Metal Carbide and Oxide Supports for Iridium-Based Oxygen Evolution Reaction Electrocatalysts for Polymer-Electrolyte-Membrane Water Electrolysis

Electrochimica Acta 246 (2017) 654–670

Contents lists available at ScienceDirect

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

Metal Carbide and Oxide Supports for Iridium-Based Oxygen Evolution Reaction Electrocatalysts for Polymer-Electrolyte-Membrane Water Electrolysis Fatemeh Karimi, Brant A. Peppley* Department of Chemical Engineering, Queen’s University, 19 Division St., Kingston, Ontario, K7L 3N6, Canada

A R T I C L E I N F O

Article history: Received 22 January 2017 Received in revised form 6 June 2017 Accepted 7 June 2017 Available online 15 June 2017 Keywords: Oxygen Evolution Reaction PEM Water Electrolysis Metal Carbide and Oxide Support Polyol Synthesis Iridium Electrocatalyst

A B S T R A C T

Iridium based materials are one of the most active electrocatalysts used for the oxygen evolution reaction (OER) in polymer electrolyte membrane (PEM) water electrolysers. To increase the utilization, the iridium electrocatalyst is typically dispersed on a high-surface area support material. This results in less iridium being required and consequently reduced catalyst cost. In this work, six metal carbides and oxides were characterized and evaluated as supports for iridium electrocatalyst. The supports studied included: tantalum carbide (TaC), niobium carbide (NbC), titanium carbide (TiC), tungsten carbide (WC), niobium oxide (NbO2), and antimony-doped tin oxide (Sb2O5-SnO2). © 2017 Elsevier Ltd. All rights reserved.

The thermal stability, electrical conductivity, specific surface area, and electrocatalytic activity of these support materials were measured before the iridium electrocatalyst was dispersed on the surface. Afterwards, 20 wt% iridium was dispersed on each of the supports using the polyol method, and the resultant BET surface area, electrical conductivity and OER activity of the synthesized supported catalysts were compared with the support material alone. The two most promising supports, TaC and NbO2, were ballmilled for durations varying between 1 to 7 days to reduce the particle size and increase the surface area. As before, 20 wt% supported iridium electrocatalysts were synthesized. Increase in the surface area of TaC increased the mass-specific OER performance by 75%, but no significant change in the performance was observed for NbO2. For both TaC and NbO2, increase in the surface area decreased the powder electrical conductivity of the supported catalysts. A series of catalysts was then prepared to study iridium loading (2, 5, 10, 20, 30 and 100 wt% Ir). BET surface area, electrical conductivity, and OER activity were measured for each catalyst loading and also for the unsupported catalyst. The mass specific

* Corresponding author. E-mail address: [email protected] (B.A. Peppley). http://dx.doi.org/10.1016/j.electacta.2017.06.048 0013-4686/© 2017 Elsevier Ltd. All rights reserved.

OER activity showed a clear maximum around 5 wt% iridium loading. 1. Introduction The increasing world energy need, the depletion of fossil fuels, and concerns about greenhouse gases emissions demand development of a new energy vector, and hydrogen as an energy carrier has been widely mentioned as a solution to world energy problems [1–3]. Polymer electrolyte membrane water electrolysis (PEMWE), where electricity splits the water to high purity oxygen and hydrogen, is a beneficial method to produce hydrogen from renewable energy sources like wind, solar and hydro. However, the widespread commercialization of PEMWE is limited due to problems associated with the substantial overpotential losses and durability issues of oxygen evolution reaction (OER) electrocatalyst in anode [4–6]. The most active material with the lowest overpotential for OER are iridium and ruthenium based electrocatalysts [7–9]. However, ruthenium has been shown to be unstable [10,11] in the acidic OER environment. Iridium has been shown to be more stable [9] and thus has been used exclusively as OER electrocatalyst [5,12–17]. Due to the high cost of iridium material different approaches are taken to increase the utilization of iridium electrocatalyst. One approach is doping the iridium electrocatalyst with some inexpensive metals, which can improve the performance and

F. Karimi, B.A. Peppley / Electrochimica Acta 246 (2017) 654–670

stability of iridium due to changes in the electronic and structural properties. Some of these doping elements include tantalum [18– 20], tin [4,21,22], tin and tantalum [23], tin and antimony [24,25], molybdenum [26], and manganese [27]. The resulting electrocatalyst usually is a binary or ternary composite of iridium oxide with the other metal oxides in the form of IrxMyNzOa, M and N being the inexpensive metals. This approach is often used to make Dimensionally Stable Anodes (DSA), where a layer of OER electrocatalyst is deposited on a substrate (usually titanium), rather than coating it on a membrane to make Membrane Electrode Assemblies (MEAs). Another approach is to use a high-surface area support for the electrocatalyst [14,28]. Using a support reduces the agglomeration of the electrocatalyst and increases the active surface area. When the supported catalyst is coated on the electrolyte membrane, larger supported particles penetrate less through the layers and stay in contact with the polymer electrolyte membranes [5]. As a result, a lower rate of electrocatalyst loss and a higher efficiency of electrocatalyst utilization is achieved. There are several requirements for a suitable support for OER electrocatalyst: 1) high availability and low cost; 2) high stability in the acidic OER environment; 3) high specific surface area; and 4) good electrical conductivity. Polonsky et al. [28]; however, have reported that the conductivity of a support is not critical and non-conductive materials might also be considered as support for IrO2 catalyst. This is because the conductivity of IrO2 alone is enough for the operation of PEMWE and a sufficient loading of IrO2 provides acceptable electrical conductivity. In addition to the abovementioned requirements for a support it is desirable that the support itself has some electrocatalytic activity, and preferably improves the electrochemical activity and durability of the electrocatalyst through influencing its electronic structure [29,30]. Various materials have been used as support for iridium for the OER in PEMWE such as: TiO2 [5], titanium suboxide (TinO2n1) [13], SnO2 [12,31], Antimony-doped SnO2 [32–36], TiC [37], SiCSi [14], and TaC [28,38]. The major disadvantage of metal oxides supports such as SnO2 and TiO2 is their low electrical conductivity. Mazur et al. [5] reported that due to the low electrical conductivity of TiO2 a high loading of iridium electrocatalyst is required to have a sufficient conductivity for the supported catalyst in PEMWE. Adamaki et al. [39] have shown that Magneli phase TinO2n-1 has significantly higher conductivity compared to TiO2. Siracusano et al. [13] have developed a procedure to synthesize titanium suboxides (TinO2n
1) with Magneli phase, which were successfully used as a support for iridium oxide. In the case of SnO2, doping it with antimony (Sb) is a common approach to increase its electrical conductivity [40,41]. Antimony-doped tin oxide (ATO) has been used in many research studies as a promising support for OER electrocatalysts [32–36,42]. ATO with high surface area has been synthesized by a colloidal method [36] or a hydrothermal method [47], and it has been reported that it can potentially improve the stability of iridium and ruthenium electrocatalyst in a supported catalyst [32,43]. However, doping tin oxide with antimony might also increase the corrosion rate in the acidic environment [44]. Antimony oxide has poor stability itself [45]. Overtime the improved conductivity caused by Sb2O5 could decrease resulting in increasing Ohmic losses in the catalyst layer. It has been reported that doping the ATO with Pt could increase its stability [46]; however, this is not a desirable option as Pt doping would significantly increase the cost of the catalyst. Transition metal carbides have attracted much attention recently as potential support materials [47,48] or as electrocatalysts [49–51]. Many transition metal carbides show good electrical conductivity, high chemical stability, good electrocatalytic activity, and excellent mechanical stability [48,51]. Some transition metal carbides such as

655

TiC [37], SiC-Si [14], and TaC [28,38] have been used as supports for iridium oxide for the OER in PEMWE. Nikiforov et al. [14] and Polonsky et al. [28,38] have used the high temperature Adams fusion method [52] to synthesize supported IrO2 using SiC-Si and TaC as support, respectively. In this method iridium nitrate, obtained from the interaction between an iridium precursor salt (H2IrCl6) and NaNO3, is heated to temperatures greater than 500  C and a IrO2 with a crystalline rutile structure is synthesized. The high temperatures used in the Adams fusion method can be tolerated by a limited number of materials without being thermally oxidized. Polonsky et al. [28] have reported that during the synthesis of IrO2/TaC catalyst by the Adams fusion method, a low conductivity surface film of NaTaO3 was formed due to oxidation of TaC by NaNO3 at high temperatures (around 500  C). Ma et al. [37] have also used a chemical reduction technique to synthesize iridium catalyst with a crystalline structure supported on TiC. It has been repeatedly stated that amorphous IrOx has higher OER activity compared to crystalline IrO2 [15,53,54]. Amorphous IrOx could be synthesized with low temperature synthesis methods such as the polyol method. The polyol method [55] is a fast, simple, and low cost technique that enables synthesis of nanosize materials including IrOx with amorphous structure and high OER activity for PEMWE [35,56,57]. However, the polyol method has not previously used to synthesize an OER electrocatalyst supported on transition metal carbides for PEMWE. In this work a number of metal carbides and metal oxides are examined for OER electrocatalyst support; The polyol method is used to synthesize the supported iridium catalysts; The measured OER performance is related to the properties of supports and the supported catalysts. The result of this work is divided into four sections. In the first section: tantalum carbide (TaC), niobium carbide (NbC), titanium carbide (TiC), tungsten carbide (WC), niobium oxide (NbO2), and antimony-doped tin oxide (Sb2O5-SnO2), in the form as supplied by Sigma-Aldrich were characterized and compared as supports for OER electrocatalysts. For the second section, the polyol method was used to synthesize 20 wt% IrOx catalyst on the supports. The specific surface area, electrical conductivity and OER activity of these supported catalysts were then compared. In the third section, the effect TaC and NbO2 surface area on the performance and OER activity of the supported catalysts was studied. Finally, in the last section, the effect of iridium loading on the conductivity, surface area and OER activity of a TaC supported catalyst was examined. 2. Experimental 2.1. Chemicals All the chemicals were purchased from Sigma-Aldrich1, unless otherwise noted. Tantalum (IV) carbide (TaC, 5 mm, 99.99%), niobium (IV) carbide (NbC, 5 mm, 97%), niobium (IV) oxide (NbO2, 5 mm, 99.99%), titanium (IV) carbide (TiC, <200 nm, 99.99%), tungsten (IV) carbide (WC, 2 mm, 99%), and antimony-doped tin oxide (Sb2O5-SnO2, <50 nm, 99.5%) were used as support for iridium catalyst. To synthesize iridium based electrocatalyst with the polyol method, the iridium (III) chloride hydrate (IrCl3.xH2O) was used as metal precursor; ethylene glycol (anhydrous, 99.8%) was used as solvent and reducing agent; and sodium hydroxide (NaOH, ACS reagent) and sulfuric acid (H2SO4, 95-98%, Fisher Scientific) were used for adjusting pH during the synthesis process. 2.2. Ball milling TaC and NbO2 were ball milled using zirconia ceramic balls (5 mm diameter) in a polyethylene vial containing approximately

656

F. Karimi, B.A. Peppley / Electrochimica Acta 246 (2017) 654–670

1 ml DI water. The ball to powder mass ratio was 15:1 with a rotational speed of 50 rpm. The TaC powder was ball milled under this condition for 1, 4, and 7 days. The NbO2 was ball milled for 4 days. The final powder slurry was dried at 80  C overnight.

surface area was measured two times and the average value is reported.

