Analysis of the electrocatalytic activity of α-molybdenum carbide thin porous electrodes toward the hydrogen evolution reaction

Electrochimica Acta 220 (2016) 363–372

Contents lists available at ScienceDirect

Electrochimica Acta journal homepage: www.elsevier.com/locate/electacta

Analysis of the electrocatalytic activity of a-molybdenum carbide thin porous electrodes toward the hydrogen evolution reaction Ana Ma . Gómez-Marín* , Edson A. Ticianelli Instituto de Química de São Carlos, Universidade de São Paulo, Caixa Postal 780, Dept. Fisico Quimica, Av. Trabalhador São-Carlense 400, CEP 13566-590, SP, Brazil

A R T I C L E I N F O

Article history: Received 6 July 2016 Received in revised form 14 October 2016 Accepted 15 October 2016 Available online 17 October 2016 Keywords: Electrocatalysis carbides sustainable fuels hydrogen evolution reaction thin porous electrodes

A B S T R A C T

In the last years, transition metal carbides have appeared as novel materials with promising catalytic properties toward important practical reactions. In this work, the use of the thin porous electrodes for evaluating the hydrogen evolution reaction (HER) of hexagonal molybdenum carbides (a-Mo2C)-based materials is analyzed and the effect of catalyst’s load, Lcat, and catalyst’s dispersion are discussed in terms of kinetic parameters calculated by employing the electrode geometric area. Catalysts were characterized by X-ray diffraction, energy dispersive X-ray analysis, X-ray photoelectron spectroscopy, cyclic voltammetry and differential electrochemical mass spectrometry. Results have shown that, even for the same catalyst, mass activities and specific activities depend on Lcat. and catalyst’s dispersion. In contrast, intrinsic kinetic parameters, calculated from double layer capacitance normalizations, can be considered rather constant. XPS analysis of samples under different electrochemical treatments reveals a surface enrichment of carbon terminated planes after the HER, suggesting a higher HER activity on these planes. An investigation of the electrochemical oxidation of a-Mo2Cand the catalyst’s HER activity show a direct correlation between active sites for HER and active sites for catalyst oxidation. Therefore, this oxidation is also used to estimate HER intrinsic parameters. Finally, the activity of a composite sample, in which a-Mo2C is the only active component, is also evaluated. ã 2016 Elsevier Ltd. All rights reserved.

1. Introduction The worldwide necessity of a renewable energy source in a sustainable way has prompted the research toward the electrochemical, or photo-electrochemical, production of hydrogen [1,2]. In this process, hydrogen is produced as a final product of the electrolysis of water, which takes place onto a proper catalyst. Unfortunately, after decades of research on new materials, platinum (Pt) continues being the most effective electrocatalyst for this reaction at present time [3,4], and, because Pt is expensive and scarce, other catalytic materials must be developed before this technology can be suitable for commercial applications. In this sense, several works have recently proposed molybdenum-based compounds, such as molybdenum sulfides, MoS2, and molybdenum carbides, Mo2C, as promising candidates to be active catalysts for the hydrogen evolution reaction (HER) in both acid and alkaline media [1,5–16].

* Corresponding author. E-mail addresses: [email protected], [email protected] (A.M. Gómez-Marín). http://dx.doi.org/10.1016/j.electacta.2016.10.101 0013-4686/ã 2016 Elsevier Ltd. All rights reserved.

However, the catalytic HER activity of molybdenum-based compounds is still not enough for substituting Pt-based materials and hence it should be improved. The most popular strategies for this goal involve either the fabrication of different nanostructures that may display a high density of the most active sites [14], or increasing the catalyst’s surface area by dispersion of nanoparticles of the active material into electronic conductors, such as carbonderived compounds [1,8,13,15,17,18]. Though, in this latter case, improvements on the HER catalytic activity have been also explained as a consequence of the better electron transport in carbon supported materials [1,8,13,15,17,18]. These approaches do not alter the electronic properties of the catalysts, but physically increase the number of active sites accessible for catalyzing per geometric area of the electrode and thus, enhance the apparent HER current density. The modification of the electronic and chemical properties of any material can be effectively reached either by the introduction of another element into its lattice, well because of the formation of heteroatom bonds (ligand effect) or owing to the alteration of the average atom-atom bond length (strain effect) [4,19]. For example, it has been reported that transition metal doping of MoS2 materials with Fe, Co and Ni may significantly improve HER activity of these

364

A.M. Gómez-Marín, E.A. Ticianelli / Electrochimica Acta 220 (2016) 363–372

materials, but not with Mn, Cu and Zn metals, specially under neutral conditions [4,10]. For Mo2C catalysts, the effect of metal doping is not that clear. Xiong et al. have found Ni-doped Mo2C more active than non-doped Mo2C for the HER activity in alkaline media [20], although in acid media, recent results suggest that it is less active [11]. In any case, in the search of new and better catalytic materials, determining the intrinsic activity of different materials is a crucial step for establishing a clear path toward the improvement of the catalyst’s performance. Only then, the effect on the catalytic activity of the different factors mentioned above, such as the structure, dispersion, and composition of the catalyst, can be unambiguously examined. However, one of the main drawbacks when comparing the catalytic activity of diverse materials, from different synthesis procedures, is to find accurate descriptors of their real catalytic activity. This is especially true for Mo2C-based compounds, for which no experimental approach has been established to assess their intrinsic catalytic activity toward the HER. In contrast, for MoS2-based compounds, their electrochemical oxidation at high potentials has been proposed for determining the surface concentration of MoS2 atoms, allowing the calculation of the intrinsic catalytic activity of these materials [5,21]. For evaluating the HER activity, thin porous electrodes are commonly employed, in which catalysts are deposited onto a glassy carbon electrode and tested in an ordinary electrochemical cell [1,5,7–9,13–17,22–25], sometimes including significant ohmic drops. Then, from these polarization curves, catalytic activities are usually extracted in terms of Tafel slopes, current densities at a reference potential (Er), jE=Er, and exchange current densities, j0, calculated by employing the electrode geometric area, or mass activities at Er, MAE=Er, for catalysts in which the amount of active material is not even reported. Under this approach, different values for the HER activity of Mo2C-based materials have been reported, under diverse set of experimental parameters [1,5,7–9,13–17,22– 24], but no intrinsic activities have been determined so far, which forbids to set any guidelines for the design of new, more active Mo2C-based materials. In this work, the HER activity of several a-Mo2C-based materials is evaluated by using thin porous electrodes and the experimental methodology is carefully analyzed, in order to determine the effect of catalyst’s dispersion and catalyst’s load on kinetic parameters calculated by employing the electrode geometric area. Additionally, the electrochemical oxidation of a-Mo2C at high potentials is also studied and later employed for estimating intrinsic kinetic parameters for the HER, which are compared to similar parameters computed from double layer capacitance measurements. Finally, intrinsic kinetic parameters for the HER activity of a composite sample, in which a-Mo2C is the only active component, are also assessed. Here, it is important to note that, considering the different possible phases for molybdenum carbides [24,25] and the lack of an unique definition in the literature, a-Mo2C corresponds to the hexagonal structure of Mo2C,

following the notation convention defined by the Joint Committee on Power Diffraction Standards (JCPDS) files and references [9,12,16,25,26]. 2. Experimental 2.1. Preparation of Catalysts All samples were prepared by a carbothermal carburization process of the metal oxide precursor, MoO3 (Sigma Aldrich, 99.5%), by using a temperature programmed reduction (TPR) method [26– 31]. The preparation conditions for all samples, including the starting materials, are summarized in Table 1. Briefly, the process consists of three steps. First, wet impregnation of MoO3 on carbon black (Vulcan XC-72, Cabot Corp., USA) was carried out in an ultrasonic bath to form a well-dispersed slurry, using isopropanol as solvent. Then, the isopropanol was completely evaporated and the solid precursor was dried at 80
C overnight. Second, the catalyst precursor powder, 200–700 mg, was preheated, in an inert atmosphere (Ar), from room temperature (RT) to 550
C, at 20
C min1, and held there for 30 min. Subsequently, the temperature of the sample was continuously raised at 1
C min1 until to the final carburization temperature, Tcarb, and held there for a fixed time, th, Table 1. For the sample named F, the last heating rate was 3
C min1, while the samples named B and G were not pre-heated, and the heating was from RT to Tcarb at a heating rate of 1
C min1 in an 10% H2/Ar atmosphere. TPR experiments were performed using a Micromeritics 2900 AutoChem II Chemisorption Analyzer Micrometrics equipped with a thermal conductivity detector (TCD). Finally, once the carburization process is finished, resulting catalysts were cooled down to RT and passivated for 90 min in a stream of 2 vol% O2/He mixture. 2.2. Catalyst characterization The diffraction patterns of all samples were measured by X-ray diffraction (XRD, RIGAKU model RU200B) in the 2u range from 20 to 80
and using CuKa radiation. The crystallite size (Table 1) was established from XRD data using the Debye-Scherrer equation, Dc = 0.9l/(bcosu ), where l is the wavelength of the X-ray radiation (l = 1.541 Å), b is the width of the peak at half-maximum and u is the Bragg angle. The 2-3 major peaks were used in calculations, which for a-Mo2C phase (JCPDS 35-0787) correspond to the {101}, {002} and {100} planes and for MoO2 phase (JCPDS 32-0671) correspond to the {-111} and {-211} planes. In all cases, the major peaks yielded similar values of Dc [24–29]. Approximate chemical compositions of all samples, were estimated by energy dispersive X-ray spectroscopy (EDX, Isis System Series 300) in a scanning electron microscope LEO, 440 SEM-EDX system (Leica-Zeiss, DSM-960) with a microanalyzer (Link analytical QX 2000) and a Si (Li) detector, using a 20 keV incident electron beam.

