Ti Oxide Water Oxidation Electrocatalysts by ... - Wiley Online Library

DOI: 10.1002/celc.201500137

Articles

Roughened Zn-Doped Ru–Ti Oxide Water Oxidation Electrocatalysts by Blending Active and Activated Passive Components Koshal Kishor,[a] Sulay Saha ,[a] Manish Kumar Gupta ,[a] Anshumaan Bajpai,[a] Moitrayee Chatterjee,[a] Sri Sivakumar ,*[a, b, c] and Raj Ganesh S. Pala *[a, b] An approach to decreasing the overpotential, increasing the stability, and optimizing the noble-metal composition of electrocatalysts for the oxygen evolution reaction (OER) in acidic media is demonstrated. Essential components of this approach are: 1) combining an active (unstable Ru) component with a dopant (Zn)-activated passive (stable Ti) element, 2) blending these elements by co-electrodeposition in an acidic environment in which dissolution of the unstable component (excess Ru) promotes roughness, and 3) further increasing the rough-

ness of the resultant electrode through chemical inhomogeneity by the incorporation of Ti and through structural inhomogeneity by incorporation of Zn in RuO2. The composition of the electrode with the maximal activity is Ru0.258Ti0.736Zn0.006Ox, and its activity is four times higher than that of RuO2. The electrochemical stability towards the OER follows the order RuTiZn > RuTi > RuZn > Ru. This design strategy provides a facile method to improve activity without compromising stability.

1. Introduction The oxygen evolution reaction (OER) involves multiple electron-transfer steps, and the associated reaction kinetics of the individual steps poses significant challenges in the design of water electrolyzers.[1–6] Overcoming these hurdles requires an effective electrocatalyst that has high rate of reaction, low overpotential, and high electrochemical stability and is compatible with highly conducting polymer electrolytes such as Nafion. A composite of ruthenium- and iridium-based oxides has been extensively used due to favorable properties such as high electronic conductivity, low resistivity, and high thermal stability.[7–13] However, these metal-oxide electrocatalysts are expensive and scarce in nature, which restricts their widespread commercial application. Replacement of expensive catalysts by cheaper earth-abundant elements is critical for increased industrial application and has been the topic of much research.[1–4, 14, 15]

[a] K. Kishor,+ S. Saha ,+ M. K. Gupta ,+ A. Bajpai, Dr. M. Chatterjee, Dr. S. Sivakumar , Dr. R. G. S. Pala Department of Chemical Engineering Indian Institute of Technology Kanpur, Kanpur 208016 (India) E-mail: [email protected] [email protected]

[b] Dr. S. Sivakumar , Dr. R. G. S. Pala Material Science Programme Indian Institute of Technology Kanpur, Kanpur 208016 (India) [c] Dr. S. Sivakumar Centre for Environmental Science and Engineering Thematic Unit of Excellence on Soft Nanofabrication Indian Institute of Technology Kanpur, Kanpur 208016 (India) [+] These authors contributed equally to this work. Supporting Information for this article is available on the WWW under http://dx.doi.org/10.1002/celc.201500137.

ChemElectroChem 2015, 2, 1839 – 1846

RuO2 suffers from a high corrosion rate despite providing the highest current density amongst all the electrocatalysts used in acidic medium.[16–18] Under the typical operating conditions (i.e. potential > 1.4 VRHE) of OER catalysts, oxidation of RuO2 to RuO4 results in corrosion of the electrode.[16, 18] The general strategy for increasing the stability involves mixing an active (i.e. RuO2) and a stabilizing oxide (e.g. IrO2, SnO2, TiO2) to obtain a better OER electrocatalyst.[19–23] With respect to RuO2, the latter oxides, which have rutile structures, show higher overpotential for the OER and lower electrical conductivity.[5, 6] However, when these oxides are mixed with RuO2, the current density is higher than for the separate phases of individual oxides, and the solid solution has higher stability. Naslund et al. have suggested a charge-transfer mechanism between metal atoms to account for the increase in activity of RuO2– TiO2 electrodes.[19] A recent study has clearly unraveled the underlying mechanism of how addition of IrO2 to RuO2 aids in increasing stability without decreasing activity.[24] An alternative strategy to increase the activity and stability of RuO2 is doping with Ta, Ce, Co, Ni, or Nb. However, the extent of doping is restricted by the limited solubility of the dopant in the host rutile oxide.[22, 25–32] Herein, we demonstrate a design strategy for an OER electrocatalyst with the key features of: 1) combining an active but unstable element (Ru) with relatively passive but stable element (Ti) that is made more active by a dopant (Zn), 2) facile blending by co-electrodeposition, which not only promotes roughness in the electrode by deposition and dissolution of excess Ru, but also provides a methodology to optimize the composition of the ternary electrode, 3) further increase in electrode roughness by blending components that are chemically inhomogeneous but structurally similar (RuO2 and TiO2

1839

Ó 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Articles have the same crystal structure) and by blending components that are both structurally and chemically inhomogeneous (RuO2/ZnO or TiO2/ZnO have dissimilar crystal structures). During sintering after co-electrodeposition, inhomogeneities increase the corrugations in the potential-energy surfaces relevant to atomic surface diffusion.[33] This decreases the extent of sintering and hence increases the active surface area of the electrocatalysts. In addition, structural and chemical inhomogeneities tend to activate the catalysts. In this study, the best-performing Zn-doped Ru–Ti ternary composite electrode outperformed Zn-doped Ru composite and Ru–Ti composite binary mixed-oxide electrodes in terms of both activity and stability. A fourfold increase in current density was observed in the best-performing ternary electrode with respect to electrodeposited RuO2. The onset potential for water oxidation in acidic medium was reduced by 0.13 V on addition of Ti and Zn additives to RuO2. We attribute this increased activity to the higher surface roughness achieved by the addition of inhomogeneous Zn and Ti additives and also to the dissolution of excess Ru during co-electrodeposition. Additionally, Zn is expected to enhance electrical conductivity due to creation of oxygen vacancies in RuO2, TiO2, and their mixed oxides.[5, 34] Overall, this study demonstrates a strategy for increasing activity without compromising stability through addition of inexpensive elements, and thereby increasing the potential for large-area OER electrodes.

