Influence of an external electric field on the catalytic CO oxidation of nanoscaled CeO₂ and its La³⁺ and Gd³⁺ congeners

Introduction

Catalytic CO oxidation is a straightforward approach to reduce CO levels. A variety of catalysts have been synthesized and applied to improve the efficiency of CO oxidation [1]. The potential applicability of CeO₂ in this reaction as well as in photocatalysis, gas sensing, solar cells, electrocatalysis, and as ultraviolet absorbent has been investigated extensively [2]. CeO₂ crystallizes in the fluorite crystal structure in which each Ce⁴⁺ is coordinated by eight oxygen anions in a face-centered cubic (fcc) lattice [3]. Characteristically, ceria often shows a mixed valence situation in which both Ce⁴⁺ and Ce³⁺ exist [4]. Although there are numerous works that have used this mixed valence situation in CeO₂ to influence catalytic properties, current studies still lack investigations of an external electric field to enhance CO oxidation in nanoscaled CeO₂ (nCeO₂) in the absence of any precious metals. An external electric field can modify a material’s crystal structure and lattice parameters; for example, it drives charge and mass separation and transport and influences heat transfer occurring in the bulk and at surfaces of dielectrics, piezoelectrics, and electrostrictive materials [5-10]. The electric-field-assisted catalytic decomposition over Pt/CeO₂ enables the conversion of methanethiol to H₂S even at low temperatures. An external electric field enhances the formation rates of CH₄ and H₂S [11]. Applying a +10 V potential to Au and Pd nanocatalysts increases the CO oxidation rate by 28% and 14%, respectively [12,13]. Under positive bias, Pt (18–54%) and Pd (14–42%) exhibited markedly higher activity than under negative bias (Pt: 4–8% and Pd: 4–12%) [14,15]. The catalytic decomposition of ammonia over organoboron nanoparticles was also examined under an external electric potential. Applying positive or negative potential allowed the decomposition rate to be tuned between +26% and −37% relative to the zero-bias condition [16], indicating that the sign of the applied potential also affects the overall reaction rate. Very recently, the influence of an electric field on the oxidation of 1,2-dichloroethane on CeO₂ and doped WO_x/CeO₂ has proven that a coupled mechanism of electronic conduction and lattice oxygen migration can be accelerated applying an external electrical field [17].

The oxygen deficiency of CeO₂ can be manipulated in a controlled manner by changing temperature and reduction conditions. Also elemental doping of the underlying fcc solid-state structure may create nonstoichiometric species, for example, Ce_1−xM_xO_2−δ (M = metal) [3,16]. Under these conditions oxygen vacancies (V_O) are formed. The incorporated dopant elements thus allow one to modify the coordination environment in the CeO₂ lattice by creating oxygen vacancies, decreasing the energy of V_O formation, and improving charge migration [18-20]. When an V_O is created in doped ceria, one of the unpaired electrons from the vacant oxygen site fills the hole left behind, and the remaining second electron reduces a tetravalent cerium atom. This contrasts with the V_O formation in the undoped ceria surface where two Ce⁴⁺ atoms are reduced. Doping ceria with various cations influences this general mechanism and changes the energy of the defects and their accessibility, as well as migration energy and migration pathways of mobile ions within the structure [18,21]. With respect to gas adsorption on the surface of ceria this situation is crucial and should vary between undoped and doped nCeO₂.

For the first time, we report herein on catalytic studies of CO oxidation on nCeO₂ and doped-nCeO₂ under an external low-power electric DC potential. Our results show that Gd³⁺- and La³⁺-doped nCeO₂ influence CO oxidation significantly, but these dopants behave differently when coupling with such low-power electrical fields. In situ FTIR studies were used to follow and monitor the influence of the electrical field by measuring the change in the final gas composition over the headspace of the sample, giving substantial information about the CO oxidation capability on solid nCeO₂ under an external electric field.

Results and Discussion

Characterization

A series of nCeO₂, containing Gd³⁺ and La³⁺ (2.5% and 5%), was synthesized by a precipitation method (see Supporting Information File 1). XRD patterns are in accord with JCPDS of 34-0394 and prove the fcc structure (Supporting Information File 1, Figures S5–S9) [22]. To investigate the effect of Gd³⁺ and La³⁺ on the doped nCeO₂, full width at half maximum (β), the values of interplanar spacing (d), crystallite size (D, Scherrer value), dislocation density (δ value), and number of particles per unit surface area (N value) were calculated based on three different measurements; these values are summarized in Tables S3–S5 (Supporting Information File 1). According to the values of δ and ε, doping of Gd³⁺ and La³⁺ causes a higher defect density in the fcc lattice of nCeO₂ as compared to the undoped nCeO₂. The structural integrity of the materials after the CO oxidation process was proven spectroscopically (Supporting Information File 1, Figures S10–S14).

