Durability of a thermally decomposed iridium(IV) oxide–tantalum(V) oxide coated titanium anode in aqueous ammonium citrate

2.1 Electrode preparation

Ti/IrO₂–Ta₂O₅ electrodes with a molar ratio of Ir:Ta = 6:4 were synthesized by thermal decomposition. First, titanium substrates (8 cm × 8 cm) were pretreated by etching after sandblasting or only etching. Subsequently, the precursor liquid was brushed onto the titanium plate. Then, the coated electrodes were dried. The preparation of the IrO₂–Ta₂O₅ electrodes by conventional method of thermal decomposition was described elsewhere in detail (Li et al. 2006). After each coating, the electrodes were calcined in ambient air for 15 min at 500 °C to complete the decomposition of the precursor and form the metal oxides. The total loading of Ti/IrO₂–Ta₂O₅ oxide coating is 2 mg/cm².

2.2 Physical characterization

The phases of IrO₂–Ta₂O₅ and IrO₂–Ta₂O₅–MnO_x oxide coatings were determined by X-ray diffraction (XRD) with Cu Kα radiation (Smartlab, Rigaku), collected by varying the 2θ angle from 10° to 80° at increments of 0.02°. The morphology and elemental composition were determined by scanning electron microscopy (SEM; JSM IT300). The elemental chemical states were determined by X-ray photoelectron spectroscopy (XPS, Axis Ultra, Kratos Analytical). Samples of the test solutions were collected every 3 d during the accelerated life test (ALT) for analysis of the dissolved elements (Ir and Ta) by inductively coupled plasma–mass spectrometry (ICP–MS). The resistance of the anode is tested by a four-probe tester (FT-341SJB).

2.3 Electrochemical measurements

An electrochemical workstation (CHI 660 E) was used. The electrolyte was 100 g/L aqueous ammonium citrate. Saturated Ag/AgCl, platinum, and the prepared IrO₂–Ta₂O₅ were used as the reference, counter, and working electrodes, respectively. Cyclic voltammetry (CV) was carried out over the potential range of 0.2–1.0 V at 20 mV s⁻¹. The temperature was maintained at 25 °C.

2.4 Accelerated life tests

By considering that the practical lifetime of this type of anode could be several years, ALTs were carried out to reveal the durability of the anodes (Comninellis and Vercesi 1991). The tests were performed in aqueous ammonium citrate at a current density of 4 kA/m² under galvanostatic conditions. The temperature was maintained at 30 °C. The cell voltage was monitored. The electrode was considered to deactivate when the cell voltage reached 15 V. The lifetime was then recorded as the time until deactivation. For each sample, one more test was performed on an identical anode at the same time to evaluate the reproducibility (Figure 1).

3.1 Cell voltage as a function of electrolysis time

Figure 2 shows the variation of the cell voltage as a function of electrolysis time during the ALT. The anode was placed in aqueous ammonium citrate and sulfuric acid. The failure behavior of the anode in aqueous sulfuric acid is as in previous research (Hu et al. 2002; Xin et al. 2010); the entire failure process can be divided into three stages: active, stable, and deactive. With the decrease of current density in sulfuric acid solution, the electrode life is extended. However, the anode failed rapidly in aqueous ammonium citrate. At the same current density, the electrode life in ammonium citrate solution is only one twentieth of that in sulfuric acid solution.

Figure 1:

Schematic of Ti/IrO₂–Ta₂O₅ anode preparation.

Figure 2:

The results of IrO₂–Ta₂O₅ anodes under different conditions during the ALT.

3.2 Crystalline phases as a function of electrolysis time

Figure 3 shows the XRD patterns of the Ti/IrO₂–Ta₂O₅ electrode as a function of time during the ALT. Diffraction peaks corresponding to metallic Ti and rutile IrO₂ are evident in the patterns of the original as well as electrolyzed anodes (Figure 3). No characteristic peaks for Ta₂O₅ were detected, indicating the absence of Ta₂O₅ crystallization during the thermal treatment of the coating. The rutile IrO₂ phase was observed with (110), (101), and (211) crystal faces. The intensity of the diffraction peaks of rutile IrO₂ decreased with increasing electrolysis time, but these peaks were observed in the failed anode. Thus, some residual coating remained on the anode after deactivation. The peak corresponding to metallic Ti increased with increasing electrolysis time, suggesting that the exposed area of the Ti substrate increased with increasing dissolution of the oxide coating. As indicated in the literature (Hu et al. 2002), oxidation of the Ti substrate can occur and form a TiO₂ layer at the coating–substrate interface during electrolysis. However, the TiO₂ phase was not observed in the entire electrolysis period, in accordance with XRD analysis. This might be attributable to the compact structure of the IrO₂–Ta₂O₅ coating, which prevented the Ti substrate from oxidizing.