Table 1 Composition of precursors, carburization conditions and main Mo-crystalline phases, identified by XRD, for different a-Mo2C samples. Catalysts

wt.% MoO3

wt.% Carbon black

Carburization gas

Tcarb (
C)

Holding time, th (min)

Mo-Crystalline phases

Crystallite size (nm)

A B C D E F G

77.3 77.2 81.3 81.3 81.3 81.3 77.2

22.7 22.8 18.7 18.7 18.7 18.7 22.8

Ar 10% H2-Ar Ar Ar Ar Ar 10% H2-Ar

750 725 800 900 700 700 625

20 30 45 30 720 0 30

a-Mo2C a-Mo2C a-Mo2C a-Mo2C a-Mo2C

35 29 40 42 41 42/35 39

*

Calculated from XRD data.

MoO2 and a-Mo2C MoO2

*

A.M. Gómez-Marín, E.A. Ticianelli / Electrochimica Acta 220 (2016) 363–372

X-ray Photoelectron Spectroscopy (XPS) was performed using a ScientaOmicron ESCA+ spectrometer equipped with a highperformance hemispheric analyzer (Argus) and with a monochromatic Al Ka (hn = 1486.6 eV) radiation as the excitation source. The operating pressure in the ultrahigh vacuum chamber (UHV) during analysis was 2  109 mbar. Energy steps of 50 and 20 eV were used for the survey and high-resolution spectra, respectively. Each peak was simulated using combined profiles of Gaussian–Lorentzian functions (GL30), and the baseline of Mo 3d envelopes was determined by the Shirley method, while a linear baseline was employed for C 1 s and O 1s. For Mo 3d XPS spectra, measurements were analyzed by considering five oxidation states: Mo6+, Mo5+, Mo4+, Mo0 and one at the binding energy (BE) 227.9 to 228.3 eV (3d5/2), assigned to molybdenum carbide, Mod+ (0 < d+ < +0.3) [32,33]. Mo 3d5/2 and Mo 3d3/2 peaks were separated by 3.1 eV with peak ratios equal to the theoretical value of 3:2, and identical Mo 3d5/2 and Mo 3d3/2 FWHM values were used [32,33]. Samples for XPS analysis were prepared in the form of a thin film by pippeting onto a gold layer (1.13 cm2 area, 50 nm thickness), obtained by Au sputtering onto a Gore-Tex1 PTFE membrane (pore size 0.02 mm), an aliquot of the ultrasonically redispersed catalyst ink (4.0 to 5.0 mg mL1). After the evaporation of the alcohol at ambient temperature, the electrodes were washed with ultrapure water and inserted in the electrochemical cell. After the electrochemical treatment, the electrodes were removed from the cell, washed with deaerated doubly distilled water, mounted onto an XPS sample holder and then transferred to the spectrometer. 2.3. Electrochemical measurements Electrochemical measurements were conducted at RT, 22
C, in a two–compartment, three electrodes, all–glass cell, using an Autolab (Nova) equipped with an interchangeable rotating disk electrode setup (Pine Instruments). Suprapure perchloric acid (Merck) was used to prepare aqueous solutions in ultrapure water (Purelab Ultra, Elga–Vivendi). H2 and Ar (N50, Air Liquid) were also employed. Potentials were measured against the Reversible Hydrogen Electrode (RHE) and a large Au plaque was used as a counter electrode. Similar results were also measured when a Pt plaque, instead of the Au plaque, was used as a counter electrode. If

365

not stated otherwise, in-situ IR ohmic-drop corrections were applied by positive feedback, when necessary, for all electrochemical measurements. Glassy carbon disks (5 mm diam., 0.196 cm2), polished to a mirror finish (0.05 mm alumina, Buehler) before each experiment, served as substrate for all samples. Electrodes were prepared by pipetting onto the glassy carbon surface an aliquot of the ultrasonically redispersed catalyst ink. The inks were prepared by suspending in an ultrasound bath a certain amount of the as prepared catalysts in isopropanol (1.0 to 2.0 mg mL1). Catalyst loads of all samples are given below (see Table 3). After evaporation of isopropanol from the glassy carbon surface, to fix the electrocatalyst, on top of the dried catalyst powder an aliquot of a diluted Nafion1 solution (Aldrich) was dropped, and let it to dry in air atmosphere, giving films of  0.15–0.20 mm thickness. Directly after preparation, the electrodes were immersed into deaerated, 0.1 M HClO4 electrolyte. Initially, as a cleaning procedure for the catalyst surface, the electrode potential was cycled several times between 0.2 to 0.4 V. The upper potential limit, Eup, was chosen to be lower than the open circuit potential of the electrode, in order to avoid catalyst oxidation (see below). The activity of catalysts toward the hydrogen evolution reaction (HER) was evaluated in H2-free and H2-saturated solutions, different to the solution in which the cleaning procedure was performed, in quiescent solutions and rotating electrodes. Current densities were normalized by either the geometric area of the glassy carbon substrate or the electrochemically active surface area (ESA) of a-Mo2C catalysts, estimated from either double layer capacitance measurements or the integrated charge of the irreversible oxidation peak attributed to Mo2C species, as explained below. With respect to HER measurements, it should be noted that 1) reported currents correspond to the steady-state currents reached after the cleaning procedure was performed; 2) the glassy carbon RDE and pure Vulcan carbon, at similar loadings to those used in a-Mo2C catalyst experiments, did not show any appreciable hydrogen production activity at potentials positive of 0.3 V, in agreement with Ref. [34]. Thus, the overall activity of Vulcan carbon present in catalysts and the glassy carbon RDE can be considered negligible inside the potential window evaluated in this work.

Table 3 Electrochemical activities toward the HER of several samples of a-Mo2C at different electrode loads* . Catalysts

Lcat/mg cm2

TafelHER/mV

j0

A (carbon-supported)

8.5 12.7 16.9 25.4 50.8 100.8

73.8 79.6 75.5 75.5 77.8 79.0

A

200.0+

B

IR-drop non-corrected C D E F * +

jE=-0.15V/mA cm2

MAE=-0.15V/mA mg1

1.9 3.5 3.0 5.1 9.5 17.4

210 310 340 550 820 1550

24.8 24.4 20.1 21.7 16.1 15.4

85.5

39.0

2280

11.4

12.3 62.8 99.7 733.3

70.0 73.7 74.4 116.0

11.4 23.6 47.2 387.9

1710 2530 5170 9850

138.9 40.3 51.9 13.4

733.3 100.2 100.0 100.4 100.2

151.0 84.8 106.0 81.0 80.1

522.2 15.6 37.4 27.0 26.2

5370 1070 980 2170 1970

7.3 14.4 11.6 26.5 19.7

Parameters were calculated by using the electrode geometric area. Equivalent electrode load of the pure sample to the carbon supported samples.

mA cm2

HER/

366

A.M. Gómez-Marín, E.A. Ticianelli / Electrochimica Acta 220 (2016) 363–372

2.4. Differential electrochemical mass spectrometry (DEMS) setup

log(I / mA)

-2

-1.5

jgeo / mA cm

On-line DEMS measurements were performed with a Pfeiffer Vacuum QMA 200 quadrupole mass spectrometer using a setup consisting of two differentially pumping chambers. More details of this method can be found in Ref. [35] and [36]. This technique allows the on-line detection of volatile and gaseous products of electrochemical reactions during the application of a potential scan. The electrochemical cell was constructed according to previous published principles [35,37]. Experimentally, currentpotential profiles were simultaneously recorded with mass intensity versus potential curves, for selected values of mass/ charge (m/z) ionic signals: H2 evolution was monitored at m/z = 2, while the electrochemical oxidation of a-Mo2C was followed at m/ z = 28 (CO) and m/z = 44 (CO2). The electrode potential was cycled at 0.01 V s1 between either 0.4 to 0.2 V, for HER polarization curves, or 0.05 to 1.2 V, for the electrochemical oxidation of the samples. Working electrodes were prepared in the same form than for XPS analysis but, in this case, the catalyst ink included 20 mL of Nafion1 solution–5 wt.%.

