In situ demonstration of anodic interface degradation during water electrolysis: Corrosion and passivation

Journal Pre-proof

In situ demonstration of anodic interface degradation during water electrolysis: Corrosion and passivation Hansaem Jang , Sunki Chung , Jaeyoung Lee PII: DOI: Reference:

S0013-4686(20)31669-8 https://doi.org/10.1016/j.electacta.2020.137276 EA 137276

To appear in:

Electrochimica Acta

Received date: Revised date: Accepted date:

28 July 2020 21 September 2020 8 October 2020

Please cite this article as: Hansaem Jang , Sunki Chung , Jaeyoung Lee , In situ demonstration of anodic interface degradation during water electrolysis: Corrosion and passivation, Electrochimica Acta (2020), doi: https://doi.org/10.1016/j.electacta.2020.137276

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2020 Elsevier Ltd. All rights reserved.

Highlights 

Interfacial degradations on anodes are observed under reacting conditions.



Structures idealized for the in situ detection are devised and fabricated.



Corrosion is observed using surface enhanced infrared absorption spectroscopy.



Passivation is monitored through controlled polarization tests.

In situ demonstration of anodic interface degradation during water electrolysis: Corrosion and passivation

Hansaem Jang a, Sunki Chung a, Jaeyoung Lee a, b, * a

Electrochemical Reaction & Technology Laboratory (ERTL), School of Earth Sciences and

Environmental Engineering, Gwangju Institute of Science and Technology (GIST), Gwangju 61005, South Korea. b

Ertl Center for Electrochemistry and Catalysis, Research Institute for Solar and Sustainable

Energies, GIST, Gwangju 61005, South Korea. *

E-mail: [email protected]

Abstract Understanding of interface-driven performance degradation in the oxygen evolution reaction (OER) is a prerequisite for developing long-term electrolytic systems operating at maximum efficiency. The question remains as to whether in situ observation of degradation during OER will be possible. We design a model system and fabricate a structure that is ideal for in situ detection of interfacial degradation that could occur at the anode, i.e. corrosion and passivation. The synthesized structure is based on carbon (C) or non-carbon (Ti); this is not only because it is a practical and general choice as an anode substrate, but also because it can represent the degradation reaction of corrosion or passivation, respectively. Interfacial responses from the ideally structured surface are observed (spectro)electrochemically under OER-operating conditions. We reveal that while electrochemical oxidation of C is inevitable over time, the passivation of Ti can be circumvented under controlled conditions. We also disclose that the use of Ti can be more beneficial for electrolytic cells, than that of C. These results highlight future research strategies required for the optimization of C- and Ti-based components used in electrolytic cells.

Keywords Water splitting; Oxygen evolution reaction; Infrared spectroscopy; Corrosion; Passivation

1. Introduction In the search for practical operation of electrolyzers and fuel cells, the designing and fabrication of effective system components used within them must be achieved. The generating unit, in which power or target chemicals are produced through electrocatalysis, consists of anode, cathode, and other supporting and auxiliary components [1-4]. As part of improving the efficiency of a generating unit, catalysts with increased number of active sites or enhanced intrinsic activity can be developed [1, 5-9]. At the same time, catalysts developed for prolonged catalytic activity must be thermodynamically stable within the electrolyte and electrochemically stable while functioning as electrocatalyst [10-12]. Three-electrode system testing in liquid electrolyte is an empirical method to examine the activity and stability of an electrocatalyst towards its target reaction. However, the excellent electrocatalytic performance of the half-cell does not always guarantee accordingly excellent performance in full-cell tests conducted using assembled cells made of solid electrolytes. Performance discrepancies can be due to differences in configuration between them. When making assembled cells using a polymer electrolyte membrane, it is usually necessary to use additional components that are not essential for half-cells [1, 3, 4, 13, 14]. Examples are flow field plates (e.g. single-sided plates, bipolar plates), backing layers, and porous transport layers (e.g. gas diffusion layer) as shown in Supplementary data §1. The introduction of these components creates additional interfaces that can influence the electrochemical reactions that take place inside the cell. Typically, the additional components are made of C- or Ti-based materials. This is because in general these materials can exist stably over a wide pH range and can be considered as electrical conductors [11, 12, 15, 16]. However, when anodic potential is

applied, these materials also participate in electrochemical reactions, which can degrade the surface. Carbon can be corroded through carbon oxidation reactions (COR) in the potential regime required for the oxygen evolution reaction (OER). Corrosion is the degradation of the material and is a major cause of instability [17, 18]; therefore, the control of the degradation rate is at the heart of the electrochemical system for long-term operation [19-21]. In the case of Ti, a passivation layer can grow on top of it, thus reducing the electronic conductivity [11]. These parasitic reactions can degrade interfaces, so these reactions are a reasonable cause of interface-driven degradation. A great deal of effort has been made by the relevant research groups to understand the oxidation of C and Ti under electrolytic conditions. Advanced characterization techniques

such

as

differential

electrochemical

mass

spectroscopy

(DEMS),

electrochemical impedance spectroscopy (EIS), surface enhanced Raman spectroscopy (SERS) and rotating ring-disk electrode (RRDE) have been thus developed to enable the in situ observation of surface and near-surface reactions [22-32]. These methods provide a posteriori knowledge of the electrochemical reactions on the electrode surface. Meanwhile, operando surface enhanced infrared absorption spectroscopy (SEIRA) can provide direct information on the species adsorbed on the electrode surface [33-36]. In this study, we aim to fabricate a model catalyst–support structure that is ideal for operando electrochemical observation of carbon and non-carbon surfaces under harsh oxidative conditions in acidic media [37]. We also devise a model system in order to rule out the catalyst’s contribution to performance degradation. In this regard, Ir-based materials known to have an optimal stability–activity trade-off relationship are used as electrocatalysts [38-41].

2. Experimental 2.1. Electrochemical carbon corrosion Iridium-embedded carbon nanofibers (Ir–CNF) were synthesized through a multi-stage heat-treatment process of electrospun products. First, a mixture for electrospinning was prepared by mixing the following chemicals for 12 h at 40 °C: 0.3 g Ir(III) acetylacetonate (Aldrich, 97 %), 2.0 g polyacrylonitrile (Aldrich, average MW 150,000), and 18 g dimethylformamide (JUNSEI, ≥ 99.5 %). The mixed solution was injected using a quintuple nozzle at a flow rate of 5 mL h−1. In the interim, the applied voltage was maintained at 24 kV. Subsequently, we acquired an electrospun web. The web was stabilized in air for 1 h at 280 °C and the ramping rate was set at 1 K min−1. The stabilized web was carbonized in N2 for 1 h at 1200 °C and the ramping rate was 5 K min−1. Additional heat-treatment processes were performed under Ar at various heattreatment temperatures (HTT), including 1500, 1800, 2100 and 2400 °C. The ramping rate was 10 K min−1 and the dwell time at the target HTT was 1 h. The physicochemical properties of the fabricated structures were characterized using the following means:

X-ray photoelectron

spectroscopy (XPS), Raman spectroscopy, X-ray diffraction (XRD), transmission electron microscopy (TEM), selected area electron diffraction (SAED), scanning transmission electron microscopy (STEM), STEM-coupled energy dispersive spectrometry (STEM–EDS), surface area analysis, scanning electron microscopy (SEM), SEM-coupled energy dispersive spectrometry (SEM–EDS), and thermogravimetric analysis (TGA). Detailed equipment specifications and further discussion can be found in Supplementary data §2. Electrochemical and electrocatalytic responses were identified using electroanalytical techniques under OER-operating conditions. The following procedure was used: First, the ink was prepared by mixing Ir–CNF with a mixture of 1200 μL deionized water, 800 μL isopropyl

alcohol (IPA; JUNSEI, ≥ 99.5 %), and 10 μL Nafion solution (Aldrich, 10 wt% in H2O). Then, the mixed solution was dispersed using a bath sonicator (8510, Branson) and a sonic dismembrator (Model 100, Fisher Scientific). Thereafter, the ink was loaded onto the working electrode and the amount of the loaded Ir–CNF was 100 μg cm−2. After drying, the loaded electrode was connected to a rotator (Pine Research Instrumentation) of a potentiostat (VSP, BioLogic). During the chronopotentiometry (CP) test, the rotator was maintained at 1600 rpm. Pt coil and Ag/AgCl (3 M KCl) were used as counter electrode and reference electrode, respectively. The electrolyte was 0.1 M HClO4 (Fluka), bubbled with N2 at 30 sccm before and during testing. Prior to testing, the working electrode was conditioned using cyclic voltammetry (CV) under the following conditions: 20 cycles of CV at 100 mV s−1 and 3 cycles of CV at 20 mV s−1 at 25 °C in a potential window between 0.05 and 0.9 V vs. RHE). The CP testing was performed at 1 mA cm−2 (Supplementary data §3). A SEIRA experiment was performed on the surface of a Si prism coated with Au. The Si prism was polished using alumina solution and then etched using NH4F. The Au coating was formed through electroless deposition. The precursor solution was prepared by sequentially mixing the following solutions: (i) a solution of 0.15 M Na2SO3, 0.05 M Na2S2O3·5H2O and 0.05 M NH4Cl, (ii) a solution of 0.015 M Na(AuCl4)·2H2O, and (iii) a solution of 2 % HF. The mixed solution was heated to 60 °C and the prism was immersed in the heated solution for a finite time. The deposition process was repeated until the deposited Au film exhibited optimal conductivity. After deposition, Ir–CNF1800 ink was loaded onto the Au surface using dropcasting. The prepared prism was connected to a cell. Pt wire and Ag/AgCl were employed as counter and reference electrodes, respectively. 0.1 M HClO4 (25 °C) was used as an electrolyte. The cell was then placed in a Fourier-transform infrared spectroscopy instrument (FT–IR; Cary 670 FTIR,

Agilent Technologies). Before initiating CP at 10 mA cm−2, the electrolyte was purged with N2 at 30 sccm. For CP testing, a potentiostat (Autolab PGSTAT302N, Metrohm AG) was utilized and the FT–IR was operated simultaneously to collect signals.

2.2. Electrochemical titanium passivation All electrochemical tests were performed using Ti foil disks (thickness: 0.125 mm, diameter: 15 mm, purity: 99.6 %; GoodFellow) as working electrode. The catalyst was loaded onto Ti disks using dropcasting or electrodeposition. Prior to loading, the disks were cleaned using a three-step sonication in a bath sonicator: (i) 15 min in acetone, (ii) 10 min in ethanol, and (iii) 5 min in H2O. When dropcasting was used, a catalyst of powdery IrO2 (Premetek Co.) was made into ink and then loaded onto Ti foil. When electrodeposition was used, iridium oxide was anodically electrodeposited on Ti foil as described below (cf. detailed procedure available in Supplementary data §4). Either case, the loading mass was 670 μg cm−2. The prepared Ti disk was combined with the current collector using a customized Teflon cap. The combined electrode was connected to a rotator (Pine Research Instrumentation) of a potentiostat (VSP, Bio-Logic). Passivation tests were performed in a cell using CP at 1 mA cm−2. During the CP test, the rotator was maintained at 1600 rpm. Pt coil and Ag/AgCl (3 M KCl) were used as counter electrode and reference electrode, respectively. The electrolyte was 25 °C 0.1 M HClO4 (Fluka), bubbled with N2 at 30 sccm before and during testing. The anodic electrodeposition (AED) of iridium oxide was carried out under certain conditions in which rapid breakdown anodization was avoided and other parameters were optimized (cf. detailed discussion available in Supplementary data §4). AED was performed

using a customized instrument. Cu plate and Pt wire were used as current collector and counter electrode, respectively. As a power source, a potentiostat (2400 Source Meter, Keithley) was used. For optimal AED, we modified the Yamanaka’s method [42]. The following is the recipe used in our experiment: (i) A quantity corresponding to 1.4929 g L−1 (4.2 to 5 mM, Table S4) of IrCl3·xH2O (reagent grade; Aldrich) was prepared and dissolved in a small amount of H2O in a volumetric flask; (ii) A quantity corresponding to 6.3 g L−1 (50 mM) of (COOH)2·2H2O (≥ 99.5 %; KANTO CHEMICAL CO., INC.) was prepared and added to the IrCl3 solution, and an additional small amount of H2O was added to the solution; (iii) A quantity corresponding to 41.46 g L−1 (300 mM) of K2CO3 (≥ 99.5 %; JUNSEI) was prepared and added to the solution; (iv) A specific amount of H2O to fill the flask up to the graduation mark was added to the solution; (v) After sufficient dissolution, the solution was transferred to a glass container and sealed with a cap. The prepared electrodeposition bath was aged up to 40 weeks to monitor the formation and degradation of the ionic precursor complex, [Ir(OH)6]3−. Detailed equipment specifications and further discussion can be found in Supplementary data §4.

3. Results and discussion 3.1. Carbon catalyst–supports and substrates Under OER-operating conditions, water molecules can participate in various reactions at the anode. The OER can be initiated on the surface of the electrocatalyst. If the substrate is composed of carbon, the COR can occur on the substrate surface by reaction with water. The theoretical onset potentials of OER (Eq. 1) and COR (Eq. 2 & 3) at 25 °C are known as [43, 44]:

H2O → ½O2 + 2H+ + 2e−, E0 = 1.23 V vs. SHE

(1)

C + 2H2O → CO2 + 4H+ + 4e−, E0 = 0.207 V vs. SHE

(2)

C + H2O → CO + 2H+ + 2e−, E0 = 0.518 V vs. SHE

(3)

where E0 is the standard electrode potential versus the standard hydrogen electrode (SHE). The thermodynamic onset potentials of the CORs are much lower than that of the OER. Therefore, under OER-operating conditions, it is expected that carbonaceous materials should undergo COR prior to OER. However, it has been reported that some advanced carbon structures can remain somewhat stable even under oxidizing conditions due to increased overpotential. Examples are glassy carbon and graphitic carbon [45-48]. Nevertheless, these carbons will eventually experience the COR when the applied potential exceeds the overpotential [45, 49]. Even under conditions where COR should occur, carbon corrosion can be circumvented. It is known that the presence of active OER electrocatalyst can inhibit COR [43, 50]. CORC/CO2 (Eq. 2) can be suppressed when the active catalyst coexists with carbon on the surface [51]. However, it is still unclear whether CORC/CO (Eq. 3) can also be avoided by the coexistence of the catalyst. We designed a specific structure that is ideal for operando detection of carbon corrosion in the coexistence of a catalyst and a carbon support. Model systems such as Scheme 1A–1C are not considered because the results are obvious: the exclusive occurrence of OER or COR. The model system in Scheme 1D (i.e. a core@shell structure) is excluded because it is evident that

