High-performance AEM unitized regenerative fuel cell using Pt-pyrochlore as bifunctional oxygen electrocatalyst

Significance

Unitized regenerative fuel cells (URFCs) combine the functionalities of both electrolyzers and fuel cells in a single cell/stack, enabling lower weight, volume, and capital cost. A fixed-gas URFC (FG-URFC) has the advantage of efficient and easy gas management. A highly active and highly stable bifunctional oxygen electrocatalyst with low bifunctionality index (BI) helps in achieving high performance and high round-trip efficiencies (RTE) in a URFC. In this context, Pt-Pb₂Ru₂O_7-x electrocatalyst is synthesized and evaluated as bifunctional oxygen electrocatalyst in alkaline medium. This electrocatalyst exhibits a low BI with highly symmetric oxygen reduction and evolution activity. The FG-URFC using Pt-Pb₂Ru₂O_7-x shows high RTE for an anion exchange membrane FG-URFC, providing a pathway for the development of FG-URFCs for bidirectional energy and fuel production.

Abstract

The performance of fixed-gas unitized regenerative fuel cells (FG-URFCs) are limited by the bifunctional activity of the oxygen electrocatalyst. It is essential to have a good bifunctional oxygen electrocatalyst which can exhibit high activity for oxygen evolution reaction (OER) and oxygen reduction reaction (ORR). In this regard, Pt-Pb₂Ru₂O_7-x is synthesized by depositing Pt on Pb₂Ru₂O_7-x wherein Pt individually exhibits high ORR while Pb₂Ru₂O_7-x shows high OER and moderate ORR activity. Pt-Pb₂Ru₂O_7-x exhibits higher OER (η_@10mAcm-2 = 0.25 ± 0.01 V) and ORR (η_@-3mAcm-2 = -0.31 ± 0.02 V) activity in comparison to benchmark OER (IrO₂, η_@10mAcm-2 = 0.35 ± 0.02 V) and ORR (Pt/C, η_@-3mAcm-2 = -0.33 ± 0.02 V) electrocatalysts, respectively. Pt-Pb₂Ru₂O_7-x shows a lower bifunctionality index (η_{@10mAcm-2, OER} ⁻ η_{@-3mAcm-2, ORR}) of 0.56 V with more symmetric OER–ORR activity profile than both Pt (>1.0 V) and Pb₂Ru₂O_7-x (0.69 V) making it more useful for the AEM (anion exchange membrane) URFC or metal-air battery applications. FG-URFC tested with Pt-Pb₂Ru₂O_7-x and Pt/C as bifunctional oxygen electrocatalyst and bifunctional hydrogen electrocatalyst, respectively, yields a mass-specific current density of 715 ± 11 A/g_cat^-1 at 1.8 V and 56 ± 2 A/g_cat^-1 at 0.9 V under electrolyzer mode and fuel-cell mode, respectively. The FG-URFC shows a round-trip efficiency of 75% at 0.1 A/cm⁻², underlying improvement in AEM FG-URFC electrocatalyst design.

Energy storage has gained increased attention for flexible electrical grid operation as conventional constant and variable-energy sources converge on the electrical grid. In this context, hydrogen has emerged as a clean energy carrier/source (energy density: 120 to 142 MJ/kg), which is produced via water splitting in an electrolyzer using external power source. The generated hydrogen is used as a fuel in a fuel cell (FC) to convert chemical energy into electrical energy. Combining FC and electrolyzer in a single device, known as a regenerative FC (RFC), offers certain advantages over conventional rechargeable batteries such as fast start-up/shut down, low self-discharge, low environmental effect, high energy density, and long duration/lifetime. To date, RFCs have been extensively used for unmanned underwater vehicles, high-altitude, long-duration aircraft, off-grid power storage, and emergency power generators (1). Unitized RFCs (URFCs) are the modification of RFCs, which offer the same benefits as RFCs with less weight, less volume, and less capital cost, as a single cell (FC + electrolyzer) in URFC does not require auxiliary equipment for an additional cell stack used in the RFCs (2–4). Despite the wide application of proton exchange membrane–URFCs (PEM-URFCs) for their high power density, anion exchange membrane (AEM)–URFCs have gained attention due to their cost-effectiveness as use of costly noble metals with high loading can be avoided (3, 4). Typically the URFCs have two different configurations: 1) fixed-gas (FG-URFC) in which oxygen evolution reaction (OER) and oxygen reduction reaction (ORR) happen at one electrode (oxygen electrode) whereas hydrogen evolution reaction (HER) and hydrogen oxidation reaction (HOR) occur on the other electrode (hydrogen electrode) under water-electrolyzer (WE) and FC mode, respectively, and 2) fixed-polarity (FP-URFC) in which OER (WE mode) and HOR (FC mode) occur at the same electrode, whereas HER (WE mode) and ORR (FC mode) occur at the other (5). The FG-URFCs offer the advantage of efficient and easy gas management over FP-URFCs. However, FG-URFCs suffer from sluggish oxygen electrode reactions (ORR during FC mode and OER during WE mode) on the same side of the cell resulting in a convergence of inefficiencies which act as a bottleneck for the wide application of AEM FG-URFCs. Therefore, the development of highly active ORR/OER bifunctional electrocatalyst is necessary for AEM FG-URFCs. An ideal bifunctional oxygen electrocatalyst should show converging OER and ORR onset potential toward an equilibrium potential of 1.23 V versus RHE with a low bifunctionality index (BI = difference between OER potential at 10 mA/cm²_geo and ORR potential at −3.0 mA/cm²_geo). However, benchmark OER (RuO₂ and IrO₂) and ORR (Pt) electrocatalyst exhibit a BI greater than 1.0 V making them unsuitable individually for use in URFC (6, 7). Hence, they are used along with other electrocatalyst which can reduce their asymmetric electrocatalytic behavior [e.g., Pt-IrO2-(RuO₂-TiO₂ {RTO}) (8), Pt-IrO₂ (9, 10), and Pt-Ru-Ir (11)]. The use of AEM-URFC enables expansion of the material space of electrocatalysts as a wide spectrum of materials is stable in alkaline medium.

