Yuanfang
Ying
a,
Jose Fernando
Godínez Salomón
b,
Luis
Lartundo-Rojas
c,
Ashley
Moreno
b,
Robert
Meyer
b,
Craig A.
Damin
b and
Christopher P.
Rhodes
*ab
aMaterials Science, Engineering and Commercialization Program, Texas State University, San Marcos, TX 78666, USA. E-mail: cprhodes@txstate.edu
bDepartment of Chemistry and Biochemistry, Texas State University, San Marcos, TX 78666, USA
cInstituto Politécnico Nacional, Centro de Nanociencias y Micro y Nanotecnologías, UPALM, Zacatenco, CP 07738, Ciudad de México, Mexico
First published on 5th February 2021
Acidic oxygen evolution reaction (OER) electrocatalysts that have high activity, extended durability, and lower costs are needed to further the development and wide-scale adoption of proton-exchange membrane electrolyzers. In this work, we report hydrous cobalt–iridium oxide two-dimensional (2D) nanoframes exhibit higher oxygen evolution activity and similar stability compared with commercial IrO2; however, the bimetallic Co–Ir catalyst undergoes a significantly different degradation process compared with the monometallic IrO2 catalyst. The bimetallic Co–Ir 2D nanoframes consist of interconnected Co–Ir alloy domains within an unsupported, carbon-free, porous nanostructure that allows three-dimensional molecular access to the catalytically active surface sites. After electrochemical conditioning within the OER potential range, the predominately bimetallic alloy surface transforms to an oxide/hydroxide surface. Oxygen evolution activities determined using a rotating disk electrode configuration show that the hydrous Co–Ir oxide nanoframes provide 17 times higher OER mass activity and 18 times higher specific activity compared to commercial IrO2. The higher OER activities of the hydrous Co–Ir nanoframes are attributed to the presence of highly active surface iridium hydroxide groups. The accelerated durability testing of IrO2 resulted in lowering of the specific activity and partial dissolution of Ir. In contrast, the durability testing of hydrous Co–Ir oxide nanoframes resulted in the combination of a higher Ir dissolution rate, an increase in the relative contribution of surface iridium hydroxide groups and an increase in specific activity. The understanding of the differences in degradation processes between bimetallic and monometallic catalysts furthers our ability to design high activity and stability acidic OER electrocatalysts.
To lower the Ir content, improvements in the activity and stability of Ir-based acidic OER catalysts are needed. A number of approaches have been explored including lowering the Ir loading on the electrode,9 using supported Ir catalysts (e.g. TiO2-supported, SnO2-supported Ir, etc.),10,11 and increasing the activity of Ir using through substitution with non-precious metals (e.g. Ni, Co, Cu, etc.).12–14 Prior density functional theory (DFT) calculations showed that IrO2 binds adsorbed oxygen too strongly which contributes to increasing the overpotential for the OER.15 A prior experimental study showed that Ni substituted within IrO2 resulted in increased OER activity attributed to leaching of Ni atoms, which promote the formation of structurally flexible, reactive OH groups that act as reactive surface intermediates for the OER catalytic reaction.16 Other work reported that the substitution of Cu within IrO2 creates oxygen vacancies and changes in the electron occupancy of the t2g and eg orbital states, resulting in lower overpotentials.17
In addition to increasing activity, the stability of OER catalyst remains a significant challenge1,7,18–20 since all known OER catalysts with reasonable activity (including Ir) dissolve under the highly acidic and oxidative potentials required for oxygen evolution.7,18,19 Within substituted IrO2-based catalysts, stability remains a major challenge since many of the metals such as Ni, Co and Cu are thermodynamically unstable in acid at high potentials, as described by Pourbaix diagrams21 and dissolve under OER testing conditions. Another factor affecting activity and stability is the surface structure. Prior work supports that there are significant surface structure, activity and stability differences between thermally prepared iridium oxides (IrO2) and amorphous, electrochemically generated iridium oxides (IrOx).19 The predominant current understanding of acidic OER catalysts is that activity and stability are inherently linked since they share a common intermediate.20,22 Prior work on noble metal monometallic oxides established a direct relationship between the high activity and low stability of acidic OER catalysts;5 however, others' work on RuO2 catalysts reports no correlation between activity and stability.23 Recent work suggests that it may be possible to obtain catalysts that are not restricted by the traditional inverse relationship between OER activity and stability.24
Cobalt is a potentially promising substituent within IrO2 for improving OER activity since doping of Co within IrO2 was shown from DFT calculations to alter the electron density and influence changes in the binding energies of intermediates resulting in structures with lower activation energies.25 A number of prior studies have evaluated developing improved acidic OER catalysts through interacting iridium and cobalt within iridium–cobalt nanowires,3 iridium–cobalt porous nanocrystals,26 iridium–cobalt oxide,27,28 IrO2–Co3O4 nanorods,29 and cobalt-doped SrIrO3.30 Nanoframe structures that combine iridium with Pt31 or Ni and Cu32 have been previously studied to enhance the activity and stability of oxygen evolution electrocatalysts. The catalyst structure and resulting OER activity and stability are highly dependent on the starting materials and chemical and electrochemical processes used to generate the active catalyst.33
Our group has developed bimetallic two-dimensional (2D) nanoframes that combine highly active bimetallic surface sites and a porous, carbon-free nanoarchitecture to provide oxygen electrocatalysts with improved activity and stability.14,34,35 The 2D nanoframe structure is synthesized from thermal reduction of precious metal-decorated metal oxide nanosheets and is composed of a unique hierarchical 2D framework containing interconnected solid metallic alloy domains within a porous matrix that provides 3D molecular accessibility. We recently reported unsupported, hydrous iridium–nickel oxide 2D nanoframes exhibit 14 times higher OER mass activity than commercial IrO2.14 In this prior study, we used interaction of Ir with Ni within the structure to significantly enhance OER activity and the porous unsupported electrocatalyst structure to provide access to the active sites and avoid support degradation and nanoparticle aggregation. However, the stability of the hydrous iridium–nickel oxide nanoframes was substantially lower than IrO2. We recently reported the combination of CoIr and NiPt 2D nanoframes function as highly efficient bifunctional oxygen reduction/evolution electrocatalysts;35 however, our recent study did not explore CoIr nanoframes separately as oxygen evolution catalysts. In this work, we report the morphological and structural characterization of CoIr nanoframes and precursor materials, electrochemical and spectroscopic characterization of the surface structure, and evaluation of the oxygen evolution activity and stability compared with commercial IrO2 to provide insight into how cobalt interacting with iridium within a 2D nanoframe structure affects OER activity and stability.