2.3. Synthesis Procedure

The electrical conductivity of the support material powder and the supported iridium catalyst powder was determined using the device shown in Fig. 1. The design of the conductivity tester was based on various designs reported in the literature [58–61]. It consisted of a 30 mm diameter acrylic cylinder with a centrally bored hole of 3 mm diameter. The acrylic cylinder was placed between two brass pistons. A given quantity of powder was weighed and placed inside the acrylic cylinder between the two brass pistons (as indicated in Fig. 1). A pressure (275 kPa) was applied on the top piston using a pneumatic cell holder. In addition to the applied pressure, the weight of the upper piston also added an additional static pressure of 180 kPa on the powder. The electrical resistivity in the system was measured by a twoprobe method [58] using a potentiostat connected to the upper and lower pistons. Linear sweep voltammetry was performed in the range of 0 and 0.5 V with 1 mV/s scan rate. The electrical resistance (Rt) between the upper and lower brass pistons was obtained from the slope of the voltage vs. current [61]. The thickness of the compressed powder inside the cylinder was measured using a digital caliper. The electrical resistance (Rt) and the thickness (t) for 3 to 4 different quantities of powder (5–150 mg) were obtained and the slope of Rt vs. thickness plot was used in Eq. 1 to calculate the powder conductivity.   t 1 s¼  ðEq:?1Þ Rt A

The synthesis was carried out in 50 ml of ethylene glycol containing 200 mg of NaOH powder. The required mass of IrCl3 was carefully weighed and added to the ethylene glycol solution and the temperature was increased to 160  C. The dispersion was left at this temperature for 2.5 hours under reflux condenser, stirring (500 rpm), and nitrogen bubbling. The temperature was then reduced to 80  C; the support was added to the dispersion; and the pH was adjusted to 1-2 using 1 M H2SO4 solution. The dispersion was left under stirring, reflux condenser, and nitrogen flow overnight. The next day, the catalyst was centrifuged and washed with DI water 4-5 times and dried at 80  C. 2.4. Thermogravimetric Analysis (TGA) The TGA results (reported in Appendix A) were obtained using a Q50 TA instrument. The samples were analyzed in air (PRAXAIR, 3.06% carbon dioxide, 21.1% oxygen balance nitrogen). A temperature ramp of 10  C/min was applied up to 950  C. Mass and differential thermal signal were recorded during the experiments. 2.5. Transmission Electron Microscopy (TEM) The TEM images (reported in Appendix B) were collected using FEI Tecnai Osiris microscope operated at 200 keV. To prepare the sample, 5 mg of catalyst was dispersed in 1 ml of isopropyl alcohol and DI water solution (with 1:1 ratio). Afterwards the dispersions were sonicated in an ultrasonic bath for 20 minutes. Then, 7 ml of the dispersion was deposited on a copper grid, covered by an amorphous carbon film, and was dried at 80  C.

2.8. Conductivity Measurements

Where s is the electrical conductivity (S/cm), Rt is the electrical resistance (V) of the compressed powder, t is the thickness (cm), and A is the cross sectional area (cm2) of the powder under compression. 2.9. Electrochemical Measurements

2.6. X-Ray Photoelectron Spectroscopy (XPS) The XPS spectra (reported in Appendix C) were measured on a Kratos Nova AXIS spectrometer equipped with an Al X-ray source. The samples were mounted onto SEM mounts (with the pins cut off) using double-sided adhesive Cu tape. After removing excess powder, the SEM mounts were attached to the coated aluminum platen using double-sided adhesive Cu tape. The samples were kept under high vacuum (10
9 Torr) overnight inside the preparation chamber before they were transferred into the analysis chamber (ultrahigh vacuum, 10
10 Torr) of the spectrometer. The XPS data were collected using AlKa radiation at 1486.69 eV (150 W, 15 kV), charge neutralizer and a delay-line detector (DLD) consisting of three multi-channel plates. Binding energies are referred to the C1 s peak at 285 eV. High resolution spectra for were recorded in the appropriate regions at a pass energy of 20 eV (number of sweeps 10-20) using a dwell time of 300 ms and energy step sizes of 0.1 eV. The spectra were measured using the Vision 2 software and processed using the CasaXPS software. The data were corrected for energy shifts due to charging of the sample under the influence of the X-rays and the spectra were corrected for background using the Shirley algorithm.

The electrochemical measurements were performed at room temperature in a three-electrode experimental cell shown in Fig. 2. To prepare the working electrode, 5 mg/ml dispersion of the powder material in iso-propyl alcohol (IPA) and DI water (with ratio of 1:2) was prepared. The dispersion was sonicated for 60 minutes and left overnight. The gold RDE electrode was first

2.7. BET Surface Area Measurement The specific surface area of supports and supported catalysts were determined by means of nitrogen adsorption BET using a Quantachrome Autosorb 1C instrument. The samples were degassed for 16 hours at 110  C prior to the measurements. The

Fig. 1. The experimental device used to measure conductivity of moderately compressed powders.

F. Karimi, B.A. Peppley / Electrochimica Acta 246 (2017) 654–670

657

Fig. 2. Experimental 3-electrode cell used for electrochemical measurements.

heated to 80  C. The dispersion was sonicated for 15 minutes immediately prior to coating RDE stub. An 8 ml aliquot of the dispersion was then deposited on the hot surface of the gold electrode and it was rotated manually while the dispersion was drying. Afterwards, 8 ml of a Nafion solution (5 wt% Nafion1, Alfa Aesar, diluted in water with the ratio of 1:5) was coated on top of the dried deposited catalyst. The electrode was dried at 80  C. The electrolyte solution (0.5 M H2SO4 solution) was bubbled with argon for 30 minutes, prior to the measurements, to remove dissolved gases. Cyclic voltammetry was performed between 0.2 and 1.25 V vs. Ag/AgCl with 100 mV/s scan rate for 10 cycles. The linear sweep voltammetry (LSV) was performed between 0.8 and 1.5 V vs. Ag/AgCl with 0.167 mV/s scan rate. The working electrode was rotated at 1000 rpm during the LSV to clear the surface from oxygen bubbles. The measurements were repeated 3-5 times, and the average results are reported with 95% confidence interval. All potentials reported in the result section are vs. Ag/AgCl reference electrode. The voltage of the reference electrode was 0.2 V vs. standard hydrogen electrode (SHE). 3. Results and Discussions 3.1. Characterization of Supports The particle size of WC, TaC, NbO2, NbC, TiC and Sb2O5-SnO2 (provided by Sigma-Aldrich) and the measured BET surface areas are shown in Table 1. The thermal stability of the support material (TGA results) is reported in Appendix A. The electrical resistance of 15 mg ATO powder was measured at various pressures and the results are depicted in Fig. 3. It is clear that the measured resistivity of the powder is very dependent on the applied pressure and decreases as the pressure is increased. It would be expected that as the pressure is increased, the void volume would decrease and the contact Table 1 Particle size and measured BET surface area of supports.

Tungsten Carbide (WC) Tantalum Carbide (TaC) Niobium Oxide (NbO2) Niobium Carbide (NbC) Titanium Carbide (TiC) Antimony Tin Oxide (Sb2O5-SnO2)

Particle Size

BET Surface Area (m2/g)

2 mm 5 mm 3 mm 5 mm 200 nm 30 nm

1.7  0.1 0.8  0.1 1.0  0.1 0.9  0.3 28.3  2.5 34.4  5.1

Fig. 3. Resistivity of 15 mg of ATO powder vs. applied pressure. The applied pressure does not include the pressure due to the weight of upper piston (180 kPa).

surface area between the grains would increase. This enhanced contact between particles throughout the powder would result in lower resistivity. Results reported by Espinola et al. [61] for carbonaceous powders show the same effect. Based on the change of resistivity with pressure observed with ATO, a pressure of 275 kPa was chosen for resistance measurement as a good compromise between achieving good contact without significant crushing of the particles. The measured conductivities of the supports are represented in Fig. 4. It should be noted that the conductivity of a powder measured using this method depends on various parameters such as: the applied pressure, the average particles size, the shape of particles, and the surface area of the particles [59,60,62]. Thus, one cannot say the conductivity of TiC was significantly lower than the conductivity of NbC, because the samples of TiC and NbC had very different particle sizes and surface areas. However, it can be said that TiC with 200 nm average particle size has significantly lower conductivity compared to NbC with 5 mm average particle size under 275 kPa compression. WC, NbC, TaC and NbO2 had similar particle size and surface area; therefore, it is reasonable to compare the measured conductivity for these samples. Among these four materials NbO2 showed the lowest electrical conductivity while NbC had the highest electrical conductivity. The transition metal carbides are known to have metal like properties with respect to electric properties [63,64]; therefore, they would

658

F. Karimi, B.A. Peppley / Electrochimica Acta 246 (2017) 654–670

The trend for the conductivity of these four material was: NbC > TaC > WC >> NbO2. The first, second, and tenth (last) cycle of the cyclic voltammograms (CVs) at 100 mV/s obtained for all supports are shown in Fig. 5. The currents in these voltammograms are normalized using the mass of supports. An anodic peak was observed in the first cycle of WC at around 0.75 V vs. Ag/AgCl, and the absence of a cathodic peak indicates irreversibility of the anodic reaction. The anodic peak did not appear in the second cycle or the other consecutive cycles after. This anodic peak corresponds to passivation of WC where an oxide layer forms on the surface according to following reaction [66]. WC + 5H2O ! WO3 + CO2 + 10H+ + 10e


Fig. 4. The support electrical conductivities under 275 kPa pressure.

be expected to have high electrical conductivity comparable to the parent transition metals. However, some transition metal oxides such as NbO2 are known to have semiconducting properties [65].

