Epitaxial oxide thin films for oxygen electrocatalysis: A tutorial review

Electrocatalysis converts electrical and chemical energies by driving a reaction at a material’s surface. Such processes offer a means to integrate renewable electrons—now at cost parity to conventional sources^1,2—into chemically storable forms. Furthermore, electrocatalysis has the potential to drive chemical conversions previously driven by thermal energy more efficiently or selectively via applied voltage.^3–7 This context urges an in-depth, predictive understanding of the electrocatalysts that drive the rate and selectivity of energy conversion/storage and electrosynthesis reactions. One example of particular importance is the oxygen evolution reaction (OER) upon water oxidation that limits the efficiency of electrolyzers for hydrogen fuel production,^8–10 catalyzed at metal oxide surfaces, as well as the oxygen reduction reaction (ORR) limiting performance in fuel cells.¹¹ 

Electrocatalysts are often fabricated as particles, dispersed in an “ink,”^12,13 or as electrodeposited films.^8,14 Such materials have a high surface area, resulting in high reaction rates, but it is challenging to compare material “intrinsic” activities normalized to the active surface area. Furthermore, these materials can have heterogeneous surface composition challenging to quantify pre- and postcatalytic testing,¹⁵ variable crystallographic phases,¹⁶ and a range of interactions with the current collector.¹⁷ In contrast, epitaxial films offer a (often atomically) smooth surface that can be quantitatively characterized by spectroscopic and microscopic techniques, uniform crystal structure, and crystallographic orientation and can be varied in thickness to control interactions with a current collector. In addition, these films offer new ways to tailor catalytic activity, such as through strain^18–21 and the formation of heterostructures.^22,23 Studies of materials in this architecture have brought new insight into understanding the catalytic activity in ways intractable in high surface area systems.²⁴ 

We here provide a review of best practices in measuring the catalytic activity of epitaxial thin films, including example cases measuring oxygen electrocatalysis (OER and ORR) on oxide surfaces. We first define catalytic activity and describe typical experimental setups for electrocatalytic measurements and the methods by which thin films are measured. We then describe types of electrochemical measurements and the information obtained from them. We include metrics for assessing activity and, importantly, highlight requirements for comparing catalytic activity in the literature. We conclude with a reflection and an outlook on the unique ways in which epitaxial films benefit the electrocatalysis community through unique avenues to tailor and understand the catalytic activity.

The activity of a catalyst reflects the rate of reaction occurring at an active site for a given energetic input. In electrocatalysis, this rate can be tracked by the current in the absence of Faradaic (involving electron transfer) side reactions. Thus, more active catalysts proceed at a higher rate (current) at a given electrochemical potential (voltage on an appropriate reference scale). Alternatively, a more active catalyst is one that lowers the energetic input (voltage) required to achieve a given rate of reaction (current).

Comparison of such activities requires a means for normalizing current to the number of catalytic sites. This site density is often approximated by the surface area of a catalyst. Considering polycrystalline surfaces, however, metal site densities can vary by a factor of ∼2 across different facets, e.g., from 8 Ru/nm² for RuO₂ (101) versus 4 Ru/nm² for RuO₂ (111) orientations.²⁵ Considering nanoparticles of different sizes, surface areas can also differ by more than an order of magnitude, e.g., 170 m²/g for 5 nm RuO₂ compared to 9 m²/g for 100 nm RuO₂ particles. Coupled with this potential uncertainty in site density are complexities associated with the surrounding environment, such as electronic transport through a percolating path of nanoparticles and ionic transport through a rough electrode/electrolyte interface with complex charge distribution. Thus, understanding intrinsic catalytic rates on particles or electrodeposited films can be challenging.

In consideration of metal electrocatalysts, such as Pt for the ORR, a wealth of the literature has studied single crystal surfaces to build a fundamental understanding of intrinsic catalytic rates,²⁶ the reaction mechanism,²⁷ and design strategies to increase activity.²⁸ This has translated to the design of high surface area particles that can support high current densities per catalyst mass (cost) for technical application.²⁹ Consideration of epitaxial oxide thin films offers the possibility for similar fundamental insight into oxygen electrocatalysis.