In Raman spectra, the characteristic peak corresponding to the F_2g mode at ≈470 cm⁻¹ is indicative of the symmetry of the lattice (O_h, Supporting Information File 1, Figures S15–S19). The fcc structure is thus confirmed again for all materials. No signals for other phases or impurities are observed [23]. The characteristic peaks related to defects, extrinsic V_O, and intrinsic V_O, appeared around 250, 540, and 600 cm⁻¹, respectively [23,24]. The ceria structure remains stable after the oxidation reaction depicting all characteristic Raman modes; however, additional peaks are observed at wavenumbers higher than 1000 cm⁻¹, indicative either of physisorption or chemisorption of gas molecules on the catalyst surface (Supporting Information File 1, Figures S20–S24).

All cerium oxide nanoparticles are strongly aggregated and form dense strongly corrugated films with an agglomerated particle size of about 50 nm in scanning electron microscopy (SEM) (Supporting Information File 1, Figure S25). Transmission electron microscopy (TEM) and high-resolution transmission electron microscopy (HRTEM) reveal highly crystalline spherical, slightly faceted nCeO₂ and 5% doped-nCeO₂ with a diameter of around 15 nm (Figure 1a–c). The observed d-spacings are characteristic of nCeO₂ (0.31 nm), Gd³⁺–nCeO₂ (0.31 nm), and La³⁺–nCeO₂ (0.29 nm) (Figure 1d–f). This is in accord with the selected-area electron diffraction (SAED) patterns in which pronounced Debye–Scherrer diffraction rings can be assigned to (111), (200), (220), and (311) reflections of cubic CeO₂ (Figure 1g–i).

The Brunauer–Emmett–Teller theory (BET) surface areas of the samples are determined in the range from 134 to 148 m²·g⁻¹ (Supporting Information File 1, Table S6).

CO oxidation on nCeO₂ and doped nCeO₂ without and with applied electrical field

Figure 2 illustrates CO consumption (I_CO) and CO₂ production () for undoped and doped nCeO₂ catalysts under field-free conditions with an applied potential of 50 V after 30, 60, 90, and 120 min. The increasing circle size indicates the progression of reaction time from 30 to 120 min. The arrows indicate the direction of change in I_CO and values upon application of the electric field. I_CO attributes to the CO to CO₂ conversion and to CO adsorption on the sample (light circles). nCeO₂ showed I_CO of 1.40 g⁻¹ and the doping process increased the value up to 2.00 g⁻¹. I_CO shows a plateau after 60 min from the beginning of the reaction for all different catalysts, under applied potential conditions as well as without electric potential.

Although applying electric potential (50 V) onto nCeO₂ and 5%Gd³⁺–nCeO₂ caused no significant change in I_CO, increased substantially (Figure 2 top panel). After 120 min, increased from 1.32 to 3.11 g⁻¹ for nCeO₂ and from 3.19 to 5.57 g⁻¹ for 5%Gd³⁺–nCeO₂, clearly indicating that the electric field promotes CO oxidation, particularly for 5%Gd³⁺–nCeO₂. It thus reveals an increase for CO₂ production of 170% compared to field-free oxidation conditions. For 2.5%Gd³⁺–nCeO₂, a higher CO uptake was again observed under 50 V; however, CO₂ production remained approximately constant ( varied from 1.72 to 1.92 g⁻¹) (Figure 2 lower panel). Interestingly and in contrast to the Gd³⁺-doped samples, applying the same electric potential affects the La³⁺-doped nCeO₂ samples unexpectedly differently. There was no change in I_CO under the same electric field for the La³⁺-doped samples (2.5%La³⁺–nCeO₂: from 2.26 to 2.05 g⁻¹ and 5%La³⁺–nCeO₂: from 2.15 to 2.07 g⁻¹). La³⁺-doped samples showed only small changes in CO₂ production under field-free and under applied-field conditions. of 2.5%La³⁺–nCeO₂ and 5%La³⁺–nCeO₂ varied from 2.84 g⁻¹ to 2.30, 1.97, and 1.72 g⁻¹ under field-free vs applied-field conditions, respectively (Figure 2 lower panel and Supporting Information File 1, Tables S7–S12 and Figures S26–S29).

Without applying potential, trivalent rare-earth doping with La³⁺ and Gd³⁺ increases the V_O concentration and is expected to increase the population of reduced defect states associated with charge compensation [25,26]. Consistent with this, the doped samples show higher CO₂ production than undoped nCeO₂ (Figure 3).

