Supporting information for: Trends in Activity and Dissolution on RuO 2 under Oxygen Evolution

Transcription

1 Supporting information for: Trends in Activity and Dissolution on RuO 2 under Oxygen Evolution Conditions: Particles versus Well-Defined Extended Surfaces Roy, Claudie a ; Rao, Reshma R; b Stoerzinger, Kelsey A. c ; Hwang, Jonathan c ; Rossmeisl, J.; d Chorkendorff, Ib; Shao-Horn, Yang b,c ; Stephens, Ifan E.L. a,e ; a Surface Physics and Catalysis, Department of Physics, Technical University of Denmark, DK-2800 Kgs. Lyngby, Denmark b Department of Mechanical Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, United States c Department of Materials Science and Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, United States d Department of Chemistry, University of Copenhagen, Copenhagen, Denmark e Department of Materials, Imperial College London, London, United Kingdom

2 Figure S1. Comparison of the Ru dissolution rate for Ru-based catalysts in acidic and alkaline electrolyte. The y axis represents the Ru dissolution normalized by hour per cm 2 of electrode. These values have been extracted from [a] Hodnik et al. for Ru metal, electrochemically prepared RuO 2 and thermally prepared RuO 2. Ru dissolution measurements performed in 0.1 M HClO 4 using ICP-MS and at maximum of 1.4 V RHE [b] Cherevko et al. for Ru metal, and thermally prepared RuO 2 in 0.1 M H 2 SO 4 and 0.05 M NaOH. Ru dissolution measured by ICP and potential scanning from 1.2 V RHE until reaching 5 ma/cm 2 [c] Chang et al. for polycrystalline Ru metal and oriented SrRuO 3 films in 0.1 M KOH. Ru dissolution measured using RRDE and scanning up to a maximum potential of ~1.485 V RHE for Ru and estimated to 1-10% for SrRuO 3. [d] Tamura et al. measured Ru dissolution for Ru metal at ~200 ma/cm 2, hydrated RuO 2 at ~50 ma/cm 2 and thermally prepared RuO 2 at ~100 ma/cm 2 in 0.5 M H 2 SO 4 using spectrophotometric methods. [e] Frydendal et al. for Ru dissolution from thermally prepared sputter deposited thin films in 0.05 M H 2 SO 4 using ICP and EQCM at a constant potential of 1.8 V RHE. [This work]. 2

3 1000 A (100) (001) (111) j (µa/cm 2 ox ) 100 (101) NPs (110)* 10 j (µa/cm 2 ox ) As prepared 1,46 1,48 1,50 1,52 1,54 1,56 1,58 1,60 B (110)* (001) (101) (111) 10 NPs 1000 After 1.6 V 1,46 1,48 1,50 1,52 1,54 1,56 1,58 1,60 C (110)* j (µa/cm 2 ox ) 100 (101) (001) (111) NPs 10 After 1.6 V 1,46 1,48 1,50 1,52 1,54 1,56 1,58 1,60 V vs RHE Figure S2. Tafel plot of OER activity in 0.05 M H 2 SO 4, measured by cyclic voltammetry at 20 mvs -1 (taking mean value of forward and backward cycle and ohmic drop corrected) on the as-prepared oriented thin films (A), after 2h at 1.6V RHE (B) and 4h at 1.6V RHE (C) 3

4 Figure S3. Overview of Ru and RuO 2 catalyst for the oxygen evolution reaction in acid electrolyte [a] Strasser and co-workers 1, Ru polycrystalline (PC) and nanoparticles (NP) in 0.1 M HClO 4 at 6 mv/s, 1600rpm [b] Mayrhofer and coworkers 2, Ru thin film in 0.1 M H 2 SO 4 at 10mV/s [c] Markovic and co-workers 3, RuO 2 (0001) single crystal in 0.1 M HClO 4 at 5mV/s, 1600 rpm [d] Chorkendorff and coworkers 4, sputter deposited RuO 2 thin film in 0.05 M H 2 SO 4 at 5 mv/s, 1600 rpm [e] Chorkendorff and co-workers 5, RuO 2 mass-selected particles in 0.05 M H 2 SO 4 at 20 mv/s, 1600 rpm and [This work] RuO2 oriented thin films and particles in 0.05 M H 2 SO 4 at 20 mv/s. 4