Figure 3:

XRD patterns of the Ti/IrO₂–Ta₂O₅ electrode as a function of electrolysis time.

3.3 Surface morphology as a function of electrolysis time

Figure 4 shows the variations of the surface morphology as a function of electrolysis time during the ALT. All of the surfaces exhibited a typical mud-crack morphological feature. Many needle-like crystals were dispersed all over the surface of the coating (Figure 4(a)). The size and quantity of the dispersed needle-like crystals decreased with increasing electrolysis time. After electrolysis for 216 h, the needle-like crystals were negligibly evident on the coating. Dissolution of the needle-like crystals decreased the quantity of active components and the active surface area for the OER, which might account for the slow increase in the potential in the ALT. Such needle-like crystals were not evident in the electrolyzed anodes because of chemical dissolution and abscission of the coating during the OER (Figure 4(c)–(j)). Increasing electrolysis time increased the quantity of pores and tracks (Alanazi et al. 2019; Hu et al. 2002; Xu et al. 2019). The porous structure of the electrolyzed coating facilitated access of the electrolyte and oxygen to the Ti substrate; resulting in less-conductive TiO₂ crystals. However, the quantity of pores and tracks did not increase. This might be attributable to the compact structure of the IrO₂–Ta₂O₅ coating, such that Ta₂O₅ does not react with the electrolyte.

Figure 4:

Variation of the surface morphology as a function of time during the ALT: (a), (b) original; (c), (d) t = 72 h; (e), (f) t = 144 h; (g), (h) t = 216 h; (i), (j) failed anode.

The main elemental composition of the coatings was Ti, O, Ir, and Ta (Table 1). The fresh anode atom ratio of Ir to Ta was ca. 6:4. After electrolysis for 72 h, the atom ratio of Ir to Ta was ca. 5.3:4.7. With increasing electrolysis time, the atom ratio of Ir to Ta of the coating gradually decreased, and finally stabilized at 3:7. During electrolysis, Ir reacted with the electrolyte and IrO₂ was gradually dissolved, such that the ratio of Ir to Ta of the coating was gradually reduced (Figure 4).

Figure 5 shows the EDS mapping of the surface of the original and failed anodes. Elemental Ir, Ta, Cl, and O were uniformly dispersed all over the surface of the original oxide coating (Figure 5(a)) while there was little elemental Ti on the surface. Some elemental Ir remained on the surface of the failed anode (Figure 5(b)). Some elemental Ti (the Ti substrate) was observed on the surface of the failed anode, detected by EDS because the coating was partially dissolved. Figure 6 shows the morphology of a cross section of the failed anode. The cross section was divided into two parts: the inside was comparatively denser with IrO₂ crystals and the outside was comparatively less dense. The thickness of Ir on the outside of the coating was ca. 3–4 µm, and the ratio of Ir to Ta was 3:7 (area 1). There was no corrosion inside the coating, and the ratio of Ir to Ta was 6:4 (area 2). In addition, no passivation layer of nonconductive TiO₂ was found in the cross section between the coating and substrate.

Figure 5:

EDS mapping of the (a) original and (b) failed anodes.

Figure 6:

SEM and EDS of a cross section of a failed anode.

3.4 Electrochemically active surface area

Figure 7 shows CV curves of the prepared sample as a function of electrolysis time. The fresh anode exhibited a shape typical of thermally decomposed anodes (Shao et al. 2015). With increasing electrolysis time, the shape of the CV curve was almost unchanged. The peak current decreased with increasing electrolysis time. The apparent change in the voltammetric response is related to the variation of the active surface area. Because of the dissolution of IrO₂, the active component was reduced and, therefore, the catalytic activity was reduced. The voltammetric charges were calculated by integrating the CV curves. This charge was considered to proportional to the number of electroactive sites, which corresponds to the electrochemically active surface area of the anode. Overall, the charge decreased with decreasing coating loading (Figure 7(b)).

Figure 7:

Electrochemical performance of the Ti/IrO₂–Ta₂O₅ anode as a function of time during the ALT: (a) in aqueous ammonium citrate and (b) in aqueous sulfuric acid.