(Reaction 1)

The gaseous CO2 leaves the system, and thus the WC cannot be regenerated through reduction of electrode and this makes the reaction irreversible. The passivation layer, WO3, is stable in the acidic environment and prevents further oxidation of the bulk material [67]. A similar anodic peak was observed in the first potential cycle of TiC at around 1 V vs. Ag/AgCl. This oxidation peak similarly corresponds to the surface reaction of TiC and formation of a TiO2

Fig. 5. First, second and last cycle of voltammograms at 100 mV/s for a) Tantalum Carbide (TaC), b) Niobium Carbide (NbC), c) Niobium Oxide (NbO2), d) Antimony Tin Oxide (Sb2O5- SnO2), e) Tungsten Carbide (WC), and f) Titanium Carbide (TiC), currents are normalized using the mass of support materials, voltages are vs. Ag/AgCl electrode.

F. Karimi, B.A. Peppley / Electrochimica Acta 246 (2017) 654–670

layer on the surface according to the following reaction [68,69]: þ
TiC þ 5H2 O ! TiO2 þ CO2
3 þ 10H þ 8e

ðReaction2Þ

The passivation behavior was not observed in the CVs of TaC, NbC, NbO2 and Sb2O5-SnO2 in the range of voltage tested (0.21.25 V vs. Ag/AgCl). However, the shape of CVs for these materials in the oxidation region (more specifically within 0.75-1.25 V) was slightly changing, as the number of cycles was increasing. This behavior was less noticeable in the CVs of TaC. The performance of these materials toward OER was compared using OER Tafel slopes obtained from the LSV shown in Fig. 6a. The Tafel slopes were obtained from the linear region of the log (current) vs. voltage graph, and the results are presented in Fig. 6b. TiC had the highest Tafel slope (312 mV), while TaC and NbC had the lowest Tafel slopes (125 mV). Therefore, one can say TiC showed the lowest activity, and NbC and TaC showed the highest activity for the OER among the material tested. The trend for the Tafel slopes was: TiC > ATO > WC  NbO2 > NbC  TaC. 3.2. Characterization of 20 wt% Iridium on Various Support Materials The BET surface area of 20 wt% supported iridium catalysts are presented in the second column of Table 2. It is clear that the surface area was significantly increased when iridium electrocatalyst was dispersed on the surface of the supports (comparing Table 1 with Table 2). It has been previously reported that the BET surface area of iridium oxide catalyst supported on Si-SiC follows the rule of mixtures [14]. Assuming that the BET surface area of the supported catalysts synthesized in this work also follows the rule of mixtures, we can use the following equation:

659

m2

Where, BET Ir/Sup (g ) is the surface area of supported iridium Ir=support 2 catalyst; BET Sup (g m ) is the surface area of support; Xw Sup (wt support %) is the weight percentage of support in the supported catalyst; 2 BET Ir (m gIr ) is the surface area of iridium electrocatalyst; and Xw Ir (wt%) is the weight percentage of iridium in supported catalyst. Using Eq. 2, the surface area specific to the iridium-species electrocatalyst (BET Ir) separate from the support is calculated. The iridium species synthesized using polyol method exist mostly in the form of iridium suboxide (IrOx, x < 4) [57]. An XPS spectrum for Ir 4f, in the iridium species synthesized by polyol method, is shown in Appendix C (Fig. C1). Therefore, one should take this into account that the actual weight ratio of iridium species in the supported catalysts is more than the elemental loading of iridium (20 wt%) which is used for the BET surface area calculations in Eq. 2. The results obtained for BET Ir are shown in Table 2. These results provide a general indication of the degree of agglomeration and the available surface area of iridium-based electrocatalyst in the catalyst on support structure. The surface area of iridium-based electrocatalyst in Ir/ATO and Ir/TiC based on Equation 2 was 148 m2/g and 62 m2/g; respectively. These values are considerably higher than areas attributed to the iridium-based catalyst on the other supports that was on the order Table 3 Conductivity of supports before and after iridium deposition. Conductivity (S/cm)

NbC TaC WC TiC Sb2O5-SnO2 NbO2

BET Ir/Sup = (BET Sup)  (Xw Sup) + (BET Ir)  (Xw Ir) (Eq. 2)

Support

Ir/Support

173  17 68  5 25  2 2.2  0.2 0.01  0.001 0.0001  0.000

9.7  0.8 10  2 5.0  0.1 3.7  0.3 1.1  0.1 0.3  0.1

Fig. 6. a) Polarization curves in logarithmic scale (Tafel curves), and b) Tafel slopes for various supports.

Table 2 The BET surface area of 20 wt% supported catalyst, and the BET surface area of iridium electrocatalyst independently from the support in supported catalysts. 20 wt% supported Iridium (BET Ir/Sup) Ir/WC Ir/TaC Ir/NbO2 Ir/NbC Ir/TiC Ir/ATO

5.7  0.8 4.8  0.5 5.1  0.6 5.2  0.6 35  0.0 57  2

m2 gIr=support

Iridium in the supported catalyst (BET Ir) 22  0.8 21  0.6 21  0.6 22  0.7 62  0.1 148  0.1

m2 gIr

660

F. Karimi, B.A. Peppley / Electrochimica Acta 246 (2017) 654–670

of 21 m2/g. The higher surface area of iridium-based electrocatalyst in Ir/TiC and Ir/ATO suggests that there is less agglomeration and better dispersion and is most likely due to the higher surface area available on these two supports. A comparison of the conductivity of the support materials before and after deposition of the iridium is given in Table 3. The conductivity of supported catalysts, had a similar trend to that of the supports (Fig. 4). The Ir/NbC and Ir/TaC catalysts showed the highest conductivity; while, the Ir/NbO2 showed the lowest conductivity. It should be noted that these conductivity measurements are performed in dry condition and not in OER condition. In the actual OER condition, some of these materials might experience certain transformations which might affect their conductivity. The conductivity of support materials before iridium deposition was in the range of 0–173 S/cm, while after iridium deposition this range reduced to 0.2–10 S/cm. The materials with a high conductivity (NbC, TaC and WC) showed significantly lower conductivity after iridium was dispersed on their surface (17 times smaller in case of NbC) while the materials with lower conductivity (TiC, ATO, NbO2) showed higher conductivity after iridium deposition (110 times higher in case of ATO). This means that support materials with low conductivity could potentially be used as a catalyst support for iridium. This agrees with previously published papers [5,28], where it has been reported that it is possible to overcome the lack of conductivity of a support by ensuring a sufficient loading of electrocatalyst on the support; this forms a continuous electronically conductive phase on the surface of the non-conductive support. The reduction in the conductivity of highly conductive support materials (NbC, TaC, WC), after depositing iridium on their surface using polyol method, was more likely due to transformation of the metal carbides when exposed to the ethylene glycol during polyol

Fig. 7. a) CVs at 100 mV/s for the 20 wt% supported iridium catalysts b) voltammograms oxidation charges in the range of 0.5–1 V vs. Ag/AgCl.

Fig. 8. a) Polarization curves for various supported iridium catalysts, b) currents at 1.48 V vs. Ag/AgCl, c) Tafel curves, and d) Tafel slopes for various supported catalysts.

F. Karimi, B.A. Peppley / Electrochimica Acta 246 (2017) 654–670

synthesis method. To investigate this further, the synthesis procedure was repeated using TaC exclusively (without adding the iridium precursor). The TaC after polyol procedure was collected and tested for conductivity measurement. The change in chemical state of tantalum after polyol synthesis is studied using XPS analysis and the results are shown in Appendix C. It was found that the TaC conductivity was reduced from 68  5 S/cm to 6  2 S/cm after polyol procedure. The conductivity of TaC material after polyol method was higher than the conductivity of the oxide supports (tin and niobium oxides) and lower than the conductivity of Ir/TaC after polyol synthesis (10  2 S/cm). CVs of the supported iridium catalysts are shown in Fig. 7a. The CV shown in the figure is the final cycle in a series and represents the steady state condition. The current is normalized using the mass of supported catalysts used for the measurements. The voltammograms of the various supported catalysts revealed a similar shape and behavior. A reduction/oxidation (redox) peak was observed at about 0.75 V vs. Ag/AgCl, which was attributed to the change in the oxidation state of iridium from +3 to +4 [16,70]. The redox peak was less predominant in the voltammogram of the Ir/WC catalyst. The charge associated with this redox reaction is directly proportional to the electrochemically active surface area (ECSA) of iridium species. The oxidation charge related to this peak in the voltage range of 0.5–1 V vs. Ag/AgCl was obtained for all supported catalysts and the results are shown in Fig. 7b. The trend obtained for the oxidation charge was: Ir/ATO  Ir/TiC > Ir/NbC  Ir/TaC  Ir/ NbO2 > Ir/WC. This result implies that the ECSA of iridium species in Ir/ATO and Ir/TiC catalysts was 2-3 times more than the ECSA of iridium species in other supported catalysts. Comparable findings were seen in the calculated surface area of iridium electrocatalyst (Table 2, third column), where the surface area of iridium species in Ir/TiC and Ir/ATO had the highest value. The polarization curves obtained for supported iridium catalysts are shown in Fig. 8a. The Ir/ATO catalyst showed the best performance, while Ir/TiC showed the lowest performance. The current at 1.48 V vs. Ag/AgCl for the various supported catalysts is shown in Fig. 8b. The order of the current from highest to lowest was: Ir/ATO > Ir/NbO2 > Ir/TaC > Ir/WC > Ir/NbC > Ir/TiC. The morphological and electrochemical characteristics of the best performing catalyst, Ir/ATO catalyst, can be seen in our previous published papers [56,57]. The Tafel slopes were also estimated from the log (current) vs. voltage graph (Fig. 8c) and the results are shown in Fig. 8d. The trend for the Tafel slope was also similar to the trend for the generated current. The Ir/TiC catalyst showed a Tafel slope of 75 mV, which was the largest Tafel slope measured for the supported catalysts.

Fig. 9. BET surface area of TaC and NbO2 versus certain days of ball milling.