In this context, an OER electrocatalyst which can also act as a support, in lieu of benchmark ORR electrocatalyst Pt, could be a good candidate as a bifunctional oxygen electrocatalyst. The support-material for Pt plays a critical role as it should provide high surface area, high OER activity, high OER–ORR stability, high electronic charge transport, efficient catalyst dispersion, and help in facet engineering (12–15). The most commonly used catalyst supports are Vulcan XC-72 (16), TiC (17, 18), RTO (8, 12, 19), Sb-doped SnO₂ (ATO) (20), doped-TiO₂ (21, 22), metal (13, 14), and Ti_nO_2n-1 (23), which offer moderate-to-high conductivity and moderate surface area but offer no OER activity of note, thereby effectively ruling them out. In this regard, use of alkaline OER-active and OER/ORR-stable electrocatalyst as a support for Pt could be a solution. Among them, lead ruthenate pyrochlore (Pb₂Ru₂O_7-x) has shown excellent OER activity–stability (24–26) and moderate ORR activity–stability in alkaline medium (24).

In this perspective, we have deposited Pt on Pb₂Ru₂O_7-x, which shows a very low BI with highly symmetric OER–ORR activity profile, making it useful for the alkaline-URFC as well as metal-air battery applications. Pt-Pb₂Ru₂O_7-x exhibits higher OER and ORR activity in comparison to IrO₂ and Pt/C, respectively. The high OER activity is ascribed to a high Ru(V): Ru(IV) ratio in Pt-Pb₂Ru₂O_7-x, which is confirmed through X-ray photoelectron spectroscopy (XPS) study. The high ORR activity of Pt-Pb₂Ru₂O_7-x is attributed to the high dispersion of Pt on Pb₂Ru₂O_7-x support. An FG-URFC tested with Pt-Pb₂Ru₂O_7-x and Pt/C as bifunctional oxygen electrocatalyst and bifunctional hydrogen electrocatalyst, respectively, yields a mass-specific current density of 715 ± 11 A/g⁻¹ at 1.8 V and 56 ± 2 A/g_cat^-1 at 0.9 V under WE mode and FC mode, respectively. The FG-URFC shows a round-trip efficiency (RTE) of 75% at 0.1 A/cm⁻², which is the highest-reported RTE in an AEM FG-URFC to our knowledge, thereby signifying the usefulness of Pt-Pb₂Ru₂O_7-x as OER/ORR bifunctional electrocatalyst for future energy and fuel production applications.

Results and Discussion

Physical Characterization.

The particle size, surface morphology, and elemental composition and mapping are evaluated using scanning electron microscopy coupled with energy dispersive spectroscopy (SEM/EDX). The SEM images reveal that the particle size of Pb₂Ru₂O_7-x and Pt-Pb₂Ru₂O_7-x samples are in the range of 70 to 140 nm and 100 to 210 nm, respectively (Fig. 1 A and B). The increase in the particle size for Pt-Pb₂Ru₂O_7-x as compared to Pb₂Ru₂O_7-x is attributed to the Pt deposition as well as additional heating during Pt-Pb₂Ru₂O_7-x synthesis. Elemental mapping with EDX for Pb₂Ru₂O_7-x shows the presence of Pb, Ru, and O elements confirming the formation of Pb-Ru oxide system, whereas EDX elemental mapping of Pt-Pb₂Ru₂O_7-x shows the presence of Pt along with Pb, Ru, and O elements confirming a Pt-deposited Pb-Ru oxide system (Fig. 1C). The amount of Pt in bulk Pt-Pb₂Ru₂O_7-x is also determined as ∼5 wt% using EDX.

Fig. 1.

SEM images of (A) Pb₂Ru₂O_7-x and (B) Pt-Pb₂Ru₂O_7-x. EDX elemental mapping of (C) Pb₂Ru₂O_7-x and Pt-Pb₂Ru₂O_7-x.

X-Ray Diffraction (XRD) is performed on the Pb₂Ru₂O_7-x and Pt-Pb₂Ru₂O_7-x samples to identify the crystal structure (Fig. 2). The XRD peaks for Pb₂Ru₂O_7-x match with the pyrochlore phase (PDF-00–002-1365, Space-group-Fd-3m) of Pb₂Ru₂O₇ (Fig. 2) without any additional peak associated with segregated Ru- and Pb-oxide or any of their mixed oxide phase (24). The probability of the formation of solid solution or doped oxide is minimal due to high size mismatch of Pb and Ru ions (27). The Rietveld refinement on Pb₂Ru₂O_7-x XRD peaks determines a lattice constant of a = b = c = 10.325 Å, which is similar to the other literature reported values (24, 26). This structure includes structural oxygen defects originating from charge imbalance between multiple oxidation states of Pb(II/IV) and Ru(IV/V) as indicated by our previous experimental and theoretical studies (24). The XRD peaks for Pt-Pb₂Ru₂O_7-x did not show the presence of prominent additional peaks associated with metallic Pt, which may be due to the low concentration (∼5 wt%) and low crystallinity of Pt as against Pb₂Ru₂O_7-x. The Rietveld refinement on the XRD peaks of Pt-Pb₂Ru₂O_7-x shows almost no change in the lattice constant (a = b = c = 10.325 Å) confirming no lattice expansion or contraction of Pb₂Ru₂O_7-x. No shift in the peaks (pyrochlore phase) is observed for Pt-Pb₂Ru₂O_7-x as compared to Pb₂Ru₂O_7-x confirming no Pt-doping into the Pb₂Ru₂O_7-x structure (Fig. 2).

Fig. 2.

XRD of Pb₂Ru₂O_7-x and Pt-Pb₂Ru₂O_7-x samples.