5 The slope of the polarization curves of the catalysts reported in Figure S3 are reported in Table S1. For the polarization curves showing 2 slopes, both values with the associated potential range are reported Samples Slope (ma/v) Ru thin film [b] 15.5 Ru PC [a] 30.6 Ru NP [a] 9.9 Ru (0001) [c] 29.2 ( V RHE ) 9.7 ( V RHE ) Sputter RuO 2 [d] 13.5 RuO 2 NP [e] 7.9 ( V RHE ) 17.5 ( V RHE ) RuO 2 thin film [b] 8.4 ( V RHE ) 12.9 ( V RHE ) RuO 2 (001) (This work) 9.5 RuO 2 (110) single crystal (This work) 10.2 RuO 2 (110) nanoparticles (This work) 12.4 ( V RHE ) 3.76 ( V RHE ) Table S1. Tafel slope of Ru and RuO 2 catalyst for the oxygen evolution in acid electrolyte. [a] Strasser and co-workers [16], Ru polycrystalline (PC) and nanoparticles (NP) in 0.1 M HClO 4 at 6 mv/s, 1600rpm [b] Mayrhofer and co-workers [17], Ru thin film in 0.1 M H 2 SO 4 at 10mV/s [c] Markovic and co-workers [4], Ru (0001) single crystal in 0.1 M HClO 4 at 5mV/s, 1600 rpm [d] Chorkendorff and co-workers [18], sputter deposited RuO 2 thin film in 0.05 M H 2 SO 4 at 5 mv/s, 1600 rpm [e] Chorkendorff and co-workers [19], RuO 2 mass-selected particles in 0.05 M H 2 SO 4 at 20 mv/s, 1600 rpm and [This work] RuO 2 oriented thin films and particles in 0.05 M H 2 SO 4 at 20 mv/s. 5

6 Figure S4. Typical cyclic voltammetry at 20 mvs -1 in 0.05 M H 2 SO 4 before (blue) and after stability test at 1.4V RHE (A), 1.5V RHE (B), 1.6V RHE (C) and 1.7 V RHE (D) 6

7 Figure S5. a) Tafel plots of RuO 2 particles in 0.05 M H 2 SO 4, measured by cyclic voltammetry at 10 mv/s based on 3 independent measurements (averaged forward and back and ohmic drop corrected) b) Potentiostatic measurements at 1.4, 1.5, 1.6 and 1.8 V RHE c) Current densities at 1.55 V RHE and 1.7 V RHE from the Tafel plots presented in a). d) Ru dissolution found from ICP-MS measurements, taken from the tests depicted in c). The electrolyte was sampled before and after the stability tests, using the first sample as a blank. The dotted horizontal line indicates a dissolution rate of zero. The Ru dissolution behavior of rutile RuO 2 particles was established. Figure S5a shows the initial Tafel plots of RuO 2 particles catalyst, in comparison to the Tafel plots recorded after 2 hours corrosion tests at 1.4, 1.5, 1.6 and 1.7 V RHE. Figure S5b shows the current density as a function of time, during the stability test. Figure S5b shows the current density recorded during these 2 hours testing; evidently, at all potentials except for 1.4V RHE, the current decreases over time. We attribute this decrease in activity to the deactivation of the surface and also bubbles formation that blocks active sites. There is a pronounced decrease in oxygen evolution current on 7

8 the Tafel plot taken after the stability test, relative to the Tafel plot taken before the stability test, as shown in Figure S5c. The electrolyte was sampled before and after the stability tests presented on Figure S5b, and the Ru was quantified using ICP-MS. The normalized Ru amount ( / ) is presented as function of the applied potential on Figure S5d. Even at potentials as low as 1.4V RHE, a small amount of Ru is detected. The Ru dissolved increases with the applied potential and seems to follow an exponential trend. 8