2. Results and Discussion 2.1. Electrocatalytic Activity and Stability of the Electrodes The activity and stability of undoped electrodeposited RuO2 and Ru composite oxide electrodes with Ti or Zn or both elements were analyzed by cyclic voltammetry (CV) and linearsweep voltammetry (LSV). The electrodes were subjected to 500 cycles to assess their stability. Furthermore, the overpotential was benchmarked with respect to the potential required for attaining a current density of 20 mA cm¢2. The overpotential needed to attain a current density of 20 mA cm¢2 and the

current density at a potential of 1.35 V (vs. Ag/AgCl) are summarized in Table 1. Both activity and electrochemical stability towards the OER of the RuO2 electrode are improved on coelectrodeposition with Zn and Ti. After five cycles, the activity for the four possible combinations with and without additives (Zn, Ti) of Ru-based co-electrodeposited electrodes follows the order RuO2 < RuTi < RuZn < RuTiZn. Among all the prepared electrodes, the ternary composite electrode with a composition of Ru0.258Ti0.736Zn0.006Ox was found to have the highest activity with a current density of 63 mA cm¢2 at a potential of 1.35 V versus Ag/AgCl (Supporting Information, Figure S1a). This is four times higher than the activity of RuO2. The overpotential of the ternary electrode is 0.13 V lower than that of the RuO2 electrode (Table 1). The best-performing co-electrodeposited RuZn and RuTi electrodes have the compositions of Ru0.96Zn0.04Ox and Ru0.33Ti0.67Ox, respectively (Figures S1 b and S1 c). The fact that a low Ru:Ti ratio drastically reduces the activity indicates that Ru is the more active component. However, in both the RuTi composite and co-electrodeposited RuTiZn electrodes, increasing the Ru:Ti ratio does not contribute to increasing the activity beyond an optimum composition. This may be correlated to the fact that increasing Ru content in the composite contributes to bulk enrichment, which does not increase the activity. A small amount of Zn doping was found to have a positive effect on the activity of the electrodes. The activities of the best-performing electrocatalysts in the categories of binary RuTi (Ru0.33Ti0.67Ox), RuZn (Ru0.96Zn0.04Ox), and ternary Ru0.258Ti0.736Zn0.006Ox electrodes over 500 cycles are compared with that of RuO2 in Figure 1 a. Chronoamperometry was performed for 108 h at a potential of 1.4 V (Ag/AgCl) for co-electrodeposited electrodes (Figure S3). The electrochemical stability towards the OER of the electrodes follows the order RuO2 < Ru0.96Zn0.04Ox < Ru0.33Ti0.67Ox < Ru0.258Ti0.736Zn0.006Ox (Figure 1 a) To further assess the stability of the electrodes, we normalized the current density with respect to the maximum current density at a potential of 1.3 V (vs. Ag/AgCl) over 500 cycles (Figure 1 b). Figure 1 b reveals the rate of loss of electrochemical activity, which is proportional to the rate of loss of active sites due to dissolution, with potential cycling. The

Table 1. Precursor concentrations for different compositions of the electrodes and catalytic activity (current density JE = 1.35 V) at 1.35 V vs. Ag/AgCl and overpotential needed to reach 20 mA cm¢2 in fifth electrochemical cycle with a scan rate of 5 mV s¢1] of the electrodes towards electrochemical water oxidation reaction in 0.5 m H2SO4 solution.

Electrodes

Composition

JE = 1.35 V [mA cm¢2]

Overpotential [V] at 20 mA cm¢2

Precursor concentration [mm] RuCl3 TiCl4

ZnSO4

co-electrodeposited RuZnTi

Ru0.11Ti0.881Zn0.009Ox Ru0.258Ti0.736Zn0.006Ox Ru0.376Ti0.616Zn0.008Ox Ru0.553Ti0.443Zn0.004Ox Ru0.10Ti0.90Ox Ru0.33Ti0.67Ox Ru0.39Ti0.61Ox Ru0.70Ti0.30Ox Ru0.98Zn0.02Ox Ru0.96Zn0.04Ox Ru0.93Zn0.07Ox Ru0.89Zn0.11Ox RuO2

10 63 59 52 7 37 21 18 32 53 47 43 17

0.42 0.22 0.23 0.25 0.47 0.27 0.31 0.33 0.3 0.25 0.27 0.28 0.35

7.5 7.5 7.5 7.5 7.5 7.5 7.5 7.5 7.5 7.5 7.5 7.5 7.5

10 10 40 5 – – – – 5 10 20 25 –

co-electrodeposited RuTi

co-electrodeposited RuZn

electrodeposited RuO2


www.chemelectrochem.org 1840

600 300 300 75 600 300 75 25 – – – – –


Articles creasing potential cycles (Table S2). The increased ECSA indicates enhancement of porosity during later stages of cycling. However despite the increase in porosity and ECSA, the current density decreases in later electrochemical cycles. This observation is rationalized by the fact that dissolution of Ru sites with exposure of Ti sites would lead to marginal lowering of activity but increasd stability during later electrochemical cycles. To further explore the effect of Ti and Zn on the Ru0.258Ti0.736Zn0.006Ox electrode, detailed structural and electrochemical analyses were performed on the best performing co-electrodeposited electrodes (i.e. Ru0.258Ti0.736Zn0.006Ox, Ru0.96Zn0.04Ox, and Ru0.33Ti0.67Ox). 2.2. XRD Analysis of the Electrodes