The XPS is performed on both Pb₂Ru₂O_7-x and Pt-Pb₂Ru₂O_7-x to determine the surface elemental composition and oxidation states of the elements. XPS of Pb₂Ru₂O_7-x shows multiple oxidation states of Pb and Ru along with the formation of oxygen vacancies. The deconvoluted Pb 4f XPS spectrum of Pb₂Ru₂O_7-x yields four subpeaks at ∼136.2 eV, ∼137.3 eV, ∼141.1 eV, and ∼142.3 eV corresponding to Pb(II) 4f_7/2, Pb(IV) 4f_7/2, Pb(II) 4f_5/2, and Pb(IV) 4f_5/2, respectively (SI Appendix, Fig. S1A) (24, 25). The peak splitting and peak area ratio (Pb(IV)/Pb(II)) are determined as ∼4.9 eV and ∼0.18 eV, respectively. The deconvoluted Ru 3p XPS spectrum also shows four subpeaks at ∼462.5 eV, ∼464.8 eV, ∼484.8 eV, and ∼487.5 eV that correspond to Ru(IV) 3p_3/2, Ru(V) 3p_3/2, Ru(IV) 3p_1/2, and Ru(V) 3p_1/2, respectively, which is similar to the literature (SI Appendix, Fig. S1B) (24, 25, 28). The peak splitting and peak area ratio (Ru(V)/Ru(IV)) are determined as ∼22.5 eV and ∼0.80, respectively. The deconvolution of O 1s spectrum for Pb₂Ru₂O_7-x generates three subpeaks at ∼528.3 eV, ∼529.9 eV, and ∼531.0 eV corresponding to the O²⁻ state of O-atoms in crystal lattice (O_lattice), an auxiliary oxidation state of the O-atom due to the creation of oxygen vacancies (O_vac), and the surface −OH state, respectively (SI Appendix, Fig. S1C). For Pt-Pb₂Ru₂O_7-x, the deconvoluted Pb 4f XPS spectrum yields four subpeaks with peak splitting and peak area ratio of ∼5.0 eV and ∼0.18 which are similar to that of Pb₂Ru₂O_7-x, confirming no change in pyrochlore structure and metal oxidation states (SI Appendix, Fig. S2A). The result is also confirmed by minimal change in peak splitting (∼22.3 eV) and peak area ratio (∼0.78) of Ru 3p XPS spectrum (SI Appendix, Fig. S2B). The deconvoluted O 1s XPS peak of Pt-Pb₂Ru₂O_7-x shows similar subpeaks as Pb₂Ru₂O_7-x with higher O_vac/O_lattice (SI Appendix, Fig. S2C). The higher O_vac/O_lattice for Pt-Pb₂Ru₂O_7-x (1.32) as compared to Pb₂Ru₂O_7-x (0.68) is ascribed to the appearance of another peak at ∼529.5 eV associated with the adsorbed −OH on Pt (29). The deconvolution of Pt 4f XPS peak for Pt-Pb₂Ru₂O_7-x yields subpeaks mainly associated with metallic Pt (Pt(0) 4f_7/2 and Pt(0) 4f_5/2 at ∼71.5 and ∼74.9 eV, respectively) with a peak orbital splitting of ∼3.4 eV.

Electrochemical Characterization.

There is a lot of ambiguity regarding the choice of the characterization technique which can appropriately quantify the active site density or specific surface area. While Brunauer–Emmett–Teller (BET) surface area is used to calculate surface area for gas-phase heterogenous catalysis reactions, its utility for electrochemical reactions is still in open to questions. In case of electrochemical reactions, double-layer capacitance (C_DL) is taken as a surrogate for electrochemical surface area (ECSA), and it is quite effective in quantifying active surface area, especially for OER (15, 30, 31). It has been found that BET surface area and ECSA computed through C_DL has similar scaling relationship for metal oxide nanoparticles (15, 31). However, if the electrode material in question is composed of a mixture of materials, then ECSA calculation becomes erroneous due to variety of specific capacitance and dielectric behavior of different materials (24). For those cases, BET surface area is suitable to calculate specific electrochemical activity (24). For reduction reactions (e.g., ORR or HER) especially involving Pt-electrocatalyst, hydrogen adsorption/desorption, or CO-stripping is being used to quantify the active surface area (32). Hence, we have employed BET surface area to calculate specific OER activity, while both the ECSA computed from H-adsorption/desorption and BET surface area are employed to calculate specific ORR activity.

Cyclic voltammetry (CV) has been performed on both Pt/C and Pt-Pb₂Ru₂O_7-x in N₂-saturated 0.1 M of potassium hydroxide (KOH) to measure Pt-specific ECSA (SI Appendix, Fig. S3). For both electrocatalysts, peaks appear signifying hydrogen adsorption/desorption on Pt between 0 and 0.6 V versus RHE followed by the formation of double-layer at 0.6 to 0.8 V versus RHE using hydrogen underpotential deposition (H_upd) method. With increase in the potential (at 0.8 to 1.1 V versus RHE), metal surface oxidation and reduction occur during forward and reverse scan, respectively (33). The Pt-specific ECSA of Pt/C and Pt-Pb₂Ru₂O_7-x are determined according to SI Appendix, Eq. S4. Despite of the low Pt-loading of Pt-Pb₂Ru₂O_7-x (∼5 wt%) as compared to Pt/C (46.5 wt%), the ECSA of Pt-Pb₂Ru₂O_7-x (135 m²/g⁻¹_Pt) is higher than that of Pt/C (57 m²/g⁻¹_Pt). This underlines the better dispersion of Pt achieved on Pb₂Ru₂O_7-x support through our synthesis protocol as compared to Pt-dispersion in commercial Pt/C electrocatalyst. Additionally, a positive and negative shift are observed for hydrogen desorption peak and hydrogen adsorption peak for Pt-Pb₂Ru₂O_7-x as compared to Pt/C signifying facile hydrogen adsorption (facilitating ORR) (34).

The C_DL of Pt-Pb₂Ru₂O_7-x and Pb₂Ru₂O_7-x are found to be 9 × 10⁻⁷ F and 2.9 × 10⁻⁶ F, respectively. The Pb₂Ru₂O_7-x and Pt-Pb₂Ru₂O_7-x show a BET surface area of 85 m²/g and 50 m²/g, respectively. The surface areas measured from BET and C_DL follow a similar trend (i.e., Pb₂Ru₂O_7-x < Pt-Pb₂Ru₂O_7-x), which is ascribed to increase in particle size of Pb₂Ru₂O_7-x during Pt deposition process as confirmed by SEM studies (Fig. 1). The discrepancy in the ratio of surface area (BET: C_DL) computed for Pt-Pb₂Ru₂O_7-x and Pb₂Ru₂O_7-x is due to the different dielectric behavior of Pt and Pb₂Ru₂O_7-x.

Apparent and Specific Electrochemical Activity.

