Methodology for Investigating Electrochemical Gas Evolution Reactions: Floating Electrode as a Means for Effective Gas Bubble Removal

The future significance of energy conversion has stimulated intense investigation of various electrocatalytic materials. Hence electrocatalysts have become the subject of electrochemical characterization on a daily basis. In certain cases of interest, when measuring electrochemical reactions beyond the onset potentials, however, appropriateness of existing electroanalytical methods may be questioned and alternative approaches need to be developed. The present study highlights some shortcomings in the electrochemical investigation of gas evolving reactions. The oxygen evolution reaction (OER) is selected as a case example with a specific focus on the electrochemical stability of a nanoparticulate iridium catalyst. When conventional electrochemical methods, such as thin film rotating disc electrodes are employed to study the materials’ stability, the intrinsic degradation is masked by oxygen bubbles, which are inherently being formed during the reaction, especially when high current densities are used. In this Letter, we present a solution to this issue, the so-called floating electrode arrangement. Its elegant usage enables fast and reliable electrochemical characterization of oxygen evolution electrocatalysts.

S-3 A vacuum pump from producer Ted Pela was used in film preparation as well as during electrochemical experiments. An easy-affordable vacuum pump (Ted Pella, PELCO® 520-220) that was used to induce vacuum suction is conventionally used to manipulate with TEM grids. Hence its vacuum suction strength cannot be determined. The pump and valve characteristics are given bellow:  The vacuum system is powered by a 220V, (50 Hz) cycle low wattage vibrator-type pump. The in-line filter is placed adjacent to the pump with the four foot plastic hose following.  Vacuum suction level: In order to regulate vacuum suction a special micrometer valve (Stainless Steel Integral Bonnet Needle Valve,0.37 Cv,1/4 in. Swagelok Tube Fitting,Regulating Stem,Part No. SS-1RS4) was used to manipulate with vacuum strength during i) electrochemical experiment and ii) during deposition of the catalyst suspension. However, we note that manipulation with a vacuum can only be performed in a relative manner. This means that if the valve is entirely opened the level is labelled as 1. If the valve is opened only till half the level is labeled as 0.5 etc.

S-4
 Tubings: In the case of vacuum suction used to prepare the catalyst layer the following tubing configuration was used (see Fig. S

S-3 Scanning electron microscopy characterization
Catalyst deposits on TEM grids were investigated using a scanning electron microscope (FE-SEM Zeiss SUPRA 35VP) with an accelerating voltage of 7 kV.
Catalyst films on the TEM grid were prepared either by dropcasting an aqueous suspension of the catalyst on the TEM grid followed by drying under i) ambient conditions (Fig.S1 a,b) or by dropcasting the aqueous suspension of the catalyst on the TEM grid and ii) subsequent drying by using a vacuum suction ( Fig.S1 c,d). The vacuum suction causes fast evaporation of water from the catalyst suspension.
Comparison of the two preparation modes i) and ii) reveals the beneficiary effect of the vacuum preparation method as the catalyst film coverage is more uniform in the latter case.
Afterwards, the dry residue was thermally treated in a 5% H 2 /Ar mixture. The temperature was increased to 450 °C with a rate of 2 °C min -1 , and then cooled to room temperature with a rate of 3 °C min -1 .
X-ray powder diffraction pattern for the analysis was collected at room temperature on a laboratory PANalytical X'Pert PRO diffractometer using CuKa radiation (1.54060 Å). Sample was loaded into a flat disc-like sample holder. The XRD data were collected in the 2θ range from 30 to 60° in steps of 0.04° 2θ with a total measurement time of 12.5 min. Phase identification was performed in the X'Pert  S-6

S-5 Transmission electron microscopy characterization
For the detailed microstructural investigation, Cs probe corrected Scanning transmission electron microscope (Jeol ARM 200 CF) with attached Jeol Centurio EDXS system with 100mm 2 SDD detector and Gatan Quantum ER DualEELS system was used.
Wire-like particle morphologies of iridium based electrocatalyst are observed in the TEM imaging. This reflects the agglomeration of primary particles during heat treatment.

