Development, ed. by E. Santos, W. Schmickler (Wiley, USA, 2011), p. 35 10. S. Trasatti, J. Electroanal. Chem. 39, 163 (1972) 11. a) J. K. Nørskov, T. Bligaard, A. Logadottir, J. R. Kitchin, J. G. Chen, S. Pandelov, U. Stimming, J. Electrochem. Soc., 152, J23 (2005), b) K.Nørskov, T. Bligaard, A. Logadottir, J. R. Kitchin, J. G. Chen, S. Pandelov, U. Stimming, J. Electrochem. Soc. 153, L33 (2006) 12. Y.Y. Andreev, Prot. Met. Phys. Chem. Surf. 48, 290 (2012) 13. N. Eliaz, E. Gileadi, in Modern Aspects of Electrochemistry, ed. by C.G. Vayenas, R.E. White, M.E. Gamboa-Aldeco, vol. 42 (Springer, New York, 2008), p. 268. 270–279 14. B.P. Hahn, R.A. May, K.J. Stevenson, Langmuir 23, 10837 (2007) 15. R.J. Magee, T.J. Cardwell, in Encyclopedia of Electrochemistry of the Elements, ed. by A.J. Bard, vol. 2 (Marcel Dekker, New York, 1973), pp. 129–131. 133-136,145-146, 149, 175–177

16. R. Schrebler, P. Cury, M. Orellana, H. Gomez, R. Cordova, E.A. Dalchiele, Electrochim. Acta 46, 4309 (2001) 17. J.O. Zerbino, A.M. Castro-Luna, C.F. Zinola, E. Méndez, M.E. Martins, J. Electroanal. Chem. 521, 168 (2002) 18. R. Schrebler, P. Cury, C. Suárez, E. Muñoz, F. Vera, R. Córdova, H. Gómez, J.R. Ramos-Barrado, D. Leinen, E.A. Dalchiele, Thin Solid Films 483, 50 (2005) 19. M.H. Miles, M.A. Thomason, J. Electrochem. Soc. 123, 1459 (1976) 20. J. Gómez, J.I. Gardiazábal, R. Schrebler, H. Gómez, R. Córdova, J. Electroanal. Chem. 260, 113 (1989) 21. M.H. Miles, J. Electroanal. Chem. 60, 89–96 (1975) 22. B.V. Tilak, A.C. Ramamurthy, B.E. Conway, Proceedings of the Indian Academy of Sciences-Chemical Sciences, vol. 97 (Springer, India, 1986), p. 366 23. G.E.F. Lundell, H.B. Knowles, H.B. Bur, Stand. J. Res. 18, 629 (1937) 24. L. Pauling, Chem. Eng. News 25, 297 (1947) 25. J.B. Bravo, E. Griswold, J. Kleinberg, J. Phys. Chem. 58, 18 (1954) 26. J.W. Cobble, J. Phys. Chem. 61, 727 (1957) 27. A.P. Ginsberg, J.M. Miller, J.R. Cavanaugh, B.P. Dailey, Nature 185, 528 (1960) 28. N.N. Greenwood, A. Earnshaw, Chemistry of the Elements, 2nd edn. (Butterworth-Heinemann, Oxford, 1997), p. 1048. 1050, 1054, 1058 29. J.E. Ellis, Inorg. Chem. 45, 3167 (2006) 30. N. de Zoubov, M. Pourbaix, in Atlas of Electrochemical Equilibria in Aqueous Solutions (NACE, Texas, 1974), pp. 300–306 31. G. Inzelt, in The Encyclopedia of Electrochemistry, ed. by A.J. Bard, M. Stratman, F. Scholz, C.J. Pickett, vol. 7 (Wiley-VCH, Weinheim, 2006), pp. 35–36 32. A.J. Bard, R. Parsons, J. Jordan, Standard Potentials in Aqueous Solution (Marcel Dekker, New York, 1985), pp. 444–451 33. M.S. Antelman, Encyclopedia of Chemical Electrode Potentials (Plenum Press, New York, 1982), p. 176 34. A.J. Dean, Lange’s Handbook of Chemistry, 15th edn. (McGraw-Hill, New York, 1999), p. 6.110 35. N. Takeno, Atlas of Eh-pH Diagrams (National Institute of Advanced Industrial Science and Technology, Japan, 2005), p. 211 36. S.G. Bratsch, J. Phys. Chem. Ref. Data 18, 1 (1989) 37. C.G. Zoski, Handbook of Electrochemistry (Elsevier, Netherlands, 2007), p. 456 38. S. Krause, in Instrumentation and Electroanalytical Chemistry, ed. by A.J. Bard, M. Stratmann, vol. 3 (Wiley-VCH, Berlin, 2003), p. 202 39. P. Wehner, J.C. Hindman, J. Am. Chem. Soc. 75, 2873 (1953) 40. C.L. Rulfs, P.J. Elving, J. Am. Chem. Soc. 73, 3284 (1951) 41. C.L. Rulfs, P.J. Elving, J. Am. Chem. Soc. 73, 3287 (1951) 42. Rhenium, ASM Handbook, vol. 2 (10th (ASM International, Ohio, 1998), pp. 4843–4844 43. H. Kita, J. Res. Inst. Catal. Hokkaido Univ 13, 151 (1965) 44. V.I. Kopylets, Mater Sci 35, 438 (1999) 45. A.J. Appleby, H. Kita, M. Chemla, G. Bronoel, in Encyclopedia of Electrochemistry of the Elements, ed. by A.J. Bard, vol. 9 (M. Dekke, New York, 1982), pp. 456–457 46. M.M. Jaksic, J. New Mater. Electrochem. Syst. 3, 153 (2000) 47. J.O. Zerbino, A.M. Castro-Luna, C.F. Zinola, E. Méndez, M.E. Martins, J. Braz. Chem. Soc. 13, 510 (2002) 48. A.V. Uscategui, E. Mosquera, L. Cifuentes, Mater. Lett. 94, 44 (2013) 49. S. Szabo, I. Bakos, J. Electroanal. Chem. 492, 103 (2000) 50. M. Castriota, E. Cazzanelli, G. Das, R. Kalendarev, A. Kuzmin, S. Marino, G. Mariotto, J. Purans, N. Scaramuzza, Mol. Cryst. Liq. Cryst. Sci. Technol. Sect. 474, 1 (2007) 51. J.S. Lord, W.G. Williams, Solid State Ion. 145, 381 (2001) 52. M. Calatayud, F. Tielens, in Catalysis in Electrochemistry: From Fundamentals to Strategies for Fuel Cell Development, ed. by E. Santos, W. Schmickler (Wiley, New York, 2011), p. 395 53. B.E. Conway, G. Jerkiewicz, Electrochim. Acta 45, 4075 (2000) 54. B.E. Conway, G. Jerkiewicz, Solid State Ion. 150, 93 (2002) 55. G. Jerkiewicz, Electrocatalysis 1, 17 (2010)