To investigate the applicability of the synthesized electrocatalysts toward the FG-URFC application, linear sweep voltammetry (LSV) in O₂-saturated 0.1-M KOH solution has been performed to evaluate the ORR and OER activities of the electrocatalysts in a RDE setup (Fig. 3). The OER activity is benchmarked by measuring the potential required to reach 10 mA/cm²_geo whereas the OER stability is measured by the change in the potential required to reach 10 mA/cm²_geo after holding the potential at 1.7 V versus RHE for 2 h (24, 31). Pt/C does not even reach 10 mA/cm²_geo at a voltage up to 1.9 V versus RHE as Pt is not an OER-active electrocatalyst which is widely documented in literature (12). Pb₂Ru₂O_7-x (η_10mA/cm2geo = 0.22 ± 0.01 V) shows much higher apparent OER activity as compared to IrO₂ (η_10mA/cm2geo = 0.33 ± 0.02 V) (Fig. 3 and Table 1). The high apparent OER activity can be ascribed to the high Ru(V)/Ru(IV) ratio (0.67) as confirmed from XPS (SI Appendix, Figs. S1 and S2) (24). Pt-Pb₂Ru₂O_7-x shows slightly lower apparent OER activity (η_{10 mA/cm2geo} = 0.25 ± 0.01 V) as compared to Pb₂Ru₂O_7-x, which is due to lower surface area and OER-inactive Pt deposition. Apparent OER activity trend is also measured at η_OER = 0.25 V (1.48 V versus RHE) overpotential for all the electrocatalysts, which is in the following order: Pb₂Ru₂O_7-x > Pt-Pb₂Ru₂O_7-x > IrO₂ > Pt/C. The current density for Pb₂Ru₂O_7-x at 0.25 V overpotential is 3.91-, 33.1-, and 65-fold higher than that for Pt-Pb₂Ru₂O_7-x, IrO₂ and Pt/C, respectively. The mass-specific and BET-specific OER activities of Pb₂Ru₂O_7-x, Pt-Pb₂Ru₂O_7-x, IrO₂, and Pt/C at 1.48 V versus RHE are listed in Table 1, which shows a trend of Pb₂Ru₂O_7-x > Pt-Pb₂Ru₂O_7-x > IrO₂ > Pt/C, confirming the excellent OER activity of Pb₂Ru₂O_7-x and Pt-Pb₂Ru₂O_7-x.

Fig. 3.

(A) LSV curves corresponding to OER and ORR on Pb₂Ru₂O_7-x, Pt-Pb₂Ru₂O_7-x, IrO₂, and Pt/C in 0.1 M KOH solution at 1,600 rpm. (B) OER and (C) ORR Tafel slopes for all the electrocatalysts. Catalyst loading = 200 µg/cm⁻². (D) The activation barrier involving each electron transfer during OER on Pb₂Ru₂O_6.5 (111) facet and Pt (111) facet. “S” corresponds to the active site.

Table 1.

Comparison of OER activity of all the electrocatalysts in 0.1 M KOH in an RDE setup

	η@10mAcm-2 (V)	Mass-specific current density at 1.48 V versus RHE (mA/mg−1catalyst)	Tafel slope (mV/dec)	BET-specific OER activity (mA/cm−2BET)
Pt-Pb2Ru2O7-x	0.25 ± 0.01	55.1 ± 2.4	80.6 ± 1.4	0.254 ± 0.01
Pb2Ru2O7-x	0.22 ± 0.01	215.5 ± 10.1	69.9 ± 1.1	0.110 ± 0.005
IrO2	0.35 ± 0.02	6.5 ± 0.3	108.4 ± 2.3	0.021 ± 0.001
Pt/C	—	3.3 ± 0.1	408.2 ± 4.2	0.001

The transition of Ru(V) → Ru(VIII) in Pb₂Ru₂O_7-x is far less energy intensive in comparison to Ir(IV) → Ir(VIII) in commercial IrO₂, leading to superior specific OER activity of Pb₂Ru₂O_7-x (24). The Ru-oxidation state–sensitive OER activity in Pb₂Ru₂O_7-x is explained by the stabilization of the higher oxidation state of the surface Ru atoms that yields to the lowering of the OER activation barrier leading to facile OER (35, 36). The Ru(V):Ru(IV) ratio remains almost the same in Pb₂Ru₂O_7-x (Ru(V)/Ru(IV) = 0.8) and Pt-Pb₂Ru₂O_7-x (Ru(V)/Ru(IV) = 0.78) as confirmed from XPS (SI Appendix, Figs. S1 and S2), which is responsible for their superior specific activity in comparison to commercial IrO₂.

The Tafel slope is used to understand the mechanistic aspects of the OER mechanism for synthesized Pb₂Ru₂O_7-x and Pt-Pb₂Ru₂O_7-x. The Tafel slopes for Pb₂Ru₂O_7-x and Pt-Pb₂Ru₂O_7-x are found to be ∼69.9 ± 1.1 mV/dec and ∼80.6 ± 1.4 mV/dec, which indicates that the third electron (S-O + HO⁻ → S-OOH + e-) transfer is the potential determining step (PDS) during OER (Fig. 3B) (37, 38). The exchange current density calculated from Tafel analysis results suggest faster OER kinetics for Pb₂Ru₂O_7-x (∼27.2 ± 1.3 µA/cm⁻²) as compared to Pt-Pb₂Ru₂O_7-x (∼16.5 ± 0.8 µA/cm⁻²) despite showing similar PDS, which can be ascribed to less-active sites due to OER-inactive Pt deposition. The oxygen vacancies present in Pb₂Ru₂O_7-x contribute to increase in the OER kinetics through increase in the bulk electronic charge transport (conductivity) (24). However, oxygen vacancies do not take part in PDS as have been shown in the previous studies (24).