661

It was unexpected to see that Ir/TiC, which had the highest value for oxidation charge (ECSA), showed a very poor OER activity and performance. The reason for the low performance of Ir/TiC, despite having a high ECSA for iridium, is probably poor activity of the TiC toward OER. TiC had a Tafel slope of 312 mV, which was the highest Tafel slope among the materials tested (Fig. 6b). Additionally, TiC could have negatively affected the activity of iridium species toward OER reaction; that is why the performance was very poor despite having 20 wt% iridium in this catalyst. The transmission electron microscopy (TEM) images of the three best performing catalysts (Ir/ ATO, Ir/NbO2, and Ir/TaC) are shown in Appendix B. 3.3. Effect of support surface area (TaC and NbO2) on OER performance TaC and NbO2 were the two material with the lowest surface area but that showed good performance as supports for the iridium based catalyst. Fig. 9 shows the increase in the BET surface area that was achieved by ball milling TaC and NbO2 for a period of up to 7 days. After 4 days of ball milling, the BET surface area of TaC increase 4 fold and that of NbO2 increased roughly 11 fold compared to the initial values. The lower increase in surface area for TaC is attributable to its higher microhardness of 1600– 2000 kg/mm2 [71] compared to the hardness of NbO2 of around 600 kg/mm2 [72]. The conductivities vs. surface area for the ball milled TaC and NbO2 are shown in Fig. 10a and b, respectively. It is seen that electrical conductivities were decreased as the surface area was increased. This is due to the increase in the contact resistances between the smaller particles in high surface area material, which caused a decrease in the conductivity. Similar findings were reported by Joshi et al. [73] for polycrystalline silicon (Si). A number of 20 wt% supported iridium catalysts were synthesized using TaC and NbO2 with increased surface area compared to the initial materials. The BET surface area of the synthesized supported catalysts vs. the BET surface area of the supports are shown in Fig. 11a. The surface area of iridium electrocatalyst in the

Fig. 10. a) The conductivity of the TaC powder versus the surface area of TaC, b) The conductivity of NbO2 powder vs. the surface area of NbO2.

662

F. Karimi, B.A. Peppley / Electrochimica Acta 246 (2017) 654–670

supported catalyst was calculated using Eq. 2, and the results are shown in Fig. 11b. It is evident that increasing the surface area of the supports improved the effective surface area of iridium electrocatalyst. The increased surface areas of the supports appeared to improve the dispersion and available surface area for the supported iridium electrocatalyst. The conductivities of Ir/TaC and Ir/NbO2 versus support surface area are shown in Fig. 12a and b, respectively. The reason for the reduction in conductivity with increasing the support surface area was most likely due to reduction in the conductivity of support. However, less degree of iridium agglomeration, existed in the catalyst with a higher support surface area, could have been another reason for the lower conductivity. Better dispersion of iridium resulted in a less electrically connected structure with a lower electrical conductivity. CVs at 100 mV/s are shown in Fig. 13a and b. The current was normalized using the mass of supported catalyst used for the measurements. The oxidation charge corresponding to the redox reaction of iridium, obtained from CVs in the voltage range of 0.5– 1 V vs. Ag/AgCl, is plotted in Fig. 13c. The oxidation charge, which is proportional to the ECSA of iridium electrocatalyst, increased with increased support surface area. This can be attributed to better iridium dispersion and consequently increased ECSA. A similar trend was also observed in the BET surface area of iridium electrocatalyst (Fig. 11b). The polarization curves for these catalysts are shown in Fig. 14a. It is evident that the performance of Ir/TaC was increased with increased support surface area. The improvement in the performance of Ir/TaC is evident in Fig. 14b, the graph of current (at 1.48 V) vs. support surface area. However, there was no significant improvement in the performance of Ir/NbO2 when support surface area was increased.

Fig. 12. a) Conductivity of 20 wt% Ir/TaC catalyst vs. the surface area of the support (TaC), b) Conductivity of 20 wt% Ir/NbO2 catalyst vs. the surface area of support (NbO2).

3.4. The Effect of Iridium Loading on the Performance of Ir/TaC

Fig. 11. a) The BET Surface area of 20 wt% Ir/TaC and 20 wt% Ir/NbO2 vs. the support surface area, b) The BET surface area of iridium electrocatalyst in the supported catalysts (calculated form Eq. 2) vs. the support surface area (TaC and NbO2).

The pristine TaC (0.85 m2/g) was used to synthesize 2, 5, 10, 20 and 30 wt% Ir/TaC catalysts, and their BET surface area are shown in Fig. 15a. It is obvious that the BET surface area of the supported catalyst was improved when iridium loading was increased. This was due to the existence of larger quantity of nano-size iridium electrocatalyst in Ir/TaC with higher iridium loading. The surface area of iridium electrocatalyst in Ir/TaC (calculated from Eq. 2) are also shown in Fig. 15b. The surface area of iridium electrocatalyst was decreased as iridium loading was increased. This is because when the loading of iridium was increased the possibility of iridium electrocatalyst to deposit on the surface of the support in multiple layers also was increased. This consequently resulted in a higher rate of agglomeration and lower surface area of iridium electrocatalyst. The conductivity of the Ir/TaC catalysts, with various iridium loading, is presented in Table 4. The conductivity of TaC support (without any iridium loading) and the conductivity of an unsupported iridium electrocatalyst (without any TaC) are also presented in Table 4. TaC and un-supported iridium electrocatalyst both had good electrical conductivities (small electrical resistances); however, the conductivity of un-supported iridium electrocatalyst (2–5 nm particle size) was 25% larger than the conductivity of TaC support (2 mm grain size). Nevertheless, it is seen that despite the fact that both TaC and Ir electrocatalyst had high electrical conductivity; when a small amount of iridium (e.g 2 wt% Ir) was deposited on the surface of TaC, the conductivity of the supported catalyst (Ir/TaC) decreased significantly (roughly 12 times smaller). The results show that when iridium loading was increased the conductivity was improved gradually.

F. Karimi, B.A. Peppley / Electrochimica Acta 246 (2017) 654–670

663

Fig. 14. a) Polarization curves obtained for Ir/TaC and Ir/NbO2 catalysts with various support surface area, b) The current (mA/mg) at 1.48 V for Ir/TaC and Ir/NbO2 vs. surface area of the support used (TaC and NbO2).

Fig. 13. a) CVs at 100 mV/s for iridium supported at TaC with various surface area. b) CVs at 100 mV/s for iridium supported at NbO2 with various surface area. c) oxidation charge in the range of 0.5–1 V vs. the surface area of supports used in the supported catalysts.

Polarization curves, normalized by the mass of iridium, are shown in Fig. 16. The un-supported iridium electrocatalyst showed the lowest performance; as without any support material, iridium electrocatalyst had a highest rate of agglomeration and consequently smallest ECSA. The 30 wt% Ir/TaC, had very similar performance compared to un-supported iridium electrocatalyst. This is because 30 wt% was a high value of iridium loading for TaC support with a small surface area (0.85 cm2/g); thus, iridium agglomeration was not any smaller than the un-supported iridium electrocatalyst. However, further decrease in iridium loading had improved the catalyst performance which was due to better dispersion of iridium on the support and its higher surface area. The best performance was seen for 5 wt% iridium, after which the performance was decreased with increasing iridium loading to 2 wt%. The performance of 2 wt% Ir/TaC catalyst was acceptable in lower voltage range, but it rapidly dropped in higher voltages. The 5 wt% Ir/TaC catalyst also reached a plateau at the higher end of the polarization curve (around 1.47 V vs. Ag/AgCl). Siracusano et al. [74] also have shown that a triple current density

Fig. 15. a) The BET surface area of supported iridium catalyst (Ir/TaC) vs. the loading of iridium in the supported catalyst, b) the BET surface area of iridium electrocatalyst in the supported catalyst (calculated from Eq. 2).

664

F. Karimi, B.A. Peppley / Electrochimica Acta 246 (2017) 654–670

Table 4 The electrical conductivity measured for TaC, Ir/TaC (catalyst with various loadings of iridium), and unsupported iridium electrocatalyst. Ir Loading (wt%)

Conductivity (S/cm)

TaC 2 wt% Ir/TaC 5 wt% Ir/TaC 10 wt% Ir/TaC 20 wt% Ir/TaC Un-supported Iridium

68  4 5.7  0.8 6.1  0.7 72 10  2 86  11

Fig. 16. Polarization curves obtained for Ir/TaC catalyst with various loading of iridium.

can be achieved using a 4-fold reduction in the precious metal loading (from 2 to 0.5 mg/cm2 MEA), while maintaining a very high conversion efficiency (>80%). They also achieved a proper durability at 3 A/cm2 using low loadings of precious metal (0.5 mg/cm2) for an MEA in PEM electrolyser. The CVs at 100 mV/s are shown in Fig. 17a for the Ir/TaC with various loadings of iridium. The average anodic charges (in the range of 0.45–1 V vs. Ag/AgCl) obtained from 3-5 CVs are shown in Fig. 17b. It is seen that the area under the voltammograms and thus the oxidation charge was increased as iridium loading was decreased. This is another indication of the increase in the ECSA of iridium electrocatalyst with reduction in loading, which was due to lower degree of iridium agglomeration. 4. Conclusions TiC, TaC, NbC, WC, NbO2 and ATO were used as support for iridium electrocatalyst for the oxygen evolution reaction (OER). The surface area, conductivity and activity of these supports were measured, 20 wt% iridium electrocatalysts supported on these material were synthesized using a polyol method and the performance of the synthesized supported catalysts was compared. Table 5 shows a summary of the key results, compared in the following order: Very good (UU), good (U), average (O), bad (), very bad (). The Ir/ATO showed the best performance compared to other supported catalysts, as the generated current at 1.48 V vs. Ag/AgCl was the highest (845 mA/mg) for the Ir/ATO. ATO had a relatively large surface area (35 m2/g), and an average activity (Tafel slope of 175 mV), but a low conductivity (0.01 S/cm). Ir/NbO2 showed the second best performance for OER (558 mA/ mg at 1.48 V). NbO2, had a suitable activity for OER reaction (Tafel slope of 155 mV), but it had a small surface area (1 m2/g) and an extremely low conductivity (roughly 0 S/cm).

Fig. 17. a) CVs at 100 mV/s for catalysts with various loading of iridium, b) average anodic charge measured, in the voltage range of 0.45–1 V vs. Ag/AgCl, from 3-5 CVs.

Ir/TaC and Ir/WC were the two supported catalysts with an average performance (350–400 mA/mg at 1.48 V), the TaC and WC had very good OER activity (Tafel slope of 120–130 mV) and good electrical conductivity (25–70 S/cm); however, both had an small surface area (0.7-1.7 m2/g). NbC, despite having an excellent electrical conductivity (170 S/ cm) and very good activity (Tafel slope of 122 mV) did not show impressive performance when iridium was dispersed on its surface. It was unexpected to see that TiC, in spite of having a very good surface area (30 m2/g) and average conductivity (2.2 S/cm), revealed the lowest performance when used as support for iridium (40 mA/mg at 1.48 V vs. Ag/AgCl). It should be noted that TiC had the poorest activity for OER itself (Tafel slope of 312 mV). Therefore, we can summarize the importance of the support properties in the following order: Support OER Activity  Support Surface Area >> Electrical Conductivity This means that in order to choose a good support for iridium electrocatalyst considering both the OER activity and the surface area of support is important. Having a high surface area exclusively (as seen for TiC) or a high activity exclusively (as seen in NbC) was not enough to assure a good OER performance for iridium supported catalysts. Nevertheless, the support conductivity did not appear to be a crucial factor, as two of the best performances were achieved using ATO and NbO2 which had the poorest electrical conductivities. Additionally, the effect of support surface area on the performance of iridium supported catalyst was studied. TaC and NbO2 with a range of surface area was obtained using different periods of ball milling. After 4 days of ball milling, the surface area of TaC and NbO2 was increased by 7-fold and 10-fold, respectively; and the conductivity was decreased by 96% and 100%, respectively. The increase in the TaC support surface area improved the performance of 20 wt% Ir/TaC by around 50%. However, the

F. Karimi, B.A. Peppley / Electrochimica Acta 246 (2017) 654–670

665

Table 5 Summary of the key results for support properties and the performance of supported catalysts.