To prepare the Co(OH)2 nanosheets, 1.05 g of Co(NO3)2·6H2O (98%, Alfa Aesar) was combined with 0.6486 g of urea (99.3%, Alfa Aesar) in 3 mL of ultrapure water (≥18 MΩ cm, in-house water purification system, Purelab Classic, Evoqua Water Technologies) and 21 mL of ethylene glycol (99%, VWR). The solution was then transferred into a 35 mL microwave vial. The reaction was conducted in a Discover SP Microwave reactor under variable microwave radiation power with a controlled temperature of 120 °C for 10 min under active magnetic stirring. The powder was separated by centrifugation (Sorvall ST16, Thermo) at 3000 RPM for 3 min, rinsed five times with ultrapure water and two times with isopropanol (HPLC grade, VWR), and then dried under ambient atmosphere at 60 °C.
To synthesize iridium-decorated Co(OH)2 nanosheets (Co(OH)2:Ir), 0.038 g of IrCl3·xH2O (99.8%, Alfa Aesar) (equivalent to 20 wt% of Ir vs. Co(OH)2) was dissolved in 25 mL of ultrapure water, and 0.1 g of Co(OH)2 was added. The suspension was then transferred into a 35 mL microwave vial. The reaction was conducted in a microwave reactor under variable microwave radiation power with a controlled temperature of 120 °C for 10 min with active stirring and then separated using an identical reactor and conditions as described above. The Co(OH)2:Ir nanosheets were thermally treated within a muffle furnace (Lindberg Blue M, Thermo Scientific) at 300 °C for 20 minutes using a ramp rate of 10 °C min−1 starting from room temperature and under 120 mL min−1 flowing H2/Ar (5/95 vol%). Then, the reactor was removed from the oven and allowed to cool under H2/Ar flow until reaching room temperature. The synthesized sample after thermal treatment is notated as Co–Ir. CoIr-CL was prepared by chemically leaching of Co–Ir using a method previously reported by our group.14,35 Briefly, 0.2 g of the as-prepared Co–Ir nanoframes were dispersed in 100 mL of 0.05 M HNO3 which was diluted from 70% HNO3 (99.999%, Aldrich). The solution was bubbled with flowing Ar for 20 min and then heated to 80 °C for 2 hours under stirring. The Ar flow was maintained until the reaction completed. After the reaction, the suspension was cooled to room temperature. The solid was separated by centrifugation at 5000 RPM for 10 min, rinsed five times with ultrapure water, rinsed two times with isopropanol, and then dried at 60 °C. The samples are noted as Co(OH)2 (Co(OH)2 nanosheets), Co(OH)2:Ir (decorated Co(OH)2 nanosheets with Ir nanoparticles), Co–Ir (Co(OH)2:Ir treated at 300 °C under H2/Ar), and CoIr-CL (Co–Ir after chemical leaching).
X-ray photoelectron spectra were obtained using a Thermo Fischer Scientific K-Alpha X-ray photoelectron spectrometer with a monochromatic Al Kα X-ray source, 1486.6 eV, microfocused at the source to give a spot size on the sample of 400 microns in diameter. Samples remained under vacuum for more than 10 h in a prechamber directly connected to the equipment and were then transferred to the analysis chamber with a base pressure of 1 × 10−9 Torr that remained constant during the experiment. Survey and high-resolution spectra were collected using an analyzer, operated in Constant Analyser Energy mode (CAE), with pass energies of 200 and 10 eV, respectively. In order to compensate effects related to charge shift, the Ir4f7/2 metal peak at 60.2 eV and the C1s adventitious carbon peak at 284.6 eV were used as internal standards. Data analysis was performed using AVANTAGE v5.91 software (Thermo Fisher Scientific) using a Shirley-type background subtraction and a pseudo-Voigt function with Gaussian (70%)–Lorentzian (30%) for each component. XPS analysis was carried out from two independent batches of each catalyst material to determine any significant changes in the number and content of chemical elements between batches of the same material. In each lot, survey and high-resolution spectra were obtained from three different zones with diameters of 400 microns. From the examination of the survey spectra, it was established that among batches of the same sample the chemical elements present were the same and that these did not present significant changes, ≤1.0, in their atomic percentage content. On the other hand, the high resolution spectra for the Ir5p–Ir4f, O1s and Co2p regions suggested that between batches of the same sample, there were some changes in the atomic percentage content but not in the number of chemical species formed during the synthesis protocol (Fig. S6 and S7†). In view of these results, it was established that both catalytic systems have a homogeneous elemental and chemical species composition. Therefore, for each material average survey and high-resolution spectra from the whole set of data obtained were generated for further analysis. XPS spectra and quantities presented correspond to an average of six measurements from different points of each sample batch.
Attenuated total reflectance Fourier-transform infrared (ATR-FT-IR) spectra of the Co(OH)2 and Co–Ir samples were collected using a Harrick Scientific (Pleasantville, NY) SplitPea ATR microsampling accessory coupled to a Bruker (Billerica, MA) Tensor II FT-IR spectrometer. The SplitPea accessory was equipped with a silicon internal reflection element and utilized a liquid nitrogen cooled mercury–cadmium–telluride detector coupled to the Tensor II spectrometer. The solid samples were brought into direct contact with the silicon internal reflection element using a 0.5 kg loading. Infrared spectra represent the average of 64 individual scans with 4 cm−1 resolution. Transmission FT-IR measurements were obtained within the instrument described above using potassium bromide (KBr) pellets prepared by combining the sample with dried KBr powder.
The electrode fabrication and RDE testing protocols were based on methods previously reported by our group14,35 and other groups.36 Catalyst inks were prepared by combining a specific mass (typically ∼3 mg) of the acid-leached catalysts (CoIr-CL) with a specific volume of a stock solution to yield a catalyst concentration of 0.43 mgcat mL−1. The stock solution was prepared by mixing 0.4 mL of Nafion suspension (Aldrich, 5 wt%, 1100 g equivalent weight), 20 mL of isopropanol, and 79.6 mL ultrapure water (≥18 MΩ cm). The inks were sonicated (Fisher, 40 kHz) in an ice-bath for 20 minutes. The ink was then immersed in a controlled temperature bath (25 °C) for 1 minute while maintaining agitation and was then used immediately. The inks were applied to a polished Au working electrode (0.196 cm2 geometric area) and allowed to dry under rotation (700 rpm) under ambient conditions. The electrode loading was controlled by depositing a specific volume of the ink onto the Au working electrode. For the CoIr-CL catalysts, 10 uL was deposited onto the Au electrode which corresponds to a loading of 15.3 μg Ir cmgeo−2. After depositing the ink on the rotating Au electrode, the ink was maintained under rotation and allowed to dry under ambient conditions.