0.0

-3.0 -4.5 -6.0 -7.5 -0.20

0

1.0x10

-1

1.0x10

-2

1.0x10 -0.20 -0.15 -0.10 -0.05 E / V (RHE)

-0.15

-0.10

-0.05

0.00

E vs. RHE / V Fig. 1. Polarization curves at 0.01 V s1 and 1000 rpm of the HER activity in Arsaturated 0.1 M HClO4 of a-Mo2C-A catalysts, at different catalyst’s loads (Table 3). Reduction currents sequentially decrease at increasing Lcat. Inset: corresponding Tafel plots (solid) and extrapolated Tafel straight lines (dashed). Current densities were calculated by employing the electrode geometric area.

3. Results and discussion 3.1. Catalyst Synthesis and Characterization Once prepared, catalysts were characterized by XRD and EDX measurements. Table 1 resumes the main crystalline phases of final products, identified by XRD. For samples A to E, XRD patterns predominantly exhibits peaks of a single-hexagonal phase of a-Mo2C (JCPDS 35-0787), and, for samples A and C, they also reveals trace amounts of MoO2 crystalline phase (JCPDS 32-0671) (Fig. S1). Similarly, Sample F is a mixture, of unknown composition, of a-Mo2C and MoO2, while Sample G is only composed by MoO2. The chemical composition of Samples A to E was estimated by EDX, and collected in Table 2. These measurements were obtained by averaging data from at least four different regions of each material. As seen by the standard deviations, good homogeneities were found for all samples. 3.2. Electrochemical measurements 3.2.1. Hydrogen Evolution on a-Molybdenum carbide Electrochemical HER activities of all as prepared, a-Mo2C samples were systematically evaluated and, for samples A to E, several catalyst’s loads were assessed. Additionally, samples A, C and E were also evaluated as a physical mixture with Vulcan carbon (VC): 40% of catalyst plus 60% of VC (carbon-supported samples), in order to establish the effect of catalyst’s dispersion and conductivity on the HER activity. Fig. 1 displays stable polarization curves (main) and corresponding Tafel plots (inset) towards the HER of sample A (a-Mo2C-A), at different catalyst’s loads, in Ar-saturated 0.1 M HClO4 solutions at1000 rpm. Initially, reduction currents were higher than those depicted in Fig. 1 for all samples but, in subsequent scans, CVs become stable. This is probably a result of Table 2 EDX compositions for different catalysts (wt.%). Catalyst

A B C D E *

Molar fraction.

Determined by EDX Mo

C

C/Mo*

62.3  4.8 57.9  4.8 87.9  0.5 85.4  1.6 83.9  0.2

37.7  4.8 42.1  4.8 12.1  0.5 14.9  1.6 16.1  0.0

4.9  1.0 5.8  1.0 1.1  0.1 1.4  0.2 1.5  0.9

either reduction/dissolution of surface molybdenum oxides, or amorphous oxides, residuals from the carburization process. Similar curves were also obtained in H2-saturated 0.1 M HClO4 solutions (not shown). As it can be clearly seen from Fig. 1, reduction currents for the same catalyst sequentially increase with the catalyst load, Lcat. In order to compare the catalyst’s performance to similar a-Mo2C catalysts reported in the literature, kinetic parameters commonly reported, such as Tafel slopes, mass activities and current densities at a reference potential (here chosen to be 0.15 V), MAE=-0.15V and jE=-0.15V, and exchange current densities, j0, calculated by employing the electrode geometric area, were determined. Results are collected in Table 3. In the case of sample G, no significant HER activity was measured in the potential interval depicted in Fig. 1, at a Lcat of 99.7 mg cm2 (see below: curve 2, inset to Fig. 6), confirming that MoO2 is not an active catalyst for the HER. From Table 3, it can be appreciated that, for a similar Lcat, 100 mg cm2 of catalyst, Tafel slopes, j0s, MAE=-0.15V and jE=-0.15V vary between 74.4 to 106 mV, 15.6 to 47.2 mA cm2, 11.6 to 51.9 mA mg1 and 980 to 5170 mA cm2, respectively; with the lowest Tafel slope and the highest j0, jE=-0.15V and MAE=-0.15V for a-Mo2C-B. Indeed, if these results are analyzed together with tcarb during carbide synthesis, Table 1, it can be clearly seen that while Tafel slopes increase at increasing tcarb, MAE=-0.15V and jE=-0.15V decrease. Therefore, variations among kinetic parameters of these samples can be probably ascribed to changes in the catalyst’s particle size and, therefore with their real active area. Here it is worth considering that kinetic parameters for a-Mo2C-B are better than for a-Mo2C-E because of the shorter time at tcarb for the former, and despite the 25
C difference of tcarb between these samples. Moreover, even for the same catalyst, all kinetic parameters actually evidence a stronger dependence on the catalyst’s load than on the particle size of different samples at the same Lcat, as analyzed previously, Table 3. In this sense, while Tafel slopes, j0s and jE=-0.15Vs for a-Mo2C-A (a-Mo2C-B) increase from 73.8 to 85.5 (70.0 to 116) mV, 1.9 to 39 (11.4 to 387.9) mA cm2 and 210 to 2280 (1710 to 9850) mA cm2; MAE=-0.15Vs decrease from 24.8 to 11.4 (138.9 to 13.43) mA mg1, when going from an Lcat of 8.5 to 200.0 (12.3 to 733.3) mg cm2. Analogously, reported kinetic parameters also comprise a diversity of values. Tafel slopes from 54 to 264 mV, j0s from 0.69 to 1690 mA cm2, jE=-0.15Vs from 100 to 31600 mA cm2 and MAE=-0.15V from  0.1 to 95.2 mA mg1 for Lcats between 102 to

A.M. Gómez-Marín, E.A. Ticianelli / Electrochimica Acta 220 (2016) 363–372

8200 mg cm2, have been calculated for a-Mo2C prepared by different procedures [1,7–9,13,22–24]. In agreement with what was discussed above, the greater value for MAE=-0.15V, 95.2 mA mg1, was estimated for one of the lowest reported Lcat = 210 mg cm2 [24]. Therefore, the 138.9 mA mg1value calculated for a-Mo2C-B sample at Lcat = 12.3 mg cm2 is, to our knowledge, the highest MAE=-0.15V value reported until now. Similarly, the highest values for the Tafel slope and j0, 264 mV and 1690 mA cm2, were reported for Lcat as high as 6300 mg cm2 [13]. Although in this latter case, measured jE=-0.15V is not the highest calculated values, and this is because of the effect of the Tafel slope on determining j0, as discussed below. The decrease in mass activities at increasing Lcat can be understood considering purely geometric diffusion limitations and therefore, an incomplete usage of the catalyst. This effect can be more noticeable at high Lcat, because of the large amount of hydrogen produced, which in turn also decreases the catalyst’s effective area. Regarding the Tafel slope, besides to the increase on its value because of diffusional effects, as it is seen from Tafel slopes of a-Mo2C-B for Lcat of 99 and 733 mg cm2, results in Table 3 evidence the necessity of ohmic drop corrections when evaluating kinetic parameters (also see Fig. 2). This is especially important in determining j0’s values, which are extrapolated from adjusted Tafel straight lines (insets to Figs. 1 and 2 A), as it is visible from the larger j0 value computed from non-corrected than from corrected data. Furthermore, calculated kinetic parameters for carbon-supported and non-supported samples, e.g. a-Mo2C-A, Table 3, may also indicate a slow electron transport in pure catalyst, as evidenced by the higher Tafel slope and smaller equivalent jE=0.15V (see below), of the latter, compared to those for carbon supported samples. Here it is important to note that the as prepared a-Mo2C-A catalyst was evaluated at Lcat = 80 mg cm2 and hence, any difference with the Tafel slope of carbon supported samples, at equivalent Lcat, probably is not a consequence of diffusional effects. A slow electron transport in non-supported a-Mo2C catalysts would also explain the apparent improved HER performance of supported a-Mo2C, found in several studies [1,8,13,15,17,18,23]. However, because in those works kinetic parameters were only calculated by employing the electrode geometric area, a proper comparison cannot be made. A slow