Support BET Surface Area (m2/g) Support Conductivity (S/cm) Support Tafel Slope (mV) 20 wt% Ir/Support Current (mA/mg) at 1.48 V vs. Ag/AgCl

TaC

TiC

NbC

NbO2

WC

ATO

 0.8  0.1 UU 68  4 UU 126  15 O 410  39

UU 28  2.5 O 2.2  0.2  313  54  42  26

 0.9  0.3 UU 173  17 UU 123  7  266  75

 1.0  0.0  0.0001  0.0000 U 156  12 U 554  39

 1.7  0.1 U 25  2 U 146  20 O 355  124

UU 35  5.1  0.01  0.0 O 175  14 UU 845  159

increase in the NbO2 support surface area did not show any significant effect on the performance of 20 wt% Ir/NbO2. Additionally, 2, 5, 10, 20, 30 and 100 wt% Ir supported on TaC was synthesized and the effect of Ir loading on the Ir/TaC catalyst performance was studied. It was found that increasing iridium loading decreased the surface area (and consequently decreased the ECSA) of iridium electrocatalyst. This was mainly due to higher agglomeration of iridium electrocatalyst resulted from the multilayer deposition of iridium on the support. The performance of the Ir/TaC improved significantly (by a factor of 8) when the loading was decreased from 100 wt% to 5 wt%, after that the performance decreased when the loading was decreased further to 2 wt%. Thus, it was seen that an optimum loading of iridium (5 wt%) on the support (TaC, with 0.85 m2/g surface area) provided an optimum utilization of iridium electrocatalyst.

Acknowledgments We would like to thank Matthew A. McNeil for his help with editing and proof reading of this document. The financial support from NSERC
CREATE Distributed Generation for Remote Communities (DGRC) program at University of Toronto is also gratefully appreciated. Appendix A. Thermogravimetric Analysis (TGA) TGA and Derivative Thermogravimetric (DTG) curves are presented in Fig. A1(1–6) . The TGA results showed significant weight gains for the metal carbides, and relatively minor changes in weight for the metal oxides. The weight change values, and the

Fig. A1. TGA curves in air for supports 1) Tungsten Carbide (WC), 2) Tantalum Carbide (TaC), 3) Niobium Oxide (NbO2), 4) Niobium Carbide (NbC), 5) Titanium Carbide (TiC) and 6) Antimony-doped Tin Oxide (Sb2O5-SnO2).

666

F. Karimi, B.A. Peppley / Electrochimica Acta 246 (2017) 654–670

Table A1 DTG peak temperature, measured weight gain, and the theoretical weight gain. Support

Peak Temp( C)

Weight Gain (wt%)

NbC TiC WC TaC NbO2 Sb2O5-SnO2

550 440 550 740 390 510

26 26 18 13 6 4 (Loss)

Metal Oxides and Theoretical Weight Gain (%) Oxide 1

Dm1 (%)

Oxide 2

Dm2 (%)

Oxide 3

Dm3 (%)

NbO TiO WO2 TaO2 – –

3.8 6.6 10.2 10.3 – –

NbO2 TiO2 WO3 Ta2O5 – –

19.0 33.3 18.3 14.5 – –

Nb2O5 Ti2O3 W2O3 – – –

26.6 20.0 6.1 – – –

temperature at which the weight change occurred, are reported in Table A1. The weight gains observed in the TGA curves of metal carbides were likely due to thermal oxidation of the carbides to CO2 and metal oxide [75–77]. The common oxides for tungsten are WO2, WO3, or W2O3, and they could have been formed by the following reactions: WC + 2O2 ! WO2 + CO2

(Reaction A.1)

5 WC þ O2 ! WO3 þ CO2 2

ðReactionA:2Þ

5 2WC þ O2 ! W2 O3 þ 2CO 2

ðReactionA:3Þ

The theoretical weight gains (Dm%) corresponding to the reactions A.1, A.2, and A.3 can be calculated using Eqs. A.1, A.2 and A.3, respectively:   MWO2 Dm1 ð%Þ ¼
1  100 ðEq:?A:1Þ MWC

Dm2 ð%Þ ¼

  MWO3
1  100 MWC

ðEq:?A:2Þ

Dm3 ð%Þ ¼

  MW2 O3
1  100 2MWC

ðEq:?A:3Þ

The values of Dm1, Dm2, and Dm2 are reported in Table A1. The actual weight gain observed in the TGA curve of WC (18%) was close to the theoretical weight gain corresponding to Reaction A.2, Dm2 (18.3%). This suggests that the product of thermal oxidation of WC during the TGA was mostly WO3. Similar findings were reported by Kurlov et al. for the TGA of WC [77]. The common forms of metal oxides for other metal carbides, and the corresponding calculated theoretical weight gain (%) are summarized in Table A1. Comparing the calculated theoretical weight gain with the actual weight gain observed in the TGA curves, one can say that the resulting oxide of thermal oxidation of NbC is more likely Nb2O5. However, in case of TiC it was not clear what form of oxides were formed, as the actual weight gain for TiC was in between the theoretical weight gain for both TiO2 and Ti2O3. Similarly, the actual weight gain for TaC was close to the theoretical weight gain for both TaO2 and Ta2O3. The weight gain (6%) observed in the TGA curve of NbO2 at 390  C could be due to oxidation of NbO2 to Nb2O5 based on the following reaction: 1 2NbO2 þ O2 ! 2Nb2 O5 2

ðReactionA:4Þ

The minor weight loss (4%) observed in the TGA of Sb2O5-SnO2, might had been partly due to deoxygenation of Sb2O5 to Sb6O13 [78], and decomposition of carbonaceous residuals/contaminations. Appendices B. Transmission Electron Microscopy (TEM) TEM images of the three best performing catalysts from Section 3.2 (20 wt% Ir/TaC, 20 wt% Ir/NbO2, and 20 wt% Ir/ATO) are shown in Fig. B1(1). Fig. B1(1–3) shows the TEM images for 20 wt% Ir/TaC. These images reveal that some of the iridium species (seen with red color solid arrows in Fig. B1(1)) have formed clusters on the surface of the support material. Demonstrated in the square of Fig B1(1), the iridium species formed a thin layer on some parts on the surface of the support. A TEM image with higher magnification of the same area is shown in Fig. B1(2), and in this image the thin layer of iridium cluster with 12 nm thickness is seen. However; in some other parts, the iridium species formed an excessive agglomeration without being properly deposited on the surface of the support. These agglomerations are shown with dotted circles in Fig. B1(1) and with arrows in Fig. B1(3). These large quantities of the iridium clusters with high rate of agglomerations were formed more likely due to the large amount of iridium loading (20 wt%) compared to the low surface area of TaC (0.85 m2/g). Similar morphological structure was observed for the 20 wt% Ir/ NbO2 (Fig. B1(4–6). In Fig. B1(4 and 5), the iridium species that were deposited on the support material are shown with red arrows. In this sample, similar to the 20 wt% Ir/TaC, a large quantity of iridium clusters with excessive agglomerations was seen. Some of these iridium agglomerations were not even deposited on any support and existed in the form of un-supported iridium species. These un-supported iridium clusters are shown with blue arrows in Fig. B1(6). The high rate of agglomeration, similar to Ir/TaC, is more likely due to the large amount of iridium loading (20 wt%) for the NbO2 with a low surface area (1 m2/g). Additionally, for this catalyst, some bare surfaces of the support material were detected where no iridium/small amount of iridium seemed to be deposited. Examples of these areas are shown with dotted arrows in Fig. B1 (5). The TEM images of 20 wt% Ir/ATO is shown in Fig. B1(7–9). Two types of particles are clearly seen in these images. The larger particles ( 10–50 nm size) with lighter contrast, were the ATO support material, which are shown with dotted arrows in Fig. B1(7–9). The smaller particles with darker contrast, which are shown with solid arrows in Fig. B1(7–9), were the iridium species distributed on the surface of ATO particles. The iridium agglomoration in this catalyst did not exist in the extend seen in the previous two catalysts, this was due to small ATO particle size and its relatively high surface area (35 m2/g). The small particle size of the ATO (compared to the NbO2 and TaC) allowed a good trasnmission of electrons through the sample during TEM analysis. This enabled detecting the crystalline lattice fringes of tin oxide in

F. Karimi, B.A. Peppley / Electrochimica Acta 246 (2017) 654–670

667

Fig. B1. 1, 2, and 3) TEM images of 20 wt% Ir/TaC, 4, 5, and 6) TEM images of 20 wt% Ir/NbO2, 7, 8, 9) TEM images of 20 wt% Ir/ATO.

the TEM images. Some of these crystalline lattice fringes are shown with squares in Fig. B1(8 and 9).

Appendix C. X-ray Photoelectron Spectroscopy (XPS) The narrow scan XPS spectrum for Ir 4f in the iridium based catalyst, synthesized with the polyol method, is shown in Fig. C1. The asymmetric XPS peaks of Ir 4f7/2 and 4f5/2 with doublet splitting energy of 2.9 eV were detected in the sample. The asymmetric shape of Ir 4f is due to high density of states at the Fermi level and high degree of core hole screening [79]. The binding energy (BE) of Ir 4f7/2 for metallic iridium (Ir0) reported in the literature is around 60.7 eV [15,80–82]. However, the BE of Ir 4f7/2 in iridium oxide (IrO2) reported in the literature has a wider range of 61.9-62.7 eV [82–85]. These two reference BEs are also indicated in Fig. C1. The BE of Ir 4f7/2 in the iridium species synthesized with the polyol method lies in between the BE of Ir0 and Ir+4. The results indicate that iridium species exist mostly in a form in which the oxidation state of iridium is smaller than +4. This

Fig. C1. Narrow Scan XPS spectrum for Ir 4f in iridium catalyst synthesized by the polyol method.