the COR is the only reaction that occurs at the beginning of CP operation and that this spherical particle can be physically disconnected from the electrode due to the carbon corrosion. Therefore, we thought of a structure in which the position of the core is biased but not concentric with the shell (Scheme 1E). The expected result of using this system is that COR occurs only at the beginning of operation until the catalyst is exposed to the electrolyte due to carbon corrosion. In this case, carbon and catalyst can coexist on the surface after sufficient carbon corrosion. However, this system is also excluded because it is hardly available to selectively control the orientation of the catalyst core at the electrode. Failure in the control can lead to corrosion-induced detachment, as shown in Scheme 1F. As an ideal structure, we thus considered the fabrication of Ir–CNFs (Scheme 1G). In particular, among various possible Ir–CNF structures, a structure in which Ir is located biasedly at the subsurface of the CNF and the Ir content of the surface is minized was considered as the target structure. The use of this structure is beneficial in that it can distinctively show two different phases: COR at the beginning and co-occurrence of COR and OER at the extended operation. Given the fact that COR in the absence of catalyst is mostly CORC/CO2, we can expect that the use of this structure will lead to CORC/CO2 at the beginning of operation. The model system of Ir–CNF has been successfully fabricated using a sequential process consisting of electrospinning and carbonization. Taking into account the chemical formula of the chemical used, it can be assumed that N is the only impurity that can coexist in the synthesized Ir–CNF structure. Oxygen was not considered a possible impurity, because the presence of O during carbonization can cause the liberation of CO

and/or CO2. Hence, Ir–CNF produced at elevated temperatures could not contain oxygen species within its structure. The presence of N impurities can affect electrochemical reactions. N-doped C is one of the well-known non-precious metal OER catalysts [52-54]. In order to remove N from Ir–CNF, we added an ultra-high heat-treatment (UHT) process performed in an inert atmosphere with continuous ventilation using Ar. Interestingly, we observed that the additional process of UHT could be used to modify the surface elemental composition of Ir–CNF. Considering both the decomposition temperature of IrO2 [55-57] and the melting temperature of Ir [58], the HTT was confined to a temperature window between a minimum of 1100 °C and a maximum of 2450 °C. Accordingly UHT was performed at 1200, 1500, 1800, 2100 and 2400 °C (Fig. 1). The low decomposition temperature of IrO2 can be considered the second reason why O is not thought of as possible impurity, in addition to the first reason described above (e.g. 400–700 °C, Fig. S8). After synthesis, the physicochemical properties of the model system were characterized. First, we studied the UHT process to remove N impurities. The XPS spectrum of Ir–CNF1200 (Fig. 2A, Figure S2) shows characteristic peaks of pyridinic N (398.2–398.8 eV) and quaternary N (401.3–401.9 eV) [59-61]. Here, the peak of quaternary N is the most pronounced. The intensity of N 1s is significantly reduced by UHT at 1500 °C (Fig. 2A). This is in good agreement with the report that N begins to be removed from CNF at ca. 1400 °C (especially for polyacrylonitrile-derived CNFs) [62]. Moreover, the N-related peaks completely disappear from the XPS spectrum of Ir– CNF1800 (Fig. 2A). Removal of N species from a layer of graphene can lead to defects,

which can affect the surface area of the material [63]. Accordingly, the following trends were found in the surface area analysis: Ir–CNF1200 < Ir–CNF1500 < Ir–CNF1800 (Table 1).

Table 1. Selected physicochemical properties of Ir–CNFs. HTT/°C a

1200

1500

1800

2100

2400

BET/m2 g−1 b

11

13

22

19

17

d002/nm c

0.361

0.353

0.344

0.344

0.343

Surface Ir/wt% d

2.16

3.38

0.32

0.16

0.00 e

Total Ir/wt% f

3.67

12.91

21.19

10.84

7.28

a

Heat-treatment temperature used to prepare an Ir–CNF sample.

b

Surface area based on the Brunauer–Emmett–Teller theory (Figure S5, Table S1).

c

Interplanar C–C spacing acquired from XRD data (Figure S3).

d

Surface Ir content converted from XPS data (Figure 2, Table S2).

e

Not detected.

f

Actual total Ir content calculated from TGA data (Figure S8, Table S3).

There are more reasons why HTT of 1800 °C or higher should be chosen when preparing a model system. One of the reasons is the corrosion resistance of CNF. Given the fact that graphene alignment starts to form within carbon structure from around 1550 °C [64], the minimum HTT required is 1550 °C. Indeed, when compared to Ir– CNF1500, Ir–CNF1800 shows graphitic features. The C–C interplanar spacing (d002) of Ir–

CNF1800 is 0.344 nm (Table 1 and Fig. S4), which is very close to that of graphite (i.e. 0.34 nm) [65]; on the other hand, the d002 of Ir–CNF1500 is 0.353 nm. Also, comparing Ir– CNF1500 and Ir–CNF1800 in Raman shift (Fig. S3A) and C 1s π–π* satellite XPS (Fig. S2E) suggests that the critical temperature for sp2 hybridization lies between 1500 °C and 1800 °C (Supplementary data §2). Another reason is that UHT at elevated temperatures can control the surface coverage of Ir. Surface Ir content is significantly reduced by UHT over 1800 °C (Fig. 2B). XPS analysis was performed to determine the surface elemental composition. Considering the mean free path of a solid material, the electrons detected in XPS come from only a few atomic layers of the sample surface [66]. Thus, the discrepancy between the XPS data and the bulk characterization results can be understood as the biased spatial distribution of the detected element. In the XPS data (Table 1), the surface Ir content of Ir–CNF1500 is 3.38 wt%, whereas that of Ir–CNF1800 is 0.32 wt%. The decrease is ascribable to the evaporation of Ir through sublimation and vaporization [67-69]. However, the bulk Ir content of Ir–CNF1800 is higher than that of Ir–CNF1500. The actual Ir content of Ir–CNF1500 is 12.91 wt%, while that of Ir–CNF1800 is 21.19 wt%. This mismatch indicates that Ir–CNF1800 contains more Ir than Ir–CNF1500, but most of Ir in Ir–CNF1800 is not located on the surface of the structure. The reason the actual Ir content of Ir–CNF1800 is higher than that of Ir–CNF1500 is that the removal of N and N-related compounds by UHT entails a reduction in total mass. UHT above 2000 °C is not optimal because the total Ir content drops significantly (Table 1). Taking these considerations into account, Ir– CNF1800 was chosen as the ideal structure. Experimental evidence for why Ir–CNF1800 is considered an ideal structure can be found in Supplementary data §3 (Fig. S9).

Operando SEIRA spectroscopy was performed using Ir–CNF1800 to observe the change of the interface upon the application of anodic potential (Fig. 4A). CORC/CO2 was mostly observed at the beginning of CP operation. As soon as CP started (Fig. 4B), CO2 formation was detected, but soon the relevant peak disappeared (Fig. 4D). After a few seconds, CO2 was detected again for a short period (Fig. 4C). Conversely, CORC/CO was not the dominating COR at the beginning, but it occurred consistently over time (i.e. C=O peak in Fig. 4D). Over time, the humps representing C=O and O–H, respectively, become apparent (Fig. 4D). Therefore, it is assumed that CORC/CO can occur continually during OER. The fact that the occurrence of CO2 peaks is concentrated mostly on the early SEIRA spectra is of significant importance. This is in good agreement with the experimental observation that CORC/CO2 occurs predominantly in the absence of catalyst at the reaction site. To demonstrate this experimentally, we measured potential–time profiles using samples with varied surface Ir content (Supplementary data §3). The results suggest that, when anodic potential is applied, COR can occur alone or with OER until a sufficient amount of Ir is exposed to the electrolyte (Fig. S10). Increase in the amount of surface Ir can reduce the overpotential by promoting catalysis towards OER (Fig. 5). In the case of Ir–CNF1800 (Fig. 5), CORC/CO2 mainly occurs until ca. 50 s. We extracted an initial SEIRA spectrum between 10 and 40 s (Fig. 4C). Within this region, the peak of CO2 appears for a short period of time and then disappears, while the humps for C=O and O–H intensify over time. The limited occurrence of the CO2 peak indicates that CORC/CO2 would be a self-controlled reaction that begins only when the number of electrocatalytically active sites on the surface is not sufficient. In particular, the initial SEIRA at 5 s (Fig. 4D) displays a clear spectrum of CO2, which is accompanied by a decrease in overpotential observed