-0.2

-0.1

0.0

0.1

0.2

log(I / mA)

jgeo / mA cm

-5.0

-10.0

2

IH / pA

0.4

1.0x10

1

1.0x10

0

1.0x10

-1

0.3

0.4

-0.2 -0.1 0.0 E vs. RHE / V

-15.0

160

0.3

A

-2

0.0

B

120 80 40 0 -0.2

-0.1

0.0 0.1 0.2 E vs. RHE / V

Fig. 2. (A) Polarization curves, iR ohmic drop-corrected (solid) and non-corrected (dashed) and (B) Mass spectrometric cyclic voltamogram, m/z = 2 H2, at 0.01 V s1 for the HER activity in Ar-saturated, 0.1 M HClO4 of a-Mo2C-B, at Lcat = 733 mg cm2. Inset to (A): Corresponding Tafel plots (thick lines) and extrapolated Tafel straight lines (thin lines). Current densities were calculated by employing the electrode geometric area.

367

electron transport in unsupported catalyst has also been reported for MoS3 based catalysts [17]. At this point, it is important to stress that the above discussion only applies if reduction currents in Fig. 1 can be unambiguously attributed to H2 production on a-Mo2C catalysts. Thus, DEMS measurements were simultaneously performed to monitor the evolution of gas components during the cyclic voltammetry of a-Mo2C-B, Fig. 2. Additionally, DEMS measurements allow unequivocally determining the onset potential for the reaction, which is a common parameter employed to characterize new catalysts, despite of its lack of kinetic meaning. As it is clearly seen, the mass spectroscopic cyclic voltammogram for H2 production, Fig. 2B, exactly follows the polarization curve, Fig. 2A, for potentials below -0.05 V, the reaction onset. Hence, reduction currents in Fig. 1 can be essentially attributed to H2 formation, and no other faradaic processes take place. Reaction onsets of 0.333 V [18], 0.244 V [16], higher than 0.100 V [7] and 0.062 V [1] have been reported for other a-Mo2C catalysts, and thus, to our knowledge, 0.05 V is the lowest value determined so far for this reaction on a-Mo2C and in acidic media. Results described in preceding paragraphs raise serious questions about the current way of evaluating and comparing the electrochemical performance toward the HER of new catalysts, and reflect the necessity of analyzing them in light of true specific parameters, calculated from IR-ohmic drop corrected data, and not through pseudo kinetic parameters computed by using the electrode geometric area. Additionally, it is clear that both, the catalyst’s conductivity and Lcat, should be also carefully adjusted for each catalyst and so, the precise evaluation conditions would depend on the catalyst activity toward the interested reaction. Otherwise, geometric diffusion limitations, incomplete catalyst’s usage or a slow electron transport inside the catalyst may dominate the electrochemical response and, it would not be possible to extract neither real kinetic parameters from polarization curves nor information about the possible rate-determining step (RDS) from the Tafel slope. The above considerations are similar to what was reported several years ago for determining the activity toward the oxygen reduction reaction, ORR, of carbon-supported Pt-based catalysts [38], and would be applicable, in principle, to evaluate any electrochemical reaction. Although, in principle, the origin of the geometric diffusion limitations for the ORR and HER may be different for those catalysts not active as platinum, as the materials in this work. In this case, while diffusion limitations for the ORR arrive for the limited O2 dissolution rate, for the HER they appear because of the filling of the catalyst’s pores by H2 bubbles, which would reduce the active area of catalysts. 3.2.2. a-Molybdenum carbide electro-oxidation and HER activity At potentials higher than 0.5 V, CVs for all a-Mo2C samples reveal a current peak in the first positive-going scan, 0.75 to 0.94 V, indicating their electrochemical oxidation, Fig. 3A. Accordingly, in the subsequent negative sweep, CVs are characterized by the appearance of new reduction signals around 0.2 to 0.4 V, possibly from reduction of molybdenum oxides species formed during the oxidation of the carbide [39–41]. In the second positivegoing excursion to high potentials, peak currents noticeably decrease. The magnitude of maximum currents and peak potentials, depend on the catalyst’s load and the scan rate. However, for a same Lcat, the integrated charge at different scan rates of the oxidation process, from 0.05 to 0.2 V s1, deviates less than 10% between one measurement and another. Similar deviations are also found between CVs taken at the same scan rate. Therefore, it can be considered that, inside this time window, the extension of catalyst oxidation is almost constant and thus, no massive bulk catalyst’s oxidation takes place.

A.M. Gómez-Marín, E.A. Ticianelli / Electrochimica Acta 220 (2016) 363–372

0.2

0.4

jgeo / mA cm

0 -5 -10 -15

45 30 15

0.6

0.8

1.0

1.2

A

-2

0.0

-2

60

jgeo / mA cm

368

-0.2 0.0 0.2 0.4 E vs. RHE / V

B

6

2

ICO / pA

0 9

3

ICO / pA

0 15 10 5 0

C

0.0

Fig. 3B indicates a linear relation with a zero intercept between integrated oxidation charges, calculated from CVs, and the catalyst’s load, and thus, a complete usage of the catalyst can be expected. From the slope of the linear approximation in this figure, and after assuming a proper reaction scheme for the oxidation process, the electrochemical active surface area (ECA), m2 gcat1, of the catalyst could be determined. After the electrochemical oxidation of a-Mo2C has occurred, the HER activity of all samples markedly decreases Fig. 3C, attaining a loss in activity > 90% in most of the samples, and confirming that HER activities in Figs. 1 and 2 are entirely because of a-Mo2C. Indeed, graphs of HER current densities of non-oxidized samples at 0.15 V, jE=-0.15V, 0.18 V, jE=-0.18V, and 0.20 V, jE=-0.20V, against the integrated charge from the electrochemical oxidation indicate a linear relationship with an almost zero intercept, Fig. 3D, confirming the proportional relationship between these two quantities. For a better understanding of the electro-oxidation process of a-Mo2C, XPS analysis and DEMS measurements were also performed for a-Mo2C-B. Fig. 4 shows DEMS data simultaneously recorded during the CV for a-Mo2C-B at Lcat = 733 mg cm2. From this figure, it is seen that the electrochemical oxidation of a-Mo2C involves the simultaneous oxidation of both molybdenum, Mo, and carbon, C, atoms, at potentials higher than 0.55 V. Initially, carbon atoms completely oxidize to CO2, Fig. 4B, but at increasing potentials, > 0.65 V, production of CO also takes place, Fig. 4C. As mentioned, the HER activity of the catalyst decreases after the electrochemical oxidation, inset to Fig. 4A. However, because of the thick thickness of the catalyst’s layer in this experiment, Lcat = 733 mg cm2, the loss in HER activity was only 51%, compared to the non-oxidized sample, and no further activity loss was registered after subsequent excursions to high potentials (dashed curve). This finding confirms the incomplete catalyst’s usage in thick catalyst’s layers and suggests that, after the electrooxidation of surface a-Mo2C, the dissolution of a thin, passivating layer slows down further catalyst’s oxidation. XPS analysis of a-Mo2C-B are depicted in Fig. 5. Four different electrochemical treatments were analyzed: 1) after submersion in 0.1 M HClO4, in which an open circuit potential, 0.4 V, is

0.4 0.6 0.8 E vs. RHE / V

1.0

1.2

Fig. 4. (A) Polarization curves and mass spectrometric cyclic voltamograms, (B) m/ z = 44, CO2 and (C) m/z = 28, CO, for the first (solid) and second (dashed) oxidation cycles of a-Mo2C-B in Ar-saturated 0.1 M HClO4, at Lcat = 733 mg cm2. Inset to (A): Polarization curves for HER activity before (thin) and after the first (solid-thick) and second (dashed-thick) catalyst’s oxidation cycles. Scan rate 0.01 V s1. Current densities were calculated by employing the electrode geometric area.

Intensity / a.u. Intensity / a.u. Intensity / a.u.