S-6 Feasibility of the Au TEM grid as the working electrode
Prior to electrochemical testing of electrocatalyst electrochemical performance of bare Au TEM grid was inspected. Characteristic features of Au were followed via cyclovoltammtery by gradually increasing the upper potential limit (UPL). From the results (Fig.S3) it is clearly seen that characteristic redox features are obtained.

V vs RHE
The feasibility of the floating electrode to efficiently remove the oxygen bubbles was tested during a potentiostatic (2.0 V vs RHE) treatment. Two modes of operation were compared. In the first case vacuum suction was present during potentiostatic treatment (blue curve) and in the second case vacuum suction was absent (black curve). Even though potentials of this range are rather severe for simulating degradation of the catalyst in our study they are used in order to maximize the evolution of bubbles.
Chronoamperometric response normalized per mass of iridium is shown in Fig. S6. The performance of the two electrode configuration (vacuum and non-vacuum-ambiental) were compared (Fig. S6).
Chronoamperometric performance (mass current densities) in the case of the vacuum case is surpassing the non-vacuum case. This can be assigned to the use of vacuum. This has triggered the appearance of two effects: i) higher utilization of the catalyst layer hence a larger effective surface area in vacuum analogue in comparison to non-vacuum case. ii) Additionally, given the fact that substantially different shape of chronoamperometric response is detected for the vacuum analogue the higher current densities of the vacuum analogue electrode should also be ascribed to more efficient removal of oxygen bubbles which leads to more exposed surface area. It is noted here that in the case of non-vacuum-ambiental experiment a complete bubble passivation of the electrode occurred hence the experiment was stopped after approximately 800 s in this case. This is why only one dash curve is shown in Fig. S6d as one cannot refer to decaj in performance in a proper sense. More specific, if such TEM grid (which is passivated with bubbles) is de-passivated mechanically (for example by removing from the electrode S-9 compartment) and the its LSV response is re-analyzed almost the same performance as prior potentiostatic treatment is obtained (data not shown).

S-8 Performance of the floating electrode vs RDE configuration:
The electrochemical performances of the floating electrode (vacuum case analogue) and RDE configuration were compared. Potentiostatic treatment in the case of the floating electrode was performed at 2.0 V and at 1.8 V (both for 2000 s) in the case of RDE. Before and after potentiostatic treatment LSV protocol was performed in order to estimate OER activity (Fig.S5). Interestingly, retention of OER activity after potentiostatic treatment is lower in the case of RDE configuration despite being treated at a lower potential (1.8 V). This trend is in line with the fact that oxygen bubbles passivate the electrode surface in the RDE configuration.
Hence the effective catalyst surface is lower in this case. However, it has to be noted that further optimization of the floating electrode architecture is necessary in order to maximize the removal of the oxygen bubbles.

S-9 Indirect confirmation of oxygen bubble effect in RDE configuration:
In order to indirectly confirm that oxygen bubbles are being evolved in the course of OER potentiostatic treatment, the following experiment was performed in RDE configuration. The catalayts was treated under three modes of potentiostatic treatment: i) Initially potential of 1.8 V was used to perform OER after which ii) second potentiostatic treatment was performed under conditions of the oxygen reduction reaction (5 min at 0.3 V vs. RHE (Fig. S6). The subsequent potential hold is not shown. The purpose of ii) was to electrochemically reduce the oxygen bubbles formed during treatment i). After treatment ii) potential was set to 1.8 V again (treatment iii) in order to compare the corresponding current response to the one of i). Alltogether, treatment i) lasted 35 min. The chronoamperometric responses of i) and iii) are shown in Fig. S6. These results clearly indicate that oxygen bubbles are formed during potentiostatic treatment at 1.8 V (Fig. S6b-