in CP tests performed concurrently at this point of time (Fig. 4B inset). This improvement in CP can be attributed to the exposure of Ir to the electrolyte, which is caused by the electrochemical removal of the surrounding C by the COR. As well as the CO2 peak, the C=O hump is also ascribable to the COR. The C=O spectra, observed in the absence of simultaneous CO2 peaks, were assigned to CORC/CO. We conjecture that CORC/CO is the main cause of performance degradation at the interface of Ir and C. The reason is that the C=O humps are continuously observed until the termination of the cell (e.g. ca. 360 s in Fig. 4B), whereas the occurrence of CO2 during that period is infrequent (Fig. 4D). The termination of the cell (or the dramatic leap in overpotential present in Fig. 4B) is a good indication of the absence of the catalyst. Thus, it is obvious that the loss or disconnection of the Ir from the C is accompanied by the CORC/CO. In this regard, we can say that the CORC/CO is the main cause of the interface-driven performance degradation during OER at the interface of C and Ir.

3.2. Non-carbon catalyst–supports and substrates When exposed to air at ambient temperature, Ti forms passive native oxide on the surface. Anodizing of Ti can also lead to the growth of the oxide layer and the resulting thickness depends on the applied potential (Scheme 2A) [70]. Dissimilar to Ti, TiO2 has a significant band gap [71]. When Ti is used as the substrate for OER, the passivation layer can become thick due to the oxidizing atmosphere, which can interfere with electron transfer. This interference is considered to be the cause of performance degradation at the interface of the catalyst and the Ti substrate [72]. In ideal circumstances, oxide growth can be suppressed if Ti is

completely shielded by suitable materials (Scheme 2B). It was experimentally indicated that in OER, the minimum thickness required to avoid reaction between the substrate and the electrolyte is a few nanometers of the shielding catalyst layer [41]. However, defective shielding can initiate Ti passivation [73]. Ti passivation can continue to grow inwards from the surface over time (Scheme 2C). Taking this into account, a straightforward electrochemical method was devised to observe interface-driven performance degradation. That is two different sets of comparative studies. Stoichiometric IrO2 was used as a catalyst because it is known to have high electrochemical stability [40, 74]. In the first set of comparative studies, we performed a controlled polarization test to investigate the effect of the growth of oxide layer during OER (Scheme 2C). In the second set of comparative studies, experiments were conducted using anodized and electrodeposited Ti samples to rule out the insulating effects originating from the presence of a native oxide layer (Scheme 2D). Here, both the anodization and electrodeposition processes were performed at 20 V. This is because using such a high potential can result in the formation of a thick passivation layer that is much thicker than the native oxide layer [70, 75-78]. The first set of comparative studies is based on the following considerations. The use of a powdery catalyst may result in imperfect or insufficient coverage of the catalyst on the surface of the Ti substrate (Scheme 2C). If there are defects such as pinholes in the catalyst shielding layer, Ti can be anodically oxidized under OER-operating conditions. In this case, the passivation layer thickens and the overpotential increases due to the impeded electron transfer. In Fig. 6a (i.e. α), CP shows the increase in overpotential over time. However, this observation can be due to other reasons such as catalyst deactivation during OER. In order to decouple the

deactivation-driven overpotential (due to IrO2 catalyst) from the passivation-driven overpotential (due to Ti substrate), a comparative study was conducted in the following manner. After the first cycle of CP, the surface of α was cleaned to remove the catalyst (Fig. 6i). Then a new layer of IrO2 (the same loading amount as for α) was loaded onto the cleaned α surface. We named this structure β. Whether Ti passivation can occur parasitically under OERoperation conditions can be ascertained by comparing the CP of α and that of β. This is because the growth of the passivation layer lowers the electron transfer. If the CP of β is the same as that of α, it can be said that α and β have the identical passivation layer thickness. On the other hand, if the CP of β exhibits an increased overpotential compared to that of α, it can be said that the oxide layer becomes thicker during OER. When compared with α, the CP performed using β shows an increased overpotential (Fig. 6b). This phenomenon is a clear indication that the interface of Ti was passivated and that passivation contributes to the interface-driven performance degradation during OER. An additional cycle of CP was performed to investigate whether the growth of the oxide layer was self-limiting under OER-operating conditions. After the second cycle of CP, the surface of β was cleaned to remove the catalyst (Fig. 6ii). Then a new layer of IrO2 (the same loading amount as for α or for β) was loaded onto the cleaned α surface. We named this structure γ. When compared with β, the CP performed using γ shows an increased overpotential but the increase in overpotential between γ and β is not as pronounced as that between β and α (Fig. 6). This suggests that there is a finite thickness for the oxide layer to grow under a given anodic condition. For the second set of comparative studies, the conditions for the AED process and CP-based degradation tests were optimized. When high potentials such as tens of volts are

used, several conditions must be met to deposit iridium oxide on the Ti surface (Supplementary data §4): First, the ionic precursor complex should be prepared to a sufficient degree by controlled hydrolysis [79-84]. To optimize the conditions for hydrolysis (i.e. to avoid the condensation of the hydrolytic products), we performed long-term aging of the precursor solution up to 40 weeks (Fig. S12, S13). Second, it is necessary to control the ionic activity and mobility of the solution to circumvent the rapid breakdown anodization of Ti [85100]. For this, the following parameters were optimized: solution temperature, ionic composition, ionic concentration, and agitation speed (Fig. S11). Taking into account the aforementioned considerations, the model system for the second set of comparative studies was prepared under the following conditions: 4 weeks of aging period, 20 V of applied potential, pH 10.35 of solution basicity, 180 rpm of agitation speed, and 30 °C of solution temperature (Fig. 7). We named the produced structure AED-IrOx/Ti (i.e. δ in Fig. 8). The CP of δ continues to function at a constant but non-negligible overpotential over time (Fig. 8b). The formation of the AED layer begins on the surface of the Ti oxide layer (Fig. S14). The presence of the Ti oxide layer is why the overpotential of δ is greater than that of α (cf. the catalyst loading amount of α is equal to that of δ). On the other hand, the constant overpotential of δ over time is due to the meticulous shielding of the Ti surface (Fig. 2D). This is also ascribable to an increase in the number of active sites contributed by the enhanced intrinsic catalytic activity of AED-IrOx when compared to the referenced stoichiometric IrO2 (Fig. S15). Lastly, we tested whether catalysis towards the OER can occur under extreme conditions that may be rarely observed during the actual functioning of electrolytic cells (i.e. in the presence of thick TiO2 layer at the interface). A layer of IrO2 (the same loading amount with the case of α) was loaded onto the TiO2/Ti substrate. The substrate was prepared by anodizing a Ti disk at 20 V