External electrical bias has been reported to influence catalytic reactions on solid surfaces by modifying interfacial charge distribution and coupled electronic and ionic defect processes, particularly near surfaces and electrode/oxide interfaces [16,27-30]. In ceria, these processes are closely connected to the Ce⁴⁺/Ce³⁺ redox couple and V_O defect chemistry. A small polaron in nCeO₂ can be described as a localized electron associated with a Ce³⁺ 4f state, and electronic charge transport is commonly discussed in terms of small-polaron hopping between neighbouring cerium sites (Figure 4, left panel) [31-34]. In parallel, oxygen vacancies may rearrange within the fcc ceria lattice under suitable conditions (Figure 4, middle panel) [35]. In the present work, these processes are used as a phenomenological framework to interpret the observed bias-dependent CO oxidation response (Figure 4, right panel).

A pronounced synergistic enhancement was also observed for 5%Gd³⁺–nCeO₂, whereas the La³⁺-doped nCeO₂ sample showed no measurable bias-induced improvement in catalytic activity. This dopant-dependent response can be rationalized in terms of the different lattice strain and defect interactions introduced by Gd³⁺ and La³⁺. Regarding Ce⁴⁺ (0.97 Å), Gd³⁺ (1.053 Å) is closer in ionic radius than La³⁺ (1.16 Å); it is therefore expected to perturb the fluorite lattice less strongly. Consistent with this interpretation, the δ values extracted for 5%La³⁺–nCeO₂ from the (111), (200), (220), and (311) reflections are higher than those obtained for nCeO₂ and 5%Gd³⁺–nCeO₂ (Figure 5), indicating stronger lattice distortion.

Applying an external potential to bare nCeO₂ nearly doubled the apparent CO oxidation efficiency compared with the field-free condition (Figure 6a,b).

The different behaviour of Gd³⁺- and La³⁺-doped nCeO₂ may also be related to their different interactions with oxygen vacancies. Gd³⁺ incorporation is expected to favour the formation of oxygen vacancies while maintaining a comparatively less distorted fluorite lattice. In contrast, the larger La³⁺ ions can promote the formation of more stable defect associates, such as [La³⁺--La³⁺] and [La³⁺--Ce³⁺], which may reduce the ability of oxygen vacancies to rearrange within the lattice or participate in near-surface redox processes [20,36].

Under electrical bias, polarization and charge redistribution at grain surfaces may therefore affect the near-surface Ce³⁺/Ce⁴⁺ population and the availability of oxygen-vacancy-related defect sites involved in CO oxidation. On this basis, the observed CO oxidation behaviour with and without electrical bias suggests that local changes in oxygen-vacancy distribution and associated Ce³⁺/Ce⁴⁺ redox states may modify the near-surface ionic and electronic environment. The pronounced response of nCeO₂ and 5%Gd³⁺–nCeO₂ to the applied potential is consistent with possible contributions from small-polaron-related charge transport and oxygen-vacancy rearrangement near the surface. In contrast, the weaker response of 2.5%Gd³⁺–nCeO₂ suggests that the dopant concentration may be insufficient to generate a strong coupled defect/electronic response, while the negligible bias-induced improvement observed for La³⁺–nCeO₂ is consistent with a reduced ability of oxygen vacancies to participate in field-assisted surface redox changes, possibly due to the formation of stable La-associated defect clusters (Figure 7).

We emphasize that this mechanistic interpretation remains descriptive and phenomenological. The present data support a correlation between dopant-dependent defect chemistry, lattice distortion, and bias-assisted CO oxidation behaviour, but they do not directly quantify oxygen-vacancy migration or small-polaron hopping during the reaction. Further systematic studies, such as voltage-dependent measurements, operando spectroscopy, oxygen-isotope exchange, conductivity measurements, or theoretical calculations, would be required to verify the microscopic contributions of these processes. Nevertheless, the present results indicate that applying an electrical bias to reducible oxide catalysts can provide a promising strategy for influencing heterogeneous oxidation reactions, even in the absence of precious metal active phases.

Conclusion

An external electric field can substantially boost room-temperature CO oxidation on nanoscaled ceria by tuning its interfacial defect/redox chemistry. It induces interfacial polarization in ceria based catalysts and boosts CO oxidation under noble-metal-free conditions. The magnitude of the enhancement depends critically on dopant-controlled defect interactions, ion mobility, and electronic polaron transport mechanisms. Undoped nCeO₂ shows the largest activity and thus response to the field, with nearly a twofold increase in apparent CO oxidation efficiency under bias, while 5%Gd³⁺–nCeO₂ exhibits a clear synergistic enhancement consistent with sufficient defect mobility and interfacial responsiveness. In contrast, 2.5%Gd³⁺–nCeO₂ appears insufficient for a strong coupled polaron vacancy response, and La³⁺-doped samples show little to no bias-induced improvement, consistent with stronger defect association/cluster formation that suppresses vacancy participation in sustained field-induced surface redox changes.