The electrochemical characterization of the catalysts was carried out in 0.1 M HClO4 electrolyte prepared with 70% HClO4 (Veritas Doubly Distilled, GFS Chemicals) (0.000001% Cl−) and ultrapure water. For the electrochemical tests, the Au working electrode was placed in a three-electrode cell with the 0.1 M HClO4 electrolyte. The electrodes were immersed in argon saturated 0.1 M HClO4 under potential control (0.1 VRHE). The catalysts were first conditioned by cycling 20 times from 0.05–1.0 VRHE at 100 mV s−1. The electrochemical surface area (ECSA) of metallic Ir was calculated by CO-stripping using a specific charge of 358 μC cm−2 as the charge corresponding to a monolayer of adsorbed CO.36 Prior to the analysis in the OER potential range, the catalysts were electrochemically conditioned by cycling between 0.05 to 1.5 VRHE for 60 cycles at a scan rate of 100 mV s−1 in argon-saturated 0.1 M HClO4, which is designated as the electrochemical oxidation (EO) step. The electrochemical surface area of IrO2 (ECSAIrO2) was then determined by measuring the pseudocapacitive charge between 0.3 V and 1.25 V obtained using a scan rate of 50 mV s−1 and by subtracting the contribution of the Au disk current collector. The electrochemical surface area was calculated based on the background subtracted pseudocapacitive charge and the coulombic conversion factor of 596 μC cmIrO2−2.37
For evaluating the oxygen evolution reaction activity, the electrode was then conditioned by cycling 10 times in the potential range 1.2–1.8 VRHE at 100 mV s−1 and 2500 rpm. Linear sweep voltammetry (LSV) in the potential range of 1.2–1.8 VRHE was performed using a scan rate of 20 mV s−1 and a rotation rate of 2500 rpm. Steady-state (iR-corrected) chronoamperometric measurements were then performed by stepping the potential at steps of 0.01 V from 1.3 to 1.6 VRHE while holding for 5 seconds at each potential and rotating the working electrode at 2500 rpm. The internal resistance (iR) values (23–27 Ω) was determined prior to every evaluation using the current interruption method at 1.6 VRHE. The data obtained from chronoamperometric measurements was used for the Tafel plots and for determining the OER mass activity and specific activity. The mass-normalized OER activity was determined by the current at a specific voltage from the chronoamperometric measurements divided by the Ir mass on the electrode. The Ir mass on the electrode was determined by the mass loading and the Ir content within the material determined by inductively coupled plasma mass spectrometry (ICP-MS) analysis for CoIr-CL and for IrO2 from the certificate of analysis reported by the vendor (84.5 wt%) which was also corroborated by EDS measurements. The actual Ir mass after EO and after accelerated durability testing (ADT) was determined by subtracting the Ir leached into the electrolyte from the initial Ir mass on the electrode. The percent of Ir leached into the solution was determined by ICP-MS after each procedure, EO and ADT, for CoIr-CL and ADT for IrO2. Inspection of the Tafel plots for linear behavior within specific voltage ranges was used to evaluate that the electrode was not within a mass-transport limited regime within the specific voltage range. Tafel slopes were determined using current obtained from chronoamperometry measurements using the iR-corrected potential between 1.47 V to 1.52 VRHE. After the evaluation of the electrocatalyst in the OER potential range, CV measurements were again obtained to determine the nature of the surface after exposure to OER potentials of up to 1.8 VRHE.
Following the measurements of the OER activity, an accelerated durability test was carried out by maintaining the electrode at a constant potential of 1.6 VRHE for 13.5 hours under a rotation rate of 2500 rpm. After the constant potential step was completed, the electrolyte was replaced, and a 15 minute argon purge was performed to help remove entrapped oxygen bubbles within the catalyst layer which can contribute the influencing the oxygen evolution reaction current using RDE measurements.38 CV, CO stripping, LSV, and chronoamperometry measurements were then obtained as described above. ICP-MS measurements (PerkinElmer NexIon 2000, Washington University in Saint Louis) were used to analyze the amount of iridium dissolved within the electrolyte after the electrochemical oxidation step and after the accelerated durability test.
The changes in the relative elemental composition and distribution of elements within the materials through the synthesis steps were followed by EDS (Fig. 2). In general, the as-prepared, intermediate materials, and final catalyst showed a homogeneous distribution of Co and Ir within the structures. The Co(OH)2:Ir material contained 23.1 ± 3.5 wt% of Ir, very close to the nominal weight ratio determined from weights of the precursors used for the synthesis (i.e. 20 wt%), which demonstrates good reaction efficiency. After thermal treatment of Co(OH)2:Ir at 300 °C under H2/Ar, within Co–Ir the elemental composition of Ir and Co increased to 36.0 ± 1.8 wt% iridium and 46.0 ± 2.9 wt% cobalt and the oxygen content decreased to 17.3 ± 4.3 wt%, which is attributed primarily to the removal of water from the structure as a product of the reaction. After the acid leaching step which selectively removed Co species, the CoIr-CL sample had a content of 88.3 ± 1.5 wt% iridium, 5.2 ± 0.6 wt% cobalt and 6.3 ± 2.0 wt% oxygen. The elemental composition of CoIr-CL determined from EDS was similar to the elemental composition determined by inductively coupled plasma atomic emission spectroscopy (ICP-AES), which yielded a content of 83 ± 2.7 wt% Ir and 9.3 ± 1.0 wt% Co for CoIr-CL. Although metallic Co is thermodynamically unstable in acid,21 we considered that Co within CoIr-CL could be stabilized from dissolution in acidic electrolyte through interaction of Co with Ir as supported by prior studies of Ni stabilization within IrO2.12,14
![]() | ||
Fig. 3 Powder X-ray diffraction (XRD) patterns of Co(OH)2, Co(OH)2:Ir, Co–Ir, CoIr-CL and patterns for references (α-Co(OH)2, Co3O4, Co, and Ir). |
After Ir deposition, the XRD pattern of Co(OH)2:Ir showed the strong peak associated with the (001) basal plane of Co(OH)2 disappeared, while small peaks around 18° and 59° associated with a Co3O4 phase were present. The disappearance of the (001) peak within Co(OH)2:Ir is attributed to the disruption of the basal plane which may be influenced by the creation of pores within the structure as observed by microscopy (Fig. S1†). The deposition of Ir onto the Co(OH)2 nanosheets may result in the displacement of oxygen by Ir species due to galvanic displacement which modifies the oxygen stoichiometry due to oxygen depletion28 or may occur via ethylene glycol reduction.42,43
After thermal treatment of Co(OH)2:Ir under hydrogen, the XRD patterns of Co–Ir showed significant morphological and structural changes. The XRD pattern of Co–Ir showed the presence of peaks consistent with Co3O4 (at ∼18°, 32°, 37° and 65°).39 The presence of Co3O4 within Co–Ir was also supported by infrared spectroscopic analysis (described below). The XRD pattern of Co–Ir showed wide peaks between 40° to 50° which may be associated with low intensity (111) reflections of small crystalline domains and/or microstrain from a metallic face-centered cubic CoIr structure.44,45 The presence of a poorly crystalline metallic Co–Ir phase is supported by the magnetic response of Co–Ir to a lab magnet; however, further work is needed to confirm the metallic character. We note that the XRD pattern of Co–Ir prepared by thermal treatment at 300 °C in H2/Ar was significantly different than the XRD previously reported of NiIr treated under identical conditions.14 In the case of NiIr, we observed the presence of peaks that correspond to a metallic NiIr phase and no NiO peaks; in contrast, for CoIr, we did not observe defined peaks corresponding to metallic CoIr phase and did observe peaks consistent with Co3O4. The differences in the thermal reduction of Ni(OH)2:Ir and Co(OH)2:Ir to NiIr and CoIr respectively suggests that oxygen binding to Co may be stronger than oxygen binding to Ni within the structure.