Electrochemical cells are systems that convert electrical energy to chemical energy (electrolytic cells) or chemical energy to electrical energy (galvanic cells). Electrolytic cells drive reactions by a difference in electrochemical potential between two electrodes. Reactions at the cathode involve electrons moving from the electrode to a reactant species, whereas the anode accepts electrons from a reactant. In measurements assessing the ability of one of these electrodes—termed a “working electrode”—to catalyze a reaction, the opposing electrode is referred to as the “counter electrode.” Applying a voltage between the working and counter electrodes will drive current flow between them and the associated electrochemical reactions; however, the distribution of this voltage between the two, and the electrochemical potential on an absolute scale, is unknown in such a setup. This two-electrode configuration thus prohibits quantification of the intrinsic activity of a material and comparison between setups or research groups.

To isolate the performance of the working electrode (in our case, the epitaxial film), a third “reference” electrode of known absolute electrochemical potential is introduced (Fig. 1). The electrochemical potential of the working electrode is then controlled relative to this known reference potential by the application of a voltage between the two in a high impedance circuit. Electric potentials can be reported relative to this reference but are also often calibrated to another, such as the reversible hydrogen electrode (⁠$2H++2e−↔H2,g$⁠). The voltage of the counter electrode will then vary such that the current at the working electrode remains balanced, controlled by a potentio/galvanostat.

A suitable counter electrode is one that is stable under the electrochemical conditions tested and of sufficient activity/active area so as not to limit the current at the working electrode, as the flow of electrons is equal and opposite at the two electrodes. Care should be taken in considering potential crosscontamination. For example, a Pt counter could exhibit oxidative dissolution at sufficiently anodic conditions, redepositing on a cathodic working electrode and artificially enhancing the measured current under ORR conditions. Readers are referred elsewhere for additional considerations in the choice of counter electrode.^30,31

Bubbling of gases in an electrochemical cell can be a means to introduce a reactant (e.g., O₂ in the ORR), fix an equilibrium potential (e.g., O₂ in the OER), remove a potential side reactant (e.g., N₂ to displace O₂ in measuring cathodic reactions other than OER), and additionally provide convection.

Different cell configurations are used depending on the way by which activity is assessed. For electrochemical cells where the metrics of catalyst performance can be detected solely by the current passed (i.e., the material is known to be selective in the number of electrons passed and the desired reaction occurring, termed Faradaic efficiency), both working and counter electrodes can reside in a single compartment, connected via an external circuit [Fig. 1(a)]. In cases where the rate of reaction or Faradaic efficiency must be quantified by the explicit detection of a reaction product, for example, by titration or spectroscopy, a two-compartment cell is employed to ensure products do not react at the counter electrode. In practice, this can be done by isolating the counter electrode with a glass frit or polymer membrane, both of which permit ionic conduction while minimizing product crossover. This isolation is also employed when products generated at the counter electrode can have a detrimental effect (physically and/or kinetically) on the working electrode, or when a distinct electrolyte or reactive environment is employed in the cathode and anode compartments. An example of such a two-compartment cell, referred to as an H-cell, is shown in Fig. 1(b).

The first step to measuring the electrocatalytic activity of an epitaxial film is to electrically contact it such that current can be passed in the electrochemical cell. The ability of charge to be transferred from the film surface to this contact can impact the current-voltage profile of the system. The contact should be ohmic, and attention paid to any interfaces (e.g., with a conductive substrate when this electrical contact is at the back of a substrate/film heterostructure) through which charge is also passed. Only the catalytic material to be evaluated, the epitaxial film, should be exposed to the electrolyte, thus the contact and any film support should be encased in a chemically and mechanically stable, electrically insulating material, such as Teflon. Contacts can be made permanent (i.e., wire bonding) or temporary (e.g., conductive tape, compression-based contacts), with temporary contacts facilitating postmortem characterization. Further, mounting the electrode on a rotating shaft can enable the separation of kinetic and transport contributions to observed current profiles by utilizing the system hydrodynamics.

One accessible means of measuring the electrocatalytic activity of an epitaxial film is affixing a wire to the front or side of a conductive film, or the back of a conductive substrate. Care should be taken that the nature of this contact be ohmic in the polarization direction of interest (with methods for this assessment in situ discussed further in Sec. III C). In examples of this approach, researchers have employed sputtered or evaporated metal contacts,^24,32,33 although conductive pastes can be used as well.^34–36 As many studies of perovskites for OER and ORR employ Nb-doped SrTiO₃ as a conductive substrate, we highlight that room temperature curable Ag paste does not form a reliable ohmic contact, which can be circumvented by the application of GaIn eutectic.^37–39 The polished surfaces of single crystals can be challenging to contact with conductive paste, circumvented by slight mechanical scratching of the surface prior to paste application.