Experimental

CO to CO₂ oxidation experiments

The custom-built gas adsorption device consists of sample cell, volume cell, circulation pump, and IR measurement cell, which are interconnected by stainless-steel tubes (6 mm), valves (Swagelok^® SS-6P4T-MM-8K), and Swagelok^® fittings [37]. The process was conducted under ambient conditions and was controlled by a K-type thermocouple (OMEGA Engineering Inc., NiCr, ±1.1 °C) inside the reactor. The outer and inner electrodes of the sample cell were isolated from each other by tubes (PEEK, 6 mm outer diameter, 1.65 mm inner diameter; 3.6 mm outer diameter, and 1.7 mm inner diameter). A Sorensen^® XEL 250 DC Power Supply was used to apply the potential to the sample, where the thermocouple inside and the steel mantle of the sample chamber were used as the inner electrode and the outer electrode, respectively. The measurement cell, composed of a steel tube and sealed by two stainless steel caps (0.9 cm window) using PEEK and two Viton^® O-rings each, was placed inside a FTIR spectrometer (Thermo Fisher^®-Nicolet iS5 with ID1 Transmission Accessory; Software Omnic 9.2.106). Each cap held 1.5 mm thick quartz lenses. The FTIR spectrometer was used to record the gas mixture spectra. A circulation pump (GK-M-12/02) and a pressure transducer (PAA33X-V-30 (0–30 bar), accuracy of 0.05% and resolution of 0.002%) were purchased from Gardner Denver Thomas GmbH, Germany and Omega Newport electronics GmbH, Germany, respectively.

The FTIR spectra were recorded from 400 to 4000 cm⁻¹, considering integration in the ranges from 2143 to 2231 cm⁻¹ and from 2283 to 2389 cm⁻¹, related to CO and CO₂ signals, respectively. The integration values for the cell under vacuum and for each measurement were calculated by OMNIC software. The schematic setup of the device is shown in Figure 8.

Oxidation reaction process

The system consisted of three valve-separated compartments, namely, an O₂ chamber (≈17 cm³), a CO chamber (≈30 cm³), and a sample chamber (≈18 cm³). These compartments are indicated by different colours in Figure 8. This configuration allowed the gases to be introduced, isolated, mixed, and exposed to the catalyst in a controlled manner. The pressure in the system was monitored using a pressure transducer, whose output signal was read with a multimeter.

The pressed catalyst pellet was placed in the sample chamber at the end of the cell. The catalyst amounts are shown in Supporting Information File 1, Table S3. Before gas dosing, the sample chamber was heated to approximately 120 °C under vacuum for 15 h to remove weakly adsorbed species, including adsorbed gases, water, and possible residual ammonia, from the sample surface and the cell environment. After this pre-treatment, the system was kept under vacuum and cooled to room temperature before gas dosing. During the vacuum pre-treatment and before gas dosing, the pressure reading was close to zero within the resolution of the pressure-monitoring system. The vacuum spectrum was recorded in this step.

The CO-containing gas mixture, CO/N₂ = 0.1/99.9, Air Liquide, Germany, was first introduced into the CO chamber up to a pressure of 1.5 bar. After an equilibration period of 15 min and reaching a nearly constant pressure signal, the CO chamber (purple region in Figure 8) was isolated by closing the valve between the CO and O₂ chambers. The O₂ chamber was then evacuated for 10 min to remove residual gas before O₂ was introduced up to a pressure of 1.5 bar. The O₂ chamber is shown as the red region in Figure 8. After a further equilibration period of 15 min, the reference CO spectrum was recorded before opening the valve between the CO and O₂ chambers, while the two gases were still kept in separate compartments.

Subsequently, the valve between the CO and O₂ chambers was opened, and the gases were mixed using a circulation pump. After 15 min of mixing/equilibration, the spectrum of the CO/O₂ gas mixture was recorded. The upstream and downstream valves of the sample holder were then opened to expose the catalyst pellet to the gas mixture (the green region in Figure 8). Due to expansion, the pressure of the gas mixture decreased to around 1 bar. Spectra were recorded after 10, 30, 60, 90, and 120 min of contact between the gas mixture and the catalyst.

For the experiments performed under an external electric field, the same gas dosing, mixing, and equilibration procedure was followed. A DC voltage of 50 V was applied using a potentiostat in a two-electrode configuration, with the inner and outer electrodes of the cell serving as the two electrodes. The applied voltage (50 V) generated a DC electric field across the catalyst-containing region. The voltage was applied before opening the upstream and downstream valves of the sample holder and was maintained throughout the full reaction period of 120 min. Spectra were recorded after 10, 30, 60, 90, and 120 min of exposure to the CO/O₂ gas mixture.

The circulation pump was switched off during spectral acquisition. After the final spectrum was collected, the system was evacuated again. A process-flow diagram summarizing the evacuation, gas dosing, equilibration, CO/O₂ mixing, and exposure of the gas mixture to the catalyst pellet is provided in Supporting Information File 1. For details about the calculation of the gas compositions, see Supporting Information File 1.