Fig. 3. (A) Cyclic voltammograms at 0.1 V s1 for the oxidation of a-Mo2C-A at Lcat 8.5, 50.8 and 100.8 mg cm2. (B) Oxidation charge of a-Mo2C-A as a function of the catalyst’s load. (C) Polarization curves at 0.01 V s1 and 1000 rpm before and after a-Mo2C-A oxidation, Lcat = 50.8 mg cm2. (D) Dependence of jE=-0.15V (squares), jE=-0.18V (circles) and jE=-0.20V (triangles) on the oxidation charge of a-Mo2C-A. All curves were taken in Ar-saturated, 0.1 M HClO4. Current densities were calculated by employing the electrode geometric area.

0.2

A

O 1s

1 2 3 4

540

530

B

520 C 1s

1 2 3 4

300

290

C1

280 Mo 3d

2 3 4

240 230 220 Binding Energy / eV

Fig. 5. XPS spectra for a-Mo2C-B prepared under different electrochemical treatment (see text): (A) O 1s, (B) C 1 s and (C) Mo 3d.

measured. 2) After measuring the HER activity (10 CVs between 0.2 to 0.4 V at 0.01 V s1) and removing the electrode at 0.4 V. 3) After one cycle to high potentials, where the electrochemical oxidation takes place (CV from 0.0 to 1.2 V), and extracting the electrode at 1.2 V. 4) After one cycle to high potentials, but removing the electrode from the electrolyte at 0.4 V. Results from the deconvolution of spectra are summarized in Table 4. Results in Fig. 5 indicates that all XPS spectra for samples after treatment no 1 and 4 almost superimpose, i.e. just submerging the electrode in the electrolyte (no. 1) and after the electrochemical oxidation and removing the electrode at 0.4 V (no. 4). This result suggest that the chemical state of the catalysts surface does not significantly change after oxidation (Table 4) and hence, a dissolution of oxidized Mo species into the solution should have taken place [6,42]. The only noticeably difference between treatments 1 and 4 is the appearance of a shoulder in the O 1 s spectra at high energies, BE > 530 eV, in the sample no 4. This

A.M. Gómez-Marín, E.A. Ticianelli / Electrochimica Acta 220 (2016) 363–372

369

Table 4 Mo 3d5/2, O 1s and C 1s binding energies for the species on surfaces of a-Mo2C-B after several electrochemical treatments. Oxidation State

Mo 3d5/2 Mo6+ Mo5+ Mo4+ Mod+ Mo0 O 1s MoxOy CO strongly bound O, OH, H2O and/or O¼C C 1s Graphitic Carbidic CO C¼O OC¼O a

Binding Energy (BE)/eVa 1

2

3

4

[13.6] 232.3 (0.7) 231.8 (95.7) 230.0 (2.3) 228.1 (1.3) – [47.1] 529.6 (53.6) 530.9 (39.5) 532.1 (6.9) [39.3] 284.4 (17.9) 283.2 (65.1) 286.2 (4.5) 287.8 (11.5) 290.4 (0.9)

[8.8] 232.5 (2.0) 231.6 (42.9) 229.8 (1.6) 228.3 (6.5) 227.4 (6.5) [29.8] 529.5 (57.7) 530.9 (26.0) 531.9 (16.4) [61.5] 284.6 (15.6) 283.3 (70.1) 286.3 (5.1) 287.8 (5.7) 290.6 (3.6)

[11.8] 232.3 (23.3) 231.8 (75.3) 230.0 (0.4) 228.2 (1.0) – [40.3] 529.8 (47.4) 530.9 (46.2) 532.0 (6.4) [47.9] 284.5 (12.6) 283.5 (74.7) 286.5 (4.7) 287.9 (6.2) 290.8 (1.8)

[11.3] 232.3 (4.3) 231.7 (93.3) 230.0 (1.2) 228.1 (1.2) – [49.6] 529.5 (38.0) 530.9 (52.5) 532.1 (9.6) [39.1] 284.7 (16.0) 283.2 (67.7) 286.2 (6.3) 287.7 (9.9) 290.9 (0.1)

The number in brackets represents the approximate, surface atomic composition, while in parentheses it represents the distribution of species of a given element.

region of BE > 530 eV is commonly assigned to C-O species and strongly adsorbed oxygen species, such as O, OH H2O and/or O¼C [43,44]. Therefore, the appearance of a shoulder in this region after the electrode oxidation may indicate that some of the oxidation products: CO and/or CO2, could have remained adsorbed onto the catalysts surface, consistent with the CV results. As expected, the lowest intensity for O 1 s XPS spectra is recorded for the sample after the HER activity measurements (no 2). This sample also presents the lowest and the highest intensities, and atomic compositions (Table 4), for Mo 3d and C 1 s XPS spectra, respectively, which reveals a surface enrichment of carbon terminated planes during the HER. This fact may indicate that HER activity of a-Mo2C based catalysts preferentially takes place on these planes, in agreement with some theoretical works that predict a higher HER activity on carbon rich than on Mo-ended surfaces, possibly due to a lower coordination number than the average value of the plane, [45]. This is because, considering the volcano-type relationship between HER activity and hydrogen adsorption energy, DEH [2–4], calculated DEH for all surface terminations of a-Mo2C locate them on the strong binding side of the volcano curve, the left side [11,45], with the C-ended surfaces closer to the top of the curve owing to a lower DEH than Mo-ended surfaces [11,45]. Finally, the deconvolution of Mo 3d spectra in Table 4 reveal Mo5+ as the predominant surface specie, with an almost negligible contribution of the Mo4+, Mo0 and Mod+ species for a-Mo2C catalysts in acid solutions, samples no 1 and 4. Only the sample no 2 evidences measurable amounts of Mo0 and Mod+ species, which points out toward a reduction of the catalyst’s surface during the HER. In addition, the increase on the Mo6+ surface concentration after the electrochemical oxidation, especially in the sample removed at 1.2 V (no 3), indicates that this oxidation process takes place through the formation of this species. An interesting result from Table 4 is the increase in the C atomic composition in sample no 3, compared to samples no 1 and 4, and it may propose a faster oxidation of Mo than C atoms, as suggested from other electrochemical studies [46]. With this information, and according to the literature [40,41,46,47], it can be considered that the oxidation of molybdenum carbide is given by Mo2 C þ 8H2 O ! 2MoO3 þ CO2 þ 16Hþ þ 16e

ð1Þ

and Mo2 C þ 7H2 O ! 2MoO3 þ CO þ 14Hþ þ 14e

ð2Þ

Nevertheless, the exact nature of this process cannot be conclusively determined from our experimental results. In this sense, MoO3 species are preferred in Eqns. (1) and (2), over MoO2þ 2 or H2 MoO4 [40,41,46,47], because it is thermodynamically stable at E > 0.4 V (pH = 1) [39], and has been detected by XPS as the main product in the transpassive region for molybdenum electrodes [48]. Moreover, MoO3 has also been detected as the final electrooxidation product of MoS2, [5], for which a similar final product would be anticipated. Nonetheless, even if MoO2þ 2 , or H2 MoO4, is the final product in Eqns. (1) and (2), a transference of 16 electrons per oxidized mol of molybdenum carbide is also expected, as for MoO3. Hence, because the number of electrons is the most relevant parameter to determine the amount of a-Mo2C present on the surface, the knowledge of the exact reaction mechanism is not crucial. Additionally, because of the low CO/CO2 ratios measured during electrochemical a-Mo2C oxidation studies [40,41,46,47], it is expected that the electro-oxidation process of a-Mo2C would be dominated by Eqn. (1), with a modest contribution from Eqn. (2). 3.2.3. Calculating the HER specific activity of a-Molybdenum carbide From what was discussed above, it is clear the necessity of real kinetic parameters for evaluating the electrochemical activity of a catalyst toward a specific reaction. In doing so, measured currents have to be normalized by the electrochemically surface area (ESA) of the material. However, estimating the ESA of an electrocatalyst is not easy, and currently there is no such a method for carbide materials. Commonly, the ESA can be approximated either by measuring the double layer (DL) capacitance, Cdl, of the material, since only the electrolyte-wetted surface-area contributes to its formation, or employing an electrochemical surface reaction, as the adsorption/desorption of H+ on Pt-based materials [38,49]. However, in this latter case, it is implicitly assumed that the active sites for the surface reaction and the evaluated faradaic, electrochemical reaction are equivalent or even the same. In our case, Cdl for all a-Mo2C samples, at each catalyst’s load, were determined as the charging current densities at E = 0.15 V divided by the scan rate [49,50]. They were calculated from CVs measured at different scan rates, from 0.01 to 0.2 V s1, in a potential range of 0.05 to 0.4 V, in which no obvious faradaic processes are observed (Fig. S2). Additionally, the linear relationship between HER activities and the integrated oxidation charge from the CV, Fig. 3D, and the catalyst’s deactivation toward the HER upon catalyst’s oxidation, Fig. 3C, reveals a direct correlation