To demonstrate the ORR/OER bifunctional activity, we have studied the ORR on all the electrocatalysts in an O₂-purged 0.1-M KOH solution using LSVs (Fig. 3). The apparent ORR activity is benchmarked by the overpotential (η_-3mA/cm2geo) required to reach −3 mA/cm²_geo (Fig. 3) (24). The apparent ORR activity is found to be in the following order: Pt-Pb₂Ru₂O_7-x > Pt/C > Pb₂Ru₂O_7-x > IrO₂ (Fig. 3 and Table 2). The higher apparent ORR activity for Pt-Pb₂Ru₂O_7-x as compared to Pt/C is due to 1) the better Pt-dispersion and subsequent higher ECSA and 2) self-supported electrocatalytic activity of Pt-Pb₂Ru₂O_7-x. The apparent ORR activity at 0.78 V versus RHE is measured as −4.34 ± 0.02, −3.84 ± 0.05, and −2.79 ± 0.12 mA/cm²_geo for Pt/C, Pt-Pb₂Ru₂O_7-x, and Pb₂Ru₂O_7-x, respectively. The apparent ORR activity is consisted of an intrinsic kinetic current from O₂ reduction on the surface of the electrocatalyst and a diffusion-limited current. The ORR activity at higher overpotentials (mass-transfer limitation) is limited by the diffusion of O₂ approaching the electrocatalyst surface. Therefore, the diffusion restrictions are decoupled from the observed current density through Koutecký–Levich analysis (SI Appendix, Eq. S2). The kinetic current density (i_k) is calculated for only Pt/C and Pt-Pb₂Ru₂O_7-x to compare the intrinsic ORR activity of the electrocatalysts as they are the highest ORR-active electrocatalysts among all. The i_k for Pt-Pb₂Ru₂O_7-x (7.61 ± 0.13 mA/cm²_geo) is found to be greater than that for Pt/C (6.61 ± 0.09 mA/cm²_geo) confirming better intrinsic ORR activity of Pt-Pb₂Ru₂O_7-x as compared to Pt/C. Though the η_-3mA/cm2geo, apparent current density, and kinetic current density are good metrics to evaluate the ORR activity, these parameters are greatly influenced by the electrocatalyst surface roughness and deviations in the mass transfer caused by the quality of the drop-casted film (39). Therefore, to compare the ORR activity, the current is measured at a fixed overpotential (in the kinetic-controlled region) in the polarization curve, followed by normalization using both BET surface area as well as Pt-specific ECSA. The BET surface area–specific ORR activity for both the electrocatalysts is determined as Pb₂Ru₂O_7-x is also ORR active (24). The BET surface area–specific ORR activity of Pt-Pb₂Ru₂O_7-x and Pt/C are determined as −0.04 ± 0.001 and −0.007 ± 0.001 mA/cm⁻²_BET at 0.78 V versus RHE, respectively. However, carbon is not ORR active, which makes comprehensive comparison of BET-specific ORR activity of the electrocatalysts difficult and leads to underestimation of specific ORR activity in Pt/C. Therefore, the specific ORR activity of both electrocatalysts is determined based on Pt-specific ECSA. The Pt-specific ECSA of Pt/C and Pt-Pb₂Ru₂O_7-x are determined as 0.8 ± 0.08 and 2.84 ± 0.14 A/m⁻²_{Pt-specific ECSA} at 0.78 V versus RHE, respectively. The results show higher (3.55-fold) Pt-specific ORR activity for Pt-Pb₂Ru₂O_7-x as compared to Pt/C, which offers an excellent ORR activity crucial for FG-URFC application. However, the reported ORR activity in Pt-Pb₂Ru₂O_7-x constitutes a combinatorial ORR activity coming out from both Pt and Pb₂Ru₂O_7-x, which may lead to overestimation of specific activity of Pt in Pt-Pb₂Ru₂O_7-x

Table 2.

Comparison of ORR activity in 0.1 M KOH of all the electrocatalysts in an RDE setup

	η@-3mAcm-2 (V)	Mass-specific current density at 0.78 V versus RHE (mA/mg−1catalyst)	Tafel slope (mV/dec)	BET-specific ORR activity at 0.78 V versus RHE (mA/cm−2BET)
Pt-Pb2Ru2O7-x	−0.31 ± 0.02	−19.2 ± 0.26	61 ± 1.6	−0.04 ± 0.001
Pb2Ru2O7-x	−0.47 ± 0.01	−13.9 ± 0.61	116 ± 3.8	−0.02 ± 0.001
IrO2	−1.03 ± 0.02	—	164 ± 2.0	—
Pt/C	−0.33 ± 0.02	−21.7 ± 0.1	71 ± 0.8	−0.007 ± 0.001

The Tafel analysis of ORR for all the samples is determined in the following order: IrO₂ > Pb₂Ru₂O_7-x > Pt/C > Pt-Pb₂Ru₂O_7-x (Table 2). The result indicates highly facile ORR for both Pt/C and Pt-Pb₂Ru₂O_7-x as compared to IrO₂ and Pb₂Ru₂O_7-x. The higher Tafel slope for Pt/C as compared to Pt-Pb₂Ru₂O_7-x indicates more facilitated ORR on Pt-Pb₂Ru₂O_7-x in comparison to Pt/C. The Tafel slope of Pb₂Ru₂O_7-x indicates that the first electron transfer step involving molecular O₂ dissociation (S + O₂ + H₂O + e⁻ → S-OOH + OH⁻) on the surface as PDS for the ORR (38). For both Pt/C and Pt-Pb₂Ru₂O_7-x, the Tafel slope value signifies second electron transfer (S-OOH + e⁻ → S-O + OH⁻) as PDS. The change in the PDS for Pt-Pb₂Ru₂O_7-x as compared to Pb₂Ru₂O_7-x is attributed to the deposition of more ORR-active Pt deposition which controls the ORR activity.

We have performed density functional theory (DFT)-based calculation (SI Appendix, section S4) on (111) facet of Pb₂Ru₂O_6.5 and Pt to calculate their OER and ORR overpotentials individually. The surface Pb and Ru atoms of Pb₂Ru₂O_6.5 (111) facet become covered by O-atoms before OER sets in Fig. 3D, which is also confirmed through operando XAS study by Kim et al. at electrochemical OER conditions (40). The OER overpotential is found to be 0.52 V considering surface Ru as an active site with third electron transfer being the PDS, which is also indicated by our previous study (24). The ORR overpotential on Pt(111) facet is found to be 0.45 V, which is close to previous literature reports (41).

Electrochemical Stability.