in the absence of an ionic Ir complex precursor. We named this structure IrO2/TiO2/Ti (i.e. ε in Fig. 8). The CP of ε shows a significant increase in overpotential compared to that of α (Fig. 6), but the pattern and trend of CP of ε over time is almost the same as that of α. This phenomenon indicates that the OER can be functioning even in the presence of a passivation layer, but in this case, electron transfer is disturbed, resulting in a significant reduction in kinetics. 4. Conclusions We successfully fabricated the carbon and non-carbon structures idealized for operando (spectro)electrochemical observation of interface-driven performance degradations. Using these model systems, it was possible to monitor the occurrence of corrosion or passivation under OERoperating conditions. Carbon corrosion occurred in two different forms depending on the circumstance. When the carbon substrate was not covered by the catalyst or was insufficiently covered, the result was the predominant occurrence of CO2 formation. On the other hand, sufficient coexistence of the catalyst and carbon exposed to the electrolyte resulted in the predominant occurrence of CO formation. The parasitic occurrence of carbon corrosion during OER led to the physical disconnection of the catalyst from the substrate. Titanium was chosen as the non-carbon material to demonstrate passivation-induced performance degradation. It has been found that passivation of the substrate can be avoided if the substrate surface is completely shielded by a sufficiently thick active catalyst layer. Another finding is that the parasitic occurrence of passivation during OER can occur in a self-limiting manner. The consequence of corrosion was an irreversible loss of the catalyst, and that of passivation was the inhibition of electron transfer. Therefore, a non-carbon substrate (i.e. Ti) is considered a better choice than a carbon substrate. However, in order to use Ti as a practical substrate, the growth of the insulating oxide layer must be circumvented. Possible strategies for

this are: (i) Shielding the Ti surface completely with a conductive film that is stable under OERoperating conditions (e.g. PtOx/Pt but not AuOx/Au since the latter could be dissolved during the OER), and (ii) Modifying the conductivity of the passivation layer through doping or other surface engineering methods (e.g. Nb-doped TiO2 but not self-doped TinO2n−1 since the latter could be oxidized into TiO2 during the OER). In this regard, future research on substrate material design should aim at the development of conductive and stable materials under oxidative conditions, and it is envisaged that the in situ methodology introduced in this study can facilitate the screening procedure for material selection. Acknowledgements This research was supported by Technology Development Program to Solve Climate Changes through the National Research Foundation of Korea (NRF) funded by the Ministry of Science, ICT (2018M1A2A2063861). References [1] Z.W. Seh, J. Kibsgaard, C.F. Dickens, I. Chorkendorff, J.K. Nørskov, T.F. Jaramillo, Combining theory and experiment in electrocatalysis: Insights into materials design, Science, 355 (2017) eaad4998. [2] A. Kirubakaran, S. Jain, R.K. Nema, A review on fuel cell technologies and power electronic interface, Renew. Sustain. Energy Rev., 13 (2009) 2430. [3] O. Schmidt, A. Gambhir, I. Staffell, A. Hawkes, J. Nelson, S. Few, Future cost and performance of water electrolysis: An expert elicitation study, Int. J. Hydrog. Energy, 42 (2017) 30470.

[4] J. Chi, H. Yu, Water electrolysis based on renewable energy for hydrogen production, Chin. J. Catal., 39 (2018) 390. [5] Z. Lei, T. Wang, B. Zhao, W. Cai, Y. Liu, S. Jiao, Q. Li, R. Cao, M. Liu, Recent progress in electrocatalysts for acidic water oxidation, Adv. Energy Mater., 10 (2020) 2000478. [6] J. Mohammed-ibrahim, H. Moussab, Tuning the electronic structure of the earth-abundant electrocatalysts for oxygen evolution reaction (OER) to achieve efficient alkaline water splitting – A review, J. Energy Chem., 56 (2021), 299. [7] J. Mohammed-Ibrahim, A review on NiFe-based electrocatalysts for efficient alkaline oxygen evolution reaction, J. Power Sources, 448 (2020) 227375. [8] M.-I. Jamesh, Y. Kuang, X. Sun, Constructing earth-abundant 3D nanoarrays for efficient overall water splitting – A review, ChemCatChem, 11 (2019) 1550. [9] J. Mohammed-Ibrahim, X. Sun, Recent progress on earth abundant electrocatalysts for hydrogen evolution reaction (HER) in alkaline medium to achieve efficient water splitting – A review, J. Energy Chem., 34 (2019) 111. [10] M. Schalenbach, A.R. Zeradjanin, O. Kasian, S. Cherevko, K.J.J. Mayrhofer, A perspective on low-temperature water electrolysis – Challenges in alkaline and acidic technology, Int. J. Electrochem. Sci., 13 (2018) 1173. [11] M. Pourbaix, Atlas of electrochemical equilibria in aqueous solutions, National Association of Corrosion Engineers, 1974. [12] W.M. Haynes, CRC Handbook of Chemistry and Physics, 96th Ed., CRC Press, 2015.

[13] A. Buttler, H. Spliethoff, Current status of water electrolysis for energy storage, grid balancing and sector coupling via power-to-gas and power-to-liquids: A review, Renew. Sustain. Energy Rev., 82 (2018) 2440. [14] C. Wei, R.R. Rao, J. Peng, B. Huang, I.E.L. Stephens, M. Risch, Z.J. Xu, Y. Shao-Horn, Recommended practices and benchmark activity for hydrogen and oxygen electrocatalysis in water splitting and fuel cells, Adv. Mater., 31 (2019) 1806296. [15] L.F. Castañeda, F.C. Walsh, J.L. Nava, C. Ponce de León, Graphite felt as a versatile electrode material: Properties, reaction environment, performance and applications, Electrochim. Acta, 258 (2017) 1115. [16] J.D. Benck, B.A. Pinaud, Y. Gorlin, T.F. Jaramillo, Substrate selection for fundamental studies of electrocatalysts and photoelectrodes: Inert potential windows in acidic, neutral, and basic electrolyte, PLOS ONE, 9 (2014) e107942. [17] M.I. Jamesh, P. Li, M.M.M. Bilek, R.L. Boxman, D.R. McKenzie, P.K. Chu, Evaluation of corrosion resistance and cytocompatibility of graded metal carbon film on Ti and NiTi prepared by hybrid cathodic arc/glow discharge plasma-assisted chemical vapor deposition, Corros. Sci., 97 (2015) 126. [18] M. Jamesh, T.S.N. Sankara Narayanan, P.K. Chu, Thermal oxidation of titanium: Evaluation of corrosion resistance as a function of cooling rate, Mater. Chem. Phys., 138 (2013) 565. [19] M.I. Jamesh, G. Wu, Y. Zhao, D.R. McKenzie, M.M.M. Bilek, P.K. Chu, Electrochemical corrosion behavior of biodegradable Mg–Y–RE and Mg–Zn–Zr alloys in Ringer’s solution and simulated body fluid, Corros. Sci., 91 (2015) 160.