To isolate the electrocatalytic activity of the epitaxial film, all contacting materials and the substrate should be encased in an inert, nonconductive epoxy. Typical epoxies include Loctite® 9462™,⁴⁰ 9460™,⁴¹ as well as Omegabond™ 101,^37–39 with room temperature full cure times of 1–3 days.⁴² The stability of the epoxy against the electrolyte and electrochemical conditions is critical in ensuring the exposed area of the sample remains well defined, that the surface of the sample is not contaminated during measurement (affecting activity), and that parasitic side reactions do not occur (affecting quantification of activity and selectivity). We highlight that Ag paste has distinct redox features at ∼1.1 (reduction) and ∼1.3 V versus RHE (oxidation) in 0.1M KOH,⁴³ the presence of which precludes certainty in quantifying the intrinsic activity of an epitaxial film. A schematic example of a wire bonded to the front of an epitaxial catalyst film is shown in Fig. 2(a).

Provided ohmic contact can be ensured, epitaxial films can be contacted with a conductive tape, such as copper. This offers the benefit of easy removal for postmortem characterization; however, we caution that residual adhesive on the surface may impact subsequent measurements of catalytic activity. The conductive tape contact must still be electrically isolated from the electrolyte, which can be done by covering with an inert, insulating layer such as the Kapton tape,⁴⁴ schematically shown in Fig. 2(b), or a custom sample holder.⁴⁵ 

For films or substrates sufficiently conductive or when a metallic pad is deposited on a film or substrate, custom sample holders can be designed to electrically contact the electrode by compression.^33,46,47 The inert sample holder then electrically isolates everything but a known area of the epitaxial film surface, which is exposed to the electrolyte for catalytic testing. This architecture offers the benefit of easy removal of the film for postmortem characterization, and a well-defined area of the electrode exposed to electrolyte via, e.g., an O-ring seal. One such example is indicated in Fig. 2(c) where an La_(1−x)Sr_(x)MnO₃/Nb:SrTiO₃/Ti/Au electrode is compressed against a rod.⁴⁵ 

After ensuring electrons (a reactant or product in electrochemistry) can reach the film surface unimpeded by appropriate electrical contacting, consideration should also be taken in the transport of other reactant species. In particular, in cases where the concentration of reactants in a liquid electrolyte is low (e.g., gases of low solubility, such as O₂ in the ORR and H⁺ in unbuffered media), care must be taken that measurements of catalytic activity minimize the influence of mass transport through the electrolyte. This can be minimized by forcing convection within the electrolyte, such as by bubbling with a reactant gas, stirring the electrolyte, or rotating the electrode itself. One can test for mass transport limitations by varying this forced convection and assessing if the rate of reaction is transport-dependent. For example, Fig. 3 shows a case where the rate of O₂ bubbling and stir rate was varied over a range of observed catalytic ORR rates (here measured by current) on a LaMnO₃ epitaxial film.³⁴ For absolute currents equal and greater than 0.04 mA/cm²_ox, variation of the reactant flux (O₂) impacts the measured rate, illustrating that the observed rate of reaction is influenced by mass transfer and does not reflect the intrinsic kinetics of the catalyst. Comparing absolute ORR currents lower than 0.04 mA/cm²_ox, employing forced convection, would thus minimize mass transfer limitations.

Rotation of the working electrode can in some cases allow for the correction of diffusion limitations in the reaction rate via known hydrodynamic equations. This setup is termed a rotating disk electrode (RDE), in which the impact of rotation rate on the diffusion-limited current is corrected for in the mixed kinetic-diffusion regime by Koutecky–Levich analysis.⁴⁸ In cases where epitaxial films can be grown on circular substrates,⁴⁹ they can be integrated into existing sample holders affixed to rotating shafts with brushes as electrical contacts to a potentio/galvanostat. Such sample holders can also include a secondary electrode with a well-defined collection efficiency for online product detection, termed a rotating ring disk electrode (RRDE).