After the acidic treatment of Co–Ir, most of the Co species that were not stabilized within the Ir structure were removed as shown from EDS analysis (Fig. 2). The XRD pattern of CoIr-CL shows peaks at 41.2° and 46.9° that are associated with the (111) planes (2θ = 40.6°) and (200) planes (2θ = 47.3°) of metallic Ir. The peak position of the Ir(111) reflection is shifted to higher 2θ values compared to the characteristic position of a metallic Ir lattice. A lattice constant of 3.771 Å was calculated from the peak position of the (111) plane using a face-centered cubic cell. The lattice constant of 3.771 Å for the CoIr-CL material is smaller than the standard metallic iridium lattice constant of 3.831 Å, indicating a ∼1.6% lattice contraction. These observations support a lattice disruption of iridium likely due to the inclusion of residual Co within the Ir structure in agreement with STEM data, as presented below. In addition, very low intensity peaks around 36° and 60° assigned to trace Co3O4 within the structure are still observed.
A prior study of α-hydroxides, including those of cobalt and nickel, supports that the materials generally contain a significant amount of intercalated anions.61 Within the ATR-FT-IR spectrum of Co(OH)2 nanosheets, Fig. 4a, surface and intercalated nitrates were assigned to the absorptions at 1381, 1345, 997, and 829 cm−1.50,52,56,57,60,63,64 The nitrate absorptions at 1381, 1345, and 829 cm−1 correspond to free NO3− ions with D3h point group symmetry.48,53,67 Upon coordination to a metal, the local symmetry of the NO3− is lowered to a C2v symmetry.68,69 The remaining nitrate absorption at 997 cm−1 is assigned to coordinated NO3− ions in the Co(OH)2 structure with C2v symmetry.65 The presence of nitrate absorptions in the ATR-FT-IR spectra of the as-prepared Co(OH)2 nanosheets suggests that NO3− from Co(NO3)2·6H2O is carried through the synthesis procedure and exists in two different environments within the material.64 The ν1 mode of NO3− is usually IR-inactive; however, the presence of intercalated water between the Co(OH)2 nanosheets could result in a NO3− environment in which the ν1 mode of NO3− becomes weakly IR-active.64 The presence of free CO32− ions with D3h point group symmetry is indicated by absorptions at 1478 and 889 cm−1.50,52,68 The presence of carbonate within the sample could result from the dissolution of CO2 in the aqueous solutions prior to urea hydrolysis. In addition, carbonate anions could be generated by the decomposition of urea, as previously proposed.47
The ATR-FT-IR spectrum of Co(OH)2:Ir is shown in Fig. 4b. Following iridium deposition onto the Co(OH)2 nanosheets, changes to the Co(OH)2 spectrum are observed. The δ(Co–O–H) peak is shifted to 632 cm−1 while a new peak with greater relative intensity appears at 510 cm−1 in the Co(OH)2:Ir spectrum. Bands for carbonate and nitrate ν3 modes were blue shifted to 1495 and 1365 cm−1 respectively. Absorptions at 3417 and 1627 cm−1 indicate the presence of surface or interlayer water. The Co–O–H deformation and Co–O stretch of Co(OH)2 are assigned to the absorptions at 633 and 401 cm−1, respectively. The absorption at 515 cm−1 is in the range of bands observed for ν(Ir–O)70 and ν(Co–O)71 modes; and further analysis is needed to determine the specific assignment of this band within Co(OH)2:Ir.
The remaining absorptions in the Co(OH)2:Ir spectrum are associated with surface and interlayer species that are present in the as-prepared Co(OH)2 precursor. The 2201 cm−1 absorption is attributed to the asymmetric OCN− stretch. The presence of free NO3− ions was determined based on the 1354 and 828 cm−1 absorptions. The absorption at 1493 cm−1 is assigned to the ν3 mode of CO32−.52 The O–H deformation and C–O stretching modes at 1303 and 1040 cm−1, respectively,66 support the presence of residual ethylene glycol within Co(OH)2:Ir. The nitrogen species band is blue shifted to 2200 cm−1 and is greatly reduced in intensity following iridium deposition. Broadening and red shifting of the ν(O–H) band, as well as increased relative intensity of the δ(H2O) band, suggest incorporation of more water molecules into the structure during the deposition process. Broadening and shifting of the ν(O–H) bands is consistent with increased hydrogen bonding within the structure.54 The reduction in the relative intensity of bands arising from ethylene glycol between the spectrum of Co(OH)2 and Co(OH)2:Ir suggests that it is possible that ethylene glycol may act as a reducing agent to reduce Ir3+ to Ir during the deposition process as supported by prior work on formation of colloidal Ir and other platinum-group metal nanoparticles,42,43 or alternatively ethylene glycol is removed during the Ir deposition process; however, further analysis is needed to determine the potential role of ethylene glycol during the deposition process.
Fig. 4c shows the ATR-FT-IR spectrum of a Co–Ir sample following a thermal treatment at 300 °C of Co(OH)2:Ir for 20 minutes. The ATR spectrum of the resulting black solid indicated that the sample strongly absorbed the infrared radiation and possessed a highly metallic character. Dehydration of the sample was apparent by the lack of water related absorptions. The effect of scattering of the infrared radiation by the Co–Ir sample is apparent by the sloping baseline. For strongly absorbing samples, light scattering within the sample causes the baseline to slope upwards with increasing wavenumbers.72 The ATR spectrum of the CoIr-CL sample following chemical leaching of the cobalt, Fig. 4d, also indicated the presence of a highly metallic phase.70
Since the ATR spectra of Co–Ir and CoIr-CL did not show clearly resolved bands relative to the baseline consistent with their metallic character as described above, transmission FT-IR measurements of samples within potassium bromide (KBr) pellets were made. Shown in Fig. 4e and (f) are the transmission FT-IR spectra of Co–Ir and CoIr-CL, respectively. The FT-IR spectrum of Co–Ir (Fig. 4e) has only two defined peaks at 670 and 578 cm−1 whose wavenumbers are generally consistent with the positions of bands from Co3O4 modes based on a prior study that reported two distinct and sharp bands at 568 (ν1) and 664 (ν2) cm−1 which originate from the stretching vibrations of the Co–O bonds within spinel Co3O4.71 The ν1 band is characteristic of a Co3+–O vibration in an octahedral site, and the ν2 band is attributed to a Co2+–O vibration in a tetrahedral site in the spinel lattice.73 Our X-ray diffraction data showed the presence of peaks consistent with Co3O4. It is also possible that the 578 cm−1 band originated from a Ir–O stretching mode based on a prior study;70 however, our X-ray diffraction data did not show the presence of any IrOx phase.