The OER stability of Pt-Pb₂Ru₂O_7-x is studied using potentiostatic-hold that operates at the potential of 1.7 V versus RHE for 2 h. LSVs have been performed before and after the potentiostatic-hold test, and both the OER and ORR activity of Pt-Pb₂Ru₂O_7-x are compared (Fig. 4A). Pt-Pb₂Ru₂O_7-x shows no change in OER overpotential (η_10mAcm-2,2h = 0.26 ± 0.01 V) after the OER-hold test, confirming high stability of Pt-Pb₂Ru₂O_7-x in alkaline medium under OER condition. However, it shows a 0.12-V increase in the ORR overpotential (η_-3mAcm-2,2h = −0.43 ± 0.01 V) after the OER-hold test. The increase in the ORR overpotential is due to the formation of passivated Pt-oxide (ORR-inactive) in Pt-Pb₂Ru₂O_7-x, which is not fully transformed to Pt during ORR (SI Appendix, section S3 and Fig. S4) (12, 42). Additionally, the oxgen vacancies in Pb₂Ru₂O_7-x are quenched during OER which are responsible for promoting ORR (24), thereby further lowering the activity. The results also confirm that Pb₂Ru₂O_7-x in Pt-Pb₂Ru₂O_7-x is mostly responsible for the OER activity with minimum contribution from Pt. Pt-Pb₂Ru₂O_7-x do not follow the inverse activity–stability relationship, which is commonly found for single-crystal Ru(0001) or RuO₂(110) (43, 44). Surface Ru atoms in Pb₂Ru₂O_6.5 (45), Ln₂Ru₂O₇ (Ln = lanthanide elements) (46), or doped oxides (e.g., dimensionally stable anodes) (47) do not show the traditional low-stability behavior found in single crystals of Ru(0001) or RuO₂(110) (43, 44), as the formation of RuO₄ or its hydrated form (48) is avoided in the pyrochlore system through charge-compensation of Ru atoms with their neighboring Pb-atoms (24, 25).

Fig. 4.

(A) LSV curves of Pt-Pb₂Ru₂O_7-x before and after the OER-hold test (1.7 V versus RHE for 2 h). (B) LSV curves of Pt-Pb₂Ru₂O_7-x before and after the ORR-hold test (0.5 V versus RHE for 2 h).

The ORR stability of Pt-Pb₂Ru₂O_7-x is also studied using potentiostatic-hold at the potential of 0.5 V versus RHE for 2 h. Similar to OER stability, LSVs have been performed before and after the ORR-hold test and both the OER and ORR activity of Pt-Pb₂Ru₂O_7-x are compared (Fig. 4B). Pt-Pb₂Ru₂O_7-x shows no change in both OER (η_10mAcm-2,2h = 0.26 ± 0.01 V) as well as ORR overpotential (η_-3mAcm-2,2h = −0.31 ± 0.02 V) after the ORR-hold test, confirming high stability of Pt-Pb₂Ru₂O_7-x in alkaline medium under ORR condition. However, there is a small decrease in the OER activity for Pt-Pb₂Ru₂O_7-x at higher potentials after the ORR-hold test, which may be due to the surface reduction (lower oxidation states of the metal cations) of Pb₂Ru₂O_7-x (OER-active). Similar reduction behavior and subsequent decrease in OER activity is observed for Pb₂Ru₂O_7-x samples in the literature (24). Inductively coupled plasma optical emission spectrometry (ICP-OES) does not detect any Ru in the solution after the OER-hold as well as ORR-hold tests for Pb₂Ru₂O_7-x and Pt-Pb₂Ru₂O_7-x samples, which is also found in our previous studies (detection limit = 0.1 ppm) (24).

Bifunctional Activity.

The OER/ORR bifunctional activity is benchmarked using BI. The BI of benchmark OER (IrO₂) and ORR (Pt/C) catalyst is greater than 1.0 V, which makes them incompatible for use in URFC (Table 3). Pb₂Ru₂O_7-x exhibits a BI of 0.69 V, which is similar to our previous reports (24). However, BI does not clearly represent the effect of individual OER and ORR activity. For example, the activity of Pb₂Ru₂O_7-x is skewed toward OER as compared to ORR, making its BI low. Pt-Pb₂Ru₂O_7-x shows a much-lower BI (0.56 V) with more symmetric OER–ORR activity profile as compared to Pb₂Ru₂O_7-x making it more useful for the alkaline-URFC or metal-air battery applications. Furthermore, the BI is lower than the best-reported bifunctional OER–ORR electrocatalyst (La_0.8Ca_0.2MnO₃) till date by 0.11 V (49), thereby underlining its superior performance (Table 3).

Table 3.

Activity of some of the state-of-art bifunctional catalysts reported in the literature

Catalysts	ORR(V) @ −3 mA/cm2geo	OER (V) @ 10 mA/cm2geo	BI (V)	Ref.
La0.8Ca0.2MnO3	0.89	1.56	0.67	(49)
Nd1.5Ba1.5CoFeMnO9−δ hybridized with N-doped r-GO	0.89	1.59	0.70	(57)
α-MnO2	0.76	1.72	0.96	(58)
RuO2/C	0.68	1.62	0.94	(6)
CoO	0.87	1.57	0.70	(59)
Pt/C	0.89	1.69	0.80	(57)
Pb2Ru2O7-x	0.74	1.43	0.69	(24)
IrO2	0.2	1.58	1.38	This work
Pb2Ru2O7-x	0.76	1.45	0.69	This work
Pt- Pb2Ru2O7-x	0.92	1.48	0.56	This work

While Pb₂Ru₂O_7-x show high OER activity, it shows moderate ORR activity. On the other hand, despite showing excellent ORR activity, Pt transforms to PtO₂ after 1.3 V versus RHE during OER thereby reducing its stability and activity (12). A reduction of overpotential of PDS in Pb₂Ru₂O_7-x and replacement of Pt with more OER-stable, ORR-active electrocatalyst will improve the bifunctional oxygen electrocatalytic properties required for FG-AEMFCs.

FG-URFC Performance.