For the square substrates often grown upon a circular area can be exposed to the electrolyte by a machined holder^45,46 or inert tape.⁴⁴ We caution, however, that robust correction of diffusion-limited currents may not apply for holders protruding >1 μm above the film surface.⁵⁰ However, commercial rotors can operate at high enough rates (1000+ rpm) that sufficient convection can be achieved in many cases that such corrections may be unnecessary at reasonable catalytic rates.

Caution should be taken in gas-evolving reactions that occur in a liquid electrolyte, such as the OER, where microbubbles can accumulate on the film surface. The presence of such bubbles renders a portion of the surface inactive, artificially reducing the measured activity. Bubbles tend to congregate at the interface between the electrode (where generated) and the hydrophobic inert material typically encasing the electrode contact (e.g., epoxy, Teflon® sample holder). With RDE setups, high rotation rates can typically dislodge such microbubbles during the experiment. With stationary electrode configurations, stirring the solution may help with bubble removal.

Multiple electrochemical techniques are employed to probe the performance of epitaxial films. Here, we highlight those typically used in the literature for assessing resistance to charge transport, capacitance, reduction-oxidation processes, and steady-state rates of reaction. We refer the reader to foundational texts for a more in-depth discussion of their application.^51,52

At the core of electrochemistry is measurement of the current-voltage profile. Together, this defines the power consumed in an electrochemical process. When quantifying the intrinsic electrocatalytic activity of a material, steps are taken to isolate the applied voltage (on a known reference scale) at the electrode surface. In doing so, the voltage drop between the working electrode and reference electrode (dependent on the distance between the two, electrolyte conductivity, and current passed) is subtracted from the applied voltage (Fig. 4). This “iR” drop is calculated by obtaining the resistance (R) from the high frequency intercept of electrochemical impedance spectroscopy (EIS).

EIS is a perturbative characterization of the dynamics of an electrochemical process. During EIS, the input signal (either voltage or current) is applied as a sinusoidal function of varying frequency, resulting in a sinusoidal response. The changes in amplitude and shift in phase angle can be used to estimate the impedance and capacitance of the electrochemical system by fitting the response to a model circuit.^53–55 Potentiostatic-EIS, PEIS, modulates the voltage by a small amount (typically 5–10 mV) such that a linear current response results. Such a measurement can determine the resistance, R, that is characteristic of the high frequency response. This resistance arises from the transport loss between the working electrode and reference along with that across the thin film or thin film/substrate assembly. Resistance measurements by PEIS are often performed around the open circuit potential (OCP, at which no net current passes between the working and counter electrode), but should be checked under operational potentials as well, as residual resistances of heterostructures can depend on the applied potential. PEIS can also be used to probe charge transfer behavior around redox features and during Faradaic processes as well, in addition to measuring the flat band potential via changes in the capacitance in photoelectrodes.^36,56,57

Cyclic voltammetry (CV) is often used to quantify the activity of epitaxial films via the current (proportional to reaction rate) obtained at a given applied potential (the driving force of the reaction). This transient technique sweeps the voltage linearly at a characteristic “scan rate,” with an example shown in Fig. 4. As such, additional processes besides the exponential electrocatalytic reaction (orange in Fig. 4) can contribute, including the reduction/oxidation or “redox” of components within the electrode (green in Fig. 4) and the “charging current” associated with the capacitance of the surface (blue in Fig. 4). Redox features can be useful in understanding the electrode material, for instance, the ease of a metal site’s oxidation and exposure of these sites at the catalyst surface,^33,38,58 but care should be taken not to include such processes in the quantification of the catalyst activity. Similarly, capacitance can help assess the exposed surface area of the electrode, as discussed further below, but does not contribute to the rate of electrolytic reaction. Capacitance can sometimes be corrected for by subtracting a background CV when the reactant (e.g., dissolved O₂) can be eliminated or by averaging the forward and reverse sweep (dashed line in Fig. 4). Alternatively, measurements of electrocatalyst activity generally use a low scan rate (e.g., 5 mV/s) to assess activity, as catalytic current is generally independent of scan rate whereas redox processes and capacitance result in increased current with scan rate, discussed in more detail in Sec. III E. We note that the initial CV sweep may differ from subsequent ones and be unrepresentative of steady-state behavior, for example, due to surface restructuring or oxidation of organic contaminants. Carbonate, resulting from such an oxidation or other ambient exposure, may persist on the surface with a site-blocking effect.^39,59