Figure 1. a) Normalized current density of Ru0.258Ti0.736Zn0.006Ox, Ru0.96Zn0.04Ox, Ru0.33Ti0.67Ox, and RuO2 at 1.3 V (vs. Ag/AgCl) in 0.5 m H2SO4 solution. b) Current densities of Ru0.258Ti0.736Zn0.006Ox, Ru0.96Zn0.04Ox, Ru0.33Ti0.67Ox, and RuO2 at 1.3 V (vs. Ag/AgCl) in 0.5 m H2SO4.

higher the loss of electrochemical activity, the greater the dissolution. Figure 1 b suggests that the rate of dissolution is high in the first few cycles ( 100 cycles) and then decreases in coelectrodeposited electrodes (Ru0.258Ti0.736Zn0.006Ox, Ru0.96Zn0.04Ox, and Ru0.33Ti0.67Ox). However, the rate of dissolution remains high throughout the cycles for the electrodeposited RuO2 electrode. On the basis of these observations, we hypothesize that the stable surface has a configuration and composition that increase the activation barrier for corrosion without having much impact on the activation energy of the water oxidation pathway. The stable bulk composition of the electrodes was quantified by X-ray fluorescence (XRF) spectroscopy after 500 CV cycles (Table S1). After 500 cycles, the ratio of Ru:Ti in the electrodeposited RuTi electrode is about 3:7. Increasing the Ti content in the co-electrodeposited RuTi electrode above this ratio results in stable but low-activity electrodes, as is evident in the very low activity of the Ru0.1Ti0.9Ox electrode. The electrodes with Ti-enriched surfaces are not active towards electrochemical water oxidation, but they are stable. The Ru:Zn ratio in the electrodeposited RuZn electrode is about 96:4, and the ternary RuTiZn electrode has an Ru:Ti:Zn ratio of about 25.8:73.6:0.6. The final compositions of the electrodes after 500 cycles indicate dissolution of Ru-rich and Zn-rich clusters (Table S1). However, the dissolution of (Zn,Ru)-rich unstable cluster contributes to increased porosity, which leads to changes in surface morphology after electrochemical cycling (Figure S5). The increase in double-layer capacitance of binary (RuTi, RuZn) and ternary electrodes (RuTiZn) in the non-faradaic region (1.0 V vs. Ag/AgCl) with increasing potential cycling also indicates an increase in accessible electrochemical surface area (ECSA; Figure S6). The increased ECSA indicates enhancement of porosity during later stages of cycling. The calculated electrochemical porosity of the electrodes increases with inChemElectroChem 2015, 2, 1839 – 1846

The crystalline structures of the fabricated electrodes on titanium supports were investigated by XRD (Figure 2). Ru0.258Ti0.736Zn0.006Ox shows the peaks corresponding to rutile phase of RuO2 along with peaks for the Ti substrate and hexagonal phase of Ru metal. This suggests that some of the Ru metal is not fully oxidized, and the XPS data (Figure 3) suggest that incompletely oxidized Ru is present in the subsurface region (see Section 2.3). The XRD data of undoped RuO2, Ru0.96Zn0.04Ox, and Ru0.33Ti0.67Ox match those of the rutile phase of RuO2. No peaks of Zn or its oxides are found for any of the Zn-containing electrodes (Ru0.258Ti0.736Zn0.006Ox and Ru0.96Zn0.04Ox), and this suggests that that Zn is substitutionally doped into these electrodes. There is no major shift of the peak for the RuO2 (110) plane at 27.28 in RuO2, Ru0.33Ti0.67Ox, Ru0.96Zn0.04Ox, and Ru0.258Ti0.736Zn0.006Ox. Hence, it can be inferred that Ru0.33Ti0.67Ox consists of a TiO2–RuO2 solid solution and Ru0.258Ti0.736Zn0.006Ox consists of a Zn-doped TiO2–RuO2 solid solution, both of which contain the rutile phase of RuO2. RuO2 (a = b = 4.52, c = 3.14 æ) and TiO2 (a = b = 4.67, c = 2.97 æ) have similar lattice structures (space group 136) and lattice con-

Figure 2. XRD patterns of Ru0.258Ti0.736Zn0.006Ox, Ru0.96Zn0.04Ox, Ru0.33Ti0.67Ox, and RuO2.



Articles

Figure 3. a) XP spectrum near Ti 2p peak positions (448–470 eV) of Ru0.258Ti0.736Zn0.006Ox, Ru0.96Zn0.04Ox, Ru0.33Ti0.67Ox, and RuO2. b) XPS spectrum near Ru 3d peak position (275–295 eV) of Ru0.258Ti0.736Zn0.006Ox, Ru0.96Zn0.04Ox, Ru0.33Ti0.67Ox, and RuO2.

stants, which result in XRD peaks in almost identical positions and makes it very difficult to identify possible segregated pure oxide phases of RuO2 and TiO2. 2.3. XPS Analysis of the Surface Composition of the Electrodes The surface elemental composition and chemical state of the components were analyzed by XPS (Figure 3 and Figure S7). The XP spectrum of RuO2 shows peaks at 280.4 and 284.5 eV, which are attributed to 3d5/2 and 3d3/2 of RuIV, respectively, consistent with previous reports.[19, 31, 35, 36] In Ru0.33Ti0.67Ox these peaks are broad and shifted towards higher binding energies with respect to Zn-doped samples (Figure 3 b). The broadened peaks have been ascribed to charge transfer between Ru and Ti (RuxTiyOz !Ruxd + Tiyd¢Oz), whereby the Ru 3d5/2 peak is shifted towards higher binding energy by 0.9 eV in comparison to RuO2 and thus signifies that Ru is in a higher oxidation state (> + 4). The existence of a charge-transfer mechanism indicates that a doped oxide has been formed rather than a mixed oxide. The extent of peak broadening and peak shift of the Ru 3d XP spectrum is much more pronounced in Ru0.33Ti0.67Ox than in Ru0.258Ti0.736Zn0.006Ox. In contrast, Ru 3d peaks of Ru0.96Zn0.04Ox are at almost the same position (280.6 and 284.7 eV) as those of Ru0.258Ti0.736Zn0.006Ox, although their intensities differ. The ternary electrode has the most intense Ru 3d3/2 peak, which signifies that a larger amount of electrocatalytically active RuIV species is present. Peaks corresponding to Ru 3d3/2 and Ru 3d5/2 are present at higher binding energies compared to undoped RuO2. The Ti 2p3/2 peaks at about 459.3 and 457.2 eV correspond to oxidation states of + 4 and + 3, respectively. Twin peaks at about 457 and 462 eV are observed for Ru0.33Ti0.67Ox, whereas a single intense peak at 459 eV is observed for ChemElectroChem 2015, 2, 1839 – 1846