370

A.M. Gómez-Marín, E.A. Ticianelli / Electrochimica Acta 220 (2016) 363–372

between active sites for HER and active sites for catalyst oxidation. So, as a first approximation, ESAs calculated from the oxidation process of a-Mo2C can also be employed to normalize jE=-0.15V and j0 parameters in Table 3. The advantage of using an electrochemical surface process of the active component for normalizing geometric kinetic parameters, as those in Table 3, rests on the possibility of directly evaluating the ESA, after assuming a suitable reaction scheme. In contrast, because the specific DL capacitance, Cspedl, of exclusively a-Mo2C in the electrodes is unknown, only relative surface areas could be directly obtained from Cdl measurements. In this latter case, for estimating effective surface areas, a specific DL capacitance value should be first assumed. Inside this framework, Table 5 compares jE=-0.15V and j0 values for the a-Mo2C samples in Table 3, normalized by the ESA calculated by both methods DL capacitance and the oxidation charge measurements. The ESA from the electrochemical oxidation of a-Mo2C was computed considering the reaction scheme given by Eqn. (1), while from Cdl measurements a specific DL capacitance of 40 mF cm2 for a-Mo2C materials was assumed [51,52]. This value was selected taking into account that a similar value of Cspedl, at least within an order of magnitude, is expected for many metallic and semiconducting materials in the same aqueous electrolyte [53]. Additionally, Cspedl values for all a-Mo2C samples, determined from Cdl data and the ESA obtained from the oxidation charge, are also given in Table 5. As can be seen, both methods of normalizing kinetic parameters provide similar, within the same order of magnitude, values of jE=0.15V and j0, for those catalysts with a considerable amount of carbon, a-Mo2C-A and a-Mo2C-B, regardless their Lcat and particle size, Table 5. Therefore, catalyst’s dispersion on carbon does not appreciably modify the HER specific activity of a-Mo2C-A. The largest deviations in jE=-0.15V and j0, from the average values, are found for a-Mo2C-B at low and high catalyst’s loads. In the former case, this is most probable because of the impossibility of the catalyst completely to coat the whole surface of the electrode and thus, inactive areas appear and the DL capacitance of the glassy carbon electrode contributes to measured currents. At high catalyst’s loads, deviations arise because of an incomplete

catalyst’s usage, as mentioned before. These results confirm the importance of a proper adjustment of the catalyst load on the electrode. Calculated average values for jE=-0.15V and j0 are 55  8 and 0.56  0.09 mA cm2, respectively, when normalized by the oxidation charge, which are too low compared to those ones for 20% Pt nanoparticles on carbon, measured in our laboratory under the same conditions: 23000 and 230 mA cm2, respectively. Indeed, even several orders of magnitude higher values are expected for polycrystalline Pt, j0 2300 mA cm2 [4], considering that estimated j0 from RDE measurements cannot be considered a real kinetic value. This is because under RDE conditions the low, hydrogen diffusion-limited current densities dominates the HER current response [54], and so, measured currents practically are indistinguishable from the Nernstian diffusion overpotential, reaching reaction currents as high as 2500 mA cm2 at only 15 mV of overpotential [54]. Here it is important to comment that the significance of a-Mo2C based materials as promising HER catalysts mainly lies on the greater availability and lower cost of these materials, compared to noble metals. Relative lower values,  –36  5 and 0.36  0.07 mA cm2, respectively, are obtained from capacitance measurements of carbon supported a-Mo2C-A samples. This is because of the difference between the real value for Cspedl and the one used to calculate the ESA in Table 5, evidenced in the higher Cspedl values summarized in this Table 5. This suggests that employing a surface reaction for determining the ESA is more suitable method than capacitance measurements for carbon supported samples, because the first approach is not affected by the capacitance of the carbon support. For catalysts with a low amount of carbon, a-Mo2C-C to a-Mo2C-E, values for jE=-0.15V and j0 from capacitance measurements are closer to jE=-0.15V and j0 of a-Mo2C-A and a-Mo2C-B described above, while values derived from the oxidation charge are larger. This finding can be also explained by considering the difference in the value assumed for Cspedl and the one calculated by employing the ESA from the oxidation charge, Table 5. If a higher value of Cspedl had considered, values in Table 5 would proportionally increase. Discrepancies between calculated Cspedl values for samples with low and high amount of carbon can be due to the

Table 5 Specific electrochemical activities toward the HER of several samples of a-Mo2C at different electrode loads* . Catalyst

Specific capacitance mF cm2

Oxidation charge j0

2 HER/mA cm

Capacitance jE=-0.15V/mA cm2

j0

mA cm2

HER/

jE=-0.15V/mA cm2

A (carbon-supported)

69 57 62 63 57 61

0.49 0.74 0.53 0.53 0.53 0.49

55 66 60 57 46 44

0.28 0.52 0.34 0.34 0.37 032

32 46 39 36 32 29

Average A B

63  5 38 48 38 50

0.56 0.09 0.66 0.73 0.41 0.56

-55  8 39 109 44 61

0.36 0.07 0.69 0.60 0.42 0.45

-36  5 41 91 45 49

Average High loading* C D E F

45  6 42.0 199 199  115 146 136

0.56  0.16 0.68 2.38 6.46  3.30 2.20 0.80

-71  34 17 163 214  73 174 60

0.49  0.10 0.65 0.48 1.13  0.17 0.59 0.24

-62  25 16 33 42  10 48 18

* Parameters were calculated considering the electrochemical active area (ESA) of the a-Mo2Csamples. The value of ESA is calculated according to the following formulas. ESA = Q/n*0.204. Where Q (mC) consumed during oxidation of a-Mo2C has been estimated by integration of the area under the a-Mo2C oxidation peak, Fig. 3A. n is the number of electrons transferred during the oxidation, according to Eqn. (1) n = 16. Qesp (mC) is the charge density of bulk a-Mo2C, taken as 0.204 mC cm2, calculated from the experimental Mo-Mo distance, 3.01 Å [12,45], and considering the formation of a monolayer of oxides on the hcp plane {100} of a-Mo2C [45]. From capacitance measurements, ESA = Cdl/ Cspedl. Where Cspedl (mF cm2) is the specific DL capacitance of a-Mo2C materials, assumed as 40 mF cm2 [51,52].

A.M. Gómez-Marín, E.A. Ticianelli / Electrochimica Acta 220 (2016) 363–372

lower conductivity of the former samples, which may slow down the electrochemical oxidation process and hence, a lower oxidation charge would be measured. Here, it is important to highlight that the use of the electrochemical oxidation for calculating the ESA is only approximation to estimate the catalysts activity. However, other oxidation schemes, alternative to Eqn. (1), may also take place, e.g. one in which the oxidation of the carbon proceeds at a lower rate than that of the metal [46], as suggested by XPS data in Fig. 5. In this case, the number of electrons transferred to the electrode would be lower than 16 and thus, higher ESAs will be calculated. In this case, the value of kinetic parameters in Table 5 would decrease. Nevertheless, the good agreement between kinetic parameters from both capacitance measurements and the electrochemical oxidation for catalysts with a considerable amount of carbon, such as a-Mo2C-A and a-Mo2C-B, strongly supports the methodological procedure followed here, at least for samples in which the electron conductivity does not determine the electron transport inside the material. Another advantage of using an electrochemical surface reaction in determining the ESA of a catalyst is the possibility to evaluate the intrinsic catalyst’s activity in composite samples [50], in which only one of the components is active toward the target reaction. To verify this approach, a mixed sample of a-Mo2C and MoO2 was prepared, Sample-F, Table 1. The electrochemical response and the HER polarization curve (inset) of this sample are shown in Fig. 6. For the sake of comparison, CVs for pure materials, a-Mo2C, a-Mo2C-B, and MoO2, Sample-G, are also given. As can be appreciated from Fig. 6, the CV of Sample-F can be perfectly described as the resultant curve from adding the currentvoltage contributions of each individual component of the sample: a-Mo2C and MoO2. Accordingly, calculated kinetic parameters, jE=-0.15V and j0, 60 and 0.80 mA cm2, respectively, by employing the integrated oxidation charge are similar, inside the experimental error, to the kinetic parameters for carbon supported a-Mo2C-A and a-Mo2C-B in Table 5. Kinetic parameters from capacitance measurements are smaller than these values because, in this case, the approximate Cdl value, from CVs at different scan rates, also includes the contribution to the DL capacitance of the MoO2 oxide present in the composite.