After demonstrating the excellent bifunctional ORR–OER activity of Pt-Pb₂Ru₂O_7-x vis-a-vis Pb₂Ru₂O_7-x, Pt/C, and IrO₂ in a RDE setup, we have fabricated membrane electrode assemblies (MEAs) (5 cm²) using Pt-Pb₂Ru₂O_7-x (1.0 mg/cm⁻²) as anode and Pt/C (0.5 mg_Pt/cm⁻²) as cathode for FG-URFC application (Fig. 5). The performance of the FG-URFC is evaluated at 80 °C under both WE and FC mode (Fig. 5). The FG-URFC shows an excellent FC performance of ∼56 ± 2 mA/cm⁻² at 0.9 V (kinetic-controlled region) and ∼450 ± 7 mA cm⁻² at 0.5 V (mass-transfer controlled region). The FG-URFC also shows an excellent WE performance of ∼715 ± 11 mA/cm⁻² at 1.8 V. Our result shows much higher performance for both WE and FC mode as compared to the literature studies (Table 4) (50–53). A proper comparison of the URFC performances with the literature is not possible due to the difference in the MEA fabrication parameters (e.g., anode and cathode catalyst loading, separator type, binder type, and binder to catalyst ratio) operating condition (our work: 80 °C and literature: 20 to 65 °C). The mass-specific current density under WE mode is determined as 715 ± 11 A/g_cat^-1 at 1.8 V. The mass-specific and Pt-specific current density of the FG-URFC under FC mode (kinetic-controlled region) is determined as 56 ± 2 A/g_cat⁻¹ and 1,120 ± 40 A/g_Pt⁻¹ at 0.9 V. The RTE for our FG-URFC (SI Appendix) is determined as ∼84%, ∼75%, and ∼40% at 0.05, 0.1, and 0.5 A/cm⁻², respectively, which is much higher than any reported value to date (Table 4). However, after 10 consecutive URFC cycles (WE + FC), the RTE is reduced from ∼84% to ∼70% at 0.05 A/cm⁻² and from ∼∼5% to ∼57% at 0.1/A cm⁻², which may be attributed to a combination of membrane (AEM) degradation and surface oxidation of Ti gas-diffusion layer leading to electrical contact issues (SI Appendix, Fig. S4) (54). The transmission electron microscopy (TEM) images show agglomeration of Pt due to 10 consecutive URFC cycle, which is also one of the factors for the AEM-FG-URFC performance loss (SI Appendix, Fig. S5). The agglomeration is because of dissolution of Pt particles in WE mode followed by its redeposition mode in the FC mode. Therefore, our work achieves a good initial AEM-FG-URFC performance with high RTE for low PGM loading both on anode and cathode side.

Fig. 5.

FG-URFC performance (WE and FC) using Pt-Pb₂Ru₂O_7-x, Pt/C, and Fumasep FAA-3–50 as anode, cathode, and separator, respectively. The anode and cathode remain unchanged for both FC and WE mode.

Table 4.

Comparison of the alkaline FG-URFC performances

BHE	BOE	GDL	Temp (°C)	PPD (mW/cm−2)	RTE(%)	Ref.
50 wt% Pt/C (1 mg/cm−2)	Cu0.6Mn0.3Co2.1O4 (3 mg/cm−2)	Carbon paper (BHE and BOE)	40	80	31.9% @100 mA cm−2	(51)
20 wt% Pt/C (0.1 mg/cm−2)	Cu0.6Mn0.3Co2.1O4 (3 mg/cm−2)	Au-coated Ti-mesh (BHE and BOE)	22	114	34% @100 mA cm−2	(50)
Ni/C (6 mg/cm−2)	Ni/C + MnOx/GC (1:5 by mass, 4 mg/cm−2)	Carbon paper (BHE and BOE)	65	16.5	40% @10 mA cm−2	(53)
46 wt% Pt/C (0.5 mg/cm−2)	MnOx (0.3 mg/cm−2)	Carbon paper (BHE) Stainless steel (BOE)	55	27	2 to 45% @20 mA cm−2	(52)
46 wt% Pt/C (0.5 mg/cm−2)	Pt-Pb2Ru2O7-x (1.0 mg/cm−2)	Carbon paper (BHE) Ti-plate (BOE)	80	253	75% @100 mA cm−2	This work
46 wt% Pt/C (0.5 mg/cm−2)	Pt-Pb2Ru2O7-x (1.0 mg/cm−2)	Carbon paper (BHE) Ti-plate (BOE)	80	253	40% @500 mA cm−2	This work

BHE, bifunctional hydrogen electrocatalyst; BOE, bifunctional oxygen electrocatalyst; GDL, gas diffusion layer; PPD, peak power density.

Material Synthesis and Methods

Synthesis of Pb₂Ru₂O_7-x.

Pb₂Ru₂O_7-x with pyrochlore structure is synthesized using a method described in our previous studies (24, 25, 55). A 5-mmol ruthenium (III) nitrosylnitrate (Ru(NO)(NO₃)₃, Ru 31.3% minimum, Alfa Aesar) is dissolved in 25 mL of de-ionized (DI) water (18.2 MΩ cm) and stirred for 10 min which is added to a 25-mL solution containing 5-mmol lead (II) nitrate (Pb(NO₃)₂, 99.99%, Sigma-Aldrich) in DI water. Then, the mixture is stirred for additional 30 min to achieve a homogeneous solution. The total 50-mL mixture is slowly added dropwise and precipitated in a 500-mL, 4-M KOH solution in a plastic conical flask to avoid any corrosion of glass containers by the KOH solution. The precipitate is crystallized by maintaining the resultant solution at 85 °C with continuous oxygen bubbling for 5 d. The solution volume and concentration are maintained by adding DI water in every 24 h. After 5 d of crystallization, the resultant solid is separated from the solution by using centrifuge (Thermo Fisher Scientific, Heraeus Multifuge ×1) at 10,000 rpm. Then, the separated solid is thoroughly washed with DI water using the centrifuge until a pH of 7 to 8 of the supernatants is achieved. After that, the material is washed three times with glacial acetic acid followed by acetone (three times) to remove any impurity. Finally, the solid is dried at 60 °C overnight in an oven, ground to powder, and used for experiments.

Synthesis of Pt-Pb₂Ru₂O_7-x.