Ru0.258Ti0.736Zn0.006Ox (Figure 3 a). A shift towards lower binding energy is attributed to accumulation of negative charge on the Ti atom. Such behavior has been observed in RuO2–TiO2 composite electrodes and has been ascribed to the charge-transfer mechanism.[19] However, the small peak shifts observed for Ru0.258Ti0.736Zn0.006Ox indicate that this type of charge-transfer mechanism is not a central feature in ternary RuTiZn electrocatalysts. The XPS depth-profile analysis of Ru0.258Ti0.736Zn0.006Ox and Ru0.96Zn0.04Ox electrodes revealed that very small amounts of < 0.1 and < 0.4 % Zn, respectively, are present on the surface (Table S3). From this we infer that most of the Zn is present in the subsurface region. Previous DFT studies indicated that subsurface Zn is responsible for lowering the activation energy of the OER.[32, 37] Furthermore, Ru0.258Ti0.736Zn0.006Ox and Ru0.33Ti0.67Ox have nearly equal amounts of Ru and Ti atoms on the surface. The increased stability of the electrocatalysts can be correlated with this stable surface composition. The dissimilar composition of the surface and bulk also confirm that co-electrodeposited electrodes are mixed homogenously. The depth-profile analysis of Ru0.258Ti0.736Zn0.006Ox and Ru0.33Ti0.67Ox also indicates that the Ru:Ti ratio is nearly same (1:1) up to 10 nm from the surface. The dissimilarity of surface/subsurface layer composition (from depth-profile analysis) and bulk composition (XRF analysis) indicates that Ru is electrodeposited in the surface or near-surface region, whereas Ti is deposited more in the bulk. We did not observe any peak corresponding to Ru metal (generally at 280 eV), which suggests that Ru metal is present only in the bulk of the electrode. The XRD and XPS studies confirm that the near-surface structure of the electrode consists of Zndoped RuO2–TiO2 for co-electrodeposited RuTiZn, Zn-doped RuO2 for co-electrodeposited RuZn, and RuO2–TiO2 solid solution for co-electrodeposited RuTi. 2.4. Surface Structure of the Electrodeposited Electrodes The SEM images of Ru0.258Ti0.736Zn0.006Ox, Ru0.96Zn0.04Ox, Ru0.33Ti0.67Ox, and RuO2 (Figure 4) suggest that they have different surface morphologies. Ru0.258Ti0.736Zn0.006Ox (Figure 4 a) has strawberry-like particles with increased surface roughness, whereas Ru0.96Zn0.04Ox (Figure 4 b) and Ru0.33Ti0.67Ox (Figure 4 c) are made of irregular-shaped particles. In contrast, RuO2 (Figure 4 d) has a thin-film morphology with cracks on the surface. Ru0.258Ti0.736Zn0.006Ox has a particle size of 150 nm with narrow particle size distribution, whereas Ru0.96Zn0.04Ox and Ru0.33Ti0.67Ox show broader size distributions. The SEM images further show that composite electrodes are more granular in nature, and addition of Ti decreases the particle size. To further quantify the surface macro-roughness of the electrodes, surface profiling was performed (Figure 5 and Figure S8). Ru0.258Ti0.736Zn0.006Ox shows higher surface roughness (1280 nm) than Ru0.96Zn0.04Ox (1070 nm), Ru0.33Ti0.67Ox (635 nm), and RuO2 (235 nm). In addition to the dissolution of excess Ru during co-electrodeposition, porosity and roughness of the mixed composite electrodes can also be enhanced by 1) hydrogen evolution during electrodeposition yielding surface cracks and 2) subsequent propagation of these cracks due to differential thermal



Articles

Figure 6. Tafel plots of Ru0.258Ti0.736Zn0.006Ox, Ru0.96Zn0.04Ox, Ru0.33Ti0.67Ox, and RuO2 for the fifth electrochemical cycle in 0.5 m H2SO4 at a scan rate of 5 mV s¢1.

Figure 4. SEM images of a) Ru0.258Ti0.736Zn0.006Ox, b) Ru0.96Zn0.04Ox, c) Ru0.33Ti0.67Ox, and d) RuO2.

Figure 5. 3D surface profiling of a) Ru0.258Ti0.736Zn0.006Ox, b) Ru0.96Zn0.04Ox, c) Ru0.33Ti0.67Ox, and d) RuO2.

