Benchmarking Perovskite Electrocatalysts’ OER Activity as Candidate Materials for Industrial Alkaline Water Electrolysis

: The selection and evaluation of electrocatalysts as candidate materials for industrial alkaline water electrolysis is fundamental in the development of promising energy storage and sustainable fuels for future energy infrastructure. However, the oxygen evolution reaction (OER) activities of various electrocatalysts already reported in previous studies are not standardized. This work reports on the use of perovskite materials (LaFeO 3 , LaCoO 3 , LaNiO 3 , PrCoO 3 , Pr 0.8 Sr 0.2 CoO 3 , and Pr 0.8 Ba 0.2 CoO 3 ) as OER electrocatalysts for alkaline water electrolysis. A facile co-precipitation technique with subsequent thermal annealing (at 700 ◦ C in air) was performed. Industrial requirements and criteria (cost and ease of scaling up) were well-considered for the selection of the materials. The highest OER activity was observed in LaNiO 3 among the La-based perovskites, and in Pr 0.8 Sr 0.2 CoO 3 among the Pr-based perovskites. Moreover, the formation of double perovskites (Pr 0.8 Sr 0.2 CoO 3 and Pr 0.8 Ba 0.2 CoO 3 ) improved the OER activity of PrCoO 3 . This work highlights that the simple characterization and electrochemical tests performed are considered the initial step in evaluating candidate catalyst materials to be used for industrial alkaline water electrolysis.


Introduction
Hydrogen could play a huge role in energy transition and in reaching transportation and climate objectives, eliminating problems created by fossil fuels [1]. The use of hydrogen has two major roles in decarbonizing major sectors of the economy and can be broadly categorized into either: (a) as a feedstock; or (b) as an energy vector that could enable energy transition [2]. However, over 95% of current hydrogen production is still fossil fuel-based. Steam-methane reforming (SMR) is the most common method of producing hydrogen due to its cost-effectivity. Other processes, such as oil and coal gasification, are also widely used, particularly in China and Australia. Only around 4% of the global hydrogen supply is produced via electrolysis [3]. This presents the major problem of using fossil fuels, which creates a significant amount of greenhouse gas (GHG) emissions. Therefore, a shift in hydrogen production via electrolysis has an extremely large potential to reduce GHG emissions as well as provide energy storage and sustainable fuels for future energy infrastructure.

Structural and Morphological Analysis
The crystalline phases of the prepared perovskite materials by co-precipitation method with subsequent calcination at 700 • C were analyzed by XRD (Figure 1   Rietveld refinement of the experimental diffraction patterns was performed, and the results of the analysis are summarized in Table 1 (for more details, see Figure S1). In this particular case, the Rietveld refinement is useful for quantifying the amount the La 2 O 3 impurities present in the LaFeO 3 sample. The results show that around 9.8% hexagonal La 2 O 3 is present in the sample. In addition, the results from the full Rietveld refinement show the formation of phase-pure orthorhombic Pr 0.8 Sr 0.2 CoO 3 and Pr 0.8 Ba 0.2 CoO 3 . However, this is not consistent with the equimolar amounts of the A-site metals (Pr and Sr; Pr and Ba) initially added for the preparation of the materials. In this case, the theoretical composition of the double perovskites is Pr 0.5 Sr 0.5 CoO 3 and Pr 0.5 Ba 0.5 CoO 3 . To verify the Rietveld refinement results, energy-dispersive X-ray spectroscopy (EDX) analysis was performed (see Table S1). The chemical composition analysis obtained from EDX confirms that Sr or Ba are present in 20 at.% of the A-site and Pr is present in 80 at.%. The reason for the formation of non-equimolar A-site metals could be due to the incomplete precipitation of Sr and/or Ba in the initial solution. Residual Sr and/or Ba salts in the supernatant led to the lesser amount of Sr or Ba on the final materials. Optimization of the co-precipitation on these materials could be performed as an extension of this work. Moreover, the calculated unit cell parameters and unit cell volume reported in Table 1 are in good agreement with the unit cell parameters of the following standards: The characteristic particle sizes obtained from the SEM images of the materials are shown in Figure 2. The calcination temperature employed in the preparation of the materials clearly affects the resulting perovskite catalyst particle sizes. In this case, the nanosized character is lost, thus obtaining particle sizes in the micron scale range (ca. 180 to 230 nm). Additionally, it can also be observed from the SEM images that the particles of the prepared materials show isometric shape without preferential surface orientation. As mentioned earlier, the chemical compositions were evaluated by EDX analysis (see Table S1). The results show that the chemical compositions match the projected ratios of lanthanum and B-site metal cations in the starting solutions. In the XRD analysis, the formation of crystalline La 2 O 3 was observed in LaFeO 3 . In the material preparation, an equimolar amount of La and Fe was added; therefore, iron-containing species (iron oxide and/or iron hydroxide) should also be present in the LaFeO 3 materials, which was confirmed by EDX analysis. The results show that A and B metal cations are present in almost the same at.%. The Fe-containing species, which were not detected by XRD analysis, suggest that these material components could be amorphous or present in a very low amount. 3 1 μm  Statistical analysis of size distribution from 400 particles.