371

4. Conclusions In this paper, hexagonal molybdenum carbides (a-Mo2C)-based materials were synthesized, characterized and evaluated, by using thin porous electrodes, as catalytic materials toward the electrochemical hydrogen evolution reaction (HER). Different experimental parameters, such as catalyst’s load, Lcat, and catalyst’s dispersion, were analyzed and compared in terms of kinetic parameters calculated by employing the geometric electrode area. Results evidence a dependency of mass activities and specific activities on Lcat and catalyst’s dispersion, and therefore a proper adjustment of the catalyst’s conductivity and Lcat, is needed in order to get meaningful and reproducible data. Precise evaluation conditions would depend on the catalyst activity toward the desired reaction. Otherwise, geometric diffusion limitations, incomplete catalyst’s usage or a slow electron transport inside the catalyst may dominate the electrochemical response. In contrast, intrinsic kinetic parameters, such as current density at Er = 0.15 V, jE=-0.15V, and the exchange current density, j0, normalized from double layer capacitance values, can be considered rather constant, regardless catalyst’s dispersion, and equal to 55  8 and 0.56  0.09 mA cm2, respectively. A reaction onset of 0.05 V is measured for the HER, the lowest reaction onset reported up to now for this reaction on a-Mo2C in acidic media. XPS measurements on samples subject to several electrochemical treatments reveal a surface enrichment of carbon-ended surfaces after the HER has been performed. This fact may indicate that the HER activity of a-Mo2C based catalysts preferentially takes place on these C-terminated planes. An investigation onto the electrochemical oxidation of a-Mo2C and the catalyst’s HER activity evidences a catalyst deactivation toward the HER after the catalyst is oxidized, revealing a direct correlation between active sites for HER and active sites for catalyst oxidation. Therefore, this oxidation process can be also used to estimate HER intrinsic parameters. A good agreement between the electrochemical surface areas calculated from capacitance measurements and the electrochemical oxidation of a-Mo2Cis found for those samples in which the electron transport is not a ratedetermining step. Finally, the activity of a composite sample, in which a-Mo2C is the only active component, is estimated and computed intrinsic kinetic parameters are similar to those ones reported for pure a-Mo2C samples. Acknowledgments

Mo2C

16 12

jgeo / mA cm

jgeo / mA cm

-2

20

-2

24 0 -3

1

MoO2

2

References

-6 -9

1

-0.3 0.0 0.3 E vs. RHE / V

8 4

3

2

0 0.0

The authors would like to thank Fundação de Amparo a Pesquisa do Estado de São Paulo (FAPESP  Procs. 2013/16930-7 and 2014/23486-9), Brazil, for financial supports.

0.2

0.4 0.6 0.8 1.0 E vs. RHE / V

1.2

Fig. 6. Cyclic voltammograms for the oxidation of a-Mo2C-F (solid-1), MoO2-G (dashed-2) and a-Mo2C-B (dashed-3) at Lcat 100.2, 99.7 and 18.5 mg cm2, respectively. Inset: HER polarization curves at 0.01 V s1 and 1000 rpm. Sample a-Mo2C-F is a mixture of a-Mo2C and MoO2. Curves were taken in Ar-saturated, 0.1 M HClO4 at 0.05 V s1. Current densities were calculated by employing the electrode geometric area.

[1] D.H. Youn, S. Han, J.Y. Kim, J.Y. Kim, H. Park, S.H. Choi, J.S. Lee, Highly active and stable hydrogen evolution electrocatalysts based on molybdenum compounds on carbon nanotube-graphene hybrid support, ACS Nano. 8 (2014) 5164–5173. [2] Y. Zheng, Y. Jiao, M. Jaroniec, S.Z. Qiao, Advancing the electrochemistry of the hydrogen-evolution reaction through combining experiment and theory, Angew. Chem. Int. Ed. 54 (2015) 52–65. [3] Roger Parsons, The rate of electrolytic hydrogen evolution and the heat of adsorption of hydrogen, Trans. Faraday Soc. 54 (1958) 1053–1063. [4] J.K. Nørskov, T. Bligaard, A. Logadottir, J.R. Kitchin, J.G. Chen, S. Pandelov, U. Stimming, Trends in the exchange current for hydrogen evolution, J. Electrochem. Soc. 152 (2005) J23–J26. [5] J. Bonde, P.G. Moses, T.F. Jaramillo, J.K. Nørskov, I. Chorkendorff, Hydrogen evolution on nano-particulate transition metal sulfides, Faraday Discuss. 140 (2008) 219–231. [6] M.C. Weidman, D.V. Esposito, Y.-C. Hsu, J.G. Chen, Comparison of electrochemical stability of transition metal carbides (WC, W2C Mo2C) over a wide pH range, J. Power Sources. 202 (2012) 11–17. [7] H. Vrubel, X. Hu, Molybdenum boride and carbide catalyze hydrogen evolution in both acidic and basic solutions, Angew. Chem. 124 (2012) 12875–12878.

372

A.M. Gómez-Marín, E.A. Ticianelli / Electrochimica Acta 220 (2016) 363–372

[8] W.-F. Chen, C.-H. Wang, K. Sasaki, N. Marinkovic, W. Xu, J.T. Muckerman, Y. Zhu, R.R. Adzic, Highly active and durable nanostructured molybdenum carbide electrocatalysts for hydrogen production, Energy Environ. Sci. 6 (2013) 943– 951. [9] K. Zhang, Y. Zhao, D. Fu, Y. Chen, Molybdenum carbide nanocrystal embedded N-doped carbon nanotubes as electrocatalysts for hydrogen generation, J. Mater. Chem. A. 3 (2015) 5783–5788. [10] D. Merki, H. Vrubel, L. Rovelli, S. Fierro, X. Hu, Fe Co, and Ni ions promote the catalytic activity of amorphous molybdenum sulfide films for hydrogen evolution, Chem. Sci. 2 (2012) 2515–2525. [11] R. Michalsky, Y.-J. Zhang, A.A. Peterson, Trends in the hydrogen evolution activity of metal carbide catalysts, ACS Catal. 4 (2014) 1274–1278. [12] C. Wan, B.M. Leonard, Iron-doped molybdenum carbide catalyst with high activity and stability for the hydrogen evolution reaction, Chem. Mater. 27 (2015) 4281–4288. [13] B. Sljukic, M. Vujkovic, L. Amaral, D.M.F. Santos, R.P. Rocha, C.A.C. Sequeira, J.L. Figueiredo, Carbon-supported Mo2C electrocatalysts for hydrogen evolution reaction, J. Mater. Chem. A. 3 (2015) 15505–15512. [14] C. Ge, P. Jiang, W. Cui, Z. Pu, Z. Xing, A.M. Asiri, A.Y. Obaid, X. Sun, J. Tian, Shapecontrollable synthesis of Mo2C nanostructures as hydrogen evolution reaction electrocatalysts with high activity, Electrochim. Acta. 134 (2014) 182–186. [15] W.-F. Chen, J.T. Muckerman, E. Fujita, Recent developments in transition metal carbides and nitrides as hydrogen evolution electrocatalysts, Chem. Commun. 49 (2013) 8896–8909. [16] Y.N. Regmi, G.R. Waetzig, K.D. Duffee, S.M. Schmuecker, J.M. Thode, B.M. Leonard, Carbides of group IVA, VA and VIA transition metals as alternative her and ORR catalysts and support materials, J. Mater. Chem. A. 3 (2015) 10085– 10091. [17] H. Vrubel, T. Moehl, M. Grätzel, X. Hu, Revealing and accelerating slow electron transport in amorphous molybdenum sulphide particles for hydrogen evolution reaction, Chem. Commun. 49 (2013) 8985–8987. [18] S. Wirth, F. Harnisch, M. Weinmann, U. Schröder, Comparative study of IVB– VIB transition metal compound electrocatalysts for the hydrogen evolution reaction, Appl. Catal. B: Environ. 126 (2012) 225–230. [19] C. Tsai, K. Chan, J.K. Nørskov, F. Abild-Pedersen, Rational design of MoS2 catalysts: tuning the structure and activity via transition metal doping, Catal. Sci. Technol. 5 (2015) 246–253. [20] K. Xiong, L. Li, L. Zhang, W. Ding, L. Peng, Y. Wang, S. Chen, S. Tan, Z. Wei, Nidoped Mo2C nanowires supported on Ni foam as a binder-free electrode for enhancing the hydrogen evolution performance, J. Mater. Chem. A. 3 (2015) 1863–1867. [21] L.R.L. Ting, Y. Deng, L. Ma, Y.-J. Zhang, A.A. Peterson, B.S. Yeo, Catalytic activities of Sulfur atoms in amorphous molybdenum sulfide for the electrochemical hydrogen evolution reaction, ACS Catal. 6 (2016) 861–867. [22] L. Ma, L.R.L. Ting, V. Molinari, C. Giordano, B.S. Yeo, Efficient hydrogen evolution reaction catalyzed by molybdenum carbide and molybdenum nitride nanocatalysts synthesized via the urea glass route, J. Mater. Chem. A. 3 (2015) 8361–8368. [23] L. Liao, S. Wang, J. Xiao, X. Bian, Y. Zhang, M.D. Scanlon, X. Hu, Y. Tang, B. Liu, H. H. Girault, A nanoporous molybdenum carbide nanowire as an electrocatalyst for hydrogen evolution reaction, Energy Environ. Sci. 7 (2014) 387–392. [24] C. Wan, Y.N. Regmi, B.M. Leonard, Multiple phases of molybdenum carbide as electrocatalysts for the hydrogen evolution reaction, Angew. Chem. Int. Ed. 53 (2014) 6407–6410. [25] A.M. Gómez-Marín, J.L. Bott-Neto, J.B. Souza, T.L. Silva, W. Beck, L.C. Varanda, E. A. Ticianelli, Electrocatalytic Activity of Different Phases of Molybdenum Carbide/Carbon and Platinum-Molybdenum Carbide/Carbon Composites toward the Oxygen Reduction Reaction, ChemElectroChem (2016), doi:http:// dx.doi.org/10.1002/celc.201600376. [26] R. Guil-López, E. Nieto, J.A. Botas, J.L.G. Fierro, On the genesis of molybdenum carbide phases during reduction-carburization reactions, J. Solid State Chem. 190 (2012) 285–295. [27] S. Chaudhury, S.K. Mukerjee, V.N. Vaidya, V. Venugopal, Kinetics and mechanism of carbothermic reduction of MoO3 to Mo2C, J. Alloys Components. 261 (1997) 105–113. [28] Z. Hang-yu, L. Zheng-bang, Y. Hai-sen, L. Lin-gen, Carbothermic reduction of MoO3 for direct alloying process, J. Iron and Steel Research, Int. 20 (2013) 51– 56. [29] J.S. Lee, S.T. Oyama, M. Boudart, Molybdenum carbide catalysts I. Synthesis of unsupported powders, J. Catal. 106 (1998) 125–133.