The high-resolution HAADF-STEM image (Fig. 5B) and fast Fourier transform (FFT) pattern (Fig. 5D) demonstrate the single-crystalline nature of domains within the material. The intensity profile (Fig. 4C) shows an average lattice spacing of 0.213 ± 0.009 nm corresponding to the (111) lattice plane of bimetallic CoIr-CL, according to a face-centered cubic (fcc) structure identified by XRD (Fig. 3), and confirmed a lattice compression with respect to the metallic iridium pattern taken as the standard (0.216 nm).35 The HAADF-STEM and corresponding EDS mapping images of CoIr-CL (Fig. S2†) show uniformly dispersed Co and Ir, consistent with the XRD data. The elemental composition of Ir, Co, and O from HAADF-STEM EDS is in a similar range as values obtained from SEM EDS analysis (Fig. 2).
Semi-quantitative analysis of the XPS average survey spectra was used to estimate the atomic composition within the surface region (Table 1). Analysis of the survey spectra of CoIr-CL and IrO2 supports that a higher degree of oxygen is present within the surface region of IrO2 (52.3 atomic% oxygen) compared to CoIr-CL (33.3 atomic% oxygen). The weight percentages of Ir, Co and O within the surface region determined from XPS are in general agreement with values obtained from SEM-EDS analysis (Fig. 2I–L); however, XPS analysis showed a slightly lower surface concentration of Ir and a slightly higher concentration of O and Co species compared to SEM-EDS analysis (Table S1†) which may be due to differences in probing of the surface region between these methods.
![]() | ||
Fig. 6 High-resolution X-ray photoelectron spectra and fitted peaks of (A) Ir4f and (B) O1s regions of commercial IrO2 and CoIr-CL samples and (C) Co2p region of CoIr-CL; the intensities of Ir4f and O1s core level spectra were normalized; details of peak fitting analysis and label are discussed in the text and in Table 2. |
Region | Sample | BE [±0.1 eV] | FWHM [eV] | Areaa [at%] | Label | Assignment |
---|---|---|---|---|---|---|
a Average content of six points. | ||||||
Ir4f | IrO2 | ≈60.8 | 4.2 | 43.2 | Ir5p1/2 | Metal, oxide and hydroxide range |
61.4 | 1.1 | 29.8 | Ir1 | IrO2 | ||
62.3 | 1.2 | 14.2 | Ir2 | Ir2O3, IrCl3·xH2O, IrClx and/or IrCl3 | ||
62.6 | 1.2 | — | — | Ir(IV) Sat I | ||
66.6 | 1.3 | — | — | Ir(IV) Sat II | ||
63.5 | 1.3 | 12.8 | Ir3 | Hydrous IrO2, Ir(OH)3 | ||
64.7 | 1.4 | — | — | Ir(III) Sat | ||
CoIr-CL | 60.9 | 3.5 | 33.8 | Ir5p1/2 | Metal, oxide and hydroxide range | |
59.6 | 1.7 | 23.8 | Co3p | Metal, oxide and hydroxide range | ||
60.9 | 0.9 | 24.3 | Ir0 | Metallic iridium | ||
61.6 | 1.1 | 10.5 | Ir1 | IrO2 | ||
62.5 | 1.2 | 4.2 | Ir2 | Ir2O3, IrCl3·xH2O, IrClx and/or IrCl3 | ||
62.9 | 1.0 | — | — | Ir(IV) Sat I | ||
66.8 | 1.1 | — | — | Ir(IV) Sat II | ||
63.4 | 1.3 | 3.4 | Ir3 | Hydrous IrO2, Ir(OH)3 | ||
64.8 | 1.4 | — | — | Ir(III) Sat | ||
O1s | IrO2 | 529.8 | 1.55 | 32.4 | IrO2 | IrO2 |
530.8 | 1.56 | 27.9 | O1 | Ir2O3, Mx(CO3)y | ||
531.6 | 1.63 | 19.8 | O2 | Ir(OH)3, IrO(OH)2 | ||
532.4 | 1.70 | 13.7 | O3 | IrCl3·xH2O, ClO3−, C–OH, C–O–C | ||
533.6 | 1.78 | 5.0 | O4 | NO3−, O![]() |
||
534.9 | 1.85 | 1.1 | O5 | –COO–, O![]() |
||
CoIr-CL | 529.4 | 1.33 | 7.5 | CoOx | CoO, Co3O4 | |
529.9 | 1.34 | 25.8 | IrO2 | IrO2 | ||
530.9 | 1.40 | 23.9 | O1 | Ir2O3, Mx(CO3)y | ||
531.5 | 1.47 | 17.5 | O2 | Ir(OH)3, IrO(OH)2 | ||
532.3 | 1.53 | 14.5 | O3 | IrCl3·xH2O, ClO3−, C–OH, C–O–C | ||
533.6 | 1.60 | 8.9 | O4 | NO3−, O![]() |
||
535.2 | 1.66 | 1.9 | O5 | –COO–, O![]() |
||
Co2p | CoIr-CL | 776.9 | 2.75 | 7.1 | Co LMM | Co Auger L3M23M45 transition |
778.2 | 1.02 | 35.8 | Co1 | Metallic cobalt | ||
781.6 | 2.52 | — | — | 1st metallic cobalt satellite | ||
783.8 | 2.77 | — | — | 2nd metallic cobalt satellite | ||
779.1 | 1.29 | 21.8 | Co2 | Co3O4 | ||
781.3 | 1.67 | — | — | 1st Co3O4 multiplet | ||
782.4 | 1.77 | — | — | 2nd Co3O4 multiplet | ||
785.5 | 2.44 | — | — | 1st Co3O satellite | ||
789.5 | 2.91 | — | — | 2nd Co3O satellite | ||
780.2 | 1.80 | 35.3 | Co3 | Co(OH)2, CoOOH, CoO | ||
783.4 | 2.69 | — | — | 1st Co(OH)2 multiplet | ||
787.2 | 2.98 | — | — | 1st Co(OH)2 satellite | ||
791.8 | 3.23 | — | — | 2nd Co(OH)2 satellite |
Similar to as observed within our prior work on NiIr nanoframes,14 the CoIr-CL catalyst showed that the predominantly metallic nature of the surface of the electrocatalyst was altered after cycling to potentials required for OER, modifying the surface functional groups. Therefore, in order to have obtain a representative surface structure upon which the OER reaction takes place, an “electrochemical oxidation” step (notated with an “EO” subscript) comprising 60 scans between 0.05–1.5 VRHE, was carried out with the CoIr-CL and IrO2 catalysts. The catalyst after the EO step is labelled as CoIr-CLEO (Fig. 7A, blue curve). After electrochemical oxidation, the peaks due to Hupd features were no longer observed, instead, two broad anodic peaks around 0.27 VRHE and 0.82 VRHE assigned to iridium hydroxide/oxide surface features were present.83 The CVs of CoIr-CLEO and IrO2-EO are compared in Fig. 7C. Within IrO2,EO, the anodic peak at ∼0.48 VRHE (labeled A0) has been attributed to formation of iridium(III) hydroxide86 and modelled as involving the oxidation of two Ir3+–OH2 groups to two Ir4+–OH groups,87 while the peak at ∼0.81 V (labeled B0) has been attributed to oxidation of iridium(III) hydroxide to tetravalent IrO2 or IrO(OH)2.86,88 Compared with the CV of IrO2-EO, the CV of CoIr-CLEO showed the anodic peak, A0, at a significantly lower potential of ∼0.27 VRHE compared to the voltage of the peak at IrO2-EO (∼0.48 VRHE) which suggests different local chemical environments for hydrated Ir–OH species at the surface of CoIr-CLEO and IrO2-EO.