Electrochemical Performance
The cyclic voltammograms and the corresponding Tafel plots of the perovskite materials are shown in Figures 3 and 4, respectively. The Tafel plots are obtained by averaging the currents in the forward and backward scans of the last cycle and subtracting the ohmic losses. For the La-based perovskites, the observed trend in the OER activity (overpotential values at 1 mA cm −2 , see Figure 5) for different B-site metal cations (highest to lowest activity) is LaNiO 3 > LaCoO 3 > LaFeO 3 , confirming the previously reported results by Bockris and Otagawa [17,21]. In this previous report, they proposed a relationship between the electronic structure and the OER activity of the perovskites by using the Catalysts 2020, 10, 1387 6 of 10 bond strength of the surface oxygenated intermediates as a descriptor. They successfully displayed this correlation in a volcano plot of activity in terms of the bond strength (linear relationship with the number of d-electrons). Therefore, LaNiO 3 appears at the top of the activity volcano based on the binding energy of HO*. However, at relatively low current densities, the formation of the NiOOH/Ni(OH) 2 (Ni 3+ /Ni 2+ ) redox couple could be observed. This is due to the deterioration of the rhombohedral LaNiO 3 structure [22,23]. shown in Figure 3 and Figure 4, respectively. The Tafel plots are obtained by averaging the currents in the forward and backward scans of the last cycle and subtracting the ohmic losses. For the Labased perovskites, the observed trend in the OER activity (overpotential values at 1 mA cm −2 , see Figure 5) for different B-site metal cations (highest to lowest activity) is LaNiO3 > LaCoO3 > LaFeO3, confirming the previously reported results by Bockris and Otagawa [17,21]. In this previous report, they proposed a relationship between the electronic structure and the OER activity of the perovskites by using the bond strength of the surface oxygenated intermediates as a descriptor. They successfully displayed this correlation in a volcano plot of activity in terms of the bond strength (linear relationship with the number of d-electrons). Therefore, LaNiO3 appears at the top of the activity volcano based on the binding energy of HO*. However, at relatively low current densities, the formation of the NiOOH/Ni(OH)2 (Ni 3+ /Ni 2+ ) redox couple could be observed. This is due to the deterioration of the rhombohedral LaNiO3 structure [22,23].   The differences in the observed Tafel slope among the La-based perovskites indicate that the OER mechanism is affected by the B-site metal cations. The Tafel slope of ca. 60 mV dec −1 (LaNiO3 and LaCoO3) indicates that the rate-determining step is the chemical reaction following the oneelectron transfer reaction, while the Tafel slope ca. 120 mV dec −1 (LaFeO3) indicates that the first electron transfer reaction is the rate-determining step [24]. Singh et al. [25] reported Tafel slopes of 40 to 65 mV dec −1 for LaNiO3, while Lopez et al. [26] reported a Tafel slope of ca. 75 mV dec −1 for LaCoO3 obtained by the sol-gel method. Lastly, the high Tafel slope ca. 100 mV dec −1 obtained from LaFeO3 is comparable to the values obtained in the lanthanum-calcium ferrites synthesized by the nitrate combustion method [27]. Therefore, the Tafel slopes reported in this study for La-based perovskites are similar or comparable to those reported in previous studies, even if the synthesis technique used was different.
For the Pr-based perovskites, the incorporation of Sr or Ba into the PrCoO3 structure improved the OER activity. The trend of OER activity (overpotential values at 1 mA cm −2 ; see Figure 5) from highest to lowest is Pr0.8Sr0.2CoO3 > Pr0.8Ba0.2CoO3 > PrCoO3. The Tafel slope slightly decreased from ca. 80 mV dec −1 for PrCoO3 to 70 mV dec −1 for Pr0.8Sr0.2CoO3 and Pr0.8Ba0.2CoO3. These values agree with the reported Tafel slope of Grimaud et al. [18] for the double perovskites obtained by the conventional solid-state route. The differences in the observed Tafel slope among the La-based perovskites indicate that the OER mechanism is affected by the B-site metal cations. The Tafel slope of ca. 60 mV dec −1 (LaNiO 3 and LaCoO 3 ) indicates that the rate-determining step is the chemical reaction following the one-electron transfer reaction, while the Tafel slope ca. 120 mV dec −1 (LaFeO 3 ) indicates that the first electron transfer reaction is the rate-determining step [24]. Singh et al. [25] reported Tafel slopes of 40 to 65 mV dec −1 for LaNiO 3 , while Lopez et al. [26] reported a Tafel slope of ca. 75 mV dec −1 for LaCoO 3 obtained by the sol-gel method. Lastly, the high Tafel slope ca. 100 mV dec −1 obtained from LaFeO 3 is comparable to the values obtained in the lanthanum-calcium ferrites synthesized by the nitrate combustion method [27]. Therefore, the Tafel slopes reported in this study for La-based perovskites are similar or comparable to those reported in previous studies, even if the synthesis technique used was different.
For the Pr-based perovskites, the incorporation of Sr or Ba into the PrCoO 3 structure improved the OER activity. The trend of OER activity (overpotential values at 1 mA cm −2 ; see Figure 5) from highest to lowest is Pr 0.8 Sr 0.2 CoO 3 > Pr 0.8 Ba 0.2 CoO 3 > PrCoO 3 . The Tafel slope slightly decreased from ca. 80 mV dec −1 for PrCoO 3 to 70 mV dec −1 for Pr 0.8 Sr 0.2 CoO 3 and Pr 0.8 Ba 0.2 CoO 3 . These values agree with the reported Tafel slope of Grimaud et al. [18] for the double perovskites obtained by the conventional solid-state route.
is comparable to the values obtained in the lanthanum-calcium ferrites synthesized by the nitrate combustion method [27]. Therefore, the Tafel slopes reported in this study for La-based perovskites are similar or comparable to those reported in previous studies, even if the synthesis technique used was different.
For the Pr-based perovskites, the incorporation of Sr or Ba into the PrCoO3 structure improved the OER activity. The trend of OER activity (overpotential values at 1 mA cm −2 ; see Figure 5) from highest to lowest is Pr0.8Sr0.2CoO3 > Pr0.8Ba0.2CoO3 > PrCoO3. The Tafel slope slightly decreased from ca. 80 mV dec −1 for PrCoO3 to 70 mV dec −1 for Pr0.8Sr0.2CoO3 and Pr0.8Ba0.2CoO3. These values agree with the reported Tafel slope of Grimaud et al. [18] for the double perovskites obtained by the conventional solid-state route.  Lastly, the OER activities (overpotential values at 1 mA cm −2 , see Figure 5) of the synthesized perovskites were compared with commercial Ir and Ru nanoparticles supported on carbon, which are usually used as OER activity benchmark catalysts [28,29]. The results show that the synthesized perovskites (Pr 0.8 Sr 0.2 CoO 3 ) have a comparable OER activity to the commercial Ru/C catalysts. This is advantageous in terms of cost since the perovskite electrocatalysts presented do not require the use of noble metals (i.e., Ir and/or Ru), yet they exhibit a comparable performance to the benchmark OER catalysts, leading to an economically viable material for industrial alkaline water electrolysis.