9 Figure S6. Potentiostatic measurements at 1.6 V RHE a) and 1.7 V RHE b) in 0.05 M H 2 SO 4 and the respective amount of Ru dissolved during these electrochemical testing for the particles, and (111), (101) and (001) surfaces c). Corrosion measurements on preferentially oriented thin films and RuO 2 particles were performed at 1.6 and 1.7 V RHE. Figure S6a and Figure S6b present the current generated from RuO 2 particles, (001), (101) and (111) RuO 2 surfaces at 1.6 and 1.7 V RHE respectively. Whereas at at 1.6 V RHE, the current gradually declines over two hours (Figure S6a) for all samples, at 1.7 V, on the (001) and (111) samples the current density fluctuates significantly. Fig. S6c shows that the amount of Ru dissolved increases in the following order particles < (111) < (101) < (001). The increase in Ru dissolution from 1.6 to 1.7 V RHE is modest on all surfaces, except (001) film, where there is a 2.5 fold enhancement. 9

10 Figure S7. Physical characterization of the commercially available rutile RuO 2 particles (Sigma-Aldrich), including transmission electron microscopy (a-b) and X-ray diffraction (c) 10

11 Figure S8. Relationship between amount of Ru dissolved and activity measured by cyclic voltammetry (a-d) and chronoamperometry (e-f), in 0.05 M H 2 SO 4. The amount of Ru dissolved was taken from stability test 1 (a,c,e) and stability test 2 (b,d,f) at 1.6 V RHE over 2 hours. The current density was taken from the cyclic voltammetry recorded on the as-prepared sample, immediately prior to stability test 1 (a), after 2h stability test at 1.6V RHE, immediately prior to stability test 2 (b-c), and after stability test 2 (d). 11

12 j (µa/cm 2 RuO2 ) j (µa/cm 2 RuO2 ) (110) -40 (001) 0,55 0,60 0,65 0,70 0,75 0,55 0,60 0,65 0,70 0, A after stability test as prepared C after stability test as prepared -40 (101) -40 (111) 0,55 0,60 0,65 0,70 0,75 0,55 0,60 0,65 0,70 0,75 V vs RHE V vs RHE Figure S9. Typical cyclic voltammogram at 100 mvs -1 in 0.05 M H 2 SO 4 before and after stability test at 1.6V RHE for 2 hours for the (110) (A), (001) (B), (101) (C) and (111) (D) for the oriented thin films D B as prepared after stability test after stability test as prepared 12

13 j (µa/cm 2 RuO2 ) A after stability test as prepared (110) 0,55 0,60 0,65 0,70 0, B as prepared after stability test (001) 0,55 0,60 0,65 0,70 0,75 j (µa/cm 2 RuO2 ) 20 0 C as prepared (101) -40 (111) 0,55 0,60 0,65 0,70 0,75 0,55 0,60 0,65 0,70 0,75 V vs RHE V vs RHE Figure S10. Typical cyclic voltammogram at 100 mvs -1 in 0.05 M H 2 SO 4 before and after stability test at 1.6V RHE for 4 hours the (110) (A), (001) (B), (101) (C) and (111) (D) for the oriented thin films after stability test 40 D 20 0 after stability test as prepared 13

14 0.7V (µa/cm 2 geo ) Initial After 1.6V After 1.6V 0 (001) (101) (111) (110) Figure S11. Average current at 0.7 V RHE taken from the cyclic voltammetry at 100 mvs -1 of the as prepared, after 2 hours and 4 hours at 1.6V RHE for the (001), (101), (111) and (110) surfaces. 14

15 Figure S12. Two-hour potentiostatic measurements at 1.6 V RHE in 0.05 M H 2 SO 4 from stability test 1 (left) and stability test 2 (right) for the RuO 2 (111), (101), (110) and (001) surfaces. 15

16 Figure S13. Tafel plot of OER activity in 0.05 M HClO 4, measured by cyclic voltammetry at 10 mv/s (averaged forward and back, line) and potentiostatic measurements (constant applied voltage, points) 16

17 Figure S14: Schematics of the stoichiometric (100) and (001) surfaces. For the (100) facet, the surface Ru atoms are five-fold coordinated, whereas for the (001) facet, the surface Ru atoms are four-fold coordinated. Blue and golden atoms represent Ru and O atoms respectively. 17