Assessment of redox features by CV can help understand processes occurring on the surfaces of epitaxial films. For example, the Ni redox feature shown in Fig. 4 shifts more cathodically with Sr incorporation in La_(1−x)Sr_(x)NiO₃, indicating that Ni becomes easier to oxidize.⁶⁰ When LaNiO₃ films are grown at higher temperatures, the charge associated with Ni redox is reduced, shown by x-ray photoelectron spectroscopy to result from increased termination with LaO_x (instead of NiO_x) and leading to reduced OER activity [Figs. 5(a) and 5(b)].³³ In RuO₂, redox features are associated with the adsorption of species such as O and OH, confirmed by ab initio calculations⁶¹ and surface x-ray scattering.^61,62 The shifts of these features with pH shed light on the free energies of adsorbed intermediates, and reduction in redox charge originates from site blocking by phosphate ions in neutral pH [Fig. 5(c)]. The charge associated with these Ru oxidation features correlates with the number of undercoordinated Ru on the surface for different film orientations [Fig. 5(d)], as well as OER activity, providing insight into the active site for catalysis.^38,58,62 Unexpected redox features can also be indicative of contamination, such as from silver paint.

Measurement of steady-state behavior can eliminate the possible convolution of activity with transient effects that may be present in CV measurements. In a potentiostatic approach, a constant electric potential is applied to the working electrode and the current measured as a function of time, termed chronoamperometry (CA), and shown as an example for RuO₂ films of different orientations in Fig. 6,⁶³ with electrode stability measurements discussed further in Sec. IV D. Chronopotentiometry (CP) in contrast is a galvanostatic technique that draws a constant current and measures the voltage over time. CA has the benefit of maintaining a constant driving force of the electrochemical reaction as well as any corrosion processes that might occur, simplifying the interpretation of any loss in activity that might occur over time. However, CP draws some parallels to device performance (perhaps less relevant in the fundamental assessment of intrinsic activity in epitaxial films) by maintaining a constant rate of reaction. We note that transient processes like capacitance and metal redox can still be present in such measurements, but typically at timescales less than a few minutes (Fig. 6),^51,64 thus excluded by considering the steady-state behavior.

Measurement of “intrinsic activity” of an electrocatalyst—dependent on the material itself rather than external factors like the size of an electrode—and comparison of such values across institutions requires

1. Certainty of the applied voltage, relative to a known reference, at the surface of the electrocatalyst, correcting for the voltage drop as discussed earlier.

2. Certainty of the area of the electrocatalyst that is “active” and exposed to the electrolyte, by which activity is normalized.

3. Isolation of the current attributed to the rate of the desired reaction, such as by steady-state measurements, in cases where the Faradaic efficiency is 100% (meaning no parasitic electrochemical processes occur).

We next discuss corrections to the applied voltage and then approaches to determine the active area of epitaxial thin films (Fig. 7).

As potentials in a three-electrode cell are applied relative to a reference electrode, care should be taken that this reference potential is well-defined and reproducible. Common aqueous references including Ag/AgCl and Hg/Hg₂Cl₂ (saturated calomel electrode, SCE) in acidic media, and Hg/HgO in alkaline have tabulated values versus the standard hydrogen electrode (SHE) used in electrochemistry. Tabulated values depend on the concentration of ions within the reference electrode, but when used in suitable electrolytes, their potentials are stable (<1 mV deviation) for days, if not weeks, of continuous use. For ORR and OER, however, the reversible hydrogen electrode (RHE) is typically employed as it has the same pH dependence as the thermodynamic potential for these reactions, unlike, e.g., Ag/AgCl which is pH-independent in reference. Thus, quoting the potentials of epitaxial films for ORR/OER against the RHE scale is useful for other researchers to replicate the same experiments as this defines the kinetic overpotential and negates the need for replicating electrolyte pH or the reference electrode salt concentration.

A reference electrode can be calibrated to the RHE in the electrolyte used for catalytic testing by measuring the relative potential at which H₂ is evolved and oxidized on a Pt wire (an excellent catalyst for these reactions). This homemade RHE is constructed with a clean Pt wire in electrolyte purged with H₂; care must be taken in its appropriate construction.⁶⁵ The open circuit potential of such a setup (typically measured by multiple CV sweeps to ensure a clean electrode and well-saturated electrolyte) is the potential of the RHE relative to the employed reference (e.g., Ag/AgCl).