[20] M. Jamesh, S. Kumar, T.S.N. Sankara Narayanan, Corrosion behavior of commercially pure Mg and ZM21 Mg alloy in Ringer’s solution – Long term evaluation by EIS, Corros. Sci., 53 (2011) 645. [21] M.I. Jamesh, Recent progress on earth abundant hydrogen evolution reaction and oxygen evolution reaction bifunctional electrocatalyst for overall water splitting in alkaline media, J. Power Sources, 333 (2016) 213. [22] C. Sandford, M.A. Edwards, K.J. Klunder, D.P. Hickey, M. Li, K. Barman, M.S. Sigman, H.S. White, S.D. Minteer, A synthetic chemist's guide to electroanalytical tools for studying reaction mechanisms, Chem. Sci., 10 (2019) 6404. [23] S.J. Ashton, M. Arenz, A DEMS study on the electrochemical oxidation of a high surface area carbon black, Electrochem. Commun., 13 (2011) 1473. [24] H.M.A. Amin, H. Baltruschat, D. Wittmaier, K.A. Friedrich, A highly efficient bifunctional catalyst for alkaline air-electrodes based on a Ag and Co3O4 hybrid: RRDE and online DEMS insights, Electrochim. Acta, 151 (2015) 332. [25] W.A. Badawy, S.S. Elegamy, K.M. Ismail, Comparative study of tantalum and titanium passive films by electrochemical impedance spectroscopy, Br. Corros. J., 28 (1993) 133. [26] O. Diaz-Morales, F. Calle-Vallejo, C. de Munck, M.T.M. Koper, Electrochemical water splitting by gold: evidence for an oxide decomposition mechanism, Chem. Sci., 4 (2013) 2334. [27] V. Pfeifer, T.E. Jones, J.J. Velasco Vélez, R. Arrigo, S. Piccinin, M. Hävecker, A. KnopGericke, R. Schlögl, In situ observation of reactive oxygen species forming on oxygen-evolving iridium surfaces, Chem. Sci., 8 (2017) 2143.

[28] Z. Pavlovic, C. Ranjan, Q. Gao, M. van Gastel, R. Schlögl, Probing the structure of a wateroxidizing anodic iridium oxide catalyst using Raman spectroscopy, ACS Catal., 6 (2016) 8098. [29] I.S. Filimonenkov, S.Y. Istomin, E.V. Antipov, G.A. Tsirlina, E.R. Savinova, Rotating ringdisk electrode as a quantitative tool for the investigation of the oxygen evolution reaction, Electrochim. Acta, 286 (2018) 304. [30] I.S. Filimonenkov, C. Bouillet, G. Kéranguéven, P.A. Simonov, G.A. Tsirlina, E.R. Savinova, Carbon materials as additives to the OER catalysts: RRDE study of carbon corrosion at high anodic potentials, Electrochim. Acta, 321 (2019) 134657. [31] K.H. Saeed, M. Forster, J.-F. Li, L.J. Hardwick, A.J. Cowan, Water oxidation intermediates on iridium oxide electrodes probed by in situ electrochemical SHINERS, Chem. Commun., 56 (2020) 1129. [32] H.-S. Oh, H.N. Nong, T. Reier, A. Bergmann, M. Gliech, J. Ferreira de Araújo, E. Willinger, R. Schlögl, D. Teschner, P. Strasser, Electrochemical catalyst–support effects and their stabilizing role for IrOx nanoparticle catalysts during the oxygen evolution reaction, J. Am. Chem. Soc., 138 (2016) 12552. [33] J. Hwang, R.R. Rao, L. Giordano, Y. Katayama, Y. Yu, Y. Shao-Horn, Perovskites in catalysis and electrocatalysis, Science, 358 (2017) 751. [34] M. Osawa, Dynamic processes in electrochemical reactions studied by surface-enhanced infrared absorption spectroscopy (SEIRAS), Bull. Chem. Soc. Jpn., 70 (1997) 2861.

[35] N. Sivasankar, W.W. Weare, H. Frei, Direct observation of a hydroperoxide surface intermediate upon visible light-driven water oxidation at an Ir oxide nanocluster catalyst by rapid-scan FT-IR spectroscopy, J. Am. Chem. Soc., 133 (2011) 12976. [36] Y. Yao, S. Hu, W. Chen, Z.-Q. Huang, W. Wei, T. Yao, R. Liu, K. Zang, X. Wang, G. Wu, W. Yuan, T. Yuan, B. Zhu, W. Liu, Z. Li, D. He, Z. Xue, Y. Wang, X. Zheng, J. Dong, C.-R. Chang, Y. Chen, X. Hong, J. Luo, S. Wei, W.-X. Li, P. Strasser, Y. Wu, Y. Li, Engineering the electronic structure of single atom Ru sites via compressive strain boosts acidic water oxidation electrocatalysis, Nat. Catal., 2 (2019) 304. [37] S. Trasatti, Electrocatalysis: understanding the success of DSA®, Electrochim. Acta, 45 (2000) 2377. [38] M. Carmo, D.L. Fritz, J. Mergel, D. Stolten, A comprehensive review on PEM water electrolysis, Int. J. Hydrog. Energy, 38 (2013) 4901. [39] C. Iwakura, K. Hirao, H. Tamura, Anodic evolution of oxygen on ruthenium in acidic solutions, Electrochim. Acta, 22 (1977) 329. [40] N. Danilovic, R. Subbaraman, K.-C. Chang, S.H. Chang, Y.J. Kang, J. Snyder, A.P. Paulikas, D. Strmcnik, Y.-T. Kim, D. Myers, V.R. Stamenkovic, N.M. Markovic, Activity– stability trends for the oxygen evolution reaction on monometallic oxides in acidic environments, J. Phys. Chem. Lett., 5 (2014) 2474. [41] H. Jang, T.T. Hieu, S.H. Kim, J. Lee, Reduction of iridium loading to the minimum level required for water oxidation electrocatalysis without sacrificing the electrochemical stability, J. Phys. Chem. C, 123 (2019) 12928.

[42] K. Yamanaka, Anodically electrodeposited iridium oxide films (AEIROF) from alkaline solutions for electrochromic display devices, Jpn. J. Appl. Phys., 28 (1989) 632. [43] L. Hu, B.K. Hong, J.-G. Oh, S. Litster, Robust operation of fuel cell systems in subfreezing conditions: A material-based solution to achieve better anode durability, ACS Appl. Energy Mater., 2 (2019) 7152. [44] H. Ju, S. Badwal, S. Giddey, A comprehensive review of carbon and hydrocarbon assisted water electrolysis for hydrogen production, Appl. Energy, 231 (2018) 502. [45] Y. Yi, G. Weinberg, M. Prenzel, M. Greiner, S. Heumann, S. Becker, R. Schlögl, Electrochemical corrosion of a glassy carbon electrode, Catal. Today, 295 (2017) 32. [46] A.F. Craievich, On the structure of glassy carbon, Mater. Res. Bull., 11 (1976) 1249. [47] G.M. Jenkins, K. Kawamura, Structure of Glassy Carbon, Nature, 231 (1971) 175. [48] B. Jeong, J.D. Ocon, J. Lee, Electrode architecture in galvanic and electrolytic energy cells, Angew. Chem. Int. Ed., 55 (2016) 4870. [49] S. Geiger, O. Kasian, A.M. Mingers, S.S. Nicley, K. Haenen, K.J.J. Mayrhofer, S. Cherevko, Catalyst stability benchmarking for the oxygen evolution reaction: The importance of backing electrode material and dissolution in accelerated aging studies, ChemSusChem, 10 (2017) 4140. [50] S. Watzele, P. Hauenstein, Y. Liang, S. Xue, J. Fichtner, B. Garlyyev, D. Scieszka, F. Claudel, F. Maillard, A.S. Bandarenka, Determination of electroactive surface area of Ni-, Co-, Fe-, and Ir-based oxide electrocatalysts, ACS Catal., 9 (2019) 9222.