Synthesis of Perovskite Materials
The synthesis of various perovskites (ABO 3 , A = La, Pr, Pr 0.8 Sr 0.2 , or Pr 0.8 Ba 0.2 and B = Fe, Co, or Ni) was based on the co-precipitation synthesis of LaMnO 3 reported by Feng et al. [30]. In this technique, a corresponding stoichiometric amount of A-site metal nitrates (La(NO 3 ) 2 ·6H 2 O, Pr(NO 3 ) 3 ·6H 2 O, Sr(NO 3 ) 2 , or Ba(NO 3 ) 2 ) and B-site metal nitrates (Fe(NO 3 ) 3 ·9H 2 O, Co(NO 3 ) 2 ·6H 2 O, or Ni(NO 3 ) 2 ·6H 2 O) was dissolved in deionized water. All the reagents were purchased from Sigma-Aldrich and Honeywell Fluka TM (for Co(NO 3 ) 2 ·6H 2 O). After stirring at room temperature for at least 1 h (to ensure complete dissolution), KOH solution was dropped in slowly while the solution was continuously stirred. The molar ratio used in this work was A nitrate:B nitrate:KOH = 1:1:6. After stirring for about 10 to 15 min, the formed precipitate was washed with deionized water (until at approximately neutral pH) and was collected using a centrifuge. The filtered precipitate was then dried in air at 60 • C for 3 h. Finally, to obtain the desired perovskite structures, the dried precipitate was calcined at 700 • C for 3 h in air. The samples were then ground into a fine powder to be used for further characterization.