The applied voltage in a three-electrode cell is measured between a known reference and the working electrode in a high impedance circuit. However, voltage drops can occur at points of electrical contact, in the travel of current through resistive components, and the travel of charge through ionic media (i.e., between the working and reference electrode). Figure 7(a) schematically shows how the applied voltage between a reference and the working electrode increases to achieve a given current when the system resistance increases, such as by greater electrode separation or lower electrolyte concentration. Correcting this loss (typically called iR correction) is important to reflect the true kinetic overpotentials exhibited by the epitaxial films. Contact resistance is measured by performing PEIS, typically at OCP conditions.⁸ The resultant PEIS trend can be fitted with an equivalent circuit model to yield a high frequency intercept representing contact and electrolyte resistance. The applied voltage at the working electrode is then corrected for the resultant voltage drop via Ohms law (V-iR).

Ex situ, film conductivity can be assessed by methods like four-point probe. However, the application of a voltage in electrochemistry can lead to dramatically different material properties under catalytic conditions, as can interaction with the electrolyte. Charge transfer under reaction conditions can be assessed using a facile redox couple, such as ferro-ferricyanide [Fe(CN)₆]^3−/4−, Fig. 8, without the complications of sluggish reaction kinetics for multistep reactions (such as OER). Such facile redox couples do not specifically adsorb to the surface and yield a pair of redox peaks whose separation (ideally 59 mV at room temperature) is proportional to the charge transfer resistance. In semiconducting films, these redox peaks can show hysteresis as they spread farther from the reversible potential, where this asymmetry can be used to assess space charge barriers.^34,37 Such an assessment is particularly important when the substrate is used for electrical contacting or when studying multilayers.⁶⁶ In films that become insulating, the redox peaks can spread dramatically,⁴² indicating that the potential applied to the epitaxial thin film is not uniform across the surface due to resistive losses. Measurements of the “catalytic activity” of such films are thus convoluted with challenges in charge transport.

The geometric area of an epitaxial film is the exposed area of the film surface. This may not necessarily be the “active” area, for example, if a surface has high roughness, this true surface area will not be captured by a projected image. If the area exposed to the electrolyte is defined by, e.g., insulating epoxy, this area must be calculated for each individual film, such as by using an image processing software like imagej.⁶⁷ Hence, when comparing activity (first obtained in terms of raw current) across a set of epitaxial films, geometric area normalization is an important step for eliminating area effects in exhibited performance. Figure 7(b) shows schematically how increases in the electrode surface area, either by geometric (projected) area or roughness, can increase the measured raw current—both capacitive and catalytic.

Certain epitaxial films may have more exposed area than the geometric area due to the surface roughness of the film. Activity for such films can be normalized using electrochemically active surface area (ECSA).^8,68 ECSA is a measure of the non-Faradaic surface contributions of an epitaxial film by CV. Performing CV sweeps around non-Faradaic regions at different scan rates would result in different surface charge accumulation (or charging current), directly proportional to sweep speed (Fig. 9).⁸ We refer the reader to the literature for a more detailed discussion on proper voltage window selection and collection settings in the measurement of this so-called double-layer capacitance,⁶⁹ and caution this capacitance can include contributions from the electrode and electrolyte. The slope yields the capacitance as shown as an example for a La_0.6Sr_0.4MnO₃ (001) film in Fig. 9,⁴⁹ which has a specific capacitance (per geometric area of a low-roughness film) of ∼77 μF/cm². For comparably smooth (RMS roughness <1 nm) LaNiO₃ (001) films, the specific capacitance was ∼68 μF/cm².⁷⁰ We caution, however, that the tip diameter for atomic force microscopy (AFM) is notably larger than that of the molecules and ions making up the electrolyte double layer. Regardless, observations of increased capacitance with cycling likely correspond to an increase in film roughness,⁷¹ and may be used for normalization in comparing intrinsic activity.