[30] J.S. Lee, L. Vope, F.H. Ribeiro, M. Boudart, Molybdenum carbide catalysts II. Topotactic synthesis of unsupported powders, J. Catal. 112 (1988) 44–53. [31] A. Hanif, T. Xiao, A.P.E. York, J. Sloan, M.L.H. Green, Study on the structure and formation mechanism of molybdenum carbides, Chem. Mater. 14 (2002) 1009–1015. [32] T.P.S. Clair, S.T. Oyama, D.F. Cox, S. Otani, Y. Ishizawa, R.-L. Lo, K. Fukui, Y. Iwasawa, Surface characterization of a-Mo2C (0001), Surf. Sci. 426 (1999) 187– 198. [33] K. Oshikawa, M. Nagai, S. Omi, Characterization of molybdenum carbides for methane reforming by TPR XRD, and XPS, J. Phys. Chem. B. 105 (2001) 9124– 9131. [34] S.T. Hunt, T.M. Kokumai, D. Zanchet, Y. Román-Leshkov, Alloying tungsten carbide nanoparticles with tantalum: Impact on electrochemical oxidation resistance and hydrogen evolution activity, J. Phys. Chem. C. 119 (2015) 13691– 13699. [35] J.P.I. de Souza, S.L. Queiroz, F.C. Nart, Uso de espectrometria de massas em medidas electroquímicas  A técnica de DEMS, Quim. Nova. 23 (2000) 384– 391. [36] B. Bittins-Cattaneo, Electroanalytical chemistry A series of advances, in: E. Cattaneo, P. Konigshoven, W. Vielstich, A.J. Bard (Eds.), Marcel Dekker, New York, 1991. [37] R. Ianniello, V.M. Schmidt, A simplified DEMS set up for electrocatalytic studies of porous PtRu alloys, Ber. Bunsenges. Phys. Chem. 99 (1995) 83–86. [38] K.J.J. Mayrhofer, D. Strmcnik, B.B. Blizanac, V. Stamenkovic, M. Arenz, N.M. Markovic, Measurement of oxygen reduction activities via the rotating disc electrode method: From Pt model surfaces to carbon-supported high surface area catalysts, Electrochim. Acta. 53 (2008) 3181–3188. [39] V.S. Saji, C.-W. Lee, Molybdenum, molybdenum oxides, and their electrochemistry, ChemSusChem 5 (2012) 1146–1161. [40] L.A. Il’icheva, M.K. Freid, V.A. Suprunov, N.A. Tyurina, Corrosion and electrochemical properties of molybdenum carbide in acid and alkaline solutions, Sov. Powder Metall. Met. Ceram. 17 (1978) 594–598. [41] L.A. Il’icheva, L.N. Frolova, M.K. Freid, V.A. Suprunov, Corrosionelectrochemical behavior of the carbide Mo2C in sulfuric acid solutions, Sov. Powder Metall. Met. Ceram. 14 (1975) 916–920. [42] E.C. Weigert, D.V. Esposito, J.G. Chen, Cyclic voltammetry and X-ray photoelectron spectroscopy studies of electrochemical stability of clean and Pt-modified tungsten and molybdenum carbide (WC and Mo2C) electrocatalysts, J.Power Sources. 193 (2009) 501–506. [43] J.-G. Choi, L.T. Thompson, XPS study of as-prepared and reduced molybdenum oxides, Appl. Surf. Sci. 93 (1996) 143–149. [44] J.A. Schaidle, A.C. Lausche, L.T. Thompson, Effects of sulfur on Mo2C and Pt/ Mo2C catalysts: Water gas shift reaction, J.Catal. 272 (2010) 235–245. [45] J.R. Kitchin, J.K. Nørskov, M.A. Barteau, J.G. Chen, Trends in the chemical properties of early transition metal carbide surfaces: A density functional study, Catal. Today. 105 (2005) 66–73. [46] R.D. Armstrong, A.F. Douglas, The anodic oxidation of the binary compounds of the transition elements in sulphuric acid, J. Appl. Electrochem. 2 (1972) 143– 149. [47] A. Maslennikov, C. Cannes, B. Fourest, N. Boudanova, V. Vivier, P. Moisy, Dissolution and electrochemical properties of molybdenum carbide (Mo2C) in basic solutions, Radiochim. Acta. 95 (2007) 399–408. [48] Y.C. Lu, C.R. Clayton, An XPS study of the passive and transpassive behavior of molybdenum in deaerated 0.1 M HCl, Corr. Sci. 29 (1989) 927–937. [49] S. Trasatti, O.A. Petrii, Real surface area measurements in electrochemistry, J. Electroanal. Chem. 327 (1992) 353–376. [50] A.J. Bard, L.R. Faulkner, Electrochemical methods: Fundamentals and applications, 2nd Ed., John Wiley and Sons, New York, 2001. [51] E. Frackowiak, F. Beguin, Carbon materials for the electrochemical storage of energy in capacitors, Carbon. 39 (2001) 937–950. [52] J.D. Benck, Z. Chen, L.Y. Kuritzky, A.J. Forman, T.F. Jaramillo, Amorphous molybdenum sulfide catalysts for electrochemical hydrogen production: insights into the origin of their catalytic activity, ACS Catal. 2 (2012) 1916– 1923. [53] B.E. Conway, V. Birss, J. Wojtowicz, The role and utilization of pseudocapacitance for energy storage by supercapacitors, J.Power Sources. 66 (1997) 1–14. [54] W. Sheng, H.A. Gasteiger, Y. Shao-Horn, Hydrogen oxidation and evolution reaction kinetics on Platinum: Acid vs alkaline electrolytes, J. Electrochem. Soc. 157 (2010) B1529–B1536.