3.5 Degradation mechanism

During the ALT, samples (5 mL) were collected every 2 d for ICP–MS analysis of the dissolved elements. Figure 8 shows the concentration of Ir in solution as a function of the ALT time vs. time in 100 g/L aqueous ammonium citrate. Figure 8 also shows the cell voltage as a function of time.

Figure 8:

Concentration of Ir in solution and the cell voltage as a function of time for the ALT in 100 g/L aqueous ammonium citrate.

The concentration of Ir increased in the first 4 days of the ALT, but afterward, the growth rate gradually decreased. In addition, Ta was not detected in the solution samples. Interestingly, throughout the ALT, ammonium citrate dissolved Ir but Ta did not react. As a result, the rate of Ir dissolution gradually decreased throughout the procedure.

To explain the aforementioned phenomenon, potential–pH diagrams were constructed to predict the dominant Ir and Ta species that are likely to be present when the coatings are exposed to ammonium citrate at various anodic potentials (Wang et al. 2020). Solid IrO₂ is the thermodynamically favorable species when the potential is >0.5 V at pH 7 (Figure 9(a)). When the potential increases to >1.5 V, oxidation of IrO₂ to IrO₄ ²⁻ is thermodynamically favorable. Ta₂O₅ is the thermodynamically favorable species when the potential is >−1 V at pH 7 (Figure 9(b)).

Figure 9:

Pourbaix diagrams of Ir (a) and Ta (b) systems in aqueous ammonium citrate.

Based on the chemistry of the system and the constructed Pourbaix diagrams, it is reasonable to conclude that the following reactions occurred during the ALT tests.

Figure 10 shows a schematic for the corrosion mechanism of Ti/IrO₂–Ta₂O₅ electrodes. As aforementioned, the stable stage is the main stage of the entire electrolysis. In this stage, the active component loaded on the coating is dissolved into the electrolyte. During this stage, the crystals are negligibly evident on the coating and the cell voltage slowly increases. The slow increase in the cell voltage is attributable to rapid dissolution of IrO₂ and the reduction of the active surface area for the OER. A sharp increase in the cell voltage is observed in the sharp deactivation stage. In this stage, a large extent of dissolution of IrO₂ results in the ratio of Ir to Ta in the coating reaching a critical value, and the conductivity of the coating decreases (the resistance of the original and failed anodes is 0.63 and 0.79 mΩ, respectively). Consequently, the synergistic effect of the substantial dissolution of the active components is the main reason for the failure of the Ti/IrO₂–Ta₂O₅ anode in aqueous ammonium citrate.

Figure 10:

Schematic for the corrosion mechanism of Ti/IrO₂–Ta₂O₅ electrodes in aqueous ammonium citrate.

3.6 XPS analysis

Figure 11 shows XPS spectra of the prepared sample as a function of electrolysis time. Tables 2 and 3 show the binding energy values and intensities, respectively, obtained from the height of the Ir 4f, Ta 4f, and O 1s peaks. The 4f region consisted of two distinguishable peaks, Ir 4f _7/2 and Ir 4f _5/2, with a spin–orbit splitting of 3.0 eV. Typically, the 4f region of rutile IrO₂ is asymmetric because of its conductive nature (Wertheim and Guggenheim 1980). Ta 4f _7/2 was located at a binding energy of ca. 26.1 eV for all samples (Figure 10(b)), which corresponds to Ta⁵⁺ species. This value is slightly lower than 26.4 to 26.9 typically found for Ta₂O₅ in the literature, which can be ascribed to the nature of Ta₂O₅ in this IrO₂–Ta₂O₅ oxide mixture (Xu et al. 2010). These results are in good agreement with the aforementioned XRD results. Furthermore, Figure 10(c) shows O 1s spectra of the coatings, indicating lattice oxygen (530 eV), hydroxyl groups and/or oxygen defects (531 eV), and adsorbed water (533 eV) (Wei et al. 2020). All five anodes had these three species in their coatings but with slight changes in binding energy, possibly because of the difference in Ir/Ta ratio. The intensity of the 4f _5/2 and 4f _7/2 peaks corresponding to Ir decreased during the ALT, suggesting that Ir was removed from the solid. Regarding Ta, the intensity of these peaks decreased only slightly. These results are in good agreement with the aforementioned Pourbaix diagram results.

Figure 11:

XPS spectra of Ir 4f (a), Ta 4f (b), and O 1s (c) levels as a function of time during the ALT.