[51] S. Möller, S. Barwe, J. Masa, D. Wintrich, S. Seisel, H. Baltruschat, W. Schuhmann, Online monitoring of electrochemical carbon corrosion in alkaline electrolytes by differential electrochemical mass spectrometry, Angew. Chem. Int. Ed., 59 (2020) 1585. [52] M. Tahir, L. Pan, F. Idrees, X. Zhang, L. Wang, J.-J. Zou, Z.L. Wang, Electrocatalytic oxygen evolution reaction for energy conversion and storage: A comprehensive review, Nano Energy, 37 (2017) 136. [53] S. Chen, J. Duan, J. Ran, S.-Z. Qiao, Paper-based N-doped carbon films for enhanced oxygen evolution electrocatalysis, Adv. Sci., 2 (2015) 1400015. [54] S. Chen, J. Duan, M. Jaroniec, S.-Z. Qiao, Nitrogen and oxygen dual-doped carbon hydrogel film as a substrate-free electrode for highly efficient oxygen evolution reaction, Adv. Mater., 26 (2014) 2925. [55] W.-P. Wu, Z.-F. Chen, Iridium coating: Processes, properties and application. Part I, Johnson Matthey Technol. Rev., 61 (2017) 16. [56] H. Jang, J. Lee, Iridium oxide fabrication and application: A review, J. Energy Chem., 46 (2020) 152. [57] S. Cotton, Chemistry of Precious Metals, Springer Netherlands, 1997. [58] L.B. Hunt, A history of iridium, Platin. Met. Rev., 31 (1987) 32. [59] L. Roldán, S. Armenise, Y. Marco, E. García-Bordejé, Control of nitrogen insertion during the growth of nitrogen-containing carbon nanofibers on cordierite monolith walls, Phys. Chem. Chem. Phys., 14 (2012) 3568.

[60] S. van Dommele, A. Romero-Izquirdo, R. Brydson, K.P. de Jong, J.H. Bitter, Tuning nitrogen functionalities in catalytically grown nitrogen-containing carbon nanotubes, Carbon, 46 (2008) 138. [61] Y. Deng, Y. Xie, K. Zou, X. Ji, Review on recent advances in nitrogen-doped carbons: preparations and applications in supercapacitors, J. Mater. Chem. A, 4 (2016) 1144. [62] L.R. Radovic, Chemistry and physics of carbon, CRC Press Taylor & Francis Group, 2012. [63] R. Daiyan, X. Tan, R. Chen, W.H. Saputera, H.A. Tahini, E. Lovell, Y.H. Ng, S.C. Smith, L. Dai, X. Lu, R. Amal, Electroreduction of CO2 to CO on a mesoporous carbon catalyst with progressively removed nitrogen moieties, ACS Energy Lett., 3 (2018) 2292. [64] A. Oberlin, Carbonization and graphitization, Carbon, 22 (1984) 521. [65] L.M. Rivera, S. Fajardo, M.D.C. Arévalo, G. García, E. Pastor, S- and N-doped graphene nanomaterials for the oxygen reduction reaction, Catalysts, 7 (2017) 278. [66] J.F. Moulder, J. Chastain, R.C. King, Handbook of X-ray photoelectron spectroscopy: A reference book of standard spectra for identification and interpretation of XPS data, Physical Electronics, 1995. [67] H. Jehn, R. Völker, M.I. Ismail, Iridium losses during oxidation, Platin. Met. Rev., 22 (1978) 92. [68] R.T. Wimber, H.G. Kraus, Oxidation of iridium, Metall. Trans., 5 (1974) 1565. [69] J. R. F. Hampson, R.F. Walker, Vapor pressures of platinum, iridium, and rhodium, J. Res. Natl. Bur. Stand.―A. Phys. Chem., 65A (1961) 289.

[70] M.V. Diamanti, B. Del Curto, M. Pedeferri, Interference colors of thin oxide layers on titanium, Color Res. Appl., 33 (2008) 221. [71] N. Serpone, Is the band gap of pristine TiO2 narrowed by anion- and cation-doping of titanium dioxide in second-generation photocatalysts?, J. Phys. Chem. B, 110 (2006) 24287. [72] M.I. Rojas, M.J. Esplandiu, L.B. Avalle, E.P.M. Leiva, V.A. Macagno, The oxygen and chlorine evolution reactions at titanium oxide electrodes modified with platinum, Electrochim. Acta, 43 (1998) 1785. [73] G.P. Vercesi, J. Rolewicz, C. Comninellis, J. Hinder, Characterization of dsa-type oxygen evolving electrodes. Choice of base metal, Thermochim. Acta, 176 (1991) 31. [74] S. Geiger, O. Kasian, M. Ledendecker, E. Pizzutilo, A.M. Mingers, W.T. Fu, O. DiazMorales, Z. Li, T. Oellers, L. Fruchter, A. Ludwig, K.J.J. Mayrhofer, M.T.M. Koper, S. Cherevko, The stability number as a metric for electrocatalyst stability benchmarking, Nat. Catal., 1 (2018) 508. [75] J. Emsley, Nature's building blocks: An A-Z guide to the elements, Oxford University Press, 2001. [76] C.J. Boxley, H.S. White, C.E. Gardner, J.V. Macpherson, Nanoscale imaging of the electronic conductivity of the native oxide film on titanium using conducting atomic force microscopy, J. Phys. Chem. B, 107 (2003) 9677. [77] M.V. Diamanti, F.C. Spreafico, M.P. Pedeferri, Production of anodic TiO2 nanofilms and their characterization, Phys. Procedia, 40 (2013) 30-37.

[78] A. Mazzarolo, M. Curioni, A. Vicenzo, P. Skeldon, G.E. Thompson, Anodic growth of titanium oxide: Electrochemical behaviour and morphological evolution, Electrochim. Acta, 75 (2012) 288. [79] R.A. Lidin, V.A. Molochko, L.L. Andreeva, Reactivity of inorganic substances: Revised and augmented wdition, Begell House Inc., 1996. [80] D. Xu, P. Diao, T. Jin, Q. Wu, X. Liu, X. Guo, H. Gong, F. Li, M. Xiang, Y. Ronghai, Iridium oxide nanoparticles and iridium/iridium oxide nanocomposites: Photochemical fabrication and application in catalytic reduction of 4-nitrophenol, ACS Appl. Mater. Interfaces, 7 (2015) 16738. [81] P. Beutler, H. Gamsjäger, Preparation and ultraviolet–visible spectrum of hexaaquairidium(III), J. Chem. Soc., Chem. Commun., (1976) 554. [82] T. Ioroi, N. Kitazawa, K. Yasuda, Y. Yamamoto, H. Takenaka, Iridium oxide/platinum electrocatalysts for unitized regenerative polymer electrolyte fuel cells, J. Electrochem. Soc., 147 (2000) 2018. [83] M. Mehdipour, S.H. Tabaian, S. Firoozi, Anodic electrodeposition of ligand-free iridium oxide on titanium with high mass loading and study of electrochemical treatments, J. Electroanal. Chem., 858 (2020) 113831. [84] Z. Yan, H. Liu, Z. Hao, M. Yu, X. Chen, J. Chen, Electrodeposition of (hydro)oxides for an oxygen evolution electrode, Chem. Sci., (2020). [85] D.J. Blackwood, L.M. Peter, D.E. Williams, Stability and open circuit breakdown of the passive oxide film on titanium, Electrochim. Acta, 33 (1988) 1143.