668

F. Karimi, B.A. Peppley / Electrochimica Acta 246 (2017) 654–670

Fig. C2. 1) Narrow scan XPS Spectrum for Ta 4f in TaC before polyol procedure, 2) Ta 4f in TaC after polyol procedure, 3) C 1S in TaC before polyol procedure, 4) C 1S in TaC after polyol procedure.

could be in the form of an iridium suboxide (IrOx, x < 2) [53,57,82] such as Ir2O3 [86]. The XPS analysis was also performed on tantalum carbide before and after polyol synthesis to analyze the change in the oxidation state of tantalum during polyol procedure. For this purpose, the polyol procedure was repeated with tantalum carbide (without adding the iridium metal precursor). The material after polyol procedure was collected and analyzed using XPS analysis. The XPS analysis was also performed on TaC in its initial form (without exposure to polyol synthesis environment) for comparison. The narrow scan XPS spectrum for Ta 4f and C 1 s in TaC before polyol synthesis are shown in Fig. C2(1 and 3). Similarly, the narrow scan XPS spectrum for Ta 4f and C 1 s in TaC after polyol synthesis are shown in Fig. C2(2 and 4). The BE of Ta 4f 7/2 in the form of elemental Ta, TaC, TaOx (x < +5), and Ta2O5 are as follows: 22.0 eV [80], 23.5 eV [87], 25.5 eV [88], and 26.3 eV [80]. Also, the BE of C 1S in the form of carbide, adventitious carbon, C

O, and C¼O are as follows: 282.8 eV, 284.5 eV, 286 eV, and 288.5 eV [80,89]. Most samples that have been exposed to the atmosphere will have a noticeable quantity of adventitious carbon contamination. These BE values were used to perform curve fitting on Ta 4f and C 1S spectra. Comparing Fig. C2(1 and 2), it’s evident that the intensity of the Ta 4f peaks in the form of carbide were decreased after polyol procedure; while, the intensity of Ta 4f peaks related to oxide species were increased. The result of curve fitting revealed that 70% of tantalum existed in the form of carbide in the initial TaC, while after polyol procedure this value reduced to 40%. On the other hand, the amount of TaOx and Ta2O5 were increased by approximately two times after polyol procedure. Additionally, comparison of Fig. C2(3 and 4) shows that the intensity of C 1S peak

related to carbide material was decreased after polyol procedure. The result of curve fitting showed that 30% of the carbon existed in the form of carbides in the initials TaC, and this value decreased to 10% after polyol procedure. These results are suggesting that some of the TaC material, more likely those existed on the top layer surface, are converting to oxides species during polyol procedure. Knowing that the oxide materials generally have lower electrical conductivity compared to metal carbides [64], the reduction in the conductivity of Ir/Carbides seen in Section 3.2 can be attributed to the formation of oxide layer on the surface of support material during polyol procedure.

Fig. C3. Narrow scan XPS Spectrum for Tin (Sn 3d) in Ir/ATO catalyst synthesized with the polyol method.

F. Karimi, B.A. Peppley / Electrochimica Acta 246 (2017) 654–670

The XPS analysis was also performed on Ir/ATO, synthesized with the polyol method, to examine if tin in the tin oxide went through any chemical transformation during polyol procedure. The XPS spectrum of Sn 3d is shown in Fig. C3. The XPS peaks of Sn 3d5/2 and 3d3/2 with doublet splitting energy of 8.5 eV [80] were detected in the sample. The BEs of Sn3d5/2 in the form of Sn, SnO, and SnO2 are 485.0 [80,90], 486.3 [80,91] and 487.1 eV [80,90], respectively (indicated in Fig. C3). The BE of Sn 3d5/2 in the catalyst exactly aligned with the BE of Sn 3d5/2 in SnO2. This confirms that tin oxide did not go through transformation during polyol procedure as the Sn element in the support material after polyol synthesis existed in the form of SnO2. References [1] A. Godula-Jopek, Hydrogen production: by electrolysis, John Wiley & Sons, 2015. [2] F. Barbir, PEM electrolysis for production of hydrogen from renewable energy sources, Sol. Energy 78 (2005) 661–669, doi:http://dx.doi.org/10.1016/j. solener.2004.09.003. [3] K. Mazloomi, C. Gomes, Hydrogen as an energy carrier: Prospects and challenges, Renew. Sustain. Energy Rev. 16 (2012) 3024–3033, doi:http://dx. doi.org/10.1016/j.rser.2012.02.028. [4] E. Mayousse, F. Maillard, F. Fouda-Onana, O. Sicardy, N. Guillet, Synthesis and characterization of electrocatalysts for the oxygen evolution in PEM water electrolysis, Int. J. Hydrogen Energy 36 (2011) 10474–10481, doi:http://dx.doi. org/10.1016/j.ijhydene.2011.05.139. [5] P. Mazúr, J. Polonský, M. Paidar, K. Bouzek, Non-conductive TiO2 as the anode catalyst support for PEM water electrolysis, Int. J. Hydrogen Energy 37 (2012) 12081–12088, doi:http://dx.doi.org/10.1016/j.ijhydene.2012.05.129. [6] M. Carmo, D.L. Fritz, J. Mergel, D. Stolten, A comprehensive review on PEM water electrolysis, Int. J. Hydrogen Energy 38 (2013) 4901–4934, doi:http://dx. doi.org/10.1016/j.ijhydene.2013.01.151. [7] S. Trasatti, Electrocatalysis in the anodic evolution of oxygen and chlorine, Electrochim. Acta 29 (1984) 1503–1512, doi:http://dx.doi.org/10.1016/00134686(84)85004-5. [8] S. Song, H. Zhang, X. Ma, Z. Shao, R.T. Baker, B. Yi, Electrochemical investigation of electrocatalysts for the oxygen evolution reaction in PEM water electrolyzers, Int. J. Hydrogen Energy 33 (2008) 4955–4961, doi:http://dx.doi. org/10.1016/j.ijhydene.2008.06.039. [9] T. Reier, M. Oezaslan, P. Strasser, Electrocatalytic Oxygen Evolution Reaction (OER) on Ru, Ir, and Pt Catalysts: A Comparative Study of Nanoparticles and Bulk Materials, ACS Catal. 2 (2012) 1765–1772, doi:http://dx.doi.org/10.1021/ cs3003098. [10] J.L. Weininger, R.R. Russell, Corrosion of the Ruthenium Oxide Catalyst at the Anode of a Solid Polymer Electrolyte Cell, J. Electrochem. Soc. 125 (1978) 1482, doi:http://dx.doi.org/10.1149/1.2131700. [11] T. Lou9 cka, The reason for the loss of activity of titanium anodes coated with a layer of RuO2 and TiO2, J. Appl. Electrochem. 7 (1977) 211–214, doi:http://dx. doi.org/10.1007/BF00618987. [12] J. Xu, G. Liu, J. Li, X. Wang, The electrocatalytic properties of an IrO2/SnO2 catalyst using SnO2 as a support and an assisting reagent for the oxygen evolution reaction, Electrochim. Acta 59 (2012) 105–112, doi:http://dx.doi.org/ 10.1016/j.electacta.2011.10.044. [13] S. Siracusano, V. Baglio, C. D’Urso, V. Antonucci, A.S. Arico, Preparation and characterization of titanium suboxides as conductive supports of IrO2 electrocatalysts for application in SPE electrolysers, Electrochim. Acta 54 (2009) 6292–6299, doi:http://dx.doi.org/10.1016/j.electacta.2009.05.094. [14] A.V. Nikiforov, A.L. Tomas Garcia, I.M., Petrushina, E., Christensen, N.J. Bjerrum, Preparation and study of IrO2/SiC-Si supported anode catalyst for high temperature PEM steam electrolysers, Int. J. Hydrogen Energy 36 (2011) 5797– 5805, doi:http://dx.doi.org/10.1016/j.ijhydene.2011.02.050. [15] T. Reier, D. Teschner, T. Lunkenbein, a Bergmann, S. Selve, R. Kraehnert, R. Schlögl, P. Strasser, Electrocatalytic Oxygen Evolution on Iridium Oxide: Uncovering Catalyst-Substrate Interactions and Active Iridium Oxide Species, J. Electrochem. Soc. 161 (2014) F876–F882, doi:http://dx.doi.org/10.1149/ 2.0411409jes. [16] E. Rasten, G. Hagen, R. Tunold, Electrocatalysis in water electrolysis with solid polymer electrolyte, Electrochim. Acta 48 (2003) 3945–3952, doi:http://dx. doi.org/10.1016/j.electacta.2003.04.001. [17] J.C. Cruz, V. Baglio, S. Siracusano, R. Ornelas, L. Ortiz-Frade, L.G. Arriaga, V. Antonucci, A.S. Arico, Nanosized IrO2 electrocatalysts for oxygen evolution reaction in an SPE electrolyzer, J. Nanoparticle Res. 13 (2011) 1639–1646, doi: http://dx.doi.org/10.1007/s11051-010-9917-2. [18] J.M. Hu, H.M. Meng, J.Q. Zhang, C.N. Cao, Degradation mechanism of long service life Ti/IrO2-Ta2O5 oxide anodes in sulphuric acid, Corros. Sci. 44 (2002) 1655–1668, doi:http://dx.doi.org/10.1016/S0010-938X(01)00165-2. [19] M. Morimitsu, R. Otogawa, M. Matsunaga, Effects of cathodizing on the morphology and composition of IrO2-Ta2O5/Ti anodes, Electrochim. Acta 46 (2000) 401–406, doi:http://dx.doi.org/10.1016/S0013-4686(00)00598-3.