The testing strategies for epitaxial films align toward the final goal of evaluating activity, stability, and selectivity in electrocatalytic reactions. We next review typical metrics for comparison and refer the reader to the literature for a more in-depth discussion of electrocatalytic processes,^11,72 benchmarking strategies,^8,30 and summaries of state-of-the-art performance.^73,74

Electrocatalytic reactions exhibit an exponential shape, described by the Nernst equation. The onset potential, where the rate of reaction takes off, describes the energetic input required to overcome the highest activation barrier of the reaction pathway, or the rate limiting step.^75,76 Additional energetic input then results in a higher rate of reaction. While the exponential tail can be influenced by factors such as mass transport, resistive losses, or bubble formation, the onset potential (and its associated low currents/rates of reaction) is typically less sensitive to such externalities. Typical metrics for describing onset potential include the voltage at which the reaction reaches, e.g., 100 μA/cm² (Fig. 10),⁷⁷ normalized by the electrode surface area and accounting for any roughness. Ideally, this current would be extracted from steady-state measurements (CA, CP) that exclude capacitive and redox current. If extracting the onset potential from CVs uncorrected for capacitance or with redox features present, care should be taken that the chosen current is above these values. The onset potential is often reported not relative to the experimental reference electrode, but relative to the thermodynamic potential of a given reaction. This is termed an “overpotential,” η, reflecting the thermodynamic and kinetic losses on a given catalyst.

We note that literature considering high surface area particle dispersions, meshes, and electrodeposited films will often consider the potential required to reach a much higher value of 10 mA/cm²_geo, normalized the projected area of the surface.^8,78 This metric is an applied one, defined by the solar flux and relevant for photoelectrocatalysis, but generally inappropriate for the atomically smooth surfaces of epitaxial films. Considering typical loadings (0.05 mg_ox/cm² or greater) and oxide nanoparticle surface areas (e.g., 120 m²/g_ox for 6 nm RuO₂),¹³ this corresponds to intrinsic current densities of ∼170 μA/cm²_ox. An intrinsic current density of 50 μA/cm²_ox is typically employed in the literature^12,24 to assess catalytic activity and is appropriate for the comparison of epitaxial films as well.

Catalysts can also be compared by the rate of reaction at a given energetic input (electrochemical potential). Care should be taken that the voltage is appropriately referenced and corrected for a resistive drop between electrodes, as discussed earlier. In cases where the products are detected directly (e.g., an O₂ sensor for OER), this is reported as a rate of product evolved per catalyst surface. When this rate is normalized to a number of active sites, such as the number of exposed metals on the surface obtained from redox features in CV,¹⁴ it is called a turnover frequency (TOF).

When the products are not directly detected, comparisons of the current at a fixed potential can be used to compare the rate of reaction. We note that this assumes catalysts have a comparable Faradaic efficiency, or selectivity to a given product, without the occurrence of parasitic reactions. Care should be taken that the current is free of contributions from capacitance and redox processes for robust comparison between catalysts.

For oxygen electrocatalysis, the 4 e^– process results in appreciable overpotentials—additional energetic input beyond a tabulated thermodynamic requirement (1.23 V versus RHE for O₂/H₂O). We suggest comparing the intrinsic activity at an overpotential of 0.33 V for OER (1.56 V versus RHE) and 0.4 V for ORR (0.83 V versus RHE), given the performance of typical oxides.^11,24

The exponential nature of catalytic activity can be captured in a logarithmic representation of current versus voltage in a so-called Tafel plot, yielding a linear slope at high overpotential compared to the thermodynamic potential,^79,80 shown for an example case in Fig. 10(b). Such a presentation is also a convenient way to compare materials measured at different current/voltage combinations or that differ widely in activity, as well as well as for steady-state data obtained from chronoamperometry or chronopotentiometry. In some cases, estimation of the Tafel slope can help indicate the rate determining step of a multistep reaction mechanism,^81,82 but it can also change as a function of potential as different equilibrium coverages result in different rate limiting steps.³⁸ We caution, however, that Tafel slopes are sensitive to any voltage drop remaining in the system (requiring particular care in multijunction heterostructures) and inherently assume all measured current originates from the desired catalytic reactions.