18 Experimental Section Samples preparation Rutile RuO 2 films of the (100), (001), (101), and (111) orientations were grown by pulsed laser deposition (PLD), as reported previously 10,41,42. The oriented thin films were deposited using pulsed laser deposition (PLD) on single-crystal (001)-oriented SrTiO3 or (101), (001) or (111) TiO2 from a polycrystalline ceramic RuO2 target. The SrTiO3 substrates were conditioned with NH4F-buffered HF and annealed at 950 in O2 atmosphere for 1 hour to obtain an atomically flat surface. For surface quality recovery, the substrates were annealed at growth temperature for 30 min. PLD was performed using a KrF excimer laser (λ = 248 nm) to obtain RuO2 oriented films of ~25 nm thickness. The deposition took place at an oxygen pressure of 50 mtorr and cooled to room temperature under 200 Torr oxygen. The XRD patterns of the oriented surfaces are presented below: Figure S15: Normal 2θ-ω scan of RuO 2 films of noted orientation, where the (100) is grown on SrTiO 3, while all others grown on TiO 2 of corresponding orientation. Intensity is on a logarithmic scale. 18

19 The nanoparticles were purchased commercially (Sigma-Aldrich, 99.9% trace metal base). The (100) film was grown on a (001)-oriented SrTiO 3 substrate and the (001), (101) and (111) films were grown on (001), (101) and (111) oriented TiO 2 substrates. For the (110) orientation, a single crystal electrode was used, which was prepared by oxidative evaporation/redeposition of RuO 2 powders as described in reference 45,50. The particle-based film electrodes were prepared by ultrasonic dispersing RuO 2 powder (99.9% trace metal basis, Sigma-Aldrich) into tetrahydrofuran ( 99.9%, Sigma-Aldrich). A small amount of alkaline Nafion solution (3.3 % wt) was added to the ink to improve the adhesion to the fluorine doped tin oxide substrate. The ink was drop-casted and the film was dried over night. Electrochemical measurements All electrochemical measurements were conducted in a homemade threecompartment glass cell. The potentiostat used was a Bio-logic, WMP-2. A gold mesh was used as counter electrode, while an Hg/HgSO 4 electrode was used as reference electrode, calibrated to the reversible hydrogen electrode (RHE) in the same electrolyte. All measurements are reported vs RHE, as denoted by V RHE. A magnetic stirrer was used to prevent bubbles formation at the surface of the electrode as much as possible. The working electrode and counter electrode compartments were separated by a Nafion 117 membrane. A schematic of the electrochemical setup used for the corrosion measurements is presented below, where the counter-electrode, the Nafion membrane, the magnetic stirrer, the working electrode and reference electrode are identified. The counter-electrode, working electrode and reference electrode were aligned to minimize the ohmic drop. The working electrode was placed in a way so it faced the counter and reference electrode. This step ensure that the potential applied was as even as possible at the surface of the catalyst. The magnetic stirrer was placed just below the surface of the catalyst to remove as many bubbles as possible. 19

20 Figure S16. Schematic of the electrochemical setup used for the stability tests. The electrodes were mounted and the open-circuit potential was monitored for few seconds. The Ohmic drop measured using electrochemical impedance spectroscopy typically ranged from Ω for the RuO 2 powder, Ω for the oriented RuO x thin films and ca. 40 Ω for the (110) single crystal. The Ohmic drop compensation was performed by online Ohmic drop correction, where 85% correction was applied M H 2 SO 4 was prepared from ultrapure water (Milli-Q, 18 MΩ-cm) with sulfuric acid (99.999%, Sigma-Aldrich). The activity of the RuO 2 particles was first evaluated by using cyclic voltammetry (CV) at 10 mv/s between 1 to 1.7 V RHE. The OER activity measured from CV at 10 mv/s is comparable to that measured from galvanostatic measurements 10,41,42. Thus, we used CV to quantify OER activity in this study. A single stability test was then conducted at a given potential for 2 hours. The activity was again evaluated using one CV at 10 mv/s between 1 and 1.7 V RHE. The activity and stability of the oriented thin films and single crystal were evaluated in a similar manner. First, the activity was recorded using CVs at 10 mv/s between 1 20