Structural and Morphological Characterization of Perovskite Materials
The phase structure of the materials was determined by performing X-ray diffraction (XRD) analysis using Cu Kα radiation on a Panalytical X'Pert PRO MPD (Multi-Purpose Diffractometer). The crystalline phases were identified by referring to the ICDD (International Center for Diffraction Data) standards. Panalytical X'pert Highscore Plus software was used to perform Rietveld refinement to evaluate the phase composition, unit cell parameters, and crystallite sizes. The surface morphology of the catalyst and the chemical composition analysis were tested by scanning electron microscopy (SEM) FEI Quanta 400. The images obtained by SEM were examined using an image analysis tool, ImageJ, where particle size measurements were performed.

Electrode Preparation and Electrochemical Tests
The oxygen evolution activities of the prepared materials were evaluated in a standard single-compartment three-electrode cell using a rotating disk electrode (RDE) setup from Pine instruments and Autolab potentiostat following the thin-film RDE approach [31]. The catalyst ink was prepared from a suspension of 5 mg of the perovskite powders in a solution mixture consisting of 0.5 mL of Milli-Q water, 2 mL of isopropanol (C 3 H 7 OH), and 10 µL of binder solution (5 wt.% of Nafion ® ). The mixture was placed in an ultrasonication bath for 30 min. Afterwards, 10 µL of prepared ink was drop casted onto the polished glassy carbon electrode (with a geometric area of 0.196 cm 2 ) for the evaluation of electrochemical performance.
All the electrochemical measurements were performed at room temperature (25 • C), 0.1 M KOH (pH = 13) electrolyte, and 1600 rpm rotation rate. Firstly, 15 full cycle voltammograms (CVs) were recorded in the potential range 1.3 to 1.9 V (vs. RHE) at 50 mV s −1 as a pre-conditioning step of the catalyst layer. The oxygen evolution reaction activities were evaluated by recording a series of CV curves in the potential range 1.3 to 1.9 V (vs. RHE) at 5 mV s −1 . Overpotential values at 1 mA cm −2 were then compared to select the candidate materials that will be used in industrial electrode manufacturing. All measured currents were converted to current densities and potentials were corrected for the ohmic-drop measured by the electrochemical impedance spectroscopy (EIS).

Conclusions
In summary, ABO 3 perovskites (A = La, Pr, Pr 0.8 Sr 0.2 , or Pr 0.8 Ba 0.2 ; B = Fe, Co, or Ni) were prepared via the co-precipitation method. This work introduces an easy strategy to screen electrocatalysts (LaFeO 3 , LaCoO 3 , LaNiO 3 , PrCoO 3 , Pr 0.8 Sr 0.2 CoO 3 , and Pr 0.8 Ba 0.2 CoO 3 ) to be used as candidate anodic materials for industrial alkaline water electrolysis. Structural and morphological analysis followed by electrochemical characterization of the materials is sufficient to select the candidate materials for industrial electrode manufacturing. The highest OER activity was observed in LaNiO 3 among the La-based perovskites, and the highest OER activity was observed in Pr 0.8 Sr 0.2 CoO 3 among the Pr-based perovskites. Additionally, this work verifies that the formation of double perovskites (Pr 0.8 Sr 0.2 CoO 3 and Pr 0.8 Ba 0.2 CoO 3 ) improves the OER activity of PrCoO 3 .
It is important to note that the scale up of industrial electrodes using the selected catalysts as active materials could introduce other factors (such as substrates used, deposition technique, and conditions, etc.). These factors should be considered, which is out of the scope of the current work. In addition, this work could be extended by evaluating the stability of the materials, because this is one of the most important criteria in developing electrodes for industrial water electrolysis.
Overall, this work highlights that the simple characterization and electrochemical tests performed are the first and most important steps in evaluating candidate catalyst materials to be used for industrial alkaline water electrolysis.