The stability of an electrode can be assessed by changes in the electrochemical performance, surface roughness/crystallinity, or dissolved electrode components⁸³ (Fig. 6)⁶³ measured over time. Stability is dependent on both electrolyte pH and electrochemical potential and can be initially considered in the context of Pourbaix diagrams. Electrochemical stability assessment primarily involves either performing chronoamperometric measurements over predefined durations of time, such as 2 h,⁸ or multiple cycles during cyclic voltammetry experiments, usually around 100.⁷⁷ Decreases in measured current at a given overpotential could indicate catalyst deactivation or loss in surface area, whereas increases in measured current could indicate transformation to a more active phase and/or increasing surface roughness. Figure 10(c) compares the voltage required to achieve 100 μA/cm²_film as a function of the CV cycle number for different coverages of Ba_0.5Sr_0.5Co_0.8Fe_0.2O₃ on the La_0.8Sr_0.2MnO₃/Nb:SrTiO₃ surface. The decrease in voltage for 0% LSMO indicates catalyst activation with cycling and the constant voltage for other films indicates stable cycling.⁷⁷ 

We caution that the low inherent roughness (and resultant surface area) of epitaxial films can lead to challenges in assessing the Faradaic efficiency of reactions. For example, a 1 cm² film in 40 ml of electrolyte would need to pass 50 μA/cm²_ox for ∼90 s before reaching the detection threshold for commercial O₂ sensors (0.3 μM),⁸ requiring hours to detect small parasitic losses. However, metal dissolution can be detected at lower concentrations (ppb) via mass spectrometry, with 1% current-to-dissolution loss detected in under 10 min for a similar setup. Higher current densities (e.g., 200 μA/cm²_ox) can be used for such stability assessments, where the comparison between CV and CA data can help assess if bubble accumulation in OER is reducing active surface area.

We conclude with a brief outlook on recent insights from epitaxial films bringing a new understanding of electrocatalytic reactions. These opportunities in building understanding primarily stem from the ability to both define the catalyst surface via growth and probe it via characterization of the catalytic surface (Fig. 11).

The atomically flat nature of most epitaxial films aids in the quantification of sites by spectroscopic techniques. For example, the Mn content on the surface of La_0.6Sr_0.4MnO₃ was found to correlate with the OER activity.⁴⁹ Layer-by-layer growth methods or chemical/thermal treatment can also probe the intrinsic activity of a given termination or identification of a specific active site in multicomponent oxides.³³ Furthermore, the ability to grow conformal layers in epitaxial heterostructures has highlighted the ability to tailor the electronic structure and resultant activity of the surface in OER.^22,23

The study of epitaxial films can also aid in understanding the role of surface structure—or its restructuring—during electrocatalytic reactions. Recent studies have considered the role of crystallographic orientation on OER, with notable differences in the activity for both perovskite^{32,39,84–86} and rutile crystal structures.^25,38,58 The ability to grow a film of known thickness and image this uniform layer cross-sectionally by transmission electron microscopy has helped identify the formation of amorphous regions, as well as regions maintaining crystallinity, on the surface after cycling.⁸⁷ The (dis)order of the surface can also be probed by diffraction- and scattering-based approaches to identify the local structure⁷¹ and the amount of amorphization.^71,88 Surface roughness can be considered postcycling by AFM.^38,71

Thickness-dependent effects in thin films can probe the extent of charge transfer with a buried layer/substrate and the electrolyte, as well as strain and its relaxation. For example, charge transfer occurs between LaMnO₃³⁴ and an Nb:SrTiO₃ substrate at thicknesses on the order of a few unit cells, reducing ORR activity. A similar effect has been leveraged to increase the activity of SrTiO₃ for OER via a buried SrRuO₃ layer.²³ Effects of strain can also be considered by the growth on different substrates,²¹ and compressive strain has been found to increase the OER activity of perovskite nickelates via changes in the electronic structure.^18,19

Here, we have reviewed best practices in the measurement of the electrocatalytic activity of epitaxial films, contextualized in the study of oxides for oxygen electrocatalysis. We have addressed approaches to electrically contact these films and measure the activity experimentally, highlighting typical techniques, data analysis, and potential pitfalls. The activity can be measured by the rate of product formation or proportional current in many cases and, additionally, compared by the energetic input required. Such measurements of epitaxial films can bring new insight into material property-performance relationships and the mechanisms of electrocatalysis, as well as precision and accuracy in the measurement of the intrinsic activity of a catalyst surface.

This material is based upon work supported by the National Science Foundation under Grant No. 2041153. Acknowledgment is made to the donors of The American Chemical Society Petroleum Research Fund for partial support of this work. K.A.S. acknowledges support from Oregon State University as a Callahan Faculty Scholar. P.A. acknowledges support from the Link Foundation Energy Fellowship.

Data sharing is not applicable to this article as no new data were created or analyzed in this study.