The deposition of Pt on Pb₂Ru₂O_7-x is performed by slightly modifying a previously reported method (56). A total 0.45 g of chloroplatinic acid hexahydrate (H₂PtCl₆, 6H₂O) (Sigma-Aldrich) is added to 180 mL of DI water under vigorous stirring. Then, 3.6 g of sodium bisulfite (NaHSO₃) is added into the solution under continuous stirring that turns the color of the solution from yellow (H₂PtCl₆) to colorless (H₃Pt(SO₃)₂OH). To dilute the solution, an additional 450 mL of DI water is added, which is followed by the dropwise addition of 0.6 M sodium carbonate (Na₂CO₃) buffer to the solution until pH = 5 is reached. After that, 135 mL hydrogen peroxide (H₂O₂) is added dropwise together with 5 wt% NaOH. The addition of H₂O₂ decreases the solution pH rapidly. Therefore, initially, H₂O₂ (2 to 3 mL) and NaOH are added slowly so that the solution pH is maintained at 5.0 to form the colloidal suspension of Pt-oxide (solution turns yellow). The reaction indicates near completion when the pH of the solution stops to decrease. Then, the rest of the H₂O₂ is added to complete the reaction. Finally, 0.56 g of Pb₂Ru₂O_7-x is added to the solution under stirring for 24 h to facilitate Pt-oxide adsorption onto Pb₂Ru₂O_7-x. The resultant black material is filtered and washed repeatedly by DI water. The washed black material is again dispersed in 100 mL DI water, stirred, and followed by the addition of 100 mL absolute ethanol (1:1 vol/vol). Then, the solution is heated at 70 °C for 2 h with stirring to reduce Pt-oxide to Pt deposited on Pb₂Ru₂O_7-x. After the solution is cooled down, the particles are filtered, washed with DI water and dried overnight at 60 °C to achieve Pt-Pb₂Ru₂O_7-x.

Fabrication of Membrane Electrode Assembly and URFC Operation.

The MEA is prepared by sandwiching an AEM (Fumasep FAA-3–50, thickness 50 µm) as separator between two gas-diffusion electrodes (GDE). The GDEs are fabricated by spraying a catalyst ink with an airbrush (Badger 150) under N₂ flow on carbon paper and Ti-sheet (to avoid corrosion) for the cathode and anode, respectively. For the cathode side, the catalyst ink containing Pt/C catalyst (Pt 46.5%, Tanaka, Japan) is prepared by sonicating 0.05 g catalyst, 3.2 g methanol/water (1/1 vol%), and 0.428 g of 5 wt % solubilized AEM binder (Fumion FAA-3, Fumatech) to achieve a catalyst to binder ratio of 70:30. For the anode side, the catalyst ink containing Pt-Pb₂Ru₂O_7-x is prepared by sonicating 0.05 g catalyst, 3.2 g methanol/water (1/1 vol %), and 0.176 g of 5 wt% solubilized AEM binder to achieve a catalyst to binder ratio of 85:15. The Pt-Pb₂Ru₂O_7-x loading is achieved as 1 mg/cm⁻² on the anode side, whereas Pt/C loading is kept as 0.5 mg_Pt cm⁻² on the cathode side. As the AEMs are received in the bromide form, all the MEAs are ion-exchanged with OH⁻ by immersing them in three batches of 1 M KOH each for 8 to 10 h for a total of 24 to 30 h followed by a thorough DI water wash to remove any extra KOH.

The FG-URFC is operated by testing the MEA under both FC and WE mode. The WE is tested by the electrolysis of ultrapure DI water at 80 °C. The DI water is heated using a hot plate/stirrer (Thermo Fisher Scientific), pumped only into the anode side, and recirculated at a flow rate of 200 mL/min⁻¹. Potential is swept anodically (stair-step protocol) from 1.25 V to 1.90 V, and the current value is recorded after a 2-min potentiostatic-hold (OER on Pt-Pb₂Ru₂O_7-x anode and HER on Pt/C cathode). Before performing the fuel-cell polarization experiments, the MEAs are conditioned by holding the cell at a constant voltage (potentiostatic-hold) of 0.55 V for 90 min to ensure stable operation. The FC polarization data are obtained at 80 °C and 90% relative humidity using O₂ as oxidant (ORR on Pt-Pb₂Ru₂O_7-x anode) and H₂ (HOR on Pt/C cathode) as fuel without any backpressure (1-atm absolute gas pressure) on both sides. The anode bipolar plate is a corrosion-resistant Ti-plate (2 × 2 mm², single parallel flow channel) to avoid carbon corrosion at high potentials during electrolyzer mode whereas the cathode is a graphite plate (1 × 1 mm², three serpentine flow channels). The gases are passed at a stoichiometric ratio of 2.0 with a minimum flow rate maintained at 100 mL/min⁻¹. The open circuit potential (OCP) is recorded, and the current is scanned from the OCP to a value in which the cell voltage drops below 0.2 V (operation is stopped), with each current density maintained for 2 min.

Conclusion

In conclusion, we have deposited Pt on Pb₂Ru₂O_7-x, which shows a record-low BI of 0.56 V with a more symmetric OER–ORR activity profile, which makes it useful for the alkaline-URFC or metal-air battery applications. Pt-Pb₂Ru₂O_7-x exhibits higher OER and ORR activity in comparison to benchmark OER (i.e., commercial IrO₂) and ORR (i.e., commercial Pt/C) electrocatalysts, respectively. The increased OER activity is ascribed to a high Ru(V): Ru(IV) ratio in Pt-Pb₂Ru₂O_7-x which is confirmed through XPS study. The high ORR activity of Pt-Pb₂Ru₂O_7-x is due to the high dispersion of Pt on Pb₂Ru₂O_7-x in Pt-Pb₂Ru₂O_7-x, which is confirmed through ECSA measurement. Furthermore, Pt-Pb₂Ru₂O_7-x shows no change in OER and ORR activity after a 2-h ORR-hold test experiment underlying its long-term stability in alkaline-ORR conditions. Additionally, Pt-Pb₂Ru₂O_7-x exhibits an unchanged OER activity after OER-hold test for 2 h. However, it shows a reduction of 0.12 V in the ORR overpotential after the OER-hold test due to the formation of passivated Pt-oxide. A FG-URFC tested with Pt-Pb₂Ru₂O_7-x and Pt/C as bifunctional oxygen electrocatalyst and bifunctional hydrogen electrocatalyst, respectively, yields a mass-specific current density of 715 ± 11 A/g⁻¹ at 1.8 V and 56 ± 2 A/g_cat^-1 at 0.9 V under electrolyzer mode and fuel-cell mode, respectively. The FG-URFC shows an RTE of 75% at 0.1 A/cm⁻², which is the highest-reported RTE in an alkaline FG-URFC till date. Therefore, Pt-Pb₂Ru₂O_7-x as OER/ORR bifunctional electrocatalyst with a low bifunctional index leads to significant improvement on electrocatalyst design and demonstrates increased viability of alkaline FG-URFCs.

Data Availability

All study data are included in the article and/or SI Appendix.