Since the surface of CoIr-CLEO was no longer metallic after the electrochemical oxidation step, we used pseudocapacitance measurements (Fig. S8†) and previously reported parameters and conversion factors37 to determine the ECSA of CoIr-CLEO (Table S5†). The ECSAIrO2 of CoIr-CLEO determined from pseudocapacitance measurements was calculated to be 30 ± 5 m2 g−1 which is much lower than the ECSAIr value determined from CO stripping (61.2 ± 3.1 m2 g−1) obtained before electrochemical oxidation, but very similar to the surface area estimated for IrO2-EO. The growth of the oxide/hydroxide surface species during the electrochemical oxidation step may result in lowering the surface area by growing within smaller pores and limiting the access of the electrolyte.74,89 Our previous study of NiIr nanoframes showed the presence of a ∼5 Å-thick oxide/hydroxide surface layer after the electrochemical oxidation step.14
Shown in Fig. 8A are the chronoamperometric polarization curves of CoIr-CLEO and IrO2,EO normalized versus the mass of Ir. The higher current observed for CoIr-CLEO within the OER chronoamperometric curve clearly shows the superior OER activity of CoIr-CLEO compared to IrO2,EO. The mass-normalized OER activities of the CoIr-CLEO, and IrO2-EO were compared at 1.51 VRHE and are shown in Fig. 8C. A potential of 1.51 VRHE was used since the Tafel plot (Fig. 8B) still shows linear behavior within this voltage region, and lower overpotential relative to the thermodynamic potential of the reaction also minimizes the contribution of mass transport. At a potential of 1.51 VRHE, the mass activity of CoIr-CLEO (243 ± 47 A gIr−1) is 17 times higher than the mass activity of IrO2-EO (13.9 ± 1.4 A gIr−1).
We determined the OER specific activity by normalizing the current to the electrochemical surface area determined from pseudocapacitance measurements. As shown in Fig. 8D, the OER specific activities at 1.51 VRHE is 0.80 ± 0.02 A cmIrO2−2 for CoIr-CLEO which is significantly (18 times) higher than the OER specific activity of 0.045 ± 0.005 A cmIrO2−2 for commercial IrO2 (IrO2-EO). The CoIr-CLEO catalyst showed a significant improvement in both mass and specific OER activity compared with the commercial IrO2-EO catalyst.
The improvement of the specific activity of CoIr-CLEO compared to IrO2-EO indicates that the catalytically active surface sites of CoIr-CLEO are more active for oxygen evolution. Our XPS measurements (Fig. 6) support the presence of multiple species within the surface region including iridium oxide, iridium oxyhydroxide, and iridium hydroxide with iridium in 4+ and 3+ oxidation states and the presence of cobalt oxides and hydroxides with cobalt in 2+ and 3+ oxidation states. However, our CV data (Fig. 7C) shows that the surface is changed after undergoing exposure to high oxidation potentials under OER conditions, and therefore the CV measurements are more representative of the OER catalyst surface. The CV of CoIr-CLEO (Fig. 7C) shows peaks consistent with the surface having Ir3+–OH and Ir4+O2 or Ir4+O(OH)2 groups. As discussed above, the potential of the oxidation peak at ∼0.27 VRHE, attributed to surface Ir3+–OH groups occurs at a lower potential compared to the peak potential of ∼0.48 VRHE for IrO2-EO. Prior work supports that iridium hydroxo (Ir–OH) surface species are strongly linked to OER activity,2 and Ir–OH groups can be considered as descriptors for OER activity.90 The higher surface activity of CoIr-CLEO compared to IrO2-EO is attributed to presence of highly active Ir-OH species that may interact with subsurface Co; however, additional analysis is needed to characterize the surface under reaction conditions. The OER in acid involves a complex, multi-step reaction with multiple intermediates (discussed below). A number of prior studies support that at OER potentials the oxidation state of iridium increases from 4+ to 5+,91–93 while other studies report only Ir4+ is present under OER conditions.94–96 The different electronegativity of Co compared to Ir may influence the oxidation states and electron density distribution within one or more of the OER steps and/or intermediates.2,14,90–96 Substituents influence local electron density and affect the OER steps as supported by our previous study that modelled the effect of Ni-substituted into IrO2 and found Ni within specific sites resulted in electron density accumulation within bridging oxygens, lowering the activation energy of the rate-determining OH bond breaking step.14 Prior DFT calculations support that cobalt doping within IrO2 modifies the electronic structure of the active site and lowers the OER activation energy.25 A previous study reported Co incorporation within SrIrO3 increased the coverage of surface hydroxyl groups, modified the Ir–O bond covalency, altered the oxygen p-band center of the material, and increased OER activity.30 We confirmed that cobalt remains within the structure after exposure to OER conditions following durability testing (Fig. S9†); however, additional analysis is needed to determine the specific nature of Co within the surface region after exposure to electrochemical potentials and how Co within IrO2 affects the electron density and OER kinetics.