expansion of the substrate and electrodeposited material, referred to as drying stress,[38] during heat treatment of the electrode. These surface cracks are desirable, since undercoordination along the crack edges results in higher conductivity. A smooth electrode surface often behaves like a two-dimensional surface, whereas a porous electrode with low diffusional resistance to the reactants and products behaves more like a three-dimensional electrode in which both the inner surface region and outer surface region are available for electrocatalytic activity. It has been suggested that the inner surface region results in higher electrochemical surface area for the OER.[38]

compensated resistance. At higher potential, the obtained current has contributions from both the OER and the dissolution reaction, which cannot be decoupled from each other. The lower the Tafel slope, the better the OER electrocatalytic activity. The Tafel slope is not only dependent on the chemical composition of the electrode, but also on exposed surface facets and surface coverage of the reactive species. Generally, two Tafel lines are observed for the OER. The lower Tafel slope corresponds to Langmuir–Hinshelwood (LH) bimolecular water oxidation when the electrodes have mixed oxide/hydroxyl coverage.[39] At higher potential, the reaction mechanism changes to the more favorable Eley–Rideal (ER) mechanism, whereby the electrode surface is covered by oxygen species. Introduction of additives (Ti and Zn) promotes surface roughness, which results in lowering of the Tafel slope. The Tafel slopes of the electrocatalysts at higher and lower current density are summarized in the Table 2. At lower current density, the Tafel slope of the ternary electrode (Ru0.258Ti0.736Zn0.006Ox) is 53 mV dec¢1, which is slightly lower than those of both Ru0.96Zn0.04Ox (58 mV dec¢1) and Ru0.33Ti0.67Ox (58 mV dec¢1). Electrodeposited RuO2 has the highest Tafel slope of about 59 mV dec¢1, which matches well with the previous studies.[40, 41] In this region, the Tafel slope indicates that deprotonation of the surface hydroxyl species (Surf¢OHads ! Surf¢Oads + H + + e¢) is the rate-determining step according to the reaction mechanism proposed by Bockris.[42] At higher current density, the Tafel slope of the ternary electrode is about 129 mV dec¢1, which is lower than those of Ru0.96Zn0.04Ox

Table 2. Tafel analysis of Ru0.258Ti0.736Zn0.006Ox, Ru0.96Zn0.04Ox, Ru0.34Ti0.67Ox, and RuO2 in 0.5 m H2SO4 solution.

2.5. Tafel Analysis of the Electrodes Figure 6 shows the Tafel plots of electrodes for the fifth CV cycle after taking into account the unChemElectroChem 2015, 2, 1839 – 1846

Electrode

Tafel slope [mV dec¢1] Lower Higher

Exchange-current density [mA cm¢2] current density < 5 mA cm¢2 current density > 5 mA cm¢2

Ru0.258Ti0.736Zn0.006Ox Ru0.96Zn0.04Ox Ru0.33Ti0.67Ox RuO2

53 58 58 59

4.7 5.9 3.3 1.4


129 139 143 150

503 440 359 56


Articles (139 mV dec¢1), Ru0.33Ti0.67Ox (143 mV dec¢1), and RuO2 (150 mV dec¢1). At higher potential, hydration of the surface becomes the rate-determining step (Surf + H2O!Surf¢OHads + H + + e¢). From the Tafel plot, exchange-current densities were calculated (Table 2). The exchange-current density gives a measure of the activation energy of the OER reaction through Equation (1):

ð EÞ

i0 ¼Fk0 exp ¢KTa

ð1Þ

where i0 is the exchange-current density of the OER catalyst, k0 is the rate constant, Ea is the activation energy of the OER in the given potential range, and F and K are the Faraday and Boltzmann constants, respectively. A high exchange-current density indicates high reaction rate, that is, low activation energy, if the electrochemical surface area is constant. Table 2 suggests that both Ti and Zn doping may activate the electrocatalysts by lowering the activation energy of the OER, which is reflected in enhanced exchange-current densities. To determine the effect of surface roughening due to dissolution in anodic electrochemical cycles, the Tafel slope was measured after 500 cycles (Figure S9). No significant change in Tafel slope was observed after 500 cycles for all the electrodes. However, the exchange-current density of all the electrocatalysts decreases in higher electrochemical cycles (Table S4). The exchange current density for the LH mechanism is in the order of RuO2 < Ru0.33Ti0.67Ox < Ru0.258Ti0.735Zn0.006Ox < Ru0.96Zn0.04Ox, and for the ER mechanism, it is in the order of RuO2 < Ru0.96Zn0.04Ox Ru0.33Ti0.67Ox < Ru0.258Ti0.735Zn0.006Ox. The decrease in exchange-current density can be ascribed to a reduced number of active sites and lowering of the dissolution kinetics in stable electrodes. 2.6. EIS Analysis of the Electrodes To understand the charge-transfer kinetics and double-layer capacitance of the electrodes, electrochemical impedance spectroscopy (EIS) of the electrodes was carried out at 1.2 V (vs. Ag/AgCl). The corresponding Nyquist plots are shown in Figure 7. In the Nyquist plot, two distinct semicircles are observed, whereby the first semicircle at higher frequency is relatively small compared to the second semicircle, which is observed at lower frequencies. The equivalent circuit (inset to Figure 7 a) comprises Rs[Rct(RfQf)CDL], where Rs is the resistance of the electrochemical cell, Rct is the charge-transfer resistance, CDL is the double-layer capacitance, Rf is the resistance of the surface film, and Qf is the constant-phase element of the film. The observed additional circuit in series with Rct may account for the rate-determining step of formation of a hydrous oxide layer over the electrode surface, which is also found in other metal-oxide electrodes.[17, 43, 44] The reactant diffuses through the porous passive outer layer and comes into contact with the active inner surface layer, and thereby the reaction rate increases. The CDL and Rct values of the electrodes are listed in Table 3. ChemElectroChem 2015, 2, 1839 – 1846

Figure 7. Nyquist plots of Ru0.258Ti0.736Zn0.006Ox, Ru0.96Zn0.04Ox, Ru0.33Ti0.67Ox, and RuO2 in 0.5 m H2SO4 at a potential of 1.2 V (Ag/AgCl). The symbols signify the experimental points, and the lines the fitted data. Inset a): Equivalent EIS circuit of electrodes in 0.5 m H2SO4 at a potential of 1.2 V (Ag/AgCl). Inset b): Frequency response in the higher-frequency regime of Ru0.258Ti0.736Zn0.006Ox, Ru0.96Zn0.04Ox, and Ru0.33Ti0.67Ox.