669

[20] A.T. Marshall, S. Sunde, M. Tsypkin, R. Tunold, Performance of a PEM water electrolysis cell using Irx Ruy Taz O2 electrocatalysts for the oxygen evolution electrode, Int. J. Hydrogen Energy 32 (2007) 2320–2324, doi:http://dx.doi.org/ 10.1016/j.ijhydene.2007.02.013. [21] C.P. De Pauli, S. Trasatti, Electrochemical surface characterization of IrO2 + SnO2 mixed oxide electrocatalysts 1, J. Electroanal. Chem. 396 (1995) 161–168. [22] A. Marshall, B. Børresen, G. Hagen, M. Tsypkin, R. Tunold, Electrochemical characterisation of IrxSn1-xO 2 powders as oxygen evolution electrocatalysts, Electrochim. Acta 51 (2006) 3161–3167, doi:http://dx.doi.org/10.1016/j. electacta.2005.09.004. [23] S. Ardizzone, C.L. Bianchi, G. Cappelletti, M. Ionita, A. Minguzzi, S. Rondinini, A. Vertova, Composite ternary SnO2-IrO2-Ta2O5 oxide electrocatalysts, Journal of Electroanalytical Chemistry (2006), doi:http://dx.doi.org/10.1016/j. jelechem.2006.02.004. [24] G. Chen, X. Chen, P.L. Yue, Electrochemical Behavior of Novel Ti/IrOx-Sb2O5SnO2 Anodes, J. Phys. Chem. B 106 (2002) 4364–4369, doi:http://dx.doi.org/ 10.1021/jp013547o. [25] N.J. Pérez-Viramontes, I.L. Escalante-García, C. Guzmán-Martínez, M. GalvánValencia, S.M. Durón-Torres, Electrochemical study of Ir-Sn-Sb-O materials as catalyst-supports for the oxygen evolution reaction, J. Appl. Electrochem. 45 (2015) 1165–1173, doi:http://dx.doi.org/10.1007/s10800-015-0875-3. [26] J. Cheng, H. Zhang, H. Ma, H. Zhong, Y. Zou, Preparation of Ir0.4Ru0.6MoxOy for oxygen evolution by modified Adams’ fusion method, Int. J. Hydrogen Energy 34 (2009) 6609–6613, doi:http://dx.doi.org/10.1016/j.ijhydene.2009.06.061. [27] Z.-G. Ye, H.-M. Meng, D. Chen, H.-Y. Yu, Z.-S. Huan, X.-D. Wang, D.-B. Sun, Structure and characteristics of Ti/IrO2(x)+MnO2(1-x) anode for oxygen evolution, Solid State Sci. 10 (2008) 346–354, doi:http://dx.doi.org/10.1016/j. solidstatesciences.2007.09.011. [28] J. Polonský, I.M. Petrushina, E. Christensen, K. Bouzek, C.B. Prag, J.E.T. Andersen, N.J. Bjerrum, Tantalum carbide as a novel support material for anode electrocatalysts in polymer electrolyte membrane water electrolysers, Int. J. Hydrogen Energy 37 (2012) 2173–2181, doi:http://dx.doi.org/10.1016/j. ijhydene.2011.11.035. [29] A. Kumar, V. Ramani, Strong Metal-Support Interactions Enhance the Activity and Durability of Platinum Supported on Tantalum-Modified Titanium Dioxide Electrocatalysts, ACS Catal. 4 (5) (2014) 1516–1525, doi:http://dx.doi.org/ 10.1021/cs500116h. [30] O.E. Haas, S.T. Briskeby, O.E. Kongstein, M. Tsypkin, R. Tunold, Synthesis and Characterisation of RuxTix-1O2 as a Catalyst Support for Polymer Electrolyte Fuel Cell, J. New Mater. Electrochem. Syst. 11 (2008) 21–29. http://www. groupes.polymtl.ca/jnmes/modules/journal/index.php/content0096.html. [31] J. Xu, Q. Li, E. Christensen, X. Wang, N.J. Bjerrum, Platinum Activated IrO2/SnO2 Nanocatalysts and Their Electrode Structures for High Performance Proton Exchange Membrane Water Electrolysis, Int. J. Electrochem. Sci. 8 (2013) 2388–2406. [32] A.T. Marshall, R.G. Haverkamp, Electrocatalytic activity of IrO2-RuO2 supported on Sb-doped SnO2 nanoparticles, Electrochim. Acta 55 (2010) 1978–1984, doi:http://dx.doi.org/10.1016/j.electacta.2009.11.018. [33] J. Xu, Q. Li, M.K. Hansen, E. Christensen, A.L. Tomas Garcia, G. Liu, X. Wang, N.J. Bjerrum, Antimony doped tin oxides and their composites with tin pyrophosphate as catalysts supports for oxygen evolution reaction in proton exchange membrane water electrolysis, Int. J. Hydrogen Energy 37 (2012) 18629–18640. [34] H.-S. Oh, H.N. Nong, T. Reier, M. Gliech, P. Strasser, Oxide-supported Ir nanodendrites with high activity and durability for the oxygen evolution reaction in acid PEM water electrolyzers, Chem. Sci. 6 (2015) 3321–3328, doi: http://dx.doi.org/10.1039/C5SC00518C. [35] M.S. Thomassen, T. Mokkelbost, E. Sheridan, A. Lind, Supported Nanostructured Ir and IrRu Electrocatalysts for Oxygen Evolution in PEM Electrolysers, ECS Trans, The Electrochemical Society, 2011, pp. 271–279, doi: http://dx.doi.org/10.1149/1.3654225. [36] V. Ávila-Vázquez, J.C. Cruz, M. Galván-Valencia, J. Ledesma-García, L.G. Arriaga, C. Guzmán, S.M. Durón-Torres, P. Escobedo, M. Querétaro, Electrochemical Study of Sb-Doped SnO2 Supports on the Oxygen Evolution Reaction: Effect of Synthesis Annealing Time, Int. J. Electrochem. Sci. Int. J. Electrochem. Sci. 8 (2013) 10586–10600 www.electrochemsci.org (Accessed 19 July 2016). [37] L. Ma, S. Sui, Y. Zhai, Preparation and characterization of Ir/TiC catalyst for oxygen evolution, J. Power Sources 177 (2008) 470–477, doi:http://dx.doi.org/ 10.1016/j.jpowsour.2007.11.106. [38] J. Polonský, P. Mazúr, M. Paidar, E. Christensen, K. Bouzek, Performance of a PEM water electrolyser using a TaC-supported iridium oxide electrocatalyst, Int. J. Hydrogen Energy 39 (2014) 3072–3078, doi:http://dx.doi.org/10.1016/j. ijhydene.2013.12.085. [39] V. Adamaki, F. Clemens, P. Ragulis, S.R. Pennock, J. Taylor, C.R. Bowen, Manufacturing and Characterization of Magnéli Phase Conductive Fibres, J. Mater. Chem. A 2 (2014) 8328–8333. [40] J. Zhang, L. Gao, Synthesis and characterization of antimony-doped tin oxide (ATO) nanoparticles by a new hydrothermal method, Mater. Chem. Phys. 87 (2004) 10–13, doi:http://dx.doi.org/10.1016/j.matchemphys.2004.06.004. [41] Z. Orel, B. Orel, M. Hodos, V. Kaucic, Conductive SnO2/Sb powder: preparation and optical properties, J. Mater. Sci. 27 (1992) 313–318. [42] X. Wu, K. Scott, RuO2 supported on Sb-doped SnO2 nanoparticles for polymer electrolyte membrane water electrolysers, Int. J. Hydrogen Energy 36 (2011) 5806–5810, doi:http://dx.doi.org/10.1016/j.ijhydene.2010.10.098.

670

F. Karimi, B.A. Peppley / Electrochimica Acta 246 (2017) 654–670

[43] R. Hutchings, K. Muller, R. Kotz, S. Stucki, A structural investigation of stabilized oxygen evolution catalysts, J. Mater. Sci. 19 (1984) 3987–3994, doi: http://dx.doi.org/10.1007/BF00980762. [44] H. Leidheiser, A.F. Rauch, E.M. Ibrahim, R.D. Granata, The Effect of Antimony on the Corrosion of Tin in Deaerated Citric Acid, (1982) , pp. 1651–1658. [45] S.S. El-Egamy, Electrochemical Behavior of Antimony and Antimony Oxide Films in Acid Solutions, CORROSION 62 (2006) 739–744, doi:http://dx.doi.org/ 10.5006/1.3278298. [46] V. Natarajan, S. Basu, K. Scott, Effect of treatment temperature on the performance of RuO2 anode electrocatalyst for high temperature proton exchange membrane water electrolysers, Int. J. Hydrogen Energy 38 (2013) 16623–16630, doi:http://dx.doi.org/10.1016/j.ijhydene.2013.05.133. [47] Y. Liu, T.G. Kelly, J.G. Chen, W.E. Mustain, Metal Carbides as Alternative Electrocatalyst Supports, ACS Catal. 3 (2013) 1184–1194, doi:http://dx.doi.org/ 10.1021/cs4001249. [48] Y.J. Wang, D.P. Wilkinson, J.J. Zhang, Noncarbon Support Materials for Polymer Electrolyte Membrane Fuel Cell Electrocatalysts, Chem. Rev. 111 (2011) 7625– 7651, doi:http://dx.doi.org/10.1021/cr100060r. [49] S. Meyer, A.V. Nikiforov, I.M. Petrushina, K. Köhler, E. Christensen, J.O. Jensen, N.J. Bjerrum, Transition metal carbides (WC, Mo2C, TaC, NbC) as potential electrocatalysts for the hydrogen evolution reaction (HER) at medium temperatures, Int. J. Hydrogen Energy 40 (2015) 2905–2911, doi:http://dx.doi. org/10.1016/j.ijhydene.2014.12.076. [50] W.-F. Chen, K. Sasaki, J.M. Schneider, C.-H. Wang, J.T. Muckerman, E. Fujita, Nitride-Stabilized Tungsten Carbide Electrocatalysts for Hydrogen Evolution Reaction, 66th Annual Meeting of International Society of Electrochemistry, At Taipei, 2015. https://www.researchgate.net/publication/284713696_NitrideStabilized_Tungsten_Carbide_Electrocatalysts_for_Hydrogen_Evolution_ Reaction. [51] D.J. Ham, J.S. Lee, Transition Metal Carbides and Nitrides as Electrode Materials for Low Temperature Fuel Cells, Energies 2 (2009) 873–899, doi:http://dx.doi. org/10.3390/en20400873. [52] R. Adams, R.L. Shriner, Platinum Oxide As a Catalyst in the Reduction of Organic Compounds. III. Preparation and Properties of the Oxide of Platinum Obtained by the Fusion of Chloroplatinic Acid with Sodum Nitrate 1, J. Am. Chem. Soc. 45 (1923) 2171–2179, doi:http://dx.doi.org/10.1021/ja01662a022. [53] V. Pfeifer, T.E. Jones, J.J. Velasco Velez, C. Massue, R. Arrigo, D. Teschner, F. Girgsdies, M. Scherzer, M.T. Greiner, J. Allan, M. Hashagen, G. Weinberg, S. Piccinin, M. Havecker, A. Knop-Gericke, R. Schlogl, The electronic structure of iridium and its oxides, Surf. Interface Anal. 48 (2016) 261–273, doi:http://dx. doi.org/10.1002/sia.5895. [54] R.D.L. Smith, M.S. Prévot, R.D. Fagan, Z. Zhang, P.A. Sedach, M.K.J. Siu, S. Trudel, C.P. Berlinguette, Photochemical route for accessing amorphous metal oxide materials for water oxidation catalysis, Science. 340 (2013) 60–63, doi:http:// dx.doi.org/10.1126/science.1233638. [55] F. Fievet, J. Lagier, B. Blin, B. Beaudoin, M. Figlaez, Homogeneous and heterogeneous nucleations in the polyol process for the preparation of micron and submicron size metal particles, Solid State Ionics 32–33 (1989) 198–205, doi:http://dx.doi.org/10.1016/0167-2738(89)90222-1. [56] F. Karimi, B.A. Peppley, Comparison of conventional versus microwave heating for polyol synthesis of supported iridium based electrocatalyst for polymer electrolyte membrane water electrolysis, Int. J. Hydrogen Energy 42 (2017) 5083–5094, doi:http://dx.doi.org/10.1016/j.ijhydene.2017.01.090. [57] F. Karimi, A. Bazylak, B.A. Peppley, Effect of Calcination Temperature on the Morphological and Electrochemical Characteristics of Supported Iridium Hydroxyoxide Electrocatalysts for the PEM Electrolyzer Anode, J. Electrochem. Soc. 164 (2017) F464–F474, doi:http://dx.doi.org/10.1149/2.0111706jes. [58] F. Wudl, M.R. Bryce, Apparatus for two-probe conductivity measurements on compressed powders, J. Chem. Educ. 67 (1990) 717, doi:http://dx.doi.org/ 10.1021/ed067p717. [59] A. Celzard, J.F. Marêché, F. Payot, G. Furdin, Electrical conductivity of carbonaceous powders, Carbon N. Y. 40 (2002) 2801–2815, doi:http://dx. doi.org/10.1016/S0008-6223(02)00196-3. [60] J. Sanchez-Gonzalez, A. Macias-Garcia, M.F. Alexandre-Franco, V. GomezSerrano, Electrical conductivity of carbon blacks under compression, Carbon N. Y. 43 (2005) 741–747, doi:http://dx.doi.org/10.1016/j.carbon.2004.10.045. [61] A. Espinola, P.M. Miguel, M.R. Salles, A.R. Pinto, Electrical properties of carbons—resistance of powder materials, Carbon N. Y. 24 (1986) 337–341, doi: http://dx.doi.org/10.1016/0008-6223(86)90235-6. [62] R. Kirchhof, The particle shape influence on the apparent density and electric conductivity of pressed powder materials, Mater. Chem. 6 (1981) 209–222, doi:http://dx.doi.org/10.1016/0390-6035(81)90043-2. [63] S. Ted Oyama, Crystal structure and chemical reactivity of transition metal carbides and nitrides, J. Solid State Chem. 96 (1992) 442–445, doi:http://dx. doi.org/10.1016/S0022-4596(05)80279-8. [64] C.C. Wang, S.A. Akbar, W. Chen, V.D. Patton, Electrical properties of hightemperature oxides, borides, carbides, and nitrides, J. Mater. Sci. 30 (1995) 1627–1641, doi:http://dx.doi.org/10.1007/BF00351591. [65] Y. Sakai, N. Tsuda, T. Sakata, Electrical Properties of Semiconducting NbO 2, J. Phys. Soc. Japan 54 (1985) 1514–1518, doi:http://dx.doi.org/10.1143/JPSJ.54.1514. [66] J.D. Voorhies, Electrochemical and Chemical Corrosion of Tungsten Carbide (WC), J. Electrochem. Soc. 119 (1972) 219, doi:http://dx.doi.org/10.1149/ 1.2404164. [67] M.C. Weidman, D.V. Esposito, I.J. Hsu, J.G. Chen, Electrochemical Stability of Tungsten and Tungsten Monocarbide (WC) Over Wide pH and Potential Ranges, J. Electrochem. Soc. 157 (2010) F179, doi:http://dx.doi.org/10.1149/1.3491341.