In addition to direct comparison of the OER activity of CoIr-CLEO with IrO2-EO from our group's tests,14 the OER mass activities of CoIr-CLEO were compared with previously reported Co and Ir based catalysts in Table S6.† A direct comparison is difficult due to differences in the experimental conditions such as the type of electrolyte, electrochemical method (i.e. linear sweep voltammetry or chronoamperometry), catalyst loading, potential of the analysis, ohmic drop correction, background subtraction etc., which can influence the final activity values.14 The CoIr-CLEO nanoframes showed one of the highest OER activities compared to similar CoIr-based materials reported to date only lower than recently reported IrCo nanowires.3 The overpotential of CoIr-CLEO nanoframes was also lower than that of previously reported IrCo oxide materials.27,28
The Tafel slopes of CoIr-CLEO and IrO2,EO (Fig. 8B) were analyzed to provide insight regarding the reaction mechanism.97,98 The Tafel slope of CoIr-CLEO (40 ± 4 mV dec−1) is similar to the Tafel slope of IrO2,EO (44 ± 3 mV dec−1), and consistent with the value of others that report a similar IrOx(OH)y surface structure.99 The similar Tafel slopes of CoIr-CLEO and IrO2-EO suggest that similar rate-determining step is controlling the reaction mechanism on both catalysts. Previous studies on OER mechanisms have proposed different electrochemical pathways including the “electrochemical oxide path” and “DFT-predicted peroxide path”, differentiated by the formation of dissimilar surface intermediate species and recombination steps.100–103 The electrochemical oxide path is described by eqn (1)–(3) below,
H2O + M → M–OH + H+ + e− | (1) |
M–OH → M–O + H+ + e− | (2) |
2M–O → 2M + O2 | (3) |
M–O + H2O → MOOH + H+ + e− | (4) |
MOOH → M + O2 + H+ + e− | (5) |
Within both reaction mechanisms (i.e. electrochemical oxide path and DFT-predicted peroxide path), considering the second proton–electron transfer step as the rate limiting reaction step results in a predicted Tafel slope of 40 mV dec−1.101 The measured values of the Tafel slopes of CoIr-CLEO and IrO2-EO suggest that the OER is limited by the second electron transfer step (eqn (2)) which involves formation of Ir–O species and that the reaction may proceed via either the electrochemical oxide path or DFT-predicted peroxide path. It is important to note that the analysis of the reaction mechanism using the Tafel slope includes the assumption that the relative surface coverage of the adsorbed species is constant.100 Results obtained by microkinetic analysis98 have demonstrated that the evolution of the concentration of intermediates must be considered; therefore, more analysis needs to be done to corroborate the actual mechanism.
The comparison of the iridium mass-normalized current, Tafel slopes, OER mass activities, and OER specific activities of CoIr-CLEO and IrO2-EO (after EO) and CoIr-CLADT and IrO2-ADT (after ADT) are presented in Fig. 9A–D. As shown in Fig. 9A, both CoIr-CLEO and IrO2-EO resulted in lower currents after the ADT protocol, which is consistent with prior reports of IrO2 that showed catalyst performance degradation occurs under similar testing conditions.36 The OER mass activity of the commercial IrO2-EO catalyst at 1.51 VRHE decreased from 13.9 to 11.7 A gIr−1 indicating a retention of 84 ± 7% of the initial mass-normalized current, and the OER mass activity of the CoIr-CLEO catalyst decreased from 243 to 192 A gIr−1 indicating a retention of 79 ± 9% of the initial mass-normalized current (Fig. 9C). After the ADT protocol, the specific activity of IrO2-ADT was reduced by ∼7% from 0.045 ± 0.005 mA cmIrO2−2 to 0.042 ± 0.005 mA cmIrO2−2, while the specific activity of CoIr-CLADT slightly increased (∼4%) from 0.80 ± 0.02 mA cmIrO2−2 to 0.83 ± 0.08 mA cmIrO2−2, (Fig. 9D). The analysis of the Tafel slopes of CoIr-CLEO and IrO2-EO before and after ADT (Fig. 9B) suggests that the reaction mechanism and possibly the nature of active sites on the surface remain similar. As explained above, it is likely that the second electron transfer remains the rate determining step before and after ADT; however, more studies are needed to support the specific reaction mechanism.
To gain insight into the electrochemical stability, understand the intrinsic relationship between mass and specific activity within CoIr-CLADT and IrO2-ADT, and evaluate the relative contributions of different factors involved in the degradation process, we further compared the surface structure, morphology, and iridium and cobalt dissolution over the ADT protocol. The mass and specific activity are related, and the relationship can be described by eqn (6),105
![]() | (6) |
To understand factors contributing to the increase in specific activity for CoIr-CLADT and the decrease in specific activity for IrO2-ADT, we first analyzed the evolution of the surface of the CoIr-CL and IrO2 catalysts before and after ADT by CV (Fig. S8†). In general, the CVs of the catalysts after EO and after ADT showed qualitatively similar profiles, suggesting that the surface chemical environment remained unchanged. However, the decrease in the coulombic charge after ADT is associated to the loss of electrochemically active surface sites, either associated to dissolution of Ir or particle agglomeration. Changes in the morphology and agglomeration were demonstrated by microscopy after ADT (Fig. S9†), and the iridium dissolution rate was determined by analyzing the concentration of Ir within the electrolyte solution after EO and ADT protocols, as discussed below.
The similarities of the Tafel slopes (Fig. 9B) and cyclic voltammograms (Fig. S8†) obtained before and after ADT discredits the idea that the changes on the nature of active sites or the reaction mechanism are primarily responsible for modifying the specific activity of CoIr-CL. Upon further examination although qualitatively the electroactive species observed within the CV seem to be the same, the relative charge of the electroactive species was determined to be different after ADT. Considering that the oxidation peak, Ao, is intimately related to the presence of highly active sites of Ir3+–OH species, which have been taken as experimental descriptors toward increasing specific activity,90 the changes of the number of these surface species are an important factor. The fitting of the CV curves allowed the quantification of the relative charge associated to the different surface species (i.e. Ir3+–OH species,2 and Ir3+ to Ir4+ species within an oxide structure87,88). The analysis of the relative ratio of coulombic charge of oxidation peaks Ao/Bo decreased from 0.12 to 0.097 for IrO2-ADT but increased from 0.05 to 0.085 for CoIr-CLADT (Fig. S8†). The reduction of highly active Ir3+–OH species may be responsible for the reduction of activity of IrO2-ADT, and the increase of Ir3+–OH species may be responsible for improving of the specific activity of CoIr-CLADT. Additional analysis is needed to consider other factors (e.g. conductivity, etc.).
The OER mass activity is intimately related to the OER specific activity, but also to the number of iridium active sites available for the reaction, as described in eqn (6). Therefore, changes of specific activity and/or ECSA, caused by Ir dissolution or changes of morphology among other factors will affect mass activity. The OER mass activity of the commercial IrO2-EO catalyst at 1.51 VRHE decreased from 13.9 to 11.7 A gIr−1 (Fig. 9A and C). The average decrease of mass activity in IrO2-ADT can be explained by the combined contribution of lower specific activity (∼7%) and partial dissolution of Ir as discussed below.