Table 3. Capacitance and resistance values estimated by fitting the EIS data to the corresponding equivalent circuit for Ru0.258Ti0.736Zn0.006Ox, Ru0.96Zn0.04Ox, Ru0.33Ti0.67Ox, and RuO2 at 1.2 V (vs. Ag/AgCl) in 0.5 m H2SO4 solution.

Electrodes

Rct [mW]

CDL [mF]

Rf [W]

Cf [mF]

Ru0.258Ti0.736Zn0.006Ox Ru0.96Zn0.04Ox Ru0.33Ti0.67Ox RuO2

43.2 113 456 1170

16.5 14.97 7.10 1.12

4.27 4.14 5.6 9.11

74.30 51.56 29.60 1.47

The CDL value is a measure of the activity of the electrocatalysts, since it is proportional to the ECSA. The CDL values follow the order RuO2 < Ru0.33Ti0.67Ox < Ru0.96Zn0.04Ox < Ru0.258Ti0.736Zn0.006Ox, which corresponds to the order of the current densities and thus corroborates the hypothesis that the increased activity is due to the increased ECSA. In addition, all co-electrodeposited electrodes showed lower Rct than the electrodeposited RuO2 electrode, consistent with the assumption that addition of Ti and Zn improves electrical conductivity along the electrode surface. The lowering of Rct improves the site activity of the electrocatalysts, and the best-performing ternary electrode has the lowest Rct. 2.7. Role of Ti and Zn in Increasing Activity and Stability in Water Oxidation Ti and Zn play twin roles of increasing activity and stability in co-electrodeposited Ru-based catalysts for OER. In this regard, the mechanisms by which Zn and Ti increase activity in the ternary RuO2 based electrocatalysts seem to be different. In the co-electrodeposited RuZn electrode, the presence of Zn atoms in the subsurface region causes distortion of the local surface



Articles structure of RuO2, which lowers the activation energy for electrocatalytic water oxidation, as suggested by DFT studies.[32] Tafel-slope analysis and the higher exchange-current density of the Zn-doped RuO2 catalyst in comparison to the RuO2 catalyst also support this hypothesis. The role of Ti atoms in binary electrodeposited RuTi electrodes is different, since their activation energy towards water oxidation is lowered through chemical inhomogeneity. In the co-electrodeposited RuTi electrode, Ti and Ru atoms interact with each other, and electron accumulation at Ti is increased while Ru becomes more positively charged, as suggested by XPS studies, and more positively charged Ru requires a lower activation energy for electrocatalytic water oxidation. In case of the best-performing ternary electrode, the activation mechanism seems to be different, and activation through the charge-transfer mechanism operating in the co-electrodeposited RuTi electrode is absent, since the XPS studies suggest that the binding energy of the Ti atom resembles that of TiIV. Incorporation of structural and chemical inhomogeneity also increases the roughness of the electrode. Previous studies indicate that M-doped (M = Cr, Ir, Mo, Mn, Co, Fe, Ni) TiO2 (110) surfaces have enhanced activity for OER compared to the pure TiO2 (110) surface.[37, 45] Our study shows that Zn-doped TiO2 has nine times higher current density than electrodeposited TiO2 (Figure S10). These results suggest that the surface Ti sites of the ternary electrode are no longer benign towards the OER, and they also provide active sites, as is corroborated by the maximum CDL (ECSA) found in co-electrodeposited ternary electrodes. The cumulative effect of increased number of active sites, lower activation energy, and porous and cracked structure of the electrode leads to improved OER performance. Under acidic conditions relevant to OER, TiO2 is passive and stable, whereas Zn is prone to corrosion. XPS and XRF studies indicate that the surface structure of the ternary electrode consists of equal amounts of Ru and Ti and is responsible for its low corrosion rate, and Zn is in the subsurface region. The hydrated outer surface layer also protects the inner oxide layer from corrosion. Although the mechanism of increasing the chemical stability by Zn doping is unclear, we believe that the stability of Zn-doped RuO2 catalyst might be due to changes in M¢O bond lengths in the surface, which may hinder the formation of reaction intermediates on the dissolution pathway.

3. Conclusions We have explored a design strategy for increasing the activity and stability of electrodes for the OER in acidic media by combining active (Ru) and passive (Ti) components, made less passive by a dopant (Zn). As the components are co-electrodeposited, the dissolution of unstable excess Ru leads to higher porosity and roughness within the electrode. The roughness is further increased by the chemical inhomogeneity of Ti in the RuO2 matrix and the structural inhomogeneity of Zn in the RuO2 matrix. The Ru0.258Ti0.735Zn0.006Ox ternary electrode was found to be most stable and has the best activity with a current density of 63 mA cm¢2 at 1.35 V (vs. Ag/AgCl) among the studied electrodes. This electrode requires 0.13 V less overpotential ChemElectroChem 2015, 2, 1839 – 1846

than the RuO2 electrode to reach 20 mA cm¢2. The increased activity is attributed to favorable electronic effects (i.e. decreased charge-transfer resistance and increased exchange-current density) and an increase in the electrochemical surface area. Zn doping lowers the activation energy of the OER, and the addition of Ti increases both the activity and the stability, while providing a greater number of active sites. Apart from these electronic factors and enhanced porosity, the cracked electrode morphology is also responsible for the enhancement of the OER, because the inner-surface region also becomes available for OER. As the enhanced activity was achieved without compromising stability by optimizing low-cost elements, the explored strategy is potentially viable for fabricating largearea electrodes on an industrial scale.

Experimental Section Chemicals All precursor salts were purchased from Sigma-Aldrich. The following chemicals were used as received as precursors of Zn, Ru, and Ti, respectively: ZnSO4 (99.9 % pure), RuCl3·x H2O) (99.9 % pure), and TiCl4 (99.9 % pure). Laboratory-grade H2SO4 (Fisher Scientific) was used as received. Boric acid powder (Fisher Scientific, ACS grade, purity 99.5 %) was dissolved in deionized (DI) water to make boric acid solutions. All aqueous solutions were prepared with DI water. The support, a titanium (99 % purity) plate of 1 mm thickness, was purchased from Alfa Aesar.