[68] N. Heide, B. Siemensmeyer, J.W. Schultze, Surface characterization and electrochemical behaviour of nitrogen- and carbon-implanted titanium, Surf. Interface Anal. 19 (1992) 423–429, doi:http://dx.doi.org/10.1002/ sia.740190179. [69] S.J. Stott, R.J. Mortimer, S.E. Dann, M. Oyama, F. Marken, Electrochemical properties of core-shell TiCTiO2 nanoparticle lms immobilized at ITO electrode surfaces Electrochemical properties of core-shell TiC-TiO2 nanoparticle films immobilized at ITO electrode surfaces, Phys. Chem. Chem. Phys. 8 (2006), doi:http://dx.doi.org/10.1039/b610391j. [70] F. Mattos-Costa, P. de Lima-Neto, S.A. Machado, L. Avaca, Characterisation of surfaces modified by sol-gel derived RuxIr1-xO2 coatings for oxygen evolution in acid medium, Electrochim. Acta 44 (1998) 1515–1523, doi:http://dx.doi.org/ 10.1016/S0013-4686(98)00275-8. [71] K.H. Stern, Electrodeposition of refractory compounds from molten salts— borides, carbides and silicides, Metall. Ceram. Prot. Coatings, Springer, Netherlands, Dordrecht, 1996, pp. 54–73, doi:http://dx.doi.org/10.1007/97894-009-1501-5_4. [72] G.C. Wood, T. Hodgkiess, The Hardness of Oxides at Ambient Temperatures, Mater. Corros. Und Korrosion. 23 (1972) 766–773, doi:http://dx.doi.org/ 10.1002/maco.19720230905. [73] D.P. Joshi, K. Sen, Effect of grain size on the resistivity of polycrystalline material, Sol. Cells. 9 (1983) 261–267, doi:http://dx.doi.org/10.1016/03796787(83)90020-0. [74] S. Siracusano, V. Baglio, N. Van Dijk, L. Merlo, A.S. Aricò, Enhanced performance and durability of low catalyst loading PEM water electrolyser based on a shortside chain perfluorosulfonic ionomer, Appl. Energy 192 (2017) 477–489, doi: http://dx.doi.org/10.1016/j.apenergy.2016.09.011. [75] K.S. Anisimov, A.A. Malkov, A.A. Malygin, Mechanism of thermal oxidation of silicon carbide modified by chromium oxide structures, Russ. J. Gen. Chem. 84 (2014) 2375–2381, doi:http://dx.doi.org/10.1134/S1070363214120032. [76] A.L. Tomas-Garcia, Q. Li, J.O. Jensen, N.J. Bjerrum, High surface area tungsten carbides: Synthesis, characterization and catalytic activity towards the hydrogen evolution reaction in phosphoric acid at elevated temperatures, Int. J. Electrochem. Sci. 9 (2014) 1016–1032. [77] A.S. Kurlov, A.I. Gusev, Effect of particle size on the oxidation of WC powders during heating, Inorg. Mater. 47 (2011) 133–138, doi:http://dx.doi.org/10.1134/ S0020168511020099. [78] C.A. Cody, L. DiCarlo, R.K. Darlington, Vibrational and thermal study of antimony oxides, Inorg. Chem. 18 (1979) 1572–1576, doi:http://dx.doi.org/ 10.1021/ic50196a036. [79] G.K. Wertheim, H.J. Guggenheim, Conduction-electron screening in metallic oxides: IrO2, Phys. Rev. B 22 (1980) 4680–4683, doi:http://dx.doi.org/10.1103/ PhysRevB.22.4680. [80] J.F. Moulder, W.F. Stickle, P.E. Sobol, K.D. Bomben, Handbook of X-ray Photoelectron Spectroscopy, (1995), pp. 255, doi:http://dx.doi.org/10.1002/ sia.740030412. [81] R. Kötz, Anodic Iridium Oxide Films, J. Electrochem. Soc. 131 (1984) 72, doi: http://dx.doi.org/10.1149/1.2115548. [82] B.V. Crist, B.V. Crist, The handbook of monochromatic XPS spectra/[1], The elements and native oxides, Wiley, 2000. [83] L. da Silva, V. Alves, S. de Castro, J.F. Boodts, XPS study of the state of iridium, platinum, titanium and oxygen in thermally formed IrO2+TiO2+PtOX films, Colloids Surfaces A Physicochem. Eng. Asp. 170 (2000) 119–126, doi:http://dx. doi.org/10.1016/S0927-7757(99)00535-X. [84] A.K. Shukla, A.M. Kannan, M.S. Hegde, J. Gopalakrishnan, Effect of counter cations on electrocatalytic activity of oxide pyrochlores towards oxygen reduction/evolution in alkaline medium: an electrochemical and spectroscopic study, J. Power Sources 35 (1991) 163–173, doi:http://dx.doi.org/ 10.1016/0378-7753(91)80033-T. [85] M. Peuckert, XPS study on thermally and electrochemically prepared oxidic adlayers on iridium, Surf. Sci. 144 (1984) 451–464, doi:http://dx.doi.org/ 10.1016/0039-6028(84)90111-0. [86] W. Banerjee, S. Maikap, C.-S. Lai, Y.-Y. Chen, T.-C. Tien, H.-Y. Lee, W.-S. Chen, F.T. Chen, M.-J. Kao, M.-J. Tsai, J.-R. Yang, Formation polarity dependent improved resistive switching memory characteristics using nanoscale (1.3 nm) coreshell IrOx nano-dots, Nanoscale Res. Lett. 7 (2012) 194, doi:http://dx.doi.org/ 10.1186/1556-276X-7-194. [87] O.Y. Khyzhun, E. Zhurakovsky, A. Sinelnichenko, V. Kolyagin, Electronic structure of tantalum subcarbides studied by XPS, XES, and XAS methods, J. Electron Spectros. Relat. Phenomena 82 (1996) 179–192, doi:http://dx.doi.org/ 10.1016/S0368-2048(96)03057-5. [88] M. Khanuja, H. Sharma, B.R. Mehta, S.M. Shivaprasad, XPS depth-profile of the suboxide distribution at the native oxide/Ta interface, J. Electron Spectros. Relat. Phenomena 169 (2009) 41–45, doi:http://dx.doi.org/10.1016/j. elspec.2008.10.004. [89] Y.-H. Chang, J.-B. Wu, P.-J. Chang, H.-T. Chiu, C. Nordling, C.M. Allocca, Chemical vapor deposition of tantalum carbide and carbonitride thin films from Me3CE Ta(CH2CMe3)3 (E=CH, N)Electronic supplementary information (ESI) available: AFM and SEM images of TaC and TaCN films deposited on Si(100) at 773, 823 and 923K, J. Mater. Chem. 13 (2003) 365–369, doi:http://dx.doi.org/ 10.1039/b208129f See http://www.rsc.org/suppdata/jm/b2/b208129f/. [90] J. Luo, C. Xu, XPS Examination of Tin Oxide on Float Glass Surface, J. Non. Cryst. Solids 119 (1990) 37–40. [91] W.-S. Choi, The Fabrication of Tin Oxide Films by Atomic Layer Deposition using Tetrakis(Ethylmethylamino) Tin Precursor, Trans. Electr. Electron. Mater. 10 (2009) 200–202.