After the ADT protocol, the OER mass activity of the CoIr-CLEO catalyst at 1.51 VRHE decreased from 243 to 192 A gIr−1 resulting in a relative stability of 79 ± 9%, which is within experimental error of the value of the commercial IrO2 catalyst. However, at higher potential of 1.55 VRHE (Table S5†), the mass activity retention of CoIr-CLEO was 94 ± 2%. In contrast the IrO2-EO catalyst showed a relative retention of mass activity at 1.55 VRHE which was lower (66 ± 9%) but within experimental error of the value at 1.51 VRHE (77 ± 7%) (Table S5†). The mass transport issues become more relevant at higher current densities due to the inability to remove reaction products fast enough.38 The improved retention of mass activity at higher voltages for the CoIr-CLEO catalyst compared with commercial IrO2 may also represent an important advantage in practical electrolyzer cells that operate at higher current densities and voltages. The post-mortem SEM analysis of the CoIr-CLADT (Fig. S9†) showed the presence of cobalt, the retention of 2D-structure, and the existence of additional morphologies that included small nanoparticles. We consider that the ability of CoIr-CLEO to maintain high activity after durability testing may result from the combination of (i) retention of the high specific surface activity by the presence of highly active surface functional groups and (ii) the partial retention of the integrated nanostructured morphology.
Multiple studies have shown that iridium dissolution occurs under the highly oxidative potentials and highly acidic conditions used for acidic oxygen evolution.19,20,83 The Ir dissolution rate is also affected by the presence of non-noble metals (i.e. Co, Ni etc.) within the structure as well as the experimental processes used during the electrochemical conditioning and testing of the electrode.3,12,14 To gain more insight into the degradation processes, we analyzed the Ir dissolution rates of IrO2 and Ir and Co cobalt dissolution rates of CoIr-CL, both after EO and after ADT using ICP-MS analysis, and the results are summarized in Table 3. Significant differences between the dissolution rates are observed between IrO2 and CoIr-CL and between the EO and ADT steps. For IrO2, the EO protocol resulted in an Ir dissolution rate of 2.4 pgIr cm−2 s−1 and barely contributed to the dissolution of Ir from IrO2 giving only 1.1 ± 0.1 wt%.14 This result is in line with the constant CV profile observed before and after EO for commercial IrO2 (data not shown). The ADT protocol resulted in a higher Ir dissolution rate of 27.8 ± 1.6 pgIr cm−2 s−1 for IrO2-ADT. The dissolution of Ir has been shown to be potential-dependent,83 and the different Ir dissolution rates of IrO2 for EO and ADT steps may result from the different protocols that involve either potential sweeps (EO step) or constant potentials (ADT).
IrO2 | CoIr-CL | ||||
---|---|---|---|---|---|
Mass loading (μgIr cmgeo−2) | Ir dissolution rate (pgIr cm−2 s−1) | Mass loading (μgIr cmgeo−2) | Ir dissolution rate (pgIr cm−2 s−1) | Co dissolution rate (pgCo cm−2 s−1) | |
a Value estimated from previous report.14 | |||||
Initial | 11.6 | — | 15.3 | — | |
EO | 11.6 | 2.4a | 13.7 | 997 ± 470 | 2990 ± 250 |
ADT | 10.5 | 27.8 ± 1.6 | 11.0 | 66 ± 23 | 4 ± 3 |
For the CoIr-CLEO catalyst, the EO step involves a high rate of Ir dissolution (997 ± 470 pgIr cm−2 s−1) and Co dissolution (2990 ± 250 pgCo cm−2 s−1) into the solution (Table 3) which can be explained by the significant surface reorganization involved in the EO step that transforms the initially metallic surface to an oxide/hydroxide surface (Fig. 7A). It is possible that the Ir and Co dissolution during the EO step may contribute to the generation of highly active Ir–OH groups. A previous study reported leaching of nickel from Ir–Ni oxide thin films resulted in formation of highly active Ir–OH surface sites.12 The higher dissolution rate of Ir within CoIr-CLEO compared to IrO2-EO is consistent with our prior study of hydrous nickel–iridium oxide OER catalysts where presence of Ni increased the dissolution of iridium compared to IrO2.14 CoIr-CLADT showed a significant reduction of the iridium corrosion rate to 66 ± 23 pgIr cm−2 s−1 compared to the value after the EO step (997 ± 470 pgIr cm−2 s−1). After the significant surface reorganization and dissolution involved in the potential sweeps of the EO step of CoIr-CL, the catalyst surface may then become passivated resulting in lower dissolution of Ir and Co over the ADT step, which is supported by the comparison of the CVs for the EO step and ADT step (Fig. S8†). However, the iridium corrosion rate for CoIr-CLADT still showed a ∼2 times higher dissolution rate compared to IrO2-ADT.
The comparison of the electrochemical stabilities of CoIr-CL and IrO2 shows significantly different degradation processes occur for these materials. While from an application perspective additional work is needed to further reduce Ir dissolution rate, despite the higher iridium dissolution of CoIr-CLADT, the increase in the relative concentration of surface Ir–OH groups and the presence of Co within the sublayers resulted in an increase of the specific activity and the retention of OER mass activities compared with IrO2.
In addition to higher activity, the hydrous Co–Ir oxide nanoframes exhibited similar retention of initial OER mass activity (79 ± 9% retention) as commercial IrO2 (84 ± 7% retention) tested with rotating disk electrode measurements using an accelerated durability testing protocol. The retention of OER mass activity over accelerated durability testing of cobalt-containing Co–Ir nanoframes was dramatically higher than similarly prepared nickel-containing Ni–Ir nanoframes14 which suggests cobalt–iridium interaction may be more stable than nickel–iridium interaction within these structures. The comparison of the factors influencing the electrochemical stabilities of CoIr-CL and IrO2 under OER conditions indicates that the catalysts undergo significantly different degradation processes. The decrease in the OER mass activity of the IrO2 catalyst can be explained by the combined contribution of lower specific activity and partial dissolution of Ir. The analysis of the degradation of hydrous Co–Ir oxide nanoframes showed a higher Ir dissolution rate for the compared with IrO2; however, the specific activity of CoIr-CLEO and the relative contribution of surface iridium hydroxide groups increased which contributes to the similar OER mass stability of CoIr-CLEO compared with IrO2 under the ADT testing conditions. Our work that shows that the bimetallic cobalt–iridium 2D nanoframes obtain a significantly higher OER mass activity and similar stability compared with a commercial benchmark IrO2; however, the bimetallic Co–Ir catalyst undergoes a significantly different degradation process compared with the monometallic IrO2 catalyst. Our work furthers the understanding of factors influencing the activity and stability of bimetallic and monometallic acidic OER electrocatalysts and contributes to the design of electrocatalysts with high activity and stability.
Footnote |
† Electronic supplementary information (ESI) available: Microscopic, spectroscopic, and electrochemical analysis. See DOI: 10.1039/d0na00912a |
This journal is © The Royal Society of Chemistry 2021 |