Electrode Preparation CV of precursor solutions was performed before electrodeposition to determine the deposition potential and showed that electrodeposition started at ¢0.3 V (vs. Ag/AgCl). To ensure the deposition of all elements, the optimum deposition potential was taken as ¢0.9 V, which is the most negative standard reduction potential of the three elements (i.e. Ti, Ru, and Zn).[46] At more negative potentials, increased H2 evolution from the electrode surface hampering the deposition process. Prior to electrodeposition, the titanium plate was cleaned with emery paper, sonicated for 20 min in acetone, treated with oxalic acid at 95 8C, and rinsed with water. The bath for electrodeposition was prepared by mixing the precursors in deionized water followed by the addition of 5 vol % H2SO4. This mixed solution was stirred at 70 8C for 1 h, after which 0.5 wt % boric acid was added and the mixture stirred for a further 30 min. The stirred, well-mixed solution was used for electrodeposition on the cleaned titanium plate in a three-electrode system with Ag/AgCl reference and Pt counter electrodes for 30 min. The compositions of precursor solutions were varied to get different composition of the electrocatalysts. The composition of the electrocatalysts was analyzed by XRF spectroscopy (Table 1). After electrodeposition, the electrode was dried in open atmosphere at room temperature for 24 h and then heated at 500 8C for 4 h. After heat treatment, the electrode was cooled to room temperature.

Material Characterization Powder XRD data were collected in the 2 q range 20–708 with CuKa (40 kV, 40 mA) radiation on a Siemens D5000 Bragg–Brentano q– 2 q diffractometer equipped with a diffracted-beam graphite mon-



Articles ochromator crystal, 2 mm (18) divergence and antiscatter slits, 0.6 mm receiving slit, and incident beam Soller slit. The compositions of the electrodes were determined through XRF spectroscopy (Rigaku ZSX Primus spectrometer). Surface morphology was determined by SEM with a Zeiss FESEM SUPRA 400VP Gemini instrument. Surface profiling of the fabricated electrode was carried out by using an optical profiling system (NanoMap-D, AepTechnology, USA). The surface elemental composition and chemical state of the components were analyzed by XPS with a PHI Versa Probe II scanning XPS microprobe.

Electrochemical Characterization All the electrochemical characterizations were performed in 0.5 m H2SO4 solution at room temperature. The working area of the electrodes was 1 cm2. All potentials reported were measured against Ag/AgCl (saturated KCl) as reference electrode and Pt mesh as counter electrode. CV and LSV of the electrodes were performed in a single-compartment cell by using a potentiostat (Autolab PGSTAT302N). The scan rate for CV measurements was 50 mV s¢1, and that for LSV measurements was 5 mV s¢1. The electrochemical characterizations were carried out in potential window of 0–1.4 V. EIS was performed with an Autolab/FRA instrument in the frequency range of 10 kHz to 0.05 Hz with an ac perturbation of 10 mV at a working-electrode potential of 1.2 V. The complex nonlinear least-squares fit of the experimental results to an appropriate equivalent circuit was performed with NOVA 1.9 software. Tafel analysis and EIS was performed after aging the electrodes for five electrochemical cycles in a potential window of 0.0–1.4 V in 0.5 m H2SO4 solution at a scan rate of 50 mV s¢1. The methodologies for computing electrochemical porosity and exchange-current density are provided in the Supporting Information. A detailed discussion of the rationale for choosing the EIS circuit is also provided.

Acknowledgements We thank the Indian Space Research Organization (ISRO) and the Technology Systems Development program of the Department of Science and Technology for supporting this work via Grant No. STC/CHE/20110043 and Grant No.DST/TSG/SH/2011/106, respectively. We also thank Dr. S. Illangovan and A. Senthil Kumar, Vikram Sarabhai Space Center, ISRO, and Dr. S. Ravichandran, Central Electro-chemical Research Institute, Karaikudi, for useful discussions during the course of this work. We are also thankful to Gyanprakash Maurya for the valuable help extended for EIS analysis. Keywords: doping · electrodeposition · oxygen evolution reaction · ruthenium · titanium [1] L. Trotochaud, S. W. Boettcher, Scr. Mater. 2014, 74, 25. [2] C. C. L. McCrory, S. Jung, J. C. Peters, T. F. Jaramillo, J. Am. Chem. Soc. 2013, 135, 16977 – 16987. [3] C. C. L. McCrory, S. Jung, I. M. Ferrer, S. Chatman, J. C. Peters, T. F. Jaramillo, J. Am. Chem. Soc. 2015, 137, 4347 – 4357. [4] J. R. Galn-Mascarûs, ChemElectroChem 2015, 2, 37. [5] I. C. Man, H.-Y. Su, F. Calle-Vallejo, H. A. Hansen, J. I. Martnez, N. G. Inoglu, J. Kitchin, T. F. Jaramillo, J. K. Norskov, J. Rossmeisl, ChemCatChem 2011, 3, 1159. [6] J. Rossmeisl, Z. W. Qu, H. Zhu, G. J. Kroes, J. K. Norskov, J. Electroanal. Chem. 2007, 607, 83.