21 and 1.6 V RHE. A stability test was then performed at a constant potential of 1.6V RHE for 2 hours. Finally, CVs at 10 mv/s between 1 and 1.6 V RHE were recorded. Optimal microscope images recorded after 2h at 1.6V RHE confirms that the samples did not delaminated (Figure S17). For the CVs, the O 2 evolution activity was measured by averaging the current at any given IR corrected potential over the forward and backward scans. Figure S17. Optical microscopy images of the a) (111), b) (101), c) (100) and d) (001) oriented thin films recorded after two-hour stability test at 1.6VRHE in 0.05 MH 2 SO 4. With the exception of the (110) single crystal, 3 independent samples were tested for each measurement. The currents were normalised to the microscopic surface area of the catalyst, which was obtained from BET measurements for the nanoparticles and the geometric area for the oriented thin films and (110) single crystal. The BET measurements were performed using a Quantachrome ChemBET Pulsar from a single-point BET analysis performed after 12h outgassing at 150 o C. 21

22 The surface area measurement were performed by measuring N 2 adsorption/desorption. Because of the small geometric area of the oriented thin films and (110) single crystal, i.e. typically between 0.05 to 0.1 cm 2, and the relatively large volume of the electrolyte in the working compartment (10 ml), we were unable to accurately measure the Ru concentration for stability tests below 1.6 V RHE. On the other hand, in the case of the oriented thin films, we found that if we subjected them to potentials positive of 1.7 V RHE, we obtained poor reproducibility, accompanied by visible signs of delamination from the underlying substrate. Presumably, these conditions led to the exposure of the backing substrates, constituting SrTiO 3 and TiO 2, and their consequent corrosion. Such substrate effects are undesirable 46. Although we tested the (100) oriented films, the high instability of these samples prevented us from measuring an accurate corrosion rate. For this reason, we only provide the initial activity herein (Figure S1a). Finally, for all samples the capacitance was measured using cyclic voltammetry at 100 mvs -1 between 0.55 and 0.75 V RHE, before and after the stability measurements. ICP-MS measurements The electrolyte was sampled three times for each sample: (i) prior to the immersion of the electrode, to ensure that the cell was not contaminated (ii) following the immersion and after recording the initial CV, to provide a baseline for the subsequent stability test (iii) after the stability test. Typically there was no difference between samples i and ii. The concentration of Ru was detected using inductively coupled 22

23 plasma mass spectrometry (ICP-MS, Fischer Scientific, model icap-qc ICP-M) and quantified using a calibration curve (see below). Figure S18. Typical calibration curve used for measuring the Ru concentration dissolved in solution, using ICP-MS. To ensure that no Ru crossed or adsorbed to the Nafion membrane, tests prior to measurement were performed, showing negative results. The amount of Ru measured was then normalized by the electrochemical surface area (ECSA) of the electrode. For all experiments, the concentration lied between 0.6 and 9.7 ppb. However, for the samples involving the particulate Ru, in around 10% of experiments, the signal intensity corresponded to a concentration 2-3 orders of magnitude higher than this value. We attributed this phenomenon to the presence of solid Ru nanoparticles being sampled by the instrument. The reliable quantification of Ru concentration using ICP-MS is only possible for dissolved species. Presumably, the particles had detached from the electrode as a result of corrosion or 23

24 bubble evolution. We performed 2-3 samples of each electrolyte involving particulate Ru; we eliminated the samples with anomalously high values from our analysis. Calculation of Faradaic efficiency towards dissolution Measurement of the Faradaic Efficiency based on the 2h chronoamperometric measurements: 1) Measurement of the total charge passed, Q total, by the integration of the current as function of time. An example for the (001) oriented thin film is presented below for the first stability test at 1.6V RHE Area=14766,62764 FWHM=0,15549 (001) Oriented thin film Stability Test 1.6 V RHE j (µa/cm 2 RuO2 ) Time (min) 2) Measurement of the charge related to the dissolution of Ru. The corrosion mechanism for RuO 2 was assumed to follow the following reaction mechanisms: RuO +2H O RuO aq + 4H +4e 24

25 Where 4 electrons are exchanged per mole of RuO 2 dissolved. As we measured the amount of Ru dissolved using ICP-MS, we can thus go back to the charge that it represents, using the following equation: Q = m F z M Where m RuO2 is the mass dissolved, F the Faraday constant, z the number of electron exchanged (4 e ) and M the molar mass of RuO 2. As ICP-MS only allows us to detect Ru, the mass of Ru has to be converted into RuO 2 by using the ratio: m = M M m The Faradaic Efficiency towards dissolution is then obtained: FE % = 100% 25