[86] C. Richter, E. Panaitescu, R. Willey, L. Menon, Titania nanotubes prepared by anodization in fluorine-free acids, J. Mater. Res., 22 (2007) 1624. [87] H. Jang, J. Lee, J.K. Lee, Manufacturing method of titanium oxide nanostructures in freestanding powder form and titanium oxide nanostructures manufactured thereby, Republic of Korea Patent, 10-1960746, 2019, pp. 1-11. [88] J.F. Vanhumbeeck, J. Proost, Current understanding of Ti anodisation: Functional, morphological, chemical and mechanical aspects, Corros. Rev., 27 (2009) 117. [89] Y.-T. Sul, C.B. Johansson, Y. Jeong, T. Albrektsson, The electrochemical oxide growth behaviour on titanium in acid and alkaline electrolytes, Med. Eng. Phys., 23 (2001) 329. [90] F. Keller, M.S. Hunter, D.L. Robinson, Structural features of oxide coatings on aluminum, J. Electrochem. Soc., 100 (1953) 411. [91] P. Roy, S. Berger, P. Schmuki, TiO2 nanotubes: synthesis and applications, Angew. Chem. Int. Ed., 50 (2011) 2904. [92] S. John, V. Balasubramanian, B.A. Shenoi, Studies on anodizing of aluminium in alkaline electrolyte using alternating current, Surf. Technol., 26 (1985) 207. [93] V.P. Parkhutik, V.I. Shershulsky, Theoretical modelling of porous oxide growth on aluminium, J. Phys. D: Appl. Phys., 25 (1992) 1258. [94] D. Kowalski, D. Kim, P. Schmuki, TiO2 nanotubes, nanochannels and mesosponge: Selforganized formation and applications, Nano Today, 8 (2013) 235. [95] M. Wang, H. Yang, Y. Liu, Current oscillations during potentiostatic anodization of tin in alkaline electrolytes, Electrochim. Acta, 56 (2011) 7051.

[96] Y. Mizutani, S.J. Kim, R. Ichino, M. Okido, Anodizing of Mg alloys in alkaline solutions, Surf. Coat. Technol., 169-170 (2003) 143. [97] N. Nyein, Z. Lockman, A. Matsuda, G. Kawamura, W.K. Tan, T.Z. Oo, Formation of TiO2 nanotube arrays in KOH added fluoride-ethylene glycol (EG) electrolyte and its photoelectrochemical response, AIP Conf. Proc., 1733 (2016) 020030. [98] M.A.A. Taib, W.K. Tan, T. Okuno, G. Kawamura, M. Jaafar, K.A. Razak, A. Matsuda, Z. Lockman, Formation of TiO2 nanotube arrays by anodic oxidation in LiOH added ethylene glycol electrolyte and the effect of thermal annealing on the photoelectrochemical properties, AIP Conf. Proc., 1733 (2016) 020025. [99] K. Wang, G. Liu, N. Hoivik, E. Johannessen, H. Jakobsen, Electrochemical engineering of hollow nanoarchitectures: pulse/step anodization (Si, Al, Ti) and their applications, Chem. Soc. Rev., 43 (2014) 1476. [100] D. Regonini, C.R. Bowen, A. Jaroenworaluck, R. Stevens, A review of growth mechanism, structure and crystallinity of anodized TiO2 nanotubes, Mater. Sci. Eng. R, 74 (2013) 377.

Figures and captions

Scheme 1. Schematic representation of the expected consequences of application of anodic potential. (A–F) Cross-sectional illustrations showing the shortcomings of these structures for operando studies. (A) Carbon film on top of the substrate can be corroded. (B) Catalyst film on

top of the substrate can catalyze the reaction without carbon corrosion. (C) Spherical carbon particle can be detached because of the carbon corrosion. (D) Core-site catalyst cannot participate in catalysis so shell-site carbon undergoes corrosion. (E) Catalyst can participate in catalysis after being exposed to the electrolyte. (F) Catalyst can hardly participate in catalysis because of a lack of arrival of reactants to the catalyst surface. (G) Three-dimensional and crosssectional illustration of the ideal structure for operando studies.

Fig. 1. Morphologies of Ir–CNFs. (A–E) Transmission electron microscopy images sharing a scale bar in (E). (F–J) Scanning electron microscopy images sharing a scale bar in (J). (A, F) Ir– CNF1200. (B, G) Ir–CNF1500. (C, H) Ir–CNF1800. (D, I) Ir–CNF2100. (E, J) Ir–CNF2400. Note: For brevity, an Ir–CNF sample prepared at a designated HTT of T °C is named Ir–CNFT.

Fig. 2. X-ray photoelectron spectroscopy of Ir–CNFs. (A) N 1s. (B) Ir 4f. Note: Temperatures in the figure is the used HTTs.

Fig. 3. Scanning transmission electron microscopy image of Ir–CNF1800 and its mapping with selected elements.

Fig. 4. Surface enhanced infrared absorption (SEIRA) spectroscopy with the placement of Ir– CNF1800 at an anode. (A) Schematic representation of the concept of the SEIRA study. (B) Chronopotentiometry performed simultaneously with the SEIRA test (the inset shows the enlarged image of the initial chronopotentiometry result near 5 s). (C) Initial SEIRA results (the hump observed between 2400 and 2300 cm−1 is responsible for CO2). (D) Overall SEIRA results including the spectra acquired under the open circuit voltage condition.

Fig. 5. Chronopotentiometry tested using Ir–CNF1800 and the schematic representation of the expected carbon corrosion mechanism over time during the OER. (i) At the early stage of the OER, CORC/CO2 is dominant. (ii) At the prolonged operation of the OER, CORC/CO is dominant. Note: Orange-colored arrows (or η↓) indicate the exposure of the catalyst to the electrolyte and thereby decreased overpotential.

Scheme 2. Schematic cross-sectional representation of the expected consequences of application of anodic potential on various structures. (A) Titanium with native oxide on its surface. (B) Titanium coated with a catalyst film. (C) Catalyst is insufficiently covering the titanium surface. (D) Catalyst is sufficiently covering the titanium surface.

Fig. 6. Chronopotentiometry tested on Ti surface in the presence of IrO2 catalyst. Schematic representation to explain experimental procedure: (a) IrO2 catalyst is loaded onto Ti electrode and then the OER is performed (i.e. model system of α). (i) After the testing, the surface is cleaned to remove the used IrO2. (b) New IrO2 catalyst is loaded onto the cleaned Ti electrode and then the OER is performed (i.e. model system of β). (ii) After the testing, the surface is cleaned to remove the used IrO2. (c) New IrO2 catalyst is loaded onto the cleaned Ti electrode and then the OER is performed (i.e. model system of γ). Note: The presence of native oxide on Ti surface is ignored for brevity.

Fig. 7. Photographs of anodically prepared samples. Note: TiO2/Ti is fabricated using an anodic treatment in the absence of ionic Ir complex precursor; A label below each image indicates the period of time spent to age a precursor solution.

Fig. 8. Chronopotentiometry tested on Ti and TiO2 surfaces in the presence of IrO2 catalyst in comparison with a sample prepared using the anodic electrodeposition (AED) of IrOx. Schematic representation to explain experimental procedures: (a) Stoichiometric IrO2 catalyst is loaded onto anodized Ti electrode whose surface is passivated by TiO2 and then, the OER is performed (i.e. model system of ε). (b) IrOx is formed on Ti electrode through AED during which the surface of Ti is passivated to leave a Ti oxide layer grown on top; afterwards, the OER is performed (i.e. model system of δ). (c) Stoichiometric IrO2 catalyst is loaded onto Ti electrode and then the OER is performed (i.e. identical to Figure 6a and thus the model system of α). Note: The presence of native oxide on Ti surface is ignored for brevity.

Graphical Abstract

CRediT author statement Hansaem Jang: Conceptualization, Methodology Validation, Formal analysis, Investigation, Writing Sunki Chung: Methodology, Formal analysis Jaeyoung Lee: Conceptualization, Writing - Review & Editing, Visualization, Supervision

Declaration of interests The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.