[7] V. Baglio, A. D. Blasi, T. Denaro, V. Antonucci, A. S. Aricý, R. Ornelas, F. Matteucci, G. Alonso, L. Morales, G. Orozco, L. G. Arriaga, J. New Mater. Electrochem. Syst. 2008, 11, 105. [8] S. Trasatti, Electrochim. Acta 1984, 29, 1503. [9] Y. Lee, J. Suntivich, K. J. May, E. E. Perry, Y. Shao-Horn, J. Phys. Chem. Lett. 2012, 3, 399. [10] E. Tsuji, A. Imanishi, K.-i. Fukui, Y. Nakato, Electrochim. Acta 2011, 56, 2009. [11] T. Reier, M. Oezaslan, P. Strasser, ACS Catal. 2012, 2, 1765. [12] H. Ma, C. Liu, J. Liao, Y. Su, X. Xue, W. Xing, J. Mol. Catal. A 2006, 247, 7. [13] W. Hu, Y. Wang, X. Hu, Y. Zhou, S. Chen, J. Mater. Chem. 2012, 22, 6010. [14] L. G. Bloor, P. I. Molina-Sanchez, M. D. Symes, L. Cronin, J. Am. Chem. Soc. 2014, 136, 3304. [15] D. Seley, K. Ayers, B. A. Parkinson, ACS Comb. Sci. 2013, 15, 82. [16] R. Kçtz, H. J. Lewerenz, S. Stucki, J. Electrochem. Soc. 1983, 130, 825. [17] M. E. G. Lyons, S. Floquet, Phys. Chem. Chem. Phys. 2011, 13, 5314. [18] N. Hodnik, P. Jovanovic, A. Pavlisic, B. Jozinovic, M. Zorko, M. Bele, V. S. Selih, M. Sala, S. B. Hocevar, M. Gaberscek, J. Phys. Chem. C 2015, 119, 10140. [19] L.-A. Nslund, C. M. Sanchez-Sanchez, A. S. Ingason, J. Backstrom, E. Herrero, J. Rosen, S. Holmin, J. Phys. Chem. C 2013, 117, 6126. [20] X. Wu, J. Tayal, S. Basu, K. Scott, Int. J. Hydrogen Energy 2011, 36, 14796. [21] K. Kadakia, M. K. Datta, O. I. Velikokhatnyi, P. Jampani, S. K. Park, S. J. Chung, P. N. Kumta, J. Power Sources 2014, 245, 362. [22] A. T. Marshall, R. G. Haverkamp, Electrochim. Acta 2010, 55, 1978. [23] J. Aromaa, O. Forsen, Electrochim. Acta 2006, 51, 6104. [24] N. Danilovic, R. Subbaraman, K. C. Chang, S. H. Chang, Y. Kang, J. Snyder, A. P. Paulikas, D. Strmcnik, Y. T. Kim, D. Myers, Angew. Chem. 2014, 126, 14240. [25] A. T. Marshall, S. Sunde, M. Tsypkin, R. Tunold, Int. J. Hydrogen Energy 2007, 32, 2320. [26] M. T. Brumbach, T. M. Alam, R. H. Nilson, P. G. Kotula, B. B. McKenzie, R. G. Tissot, B. C. Bunker, Mater. Chem. Phys. 2010, 124, 359. [27] R. Forgie, G. Bugosh, K. C. Neyerlin, Z. Liu, P. Strasser, Electrochem. SolidState Lett. 2010, 13, B36. [28] A. J. Terezo, E. C. Pereira, Electrochim. Acta 1999, 44, 4507. [29] M. H. P. Santana, L. A. De Faria, Electrochim. Acta 2006, 51, 3578. [30] V. Petrykin, K. Macounov, M. Okube, S. Mukerjee, P. Krtil, Catal. Today 2013, 202, 63. [31] V. Petrykin, Z. Bastl, J. Franc, K. Macounova, M. Makarova, S. Mukerjee, N. Ramaswamy, I. Spirovova, P. Krtil, J. Phys. Chem. C 2009, 113, 21657. [32] V. Petrykin, K. Macounova, J. Franc, O. Shlyakhtin, M. Klementova, S. Mukerjee, P. Krtil, Chem. Mater. 2011, 23, 200. [33] E. W. McFarland, H. Metiu, Chem. Rev. 2013, 113, 4391. [34] M. Garca-Mota, A. Vojvodic, F. Abild-Pedersen, J. K. Norskov, J. Phys. Chem. C 2013, 117, 460. [35] S. Bhaskar, P. S. Dobal, S. B. Majumder, R. S. Katiyar, J. Appl. Phys. 2001, 89, 2987. [36] D. Rochefort, P. Dabo, D. Guay, P. M. A. Sherwood, Electrochim. Acta 2003, 48, 4245. [37] M. Garca-Mota, A. Vojvodic, H. Metiu, I. C. Man, H.-Y. Su, J. Rossmeisl, J. K. Nørskov, ChemCatChem 2011, 3, 1607. [38] A. R. Zeradjanin, A. A. Topalov, Q. Van Overmeere, S. Cherevko, X. Chen, E. Ventosa, W. Schuhmann, K. J. J. Mayrhofer, RSC Adv. 2014, 4, 9579. [39] Y.-H. Fang, Z.-P. Liu, J. Am. Chem. Soc. 2010, 132, 18214. [40] C. Iwakura, K. Hirao, H. Tamura, Electrochim. Acta 1977, 22, 329. [41] L. M. Da Silva, J. F. C. Boodts, L. A. De Faria, Electrochim. Acta 2001, 46, 1369. [42] J. M. Bockris, J. Chem. Phys. 1956, 24, 817. [43] M. E. G. Lyons, M. P. Brandon, J. Electroanal. Chem. 2010, 641, 119. [44] J.-M. Hu, J.-Q. Zhang, C.-N. Cao, Int. J. Hydrogen Energy. 2004, 29, 791. [45] D. M. Jang, I. H. Kwak, E. L. Kwon, C. S. Jung, H. S. Im, K. Park, J. Park, J. Phys. Chem. C 2015, 119, 1921. [46] A. J. Bard, L. R. Faulkner, Electrochemical Methods: Fundamentals and Applications, Vol. 2, 2nd ed., Wiley, New York, 1980, pp. 808 – 810.

Manuscript received: March 31, 2015 Revised: May 11, 2015 Final Article published: July 20, 2015

